&EPA
United States
Envirofunmlal Protection
Agnncy
Policy Assessment for the Review of the
Ozone National Ambient Air Quality
Standards
-------
DISCLAIMER
This document has been reviewed by the Office of Air Quality Planning and Standards
(OAQPS), U.S. Environmental Protection Agency (EPA), and approved for publication. This
OAQPS Policy Assessment contains conclusions of the staff of the OAQPS and does not
necessarily reflect the views of the Agency. Mention of trade names or commercial products is
not intended to constitute endorsement or recommendation for use.
-------
EPA-452/R-14-006
August 2014
Policy Assessment for the Review of the Ozone National Ambient Air
Quality Standards
U.S. Environmental Protection Agency
Office of Air and Radiation
Office of Air Quality Planning and Standards
Health and Environmental Impacts Division
Ambient Standards Group
Research Triangle Park, North Carolina 27711
-------
This page left intentionally blank.
-------
TABLE OF CONTENTS
LISTS OF FIGURES v
LIST OF TABLES ix
LIST OF ACRONYMS AND ABBREVIATIONS xi
EXECUTIVE SUMMARY ES-1
1 INTRODUCTION 1-1
1.1 PURPOSE 1-1
1.2 BACKGROUND 1-3
1.2.1 Legislative Requirements 1-3
1.2.2 History of OsNAAQS Reviews 1-5
1.2.3 Current Os NAAQS Review 1-10
1.3 GENERAL APPROACH FOR REVIEW OF THE STANDARDS 1-12
1.3.1 Approach for the Primary Standard 1-13
1.3.1.1 Approach Used in the Last Review 1-14
1.3.1.2 Approach for the Current Review 1-17
1.3.1.2.1 Consideration of the Scientific Evidence 1-20
1.3.1.2.2 Consideration of Exposure and Risk Estimates 1-24
1.3.1.2.3 Considerations Regarding Ambient Os Concentration
Estimates Attributable to Background Sources 1-26
1.3.2 Approach for the Secondary Standard 1-27
1.3.2.1 Approach Used in the Last Review 1-28
1.3.2.2 Approach for the Current Review 1-33
1.3.2.2.1 Consideration of the Scientific Evidence 1-36
1.3.2.2.2 Consideration of Exposure and Risk Estimates and Air Quality
Analyses 1-39
1.3.2.2.1 Considerations Regarding Ambient Os Concentration Estimates
Attributable to Background Sources 1-41
1.3.3 Organization of this Document 1-42
1.4 REFERENCES 1-43
2 O3 MONITORING AND AIR QUALITY 2-1
2.1O3 MONITORING 2-1
2.1.1 Os Monitoring Network 2-1
2.1.2 Recent Os Monitoring Data and Trends 2-3
2.2 EMISSIONS AND ATMOSPHERIC CHEMISTRY 2-9
2.3 AIR QUALITY CONCENTRATIONS 2-10
2.4 BACKGROUND Os 2-12
2.4.1 Seasonal Mean Background Os in the U.S 2-17
2.4.2 Seasonal Mean Background Os in the U.S. as a Proportion of Total Os ..2-19
2.4.3 Daily Distributions of Background Os within the Seasonal Mean 2-20
-------
2.4.4 Proportion of Background Os in 12 Urban Case Study Areas 2-24
2.4.5 Influence of Background Os on W126 levels 2-25
2.4.6 Estimated Magnitude of Individual Components of Background Os 2-27
2.4.7 Summary 2-30
2.5 REFERENCES 2-32
3 ADEQUACY OF THE CURRENT PRIMARY STANDARD 3-1
3.1 EVIDENCE-BASED CONSIDERATIONS 3-2
3.1.1 Modes of Action 3-2
3.1.2 Nature of Effects 3-7
3.1.2.1 Respiratory Effects - Short-term Exposures 3-8
3.1.2.2Respiratory Effects -Long-term Exposures 3-36
3.1.2.3 Total Mortality - Short-term Exposures 3-45
3.1.2.4 Cardiovascular effects - Short-term Exposure 3-49
3.1.3 Adversity of Effects 3-52
3.1.4 Ozone Concentrations Associated With Health Effects 3-56
3.1.4.1 Concentrations in Controlled Human Exposure Studies and in
Epidemiologic Panel Studies 3-56
3.1.4.2 Concentrations in Epidemiologic Studies - Short-term Metrics 3-60
3.1.4.3 Concentrations in Epidemiologic Studies - "Long-term" Metrics .. 3-74
3.1.5 Public Health Implications 3-77
3.1.5.1 At-Risk Populations 3-77
3.1.5.2 Size of At-Risk Populations and Lifestages in the United States ....3-87
3.1.5.3 Averting Behavior 3-89
3.2 AIR QUALITY-, EXPOSURE-, AND RISK-BASED CONSIDERATIONS.. 3-90
3.2.1 Consideration of the Adjusted Air Quality Used in Exposure and Risk
Assessments 3-90
3.2.2 Exposure-Based Considerations 3-93
3.2.3 Risk-Based Considerations 3-103
3.2.3.1 Risk of Lung Function Decrements 3-103
3.2.3.2 Estimated Health Risks Associated with Short- or Long-Term Os
Exposures, Based on Epidemiologic Studies 3-113
3.3 CASAC ADVICE AND PUBLIC COMMENTERS' VIEWS ON THE
ADEQUACY _OF THE CURRENT STANDARD 3-124
3.4 STAFF CONCLUSIONS ON ADEQUACY OF PRIMARY STANDARD ..3-128
3.5 REFERENCES 3-137
4 CONSIDERATION OF ALTERNATIVE PRIMARY STANDARDS 4-1
4.1 INDICATOR 4-1
4.2 AVERAGING TIME 4-2
4.3 FORM 4-5
4.4 LEVEL 4-8
4.4.1 Evidence-based Considerations 4-9
4.4.2 Air Quality-, Exposure-, and Risk-Based Considerations 4-21
4.4.2.1 Exposure-Based Considerations 4-21
4.4.2.2 Risk-Based Considerations: Lung Function 4-29
ii
-------
4.4.2.3 Risk-Based Considerations: Epidemiology-Based Mortality and
Morbidity 4-36
4.5 CASAC ADVICE AND PUBLIC COMMENTERS' VIEWS ON
ALTERNATIVE STANDARDS 4-46
4.6 STAFF CONCLUSIONS ON ALTERNATIVE PRIMARY STANDARDS
FOR CONSIDERATION 4-48
4.7 KEY UNCERTAINTIES AND AREAS FOR FUTURE RESEARCH AND
DATA COLLECTION 4-70
4.8 SUMMARY OF STAFF CONCLUSIONS ON PRIMARY STANDARD.. 4-73
4.9 REFERENCES 4-76
5 ADEQUACY OF THE CURRENT SECONDARY STANDARD 5-1
5.1 NATURE OF EFFECTS AND BIOLOGICALLY RELEVANT EXPOSURE
METRIC 5-1
5.2 FOREST TREE GROWTH, PRODUCTIVITY AND CARBON
STORAGE 5-11
5.2.1 Evidence-based Considerations 5-12
5.2.2 Exposure/Risk-based Considerations 5-29
5.3 CROP YIELD LOSS 5-43
5.3.1 Evidence-based Considerations 5-43
5.3.2 Exposure/Risk-based Considerations 5-48
5.4 VISIBLE FOLIAR INJURY 5-51
5.4.1 Evidence-based Considerations 5-53
5.4.2 Exposure- and Risk-based Considerations 5-60
5.5 OTHER WELFARE EFFECTS 5-69
5.5.1 Forest Susceptibility to Insect Infestation 5-69
5.5.2 Fire Regulation 5-70
5.5.3 Ozone Effects on Climate 5-71
5.5.4 Additional Effects 5-72
5.6 CASAC ADVICE 5-73
5.7 STAFF CONCLUSIONS ON ADEQUACY OF SECONDARY
STANDARD 5-75
5.8 REFERENCES 5-89
6 CONSIDERATION OF ALTERNATIVE SECONDARY STANDARDS 6-1
6.1 INDICATOR 6-1
6.2 FORM AND AVERAGING TIME 6-2
6.3 LEVEL 6-8
6.4 CONSIDERATION OF PROTECTIVENESS OF REVISED PRIMARY
STANDARD 6-37
6.5 CASAC ADVICE 6-40
6.6 STAFF CONCLUSIONS ON ALTERNATIVE STANDARDS 6-44
6.7 SUMMARY OF CONCLUSIONS ON THE SECONDARY STANDARD 6-57
6.8 KEY UNCERTAINTIES AND AREAS FOR FUTURE RESEARCH AND
DATA COLLECTION 6-58
6.9 REFERENCES 6-62
iii
-------
APPENDICES
Appendix 2A. Supplemental Air Quality Modeling Analyses of Background Os 2A-1
Appendix 2B. Monitoring Data Analysis of Relationships Between Current Standard and W126
Metric 2B-1
Appendix 2C. Inter-annual Variability in W126 Index Values: Comparing Annual and 3-Year
Average Metrics (2008-2010) 2C-1
Appendix 3 A. Recent Studies of Respiratory-Related Emergency Department Visits and Hospital
Admissions 3A-1
Appendix 3B. Ambient Os Concentrations in Locations of Health Studies 3B-1
Appendix 5A. Os-Sensitive Plant Species Used by Some Tribes 5A-1
Appendix 5B. Class I Areas Below Current Standard And Above 15 ppm-hrs 5B-1
Appendix 5C. Expanded Evaluation of Relative Biomass and Yield Loss 5C-1
IV
-------
List of Figures
Figure 1-1. Overview of approach to reviewing the primary standard 1-19
Figure 1 -2. Overview of approach to reviewing the secondary standard 1-35
Figure 2-1. Map of U.S. ambient Os monitoring sites reporting data to EPA during the 2006-
2010 period 2-3
Figure 2-2. Trend in U.S. annual 4th highest daily maximum 8-hour Os concentrations in ppb,
2000 to 2012. Solid center line represents the median value across monitoring sites,
dashed lines represent 25th and 75th percentile values, and top/bottom lines
represent 10th and 90th percentile values 2-4
Figure 2-3. Map of 8-hour Os design values in ppb for the 2009-2011 period 2-5
Figure 2-4. Map of 8-hour Os design values in ppb for the 2010-2012 period 2-5
Figure 2-5. Trend in U.S. annual W126 concentrations in ppm-hrs, 2000 to 2012. Solid center
line represents the median value across monitoring sites, dashed lines represent
25th and 75th percentile values, and top/bottom lines represent 10th and 90th
percentile values 2-6
Figure 2-6. Map of 2009-2011 average annual W126 values in ppm-hrs 2-7
Figure 2-7. Map of 2010-2012 average annual W126 values in ppm-hrs 2-7
Figure 2-8. Trend in the May to September mean of the daily maximum 8-hour ozone
concentrations before (dotted red line) and after (solid blue line) adjusting for year-
to-year variability in meteorology 2-8
Figure 2-9. Map of 2007 CMAQ-estimated seasonal mean natural background Os levels (ppb)
from zero-out modeling 2-18
Figure 2-10. Map of 2007 CMAQ-estimated seasonal mean North American background Os
levels (ppb) from zero-out modeling 2-19
Figure 2-11. Map of 2007 CMAQ-estimated seasonal mean United States background Os levels
(ppb) from zero-out modeling 2-19
Figure 2-12. Map of site-specific ratios of U.S. background to total seasonal mean Os based on
2007 CMAQ zero-out modeling 2-22
Figure 2-13. Map of site-specific ratios of apportionment-based U.S. background to seasonal
mean Os based on 2007 CAMx source apportionment modeling 2-22
Figure 2-14. Distributions of absolute estimates of apportionment-based U.S. Background (all
site-days), binned by modeled MDA8 from the 2007 source apportionment
simulation 2-23
Figure 2-15. Distributions of the relative proportion of apportionment-based U.S. Background to
total Os (all site-days), binned by modeled MDA8 from the 2007 source
apportionment simulation 2-23
Figure 2-16. Fractional influence of background sources to W126 levels in four sample
locations. Model estimates based on 2007 CMAQ zero-out modeling 2-27
Figure 2-17. Differences in seasonal mean Os (ppb) between the NAB and NB scenarios. 2-28
Figure 2-18. Percent contribution of U.S. anthropogenic emissions to total seasonal mean
MDA8 Os levels, based on 2007 source apportionment modeling 2-30
Figure 3-1. Modes of action/possible pathways underlying the health effects resulting from
inhalation exposure to Os. (Adapted from U.S. EPA, 2013, Figure 5-8) 3-4
-------
Figure 3-2. Percent increase in respiratory-related hospital admission and emergency
department visits in studies that presented all-year and/or seasonal results 3-34
Figure 3-3. Summary of mortality risk estimates for short-term Os and all-cause
(nonaccidental) mortality 3-47
Figure 3-4. Concentration-response function for asthma hospital admissions over the
distribution of area-wide averaged Os concentrations (adapted from Silverman and
Ito, 2010) 3-66
Figure 3-5. Concentration-response function for pediatric asthma emergency department visits
over the distribution of averaged, population-weighted 8-hour Os concentrations
(reprinted from Strickland et al., 2010) 3-68
Figure 3-6. Exposure-Response relationship between risk of death from respiratory causes and
ambient Os concentration study metric (Jerrett et al., 2009) 3-76
Figure 3-7. Percent of children estimated to experience one or more exposures of concern at or
above 60, 70, 80 ppb with air quality adjusted to just meet the current standard -
Averaged Over 2006 to 2010 3-96
Figure 3-8. Percent of children estimated to experience one or more exposures of concern at or
above 60, 70, 80 ppb with air quality adjusted to just meet the current standard -
Worst-Case Year from 2006 to 2010 3-97
Figure 3-9. Percent of children estimated to experience two or more exposures of concern at or
above 60, 70, 80 ppb with air quality adjusted to just meet the current standard -
Averaged Over 2006 to 2010 3-98
Figure 3-10. Percent of children estimated to experience two or more exposures of concern at or
above 60, 70, 80 ppb with air quality adjusted to just meet the current standard -
Worst-Case Year from 2006 to 2010 3-99
Figure 3-11. Percent of school-aged children (5-18 yrs) estimated to experience one or more
days with FEVi decrements > 10, 15, or 20% with air quality adjusted to just meet
the current standard -Averaged over 2006 to 2010 3-107
Figure 3-12. Percent of school-aged children (5-18 yrs) estimated to experience one or more
days with FEVi decrements > 10, 15, or 20% with air quality adjusted to just meet
the current standard-Worst-Case Year from 2006 to 2010 3-108
Figure 3-13. Percent of school-aged children (aged 5-18 yrs) estimated to experience two or
more days with FEVi decrements > 10, 15, or 20% with air quality adjusted to just
meet the current standard- Averaged over 2006 to 2010 3-109
Figure 3-14. Percent of school-aged children (5-18 yrs) estimated to experience two or more
days with FEVi decrements > 10, 15, or 20% with air quality adjusted to just meet
the current standard- Worst-Case Year from 2006 to 2010 3-110
Figure 3-15. Percent of all-cause mortality associated with Os for air quality adjusted to just
meet the current standard 3-115
Figure 3-16. Estimated Os-associated deaths attributable to various area-wide average Os
concentrations, with air quality adjusted to just meet current standard 3-116
Figure 3-17. Percent of baseline respiratory mortality estimated to be associated with long-term
O3 3-118
Figure 4-1. Percent of children estimated to experience one or more exposures of concern at or
above 60, 70, or 80 ppb for air quality adjusted to just meet the current and
potential alternative standards (averaged over 2006 to 2010) 4-23
VI
-------
Figure 4-2. Percent of children estimated to experience one or more exposures of concern at or
above 60, 70, or 80 ppb for air quality adjusted to just meet the current and
potential alternative standards (worst-case year from 2006 to 2010) 4-24
Figure 4-3. Percent of children estimated to experience two or more exposures of concern at or
above 60, 70, or 80 ppb for air quality adjusted to just meet the current and
potential alternative standards (averaged over 2006 to 2010) 4-25
Figure 4-4. Percent of children estimated to experience two or more exposures of concern at or
above 60, 70, or 80 ppb for air quality adjusted to just meet the current and
potential alternative standards (worst-case year from 2006 to 2010) 4-26
Figure 4-5. Percent of children estimated to experience one or more Os-induced lung function
decrements greater than 10, 15, or 20% for air quality adjusted to just meet the
current and potential alternative standards (averaged over 2006 to 2010) 4-30
Figure 4-6. Percent of children estimated to experience one or more Os-induced lung function
decrements greater than 10, 15, or 20% for air quality adjusted to just meet the
current and potential alternative standards (worst-case year from 2006 to 2010)
4-31
Figure 4-7. Percent of children estimated to experience two or more Ch-induced lung function
decrements greater than 10, 15, or 20% for air quality adjusted to just meet the
current and potential alternative standards (averaged over 2006 to 2010) 4-32
Figure 4-8. Percent of children estimated to experience two or more Ch-induced lung function
decrements greater than 10, 15, or 20% for air quality adjusted to just meet the
current and potential alternative standards (worst-case year from 2006 to 2010)
4-33
Figure 4-9. Estimates of Total Mortality Associated with Short-Term Os Concentrations in
Urban Case Study Areas (Air Quality Adjusted to Current and Potential alternative
standard levels) - Total Risk 4-38
Figure 4-10. Estimates of Os-Associated Deaths Attributable to Full Distribution of 8-Hour
Area-Wide Os Concentrations and to Concentrations at or above 20, 40, or 60 ppb
- Deaths Summed Across Urban Case Study Areas 4-40
Figure 4-11. Estimates of Respiratory Hospital Admissions Associated with Short-Term Os
Concentrations in Urban Case Study Areas (Air Quality Adjusted to Current and
Potential alternative standard levels) - Total Risk 4-41
Figure 4-12. Estimates of Respiratory Mortality Associated with long-term Os Concentrations
in Urban Case Study Areas (Air Quality Adjusted to Current and Potential
alternative standard levels) - Total Risk 4-43
Figure 4-13. Estimates of Os-Associated Deaths Attributable to Full Distributions of 8-Hour
Area-Wide Os Concentrations and to Concentrations at or above 20, 40, or 60 ppb
Os - Deaths Summed Across Urban Case Study Areas and Expressed Relative to a
Standard with a Level of 75 ppb 4-51
Figure 5-1. A) Relative biomass loss in seedlings for 12 studied species using composite
functions in response to seasonal Os concentrations in terms of seasonal W126
index values, Y-axis scale for RBL values represents 0% up to 100% (U.S. EPA
2014, Figure 6-2) 5-25
Figure 5-1. B) Expanded view of relative biomass loss in seedlings for 12 studied species
using composite functions in response to seasonal Os concentrations in terms of
vii
-------
lower range of seasonal W126 index values, Y-axis scale for RBL values
represents 0% up to 10% (U.S. EPA 2014, Figure 6-2) 5-26
Figure 5-2. Relationship of tree seedling percent biomass loss with seasonal W126 index.
(From U.S. EPA 2014, Figure 6-5) 5-33
Figure 5-3. Relative biomass loss of Ponderosa Pine for air quality adjusted to just meet the
current standard (U.S. EPA 2014, Figure 6-8) 5-34
Figure 5-4. Relative yield loss in crops using the composite functions for 10 studied species in
response to seasonal Os concentrations in terms of seasonal W126 index values, Y-
axis scale for RYL values represents 0% up to 100% (U.S. EPA 2014, Figure 6-
3) 5-45
Figure 5-5. Cumulative proportion of biosites with any foliar injury present, by moisture
category (U.S. EPA 2014, Figure 7-10) 5-64
Vlll
-------
List of Tables
Table 1-1. Summary of primary and secondary Os NAAQS promulgated during the period from
1971 to 2008 1-5
Table 2-1 Comparison of the two model methodologies used to characterize background ozone
levels 2-15
Table 2-2. Seasonal mean MDA8 Os (ppb), seasonal mean apportionment-based USB
contribution (ppb), and fractional apportionment-based USB contribution to total Cb
(all site-days) in the 12 REA urban case study areas (%) 2-25
Table 2-3. Seasonal mean MDA8 Os (ppb), seasonal mean apportionment-based USB
contribution (ppb), and fractional apportionment-based USB contribution to total Os
(site-days > 60 ppb) in the 12 REA urban study areas (%) 2-25
Table 2-4. Fractional contribution of apportionment-based USB in the 12 REA urban study areas
(%), using the means and medians of daily MDA8 fractions (instead of fractions of
seasonal means) 2-25
Table 2-5. Seasonal mean MDA8 Os (ppb), seasonal mean USB (ppb), and USB/Total ratio (all
site-days) in the 12 REA urban case study areas (%) 2-25
Table 3-1. Group mean results of controlled human exposure studies that have evaluated
exposures to ozone concentrations below 75 ppb in young, healthy adults 3-58
Table 3-2. Panel studies of lung function decrements with analyses restricted to Os
concentrations below 75 ppb 3-60
Table 3-3. U.S. and Canadian epidemiologic studies reporting Os health effect associations in
locations that would have met the current standard during study periods 3-63
Table 3-4. Distributions of daily 8-hour maximum ozone concentrations from highest monitors
over range of 2-day moving averages from composite monitors (for study area
evaluated by Silverman and Ito, 2010) 3-67
Table 3-5. Distribution of daily 8-hour maximum ozone concentrations from highest monitors
over range of 3-day moving averages of population-weighted concentrations (for
study area evaluated by Strickland etal., 2010) 3-69
Table 3-6. Number of study cities with 4th highest daily maximum 8-hour concentrations greater
than 75 ppb, for various cut-point analyses presented in Bell et al. (2006) 3-74
Table 3-7. Prevalence of asthma by ageintheU.S 3-88
Table 4-1 Numbers of epidemiologic study locations likely to have met potential alternative
standards with levels of 70, 65, and 60 ppb 4-14
Table 4-2 Number of study cities with 3-year averages of 4th highest 8-hour daily max
concentrations greater than 70, 65, or 60 ppb, for various cut-point analyses presented
in Bell et al. (2006) 4-16
Table 4-3 Seasonal averages of 1-hour daily max Os concentrations in U.S. urban case study
areas for recent air quality and air quality adjusted to just meet the current and
potential alternative standards 4-20
Table 4-4 Summary of Estimated Exposures of Concern for Potential Alternative Os Standard
Levels of 70, 65, and 60 ppb in Urban Case Study Areas 4-49
Table 4-5 Summary of Estimated Lung Function Decrements for Potential Alternative Os
Standard Levels of 70, 65, and 60 ppb in Urban Case Study Areas 4-50
ix
-------
Table 5-1. Cb-Sensitive Trees, Their Uses and Relative Sensitivity 5-20
Table 5-2. Cb concentrations in Class I areas during period from 1998 to 2012 that met the
current standard and where three-year average W126 index value was at or above 15
ppm-hrs 5-28
Table 5-3. Exposure, risk and ecosystem services analyses related to tree growth, productivity
and carbon storage 5-30
Table 5-4. Summary of methodology by which national surface of 3-year average W126 index
values was derived for each air quality scenario 5-31
Table 5-5. Number of Counties with Tree Species Exceeding 2% Relative Biomass Loss 5-37
Table 5-6. Exposure, risk and ecosystem services analyses related to crop yield 5-48
Table 5-7. Visible foliar injury incidence in four National Wildlife Refuges 5-57
Table 5-8. Exposure, risk and ecosystem services analyses related to visible foliar injury 5-61
Table 5-9. Benchmark criteria for Os exposure and relative soil moisture used in screening-level
assessment of parks (from U.S. EPA 2014, Table 7-6) 5-65
Table 6-1. Tree seedling biomass loss and crop yield loss estimated for Os exposure over a
season 6-11
Table 6-2. Percent of assessed geographic area exceeding 2% weighted relative biomass loss in
WREA air quality scenarios 6-20
Table 6-3. Number of Class I areas (of 145 assessed) with weighted relative biomass loss
greater than 2% 6-22
Table 6-4. Estimated mean yield loss (and range across states) due to Os exposure for two
important crops 6-23
Table 6-5. Estimated effect of Os-sensitive tree growth-related impacts on the ecosystem
services of air pollutant removal and carbon sequestration in five urban case study
areas 6-25
-------
List of Acronyms and Abbreviations
Act
ACS
AHR
ANF
AOT40
AOT60
APEX
APHEA
APHENA
AQCD
AQI
AQRV
AQS
ARG
AspenFACE
ATS
BALF
BI
BRFSS
C
CAA
CALFIRE
CAMx
CAMP
CAR
CASAC
CASTNET
CAT
CBS A
CD
CDC
CFR
CH4
CHD
CHF
CHS
CI
CMAQ
CO
CO2
COPD
C-R
Clean Air Act
American Cancer Society
Airway hyperresponsiveness
Atrial natriuretic factor
Seasonal sum of the difference between an hourly concentration at the
threshold value of 40 ppb, minus the threshold value of 40 ppb
Seasonal sum of the difference between an hourly concentration at the
threshold value of 60 ppb, minus the threshold value of 60 ppb
Air Pollutants Exposure model
Air Pollution and Health: A European Approach
Air Pollution and Health: A European and North American Approach
Air Quality Criteria Document
Air Quality Index
Air quality related value
Air Quality System
Arginase
Aspen Free Air gas Concentration Enrichment Facility
American Thoracic Society
Bronchoalveolar Lavage Fluid
Biosite Index
Behavioral Risk Factor Surveillance System
Concentration
Clean Air Act
California Department of Forestry and Fire Protection
Comprehensive Air Quality Model with Extensions
Childhood Asthma Management Program
Centriacinar region
Clean Air Scientific Advisory Committee
Clean Air Status and Trends Network
Catalase
Core-based statistical area
Criteria Document
Centers for Disease Control
Code of Federal Regulations
Methane
Coronary Heart Disease
Congestive Heart Failure
Children's Health Study
Confidence interval
Community Multi-scale Air Quality Model
Carbon monoxide
Carbon dioxide
Chronic obstructive pulmonary disease
Concentration-response
XI
-------
CSA
CSTR
CVD
DHEW
ED
ELF
EPA
E-R
eVNA
FACE
FASOMGHG
FEM
FeNO
FEVi
FHM/FIA
FHWAR
FIA
FLAG
FR
FRM
FVC
GEOS
GIS
GRSM
GSTM1
GSTP1
HA
HDDM
HEI
HMOX1
HO
HR
HREA
HRV
ICD-9
ICU
IgE
IL
I/R
ISA
Max
MDA8
MMTCO2e
MPO
MSA
NAAQS
Combined Statistical Area
Continuous stirred tank reactor
Cardiovascular disease
Department of Health, Education, and Welfare
Emergency department
Extracellular Lining Fluid
Environmental Protection Agency
Exposure-response
Enhanced Voronoi Neighbor Averaging
Free-air CO2 (and ozone) enrichment system
Forest and Agricultural Optimization Model - Greenhouse gas version
Federal Equivalent Method
Exhaled nitric oxide fraction
Forced Expiratory Volume for 1 second
Forest Health Monitoring /Forest Inventory and Analysis Program
National Survey of Fishing, Hunting, and Wildlife-Associated Recreation
USDA Forest Inventory and Analysis Program
Federal Land Managers' Air Quality Related Values Workgroup
Federal Register
Federal Reference Method
Forced Vital Capacity
Goddard Earth Observing System
Geographic Information Systems
Great Smoky Mountains National Park
Glutathione-S-transferase polymorphism Ml genotypes
Glutathione-S-transferase polymorphism PI genotypes
Hospital Admission
Higher Order Direct Decoupled Method
Health Effects Institute
Heme oxygenase-1 polymorphism
Heme oxygenase
Heart rate
Health Risk and Exposure Assessment
Heart rate variability
International Classification of Disease - 9th revision
Intensive care unit
Immunoglobulin E
Interleukin
Ischemia-reperfusion
Integrated Science Assessment
maximum
Maximum daily 8-hour ozone average
Million metric tonnes of carbon dioxide equivalents
Myeloperoxidase
Metropolitan Statistical Area
National ambient air quality standards
xn
-------
NAB
NB
NCDC
NCLAN
NCORE sites
NHLBI
NMMAPS
NO
NO2
NOx
NOAA
NQO1
NRC
NRCS
NSRE
NWR
O3
OC
OIF
OIRA
OMB
OR
OTC
PA
PAMS
PAN
PAPA
PDSI
PEFR
ppm
ppm-hrs
ppb
PM
PM2.5
PMio
PMN
POMS
PRB
QA
RBL
ROMO
RR
SEKI
SES
SIP
SLAMS
North American background
Natural background
National Climatic Data Center
National Crop Loss Assessment Network
National Core multi-pollutant monitoring sites
National Heart, Lung, and Blood Institute
National Morbidity, Mortality, and Air Pollution Study
Nitric oxide
Nitrogen Dioxide
Nitrogen Oxides
National Oceanic and Atmospheric Administration
NAD(P)H-quinone oxidoreductase genotype
National Research Council
Natural Resources Conservation Service
National Survey on Recreation and the Environment
National wildlife refuges
Ozone
Organic carbon
Outdoor Industry Foundation
Office of Information and Regulatory Affairs
Office of Management and Budget
Odds ratio
Open-top chamber
Policy Assessment
Photochemical Assessment Monitoring Stations
Peroxyacetyl nitrate
Public Health and Air Pollution in Asia
Palmer Drought Severity Index
Peak Expiratory Flow Rate
Parts per million
part per million-hours
Parts per billion
Particulate matter
Particles generally less than or equal to 2.5 jim in diameter
Particles generally less than or equal to 10 micrometers (|im) in diameter
Polymorphonuclear leukocyte
Portable Ozone Monitoring System
Policy relevant background
Quality assurance
Relative biomass loss
Rocky Mountain National Park
Relative risk
Sequoia and Kings Canyon National Parks
Socioeconomic status
State Implementation Plans
State and Local Monitoring Stations
xni
-------
SO2
SOx
SoyFACE
SP
SPMS
STE
SUMOO
SUM06
TLC
TNF
UNEP
U.S.
USB
UV
VNA
voc
W126
WHO
wRBL
WREA
WTP
Sulfur Dioxide
Sulfur Oxides
Soybean Free Air gas Concentration Enrichment Facility
Staff Paper
Special Purpose Monitoring Stations
Stratospheric-tropospheric exchange
Season sum of all hourly average concentrations
Seasonal sum of all hourly average concentrations > 0.06 ppm
Total Lung Capacity
Tumor Necrosis Factor
United Nations Environmental Programme
United States
United States background
Ultraviolet
Ventilation rate
Voronoi neighbor Averaging
Volatile Organic Compounds
Cumulative integrated exposure index with a sigmoidal weighting function
World Health Organization
Weighted relative biomass loss
Welfare Risk and Exposure Assessment
Willingness to pay
xiv
-------
EXECUTIVE SUMMARY
This Policy Assessment (PA) has been prepared by staff in the Environmental Protection
Agency's (EPA) Office of Air Quality Planning and Standards (OAQPS) as part of the Agency's
review of the primary (health-based) and secondary (welfare-based) national ambient air quality
standards (NAAQS) for ozone (Os). The current Os standards were established in 2008 at the end
of the previous review cycle. These standards include a primary Os standard of 75 ppb,1 and a
secondary Os standard set identical to the primary standard. These 2008 standards are now under
review, as required by sections 108 and 109 of the Clean Air Act (Act). The PA presents
analyses and staff conclusions regarding the policy implications of the key scientific and
technical information that informs this review. Staff conclusions are presented regarding the
adequacy of the current standards and potential alternative standards appropriate for
consideration. Staff analyses in this PA are based on the scientific and technical information,
including the uncertainties and limitations related to this information, assessed and presented in
the Integrated Science Assessment for Ozone (ISA), the Health Risk and Exposure Assessment
for Ozone (HREA), and the Welfare Risk and Exposure Assessment for Ozone (WREA). The
PA is intended to "bridge the gap" between the relevant scientific evidence and technical
information and the judgments required of the EPA Administrator in determining whether to
retain or revise the current standards. Development of the PA is also intended to facilitate advice
and recommendations on the standards to the Administrator from an independent scientific
review committee, the Clean Air Scientific Advisory Committee (CASAC), as provided for in
the Act. Staff considerations and conclusions in this final PA have been informed by comments
and recommendations from CASAC, and by public comments.
Health Effects and Review of the Primary Standard
A longstanding and comprehensive evidence base, stronger now than in the last review,
documents the effects of Os exposures on human health. It is well-under stood that secondary
oxidation products, which develop as a result of Os exposure, initiate numerous responses at the
cellular, tissue, and whole organ level of the respiratory system. These key initiating events have
the potential to result in a variety of adverse respiratory effects, as well as effects outside the
respiratory system (e.g., cardiovascular effects). Ozone inhalation poses the greatest risk to
people in certain lifestages (i.e., children, older adults), people with asthma, people with certain
genetic variants (related to oxidative stress and inflammation), people with diets limited in
1 The level of the O3 standard is specified as 0.075 ppm rather than 75 ppb. However, in this PA we refer to ppb,
which is most often used in the scientific literature and in the ISA, in order to avoid the confusion that could result
from switching units when discussing the evidence in relation to the standard level.
ES-1
-------
certain nutrients (antioxidant vitamins C and E), and people experiencing the largest exposures
(e.g., outdoor workers, children). The evidence from animal toxicology and controlled human
exposure studies indicates that higher exposure concentrations and repeated exposures lead to a
greater prevalence of effects and increasingly severe effects, including increased susceptibility to
other respiratory stressors, among exposed populations, especially these at-risk populations.
As an initial matter in this PA, staff concludes that reducing ambient Os concentrations to
meet the current standard of 75 ppb will provide important improvements in public health
protection. This initial conclusion is based on (1) the strong body of scientific evidence
indicating a wide range of adverse health outcomes attributable to exposures to Os
concentrations found in the ambient air and (2) estimates indicating decreased Os exposures and
health risks upon meeting the current standard, compared to recent air quality.
Strong support for this initial conclusion is provided by controlled human exposure
studies of respiratory effects, and by quantitative estimates of exposures of concern and lung
function decrements based on information in these studies. Analyses in the HREA estimate that
the percentages of children (i.e., all children and children with asthma) in urban case study areas2
experiencing exposures of concern, or experiencing abnormal and potentially adverse lung
function decrements, are consistently lower for air quality that just meets the current Os standard
than for recent air quality. The HREA estimates such reductions consistently across the urban
case study areas and across years evaluated, and throughout various portions of individual urban
case study areas, including in urban cores and in the portions of case study areas surrounding
urban cores. These reductions in exposures of concern and Os-induced lung function decrements
reflect consistent reductions in relatively high Os concentrations (i.e., those in the upper portions
of the distribution of ambient concentrations) following reductions in precursor emissions to
meet the current standard. Thus, populations in both urban and non-urban areas would be
expected to experience important reductions in Ch exposures and Os-induced lung function risks
upon meeting the current standard.
Support for this initial conclusion is also provided by estimates of Os-associated mortality
and morbidity based on application of concentration-response relationships from epidemiologic
studies to air quality adjusted to just meet the current standard. While these estimates are
associated with uncertainties that complicate their interpretation, they suggest that Os-associated
mortality and morbidity would be expected to decrease nationwide following reductions in
precursor emissions to meet the current Os standard.
2 HREA analyses for exposures of concern and for risk of moderate or large lung function decrements covered 15
urban case study areas. HREA analyses of mortality and morbidity endpoints from epidemiologic studies covered
12 urban case study areas. Exposures and risks were evaluated for the years 2006 through 2010.
ES-2
-------
While meeting the current Os standard is estimated to result in important public health
improvements compared to recent air quality, staff further concludes that the Os-attributable
health effects estimated to be allowed by air quality that meets the current primary standard can
reasonably be judged important from a public health perspective. This conclusion is based on
consideration of: (1) the scientific evidence discussed in the ISA, including controlled human
exposure studies reporting abnormal or adverse respiratory effects following exposures to Os
concentrations below the level of the current standard and epidemiologic studies indicating
associations with morbidity and mortality for air quality that would likely meet the current
standard; (2) HREA estimates of Os exposures of concern, Os-induced lung function risks, and
Os-associated morbidity and mortality risks; (3) advice received from CASAC based on their
review of draft versions of the ISA, HREA, and PA, and advice received in previous reviews;
and (4) staff consideration of public comments. Staff reaches the overall conclusion that the
available health evidence and exposure/risk information call into question the adequacy of the
public health protection provided by the current standard.
Given this conclusion regarding the adequacy of the current standard, staff also reaches
conclusions for the Administrator's consideration regarding the elements of potential alternative
primary Os standards that could be supported by the available evidence and exposure/risk
information. Any such potential alternative standards should protect public health against effects
associated with exposures to Os, alone or in combination with related photochemical oxidants,
taking into account the available scientific evidence and exposure/risk information. In reaching
conclusions about the range of potential alternative standards appropriate for consideration, staff
is mindful that the Act requires primary standards that, in the judgment of the Administrator, are
requisite to protect public health with an adequate margin of safety. In setting a primary standard
that is "requisite" to protect public health, the EPA's task is to establish standards that are neither
more nor less stringent than necessary. The requirement that primary standards provide an
"adequate margin of safety" is intended to address uncertainties associated with inconclusive
scientific and technical information. Thus, the Act does not require that primary NAAQS be set
at zero-risk levels, but rather at levels that reduce risk sufficiently to protect public health with an
adequate margin of safety.
The degree of public health protection provided by any NAAQS results from the
collective impact of the elements of the standard, including the indicator, averaging time, form,
and level. Staffs conclusions on each of these elements are summarized below.
(1) Indicator: It is appropriate to continue to use Os as the indicator for a standard that is
intended to address effects associated with exposure to Os, alone or in combination with
related photochemical oxidants. Based on the available information, staff concludes that
ES-2
-------
there is no basis for considering any alternative indicator at this time. Meeting an Os
standard can be expected to provide some degree of protection against potential health
effects that may be independently associated with other photochemical oxidants, even
though such effects are not discernible from currently available studies indexed by Os
alone. Staff notes that control of ambient Os concentrations is generally understood to
provide the best means of controlling photochemical oxidants, and thus of protecting
against effects that may be associated with individual species and/or the broader mix of
photochemical oxidants. CASAC concurred with these conclusions.
(2) Averaging time: It is appropriate to consider retaining the current 8-hour averaging time
for the primary Os standard.
(a) Staff concludes that an 8-hour averaging time remains appropriate for addressing
health effects associated with short-term exposures to ambient Os. An 8-hour
averaging time is similar to the exposure periods evaluated in controlled human
exposure studies, including recent studies reporting respiratory effects following
exposures to Os concentrations below the level of the current standard. In
addition, epidemiologic studies provide evidence for health effect associations
with 8-hour Os concentrations, as well as with 1-hour and 24-hour concentrations.
Staff concludes that a standard with an 8-hour averaging time (combined with an
appropriate standard form and level) would be expected to provide substantial
protection against health effects attributable to 1- and 24-hour exposures. CASAC
concurred, concluding that the current 8-hour averaging time is justified by the
combined evidence from epidemiologic and clinical studies.
(b) Staff also concludes that a standard with an 8-hour averaging time can provide
protection against respiratory effects associated with longer term Os exposures.
Air quality analyses indicate that just meeting an 8-hour standard with an
appropriate level (i.e., 70 to 60 ppb, as discussed below) would be expected to
maintain long-term Os concentrations (i.e., seasonal average of 1-hour daily max)
below those where a key study indicates the most confidence in the concentration-
response relationship with respiratory mortality. In addition, risk analyses in the
HREA estimate that just meeting such alternative 8-hour standards would be
expected to decrease the incidence of respiratory mortality associated with long-
term Os concentrations. In considering other long-term Os metrics evaluated in
recent health studies, analyses in the HREA indicate that the large majority of the
U.S. population lives in locations where reducing precursor emissions would be
ES-4
-------
expected to decrease warm season averages of daily 8-hour ambient Os
concentrations, a long-term metric used in several recent studies reporting
associations with respiratory morbidity. Taken together, these analyses suggest
that a standard with an 8-hour averaging time, coupled with the current 4th-highest
form and an appropriate level (discussed below), could provide appropriate
protection against the long-term Os concentrations reported to be associated with
respiratory morbidity and mortality. CASAC concurred, concluding that the 8-
hour averaging time provides protection against the adverse impacts of long-term
Os exposures.
(3) Form: For an 8-hour Os standard with a revised level, as described below, it is
appropriate to consider retaining the current form, defined as the 3-year average of the
annual 4th-highest daily maximum concentration. Staff notes that this form was selected
in 1997 and 2008 in recognition of the public health protection provided, when coupled
with an appropriate averaging time and level, combined with the stability provided for
implementation programs. The currently available evidence and exposure/risk
information do not call into question these conclusions from previous reviews. CASAC
concurred with this conclusion, agreeing that the current form, combined with the current
8-hour averaging time, provides health protection while allowing for atypical
meteorological conditions that can lead to abnormally high ambient Os concentrations
which, in turn, provides programmatic stability.
(4) Level: The available scientific evidence and exposure/risk information provide strong
support for considering a primary Os standard with a revised level in order to increase
public health protection, including for at-risk populations and lifestages. Staff concludes
that it is appropriate in this review to consider a revised primary Os standard level within
the range of 70 ppb to 60 ppb. A standard set within this range would result in important
improvements in public protection, compared to the current standard, and could
reasonably be judged to provide an appropriate degree of public health protection,
including for at-risk populations and lifestages. In its advice to the Administrator,
CASAC also concluded that the scientific evidence and exposure/risk information
support consideration of standard levels from 70 to 60 ppb. Within this range, CASAC
concluded that a level of 70 ppb would provide little margin of safety and, therefore,
provided the policy advice that the level of the Os standard should be set below 70 ppb.
ES-5
-------
The Administrator's consideration of specific standard levels will reflect her judgments
as to the appropriate weight to be given to various aspects of the scientific evidence and
exposure/risk information, including the appropriate weight to be given to important
uncertainties. To inform these judgments, staff considers what the evidence and
information indicate with regard to the degree of public health protection that could be
achieved with levels from the upper (70 ppb), middle (65 ppb), and lower (60 ppb) parts
of the range.
A level of 70 ppb is below the Os exposure concentration that has been reported to elicit a
broad range of respiratory effects that includes airway hyperresponsiveness and
decreased lung host defense, in addition to lung function decrements, airway
inflammation, and respiratory symptoms (i.e., 80 ppb). A level of 70 ppb is also just
below the lowest exposure concentration at which the combined occurrence of respiratory
symptoms and lung function decrements have been reported (i.e., 72 ppb), a combination
judged adverse by the ATS (section 3.1.3). A level of 70 ppb is above the lowest
exposure concentration demonstrated to result in lung function decrements and
pulmonary inflammation (i.e., 60 ppb). Compared to the current standard, a revised Os
standard with a level of 70 ppb would be expected to (1) reduce the occurrence of
exposures of concern to Cb concentrations that result in respiratory effects in healthy
adults (at or above 60 and 70 ppb) by about 45 to 95%, almost eliminating the
occurrence of multiple exposures at or above 70 ppb; (2) reduce the occurrence of
moderate-to-large Os-induced lung function decrements (FEVi decrements > 10, 15,
20%) by about 15 to 35%, most effectively limiting the occurrence of multiple
decrements and decrements > 15, 20%; (3) more effectively maintain short- and long-
term Os concentrations below those present in the epidemiologic studies that reported
significant Os health effect associations in locations likely to have met the current
standard;3 and (4) reduce the risk of Os-associated mortality and morbidity, particularly
the risk associated with the upper portions of the distributions of ambient Os
concentrations.
A level of 65 ppb is well below the Os exposure concentration that has been reported to
elicit the wide range of potentially adverse respiratory effects noted above, and is below
the lowest exposure concentration at which the combined occurrence of respiratory
3 Though epidemiologic studies also provide evidence for O3 health effect associations in locations likely to have
met a standard with a level of 70 ppb, as discussed below for lower standard levels.
ES-6
-------
symptoms and lung function decrements has been reported. As noted above for 70 ppb, a
level of 65 ppb is above the lowest exposure concentration demonstrated to result in lung
function decrements and pulmonary inflammation. Compared to a standard with a level
of 70 ppb, a revised standard with a level of 65 ppb would be expected to (1) further
reduce the occurrence of exposures of concern (by about 80 to 100% compared to the
current standard), decreasing exposures at or above 60 ppb and almost eliminating
exposures at or above 70 and 80 ppb; (2) further reduce the occurrence of FEVi
decrements > 10, 15, and 20% (by about 30 to 65%, compared to the current standard);
(3) more effectively maintain short- and long-term Os concentrations below those present
in the epidemiologic studies that reported significant Os health effect associations in
locations likely to have met the current standard;4 and (4) further reduce the risk of Os-
associated mortality and morbidity, particularly the risk associated with the upper portion
of the distribution of ambient Os concentrations.
A level of 60 ppb is well below the Os exposure concentration shown to result in the
combined occurrence of respiratory symptoms and lung function decrements, and
corresponds to the lowest exposure concentration demonstrated to result in lung function
decrements and pulmonary inflammation. Compared to a standard with a level of 70 or
65 ppb, a revised standard with a level of 60 ppb would be expected to (1) further reduce
the occurrence of exposures of concern (by about 95 to 100% compared to the current
standard), almost eliminating exposures at or above 60 ppb; (2) further reduce the
occurrence of FEVi decrements > 10, 15, and 20%, (by about 45 to 85% compared to the
current standard); (3) more effectively maintain short- and long-term Os concentrations
below those present in the epidemiologic studies that reported significant Os health effect
associations in locations likely to have met the current standard;5 and (4) further reduce
the risk of Os-associated mortality and morbidity, particularly the risk associated with the
upper portion of the distribution of ambient Os concentrations.
Welfare Effects and Review of the Secondary Standard
The longstanding and comprehensive evidence base, stronger than in the last review,
documents the vegetation and ecosystem-related effects of Os in ambient air. In particular, recent
controlled studies at the molecular, biochemical and cellular scales have increased the
4 Though epidemiologic studies also provide evidence for O3 health effect associations in locations likely to have
met a standard with a level of 65 ppb.
5 Epidemiologic studies have not evaluated O3 health effect associations based primarily on air quality in locations
likely to have met a standard with a level of 60 ppb.
ES-7
-------
mechanistic understanding of the basic biology of how plants are affected by oxidative stress.
These studies have focused on a variety of plant responses to Os including: 1) reduced carbon
dioxide uptake due to stomatal closure; 2) the upregulation of genes associated with plant
defense, signaling, hormone synthesis and secondary metabolism; 3) the down regulation of
genes related to photosynthesis and general metabolism; 4) the loss of carbon assimilation
capacity due to declines in the quantity and activity of key proteins and enzymes; and 5) the
negative impacts on the efficiency of the photosynthetic light reactions. In addition, these effects
at the plant scale can be linked to an array of effects at larger scales, as shown in recent field
studies, together with previously available evidence. Specifically, plant-scale effects, such as
altered rates of leaf gas exchange, growth, and reproduction at the individual plant level, can
result in larger scale effects in ecosystems, such as alterations in productivity, carbon storage,
water cycling, nutrient cycling, and community composition. The available information also
demonstrates a relationship between changes in tropospheric Os concentrations and radiative
forcing, and between changes in tropospheric Os concentrations and effects on climate.
The long-standing body of available evidence also provides a wealth of information on
aspects of Os exposure that are important in influencing plant response. These include support
for the conclusions that: Os effects in plants are cumulative; higher Os concentrations appear to
be more important than lower concentrations in eliciting a response; plant sensitivity to Os varies
with time of day and plant development stage; and quantifying exposure with indices that
cumulate hourly Os concentrations and preferentially weight the higher concentrations improves
the explanatory power of exposure/response models for growth and yield, over using indices
based on mean and peak exposure values.
As an initial matter in this PA, staff concludes that reducing ambient Os concentrations to
meet the current standard of 75 ppb will provide important improvements in public welfare
protection. This initial conclusion is based on (1) the strong body of scientific evidence
indicating a wide range of effects to sensitive vegetation, including tree biomass loss, crop yield
loss, and visible foliar injury, and associated ecosystems and services attributable to cumulative
exposures to Os concentrations found in the ambient air and (2) estimates indicating decreased
cumulative Os exposures and welfare risks upon meeting the current standard, compared to
recent air quality. Strong support for this conclusion is provided by the available welfare
evidence; by WREA estimates of cumulative exposures to Os concentrations shown to result in
decreased biomass loss, crop yield loss, and visible foliar injury incidence under just meeting the
current secondary standard; and by WREA estimates of improvements in carbon storage and air
pollution removal in urban and commercial forests. Support for this conclusion is also provided
by WREA estimates of increased protection for Class I areas from Os-associated visible foliar
injury and tree biomass loss.
ES-8
-------
Staff further concludes that the Os-attributable welfare effects estimated to be allowed by
air quality that meets the current secondary standard call into question the adequacy of the public
welfare protection provided by the current standard. In addition, staff also concludes that the
public welfare protection is most appropriately judged through the use of a more biologically
relevant form, such as the cumulative, seasonal W126-metric. These conclusions are based on
consideration of: (1) the scientific evidence, including controlled exposure studies reporting
effects on plant growth, productivity and carbon storage, crop yield loss, and visible foliar injury
following exposures to Os concentrations below the level of the current standard and field based
studies that support these conclusions for air quality that would likely meet the current standard;
(2) the longstanding and extensive evidence demonstrating that the risk to vegetation comes from
cumulative seasonal exposures; (3) evidence suggesting that in Class I areas meeting the current
standard, cumulative seasonal Os exposures occur that are associated with estimates of tree
growth impacts of a magnitude that are reasonably considered important to public welfare; (4)
WREA estimates of reductions in biomass loss, crop yield loss, and visible foliar injury
incidence, and improvements in carbon storage and air pollution removal in urban and
commercial forests when meeting alternative W126 levels; (5) advice received from CASAC
based on their review of draft versions of the ISA, WREA, and PA, and advice received in
previous reviews; and (6) public comments. Staff reaches the overall conclusion that the
available vegetation and ecosystem effects evidence and exposure/risk information, including for
associated ecosystem services important from a public welfare perspective, call into question the
adequacy of the public welfare protection provided by the current standard. Based on the
evaluation presented in this PA, staff concludes that consideration should be given to revising the
standard to provide increased public welfare protection. CASAC agreed with this conclusion.
Given this conclusion regarding the adequacy of the current standard, staff also reaches
conclusions for the Administrator's consideration regarding the elements of potential alternative
secondary Os standards that could be supported by the available evidence and exposure/risk
information. Any such potential alternative standards should protect public welfare against
known or anticipated adverse environmental effects associated with exposures to Os, alone or in
combination with related photochemical oxidants, taking into account the available scientific
evidence and exposure/risk information. In reaching conclusions about the range of potential
alternative standards appropriate for consideration, staff is mindful that the Act requires
secondary standards that are at "a level of air quality the attainment and maintenance of which"
in the "judgment of the Administrator", are "requisite to protect public welfare from any known
or anticipated adverse effects". In setting a secondary standard that is "requisite" to protect
public welfare, the EPA's task is to establish standards that are neither more nor less stringent
ES-9
-------
than necessary. Thus, the Act does not require that NAAQS be set at zero-risk levels, but rather
at levels that reduce risk sufficiently to protect public welfare from adverse effects.
The degree of public welfare protection provided by any NAAQS results from the
collective impact of the elements of the standard, including the indicator, averaging time, form,
and level. Staffs conclusions on each of these elements are summarized below.
(1) Indicator: Staff concludes that it is appropriate to continue to use Cb as the indicator
for a standard that is intended to address welfare effects associated with exposure to
Os, alone or in combination with related photochemical oxidants. Based on the
available information, staff concludes that there is no basis for considering an
alternative indicator at this time. CASAC concurred with these conclusions.
(2) Averaging time and form: Staff concludes that it is appropriate to consider a revised
secondary standard in terms of the cumulative, seasonal, concentration-weighted
form, called the W126 index. This is supported by strong scientific evidence that
cumulative Os exposures drive plant response and can cause reduced tree growth,
productivity, and carbon storage; crop yield loss; visible foliar injury; and other
welfare effects. With regard to the appropriate form and averaging times, staff
reaches the following additional conclusions:
(a) It is appropriate to consider the consecutive 3-month period within the Os
season with the maximum index value as the seasonal period over which
to cumulate hourly Cb exposures. Staff notes that the maximum 3-month
period generally coincides with maximum biological activity for most
vegetation, making the 3-month duration a suitable surrogate for longer
growing seasons.
(b) It is appropriate to cumulate daily exposures for the 12-hour period from
8:00 am to 8:00 pm, generally representing the daylight period during the
3-month period identified above.
To the extent the Administrator finds it useful to consider the extent of public welfare
protection that might be afforded by a revised primary standard, staff concludes that
public welfare protection is appropriately judged through the use of the cumulative,
seasonal W126 index form, as described above. CASAC agreed that it was
ES-10
-------
appropriate to establish a revised form of the secondary standard and that the W126
index was a more biologically relevant form than the current form of the standard.
With regard to the number of years over which it is appropriate to average, staff
notes the that there is limited information to discern between the level of protection
provided by an annual form or a 3-year average form of a W126 standard for crop
yield loss or foliar injury, and that a multiple year form could be considered to
provide a more consistent target level of protection for this endpoint. Such a form
might also be appropriate for a standard intended to achieve the desired level of
protection from longer-term effects, including those associated with potential
compounding of biomass loss over multiple years. Further, such a form might be
concluded to contribute to greater stability in air quality management programs, and
thus, greater effectiveness in achieving the desired level of public welfare protection,
than that that might result from a single year form. Therefore, to the extent that the
greater emphasis is placed on protecting against effects associated with multi-year
exposures and maintaining more year-to-year stability of public welfare protection,
staff concludes that it is appropriate to consider a secondary standard form that
averages the seasonal W126 index values across three consecutive years. CAS AC
recommended that if a 3-year averaging period is selected, the level should be set
lower than if a 1-year averaging period is selected in order to provide greater
protection for annual crops and against cumulative effects on perennial species.
(3) Level: With regard to level for a revised secondary standard, staff concludes that it is
appropriate to give consideration to a range of levels from 17 to 7 ppm-hrs, expressed
in terms of the W126 index. In so doing, we primarily consider the evidence- and
exposure/risk-based information for cumulative seasonal Os exposures represented by
W126 index values (including those represented by the WREA average W126
scenarios) associated with biomass loss in studied tree species, both in and outside
areas that have been afforded special protections. We note CASAC's advice that a 6%
median RBL is unacceptably high, and that the 2% median RBL is an important
benchmark to consider. We further note that for the lower level of 7 ppm-hrs the
median tree species biomass loss is at or below 2% and that for the upper level of 17
ppm-hrs the median tree biomass loss is below 6%.6 We also note that a level of 17
ppm-hrs reduces the percent of total area having weighted RBL greater than 2% to
' We note that a W126 index value of 19 ppm-hrs is estimated to result in a median RBL value of 6%.
ES-11
-------
0.2%, and reduces the number of Class I areas with weighted RBL greater than 2% to
2 of the 145 assessed nationally protected Class I areas.
We also note that tree biomass loss can be an indicator of more significant ecosystem-
wide effects which might reasonably be concluded to be significant to public welfare.
For example, when it occurs over multiple years at a sufficient magnitude, biomass
loss is linked to an array of effects on other ecosystem-level processes such as
nutrient and water cycles, changes in above and below ground communities, and
carbon storage and air pollution removal. These effects have the potential to be
adverse to the public welfare.
In addition, a range of levels from 17 to 7 ppm-hrs would protect at least half of the
crop species from a yield loss of greater than 5%. A W126 level of 10 ppm-hrs or less
would also reduce prevalence of visible foliar injury and promote appreciable gains in
carbon sequestration and pollutant removal.
CAS AC recommended a range of W126 values of 15 ppm-hrs to 7 ppm-hrs and did
not recommend levels above 15 ppm-hrs. CAS AC noted that a level of 15 ppm-hrs is
requisite to protect median crop yield loss to no more than 5% and that a level below
10 ppm-hrs is required to reduce foliar injury prevalence. CAS AC also noted that a
W126 level of 7 ppm-hrs limits median relative biomass loss for trees to no greater
than 2% and offers additional protection against crop yield loss and foliar injury.
The Administrator's consideration of a particular level within the range of 17 to 7 ppm-
hrs would reflect judgments as to the appropriate weight to be given to various aspects of the
scientific evidence and exposure/risk information, with appropriate weight given to important
uncertainties and with particular consideration of the support provided by this evidence and
information regarding the protection of public welfare. To the extent the Administrator finds it
useful to consider the extent of public welfare protection that might be afforded by a revised
primary standard, staff concludes that public welfare protection is appropriately judged through
the use of the cumulative seasonal W126-based metric.
ES-12
-------
1 INTRODUCTION
1.1 PURPOSE
The U.S. Environmental Protection Agency (EPA) is presently conducting a review of
the primary (health-based) and secondary (welfare-based) national ambient air quality standards
(NAAQS) for ozone (Os). The overall plan for this review was presented in the Integrated
Review Plan for the Os National Ambient Air Quality Standards (IRP, U.S. EPA, 201 la). The
IRP also identified key policy-relevant issues to be addressed in this review and discussed the
key documents that generally inform NAAQS reviews, including an Integrated Science
Assessment (ISA), Risk and Exposure Assessments (REAs), and a Policy Assessment (PA). The
PA is prepared by the staff in EPA's Office of Air Quality Planning and Standards (OAQPS). It
presents a staff evaluation of the policy implications of the key scientific and technical
information in the ISA and REAs for EPA's consideration.1 The PA provides a transparent
evaluation, and staff conclusions, regarding policy considerations related to reaching judgments
about the adequacy of the current standards, and if revision is considered, what revisions may be
appropriate to consider.
The PA is intended to help "bridge the gap" between the Agency's scientific assessments
presented in the ISA and REAs, and the judgments required of the EPA Administrator in
determining whether it is appropriate to retain or revise the NAAQS.2 In evaluating the
adequacy of the current standard and whether it is appropriate to consider potential alternative
standards, the PA focuses on information that is most pertinent to evaluating the basic elements
of the NAAQS: indicator,3 averaging time, form,4 and level. These elements, which together
serve to define each standard, must be considered collectively in evaluating the health and
welfare protection afforded by the Os standards. The PA integrates and interprets the information
from the ISA and REAs to frame policy options for consideration by the Administrator. In so
doing, the PA recognizes that the selection of a specific approach to reaching final decisions on
the primary and secondary Os standards will reflect the judgments of the Administrator.
JThe terms "staff and "we" through this document refer to personnel in the EPA's Office of Air Quality Planning
and Standards (OAQPS).
2American Farm Bureau Federation v. EPA. 559 F. 3d 512, 521 (D.C. Cir. 2009); Natural Resources Defense
Council v. EPA. 902 F. 2d 962, 967-68, 970 (D.C. Cir. 1990).
3The "indicator" of a standard defines the chemical species or mixture that is to be measured in determining whether
an area attains the standard. The indicator for photochemical oxidants is ozone.
4The "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 current 8-hour Os NAAQS is the 3-
year average of the annual fourth-highest daily maximum 8-hour average.
1-1
-------
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. As
discussed below in section 1.2.1, the CASAC is to advise not only on the Agency's assessment
of the relevant scientific information, but also on the adequacy of the existing standards, and to
make recommendations as to any revisions of the standards that may be appropriate. The EPA
facilitates CASAC advice and recommendations, as well as public input and comment, by
requesting CASAC review and public comment on one or more drafts of the PA.
In this PA for the review of the Os NAAQS, we5 consider the scientific and technical
information available in this review as assessed in the Integrated Science Assessment for Os and
Related Photochemical Oxidants (ISA, U.S. EPA, 2013), prepared by EPA's National Center for
Environmental Assessment (NCEA), and the quantitative human exposure and health risk
assessment and welfare risk assessment documents (HREA, U.S. EPA, 2014a; WREA, U.S.
EPA, 2014b). The evaluation and staff conclusions presented in this PA have been informed by
comments and advice received from CASAC in their reviews of draft versions of the PA, and in
their reviews of the other draft Agency documents prepared for this NAAQS review.
Beyond informing the EPA Administrator and facilitating the advice and
recommendations of CASAC and the public, the PA is also intended to be a useful reference to
all parties interested in the NAAQS review. In these roles, it is intended to serve as a single
source of the most policy-relevant information that informs the Agency's review of the NAAQS,
and it is written to be understandable to a broad audience.
The remainder of chapter 1 summarizes information on the NAAQS legislative
requirements and on the history of the Os NAAQS (section 1.2), and summarizes our general
approaches to reviewing the current Os NAAQS (section 1.3). Chapter 2 of this PA provides an
overview of the Os ambient monitoring network and Os air quality, including estimates of Os
concentrations attributable to background sources. The remaining chapters are organized into
two main parts. Chapters 3 and 4 focus on the review of the primary Os NAAQS while chapters
5 and 6 focus on the review of the secondary Os NAAQS. Staffs considerations and conclusions
related to the current primary and secondary standards are discussed in chapters 3 and 5,
respectively. Staffs considerations and conclusions related to potential alternative primary and
secondary standards are discussed in chapters 4 and 6, respectively. Key uncertainties in the
review and areas for future research and data collection are additionally identified in chapters 4
and 6 for the two types of standards.
5 As noted above, the term "we" through this document refer to personnel in the EPA's Office of Air Quality
Planning and Standards (OAQPS).
1-2
-------
1.2 BACKGROUND
1.2.1 Legislative Requirements
Two sections of the Clean Air Act (CAA) govern the establishment and revision of the
NAAQS. Section 108 (42 U.S.C. section 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 air pollutants that in her "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 . . .
[the Administrator] plans to issue air quality criteria...." 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(b). 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. Section 109(b)(l) defines a primary standard as one "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."6
A secondary standard, as defined in section 109(b)(2), 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." 7
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 State of Mississippi
v. EPA. 744 F. 3d 1334, 1353 (D.C. Cir. 2012) ("By requiring an 'adequate margin of safety',
Congress was directing EPA to build a buffer to protect against uncertain and unknown dangers
to human health"). See also Lead Industries Association v. EPA. 647 F.2d 1130, 1154 (D.C. Cir
1980); American Petroleum Institute v. Costle. 665 F.2d 1176, 1186 (D.C. Cir. 1981); American
Farm Bureau Federation v. EPA. 559 F. 3d 512, 533 (D.C. Cir. 2009); Association of Battery
6 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).
7 Welfare effects as defined in section 302(h) (42 U.S.C. § 7602(h)) include, but are not limited to, "effects on soils,
water, crops, vegetation, man-made 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
-------
Recvclers v. EPA. 604 F. 3d 613, 617-18 (D.C. Cir. 2010). 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 provide 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. 647F.2dat 1156 n.51: State of Mississippi v. EPA. 744 F. 3d at 1343, 1351, but rather
at a level that reduces risk sufficiently so as to protect public health with an adequate margin of
safely.
In addressing the requirement for an adequate margin of safety, the EPA considers such
factors as the nature and severity of the health effects, the size of sensitive population(s)8 at risk,
and the kind and degree of the uncertainties that must be addressed. The selection of any
particular approach for providing an adequate margin of safety is a policy choice left specifically
to the Administrator's judgment. See Lead Industries Association v. EPA, 647 F.2d at 1161-62;
State of Mississippi. 744 F. 3d at 1353.
In setting primary and secondary standards that are "requisite" to protect public health
and welfare, respectively, as provided in section 109(b), EPA's task is to establish standards that
are neither more nor less stringent than necessary for these purposes. In so doing, the EPA may
not consider the costs of implementing the standards. See generally, Whitman v. American
Trucking Associations. 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 at 1185.
Section 109(d)(l) requires that "not later than December 31, 1980, and at 5-year intervals
thereafter, the Administrator shall complete a thorough review of the criteria published under
section 108 and the national ambient air quality standards . . . and shall make such revisions in
such criteria and standards and promulgate such new standards as may be appropriate
Section 109(d)(2) requires that an 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
8 As used here and similarly throughout this document, the term population refers to persons having a quality or
characteristic in common, including a specific pre-existing illness or a specific age or life stage.
1-4
-------
criteria and standards as may be appropriate . . . ." Since the early 1980's, the Clean Air
Scientific Advisory Committee (CASAC) has performed this independent review function.9
1.2.2 History of O3 NAAQS Reviews
Table 1-1 summarizes the O3 NAAQS that the EPA has promulgated to date. In each
review, the EPA set the secondary standard at a level identical to the primary standard. These
reviews are briefly described below.
Table 1-1. Summary of primary and secondary Os NAAQS promulgated during the
period from 1971 to 2008.
Final Rule
1971 (36 FR 81 86)
1979 (44 FR 8202)
1 993 (58 FR 13008)
1997 (62 FR 38856)
2008 (73 FR 16483)
Indicator
Total photochemical
oxidants
03
Averaging Time
1 hour
1 hour
Level (ppm)
0.08
0.12
Form
Not to be exceeded more than
one hour per year
Attainment is defined when the
expected number of days per
calendar year, with maximum
hourly average concentration
greater than 0.12 ppm, is
equal to or less than 1
The EPA decided that revisions to the standards were not warranted at the time.
03
03
8 hours
8 hours
0.08
0.075
Annual fourth-highest daily
maximum 8-hour
concentration, averaged over 3
years
Form of the standards
remained unchanged relative
to the 1997 standard
The EPA first established primary and secondary NAAQS for photochemical oxidants in
1971 (36 FR 8186, April 30, 1971). The EPA set both primary and secondary standards at a level
of 0.08 parts per million (ppm), 1-hr average, total photochemical oxidants, not to be exceeded
more than one hour per year. The EPA based the standards on scientific information contained in
the 1970 Air Quality Criteria for Photochemical Oxidants (U.S. DHEW, 1970). We initiated the
first periodic review of the NAAQS for photochemical oxidants in 1977. Based on the 1978 Air
9 Lists of CASAC members and of members of the CASAC Ozone Review Panel are available at:
http://yosemite.epa.gov/sab/sabpeople.nsfAVebCommitteesSubCommittees/Ozone%20Review%20Panel.
1-5
-------
Quality Criteria for Ozone and Other Photochemical Oxidants (U.S. EPA, 1978), the EPA
published proposed revisions to the original NAAQS in 1978 (43 FR 16962) and final revisions
in 1979 (44 FR 8202). At that time, the EPA revised the level of the primary and secondary
standards from 0.08 to 0.12 ppm and changed the indicator from photochemical oxidants to Os,
and the form of the standards from a deterministic to a statistical form. This statistical form
defined attainment of the standards as occurring when the expected number of days per calendar
year with maximum hourly average concentration greater than 0.12 ppm equaled one or less.
Following the final decision in the 1979 review, the City of Houston challenged the
Administrator's decision arguing that the standard was arbitrary and capricious because natural
Os 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 above) that attainability and technological
feasibility are not relevant considerations in the promulgation of the NAAQS. 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 Act. See API v. Costle, 665
F.2d 1176, 1184-6 (D.C. Cir. 1981).
In 1982, we announced plans to revise the 1978 Air Quality Criteria document (47 FR
11561), and in 1983, we initiated the second periodic review of the Os NAAQS (48 FR 38009).
We subsequently published the 1986 Air Quality Criteria for Ozone and Other Photochemical
Oxidants (U.S. EPA, 1986) and the 1989 Staff Paper (U.S. EPA, 1989). Following publication of
the 1986 Air Quality Criteria Document (AQCD), a number of scientific abstracts and articles
were published that appeared to be of sufficient importance concerning potential health and
welfare effects of Os to warrant preparation of a Supplement. On August 10, 1992, under the
terms of a court order, the EPA published a proposed decision to retain the existing primary and
secondary standards (57 FR 35542). The notice explained that the proposed decision would
complete EPA's review of information on health and welfare effects of Os assembled over a 7-
year period and contained in the 1986 AQCD and its 1992 Supplement. The proposal also
announced EPA's intention to proceed as rapidly as possible with the next review of the air
quality criteria and standards for Os in light of emerging evidence of health effects related to 6-
to 8-hour Os exposures. On March 9, 1993, the EPA concluded the review by affirming its
proposed decision to retain the existing primary and secondary standards. (58 FR 13008).
In August 1992, we announced plans to initiate the third periodic review of the air quality
criteria and Os NAAQS (57 FR 35542). In December 1996, the EPA proposed to replace the then
existing 1-hour primary and secondary standards with 8-hour average Os standards set at a level
of 0.08 ppm (equivalent to 0.084 ppm using standard rounding conventions) (61 FR 65716). The
1-6
-------
EPA also proposed to establish a new distinct secondary standard using a biologically-based
cumulative, seasonal form. The EPA completed this review on July 18, 1997 (62 FR 38856) by
setting the primary standard at a level of 0.08 ppm, based on the annual fourth-highest daily
maximum 8-hr average concentration, averaged over three years, and setting the secondary
standard identical to the revised primary standard. In reaching this decision, the EPA identified
several reasons supporting its decision to reject a potential alternate standard set at 0.07 ppm.
Most importantly, the EPA pointed out the scientific uncertainty at lower concentrations and
placed significant weight on the fact that no CASAC panel member supported a standard level
set lower than 0.08 ppm (62 FR 38868). In addition to noting the uncertainties in the health
evidence for exposure concentrations below 0.08 ppm and the advice of CASAC, the EPA noted
that a standard set at a level of 0.07 ppm would be closer to peak background concentrations that
infrequently occur in some areas due to nonanthropogenic sources of Os precursors (62 FR
38856, 38868; July 18, 1997).
On May 14, 1999, in response to challenges by industry and others to EPA's 1997
decision, the U.S. Court of Appeals for the District of Columbia Circuit remanded the Os
NAAQS to the EPA, finding that section 109 of the Act, as interpreted by the EPA, effected an
unconstitutional delegation of legislative authority. American Trucking Assoc. vs. EPA, 175
F.3d 1027, 1034-1040(D.C. Cir. 1999) ("ATA I"). In addition, the court directed that, in
responding to the remand, the EPA should consider the potential beneficial health effects of Os
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 petitioned for rehearing en bane on
several issues related to that decision. The court granted the request for rehearing in part and
denied it in part, but declined to review its ruling with regard to the potential beneficial effects of
O3 pollution. 195 F3d 4, 10 (D.C Cir., 1999) ("ATAII"). 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 Os. 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 Assoc., 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 Os 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 remand, finding the
1997 Os NAAQS to be "neither arbitrary nor capricious," and so denying the remaining petitions
for review. American Trucking Associations, Inc. v EPA, 283 F.3d 355, 379 (D.C Cir., 2002)
("ATA III").
1-7
-------
Specifically, in ATA III, the D.C. Circuit upheld EPA's decision on the 1997 Os standard
as the product of reasoned decision-making. The Court made clear that the most important
support for EPA's decision was the health evidence and the concerns it raised about setting a
standard level below 0.08 ppm. ("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). 283 F 3d at 378. The Court also pointed to the significant
weight that the EPA properly placed on the advice it received from CASAC. Id at 379. The
court further noted that "although relative proximity to peak background ozone concentrations
did not, in itself, necessitate a level of 0.08, EPA could consider that factor when choosing
among the three alternative levels." Id.
Independently of the litigation, the EPA also responded to the Court's remand to
consider the potential beneficial health effects of Os pollution in shielding the public from effects
of solar (ultraviolet or UV-B) radiation. The EPA provisionally determined that the information
linking changes in patterns of ground-level Os concentrations to changes in relevant patterns of
exposures to ultraviolet (UV-B) radiation of concern to public health was too uncertain, at that
time, to warrant any relaxation in 1997 Os NAAQS. The EPA also expressed the view that any
plausible changes in UV-B radiation exposures from changes in patterns of ground-level Os
concentrations would likely be very small from a public health perspective. In view of these
findings, the EPA proposed to leave the 1997 8-hour NAAQS unchanged (66 FR 57268, Nov.
14, 2001). After considering public comment on the proposed decision, the EPA published its
final response to this remand on January 6, 2003, re-affirming the 8-hour Os NAAQS set in 1997
(68FR614).
The EPA initiated the fourth periodic review of the air quality criteria and Ch standards in
September 2000 with a call for information (65 FR 57810). The schedule for completion of that
review was ultimately governed by a consent decree resolving a lawsuit filed in March 2003 by
plaintiffs representing national environmental and public health organizations, who maintained
that EPA was in breach of a mandatory legal duty to complete review of the Os NAAQS within a
statutorily-mandated deadline. On July 11, 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). Documents supporting
this proposed decision included the Air Quality Criteria for Ozone and Other Photochemical
Oxidcmts (U.S. EPA, 2006) and the Staff Paper (U.S EPA, 2007) and related technical support
documents. The EPA also proposed two options for revising the secondary standard: (1) replace
the current standard with a cumulative, seasonal standard, expressed as an index of the annual
sum of weighted hourly concentrations cumulated over 12 daylight hours during the consecutive
3-month period within the Os season with the maximum index value, set at a level within the
range of 7 to 21 ppm-hrs, and (2) set the secondary standard identical to the proposed primary
1-8
-------
standard. The EPA completed the review with publication of a final decision on March 27, 2008
(73 FR 16436). In that final rule, the EPA revised the NAAQS by lowering the level of the 8-
hour primary Os standard from 0.08 ppm to 0.075 ppm, not otherwise revising the primary
standard, and adopting a secondary standard identical to the revised primary standard. In May
2008, state, public health, environmental, and industry petitioners filed suit challenging EPA's
final decision on the 2008 Os standards. On September 16, 2009, the EPA announced its
intention to reconsider the 2008 Cb standards, and initiated a rulemaking to do so. At EPA's
request, the Court held the consolidated cases in abeyance pending EPA's reconsideration of the
2008 decision.
On January 19, 2010 (75 FR 2938), the EPA issued a notice of proposed rulemaking to
reconsider the 2008 final decision. 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 decrease the level of the 2008 8-hour primary
standard from 0.075 ppm to a level within the range of 0.060 to 0.070 ppm, and to change the
secondary standard to a new cumulative, seasonal standard expressed as an annual index of the
sum of weighted hourly concentrations, cumulated over 12 hours per day (8 am to 8 pm), during
the consecutive 3-month period within the Os season, with a maximum index value set at a level
within the range of 7 to 15 ppm-hours. The Agency also solicited CAS AC review of the
proposed rule on January 25, 2010 and solicited additional CASAC advice on January 26, 2011.
After considering comments from CASAC and the public, the EPA prepared a draft final rule,
which was submitted for interagency review pursuant to Executive Order 12866. On September
2, 2011, consistent with the direction of the President, the Administrator of the Office of
Information and Regulatory Affairs ("OIRA"), Office of Management and Budget ("OMB"),
returned the draft final rule to the EPA for further consideration. In view of this return and the
timing of the Agency's ongoing periodic review of the Os NAAQS required under Clean Air Act
section 109 (as announced on September 29, 2008), the EPA decided to coordinate further
proceedings on its voluntary rulemaking on reconsideration with that ongoing periodic review,
by deferring the completion of its voluntary rulemaking on reconsideration until it completes its
statutorily-required periodic review.
In light of EPA's decision to consolidate the reconsideration with the current review, the
Court proceeded with the litigation on the 2008 final decision. On July 23, 2013, the D.C. Circuit
Court of Appeals upheld EPA's 2008 primary Os standard, but remanded the 2008 secondary
standard to the EPA. State of Mississippi v. EPA, 744 F. 3d 1334. With respect to the primary
standard, the court first held that the EPA reasonably determined that the existing standard was
not requisite to protect public health with an adequate margin of safety, and consequently
required revision. Specifically, the court noted that there were "numerous epidemiological
1-9
-------
studies linking health effects to exposure to ozone levels below 0.08 ppm and clinical human
exposure studies finding a causal relationship between health effects and exposure to ozone
levels at and below 0.08 ppm". 744 F. 3d at 1345. The court also specifically endorsed the
weight of evidence approach utilized by EPA in its deliberations. Id at 1344.
The court went on to reject arguments that EPA should have adopted a more stringent
primary standard. Dismissing arguments that a clinical study (as properly interpreted by EPA) to
show effects at 0.06 ppm necessitated a standard level lower than that selected, the court noted
that this was a single, limited study. Id. at 1350. With respect to the epidemiologic evidence, the
court accepted EPA's argument that there could be legitimate uncertainty that a causal
relationship between Os and 8-hour exposures less than 0.075 ppm exists, so that associations at
lower levels reported in epidemiologic studies did not necessitate a more stringent standard. Id.
at 1351-52.10
The court also rejected arguments that an 8-hour primary standard of 0.075 ppm failed to
provide an adequate margin of safety, noting that margin of safety considerations involved policy
judgments by the agency, and that by setting a standard "appreciably below" the level of the
current standard (0.08 ppm), the agency had made a reasonable policy choice. Id. Finally, the
court rejected arguments that EPA's decision was inconsistent with CAS AC's scientific
recommendations because CASAC had been insufficiently clear in its recommendations whether
it was providing scientific or policy recommendations, and EPA had reasonably addressed
CASAC's policy recommendations. Id. at 1357-58.
With respect to the secondary standard, the court held that because EPA had failed to
identify a level of air quality requisite to protect public welfare, EPA's comparison between the
primary and secondary standards for determining if requisite protection for public welfare was
afforded by the primary standard was inherently arbitrary. The court thus rejected EPA's
determination that the revised 8-hour primary standard afforded sufficient protection of public
welfare, and remanded the standard to EPA. Id. at 1360-62.
1.2.3 Current O3 NAAQS Review
On September 29, 2008, the EPA announced the initiation of a new periodic review of
the air quality criteria for Os and related photochemical oxidants and issued a call for
information in the Federal Register (73 FR 56581, Sept. 29, 2008). A wide range of external
10 The court cautioned, however, that "perhaps more [clinical] studies like the Adams studies will
yet reveal that the 0.060 ppm level produces significant adverse decrements that simply cannot
be attributed to normal variation in lung function", and further cautioned that "agencies may not
merely recite the terms 'substantial uncertainty' as a justification for their actions'". Id. at 1350,
1357 (internal citations omitted).
1-10
-------
experts, as well as EPA staff, representing a variety of areas of expertise (e.g., epidemiology,
human and animal toxicology, statistics, risk/exposure analysis, atmospheric science, ecology,
biology, plant science, ecosystem services) participated in a workshop. This workshop was held
on October 28-29, 2008 in Research Triangle Park, NC. The workshop provided an opportunity
for a public discussion of the key policy-relevant issues around which the EPA would structure
this Os NAAQS review and the most meaningful new science that would be available to inform
our understanding of these issues.
Based in part on the workshop discussions, the EPA developed a draft Integrated Review
Plan outlining the schedule, process, and key policy-relevant questions that would guide the
evaluation of the air quality criteria for Os and the review of the primary and secondary Os
NAAQS. A draft of the IRP was released for public review and comment in September 2009.
This IRP was the subject of a consultation with the CASAC on November 13, 2009 (74 FR
54562; October 22, 2009).u We considered comments received from that consultation and from
the public in finalizing the plan and in beginning the review of the air quality criteria. The EPA's
overall plan and schedule for this review is presented in the Integrated Review Plan for the
Ozone National Ambient Air Quality Standards. u
As part of the process of preparing the Os ISA, NCEA hosted a peer review workshop in
October 29-30, 2008 (73 FR 56581, September 29, 2008) on preliminary drafts of key ISA
chapters. The CASAC and the public reviewed the first external review draft ISA (U.S. EPA,
201 Ib; 76 FR 10893, February 28, 2011) at a meeting held in May 19-20, 2011 (76 FR 23809;
April 28, 2011). Based on CASAC and public comments, NCEA prepared a second draft ISA
(U.S. EPA, 201 Ic; 76 FR 60820, September 30, 2011). CASAC and the public reviewed this
draft at a January 9-10, 2012 (76 FR 236, December 8, 2011) meeting. Based on CASAC and
public comments, NCEA prepared a third draft ISA (U.S. EPA 2012a; 77 FR 36534; June 19,
2012), which was reviewed at a CASAC meeting in September 2012. The EPA released the final
IS A in February 2013.
The EPA presented its plans for conducting the Risk and Exposure Assessments (REAs)
that build on the scientific evidence presented in the ISA, in two planning documents titled
Ozone National Ambient Air Quality Standards: Scope and Methods Plan for Health Risk and
Exposure Assessment and Ozone National Ambient Air Quality Standards: Scope and Methods
11 See http://vosemite.epa.gov/sab/sabproduct.nsf/WebProjectsbvTopicCASACIOpenView for more information on
CASAC activities related to the current O3 NAAQS review.
12 EPA 452/R-11-006; April 2011; Available:
http://www.epa.gov/ttn/naaqs/standards/ozone/data/2011 04 OzoneIRP.pdf
1-11
-------
Plan for Welfare Risk and Exposure Assessment (henceforth, Scope and Methods Plans).13
These planning documents outlined the scope and approaches that staff planned to use in
conducting quantitative assessments, as well as key issues that would be addressed as part of the
assessments. We released these documents for public comment in April 2011, and consulted with
CASAC on May 19-20, 2011 (76 FR 23809; April 28, 2011). In designing and conducting the
initial health risk and welfare risk assessments, we considered CASAC comments (Samet 2011)
on the Scope and Methods Plans and also considered public comments. In May 2012, we issued
a memo titled Updates to Information Presented in the Scope and Methods Plans for the Ozone
NAAQS Health and Welfare Risk and Exposure Assessments that described changes to elements
of the scope and methods plans and provided a brief explanation of each change and the reason
for it.
In July 2012, EPA made the first drafts of the Health and Welfare REAs available for
CASAC review and public comment (77 FR 42495, July 19, 2012). The first draft PA was made
available for CASAC review and public comment in August 2012. These documents were
reviewed by the CASAC Os Panel at a public meeting in September 2012. The second draft
REAs and PA were prepared by EPA in consideration of CASAC (Frey and Samet, 2012a,
2012b) and public comment and were reviewed by the CASAC Os Panel at a public meeting on
March 25-27, 2014. This final PA reflects staffs consideration of the comments and
recommendations made by CASAC, and comments made by members of the public, in their
review of draft versions of the PA.
1.3 GENERAL APPROACH FOR REVIEW OF THE STANDARDS
As described in section 1.1 above, this PA presents a transparent evaluation and staff
conclusions regarding policy considerations related to reaching judgments about the adequacy of
the current standards and the revisions that are appropriate to consider. Staff considerations and
conclusions in this document are based on the available body of scientific evidence assessed in
the ISA (U.S. EPA, 2013), exposure and risk analyses presented in the REAs (U.S. EPA, 2014a,
b), advice and recommendations from CASAC on the first and second draft REAs and PA and
other draft and final EPA documents in this review, as well as on public comments. This
evaluation and associated conclusions on the range of policy options that, in staffs view, are
supported by the available scientific evidence and exposure/risk information will inform the
Administrator's decisions as to whether the existing primary and/or secondary Os standards
should be revised and, if so, what revised standard or standards is/are appropriate.
13 EPA-452/P-11-001 and -002; April 2011; Available:
http://www.epa.gOv/ttn/naaqs/standards/ozone/s o3 2008_pd.html
1-12
-------
Staffs considerations and conclusions related to the current and alternative primary and
secondary Os standards are framed by a series of key policy-relevant questions, expanding upon
those presented in the IRP at the outset of this review (U.S. EPA, 201 la). Answers to these
questions in this final PA will inform the Administrator's decisions as to whether, and if so how,
to revise the current Os standards. The first overarching question is as follows.
• Do the currently available scientific evidence and exposure/risk information, as
reflected in the ISA and REAs, support or call into question the adequacy of the
protection afforded by the current Os standards?
If the answer to this question, which is informed by staffs consideration of more specific
questions related to the primary and secondary standards, suggests that revision of the current
standards may be appropriate, then staff further considers the currently available evidence and
information with regard to the following question.
• What range of potential alternative standards is appropriate to consider based on
the scientific evidence, air quality analyses, and exposure/risk-based
information?
The general approaches for consideration of these overarching questions in review of the primary
and secondary standards are described separately in sections 1.3.1 and 1.3.2 below.
1.3.1 Approach for the Primary Standard
Staffs approach in this review of the current primary Os standard takes into
consideration the approaches used in previous Os NAAQS reviews. The past and current
approaches described below are both based, fundamentally, on using EPA's assessment of the
current scientific evidence and associated quantitative analyses to inform the Administrator's
judgment regarding a primary standard for Os that is "requisite" (i.e., neither more nor less
stringent than necessary) to protect public health with an adequate margin of safety.
In reaching conclusions on options for the Administrator's consideration, we note that the
final decision to retain or revise the current primary Os standard is a public health policy
judgment to be made by the Administrator. This final decision by the Administrator will draw
upon the available scientific evidence for Os-attributable health effects, and on analyses of
population exposures and health risks, including judgments about the appropriate weight to
assign the range of uncertainties inherent in the evidence and analyses. Our general approach to
informing these judgments, discussed more fully below, recognizes that the available health
effects evidence reflects a continuum from relatively higher Os concentrations, at which
scientists generally agree that health effects are likely to occur, through lower concentrations, at
which the likelihood and magnitude of a response become increasingly uncertain. Therefore, in
developing conclusions in this PA, we are mindful that the Administrator's ultimate judgments
1-13
-------
on the primary standard will most appropriately reflect an interpretation of the available
scientific evidence and exposure/risk information that neither overstates nor understates the
strengths and limitations of that evidence and information. This approach is consistent with the
requirements of sections 108 and 109 of the Act, as well as with how the EPA and the courts
have historically interpreted the Act.
Section 1.3.1.1 below provides an overview of the general approach taken in the last
review of the primary Os NAAQS (i.e., the 2008 review), and a summary of the rationale for the
decision on the level of the standard in that review (73 FR 16436). Section 1.3.1.2 presents our
approach in the current review, including our approach to considering the health evidence and
exposure/risk information, and considerations regarding ambient Os concentrations attributable
to background sources.
1.3.1.1 Approach Used in the Last Review
In the 2008 review of the Os NAAQS, the Administrator considered the available
scientific evidence and exposure/risk information, the advice and recommendations of CASAC,
and comments from the public. Based on this, he revised the level of the 8-hour primary Os
standard from 0.08 ppm14 to 0.075 ppm (75 ppb15). In reaching a decision to revise the 1997 8-
hour primary Os standard, the Administrator noted that much new evidence had become
available since the 1997 review. He noted that this body of scientific evidence was very robust
and provided consistent and coherent evidence of an array of Os-related respiratory morbidity
effects, and possibly cardiovascular-related morbidity, as well as total nonaccidental and
cardiorespiratory mortality. The Administrator specifically observed that (1) the evidence of a
range of respiratory-related morbidity effects had been considerably strengthened; (2) newly
available evidence from controlled human exposure and epidemiologic studies identified people
with asthma as an important susceptible population for which estimates of respiratory effects in
the general population likely underestimate the magnitude or importance of these effects; (3)
newly available evidence about mechanisms of toxicity more completely explained the
biological plausibility of Os-induced respiratory effects and was beginning to suggest
mechanisms that may link Os exposure to cardiovascular effects; and (4) there was relatively
strong evidence for associations between short-term Os concentrations and total nonaccidental
and cardiopulmonary mortality. In the opinion of the Administrator, this very robust body of
evidence enhanced our understanding of Os- related effects and provided increased confidence
14 Due to rounding convention, the 1997 standard level of 0.08 ppm corresponded to 0.084 ppm (84 ppb).
15 The level of the O3 standard is specified as 0.075 ppm rather than 75 ppb. However, in this PA we refer to ppb,
which is most often used in the scientific literature and in the ISA, in order to avoid the confusion that could result
from switching units when discussing the evidence in relation to the standard level.
1-14
-------
that various respiratory morbidity effects and other effects marked by indicators of respiratory
morbidity are causally related to Cb exposures, and the evidence was highly suggestive that Os
exposures during the warm Os season contribute to premature mortality.16
The Administrator also noted important new health evidence reporting a broad array of
adverse effects following short-term exposures to Os concentrations below the level of the 1997
standard, and concerns for such or related effects in at-risk populations,17 including people with
asthma or other lung diseases, older adults with increased susceptibility, and those who are likely
to be vulnerable as a result of spending a lot of time outdoors engaged in physical activity (e.g.,
especially active children and outdoor workers).
He specifically noted new scientific evidence, which built upon existing evidence,
demonstrating Os-induced lung function effects and respiratory symptoms in some healthy
individuals following exposures down to 80 ppb. He also noted very limited new evidence
demonstrating such effects at exposure concentrations well below 80 ppb. In addition, the
Administrator noted (1) epidemiologic evidence of statistically significant associations with Os-
related health effects in areas that likely would have met the then-current standard; (2)
epidemiologic studies conducted in areas that likely would have violated the existing standard
but which nonetheless reported statistically significant associations that generally extended down
to ambient Os concentrations below the level of that standard; (3) the few studies that had
reported statistically significant associations with respiratory morbidity outcomes and mortality
in subsets of data that included only days with ambient Os concentrations below the level of the
existing standard; and (4) controlled human exposure studies, together with animal toxicological
studies, that provided considerable support for the biological plausibility of the respiratory
morbidity associations observed in the epidemiologic studies. Based on the available evidence,
the Administrator agreed with the CAS AC and the majority of public commenters that the
existing standard was not requisite to protect public health with an adequate margin of safety (73
FR 16471).
16 73 FR 16470-16471 (March 27, 2008)
17 Here, as in the ISA, the term "at-risk population" is used to encompass populations or lifestages that have a
greater likelihood of experiencing health effects related to exposure to an air pollutant due to a variety of factors;
other terms used in the literature include susceptible, vulnerable, and sensitive. These factors may be intrinsic, such
as genetic factors, lifestage, or the presence of preexisting diseases, or they may be extrinsic, such as socioeconomic
status (SES), activity pattern and exercise level, or increased pollutant exposures (U.S. EPA 2013, p. Ixx, 8-1, 8-2).
The courts and the Act's legislative history refer to these at-risk subpopulations as "susceptible" or "sensitive"
populations. See, e.g., American Lung Ass'n v. EPA. 134 F. 3d 388, 389 (D.C. Cir. 1998) ("NAAQS must protect
not only average health individuals, but also 'sensitive citizens' - children, for example, or people with asthma,
emphysema, or other conditions rendering them particularly vulnerable to air pollution" (quoting S. Rep. No. 91-
1196 at 10).
1-15
-------
Beyond this focus on the available health evidence, the Administrator also considered
estimates of Os exposures and health risks based on analyses where air quality was adjusted to
simulate just meeting the existing and potential alternative standards. For the various air quality
simulations, he specifically considered the pattern of estimated reductions in Os exposures across
health benchmark concentrations of 80, 70, and 60 ppb. The 80 ppb benchmark reflected an
exposure concentration for which there was strong evidence for respiratory effects in healthy
people, including airway inflammation, respiratory symptoms, airway hyperresponsiveness, and
impaired lung host defense (U.S. EPA, 2007, section 4.7). The 60 ppb benchmark reflected an
exposure concentration for which the Administrator judged the evidence of such effects to be
very limited (73 FR 16471).
The Administrator took note of the magnitudes of estimated health risks for a range of
health effects, including moderate and large lung function decrements, respiratory symptoms,
respiratory-related hospital admissions, and nonaccidental and cardiorespiratory mortality. He
recognized that these quantitative risk estimates for a limited number of specific health effects
were indicative of a much broader array of Os-related effects, including various indicators of
morbidity in at-risk populations that we could not analyze in the risk assessment (e.g., school
absences, increased medication use, emergency department visits). The Administrator concluded
that quantitative exposure and risk estimates, as well as the broader array of Os-related health
endpoints that could not be quantified, provided additional support for the evidence-based
conclusion that the existing standard needed to be revised (73 FR 16472).
Based on the above considerations, and consistent with CAS AC's unanimous conclusion
that there was no scientific justification for retaining the existing standard, the Administrator
concluded that the primary Os standard set in 1997 was not sufficient and thus not requisite to
protect public health with an adequate margin of safety. He further concluded that revision of
this standard was needed to provide increased public health protection (73 FR 16472).
Throughout the 2008 review, CASAC supported a standard level in the range of 60 to 70
ppb (without change to the form, indicator, or averaging time). In a letter to the Administrator on
the second draft Staff Paper, CASAC unanimously recommended "that the current primary
ozone standard be revised and that the level that should be considered for the revised standard be
from 0.060 to 0.070 ppm" (60 to 70 ppb) (Henderson, 2006, p. 5). This recommendation, based
in part on the placement of more weight on the evidence for effects following exposures to 60
ppb Os, followed from the CASAC's more general recommendation that the 1997 standard
needed to be made substantially more protective of human health, particularly for at-risk
populations. In a subsequent letter sent specifically to offer advice to aid the Administrator and
Agency staff in developing the 2007 Cb proposal, CASAC reiterated that Panel members "were
1-16
-------
unanimous in recommending that the level of the current primary ozone standard should be
lowered from 0.08 ppm to no greater than 0.070 ppm" (Henderson, 2007, p. 2).18
After considering CASACs comments, the Administrator judged that the appropriate
balance to draw, based on the entire body of evidence and information available in the 2008
review, was a standard set at a level of 75 ppb (and leaving all other elements of the NAAQS
unchanged). In making this decision, the Administrator placed primary emphasis on the body of
available scientific evidence, while viewing the results of exposure and risk assessments as
providing supporting information. Specifically, the Administrator judged that a standard set at
75 ppb would be appreciably below 80 ppb, the level in controlled human exposure studies at
which adverse effects had been demonstrated at the time, and would provide a significant
increase in protection compared to the then-current standard. Based on results of the exposure
assessment, he also noted that exposures to Os concentrations at and above a benchmark level of
80 ppb would be essentially eliminated with a standard level of 75 ppb, and that exposures at and
above a 70 ppb benchmark level would be substantially reduced or eliminated for the vast
majority of people in at-risk groups. In addition, the Administrator concluded that the body of
evidence did not support setting a lower standard level, specifically judging that the available
evidence for effects following exposures to Os concentrations of 60 ppb was "too limited to
support a primary focus at this level" (75 FR 2938). With respect to the epidemiologic evidence,
the Administrator stated that a standard set at a level lower than 75 ppb "would only result in
significant further public health protection if, in fact, there is a continuum of health risks in areas
with 8-hour average Cb concentrations that are well below the concentrations observed in the key
controlled human exposure studies and if the reported associations observed in the
epidemiological studies are, in fact, causally related to Os at those lower levels" (73 FR 16483).
In making his final decision about the level of the primary Os standard, the Administrator
noted that the level of 75 ppb was above the range recommended by CASAC (i.e., 70 to 60 ppb).
He concluded that "CASAC's recommendation appeared to be a mixture of scientific and policy
considerations" (75 FR 2992). The Administrator reached a different policy judgment than the
CASAC Panel, placing less weight than CASAC on the available controlled human exposure
studies reporting effects following exposures to 60 ppb Os and less weight on the results from
exposure and risk assessments, particularly on estimates of exposures to Os concentrations at or
above 60 ppb (73 FR 16482-3).
18 The D.C. Circuit, in its review of the 2008 primary standard, stated that it was unclear whether CASAC's advice
reflected issues of pure science or issues of science and policy. That is, the court was unable to determine whether
CASAC's conclusion in its 2007 letter that the standard be set no higher than 70 ppb "was based on its scientific
judgment that adverse effects would occur at that level or instead based on its more qualitative judgment that the
range it proposed would be more appropriately protective of human health with an adequate margin of safety."
Mississippi. 744 F. 3d at 1357.
1-17
-------
1.3.1.2 Approach for the Current Review
To identify the range of options appropriate for the Administrator to consider in the
current review, we apply an approach that builds upon the general approach used in the last
review (and in the 2010 reconsideration proposal) and that reflects the broader body of scientific
evidence, updated exposure/risk information, and advances in Os air quality modeling now
available. As summarized above, the Administrator's decisions in the prior review were based on
an integration of information on health effects associated with exposure to Os, judgments on the
adversity and public health significance of key health effects, and expert and policy judgments as
to when the standard is requisite to protect public health with an adequate margin of safety.
Staffs conclusions on the primary Os standard reflect our consideration of the available
scientific evidence, exposure/risk information, and air quality modeling information, within the
context of the overarching questions related to: (1) the adequacy of the current primary Os
standard to protect against effects associated with both short- and long-term exposures and (2)
potential alternative standards that are appropriate to consider in this review. In addressing these
broad questions, we organize the discussions in chapters 3 and 4 of this document around a series
of more specific questions reflecting different aspects of each overarching question. When
evaluating the health protection afforded by the current and potential alternative standards, we
take into account the four basic elements of the NAAQS: the indicator, averaging time, form,
and level.
Figure 1-1 below provides an overview of our approach in this review. We believe that
the general approach summarized in this section, and outlined in Figure 1-1, provides a
comprehensive basis to help inform the judgments required of the Administrator in reaching
decisions about the current and potential alternative primary Os standards. In the subsections
below, we describe our general approaches to considering the scientific evidence (evidence-
based considerations) and to considering the human exposure- and health risk information
(exposure- and risk-based considerations). We also recognize considerations related to ambient
Os attributable to background sources.
1-18
-------
Adequacy of Current 8-Hour Primary 03 Standard
Evidence-Based Considerations
> ISA weight-of-evidence conclusions for health effects and at-risk
populations
>Controlled human exposure and animal toxicology studies: Nature,
magnitude, likely adversity of effects; consistency across studies?
>Epidemiologic studies: Statistical precision; confounding; link effects to 03
air quality in specific locations?
>Health effect associations reported in locations meeting current
standard?
>Confidence in concentration-response relationships over distributions
of ambient concentrations, including concentrations below level of
current standard?
> Uncertainties in evidence for 03-attriubutable effects across distributions
of exposure concentrations and ambient concentrations
Exposure-/Risk-Based
Considerations
>Nature, magnitude, and importance of
estimated exposures and risks
associated with current 03 standard
> Focus on at-risk populations
> Uncertainties in the exposure and risk
estimates
Does information call
into question adequacy
of current 8-hour
Primary 03 standard?
Consider retaining
current 8-hour 03
standard
Consider Potential Alternative Standards
I
Indicator
>Support for retaining 03?
Averaging Time
>Support for 8-hour only?
>Support for different
averaging time?
Form
>Supportfor retaining annual 4th
highest?
Level
>Evidence-based considerations : Consider controlled human exposure studies and epidemiologic
studies, including uncertainties
>Exposure- and risk-based considerations: Consider exposure and risk reductions for alternative 03
standards, exposures and risks estimated to remain upon meeting alternatives, and uncertainties and
limitations in exposure/risk estimates
Identify range of potential alternative standards for consideration
Figure 1 -1. Overview of approach to reviewing the primary standard.
1-19
-------
1.3.1.2,1 Consideration of the Scientific Evidence
Our approach in this review draws upon an integrative synthesis of the entire body of
available scientific evidence for Os-related health effects, including the evidence newly available
in the current review and the evidence from previous reviews, as presented in the ISA (U.S.
EPA, 2013).19 Our approach to considering the scientific evidence is based fundamentally on
using information from controlled human exposure and epidemiologic studies, supplemented by
information from animal toxicology studies. Such evidence informs our consideration of the
health endpoints and at-risk populations on which to focus the current review, and our
consideration of the Os concentrations at which various health effects can occur.
Since the 2008 review of the Os NAAQS, the Agency has developed formal frameworks
for characterizing the strength of the scientific evidence with regard to health effects associated
with exposures to Os in ambient air and factors that may increase risk in some populations or
lifestages (U.S. EPA, 2013, Preamble; Chapter 8). These frameworks provide the basis for
robust, consistent, and transparent processes for evaluating the scientific evidence, including
uncertainties in the evidence, and for drawing weight-of-evidence conclusions on air pollution-
related health effects and at-risk populations.
With regard to characterization of health effects, the ISA uses a five-level hierarchy to
classify the overall weight-of-evidence into one of the following categories: causal relationship,
likely to be a causal relationship, suggestive of a causal relationship, inadequate to infer a causal
relationship, and not likely to be a causal relationship (U.S. EPA, 2013, Preamble Table II). In
using the weight of evidence approach to inform judgments about the degree of confidence that
various health effects are likely to be caused by exposure to Os, confidence increases as the
number of studies consistently reporting a particular health endpoint grows and as other factors,
such as biological plausibility and strength, consistency and coherence of evidence increases.
Conclusions about biological plausibility, consistency and coherence of Os-related health effects
are drawn from the integration of epidemiologic studies with mechanistic information from
controlled human exposure and animal toxicological studies, as discussed in the ISA (U.S. EPA,
2013, EPA Framework for Causal Determination, p. Iviii). In this PA, we place the greatest
19Selection of studies for inclusion in the ISA is based on the general scientific quality of the study, and
consideration of the extent to which the study is informative and policy-relevant. Policy relevant and
informative studies include those that provide a basis for or describe the relationship between the criteria
pollutant and effects. This includes studies that offer innovation in method or design and studies that
reduce uncertainty on critical issues, such as analyses of confounding or effect modification by
copollutants or other variables, analyses of concentration-response or dose-response relationships, or
analyses related to time between exposure and response. Review articles, by contrast, are generally not
included because they typically present summaries or interpretations of existing studies. The specific
criteria applied to the various types of studies are discussed in more detail in the Preamble to the ISA
(U.S. EPA, 2013, Preamble).
1-20
-------
weight on the health effects for which the evidence has been judged in the ISA to demonstrate a
causal or a "likely to be" causal relationship with Os exposures. Our consideration of the
available evidence for such effects is presented below in Chapter 3 (consideration of the
adequacy of the current standard) and in Chapter 4 (consideration of potential alternative
standards).
As discussed below, we further consider the evidence base assessed in the ISA with
regard to the types and levels of exposure at which health effects are indicated. This further
consideration of the evidence, which directly informs EPA's conclusions regarding the adequacy
of current or potential alternative standards in providing requisite public health protection, differs
from consideration of the evidence in the ISA with regard to overarching determinations of
causality. Therefore, studies that inform determinations of causality may or may not be
concluded to be informative with regard to the adequacy of the current or potential alternative
standards.20
As with health endpoints, the ISA's characterization of the weight-of-evidence for
potential at-risk populations is based on the evaluation and synthesis of evidence from across
scientific disciplines. The ISA characterizes the evidence for a number of "factors" that have the
potential to place populations at increased risk for Os-related effects. The categories considered
in evaluating the evidence for these potential at-risk factors are "adequate evidence," "suggestive
evidence," "inadequate evidence," and "evidence of no effect." These categories are discussed
in more detail in the ISA (U.S. EPA, 2013, chapter 8, Table 8-1). In this PA, we focus our
consideration of potential at-risk populations on those factors for which the ISA judges there is
"adequate" evidence (U.S. EPA, 2013, Table 8-6). At-risk populations are discussed in more
detail in section 3.2.1, below.
Using the available scientific evidence to inform conclusions on the current and potential
alternative standards is complicated by the recognition that a population-level threshold has not
been identified, below which it can be concluded with confidence that Cb-attributable effects do
not occur (U.S. EPA, 2013, section 2.5.4.4). In the absence of a discernible threshold, our
general approach to considering the available Os health evidence involves characterizing our
confidence in the extent to which Os-attributable effects occur, and the extent to which such
effects are adverse, over the ranges of Os exposure concentrations evaluated in controlled human
exposure studies and over the distributions of ambient Os concentrations in locations where
epidemiologic studies have been conducted. As noted above, we recognize that the available
20For example, as discussed further in this section and in Chapters 3 and 4 of this PA, we judge that health studies
evaluating exposure concentrations near or below the level of the current standard and epidemiologic studies
conducted in locations meeting the current standard are particularly informative when considering the adequacy of
the public health protection provided by the current standard.
1-21
-------
health effects evidence reflects a continuum from relatively high Os concentrations, at which
scientists generally agree that adverse health effects are likely to occur, through lower
concentrations, at which the likelihood and magnitude of a response become increasingly
uncertain. Aspects of our approach particular to evidence from controlled human exposure and
epidemiologic studies, respectively, are discussed below.
Controlled Human Exposure Studies
Controlled human exposure studies provide direct evidence of relationships between
pollutant exposures and human health effects (U.S. EPA, 2013, p.lx). Controlled human
exposure studies provide data with the highest level of confidence since they provide human
effects data under closely monitored conditions and can provide exposure response relationships.
Such studies are particularly useful in defining the specific conditions under which pollutant
exposures can result in health impacts, including the exposure concentrations, durations, and
ventilation rates under which effects can occur. As discussed in the ISA, controlled human
exposure studies provide clear and compelling evidence for an array of human health effects that
are directly attributable to acute exposures to Qiper se (i.e., as opposed to Os and other
photochemical oxidants, for which Os is an indicator, or other co-occurring pollutants) (U.S.
EPA, 2013, Chapter 6). Together with animal toxicological studies, which can provide
information about more serious health outcomes as well as the effects of long-term exposures
and mode of action, controlled human exposure studies also help to provide biological
plausibility for health effects observed in epidemiologic studies.
In this PA, we consider the evidence from controlled human exposure studies in two
ways. First, we consider the extent to which controlled human exposure studies provide evidence
for health effects following exposures to different Os concentrations, down to the lowest-
observed-effects levels in those studies. Second, we use such studies to inform our evaluation of
the extent to which we have confidence in health effect associations reported in epidemiologic
studies down through lower ambient Os concentrations, where the likelihood and magnitude of
Os-attributable effects become increasingly uncertain.
We consider the range of Os exposure concentrations evaluated in controlled human
exposure studies, including concentrations near or below the level of the current standard. We
consider both group mean responses, which provide insight into the extent to which observed
changes are due to Os exposures rather than to chance alone, and inter-individual variability in
responses, which provides insight into the fraction of the population that might be affected by
such Os exposures (U.S. EPA, 2013, section 6.2.1.1). When considering the relative weight to
place on various controlled human exposure studies, we consider the exposure conditions
evaluated (e.g., exercising versus resting, exposure duration); the nature, magnitude, and likely
adversity of effects over the range of reported Os exposure concentrations; the statistical
1-22
-------
precision of reported effects; and the consistency of results across studies for a given health
endpoint and exposure concentration. In addition, because controlled human exposure studies
typically involve healthy individuals and do not evaluate the most sensitive individuals in the
population (U.S. EPA, 2013, Preamble p. Ix), when considering the implications of these studies
for our evaluation of the current and potential alternative standards, we also consider the extent
to which reported effects are likely to reflect the magnitude and/or severity of effects in at-risk
groups.
Epidemiologic Studies
We also consider epidemiologic studies of short- and long-term Os concentrations in
ambient air. Epidemiologic studies provide information on associations between variability in
ambient Os concentrations and variability in various health outcomes, including lung function
decrements, respiratory symptoms, school absences, hospital admissions, emergency department
visits, and premature mortality (U.S. EPA, 2013, Chapters 6 and 7). Epidemiologic studies can
inform our understanding of the effects in the study population (which may include at-risk
groups) of real-world exposures to the range of Os concentrations in ambient air, and can provide
evidence of associations between ambient Os levels and serious acute and chronic health effects
that cannot be assessed in controlled human exposure studies. For these studies, the degree of
uncertainty introduced by confounding variables (e.g., other pollutants, temperature) and other
factors affects the level of confidence that the health effects being investigated are attributable to
Os exposures, alone and in combination with copollutants.
Available studies have generally not indicated a discernible population threshold, below
which Os is no longer associated with health effects (U.S. EPA, 2013, section 2.5.4.4). However,
the currently available epidemiologic evidence indicates decreased confidence in reported
concentration-response relationships for Os concentrations at the lower ends of ambient
distributions (U.S. EPA, 2013, section 2.5.4.4). Therefore, our general approach to considering
the results of epidemiologic studies within the context of the current and potential alternative
standards focuses on characterizing the range of ambient Os concentrations over which we have
the most confidence in Os-associated health effects, and the concentrations below which our
confidence in such health effect associations becomes appreciably lower. In doing so, we
consider the statistical precision of Os health effect associations reported in study locations with
various ambient Os concentrations; confidence intervals around concentration-response functions
reported over distributions of ambient Os (where available); and the extent to which the
biological plausibility of associations at various ambient Os concentrations is supported by
evidence from controlled human exposure and/or animal toxicological studies.
We consider both multi-city and single-city studies assessed in the ISA, each of which
have strengths and limitations. Multi-city studies evaluate large populations and provide greater
1-23
-------
statistical power than single-city studies. Multi-city studies also reflect Os-associated health
impacts across a range of diverse locations, providing spatial coverage for different regions and
reflecting differences in exposure-related factors that could impact Os risks. In addition,
compared to single-city studies, multi-city studies are less prone to publication bias and they
afford the possibility of generalizing to the broader national population (U.S. EPA, 2004, p. 8-
30). In contrast, while single-city studies are more limited than multicity studies in terms of
statistical power and geographic coverage, conclusions regarding the extent to which air quality
met the current or potential alternative standards in the cities for which associations have been
reported can be made with greater certainty for single-city studies (compared to multicity studies
reporting only multicity effect estimates) because the associations are reported for city-specific
analyses (U.S. EPA, 201 Id, section 2.3.4.1).21 In some cases, single-city studies can also
provide evidence for locations or population-specific characteristics not reflected in multicity
studies (U.S. EPA, 2013, section 6.2.7.1). Therefore, when considering available epidemiologic
studies we evaluate both multi-city and single-city studies, recognizing the strengths and
limitations of each.
In placing emphasis on specific epidemiologic studies, we focus on studies conducted in
the U.S. and Canada. Such studies reflect air quality and exposure patterns that are likely more
typical of the U.S. population than the air quality and exposure patterns reflected in studies
conducted outside the U.S. and Canada.22 We also focus on studies reporting associations with
effects judged in the ISA to be robust to confounding by other factors, including co-occurring air
pollutants.
1.3.1.2.2 Consideration of Exposure and Risk Estimates
To put judgments about Os-related health effects into a broader public health context, we
consider exposure and risk estimates from the HREA, which develops and applies models to
estimate human exposures to Os and Ch-related health risks in urban case study areas across the
United States (U.S. EPA, 2014a). The HREA estimates exposures of concern, based on
interpreting quantitative exposure estimates within the context of controlled human exposure
study results; lung function risks, based on applying exposure-response relationships from
controlled human exposure studies to quantitative estimates of exposures; and epidemiologic-
based risk estimates, based on applying concentration-response relationships drawn from
epidemiologic studies to adjusted air quality. Each of these types of assessments is discussed
briefly below.
21 Though in some cases multicity studies present single-city effect estimates in addition to multi-city estimates.
22 Though we recognize that a broader body of studies, including international studies, inform the causal
determinations in the ISA.
1-24
-------
As in the 2008 review, the HREA estimates exposures at or above benchmark
concentrations of 60, 70, and 80 ppb, reflecting exposure concentrations of concern based on the
available health evidence.23 Estimates of exposures of concern, defined as personal exposures
while at moderate or greater exertion to 8-hour average ambient Os levels, at or above these
discrete benchmark concentrations provide perspective on the public health impacts of Os-related
health effects that are plausibly linked to the more serious effects seen in epidemiological
studies, but cannot be evaluated in quantitative risk assessments. They also help elucidate the
extent to which such impacts may be reduced by meeting the current and alternative standards.
Estimates of the number of people likely to experience exposures of concern cannot be directly
translated into quantitative estimates of the number of people likely to experience specific health
effects due to individual variability in responsiveness. Only a subset of individuals can be
expected to experience such adverse health effects, and at-risk populations or lifestages, such as
people with asthma or children, are expected to be affected more by such exposures than healthy
adults. Though this analysis is conducted using discrete benchmark concentrations, health-
relevant exposures are more appropriately viewed as a continuum with greater confidence and
less uncertainty about the existence of health effects at higher Os exposure concentrations and
less confidence and greater uncertainty at lower exposure concentrations. This approach
recognizes that there is no sharp breakpoint within the exposure-response relationship for
exposure concentrations at and above 80 ppb down to 60 ppb.
The HREA also generates quantitative estimates of Os health risks for air quality adjusted
to just meet the current and potential alternative standards. As noted above, one approach to
estimating Os health risks is to combine modeled exposure estimates with exposure-response
relationships derived from controlled human exposure studies of Os-induced health effects. The
HREA uses this approach to estimate the occurrence of Os-induced lung function decrements in
at-risk populations in urban case study areas, including school-age children, school-age children
with asthma, adults with asthma, and older adults. The available exposure-response information
does not support this approach for other endpoints evaluated in controlled human exposure
studies (U.S. EPA, 2014a, section 2.2.5 to 2.2.7).
Another approach to estimating Os-associated health risks is to apply concentration-
response relationships derived from short- and/or long-term epidemiologic studies to air quality
adjusted to just meet current and potential alternative standards. The concentration-response
relationships drawn from epidemiologic studies are based on population exposure surrogates,
such as 8-hour concentrations averaged across monitors and over more than one day
23 For example, see 75 FR 2945-2946 (January 19, 2010) and 73 FR 16441-16442 (March 27, 2008) discussing
"exposures of concern".
1-25
-------
(incorporation of lag) (U.S. EPA, 2013, Chapter 6). The HREA presents epidemiologic-based
risk estimates for Os-associated mortality, hospital admissions, emergency department visits, and
respiratory symptoms (U.S. EPA, 2014a, Chapter 7). These estimates are derived from the full
distributions of ambient Os concentrations estimated for the study locations.24 In addition, the
HREA estimates mortality risks associated with various portions of distributions of short-term Cb
concentrations (U.S. EPA, 2014a). In this PA we consider risk estimates based on the full
distributions of ambient Os concentrations and, when available, estimates of the risk associated
with various portions of those ambient distributions.25 In doing so, we take note of the ISA
conclusions regarding confidence in linear concentration-response relationships over
distributions of ambient concentrations, and of the extent to which health effect associations at
various ambient Os concentrations are supported by the evidence from experimental studies for
effects following specific Os exposures.
1.3,1.2,3 Considerations Regarding Ambient Os Concentration Estimates
Attributable to Background Sources
As noted above, our approach in this review utilizes recent advances in modeling
techniques to estimate the contributions of U.S. anthropogenic, international anthropogenic, and
natural sources to ambient Os (discussed in detail in Chapter 2 of this PA). Such model estimates
can provide insights into the extent to which different types of background emissions sources
contribute to total ambient Os concentrations. Consideration of this issue in the current review is
informed by the approaches taken in previous reviews, as well as by court decisions in
subsequent litigation.
In 1979, the EPA set a 1-hour Os standard with a level of 0.12 ppm. Following the final
decision in that review, the City of Houston argued that the standard was arbitrary and capricious
because natural Os concentrations and other physical phenomena in the Houston area made the
standard unattainable in that area. The D.C. Circuit rejected this argument, stating that
attainability and technological feasibility are not relevant considerations in the promulgation of
the NAAQS. 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
24 In previous reviews, including the 2008 review and reconsideration, such risks were separately estimated for Os
concentrations characterized as above policy-relevant background concentrations. Policy-relevant background
concentrations were defined as the distribution of ozone concentrations attributable to sources other than
anthropogenic emissions of ozone precursor emissions (e.g., VOC, CO, NOx) in the U.S., Canada, and Mexico. The
decision to estimate total risk across the full range of Os concentrations reflects current OAQPS views and
consideration of advice from CAS AC (Frey and Samet, 2012b).
25In a series of sensitivity analyses, the HREA also evaluates a series of threshold models for respiratory mortality
associated with long-term Os concentrations. In this PA we consider these risk estimates based on threshold models,
in addition to HREA core estimates based on the linear model (sections 3.2.3.2, 4.4.2.3).
1-26
-------
locations and had addressed this difficulty through various compliance related provisions in the
Act. See API v. Costle. 665 F.2d 1176, 1184-6 (D.C. Cir. 1981).
More recently, in the 1997 review of the Os NAAQS, the Administrator set an 8-hour
standard with a level of 0.08 ppm (84 ppb). In reaching this decision, the EPA identified several
reasons supporting its decision to reject a more stringent standard of 0.07 ppm. Most
importantly, the EPA pointed out the scientific uncertainty at lower concentrations and placed
significant weight on the fact that no CASAC panel member supported a standard level set lower
than 0.08 ppm (62 FR 38868). In addition to noting the uncertainties in the health evidence for
exposure concentrations below 0.08 ppm and the advice of CASAC, the EPA noted that a
standard set at a level of 0.07 ppm would be closer to peak background concentrations that
infrequently occur in some areas due to nonanthropogenic sources of Os precursors (62 FR
38856, 38868; July 18, 1997).
In subsequent litigation, the D.C. Circuit upheld the EPA's decision as the product of
reasoned decision-making. The Court made clear that the most important support for the EPA's
decision was the health evidence and the concerns it raised about setting a standard level below
0.08 ppm. The Court also pointed to the significant weight that the EPA properly placed on the
advice it received from CASAC. Finally (as discussed in section 1.2.2 above), the Court noted
that the EPA could also consider relative proximity to peak natural background Os when
considering alternatives within the range of reasonable values supported by the scientific
evidence and judgments of the Administrator. See ATA III. 283 F.3d at 379 (D.C. Cir. 2002).
These cases provide a framework for considering the contributions of U.S.
anthropogenic, international anthropogenic, and natural sources within the context of considering
the health evidence and CASAC advice, when evaluating various potential alternative standards.
Consistent with such a framework, this PA identifies the range of policy options for the primary
Os standard that staff concludes are appropriate to consider in light of the available scientific
evidence and exposure/risk information, and the advice of CASAC. In identifying the range of
policy options supported by the evidence and information, staff has not considered proximity to
background Os concentrations. The Administrator, when evaluating the range of possible
standards that are supported by the scientific evidence, could consider proximity to background
O3 concentrations as one factor in selecting the appropriate standard.
1.3.2 Approach for the Secondary Standard
Staffs approach in this review of the current secondary standard takes into consideration
aspects of the approaches used in past Os NAAQS reviews. The past and current approaches,
generally described below, are both based fundamentally on using EPA's assessment of the
current scientific evidence and associated quantitative analyses to inform the Administrator's
1-27
-------
judgment regarding a secondary standard for Os that is requisite (i.e., neither more nor less
stringent than necessary) to protect public welfare.
In reaching conclusions on options for the Administrator's consideration, we note that the
final decision to retain or revise the current secondary Os standard is a public welfare policy
judgment to be made by the Administrator. This final decision will draw upon the available
scientific evidence for Os-attributable welfare effects and on analyses of vegetation and
ecosystem exposures and public welfare risks based on impacts to vegetation, ecosystems and
their associated services, including judgments about the appropriate weight to place on the range
of uncertainties inherent in the evidence and analyses. In determining the requisite level of
protection for crops and trees, the Administrator will need to weigh the importance of the
predicted risks of these effects in the overall context of public welfare protection, along with a
determination as to the appropriate weight to place on the associated uncertainties and limitations
of this information. Our general approach to informing these judgments, discussed more fully
below, recognizes that the available evidence demonstrates a range of Os sensitivity across
studied plant species and documents an array of Os-induced effects that extend from lower to
higher levels of biological organization. These effects range from those affecting cell processes
and individual plant leaves to effects on the physiology of whole plants, species effects and
effects on plant communities to effects on related ecosystem processes and services. Given this
evidence, it is not possible to generalize across all studied species regarding which cumulative
exposures are of greatest concern, as this can vary by situation due to differences in exposed
species sensitivity, the importance of the observed or predicted Os-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. At the same time, the evidence also demonstrates
that though effects of concern can occur at very low exposures in sensitive species, at higher
cumulative exposures those effects would likely occur at a greater magnitude and/or higher
levels of biological organization and additional species would likely be impacted. It is important
to note, however, due to the variability in the importance of the associated ecosystem services
provided by different species at different exposures and in different locations, as well as
differences in associated uncertainties and limitations, adverse effects observed or predicted at
lower exposures along the exposure continuum may or may not have less public welfare
significance than those observed at higher cumulative exposures. Therefore, in developing
conclusions in this final PA, we take note of the complexity of judgments to be made by the
Administrator regarding the adversity of known and anticipated effects to the public welfare and
are mindful that the Administrator's ultimate judgments on the secondary standard will most
appropriately reflect an interpretation of the available scientific evidence and exposure/risk
1-28
-------
information that neither overstates nor understates the strengths and limitations of that evidence
and information.
Section 1.3.2.1 below provides an overview of the general approach taken in the last
review of the secondary standard for Os (i.e., the 2008 review), and a summary of the rationale
for the decision on the standard in that review (73 FR 16436). Section 1.3.2.2 presents our
approach in the current review, including our approach to considering the vegetation effects
evidence and exposure/risk information, and considerations regarding ambient Os concentrations
attributable to background sources.
1.3.2.1 Approach Used in the Last Review
In the 2008 review of the secondary NAAQS for Os, the Administrator relied upon
consideration of the available scientific evidence and exposure/risk information, information
regarding biologically-relevant exposure indices, air quality information regarding the degree of
overlap between different exposure index forms, the advice and recommendations of CAS AC,
considerations regarding adversity, and comments from the public. Based on all of this, he
revised the level of the secondary Os standard from 0.08 ppm26 to 0.075 ppm (75 ppb27).
In reaching a decision to revise the 1997 8-hour secondary standard, the Administrator
found, after carefully considering the public comments, that the fundamental scientific
conclusions on the effects of Os on vegetation and sensitive ecosystems reached in the 2006
Criteria Document and 2007 Staff Paper, as discussed in section IV. A of the final rule remained
valid (73 FR 16496). He further recognized that several additional lines of evidence had
progressed sufficiently since the 1997 review to provide a more complete and coherent picture of
the scope of Os-related vegetation risks (i.e., visible foliar injury, tree biomass loss, crop yield
loss, and others), especially those faced by sensitive seedling, sapling and mature growth stage
tree species growing in field settings, and their associated forested ecosystems. This new
research reflected an increased emphasis on field-based exposure methods (e.g., free-air, ambient
gradient and biomonitoring surveys) (73 FR 16490) in addition to the more traditional controlled
open-top chamber (OTC) studies (73 FR 16485), and began to address one of the key data gaps
cited by the Administrator in the 1997 review (73 FR 16486). Specifically, by providing
additional evidence that Os-induced crop yield loss and tree seedling biomass loss effects
observed in chambers also occurs in the field, this new research qualitatively increased support
26 As noted earlier, due to rounding convention, the 1997 standard level of 0.08 ppm corresponded to 0.084 ppm (84
ppb).
27 As explained above, the level of the O3 standard is specified as 0.075 ppm rather than 75 ppb. However, in this
draft PA we refer to ppb, which is most often used in the scientific literature and in the ISA, in order to avoid the
confusion that could result from switching units when discussing the evidence in relation to the standard level.
1-29
-------
for, and confidence in, the continued use of OTC-derived crop and tree seedling exposure-
response (E-R) functions developed in the National Crop Loss Assessment Network (NCLAN)
and National Health and Environmental Effects Research Laboratory - Western Ecology
Division (NHEERL-WED) studies, respectively, to predict Os-induced impacts on crops and tree
seedlings in the field (72 FR 37886). All of these areas were considered together, along with
associated uncertainties, in an integrated weight-of-evidence approach (73 FR 16490).
Beyond the available vegetation effects evidence, the Administrator also considered
estimates of Os exposures and risks when air quality was adjusted to simulate just meeting the
existing and potential alternative standards. On the basis of these assessments, the Administrator
concluded that Cb exposures that would be expected to remain after meeting the existing
standard would be sufficient to cause visible foliar injury and seedling and mature tree biomass
loss in Os-sensitive vegetation (73 FR 16496) and would still allow Os-related yield loss to occur
in some commodity crop species and fruit and vegetable species grown in the U.S. (73 FR
16489). Other Os-induced effects described in the literature, including an impaired ability of
many sensitive species and genotypes within species to adapt to or withstand other
environmental stresses, such as freezing temperatures, pest infestations and/or disease, and to
compete for available resources, would also be anticipated to occur. In the long run, the result of
these impairments (e.g., loss in vigor) could lead to premature plant death in Os sensitive species.
Though effects on other ecosystem components had only been examined in isolated cases, the
Administrator noted effects such as those described above could have significant implications for
plant community and associated species biodiversity and the structure and function of whole
ecosystems (73 FR 16496).
Although the Administrator concluded that the then-current standard was not sufficient to
protect against the known and anticipated effects described above, he also recognized that the
secondary standard is not meant to protect against all known observed or anticipated Os-related
effects, but only those that can reasonably be judged to be adverse to the public welfare. The
Administrator found that the degree to which such effects should be considered to be adverse
depended on the intended use of the vegetation and its significance to the public welfare (73 FR
16496). In this regard, he 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. Based on these considerations, and taking into consideration the advice and
recommendations of CASAC, the Administrator concluded that the protection afforded by the
existing standard was not sufficient, and that the standard needed to be revised to provide
1-30
-------
additional protection from known and anticipated adverse effects on sensitive natural vegetation
and ecosystems (73 FR 16497).
Given this judgment, the Administrator then considered what revisions to the standard
were requisite to protect public welfare. Regarding the form of the standard, the Administrator
took note that at the conclusion of the 1997 review, the biological basis for a cumulative,
seasonal form was not in dispute28 and that the 2006 Criteria Document also concluded that Os
exposure indices that cumulate differentially-weighted hourly concentrations are the best
candidates for relating exposure to plant growth responses (EPA, 2006) (61 FR 65716; 73 FR
16486). The CAS AC, in its letter to the Administrator following its review of the second draft
Staff Paper, stated that "there is a clear need for a secondary standard which is distinctly
different from the primary standard in averaging time, level and form" and that "the CASAC
unanimously agrees that it is not appropriate to try to protect vegetation from the substantial,
known or anticipated, direct and/or indirect, adverse effects of ambient ozone by continuing to
promulgate identical primary and secondary standards for ozone" (Henderson, 2006, pp. 5-7).
Although many possible cumulative, seasonal concentration-weighted exposure metrics exist, the
Staff Paper and the CASAC Panel concluded that the W12629 form is the most biologically-
relevant cumulative, seasonal form appropriate to consider in the context of the secondary
standard review (73 FR 16486-87).30
Although agreeing with the Criteria Document, Staff Paper and CASAC conclusions that
a cumulative exposure index that differentially weights Os concentrations could represent a
reasonable policy choice for a seasonal secondary standard to protect against the effects of Os on
vegetation and that the most appropriate cumulative, concentration-weighted form to consider
was the sigmoidally-weighted W126 form (73 FR 16498), the Administrator also took note of
the 1997 decision to make the revised secondary standard identical to a revised primary standard
after similar considerations (73 FR 16498). In considering the rationale for the 1997 decision, the
Administrator observed that it was based in part on an analysis that compared the degree of
28 In the 1997 review, a different cumulative metric (SUM06) was proposed.
29 W126 is a cumulative exposure index that is biologically based. The W126 index focuses on the higher hourly
average concentrations, while retaining the mid-and lower-level values. It is defined as the sum of sigmoidally-
weighted hourly Os concentrations over a specified period, where the daily sigmoidal weighting function is defined
as: l-
30 In a subsequent letter offering unsolicited advice to the Administrator and Agency staff on development of the
proposed rulemaking, the CASAC reiterated that Panel members ' 'were unanimous in supporting the
recommendation in the Final Ozone Staff Paper that protection of managed agricultural crops and natural terrestrial
ecosystems requires a secondary Ozone NAAQS that is substantially different from the primary ozone standard in
averaging time, level and form". ..and "[tjhe recommended metric for the secondary ozone standard is the
(sigmoidally-weighted) W126 index, accumulated over at least the 12 'daylight' hours and over at least the three
maximum ozone months of the summer 'growing season" (Henderson, March 26, 2007, p.3).
1-31
-------
overlap in county-level air quality measured in terms of alternative standard forms (62 FR
38876). Recognizing that significant uncertainty remained in 1997 regarding conclusions drawn
from such analyses, the Administrator also considered the results of a similar analysis of recent
monitoring data undertaken in the 2007 Staff Paper to assess the degree of overlap expected
between the existing standard (4th high, daily maximum 8-hour concentration averaged over
three years) and potential alternative standards based on W126 cumulative seasonal forms.
The Administrator noted that this analysis showed significant overlap between the 8-hour
secondary standard and selected levels of W126 standard forms, with the degree of overlap
between these potential alternative standards depending greatly on the W126 level selected and
the distribution of hourly Os concentrations within the annual and/or 3-year average period.
From this analysis, the Administrator recognized that a secondary standard set identical to a
revised primary standard would provide a significant degree of additional protection for
vegetation as compared to that provided by the existing secondary standard. In further
considering the significant uncertainties in the available body of evidence and in the exposure
and risk analyses, and the difficulty in determining at what point various types of vegetation
effects become adverse for sensitive vegetation and ecosystems, the Administrator focused his
consideration on a level for an alternative W126 standard (with an annual form) at the upper end
of the proposed range (i.e., 21 ppm-hours). The Staff Paper analysis showed that at a W126 level
of 21 ppm-hours, there would be essentially no counties with air quality expected both to exceed
such an alternative W126 standard and to meet the revised 8-hour primary standard—that is,
based on this analysis of counties with ambient Os monitors, a W126-based level of 21 ppm-
hours would be unlikely to provide additional protection in any areas beyond that likely to be
provided by the revised 2008 primary standard (73 FR 16499/500).
The Administrator also considered the Staff Paper finding that the degree of overlap
between counties (with areas of concern for vegetation) expected to meet an 8-hour level for the
form of the existing standard and potential alternative levels of a W126-based standard was
inconsistent across years analyzed. This variation depended greatly on levels selected for a
W126-based standard and a 3-year average 4th high daily maximum 8-hour standard,
respectively, and the distribution of hourly Os concentrations within the annual and/or 3-year
average period. From this, the Staff Paper recognized the need for caution in evaluating the
likely vegetation impacts associated with a given level of air quality expressed in terms of the
existing 8-hour average standard in the absence of parallel W126 information. In considering
these findings, the Administrator "recognize[d] that the general lack of rural monitoring data
made uncertain the degree to which the revised 8-hour standard or an alternative W126 standard
would be protective, and that there was the potential for not providing the appropriate degree of
protection for vegetation in areas with air quality distributions that resulted in a high cumulative,
1-32
-------
seasonal exposure but did not result in high 8-hour average exposures" (73 FR 16500). With
regard to the 8-hour standard, he also noted that "[w]hile this potential for under-protection was
clear, the number and size of areas [then] at issue and the degree of risk [was] hard to determine.
However, such a standard would also tend to avoid the potential for providing more protection
than is necessary, a risk that would have arisen from moving to a new form for the secondary
standard despite the significant uncertainty in determining the degree of risk for any exposure
level and the appropriate level of protection, as well as uncertainty in predicting exposure and
risk patterns" (73 FR 16500).
Thus, although the Administrator agreed with the views and recommendations of
CASAC that a cumulative, seasonal standard was the most biologically relevant way to relate
exposure to plant growth response, he also recognized that there remained significant
uncertainties in determining or quantifying the degree of risk attributable to varying levels of Os
exposure, the degree of protection that any specific cumulative, seasonal standard would
produce, and the associated potential for error in determining the secondary standard that would
provide a requisite degree of protection—i.e., sufficient but not more than what is necessary.
Given these significant uncertainties, the Administrator concluded that establishing a new
secondary standard with a cumulative, seasonal form, at that time, would have resulted in
uncertain benefits beyond those afforded by the revised primary standard, and therefore, might
have been more than necessary to provide the requisite degree of protection (73 FR 16500).
Based on his consideration of these issues (73 FR 16497), the Administrator judged that the
appropriate balance to be drawn was to set a secondary standard identical in every way to the
revised 8-hour primary standard of 0.075 ppm. The Administrator believed that such a standard
would be sufficient to protect public welfare from known or anticipated adverse effects, and did
not believe that an alternative cumulative, seasonal standard was needed to provide this degree of
protection (73 FR 16500).
As noted above, on July 23, 2013 the D.C. Circuit found this approach to be contrary to
law because the EPA had failed to identify a level of air quality requisite to protect public
welfare and, therefore, the EPA's comparison between the primary and secondary standards for
determining if requisite protection for public welfare was afforded by the primary standard was
inherently arbitrary. The court remanded the secondary standard to the EPA for further
consideration. 744 F. 3d at 1360-62.
1.3.2.2 Approach for the Current Review
To identify the range of options appropriate for the Administrator to consider in the
current review, we apply an approach that builds upon the general approach used in the 2008
review (and in the 2010 reconsideration proposal), and that reflects the broader body of scientific
1-33
-------
evidence, updated exposure/risk information, and advances in Os air quality modeling now
available. As summarized above, the Administrator's decisions in the prior review were based on
an integration of information on welfare effects associated with exposure to Os, judgments on the
adversity and public welfare significance of key effects, and, expert and policy judgments as to
when the standard is requisite to protect public welfare. These considerations were informed by
air quality and related analyses, quantitative exposure and risk assessments, and qualitative
assessment of impacts that could not be quantified. In performing the evaluation in this
document, we are additionally mindful of the recent remand of the secondary standard by the
D.C. Circuit and our approach in the current review incorporates our response to this remand.
Our approach in this review of the secondary Os standard also reflects our consideration
of the available scientific evidence, information on biologically-relevant exposure indices,
exposure/risk information, and air quality modeling information, within the context of
overarching questions related to: (1) the adequacy of the current secondary Os standard to protect
against effects associated with cumulative, seasonal exposures and (2) potential alternative
standards, if any, that are appropriate to consider in this review. In addressing these broad
questions, we have organized the discussions in chapters 5 and 6 of this document around a
series of more specific questions reflecting different aspects of each overarching question. When
evaluating the welfare protection afforded by the current or potential alternative standards, we
take into account the four basic elements of the NAAQS: the indicator, averaging time, form,
and level.
Figure 1-2 below provides an overview of our approach in this review. We believe that
the general approach summarized in this section, and outlined in Figure 1-2, provides a
comprehensive basis to help inform the judgments required of the Administrator in reaching
decisions about the current and potential alternative secondary Os standards. In the subsections
below, we summarize our general approaches to considering the scientific evidence (evidence-
based considerations) and to considering the exposure and risk information (exposure- and risk-
based considerations).
1-34
-------
Adequacy of Current 8-Hour Secondary 03 Standard
Evidence-Based Considerations
>ISA conclusions on causality
>Evidence for vegetation effects from cumulative
exposures allowed by the current standard
>Evidence for vegetation effects in locations that would
likely have met the current standard
Exposure-/Risk-Based Considerations
>Nature, magnitude, and importance of
estimated exposures and risks associated with
current 03 standard
Uncertainties in the exposure and risk
estimates?
Joes informatio^
call into question
the adequacy and
appropriateness of
currentS-hour
secondary 03
standard?
NO
Consider retaining
current 8-hour 03
standard
YES
Consider Potential Alternative Standards
Indicator
>Support for retaining 03?
Form
>Support for cumulative, seasonal
form?
>Support for peak-weighted form?
Averaging Time
>Supportfor 12 hour diurnal period?
>Support for 3-month seasonal
period?
>Supportfor annual vs. 3-year period?
Level
>Evidence-based considerations: Consider 03 exposure concentrations in OTC and field-based studies,
including uncertainties
>Exposure and risk considerations: Consider the nature, magnitude, and importance of estimated
exposures and risks associated with potential alternative 03 standards, 03 air quality projections across
U.S. and case study locations, including uncertainties
Identify range of potential alternative standards for consideration
Figure 1-2. Overview of approach to reviewing the secondary standard.
1-35
-------
1.3.2.2,1 Consideration of the Scientific Evidence
Our approach in this review draws upon an integrative synthesis of the entire body of
available scientific evidence for Os-related welfare effects, including the evidence newly
available in the current review and the evidence from previous reviews, as presented in the ISA
(U.S. EPA, 2013). Our approach to considering the scientific evidence for effects on vegetation
is based fundamentally on using information from controlled chamber studies and field-based
studies. Such evidence informs our consideration of welfare endpoints and at-risk species and
ecosystems on which to focus the current review, and our consideration of the ambient Os
conditions under which various welfare effects can occur.
As in each NAAQS review, we consider the entire body of evidence for the subject
criteria pollutant. With regard to identification of the welfare effects that could be caused by a
pollutant, we look to controlled exposure studies using chamber or free air methodologies and
field-based observational, survey and gradient studies. Evaluating all of the evidence together,
the ISA makes a determination with regard to the strength of the evidence for a causal
relationship between the air pollutant and specific welfare effects. These determinations inform
our identification of welfare effects for which the NAAQS may provide protection.
Since the 2008 review of the Os NAAQS, the Agency has developed a formal framework
for characterizing the strength of the scientific evidence with regard to a causal relationship
between ambient Os and welfare effects (U.S. EPA, 2013, Preamble; Chapter 9). This framework
provides the basis for a robust, consistent, and transparent process for evaluating the scientific
evidence, including uncertainties in the evidence, and for drawing weight-of-evidence
conclusions regarding air pollution-related welfare effects. In so doing, the ISA uses a five-level
hierarchy, classifying the overall weight of evidence into one of the following categories: causal
relationship, likely to be a causal relationship, suggestive of a causal relationship, inadequate to
infer a causal relationship, and not likely to be a causal relationship (U.S. EPA, 2013, Preamble
Table II). In our approach here, we place the greatest weight on the evidence for welfare effects
that have been judged in the ISA to be caused by, or likely to be caused by, Os exposures. Our
consideration of the available evidence for such effects is presented below in Chapter 5
(consideration of the adequacy of the current standard) and in Chapter 6 (consideration of
potential alternative standards).
We further consider the evidence base, as assessed in the ISA, with regard to the types
and levels of exposure at which welfare effects are indicated. This further consideration of the
evidence base, which directly informs the EPA's conclusions regarding the adequacy of current
or potential alternative standards in providing requisite public welfare protection, differs from
consideration of the evidence in the ISA with regard to overarching determinations of causality.
1-36
-------
Studies that have informed determinations of causality may or may not be concluded to be
informative with regard to the adequacy of the current or potential alternative standards.
Our approach in this review, as in past reviews, included recognition that the available
evidence has not provided identification of a threshold in exposure or ambient Os concentrations
below which it can be concluded with confidence that Os-attributable vegetation effects do not
occur across the broad range of Os-sensitive plant species growing within the U.S. This is due in
part to the fact that research shows that there is variability in sensitivity between and within
species and that numerous factors, i.e. chemical, physical, biological, and genetic, can influence
the direction and magnitude of the studied effect (U.S. EPA, 2013, section 9.4.8). In the absence
of a discernible threshold, our general approach to considering the available Os welfare evidence
involves characterizing our confidence in conclusions regarding Os-attributable vegetation
effects over the ranges of cumulative seasonal Cb exposure values evaluated in chamber studies
and in field studies in areas where Os-sensitive vegetation are known to occur, as well as
characterizing the extent to which these effects can be considered adverse. In addition, because
Os can indirectly affect other ecosystem components (such as soils, water, and wildlife, and their
associated goods and services, through its effects on vegetation) our approach also considers
those indirect effects for which the ISA concludes, based on multiple lines of evidence, including
mechanistic and physiological processes, to have a causal or likely to be a causal relationship.
With respect to ecosystem services for which we may have only limited or qualitative
information regarding an association with Os exposures, our approach is to consider their policy-
relevance in the context of section 109(b) (2) of the CAA which specifies that secondary
standards provide requisite protection of "public welfare from any ... known or anticipated
adverse effects associated with the presence of [the] pollutant in the ambient air". As noted
above, our approach to informing these judgments, discussed more fully below, recognizes that
the available welfare effects evidence demonstrates a wide range in Os sensitivities across
studied plant species. As a result, at relatively high cumulative Os exposures, a greater number
of plant species will show effects, and the magnitude of the observed effects will be greater,
particularly on the more sensitive species, while at lower cumulative Os exposures, fewer species
will demonstrate effects and the magnitude of those observed effects will be less.
In this review, the evidence base includes quantitative information across a broad array of
vegetation effects (e.g., growth impairment during seedlings, saplings and mature tree growth
stages, visible foliar injury, and yield loss in annual crops) and across a diverse set of exposure
methods from laboratory and field studies. These methods include the more traditional OTC
studies, as well as field-based exposure studies. While we consider the full breadth of
information available, we place greater weight on U.S. studies due to the often species-, site-,
and climate-specific nature of Os-related vegetation responses. We especially weight those
1-37
-------
studies that include Os exposures that fall within the range of those likely to occur in the ambient
air. Further, our approach in the context of the quantitative exposure and risk assessments
(discussed below), places greatest emphasis on studies that have evaluated plant response over
multiple exposure levels and developed exposure-response relationships that allow the prediction
(estimation) of plant responses over the range of potential alternative standards being assessed.
In considering the evidence, we recognize differences across different study types in what
information they provide. For example, because conditions can be controlled in laboratory
studies, responses in such studies may be less variable and smaller differences may be easier to
detect. However, the control conditions may limit the range of responses or incompletely reflect
pollutant bioavailability, so they may not reflect responses that would occur in the natural
environment. Alternatively, field data can provide important information for assessments of
multiple stressors or where site-specific factors significantly influence exposure. They are also
often useful for analyses of larger geographic scales and higher levels of biological organization.
However, because most field study conditions cannot be controlled, variability is expected to be
higher and differences harder to detect. The presence of confounding factors can also make it
difficult to attribute observed effects to specific stressors.
In considering information from across multiple lines of evidence, our approach is to first
integrate the evidence from both controlled and field-based studies and assess the coherence and
consistency across the available evidence for each effect. We then consider the extent to which
these identified effects should be considered adverse to the public welfare, relying largely on the
paradigm used in the 2008 review and 2010 proposed reconsideration (e.g., 75 FR 3006). This
paradigm recognizes that the significance to the public welfare of Os-induced effects on sensitive
vegetation growing within the U.S. can vary 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. Accordingly, any given Os-related effect on vegetation
and ecosystems (e.g., biomass loss, crop yield loss, foliar injury) may be judged to have a
different degree of impact on the public welfare depending, for example, on whether that effect
occurs in a Class I area, a city park, or commercial cropland. Our approach takes this variation in
the significance of Os-related vegetation effects into account in evaluating the currently available
evidence with regard to the extent to which it calls into question the adequacy of the current
standard and, as appropriate, indicates potential alternative standards that would be appropriate
for the Administrator to consider. In the 2010 proposed reconsideration, the Administrator
proposed to place the highest priority and significance on vegetation and ecosystem effects to
sensitive species that are known to or are likely to occur in federally protected areas such as
national parks and other Class I areas, or on lands set aside by States, Tribes and public interest
groups to provide similar benefits to the public welfare (75 FR 3023/24). Our approach in this
1-38
-------
review considers whether newly available information would suggest any evolution to this
paradigm, in particular in the context of considering associated ecosystem services.
Finally, our approach continues to give great weight to the scientific evidence available in
this and previous reviews indicating the relevance of cumulative, seasonal, concentration-
weighted exposures in inducing vegetation effects. More specifically, in the 2008 and 2010
reviews, the EPA concluded and the CAS AC agreed that the W126 cumulative exposure metric
was the most appropriate to use in this review to evaluate both the adequacy of the current
secondary standard and the appropriateness of any potential revisions. As discussed in chapter 5
in this PA, the information available in this review continues to support the use of such a metric
and does not call into question the appropriateness of using the W126 metric in this context.
Therefore, both the WREA and PA continue to express exposures in terms of the W126 index,
and continue to consider the important policy implications regarding selection of an appropriate
exposure index for vegetation. Our approach also places primary emphasis on studies that
evaluated plant response to exposures that were or can be described using such an index. The
policy-relevant discussions in chapters 5 and 6 focus on vegetation effects evidence and
exposure/risk information that can be associated with cumulative, seasonal peak-weighted
exposures, where possible. Discussions pertaining to the adequacy of the current secondary
standard will consider what cumulative seasonal exposures would be allowed under air quality
that would just meet the current standard.
1.3.2.2.2 Consideration of Exposure and Risk Estimates and Air Quality Analyses
To put judgments about Os-related vegetation and ecosystem effects and services into a
broader public welfare context, we consider national scale exposure and risk assessments
described in the WREA (U.S. EPA, 2014b). We particularly focused on the WREA quantitative
risks related to three types of vegetation effects: foliar injury, biomass loss, and crop yield loss.
These risks were assessed in a range of WREA analyses variously involving recent Os
monitoring data and/or national-scale adjusted air quality scenarios for the current secondary
standard and, in some analyses, for a cumulative, seasonal W126 form at one or more levels (15,
11 and 7 ppm-hours). Our consideration of these WREA results provide insight into the extent to
which the current or potential alternative standards would be expected to maintain distributions
of cumulative, seasonal Os exposures below those associated with adverse vegetation effects.
With regard to quantitative Os risks related to welfare effects and ecosystem services for
foliar injury, we consider two main analyses in the WREA: a screening-level assessment of 214
National Parks and a case study focused on three National Parks. In the screening-level
assessment, Os concentrations in national parks are assessed using criteria developed from a U.S.
Forest Service nationwide dataset on foliar injury, ambient Os concentrations (in terms of W126
1-39
-------
index) and soil moisture (which can influence susceptibility of vegetation to foliar injury).
Additionally, we consider a case study for three Class I areas (Great Smoky Mountain National
Park, Rocky Mountain National Park, and Sequoia/Kings Canyon National Park). We consider
results from this case study for three metrics: 1) percent of vegetation cover affected by foliar
injury; 2) percent of trails affected by foliar injury; 3) estimates of species specific biomass loss
within the case study area. We also consider qualitative analyses on ecosystem services effects
for this endpoint. For example, the WREA uses GIS mapping to illustrate where effects may be
occurring and relates those areas to national scale statistics for recreational use and data on
hiking trails, campgrounds and other park amenities that intersect with potentially affected areas.
These are used to identify impacts on ecosystem services associated with recreation in national
parks. We additionally consider analyses showing associations between elevated Cb
concentrations and increased vulnerability to fire risk regimes, insect attacks and impacts on
hydrological cycles.
With regard to risks related to biomass and crop yield loss, we consider WREA results
based on exposure-response functions for tree and crop species that predict the growth or yield
response of each species, based on the exposure patterns estimated within its growing region. To
compare exposure-response across species, genotypes or experiments for which absolute
response values may vary greatly, the WREA instead uses estimates of relative biomass loss for
trees or yield loss for crops. The WREA develops such estimates nationally and separately for
more than 100 federally designated Class I areas. Additionally, we consider WREA-developed
estimates of associated impacts on the agriculture and forestry sectors quantifying how Os
exposure to vegetation is estimated to affect the provision of timber and crops and carbon
sequestration. We consider estimates for impacts related to tree biomass loss on ecosystem
services such as pollution removal, carbon storage and sequestration in five urban case study
areas. We consider biomass and crop yield loss estimates in light of advice from CASAC, as
discussed in sections 5.3 and 5.4 below.
In considering the amount of weight to place on the estimates of exposures and risks at or
above specific W126 values described in the WREA, our approach: 1) evaluates the weight of
the scientific evidence concerning vegetation effects associated with those Os exposures; 2)
considers the importance, from a public welfare perspective, of the Os-induced effects on
sensitive vegetation and associated ecosystem services that are known or anticipated to occur as
a result of exposures at selected W126 values; and, 3) recognizes that predictions of effects
associated with any given Os exposure may be mitigated or exacerbated by actual conditions in
the field (i.e., co-occurring modifying environmental and genetic factors). When considering
analyses in the WREA that involve discrete exposure levels or varying levels of severity of
effects, our approach to informing these judgments recognizes that the available welfare effects
1-40
-------
evidence demonstrates a wide range in Cb sensitivities across studied plant species. As a result,
at relatively high cumulative Os exposures, a greater number of plant species will show effects,
and the magnitude of the observed effects will be greater, particularly on the more sensitive
species, while at lower cumulative Os exposures, fewer species will demonstrate effects and the
magnitude of those observed effects will be less. We recognize that there is no sharp breakpoint
along this continuum of effects incidence and severity, ranging from concentrations at and above
the level of the current secondary standard down to the lowest cumulative, seasonal W126 value
assessed. In considering these results in this PA, we consider both the potential for welfare
effects and their severity and our understanding of the likelihood of such effects at different Os
exposures.
1.3.2.2.1 Considerations Regarding Ambient Os Concentration Estimates
Attributable to Background Sources
As noted above, our approach in this review utilizes recent advances in modeling
techniques to estimate the contributions of U.S. anthropogenic, international anthropogenic, and
natural sources to ambient Os (discussed in detail in Chapter 2 of this document). Such model
estimates can provide insights into the extent to which different types of emissions sources
contribute to total ambient Os concentrations. Our consideration of this issue in the current
review is informed by the approaches taken in previous reviews, and by court decisions on
subsequent litigation, as discussed in section 1.3.1.2.3 above. Further, in the 1996 proposal, Os
background concentrations were one of the factors the Administrator considered in selecting the
SUM06 index as a form for an alternative secondary standard. This and other cumulative
exposure indices under consideration were judged to be equally capable at estimating exposures
relevant to vegetation, given the lack of evidence for a discernible threshold for vegetation
effects in general (U.S. EPA 1996, p. 225), which might have provided a scientific basis for
selecting among different cumulative exposure indices. At that time, the SUM06 metric was
selected over the W126 metric because it focused on the policy-relevant (above background)
portion of the total cumulative seasonal exposures reaching plants (62 FR 38856). At the
conclusion of that review, the Administrator ultimately chose to set the secondary standard
identical to the primary standard, including using the 8-hour average instead of a cumulative
seasonal form (62 FR 38868). In the 2008 review, staff analyses concluded that the W126 index
was more biologically-relevant based on the available science; staff additionally noted, based on
then-available estimates of background, that this form was also not likely to be significantly
impacted by background concentrations given the very low weight assigned to lower Cb
concentrations by the W126 index (U.S. EPA, 2007 SP, 7-22; 72 FR 37893). In this review, the
degree to which the total value of the W126 index could be contributed by background
concentrations was again assessed. Based on a limited analysis, described in chapter 2 of the
1-41
-------
PA, background Os (BGO3) can comprise a non-negligible portion of W126 across the U.S.,
have the greatest contributions to W126 in the intermountain western U.S., and because of the
sigmoidal weighting function that emphasizes the background contributions to the highest hourly
ozone values (when BGO3 contributions are generally lowest), proportionally contribute slightly
less for the W126 than for seasonal means of maximum daily 8 hour average values. As with the
primary standard, in identifying the range of policy options supported by the evidence and
information, staff has not considered proximity to background Cb concentrations. The
Administrator, when evaluating the range of possible standards that are supported by the
scientific evidence, could consider proximity to background O3 concentrations as one factor in
selecting the appropriate standard.
1.3.3 Organization of this Document
Chapter 2 of this PA provides an overview of the Os ambient monitoring network and Os
air quality, including estimates of Os concentrations attributable to background sources. The
remaining chapters are organized into two main parts. Chapters 3 and 4 focus on the review of
the primary Os NAAQS while chapters 5 and 6 focus on the review of the secondary Os
NAAQS. Staffs considerations and conclusions related to the current primary and secondary
standards are discussed in chapters 3 and 5, respectively. Staffs considerations and conclusions
related to potential alternative primary and secondary standards are discussed in chapters 4 and
6, respectively. Key uncertainties in the review and areas for future research and data collection
are additionally identified in chapters 4 and 6 for the two types of standards.
1-42
-------
1.4 REFERENCES
Frey, H.C. and Samet, J.M. (2012a) 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). EPA-CASAC-13-
003. November 26, 2012.
Frey, H.C. and Samet, J.M. (2012b) CASAC Review of the EPA's Health Risk and Exposure Assessment for Ozone
(First External Review Draft - Updated August 2012) and Welfare Risk and Exposure Assessment for
Ozone (First External Review Draft - Updated August 2012). EPA-CASAC-13-002. November 19, 2012
Henderson, R. (2006) Letter from CASAC Chairman Rogene Henderson to EPA Administrator Stephen Johnson.
October 24, 2006, EPA-CASAC-07-001.
Henderson, R. (2007) Letter from CASAC Chairman Rogene Henderson to EPA Administrator Stephen Johnson.
March 26, 2007, EPA-CASAC-07-002.
Henderson, R. (2008) Letter from CASAC Chairman Rogene Henderson to EPA Administrator Stephen Johnson.
April 7, 2008, EPA-CASAC-08-001.
Samet, J.M. (2011) Clean Air Scientific Advisory Committee (CASAC) Response to Charge Questions on the
Reconsideration of the 2008 Ozone National Ambient Air Quality Standards. EPA-CASAC-11-004.
March 30, 2011. Available online at:
http://yosemite.epa.gov/sab/sabproduct.nsf/0/F08BEB48C1139E2A8525785E006909AC/$File/EPA-
CASAC-1 l-004-unsigned+.pdf
U.S. DHEW (Department of Health, Education, and Welfare) (1970). Air Quality Criteria for Photochemical
Oxidants. Washington, D.C.: National Air Pollution Control Administration; publication no. AP-63.
Available from: NTIS, Springfield, VA; PB-190262/BA.
U.S. EPA (U.S. Environmental Protection Agency). (1978). Air quality criteria for ozone and other photochemical
oxidants [EPA Report]. (EPA/600/8-78/004). Washington, D.C..
U.S. EPA (U.S. Environmental Protection Agency). (1982). Air quality criteria document for ozone and other
photochemical oxidants. Fed Reg 47: 11561.
U.S. EPA (U.S. Environmental Protection Agency). (1986). Air quality criteria for ozone and other photochemical
oxidants [EPA Report]. (EPA-600/8-84-020aF - EPA-600/8-84-020eF). Research Triangle Park, NC.
http://www.ntis.gov/search/product.aspx? ABBR=PB87142949
U.S. Environmental Protection Agency. (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. (1996) Air Quality Criteria for Ozone and Related Photochemical Oxidants Volume I of III (Final, 1996).
U.S. Environmental Protection Agency, Washington, D.C., EPA/600/AP-93/004aF (NTIS PB94173127).
U.S. EPA. (2004) Air Quality Criteria for Paniculate Matter (Final Report). U.S. Environmental Protection Agency,
Washington, D.C., EPA 600/P-99/002aF-bF, 2004.
U.S. Environmental Protection Agency. (2006). Air quality criteria for ozone and related photochemical oxidants
[EPA Report]. (EPA/600/R-05/004AF). Research Triangle Park, NC.
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=149923
1-43
-------
U.S. Environmental Protection Agency. (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. EPA-452/R-07-007.
U.S. Environmental Protection Agency. (201 la). Integrated Review Plan for the O3 National Ambient Air Quality
Standards (IRP) U. S. Environmental Protection Agency, National Center for Environmental Assessment
Office of Research and Development and Office of Air Quality Planning and Standards Office of Air and
Radiation, Research Triangle Park, North Carolina EPA. 452/R-l 1-006.
U.S. Environmental Protection Agency. (201 Ib) Integrated Science Assessment for Ozone and Related
Photochemical Oxidants: First External Review Draft, U.S. Environmental Protection Agency,
Washington, D.C., EPA/600/R-10/076A.
U.S. Environmental Protection Agency. (201 Ic) Integrated Science Assessment of Ozone and Related
Photochemical Oxidants (Second External Review Draft). U.S. Environmental Protection Agency,
Washington, D.C., EPA/600/R-10/076B.
U.S. Environmental Protection Agency. (201 Id) Policy Assessment for the Review of the Paniculate Matter
National Ambient Air Quality Standards. Office of Air Quality Planning and Standards, Research Triangle
Park, NC. EPA 452/R-l 1-003.
U.S. Environmental Protection Agency. (2012a). Integrated Science Assessment for Ozone and Related
Photochemical Oxidants: Third External Review Draft, U.S. Environmental Protection Agency, Research
Triangle Park, NC. EPA/600/R-10/076C
U.S. Environmental Protection Agency. (2012b). Health Risk and Exposure Assessment for Ozone, First External
Review Draft, U.S. Environmental Protection Agency, Research Triangle Park, NC. EPA 452/P-12-001.
U.S. Environmental Protection Agency. (2013). Integrated Science Assessment for Ozone and Related
Photochemical Oxidants (Final Report). U.S. Environmental Protection Agency, Washington, D.C.,
EPA/600/R-10/076F, 2013
U.S. Environmental Protection Agency. (2014a). Health Risk and Exposure Assessment for Ozone, Second External
Review Draft. Office of Air Quality Planning and Standards, Research Triangle Park, NC. EPA-452/P-14-
004a.
U.S. Environmental Protection Agency. (2014b). Welfare Risk and Exposure Assessment for Ozone, Second
External Review Draft. Office of Air Quality Planning and Standards, Research Triangle Park, NC. EPA-
452/P-14-003a.
1-44
-------
2 O3 MONITORING AND AIR QUALITY
This section provides overviews of ambient Os monitoring in the U.S. (section 2.1); Ch
precursor emissions and atmospheric chemistry (section 2.2); ambient Os concentrations (section
2.3); and available evidence and information related to background Os (section 2.4). These issues
are also discussed in detail in chapter 3 of the Integrated Science Assessment (ISA) (US EPA,
2013).
2.1 O3 MONITORING
2.1.1 Os Monitoring Network
To monitor compliance with the NAAQS, state and local environmental agencies operate
Os monitoring sites at various locations, depending on the population of the area and typical peak
Os concentrations.1 All of the state and local monitoring stations that report data to the EPA AQS
use ultraviolet (UV) Federal Equivalent Methods (FEMs). The Federal Reference Method (FRM)
is no longer used due to lack of availability and safety concerns.2 In 2010, there were over 1,300
state, local, and tribal Os monitors reporting concentrations to EPA. The "State and Local
Monitoring Stations" (SLAMS) minimum monitoring requirements to meet the Os design criteria
are specified in 40 CFR Part 58, Appendix D. The requirements are both population and design
value based.3 The minimum number of Os monitors required in a Metropolitan Statistical Area
(MSA) ranges from zero for areas with a population of at least 50,000 and under 350,000 with no
recent history of an Os design value greater than 85 percent of the NAAQS, to four for areas with
a population greater than 10 million and an Os design value greater than 85 percent of the NAAQS.
Within an Os network, at least one site for each MSA, or Combined Statistical Area (CSA) if
multiple MSAs are involved, must be designed to record the maximum concentration for that
particular metropolitan area. Since highest Os concentrations tend to be associated with particular
seasons for various locations, EPA requires ozone monitoring during specific ozone monitoring
seasons which vary by state.4
1 The minimum Os monitoring network requirements for urban areas are listed in Table D-2 of Appendix D to 40
CFR Part 58.
2 EPA is developing a new Os Federal Reference Method (FRM) and proposed changes to the FEM testing
requirements to reflect new and improved measurement technology.
3 A design value is a statistic that describes the air quality status of a given area relative to the level of the NAAQS.
Design values are typically used to classify nonattainment areas, assess progress towards meeting the NAAQS, and
develop control strategies. See http://epa.gov/airtrends/values.html (U, 2010, 677582) for guidance on how these
values are defined.
4 The required O3 monitoring seasons for each state are listed in Table D-3 of Appendix D to 40 CFR Part 58. EPA
plans to complete an analysis using certified data for the years of 2010-2012 to determine if any changes to the
length of the required Os monitoring seasons would be needed to support a revised NAAQS.
2-1
-------
Figure 2-1 shows the locations of the U.S. ambient Os monitoring sites reporting data to
EPA at any time during the 2006-2010 period. The gray dots which make up over 80% of the Os
monitoring network are SLAMS monitors, which are operated by state and local governments to
meet regulatory requirements and provide air quality information to public health agencies. Thus,
the SLAMS monitoring sites are largely focused on urban and suburban areas. The blue dots
highlight two important subsets of monitoring sites within the SLAMS network: the "National
Core" (NCore) multi-pollutant monitoring network and the "Photochemical Assessment
Monitoring Stations" (PAMS) network.5
While the existing U.S. Cb monitoring network has a largely urban focus, to address
ecosystem impacts of Os such as biomass loss and foliar injury, it is equally important to focus on
Os monitoring in rural areas. The green dots in Figure 2-1 represent the Clean Air Status and
Trends Network (CASTNET) monitors which are located in rural areas. There were about 80
CASTNET sites operating in 2010, with sites in the eastern U.S. being operated by EPA and sites
in the western U.S. being operated by the National Park Service (NPS). Finally, the black dots
represent "Special Purpose Monitoring Stations" (SPMS), which include about 20 rural monitors
as part of the "Portable Os Monitoring System" (POMS) network operated by the NPS. Between
the CASTNET, NCore, and POMS networks, there were about 120 rural Os monitoring sites
operating in the U.S. in 2010.
5 EPA is currently developing proposed revisions to the PAMS network design intended to increase coverage and
allow for more locale-specific flexibility.
2-2
-------
• SLAMS
CASTNET
• NCORE/PAMS • SPMS/OTHER
Figure 2-1. Map of U.S. ambient Os monitoring sites reporting data to EPA during the
2006-2010 period.
2.1.2 Recent Os Monitoring Data and Trends
To determine whether or not the Os NAAQS has been met at an ambient monitoring site,
a statistic commonly referred to as a "design value" must be calculated based on three consecutive
years of data collected from that site. The form of the existing Os NAAQS design value statistic is
the 3-year average of the annual 4th highest daily maximum 8-hour Os concentration in parts per
billion (ppb), with decimal digits truncated. The existing primary and secondary Cb NAAQS are
met at an ambient monitoring site when the design value is less than or equal to 75 ppb.6 In counties
or other geographic areas with multiple monitors, the area-wide design value is defined as the
design value at the highest individual monitoring site, and the area is said to have met the NAAQS
if all monitors in the area are meeting the NAAQS.
Figure 2-2 shows the trend in the annual 4th highest daily maximum 8-hour Os
concentrations in ppb based on 933 "trends" sites with complete data records over the 2000 to
2012 period. The center line in this figure represents the median value across the trends sites, while
6 For more details on the data handling procedures used to calculate design values for the existing O3 NAAQS, see
40 CFR Part 50, Appendix P.
2-3
-------
the dashed lines represent the 25th and 75th percentiles, and the bottom and top lines represent the
10th and 90th percentiles. Figures 2-3 and 2-4 show maps of the Os design values (ppb) at all U.S.
monitoring sites for the 2009-2011 and 2010-2012 periods, respectively. The trend figure shows
that the annual 4th highest daily maximum values decreased for the vast majority of monitoring
sites in the U.S. between 2000 and 2009. The decreasing trend is especially sharp from 2002 to
2004, when EPA implemented the "NOx SIP Call", a program designed to reduce summertime
emissions of NOx in the eastern U.S., but has been relatively flat since then.
The trends also show a modest increase in the 4th highest daily maximum values from 2009
to 2012. This is reflected in the design value maps, which show an increase in the number of
monitors violating the existing Os standard in 2010-2012 relative to 2009-2011. Meteorology
played an important role in these short-term trends. Os concentrations tend to be higher on days
with hot and stagnant conditions and lower on days with cool or wet conditions. According to the
National Oceanic and Atmospheric Administration's National Climatic Data Center (NOAA-
NCDC), the summer of 2009 was cooler and wetter than average over most of the eastern U.S.,
while conversely the summers, of 2010, 2011, and 2012 were all much warmer than average. In
particular, the central and eastern U.S. experienced a 2-week period of record-breaking heat in late
June and early July of 2012, which contributed to hundreds of violations of the existing Os
standard. In contrast, the most recent climatological information available from NOAA-NCDC
(http://www.ncdc.noaa.gov/sotc/) shows that the summer of 2013 was cooler and wetter than
average for much of the U.S. Thus, EPA does not expect the recent increasing trend in the 4th
highest daily maximum Os concentrations to continue in 2013.
Trend in Annual 4th Highest Daily Maximum 8-hour O3 Concentrations
o r-
O
National Trend Based on 933 Monitoring Sites
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
Year
Figure 2-2. Trend in U.S. annual 4th highest daily maximum 8-hour Os concentrations in
ppb, 2000 to 2012. Solid center line represents the median value across
monitoring sites, dashed lines represent 25th and 75th percentile values, and
top/bottom lines represent 10th and 90th percentile values.
2-4
-------
8-hour Ozone Design Values
• 41 - 60 ppb (92 sites)
©61-65 ppb (179 sites)
O 66 - 70 ppb (324 sites)
71 - 75 ppb (302 sites)
76- 107 ppb (205 sites)
Figure 2-3. Map of 8-hour Os design values in ppb for the 2009-2011 period.
8-hour Ozone Design Values
- 60 ppb (87 sites)
-65 ppb (130 sites)
- 70 ppb (237 sites)
- 75 ppb (309 sites)
76- 106 ppb (363 sites)
Figure 2-4. Map of 8-hour Os design values in ppb for the 2010-2012 period.
2-5
-------
In addition, EPA focused our analyses of welfare and ecosystem effects on a W126 Os
exposure metric in this review. The W126 metric7 is a seasonal aggregate of daytime (8:00 AM to
8:00 PM) hourly Os concentrations, designed to measure the cumulative effects of Os exposure on
plant and tree species, with units in parts per million-hours (ppm-hrs). The W126 metric uses a
logistic weighting function to place less emphasis on exposure to low hourly Os concentrations
and more emphasis on exposure to high hourly Cb concentrations (Lefohn et al, 1988).
Figure 2-5 shows the trend in annual W126 concentrations in ppm-hrs based on 933
"trends" sites with complete data records over the 2000 to 2012 period. Figures 2-6 and 2-7 show
maps of the 3-year average annual W126 concentrations in ppm-hrs at all U.S. monitoring sites
for the 2009-2011 and 2010-2012 periods, respectively. The general patterns seen in these figures
are similar to those seen in the design value metric for the existing standard.
Trend in Annual W126 Concentrations
National Trend Based on 933 Monitoring Sites
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
Year
Figure 2-5. Trend in U.S. annual W126 concentrations in ppm-hrs, 2000 to 2012. Solid
center line represents the median value across monitoring sites, dashed lines
represent 25th and 75th percentile values, and top/bottom lines represent
10th and 90th percentile values.
7 Details on the procedure used to calculate the W126 metric are provided in Chapter 4 of the welfare Risk and
Exposure Assessment.
2-6
-------
Average Annual W126 Values
• 0-3 ppm-hr (79 sites)
04-7 ppm-hr (328 sites)
O 8- 11 ppm-hr(367 sites)
• 12- 15 ppm-hr (194 sites)
• 16-58 ppm-hr (128 sites)
Figure 2-6. Map of 2009-2011 average annual W126 values in ppm-hrs.
Average Annual W126 Values
• 0-3 pprn-hr(68 sites)
• 4- 7 pprn-hr(217 sites)
O 8- 11 ppm-hr(314 sites)
• 12- 15 ppm-hr (260 sites)
• 16-59 ppm-hr (257 sites)
Figure 2-7. Map of 2010-2012 average annual W126 values in ppm-hrs.
2-7
-------
Variations in meteorological conditions play an important role in determining ozone
concentrations. Ozone is more readily formed on warm, sunny days when the air is stagnant.
Conversely, ozone generation is more limited when it is cool, rainy, cloudy, or windy. EPA uses
a statistical model to adjust for the variability in seasonal average ozone concentrations due to
weather conditions to provide a more accurate assessment of the underlying trend in ozone
caused by emissions (Camalier, 2007). Figure 2-8 shows the national trend in the May to
September average of the daily maximum 8-hour ozone concentrations from 2000 to 2012 in 112
urban locations. The dotted red line shows the trend in observed ozone concentrations at selected
monitoring sites, while the solid blue line shows the underlying ozone trend at those sites after
removing the effects of weather. The solid blue lines represent ozone levels anticipated under
"typical" weather conditions and serve as a more accurate assessment of the trend in ozone due
to changes in precursor emissions.
National Urban Ozone Trend (112 Locations)
-g.eoH
£t
I45H
0>
(O
ay
-F-
o
i
Adjusted for Weather
Unadjusted for Weather
O
O
O
CN
0
0
CM
(N
0
0
04
oo
O
O
CM
O
e
CM
m
0
o
CM
0
o
CM
h-
0
0
CM
00
0
O
CM
O1
0
0
CM
0 T-
O
CM
O
CM
CM
O
CM
Figure 2-8. Trend in the May to September mean of the daily maximum 8-hour ozone
concentrations before (dotted red line) and after (solid blue line) adjusting for
year-to-year variability in meteorology8.
Figure 2-8 shows that after adjusting for the year-to-year variability in meteorology, the
overall trend in seasonal average ozone concentrations is much smoother. The adjusted trend
' More detailed information on these trends is available at: http://www.epa.gov/airtrends/weather.html
2-8
-------
clearly shows that the NOx SIP Call program resulted in a sharp decrease in summertime ozone
concentrations starting in 2004. The adjusted trend also indicates that ozone levels continued to
decrease between 2004 and 2009, and while there is still some evidence of an increasing trend
from 2009 to 2012, there is also evidence that much of the recent increase in ozone levels is due
to meteorological conditions which were more favorable to ozone formation than normal.
2.2 EMISSIONS AND ATMOSPHERIC CHEMISTRY
Os is formed by photochemical reactions of precursor gases and is not directly emitted
from specific sources. In the stratosphere, Os occurs naturally and provides protection against
harmful solar ultraviolet radiation. In the troposphere, near ground level, Os forms through
atmospheric reactions involving two main classes of precursor pollutants: volatile organic
compounds (VOCs) and nitrogen oxides (NOx). Carbon monoxide (CO) and methane (CIHLt) are
also important for Os formation over longer time periods (US EPA, 2013, section 3.2.2).
Emissions of Os precursor compounds can be divided into anthropogenic and natural
source categories, with natural sources further divided into biogenic emissions (from vegetation,
microbes, and animals) and abiotic emissions (from biomass burning, lightning, and geogenic
sources). Anthropogenic sources, including mobile sources and power plants, account for the
majority of NOx and CO emissions. Anthropogenic sources are also important for VOC
emissions, though in some locations and at certain times of the year (e.g., southern states during
summer) the majority of VOC emissions come from vegetation (US EPA, 2013, section 3.2.1).
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, 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
(US EPA, 2013, sections 3.2 and 3.7.1).
Rather than varying directly with emissions of its precursors, Os changes in a nonlinear
fashion with the concentrations of its precursors. NOx emissions lead to both the formation and
destruction of Os, depending on the local quantities of NOx, VOC, radicals, and sunlight. In
areas dominated by fresh emissions of NOx, radicals are removed, which lowers the Os
formation rate. In addition, the scavenging of Os by reaction with NO is called "titration" and is
often found in downtown metropolitan areas, especially near busy streets and roads, as well as in
power plant plumes. This short-lived titration results in localized areas in which Os
concentrations are suppressed compared to surrounding areas, but which contain NO2 that
contributes to subsequent Os formation further downwind. Consequently, Os response to
reductions in NOx emissions is complex and may include Os decreases at some times and
2-9
-------
locations and increases of Os at other times and locations. In areas with relatively low NOx
concentrations, such as those found in remote continental areas and rural and suburban areas
downwind of urban centers, Os production typically varies directly with NOx concentrations
(e.g. decreases with decreasing NOx emissions). The NOx titration effect is most pronounced in
urban core areas which have higher volume of mobile source NOx emissions from vehicles than
do the surrounding areas. It should be noted that such locations, which are heavily NOx saturated
(or radical limited), tend to have much lower observed Os concentrations than downwind areas.
As a general rule, as NOx emissions reductions occur, one can expect lower Os values to
increase while the higher ozone values would be expected to decrease. NOx reductions are
expected to result in a compressed Os distribution, relative to current conditions.
The formation of Os from precursor emissions is also affected by meteorological
parameters such as the intensity of sunlight and atmospheric mixing. Major episodes of high
ground-level Os concentrations in the eastern United States are associated with slow-moving
high pressure systems. 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 Os precursors. The combination of inhibited vertical mixing and light winds
minimizes the dispersal of pollutants, allowing their concentrations to build up. In addition, in
some parts of the United States (e.g., in Los Angeles), mountain barriers limit mixing and result
in a higher frequency and duration of days with elevated Os concentrations. Photochemical
activity involving precursors is enhanced during warmer seasons because of the greater
availability of sunlight and higher temperatures (US EPA, 2013, section 3.2).
Os concentrations in a region are affected both by local formation and by transport of Os
and its precursors from upwind areas. Os transport occurs on many spatial scales including local
transport between cities, regional transport over large regions of the U.S. and international/long-
range transport. In addition, Os can be transferred into the troposphere from the stratosphere,
which is rich in Os, through stratosphere-troposphere exchange (STE). These intrusions usually
occur behind cold fronts, bringing stratospheric air with them and typically affect Os
concentrations in higher elevation areas (e.g. > 1500 m) more than areas at lower elevations
(U.S. EPA, 2012, section 3.4.1.1). The role of long-range transport of ozone and other elements
of ozone background is discussed in more detail in Section 2.4.
2.3 AIR QUALITY CONCENTRATIONS
Because Os is a secondary pollutant formed in the atmosphere from precursor emissions,
concentrations are generally more regionally homogeneous than concentrations of primary
pollutants emitted directly from stationary and mobile sources (US EPA, 2013, section 3.6.2.1).
2-10
-------
However, variation in local emissions characteristics, meteorological conditions, and topography
can result in daily and seasonal temporal variability in ambient Os concentrations, as well as
local and national-scale spatial variability.
Temporal variation in ambient Os concentrations results largely from daily and seasonal
patterns in sunlight, precursor emissions, atmospheric stability, wind direction, and temperature
(US EPA, 2013, section 3.7.5). On average, ambient Os concentrations follow well-recognized
daily and seasonal patterns, particularly in urban areas. Specifically, daily maximum 1-hr Os
concentrations in urban areas tend to occur in mid-afternoon, with more pronounced peaks in the
warm months of the Os season than in the colder months (US EPA, 2013, Figures 3-54, 3-156 to
3-157). Rural sites also follow this general pattern, though it is less pronounced in colder months
(US EPA, 2013, Figure 3-55). With regard to day-to-day variability, median maximum daily 8-
hour average (MDA8) Os concentrations in U.S. cities in 2007-2009 were approximately 47 ppb,
with typical ranges between 35 to 60 ppb and the highest MDA8 concentrations above 100 ppb
in several U.S. cities (as noted further below).
In addition to temporal variability, there is considerable spatial variability in ambient Os
concentrations within cities and across different cities in the United States. With regard to spatial
variability within a city, local emissions characteristics, geography, and topography can have
important impacts. For example, as noted above, fresh NO emissions from, for example, motor
vehicles titrate Os present in the urban background air, resulting in an Os gradient around
roadways with Os concentrations increasing as distance from the road increases (US EPA, 2013,
section 3.6.2.1). In comparing urban areas, the ISA notes that measured Os concentrations are
relatively uniform and well-correlated within some cities (e.g., Atlanta) while they are more
variable in others (e.g., Los Angeles) (US EPA, 2013, section 3.6.2.1 and Figures 3-28 to 3-36).
With regard to variability across cities, when the ISA evaluated the distributions of 8-
hour Os concentrations for the years 2007 to 2009 in 20 cities, the highest concentrations were
reported in Los Angeles, with high concentrations also reported in several eastern and southern
cities. The maximum recorded MDA8 was 137 ppb in Los Angeles, and was near or above 120
ppb in Atlanta, Baltimore, Dallas, New York City, Philadelphia, and St. Louis (US EPA, 2013,
Table 3-10). The pattern was similar for the 98th percentile of the distribution of MDA8
concentrations9, with Los Angeles recording the highest 98th percentile concentration (91 ppb)
and many eastern and southern cities reporting 98th percentile concentrations near or above 75
ppb. In contrast, somewhat lower 98th percentile Os concentrations were recorded in cities in the
western United States outside of California (US EPA, 2013, Table 3-10).
9 Table 3-10 in the ISA analyzes the warm season. Therefore, the 98th percentile values would be an approximation
of the 4th highest value.
2-11
-------
Rural sites can be affected by transport of Os or Cb precursors from upwind urban areas
and by local anthropogenic sources such as motor vehicles, power generation, biomass
combustion, or oil and gas operations (US EPA, 2013, section 3.6.2.2). In addition, Os tends to
persist longer in rural than in urban areas due to lower rates of chemical scavenging in non-urban
environments. At higher elevations, increased Os concentrations can also result from
stratospheric intrusions (US EPA, 2013, sections 3.4, 3.6.2.2). As a result, Os concentrations
measured in some rural sites can be higher than those measured in nearby urban areas (US EPA,
2013, section 3.6.2.2), and the ISA concludes that cumulative exposures for humans and
vegetation in rural areas can be substantial, often higher than cumulative exposures in urban
areas (US EPA, 2013, section 3.7.5).
2.4 BACKGROUND O3
One of the aspects of ozone that is unusual relative to the other pollutants with National
Ambient Air Quality Standards (NAAQS) is that, periodically, in some locations, an appreciable
fraction of the observed ozone results from sources or processes other than local and domestic
regional anthropogenic emissions of ozone precursors (Fiore etal., 2002). Any ozone formed by
processes other than the chemical conversion of local or regional ozone precursor emissions is
generically referred to as "background" ozone. Background Os can originate from natural
sources of Os and Os precursors, as well as from manmade international emissions of Os
precursors. Natural sources of Os precursor emissions such as wildfires, lightning, and vegetation
can lead to Os formation by chemical reactions with other natural sources. Another important
component of background is Os that is naturally formed in the stratosphere through interactions
of ultraviolet light with molecular oxygen. Stratospheric Os can mix down to the surface at high
concentrations in discrete events called intrusions, especially at higher-altitude locations. The
manmade portion of the background includes any Os formed due to anthropogenic sources of Os
precursors emitted far away from the local area (e.g., international emissions). Finally, both
biogenic and international anthropogenic emissions of methane, which can be chemically
converted to Os over relatively long time scales, can also contribute to global background Os
levels. Away from the surface, ozone can have an atmospheric lifetime on the order of weeks. As
a result, background ozone can be transported long distances in the upper troposphere and, when
meteorological conditions are favorable, be available to mix down to the surface and add to the
ozone loading from non-background sources.
As indicated in the first draft policy assessment (US EPA, 2012, sections 1.3.4 and 3),
EPA has updated several aspects of our methodology for estimating the change in health risk and
exposure that would result from a revision to the Os NAAQS. First, risk estimates are now based
on total Os concentrations, as opposed to previous reviews which only considered risk above
2-12
-------
background levels. Second, EPA is now using air quality modeling to estimate the spatial
patterns of Os that would result from attaining various levels of the NAAQS, as opposed to a
quadratic rollback approach that required the estimation of a background "floor" beyond which
the rollback would not take place. Both of these revisions have had the indirect effect of reducing
the need for estimates of background Os levels as part of the Os risk and exposure assessment
(REA). Regardless, EPA expects that a well-founded understanding of the fractional contribution
of background sources and processes to surface Os levels will be valuable. Accordingly, in this
section, we briefly summarize existing results on background Os from the ISA (US EPA, 2013,
section 3.4) as supplemented by additional EPA modeling recently conducted for a 2007 base
year. The summary will focus on national estimates of the:
• seasonal mean background Os values for three specific definitions of background Os,
• relative proportion of background Os to total Os for the same three definitions from a
seasonal mean perspective,
• distributions of background Os within a seasonal mean,
• ratio of background Os to total modeled ozone in the 12 REA case study areas,
• relative proportion of background Os concentrations to total W126 ozone, and
• relative contribution of different components of background to total background Os.
The definition of background Os can vary depending upon context, but it generally refers
to Os that is formed by sources or processes that cannot be influenced by actions within the
jurisdiction of concern. In the first draft policy assessment document (US EPA, 2012), EPA
identified three specific definitions of background Os: natural background (NB), North American
background (NAB), and United States background (USB). Natural background is the narrowest
definition of background, and it is defined as the Os that would exist in the absence of any
manmade Os precursor emissions. The other two definitions of background are based on a
presumption that the U.S. has little influence over anthropogenic emissions outside either our
continental or domestic borders. North American background is defined as that Os that would
exist in the absence of any manmade Os precursor emissions from North America. U.S.
background is defined as that Os that would exist in the absence of any manmade emissions
inside the United States.
Each of these three definitions of background Os requires photochemical modeling
simulations to estimate what the residual Os concentrations would be were the various
anthropogenic emissions to be removed. EPA exclusively uses modeling estimates to
characterize background, as opposed to using observed concentrations from a remote site,
because even the most remote monitors within the U.S. can be periodically affected by U.S.
2-13
-------
anthropogenic emissions. In most situations, without special monitoring it is impossible to
determine how much of the ozone measured by a monitor originated from background sources.
Prior to using the new 2007-based model simulations to estimate background Cb levels over the
U.S., EPA confirmed that this modeling was able to reproduce historical Os levels and that there
was limited correlation between model errors and the background estimates. This evaluation is
described more fully in the appendix (Appendix A) to this chapter.
Previous modeling studies have estimated what background levels would be in the
absence of certain sets of emissions by simply assessing the remaining Os in a simulation in
which certain emissions were removed (Zhang et al. (2011), Emery et al. (2012)). This basic
approach is often referred to as "zero-out" modeling or "emissions perturbation" modeling.
While the zero-out approach has traditionally been used to estimate natural background, North
American background, and U.S. background, the methodology has an acknowledged limitation.
It cannot answer the question of how much of the existing observed ozone results from
background sources or processes.
A separate modeling technique can be used to estimate the contribution of background
ozone and other contributing source terms to total Os within a model. This approach, referred to
as "source apportionment" modeling, has been described and evaluated in the peer-reviewed
literature (Dunker et al., 2002; Kemball-Cook et al., 2009). Source apportionment modeling has
frequently been used in other regulatory settings to estimate the "contribution" to ozone of
certain sets of emissions (EPA 2005, EPA 2011). The source apportionment technique provides a
means of estimating the contributions of each user-identified source category to ozone formation
in a single model simulation. This is achieved by using multiple tracer species to track the fate of
ozone precursor emissions (VOC and NOx) and the ozone formation resulting from these
emissions. The methodology is designed so that all ozone and precursor concentrations are
tracked and apportioned to the selected source categories at all times without perturbing the
inherent chemistry. The primary limitation of the source apportionment modeling is that its
estimations of background are explicitly linked to the emissions scenarios modeled and would
change with different emissions scenarios.
EPA recently completed updated zero-out and source apportionment modeling for a 2007
base year to supplement the characterization of background Os that was provided in the ISA.
Both of these approaches have value in assessing the potential impacts of background Os, but
they are used separately as described in Table 2-1.
2-14
-------
Estimation
Methodology
Question addressed
Background
Quantities
Strengths and Limitations
Zero-out
How much ozone would remain if
controllable emissions were
completely removed?
NB/NAB/
USB
Strength: The approach is simple
to implement and provides an
estimate of the lowest O3 levels
that can be attained by
eliminating all U.S.
anthropogenic emissions.
Limitation: Estimates are based
on a counterfactual, represents a
quantity never to occur in real
atmosphere. Additionally,
sensitivity approaches can be
unreliable for evaluating mass
contributions to Os production
because of non-linearity in the
chemistry.
Source
Apportionment
How much of the current ozone
can be attributed to sources other
than U.S. anthropogenic sources?
Apportionment-
based USB
Strength: Provides a direct
estimate of the amount of Os
contributed by each source
category while avoiding artifacts
caused by non-linearity in the
chemistry.
Limitation: While this approach
identifies important sources that
contribute to Os, it does not
predict quantitatively how O3 will
respond to specific emissions
reduction scenarios.
Table 2-1 Comparison of the two model methodologies used to characterize background
ozone levels.
The key configuration elements of the updated modeling are described below; a more
detailed description of the modeling is provided in Appendix A. The zero-out modeling was
based on a model configuration that nested a regional-scale air quality model (CMAQ at 12 km
horizontal grid resolution) within a global scale air quality model (GEOS-Chem at 2.0 x 2.5
2-15
-------
degree horizontal grid resolution). The lateral boundary conditions from the global model were
used as inputs for the regional simulation. Four scenarios were modeled:
• a 2007 base case simulation which was also the basis of the air quality modeling
performed for the 2nd draft ozone REA and is described in more detail in Appendix 4b
ofEPA(2013b)
• a natural background run with anthropogenic ozone precursor emissions10 removed in
both the global and regional models11
• a North American background run with anthropogenic ozone precursor emissions
removed across North America in both the global and regional model simulations
• a U.S. background run with anthropogenic ozone precursor emissions removed over
the U.S. in both the global and regional model simulations
The source apportionment modeling was also based on a regional scale air quality model
(CAMx at 12 km horizontal grid resolution) that used the same lateral boundary conditions from
the 2007 base global modeling scenario. EPA used the Anthropogenic Precursor Culpability
Assessment (APCA) tool in this analysis. The APCA tool attributes ozone production to
manmade sources whenever ozone is determined to result from a combination of anthropogenic
and biogenic emissions (Environ, 2011). The APCA methodology calculates natural ozone as the
production resulting from the interaction of biogenic VOC with biogenic NOx emissions. Eleven
separate source categories were tracked in the EPA source apportionment analysis, including five
boundary condition terms and six in-domain sectors:
• Boundary condition terms:12
o Northern edge
o Eastern edge
o Southern edge
o Western edge
o Top boundary
10 In the global model, only emissions from natural sources were used (i.e., VOC, NOx, CO) and methane was reset
to pre-industrial levels (700 ppb) to reflect natural contributions. In the regional modeling, the methane levels were
left unchanged.
11 Note that methane is not modeled as an explicit species in CMAQ but instead is treated as having a constant
concentration. Therefore, although methane was reduced to pre-industrial levels in the global GEOS-Chem run used
to create boundary conditions, methane was assumed to be equal to modern-day levels in the CMAQ model runs.
Since methane reactions occur on relatively long timescales, this discrepancy is not expected to have a large impact
on modeled background ozone levels.
12 It should be noted that although boundary conditions are treated as part of apportionment-based USB for this
analysis, in some cases they may be influenced by US anthropogenic emissions that are advected out of the model
domain and recirculated back into the U.S. This is not expected to make a substantial impact on results.
2-16
-------
• In-domain sectors:
o U.S. anthropogenic emissions
o Point sources located within the Gulf of Mexico
o Category 3 marine vessels outside State boundaries
o Climatologically-averaged wildfire emissions
o Biogenic emissions
o Canada/Mexico emissions (only those sources within the domain)
2.4.1 Seasonal Mean Background Os in the U.S.
The ISA (US EPA 2013, section 3.4) previously established that background
concentrations vary spatially and temporally and that simulated mean background concentrations
are highest at high-elevation sites within the western U.S. Background levels typically are
greatest over the U.S. in the spring and early summer. The updated EPA modeling focused on
the months from April to October. Figure 2-9 displays the spatial patterns of seasonal mean
natural background Os as estimated by a 2007 zero-out scenario. Seasonal means are computed
over those seven months. This figure shows the average daily maximum 8-hour Os concentration
(MDA8) that would exist in the absence of any anthropogenic Os precursor emissions at monitor
locations. As shown, seasonal mean NB levels range from approximately 15-35 ppb (i.e., +/- 1
standard deviation) with the highest values at higher-elevation sites in the western U.S. The
median value over these locations is 24.2 ppb, and more than 50 percent of the locations have
natural background levels of 20-25 ppb. The highest modeled estimate of seasonal average,
natural background, MDA8 Os is 34.3 ppb at the high-elevation CASTNET site (Gothic) in
Gunnison County, CO. Natural background levels are higher at these high-elevation locations
primarily because natural stratospheric Os impacts and international transport impacts increase
with altitude (where Os lifetimes are longer).
2-17
-------
Ozone (ppb)
--20(104)
20-25 (740)
23-30(331)
30-35(119)
35-40 (0)
O 40-45 (0)
O 45-50(0)
50 - 55 (0)
55 • 60 (0)
> 60 (0)
Figure 2-9. Map of 2007 CMAQ-estimated seasonal mean natural background Os levels
(ppb) from zero-out modeling.
Figures 2-10 and 2-11 show the same information for the NAB and USB scenarios. In
these model runs, all anthropogenic Os precursor emissions were removed from the U.S.,
Canada, and Mexico portions of the modeling domain (NAB scenario) and then only from the
U.S. (USB scenario). The figures show that there is not a large difference between the NAB and
USB scenarios. Seasonal mean NAB and USB Os levels range from 25-50 ppb, with the most
frequent values estimated in the 30-35 ppb bin. The median seasonal mean background levels are
31.5 and 32.7 ppb (NAB and USB, respectively). Again, the highest levels of seasonal mean
background are predicted over the intermountain western U.S. Locations with NAB and USB
concentrations greater than 40 ppb are confined to Colorado, Nevada, Utah, Wyoming, northern
Arizona, eastern California, and parts of New Mexico. The 2007 EPA modeling suggests that
seasonal mean USB concentrations are on average 1-3 ppb higher than NAB background. These
results were similar to those reported by Wang et al. (2009). From a seasonal mean perspective,
background levels are typically well-below the NAAQS thresholds.
2-18
-------
Ozone (ppb)
• '- 20 (0]
20 - 25 (0)
25 - 30 (396)
30 - 35 (628)
35-40(147)
O 40-45(121)
O 15.50(2)
SO - 55 (0)
55 • 60 (0)
=- 60 (0)
Figure 2-10. Map of 2007 CMAQ-estimated seasonal mean North American background
Os levels (ppb) from zero-out modeling.
Ozone (ppb)
--20(0}
20-25 (0)
25-30(127)
30-35 (842)
35-40(186)
40-45(132)
O 45 . SO (5)
50 - 55 (0)
55 • 60 (0)
•- 60 (0)
Figure 2-11. Map of 2007 CMAQ-estimated seasonal mean United States background Os
levels (ppb) from zero-out modeling.
2.4.2 Seasonal Mean Background Os in the U.S. as a Proportion of Total Os
Another informative way to assess the importance of background as part of seasonal
mean Os levels across the U.S. is to consider the ratios of NB, NAB, and USB to total modeled
Os at each monitoring location. Considering the proportional impact of background allows for an
2-19
-------
initial assessment of the relative importance of background and non-background sources.
Because ozone chemistry is non-linear, one should not assume that individual perturbations (e.g.,
zero out runs) are additive in all locations. Figures 2-12 and 2-13 show the ratio of U.S.
anthropogenic sources to total Os using the metric of the seasonal mean MDA8 Os
concentrations as estimated by both the zero-out and source apportionment modeling
methodologies. Recall that the terms NB, NAB, and USB are explicitly linked to the zero-out
modeling approach. For comparison sake, in Figure 2-13 we are extending the definition of USB
to also include the source apportionment model estimates of the Os that is attributable to sources
other than U.S. anthropogenic emissions. To preserve the original definition of USB, this second
term will be hereafter referred to as "apportionment-based USB". As noted earlier, the advantage
of the source apportionment modeling is that all of the modeled Os is attributed to various source
terms without perturbing the inherent chemistry. Thus, this approach is not affected by the
confounding occurrences of background Os values exceeding the base Os values as can happen
in the zero-out modeling (i.e., background proportions > 100%). Consequently, one would
expect the fractional background levels to be lower in the source apportionment methodology as
a result of removing this artifact.
When averaged over all sites, Os from sources other than U.S. anthropogenic emissions is
estimated to comprise 66 (zero-out) and 59 (source apportionment) percent of the total seasonal
Os mean. The spatial patterns of USB and apportionment-based USB are similar across the two
modeling exercises. Background Os is a relatively larger percentage (e.g., 70-80%) of the total
seasonal mean Os in locations within the intermountain western U.S. and along the U.S. border.
In locations where Os levels are generally higher, like California and the eastern U.S., the
seasonal mean background fractions are relatively smaller (e.g., 40-60%). The additional 2007
modeling confirms that background ozone, while generally not approaching levels of the ozone
standard, can comprise a considerable fraction of total seasonal mean ozone across the U.S.
2.4.3 Daily Distributions of Background Os within the Seasonal Mean
As a first-order understanding, it is valuable to be able to characterize seasonal mean
levels of background Os. However, it is well established that background levels can vary
substantially from day-to-day within the seasonal mean. From an implementation perspective,
the values of background Os on possible exceedance days are a more meaningful consideration.
The first draft policy assessment (US EPA, 2012) considered this issue in detail, via summaries
of the existing 2006 zero-out modeling (Henderson et al., 2012), and concluded that "results
suggest that background concentrations on the days with the highest total Os concentrations are
not dramatically higher than typical seasonal average background concentrations." Based on this
finding, EPA determined that "anthropogenic sources within the U.S. are largely responsible for
2-20
-------
4th highest 8-hour daily maximum Cb concentrations." The recent EPA modeling using a 2007
base year and the two distinct modeling methodologies described above, supports the finding
from the previous 2006-based modeling analyses. That is, the highest modeled Os site-days tend
to have background Os levels similar to mid-range Os days. Figure 2-14 and 2-15 show the
distribution of daily MDA8 apportionment-based USB levels (absolute magnitudes and relative
fractions, respectively) from the source apportionment simulation13. Again, the 2007 modeling
shows that the days with highest Os levels have similar distributions (i.e., means, inter-quartile
ranges) of background levels as days with lower values, down to approximately 40 ppb. As a
result, the proportion of total Os that has background origins is smaller on high Os days (e.g.,
days > 60 ppb) than on the more common lower Os days that tend to drive seasonal means. This
helps put the results from section 2.4.2 into better context. For example, for site-days in which
base Os is between 70-75 ppb, the source apportionment modeling estimates that approximately
37 percent of those Os levels originate from sources other than U.S. anthropogenic emissions
(i.e., apportionment-based USB). Figure 2-15 also indicates that there are cases in which the
model predicts much larger background proportions, as shown by the upper outliers in the figure.
These infrequent episodes usually occur in relation to a specific event, and occur more often in
specific geographical locations, such as at high elevations or wildfire prone areas during the local
dry season.
It should be noted here that EPA has policies for treatment of air quality monitoring data
affected by these types of events. EPA's exceptional events policy allows exclusion of certain air
quality monitoring data from regulatory determinations if a State adequately demonstrates that an
exceptional event has caused the exceedance or violation of a NAAQS. In addition, Section
179B of the CAA also provides for treatment of air quality data from international transport
when an exceedance or violation of a NAAQS would not have occurred but for the emissions
emanating from outside of the United States.
13 Similar plots from the zero-out modeling for natural background, North American background, and U.S.
background are provided in Appendix A.
2-21
-------
-=40%lOl
M-50%<2|
50 -60% (393)
60-70%(551>
70-80% (263)
> 80% 1341
Figure 2-12. Map of site-specific ratios of U.S. background to total seasonal mean Os based
on 2007 CMAQ zero-out modeling.
< 40% (0)
40 - 50% (343)
SO - 60% (483)
60 - 70% (237)
70-ao%(l7S>
> 80% (52)
Figure 2-13. Map of site-specific ratios of apportionment-based U.S. background to
seasonal mean Os based on 2007 CAMx source apportionment modeling.
2-22
-------
BBBB
<25 25-30 30-36 36-40 40-45 45-50 50-55 55-60 60-65 65-70 70-75 75-SO S3-BE BE-SO 90-95 95-100 > 100
Bins of Base Model MDA8 Ozone (ppb)
Figure 2-14. Distributions of absolute estimates of apportionment-based U.S. Background
(all site-days), binned by modeled MDA8 from the 2007 source
apportionment simulation.
; : : : :•: • :• - :• • r : r : r: :.- ^'. \ : :: ;•: ;-: ~: ~: •-": ":-:: •:: ; :
Bins of Base Model MDA8 Ozone (ppb)
-: : - =o-=e ; . ; : : • : •:
Figure 2-15. Distributions of the relative proportion of apportionment-based U.S.
Background to total Os (all site-days), binned by modeled MDA8 from the
2007 source apportionment simulation.
2-23
-------
2.4.4 Proportion of Background Os in 12 Urban Case Study Areas
This section presents estimates of the overall fraction of Os that is estimated to result
from background sources or processes in each of the 12 urban case study areas considered in the
epidemiological-based risk assessment of the REA (US EPA 2014, Chapter 7). The results are
based on the recent EPA 2007 source apportionment modeling. Table 2-1 summarizes the
estimated ratios of sources other than U.S. anthropogenic emissions (i.e., apportionment-based
USB) to total seasonal mean MDA8 Os in each of the 12 urban case study areas. The table shows
that the fractional contributions from sources other than anthropogenic emissions within the U.S.
can range from 43 to 66 percent across these 12 urban areas. These fractions are consistent with
the national ratios summarized in section 2.4.2, although the fractions of background are
generally smaller at urban sites than at rural sites.
As shown in section 2.4.3, the background-to-total ratios are smaller on days with high
modeled Os (i.e., days that may exceed the level of the NAAQS). Table 2-2 provides the
fractional contributions from apportionment-based USB, only considering days in which base
model MDA8 Os was greater than 60 ppb. As expected, the fractional background contributions
are smaller, ranging from 31 to 55 percent.
Rather than taking the fractions of the seasonal means (as in Table 2-1), Table 2-3
displays the mean and median daily MDA8 background fractions. These metrics may be more
appropriate for application to health studies. The fractional contributions to background
calculated via this approach are very similar to the Table 2-1 calculations. For completeness, we
also provide USB ratios based on the zero-out modeling for the 12 cities (see Table 2-4). The
results are similar to the source apportionment findings (Table 2-1), though the zero-out
technique provides slightly higher background proportions, as expected. It should be noted that
all fractional contributions are based on recent conditions from 2007. These ratios would be
expected to change as anthropogenic emissions and Os levels are lowered.
Based on the source apportionment modeling for these 12 areas, there is evidence that
background levels comprise a non-negligible fraction of the total ozone observed within these
locations. However, for site-days in which model MDA8 ozone exceeds 60 ppb, ozone formed
from U.S. anthropogenic emissions comprise a larger fraction of the total ozone in 11 of the 12
areas (all but Denver). The major metropolitan areas in the eastern U.S. (e.g., Atlanta, New York
City, Philadelphia) are less influenced by background sources than a higher-elevation, western
U.S., location like Denver. Even in Denver, though, U.S. anthropogenic emissions have a large
influence on total ozone (45 percent).
2-24
-------
Table 2-2 Seasonal mean MDA8 Os (ppb), seasonal mean apportionment-based USB
contribution (ppb), and fractional apportionment-based USB contribution to
total Os (all site-days) in the 12 REA urban case study areas (%).
All days, CAMx
Model MDA8 seasonal mean
Model MDA8 seasonal mean
from emissions other than
U.S. anthropogenic sources
Fractional contribution from
background
ATL
59.3
25.3
0.43
BAL
54.4
25.9
0.48
BOS
43.0
26.2
0.61
CLE
48.9
25.7
0.53
DEN
47.3
31.3
0.66
DET
39.1
23.3
0.60
HOU
48.5
27.0
0.56
LA
51.1
29.1
0.57
NYC
45.4
24.5
0.54
PHI
48.7
24.2
0.50
SAC
46.4
29.7
0.64
STL
49.8
24.3
0.49
Table 2-3 Seasonal mean MDA8 Os (ppb), seasonal mean apportionment-based USB
contribution (ppb), and fractional apportionment-based USB contribution to
total Os (site-days > 60 ppb) in the 12 REA urban study areas (%).
Only days w/ base
MDA8 > 60 ppb
Model MDA8 seasonal mean
Model MDA8 seasonal mean
from emissions other than
U.S. anthropogenic sources
Fractional contribution from
background
ATL
74.0
25.4
0.34
BAL
75.3
23.7
0.31
BOS
70.7
24.4
0.35
CLE
72.0
25.4
0.35
DEN
67.5
37.3
0.55
DET
68.9
24.4
0.35
HOU
70.3
28.0
0.40
LA
74.4
31.9
0.43
NYC
74.1
23.5
0.32
PHI
74.0
22.9
0.31
SAC
68.3
32.1
0.47
STL
70.0
25.4
0.36
Table 2-4 Fractional contribution of apportionment-based USB in the 12 REA urban
study areas (%), using the means and medians of daily MDA8 fractions
(instead of fractions of seasonal means).
Mean of daily MDA8
background fractions
Median of daily MDA8
background fractions
ATL
0.46
0.43
BAL
0.53
0.51
BOS
0.68
0.73
CLE
0.58
0.54
DEN
0.69
0.69
DET
0.64
0.66
HOU
0.59
0.59
LA
0.61
0.60
NYC
0.61
0.63
PHI
0.56
0.54
SAC
0.67
0.66
STL
0.52
0.49
Table 2-5 Seasonal mean MDA8 Os (ppb), seasonal mean USB (ppb), and USB/Total ratio
(all site-days) in the 12 REA urban case study areas (%).
All days, CMAQ
Model MDA8 seasonal mean
Model MDA8 seasonal mean
USB
Ratio of USB/Total MDA8 03
ATL
58.6
30.0
0.51
BAL
55.6
29.9
0.54
BOS
45.2
28.5
0.63
CLE
51.8
31.6
0.61
DEN
57.1
42.2
0.74
DET
43.5
31.7
0.73
HOU
49.4
33.0
0.67
LA
54.8
33.3
0.61
NYC
47.7
29.1
0.61
PHI
50.5
29.4
0.58
SAC
51.9
34.4
0.66
STL
52.6
32.0
0.61
2.4.5 Influence of Background Os on W126 levels
EPA also conducted a limited assessment of the impacts of background Os sources on the
W126 metric. The W126 metric (LeFohn et al., 1988) is a cumulative peak-weighted index
designed to estimate longer-term effects of daytime ozone levels on sensitive vegetation and
ecosystems. EPA used the 2007 zero-out modeling to assess NB, NAB, and USB influences at
2-25
-------
four sample locations: Atlanta GA, Denver CO, Farmington NM, and Riverside CA. Each of the
four analyses locations had relatively high observed values of W126 in 2007, as averaged over
all sites in the area: Atlanta (25.1 ppm-hrs), Denver (19.6 ppm-hrs), Farmington (20.2 ppm-hrs),
and Riverside (36.0 ppm-hrs).
As discussed above, in the current review, EPA is supplementing the counterfactual
assessment used in previous reviews (zero out modeling) with analyses that estimate the portion
of the existing ozone that is due to background sources (source apportionment). This has
important ramifications for assessing the influence of background on W126 concentrations,
because of the non-linear weighting function used in the metric, which emphasizes high ozone
hours (e.g., periods in which ozone is greater than -60 ppb). As an example, consider a sample
site in the intermountain western U.S. region with very high modeled estimates of U.S
background (e.g., seasonal mean of 45 ppb with some days as high as 65 ppb). Even at this high
background location, the USB simulations estimate annual W126 (USB) values that are quite
low, on the order of 3 ppm-hrs. Sites in the domain with lower U.S. background levels have even
smaller USB W126 values, on the order of the 1 ppm-hrs, which is consistent with values
mentioned in past reviews (US EPA, 2007). Using the counterfactual scenarios, background
ozone has a relatively small impact on W126 levels across the U.S.
However, because of the non-linear weighting function used in the W126 calculation, the
sum of the W126 from the USB scenario and the W126 resulting from U.S. anthropogenic
sources will not equal the total W126. In most cases, the sum of those two components will be
substantially less than total W126. As a result, EPA believes it is more informative to estimate
the fractional influence of background ozone to W126 levels. Using a methodology that is
described in more detail in Appendix A, EPA considered the fractional influence of background
ozone on annual W126 levels in four locations. The fractional influence methodology utilizes the
2007 zero-out modeling but places higher weights on background fractions on days that are
going to contribute most substantially to the yearly W126 value. Figure 2-16 shows the results.
Based on the fractional influence methodology, natural background sources are estimated to
contribute 29-50% of the total modeled W126 with the highest relative influence in the
intermountain western U.S. (i.e., Farmington NM) and the lowest relative influence in the eastern
U.S. (i.e., Atlanta). U.S. background is estimated to contribute 37-65% of the total modeled
W126. The proportional impacts of background are slightly less for the W126 metric than for
seasonal mean MDA8 (discussed in section 2.4.2), because of the sigmoidal weighting function
that places more emphasis on higher ozone days when background fractions are generally lower.
The key conclusion from this cursory analysis is that background ozone can comprise a
non-negligible portion of current W126 levels across the U.S. These fractional influences are
greatest in the intermountain western U.S. and are slightly smaller than the seasonal mean
2-26
-------
MDA8 metric. This conclusion was also reached in a separate analysis recently completed by
Lapinaetal. (2014).
INB
I NAB
USB
Farmington
Denver
Riverside
Atlanta
Figure 2-16. Fractional influence of background sources to W126 levels in four sample
locations. Model estimates based on 2007 CMAQ zero-out modeling.
2.4.6 Estimated Magnitude of Individual Components of Background Os
To provide a fuller characterization of background Os levels, it is useful to develop an
understanding of the relative contributions of various background elements to total background
Os. This section will utilize the supplemental 2007 air quality modeling estimates to consider the
relative contribution of specific elements of background Os. Several background elements were
isolated in either the zero-out or source apportionment modeling. Appendix A provides more
detail on these analyses. In conjunction with the previous analyses summarized in the ISA, some
broad characterizations of the individual components of background Os can be developed.
The recent 2007 EPA modeling confirms the importance of methane emissions and
international ozone precursor emissions in contributing to background Os. Methane has an
atmospheric lifetime of about a decade and can be an important contributor to ozone on longer
time scales. Anthropogenic methane emission sources include agriculture, coal mines, landfills,
and natural gas and oil systems. The difference between the NAB and NB zero-out scenarios
provides an estimate of contributions from international anthropogenic emissions and
anthropogenic methane, which is modeled by reducing model concentrations from present-day
values to pre-industrial levels. The ISA (US EPA, 2013, section 3.4) estimated that roughly half
of the difference between the NB and NAB scenarios resulted from the removal of anthropogenic
2-27
-------
methane emissions and that the other half resulted from international anthropogenic emissions of
shorter-lived Os precursors (e.g., NOx and VOC). Figure 2-17 shows the difference in seasonal
mean MDA8 Os levels between the NB and NAB scenarios. North American seasonal mean
background is 6-15 ppb higher than comparable natural background levels. The most frequent
increment is an 8-10 ppb increase when the methane is increased and the non-North American
emissions are added. These estimates of seasonal-mean external enhancement are similar to
previous estimates summarized in the ISA (e.g., Fiore et al., 2009; Zhang et al., 2011). It is not
possible via the EPA 2007 modeling to parse out what fraction of this change is due to emissions
outside of North America, as opposed to what fraction is due to anthropogenic methane
emissions, but the modeling suggests that any control measures to reduce emissions from either
of these terms have the potential to contribute in an important way to reduce average background
Os levels in the U.S.
The difference between the NAB and USB scenarios is easier to interpret as it only
involves one change, the inclusion of anthropogenic emissions from the in-domain portion of
Canada and Mexico. These emissions (not shown here, but depicted in Appendix A) contribute
less than 2 ppb to the seasonal mean MDA8 Os levels over most of the U.S. There are 70 sites,
near an international border, where the modeling estimates Canadian/Mexican seasonal average
impacts of 2-4 ppb. Peak single-day MDA8 impacts from these specific international emissions
sources can approach 25 ppb (e.g., San Diego, Buffalo NY).
Figure 2-17. Differences in seasonal mean Os (ppb) between the NAB and NB scenarios.
2-28
-------
Eleven separate source categories were tracked in the source apportionment modeling,
including five boundary condition terms (East, South, West, North, and top) and six emissions
sectors within the domain. The contributions of each of these terms is provided in the Appendix
and summarized below. At most locations, the five model boundary terms contributed an
aggregate 40-60 percent of the total seasonal mean MDA8 Cb across the U.S. The highest
proportional impacts from the boundary conditions are along the coastlines and the
intermountain West. The Os entering the model domain via the boundary conditions can have a
variety of origins including: a) natural sources of Os and precursors emanating from outside the
domain, b) anthropogenic sources of Os precursors emanating from outside the domain, and c)
some fraction of U.S. emissions (natural and anthropogenic) which exit the regional model
domain but get re-imported into the domain via synoptic-scale recirculation. Accordingly, it is
not possible to relate the boundary condition contribution to any specific background element.
The single largest sector that was tracked in the source apportionment modeling was U.S.
anthropogenic emissions. Figure 2-18 shows the contributions from this sector to seasonal mean
MDA8 Os levels. Over most of the U.S. this term contributes 40-60 percent to the total seasonal
mean Os. As discussed in section 2.4.3, these contributions are even higher when only high Cb
days are considered. International shipping emissions, as well as fires and other biogenic
emissions also contribute in a non-negligible way to background Os over the U.S. The key
finding from this analysis is that air quality planning efforts to reduce anthropogenic methane
emissions and international NOx/VOC emissions (e.g., migrating from Asia, Canada, and
Mexico; and from commercial shipping) have the potential to lower background Cb levels.
2-29
-------
Mexico
USGS. EGRI.VTANA:'AND. Soul
O 10-20% (59)
O 20-40% (441)
0 40 - 60% (791)
A > 60% lOI
Figure 2-18. Percent contribution of U.S. anthropogenic emissions to total seasonal mean
MDA8 Os levels, based on 2007 source apportionment modeling.
2.4.7 Summary
For a variety of reasons, it is challenging to present a comprehensive summary of all the
components and implications of background Os. In many forums the term "background" is used
generically and the lack of specificity can lead to confusion as to what sources are being
considered. Additionally, it is well established that the impacts of background sources can vary
greatly over space and time which makes it difficult to present a simple summary of background
Os levels. Further, background Os can be generated by a variety of processes, each of which can
lead to differential patterns in space and time, and which often have different regulatory
ramifications. Finally, background Os is difficult to measure and thus, typically requires air
quality modeling which has inherent uncertainties and potential errors and biases.
That said, EPA believes the following concise and three step summary of the implications
of background Os on the NAAQS review is appropriate, as based on previous modeling exercises
and the more recent EPA analyses summarized herein. First, background Os exists and can
comprise a considerable fraction of total seasonal mean MDA8 Os and W126 across the U.S. Air
quality models can estimate the fractional contribution of background sources to total Os in an
individual area. The largest absolute values of background (NB, NAB, USB, or apportionment-
2-30
-------
based USB) are modeled to occur at locations in the intermountain western U.S. and are
maximized in the spring and early summer seasons. Second, the modeling indicates that U.S.
anthropogenic emission sources are the dominant contributor to the majority of modeled Os
exceedances of the NAAQS. Higher Os days generally have smaller fractional contributions
from background. This finding indicates that the relative importance of background Os would
increase were Os concentrations to decrease with a lower level of the Os NAAQS. Third and
finally, while the majority of modeled Os exceedances have local and domestic regional
emissions as their primary cause, there can be events where Os levels approach or exceed 60-75
ppb due to the influence of background sources. These events are relatively infrequent and EPA
has policies that could allow for the exclusion of air quality monitoring data affected by these
types of events from design value calculations.
2-31
-------
2.5 REFERENCES
Camalier, L.; Cox, W.; Dolwick, P. (2007). The Effects of Meteorology on Ozone in Urban Areas and their use in
Assessing Ozone Trends. Atmos Environ 41: 7127-7137.
Dunker, A.M.; Yarwood, G; Ortmann, J.P.; Wilson, G.M. (2002). Comparison of source apportionment and source
sensitivity of ozone in a three-dimensional air quality model. Environmental Science & Technology 36:
2953-2964.
Emery, C; Jung, J; Downey, N; Johnson, J; Jimenez, M; Yarwood, G; Morris, R. (2012). Regional and global
modeling estimates of policy relevant background ozone over the United States. Atmos Environ 47: 206-
217. http://dx.doi.0rg/10.1016/j.atmosenv.2011.ll.012.
LapinaK., Henze D.K., Milford J.B., Huang M., LinM., Fiore A.M., Carmichael G., Pfister G.G., Bowman K.
(2014). Assessment of source contributions to seasonal vegetative exposure to ozone in the U.S. Journal of
Geophysical Research-Atmospheres, 119: DOI: 10.1002/2013JD020905.
Lefohn, A. S.; Laurence, J. A.; Kohut, R. J. (1988). A comparison of indices that describe the relationship between
exposure to ozone and reduction in the yield of agricultural crops. Atmos. Environ. 22: 1229-1240.
Henderson, B.H., Possiel, N., Akhtar, F., Simon, H.A. (2012). Regional and Seasonal Analysis of North American
Background Ozone Estimates from Two Studies. Available on the Internet at:
http://www.epa.gov/ttn/naaqs/standards/ozone/s_o3_td.html
Kemball-Cook, S.; Parrish, D.; Ryerson, T; Nopmongcol, U.; Johnson, J.; Tai, E.; Yarwood, G. (2009).
Contributions of regional transport and local sources to ozone exceedances in Houston and Dallas:
Comparison of results from a photochemical grid model to aircraft and surface measurements. Journal of
Geophysical Research-Atmospheres, 114: DOOF02. DOI: 10.1029/2008JDO10248.
U.S. Environmental Protection Agency (2005). Technical Support Document for the Final Clean Air Interstate Rule
Air Quality Modeling. Office of Air Quality Planning and Standards, Research Triangle Park, NC, 285pp.
http://www.epa.gov/cair/technical.html.
U.S. Environmental Protection Agency. (2007). Review of the National Ambient Air Quality Standards for Ozone:
Policy Assessment of Scientific and Technical Information - OAQPS Staff Paper, U.S. Environmental
Protection Agency, Research Triangle Park, NC. EPA 452/R-07-007.
U.S. Environmental Protection Agency (2011). Air Quality Modeling Final Rule Technical Support Document.
Office of Air Quality Planning and Standards, Research Triangle Park, NC, 363pp.
http://www.epa.gov/airtransport/CSAPR/techinfo.html.
U.S. Environmental Protection Agency. (2012). Policy Assessment for the Review of the Ozone National Ambient
Air Quality Standards, First External Review Draft, U.S. Environmental Protection Agency, Research
Triangle Park, NC. EPA 452/P-12-002.
U.S. Environmental Protection Agency. (2013). Integrated Science Assessment for Ozone and Related
Photochemical Oxidants, U.S. Environmental Protection Agency, Research Triangle Park, NC.
EPA/600/R-10/076.
U.S. Environmental Protection Agency. (2014). Health Risk and Exposure Assessment for Ozone, Second External
Review Draft, U.S. Environmental Protection Agency, Research Triangle Park, NC. EPA-452/P-14-004a.
Wang, HQ; Jacob, DJ; Le Sager, P; Streets, DG; Park, RJ; Gilliland, AB; van Donkelaar, A. (2009). Surface ozone
background in the United States: Canadian and Mexican pollution influences. Atmos Environ 43: 1310-
1319. http://dx.doi.0rg/10.1016/j.atmosenv.2008.ll.036.
2-32
-------
Zhang, L; Jacob, DJ; Downey, NV; Wood, DA; Blewitt, D; Carouge, CC; Van donkelaar, A; Jones, DBA; Murray,
LT; 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: 6769-6776. http://dx.doi.0rg/10.1016/j.atmosenv.2011.07.054.
2-33
-------
3 Adequacy of the Current Primary Standard
This chapter presents staffs considerations and conclusions regarding the adequacy of
the current primary Os NAAQS. In doing so, we pose the following overarching question:
Does the currently available scientific evidence and exposure/risk information, as
reflected in the ISA and HREA, support or call into question the adequacy of the
current Os standard?
As discussed more fully in this chapter, staff reaches the conclusion that the available
evidence and exposure and risk information clearly calls into question the adequacy of public
health protection provided by the current primary standard. This evidence and information
provides strong support for the occurrence of a range of adverse respiratory effects, and
mortality, under air quality conditions that would meet the current standard. Based on the
analyses in the HREA, we conclude that the exposures and risks projected to remain upon
meeting the current standard are indicative of risks that can reasonably be judged to be important
from a public health perspective. Thus, staff concludes that the evidence and information
provides strong support for giving consideration to revising the current primary standard in order
to provide increased public health protection against an array of adverse health effects that range
from decreased lung function and respiratory symptoms to more serious indicators of morbidity
(e.g., including emergency department visits and hospital admissions), and mortality. The
remainder of this chapter discusses the evidence and exposure/risk information, and the
considerations and conclusions based on that evidence and information, supporting staffs
overarching conclusion regarding the adequacy of the current primary Os standard.
In addressing the overarching question for this chapter, we pose a series of more specific
questions, as discussed in sections 3.1 through 3.4 below. Section 3.1 presents our consideration
of the available scientific evidence (i.e., evidence-based considerations) about the health effects
associated with short- and long-term Os exposures. Section 3.2 presents our consideration of
available estimates of Os exposures and health risks (exposure- and risk-based considerations).
Section 3.3 discusses the advice and recommendations that we have received from the CAS AC
on the first draft Os PA, and on documents from previous reviews of the Os NAAQS. Section 3.4
revisits the overarching question of this section, and presents staffs conclusions regarding the
adequacy of the current primary Os NAAQS.
3-1
-------
3.1 EVIDENCE-BASED CONSIDERATIONS
This section presents our consideration of the available scientific evidence with regard to
the adequacy of the current Os standard. Our approach, as summarized in section 1.3.1 above, is
based on the full body of evidence in this review. We use information from the full evidence
base to characterize our confidence in the extent to which Os-attributable effects occur, and the
extent to which such effects are adverse, over the ranges of Os exposure concentrations evaluated
in controlled human exposure studies and over the distributions of ambient Os concentrations in
locations where epidemiologic studies have been conducted. In doing so, we recognize that the
available health effects evidence reflects a continuum from relatively high Os concentrations, at
which scientists generally agree that adverse health effects are likely to occur, through lower
concentrations, at which the likelihood and magnitude of a response become increasingly
uncertain.
Section 3.1.1 summarizes a mode of action framework for understanding the effects of
both short- and long-term Os exposures, based on Chapter 5 of the ISA (U.S. EPA, 2013).
Section 3.1.2 presents our consideration of the evidence for health effects attributable to short-
term and long-term Os exposures. Section 3.1.3 discusses the adversity of the effects. Section
3.1.4 presents our consideration of evidence with regard to concentrations associated with health
effects and section 3.1.5 presents our consideration of the public health implications of exposures
to Os, including the adversity of effects and evidence for at-risk populations and lifestages.1
3.1.1 Modes of Action
Our consideration of the evidence of effects attributable to short-and long-term exposures
and the factors that increase risk in populations and lifestages builds upon evidence about the
modes of action by which Os exerts effects (U.S. EPA, 2013; section 5.3). Mode of action refers
to a sequence of key events and processes that result in a given toxic effect; elucidation of
mechanisms provides a more detailed understanding of these key events and processes. The
purpose of this section is to describe the key events and pathways that contribute to health effects
resulting from both short-term and long-term exposures to Os. The extensive research carried out
1 Here, as in the ISA, the term "at-risk population" is used to encompass populations or lifestages that have a greater
likelihood of experiencing health effects related to exposure to an air pollutant due to a variety of factors; other
terms used in the literature include susceptible, vulnerable, and sensitive. These factors may be intrinsic, such as
genetic factors, lifestage, or the presence of preexisting diseases, or they may be extrinsic, such as socioeconomic
status (SES), activity pattern and exercise level, or increased pollutant exposures (U.S. EPA, 2013, p. Ixx, 8-1, 8-2).
The courts and the Act's legislative history refer to these at-risk subpopulations as "susceptible" or "sensitive"
populations. See, e.g., American Lung Ass'n v. EPA. 134 F. 3d 388, 389 (D.C. Cir. 1998) ("NAAQS must protect
not only average health individuals, but also 'sensitive citizens' - children, for example, or people with asthma,
emphysema, or other conditions rendering them particularly vulnerable to air pollution" (quoting S. Rep. No. 91-
1196 at 10).
3-2
-------
over several decades in humans and animals has yielded numerous studies on mechanisms by
which Os exerts its effects. It is well-understood that secondary oxidation products, which form
as a result of Os exposure, initiate numerous responses at the cellular, tissue and whole organ
level of the respiratory system. These responses include the activation of neural reflexes,
initiation of inflammation, alteration of barrier epithelial function, sensitization of bronchial
smooth muscle, modification of lung host defenses, and airways remodeling, as discussed below.
These key events have the potential to affect other organ systems such as the cardiovascular
system. It has been proposed that secondary oxidation products, which are bioactive and
cytotoxic in the respiratory system, are also responsible for systemic effects. Recent studies in
animal models show that inhalation of Os results in systematic oxidative stress.
Figure 3.1 below, adapted from Figure 5-8 of the ISA (ISA, Section 5.3.10, U.S. EPA,
2013), shows key events in the toxicity pathway of Os that are described in more detail below.
The initial key event in the toxicity pathway of Cb is the formation of secondary oxidation
products in the respiratory tract (ISA, section 5.3, U.S. EPA, 2013). This mainly involves direct
reactions with components of the extracellular lining fluid (ELF). Although the ELF has inherent
capacity to quench (based on individual antioxidant capacity), this capacity can be overwhelmed,
especially with exposure to elevated concentrations ofOs2 The resulting secondary oxidation
products transmit signals to the epithelium, pain receptive nerve fibers and, if present, immune
cells (i.e., eosinophils, dendritic cells and mast cells) involved in allergic responses. Thus, the
effects of Os are mediated by components of ELF and by the multiple cell types found in the
respiratory tract. Further, oxidative stress3 is an implicit part of this initial key event.
2 The ELF is a complex mixture of lipids (fats), proteins, and antioxidants that serve as the first barrier and target for
inhaled Os. The quenching ability of antioxidant substances present in the ELF appear in most cases to limit
interaction of Os with underlying tissues and to prevent penetration of Cb deeper into the lung. However, as the ELF
thickness decreases and becomes ultra thin in the alveolar region, it may be possible for direct interaction of Os with
the underlying epithelial cells to occur. The formation of secondary oxidation products is likely related to the
concentration of antioxidants present and the quenching ability of the lining fluid.
3 Oxidative stress reflects an imbalance between the systemic manifestation of reactive oxygen species, such as
superoxide, and a biological system's ability to readily detoxify the reactive intermediates or to repair the resulting
damage.
3-3
-------
Activation
of neural
reflexes
Mode of Action/Possible Pathways
Ozone + Respiratory Tract
i
Formation of secondary oxidation products
I
r T
Initiation of
inflammation
Sensitization
of bronchial
smooth muscle
Systemic inflammation and
oxidative/nitrosative stress
Extrapulmonary Effects
Decrements in pulmonary function
Pulmonary inflammation/oxidative stress
Increases in airways permeability
Airways hyperresponsiveness
Exacerbation/induction of asthma
Decreased host defenses
Epithelial metaplasia and fibrotic changes
Altered lung development
Figure 3-1. Modes of action/possible pathways underlying the health effects resulting from
inhalation exposure to Os. (Adapted from U.S. EPA, 2013, Figure 5-8)
Another key event in the toxicity pathway of Os is the activation of neural reflexes which
lead to lung function decrements. Evidence is accumulating that secondary oxidation products
are responsible for this effect. Different receptors on bronchial sensory nerves (i.e., C-fibers)
have been shown to mediate separate effects of Os on pulmonary function. For example, pain
(i.e., nociceptive) sensory nerves are involved in the involuntary truncation of inspiration which
results in decreases in FVC, FEVi, tidal volume and pain upon deep inspiration. Ozone exposure
also results in activation of vagal sensory nerves and a mild increase in airway obstruction
measured as increased sRaw. Activation of neural reflexes also results in extrapulmonary effects
such as slow resting heart rate (i.e., bradycardia).
Initiation of inflammation is also a key event in the toxicity pathway of Os. Secondary
oxidation products, as well as cell signaling molecules (i.e., chemokines and cytokines) from
airway epithelial cells and white blood cells (i.e., macrophages), have been implicated in the
initiation of inflammation. Airways neutrophilia has been demonstrated in bronchoalveolar
lavage fluid (BALF), mucosal biopsy and induced sputum samples. Influx of other cells (i.e.,
mast cells, monocytes and macrophages) also occur. Inflammation further contributes to
3-4
-------
Os-mediated oxidative stress. It should be noted that inflammation, as measured by airways
neutrophilia, is not correlated with decrements in pulmonary function as measured by
spirometry.
A fourth key event in the toxicity pathway of Os is alteration of epithelial barrier
function. Increased permeability4 occurs as a result of damage to tight junctions between
epithelial cells subsequent to Os-induced injury and inflammation. It may play a role in allergic
sensitization and in airway hyperresponsiveness (AHR). Genetic susceptibility has been
associated with this pathway.
A fifth key event in the toxicity pathway of Os is the sensitization of bronchial smooth
muscle. Airway hyperresponsiveness, or increased bronchial reactivity, can be both a rapidly
occurring and a persistent response. The mechanisms responsible for AHR are not well-
understood. Tachykinins, peptides that can excite neurons and cause smooth muscle contraction,
and the secondary oxidation products of Os have been proposed as mediators of the early
response and inflammation-derived products have been proposed as mediators of the later
response. Other chemical signaling molecules (i.e., cytokines and chemokines) have been
implicated in the AHR response to Cb in animal models. Antioxidants may confer protection.
A sixth key event in the toxicity pathway of Os is the modification of innate/adaptive
immunity. While the majority of evidence for this key event comes from animal studies, there
are several studies suggesting that this pathway may also be relevant in humans. Ozone exposure
of human subjects resulted in recruitment of activated innate immune cells to the airways.
Animal studies further linked Os-mediated activation of the innate immune system to the
development of nonspecific AHR, demonstrated an interaction between allergen and Os in the
induction of nonspecific AHR, and found that Os acted as an adjuvant for allergic sensitization
through the activation of both innate and adaptive immunity. These studies provide evidence that
Os can alter host immunologic response and lead to immune system dysfunction. These
mechanisms may underlie the exacerbation and induction of asthma, as well as decreases in lung
host defense.
Another key event in the toxicity pathway of Os is airways remodeling. Persistent
inflammation and injury, which are observed in animal models of chronic and intermittent
4 Cells in epithelium are very densely packed together, leaving very little intercellular space. All epithelial cells rest
on a basement membrane, a thin sheet of fibers that acts as scaffolding on which epithelium can grow and regenerate
after injuries. Epithelial tissue is innervated but avascular; it must be nourished by substances diffusing from the
blood vessels in the underlying tissue. Injury to epithelial cells, such as caused by oxidative stress, can cause the
epithelium to become more permeable to substances in the underlying vasculature.
3-5
-------
exposure to Cb, are associated with morphologic changes such as mucous cell metaplasia5 of
nasal epithelium, bronchiolar metaplasia of alveolar ducts and fibrotic changes in small airways
(see Section 7.2.3 of the ISA, U.S. EPA, 2013). Mechanisms responsible for these responses are
not well-understood. However, a recent study in mice demonstrated a key role for a signaling
pathway in the deposition of collagen in the airway wall following chronic intermittent exposure
to Os. Chronic intermittent exposure to Os has also been shown to result in effects on the
developing lung and immune system.
Systemic inflammation and vascular oxidative/nitrosative stress are also key events in the
toxicity pathway of Os. Extrapulmonary effects of Os occur in numerous organ systems,
including the cardiovascular, central nervous, reproductive, and hepatic systems (see
Sections 6.3 to 6.5 and Sections 7.3 to 7.5 of the ISA, U.S. EPA, 2013). It has been proposed that
lipid oxidation products resulting from reaction of Os with lipids and/or cellular membranes in
the ELF are responsible for systemic responses; however, it is not known whether they gain
access to the circulation. Alternatively, release of diffusible mediators from the lung into the
circulation may initiate or propagate inflammatory responses in the circulation or other organ
systems.
Responses to Os exposure are variable within the population. Although studies have
shown a large range of pulmonary function (i.e., spirometric) responses to Os among healthy
young adults, responses within an individual are relatively consistent over time. Other responses
to Os have also been characterized by a large degree of inter-individual variability. For example,
a 3- to 20-fold difference among subjects in their studies in airways inflammation (i.e.,
neutrophilia influx) following Os exposure has been reported (Schelegle et al., 1991 and Devlin
et al., 1991, respectively). Reproducibility of an individual's inflammatory response to Os
exposure in humans, measured as sputum neutrophilia, was demonstrated by Holz et al. (1999).
Since individual inflammatory responses were relatively consistent across time, it was thought
that inflammatory responsiveness reflected an intrinsic characteristic of the subject (Mudway and
Kelly, 2000). While the basis for the observed inter-individual variability in responsiveness to Os
is not clear, section 5.4.2 of the ISA (U.S. EPA, 2013) discusses mechanisms that may underlie
the variability in responses seen among individuals. Certain functional genetic polymorphisms,
pre-existing conditions or diseases, nutritional status, lifestages, and co-exposures contribute to
altered risk of Os-induced effects.
5 Metaplasia is the reversible replacement of one differentiated cell type with another mature differentiated cell type.
The change from one type of cell to another may generally be a part of normal maturation process or caused by
some sort of abnormal stimulus. In simplistic terms, it is as if the original cells are not robust enough to withstand
the new environment, and so they change into another type more suited to the new environment. If the stimulus that
caused metaplasia is removed or ceases, tissues return to their normal pattern of differentiation.
3-6
-------
Experimental evidence for such Os-induced changes contributes to our understanding of
the biological plausibility of adverse Os-related health effects, including a range of respiratory
effects as well as effects outside the respiratory system (e.g., cardiovascular effects) (U.S. EPA,
2013, Chapters 6 and 7).
3.1.2 Nature of Effects
• To what extent does the currently available scientific evidence alter or strengthen
our conclusions from the last review regarding health effects attributable to Os
exposure in ambient air? Are previously identified uncertainties reduced or do
important uncertainties remain?
The health effects of ozone are described in detail in the assessment of the evidence
available in this review which is largely consistent with conclusions of past Air Quality Criteria
Documents (AQCD). In some categories of health effects, there is newly available evidence
regarding some aspects of the effects described in the last review or that strengthens our
conclusions regarding aspects of Os toxicity on a particular physiological system (U.S. EPA,
2013, Table 1-1). A sizeable number of studies on Os health effects are newly available in this
review and are critically assessed in the ISA as part of the full body of evidence. Based on this
assessment, the ISA determined that a causal relationship6 exists between short-term exposure to
Os in ambient air7 and effects on the respiratory system and that a likely to be causal relationship8
exists between long-term exposure to Os in ambient air and respiratory effects (U.S. EPA, 2013,
pp. 1-6 to 1-7). As stated in the ISA, "[c]ollectively, a very large amount of evidence spanning
several decades supports a relationship between exposure to Cb and a broad range of respiratory
effects" (ISA, p. 1-6). Additionally, the ISA determined that the relationships between short-term
exposures to Os in ambient air and both total mortality and cardiovascular effects are likely to be
causal, based on expanded evidence bases in the current review (U.S. EPA, 2013, pp. 1-7 to 1-8).
In the ISA, EPA additionally determined that the currently available evidence for additional
endpoints is suggestive of causal relationships between short-term (central nervous system
effects) and long-term exposure (cardiovascular effects, central nervous system effects and total
mortality) to ambient Os. Consistent with emphasis in past reviews on Os health effects for which
6 Since the last O3 NAAQS review, the ISAs which have replaced CDs in documenting each review of the scientific
evidence (or air quality criteria) employ a systematic framework for weighing the evidence and describing
associated conclusions with regard to causality, using established descriptors, as summarized in section 1.3.1 above
(U.S. EPA, 2013, Preamble).
7 In determining that a causal relationship exists for Os with specific health effects, EPA has concluded that
"[e]vidence is sufficient to conclude that there is a causal relationship with relevant pollutant exposures" (ISA, p.
Ixiv).
8 In determining a likely to be a causal relationship exists for O3 with specific health effects, EPA has concluded that
"[e]vidence is sufficient to conclude that a causal relationship is likely to exist with relevant pollutant exposures, but
important uncertainties remain" (ISA, p. Ixiv).
3-7
-------
the evidence is strongest, we place the greatest emphasis on studies of health effects that have
been judged in the ISA to be caused by, or likely to be caused by, Os exposures (U.S. EPA,
2013, section 2.5.2).
This section presents our consideration of the evidence for health effects attributable to
Os exposures, including respiratory morbidity and mortality effects attributable to short- and
long-term exposures, and cardiovascular system effects (including mortality) and total mortality
attributable to short-term exposures. We focus particularly on considering the extent to which the
scientific evidence available in the current review has been strengthened since the last review,
and the extent to which important uncertainties and limitations in the evidence from the last
review have been addressed. In section 3.1.2.2, we then consider the extent to which the
available evidence indicates health effects may be attributable to ambient Os concentrations
likely to be allowed by the current Os NAAQS. In this section, we address the following specific
question for each category of health effects considering the evidence available in the 2008
review of the standard as well as evidence that has become available since then. The ISA
summarizes the longstanding body of evidence for Os respiratory effects as follows (U.S. EPA,
2013, p. 1-6).
The clearest evidence for health effects associated with exposure to Oi is provided
by studies of respiratory effects. Collectively, a very large amount of evidence
spanning several decades supports a relationship between exposure to O3 and a
broad range of respiratory effects (see Section 6.2.9 and Section 7.2.8). The
majority of this evidence is derived from studies investigating short-term
exposures (i.e., hours to weeks) to Os, although animal toxicological studies and
recent epidemiologic evidence demonstrate that long-term exposure (i.e., months
to years) may also harm the respiratory system.
The extensive body of evidence supporting a causal relationship between short-term Os
exposures and respiratory effects is discussed in detail in Chapter 6 of the ISA (U.S. EPA, 2013),
while evidence for respiratory effects associated with long-term or repeated Os exposures are
discussed in chapter 7 of that document (U.S., EPA, 2013).
3.1.2.1 Respiratory Effects - Short-term Exposures
• To what extent does the currently available scientific evidence, including related
uncertainties, strengthen or alter our understanding from the last review of respiratory
effects attributable to short-term Os exposures?
The 2006 Os AQCD concluded that there was clear, consistent evidence of a causal
relationship between short-term Os exposure and respiratory effects (U.S. EPA, 2006). This
conclusion was substantiated by evidence from controlled human exposure and toxicological
studies indicating a range of respiratory effects in response to short-term Os exposures, including
pulmonary function decrements and increases in respiratory symptoms, lung inflammation, lung
3-8
-------
permeability, and airway hyperresponsiveness. Toxicological studies provided additional
evidence for Os-induced impairment of host defenses. Combined, these findings from
experimental studies provided support for epidemiologic evidence, in which short-term increases
in ambient Os concentration were consistently associated with decreases in lung function in
populations with increased outdoor exposures, especially children with asthma and healthy
children; increases in respiratory symptoms and asthma medication use in children with asthma;
and increases in respiratory-related hospital admissions and asthma-related ED visits (U.S. EPA,
2013, pp. 6-1 to 6-2).
As discussed in detail in the ISA (U.S. EPA, 2013, section 6.2.9), studies evaluated since
the completion of the 2006 Os AQCD support and expand upon the strong body of evidence that,
in the last review, indicated a causal relationship between short-term Os exposures and
respiratory health effects. Recent controlled human exposure studies conducted in young, healthy
adults with moderate exertion have reported FEVi decrements and pulmonary inflammation
following prolonged exposures to Os concentrations as low as 60 ppb, and respiratory symptoms
following exposures to concentrations as low as 70 ppb.9 Epidemiologic studies provide
evidence that increases in ambient Os exposures can result in lung function decrements, increases
in respiratory symptoms, and pulmonary inflammation in children with asthma; increases in
respiratory-related hospital admissions and emergency department visits; and increases in
respiratory mortality. Some of these studies report such associations even for Os concentrations
at the low end of the distribution of daily concentrations. Recent epidemiologic studies report
that associations with respiratory morbidity and mortality are stronger during the warm/summer
months and remain robust after adjustment for copollutants. Recent toxicological studies
reporting Os-induced inflammation, airway hyperresponsiveness, and impaired lung host defense
continue to support the biological plausibility and modes of action for the Os-induced respiratory
effects observed in the controlled human exposure and epidemiologic studies. Further support is
provided by recent studies that found Os-associated increases in indicators of airway
inflammation and oxidative stress in children with asthma (U.S. EPA, 2013, section 6.2.9).
Together, epidemiologic and experimental studies support a continuum of respiratory effects
associated with Cb exposure that can result in respiratory-related emergency department visits,
hospital admissions, and/or mortality (U.S. EPA, 2013, section 6.2.9).
Across respiratory endpoints, evidence indicates antioxidant capacity may modify the
risk of respiratory morbidity associated with Os exposure (U.S. EPA, 2013, section 6.2.9, p. 6-
161) (section 3.1.1, above). The potentially elevated risk of populations with diminished
9 Schelegle et al. (2009) reported a statistically significant increase in respiratory symptoms in healthy adults at a
target Os exposure concentration of 70 ppb. For this 70 ppb target, Schelegle et al. (2009) reported an actual
exposure concentration, averaged over the study period, of 72 ppb.
3-9
-------
antioxidant capacity and the reduced risk of populations with sufficient antioxidant capacity is
supported by epidemiologic and controlled human exposure studies. Additional evidence
characterizes Os-induced decreases in antioxidant levels as a key event in the mode of action for
downstream effects.
We describe key aspects of this evidence below with regard to lung function decrements;
pulmonary inflammation, injury, and oxidative stress; airway hyperresponsiveness; respiratory
symptoms and medication use; lung host defense; allergic and asthma-related responses; hospital
admissions and emergency department visits; and respiratory mortality.
Lung Function Decrements
In the 2008 review, a large number of controlled human exposure studies reported Os-
induced lung function decrements in young, healthy adults engaged in intermittent, moderate
exertion following 6.6 hour exposures to Cb concentrations at or above 80 ppb. Although two
studies also reported effects following exposures to lower concentrations, an important
uncertainty in the last review was the extent to which exposures to Os concentrations below 80
ppb result in lung function decrements. In addition, in the last review epidemiologic panel
studies had reported Cb-associated lung function decrements in a variety of different populations
(e.g., children, outdoor workers) likely to experience increased exposures. In the current review,
additional controlled human exposure studies are available that have evaluated exposures to Os
concentrations of 60 or 70 ppb. The available evidence from controlled human exposure and
panel studies is assessed in detail in the ISA (U.S. EPA, 2013, section 6.2.1) and is summarized
below.
Controlled exposures to Os concentrations that can be found in the ambient air can result
in a number of lung function effects, including decreased inspiratory capacity; mild
bronchoconstriction; and rapid, shallow breathing patterns during exercise. Reflex inhibition of
inspiration results in a decrease in forced vital capacity (FVC) and total lung capacity (TLC) and,
in combination with mild bronchoconstriction, contributes to a decrease in the forced expiratory
volume in 1 second (FEVi) (U.S. EPA, 2013, section 6.2.1.1).10 Accumulating evidence
indicates that such effects are mediated by activation of sensory nerves, resulting in the
involuntary truncation of inspiration and a mild increase in airway obstruction due to
bronchoconstriction (U.S. EPA, 2013, section 5.3.10).
10 The controlled human exposure studies emphasized in this PA utilize only healthy adult subjects. In the near
absence of controlled human exposure data for children, HREA estimates of lung function decrements are based on
the assumption that children exhibit the same lung function responses following Os exposures as healthy 18 year
olds (U.S. EPA, 2014, sections 6.2.4 and 6.5). This assumption is justified in part by the findings of McDonnell et
al. (1985), who reported that children 8-11 year old experienced FEVi responses similar to those observed in adults
18-35 years old. Thus, the conclusions about the occurrence of lung function decrements that follow generally apply
to children as well as to adults.
3-10
-------
Data from controlled human exposure studies indicate that increasing the duration of Os
exposures and increasing ventilation rates decreases the Os exposure concentrations required to
impair lung function. Ozone exposure concentrations well above those typically found in
ambient air are required to impair lung function in healthy resting adults, while exposure to Cb
concentrations at or below those in the ambient air have been reported to impair lung function in
healthy adults exposed for longer durations while undergoing intermittent, moderate exertion
(U.S. EPA, 2013, section 6.2.1.1). With repeated Os exposures over several days, FEVi
responses become attenuated in both healthy adults and adults with mild asthma, though this
attenuation of response is lost after about a week without exposure (U.S. EPA, 2013, section
6.2.1.1; page 6-27).
When considering controlled human exposures studies of Os-induced lung function
decrements we evaluate both group mean changes in lung function and the interindividual
variability in the magnitude of responses. An advantage of Os controlled human exposure studies
(i.e., compared to the epidemiologic panel studies discussed below) is that reported effects
necessarily result from exposures to Os itself.11 To the extent studies report statistically
significant decrements in mean lung function following Os exposures after controlling for other
factors, we have more confidence that measured decrements are due to the Os exposure itself,
rather than to chance alone. As discussed below, group mean changes in lung function are often
small, especially following exposures to relatively low Os concentrations (e.g., 60 ppb).
However, even when group mean decrements in lung function are small, some individuals could
experience decrements that are "clinically meaningful" (Pellegrino et al., 2005; ATS, 1991) with
respect to criteria for spirometric testing, and/or that could be considered "adverse" with respect
to public health policy decisions (section 3.1.3 below).
At the time of the last review, a number of controlled human exposure studies had
reported lung function decrements in young, healthy adults following prolonged (6.6-hour)
exposures while at moderate exertion to Os concentrations at and above 80 ppb. In addition,
there were two controlled human exposure studies by Adams (2002, 2006) that examined lung
function effects following exposures to Os concentrations of 60 ppb. The EPA's analysis of the
data from the Adams (2006) study reported a small but statistically significant Os-induced
decrement in group mean FEVi following exposures of young, healthy adults, while at moderate
exertion, to 60 ppb Os, when compared with filtered air controls (Brown, 2008).12 Further
11 The ISA notes that the use of filtered air responses as a control for the assessment of responses following O3
exposure in controlled human exposure studies serves to eliminate alternative explanations other than O3 itself in
causing the measured responses (U.S. EPA, 2013, section 6.2.1.1).
1 9
Adams (2006) did not find effects on FEVi at 60 ppb to be statistically significant. In an analysis of the Adams
(2006) data, even after removal of potential outliers, Brown et al. (2008) found the average effect on FEVI at 60 ppb
to be small, but highly statistically significant (p < 0.002) using several common statistical tests.
3-11
-------
examination of the post-exposure FEVi data, and mean data for other time points and other
concentrations, indicated that the temporal pattern of the response to 60 ppb Os was generally
consistent with the temporal patterns of responses to higher Os concentrations in this and other
studies. (75 FR 2950, January 19, 2010). This suggested a pattern of response following
exposures to 60 ppb Os that was consistent with a dose-response relationship, rather than random
variability. See also State of Mississippi v. EPA., 744 F. 3d at 1347 (upholding EPA's
interpretation of the Adams studies).
Figure 6-1 in the ISA summarizes the currently available evidence from multiple
controlled human exposure studies evaluating group mean changes in FEVi following prolonged
Os exposures (i.e., 6.6 hours) in young, healthy adults engaged in moderate levels of physical
activity (U.S. EPA, 2013, section 6.2.1.1). With regard to the group mean changes reported in
these studies, the ISA specifically notes the following (U.S. EPA, 2013, section 6.2.1.1, Figure
6-1):
1. Prolonged exposure to 40 ppb Os 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).
2. Prolonged exposure to an average Os concentration of 60 ppb results in group mean FEVi
decrements ranging from 1.8% to 3.6% (Adams 2002; Adams, 2006;13 Schelegle et al.,
2009;14 Kim et al., 2011). Based on data from multiple studies, the weighted average
group mean decrement was 2.7%. In some analyses, these group mean decrements in
lung function were statistically significant (Brown et al., 2008; Kim et al., 2011), while in
other analyses they were not (Adams, 2006; Schelegle et al., 2009).15
3. Prolonged exposure to an average Os concentration of 70 ppb results in a statistically
significant group mean decrement in FEVi of about 6% (Schelegle et al., 2009).16
4. Prolonged square-wave exposure to average Os concentrations of 80 ppb, 100 ppb, or 120
ppb Os results in statistically significant group mean decrements in FEVi ranging from 6
to 8%, 8 to 14%, and 13 to 16%, respectively (Folinsbee et al., 1988; Horstman et al.,
1990; McDonnell et al., 1991; Adams, 2002; Adams, 2003; Adams, 2006).
13 Adams (2006; 2002) both provide data for an additional group of 30 healthy subjects that were exposed via
facemask to 60 ppb (square-wave) Os for 6.6 hours with moderate exercise (VE = 23 L/min per m2 BSA). These
subj ects are described on page 13 3 of Adams (2006) and pages 747 and 761 of Adams (2002). The FEVi decrement
may be somewhat increased due to a target VE of 23 L/min per m2 BSA relative to other studies having the target VE
of 20 L/min per m2 BSA. The facemask exposure is not expected to affect the FEVi responses relative to a chamber
exposure.
14 Schelegle et al. (2009) reported an actual mean exposure concentration of 63 ppb for the target of 60 ppb.
15 Adams (2006) did not find effects on FEVi at 60 ppb to be statistically significant. In an analysis of the Adams
(2006) data, Brown et al. (2008) addressed the more fundamental question of whether there were statistically
significant differences in responses before and after the 6.6 hour exposure period and found the average effect on
FEVi at 60 ppb to be small, but highly statistically significant using several common statistical tests, even after
removal of potential outliers.
16 Schelegle et al. (2009) reported an actual mean exposure concentration of 72 ppb for the target of 70 ppb.
3-12
-------
As illustrated in Figure 6-1 of the ISA, there is a smooth dose-response curve without
evidence of a threshold for exposures between 40 and 120 ppb Os (U.S. EPA, 2013, Figure 6-1).
When these data are taken together, the ISA concludes that "mean FEVi is clearly decreased by
6.6-h exposures to 60 ppb Os and higher concentrations in [healthy, young adult] subjects
performing moderate exercise" (U.S. EPA, 2013, p. 6-9).
With respect to interindividual variability in lung function, in an individual with
relatively "normal" lung function, with recognition of the technical and biological variability in
measurements, within-day changes in FEVi of > 5% are clinically meaningful (Pellegrino et al.,
2005; ATS, 1991). The ISA (U.S. EPA, 2013, section 6.1.) focuses on individuals with >10%
decrements in FEVi for two reasons. A 10% FEVi decrement is accepted by the American
Thoracic Society (ATS) as an abnormal response and a reasonable criterion for assessing
exercise-induced bronchoconstriction (Dryden et al., 2010; ATS, 2000). (U.S. EPA, 2013,
section 6.2.1.1). Also, some individuals in the Schelegle et al. (2009) study experienced 5-10%
FEVi decrements following exposure to filtered air.
In previous NAAQS reviews, the EPA has made judgments regarding the potential
implications for individuals experiencing FEVi decrements of varying degrees of severity.17 For
people with lung disease, the EPA judged that moderate functional decrements (e.g., FEVi
decrements > 10 percent but < 20 percent, lasting up to 24 hours) would likely interfere with
normal activity for many individuals, and would likely result in more frequent use of medication
(75 FR 2973, January 19, 2010). In previous reviews CAS AC has endorsed these conclusions. In
the context of standard setting, in the last review Os review CASAC indicated that it is
appropriate to focus on the lower end of the range of moderate functional responses (e.g., FEVi
decrements > 10 percent) when estimating potentially adverse lung function decrements in
people with lung disease, especially children with asthma (Henderson, 2006). More specifically,
CASAC stated that "[a] 10% decrement in FEVi can lead to respiratory symptoms, especially in
individuals with pre-existing pulmonary or cardiac disease. For example, people with chronic
obstructive pulmonary disease have decreased ventilatory reserve (i.e., decreased baseline FEVi)
such that a > 10% decrement could lead to moderate to severe respiratory symptoms" (Samet,
2011). In this review, CASAC reiterated its support for this conclusion, stating that "[a]n FEVi
decrement of >10% is a scientifically relevant surrogate for adverse health outcomes for people with
asthma and lung disease" (Frey, 2014 p. 3). Therefore, in considering interindividual variability in
17 Such judgments have been made for decrements in FEVi as well as for increased airway responsiveness and
symptomatic responses (e.g., cough, chest pain, wheeze). Ranges of pulmonary responses and their associated
potential impacts are presented in Tables 3-2 and 3-3 of the Staff Paper (U.S. EPA, 2007).
3-13
-------
Os-induced lung function decrements in the current review, we also focus on the extent to which
individuals were reported to experience FEVi decrements of 10% or greater.
New studies (Schelegle et al., 2009; Kim et al., 2011) add to the previously available
evidence for interindividual variability in the responses of healthy adults following exposures to
Os. Following prolonged exposures to 80 ppb Os while at moderate exertion, the proportion of
healthy adults experiencing FEVi decrements greater than 10% was 17% by Adams (2006), 26%
by McDonnell (1996), and 29% by Schelegle et al. (2009). Following exposures to 60 ppb Os,
that proportion was 20% by Adams (2002), 3% by Adams (2006), 16% by Schelegle et al.
(2009), and 5% by Kim et al. (2011). Based on these studies, the weighted average proportion of
young, healthy adults with >10% FEVi decrements is 25% following exposure to 80 ppb Os and
10% following exposure to 60 ppb Os (U.S. EPA, 2013, page 6-19).18 The ISA notes that
responses within an individual tend to be reproducible over a period of several months,
indicating that interindividual differences reflect differences in intrinsic responsiveness. Given
this, the ISA concludes that "a considerable fraction" of healthy individuals experience clinically
meaningful decrements in lung function when exposed for 6.6 hours to 60 ppb Os during quasi
continuous, moderate exertion (U.S. EPA, 2013, section 6.2.1.1, p. 6-20).
As discussed above (Figure 3-1) and in the ISA (U.S EPA, 2013, Section 5.3.2),
secondary oxidation products formed following Os exposures can activate neural reflexes leading
to decreased lung function. Two new quantitative models, discussed in section 6.2.1.1 of the ISA
(U.S. EPA, 2013, p. 6-15), included mathematical approaches to simulate the protective effect of
antioxidants in the ELF at lower ambient Os concentrations, and include a threshold below which
changes in lung function do not occur (McDonnell et al., 2012; Schelegle et al., 2012).
McDonnell et al. (2012) and Schelegle et al. (2012) developed models using data on Os
exposure concentrations, ventilation rates, duration of exposures, and lung function responses
from a number of controlled human exposure studies. The McDonnell et al. (2012) and Schelegle
et al. (2012) studies analyzed large datasets to fit compartmental models that included the
concept of a dose of onset in lung function response or a response threshold based upon the
inhaled Os dose. The first compartment in the McDonnell et al. (2012) model considers the level
of oxidant stress in response to Os exposure to increase over time as a function of dose rate
(C>
-------
as a sigmold-shaped function. In the Schelegle et al. (2012) model, a first compartment acts as a
reservoir in which oxidant stress builds up until the dose of onset, at which time it spills over into
a second compartment. The second compartment is identical to the first compartment in
McDonnell et al. (2012) model. The oxidant levels in the second compartment were multiplied
by a responsiveness coefficient to predict FEVi responses for the Schelegle et al. (2012) model.
The McDonnell et al. (2012) model was fit to a large dataset consisting of the FEVi
responses of 741 young, healthy adults (18-35 years of age) from 23 individual controlled
exposure studies. Concentrations across individual studies ranged from 40 ppb to 400 ppb,
activity level ranged from rest to heavy exercise, duration of exposure was from 2 to 7.6 hours.
The extension of the McDonnell et al. (2012) model to children and older adults is discussed in
section 6.2.1 of the ISA (U.S. EPA, 2013). Schelegle et al. (2012) also analyzed a large dataset
with substantial overlap to that used by McDonnell et al. (2012). The Schelegle et al. (2012)
model was fit to the FEVi responses of 220 young healthy adults (taken from a dataset of 704
individuals) from 21 individual controlled exposure studies. The resulting empirical models can
estimate the frequency distribution of individual lung function responses for any exposure
scenario as well as summary measures of the distribution such as the mean or median response
and the proportions of individuals with FEVi decrements > 10%, 15%, and 20%.
The predictions of the McDonnell and Schelegle models are consistent with the observed
results from the individual studies of Os-induced FEVi decrements. Specifically, the model
developed by McDonnell et al. (2012) predicts that 9% of healthy exercising adults would
experience FEVi decrements greater than 10% following 6.6 hour exposure to 60 ppb Os, and
that 22% would experience such decrements following exposure to 80 ppb Os (U.S. EPA, 2013,
p. 6-18 and Figure 6-3). The model developed by Schelegle et al. (2012) predicts that, for a
prolonged (6.6 hours) Os exposure with moderate, quasi continuous exercise, the average dose of
onset for FEVi decrement would be reached following 4 to 5 hours of exposure to 60 ppb, and
following 3 to 4 hours of exposure to 80 ppb. However, 14% of the individuals had a dose of
onset that was less than 40% of the average. Those individuals would reach their dose of onset
following 1 to 2 hours of exposure to 50 to 80 ppb Os (U.S. EPA, 2013, p. 6-16), which is
consistent with the threshold FEVi responses reported by McDonnell et al. (2012).
Epidemiologic studies19 have consistently linked short-term increases in ambient Os
concentrations with lung function decrements in diverse populations and lifestages, including
children attending summer camps, adults exercising or working outdoors, and groups with pre-
existing respiratory diseases such as asthmatic children (U.S. EPA, 2013, section 6.2.1.2). Some
19
Unless otherwise specified, the epidemiologic studies discussed in this PA evaluate only adults.
3-15
-------
of these studies reported ozone-associated lung function decrements accompanied by respiratory
symptoms20 in asthmatic children (Just et al., 2002; Mortimer et al., 2002; Ross et al., 2002;
Gielen et al., 1997; Romieu et al., 1997; Thurston et al., 1997; Romieu et al., 1996). In contrast,
studies of children in the general population have reported similar Os-associated lung function
decrements but without accompanying respiratory symptoms (Ward et al., 2002; Gold et al.,
1999; Linn et al., 1996) (U.S. EPA, 2013, section 6.2.1.2).
Several epidemiologic panel studies reported that associations with lung function
decrements persisted at relatively low ambient Os concentrations. For outdoor recreation or
exercise, associations were reported in analyses restricted to 1-hour average Os concentrations
less than 80 ppb (Spektor et al., 1988a; Spektor et al., 1988b), 60 ppb (Brunekreef et al., 1994;
Spektor et al., 1988a), and 50 ppb (Brunekreef et al., 1994). Among outdoor workers, Brauer et
al. (1996) found a robust association using daily 1-hour max Os concentrations less than 40 ppb.
Ulmer et al. (1997) found a robust association in schoolchildren using 30-minute maximum Os
concentrations less than 60 ppb. For 8-hour average Os concentrations, associations with lung
function decrements in children with asthma were found to persist at concentrations less than
80 ppb in a U.S. multicity study (Mortimer et al., 2002) and less than 51 ppb in a study
conducted in the Netherlands (Gielen et al., 1997).
Epidemiologic panel studies investigating the effects of short-term exposure to Os
provided information on potential confounding by copollutants such as PM2.5, PMio, NCh, or
SCh. These studies varied in how they evaluated confounding. Some studies of subjects
exercising outdoors indicated that ambient concentrations of copollutants such as NCh, SCh, or
acid aerosol were low, and thus not likely to confound associations observed for Os (Hoppe et
al., 2003; Brunekreef et al., 1994; Hoek et al., 1993). In other studies of children with increased
outdoor exposures, Os was consistently associated with decreases in lung function, whereas other
pollutants such as PM2.5, sulfate, and acid aerosol individually showed variable associations
across studies (Thurston et al., 1997; Castillejos et al., 1995; Berry et al., 1991; Avol et al., 1990;
Spektor et al., 1988a). Studies that conducted copollutant modeling generally found
Os-associated lung function decrements to be robust (i.e., most copollutant-adjusted effect
estimates fell within the 95% CI of the single-pollutant effect estimates) (U.S. EPA, 2013, Figure
6-10 and Table 6-14). Most Os effect estimates for lung function were robust to adjustment for
temperature, humidity, and copollutants such as PM2.5, PMio, NCh, or SCh. Although examined
20 Reversible loss of lung function in combination with the presence of symptoms meets the ATS definition of
adversity (ATS, 2000).
3-16
-------
in only a few epidemiologic studies, Os also remained associated with decreases in lung function
with adjustment for pollen or acid aerosols (U.S. EPA, 2013, section 6.2.1.2).
Several epidemiologic studies demonstrated the protective effects of vitamin E and
vitamin C supplementation, and increased dietary antioxidant intake, on Os-induced lung
function decrements (Romieu et al., 2002) (U.S. EPA, 2013, Figure 6-7 and Table 6-8).21 These
results provide support for the new, quantitative models (McDonnell et al., 2012; Schelegle et
al., 2012), discussed above, which make use of the concept of oxidant stress to estimate the
occurrence of lung function decrements following exposures to relatively low Os concentrations.
In conclusion, new information from controlled human exposure studies considerably
strengthens the evidence and reduces the uncertainties, relative to the evidence that was available
at the time of the 2008 review, regarding the presence and magnitude of lung function
decrements in healthy adults following prolonged exposures to Os concentrations below 80 ppb.
As discussed in Section 6.2.1.1 in the ISA (U.S. EPA, 2013, p. 6-12), there is information
available from four separate studies that evaluated exposures to 60 ppb Ch (Kim et al., 2011;
Schelegle et al., 2009; Adams 2002; 2006). Although not consistently statistically significant,
group mean FEVi decrements following exposures to 60 ppb Cb are consistent among these
studies. Moreover, as is illustrated in Figure 6-1 of the ISA (U.S. EPA, 2013), the group mean
FEVi responses at 60 ppb fall on a smooth intake dose-response curve for exposures between 40
and 120 ppb Os. These studies also indicate that, on average, 10% of young, healthy adults
experience clinically meaningful decrements in lung function when exposed for 6.6 hours to 60
ppb Os during intermittent, moderate exertion. One recent study has also reported statistically
significant decrements following exposures to 70 ppb Os (Schelegle et al., 2009). Predictions
from newly developed quantitative models, based on the concept that Os-induced oxidation
results in lung function decrements, are consistent with these experimental results. Additionally,
as discussed in more detail in section 3.1.4 below, epidemiologic studies continue to provide
evidence of lung function decrements in people who are active outdoors, including people
engaged in outdoor recreation or exercise, children, and outdoor workers, at low ambient Os
concentrations. While few new epidemiologic studies of Os-associated lung function decrements
are available in this review, previously available studies have reported associations with
decrements, including at relatively low ambient Os concentrations.
Pulmonary Inflammation, Injury, and Oxidative Stress
Ozone exposures result in increased respiratory tract inflammation and epithelial
permeability. Inflammation is a host response to injury, and the induction of inflammation is
21 Evidence from controlled human exposure studies is mixed, suggesting that supplementation may be ineffective
in the absence of antioxidant deficiency (U.S. EPA, 2013, p. 5-63).
3-17
-------
evidence that injury has occurred. Oxidative stress has been shown to play a key role in initiating
and sustaining Os-induced inflammation. Secondary oxidation products formed as a result of
reactions between Os and components of the ELF can increase the expression of molecules (i.e.,
cytokines, chemokines, and adhesion molecules) that can enhance airway epithelium
permeability (U.S. EPA, 2013, Sections 5.3.3 and 5.3.4). As discussed in detail in the ISA (U.S.
EPA, 2013, section 6.2.3), Os exposures can initiate an acute inflammatory response throughout
the respiratory tract that has been reported to persist for at least 18-24 hours after exposure.
Inflammation induced by exposure of humans to Os 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 asthma; (4) inflammation can alter the body's host
defense response to inhaled microorganisms, particularly in potentially at-risk populations or
lifestages such as the very young and old; and (5) inflammation can alter the lung's response to
other agents such as allergens or toxins (U.S. EPA, 2013, Section 6.2.3). Thus, lung injury and
the resulting inflammation provide a mechanism by which Cb may cause other more serious
morbidity effects (e.g., asthma exacerbations).
In the last review, controlled human exposure studies reported Os-induced airway
inflammation following exposures at or above 80 ppb. In the current review, the link between Os
exposures and airway inflammation and injury has been evaluated in additional controlled human
exposure studies, as well as in recent epidemiologic studies. Controlled human exposure studies
have generally been conducted in young, healthy adults or in adults with asthma using lavage
(proximal airway and bronchoalveolar), bronchial biopsy, and more recently, induced sputum.
These studies have evaluated one or more indicators of inflammation, including neutrophil22
(PMN) influx, markers of eosinophilic inflammation, increased permeability of the respiratory
epithelium, and/or prevalence of proinflammatory molecules (U.S. EPA, 2013, section 6.2.3.1).
Epidemiologic studies have generally evaluated associations between ambient Os and markers of
inflammation and/or oxidative stress, which plays a key role in initiating and sustaining
inflammation (U.S. EPA, 2013, section 6.2.3.2).
There is an extensive body of evidence from controlled human exposure studies
indicating that short-term exposures to Os can cause pulmonary inflammation. Previously
22 Referred to as either neutrophils or polymorphonuclear neutrophils (or PMNs), these are the most abundant type
of white blood cells in mammals. PMNs are recruited to the site of injury following trauma and are the hallmark of
acute inflammation. The presence of PMNs in the lung has long been accepted as a hallmark of inflammation and is
an important indicator that Os causes inflammation in the lungs. Neutrophilic inflammation of tissues indicates
activation of the innate immune system and requires a complex series of events, that then are normally followed by
processes that clear the evidence of acute inflammation.
3-18
-------
available evidence indicated that Os causes an inflammatory response in the lungs (U.S. EPA,
1996). A single acute exposure (1-4 hours) of humans to moderate concentrations of Os (200-
600 ppb) while exercising at moderate to heavy intensities resulted in a number of cellular and
biochemical changes in the lung, including inflammation characterized by increased numbers of
PMNs, increased permeability of the epithelial lining of the respiratory tract, cell damage, and
production of proinflammatory molecules (i.e., cytokines and prostaglandins, U.S. EPA, 2006).
A meta-analysis of 21 controlled human exposure studies (Mudway and Kelly, 2004) using
varied experimental protocols (80-600 ppb Os exposures; 1-6.6 hours exposure duration; light to
heavy exercise; bronchoscopy at 0-24 hours post-Os exposure) reported that PMN influx in
healthy subjects is linearly associated with total Os dose. Animal toxicological studies also
provided evidence for increases in inflammation and permeability in rabbits at levels as low as
100 ppb O3 (Section 2.5.3.1, ISA, U.S. EPA, 2013).
Several studies, including one published since the last review (Alexis et al., 2010), have
reported Os-induced increases in PMN influx and permeability following exposures at or above
80 ppb (Alexis et al., 2010; Peden et al., 1997; Devlin et al., 1991), and eosinophilic
inflammation following exposures at or above 160 ppb (Scannell et al., 1996; Peden et al., 1997;
Hiltermann et al., 1999; Vagaggini et al., 2002). In addition, one recent controlled human
exposure study has reported Os-induced PMN influx following exposures of healthy adults to Os
concentrations of 60 ppb (Kim et al., 2011), the lowest concentration at which inflammatory
responses have been evaluated in human studies.
As with FEVi responses to Os, inflammatory responses to Os are generally reproducible
within individuals, with some individuals experiencing more severe Os-induced airway
inflammation than indicated by group averages (Holz et al., 2005; Holz et al., 1999). Unlike Os-
induced decrements in lung function, which are attenuated following repeated exposures over
several days (U.S. EPA, 2013, section 6.2.1.1), some markers of Os-induced inflammation and
tissue damage remain elevated during repeated exposures, indicating ongoing damage to the
respiratory system (U.S. EPA, 2013, section 6.2.3.1, p. 6-81).
Most controlled human exposure studies have reported that asthmatics experience larger
Os-induced inflammatory responses than non-asthmatics. Specifically, asthmatics exposed to
200 ppb Os for 4-6 hours with exercise show significantly more neutrophils in bronchoalveolar
lavage fluid (BALF) than similarly exposed healthy individuals (Scannell et al., 1996; Basha et
al., 1994). Bosson et al. (2003) reported significantly greater expression of a variety of pro-
inflammatory cytokines in asthmatics, compared to healthy subjects, following exposure to
200 ppb Os for 2 hours. In addition, research available in the last review, combined with a recent
study newly available in this review, indicates that pretreatment of asthmatics with
corticosteroids can prevent the Os-induced inflammatory response in induced sputum, though
3-19
-------
pretreatment did not prevent FEVi decrements (Vagaggini et al., 2001; 2007). In contrast,
Stenfors et al. (2002) did not detect a difference in the Os-induced increases in neutrophil
numbers between 15 subjects with mild asthma and 15 healthy subjects by bronchial wash at the
6 hours postexposure time point, although the neutrophil increase in the asthmatic group was on
top of an elevated baseline.
In people with allergic airway disease, including people with rhinitis and asthma,
evidence available in the last review indicated that proinflammatory mediators also cause
accumulation of eosinophils in the airways (Torres et al., 1996; Peden et al.,1995 and 1997;
Frampton et al., 1997; Hiltermann et al., 1999; Holz et al., 2002; Vagaggini et al., 2002). The
eosinophil, which increases inflammation and allergic responses, is the cell most frequently
associated with exacerbations of asthma (75 CFR 2969, January 19, 2010).
Studies reporting inflammatory responses and markers of lung injury have clearly
demonstrated that there is important variation in the responses of exposed subjects (75 FR 2953,
January 19, 2010). Some individuals also appear to be intrinsically more susceptible to increased
inflammatory responses from Os exposure (Holz et al., 2005). In healthy adults exposed to each
80 and 100 ppb Os, Devlin et al. (1991) observed group average increases in neutrophilic
inflammation of 2.1- and 3.8-fold, respectively. However, there was a 20-fold range in
inflammatory responses between individuals at both concentrations. Relative to an earlier,
similar study conducted at 400 ppb (Koren et al., 1989), Devlin et al. (1991) noted that although
some of the study population showed little or no increase in inflammatory and cellular injury
indicators analyzed after exposures to lower levels of Os (i.e., 80 and 100 ppb), others had
changes that were as large as those seen when subjects were exposed to 400 ppb Os. The data
suggest that as a whole the healthy population, on average, may have small inflammatory
responses to near-ambient levels of Os, though there may be a substantial subpopulation that is
very sensitive to low levels of Os. Devlin et al. (1991) expressed the view that "susceptible
subpopulations such as the very young, elderly, and people with pulmonary impairment or
disease may be even more affected."
A number of studies report that Os exposures increase epithelial permeability. Increased
BALF protein, suggesting Os-induced changes in epithelial permeability, has been reported at
1 hour and 18 hours postexposure (Devlin et al., 1997; Balmes et al., 1996). A meta-analysis of
results from 21 publications (Mudway and Kelly, 2004) for varied experimental protocols (80-
600 ppb Os; 1-6.6 hours duration; light to heavy exercise; bronchoscopy at 0-24 hours post-Os
exposure; healthy subjects), showed that increased BALF protein is associated with total inhaled
Os dose (i.e., the product of Os concentration, exposure duration, and VE). As noted in the 2009
PM ISA (U.S. EPA, 2009), it has been postulated that changes in permeability associated with
acute inflammation may provide increased access of inhaled antigens, particles, and other
3-20
-------
inhaled substances deposited on lung surfaces to the smooth muscle, interstitial cells, immune
cells underlying the epithelium, and the blood (U.S. EPA, 2013, sections 5.3.4, 5.3.5). Animal
toxicology studies have provided some support for this hypothesis (Adamson and Prieditis, 1995;
Chen et al., 2006), though these studies did not specifically evaluate Os exposures (U.S. EPA,
2009). Because of this potentially increased access, it has been postulated that increases in
epithelial permeability following Os exposure might lead to increases in airway responsiveness
to specific and nonspecific agents. In a recent study, Que et al. (2011) investigated this
hypothesis in healthy young adults (83M, 55 F) exposed to 220 ppb Os for 2.25 hours
(alternating 15 min periods of rest and brisk treadmill walking). As has been observed for FEVi
responses, within-individual changes in permeability were correlated between sequential Os
exposures, indicating intrinsic differences among individuals in susceptibility to epithelial
damage following Os exposures. However, increases in epithelial permeability at 1 day post-Os
exposure were not correlated with with changes in airway responsiveness assessed 1 day post-Os
exposure. The authors concluded that changes in epithelial permeability is relatively constant
over time in young healthy adults, although changes in permeability and AHR appear to be
mediated by different physiologic pathways.
The limited epidemiologic evidence reviewed in the 2006 Os AQCD (U.S. EPA, 2006)
demonstrated an association between short-term increases in ambient Os concentrations and
airways inflammation in children (1-hour max Os of approximately 100 ppb). In the 2006 Os
AQCD (U.S. EPA, 2006), there was limited evidence for increases in nasal lavage levels of
inflammatory cell counts and molecules released by inflammatory cells (i.e., eosinophilic
cationic protein, and myeloperoxidases). Since 2006, as a result of the development of less
invasive methods, there has been a large increase in the number of studies assessing ambient Os-
associated changes in airway inflammation and oxidative stress, the types of biological samples
collected, and the types of indicators. Most of these recent studies have evaluated biomarkers of
inflammation or oxidative stress in exhaled breath, nasal lavage fluid, or induced sputum (U.S.
EPA, 2013, section 6.2.3.2). These recent studies form a larger database to establish coherence
with findings from controlled human exposure and animal studies that have measured the same
or related biological markers. Additionally, results from these studies provide further biological
plausibility for the associations observed between ambient Os concentrations and respiratory
symptoms and asthma exacerbations.
A number of epidemiologic studies provide evidence that short-term increases in ambient
Os exposure increase pulmonary inflammation and oxidative stress in children, including those
with asthma (Sienra-Monge et al., 2004; Barraza-Villarreal et al., 2008; Romieu et al., 2008;
3-21
-------
Berhane et al., 2011). Multiple studies examined and found increases in exhaled NO (eNO)23
(Berhane et al., 2011; Khatri et al., 2009; Barraza-Villarreal et al., 2008). In some studies of
subjects with asthma, increases in ambient Os concentration at the same lag were associated with
both increases in pulmonary inflammation and respiratory symptoms (Khatri et al., 2009;
Barraza-Villarreal et al., 2008). Although more limited in number, epidemiologic studies also
found associations with cytokines such as IL-6 or IL-8 (Barraza-Villarreal et al., 2008; Sienra-
Monge et al., 2004), eosinophils (Khatri et al., 2009), antioxidants (Sienra-Monge et al., 2004),
and indicators of oxidative stress (Romieu et al., 2008) (ISA, Section 6.2.3.2, U.S. EPA, 2013).
Because associations with inflammation were attenuated with higher antioxidant intake the study
by Sienra-Monge et al. (2004) provides additional evidence that inhaled Os is likely to be an
important source of reactive oxygen species in airways and/or may increase pulmonary
inflammation via oxidative stress-mediated mechanisms among all age groups. Limitations in
some recent studies have contributed to inconsistent results in adults (U.S. EPA, 2013, section
6.2.3.2).
Exposure to ambient Os on multiple days can result in larger increases in pulmonary
inflammation and oxidative stress, as discussed in section 6.2.3.2 of the ISA (U.S. EPA, 2013).
In studies that examined multiple Os lags, multiday averages of 8-hour maximum or
8-hour average concentrations were associated with larger increases in pulmonary inflammation
and oxidative stress (Berhane et al., 2011; Delfino et al., 2010a; Sienra-Monge et al., 2004),
consistent with controlled human exposure (U.S. EPA, 2013, section 6.2.3.1) and animal studies
(U.S. EPA, 2013, section 6.2.3.3) reporting that some markers of pulmonary inflammation
remain elevated with Os exposures repeated over multiple days. Evidence from animal
toxicological studies also clearly indicates that Os exposures result in damage and inflammation
in the lung (ISA, Section 5.3, U.S. EPA, 2013). In the few studies that evaluated the potential for
confounding, Os effect estimates were not confounded by temperature or humidity, and were
robust to adjustment for PM2.5 or PMio (Barraza-Villarreal et al., 2008; Romieu et al., 2008;
Sienra-Monge et al., 2004).
In conclusion, a relatively small number of controlled human exposure studies evaluating
Os-induced airway inflammation have become available since the last review. For purposes of
reviewing the current Os NAAQS, the most important of these recent studies reported a
statistically significant increase in airway inflammation in healthy adults at moderate exertion
following exposures to 60 ppb Os, the lowest concentration that has been evaluated for
inflammation. In addition, a number of recent epidemiologic studies report Os-associated
23 Exhaled NO has been shown to be a useful biomarker for airway inflammation in large population-based studies
(ISA, U.S. EPA, 2013, Section 7.2.4).
3-22
-------
increases in markers of pulmonary inflammation, particularly in children. Thus, recent studies
continue to support the evidence for airway inflammation and injury that was available in
previous reviews, with new evidence for such effects following exposures to lower
concentrations than had been evaluated previously.
Airway Hyperresponsiveness
Airway hyperresponsiveness (AHR) refers to a condition in which the conducting
airways undergo enhanced bronchoconstriction in response to a variety of stimuli. Airway
hyperresponsiveness is an important consequence of exposure to ambient Os because its presence
reflects a change in airway smooth muscle reactivity, and indicates that the airways are
predisposed to narrowing upon inhalation of a variety of ambient stimuli including specific
triggers (i.e., allergens) and nonspecific triggers (e.g., SCh, and cold air). People with asthma are
generally more sensitive to bronchoconstricting agents than those without asthma, and the use of
an airway challenge to inhaled bronchoconstricting agents is a diagnostic test in asthma.
Standards for airway responsiveness testing have been developed for the clinical laboratory
(ATS, 2000), although variation in the methodology for administering the bronchoconstricting
agent may affect the results (Cockcroft et al., 2005). There is a wide range of airway
responsiveness in people without asthma, and responsiveness is influenced by a number of
factors, including cigarette smoke, pollutant exposures, respiratory infections, occupational
exposures, and respiratory irritants. Dietary antioxidants have been reported to attenuate Os-
induced bronchial hyperresponsiveness in people with asthma (Trenga et al., 2001).
Evidence for airway hyperresponsiveness following Os exposures is derived primarily
from controlled human exposure and toxicological studies (U.S. EPA, 2013, section 6.2.2).
Airway responsiveness is often quantified by measuring changes in pulmonary function
following the inhalation of an aerosolized allergen or a nonspecific bronchoconstricting agent
(e.g., methacholine), or following exposure to a bronchoconstricting stimulus such as cold air. In
the last review, controlled human exposure studies of mostly adults (> 18 years of age) had
shown that exposures to Cb concentrations at or above 80 ppb increase airway responsiveness, as
indicated by a reduction in the concentration of specific (e.g., ragweed) and non-specific (e.g.,
methacholine) agents required to produce a given reduction in lung function (e.g., as measured
by FEVi or specific airway resistance) (U.S. EPA, 2013, section 6.2.2.1). This Os-induced AHR
has been reported to be dose-dependent (Horstman et al., 1990). Animal toxicology studies have
reported Os-induced airway hyperresponsiveness in a number of species, with some rat strains
exhibiting hyperresponsiveness following 4-hour exposures to Os concentrations as low as 50
ppb (Depuydt et al., 1999). Since the last review, there have been relatively few new controlled
human exposure and animal toxicology studies of Os and airway hyperresponsiveness, and no
3-23
-------
new studies have evaluated exposures to Os concentrations at or below 80 ppb (U.S. EPA, 2013,
section 6.2.2.1)
Airway hyperresponsiveness is linked with the accumulation and/or activation of
eosinophils in the airways of asthmatics, which is followed by production of mucus and a late-
phase asthmatic response (75 FR 2970, January 19, 2010). In a study of 16 intermittent
asthmatics, Hiltermann et al. (1999) found that there was a significant inverse correlation
between the Os-induced change in the percentage of eosinophils in induced sputum and the
concentration of methacholine causing a 20% decrease in FEVi. Hiltermann et al. (1999)
concluded that the results point to the role of eosinophils in Os-induced airway
hyperresponsiveness. Increases in Os-induced nonspecific airway responsiveness incidence and
duration could have important clinical implications for children and adults with asthma, such as
exacerbations of their disease.
Airway hyperresponsiveness after Os exposure appears to resolve more slowly than
changes in FEVi or respiratory symptoms (Folinsbee and Hazucha, 2000). Studies suggest that
Os-induced AHR usually resolves 18 to 24 hours after exposure, but may persist in some
individuals for longer periods (Folinsbee and Hazucha, 1989). Furthermore, in studies of
repeated exposure to Os, changes in AHR tend to be somewhat less susceptible to attenuation
with consecutive exposures than changes in FEVi (Gong et al., 1997; Folinsbee et al., 1994;
Kulle et al., 1982; Dimeo et al., 1981) (U.S. EPA, 2013, section 6.2.2). In animal studies a 3-day
continuous exposure resulted in attenuation of Os-induced airway hyperresponsiveness (Johnston
et al., 2005) while repeated exposures for 2 hours per day over 10 days did not (Chhabra et al.,
2010), suggesting that attenuation could be lost when repeated exposures are interspersed with
periods of rest (U.S. EPA, 2013, section 6.2.2.2).
Increases in airway responsiveness do not appear to be strongly associated with
decrements in lung function or increases in symptoms (Aris et al., 1995). Recently, Que et al.
(2011) assessed methacholine responsiveness in healthy young adults (83M, 55 F) one day after
exposure to 220 ppb Os and filtered air for 2.25 hours (alternating 15 minute periods of rest and
brisk treadmill walking). Increases in airways responsiveness at 1 day post-Os exposure were not
correlated with FEVi responses immediately following the Os exposure or with changes in
epithelial permeability assessed 1-day post-Os exposure. This indicates that airway hyper-
responsiveness also appears to be mediated by a differing physiologic pathway.
As mentioned above, in addition to human subjects a number of species, including
nonhuman primates, dogs, cats, rabbits, and rodents, have been used to examine the effect of Os
exposure on airway hyperresponsiveness (see Table 6-14, (U.S. EPA, 1996) of the 1996 Os
AQCD and Annex Table AX5-12 on page AX5-36 (U.S. EPA, 2006) of the 2006 O3 AQCD). A
body of animal toxicology studies, including some recent studies conducted since the last review,
3-24
-------
provides support for the Os-induced AHR reported in humans (U.S. EPA, 2013, section 6.2.2.2).
Although most of these studies evaluated Os concentrations above those typically found in
ambient air in cities in the United States (i.e., most studies evaluated Os concentrations of 100
ppb or greater), one study reported that a very low exposure concentration (50 ppb for 4 hours)
induced AHR in some rat strains (Depuydt et al., 1999). Additional recent rodent studies
reported Os-induced AHR following exposures to Os concentrations from 100 to 500 ppb
(Johnston et al., 2005; Chhabra et al., 2010; Larsen et al., 2010). In characterizing the relevance
of these exposure concentrations, the ISA noted that a study using radiolabeled Os suggests that
even very high Os exposure concentrations in rodents could be equivalent to much lower
exposure concentrations in humans. Specifically, a 2000 ppb (2 ppm) Os exposure concentration
in resting rats was reported to be roughly equivalent to a 400 ppb exposure concentration in
exercising humans (Hatch et al., 1994). Given this relationship, the ISA noted that animal data
obtained in resting conditions could underestimate the risk of effects for humans (U.S. EPA,
2013, section 2.4, p. 2-14).
The 2006 AQCD (U.S. EPA, 2006, p. 6-34) concluded that spirometric responses to Os
are independent of inflammatory responses and markers of epithelial injury or integrity (Balmes
et al., 1996; Blomberg et al., 1999; Torres et al., 1997). Significant inflammatory responses to Os
exposures that did not elicit significant spirometric responses have been reported (Holz et al.,
2005). A recent study (Que et al., 2011) indicates that airway hyper-responsiveness also appears
to be mediated by a differing physiologic pathway. These results from controlled human
exposure studies indicate that sub-populations of healthy study subjects consistently experience
larger than average lung function decrements, greater than average inflammatory responses and
pulmonary injury as expressed by increased epithelial permeability, and greater than average
airway responsiveness, and that these effects are mediated by apparently different physiologic
pathways. Except for lung function decrements, we do not have the concentration- or exposure-
response function information about the other, potentially more sensitive,24 clinical endpoints
(i.e., inflammation, increased epithelial permeability, airway hyperresponsiveness) that would
allow us to quantitatively estimate the size of the population affected and the magnitude of their
responses. Moreover, some uncertainties about the exact physiological pathways underlying
these endpoints prevents us from knowing whether the exaggerated responses are distributed in
sub-populations evenly across the population, or may be clustered with more than one type of
exaggerated response in particular sub-populations, or both.
24 CAS AC noted that "... [W]hile measures of FEVi are quantitative and readily obtainable in humans, they are not
the only measures — and perhaps not the most sensitive measures — of the adverse health effects induced by ozone
exposure." (Henderson, 2006).
3-25
-------
In summary, a strong body of controlled human exposure and animal toxicological
studies, most of which were available in the last review of the Os NAAQS, report Os-induced
airway hyperresponsiveness after either acute or repeated exposures (U.S. EPA, 2013, section
6.2.2.2). People with asthma often exhibit increased airway responsiveness at baseline relative to
healthy controls, and they can experience further increases in responsiveness following
exposures to Os. Studies reporting increased airway responsiveness after Os exposure contribute
to a plausible link between ambient Os exposures and increased respiratory symptoms in
asthmatics, and increased hospital admissions and emergency department visits for asthma (U.S.
EPA, 2013, section 6.2.2.2).
Respiratory Symptoms and Medication Use
Because respiratory symptoms are associated with adverse outcomes such as limitations
in activity, and are the primary reason for people with asthma to use quick relief medication and
seek medical care, studies evaluating the link between Os exposures and such symptoms allow a
more direct characterization of the clinical and public health significance of ambient Os exposure
than measures of lung function decrements and pulmonary inflammation. Controlled human
exposure and toxicological studies have described modes of action through which short-term Os
exposures may increase respiratory symptoms by demonstrating Os-induced airway
hyperresponsiveness (U.S. EPA, 2013, section 6.2.2) and pulmonary inflammation (U.S. EPA,
2013, section 6.2.3).
The link between subjective respiratory symptoms and Os exposures has been evaluated
in both controlled human exposure and epidemiologic studies, and the link with medication use
has been evaluated in epidemiologic studies. In the last review, several controlled human
exposure studies reported respiratory symptoms following exposures to Os concentrations at or
above 80 ppb. In addition, one study reported such symptoms following exposures to 60 ppb Os,
though the increase was not statistically different from filtered air controls. Epidemiologic
studies reported associations between ambient Cb and respiratory symptoms and medication use
in a variety of locations and populations, including asthmatic children living in U.S. cities. In the
current review, additional controlled human exposure studies have evaluated respiratory
symptoms following exposures to Os concentrations below 80 ppb and recent epidemiologic
studies have evaluated associations with respiratory symptoms and medication use (U.S. EPA,
2013, sections 6.2.1, 6.2.4).
In controlled human exposure studies available in the last review as well as newly
available studies, statistically significant increases in respiratory symptoms have been
consistently reported in healthy adult volunteers engaged in intermittent, moderate exertion
following 6.6 hour exposures to average Os concentrations at or above 80 ppb (Adams, 2003;
Adams, 2006; Schelegle et al., 2009). Such symptoms have been reported to increase with
3-26
-------
increasing Os exposure concentrations, duration of exposure, and activity level (McDonnell et
al., 1999). For example, in a study available during the last review, Adams (2006) reported an
increase in respiratory symptoms in healthy adults during a 6.6 hour exposure protocol with an
average Os exposure concentration of 60 ppb. This increase was significantly different from
initial respiratory symptoms, but not from filtered air controls. Two recent controlled human
exposure studies that have become available since the last review did not report statistically
significant increases in respiratory symptoms following exposures of healthy adults to 60 ppb Os
(Schelegle et al., 2009; Kim et al., 2011). A recent study by Schelegle et al. (2009) did report a
statistically significant increase in respiratory symptoms in healthy adults following 6.6 hour
exposures to an average Os concentration of 70 ppb. The findings for Os-induced respiratory
symptoms in controlled human exposure studies, and the evidence integrated across disciplines
describing underlying modes of action, provide biological plausibility for epidemiologic
associations observed between short-term increases in ambient Os concentration and increases in
respiratory symptoms (U.S. EPA, 2013, section 6.2.4).
In epidemiologic studies of respiratory symptoms, data typically are collected by having
subjects (or their parents) record symptoms and medication use in a diary without direct
supervision by study staff. Several limitations of symptom reports are well recognized, as
described in the ISA (U.S. EPA, 2013, section 6.2.4). Nonetheless, symptom diaries remain a
convenient tool to collect individual-lev el data from a large number of subjects and allow
modeling of associations between daily changes in Os concentration and daily changes in
respiratory morbidity over multiple weeks or months. Importantly, many of the limitations in
these studies are sources of random measurement error that can bias effect estimates to the null
or increase the uncertainty around effect estimates (U.S. EPA, 2013, Section 6.2.4). Because
respiratory symptoms are associated with limitations in activity and daily function and are the
primary reason for using medication and seeking medical care, the evidence is directly coherent
with the associations consistently observed between increases in ambient Os concentration and
increases in asthma emergency department visits, discussed below (U.S. EPA, 2013, Section
6.2.4).
Most epidemiologic studies of Os and respiratory symptoms and medication use have
been conducted in children and/or adults with asthma, with fewer studies, and less consistent
results, in non-asthmatic populations (U.S. EPA, 2013, section 6.2.4). The 2006 AQCD (U.S.
EPA, 2006, U.S. EPA, 2013, section 6.2.4) concluded that the collective body of epidemiologic
evidence indicated that short-term increases in ambient Os concentrations are associated with
increases in respiratory symptoms in children with asthma. A large body of single-city and
single-region studies of asthmatic children provides consistent evidence for associations between
3-27
-------
short-term increases in ambient Os concentrations and increased respiratory symptoms and
asthma medication use in children with asthma (U.S. EPA, 2013, Figure 6-12, Table 6-20).
Methodological differences among studies make comparisons across recent multicity
studies of respiratory symptoms difficult. Because of fewer person-days of data (Schildcrout et
al., 2006) or examination of 19-day averages of ambient Os concentrations (O'Connor et al.,
2008), the ISA did not give greater weight to results from recent multicity studies than results
from single-city studies (U.S. EPA, 2013, section 6.2.4.5). While evidence from the few
available U.S. multicity studies is less consistent (O'Connor et al., 2008; Schildcrout et al., 2006;
Mortimer et al., 2002), the overall body of epidemiologic evidence with respect to the
association betweeen exposure to Os and respiratory symptoms in asthmatic children remains
compelling (U.S. EPA, 2013, section 6.2.4.1). Findings from a small body of studies indicate that
Os is also associated with increased respiratory symptoms in adults with asthma (Khatri et al.,
2009; Feo Brito et al., 2007; Ross et al., 2002) (U.S. EPA, 2013, section 6.2.4.2).
Available evidence indicates that Os-associated increases in respiratory symptoms are not
confounded by temperature, pollen, or copollutants (primarily PM) (U.S. EPA, 2013, section
6.2.4.5; Table 6-25; Romieu et al., 1996; Romieu et al., 1997; Thurston et al., 1997; Gent et al.,
2003). However, identifying the independent effects of Os in some studies was complicated due
to the high correlations observed between Os and PM or different lags and averaging times
examined for copollutants. Nonetheless, the ISA noted that the robustness of associations in
some studies of individuals with asthma, combined with findings from controlled human
exposure studies for the direct effects of Os exposure, provide substantial evidence supporting
the independent effects of short-term ambient Os exposure on respiratory symptoms (U.S. EPA,
2013, section 6.2.4.5).
Epidemiologic studies of medication use have reported associations with
1-hour maximum Os concentrations and with multiday average Os concentrations (Romieu et al.,
2006; Just et al., 2002). Some studies reported Os associations for both respiratory symptoms and
asthma medication use (Escamilla-Nufiez et al., 2008; Romieu et al., 2006; Schildcrout et al.,
2006; Jalaludin et al., 2004; Romieu et al., 1997; Thurston et al., 1997) while others reported
associations for either respiratory symptoms or medication use (Romieu et al., 1996; Rabinovitch
et al., 2004; Just et al., 2002; Ostro et al., 2001).
In summary, both controlled human exposure and epidemiologic studies have reported
respiratory symptoms attributable to short-term Os exposures. In the last review, the majority of
the evidence from controlled human exposure studies in young, healthy adults was for symptoms
following exposures to Os concentrations at or above 80 ppb. Although studies that have become
available since the last review have not reported respiratory symptoms in young, healthy adults
following exposures with moderate exertion to 60 ppb, one recent study has reported increased
3-28
-------
symptoms in young, healthy adults while at moderate exertion following exposures to Os
concentrations as low as 70 ppb.25 As was concluded in the 2006 Os AQCD (U.S. EPA, 2006,
1996), the collective body of epidemiologic evidence indicates that short-term increases in
ambient Os concentration are associated with increases in respiratory symptoms in children with
asthma (U.S. EPA, 2013, section 6.2.4). Recent studies of respiratory symptoms and medication
use, primarily in asthmatic children, add to this evidence. In a smaller body of studies, increases
in ambient Os concentration were associated with increases in respiratory symptoms in adults
with asthma.
Lung Host Defense
The mammalian respiratory tract has a number of closely integrated defense mechanisms
that, when functioning normally, provide protection from the potential health effects of
exposures to a wide variety of inhaled particles and microbes. These defense mechanisms
include mucociliary clearance, alveolobronchiolar transport mechanism, alveolar macrophages26,
and adaptive immunity27 (U.S. EPA, 2013, section 6.2.5). The previous Os AQCD (U.S. EPA,
2006) concluded that animal toxicological studies provided evidence that acute exposure to Cb
concentrations as low as 100 to 500 ppb can increase susceptibility to infectious diseases due to
modulation of these lung host defenses. This conclusion was based in large part on animal
studies of alveolar macrophage functioning and mucociliary clearance (U.S. EPA, 2013, section
6.2.5).
With regard to alveolar macrophage functioning, the previous Os AQCD (U.S. EPA,
2006) concluded that short periods of Os exposure can cause a reduction in the number of free
alveolar macrophages available for pulmonary defense, and that these alveolar macrophages are
more fragile, less able to engulf particles (i.e., phagocytic), and exhibit decreased lysosomal28
enzyme activities (U.S. EPA, 2013, section 6.2.5). These conclusions were based largely on
studies conducted in animals exposed for several hours up to several weeks to Os concentrations
from 100 to 250 ppb (Hurst et al., 1970; Driscoll et al., 1987; Cohen et al., 2002). Consistent
with the animal evidence, a controlled human exposure study available in the last review had
reported decrements in the ability of alveolar macrophages to phagocytize yeast following
exposures of healthy volunteers to Os concentrations of 80 and 100 ppb for 6.6-hour during
25As noted above, for the 70 ppb exposure concentration Schelegle et al. (2009) reported that the actual mean
exposure concentration was 72 ppb.
26 Phagocytic white blood cells within the alveoli of the lungs that ingest inhaled particles.
27 The adaptive immune system, is also known as the acquired immune system. Acquired immunity creates
immunological memory after an initial response to a specific pathogen, leading to an enhanced response to
subsequent encounters with that same pathogen.
28 Lysosomes are cellular organelles that contain acid hydrolase enzymes that break down waste materials and
cellular debris.
3-29
-------
moderate exercise (Devlin et al., 1991). Integrating the animal study results with human
exposure evidence available in the 1996 Criteria Document, the 2006 Criteria Document
concluded that available evidence indicates that short-term Os exposures have the potential to
impair host defenses in humans, primarily by interfering with alveolar macrophage function. Any
impairment in alveolar macrophage function may lead to decreased clearance of microorganisms
or nonviable particles. Compromised alveolar macrophage functions in asthmatics may increase
their susceptibility to other Os effects, the effects of particles, and respiratory infections (EPA,
2006, p. 8-26).
With regard to mucociliary clearance, in the last review a number of studies conducted in
different animal species had reported morphological damage to the cells of the tracheobronchial
tree from acute and sub-chronic exposure to Os concentrations at or above 200 ppb. The cilia
were either completely absent or had become noticeably shorter or blunt. In general, functional
studies of mucociliary transport had observed a delay in particle clearance soon after acute
exposure, with decreased clearance more evident at higher doses (1 ppm) (U.S. EPA, 2013,
section 6.2.5.1).
Alveolobronchiolar transport mechanisms refers to the transport of particles deposited in
the deep lung (alveoli) which may be removed either up through the respiratory tract (bronchi)
by alveolobronchiolar transport or through the lymphatic system. The pivotal mechanism of
alveolobronchiolar transport involves the movement of alveolar macrophages with ingested
particles to the bottom of the conducting airways. These airways are lined with ciliated epithelial
cells and cells that produce mucous, which surrounds the macrophages. The ciliated epithelial
cells move the mucous packets up the resiratory tract, hence the term "mucociliary escalator."
Although some studies show reduced tracheobronchial clearance after Os exposure, alveolar
clearance of deposited material is accelerated, presumably due to macrophage influx, which in
itself can be damaging.
With regard to adaptive immunity, a limited number of epidemiologic studies have
examined associations between Os exposure and hospital admissions or ED visits for respiratory
infection, pneumonia, or influenza. Results have been mixed, and in some cases conflicting (U.S.
EPA, 2013, Sections 6.2.7.2 and 6.2.7.3). With the exception of influenza, it is difficult to
ascertain whether cases of respiratory infection or pneumonia are of viral or bacterial etiology.
A recent study that examined the association between Os exposure and respiratory hospital
admissions in response to an increase in influenza intensity did observe an increase in respiratory
hospital admissions (Wong et al., 2009), but information from toxicological studies of Os and
viral infections is ambiguous.
In summary, relatively few studies conducted since the last review have evaluated the
effects of Os exposures on lung host defense. When the available evidence is taken as a whole,
3-30
-------
the ISA concludes that acute Os exposures impair the host defense capability of animals,
primarily by depressing alveolar macrophage function and perhaps also by decreasing
mucociliary clearance of inhaled particles and microorganisms. Coupled with limited evidence
from controlled human exposure studies, this suggests that humans exposed to Cb could be
predisposed to bacterial infections in the lower respiratory tract (EPA, 2013, section 6.2.5.5).
The seriousness of such infections may depend on how quickly bacteria develop virulence
factors and how rapidly PMNs are mobilized to compensate for the deficit in alveolar
macrophage function.
Allergic and Asthma-Related Responses
Effects resulting from combined exposures to Os and allergens have been studied in a
variety of animal species, generally as models of experimental asthma. Pulmonary function and
AHR in animal models of asthma are discussed in detail in Section 6.2.1.3 and Section 6.2.2.2,
respectively, in the ISA (U.S. EPA, 2013). Studies of allergic and asthma-related responses are
discussed in detail in sections 5.3.6 and 6.2.6 of the ISA (U.S. EPA, 2013).
Evidence available in the last review indicates that Os exposure skews immune responses
toward an allergic phenotype. For example, Gershwin et al. (1981) reported that Os (800 and 500
ppb for 4 days) exposure caused a 34-fold increase in the number of IgE (allergic antibody)-
containing cells in the lungs of mice. In general, the number of IgE-containing cells correlated
positively with levels of anaphylactic sensitivity. In humans, allergic rhinoconjunctivitis
symptoms are associated with increases in ambient Os concentrations (Riediker et al., 2001).
Controlled human exposure studies have observed Os-induced changes indicating allergic
skewing. Airway eosinophils, which are white blood cells that participate in allergic disease and
inflammation, were observed to increase in volunteers with atopy29 and mild asthma (Peden et
al., 1997). In a more recent study, expression of IL-5, a cytokine involved in eosinophil
recruitment and activation, was increased in subjects with atopy but not in healthy subjects
(Hernandez et al., 2010). Epidemiologic studies describe associations between eosinophils in
both short- (U.S. EPA, 2013, Section 6.2.3.2) and long-term (U.S. EPA, 2013, Section 7.2.5) Os
exposure, as do chronic exposure studies in non-human primates. Collectively, findings from
these studies suggest that Os can induce or enhance certain components of allergic inflammation
in individuals with allergy or allergic asthma.
Evidence available in the last review indicates that ozone may also increase AHR to
specific allergen triggers (75 FR 2970, January 19, 2010). Two studies (Torres et al., 1996; Holz
et al., 2002) observed increased airway responsiveness to Os exposure with bronchial allergen
29 Atopy is a predisposition toward developing certain allergic hypersensitivity reactions. A person with atopy
typically presents with one or more of the following: eczema (atopic dermatitis), allergic rhinitis (hay fever), allergic
conjunctivitis, or allergic asthma.
3-31
-------
challenge in subjects with preexisting allergic airway disease. Ozone-induced exacerbation of
airway responsiveness persists longer and attenuates more slowly than Os-induced lung function
decrements and respiratory symptom responses and can have important clinical implications for
asthmatics.
Animal toxicology studies indicate that Os enhances inflammatory and allergic responses
to allergen challenge in sensitized animals. In addition to exacerbating existing allergic
responses, toxicology studies indicate that Os can also act as an adjuvant to produce sensitization
in the respiratory tract. Along with its pro-allergic effects (inducing or enhancing certain
components of allergic inflammation in individuals with allergy or allergic asthma), Os could
also make airborne allergens more allergenic. When combined with NCh, Os has been shown to
enhance nitration of common protein allergens, which may increase their allergenicity Franze et
al. (2005).
Hospital Admissions and Emergency Department Visits
The 2006 Os AQCD evaluated numerous studies of respiratory-related emergency
department visits and hospital admissions. These were primarily time-series studies conducted in
the U.S., Canada, Europe, South America, Australia, and Asia. Based on such studies, the 2006
Os AQCD concluded that "the overall evidence supports a causal relationship between acute
ambient Os exposures and increased respiratory morbidity resulting in increased ED visits and
[hospital admissions] during the warm season30" (U.S. EPA, 2006). This conclusion was
"strongly supported by the human clinical, animal toxicologic[al], and epidemiologic evidence
for [Os-induced] lung function decrements, increased respiratory symptoms, airway
inflammation, and airway hyperreactivity" (U.S. EPA, 2006).
The results of recent studies largely support the conclusions of the 2006 Os AQCD (U.S.
EPA, 2013, section 6.2.7). Since the completion of the 2006 Os AQCD, relatively fewer studies
conducted in the U.S., Canada, and Europe have evaluated associations between short-term Os
concentrations and respiratory hospital admissions and emergency department visits, with a
growing number of studies conducted in Asia. This epidemiologic evidence is summarized in
Appendix 3A and discussed in detail in the ISA (U.S. EPA, 2013, section 6.2.7).
In considering this body of evidence, the ISA focused primarily on multicity studies
because they examine associations with respiratory-related hospital admissions and emergency
department visits over large geographic areas using consistent statistical methodologies (U.S.
EPA, 2013, section 6.2.7.1). The ISA also focused on single-city studies that encompassed a
large number of daily hospital admissions or emergency department visits, included long study -
30Epidemiologic associations for O3 are more robust during the warm season than during cooler months (e.g.,
smaller measurement error, less potential confounding by copollutants). Rationale for focusing on warm season
epidemiologic studies for Os can be found at 72 FR 37838-37840.
3-32
-------
durations, were conducted in locations not represented by the larger studies, or examined
population-specific characteristics that may increase the risk of Os-related health effects but were
not evaluated in the larger studies (U.S. EPA, 2013, section 6.2.7.1). When examining the
association between short-term Os exposure and respiratory health effects that require medical
attention, the ISA distinguishes between hospital admissions and emergency department visits
because it is likely that a small percentage of respiratory emergency department visits will be
admitted to the hospital; therefore, respiratory emergency department visits may represent
potentially less serious, but more common outcomes (U.S. EPA, 2013, section 6.2.7.1).
Several recent multicity studies (e.g., Cakmak et al., 2006; Dales et al., 2006) and a
multi-continent study (Katsouyanni et al., 2009) report associations between short-term Os
concentrations and increased respiratory-related hospital admissions and emergency department
visits. These multicity studies are supported by single-city studies also reporting consistent
positive associations using different exposure assignment approaches (i.e., average of multiple
monitors, single monitor, population-weighted average) and averaging times (i.e., 1-hour max
and 8-hour max) (U.S. EPA, 2013, sections 6.2.7.1 to 6.2.7.5). When examining cause-specific
respiratory outcomes, recent studies report positive associations with hospital admissions and
emergency department visits for asthma (Strickland et al., 2010; Stieb et al., 2009) and COPD
(Stieb et al., 2009; Medina-Ramon et al., 2006), with more limited evidence for pneumonia
(Medina-Ramon et al., 2006; Zanobetti and Schwartz, 2006). In seasonal analyses (Figure 3-2
below; U.S. EPA, 2013, Figure 6-19, Table 6-28), stronger associations were reported in the
warm season or summer months (red circles), when Os concentrations are higher, compared to
the cold season (blue circles), particularly for asthma (Strickland et al., 2010; Ito et al., 2007) and
COPD (Medina-Ramon et al., 2006).31 The available evidence indicates that children are at
greatest risk for Os-induced respiratory effects (Silverman and Ito, 2010; Strickland et al., 2010;
Mar and Koenig, 2009; Villeneuve et al., 2007; Dales et al., 2006).
Although the collective evidence across studies indicates a mostly consistent positive
association between Os exposure and respiratory-related hospital admissions and ED visits, the
magnitude of these associations may be underestimated due to behavioral modification in
response to air quality forecasts (U.S. EPA, 2013, Section 4.6.6).
31 The study by Strickland et al. (2010) is discussed in more detail in section 3.1.4.2, below.
3-33
-------
Study
Wongetal. (2009)
Cakmaketal. (2006)
Dales etal. (2006)
Orazzoet al. (2009)a
Katsouyanni et al. (2009)
Darrowet al. (2009)
.
Tplbert etal. (2007)
Biggerietal. (2005)c
Katsouyanni et al. (2009)
Stiebetal. (2009)
Villeneuveetal. (2007)
Strickland etal. (2010)
Silverman and Ito (2010)d
Itoetal. (2007
Villeneuveetal. (2007)
Mar and Koenig [2009
Strickland etal (2010)
Silverman and Ito (_2010)d
Mar and Koenig (2009)
Ito etal. (2007)
Villeneuveetal. (2007)
Strickland etal. (2010)
Wongetal. (2009)
Stiebetal. (2009)
Yang etal. (2006)
Medina-Ramon etal. (2006)
Stiebetal. (2009)e
Medina-Ramon etal. (2006)
Zanobettiand Schwartz (2006)
Medina-Ramon etal. (2006)
Location
Hong Kong
10 Canadian cities
11 Canadian cities
6 Italian cities
APHENA-Europe
APHENA-U.S.
APHENA-Canada
APHENA-Canada
Atlanta
Atlanta
8 Italian cities
APHENA-Europe
APHENA-U.S.
APHENA-Canada
APHENA-Canada
7 Canadian Cities
Alberta, CAN
Atlanta
New York
New York
Alberta, CAN
Seattle, WA
Atlanta
New York
Seattle, WA
New York
Alberta, CAN
Atlanta
Hong Kong
7 Canadian Cities
Vancouver
36 U.S. cities
7 Canadian Cities
36 U.S. cities
36 U.S. cities
Boston
36 U.S. cities
36 U.S. cities
36 U.S. cities
Visit Type
HA
HA
HA
ED
HA
HA
HA
HA
ED
ED
HA
HA
HA
HA
HA
ED
ED
ED
HA
ED
ED
ED
ED
HA
ED
ED
ED
ED
HA
ED
HA
HA
ED
HA
HA
HA
HA
HA
HA
Age
Lag
All
All
0-27 days
0-2
65+
65+
65+
65+
All
All
All
65+
65+
65+
65+
All
>2
Children
All
All
>2
18+
Children
6-18
<18
All
>2
Children
0-1
1.2
2
0-6
0-1
0-1
DLjO-2)
DL 0-2 b
1
0-2
0-3
0-1
0-1
DLjO-2)
DL(0-2)b
2
0-2
0-2
0-1
0-1
0-2
2
0-2
0-1
0
0-1
0-2
0-2
Respiratory
Asthma
All
All
65+
65+
All
65+
65+
65+
65+
65+
65+
0-1
0-3
DL(0-1)
NR
DL(O-l)
DL 0-1
0-1
COPD
Pneumonia
-25 -20 -15 -10 -50 5 10 15 20 25 30 35 40
% Increase
Note: Effect estimates are for a 20 ppb increase in 24-hour; 30 ppb increase in 8-hour max; and 40 ppb increase in 1-hour max O3 concentrations. HA=hospital admission;
ED=emergency department. Black=AII-year analysis; Red=Summer only analysis; Blue=Winter only analysis.
aWheeze used as indicator of lower respiratory disease.
b APHENA-Canada results standardized to approximate IQR of 5.1 ppb for 1-h max O3 concentrations.
0 Study included 8 cities; but of those 8, only 4 had O3 data.
dnon-ICU effect estimates.
eThe study did not specify the lag day of the summer season estimate.
Figure 3-2. Percent increase in respiratory-related hospital admission and emergency department visits in studies that
presented all-year and/or seasonal results.
3-34
-------
Studies examining the potential confounding effects of copollutants have reported that Os
effect estimates remained relatively robust upon the inclusion of PM and gaseous pollutants in
two-pollutant models (U.S. 2013, Figure 6-20, Table 6-29). Additional studies that conducted
copollutant analyses, but did not present quantitative results, also support these conclusions
(Strickland et al., 2010; Tolbert et al., 2007; Medina-Ramon et al., 2006) (U.S. 2013, section
6.2.7.5).
In the last review, studies had not evaluated the concentration-response relationship
between short-term Os exposure and respiratory-related hospital admissions and emergency
department visits. A preliminary examination of this relationship in studies that have become
available since the last review found no evidence of a deviation from linearity when examining
the association between short-term Os exposure and asthma hospital admissions (U.S. EPA,
2013, page 6-157, and Silverman and Ito, 2010). In addition, an examination of the
concentration-response relationship for Os exposure and pediatric asthma emergency department
visits found no evidence of a threshold at Os concentrations as low as 30 ppb (for daily
maximum 8-hour concentrations) (Strickland et al., 2010). However, in both studies there is
uncertainty in the shape of the concentration-response curve at the lower end of the distribution
of Os concentrations due to the low density of data in this range (U.S. 2013, page 6-157).
Respiratory Mortality
The controlled human exposure, epidemiologic, and toxicological studies discussed in
section 6.2 of the ISA (U.S. EPA, 2013, section 6.2) provide strong evidence for respiratory
morbidity effects, including ED visits and hospital admissions, in response to short-term Os
exposures. Moreover, evidence from experimental studies indicates multiple potential pathways
of respiratory effects from short-term Os exposures, which support the continuum of respiratory
effects that could potentially result in respiratory-related mortality in adults (U.S. EPA, 2013,
section 6.2.8). The 2006 Os AQCD found inconsistent evidence for associations between short-
term Os concentrations and respiratory mortality (U.S. EPA, 2006). Although some studies
reported a strong positive association between Os and respiratory mortality, additional studies
reported small associations or no associations. New epidemiologic evidence for respiratory
mortality is discussed in detail in section 6.2.8 of the ISA (U.S. EPA, 2013). The majority of
recent multicity studies have reported positive associations between short-term Os exposures and
respiratory mortality, particularly during the summer months (U.S. EPA, 2013, Figure 6-36).
Specifically, recent multicity studies from the U.S. (Zanobetti and Schwartz, 2008b),
Europe (Samoli et al., 2009), Italy (Stafoggia et al., 2010), and Asia (Wong et al., 2010), as well
as a multi-continent study (Katsouyanni et al., 2009), reported associations between short-term
Os concentrations and respiratory mortality (U.S. EPA, 2013, Figure 6-37, page 6-259). With
3-35
-------
respect to respiratory mortality, summer-only analyses were consistently positive and most were
statistically significant. In all-year analyses associations were positive, but smaller in magnitude.
Of the studies evaluated, only the studies by Katsouyanni et al. (2009) and by Stafoggia
et al. (2010) analyzed the potential for copollutant confounding of the Os-respiratory mortality
relationship. Based on the results of these analyses, the ISA concluded that Os respiratory
mortality risk estimates appear to be moderately to substantially sensitive (e.g., increased or
attenuated) to inclusion of PMio. However, in the APHENA study (Katsouyanni et al., 2009), the
mostly every-6th-day sampling schedule for PMio in the Canadian and U.S. datasets greatly
reduced their sample size and limits the interpretation of these results (U.S. EPA, 2013, section
6.2.8).
In summary, recent epidemiologic studies support and reinforce the epidemiologic
evidence for Os-associated respiratory hospital admissions and emergency department visits
from the last review. In addition, the evidence for associations with respiratory mortality has
been strengthened considerably since the last review, with the addition of several large multicity
studies. The biological plausibility of the associations reported in these studies is supported by
the experimental evidence for respiratory effects.
3.1.2.2 Respiratory Effects - Long-term Exposures
• To what extent does the currently available scientific evidence, including related
uncertainties, strengthen or alter our understanding from the last review of
respiratory effects attributable to long-term Os exposures?
As recognized in section 3.1.2.1, "the clearest evidence for health effects associated with
exposure to Os is provided by studies of respiratory effects" (U.S. EPA, 2013, section 1, p. 1-6).
Collectively, there is a vast amount of evidence spanning several decades that supports a causal
association between exposure to Os and a continuum of respiratory effects (U.S. EPA, 2013,
section 2.5). While the majority of this evidence is derived from studies investigating short-term
exposures, evidence from animal toxicological studies and recent epidemiologic evidence
indicate that long-term exposures (i.e., months to years) may also be detrimental to the
respiratory system. Across this evidence, particularly the epidemiologic evidence, the exposures
of focus vary, often involving repeated short concentrations extending over a long period, rather
than a continuous long-term exposure period.
In the 2006 Os AQCD, evidence was examined for relationships between long-term Os
exposure and effects on respiratory health outcomes including declines in lung function,
increases in inflammation, and development of asthma in children and adults. Animal toxicology
data provided a clearer picture indicating that long-term Os exposure may have lasting effects.
3-36
-------
Chronic32 exposure studies in animals have reported biochemical and morphological changes
suggestive of irreversible long-term Os impacts on the lung. In contrast to supportive evidence
from chronic animal studies, the epidemiologic studies on longer-term (annual) lung function
declines, inflammation, and new asthma development remained inconclusive.
Several epidemiologic studies collectively indicated that Os exposure averaged over
several summer months was associated with smaller increases in lung function growth in
children. For longer averaging periods (annual), the analysis in the Children's Health Study
(CHS) reported by Gauderman et al. (2004) provided little evidence that such long-term
exposure to ambient Os was associated with significant deficits in the growth rate of lung
function in children. Limited epidemiologic research examined the relationship between long-
term Os exposures and inflammation. Cross-sectional studies detected no associations between
long-term Os exposures and asthma prevalence, asthma-related symptoms or allergy to common
aeroallergens in children. However, longitudinal studies provided evidence that long-term Os
exposure influences the risk of asthma development in children and adults.
The currently available body of evidence supporting a relationship between long-term Os
exposures and adverse respiratory health effects that is likely to be causal is discussed in detail in
the ISA (EPA 2013, section 7.2). New evidence reports interactions between genetic variants and
long-term Os exposure affect the occurrence of new-onset asthma in multi-community, U.S.
cohort studies where protection by specific oxidant gene variants was restricted to children living
in low Os communities. A new line of evidence reports a positive concentration-response
relationship between first asthma hospitalization and long-term Os exposure. Related studies
report coherent relationships between asthma severity and control, and respiratory symptoms
among asthmatics and long-term Os exposure. There is also limited evidence for an association
between long-term exposure to ambient Os concentrations and respiratory mortality. These
studies are summarized briefly below for new-onset asthma and asthma prevalence, asthma
hospital admissions and other morbidity effects, pulmonary structure and function, and
respiratory mortality.
Currently available scientific evidence of the adverse health effects attributable to long-
term Os exposures, even considering related uncertainties, is much stronger than the body of
evidence available at the time of the 2008 review of the Cb standard. The 2006 Os AQCD (U.S.
EPA, 2006) concluded that epidemiologic studies provided no evidence of associations between
long-term (annual) Os exposures and asthma-related symptoms, asthma prevalence, or allergy to
common allergens after controlling for covariates. It found limited evidence for a relationship
32 Unless otherwise specified, the term "chronic" generally refers to an annual exposure duration for epidemiologic
studies and a duration of greater than 10% of the lifespan of the animal in lexicological studies.
3-37
-------
between long-term exposures to ambient Cb and deficits in the growth rate of lung-function in
children, pulmonary inflammation and other endpoints. Episodic exposures were also known to
cause more severe pulmonary morphological changes than continuous exposure.
The evidence base available in this review includes additional epidemiologic studies
using a variety of designs and analysis methods evaluating the relationship between long-term Os
exposures and measures of respiratory morbidity and mortality effects conducted by different
research groups in different locations. The ISA (U.S. EPA, 2013, p. 7-33), in Table 7-2 presents
selected key new longitudinal and cross-sectional studies of respiratory health effects and
associated Os concentrations. The positive results from various designs and locations support a
relationship between long-term exposure to ambient Os and respiratory health effects and
mortality.
In this review, the evidence of effects associated with long-term exposures strengthens
the relationship between Os exposure and health effects defined as adverse by the ATS, a
definition that has been used in previous reviews of the Os standard. As discussed in more detail
in section 3.1.3 below, the ATS (1985) defined adverse as "medically significant physiologic or
pathologic changes generally evidenced by one or more of the following: (1) interference with
the normal activity of the affected person or persons, (2) episodic respiratory illness, (3)
incapacitating illness, (4) permanent respiratory injury, and/or (5) progressive respiratory
dysfunction." As discussed below, in this review there is now credible evidence of respiratory
health effects associated with long-term Os exposures that would fall in to each of these five
categories that define adversity.
From a policy perspective, the recent epidemiologic studies from the CHS of long-term
Os exposures that shed light on the interaction between genetic variability, Cb exposures, and
health effects in children are important, not only because they help clarify previous findings, but
also because the effects evaluated, such as new-onset asthma, are clearly adverse. The ISA (U.S.
EPA, 2013, p. 7-12) notes that the collective evidence from CHS provides an important
demonstration of gene-environment interactions. It further notes that in the complex
gene-environment setting a modifying effect might not be reflected in an exposure main effect
and that the simultaneous occurrence of main effect and interaction effect can occur. Moreover,
the study of gene-environment interactions elucidates disease mechanisms in humans by using
information on susceptibility genes to focus on the biological pathways that are most relevant to
that disease.
In the CHS cohort of children in 12 Southern California communities, long-term
exposure to Cb concentrations was not associated with increased risk of developing asthma
(McConnell et al., 2010); however, greater outdoor exercise was associated with development of
asthma in children living in communities with higher ambient Cb concentrations (McConnell et
3-38
-------
al., 2002). Recent CHS studies examined interactions among genetic variants, long-term Os
exposure, and new onset asthma in children. These prospective cohort studies are
methodologically rigorous epidemiologic studies, and evidence indicates gene-Os interactions.
These studies have provided data supporting decreased risk of certain genetic variants on new
onset asthma (e.g., HMOX-1, ARG) that is limited to children either in low (Islam et al., 2008)
or high (Salam et al., 2009) Os communities. Gene-environment interaction also was
demonstrated with findings that greater outdoor exercise increased risk of asthma in GSTP1
lie/lie children living in high Os communities (Islam et al., 2009). Biological plausibility for
these gene-Os environment interactions is provided by evidence that these enzymes have
antioxidant and/or anti-inflammatory activity and participate in well recognized modes of action
in asthma pathogenesis. As Os is a source of oxidants in the airways, oxidative stress serves as
the link among Os exposure, enzyme activity, and asthma. Cross-sectional studies by Akinbami
et al. (2010) and Hwang et al. (2005) provide further evidence relating Os exposures with asthma
prevalence.
Studies using a cross-sectional design provide support for a relationship between long-
term Os exposure and adverse health effects in asthmatics, including: bronchitic symptoms
(related to TNF-308 genotype in asthmatic children) (Lee et al., 2009); asthma severity (Rage et
al., 2009) and asthma control (Jacquemin et al., in press) in an adult cohort; respiratory-related
school absences (related to CAT and MPO variant genes) (Wenten et al., 2009); asthma ED
visits in adults (Meng et al., 2010); and, asthma hospital admissions in adults and children (Lin et
al., 2008b; Meng et al., 2010; Moore et al., 2008). Several studies, shown in Table 7-3 (ISA, U.S.
EPA, 2013, p. 7-35), provide results adjusted for potential confounders presenting results for
both Os and PM (in single and multipollutant models) as well as other pollutants where PM
effects were not provided. As shown in this table, Os associations were generally robust to
adjustment by potential confounding by PM.
Information from toxicological studies in nonhuman primates indicates that long term
exposure to Os during gestation or development can result in irreversible morphological changes
in the lung, which in turn can influence the function of the respiratory tract. This nonhuman
primate evidence of an Os-induced change in airway responsiveness supports the biologic
plausibility of long term exposure to Os contributing to effects of asthma in children. However,
results from epidemiologic studies examining long-term Os exposure and pulmonary function
effects are inconclusive with some new studies relating effects at higher exposure levels.
The ISA (U.S. EPA, 2013, p. 7-31) concludes that there is limited evidence for an
association between long-term exposure to ambient Os concentrations and respiratory mortality
in adults (Jerrett et al., 2009). This effect was robust to the inclusion of PM2.5 and insensitive to a
number of different model specifications. Moreover, there is evidence that long-term exposure to
3-39
-------
Os is associated with mortality among individuals that had previously experienced an emergency
hospital admission due to COPD (Zanobetti and Schwartz, 2011).
In conclusion, since the last review, the body of evidence about the effects of long-term
Os exposure has been considerably strengthened. The scientific evidence available for this
review, including related uncertainties, provides an overall strong body of evidence of adverse
health effects attributable to long-term Os exposures. These include a coherent range of asthma
morbidity effects such as new-onset asthma, asthma prevalence, symptoms, school absences, ED
visits and hospital admissions. There is also new evidence of respiratory mortality associated
with long-term Os exposure. Further discussion of key studies is below.
New-onset Asthma and Asthma Prevalence
Asthma is a heterogeneous disease with a high degree of temporal variability. The on-set,
progression, and symptoms can vary within an individual's lifetime, and the course of asthma
may vary markedly in young children, older children, adolescents, and adults. In the previous
review, longitudinal cohort studies that examined associations between long-term Os exposures
and the onset of asthma in adults and children indicated a direct effect of long-term Os exposures
on asthma risk in adults (McDonnell et al., 1999, 15-year follow-up; Greer et al., 1993, 10-year
follow-up) and effect modification by Os in children (McConnell et al., 2002). Since that review,
new evidence has become available about the association between long-term exposures to Os and
new-onset asthma that has increased our understanding of the gene-environment interaction and
the mechanisms and biological pathways most relevant to assessing Os-related effects.
In children, the relationship between long-term Os exposure and new-onset asthma has
been extensively studied in the CHS; a long-term study that was initiated in the early 1990's
which has evaluated effects in several cohorts of children. The CHS was initially designed to
examine whether long-term exposure to ambient pollution was related to chronic respiratory
outcomes in children in 12 communities in southern California. In the CHS, new-onset asthma
was classified as having no prior history of asthma at study entry with subsequent report of
physician-diagnosed asthma at follow-up, with the date of onset assigned to be the midpoint of
the interval between the interview date when asthma diagnosis was first reported and the
previous interview date. The results of one study (McConnell et al., 2002) available in the
previous review indicated that within high Os communities, asthma risk was 3.3 times greater for
children who played three or more outdoor sports as compared with children who played no
sports.
For this review, as discussed in section 7.2.1.1 of the ISA (U.S. EPA, 2013), recent
studies from the CHS provide evidence for gene-environment interactions in effects on new-
onset asthma by indicating that the lower risks associated with specific genetic variants are found
in children who live in lower Os communities. These studies indicate that the risk for new-onset
3-40
-------
asthma is related in part to genetic susceptibility, as well as behavioral factors and environmental
exposure. The onset of a chronic disease, such as asthma, is partially the result of a sequence of
biochemical reactions involving exposures to various environmental agents metabolized by
enzymes related to a number of different genes. Oxidative stress has been proposed to underlie
the mechanistic hypotheses related to Os exposure. Genetic variants may impact disease risk
directly, or modify disease risk by affecting internal dose of pollutants and other environmental
agents and/or their reaction products, or by altering cellular and molecular modes of action.
Understanding the relation between genetic polymorphisms and environmental exposure can
help identify high-risk subgroups in the population and provide better insight into pathway
mechanisms for these complex diseases.
The CHS analyses (Islam et al., 2008; Islam et al. 2009; Salam et al., 2009) have found
that asthma risk is related to interactions between Os and variants in genes for enzymes such as
heme-oxygenase (HO-1), arginases (ARG1 and 2), and glutathione S transferase PI (GSTP1).
Biological plausibility for these findings is provided by evidence that these enzymes have
antioxidant and/or anti-inflammatory activity and participate in well-recognized modes of action
in asthma pathogenesis. Further, several lines of evidence demonstrate that secondary oxidation
products of Os initiate the key modes of action that mediate downstream health effects (ISA,
Section 5.3, U.S. EPA, 2013). For example, HO-1 responds rapidly to oxidants, has anti-
inflammatory and anti-oxidant effects, relaxes airway smooth muscle, and is induced in the
airways during asthma. Gene-environment interactions are discussed in detail in Section 5.4.2.1
in the ISA (U.S. EPA, 2013).
Asthma Hospital Admissions
In the 2006 AQCD, studies on Os-related hospital discharges and emergency department
(ED) visits for asthma and respiratory disease mainly looked at short-term (daily) metrics. The
short-term Os studies presented in section 6.2.7.5 of the ISA (U.S. EPA, 2013) and discussed
above in section 3.1.2.1 continue to indicate that there is evidence for increases in both hospital
admissions and ED visits in children and adults related to all respiratory outcomes, including
asthma, with stronger associations in the warm months. New studies, discussed in section 7.2.2
of the ISA (U.S. EPA, 2013) also evaluated long-term Os exposure metrics, providing a new line
of evidence that suggests a positive exposure-response relationship between the first hospital
admission for asthma and long-term Os exposure, although the ISA cautions in attributing the
associations in that study to long-term exposures since there is potential for short-term exposures
to contribute to the observed associations.
Evidence associating long-term Os exposure to first asthma hospital admission in a
positive concentration-response relationship is provided in a retrospective cohort study (Lin et
al., 2008b). This study investigated the association between chronic exposure to Os and
3-41
-------
childhood asthma admissions by following a birth cohort of more than 1.2 million babies born in
New York State (1995-1999) to first asthma admission or until 31 December 2000. Three annual
indicators (all 8-hour maximum from 10:00 a.m. to 6:00 p.m.) were used to define chronic Os
exposure: (1) mean concentration during the follow-up period (41.06 ppb); (2) mean
concentration during the Os season (50.62 ppb); and (3) proportion of follow-up days with Os
levels >70 ppb. The effects of co-pollutants were controlled, and interaction terms were used to
assess potential effect modifications. A positive association between chronic exposure to Os and
childhood asthma hospital admissions was observed, indicating that children exposed to high Os
levels over time are more likely to develop asthma severe enough to be admitted to the hospital.
The various factors were examined and differences were found for younger children (1-2 years),
poor neighborhoods, Medicald/self-paid births, geographic region and others. As shown in the
ISA, Figure 7-3 (U.S. EPA, 2013, p. 7-16), positive concentration-response relationships were
observed. Asthma admissions were significantly associated with increased Os levels for all
chronic exposure indicators.
In considering the relationship between long-term pollutant exposures and chronic
disease heath endpoints, where chronic pathologies are found with acute expression of chronic
disease, Kiinzli (2012) hypothesizes that if the associations of pollution with events are much
larger in the long-term studies, it provides some indirect evidence that air pollution increases the
pool of subjects with chronic disease, and that more acute events are to be expected to be seen
for higher exposures. The results of Lin et al. (2008b) for first asthma hospital admission,
presented in Figure 7-3 (U.S. EPA, 2013, p. 7-16), show effects estimates that are larger than
those reported in a study of childhood asthma hospital admission in New York state (Silverman
and Ito, 2010), discussed in section 3.1.2.1 and 3.1.2.2 above. The ISA (U.S. EPA, 2013, p. 7-16)
notes that this provides some support for the hypothesis that Os exposure may not only have
triggered the events but also increased the pool of asthmatic children, but cautions in attributing
the associations in Lin et al. (2008b) study to long-term exposures since there is potential for
short-term exposures to contribute to the observed associations.
Pulmonary structure and function
In the 2006 Os AQCD, few epidemiologic studies had investigated the effect of chronic
Os exposure on pulmonary function. The definitive 8-year follow-up analysis of the first cohort
of the CHS (U.S. EPA, 2013, section 7.2.3.1) provided little evidence that long-term exposure to
ambient Os was associated with significant deficits in the growth rate of lung function in
children. The strongest evidence was for medium-term effects of extended Os exposures over
several summer months on lung function (FEVi) in children, i.e., reduced lung function growth
being associated with higher ambient Os levels. Short-term Os exposure studies presented in ISA
(U.S. EPA, 2013, Section 6.2.1.2), and above in section 3.1.2.1, provide a cumulative body of
3-42
-------
epidemiologic evidence that strongly supports associations between ambient Os exposure and
decrements in lung function among children. A later CHS study (Islam et al., 2007) included in
this review (U.S. EPA, 2013, section 7.2.3.1) also reported no substantial differences in the
effect of Os on lung function. However, in a more recent CHS study, Breton et al. (2011)
hypothesized that genetic variation in genes on the glutathione metabolic pathway may influence
the association between ambient air pollutant exposures and lung function growth in children,
and found that variation in the GSS locus was associated with differences in risk of children for
lung function growth deficits associated ambient air pollutants, including Os. A recent study
(Rojas-Martinez et al., 2007) of long-term exposure to Os, described in section 7.2.3.1 of the ISA
(U.S. EPA, 2013, p. 7-19), observed a relationship with pulmonary function declines in school-
aged children where Os and other pollutant levels were higher (90 ppb at high end of the range)
than those in the CHS. Two studies of adult cohorts provide mixed results where long-term
exposures were at the high end of the range.
Long-term studies in animals allow for greater insight into the potential effects of
prolonged exposure to Os that may not be easily measured in humans, such as structural changes
in the respiratory tract. Despite uncertainties, epidemiologic studies observing associations of Os
exposure with functional changes in humans can attain biological plausibility in conjunction with
long-term toxicological studies, particularly Os-inhalation studies performed in non-human
primates whose respiratory systems most closely resembles that of the human. An important
series of studies, discussed in section 7.2.3.2 of the ISA (U.S. EPA, 2013), have used nonhuman
primates to examine the effect of Os alone, or in combination with an inhaled allergen, house
dust mite antigen (HDMA), on morphology and lung function. These animals exhibit the
hallmarks of allergic asthma defined for humans, including: a positive skin test for HDMA with
elevated levels of IgE in serum and IgE-positive cells within the tracheobronchial airway walls;
impaired airflow which is reversible by treatment with aerosolized albuterol; increased
abundance of immune cells, especially eosinophils, in airway exudates and bronchial lavage; and
development of nonspecific airway responsiveness (NHLBI, 2007). These studies and others
have demonstrated changes in pulmonary function and airway morphology in adult and infant
nonhuman primates repeatedly exposed to environmentally relevant concentrations of Os (ISA,
U.S. EPA, 2013, section 7.2.3.2).
The initial observations in adult nonhuman primates have been expanded in a series of
experiments using infant rhesus monkeys repeatedly exposed to 0.5 ppm Os starting at 1 month
of age (Plopper et al., 2007). The purpose of these studies was to determine if a cyclic regimen of
Os inhalation would amplify the allergic responses and structural remodeling associated with
allergic sensitization and inhalation in the infant rhesus monkey. After several episodic
exposures of infant monkeys to Os, they observed a significant increase in the baseline airway
3-43
-------
resistance, which was accompanied by a small increase in airway responsiveness to inhaled
histamine (Schelegle et al., 2003), although neither measurement was statistically different from
filtered air control values. Exposure of animals to inhaled house dust mite antigen alone also
produced small but not statistically significant changes in baseline airway resistance and airway
responsiveness, whereas the combined exposure to both (Os + antigen) produced statistically
significant and greater than additive changes in both functional measurements. This nonhuman
primate evidence of an Os-induced change in airway resistance and responsiveness provides
biological plausibility of long-term exposure, or repeated short-term exposures, to Os
contributing to the effects of asthma in children.
To understand which conducting airways and inflammatory mechanisms are involved in
Os-induced airway hyperresponsiveness in the infant rhesus monkey, results of a follow-up study
(load et al., 2006) suggest that effect of Os on airway responsiveness occurs predominantly in
the smaller bronchioles, where dosimetric models indicate the dose would be higher.
The functional changes in the conducting airways were accompanied by a number of cellular and
morphological changes, including a significant 4-fold increase in eosinophils. Thus, these studies
demonstrate both functional and cellular changes in the lung of infant monkeys after cyclic
exposure to 0.5 ppm Os, providing relevant information to understanding the potentially
damaging effects of ambient Os exposure on the respiratory tract of children.
In addition, noteworthy structural changes in the respiratory tract development, during
which conducting airways increase in diameter and length, have been observed in infant rhesus
monkeys after cyclic exposure to Os (Fanucchi et al., 2006). Observed changes included more
proximal first alveolar outpocketing, decreases in the diameter and length of the terminal and
respiratory bronchioles, increases in mucus-producing goblet cell mass, alterations in smooth
muscle orientation in the respiratory bronchioles, epithelial nerve fiber distribution, and
basement membrane zone morphometry. The latter effects are important because of their
potential contribution to airway obstruction and airway hyperresponsiveness which are central
features of asthma. A number of studies in both non-human primates and rodents demonstrate
that Os exposure can increase collagen synthesis and deposition, including fibrotic-like changes
in the lung (ISA, U.S. EPA, 2013, section 7.2.3.2,).
Collectively, evidence from animal studies strongly suggests that chronic Os exposure is
capable of damaging the distal airways and proximal alveoli, resulting in lung tissue remodeling
and leading to apparent irreversible changes. Potentially, persistent inflammation and interstitial
remodeling play an important role in the progression and development of chronic lung disease.
Further discussion of the modes of action that lead to Os-induced morphological changes can be
found in Section 5.3.7 of the ISA (U.S. EPA, 2013). Discussion of mechanisms involved in
lifestage susceptibility and developmental effects can be found in Section 5.4.2.4 of the ISA
3-44
-------
(U.S. EPA, 2013). The findings reported in chronic animal studies offer insight into potential
biological mechanisms for the suggested association between seasonal Os exposure and reduced
lung function development in children as observed in epidemiologic studies (see Section 7.2.3.1).
Respiratory Mortality
A limited number of epidemiologic studies have assessed the relationship between long-
term exposure to Os and mortality in adults. The 2006 Os AQCD concluded that an insufficient
amount of evidence existed "to suggest a causal relationship between chronic Os exposure and
increased risk for mortality in humans" (U.S. EPA, 2006). Though total and cardio-pulmonary
mortality were considered in these studies, respiratory mortality was not specifically considered.
In the most recent follow-up analysis of the ACS cohort (Jerrett et al., 2009), cardiopulmonary
deaths were separately subdivided into respiratory and cardiovascular deaths, rather than
combined as in the Pope et al. (2002) work. Increased Os exposure was associated with the risk
of death from respiratory causes, and this effect was robust to the inclusion of PM2.5. The
association between increased Os concentrations and increased risk of death from respiratory
causes was insensitive to the use of different models and to adjustment for several ecologic
variables considered individually. Additionally, a recent multi-city time series study (Zanobetti
and Schwartz, 2011), which followed (from 1985 to 2006) four cohorts of Medicare enrollees
with chronic conditions that might predispose to Os-related effects, observed an association
between long-term (warm season) exposure to Os and elevated risk of mortality in the cohort that
had previously experienced an emergency hospital admission due to COPD. A key limitation of
this study is the inability to control for PM2.5, because data were not available in these cities until
1999.
3.1.2.3 Total Mortality - Short-term Exposures
• To what extent does the currently available scientific evidence, including related
uncertainties, strengthen or alter our understanding from the last review of
mortality attributable to short-term Os exposures?
The 2006 Os AQCD concluded that the overall body of evidence was highly suggestive
that short-term exposure to Os directly or indirectly contributes to nonaccidental and
cardiopulmonary-related mortality in adults, but additional research was needed to more fully
establish underlying mechanisms by which such effects occur (U.S. EPA, 2006; U.S. EPA, 2013,
p. 2-18). In building on the 2006 evidence, the ISA states the following (U.S. EPA, 2013, p. 6-
261).
The evaluation of new multicity studies that examined the association between
short-term Os exposures and mortality found evidence that supports the
conclusions of the 2006 AQCD. These new studies reported consistent positive
associations between short-term Os exposure and all-cause (non-accidental)
3-45
-------
mortality, with associations persisting or increasing in magnitude during the
warm season, and provide additional support for associations between Oi
exposure and cardiovascular and respiratory mortality
The 2006 Os AQCD reviewed a large number of time-series studies of associations
between short-term Os exposures and total mortality including single- and multicity studies, and
meta-analyses. In the large U.S. multicity studies that examined all-year data, effect estimates
corresponding to single-day lags ranged from a 0.5-1% increase in all-cause (nonaccidental) total
mortality per a 20 ppb (24-hour), 30 ppb (8-hour maximum), or 40 ppb (1-hour maximum)
increase in ambient Os (U.S. EPA, 2013, section 6.6.2). Available studies reported some
evidence for heterogeneity in Os mortality risk estimates across cities and across studies. Studies
that conducted seasonal analyses reported larger Os mortality risk estimates during the warm
season. Overall, the 2006 Os AQCD identified robust associations between various measures of
daily ambient Os concentrations and all-cause mortality, which could not be readily explained by
confounding due to time, weather, or copollutants. With regard to cause-specific mortality,
consistent positive associations were reported between short-term Os exposure and
cardiovascular mortality, with less consistent evidence for associations with respiratory
mortality. The majority of the evidence for associations between Os and cause-specific mortality
were from single-city studies, which had small daily mortality counts and subsequently limited
statistical power to detect associations. The 2006 Os AQCD concluded that "the overall body of
evidence is highly suggestive that Cb directly or indirectly contributes to non-accidental and
cardiopulmonary-related mortality" (U.S. EPA, 2013, section 6.6.1).
Recent studies have strengthened the body of evidence that supports the association
between short-term Os concentrations and mortality in adults. This evidence includes a number
of studies reporting associations with non-accidental as well as cause-specific mortality. Multi-
continent and multicity studies have consistently reported positive and statistically significant
associations between short-term Os concentrations and all-cause mortality, with evidence for
larger mortality risk estimates during the warm or summer months (Figure 3-3 below, reprinted
from the ISA) (U.S. EPA, 2013, Figure 6-27; Table 6-42). Similarly, evaluations of cause-
specific mortality have reported consistently positive associations with Os, particularly in
analyses restricted to the warm season (U.S. EPA, 2013, Figure 6-37; Table 6-53).33
! Respiratory mortality is discussed in more detail above.
3-46
-------
Study
Gryparisetal.{2004;57276)
Bell etal. (2007; 93256)
Schwartz (2005; 57333)
Bell and Dominici(2008; 193828)
Bell etal. (2004; 94417)a
Levy etal. (2005; 74 34 7)a
Katsouyanni et al. (2009; 199899)
Bell etal. (2005; 74345)a
ltd etal. (2005; 74346)a
Wongetal. (2010; 732535)
Katsouyanni et al. (2009; 199899)
Cakmak etal. (2011, 699135)
Katsouyanni et al. (2009; 199899)
Katsouyanni et al. (2009; 199899Jb
Samoli etal. (2009; 195855)
Bell etal. (2004; 94417)a
Schwartz (2005; 57333)
Zanobetti and Schwartz(2008; 195755)
Zanobetti and Schwartz(2008; 101596)
Franklin and Schwartz (2008; 156448)
Gryparisetal.(2004;57276)
Medina-Ramon and Schwartz (2008)
Katsouyanni et al. (2009; 199899)
Bell etal. (2005; 74345)a
Katsouyanni et al. (2009; 199899)
Katsouyanni et al. (2009; 199899Jb
Levy etal. (2005; 74 34 7)a
Ito etal. (2005; 74346)a
Katsouyanni et al. (2009; 199899)
Stafoggia et al. (2010; 625034)
Location
APHEA2 (23 cities)
98 U.S. communities
14 U.S. cities
98 U.S. communjtjes
95 U.S. communities
U.S. and Non-U.S.
APHENA-Europe
U.S. and Non-U.S.
U.S. and Non-U.S.
PAPA (4 cities
APHENA-U.S.
7 Chilean cities
APHENA-Canada
APHENA-Canada
21 European cities
95 U.S. communities
14 U.S. cities
48 U.S. cities
48 U.S. cities
18 U.S. communities
APHEA2 (21 cities)
48 U.S. cities
APHENA-Europe
U.S. and Non-U.S.
APHENA-Canada
APHENA-Canada
U.S. and Non-U.S.
U.S. and Non-U.S.
APHENA-U.S.
10 Italian cities
Lag
0-1
0-1
0
0-6
0-6
DL(0-2)
0-1
DL(0-2)
All-Year
DL
DL
DL
0-6'
0-2
0-2
0-1
0-6
0
0
0-3
0
0-1
DL(0-2)
DL(0-2
DL(0-2
DL(0-2
DL 0-5
Summer
-1
5 7
% Increase
11
Figure 3-3. Summary of mortality risk estimates for short-term Os and all-cause (nonaccidental) mortality.
34
34
Reprinted from the ISA (U.S. EPA, 2013, Figure 6-27).
3-47
-------
In assessing the evidence for Os-related mortality, the 2006 AQCD also noted that
multiple uncertainties remained regarding the relationship between short-term Os concentrations
and mortality, including the extent of residual confounding by co-pollutants; characterization of
the factors that modify the Os-mortality association; the appropriate lag structure for identifying
Os-mortality effects; and the shape of the Os-mortality concentration-response function and
whether a threshold exists. Many of the studies, published since the last review, have attempted
to address one or more of these uncertainties. The ISA (U.S. EPA, 2013, Section 6.6.2) discusses
the extent to which recent studies have evaluated these uncertainties in the relationship between
Os and mortality.
In particular, recent studies have evaluated different statistical approaches to examine the
shape of the Os-mortality concentration-response relationship and to evaluate whether a
threshold exists for Os-related mortality. In an analysis of the NMMAPS data, Bell et al. (2006)
evaluated the potential for a threshold in the Os-mortality relationship. The authors reported
positive and statistically significant associations with mortality in a variety of subset analyses,
including analyses restricted to days with 24-hour area-wide average Os concentrations below
60, 55, 50, 45, 40, 35, and 30 ppb. In these restricted analyses Os effect estimates were of similar
magnitude, were statistically significant, and had similar statistical precision. In analyses
restricted to days with 24-hour average Os concentrations below 25 ppb, the Os effect estimate
was similar in magnitude to the effect estimates resulting from analyses with the higher cutoffs,
but had somewhat lower statistical precision, with the estimate approaching statistical
significance (i.e., based on observation of Figure 2 in Bell et al., 2006). In analyses restricted to
days with lower 24-hour average Os concentrations, effect estimates were not statistically
significant (i.e., based on observation of Figure 2 in Bell et al., 2006).
Bell et al. (2006) also evaluated the shape of the concentration-response relationship
between Os and mortality. Although the results of this analysis suggested the lack of threshold in
the Os-mortality relationship, the ISA noted that it is difficult to interpret such a curve because:
(1) there is uncertainty around the shape of the concentration-response curve at 24-hour average
Os concentrations generally below 20 ppb and (2) the concentration-response curve does not take
into consideration the heterogeneity in Os-mortality risk estimates across cities (U.S. EPA, 2013,
section 6.6.2.3).
Several additional studies have used the NMMAPS dataset to evaluate the concentration-
response relationship between short-term Os concentrations and mortality. For example, using
the same data as Bell et al. (2006), Smith et al. (2009) conducted a subset analysis, but instead of
restricting the analysis to days with Os concentrations below a cutoff the authors only included
days above a defined cutoff. The results of this analysis were consistent with those reported by
Bell et al. (2006). Specifically, the authors reported consistent positive associations for all cutoff
3-48
-------
concentrations up to concentrations where the total number of days available were so limited that
the variability around the central estimate was increased (U.S. EPA, 2013, section 6.6.2.3). In
addition, using NMMAPS data for 1987-1994 for Chicago, Pittsburgh, and El Paso, Xia and
long (2006) reported evidence for a threshold around a 24-hour average Os concentration of
25 ppb, though the threshold values estimated in the analysis were sometimes in the range of
where data density was low (U.S. EPA, 2013, section 6.6.2.3). Stylianou and Nicolich (2009)
examined the existence of thresholds following an approach similar to Xia and Tong (2006)
using data from NMMAPS for nine major U.S. cities (i.e., Baltimore, Chicago, Dallas/Fort
Worth, Los Angeles, Miami, New York, Philadelphia, Pittsburgh, and Seattle) for the years
1987-2000. The authors reported that the estimated Os-mortality risks varied across the nine
cities, with the models exhibiting apparent thresholds in the 10-45 ppb range for Os (24-hour
average). Additional studies in Europe, Canada, and Asia did not report evidence for a threshold
(Katsouyanni et al., 2009).
3.1.2.4 Cardiovascular effects - Short-term Exposure
• To what extent does the currently available scientific evidence, including related
uncertainties, strengthen or alter our understanding from the last review of
cardiovascular effects attributable to short-term Os exposures?
A relatively small number of studies have examined the potential effect of short-term Os
exposure on the cardiovascular system. The 2006 Os AQCD (U.S. EPA, 2006, p. 8-77)
concluded that "Os directly and/or indirectly contributes to cardiovascular-related morbidity" but
added that the body of evidence was limited. This conclusion was based on a controlled human
exposure study that included hypertensive adult males; a few epidemiologic studies of
physiologic effects, heart rate variability, arrhythmias, myocardial infarctions, and hospital
admissions; and toxicological studies of heart rate, heart rhythm, and blood pressure.
More recently, the body of scientific evidence available that has examined the effect of
Os on the cardiovascular system has expanded. There is an emerging body of animal
toxicological evidence demonstrating that short-term exposure to Os can lead to autonomic
nervous system alterations (in heart rate and/or heart rate variability) and suggesting that
proinflammatory signals may mediate cardiovascular effects. Interactions of Cb with respiratory
tract components result in secondary oxidation product formation and subsequent production of
inflammatory mediators, which have the potential to penetrate the epithelial barrier and to initiate
toxic effects systemically. In addition, animal toxicological studies of long-term exposure to Os
provide evidence of enhanced atherosclerosis and ischemia/reperfusion (I/R) injury,
corresponding with development of a systemic oxidative, proinflammatory environment. Recent
experimental and epidemiologic studies have investigated Os-related cardiovascular events and
3-49
-------
are summarized in Section 6.3 of the ISA (U.S. EPA, 2013, Section 6.3). Overall, the ISA
summarized the evidence in this review as follows (U.S. EPA, 2013, p. 6-211).
In conclusion, animal toxicological studies demonstrate Os-induced
cardiovascular effects, and support the strong body of epidemiologic evidence
indicating Os-induced cardiovascular mortality. Animal toxicological and
controlled human exposure studies provide evidence for biologically plausible
mechanisms underlying these Os-induced cardiovascular effects. However, a lack
of coherence with epidemiologic studies of cardiovascular morbidity remains an
important uncertainty.
Animal toxicological studies support that short-term Os exposure can lead to
cardiovascular morbidity. Animal studies provide evidence for both increased and decreased
heart rate (HR), however it is uncertain if Os-induced reductions in HR are relevant to humans.
Animal studies also provide evidence for increased heart rate variability (HRV), arrhythmias,
vascular disease and injury following short-term Os exposure. In addition, a series of studies
highlight the role of genetic variability and age in the induction of effects and attenuation of
responses to Os exposure.
Biologically plausible mechanisms have been described for the cardiovascular effects
observed in animal exposure studies (U.S. EPA, 2013, Section 5.3.8). Evidence that
parasympathetic pathways may underlie cardiac effects is described in more detail in Section
5.3.2 of the ISA (U.S. EPA, 2013). Recent studies suggest that Os exposure may disrupt the
endothelin system that constricts blood vessels and increase blood pressure, which can result in
an increase in HR, HRV; and disrupt the NO'system and the production of atrial natriuretic
factor (ANF), vasodilators that reduce blood pressure. Additionally, Os may increase oxidative
stress and vascular inflammation promoting the progression of atherosclerosis and leading to
increased susceptibility to I/R injury. As Os reacts quickly with the ELF and does not translocate
to the heart and large vessels, studies suggest that the cardiovascular effects exhibited could be
caused by secondary oxidation products resulting from Os exposure. However, direct evidence of
translocation of Os reaction products to the cardiovascular system has not been demonstrated in
vivo. Alternatively, extrapulmonary release of diffusible mediators (such as cytokines or
endothelins) may initiate or propagate inflammatory responses throughout the body leading to
the cardiovascular effects reported in toxicology studies. Ozone reacts within the lung to induce
pulmonary inflammation and the influx and activation of inflammatory cells, resulting in a
cascading proinflammatory state, and may lead to the extrapulmonary release of diffusible
mediators that could result in cardiovascular injury.
Controlled human exposures studies discussed in previous AQCDs have not
demonstrated any consistent extrapulmonary effects. In this review, evidence from controlled
human exposure studies suggests cardiovascular effects in response to short-term Os exposure
3-50
-------
(see ISA, U.S. EPA, 2013, Section 6.3.1) and provides some coherence with evidence from
animal toxicology studies. Controlled human exposure studies also support the animal
toxicological studies by demonstrating Os-induced effects on blood biomarkers of systemic
inflammation and oxidative stress, as well as changes in biomarkers that can indicate a
prothrombogenic response to Os. Increases and decreases in high frequency HRV have been
reported following relatively low (120 ppb during rest) and high (300 ppb with exercise) Os
exposures, respectively. These changes in cardiac function observed in animal and human studies
provide preliminary evidence for Os-induced modulation of the autonomic nervous system
through the activation of neural reflexes in the lung (see ISA, U.S. EPA, 2013, Section 5.3.2).
Overall, the ISA concludes that the available body of epidemiologic evidence examining
the relationship between short-term exposures to Os concentrations and cardiovascular morbidity
is inconsistent (U.S. EPA, 2013, Section 6.3.2.9). Across studies, different definitions, (i.e., ICD-
9 diagnostic codes) were used for both all-cause and cause-specific cardiovascular morbidity
(ISA, U.S. EPA, 2013, see Tables 6-35 to 6-39), which may contribute to inconsistency in
results. However, within diagnostic categories, no consistent pattern of association was found
with Os. Generally, the epidemiologic studies used nearest air monitors to assess Os
concentrations, with a few exceptions that used modeling or personal exposure monitors.
The inconsistencies in the associations observed between short-term Os and cardiovascular
disease (CVD) morbidities are unlikely to be explained by the different exposure assignment
methods used (see Section 4.6, ISA, U.S. EPA, 2013). The wide variety of biomarkers
considered and the lack of consistency among definitions used for specific cardiovascular disease
endpoints (e.g., arrhythmias, HRV) make comparisons across studies difficult.
Despite the inconsistent evidence for an association between Os concentration and CVD
morbidity, mortality studies indicate a consistent positive association between short-term Os
exposure and cardiovascular mortality in multicity studies and in a multicontinent study. When
examining mortality due to cardiovascular disease, epidemiologic studies consistently observe
positive associations with short-term exposure to Os. Additionally, there is some evidence for an
association between long-term exposure to Os and mortality, although the association between
long-term ambient Os concentrations and cardiovascular mortality can be confounded by other
pollutants as evident by a study of cardiovascular mortality that reported no association after
adjustment for PM2.5 concentrations. The ISA (U.S. EPA, 2013, section 6.3.4) states that taken
together, the overall body of evidence across the animal and human studies is sufficient to
conclude that there is likely to be a causal reationship between relevant short-term exposures to
O3 and cardiovascular system effects.
3-51
-------
3.1.3 Adversity of Effects
In this section we address the following question:
• To what extent does the currently available scientific evidence expand our
understanding of the adversity of Os-related health effects?
In making judgments as to when various Cb-related effects become regarded as adverse
to the health of individuals, in previous NAAQS reviews staff has relied upon the guidelines
published by the ATS and the advice of CASAC. In 2000, the ATS published an official
statement on "What Constitutes an Adverse Health Effect of Air Pollution?" (ATS, 2000), which
updated and built upon its earlier guidance (ATS, 1985). The earlier guidance defined adverse
respiratory health effects as "medically significant physiologic changes generally evidenced by
one or more of the following: (1) interference with the normal activity of the affected person or
persons, (2) episodic respiratory illness, (3) incapacitating illness, (4) permanent respiratory
injury, and/or (5) progressive respiratory dysfunction", while recognizing that perceptions of
"medical significance" and "normal activity" may differ among physicians, lung physiologists
and experimental subjects (ATS, 1985). The 2000 ATS guidance builds upon and expands the
1985 definition of adversity in several ways. The guidance concludes that transient, reversible
loss of lung function in combination with respiratory symptoms should be considered adverse.
There is also a more specific consideration of population risk (ATS, 2000). Exposure to air
pollution that increases the risk of an adverse effect to the entire population is adverse, even
though it may not increase the risk of any individual to an unacceptable level. For example, a
population of asthmatics could have a distribution of lung function such that no individual has a
level associated with clinical impairment. Exposure to air pollution could shift the distribution to
lower levels that still do not bring any individual to a level that is associated with clinically
relevant effects. However, this would be considered to be adverse because individuals within the
population would have diminished reserve function, and therefore would be at increased risk to
further environmental insult (U.S. EPA, 2013, p. Ixxi; and 75 FR at 35526/2, June 22, 2010).
The ATS also concluded that elevations of biomarkers such as cell types, cytokines and
reactive oxygen species may signal risk for ongoing injury and more serious effects or may
simply represent transient responses, illustrating the lack of clear boundaries that separate
adverse from nonadverse events. More subtle health outcomes also may be connected
mechanistically to health effects that are clearly adverse, so that small changes in physiological
measures may not appear clearly adverse when considered alone, but may be part of a coherent
and biologically plausible chain of related health outcomes that include responses that are clearly
adverse, such as mortality (section 3.1.2.1, above).
3-52
-------
In this review, the new evidence provides further support for relationships between Os
exposures and a spectrum of health effects, including effects that meet the ATS criteria for being
adverse (ATS, 1985 and 2000). The ISA judgment that there is a causal relationship between
short-term Os exposure and a full range of respiratory effects, including respiratory morbidity
(e.g., lung function decrements, respiratory symptoms, inflammation, hospital admissions, and
emergency department visits) and mortality, provides support for concluding that short-term Os
exposure is associated with adverse effects (U.S. EPA, 2013, section 2.5.2). Overall, including
new evidence of cardiovascular system effects, the evidence supporting an association between
short-term Os exposures and total (non-accidental, cardiopulmonary) respiratory mortality is
stronger in this review (U.S. EPA, 2013, section 2.5.2). And the judgment of likely causal
associations between long-term measures of Cb exposure and respiratory effects such as new-
onset asthma, prevalence of asthma, asthma symptoms and control, and asthma hospital
admissions provides support for concluding that long-term Cb exposure is associated with
adverse effects ranging from episodic respiratory illness to permanent respiratory injury or
progressive respiratory decline (U.S. EPA, 2013, section 7.2.8).
This review provides additional evidence of Os-attributable effects that are clearly
adverse, including premature mortality. Application of the ATS guidelines to the least serious
category of effects related to ambient Os exposures, which are also the most numerous and
therefore are also potentially important from a public health perspective, involves judgments
about which medical experts on CASAC panels and public commenters have in the past
expressed diverse views. To help frame such judgments, EPA staff defined gradations of
individual functional responses (e.g., decrements in FEVi and airway responsiveness) and
symptomatic responses (e.g., cough, chest pain, wheeze), together with judgments as to the
potential impact on individuals experiencing varying degrees of severity of these responses.
These gradations were used in the 1997 Os NAAQS review and slightly revised in the 2008
review (U.S. EPA, 1996, p. 59; 2007, p. 3-72; 72 FR 37849, July 11, 2007). These gradations
and impacts are summarized in Tables 3-2 and 3-3 in the 2007 Os Staff Paper (U.S. EPA, 2007,
pp. 3-74 to 3-75).
For active healthy people, including children, moderate levels of functional responses
(e.g., FEVi decrements of > 10% but < 20%, lasting 4 to 24 hours) and/or moderate symptomatic
responses (e.g., frequent spontaneous cough, marked discomfort on exercise or deep breath,
lasting 4 to 24 hours) would likely interfere with normal activity for relatively few sensitive
individuals (U.S. EPA, 2007, p. 3-72; 72 FR 37849, July 11, 2007); whereas large functional
responses (e.g., FEVi decrements > 20%, lasting longer than 24 hours) and/or severe
symptomatic responses (e.g., persistent uncontrollable cough, severe discomfort on exercise or
deep breath, lasting longer than 24 hours) would likely interfere with normal activities for many
3-53
-------
sensitive individuals (U.S. EPA, 2007, p. 3-72; 72 FR 37849, July 11, 2007) and therefore would
be considered adverse under ATS guidelines. For the purpose of estimating potentially adverse
lung function decrements in active healthy people in the 2008 Os NAAQS review, the CASAC
panel for that review indicated that a focus on the mid to upper end of the range of moderate
levels of functional responses is most appropriate (e.g., FEVi decrements > 15% but < 20%)
(Henderson, 2006; U.S. EPA, 2007, p. 3-76). In this review, CASAC concurred that the
"[estimation of FEVi decrements of >15% is appropriate as a scientifically relevant surrogate
for adverse health outcomes in active healthy adults" (Frey, 2014, p. 3). However, for children
and adults with lung disease, even moderate functional (e.g., FEVi decrements > 10% but <
20%, lasting up to 24 hours) or symptomatic responses (e.g., frequent spontaneous cough,
marked discomfort on exercise or with deep breath, wheeze accompanied by shortness of breath,
lasting up to 24 hours) would likely interfere with normal activity for many individuals, and
would likely result in additional and more frequent use of medication (U.S. EPA, 2007, p.3-72;
72 FR 37849, July 11, 2007). For people with lung disease, large functional responses (e.g.,
FEVi decrements > 20%, lasting longer than 24 hours) and/or severe symptomatic responses
(e.g., persistent uncontrollable cough, severe discomfort on exercise or deep breath, persistent
wheeze accompanied by shortness of breath, lasting longer than 24 hours) would likely interfere
with normal activity for most individuals and would increase the likelihood that these individuals
would seek medical treatment (U.S. EPA, 2007, p.3-72; 72 FR 37849, July 11, 2007). In the last
Os NAAQS review, for the purpose of estimating potentially adverse lung function decrements
in people with lung disease the CASAC panel indicated that a focus on the lower end of the
range of moderate levels of functional responses is most appropriate (e.g., FEVi decrements
>10%) (Henderson, 2006; U.S. EPA, 2007, p. 3-76). In addition, in the reconsideration of the
2008 final decision, CASAC stated that "[a] 10% decrement in FEVI can lead to respiratory
symptoms, especially in individuals with pre-existing pulmonary or cardiac disease. For
example, people with chronic obstructive pulmonary disease have decreased ventilatory reserve
(i.e., decreased baseline FEVi) such that a >10% decrement could lead to moderate to severe
respiratory symptoms" (Samet, 2011) (section 3.1.2.1, above). In this review, CASAC concurred
that "[a]n FEVi decrement of >10% is a scientifically relevant surrogate for adverse health outcomes
for people with asthma and lung disease" (Frey, 2014, p. 3).
In judging the extent to which these impacts represent effects that should be regarded as
adverse to the health status of individuals, in previous NAAQS reviews we also considered
whether effects were experienced repeatedly during the course of a year or only on a single
occasion (Staff Paper, U.S. EPA, 2007). Although some experts would judge single occurrences
of moderate responses to be a "nuisance," especially for healthy individuals, a more general
consensus view of the adversity of such moderate responses emerges as the frequency of
3-54
-------
occurrence increases. Thus it has been judged that repeated occurrences of moderate responses,
even in otherwise healthy individuals, may be considered to be adverse since they could well set
the stage for more serious illness (61 FR 65723). The CAS AC panel in the 1997 NAAQS review
expressed a consensus view that these "criteria for the determination of an adverse physiological
response were reasonable" (Wolff, 1995). In the review completed in 2008, estimates of repeated
occurrences continued to be an important public health policy factor in judging the adversity of
moderate lung function decrements in healthy and asthmatic people (72 FR 37850, July 11,
2007).
Evidence new to this review indicates that 6.6-hour exposures to 60 ppb Os during
moderate exertion can result in pulmonary inflammation in healthy adults. As discussed in
section 3.1.2 above, the initiation of inflammation can be considered as evidence that injury has
occurred. Inflammation induced by a single Os exposure can resolve entirely, but continued
acute inflammation can evolve into a chronic inflammatory state (ISA, U.S. EPA, 2013, p. 6-76),
which is clearly adverse. Therefore, like moderate lung function decrements, whether
inflammation is experienced repeatedly during the course of a year or only on a single occasion
is judged by staff to be reasonable criteria for determining adverse inflammatory effects
attributable to Ch exposures at 60 ppb.
Responses measured in controlled human exposure studies indicate that the range of
effects elicited in humans exposed to ambient Os concentrations include: decreased inspiratory
capacity; mild bronchoconstriction; rapid, shallow breathing pattern during exercise; and
symptoms of cough and pain on deep inspiration (EPA, 2013, section 6.2.1.1). Some young,
healthy adults exposed to Os concentrations > 60 ppb, while engaged in 6.6 hours of intermittent
moderate exertion, develop reversible, transient decrements in lung function, symptoms of
breathing discomfort, and inflammation if minute ventilation or duration of exposure is increased
sufficiently (EPA, 2013, section 6.2.1.1). Among healthy subjects there is considerable
interindividual variability in the magnitude of the FEVi responses, but averaged across studies at
60 ppb (EPA, 2013, pp. 6-17 to 6-18), 10% of healthy subjects had >10% FEVi decrements.
Moreover, consistent with the findings of the ISA (EPA, 2013, section 6.2.1.1), CASAC
concluded that "[a]sthmatic subjects appear to be at least as sensitive, if not more sensitive, than
non-asthmatic subjects in manifesting ozone-induced pulmonary function decrements" (Frey, 2014,
p. 4). The combination of lung function decrements and respiratory symptoms, which has been
considered adverse in previous reviews, has been demonstrated in healthy adults following
prolonged (6.6 hour) exposures, while at intermittent moderate exertion, to 70 ppb. For these
types of effects, information from controlled human exposure studies, which provides an
indication of the magnitude and thus adversity of effects at different Os concentrations,
3-55
-------
combined with estimates of occurrences in the population from the HREA, provide information
about their importance from a policy perspective.
3.1.4 Ozone Concentrations Associated With Health Effects
In evaluating Os exposure concentrations reported to result in health effects, within the
context of the adequacy of the current standard, we first consider the following specific question:
• To what extent does the currently available scientific evidence indicate morbidity
and/or mortality attributable to exposures to Os concentrations lower than
previously reported or that would meet the current standard?
In addressing this question, we characterize the extent to which Os-attributable effects have been
reported over the ranges of Os exposure concentrations evaluated in controlled human exposure
studies and over the distributions of ambient Os concentrations in locations where epidemiologic
studies have been conducted.
3.1.4.1 Concentrations in Controlled Human Exposure Studies and in Epidemiologic
Panel Studies
In considering what the currently available evidence indicates with regard to effects
associated with exposure concentrations lower than those identified in the last review, or that
could meet the current standard, we first consider the evidence from controlled human exposure
studies and epidemiologic panel studies. This evidence is assessed in section 6.2 of the ISA and
is summarized in section 3.1.2 above. Controlled human exposure studies have generally been
conducted with young, healthy adults, and have evaluated exposure durations less than 8 hours.
Epidemiologic panel studies have evaluated a wider range of study populations, including
children, and have generally evaluated associations with Os concentrations averaged over several
hours (U.S. EPA, 2013, section 6.2.1.2).35
As summarized above (section 3.1.2.1), and as discussed in detail in the ISA (U.S. EPA,
2013, section 6.2), a large number of controlled human exposure studies have reported lung
function decrements, respiratory symptoms, airway inflammation, airway hyperresponsiveness,
and/or impaired lung host defense in young, healthy adults engaged in moderate, intermittent
exertion, following 6.6-hour Os exposures. These studies have consistently reported such effects
following exposures to Os concentrations of 80 ppb or greater. Available studies have also
evaluated some of these effects (i.e., lung function decrements, respiratory symptoms, airway
inflammation) following exposures to Os concentrations below 75 ppb. Table 3-1 highlights the
35 In this section we focus on panel studies that used on-site monitoring, and that are highlighted in the ISA for the
extent to which monitored ambient Os concentrations reflect exposure concentrations in their study populations
(U.S. EPA, 2013, section 6.2.1.2).
3-56
-------
group mean results of individual controlled human exposure studies that have evaluated
exposures of healthy adults to Os concentrations below 75 ppb. The studies included in Table 3-1
indicate lung function decrements, airway inflammation, and respiratory symptoms in healthy
adults following exposures to Os concentrations below 75 ppb.
3-57
-------
Table 3-1. Group mean results of controlled human exposure studies that have evaluated
exposures to ozone concentrations below 75 ppb in young, healthy adults.
Endpoint
FEVi decrements
Os Exposure
Concentration
70 ppb
60 ppb
40 ppb
Study
Schelegle et al, 200937
Kim etal., 2011
Schelegle et al., 200938
Adams, 2006
Adams, 2002
Adams, 2006
Adams, 2002
Statistically
Significant Os-
Induced
Effect36
yes
yes
no
yes39
no
no
no
Respiratory
Symptoms
70 ppb
60 ppb
40 ppb
Schelegle et al., 2009
Kim etal. ,2011
Schelegle et al., 2009
Adams, 2006
Adams, 2006
Adams, 2002
yes
no
no
no40
no
no
Airway
Inflammation
(neutrophil influx)
60 ppb
Kim etal. ,2011
yes
In further evaluating Os-induced FEVi decrements following exposures to Os
concentrations below 75 ppb, the ISA also combined the individual data from multiple studies of
healthy adults exposed for 6.6 hours to 60 ppb Os (Kim et al., 2011; Schelegle et al., 2009;
Adams, 2006, 2002, 1998). Based on these data, the ISA reports that 10% of exposed subjects
experienced FEVi decrements of 10% or more (i.e., abnormal and large enough to be potentially
adverse for people with pulmonary disease, based on past CASAC advice (section 3.1.3,
above))41 (U.S. EPA, 2013, section 6.2.1.1). Consistent with these findings, recently developed
36 Based on study population means.
37 As noted above, for the 70 ppb exposure concentration Schelegle et al. (2009) reported that the actual mean
exposure concentration was 72 ppb.
38 As noted above, for the 60 ppb exposure concentration Schelegle et al. (2009) reported that the actual mean
exposure concentration was 63 ppb.
39 In an analysis of the Adams (2006) data for square-wave chamber exposures, even after removal of potential
outliers, Brown et al. (2008) reported the average effect on FEVI at 60 ppb to be statistically significant (p < 0.002)
using several common statistical tests (U.S. EPA, 2013, section 6.2.1.1) (section 3.1.2.1, above).
40 Adams (2006) reported increased respiratory symptoms during a 6.6 hour exposure protocol with an average Cb
exposure concentration of 60 ppb. The increase in symptoms was reported to be statistically different from initial
respiratory symptoms, though not statistically different from filtered air controls.
41 As noted above (section 3.1.3), CASAC has previously stated that "[a] 10% decrement in FEVI can lead to
respiratory symptoms, especially in individuals with pre-existing pulmonary or cardiac disease. For example, people
with chronic obstructive pulmonary disease have decreased ventilatory reserve (i.e., decreased baseline FEVI) such
that a >10% decrement could lead to moderate to severe respiratory symptoms" (Samet, 2011).
3-58
-------
empirical models predict that the onset of Cb-induced FEVi decrements in healthy adults occurs
following exposures to 60 ppb Os for 4 to 5 hours while at moderate, intermittent exertion
(Schelegle et al., 2012), and that 9% of healthy adults exposed to 60 ppb Os for 6.6 hours would
experience FEVi decrements greater than or equal to 10% (McDonnell et al., 2012) (U.S. EPA,
2013, section 6.2.1.1; section 3.1.2.1, above). When the evidence for Os-induced lung function
decrements was taken together, the ISA concluded that (1) "mean FEVi is clearly decreased by
6.6-h exposures to 60 ppb Os and higher concentrations in subjects performing moderate
exercise" (U.S. EPA, 2013, p. 6-9) and (2) although group mean decrements following exposures
to 60 ppb Os are biologically small, "a considerable fraction of exposed individuals experience
clinically meaningful decrements in lung function" (U.S. EPA, 2013, p. 6-20).
In considering the specific question above, controlled human exposure studies have
reported decreased lung function, increased airway inflammation, and increased respiratory
symptoms in healthy adults following exposures to Os concentrations below 75 ppb. Such
impairments in respiratory function have the potential to be adverse, based on ATS guidelines
for adversity and based on previous advice from CASAC (section 3.1.3, above). In addition, if
they become serious enough, these respiratory effects could lead to the types of clearly adverse
effects commonly reported in Os epidemiologic studies (e.g., respiratory emergency department
visits, hospital admissions). Therefore, following exposures to Os concentrations lower than 75
ppb, controlled human exposure studies have reported respiratory effects that could be adverse in
some individuals, particularly if experienced by members of at-risk populations (e.g., asthmatics,
children).42
In further considering effects following exposures to Os concentrations below 75 ppb, we
also note that the ISA highlights some epidemiologic panel studies for the extent to which
monitored ambient Os concentrations reflect exposure concentrations in their study populations
(U.S. EPA, 2013, section 6.2.1.2). Specifically, Table 3-2 below includes Os panel studies that
have evaluated associations with lung function decrements for Os concentrations at or below 75
ppb, and that measured Os concentrations with monitors located in the areas where study
subjects were active (e.g., on site at summer camps or in locations where exercise took place)
(U.S. EPA, 2013, section 6.2.1.2 and Table 6-6). Epidemiologic panel studies have evaluated a
wider range of populations and lifestages than controlled human exposure studies of Os
concentrations below 75 ppb (e.g., including children).
42 These effects were reported in healthy individuals. Consistent with past CASAC advice (Samet, 2011), and
evidence in the ISA (U.S. EPA, 2013, p. 6-77), it is a reasonable inference that the effects would be greater in
magnitude and potential severity for at-risk groups. See National Environmental Development Ass 'n Clean Air
Project v. EPA, 686 F. 3d 803, 811 (D.C. Cir. (2012) (making this point).
3-59
-------
Table 3-2. Panel studies of lung function decrements with analyses restricted to
concentrations below 75 ppb.
Study
Spektor et al.
(1988a)
Chan and
Wu (2005)
Korrick et al.
(1998)
Brauer et al.
(1996)
Brunekreef
etal. (1994)
Population
Children at
summer camp
Mail carriers
Adult hikers
Farm workers
Exercising
adults
Os Concentrations
Restricted to 1 -hour concentrations
below 60 ppb
Maximum 8-hour average was 65
ppb
2- to 12-hour average from 40 to
74 ppb during hikes
Restricted to 1 -hour maximum
below 40 ppb
Restricted to 1 -hour maximum
below 30 ppb
Restricted to 10-minute to 2.4-hour
averages below 61 ppb
Restricted to 10-minute to 2.4-hour
averages below 5 1 ppb
Restricted to 10-minute to 2.4-hour
averages below 41 ppb
Statistically Significant Association
with Lung Function Decrements
Yes
Yes
Yes
Yes
No
No
No
No
Although these studies report health effect associations for different averaging times, and it is not
clear the extent to which specific Os exposure conditions (i.e., concentrations, durations of
exposure, degrees of activity) were responsible for eliciting reported decrements, they are
consistent with the findings of the controlled human exposure studies discussed above.
Specifically, the epidemiologic panel studies in Table 3-2 indicate Os-associated lung function
decrements when on-site monitored concentrations (ranging from minutes to hours) were below
75 ppb, with the evidence becoming less consistent at lower Os concentrations.
3.1.4.2 Concentrations in Epidemiologic Studies - Short-term Metrics
We next consider distributions of ambient Os concentrations in locations where
epidemiologic studies have evaluated Os-associated hospital admissions, emergency department
visits, and/or mortality. When considering epidemiologic studies within the context of the current
standard, we emphasize those studies conducted in the U.S. and Canada. Such studies reflect air
quality and exposure patterns that are likely more typical of the U.S. population than the air
quality and exposure patterns reflected in studies conducted outside the U.S. and Canada (section
1.3.1.2, above).43 We also emphasize studies reporting associations with effects judged in the
ISA to be robust to confounding by other factors, including co-occurring air pollutants. In
addition to these factors, we consider the statistical precision of study results, the extent to which
43 Nonetheless, we recognize the importance of all studies, including international studies, in the ISA's assessment
of the weight of the evidence that informs causality determinations.
3-60
-------
studies report associations in at-risk populations, and the extent to which the biological
plausibility of associations at various ambient Os concentrations is supported by controlled
human exposure and/or animal toxicological studies. These considerations help inform the range
of ambient Os concentrations over which we have the most confidence in Os-associated health
effects, and the range of concentrations over which our confidence in such associations is
appreciably lower. We place particular emphasis on characterizing those portions of distributions
of ambient Os concentrations likely to meet the current standard.
In our consideration of these issues, we first address the following question:
• To what extent have U.S. and Canadian epidemiologic studies reported
associations with mortality or morbidity in locations that would have met the
current Os standard during the study period?
Addressing this question can provide important insights into the extent to which Os-health effect
associations are present for distributions of ambient Os concentrations that would be allowed by
the current standard. To the extent Os health effect associations are reported in study areas that
would have met the current standard, we have greater confidence that the current standard could
allow the clearly adverse Os-associated effects indicated by those studies (e.g., mortality,
hospital admissions, emergency department visits).44
Epidemiologic studies evaluate statistical associations between variation in the incidence
of health outcomes and variation in ambient Os concentrations. In many of the Os epidemiologic
studies assessed in the ISA, ambient concentrations are averaged across multiple monitors within
study areas, and in some cases over multiple days. These averages are used as surrogates for the
spatial and temporal patterns of Os exposures in study populations. In this PA, we refer to these
averaged concentrations as "area-wide" Os concentrations.
The area-wide concentrations reported in many epidemiologic studies do not identify the
actual Os exposures that may be eliciting the observed health outcomes. Thus, in considering
epidemiologic studies of mortality and morbidity, we are not drawing conclusions regarding
single short-duration Os concentrations in ambient air that, alone, are eliciting the reported health
outcomes. Rather, our focus in this section is to consider what these studies convey regarding the
extent to which health effects may be occurring (i.e., as indicated by associations) under air
quality conditions meeting the current standard.
44 See ATA III, 283 F.3d at 370 (EPA justified in revising NAAQS when health effect associations are observed at
levels allowed by the NAAQS).
3-61
-------
In order to facilitate consideration of the question above, we have identified U.S. and
Canadian studies of respiratory hospital admissions, respiratory emergency department visits,45
and mortality (total, respiratory, cardiovascular) from the ISA (studies identified from U.S. EPA,
2013, Tables 6-28, 6-42, and 6-53, and section 6.2.8) (Appendix 3B). For each monitor in the
areas evaluated by these studies, we have identified the 3-year averages of the annual 4th highest
daily maximum 8-hour Ch concentrations (Appendix 3B).46 To provide perspective on whether
study cities would have met or violated the current Os NAAQS during the study period, these Os
concentrations were compared to the level of the current standard. Based on this approach, a
study city was judged to have met the current standard during the study period if all of the 3-year
averages of annual 4th highest 8-hour Os concentrations in that area were at or below 75 ppb.
Based on these analyses, the large majority of epidemiologic study areas evaluated would
have violated the current standard during study periods (Appendix 3B). Table 3-3 below
highlights the subset of U.S. and Canadian studies that evaluated Os health effect associations in
locations that would have met the current standard during study periods. This includes a U.S.
single-city study that would have met the current standard over the entire study period (Mar and
Koenig, 2009) and four Canadian multicity studies for which the majority of study cities would
have met the current standard over the entire study periods (Cakmak et al., 2006; Dales et al.,
2006; Katsouyanni et al., 2009; Stieb et al., 2009).47
45 Given the inconsistency in results across cardiovascular morbidity studies (U.S. EPA, 2013, section 6.3.2.9), our
consideration of the morbidity evidence in this section focuses on studies of respiratory hospital admissions and
emergency department visits.
46 These concentrations are referred to as "design values." A design value is a statistic that is calculated at individual
monitors and based on 3 consecutive years of data collected from that site. In the case of Os, the design value for a
monitor is based on the 3-year average of the annual 4th highest daily maximum 8-hour Os concentration in parts
per billion (ppb). For U.S. study areas, we used EPA's Air Quality System (AQS)
(http://www.epa.gov/ttn/airs/airsaqs/) to identify design values. For Canadian study areas, we used publically
available air quality data from the Environment Canada National Air Pollution Surveillance Network
(http://www.etc-cte.ec.gc.ca/napsdata/main.aspx). We followed the data handling protocols for calculating design
values as detailed in 40 CFR Part 50, Appendix P.
47 In addition, a study by Vedal et al. (2003) was included in the 2006 CD (U.S. EPA, 2006). This study reported
positive and statistically significant associations with mortality in Vancouver during a time period when the study
area would have met the current standard (U.S. EPA, 2007). This study was not highlighted in the ISA in the current
review (U.S. EPA, 2013).
3-62
-------
Table 3-3. U.S. and Canadian epidemiologic studies reporting Os health effect
associations in locations that would have met the current standard during
study periods.
Authors
Cakmak et al.
(2006)
Dales et al.
(2006)
Katsouyanni et
al. (2009)
Katsouyanni et
al. (2009)
Mar and
Koenig (2009)
Stieb et al.
(2009)
Study Results
Positive and statistically significant association
with respiratory hospital admissions
Positive and statistically significant association
with respiratory hospital admissions
Positive and statistically significant associations
with respiratory hospital admissions
Positive and statistically significant associations
with all-cause and cardiovascular mortality48
Positive and statistically significant associations
with asthma emergency department visits in
children (< 18 years) and adults (> 18 years)
Positive and statistically significant association
with asthma emergency department visits
Cities
10
Canadian
cities
11
Canadian
cities
12
Canadian
cities
12
Canadian
cities
Seattle
7 Canadian
cities
Number of cities meeting
the current standard over
entire study period
7
7
10
8
1
5
As illustrated in Table 3-3, Mar and Koenig reported health effect associations with
asthma emergency department visits in a location that would have met the current standard over
the entire study period. This analysis indicates that the current standard would allow the
distribution of ambient Os concentrations that provided the basis for reported associations with
respiratory emergency department visits.
In addition, four multicity studies reported associations with mortality or morbidity when
the majority of study locations would have met the current standard over the entire study periods.
Thus, the current standard would allow the majority of the distributions of ambient Os
concentrations that provided the basis for positive and statistically significant associations with
mortality or morbidity. Our interpretation of these results is complicated by uncertainties in the
extent to which multicity effect estimates (i.e., which are based on combining estimates from
multiple study locations) can be attributed to ambient Os in the subset of locations that would
have met the current standard, versus Os in the smaller number of locations that would have
violated the standard. While there is uncertainty in ascribing the multicity effect estimates
reported in these Canadian studies to ambient concentrations that would have met the current
48 Katsouyanni et al. (2009) report a positive and statistically significant association with cardiovascular mortality
for people aged 75 years or older.
3-63
-------
standard, the information in Table 3-3 suggests that reported multicity effect estimates are
largely influenced by locations meeting the current standard (i.e., given that most study areas
would have met this standard). Together, these U.S. and Canadian epidemiologic studies suggest
a relatively high degree of confidence in the presence of associations with mortality and
morbidity for ambient Os concentrations meeting the current standard.
We next consider the extent to which additional epidemiologic studies of mortality or
morbidity, specifically those conducted in locations that violated the current standard, can also
inform our consideration of adequacy of the current standard. In doing so, we note that health
effect associations reported in epidemiologic studies are influenced by the full distributions of
ambient Os concentrations, including concentrations below the level of the current standard. We
focus on studies that have explicitly characterized such Os health effect associations, including
confidence in those associations, for various portions of distributions of ambient Os
concentrations. In doing so, we consider the following question:
• To what extent do analyses from epidemiologic studies indicate confidence in health
effect associations over distributions of ambient Os concentrations, including at
concentrations lower than previously identified or below the current standard?
We first focus on those studies that have reported confidence intervals around
concentration-response functions over distributions of ambient Os concentrations. Confidence
intervals around concentration-response functions can provide insights into the range of ambient
concentrations over which the study indicates the most confidence in the reported health effect
associations (i.e., where confidence intervals are narrowest), and into the range of ambient
concentrations below which the study indicates that uncertainty in the nature of such associations
becomes notably greater (i.e., where confidence intervals become markedly wider). The
concentrations below which confidence intervals become markedly wider in such analyses are
intrinsically related to data density, and do not necessarily indicate the absence of an association.
The ISA identifies several epidemiologic studies that have reported confidence intervals
around concentration-response functions in U.S. cities. The ISA concludes that studies generally
indicate a linear concentration-response relationship "across the range of 8-h max and 24 h avg
Os concentrations most commonly observed in the U.S. during the Os season" and that "there is
less certainty in the shape of the C-R curve at the lower end of the distribution of Os
concentrations" (U.S. EPA, 2013, pp. 2-32 to 2-34). In characterizing the Os concentrations
below which such certainty decreases, the ISA discusses area-wide Os concentrations as low as
20 ppb and as high as 40 ppb (U.S. EPA, 2013, section 2.5.4.4).
Consistent with these conclusions, the range of ambient concentrations over which the
evidence indicates the most certainty in concentration-response relationships can vary across
3-64
-------
studies. Such variation is likely due at least in part to differences in the Os metrics evaluated and
differences in the distributions of ambient concentrations and health events. Thus, although
consideration of confidence intervals around concentration-response functions can provide
valuable insights into the ranges of ambient concentrations over which studies indicate the most
confidence in reported health effect associations, there are limitations in the extent to which
these analyses can be generalized across Os metrics, study locations, study populations, and
health endpoints.
The ISA emphasizes two U.S. single-city studies that have reported confidence intervals
around concentration-response functions (Silverman and Ito, 2010; Strickland et al., 2010).
These studies, and their associated Os air quality, are discussed below.
Silverman and Ito (2010) evaluated associations between 2-day rolling average Cb
concentrations49 and asthma hospital admissions in New York City from 1999 to 2006 (a time
period when the study area would have violated the current standard, Appendix 3B). As part of
their analysis, the authors evaluated the shape of the concentration-response relationship for Os
using a co-pollutant model that included PM2.5 (reprinted in Figure 3-4, below). Based on their
analyses, Silverman and Ito (2010) concluded a linear relationship between Cb and hospital
admissions is a reasonable approximation of the concentration-response function throughout
much of the range of ambient Os concentrations. Based on visual inspection of Figure 3-4 below
(Figure 3 from published study), we note that confidence in the reported concentration-response
relationship is highest for area-wide average Os concentrations around 40 ppb (i.e., near the
reported median of 41 ppb), and decreases notably for concentrations at and below about 20 ppb.
49 2-day rolling averages of daily maximum 8-hour O3 concentrations were calculated throughout the study period,
averaged across study monitors.
3-65
-------
Ozone: All Ages
cc
(L) ^
(L)
CC
inn
:i mi i ii 11 i M
20 40 60 80
0 Concentration (ppb)
100
Figure 3-4. Concentration-response function for asthma hospital admissions over the
distribution of area-wide averaged Os concentrations (adapted from Silverman
and Ito, 2010).50
In considering the concentration-response function presented by Silverman and Ito (2010)
within the context of the adequacy of the current standard, we recognize that true design values
cannot be identified for the subsets of air quality data contributing to various portions of the
concentration-response function.51 Therefore, to use this analysis to inform our consideration of
the adequacy of the current standard we evaluate the extent to which the concentration-response
function indicates a relatively high degree of confidence in the reported health effect association
on days when all monitored 8-hour Os concentrations were below 75 ppb (Table 3-4, below).
This approach can provide insight into the extent to which the reported Os health effect
association is present when all monitored Os concentrations are below the level of the current
standard.
Based on the information in Table 3-4 below, when 2-day averaged Os concentrations
ranged from 36 to 45 ppb (i.e., around the median, where confidence intervals are narrowest),
there were 3 days (out of 432) with at least one monitor recording a daily maximum 8-hour Os
concentration above the level of the current standard (approximately 0.7% of days). When 2-day
50This figure was also reprinted in the ISA (U.S. EPA, 2013; Figure 6-16).
51 As discussed above, O3 design values are calculated using all data available from a monitor.
3-66
-------
averaged Os concentrations ranged from 26 to 45 ppb (i.e., extending to concentrations below the
median, but still above the concentrations where confidence intervals widen notably), there were
4 days (out of 816) with at least one monitor recording a daily maximum 8-hour Os
concentration above the level of the current standard (approximately 0.5% of days). Thus, on
over 99% of the days when area-wide "averaged" Os concentrations were between 26 and 45
ppb, the highest daily maximum 8-hour Os concentrations were below 75 ppb. For comparison,
the annual 4th highest daily maximum 8-hour Os concentration generally corresponds to the 98th
or 99th percentile of the seasonal distribution, depending on the length of the Os season.
Table 3-4. Distributions of daily 8-hour maximum ozone concentrations from highest
monitors over range of 2-day moving averages from composite monitors (for
study area evaluated by Silverman and Ito, 2010)
Distribution of 8-hr
max from highest
monitors
Min
5th
25th
50th
75th
95th
98th
99th
Max
Days > 75 ppb
2-day moving average across monitors (ppb)
11 to 20
(62 days)
15
16
20
24
29
37
41
41
42
0
21 to 25
(92 days)
21
23
28
31
36
49
56
57
59
0
26 to 30
(178 days)
19
25
32
36
42
50
60
67
80
1
31 to 35
(206 days)
26
33
38
43
47
55
71
75
75
0
36 to 40
(236 days)
25
34
42
47
52
61
67
69
79
1
41 to 45
(196 days)
15
37
46
52
59
72
75
87
91
2
46 to 50
(153 days)
31
38
51
59
65
77
85
91
97
9
51 to 55
(111 days)
30
37
54
62
69
80
89
94
94
15
56 to 60
(71 days)
41
46
60
68
78
90
93
93
93
20
In a separate study, Strickland et al. (2010) evaluated associations between 3-day rolling
average Os concentrations52 and asthma hospital admissions in Atlanta during the warm season
from 1994 to 2004 (a time period when the study area would have violated the current standard,
Appendix 3B). As part of this analysis, Strickland et al. (2010) evaluated the concentration-
response relationship for Os and pediatric asthma emergency department visits. The authors
reported the shape of the concentration-response function to be approximately linear with no
evidence of a threshold when 3-day averaged daily maximum 8-hour Os concentrations were
approximately 30 to 80 ppb (Figure 3-5 below and U.S. EPA, 2013, Figure 6-18). Figure 3-5
below illustrates that the confidence intervals around the concentration-response function are
52 Three-day rolling averages of population-weighted daily maximum 8-hour O3 concentrations were calculated
throughout the study period (Strickland et al., 2010).
3-67
-------
narrowest around the study mean (i.e., 55 ppb), and that these confidence intervals do not widen
notably for "averaged" Os concentrations as low as about 30 ppb.
Ozone Warm Season
CM -
LT>
ro
OH
-------
Thus, on over 99% of the days when "averaged" Os concentrations were between 31 and 45 ppb,
all monitors measured daily maximum 8-hour Cb concentrations below 75 ppb.
Table 3-5. Distribution of daily 8-hour maximum ozone concentrations from highest
monitors over range of 3-day moving averages of population-weighted
concentrations (for study area evaluated by Strickland et al., 2010)
Distribution of 8-
hr max from
hiehest monitors
Min
5th
25th
50th
75th
95th
98th
99th
Max
Days > 75
3 -day moving average across monitors (j
26-30
(75 days)
15
20
29
33
40
53
67
68
70
0
31-35
(144 days)
18
22
31
37
44
55
59
63
64
0
36-40
(165 days)
21
28
35
42
49
68
74
78
82
2
41-45
(210 days)
19
27
37
47
56
69
72
78
90
2
46-50
(235 days)
25
35
46
52
62
75
83
85
92
10
wb)
51-55
(244 days)
23
38
52
60
67
80
88
95
97
24
56-60
(272 days)
25
39
54
63
73
88
95
99
106
53
In summary, analyses of air quality data from the study locations evaluated by Silverman
and Ito (2010) and Strickland et al. (2010) indicate a relatively high degree of confidence in
reported statistical associations with respiratory health outcomes on days when virtually all
monitored 8-hour Os concentrations were 75 ppb or below. Though these analyses do not
identify true design values, the presence of Os-associated respiratory effects on such days
provides insight into the types of health effects that could occur in locations with maximum
ambient Os concentrations below the level of the current standard.
We next consider the following question:
• To what extent are there important uncertainties in analyses of confidence in
concentration-response functions?
There are several important uncertainties that are specifically related to our analyses of
distributions of Os air quality in the study locations evaluated by Silverman and Ito (2010) and
Strickland et al. (2010). Although these studies report health effect associations with two-day
(Silverman and Ito) and three-day (Strickland) averages of daily Os concentrations, it is possible
that the respiratory morbidity effects reported in these studies were also at least partly
attributable to the days immediately preceding these two- and three-day periods. In support of
this possibility, Strickland et al. reported positive and statistically significant associations with
emergency department visits for multiple lag periods, including lag periods exceeding three days.
3-69
-------
Our analysis of highest monitored concentrations focuses on two- and three- day periods, as used
in the published study to generate concentration-response functions. This could have important
implications for our interpretation of the reported concentration-response functions if a 2-day
period with no monitors measuring 8-hour concentrations at or above 75 ppb is immediately
preceded by one or more days with monitors that do exceed 75 ppb. Although we do not know
the extent to which Os concentrations on a larger number of days could have contributed to
reported health effect associations, we note this as a potentially important uncertainty in our
consideration of concentration-response functions within the context of the current standard.
In addition, an important uncertainty that applies to epidemiologic studies in general is
the extent to which reported health effects are caused by exposures to Cb itself, as opposed to
other factors such as co-occurring pollutants or other pollutant mixtures. Although both of the
studies evaluated above reported health effect associations in co-pollutant models, this
uncertainty becomes an increasingly important consideration as health effect associations are
evaluated at lower ambient Os concentrations (i.e., presumably corresponding to lower exposure
concentrations).
One approach to considering the potential importance of this uncertainty in
epidemiologic studies is to evaluate the extent to which there is coherence with the results of
experimental studies (i.e., in which the study design dictates that exposures to Os itself are
responsible for reported effects). Therefore, in further considering uncertainties associated with
the above air quality analyses for the study areas evaluated by Silverman and Ito (2010) and
Strickland et al. (2010), we evaluate the following question:
• To what extent is there coherence between evidence from controlled human
exposure studies and epidemiologic studies supporting the occurrence of Os-
attributable respiratory effects when daily maximum 8-hour ambient Os
concentrations are at or below 75 ppb?
As summarized above and as discussed in the ISA (U.S. EPA, 2013, section 6.2),
controlled human exposure studies demonstrate the occurrence of respiratory effects in an
appreciable percentage of healthy adults following single short-term exposures to Os
concentrations as low as 60 ppb. As Os exposure concentrations exceed 60 ppb: 1) effects in
healthy adults become larger and more serious; 2) a broader range of effects are observed in a
greater percentage of exposed individuals; and 3) effects are reported more consistently across
studies. In addition, exposure concentrations below 60 ppb could potentially result in respiratory
effects, particularly in at-risk populations such as children and asthmatics. Thus, as the potential
increases for portions of epidemiologic study populations to be exposed to Os concentrations
approaching or exceeding 60 ppb, our confidence increases that reported respiratory health
effects could be caused by exposures to the ambient Os concentrations present in study locations.
3-70
-------
As discussed above, for the study by Silverman and Ito (2010), 26 to 45 ppb represents
the lower end of the range of "averaged" concentrations over which the study indicates a
relatively high degree of confidence in the statistical association with respiratory hospital
admissions (and for which virtually all monitored concentrations were 75 ppb or below). As
averaged concentrations increase from 26 to 45 ppb, the number of days with maximum
monitored concentrations approaching or exceeding 60 ppb increases (Table 3-4, above).55 For
example, of the 178 days with area-wide average concentrations from 26 to 30 ppb, only about
5% had monitors recording ambient concentrations of 50 ppb or greater and about 2% had
monitors recording concentrations of 60 ppb or greater. In contrast, of the 196 days with area-
wide average concentrations from 41 to 45 ppb, about half had one or more monitors recording
ambient concentrations above 50 ppb and about 25% had monitors recording concentrations at or
above 60 ppb. On a small number of these days, at least one monitored concentration exceeded
70 or 80 ppb. Thus as averaged concentrations approach 45 ppb there is an increasing likelihood
that at least some portion of the Silverman and Ito study population could have been exposed to
Os concentrations near or above those shown to cause respiratory effects in healthy adults. If
these effects become serious enough (e.g., in people with asthma) they could lead to the
respiratory-related hospital admissions reported in the study. This analysis is consistent with the
occurrence of Os-attributable respiratory hospital admissions, even when virtually all monitored
concentrations were below the level of the current standard. Similar results were obtained for the
study by Strickland et al. (2010) (Table 3-5, above).
In further evaluating Os concentration-response relationships within the context of the
adequacy of the current standard, we note that some epidemiologic studies report health effect
associations for air quality subsets restricted to ambient pollutant concentrations below one or
more predetermined cut points. Such "cut point" analyses can provide information on the
magnitude and statistical precision of effect estimates for defined distributions of ambient
concentrations, which may in some cases include distributions that would meet the current
standard. Therefore, we next consider the following question:
• To what extent do cut-point analyses from epidemiologic studies report health effect
associations at ambient Os concentrations lower than previously identified or that
would likely meet the current standard?
By considering the magnitude and statistical significance of effect estimates for restricted
air quality distributions, cut-point analyses can provide insight into the extent to which health
55 Though, as noted above, the epidemiologic studies by Silverman and Ito (2010) and Strickland et al. (2010) do
not provide information on the extent to which reported health effects result from exposures to any specific Os
concentrations.
3-71
-------
effect associations are driven by ambient concentrations above the cut point, versus
concentrations below the cut point. For studies that evaluate multiple cut points, these analyses
can provide insights into the magnitude and statistical precision of health effect associations for
different portions of the distribution of ambient concentrations, including insights into the
ambient concentrations below which uncertainty in reported associations becomes notably
greater. As with analyses of concentration-response functions, discussed above, the cut points
below which confidence intervals become notably wider depend in large part on data density.56
In the U.S. multicity study by Bell et al. (2006), study authors used the NMMAPS data
set to evaluate associations between 2-day rolling average Os concentrations57 and total (non-
accidental) mortality in 98 U.S. cities from 1987 to 2000. Based on the full distributions of
ambient Os concentrations in study cities, the large majority of the NMMAPS cities would have
violated the current standard during the study period (Appendix 3B). However, Bell et al. (2006)
also reported health effect associations in a series of cut-point analyses, with effect estimates
based only on the subsets of days contributing to "averaged" Os concentrations below cut points
ranging from 5 to 60 ppb (see Figure 2 in Bell et al., 2006). The lowest cut-point for which the
association between Os and mortality was reported to be statistically significant was 30 ppb
(based on visual inspection of Figure 2 in the published study). As with the studies by Silverman
and Ito (2010) and Strickland et al. (2010), discussed above, we consider what these cut point
analyses indicate with regard to the potential for health effect associations to extend to ambient
Os concentrations likely to be allowed by the current Os NAAQS.
We attempted to recreate the subsets of air quality data used in the cut point analyses
presented by Bell et al. (2006). In doing so, we applied the criteria described in the published
study to generate air quality subsets corresponding to those defined by the cut points evaluated
by study authors.58 From the days with averaged Os concentrations below each cut point, we
identified 3-year averages of annual 4th highest daily maximum 8-hour Os concentrations in each
study area. We then compared these 4th highest Os concentrations to the level of the current
standard in order to provide insight into the extent to which the air quality distributions included
in various cut point analyses would likely have met the current standard.
56 As such, these analyses provide insight into the ambient concentrations below which the available air quality
information becomes too sparse to support conclusions about the nature of concentration-response relationships,
with a high degree of confidence.
57 Two-day rolling averages of 24-hour average Os concentrations were calculated throughout the study period. This
calculation was done across study monitors in study cities with multiple monitors.
58 We were unable to obtain the air quality data used to generate the cut-point analyses in the study published by
Bell et al. (2006). Therefore, we generated 2-day averages of 24-hour Os concentrations in study locations using the
air quality data available in AQS, combined with the published description of study area definitions. In doing so, we
did not recreate the trimmed means used by Bell. As discussed below, this represents an important uncertainty in our
analysis.
3-72
-------
We particularly focus on the lowest cut-point for which the association between Os and
mortality was reported in this study to be statistically significant (i.e., 30 ppb, as noted above).
Based on the Os air quality concentrations that met the criteria for inclusion in the 30 ppb cut
point analysis, 95% of study areas had 3-year averages of annual 4th highest daily maximum 8-
hour Os concentration at or below 75 ppb over the entire study period. For the 35 ppb cut point,
which also resulted in a statistically significant association with mortality, 68% of study areas
had 3-year averages of annual 4th highest daily maximum 8-hour Os concentration at or below 75
ppb. This suggests that the large majority of air quality distributions that provided the basis for
positive and statistically significant associations with mortality (i.e., for the 30 and 35 ppb cut
points) would likely have met the current Os standard. For higher cut points, all of which also
resulted in statistically significant associations with mortality, the majority of study cities had 3-
year averages of annual 4th highest daily maximum 8-hour concentrations greater than 75 ppb.
3-73
-------
Table 3-6.
Number of study cities with 4th highest daily maximum 8-hour concentrations
greater than 75 ppb, for various cut-point analyses presented in Bell et al.
(2006)
Number (%) of
Cities with 4th
highest >75 (any
3 -yr period;
1987-2000)
Cut-point for 2-day moving average across monitors and cities (24-hour average)
25
0 (0%)
30
5 (5%)
35
31
(32%)
40
70
(71%)
45
86
(88%)
50
88
(90%)
55
92
(94%)
60
92
(94%)
All
92
(94%)
In addition to the uncertainties noted above for our analysis of the single-city studies by
Silverman and Ito (2010) and Strickland et al. (2010) (e.g., attributing effects specifically to air
quality included in various subsets), an important uncertainty related to this analysis is that we
were unable to obtain the air quality data used to generate the cut-point analyses in the study
published by Bell et al. (2006). Therefore, as noted above, we generated 2-day averages of 24-
hour Os concentrations in study locations using the air quality data available in AQS, combined
with the published description of study area definitions. In doing so, we did not recreate the
trimmed means used by Bell. An important uncertainty in this approach is the extent to which we
were able to appropriately recreate the cut-point analyses in the published study.
The ISA also notes important uncertainties inherent in multicity studies that evaluate the
potential for thresholds to exist, as was done in the study by Bell et al. (2006). Specifically, the
ISA highlights the regional heterogeneity in Os health effect associations as a factor that could
obscure the presence of thresholds, should they exist, in multicity studies (U.S. EPA, 2013,
sections 2.5.4.4 and 2.5.4.5). The ISA notes that community characteristics (e.g., activity
patterns, housing type, age distribution, prevalence of air conditioning) could be important
contributors to reported regional heterogeneity (U.S. EPA, 2013, section 2.5.4.5). Given this
heterogeneity, the ISA concludes that "a national or combined analysis may not be appropriate to
identify whether a threshold exists in the Os-mortality C-R relationship" (U.S. EPA, 2013, p. 2-
33). This represents an important source of uncertainty when characterizing our confidence in
reported concentration-response relationships over distributions of ambient Os concentrations,
based on multicity studies. This uncertainty becomes increasingly important when interpreting
concentration-response relationships at lower ambient Os concentrations, particularly those
concentrations corresponding to portions of distributions where data density decreases notably.
3.1.4.3 Concentrations in Epidemiologic Studies - "Long-term" Metrics
We next consider the extent to which epidemiologic studies employing longer-term
ambient Os concentration metrics inform our understanding of the air quality conditions
3-74
-------
associated with Ch-attributable health effects, and specifically inform consideration of the extent
to which such effects could occur under air quality conditions meeting the current standard.
Unlike for the studies of short-term Os discussed above, the available U.S. and Canadian
epidemiologic studies evaluating long-term ambient Os concentration metrics have not been
conducted in locations likely to have met the current 8-hour Os standard during the study period
(Appendix 3B). Therefore, although these studies contribute to our understanding of health
effects associated with long-term or repeated exposures to ambient Os (as summarized in section
3.1.2 above), consideration of study area design values does not inform our consideration of the
extent to which those health effects may be occurring in locations that met the current standard.
In further considering epidemiologic studies of long-term Os concentrations, we also
evaluate the extent to which concentration-response functions, including associated confidence
intervals, have been characterized for distributions of ambient Os, and what those functions can
tell us about health effect associations for Os concentrations likely to be allowed by the current
standard. Specifically, we consider the following question:
• To what extent do confidence intervals around concentration-response functions
indicate Os-associated health outcomes at ambient concentrations meeting the
current Os standard?
The ISA identifies a single epidemiologic study reporting confidence intervals around a
concentration-response function for "long-term" Os concentrations and respiratory mortality
(Jerrett et al., 2009; U.S. EPA, 2013, sections 7.2.7, 7.2.8 and 7.7). Jerrett et al. (2009) reported
that when seasonal averages of 1-hour daily maximum Os concentrations59 ranged from 33 to
104 ppb, there was no statistical deviation from a linear concentration-response relationship
between Os and respiratory mortality across 96 U.S. cities (U.S. EPA, 2013, section 7.7).
However, the authors reported "limited evidence" for an effect threshold at an Os concentration
of 56 ppb (p=0.06).60 Visual inspection of this concentration-response function (Figure 3-6)
confirms the possibility of an inflection point just below 60 ppb, which is close to the median
concentration across cities (i.e., 57 ppb).
59 Jerrett et al. (2009) evaluated the April to September averages of 1-hour daily maximum Os concentrations across
96 U.S. metropolitan areas from 1977- 2000. In urban areas with multiple monitors, April to September 1-hour daily
maximum concentrations from each individual monitor were averaged. This step was repeated for each year in the
study period. Finally, each yearly averaged Os concentrations was then averaged again to yield the single averaged
1-hour daily maximum Os concentration depicted on the x axis of Figure 3-6 below.
60 This issue is discussed further in section 3.2.3.2, below.
3-75
-------
0.2-
40 60 80 100
Daily 1-hour Maximum O3 Level (ppb). I9~~-2000
Figure 3-6. Exposure-Response relationship between risk of death from respiratory causes
and ambient Os concentration study metric (Jerrett et al., 2009).
We consider the extent to which this concentration-response function indicates
confidence in the reported health effect association at various ambient Os concentrations. In
identifying the concentrations over which we have the greatest confidence, we note the
following: (1) most of the study cities had Os concentrations above 53.1 ppb (i.e., the upper
bound of the first quartile), accounting for approximately 72% of the respiratory deaths in the
cohort (Table 2 in Jerrett et al. 2009); (2) confidence intervals widen notably for Os
concentrations in the first quartile (based on visual inspection of Figure 3-6); and (3) study
authors noted limited evidence for a threshold at 56 ppb.61 In considering this information, we
conclude that the analysis reported by Jerrett indicates a relatively high degree of confidence in
the linear concentration-response function for "long-term" Os concentrations at least as low as 56
ppb, and notably decreased confidence in the linear function for concentrations at or below about
53 ppb (i.e., the upper bound of the first quartile of Os concentrations).
Based on information in the published study (Figure 1 in Jerrett et al., 2009), we
identified 72 of the 96 study cities as having ambient Os concentrations in the highest three
quartiles (Appendix 3B). As noted above, these 72 cities account for approximately 72% of the
respiratory deaths in the cohort (Table 2 in Jerrett et al. 2009). Of these 72 cities, 71 had 3-year
averages of annual 4th highest daily maximum 8-hour Os concentrations above 75 ppb (Appendix
3B). Thus, the current 8-hour NAAQS would have been violated during the study period in
virtually all of the study cities that contribute to the range of long-term Os concentrations over
which we have the greatest confidence in the reported relationship with respiratory mortality.
61 The ISA does not reach conclusions regarding the potential for a threshold in the association between "long-term"
Os concentrations and respiratory mortality.
3-76
-------
Thus, while the study by Jerrett et al. (2009) contributes to our understanding of health effects
associated with ambient Os (as summarized in section 3.1.2 above), it is less informative
regarding the extent to which those health effects may be occurring under air quality conditions
allowed by the current standard.
3.1.5 Public Health Implications
In this section, we address the public health implications of Cb exposures with respect to
the factors that put populations at increased risk from exposures (section 3.1.5.1), the size of at-
risk populations (section 3.1.5.2), and the potential effects of averting behavior on reducing Os
exposures and associated health effects (section 3.1.5.3). Providing appropriate public health
protection requires consideration of the factors that put populations at greater risk from Os
exposure. In order to estimate potential public health impacts, it is important to consider not only
the adversity of the health effects, but also the populations at greater risk and potential behaviors
that may reduce exposure.
3.1.5.1 At-Risk Populations
In this section we address the following question:
• To what extent does the currently available scientific evidence expand our
understanding of at-risk populations?
The currently available evidence expands our understanding of populations that were
identified to be at greater risk of Os-related health effects at the time of the last review (i.e.,
people who are active outdoors, people with lung disease, children and older adults and people
with increased responsiveness to Os) and supports the identification of additional factors that
may lead to increased risk (U.S. EPA, 2006, section 3.6.2; U.S. EPA, 2013, chapter 8).
Populations and lifestages may be at greater risk for Os-related health effects due to factors that
contribute to their susceptibility and/or vulnerability to ozone. The definitions of susceptibility
and vulnerability have been found to vary across studies, but in most instances "susceptibility"
refers to biological or intrinsic factors (e.g., lifestage, sex, preexisting disease/conditions) while
"vulnerability" refers to non-biological or extrinsic factors (e.g., socioeconomic status [SES])
(U.S. EPA, 2013, p. 8-1). In some cases, the terms "at-risk" and "sensitive" have been used to
encompass these concepts more generally. In the ISA and this PA, "at-risk" is the all-
encompassing term used to define groups with specific factors that increase their risk of
Os-related health effects. Further discussion of at-risk populations can be found below.
3-77
-------
There are multiple avenues by which groups may experience increased risk for Os-related
health effects. A population or lifestage62 may exhibit greater effects than other populations or
lifestages exposed to the same concentration or dose, or they may be at greater risk due to
increased exposure to an air pollutant (e.g., time spent outdoors). A group with intrinsically
increased risk would have some factor(s) that increases risk through a biological mechanism and,
in general, would have a steeper concentration-risk relationship, compared to those not in the
group. Factors that are often considered intrinsic include pre-existing asthma, genetic
background, and lifestage. A group of people could also have extrinsically increased risk, which
would be through an external, non-biological factor, such as socioeconomic status (SES) and
diet. Some groups are at risk of increased internal dose at a given exposure concentration, for
example, because of breathing patterns. This category would include people who work or
exercise outdoors. Finally, there are those who might be placed at increased risk for experiencing
greater exposures by being exposed to higher Os concentrations. This would include, for
example, groups of people with greater exposure to ambient Os due to less availability or use of
home air conditioners such that they are more likely to be in locations with open windows on
high ozone days. Some groups may be at increased risk of Os-related health effects through a
combination of factors. For example, children tend to spend more time outdoors when Os levels
are high, and at higher levels of activity than adults, which leads to increased exposure and dose,
and they also have biological, or intrinsic, risk factors (e.g., their lungs are still developing) (U.S.
EPA, 2013, Chapter 8). An at-risk population or lifestage is more likely to experience adverse
health effects related to Os exposures and/or, develop more severe effects from exposure than the
general population.
People with Specific Genetic Variants
Overall, for variants in multiple genes there is adequate evidence for involvement in
populations being more at-risk than others to the effects of Cb exposure on health (U.S. EPA,
2013, section 8.1). Controlled human exposure and epidemiologic studies have reported evidence
of Os-related increases in respiratory symptoms or decreases in lung function with variants
including GSTM1, GSTP1, HMOX1, and NQO1. NQO1 deficient mice were found to be
resistant to Os-induced AHR and inflammation, providing biological plausibility for results of
studies in humans. Additionally, studies of rodents have identified a number of other genes that
may affect Os-related health outcomes, including genes related to innate immune signaling and
pro- and anti-inflammatory genes, which have not been investigated in human studies.
People with Asthma
62 Lifestages, which in this case includes childhood and older adulthood, are experienced by most people over the
course of a lifetime, unlike other factors associated with at-risk populations.
3-78
-------
Previous Os AQCDs identified individuals with asthma as a population at increased risk
of Os-related health effects. Multiple new epidemiologic studies included in the ISA have
evaluated the potential for increased risk of Os-related health effects in people with asthma,
including: lung function; symptoms; medication use; airway hyperresponsiveness (AHR); and
airway inflammation (also measured as exhaled nitric oxide fraction, or FeNO). A study of
lifeguards in Texas reported decreased lung function with short-term Os exposure among both
individuals with and without asthma, however, the decrease was greater among those with
asthma (Thaller et al., 2008). A Mexican study of children ages 6-14 detected an association
between short-term Os exposure and wheeze, cough, and bronchodilator use among asthmatics
but not non-asthmatics, although this may have been the result of a small non-asthmatic
population (Escamilla-Nufiez et al., 2008). A study of modification by AHR (an obligate
condition among asthmatics) reported greater short-term Os-associated decreases in lung
function in elderly individuals with AHR, especially among those who were obese (Alexeeff et
al., 2007). With respect to airway inflammation, in one study, a positive association was reported
for airway inflammation among asthmatic children following short-term Os exposure, but the
observed association was similar in magnitude to that of non-asthmatics (Barraza-Villarreal et
al., 2008). Similarly, another study of children in California reported an association between Os
concentration and FeNO that persisted both among children with and without asthma as well as
those with and without respiratory allergy (Berhane et al., 2011). Finally, Khatri et al. (2009)
found no association between short-term Os exposure and altered lung function for either
asthmatic or non-asthmatic adults, but did note a decrease in lung function among individuals
with allergies.
New evidence for difference in effects among asthmatics has been observed in studies
that examined the association between Os exposure and altered lung function by asthma
medication use. A study of children with asthma living in Detroit reported a greater association
between short-term Os and lung function for corticosteroid users compared with
noncorticosteroid users (Lewis et al., 2005). Conversely, another study found decreased lung
function among noncorticosteroid users compared to users, although in this study, a large
proportion of non-users were considered to be persistent asthmatics (Hernandez-Cadena et al.,
2009). Lung function was not related to short-term Os exposure among corticosteroid users and
non-users in a study taking place during the winter months in Canada (Liu et al., 2009).
Additionally, a study of airway inflammation reported a counterintuitive inverse association with
Os of similar magnitude for all groups of corticosteroid users and non-users (Qian et al., 2009).
Controlled human exposure studies that have examined the effects of Os on adults with
asthma and healthy controls are limited. Based on studies reviewed in the 1996 and 2006 Os
AQCDs, subjects with asthma appeared to be more sensitive to acute effects of Os in terms of
3-79
-------
FEVi and inflammatory responses than healthy non-asthmatic subjects. For instance, Horstman
et al. (1995) observed that mild-to-moderate asthmatics, on average, experienced double the
Cb-induced FEVi decrement of healthy subjects (19% versus 10%, respectively, p = 0.04).
Moreover, a statistically significant positive correlation between FEVi responses to Cb exposure
and baseline lung function was observed in individuals with asthma, i.e., responses increased
with severity of disease. Minimal evidence exists suggesting that individuals with asthma have
smaller Cb-induced FEVi decrements than healthy subjects (3% versus 8%, respectively)
(Mudway et al., 2001). However, the asthmatics in that study also tended to be older than the
healthy subjects, which could partially explain their lesser response since FEVi responses to Cb
exposure diminish with age. Individuals with asthma also had significantly more neutrophils in
the BALF (18 hours postexposure) than similarly exposed healthy individuals (Peden et al.,
1997; Scannell et al., 1996; Basha et al., 1994). Furthermore, a study examining the effects of Cb
on individuals with atopic asthma and healthy controls reported that greater numbers of
neutrophils, higher levels of cytokines and hyaluronan, and greater expression of macrophage
cell-surface markers were observed in induced sputum of atopic asthmatics compared with
healthy controls (Hernandez et al., 2010). Differences in Cb-induced epithelial cytokine
expression were noted in bronchial biopsy samples from asthmatics and healthy controls (Bosson
et al., 2003). Cell-surface marker and cytokine expression results, and the presence of
hyaluronan, are consistent with Cb having greater effects on innate and adaptive immunity in
these asthmatic individuals. In addition, studies have demonstrated that Cb exposure leads to
increased bronchial reactivity to inhaled allergens in mild allergic asthmatics (Kehrl et al., 1999;
Torres et al., 1996) and to the influx of eosinophils in individuals with pre-existing allergic
disease (Vagaggini et al., 2002; Peden et al., 1995). Taken together, these results point to several
mechanistic pathways which could account for the enhanced sensitivity to Cb in subjects with
asthma (see Section 5.4.2.2 in the ISA).
Toxicological studies provide additional evidence of the biological basis for the greater
effects of Cb among those with asthma or AHR (U.S. EPA, 2013, section 8.2.2). In animal
toxicological studies, an asthmatic phenotype is modeled by allergic sensitization of the
respiratory tract. Many of the studies that provide evidence that Cb exposure is an inducer of
AHR and remodeling utilize these types of animal models. For example, a series of experiments
in infant rhesus monkeys have shown these effects, but only in monkeys sensitized to house dust
mite allergen. Similarly, adverse changes in pulmonary function were demonstrated in mice
exposed to Cb; enhanced inflammatory responses were in rats exposed to Cb, but only in animals
sensitized to allergen. In general, it is the combined effects of Cb and allergic sensitization which
result in measurable effects on pulmonary function. In a pulmonary fibrosis model, exposure Cb
for 5 days increased pulmonary inflammation and fibrosis, along with the frequency of
3-80
-------
bronchopneumonia in rats. Thus, short-term exposure to Os may enhance damage in a previously
injured lung (U.S. EPA, 2013, section 8.2.2).
In the 2006 Os AQCD, the potential for individuals with asthma to have greater risk of
Os-related health effects was supported by a number of controlled human exposure studies,
evidence from toxicological studies, and a limited number of epidemiologic studies. In section
8.2.2, the ISA reports that in the recent epidemiologic literature some, but not all, studies report
greater risk of health effects among individuals with asthma. Studies examining effect measure
modification of the relationship between short-term Os exposure and altered lung function by
corticosteroid use provided limited evidence of Os-related health effects. However, recent studies
of behavioral responses have found that studies do not take into account individual behavioral
adaptations to forecasted air pollution levels (such as avoidance and reduced time outdoors),
which may underestimate the observed associations in studies that examined the effect of Os
exposure on respiratory health (Neidell and Kinney, 2010). This could explain some
inconsistency observed among recent epidemiologic studies. The evidence from controlled
human exposure studies provides support for increased detriments in FEVi and greater
inflammatory responses to Os in individuals with asthma than in healthy individuals without a
history of asthma. The collective evidence for increased risk of Os-related health effects among
individuals with asthma from controlled human exposure studies is supported by recent
toxicological studies which provide biological plausibility for heightened risk of asthmatics to
respiratory effects due to Os exposure. Overall, the ISA finds there is adequate evidence for
asthmatics to be an at-risk population.
Children
Children are considered to be at greater risk from Os exposure because their respiratory
systems undergo lung growth until about 18-20 years of age and are therefore thought to be
intrinsically more at risk for Os-induced damage (U.S. EPA, 2006). It is generally recognized
that children spend more time outdoors than adults, and therefore would be expected to have
higher exposure to Os than adults. The ventilation rates also vary between children and adults,
particularly during moderate/heavy activity. Children aged 11 years and older and adults have
higher absolute ventilation rates than children aged 1-11 years. However, children have higher
ventilation rates relative to their lung volumes, which tends to increase dose normalized to lung
surface area. Exercise intensity has a substantial effect on ventilation rate, with high intensity
activities resulting in nearly double the ventilation rate during moderate activity among children
and those adults less than 31 years of age. For more information on time spent outdoors and
ventilation rate differences by age group, see Section 4.4.1 in the ISA (U.S. EPA, 2013).
The 1996 Os AQCD reported clinical evidence that children, adolescents, and young
adults (<18 years of age) appear, on average, to have nearly equivalent spirometric responses to
3-81
-------
Os exposure, but have greater responses than middle-aged and older adults (U.S. EPA, 1996).
Symptomatic responses (e.g., cough, shortness of breath, pain on deep inspiration) to Os
exposure, however, appear to increase with age until early adulthood and then gradually decrease
with increasing age (U.S. EPA, 1996). Complete lung growth and development is not achieved
until 18-20 years of age in women and the early 20s for men; pulmonary function is at its
maximum during this time as well.
Recent epidemiologic studies have examined different age groups and their risk to
Os-related respiratory hospital admissions and emergency department (ED) visits. Evidence for
greater risk in children was reported in several studies. A study in Cyprus of short-term Os
concentrations and respiratory hospital admissions (HA) detected possible effect measure
modification by age with a larger association among individuals < 15 years of age compared
with those > 15 years of age; the effect was apparent only with a 2-day lag (Middleton et al.,
2008). Similarly, a Canadian study of asthma-ED visits reported the strongest Os-related
associations among 5- to 14-year olds compared to the other age groups (ages examined 0-75+)
(Villeneuve et al., 2007). Greater Os-associated risk in asthma-related ED visits were also
reported among children (<15 years) as compared to adults (15 to 64 years) in a study from
Finland (Halonen et al., 2009). A study of New York City hospital admissions demonstrated an
increase in the association between Os exposure and asthma-related hospital admissions for 6- to
18-year olds compared to those < 6 years old and those > 18 years old (Silverman and Ito, 2010).
When examining long-term Os exposure and asthma HA among children, associations were
determined to be larger among children 1 to 2 years old compared to children 2 to 6 years old
(Lin et al., 2008b). A few studies reported positive associations among both children and adults
and no modification of the effect by age.
The evidence reported in epidemiologic studies is supported by recent toxicological
studies which observed Os-induced health effects in immature animals. Early life exposures of
multiple species of laboratory animals, including infant monkeys, resulted in changes in
conducting airways at the cellular, functional, ultra-structural, and morphological levels. The
studies conducted on infant monkeys are most relevant for assessing effects in children. Carey et
al. (2007) conducted a study of Os exposure in infant rhesus macaques, whose respiratory tract
closely resemble that of humans. Monkeys were exposed either acutely or in episodes designed
to mimic human exposure. All monkeys acutely exposed to Os had moderate to marked
necrotizing rhinitis, with focal regions of epithelial exfoliation, numerous infiltrating neutrophils,
and some eosinophils. The distribution, character, and severity of lesions in episodically exposed
infant monkeys were similar to that of acutely exposed animals. Neither exposure protocol for
the infant monkeys produced mucous cell metaplasia proximal to the lesions, an adaptation
observed in adult monkeys exposed in another study (Harkema et al., 1987). Functional and
3-82
-------
cellular changes in conducting airways were common manifestations of exposure to Os among
both the adult and infant monkeys (Plopper et al., 2007). In addition, the lung structure of the
conducting airways in the infant monkeys was significantly stunted by Os and this aberrant
development was persistent 6 months postexposure (Fanucchi et al., 2006).
Age may also affect the inflammatory response to Os exposure. Toxicological studies
reported that the difference in effects among younger lifestage test animals may be due to
age-related changes in antioxidants levels and sensitivity to oxidative stress. Further discussion
of these studies may be found in section 8.3.1.1 of the ISA (U.S. EPA, 2013, p. 8-18).
The previous and recent human clinical and toxicological studies reported evidence of
increased risk from Os exposure for younger ages, which provides coherence and biological
plausibility for the findings from epidemiologic studies. Although there was some inconsistency,
generally, the epidemiologic studies reported positive associations among both children and
adults or just among children. The interpretation of these studies is limited by the lack of
consistency in comparison age groups and outcomes examined. However, overall, the
epidemiologic, controlled human exposure, and toxicological studies provide adequate evidence
that children are potentially at increased risk of Os-related health effects.
Older Adults
The ISA notes that older adults are at greater risk of health effects associated with Os
exposure through a variety of intrinsic pathways (U.S. EPA, 2013, section 8.3.1.2). In addition,
older adults may differ in their exposure and internal dose. Older adults were outdoors for a
slightly longer proportion of the day than adults aged 18-64 years. Older adults also have
somewhat lower ventilation rates than adults aged 31 - less than 61 years. For more information
on time spent outdoors and ventilation rate differences by age group, see Section 4.4 in the ISA
(U.S. EPA, 2013). The gradual decline in physiological processes that occur with aging may lead
to increased risk of Os-related health effects (U.S. EPA, 2006). Respiratory symptom responses
to Os exposure appears to increase with age until early adulthood and then gradually decrease
with increasing age (U.S. EPA, 1996); lung function responses to Os exposure also decline from
early adulthood (U.S. EPA, 1996). The reductions of these responses with age may put older
adults at increased risk for continued Os exposure. In addition, older adults, in general, have a
higher prevalence of preexisting diseases compared to younger age groups and this may also lead
to increased risk of Os-related health effects (U.S. EPA, 2013, section 8.3.1.2). With the number
of older Americans increasing in upcoming years (estimated to increase from 12.4% of the U.S.
population to 19.7% between 2000 to 2030, which is approximately 35 million and 71.5 million
individuals, respectively) this group represents a large population potentially at risk of Os-related
health effects (SSDAN CensusScope, 2010; DeNavas-Walt et al., 2011).
3-83
-------
The majority of recent studies reported greater effects of short-term Os exposure and
mortality among older adults, which is consistent with the findings of the 2006 Os AQCD. A
study (Medina-Ramon and Schwartz, 2008) conducted in 48 cities across the U.S. reported larger
effects among adults > 65 years old compared to those < 65 years; further investigation of this
study population revealed a trend of Os-related mortality risk that gets larger with increasing age
starting at age 51 (Zanobetti and Schwartz, 2008a). Another study conducted in 7 urban centers
in Chile reported similar results, with greater effects in adults > 65 years old (Cakmak et al.,
2007). More recently, a study conducted in the same area reported similar associations between
Os exposure and mortality in adults aged < 64 years old and 65 to 74 years old, but the risk was
increased among older age groups (Cakmak et al., 2011). A study performed in China reported
greater effects in populations > 45 years old (compared to 5 to 44 year olds), with statistically
significant effects present only among those > 65 years old (Kan et al., 2008). An Italian study
reported higher risk of all-cause mortality associated with increased Os concentrations among
individuals >85 year old as compared to those 35 to 84 years old (Stafoggia et al., 2010). The Air
Pollution and Health: A European and North American Approach (APHENA) project examined
the association between Os exposure and mortality for those <75 and > 75 years of age. In
Canada, the associations for all-cause and cardiovascular mortality were greater among those
>75 years old. In the U.S., the association for all-cause mortality was slightly greater for those
<75 years of age compared to those >75 years old in summer-only analyses. No consistent
pattern was observed for CVD mortality. In Europe, slightly larger associations for all-cause
mortality were observed in those <75 years old in all-year and summer-only analyses. Larger
associations were reported among those <75 years for CVD mortality in all-year analyses, but
the reverse was true for summer-only analyses (Katsouyanni et al., 2009).
With respect to epidemiologic studies of Os exposure and hospital admissions, a positive
association was reported between short-term Os exposure and respiratory hospital admissions for
adults > 65 years old but not for those adults aged 15 to 64 years (Halonen et al., 2009). In the
same study, no association was observed between Os concentration and respiratory mortality
among those > 65 years old or those 15 to 64 years old. No modification by age (40 to 64 year
olds versus > 64 year olds) was observed in a study from Brazil examining Os levels and COPD
ED visits.
Although some outcomes reported mixed findings regarding an increase in risk for older
adults, recent epidemiologic studies report consistent positive associations between short-term
Os exposure and mortality in older adults. The evidence from mortality studies is consistent with
the results reported in the 2006 Os AQCD and is supported by toxicological studies providing
biological plausibility for increased risk of effects in older adults. Also, older adults may be
experiencing increased exposure compared to younger adults. Overall, the ISA (U.S. EPA, 2013)
3-84
-------
concludes adequate evidence is available indicating that older adults are at increased risk of
Os-related health effects.
People with Diets Lower in Vitamins C and E
Diet was not examined as a factor potentially affecting risk in previous Os AQCDs, but
recent studies have examined modification of the association between Os and health effects by
dietary factors. Because Os mediates some of its toxic effects through oxidative stress, the
antioxidant status of an individual is an important factor that may contribute to increased risk of
Os-related health effects. Supplementation with vitamins C and E has been investigated in a
number of studies as a means of inhibiting Os-mediated damage.
Two epidemiologic studies have examined effect measure modification by diet and found
evidence that certain dietary components are related to the effect Os has on respiratory outcomes.
In one recent study the effects of fruit/vegetable intake and Mediterranean diet were examined.
Increases in these food patterns, which have been noted for their high vitamins C and E and
omega-3 fatty acid content, were positively related to lung function in asthmatic children living
in Mexico City, and modified by Os exposure (Romieu et al., 2009). Another study examined
supplementation of the diets of asthmatic children in Mexico with vitamins C and E (Sienra-
Monge et al., 2004). Associations were detected between short-term Os exposure and nasal
airway inflammation among children in the placebo group but not in those receiving the
supplementation.
The epidemiologic evidence is supported by controlled human exposure studies,
discussed in section 8.4.1 of the ISA (U.S. EPA, 2013), that have shown that the first line of
defense against oxidative stress is antioxidants-rich extracellular lining fluid (ELF) which
scavenge free radicals and limit lipid peroxidation. Exposure to Os depletes antioxidant levels in
nasal ELF probably due to scrubbing of Os; however, the concentration and the activity of
antioxidant enzymes either in ELF or plasma do not appear to be related to Os responsiveness.
Controlled studies of dietary antioxidant supplementation have demonstrated some protective
effects of a-tocopherol (a form of vitamin E) and ascorbate (vitamin C) on spirometric measures
of lung function after Os exposure but not on the intensity of subjective symptoms and
inflammatory responses. Dietary antioxidants have also afforded partial protection to asthmatics
by attenuating postexposure bronchial hyperresponsiveness. Toxicological studies discussed in
section 8.4.1 of the ISA (U.S. EPA, 2013) provide evidence of biological plausibility to the
epidemiologic and controlled human exposure studies.
There is adequate evidence that individuals with diets lower in vitamins C and E are at
risk for Os-related health effects. The evidence from epidemiologic studies is supported by
controlled human exposure and toxicological studies.
3-85
-------
Outdoor Workers
Studies included in the 2006 Os AQCD reported that individuals who participate in
outdoor activities or work outside to be a population at increased risk based on consistently
reported associations between Os exposure and respiratory health outcomes in these groups (U.S.
EPA, 2006). Outdoor workers are exposed to ambient Os concentrations for a greater period of
time than individuals who spend their days indoors. As discussed in Section 4.7 of the ISA (U.S.
EPA, 2013) outdoor workers sampled during the work shift had a higher ratio of personal
exposure to fixed-site monitor concentrations than health clinic workers who spent most of their
time indoors. Additionally, an increase in dose to the lower airways is possible during outdoor
exercise due to both increases in the amount of air breathed (i.e., minute ventilation) and a shift
from nasal to oronasal breathing. The association between FEVi responses to Os exposure and
minute ventilation is discussed more fully in Section 6.2.3.1 of the 2006 Os AQCD.
Previous studies have shown that increased exposure to Os due to outdoor work leads to
increased risk of Os-related health effects, specifically decrements in lung function (U.S. EPA,
2006). The strong evidence from the 2006 Os AQCD which demonstrated increased exposure,
dose, and ultimately risk of Os-related health effects in this population supports the conclusion
that there is adequate evidence to indicate that increased exposure to Os through outdoor work
increases the risk of Os-related health effects.
In some cases, it is difficult to determine a factor that results in increased risk of effects.
For example, previous assessments have included controlled human exposure studies in which
some healthy individuals demonstrate greater Os-related health effects compared to other healthy
individuals. Intersubject variability has been observed for lung function decrements,
symptomatic responses, pulmonary inflammation, AHR, and altered epithelial permeability in
healthy adults exposed to Os and these results tend to be reproducible within a given individual
over a period of several months indicating differences in the intrinsic responsiveness. In many
cases the reasons for the variability is not clear. This may be because one or some of the factors
described above have not been evaluated in studies, or it may be that additional, unidentified
factors influence individual responses to Os (U.S. EPA, 2013, section 8.5).
As discussed in chapter 8 of the ISA the challenges and limitations in evaluating the
factors that can increase risk for experiencing Os-related health effects may contribute to a lack
of information about the factors that may increase risk from Os exposures. This lack of
information may contribute to conclusions that evidence for some factors, such as sex, SES, and
obesity provided "suggestive" evidence of increased risk, or that for a number of factors the
evidence was inadequate to draw conclusions about potential increase in risk of effects. Overall,
the factors for which the ISA concludes there is adequate evidence of increased risk for
3-86
-------
experiencing Os-related effects were related to asthma, lifestage (children and older adults),
genetic variability, dietary factors, and working outdoors.
3.1.5.2 Size of At-Risk Populations and Lifestages in the United States
One consideration in the assessment of potential public health impacts is the size of
various population groups for which there is adequate evidence of increased risk for health
effects associated with Os-related air pollution exposure. The factors for which the ISA judged
the evidence to be "adequate" with respect to contributing to increased risk of Os-related effects
among various populations and lifestages included: asthma; childhood and older adulthood; diets
lower in vitamins C and E; certain genetic variants and, working outdoors (U.S. EPA, 2013,
section 8.5).
With regard to asthma, Table 3-7 below summarizes information on the prevalence of
current asthma by age in the U.S. adult population in 2010 (Schiller et al., 2012; children -
Bloom et al., 2011). Individuals with current asthma constitute a fairly large proportion of the
population, including more than 25 million people. Asthma prevalence tends to be higher in
children than adults.
Within the U.S., approximately 8.2% of adults have reported currently having asthma
(Schiller et al., 2012) and 9.5% of children have reported currently having asthma (Bloom et al.,
2011). Table 3-12 below provides more detailed information on prevalence of asthma by age in
the U.S.
3-87
-------
Table 3-7. Prevalence of asthma by age in the U.S.
Age (years)
0-4
5-11
12-17
18-44
45-64
65-74
75+
Asthma prevalence is reported for "still
Source: Statistics for adults: Schiller et
N (in thousands)
1,285
3,020
2,672
8,902
6,704
1,849
1,279
has asthma"
al. (2012); Statistics for children:
Percent
6.0
10.5
10.9
8.1
8.4
8.7
7.4
Bloom etal. (2011)
With regard to lifestages, based on U.S. census data from 2010 (Howden and Meyer,
2011), about 74 million people, or 24% of the U.S. population, are under 18 years of age and
more than 40 million people, or about 13% of the U.S. population, are 65 years of age or older.
Hence, a large proportion of the U.S. population, more than 33%, is included in age groups that
are considered likely to be at increased risk for health effects from ambient Os exposure.
With regard to dietary factors, no statistics are available to estimate the size of an at-risk
population based on nutritional status.
With regard to outdoor workers, in 2010 approximately 11.7% of the total number of
people (143 million people) employed, or about 16.8 million people, worked outdoors one or
more day per week (based on worker surveys).63 Of these approximately 7.4% of the workforce,
or about 7.8 million people, worked outdoors three or more days per week.
The health statistics data illustrate what is known as the "pyramid" of effects. At the top
of the pyramid, there are approximately 2.5 million deaths from all causes per year in the U.S.
population, with about 250 thousand respiratory-related deaths (CDC-WONDER64). For
respiratory health diseases, there are nearly 3.3 million hospital discharges per year (HCUP65),
63 The O*NET program is the nation's primary source of occupational information. Central to the project is the
O*NET database, containing information on hundreds of standardized and occupation-specific descriptors. The
database, which is available to the public at no cost, is continually updated by surveying a broad range of workers
from each occupation, http://www.onetcenter.org/overview.html
http://www.onetonline.org/find/descriptor/browse/Work_Context/4.C.2/
64 http://wonder.cdc.gov/
65 http://www.hcup-us.ahrq.gov/
3-88
-------
8.7 million respiratory ED visits (HCUP, 2007), 112 million ambulatory care visits (Woodwell
and Cherry, 2004), and an estimated 700 million restricted activity days per year due to
respiratory conditions (Adams et al., 1999). Combining small risk estimates with relatively large
baseline levels of health outcomes can result in quite large public health impacts. Thus, even a
small percentage reduction in Os health impacts on cardiopulmonary diseases would reflect a
large number of avoided cases.
3.1.5.3 Averting Behavior
The activity pattern of individuals is an important determinant of their exposure (ISA,
U.S. EPA, 2013, section 4.4.1). Variation in Os concentrations among various
microenvironments means that the amount of time spent in each location, as well as the level of
activity, will influence an individual's exposure to ambient Os. Activity patterns vary both
among and within individuals, resulting in corresponding variations in exposure across a
population and over time. Individuals can reduce their exposure to Os by altering their behaviors,
such as by staying indoors, being active outdoors when air quality is better, and by reducing their
activity levels or reducing the time being active outdoors on high-Os days (U.S. EPA, 2013,
section 4.4.2). The evidence in this topic area, while not addressed in the 2006 AQCD, is
evaluated in the ISA for this review.
The widely reported Air Quality Index (AQI) conveys advice to the public, and
particularly at-risk populations, on reducing exposure on days when ambient levels of common
air pollutants are elevated (www.airnow.gov). The AQI describes the potential for health effects
from Os (and other individual pollutants) in six color-coded categories of air-quality, ranging
from Good (green), Moderate (yellow), Unhealthy for Sensitive Groups (orange), Unhealthy
(red), and Very Unhealthy (purple), and Hazardous (maroon). Levels in the unhealthy ranges
(i.e., Unhealthy for Sensitive Groups and above) come with recommendations about reducing
exposure. Forecasted and actual AQI values for Os are reported to the public during the Os
season. The AQI advisories explicitly state that children, older adults, people with lung disease,
and people who are active outdoors, may be at greater risk from exposure to Os. People are
advised to reduce exposure depending on the predicted Os levels and the likelihood of risk. This
advice includes being active outdoors when air quality is better, and reducing activity levels or
reducing the time being active outdoors on high-Os days. Staying indoors to reduce exposure is
not recommended until air quality reaches the Very Unhealthy or Hazardous categories.
Evidence of individual averting behaviors in response to AQI advisories has been found
in several studies, including activity pattern and epidemiologic studies, especially for the at-risk
populations, such as children, older adults, and people with asthma, who are targeted by the
advisories. Such effects are less pronounced in the general population, possibly due to the
opportunity cost of behavior modification. Epidemiologic evidence from a study (Neidell and
3-89
-------
Kinney, 2010) conducted in the 1990's in Los Angeles, CA reports increased asthma hospital
admissions among children and older adults when Os alert days (1-hour max Os concentration
>200 ppb) were excluded from the analysis of daily hospital admissions and Os concentrations
(presumably thereby eliminating averting behavior based on high Os forecasts). The lower rate of
admissions observed when alert days were included in the analysis suggests that estimates of
health effects based on concentration-response functions that do not account for averting
behavior may be biased towards the null (U.S. EPA, 2013, section 4.4.2).
3.2 AIR QUALITY-, EXPOSURE-, AND RISK-BASED CONSIDERATIONS
In order to inform judgments about the public health impacts of Os-related health effects,
the HREA has developed and applied models to estimate human exposures to Cb and Os-
associated health risks across the United States, with a specific focus on urban case study areas
(U.S. EPA, 2014).66 The HREA uses photochemical modeling to adjust air quality from the
2006-2010 Os seasons to just meet the current and alternative standards for the 2006-2008 and
2008-2010 periods.67 In this section, staff considers estimates of short-term Os exposures and
estimates of health risks associated with short- and long-term Os exposures, for air quality
adjusted to just meet the current Os standard. In section 3.2.1, we consider the implications for
exposure and risk estimates of the approach used in the HREA to adjust air quality. Sections
3.2.2 and 3.2.3 discuss our exposure-based and risk-based considerations, respectively. In these
sections we specifically consider the following question:
• What are the nature and magnitude of Os exposures and health risks remaining
upon adjusting recent air quality to just meet the current Os standard, and what are
the important uncertainties associated with those exposure and risk estimates?
3.2.1 Consideration of the Adjusted Air Quality Used in Exposure and Risk
Assessments
In the first draft HREA for this review, as in the last review, the EPA relied upon
quadratic rollback to adjust hourly Os concentrations in urban case study areas to just meet the
current Os standard (U.S. EPA, 2014). Although the quadratic rollback method reproduces
66 The 15 urban case study areas analyzed for exposures are Atlanta, Baltimore, Boston, Chicago, Cleveland, Dallas,
Denver, Detroit, Houston, Los Angeles, New York, Philadelphia, Sacramento, St. Louis, and Washington, DC.
Morbidity and mortality risk estimates are presented for these same areas, with the exception of Chicago, Dallas,
and Washington, DC. The HREA also presents a national scale mortality risk assessment for unadjusted (recent) air
quality. This national-scale assessment, which focuses on existing air quality conditions and does not estimate the
health risks associated with just meeting the current or alternative standards, can provide perspective on the
relationship between national-scale Os public health impacts and impacts estimated in specific urban areas.
67 Three-year periods are used recognizing that the current standard is the average across three years of the annual
fourth-highest daily maximum 8-hour average concentration.
3-90
-------
historical patterns of air quality changes better than some alternative methods, it relies on
statistical relationships without explicitly accounting for atmospheric chemistry and precursor
emissions (U.S. EPA, 2014, Chapter 4). An important drawback of the quadratic rollback
approach, recognized in the first draft HREA (U.S. EPA, 2012b), is that it forces all monitors in
an assessment area to exhibit the same response when air quality is adjusted. It does not allow for
the spatial or temporal heterogeneity in responses that result from the non-linear atmospheric
chemistry that influences ambient Os concentrations (U.S. EPA, 2014, Chapter 4). Because
quadratic rollback does not account for physical and chemical atmospheric processes, or the
sources of emissions precursors that lead to Os formation, a backstop or "floor" must be used
when applying quadratic rollback to just meet current or alternative standards to ensure that
estimated Os is not reduced in a manner inconsistent with Os chemistry, such as to reduce
concentrations below that associated with background sources (U.S. EPA, 2014, Chapter 4).
Consistent with recommendations from the National Research Council of the National
Academies (NRC, 2008), the HREA uses a photochemical model to estimate sensitivities of Os
to changes in precursor emissions, in order to estimate ambient Os concentrations that would just
meet the current and alternative standards (U.S. EPA, 2014, Chapter 4).68 For the urban case
study areas evaluated in the HREA, this model-based adjustment approach was set up to estimate
hourly Os concentrations at each monitor location when modeled U.S. anthropogenic precursor
emissions (i.e., NOx, VOC)69 were reduced to estimate air quality that just meets the current and
alternative Os standards.70
As discussed in Chapter 4 of the HREA (U.S. EPA, 2014), this approach models the
physical and chemical atmospheric processes that influence ambient Os concentrations.
Compared to the quadratic rollback approach, it provides more realistic estimates of the spatial
and temporal responses of Os to reductions in precursor emissions. These improved estimates
avoid many of the limitations inherent in the quadratic rollback method, including the
requirement that all monitors in an assessment area exhibit the same response upon air quality
The HREA uses the CMAQ photochemical model instrumented with the higher order direct decoupled method
(HDDM) to estimate ozone concentrations that would occur with the achievement of the current and alternative Os
standards (U.S. EPA, 2014, Chapter 4).
69 Exposure and risk analyses for most urban case study areas focus on reducing NOX emissions alone (NOX
emissions were reduced by about 10 to 85% for the current standard, and up to about 95% for alternatives). In most
of the urban case study areas, reducing VOC emissions did not alter the NOx emissions reductions required to just
meet the current or alternative standards. However, in Chicago and Denver, reductions in VOC emissions allowed
for smaller NOx emissions reductions. Therefore, exposure and risk analyses for Chicago and Denver are based on
reductions in emissions of both NOx and VOC (U.S. EPA, 2014, section 4.3.3.1, Table 4-3).
70 Although this chapter focuses on the current standard, our overarching considerations regarding adjusted air
quality also apply to alternative standards simulated in the HREA. Alternative standards are discussed in chapter 4
of this PA.
3-91
-------
adjustment to the current and/or alternative standards. Because adjusted air quality scenarios are
based on reducing only U.S. anthropogenic emissions, this approach also does not require the
specification of background concentrations as a rollback "floor" (U.S. EPA, 2014, section 4.3.3).
The use of this model-based air quality adjustment approach in the HREA has important
implications for the patterns of ambient Os concentrations estimated in urban case study areas.
Specifically, in locations and time periods when NOx is predominantly contributing to Os
formation (e.g., downwind of important NOx sources, where the highest Os concentrations often
occur), model-based adjustment to the current and alternative standards decreases estimated
ambient Os concentrations compared to recent monitored concentrations (U.S. EPA, 2014,
section 4.3.3.2). In contrast, in locations and time periods when NOx is predominantly
contributing to Os titration (e.g., in urban centers with high concentrations of NOx emissions,
where ambient Os concentrations are often suppressed and thus relatively low71), model-based
adjustment increases ambient Os concentrations compared to recent measured concentrations
(U.S. EPA, 2014, section 4.3.3.2) (Chapter 2, above).
Within urban case study areas, the overall impacts of model-based air quality adjustment
are to reduce relatively high ambient Os concentrations (i.e., concentrations at the upper ends of
ambient distributions) and to increase relatively low Os concentrations (i.e., concentrations at the
lower ends of ambient distributions) (U.S. EPA, 2014, section 4.3.3.2, Figures 4-9 and 4-10).
Seasonal means of daily concentrations generally exhibit only modest changes upon air quality
adjustment, reflecting the seasonal balance between daily decreases and increases in ambient
concentrations (U.S. EPA, 2014, Figures 4-9 and 4-10). The resulting compression in
distributions of ambient Os concentrations is evident in all of the urban case study areas
evaluated, though the degree of compression varies considerably across areas (U.S. EPA, 2014,
Figures 4-9 and 4-10).
Adjusted patterns of Os air quality have important implications for exposure and risk
estimates in urban case study areas. Estimates influenced largely by the upper ends of the
distribution of ambient concentrations (i.e., exposures of concern and lung function risk
estimates, as discussed in sections 3.2.2 and 3.2.3.1 below) will decrease with model-adjustment
to the current and alternative standards. In contrast, seasonal risk estimates influenced by the full
distribution of ambient Os concentrations (i.e., epidemiology-based risk estimates, as discussed
in section 3.2.3.2 below) either increase or decrease in response to air quality adjustment,
71 Titration is also prominent during time periods when photochemistry is limited, such as at night and on cool,
cloudy days.
3-92
-------
depending on the balance between the daily decreases in high Os concentrations and increases in
low Os concentrations.72
We further consider the implications of the spatial and temporal patterns of adjusted air
quality within the context of exposure (section 3.2.2) and risk (section 3.2.3) estimates for Os
concentrations adjusted to just meet the current standard. As discussed below (section 3.2.3.2),
these altered patterns are particularly important to consider when interpreting epidemiology-
based risk estimates.
3.2.2 Exposure-Based Considerations
The exposure assessment presented in the HREA (U.S. EPA, 2014, Chapter 5) provides
estimates of the number of people exposed to various concentrations of ambient Os, while at
specified exertion levels. The HREA estimates exposures in 15 urban case study areas for
school-age children (ages 5 to 18), asthmatic school-age children, asthmatic adults, and older
adults, reflecting the strong evidence indicating that these populations are potentially at increased
risk for Os-attributable effects (EPA, 2013, Chapter 8; section 3.1.2, above). An important
purpose of these exposure estimates is to provide perspective on the extent to which air quality
adjusted to just meet the current Os NAAQS could be associated with exposures to Cb
concentrations reported to result in respiratory effects.73 Estimates of such "exposures of
concern" provide perspective on the potential public health impacts of Os-related effects,
including for effects that cannot currently be evaluated in a quantitative risk assessment (e.g.,
airway inflammation).
In the absence of large scale exposure studies that encompass the general population, as
well as at-risk populations, modeling is the preferred approach to estimating exposures to Cb.
The use of exposure modeling also facilitates the estimation of exposures resulting from ambient
air concentrations differing from those in exposure studies (e.g., concentrations just meeting the
current standard). In the HREA, population exposures to ambient Os concentrations are
estimated using the current version of the Air Pollutants Exposure (APEX) model. The APEX
model simulates the movement of individuals through time and space and estimates their
exposures to a given pollutant in indoor, outdoor, and in-vehicle microenvironments (U.S. EPA,
2014, section 5.1.3). APEX takes into account the most critical factors that contribute to total
72 In addition, because epidemiology-based risk estimates use "area-wide" average Os concentrations, calculated by
averaging concentrations across multiple monitors in urban case study areas (section 3.2.3.2 below), risk estimates
on a given day depend on the daily balance between increasing and decreasing Os concentrations at individual
monitors.
73In addition, the range of modeled personal exposures to ambient Os provide an essential input to the portion of the
health risk assessment based on exposure-response functions (for lung function decrements) from controlled human
exposure studies. The health risk assessment based on exposure-response information is discussed in section 3.2.3,
below.
3-93
-------
human exposure to ambient Os, including the temporal and spatial distributions of people and Os
concentrations throughout an urban area, the variation of Os concentrations within various
microenvironments, and the effects of exertion on breathing rate in exposed individuals (U.S.
EPA, 2014, section 5.1.3). To the extent spatial and/or temporal patterns of ambient Os
concentrations are altered upon air quality adjustment, as discussed above, exposure estimates
reflect population exposures to those altered patterns.
The HREA estimates 8-hour exposures at or above benchmark concentrations of 60, 70,
and 80 ppb for individuals engaged in moderate or greater exertion. Benchmarks reflect exposure
concentrations at which Os-induced respiratory effects are known to occur in some healthy adults
engaged in moderate, intermittent exertion, based on evidence from controlled human exposure
studies (section 3.1.2.1 above and U.S. EPA, 2013, section 6.2). The amount of weight to place
on the estimates of exposures at or above specific benchmark concentrations depends in part on
the weight of the scientific evidence concerning health effects associated with Os exposures at
that concentration. It also depends on judgments about the importance, from a public health
perspective, of the health effects that are known or can reasonably be inferred to occur as a result
of exposures at benchmark concentrations (sections 3.1.3, 3.1.5 above).
As discussed in more detail above (section 3.1.2.1), the health evidence that supports
evaluating exposures of concern at or above benchmark concentrations of 60, 70, and 80 ppb
comes from a large body of controlled human exposure studies reporting a variety of respiratory
effects in healthy adults. The lowest Os exposure concentration for which controlled human
exposure studies have reported respiratory effects in healthy adults is 60 ppb, with more
evidence supporting this benchmark concentration in the current review than in the last review.
In healthy adults, exposures to 60 ppb Os have been reported to decrease lung function and to
increase airway inflammation. Exposures of healthy adults to 70 ppb Os have been reported to
result in larger lung function decrements, compared to 60 ppb, as well as in increased respiratory
symptoms.74 Exposures of healthy adults to 80 ppb Os have been reported to result in larger lung
function decrements than following exposures to 60 or 70 ppb, increased airway inflammation,
increased respiratory symptoms, increased airways responsiveness, and decreased lung host
defense (section 3.1.2.1, above). As discussed above (section 3.1.3), respiratory effects reported
following exposures to Os concentrations of 60, 70, or 80 ppb meet ATS criteria for adverse
effects, result in effects judged important by CASAC in past reviews, and/or could contribute to
the clearly adverse effects reported in epidemiologic studies evaluating broader populations.
Compared to the healthy individuals included in the studies that provided the basis for the
74 As noted above, for the 70 ppb exposure concentration Schelegle et al. (2009) reported that the actual mean
exposure concentration was 72 ppb.
3-94
-------
benchmarks, at-risk populations (e.g., asthmatics, children) are more likely to experience larger
and/or more serious effects (e.g., U.S. EPA, 2013, p. 6-21).
In considering estimates of Os exposures of concern at or above benchmarks of 60, 70,
and 80 ppb, within the context of the adequacy of the current standard, we first address the
following specific question:
• What are the nature and magnitude of the short-term Os exposures of concern
remaining upon adjustment of air quality to just meet the current Os standard?
In addressing this question, we focus on modeled exposures for school-age children (ages
5-18) and asthmatic school-age children, two of the at-risk populations identified in the ISA
(section 3.1.5 above). The percentages of children estimated to experience exposures of concern
are larger than the percentages estimated for adult populations (i.e., approximately 3-fold larger
across cities) (U.S. EPA, 2014, sections 5.3.2, 5.3.3 and Figures 5-5 to 5-8). The larger exposure
estimates for children are due primarily to the larger percentage of children estimated to spend an
extended period of time being physically active outdoors when Os concentrations are elevated
(U.S. EPA, 2014, sections 5.3.2 and 5.4.1).
Although exposure estimates differ between children and adults, the patterns of results
across the cities and years are similar among all of the populations evaluated (U.S. EPA, 2014,
Figures 5-5 to 5-8). Therefore, while we highlight estimates in children, we also note that the
patterns of exposures estimated for children represent the patterns estimated for adult asthmatics
and older adults.
Key results for children are summarized below for air quality adjusted to simulate just
meeting the current Cb NAAQS (Figures 3-7 to 3-10).75 Estimates for all children and asthmatic
children are virtually indistinguishable (U.S. EPA, 2014, section 5.3.2). The estimates presented
in Figures 3-7 to 3-10 below reflect consistent reductions in estimated exposures of concern
across urban case study areas, relative to recent (i.e., unadjusted) air quality (U.S. EPA, 2014,
Appendix 5F). When averaged over the years evaluated in the UREA, reductions of up to about
70% were estimated, compared to recent air quality. These reductions in estimated exposures of
concern, relative to unadjusted air quality, reflect the consistent reductions in the highest ambient
Os concentrations upon air quality adjustment to just meet the current standard (section 3.2.1
above; U.S. EPA, 2014, Chapter 4). Such reductions in estimated exposures of concern are
evident throughout urban case study areas, including in urban cores and in surrounding areas
(U.S. EPA, 2014, section 9.6, Appendix 9A). Figures 3-7 (Average over years) and 3-8 (Worst-
Case Years) present estimates of one or more exposures of concern. Figures 3-9 (Average over
years) and 3-10 (Worst-Case Years) present estimates of two or more exposures of concern.
75 Figures 3-7 and 3-8 present estimates of one or more exposures of concern. Figures 3-9 and 3-10 present estimates
of two or more exposures of concern.
3-95
-------
:_
o
;_
o
CU
'-
o
{•
.^^
^«-
5
c
0!
._
^
—
f -1
-D
0
60
ffi
o3
<^
w
c
o
'•M
fT5
^^
C
0)
u
c
o
u
\^f
_£.
CD
E
(j
c
Ol
_C:
V)
^
CD
0)
>-
_
OJ
>
0
30
25
20
15
10
5
01
Q_
fT3
TO
0
• 60ppb
• 70 ppb
80 ppb
Figure 3-7. Percent of children estimated to experience one or more exposures of concern at or above 60, 70, 80 ppb with air
quality adjusted to just meet the current standard - Averaged Over 2006 to 2010.
5-96
-------
$* f aP .j* ^ JF ^ ^ £
^ ^ ^ ^° sr cr ^ ^ J?
60ppb
70ppb
80 ppb
Figure 3-8. Percent of children estimated to experience one or more exposures of concern at or above 60, 70, 80 ppb with air
quality adjusted to just meet the current standard - Worst-Case Year from 2006 to 2010.
5-97
-------
:
-^ O!
I § 10
o;
o
-Q
TO
TO
0
#
eoppb
70ppb
SOppb
Figure 3-9. Percent of children estimated to experience two or more exposures of concern at or above 60, 70, 80 ppb with air
quality adjusted to just meet the current standard - Averaged Over 2006 to 2010.
5-98
-------
W I
QJ +J
s_ in
U
a*
X in
OJ C
a; o
s- £
o E
E £
!_ =
o g S2
O c ™
5 O «
|S o >"
O
M
•r re re
5 E u
= ^
QJ (J
||
•g j
OJ -"
U re
•^ ""Sk
o; -^
o_ re
_^:
s_
30
25
20
15
10
0
60ppb
70 ppb
SOppb
^V
-------
Based on Figures 3-7 to 3-10 and the associated details described in the HREA (U.S.
EPA, 2014, Chapter 5), we take note of the following with regard to exposures that are estimated
to be allowed by the current standard:
1. For exposures of concern at or above 60 ppb:
a. On average over the years 2006 to 2010, the current standard is estimated to allow
approximately 10 to 17% of children in urban case study areas to experience one or
more exposures of concern at or above 60 ppb. Summing across urban case study
areas, these percentages correspond to almost 2.5 million children experiencing
approximately 4 million exposures of concern at or above 60 ppb during a single Cb
season. Of these children, almost 250,000 are asthmatics.
b. On average over the years 2006 to 2010, the current standard is estimated to allow
approximately 3 to 8% of children in urban case study areas to experience two or
more exposures of concern to Os concentrations at or above 60 ppb. Summing across
the urban case study areas, these percentages correspond to almost 900,000 children
(including about 90,000 asthmatic children) estimated to experience at least two Os
exposure concentrations at or above 60 ppb during a single Os season.
c. In the worst-case years (i.e., those with the largest exposure estimates), the current
standard is estimated to allow approximately 10 to 26% of children to experience one
or more exposures of concern at or above 60 ppb, and approximately 4 to 14% to
experience two or more exposures of concern at or above 60 ppb.
2. For exposures of concern at or above 70 ppb:
a. On average over the years 2006 to 2010, the current standard is estimated to allow up
to approximately 3% of children in urban case study areas to experience one or more
exposures of concern at or above 70 ppb. Summing across urban case study areas,
more than 350,000 children (including about 40,000 asthmatic children) are estimated
to experience Os exposure concentrations at or above 70 ppb during a single Os
season.
b. On average over the years 2006 to 2010, the current standard is estimated to allow
less than 1% of children in urban case study areas to experience two or more
exposures of concern to Os concentrations at or above 70 ppb.
c. In the worst-case years, the current standard is estimated to allow approximately 1 to
8% of children to experience one or more exposures of concern at or above 70 ppb,
and up to approximately 2% to experience two or more exposures of concern, at or
above 70 ppb.
3. For exposures of concern at or above 80 ppb: The current standard is estimated to allow
about 1% or fewer children in urban case study areas to experience exposures of concern at
or above 80 ppb, even in years with the highest exposure estimates.
3-100
-------
In further evaluating estimated exposures of concern from the HREA, we next consider
the following question:
• What are the important sources of uncertainty associated with exposure estimates?
Due to variability in responsiveness, only a subset of individuals who experience
exposures at or above a benchmark concentration can be expected to experience health effects.
Given the lack of sufficient exposure-response information for most of the health effects that
informed benchmark concentrations, estimates of the number of people likely to experience
exposures at or above benchmark concentrations generally cannot be translated into quantitative
estimates of the number of people likely to experience specific health effects.76 We view health-
relevant exposures as a continuum with greater confidence and certainty about the existence of
adverse health effects at higher Os exposure concentrations, and less confidence and greater
uncertainty as one considers lower exposure concentrations. This view draws from the overall
body of available health evidence, which indicates that as exposure concentrations increase the
incidence, magnitude, and severity of effects increases.
Though we have less confidence in the likelihood of adverse health effects as Os
exposure concentrations decrease, we also note that the controlled human exposure studies that
provided the basis for health benchmark concentrations have not evaluated at-risk populations.
Compared to the healthy individuals included in controlled human exposure studies, members of
at-risk populations (e.g., asthmatics, children) could be more likely to experience adverse effects,
could experience larger and/or more serious effects, and/or could experience effects following
exposures to lower Os concentrations. In considering estimated exposures of concern within the
context of drawing conclusions on the adequacy of the current standard (section 3.4, below), we
balance concerns about the potential for adverse health effects, including effects in at-risk
populations, with our increasing uncertainty regarding the likelihood of such effects following
exposures to lower Os concentrations.
Uncertainties associated with the APEX exposure modeling also have the potential to be
important in our consideration of the adequacy of the current standard (U.S. EPA, 2014, section
5.5.2, Table 5-10). For example, the HREA concludes that exposures of concern could be
underestimated for some individuals who are frequently and routinely active outdoors during the
warm season (U.S. EPA, section 5.5.2). This could include outdoor workers and children who
are frequently active outdoors. The HREA specifically notes that long-term diary profiles (i.e.,
monthly, annual) do not exist for such populations, limiting the extent to which APEX outputs
reflect people who follow similar daily routines resulting in high exposures, over extended
76 The exception to this is lung function decrements, as discussed below (section 3.2.3.1).
3-101
-------
periods of time. Thus, exposure estimates generated from the general pool of available diary
profiles likely do not reflect the most highly exposed individuals in the population.
In order to evaluate the potential implications of this uncertainly for exposure estimates,
the HREA reports the results of limited sensitivity analyses using subsets of diaries specifically
selected to reflect groups spending a larger proportion of time being active outdoors during the
Os season. When diaries were selected to mimic exposures that could be experienced by outdoor
workers, the percent of modeled individuals estimated to experience exposures of concern
increased compared to other adult populations evaluated. The percent of outdoor workers
estimated to experience exposures of concern were generally similar to the percentages estimated
for children (i.e., using the full database of diary profiles) in the worst-case cities and years (i.e.,
cities and years with the highest exposure estimates) (U.S. EPA, 2014, section 5.4.3.2, Figure 5-
14). In addition, when diaries were restricted to children who did not report any time spent inside
a school or performing paid work (i.e., to mimic children spending particularly large portions of
their time outdoors during the summer), the number experiencing exposures of concern increased
by approximately 30% (U.S. EPA, 2014, section 5.4.3.1). Though these sensitivity analyses are
limited to single urban case study areas, and though there is uncertainty associated with diary
selection approaches to mimic highly exposed populations, they suggest the possibility that some
at-risk groups could experience more frequent exposures of concern than indicated by estimates
based on the full database of activity diary profiles.
In further considering activity diaries, the HREA also notes growing evidence indicating
that people can change their behavior in response to high Os concentrations, reducing the time
spent being active outdoors (U.S. EPA, 2014, section 5.4.3.3). Commonly termed "averting
behaviors," these altered activity patterns could reduce personal exposure concentrations.
Therefore, the HREA also performed limited sensitivity analyses to evaluate the potential
implications of averting behavior for estimated exposures of concern. These analyses suggest
that averting behavior could reduce the percentages of children estimated to experience
exposures of concern at or above the 60 or 70 ppb benchmark concentrations by approximately
10 to 30%, with larger reductions possible for the 80 ppb benchmark (U.S. EPA, 2014, Figure 5-
15). As discussed above for other sensitivity analyses, these analyses are limited to a single
urban case study area and are subject to uncertainties associated with assumptions about the
prevalence and duration of averting behaviors. However, the results suggest that exposures of
concern could be overestimated, particularly in children (Neidell et al., 2009; U.S. EPA, 2013,
Figures 4-7 and 4-8), if the possibility for averting behavior is not incorporated into estimates.
3-102
-------
3.2.3 Risk-Based Considerations
For some health endpoints, there is sufficient scientific evidence and information
available to support the development of quantitative estimates of Os-related health risks. In the
last review of the Os NAAQS, the quantitative health risk assessment estimated Cb-related lung
function decrements, respiratory symptoms, respiratory-related hospital admissions, and non-
accidental and cardiorespiratory-related mortality (U.S. EPA, 2007). In those analyses, both
controlled human exposure and epidemiologic studies were used for the quantitative assessment
of Os-related human health risks.
In the current review, for short-term Os concentrations the HREA estimates lung function
decrements; respiratory symptoms in asthmatics; hospital admissions and emergency department
visits for respiratory causes; and all-cause mortality (U.S. EPA, 2014, Chapters 6 and 7). For
"long-term" Os concentrations, the UREA estimates respiratory mortality (U.S. EPA, 2014,
Chapter 7).77 Estimates of Os-induced lung function decrements are based on exposure modeling,
as noted above, combined with exposure-response relationships from controlled human exposure
studies (U.S. EPA, 2014, Chapter 6). Estimates of Cb-associated respiratory symptoms; hospital
admissions and emergency department visits; and mortality are based on concentration-response
relationships from epidemiologic studies (U.S. EPA, 2014, Chapter 7). As with the exposure
assessment discussed above, Os-associated health risks are estimated for recent air quality and
for ambient concentrations adjusted to just meet the current 8-hour Os NAAQS, based on 2006-
2010 air quality and adjusted precursor emissions.
Section 3.2.3.1 below discusses risk results for Os-induced lung function decrements
following short-term exposures, based on exposure modeling results and exposure-response
relationships from controlled human exposure studies. Section 3.2.3.2 discusses epidemiology-
based risk estimates, with a focus on all-cause mortality (short-term Cb concentrations);
respiratory-related morbidity outcomes (short-term Os concentrations); and respiratory mortality
(long-term Os concentrations).
3.2.3.1 Risk of Lung Function Decrements
In the last review, EPA conducted a health risk assessment that produced risk estimates
for the number and percent of school-aged children, asthmatic school-aged children, and the
general population experiencing lung function decrements associated with Os exposures for 12
urban areas. These estimates were based on exposure-response relationships developed from
77 Risk estimates for "long-term" concentrations are based on the concentration-response relationship identified in
the study by Jerrett et al. (2009). As discussed above, study authors used April to September averages of 1-hour
daily maximum Os concentrations as surrogates for "long-term" exposures.
3-103
-------
analysis of data from several controlled human exposure studies, combined with exposure
estimates developed for children and adults (U.S. EPA, 2007).
In the current review, the HREA estimates risks of lung function decrements in school-
aged children (ages 5 to 18), asthmatic school-aged children, and the general adult population for
15 urban case study areas.78 The results presented in the HREA are based on an updated dose-
threshold model that estimates FEVi responses for individuals following short-term exposures to
Os (McDonnell et al., 2012), reflecting methodological improvements since the last review (U.S.
EPA, 2014, section 6.2.4; section 3.1.2.1, above). The impact of the dose threshold is that Os-
induced FEVi decrements result primarily from exposures on days with ambient Os
concentrations above about 40 ppb (U.S. EPA, 2013, section 6.3.1, Figure 6-9).79
As discussed above (section 3.1.3), for effects such as lung function decrements, which
are transient and reversible, aspects such as the likelihood that these effects would occur
repeatedly and would interfere with normal activities are important to consider in making
judgments about adversity to individuals. As stated in the 2006 Criteria Document (Table 8-3, p.
8-68), for people with lung disease even moderate functional responses (e.g., FEVi decrements >
10% but < 20%) would likely interfere with normal activities for many individuals, and would
likely result in more frequent medication use. Moreover, as noted above, in a recent letter to the
Administrator, the CASAC Os Panel stated that '"[c]linically relevant effects are decrements >
10%, a decrease in lung function considered clinically relevant by the American Thoracic
Society" (Samet, 2011, p.2). The CASAC Os Panel also stated that:
[A] 10% decrement in FEVi can lead to respiratory symptoms, especially in
individuals with pre-existing pulmonary or cardiac disease. For example, people
with chronic obstructive pulmonary disease have decreased ventilatory reserve
(i.e., decreased baseline FEVi) such that a > 10% decrement could lead to
moderate to severe respiratory symptoms (Samet, 2011, p.7).
Consistent with this advice from the last review, in the current review CASAC has concluded
that "estimation of FEVi decrements of >15% is appropriate as a scientifically relevant surrogate
for adverse health outcomes in active healthy adults, whereas an FEVi decrement of >10% is a
scientifically relevant surrogate for adverse health outcomes for people with asthma and lung
disease" (Frey, 2014, p. 3).
78As noted for the exposure assessment above, the 15 cities assessed are Atlanta, Baltimore, Boston, Chicago,
Cleveland, Dallas, Denver, Detroit, Houston, Los Angeles, New York, Philadelphia, Sacramento, St. Louis, and
Washington, DC.
79 Error! Reference source not found, in the HREA shows that more than 90% of daily instances of FEVi
decrements > 10% occur when 8-hr average ambient concentrations are above 40 ppb for 2006 air quality in Los
Angeles. The distribution of decrements will be different for different study areas, years, and air quality scenarios
(U.S. EPA, 2014, section 6.3.1).
3-104
-------
In judging the extent to which moderate lung function decrements represent effects that
should be regarded as adverse to the health status of individuals, in previous NAAQS reviews we
have also considered the extent to which decrements were experienced repeatedly during the
course of a year (U.S. EPA, 2007). Although some experts would judge single occurrences of
moderate responses to be a "nuisance," especially for healthy individuals,80 a more general
consensus view of the adversity of such moderate responses emerges as the frequency of
occurrence increases. Thus in the past EPA has judged that repeated occurrences of moderate
responses, even in otherwise healthy individuals, may be considered to be adverse since they
could well set the stage for more serious illness (61 FR 65723). The CASAC panel in the 1997
NAAQS review expressed a consensus view that these "criteria for the determination of an
adverse physiological response were reasonable" (Wolff, 1995).
The HREA estimates risks of moderate to large lung function decrements, defined as
FEVi decrements > 10%, > 15%, or > 20%. In evaluating these lung function risk estimates
within the context of considering the adequacy of the current Os standard, we first consider the
following specific question:
• What are the nature and magnitude of lung function risks remaining upon just meeting
the current Os standard?
In considering risks of Os-induced FEVi decrements, we focus on the percent of children
estimated to experience decrements > 10, 15, and 20%, noting that the percentage of asthmatic
children estimated to experience such decrements is virtually indistinguishable from the
percentage estimated for all children.81 Compared to children, smaller percentages of adults are
estimated to experience Os-induced FEVi decrements (U.S. EPA, 2014, section 6.3.1, Table 6-
4). As for exposures of concern (see above), the patterns of results across urban case study areas
and over the years evaluated are similar in children and adults (U.S. EPA, 2014, Appendix 6E).
Therefore, while we highlight estimates in children, we note that these results are also
representative of the patterns estimated for adult populations.
Key results for children are summarized below for air quality adjusted to just meet the
current Os NAAQS (Figures 3-11 to 3-14).82 The estimates presented in Figures 3-11 to 3-14
below reflect consistent reductions across urban case study areas in the percent of children
estimated to experience Os-induced lung function decrements, relative to recent (i.e., unadjusted)
air quality (U.S. EPA, 2014, Appendix 6B). When averaged over the years evaluated in the
80Though not all experts, as indicated by the advice received on this issue from past CASAC Os Panels (Samet,
2011).
81 Though see below for discussion of uncertainty in lung function responses of children and asthmatics.
82 Figures 3-11 and 3-12 present estimates of one or more decrements. Figures 3-13 and 3-14 present estimates of
two or more decrements.
3-105
-------
HREA, reductions of up to about 40% were estimated compared to recent air quality. These
reductions reflect the consistent decreases in relatively high ambient Os concentrations upon
adjustment to just meet the current standard (section 3.2.1 above; U.S. EPA, 2014, Chapter 4).83
Such reductions in estimated lung function risks are evident throughout urban case study areas,
including in urban cores and in surrounding areas (U.S. EPA, 2014, section 9.6). Figures 3-11
(Average over years) and 3-12 (Worst-Case Years) present estimates of one or more Os-induced
lung function decrements. Figures 3-13 (Average over years) and 3-14 (Worst-Case Years)
present estimates of two or more decrements.
83 As noted above, the impact of the dose threshold is that O3-induced FEVi decrements result primarily from days
with average ambient Os concentrations above about 40 ppb.
3-106
-------
£
25
e c
o o
E S2o
°£
g£
•o £
E ^ -1C
O) QJ -I -'
11
X u
LU O)
c Q
e 510
T3 !?"
O T3
*- "
O °
= 1 5
O) C
y -
£
> 20%
>
Figure 3-11. Percent of school-aged children (5-18 yrs) estimated to experience one or more days with FEVi decrements > 10,
15, or 20% with air quality adjusted to just meet the current standard - Averaged over 2006 to 2010
5-107
-------
>20%
v^
Figure 3-12. Percent of school-aged children (5-18 yrs) estimated to experience one or more days with FEVi decrements > 10,
15, or 20% with air quality adjusted to just meet the current standard - Worst-Case Year from 2006 to 2010
5-108
-------
>20%
Figure 3-13. Percent of school-aged children (aged 5-18 yrs) estimated to experience two or more days with FEVi decrements >
10,15, or 20% with air quality adjusted to just meet the current standard - Averaged over 2006 to 2010
5-109
-------
o
E
o
w
CO
-
.
X
flj
CL
4) T-
2 S
IE "J-
O "o
»_ «
o ii
25
20
15
10
£ 0
I I I
- , fli-' wO>
>20%
Figure 3-14. Percent of school-aged children (5-18 yrs) estimated to experience two or more days with FEVi decrements > 10,
15, or 20% with air quality adjusted to just meet the current standard - Worst-Case Year from 2006 to 2010
5-110
-------
Based on Figures 3-11 to 3-14 and the associated details described in the HREA (U.S.
EPA, 2014, Chapter 6), we take note of the following with regard to lung function decrements
estimated to be allowed by the current standard:
1. For FEVi decrements > 10%:
a. On average over the years 2006 to 2010, the current standard is estimated to allow
approximately 14 to 19% of children in urban case study areas to experience one or
more lung function decrements > 10%. Summing across urban case study areas, this
corresponds to approximately 3 million children experiencing 15 million Cb-induced
lung function decrements > 10% during a single Os season. Of these children, about
300,000 are asthmatics.
b. On average over the years 2006 to 2010, the current standard is estimated to allow
approximately 8 to 12% of children in urban case study areas to experience two or
more Os-induced lung function decrements > 10%. Summing across the urban case
study areas, this corresponds to almost 2 million children (including almost 200,000
asthmatic children) estimated to experience two or more Os-induced lung function
decrements greater than 10% during a single Os season.
c. In the worst-case years, the current standard is estimated to allow approximately 17 to
22% of children in urban case study areas to experience one or more lung function
decrements > 10%, and approximately 10 to 14% to experience two or more Os-
induced lung function decrements > 10%.
2. For FEVi decrements > 15%:
a. On average over the years 2006 to 2010, the current standard is estimated to allow
approximately 3 to 5% of children in urban case study areas to experience one or
more lung function decrements > 15%. Summing across urban case study areas, this
corresponds to over 750,000 children (including approximately 80,000 asthmatic
children) estimated to experience at least one Os-induced lung function decrement >
15% during a single Os season.
b. On average over the years 2006 to 2010, the current standard is estimated to allow
approximately 2 to 3% of children in urban case study areas to experience two or
more Os-induced lung function decrements > 15%.
c. In the worst-case years, the current standard is estimated to allow approximately 4 to
7% of children in urban case study areas to experience one or more lung function
decrements > 15%, and approximately 2 to 4% to experience two or more Os-induced
lung function decrements > 15%.
3. For FEVi decrements > 20%:
a. On average over the years 2006 to 2010, the current standard is estimated to allow
approximately 1 to 2% of children in urban case study areas to experience one or
more lung function decrements > 20%. Summing across urban case study areas, this
corresponds to almost 300,000 children (including approximately 30,000 asthmatic
3-111
-------
children) estimated to experience at least one Os-induced lung function decrement >
20% during a single Os season.
b. On average over the years 2006 to 2010, the current standard is estimated to allow
less than about 1% of children in urban case study areas to experience two or more
Os-induced lung function decrements > 20%.
c. In the worst-case years, the current standard is estimated to allow approximately 2 to
3% of children to experience one or more lung function decrements > 20%, and less
than 2% to experience two or more Os-induced lung function decrements > 20%.
In further considering estimated lung function risks from the HREA, we next consider the
following question:
• What are the important sources of uncertainty associated with lung function risk
estimates?
In addition to the uncertainties noted above for exposure estimates, the HREA identifies
several key uncertainties associated with estimates of Os-induced lung function decrements. An
uncertainty with particular potential to impact our consideration of risk estimates in this Policy
Assessment stems from the lack of exposure-response information in children. In the absence of
controlled human exposure data for children, risk estimates are based on the assumption that
children exhibit the same lung function response following Os exposures as healthy 18 year olds
(i.e., the youngest age for which controlled human exposure data is available) (U.S. EPA, 2014,
section 6.5.3). This assumption was justified in part by the findings of McDonnell et al. (1985),
who reported that children 8-11 year old experienced FEVi responses similar to those observed
in adults 18-35 years old. In addition, as discussed in the ISA (U.S. EPA, 2013, section 6.2.1),
summer camp studies of school-aged children reported Os-induced lung function decrements
similar in magnitude to those observed in controlled human exposure studies using adults. In
extending the risk model to children, the HREA fixes the age term in the model at its highest
value, the value for age 18. This approach could result in either over- or underestimates of Os-
induced lung function decrements in children, depending on how children compare to the adults
used in controlled human exposure studies (U.S. EPA, 2014, section 6.5.3).
A related source of uncertainty is that the risk assessment estimates Os-induced
decrements in asthmatics using the exposure-response relationship developed from data collected
from healthy individuals. Although the evidence has been mixed (U.S. EPA, 2013, section
6.2.1.1), several studies have reported larger Os-induced lung function decrements in asthmatics
than in non-asthmatics (Kreit et al., 1989; Horstman et al., 1995; Torres et al., 1996; Alexis et al.,
2000). Consistent with the findings of the ISA (U.S. EPA, 2013, section 6.2.1.1), CASAC noted
that "[a]sthmatic subjects appear to be at least as sensitive, if not more sensitive, than non-
asthmatic subjects in manifesting ozone-induced pulmonary function decrements" (Frey, 2014,
3-112
-------
p. 4). To the extent asthmatics experience larger Os-induced lung function decrements than the
healthy adults used to develop exposure-response relationships, the HREA could underestimate
the impacts of Os exposures on lung function in asthmatics, including asthmatic children. The
HREA notes that the magnitude this uncertainty might have on risk estimates remains unknown
at this time (U.S. EPA, 2014, section 6.5.4).
3.2.3.2 Estimated Health Risks Associated with Short- or Long-Term Os Exposures,
Based on Epidemiologic Studies
Risk estimates based on epidemiologic studies can provide perspective on the most
serious Cb-associated public health outcomes (e.g., mortality, hospital admissions, emergency
department visits) in populations that often include at-risk groups. The HREA estimates Os-
associated risks in 12 urban case study areas84 using concentration-response relationships drawn
from epidemiologic studies. These concentration-response relationships are based on "area-
wide" average Os concentrations.85 The HREA estimates risks for the years 2007 and 2009 in
order to provide estimates of risk for a year with generally higher Os concentrations (2007) and a
year with generally lower Os concentrations (2009) (U.S. EPA, 2014, section 7.1.1).
In the last review, epidemiologic-based risks were estimated for Os concentrations above
mean "policy-relevant background concentrations." As discussed above (Chapter 2), policy-
relevant background (PRB) concentrations were defined as the distribution of ozone
concentrations that would be observed in the U.S. in the absence of anthropogenic (man-made)
emissions of ozone precursor emissions (e.g., VOC, CO, NOx) in the U.S., Canada, and Mexico.
This approach provided a focus on Os concentrations "that can be controlled by U.S. regulations
(or through international agreements with neighboring countries)" (U.S. EPA, 2007, pp. 2-48 to
2-54).
As in the last review, we recognize that ambient Os concentrations, and therefore Os-
associated health risks, result from precursor emissions from various types of sources. Based on
the air quality modeling discussed above in chapter 2, approximately 30 to 60% of average
daytime Os during the warm season (i.e., daily maximum 8-hour concentrations averaged from
April to October) is attributable to precursor emissions from U.S. anthropogenic sources (section
2.4.4). The remainder is attributable to precursor emissions from international anthropogenic
84 The 12 urban areas evaluated are Atlanta, Baltimore, Boston, Cleveland, Denver, Detroit, Houston, Los Angeles,
New York, Philadelphia, Sacramento, and St. Louis.
85 In the epidemiologic studies that provide the health basis for HREA risk assessments, concentration-response
relationships are based on daytime Cb concentrations, averaged across multiple monitors within study areas. These
daily averages are used as surrogates for the spatial and temporal patterns of exposures in study populations.
Consistent with this approach, the HREA epidemiologic-based risk estimates also utilize daytime Os concentrations,
averaged across monitors, as surrogates for population exposures. In this PA, we refer to these averaged
concentrations as "area-wide" Os concentrations. Area-wide concentrations are discussed in more detail in section
3.1.4, above.
3-113
-------
sources and natural sources. Because the HREA characterizes health risks from all Cb, regardless
of source, risk estimates reflect emissions from U.S. anthropogenic, international anthropogenic,
and natural sources.
In evaluating epidemiology-based risk estimates within the context of the adequacy of the
current standard, we first consider the following question:
• What are the nature and magnitude of the Os-associated mortality and morbidity risks
remaining upon adjustment of air quality to just meet the current Os standard?
In addressing this question, we note that the HREA estimates mortality and morbidity
risks associated with just meeting the current standard by applying concentration-response
relationships from epidemiologic studies to the entire distributions of adjusted "area-wide"
average Os concentrations present in urban case study areas (U.S. EPA, 2014, Chapter 7).
Implicit in this approach to estimating risks is the assumption that concentration-response
relationships are linear over those distributions. Therefore, as noted in section 3.2.1, when air
quality is adjusted to just meet the current standard, risk estimates are influenced by the
decreases in area-wide Os concentrations at the upper ends of warm season distributions and the
increases in area-wide Os concentrations at the lower ends of those distributions (U.S. EPA,
2014, section 4.3.3.2, Figures 4-9 and 4-10).86 When the decreases and increases are of the same
magnitude, they result in the same degree of change in estimated risks, though opposite in
direction. Therefore, seasonal estimates of Os-associated mortality and morbidity risks either
increase or decrease in response to air quality adjustment, depending on the seasonal balance
between the modeled daily decreases in high area-wide Os concentrations and increases in low
area-wide Os concentrations. One consequence is that the estimated impacts on mortality and
morbidity risks of adjusting air quality to just meet the current standard are more modest, and
less directionally consistent across urban case study areas, than on either exposures of concern or
Os-induced lung function decrements.
In the remainder of this section, we consider estimates of total (non-accidental) mortality
and respiratory morbidity associated with short-term Os concentrations, and respiratory mortality
associated with "long-term" Os concentrations.
Total Mortality - Short-Term O3
Risk estimates for total mortality are based on concentration-response relationships
described by Smith et al. (2009). To generate risk estimates, the HREA uses "area-wide"
86 On a given day, area-wide Cb concentrations and estimated risks decrease when the sum of the changes at
monitors with decreasing Os (e.g., downwind of important NOX sources, where the highest O3 concentrations often
occur) are larger than the sum of the changes at monitors with increasing Os (e.g., often in urban centers with high
concentrations of NOx emissions, where ambient O3 concentrations are suppressed and thus relatively low). Area-
wide Os concentrations and estimated risks increase when the opposite occurs.
3-114
-------
averages of daily maximum 8-hour Os concentrations over the full monitoring periods in urban
case study areas. When 2007 air quality was adjusted to the current standard (the year with
generally "higher" Os-associated risks), 10 of 12 urban case study areas exhibited either
decreases or virtually no change in estimates of the number of Os-associated deaths (U.S. EPA,
2014, Appendix 7B). Increases were estimated in two of the urban case study areas (Houston,
Los Angeles) (U.S. EPA, 2014, Appendix 7B).87
Figure 3-15 below presents estimates of Os-associated all-cause mortality in urban case
study areas for 2007 and 2009, with air quality adjusted to just meet the current Os standard. The
HREA estimates that upon just meeting the current standard, Os could be associated with from
0.8 to 4.1% of all-cause mortality across the urban case study areas. This corresponds to
approximately 60 to 3,200 Os-associated deaths per season in individual urban case study areas,
and approximately 7,000 to 7,500 Os-associated deaths per season summed over the 12 urban
case study areas (U.S. EPA, 2014, Tables 7-7 and 7-8).
Associated
45
35
TO fn
£ o
o ,-
25
•= 2
ra
*-/
o
5
u
b
CL
12007
12009
<& ^
/ y
^
Figure 3-15. Percent of all-cause mortality associated with Os for air quality adjusted to
just meet the current standard.
In considering the risk estimates presented in Figure 3-15, which are based on applying
linear concentration-response relationships to the full distributions of daily 8-hour "area-wide"
87 For 2009 (i.e., the year with generally lower O3 concentrations), changes in risk were generally smaller than in
2007 (i.e., most changes about 2% or smaller). Increases were estimated for Houston, Los Angeles, and New York
City.
3-115
-------
Os concentrations, we note the ISA conclusion that there is less certainty in the shape of
concentration-response functions for area-wide Os concentrations at the lower ends of warm
season distributions (i.e., below about 20 to 40 ppb depending on the Os metric, health endpoint,
and study population) (U.S. EPA, 2013, section 2.5.4.4). We also recognize that for the range of
health endpoints evaluated, controlled human exposure and animal toxicological studies provide
greater certainty in the increased incidence, magnitude, and severity of effects at higher exposure
concentrations (discussed in sections 3.1.2.2 and 3.1.4.2, above).88 Thus, in addition to
considering estimates of total Os-associated risks, we also consider the extent to which risks are
associated with days with higher, versus lower, area-wide Os concentrations.
Figure 3-16 presents risk estimates, summed across urban case study areas, for days with
area-wide concentrations at or above 20, 40, and 60 ppb. Daytime Os concentrations in the upper
portion of the distribution of area-wide concentrations (e.g., at or above 40 or 60 ppb) are
estimated to be associated with hundreds to thousands of deaths per year in urban case study
QQ
areas.
to £l
CO O3
o -o
g ji
-O «>
CD -*—
II
8,000
7JOOO
6,000
CO
of
(—
o>
"O
"S
"ro
"o
0
CO
CO
0!
6"
~£i
CO
O3
JD
CD
^>^
"0
ZJ
to
CD
10
03
O
f
O3
.0
SflOO
4,000
3,000
2,000
1,000
I Total risk
I 20+ ppb
I 4Q+ ppb
60+ ppb
2007
2009
Figure 3-16. Estimated Os-associated deaths attributable to various area-wide average
concentrations, with air quality adjusted to just meet current standard.
88 As discussed in section 3.1.4.2, as ambient concentrations increase the potential for exposures to higher Os
concentrations also increases. Thus with increasing ambient concentrations, controlled human exposure and animal
toxicological studies support the increased incidence, magnitude, and severity of Os-attributable effects.
89 The relatively small proportion of O3-associated deaths attributable to days with area-wide concentrations of 60
ppb or greater reflects the relatively small proportion of days with such elevated area-wide concentrations.
3-116
-------
Respiratory Mortality - "Long-Term" Os
The HREA estimates the risk of respiratory mortality associated with long-term Os
exposures, based on the study by Jerrett et al. (2009) (U.S. EPA, 2014, Chapter 7). As discussed
above (section 3.1.4.3), Jerrett et al. (2009) reported that when seasonal averages of 1-hour daily
maximum Os concentrations ranged from 33 to 104 ppb, there was no statistically significant
deviation from a linear concentration-response relationship between Os and respiratory mortality
across 96 U.S. cities (U.S. EPA, 2013, section 7.7). However, the authors reported "limited
evidence" for an effect threshold at an Os concentration of 56 ppb (p=0.06). In communications
with EPA staff (described in Sasser, 2014), the study authors indicated that it is not clear whether
a threshold model is a better predictor of respiratory mortality than the linear model, and that
"considerable caution should be exercised in accepting any specific threshold." Consistent with
this communication, the HREA estimated respiratory mortality associated with long-term Os
concentrations based on the linear model from the published study, and in a series of sensitivity
analyses with models that included thresholds ranging from 40 to 60 ppb (U.S. EPA, 2014,
Figure 7-9).
To generate risk estimates, the HREA uses "area-wide" averages of 1-hour daily
maximum Os concentrations during the warm season (April to September). When 2007 air
quality was adjusted to just meet the current standard (i.e., the year with generally higher Os
concentrations) all 12 of the urban case study areas exhibited decreases in estimated Os-
associated respiratory mortality (i.e., compared to recent, unadjusted air quality). For 2009
adjusted air quality (i.e., the year with generally lower Os concentrations), urban case study areas
exhibited either no change in estimated risk, or decreases in risk that were smaller than those for
2007 (U.S. EPA, 2014, Appendix 7B, Tables 7B-6 and 7B-7). Risk estimates based on the linear
model, for air quality adjusted to just meet the current standard, are presented below in Figure 3-
17.
3-117
-------
25
-ffl _
ns ~o
a |
tn C
< 3
-
nj j-
•"
12007
2009
•/
Figure 3-17. Percent of baseline respiratory mortality estimated to be associated with long-
term Os.
Based on a concentration-response function that is linear over the entire distribution of
long-term Os concentrations, Os is estimated to be associated with approximately 16 to 21% of
respiratory deaths in urban case study areas during the warm season. This corresponds to
approximately 300 to 2,100 Os-associated deaths per season in individual urban case study areas,
and a total of approximately 8,000 to 9,000 Os-associated deaths summed across all 12 case
study areas. Based on threshold models, HREA sensitivity analyses indicate that the number of
respiratory deaths associated with long-term Os concentrations could potentially be considerably
lower (i.e., by more than 75% if a threshold exists at 40 ppb, and by about 98% if a threshold
exists at 56 ppb) (U.S. EPA, 2014, Figure 7-9).
Hospital Admissions, Emergency Department Visits, and Asthma Exacerbations
Risk estimates for respiratory-related hospital admissions, emergency department visits,
and asthma exacerbations associated with air quality adjusted to just meet the current standard
are based on several studies, as presented in Table 7-2 of the HREA (U.S. EPA, 2014).90
Estimates indicate that Os-associated respiratory-related hospital admissions generally account
for approximately 2 to 3% of total respiratory-related admissions in urban case study locations.
Depending on the city, this corresponds to 10's to 100's of Os-associated hospital admissions per
season. Estimates indicate that Cb-associated respiratory-related emergency department visits
90As with respiratory mortality above, the HREA does not characterize distributions of respiratory morbidity risks
over distributions of ambient Os concentrations. Therefore, in considering respiratory morbidity risks we evaluate
estimates of total risk.
3-118
-------
account for approximately 3 to 20% of total respiratory-related emergency department visits in
Atlanta and New York City (corresponding to thousands of visits per season in these two cities),
and that Cb-associated asthma exacerbations account for approximately 15 to 30% of total
exacerbations in Boston (30,000 to 80,000 exacerbations per season). Full estimates are
presented in Tables 7-9 to 7-11 in the HREA (U.S. EPA, 2014).
Based on the detailed information presented in Chapter 7 of the HREA (U.S. EPA, 2014),
we note the following key observations:
1. In focusing on total risk, the current standard is estimated to allow thousands of Os-
associated deaths per year in the urban case study areas. These estimates are based on
concentration-response functions from epidemiologic studies that used either 8-hour daily
Os concentrations (total mortality associated with short-term Os) or seasonal averages of
1-hour daily Os concentrations (respiratory mortality associated with long-term Os).
2. In focusing on the risks associated with the upper portions of distributions of ambient
concentrations, the current standard is estimated to allow hundreds to thousands of Cb-
associated deaths per year in the urban case study areas. These estimates are based on
concentration-response functions from an epidemiologic study that evaluated associations
between 8-hour daily Os concentrations and total mortality.
3. In urban case study areas, the current standard is estimated to allow tens to thousands of
Os-associated morbidity events per year. Distributions of Os-associated morbidity over
distributions of ambient Os concentrations would likely be similar to mortality, though
the HREA did not analyze such distributions for morbidity endpoints.
In further considering estimated Os-associated mortality and morbidity risks from the
HREA, we next consider the following question:
• What are the important sources of uncertainty associated with mortality and morbidity
risk estimates?
Compared to estimates of Os exposures of concern and estimates of Os-induced lung
function decrements (discussed above), the HREA conclusions reflect somewhat lower
confidence in epidemiologic-based risk estimates, given important uncertainties. In particular,
the HREA highlights the unexplained 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 Os concentrations (U.S. EPA, 2014, section 9.6). The
HREA also concludes that lower confidence should be placed in the results of the assessment of
respiratory mortality risks associated with long-term Os exposures, primarily because that
analysis is based on only one study (even though that study is well-designed) and because of the
uncertainty in that study about the existence and level of a potential threshold in the
concentration-response function (U.S. EPA, 2014, section 9.6). This section discusses some of
the key uncertainties in epidemiologic-based risk estimates, with a focus on uncertainties that can
3-119
-------
have particularly important implications for our consideration of epidemiology-based risk
estimates in this PA
Estimating air quality that just meets the current standard based on modeled responses to
reductions in NOx emissions generally reduces Cb-associated mortality and morbidity risks in
locations and time periods with relatively high ambient Os concentrations and increases risks in
locations and time periods with relatively low concentrations. When evaluating uncertainties in
epidemiologic risk estimates, it is important to consider the extent to which the pattern of air
quality changes in response to reductions in NOx emissions is representative of trends in ambient
Os; the extent to which estimated changes in risks in urban case study areas are representative of
the changes that would be experienced broadly across the U.S. population; and the extent to
which the Os response to reductions in precursor emissions could differ with emissions reduction
strategies that are different from those used in REA risk estimates.
To evaluate the first issue, the HREA conducted a national analysis evaluating trends in
monitored ambient Os concentrations during a time period when the U.S. experienced large-scale
reductions in NOx emissions (i.e., 2001 to 2010). Analyses of trends in monitored Os indicate
that over such a time period, the upper end of the distribution of monitored Os concentrations
(i.e., indicated by the 95th percentile) generally decreased in urban and non-urban locations
across the U.S. (U.S. EPA, 2014, Figure 8-29). During this same time period, median Os
concentrations decreased in suburban and rural locations, and in some urban locations. However,
median concentrations increased in some large urban centers (U.S. EPA, 2014, Figure 8-28). As
discussed in the REA, and above (II.C.I), these increases in median concentrations likely reflect
the increases in relatively low Os concentrations that can occur near important sources of NOx
upon reductions in NOx emissions (U.S. EPA, 2014, section 8.2.3.1). These patterns of
monitored Os during a period when the U.S. experienced large reductions in NOx emissions are
qualitatively consistent with the modeled responses of Os to reductions in NOx emissions.
To evaluate the second issue, the HREA conducted national air quality modeling
analyses. These analyses estimated the proportion of the U.S. population living in locations
where seasonal averages of daily Os concentrations are estimated to decrease in response to
reductions in NOx emissions, and the proportion living in locations where such seasonal
averages are estimated to increase. Given the strong relationship between changes in seasonal
averages of daily Os concentrations and changes in seasonal risk estimates, this analysis informs
consideration of the extent to which the risk results in urban case study areas represent the U.S.
population as a whole. This representativeness analysis indicates that the 12 urban case study
areas may not represent the response of Os in other populated areas of the U.S., including
suburban areas, smaller urban areas, and rural areas. This analysis also indicates that the majority
of the U.S. population lives in locations where reducing NOx emissions would be expected to
3-120
-------
result in decreases in warm season averages of daily maximum 8-hour ambient Os
concentrations. One implication of these results is that HREA risk estimates for the urban case
study areas may understate the average reduction in Os-associated mortality and morbidity risk
that would be experienced across the U.S. population upon reducing NOx emissions (U.S. EPA,
2014, sections 8.2.3.2 and 8.4).
To evaluate the third issue, the HREA assesses the Os air quality response to reducing
both NOx and VOC (i.e., in addition to assessing reductions in NOx emissions alone) for a
subset of seven urban case study areas. As noted above (section 3.2.1), in most of these urban
case study areas the inclusion of VOC emissions reductions did not alter the NOx emissions
reductions required to meet the current or alternative standards.91 However, the addition of VOC
reductions generally resulted in larger decreases in mid-range Os concentrations (25th to 75th
percentiles) (U.S. EPA, 2014, Appendix 4D, section 4D-4.7).92 In addition, in all seven of the
urban case study areas evaluated, the increases in low Os concentrations were smaller for the
NOx/VOC scenarios than the NOx alone scenarios (U.S. EPA, 2014, Appendix 4D, section 4D-
4.7). This was most apparent for Denver, Houston, Los Angeles, New York, and Philadelphia.
Given the impacts on total risk estimates of increases in low Os concentrations, these results
suggest that in some locations better-optimized emissions reduction strategies could result in
larger reductions in Os-associated mortality and morbidity than indicated in the HREA core
estimates.
Section 7.4 of the HREA also highlights some additional uncertainties associated with
epidemiologic-based risk estimates (U.S. EPA, 2014). This section of the HREA identifies and
discusses sources of uncertainty and presents a qualitative evaluation of key parameters that can
introduce uncertainty into risk estimates (U.S. EPA, 2014, Table 7-4). For several of these
parameters the HREA also presents quantitative sensitivity analyses (U.S. EPA, 2014, section
7.5.3). Of the uncertainties discussed in Chapter 7 of the HREA, those related to the application
of concentration-response functions from epidemiologic studies can have particularly important
implications for our consideration of epidemiology-based risk estimates in this PA, as discussed
below.
An important uncertainty is the shape of concentration-response functions at low ambient
Os concentrations (U.S. EPA, 2014, Table 7-4).93 Consistent with the ISA conclusion that there
91 The exception is Chicago and Denver, for which the HREA risk estimates are based on reductions in both NOx
and VOC (U.S. EPA, 2014, section 4.3.3.1). Emissions of NOx and VOC were reduced by equal percentages, a
scenario not likely to reflect the optimal combination for reducing risks.
92 This was the case for all of the urban case study areas evaluated, with the exception of New York (U.S. EPA,
2014, Appendix 4D, section 4D-4.7).
93 A related uncertainty is the existence, or not, of a threshold. The HREA addresses this issue for long-term O3 by
evaluating risks in models that include potential thresholds (see above).
3-121
-------
is no discernible population threshold in Cb-associated health effects, the HREA estimates
epidemiology-based mortality and morbidity risks for entire distributions of ambient Os
concentrations, with the assumption that concentration-response relationships remain linear over
those distributions. In addition, in recognition of the ISA conclusion that certainty in the shape of
Os concentration-response functions decreases at low ambient concentrations, the HREA also
estimates distributions of total mortality incidence for various portions of the distribution of
ambient Os concentrations. In this PA, we consider both types of risk estimates while
recognizing that we have greater certainty in the increased incidence and severity of Os-
attributable effects at higher ambient Os concentrations (which drive higher exposure
concentrations, section 3.2.2 above), as compared to lower concentrations.94
The HREA also notes important uncertainties associated with using a concentration-
response relationship developed for a particular population in a particular location to estimate
health risks in different populations and locations (U.S. EPA, 2014, Table 7-4). As discussed
above, concentration-response relationships derived from epidemiologic studies reflect the
spatial and temporal patterns of population exposures during the study. The HREA applies
concentration-response relationships from epidemiologic studies to adjusted air quality in study
areas that are different from, and often larger in spatial extent than, the areas used to generate the
relationships. This approach ensures the inclusion of the actual non-attainment monitors that
often determine the magnitude of emissions reductions for the air quality adjustments throughout
the urban case study areas. This approach also allows the HREA to estimate patterns of health
risks more broadly across a larger area, including a broader range of air quality concentrations
and a larger population. The HREA notes that it is not possible to quantify the impacts of this
uncertainty on risk estimates in most urban case study locations, though the HREA notes that
mortality effect estimates for different portions of the New York City CBSA-based assessment
area vary by a factor of almost 10 (U.S. EPA, 2014, section 7.5.3).
An additional, related uncertainty is that associated with applying concentration-response
functions from epidemiologic studies to adjusted air quality. Concentration-response functions
from the Os epidemiologic studies used in the HREA are based on associations between day to
day variation in "area-wide" Os concentrations (i.e., averaged across multiple monitors) and
variation in health effects. Epidemiologic studies use these area-wide Os concentrations, which
reflect the particular spatial and temporal patterns of ambient Os present in study locations, as
surrogates for the pattern of Os exposures experienced by study populations. To the extent
adjusting Os concentrations to just meet the current standard results in important alterations in
94 As discussed above, we also consider the potential implications of the existence of a threshold in the association
between long-term Os and respiratory mortality.
3-122
-------
the spatial and/or temporal patterns of ambient Cb, there is uncertainty in the appropriateness of
applying concentration-response functions from epidemiologic studies to estimate health risks
associated with adjusted Cb air quality.95 In particular, this uncertainty could be important to the
extent that (1) factors associated with space modify the effects of Os on health or (2) spatial
mobility is a key driver of individual-level exposures. Although the impact of this uncertainty on
risk estimates cannot be quantified (U.S. EPA, 2014, Table 7-4), it has the potential to become
more important as air quality adjustment results in larger changes in spatial and temporal patterns
of ambient Os concentrations across urban case study areas.
There is also uncertainty related specifically to the public health importance of the
increases in relatively low Os concentrations following air quality adjustment. This uncertainty
relates to the fact that HREA risk estimates are equally influenced by decreasing high
concentrations and increasing low concentrations, when the increases and decreases are of equal
magnitude. Even on days with increases in relatively low area-wide average concentrations,
resulting in increases in estimated risks, some portions of the urban case study areas could
experience decreases in high Os concentrations. To the extent it is reasonable to conclude that
Os-attributable effects are more strongly supported for higher ambient concentrations (see
above), likely resulting in higher exposure concentrations for some portions of study areas, the
impacts on risk estimates of increasing low Os concentrations reflect an important source of
uncertainty.
The use of a national concentration-response function to estimate respiratory mortality
associated with long-term Os is a source of uncertainty. Risk estimates generated in sensitivity
analyses using region-specific effect estimates differ substantially from the core estimates based
on a single national-level effect estimate (U.S. EPA, 2014; Table 7-14). Furthermore, the risk
estimates generated using the regional effect estimates display considerable variability across
urban case study areas (U.S. EPA, 2014; Table 7-14), reflecting the substantial variability in the
underlying effect estimates (see Jerrett et al., 2009, Table 4). While the results of the HREA
sensitivity analyses evaluating this uncertainty point to the potential for regional heterogeneity in
the long-term risk estimates, the relatively large confidence intervals associated with regional
effect estimates resulted in the HREA conclusion that staff does not have confidence in the
regionally-based risk estimates themselves.
Finally, we note the HREA does not quantify any reductions in risk that could be
associated with reductions in the ambient concentrations of pollutants other than Os, resulting
from control of NOx. For example as discussed in chapter 2 of this PA, NOx emissions
contribute to ambient NCh, and NOx and VOCs can contribute to secondary formation of PM2.5
95 As discussed above (section 3.2.1), decreasing modeled NOX emissions to just meet the current standard can
dramatically alter the spatial and temporal patterns of ambient Os concentrations across urban case study areas.
3-123
-------
constituents, including ammonium sulfate (NFLtSO^, ammonium nitrate (NFLtNOs), and organic
carbon (OC). Therefore, at some times and in some locations, control strategies that would
reduce NOx emissions (i.e., to meet an Os standard) could reduce ambient concentrations of NCh
and PM2.5, resulting in health benefits beyond those directly associated with reducing ambient Os
concentrations.96
3.3 CASAC ADVICE AND PUBLIC COMMENTERS' VIEWS ON THE
ADEQUACY OF THE CURRENT STANDARD
Beyond the evidence- and risk/exposure-based information discussed above, staff has
also taken into account the comments and advice of CASAC, based on their reviews of the ISA,
the HREA and PA, as well as comments provided by public commenters. The range of views
summarized here generally reflects differing judgments as to the relative weight to place on
various types of evidence, the exposure- and risk-based information, and the associated
uncertainties, as well as differing judgments about the importance of various Os-related health
effects from a public health perspective.
Following the 2008 decision to revise the primary Os standard by setting the level at
0.075 ppm (75 ppb), CASAC strongly questioned whether the standard met the requirements of
the CAA, further described below. In September 2009, EPA announced its intention to
reconsider the 2008 standards, issuing a notice of proposed rulemaking in January 2010 (FR 75
2938). Soon after, EPA solicited CASAC review of that proposed rule and in January 2011
solicited additional advice. This proposal was based on the scientific and technical record from
the 2008 rulemaking, including public comments and CASAC advice and recommendations. As
further described in section 1.2.2 above, EPA in the fall of 2011 did not revise the standard as
part of the reconsideration process but decided to coordinate further proceedings on the
reconsideration rulemaking with this ongoing periodic review. Accordingly, in this section we
describe CASAC's advice related to the 2008 final decision and the subsequent reconsideration,
as well as its advice on the NAAQS review that was initiated in September 2008.
In April 2008, the members of the CASAC Ozone Review Panel sent a letter to EPA
stating "[I]n our most-recent letters to you on this subject—dated October 2006 and March
2007—the CASAC unanimously recommended selection of an 8-hour average Ozone NAAQS
within the range of 0.060 to 0.070 parts per million [60 to 70 ppb] for the primary (human
health-based) Ozone NAAQS" (Henderson, 2008). The letter continued:
96 We expect little focus by states on controlling NOX for purposes of controlling PM2 5 given the more efficient
control of PM25 through reduction of SO2 and direct PM25 emissions in most locations. Thus, consideration in this
review of reductions in ambient PM2 5 resulting from putative NOX control would not double-count PM2 5 emission
reductions.
3-124
-------
The CASAC now wishes to convey, by means of this letter, its additional,
unsolicited advice with regard to the primary and secondary Ozone NAAQS. In
doing so, the participating members of the CASAC Ozone Review Panel are
unanimous in strongly urging you or your successor as EPA Administrator to
ensure that these recommendations be considered during the next review cycle for
the Ozone NAAQS that will begin next year ... numerous medical organizations
and public health groups have also expressed their support of these CASAC
recommendations' ... [The CASAC did] not endorse the new primary ozone
standard as being sufficiently protective of public health. The CASAC—as the
Agency's statutorily-established science advisory committee for advising you on
the national ambient air quality standards—unanimously recommended
decreasing the primary standard to within the range of 0.060 0.0 70 ppm [60 to
70 ppb]. It is the Committee's consensus scientific opinion that your decision to
set the primary ozone standard above this range fails to satisfy the explicit
stipulations of the Clean Air Act that you ensure an adequate margin of safety for
all individuals, including sensitive populations.
In response to EPA's solicitation of their advice on the Agency's proposed rulemaking as
part of the reconsideration, CASAC conveyed support (Samet, 2011).
CASA C fully supports EPA 's proposed range of'0.060 - 0.070 parts per million
(ppm) for the 8-hour primary ozone standard. CASAC considers this range to be
justified by the scientific evidence as presented in the Air Quality Criteria for
Ozone and Related Photochemical Oxidants (March 2006) and Review of the
National Ambient Air Quality Standards for Ozone: Policy Assessment of
Scientific and Technical Information, OAQPS Staff Paper (July 2007). As stated
in our letters of October 24, 2006, March 26, 2007 and April 7, 2008 to former
Administrator Stephen L. Johnson, CASAC unanimously recommended selection
of an 8-hour average ozone NAAQS within the range proposed by EPA (0.060 to
0.070 ppm). In proposing this range, EPA has recognized the large body of data
and risk analyses demonstrating that retention of the current standard would
leave large numbers of individuals at risk for respiratory effects and/or other
significant health impacts including asthma exacerbations, emergency room
visits, hospital admissions and mortality.
In response to EPA's request for additional advice on the reconsideration in 2011,
CASAC reaffirmed their conclusion that "the evidence from controlled human and
epidemiological studies strongly supports the selection of a new primary ozone standard within
the 60 - 70 ppb range for an 8-hour averaging time" (Samet, 2011). As requested by EPA,
CASAC's advice and recommendations were based on the scientific and technical record from
the 2008 rulemaking. In considering the record for the 2008 rulemaking, CASAC stated the
following to summarize the basis for their conclusions (Samet, 2011, pp. ii to iii).
• The evidence available on dose-response for effects of ozone shows
associations extending to levels within the range of concentrations
currently experienced in the United States.
3-125
-------
• There is scientific certainty that 6.6-hour exposures with exercise of
young, healthy, non-smoking adult volunteers to concentrations > 80 ppb
cause clinically relevant decrements of lung function.
• Some healthy individuals have been shown to have clinically relevant
responses, even at 60 ppb.
• Since the majority of clinical studies involve young, healthy adult
populations, less is known about health effects in such potentially ozone
sensitive populations as the elderly, children and those with
cardiopulmonary disease. For these susceptible groups, decrements in
lung function may be greater than in healthy volunteers and are likely to
have a greater clinical significance.
• Children and adults with asthma are at increased risk of acute
exacerbations on or shortly after days when elevated ozone concentrations
occur, even when exposures do not exceed the NAAQS concentration of 75
ppb.
• Large segments of the population fall into what EPA terms a "sensitive
population group,'' i.e., those at increased risk because they are more
intrinsically susceptible (children, the elderly, and individuals with
chronic lung disease) and those who are more vulnerable due to increased
exposure because they work outside or live in areas that are more polluted
than the mean levels in their communities.
With respect to evidence from epidemiologic studies, CASAC stated "while epidemiological
studies are inherently more uncertain as exposures and risk estimates decrease (due to the greater
potential for biases to dominate small effect estimates), specific evidence in the literature does
not suggest that our confidence on the specific attribution of the estimated effects of ozone on
health outcomes differs over the proposed range of 60-70 ppb." (Samet, 2011, p. 10).
Following their review of the second draft PA in the current review, which considers an
updated scientific and technical record since the 2008 rulemaking, CASAC concluded that "there
is clear scientific support for the need to revise the standard" (Frey, 2014, p. ii). In particular,
CASAC noted the following (Frey, 2014, p. 5):
[T]he scientific evidence provides strong support for the occurrence of a range of
adverse respiratory effects and mortality under air quality conditions that would
meet the current standard. Therefore, CASAC unanimously recommends that the
Administrator revise the current primary ozone standard to protect public health.
In supporting these conclusions, CASAC judged that the strongest evidence comes from
controlled human exposure studies of respiratory effects. The Committee specifically noted that
"the combination of decrements in FEVi together with the statistically significant alterations in
symptoms in human subjects exposed to 72 ppb ozone meets the American Thoracic Society's
definition of an adverse health effect" (Frey, 2014, p. 5). CASAC further judged that "the level at
3-126
-------
which adverse effects might be observed would likely be lower for more sensitive subgroups,
such as those with asthma" (Frey, 2014, p. 5).
With regard to lung function risk estimates based on information from controlled human
exposure studies, CASAC concluded that "estimation of FEV1 decrements of >15% is
appropriate as a scientifically relevant surrogate for adverse health outcomes in active healthy
adults, whereas an FEV1 decrement of >10% is a scientifically relevant surrogate for adverse
health outcomes for people with asthma and lung disease" (Frey, 2014, p. 3). The Committee
further concluded that "[a]sthmatic subjects appear to be at least as sensitive, if not more
sensitive, than non-asthmatic subjects in manifesting Os-induced pulmonary function
decrements" (Frey, 2014, p. 4). In considering estimates of the occurrence of these decrements in
urban case study areas, CASAC specifically noted that the current standard is estimated to allow
11 to 22% of school age children to experience at least one day with an FEVi decrement > 10%.
While CASAC judged that controlled human exposure studies of respiratory effects
provide the strongest evidence supporting their conclusion on the current standard, the
Committee judged that there is also "sufficient scientific evidence based on epidemiologic
studies for mortality and morbidity associated with short-term exposure to ozone at the level of
the current standard" (Frey, 2014, p. 5). In support of the biological plausibility of the
associations reported in these epidemiologic studies, CASAC noted that "[rjecent animal
toxicological studies support identification of modes of action and, therefore, the biological
plausibility associated with the epidemiological findings" (Frey, 2014, p. 5).
Consistent with the advice of CASAC, several public commenters supported revising the
primary Os standard to provide increased public health protection. In considering the available
evidence as a basis for their views, these commenters generally noted that the health evidence is
stronger in the current review than in past reviews. These commenters often noted that causal
determinations were strengthened to "likely causal" for total mortality and cardiovascular effects
from short-term Os exposures, and for respiratory effects from long-term Os exposures. These
commenters also noted the increase in controlled human exposure studies showing lung function
decrements and new evidence of inflammation in healthy young adults at 60 ppb Os, as well as
the increase in the number of epidemiologic studies showing consistent, positive associations
between Os exposures and hospital admissions, emergency department visits, and premature
mortality. Some commenters noted that children have long been known to be more vulnerable
than adults to the effects of air pollution due to ongoing lung development, the greater
permeability of their airways epithelial layer, and greater resting minute ventilation (when
normalized to body mass or lung volume) resulting in increased exposure compared with adults.
These commenters noted that adverse effects have been described on early lung development and
the evidence for Os as a contributor to childhood respiratory disease is extremely strong. They
3-127
-------
expressed the view that Os in particular has long been known to induce asthma exacerbations in
children and, in one well characterized population-based cohort study in California, exposure to
ozone was associated with the development of asthma. Some commenters expressed the view
that young children and small infants should be included in the exposure and risk assessment.
Other commenters noted that the health endpoints considered in the HREA are limited, and do
not represent the comprehensive array of health effects attributable to Cb exposure.
In contrast to the views discussed above, other public commenters opposed considering
revised standards. These commenters discussed a variety of reasons for their views. A number of
commenters expressed the view that EPA should not lower the level of the standard because a
lower level would be closer to background Os concentrations. In addition, several commenters
challenged the interpretation of the evidence presented in the ISA. For example, some
commenters questioned the ISA's judgments regarding the strength of evidence for
cardiovascular system effects from short-term Cb exposures. With respect to the risk assessment,
several commenters expressed the view that the EPA should only estimate risks above Os
background concentrations, or above threshold concentrations. In some cases these commenters
noted that (1) the Os mode of action indicates that there are thresholds for Os effects; (2) that
these thresholds are considered in the lung function risk assessment; and (3) that there is no
reason to believe that similar thresholds would not also be associated with other health effects,
particularly more serious effects. Some commenters also expressed the view that, based on the
mortality and morbidity risk estimates in the HREA, there is little to no difference between the
risks estimated for the current Os standard and the risks estimated for revised standards with
lower levels. These commenters concluded that the HREA and PA have not shown that the
public health improvements likely to be achieved by a revised Os standard would be greater than
the improvements likely to be achieved by the current standard.
3.4 STAFF CONCLUSIONS ON ADEQUACY OF PRIMARY STANDARD
This section presents staffs conclusions for the Administrator to consider in deciding
whether it is appropriate to revise the existing primary Os standard. Staff conclusions are based
on our consideration of the assessment and integrative synthesis of the evidence presented in the
ISA, the air quality distributions in locations of selected epidemiologic studies, exposure and risk
analyses in the HREA, the advice of CAS AC, and comments received from members of the
public.
As an initial matter, staff concludes that reducing precursor emissions to achieve Os
concentrations that meet the current standard will provide important improvements in public
health protection. This initial conclusion is based on (1) the strong body of scientific evidence
indicating a wide range of adverse health outcomes attributable to exposures to Os
3-128
-------
concentrations commonly found in the ambient air and (2) estimates indicating decreased
occurrences of Os exposures of concern and decreased health risks upon meeting the current
standard, compared to recent air quality.
Strong support for this initial conclusion is provided by controlled human exposure
studies of respiratory effects, and by quantitative estimates of exposures of concern and lung
function decrements based on information in these studies. Analyses in the HREA estimate that
the percentages of children (i.e., all children and children with asthma) in urban case study areas
experiencing exposures of concern, or experiencing abnormal and potentially adverse lung
function decrements, are consistently lower for air quality that just meets the current Os standard
than for recent air quality. The HREA estimates such reductions consistently across the urban
case study areas evaluated and throughout various portions of individual urban case study areas,
including in urban cores and the portions of case study areas surrounding urban cores. These
reductions in exposures of concern and Os-induced lung function decrements reflect the
consistent decreases in the highest Cb concentrations following reductions in precursor emissions
to meet the current standard. Thus, populations in both urban and non-urban areas would be
expected to experience important reductions in Cb exposures and Os-induced lung function risks
upon meeting the current standard.
Support for this initial conclusion is also provided by estimates of Os-associated mortality
and morbidity based on application of concentration-response relationships from epidemiologic
studies to air quality adjusted to just meet the current standard. These estimates, which are based
on the assumption that concentration-response relationships are linear over entire distributions of
ambient Os concentrations, are associated with uncertainties that complicate their interpretation
(discussed below). However, risk estimates for effects associated with short- and long-term Os
exposures, combined with the HREA's national analysis of Os responsiveness to reductions in
precursor emissions and the consistent reductions estimated for the highest ambient Os
concentrations, suggest that Cb-associated mortality and morbidity would be expected to
decrease nationwide following reductions in precursor emissions to meet the current Os standard.
As discussed in section 3.2.3.2, reductions in Cb precursor emissions (i.e., NOx) could
also increase public health protection by reducing the ambient concentrations of pollutants other
than Cb. For example, NOx emissions contribute to ambient NCh, and NOx and VOCs can
contribute to secondary formation of PM2.5 constituents, including ammonium sulfate (NH4SO4),
ammonium nitrate (NHtNOs), and organic carbon (OC). Therefore, at some times and in some
locations, control strategies that would reduce NOx emissions (i.e., to meet an Os standard) could
reduce ambient concentrations of NO2 and PM2.5, resulting in health benefits beyond those
directly associated with reducing ambient Os concentrations.
3-129
-------
We next revisit the overarching policy question for this chapter, taking into consideration
the responses to the specific questions focused on the adequacy of the current primary Os
standard, as discussed above.
• Does the currently available scientific evidence and exposure/risk information,
as reflected in the ISA and HREA, support or call into question the adequacy
of the protection afforded by the current primary Os standard?
In considering the available evidence and information, staff concludes that the Ch-
attributable health effects estimated to be allowed by air quality that meets the current primary
standard for Os can reasonably be judged important from a public health perspective. Thus, we
conclude that the available health evidence and exposure/risk information call into question the
adequacy of the public health protection provided by the current standard. We further conclude
that it is appropriate in this review to consider alternative standards that would increase public
health protection, compared to the current standard. The basis for these conclusions is discussed
below.
Studies evaluated since the completion of the 2006 Os AQCD support and expand upon
the strong body of evidence that, in the last review, indicated a causal relationship between short-
term Os exposures and respiratory health effects. Together, experimental and epidemiologic
studies support conclusions regarding a continuum of Os respiratory effects ranging from small
reversible changes in pulmonary function to more serious effects that can result in respiratory-
related emergency department visits, hospital admissions, and/or mortality. Recent animal
toxicological studies support descriptions of modes of action for these respiratory effects and
augment support for biological plausibility for the role of Os in reported effects. With regard to
mode of action, evidence indicates that antioxidant capacity may modify the risk of respiratory
morbidity associated with Os exposure. In addition, based on the consistency of findings across
studies and evidence for the coherence of results from different scientific disciplines, strong
evidence indicates that certain populations are at increased risk of experiencing Os-related
effects. These include populations and lifestages identified in previous reviews (i.e., people with
asthma, children, older adults, outdoor workers) and populations identified since the last review
(i.e., people with certain genotypes related to anti-oxidant and/or anti-inflammatory status;
people with reduced intake of certain nutrients, such as Vitamins C and E).
Evidence for adverse respiratory health effects attributable to "long-term" or repeated
daily Os exposures is much stronger than in previous reviews, and the ISA concludes that the
evidence supports a "likely to be" causal relationship between such Os exposures and adverse
respiratory health effects. Uncertainties related to the extrapolation of data generated by rodent
toxicology studies to the understanding of health effects in humans have been reduced by studies
3-130
-------
in non-human primates and by recent epidemiologic studies. The evidence available in this
review includes new epidemiologic studies using a variety of designs and analysis methods,
conducted by different research groups in different locations, evaluating the relationships
between long-term Os exposures and measures of respiratory morbidity and mortality. New
evidence supports associations between long-term or repeated Os exposures and the development
of asthma in children, with several studies reporting interactions between genetic variants and
such Os exposures. Studies also report associations between long-term or repeated Os exposure
and asthma prevalence, asthma severity and control, respiratory symptoms among asthmatics,
and respiratory mortality.
In considering the Os exposure concentrations reported to elicit respiratory effects, we
note that controlled human exposure studies provide the most certain evidence indicating the
occurrence of health effects in humans following exposures to specific Os concentrations.
Consistent with this, CASAC also concluded that "the scientific evidence supporting the finding
that the current standard is inadequate to protect public health is strongest based on the
controlled human exposure studies of respiratory effects" (Frey, 2014, p. 5). As discussed above,
recent evidence includes controlled human exposure studies reporting lung function decrements
and pulmonary inflammation in healthy adults engaged in intermittent, moderate exertion
following 6.6 hour exposures to Os concentrations as low as 60 ppb, and lung function
decrements and respiratory symptoms following exposures to concentrations as low as 72 ppb.97
Compared to the evidence available in the last review, these studies have strengthened support
for the occurrence of abnormal and adverse respiratory effects attributable to short-term
exposures to Os concentrations below 80 ppb.98 Consistent with CASAC advice, we conclude
that exposures to such Os concentrations are potentially important from a public health
perspective given the following:
1. The respiratory effects reported to occur in healthy adults following exposures to Os
concentrations of 60 and 72 ppb, while at moderate exertion, can reasonably be judged
adverse based on ATS criteria and advice from CASAC. In considering the 72 ppb
exposure concentration, CASAC noted that "the combination of decrements in FEVi
together with the statistically significant alterations in symptoms in human subjects
exposed to 72 ppb ozone meets the American Thoracic Society's definition of an adverse
health effect" (Frey, 2014, p. 5). With regard to 60 ppb Os, CASAC agreed that "a level
of 60 ppb corresponds to the lowest exposure concentration demonstrated to result in
97 As noted above, for the 70 ppb exposure concentration Schelegle et al. (2009) reported that the actual mean
exposure concentration was 72 ppb.
98 Cf. Misisssippi. 744 F.3d at 1350 ("Perhaps more studies like the Adams studies will yet reveal that the 0.060 ppm
level produces significant adverse decrements that simply cannot be attributed to normal variation in lung
function.").
3-131
-------
lung function decrements large enough to be judged an abnormal response by ATS and
that could be adverse in individuals with lung disease" (Frey, 2014, p. 7). CASAC further
noted that "a level of 60 ppb also corresponds to the lowest exposure concentration at
which pulmonary inflammation has been reported" (Frey, 2014, p. 7).
2. The controlled human exposure studies reporting these respiratory effects were conducted
in healthy adults, while at-risk groups (e.g., children, people with asthma) could
experience larger and/or more serious effects. In their advice to the Administrator,
CASAC concurred with this conclusion (Frey, 2014, p. 5).
3. These respiratory effects are coherent with the serious health outcomes that have been
reported in epidemiologic studies (e.g., respiratory-related hospital admissions,
emergency department visits, and mortality).
Given the above considerations, our conclusions regarding the adequacy of the current primary
Os standard place a large amount of weight on the results of controlled human exposure studies
conducted at 60 and 72 ppb, and on HREA analyses based on information from controlled
human exposure studies (i.e., exposures of concern to Os concentrations at or above 60, 70, and
80 ppb and Os-induced FEVi decrements > 10%, 15%, and 20%).
Recent epidemiologic studies also provide support, beyond that available in the last
review, for associations between short-term Os exposures and a wide range of adverse
respiratory outcomes (including respiratory-related hospital admissions, emergency department
visits, and mortality) and with total mortality. Associations with morbidity and mortality are
stronger during the warm or summer months, and remain robust after adjustment for co-
pollutants. Many epidemiologic studies of morbidity effects and mortality were conducted in
locations that did not meet the current standard. However, in one U.S. single-city study
associations with respiratory morbidity were reported in a location that would likely have met the
current Os standard over the entire study period, suggesting that health effect associations persist
in locations meeting the current standard. In addition, associations with respiratory morbidity or
mortality were reported in several Canadian multicity studies, and in cut point analyses included
in a U.S. multicity study, when the majority of study locations would likely have met the current
Os standard. While there is additional uncertainty in interpreting the relationship between air
quality meeting the current standard and health effects in these multicity studies (i.e., compared
to single-city studies), they provide supporting evidence for the occurrence of health effect
associations in locations that meet the current standard. Even in some study locations where the
current standard was likely not met, considering reported concentration-response functions in the
context of available air quality data support the occurrence of Os-health effect associations on the
subsets of days with ambient Os concentrations below the level of the current standard. Taken
together, these studies and associated air quality data support the occurrence of Os-associated
3-132
-------
hospital admissions, emergency department visits, and mortality at ambient concentrations that
meet the current standard.
Beyond our consideration of the evidence, we also consider the results of the HREA
exposure and risk analyses in reaching conclusions regarding the adequacy of the current
primary Os standard. In doing so, we focus primarily on estimates of the occurrence of exposures
of concern to Os concentrations at or above 60 and 70 ppb and lung function decrements > 10%,
15% and 20%. We place relatively less weight on epidemiologic-based risk estimates, noting that
the overall conclusions from the HREA likewise reflect less confidence in estimates of
epidemiologic-based risks than in estimates of exposures and lung function risks (U.S. EPA,
2014, section 9.6). Our determination to attach less weight to the epidemiologic-based estimates
reflects the uncertainties associated with mortality and morbidity risk estimates, including 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 Os concentrations. The HREA also concludes that lower confidence should be placed in
the results of the assessment of respiratory mortality risks associated with long-term Os
exposures, primarily because that analysis is based on only one study (even though that study is
well-designed) and because of the uncertainty in that study about the existence and level of a
potential threshold in the concentration-response function (U.S. EPA, 2014, section 9.6).
With regard to HREA estimates of exposures of concern we note the CAS AC conclusion
that 60 ppb is an appropriate exposure of concern for asthmatic children (Frey, 2014, p. 8).
Exposure estimates from the HREA indicate that, if the 15 urban case study areas were to just
meet the current Os standard, approximately 10 to 20% of children (on average over the years of
analysis) in those areas, including asthmatic children, could experience one or more exposures of
concern to Os concentrations of 60 ppb or above. In the case study areas evaluated, this
corresponds to over 2 million children (including over 200,000 asthmatic children) experiencing
approximately 4 million such exposures. Nationally, far more children would be expected to
experience such exposures of concern. On average over the years evaluated in the HREA,
approximately 3 to 8% of children are estimated to experience two or more exposures of concern
to Os concentrations of 60 ppb or greater. For the worst-case years in the worst-case locations
(i.e., years and locations with air quality patterns resulting in the largest exposure estimates),
approximately 25% of children are estimated to experience one or more exposures of concern at
or above 60 ppb, and about 14% are estimated to experience two or more such exposures.
Although the current standard more effectively limits exposures of concern at or above higher Os
concentrations (i.e., 70, 80 ppb), we note that in the worst-case year and location about 8% of
children are estimated to experience one or more exposures of concern at or above 70 ppb and
about 2% of children are estimated to experience two or more such exposures.
3-133
-------
Though we focus on children in these analyses of Os exposures, we also recognize that
exposures to 8-hour average Cb concentrations at or above 60, 70, or 80 ppb could be of concern
for adult populations as well. As discussed above, the patterns of exposure estimates over years
and across cities are similar in adult asthmatics, older adults, and children, though smaller
percentages of adult populations are estimated to experience exposures of concern. Thus, the
results for children are one part of a broader range of at-risk populations that also includes
asthmatic adults and older adults.
Consistent with estimates of exposures of concern, the HREA also estimates that under
air quality conditions just meeting the current Os NAAQS, hundreds of thousands of asthmatic
children would be expected to experience Os-induced lung function decrements that are large
enough to be potentially adverse in people with lung disease. On average over the years
evaluated in the HREA, the current standard is estimated to allow about 14% to 19% of children
in the 15 urban case study areas, including asthmatic children, to experience one or more Cb-
induced lung function decrements > 10% (a decrement judged by CASAC to be a "scientifically-
relevant surrogate for adverse health outcomes for people with asthma and lung disease" (Frey,
2014, p. 4)). This corresponds to about 300,000 asthmatic children. Nationally, far more children
would be expected to experience such Os-induced lung function decrements. Across the 15 urban
areas, about 8% to 12% of children are estimated to experience two or more decrements > 10%,
on average over the analysis years. In the worst-case year and location, approximately 22% of
children are estimated to experience one or more decrements > 10% and about 14% are
estimated to experience two or more such decrements. As with exposures of concern, the current
standard more effectively limits the larger Os-induced lung function decrements evaluated (i.e., >
15%, 20%). However, about 7% of children are estimated to experience one or more Os-induced
decrements > 15% in the worst-case city and year analyzed in the HREA, and about 4% are
estimated to experience two or more such decrements. As discussed above, CASAC judged
decrements > 15% to be an appropriate "surrogate for adverse health outcomes in active healthy
adults" (Frey, 2014, p. 4).
As noted above, compared to the weight given to HREA estimates of exposures of
concern and lung function risks, we place relatively less weight on epidemiologic-based risk
estimates. For epidemiology-based risk estimates, we consider total risks (i.e., based on the full
distributions of ambient Os concentrations) and risks associated with Os concentrations in the
upper portions of ambient distributions. A focus on estimates of total risks places greater weight
on the possibility that concentration-response relationships remain linear over the entire
distributions of ambient Os concentrations. With regard to total risks, the HREA estimates
thousands of Os-associated hospital admissions, emergency department visits, and deaths per
year for air quality conditions associated with just meeting the current standard in the 12 urban
3-134
-------
case study areas evaluated. A focus on risks associated with Os concentrations in the upper
portions of ambient distributions places greater weight on the uncertainty associated with the
shapes of concentration-response curves for Os concentrations in the lower portions of ambient
distributions (section 3.2.3.2). Based on area-wide O3 concentrations from the upper portions of
seasonal distributions, the current standard is estimated to allow hundreds to thousands of Os-
associated deaths per year in urban case study areas. As with the exposures of concern and lung
function risks, this number would be much greater if risks were assessed across the entire U.S.
population.
Although we note the HREA conclusions indicating somewhat less confidence in
estimates of Os-associated mortality and morbidity risks compared to estimates of exposures of
concern and lung function risks, we conclude that the general magnitude of mortality and
morbidity risk estimates suggests the potential for a substantial number of Os-associated deaths
and adverse respiratory events nationally when the current standard is met. This is the case even
based on the risks associated with the upper ends of distributions of ambient Os concentrations,
where experimental evidence indicates increasing support for the occurrence of adverse effects
attributable to Cb exposures.
In addition to the evidence and exposure/risk information discussed above, we also take
note of the CAS AC advice provided to the EPA Administrator on the proposed reconsideration
of the 2008 decision establishing the current standard and the advice of CAS AC in the current
review. In commenting on the proposed reconsideration, the prior CASAC Os Panel
recommended revision of the standard to one with a lower level based on the evidence and
information in the record for the 2008 standard (Samet, 2011), which has been substantially
strengthened in the current review. As discussed in more detail above, the current CASAC also
"unanimously recommends that the Administrator revise the current primary ozone standard to
protect public health" (Frey, 2014, p. 6).
In consideration of all of the above, staff reaches the conclusion that the available
evidence and exposure and risk information clearly calls into question the adequacy of public
health protection provided by the current primary standard. The evidence from controlled human
exposure studies provides strong support for the occurrence of adverse respiratory effects
following exposures to Os concentrations below the level of the current standard. Epidemiologic
studies provide support for the occurrence of adverse respiratory effects and mortality under air
quality conditions that would likely meet the current standard. In addition, based on the analyses
in the HREA, we conclude that the exposures and risks projected to remain upon meeting the
current standard are indicative of risks that can reasonably be judged to be important from a
public health perspective. Thus, staff concludes that the evidence and information provides
strong support for giving consideration to revising the current primary standard in order to
3-135
-------
provide increased public health protection against an array of adverse health effects that range
from decreased lung function and respiratory symptoms to more serious indicators of morbidity
(e.g., including emergency department visits and hospital admissions), and mortality. In
consideration of all of the above, staff draws the conclusion that it is appropriate for the
Administrator to consider revision of the current primary Os standard to provide increased public
health protection.
3-136
-------
3.5 REFERENCES
Adams, W. C. (2006) Comparison of chamber 6.6 hour exposures to 0.04-0.08 ppm ozone via square-wave and
triangular profiles on pulmonary responses. Inhalation Toxicol. 18: 127-136.
http://dx.doi.org/10.1080/08958370500306107
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: 265-281.
Adams, WC. (2002). Comparison of chamber and face-mask 6.6-hour exposures to ozone on pulmonary function
and symptoms responses. Inhal Toxicol 14: 745-764. http://dx.doi.org/10.1080/08958370290084610
Adams, WC. (1998). Dose-response effect of varied equivalent minute ventilation rates on pulmonary function
responses during exposure to ozone. Washington, DC: American Petroleum Institute.
Adams, P. F.; Hendershot, G. E.; Marano, M. A. (1999) Current estimates from the National Health Interview
Survey, 1996. Hyattsville, MD: U.S. Department of Health and Human Services, Public Health Service,
National Center for Health Statistics; DHHS publication no. (PHS) 99-1528. (Vital and health statistics: v.
10, data from the National Health Survey, no. 200). Available:
http://www.cdc.gov/nchs/products/pubs/pubd/series/srlO/pre-200/pre-200.htm [12 March, 2001].
Adamson, I; Prieditis, H. (1995). Response of mouse lung to carbon deposition during injury and repair.
Environmental Health Perspectives. 103: 1: 72-76.
Akinbami, LJ; Lynch, CD; Parker, JD; Woodruff, TJ. (2010). The association between childhood asthma prevalence
and monitored air pollutants in metropolitan areas, United States, 2001-2004. Environ Res 110: 294-301.
http://dx.doi.0rg/10.1016/j.envres.2010.01.001
Alexeeff, SE; Litonjua, AA; Suh, H; Sparrow, D; Vokonas, PS; Schwartz, J. (2007). Ozone exposure and lung
function: Effect modified by obesity and airways hyperresponsiveness in the VA Normative Aging Study.
Chest 132: 1890-1897. http://dx.doi.org/10.1378/chest.07-1126
Alexis, NE; Lay, JC; Hazucha, M; Harris, B; Hernandez, ML; Bromberg, PA; Kehrl, H; Diaz-Sanchez, D; Kim, C;
Devlin, RB; Peden, DB. (2010). Low-level ozone exposure induces airways inflammation and modifies cell
surface phenotypes in healthy humans. Inhal Toxicol 22: 593-600.
Alexis, N; Urch, B; Tarlo, S; Corey, P; Pengelly, D; O'Byrne, P; Silverman, F. (2000). Cyclooxygenase metabolites
play a different role in ozone-induced pulmonary function decline in asthmatics compared to normals. Inhal
Toxicol 12: 1205-1224.
Aris, RM; Tager, I; Christian, D; Kelly, T; Balmes, JR. (1995). Methacholine responsiveness is not associated with
Os-induced decreases inFEVi. Chest 107: 621-628.
ATS (American Thoracic Society). (2000). What constitutes an adverse health effect of air pollution? Am. J. Respir.
Crit. Care Med. 161: 665-673.
ATS (American Thoracic Society). (1991). Lung function testing: selection of reference values and interpretive
strategies. Am. Rev. Respir. Dis. 144: 1202-1218.
ATS (American Thoracic Society). (1985) Guidelines as to what constitutes an adverse respiratory health effect,
with special reference to epidemiological studies of air pollution. Am. Rev. Respir. Dis. 131: 666-668.
Avol, EL; Trim, SC; Little, DE; Spier, CE; Smith, MN; Peng, RC; Linn, WS; Hackney, JD; Gross, KB; D'Arcy, JB;
Gibbons, D; Higgins, ITT. (1990). Ozone exposure and lung function in children attending a southern
3-137
-------
California summer camp. In Proceedings of the 83rd A&WMA Annual Meeting. Pittsburgh, PA: Air &
Waste Management Association.
Balmes, JR; Chen, LL; Scannell, C; Tager, I; Christian, D; Hearne, PQ; Kelly, T; Aris, RM. (1996). Ozoneinduced
decrements in FEV1 and FVC do not correlate with measures of inflammation. Am J Respir Crit Care Med
153:904-909.
Barraza-Villarreal, A; Sunyer, J; Hernandez-Cadena, L; Escamilla-Nunez, MC; Sienra-Monge, JJ; Ramirez-Aguilar,
M; Cortez-Lugo, M; Holguin, F; Diaz-Sanchez, D; Olin, AC; Romieu, I. (2008). Air pollution, airway
inflammation, and lung function in a cohort study of Mexico City schoolchildren. Environ Health Perspect
116: 832-838. http://dx.doi.org/10.1289/ehp.10926
Basha, MA; Gross, KB; Gwizdala, CJ; Haidar, AH; Popovich, J, Jr. (1994). Bronchoalveolar lavage neutrophilia in
asthmatic and healthy volunteers after controlled exposure to ozone and filtered purified air. Chest 106:
1757-1765.
Bell, ML; Peng, RD; Dominici, F. (2006). The exposure-response curve for ozone and risk of mortality and the
adequacy of current ozone regulations. Environ Health Perspect 114: 532-536.
Berhane, K; Zhang, Y; Linn, WS; Rappaport, EB; Bastain, TM; Salam, MT; Islam, T; Lurmann, F; Gilliland, FD.
(2011). The effect of ambient air pollution on exhaled nitric oxide in the Children's Health Study. Eur
Respir J 37: 1029-1036. http://dx.doi.org/10.1183/09031936.00081410
Berry, M; Lioy, PJ; Gelperin, K; Buckler, G; Klotz, J. (1991). Accumulated exposure to ozone and measurement of
health effects in children and counselors at two summer camps. Environ Res 54: 135-150.
Blomberg, A; Mudway, IS; Nordenhall, C; Hedenstrom, H; Kelly, FJ; Frew, AJ; Holgate, ST; Sandstrom, T. (1999).
Ozone-induced lung function decrements do not correlate with early airway inflammatory or antioxidant
responses. Eur Respir J 13: 1418-1428.
Bloom B, Cohen RA, Freeman G. (2011). Summary health statistics for U.S. children: National Health Interview
Survey, 2010. National Center for Health Statistics. Vital Health Stat 10(250).
http://www.cdc.gov/nchs/data/series/sr_10/srlO_250.pdf
Bosson, J; Stenfors, N; Bucht, A; Helleday, R; Pourazar, J; Holgate, ST; Kelly, FJ; Sandstrom, T; Wilson, S; Frew,
AJ; Blomberg, A. (2003). Ozone-induced bronchial epithelial cytokine expression differs between healthy
and asthmatic subjects. ClinExp Allergy 33: 777-782.
Brauer, M; Blair, J; Vedal, S. (1996). Effect of ambient ozone exposure on lung function in farm workers. Am J
Respir Crit Care Med 154: 981-987.
Breton, CV; Salam, MT; Vora, H; Gauderman, WJ; Gilliland, FD. (2011). Genetic variation in the glutathione
synthesis pathway, air pollution, and children's lung function growth. Am J Respir Crit Care Med 183: 243-
248. http://dx.doi.org/10.1164/rccm.201006-0849OC
Brown, JS; Bateson, TF; McDonnell, WF. (2008). Effects of exposure to 0.06 ppm ozone on FEVi in humans: A
secondary analysis of existing data. Environ Health Perspect 116: 1023-1026.
http://dx.doi.org/10.1289/ehp. 11396
Brunekreef, B; Hoek, G; Breugelmans, O; Leentvaar, M. (1994). Respiratory effects of low-level photochemical air
pollution in amateur cyclists. Am J Respir Crit Care Med 150: 962-966.
Cakmak, S; Dales, RE; Angelica Rubio, M; Blanco Vidal, C. (2011). The risk of dying on days of higher air
pollution among the socially disadvantaged elderly. EnvironRes 111: 388-393.
http://dx.doi.0rg/10.1016/j.envres.2011.01.003
3-138
-------
Cakmak, S; Dales, RE; Vidal, CB. (2007). Air pollution and mortality in Chile: Susceptibility among the elderly.
Environ Health Perspect 115: 524-527.
Cakmak, S; Dales, RE; Judek, S. (2006). Respiratory health effects of air pollution gases: Modification by education
and income. Arch Environ Occup Health 61: 5-10. http://dx.doi.org/10.3200/AEOH.61.L5-10
Carey, SA; Minard, KR; Trease, LL; Wagner, JG; Garcia, GJ; Ballinger, CA; Kimbell, JS; Plopper, CG; Corley,
RA; Postlethwait, EM; Harkema, JR; Einstein, DR. (2007). Three-dimensional mapping of ozone-induced
injury in the nasal airways of monkeys using magnetic resonance imaging and morphometric techniques.
Toxicol Pathol 35: 27-40. http://dx.doi.org/10.1080/01926230601072343
Castillejos, M; Gold, DR; Damokosh, AI; Serrano, P; Allen, G; McDonnell, WF; Dockery, D; Velasco, SR;
Hernandez, M; Hayes, C. (1995). Acute effects of ozone on the pulmonary function of exercising
schoolchildren from Mexico City. Am J Respir Crit Care Med 152: 1501-1507.
Chan, CC; Wu, TH. (2005). Effects of ambient ozone exposure on mail carriers' peak expiratory flow rates. Environ
Health Perspect 113: 735-738. http://dx.doi.org/10.1289/ehp.7636
Chen, J; Tan, M; Nemmar, A; Song, W; Dong, M; Zhang, G; Li, Y. (2006). Quantification of extrapulmonary
translocation of intratracheal-instilled particles in vivo in rats: effect of lipopolysaccharide. Toxicology
222:195-201.
Chhabra, SK; Yasir, A; Chaudhry, K; Shah, B. (2010). Effect of ozone on response to ovalbumin & its modulation
by vitamins C & E in sensitized guinea pigs. Indian J Med Res 132: 87-93.
Cockcroft, DW; Davis, BE; Todd, DC; Smycniuk, AJ. (2005). Methacholine challenge: Comparison of two
methods. Chest 127: 839-844.
Cohen, MD; Sisco, M; Baker, K; Li, Y; Lawrence, D; VanLoveren, H; Zelikoff, JT; Schlesinger, RB. (2002).
Effects of inhaled ozone on pulmonary immune cells critical to antibacterial responses in situ. Inhal
Toxicol 14: 599-619. http://dx.doi.org/10.1080/08958370290084520
Dales, RE; Cakmak, S; Doiron, MS. (2006). Gaseous air pollutants and hospitalization for respiratory disease in the
neonatal period. Environ Health Perspect 114: 1751-1754. http://dx.doi.org/10.1289/ehp.9044
Delfino, RJ; Staimer, N; Tjoa, T; Arhami, M; Polidori, A; Gillen, DL; George, SC; Shafer, MM; Schauer, JJ;
Sioutas, C. (2010a). Associations of primary and secondary organic aerosols with airway and systemic
inflammation in an elderly panel cohort. Epidemiology 21: 892-902.
http://dx.doi.org/10.1097/EDE.Ob013e3181f20e6c
DeNavas-Walt, Carmen, Bernadette D. Proctor, and Jessica C. Smith. (2011) U.S. Census Bureau, Current
Population Reports, P60-239, Income, Poverty, and Health Insurance Coverage in the United States: 2010,
U.S. Government Printing Office, Washington, DC. http://www.census.gov/prod/201 Ipubs/p60-239.pdf
Depuydt, P; Joos, GF; Pauwels, RA. (1999). Ambient ozone concentrations induce airway hyperresponsiveness in
some rat strains. Eur Respir J 14: 125-131.
Devlin, RB; Mcdonnell, WF; Mann, R; Becker, S; House, DE; Schreinemachers, D; 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: 72-81.
Devlin, RB; Folinsbee, LJ; Biscardi, F; Hatch, G; Becker, S; Madden, MC; Robbins, M; Koren, HS. (1997).
Inflammation and cell damage induced by repeated exposure of humans to ozone. Inhal Toxicol 9: 211-
235.
3-139
-------
Dimeo, MJ; Glenn, MG; Holtzman, MJ; Shelter, JR; Nadel, JA; Boushey, HA. (1981). Threshold concentration of
ozone causing an increase in bronchial reactivity in humans and adaptation with repeated exposures. Am
Rev RespirDis 124: 245-248.
Driscoll, KE; Vollmuth, TA; Schlesinger, RB. (1987). Acute and subchronic ozone inhalation in the rabbit:
Response of alveolar macrophages. J Toxicol Environ Health 21: 27-43.
http://dx.doi.org/10.1080/15287398709531000
Dryden, DM; Spooner, CH; Stickland, MK; Vandermeer, B; Tjosvold, L; Bialy, L; Wong, K; Rowe, BH. (2010).
Exercise-induced bronchoconstriction and asthma. (AHRQ Publication No. 10-E001). Rockville, MD:
Agency for Healthcare Research and Quality.
Escamilla-Nunez, MC; Barraza-Villarreal, A; Hernandez-Cadena, L; Moreno-Macias, H; Ramirez-Aguilar, M;
Sienra-Monge, JJ; Cortez-Lugo, M; Texcalac, JL; del Rio-Navarro, B; Romieu, I. (2008). Traffic-related
air pollution and respiratory symptoms among asthmatic children, resident in Mexico City: The EVA
cohort study. RespirRes 9: 74. http://dx.doi.org/10.1186/1465-9921-9-74
Fanucchi, MV; Plopper, CG; Evans, MJ; Hyde, DM; Van Winkle, LS; Gershwin, LJ; Schelegle, ES. (2006). Cyclic
exposure to ozone alters distal airway development in infant rhesus monkeys. Am J Physiol Lung Cell Mol
Physiol 291: L644-L650. http://dx.doi.org/10.1152/ajplung.00027.2006
Feo Brito, F; Mur Gimeno, P; Martinez, C; Tobias, A; Suarez, L; Guerra, F; Borja, JM; Alonso, AM. (2007). Air
pollution and seasonal asthma during the pollen season: A cohort study in Puertollano and Ciudad Real
(Spain). Allergy 62: 1152-1157. http://dx.doi.0rg/10.llll/i.1398-9995.2007.01438.x
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. J Air Waste Manag Assoc 38: 28-35.
Folinsbee, LJ; Hazucha, MJ. (2000). Time course of response to ozone exposure in healthy adult females. Inhal
Toxicol 12: 151-167.
Folinsbee, LJ; Hazucha, MJ. (1989). Persistence of ozone-induced changes in lung function and airway
responsiveness. In Atmospheric ozone research and its policy implications. Amsterdam, The Netherlands:
Elsevier.
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: 98-105.
Frampton, MW; Morrow, PE; Torres, A; Cox, C; Voter, KZ; Utell, MJ; Gibb, FR; Speers, DM. (1997). Ozone
responsiveness in smokers and nonsmokers. Am J Respir Crit Care Med 155: 116-121.
Franze, T; Weller, MG; Niessner, R; Poschl, U. (2005). Protein nitration by polluted air. Environ Sci Technol 39:
1673-1678. http://dx.doi.org/10.1021/es0488737
Frey, HC. CAS AC Review of the EPA's Second Draft Policy Assessment for the Review of the Ozone National
Ambient Air Quality Standards EPA-CASAC-14-004. June 26, 2014. Available online at:
http://yosemite.epa.gov/sab/sabproduct.nsf/264cbl227d55e02c85257402007446a4/5EFA320CCAD326E8
85257D030071531C/$File/EPA-CASAC-14-004+unsigned.pdf
Gauderman, WJ; Avol, E; Gilliland, F; Vora, H; Thomas, D; Berhane, K; McConnell, R; Kuenzli, N; Lurmann, F;
Rappaport, E; Margolis, H; Bates, D; Peters, J. (2004). The effect of air pollution on lung development
from 10 to 18 years of age. N Engl J Med 351: 1057-1067.
Gershwin, LJ; Osebold, JW; Zee, YC. (1981). Immunoglobulin E-containing cells in mouse lung following allergen
inhalation and ozone exposure. International Arch Allergy Appl Immunol 65: 266-277'.
3-140
-------
Gielen, MH; Van Der Zee, SC; Van Wijnen, JH; Van Steen, CJ; Brunekreef, B. (1997). Acute effects of summer air
pollution on respiratory health of asthmatic children. Am J Respir Crit Care Med 155: 2105-2108.
Gold, DR; Damokosh, AI; III, PC; Dockery, DW; McDonnell, WF; Serrano, P; Retama, A; Castillejos, M. (1999).
Paniculate and ozone pollutant effects on the respiratory function of children in southwest Mexico City.
Epidemiology 10: 8-16.
Gong, H, Jr; McManus, MS; Linn, WS. (1997). Attenuated response to repeated daily ozone exposures in asthmatic
subjects. Arch Environ Occup Health 52: 34-41.
Greer, JR; Abbey, DE; Burchette, RJ. (1993). Asthma related to occupational and ambient air pollutants in
nonsmokers. J Occup Environ Med 35: 909-915.
Harkema, JR; Plopper, CG; Hyde, DM; St George, JA; Dungworth, DL. (1987). Effects of an ambient level of ozone
on primate nasal epithelial mucosubstances: quantitative histochemistry. Am J Pathol 127: 90-96.
Hatch, GE; Slade, R; Harris, LP; Mcdonnell, WF; Devlin, RB; Koren, HS; Costa, DL; Mckee, J. (1994). Ozone dose
and effect in humans and rats: A comparison using oxygen-18 labeling and bronchoalveolar lavage. Am J
Respir Crit Care Med 150: 676-683.
Henderson, R. (2008) Letter from CASAC Chairman Rogene Henderson to EPA Administrator Stephen Johnson.
April 7, 2008, EPA-CASAC-08-001.
Henderson, R. (2006) Letter from CASAC Chairman Rogene Henderson to EPA Administrator Stephen Johnson.
October 24, 2006, EPA-CASAC-07-001.
Hernandez, ML; Lay, JC; Harris, B; Esther, CR; Brickey, WJ; Bromberg, PA; Diaz-Sanchez, D; Devlin, RB;
Kleeberger, SR; Alexis, NE; Peden, DB. (2010). Atopic asthmatic subjects but not atopic subjects without
asthma have enhanced inflammatory response to ozone. J Allergy Clin Immunol 126: 537-544.
http://dx.doi.0rg/10.1016/j.jaci.2010.06.043
Hernandez-Cadena, L; Holguin, F; Barraza-Villarreal, A; Del Rio-Navarro, BE; Sienra-Monge, JJ; Romieu, I.
(2009). Increased levels of outdoor air pollutants are associated with reduced bronchodilation in children
with asthma. Chest 136: 1529-1536. http://dx.doi.org/10.1378/chest.08-1463
Halonen, JI; Lanki, T; Tiittanen, P; Niemi, JV; Loh, M; Pekkanen, J. (2009). Ozone and cause-specific
cardiorespiratory morbidity and mortality. J Epidemiol Community Health 64: 814-820.
http://dx.doi.org/10.1136/jech.2009.087106
Hiltermann, JTN; Lapperre, TS; VanBree, L; Steerenberg, PA; Brahim, JJ; Sont, JK; Sterk, PJ; Hiemstra, PS; 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 Radic Biol Med 27: 1448-
1454.
Hoek, G; Brunekreef, B; Kosterink, P; Van den Berg, R; Hofschreuder, P. (1993). Effect of ambient ozone on peak
expiratory flow of exercising children in the Netherlands. Arch Environ Occup Health 48: 27-32.
http://dx.doi.org/10.1080/00039896.1993.9938390
Holz, O; Jorres, RA; Timm, P; Mucke, M; Richter, K; Koschyk, S; Magnussen, H. (1999). Ozone-induced airway
inflammatory changes differ between individuals and are reproducible. Am J Respir Crit Care Med 159:
776-784.
Holz, O; Mucke, M; Paasch, K; Bohme, S; Timm, P; Richter, K; Magnussen, H; Jorres, RA. (2002). Repeated ozone
exposures enhance bronchial allergen responses in subjects with rhinitis or asthma. Clin Exp Allergy 32:
681-689.
3-141
-------
Holz, O; Tal-Singer, R; Karmiess, F; Simpson, KJ; Gibson, A; Vessey, RSJ; Janicki, S; Magnussen, H; Torres, RA;
Richter, K. (2005). Validation of the human ozone challenge model as a tool for assessing
antiinflammatory drugs in early development. J Clin Pharmacol 45: 498-503.
Hoppe, P; Peters, A; Rabe, G; Praml, G; Lindner, J; Jakobi, G; Fruhmann, G; Nowak, D. (2003). Environmental
ozone effects in different population subgroups. Int J Hyg Environ Health 206: 505-516.
http://dx.doi.org/10.1078/1438-4639-00250
Horstman, DH; Ball, BA; Brown, J; Gerrity, T; 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: 369-385.
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 J Respir Crit Care Med 142: 1158-1163.
Howden, Lindsay M. and Meyer, Julie A. (2011). U.S. Census Bureau, 2010 Census Briefs, C2010BR-03, Age and
Sex Composition: 2010, U.S. Department of Commerce, Economics and Statistics Administration, U.S.
Census Bureau, Washington, DC 20233. http://www.census.gov/prod/cen2010/briefs/c2010br-03.pdf
Hurst, DJ; Gardner, DE; Coffin, DL. (1970). Effect of ozone on acid hydrolases of the pulmonary alveolar
macrophage. J Reticuloendothel Soc 8: 288-300.
Hwang, BF; Lee, YL; Lin, YC; Jaakkola, JJK; Guo, YL. (2005). Traffic related air pollution as a determinant of
asthma among Taiwanese school children. Thorax 60: 467-473.
Islam, T; Berhane, K; McConnell, R; Gauderman, WJ; Avol, E; Peters, JM; Gilliland, FD. (2009). Glutathione-S-
transferase (GST) PI, GSTM1, exercise, ozone and asthma incidence in school children. Thorax 64: 197-
202. http://dx.doi.org/10.1136/thx.2008.099366
Islam, T; McConnell, R; Gauderman, WJ; Avol, E; Peters, JM; Gilliland, FD. (2008). Ozone, oxidant defense genes
and risk of asthma during adolescence. Am J Respir Crit Care Med 177: 388-395.
http://dx.doi.org/10.1164/rccm.200706-863OC
Islam, T; Gauderman, WJ; Berhane, K; McConnell, R; Avol, E; Peters, JM; Gilliland, FD. (2007). The relationship
between air pollution, lung function and asthma in adolescents. Thorax 62: 957-963.
Ito, K; Thurston, GD; Silverman, RA. (2007). Characterization of PM2.5, gaseous pollutants, and meteorological
interactions in the context of time-series health effects models. J Expo Sci Environ Epidemiol 17: S45-S60.
Jacquemin, B; Kauffmann, F; Pin, I; Le Moual, N; Bousquet, J; Gormand, F; Just, J; Nadif, R; Pison, C; Vervloet,
D; Ktinzli, N; Siroux, V. (In Press) Air pollution and asthma control in the epidemiological study on the
genetics and environment of asthma. J Epidemiol Community Health.
http://dx.doi.org/10.1136/jech.2010.130229
Jalaludin, BB; O'Toole, BI; Leeder, S. R. (2004). Acute effects of urban ambient air pollution on respiratory
symptoms, asthma medication use, and doctor visits for asthma in a cohort of Australian children. Environ
Res 95: 32-42. http://dx.doi.org/10.1016/S0013-9351 (03)00038-0
Jerrett, M; Burnett, RT; Pope, CA, III; Ito, K; Thurston, G; Krewski, D; Shi, Y; Calle, E; Thun, M. (2009). Long-
term ozone exposure and mortality. N Engl J Med 360: 1085-1095.
http://dx.doi.org/10.1056/NEJMoa0803894
3-142
-------
Joad, JP; Kott, KS; Brie, JM; Peake, JL; Plopper, CG; Schelegle, ES; Gershwin, LJ; Pinkerton, KE. (2006).
Structural and functional localization of airway effects from episodic exposure of infant monkeys to
allergen and/or ozone. Toxicol Appl Pharmacol 214: 237-243.
Johnston, RA; Schwartzman, IN; Flynt, L; Shore, SA. (2005). Role of interleukin-6 in murine airway responses to
ozone. Am J Physiol Lung Cell Mol Physiol 288: L390-L397
Torres, R; Nowak, D; Magnussen, H; Speckin, P; Koschyk, S. (1996). The effect of ozone exposure on allergen
responsiveness in subjects with asthma or rhinitis. Am J Respir Crit Care Med 153: 56-64.
Just, J; Segala, C; Sahraoui, F; Priol, G; Grimfeld, A; Neukirch, F. (2002). Short-term health effects of paniculate
and photochemical air pollution in asthmatic children. Eur Respir J 20: 899-906.
http://dx.doi.org/10.1183/09031936.02.00236902
Kan, H; London, SJ; Chen, G; Zhang, Y; Song, G; Zhao, N; Jiang, L; Chen, B. (2008). Season, sex, age, and
education as modifiers of the effects of outdoor air pollution on daily mortality in Shanghai, China: The
Public Health and Air Pollution in Asia (PAPA) Study. Environ Health Perspect 116: 1183-1188.
http://dx.doi.org/10.1289/ehp. 10851
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; Zanobetti, A. (2009). Air
pollution and health: A European and North American approach (APHENA). (Research Report 142).
Boston, MA: Health Effects Institute. http://pubs.healtheffects.org/view.php?id=327
Kehrl, HR; Peden, DB; Ball, BA; Folinsbee, LJ; Horstman, DH. (1999). Increased specific airway reactivity of
persons with mild allergic asthma after 7.6 hours of exposure to 0.16 ppm ozone. J Allergy Clin Immunol
104: 1198-1204.
Khatri, SB; Holguin, FC; Ryan, PB; Mannino, D; Erzurum, SC; Teague, WG. (2009). Association of ambient ozone
exposure with airway inflammation and allergy in adults with asthma. J Asthma 46: 777-785.
http://dx.doi.org/10.1080/02770900902779284
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: 1215-1221.
http://dx.doi.org/10.1164/rccm.201011-1813OC
Koren, HS, RB Devlin, DE Graham; R Mann; MP Mcgee; DH Horstman; WJ Kozumbo; S Becker; DE House; WF
McDonnell; PA Bromberg. (1989). Ozone-induced inflammation in the lower airways of human subjects.
Am. Rev. Respir. Dis. 139:407-415.
Korrick, SA; Neas, LM; Dockery, DW; Gold, DR; Allen, GA; Hill, LB; Kimball, KD; Rosner, BA; Speizer, FE.
(1998). Effects of ozone and other pollutants on the pulmonary function of adult hikers. Environ Health
Perspect 106: 93-99. http://dx.doi.org/10.1289/ehp.9810693
Kreit, JW; Gross, KB; Moore, TB; Lorenzen, TJ; D'Arcy, J; Eschenbacher, WL. (1989). Ozone-induced changes in
pulmonary function and bronchial responsiveness in asthmatics. J Appl Physiol 66: 217-222.
Kulle, TJ; Sauder, LR; Kerr, HD; Farrell, BP; Bermel, MS; Smith, DM. (1982). Duration of pulmonary function
adaptation to ozone in humans. Am Ind Hyg Assoc J 43: 832-837.
Ktinzli, N. (2012). Is air pollution of the 20th century a cause of current asthma hospitalisations? [Editorial]. Thorax
67: 2-3. http://dx.doi.org/10.1136/thoraxjnl-2011-200919
3-143
-------
Larsen, ST; Matsubara, S; Mcconville, G; Poulsen, SS; Gelfand, EW. (2010). Ozone increases airway
hyperreactivity and mucus hyperproduction in mice previously exposed to allergen. J Toxicol Environ
HealthA73: 738-747. http://dx.doi.org/10.1080/15287391003614034
Lee, YL; McConnell, R; Berhane, K; Gilliland, FD. (2009). Ambient ozone modifies the effect of tumor necrosis
factor G-308A onbronchitic symptoms among children with asthma. Allergy 64: 1342-1348.
http://dx.doi.org/10.1111/j. 1398-9995.2009.02014.X
Lewis, TC; Robins, TG; Dvonch, JT; Keeler, GJ; Yip, FY; Mentz, GB; Lin, X; Parker, EA; Israel, BA; Gonzalez, L;
Hill, Y. (2005). Air pollution-associated changes in lung function among asthmatic children in Detroit.
Environ Health Perspect 113: 1068-1075.
Lin, S; Bell, EM; Liu, W; Walker, RJ; Kim, NK; Hwang, SA. (2008a). Ambient ozone concentration and hospital
admissions due to childhood respiratory diseases in New York State, 1991-2001. Environ Res 108: 42-47.
http://dx.doi.0rg/10.1016/j.envres.2008.06.007
Lin, S; Liu, X; Le, LH; Hwang, SA. (2008b). Chronic exposure to ambient ozone and asthma hospital admissions
among children. Environ Health Perspect 116: 1725-1730. http://dx.doi.org/10.1289/ehp. 11184
Linn, WS; Shamoo, DA; Anderson, KR; Peng, RC; Avol, EL; Hackney, JD; Gong, H, Jr. (1996). Short-term air
pollution exposures and responses in Los Angeles area schoolchildren. J Expo Sci Environ Epidemiol 6:
449-472.
Liu, L; Poon, R; Chen, L; Frescura, AM; Montuschi, P; Ciabattoni, G; Wheeler, A; Dales, R. (2009). Acute effects
of air pollution on pulmonary function, airway inflammation, and oxidative stress in asthmatic children.
Environ Health Perspect 117: 668-674. http://dx.doi.org/10.1289/ehpll
Mar, TF; Koenig, JQ. (2009). Relationship between visits to emergency departments for asthma and ozone exposure
in greater Seattle, Washington. Ann Allergy Asthma Immunol 103: 474-479.
McConnell, R; Berhane, K; Gilliland, F; London, SJ; Islam, T; Gauderman, WJ; Avol, E; Margolis, HG; Peters, JM.
(2002). Asthma in exercising children exposed to ozone: A cohort study. Lancet 359: 386-391.
McConnell, R; Islam, T; Shankardass, K; Jerrett, M; Lurmann, F; Gilliland, F; Gauderman, J; Avol, E; Kuenzli, N;
Yao, L; Peters, J; Berhane, K. (2010). Childhood incident asthma and traffic-related air pollution at home
and school. Environ Health Perspect 118: 1021-1026.
McDonnell, WF. (1996). Individual variability in human lung function responses to ozone exposure. Environ
Toxicol Pharmacol 2: 171-175.
McDonnell, WF; Abbey, DE; Nishino, N; Lebowitz, MD. (1999). Long-term ambient ozone concentration and the
incidence of asthma in nonsmoking adults: the Ahsmog study. Environ Res 80: 110-121.
McDonnell, WF; Chapman, RS; Horstman, DH; Leigh, MW; Abdul-Salaam, S. (1985). A comparison of the
responses of children and adults to acute ozone exposure.
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 ozone for 6.6 hours. Arch Environ Occup
Health 46: 145-150.
McDonnell, WF; Stewart, PW; Smith, MV; Kim, CS; Schelegle, ES. (2012). Prediction of lung function response
for populations exposed to a wide range of ozone conditions. Inhal Toxicol 24: 619-633.
Medina-Ramon, M; Schwartz, J. (2008). Who is more vulnerable to die from ozone air pollution? Epidemiology 19:
672-679.
3-144
-------
Medina-Ramon, M; Zanobetti, A; Schwartz, J. (2006). The effect of ozone and PM10 on hospital admissions for
pneumonia and chronic obstructive pulmonary disease: A national multicity study. Am J Epidemiol 163:
579-588. http://dx.doi.org/10.1093/aje/kwj078
Meng, YY; Hull, RP; Wilhelm, M; Lombardi, C; Balmes, J; Ritz, B. (2010). Outdoor air pollution and uncontrolled
asthma in the San Joaquin Valley, California. J Epidemiol Community Health 64: 142-147.
http://dx.doi.org/10.1136/jech.2008.083576
Middleton, N; Yiallouros, P; Kleanthous, S; Kolokotroni, O; Schwartz, J; Dockery, DW; Demokritou, P; Koutrakis,
P. (2008). A 10-year time-series analysis of respiratory and cardiovascular morbidity in Nicosia, Cyprus:
The effect of short-term changes in air pollution and dust storms. Environ Health 7:39.
Moore, K; Neugebauer, R; Lurmann, F; Hall, J; Brajer, V; Alcorn, S; Tager, I. (2008). Ambient ozone
concentrations cause increased hospitalizations for asthma in children: An 18-year study in Southern
California. Environ Health Perspect 116: 1063-1070. http://dx.doi.org/10.1289/ehp.10497
Mortimer, KM; Neas, LM; Dockery, DW; Redline, S; Tager, IB. (2002). The effect of air pollution on inner-city
children with asthma. EurRespir J 19: 699-705. http://dx.doi.org/10.1183/09031936.02.00247102
Mudway, IS; Kelly, FJ. (2000). Ozone and the lung: A sensitive issue. Mol Aspects Med 21: 1-48.
Mudway, IS; Stenfors, N; Blomberg, A; Helleday, R; Dunster, C; Marklund, SL; Frew, AJ; Sandstrom, T; 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 Radic Biol Med 31:
962-974.
Mudway, IS; Kelly, FJ. (2004). An investigation of inhaled ozone dose and the magnitude of airway inflammation in
healthy adults. Am J Respir Crit Care Med 169: 1089-1095.
Neidell, M. (2009). Information, avoidance behavior, and health: The effect of ozone on asthma hospitalizations.
Journal of Human Resources 44: 450-478.
Neidell, M; Kinney, PL. (2010). Estimates of the association between ozone and asthma hospitalizations that
account for behavioral responses to air quality information. Environ Sci Pol 13: 97-103.
http://dx.doi.0rg/10.1016/j.envsci.2009.12.006
NHLBI (National Institutes of Health, National Heart Lung and Blood Institute). (2007). Expert panel report 3:
guidelines for the diagnosis and management of asthma. (07-4051). Bethesda, MD: National Institute of
Health.
NRC (2008). Estimating Mortality Risk Reduction and Economic Benefits from Controlling Ozone Air Pollution.
Washington, DC: The National Academies Press.
O'Connor, GT; Neas, L; Vaughn, B; Kattan, M; Mitchell, H; Crain, EF; III, ER; Gruchalla, R; Morgan, W; Stout, J;
Adams, GK; Lippmann, M. (2008). Acute respiratory health effects of air pollution on children with asthma
in U.S. inner cities. J Allergy Clinlmmunol 121: 1133-1139. http://dx.doi.org/10.1016/jjaci.2008.02.020
O*Net OnLine (2012). Work Context — Outdoors, Exposed to Weather.
http://www.onetonline.Org/find/descriptor/result/4.C.2.a.l.c?a=l
Ostro, B; Lipsett, M; Mann, J; Braxton-Owens, H; White, M. (2001). Air pollution and exacerbation of asthma in
African-American children in Los Angeles. Epidemiology 12: 200-208.
Peden, DB; Boehlecke, B; Horstman, D; Devlin, R. (1997). Prolonged acute exposure to 0.16 ppm ozone induces
eosinophilic airway inflammation in asthmatic subjects with allergies. J Allergy Clin Immunol 100: 802-
808.
3-145
-------
Peden, DB; Setzer, RW, Jr; 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: 1336-1345.
Pellegrino, R; Viegi, G; Brusasco, V; Crapo, RO; Burgos, F; Casaburi, R; Coates, A; van der Grinten, CP;
Gustafsson, P; Hankinson, J; Jensen, R; Johnson, DC; Maclntyre, N; McKay, R; Miller, MR; Navajas, D;
Pedersen, OF; Wanger, J. (2005). Interpretative strategies for lung function tests. Eur Respir J 26: 948-968.
http://dx.doi.org/10.1183/09031936.05.00035205
Plopper, CG; Smiley-Jewell, SM; Miller, LA; Fanucchi, MV; Evans, MJ; Buckpitt, AR; Avdalovic, M; Gershwin,
LJ; Joad, JP; Kajekar, R; Larson, S; Pinkerton, KE; Van Winkle, LS; Schelegle, ES; Pieczarka, EM; Wu,
R; Hyde, DM. (2007). Asthma/allergic airways disease: Does postnatal exposure to environmental
toxicants promote airway pathobiology? Toxicol Pathol 35: 97-110.
http://dx.doi.org/10.1080/01926230601132030
Pope, CA, III; Burnett, RT; Thun, MJ; Calle, EE; Krewski, D; Ito, K; Thurston, GD. (2002). Lung cancer,
cardiopulmonary mortality, and long-term exposure to fine paniculate air pollution. JAMA 287: 1132-
1141.
Qian, Z; Lin, HM; Chinchilli, VM; Lehman, EB; Duan, Y; Craig, TJ; Wilson, WE; Liao, D; Lazarus, SC; Bascom,
R. (2009). Interaction of ambient air pollution with asthma medication on exhaled nitric oxide among
asthmatics. Arch Environ Occup Health 64: 168-176. http://dx.doi.org/10.1080/19338240903240616
Que, LG; Stiles, JV; Sundy, JS; Foster, WM. (2011). Pulmonary function, bronchial reactivity, and epithelial
permeability are response phenotypes to ozone and develop differentially in healthy humans. J Appl
Physiol 111: 679-687. http://dx.doi.org/10.1152/japplphysiol.00337.2011
Rabinovitch, N; Zhang, LN; Murphy, JR; Vedal, S; Dutton, SJ; Gelfand, EW. (2004). Effects of wintertime ambient
air pollutants on asthma exacerbations in urban minority children with moderate to severe disease. J
Allergy Clinlmmunol 114: 1131-1137. http://dx.doi.0rg/10.1016/j.jaci.2004.08.026
Rage, E; Siroux, V; Kunzli, N; Pin, I; Kauffmann, F. (2009). Air pollution and asthma severity in adults. Occup
EnvironMed66: 182-188. http://dx.doi.org/10.1136/oem.2007.038349
Riediker, M; Monn, C; Koller, T; Stahel, WA; Wuthrich, B. (2001). Air pollutants enhance rhinoconjunctivitis
symptoms in pollen-allergic individuals. Ann Allergy Asthma Immunol 87: 311-318.
http://dx.doi.org/10.1016/S1081-1206 (10)62246-6
Rojas-Martinez, R; Perez-Padilla, R; Olaiz-Fernandez, G; Mendoza-Alvarado, L; Moreno-Macias, H; Fortoul, T;
Mcdonnell, W; Loomis, D; Romieu, I. (2007). Lung function growth in children with long-term exposure
to air pollutants in Mexico City. Am J Respir Crit Care Med 176: 377-384.
Romieu, I; Meneses, F; Ruiz, S; Huerta, J; Sienra, JJ; White, M; Etzel, R; Hernandez, M. (1997). Effects of
intermittent ozone exposure on peak expiratory flow and respiratory symptoms among asthmatic children
in Mexico City. Arch Environ Occup Health 52: 368-376.
Romieu, I; Meneses, F; Ruiz, S; Sienra, JJ; Huerta, J; White, MC; Etzel, RA. (1996). Effects of air pollution of the
respiratory health of asthmatic children living in Mexico City. Am J Respir Crit Care Med 154: 300-307.
Romieu, I; Ramirez-Aguilar, M; Sienra-Monge, JJ; Moreno-Macias, H; Del Rio-Navarro, BE; David, G; Marzec, J;
Hernandez-Avila, M; London, S. (2006). GSTM1 and GSTP1 and respiratory health in asthmatic children
exposed to ozone. Eur Respir J28: 953-959. http://dx.doi.org/10.1183/09031936.06.00114905
3-146
-------
Romieu, I; Barraza-Villarreal, A; Escamilla-Nunez, C; Texcalac-Sangrador, JL; Hernandez-Cadena, L; Diaz-
Sanchez, D; De Batlle, J; Del Rio-Navarro, BE. (2009). Dietary intake, lung function and airway
inflammation in Mexico City school children exposed to air pollutants. Respir Res 10: 122.
Romieu, I; Barraza-Villarreal, A; Escamilla-Nunez, C; Almstrand, AC; Diaz-Sanchez, D; Sly, PD; Olin, AC. (2008).
Exhaled breath malondialdehyde as a marker of effect of exposure to air pollution in children with asthma.
J Allergy Clinlmmunol 121: 903-909. http://dx.doi.org/10.1016/jjaci.2007.12.004
Romieu, I; Sienra-Monge, JJ; Ramirez-Aguilar, M; Tellez-Rojo, MM; Moreno-Macias, H; Reyes-Ruiz, NI; Del Rio-
Navarro, BE; Ruiz-Navarro, MX; Hatch, G; Slade, R; Hernandez-Avila, M. (2002). Antioxidant
supplementation and lung functions among children with asthma exposed to high levels of air pollutants.
Am J Respir Crit Care Med 166: 703-709.
Ross, MA; Persky, VW; Scheff, PA; Chung, J; Curtis, L; Ramakrishnan, V; Wadden, RA; Hryhorczuk, DO. (2002).
Effect of ozone and aeroallergens on the respiratory health of asthmatics. Arch Environ Occup Health 57:
568-578. http://dx.doi.org/10.1080/00039890209602090
Salam, MT; Islam, T; Gauderman, WJ; Gilliland, FD. (2009). Roles of arginase variants, atopy, and ozone in
childhood asthma. J Allergy Clinlmmunol 123: 596-602. http://dx.doi.org/10.1016/jjaci.2008.12.020
Samet, J.M. (2011). Clean Air Scientific Advisory Committee (CASAC) Response to Charge Questions on the
Reconsideration of the 2008 Ozone National Ambient Air Quality Standards. EPA-CAS AC-11-004.
March 30, 2011. Available online at:
http://yosemite.epa.gov/sab/sabproduct.nsf/0/F08BEB48C1139E2A8525785E006909AC/$File/EPA-
CASAC-1 l-004-unsigned+.pdf
Samoli, E; Zanobetti, A; Schwartz, J; Atkinson, R; Le Tertre, A; Schindler, C; Perez, L; Cadum, E; Pekkanen, J;
Paldy, A; Touloumi, G; Katsouyanni, K. (2009). The temporal pattern of mortality responses to ambient
ozone in the APHEA project. J Epidemiol Community Health 63: 960-966.
http://dx.doi.org/10.1136/jech.2008.084012
Sasser, E. (2014) Memo Responding to Request for Revised Ozone HREA Chapter 7 Appendix Tables. May 9,
2014. Available at http://www.epa.gOv/ttn/naaqs/standards/ozone/s o3 2008 rea.html
Scannell, C; Chen, L; Aris, RM; Tager, I; Christian, D; Ferrando, R; Welch, B; Kelly, T; Balmes, JR. (1996).
Greater ozone-induced inflammatory responses in subjects with asthma. Am J Respir Crit Care Med 154:
24-29.
Schelegle, ES; Adams, WC; Walby, WF; Marion, MS. (2012). Modelling of individual subject ozone exposure
response kinetics. Inhal Toxicol 24: 401-415. http://dx.doi.org/10.3109/08958378.2012.683891
Schelegle, ES; Miller, LA; Gershwin, LJ; Fanucchi, MV; Van Winkle, LS; Gerriets, JE; Walby, WF; Mitchell, V;
Tarkington, BK; Wong, VJ; Baker, GL; Pantle, LM; Joad, JP; Pinkerton, KE; Wu, R; Evans, MJ; Hyde,
DM; Plopper, CG. (2003). Repeated episodes of ozone inhalation amplifies the effects of allergen
sensitization and inhalation on airway immune and structural development in Rhesus monkeys. Toxicol
Appl Pharmacol 191: 74-85.
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: 265-272.
http://dx.doi.org/10.1164/rccm.200809-1484OC
Schelegle, ES; Siefkin, AD; McDonald, RJ. (1991). Time course of ozone-induced neutrophilia in normal humans.
Am J Respir Crit Care Med 143: 1353-1358.
3-147
-------
Schildcrout, JS; Sheppard, L; Lumley, T; Slaughter, JC; Koenig, JQ; Shapiro, GG. (2006). Ambient air pollution
and asthma exacerbations in children: An eight-city analysis. Am J Epidemiol 164: 505-517.
http://dx.doi.org/10.1093/aje/kwj225
Schiller JS, Lucas JW, Ward BW, Peregoy JA. (2012). Summary health statistics for U.S. adults: National Health
Interview Survey, 2010. National Center for Health Statistics. Vital Health Stat 10(252).
http://www.cdc.gov/nchs/data/series/sr_10/srlO_252.pdf
Sienra-Monge, JJ; Ramirez-Aguilar, M; Moreno-Macias, H; Reyes-Ruiz, NI; Del Rio-Navarro, BE; Ruiz-Navarro,
MX; Hatch, G; Crissman, K; Slade, R; Devlin, RB; Romieu, I. (2004). Antioxidant supplementation and
nasal inflammatory responses among young asthmatics exposed to high levels of ozone. Clin Exp Immunol
138: 317-322. http://dx.doi.0rg/10.llll/j.1365-2249.2004.02606.x
Silverman, RA; Ito, K. (2010). Age-related association of fine particles and ozone with severe acute asthma in New
York City. J Allergy Clin Immunol 125: 367-373. http://dx.doi.org/10.1016/jjaci.2009.10.061
Smith, RL; Xu, B; Switzer, P. (2009). Reassessing the relationship between ozone and short-term mortality in U.S.
urban communities. InhalToxicol21: 37-61. http://dx.doi.org/10.1080/08958370903161612
Spektor, DM; Lippmann, M; Lioy, PJ; Thurston, GD; Citak, K; James, DJ; Bock, N; Speizer, FE; Hayes, C.
(1988a). Effects of ambient ozone on respiratory function in active, normal children. Am Rev Respir Dis
137:313-320.
Spektor, DM; Lippmann, M; Thurston, GD; Lioy, PJ; Stecko, J; O'Connor, G; Garshick, E; Speizer, FE; Hayes, C.
(1988b). Effects of ambient ozone on respiratory function in healthy adults exercising outdoors. Am Rev
Respir Dis 138: 821-828.
SSDAN CensusScope (Social Science Data Analysis Network, CensusScope). (2010). United States: Age
distribution [Database]. Ann Arbor, Michigan: Social Science Data Analysis Network. Retrieved from
http://www.censusscope.org/us/chart_age.html
Stafoggia, M; Forastiere, F; Faustini, A; Biggeri, A; Bisanti, L; Cadum, E; Cernigliaro, A; Mallone, S; Pandolfi, P;
Serinelli, M; Tessari, R; Vigotti, MA; Perucci, CA. (2010). Susceptibility factors to ozone-related
mortality: A population-based case-crossover analysis. Am J Respir Crit Care Med 182: 376-384.
http://dx.doi.org/10.1164/rccm.200908-1269OC
Stenfors, N; Pourazar, J; Blomberg, A; Krishna, MT; Mudway, I; Helleday, R; Kelly, FJ; Frew, AJ; Sandstrom, T.
(2002). Effect of ozone on bronchial mucosal inflammation in asthmatic and healthy subjects. Respir Med
96: 352-358.
Stieb, DM; Szyszkowicz, M; Rowe, BH; Leech, JA. (2009). Air pollution and emergency department visits for
cardiac and respiratory conditions: A multi-city time-series analysis. Environ Health Global Access Sci
Source 8: 25. http://dx.doi.org/10.1186/1476-069X-8-25
Strickland, MJ; Darrow, LA; Klein, M; Flanders, WD; Sarnat, JA; Waller, LA; Sarnat, SE; Mulholland, JA; Tolbert,
PE. (2010). Short-term associations between ambient air pollutants and pediatric asthma emergency
department visits. Am J Respir Crit Care Med 182: 307-316. http://dx.doi.org/10.1164/rccm.200908-
1201OC
Stylianou, M; Nicolich, MJ. (2009). Cumulative effects and threshold levels in air pollution mortality: Data analysis
of nine large US cities using the NMMAPS dataset. Environ Pollut 157: 2216-2223.
http://dx.doi.0rg/10.1016/j.envpol.2009.04.011
3-148
-------
Thaller, El; Petronella, SA; Hochman, D; Howard, S; Chhikara, RS; Brooks, EG. (2008). Moderate increases in
ambient PM2.5 and ozone are associated with lung function decreases in beach lifeguards. J Occup Environ
Med 50: 202-211. http://dx.doi.org/10.1097/JOM.Ob013e31816386b4
Thurston, GD; Lippmann, M; Scott, MB; Fine, JM. (1997). Summertime haze air pollution and children with
asthma. Am J Respir Crit Care Med 155: 654-660.
Tolbert, PE; Klein, M; Peel, JL; Sarnat, SE; 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: S29-S35.
http://dx.doi.org/10.1038/sj.jes.7500625
Torres, A; Utell, MJ; Morow, PE; Voter, KZ; Whitin, JC; Cox, C; Looney, RJ; Speers, DM; Tsai, Y; Frampton,
MW. (1997). Airway inflammation in smokers and nonsmokers with varying responsiveness to ozone. Am
J Respir Crit Care Med 156: 728-736.
Trenga, CA; Koenig, JQ; Williams, PV. (2001). Dietary antioxidants and ozone-induced bronchial
hyperresponsiveness in adults with asthma. Arch Environ Occup Health 56: 242-249.
U.S. EPA (1996). Air quality criteria for ozone and related photochemical oxidants [EPA Report]. (EPA/600/P-
93/004aF, cF). U.S. Environmental Protection Agency, Research Triangle Park, NC.
U.S. EPA (2006). Air Quality Criteria for Ozone and Related Photochemical Oxidants (2006 Final). U.S.
Environmental Protection Agency, Washington, DC. EPA/600/R-05/004aF-cF. March 2006. Available at:
http://www.epa.gov/ttn/naaqs/standards/ozone/s_o3_cr_cd.html
U.S. EPA (2007). Review of the National Ambient Air Quality Standards for Ozone: Policy Assessment of
Scientific and Technical Information, OAQPS Staff Paper. EPA-452/R-07-007
U.S. EPA (2009). Integrated Science Assessment for Paniculate Matter (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-08/139F, 2009.
U.S. EPA (2013). Integrated Science Assessment of Ozone and Related Photochemical Oxidants (Final Report).
U.S. Environmental Protection Agency, Washington, DC. EPA/600/R-10/076F. Available at:
http://www.epa.gov/ttn/naaqs/standards/ozone/s_o3_2008_isa.html
U.S. EPA (2014). Health Risk and Exposure Assessment for Ozone. Office of Air Quality Planning and Standards,
Research Triangle Park, NC, 27711. EPA-452/R-14-004a. August 2014. Available at:
http://www.epa.gov/ttn/naaqs/standards/ozone/s_o3_index.html
Ulmer, C; Kopp, M; Ihorst, G; Frischer, T; Forster, J; Kuehr, J. (1997). Effects of ambient ozone exposures during
the spring and summer of 1994 on pulmonary function of schoolchildren. Pediatr Pulmonol 23: 344-353.
http://dx.doi.org/10.1002/(SICI)1099-0496(199705)23:5<344::AID-PPUL6>3.0.CO;2-K
Vagaggini, B; Cianchetti, S; Bartoli, M; Ricci, M; Bacci, E; Dente, FL; Di Franco, A; Paggiaro, P. (2007).
Prednisone blunts airway neutrophilic inflammatory response due to ozone exposure in asthmatic subjects.
Respiration 74: 61-58. http://dx.doi.org/10.1159/000096078
Vagaggini, B; Taccola, M; Cianchetti, S; Carnevali, S; Bartoli, ML; Bacci, E; Dente, FL; Di Franco, A; Giannini, D;
Paggiaro, PL. (2002). Ozone exposure increases eosinophilic airway response induced by previous allergen
challenge. Am J Respir Crit Care Med 166: 1073-1077.
Vagaggini, B; Taccola, M; Conti, I; Carnevali, S; Cianchetti, S; Bartoli, ML; Bacci, E; Dente, FL; Di Franco, A;
Giannini, D; Paggiaro, PL. (2001). Budesonide reduces neutrophilic but not functional airway response to
ozone in mild asthmatics. Am J Respir Crit Care Med 164: 2172-2176.
3-149
-------
Vedal, S.; Brauer, M.; White, R.; Petkau, J. (2003) Air pollution and daily mortality in a city with low levels of
pollution. Environ. Health Perspect. Ill: 45-51.
Villeneuve, PJ; Chen, L; Rowe, BH; 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
Global Access Sci Source 6: 40. http://dx.doi.org/10.1186/1476-069X-6-40
Ward, DJ; Roberts, KT; Jones, N; Harrison, RM; Ayres, JG; Hussain, S; Walters, S. (2002). Effects of daily
variation in outdoor particulates and ambient acid species in normal and asthmatic children. Thorax 57:
489-502. http://dx.doi.0rg/10.1136/thorax.57.6.489
Wenten, M; Gauderman, WJ; Berhane, K; Lin, PC; Peters, J; Gilliland, FD. (2009). Functional variants in the
catalase and myeloperoxidase genes, ambient air pollution, and respiratory-related school absences: An
example of epistasis in gene-environment interactions. Am J Epidemiol 170: 1494-1501.
http://dx.doi.org/10.1093/aje/kwp310
Wolff, G.T. (1995) Letter to EPA Administrator Carol Browner: "CASAC Closure on the Primary Standard Portion
of the Staff Paper for Ozone" EPA-SAB-CASAC-LTR-96-002, November 30, 1995.
Wong, CM; Vichit-Vadakan, N; Vajanapoom, N; Ostro, B; Thach, TQ; Chau, PY; Chan, EK; Chung, RY; Ou, CQ;
Yang, L; Peiris, JS; Thomas, GN; Lam, TH; Wong, TW; Hedley, AJ; Kan, H; Chen, B; Zhao, N; London,
SJ; Song, G; Chen, G; Zhang, Y; Jiang, L; Qian, Z; He, Q; Lin, HM; Kong, L; Zhou, D; Liang, S; Zhu, Z;
Liao, D; Liu, W; Bentley, CM; Dan, J; Wang, B; Yang, N; Xu, S; Gong, J; Wei, H; Sun, H; Qin, Z. (2010).
Part 5. Public health and air pollution in Asia (PAPA): A combined analysis of four studies of air pollution
and mortality. In Public Health and Air Pollution in Asia (PAPA): Coordinated Studies of Short-Term
Exposure to Air Pollution and Daily Mortality in Four Cities (pp. 377-418). Boston, MA: Health Effects
Institute. http://pubs.healtheffects.org/view.php?id=348
Wong, CM; Yang, L; Thach, TQ; Chau, PY; Chan, KP; Thomas, GN; Lam, TH; Wong, TW; Hedley, AJ; Peiris, JS.
(2009). Modification by influenza on health effects of air pollution in Hong Kong. Environ Health Perspect
117: 248-253. http://dx.doi.org/10.1289/ehp.11605
Woodwell, D. A.; Cherry, D. K. (2004) National Ambulatory Medical Care Survey: 2002 summary. Hyattsville,
MD: National Center for Health Statistics; DHHS publication no. (PHS) 2004-1250. (Advance data from
vital and health statistics; no. 346). Available: http://www.cdc.gov/nchs/data/ad/ad346.pdf [3 August,
2005]
Xia, Y; Tong, H. (2006). Cumulative effects of air pollution on public health. StatMed 25: 3548-3559.
http://dx.doi.org/10.1002/sim.2446
Yang, Q; Chen, Y; Krewski, D; Burnett, RT; Shi, Y; Mcgrail, KM. (2005). Effect of short-term exposure to low
levels of gaseous pollutants on chronic obstructive pulmonary disease hospitalizations. Environ Res 99: 99-
105. http://dx.doi.0rg/10.1016/j.envres.2004.09.014
Zanobetti, A; Schwartz, J. (2011). Ozone and survival in four cohorts with potentially predisposing diseases. Am J
Respir Crit Care Med 184: 836-841. http://dx.doi.org/10.1164/rccm.201102-0227OC
Zanobetti, A; Schwartz, J. (2008a). Is there adaptation in the ozone mortality relationship: A multi-city case-
crossover analysis. Environ Health 7: 22. http://dx.doi.org/10.1186/1476-069X-7-22
Zanobetti, A; Schwartz, J. (2008b). 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: 184-189.
http://dx.doi.org/10.1164/rccm.200706-823OC
3-150
-------
Zanobetti, A; Schwartz, J. (2006). Air pollution and emergency admissions in Boston, MA. J Epidemiol Community
Health 60: 890-895. http://dx.doi.org/10.1136/iech.2005.039834
3-151
-------
4 CONSIDERATION OF ALTERNATIVE PRIMARY STANDARDS
Having reached the conclusion that the currently available scientific evidence and
exposure/risk information calls into question the adequacy of the current Os standard, we next
consider the following overarching question:
• What is the range of potential alternative standards that are supported by the
currently available scientific evidence and exposure/risk information, as reflected
in the ISA and HREA respectively?
To address this overarching question, in the sections below we evaluate a series of more specific
questions related to the major elements of the NAAQS: indicator (section 4.1), averaging time
(section 4.2), form (section 4.3), and level (section 4.4). In addressing these questions, we
consider the currently available scientific evidence and exposure/risk information, including the
evidence and information available at the time of the last review and that newly available in the
current review, as assessed in the ISA and the HREA. In so doing, we note that the final decision
by the Administrator in this review will consider these elements collectively in evaluating the
health protection afforded by the primary standard.1
4.1 INDICATOR
In the last review, EPA focused on Os as the most appropriate indicator for a standard
meant to provide protection against ambient photochemical oxidants. In this review, while the
complex atmospheric chemistry in which Os plays a key role has been highlighted, no
alternatives to Os have been advanced as being a more appropriate indicator for ambient
photochemical oxidants. More specifically, the ISA noted that Os is the only photochemical
oxidant (other than NCh) that is routinely monitored and for which a comprehensive database
exists (ISA section 3.6). Data for other photochemical oxidants (e.g., PAN, H2Ch, etc.) typically
have been obtained only as part of special field studies. Consequently, no data on nationwide
patterns of occurrence are available for these other oxidants; nor are extensive data available on
the relationships of concentrations and patterns of these oxidants to those of Os (U.S. EPA, 2013,
section 3.6). In its review of the second draft PA, CASAC concurred, stating "The indicator of
ozone is appropriate 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,
2014, p. ii).
1 We also take note of the 1997 review (discussed in section 1.3.1.2.3), in which O3 background concentrations were
an additional consideration in EPA's selection of a standard from among a range of scientifically acceptable
alternatives. Background Os is discussed in more detail in chapter 2 of this PA.
4-1
-------
We further note that meeting an Os standard can be expected to provide some degree of
protection against potential health effects that may be independently associated with other
photochemical oxidants, even though such effects are not discernible from currently available
studies indexed by Os alone. That is, since the precursor emissions that lead to the formation of
Os generally also lead to the formation of other photochemical oxidants, measures leading to
reductions in population exposures to Os can generally be expected to lead to reductions in
population exposures to other photochemical oxidants. Taken together, we conclude that Cb
remains the most appropriate indicator for a standard meant to provide protection against
photochemical oxidants.2
4.2 AVERAGING TIME
The EPA established the current 8-hour averaging time3 for the primary Os NAAQS in
1997 (62 FR 38856). The decision on averaging time in that review was based on numerous
controlled human exposure and epidemiologic studies reporting associations between 6 to 8 hour
Os concentrations and adverse respiratory effects (62 FR 38861). It was also noted that a
standard with a max 8-hour averaging time is likely to provide substantial protection against
respiratory effects associated with 1-hour peak Os concentrations. Similar conclusions were
reached in the last Os NAAQS review and thus, the 8-hour averaging time was retained in 2008.
In the current review, we first consider the following question related to averaging time:
• To what extent does the available evidence continue to support the
appropriateness of a standard with an 8-hour averaging time?
In reaching conclusions related to this question, staff considers causality judgments from the
ISA, as well as results from the specific controlled human exposure and epidemiologic studies
that informed those judgments. These considerations are described below in more detail.
As an initial consideration with respect to the most appropriate averaging time for the Os
NAAQS, we note that the strongest evidence for Os-associated health effects is for respiratory
effects following short-term exposures. More specifically, the ISA concludes that evidence
relating short-term Os exposures to respiratory effects is "sufficient to infer a causal
relationship." The ISA also judges that the evidence for short-term exposures to Os indicates
"likely to be" causal relationships with both cardiovascular effects and mortality (U.S. EPA,
2013, section 2.5.2). Therefore, as in past reviews, the strength of the available scientific
2The D.C. Circuit upheld the use of O3 as the indicator for photochemical oxidants based on these same
considerations. American Petroleum Inst. v. Costle, 665 F. 2d 1176, 1186 (D.C. Cir. 1981).
3This 8-hour averaging time reflects daily max 8-hour average Os concentrations.
4-2
-------
evidence provides strong support for a standard that protects the public health against short-term
exposures to Os.
In first considering the level of support available for specific short-term averaging times,
we note the evidence available from controlled human exposure studies. As discussed in more
detail in chapter 3 of this PA, substantial health effects evidence from controlled human
exposure studies demonstrates 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 6.6 hour
exposures to Os (EPA 2013, section 6.2.1.1). Compared to shorter exposure durations (e.g., 1-
hour), studies evaluating 6.6 hour exposures in healthy adults have reported respiratory effects at
lower Os exposure concentrations and at more moderate levels of exertion.
We also note the strength of evidence from epidemiologic studies that have evaluated a
wide variety of populations (e.g., including at-risk lifestages and populations, such as children
and people with asthma, respectively). A number of different averaging times are used in Os
epidemiologic studies, with the most common being the max 1-hour concentration within a 24-
hour period (1-hour max), the max 8-hour average concentration within a 24-hour period (8-hour
max), and the 24-hour average. These studies are discussed in chapter 3 of this PA, and are
assessed in detail in chapter 6 of the ISA (U.S. EPA, 2013). Limited evidence from time-series
and panel epidemiologic studies comparing risk estimates across averaging times does not
indicate that one exposure metric is more consistently or strongly associated with respiratory
health effects or mortality, though the ISA notes some evidence for "smaller Os risk estimates
when using a 24-hour average exposure metric" (EPA 2013, section 2.5.4.2; p. 2-31). For single-
and multi-day average Os concentrations, lung function decrements were associated with 1-hour
max, 8-hour max, and 24-hour average ambient Os concentrations, with no strong difference in
the consistency or magnitude of association among the averaging times (EPA 2013, p. 6-71).
Similarly, in studies of short-term exposure to Os and mortality, Smith et al. (2009) and Darrow
et al. (2011) have reported high correlations between risk estimates calculated using 24-hour
average, 8-hour max, and 1-hour max averaging times (EPA 2013, p. 6-253). Thus, the
epidemiologic evidence alone does not provide a strong basis for distinguishing between the
appropriateness of 1-hour, 8-hour, and 24-hour averaging times.
Considering the health information discussed above, we conclude that an 8-hour
averaging time remains appropriate for addressing health effects associated with short-term
exposures to ambient Os. An 8-hour averaging time is similar to the exposure periods evaluated
in controlled human exposure studies, including recent studies that provide evidence for
respiratory effects following exposures to Os concentrations below the level of the current
standard. In addition, epidemiologic studies provide evidence for health effect associations with
4-3
-------
8-hour Os concentrations, as well as with 1-hour and 24-hour concentrations. As in previous
reviews, we note that a standard with an 8-hour averaging time (combined with an appropriate
standard form and level) would also be expected to provide substantial protection against health
effects attributable to 1-hour and 24-hour exposures (e.g., 62 FR 38861, July 18, 1997). In its
review of the second draft PA, CAS AC concurred stating that "the current 8-hour averaging time
is justified by the combined evidence from epidemiologic and clinical studies" (Frey, 2014, p. 6).
The ISA also concludes that the evidence for long-term Os exposures indicates that there
is "likely to be a causal relationship" with respiratory effects (US EPA, 2013, chapter 7). Thus,
in this review we also consider the extent to which currently available evidence and
exposure/risk information suggests that a standard with an 8-hour averaging time can provide
protection against respiratory effects associated with longer term exposures to ambient Cb. In
doing so, staff considers the following question:
• To what extent does the available evidence and exposure/risk information indicate
that a standard with the current 8-hour averaging time could provide protection
against long-term exposures to ambient Os?
In considering this issue in the last review of the Os NAAQS, staff noted that "because long-term
air quality patterns would be improved in areas coming into attainment with an 8-hr standard, the
potential risk of health effects associated with long-term exposures would be reduced in any area
meeting an 8-hr standard" (U.S. EPA, 2007, p. 6-57).
In the current review, we further evaluate this issue, with a focus on the "long-term" Os
metrics reported to be associated with mortality or morbidity in recent epidemiologic studies. As
discussed in section 3.1.3, much of the recent evidence for such associations is based on studies
that defined long-term Os in terms of seasonal averages of daily max concentrations (e.g.,
seasonal averages of 1-hour or 8-hour daily max concentrations).
As an initial consideration, we note the risk results from the HREA for respiratory
mortality associated with long-term Os concentrations. As discussed in section 3.2.3.2, UREA
analyses indicate that as air quality is adjusted to just meet the current 8-hour standard, most
urban case study areas are estimated to experience reductions in respiratory mortality associated
with long-term Os concentrations based on the seasonal averages of 1-hour daily max Os
concentrations evaluated in the study by Jerrett et al. (2009) (U.S. EPA, 2014, chapter 7). As air
quality is adjusted to meet lower potential alternative standard levels, for standards based on 3-
year averages of the annual fourth-highest daily max 8-hour Os concentrations, respiratory
mortality risks are estimated to be reduced further in urban case study areas (section
4.4.2.3,below). This analysis indicates that an Os standard with an 8-hour averaging time, when
coupled with an appropriate form and level, can reduce respiratory mortality reported to be
associated with "long-term" Os concentrations.
4-4
-------
In further considering the study by Jerrett et al. (2009), we compare long-term Os
concentrations following air quality adjustment in urban case study areas (i.e., adjusted to meet
the current and potential alternative 8-hour standards) to the concentrations present in study
cities that provided the basis for the positive and statistically significant association with
respiratory mortality. As indicated below (Table 4-3), this comparison suggests that a standard
with an 8-hour averaging time can decrease seasonal averages of 1-hour daily max Os
concentrations, and can maintain those Os concentrations below the seasonal average where we
have the most confidence in the reported concentration-response relationship with respiratory
mortality (see section 4.4.1 for further discussion).
The HREA also conducted analyses evaluating the impacts of reducing regional NOx
emissions on the seasonal averages of 8-hour daily max Os concentrations.4 Seasonal averages of
8-hour daily max Os concentrations reflect long-term metrics that have been reported to be
associated with respiratory morbidity effects in several recent Os epidemiologic studies (e.g.,
Islam et al., 2008; Lin et al., 2008; Salam et al., 2009). The HREA analyses indicate that the
large majority of the U.S. population lives in locations where reducing NOx emissions would be
expected to result in decreases in seasonal averages of daily max 8-hour ambient Os
concentrations (U.S. EPA, 2014, section 8.2.3.2). Thus, consistent with the respiratory mortality
risk estimates noted above, this analysis suggests that reductions in Os precursor emissions in
order to meet a standard with an 8-hour averaging time would also be expected to reduce the
long-term Os concentrations that have been reported in recent epidemiologic studies to be
associated with respiratory morbidity.
Taken together, we conclude that a standard with an 8-hour averaging time, coupled with
the current 4th high form and an appropriate level, would be expected to provide appropriate
protection against the long-term Os concentrations that have been reported to be associated with
respiratory morbidity and mortality. In its review of the second draft PA, CAS AC concurred,
stating that "The 8-hour averaging window also provides protection against the adverse impacts
of long-term ozone exposures, which were found to be "likely causal" for respiratory effects and
premature mortality" (Frey, 2014, p. 6). This issue is considered further, within the context of
specific potential alternative standard levels, in section 4.4 below.
4.3 FORM
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. The foremost
consideration in selecting a form for potential alternative primary standards is the adequacy of
4Analyses are based on regional NOX reductions, which are effective in bringing down peak ambient O3
concentrations, but can have variable impacts on seasonal mean concentrations.
4-5
-------
the public health protection provided by the combination of the form and the other elements of
the standard. As such, in reaching staff conclusions regarding the appropriate form(s) to consider
for a potential alternative primary Os standard, we consider the following question:
• To what extent do the available evidence and/or information continue to support
the appropriateness of a standard with a form defined by the 3-year average of
annual 4th-highest 8-hour daily max Os concentrations?
The EPA established the current form of the primary Os NAAQS in 1997 (62 FR 38856).
Prior to that time, the standard had a "1-expected-exceedance" form.5 An advantage of the
current concentration-based form recognized in the 1997 review is that such a form better
reflects the continuum of health effects associated with increasing ambient Os concentrations.
Unlike an expected exceedance form, a concentration-based form gives proportionally more
weight to years when 8-hour Os concentrations are well above the level of the standard than to
years when 8-hour Os concentrations are just above the level of the standard. It was judged
appropriate to give more weight to higher Os concentrations, given that available health evidence
indicated a continuum of effects associated with exposures to varying concentrations of Os, and
given that the extent to which public health is affected by exposure to ambient Os is related to the
actual magnitude of the Os concentration, not just whether the concentration is above a specified
level.
During the 1997 review, EPA considered a range of alternative "concentration-based"
forms, including the second-, third-, fourth- and fifth-highest daily max 8-hour concentrations in
an Os season. The fourth-highest daily max was selected, recognizing that a less restrictive form
(e.g., fifth highest) would allow a relatively large percentage of sites to experience Cb peaks well
above the level of the standard, and would allow more days on which the level of the standard
may be exceeded when attaining the standard (62 FR 38856). Consideration was also given to
setting a standard with a form that would provide a margin of safety against possible but
uncertain chronic effects, and would provide greater stability to ongoing control programs.6 A
more restrictive form was not selected, recognizing that the differences in the degree of
protection afforded by the alternatives were not well enough understood to use any such
differences as a basis for choosing the most restrictive forms (62 FR 38856).
In the 2008 review, EPA additionally considered the potential value of a percentile-based
form. In doing so, EPA recognized that such a statistic is useful for comparing datasets of
5For a standard with a 1-expected-exceedance form to be met at an air quality monitoring site, the fourth-highest air
quality value in 3 years, given adjustments for missing data, must be less than or equal to the level of the standard.
6 See American Trucking Assn's v. EPA, 283 F. 3d 355, 374-75 (D.C. Cir. 2002) (less stable implementation
programs may be less effective, and therefore EPA can consider programmatic stability in determining the form of a
NAAQS).
4-6
-------
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. However, EPA concluded
that 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 in locations with longer Os seasons relative to places with
shorter Os seasons. Thus, in the 2008 review EPA concluded that a form based on the nth-highest
max Os concentration would more effectively ensure that people who live in areas with different
length Os seasons receive the same degree of public health protection.
Based on analyses for forms specified in terms of an nth-highest concentration (n ranged
from 3 to 5), advice from CASAC, and public comment,7 the Administrator concluded that a 4th-
highest daily max should be retained (73 FR 16465). In reaching this decision, the Administrator
recognized that "there is not a clear health-based threshold for selecting a particular nth-highest
daily maximum form of the standard" and that "the adequacy of the public health protection
provided by the combination of the level and form is a foremost consideration" (73 FR 16475).
Based on this, the Administrator judged that the existing form (4th-highest daily maximum 8-
hour average concentration) should be retained, recognizing the increase in public health
protection provided by combining this form with a lower standard level (i.e., 75 ppb).
The Administrator also recognized that it is important to have a form that provides
stability with regard to implementation of the standard. In the case of Cb, for example, he noted
the importance of a form insulated from the impacts of the meteorological events that are
conducive to Os formation. 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. The Administrator noted that such frequent shifting could disrupt an
area's ongoing implementation plans and associated control programs (73 FR 16474). In his final
decision, the Administrator judged that a "4th high form provides a stable target for implementing
programs to improve air quality" (73 FR 16475).
In the current review, we consider the extent to which newly available information
provides support for consideration of alternative forms. In so doing, we take note of the
conclusions of prior reviews summarized above. We recognize the value of an nth-high statistic
over that of an expected exceedance or percentile-based form in the case of the Os standard, for
7In the 2008 review, one group of commenters expressed the view that the standard was not adequate and supported
a more health-protective form (e.g., a second- or third-highest daily max form). Another group of commenters
expressed the view that the standard was adequate and did not provide any views on alternative forms that would be
appropriate should the Administrator consider revisions to the standard. The Administrator considered the protection
afforded by the combination of level and form in revising the standard in 2008 to 75 ppb, as a 3-year average of the
annual fourth-highest daily max 8-hour concentrations (73 FR 16475).
4-7
-------
the reasons summarized above. We additionally take note of the importance of stability in
implementation to achieving the level of protection specified by the NAAQS. Specifically, we
note that to the extent that areas engaged in implementing the Os NAAQS frequently shift from
meeting to violating the standard, it is possible that ongoing implementation plans and associated
control programs could be disrupted, thereby reducing public health protection.
In light of this, while giving foremost consideration to the adequacy of public health
protection provided by the combination of all elements of the standard, including the form, we
consider particularly findings from prior reviews with regard to the use of the nth-high metric.
As noted above, the 4th-highest daily max was selected in recognition of the public health
protection provided by this form, when coupled with an appropriate averaging time and level,
and recognizing that such a form can provide stability for implementation programs. The
currently available evidence and information does not call into question these conclusions from
previous reviews. Moreover, in its review of the second draft PA, CAS AC concurred that the Cb
standard should be based on the fourth highest, daily maximum 8-hour average value (averaged
over three years), 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). Thus a standard with the current 4th
high form, coupled with a level lower than 75 ppb as discussed below, would be expected to
increase public health protection relative to the current standard while continuing to provide
stability for implementation programs. Therefore, we conclude that it would be appropriate to
consider retaining the current 4th-highest daily max form for an Cb standard with an 8-hour
averaging time and a revised level, as discussed below.
4.4 LEVEL
In considering potential alternative standards levels to provide greater protection than that
afforded by the current standard against Os-related adverse health effects, we address the
following overarching question.
• For an Os standard defined in terms of the current indicator, averaging time, and
form, what alternative levels are appropriate to consider in order to provide
adequate public health protection against short- and long- term exposures to Os
in ambient air?
In considering this question, we take into account the experimental and epidemiologic evidence
as presented in the ISA, as well as the uncertainties and limitations associated with this evidence
(section 4.4.1). In addition, we consider the quantitative estimates of exposure and risk provided
by the HREA, as well as the uncertainties and limitations associated with these risk estimates
(section 4.4.2).
4-8
-------
4.4.1 Evidence-based Considerations
In this section, we consider the available evidence from controlled human exposure and
epidemiologic studies, including the uncertainties and limitations associated with that evidence,
within the context of potential alternative standard levels. We consider both the exposure
concentrations at which controlled human exposure studies provide evidence for health effects,
and the ambient Os concentrations present in locations where epidemiologic studies have
reported health effect associations (see also section 3.1).
Controlled human exposure studies and epidemiologic panel studies
We consider the following question related to controlled human exposure studies and
panel studies:
• To what extent does the available evidence from controlled human exposure
studies and panel studies provide support for consideration of potential
alternative standard levels lower than 75 ppb?
To inform our conclusions regarding this question, we consider the lowest Os concentrations at
which various effects have been evaluated and statistically significant effects reported. We also
consider the potential for reported effects to be adverse, including in at-risk populations.
As discussed in section 3.1.2.1, data from controlled human exposure studies show that
group mean Os-induced lung function decrements in healthy adults exhibit a smooth dose-
response relationship without evidence of a threshold from 40 to 120 ppb Os (US EPA, 2013,
Figure 6-1). The lowest Cb exposure concentration for which statistically significant decrements
have been reported is 60 ppb (Brown, 2008; Kim et al., 2011). The ISA concludes that mean
FEVi is clearly decreased by 6.6-hour exposures to Os concentrations of 60 ppb and higher in
young, healthy adults during moderate exertion (US EPA, 2013, p. 6-9). As discussed in section
3.1.3, such a decrease in mean lung function meets the ATS criteria for an adverse response
given that a downward shift in the distribution of FEVi would result in diminished reserve
function, and therefore would increase risk from further environmental insult. In addition, based
on data from studies by Kim et al. (2011), Schelegle et al. (2009), Adams (2006), and Adams
(1998), the ISA notes that following exposures to 60 ppb Os 10% of healthy adults experience
FEVi decrements > 10% (U.S. EPA, 2013, page 6-19).8 A 10% decrement in FEVi is accepted
8As discussed in Chapter 3 of this PA (section 3.1.2.1), these estimates are consistent with the predictions
of quantitative models developed by McDonnell et al. (2012) and Schelegle et al. (2012). The McDonnell
model, as discussed in McDonnell et al. (2010), provides the basis for lung function risk estimates in the
HREA (section 4.4.2.2, below). For the target of 60 ppb, Schelegle et al. (2009) reported an actual mean
exposure concentration of 63 ppb.
4-9
-------
by ATS as an abnormal response. Based on advice received from CASAC in this (Frey, 2014, p.
3) and previous reviews, such decrements could be adverse in people with lung disease (section
3.1.3). Moreover, as discussed in section 3.1.3 of this PA, repeated occurrences of moderate
responses may be considered adverse since they could set the stage for more serious effects.
One recent controlled human exposure study has reported Os-induced pulmonary
inflammation (PMN increased in sputum from lower airways) following exposures of young,
healthy adults to Os concentrations of 60 ppb (Kim et al., 2011), the lowest concentration at
which inflammatory responses have been evaluated in human studies (see discussion in section
3.1.2.1). Induction of pulmonary inflammation is evidence that injury has occurred. The
possibility of chronic effects due to repeated inflammatory events has been evaluated in animal
studies. Repeated events of acute inflammation can have several potentially adverse outcomes
including: induction of a chronic inflammatory state; altered pulmonary structure and function,
leading to diseases such as asthma; altered lung host defense response to inhaled
microorganisms, particularly in potentially at-risk populations such as the very young and old;
and, altered lung response to other agents such as allergens or toxins (U.S. EPA, 2013, Section
6.2.3). Thus, lung injury and the resulting inflammation, particularly if experienced repeatedly,
provide a mechanism by which Os may cause other more serious respiratory effects (e.g., asthma
exacerbations) and possibly extrapulmonary effects.
With respect to respiratory symptoms, a recent study by Schelegle et al. (2009) reported a
statistically significant increase in respiratory symptoms in young, healthy adults following 6.6
hour exposures to an average Os concentration of 70 ppb.9 This study also reported a statistically
significant decrease in FEVi following such exposures. As discussed in section 3.1.3, the
occurrence of both lung function decrements and respiratory symptoms meets criteria established
by the ATS defining an "adverse" respiratory response. Although some studies have reported
that respiratory symptoms develop during exposures at 60 ppb, the increases in symptoms in
these studies have not reached statistical significance by the end of the 6.6 hour exposures
(Adams 2006; Schelegle et al., 2009).10
Based on the results discussed above and in section 3.1.2.1, we conclude that controlled
human exposure studies provide evidence of potentially adverse lung function decrements and
airway inflammation in healthy adults following exposures to 60 ppb Os, and evidence of
9 For the target of 70 ppb, Schelegle et al. (2009) reported an actual mean exposure concentration of 72 ppb.
10Adams (2006) reported an increase in respiratory symptoms in healthy adults during a 6.6 hour exposure protocol
with an average Os exposure concentration of 60 ppb. This increase was significantly different from initial
respiratory symptoms, but not from the filtered air control day. For the target of 60 ppb, Schelegle et al. (2009)
reported an actual mean exposure concentration of 63 ppb and did not observe a statistically significant increase in
respiratory symptoms.
4-10
-------
respiratory symptoms combined with lung function decrements (an "adverse" response based on
ATS criteria) following exposures to 70 ppb. In reaching these conclusions, we recognize that
most studies have not evaluated exposure concentrations below 60 ppb, and that 60 ppb does not
necessarily reflect an exposure concentration below which effects no longer occur. Specifically,
given the occurrence of airway inflammation in healthy adults following exposures to 60 ppb and
higher, it may be reasonable to expect that inflammation would also occur following exposures
to Os concentrations somewhat below 60 ppb. Although some studies show that respiratory
symptoms develop during exposures at 60 ppb, they have not reached statistical significance by
the end of the 6.6 hour exposures (Adams 2006; Schelegle et al. 2009). Thus, respiratory
symptoms combined with lung function decrements are likely to occur to some degree in healthy
adults with 6.6-hour exposures to concentrations below 70 ppb, and are more likely to occur with
8-hour exposures to 70 ppb and below. Further, we note that these controlled human exposure
studies were conducted in healthy adults and that people with asthma, including asthmatic
children, are likely to be more sensitive to Os-induced respiratory effects. Therefore, these
exposure concentrations are more likely to cause adverse respiratory effects in children and
adults with asthma, and more generally in people with respiratory disease.
With regard to other Os-induced effects, we note that airway hyperresponsiveness and
impaired lung host defense capabilities have been reported in healthy adults engaged in moderate
exertion following exposures to Os concentrations as low as 80 ppb, the lowest concentration
evaluated for these effects.11 As discussed in section 3.1.2.1, these physiological effects have
been linked to aggravation of asthma and increased susceptibility to respiratory infection,
potentially leading to increased medication use, increased school and work absences, increased
visits to doctors' offices and emergency departments, and increased hospital admissions. These
are all indicators of adverse Os-related morbidity effects, which are consistent with, and provide
plausibility for, the adverse morbidity effects and mortality effects observed in epidemiologic
studies.
In further considering effects following exposures to Os concentrations below 75 ppb, in
section 3.1.4.1 we discuss panel studies highlighted in the ISA for the extent to which monitored
ambient Os concentrations reflect exposure concentrations in their study populations (U.S. EPA,
2013, section 6.2.1.2). These panel studies used on-site monitoring to evaluate Cb-attributable
lung function decrements in people engaged in outdoor recreation, exercise, or work. Table 3-2
includes Os panel studies that report analyses of Os-attributable lung function decrements for Os
concentrations at or below 75 ppb, and that measure Os concentrations with monitors located in
the areas where study subjects were active (e.g., on site at summer camps or in locations where
"There is no evidence that 80 ppb is a threshold for these effects (72 FR 37878, July 11, 2007).
4-11
-------
exercise took place). Consistent with the results of controlled human exposure studies discussed
above, these panel studies report associations with lung function decrements for subjects exposed
to on-site monitored Os concentrations below 75 ppb. Associations in panel studies have been
reported for a wider range of populations than has been evaluated in controlled human exposure
studies, including children.
With regard to the question above, we conclude that the available controlled human
exposure evidence and evidence from panel studies supports an upper end of the range of
potential alternative standard levels for consideration no higher than 70 ppb. As just discussed,
6.6-hour exposures of healthy adults to 70 ppb Os result in lung function decrements and
respiratory symptoms, a combination of effects that meet ATS criteria for an adverse response
(as discussed in section 3.1.3).12 In addition, while 70 ppb is below the 80 ppb concentration
shown in 6.6-hour exposure studies to cause potentially adverse respiratory effects such as
airway hyperresponsiveness and impaired host-defense capabilities, these effects have not been
evaluated at exposure concentrations below 80 ppb and there is no reason to believe that 80 ppb
represents a threshold for such effects. As discussed in section 3.1.2.1 of this PA, the
physiological effects reported in controlled human exposure studies down to 60 ppb Os have
been linked to aggravation of asthma and increased susceptibility to respiratory infection,
potentially leading to increased medication use, increased school and work absences, increased
visits to doctors' offices and emergency departments, and increased hospital admissions.
Based on the above considerations, we also conclude that the evidence from controlled
human exposure studies and panel studies supports considering alternative Os standard levels at
least as low as 60 ppb. Potentially adverse lung function decrements and pulmonary
inflammation have been demonstrated to occur in healthy adults at 60 ppb, with little evidence
for potentially adverse effects following exposures to Os concentrations below 60 ppb. Thus, 60
ppb is a short-term exposure concentration that may be reasonably concluded to elicit adverse
effects in at-risk groups. Pulmonary inflammation, particularly if experienced repeatedly,
provides a mechanism by which Os may cause other more serious respiratory morbidity effects
(e.g., asthma exacerbations) and possibly extrapulmonary effects.
Epidemiologic evidence
We also consider what the information from epidemiologic studies indicates with regard
to potential alternative standard levels appropriate for consideration. Based on the information in
12 Based on the Schelegle et al. (2009) study, CAS AC observed that, "adverse health effects in young healthy adults
occur with exposures to 72 ppb of ozone for 6.6 hours" and that "It is the judgment of CASAC that if subjects had
been exposed to ozone using the 8-hour averaging period used in the standard, adverse effects could have occurred
at [a] lower concentration. Further, in our judgment, the level at which adverse effects might be observed would
likely be lower for more sensitive subgroups, such as those with asthma" (Frey, 2014, p. 5).
4-12
-------
section 3.1.4.2 of this PA (see Table 3-3), we first note that several epidemiologic studies have
reported positive and statistically significant associations with hospital admissions, emergency
department visits, and/or mortality in study areas where ambient Os concentrations would have
met the current standard (i.e., with its level of 75 ppb). This includes Canadian multicity studies
in which the majority of study cities would have met the current standard over entire study
periods (Cakmak et al., 2006; Dales et al., 2006; Katsouyanni et al., 2009; Stieb et al., 2009), and
a U.S. single-city study conducted in a location likely to have met the current standard over the
entire study period (Mar and Koenig, 2009).
In further evaluating these studies, and building upon our conclusions based on controlled
human exposures studies, as discussed above, we consider the following question related to the
epidemiologic evidence:
• To what extent have U.S. and Canadian epidemiologic studies reported associations
with mortality or morbidity in locations likely to have met potential alternative Os
standards with levels from 70 to 60 ppb?
Our focus in addressing this question is on what epidemiologic studies convey regarding the
extent to which Os-associated health effects may be occurring (i.e., as indicated by associations)
under air quality conditions allowed by potential alternative standards with levels of 70, 65, and
60 ppb (Table 4-1).13
13See^7M ///, 283 F.3d at 370 (EPA justified in revising NAAQS when health effect associations are observed at
levels allowed by the NAAQS).
4-13
-------
Table 4-1 Numbers of epidemiologic study locations likely to have met potential
alternative standards with levels of 70, 65, and 60 ppb
Study
Cakmak et al.
(2006)
Dales et al.
(2006)
Katsouyanni
et al. (2009)
Katsouyanni
et al. (2009)
Mar and
Koenig
(2009)
Stieb et al.
(2009)
Result
Positive and statistically
significant association with
respiratory hospital
admissions
Positive and statistically
significant association with
respiratory hospital
admissions
Positive and statistically
significant associations with
respiratory hospital
admissions
Positive and statistically
significant associations with
total and cardiovascular
mortality
Positive and statistically
significant associations with
asthma emergency
department visits
Positive and statistically
significant association with
respiratory emergency
department visits
Cities
10 Canadian
cities
1 1 Canadian
cities
12 Canadian
cities
12 Canadian
cities
Single city:
Seattle
7 Canadian
cities
Number of study cities meeting potential
alternative standards during entire study
period
70 ppb
7
5
9
7
0
5
65 ppb
6
4
9
5
0
4
60 ppb
2
0
5
1
0
3
As discussed in section 3.1.4.2, the single-city study by Mar and Koenig reported
associations with respiratory emergency department visits in a location that would have met the
current standard over the entire study period. In contrast, over at least part of the study period
this area would have violated alternative Os standards with levels of 70 ppb or below. Thus,
while this study indicates that the current standard would allow the reported associations with
respiratory emergency department visits, it does not provide information on the extent to which
those health effect associations would be present if ambient Os concentrations were reduced to
meet a revised standard with a level at or below 70 ppb.
With regard to the multicity studies included in Table 4-1, none were conducted in study
locations that all would have met an Os standard with a level at or below 70 ppb. However, for
the studies by Cakmak et al. (2006), Katsouyanni et al. (2009), and Stieb et al. (2009), the
majority of study locations would likely have met a standard with a level of either 70 or 65 ppb
(Cakmak et al., 2006; Katsouyanni et al., 2009; Stieb et al., 2009). Thus the majority of the
distributions of ambient Os concentrations that provided the basis for positive and statistically
significant associations with mortality or morbidity in these studies would likely be allowed
4-14
-------
under alternative standards with levels of 70 or 65 ppb, though not 60 ppb. However, our
interpretation of these results is complicated by uncertainties in the extent to which multicity
effect estimates can be attributed to ambient Os in the majority of locations, which would have
met alternative standards, versus Os in the smaller number of locations that would have violated
those alternatives.
As with our consideration of the current standard (section 3.1.4.2), we next consider the
extent to which epidemiologic studies have characterized Os health effect associations, including
confidence in those associations, for various portions of distributions of ambient Os
concentrations. In considering such analyses within the context of potential alternative standards,
we focus on the extent to which epidemiologic studies report health effect associations for air
quality distributions restricted to ambient pollutant concentrations below one or more
predetermined cut-points. As discussed in section 3.1.4.2, such "cut-point" analyses can provide
information on the magnitude and statistical precision of effect estimates for defined
distributions of ambient concentrations, which may in some cases include distributions that
would be allowed by potential alternative standards. Specifically, we consider the following
question:
• To what extent do cut-point analyses from epidemiologic studies report health effect
associations at ambient Os concentrations that are likely to be allowed by potential
alternative standards with levels from 70 to 60 ppb?
As with our consideration of the current standard in section 3.1.4.2 of this PA, we
evaluate the cut-point analyses presented in the U.S. multicity study by Bell et al. (2006). These
cut-point analyses can provide insights into the magnitude and statistical precision of health
effect associations for different portions of the distribution of ambient concentrations, including
insights into the ambient concentrations below which uncertainty in reported associations
becomes notably greater. Our analysis of air quality data associated with the cut-points evaluated
by Bell et al., and uncertainties associated with that analysis, is described elsewhere in this
document (section 3.1.4.2). In this section, we consider what these cut-point analyses indicate
with regard to the potential for health effect associations to extend to ambient Os concentrations
likely to be allowed by a revised Os NAAQS with a level below 75 ppb.
We particularly focus on the lowest cut-point for which the association between Os and
mortality was reported to be statistically significant (i.e., 30 ppb, as discussed in section 3.1.4.2).
Based on the Os air quality concentrations that met the criteria for inclusion in the 30 ppb cut-
point analysis, 84% of study areas had 3-year averages of annual 4th highest 8-hour daily max Os
concentrations at or below 70 ppb over the entire study period (Table 4-2). In addition, 64% of
study areas had 3-year averages of annual 4th highest 8-hour daily max Os concentrations at or
4-15
-------
below 65 ppb (Table 4-2). In contrast, the majority of study areas had 4th highest concentrations
above 60 ppb.
Consistent with our interpretation of multicity effect estimates discussed above, these
results suggest that the majority of the air quality distributions included in the 30 ppb Os cut
point would have been allowed by a standard with a level of 70 or 65 ppb. Thus the majority of
the distributions of ambient Os concentrations that provided the basis for a positive and
statistically significant association with mortality would be allowed by alternative standards with
levels of 70 or 65 ppb, but not 60 ppb. However, as discussed below our interpretation of these
cut point analyses is complicated by important uncertainties.
Table 4-2 Number of study cities with 3-year averages of 4th highest 8-hour daily max
concentrations greater than 70, 65, or 60 ppb, for various cut-point analyses
presented in Bell et al. (2006)
Number (%) of
Cities with 4th
highest >70 (any
3 -yr period; 1987-
2000)
Number (%) of
Cities with 4th
highest >65 (any
3 -yr period; 1987-
2000)
Number (%) of
Cities with 4th
highest >60 (any
3 -yr period; 1987-
2000)
Cut-point for 2-day moving average across monitors and cities (24-h avg)14
25
0 (0%)
3 (3%)
16
(16%)
30
16
(16%)
35
(36%)
61
(62%)
35
55
(56%)
77
(79%)
86
(88%)
40
82
(84%)
89
(91%)
94
(96%)
45
89
(91%)
94
(96%)
95
(97%)
50
92
(94%)
95
(97%)
96
(8%)
55
94
(96%)
95
(97%)
96
(8%)
60
95
(97%)
95
(97%)
96
(8%)
All
95
(97%)
95
(97%)
96
(8%)
In further considering the implications of Tables 4-1 and 4-2 for potential alternative
standard levels, we also note the important uncertainties described in section 3.1.4 of this
document. General uncertainties include the geographic heterogeneity in effect estimates, which
could obscure presence of potential thresholds in multicity studies; uncertainty in the extent to
which multicity effect estimates can be attributed to ambient Os in the majority of locations,
which would have met alternative standards with levels of 70 or 65 ppb, versus Os in the smaller
number of locations that would have violated those alternatives; and uncertainty in the extent to
which the relatively low ambient Os concentrations present in some study areas caused or
14Cut point analyses presented in the study by Bell et al. (2006) are described in more detail in sections 3.1.2.3 and
3.1.4.2 of this document.
4-16
-------
contributed to reported effects. Additional uncertainties specific to our analysis of the cut points
presented by Bell et al. (2006) include the appropriateness of identifying 4th highest
concentrations from air quality subsets, rather than the entire air quality distributions that existed
in study locations, and uncertainty associated with the air quality data used to re-create the cut-
point analyses from the published study. With regard to this second uncertainty, as described in
more detail in section 3.1.4.2 of this document, our re-creation of the cut points was based on air
quality data available in AQS, combined with the published descriptions of cut point criteria and
study area definitions. In doing so, we did not recreate the trimmed means used by Bell.
Therefore, an important uncertainty in this approach is the extent to which we were able to
appropriately re-create the cut-point analyses in the published study.
Overall, our analyses of air quality in U.S. and Canadian epidemiologic study locations
indicate that (1) single-city studies have not been conducted in locations that would have met
alternative Os standards with levels of 70 ppb or below and that (2) multicity epidemiologic
studies report positive and statistically significant associations with mortality and morbidity
based largely on distributions of ambient Os that would have been allowed by alternative
standards with levels of 70 or 65 ppb, but not 60 ppb. While important uncertainties, mentioned
above, complicate our interpretation of the multicity studies, at a minimum these results suggest
that an alternative standard level of 60 ppb would not allow the distributions of ambient Os
concentrations present in the majority of study locations that provided the basis for statistically
significant health effect associations. While the potential implications for alternative standard
levels of 70 and 65 ppb are less clear, given the important uncertainties in these analyses, the
results suggest that positive and statistically significant associations with mortality or morbidity
in some studies were largely influenced by air quality distributions that would be allowed under
alternative standards with such levels.
We next consider the extent to which epidemiologic studies employing longer-term
ambient Os concentration metrics can inform our consideration of potential alternative standard
levels. In doing so, we consider the following question:
• To what extent does the available evidence indicate that an Os standard with a
level from 70 to 60 ppb, combined with the current 8-hour averaging time and 4th
high form, could provide protection from long-term exposures to ambient Os
concentrations for which there is evidence of health effects?
We first note that, as discussed in section 3.1.4.3 of this PA, virtually all of the study
cities that provided the basis for the positive and statistically significant association between
long-term Os and respiratory mortality (Jerrett et al., 2009) would have violated the current
standard, and therefore potential alternative standards with lower levels. Thus, as with our
4-17
-------
consideration of the current standard in section 3.1.4.3, while the study by Jerrett et al. (2009)
contributes to our understanding of health effects associated with ambient Os (summarized in
section 3.1.2), it is less informative regarding the extent to which those health effects may be
occurring under air quality conditions that would meet potential alternative standards.
To further evaluate this issue, we use the adjusted air quality in urban case study areas, as
described in the HREA, to consider the extent to which just meeting alternative Os standards
with levels of 70, 65, and 60 ppb could maintain long-term Os concentrations below those in the
cities that provided the basis for the positive and statistically significant association with
respiratory mortality reported by Jerrett et al. (2009).15 Upon adjustment of air quality in U.S.
urban case study areas to meet the current and potential alternative 8-hour standards, seasonal
average 1-hour daily max concentrations were calculated and compared to the concentrations in
study cities.
As discussed in section 3.1.4.3, Jerrett et al. (2009) reported that when seasonal averages
of 1-hour daily max Os concentrations16 ranged from 33 to 104 ppb, there was no statistical
deviation from a linear concentration-response relationship between Os and respiratory mortality
across 96 U.S. cities (U.S. EPA, 2013, section 7.7). However, as discussed in section 3.1.4.3, the
study suggests notably decreased confidence in the reported linear concentration-response
function for "long-term" Os concentrations in the first quartile (i.e., at or below about 53 ppb),
given the widening in confidence intervals for lower concentrations (based on visual inspection
of Figure 3-6 in section 3.1.4.3); the fact that most study cities contributing to the linear function
had Os concentrations in the highest three quartiles, accounting for approximately 72% of the
respiratory deaths in the cohort (based on Table 2 in the published study); and the limited
evidence presented in the published study for a threshold at or near 56 ppb.17
Given the above, we note the extent to which long-term Os concentrations (i.e., seasonal
average of 1-hour daily max) in urban case study areas are estimated to be at or below 53 ppb
following air quality adjustment to meet potential alternative standards with levels of 70, 65, and
60 ppb. To the extent air quality adjustment to just meet potential alternative short-term
standards results in long-term concentrations near or below 53 ppb, we have greater confidence
15Air quality in U.S. urban case study areas was adjusted to just meet the current 8-hour standard at 75 ppb, as well
as potential potential alternative 8-hour standards at 70 ppb, 65 ppb, and 60 ppb, as described in the HREA (chapter
4). After a given adjustment, seasonal average 1-hour daily max concentrations were calculated.
16Jerrett et al. (2009) evaluated the April to September averages of 1-hour daily max Cb concentrations across 96
U.S. metropolitan areas from 1977- 2000. In urban areas with multiple monitors, April to September 1-hour daily
max concentrations from each individual monitor were averaged. This step was repeated for each year in the study
period. Finally, each yearly averaged Os concentrations was then averaged again to yield the single averaged 1-hour
daily max Os concentration depicted on the x-axis of Figure 3-6 below.
17The issue of potential thresholds based on the Jerrett study is discussed in more detail in section 3.2.3.2 of this PA.
4-18
-------
in the degree to which those short-term standards could protect against the health effects
associated with longer term Os exposures. Though there is uncertainty associated with these
comparisons (e.g., due to uncertainty in the potential for a threshold to exist; uncertainty in the
identification of such a threshold, should one exist; uncertainty in the long-term concentration
below which confidence intervals widen notably, based on visual inspection of concentration-
response function in the published study; and the limited number of urban case study areas for
which adjusted air quality is available), this analysis can provide insight into the extent to which
various alternative short-term standards would be expected to maintain long-term Cb
concentrations below those where we have the most confidence in the reported concentration-
response relationship with respiratory mortality.
Table 4-3 indicates that when considering recent (i.e., unadjusted) air quality, 2 of 12
urban case study areas had seasonal average 1-hour daily max Os concentrations at or below 53
ppb in all of the years examined. When air quality was adjusted to just meet the current 8-hour
standard (75 ppb in Table 4-3), 6 of 12 urban case study areas had seasonal average 1-hour daily
max Os concentrations at or below 53 ppb in all of the years examined. When air quality is
further adjusted to just meet potential alternative standards with lower levels, seasonal averages
of 1-hour daily max Os concentrations are estimated to be at or below 53 ppb in 9 of 12 urban
case study areas (70 ppb level), 10 of 12 urban case study areas (65 ppb level), and 11 of 11
urban case study areas (60 ppb level).18 Though as noted above there are important uncertainties
associated with interpreting these comparisons, they suggest that in many locations across the
U.S. a standard with an 8-hour averaging time, when combined with the current 4th high form
and an appropriate standard level, would be expected to maintain seasonal averages of 1-hour
daily max Os concentrations below those where analyses indicate the most confidence in the
concentration-response relationship with respiratory mortality reported by Jerrett et al. (2009).
18As described in the HREA, a standard level of 60 ppb was not evaluated in New York City (U.S. EPA, 2014,
chapter 4).
4-19
-------
Table 4-3 Seasonal averages of 1-hour daily max Os concentrations in U.S. urban case
study areas for recent air quality and air quality adjusted to just meet the
current and potential alternative standards.
Atlanta
Baltimore
Boston
Cleveland
Denver
Detroit
Houston
Los Angeles
New York City
Philadelphia
Sacramento
Saint Louis
Air Quality
Adjusted to:
Recent
75
70
65
60
Recent
75
70
65
60
Recent
75
70
65
60
Recent
75
70
65
60
Recent
75
70
65
60
Recent
75
70
65
60
Recent
75
70
65
60
Recent
75
70
65
60
Recent
75
70
65
60
Recent
75
70
65
60
Recent
75
70
65
60
Recent
75
70
65
60
2006
(AdjYrs 2006- 2008)
65
53
50
47
45
60
54
52
49
46
49
48
46
44
43
51
49
47
45
41
63
62
60
58
53
50
50
48
47
45
53
48
47
46
45
65
58
55
52
50
53
47
44
36
NA
56
51
49
47
45
66
55
52
50
47
58
53
50
47
44
2007
(AdjYrs 2006-2008)
63
52
49
46
44
59
54
51
49
46
50
49
47
45
43
52
50
48
45
41
63
61
59
58
53
54
52
50
49
46
48
46
45
44
43
61
59
56
53
51
54
47
45
36
NA
59
52
50
48
46
59
50
48
46
44
58
53
51
48
45
2008
(AdjYrs 2008-2010)
57
53
49
46
44
57
53
51
48
46
46
49
48
46
44
53
51
48
45
41
63
63
62
59
53
51
NA
51
49
46
47
47
46
45
43
64
60
57
54
52
55
51
48
39
NA
57
54
51
49
47
65
54
51
49
46
52
51
50
48
45
2009
(Adj Yrs 2008-2010)
50
47
44
42
40
52
49
48
46
44
45
45
44
43
41
49
47
45
43
40
58
58
58
56
51
48
NA
49
47
45
47
48
47
46
44
62
60
58
54
52
48
47
45
38
NA
51
49
47
45
43
61
51
49
47
44
51
50
48
46
43
2010
(Adj Yrs 2008-2010)
56
52
49
46
44
60
55
53
50
48
49
48
48
46
44
54
51
48
45
42
60
60
58
55
50
52
NA
52
50
47
46
46
46
45
44
57
58
56
53
50
55
51
48
39
NA
58
54
52
49
47
55
48
46
44
42
55
54
52
49
46
4-20
-------
Based on the above analyses, we conclude that the available epidemiologic evidence is
consistent with the available evidence from controlled human exposure studies in providing
support for consideration of an Os standard level in the range of 70 to 60 ppb. Compared to the
current standard, a standard level from within this range would expected to be more effective at
maintaining short-term and long-term ambient Os concentrations below those present in studies
reporting Os-associated mortality and/or morbidity.
In reaching overall staff conclusions about an appropriate range of standard levels for
consideration, we further evaluate the results of the exposure and risk assessments that are based
on modeling changes in the entire distribution of ambient Os concentrations to simulate just
meeting potential alternative standards. These results are discussed below in section 4.4.2.
4.4.2 Air Quality-, Exposure-, and Risk-Based Considerations
Beyond considering the available evidence, we also consider the extent to which specific
potential alternative standard levels, in conjunction with the current averaging time and form (3-
year average of annual 4th highest 8-hour daily max), could reduce estimated Os exposures and
health risks. In the first draft PA (U.S. EPA, 2012b), we concluded that the available evidence
supports conducting further exposure and risk analyses of potential alternative Os standard levels
in the range of 70 down to 60 ppb. Based on these conclusions, the HREA evaluates exposures
and risks estimated to be associated with potential alternative standard levels from the upper (70
ppb), middle (65 ppb), and lower (60 ppb) portions of this range. In considering these analyses in
this PA, we consider the following question:
• To what extent does the available exposure and risk information provide support
for considering potential alternative standard levels from 70 to 60 ppb, when
combined with the current 8-hour averaging time and 4th high form?
In considering exposure and risk analyses, we emphasize the nature and magnitude of the Os
exposures and health risks estimated to remain upon just meeting each alternative standard level,
and the changes in exposures and risks estimated for each alternative level when compared to the
current standard. Section 4.4.2.1 below discusses our exposure-based considerations. Sections
4.4.2.2 and 4.4.2.3 discuss our consideration of estimates of lung function risks and estimates of
epidemiology-based mortality/morbidity risks, respectively.
4.4.2.1 Exposure-Based Considerations
As discussed in more detail in section 3.2.2 of this PA, the exposure assessment
presented in the HREA (U.S. EPA, 2014, Chapter 5) provides estimates of the number and
percent of people exposed to Os concentrations at or above benchmark concentrations of 60, 70,
and 80 ppb, while at moderate or greater exertion. Estimates of such "exposures of concern"
4-21
-------
provide perspective on the potential public health impacts of Ch-related effects, including for
effects that cannot currently be evaluated in a quantitative risk assessment. The approach taken
in the HREA to estimating exposures of concern, and the key uncertainties associated with
exposure estimates, are summarized in section 3.2.2 for air quality adjusted to just meet the
current standard and are discussed in more detail in chapter 5 of the HREA (U.S. EPA, 2014). As
discussed in section 3.2.2, when evaluating potential alternative standard levels we focus on
modeled exposures for school-age children (ages 5-18), noting that percentages of asthmatic
school-age children estimated to experience exposures of concern are virtually indistinguishable
from those for all children, and that patterns of exposure in children represent a broader range of
at-risk populations, which includes adult asthmatics and older adults. In this review, CASAC
advised EPA to focus on the 60 ppb benchmark as being relevant for considering adverse effects
on people with asthma (Frey, 2014, p. 6).
In this section, we consider the following question:
• To what extent are potential alternative standards with revised levels estimated to
reduce the occurrence of Os exposures of concern, compared to the current
standard, and what are the nature and magnitude of the exposures remaining for
each alternative standard level evaluated?
Key results related to this question are summarized below (Figures 4-1 to 4-4). Figures 4-1
(estimates averaged over years) and 4-2 (estimates from worst-case years) present estimates of
one or more exposures of concern, and Figures 4-3 (estimates averaged over years) and 4-4
(estimates from worst-case years) present estimates of two or more exposures of concern.
4-22
-------
Figure 4-1. Percent of children estimated to experience one or more exposures of concern at or above 60, 70, or 80 ppb for air
quality adjusted to just meet the current and potential alternative standards (averaged over 2006 to 2010)
60 ppb benchmark
70 ppb benchmark
80 ppb benchmark
75 ppb 70 ppb 65 ppb 60ppb 75 ppb 70 ppb 65 ppb 60ppb 75 ppb 70 ppb 65 ppb 60ppb
—»—Atlanta
—•—Baltimore
—*—Boston
—W—Chicago
!\ Cleveland
—•—Dallas
—I—Denver
Detroit
Houston
9 Los Angeles
-•-New York
Philadelphia
Sacramento
I St. Louis
Washington
4-23
-------
Figure 4-2. Percent of children estimated to experience one or more exposures of concern at or above 60, 70, or 80 ppb for air
quality adjusted to just meet the current and potential alternative standards (worst-case year from 2006 to
201019)
60 ppb benchmark
70 ppb benchmark
80 ppb benchmark
75ppl> 70 ppl> &5ppb 60ppb
75 p|)l> 70 ppb
Standard Level
60ppb
75 ppb 70 ppb 65 ppb 60ppb
-Atlanta
-Baltimore
= Boston
-Chicago
- Cleveland
-Dallas
- Denver
- Detroit
Houston
-Los Angeles
-New York
Philadelphia
Sacramento
St Louis
Washington
19"Worst-case" year refers to the year in each urban case study area with the largest percentage of children estimated to experience exposures of concern.
4-24
-------
Figure 4-3. Percent of children estimated to experience two or more exposures of concern at or above 60, 70, or 80 ppb for air
quality adjusted to just meet the current and potential alternative standards (averaged over 2006 to 2010)
60 ppb benchmark
70 ppb benchmark
80 ppb benchmark
75ppb 70ppb 65ppb 60ppb 75ppb 70ppb 65ppb 60ppb 75ppb 70ppb 65ppb 60ppb
-Atlanta
-Baltimore
-Boston
-Chicago
-Cleveland
-Dallas
-Denver
-Detroit
Houston
-Los Angeles
-New York
Philadelphia
Sacramento
St. Louis
Washington
4-25
-------
Figure 4-4. Percent of children estimated to experience two or more exposures of concern at or above 60, 70, or 80 ppb for air
quality adjusted to just meet the current and potential alternative standards (worst-case year from 2006 to 2010)
60 ppb benchmark
70 ppb benchmark
80 ppb benchmark
'
75 ppb 70ppb 65ppb 60ppb 75ppb 70ppb 65ppb 60ppb 75ppb
Standard Level
70 ppb 6 5 ppb 60 ppb
-Atlanta
-Baltimore
-Boston
-Chicago
-Cleveland
-Dallas
-Denver
-Detroit
Houston
-Los Angeles
-New York
Philadelphia
Sacramento
St. Louis
Washington
4-26
-------
As illustrated above in Figures 4-1 to 4-4, adjusting air quality to just meet progressively
lower potential alternative standard levels reduces estimated exposures of concern consistently
across urban case study areas. These results reflect the consistent reductions in the highest
ambient Os concentrations upon air quality adjustment, as summarized in section 3.2.1 and as
discussed in more detail in the HREA (U.S. EPA, 2014, chapter 4). Based on Figures 4-1 to 4-4
and the associated details described in the HREA (U.S. EPA 2014, chapter 5), we take note of
the following with regard to exposures of concern for specific potential alternative standard
levels:
1. For an Os standard level of 70 ppb:
a. On average over the years 2006 to 2010, a standard with a level of 70 ppb is
estimated to allow approximately 3 to 10% of children in urban case study areas to
experience one or more exposures of concern at or above 60 ppb (approximately 30 to
70% reduction, relative to current standard). Summing across urban case study areas,
these percentages correspond to over 1 million children experiencing over 1.5 million
exposures of concern at or above 60 ppb during a single Os season. Of these children,
over 100,000 are asthmatics.
b. On average over the years 2006 to 2010, a standard with a level of 70 ppb is
estimated to allow approximately 0.5 to 3.5% of children in urban case study areas to
experience two or more exposures of concern at or above 60 ppb (approximately 50
to 85% reduction, relative to current standard).
c. In the worst-case years (i.e., those with the largest exposure estimates), a standard
with a level of 70 ppb is estimated to allow approximately 5 to 19% of children in
urban case study areas to experience one or more exposures of concern at or above 60
ppb, and approximately 2 to 9% to experience two or more.
d. On average over the years 2006 to 2010, a standard with a level of 70 ppb is
estimated to allow approximately 1% or less of children to experience one or more
exposures of concern at or above 70 ppb (approximately 55 to 90% reduction, relative
to current standard), and far less than 1% to experience two or more such exposures
(approximately 65 to 100% reduction, relative to current standard).
e. In the worst-case years, approximately 3% or less of children are estimated to
experience one or more exposures of concern at or above 70 ppb, and less than 1%
are estimated to experience two or more such exposures.
f A standard with a level of 70 ppb is estimated to allow less than 1% of children to
experience one or more exposures of concern at or above 80 ppb, even in the worst-
case years. No children are estimated to experience two or more such exposures.
2. For an Os standard level of 65 ppb:
4-27
-------
a. On average over the years 2006 to 2010, a standard with a level of 65 ppb is
estimated to allow approximately 4% or less of children in urban case study areas to
experience one or more exposures of concern at or above 60 ppb (approximately 70 to
100% reduction, relative to current standard). Summing across urban case study
areas, these percentages correspond to almost 400,000 children experiencing almost
500,000 exposures of concern at or above 60 ppb during a single Os season. Of these
children, about 40,000 are asthmatics.
b. On average over the years 2006 to 2010, a standard with a level of 65 ppb is
estimated to allow less than 1% of children to experience two or more exposures of
concern at or above 60 ppb (approximately 85 to 100% reduction, relative to current
standard).
c. In the worst-case years, a standard with a level of 65 ppb is estimated to allow
approximately 10% or less of children to experience one or more exposures of
concern at or above 60 ppb, and approximately 3% or less to experience two or more
such exposures.
d. On average over the years 2006 to 2010, a standard with a level of 65 ppb is
estimated to allow approximately 1% or less of children to experience one or more
exposures of concern at or above 70 ppb (approximately 90 to 100% reduction,
relative to current standard), and almost no children to experience two or more such
exposures. Even in the worst-case years, a level of 65 ppb is estimated to allow less
than 1% of children to experience exposures of concern at or above 70 ppb.
e. A standard with a level of 65 ppb is estimated to allow virtually no children to
experience exposures of concern at or above 80 ppb, even in the worst-case years.
3. For an Os standard level of 60 ppb:
a. On average over the years 2006 to 2010, a standard with a level of 60 ppb is
estimated to allow approximately 1% or less of children to experience one or more
exposures of concern at or above 60 ppb (approximately 90 to 100% reduction,
relative to current standard), and virtually no children to experience multiple such
exposures.
b. In the worst-case years, a standard with a level of 60 ppb is estimated to allow
approximately 2% or less of children to experience one or more exposures of concern
at or above 60 ppb, and almost no children to experience multiple such exposures.
c. On average over the years 2006 to 2010, a standard with a level of 60 ppb is
estimated to almost eliminate exposures of concern at or above 70 ppb or 80 ppb.
Even in years with the highest exposure estimates, virtually no children are estimated
to experience such exposures.
In further considering these exposure estimates, we take note of the associated
uncertainties, as discussed in more detail in section 3.2.2 of this PA (and in Chapter 5 of the
4-28
-------
HREA, U.S. EPA, 2014). These include (1) individual variability in responsiveness to Os
exposures such that only a subset of individuals who experience exposures at (or above) a
benchmark concentration would experience health effects; (2) potential to underestimate
exposures in most highly exposed populations; and (3) potential to overestimate exposures in
populations who alter behavior in response to high Os days (i.e., spend less time being active
outdoors). The implications of estimated exposures of concern for potential alternative standard
levels are discussed below in section 4.6.
4.4.2.2 Risk-Based Considerations: Lung Function
As discussed above in more detail in section 3.2.3.1 of this PA, the assessment of lung
function risks presented in the HREA (U.S. EPA, 2014, Chapter 6) provides estimates of the
number and percent of people experiencing Os-induced lung function decrements greater than or
equal to 10, 15, and 20%. In the current and past reviews, CAS AC has advised EPA to focus on
decrements of 10% or greater when considering people with pre-existing lung disease (Frey,
2014; Samet, 2011).
Lung function risk estimates are based on an updated dose-threshold model that estimates
FEVi responses for healthy adults following short-term exposures to Os (McDonnell, Stewart,
and Smith, 2010), reflecting methodological improvements since the last review (U.S. EPA,
2014, section 6.2.4). The approach taken in the HREA to estimating Os-induced lung function
decrements, and the key uncertainties associated with these estimates, are summarized in section
3.2.3.1 for air quality adjusted to just meet the current standard and are discussed in more detail
in chapter 6 of the HREA (U.S. EPA, 2014).
As discussed in section 3.2.3.1, in evaluating potential alternative standard levels we
focus on modeled exposures for school-age children, with an emphasis on asthmatic children. As
with exposures of concern, the percentages of all school age children and asthmatic school age
children estimated to experience particular Os-induced lung function decrements are virtually
indistinguishable.
In this section, we consider the following question:
• To what extent are potential alternative standards with revised levels estimated to
decrease the occurrence of Os-induced lung function decrements, compared to
the current standard, and what are the nature and magnitude of the decrements
remaining for each alternative standard level evaluated?
Key results related to this question are summarized below (Figures 4-5 to 4-8). Figures 4-5
(estimates averaged over years) and 4-6 (estimates from worst-case years) present estimates of
one or more Os-induced lung function decrements, and Figures 4-7 (estimates averaged over
years) and 4-8 (estimates from worst-case years) present estimates of two or more decrements.
4-29
-------
Figure 4-5. Percent of children estimated to experience one or more Os-induced lung function decrements greater than 10,15,
or 20% for air quality adjusted to just meet the current and potential alternative standards (averaged over 2006
to 2010)
Decrements > 10%
Decrements > 15%
Decrements > 20%
75ppb 70ppb 65ppb
6Qppb 75ppb 70ppb 65ppb
Standard Level
60ppb 75ppb 70ppb 6$ppb
60ppb
-Atlanta
-Baltimore
- Boston
-Chicago
-Cleveland
Dallas
- Denver
-Detroit
Houston
-Los Angeles
-New York
Philadelphia
Sacramento
St Louis
Washington
4-30
-------
Figure 4-6. Percent of children estimated to experience one or more Os-induced lung function decrements greater than 10,15,
or 20% for air quality adjusted to just meet the current and potential alternative standards (worst-case year from
2006 to 2010)
10%
Decrements > 15%
Decrements > 20%
75ppb 70ppb 65ppb 60ppb 75ppb 70ppb 6Sppb 60ppb 75ppb 70ppb 65ppb
Standard Level
GO ppl)
-Atlanta
-Baltimore
• Boston
-Chicago
-Cleveland
-Dallas
- Denver
- Detroit
Houston
-Los Angeles
-New York
Philadelphia
Sacramento
St Louis
Washington
4-31
-------
Figure 4-7. Percent of children estimated to experience two or more Os-induced lung function decrements greater than 10,
15, or 20% for air quality adjusted to just meet the current and potential alternative standards (averaged over
2006 to 2010)
g C
ji o
~
Decrements > 10%
Decrements > 15%
Decrements > 20%
75ppb
70ppb GSppb
60 ppb 75 ppb
70 ppb 65 ppb
Standard Level
60 ppb 75ppb 70ppb 65 ppb
60 ppb
-Atlanta
-Baltimore
-Boston
-Chicago
-Cleveland
-Dallas
-Denver
-Detroit
Houston
-Los Angeles
-NewYork
Philadelphia
Sacramento
St Louis
Washington
4-32
-------
Figure 4-8. Percent of children estimated to experience two or more Os-induced lung function decrements greater than 10,
15, or 20% for air quality adjusted to just meet the current and potential alternative standards (worst-case year
from 2006 to 2010)
~C a1
v S
o
V
Decrements > 10%
Decrements > 15%
Decrements > 20%
75ppb 70ppb
65ppb 60ppb 75ppb 70ppb 65ppb 60ppb 75ppb 70ppb 65ppb 60 ppb
Standard Level
• Atlanta
—•—Baltimore
* Boston
—M—Chicago
.'i1 Cleveland
• Dallas
I Denver
Detroit
Houston
» Los Angeles
-•-New York
Philadelphia
Sacramento
I St Louis
Washington
4-33
-------
As illustrated above in Figures 4-5 to 4-8, adjusting air quality to just meet progressively
lower potential alternative standard levels consistently reduces the percent of children estimated
to experience potentially adverse lung function decrements. These results reflect the consistent
reductions in the highest ambient Os concentrations upon air quality adjustment (section 3.2.1;
U.S. EPA, 2014, chapter 4).20 Based on Figures 4-5 to 4-8 and the associated details described in
the HREA (U.S. EPA 2014, chapter 6), we take note of the following with regard to specific
potential alternative standard levels:
1. For an Os standard level of 70 ppb:
a. On average over the years 2006 to 2010, a standard with a level of 70 ppb is
estimated to allow approximately 11 to 17% of children in urban case study areas,
including asthmatic children, to experience one or more Os-induced lung function
decrements > 10% (approximately 6 to 27% reduction, relative to current
standard) per season. Summing across case study areas, these percentages
correspond to approximately 260,000 asthmatic children experiencing
approximately 1 million total occurrences of Os-induced lung function
decrements greater than or equal to 10%.
b. On average over the years 2006 to 2010, a standard with a level of 70 ppb is
estimated to allow approximately 6 to 11% of children, including asthmatic
children, to experience two or more Os-induced lung function decrements > 10%
(approximately 8 to 30% reduction, relative to current standard).
c. In the worst-case years, a standard with a level of 70 ppb is estimated to allow
approximately 14 to 20% of children, including asthmatic children, to experience
one or more Os-induced lung function decrements >10%, and approximately 7 to
13% to experience two or more such decrements.
d. On average over the years 2006 to 2010, a standard with a level of 70 ppb is
estimated to allow approximately 2 to 4% of children, including asthmatic
children, to experience one or more Os-induced lung function decrements > 15%,
and approximately 1 to 2.5% of children to experience two or more such Os-
induced decrements. In the worst-case years, approximately 3 to 5% of children
are estimated to experience one or more Os-induced lung function decrements
>15%, and approximately 1 to 3% are estimated to experience two or more such
decrements.
e. A standard with a level of 70 ppb is estimated to allow 2% or fewer children to
experience any Os-induced lung function decrements > 20%, even in the worst-
case years. Approximately 1% or fewer children are estimated to experience two
20 As discussed in section 3.2.3.1, the impact of the dose threshold in the lung function risk model is that O3-induced
FEVi decrements result primarily from exposures on days with average ambient O3 concentrations above about 40
ppb (US EPA, 2014, section 6.3.1, Figure 6-9).
4-34
-------
or more Os-induced lung function decrements > 20%, even in the worst-case
years.
2. For an Os standard level of 65 ppb:
a. On average over the years 2006 to 2010, a standard with a level of 65 ppb is
estimated to allow approximately 3 to 15% of children, including asthmatic
children, to experience one or more Os-induced lung function decrements > 10%
(approximately 20 to 77% reduction, relative to current standard). Summing
across urban case study areas, these percentages correspond to approximately
190,000 asthmatic children experiencing almost 750,000 total occurrences of Cb-
induced lung function decrements > 10%.
b. On average over the years 2006 to 2010, a standard with a level of 65 ppb is
estimated to allow approximately 1 to 9% of children, including asthmatic
children, to experience two or more Os-induced lung function decrements > 10%
(approximately 20 to 80% reduction, relative to current standard).
c. In the worst-case years, a standard with a level of 65 ppb is estimated to allow
approximately 4 to 18% of children to experience one or more Os-induced lung
function decrements > 10%, and approximately 2 to 11% to experience two or
more such decrements.
d. On average over the years 2006 to 2010, a standard with a level of 65 ppb is
estimated to allow approximately 3% or less of children to experience one or
more Os-induced lung function decrements > 15%, and approximately 2% or less
of children to experience two or more such Os-induced decrements. In the worst-
case years, approximately 4% or less of children are estimated to experience one
or more Os-induced lung function decrements > 15%, and up to approximately
2% are estimated to experience two or more such decrements.
e. A standard with a level of 65 ppb is estimated to allow less than 1.5% of children
to experience any Os-induced lung function decrements > 20%, even in the worst-
case years. A standard with a level of 65 ppb is estimated to allow less than 1% of
children to experience two or more Os-induced lung function decrements > 20%,
even in the worst-case years.
3. For an Os standard level of 60 ppb:
a. On average over the years 2006 to 2010, a standard with a level of 60 ppb is
estimated to allow approximately 5 to 11% of children, including asthmatic
children, to experience one or more Os-induced lung function decrements > 10%
(approximately 35 to 77% reduction, relative to current standard). Summing
across urban case study areas, these percentages correspond to approximately
140,000 asthmatic children experiencing approximately 500,000 total occurrences
of Os-induced lung function decrements > 10%.
4-35
-------
b. On average over the years 2006 to 2010, a standard with a level of 60 ppb is
estimated to allow approximately 2 to 6% of children to experience two or more
Os-induced lung function decrements > 10% (approximately 40 to 70% reduction,
relative to current standard).
c. In the worst-case years, a standard with a level of 60 ppb is estimated to allow
approximately 5 to 13% of children to experience one or more Os-induced lung
function decrements > 10%, and approximately 2 to 7% to experience two or
more such decrements.
d. A standard with a level of 60 ppb is estimated to allow less than about 3% of
children to experience any Os-induced lung function decrements > 15% and less
than 1% to experience decrements greater than 20%, even in the worst-case years.
A standard with a level of 60 ppb is estimated to allow less than 1.5% of children
to experience two or more Os-induced lung function decrements > 15% and less
than 0.5% to experience two or more decrements > 20%, even in the worst-case
years.
In further considering these exposure estimates, we take note of the associated
uncertainties, as discussed in more detail in section 3.2.2 of this PA. In addition to the
uncertainties in exposure estimates noted above, these include the relative lack of exposure-
response information for key at-risk populations (i.e., children and asthmatics), since most
controlled human exposures studies are conducted in healthy adults. Section 4.6 (below)
discusses the implications of estimates of the occurrence of Cb-induced lung function decrements
for potential alternative standard levels.
4.4.2.3 Risk-Based Considerations: Epidemiology-Based Mortality and Morbidity
The epidemiology-based risk assessments presented in the HREA (U.S. EPA, 2014,
chapter 7) provide estimates of total mortality, respiratory hospital admissions and emergency
department visits, and asthma exacerbations associated with short-term Os concentrations. The
HREA also presents estimates of respiratory mortality associated with long-term21
concentrations. In evaluating these risk estimates, we consider the following question:
9 1
Estimates of respiratory mortality associated with long-term Os concentrations are based on the study by Jerrett et
al. (2009). Consistent with the Os metric used in the study, risk estimates are based on seasonal averages of 1-hour
daily max Os concentrations.
4-36
-------
• To what extent are potential alternative standards with revised levels estimated to
decrease Os health risks, compared to the current standard, and what are the
nature and magnitude of the health risks remaining for each alternative standard
level evaluated?
As discussed in more detail in section 3.2.3.2 of this PA, in considering this question we
are mindful that the model-based approach used to adjust air quality in the HREA has important
implications for risk estimates developed by applying concentration-response relationships from
epidemiologic studies (section 3.2.1). In particular, given the use of linear concentration-
response relationships, risk estimates are equally influenced by decreasing high Os
concentrations and increasing low Os concentrations following air quality adjustment, when the
increases and decreases are of equal magnitude. This and other uncertainties associated with risk
estimates are discussed in section 3.2.3.2.
Key results from the HREA (U.S. EPA, 2014, chapter 7) are summarized below for
estimates of total mortality associated with short-term Os concentrations (Figures 4-9 and 4-10)
and respiratory hospital admissions associated with short-term Os concentrations (Figure 4-11).
The other morbidity effects evaluated in the HREA (i.e., respiratory emergency department visits
and asthma symptoms associated with short-term concentrations) exhibit patterns across standard
levels that are similar to those reported for total mortality and respiratory hospital admissions
(U.S. EPA, 2014, chapter?).
4-37
-------
Figure 4-9. Estimates of Total Mortality Associated with Short-Term Os Concentrations
in Urban Case Study Areas (Air Quality Adjusted to Current and Potential
alternative standard levels) - Total Risk
2007 Simulation year
20
0)
S 16
R 1/1
1
g.12
^ 10 ••••
o
E „
-D 8
OJ
ro
•5 6
i
° 4
o
"ro
o 2
Trend in ozone-related mortality across standard
levels (deaths per 100,000)
Atlanta, GA
-v-.— .— -— m Baltimore, MD
"^^^^^^-^^^ — *^^~*^^^^_^ ' "• Boston, MA
yt—^^,,^^^^^ -^-Cleveland, OH
'•^--'^^^ Denver CO
' — .'^Detroit, Ml
^^— Los Angeles, CA
-:* " i&
Sacramento, CA
St. Louis, MO
75ppb
70ppb
65ppb
60ppb
2009 Simulation year
Trend in ozone-related mortality across standard levels
(deaths per 100,000)
E 18
cu
1 16
O
I H
S
a 12
I8
Tu 6
—1—Atlanta, GA
^^Baltimore, MD
Boston, MA
-**-Cleveland, OH
Denver, CO
^.—Detroit, Ml
Houston, TX
^^Los Angeles, CA
New York, NY
—a— Philadelphia, PA
Sacramento, CA
St. Louis, MO
75ppb 70ppb 65ppb 60ppb
The risk estimates presented in Figure 4-9 above are based on applying linear
concentration-response relationships to the full distributions of daily 8-hour "area-wide" Os
concentrations. However, as in section 3.2.3.2 we note the ISA conclusion that there is less
certainty in the shape of concentration-response functions for area-wide Os concentrations at the
lower ends of warm season distributions (i.e., below about 20 to 40 ppb depending on the Os
4-38
-------
metric, health endpoint, and study population) (U.S. EPA, 2013, section 2.5.4.4). We also
recognize that for the range of health endpoints evaluated, controlled human exposure and
animal toxicological studies provide greater certainty in the increased incidence, magnitude, and
severity of effects at higher exposure concentrations (discussed in sections 3.1.2.2 and 3.1.4.2 of
this document).22 Thus, in addition to considering estimates of total Os-associated risks, we also
consider the extent to which risks are associated with days with higher, versus lower, area-wide
Os concentrations.
Figure 4-10 presents estimates of Os-associated deaths, summed across urban case study
areas, for days with area-wide concentrations at or above 20, 40, and 60 ppb. As discussed in
more detail in section 3.2.1 of this document, daytime Os concentrations in the upper portions of
the distributions of area-wide concentrations tend to decrease upon adjustment to meet lower
potential alternative standard levels, while concentrations in the lower portions of these
distributions tend to increase. As a result, lower standard levels are estimated to be more
effective at reducing deaths associated with the upper portions of these distributions of ambient
Os concentrations than deaths associated with the full distributions.23
22As discussed in section 3.1.4.2, as ambient concentrations increase the potential for exposures to higher Os
concentrations also increases. Thus with increasing ambient concentrations, controlled human exposure and animal
toxicological studies provide greater certainty in the increased incidence, magnitude, and severity of Os-attributable
effects.
23The relatively small proportion of O3-associated deaths attributable to days with area-wide concentrations of 60
ppb or greater reflects the relatively small proportion of days with such elevated area-wide concentrations.
4-39
-------
Figure 4-10. Estimates of Os-Associated Deaths Attributable to Full Distribution of 8-Hour Area-Wide Os Concentrations and
to Concentrations at or above 20, 40, or 60 ppb - Deaths Summed Across Urban Case Study Areas24
2007 Model Adjustment
2009 Model Adjustment
tJ
ID
E
E
zs
to
to"
£
Jr
T3
In
"ro
"o
o
to
to
TO
TO
O
CO
TO
CD
ro
-_T
"to
-------
Figure 4-11. Estimates of Respiratory Hospital Admissions Associated with Short-Term Os
Concentrations in Urban Case Study Areas (Air Quality Adjusted to Current
and Potential alternative standard levels) - Total Risk
2007 Simulation year
Trend in ozone-related HA across standard levels
(HA per 100,000)
•Atlanta, GA
-Baltimore, MD
Boston, MA
-Cleveland, OH
-Denver, CO
Detroit, Ml
-Houston, TX
• Los Angeles, CA
-New York, NY
-Philadelphia, PA
Sacramento, CA
St. Louis, MO
75ppb
70ppb
65ppb
60ppb
2009 Simulation year
Trend in ozone-related HA across standard levels
(HA per 100,000)
-Atlanta, GA
-Baltimore, MD
Boston, MA
-Cleveland, OH
-Denver, CO
Detroit, Ml
-Houston, TX
-Los Angeles, CA
New York, NY
-Philadelphia, PA
Sacramento, CA
St. Louis, MO
75ppb
70ppb
65ppb
60ppb
4-41
-------
Key results from the HREA (U.S. EPA, 2014, chapter 7) are summarized in Figure 4-12
below for estimates of respiratory mortality associated with long-term Os concentrations, based
on the study by Jerrett et al. (2009). As discussed in section 3.2.3.2 of this PA, Jerrett et al.
(2009) reported that when seasonal averages of 1-hour daily maximum Os concentrations ranged
from 33 to 104 ppb, there was no statistical deviation from a linear concentration-response
relationship between Os and respiratory mortality across 96 U.S. cities (U.S. EPA, 2013, section
7.7). However, the authors reported "limited evidence" for an effect threshold at an Os
concentration of 56 ppb (p=0.06). In communications with EPA staff (described in Sasser, 2014),
the study authors indicated that it is not clear whether a threshold model is a better predictor of
respiratory mortality than the linear model, and that "considerable caution should be exercised in
accepting any specific threshold." Consistent with this communication, the HREA estimated
respiratory mortality associated with long-term Os concentrations based on the linear model from
the published study, and in a series of sensitivity analyses with models that included thresholds
ranging from 40 to 60 ppb (U.S. EPA, 2014, Figure 7-9).
Figure 4-12 presents estimates of total Os-associated respiratory deaths, based on a linear
concentration-response relationship. As discussed for the current standard (section 3.2.3.2),
HREA sensitivity analyses indicate that, if a threshold exists between 40 and 60 ppb, the number
of respiratory deaths associated with long-term Os concentrations could potentially be
considerably smaller than indicated by the no threshold model (U.S. EPA, 2014, Figure 7-9).
4-42
-------
Figure 4-12. Estimates of Respiratory Mortality Associated with long-term Os
Concentrations in Urban Case Study Areas (Air Quality Adjusted to Current
and Potential alternative standard levels) - Total Risk
2007 Simulation Year
Trend in ozone-related mortality across standard
levels (deaths per 100,000)
t
I-
-*-.-'.ta'ta.w/.
• ba [.more, MD
—*- toston. MA
•Cleveland, OH
-Denver. CO
-Detroit, Ml
-Houston,!*
-LosAngeles.CA
-New 10':. I-.1
—PhllaQ'-:: D1 3. !'•"'
-Sacramento, DV
-St. Lous, MO
2009 Simulation Year
Trend in ozone-related mortality across standard
levels (deaths per 100,000)
—*— Atlanta <;-
-•-EaltinBre, MD
—*-Boston Mi
—:k slaiyl OH
—•—terr-'sr CO
-*-Detroit Ml
—+—Houston T>;
——Los ingles C^.
— (tew tonV, Mv
HlllH-l-l|-llB P -
--,;-Sacram=nto,<:i
St. Louis, MO
75ppb
-oppb
;ppb
60ppb
Based on Figures 4-9 to 4-12 and the associated details described in the HREA (U.S.
EPA 2014, chapter 7), we take note of the following for an Os standard level of 70 ppb:
1. Total mortality associated with short-term Os concentrations:
4-43
-------
a. Across urban case study areas, risks are estimated to decrease by up to
approximately 5% for a standard level of 70 ppb, compared to the current
standard. Risk reductions are estimated consistently for the model year with
generally higher Os-associated risks (2007). In the year with generally lower risks
(2009), a standard level of 70 ppb results in either no change or more modest
reductions in estimated risks in most urban case study areas. In one area (Detroit)
for the 2009 model year, Os-associated mortality is estimated to increase by
approximately 4%, compared to the current standard (see section 3.2.3.2 for
further discussion of increased risk estimates following air quality adjustment25).
b. When summed across urban case study areas, a standard level of 70 ppb is
estimated to reduce Os-associated deaths by approximately 4% (2007 model year)
and 2% (2009 model year), compared to the current standard. For area-wide
concentrations at or above 40 ppb, a standard level of 70 ppb is estimated to
reduce Os-associated deaths by approximately 10% (2007 model year) and 9%
(2009 model year). For area-wide concentrations at or above 60 ppb, a standard
level of 70 ppb is estimated to reduce Os-associated deaths by approximately 50%
(2007 model year) and 70% (2009 model year).26
2. Respiratory hospital admissions associated with short-term Os concentrations: Compared to
the current standard, changes in total risk estimated for a standard level of 70 ppb are similar
to the changes in total risks estimated for total mortality (U.S. EPA, 2014, chapter 7).
3. Respiratory mortality associated with long-term Os concentrations: A standard level of 70
ppb reduces total risk, compared to the current standard. Across urban case study areas, risks
are estimated to decrease by up to approximately 6%. These risk reductions are estimated
most consistently for the model year with generally higher Cb-associated risks (2007). In the
year with generally lower Os concentrations (2009), a standard level of 70 ppb results in
smaller reductions in estimated risks in most urban case study areas. In one area (Detroit) for
the 2009 model year, Os-associated mortality is estimated to increase by approximately 1%,
compared to the current standard.
Based on Figures 4-9 to 4-12 and the associated details described in the HREA (U.S.
EPA 2014, chapter 7), we take note of the following for an Os standard level of 65 ppb:
1. Total mortality associated with short-term Os concentrations:
a. Across most urban case study areas, risks are estimated to decrease by up to
approximately 9% for a standard level of 65 ppb, compared to the current
25 As discussed in more detail above (section 3.2.3.2), because of the influence of the entire distribution of ambient
Os concentrations on total risk estimates, the impacts of adjusting air quality to just meet potential alternative
standards are more modest, and are less directionally consistent across urban case study areas, than observed for
exposures of concern or Os-induced lung function decrements.
26These results reflect the fact that increases in area-wide Os concentrations upon air quality adjustment occur
primarily at relatively low concentrations (i.e., on days with initial Os concentrations in the range of 10 to 40) (U.S.
EPA, 2014, section 4.3.3.2 and appendix 7B, section 9.6).
4-44
-------
standard. In one area (New York City), risks are estimated to decrease by up to
approximately 22%.27 These risk reductions are estimated most consistently for
the model year with generally higher Os-associated risks (2007). In the year with
generally lower risks (2009), a standard level of 65 ppb results in smaller
reductions in estimated risks in most urban case study areas. In one area (Detroit)
for the 2009 model year, Os-associated mortality is estimated to increase by
approximately 1% compared to the current standard.
b. When summed across urban case study areas, a standard level of 65 ppb is
estimated to reduce Os-associated deaths by approximately 13% (2007 model
year) and 9% (2009 model year), compared to the current standard. For area-wide
concentrations at or above 40 ppb, a standard level of 65 ppb is estimated to
reduce Os-associated deaths by approximately 47% (2007) and 46% (2009). For
area-wide concentrations at or above 60 ppb, a standard level of 65 ppb is
estimated to reduce Os-associated deaths by over 80% (2007 and 2009 model
years).
2. Respiratory hospital admissions associated with short-term Os concentrations: Compared to
the current standard, changes in total risk estimated for a standard level of 65 ppb are similar
to the changes in total risk estimated for total mortality (U.S. EPA, 2014, chapter 7).
3. Respiratory mortality associated with long-term Os concentrations: A standard level of 65
ppb reduces total risk, compared to the current standard. Across most urban case study areas,
risks are estimated to decrease by up to approximately 10%. In one area (New York City),
risks are estimated to decrease by up to approximately 24%. Risk reductions are estimated
across all urban case study areas and in both model years evaluated, with larger reductions
estimated for 2007 (i.e., the model year with generally higher Os-associated risks).
Based on Figures 4-9 to 4-12 and the associated details described in the HREA (U.S.
EPA 2014, chapter 7), we take note of the following for an Os standard level of 60 ppb:
1. Total mortality associated with short-term Os concentrations:
a. A standard level of 60 ppb is estimated to reduce total risk, compared to the
current standard, in all urban case study areas. Across urban case study areas,
risks are estimated to decrease by up to approximately 14%. Estimated risk
reductions are larger for the model year with generally higher Os-associated risks
(2007).
b. When summed across urban case study areas, a standard level of 60 ppb is
estimated to reduce Os-associated deaths by approximately 15% (2007 model
year) and 11% (2009 model year), compared to the current standard. For area-
wide concentrations at or above 40 ppb, a standard level of 60 ppb is estimated to
27 Because of the approach to adjusting air quality in New York (and Los Angeles), which differed from other urban
case study areas (U.S. EPA, 2014, sections 4.3.3.1, 4.5), the HREA notes less overall confidence in results for these
areas.
4-45
-------
reduce Os-associated deaths by almost 60% (2007 and 2009 model years). For
area-wide concentrations at or above 60 ppb, a standard level of 60 ppb is
estimated to reduce Os-associated deaths by over 95% (2007 and 2009 model
years).
2. Respiratory hospital admissions associated with short-term Os concentrations: Compared to
the current standard, changes in total risk estimated for a standard level of 60 ppb are similar
to the changes in total risk estimated for total mortality (U.S. EPA, 2014, chapter 7).
3. Respiratory mortality associated with long-term Os concentrations: A standard level of 60
ppb reduces total risk, compared to the current standard. Across urban case study areas, risks
are estimated to decrease by up to approximately 17%. Risk reductions are estimated across
all urban case study areas and in both model years evaluated, with larger reductions
estimated for 2007 (i.e., the model year with generally higher Os-associated risks).
In further considering these risk estimates, we take note of the associated uncertainties, as
discussed in more detail in section 3.2.3.2 of this PA. In particular, these include (1) the national
representativeness of urban case study areas in terms of the Os response to reductions in NOx
emissions; (2) the representativeness of risk changes based primarily on reductions in NOx
emissions versus changes that could be achieved with better-optimized emissions reduction
strategies; (3) the shape of the concentration-response function at lower ambient concentrations,
including the potential for a threshold in the association between long-term Os and respiratory
mortality; (4) the presence of unexplained heterogeneity in effect estimates between locations;
(5) the potential for exposure measurement errors; and (6) the possibility for reductions in risk
associated with reductions in PM and/or NCh resulting from control of NOx.
4.5 CASAC ADVICE AND PUBLIC COMMENTERS' VIEWS ON
ALTERNATIVE STANDARDS
As discussed in section 3.3, staff recognizes that decisions regarding the weight to place
on various types of evidence, exposure/risk information, and associated uncertainties reflect
public health policy judgments that are ultimately left to the Administrator. To help inform those
judgments with regard to the range of alternative primary Os standards appropriate for
consideration, CASAC has provided advice to the Administrator based on their reviews of the Os
ISA, HREA, and PA. This section summarizes the advice provided by CASAC regarding
potential alternative standards, as well as the views expressed at the CASAC meetings by public
commenters.
In the fall of 2011, rather than revising the Os NAAQS as part of the reconsideration
process, EPA elected to coordinate further proceedings on the reconsideration rulemaking with
the current ongoing periodic review. Accordingly, in this section we briefly describe CASAC
advice from the reconsideration of the 2008 final decision on the level of the standard, as well as
4-46
-------
CASAC advice received during the current review as it pertains to potential alternative
standards.
Consistent with their advice in 2008, CASAC reiterated during the reconsideration its
support for an 8-hour primary Os standard with a level ranging from 60 to 70 ppb, combined
with the current indicator, averaging time, and form. Specifically, in response to EPA's
solicitation of their advice during the reconsideration, the CASAC letter (Samet 2010) to the
Administrator stated:
CASAC fully supports EPA 's proposed range of 0.060 - 0.070 parts per million
(ppm) for the 8-hour primary ozone standard. CASAC considers this range to be
justified by the scientific evidence as presented in the Air Quality Criteria for
Ozone and Related Photochemical Oxidants (March 2006) and Review of the
National Ambient Air Quality Standards for Ozone: Policy Assessment of
Scientific and Technical Information, OAQPS Staff Paper (July 2007).
Similarly, in response to EPA's request for additional advice on the reconsideration in
2011, CASAC reaffirmed their conclusion that "the evidence from controlled human and
epidemiological studies strongly supports the selection of a new primary ozone standard within
the 60 - 70 ppb range for an 8-hour averaging time" (Samet, 2011). CASAC further concluded
that this range "would provide little margin of safety at its upper end" (Samet, 2011, p. 2).
In the current review of the Second Draft PA, as noted above, CASAC concurred with
staffs conclusions that it is appropriate to consider retaining the current indicator (Os), averaging
time (8-hour average) and form (three-year average of the 4th highest maximum daily 8-hour
average. With regard to level, CASAC stated the following (Frey, 2014, p. ii to iii):
The CASAC further concludes that there is adequate scientific evidence to
recommend a range of levels for a revised primary ozone standard from 70 ppb to
60 ppb. The CASAC reached this conclusion based on the scientific evidence from
clinical studies, epidemiologic studies, and animal toxicology studies, as
summarized in the Integrated Science Assessment (ISA), the findings from the
exposure and risk assessments as summarized in the HREA, and the interpretation
of the implications of these sources of information as given in the Second Draft
PA.
The CASAC acknowledges that the choice of a level within the range
recommended based on scientific evidence [i.e., 70 to 60 ppb] is a policy
judgment under the statutory mandate of the Clean Air Act. The CASAC advises
that, based on the scientific evidence, a level of 70 ppb provides little margin of
safety for the protection of public health, particularly for sensitive
subpopulations.
Thus, our policy advice is to set the level of the standard lower than 70 ppb within
a range down to 60 ppb, taking into account your judgment regarding the desired
4-47
-------
margin of safety to protect public health, and taking into account that lower levels
will provide incrementally greater margins of safety.
The public commenters who expressed the view that the current primary Os standard is
not adequate (section 3.3) also submitted comments that supported revising the level of the
primary Os standard. Several of these commenters expressed the view that the level should be
revised to the lower end of the range of 70 to 60 ppb, or in some cases to a level below 60 ppb.
The basis for these commenters' views on the level of the standard is generally reflected in the
rationale given by CASAC for their advice, and is discussed in section 3.3 of this PA. Public
commenters who expressed the view that revision of the current standard is not necessary did not
provide any provisional views on alternative levels that would be appropriate for consideration
should the Administrator consider revisions to the standard. These views are also discussed in
section 3.3 of this PA.
4.6 STAFF CONCLUSIONS ON ALTERNATIVE PRIMARY STANDARDS FOR
CONSIDERATION
Staffs consideration of alternative primary Os standards builds upon our conclusion,
discussed in section 3.4, that the overall body of evidence and exposure/risk information call into
question the adequacy of public health protection afforded by the current standard, particularly
for at-risk populations. We further conclude that it is appropriate in this review to consider
alternative standards that would increase public health protection, compared to the current
standard.
As discussed in sections 4.1 to 4.3 above, in the current review we conclude that it is
appropriate for the Administrator to consider retaining Os as the indicator for the standard that
protects against exposures to ambient Os and other photochemical oxidants (section 4.1), and to
consider retaining the current averaging time (section 4.2) and form (section 4.3) for the primary
Os standard. For a primary Os standard that is defined in terms of the current indicator, averaging
time, and form, we reach the conclusion that, depending on the public health policy judgments
made by the Administrator, the scientific evidence and exposure/risk information available in
this review support considering alternative Os standard levels from 70 down to 60 ppb. The basis
for this conclusion is discussed in detail in section 4.4 of this PA, and is summarized in this
section.
Below, we summarize our approach to considering the scientific evidence and
exposure/risk information, and the specific evidence and information that supports the range of
levels from 70 to 60 ppb. In doing so, we focus particularly on the evidence and information as it
relates to the upper (70 ppb), middle (65 ppb), and lower (60 ppb) portions of this range. Key
exposure/risk information is summarized in Tables 4-4, and 4-5, and Figure 4-13.
4-48
-------
Table 4-4 Summary of Estimated Exposures of Concern for Potential Alternative Os
Standard Levels of 70, 65, and 60 ppb in Urban Case Study Areas28
Benchmark
Level
> 70 ppb
> 60 ppb
> 70 ppb
> 60 ppb
Alternative
Standard
Level (ppb)
70
65
60
70
65
60
70
65
60
70
65
60
Average %
Children
Exposed29
One or
0.1-1.2
0-0.2
O31
3.3-10.2
0-4.2
0-1.2
Two or
0-0.1
0
0
0.5-3.5
0-0.8
0-0.2
Number of Children (5
to 18 years)
[Number of Asthmatic
Children]30
more exposures of concern
94,000 [10,000]
14,000 [2,000]
1,400 [200]32
1,176,000 [126,000]
392,000 [42,000]
70,000 [8,000]
more exposures of concern
5,400 [600]
300 [100]
0[0]
320,000 [35,000]
67,000 [7,500]
5,100 [700]
Average %
Reduction from
Current
Standard
per season
73
95
100
46
80
96
per season
95
100
100
61
92
100
% Children -
Worst Year and
Worst Area
3.2
0.5
0.1
18.9
9.5
2.2
0.4
0
0
9.2
2.8
0.3
28 As illustrated above in Figures 4-1 to 4-4, all alternative standard level s evaluated in the HREA were effective at
limiting exposures of concern at or above 80 ppb. Therefore, Table 4-4 focuses on exposures of concern at or above
the 70 and 60 ppb benchmark concentrations.
29 Estimates for each urban case study area were averaged for the years evaluated in the HREA (2006 to 2010).
Ranges reflect the ranges across urban case study areas.
30 Numbers of children exposed in each urban case study area were averaged over the years 2006 to 2010. These
averages were then summed across urban case study areas. Numbers are rounded to nearest thousand unless
otherwise indicated.
31 Estimates smaller than 0.1% were rounded to zero.
32As discussed in section 4.3.3 of the HREA, the model-based air quality adjustment approach used to estimate risks
associated with the current and alternative standards was unable to estimate the distribution of ambient Os
concentrations in New York City upon just meeting an alternative standard with a level of 60 ppb. Therefore, for the
60 ppb standard level the numbers of children and asthmatic children reflect all of the urban case study areas except
New York.
4-49
-------
Table 4-5 Summary of Estimated Lung Function Decrements for Potential Alternative
Os Standard Levels of 70, 65, and 60 ppb in Urban Case Study Areas
Lung
Function
Decrement
> 10%
> 15%
> 20%
> 10%
> 15%
> 20%
Alternative
Standard
Level
70
65
60
70
65
60
70
65
60
70
65
60
70
65
60
70
65
60
Average %
Children33
11-17
3-15
5-11
2-4
0-3
1-2
1-2
0-1
0-1
5.5-11
1.3-8.8
2.1-6.4
0.9-2.4
0.1-1.8
0.2-1.0
0.3-0.8
0-0.5
0-0.2
Number of Children (5
to 18 years) [Number of
Asthmatic Children]34
One or more decrements
2,527,000 [261,000]
1,896,000 [191,000]
1,404,000 [139,000]35
562,000 [58,000]
356,000 [36,000]
225,000 [22,000]
189,000 [20,000]
106,000 [11,000]
57,000 [6,000]
Two or more decrements
1,414,000 [145,000]
1,023,000 [102,000]
741,000 [73,000]
276,000 [28,000]
168,000 [17,000]
101,000 [10,000]
81,000 [8,000]
43,000 [4,000]
21,000 [2,000]
Average %
Reduction from
Current Standard
per season
15
31
45
26
50
67
32
59
77
per season
17
37
51
29
54
71
34
66
83
% Children
Worst Year and
Area
20
18
13
5
4
3
2.1
1.4
0.7
13
11
7.3
3.1
2.3
1.4
1.1
0.8
0.4
33 Estimates in each urban case study area were averaged for the years evaluated in the HREA (2006 to 2010).
Ranges reflect the ranges across urban case study areas.
34 Numbers of children estimated to experience decrements in each study urban case study area were averaged over
2006 to 2010. These averages were then summed across urban case study areas. Numbers are rounded to nearest
thousand unless otherwise indicated. As discussed above, for the 60 ppb standard level the numbers of children and
asthmatic children included in Table 4-5 reflect all of the urban case study areas except New York.
35As discussed in section 4.3.3 of the HREA, the model-based air quality adjustment approach used to estimate risks
associated with the current and alternative standards was unable to estimate the distribution of ambient Os
concentrations in New York City upon just meeting an alternative standard with a level of 60 ppb. Therefore, for the
60 ppb standard level the numbers of children and asthmatic children reflect all of the urban case study areas except
New York.
4-50
-------
Figure 4-13. Estimates of Os-Associated Deaths Attributable to Full Distributions of 8-
Hour Area-Wide Os Concentrations and to Concentrations at or above 20, 40, or 6036 ppb
Os - Deaths Summed Across Urban Case Study Areas and Expressed Relative to a
Standard with a Level of 75 ppb
1.2
I 75 ppb
I 70 ppb
i65ppb
, 00 ppb
Total
20+ ppb
40+ ppb
60+ ppb
36As discussed in section 4.3.3 of the HREA, the model-based air quality adjustment approach used to estimate risks
associated with the current and alternative standards was unable to estimate the distribution of ambient Os
concentrations in New York City upon just meeting an alternative standard with a level of 60 ppb. Therefore, the
total number of deaths indicated for the 60 ppb standard level in Figure 4-10 reflects the 60 ppb estimates for all
urban case study areas except New York City. For New York City, the estimated number of Os-associated deaths for
the 65 ppb standard level was assumed.
4-51
-------
Summary of approach to reaching conclusions on alternative standard levels
In this PA, our approach to reaching conclusions on alternative standard levels focuses on
the evidence from controlled human exposure and epidemiologic studies, as assessed in the ISA
(U.S. EPA, 2013), and the exposure and health risk analyses presented in the HREA (U.S. EPA,
2014). This approach is discussed in detail in Chapter 1 (section 1.3), and is summarized below.
As an initial matter, we note that controlled human exposure studies provide the most
certain evidence indicating the occurrence of health effects in humans following exposures to
specific Os concentrations. Consistent with this, CASAC concluded that "the scientific evidence
supporting the finding that the current standard is inadequate to protect public health is strongest
based on the controlled human exposure studies of respiratory effects" (Frey, 2014, p. 5). As
discussed above and in section 3.1.2.1, controlled human exposure studies have reported a
variety of respiratory effects in healthy adults following exposures to Os concentrations of 60,
72,37 or 80 ppb, and higher. The largest respiratory effects, and the broadest range of effects,
have been studied and reported following exposures of healthy adults to 80 ppb Os or higher,
with most exposure studies conducted at these higher concentrations. Exposures to Os
concentrations of 80 ppb or higher have been reported to decrease lung function, increase airway
inflammation, increase respiratory symptoms, result in airway hyperresponsiveness, and decrease
lung host defenses in healthy adults.
Most of these effects have also been reported in healthy adults following exposures to Os
concentrations below 80 ppb.38 Exposures to Os concentrations of 72 ppb have been reported to
decrease lung function and increase respiratory symptoms, a combination that meets the ATS
criteria for an "adverse" response (section 3.1.3). Exposures to Os concentrations of 60 ppb have
been demonstrated to decrease lung function, with decrements in some people large enough to be
judged an abnormal response by ATS, and which CASAC has indicated could be adverse to
people with lung disease.39 In addition, as discussed in section 3.1.3, such a decrease in mean
lung function meets the ATS criteria for an adverse response given that a downward shift in the
distribution of FEVi would result in diminished reserve function, and therefore would increase
risk from further environmental insult. Exposures to Os concentrations of 60 ppb have also been
reported in one study (Kim et al., 2011) to increase airway inflammation, which provides a
37 As noted above, for the 70 ppb exposure concentration Schelegle et al. (2009) reported that the actual mean
exposure concentration was 72 ppb.
38 Airway hyperresponsiveness and reductions in lung host defense have not been evaluated following exposures to
Os concentrations below 80 ppb. The extent to which these respiratory effects occur following lower exposure
concentrations is not clear from the available evidence, though we have no basis for concluding that an exposure
concentration of 80 ppb reflects an effects threshold.
39 In their advice to the Administrator based on the second draft PA, the CASAC indicated that "60 ppb is an
appropriate exposure of concern for asthmatic children" (Frey, 2014).
4-52
-------
mechanism by which Os may cause other more serious respiratory effects (e.g., asthma
exacerbations).
Given the evidence for respiratory effects from controlled human exposure studies, we
consider the extent to which standards with revised levels would be estimated to protect at-risk
populations against exposures of concern to Os concentrations at or above the health benchmark
concentrations of 60, 70, and 80 ppb (i.e., based on HREA estimates of one or more and two or
more exposures of concern). In doing so, we note that, due to individual variability in
responsiveness, only a subset of people who experience exposures at or above the three
benchmark concentrations can be expected to experience associated health effects, and that
available data are not sufficient to quantify that subset of people. We view the health effects
evidence as a continuum with greater confidence and less uncertainty about the occurrence of
adverse health effects at higher Os exposure concentrations, and less confidence and greater
uncertainty as one considers lower exposure concentrations (discussed in more detail in section
3.2.2).
While there is greater uncertainty regarding the occurrence of adverse health effects at
lower concentrations, we also note that the controlled human exposure studies that provided the
basis for benchmark concentrations have not evaluated responses in populations at the greatest
risk from exposures to Os. Thus, the effects reported in healthy adults at each of the benchmark
concentrations may underestimate effects in these at-risk groups. Compared to the healthy people
included in most controlled human exposure studies, members of at-risk populations, including
lifestages, (e.g., asthmatics, children) are at greater risk of experiencing adverse effects. In
considering the health evidence within the context of drawing conclusions on potential
alternative standard levels, we balance concerns about the potential for adverse health effects,
especially in at-risk populations, with our increasing uncertainty regarding the likelihood of such
effects following exposures to lower Os concentrations.
With respect to the lung function decrements that have been evaluated in controlled
human exposure studies, we consider the extent to which standards with revised levels would be
estimated to protect healthy and at-risk populations against Os-induced lung function decrements
large enough to be adverse in some people (based on quantitative risk estimates in the HREA).
As discussed in section 3.1.3, although some experts would judge single occurrences of moderate
responses to be a "nuisance," especially for healthy individuals, a more general consensus view
of the adversity of moderate lung function decrements emerges as the frequency of occurrence
increases. Repeated occurrences of moderate responses, even in otherwise healthy individuals,
may be considered to be adverse, since they could well set the stage for more serious illness (61
FR 65723). For the purpose of estimating potentially adverse lung function decrements in active,
healthy people, in the 2008 review the CASAC panel indicated that a focus on the mid to upper
4-53
-------
end of the range of moderate (i.e., FEVi decrements > 15%) functional responses is appropriate.
However, for children and adults with lung disease, FEVi decrements > 10% could lead to
respiratory symptoms, would likely interfere with normal activities for many individuals, and
therefore could be adverse. Large (i.e., FEVi decrements > 20%) lung function decrements
would likely interfere with normal activities for most people with lung disease and would
increase the likelihood that they would seek medical attention. In the current review, CASAC
judges that an FEVi decrement > 15% is an appropriate surrogate for adverse health outcomes in
active healthy adults, while a decrement > 10% is a scientifically relevant surrogate for adverse
health outcomes for people with asthma and lung disease (Frey, 2014). In reaching conclusions
on alternative standard levels, we consider the extent to which standards with revised levels
would be estimated to protect healthy and at-risk populations against one or more, and two or
more, moderate (i.e., FEVi decrements > 10% and > 15%) and large (i.e., FEVi decrements >
20%) lung function decrements.
In evaluating the epidemiologic evidence within the context of drawing conclusions on
potential alternative standard levels, we consider the extent to which available studies have
reported associations with emergency department visits, hospital admissions, and/or mortality in
locations that would likely have met potential alternative standards with levels below 75 ppb
(based on analyses presented in section 4.4.1). In evaluating the epidemiologic evidence in this
way, we consider both multicity and single-city studies, recognizing the strengths and limitations
of each. Specifically, multicity studies evaluate large populations and provide greater statistical
power than single-city studies; multicity studies reflect Os-associated health impacts across a
range of diverse locations, providing spatial coverage for different regions across the country and
reflecting differences in exposure-related factors that could impact Os risks; and multicity studies
afford a greater possibility of generalizing to the national population. In contrast, while single-
city studies are more limited than multicity studies in terms of statistical power and geographic
coverage, conclusions linking air quality in a specific area with health effect associations in that
same area can be made with greater certainty for single-city studies (i.e., compared to multicity
studies reporting only multicity effect estimates).
We also consider the epidemiologic evidence within the context of epidemiology-based
risk estimates. Compared to the weight given to HREA estimates of exposures of concern and
lung function risks (sections 4.4.2.1 and 4.4.2.2, above), and the weight given to the evidence
(section 4.4.1), we place relatively less weight on epidemiologic-based risk estimates. In doing
so, we note that the overall conclusions from the HREA likewise reflect less confidence in
estimates of epidemiologic-based risks than in estimates of exposures and lung function risks.
Our determination to attach less weight to the epidemiologic-based estimates reflects the
uncertainties associated with mortality and morbidity risk estimates, including the heterogeneity
4-54
-------
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 Os
concentrations (U.S. EPA, 2014, section 9.6). The HREA also concludes that lower confidence
should be placed in the results of the assessment of respiratory mortality risks associated with
long-term Os exposures, primarily because that analysis is based on only one study (even though
that study is well-designed) and because of the uncertainty in that study about the existence and
level of a potential threshold in the concentration-response function (U.S. EPA, 2014, section
9.6).
In considering the epidemiology-based risk estimates, we focus on the extent to which
potential alternative Os standards with levels below 75 ppb are estimated to reduce the risk of
Os-associated mortality (based on the HREA results summarized in section 4.4.2.3).40 As
discussed in section 3.4 for the current standard, we consider estimates of total risk (i.e., based
on the full distributions of ambient Os concentrations) and estimates of risk associated with Os
concentrations in the upper portions of ambient distributions. A focus on estimates of total risks
would place greater weight on the possibility that concentration-response relationships remain
linear over the entire distribution of ambient Os concentrations, and thus on the potential for
mortality and morbidity to be affected by changes in relatively low Os concentrations. A focus
on risks associated with Os concentrations in the upper portions of the ambient distribution
would place greater weight on the uncertainty associated with the shapes of concentration-
response curves for Os concentrations in the lower portions of the distribution. Given that both
types of risk estimates could reasonably inform a decision on standard level, depending on the
weight placed on uncertainties in the occurrence and the estimation of Os-attributable effects at
relatively low Os concentrations, in reaching conclusions we consider what both types of
estimates indicate with regard to potential alternative levels.
Staff conclusions on the range of levels appropriate for consideration
Using the approach discussed above to consider the scientific evidence and exposure/risk
information, we reach the conclusion that it is appropriate for the Administrator to consider
alternative primary Os standard levels from 70 to 60 ppb. The basis for this conclusion is
discussed in detail in sections 4.4.1 and 4.4.2 above, and is summarized below.
With regard to controlled human exposure studies, we consider the lowest Os exposure
concentrations at which various effects have been evaluated and statistically significant effects
reported. We also consider the potential for reported effects to be adverse, including in at-risk
populations and lifestages. As discussed in section 3.1.2.1, controlled human exposure studies
40 Differences in estimated respiratory morbidity risks between potential alternative standard levels are similar to the
differences estimated for total mortality associated with short-term Os concentrations.
4-55
-------
provide evidence of respiratory symptoms combined with lung function decrements (an
"adverse" response based on ATS criteria) in healthy adults following exposures to Os
concentrations as low as 72 ppb, and evidence of potentially adverse lung function decrements
and airway inflammation following exposures to Os concentrations as low as 60 ppb. Although
some studies show that respiratory symptoms also develop during exposures to 60 ppb Os, the
increase in symptoms has not been reported to reach statistical significance by the end of the 6.6
hour exposure period (Adams 2006; Schelegle et al. 2009). Thus, while significant increases in
respiratory symptoms combined with lung function decrements have not been reported following
exposures to 60 ppb Os, this combination of effects is likely to occur to some degree in healthy
adults with 6.6-hour exposures to concentrations below 72 ppb, and also are more likely to occur
with longer (i.e., 8-hour) exposures.41
With regard to the lowest exposure concentration shown to cause respiratory effects (i.e.,
60 ppb), we note that most controlled human exposure studies have not evaluated Os
concentrations below 60 ppb. Therefore, 60 ppb does not necessarily reflect an exposure
concentration below which effects such as lung function decrements and airway inflammation no
longer occur. This is particularly the case given that controlled human exposure studies were
conducted in healthy adults, while people with asthma, including asthmatic children, are likely to
be more sensitive to Os-induced respiratory effects. In support of this, some epidemiologic panel
studies, which can include members of at-risk groups such as children and outdoor workers, have
found respiratory effects at ambient concentrations lower than 60 ppb (section 3.1.2.1).
With regard to other Os-induced effects, we note that airway hyperresponsiveness and
impaired lung host defense capabilities have been reported in healthy adults engaged in moderate
exertion following exposures to Os concentrations as low as 80 ppb, the lowest concentration
evaluated for these effects. As discussed in section 3.1.2.1, these physiological effects have been
linked to aggravation of asthma and increased susceptibility to respiratory infection, potentially
leading to increased medication use, increased school and work absences, increased visits to
doctors' offices and emergency departments, and increased hospital admissions. These are all
indicators of adverse Os-related morbidity effects, which are consistent with, and provide
plausibility for, the adverse morbidity effects and mortality effects observed in epidemiologic
studies.
41 In addition, CASAC observed that, "adverse health effects in young healthy adults occur with exposures to 72 ppb
of ozone for 6.6 hours" and that "It is the judgment of CASAC that if subjects had been exposed to ozone using the
8-hour averaging period used in the standard, adverse effects could have occurred at [a] lower concentration.
Further, in our judgment, the level at which adverse effects might be observed would likely be lower for more
sensitive subgroups, such as those with asthma" (Frey, 2014, p. 5).
4-56
-------
Based on consideration of the above evidence, we conclude that available controlled
human exposure studies support a level no higher than 70 ppb as the upper end of the range for
consideration in the current review. In reaching this conclusion, we note that 70 ppb is just below
the Os exposure concentration reported to result in lung function decrements and respiratory
symptoms in healthy adults (i.e., 72 ppb), a combination of effects that meet ATS criteria for an
adverse response. In addition, while 70 ppb is well below the 80 ppb exposure concentration
shown to cause potentially adverse respiratory effects such as airway hyperresponsiveness and
impaired host-defense capabilities, these effects have not been evaluated at exposure
concentrations below 80 ppb and there is no reason to believe that 80 ppb represents a threshold
for such effects.
We further conclude that the evidence from controlled human exposure studies42 supports
considering alternative Os standard levels at least as low as 60 ppb. Potentially adverse lung
function decrements and pulmonary inflammation have been demonstrated to occur in healthy
adults at 60 ppb. Thus, 60 ppb is a short-term exposure concentration that may be reasonably
concluded to elicit adverse effects in at-risk groups. Pulmonary inflammation, particularly if
experienced repeatedly, provides a mechanism by which Os may cause other more serious
respiratory morbidity effects (e.g., asthma exacerbations) and possibly extrapulmonary effects.
As discussed in section 3.1.2.1, the physiological effects reported in controlled human exposure
studies down to 60 ppb Cb have been linked to aggravation of asthma and increased
susceptibility to respiratory infection, potentially leading to increased medication use, increased
school and work absences, increased visits to doctors' offices and emergency departments, and
increased hospital admissions.
We further note that the range of alternative levels from 70 to 60 ppb is supported by
evidence from epidemiologic studies and by exposure and risk estimates from the HREA. This
evidence and exposure/risk information indicate that a level from anywhere in the range of 70 to
60 ppb would be expected to result in important public health improvements over the current
standard. In particular, compared to the current standard a revised standard with a level from 70
to 60 ppb would be expected to (1) more effectively maintain short- and long-term Cb
concentrations below those present in the epidemiologic studies that reported significant Os
health effect associations in locations likely to have met the current standard; (2) reduce the
occurrence of exposures of concern to Os concentrations that result in respiratory effects in
healthy adults (at or above 60, 70, and 80 ppb); (3) reduce the occurrence of moderate-to-large
Os-induced lung function decrements; and (4) reduce the risk of Os-associated mortality and
42 As discussed in sections 3.1.2.1 and 4.4. labove, panel studies also provide supporting evidence for these
conclusions.
4-57
-------
morbidity, particularly the risk associated with the upper portions of the distributions of ambient
Os concentrations.
In reaching a conclusion on whether it is appropriate to consider alternative standard
levels below 60 ppb, we note the following:
• While controlled human exposure studies provide evidence for Os-induced respiratory
effects following exposures to Os concentrations as low as 60 ppb, they do not provide
evidence for adverse effects following exposures to lower concentrations. On this issue,
CASAC concurred that 60 ppb Os is an appropriate and justifiable scientifically based
lower bound for a revised primary standard, based upon findings of "adverse effects,
including clinically significant lung function decrements and airway inflammation, after
exposures to 60 ppb ozone in healthy adults with moderate exertion (Adams 2006; Schelegle
et al, 2009; Brown et al. 2008; Kim et al, 2011), with limited evidence of adverse effects
below 60 ppb" (Frey, 2014, p. 7).
• Based on the HREA results, meeting an Os standard with a level of 60 ppb would be
expected to almost eliminate exposures of concern to Cb concentrations at or above 60
ppb. To the extent lower exposure concentrations may result in adverse health effects in
some people, a standard level of 60 ppb would be expected to also reduce exposures to
Os concentrations below 60 ppb.
• U.S. and Canadian epidemiologic studies have not reported Os health effect associations
based primarily on study locations likely to have met a standard with a level of 60 ppb.
• In all of the urban case study areas evaluated, a standard with a level of 60 ppb would be
expected to maintain long-term Os concentrations below those where a key study
indicates the most confidence in a linear concentration-response relationship with
respiratory mortality.
Beyond the above considerations, we also note the HREA estimates indicating that
meeting an Os standard with a level of 60 ppb would result in important reductions in the risk of
Os-induced lung function decrements and Cb-associated mortality and morbidity. Although some
risk is estimated to remain based on these metrics, even with a level of 60 ppb, we have
decreasing confidence in further public health improvements with levels below 60 ppb. We reach
this conclusion because, as noted above, at a level of 60 ppb virtually no one in the population
would be expected to experience exposures to Os concentrations at or above 60 ppb under
conditions demonstrated in controlled human exposure studies to result in respiratory effects, and
because epidemiologic studies have not reported Os health effect associations based primarily on
study locations likely to have met a standard with a level of 60 ppb. Given all of the above
considerations we conclude that, compared to standards with levels from 70 to 60 ppb, the extent
to which standards with levels below 60 ppb could result in further public health improvements
4-58
-------
becomes notably less certain. Therefore, we conclude that it is not appropriate in this review to
consider standard levels below 60 ppb.
The range of levels from 70 to 60 ppb corresponds to the range of levels recommended
for consideration by CASAC, based on the available evidence and information (Frey, 2014).
While CASAC further offered the "policy advice" to set the level below 70 ppb, based on margin
of safety considerations, the Committee acknowledged that "the choice of a level within the
range recommended based on scientific evidence [i.e., 70 to 60 ppb] is a policy judgment under
the statutory mandate of the Clean Air Act" (Frey, 2014). Therefore, we note that our
conclusions on the appropriate range for alternative primary Os standard levels are consistent
with CASAC conclusions that the scientific evidence and exposure/risk information supports
consideration of levels from 70 to 60 ppb, and that the ultimate identification of a standard that
protects public health with an adequate margin of safety will reflect policy judgments that are
explicitly reserved for the Administrator (section 1.2.1).
The following sections summarize the specific scientific evidence and exposure/risk
information as they relate to revised Os standards with levels from the upper (70 ppb), middle
(65 ppb), and lower (60 ppb) portions of the range of 70 to 60 ppb.
Os standard level of 70 ppb
A level of 70 ppb is below 80 ppb, an Os exposure concentration that has been reported to
elicit a range of respiratory effects that includes airway hyperresponsiveness and decreased lung
host defense, in addition to lung function decrements, airway inflammation, and respiratory
symptoms. A level of 70 ppb is also below the lowest exposure concentration at which the
combined occurrence of respiratory symptoms and lung function decrements have been reported
(i.e., 72 ppb), a combination judged adverse by the ATS (section 3.1.3). A level of 70 ppb is
above the lowest exposure concentration demonstrated to result in lung function decrements
large enough to be judged an abnormal response by ATS and above the lowest exposure
concentration demonstrated to result in pulmonary inflammation (i.e., 60 ppb).
Compared to the current standard, the FIRE A estimates that a revised Os standard with a
level of 70 ppb would reduce exposures of concern to Os concentrations of 60, 70, and 80 ppb in
urban case study areas, with such a standard level estimated to be most effective at limiting
exposures at or above the higher health benchmark concentrations and at limiting multiple
occurrences of such exposures. On average over the years 2006 to 2010, a standard with a level
of 70 ppb is estimated to allow only up to about 1% of children (i.e., ages 5 to 18) to experience
exposures of concern at or above 70 ppb (73% reduction, compared to current standard), and far
less than 1% to experience two or more such exposures (95% reduction, compared to current
standard). In the worst-case location and year (i.e., location and year with the largest exposure
4-59
-------
estimate), about 3% of children are estimated to experience one or more exposures of concern at
or above 70 ppb, and less than 1% are estimated to experience two or more. A standard with a
level of 70 ppb is estimated to allow far less than 1% of children to experience exposures of
concern at or above the 80 ppb benchmark concentration, even in the worst-case year (Table 4-
4).43
An Os standard with a level of 70 ppb is estimated to allow about 3 to 10% of children,
including asthmatic children, to experience one or more exposures of concern at or above 60 ppb
in a single Os season. As noted above, CASAC advised EPA that 60 ppb is an appropriate
exposure of concern with respect to adverse effects on people with asthma, including children
(Frey, 2014, p. 6, 8). Compared to the current standard, this reflects about a 46% reduction, on
average across the urban case study areas. A standard with a level of 70 ppb is estimated to allow
about 1% to 4% of children to experience two or more exposures of concern at or above 60 ppb.
In the worst-case location and year, a standard set at 70 ppb is estimated to allow about 19% of
children to experience one or more exposures of concern at or above 60 ppb, and 9% to
experience two or more such exposures (Table 4-4).
Compared to the current standard, the HREA estimates that a revised Os standard with a
level of 70 ppb would also reduce Os-induced lung function decrements in children. A level of
70 ppb is estimated to be most effective at limiting the occurrences of moderate and large lung
function decrements (i.e., FEVi decrements > 15% and > 20%, respectively), and at limiting
multiple occurrences of Cb-induced decrements. On average over the years 2006 to 2010, a
standard with a level of 70 ppb is estimated to allow about 2 to 4% of children in the urban case
study areas to experience one or more moderate Os-induced lung function decrements (i.e., FEVi
decrement > 15%), which would be of concern for healthy people, and about 1 to 2.5% of
children to experience two or more such decrements (approximately 30% reduction, compared to
the current standard). In the worst-case location and year, up to 5% of children are estimated to
experience one or more Os-induced lung function decrements > 15%, and up to 3% are estimated
to experience two or more such decrements. A standard set at 70 ppb is estimated to allow about
2% or fewer children to experience large Os-induced lung function decrements (i.e., FEVi
decrement > 20%), and to allow about 1% or fewer children to experience two or more such
decrements, even in the worst-case years and locations (Table 4-5).
On average over the years 2006 to 2010, an Os standard set at 70 ppb is estimated to
allow about 11 to 17% of children in the urban case study areas to experience one or more
moderate Os-induced lung function decrements (i.e., FEVi decrement > 10%), which could be
43 As noted above, due to interindividual variability, children (or adults) exposed at these levels will not necessarily
experience health effects; the information available for some health effects is not sufficient to quantify the numbers
of children in the urban case study areas who might experience these effects.
4-60
-------
adverse for people with lung disease. This reflects an average reduction of about 15%, compared
to the current standard. A standard with a level of 70 ppb is also estimated to allow about 6 to
11% of children to experience two or more such decrements (17% reduction, compared to
current standard). In the worst-case location and year, a standard set at 70 ppb is estimated to
allow about 20% of children in the urban case study areas to experience one or more Os-induced
lung function decrements > 10%, and 13% to experience two or more such decrements (Table 4-
5).
Compared to the current standard, a revised standard with a level of 70 ppb would also
more effectively maintain short-term ambient Os concentrations below those present in the
epidemiologic studies that reported significant Os health effect associations in locations likely to
have met the current standard. In particular, the single-city study by Mar and Koenig (2009)
reported positive and statistically significant associations with respiratory emergency department
visits in children and adults in a location that likely would have met the current Os standard over
the entire study period but violated a revised standard with a level of 70 ppb or below. Thus,
none of the single-city studies evaluated in section 4.4.1 provide evidence for Os health effect
associations in locations meeting a standard with a level of 70 ppb or below. While this analysis
does not provide information on the extent to which the reported Os-associated emergency
department visits would persist upon meeting an Os standard with a level of 70 ppb, or on the
extent to which standard levels below 70 ppb could further reduce the incidence of such
emergency department visits, it suggests that a revised Os standard with a level at or below 70
ppb would require reductions in the ambient Os concentrations that provided the basis for the
health effect associations reported by Mar and Koenig.
As discussed above, compared to single-city studies, there is greater uncertainty in
linking air quality concentrations from individual study cities to multicity effect estimates. With
regard to multicity studies, we note that Dales et al. (2006) reported significant associations with
respiratory hospital admissions based on air quality in 11 Canadian cities, most of which would
likely have met the current standard over the entire study period but violated a revised standard
with a level of 70 ppb or below over at least part of that period (Table 4-1). This analysis
suggests that while the current standard would allow the ambient Os concentrations in most of
the study locations that provided the basis for the association with hospital admissions, a revised
Os standard with a level at or below 70 ppb would require reductions in those ambient Os
concentrations. As with the study by Mar and Koenig, this analysis does not provide information
on the extent to which the reported Os-associated hospital admissions would persist upon
4-61
-------
meeting an Os standard with a level of 70 ppb, or on the extent to which standard levels below 70
ppb could further reduce the incidence of such hospital admissions.44
With regard to long-term Os concentrations, we evaluated the "long-term" Os metrics
reported to be associated with mortality or morbidity in recent epidemiologic studies (e.g.,
seasonal averages of 1-hour or 8-hour daily max concentrations). Compared to the current
standard, a revised standard with a level of 70 ppb would be expected to reduce the risk of
respiratory mortality associated with long-term Os concentrations, based on information from the
study by Jerrett et al. (2009), though we note the HREA conclusion, discussed above, that lower
confidence should be placed in respiratory mortality risk estimates based on this study (U.S.
EPA, 2014, section 9.6). In addition, a standard with a level of 70 ppb would be expected to
more effectively maintain long-term Os concentrations below those where the study by Jerrett et
al. (2009) indicates the most confidence in the reported association with respiratory mortality.
Specifically, air quality analyses indicate this to be the case in 9 out of the 12 urban case study
areas for a level of 70 ppb, compared to 6 out of 12 areas for the current standard. Finally, a
revised standard with a level of 70 ppb would be expected to reduce long-term Os concentrations
based on the types of metrics that have been reported in recent epidemiologic studies to be
associated with respiratory morbidity (i.e., seasonal averages of daily maximum 8-hour
concentrations).
In further considering the potential implications of epidemiology studies for alternative
standard levels, we note estimates of total mortality associated with short-term Os
concentrations.45 As discussed above, we consider estimates of total risk (i.e., based on the full
distributions of ambient Os concentrations) and estimates of risk associated with Os
concentrations in the upper portions of ambient distributions. With regard to total risk we note
that, when summed across urban case study areas, a standard with a level of 70 ppb is estimated
to reduce the number of deaths associated with short-term Os concentrations by about 4% (2007)
and 2% (2009), compared to the current standard.46 Based on a national modeling analysis, the
majority of the U.S. population would be expected to experience reductions in such risks upon
reducing precursor emissions.
44 In addition, for the other multicity studies identified in Table 4-1 (Cakmak et al., 2006; Stieb et al., 2009;
Katsouyanni et al., 2009), and for the study by Bell et al. (2006) (for the 30 ppb cut point) (Table 4-2), the majority
of study locations would likely have met a standard with a level of 70 ppb.
45 As discussed above, compared to the weight given to the evidence and to HREA estimates of exposures of
concern and lung function risks, we place relatively less weight on epidemiologic-based risk estimates.
46 A standard with a level of 70 ppb is also estimated to reduce respiratory mortality associated with long-term O3
concentrations in urban case study areas. However, given uncertainties associated with these risk estimates, as
discussed above, we give them limited weight.
4-62
-------
Compared to the total risk estimates noted above, an Os standard with a level of 70 ppb is
estimated to be more effective at reducing the number of deaths associated with short-term Os
concentrations at the upper ends of ambient distributions. Specifically, for area-wide Os
concentrations at or above 40 ppb, a standard with a level of 70 ppb is estimated to reduce the
number of deaths associated with short-term Os concentrations by about 10% compared to the
current standard. In addition, for area-wide concentrations at or above 60 ppb, a standard with a
level of 70 ppb is estimated to reduce Os-associated deaths by about 50% to 70% (Figure 4-13).
As discussed above, CASAC concluded that there is adequate scientific evidence to
consider a range of levels for a primary standard that includes an upper end at 70 ppb. However,
CASAC differentiated its advice from the conclusions in the second draft PA by also advising
that a level of 70 ppb would provide little margin of safety for protection of public health,
particularly for sensitive subpopulations (Frey, 2014, p. 8). In particular, CASAC stated that:
At 70 ppb, there is substantial scientific certainty of a variety of adverse effects,
including decrease in lung function, increase in respiratory symptoms, and
increase in airway inflammation. Although a level of 70 ppb is more protective of
public health than the current standard, it may not meet the statutory requirement
to protect public health with an adequate margin of safety (Frey, 2014, p.8).
However, the committee also acknowledged that "the choice of a level within the range
recommended based on scientific evidence [i.e., 70 to 60 ppb] is a policy judgment under the
statutory mandate of the Clean Air Act" (Frey, 2014).
In summary, compared to the current standard, a revised Os standard with a level of 70
ppb would be expected to (1) reduce the occurrence of exposures of concern to Os concentrations
that result in respiratory effects in healthy adults (at or above 60 and 70 ppb) by about 45 to
95%, almost eliminating the occurrence of multiple exposures at or above 70 ppb; (2) reduce the
occurrence of moderate-to-large Os-induced lung function decrements (FEVi decrements > 10,
15, 20%) by about 15 to 35%, most effectively limiting the occurrence of multiple decrements
and decrements > 15, 20%; (3) more effectively maintain short- and long-term Os concentrations
below those present in the epidemiologic studies that reported significant Os health effect
associations in locations likely to have met the current standard;47 and (4) reduce the risk of Os-
associated mortality and morbidity, particularly the risk associated with the upper portions of the
distributions of ambient Os concentrations.
47 Though epidemiologic studies also provide evidence for O3 health effect associations in locations likely to have
met a standard with a level of 70 ppb, as discussed below for lower standard levels.
4-63
-------
Os standard level of 65 ppb
Next, we consider a standard with a level of 65 ppb. A level of 65 ppb is well below 80
ppb, an Os exposure concentration that has been reported to elicit a range of respiratory effects
that includes airway hyperresponsiveness and decreased lung host defense, in addition to lung
function decrements, airway inflammation, and respiratory symptoms. A standard level of 65
ppb is also below the lowest exposure concentration at which the combined occurrence of
respiratory symptoms and lung function decrements has been reported (i.e., 72 ppb), a
combination judged adverse by the ATS (section 3.1.3). A level of 65 ppb is above the lowest
exposure concentration demonstrated to result in lung function decrements large enough to be
judged an abnormal response by ATS, where statistically significant changes in group mean
responses would be judged to be adverse by ATS, and which the CASAC has indicted could be
adverse in people with lung disease (i.e., 60 ppb). A level of 65 ppb is also above the lowest
exposure concentration at which pulmonary inflammation has been reported in healthy adults
(i.e., 60 ppb).
Compared to the current standard and a revised standard with a level of 70 ppb, the
HREA estimates that a standard with a level of 65 ppb would reduce exposures of concern to the
range of Os benchmark concentrations analyzed (i.e., 60, 70, and 80 ppb). The HREA estimates
that meeting a standard with a level of 65 ppb would eliminate exposures of concern at or above
80 ppb in the urban case study areas. Such a standard is estimated to allow far less than 1% of
children in the urban case study areas to experience one or more exposures of concern at or
above the 70 ppb benchmark level, even in the worst-case years and locations, and is estimated
to eliminate the occurrence of two or more exposures at or above 70 ppb (Table 4-4).
In addition, on average over the years 2006 to 2010, a standard with a level of 65 ppb is
estimated to allow between 0 and about 4% of children (including asthmatic children) in urban
case study areas to experience exposures of concern at or above 60 ppb, which CASAC has
indicated is an appropriate exposure of concern for people with asthma, including children. This
reflects an 80% reduction (on average across areas), relative to the current standard. A standard
with a level of 65 ppb is estimated to allow less than 1% of children to experience two or more
exposures of concern at or above 60 ppb (> 90% reduction, compared to current standard). In the
worst-case location and year, about 10% of children are estimated to experience one or more
exposures of concern at or above 60 ppb, with about 3% estimated to experience two or more
such exposures (Table 4-4).
Compared to the current standard and a revised standard with a level of 70 ppb, the
HREA estimates that a standard with a level of 65 ppb would also reduce the occurrence of Os-
induced lung function decrements. A level of 65 ppb is estimated to allow about 4% or less of
children to experience moderate Os-induced FEVi decrements > 15% (50% reduction, compared
4-64
-------
to current standard), even considering the worst-case location and year. Such a standard is
estimated to allow about 2% or less of children to experience two or more such decrements. A
standard set at 65 ppb is estimated to allow about 1% or less of children to experience large Os-
induced lung function decrements (i.e., FEVi decrement > 20%), even in the worst-case year and
location (Table 4-5).
On average over the years 2006 to 2010, a standard with a level of 65 ppb is estimated to
allow about 3 to 15% of children to experience one or more moderate Os-induced lung function
decrements (i.e., FEVi decrement > 10%), which CASAC has indicated could be adverse for
people with lung disease. This reflects an average reduction of about 30%, relative to the current
standard. A standard with a level of 65 ppb is also estimated to allow about 1 to 9% of children
in the urban case study areas to experience two or more such decrements (37% reduction,
compared to current standard). In the worst-case location and year, a standard set at 65 ppb is
estimated to allow up to about 18% of these children to experience one or more moderate Os-
induced lung function decrements > 10%, and up to 11% to experience two or more such
decrements (Table 4-5).
With regard to Os epidemiologic studies we note that, compared to a standard with a level
of 70 ppb, a revised standard with a level of 65 ppb would more effectively maintain short-term
Os concentrations below those present in the epidemiologic studies that reported significant Os
health effect associations in locations likely to have met the current standard. In particular,
Katsouyanni et al. (2009) reported statistically significant associations with mortality based on
air quality in 12 Canadian cities, most of which would likely have met a standard with a level of
70 ppb over the entire study period but violated a revised standard with a level of 65 ppb or
below over at least part of that period (Table 4-1). This analysis suggests that while the current
standard or a standard with a level of 70 ppb would allow the ambient Os concentrations in most
of the study locations that provided the basis for the association with mortality in this study, a
revised Os standard with a level at or below 65 ppb would require reductions in those ambient Os
concentrations. As discussed above for a level of 70 ppb, this analysis does not provide
information on the extent to which Os-associated mortality would persist upon meeting an Os
standard with a level of 65 ppb, or on the extent to which standard levels below 65 ppb could
further reduce the incidence of this mortality.48
With regard to long-term Os concentrations, as for 70 ppb (above) we evaluate the "long-
term" Os metrics reported to be associated with mortality or morbidity in recent epidemiologic
48 For the other multicity studies identified in Table 4-1 (Cakmak et al., 2006; Stieb et al., 2009; Katsouyanni et al.,
2009 (for hospital admissions)), and for the study by Bell et al. (2006) (for the 30 ppb cut point) (Table 4-2), the
majority of study locations would have met a standard with a level of 65 ppb.
4-65
-------
studies (e.g., seasonal averages of 1-hour or 8-hour daily max concentrations). Compared to the
current standard or a revised Os standard with a level of 70 ppb, a revised standard with a level
of 65 ppb would be expected to further reduce the risk of respiratory mortality associated with
long-term Os concentrations, based on information from the study by Jerrett et al. (2009).49 In
addition, a standard with a level of 65 ppb would be expected to more effectively maintain long-
term Os concentrations below those where the study by Jerrett et al. (2009) indicates the most
confidence in the reported association with respiratory mortality. Specifically, air quality
analyses indicate this to be the case in 10 out of the 12 urban case study areas for a level of 65
ppb, compared to 6 out of 12 areas for the current standard and 9 out of 12 for a standard with a
level of 70 ppb (Table 4-3). Finally, a revised standard with a level of 65 ppb would be expected
to further reduce long-term Os concentrations based on the types of metrics that have been
reported in recent epidemiologic studies to be associated with respiratory morbidity (i.e.,
seasonal averages of daily maximum 8-hour concentrations).
In further considering the potential implications of epidemiology studies for alternative
standard levels, we note estimates of total mortality associated with short-term Os.50 As
discussed above, we consider estimates of total risk (i.e., based on the full distributions of
ambient Os concentrations) and estimates of risk associated with Os concentrations in the upper
portions of ambient distributions. With regard to total risk we note that, when summed across
urban case study areas, a standard with a level of 65 ppb is estimated to reduce the number of
deaths associated with short-term Os exposures by about 13% (2007) and 9% (2009), compared
to the current standard.51 For area-wide concentrations at or above 40 ppb, a standard level of 65
ppb is estimated to reduce Os-associated deaths by almost 50% compared to the current standard,
when summed across cities. For area-wide concentrations at or above 60 ppb, a standard level of
65 ppb is estimated to reduce Os-associated deaths by more than 80% (Figure 4-13).
As discussed above, although CASAC concluded that the scientific evidence supports
considering standard levels as high as 70 ppb, it also concluded that a level of 70 ppb would
provide little margin of safety (Frey, 2014, p. 8). In support of its policy advice that the level
should be set below 70 ppb, CASAC noted that an alternative standard with a level of 65 ppb
would further reduce, though not eliminate, the frequency of lung function decrements > 15%
49 Though as discussed above, we note the lower confidence we place in these risk results (U.S. EPA, 2014a, section
9.6).
50 As discussed above, compared to the weight given to the evidence and to HREA estimates of exposures of
concern and lung function risks, we place relatively less weight on epidemiologic-based risk estimates.
51 A standard with a level of 65 ppb is also estimated to reduce respiratory mortality associated with long-term O3
concentrations in urban case study areas. However, given uncertainties associated with these risk estimates, as
discussed above, we give them limited weight.
4-66
-------
and would lead to lower frequency of short-term premature mortality (i.e., compared to a
standard with a level of 70 ppb) (Frey, 2014, p. 8).
In summary, compared to a standard with a level of 70 ppb, a revised standard with a
level of 65 ppb would be expected to (1) further reduce the occurrence of exposures of concern
(by about 80 to 100% compared to the current standard), decreasing exposures at or above 60
ppb and almost eliminating exposures at or above 70 and 80 ppb; (2) further reduce the
occurrence of FEVi decrements > 10, 15, and 20% (by about 30 to 65%, compared to the current
standard); (3) more effectively maintain short- and long-term Os concentrations below those
present in the epidemiologic studies that reported significant Os health effect associations in
locations likely to have met the current standard;52 and (4) further reduce the risk of Os-
associated mortality and morbidity, particularly the risk associated with the upper portion of the
distribution of ambient Os concentrations.
Os standard level of 60 ppb
We next consider a standard with a level of 60 ppb. A level of 60 ppb is well below the
Os exposure concentration that has been reported to elicit a wide range of potentially adverse
respiratory effects in healthy adults (i.e., 80 ppb). A level of 60 ppb is also below the
concentration where the combined occurrence of respiratory symptoms and lung function
decrements was observed, a combination judged adverse by the ATS (i.e., 72 ppb, discussed in
section 3.1.3). A level of 60 ppb corresponds to the lowest exposure concentration demonstrated
to result in lung function decrements that are large enough to be judged an abnormal response by
ATS, that meet ATS criteria for adversity based on a downward shift in the distribution of FEVI,
and that the CASAC indicated could be adverse in people with lung disease. A level of 60 ppb
also corresponds to the lowest exposure concentration at which pulmonary inflammation has
been reported in controlled human exposure studies.
Based on the FtREA analyses of Os exposures of concern, a standard with a level of 60
ppb is estimated to eliminate exposures of concern at or above the 70 and 80 ppb benchmark
concentrations and to be more effective than the higher standard levels at limiting exposures of
concern at or above 60 ppb. On average over the years 2006 to 2010, a standard with a level of
60 ppb is estimated to allow between 0 and about 1% of children, including asthmatic children,
in urban case study areas to experience exposures of concern at or above 60 ppb, which CASAC
indicated is an appropriate exposure of concern for asthmatic children. This reflects a 96%
reduction (on average across areas), compared to the current standard. A standard with a level of
52Though epidemiologic studies also provide evidence for O3 health effect associations in locations likely to have
met a standard with a level of 65 ppb, as discussed below for a level of 60 ppb.
4-67
-------
60 ppb is estimated to allow virtually no children to experience two or more exposures of
concern at or above 60 ppb. In the worst-case location and year, about 2% of children are
estimated to experience exposures of concern at or above 60 ppb, with far less than 1% estimated
to experience two or more such exposures (Table 4-4).
Based on the HREA analyses of Os-induced lung function decrements, a standard with a
level of 60 ppb would be expected to be more effective than a level of 70 or 65 ppb at limiting
the occurrence of Os-induced lung function decrements. A standard with a level of 60 ppb is
estimated to allow about 2% or less of children in the urban case study areas to experience one or
more moderate Os-induced FEVi decrements > 15% (almost 70% reduction, compared to current
standard), and about 1% or less to experience two or more such decrements (3% in the location
and year with the largest estimates). A standard set at 60 ppb is estimated to allow about 1% or
less of children to experience large Os-induced lung function decrements (i.e., FEVi decrement >
20%), even in the worst-case locations and year (Table 4-5).
On average over the years 2006 to 2010, a standard with a level of 60 ppb is estimated to
allow about 5 to 11% of children in the urban case study areas to experience one or more
moderate Os-induced lung function decrements that CASAC indicated could be adverse for
people with lung disease (i.e., FEVi decrements > 10%). This reflects an average reduction of
about 45%, compared to current standard. A standard with a level of 60 ppb is also estimated to
allow about 2 to 6% of children in these areas to experience two or more such decrements (51%
reduction, compared to current standard). In the worst-case location and year, a standard set at 60
ppb is estimated to allow up to about 13% of children to experience one or more moderate Os-
induced FEVi decrements > 10%, and 7% to experience two or more such decrements (Table 4-
5).
With regard to Os epidemiologic studies we note that, compared to a standard with a level
of 70 or 65 ppb, a revised standard with a level of 60 ppb would more effectively maintain short-
term Os concentrations below those present in the epidemiologic studies that reported significant
Os health effect associations in locations likely to have met the current standard. Specifically, in
all of the U.S. and Canadian epidemiologic studies evaluated, the majority of study cities had
ambient Os concentrations that would likely have violated a standard with a level of 60 ppb.
Thus, none of the U.S. and Canadian epidemiologic studies analyzed provide evidence for Os
health effect associations when the majority of study locations would likely have met a standard
with a level of 60 ppb (Tables 4-1 and 4-2). As discussed above, while this analysis does not
provide information on the extent to which the Cb-associated morbidity or mortality would
persist upon meeting an Os standard with a level of 60 ppb, it suggests that a revised Os standard
with a level of 60 ppb would require reductions in the ambient Os concentrations that provided
the basis for those health effect associations.
4-68
-------
With regard to long-term Os concentrations, compared to the current standard or a revised
Os standard with a level of 70 or 65 ppb, a revised standard with a level of 60 ppb would be
expected to further reduce the risk of respiratory mortality associated with long-term Os
concentrations, based on information from the study by Jerrett et al. (2009).53 In addition, a
standard with a level of 60 ppb would be expected to more effectively maintain long-term Os
concentrations below those where the study by Jerrett et al. (2009) indicates the most confidence
in the reported association with respiratory mortality. Specifically, air quality analyses indicate
this to be the case in all of the urban case study areas evaluated at a level of 60 ppb, compared to
6 out of 12 areas for the current standard, 9 out of 12 for a standard with a level of 70 ppb, and
10 out of 12 for a standard with a level of 65 ppb (Table 4-3). Finally, a revised standard with a
level of 60 ppb would be expected to further reduce long-term Os concentrations based on the
types of metrics that have been reported in recent epidemiologic studies to be associated with
respiratory morbidity (i.e., seasonal averages of daily maximum 8-hour concentrations).
In further considering the potential implications of epidemiology studies for alternative
standard levels, we note estimates of total mortality associated with short-term Os
concentrations.54 As discussed above, we consider estimates of total risk (i.e., based on the full
distributions of ambient Os concentrations) and estimates of risk associated with Os
concentrations in the upper portions of ambient distributions. With regard to total risk we note
that, when summed across urban case study areas, a standard with a level of 60 ppb is estimated
to reduce the number of deaths associated with short-term Os exposures by about 15% (2007)
and 11% (2009), compared to the current standard (Figure 4-13).55 For area-wide concentrations
at or above 40 ppb, a standard with a level set at 60 ppb is estimated to reduce Os-associated
deaths by almost 60% compared to the current standard. For area-wide concentrations at or
above 60 ppb, a standard level of 60 ppb is estimated to reduce Os-associated deaths by over
95% compared to the current standard (Figure 4-13).
Relative to the current standard, or alternative Os standards with levels of 70 or 65 ppb,
CASAC stated the following:
The frequency of lung function decrements and premature mortality from short-
term exposure to ozone decreases even further when the alternative standard is
lowered to 60 ppb (Frey, 2014, p. 8).
53 Though as discussed above, we note the lower confidence we place in these risk results (U.S. EPA, 2014a, section
9.6).
54 As discussed above, compared to the weight given to the evidence and to HREA estimates of exposures of
concern and lung function risks, we place relatively less weight on epidemiologic-based risk estimates.
55 A standard with a level of 60 ppb is also estimated to reduce respiratory mortality associated with long-term O3
concentrations in urban case study areas. However, given uncertainties associated with these risk estimates, as
discussed above, we give them limited weight.
4-69
-------
CASAC also concluded that "the recommended lower bound of 60 ppb would certainly offer
more public health protection than levels of 70 ppb or 65 ppb and would provide an adequate
margin of safety" (Frey, 2014, p. ii).
In summary, compared to a standard with a level of 70 or 65 ppb, a revised standard with
a level of 60 ppb would be expected to (1) further reduce the occurrence of exposures of concern
(by about 95 to 100% compared to the current standard), almost eliminating exposures at or
above 60 ppb; (2) further reduce the occurrence of FEVi decrements > 10, 15, and 20%, (by
about 45 to 85% compared to the current standard); (3) more effectively maintain short- and
long-term Os concentrations below those present in the epidemiologic studies that reported
significant Os health effect associations in locations likely to have met the current standard;56 and
(4) further reduce the risk of Os-associated mortality and morbidity, particularly the risk
associated with the upper portion of the distribution of ambient Os concentrations.
4.7 KEY UNCERTAINTIES AND AREAS FOR FUTURE RESEARCH AND
DATA COLLECTION
It is important to highlight the uncertainties associated with establishing standards for Os
during and after completion of the NAAQS review process. Research needs go beyond what is
necessary to understand health effects, population exposures, and risks of exposure for purposes
of setting standards. Research can also support the development of more efficient and effective
control strategies. 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 body of scientific evidence.
As has been presented and discussed in the ISA, particularly chapters 4 through 7, the
scientific body of evidence informing our understanding of health effects associated with long-
and short-term exposures to Os has been broadened and strengthened since the Os NAAQS
review completed in 2008. Still, we have concluded that Os health research needs and priorities
have not changed substantially since the 2007 Os Staff Paper (EPA 2007). Key uncertainties and
research needs that continue to be high priority for future reviews of the health-based standards
are identified below:
(1) An important aspect of risk characterization and decision making for air quality
standard levels for the Os NAAQS is the characterization of the shape of exposure-response
functions for Os, including the identification of potential population threshold levels. Recent
controlled human exposure studies of measurable lung function effects provide evidence for a
smooth dose-response curve without evidence of a threshold for exposures between 40 and 120
56As discussed above, these studies do not provide information on the extent to which O3 health effect associations
would persist following reductions in ambient Os concentrations in order to meet a standard with a level of 60 ppb.
4-70
-------
ppb Os (US EPA, 2013, Figure 6-1). Considering the importance of estimating health risks in the
range below 80 ppb Os, additional research is needed to evaluate responses in healthy and
especially people with asthma in the range of 40 to 70 ppb for 6-8 hour exposures while engaged
in moderate exertion.
(2) Similarly, for health endpoints reported in epidemiologic studies such as hospital
admissions, ED visits, and premature mortality, an important aspect of characterizing risk is the
shape of concentration-response functions for Os, including identification of potential population
threshold levels. Most of the recent studies and analyses continue to show no evidence for a clear
threshold in the relationships between Os concentrations commonly observed in the U.S. during
the Os season and these health endpoints, though evidence indicates less certainty in the shape of
the concentration-response curve at the lower end of the distribution of Os concentrations.
However, there continues to be heterogeneity in the Os-mortality relationship across cities (or
regions), including effect modifiers that are also expected to vary regionally, which are sources
of uncertainty. Additionally, whether or not exposure errors, misclassification of exposure, or
potential impacts of other copollutants may be obscuring potential population thresholds is still
unknown.
(3) The extent to which the broad mix of photochemical oxidants and more generally
other copollutants in the ambient air (e.g., PM, NCh, SCh, etc.) may play a role in modifying or
contributing to the observed associations between ambient Os and various morbidity effects and
mortality continues to be an important research question. Ozone has long been known as an
indicator of health effects of the entire photochemical oxidant mix in the ambient air and has
served as a surrogate for control purposes. A better understanding of sources of the broader
pollutant mix, of human exposures, and of how other pollutants may modify or contribute to the
health effects of Os in the ambient air, and vice versa, is needed to better inform future NAAQS
reviews.
(4) As epidemiologic research has continued to be an important factor in assessing the
public health impacts of Os, methodological issues in epidemiologic studies have received
greater visibility and scrutiny. There remains a need to further examine alternative modeling
specifications and control of time-varying factors, and to better understand the role of
copollutants in the ambient air. Additionally, there remains uncertainty around the role of
temperature as a potential confounder or effect modifier in epidemiologic models.
(5) Recent animal toxicological evidence, combined with limited evidence from
controlled human exposure studies of cardiovascular morbidity and epidemiologic studies of
cardiovascular mortality, have provided evidence of both direct and indirect effects on the
cardiovascular system. However, additional work will need to examine biologically plausible
mechanisms of cardiovascular effects, expand upon preliminary evidence from controlled human
4-71
-------
exposure studies, address inconsistencies observed in epidemiologic studies of cardiovascular
morbidity, and determine the extent to which Os is directly implicated or works together with
other pollutants in causing adverse cardiovascular effects in both at-risk and the general
populations.
(6) Most epidemiologic studies of short-term exposure effects have employed time-series
or case-crossover study designs and have been conducted in large populations. These study
designs remain subject to uncertainty due to use of ambient fixed-site data serving as a surrogate
for ambient exposures, and to the difficulty of determining the impact of any single pollutant
among the mix of pollutants in the ambient air. 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. Also,
additional research is needed to improve the characterization of the degree to which discrepancy
between stationary monitor measurements and actual pollutant exposures introduces error into
statistical estimates of pollutant effects in epidemiologic studies.
(7) Recent studies of "long-term" Os often evaluate associations with daily maximum
concentrations, averaged over the Os season. Research is needed to better understand the extent
to which health effects associated with such long-term metrics are attributable to long-term
average concentrations versus the repeated occurrence of daily maximum concentrations.
(8) Improved understanding of human exposures to ambient Os and to related
copollutants is an important research need. Population-based information on human exposure for
healthy adults and children and at-risk populations, including people with asthma, to ambient Os
concentrations, including exposure information in various microenvironments, is needed to better
evaluate current and future Os exposure models. Such information is needed for sufficient
periods to facilitate evaluation of exposure models throughout the Os season.
(9) Information is needed to improve inputs to current and future population-based Os
exposure and health risk assessment models. Collection of time-activity data over longer time
periods is needed to reduce uncertainty in the modeled exposure distributions that form an
important part of the basis for decisions regarding NAAQS for Os 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.
(10) An important consideration in the Os NAAQS review is the characterization of
background levels. There still remain substantial uncertainties in the characterization of 8-hour
daily max Os background concentrations. Further research to improve the evaluation of the
global and regional models which have been used to characterize estimates of background levels
4-72
-------
would improve understanding of the role of non-U.S. anthropogenic emissions on Os levels over
the U.S.
4.8 SUMMARY OF STAFF CONCLUSIONS ON PRIMARY STANDARD
In this section, we summarize our conclusions regarding the primary Os standard. These
conclusions are informed by our consideration of the available scientific evidence as assessed in
the ISA, air quality/exposure/risk information assessed in the final HREA, recommendations and
advice received from CAS AC, and comments received from members of the public.
As an initial matter in this PA, staff concludes that reducing ambient Os concentrations to
meet the current standard will provide important improvements in public health protection. This
initial conclusion is based on (1) the strong body of scientific evidence indicating a wide range of
adverse health outcomes attributable to exposures to Cb concentrations found in the ambient air
and (2) estimates indicating decreased Os exposures and health risks upon meeting the current
standard, compared to recent air quality. Strong support for this conclusion is provided by the
available health evidence, and by HREA estimates of exposures to Os concentrations shown to
result in respiratory effects in healthy adults (exposures of concern > 60, 70 and 80 ppb); Os-
induced lung function risks (FEVi decrements > 10, 15 and 20%); and Os-associated mortality
and morbidity risks.
Staff further concludes that the Os-attributable health effects estimated to be allowed by
air quality that meets the current primary standard can reasonably be judged important from a
public health perspective. This conclusion is based on consideration of the scientific evidence
assessed in the ISA, including controlled human exposure studies reporting abnormal or adverse
respiratory effects following exposures to Os concentrations below the level of the current
standard and epidemiologic studies indicating associations with morbidity and mortality for air
quality that would meet the current standard. This conclusion is also based on the HREA
estimates of exposures of concern, lung function risks, and morbidity and mortality risks; on
advice received from CAS AC in their review of draft versions of the PA; on CAS AC advice
received in previous reviews; and on consideration of public comments. Staff reaches the overall
conclusion that the available health evidence and exposure/risk information calls into question
the adequacy of the public health protection provided by the current standard.
Given this conclusion regarding the adequacy of the current standard, staff also reaches
conclusions for the Administrator's consideration regarding the elements of alternative primary
Os standards that could be supported by the available evidence and exposure/risk information. In
reaching conclusions about the range of potential alternative standards appropriate for
consideration, staff is mindful that the Act requires primary standards that, in the judgment of the
Administrator, are requisite to protect public health with an adequate margin of safety. The
4-73
-------
primary standards are to be neither more nor less stringent than necessary. Thus, the Act does not
require that primary NAAQS be set at zero-risk levels, but rather at levels that reduce risk
sufficiently to protect public health with an adequate margin of safety.
The degree of public health protection provided by any NAAQS results from the
collective impact of the elements of the standard, including the indicator, averaging time, level,
and form. Staffs conclusions on each of these elements are summarized below.
(1) Indicator: It is appropriate to continue to use Os as the indicator for a standard that is
intended to address effects associated with exposure to Os, alone or in combination with
related photochemical oxidants. Based on the available information, staff concludes that
there is no basis for considering any alternative indicator at this time. Meeting an Os
standard can be expected to provide some degree of protection against potential health
effects that may be independently associated with other photochemical oxidants, even
though such effects are not discernible from currently available studies indexed by Os
alone. Staff notes that control of ambient Os levels is generally understood to provide the
best means of controlling photochemical oxidants, and thus of protecting against effects
that may be associated with individual species and/or the broader mix of photochemical
oxidants.
(2) Averaging time: It is appropriate to consider continuing to use an 8-hour averaging time
for the primary Os standard.
(a) Staff concludes that an 8-hour averaging time remains appropriate for addressing
health effects associated with short-term exposures to ambient Os. An 8-hour
averaging time is similar to the exposure periods evaluated in controlled human
exposure studies, including recent studies reporting respiratory effects following
exposures to Os concentrations below the level of the current standard. In
addition, epidemiologic studies provide evidence for health effect associations
with 8-hour Os concentrations, as well as with 1-hour and 24-hour concentrations.
A standard with an 8-hour averaging time (combined with an appropriate standard
form and level) would also be expected to provide substantial protection against
health effects attributable to 1- and 24-hour exposures.
(b) Staff also concludes that a standard with an 8-hour averaging time can provide
protection against respiratory effects associated with longer term Os exposures.
Analyses in the HREA show that as air quality is adjusted to just meet the current
4-74
-------
or alternative 8-hour standards, most study areas are estimated to experience
reductions in respiratory mortality associated with long-term Os concentrations.
In addition, analyses in this PA indicate that just meeting an 8-hour standard with
an appropriate level would be expected to maintain long-term Os concentrations
(i.e., seasonal average of 1-hour daily max) below those where a key study
indicates the most confidence in the concentration-response relationship with
respiratory mortality. In considering other long-term Os metrics evaluated in
recent health studies, analyses in the HREA indicate that the large majority of the
U.S. population lives in locations where reducing NOx emissions would be
expected to decrease warm season averages of daily 8-hour ambient Os
concentrations, a long-term metric used in several recent studies reporting
associations with respiratory morbidity. Taken together, these analyses suggest
that a standard with an 8-hour averaging time, coupled with the current 4th-highest
form and an appropriate level, could provide appropriate protection against the
long-term Os concentrations reported to be associated with respiratory morbidity
and mortality.
(3) Form: For an 8-hour Os standard with a revised level (as discussed below), it is
appropriate to consider retaining the current form, defined as the 3-year average of the
annual 4th-highest daily maximum concentration. Staff notes that this form was selected
in 1997 and 2008 in recognition of the public health protection provided, when coupled
with an appropriate averaging time and level, combined with the stability provided for
implementation programs. The currently available evidence and exposure/risk
information does not call into question these conclusions from previous reviews.
(4) Level: The available scientific evidence and exposure/risk information provide strong
support for considering an Os standard with a revised level in order to increase public
health protection. Staff concludes that it is appropriate in this review to consider a revised
standard level within the range of 70 ppb to 60 ppb, reflecting the judgment that a
standard set within this range could provide an appropriate degree of public health
protection and would result in important improvements in protection for at-risk
populations and lifestages.
4-75
-------
4.9 REFERENCES
Adams, WC. (1998). Dose-response effect of varied equivalent minute ventilation rates on pulmonary function
responses during exposure to ozone. Washington, DC: American Petroleum Institute.
Adams, W. C. (2006) Comparison of chamber 6.6 hour exposures to 0.04-0.08 ppm ozone via square-wave and
triangular profiles on pulmonary responses. Inhalation Toxicol. 18: 127-136.
http://dx.doi.org/10.1080/08958370500306107
Bell, ML; Peng, RD; Dominici, F. (2006). The exposure-response curve for ozone and risk of mortality and the
adequacy of current ozone regulations. Environ Health Perspect 114: 532-536.
Brown, JS; Bateson, TF; 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: 1023-1026.
http://dx.doi.org/10.1289/ehp. 11396
Cakmak, S; Dales, RE; Judek, S. (2006). Respiratory health effects of air pollution gases: Modification by education
and income. Arch Environ Occup Health 61: 5-10. http://dx.doi.org/10.3200/AEOH.61.L5-10
Dales, RE; Cakmak, S; Doiron, MS. (2006). Gaseous air pollutants and hospitalization for respiratory disease in the
neonatal period. Environ Health Perspect 114: 1751-1754. http://dx.doi.org/10.1289/ehp.9044
Darrow, LA; Klein, M; Sarnat, JA; Mulholland, JA; Strickland, MJ; Sarnat, SE; Russell, AG; 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: 10-19.
http://dx.doi.org/10.1038/jes.2009.49
Frey, C. (2014) CAS AC Review of the EPA's Second Draft Policy Assessmentfor the Review of the Ozone National
Ambient Air Quality Standards EPA-CASAC-14-004. June 26, 2014. Available online at:
http://yosemite.epa.gov/sab/sabproduct.nsf/264cbl227d55e02c85257402007446a4/5EFA320CCAD326E8
85257D030071531C/$File/EPA-CASAC-14-004+unsigned.pdf
Frey, C.; Samet, J.M. (2012). 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). EPA-CASAC-13-003.
November 26, 2012. Available online at:
Islam, T; McConnell, R; Gauderman, WJ; Avol, E; Peters, JM; Gilliland, FD. (2008). Ozone, oxidant defense genes
and risk of asthma during adolescence. Am J Respir Crit Care Med 177: 388-395.
http://dx.doi.org/10.1164/rccm.200706-863OC
Jerrett, M; Burnett, RT; Pope, CA, III; Ito, K; Thurston, G; Krewski, D; Shi, Y; Calle, E; Thun, M. (2009). Long-
term ozone exposure and mortality. N Engl J Med 360: 1085-1095.
http://dx.doi.org/10.1056/NEJMoa0803894
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; Zanobetti, A. (2009). Air
pollution and health: A European and North American approach (APHENA). (Research Report 142).
Boston, MA: Health Effects Institute. http://pubs.healtheffects.org/view.php?id=327
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: 1215-1221.
http://dx.doi.org/10.1164/rccm.201011-1813OC
Lin, S; Liu, X; Le, LH; Hwang, SA. (2008b). Chronic exposure to ambient ozone and asthma hospital admissions
among children. Environ Health Perspect 116: 1725-1730. http://dx.doi.org/10.1289/ehp. 11184
4-76
-------
Mar, TF; Koenig, JQ. (2009). Relationship between visits to emergency departments for asthma and ozone exposure
in greater Seattle, Washington. Ann Allergy Asthma Immunol 103: 474-479.
McDonnell, WF; Stewart, PW; Smith, MV. (2010). Prediction of ozone-induced lung function responses in humans.
Inhal Toxicol 22: 160-168. http://dx.doi.org/10.3109/08958370903089557
Salam, MT; Islam, T; Gauderman, WJ; Gilliland, FD. (2009). Roles of arginase variants, atopy, and ozone in
childhood asthma. J Allergy Clin Immunol 123: 596-602. http://dx.doi.0rg/10.1016/j.jaci.2008.12.020
Samet, J.M. (2010) Review of EPA's Proposed Ozone National Ambient Air Quality Standard (Federal Register,
Vol. 75, Nov. 11, January 19, 2010). EPA-CASAC-10-007. February 19, 2010. Available online at:
http://yosemite.epa.gov/sab/sabproduct.nsf/264cbl227d55e02c85257402007446a4/610BB57CFAC8A
41C852576CF007076BD/$File/EP A-CASAC-10-007-unsigned.pdf
Samet, J.M. (2011) Clean Air Scientific Advisory Committee (CASAC) Response to Charge Questions on the
Reconsideration of the 2008 Ozone National Ambient Air Quality Standards. EPA-CASAC-11-004.
March 30, 2011. Available online at:
http://yosemite.epa.gov/sab/sabproduct.nsf/0/F08BEB48C1139E2A8525785E006909AC/$File/EPA-
CASAC-1 l-004-unsigned+.pdf
Sasser, E. (2014) Memo Responding to Request for Revised Ozone HREA Chapter 7 Appendix Tables. May 9,
2014. Available at http://www.epa.gov/ttn/naaqs/standards/ozone/s_o3_2008_rea.html
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: 265-272.
http://dx.doi.org/10.1164/rccm.200809-1484OC
Smith, RL; Xu, B; Switzer, P. (2009). Reassessing the relationship between ozone and short-term mortality in U.S.
urban communities. Inhal Toxicol 21: 37-61. http://dx.doi.org/10.1080/08958370903161612
Stieb, DM; Szyszkowicz, M; Rowe, BH; Leech, JA. (2009). Air pollution and emergency department visits for
cardiac and respiratory conditions: A multi-city time-series analysis. Environ Health Global Access Sci
Source 8: 25. http://dx.doi.org/10.1186/1476-069X-8-25
U.S. EPA (2007). Review of the National Ambient Air Quality Standards for Ozone: Policy Assessment of
Scientific and Technical Information, OAQPS Staff Paper. EPA-452/R-07-007
U.S. Environmental Protection Agency. (2012). Health Risk and Exposure Assessment for Ozone, First External
Review Draft, U.S. Environmental Protection Agency, Research Triangle Park, NC. EPA 452/P-12-001.
U.S. Environmental Protection Agency. (2013). Integrated Science Assessment for Ozone and Related
Photochemical Oxidants (Final Report). U.S. Environmental Protection Agency, Washington, DC,
EPA/600/R-10/076F.
U.S. Environmental Protection Agency. (2014). Health Risk and Exposure Assessment for Ozone. Office of Air
Quality Planning and Standards, Research Triangle Park, NC. EPA-452/R-14-004a. Available at
http://www.epa.gov/ttn/naaqs/standards/ozone/s_o3_index.html
4-77
-------
5 ADEQUACY OF THE CURRENT SECONDARY STANDARD
This chapter presents staffs considerations and conclusions regarding the adequacy of
the current secondary Os NAAQS. In doing so, we pose the following overarching question:
• Does the currently available scientific evidence- and exposure/risk-based
information, as reflected in the ISA and WREA, support or call into question the
adequacy and appropriateness of the protection afforded by the current
secondary Os standard?
In addressing this overarching question, we pose a series of more specific questions, as
discussed in sections 5.1 through 5.5 below. We consider the nature of Os-induced effects,
including the nature of the exposures that drive the biological and ecological response and
related biologically relevant exposure metrics (section 5.1); the scientific evidence and
exposure/risk information, including that for associated ecosystem services, regarding (a) tree
growth, productivity and carbon storage (section 5.2), (b) crop yield loss (section 5.3), (c) visible
foliar injury (section 5.4), and (d) other welfare effects (section 5.5). Section 5.6 describes
advice and recommendations received from CASAC. In section 5.7, we revisit the overarching
question of this chapter and present staff conclusions on the adequacy and appropriateness of the
current secondary standard.
5.1 NATURE OF EFFECTS AND BIOLOGICALLY RELEVANT EXPOSURE
METRIC
• Does the current evidence alter our conclusions from the previous review
regarding the nature of Os-induced welfare effects?
As discussed further below, the current body of Os welfare effects evidence confirms and
strengthens the conclusions reached in the last review on the nature of Os-induced welfare
effects. Ozone's phytotoxic effects were first identified on grape leaves in a study published in
1958 (Richards et al., 1958). In the more than fifty years that have followed, extensive research
has been conducted both in and outside of the U.S. to examine the impacts of Os on plants and
their associated ecosystems, since "of the phytotoxic compounds commonly found in the
ambient air, Os is the most prevalent, impairing crop production and injuring native vegetation
and ecosystems more than any other air pollutant" (U.S. EPA, 1989, 1996). Recent studies,
assessed in the ISA, together with this longstanding and well established vegetation effects
literature, further contribute to the coherence and consistency of the vegetation effects evidence.
In assessing the strength of the evidence, it is important to note that different types of
studies can provide different types of information, each with different associated uncertainties
5-1
-------
(U.S. EPA, 2013, Chapter 9, section 9.2). Controlled chamber studies are the best method for
isolating or characterizing the role of Os in inducing the observed plant effects, and in assessing
plant response to Os at the finer scales (U.S. EPA, 2013, Chapter 9, section 9.3). Recent
controlled studies have focused on a variety of plant responses to Os including: 1) the underlying
mechanisms as they relate to growth, productivity and carbon storage including: reduced carbon
dioxide uptake due to stomatal closure (U.S. EPA 2013, section 9.3.2.1); 2) the upregulation of
genes associated with plant defense, signaling, hormone synthesis and secondary metabolism
(U.S. EPA 2013, section 9.3.3.2); 3) the down regulation of genes related to photosynthesis and
general metabolism (U.S. EPA 2013, section 9.3.3.2); 4) the loss of carbon assimilation capacity
due to declines in the quantity and activity of key proteins and enzymes (U.S. EPA, 2013, section
9.3.5.1); and 5) the negative impacts on the efficiency of the photosynthetic light reactions (U.S.
EPA, 2013, section 9.3.5.1). As described in the ISA, these new studies "have increased
knowledge of the molecular, biochemical and cellular mechanisms occurring in plants in
response to Os", adding "to the understanding of the basic biology of how plants are affected by
oxidative stress..." (U.S. EPA, 2013, p. 9-11). The ISA further concluded that controlled studies
"have clearly shown that exposure to Os is causally linked to visible foliar injury, decreased
photosynthesis, changes in reproduction, and decreased growth" in many species of vegetation
(U.S. EPA 2013, p. 1-15).
Such effects at the plant scale can also be linked to an array of effects at larger spatial
scales. For example, recent field studies at larger spatial scales, together with previously
available evidence, support the controlled exposure study results and indicate that "ambient Os
exposures can affect ecosystem productivity, crop yield, water cycling, and ecosystem
community composition" (U.S. EPA, 2013, p. 1-15; Chapter 9, section 9.4).
The ISA summarizes the coherence across the full range of effects, from the least serious
to the most serious, as follows (U.S. EPA, 2013, p. 1-8):
The welfare effects ofOs can be observed across spatial scales, starting at the
subcellular and cellular level, then the whole plant and finally, ecosystem-level
processes. Ozone effects at small spatial scales, such as the leaf of an individual
plant, can result in effects along a continuum of larger spatial scales. These
effects include altered rates of leaf gas exchange, growth, and reproduction at the
individual plant level, and can result in broad changes in ecosystems, such as
productivity, carbon storage, water cycling, nutrient cycling, and community
composition.
Based on its assessment of this extensive body of science, the ISA determined that, with
respect to vegetation and ecosystems, a causal relationship exists between exposure to Os in
ambient air and visible foliar injury effects on vegetation, reduced vegetation growth, reduced
5-2
-------
productivity in terrestrial ecosystems, reduced yield and quality of agricultural crops and
alteration of below-ground biogeochemical cycles (U.S. EPA 2013, Table 1-2). Additionally, the
ISA determined that a likely to be causal relationship exists between exposures to Os in ambient
air and reduced carbon sequestration in terrestrial ecosystems, alteration of terrestrial ecosystem
water cycling and alteration of terrestrial community composition (U.S. EPA, 2013, Table 1-2).
With regard to the relationship between Os and radiative forcing and climate change, the ISA
determined that there is a causal relationship between changes in tropospheric Os concentrations
and radiative forcing, and likely to be a causal relationship between changes in tropospheric Os
concentrations and effects on climate (U.S. EPA, 2013, p. 1-13, and Table 1-3). From this set of
effects that the ISA has concluded to be causally or likely causally related to Os in ambient air,
we focus the discussion in the PA primarily on: 1) impacts on tree growth, productivity and
carbon storage; 2) crop yield loss; 3) visible foliar injury. Each of these discussions also
includes where appropriate, a discussion of any known or anticipated impacts that such
individual plant or species level effects could have at larger scales, including ecosystems, and on
associated ecosystem services.
In considering the available vegetation effects evidence, we make note of several
important contextual features that frame our understanding of the science and how it informs our
evaluation of the adequacy of the protection afforded by the current secondary NAAQS. First,
we acknowledge that under natural conditions, a variety of factors can either mitigate or
exacerbate the predicted Os-plant interactions and are recognized sources of uncertainty and
variability. These include: 1) multiple genetically influenced determinants of Os sensitivity; 2)
changing sensitivity to Os across vegetative growth stages; 3) co-occurring stressors and/or
modifying environmental factors (U.S. EPA, 2013, section 9.4.8).
Second, we acknowledge that the species that have been studied for Os sensitivity
represent only a fraction of the tens of thousands of plant species that grow in the U.S. (USD A
NRCS, 2014)1, and that these species were typically selected because of their commercial
importance (e.g., commodity crop or timber species) or because of observed Os-induced visible
foliar injury in the field. Of the species known to be sensitive to Os for foliar injury, 66 species
have been identified on National Park Service (NFS) and U.S. Fish and Wildlife Service lands2
and a subset of these are used in the USFS biomonitoring program (discussed in section 5.4
below). A number of these species have also been identified as important to tribal cultural
practices (see Appendix 5-A). Appendix 7J of the 2007 Staff Paper showed that no state in the
1 USDA NRCS. 2014. The PLANTS Database (http://plants.usda.gov. 3 January 2014). National Plant Data
Team, Greensboro, NC 27401-4901 USA.
2 See http://www2.nature.nps.gov/air/Pubs/pdf/flag/NPSozonesensppFLAG06.pdf
5-3
-------
lower 48 states had less than seven known Os-sensitive plant species, with the majority of states
having between 11 and 30 (see Appendix 7J-2 in U.S. EPA, 2007). We would not expect this
information to have changed since the previous review because there has been very little change
in the list of sensitive species and the occurrence of any of these plant species within a state
would not be expected to change. With respect to agricultural species, a number of important
commodity crops such as soybean and additional fruit and vegetable species such as lettuce have
been shown to be sensitive to Os for either foliar injury or yield loss (U.S. EPA, 2013, section
9.4.4.1; Abt Associates, Inc., 1995).
Third, we acknowledge that out of the group of species known to be sensitive to Cb, we
have chosen to focus primarily on species for which we have robust exposure-response (E-R)
functions for biomass loss and yield loss using the W126 form (i.e., 11 tree and 10 crop species)
in order to be able to quantitatively relate predicted changes in Os to predicted changes in plant
exposures, responses and associated risks.3 However, while we recognize that this small group
represents only a fraction of all species known or anticipated to be sensitive to Os in the U.S., we
also note, as did CASAC, that among the studied species, there is a fairly large range of Os
sensitivities represented, so that it could be reasonable to assume that other non-studied species
might have sensitivities that fall within or near this range. Specifically, CASAC states "[i]t
should not be assumed that species of unknown sensitivity are tolerant to ozone. It 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).
Fourth, we acknowledge that in addition to the well-studied effects of biomass loss in
trees and crops and visible foliar injury in bioindicator plants that we can quantify, numerous
other more subtle and less easily observed effects occur along the continuum of spatial scales
that lead to ecosystem effects. While these effects are more difficult to quantify, we
acknowledge that any secondary standard set to protect the public welfare against the known and
quantifiable adverse effects to vegetation should also consider the anticipated, but currently
unquantifiable, potential adverse effects on vegetation, ecosystems and associated services.
Finally, we further acknowledge that in light of the above, when considering the available
evidence, we seek to find the right balance between placing weight on the associated
uncertainties and limitations of the evidence and placing weight on its well-established strength,
coherence and consistency. In so doing, we note that CASAC, in commenting on section 6.7
which describes key uncertainties and future research areas, states that "[w]hile these scientific
research priorities will enhance future scientific reviews of the ozone primary and secondary
3 '
' There is an E-R function available for a 12th tree species (cottonwood), but this E-R function is considered
less robust because it is based on the results of a single gradient study (Gregg et al., 2003).
5-4
-------
standards, we also make clear that there is sufficient scientific evidence, and sufficient
confidence in the available research results, to support the advice we have given above for this
review cycle of the primary and secondary standards" (Frey, 2014, p. iv).
• Does the current evidence continue to support a cumulative, seasonal exposure
index as a biological-relevant and appropriate metric for assessment of the
evidence and exposure/risk information?
In this review, the ISA assessment of the full body of currently available evidence stated
the following regarding biological indices (U.S. EPA, 2013, p. 2-44):
The main conclusions from the 1996 and 2006 Os AQCDs [Air Quality Criteria
Documents] regarding indices based on ambient exposure remain valid. These
key conclusions can be restated as follows:
• ozone effects in plants are cumulative;
• higher Os concentrations appear to be more important than lower
concentrations in eliciting a response;
• plant sensitivity to Os varies with time of day and plant development
stage;
• quantifying exposure with indices that cumulate hourly Os concentrations
and preferentially weight the higher concentrations improves the
explanatory power of exposure/response models for growth and yield,
over using indices based on mean and peak exposure values.
The long-standing body of available evidence upon which these conclusions are based,
provides a wealth of information on aspects of Os exposure that are important in influencing
plant response. Specifically, a variety of "factors with known or suspected bearing on the
exposure-response relationship, including concentration, time of day, respite time, frequency of
peak occurrence, plant phenology, predisposition, etc.," have been identified (U.S. EPA, 2013,
section 9.5.2). In addition, the importance of the duration of the exposure and the relatively
greater importance of higher concentrations over lower in determining plant response to Os have
been consistently well documented (U.S. EPA, 2013, section 9.5.3). Much of this evidence was
assessed in the 1996 Criteria Document (CD) (U.S. EPA, 1996), while more recent work
substantiating this evidence is assessed in the subsequent 2006 CD and 2013 ISA.
Understanding of the biological basis for plant response to Os exposure led to the
development of a large number of "mathematical approaches for summarizing ambient air
quality information in biologically meaningful forms for Os vegetation effects assessment
purposes ..." (U.S. EPA, 2013, section 9.5.3), including those that cumulate exposures over
some specified period while weighting higher concentrations more than lower (U.S. EPA, 2013,
section 9.5.2). As with any summary statistic, these exposure indices retain information on
5-5
-------
some, but not all, characteristics of the original observations. As discussed in greater detail in
section 6.2 below, the 1996 CD contained an extensive review of the published literature on
different types of exposure-response metrics, including comparisons between metrics, from
which the 1996 Staff Paper built its assessment of forms appropriate to consider in the context of
the secondary NAAQS review. The result of these assessments was a decision by the EPA to
focus on cumulative, concentration-weighted indices, which were recognized as the most
appropriate biologically based metrics to consider in this context, with attention given primarily
to two cumulative, concentration weighted index forms: SUM06 and W126. The SUM06 index
is a threshold-based approach described as the sum of all hourly Os concentrations greater or
equal to 0.06 ppm observed during a specified daily and seasonal time window (U.S. EPA, 2013,
section 9.5.2). The W126 index is a non-threshold approach described as the sigmoidally
weighted sum of all hourly Cb concentrations observed during a specified daily and seasonal
time window, where each hourly Os concentration is given a weight that increases from 0 to 1
with increasing concentration (Lefohn et al., 1988; Lefohn and Runeckles, 1987; U.S. EPA,
2013, section 9.5.2).
In both the 1997 and 2008 reviews, the EPA concluded that the risk to vegetation comes
primarily from cumulative exposures to Os over a season or seasons4 and proposed, as one policy
alternative, a secondary standard set in terms of such a form: SUM06 (61 FR 65716) and W126
(72 FR 37818) in the 1997 and 2008 reviews, respectively. Although in both reviews the policy
decision was made to set the secondary standard to be identical to a revised primary standard
(with an 8-hour averaging time), the Administrator, in both cases, also concluded, consistent
with CASAC advice, that a cumulative, seasonal index was the most biologically relevant way to
relate exposure to plant growth response (62 FR 38856, 73 FR 16436). Similarly, in the 2010
proposed reconsideration of the 2008 decision, the EPA proposed to conclude that Os exposure
indices that cumulate differentially weighted hourly concentrations are the best candidates for
relating exposure to plant growth responses and proposed as the only policy option to set the
secondary standard in terms of one such form, the W126 (75 FR 2938). This approach of
establishing a secondary standard that was separate and distinct from the primary standard and in
particular using a cumulative seasonal exposure index such as W126 received strong support
from CASAC in both 2008 and 2010 reviews (Henderson, 2006, 2008; Samet, 2010), as it has
again in this review, as discussed in section 5.6 below.
An alternative to using ambient exposure durations and concentrations to predict plant
response has been developed in recent years, primarily in Europe, i.e., flux models. While
4 In describing the form as "seasonal", the EPA is referring generally to the growing season of O3-sensitive
vegetation, not to the seasons of the year (i.e., spring, summer, fall, winter).
5-6
-------
"some researchers have claimed that using flux models can be used to better predict vegetation
responses to Os than exposure-based approaches..." because flux models estimate the ambient
Os concentration that actually enters the leaf (i.e., flux or deposition) (U.S. EPA, 2013, p. 9-114),
it is important to note that "[f]lux calculations are data intensive and must be carefully
implemented" (U.S. EPA, 2013, p. 9-114). Further, "[t]his uptake-based approach to quantify
the vegetation impact of Os requires inclusion of those factors that control the diurnal and
seasonal Os flux to vegetation (e.g., climate patterns, species and/or vegetation-type factors and
site-specific factors)" (U.S. EPA, 2013, p. 9-114). In addition to these data requirements, each
species has different amounts of internal detoxification potential that may protect species to
differing degrees. This balance between Os flux and detoxification processes has been termed
the "effective flux". Accordingly, the "models have to distinguish between stomatal and non-
stomatal components of Os deposition to adequately estimate actual concentration reaching the
target tissue of a plant to elicit a response" and " ultimately the 'effective' flux" (U.S. EPA,
2013, pp. 9-114). 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 have continued to make this
technique less viable for use in vulnerability and risk assessments at the national scale in the U.S.
(U.S. EPA, 2013, section 9.5.4).
Therefore, consistent with the ISA conclusions regarding the appropriateness of
considering cumulative exposure indices that preferentially weight higher concentrations over
lower for predicting Os effects of concern based on the long-established conclusions and long-
standing supporting evidence described above, and in light of continued CASAC support, we
continue to focus on the aspects of ambient Os exposures that have biological relevance and the
biologically relevant exposure indices or metrics that have been designed in light of this
consideration, i.e., cumulative concentration-weighted indices. In addition, given the lack of any
information in the current review to the contrary, we therefore again conclude that the current
evidence, as in recent reviews, continues to support a cumulative, seasonal exposure index as a
biologically relevant and appropriate metric for assessment of the evidence and exposure/risk
information, and in particular, the W126 cumulative, seasonal metric (U.S. EPA, 2013, section
2.6.6.1, section 9.5.2). Such a metric, as stated above, has an "explanatory power" that is
improved "over using indices based on mean and peak exposure values" (U.S. EPA, 2013,
section 2.6.6.1, p. 2-44). Thus, as in the WREA, discussions of the effects evidence and
exposure/risk results in sections 5.2 through 5.5 of this PA are provided in terms of the W126
index, where available.
• What paradigm is being used to consider which of the known or anticipated
Os-induced effects have the potential to be adverse to the public welfare?
5-7
-------
The Clean Air Act (CAA), in section 109 (b) (2) requires that "[a]ny national secondary
ambient air quality standard... shall 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
such air pollutant in the ambient air." The "criteria" referred to in this text are defined earlier in
CAA section 108 (a) (2) which states in part that "[a]ir quality criteria for an air pollutant shall
accurately reflect the latest scientific knowledge useful in indicating the kind and extent of all
identifiable effects on public health or welfare which may be expected from the presence of such
pollutant in the ambient air, in varying quantities." Thus, while the criteria include "all
identifiable effects", Congress directed the EPA to establish the secondary NAAQS based on the
Administrator's judgement of what is requisite to protect against "adverse effects" in the context
of the public welfare. However, the CAA does not provide specific standards for determining
what constitutes an effect that is adverse to the public welfare leaving these determinations
instead to the "judgment of the Administrator". As stated above in section 1.1, the PA is intended
to help "bridge the gap" between the Agency's scientific assessments presented in the ISA and
REAs, which constitute "the criteria" and the judgments required of the EPA Administrator
regarding whether it is appropriate to retain or revise the NAAQS.5 In the context of the
secondary standard, the PA thus serves the function of translating the information assessed in
both the ISA and WREA into the public welfare policy context. In order to do this, the PA
applies a specific approach or paradigm (see also section 1.3.2. above), that guides staffs
consideration and interpretation of the available information and which then informs staff
conclusions regarding policy options that are appropriate for the Administrator to consider. The
following discussion describes the evolution of this paradigm throughout the last several reviews
and into the current review.
In the 1997 secondary Os NAAQS review, a policy-relevant distinction was made
between the terms "injury" and "damage". Specifically, Os-induced "injury" to vegetation was
defined as encompassing all plant reactions, including reversible changes or changes in plant
metabolism (e.g., altered photosynthetic rate), altered plant quality or reduced growth. In
contrast, "damage" was defined to include only those injury effects that reach sufficient
magnitude as to also reduce or impair the intended use or value of the plant, thus potentially
being adverse to the public welfare. In published scientific literature, on the other hand, the terms
"adverse", "injury" or "damage" continue to be used interchangeably. The early Os NAAQS
reviews focused primarily on Os-induced effects at the individual and species level. In such
5American Farm Bureau Federation v. EPA. 559 F. 3d 512, 521 (D.C. Cir. 2009); Natural Resources
Defense Council v. EPA. 902 F. 2d 962, 967-68, 970 (D.C. Cir. 1990).
5-8
-------
cases, examples of vegetation effects that were also classified as damage included reductions in
aesthetic values (e.g., visible foliar injury in ornamental species or occurring in valued natural
landscapes such as national parks) and tree growth/biomass and crop yield losses (i.e., in terms
of weight, number, quality, appearance, or size of harvestable crop or timber species). In the
context of evaluating effects on single plants or species grown in monocultures such as managed
forests, this construct continues to remain useful (73 FR 16492/96).
In subsequent reviews, however, the scientific literature linking Os effects on plants or
species to effects at the community or ecosystem level continued to increase. As a result, more
recent reviews have considered a more expansive construct or paradigm of what appropriately
constitutes Os "damage" to extend beyond that of the individual or species level. A number of
these broader paradigms have been discussed in the literature (72 FR 37890; Hogsett et al., 1997;
Young and Sanzone, 2002). Thus, in the 2008 review, the Administrator, while continuing to
express support for relying on a definition of "adverse" discussed in section IV.A.3 of the
proposal (72 FR 37889-37890) that embeds "the concept of 'intended use' of the ecological
receptors and resources that are affected", also supported applying "that concept beyond the
species level to the ecosystem level" (73 FR 16496). In so doing, 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).
Thus, this paradigm recognized that the significance to the public welfare of Os-induced effects
on sensitive vegetation growing within the U.S. can vary 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. Accordingly, any given Os-related effect on
vegetation and ecosystems (e.g., biomass loss, crop yield loss, visible foliar injury) may be
judged to have a different degree of impact on the public welfare depending, for example, on
whether that effect occurs in a Class I area, a city park, or commercial cropland. In the 2010
proposed reconsideration, the Administrator proposed to place the highest priority and
significance on vegetation and ecosystem effects to sensitive species that are known to or are
likely to occur in federally protected areas such as national parks and other Class I areas, or on
lands set aside by states, tribes and public interest groups to provide similar benefits to the public
welfare (75 FR 3023/24). Effects occurring in such areas would likely have the highest potential
for being classified as adverse to the public welfare, due to the expectation that these areas need
to be maintained in a more pristine condition to ensure their intended use is met. In contrast, in
that proposal, the Administrator considered it less clear the degree to which Os vegetation
impacts potentially predicted to occur in areas and on species that are already heavily managed
5-9
-------
to obtain a particular output (such as commodity crops or commercial timber production), would
impair the intended use at a level that would be judged adverse to the public welfare and also
noted that these species would likely receive some protection for a standard set to provide
protection in areas set aside to be maintained in a more pristine condition (75 FR 3024).
In the current review, we revisited the appropriateness of using this paradigm and
whether the available information supported any further evolution. In so doing, we noted the
ISA text, which states that "[o]n a broader scale, ecosystem services may provide indicators for
ecological impacts. Ecosystem services are the benefits that people obtain from ecosystems
(UNEP, 2003)" (U.S. EPA, 2013, Preamble, p. Ixxii) and the ISA list of a number of ecosystem
services that can be affected by Os-induced effects on plants and ecosystems, including
decreased productivity, decreased carbon sequestration, altered water cycling, and altered
community composition (U.S. EPA, 2013, Figure 2-2, pp. 2-36; Figure 9-1, p. 9-3). We further
noted that other recent EPA documents have already incorporated this concept. For example, the
recent review of the secondary NAAQS for oxides of nitrogen and sulfur recognized that
changes in ecosystem services may be used to aid in characterizing a known or anticipated
adverse effect to public welfare and that an evaluation of adversity to the public welfare might
consider the likelihood, type, magnitude, and spatial scale of the effect, as well as the potential
for recovery and any uncertainties relating to these conditions (77 FR 20232). Similarly, the
EPA document, Ecological Benefits Assessment Strategic Plan, includes a definition of
ecological goods and services used for the purposes of benefits assessment that EPA has relied
upon in regulatory impact analyses for previous rulemakings. This definition states that
ecological goods and services are the "outputs of ecological functions or processes that directly
or indirectly contribute to social welfare or have the potential to do so in the future"... and that
"[s]ome outputs may be bought and sold, but most are not marketed" (U.S. EPA, 2006b).
After considering this information, and given the accepted use of these concepts and their
clear applicability to the secondary NAAQS review, we concluded that while it is still
appropriate to apply the paradigm used in the 2010 reconsideration that takes into account the
variation in public welfare significance of Os-related vegetation effects when evaluating the
potential adversity of the currently available evidence, there is also sufficient support for an
expansion of this paradigm to explicitly include consideration of impacts to ecosystem goods and
services. Doing so can help clarify the relationship between predicted Os-induced vegetation
effects and anticipated impacts on public welfare benefits received from those impacted species
or ecosystems, and, as was done in the WREA, clarify how those services might be expected to
change under air quality scenarios representing the current and potential alternative secondary
standards (U.S. EPA, 2014a, chapter 5). The expansion of this paradigm to include ecosystem
5-10
-------
goods and services brings with it a number of additional considerations. Specifically, when
considering the potential public welfare benefits from these goods and services, it is important to
note that they can accrue across a range of dimensions, including spatial, temporal, and social,
and these likely will vary depending on the type of effect being characterized. For example,
ecosystems can cover a range of spatial scales, and the services they provide can accrue locally
or be distributed more broadly such as when crops are sold and eaten locally and/or also sold in
regional, national and world markets. Ecosystem services can likewise be realized over a range
of temporal scales from immediate up to long term (e.g., the removal of air pollutants that have a
short-term impact on human health but are also climate forcers with long atmospheric lifetimes,
such that their removal may have immediate as well as long-term benefits). The size of the
societal unit receiving benefits from ecosystem services can also vary dramatically. 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 (U.S. EPA, 2014a, chapter 5,
section 5.5.1).
We thus conclude that it is appropriate for the Administrator, in specifying what "level of
air quality" for a pollutant "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,"
to evaluate the scientific evidence regarding these effects in the context of the most recent
paradigm discussed above. This paradigm integrates the concepts of: 1) variability in public
welfare significance given intended use and value of the affected entity such as an individual
species; 2) relevance of associated ecosystem services to public welfare; 3) variability in spatial,
temporal, and social distribution of ecosystem services associated with known and anticipated
welfare effects. In so doing, we recognize that there is no bright-line rule delineating the set of
conditions or scales at which known or anticipated effects become adverse to public welfare.
Thus, the evidence and exposure/risk information discussed in this chapter will be further
evaluated in Chapter 6 using the concepts incorporated in this paradigm to help inform the
Administrator's judgments with respect to the adversity of the effects to the public welfare and
what is considered requisite protection.
5.2 FOREST TREE GROWTH, PRODUCTIVITY AND CARBON STORAGE
Trees merit consideration from a public welfare perspective because they provide many
services that people value, including aesthetic value (also discussed in section 5.4 below), food,
fiber, timber, other forest products, habitat, recreational opportunities, climate regulation, erosion
control, air pollution removal, hydrologic and fire regime stabilization (U.S. EPA, 2014a, section
5-11
-------
6.1, Figure 6-1, section 6.4, Table 6-13). One source identifies as many as 1,497 native tree
species growing in the lower 48 of the U.S.6 Ozone has been shown to be phytotoxic to a number
of important U.S. tree species with respect to growth, productivity, and carbon storage, including
for cumulative exposures that have occurred under recent U.S. air quality. This section includes
a discussion of the policy-relevant evidence and weight-of-evidence conclusions discussed in the
ISA (section 5.2.1) and the exposure/risk results, including both quantitative and qualitative
results for these effects, as well as associated ecosystem services (section 5.2.2) as described in
the final WREA (U.S. EPA, 2014a). Important uncertainties and limitations in the available
information are also discussed in each section. These discussions highlight the information we
consider relevant to answering the overarching question and associated policy-relevant questions
included in this section.
5.2.1 Evidence-based Considerations
• To what extent has scientific information become available that alters or
substantiates our prior conclusions of Os-related effects on forest tree growth,
productivity and carbon storage and of factors that influence associations between
Os concentrations and these effects?
Research published since the 2006 CD substantiates prior conclusions regarding Os-
related effects on forest tree growth, productivity and carbon storage. The ISA states that
"previous Os AQCDs concluded that there is strong evidence that exposures to Os decreases
photosynthesis and growth in numerous plant species" and that "[s]tudies published since the
2008 review support those conclusions" (U.S. EPA, 2013, p. 9-42). The recent studies that
support the previous conclusions come from a variety of different study types that cover an array
of different species, effects endpoints, levels of biological organization and exposure methods
and durations. As stated in Chapter 1, and above, the documentation of Os-induced species-
specific responses across multiple lines of evidence, and over the full range of levels of
biological organization highlights and strengthens the consistency and coherence of the evidence
available in this review.
The previously available strong evidence for trees includes robust exposure-response (E-
R) functions for seedling biomass loss in 11 species developed under the National Health and
Environmental Effects Research Laboratory-Western Ecology Division (NHEERL-WED)
program. This series of experiments used open-top-chambers (OTC) to study seedling growth
response for a single growing season under a variety of Os exposures (ranging from near
6 USD A, NRCS. 2014. The PLANTS Database (http://plants.usda.gov. 7 July 2014). National Plant Data
Team, Greensboro, NC 27401-4901 USA.
5-12
-------
background to well above current ambient concentrations) and growing conditions (U.S. EPA
2013, section 9.6.2, Lee and Hogsett, 1996). The evidence from these studies shows that there is
a wide range in sensitivity across the studied species in the seedling growth stage over the course
of a single growing season, with some species being extremely sensitive and others being very
insensitive, or alternatively quite tolerant, over the range of cumulative Os exposures studied
(See Figure 5-1, below).
In addition, field-based studies of species growing in natural stands have compared
observed plant response across a number of different sites and/or years when exposed to varying
ambient Os exposure conditions only. For example, a study conducted in forest stands in the
southern Appalachian Mountains found that the cumulative effects of ambient levels of Os
decreased seasonal stem growth (measured as a change in circumference) by 30-50% for most of
the examined tree species (i.e., tulip poplar, black cherry, red maple, sugar maple) in a high Os
year in comparison to a low Os year (McLaughlin et al., 2007a). The authors also reported that
high ambient Os concentrations can increase whole-tree water use and in turn reduce late-season
streamflow (McLaughlin et al., 2007b) (U.S. EPA, 2013, p. 9-43). This study used ambient Os
conditions found at several different sites to create the variation in Os exposures.
Because trees and other perennials are long lived, it is important to consider the potential
for impacts beyond a single year. Limited evidence in previous reviews reported that vegetation
effects from a single year of exposure to elevated Os could be observed in the following year.
For example, growth affected by a reduction in carbohydrate storage in one year may result in
the limitation of growth in the following year. Such "carry-over" effects have been documented
in the growth of some tree seedlings and in roots (U.S. EPA, 2013, section 9.4.8; Andersen, et
al., 1997). In the current review, additional field-based evidence expands our understanding of
the consequences of single and multi-year Os exposures in subsequent years. A number of
studies were conducted at a planted forest at the Aspen Free-Air Carbon Dioxide Enrichment
(FACE) site in Wisconsin. These studies, which occurred in a field setting more similar to
natural forest stands than OTC studies, observed tree growth responses when grown in single or
two species stands within 30-m diameter rings and exposed to ambient and above ambient
conditions over a period often years. Some researchers similarly recognized the potential for
carry-over effects when they observed that the effects of Os on birch seeds (reduced weight,
germination, and starch levels) could lead to a negative impact on species regeneration in
subsequent years, and that the effect of reduced aspen bud size might have been related to the
observed delay in spring leaf development. These effects suggest that elevated Os exposures
have the potential to alter carbon metabolism of overwintering buds, which may have subsequent
effects in the following year (Darbah, et al., 2008, 2007; Riikonen et al., 2008; U.S. EPA, 2013,
section 9.4.3). Other studies found that, in addition to affecting tree heights, diameters, and main
5-13
-------
stem volumes in the aspen community, elevated Os over a 7-year study period was reported to
increase the rate of conversion from a mixed aspen-birch community to a community dominated
by the more tolerant birch, leading the authors to conclude that elevated Os may alter intra- and
inter-species competition within a forest stand (Kubiske et al., 2006; Kubiske et al., 2007) (U.S.
EPA, 2013, section 9.4.3). These studies confirm earlier FACE results showing large decreases
in growth for aspen over a 6-7 year period when exposed to elevated Os (King et al., 2005) and
that yearly biomass loss cumulated over that timeframe.
In addition to individual studies, recent meta-analyses have quantified the effect of Os on
trees across large numbers of studies. In particular, a recent meta-analysis (Wittig, et al., 2007)
indicates a relationship between Os concentrations in the northern hemisphere and stomatal
conductance and photosynthesis, which decrease growth (U.S. EPA, 2013, section 9.4.3.1;
Wittig et al., 2007). 7 This analysis reported that recent Os concentrations in the northern
hemisphere are decreasing stomatal conductance (13%) and photosynthesis (11%) across tree
species. It also found that younger trees (<4 years) were affected less by Os than older trees
(Wittig, et al., 2007). A second meta-analysis, Wittig, et al. (2009), which quantitatively
compiled peer-reviewed studies from the past 40 years, found that ambient Os concentrations
reported in those studies significantly decreased annual total biomass growth (7%) across the 263
studies (U.S. EPA, 2013, section 9.4.3.1). The ISA states that this meta-analysis demonstrates
the coherence of Os effects across numerous studies and species that used a variety of
experimental techniques, and these results support the conclusion of the previous CD that
exposure to Os decreases plant growth. Other meta-analyses have examined the effect of Os
exposure on root growth and generally found that Os exposure reduced carbon allocated to roots.
For example, Grantz et al. (2006) found that Os exposure reduced the ratio between the relative
growth rate of the root and shoot by 5.6% (U.S. EPA, 2013, pp. 9-45 to 9-46).
In our consideration of the recent studies discussed above, in combination with the entire
body of available evidence, we note that the recent scientific literature further strengthens and
contributes to the consistency and coherence of the evidence base by substantiating and
expanding prior conclusions regarding Os-related effects on tree growth, productivity and carbon
storage, including mixed species forest stands and the ecosystems and services that derive from
them, as discussed more fully below. We also note that the ISA concludes that the currently
available evidence supports causal determinations regarding Os effects on tree growth and
productivity and the associated effects of altered carbon allocation to below ground tissues, rates
of leaf and root production, turnover and decomposition that can alter below-ground
7 Meta-analysis allows for the objective development of a quantitative consensus of the effects of a
treatment across a wide body of literature.
5-14
-------
biogeochemical cycles, as well as the likely to be a causal relationship with reduced carbon
sequestration and alteration of terrestrial community composition and water cycling (U.S. EPA,
2013, Table 2-2; 9-19). Finally, we note that except for the recent limited information on
cottonwood in the ISA (U.S. EPA, 2013, section 9.6.3.3), there has not been an expansion in the
number of tree species for which we have E-R functions, so only 12 species have available E-R
functions for use in quantitative exposure and risk analyses and for predicting tree seedling
response under a range of Os exposure conditions/scenarios. While these 12 species represent
only a small fraction (0.8%) of the total number of native tree species in the contiguous U.S.
(1,497), 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 Os (Figure
5-1 below, U.S. EPA 2013, section 9.6.2; U.S. EPA, 2014a, section 6.2, Figure 6-2, Table 6-1).
The CASAC states in their letter to the Administrator on the second draft PA, that while "[fjhere
is considerable uncertainty in extrapolating from the 12 forest tree species to all forest tree
species in the U.S...[i]t is scientifically justifiable to extrapolate from the known E-R curves,
assuming that they are representative of the un-sampled population" (Frey, 2014, p. 15).
As we further consider the results from the quantitative exposure and risk analyses,
described below and in the WREA (U.S. EPA, 2014a), that in addition to the quantifiable portion
of risks associated with the robust information on tree species, it is also reasonable to consider,
based on the long-standing evidence and recent CASAC advice, the anticipated risks to other tree
species that have not had their sensitivity to Os studied in a robust quantifiable way but that
potentially have Os sensitivities that fall within the range for known species (see U.S. EPA,
2007, Table 7J-1 in Appendix 7J and Table 7J-2).
• To what extent have important uncertainties in the evidence identified in the last
review been reduced and/or new uncertainties emerged?
As stated above, the ISA concludes that the new evidence confirms, strengthens and
expands our understanding of Os effects on plants. Much of this new evidence is focused on the
molecular and genetic level, providing important new mechanistic information that in some cases
enhances our understanding of the complexity of the Os-plant response. This information has, in
general, reduced overall uncertainties at the subcellular and cellular scales (U.S. EPA, 2013,
section 9.3.6).
Other recent information has also reduced some associated uncertainties regarding Os
impacts at the whole plant, species, and ecosystem scales. Importantly, one key uncertainty
related to the potential broader applicability of OTC-generated tree seedling E-R functions to
estimate biomass loss under different (i.e., field) Os exposure conditions has been significantly
reduced (U.S. EPA, 2013, section 9.6). Using recent field-based information available in the
5-15
-------
current review, we conducted an analysis comparing OTC data with FACE data for one crop and
one tree species (U.S. EPA, 2013, section 9.6.3.2). One comparison was done using soybean
OTC data from the National Crop Loss Assessment Network (NCLAN)8 and more recent field-
based data from the SoyFACE experiment, as discussed in section 5.3 below. The second was
done using aspen seedling OTC data from the NHEERL-WED studies and more recent field-
based data from the Aspen FACE study site. The result of the aspen analysis showed very close
agreement between the biomass loss predictions based on OTC data and Aspen FACE
observations, even when comparing the results of experiments that used different exposure
methodologies, different genotypes, locations, and durations. The soybean analysis showed
similar agreement between the OTC data and the SoyFACE experiment. Based on this analysis,
the ISA concluded that "[o]verall, the studies at the Aspen FACE experiment were consistent
with many of the open-top chamber (OTC) studies that were the foundation of previous Os
NAAQS reviews" and that "[tjhese {recent} results strengthen the understanding of Os effects on
forests and demonstrate the relevance of the knowledge gained from trees grown in OTC
studies" (U.S. EPA 2013, p. 2-38, Section 9.6.3). The ISA additionally notes that with respect to
aspen, "the function based on one year of growth was applicable to subsequent years" (of the six-
year dataset) (U.S. EPA, 2013, section 9.6.3.2). This result is significant in that it shows that at
least for this species, the seedling E-R function was able to predict responses beyond the seedling
growth stage. While recognizing that some uncertainties remain for E-R functions for some
individual species for which the database is relatively less robust, taken together, this
information substantially reduces uncertainties associated with use of the tree seedling OTC-
derived E-R functions to predict the response of tree seedlings in field settings and in some cases
beyond the seedling growth stage. This information in combination with results from recent
meta-analyses, as discussed above, reduces the uncertainties associated with potential impacts of
other experimental factors on the Os-plant response. Thus, in the current review, we have greater
confidence than in the last review in using these E-R functions to estimate tree growth response
outside the chamber setting (U.S. EPA, 2013, section 9.6.2; U.S. EPA, 2014a, section 6.2).
Several uncertainties are specific to studying or modeling Os impacts on trees, and derive
from the long lifespan of trees, which can range from decades to centuries. Because most studies
are designed to take place within an annual or 2-3 year timeframe, typically information is
available for only a small fraction of the lifetime of a tree. Given this reality, one uncertainty
8 The NCLAN program was conducted from 1980 to 1987 at five different locations across the US. At
each site, open top chambers were used to expose plants to Os treatments that represented the range of
concentrations that occur in different areas of the world. The NCLAN focused on the most important U.S
agricultural crops (Heagle et al, 1989; http://www.ars.usda.gov/Main/docs.htm?docid=12462).
5-16
-------
that remains is the degree to which exposures in a single year or over multiple years affect trees
over the longer term. However, as discussed above, recent studies from the Aspen FACE site
have reduced this uncertainty by providing additional evidence that demonstrates that exposures
in one year have the potential to cause effects in a subsequent year (carry-over effects) and that
the annual effects from exposures over multiple years have the potential to compound (U.S.
EPA, 2013, 9.4.3, pp. 9-42 to 9-47). Such effects, when they cumulate from one or more years
of elevated Os exposures, can lead to more serious longer-term impacts on growth, reproduction,
recruitment, and competitive interactions within forest stands, and at larger spatial scales (U.S.
EPA, 2013, p. 1-8), which would also have ramifications for any associated ecosystem services.
In recognition of this recent evidence, the current CASAC Panel advised that "[a] 2% biomass
loss is an appropriate scientifically based value to consider as a benchmark of adverse impact for
long-lived perennial species such as trees, because effects are cumulative over multiple years"
and stated that in its "scientific judgment, it is appropriate to identify a range of levels of
alternative W126-based standards that includes levels that aim for not greater than 2% RBL for
the median tree species" (Frey, 2014, p. 14). The CASAC further states that it "considers it
significant that a similar value of 1% - 2% for tree seedling biomass loss was recommended
previously by a consensus meeting of experts on ecological effects of ozone (Heck and Cowling,
1997)" (Frey, 2014, p. 14).
A related uncertainly comes from the limited evidence showing that sensitivity to Os can
vary over the lifespan of trees and that this variation in growth-stage sensitivity is species-
specific. For example, some species have been shown to be more sensitive during younger
growth stages (i.e., seedling/sapling) while other species may be more sensitive as adults.
Though a few studies have examined tree growth beyond the seedling stage (e.g., aspen) and in
some species has been measured for both seedling and mature trees within a species (e.g., red
oak), for most studied tree species it remains uncertain to what degree effects observed during
one growth stage can be extrapolated to other growth stages. An analysis in the WREA
comparing seedling to adult tree biomass loss, discussed in 5.2.2 below, informs our
consideration of this remaining uncertainty (U.S. EPA, 2014a, section 6.2.1.1).
These uncertainties are taken into account when we consider how much weight to put on
predictions of risks for known effects and how precautionary it is appropriate to be in light of the
potential for cumulative effects from multiple year exposures that could reasonably be
anticipated to occur, based on the evidence above.
5-17
-------
• To what extent does currently available evidence suggest locations where the
vulnerability of sensitive species, ecosystems and/or their associated services to
Os-related effects on tree growth, productivity and carbon storage would have
special significance to the public welfare?
A number of different types of locations provide services of special significance to the
public welfare. These services can flow in part or entirely from the vegetation that grows there
(see also discussion under section 5.1 above). With respect to forested lands, the WREA notes
that there are approximately 751 million acres of forest lands in the U.S., one third of which (250
million acres) is federally owned (U.S. EPA, 2014a, p. 5-15). In order to identify what types of
forest locations have special significance from a public welfare perspective, it is first useful to
consider the types of services that can flow from forested areas, and more specifically, from
forested areas with trees that are sensitive to Os. Some sensitive tree species provide public
welfare benefits based on their cultural significance, and some lands are important to the public
welfare for their cultural value. For example, tribal lands, federally designated Class I areas,
non-Class I national parks and wilderness areas, and other areas set aside to provide similar
public welfare benefits, are valued for their cultural services such as outdoor recreation and
aesthetics. Appendix 5A includes a table listing known Os-sensitive species, including some
trees that have been identified as having cultural importance to some tribes (U.S. EPA, 2014a,
section 6.4.2). Locations where these species are growing and are used by tribes to support
cultural practices would thus be potentially vulnerable to impacts from elevated cumulative Cb
exposures, which could result in the loss of those associated cultural services, including those
associated with sensitive tree species. Class I areas and other parks have also been afforded
special federal protection to preserve services such as a healthy natural environment that
provides for the enjoyment of these resources unimpaired for current and future generations,
sustainable native plant and wildlife populations, and unique recreational opportunities. As
mentioned above, 66 Cb-sensitive species have been identified on NPS and U.S. Fish and
Wildlife Service lands).9 Other forested lands, both public and private, where trees are grown for
timber production could also be at risk, especially in a single timber species stand that is
sensitive to Os (i.e., Ponderosa pine) (see WREA section 6.3 and section 5.2.2 below). Urban
forests provide a number of important services to the public, such as air pollution removal,
cooling of the heat island effect, and beautification (U.S. EPA, 2014a, section 6.6.2). These
9 See http://www2.nature.nps.gov/air/Pubs/pdf/flag/NPSozonesensppFLAG06.pdf
5-18
-------
urban forests have also been recognized as important to environmental justice communities.10
Because urban forests can include Os-sensitive trees (e.g., black cherry), Cb exposures have the
potential to reduce the services they provide. The WREA analysis of five urban case study areas,
discussed below, quantified the Os impacts on air pollution removal and carbon sequestration in
those urban areas (WREA, sections 6.6.2 and 6.7; section 5.2.2 below). Black cherry, for
example, was one of the top ten occurring species in four of the five case study areas.
The above types of forested lands have clearly designated purposes or intended uses that
help define the types of services that might be recognized as important from a public welfare
perspective. In addition, other services provided by trees are potentially extremely valuable, but
limited information is available to quantitatively value the extent of these services. Perhaps one
of the most significant of these ecosystem services is climate regulation, which provides
widespread and long-lasting public welfare benefits that the ISA determined is likely being
compromised by the phytotoxic effects of Os on tree growth, productivity and carbon storage.
By reducing the amount of carbon taken up by plants, more CCh is allowed to remain in the
atmosphere where it potentially exacerbates the effects of climate change. In contrast to the
location-specific discussion of services above, this service is potentially important to the public
welfare no matter in what location the sensitive trees are growing, or what their intended current
or future use. In other words, the benefit exists as long as the tree is growing, regardless of what
additional functions and services it provides.
In addition to identifying forested locations that provide ecosystem services that are
important to the public welfare, we must also consider to what extent there is the potential for Os
to affect sensitive tree species growing on those lands to a degree sufficient to affect the public
welfare. In so doing we first note that not all tree species are equally sensitive to Os and thus not
equally vulnerable to current ambient Os exposures or those anticipated under various air quality
scenarios. In further considering the degree to which Os-induced impacts to ecosystem services
associated with such trees might be expected to occur, we first focused on the 12 species of trees
for which we have E-R functions. While all of these species provide goods and services that are
important to the public welfare, not all species are equally sensitive to Os under recent ambient
exposure conditions or conditions projected for adjusted air quality. Table 5-1 below (modified
from WREA Table 6-13), provides a more detailed description of the ecosystem services
provided by each of these species that benefit the public welfare. For the purposes of this
10 See http://www.fs.fed.us/research/urban/environmental-justice.php and Federal Interagency Working
Group on Environmental Justice. (2011). Community-Based Federal Environmental Justice Resource Guide.
August. Available at http://www.epa.gov/environmentaljustice/resources/publications/interagencv/resource-
guide.pdf
5-19
-------
discussion we have ordered the species in the table to go in descending order from most to least
sensitive (based on their predicted relative biomass loss (RBL) at a W126 of 15 ppm-hrs).
Table 5-1. Os-Sensitive Trees, Their Uses and Relative Sensitivity
Tree Species Os Effect
Role in Ecosystems and Public Welfare Uses
Eastern
Cottonwood11
Populus
deltoides
Biomass loss
Containers, pulp, and plywood
Erosion control and windbreaks
Quick shade for recreation areas
Beaver dams and food
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 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 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 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 for many bird species, including the red-winged blackbird, chickadee,
finches, and nuthatches
11
The E-R function for cottonwood is considered less robust because it is based on the results of a single
gradient study (Gregg et al., 2003).
5-20
-------
Tree Species
Os Effect
Role in Ecosystems and Public Welfare Uses
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 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 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 important food source for some species of butterflies and
moths.
Virginia Pine
Pinus
virginiana
Biomass loss,
Visible foliar
injury
Pulpwood, strip mine spoil banks and severely eroded soils
Nesting for woodpeckers, food 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
Medicinal uses, spiritual and cultural uses for several Native American
tribes
Spotted owl habitat
Food for mammals including antelope and mountain sheep
* Sensitivity categories added by EPA staff but not based on official designations.
Sources: USDA-NRCS, 2013; Burns, 1990; Hall and Braham, 1998. ARed maple information from
http://www.na.fs.fed.us/pubs/silvics manual/volume 2/acer/rubrum.htm. *Loblolly pine use information from
http://www.ncsu.edu./project/dendrologv/index/plantae/vascular/seedplants/gvmnosperms/conifers/pine/pinus/austra
les/loblollypine/html.
5-21
-------
While we recognized that there are important ecosystem services provided by those
species that are less sensitive to Cb, those species would likely receive less benefit from
additional protection below the current standard. In contrast, the other species would likely see
improvements in their associated ecosystem services, some significantly, from an improvement
in air quality. However, at the highest end of the known sensitivity spectrum, there are different
issues that must be considered when evaluating the usefulness of this information in answering
the above questions. The E-R function that is available for cottonwood is based on the results of
a single gradient study (Gregg et al., 2003) and is considered less robust than the other E-R
functions developed in OTCs. That combined with its apparent extreme response to Os
prompted CASAC to advise the Administrator to not place too much emphasis on cottonwood in
the review of the secondary standard (Frey, 2014, p. 10). As a result, we have decided it would
not be appropriate to use the cottonwood biomass loss estimates when considering what levels of
W126 should be considered protective of median species biomass loss (see Table 5C-3).
However, in this discussion of ecosystem services, we believe it is important to include
cottonwood, given the many ecosystem services cottonwood provides (see Table 5-1 above), and
several unique features that potentially make it and its associated ecosystem goods and services
particularly vulnerable to impacts from Os. Specifically we note that cottonwood: 1) is often
found growing along streams in riparian zones under well watered conditions that make it more
susceptible to injury than species growing in areas that experience drier conditions in
conjunction with higher Os exposures; 2) can be the only tree species growing in certain types of
ecosystems, thus providing important habitat for some organisms; 3) is fast growing and used
commercially for pulpwood, manufacturing furniture and as a possible source for energy
biomass (Burns and Hankola, 1990); 4) has provided limited, though still uncorroborated,
evidence of the potential for the existence of extremely sensitive plant species which can
reasonably be anticipated to exist and that could be impacted at similar cumulative exposures.
With regard to the latter, we observe that CASAC also expressed the view that it "should be
anticipated that there are species of vegetation that are highly sensitive to ozone that do not have
E-R functions, and others that are insensitive. It is scientifically justifiable to extrapolate from
the known E-R curves, assuming that they are representative of the un-sampled population"
(Frey, 2014, p. 16). We also note that upon revisiting the available literature in the ISA
following CAS AC's review of the second draft WREA and PA, we found two studies on a
related European species (Populus nigra) that showed that this species had an Os sensitivity that
appears similar in magnitude to the U.S. cottonwood (Populus deltoides) based on its response
for other growth endpoints as compared with the response of the other study species (Bortier, et
al., 2000; Novak, et al., 2007) (U.S. EPA, 2006, AX9, pp. 91, 240; U.S. EPA, 2014a, Table 6-5).
5-22
-------
This additional limited evidence ofPopulus seedling/sapling growth response, though not
directly comparable to the U.S. study (Gregg et al., 2003) with respect to species, exposure
methods, measurement endpoints and exposure values, does, in our judgment, lend some support
to the observed magnitude of the reported U.S. cottonwood response.
In addition to the information provided here on these 12 species, we note that there are
many other species of trees with known or suspected Os-sensitive vegetation, such as those
included in the 66 species identified on NFS and US Fish and Wildlife Service lands),12 species
used in the USFS biomonitoring network, and various ornamental and agricultural species (i.e.,
Christmas trees, fruit and nut trees) that currently provide ecosystem services important to the
public welfare, but whose vulnerability to impacts from Os on tree growth, productivity and
carbon storage has not been sufficiently characterized to allow it to directly inform our
quantitative assessments (U.S. EPA, 2014a, Chapter 6; Abt Associates, 1995). However, as
noted by CASAC, the anticipated impacts on these and other unstudied species should not be
ignored or assumed insignificant. It is more likely that the range of Os sensitivities found in the
studies tree species likely reflects the range of Os sensitivities in all tree species.
Other factors that should be taken into account when considering the potential degree to
which Os might affect the ecosystem service flows from forested ecosystems are 1) the type of
stand or community in which the sensitive species is found (i.e., single species versus mixed
canopy); 2) the role or position the species has in the stand (i.e., dominant, sub-dominant,
canopy, understory); 3) the Os sensitivity of the other co-occurring species (Os sensitive or
tolerant); 4) environmental factors (drought or well watered conditions, other stressors).
In light of the above discussion, it is clear that there are numerous locations where the
vulnerability of Os-sensitive tree species to impacts from Os on tree growth, productivity and
carbon storage and their associated ecosystems and services could have special significance to
the public welfare. Confirmation that the American public values healthy forests is provided in
the WREA, which shows that Americans are willing to pay to protect forests from the damaging
effects of air pollutants (U.S. EPA, 2014a, Chapter 5, pp. 5-16). Data provided by the National
Survey on Recreation and the Environment (NSRE) indicates that Americans have very strong
preferences for the non-use values of existence, bequest, and option services related to forests.
Studies (Haefele et al., 1991, Holmes and Kramer, 1995) assess willingness-to-pay (WTP) for
spruce-fir forest protection in the southeastern U.S. from air pollution and insect damage and
confirm that the non-use values held by the survey respondents were in fact greater than the use
or recreation values. The results of this survey showed that median household WTP was
12 See http://www2.nature.nps.gov/air/Pubs/pdf/flag/NPSozonesensppFLAG06.pdf
5-23
-------
estimated to be roughly $29 (in 2007 dollars) for the minimal protection program and $44 for the
more extensive program. After decomposing their value for the extensive program into use,
bequest, and existence values, the results were 13 percent for use value, 30 percent for bequest,
and 57 percent for existence value (See U.S. EPA, 2014a, Table 5-6). These services may be at
risk in areas where Os-sensitive trees are found.
• To what extent does the available evidence indicate the occurrence of Os-related
effects on forest growth, productivity and carbon storage attributable to
cumulative exposures lower than previously established or that might be expected
to occur under the current standard?
The evidence base available in this review, as in the previous review, indicates that Cb-
induced effects on tree growth, productivity and carbon storage can occur across a range of
cumulative exposures, including those lower than previously established and that would be
expected to occur under the current standard. In reaching this determination, we first consider
the 11 tree seedling species for which robust E-R functions have been developed from the
extensive evidence base of Os-induced growth effects that was also available and relied upon in
the previous review. Each of these species were studied in OTCs, with most species studied
multiple times under a wide range of exposure and/or growing conditions, with separate E-R
functions developed for each species/exposure condition/growing condition scenario
combination or case. Using all the information available from these multiple study cases (52
cases in all), a robust composite E-R function was developed for all species combined and
separate individual composite functions were derived for each species using cases that were
available on individual species. These species-specific composite E-R functions have been
successfully used to predict tree seedling species biomass loss response over a range of
cumulative exposure conditions. Figure 5-1 A, B below, which includes the 11 robust composite
E-R functions available in the last review and the E-R for cottonwood (also described in U.S.
EPA 2013, section 9.6.2 and U.S. EPA, 2014a, section 6.2, Table 6-1 and Figure 6-2), illustrates
the appreciable variability in sensitivity that exists across the 12 studied species, and shows that
biomass loss can occur over a wide range of cumulative exposures, including those previously
established. This figure further shows that for some species biomass loss would be predicted to
occur at very low cumulative exposures that can occur under air quality conditions that meet or
are below the current standard (see Table 5-2 below). While we put less emphasis on
cottonwood (as explained above), we do note that in answering the question above, it does
provide limited recent evidence of the potential for effects of a greater magnitude and at lower
cumulative exposures to occur than those considered in the last review and at exposures that
would be allowed by the current standard. To the extent that such effects could be anticipated,
5-24
-------
the cumulative exposures that could be allowed by the current secondary standard would not be
protective.
cq
d>
CD
CJ
DQ
Ct.
CNj
d>
O
CD
n
n
n
•
n
n
n
n
Red Maple
Sugar Maple
Red Alder
Tulip Poplar
Ponderosa Pine
White Pine
Loblolly Pine
Virginia Pine
Cottonwood
Aspen
Black Cherry
Douglas Fir
0
10
20
30
40
50
W126
Figure 5-1. A) Relative biomass loss in seedlings for 12 studied species using composite
functions in response to seasonal Os concentrations in terms of seasonal W126 index
values, Y-axis scale for RBL values represents 0% up to 100% (U.S. EPA 2014a, Figure
6-2).
5-25
-------
CO
cr
o
CD
CO
O
CD
O
o
o'
CN
O -I
O
o J
Red Maple
Sugar Maple
Red Alder
Tulip Poplar
Ponderosa Pine
White Pine
Loblolly Pine
Virginia Pine
Cottonwood
Aspen
Black Cherry
Douglas Fir
0
10
W126 (ppm-hrs)
15
20
B) Expanded view of relative biomass loss in seedlings for 12 studied species using
composite functions in response to seasonal Os concentrations in terms of lower range of
seasonal W126 index values, Y-axis scale for RBL values represents 0% up to 10% (U.S.
EPA 2014a, Figure 6-2).
In further answering the question above, we note CASAC's advice that a 6% median
RBL is unacceptably high, and that the 2% median RBL is an important benchmark to consider.
Based on the information above, the median RBL is at or below 2% at the lowest W126 level
assessed, 7 ppm-hrs. As the W126 level is incrementally increased, median RBL also increases
incrementally, so that at W126 index values of 9, 11, 13, 15, 17, 19 and 21, the median RBL
increases to 2.4%, 3.1%, 3.8%, 4.5%, 5.3%, 6.0% and 6.8%, respectively. Based on air quality
analyses of 2009-2011 (Appendix 2B), there are approximately 342, 199, 92, 43, 24, 9, 3 and 0
monitors with 3-year average W126 index values above 7, 9, 11, 13, 15, 17, 19 and 21 ppm-hrs
when meeting the current standard. We note that these counts of monitors are based on those
meeting the current standard and that there are many monitors for the 2009-2011 period that do
not meet the current standard and also are above the W126 values of 7-21 ppm-hrs.
We also consider it informative to examine the individual species responses and RBL
over the same W126 range. We first note, based on Figure 5-l(B) above that over the range of 7
to 17 ppm-hrs, 5 species maintain RBLs of less than 2%. These more tolerant species include
5-26
-------
Douglas fir, loblolly pine, Virginia pine, sugar maple and red maple. Two of these species (red
maple and sugar maple) are estimated to have RBL levels above 2% at a W126 of 21. Black
cherry, the most sensitive of the remaining six species, has RBL ranging from 35.57% at W126
of 17 down to 16.67% at the W126 index value of 7 ppm-hrs.
Additional evidence of the potential for Os-induced effects on tree seedling growth,
productivity and carbon storage occurring under air quality scenarios allowed by the current
standard is shown in Table 5-2 below. Specifically, all monitor sites in Table 5-2 have 3-year 8-
hour average values that meet the current standard, ranging from 67 to 75 ppb, have 3-year
average W126 index values that are above 15 ppm-hrs, and are located in Class I areas. Across
these 22 Class I areas, the highest single-year W126 index values for these three-year periods
ranged from 17.4 to 29.0 ppm-hrs. In 20 of the areas, distributed across eight states (AZ, CA,
CO, KY, NM, SD, UT, WY) and four regions (west, southwest, west/north central and central),
this range was 19.1 to 29.0 ppm-hrs, exposure values for which the corresponding median
species RBL estimates equal or exceed 6%, which CAS AC termed "unacceptably high". In
addition, given that other environmental factors can influence the extent to which Os may have
the impact predicted by the E-R functions in any given year, we also note that the highest three
year periods, that include these highest annual values for the 21 areas, are at or above 19 ppm-
hrs, ranging up to 22.5 ppm-hrs (which the median species RBL estimate is above 7%).
Additionally, the highest three-year average W126 index value for each of the 22 areas (during
periods meeting the current standard) was at or above 19 (ranging up to 22.5 ppm-hrs) in 11
areas, distributed among five states in the west and southwest regions (U.S. EPA, 2014c, Table
5-2, Appendix 5B).
In addition, as data permit, Table 5-2 shows the studied tree species that are found in each
of these Class I areas. Quaking aspen and ponderosa pine are two tree species that are found in
many of these 22 parks and have a sensitivity to Os exposure that places them near the middle of
the group for which E-R functions have been established. In the areas where ponderosa pine is
present, the highest single year W127 index values ranged from 18.7 to 29.0 and the highest 3-
year average W126 values in which these single year values are represented ranged from 15 to
22.5, with these three-year values above 19 ppm-hrs in eight areas across five states. The
ponderosa pine RBL estimates for 29 and 22.5 ppm-hrs are approximately 12% and 9%,
respectively. In areas where quaking aspen is present, the highest single year W127 index values
ranged from 19.2 to 26.7 ppm-hrs and the highest 3-year average W126 values in which these
single year values are represented ranged from 15.0 to 22.2, with values above 19 ppm-hrs in
eight areas across five states. The quaking aspen RBL estimates for 26.7 and 22.2 ppm-hrs are
approximately 16% and 13%, respectively. Based on this, we note growth effects associated with
5-27
-------
exposure concentrations occurring during periods where the current standard is met in many of
these Class I areas. On the basis of such information, Table 5-2 provides evidence of the
potential for significant growth loss in locations where ambient conditions meet the current
standard.
Table 5-2. Os concentrations in Class I areas during period from 1998 to 2012 that met
the current standard and where three-year average W126 index value was at or above
15 ppm-hrs.*
Class I Area
Bandelier Wilderness Area
QA, DF, PP
Bridger Wilderness Area
QA.DF
Canyonlands National Park
QA, DF, PP,
Carlsbad Caverns National
Parkpp
Chiricahua National
Monument DF>PP
Grand Canyon National
parkQA,DF, PP
John Muir Wilderness Area
QA, DF, PP
Lassen Volcanic National
parkDF,PP
Mammoth Cave National
Park BC> c> LPi RM' SM> w' """
Mesa Verde National Park
DF
Mokelumne Wilderness
AreaDF'pp
Petrified Forest National
Park
Pinnacles National
Monument
Rocky Mountain National
ParkQA,DF,PP
Saguaro National ParkDF'pp
Sierra Ancha Wilderness
AreaDF'pp
Superstition Wilderness
Areapp
State /
County
MM/
Sandoval
WY/Sublette
UT / San Juan
NM/Eddy
AZ / Cochise
AZ/
Coconino
CA/Inyo
CA / Shasta
KY/
Edmonson
CO/
Montezuma
CA / Amador
AZ /Navajo
CA/San
Benito
CO /Boulder
CO /
Larimer
AZ / Pima
AZ / Gila
AZ/
Maricopa
Design
Value
(ppb)*
70-74
69-72
69-73
69
69-73
68-74
71-72
75
74
67-73
74
70
74
73-75
74
69-74
72-75
75
3-year Average
W126
(ppm-hrs)*
(#> 19 ppm-hrs,
range)
15.8-20.8 (2, 20.0-
20.8)
15.1-17.4
15.0-20.5 (2, 19.8-
20.5)
15.0-15.3
15.7-18.0
15.6-22.2 (7, 19.2-
22.3)
16.5-18.6
15.3
15.9
15.5-21.0 (2, 19.0-
21.0)
17.6
15.7
15.1
15.1-19.3(1, 19.3)
15.0-18.3
15.4-18.9
17.9-22.4 (3, 20.2-
22.4)
22.4 (1, 22.4)
Annual W126
(ppm-hrs)*
(# > 19 ppm-hrs, range)
12.1-25.3 (4, 19.2-25.3)
9.9-19.2(1, 19.2)
9.9-24.8 (5, 19.3-24.8)
8.6-26.7 (1, 26.7)
13.2-21.6 (2, 19.3-21.6)
11.3-26.7(7, 19.8-26.7)
10.1-25.8(2,23.9-25.8)
13.6-18.7(1, 18.7)
12.5-22.5 (1, 22.5)
10.7-23.6 (4, 19.7-23.6)
14.8-22.6 (1, 22.6)
12.9-19.2 (1, 19.2)
13.1-17.4
9.5-25.1(5,20.7-25.1)
11.1-25.8(3, 19.1-25.8)
11.0-23.1(3,20.0-23.1)
14.8-27.5 (4, 20.3-27.5)
14.5-28.6 (2, 27.4-28.6)
Number of
3-year
Periods
8
5
9
3
7
12
3
1
1
10
1
1
1
6
3
6
4
1
5-28
-------
Weminuche Wilderness
AreaQA,DF,PP
West Elk Wilderness Area
QA, DF
Wind Cave National Park
QA, PP
Yosemite National Park QA>
DF, PP
Zion National Park QA' DF- pp
AZ / Final
CO / La Plata
CO/
Gunnison
SD / Custer
CA/
Tuolumne
UT/
Washington
73-75
70-74
68-73
70
73-74
70-73
18.7-22.5 (2, 20.9-
22.5)
15.0-19.1(1, 19.1)
15.6-20.1 (1,20.1)
15.4
20.7-20.8 (2, 20.7-
20.8)
17.8-21.1 (2,20.3-
21.1)
14.8-29.0 (3, 22.6-29.0)
10.9-21.0 (2, 20.8-21.0)
12.9-23.9(3,21.1-23.9)
12.2-20.6 (1, 20.6)
19.7-22.1(4, 19.7-22.1)
14.9-24.2 (5, 19.3-24.2)
3
5
8
1
2
4
*Based on data from http://www.epa.sov/ttn/airs/airsaqs/detaildata/downloadaqsdata.htm (US EPA, 2014c). W126
values are truncated after first decimal place.
Superscript letters refer to species present for which E-R functions have been developed.
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 for presence of species include U.S. Department of Agriculture databases in 2014
http://www.fs.fed.us/foresthealth/technology/nidrm2012.shtml, http://plants.usda.gov,
http://www.wilderness.net/printF actSheet.cfm?WID=583
In answering the above question, we note that less information is available from field-
based studies (e.g., FACE, gradient) due to the absence of robust E-R functions, the limited
range of exposure scenarios evaluated, and unavailability of study exposures in terms of daily 8-
hour averages.
Taken together, the information described above provides consistent and coherent
evidence that Os-induced impacts on tree seedling growth, productivity and carbon storage are
occurring at cumulative exposures allowed by the current standard. In particular, this information
provides clear evidence of the potential for significant growth loss in Class I locations where
ambient conditions meet the current standard.
5.2.2 Exposure/Risk-based Considerations
The WREA presents a number of quantitative analyses of exposure and risk related to
tree growth, productivity and carbon storage intended to inform our consideration of exposure
and risk associated with the current and potential alternative standards (Table 5-3 below; U.S.
EPA, 2014a, Chapter 6).
5-29
-------
Table 5-3. Exposure, risk and ecosystem services analyses related to tree growth,
productivity and carbon storage.
WREA
estimatesA
Species Level Effects
Derivation of median biomass
loss values from individual
species E-R functions
Comparison of tree seeding
growth to that of mature trees
Ecosystem Level Effects
Percent of total geographic
areaB with annual relative
biomass loss above 2%
Number of assessed Class I
areas with annual relative
biomass loss above 2%
Ecosystem Services
• Economic surplus to timber
producers and consumers
(WREA, Table 6-12)
• Carbon storage, nationally
(WREA, Table 6-19)
• Carbon storage, in 5 urban
areas (WREA, Table 6-21)
• Air pollutant removal in 5
urban areas (WREA, Table
6-22)
A See WREA Chapter 6 (U.S. EPA, 2014a).
B The total geographic area includes only the contiguous U.S.
The relevant quantitative exposure and risk analyses for tree biomass loss, productivity
and carbon storage include:
1) Species-specific and median biomass loss estimates from composite functions.
2) National-scale assessments for: a) basal area weighted relative biomass loss for tree
seedlings; b) timber production; c) carbon sequestration.
3) Case study-scale assessments for: a) carbon sequestration; b) air pollution removal.
• For what air quality scenarios were exposures and risks estimated? What
approaches were used to estimate W126 exposures for those conditions? What
are associated limitations and uncertainties?
Quantitative exposure and risk analyses were conducted to evaluate the effects on tree
growth, productivity and carbon storage, and associated ecosystem services, that would be
predicted under five air quality scenarios (recent ambient, just meeting the current standard, and
W126 potential alternative standards of 15, 11, and 7 ppm-hrs). Table 5-5 summarizes the
methodology used to develop the quantitative estimates for each of the five air quality scenarios.
In general, this methodology involved two steps. The first is derivation of the average W126
index value (across the three years) at each monitor location. This value is based on unadjusted
data for recent conditions and adjusted concentrations for the four other scenarios. The
development of adjusted concentrations was done for each of 9 regions independently (see U.S.
EPA, 2014a, section 4.3.4.1). In the second step, national-scale spatial surfaces (W126 index
values for each 12x12 km2 grid cell from the Community Multi-scale Air Quality (CMAQ)
model) were created using the monitor-location values and the Voronoi Neighbor Averaging
(VNA) spatial interpolation technique (details on the VNA technique are presented in U.S. EPA,
2014a, Appendix 4A).
5-30
-------
Table 5-4. Summary of methodology by which national surface of 3-year average W126
index values was derived for each air quality scenario.
Scenario
Development of W126 index values for Each Air Quality Scenario
Monitor-location-specific calculations
and any model-based adjustment
Derivation of national surface of
average W126 index values
Recent
Conditions
(2006-2008)
An annual W126 index value is calculated for each year at
each monitor location, using the highest 3-month period. A
location-specific 3-year W126 was calculated by averaging
annual W126 index values from 3 consecutive years which
may have used different 3-month periods.
Current
Standard
2006-2008 hourly Os concentrations at each monitor location
are adjustedA to create a three year record of Os
concentrations that just meets the current standard (see
WREA, section 4.3.4). This results in air quality at other
monitors well below the level of the controlling monitor.
A seasonal W126 index value is calculated for each year at
each monitor location using the same 3-month period for
each year (which is the highest as a 3-yr average and is
highest in at least one of the years). A location-specific
average is derived from these three index values.
Average
W126 Index
of 15 ppm-hrs
Average
W126 Index
of 11 ppm-hrs
Average
W126 Index
of 7 ppm-hrs
First, hourly Os concentrations were adjusted to just meet the
current standard. Second, hourly Cb concentrations at each
monitor location, within each modeling region, are adjusted
to create a record for which the highest location-specific
average index value in the region (the controlling location)
just meets the scenario target index value.
A seasonal W126 index value is calculated for each year (of
2006-2008 period) at each monitor location, using the same
3-month period for each year (which is the highest in at least
one of the years). A location-specific average is derived
from these three index values.
The VNA method is applied to the
monitor-location average W126
index values to create a national
distribution of W126 index values
within model grid-cells for each
scenario.
A The model-based adjustment approach is based on regional emission reduction scenarios at monitor sites
followed by spatial interpolation for broader spatial coverage. See WREA, chapters 3 and 4, and Appendix 4A.
During the recent conditions period (2006 through 2008), the average W126 index values
(across the three-year recent conditions period) at the monitor locations ranged from below 5
ppm-hrs to 48.6 ppm-hrs (U.S. EPA 2014a, Figure 4-4 and Table 4-3). Across the nine modeling
regions, the maximum average W126 index values ranged from 48.6 ppm-hrs in the west region
down to 6.6 ppm-hrs in the northwest region. After adjusting the 2006-2008 data to just meet the
current standard in each region, the region-specific maximum values range from 18.9 ppm-hrs in
the west region to 2.6 ppm-hrs in the northeast region (U.S. EPA, 2014a, Table 4-3). After
application of the VNA technique to the current standard scenario monitor location values, the
average W126 index values were below 15 ppm-hrs across the national surface with the
5-31
-------
exception of a very small area of the southwest region (near Phoenix) where the average W126
index values was near or just above 15 ppm-hrs. Thus, it can be seen that application of the
interpolation method to estimate W126 index values at the centroid of every 12 x 12 km2 grid
cell rather than only at each monitor location results in a lowering of the highest values.
• What are the nature and magnitude of exposure- and risk-related estimates for
tree growth, productivity, and carbon storage under recent conditions or
conditions remaining upon meeting the current standard? To what extent are
these exposures and risks important from a public welfare perspective?
In answering the above question, the WREA performed a number of different
assessments to estimate the exposures and risks predicted under the five air quality scenarios
across a range of spatial scales. These assessments include those for individual species response
as well as the median species response for studied species ranging from the county scale up to
estimations of exposures and risks to ecosystem services associated with forests at the urban,
park and national scales.
Before conducting the exposure and risk assessments, the WREA examined three
approaches for characterizing the median response, as shown in Figure 5-2 below (U.S. EPA,
2014a, section 6.2.1.2 and Figure 6-5). These approaches use the 11 robust E-R functions for
tree seedlings from the OTC research and the cottonwood E-R function. For some species, only
one study was available (e.g., red maple), and for other species there were as many as 11 studies
available (e.g., ponderosa pine). The first approach plotted the median (red line) of all 52 tree
seedling studies available (across the 12 species). In this first approach, species with multiple
studies would be represented more than once in the median. The second approach characterized
the median (green line) by combining the composite E-R functions, when available for species
with multiple studies, with the E-R functions for species with a single study available13 for each
of the 12 tree species. In this second approach, each species is represented only once in the
median. The third approach used a stochastic sampling method to randomly select a single E-R
function from the studies available for each of the 12 species. The process was repeated 1,000
times (grey lines), and the median value was plotted for biomass loss values of 1% to 7%, and
10% (red dots; the bar associated with each median point denotes the 25th and 75th percentile
values). This third approach illustrates the effect of within-species variability on estimates of the
median response. The median W126 index values are similar when using the first two
approaches; however, the median value is higher when within-species variability is included
(U.S. EPA, 2014a, section 6.2.1.2). Across these three approaches, the median seasonal W126
index value for which a two percent biomass loss is estimated in seedlings for the studied species
13 For some species, only one study was conducted so that E-R function was used.
5-32
-------
ranges between approximately 7 and 14 ppm-hrs. Using the green line, the seasonal W126 index
value for which a two percent biomass loss is estimated in seedlings for the median of the
composite functions for the 12 studied species is approximately 7 ppm-hrs. After reviewing these
three approaches, the CASAC stated "[t]he Monte Carlo analysis (red dots, Figure 5-2) should
not be used in evaluating the effect of ozone on RBL of tree seedlings. This analysis
overemphasizes the species for which relatively few E-R functions are available, is biased
toward the few less sensitive response functions available for some individual species, makes
unsupported assumptions regarding the representativeness of available response functions, and
confounds intra- and inter-species variability in unquantifiable ways. We favor using a measure
of central tendency of the data, specifically the median across species (the green line in Figure 5-
2). This analysis provides the median of best available estimates within each species, and the
median across species with all species treated equally" (Frey, 2014, p. 14). Given this advice, in
selecting an approach for use in later analyses, we have chosen to use the green line because the
approach that generated it incorporates all the information in a way that gives equal weight to
each studied species without losing any of the available data.
LIT
I
P
E.
Q.
5 6 7 8 9
Percent Biomass Loss
I I I I
10 11 12 13 14 15
Figure 5-2. Relationship of tree seedling percent biomass loss with seasonal W126 index.
(From U.S. EPA 2014a, Figure 6-5)
The WREA used the E-R functions for 12 species described above with information on
the distribution of those species across the U.S., and average W126 exposure estimates to
5-33
-------
estimate relative biomass loss for each of the studied species for each national air quality
scenario (U.S. EPA, 2014a, section 6.2.1.3 and Appendix 6A). For example, the estimates of
relative biomass loss of ponderosa pine for air quality adjusted to just meet the current standard
are illustrated in Figure 5-3 below. While relative biomass loss below 2% is estimated for most
areas where this species is found, estimates in some areas of the southwest fall above 2%
biomass loss (U.S. EPA 2014a, Figure 6-8 and Appendix 6A).
Ponderosa Pine (Pinus ponderosa) (Current Standard)
RBL
0.000767 - 0 003953
0.003954 - 0.005914
0.005915-0.008482
0.008483 - 0.011989
0.011990-0.015976
0.015977-0.021390
0.021391 - 0.028521
0.028522 - 0.040549
Figure 5-3. Relative biomass loss of Ponderosa Pine for air quality adjusted to just meet
the current standard (U.S. EPA 2014a, Figure 6-8).
The WREA also developed national-scale estimates of Os biomass ecosystem-level
impacts considering the 12 studied species together (U.S. EPA 2014a, section 6.8, Table 6-25).
This was done using the species-specific biomass loss E-R functions, information on prevalence
of the studied species across the U.S., and a weighting approach based on proportion of the basal
area within each grid cell that each species contributes. The RBL values for multiple tree species
were weighted by their basal area and combined into a weighted RBL value (wRBL). The
wRBL is intended to inform our understanding of the potential magnitude of the ecological effect
that could occur in some ecosystems. Specifically, the more basal area that is affected in a given
5-34
-------
ecosystem, the larger the potential ecological effect. A wRBL value for each grid cell is
generated by weighting the RBL value for each studied tree species found within that grid cell by
the proportion of basal area it contributes to the total basal area of all tree species within the grid
cell, and then summing those individual wRBLs. The percent of total basal area that exceeds a
2% weighted relative biomass loss in the recent conditions scenario is 10.1% (U.S. EPA 2014a,
Table 6-25). Based on the average W126 index values estimated for the air quality scenario just
meeting the current standard across the contiguous U.S., the WREA estimates 0.2% of the total
geographic area to have a wRBL above 2% based on the E-R functions for the 11 tree species
and 0.8% based on 12 tree species (U.S. EPA 2014a, Table 6-25). We recognize that these
estimates are likely biased low as there may be other unstudied Os-sensitive tree species in some
areas that are also being impacted at those levels. Further, this analysis does not take into
account the effects of competition, which could further increase biomass loss in Os-sensitive
species.
In addition, the WREA characterized the number of counties where there would be one or
more studied tree species showed a 2% biomass loss (U.S. EPA, 2014a, Table 6-7), which is
shown in Table 5-5 below. This is consistent with CASAC advice that "rather than focusing
solely on the median relative biomass loss (RBL), the number of counties containing sensitive
tree species that are expected to have growth loss of greater than 2% should be quantified."
(Frey, 2014, p. 11). These data are presented as the number of U.S. counties in which any of the
12 studied tree species exceeds 2% RBL, further categorized by the number of studied species
that exceed that benchmark for each of the five air quality scenarios using 3-year average W126
index values. In addition, this table provides the total number of counties (out of 3,109 total
counties) for each exposure scenario with at least one species exceeding 2% RBL and the
number of counties where the median of the composite functions for each species exceeds 2%
RBL. The maximum number of species that exceed 2% RBL in any one county is five species,
which only occurs under recent Os conditions. After meeting the current standard, the maximum
number of species in any one county is four. Because cottonwood and black cherry are highly
sensitive species and to provide a reference for the effect of these species, the data are also
presented excluding cottonwood and excluding cottonwood and black cherry.
This information shows that a number of counties have more than one Os-sensitive
species growing in it, potentially together in the same forest stands, whose RBLs are above 2%.
Under recent conditions, the proportion of total counties of 3,109, with 1 or more species with an
RBL greater than 2% is 89% (2,761 counties) for the scenario inclusive of cottonwood and black
cherry. When air quality is adjusted to just meet the current standard, that proportion dropped to
74% (2,313 counties). When air quality is adjusted to just meet a 3-year average W126 index
5-35
-------
value of 15, 11 and 7 ppm-hrs, the proportion is 73%, 72% and 71%, respectively. For median
RBL values, under recent conditions, 72% of the counties have median RBLs above 2%. When
air quality is adjusted to the current standard, that proportion drops to 22% and further decreases
to 20% for air quality adjusted to just meet a 3-year average W126 level of 7 ppm-hrs.
Given CAS AC's advice to put less emphasis on cottonwood, we focus on the rows of this
table that excluded cottonwood. Under recent air quality conditions, the proportion of counties
with 1 or more species with an RBL greater than 2% is 78% (2,418 counties). As air quality is
adjusted to just meet the current standard and the alternative W126 index value of 7 ppm-hrs,
this number drops to 62% and 58%, respectively. In addition, under recent conditions, 52% of
the counties have median RBLs above 2%. When air quality is adjusted to the current standard,
that proportion drops to 8% and further decreases to 6% for air quality adjusted to just meet a 3-
year average W126 level of 7 ppm-hrs.
Table 5-5 also provides information on the influence of black cherry on the estimates
and shows that black cherry is a very sensitive species that is widespread in the Eastern U.S. We
note that of the 1,929 counties estimated to have 1 or more species with an RBL greater than 2%
when meeting the current standard, 1,805 of those counties are estimated to have black cherry as
the only specie estimated to experience this level of biomass loss. With respect to median RBL
values, of the 239 counties estimated to have a median RBL above 2% when meeting the current
standard, 203 of those counties have a RBL above 2% because of the presence of black cherry.
5-36
-------
Table 5-5. Number of Counties with Tree Species Exceeding 2% Relative Biomass Loss.
Number of species
exceeding 2% RBL
Number of Counties (3,109 Total)
Recent
Conditions
75ppb
15 ppm-hrs
11 ppm-hrs
7 ppm-hrs
All 12 tree species with E-R functions
5
4
3
2
1
0
Total counties exceeding
Counties exceeding for
the median species
134
387
765
882
593
348
2,761
2,237
-
3
24
994
1,292
796
2,313
685
-
3
22
981
1,273
830
2,279
670
-
-
14
972
1,238
885
2,224
651
-
-
5
924
1,277
903
2,206
627
11 tree species with E-R functions excluding cottonwood
5
4
3
2
1
0
Total counties exceeding
Counties exceeding for
the median species
15
180
680
933
610
691
2,418
1,604
-
-
3
46
1,880
1,180
1,929
239
-
-
3
32
1,857
1,217
1,892
221
-
-
-
14
1,818
1,277
1,832
204
-
-
-
5
1,812
1,292
1,817
172
10 tree species with E-R functions excluding cottonwood and black cherry
5
4
3
2
1
0
Total counties exceeding
Counties exceeding for
the median species
-
15
187
856
920
1,131
1,978
666
-
-
-
29
95
2,985
124
36
-
-
-
15
72
3,022
87
18
-
-
-
2
19
3,088
21
6
-
-
-
1
8
3,100
9
2
We also consider WREA estimates (quantitative and qualitative) of effects on several
ecosystem services. First, impacts on growth related to cumulative Cb exposure values in
federally designated Class I areas were derived from an average wRBL value (discussed above)
for 145 of the 156 Class I areas (U.S. EPA 2014a, section 6.8.1). Given established objectives
for Class I areas (e.g., to maintain in perpetuity), effects in Class I areas may be considered to
5-37
-------
have the potential to adversely affect the intended use of the ecosystem, e.g., to leave them
unimpaired and preserve them for the enjoyment of future generations. Under recent conditions,
this analysis estimates that 13 Class I areas have wRBL values above 2%. Further, this analysis
estimates that based on average W126 index values estimated for the air quality scenario just
meeting the current standard, 2 of the 145 Class I areas assessed would be expected to have
multiple-species, wRBL values above 2% (U.S. EPA 2014a, Table 6-26). However, we
recognize that this analysis is limited to the 12 studied tree species, and therefore could
underestimate other Os-sensitive species without E-R functions.
The WREA also presents national-scale estimates of the effects of biomass loss on timber
production and agricultural harvesting, as well as on carbon sequestration. The WREA used the
Os E-R functions for tree seedlings to calculate relative yield loss (equivalent to biomass loss)
across the trees' entire life spans. Because the forestry and agriculture sectors are related and
trade-offs occur between the sectors, the WREA also calculated the resulting market-based
welfare effects of Os exposure in the forestry and agriculture sectors.14 In the analyses for
commercial timber production, based on the 3-year average W126 index values estimated for the
air quality scenario just meeting the existing standard, RYL estimates for timber were below one
percent with the exception of the Southwest, Southeast, Central, and South regions (U.S. EPA,
2014a, section 6.3, Table 6-9) (see U.S. EPA, 2014a, Table 6-8 for clarification on region
names). At the current standard the highest yield loss occurs in upland hardwood forests in the
South Central and Southeast regions at over 3% per year and in Corn Belt hardwoods at just over
2% loss per year. Relative yield losses for timber remain above one percent for the 3-year
average W126 scenarios for 15 and 11 ppm-hrs in parts of the Southeast, Central, and South
regions, and for the 7 ppm-hrs scenario in the Southeast and South regions (U.S. EPA, 2014a,
section 6.3, Table 6-9). In addition, relative yield losses for timber were above two percent in
parts of the Southeast and Central U.S. after just meeting the existing standard as well as in the
15 ppm-hrs and 11 ppm-hrs scenario.
In addition to estimating changes in forestry and agricultural yields, the WREA presents
estimated changes in consumer and producer/farmer surplus associated with the change in yields.
Changes in biomass affect individual tree species, but the overall effect on forest ecosystem
productivity depends on the composition of forest stands and the relative sensitivity of trees
14 The WREA used the Forest and Agricultural Sector Optimization Model with Greenhouse Gases
(FASOMGHG). FASOMGHG is a national-scale model that provides a complete representation of the U.S. forest
and agricultural sectors' impacts of meeting alternative standards. FASOMGHG simulates the allocation of land
over time to competing activities in both the forest and agricultural sectors. FASOMGHG results include multi-
period, multi-commodity results over 60 to 100 years in 5-year time intervals when running the combined forest-
agriculture version of the model.
5-38
-------
within those stands. Economic welfare impacts resulting from just meeting the existing and
alternative standards were largely similar between the forestry and agricultural sectors —
consumer surplus, or consumer gains, generally increased in both sectors because higher
productivity under lower Os concentrations increased total yields and reduced market prices.
Comparisons are not straightforward to interpret due to market dynamics. For example, because
demand for most forestry and agricultural commodities is not highly responsive to changes in
price, there were more examples for which producer surplus (i.e., producer gains) declines.15 In
some cases, lower prices reduce producer gains more than can be offset by higher yields. The
increase in consumer welfare is much larger than the loss of producer welfare resulting in net
welfare gains in the forestry sector nationally. The national-scale analysis of carbon dioxide
(CCh) sequestration estimates more storage under the current standard compared to recent
conditions (U.S. EPA 2014a, Appendix 6B, Table B-10). In considering the significance of the
potential climate and ecosystem service impact, we also note the large uncertainties associated
with this analysis (see U.S. EPA 2014a, Table 6-27).
We additionally consider the WREA estimates of tree growth and ecosystem services
provided by urban trees over a 25-year period for five urban areas based on case-study scale
analyses that quantified the effects of biomass loss on carbon sequestration and pollution
removal (U.S. EPA 2014a, sections 6.6.2 and 6.7).16 The urban areas included in this analysis
represent diverse geography in the Northeast, Southeast, and Central regions, although they do
not include an urban area in the western U.S. Estimates of the effects of Cb-related biomass loss
on carbon sequestration indicate the potential for an increase of somewhat more than a million
metric tons of CCh equivalents for average W126 index values associated with meeting for the
current standard scenario as compared to recent conditions. Somewhat smaller additional
increases are estimated for the three W126 scenarios in comparison to the current standard
scenario (U.S. EPA 2014a, section 6.6.2 and Appendix 6D).
In addition to the quantitative assessments discussed above, qualitative assessments for
some ecosystem services, were also conducted, such as commercial non-timber forest products
and recreation (U.S. EPA, 2014a, section 6.4), aesthetic and non-use values (U.S. EPA, 2014a,
section 6.4), increased susceptibility to insect attack and fire damage (U.S. EPA, 2014a, sections
5.3 and 5.4, respectively). Other ecological effects that are causally or likely causally associated
with Os exposure such as terrestrial productivity, water cycle, biogeochemical cycle, and
15 See Chapter 6, Section 6.3 of the WREA for a discussion of economic welfare and consumer and
producer surplus.
16 The WREA used the i-Tree model for the urban case studies. i-Tree is a peer-reviewed suite of software
tools provided by USFS.
5-39
-------
community composition (U.S. EPA 2013, Table 9-19) were not directly addressed in the WREA
due to a lack of sufficient quantitative information.
There is substantial heterogeneity in plant responses to Cb, both within species, between
species, and across regions of the U.S. The Os-sensitive tree species are different in the eastern
and western U.S. — the eastern U.S. has far more species. Ozone exposure and risk is somewhat
easier to assess in the eastern U.S. because of the availability of more data and the greater
number of species to analyze. In addition, there are more Os monitors in the eastern U.S. but
fewer national parks (U.S. EPA, 2014a, chapter 8).
• What are the uncertainties associated with both quantitative and qualitative
information?
Several key limitations and uncertainties, which may have a large impact on both overall
confidence and confidence in individual analyses, are discussed here. Despite these
uncertainties, the overall body of scientific evidence underlying the ecological effects and
associated ecosystem services evaluated in the WREA is strong, and the methods used to
quantify associated risks are scientifically sound (Frey, 2014). Key uncertainties associated with
the assessment of impacts on ecosystem services at the national and case-study scales, as well as
across species, U.S. geographic regions and future years include those associated with the
interpolated and adjusted Os concentrations used to estimate W126 exposures in the WREA air
quality scenarios and those associated with the available seedling E-R functions.
The WREA identifies sources of uncertainty for the W126 estimates for each air quality
scenario and qualitatively characterizes the magnitude of uncertainty and potential for directional
bias (U.S. EPA, 2014a, Table 4-5). These sources of uncertainty are described in more detail in
the WREA Chapter 4 and summarized below.
An important large uncertainty in the analyses is the assumed response of the W126
concentrations to emissions reductions needed to meet the existing standard (U.S. EPA, 2014a,
section 8.5.1). We note that any approach to characterizing Os over broad geographic areas based
on concentrations at monitor locations will convey inherent uncertainty. The model-based
adjustments, based on U.S.-wide emissions reductions in oxides of nitrogen (NOx), do not
represent air quality distributions from an optimized control scenario that just meets the current
standard (or target W126 index values for other scenarios), but rather characterize one potential
distribution of air quality across a region when all monitor locations meet the standard (U.S.
EPA 2014a, section 4.3.4.2).17 An additional uncertainty comes from the creation of a national
17 Because our analyses used U.S.-wide NOx emissions reductions to simulate just meeting the existing
standard independently in each region, there are broad regional reductions in O3 even in meeting standards in urban
areas when targeting a few high-O3 urban monitors for reductions. However, the assumption of broad regional or
national NOx reductions are not unreasonable given EPA regulations such as the NOx SIP Call program
5-40
-------
W126 surface using the VNA technique to interpolate recent air quality measurements of Os. In
general, spatial interpolation techniques perform better in areas where the Os monitoring network
is denser. Therefore, the W126 estimated in the rural areas in the West, Northwest, Southwest,
and West North Central with few or no monitors (Figure 2-1) are more uncertain than those
estimated for areas with denser monitoring. Additionally, the surface is created from the three-
year average at the monitor locations, rather than creating a surface for each year and then
averaging across years at each grid cell; the potential impact of this on the resultant estimates is
considered in the WREA (U.S. EPA, 2014a, Appendix 4A).
Because the W126 estimates generated in the air quality analyses are inputs to the
vegetation risk analyses for biomass loss, any uncertainties in the air quality analyses are
propagated into the those analyses (U.S. EPA 2014a, section 8.5). In its letter to the
Administrator following its review of the second draft WREA CAS AC notes that:
"The currently reported finding of only small differences in risk between just meeting the
current standard and a W126-based level of 15 ppm-hrs must not be interpreted to mean
that just meeting the current standard will be as protective as meeting a W126-based
standard at 15 ppm-hrs. There are two key factors that must be considered when making
this comparison. First, air quality was simulated in the Second Draft WREA based on the
magnitude of across-the-board reductions in NOx emissions required to bring the highest
monitor down to the target level. Meeting a target level at the highest monitor requires
substantial reductions below the targeted level through the rest of the region. This
artificial simulation does not represent an actual control strategy and may conflate
differences in control strategies required to meet different standards and different targets.
As a result, there may be a number of monitors that meet the current standard but would
not meet an alternative W126 standard. Second, and equally important, the current form
of the standard is much less biologically relevant for protecting vegetation than is a
seasonal, peak weighted index such as the W126, which was designed to measure the
cumulative effects of ozone exposure." (Frey, 2014, pp. 11-12).
With regard to the robust seedling E-R functions, the description of Figure 5-2 above
provides some characterization of the variability of individual study results and the impact of
that on estimates of W126 index values that might elicit different percentages of biomass loss in
tree seedlings (U.S. EPA, 2014a, section 6.2.1.2). Even though the evidence shows that there
are additional species adversely affected by Os-related biomass loss, the WREA only has E-R
functions available to quantify this loss for 12 tree species. This absence of information only
implemented to help areas meet the 1997 O3 standard resulting in substantial reductions in power plant NOx
emissions from states across the eastern U.S., and the multitude of onroad and offroad mobile source rules that will
lead to reduction in NOx from these sources across the country in future years.
5-41
-------
allows a partial characterization of the Os-related biomass loss impacts in trees associated with
recent Os index values and with just meeting the existing and potential alternative secondary
standards. In addition, there are uncertainties inherent in these E-R functions, including the
extrapolation of relative biomass loss rates from tree seedlings to adult trees and information
regarding within-species variability. The overall confidence in the E-R function varies by
species based on the number of studies available for that species. Some species have low
within-species variability (e.g., many agricultural crops) and high seedling/adult comparability
(e.g., aspen), while other species do not (e.g., black cherry). The uncertainties in the E-R
functions for biomass loss and in the air quality analyses are propagated into the analysis of the
impact of biomass loss on ecosystem services, including provisioning and regulating services
(U.S. EPA, 2014a, Table 6-27). The WREA characterizes the direction of potential influence of
E-R function uncertainty as unknown, yet its magnitude as high, concluding that further studies
are needed to determine how accurately the assessed species reflect the larger suite of Os-
sensitive tree species in the U.S. (U.S. EPA, 2014a, Table 6-27).
Another uncertainty associated with interpretation of the WREA biomass loss-related
estimates concerns the potential for underestimation of compounding of growth effects across
multiple years of varying concentrations. Though tree biomass loss impacts were estimated
using air quality scenarios of 3-year average W126 index values, the WREA also conducted an
analysis to compare the impact of using a variable compounding rate based on yearly variations
in W126 exposures to that of using a W126 index value averaged across three years. The WREA
compared the compounded values for each region, except for the South. In these examples, one
species was chosen that occurred within that region. Air quality values associated with just
meeting the existing standard of 75 ppb were used. Within each region the WREA analysis used
both the W126 index value at each monitor in the region for each year and the three-year average
W126 index value using the method described in Chapter 4. The results show that the use of the
three-year average W126 index value may underestimate RBL values slightly. However, it
should be noted that the approach does not account for moisture levels or other environmental
factors that could affect biomass loss (U.S. EPA, 2014a, section 6.2.1.4 and Figure 6-14). In
considering these results, we note that in these regions and in all three years, the three-year
average W126 index value is sometimes above and sometimes below the individual year W126
index value.
In the national-scale analyses of timber production, agricultural harvesting, and carbon
sequestration, the WREA used the FASOMGHG model, which includes functions for carbon
sequestration, assumptions regarding proxy species, and non-W126 E-R functions for three
crops. However, FASOMGHG does not include agriculture and forestry on public lands,
changes in exports due to Cb into international trade projections, or forest adaptation. Despite
5-42
-------
the inherent limitations and uncertainties, the WREA concludes that the FASOMGHG model
reflects reasonable and appropriate assumptions for a national-scale assessment of changes in
the agricultural and forestry sectors due to changes in vegetation biomass associated with Os
exposure (U.S. EPA, 2014a, sections 6.3, 6.5, 6.6, and 8.5.2).
In the case study analyses of five urban areas, the WREA used the i-Tree model, which
includes an urban tree inventory for each area and species-specific pollution removal and carbon
sequestration functions. However, i-Tree does not account for the potential additional VOC
emissions from tree growth, which could contribute to Os formation. Despite the inherent
limitations and uncertainties, the WREA concludes that the i-Tree model reflects reasonable and
appropriate assumptions for a case study assessment of pollution removal and carbon
sequestration for changes in biomass associated with Os exposure (U.S. EPA, 2014a, sections
6.6.2, 6.7, and 8.5.2).
The overall effect of the combined set of uncertainties on confidence in the interpretation
of the WREA results is difficult to quantify. Due to differences in available information, the
degree to which each analysis was able to incorporate quantitative assessments of uncertainty
differed. Despite these uncertainties, the overall body of scientific evidence underlying the
ecological effects and associated ecosystem services evaluated in the WREA is strong, and the
methods used to quantify associated risks are scientifically sound (Frey, 2014).
5.3 CROP YIELD LOSS
This section considers the current evidence and exposure/risk information to inform
consideration of the adequacy of the protection provided by the current standard from known and
anticipated adverse welfare effects of Os related to crop yield and other associated effects. Crops
warrant consideration from a public welfare perspective because they provide food and fiber
services to humans. This section includes a discussion of the policy-relevant science and weight-
of-evidence conclusions discussed in the ISA (section 5.3.1) and the exposure/risk results
(section 5.3.2) described in the final WREA. Important uncertainties and limitations in the
available information are discussed under the related question below. These discussions
highlight the information we consider relevant to answering the overarching question and
associated policy-relevant questions included in this section.
5.3.1 Evidence-based Considerations
Ozone can interfere with carbon gain (photosynthesis) and allocation of carbon. As a
result of decreased carbohydrate availability, fewer carbohydrates are available for plant growth,
reproduction, and/or yield. For seed-bearing plants, these reproductive effects will culminate in
5-43
-------
reduced seed production or yield. The detrimental effect of Os on crop production has been
recognized since the 1960s, and current Os concentrations in many areas across the U.S. are high
enough to cause yield loss in a variety of agricultural crops including, but not limited to,
soybeans, wheat, cotton, potatoes, watermelons, beans, turnips, onions, lettuces, and tomatoes.
Increases in Os concentration may further decrease yield in these sensitive crops while also
causing yield losses in less sensitive crops (U.S. EPA 2013, section 9.4.4). The ISA concluded
that the evidence is sufficient to determine that there is a causal relationship between Os
exposure and reduced yield and quality of agricultural crops (U.S. EPA 2013, Table 2-2).
• To what extent has scientific information become available that alters or
substantiates our prior conclusions regarding Os-related crop yield loss and of
factors that influence associations between Os levels and crop yield loss?
In general, the vast majority of the new scientific information has substantiated our prior
conclusions regarding Os crop yield loss. On the whole, this evidence supports previous
conclusions that exposure to Cb decreases growth and yield of crops. The ISA describes average
yield loss reported across a number of meta-analytic studies have been published recently for
soybean wheat, rice, semi-natural vegetation, potato, bean and barley (U.S. EPA 2013, section
9.4.4.1). Further, several new exposure studies continue to show decreasing yield and biomass in
a variety of crops with increased Os exposure (U.S. EPA 2013, section 9.4.4.1, Table 9-17).
Research has linked increasing Os concentration to decreased photosynthetic rates and
accelerated aging (U.S. EPA 2013, section 9.4.4) in leaves, which are related to yield. Recent
research has highlighted the effects of Os on crop quality. Increasing Os concentration can also
decrease nutritive quality of grasses and macro- and micro-nutrient concentrations in fruits and
vegetable crops (U.S. EPA 2013, section 9.4.4). The findings of these studies did not change our
understanding of Os-related crop loss since the last review and little information has emerged on
factors that influence associations between Os levels and crop yield loss.
• To what extent have important uncertainties identified in the last review been
reduced and/or new uncertainties emerged?
Important uncertainties have been reduced regarding crop E-R functions, especially for
soybean. In general, the ISA reports consistent results across exposure estimation techniques
and across crop varieties. Figure 5-4 below illustrates the composite E-R functions for the 10
crop species assessed in the WREA (U.S. EPA, 2014a, Figure 6-3).
5-44
-------
CO
0
CD
O
CN
O
O
O
0
10
20 30
W126 (ppm-hrs)
40
50
Figure 5-4. Relative yield loss in crops using the composite functions for 10 studied
species in response to seasonal Os concentrations in terms of seasonal W126 index
values, Y-axis scale for RYL values represents 0% up to 100% (U.S. EPA 2014a, Figure
6-3).
Two important uncertainties have been reduced regarding the E-R functions for yield
effects of Os in crop species, especially for soybean. First, in the last several reviews, the extent
to which E-R functions developed in OTC predicted plant responses in the field and under
different exposure conditions was not clear. In this review, the ISA included an analysis
comparing OTC data with field-based data for one crop and one tree species (U.S. EPA, 2013,
section 9.6.3.2). The crop comparison was done using soybean OTC data from NCLAN and
field-based data from SoyFACE. The NCLAN program, which was undertaken in the early to
mid-1980s, assessed multiple U.S. crops, locations, and Os exposure levels, using consistent
methods, to provide the largest, most uniform database on the effects of Os on agricultural crop
yields (U.S. EPA 1996; U.S. EPA 2006; U.S. EPA 2013, sections 9.2, 9.4, and 9.6, Frey, 2014,
5-45
-------
p. 9).18 The SoyFACE experiment was a chamberless field-based exposure study in Illinois that
was conducted from 2001 - 2009 (U.S. EPA 2013, section 9.2.4). Yield loss in soybean from Os
exposure at the SoyFACE field experiment was reliably predicted by soybean E-R functions
developed in NCLAN (U.S. EPA, 2013, Section 9.6). This analysis supports the robustness and
use of the E-R functions developed in NCLAN to predict relative yield loss to Os exposure in a
realistic agricultural setting.
A second area of uncertainty that was reduced is that regarding the appropriateness of
applying the NCLAN E-R functions to more recent cultivars that are currently being grown.
Because recent studies continue to find yield loss levels in crop species studied previously under
NCLAN that reflect the earlier findings, the ISA concluded that there has been little new
evidence that crops are becoming more tolerant of Os (U.S. EPA, 2006a; U.S. EPA 2013). This
is especially evident in the research on soybean. In a meta-analysis of 53 studies, Morgan et al.
(2003) found consistent deleterious effects of Os exposures on soybean from studies published
between 1973 and 2001. Further, Betzelberger et al. (2010) recently utilized the SoyFACE
facility to compare the impact of elevated Os concentrations across 10 soybean cultivars to
investigate intraspecific variability of the Os response. The E-R functions derived for these 10
current cultivars were similar to the response functions derived from the NCLAN studies
(Heagle, 1989), suggesting there has not been any selection for increased tolerance to Os in more
recent cultivars. The 2013 ISA reported comparisons between yield predictions based on data
from cultivars used in NCLAN studies, and yield data for modern cultivars from SoyFACE (U.S.
EPA, 2013, section 9.6.3). They confirm that the average response of soybean yield to Os
exposure has not changed in current cultivars. Thus, staff concludes that at least for soybean,
uncertainties associated with use of the NCLAN generated E-R functions to estimate biomass
loss in recent cultivars has been reduced.
• To what extent does the available evidence indicate the occurrence of Os-related
effects on crop yield loss attributable to cumulative exposures at lower ambient Os
concentrations than previously established or to exposures at or below the level of
the current standard?
Little scientific evidence has emerged to indicate a lower W126 index value for
cumulative exposures that can affect crop yield than previously established. However, as
18 The NCLAN protocol was designed to produce crop exposure-response data representative of the areas
in the U.S. where the crops were typically grown. In total, 15 species (e.g., corn, soybean, winter wheat, tobacco,
sorghum, cotton, barley, peanuts, dry beans, potato, lettuce, turnip, and hay [alfalfa, clover, and fescue]), accounting
for greater than 85 percent of U.S. agricultural acreage planted at that time, were studied. Of these 15 species, 13
species including 38 different cultivars were combined in 54 cases representing unique combinations of cultivars,
sites, water regimes, and exposure conditions. Crops were grown under typical farm conditions and exposed in
open-top chambers to ambient Os, sub-ambient Os, and above ambient Os.
5-46
-------
discussed below, CASAC has provided a target benchmark protection for crop yield loss that can
help better focus a discussion of the level of exposure that Os related effects on crops can occur
(levels of concern). Currently available evidence supports effects on crop yield at cumulative
exposures at and below the level of the current standard. As described above, the new evidence
has strengthened the basis for using the information from the E-R functions.
Based on the 10 robust E-R functions (i.e., barley, lettuce, field corn, grain sorghum,
peanut, winter wheat, cotton, soybean, potato and kidney bean) described in the ISA and
additionally analyzed in the WREA (Figure 5-4), Table 5C-3 shows that for the CASAC
recommended target benchmark protection level of 5% for median crop relative yield loss
(RYL), W126 index values ranging from 7 to 17 ppm-hrs are protective. However, when
individual species are considered over this same range, the proportion of crops protected varies
from 5/10, 6/10, 7/10, 9/10, 10/10, and 10/10 at the W126 levels of 17, 15, 13, 11, 9, and 7 ppm-
hrs. To the extent a given species is judged as having particular importance to the public
welfare, breaking the information down by species can be helpful. For example, less than 5%
yield loss was estimated for soybeans at the W126 index value of 12 ppm-hrs (U.S. EPA 2014a,
Figure 6-3). Four of the studied crop species (barley, lettuce, field corn, and grain sorghum) are
more tolerant, with RYL under 1% over the W126 range from 7 to 17 ppm-hrs. Peanut also
remained under 4% RYL over the same W126 range. Other species differed regarding the W126
level at which RYL reached or fell below 5%. Specifically, for winter wheat, cotton, soybean,
kidney bean and potato, the relevant W126 index values at which RYLs were below 5% are 15,
13, 11, 11, and 9 ppm-hrs.
Where the current evidence on crop yield loss is not in terms of parts per billion
concentrations over a specific exposure period such as eight hours, assessing whether Os
concentrations associated with meeting the current standard would allow crop yield effects is
more complex. In order to characterize the Os exposures associated with crop yield loss in terms
of seasonal W126 index and to consider the extent to which such index values might be expected
to occur in agricultural locations that meet the current standard, we evaluated two agricultural
counties in Kansas using Os monitoring data from EPA's Air Quality System (AQS) combined
with the E-R function for soybeans. Sedgwick and Sumner counties both met the level of the 3-
year, 8-hour standard of 75 ppb in 2009-2011, but both still had a maximum annual W126 level
of 19 ppm-hours in 2011. At that annual W126 index value, soybean yield loss would be
predicted to be 9% in those counties in that year.
5-47
-------
• To what extent does currently available evidence suggest locations where the
vulnerability of sensitive species, ecosystems and/or their associated services to
Os-related crop yield loss would have special significance to the public welfare?
During the previous NAAQS reviews, there were very few studies that estimated Os
impacts on crop yields at large geographical scales (i.e., regional, national or global). Recent
modeling studies of the historical impact of Os concentrations found that increased Os generally
reduced crop yield, but the impacts varied across regions and crop species (U.S. EPA, 2013,
Section 9.4.4.1). The largest Cb-induced crop yield losses were estimated to occur in high-
production areas exposed to elevated Cb concentrations, such as the Midwest and the Mississippi
Valley regions of the United States. Among crop species, the estimated yield loss for wheat and
soybean were higher than rice and maize. Additionally, satellite and ground-based Cb
measurements have been used to assess yield loss caused by Cb over the continuous tri-state area
of Illinois, Iowa, and Wisconsin. The results indicate that Cb concentrations during the assessed
period reduced soybean yield, which correlates well with the previous results from FACE- and
OTC-type experiments (U.S. EPA 2013, section 9.4.4.1).
Thus, the recent scientific literature in the ISA continues to support the conclusions of the
1996 and 2006 CDs that ambient Cb concentrations can reduce the yield of major commodity
crops in the U.S. and to support the use of crop E-R functions based on OTC experiments.
Agricultural areas that would be likely to have the most significance to the public welfare would
be those high production areas for sensitive crops that also are exposed to high Cb
concentrations, such as areas in the Midwest and Mississippi Valley regions.
5.3.2 Exposure/Risk-based Considerations
Two main analyses are conducted in the WREA to estimate Cb impacts related to crop
yield. Annual yield losses are estimated for 10 commodity crops and these estimates are then
additionally used to estimate Cb impacts on producer and consumer economic surpluses (Table
5-6 below; U.S. EPA, 2014a, sections 6.2, 6.5).
Table 5-6. Exposure, risk and ecosystem services analyses related to crop yield.
WREA estimatesA
Crop-level impact8
Annual Relative Yield Loss for
Corn, Cotton, Potato, Sorghum,
Soybean, Winter Wheat
Agri-Ecosystem Services0
Economic surplus to crop producers
and consumers
A See Chapter 4 WREA; B See Section 6.2 WREA; c See Sections 6.4 and 6.5 WREA
5-48
-------
• For what air quality scenarios were exposures and risks estimated? What
approaches were used to estimate W126 exposures for those conditions? What
are associated limitations and uncertainties?
The WREA crop analyses described here were performed for five air quality scenarios
using the methodology summarized in Table 5-4 above. In general, this methodology is identical
to the air quality scenarios for the biomass loss analyses and have the same uncertainties and
limitations summarized in section 5.2.2 above. These air quality scenarios described in more
detail in the WREA (U.S. EPA, 2014a, chapter 4 and Appendix 4A).
• What is the nature and magnitude of the cumulative exposure- and risk-related
estimates for crop yield loss associated with remaining upon simulating just
meeting the current Os standard? What are the uncertainties associated this
information?
The WREA presents estimates of crop yield loss for the five air quality scenarios
described above using 10 robust E-R functions for commodity crops that are grown across the
U.S. (U.S. EPA, 2014a, section 6.5). The largest reduction in Cb induced crop yield loss occurs
when moving from the recent conditions scenario to that for just meeting the current standard
(U.S. EPA, 2014a, section 6.5). In the analyses for agricultural harvest, the largest estimates of
yield changes also occur when comparing the recent conditions scenario to that for the current
standard. Under recent conditions, the West, Southwest, and Northeast regions generally have
the highest yield losses. For the 3-year average W126 scenarios, relative yield losses for winter
wheat19 are less than one percent. For soybeans, yield losses for these scenarios range from just
above 1 percent to below one percent (U.S. EPA 2014a, section 6.5). However, when evaluated
at the county level, 99% of soybean producing counties (1,718) have greater than 5% yield loss
under recent conditions, while no counties show yield loss at or above this level when air quality
is adjusted to just meet the current standard (U.S. EPA, 2014a, section 6.5).
The WREA estimates of Os-attributable percent yield loss based on 3-year average W126
index values estimated after just meeting the current standard are relatively small (0.0 - 2.72%)
across the 10 crop species analyzed, U.S. EPA 2014a, section 6.5, Appendix 6B). In considering
these estimates, we recognize the significant uncertainties associated with several aspects of the
analyses. Because the W126 estimates generated in the air quality analyses are inputs to the
vegetation risk analyses for crop yield loss, any uncertainties in the air quality analyses are
propagated into the those analyses (U.S. EPA 2014a, Table 6-27, section 8.5).
19 Among the major crops, because winter wheat and soybeans are more sensitive to ambient O3 levels than
other crops we focus on these crops for this discussion.
5-49
-------
• To what extent are the exposures and risks remaining upon simulating just
meeting the current Os standard important from a public welfare perspective?
From a public welfare prospective, the Os attributable risks to crops estimated for
conditions that just meet the current standard are small. However, it is unclear how much weight
to put on these results given the multiple areas of uncertainty associated with these estimates as
discussed in the WREA and summarized above, including those associated with the model-based
adjustment methodology and those associated with projection of yield loss at the estimated Os
concentrations (U.S. EPA, 2014a, Table 6-27, section 8.5). In addition we note that while having
sufficient crop yields is of high public welfare value, important commodity crops are typically
heavily managed to produce optimum yields. Given all of the inputs that go into achieving these
yields, such as fertilizer, herbicides, pesticides, and irrigation, it difficult to determine at what
point Cb-induced yield loss creates an adverse impact for the producer in the way of requiring
increased inputs in order to maintain the desired yields. In contrast, 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 benefit consumers. However, due to
pre-existing market forces and subsidies, it is not clear that such benefits would be realized by
the consumer. Given these competing impacts on producers and consumers, it is unclear how to
determine what type of effect may be adverse to the public welfare. In considering this issue,
CASAC states that "calculation of consumer and producer surpluses is a useful contribution to
quantification of welfare effects. However, this national-level approach does not adequately
account for negative effects on individual farmers and forest owners in high-ozone areas..."
(Frey, 2014, p. 10). Instead, CASAC states that "[a] county scale is appropriate for assessing
crop yield loss. Calculating producer and consumer surpluses at national or large region scales
does not provide adequate protection. Farmers growing sensitive crops in high ozone locations
can be considered a 'sensitive population' for welfare impacts, and crop yields under these
conditions should be protected." (Frey, 2014, pp. 14 - 15). The final WREA includes a county-
level analysis in Appendix 6B finding that 99 percent of soybean-producing counties, for
example, exceed 5% yield loss under recent conditions, while no counties have relative yield
losses above 5% for any crop after adjusting air quality scenarios to just meet the current
standard.
• What are the ecosystem services potentially affected by Os-related crop yield loss
and to what extent are they important from a public welfare perspective? To
what degree can the magnitude of the Os effect on these services be qualitatively
or quantitatively characterized?
The WREA presents national-scale estimates of the effects of biomass loss on timber
production and agricultural harvesting, which supply provisioning services of food and fiber, as
5-50
-------
well as on carbon sequestration (U.S. EPA 2014a, section 6.5). Because the forestry and
agriculture sectors are related and trade-offs occur between the sectors, the WREA also
calculated the resulting market-based welfare effects of Os exposure in the forestry and
agriculture sectors. Overall effect on agricultural yields and producer and consumer surplus
depends on the (1) ability of producers/farmers to substitute other crops that are less Os sensitive,
and (2) responsiveness, or elasticity, of demand and supply (U.S. EPA, 2014a, sections 6.5,
8.2.1.3). Estimated Os-attributable economic welfare impacts on agricultural sectors associated
with air quality conditions adjusted to just meet the existing and potential alternative W126
standard levels were largely similar between the forestry and agricultural sectors. Estimates of
consumer surplus, or consumer gains, were generally higher under those conditions (compared to
recent conditions) in both sectors because higher productivity under lower Os concentrations
increased total yields and reduced market prices (U.S. EPA 2014a, Table 6-18). Because
demand for most forestry and agricultural commodities is not highly responsive to changes in
price, generally producer surplus, or producer gains, decline. For agricultural welfare, annualized
combined consumer and producer surplus gains were estimated to be $2.6 trillion for model-
based adjustment to meet the current standard. Combined gains were essentially unchanged in
comparisons of the current standard scenario to the average W126 scenario for 15 ppm-hrs, but
gains increased by $21 million for the W126 scenario for 11 ppm-hrs and $231 million for the
W126 scenario for 7 ppm-hrs. In some cases, lower prices reduce producer gains more than can
be offset by higher yields (U.S. EPA, 2014a, Table 6-18).
The WREA discusses multiple areas of uncertainty associated with these estimates (also
summarized above), including those associated with the model-based adjustment methodology as
well as those associated with projection of yield loss at the estimated Os concentrations (U.S.
EPA, 2014a, Table 6-27, section 8.5).
5.4 VISIBLE FOLIAR INJURY
Visible foliar injury resulting from exposure to Os has been well characterized and
documented over several decades of research on many tree, shrub, herbaceous, and crop species
(U.S. EPA, 2013, 2006, 1996, 1984, 1978). The significance of Os injury at the leaf and whole
plant levels depends on an array of factors, and therefore, it may be difficult to quantitatively
relate visible foliar injury symptoms to other vegetation effects such as individual tree growth, or
effects at population or ecosystem levels (U.S. EPA, 2013, p. 9-39). Visible foliar injury by
itself, however, is an indication of phytotoxicity due to Os exposure and can impact the public
welfare through damaging or impairing the intended use of the affected entity or the service it
provides. For example, ways by which Os-induced visible foliar injury may impact the public
welfare include: 1) visible damage to ornamental species used in landscaping or leafy crops
5-51
-------
(spinach, lettuce, tobacco) that affects the economic value, yield, or usability of that plant (U.S.
EPA 2007, section 7.4.1; Abt Associates, Inc., 1995); 2) visible damage to plants with special
cultural significance (e.g., those used in tribal practices); 3) visible damage to species occurring
in natural settings valued for their scenic beauty and/or recreational appeal, including in areas
specially designated for more protection (e.g., federal Class I areas) (73 FR 16490). Given
limitations in the available information pertaining to the first two categories,20 the discussions of
the evidence and exposure/risk information in sections 5.4.1 and 5.4.2 below focus primarily on
what is known about visible foliar injury that has been shown to occur in natural settings valued
for their scenic beauty and/or recreational appeal.
At the time of the last review, the following was known:
1) Ozone causes diagnostic visible injury symptoms on studied bioindicator species.
2) Soil moisture is a major confounding effect that can decrease the incidence and
severity of visible foliar injury under dry conditions and vice versa.
3) The most extensive dataset regarding visible foliar injury incidence across the U.S.
was that collected by the USFSFHM/FIA Program.
4) Visible foliar injury incidence was considered to be widespread in both the eastern and
western U.S. based on staff analyses of county level air quality data and USFS
biomonitoring data which showed that for each year within a four year period (2001 -
2004) the percentage of counties having a biosite with visible foliar injury ranged
between 11-30% at an 8-hour average annual level of 0.074 ppm (U.S. EPA, 2007,
section 7.6.3.2).
In the remainder of this section, we consider how the currently available evidence and
exposure/risk information informs our understanding of the relationship that exists between
visible foliar injury and exposures to Os in ambient air and consideration of the adequacy of
protection provided by the current standard. The policy-relevant evidence and weight-of-
evidence conclusions drawn from the ISA are discussed in section 5.4.1, and the exposure/risk
and associated ecosystem services estimates from the WREA, are discussed in section 5.4.2.
Important uncertainties and limitations in each type of available information are also discussed in
these two sections.
20 Qualitative information regarding potential cultural impacts of O3-induced visible foliar injury is noted in
section 5.5 and Appendix 5-A).
5-52
-------
5.4.1 Evidence-based Considerations
• To what extent has scientific information become available that alters or
substantiates our previous conclusions of Os-related visible foliar injury and of
factors that influence associations between Os exposures or concentrations and
visible foliar injury?
Recent research continues to build and substantiate the previous conclusions and findings
drawn from several decades of research on many tree, shrub, herbaceous, and crop species (U.S.
EPA, 2013, 2006, 1996, 1984, 1978) that Os-induced visible foliar injury symptoms are well
characterized and considered diagnostic on certain bioindicator plant species. Diagnostic usage
for these plants has been verified experimentally in exposure-response studies, using exposure
methodologies such as continuous stirred tank reactors (CSTRs), open-top chambers (OTCs),
and free-air fumigation (FACE). Although there remains a lack of robust exposure-response
functions that would allow prediction of visible foliar injury severity and incidence under
varying air quality and environmental conditions, experimental and observational evidence has
clearly established a consistent association of the presence of visible injury symptoms with Os
exposure, with greater exposure often resulting in greater and more prevalent injury (U.S. EPA
2013, section 9.4.2). This new research includes: 1) controlled exposure studies conducted to
test and verify the Os sensitivity and response of potential new bioindicator plant species; 2)
multi-year field surveys in several National Wildlife Refuges (NWR) documenting the presence
of foliar injury in valued areas; 3) ongoing data collection and assessment by the USFS
FHM/FIA program, including multi-year trend analysis (U.S. EPA 2013, section 9.4.2). These
recent studies, in combination with the entire body of available evidence, thus form the basis for
the ISA determinations of a causal relationship between ambient Os exposure and the occurrence
of Os-induced visible foliar injury on sensitive vegetation across the U.S. (U.S. EPA 2013, p. 9-
42).
With regard to evidence from controlled exposure studies, a recent study of 28 plant
species confirmed prior findings of Os causing predictable diagnostic visible foliar injury
symptoms on some species of plants. This study selected 28 plant species, most of which grow
naturally throughout the northeast and midwest US, including in national parks and wilderness
areas, that were suspected of being Os-sensitive, and exposed them to four different Os
concentrations (30, 60, 90, and 120 ppb) in CSTR chambers (Kline et al., 2008). Two
experiments were conducted in each year of the study (2003 and 2004). Plants were exposed to
Os for 7 hours a day, five days a week over the course of each experiment. Specifically, in 2003,
the first experiment lasted from July 14 to August 21 and included 29 days of Os exposure and
the second from September 9 to 30 and included 16 exposure days. In 2004, the first experiment
was conducted from July 13 to August 10 with 21 Os exposure days and the second from August
5-53
-------
27 to September 24, including 21 days of Os exposure. Though the exposures were cumulative
over the course of the study, exposures were reported only in terms of the target exposure
concentration for each experiment. The study reported Os-induced responses in 12, 20, 28 and
28 of the 28 tested species at the 30, 60, 90 and 120 ppb exposure concentrations21, respectively.
Based on their findings, the authors suggest that American sycamore, aromatic sumac, bee balm,
buttonbush, common milkweed, European dwarf elderberry, New England aster, snowberry and
swamp milkweed would make the most useful bioindicator species. Some of these species are
native to Class 1 areas (discussed further below). The staff additionally concludes that given that
the exposure protocol was designed to create a continuous exposure level, not a fluctuating one,
this study shows that Os-induced foliar injury can occur from 7-hour exposures repeated over
multiple days at Os concentrations that are below the 75 ppb level of the current standard.22
While this type of controlled study provides clear evidence of cause and effect, it also has
limitations. The authors, recognizing this cautioned that "extrapolation of these CSTR results to
the field must be done carefully, since CSTR/greenhouse conditions ... are not representative of
natural environmental conditions" (Kline et al., 2008).
A string of recently published multi-year field studies provide a complimentary line of
field-based evidence by documenting the incidence of visible foliar injury symptoms on a variety
of Os-sensitive species over multiple years and across a range of cumulative, seasonal exposure
values in several eastern and midwestern NWRs (U.S. EPA 2013, section 9.4.2.1; Davis and
Orendovici, 2006; Davis, 2007a, b; Davis, 2009). Some of these studies also included
information regarding soil moisture stress using the Palmer Drought Severity Index (PDSI).
While environmental conditions and species varied across the four NWRs, visible foliar injury
was documented to a greater or lesser degree at each site. As discussed further below, visible
foliar injury incidence in these types of areas has greater significance to the public welfare.
• To what extent have important uncertainties identified in the last review been
reduced and/or new uncertainties emerged?
The studies mentioned above also provide additional information regarding an important
uncertainty identified in the previous review, i.e., the role of soil moisture in influencing visible
foliar injury response (U.S. EPA 2013, section 9.4.2). These studies confirm that adequate soil
moisture creates an environment conducive to greater visible foliar injury in the presence of Cb
21 Two of the target exposure levels, 30 and 60 ppb, fall below the level of the current standard (75 ppb).
The mean exposure concentrations achieved in the CTSRs for the 30 ppb target level for each year and study were
27.9, 26.3, 27.1, and 29.3 ppb and for the 60 ppb target level were 56.6, 55.8, 57.9, and 59.0 ppb, for 2003 study 1,
2003 study 2, 2004 study 1, and 2004 study 2, respectively.
22 The current standard is met when the 3-year average of the 4th highest daily maximum 8-hour average
concentrations is at or below 75 ppb.
5-54
-------
than drier conditions. As stated in the ISA, "[a] major modifying factor for Os-induced visible
foliar injury is the amount of soil moisture available to a plant during the year that the visible
foliar injury is being assessed ... because lack of soil moisture generally decreases stomatal
conductance of plants and, therefore, limits the amount of Os entering the leaf that can cause
injury" (U.S. EPA, 2013, p. 9-39). As a result, "many studies have shown that dry periods in
local areas tend to decrease the incidence and severity of Os-induced visible foliar injury;
therefore, the incidence of visible foliar injury is not always higher in years and areas with higher
Os, especially with co-occurring drought (Smith, 2012; Smith et al., 2003)" (U.S. EPA, 2013, p.
9-39). This ".. .partial 'protection' against the effects of Os afforded by drought has been
observed in field experiments (Low et al., 2006) and modeled in computer simulations
(Broadmeadow and Jackson, 2000)" (U.S. EPA, 2013, p. 9-87). In considering the extent of any
protective role of drought conditions, however, the ISA also notes that other studies have shown
that "drought may exacerbate the effects of Os on plants (Pollastrini et al., 2010; Grulke et al.,
2003)" and that "[tjhere is also some evidence that Os can predispose plants to drought stress
(Maier-Maercker, 1998)". Accordingly, the ISA concludes that "the nature of the response is
largely species-specific and will depend to some extent upon the sequence in which the stressors
occur" (U.S. EPA, 2013, p. 9-87). Such uncertainties associated with describing the potential for
foliar injury and its severity or extent of occurrence for any given air quality scenario due to
confounding by soil moisture levels make it difficult to identify an appropriate degree of annual
protection (as well as ambient Os exposure conditions that might be expected to provide that
protection).
• To what extent does the available evidence indicate the occurrence of Os-related
visible foliar injury attributable to cumulative exposures at lower ambient Os
concentrations than previously established or to exposures at or below the level of
the current standard?
Recently available evidence confirms the evidence available in previous reviews that
visible foliar injury can occur when sensitive plants are exposed to elevated Os concentrations in
a predisposing environment (i.e., adequate soil moisture (U.S. EPA, 2013, section 9.4.2). Recent
evidence also continues to indicate the occurrence of visible foliar injury at cumulative ambient
Os exposures previously established. Since the 2006 Os CD, results from several multi-year field
surveys and experimental screenings of Os-induced visible foliar injury on vegetation also show
that visible foliar injury can occur under conditions where the annual 8-hour average Os
concentrations are at or below the level of the current standard, as discussed here. Limited
information exists regarding the incidence of visible foliar injury occurring in areas that have
design values that meet the current 3-year average 8-hour standard.
5-55
-------
To facilitate comparison with other studies reporting foliar injury response to W126
cumulative exposures, we obtained air quality data from the EPA's AQS database for monitors
in each study location and calculated the 12-hr W126 index values and obtained the maximum
4th highest 8-hour average values for a subset of the most recent years included in each study
(Table 5-7). As the shaded rows in Table 5-7 below show, in the years 2002/2003 and 2004 in
the Cape Remain NWR in South Carolina, and the Seney NWR in Michigan, respectively, the 4th
highest daily maximum 8-hour average Os concentrations were at or below the level of the
current standard. We additionally note that the Cape Romain site met the current standard of 75
in every 3-year period during the study and has consistently met the standard from 2001 to
2012.23 Under these air quality conditions, three species (i.e., winged sumac, Chinese tallow
tree, and wild grape) exhibited Os-induced stipple. In 2002, 32% of the examined wild grape
plants, 20% of the winged sumac plants, and 4.6% of the Chinese tallow tree plants, respectively,
were symptomatic (Davis, 2009). At the same time, the 12-hour W126 index value was 20 ppm-
hrs. In 2003, when air quality was somewhat improved, foliar injury declined, with only 13.3%
of wild grape showing Os stipple at a maximum 4th highest 8-hour of 74 ppb and a W126 index
value of 11 ppm-hrs. The PDSI values were 0.27 and 2.45 in 2002 and 2003, respectively.
These values show that 2003 was a wetter year than 2002, though 2002 would have been
considered within the normal soil moisture range.
At the Seney NWR site, by comparison, the annual W126 level was similar in 2004 to
that at Cape Romain in 2003, and the annual 8-hour average level was below that of the current
standard, though the 3-year average design values were above that of the current standard for that
year. Not surprisingly, given the lower Os air quality in 2004, the Seney study reported injury
ranging from about 2% on common milkweed to about 6% on spreading dogbane. Though this
study does not provide the PDSI values, the authors provided some discussion of a possible
relationship stating that "the incidence of ozone injury on spreading dogbane, but not other
species, was weakly, but not significantly, related to the drought index (PDSI)... .However this
relationship was too weak to be used for predictive purposes" (Davis, 2007b). The authors then
conclude that "[nevertheless, the threshold SUM06 ozone level needed to induce stipple on
sensitive plants within the Seney refuge is likely 5000 ppb-hrs under the environmental
conditions of these surveys" (Davis, 2007b). On the basis of the above, the staff concludes that
these studies confirm that visible foliar injury has been shown to occur in the field at W126
index values ranging down to 10 ppm-hrs and provide limited evidence that such foliar injury
23 Design values (concentrations in the form of the standard) for this monitoring site during this period are
presented in the file available at: http://www.epa.gov/airtrends/values.html (US EPA, 2014d).
5-56
-------
can occur in areas with special public welfare significance during periods that meet the current
standard.
Table 5-7. Visible foliar injury incidence in four National Wildlife Refuges.
Name/ Site #/Ref.A
Cape Remain NWR, South Carolina /
450190046 (Davis, 2009)
MoosehornNWR, Maine/ 230090102
(Davis, 2007a)
Seney NWR, Michigan/ 261530001
(Davis, 2007b)
Brigantine NWR, New Jersey /
340010005/
(Davis and Orendovici, 2006)
A Studies (cited above) reported exposure
monitors, calculated exposures in terms o
http://www.epa.sov/ttn/airs/airsaqs/ (US
B Only recent years with available W126
Year8
2002
2003
2002
2003
2004
2002
2003
2004
2001
2002
2003
4th highest daily maximum
8-hour average
0.075 ppm
0.074 ppm
0.1 ppm
0.083 ppm
0.082 ppm
0.083 ppm
0.076 ppm
0.074 ppm
0.095 ppm
0.092 ppm
0.085 ppm
12-hour
W126
20 ppm-hrs
1 1 ppm-hrs
24 ppm-hrs
22 ppm-hrs
14 ppm-hrs
1 1 ppm-hrs
15 ppm-hrs
10 ppm-hrs
39 ppm-hrs
53 ppm-hrs
36 ppm-hrs
% Plants with
visible injury
5-32
3-13
0-17
0-13
3-10
0-13
1-6
2-6
0-45
0-4
0-4
s in terms of SUM06 form. EPA staff, using AQS data for the same
f the current 8-hour and W126 forms:
EPA, 2014b)
data were included in the table.
By far the most extensive field-based dataset of visible foliar injury incidence is that
obtained by of the USFS FHM/FIA biomonitoring network program. A trend analysis of data
from the sites located in the Northeast and North Central U.S. for the 16 year period (1994-2009)
(Smith, 2012) provides additional evidence of foliar injury occurrence in the field as well as
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/soil
moisture scenarios (Smith, 2012; U.S. EPA 2013, section 9.2.4.1). In this study ambient
exposures were expressed in terms of the SUM06 cumulative index coupled with a measure of
the number of peak hourly concentrations above 100 ppb (N100). Soil moisture estimates were
generated using both the PDSI and the plant moisture availability index (MI). Foliar injury was
expressed in terms of the biosite index (BI)24. The authors observed that over this 16-year
24 Biosite index (BI) is the average score (proportion of leaves with injury ["amount"] x mean severity of
symptoms on injured leaves ["severity"]) for each species averaged across all species on the biosite multiplied by
1,000.
5-57
-------
period, "injury indices have fluctuated annually in response to seasonal ozone concentrations and
site moisture conditions. Sites with and without injury occur at all ozone exposures but when
ambient concentrations are relatively low, the percentage of uninjured sites is much greater than
the percentage of injured sites; and regardless of ozone exposure, when drought conditions
prevail, the percentage of uninjured sites is much greater than the percentage of injured sites"
(Smith, 2012). The authors further note that while "both site moisture and ozone exposure play a
role in foliar injury expression ... the interplay among these three factors is unique for each year
and possibly each site. Extreme moisture deficits decrease foliar injury, ... [and] ... [i]n no year
do high ozone exposures override the controlling effect of site moisture, although at the other
end of the scale, injury severity is minimized under conditions of low ozone exposure regardless
of site moisture conditions. This implies a necessary threshold of ozone exposure for injury to
occur...." "In a similar analysis, Rose and Coulston (2009) reported a high percentage of
biosites with injury across the Southern region in 2003, a year when SUM06 values >10 ppm-h
were widespread at the same time that the land area was in moisture surplus or balance." Thus,
Rose and Coulston (2009) also "found evidence that it is the co-occurrence of sufficient moisture
and elevated ozone that determine whether injury occurs to bioindicator plants, not ozone
exposure alone." Regarding the role of peak Os concentrations (>100 ppb Os), Smith (2012)
reported that over the 16-year period concentrations above 100 ppb have declined, and that this
"... may account for the observed decrease in the severity of ozone-induced foliar injury to ozone
sensitive bioindicator plants in eastern forests." They also note that "[fjhere is no compelling
evidence, however, that moderate ozone concentrations, as reflected in seasonal mean SUM06
data, are on the decline" and "[fjhis may explain why injury continues to be detected on many of
the same sites every year" (Smith, 2012). The authors thus conclude that, "[although it is
reasonable to remain concerned about long-term impacts of ozone pollution on our forest
ecosystems, the findings of this biomonitoring survey point to a declining risk of probable
impact on eastern forests over the 16-year period from 1994 to 2009" (Smith, 2012).
In a similar assessment of the USFS FHM/FIA data in the West, six years (2000 to 2005)
of biomonitoring data for Os injury were evaluated for the three coastal states of California,
Oregon and Washington (Campbell et al., 2007; U.S. EPA 2013, section 9.4.2.1). Campbell et
al., 2007 found that "... ozone injury occurs frequently (25 to 37 percent of sampled biosites) in
California forested ecosystems demonstrating that ozone is present at phytotoxic levels." This
study concluded that, "in California, an estimated 1.3 million acres of forest land and 596 million
cubic feet of wood are at moderate to high risk to impacts from ozone. However, [m]ore years of
data are needed to discern any trends" (Campbell et al, 2007). Though this study does not
discuss the role of soil moisture in describing the results, the criteria used to select the
biomonitoring sites include one that considers soil conditions. The best sites are identified as
5-58
-------
those with low drought potential and good fertility. Thus, given the relatively high Os
concentrations that occur in California and the likelihood that many of the biomonitoring sites
occur in areas that have sufficient soil moisture, the high percentage of sampled biosites with
foliar injury is not unexpected.25
These recent studies continue to provide evidence of Os-induced foliar injury occurring in
many areas across the U.S. and augment our understanding of Ch-related visible foliar injury and
of factors that influence associations between Os exposures or concentrations and visible foliar
injury such as soil moisture.
• To what extent does currently available evidence suggest locations where the
vulnerability of sensitive species, ecosystems and/or their associated services to
Os-related visible foliar injury would have special significance to the public
welfare?
As mentioned above, federally designated Class I areas are afforded stringent protections
under the 1977 amendments to the CAA. The CAA gives federal land managers of Class I areas
"the responsibility to protect all air quality related values (AQRVs).. .from deterioration.... In
order to determine if deterioration is occurring, baseline AQRVs must be established" (Davis,
2009). Because of this need and the significance of these areas, studies often focus on these
sites. For example, a study by Kohut (2007) was undertaken to assess the risks of Os-induced
visible foliar injury on Os-sensitive vegetation in 244 parks managed by the NFS (U.S. EPA,
2013, pp. 9-40 to 9-41, U.S. EPA, 2014a, pp. 7-19 to 7-20). Kohut (2007) concluded that the
risk of visible foliar injury was high in 65 parks (27 percent), moderate in 46 parks (19 percent),
and low in 131 parks (54 percent). Thus, while this study suggests that there may be a reason for
concern in as much as 46% of the parks, there were a number of important limitations associated
with this study (described in footnotes 8 and 9 below) that weakened this conclusion. Given the
importance of this kind of assessment, the WREA used Kohut (2007) as the conceptual basis for
the subsequent WREA screening-level assessment, though numerous modifications were made
to the approach to make it applicable to the context of this Os NAAQS review (see section 5.4.2
below).
In addition, as described above, several recently published studies (U.S. EPA 2013,
section 9.4.2.1; Davis and Orendovici, 2006; Davis, 2007a,b; Davis, 2009, Kohut, 2007) were
conducted in federally protected areas including federally designated Class I areas such as
national parks. These studies confirm that visible foliar injury has been observed in these areas
under annual air quality conditions with ambient concentrations at or below the level of the
25 Staff additionally notes that a large proportion of O3 monitoring sites in California did not meet the
current standard during the study period (see: http://www.epa.gov/airtrends/values.html') (US EPA, 2014d).
5-59
-------
current standard and at W126 index values within the CAS AC range recommended in past
reviews. This evidence continues to suggest that Os-sensitive species and their associated
ecosystems and services continue to remain vulnerable to visible foliar injury incidence in areas
that have been afforded special protection by Congress and that have special significance to the
public welfare.
5.4.2 Exposure- and Risk-based Considerations
The WREA presents a number of analyses considering air quality conditions associated
with increased prevalence of visible foliar injury and potential associated welfare impacts (see
Table 5-8 below, U.S. EPA, 2014a, Chapter 7). An initial analysis included the development of
benchmark criteria reflecting different prevalence of visible foliar injury in conjunction with
different W126 exposures and in some cases, soil moisture conditions. These criteria were then
used in a screening-level assessment to characterize potential risk of foliar injury incidence under
2006-2010 conditions in 214 national parks. The last analysis was a case study assessment on
three national parks, which also provides limited characterization of the associated ecosystem
services. Despite the limitations and uncertainties associated with these analyses, and
recognizing that the recent air quality conditions in most cases (prior to any model-based
adjustment) did not meet the current standard, we believe that they help inform our
understanding of the relationship between soil moisture and foliar injury incidence, as well as
provide limited support for our conclusions regarding risk of visible foliar injury incidence under
air quality conditions likely to meet the current standard in areas of special significance to the
public welfare.
5-60
-------
Table 5-8.
injury.
Exposure, risk and ecosystem services analyses related to visible foliar
WREA
estimates
Ecosystem Level Effects
Proportion of FHM/FIA biosites with
foliar injury incidence at various
W126 index values and soil moisture
levels
Percent of 214 national parks
exceeding various W126 benchmarks
derived from FHM/FIA biosite
analysis A
Ecosystem Services
Case study of 3 national parks characterized impacts
using available visitor and use data, including monetary
data for activities and visitor expenditures:
• Utilized Willingness-to-Pay studies for scenic
impairment;
• Assessed the overall cover of sensitive species;
• Compared sensitive species cover to trails and
overlooks; and,
• Estimated percent of park area with Os concentrations
above different W126 index values averaged over
three consecutive years.
A The screening-level assessment of 214 national parks additionally included observations based on the model-
based adjustments to just meet the current standard and targets for the three W126 scenarios (discussed below)
but did not conduct a full analysis using these data.
• For what air quality scenarios were exposures and risks estimated? What
approaches were used to estimate W126 exposures for those conditions? What
are associated limitations and uncertainties?
Three types of foliar injury analyses were performed in the WREA and are considered
below. They include an analysis using USFS FHM/FIA biosite data, a screening-level
assessment in 214 national parks, and case studies of three national parks. The analysis of USFS
biosite data was done using Cb concentrations estimated for a national-scale surface of
concentrations (at a 12 x 12 km2 grid cell resolution in contiguous U.S.) using interpolation
methodology applied to concentrations at Os monitor locations (U.S. EPA, 2014a, section 4.3.2,
Appendix 4A). The analysis of USFS FHM/FIA data and the screening-level analysis using
W126 benchmarks derived from these data used surfaces for each year from 2006 through 2010
(U.S. EPA, 2014a, Appendix 4A, section 4.2). In the National Park case study analyses,
observations related to air quality were made for five air quality scenarios by the methodology
summarized in Table 5-4 above.26
The W126 index values in the individual years from 2006 to 2010 at monitors ranged
from less than 5 ppm-hrs up to above 48 ppm-hrs (U.S. EPA, 2014a, Figure 4-4 and Table 4-3).
26 In general, this methodology involved two steps. The first is derivation of the average W126 value
(across the three years) at each monitor location. This value is based on unadjusted data for recent conditions and
adjusted concentrations for the 4 other scenarios. The development of adjusted concentrations was done for each of
9 regions independently (see U.S. EPA, 2014a, section 4.3.4.1). In the second step, national-scale spatial surfaces
(W126 values for each model grid cell) were created using the monitor-location values and the VNA spatial
interpolation technique (details on the VNA technique are presented in U.S. EPA, 2014a Appendix 4A).
5-61
-------
Concentration estimates varied appreciably across the five years with the median index values
across grid cells ranging from a low of 5.5 ppm-hrs in 2009 up to 11 ppm-hrs in 2006 (U.S. EPA,
2014a, Appendix 4A, section 4.2). During the recent conditions period (2006 through 2008), the
average W126 index values (across the three-year recent conditions period) at the monitor
locations ranged from below 10 ppm-hrs to 48.6 ppm-hrs (U.S. EPA 2014a, Figure 4-4 and
Table 4-3). After adjusting the 2006-2008 data to just meet the current standard in each region,
and subsequent application of the VNA technique to the current standard scenario monitor
location values, the average W126 index values were below 15 ppm-hrs across the national
surface with the exception of a very small area of the southwest region (near Phoenix) where the
average W126 index values was near or just above 15 ppm-hrs. A lowering of the highest values
occurred with application of the interpolation method as a result of estimating W126 index
values at a 12 x 12 km2 grid resolution rather that at the exact location of a monitor. This
indicates one uncertainty associated with this aspect of the approach to estimating W126 index
values for the adjusted air quality just meeting the current standard. Other uncertainties are
summarized in section 5.2.2 above.
• What are the nature and magnitude of the cumulative exposure- and risk-related
estimates for visible foliar injury under recent conditions or conditions meeting
the current Os standard?
As an initial matter, we consider the analysis of the biomonitoring site data from the
USFS FHM/FIA Network, described in section 7.2 of the WREA.27 Using this dataset and
associated data for soil moisture during the sample years along with ambient air Os
concentrations based on monitoring data from 2006 to 2010 and spatial interpolation
methodology (as described above), the proportion of biosites with any foliar injury are observed
to increase with increasing annual W126 index values up to specific values after which there is
little change in proportion of affected biosites with higher W126 index values (see Figure 5-5
below; U.S. EPA, 2014a, section 7.2, Figure 7-10). The proportion of biosites metric is derived
by first ordering the data (across biosites and sample years) by W126 index value estimated for
that biosite and year. Then for each W126 index value the proportion of biosites exceeding the
selected biosite index value for all observations at or below that W126 index value is calculated.
The WREA repeated this using a biosite index value greater than zero, indicating presence of any
foliar injury (USFS, 2011).
When looking only at presence or absence of foliar injury ("any injury") with the
exception of 2008, the proportion of biosites across all W126 index values with foliar injury
27 Data were not available for several western states (Montana, Idaho, Wyoming, Nevada, Utah, Colorado,
Arizona, New Mexico, Oklahoma, and portions of Texas).
5-62
-------
exceeds 15 percent; in 2006, it exceeds 20 percent, while in 2008 the proportion of biosites with
foliar injury across all W126 index values was just below 15 percent (U.S. EPA, 2014a, section
7.2.3, Figure 7-9). When categorized by moisture levels, the data demonstrate a distinct pattern.
In general, the WREA concludes that the results of these foliar injury analyses demonstrate a
similar pattern -the proportion of biosites showing the presence of any foliar injury (biosite
index >0) increases from zero to about 20% (Figure 5-5 below). This increase occurs with
increasing W126 index values up to approximately 10 ppm-hrs for any foliar injury (biosite
index >0), with little change in proportion of biosites with any injury at higher W126 index
values. The data for biosites during normal moisture years are very similar to the dataset as a
whole, with an overall proportion of close to 18 percent for presence of any foliar injury. Among
the biosites with a relatively wet season (average Palmer Z => 1), the proportion of biosites
showing injury is much higher and the relationship with annual W126 index value is much
steeper. Much lower proportions of biosites are reached for the any injury category at biosites
with relatively dry seasons (average Palmer Z < -1.24), potentially indicating that drought may
provide some protection from foliar injury as discussed in the ISA (U.S. EPA, 2014a, section
7.2.3, Figures 7-10). This information provides insight into the relationship between soil
moisture and foliar injury and the issue of whether drought provides protection from foliar
injury. Thus, there is relatively little change in the proportion of biosites beyond a W126 index
value of 10 ppm-hrs. There are two important observations that can be made from these analyses:
(1) the proportion of biosites exhibiting foliar injury rises rapidly at increasing W126 index
values below approximately 10 ppm-hrs, and (2) there is relatively little change in the
proportions above W126 index values of approximately 10 ppm-hrs.
5-63
-------
Biosite Index > 0
2
Q_
DO 0°
10
20
W126 (ppm-hrs)
30
40
Figure 5-5. Cumulative proportion of biosites with any foliar injury present, by moisture
category (U.S. EPA 2014a, Figure 7-10).
We additionally consider the WREA screening-level assessment in 214 parks in the
contiguous U.S. that employed benchmark criteria developed from the above analysis (Table 5-
9).28,29 For example, annual Os concentrations corresponding to a W126 index value of 10.46
ppm-hrs represents the Os exposure concentration where the slope of exposure-response
relationship changes for FHM biosites. The WREA refers to this as the "base scenario"
benchmark. Above this index value, the percentage of FFDVI biosites showing foliar injury
remains relatively constant. The W126 benchmarks across the five scenarios range from 3.05
ppm-hrs (five percent of biosites, normal moisture, any injury) up to 24.61 ppm-hrs (10% of
biosites, dry, any injury). For the scenario of 10% biosites with injury, W126 index values were
approximately 4, 6, and 25 ppm-hours for wet, normal and dry years, respectively. The national-
28 The parks assessed here include lands managed by the NFS in the continental U.S., which includes
National Parks, Monuments, Seashores, Scenic Rivers, Historic Parks, Battlefields, Reservations, Recreation Areas,
Memorials, Parkways, Military Parks, Preserves, and Scenic Trails.
29 The WREA applied different foliar injury benchmarks in this assessment after further investigation into
the benchmarks applied in Kohut (2007), which were derived from biomass loss rather than visible foliar injury.
Kohut cited a threshold of 5.9 ppm-hrs for highly sensitive species from Lefohn et al. (1997), which was based on
the lowest W126 estimate corresponding to a 10% growth loss for black cherry. For soil moisture, Kohut (2007)
qualitatively assessed whether there appeared to be an inverse relationship between soil moisture and high O3
exposure.
5-64
-------
scale screening-level assessment includes 42 parks with Os monitors and a total of 214 parks
with Os exposure estimated from the interpolated Os surface for individual years from 2006 to
2010 (U.S. EPA, 2014a, Appendix 7A). These data were combined with lists from the NFS of
the parks containing Os-sensitive vegetation species (NFS, 2003, 2006). Based on NFS lists, 95
percent of the parks in this assessment contain at least one Cb-sensitive species. This analysis for
recent air quality conditions, estimates that 58 percent of parks exceeded the benchmark criteria
corresponding to the base scenario (W126>10.46 ppm-hrs, 17.7 percent of biosites, all moisture
categories, any injury) for at least three years in the period from 2006 to 2010 (U.S. EPA, 2014a,
section 7.3.2). Based on model-based adjustments to meet the current standard, none of the 214
parks have average W126 index values that would exceed the annual benchmark criteria for the
base scenario (W126 >10.46 ppm-hrs) (U.S. EPA, 2014a, section 7.3.3.3).
Table 5-9. Benchmark criteria for Os exposure and relative soil moisture used in
screening-level assessment of parks (from U.S. EPA 2014a, Table 7-6).
Scenario
Base
5% of
biosites
10% of
biosites
15% of
biosites
20% of
biosites
Description
17.7% of all FHM biosites showed any
injury (higher W126 index values have a
relatively constant percentage of FHM
biosites showing injury)
5% of FHM biosites showed any injury,
reflects soil moisture categorization
10% of FHM biosites showed any injury,
reflects soil moisture categorization
15% of FHM biosites showed any injury,
reflects soil moisture categorization
20% of FHM biosites showed any injury,
reflects soil moisture categorization
W126 Benchmark (in ppm-hrs)
Wet
(Palmer Z>1)
Normal Moisture
(Palmer Z between
-1.25 and 1)
Dry
(Palmer Z < -
1.25)
10.46
(soil moisture not considered)
3.76
4.42
4.69
5.65
3.05
5.94
8.18
N/A
6.16
24.61
N/A
N/A
Lastly, we consider the WREA case study analysis that focused on characterizing the
ecosystem services potentially associated with visible foliar injury in three specific national
parks (case study assessment). The parks included were Great Smoky Mountains National Park
(GRSM), Rocky Mountain National Park (ROMO), and Sequoia/Kings National Parks (SEKI).
For each park, the potential impact of Os-related foliar injury on recreation (cultural services)
was considered in light of information on visitation patterns, recreational activities and visitor
expenditures. For example, visitor spending in 2011 exceeded $800 million, $170 million and
5-65
-------
$97 million dollars in GRSM, ROMO and SEKI, respectively. This assessment also included
percent cover of species sensitive to foliar injury and focused on the overlap between recreation
areas within the park and elevated W126 concentrations. Ozone concentrations in GRSM have
been among the highest in the eastern U.S. In the recent conditions scenario, the grid cells
representing 44 percent of GRSM had three year average W126 index value above 15 ppm-hrs.
After adjustments to just meet the current standard, no grid cell had a three-year average W126
index value above 7 ppm-hrs. In the recent conditions scenario for ROMO, three-year average
W126 index values for all grid cells were above 15 ppm-hrs. In the current standard scenario,
values for 59 percent of the park were below 7 ppm-hrs. For SEKI, three-year average W126
index values for all grid cells were above 15 ppm-hrs in the recent conditions scenario, but
dropped below 7 ppm-hrs for the current standard scenario (U.S. EPA, 2014a, section 7.4).
In summary, these analyses indicate that Os concentrations in U.S. national parks in
recent years correspond to W126 index values at which some foliar injury may occur, with
variation associated with relative soil moisture conditions. None of the 214 parks assessed are
estimated to exceed the annual benchmark criteria for the base scenario (W126 >10.46 ppm-hrs)
after adjusting air quality to meet the current standard. Although adjusted scenarios to just meet
the current standard indicate substantial reductions in three-year average W126 index values
estimated by the VNA approach, some individual year values may range higher. The case study
analysis of three parks indicates the potential for appreciable ecosystem services impact
associated with foliar injury. While these analyses indicate the potential for foliar injury to occur
under conditions that meet the current standard, the extent of foliar injury that might be expected
under such conditions is unclear from these analyses.
• To what extent are the exposures and risks remaining upon simulating just
meeting the current Os standard important from a public welfare perspective?
The screening level assessment, as described above, indicates that risk of visible foliar
injury is likely to be lower in most national parks after simulating just meeting the current
standard. Based on the national-scale analysis, visible foliar injury would likely continue to
occur at lower Os exposures, including some sensitive species growing in areas (e.g., National
Parks and other Class I areas) that may provide important cultural ecosystem services to the
public. Staff notes that such occurrences might reasonably be considered to have some
importance from a public welfare perspective, as discussed in section 5.1 above.
5-66
-------
• What are the ecosystem services potentially affected by visible foliar injury, to
what degree can the magnitude of these effects be qualitatively or quantitatively
characterized, and to what extent are they important from a public welfare
perspective?
The ecosystem services most likely to be affected by Os-induced foliar injury are cultural
services, including aesthetic value and outdoor recreation. Aesthetic value and outdoor
recreation depend on the perceived scenic beauty of the environment. Many outdoor recreation
activities directly depend on the scenic value of the area, in particular scenic viewing, wildlife-
watching, hiking, and camping. These activities and services are of significant importance to
public welfare as they are enjoyed by millions of Americans every year and generate millions of
dollars in economic value (U.S. EPA, 2014a, Chapter 5, Chapter 7). These aesthetic values are at
risk of impairment because of Os-induced damage directly due to foliar injury. Other ecosystem
services that have also been found to be associated with Os-sensitive plants include those that fall
under the categories of provisioning. For example, several tribes have indicated that many of the
known confirmed Os-sensitive species (including bioindicator species) are culturally significant
(see Appendix 5-A). Although data are not available to explicitly quantify these negative effects
on ecosystem services, several qualitative analyses conducted in the WREA are summarized
below.
To assess the effects of visible foliar injury on recreation, the WREA reviewed the
NSRE, as well as the 2006 National Survey of Fishing, Hunting, and Wildlife-Associated
Recreation (FHWAR) and a 2006 analysis done for the Outdoor Industry Foundation (OIF).
According to the NSRE, some of the most popular outdoor activities are walking, including day
hiking and backpacking; camping; bird watching; wildlife watching; and nature viewing.
Participant satisfaction with these activities can depend on the quality of the natural scenery,
which can be adversely affected by Os-related visible foliar injury. According to the FHWAR
and the OIF reports, the total expenditures across wildlife watching activities, trail-based
activities, and camp-based activities are approximately $230 billion dollars annually. While the
WREA could not quantify the magnitude of the impacts of Os damage to the scenic beauty and
outdoor recreation, the existing losses associated with current Os-related foliar injury are
reflected in reduced outdoor recreation expenditures (U.S. EPA, 2014a, section 7.1).
The WREA also assessed Os impacts on cultural ecosystem services related to foliar
injury at three national parks - Great Smoky Mountains National Park, Rocky Mountain
National Park, and Sequoia/Kings National Parks - by considering information on visitation
patterns, recreational activities and visitor expenditures. The analysis included percent cover of
species sensitive to foliar injury and focused on the overlap between recreation areas within the
park and elevated W126 concentrations. All three of these park units are in areas that are known
5-67
-------
to have high ambient Os concentrations, have vegetation maps, and have species that are
considered Os-sensitive. Using GIS, the NFS vegetation maps were compared to the national Os
surface to illustrate where foliar injury may be occurring, particularly with respect to park
amenities such as trails (U.S. EPA, 2014a, section 7.4).
Great Smoky Mountains National Park is prized, in part, for its rich species diversity.
The large mix of species includes 37 Ch-sensitive species and many areas contain several
sensitive species. With 3.8 million hikers using the trails every year and hikers' WTP over $266
million for that activity, even a small benefit of reducing Os damage in the park could result in a
significant economic value. Ozone concentrations in Great Smoky Mountains National Park
have been among the highest in the eastern U.S. - at times twice as high as neighboring cities
such as Atlanta (U.S. EPA, 2014a, p. 7-52). Unlike Great Smoky Mountains National Park,
sensitive species cover in Rocky Mountain National Park is driven by a few Cb-sensitive species
(7 species) and most notably by Quaking Aspen. This is significant in that many of the visitors
to Rocky Mountain National Park visit specifically to see this tree in its fall foliage. Given 1.5
million hikers in Rocky Mountain National Park and their $70 million WTP for the hiking
experience, even a small improvement in the scenic value could be economically significant
(U.S. EPA, 2014a, section 7.4.2, p. 7-56). Sequoia/Kings National Parks is home to 12 identified
sensitive species. Again, although the EPA is not able to quantify the impact of this scenic
damage on hiker satisfaction for hikers in Sequoia/Kings National Parks and their $26 million
WTP for the experience, even a small improvement in the scenic value could be economically
significant ((U.S. EPA, 2014a, section 7.4.3, p. 7-63).
• What are the uncertainties associated with this information and what is the level
of confidence associated with those estimates?
Uncertainties associated with these analyses are discussed in the WREA, sections 7.5 and
8.5.3, and in WREA Table 7-24. As discussed in the WREA (section 8.5.3), evaluating soil
moisture is more subjective than evaluating Os exposure because of its high spatial and temporal
variability within the Os season, and there is considerable subjectivity in the categorization of
relative drought. The WREA generally concludes that the spatial and temporal resolution for the
soil moisture data is likely to underestimate the potential of foliar injury that could occur in some
areas. In addition, there is lack of a clear threshold for drought below which visible foliar injury
would not occur. In general, low soil moisture reduces the potential for foliar injury, but injury
could still occur, and the degree of drought necessary to reduce potential injury is not clear. Due
to the absence of biosite injury data in the Southwest region and limited biosite data in the West
and West North Central regions, the benchmarks applied may not be applicable to these regions.
The WREA applied the benchmarks from the national-scale analysis to a screening-level analysis
5-68
-------
of 214 national parks and case studies of three national parks. Therefore, uncertainties in the
foliar injury benchmarks are propagated into the national park analyses.
There are also important uncertainties in the estimated Os concentrations for the different
air quality scenarios evaluated (U.S. EPA, 2014a, section 8.5), as discussed earlier in this
section. These uncertainties only apply to the national park case studies because these are the
only foliar injury analyses that rely on the air quality scenarios, but any uncertainties in the air
quality analyses are propagated into the those analyses. Additional uncertainties are associated
with the national park case studies. Specifically, there is uncertainty inherent in survey estimates
of participation rates, visitor spending/economic impacts, and willingness-to-pay. These surveys
potentially double-count impacts based on the allocation of expenditures across activities but
also potentially exclude other activities with economic value. In general, the national level
surveys apply standard approaches, which minimize potential bias. Other sources of uncertainty
are associated with the mapping, including park boundaries, vegetation species cover, and park
amenities, such as scenic overlooks and trails. In general, the WREA concludes that there is
high confidence in the park mapping (U.S. EPA, 2014, Table 7-24).
5.5 OTHER WELFARE EFFECTS
In addition to the welfare effects discussed in the previous sections, there is evidence of
other Os effects, such as those related to climate impacts that we consider here. In this section,
the WREA national-scale analyses of the effects of insect damage to forests related to elevated
Os exposures are considered in section 5.5.1, and a case study-scale characterization of the effect
community composition changes on forest susceptibility and fire regulation in California is
considered in section 5.5.2. As above, these sections, where possible, consider the WREA
information regarding risk remaining under adjusted conditions just meeting the current standard
and associated uncertainties (U.S. EPA 2014a, section 8.5). Chapters 5, 6, and 7 of the WREA
also qualitatively assessed additional ecosystem services, including regulating services such as
hydrologic cycle and pollination; provisioning services such as commercial non-timber forest
products; and cultural services with aesthetic and non-use values. The information associated
with these latter effects is insufficient to inform the target protection of the standard. The effects
of Os on climate are also considered in section 5.5.3 below, drawing primarily on the evidence
presented in the ISA (U.S. EPA 2013, chapter 10).
5.5.1 Forest Susceptibility to Insect Infestation
Ozone in ambient air can contribute to increased susceptibility of some forests to
infestation by some chewing insects, including the southern pine beetle and western bark beetle
(U.S. EPA 2013, chapter 9; U.S. EPA 2014a, sections 5.3.3 and 5.4). These infestations can
5-69
-------
cause economically significant damage to tree stands and the associated timber production. The
WREA described the potential impacts of this effect on timber markets (U.S. EPA 2014a, section
5.4). In the short-term, the immediate increase in timber supply that results from the additional
harvesting of damaged timber depresses prices for timber and benefits consumers. In the longer-
term, the decrease in timber available for harvest raises timber prices, potentially benefitting
producers. The USFS reports timber producers have incurred losses of about $1.4 billion
(2010$), and wood-using firms have gained about $966 million, due to beetle outbreaks between
1977 and 2004. It is not possible to attribute a portion of these impacts resulting from the effect
of Os on trees' susceptibility to insect attack; however, the losses are embedded in the estimates
cited and any welfare gains from decreased Os would positively impact the net economic impact.
However, it is important to note that CASAC clarified that spatial association is not causation,
even though expert opinion relates Os-exposure to bark beetle infestation (Frey, 2014, p. 12).
To provide some quantitative estimates related to insect infestation-related risks, the
WREA reported the estimates of 3-year average W126 index values in areas estimated to be at
risk of greater than 25% timber loss (high loss) due to pine beetle infestation. This was done for
all six WREA air quality scenarios. For example, for the recent conditions scenario,
approximately 57 percent of the at-risk area has W126 estimates above 15 ppm-hrs, with the
percentage dropping to approximately five percent in the current standard scenario (U.S. EPA
2014a, section 5.4).
5.5.2 Fire Regulation
Evidence indicates that fire regime regulation may also be negatively affected by Os
exposure (U.S. EPA 2013, chapter 9; U.S. EPA 2014a, section 5.3.3). For example, Grulke et al.
(2008) reported various lines of evidence indicating that Os exposure may contribute to southern
California forest susceptibility to wildfires by increasing leaf turnover rates and litter, increasing
fuel loads on the forest floor. According to the National Interagency Fire Center, in the U.S. in
2010 over 3 million acres burned in wildland fires and an additional 2 million acres were burned
in prescribed fires. From 2004 to 2008, Southern California alone experienced, on average, over
4,000 fires per year burning, on average, over 400,000 acres per fire. The California Department
of Forestry and Fire Protection (CAL FIRE) estimated that losses to homes due to wildfire were
over $250 million in 2007 (CAL FIRE, 2008). In 2008, CAL FIRE's costs for fire suppression
activities were nearly $300 million (CAL FIRE, 2008).
The WREA developed maps that overlay the mixed conifer forest area of California with
areas of moderate or high fire risk defined by CAL FIRE and with recent W126 concentrations
5-70
-------
and surfaces adjusted to just meet existing and alternative standards. The highest fire risk and
highest Os concentrations overlap with each other, as well as with significant portions of mixed
conifer forest. In the recent concentrations scenario, over 97 percent of mixed conifer forest area
has average W126 index values over 7 ppm-hrs with a moderate to severe fire risk, and 74
percent has average W126 index values over 15 ppm-hrs with a moderate to severe fire risk. The
scenario for air quality adjusted to just meet the current standard, almost all of the mixed conifer
forest area with a moderate to high fire risk shows a reduction in Os to below a W126 index
value of 7 ppm-hrs (average across three years of scenario). In the scenario for an average W126
index value of 15 ppm-hrs, all but 0.18 percent of the area has average index values below 7
ppm-hrs, and for the W126 scenarios of 11 and 7 ppm-hrs, all of the moderate to high fire threat
area has estimated average W126 index values below 7 ppm-hrs (U.S. EPA 2014a, section 5.3.3,
Figure 5-3). However, it is important to note that CASAC clarified that spatial association is not
causation, but expert opinion relates Os-exposure to fire risk (Frey, 2014, p. 12).
5.5.3 Ozone Effects on Climate
Tropospheric Os is a major greenhouse gas, third in importance after carbon dioxide
(CCh) and methane (CFLt). While the developed world has successfully reduced emissions of Os
precursors in recent decades, many developing countries have experienced large increases in
precursor emissions and these trends are expected to continue, at least in the near term (U.S.
EPA 2013, section 10.3.6.2). Projections of radiative forcing due to changing Os over the 21st
century show wide variation, due in large part to the uncertainty of future emissions of source
gases (U.S. EPA 2013, section 10.3.6.2). In the near-term (2000-2030), projections of Os
radiative forcing range from near zero to +0.3 W/m2, depending on the emissions scenario (U.S.
EPA 2013, section 10.3.6.2; Stevenson et al., 2006). Reduction of tropospheric Os
concentrations could therefore provide an important means to slow climate change in addition to
the added benefit of improving surface air quality (U.S. EPA, 2013, section 10.5).
It is clear that increases in tropospheric Os lead to warming. However the precursors of
Os also have competing effects on the greenhouse gas CH4, complicating emissions reduction
strategies. A decrease in CO or VOC emissions would enhance OH concentrations, shortening
the lifetime of CH4, while a decrease in NOx emissions could depress OH concentrations in
certain regions and lengthen the CH4lifetime (U.S. EPA, 2013, section 10.5).
Abatement of CH4 emissions would likely provide the most straightforward means to
address Os-related climate change since CH4 is itself an important precursor of background Os
(West et al., 2007; West et al., 2006; Fiore et al., 2002). A set of global abatement measures
identified by West and Fiore (2005) could reduce CH4 emissions by 10% at a cost savings,
decrease background Os by about 1 ppb in the Northern Hemisphere summer, and lead to a
5-71
-------
global net cooling of 0.12 W/m2. West et al. (2007) explored further the benefits of CH4
abatement, finding that a 20% reduction in global CH4 emissions would lead to greater cooling
per unit reduction in surface Os, compared to 20% reductions in VOCs or CO (U.S. EPA, 2013,
section 10.5).
Important uncertainties remain regarding the effect of tropospheric Os on future climate
change. To address these uncertainties, further research is needed to: (1) improve knowledge of
the natural atmosphere; (2) interpret observed trends of Os in the free troposphere and remote
regions; (3) improve understanding of the CH4 budget, especially emissions from wetlands and
agricultural sources, (4) understand the relationship between regional Os radiative forcing and
regional climate change; and (5) determine the optimal mix of emissions reductions that would
act to limit future climate change (U.S. EPA, 2013, section 10.5).
The IPCC has estimated the effect of the tropospheric Os change since preindustrial times
on climate to be about 25-40% of the anthropogenic CCh effect and about 75% of the
anthropogenic CH4 effect (IPCC, 2007). There are large uncertainties in the radiative forcing
estimate attributed to tropospheric Os, making the effect of tropospheric Os on climate more
uncertain than the effect of the long-lived greenhouse gases (U.S. EPA, 2013, section 10.5).
Radiative forcing does not take into account the climate feedbacks that could amplify or
dampen the actual surface temperature response. Quantifying the change in surface temperature
requires a complex climate simulation in which all important feedbacks and interactions are
accounted for. As these processes are not well understood or easily modeled, the surface
temperature response to a given radiative forcing is highly uncertain and can vary greatly among
models and from region to region within the same model (U.S. EPA, 2013, section 10.5).
As discussed in section 5.2 above, Os exposure is associated with reduced forest tree
growth, productivity, and carbon storage. Therefore, reducing Os exposure would potentially
increase carbon storage in Os-sensitive trees, which could also have climate effects.
5.5.4 Additional Effects
Recent information available since the last review considers the effects of Os on chemical
signaling in insect and wildlife interactions. Specifically, studies on Os effects on pollination and
seed dispersal, defenses against herbivory and predator-prey interactions all consider the ability
of Os to alter the chemical signature of VOCs emitted during these pheromone-mediated events.
The effects of Os on chemical signaling between plants, herbivores and pollinators as well as
interactions between multiple trophic levels is an emerging area of study that may result in
further elucidation of Os effects at the species, community and ecosystem-level (U.S. EPA, 2013,
p. 9-98).
5-72
-------
5.6 CASAC ADVICE
This section discusses CASAC advice regarding the adequacy of the existing secondary
standard with respect to the 2008 review, the 2010 reconsideration of the 2008 review, and most
recently its advice in this review, initiated in September 2008, in its letter to the Administrator on
the second draft WREA and PA. To give an overview, following the 2008 decision to revise the
secondary standard by setting it identical to the revised primary standard, CASAC conveyed
additional advice to the Administrator regarding that decision. Shortly after that, several
petitioners filed suit challenging the decision and in September 2009, the EPA announced its
intention to reconsider the 2008 standards, issuing a notice of proposed rulemaking in January
2010 (75 FR 2938). Soon after, the EPA solicited CASAC review of that proposed rule and in
January 2011 solicited additional advice. This proposal was based on the scientific and technical
record from the 2008 rulemaking, including public comments and CASAC advice and
recommendations. As further described in section 1.2.2 above, the EPA in the fall of 2011 did
not promulgate final rulemaking in that process but decided to coordinate further proceedings on
the reconsideration rulemaking with this ongoing periodic review.
More specifically, in April 2008, the members of the CASAC Ozone Review Panel sent a
letter to EPA stating that "[i]n our most-recent letters to you on this subject - dated October
2006 and March 2007 - ... the Committee recommended an alternative secondary standard of
cumulative form that is substantially different from the primary Ozone NAAQS in averaging
time, level and form — specifically, the W126 index within the range of 7 to 15 ppm-hours,
accumulated over at least the 12 'daylight' hours and the three maximum ozone months of the
summer growing season" (Henderson, 2008). The letter continued:
The CASAC now wishes to convey, by means of this letter, its additional,
unsolicited advice with regard to the primary and secondary Ozone NAAQS. In
doing so, the participating members of the CASAC Ozone Review Panel are
unanimous in strongly urging you or your successor as EPA Administrator to
ensure that these recommendations be considered during the next review cycle for
the Ozone NAAQS that will begin next year ... The CASAC was also greatly
disappointed that you failed to change the form of the secondary standard to
make it different from the primary standard. As stated in the preamble to the
Final Rule, even in the previous 1996 ozone review, "there was general
agreement between the EPA staff, CASAC, and the Administrator, ... that a
cumulative, seasonal form was more biologically relevant than the previous 1-
hour and new 8-hour average forms (61 FR 65716) "for the secondary
standard. Unfortunately, this scientifically-sound approach of using a
cumulative exposure index for welfare effects was not adopted...
5-73
-------
In response to the EPA's solicitation of their advice on the Agency's proposed
rulemaking as part of the reconsideration, CAS AC conveyed their support for the proposed
approach as follows (Samet, 2010).
CASAC also supports EPA 's secondary ozone standard as proposed: a new
cumulative, seasonal standard expressed as an annual index of the sum of
weighted hourly concentrations (i.e., the W 126 form), cumulated over 12 hours
per day (Sam to 8pm) during the consecutive 3-month period within the ozone
season with the maximum index value, set as a level within the range of 7 to [1]5
ppm-hours. This W126 metric can be supported as an appropriate option for
relating ozone exposure to vegetation responses, such as visible foliar injury and
reductions in plant growth. We found the Agency's reasoning, as stated in the
Federal Register notice of January 19, 2010, to be supported by the extensive
scientific evidence considered in the last review cycle. In choosing the Wl26 form
for the secondary standard, the Agency acknowledges the distinction between the
effects of acute exposures to ozone on human health and the effects of chronic
ozone exposures on welfare, namely that vegetation effects are more dependent on
the cumulative exposure to, and uptake of, ozone over the course of the entire
growing season (defined to be a minimum of at least three months).
In its advice offered early in the current review, based on the updated scientific and
technical record since the 2008 rulemaking, CASAC indicated that a conclusion that the current
standard is inadequate to protect vegetation and ecosystems is "warranted" although it stated that
the foundation needs to be broadened beyond analysis focused on Class I areas and trees to
include "effects on sensitive crops, trees in regions outside of Class I areas, and additional
ecosystem impacts" (Frey and Samet, 2012, p. 2). The Panel additionally endorsed the first draft
PA discussions and conclusions on biologically relevant exposure metrics, stating that "the focus
on the W126 form is appropriate" and that "there is a strong justification made for using a
cumulative and weighted exposure standard for welfare effects (i.e., the W126)..." (Frey and
Samet, 2012, p. 2).
In its letter dated June 26, 2014, CASAC again concluded that "the current secondary
standard is not adequate to protect against current and anticipated welfare effects of ozone on
vegetation..." (Frey, 2014, p. iii) and that "the form of the standard should be changed from the
current 8-hour form to the cumulative W126 index and... that the discussion provides an
appropriate and sufficient rationale" (Frey, 2014, p. 12). CASAC then further states that
"[t]hus, based on identification of known or anticipated ozone effects that are
adverse to public welfare, taking into account the weight of evidence for causality of
exposure to ozone and adverse welfare effects as given in Table 2-4 of the Integrated
Science Assessment; results of the Second Draft WREA with regard to assessment of relative
5-74
-------
biomass loss for tree species, foliar injury, and crop yield loss; and the breadth of adverse
welfare effects for ecosystem services, foliar injury, and crop loss, the CASAC recommends
that the secondary standard for ozone be revised as follows: (1) ozone should be the
indicator; (2) the form and summation time of the standard should be the W126 index
summed over the highest three-month interval during a year, based on accumulation over the
08:00 a.m. - 08:00 p.m. daytime 12-hour period; and (3) the level of the standard should be
between 7 ppm-hrs and 15 ppm-hrs. These recommendations are based on scientific
evidence of adverse effect associated with the presence of ozone in ambient air. Note that
these levels are based on an annual form of the standard." (Frey, 2014, p. 15).
With respect to the averaging time, CASAC additionally states that it "does not
recommend the use of a three-year averaging period. We favor a single-year averaging
period, which will provide more protection for annual crops and for the anticipated
cumulative effects on perennial species. The scientific analyses considered in this review,
and the evidence upon which they are based, are from single-year results. If a 3-year
averaging period is established, then the upper limit will need to be reduced to protect
against one-year ozone peaks" (Frey, 2014, p. 13).
5.7 STAFF CONCLUSIONS ON ADEQUACY OF SECONDARY STANDARD
This section presents staff conclusions for the Administrator to consider in deciding
whether the existing secondary Os standard is adequate and whether it should be retained or
revised. Our conclusions are based on consideration of the assessment and integrative synthesis
of information presented in the ISA, as well as our analyses of air quality distributions; analyses
in the WREA; and the comments and advice of CASAC and public comment on earlier drafts of
this document and on the ISA and WREA, as discussed above. Taking into consideration the
responses to specific questions discussed above, we revisit the overarching policy question for
this chapter:
• Does the currently available scientific evidence and exposure/risk information, as
reflected in the ISA and WREA, support or call into question the adequacy and/or
appropriateness of the protection afforded by the current secondary Os standard?
As an initial matter, we note that the CAA does not require that a secondary standard be
protective of all effects associated with a pollutant in the ambient air, but only those considered
adverse to the public welfare (as described in section 1.3.2 above). In helping inform the
Administrator's judgments with respect to the adversity of the effects to public welfare, we have
considered the scientific evidence and risk/exposure information in light of the paradigm used in
the last review that takes into account the variation in public welfare significance of Os-related
5-75
-------
vegetation effects when evaluating the potential adversity of the currently available evidence. As
discussed in Section 5.1, this paradigm recognized that the significance to the public welfare of
Os-induced effects on sensitive vegetation growing within the U.S. can vary 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. Accordingly, any
given Os-related effect on vegetation and ecosystems (e.g., biomass loss, crop yield loss, visible
foliar injury) may be judged to have a different degree of impact on the public welfare
depending, for example, on whether that effect occurs in a Class I area, a city park, or
commercial cropland. In the 2010 proposed reconsideration, the Administrator proposed to place
the highest priority and significance on vegetation and ecosystem effects to sensitive species that
are known to or are likely to occur in federally protected areas such as national parks and other
Class I areas, or on lands set aside by states, tribes and public interest groups to provide similar
benefits to the public welfare (75 FR 3023/24), recognizing that effects occurring in such areas
would likely have the highest potential for being classified as adverse to the public welfare, due
to the expectation that these areas need to be maintained in a more pristine condition to ensure
their intended use is met.
In addition, there is also sufficient support to explicitly include consideration of impacts
to ecosystem goods and services. Although ecosystem services were not explicitly considered in
the Administrator's decision in the last review, they were recognized as an important category of
public welfare effects (73 FR 16492). The CAS AC letter also provides support for this approach.
The inclusion of ecosystem goods and services in this paradigm brings with it a number of
additional considerations. Specifically, when considering the public welfare benefits from these
goods and services, it is important to note that they can accrue across a range of dimensions,
including spatial, temporal, and social, and these likely will vary depending on the type of effect
being characterized. For example, ecosystems can cover a range of spatial scales, and the
services they provide can accrue locally or be distributed more broadly such as when crops are
sold and eaten locally and/or also sold in regional, national and world markets. Ecosystem
services can likewise be realized over a range of temporal scales from immediate up to long term
(e.g., the removal of air pollutants that have a short-term impact on human health but are also
climate forcers with long atmospheric lifetimes, which the removal of may have immediate as
well as long-term benefits). The size of the societal unit receiving benefits from ecosystem
services can also vary dramatically. 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 (U.S. EPA, 2014a, chapter 5, section 5.5.1).
5-76
-------
We thus recognize the usefulness of evaluating the scientific evidence regarding these
effects in the context of the most recent paradigm discussed above. This paradigm integrates the
concepts of: 1) variability in public welfare significance given intended use and value of the
affected entity, such as individual species; 2) relevance of associated ecosystem services to
public welfare; and 3) variability in spatial, temporal, and social distribution of ecosystem
services associated with known and anticipated welfare effects. In so doing, we recognize that
there is no bright-line rule delineating the set of conditions or scales at which known or
anticipated effects become adverse to public welfare. Thus, the evidence and exposure/risk
information discussed in this chapter will be further evaluated in Chapter 6 in light of the
concepts incorporated in this paradigm to help inform the Administrator's judgments with
respect to the potential adversity of the effects to the public welfare.
With respect to the scientific evidence, the longstanding evidence base on the phytotoxic
effects of Os demonstrates that Os -induced effects that occur at the subcellular and cellular
levels, at sufficient magnitudes propagate up to larger spatial scales. The ISA summarizes the
coherence across the full range of effects, from the least serious to the most serious, as follows
(U.S. EPA, 2013, p. 1-8):
The welfare effects ofOs can be observed across spatial scales, starting at the
subcellular and cellular level, then the whole plant and finally, ecosystem-level
processes. Ozone effects at small spatial scales, such as the leaf of an individual
plant, can result in effects along a continuum of larger spatial scales. These
effects include altered rates of leaf gas exchange, growth, and reproduction at the
individual plant level, and can result in broad changes in ecosystems, such as
productivity, carbon storage, water cycling, nutrient cycling, and community
composition.
Many of the recent studies evaluated in this review have focused on and further increased
our understanding of the molecular, biochemical and physiological mechanisms that explain how
plants are affected by Os, in the absence of other stressors, particularly in the area of genomics
(U.S. EPA, 2013, Chapter 9, section 9.3). These recent studies, in combination with the
extensive and long-standing evidence, have further strengthened the coherence and consistency
of the entire body of research, so that our confidence in the supporting science is stronger than in
the previous review.
Based on its assessment of the strength of the science, the ISA determined that the
relationship that exists between exposure to Os in ambient air and visible foliar injury effects on
vegetation, reduced vegetation growth, reduced productivity in terrestrial ecosystems, reduced
yield and quality of agricultural crops and alteration of below-ground biogeochemical cycles
(U.S. EPA 2013, Table 1-2) is causal. Additionally, the ISA determined that a likely to be causal
5-77
-------
relationship exists between exposures to Os in ambient air and reduced carbon sequestration in
terrestrial ecosystems, alteration of terrestrial ecosystem water cycling and alteration of
terrestrial community composition (U.S. EPA, 2013, Table 1-2).
Recent studies also continue to provide strong and consistent evidence that adverse
vegetation effects are attributable to cumulative seasonal Os exposures. On the basis of the
entire body of evidence in this regard, the ISA concludes that "quantifying exposure with indices
that cumulate hourly Os concentrations and preferentially weight the higher concentrations
improves the explanatory power of exposure/response models for growth and yield, over using
indices based on mean and peak exposure values" (U.S. EPA, 2013, p. 2-44). Thus, as in other
recent reviews, the evidence continues to provide a strong basis for concluding that it is
appropriate to judge impacts of Os on vegetation, related effects and services, and the level of
public welfare protection achieved, using a cumulative, seasonal exposure metric, such as the
W126-based metric. In addition, CASAC has consistently since the 1997 review expressed
support for the use of such a metric as the most appropriate form for the secondary NAAQS. In
its most recent letter on the second draft PA, CASAC states that it "concurs with the justification
in this section that the form of the standard should be changed from the current 8-hr form to the
cumulative W126 index and finds that the discussion provides an appropriate and sufficient
rationale" (Frey, 2014, p. 12). Thus, based on the consistent and well-established evidence
described above, we conclude that the most appropriate and biologically relevant way to relate
Os exposure to plant growth, and to determine what would be adequate protection for public
welfare effects attributable to the presence of Os in the ambient air, is to characterize exposures
in terms of a cumulative seasonal form, and in particular the W126 metric.
Accordingly, in considering the current evidence and exposure/risk information with
regard to the adequacy of public welfare protection it affords, we have considered both the
evidence of vegetation and welfare impacts in areas of the U.S. likely to have met the current
standard, as well as air quality information regarding W126 index values in such areas. In
evaluating the adequacy of the current secondary standard, we first considered Os effects on tree
growth, productivity and carbon storage and associated ecosystems and services. Recent studies
confirm and extend the evidence of Os-related effects on tree growth, productivity and carbon
storage. Analysis of existing data conducted by the EPA staff and discussed in the ISA has
substantially reduced the uncertainty associated with using OTC E-R functions to predict tree
growth effects in the field, as described in section 5.2.1 above (U.S. EPA, 2013, section 9.6.3.2).
The median of the composite E-R functions (green line), (U.S. EPA, 2014a, Figure 6-5, section
6.2.1.2) shows RBL for tree seedlings. We note CAS AC's advice that a 6% median RBL is
unacceptably high, and that the 2% median RBL is an important benchmark to consider. The
5-78
-------
median RBL is at or below 2% at the lowest W126 level assessed, 7 ppm-hrs. As the W126
level is incrementally increased, median RBL also increases incrementally, so that at W126
index values of 9, 11, 13, 15, 17, 19 and 21, the median RBL increases to 2.4%, 3.1%, 3.8%,
4.5%, 5.3%, 6.0% and 6.8%, respectively. Based on air quality analyses of 2009-2011 (Appendix
2B), there are approximately 342, 199, 92, 43, 24, 9, 3 and 0 monitors with 3-year average W126
index values above 7, 9, 11, 13, 15, 17, 19 and 21 ppm-hrs when meeting the current standard.
We note that these counts of monitors are based on those meeting the current standard and that
there are many monitors for the 2009-2011 period that do not meet the current standard and also
are above the W126 values of 7-21 ppm-hrs.
We also consider it informative to examine the individual species responses and RBL
over the same W126 range. We first note, based on Figure 5-1 (B) above that over the range of 7
to 17 ppm-hrs, 5 species maintain RBLs of less than 2%. These more tolerant species include
Douglas fir, loblolly pine, Virginia pine, sugar maple and red maple. Two of these species (red
maple and sugar maple) are estimated to have RBL levels above 2% at a W126 of 21. Black
cherry, the most sensitive of the remaining six species, has RBL ranging from 35.57% at W126
of 17 down to 16.67% at the W126 index value of 7 ppm-hrs.
In addition, we also consider the growth effects associated with exposure concentrations
at or below that of the current standard in Class I areas. Specifically, we found that there were 22
Class I areas that had monitor sites that have design values that meet the current standard,
ranging from 67 to 75 ppb, and have 3-year average W126 index values that are above 15 ppm-
hrs between the years of 1998 and 2012 (Table 5-2). Across these 22 Class I areas, the highest
single-year W126 index values for these three-year periods ranged from 17.4 to 29.0 ppm-hrs. In
20 of the areas, distributed across eight states (AZ, CA, CO, KY, NM, SD, UT, WY) and four
regions (west, southwest, west/north central and central), this range was 19.1 to 29.0 ppm-hrs,
exposure values for which the corresponding median species RBL estimates equal or exceed 6%,
which CAS AC termed "unacceptably high". In addition, given that other environmental factors
can influence the extent to which Os may have the impact predicted by the E-R functions in any
given year, we also note that the highest three year periods, that include these highest annual
values for the 21 areas, are at or above 19 ppm-hrs, ranging up to 22.5 ppm-hrs (for which the
median species RBL estimate is above 7%). Additionally, the highest three-year average W126
index value for each of the 22 areas (during periods meeting the current standard) was at or
above 19 (ranging up to 22.5 ppm-hrs) in 11 areas, distributed among five states in the west and
southwest regions (U.S. EPA, 2014c, Table 5-2, Appendix 5B).
In addition, quaking aspen and ponderosa pine are two tree species that are found in most
of these 22 parks and have a sensitivity to Os exposure that places them near the middle of the
5-79
-------
group for which E-R functions have been established. In the areas where ponderosa pine is
present, the highest single year W127 index values ranged from 18.7 to 29.0 and the highest 3-
year average W126 values in which these single year values are represented ranged from 15 to
22.5, with these three-year values above 19 ppm-hrs in eight areas across five states. The
ponderosa pine RBL estimates for 29 and 22.5 ppm-hrs are approximately 12% and 9%,
respectively. In the areas where quaking aspen is present, the highest single year W127 index
values ranged from 19.2 to 26.7 ppm-hrs and the highest 3-year average W126 values in which
these single year values are represented ranged from 15.0 to 22.2, with values above 19 ppm-hrs
in eight areas across five states. The quaking aspen RBL estimates for 26.7 and 22.2 ppm-hrs are
approximately 16% and 13%, respectively. Based on this, we predict growth effects associated
with exposure concentrations at or below that of the current standard for most of these Class I
areas. On the basis of such information, Table 5-2 provides evidence of the potential for
significant growth loss in locations where ambient conditions meet the current standard. Based
on this evidence, we note the occurrence in Class I areas, during periods where the current
standard is met, of cumulative seasonal Os exposures of a magnitude that might reasonably be
concluded to be important to public welfare.
Recent studies have provided additional evidence on tree biomass or growth effects
associated with multiple year exposures in the field, including the potential for cumulative and
carry-over effects. For example, a number of studies were conducted at a planted forest at the
Aspen FACE site in Wisconsin where some researchers observed that the effects of Cb on birch
seeds (reduced weight, germination, and starch levels) could lead to a negative impact on species
regeneration in subsequent years, and that the effect of reduced aspen bud size may have been
related to the observed delay in spring leaf development. These effects suggest that elevated Os
exposures have the potential to alter carbon metabolism of overwintering buds which may have
subsequent effects in the following year. Other studies found that, in addition to affecting tree
heights, diameters, and main stem volumes in the aspen community, elevated Os over a 7-year
study period was reported to increase the rate of conversion from a mixed aspen-birch
community to a community dominated by the more tolerant birch, leading the authors to
conclude that elevated Os may alter intra- and inter-species competition within a forest stand
(U.S. EPA, 2013, section 9.4.3).
While it is not possible at this time to identify the extent or magnitude of such effects in
the field under exposures that may be associated with the current standard, their occurrence, on
federal lands with special protections might reasonably be concluded to be an important public
welfare consideration. We note here that 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.
5-80
-------
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" (Frey, 2014, pp. 9-10).
In regard to the WREA analyses for risks for associated ecosystem services, we note that
the WREA presents estimated changes in consumer and producer/farmer surplus associated with
the change in forestry and agricultural yields. Changes in biomass affect individual tree species
differently, and the overall effect on forest ecosystem productivity depends on the composition
of forest stands and the relative sensitivity of trees within those stands. Economic welfare
impacts resulting from just meeting the existing and alternative standards were largely similar
between the forestry and agricultural sectors — consumer surplus, or consumer gains, generally
increased in both sectors because higher productivity under lower Os concentrations increased
total yields and reduced market prices. Comparisons are not straightforward to interpret due to
market dynamics. The national-scale analysis of carbon dioxide (CCh) sequestration estimates
more storage under the current standard compared to recent conditions, with somewhat smaller
additional increases for the three W126 scenarios in comparison to the current standard scenario
(U.S. EPA 2014a, Appendix 6B, Table B-10).
We additionally consider the WREA estimates of tree growth and ecosystem services
provided by urban trees over a 25-year period for five urban areas based on case-study scale
analyses that quantified the effects of biomass loss on carbon sequestration and pollution
removal (U.S. EPA 2014a, sections 6.6.2 and 6.7).30 The urban areas included in this analysis
represent diverse geography in the Northeast, Southeast, and Central regions, although they do
not include an urban area in the western U.S. Estimates of the effects of Cb-related biomass loss
on carbon sequestration, for example, indicate the potential for an increase of somewhat more
than a million metric tons of CCh equivalents for average W126 index values associated with
meeting the current standard scenario as compared to recent conditions. Somewhat smaller
additional increases are estimated for the three W126 scenarios in comparison to the current
standard scenario (U.S. EPA 2014a, section 6.6.2 and Appendix 6D).
In considering the significance of these WREA analyses of risks for the associated
ecosystem services for timber production, air pollution removal, and carbon sequestration, we
note the large uncertainties associated with these analyses (see U.S. EPA 2014a, Table 6-27), and
the potential to underestimate the response at the national scale. Thus, while we note that it is
appropriate to consider predicted and anticipated impacts to these services in determining the
adequacy of the protection afforded by the current standard, we also note that we place limited
30 The WREA used the i-Tree model for the urban case studies. i-Tree is a peer-reviewed suite of software
tools provided by USFS.
5-81
-------
weight on the absolute magnitude of the risk results for these ecosystem service endpoints in
light of these significant associated uncertainties.
In reaching conclusions regarding support for the adequacy of the current secondary
standard provided by the currently available information on Cb-induced effects on trees and
associated ecosystem services we note that: 1) there is robust evidence supporting the causal
relationship between cumulative Cb exposures and effects on tree growth, productivity, and
carbon storage (U.S. EPA, 2013) and causal and likely to be causal relationships for several
associated ecosystem services; 2) the tree seedling E-R functions evidence, which has been
strengthened, demonstrates variability in sensitivity to Cb across species; 3) estimated median
RBLs are at or above 6%, a key CASAC benchmark, in several areas when air quality was at or
below that of the current standard; 4) growth effects associated with exposure concentrations are
predicted to occur in several Class I areas based on air quality from 1998-2012 that was at or
below that of the current standard; 5) impacts from single year exposures can carry over to the
subsequent year and/or cumulate over multiple years with repeated annual exposures; 6)
evidence from both recent controlled chamber mechanism studies and field based exposure
studies support earlier findings from OTC studies; and 7) WREA analyses show that Cb-induced
biomass loss can impact ecosystem services provided by forests, including timber production,
carbon storage, and air pollution removal, even when air quality is adjusted to just meet the
current standard. Given the above, and noting CASAC views described above, staff concludes
that the current evidence/risk information calls into question the adequacy of the public welfare
protection afforded by the current standard from the known and anticipated adverse effects
associated with Cb-induced impacts on tree growth, productivity and carbon storage, including
the associated ecosystem services assessed in this review, and therefore it is appropriate to
consider revision to provide increased protection.
With respect to crops, the detrimental effect of Cb on crop production has been
recognized since the 1960s, and recent Cb concentrations in many areas across the U.S. are high
enough that they might be expected to cause yield loss in a variety of agricultural crops
including, but not limited to, soybeans, wheat, potatoes, watermelons, beans, turnips, onions,
lettuces, and tomatoes (U.S. EPA, 2013, section 9.4.4). In general, the vast majority of the new
scientific information confirms prior conclusions that exposure to Cb can decrease growth and
yield of crops. Recent research has highlighted the effects of Cb on crop quality. Increasing Cb
concentration decreases nutritive quality of grasses, and decreases macro- and micro-nutrient
concentrations in fruits and vegetable crops (U.S. EPA 2013, section 9.4.4). Recent studies
continue to find yield loss levels in crop species studied previously under NCLAN that reflect
the earlier findings. There has been little published evidence that crops are becoming more
tolerant of Cb (U.S. EPA, 2006a; U.S. EPA 2013). This is especially evident in the research on
5-82
-------
soybean. The 2013 ISA reported comparisons between yield predictions based on data from
cultivars used in NCLAN studies, and yield data for modern cultivars from SoyFACE (U.S.
EPA, 2013, section 9.6.3). They confirm that the average response of soybean yield to Os
exposure has not changed in current cultivars. In addition, satellite and ground-based Os
measurements have been used to assess yield loss caused by Cb over the continuous tri-state area
of Illinois, Iowa, and Wisconsin. The results showed that Os concentrations reduced soybean
yield, which correlates well with the previous results from FACE- and OTC-type experiments
(U.S. EPA, 2013, section 9.4.4.1). Thus, the recently available evidence, as assessed in the ISA,
continues to support the conclusions of the 1996 and 2006 CDs that ambient Os concentrations
can reduce the yield of major commodity crops in the U.S.
The currently available evidence, as assessed in the ISA, continues to support the use of
E-R functions for crops based on OTC experiments. Further, important uncertainties have been
reduced regarding the E-R functions for crop yield loss, especially for soybean. In general, the
ISA reports consistent results across exposure techniques and across crop varieties (U.S. EPA
2013, section 9.6.3.2). Soybean, which is the second-most planted field crop in the U.S.,31 would
be predicted to have no more than 5% RYL at a W126 index value of 12 ppm-hrs, based on the
E-R function. Staff analyses of recent monitoring data (2009-2011) indicate that Os
concentrations in multiple agricultural areas in the U.S. that meet the current standard correspond
to W126 index levels above 12 ppm-hrs. With regard to crops, CASAC states that it "concurs
that another important surrogate for damage that is adverse to public welfare is crop loss. Crops
provide food and fiber services to humans. Evaluation of market-based welfare effects of ozone
exposure in forestry and agricultural sectors is an appropriate approach to take into account
damage that is adverse to public welfare" (Frey, 2014, p. 10). However, as we describe in
section 5.3 above, determining at what point Os-induced crop yield loss becomes adverse to the
public welfare is still unclear, given that it is heavily managed with additional inputs that have
their own associated markets and that benefits can be unevenly distributed between producers
and consumers. We further note that a standard set to provide requisite protection for trees could
also potentially achieve appropriate protection for commodity crops.
In reaching conclusions regarding support for the adequacy of the current secondary
standard provided by the currently available information on Cb-induced effects on crops, we note
that 1) there is clear and robust evidence supporting the causal relationship between cumulative
Os exposures and effects on crop yields and quality (U.S. EPA, 2013); 2) the crop E-R functions
evidence, which has been strengthened, demonstrates variability in sensitivity to Os across
species; 3) evidence from both recent controlled chamber mechanism studies and field based
31 http://www.ers.usda.gov/topics/crops/soybeans-oil-crops/background.aspx
5-83
-------
exposure studies support earlier findings from OTC studies; 4) evidence continues to show that
crops, and in particular soybean, has not become more tolerant of Os (U.S. EPA, 2013, section
9.6.3, 9.4.4.1); 5) WREA analyses show that Os-induced crop yield loss can impact producer and
consumer surpluses and the interaction between agriculture and timber production.
Given the above, and noting CASAC views described above as well as the difficulty in
assessing adversity to public welfare of these effects, staff concludes that the current
evidence/risk information calls into question the adequacy of the public welfare protection
afforded by the current standard from the known and anticipated adverse effects associated with
Os-induced impacts on crop yields and associated services assessed in this review, and therefore,
it is appropriate to consider revision to provide increased protection.
With respect to foliar injury, visible foliar injury surveys are used by the federal land
managers to assess potential Os impacts in Class I areas (USFS, NFS, FWS, 2010). Given this
focus on visible foliar injury, Os-induced impacts have the potential to impact the public welfare
in scenic and/or recreational areas on an annual basis. Visible foliar injury is associated with
important cultural and recreational ecosystem services to the public, such as scenic viewing,
wildlife-watching, hiking, and camping, that are of significance to the public welfare and
enjoyed by millions of Americans every year, generating millions of dollars in economic value
(U.S. EPA 2014a, section 7.1). In addition, several tribes have indicated that many of the known
confirmed Os-sensitive species (including bioindicator species) are culturally significant (see
Appendix 5-A). We further note that CASAC "concurs that visible foliar injury can impact
public welfare by damaging or impairing the intended use or service of a resource. Visible foliar
injury that is adverse to public welfare can include: visible damage to ornamental or leafy crops
that affects their economic value, yield, or usability; visible damage to plants with special
cultural significance; and visible damage to species occurring in natural settings valued for
scenic beauty or recreational appeal" (Frey, 2014, p. 10).
New research on visible foliar injury includes: 1) controlled exposure studies; 2) multi-
year field surveys; and 3) USFS FHM/FIA biomonitoring program surveys and assessments. In
addition to supporting the ISA causal determination, these studies also address some
uncertainties identified in the previous review (i.e., influence of soil moisture on visible foliar
injury development) and further show that visible foliar injury can occur under conditions where
8-hour average Os concentrations are or would be expected to be below the level of the current
standard (e.g., Kline et al., 2008, as discussed in section 5.4.1 above). Incidence of visible foliar
injury symptoms on Cb-sensitive vegetation has also been documented in the field in federally
protected areas that have met the current standard. Importantly, these Cb-induced vegetation
effects have been identified by the federal land managers as a diagnostic tool for informing
conclusions regarding potential ozone impacts on potentially sensitive AQRVs and were found
5-84
-------
in Class I areas that have particular public welfare significance in light of direction from
Congress that these areas as merit a high level of protection (75 FR 3023/3024).
The studies mentioned above also provide additional information regarding the role of
soil moisture in influencing visible foliar injury response (U.S. EPA 2013, section 9.4.2). These
studies confirm that adequate soil moisture creates an environment conducive to greater visible
foliar injury in the presence of Os than drier conditions. As stated in the ISA, "[a] major
modifying factor for Os-induced visible foliar injury is the amount of soil moisture available to a
plant during the year that the visible foliar injury is being assessed ... because lack of soil
moisture generally decreases stomatal conductance of plants and, therefore, limits the amount of
Os entering the leaf that can cause injury" (U.S. EPA, 2013, p. 9-39). As a result, "many studies
have shown that dry periods in local areas tend to decrease the incidence and severity of Os-
induced visible foliar injury; therefore, the incidence of visible foliar injury is not always higher
in years and areas with higher Os, especially with co-occurring drought (Smith, 2012; Smith et
al., 2003)" (U.S. EPA, 2013, p. 9-39). This "...partial 'protection' against the effects of Os
afforded by drought has been observed in field experiments (Low et al., 2006) and modeled in
computer simulations (Broadmeadow and Jackson, 2000)" (U.S. EPA, 2013, p. 9-87). In
considering the extent of any protective role of drought conditions, however, the ISA also notes
that other studies have shown that "drought may exacerbate the effects of Os on plants
(Pollastrini et al., 2010; Grulke et al., 2003)" and that "[tjhere is also some evidence that Os can
predispose plants to drought stress (Maier-Maercker, 1998)". Accordingly, the ISA concludes
that "the nature of the response is largely species-specific and will depend to some extent upon
the sequence in which the stressors occur" (U.S. EPA, 2013, p. 9-87). However, such
uncertainties associated with describing the potential for foliar injury and its severity or extent of
occurrence for any given air quality scenario due to confounding by soil moisture levels make it
difficult to identify an appropriate degree of protection (as well as ambient Os exposure
conditions that might be expected to provide that protection).
We note the WREA analyses of the nationwide dataset (2006-2010) for USFS FHM/FIA
biosites described in section 5.4.2 above, including the observation that the proportion of biosites
with injury varies with soil moisture conditions and Os W126 index values (U.S. EPA 2014a,
Chapter 7, Figure 7-10). The evidence of Cb-attributable visible foliar injury incidence occurring
in USFS FHM/FIA biosites shows that the proportion of biosites showing foliar injury incidence
increases steeply with W126 index values up to approximately 10 ppm-hrs. At W126 index
levels greater than approximately 10 ppm-hrs, the proportion of sites showing foliar injury
incidence is relatively constant. The air quality assessment discussed above identified Class I
areas with recent air quality that met the current standard but were above a W126 index value of
5-85
-------
15 ppm-hrs (Table 5-2). There were 22 Class I areas in this table and most of these areas had 3-
year average W126 index values above 15 ppm-hrs for multiple 3-year periods. Given evidence
of the potential occurrence of visible foliar injury at W126 index values of this magnitude, we
note the ecosystem services that are at risk of impairment because of Os-induced damage directly
due to foliar injury, though data is not available to explicitly quantify these negative effects.
Therefore, staff concludes that air quality levels that are at or below the level of the current
standard may allow levels of visible foliar injury incidence to occur in areas of special
significance to the public welfare.
In reaching conclusions regarding support for the adequacy of the current secondary
standard provided by the currently available information on Ch-induced visible foliar injury we
note that: 1) many species of native plants, including trees, have been observed to have visible
foliar injury symptoms in both OTC and field settings, some of which have also been identified
as bioindicators of Os exposure by the USFS; 2) visible foliar injury has been identified by the
federal land managers as a diagnostic tool for informing conclusions regarding potential Os
impacts on potentially sensitive AQRVs (USFS, NFS, FWS, 2010); 3) visible foliar injury
incidence can occur for some species at very low cumulative exposures, but due to confounding
by soil moisture and other factors, it difficult to predictively relate a given Os exposure to plant
response; and 4) WREA analyses show that based on USFS biosite data, the proportion of
biosites showing foliar injury incidence drops when W126 index values drop below
approximately 10 ppm-hrs. However, we note that, with respect to visible foliar injury, we are
unaware of any guidance for federal land managers regarding at what spatial scale or what
degree of severity visible foliar injury might be sufficient to trigger protective action based on
this potential impact on AQRVs. Further, there does not appear to be any consensus in the
literature in this regard, and CASAC, while identifying target percent biomass loss and yield loss
benchmarks for tree seedlings and commodity crops, respectively, did not provide a similar
recommendation for this endpoint. Likewise, as in previous reviews, the ISA notes 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, 2013, section 9.4.2, p. 9-39) and
further noted that the full body of evidence indicates that there is wide variability in this
endpoint, such that although evidence shows visible foliar injury can occur under very low
cumulative Os concentrations, ".. .the degree and extent of visible foliar injury development
varies from year to year and site to site..., even among co-members of a population exposed to
similar Os levels, due to the influence of co-occurring environmental and genetic factors" (U.S.
EPA 2013, section 9.4.2, p. 9-38).
Given the above, and taking note of CASAC views, we recognize foliar injury as an
important Os effect which, depending somewhat on severity and spatial extent, may reasonably
5-86
-------
be concluded to be of public welfare significance, especially when occurring in nationally
protected areas. However, we also note the uncertainties associated with describing the potential
for foliar injury and its severity or extent of occurrence for any given air quality scenario due to
confounding by soil moisture levels and the difficulty in determining what degree of visible
foliar injury incidence is likely to occur under different air quality conditions, and in particular
on lands with special public welfare significance. We therefore conclude that the current
standard may not adequately protect the public welfare from the known and anticipated adverse
effects associated with Os-induced impacts on visible foliar injury and associated services
assessed in this review. Therefore, it may be appropriate to consider revising the standard to
provide increased public welfare protection, though it is uncertain to what degree these Os-
induced impacts on visible foliar injury would be appropriately judged as important and adverse
from a public welfare perspective.
The information for other welfare effects, including those with causal or likely causal
relationships with Os (e.g., alteration of ecosystem water cycling, changes in climate), is limited
with regard to our ability to consider potential impacts of air quality conditions associated with
the current standard, although the WREA provides some perspective on this issue with regard to
susceptibility to insect attack and fire regime change. We note, however, the importance of these
effects categories to the public welfare.
As noted in section 1.3.2 above, our general approach to informing the Administrator's
judgments recognizes that the available welfare effects evidence demonstrates a range of Os
sensitivity across studied plant species and documents an array of Os-induced effects that extend
from lower to higher levels of biological organization. These effects range from those affecting
cell processes and individual plant leaves to effects on the physiology of whole plants, species
effects and effects on plant communities to effects on related ecosystem processes and services.
Given this evidence, it is not possible to generalize across all studied species regarding which
cumulative exposures are of greatest concern, as this can vary by situation due to differences in
exposed species sensitivity, the importance of the observed or predicted Os-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. At the same time, the evidence also
demonstrates that though effects of concern can occur at very low exposures in sensitive species,
at higher cumulative exposures those effects would likely occur at a greater magnitude and/or
higher levels of biological organization and additional species would likely be impacted. It is
important to note, however, due to the variability in the importance of the associated ecosystem
services provided by different species at different exposures and in different locations, as well as
differences in associated uncertainties and limitations, that, in addition to the magnitude of the
5-87
-------
ambient concentrations, the species present and their public welfare significance are essential
considerations in drawing conclusions regarding the significance of public welfare impact.
Therefore, in developing conclusions in this final PA, we note the complexity of
judgments to be made by the Administrator regarding the adversity of known and anticipated
effects to the public welfare and are mindful that the Administrator's ultimate judgments on the
secondary standard will most appropriately reflect an interpretation of the available scientific
evidence and exposure/risk information that neither overstates nor understates the strengths and
limitations of that evidence and information.
Given all of the above, we reach the conclusion that the available evidence and exposure
and risk information call into question the adequacy of public welfare protection provided by the
current standard, and provides support for consideration of revisions to the current secondary
standard to provide increased public welfare protection. More specifically, staff concludes that it
is appropriate for the Administrator to consider revision of the current secondary Os standard to
increase protection against Os-attributable tree biomass loss, crop yield loss, and visible foliar
injury, and their associated services, and particularly for those effects associated with
cumulative, seasonal exposures that occur in Class I and similarly protected natural areas.
In reaching conclusions on options for the Administrator's consideration, we note that the
final decision to retain or revise the current secondary Os standard is a public welfare policy
judgment to be made by the Administrator, based on her judgment as to what degree of
protection would be requisite (i.e., neither more nor less stringent than necessary) to protect the
public welfare from any known or anticipated adverse effects. This final decision will draw upon
the available scientific evidence for Os-attributable welfare effects, and on quantitative analyses
of vegetation and ecosystem exposures and associated risks to vegetation, ecosystems and their
associated services, and judgments about the appropriate weight to place on the range of
uncertainties inherent in the evidence and analyses. In making this decision, the Administrator
will also need to weigh the importance of these effects and their associated ecosystem services in
the overall context of public welfare protection.
Based on the considerations described in the sections above and summarized below, we
therefore conclude that the currently available evidence and exposure/risk information call into
question the adequacy of the public welfare protection provided by the current standard and
provides support for considering potential alternative standards to achieve increased public
welfare protection, especially for sensitive vegetation and ecosystems in federally protected
Class I and similar areas. In this conclusion, we give particular weight to the evidence indicating
the occurrence in Class I areas that meet the current standard of cumulative seasonal Os
exposures associated with estimates of tree growth impacts of a magnitude that are reasonably
considered important to public welfare.
5-88
-------
5.8 REFERENCES
Abt Associates, Inc. (1995). Ozone NAAQS benefits analysis: California crops. Report to U.S. EPA, July.
Andersen, CP; Wilson, R; Plocher, M; Hogsett, WE. (1997). Carry-over effects of ozone on root growth and
carbohydrate concentrations of ponderosa pine seedlings. Tree Physiol 17: 805-811.
Betzelberger, AM; Gillespie, KM; McGrath, JM; Koester, RP; Nelson, RL; Ainsworth, EA. (2010). Effects of
chronic elevated ozone concentration on antioxidant capacity, photosynthesis and seed yield of 10 soybean
cultivars. Plant Cell Environ 33: 1569-1581. http://dx.doi.0rg/10.llll/j.1365-3040.2010.02165.x
Border, K.; DeTemmerman, L.; Ceulemans, R. (2000b) Effects of ozone exposure in open-top chambers on poplar
(Populus nigrd) and beech (Fagus sylvatica): a comparison. Environ. Pollut. 109: 509-516.
Broadmeadow, MSJ; Jackson, SB. (2000). Growth responses of Quercus petraea, Fraxinus excelsior and
Pinussylvestris to elevated carbon dioxide, ozone and water supply. New Phytol 146: 437-
451.http://dx.doi.org/10.1046/j.l469-8137.2000.00665.x
Burns, Russell M., and Barbara H. Honkala, tech. coords. (1990). Silvics of North America: 1. Conifers; 2.
Hardwoods. Agriculture Handbook 654. U.S. Department of Agriculture, Forest Service, Washington, DC.
vol.2, 877 p.
Burns, Russell M., and Barbara H. Honkala, tech. coords. (1990). Silvics of North America: 1. Conifers; 2.
Hardwoods. Agriculture Handbook 654. U.S. Department of Agriculture, Forest Service, Washington, DC.
vol.2, 877 p.
CAL FIRE (California Department of Forestry and Fire Protection). (2008). CAL FIRE 2007 Wildland Fire
Summary. Available at www.fire.ca.gov/communications/downloads/fact_sheets/2007Summary.pdf.
Campbell, SJ; Wanek, R; Coulston, JW. (2007). Ozone injury in west coast forests: 6 years of monitoring -
Introduction. Portland, OR: U.S. Department of Agriculture.
Darbah, JNT; Kubiske, ME; Neilson, N; Oksanen, E; Vaapavuori, E; Karnosky, DF. (2007). Impacts of elevated
atmospheric CO2 and Os on paper birch (Betula papyrifera): Reproductive fitness. Scientific World Journal
7: 240-246. http://dx.doi.org/10.1100/tsw.2007.42
Darbah, JNT; Kubiske, ME; Nelson, N; Oksanen, E; Vapaavuori, E; Kamosky, DF. (2008). Effects of decadal
exposure to interacting elevated CO2 and/or O-3 on paper birch (Betula papyrifera) reproduction. Environ
Pollut 155: 446-452. http://dx.doi.Org/10.1016/j.envpol.2008.01.033Davis, DD. (2007a). Ozone-induced
symptoms on vegetation within the Moosehorn National Wildlife Refuge in Maine. Northeast Nat 14: 403-
414. http://dx.doi.org/10.1656/1092-6194(2007)14[403:OSOVWT]2.0.CO;2
Davis, DD. (2007a). Ozone-induced symptoms on vegetation within the Moosehorn National Wildlife Refuge in 3
Maine. Northeast Nat 14: 403-414. http://dx.doi.org/10.1656/1092-
6194(2007)14[403:OSOVWT]2.0.CO;2
Davis, DD. (2007b). Ozone injury to plants within the Seney National Wildlife Refuge in northern Michigan.
Northeast Nat 14: 415-424.
Davis, DD. (2009). Ozone-induced stipple on plants in the Cape Romain National Wildlife Refuge, South Carolina.
Southeastern Naturalist 8: 471-478.
Davis, DD; Orendovici, T. (2006). Incidence of ozone symptoms on vegetation within a National Wildlife Refuge in
New Jersey, USA. Environ Pollut 143: 555-564. http://dx.doi.0rg/10.1016/j.envpol.2005.10.051
5-89
-------
Federal Register 1996. National Ambient Air Quality Standards for Ozone; Proposed Rule. 40 CFR 50; Federal
Register 61: 65716
Federal Register 1997. National Ambient Air Quality Standards for Ozone; Final Rule. 40 CFR 50; Federal Register
62:38856
Federal Register 2007. National Ambient Air Quality Standards for Ozone; Proposed Rule. 40 CFR 50; Federal
Register 72: 37818
Federal Register 2008. National Ambient Air Quality Standards for Ozone; Final Rule. 40 CFR parts 50 and 58;
Federal Register 73:16436
Federal Register 2010. National Ambient Air Quality Standards for Ozone; Proposed Rule. 40 CFR 50 and 58;
Federal Register 75 FR: 2938
Federal Register 2012. National Ambient Air Quality Standards for Oxides of Nitrogen and Sulfur; Final Rule. 40
CFR 50; Federal Register 77 FR 20218
Fiore, AM; Jacob, DJ; Field, BD; Streets, DG; Fernandes, SD; Jang, C. (2002). Linking ozone pollution and climate
change: The case for controlling methane. Geophys Res Lett 29: 1919.
http://dx.doi.org/10.1029/2002GL015601
Frey, C. and Samet, J.M. (2012) 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). EPA-CASAC-13-
003. November 26, 2012.
Frey, C. (2014) CASAC Review of the EPA's Second Draft Policy Assessmentfor the Review of the Ozone National
Ambient Air Quality Standards. EPA-CASAC-14-004. June 26, 2014.
Grantz, DA; Gunn, S; Vu, HB. (2006). O3 impacts on plant development: A meta-analysis of root/shoot allocation
and growth. Plant Cell Environ 29: 1193-1209. http://dx.doi.0rg/10.llll/j.1365-3040.2006.01521.x
Gregg, JW; Jones, CG; Dawson, TE. (2003). Urbanization effects on tree growth in the vicinity of New York City
[Letter]. Nature 424: 183-187. http://dx.doi.org/10.1038/nature01728
Grulke, NE; Johnson, R; Esperanza, A; Jones, D; Nguyen, T; Posch, S; Tausz, M. (2003). Canopy transpiration of
Jeffrey pine in mesic and xeric microsites: Os uptake and injury response. Trees Struct Funct 17: 292-298.
Grulke, NE; Minnich, RA; Paine, TD; Seybold, SJ; Chavez, DJ; Fenn, ME; Riggan, PJ; Dunn, A. (2008).
Airpollution increases forest susceptibility to wildfires: A case study in the San Bernardino Mountains in
southern California. In A Bytnerowicz; MJ Arbaugh; AR Riebau; C Anderson (Eds.), Wildland fires and
air pollution; Section III: Ecological Impacts of Forest Fires and Air Pollution (pp. 365-403). Amsterdam,
The Netherlands: ElsevierLtd. http://dx.doi.org/10.1016/S1474-8177(08)00017-X
Haefele, M., R.A. Kramer, and T.P. Holmes. (1991). Estimating the Total Value of a Forest Quality in High-
Elevation Spruce-Fir Forests. The Economic Value of Wilderness: Proceedings of the Conference. Gen.
Tech. Rep. SE-78 (pp. 91-96). Southeastern For. Exper. Station. Asheville, NC: USDA Forest Service.
Hall, Karen R. and Richard R. Braham. (1998). Native Pines of Eastern North America. North Carolina State
University Department of Forestry. Available at http://www.ncsu.edu/project/dendrology/
Holmes, T., and R. Kramer. (1995). "An Independent Sample Test of Yea-Saying and StartingPoint Bias in
Dichotomous-Choice Contingent Valuation." Journal of Environmental Economics and Management
28:121-132.
5-90
-------
Heagle, AS. (1989). Ozone and crop yield*. Annu Rev Phytopathol 27: 397-423.
http://dx.doi.org/10.1146/annurev.py.27.090189.002145
Heck, WW; Cowling, EB. (1997).The need for a long term cumulative secondary ozone standard - An ecological
perspective. EM January: 23-33.
Henderson, R. (2006) Letter from CASAC Chairman Rogene Henderson to EPA Administrator Stephen Johnson.
October 24, 2006, EPA-CASAC-07-001.
Henderson, R. (2008) Letter from CASAC Chairman Rogene Henderson to EPA Administrator Stephen Johnson.
April 7,2008, EPA-CASAC-08-009.
Hogsett, WE; Weber, JE; Tingey, D; Herstrom, A; Lee, EH; Laurence, JA. (1997). Environmental auditing: An
approach for characterizing tropospheric ozone risk to forests. J Environ Manage 21: 105-120.
http://dx.doi.org/10.1007/s002679900010
IPCC (Intergovernmental Panel on Climate Change). (2007). Climate change 2007: Impacts, adaptation and
vulnerability. Cambridge, UK: Cambridge University Press.
King, JS; Kubiske, ME; Pregitzer, KS; Hendrey, GR; McDonald, EP; Giardina, CP; Quinn, VS; Karnosky,
DF.(2005). Tropospheric O3 compromises net primary production in young stands of trembling aspen,
paper birch and sugar maple in response to elevated atmospheric CO2. New Phytol 168: 623-635.
http://dx.doi.0rg/10.llll/j.1469-8137.2005.01557.x
Kline, LJ; Davis, DD; Skelly, JM; Savage, JE; Ferdinand, J. (2008). Ozone sensitivity of 28 plant selections exposed
to ozone under controlled conditions. Northeast Nat 15: 57-66. http://dx.doi.org/10.1656/1092-
6194(2008)15[57:OSOPSE]2.0.CO;2
Kohut, R. (2007). Assessing the risk of foliar injury from ozone on vegetation in parks in the US National Park
Service's Vital Signs Network. Environ Pollut 149: 348-357.
Kubiske, ME; Quinn, VS; Heilman, WE; McDonald, EP; Marquardt, PE; Teclaw, RM; Friend, AL; Karnoskey, DF.
(2006). Interannual climatic variation mediates elevated CO2 and Os effects on forest growth. Global
Change Biol 12: 1054-1068. http://dx.doi.0rg/10.llll/j.1365-2486.2006.01152.x
Kubiske, ME; Quinn, VS; Marquardt, PE; Karnosky, DF. (2007). Effects of elevated atmospheric CO2 and/or Os on
intra- and interspecific competitive ability of aspen. Plant Biol (Stuttg) 9: 342-355.
http://dx.doi.org/10.1055/s-2006-924760
Lee, EH; Hogsett, WE. (1996). Methodology for calculating inputs for ozone secondary standard benefits analysis:
Part II. Research Triangle Park, NC: U.S. Environmental Protection Agency.
Lefohn, AS; Laurence, JA; Kohut, RJ. (1988). A comparison of indices that describe the relationship between
exposure to ozone and reduction in the yield of agricultural crops. Atmos Environ 22: 1229-1240.
http://dx.doi.org/10.1016/0004-6981(88)90353-8
Lefohn, AS; Runeckles, VC. (1987). Establishing standards to protect vegetation - ozone exposure/dose
considerations. Atmos Environ 21: 561-568. http://dx.doi.org/10.1016/0004-6981(87)90038-2
Lefohn, AS; Jackson, W; Shadwick, DS; 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: 1695-1708. http://dx.doi.org/10.1016/S1352-2310(96)00258-0
5-91
-------
Low, M; Harbinger, K; Nunn, AJ; Haberle, KH; Leuchner, M; Heerdt, C; Werner, H; Wipfler, P; Pretzsch, H;
Tausz, M; Matyssek, R. (2006). Extraordinary drought of 2003 overrules ozone impact on adult beech trees
(Fagus sylvatica). Trees Struct Funct 20: 539-548. http://dx.doi.org/10.1007/s00468-006-0069-z
Maier-Maercker, U. (1998). Predisposition of trees to drought stress by ozone. Tree Physiol 19: 71-78.
McLaughlin, SB; Nosal, M; Wullschleger, SD; Sun, G. (2007a). Interactive effects of ozone and climate on tree
growth and water use in a southern Appalachian forest in the USA. New Phytol 174: 109-124.
http://dx.doi.0rg/10.llll/j.1469-8137.2007.02018.x
McLaughlin, SB; Wullschleger, SD; Sun, G; Nosal, M. (2007b). Interactive effects of ozone and climate on water
use, soil moisture content and streamflow in a southern Appalachian forest in the USA. New Phytol 174:
125-136. http://dx.doi.0rg/10.llll/j.1469-8137.2007.01970.x
Morgan, PB; Ainsworth, EA; Long, SP. (2003). How does elevated ozone impact soybean? Ameta-analysis of
photosynthesis, growth and yield. Plant Cell Environ 26: 1317-1328.
Novak, K; Cherubini, P; Saurer, M; Fuhrer, J; Skelly, JM; Krauchi, N; Schaub, M. (2007). Ozone air pollution
effects on tree-ring growth, delta(13)C, visible foliar injury and leaf gas exchange in three ozone-sensitive
woody plant species. Tree Physiol 27: 941-949.
National Park Service (NFS). (2006). Ozone Sensitive Plant Species in NFS Class I Areas. Available at
http://www2.nature.nps.gov/air/Pubs/pdf/flag/NPSozonesensppFLAG06.pdf.
NCSU College of Forest Resources (). Loblolly Pine. In Native Pines of Eastern North America. Available at
http://www.ncsu.edu./project/dendrologv/index/plantae/vascular/seedplants/gvmnosperms/conifers/pine/pin
us/australes/loblollypine/html.
Pollastrini, M; Desotgiu, R; Cascio, C; Bussotti, F; Cherubini, P; Saurer, M; Gerosa, G; Marzuoli, R. (2010).
Growth and physiological responses to ozone and mild drought stress of tree species with different
ecological requirements. Trees Struct Funct 24: 695-704. http://dx.doi.org/10.1007/s00468-010-0439-4
Richards, B.L., Middleton, J.T., and Hewitt, W.B. (1958). Air pollution with relation to agronomic crops. V.
Oxidant stipple on grape. Agron. J. 50:559-561.
Riikonen, J; Kets, K; Darbah, J; Oksanen, E; Sober, A; Vapaavuori, E; Kubiske, ME; Nelson, N; Karnosky, DF.
(2008). Carbon gain and bud physiology in Populus tremuloides and Betula papyrifera grown under
longterm exposure to elevated concentrations of CO2 and O3. Tree Physiol 28: 243-254.
http://dx.doi.Org/10.1093/treephys/28.2.243 Rose, A and Coulston, J.W. (2009). Ozone injury across the
Southern United States, 2002 - 2006. Gen. Tech. Rep. SRS-118. Asheville, N.C.: US Department of
Agriculture, Forest Service, Southern Research Station, pp. 25.
Rose, A and Coulston, J.W. (2009). Ozone injury across the Southern United States, 2002 - 2006. Gen. Tech. Rep.
18 SRS-118. Asheville, N.C.: US Department of Agriculture, Forest Service, Southern Research Station,
pp. 1925.
Samet, J.M. (2010) Review of EPA's Proposed Ozone National Ambient Air Quality Standard (Federal Register,
Vol. 75, Nov. 11, January 19, 2010). EPA-CASAC-10-007. February 19, 2010. Available online at:
http://yosemite.epa.gov/sab/sabproduct.nsf/264cbl227d55e02c85257402007446a4/610BB57CFAC8A41C
852576CF007076BD/$File/EPA-CASAC-10-007-unsigned.pdf
Smith, G; Coulston, J; Jepsen, E; 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: 271-291.
5-92
-------
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. http://dx.doi.org/10.1007/sl0661-011-2243-z
Stevenson, DS; Dentener, FJ; Schultz, MG; Ellingsen, K; VanNoije, TPC; Wild, O; Zeng, G; Amann, M; Atherton,
CS; Bell, N; Bergmann, DJ; Bey, I; Butler, T; Cofala, J; Collins, WJ; Derwent, RG; Doherty, RM; Drevet,
J; Eskes, HJ; Fiore, AM; Gauss, M; Hauglustaine, DA; Horowitz, LW; Isaksen, ISA; Krol, MC; Lamarque,
JF; Lawrence, MG; Montanaro, V; Muller, JF; Pitari, G; Prather, MJ; Pyle, JA; Rast, S; Rodriguez, JM;
Sanderson, MG; Savage, NH; Shindell, DT; Strahan, SE; Sudo, K; Szopa, S. (2006). Multimodel ensemble
simulations of present-day and near-future tropospheric ozone. J Geophys Res 111: D08301.
http://dx.doi.org/10.1029/2005JD006338
UNEP (United Nations Environment Programme). (2003). Millennium Ecosystem Assessment: Ecosystems and
human well-being: A framework for assessment. Washington, DC: Island Press.
University of Montana, College of Forest Conservation Wilderness Institute (2014).
http://www.wilderness.net/printFactSheet.cfm?WID=583. accessed in 2014. U.M. College of Forestry and
Conservation's Wilderness Institute in collaboration with the Arthur Carhart National Wilderness Training
Center and Aldo Leopold Wilderness Research Institute.
USD A, Forest Service (2014) Tree basal area data, http://www.fs.fed.us/foresthealth/technologv/nidrm2012.shtml
USD A, National Resources Conservation Service (2014), The PLANTS Database (http://plants.usda.gov. 2014),
National Plant Data Center, Baton Rouge, LA.USDA Agricultural Research Service. (2012). Effects of
Ozone Air Pollution on Plants. Available at http://www.ars.usda.gov/Main/docs.htm?docid=12462.
USDA Economic Research Service. (2012). Background. Soil & Oilseed Crops. Available at
http://www.ers.usda.gov/topics/crops/sovbeans-oil-crops/background.aspx.
USDA US Forest Service. (1990). Acer rubrum L. Red Maple. In Silvics of North America. Agriculture Handbook
654, vol. 2 Hardwoods, Washington, DC. 877 p. Available at
http://www.na.fs.fed.us/pubs/silvics manual/volume 2/acer/rubrum.htm.
USDA US Forest Service. (2014). Environmental Justice. Available at
http://www.fs.fed.us/research/urban/environmental-justice.php.
U.S. EPA (1989). 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 (1996). Air quality criteria for ozone and related photochemical oxidants [EPA Report]. (EPA/600/P-
93/004AF). U.S. Environmental Protection Agency, Research Triangle Park, NC.
U.S. EPA (2006a). Air Quality Criteria for Ozone and Related Photochemical Oxidants (2006 Final). U.S.
Environmental Protection Agency, Washington, DC. EPA/600/R-05/004aF-cF. March 2006. Available at:
http://www.epa.gOv/ttn/naaqs/standards/ozone/s o3crcd.html
U.S. EPA (2006b). Ecological Benefits Assessment Strategic Plan. EPA-240-R-06-001. Office of Administrator,
Washington, D.C. Available at http://yosemite.epa.gov/ee/epa/eed.nsf/webpages/ecologbenefitsplan.html
U.S. EPA (U.S. Environmental Protection Agency). (2007). Review of the national ambient air quality standards for
ozone: Policy assessment of scientific and technical information: OAQPS staff paper [EPA Report].
(EPA/452/R-07/003). Research Triangle Park, NC.
http://www.epa.gov/ttn/naaqs/standards/ozone/data/2007_01_ozone_staff_paper
5-93
-------
U.S. EPA (2013). Integrated Science Assessment of Ozone and Related Photochemical Oxidants (Final). U.S.
Environmental Protection Agency, Washington, DC. EPA/600/R-10/076F
U.S. EPA (2014a). Welfare Risk and Exposure Assessment for Ozone. Office of Air Quality Planning and
Standards, Research Triangle Park, NC, 27711. EPA-452/P-14-005a
U.S. EPA. (2014b). Technology Transfer Network (TTN) Air Quality System (AQS). Available at
http://www.epa.gov/ttn/airs/airsaqs.
U.S. EPA. (2014c). Technology Transfer Network (TTN) Air Quality System (AQS). Download Detailed AQS
Data. Available at http://www.epa.gov/ttn/airs/airsaqs/detaildata/downloadaqsdata.htm.
U.S. EPA. (2014d). Air Trends: Design Values. Available at http://www.epa.gov/airtrends/values.html.
US Forest Service, National Park Service, and U.S. Fish and Wildlife Service. 2010. Federal land managers' air
quality related values work group (FLAG): phase I report—revised (2010). Natural Resource Report
NPS/NRPC/NRR—2010/232. National Park Service, Denver, Colorado.
http://www.nature.nps.gov/air/Pubs/pdf/flag/FLAG 20IQ.pdfWest JJ; Fiore, AM; Horowitz, LW;
Mauzerall, DL. (2006). Global health benefits of mitigating ozone pollution with methane emission
controls. PNAS 103: 3988-3993. http://dx.doi.org/10.1073/pnas.0600201103
West, JJ; Fiore, AM; Naik, V; Horowitz, LW; Schwarzkopf, MD; Mauzerall, DL. (2007). Ozone air quality and
radiative forcing consequences of changes in ozone precursor emissions. Geophys Res Lett 34: L06806.
http://dx.doi.org/10.1029/2006GL029173
Wittig, VE; Ainsworth, EA; Long, SP. (2007). To what extent do current and projected increases in surface ozone
affect photosynthesis and stomatal conductance of trees? A meta-analytic review of the last 3 decades of
experiments [Review]. Plant Cell Environ 30: 1150-1162. http://dx.doi.0rg/10.llll/j.1365-
3040.2007.01717.x
Wittig, VE; Ainsworth, EA; Naidu, SL; Karnosky, DF; Long, SP. (2009). Quantifying the impact of current and
future tropospheric ozone on tree biomass, growth, physiology and biochemistry: A quantitative meta-
analysis. Global Change Biol 15: 396-424. http://dx.doi.0rg/10.llll/j.1365-2486.2008.01774.x
Young, T. F.; Sanzone, S., eds. (2002). A framework for assessing and reporting on ecological condition: an SAB
report. Washington, DC: U.S. Environmental Protection Agency, Science Advisory Board; report no.
EPASAB- EPEC-02-009. Available: http://www.epa.gov/sab/pdf/epec02009.pdf [9 December, 2003].
5-94
-------
6 CONSIDERATION OF ALTERNATIVE SECONDARY
STANDARDS
Chapter 5 reached the conclusion that the available evidence and exposure and risk
information call into question the adequacy of public welfare protection provided by the current
standard, and that it is appropriate for the Administrator to consider revising the current
secondary standard to provide increased public welfare protection against Os-attributable effects
on tree biomass loss, crop yield loss, and visible foliar injury, and their associated services,
particularly for those effects associated with cumulative, seasonal exposures, to the extent these
effects are judged adverse to the public welfare. Given that conclusion, this chapter describes
the staff evaluation of the available body of evidence, and exposure, risk and air quality
information with regard to support for consideration of alternative standards, as articulated by the
following overarching question:
• What alternative secondary standards are supported by the currently available
scientific evidence, exposure/risk information and air quality analyses?
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 in sections 6.1, 6.2 and 6.3 below. We consider both the scientific
and technical information available at the time of the last review and information newly available
since the last review which has been critically analyzed and characterized in the ISA.
Specifically, we consider the currently available scientific evidence and technical information in
the context of decisions regarding the basic elements of the NAAQS: indicator (section 6.1);
averaging time and form (section 6.2); and level (section 6.3). CASAC advice on potential
alternative standards is described in section 6.4 and staff conclusions on potential alternative
standards are discussed in section 6.5. Section 6.6 summarizes staff conclusions on the adequacy
of the current standard and the alternative standards that are appropriate for the Administrator to
consider. Key uncertainties in this review and areas in which future research and data collection
would better inform the next review are identified in section 6.7.
6.1 INDICATOR
With regard to the selection of an appropriate indicator for alternative secondary
standards, we consider the following question.
6-1
-------
• Does the information available in this review continue to support Os as the
indicator for ambient air photochemical oxidants?
In the last review of the air quality for Os and other photochemical oxidants and of the Os
standard, as in other prior reviews, the EPA focused on a standard for Os as the most appropriate
surrogate for ambient photochemical oxidants. Ozone is a long-established surrogate for
ambient photochemical oxidants, among which it is by far the most widely studied with regard to
effects on welfare and specifically on vegetation. The information available in this review adds
to our understanding of the atmospheric chemistry for photochemical oxidants and Os in
particular (as described in the ISA, sections 3.2 and 3.6, and summarized in section 2.2 in this
document). The 1996 Staff Paper noted that the database on vegetation effects is generally
considered to raise concern at levels found in the ambient air for Os and, therefore, control of
ambient Os levels has previously been concluded to provide the best means of controlling other
photochemical oxidants of potential welfare concern (U.S. EPA, 1996b, p. 277). In the current
review, while the complex atmospheric chemistry in which Os plays a key role has been
highlighted, no alternatives to Os have been advanced as being a more appropriate surrogate for
ambient photochemical oxidants. Ozone continues to be the only photochemical oxidant other
than nitrogen dioxide that is routinely monitored and for which a comprehensive database exists
(U.S. EPA, 2013, section 3.6). Thus, staff concludes that Os remains the appropriate pollutant
indicator for use in a secondary NAAQS that provides protection for public welfare from
exposure to all photochemical oxidants.
6.2 FORM AND AVERAGING TIME
In considering potential forms and averaging times alternative to that of the current
secondary standard (i.e., 4th highest daily maximum 8-hour average, averaged over 3 years), we
address several specific questions.
• To what extent does the currently available information provide support for
considering forms different from that of the current secondary standard?
In characterizing the current evidence, the ISA states that "[n]o recent information is
available since 2006 that alters the basic conclusions put forth in the 2006 and 1996 Os AQCDs"
with regard to biologically relevant exposure indices (U.S. EPA, 2013, section 2.6.6.1, p. 2-43).
Based on the current state of knowledge and the best available data assessed in this review, the
ISA therefore concludes that exposure indices that cumulate and differentially weight the higher
hourly average concentrations over a season and also include the mid-level values continue to
offer the most scientifically defensible approach for use in developing response functions and in
defining indices for vegetation protection. Quantifying exposures with indices that cumulate
hourly Os concentrations and preferentially weight the higher concentrations improves the
6-2
-------
explanatory power of exposure/response models for growth and yield, over using indices based
on mean and peak exposure values (U.S. EPA, 2013, section 2.6.6.1). These conclusions are
based on the available body of evidence which provides a wealth of information, compiled over
several decades, on the aspects of Os exposure that are most important in influencing plant
response. As discussed in the ISA, the importance of the duration of the exposure and the
relatively greater importance of higher concentrations (over lower concentrations) in determining
plant response to Os have been well documented (U.S. EPA, 2013, section 9.5.3). Building on
this research, other work has focused on developing "mathematical approaches for summarizing
ambient air quality information in biologically meaningful forms for Os vegetation effects
assessment purposes ..." (U.S. EPA, 2013, section 9.5.2, p. 9-99), including those known as
cumulative, concentration weighted forms (i.e., SUM06, W126). Much of this work was
completed by the mid-1990s, and was summarized in the 1996 Criteria Document (CD) (U.S.
EPA, 1996a, section 5.5).
On the basis of this longstanding and extensive evidence demonstrating that the risk to
vegetation comes from cumulative seasonal exposures, the EPA in the 1997 and 2008 reviews, as
well as in the 2010 proposed rulemaking to reconsider the 2008 decision, recognized the
importance of cumulative, seasonal exposures as a primary determinant of plant responses to Os
in ambient air (61 FR 65741-42; 62 FR 38878; 72 FR 37893, 37896, 37900, 37904; 73 FR
16488-90, 16493-94; 75 FR 3000, 3010, 3012). For example, in the 1996 notice of proposed
rulemaking, the Administrator recognized that the scientific evidence supported the conclusion
that "a cumulative seasonal exposure index is more biologically relevant than a single event or
mean index" (61 FR 65742). In the 2008 review, CAS AC recognized that an important
difference between the effects of short-term exposures to Cb on human health and the effects of
Os exposures on welfare is that "vegetation effects are more dependent on the cumulative
exposure to, and uptake of, Os over the course of the entire growing season" (Henderson, 2006,
p. 5). In that review, the CASAC Cb Panel members were unanimous in supporting the final
Staff Paper recommendation that "protection of managed agricultural crops and natural terrestrial
ecosystems requires a secondary Ozone NAAQS that is substantially different from the primary
ozone standard in averaging time, level and form" (Henderson, 2007, p. 3). Accordingly, in both
the 1997 and 2008 reviews as well as the 2010 reconsideration, the Administrator proposed a
secondary standard with a cumulative seasonal form as an appropriate policy option (61 FR
65742-44; 72 FR 37899-905; 75 FR 3012-3027).
In considering which exposure index was best suited for use as a form for the secondary
Os NAAQS, the 1996 CD and 1996 Staff Paper evaluated a variety of different types of forms.
These documents noted that a number of forms (e.g., the one event, mean and unweighted
cumulative SUMOO) are unable to reliably predict plant response because they either ignore the
6-3
-------
role of duration or ignore the disproportionate impact of higher concentrations by weighting all
concentrations equally (U.S. EPA, 1996b, p. 224). Other forms that were considered at that time
included multicomponent forms which take into account many other relevant factors (e.g., plant
growth stage, predisposition from earlier exposures). Of all the different exposure forms, these
multicomponent forms consistently predict plant response best. However, due to being species-
specific and highly complex, they were not considered suitable for more general application in
the context of standard setting (U.S. EPA, 1996b, pp. 224-225). On the other hand,
concentration-weighted forms that take into account the role of duration and concentration
perform almost as well as the multicomponent forms. These forms include several threshold
forms (e.g., SUM06, AOT60) and sigmoidally weighted cumulative indices (e.g., W1261) (U.S.
EPA, 1996a, pp. 5-84 to 5-136; U.S. EPA, 1996b, pp. 223-227). Given that these cumulative
concentration-weighted forms were able to similarly predict plant response on the datasets for
which they were evaluated (e.g., NCLAN), it was not possible to distinguish between them on
this basis. Partly as a result, CASAC deliberations in 1995 did not produce a consensus on
which cumulative concentration-weighted form would be best suited for a secondary NAAQS.
As discussed further in 6.3 below, a workshop held in January of 1996 provided a consensus
recommendation on the SUM06 form as appropriate for use in secondary standards, while also
recognizing that a W126 form could also be appropriate (Heck and Cowling, 1997). Subsequent
to this, the final 1996 Staff Paper and 1996 proposal notice both identified the SUM06 form as
appropriate to consider and propose, respectively (U.S. EPA, 1996b, p. 285, 61 FR 65716). In
selecting the SUM06 form that imposed a threshold despite the lack of scientific evidence for a
discernible threshold for Os-related vegetation effects across the range of studied species, the
EPA noted that it had the benefit of not including concentrations that were considered at the time
to be within the range of background, which was considered to be an important feature (U.S.
EPA, 1996b, pp. 223-227).
In the subsequent review, the form of the standard was revisited in light of continued
evidence that there remained a lack of discernible threshold for vegetation effects in general, and
newer estimates of Os concentrations associated with background sources were lower than in the
previous review such that their inclusion was less of a concern. On these bases, the 2007 Staff
Paper recommended consideration of the W126 index as the basis for the form of a distinct
secondary standard (U.S. EPA, 2006, pp. 9-11 to 9-15 and pp. AX9-159 to AX9-187; U.S. EPA,
2007, pp. 7-15/16). The EPA then proposed two options for the secondary standard, one of
1 The W126 is a non-threshold approach described as the sigmoidally weighted sum of all hourly O3
concentrations observed during a specified diurnal and seasonal exposure period, where each hourly O3
concentration is given a weight that increases from 0 to 1 with increasing concentration (Lefohn et al, 1988; Lefohn
and Runeckles, 1987; U.S. EPA, 2013, section 9.5.2).
6-4
-------
which was to adopt a cumulative, seasonal standard based on the W126 index, while the other
option was a secondary standard identical to the proposed revised primary standard (72 FR
37818). The CASAC Panel in that review expressed preference for the W126 index (Henderson,
2006). In deciding to reconsider the 2008 decision, the Administrator noted that past arguments
or reasons for not moving to a cumulative, seasonal form, with appropriate exposure periods,
were not based on disagreement over the biological relevance of the cumulative, seasonal form,
or the recognized disadvantages of an 8-hour standard in measuring and identifying a specified
cumulative, seasonal exposure pattern but were based on concerns over whether the EPA had an
adequate basis to determine an appropriate level for a cumulative, seasonal secondary standard
(75 FR 3019). Having reached the conclusion that such a level could be identified from within
the range of levels proposed, the Agency proposed to set a secondary NAAQS in terms of a
cumulative, seasonal standard form based on the W126 function (75 FR 2938). The CASAC
also stated its support for this proposal, noting that it found the Agency's reasoning to be
supported by the extensive scientific evidence considered in the last review (Samet, 2010).
In this review, we conclude that specific features associated with the W126 index still
make it the most appropriate and biologically relevant cumulative concentration-weighted form
for use in the context of the secondary Os NAAQS review. In particular, the W126 index does
not apply an arbitrary exposure threshold below which concentrations are not included. Given
the acknowledged variability in vegetation sensitivity, including evidence that some species are
sensitive at very low cumulative exposures, and the continued lack of evidence of an exposure
threshold for effects above a W126 index of zero, such a feature is scientifically justifiable and
desirable. Thus, we conclude that the W126 form is best matched to the evidence associated
with vegetation effects, as well as addressing the policy-relevant issue of how to weight
exposures associated with background sources.
• To what extent does the currently available information provide support for
consideration of a cumulative seasonal form derived as a sum of weighted Os
concentrations over daylight hours (8:00 am to 8:00 pm) and over the consecutive
3-month period having the highest sum within the Os season?
As discussed in Chapter 5, mechanistic studies, including those recently assessed in this
review, provide biological plausibility for the conclusions reached in the ISA that Os-induced
effects on plants are cumulative, that higher concentrations appear to be more important in
eliciting a response than lower concentrations, and that plant sensitivity to Os can vary with time
of day (U.S. EPA, 2013, p. 2-44). In particular, studies have shown that plants take up and
respire gases through openings in their leaves, called stomata, which in general, are most open
during daylight hours in order to allow sufficient CCh uptake for use in carbohydrate production
6-5
-------
through the light-driven process of photosynthesis. Ozone, when present in sufficient amounts,
is taken up along with the CCh, where it and its derivatives can inhibit photosynthesis, leading to
reduced carbohydrate production needed for growth, reproduction and repair (U.S. EPA, 2013,
section 9.3.6; 75 FR 3013). Since plants are photosynthesizing during daylight hours and
continue to grow throughout their growing season, the effects of repeated Os exposures continue
to accumulate, both on a diurnal and seasonal basis. Thus, for vegetation, the element of
"averaging time" has more appropriately been considered in terms of relevant exposure periods -
diurnal and seasonal — over which exposures are cumulated, or summed.
In the EPA's consideration of such exposure periods in both the 1997 and 2008 reviews,
and the 2010 reconsideration, the EPA identified the 12-hour daylight period from 8:00 am to
8:00 pm as appropriately capturing the diurnal window with most relevance to the photosynthetic
process (61 FR 35716; 72 FR 37900; 75 FR 2938, 3013). In so doing, the EPA recognized, as did
CAS AC, that in some parts of the country this period may not include all daytime hours or all
exposures of importance to vegetation, thus potentially underestimating the impact of Os at those
sites (72 FR 37900-01; 75 FR 3013-14; Henderson, 2007, p. 3, pp. C-22-23). The evidence
available in this review continues to provide support for focusing on the daylight hours, since for
the majority of plants, the diurnal conditions of maximum Os uptake occur mainly during the
daytime hours (U.S. EPA, 2013, section 9.5.3.2). This evidence shows that, in general, (1) plants
have the highest stomatal conductance during the daytime; (2) atmospheric turbulent mixing is
greatest during the day in many areas; and (3) the high temperature and high light conditions that
occur during the day and that typically promote the formation of tropospheric Os also promote
physiological activity in vegetation (U.S. EPA, 2013, section 9.5.3.2).
In addition, as in past reviews, we have also considered the evidence available from a
number of studies that have reported Os uptake at night in some species (U.S. EPA, 2013, section
9.5.3.2). Typically the rate of stomatal conductance at night is much lower than during the day.
Across the studies discussed in the ISA, nocturnal conductance ranged from negligible to 25% of
daytime values (U.S. EPA, 2013, section 9.5.3.2), and, in some studies, varied by season and
drought conditions. However, many of these studies did not link the night-time flux to measured
effects on plants, making it difficult to know in those studies whether the impacts on the plant
from nocturnal exposures are greater or less than those from similar daytime exposures, and
whether or not they should be considered as separate impacts or as additive or synergistic with
impacts from the preceding or subsequent daytime exposure.
Further, there are also uncertainties associated with the extent of the occurrence of high
exposure to Os at night. For significant nocturnal stomatal flux and Os effects to occur, a
susceptible plant with nocturnal stomatal conductance and low defenses must be growing in an
6-6
-------
area with relatively high nighttime Os concentrations (often high elevation sites) and appreciable
nocturnal atmospheric turbulence. It is unclear how many areas there are in the U.S. where these
atmospheric conditions occur. It may be possible that these conditions exist in mountainous
areas of southern California, front-range of Colorado and the Great Smoky Mountains of North
Carolina and Tennessee. However, more information is needed in locations with high nighttime
Os to assess the local Os patterns, micrometeorology and responses of potentially vulnerable
plant species (U.S. EPA, 2013, section 9.5.3.2).
In consideration of the uncertainties that remain regarding the importance and extent of
nocturnal exposures associated with plant uptake, and whether and how they might be
incorporated into a national index, we conclude that it is appropriate to continue to focus on the
12-hour daylight exposure period of 8:00 am to 8:00 pm. We note that available monitoring data
indicates that the daily increase in Os concentrations generally does not begin until after 8:00 am
(U.S. EPA, 2013, section 3.6.3.2). In regard to this staff conclusion on an appropriate diurnal
exposure period, CASAC states that "[accumulation over the 08:00 a.m. - 08:00 p. m. daytime
12-hour period is a scientifically acceptable and recommended means of generalizing across
latitudes and seasons" (Frey, 2014a, p. 13).
With regard to identification of the seasonal period over which to cumulate exposures, we
note that a plant is vulnerable to Os pollution as long as it has foliage and is physiologically
active (U.S. EPA, 2013, section 9.5.3, p. 9-112), i.e., during its growing season. The length of
vegetative growing seasons varies depending on the type or species of vegetation and where it
grows. For example, as discussed in the ISA, annual crops are typically grown for periods of
two to three months while perennial species may be photosynthetically active longer, up to 12
months each year for some species (U.S. EPA 2013, section 9.5.3, p. 9-112). In general, the
period of maximum physiological activity and thus potential Os uptake for vegetation coincides
with some or all of the intra-annual period defined as the Os season, which can vary on a state-
by-state basis (U.S. EPA, 2013, Figure 3-24, p. 3-83). This is because the high temperature and
high light conditions, which can vary geographically, typically promote the formation of
tropospheric Cb, as well as physiological activity in vegetation (U.S. EPA, 2013, section 9.5.3, p.
9-112).
The exposure periods used in studies of Os effects on vegetation reflect this
understanding, with crop studies typically using shorter seasonal exposure periods and studies of
longer lived trees and other perennial vegetation often extending for the entire annual growing
season or in some cases over multiple growing seasons. Specifically, the ISA notes that "[m]ost
of the crop studies done through NCLAN had exposures less than three months with an average
of 77 days. Open-top chamber studies of tree seedlings, compiled by the EPA, had an average
exposure of just over three months or 99 days. In more recent FACE experiments, SoyFACE
6-7
-------
exposed soybeans for an average of approximately 120 days per year and the Aspen FACE
experiment exposed trees to an average of approximately 145 days per year of elevated Cb,
which included the entire growing season at those particular sites" (U.S, EPA, 2013, section
9.5.3.2, p. 9-112). Further, the U.S. Forest Service and federal land managers have typically
used the 6 months from April through September as the accumulation period (U.S, EPA, 2013,
section 9.5.3.2, p. 9-112). However, despite the possibility that plants may be exposed to
ambient Os longer than 3 months in some locations, the ISA notes that "[t]he exposure period in
the vast majority of Os exposure studies conducted in the U.S. has been much shorter than 6
months..." and "there is generally a lack of exposure experiments conducted for longer than 3
months" (U.S. EPA, 2013, section 9.5.3.2, p. 9-112). As a result, analyses of effects in terms of
the W126 exposure index have typically defined the index in terms of a 3-month exposure period
or at least in terms of periods shorter than 6 months (e.g., SoyFACE, Aspen FACE) (U.S, EPA,
2013, p. 9-112).
In the current review, the EPA conducted a new analysis to further inform the
consideration of the most appropriate seasonal accumulation period (U.S. EPA, 2013, section
9.5.3). This analysis calculated and compared the 3- and 6-month maximum W126 index values
for over 1,200 AQS and CASTNET EPA monitoring sites for the years 2008-2009. The two
accumulation periods were found to be highly correlated metrics (U.S. EPA, 2013, Figure 9-13;
section 9.5.3). The analysis indicates that in the U.S., W126 cumulated over 3 months and W126
cumulated over 6 months could be proxies of one another, as long as the period in which daily
W126 is accumulated corresponds to the seasonal maximum. Therefore, it is expected that either
statistic will predict vegetation response equally well. Given the above information, and in
particular the results of the EPA analysis showing the maximum 3-month period is highly
correlated with the longer 6-month maximum period, we again conclude that it is appropriate to
continue to focus on the consecutive 3-month period with the highest cumulative exposure value
within the monitored Os season as the seasonal exposure period with most relevance to
vegetation. Given its review of the available science, CAS AC also expressed support for this
seasonal period, stating that "[t]he Second Draft PA makes a very strong case, consistent with
previous CASAC judgment, for changing the secondary metric to the W126 averaged over the
highest three-month interval" (Frey, 2014a, p. 13).
6.3 LEVEL
In considering potential levels for alternative secondary standards, we again find it useful
to note that the protection provided by the secondary standard derives from the combination of
all elements of the standard (indicator, form, averaging time, and level). Thus, in light of the
discussions in section 6.2 above, we first consider what level or range of levels can reasonably be
6-8
-------
judged to provide a requisite degree of public welfare protection when combined with a W126
index form of cumulatively weighted concentrations from 8:00 am to 8:00 pm over a maximum
consecutive 3-month period. In addition to considering the information in the context of a single
growing season, we also consider it in the context of a form for this W126 metric averaged
across three consecutive growing seasons for reasons discussed in section 6.2 above.
In the discussion below, we turn first to consideration of the currently available
scientific evidence as assessed and characterized in the ISA. We then consider the WREA
findings with regard to vegetation, ecosystem effects and services estimated for different air
quality scenarios. We additionally take note of important uncertainties and limitations in the
evidence and exposure/risk analyses, as well as considerations related to interpreting these
impacts in light of the additional policy considerations described in the adversity paradigm.
Lastly, we take note of judgments to be made by the Administrator in drawing conclusions
regarding effects and risks that represent adverse effects to public welfare. In so doing, we
identify key considerations with regard to the currently available evidence, exposure/risk
information and associated uncertainties in identifying the range of levels that may be
appropriate to consider for a cumulative seasonal secondary standard. Such levels are described
in section 6.5 below, which describes staff conclusions regarding alternative secondary standards
appropriate to consider in this review.
• What does the currently available evidence indicate with regard to the range of
W126-based index values that may provide protection from vegetation effects of
03?
In answering this question, we first consider quantitative evidence for Os exposure effects
on plant growth, productivity and related endpoints. In so doing, we draw primarily on the
robust E-R functions developed in OTC studies for tree seedling and crop species as described in
the ISA (U.S. EPA 2013, section 9.6), and as used in the WREA exposure and risk analyses
(U.S. EPA, 2014b, section 6.2), and discussed in Chapter 5 of this document (Figures 5-1 and 5-
4). It is important to note that these functions are used to provide estimates of growth and yield
reduction in tree seedlings and crops that might be expected to result from exposure over a single
growing season to various Os concentrations expressed in terms of a W126 index. We also
consider the available evidence and exposure/risk information for visible foliar injury.
As a point of clarification, we note that CAS AC commented that it "concurs that relative
biomass loss for tree species, crop yield loss, and visible foliar injury are appropriate surrogates
of a wide range of damage that is adverse to public welfare." (Frey, 2014a, p. 10). While we
agree that effects at the individual tree, crop, or other plant species level, in and of themselves,
can be directly related to effects on the public welfare when they occur to a sufficient degree on
lands with an intended use that can be affected by Os-induced vegetation effects (e.g., timber
6-9
-------
production, AQRVs in Class I areas), we also caution that not all predicted effects on vegetation
occur to such a degree or occur on lands within this category. Thus, in considering the predicted
effects on studied tree and crop species under various W126 exposures, we are mindful of the
need to further determine under what conditions they can be considered surrogates for impacts
that are important in the public welfare policy context.
Table 6-1 below presents estimates of relative biomass and yield loss for the 11 and 10
studied species of tree seedlings and crops, respectively, for which we have robust E-R functions
developed in OTC studies, for a single growing season exposure to a number of W126 index
values. In this table, we have included observations related to median and individual species
relative biomass loss in tree seedlings and relative crop yield loss, at the target benchmark levels
of 2% and 5%, respectively. These benchmarks are consistent with the 2% and 5% benchmarks
for tree seedlings and crops, respectively, as advised by CASAC in this review (Frey, 2014a;
section 6.4 below), and with values given focus in the 1996 expert consensus workshop. We
have also included information on the number of studied species with estimates below other
benchmarks that may also be of interest (i.e., 5%, 10%, and 15% for trees and 10% for crops).
CASAC has placed most emphasis on the median species response in recommending a range of
scientifically supportable levels. For example, CASAC noted that "[i]n our scientific judgment, it
is appropriate to identify a range of levels of alternative W126-based standards that includes
levels that aim for not greater than 2% RBL for the median tree species" (Frey, 2014a, p. 14)."
CASAC also recognizes that as a policy matter the Administrator may find it useful to also
consider information related to individual species responses, to the degree that they have special
significance to the public welfare, when selecting an appropriate level or range of levels.
Specifically, CASAC states that "[a]s a policy recommendation, separate from its advice above
6-10
-------
Table 6-1. Tree seedling biomass loss and crop yield loss estimated for
season.
exposure over a
W126 value
for exposure
period
21 ppm-hrs
19ppm-hrs
17 ppm-hrs
15 ppm-hrs
13 ppm-hrs
11 ppm-hrs
9 ppm-hrs
7 ppm-hrs
Tree seedling biomass lossA
Median Value
Median species w.
6.8% loss B
Median species w.
6.0% loss B
Median species w.
5.3% loss B
Median species w.
4.5% loss B
Median species w.
3.8% loss B
Median species w.
3.1% loss6
Median species w.
2.4% loss B
Median species w.
<2% loss B
Individual Species
< 2% loss: 3/11 species
< 5% loss: 5/1 1 species
<10% loss: 7/11 species
< 15% loss: 10/11 species
>40%loss: 1/11 species
< 2% loss: 3/11 species
<5% loss: 5/1 1 species
< 10% loss: 7/11 species
< 15% loss: 10/11 species
>30%loss: 1/11 species
< 2% loss: 5/11 species
<5% loss: 5/1 1 species
< 10% loss: 9/11 species
< 15% loss: 10/11 species
>30%loss: 1/11 species
< 2% loss: 5/11 species
<5% loss: 6/1 1 species
< 10% loss: 10/11 species
>30%loss: 1/11 species
< 2% loss: 5/11 species
<5% loss: 7/1 1 species
< 10% loss: 10/11 species
>20%loss: 1/11 species
< 2% loss: 5/11 species
<5% loss: 8/1 1 species
< 10% loss: 10/11 species
>20%loss: 1/11 species
< 2% loss: 5/11 species
<5% loss: 10/11 species
>20%loss: 1/11 species
< 2% loss: 7/11 species
<5% loss: 10/11 species
>15% loss: 1/11 species
Crop yield lossc
Median Value
Median species w.
7.7 % loss D
Median species w.
6.4 % loss D
Median species w.
5.1 %loss°
Median species
w.<5% loss D
Median species
w.<5% loss D
Median species
w. <5% loss D
Median species
w. <5% loss D
Median species
w. <5% loss D
Individual Species
< 5% loss: 4/10 species
>5,< 10% loss: 3/10 species
>10,<20% loss: 3/10 species
< 5% loss: 5/10 species
>5,<10% loss: 3/10 species
>10,<20% loss: 2/10 species
< 5% loss: 5/10 species
>5,<10% loss: 3/10 species
>10,<20% loss: 2/10 species
< 5% loss: 6/10 species
>5,<10% loss: 4/10 species
< 5% loss: 6/10 species
>5,<10% loss: 4/10 species
< 5% loss: 9/10 species
>5, <10%loss: 1/10 species
< 5% loss: all species
< 5% loss: all species
A Estimates here are based on the 1 1 E-R functions for tree seedlings described in WREA, Appendix 6F and discussed in section
5.2.1, with the exclusion of cottonwood. See CASAC comments (Frey, 2014a).
B This median value is the median of the composite E-R functions for 1 1 tree species in the WREA, Appendix 6F (also discussed
in section 5.2.1).
C Estimates here are based on the 10 E-R functions for crops described in Appendix 6F and discussed in section 5.3.1.
D This median value is the median of the composite E-R functions for 10 crops from WREA, Appendix 6F (also discussed in
section 5.3.1).
6-11
-------
regarding scientific findings, the CASAC advises that a level of 15 ppm-hrs for the highest 3-
month sum in a single year is requisite to protect crop yield loss, but that lower levels provide
additional protection against crop yield loss. Furthermore, there are specific economically
significant crops, such as soybeans, that may not be protected at 15 ppm-hrs but would be
protected at lower levels" (Frey, 2014a, p. iii).
From Table 6-1, we see that median tree species biomass loss is at or below 2% only at
the lowest W126 index value assessed, 7 ppm-hrs. As the W126 index value is incrementally
increased, median RBL also increases incrementally, so that at W126 index values of 9, 11, 13,
15, and 17, the median RBL increases to 2.4%, 3.1%, 3.8%, 4.5%, and 5.3%, respectively. Thus
over the W126 range of 7 to 17 ppm-hrs, median species biomass loss ranges from
approximately 2% to approximately 5%.
We also believe it is informative to examine the individual species responses and RBL
over the same W126 range (7 to 17 ppm-hrs). We first note, based on Figure 5-1 (B) in chapter
5, that over this range, five species maintain RBLs of less than 2%. These more tolerant species
include Douglas fir, loblolly pine, Virginia pine, sugar maple and red maple. Thus, little
additional protection would be achieved for these species below the W126 index value of 17
ppm-hrs. Two of these species (red maple and sugar maple) would only exceed 2% RBL at 21
ppm-hrs. In contrast, black cherry, the most sensitive of the remaining six species, has RBL
ranging from approximately 36% at 17 ppm-hrs down to approximately 17% at 7 ppm-hrs.
Thus, given that the magnitude of predicted black cherry RBL could be judged adverse over this
range, it is not clear to what extent this information informs the selection of an appropriate level
(Table 6-1; U.S. EPA, 2014b, section 6.2, Appendix 6A), though clearly protection would be
expected to be greater at lower W126 index values. We further note that CASAC based their
recommendation of an appropriate W126 level by considering the median tree species RBL of no
more than 2%, but some levels within CASAC's recommended range allow for the possibility
for individual species RBL to go much higher.
Because Table 6-1 was updated in this final PA to deemphasize cottonwood, based on
staff s understanding of CASAC advice in that regard, we note that the CASAC advice based on
the numbers of species protected to no more than 2% RBL and median RBL values for tree
seedlings associated with various W126 levels, as shown in the in the second draft PA table
(U.S. EPA, 2014a), is no longer consistent in some cases with the revised Table 6-1. For
example, in commenting on the version of Table 6-1 in the second draft PA CASAC states that
"[tjable 6-1 presents the RBL results for individual species for different levels of W126. This
table demonstrates that a range of 7 ppm-hrs to 15 ppm-hrs will protect against RBL of 2% for at
6-12
-------
least 5 of the 12 species"2 (Frey, 2014a, p. 14). In addition, CASAC states that "[t]he CASAC
does not support a level higher than 15 ppm-hrs. For example, at 17 ppm-hrs, the median tree
species has 6% relative biomass loss.... These levels are unacceptably high" (Frey, 2014a, p. iii).
While we continue to place weight on CASAC's scientific judgments that a 6% median RBL is
unacceptably high, and that the 2% median RBL is an important benchmark to consider, we also
note that the updated median RBL for a W126 level of 17 ppm-hrs is now 5.3%.
We further note that CASAC does not provide additional clarification regarding its views
on the acceptability of median tree species RBL levels between 2% and 6%, beyond noting that
values closer to the lower end of the range (W126 index value of 7 ppm-hrs) would provide
greater protection for more sensitive tree species, and that the levels within CASAC's
recommended range allow for the possibility for individual species RBL to go much higher than
2% and 6%. Given the nature of this input, we then considered the RBL information available for
the remaining five species (i.e., eastern white pine, aspen, tulip poplar, ponderosa pine, red alder)
to further inform our evaluation of the additional protection that potentially could be achieved at
different W126 levels within the range being considered. We thus note that at the W126 index
value of 17 ppm-hrs, one species (eastern white pine) has RBL above 10% and one species (red
alder) has RBL of 5.3% (below 6%) while the other three fall between approximately 6.7% and
9.8%. At the W126 index value of 15 ppm-hrs, two (i.e., red alder and ponderosa pine) of the
five species fall below 6% RBL, while the remaining 3 species have RBLs that range from 7.4%
to 8.8%. At the W126 index value of 13 ppm-hrs, three species (i.e., tulip poplar, ponderosa
pine, red alder) fall below 6% RBL, while the remaining two have RBLs of 7.0% and 7.1%. At
the W126 index value of 11 ppm-hrs, all five species have RBLs below 6%. Taken together with
the more tolerant species, the proportion of the studied tree species with RBLs below 6% are
6/11, 7/11, 8/11, and 10/11 at W126 index values of 17, 15, 13, and 11 ppm-hrs, respectively.
To the extent the focus is placed on different % RBL benchmarks and the proportion of studied
trees protected at those levels, as well the expected impacts to associated ecosystem services, in
regard to the identification of the appropriate level or range of levels, this information may be
appropriate to consider.
With respect to crops, based on the 10 robust E-R functions (i.e., barley, lettuce, field
corn, grain sorghum, peanut, winter wheat, cotton, soybean, potato and kidney bean) described in
the ISA and additionally analyzed in the WREA (Figure 5-4), Table 6-1 shows that for the
CASAC recommended target benchmark protection level of 5% for median crop relative yield
loss (RYL), W126 index values ranging from 7 to 17 ppm-hrs are protective. However, when
2 We note that the updated table shows that a range of 7 ppm-hrs to 17 ppm-hrs will protect against RBL of
2% for at least 5 of the 11 species.
6-13
-------
individual species are considered over this same range, the proportion of crops protected varies
from 5/10, 6/10, 7/10, 9/10, 10/10, and 10/10 at the W126 levels of 17, 15, 13, 11, 9, and 7 ppm-
hrs. To the extent a given species is judged as having particular importance to the public
welfare, breaking the information down by species can be helpful. For example, less than 5%
yield loss was estimated for soybeans at the W126 index value of 12 ppm-hrs (U.S. EPA 2014,
Figure 6-3). Four of the studied crop species (barley, lettuce, field corn, and grain sorghum) are
more tolerant, with RYL under 1% over the W126 range from 7 to 17 ppm-hrs. Peanut also
remained under 4% RYL over the same W126 range. Other species differed regarding the W126
level at which RYL reached or fell below 5%. Specifically, for winter wheat, cotton, soybean,
kidney bean and potato, the relevant W126 index values at which RYLs were below 5% are 15,
13, 11, 11, and 9 ppm-hrs. As noted in Chapter 5, the significance of these predicted RYLs to
the public welfare could be informed by the recognition that crops are heavily managed to obtain
the desired yield, and the extent to which yield reductions in any specific crop in a particular
location are considered adverse to public welfare could depend on a number of economic factors,
including crop prices, crop substitution, and the welfare importance of relative changes in
consumer and producer surplus.
With respect to considering visible foliar injury, and its ability to inform selection of an
appropriate target level or range of levels for public welfare protection, we first recognize its
value as a long-standing and well-established bioindicator of Os exposure, as described in the
ISA (U.S. EPA 2013, section 9.4.2). In addition to the role of visible foliar injury as an
indicator, we note that the aesthetic aspects of visible foliar injury itself have the potential to be
important to public welfare (as described in section 5.4). CAS AC also "concurs that visible
foliar injury can impact public welfare by damaging or impairing the intended use or service of a
resource. Visible foliar injury that is adverse to public welfare can include: visible damage to
ornamental or leafy crops that affects their economic value, yield, or usability; visible damage to
plants with special cultural significance; and visible damage to species occurring in natural
settings valued for scenic beauty or recreational appeal" (Frey, 2014a, p. 10). In this regard, we
first note that several tribes have identified a number of Os sensitive species that are important to
their cultural practices (Appendix 5A). These species have many cultural uses such as food,
medicines, dyes, tools/textiles, spiritual, and commercial. In addition, visible foliar injury has
been identified by the federal land managers (FLMs) as a diagnostic tool for informing
conclusions regarding potential ozone impacts on potentially sensitive AQRVs (USFS, NPS,
FWS, 2010), and the evidence shows that injury has been documented in such areas under recent
air quality conditions.
6-14
-------
Despite its recognized importance to the public welfare as a general matter, we are
unaware of any injury benchmarks or criteria that have been identified by the FLMs as to what
extent and/or severity of observed foliar injury warrants protection efforts. In considering
CASAC comments in this regard we note that while it states that "[a] level below 10 ppm-hrs is
required to reduce foliar injury" (Frey, 2014a, p. iii), CASAC does not provide any additional
information regarding the public welfare significance of this degree of injury or what an
appropriate target benchmark or range of benchmarks would be for foliar injury in relation to
what could be considered adverse to the public welfare. Given that there is substantial variability
in this endpoint, such that "the degree and extent of visible foliar injury development varies from
year to year and site to site ... even among co-members of a population exposed to similar Os
levels, due to the influence of co-occurring environmental and genetic factors" (U.S. EPA, 2013,
p. 9-38), staff recognizes the lack of a consistent or generally predictable relationship between
particular W126 exposures and visible foliar injury incidence. We additionally note uncertainty
in what can be concluded from foliar injury in relation to plant health, productivity and
ecological function as "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" (U.S. EPA, 2013, p. 9-39). However, we do recognize the Congressional mandate,
provided in the CAA amendments of 1977 that establish additional protections for Class I areas.
The 1997 Consensus Workshop Report (Heck and Cowling, 1997) discussed below, noted the
potential for visible foliar injury to occur at very low levels. CASAC also stated that "[v]isible
foliar injury is even more sensitive than RBL of 2%, with W126 index values below 10 ppm-hrs
required to reduce the number of sites showing visible foliar symptoms" (Frey, 2014, p. 14).We
further note that the information discussed here regarding incidence of visible foliar injury does
not include information regarding the severity of the observed symptoms and the degree to which
the public welfare impacts from different severity benchmarks might vary. Thus, there is
additional uncertainty regarding the potential variability in the severity of the symptoms across
species and locations and to what degree this would affect the public welfare significance of
these effects so that the appropriate range of W126 index values to protect against this effect is
difficult and complicated to identify.
In further considering the available information pertaining to the question above, we
additionally recognize conclusions that have been drawn by expert committees with regard to
these endpoints (i.e., tree seedling growth, crop yields and visible foliar injury). For example, in
their review of staff documents during the Os NAAQS review completed in 1997, the CASAC
Os panel members expressed a wide range of opinions on aspects of the evidence important to
consider in judging the adequacy of the Os secondary standard and in considering the form and
level that would be appropriate for a secondary Os standard (Wolff, 1996). Subsequent to
6-15
-------
CAS AC meetings in 1995 on this topic, a consensus-building workshop sponsored by the
Southern Oxidant Study group was held on the topic of the Os secondary standard in January
1996 (Heck and Cowling, 1997). This workshop was attended by 16 scientists with backgrounds
in agriculture, managed forest, natural systems, and air quality, all of whom were leaders in their
fields and whose research formed the basis of much of the research examined in the 1996
Criteria Document. These scientists expressed their judgments on what standard level(s) would
provide vegetation with protection from Os-related adverse effects that would be adequate, in
their view.3'4 As the 1997 workshop publication indicates, the scientists at the 1996 workshop
also reached consensus views regarding the types of exposures that were important in eliciting
plant response and the types of metrics that were best at predicting these responses (Heck and
Cowling, 1997). Before coming to agreement on daily and seasonal durations and forms
pertinent to a distinct secondary standard, the participants discussed and identified endpoints to
consider for natural, forest and agricultural ecosystems.5 With regard to form of the standard,
participants concurred with either the SUM06 or W126 metrics, with consensus finally reached
for SUM06, with some qualification regarding implications for a threshold. The participants
identified the ranges they felt should be considered for each of three endpoints. Overall, the
SUM06 values ranged from 8 to 20 ppm-hrs corresponding to W126 index values ranging from 5
to 17 ppm-hrs, based on the EPA analysis focused on conditions in NCLAN studies.6 This
overall range reflected ranges for each of the three endpoints, with the following considerations
(Heck and Cowling, 1997).
- Crops (yield reductions): SUM06 of 15-20 ppm-hrs (13 to 17 ppm-hrs, W126).
This range was recognized to generally consider <10% yield loss in more than
75% of species.
- Trees (growth effects): SUM06 of 10-16 ppm-hrs (7 to 14 ppm-hrs, W126). This
range was recognized to generally consider 1-2% per year growth reduction; in so
3 At the time of the workshop, the secondary Os standard being reviewed by EPA was a 1-hour average of
0.12 ppm (identical to the primary standard at that time). In 1997, EPA concluded the review by revising both
standards to a longer averaging time of 8 hours with a level of 0.08 ppm (62 FR 38856).
4 The workshop publication describes the primary objective for the workshop as having been to assemble
knowledgeable scientists to develop a group consensus on "various critical components associated with a possible
revised secondary ozone standard" (Heck and Cowling, 1997).
5 For natural ecosystems, they focused on foliar injury as an indicator. For forest ecosystems, they
concluded the data did not support selection of an indicator of effects on forest structure or function. As a result,
they identified two indicators pertinent to the systems: growth effects on seedlings from species of natural forest
stands (1-2% per year reduction), and growth effects on seedlings and saplings from tree plantations (1-2% per year
reduction). For agricultural systems, the participants focused on protection against crop yield reductions, with their
acknowledgment of high uncertainties at 5% leading them to a crop yield endpoint of 10% yield reduction (Heck
and Cowling, 1997).
6During the last review, W126 index values corresponding to the SUM06 values cited in the report were
estimated using the NCLAN crop loss data, a key dataset considered by workshop participants (see Appendix 7B of
2007 Staff Paper; Appendix 6 A of this document).
6-16
-------
doing, the group identified a need to consider the potential for year-to-year
compounding of impacts in long-lived perennial species.
- Visible Foliar Injury: SUM06 of 8 to 12 ppm-hrs (5 to 9 ppm-hrs, W126).
Since the publication of 1996 workshop report and conclusion of the 1997 NAAQS
review, the evidence base has continued to expand as described in the 2006 CD and ISA (U.S.
EPA, 2006; U.S. EPA, 2013). With regard to tree growth effects and crop yield reductions,
results of additional studies conducted in the field have confirmed the tree seedling biomass loss
and crop yield loss E-R relationships derived from earlier studies that used OTC (U.S. EPA
2013, section 9.6).
In the 2008 review, CASAC provided comments related to a cumulative seasonal
secondary standard in the context of their comments on the draft and final Staff Papers and on
the final decision (Henderson, 2006; Henderson, 2007; Henderson, 2008). In all instances, they
conveyed support for establishment of a distinct secondary standard with a cumulative seasonal
form. While the EPA, in the 2007 Staff Paper and 2007 notice of proposed rulemaking,
recognized a broader range of W126 index values as appropriate for consideration with regard to
a distinct secondary standard, the CASAC Panel focused on a range they described as
approximately equivalent to that identified by the 1996 workshop participants (Henderson, 2007,
pp. 3, C-27).7 In the CASAC Panel 2006-2007 advice on levels for such a standard, their
suggestion was a focus on levels for a W126 index approximately equivalent to a SUM06 range
of 10 to 20 ppm-hrs (Henderson, 2006, 2007, 2008), which they estimated in 2007 to be a range
from 7 (or 7.5) to 15 ppm-hrs. Based on their consideration of the information available in that
review (with regard to potential magnitude of effects across multiple years), the CASAC Panel
further advised that "[i]f multi-year averaging is employed to increase the stability of the
secondary standard, the level of the standard should be revised downward to assure that the
desired threshold is not exceeded in individual years" (Henderson, 2007, p. 3). The CASAC
advice provided on the 2010 proposed reconsideration and in this review is summarized in
section 6.4 below.
In considering the evidence briefly summarized above in the context of levels for a
W126-based standard, we recognize that given the different types of Os-induced effects, genetic
variability within and between species, and environmental modifiers of effects that also
contribute to variability, it is not feasible to identify a range of cumulative seasonal exposures
from the vegetation effects evidence which would provide a consistent degree of protection for
7 Appendix C of the March 26, 2007 CASAC letter (Henderson, 2007) used a 2001 ambient concentration
dataset and other factors, rather than study data considered in the 1996 workshop, in estimating an "equivalency"
between the two indices.
6-17
-------
all species. Thus, in our consideration of the evidence, we note the importance of considering
several dimensions that pertain to judgments regarding public welfare significance. For
example, we take note of the usefulness of considering the cumulative seasonal exposure at
which the median species response or the majority of the species' responses are expected to be
below minimal response benchmarks of interest and at which only a very few species' responses
are expected to exceed more substantial response benchmarks. Before articulating such
considerations with regard to specific benchmarks and index values, we first consider the WREA
findings in the context of the following question.
• What are the nature and magnitude of risks to vegetation estimated for the
average W126 index scenarios evaluated in the WREA, and what is the
magnitude of risk reduction from risks estimated for air quality conditions
estimated for the current standard?
The WREA provides a characterization of ambient Os exposure and its relationship to
ecological effects, and estimates of the resulting impacts to several ecosystem services. In
considering the question posed above, we focus particularly on WREA estimates related to Os
effects on plant biomass and associated ecosystem services effects. The WREA analyses provide
information on the geographical extent of the effects of Os exposure on plant biomass for
different air quality scenarios. We also note the relationships among effects on individual plants
to other ecosystem components and functions, such as carbon sequestration and air pollutant
removal (U.S. EPA 2013, section 9.4.3.4; U.S. EPA, 2014b, sections 6.6 and 6.7), as well as
market responses to changes in timber and agricultural production (U.S. EPA, 2014b, sections
6.3 and 6.5). We additionally recognize the potential for Os to impact other biomass-related
responses, such as the supply of non-timber forest products and other ecosystem responses for
which we have primarily qualitative characterizations of impacts (U.S. EPA, 2014b, chapter 5).
We turn first to the WREA estimates for a range of effects related to biomass loss, which
are based on application of the robust E-R functions for seedlings of 11 tree species described in
the ISA (U.S. EPA, 2013, section 9.6.2) and the WREA (U.S. EPA, 2014b, Appendix 6F)8.
First, we note (as considered above) the range of responses for the individual species for which
robust E-R functions have been developed. These eleven species vary appreciably in sensitivity
of growth reduction (in terms of relative biomass loss, or RBL) in response to Os exposure.
8 There is an E-R function available for a 12th tree species (cottonwood), but this E-R function is considered
less robust because it is based on the results of a single gradient study (Gregg et al, 2003). That combined with its
apparent extreme response to Os prompted CAS AC to advise the Administrator to not place too much emphasis on
cottonwood in the review of the secondary standard (Frey, 2014, p. 10). As a result, while we do include
cottonwood in some of the analyses, we have decided it would not be appropriate to put less weight on the
cottonwood biomass loss estimates when considering what levels of W126 should be considered protective of
median species biomass loss.
6-18
-------
Based on the 11 individual tree species with robust seedling E-R functions, six of the 11 species
show 2% seedling biomass loss at a W126 index value below 8 ppm-hrs and in the other five
species at a W126 index value above 18 ppm-hrs. Within the group of six more sensitive
species, the most sensitive is black cherry (see Figure 5-1B).
In Appendix 6F, the WREA presents individual and median response across the studied
tree and crop species (U.S. EPA 2014, Appendix 6F). This appendix includes an analysis of the
median of the composite exposure-response (E-R) functions for tree seedlings and crops.
Specifically, Tables 6F-1 and 6F-2 provide estimates of the relative loss for trees and crops
respectively at various W126 index values using the composite E-R functions for each species.
The median of the composite functions is calculated for all 12 tree species as well as for the 11
tree species excluding cottonwood. The median of the composite functions for all 12 tree species
and all 10 crop species is consistent with the green line shown in Figures 6-5 and 6-6 (U.S. EPA
2014, section 6.2.1.2. Tables 6F-3 and 6F-4 provide estimates of the number of species for trees
and crops respectively that would be below various benchmarks (e.g., 2% biomass loss for trees)
at various W126 index values. Based on the median composite E-R function developed for the
11 tree species depicted in WREA Table 6F-1, median tree species biomass loss ranges from less
than 1.5% to 5.3% over the W126 index value range of 7 to 17 ppm-hrs (U.S. EPA, 2014b,
Appendix 6F).
We additionally consider the WREA estimates of overall ecosystem-level effects from
biomass loss considering the studied species together (U.S. EPA 2014, section 6.8). The WREA
analysis used the species-specific biomass loss E-R functions, information on prevalence of the
studied species across the U.S., and a weighting approach based on proportion of the basal area
within each grid cell that each species contributes. The WREA analyses use information from
the individual and median E-R functions for tree seedlings to provide information on the
geographical extent of the effects of Os exposure on growth reduction for different air quality
scenarios. It provides information on the location and number of species affected, as well as
information about the estimated effects in Class I areas. We note that some of these analyses
continue to include cottonwood and where this is the case, it is so noted. In the WREA analyses,
the largest reductions in Os concentrations occur when air quality is adjusted from recent
conditions to meeting the current standard. Smaller changes in Os concentrations occur when air
quality is adjusted for the W126 air quality scenarios for 15, 11, and 7 ppm-hrs (average across
three years), relative to meeting the current standard.
A weighted RBL value for each grid cell is generated by weighting the RBL value for
each studied tree species found within that grid cell by the proportion of basal area it contributes
to the total basal area of all (unstudied and studied) tree species within the grid cell, and then
6-19
-------
summing those individual weighted RBLs. Table 6-2 below describes the percent of assessed
geographic area with RBL exceeding 2% for 11 species based on the average W126 index values
estimated for five air quality scenarios. Under recent conditions, 7.6 % of the total geographic
area has a wRBL above 2% while just meeting the current standard across the contiguous U.S.,
the WREA estimates 0.2% of the total geographic area to have a weighted relative biomass loss
above 2% for the 7 ppm-hrs scenario (Table 6-2 below; U.S. EPA 2014, Table 6-25). In the
W126 air quality scenarios for 15, 11, and 7 ppm-hrs (average across three years), the percent of
total area having weighted relative biomass loss greater than two percent was 0.2 percent, 0.1
percent and <0.1 percent, respectively (Table 6-2 below; U.S. EPA 2014, Table 6-25). In
considering these estimates, however, we note that the values for percentages of basal area
include many grid cells in which none of the 11 studied species are found, and thus these values
are likely to be low. In addition, the ecosystem level impacts from Os-induced effects on
biomass loss in each grid cell would also depend on the interaction between the studied species
with known Os-sensitivities and the other species that are also contributing to the total basal area
which have unstudied Cb-sensitivities. Given these and other potential uncertainties and
limitations associated with this analysis (U.S. EPA, 2014b, section 6.8), which were also
commented on by CASAC (Frey, 2014a, p. A-40), we thus conclude that this analysis does little
to inform the nature and degree of risk likely to be experienced by Os-sensitive species growing
in mixed-species forests, which are wide-spread in the eastern U.S. These values may be more
appropriate for western forests which more often are composed of a single species (i.e.,
ponderosa pine, aspen forests).
Table 6-2. Percent of assessed geographic area exceeding 2% weighted relative biomass
loss in WREA air quality scenarios.
Percent of total
area with
wRBL>2%
Using all 12
Species
Using 11
species
(excluding
cottonwood)
Air Quality Scenarios
Recent Conditions
(2006-2008)
10.8%
7.6%
Conditions just
meeting the
current standard*
0.8 %
0.2%
W126 index scenarios6
15 ppm-hrs
0.7 %
0.2%
11 ppm-hrs
0.5 %
0.1%
7 ppm-hrs
0.2 %
<0.1%
A This analysis uses air quality values that are estimated per model grid cell using the W126 index value assigned to the grid
cell based on application of the VNA method to the monitor-location W126 index values that are the average at that location
across the 3 years of W126 index values for the adjusted dataset that just meets the current standard (4th highest daily
maximum 8-hour concentration, averaged over 3 consecutive years of 75 ppb).
6-20
-------
B The national distribution of W126 index values within model grid-cells for each scenario reflects model-based adjustment of
2006-2008 Os concentrations at monitoring sites such that the average W126 index at the controlling location in each of the
modeling regions just meets the scenario target index value, followed by application of the VNA interpolation methodology
(see U.S. EPA, 2014b, section 4.3.4.1 and Appendix 4A).
To further inform this issue, the WREA characterized the number of counties where the
median RBLs were above 2% (U.S. EPA, 2014b, Table 6-7), as shown in Table 5-5. Given
CASAC's advice to put less emphasis on cottonwood, we focus on the rows of this table that
excluded cottonwood. Under recent conditions, 52% of the counties have median RBLs above
2%. When air quality is adjusted to the current standard, that proportion drops to 8% and further
decreases to 6% for air quality adjusted to just meet a 3-year average W126 level of 7 ppm-hrs.
With respect to median RBL values, of the 239 counties (8% of counties) estimated to have a
median RBL above 2% when meeting the current standard, 203 of those counties have a RBL
above 2% because of the presence of black cherry. Thus, as also discussed above in Section 6.2,
given the magnitude of estimated RBL for black cherry over the entire range assessed, it is not
clear to what extent the information for black cherry informs the selection of an appropriate
level.
In addition, the WREA also characterized the number of counties where one or more
individual studied tree species showed a 2% biomass loss (U.S. EPA, 2014b, Table 6-7), as also
shown in Table 5-5. This is consistent with CAS AC advice that "rather than focusing solely on
the median relative biomass loss (RBL), the number of counties containing sensitive tree species
that are expected to have growth loss of greater than 2% should be quantified" (Frey, 2014a, p.
11). The maximum number of species that exceed 2% RBL in any one county is five species,
which only occurs under recent Os conditions. After meeting the current standard, the maximum
number of species in any one county is four. This information shows that a number of counties
have more than one Os-sensitive species growing in it, potentially together in the same forest
stands, whose RBLs are above 2%. Given CASAC's advice to put less emphasis on cottonwood,
we focus on the rows of this table that excluded cottonwood. Under recent air quality conditions,
the proportion of counties with 1 or more species with an RBL greater than 2% is 78% (2,418
counties). As air quality is adjusted to just meet the current standard and the alternative W126
index value of 7 ppm-hrs, this number drops to 62% and 58%, respectively. We note that of the
1929 counties estimated to have 1 or more species with an RBL greater than 2% when meeting
the current standard9, 1805 of those counties are estimated to have black cherry as the only
species estimated to experience this level of biomass loss. Thus, as also discussed above in
Section 6.2, given the magnitude of estimated RBL for black cherry over the entire range
9 Excluding cottonwood, as discussed above.
6-21
-------
assessed, it is not clear to what extent the information for black cherry informs the selection of an
appropriate level. We next consider the wRBL estimates from the WREA analysis of 145 (of the
155) federally designated Class I areas for which there was sufficient information regarding Os-
sensitive species (U.S. EPA, 2014b, section 6.8.1, Table 6-26, Appendix 6E). These 145 parks
had at least one Os-sensitive tree species for which an E-R function for RBL was available.
Using the E-R functions for the species found within each park, the WREA calculated an average
wRBL value for each park for the 3-year average W126 index values estimated in those locations
for the current standard and three W126 air quality scenarios. Under conditions adjusted to just
meet the current standard, the average wRBL in 2 of the 145 parks is estimated to be above 2%,
as presented in Table 6-3 below. These two parks are Badlands National Park, driven by
sensitivity of cottonwood, and Wind Cave National Park, driven by sensitivity of ponderosa pine.
We compare this estimate to those for the W126 scenarios. For the W126 scenarios of 15 and 11
ppm-hrs, the estimated weighted RBL is greater than 2% in these same two of the 145 parks,
while it is greater than 2% in only 1 park (Wind Cave) for the 7 ppm-hrs scenario.
Table 6-3. Number of Class I areas (of 145 assessed) with weighted relative biomass loss
greater than 2%.
Number of Class 1 areas with
wRBL>2%
Air Quality Scenarios
Conditions just meeting
the current standard*
2
3- Year Average W126 index scenarios6
15 ppm-hrs
2
11 ppm-hrs
2
7 ppm-hrs
1
A The wRBL is estimated per model grid cell (in which there are any of the 12 studied species) from W126 index value assigned
to the grid cell based on application of the VNA method to the monitor-location W126 index values that are the average at that
location across the 3 years of W126 index values for the adjusted dataset that just meets the current standard (4th highest daily
maximum 8-hour concentration, averaged over 3 consecutive years of 75 ppb).
B The national distribution of W126 index values within model grid-cells for each scenario reflects model-based adjustment of
2006-2008 Os concentrations at monitoring sites such that the average W126 index at the controlling location in each of the
modeling regions just meets the scenario target index value, followed by application of the VNA interpolation methodology (see
U.S. EPA, 2014b, section 4.3.4.1 and Appendix 4A).
The WREA estimates of crop yield loss for the modeled air quality scenarios are
summarized in Table 6-4 below (details are provided in U.S. EPA 2014, section 6.5.1 and
Appendix 6B). For the recent air quality conditions scenario, the means for all crops were less
than 5% loss across all states. Crop yield loss estimates for all states were also less than 5% in
the air quality scenario representing conditions just meeting the current standard (U.S. EPAb,
2014, section 6.5.1 and Appendix 6B).
6-22
-------
Table 6-4. Estimated mean yield loss (and range across states) due to
important crops.
exposure for two
Crop
Corn
Soybean
Air r^iolitw ^f*onorirto
rMl \XUQIILy WwCIIQI IWO ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Recent
Conditions
(2006-2008)
<5%c
(0.01-0.88)
<5%
(0.69-8.30)
Conditions just
meeting the current
standard*
<5%
(0.0-0.01)
<5%
(0.01-1.39)
Average W126 index scenarios6
15ppm-hrs
<5%
(0.0-0.01)
<5%
(0.01-1.13)
11 ppm-hrs
<5%
(0.0-0.0)
<5%
(0.01-0.75)
7 ppm-hrs
<5%
(0.0-0.0)
<5%
(0.01-0.59)
A The crop yield loss is estimated per grid cell (and per FASOMGHG region) from W126 index value assigned to the cell based
on application of the VNA method to the monitor-location W126 index values that are the average at that location across the 3
years of W126 index values for the adjusted dataset that just meets the current standard (4th highest daily maximum 8-hour
concentration, averaged over 3 consecutive years of 75 ppb).
B The national distribution of W126 index values within grid cells for each scenario reflects model-based adjustment of 2006-
2008 Os concentrations at monitoring sites such that the average W126 index at the controlling location in each of the
modeling regions just meets the scenario target index value, followed by application of the VNA interpolation methodology (see
U.S. EPA 2014 section 4.3.4.1 and Appendix 4A).
c Mean yield loss is the mean across modeling units. The range presented in parentheses below the mean represents the
minimum and maximum estimates across modeling units (U.S. EPA 2014, Appendix 6B).
The WREA also analyzes market responses to changes in timber and agricultural
production (U.S. EPA, 2014b, sections 6.3 and 6.5). As explained above, however, comparisons
of the WREA's air quality scenarios for the national-scale estimates of timber production and
consumer and producer surpluses are not straightforward to interpret due to market dynamics.
Estimates for the recent conditions and current standard scenarios are compared to the three
W126 scenarios. In general, substantially greater economic surpluses (approximately $51
billion) are estimated from the comparison of the recent conditions (2006-2008) scenario to the
current standard scenario. The vast majority of these economic surpluses are estimated for
agricultural production. Differences of the average W126 scenarios from the current standard
scenario are much smaller (U.S. EPA 2014, Appendix 6B).
Because increases in timber production represent increased tree growth and concurrent
carbon sequestration, we also consider WREA estimates of the potential increase in carbon
storage that potentially could occur for different air quality scenarios (U.S. EPA 2014, section
6.6.1). Comparisons of the W126 scenarios to the current standard scenario with regard to
carbon sequestration estimates do not indicate an appreciable difference for the W126 scenario
of 15 ppm-hrs beyond that achieved by just meeting the current standard. The majority of the
enhanced carbon sequestration potential resulting from increases in forest biomass is predicted to
occur for the W126 scenarios of 11 and 7 ppm-hrs. Over 30 years, the current standard scenario
6-23
-------
projection is 89,184 million metric tons of CCh equivalents (MMtCChe).10 The WREA estimates
additional sequestration potential of 13, 593 and 1,600 MMtCChe, for the W126 scenarios of 15,
11 and 7 ppm-hrs, respectively, as compared to the current standard (U.S. EPA 2014, Table 6-
18). We also take note of the relatively smaller estimates for carbon sequestration associated
with improved crop yields (over 30 years) in the agricultural sector, which indicate little
difference among the different W126 scenarios, beyond that achieved by just meeting the current
standard.
We additionally consider the WREA estimates for five urban areas of how reduced
growth of Cb-sensitive trees in urban forests may affect the ecosystem services of air pollutant
removal and carbon sequestration (U.S. EPA, 2014b, sections 6.6.2 and 6.7 and Appendix 6D).
With regard to air pollutant removal, the WREA estimated metric tons of carbon monoxide,
nitrogen dioxide, ozone and sulfur dioxide removed under the W126 scenarios. In considering
these estimates we note the general assumptions made to estimate order of magnitude effects of
Os removal by trees on Cb concentrations in the five urban areas and the associated uncertainties
(U.S. EPA 2014, sections 6.7 and 6.9 and Appendix 6D). Estimates for all five case study areas
indicate increased pollutant removal from the recent conditions scenario to just meeting the
current standard scenario, with much smaller differences between the current standard and the
three W126 scenarios (Table 6-5 below). The largest difference in carbon sequestration is
between the existing conditions scenario and the current standard scenario (Table 6-5). In
addition to the small differences in W126 index values among the three W126 air quality
scenarios relative to the current standard for these five areas, only 2 or 3 tree species were able to
be assessed in each city. Therefore, these results may underestimate the overall impacts in these
areas and nationally, although other areas of uncertainty (recognized below) may tend to
contribute to the opposite potential (U.S. EPA 2014, Table 6-27).
101 MMtCO2e is equivalent to 208,000 passenger vehicles or the electricity to run 138,000 homes for 1
year as calculated by the EPA Greenhouse Gas Equivalencies Calculator (updated September 2013 and available at
http://www.epa.gov/cleanenergy/energy-resources/calculator.html).
6-24
-------
Table 6-5. Estimated effect of Os-sensitive tree growth-related impacts on the ecosystem
services of air pollutant removal and carbon sequestration in five urban case
study areas.
Case Study Area
Atlanta
Baltimore
Chicago
Syracuse
Tennessee urban
Atlanta
Baltimore
Chicago
Syracuse
Tennessee urban
Air Oii^ilitw Qf*onarinc
^^1 1 \XUCIII LV WwGI ICil 1 vw
Recent
Conditions
(2006-2008)
Conditions just
meeting the
current
standard*
Average W126 index scenarios6
15ppm-hrs
11 ppm-hrs
7 ppm-hrs
Air Pollutant Removal (metric tons, CO, N02, 03, S02)
33,000
8,500
355,000
1,500
474,000
35,800
9,200
359,000
1,700
511,000
35,800
9,200
359,000
1,700
511,000
36,000
9,200
361,000
1,700
516,000
36,300
9,200
365,000
1,700
522,000
Carbon Storage (million metric tons of C02 equivalents, cumulative over 25 years)
1.2
0.5
16.9
0.14
18.0
1.32
0.57
17.05
0.17
19.67
1.32
0.57
17.05
0.17
19.67
1.32
0.57
17.10
0.17
19.89
1.34
0.57
17.21
0.17
20.16
A Results are derived from estimates per model grid cell (in which there are any of the 12 studied species) from W126 index
value assigned to the grid cell based on application of the VNA method to the monitor-location W126 index values that are
the average at that location across the 3 years of W126 index values for the adjusted dataset that just meets the current
standard (4th highest daily maximum 8-hour concentration, averaged over 3 consecutive years of 75 ppb).
B The national distribution of W126 index values within model grid-cells for each scenario reflects model-based adjustment of
2006-2008 Os concentrations at monitoring sites such that the average W126 index at the controlling location in each of the
modeling regions just meets the scenario target index value, followed by application of the VNA interpolation methodology
(see U.S. EPA, 2014b, section 4.3.4.1 and Appendix 4A).
With regard to foliar injury, we take note of the WREA analyses of the nationwide
dataset (2006- 2010) for USFS/FHM biosites described in section 5.4.2 above, including the
observation that the proportion of biosites with injury varies with soil moisture conditions and Os
W126 index values (U.S. EPA 2014, Chapter 7, Figure 7-10). The evidence of Os-attributable
visible foliar injury incidence occurring in USFS/FHM biosites shows that the proportion of
biosites showing foliar injury incidence increases steeply with W126 index values up to
approximately 10 ppm-hrs. At W126 index levels greater than approximately 10 ppm-hrs, the
proportion of sites showing foliar injury incidence is relatively constant.
In reflecting across the range of W126 index values evaluated in various WREA
analyses, we first note the substantial reductions in biomass-related risks estimated for air quality
6-25
-------
adjusted to just meet the current standard scenario. Additional incremental risk reductions are
estimated across the W126 scenarios, although these risk reductions are substantially smaller.
In considering the WREA estimates here, we take note of uncertainties in the adjusted
estimates. Adjustments were made to recent air quality to reflect just meeting the current
standard and three W126 levels. These adjustments were based on air quality modeling
simulations reflecting across-the-board reductions in NOx emissions required to bring the highest
monitor down to the target level in different regions of the country. In some areas, meeting a target
level at the highest monitor in the region had the effect of substantially reducing concentrations
below the targeted level in other parts of the region. This adjustment approach is not meant to
represent an actual control strategy but to provide an approximation of the spatial variability of Os
across an area when just meeting the current standard and three W126 levels.
We also note potential uncertainties in the extent to which the results for each modeled
air quality scenario represent cumulative seasonal Os exposures that would be expected to occur
across the three years represented in each scenario. In general, each scenario is represented by a
dataset of 3-year average W126 index values across the national modeling area. Thus, the results
estimated for the various analyses that use these scenarios do not reflect any year-to-year
variability that would be expected in single year results. Rather, they reflect average estimates
for the three year period modeled. Analyses in the WREA describe the potential for the WREA
estimates to underestimate cumulative biomass-related effects in perennial species (as noted in
sections 6.2 and 5.2.2 above and described in detail in U.S. EPA, 2014b, chapter 6, 6.2.1.4). This
potential for underestimation is recognized in the context of the uncertainties associated with
other aspects of the different analyses in section 6.9 of the WREA (e.g., U.S. EPA, 2014b, Table
6-27). We additionally note that the WREA compounding analyses do not take into account
other variables that can affect the magnitude of these effects in the field. In considering this
information discussed above in the context of identifying levels appropriate to consider for a
W126-based standard, we take note of additional associated uncertainties as discussed under the
following question.
• What are important uncertainties and limitations in the evidence and
exposure/risk analyses?
In considering the evidence and exposure/risk information summarized above and the
weight to place on this information, we are mindful of the uncertainties and limitations
associated with several key aspects of this information. We first consider the uncertainties
associated with the evidence underlying the tree seedling and crop E-R functions, given the
importance of these functions for many of the ecosystem service analyses described in the
6-26
-------
WREA. Several key uncertainties associated with this information are listed below and
described in more detail in the WREA (U.S. EPA, 2014b).
• Uncertainty regarding the extent to which the subset of studied tree and crop species
encompass the total number of Os sensitive species in the nation and to what extent it
is representative of U.S. vegetation as a whole, given that information is available for
only a small fraction of the number of total species of trees and crops grown in the
U.S. (U.S. EPA, 2013, section 9.6, U.S. EPA, 2014b, Table 6-27).
• Uncertainties regarding intra-species variability due to the different numbers of studies
that exist for different species so that the weight of evidence is not the same for each
species. Those species with more than one study show variability in response and E-R
functions. The potential variability in less well-studied species is, however, unknown
(U.S. EPA, 2013, pp. 9-123/125, U.S. EPA, 2014b, section 6.2.1.2, and Table 6-27).
• Uncertainty regarding the extent to which tree seedling E-R functions can be used to
represent mature trees since seedling sensitivity has been shown in some cases to not
reflect mature tree Os sensitivity in the same species (U.S. EPA, 2013, section 9.6,
U.S. EPA, 2014b, section 6.2.1.1 and Tables 6-5 and 6-27).
• Uncertainty in the relationship of Os effects on tree seedlings (e.g., relative biomass
loss) in one or a few growing seasons to effects that might be expected to accrue over
the life of the trees extending into adulthood (U.S. EPA, 2013, pp. 9-52/53, U.S.
EPA, 2014b, section 6.2.1.4 and Table 6-27).
• Uncertainties associated with estimating the national scale ecosystem-level impacts
using weighted relative biomass loss (U.S. EPA, 2014b, section 6.8, and Table 6-27).
• Uncertainties associated with potential biomass loss in federally designated Class I
areas (U.S. EPA, 2014b, section 6.8. and Table 6-27).
Turning to consideration of the air quality conditions estimated for the various air quality
scenarios, we take note of the following uncertainties associated particularly with estimates of Os
exposures in rural areas nationally. These are described more completely in chapter 4 of the
WREA (see for example, U.S. EPA, 2014b, section 4.4) and summarized in chapter 8 of the
REA (U.S. EPA, 2014b, section 8.5).
• Uncertainties in Os exposures due to a lack of rural monitors, especially in the western
U.S. and at high elevation sites.
• Uncertainties associated with the method (VNA) used to interpolate monitor values to
estimate W126 index values in locations without monitors.
• Uncertainties in adjusted estimates of Os concentrations associated with meeting the
current standard and potential alternative W126-based standards.
Numerous ecosystem services assessments were described in the WREA. These
assessments relied heavily on models, which also relied on the inputs of the tree seedling and
crop E-R functions and adjusted air quality estimates. Thus, including the uncertainties from the
6-27
-------
first two categories discussed above, additional uncertainties associated with the ecosystem
services models include the following.
• Uncertainties associated with use of the i-Tree model to estimate pollution removal
and carbon storage in five urban area case studies, including uncertainties in the base
inventory of city trees, the functions used for air pollution removal and carbon storage
(U.S. EPA, 2014b, sections 6.6.2, 6.7, and Table 6-27).
• Uncertainties associated with use of the FASOMGHG model for national timber and
crop production, including use of median E-R functions for crops in FASOM and
crop proxy and forest type assumptions to fill in where there was insufficient data
(U.S. EPA, 2014b, sections 6.3, 6.5, 6.6.1, and Table 6-27).
• Uncertainties associated with use of the FASOMGHG model to estimate national scale
carbon sequestration, including those associated with the functions for carbon
sequestration (U.S. EPA, 2014b, sections 6.2.1.1, 6.6.1, and Table 6-27).
In addition, the WREA estimates the incidence and of Os-induced visible foliar injury,
both at the national and national park scales. Numerous uncertainties are associated with these
assessments and include the following.
• Uncertainties associated with our understanding of the number and sensitivity of Os
sensitive species (U.S. EPA, 2014b, sections 7.2.1, 7.5 and Table 7-22).
• Uncertainties associated with spatial assignment of foliar injury biosite data to 12x12
km grids (U.S. EPA, 2014b, sections 7.2.1, 7.5 and Table 7-22).
• Uncertainties associated with availability of biosite sampling data in some locations in
the western U.S. (U.S. EPA, 2014b, sections 7.2.1, 7.5 and Table 7-22).
• Uncertainties associated with soil moisture threshold for foliar injury (U.S. EPA,
2014b, sections 7.2.2, 7.2.3, 7.5 and Table 7-22).
• Uncertainties associated with spatial resolution of soil moisture data, time period for
soil moisture data, drought categories and the combination of soil moisture and
biosite data (U.S. EPA, 2014b, sections 7.3.3.2, 7.5 and Table 7-22).
• Uncertainties associated with Os exposure data of vegetation and recreational areas
within parks (U.S. EPA, 2014b, sections 7.4, 7.5 and Table 7-22).
• Uncertainties associated with surveys of recreational activities (U.S. EPA, 2014b,
sections 7.1.1.2, 7.5 and Table 7-22).
Additionally, there is uncertainty associated with the extent to which the endpoints and
associated risk estimates considered above represent effects reasonably judged adverse in the
context of public welfare. Despite these uncertainties, the overall body of scientific evidence
underlying the ecological effects and associated ecosystem services evaluated in the WREA is
strong, and the methods used to quantify associated risks are scientifically sound (Frey, 2014b).
6-28
-------
All of these uncertainties are important to considerations below in the context of target levels of
protection with regard to weight to be placed on various lines of evidence and assessment results.
• Are there other aspects of the form that affect consideration of the welfare
protection provided by the level of the cumulative seasonal standard?
Although cumulative, seasonal exposure indices of interest for vegetation effects are
often expressed in terms of a single season, we recognize that it can also be appropriate to
consider a form that is evaluated over a multiple-year period, such as three years (U.S. EPA,
2007; 72 FR 37901; 75 FR 3021). The current form of the secondary standard is a 3-year
average, and we recognize that the protection provided by the secondary standard derives from
the combination of all elements of the standard (indicator, form, averaging time(s), and level).
Thus, we find it appropriate to evaluate the protection that might be afforded by a form limited to
a single year or one that is based on evaluation of exposures across multiple years. Although
cumulative, seasonal exposure indices of interest for vegetation effects are often expressed in
terms of a single season, we recognize that it can also be appropriate to consider a form that is
evaluated over a multiple-year period, such as three years (U.S. EPA, 2007; 72 FR 37901; 75 FR
3021). Accordingly, this discussion explores the information relevant to consider in conjunction
with the above identification of the W126 index form, 12-hour daylight averaging time and
maximum consecutive 3-month seasonal exposure period, and the subsequent discussion on level
below, when considering support in the current information for single and/or multiple-year
options.
We additionally take note of advice from CASAC on this topic in the current and prior
reviews. Specifically, in this review, CASAC stated that it "does not recommend the use of a
three-year averaging period for the secondary standard. We favor a single-year period for
determining the highest three-month summation which will provide more protection for annual
crops and for the anticipated cumulative effects on perennial species. The scientific analyses
considered in this review, and the evidence upon which they are based, are from single-year
results. If, as a policy matter, the Administrator prefers to base the secondary standard on a three-
year averaging period for the purpose of program stability, then the level of the standard should
be revised downward such that the level for the highest three-month summation in any given
year of the three-year period would not exceed the scientifically recommended range of 7 ppm-
hrs to 15 ppm-hrs" (Frey, 2014a, p. iii).
In considering an annual form of a standard, we particularly take note of Cb-induced
vegetation effects that can occur as a result of a single year's exposure. These include visible
foliar injury symptoms, growth reduction in annual and perennial species, and yield loss in
annual crops. The following discussion considers these effects, in the context of their potential
public welfare significance, and in regard to the extent to which a W126-based standard with an
6-29
-------
annual form or one based on evaluation across multiple years may be able to provide appropriate
protection
In the case of foliar injury, the ISA notes that the full body of evidence indicates that
there is wide variability in this endpoint, such that although evidence shows visible foliar injury
can occur under very low cumulative Os concentrations, ".. .the degree and extent of visible
foliar injury development varies from year to year and site to site... even among co-members of
a population exposed to similar Os levels, due to the influence of co-occurring environmental
and genetic factors" (U.S. EPA 2013, section 9.4.2, p. 9-38). In addition, the WREA assessment
of foliar injury showed the difficulty and complexity associated with identifying W126 index
values that would consistently provide appropriate protection on an annual basis for this
endpoint. We thus conclude that there is limited information to discern between the level of
protection provided by an annual form or a 3-year average form of a W126 standard for this
endpoint, and that a multiple year form could be considered to provide a more consistent target
level of protection for this endpoint, given likely fluctuations in annual Os and soil moisture
conditions.
In the case of annual commodity crops, the overall welfare effect of annual changes in
yields due to Os exposures is not straightforward. As noted above, determining at what point Os-
induced crop yield loss becomes adverse to the public welfare is still unclear, given that it is
heavily managed with additional inputs that have their own associated markets and that benefits
can be unevenly distributed between producers and consumers. We thus conclude that there is
limited information to discern between the level of protection provided by an annual form or a 3-
year average form of a W126 standard for this endpoint. As with foliar injury, we thus conclude
that it is appropriate to consider a level of protection for annual commodity crops that would be
achieved, on average, using a multiple year form, to provide a more consistent target, given
likely fluctuations in environmental and economic conditions.
In contrast to impacts on annual species that accrue in the single growing season in which
the Os exposures occur, annual effects in perennial species can be "carried over" into the
subsequent year where they affect growth and reproduction (U.S. EPA, 2013, pp. 9-43 to 9-44
and p. 9-86). In addition, when these effects occur over multiple years due to elevated Os
exposures across several years, they accumulate and potentially compound, increasing the
potential for effects at the ecosystem level and associated ecosystem services that may be of
significance to the public welfare.
Effects from elevated Os years on perennial plants, when they occur over several years,
can be propagated up to higher spatial scales where they can contribute to effects on ecosystem
services, e.g., alteration of below-ground biogeochemical cycles, and alteration of both above-
and below- ground terrestrial community composition and terrestrial ecosystem water cycling
6-30
-------
(U.S. EPA, 2013, Table 9-19). Ozone has also been shown to affect plant reproduction in
numerous ways (U.S. EPA, 2007, 7.3.3.3; U.S. EPA, 2013, 9.4.3.1). These effects, when they
occur at sufficient magnitude for a single species, may result in impaired recruitment and loss of
the species from the stand or community. This has the potential to change the community
composition and biodiversity. If these effects occur in multiple plant species and/or over
multiple years, they can result in a reduction in the productivity and carbon sequestration of
terrestrial ecosystems. Such ecosystem-related effects and others discussed in the ISA may be
considered to reflect impacts of critical Os exposures over the longer term. We additionally note
that as compared to intermittent (or single year) critical Os exposures, multiple years of such
exposures might be expected to result in larger impacts on forested areas, e.g., increased
susceptibility to other stressors such as insect pests, disease, co-occurring pollutants and harsh
weather, due to the potential for compounding or carry-over effects on tree growth.
Given the above, we find it reasonable to conclude that the public welfare significance of
the effects that can accumulate as a result of multiple-year Os exposures have the potential to be
greater and more certain than those that are realized in an individual year. Thus, to the extent
that the focus for public welfare protection is on long-term effects that occur in sensitive tree
species in natural forested ecosystems, including in federally protected areas such as Class I
areas or on lands set aside by states, tribes and public interest groups to provide similar benefits
to the public welfare, a cumulative seasonal standard that evaluates exposures across multiple
years (in combination with an appropriate level) might be a more appropriate match to provide
the requisite protection for those Cb-related effects on vegetation that when accumulated across
years, are potentially significant and adverse to the public welfare.
Additionally, we address the potential for cumulative impacts on biomass loss over a 3-
year period versus a 1-year period. First it is important to note that the WREA analyses that
characterize plant biomass and associated ecosystem services effects, discussed above in this
section, are based on a 3-year average. The WREA analysis examined the potential for biomass
loss estimates based on a 3-year average W126 index value to underestimate the cumulative
impact on growth based on the biomass loss that would be predicted in each of the 3 years, based
on the yearly W126 index values. The results show that the use of the three-year average W126
index value may underestimate RBL values slightly. However, it should be noted that the
approach does not account for moisture levels or other environmental factors that could affect
biomass loss (U.S. EPA, 2014b, section 6.2.1.4 and Figure 6-14). In considering these results,
we note that in these regions and in all three years, the three-year average W126 index value is
sometimes above and sometimes below the individual year W126 index value.
In addition to the vegetation effects considerations described above, there are other
policy-relevant factors that can be useful to consider. For example, under a standard with a
6-31
-------
single year form, a monitor may be judged to meet the standard based on a single year of data,
while under a standard with a form requiring evaluation over a multi-year period, a monitor is
not judged to have met the standard until a complete multi-year record is available. For a W126-
based potential standard, the multi-year form identified for consideration in the last review was
the average cumulative seasonal metric over three consecutive years (75 FR 3027). Such a
multi-year form remains appropriate to consider to provide stability to an alternative secondary
standard, just as the multi-year form provides for the current standard (average over three years
of annual fourth-highest daily maximum 8-hour average Os concentrations).11 In considering the
issue of stability in the context of such a form, we first note the inter-annual variability of
seasonal W126 index, which is not unexpected given the logistic weighting function and also
inter-annual variability in meteorological conditions that contribute to Os formation (see
Appendix 2C). The staff analysis in Appendix 2C describes the variability in annual W126
index values in relation to variability in the 3-year average, which indicates that a standard based
on an annual W126 index would be expected to have a lower degree of year-to-year stability
relative to a standard based on a form that averages seasonal indices across three consecutive
years. A more stable standard can be expected to contribute to greater public welfare protection
by limiting year-to-year disruptions in ongoing control programs that would occur if an area was
frequently shifting in and out of attainment due to extreme year-to-year variations in
meteorological conditions. This greater stability in air quality management programs thus
facilitates achievement of the protection intended by a standard. In light of this relationship, we
conclude that a 3-year average form has the desirable feature of providing greater stability in air
quality management programs and thus facilitating the achievement of the protection intended by
a standard. Thus, we recognize the public welfare benefits of having a standard of a 3-year
average form.
CAS AC has asked that the PA quantify the ratio of the 3-year average of the highest
three-month summations in each year to the highest three-month summation in the highest year
within that same 3-year average period. This information is provided in a technical
memorandum titled "Relationship between W126 annual values and three-year averages" (EPA-
HQ-OAR-2005-0172) and in the analyses included in Appendix 2C. In the technical
memorandum, an analysis summarized the relationship between annual W126 index values and
the three-year averages of the annual values based on 2007-2009 air quality data. Based on the
air quality data, 79 percent of counties meeting a three-year average W126 index value of 13
ppm-hrs would also not have annual W126 index values above 15 ppm-hrs. In addition, in terms
11 See ATA III, 283 F. 3d at 374-75 (recognizing programmatic stability as a legitimate consideration in the
NAAQS standard-setting process).
6-32
-------
of county-years (i.e., the number of counties times the number of years in the analysis), 93
percent of the county-years meeting a three-year average W126 index value of 13-ppm-hrs
would also meet an annual W126 index value of 15 ppm-hrs.
In addition, Appendix 2C compared annual W126 index values to three-year average
W126 index values for 2008-2010 air quality data. It concluded that the data analysis "shows that
the inter-annual variability in the annual W126 index tends to decrease with decreasing W126
levels. Thus, it is expected that reductions in NOx emissions will not only result in lower 3-year
average W126 levels, but also result in less inter-annual variability associated with annual W126
levels." Appendix 2C also concludes that the inter-annual variability in the W126 index increases
and decreases along with the three-year average.
These analyses suggest that meeting a 3-year average W126 index value of 13 ppm-hrs
would mean that for most years and monitoring sites the annual W126 index value would be
below 15 ppm-hrs. In addition, the relationship between 3-year average W126 index values and
annual W126 index values is dynamic and varies with the three-year average W126 index value
and will continue to change in the future with changing pollution levels.
Accordingly, in considering all elements for a revised standard, including level and form,
we note that a standard with a form that averages across three years can also control for year-to-
year variability and individual year concentrations. The appropriate level and form combination
will depend on which effects endpoints are considered to warrant additional public welfare
protection and what is considered to be the requisite range of target levels of protection. In
articulating these objectives it may be appropriate to evaluate the nature of the Os induced effects
and their significance or importance to the public welfare, as well as the role that year-to-year
exposure variability can play in public welfare impacts.
• What considerations may be important to the Administrator's judgments on the
public welfare significance of Os associated vegetation effects that may be
expected under air quality conditions associated with different levels for a
seasonal cumulative standard?
Our consideration of this question is intended to provide a public welfare context for
consideration of the evidence and exposure/risk information discussed above, which includes the
nature and magnitude of observed and predicted effects at various levels of cumulative seasonal
exposures. We also note the importance of considering information in an integrated manner,
rather than focusing only on results from any one analysis. For example, we find it appropriate,
in considering the evidence with regard to seedling growth reduction (or biomass loss), to
consider the WREA estimates of affected area based on tree basal area together with estimates of
individual species responses based simply on the evidence-based E-R functions, and in light of
other potential impacts summarized above. In so doing in section 6.5 below, we take into
6-33
-------
account considerations relevant to public welfare policy judgments required of the
Administrator, such as those described here.
As recognized in sections 1.3.2 and 5.1, the Clean Air Act specifies that secondary
standards specify a level of air quality that is requisite to protect against known or anticipated
adverse effects to public welfare. In the Administrator's judgment as to the standards that would
be requisite (i.e., neither more nor less stringent than necessary) to protect the public welfare
under the Act, she may consider a number of factors including 1) what should be considered to
constitute an adverse effect to the public welfare; 2) the nature and magnitude of the effects and
the risks that remain after meeting the level of the current standard; and, 3) what is necessary to
achieve the requisite (no more and no less) degree of public welfare protection. In the 2008
decision by which the current standard was established, the Administrator considered these
factors in judging the previously existing standard to not provide the requisite public welfare
protection. At that time the Administrator found that the exposure- and risk-based analyses
available in that review indicated that adverse effects to vegetation would be predicted to occur
under air quality conditions associated with just meeting the then-current standard. The effects
identified were "visible foliar injury and seedling and mature tree biomass loss in Os-sensitive
vegetation" (73 FR 16496). In so noting, the Administrator indicated that he believed that "the
degree to which such effects should be considered to be adverse depends on the intended use of
the vegetation and its significance to public welfare" (73 FR 16496). With regard to
consideration of intended use, the Administrator took note of the specific uses of public lands set
aside by Congress and 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 for future generations" such
as Class I areas (73 FR 16496). The Administrator also recognized areas set aside by states,
tribes and public interest groups with the intent "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).12
In the Administrator's judgments in the 2008 review, he did not identify specific criteria
or benchmarks or a specific level of protection from adverse environmental effects to public
welfare judged to be requisite under the Act.13 As noted above, the scientists at the 1996
12 In considering areas that have not been afforded such special protection, ranging from vegetation used
for residential or commercial ornamental purposes, such as land use categories that are heavily managed for
commercial production of commodities such as agricultural crops, timer and ornamental vegetation, the
Administrator indicated his expectation that protection of sensitive natural vegetation and ecosystems might be
expected to also provide some degree of additional protection for heavily managed commercial vegetation (73 FR
16496).
13 In remanding the 2008 decision on the secondary standard back to the EPA, the Court of Appeals for the
D.C. Circuit determined that EPA did not specify what level of air quality was requisite to protect public welfare
6-34
-------
workshop identified ranges of cumulative seasonal index values (e.g., in terms of SUM06 or
W126) in the context of considering a degree of protection for vegetation effects defined in terms
of relative yield loss in crops and relative biomass loss in tree seedlings. Considering this
information in the context of a secondary standard entails policy judgments by the Administrator
with regard to the degree that impacts exceeding these or other benchmarks and other effects
should be judged adverse to the public welfare. In considering levels for a W126-based
secondary standard that may be appropriate to consider, we recognize that the statute requires
that a secondary standard be protective against only those known or anticipated Cb effects that
are "adverse" to the public welfare, not all identifiable Cb-induced effects. Thus, we recognize
both the importance of scientific consensus statements that have been made regarding
vegetation-related endpoints and Cb exposure levels that might protect against such key
endpoints and the importance of placing such conclusions in the context of consideration of the
public welfare more broadly.
As discussed in section 5.1 and recognized by the EPA in prior reviews, staff recognizes
the importance of a more expansive construct or paradigm that addresses what constitutes
adverse effects of Cbto public welfare. In so doing, we also recognize several aspects or
dimensions of vegetation effects for consideration within this paradigm. These include the
likelihood, type, magnitude, and spatial scale of the effect, as well as the potential for recovery
and any uncertainties relating to these conditions (77 FR 20231). As in the last review, we also
continue to recognize that the public welfare significance of Cb-induced effects on sensitive
vegetation growing within the U.S. can vary, 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. Any given Cb-related effect on vegetation and
ecosystems (e.g., biomass loss, foliar injury), therefore, may be judged to have a different degree
of impact on the public welfare depending, for example, on whether that effect occurs in a Class
I area, a city park, or in commercial cropland. In the 2008 review, the Administrator judged it
appropriate that this variation in the significance of Cb-related vegetation effects should be taken
into consideration in judging the level of ambient Cb that is requisite to protect the public welfare
from any known or anticipated adverse effects (73 FR 16496). For example, in considering
visible foliar injury and seedling and mature tree biomass loss in Cb-sensitive vegetation
expected under alternative air quality scenarios, the Administrator noted that "the degree to
which such effects should be considered to be adverse depends on the intended use of the
vegetation and its significance to the public welfare" (73 FR 16496). Further, the rulemaking
from adverse public welfare effects or explain why any such level would be requisite, as described in section 1.2.2
above. Mississippi, 744 F.3d at 272-73.
6-35
-------
notice stated that "[i]n considering what constitutes a vegetation effect that is adverse from a
public welfare perspective, the Administrator believes it is appropriate to continue to rely on the
definition of 'adverse,' ... that imbeds the concept of 'intended use' of the ecological receptors
and resources that are affected, and applies that concept beyond the species level to the
ecosystem level" (73 FR 16496). The notice went on to state that "[i]n so doing, 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). Such public lands that are protected areas of national interest
include national parks and forests, wildlife refuges, and wilderness areas.
We also consider effects on ecosystem services in considering adversity to public
welfare. For example, the WREA has evaluated the economic value of ecosystem services
affected by Os and how those services might be expected to change under different air quality
scenarios representing the current and potential alternative standards (U.S. EPA, 2014b, chapters
6 and 7).
Lastly, we recognize several important considerations in evaluating levels of protection
and levels for a cumulative seasonal W126-based standard including: the extent of areas
expected to be affected nationwide and the magnitude of those effects; the extent of effects in
areas of national significance; the extent to which these impacts might be judged significant from
a public welfare perspective and associated uncertainties in the information. Accordingly, we
recognize that the range of alternative standard levels that may be appropriate to consider differs
based on the weight placed on different aspects of the evidence and on different aspects of the
quantitative exposure/risk information, and the associated uncertainties, as well as on public
welfare policy decisions regarding the public welfare significance of the effects considered and
the approaches for considering benchmarks for growth or biomass loss and other vegetation
effects of Os. As described in chapter 1, our objective is to identify the range of policy options
supported by the current evidence- and exposure/risk-based information and with consideration
of the role of the Administrator's public welfare judgments. In so doing, we recognize support
for consideration of a broad range of W126 index values, which we discuss in section 6.5, with
recognition of the different judgments that might provide support for different parts of such a
range.
6-36
-------
6.4 CONSIDERATION OF PROTECTIVENESS OF REVISED PRIMARY
STANDARD
In staff consideration of the primary standard in chapter 4, staff concludes it is
appropriate to consider alternative primary standards of the same form and averaging time as the
current primary standard and a lower standard level within the range of 60 to 70 ppb. Thus,
although the discussion in this chapter, with regard to the secondary standard, indicates the
appropriateness of considering an alternative secondary standard with a cumulative, seasonal
form, we also recognize that, to the extent that the Administrator may find it effective to control
air quality using the same form for both the primary and secondary standards, it may be practical
to consider the extent to which a standard in the form of the primary standard might be expected
to also reduce and provide protection from cumulative seasonal exposures of concern. For
example, if a clear and robust relationship was found to exist between 8-hour daily peak Os
concentrations and cumulative, seasonal exposures, the averaging time and form of the current
standard might be concluded to have the potential to be effective as a surrogate. In response to
this, we ask the following question:
• What does the available information indicate with regard to protection of welfare
from cumulative Os exposures that might be afforded by alternative secondary
standards based on the form of the current standard (a 3-year average of 4th
highest 8-hour average concentrations)?
Addressing this point, the ISA describes the results of a recent focus study that examined
the diel14 variability in Os concentrations in six rural areas between 2007 and 2009 (U.S. EPA,
2013, pp. 3-131 to 3-133). The ISA reported that "[tjhere was considerable variability in the diel
patterns observed in the six rural focus areas" with the three mountainous eastern sites exhibiting
a "generally flat profile with little hourly variability in the median concentration and the upper
percentiles", while the three western rural areas demonstrated a "clear diel pattern to the hourly
Os data with a peak in concentration in the afternoon similar to those seen in the urban areas",
which was especially obvious at the San Bernardino National Forest site, 90 km east of Los
Angeles at an elevation of 1,384 meters (U.S. EPA, 2013, p. 3-132). Thus, while the western
sites that are influenced by upwind urban plumes may have increased cumulative seasonal values
coincident with increased daily 8-hour peak Os concentrations, this analysis indicates that, in
sites without such an urban influence (the eastern sites in this analysis), such a relationship does
not occur (U.S. EPA, 2013, section 3.6.3.2). Thus, the lack of such a relationship indicates that
in some locations, Os air quality patterns can lead to elevated cumulative, seasonal Os exposures
without the occurrence of elevated daily maximum 8-hour average Ch concentrations (U.S. EPA,
1 involving a 24-hr period
6-37
-------
2013, section 3.6.3.2). Further, staff notes that the prevalence and geographic extent of such
locations is unclear, since as in the last review, there continue to be relatively fewer monitors in
the West, including in high elevation remote sites. In considering the findings of this analysis,
we additionally recognize, however, that the cumulative seasonal values for the eastern rural
sites, where cumulative seasonal Os concentrations appear to be relatively less related to daily
maximum 8-hour concentrations, are lower in general than those of the western, urban-
influenced sites.
In addition to the focus study described in the ISA (U.S. EPA, 2013, section 3.6.3.2), we
considered analyses of air quality monitoring data and air quality modeling analyses. Chapter 2
of this document characterizes recent monitoring data on Os air quality in rural areas. While
approximately 80% of the Os monitoring network is urban focused, about 120 rural monitors are
divided among CASTNET, NCore, and portable ozone monitors (POMs) sites (Chapter 2, pp. 2-
2 to 2-3, Figure 2.1). Specifically, as stated in chapter 2 "[although rural monitoring sites tend
to be less directly affected by anthropogenic pollution sources than urban sites, rural sites can be
affected by transport of Os or Os precursors from upwind urban areas and by local anthropogenic
sources such as motor vehicles, power generation, biomass combustion, or oil and gas
operations" (U.S. EPA, 2013, section 3.6.2.2). In addition, Os tends to persist longer in rural
than in urban areas due to lower rates of chemical scavenging in non-urban environments. At
higher elevations, increased Os concentrations can also result from stratospheric intrusions (U.S.
EPA, 2013, sections 3.4, 3.6.2.2). As a result, Os concentrations measured in some rural sites
can be higher than those measured in nearby urban areas (U.S. EPA, 2013, section 3.6.2.2).
These known differences between urban and rural sites suggest that there is the potential for 8-
hour daily peak Os concentrations and cumulative, seasonal exposures to not correlate well in
those areas. However, while these metrics may not be directly correlated, reductions in NOx
emissions that occur in urban areas to attain primary standards would also have the effect of
reducing downwind, rural concentrations over the season.
In addition, as was done in both the 1997 and 2008 reviews, staff has analyzed
relationships between Os levels in terms of the current averaging time and form and a W126
cumulative form, based on recent air quality data. One analysis describes the W126 index values
and current standard design values at each monitor for two periods: 2001-2003 and 2009-2011
(e.g., Appendix 2B, Figures 2B-2 and 2B-3). This shows that between the two periods, during
which broad scale Os precursor emission reductions occurred, Os concentrations in terms of both
metrics were reduced. There is a fairly strong, positive degree of correlation between the two
6-38
-------
metrics (Appendix 2B).15 Focusing only on the latter dataset (2009-2011), it can be seen that at
monitors just meeting the current standard (3-year average fourth-highest daily maximum 8-hour
average concentration equal to 0.075 ppm), W126 index values (in this case 3-year averages)
varied from less than 3 ppm-hrs to approximately 20 ppm-hrs (Appendix 2B, Figure 2B-3b). At
sites with a 3-year average fourth-highest daily maximum 8-hour average concentration at or
below a potential alternative primary standard level of 70 ppb, 3-year W126 index values were
above 17 ppm-hrs at no monitors, above 15 ppm-hrs at one monitor, and above 13 ppm-hrs at 8
monitors. At sites with a 3-year average fourth-highest daily maximum 8-hour average
concentration at or below a potential alternative primary standard level of 65 ppb, 3-year W126
index values were above 13 ppm-hrs at no monitors, above 11 ppm-hrs at three monitors, and
above 7 ppm-hrs at 9 monitors. The majority of these monitoring sites are located in the West
and Southwest and include the states of Arizona, California, Colorado, Nevada, New Mexico,
and Utah. At sites with a 3-year average fourth-highest daily maximum 8-hour average
concentration at or below a potential alternative primary standard level of 60 ppb, 3-year W126
index values were at or below 7 ppm-hrs at all monitors.
An additional analysis presents the data for sets of recent 3-year periods back to 2006 -
2008 and indicates that among the counties with Os concentrations that met the current standard,
the number of counties with 3-year W126 index values above 15 ppm-hrs ranges from fewer
than 10 to 24 (Appendix 2B, Figure 2B-9). In general during this longer period, W126 index
values above 15 ppm-hrs and meeting the current standard were pre-dominantly in Southwest
region. As the first analysis in Appendix 2B (for the 2001-2003 and 2009-2011 periods)
indicates, monitors in the West and Southwest tend to have higher W126 index values relative to
their design values than do monitors in other regions. This pattern is noteworthy because the
Southwest region has a less dense monitoring network than regions in the Eastern U.S. (see
Figure 2-1), so that the extent to which this pattern occurs throughout these regions is uncertain.
Although single-year W126 index values were not separately analyzed in this analysis of the
monitor data, it indicates appreciable variation in cumulative, seasonal Os concentrations among
monitor locations meeting different levels of a standard of the current form.
Analyses of the WREA air quality scenarios indicate the potential for Os precursor
emission reductions achieving Os concentrations that just meet different 8-hour standards to
produce a significant reduction in 3-year W126 index values. For example, for the current
standard scenario, nearly all adjusted monitors are at or below an estimated 3-year average W126
index value of 15 ppm-hrs (as summarized in section 5.2.2 and described in U.S. EPA, 2014b,
15 Appendix 2B additionally observes that the program implemented for reducing precursor emissions,
especially NOx, appears to have been an effective strategy for lowering both design values and W126 index values.
6-39
-------
Table 4-1). Those monitors above 15 ppm-hrs would be limited to large urban areas in the
southwestern U.S. (i.e., Phoenix, Los Angeles and Denver). When meeting a 4th highest 8-hour
average scenario of 70 ppb averaged across 3 years, nearly all monitors in the U.S. would meet a
3-year W126 index value of 11 ppm-hrs, though some monitors in the southwest would remain
between 11 and 15 ppm-hrs. At 65 ppb, all locations are at or below 11 ppm-hrs. Thus, similar
to the monitoring analysis, the modeling analysis generally indicates reductions in W126 levels
with reduced Os concentrations in terms of the current standard averaging time and form. This
suggests that depending on the level for a standard of the current averaging time and form, a
degree of welfare protection may be afforded. The extent to which such protection provides
adequate public welfare protection additionally depends on the level of protection identified by
the Administrator as requisite to protect the public welfare from any known or anticipated
adverse effects. In so noting, however, we recognize the importance of also considering
uncertainties in both the model-based adjustment analyses and those based on monitoring data.
These uncertainties, including those related to monitor coverage, the extent to which recent data
can be expected to describe future relationships, and modeling approaches16, among others,
should be kept in mind when assessing the strength of this apparent relationship.
6.5 CASACADVICE
In our consideration of potential alternative standards, in addition to the evidence-based,
risk/exposure-based, and air quality information discussed above, we also consider the advice
and recommendations of CAS AC in EPA's proposed 2010 reconsideration of the 2008 decision,
as well as comments received in the current review, in the context of its review of the ISA, and
the WREA and PA. Some of this advice on specific aspects of the evidence and exposure/risk
information has already been discussed in the relevant sections above. This section specifically
considers CASAC's scientific advice on the appropriate form, averaging times and level(s)
associated with a secondary standard and other related science and policy advice. We have
additionally considered public comments received to date, some of which have suggested a lack
of new information to support a distinct secondary standard and others that urge the
consideration of a secondary standard with a cumulative seasonal form using the W126 metric
and a level within the range of 7 to 15 ppm-hrs.17
16 One uncertainty associated with the modeling approach, as noted in Chapter 5, relates to the lowering of
the highest monitored values as a result of the application of the interpolation method used to estimate W126 index
values at the centroid of every 12X12 km2 grid resolution, rather than only at the exact location of a monitor.
17 Public comment received thus far in this review are in the docket EPA-HQ-OAR-2008-0699, accessible
at www.regulations.gov.
6-40
-------
In response to the EPA's solicitation of CAS AC's advice on the Agency's proposed
rulemaking as part of the reconsideration,18 CASAC conveyed their support for a secondary
standard distinct from the primary standard, stating that it "also supports EPA's secondary ozone
standard as proposed: a new cumulative, seasonal standard expressed as an annual index of the
sum of weighted hourly concentrations (i.e., the W126 form), cumulated over 12 hours per day
(Sam to 8pm) during the consecutive 3-month period within the ozone season with the maximum
index value, set as a level within the range of 7 to [1]5 ppm-hours. This W126 metric can be
supported as an appropriate option for relating ozone exposure to vegetation responses, such as
visible foliar injury and reductions in plant growth. We found the Agency's reasoning ... to be
supported by the extensive scientific evidence considered in the last review cycle. In choosing
the W126 form for the secondary standard, the Agency acknowledges the distinction between the
effects of acute exposures to ozone on human health and the effects of chronic ozone exposures
on welfare, namely that vegetation effects are more dependent on the cumulative exposure to,
and uptake of, ozone over the course of the entire growing season (defined to be a minimum of at
least three months). In this proposal, the Agency is responding to the clear need for a secondary
standard that is different from the primary standard in averaging time, level and form" (Samet,
2010, p. i-ii).
In advice offered in the current review, which considers an updated scientific and
technical record since the 2008 rulemaking, the CASAC reiterated its earlier conclusions
regarding the appropriate form and averaging times for a secondary Os NAAQS at several points
in its letter to the Administrator. In stating the basis for its conclusion, CASAC notes that "[i]n
reaching its scientific judgment regarding the indicator, form, summation time, and range of
levels for a revised secondary standard, the CASAC has focused on the scientific evidence for
the identification of the kind and extent of adverse effects on public welfare" (Frey, 2014a, p.
iii), and further that "[tjhese recommendations are based on scientific evidence of adverse effect
associated with the presence of ozone in ambient air" (Frey, 2014a, p. 15). On this basis,
CASAC reached its conclusions on the appropriate form for the secondary standard stating "[t]he
CASAC supports the scientific conclusion in the Second Draft PA that the current secondary
standard is not adequate to protect against current and anticipated welfare effects of ozone on
vegetation. We recommend retaining the current indicator (ozone) but establishing a revised
form of the secondary standard to be the biologically relevant W126 index accumulated over a
18 The reconsideration proposal included a proposed new cumulative, seasonal secondary standard,
expressed as an index of the annual sum of weighted hourly concentrations (the W126 index), cumulated over 12
hours per day during the consecutive 3-month period within the O3 season with the maximum index value, averaged
over three years, set within a range of 7 to 15 ppm-hrs (75 FR 3027).
6-41
-------
12-hour period (8 a.m. - 8 p.m.) over the 3-month summation period of a single year resulting in
the maximum value of W126" (Frey, 2014a, p. iii).
In addition, we take note of the scientific advice provided by CAS AC regarding its
scientific judgments regarding appropriate target benchmarks of protection and the range of
W126 index values that in its scientific judgment provides appropriate protection for these
benchmarks. CASAC states that "[a] 2% biomass loss is an appropriate scientifically based
value to consider as a benchmark of adverse impact for long-lived perennial species such as
trees, because effects are cumulative over multiple years" and "[c]rop loss appears to be less
sensitive than these other indicators, largely because of the CASAC judgment that a 5% yield
loss represents an adverse impact, and in part due to more opportunities to alter management of
annual crops" (Frey, 2014a, p. 14).
Given these benchmarks, CASAC provided further advice regarding an appropriate range
of W126 levels that it considered appropriately protective. Specifically, "[t]he CASAC
recommends that the level associated with this form be within the range of 7 ppm-hrs to 15 ppm-
hrs to protect against current and anticipated welfare effects of ozone. The CASAC does not
support a level higher than 15 ppm-hrs. For example, at 17 ppm-hrs, the median tree species has
6% relative biomass loss, and the median crop species has over 5% yield loss. These levels are
unacceptably high" (Frey, 2014a, p. iii)19. CASAC further noted that "[w]ith compounding over
the harvest cycle or life span of these species, this will result in considerably greater cumulative
RBL as discussed above. For the more sensitive tree seedlings, a value closer to the lower end of
the range (7 ppm-hrs) would be more appropriate. The level of 7 ppm-hrs is the only level
analyzed for which the relative biomass loss for the median tree species is less than or equal to 2
percent. At 7 ppm-hrs, 7 of the 12 analyzed species have relative biomass loss of less than 2%"
(Frey, 2014a, p. 14).
CASAC further noted that "the correlative similarity between the current standard and a
level of the W126 index of 15 ppm-hrs must not be interpreted to mean that just meeting the
current standard is equivalent to just meeting a W126 level of 15 ppm-hrs. Most of the analyses
found effects below 15 ppm-hrs (many at 10 or even 7 ppm-hrs)" (Frey, 2014a, p. 12).
CASAC also recognized that there were policy choices left to the Administrator with
respect to determining an appropriate level of protection. In so doing "[t]he CASAC
acknowledges 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. Specifically, the
19 As noted in Section 6.3, the numbers for RBL for the median tree species have been updated between the
second and final PA to deemphasize cottonwood, based on staff's understanding of CASAC advice in that regard.
We note that CASAC advice based on what is shown in Table 6-1 is no longer consistent in some cases with the
revised table, and in particular with regard to median tree seedling RBL values.
6-42
-------
Clean Air Act grants discretion to the Administrator to specify a standard that 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'... (Frey, 2014a, p. iii). In addition, CASAC also
offered its policy advice regarding selection of an appropriate level within its scientifically
recommended range, stating that "[a]s a policy recommendation, separate from its advice above
regarding scientific findings, the CASAC advises that a level of 15 ppm-hrs for the highest 3-
month sum in a single year is requisite to protect crop yield loss, but that lower levels provide
additional protection against crop yield loss. Furthermore, there are specific economically
significant crops, such as soybeans, that may not be protected at 15 ppm-hrs but would be
protected at lower levels. A level below 10 ppm-hrs is required to reduce foliar injury. A level of
7 ppm-hrs is protective of relative biomass loss for trees and offers additional protection against
crop yield loss and foliar injury. Therefore, 7 ppm-hrs is protective of ecosystem services. Thus,
lower levels within the recommended range offer a greater degree of protection of more
endpoints than do higher levels within the range" (Frey, 2014a, p. iii).
Additionally, in regard to the 3-year average option discussed in the second draft PA,
CASAC thus notes that "[i]f, as a policy matter, the Administrator prefers to base the secondary
standard on a three-year averaging period for the purpose of program stability, then the level of
the standard should be revised downward such that the level for the highest three-month
summation in any given year of the three-year period would not exceed the scientifically
recommended range of 7 ppm-hrs to 15 ppm-hrs. .. .The final Policy Assessment should quantify
the ratio of the three-year average of the highest three-month summations in each year to the
highest three-month summation in the highest year. This ratio should be used to determine what
downward adjustment from the three-month summation in one year recommended here is needed
if a three-year form is selected" (Frey, 2014a, pp. iii and iv).20
Finally, we note that in commenting on the significance of the uncertainties associated
with the evidence and exposure and risk analyses that remain, CASAC concludes that "[w]hile
these scientific research priorities will enhance future scientific reviews of the ozone primary and
secondary standards, we also make clear that there is sufficient scientific evidence, and sufficient
confidence in the available research results, to support the advice we have given above for this
review cycle of the primary and secondary standards" (Frey, 2014a, p. iv).
20 See Section 6.4 and Chapter 2 for more discussion on the relationship between one-year and three-year
average W126 index values.
6-43
-------
6.6 STAFF CONCLUSIONS ON ALTERNATIVE STANDARDS
Staffs consideration of alternative secondary Os standards builds on our conclusion from
section 5.7 above that the body of evidence, in combination with the results of the WREA
analyses, calls into question the adequacy of the current secondary standard and provides support
for consideration of alternative standards. In sections 6.1 to 6.3 above, we consider how the
currently available scientific evidence and exposure/risk information informs staff conclusions
regarding the basic elements of the NAAQS: indicator (6.1), form and averaging time (6.2), and
level (6.3). In so doing, we consider both the information available at the time of the last review
and information newly available since the last review that has been critically analyzed and
characterized in the 2013 ISA. As an initial matter, with regard to the indicator, we conclude
that based on the available science it is still appropriate to continue to use measurements of Os in
accordance with federal reference methods as the indicator to address effects associated with
exposure to ambient Os alone or in combination with related photochemical oxidants.
In considering alternative standards, staff has considered the available body of evidence
as comprehensively assessed in the ISA, the risk and exposure information presented in the
WREA, and CASAC advice and public comment in this review with regard to support for
consideration of options that are different from the current standard, as articulated by the
following overarching question:
• To what extent does the currently available scientific evidence- and exposure/risk-
based information, as reflected in the ISA and WREA, support consideration of
alternatives to the current Os standard to provide increased protection from
ambient Os exposures?
In considering potential forms alternative to that of the current standard, we note that the
form for the current secondary standard is the 4th highest daily maximum 8-hour average,
averaged over three years. As discussed in chapter 5 and section 6.2 above, the longstanding
evidence regarding the fundamental aspects of Os exposure that are directly responsible for
inducing vegetation response indicates that plant response to Cb is driven by the cumulative
exposure to Cb during the growing season (U.S. EPA, 2013, section 2.6.6.1). This cumulative
exposure depends on both the total duration of the exposure (from repeated Os episodes) and the
concentrations of those exposures (higher concentrations having a disproportionate impact as
compared to lower concentrations). On the basis of this longstanding and extensive evidence,
the ISA concludes that exposure indices that cumulate and differentially weight the higher hourly
average concentrations over a season and also include the mid-level values offer the most
scientifically defensible approach for use in developing response functions and in defining
indices for vegetation protection (U.S. EPA, 2013, section 2.6.6.1).
6-44
-------
CAS AC advice in the 2008 review and on the 2010 proposed reconsideration also
recognized that the nature of the exposures relevant to vegetation response is well described by a
cumulative seasonal form and has supported the use of such a form for a secondary Os standard
(Henderson, 2006; Samet, 2010). The current CAS AC Os Panel has expressed similar views.
We also note that on the basis of the evidence and exposure/risk information available in the two
previous reviews, and in consideration of CASAC advice, the Administrator has recognized the
importance of protecting vegetation from cumulative, seasonal exposures and proposed such a
form as an appropriate, reasonable policy option (61 FR 65741-44; 72 FR 37899-905; 75 FR
3012-3027).
Thus, in considering alternative forms of the standard we conclude that it is reasonable
and appropriate to consider a cumulative, concentration-weighted form to provide protection
against cumulative, seasonal exposures to Os that are known or anticipated to harm sensitive
vegetation or ecosystems. Such a form is specifically designed to focus on the kind of Os
exposures that have been shown to cause harm to vegetation and would have a distinct advantage
over the form of the current standard in characterizing air quality conditions potentially of
concern for vegetation and in more directly demonstrating that the desired degree of protection
against those conditions was being achieved.
In considering the appropriate index for a cumulative seasonal form, we recognize that a
number of different cumulative concentration weighted indices have been developed and have
been evaluated in the scientific literature and in past NAAQS reviews in terms of their ability to
predict vegetation response and their usefulness in the NAAQS context (U.S. EPA, 2006, pp. 9-
11 to 9-15 and pp. AX9-159 to AX9-187; U.S. EPA, 2007, pp. 7-15/16). While these various
forms have different strengths and limitations, as noted in the ISA (U.S. EPA, 2013, section 9.5),
the W126 index21 has some important advantages over other non-sigmoidally weighted
cumulative indices. For example, given the lack of a discernible threshold for vegetation effects
in general, we recognize the fact that the W126 metric does not have a cut-off in its weighting
scheme (down to about 30 ppb below which the weighting factor is effectively zero), such that it
includes consideration of potentially damaging lower Os concentrations. Additionally, the W126
metric adds increasing weight to hourly concentrations from about 40 ppb to about 100 ppb, an
important feature because "as hourly concentrations become higher, they become increasingly
likely to overwhelm plant defenses and are known to be more detrimental to vegetation" (U.S.
21 The W126 is a non-threshold approach described as the sigmoidally weighted sum of all hourly Os
concentrations observed during a specified diurnal and seasonal exposure period, where each hourly O3
concentration is given a weight that increases from 0 to 1 with increasing concentration (Lefohn et al, 1988; Lefohn
and Runeckles, 1987; U.S. EPA, 2013, section 9.5.2).
6-45
-------
EPA, 2013, p. 9-104). We additionally take note of CASAC advice in the 2008 review and on
the 2010 proposed reconsideration recommending the use of the W126 index for a cumulative
seasonal form for a secondary Os standard (Henderson, 2006; Samet, 2010). Similarly, the
current CASAC Os Panel has indicated that a focus on a W126 form is appropriate (Frey, 2014a,
p. iii). Therefore, on the basis of the strength of the evidence and advice from CASAC, we
conclude that the W126 index is the most appropriate cumulative seasonal form to consider in
the context of the secondary Os NAAQS review.
We next turn to the exposure periods - diurnal and seasonal - over which the W126
index would be summed in any given year. As discussed in section 6.2 above, the currently
available information continues to provide support for a definition of the diurnal period of
interest as the 12-hour period from 8:00 am to 8:00 pm (U.S. EPA, 2013, section 9.5.3). In prior
reviews, the EPA has identified the 12-hour period from 8:00 am to 8:00 pm as appropriately
capturing the diurnal window with most relevance to the photosynthetic process (72 FR 37900;
75 FR 3013), and CASAC has generally supported the 12-hour daylight period (Henderson,
2006, 2007). In light of the continued support in the evidence base and no evidence on this issue
differing from that in previous reviews, we again conclude that it is appropriate to use the 12-
hour period from 8:00 am to 8:00 pm to cumulate daily Os exposures. On this basis, we
conclude that the 12-hour daylight window (8:00 am to 8:00 pm) represents the portion of the
diurnal exposure period that is most relevant to predicting or inducing plant effects related to
photosynthesis and growth and thus is an appropriate diurnal period to use in conjunction with a
W126 cumulative metric.
With regard to a seasonal period of interest, the current evidence base continues to
provide support for a seasonal period with a minimum duration of three months (U.S. EPA,
2013, section 9.5.3). We note that a plant is vulnerable to Os pollution as long as it has foliage
and is physiologically active (U.S. EPA, 2013, section 9.5.3, p. 9-112), i.e., during its growing
season. The exposure periods used in studies of Os effects on vegetation reflect this
understanding and typically focus on study periods of 3-6 months. Included in the currently
available evidence is a new analysis that compared 3- and 6-month maximum W126 index
values for over 1,200 AQS and CASTNET EPA monitoring sites for the years 2008-2009 that
found that the two accumulation periods were highly correlated (U.S. EPA, 2013, section 9.5.3,
Figure 9-13). Thus, although we recognize that the selection of a single seasonal time period
over which to cumulate Os exposures for a national standard necessarily represents a balance of
factors, given the significant variability in growth patterns and lengths of growing season among
vegetative species growing within the U.S., we conclude it is appropriate to identify the seasonal
W126 index value as that derived from the consecutive 3-month period within the Os season
with the highest W126 index value. We note that such a 3-month exposure period was also
6-46
-------
supported by CAS AC in advice provided during the last review and the 2010 proposed
reconsideration (Henderson, 2006; Samet, 2010).
With regard to form, we additionally consider the period of time over which a cumulative
seasonal W126-based standard should be evaluated. In so doing, we have considered the support
for both a single year form and a form averaged over three years (section 6.2). We note
comments from CASAC on this matter, in particular their comment in the current review that
"[t]he CASAC does not recommend the use of a three-year averaging period for the secondary
standard. We favor a single-year period for determining the highest three-month summation
which will provide more protection for annual crops and for the anticipated cumulative effects on
perennial species. The scientific analyses considered in this review, and the evidence upon which
they are based, are from single-year results" (Frey, 2014a, p. iii).
We recognize that there are a number of Os-induced effects that have the potential for
public welfare significance within the annual timeframe. These effects mainly include reduced
crop yields and visible foliar injury, as noted in section 6.2 above. There are uncertainties
associated with these effects that make it difficult to determine the degree of annual protection
needed to protect the public welfare from any known or anticipated adverse effects. There are
also annual effects in perennial species that may result from a single year exposure and can be
"carried over" into the subsequent year where they affect growth and reproduction (U.S. EPA,
2013, pp. 9-43 to 9-44 and p. 9-86). When such annual effects due to elevated Os exposures
occur over multiple years, they have the further potential to be compounded, increasing the
potential for effects at larger scales (e.g., population, ecosystem), including effects on associated
services that may be of significance to the public welfare. These ecosystem services effects can
include alteration of below-ground biogeochemical cycles, and alteration of both above- and
below-ground terrestrial community composition and terrestrial ecosystem water cycling (U.S.
EPA, 2013, Table 9-19) and reductions in productivity and carbon sequestration in terrestrial
ecosystems. We additionally note that multiple consecutive years of critical Os exposures might
be expected to result in larger impacts on forested areas (e.g., increased susceptibility to other
stressors such as insect pests, disease, co-occurring pollutants and harsh weather) than
intermittent occurrences of such exposures due to the potential for compounding or carry-over
effects on tree growth.
Given the above, we conclude that the public welfare significance of the effects that can
occur as a result of three-year Os exposures are potentially greater than those associated with a
single year of such exposure. Thus, to the extent that the focus for public welfare protection to
be afforded by the secondary Os standard is on long-term effects that occur in sensitive tree
species in natural forested ecosystems, including federally protected areas such as Class I areas
or on lands set aside by States, Tribes and public interest groups to provide similar benefits to the
6-47
-------
public welfare, a standard with a form that evaluates the cumulative seasonal index across
multiple years might be considered to provide a more appropriate match to the nature of Os-
related effects on vegetation upon which the secondary Os standard is focused. In considering
such forms, we focus on one that averages the W126 index values across three years, as
discussed in section 6.2 above.
In addition to the vegetation effects considerations described above, there are other
policy-relevant factors that can be useful to consider. For example, under a standard with a
single year form, a monitor may be judged to meet the standard based on a single year of data,
while under a standard with a form requiring evaluation over a multi-year period, a monitor is
not judged to have met the standard until a complete multi-year record is available. For a W126-
based potential standard, the multi-year form identified for consideration in the last review was
the average cumulative seasonal metric over three consecutive years (75 FR 3027). Such a
multi-year form remains appropriate to consider to provide stability to an alternative secondary
standard, just as the multi-year form provides for the current standard (average over three years
of annual fourth-highest daily maximum 8-hour average Os concentrations).22 In considering the
issue of stability in the context of such a form, we first note the inter-annual variability of
seasonal W126 index, which is not unexpected given the logistic weighting function and also
inter-annual variability in meteorological conditions that contribute to Os formation (see
Appendix 2C). The staff analysis in Appendix 2C describes the variability in annual W126
index values in relation to variability in the 3-year average, which indicates that a standard based
on an annual W126 index would be expected to have a lower degree of year-to-year stability
relative to a standard based on a form that averages seasonal indices across three consecutive
years. A more stable standard can be expected to contribute to greater public welfare protection
by limiting year-to-year disruptions in ongoing control programs that would occur if an area was
frequently shifting in and out of attainment due to extreme year-to-year variations in
meteorological conditions. This greater stability in air quality management programs thus
facilitates achievement of the protection intended by a standard. In light of this relationship, we
conclude that a 3-year average form has the desirable feature of providing greater stability in air
quality management programs and thus facilitating the achievement of the protection intended by
a standard. Thus, we recognize the public welfare benefits of having a standard of a 3-year
average form.
Thus, to the extent that the greater emphasis is placed on protecting against effects
associated with multi-year exposures and maintaining more year-to-year stability of public
22 See ATA III, 283 F. 3d at 374-75 (recognizing programmatic stability as a legitimate consideration in the
NAAQS standard-setting process).
6-48
-------
welfare protection, we conclude that it is appropriate to consider a secondary standard form that
averages the seasonal W126 index values across three consecutive years. We conclude that such
a form might be appropriate for a standard intended to achieve the desired level of protection
from longer-term effects, including those associated with potential compounding. Further, such
a form might be concluded to contribute to greater stability in air quality management programs,
and thus, greater effectiveness in achieving the desired level of public welfare protection, than
that that might result from a single year form.
Turning to consideration of an appropriate range of levels for a W126-based standard, we
first note that our general approach to informing these judgments recognizes that the available
evidence demonstrates a range of Os sensitivity across studied plant species and documents an
array of Cb-induced effects that extend from lower to higher levels of biological organization.
These effects range from those affecting cell processes and individual plant leaves to effects on
the physiology of whole plants, species effects and effects on plant communities to effects on
related ecosystem processes and services. Given this evidence, it is not possible to generalize
across all studied species regarding which cumulative exposures are of greatest concern, as this
can vary by situation due to differences in exposed species sensitivity, the importance of the
observed or predicted Os-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. At the same time, the evidence also demonstrates that though effects of concern can
occur at very low exposures in sensitive species, at higher cumulative exposures those effects
would likely occur at a greater magnitude and/or higher levels of biological organization and
additional species would likely be impacted. It is important to note, however, that due to the
variability in the importance of the associated ecosystem services provided by different species
at different exposures and in different locations, as well as differences in associated uncertainties
and limitations, that, in addition to the magnitude of the ambient concentrations, both the species
present and their public welfare significance are essential considerations in drawing conclusions
regarding the significance or magnitude of public welfare impact.
Therefore, in developing conclusions in this PA, we take note of the complexity of
judgments to be made by the Administrator regarding the adversity of known and anticipated
effects to the public welfare and are mindful that the Administrator's ultimate judgments on the
secondary standard will, as appropriate, reflect an interpretation of the available scientific
evidence and exposure/risk information that neither overstates nor understates the strengths and
limitations of that evidence and information.
As described above in section 5.1, we employ a paradigm to assist in putting the available
science and exposure/risk information into the public welfare context. This paradigm has
6-49
-------
evolved over the course of the Os NAAQS reviews and has also been informed by similar
constructs developed for other secondary NAAQS reviews. As discussed in Section 5.1, this
paradigm recognizes that the significance to the public welfare of Os-induced effects on sensitive
vegetation growing within the U.S. can vary 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. Accordingly, any given Os-related effect on vegetation
and ecosystems (e.g., biomass loss, crop yield loss, visible foliar injury) may be judged to have a
different degree of impact on or significance to the public welfare depending, for example, on
whether that effect occurs in a Class I area, a city park, or commercial cropland. In the last
review, the Administrator placed the highest priority and significance on vegetation and
ecosystem effects to sensitive species that are known to or are likely to occur in federally
protected areas such as national parks and other Class I areas, or on lands set aside by states,
tribes and public interest groups to provide similar benefits to the public welfare (75 FR 3023-
24; 73 FR 16496), recognizing that effects occurring in such areas would likely have the highest
potential for being classified as adverse to the public welfare, due to the expectation that these
areas need to be maintained in pristine or near pristine conditions to ensure their intended use is
met. This approach also includes consideration of impacts to ecosystem goods and services.
Although ecosystem services were not explicitly considered in the Administrator's decision in
the last review, they were explicitly recognized as an important category of public welfare
effects and they have an obvious relationship to consideration of intended use (73 FR 16492). In
employing this approach, we note the support for it provided by CASAC advice in this review
(Frey, 2014a).
In considering potential levels for an alternative standard based on the W126 metric, we
focus the discussion primarily on: 1) impacts on tree growth, productivity and carbon storage; 2)
crop yield loss; and 3) visible foliar injury. With respect to tree growth, we find it useful to
consider the summary of relative biomass loss estimates in Table 6-1 above and the WREA
risk/exposure estimates discussed in Section 6.3 and Appendix 6F. In Table 6-1, we take note of
the different index value estimates with regard to the number of studied species below different
response benchmarks, as well as with regard to the median response. We additionally consider
the WREA estimates regarding: (1) percent of assessed geographic area exceeding 2% weighted
relative biomass (Table 6-2); (2) number of assessed Class I areas with tree seedling weighted
relative biomass loss estimates above 2% (Table 6-3); and (3) the percent median biomass loss
across counties for different air quality scenarios (Table 5-5). Further, we note other WREA
estimates for effects on ecosystem services related to public welfare, such as carbon
sequestration and air pollutant removal. With respect to crop yield loss, we note the summary of
crop yield loss estimates in Table 6-1 and the WREA risk/exposure estimates discussed in
6-50
-------
Section 6.3 and Appendix 6F, which include individual species and median response. We also
note information available on visible foliar damage to species occurring in natural settings, such
as federal Class I areas, and the analyses in the WREA evaluating biosite data and several
benchmarks of injury as summarized in section 5.4.2.
In focusing on trees and their associated ecosystem services, we first note that the studied
tree species vary widely in their sensitivity to Os-induced relative biomass loss. We thus find it
informative to consider both median species values and individual species responses and RBL
over the same W126 range. We note CASAC's advice regarding RBL levels, specifically their
emphasis on a benchmark of median relative tree biomass loss at or below 2% and their view
that a 6% median relative biomass loss is "unacceptably high". From Table 6-1 we see that
median tree species biomass loss is at or below 2% only at the lowest W126 level assessed, 7
ppm-hrs. As the W126 level is incrementally increased, median RBL also increases
incrementally, so that at W126 index values of 9, 11, 13, 15, 17 and 19, the median RBL
increases to 2.4%, 3.1%, 3.8%, 4.5%, 5.3% and 6.0%, respectively. Thus over the W126 range
of 7 - 17 ppm-hrs, median species biomass loss ranges from approximately 2% to approximately
5%.
We next take note of the number of individual species' RBLs that fall below those same
benchmarks assessed for median species values. We also note the value of additionally
characterizing the RBL estimates in comparison to higher loss levels such as 10% or 15%,
especially for individual tree species. Based on Figure 5-1 (B) in Chapter 5, and as shown in
Table 6-1, for W126 values at or below 17 ppm-hrs, the RBLs for each of 5 species is less than
2%. Thus, over the full range of alternative levels considered, the same level of protection
relative to the 2% benchmark is achieved for these species. We therefore turn our attention to
the remaining 6 studied species to see if additional information might be available to help inform
consideration of an appropriate degree of protection. Specifically, we consider the RBL
information available for the other species (i.e., eastern white pine, aspen, tulip poplar,
ponderosa pine, red alder, and black cherry) to further inform our evaluation of the additional
protection that potentially could be achieved at different W126 levels within the range identified.
We note that, for W126 levels of 17 to 7 ppm-hrs, biomass loss decreases for these individual
species with decreasing W126 levels such that at the W126 level of 17 ppm-hrs, five species
have RBL above 6% while at the W126 level of 7 ppm-hrs, one species (black cherry) has an
RBL above 6%. Taken together with the more tolerant species, the proportion of the studied tree
species with RBLs below 6% are 6/11, 7/11, 8/11, and 10/11 at W126 index values of 17, 15,
13, and 11 ppm-hrs, respectively.
In consideration of other benchmark levels, 9/11 studied tree species have a predicted
RBL below 10% at the W126 level of 17 ppm-hrs, while 10/11 species have a predicted RBL
6-51
-------
below 10% for W126 levels of 15 to 7 ppm-hrs. In addition, 10/11 studied tree species have a
predicted RBL below 15% for W126 levels of 17 to 7 ppm-hrs. We note that black cherry, the
most sensitive of the 11 species, has RBLs ranging from approximately 36% at W126 index
value of 17 down to approximately 17% at the W126 index value of 7 ppm-hrs. Thus, the
predicted RBL for black cherry remains above 15% for W126 levels of 17 to 7 ppm-hrs, and it is
not clear to what extent those predicted RBL values might inform consideration of the level of
protection achieved for different W126 exposures within this range (Table 6-1; U.S. EPA,
2014b, section 6.2, Appendix 6A).
To further inform this issue, the WREA also characterizes the number of counties where
the median RBLs were greater than 2% (U.S. EPA, 2014b, Table 6-7), as shown in Table 5-5.
When air quality is adjusted to the current standard, 8% of the counties have median RBLs
greater than 2%. That proportion drops to 7% for air quality adjusted to just meet a 3-year
average W126 level of 15 ppm-hrs and to 6% for air quality adjusted to just meet a 3-year
average W126 level of 7 ppm-hrs. Of the 239 counties (8% of counties) estimated to have a
median RBL above 2% when meeting the current standard, 203 of those counties have a RBL
greater than 2% because of the presence of black cherry. Thus, as also discussed above in
Section 6.2, given the large magnitude of estimated RBL for black cherry over the entire range
assessed, it is not clear to what extent the information for black cherry informs consideration of
the overall level of protection achieved across the identified range.
In considering the potential magnitude of the ecosystem impact of reduced biomass in
trees, we focus on the WREA estimates of weighted RBL for the W126 air quality scenarios
(U.S. EPA, 2014b, section 6.8), focusing particularly on impacts in Class I areas. For the current
standard and the three W126 scenarios (15 ppm-hr, 11 ppm-hr, and 7 ppm-hr), the percent of
total national land-area having weighted RBL greater than 2% was 0.2%, 0.2%, 0.1% and
<0.1%, respectively (Table 6-2; U.S. EPA 2014, Table 6-25). In addition, the WREA estimates
indicate weighted RBL greater than 2% in 1-2 of 145 assessed nationally protected Class I areas
for the current standard and all three W126 scenarios. To the extent that emphasis is given to
such estimates for nationally protected Class I areas and for appreciable percentages of forested
areas nationwide, a W126 index value extending up to 17 ppm23 may be appropriate to consider.
The WREA provides qualitative and semi-quantitative information regarding the types
and potential magnitude of Os impacts on ecosystem services. In noting the potential ecosystem
23 While the WREA analyses did not include an air quality scenario for 17 ppm-hrs, the data suggests that,
to adjust air quality to meet a W126 index value of 17 ppm-hrs, additional emissions reductions would have been
needed relative to just meeting the current standard. Therefore, because the air quality scenarios for meeting the
current standard and meeting a W126 index value of 15 ppm-hrs both indicate weighted relative biomass loss less
than or equal to 2% in 143 of 145 assessed nationally protected Class I areas, the same would be true for an air
quality scenario for just meeting a W126 index value of 17 ppm-hrs.
6-52
-------
services benefits related to reductions in tree biomass loss resulting from just meeting potential
alternative W126-based standards, we recognize, in particular, that impacts on climate regulation
can reasonably be concluded to be potentially significant from a public welfare perspective and
carbon sequestration has been identified as a potentially important tool for managing
anthropogenic impacts on climate. The WREA estimates the potential increase in carbon storage
that potentially could occur for different air quality scenarios (U.S. EPA 2014, section 6.6.1).
Comparisons of the W126 scenarios to the current standard scenario with regard to carbon
sequestration estimates do not indicate an appreciable difference for the W126 scenario of 15
ppm-hrs beyond that achieved by just meeting the current standard. The majority of the
enhanced carbon sequestration potential in forests over time is predicted to occur for the
alternative W126 scenarios of 11 and 7 ppm-hrs. Over 30 years, the current standard scenario
projection is 89,184 million metric tons of CCh equivalents (MMtCChe).24 The WREA estimates
additional sequestration potential of 13, 593 and 1,600 MMtCChe for the W126 scenarios of 15,
11 and 7 ppm-hrs, respectively, as compared to the current standard (U.S. EPA 2014, Table 6-
18). We additionally consider the WREA estimates for five urban areas of how reduced growth
of Os-sensitive trees in urban forests may affect air pollutant removal (U.S. EPA, 2014b, sections
6.6.2 and 6.7 and Appendix 6D). Estimates for all five case study areas indicate increased
pollutant removal from the recent conditions to just meeting the current standard, with much
smaller differences between the current standard and the three W126 scenarios (Table 6-5).
However, we additionally take note of significant uncertainties and limitations associated with
WREA estimates related to carbon sequestration and air pollution removal. Thus, we note that
an identification of the requisite protection for forest trees and their associated ecosystem
services would likely involve policy judgments regarding the appropriate weight to place on
potential impacts to the public welfare with respect to estimated effects on the ecosystem
services of carbon storage and urban air pollution removal associated with tree growth, as well as
on the uncertainties associated with this information.
With respect to crops, we focus on the 10 robust E-R functions (barley, lettuce, field
corn, grain sorghum, peanut, winter wheat, cotton, soybean, potato and kidney bean) described in
the ISA and additionally analyzed in the WREA (Figure 5-4). We also note CASAC's advice
regarding a recommended target benchmark protection level of 5% for median crop relative yield
loss (RYL) and that, as shown in Table 6-1, W126 index values ranging from 7 to 17 ppm-hrs
241 million metric tons of carbon dioxide equivalents (MMtCO2e) is equivalent to 208,000 passenger
vehicles or the electricity to run 138,000 homes for 1 year as calculated by the EPA Greenhouse Gas Equivalencies
Calculator (updated September 2013 and available at http://www.epa.gov/cleanenergy/energy-
resources/calculator.html).
6-53
-------
are estimated to have median crop RYL of less than or equal to approximately 5%. Given this, it
is not clear to what extent this information informs the selection of an appropriate level.
When individual species are considered over this same range, the proportion of crops
protected varies from 5/10, 6/10, 6/10, 9/10, 10/10, and 10/10 at the W126 levels of 17, 15, 13,
11, 9, and 7 ppm-hrs. To the extent a given species is judged as having particular importance to
the public welfare, breaking the information down by species can be helpful. For example, less
than 5% yield loss was estimated for soybeans at the W126 index value of 12 ppm-hrs (U.S.
EPA 2014, Figure 6-3). Four of the studied crop species (i.e., barley, lettuce, field corn, and
grain sorghum) are more tolerant, with RYL under 1% over the W126 range from 7 to 17 ppm-
hrs. Peanut also remained under 4% RYL over the same W126 range. Other species differed
regarding the W126 level at which RYL reached or fell below 5%. Specifically, for winter
wheat, cotton, soybean, kidney bean and potato, the relevant W126 index values at which RYLs
were below 5% are 15, 13, 11,9 and 7 ppm-hrs. As noted in Chapter 5, and in early discussions
in this chapter, the significance of these predicted RYLs to the public welfare could be informed
by the recognition that crops are heavily managed to obtain the desired yield and the potential
adversity to public welfare from yield reductions in any specific crop in a particular location
would depend on a number of economic factors, including crop prices, crop substitution, and the
welfare importance of relative changes in consumer and producer surplus. We also note that
these crop species would likely receive some protection from a standard set, for example, to
provide protection against tree biomass loss, such as in areas set aside to be maintained in a more
pristine condition (75 FR 3024).
Visible foliar injury has been identified by the FLMs as a diagnostic tool for informing
conclusions regarding potential ozone impacts on potentially sensitive AQRVs (USFS, NPS,
FWS, 2010), indicating that such Os-induced impacts might be considered to have the potential
to impact the public welfare in scenic and/or recreational areas during years they occur. We take
note of the WREA analyses of the nationwide dataset (2006-2010) for USFS/FHM biosites
described in section 5.4.2 above, including the observation that the proportion of biosites with
injury varies with soil moisture conditions and Os W126 index values (U.S. EPA 2014, Chapter
7, Figure 7-10). These analyses also show that foliar injury incidence increases steeply with
W126 index values up to approximately 10 ppm-hrs. At W126 index levels greater than that,
little or no further increase in proportion of sites showing foliar injury occurs.
With respect to visible foliar injury, we are unaware of any guidance for federal land
managers regarding at what spatial scale or what degree of severity visible foliar injury is
sufficient to trigger protective action based on this potential impact on AQRVs. Further, there
does not appear to be any consensus in the literature regarding severity of foliar injury and risks
to plant functions or services, and CASAC, while identifying target percent biomass loss and
6-54
-------
yield loss benchmarks for tree seedlings and commodity crops, respectively, did not provide a
similar recommendation for this endpoint. Likewise, as in previous reviews, the ISA notes 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, 2013, section 9.4.2, p. 9-39)
and further noted that the full body of evidence indicates that there is wide variability in this
endpoint, such that although evidence shows visible foliar injury can occur under very low
cumulative Os concentrations, ".. .the degree and extent of visible foliar injury development
varies from year to year and site to site..., even among co-members of a population exposed to
similar Os levels, due to the influence of co-occurring environmental and genetic factors" (U.S.
EPA 2013, section 9.4.2, p. 9-38). Given this, it is not clear to what extent this information
informs the selection of an appropriate level.
On the basis of all the considerations described above, including the evidence and
exposure/risk analyses, and advice from CASAC, we conclude that a range of W126 index
values appropriate for the Administrator to consider extends from 7 to 17 ppm-hrs. In so doing,
however, we note, as recognized above, the role of judgments by the Administrator in such
decisions. In selecting the range identified here, we primarily consider the evidence- and
exposure/risk-based information for cumulative seasonal Os exposures represented by W126
index values (including those represented by the WREA average W126 scenarios) associated
with biomass loss in studied tree species, both in and outside areas that have been afforded
special protections. We note CAS AC's advice that a 6% median RBL is unacceptably high, that
the 2% median RBL is an important benchmark to consider, that for the lower W126 value of 7
ppm-hrs that the median tree species biomass loss is at or below 2%, and that for the upper value
of 17 ppm-hrs the median tree biomass loss is below 6%25. We also note the estimates indicating
that a W126 level of 17 ppm-hrs reduces the percent of total nationwide land-area having
weighted RBL greater than 2% to 0.2% (Table 6-2) and the number of Class I areas with
weighted RBL greater than 2% to 2 of the 145 assessed nationally protected Class I areas.
We also note that tree biomass loss can be an indicator of more significant ecosystem-
wide effects which might reasonably be concluded to be significant to public welfare. For
example, when it occurs over multiple years at a sufficient magnitude, it is linked to an array of
effects on other ecosystem-level processes, such as nutrient and water cycles, changes in above
and below ground communities, carbon storage and air pollution removal (U.S. EPA, 2014b,
Figure 5-1), that have the potential to be adverse to the public welfare.
Thus, in staff s view, the evidence- and exposure/risk-based information relevant to tree
biomass loss and the associated ecosystem services important to the public welfare support
25 We note that a W126 index value of 19 ppm-hrs is estimated to result in a median RBL value of 6%.
6-55
-------
consideration of a W126-based secondary standard with index values within the range of 7-17
ppm-hrs. We consider such a range for a potential alternative cumulative seasonal W126-based
standard, averaged over three years, based on our analysis of the small effect of year to year
variability on the cumulative biomass loss associated with multiple years of exposure, and the
benefits of improved stability of the W126 standard when evaluated using the 3-year average
form. Lastly, we are mindful of the policy judgments required of the Administrator with regard
to the public welfare significance of identified effects and the requisite level of protection, as
well as the appropriate weight to assign the range of uncertainties inherent in the evidence and
analyses.
While we additionally recognize foliar injury as an important Os effect which, depending
on severity and spatial extent, most particularly in nationally protected areas such as Class I
areas, may reasonably be concluded to be of public welfare significance, we take note of the
appreciable variability in this endpoint, as summarized in chapter 5 and section 6.3 above, which
poses challenges to giving it primary emphasis in identifying potential alternative standard
levels. Similarly, we give less emphasis to consideration of crop yield loss in our consideration
of potential standard levels here and in section 6.3 above, noting the median estimates of
approximately 5% or lower for W126 index levels at and below 17 ppm-hrs. We also note the
range of factors affecting annual crop yields, including those related to the role of management
strategies as recognized in sections 5.3 and 6.2 above which complicate the identification of a
degree of impact that can be considered adverse to the public welfare.
We further recognize the role of policy judgments by the Administrator, as described
above, in identifying a target level of protection for the secondary Os standard. For example, to
the extent effects associated with cumulative multi-year exposures are judged important to the
public welfare, more weight may be placed on such effects, as well as the role that year-to-year
exposure variability can play in realizing the potential public welfare impacts.
Lastly, we also conclude that, to the extent the Administrator finds it useful to consider
the public welfare protection that might be afforded by a revised primary standard, this is
appropriately judged by evaluating how the cumulative seasonal W126-based exposure metric is
affected by attainment with such a revised primary standard. For example, comparison of the air
quality conditions expected to result from a revised primary standard, with those conditions
expressed in terms of W126 exposures, to the W126 levels concluded to provide the desired level
of public welfare protection could inform a judgment of whether a secondary standard set
identical to a revised primary standard would be expected to achieve the level of public welfare
protection concluded to be requisite under the Act. In this type of evaluation, such as through
the overlap analyses discussed in section 6.4 above, staff further concludes it is important to take
into account associated uncertainties, including those associated with the limited monitor
6-56
-------
coverage in many rural areas, including those in the west and southwest and at high elevation
sites.
6.7 SUMMARY OF CONCLUSIONS ON THE SECONDARY STANDARD
Staff conclusions are informed by our consideration of the available scientific evidence as
assessed in the ISA, the air quality/exposure/risk information in the WREA, advice from
CASAC in this review and in prior reviews, and public comment in this review.
Staff conclusions on policy options that are appropriate for the Administrator's
consideration in making decisions on the secondary standards for Os, together with supporting
conclusions from sections 5.7 and 6.5 above, are briefly summarized below. In reaching
conclusions on alternative standards to provide requisite protection for public welfare effects
associated with ambient Os exposures, staff has considered these standards in terms of the basic
elements of the NAAQS: indicator, form, averaging time, and level. In drawing these
conclusions, we are mindful that the Act requires secondary standards to be set so that, in the
Administrator's judgment, they are requisite to protect public welfare from known or anticipated
adverse effects, such that the standards are to be neither more nor less stringent than necessary.
Thus, the Act does not require that NAAQS be set at zero-risk or background levels, but rather at
levels that reduce risk sufficiently to protect public welfare from adverse effects.
(1) Staff concludes, based on the combined consideration of the body of evidence and
the results from the quantitative exposure/risk assessment, that the available
evidence and exposure/risk information call into question the adequacy of the
public welfare protection provided by the current standard and that it is
appropriate to consider revising the standard to provide greater public welfare
protection.
(2) With regard to indicator, staff concludes that it is appropriate to continue to use
Os as the indicator for a standard that is intended to address welfare effects
associated with exposure to Os, alone or in combination with related
photochemical oxidants. Based on the available information, staff concludes that
there is no basis for considering an alternative indicator at this time.
(3) With regard to averaging time and form, staff concludes that it is appropriate to
consider a revised secondary standard in terms of the cumulative, seasonal,
concentration-weighted form, the W126 index. With regard to definition of the
W126 index for this purpose, staff makes the additional conclusions:
a. It is appropriate to consider the consecutive 3-month period within the Os
season with the maximum index value as the seasonal period over which
6-57
-------
to cumulate hourly Cb exposures. Staff notes that the maximum 3-month
period generally coincides with maximum biological activity for most
vegetation, making the 3-month duration a suitable surrogate for longer
growing seasons.
b. It is appropriate to cumulate daily exposures for the 12-hour period from
8:00 am to 8:00 pm, generally representing the daylight period during the
3-month period identified above.
c. It is appropriate to consider a form that averages W126 index values
across three consecutive years. Staff concludes it is appropriate to
consider this form in conjunction with appropriate levels in order to
provide the desired degree of public welfare protection from Os effects
across multiple years.
(4) With regard to a target level of protection for a revised standard, staff concludes
that it is appropriate to give consideration to a range of levels from 17 ppm-hrs to
7 ppm-hrs, expressed in terms of the W126 index averaged across three
consecutive years.
a. To the extent the Administrator finds it useful to consider the extent of
public welfare protection that might be afforded by a revised primary
standard, staff concludes that public welfare protection is appropriately
judged through the use of the cumulative seasonal W126-based metric.
Staff additionally notes that, consideration of the support provided by the information
available in this review will depend on public welfare policy judgments by the Administrator
regarding the protection of public welfare. This range reflects staff judgment that a standard set
within this range could provide an appropriate degree of public welfare protection.
6.8 KEY UNCERTAINTIES AND AREAS FOR FUTURE RESEARCH AND DATA
COLLECTION
Staff believes it is important to highlight key uncertainties and recommendations for
welfare-related research, including model development and data gathering, associated with
secondary standards for Os. Based on items highlighted in chapter 9 of the ISA, chapters 5 and 6
herein, and CASAC advice, we have identified the following areas for future research and data
collection based on key uncertainties, research questions and data gaps that have been
highlighted in this review of the secondary standard. The first research area addresses the key
uncertainties associated with the extrapolation to plant species and environments outside of
specific experimental or field study conditions. The second area of research pertains to the
6-58
-------
assessment of the impact of Os on other welfare effects categories such as climate, ecosystem
components, and whole ecosystem structure and function. A third area of research would
support the development of approaches, tools, or methodologies useful in characterizing Os
exposures in rural, remote, high elevation and/or complex terrain areas and in characterizing
ecosystem services and their importance to the public welfare. These three areas are described
below.
With regard to the first research area we note that while there have been five decades of
research regarding Os effects on plants and much information has been compiled in previous
reviews, a number of key uncertainties remain. For example, while national visible foliar injury
surveys can indicate how widespread Os effects may be within the U.S., there remain
uncertainties associated with estimating the risk to vegetation of differing amounts of Os-induced
visible foliar injury over the plant's leaf area and the relationship between relative soil moisture
and the incidence and severity of foliar injury in sensitive species, as well as the extent to which
different degrees of visible foliar injury can impact ecosystem services (e.g., tourism). Research
to better characterize the relationship between Os, soil moisture and foliar injury and to
determine if there is an injury threshold or quantifiable relationship between these factors could
help inform policy. Additionally, research to understand the connection between Os-related
foliar injury and other physiological effects and ecosystem services could also be useful. We
further note that while this review relied on the robust E-R functions that are available for 11
tree and 10 crop species, there are tens of thousands of plant species in the U.S. (USD A, NRCS,
2014),26 66 of which have also been identified as Cb sensitive on National Park Service and US
Fish and Wildlife Service lands27. Research on additional tree as well as non-tree species that
would support the development of robust E-R functions would improve our understanding of the
full range of response of plant species to Os and our understanding of the overall risk to
vegetation. For example, studies using large numbers of native plant species across regions
where those species are indigenous, might be expected to reduce uncertainties associated with
extrapolating plant response for a given level of Os using composite response functions across
differing regions and climates. Studies focused on fruits and vegetables might assist in reducing
uncertainties associated with Os effects on agriculture. Particular focus is suggested on
organically grown vegetables that may receive less intensive management than conventionally
grown crops. Recent studies indicate that watermelons may be particularly sensitive to Os
26 USD A, NRCS. 2014. The PLANTS Database (http://plants.usda.gov. 3 January 2014). National Plant
Data Team, Greensboro, NC 27401-4901 USA.
27See http://www2.nature.nps.gov/air/Pubs/pdf/flag/NPSozonesensppFLAG06.pdf
6-59
-------
exposure (U.S. EPA, 2013, section 9.4.4.1) and older studies indicate grapes, honeydew melon,
lemons and oranges may also be Os sensitive (Abt Associates Inc., 1995).
Some new information has emerged linking effects on tree seedlings with larger trees and
similarities in results between exposure techniques (U.S. EPA 2013, section 9.6). Uncertainties
remain in this area as well as uncertainties in extrapolating from Os effects on juvenile to mature
trees and from trees grown in the open versus those in a closed forest canopy in a competitive
environment. The relationship between nocturnal exposures and plant uptake and response is
also an important subject for further research.
With respect to the second research area pertaining to the impact of Os on other welfare
effects categories such as climate, ecosystem components, and whole ecosystem structure and
function, 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 Os, plant response and
multiple biotic and abiotic stressors, including those associated with the ecosystem services that
would be affected (e.g., hydrology, productivity, carbon sequestration). Because these
uncertainties are multiple and significant due to the complex interactions involved, new research
will likely require a combination of manipulative experiments with model ecosystems,
community and ecosystem studies along natural Os gradients, and extensive modeling efforts to
project landscape-level, regional, national and international impacts of Os.
Uncertainties associated with projections of the effects of Os on the ecosystem processes
of water, carbon, and nutrient cycling, particularly at the stand and community levels might be
addressed through research on the effects on below ground ecosystem processes in response to
Os exposure alone and in combination with other stressors. These below-ground processes
include interactions of roots with the soil or microorganisms, effects of Os 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 Os exposures might improve our understanding of how Os may affect
biodiversity or genetic diversity. Such research could be strengthened by modern molecular
methods to quantify impacts on diversity. More tools and research would improve our
understanding of relationships between Os exposure and stressors such as insect infestations,
plant diseases, drought and potential stressors from climate change. It is also important to
understand how such interactions may affect ecosystem services such as CCh sequestration; food
and fiber production; wildlife habitat and water resources.
With respect to the research areas related to the development of approaches, tools, or
methodologies useful in characterizing Cb exposures and the relationship between Os-induced
6-60
-------
effects and associated ecosystem services and public welfare in a policy context, we note that
one of the most important uncertainties in this review is the characterization of air quality in rural
areas where there is limited monitoring. More comprehensive monitoring in these areas would
reduce uncertainties associated with Os exposures in many rural areas. Areas of particular
uncertainty include protected natural areas in the western U.S, including those at high elevation,
as well as those downwind of recently expanded oil and gas development areas. Uncertainties
associated with quantifying exposure in areas with and without monitors might be addressed
through additional work on interpolation methods and air quality models that are tailored to
estimating cumulative seasonal exposures, as well as improved model capabilities that use more
refined spatial grids and are better able to handle Cb movement in complex terrain.
Uncertainties related to characterizing the potential public welfare significance of Cb-
induced effects and impacts to associated ecosystem services could also be informed by research,
such as research intended to clarify the relationship between Cb exposure and fire risk and Cb
exposure and forest susceptibility to bark beetle infestation. Research relating known Cb
ecological effects such as reproductive effects 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.
6-61
-------
6.9 REFERENCES
Abt Associates, Inc. (1995). Ozone NAAQS benefits analysis: California crops. Report to U.S. EPA, July.
Federal Register 1996. National Ambient Air Quality Standards for Ozone; Proposed Rule. 40 CFR 50; Federal
Register 61: 65716
Federal Register 1997. National Ambient Air Quality Standards for Ozone; Final Rule. 40 CFR 50; Federal Register
62:38856
Federal Register 2007. National Ambient Air Quality Standards for Ozone; Proposed Rule. 40 CFR 50; Federal
Register 72: 37818
Federal Register 2008. National Ambient Air Quality Standards for Ozone; Final Rule. 40 CFR 50 and 58; Federal
Register 73:16436
Federal Register 2010. National Ambient Air Quality Standards for Ozone; Proposed Rule. 40 CFR 50 and 58;
Federal Register 75 FR: 2938
Federal Register 2012. National Ambient Air Quality Standards for Oxides of Nitrogen and Sulfur; Final Rule. 40
CFR 50; Federal Register 77 FR 20218
Frey, C., (2014a) CASAC Review of the EPA's Second Draft Policy Assessmentfor the Review of the Ozone
National Ambient Air Quality Standards. EPA-CASAC-14-004. June 26, 2014.
Frey, C. (2014b) CAS AC Review of the EPA's Welfare Risk and Exposure Assessment for Ozone (Second
External Review Draft). EPA-CASAC-14-003. June 18, 2014.
Gregg, JW; Jones, CG; Dawson, TE. (2003). Urbanization effects on tree growth in the vicinity of New York City
[Letter]. Nature 424:183-187. http://dx.doi.org/10.1038/nature01728
Heck, WW; Cowling, EB. (1997).The need for a long term cumulative secondary ozone standard - An ecological
perspective. EM January: 23-33.
Henderson, R. (2006) Letter from CASAC Chairman Rogene Henderson to EPA Administrator Stephen Johnson.
October 24, 2006, EPA-CASAC-07-001.
Henderson, R. (2007) Letter from CASAC Chairman Rogene Henderson to EPA Administrator Stephen Johnson.
March 26, 2007, EPA-CASAC-07-002.
Henderson, R. (2008) Letter from CASAC Chairman Rogene Henderson to EPA Administrator Stephen Johnson.
April 7,2008, EPA-CASAC-08-009.
Lefohn, AS; Laurence, JA; Kohut, RJ. (1988). A comparison of indices that describe the relationship between
exposure to ozone and reduction in the yield of agricultural crops. Atmos Environ 22: 1229-1240.
http://dx.doi.org/10.1016/0004-6981(88)90353-8
Lefohn, AS; Runeckles, VC. (1987). Establishing standards to protect vegetation - ozone exposure/dose
considerations. Atmos Environ 21: 561-568. http://dx.doi.org/10.1016/0004-6981(87)90038-2
Samet, J.M. (2010) Clean Air Scientific Advisory Committee (CASAC) Review of EPA's Proposed Ozone National
Ambient Air Quality Standard. EPA-CASAC-10-007. January 19, 2010. Available online at:
http://yosemite.epa.gov/sab/sabproduct.nsf/264cbl227d55e02c85257402007446a4/610BB57CFAC8A41C
852576CF007076BD/$File/EPA-CASAC-10-007-unsigned.pdf
6-62
-------
USD A, NRCS. 2014. The PLANTS Database (http://plants.usda.gov. 3 January 2014). National Plant Data Team,
Greensboro, NC 27401-4901 USA
U.S. EPA (1996a). Air quality criteria for ozone and related photochemical oxidants [EPA Report]. (EPA/600/P-
93/004AF). U.S. Environmental Protection Agency, Research Triangle Park, NC.
U.S. EPA (1996b). Review of national ambient air quality standards for ozone: Assessment of scientific and
technical information: OAQPS staff paper [EPA Report]. (EPA/452/R-96/007). Research Triangle Park,
NC. http://www.ntis.gov/search/product.aspx?ABBR=PB96203435
U.S. EPA (2006). Air Quality Criteria for Ozone and Related Photochemical Oxidants (2006 Final). U.S.
Environmental Protection Agency, Washington, DC. EPA/600/R-05/004aF-cF. March 2006. Available at:
http://www.epa.gOv/ttn/naaqs/standards/ozone/s o3crcd.html
U.S. EPA (2007). Review of the national ambient air quality standards for ozone: Policy assessment of scientific and
technical information: OAQPS staff paper [EPA Report]. (EPA/452/R-07/003). Research Triangle Park,
NC. http://www.epa.gov/ttn/naaqs/standards/ozone/data/2007_01_ozone_staff_paper
U.S. EPA (2013). Integrated Science Assessment of Ozone and Related Photochemical Oxidants (Final). U.S.
Environmental Protection Agency, Washington, DC. EPA/600/R-10/076F
U.S. EPA (2014a). Policy Assessment for the Review of National Ambient Air Quality Standards for Ozone,
Second External Review Draft. Office of Air Quality Planning and Standards, Research Triangle Park, NC,
27711. EPA-452/P-14-002
U.S. EPA (2014b). Welfare Risk and Exposure Assessment for Ozone, Second External Review Draft. Office of
Air Quality Planning and Standards, Research Triangle Park, NC, 27711. EPA-452/P-14-003a
U.S. Forest Service, National Park Service, and U.S. Fish and Wildlife Service. 2010. Federal land managers' air
quality related values work group (FLAG): phase I report—revised (2010). Natural Resource Report
NPS/NRPC/NRR—2010/232. National Park Service, Denver, Colorado.
Wolff, G.T. (1996) Letter to EPA Administrator Carol Browner, RE: Closure by the Clean Air Scientific Advisory
Committee (CASAC) on the Secondary Standard Portion of the Staff Paper for Ozone. EPA-SAB-CASAC-
LTR-96-002, April 4, 1996.
6-63
-------
APPENDICES
Appendix 2A. Supplemental Air Quality Modeling Analyses of Background Os 2A-1
Appendix 2B. Monitoring Data Analysis of Relationships Between Current Standard and W126
Metric 2B-1
Appendix 2C. Inter-annual Variability in W126 Index Values: Comparing Annual and 3-Year Average
Metrics (2008-2010) 2C-1
Appendix 3 A. Recent Studies of Respiratory-Related Emergency Department Visits and Hospital
Admissions 3A-1
Appendix 3B. Ambient Os Concentrations in Locations of Health Studies 3B-1
Appendix 5A. Os-Sensitive Plant Species Used by Some Tribes 5A-1
Appendix 5B. Class I Areas Below Current Standard And Above 15 ppm-hrs 5B-1
Appendix 5C. Expanded Evaluation of Relative Biomass and Yield Loss 5C-1
-------
APPENDIX 2A
SUPPLEMENTAL AIR QUALITY MODELING ANALYSES OF
BACKGROUND OZONE
Table of Contents
List of Figures 2
List of Tables 4
1. Introduction 5
2. Description of modeling methodologies 6
a. 2007 GEOS-Chem/CMAQ zero-out modeling 7
b. 2007 GEOS-Chem/CAMx source apportionment modeling 12
3. Estimates of seasonal-average background ozone levels 15
4. Distributions of background ozone levels 22
5. Contribution of various processes and sources to total background ozone 30
6. Estimates of the fractional background contribution to total ozone in 12 specific
areas 37
7. Background ozone and W126 39
8. Summary 39
9. References 42
2A-1
-------
List of Figures
Figure la. Modeling domain used in 2007 CMAQand CAMx modeling 10
Figure Ib. Density scatterplot comparing CMAQ base daily peak 8-hour ozone predictions
against observed 8-hour ozone peaks paired in space and time for all sites during April-
October 2007 10
Figure Ic. Bias in seasonal mean (April-October) maximum daily 8-hour ozone predictions in
the 2007 CMAQ base simulation 11
Figure Id. Relationship between CMAQ estimations of MDA8 natural background ozone and
daily model biases 11
Figure 2a. Density scatterplot comparing CAMx base daily peak 8-hour ozone predictions
against observed 8-hour ozone peaks paired in space and time for all sites during April-October
2007 14
Figure 2b. Bias in seasonal mean (April-October) maximum daily 8-hour ozone predictions in
the 2007 CAMx base simulation 14
Figure 3a. April-October average MDA8 ozone (ppb) at monitoring locations across the U.S. as
estimated by a 2007 CMAQ base simulation 18
Figure 3b. April-October average natural background MDA8 ozone (ppb) at monitoring
locations across the U.S. as estimated by a 2007 CMAQ zero out simulation 18
Figure 3c. April-October average North American background MDA8 ozone (ppb) at monitoring
locations across the U.S. as estimated by a 2007 CMAQ zero out simulation 19
Figure 3d. April-October average United States background MDA8 ozone (ppb) at monitoring
locations across the U.S. as estimated by a 2007 CMAQ zero out simulation 19
Figure 4a. Ratio of natural background to total April-October average MDA8 ozone at
monitoring locations across the U.S. as estimated based on 2007 CMAQ simulations 20
Figure 4b. Ratio of N. American background to total April-October average MDA8 ozone at
monitoring locations across the U.S. as estimated based on 2007 CMAQ simulations 20
Figure 4c. Ratio of U.S. background to total April-October average MDA8 ozone at
monitoring locations across the U.S. as estimated based on 2007 CMAQ simulations 21
Figure 4d. Ratio of sources other than U.S. anthropogenic emissions to total April-October
average MDA8 ozone at monitoring locations across the U.S. as estimated by a 2007 CAMx
source apportionment simulation 21
Figure 5a. Distribution of natural background MDA8 ozone (ppb) at monitoring locations
across the U.S. (Apr-Oct), binned by base modeled site-day MDA8, as estimated by 2007
2A-2
-------
CMAQ simulations 25
Figure 5b. Distribution of N. American background MDAS ozone (ppb) at monitoring
locations across the U.S. (Apr-Oct), binned by base modeled site-day MDAS, as estimated by
2007 CMAQ simulations 25
Figure 5c. Distribution of U.S. background MDAS ozone (ppb) at monitoring locations across
the U.S. (Apr-Oct), binned by base modeled site-day MDAS, as estimated by 2007 CMAQ
simulations 26
Figure 5d. Distribution of MDAS ozone contributions from non-U.S. manmade sources (ppb)
at monitoring locations across the U.S. (Apr-Oct), binned by base modeled site-day MDAS, as
estimated by 2007 CAMx simulations 26
Figure 6a. Distribution of natural background MDAS ozone fractions at monitoring locations
across the U.S. (Apr-Oct), binned by base modeled site-day MDAS, as estimated by 2007
CMAQ simulations 27
Figure 6b. Distribution of N. American background MDAS ozone fractions at monitoring
locations across the U.S. (Apr-Oct), binned by base modeled site-day MDAS, as estimated by
2007 CMAQ simulations 27
Figure 6c. Distribution of U.S. background MDAS ozone fractions at monitoring locations
across the U.S. (Apr-Oct), binned by base modeled site-day MDAS, as estimated by 2007
CMAQ simulations 28
Figure 6d. Distribution of MDAS ozone fractions from non-U.S. anthropogenic sources at
monitoring locations across the U.S. (Apr-Oct), binned by base modeled site-day MDAS, as
estimated by the 2007 CAMx simulation 28
Figure 7. April-October 95th percentile United States background MDAS ozone (ppb) at
monitoring locations across the U.S. as estimated by a 2007 CMAQ base simulation 29
Figure 8a. Difference in April-October average MDAS ozone (ppb) at monitoring locations
across the U.S. between the USB scenario and the NAB scenario 32
Figure 8b. Difference in April-October average MDAS ozone (ppb) at monitoring locations
across the U.S. between the NAB scenario and the NB scenario 32
Figure 9a. Percentage of April-October average MDAS ozone that is apportioned to
boundary conditions as estimated at monitoring locations by a 2007 CAMx simulation 33
Figure 9b. Percentage of April-October average MDAS ozone that is apportioned to U.S.
anthropogenic sources as estimated at monitoring locations by a 2007 CAMx simulation 33
Figure 9c. Percentage of April-October average MDAS ozone that is apportioned to purely
biogenic emissions as estimated at monitoring locations by a 2007 CAMx simulation 34
2A-3
-------
Figure 9d. Percentage of April-October average IV1DA8 ozone that is apportioned to
climatologicalfire emissions as estimated at monitoring locations by a 2007 CAMx simulation ... 34
Figure 9e. Percentage of April-October average IV1DA8 ozone that is apportioned to
anthropogenic emissions from in-domain Canadian and Mexican sources as estimated at
monitoring locations by a 2007 CAMx simulation 35
Figure 9f. Percentage of April-October average IV1DA8 ozone that is apportioned to Category 3
marine vessel emissions beyond U.S. territorial waters as estimated at monitoring locations by
a 2007 CAMx simulation 35
Figure 9g. Percentage of April-October average MDA8 ozone that is apportioned to Gulf of
Mexico point sources as estimated at monitoring locations by a 2007 CAMx simulation 36
List of Tables
Table la. April-October average MDA8 ozone, average MDA8 ozone from sources other than
U.S. manmade emissions, and the fractional contribution of these background sources in the
12 REA urban study areas, as estimated by a 2007 CAMx simulation 37
Table Ib. Average MDA8 ozone, average MDA8 ozone from sources other than U.S.
manmade emissions, and the fractional contribution of these background sources in the 12 REA
areas, as estimated by a 2007 CAMx simulation using site-days in which base MDA8 ozone
exceeded 60 ppb 38
Table Ic. Fractional contribution of non-U.S. manmade emissions sources in the 12 REA urban
study areas, as estimated by a 2007 CAMx simulation using means and medians of daily MDA8
fractions 38
Table Id. April-October average MDA8 ozone, average MDA8 ozone from USB, and the
fractional contribution of these background sources in the 12 REA urban study areas, as
estimated by two separate 2007 CMAQ simulations 38
2A-4
-------
1. Introduction
One of the aspects of ozone that is unusual relative to the other pollutants with National
Ambient Air Quality Standards (NAAQS) is that, periodically, in some locations, an appreciable fraction of
the observed ozone results from sources or processes other than local and regional anthropogenic
emissions of ozone precursors (Fiore et al., 2002). Any ozone formed by processes other than the
chemical conversion of local or regional ozone precursor emissions, such as nitrogen oxides (NOx) or
volatile organic emissions (VOC), is generically referred to as "background" ozone. As part of this review
of the ozone NAAQS, EPA completed an extensive review of the known aspects of background ozone
and summarized the findings in the Integrated Science Assessment (ISA) in March 2013 (USEPA, 2013).
The purpose of this appendix is to present the results from supplemental air quality modeling analyses
related to background ozone that were completed by EPA subsequent to the ISA. While these updated
analyses use a recent base year (2007) and consider an alternative modeling methodology which can
better account for non-linear ozone chemistry in some conditions, the results are largely consistent with
previous determinations about the magnitude of background ozone contributions across the U.S.
Away from the surface, ozone can have an atmospheric lifetime on the order of weeks. As a
result, background ozone can be transported long distances at heights above the boundary layer and,
when meteorological conditions are favorable, be available to mix down to the surface and add to the
total ozone loading from non-background sources. Generically, background ozone can originate from
natural sources of ozone and ozone precursors, as well as from far upwind manmade emissions of ozone
precursors. Natural sources of ozone precursor emissions such as wildfires, lightning, and vegetation
can lead to ozone formation by chemical reactions with other natural sources1. Another important
natural component of background is ozone that is naturally formed in the stratosphere through
interactions of UV light with atomic oxygen (O2). Stratospheric ozone can periodically mix down to the
surface at high concentrations, especially at higher altitude locations. The manmade portion of the
background includes any ozone formed due to anthropogenic sources of ozone precursors emitted far
away from the local area (e.g., international emissions). Finally, both biogenic and international
anthropogenic emissions of methane, which can be chemically converted to ozone over relatively long
time scales, can also contribute to global background ozone levels.
The precise definition of background ozone can vary depending upon context, but it generally
refers to ozone that is formed by sources or processes that cannot be influenced by actions within the
jurisdiction of concern. In the first draft policy assessment document (EPA, 2012), EPA presented three
specific definitions of background ozone: natural background, North American background, and U.S.
background. Natural background (NB) was the narrowest definition of background and it was defined as
the ozone that would exist in the absence of any manmade ozone precursor emissions. The other two
previously-established definitions of background presume that the U.S. has little influence over
anthropogenic emissions outside our continental or domestic borders. North American background
(NAB) is defined as that ozone that would exist in the absence of any manmade ozone precursor
1 Ozone formed through reactions between natural emissions and local anthropogenic emissions (e.g., biogenic
VOC with man-made NOx) is generally not considered to be background ozone.
2A-5
-------
emissions inside of North America. U.S. background (USB) is defined as that ozone that would exist in
the absence of any manmade emissions inside the United States. It is important to note that each of
these three definitions of background ozone requires photochemical modeling simulations to estimate
what the residual ozone concentrations would be were the various anthropogenic emissions to be
removed.
As noted in the first draft policy assessment, EPA has revised several aspects of our
methodology for estimating the change in health risk and exposure that would result from a revision to
the ozone NAAQS. First, risk estimates are now based on total ozone concentrations as opposed
previous reviews which only considered risk above background levels. Second, EPA is now using air
quality models to estimate the spatial patterns of ozone that would result from attaining various levels
of the NAAQS, as opposed to simplistic rollback techniques that required the estimation of a background
ozone "floor" beyond which the rollback would not take place. Both of these revisions have had the
indirect effect of obviating the need for estimating background ozone levels as part of the ozone risk
and exposure assessment (REA). Regardless, EPA expects that a well-founded understanding of the
fractional contribution of background sources and processes to surface ozone levels will be valuable
towards informing policy decisions about the NAAQS. Section 2 of this document will describe the
supplemental air quality modeling simulations that have recently been completed by EPA to bolster our
understanding of background ozone. Section 3 will present the results from the updated analyses and
provide estimates of average background ozone levels, and how they can vary in time and space across
the U.S. Based on the same modeling, Section 4 will consider the entire spectrum of variable
background ozone levels with special emphasis on areas and times in which background can approach or
exceed the level of the NAAQS. Section 5 will utilize the supplemental air quality modeling estimates to
determine the relative importance of specific components of background ozone. Section 6 will present
estimates of the overall fraction of ozone that is estimated to result from background sources or
processes in each of the 12 urban case study areas in the epidemiology study based analyses in Chapter
7 of the Risk and Exposure Assessment (REA) (EPA, 2014) based on the updated modeling. Finally,
Section 7 will conclude with a limited analysis of how background ozone levels impact longer-term
ozone metrics that may be important from a welfare perspective (i.e., W126).
2. Description of modeling methodologies
As noted above, air quality models are typically used to estimate background ozone as it is quite
difficult to measure directly. Without special monitoring, it is impossible to determine how much of the
ozone measured by a monitor originated from sources that are considered background. Even the most
remote monitors within the U.S. can periodically be affected by U.S. anthropogenic emissions. Previous
modeling studies have estimated what background levels would be in the absence of certain sets of
emissions by simply comparing the ozone differences between a base model simulation and a control
simulation in which emissions were removed. This basic approach is often referred to as "zero out"
modeling or "emissions perturbation" modeling. Examples of zero out modeling include the three major
studies summarized in the ISA (Zhang et al., 2011; Emery et al., 2012, Lin et al., 2013). It is important to
note that the specific concepts of NB, NAB, and USB are all explicitly tied to zero-out modeling, as those
definitions are based on estimating what remains in the absence o/specific sets of man-made emissions.
2A-6
-------
EPA has conducted and will describe updated air quality modeling for a 2007 base year that employs a
regional air quality model nested within a coarser-scale global chemical transport model to estimate NB,
NAB, and USB levels when the respective manmade emissions are zeroed. This modeling is described in
detail in section 2a.
While the zero-out approach has traditionally been used to estimate background ozone levels,
the methodology has some acknowledged limitations. First, from a policy perspective, the purely
hypothetical and ultimately unrealizable zero manmade emissions scenarios have limited application in
this regard. Secondly, the assumption that background ozone is what is left after specific emissions have
been removed within the model simulation can be misleading in locations where ozone chemistry is
highly non-linear. Depending upon the local composition of ozone precursors, NOx emissions
reductions can either increase or decrease ozone levels in the immediate vicinity of those reductions.
For those specific urban areas in which NOx titration of ozone can be significant, zero-out modeling can
result in inflated estimates of background ozone when these NOx emissions are completely and
unrealistically removed. Paradoxically, in certain times and locations in a zero-out scenario there can be
more background ozone than actual ozone within the model (EPA, 2014).
A separate modeling technique attempts to circumvent these limitations by apportioning the
total ozone within the model to its contributing source terms. This basic approach is referred to as
"source apportionment" modeling. While source apportionment modeling has not been previously used
in the context of estimating background ozone levels as part of an ozone NAAQS review, it has
frequently been used in other regulatory settings to estimate the "contribution" to ozone of certain sets
of emissions (EPA 2005, EPA 2011). The source apportionment technique provides a means of
estimating the contributions of user-identified source categories to ozone formation in a single model
simulation. This is achieved by using multiple tracer species to track the fate of ozone precursor
emissions (VOC and NOx) and the ozone formation caused by these emissions. The methodology is
designed so that all ozone and precursor concentrations are attributed to the selected source categories
at all times without perturbing the inherent chemistry. The zero out modeling attempts to determine
what ozone be in the absence of background sources. The source apportionment modeling attempts to
determine how much of the modeled ozone has resulted from background sources. EPA has conducted
and will describe new source apportionment modeling that employs a regional air quality model nested
within a coarser-scale global chemical transport model to assess the contributions of boundary
conditions and other potential background sources (e.g., wildfires, biogenic emissions, and
Canadian/Mexican emissions). This modeling is described in detail in section 2b.
a. 2007 GEOS-Chem/CMAQ zero-out modeling:
In order to provide estimates of the overall fraction of ozone that is estimated to result from
background sources in each of the 12 REA urban study areas, EPA conducted new modeling that utilized
the same model base year (2007) as was used in the ozone modeling that inform the risk and exposure
analyses (EPA, 2014, Appendix 4b). The EPA modeling used a model configuration similar to that of
Emery (2012), in that it nested a regional-scale (12 km) air quality model inside a global air quality model
2A-7
-------
simulation with a much coarser horizontal grid resolution (2.0 by 2.5 degrees). Figure la shows a map of
the model domain.
The global scale simulation utilized the GEOS-Chem model, version v8-03-02, except for the
chemistry package which was from version v8-02-01. The emissions estimates used in the 2007 base
year modeling were aggregated from a variety of sources, starting with the global Emissions Database
for Global Atmospheric Research (EDGAR) emission inventory. These initial estimates were then
improved by utilizing various area-specific inventories, such as the 2005 National Emissions Inventory
(NEI) for the U.S. portions of the domain, and available inventories for Asia, Canada, Europe, and
Mexico. In addition to the anthropogenic estimates, emissions were specified for a variety of
background sources including: lightning NO, soil NOx, wildfires, and biogenic VOC emissions. The
wildfire data is from the Global Fire Emissions Database (GFED). The biogenic VOC estimates were
simulated by the Model of Emissions of Gases and Aerosols from Nature (MEGAN) version 2.1. The
meteorological data is based on the Goddard Earth Observing System Model, Version 5 (GEOS-5)
analysis fields. More information on the global simulation is available within Henderson et al. (2013).
This reference also provides a broad evaluation of the ability of this specific GEOS-Chem configuration to
provide accurate lateral boundary conditions of ozone to finer-scale regional simulations. Using satellite
retrievals from the Tropospheric Emissions Spectrometer (TES), Henderson et al. (2013) concluded that
the GEOS-Chem ozone prediction biases and errors are generally within TES uncertainty estimates. For
instance, for the ozone season month of August, model predictions are within plus or minus 20 percent
of the satellite estimates between nearly 80 percent of the time, with slightly better performance along
the southern boundary.
The lateral boundary conditions from the global model were then used as inputs for a 12 km
horizontal resolution, CMAQ version 4.7.1, model simulation. Four scenarios were modeled: 1) a 2007
base case simulation which was the basis of the air quality modeling performed for the 2nd draft ozone
REA and is described in more detail in Appendix 4b of EPA (2014), 2) a natural background run with
anthropogenic ozone precursor emissions2 removed in both the global and regional models, 3) a North
American background run with anthropogenic ozone precursor emissions removed across North
America (global and regional model simulations), and 4) a U.S. background run with anthropogenic
ozone precursor emissions were removed over the U.S (global and regional model simulations).
Detailed analyses of EPA's 2007 zero out modeling results are provided in sections 3 through 6 of this
appendix.
An operational model performance evaluation was completed for surface ozone in the 2007
base simulation as described separately (EPA 2014, Appendix 4b). For the purposes of this analysis, EPA
assessed the model ability to reproduce measured daily maximum 8-hour (MDA8) ozone values and
seasonal mean MDA8 ozone concentrations for the period April to October 2007. As noted earlier, the
base year modeling in this analysis used climatological monthly-average wildfire emissions which are not
2 In the global model all ozone precursor species were removed (i.e., VOC, NOx, CO), except for methane which
was reset to pre-industrial levels to reflect natural contributions. In the regional modeling, the methane levels
were left unchanged.
2A-8
-------
intended to capture discrete events from specific fires that occurred in 2007, so perfect correlation
between observations and model predictions should not be expected. Figure Ib provides a density
scatterplot of the observed and predicted daily 8-hour ozone peaks paired in space and time for the
2007 CMAQ base. As can be seen, the majority of pairs line up along the 1:1 line. There is a tendency
for the model to overestimate site-days with low 8-hour ozone peaks, and underestimate the site-days
with higher peak ozone values. Modeled 8-hour ozone peak concentrations exhibited relatively small
bias and error compared to the observations. The average bias in IV1DA8 ozone estimates was 3.5 ppb.
Figure Ic depicts the spatial bias patterns in IV1DA8 ozone at all sites that measured valid ozone data for
at least 100 days during the April-October period. CMAQ overestimations are greatest along the Gulf
Coast region, along the Atlantic coastline, and over the central U.S. The majority of underestimated
seasonal mean MDA8 occurs within southern California. The model performance for the 2007 base
simulation is equivalent or better than typical state-of-the-science photochemical model performance
recently reported in the literature (Simon et al, 2012).
Certainly some remote monitoring locations are more affected by background sources than
other locations in the network. However, this and numerous other analyses have shown that even the
most remote ozone monitoring locations in the U.S. are periodically affected by U.S. manmade
emissions. In this analysis we carefully assess model performance to ensure that model error does not
influence the characterization of background ozone. As noted in the recent ISA (EPA, 2013), there is
greater confidence in the ability of the model to predict mean contributions from background sources
rather than individual events. Beyond the statistical analyses summarized in the previous paragraph and
in appendix 4b of the 2nd draft ozone REA (EPA, 2014), it is valuable to attempt to diagnose the model
ability to account for background ozone within the simulation. EPA assessed whether any correlation
existed between daily model biases and daily background ozone estimates. Figure Id shows that at
high-elevation sites (i.e., sites more than 1km above sea level) the highest estimates of natural
background ozone tend occur on days with greatest overestimation. Conversely, the site-days with the
lowest natural background estimates tend to occur when the model underestimates the observed daily
peaks at these sites. This relationship between background estimates and simulation bias appears to be
constrained to the mountainous portion of the Western U.S. Figure Id also shows that estimates of
natural background ozone greater than 60 ppb are associated with large over-predictions. However,
based on the relatively low model bias and the general lack of correlation between daily bias values and
background estimates, EPA believes that these model estimates can be used to help characterize
background ozone levels over the U.S. Although the highest background estimates should be
considered with caution.
2A-9
-------
x,y origin: -M12MOiri. ylGZQOuOm ... ' -.
col: 396 row:Z46 / \ X f ^
Figure la. Modeling domain used in 2007 CMAQand CAMx modeling.
Q 9 -
50
150
MDA8 observations (ppb)
Figure Ib. Density scatterplot comparing CMAQ base daily peak 8-hour ozone predictions against
observed 8-hour ozone peaks paired in space and time for all sites during April-October 2007.
2 A-10
-------
Figure Ic. Bias in seasonal mean (April-October) maximum daily 8-hour ozone predictions in the 2007
CMAQ base simulation.
_BO- t
•0 t
0. . *
0. .
§5
§
!*-
c
O)
l n
i
1 T
20-
*
r
1 q
1 J L_
• • • I
* • *
, * *
t
t '
—
•
• . i
JTT'
it
!
1 T
Model Bias Ippb)
h~p > 10 ppb under
^,
t— H 6-10 ppb under
t^j 0-3 ppb under
(— H 0-3 ppb aver
^ 3-6 ppb over
f— I 6-10 ppb over
[— ~j > 10 ppb over
* ,
* * '
Low Med High
Elevation of site \m MSU: Low r 500): Med (500-1000): High i 1000)
Figure Id. Relationship between CMAQ estimations of MDA8 natural background ozone and daily
model biases.
2 A-11
-------
b. 2007 GEOS-Chem/CAMx source apportionment modeling:
The same global modeling described above was used to assign lateral boundary conditions to
the regional-scale (12 km) CAMx v5.0 source apportionment simulations. Wherever possible, the
emissions and meteorological inputs in the CAMx modeling were chosen to mimic the 2007 base CMAQ
simulation described earlier. Figure la shows a map of the model domain.
As with the CMAQ base case, a limited operational model performance evaluation was also
completed for surface ozone in the 2007 base simulation. For the purposes of this analysis, EPA
assessed the model ability to reproduce measured daily maximum 8-hour (MDA8) ozone values and
seasonal mean MDA8 ozone concentrations for the period April to October 2007. Figure 2a provides a
density scatterplot of the observed and predicted daily 8-hour ozone peaks paired in space and time for
the 2007 CAMx base simulation. As can be seen, the majority of pairs line up along the 1:1 line. Again,
there is a tendency for the model to overestimate site-days with low 8-hour ozone peaks and
underestimate the site-days with higher peak ozone values. Modeled 8-hour ozone peak concentrations
exhibited relatively small bias and error compared to the observations. The average bias in MDA8 ozone
estimates was 3.5 ppb. Figure 2b depicts the spatial bias patterns in MDA8 ozone at all sites that
measured valid ozone data for at least 100 days during the April-October period. CAMx overestimations
are greatest along the Gulf Coast region, along the Atlantic and Pacific coastlines, and within the
southeastern U.S. The majority of underestimated seasonal mean MDA8 occurs in California away from
the coastline.
The apportionment tools in CAMx utilized here to estimate the contribution of background
sources are well-established and have previously been peer-reviewed (UNC, 2009). EPA used the
Anthropogenic Precursor Culpability Assessment (APCA) tool in this analysis. The APCA tool attributes
ozone production to manmade sources whenever ozone is determined to result from a combination of
anthropogenic and biogenic emissions (Environ, 2011). The APCA methodology defines natural ozone as
the production resulting from the interaction of biogenic VOC with biogenic NOx emissions. Eleven
separate source categories were tracked in the source apportionment analysis, including five boundary
condition terms and six in-domain sectors:
• Boundary condition terms:
o Northern edge
o Eastern edge
o Southern edge
o Western edge
o Top boundary
• In-domain sectors:
o U.S. anthropogenic emissions
o Point sources located within the Gulf of Mexico
o Category 3 marine vessels outside State boundaries
o Climatologically-averaged wildfire emissions
2 A-12
-------
o Biogenic emissions
o Canada/Mexico emissions (only those sources within the domain)
It should be noted that the source apportionment modeling conducted here does not allow for
replication of natural background because of the construct of boundary conditions. The boundary
conditions for our applications can include ozone and/or ozone precursors that were originally
generated by natural sources, as well as ozone produced from far upstream anthropogenic emissions
(e.g., Asia). It is not possible to disentangle these two terms. Instead, the source apportionment
modeling is primarily used to help estimate background into the U.S., which is assumed to be the
contributions from nine of the modeled sectors; that is, everything except U.S. anthropogenic emissions
and point sources located within the Gulf of Mexico.
<
Q
50
100
150
MDA8 observations (ppb)
Figure 2a. Density scatterplot comparing CAMx base daily peak 8-hour ozone predictions against
observed 8-hour ozone peaks paired in space and time for all sites during April-October 2007.
2 A-13
-------
°£ 8>*
o
2b. Bias in seasonal mean (April-October) maximum daily 8-hour ozone predictions in the 2007
CAMx base simulation.
3. Estimates of seasonal-average background ozone levels
This section of the appendix provides estimates of seasonal average background ozone levels
over the U.S. As noted in the introduction and as discussed in detail in the ISA, background ozone values
can vary significantly in space and time. There can be atypical episodes of higher background ozone
concentrations amidst the routine days that drive seasonal average background. The highest
background episodic concentrations are typically associated with stratospheric intrusions or wildfires.
These background "events" can be difficult to model as they require event-specific model inputs. The
primary goal of the EPA modeling was to estimate the seasonal average background concentrations
between April and October 2007. Previous analyses have shown that this is the period in which average
background levels are highest (Zhang et al., 2011). This section of the appendix focuses on seasonal
mean levels of background. (Section 4 will consider the upper range of possible background ozone.)
The analysis focus on the maximum daily 8-hour ozone average in ppb. This metric is referred to
as MDA8. This section will first present model estimates of seasonal mean ozone levels in the base
simulation. This will be followed by estimates of NB, NAB, and USB from the CMAQ zero out modeling.
After discussing the magnitudes of background levels, the section switches to a consideration of the
relative percentage of background to total ozone across the U.S. This portion of the text will utilize both
the CMAQ zero out and CAMx source apportionment modeling.
Figure 3a displays the 2007 base case, CMAQ model-predicted, seasonal mean (April-October)
MDA8 ozone concentrations in grid cells with active monitoring locations over the U.S. The model
results are shown at the monitoring site level as opposed to in the default gridded format to foster
subsequent site-level estimates of background magnitudes. Each grid cell containing an Air Quality
System (AQS) ozone monitor that was collecting valid data in 2007 was identified and the model
2 A- 14
-------
background estimates were extracted for those grid cells and displayed accordingly. The base
predictions are provided for context to allow easier interpretation of the following plots which isolate
specific background levels. As can be seen, most of the U.S. experiences seasonal mean IV1DA8 ozone
levels greater than 50 ppb in the base case simulation. The median value over the 1,294 monitoring
locations is 52.5 ppb.
Figure 3b provides an estimate of what seasonal-average IV1DA8 would be in a natural
background scenario, using the 2007 EPA zero out modeling. Again, in this GEOS-Chem/CMAQ
simulation, all anthropogenic ozone precursor emissions were removed from both the global and
regional simulations, and methane levels were adjusted to pre-industrial levels in the global simulation.
As shown, natural background ozone levels range from approximately 15-35 ppb with the highest values
occurring over the higher-elevation sites in the western U.S. The median value over these locations is
24.2 ppb, and more than 50 percent of the sites have natural background levels of 20-25 ppb. The
highest modeled estimate of seasonal average, natural background, IV1DA8 ozone is 34.3 ppb at the
high-elevation CASTNETsite (Gothic) in Gunnison County, CO.
Figures 3c and 3d show the same information for the North American and U.S. background
scenarios. In these model runs, all anthropogenic ozone precursor emissions were removed from the
U.S., Canada, and Mexico (NAB scenario) and then only the U.S. (USB scenario). The figures show that
there is not a large difference between the NAB and USB scenarios. Seasonal mean IV1DA8 NAB and USB
ozone levels range from 25-50 ppb, with the most frequent values estimated in the 30-35 ppb bin. The
median seasonal mean background levels are 31.5 and 32.7 ppb (NAB and USB, respectively). Again, the
highest levels of background are predicted over the intermountain western U.S. Locations with NAB and
USB concentrations greater than 40 ppb are confined to Colorado, Nevada, Utah, Wyoming, northern
Arizona, eastern California, and parts of New Mexico. Similar to NB, the highest NAB and USB levels
were modeled to occur at the Gothic CO site (46.7/47.7). This remote rural site is located 2,926 meters
(9,600 feet) above mean sea level and should not be considered representative of background ozone at
lower-altitude, more-populated regions. The high USB and NAB values along the Gulf Coast are most
likely due to model biases.
Absolute model estimates of various background definitions are useful, but they can be
influenced by any local biases and errors in the modeling. A separate way to look at the role of
background in seasonal mean ozone levels is to consider the fractional contribution of NB, NAB, and USB
to total ozone at each location. Considering the proportional role of background allows for an
informative comparison between the two modeling approaches without having to account for the
differences in base case biases and errors. Figures 4a, 4b, and 4c show the estimated fractional
contribution of NB, NAB, and USB to total seasonal average MDA8 ozone levels at the monitoring
locations from the CMAQ zero out modeling. The modeling estimates that approximately 35-80 percent
of the seasonal average MDA8 ozone at monitoring locations is due to natural background sources. A
majority sites have NB fractions between 40 and 60 percent. The mean natural background proportion
over all sites is 47 percent. That is, when all global anthropogenic emissions are removed and global
methane levels in GEOS-Chem are restored to pre-industrial levels, seasonal average MDA8 levels are
reduced by approximately half. The fractional proportions of NAB and USB are very similar. In both
2 A-15
-------
cases, most sites have background fractions that range from 50 to 80 percent. The mean NAB fraction
(to seasonal mean MDA8) is 63 percent. The mean USB fraction is 66 percent.
As noted in the introduction, the advantage of the source apportionment modeling is that all of
the modeled ozone is attributed to various source terms and thus this approach is not affected by the
confounding occurrences of background ozone values exceeding the base ozone values as can happen in
the zero out modeling (i.e., background proportions > 100%). Consequently, one would expect the
fractional background levels to be lower in the source apportionment methodology as a result of
removing this artifact. It is also important to remember that the terms NB, NAB, and USB are explicitly
linked to the zero out modeling approach. (USB is the ozone that would exist in the absence of U.S.
anthropogenic emissions.) In contrast, the source apportionment modeling performed here provides
estimates the amount of IV1DA8 ozone that is attributable to U.S. anthropogenic emissions relative to
total base model ozone. Figure 4d shows the relative contribution from sources other than U.S.
anthropogenic emissions to total seasonal mean MDAS ozone based on the 2007 source apportionment
modeling. The fractional contribution fields between CMAQ zero out USB estimates and CAMx source
apportionment estimates of source other than U.S. anthropogenic emissions are quite similar. The
spatial patterns in Figures 4c and 4d are consistent, with the highest fractional contributions from
sources other than U.S. anthropogenic emissions occurring along U.S. borders and over the
intermountain western States. The source apportionment modeling estimates that approximately 40-
80% of the seasonal average MDAS ozone at monitoring locations is due to sources other than
manmade ozone precursor emissions from the U.S. A majority of sites have non-U.S. fractions between
40 and 70 percent. The mean proportion attributable to international and natural sources over all sites
is 59 percent. Despite the differences in the methodologies this is very similar to the mean USB
estimate of 66 percent from the zero out modeling.
2 A-16
-------
9 80 o°0
Ozone (ppb)
• < 20 (0)
20 - 25 (0)
25 - 30 (6)
30-35(10)
35 - 40 (38)
O 40-45(112)
O 45-50 (264)
50 - 55 (488)
55 - 60 (300)
> 60 (75)
Figure 3a. April-October average MDA8 ozone (ppb) at monitoring locations across the U.S. as
estimated by a 2007 CMAQ base simulation.
Ozone (ppb)
< 20 (104)
20 - 25 (740)
25-30(331)
30- 35 (119)
35 - 40 (0)
O 40-45(0)
O 45-50(0)
50 - 55 (0)
55 - 60 (0)
> 60 (0)
Sources: USG:.. E;FI J-II-. -'.I ID. Conic*
Figure 3b. April-October average natural background MDA8 ozone (ppb) at monitoring locations
across the U.S. as estimated by a 2007 CMAQ zero out simulation.
2 A-17
-------
Figure 3c. April-October average North American background MDA8 ozone (ppb) at monitoring
locations across the U.S. as estimated by a 2007 CMAQ zero out simulation.
Ozone (ppb)
< 20 (0)
20 - 25 (0)
25-30 (127)
30 - 35 (842)
35-40(188)
O 40-45(132)
O 45-50 (5)
50 - 55 (0)
55 - 60 (0)
> 60 (0)
Figure 3d. April-October average United States background MDA8 ozone (ppb) at monitoring
locations across the U.S. as estimated by a 2007 CMAQ zero out simulation.
2 A-18
-------
< 40% (1541
40-50% (752)
50-60% (371)
60-70% (12)
70 - 80% (3)
Figure 4a. Ratio of natural background to total April-October average MDA8 ozone at monitoring
locations across the U.S. as estimated based on 2007 CMAQ simulations.
4
OrQO r-rv •
^to k-D ^ o = .1
£.-.-:. °0
X"
<40%(0)
40-50% 121)
50-60% (539)
60-70% (475)
70 - 80% (223)
> 80% (35)
Figure 4b. Ratio of N. American background to total April-October average MDA8 ozone at
monitoring locations across the U.S. as estimated based on 2007 CMAQ simulations.
2 A-19
-------
<40%(0>
40 -50% (2)
50-60% (393)
60-70% (551)
70 - 80% (263)
> 50% (84)
Figure 4c. Ratio of U.S. background to total April-October average MDA8 ozone at monitoring
locations across the U.S. as estimated based on 2007 CMAQ simulations.
M e i i i- o
Sourcts:USGS. ESRl!!T«1«L«MD. S
< 40% (0)
40 - 50% (343)
50 - 60% (483)
60 - 70% <237)
70-80% (178)
> 80% (52)
Figure 4d. Ratio of sources other than U.S. anthropogenic emissions to total April-October average
MDA8 ozone at monitoring locations across the U.S. as estimated by a 2007 CAMx source
apportionment simulation.
2A-20
-------
4. Distributions of background ozone levels
As a first-order understanding, it is valuable to be able to characterize seasonal mean levels of
background ozone. However, it is well established that background levels can vary substantially from
day-to-day. From an implementation perspective, the values of background ozone on possible
exceedance days is a more meaningful distinction. The first draft policy assessment (EPA, 2012)
considered this issue in detail, via summaries of the existing 2006 zero out modeling (Henderson et al.,
2012), and concluded that "results suggest that background concentrations on the days with the highest
total ozone concentrations are not dramatically higher than typical seasonal average background
concentrations." Based on this finding, the 1st draft policy assessment determined that "anthropogenic
sources within the U.S. are largely responsible for 4th highest 8-hour daily maximum ozone
concentrations." This portion of the appendix will consider the entire spectrum of variable background
ozone levels with special emphasis on days in which base model ozone concentrations approach or
exceed the level of the NAAQS.
The 2007 modeling agrees with the finding from the previous 2006-based modeling analyses
that the highest modeled ozone site-days tend to have background ozone levels similar to mid-range
ozone days. Figures 5a-5c show the distribution of April-October IV1DA8 background levels (NB, NAB,
and USB, respectively) from the CMAQ zero out runs. As noted in section 2, zeroing out emissions can
remove the effects of local NOx titration and result in modeled background values that are higher than
the base model ozone. The "box and whisker" plots shown in these figures display four key features of
the distributions:
a. the median concentration (black horizontal line) per bin,
b. the inter-quartile range (blue colored box) which represents the 25th-75th percentile range in
values within the distribution,
c. the "whiskers" (dark gray vertical lines with top and bottom whiskers) which represent the
range of values within 1.5 times the inter-quartile range, and
d. the "outliers" (gray points) which are any values outside the whiskers.
As can be seen in Figure 5a, natural background values do not vary greatly as a function of the
base modeled ozone. Recall that the seasonal average natural background IV1DA8 ozone values were
modeled to range from 15-35 ppb across the U.S. with a median value of 24 ppb. The highest values
were at the high-elevation sites in the western U.S. Based on the distributional analysis, the 75th
percentile values are on the order of 30 ppb. Natural background levels exceeding 40-45 ppb are
considered to be statistical outliers, due to their infrequency. Figure 5b shows the same type of
distributions but for NAB instead of NB. NAB values are generally 6-12 ppb higher than their NB
counterparts, due to the affect of higher global methane values and the influence of anthropogenic
emissions from Asia. It was previously reported (in section 3) that the median seasonal average NAB
IV1DA8 values were 31.5 ppb. Based on the distributions, it can be seen that 75th percentile values are
2A-21
-------
approximately 40 ppb; it is rare for NAB MDAS values to exceed 50-55 ppb. NAB values are constant in
magnitude once the base ozone exceeds 50 ppb indicating that the higher base ozone values are driven
by non-NAB sources (i.e., North American emissions). Finally in Figure 5c, the USB MDAS distributions
by base model MDAS are shown. The results are similar to NAB.
Figure 5d shows the results from the source apportionment modeling of non-U.S. anthropogenic
source contributions to MDAS ozone (i.e., the nine source apportionment categories other than U.S.
anthropogenic emissions and Gulf of Mexico point sources). This non-counterfactual approach is
expected to give a better indication of background levels at low concentrations. At low levels, almost all
of the ozone is determined to be from background origins. The CAMx modeling shows that
contributions from non-U.S. anthropogenic emissions peak when base ozone ranges from 45-55 ppb and
then drop off slightly at higher base MDAS values. The source apportionment modeling of non-U.S.
impacts (similar to USB) indicates slightly lower background levels than the zero out modeling. The 75th
percentile values are generally less than 35 ppb, compared to 40 ppb in the zero out modeling. It is rare
to have background impacts greater than 55ppb. Interestingly, when base model MDAS ozone exceeds
70 ppb, it is rare to have background impacts greater than 45 ppb in the CAMx source apportionment
modeling.
Figures 6a-6d show the equivalent plots as 5a-5d, but use background fractions (background
MDAS/ base MDAS) as the dependent variable instead of the absolute background concentrations.
These plots show the same effect; that is, the proportional relative contribution of background sources
and processes decreases as peak ozone increases. For natural background (Figure 6a), the median
fractions drop from 50% background for values between 45-50 ppb to only 35% background for base
MDAS values between 70-75 ppb. For NAB and USB (Figures 6b and 6c), the median fractions drop from
70% background for values between 45-50 ppb to only 45% background for base MDAS values between
70-75 ppb. The source apportionment modeling (Figure 6d) estimates less of a proportional role of non-
U.S. anthropogenic emissions. In that modeling, the median fractions drop from 65% background for
values between 45-50 ppb to only 35% background for base MDAS values between 70-75 ppb. A key
observation, as noted in the first draft policy assessment document, is that the relative importance of
background decreases on days most likely to violate the NAAQS. An additional policy-relevant finding
from the distributional analyses is that the relative role of background sources would be increased if the
level of the NAAQS were lowered. At 60 ppb, the modeling suggests that the median fractional
contribution from background is 45-55 percent, but there can be cases where background comprises 80-
90 percent of the total ozone.
Many of the cases when background ozone is estimated to contribute in large proportions to
relatively high ozone days may be eligible for consideration as exceptional events, but again, this
modeling is not designed to resolve specific events that occurred in 2007. While there is greater
confidence in the model's ability to predict mean contributions from background sources than from
individual events, it is also useful to briefly consider the upper end of the background ozone
2A-22
-------
distributions. Figure 7 shows the 95th percentile3 USB estimates from the zero out modeling. The 95th
percentile MDAS USB ozone levels range from 35-60 ppb, with the most frequent values residing in the
35-40 and 40-45 ppb bins. The median 95th percentile background USB ozone level is 42.0 ppb. As with
the seasonal mean MDAS USB, the highest levels of high background days (i.e., 95th percentile days) are
observed over the intermountain western U.S. At these locations, 95th percentile USB levels can exceed
50 ppb. Background values at the 95th percentile end of the distribution are 4-12 ppb higher than the
mean background values at the same locations.
3 During the April-October period, there were 214 days of modeling results. Thus, the 95th percentile values
represent approximately the 10th highest days from the distribution.
2A-23
-------
n
a.
0.70-
co
•o
a
re
m
n
<2E 25-30 30-35 25-43 ^0-^5 45-50 50-55 55-73 73-55 65-73 70-75 75-80 80-55 85-30 30-S5 S5-100 >100
Bins of Base Model MDA8 Ozone (ppb)
Figure 5a. Distribution of natural background MDA8 ozone (ppb) at monitoring locations across the
U.S. (Apr-Oct), binned by base modeled site-day MDA8, as estimated by 2007 CMAQ simulations.
•o
I
X
U
re
CD
c
re
a
<25 25-30 30-35 35-40 40-45 45-50 50-55 55-60 60-65 65-70 70-75 75-80 80-85 85-90 90-95 95-100 > 100
Bins of Base Model MDA8 Ozone (ppb)
Figure 5b. Distribution of N. American background MDA8 ozone (ppb) at monitoring locations across
the U.S. (Apr-Oct), binned by base modeled site-day MDA8, as estimated by 2007 CMAQ simulations.
2A-24
-------
a.' :•
Q
•D
i
o
O)
CD
w
o-
<25 25-30 30-35 35-40 40-46
7E-BO 80-85 85-90 90-95 95-100 > 130
Bins of Base Model MDA8 Ozone (ppb)
Figure 5c. Distribution of U.S. background MDA8 ozone (ppb) at monitoring locations across the U.S.
(Apr-Oct), binned by base modeled site-day MDA8, as estimated by 2007 CMAQ simulations.
0BBB
<2S 25-30 30-35 35-40 40-45 45-50 50-55 55-60 60-6.5 65-70 70-75 75-SO SO-85 85-90 90-95 95-100 >100
Bins of Base Model MDA8 Ozone (ppb)
Figure 5d. Distribution of MDA8 ozone contributions from non-U.S. manmade sources (ppb) at
monitoring locations across the U.S. (Apr-Oct), binned by base modeled site-day MDA8, as estimated
by 2007 CAMx simulations.
2A-25
-------
1.25-
<25 26-30 30-35 36-40 40-45 46-50 50-55 55-60 60-66 66-70 70-75 75-80 80-85 86-90 90-95 95-100 > 100
Bins of Base Model P.IDAS Ozone (ppb)
Figure 6a. Distribution of natural background MDA8 ozone fractions at monitoring locations across
the U.S. (Apr-Oct), binned by base modeled site-day MDA8, as estimated by 2007 CMAQ simulations.
1.25-
m
I
<25 25-30 20-3E 3E-4D
:3-:; 65-70 70-75 75-80 80-86 85-90 90-95 95-100
Bins of Base Model MDA8 Ozone (ppb)
Figure 6b. Distribution of N. American background MDA8 ozone fractions at monitoring locations
across the U.S. (Apr-Oct), binned by base modeled site-day MDA8, as estimated by 2007 CMAQ
simulations.
2A-26
-------
1.25-
«1.D3-
II
a
CD
w
a
E D'7" ~
I
§
£ o.2f-
o
—*
u
ffi
iL
o.oo-
«
<25 2S-23 30-3!
45-50 50-55 55-60 60-65 65-70 70-75 75-80 80-85 85-90 90-95 95-100 > 100
Bins of Base Model MDA8 Ozone (ppb)
Figure Be. Distribution of U.S. background MDA8 ozone fractions at monitoring locations across the
U.S. (Apr-Oct), binned by base modeled site-day MDA8, as estimated by 2007 CMAQ simulations.
1.25 -
o.oo-
e25 26-30 30-15 35-40 40-45 4E-E3 EO-EE EE-60 fO-SE 65-70 70-75 75-80 80-85 85-90 90-95 95-100 > 100
Bins of Base Model MDA8 Ozone (ppb)
Figure 6d. Distribution of MDA8 ozone fractions from non-U.S. anthropogenic sources at monitoring
locations across the U.S. (Apr-Oct), binned by base modeled site-day MDA8, as estimated by the 2007
CAMx simulation.
2A-27
-------
Ozone(ppb)
< 20 iOi
20-25(0)
25-30(0)
30-35(0)
35-40(485]
40-45(571)
Q 45-50(116)
50-55(116)
55-60(51
Figure 7. April-October 95th percentile United States background MDA8 ozone (ppb) at monitoring
locations across the U.S. as estimated by a 2007 CMAQ base simulation.
2A-28
-------
5. Contribution of various processes and sources to total background ozone
This section will utilize the supplemental 2007 air quality modeling estimates to determine the
relative importance of specific elements of background ozone. Comparing the differences between the
three zero out scenarios can provide some information about the role of certain sets of emissions.
Figure 8a compares the NAB (zero out North American manmade emissions) and USB (zero out U.S.
manmade emissions) scenarios. The difference between these two runs is the inclusion of
anthropogenic emissions within the Canada and Mexico portions of the modeling domain. These
emissions contribute less than 2 ppb to the seasonal mean IV1DA8 ozone levels over most of the U.S.
There are 70 sites, near an international border, where the modeling estimates Canadian/Mexican
seasonal average impacts of 2-4 ppb. While not shown, the modeled peak single day impacts from
these specific international emissions sources can approach 25 ppb (e.g., San Diego, Buffalo NY). Figure
8b compares the NB (zero out all manmade emissions and reset GEOS-Chem methane values to pre-
industrial levels) to the NAB. The difference between these two runs is the inclusion of global methane
emissions related to recent human activity as well as anthropogenic emissions outside of North America.
These emissions are estimated to contribute 6-15 ppb to seasonal mean ozone levels over the U.S. The
most frequent bin is the 8-10 ppb increase. It is not possible via these runs to parse out what fraction of
this change is due to international emissions as opposed to methane emissions, but the ISA summarized
existing modeling (Zhang et a/., 2011) that suggested that the rise in methane from pre-industrial levels
to present-day levels led to increases in seasonal average ozone levels of 4-5 ppb. The greatest impacts
from these sources occurs over the western U.S., where international emissions would be expected to
have the largest impacts.
Figures 9a-9g show the fractional contribution to total seasonal mean IV1DA8 values of
individual source sectors that were tracked in the CAMx source apportionment modeling. Figure 9a
shows the impact from the regional model boundary conditions. The ozone entering the model domain
via the boundary conditions could have a variety of origins including: a) natural sources of ozone and
ozone precursors (including methane) emanating from outside the domain, b) anthropogenic sources of
ozone precursors (including methane) from international emitters, and c) some fraction of U.S.
emissions (natural and anthropogenic) which are exported and then re-imported into the domain via
synoptic-scale recirculation. Thus, one should not presume that the boundary condition contribution is
directly tied to any particular background definition. At most locations, boundary conditions
contributed 40-60 percent of the total IV1DA8 seasonal mean at sites across the U.S. The highest
proportional impacts from the boundary conditions (the top boundary contributes negligibly) are along
the coastlines and the intermountain West.
Figure 9b shows the source apportionment contribution (to seasonal mean MDA8) from the
most significant sector that was tracked: U.S. anthropogenic ozone precursor emissions. Again the most
common outcome at an individual site was that 40-60% of the seasonal mean ozone values originated
from U.S. anthropogenic emissions. The locations with smaller fractional contributions (e.g., 10-20
percent) from U.S. sources are generally located in places where ozone values are typically low such as
the Pacific Northwest. Figures 9c-9g display the fractional contributions from the other five in-domain
sectors listed in section 2. The impacts from these sectors are briefly summarized below:
2A-29
-------
Biogenic emissions:
o Most frequent bin: 3-5 percent
o Highest site-specific contribution: 10-20 percent
o Region with greatest impacts: Great Plains states where soil NOx emissions are large
Climatologically-average fire emissions:
o Most frequent bin: 0-1 percent
o Highest site-specific contribution: 3-5 percent
o Region with greatest impacts: California, Kansas/Oklahoma region
Within-domain Canadian/Mexican manmade emissions:
o Most frequent bin: 0-1 percent
o Highest site-specific contribution: 10-20 percent
o Region with greatest impacts: Sites along international borders (NY, VT, CA, AZ, TX)
Category 3 marine vessels outside U.S. territorial waters:
o Most frequent bin: 0-1 percent
o Highest site-specific contribution: 10-20 percent
o Region with greatest impacts: Coastal sites (especially southern CA)
Gulf of Mexico point sources4:
o Most frequent bin: 0-1 percent
o Highest site-specific contribution: 1-3 percent
o Region with greatest impacts: Sites in southeast TX and southern LA
4 This sector was also included as part of U.S. anthropogenic source impacts in Figure 9b, but is broken out
separately in Figure 9g.
2A-30
-------
•«
Oregon ^^ i,1.ij^
• •
? E 4 AM^-I> _-ti,,,, ,|^
*,K *••>•
* M•' '
^••-V ::;,,*»•
• .••! •
Mexico
Figure 8a. Difference in April-October average MDA8 ozone (ppb) at monitoring locations across the
U.S. between the USB scenario and the NAB scenario. The difference between these two runs isolates
the impact of within-the-domain anthropogenic emissions from Canada and Mexico.
Ozone(ppb)
<1(0)
1 - 2 (0)
2 - 3 (0)
3-4(0)
4-6(0)
O 6-8(141)
O 8-10(821)
O 10-12(267)
12-15 (65)
Souiees: USGS. ESRIPIWW. iND Souices:
Figure 8b. Difference in April-October average MDA8 ozone (ppb) at monitoring locations across the
U.S. between the NAB scenario and the NB scenario. The difference between these two runs isolates
the impact of the rise in global methane emissions from the pre-industrial and anthropogenic
emissions from outside North America.
2A-31
-------
3-5% IOI
5-10%(0)
O 10-20%(0)
20-40% (212)
40-60% (849)
> 60% (232)
Figure 9a. Percentage of April-October average MDA8 ozone that is apportioned to boundary
conditions as estimated at monitoring locations by a 2007 CAMx simulation.
O 10-20% (59)
20-40% (441)
40 - 60% (791)
>60%(0)
Figure 9b. Percentage of April-October average MDA8 ozone that is apportioned to U.S.
anthropogenic sources as estimated at monitoring locations by a 2007 CAMx simulation.
2A-32
-------
0-1% (42)
1 - 3% (332)
3 - 5% (527)
O 5-10% (362)
O 10-20% (30)
20-40%(0)
40-60%|0)
•' 60% (0)
Figure 9c. Percentage of April-October average MDA8 ozone that is apportioned to purely biogenic
emissions as estimated at monitoring locations by a 2007 CAMx simulation.
0-1% (1.104)
1-3% (181)
3 - 5% (8)
5-10%(0)
10-20%(0)
20-40%(0)
40-60% (0)
>60%IO)
Figure 9d. Percentage of April-October average MDA8 ozone that is apportioned to dimatological fire
emissions as estimated at monitoring locations by a 2007 CAMx simulation.
2A-33
-------
0- 1%(704)
1 - 3% (407)
3-5% (124)
5-10% (49!
O 10-20% (9)
20-40%(0)
40-60%(0)
> 60% (0)
Figure 9e. Percentage of April-October average MDA8 ozone that is apportioned to anthropogenic
emissions from in-domain Canadian and Mexican sources as estimated at monitoring locations by a
2007 CAMx simulation.
0-1% (9741
1-3% (193)
3 -5% (66)
5-10% (56)
O 10-20% (4)
20-40%(0)
40-60% (01
>60%(0)
Figure 9f. Percentage of April-October average MDA8 ozone that is apportioned to Category 3 marine
vessel emissions beyond U.S. territorial waters as estimated at monitoring locations by a 2007 CAMx
simulation.
2A-34
-------
7IXSS
• €9^**
• €
3-5%(0)
5-10%(0)
O 10-20%(0)
20-40%(0)
40-60%(0)
> 60% (0)
Figure 9g. Percentage of April-October average MDA8 ozone that is apportioned to Gulf of Mexico
point sources as estimated at monitoring locations by a 2007 CAMx simulation.
2A-35
-------
6. Estimates of the fractional background contribution to total ozone in 12 specific areas
This penultimate section of the appendix presents estimates of the overall fraction of ozone that
is estimated to result from background sources or processes based on the updated modeling in each of
the 12 urban case study areas in the epidemiology study based analyses in Chapter 7 of the Risk and
Exposure Assessment (REA). Tables la-lc summarize the CAMx-estimated fractional contributions of
sources other than U.S. anthropogenic emissions to total ozone in each of the 12 areas. Table la shows
that the fractional contributions from sources other than anthropogenic emissions within the U.S. to
seasonal mean IV1DA8 levels can range from 43 to 66 percent across these 12 urban areas. These
fractions are consistent with the national ratios summarized in section 3, although the urban fractions of
background tend to be smaller than at rural sites. As shown in section 4, the fractional contributions
from background are smaller on days with high modeled ozone (i.e., days that may exceed the level of
the NAAQS). Table Ib provides the fractional contributions from these non-U.S. sources, only
considering days in which base model IV1DA8 ozone was greater than 60 ppb. As expected, the fractional
background contributions are less and range from 31 to 55 percent. Rather than taking the fractions of
the seasonal means (as in Table la), Table Ic displays the mean and median daily IV1DA8 background
fractions. These metrics may be more appropriate for application to health studies, but as can be seen
the fractional contribution to backgrounds calculated via this approach are very similar to the Table la
calculations. For completeness sake, although EPA expects the source apportionment results to provide
a more realistic estimate of fractional background values, for completeness, we also provide USB
fractions based on zero out modeling for the 12 cities (see Table Id). The results are similar to the
source apportionment findings (compare against Table la), but the zero out technique provides slightly
higher background proportions.
All days, CAMx
Model MDAS seasonal mean
Model MDAS seasonal mean
from emissions other than
U.S. anthropogenic sources
Fractional contribution from
background
ATL
59.3
25.3
0.43
BAL
54.4
25.9
0.48
BOS
43.0
26.2
0.61
CLE
48.9
25.7
0.52
DEN
47.3
31.3
0.66
DET
39.1
23.3
0.60
HOU
48.5
27.0
0.56
LA
51.1
29.1
0.57
NYC
45.4
24.5
0.54
PHI
48.7
24.2
0.50
SAC
46.4
29.7
0.64
STL
49.8
24.3
0.49
Table la. April-October average MDAS ozone, average MDAS ozone from sources other than U.S.
manmade emissions, and the fractional contribution of these background sources in the 12 REA urban
study areas, as estimated by a 2007 CAMx simulation.
2A-36
-------
Only days w/ base
MDA8 > 60 ppb
Model MDAS seasonal mean
Model MDAS seasonal mean
from emissions other than
U.S. anthropogenic sources
Fractional contribution from
background
ATL
74.0
25.4
0.34
BAL
75.3
23.7
0.31
BOS
70.7
24.4
0.35
CLE
72.0
25.4
0.35
DEN
67.5
37.3
0.55
DET
68.9
24.4
0.35
HOU
70.3
28.0
0.40
LA
74.4
31.9
0.43
NYC
74.1
23.5
0.32
PHI
74.0
22.9
0.31
SAC
68.3
32.1
0.47
STL
70.0
25.4
0.36
Table Ib. Average MDAS ozone, average MDAS ozone from sources other than U.S. manmade
emissions, and the fractional contribution of these background sources in the 12 REA areas, as
estimated by a 2007 CAMx simulation using site-days in which base MDAS ozone exceeded 60 ppb.
Mean of daily MDAS
background fractions
Median of daily MDAS
background fractions
ATL
0.46
0.43
BAL
0.53
0.51
BOS
0.68
0.73
CLE
0.58
0.54
DEN
0.69
0.69
DET
0.64
0.66
HOU
0.59
0.59
LA
0.61
0.60
NYC
0.61
0.63
PHI
0.56
0.54
SAC
0.67
0.66
STL
0.52
0.49
Table Ic. Fractional contribution of non-U.S. manmade emissions sources in the 12 REA urban study
areas, as estimated by a 2007 CAMx simulation using means and medians of daily MDAS fractions.
All days, CMAQ
Model MDAS seasonal mean
Model MDAS seasonal mean
from USB emissions
Fractional contribution from
background
ATL
58.6
30.0
0.51
BAL
55.6
29.9
0.54
BOS
45.2
28.5
0.63
CLE
51.8
31.6
0.61
DEN
57.1
42.2
0.74
DET
43.5
31.7
0.73
HOU
49.4
33.0
0.67
LA
54.8
33.3
0.61
NYC
47.7
29.1
0.61
PHI
50.5
29.4
0.58
SAC
51.9
34.4
0.66
STL
52.6
32.0
0.61
Table Id. April-October average MDAS ozone, average MDAS ozone from USB, and the fractional
contribution of these background sources in the 12 REA urban study areas, as estimated by two
separate 2007 CMAQ simulations.
2A-37
-------
7. Background ozone and W126
As discussed in section 5 of the second draft policy assessment, EPA is considering the adequacy
of the current secondary standard to protect against welfare effects. One metric that has been
considered previously as a potential cumulative seasonal index is the W126 metric. The W126 index is a
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 0 to 1 with
increasing concentration (Lefohn et al, 1988). The weights are defined such that values of 0.060 ppm
get a weight of ~0.3; 0.070 ppm values get a weight of ~0.6; and 0.085 ppm values get a weight of ~0.9.
The remainder of this section uses the 2007 zero out modeling to conduct a limited assessment of the
role of background ozone on W126 levels over the U.S.
The analysis of background influence on W126 is not as detailed as the analyses related to
seasonal mean IV1DA8 ozone. Instead of considering impacts at every monitoring location, EPA assessed
NB, NAB, and USB influences at four sample locations: Atlanta GA, Denver CO, Farmington NM, and
Riverside CA. Each of these four locations had relatively high observed values of W126 in 2010-2012.
Atlanta is an urban area in the Eastern U.S. with high primary ozone design values but relatively low
levels of seasonal background ozone. Riverside and Denver also have high primary ozone design values
but are in the Western U.S. where background ozone levels are generally higher. Farmington NM was
chosen as a site that has relatively lower primary ozone design values along with its relatively high W126
levels. The varying characteristics of each of these locations perhaps allows broader national
extrapolation of the 4-site results.
In previous EPA reviews of the O3 NAAQS, the influence of background ozone was estimated
according to a counterfactual (i.e., how much ozone would exist in the absence of certain sets of
emissions). In the current review, EPA is supplementing the counterfactual assessment with analyses
that estimate the fraction of the existing ozone that is due to background sources. This has important
ramifications for assessing the influence of background on W126 concentrations, because of the non-
linear weighting function used in the metric which emphasizes high ozone hours (e.g., periods in which
ozone is greater than ~60 ppb). As an example, consider a sample site in the intermountain western
U.S. region with very high modeled estimates of U.S background (e.g., seasonal mean USB of 45 ppb
with some days as high as 65 ppb). Even at this high background location, the calculated annual W126
values in the USB scenario are quite low, on the order of 3 ppm-hrs. Most sites in the domain where
background levels are lower than the location cited above will have even smaller background W126
estimates, on the order of 1 ppm-hrs, which is consistent with values mentioned in past reviews (USEPA,
2007). Using the counterfactual scenarios, background ozone has a relatively small impact on W126
levels across the U.S.
However, because of the non-linear weighting function used in the W126 calculation, the sum of
the W126 from the USB scenario and the W126 resulting from US anthropogenic sources will not equal
the total W126. In most cases, the sum of those two components will be substantially less than total
W126. As a result, EPA believes it is more informative to estimate the fractional contribution of
background ozone to W126 levels. The 5-step methodology for assessing the fractional influence of
2A-38
-------
background ozone to annual W126 levels in the four locations is described below. The fractional
influence methodology essentially places higher weights on background fractions on days that are going
to contribute most substantially to the yearly W126 value.
• Step la: Calculate the IV1DA8 ozone values from the base and the three zero out
modeling scenarios at each grid cell containing a site in an area.
• Step Ib: Calculate the W126 daily index for the base model scenario.
• Step 2: For each site, find the three months with highest summed W126 daily indices.
• Step 3: Normalize the daily IV1DA8 values in the base, NB, NAB, and USB scenarios by the
corresponding W126 daily index from the base scenario.
• Step 4: Calculate the average W126-weighted IV1DA8 values over the three month
period for each of the four scenarios (base, NB, NAB, USB).
• Step 5: Calculate the NB/Base, NAB/Base, and USB/Base ratios based on step 4 outputs.
These values represent an estimate of the fractional influence of background ozone on
modeled W126 levels.
Figure 7a shows the estimated fractional influence of the three background definitions on W126
levels in Atlanta, Denver, Farmington, and Riverside. Based on this limited assessment, natural
background sources are estimated to contribute 29-50% of the total modeled W126 with the highest
relative influence in the intermountain western U.S. (e.g., Farmington NM) and the lowest relative
influence in the eastern U.S. (e.g., Atlanta). U.S. background is estimated to contribute 37-65% of the
total modeled W126. Figure 7b compares the relative influence of background on W126 versus seasonal
mean IV1DA8 ozone. The proportional impacts of background are slightly less for the W126 metric than
for seasonal mean IV1DA8 (discussed in section 2.4.2), because of the weighting function that places
more emphasis on higher ozone days when background fractions are generally lower.
There are several caveats associated with this analysis. First, only the zero out modeling was
used to assess the fractional influence of background sources on W126. The source apportionment
approach estimated slightly smaller relative contributions for seasonal mean IV1DA8 levels, so from that
perspective the zero out estimates could represent the high end of background influence on W126.
Additionally, the methodology used for this analysis relies on daily IV1DA8 values as a surrogate (the data
were readily available) for the 8a-8p time period relevant to the W126 metric. The key conclusion from
this cursory analysis is that background ozone may comprise a non-negligible portion of current W126
levels across the U.S. This fractional influence is greatest in the intermountain western U.S. and are
slightly smaller than the seasonal mean IV1DA8 metric. In the counterfactual cases, when non
background sources are completely removed, the remaining W126 levels are low (< 3 ppm-hrs).
2A-39
-------
100%
Farmington
Denver
Riverside
Atlanta
Figure 7a. Fractional contribution of background sources to W126 levels in four sample locations.
Model estimates based on 2007 CMAQzero out modeling.
100%
MDA8
IW126
Farmington
Denver
Riverside
Atlanta
Figure 7b. Fractional contribution of U.S. background to seasonal mean MDA8 ozone and W126 levels
in four sample locations. Model estimates based on 2007 CMAQ zero out modeling.
2A-40
-------
8. Summary
The precise definition of background ozone can vary depending upon context, but it generally
refers to ozone that is formed by sources or processes that cannot be influenced by local control
measures. Background ozone can originate from natural sources of ozone and ozone precursors, as well
as from upwind manmade emissions of ozone precursors. In order to help further characterize
background ozone levels over the U.S., EPA has completed additional air quality modeling analyses
subsequent to the lst-draft policy assessment. As shown above, the results are largely consistent with
previous determinations about the magnitude of background ozone contributions across the U.S.
For a variety of reasons, it is challenging to present a comprehensive summary of all the
components and implications of background ozone. In many forums the term "background" is used
generically and the lack of specificity can lead to confusion as to what sources are being considered.
Additionally, it is well established that the impacts of background sources can vary greatly over space
and time which makes it difficult to present a simple summary of background ozone levels. Further,
background ozone can be generated by a variety of processes, each of which can lead to differential
patterns in space and time, and which often have different regulatory ramifications. Finally, background
ozone is difficult to measure and thus, typically requires air quality modeling which has inherent
uncertainties and potential errors and biases. Even with all of these complexities in mind, EPA believes
the following concise and step-wise summary of background ozone is appropriate as based on previous
modeling exercises and the more recent EPA analyses summarized herein.
• The most fundamental definition of background is "natural background" (NB). NB ozone is that
which is produced by processes other than manmade emissions. Examples of sources of natural
background include: stratospheric ozone intrusions, wildfire emissions, and biogenic emissions
from vegetation and soils. To date, NB ozone has been estimated to be that ozone that would
exist in the absence of anthropogenic ozone precursor emissions worldwide. Modeling analyses
have shown that NB levels can vary in time and space. As shown in Section 3, April-October
average NB levels range from approximately 15-35 ppb with the highest values in the spring and
at higher-elevation sites.
• More expansive definitions of background include North American background (NAB) and U.S.
background (USB). These definitions represent the ozone that originates from sources and
processes other than North American or U.S. anthropogenic sources. Sources of NAB and USB
include all the same sources of natural background, plus manmade ozone precursors emitted
outside the North America or the U.S. Modeling analyses have shown that NAB and USB
background levels can vary in time and space. As discussed in Section 3, seasonal mean NAB
and USB background levels range from approximately 25-45 ppb with the highest values in the
spring and at higher-elevation sites. USB levels are slightly higher than NAB, usually by less than
2 ppb.
2A-41
-------
• Estimates of seasonal mean background ozone levels are valuable in terms of a first-order
characterization, however because levels can vary significantly from day-to-day, it is also
instructive to consider the distribution of daily model estimates of background ozone over a
season. Typically, model background is slightly higher in the April-June period than in the later
portion of the ozone season (July-October) (EPA, 2012). More importantly, the modeling shows
that the days with highest ozone levels, on average, have similar background levels to days with
lower values. As a result, the proportion of total ozone that has background origins is smaller
on high ozone days (e.g., days > 70 ppb) than the more common lower ozone days that drive
seasonal means. Section 4 provides information about the distribution of background ozone
fractions. Based on the source apportionment modeling, it is shown that U.S. anthropogenic
emissions typically comprise the majority of the total ozone on site-days with base modeled
ozone IV1DA8 values greater than 60 ppb.
• While it is important to recognize that most high ozone days (i.e., potential exceedance days)
are estimated to be driven predominantly by non-background emissions, the recent EPA
modeling also shows times and locations in which background contributions are estimated to
approach 60-80 ppb. As described in Sections 4 and 6 of this document, these occurrences are
relatively infrequent. While the modeling was not expressly developed to capture these types
of events, ambient observations have also shown relatively rare events where background
ozone sources (wildfires, stratospheric intrusions) have overwhelmingly contributed to an
ozone exceedance. From a policy perspective, these background events must be viewed in the
context of their relative infrequency and the existing mechanisms within the Clean Air Act (e.g.,
exceptional event policy, 179B international determinations) that help ensure States are not
required to control for events that are inherently outside their ability to influence. While
background ozone levels can approach and periodically exceed the NAAQS at some locations,
these conditions are not a constraining factor in the selection of a NAAQS. The Clean Air Act
requires the NAAQS to be set at a level requisite to protect public health and welfare. Case law
makes it clear that attainability and technical feasibility are not relevant considerations. In
previous reviews, EPA assessed the proximity of potential levels to peak background levels as a
secondary consideration between levels where health and welfare was protected.
• Section 5 shows that the contributions to background are multi-dimensional. Daily peak 8-hour
ozone values over the U.S. are a function of local and regional anthropogenic emissions,
anthropogenic emissions from outside the U.S. (including shipping emissions), natural and
anthropogenic methane emissions, wildfire emissions, and purely natural sources. While local
and regional controls are still considered to be the most effective at reducing local ozone levels,
any measures to reduce the international contributions or methane-induced background will
also be valuable.
• In previous ozone NAAQS reviews, EPA estimated risk from exposure only to ozone
concentrations above background. In the first drafts of the REA and PA for the current ozone
2A-42
-------
review, EPA estimated risk from exposure to total measured ozone concentrations, which
include those concentrations from background sources. EPA will continue to provide estimates
of risk from exposure to total ozone, consistent with CASAC advice, in the second draft policy
assessment. The recent EPA modeling was completed to assist in determining, in a limited
sense, the risk attributable to background ozone. The fractional values of background
contributions in the 12 REA study areas (43-66 percent) could be used as first order
approximations of the risk due to ozone background.
9. References
Arunachalam, S. (2009). Peer Review of Source Apportionment Tools in CAMx and CMAQ. Prepared
under EPA Contract #EP-D-07-102, Institute of the Environment, University of North Carolina,
Chapel Hill, NC, 42pp,
http://www.epa.gov/scram001/reports/SourceApportionmentPeerReview.pdf.
Emery, C; Jung, J; Downey, N; Johnson, J; Jimenez, M; Yarwood, G; Morris, R. (2012). Regional and global
modeling estimates of policy relevant background ozone over the United States. Atmospheric
Environment, 47: 206-217. http://dx.doi.Org/10.1016/i.atmosenv.2011.ll.012.
Environ (2011). User's Guide: Comprehensive Air Quality Model with Extensions, Version 5.40; Novato
CA, 306pp. http://www.camx.com/files/camxusersguide v5-40.aspx.
Fiore, A; Jacob, DJ; Liu, H; Yantosca, RM; Fairlie, TD; Li, Q. (2003). Variability in surface ozone background
over the United States: Implications for air quality policy. J Geophys Res, 108: 4787.
http://dx.doi.org/10.1029/2003JD003855.
Henderson, BH; Possiel N; Akhtar F; Simon H. (2012). Memo to the Ozone NAAQS Review Docket EPA-
HQ-OAR-2012-0699: Regional and Seasonal Analysis of North American Background Ozone
Estimates from Two Studies.
http://www.epa.gOV/ttn/naaqs/standards/ozone/s o3 2008 td.html.
Lefohn, A. S.; Laurence, J. A.; Kohut, R. J. (1988). A comparison of indices that describe the relationship
between exposure to ozone and reduction in the yield of agricultural crops. Atmos. Environ. 22:
1229-1240.
Simon, H., Baker, K.P., Phillips, S. (2012) Compilation and interpretation of photochemical model
performance statistics published between 2006 and 2012. Atmospheric Environment, 61, 124-
139.
U.S. Environmental Protection Agency (2005). Technical Support Document for the Final Clean Air
Interstate Rule Air Quality Modeling. Office of Air Quality Planning and Standards, Research
Triangle Park, NC, 285pp. http://www.epa.gov/cair/technical.html.
U.S. Environmental Protection Agency. (2007). Review of the National Ambient Air Quality Standards for
Ozone: Policy Assessment of Scientific and Technical Information - OAQPS Staff Paper, U.S.
Environmental Protection Agency, Research Triangle Park, NC. EPA 452/R-07-007.
2A-43
-------
U.S. Environmental Protection Agency (2011). Air Quality Modeling Final Rule Technical Support
Document. Office of Air Quality Planning and Standards, Research Triangle Park, NC, 363pp.
http://www.epa.gov/airtransport/CSAPR/techinfo.html.
U.S. Environmental Protection Agency (2012). Policy Assessment for the Review of the Ozone National
Ambient Air Quality Standards: First External Review Draft. EPA-452/P-12-002, 297pp.
U.S. Environmental Protection Agency. (2013). Integrated Science Assessment for Ozone and Related
Photochemical Oxidants, U.S. Environmental Protection Agency, Research Triangle Park, NC.
EPA/600/R-10/076.
U.S. Environmental Protection Agency. (2014). Health Risk and Exposure Assessment for Ozone, Second
External Review Draft, U.S. Environmental Protection Agency, Research Triangle Park, NC. EPA
xxx/P-xx-xxx.
Zhang, L; Jacob, DJ; Downey, NV; Wood, DA; Blewitt, D; Carouge, CC; Van donkelaar, A; Jones, DBA;
Murray, LT; Wang, Y. (2011). Improved estimate of the policy-relevant background ozone in the
United States using the GEOS-Chem global model with 1/2 2/3 horizontal resolution over North
America. Atmospheric Environment 45: 6769-6776.
http://dx.doi.0rg/10.1016/i.atmosenv.2011.07.054.
2A-44
-------
APPENDIX 2B
MONITORING DATA ANALYSIS OF RELATIONSHIPS
BETWEEN CURRENT STANDARD AND W126 METRIC
Presented here are monitoring data analyses evaluating relationship between ozone (Os)
concentrations in the averaging time and form of the current secondary standard (3-year average
of the annual 4th highest daily maximum 8-hour concentrations, in parts per billion), and a three-
year W126 metric (3-year average of the annual maximum 3-month sum of weighted daytime
concentrations, in parts per million-hours). We also consider the responsiveness of these two
metrics to historical changes in air quality related to ozone precursor emissions.
For this analysis, we chose to examine monitoring data from a base period (2001-2003)
as well as a recent period (2009-2011). The base period was chosen to represent air quality
conditions before the implementation of the 1997 national ambient air quality standard
(NAAQS) for Os (0.08 ppm). In 2004, EPA designated 113 areas as nonattainment for the 1997
standard, which required many areas to begin precursor emissions control programs for the first
time. At about the same time, EPA began implementation of the NOx Budget Trading Program
under the NOx State Implementation Plan, also known as the "NOx SIP Call1," which required
summertime reductions in NOx emissions from power plants and other large sources throughout
the Eastern U.S. These programs were successful in reducing peak O3 concentrations, especially
in the Eastern U.S., and as a result only 8 of the original 113 nonattainment areas were still
violating the 1997 O3 NAAQS during the 2009-2011 period.
Hourly O3 concentration data were retrieved from EPA's Air Quality System (AQS)
database2 for both periods, and used to calculate design values for the current standard as well as
3-year average W126 values for both periods. The procedures for calculating design values for
the current standard from hourly Os concentration data are described in 40 CFR Part 50,
Appendix P, and the procedures for calculating the 3-year average W126 values are described in
section 4.3.1. of the 2nd draft Welfare Risk and Exposure Assessment (WREA). There were 838
monitoring sites with sufficient data to calculate these values for both periods. In order to
identify regional patterns in the relationships, these sites were grouped into the nine NOAA
1 http://www.epa.gov/airmarkets/progsregs/nox/sip.html
2 EPA's Air Quality System (AQS) database is a national repository for many types of air quality and
related monitoring data. AQS contains monitoring data for the six criteria pollutants dating back to the 1970's, as
well as more recent additions such as PM2.5 speciation, air toxics, and meteorology data. At present, AQS receives
hourly O3 monitoring data collected from nearly 1,400 monitors operated by over 100 state, local, and tribal air
quality monitoring agencies.
2B-1
-------
climate regions (Karl and Koss, 1984) used in the WREA. Figure 2B-1 presents a map of these
regions, which are color-coded to match the scatter plots in the subsequent figures.
Figures 2B-2a, 2B-2b, 2B-3a and 2B-3b show scatter plots of the design values for the
current standard (x-axis) versus 3-year average W126 values (y-axis) for the base period and
recent period, respectively. Most monitors in the U.S. both exceeded the current standard of 75
ppb and a three-year average W126 value of 15 ppm-hrs during the base period. During the
recent period, both the design values and 3-year average W126 values were much lower, and
there also appears to be less scatter between the two metrics. In both periods, the highest design
values and W126 values occurred in the West region which includes California. Finally, it is
worth noting that monitors in the Southwest and West regions tend to have higher W126 values
relative to their design values than in other regions.
Figure 2B-4 shows a scatter plot of the design values for the current standard for the base
period (x-axis) versus for the recent period (y-axis), while Figure 2B-5 shows this same
relationship based on the 3-year average W126 values. The relationship between the two periods
appears to be fairly linear for both metrics, indicating that larger decreases in these metrics
tended to occur at monitors with higher base values. Figures 2B-6 and 2B-7 show design values
for the current standard and 3-year average W126 values, respectively, compared to the unit
changes in those values between the base period and recent period. Figures 2B-6 and 2B-7 show
the difference between each point and the one-to-one lines in Figures 2B-4 and 2B-5,
respectively. In particular, these figures highlight that there were some monitors where design
values for the current standard and/or W126 values increased. However, those monitors also
tended to have lower base values, and were mostly located outside of areas subject to emissions
controls under the 1997 standard.
Finally, Figure 2B-8 compares the unit change in design values (in ppb; x-axis) to the
unit change in 3-year average W126 values (in ppm-hrs; y-axis). This figure shows that in most
locations, the current standard metric and the W126 metric exhibit similar responses to changes
in precursor emissions. In particular, the NOx SIP Call, which was implemented in the states
east of the Mississippi River, was effective at reducing both design values and W126 values at
nearly all monitors in the Eastern U.S. The relationship was much more variable in the
remaining regions, where emissions control programs were mostly local and limited to areas
which were violating the NAAQS.
Based on this analysis of ambient monitoring data, we can make the following general
conclusions about the relationship between the design value metric for the current Os standard
and the 3-year average W126 metric:
1. There is a fairly strong, positive degree of correlation between the two metrics.
2B-2
-------
2. Monitors in the West and Southwest regions tend to have higher W126 values relative to
their design values than in other regions.
3. Reducing precursor emissions, especially NOx, is an effective strategy for lowering both
design values and W126 values. In particular, regional control programs such as the NOx
SIP call are effective at reducing both metrics over a broad area.
In addition, Figure 2B-9 examines the number of counties with 8-hour design values
meeting the current standard and 3-year average W126 index values greater than 15 ppm-hrs.
Most of these counties were located in the Southwest region of the country. There were no
counties in any of the studied 3-year periods that had design values less than or equal to 65 ppb
and 3-year average W126 index values greater than 15 ppm-hrs.
Central
EastNorthCentral
NorthEast
Northwest
South
SouthEast
Southwest
West
WestNorthCentral
Figure 2B-l.Map of the 9 NOAA climate regions (Karl and Koss, 1984), color coded to
match the subsequent scatter plots.
2B-3
-------
CD
r--
CD
CD
CD
rn
CD
CN
• Central
* EastNorthCentral
NorthEast
* NorthWest
; South
SouthEast
• SouthWest
* West
* WestNorthCentral
* #
40
I
60
1
80
100
120
Figure 2B-2a. Design values for the current Os standard in ppb (x-axis) versus 3-year
average W126 values in ppm-hrs (y-axis) based on ambient monitoring data
for 2001-2003.
2B-4
-------
CM
T
o
CN
Central
* EastNorthCentral
• NorthEast
• NorthWest
South
SouthEast
SouthWest
• West
* WestNoithCentral
••
•MA
t •
1
•
: * *
*
* **
•«
*•
* •
' • * •, '
•
GO
65
I
70
i
75
Figure 2B-2b. Design values for the current Os standard in ppb (x-axis) versus 3-year
average W126 values in ppm-hrs (y-axis) based on ambient monitoring data
for 2001-2003 with a focus on monitors with 2001-2003 design values below 75
ppb.
2B-5
-------
CD
r--
CD
CD
CD
rn
CD
CN
• Central
* EastNorthCentral
NorthEast
* NorthWest
; South
SouthEast
SouthWest
* West
* WestNorthCentral
.*
* *
*
.*.
****
»» *•**"
40
I
60
1
80
100
120
Figure 2B-3a. Design values for the current Os standard in ppb (x-axis) versus 3-year
average W126 values in ppm-hrs (y-axis) based on ambient monitoring data
for 2009-2011.
2B-6
-------
CM
o
CN
• Central
* EastNorthCentral
• NorthEast
* NorthWest
• South
SouthEast
SouthWest
* West
* WestNorthCentral
*•
•
• • ••
A
• * t •*!.*.;
ft 4 rt
* ( * v J i
ft A s u
— 1 ^ 4
* * t * '
* _
•
• * *
. * *
• •
* !••!••
* * 1
• • *<
• • ': .*.*
* * S 1 **
*~i ff **li**l<
S .* J* .,
J* •*
f * *
_ • *
• * *
, *
*
•
*
* » *
*•
4
• *•*
*
* ~ «
• 1
"**•'
• |* j*
•**. I1!!:**!1
'.§•!* *.*;
i ^ • * * #
I*!1** . lf
!•.*••
• *
i * *
• * * •
*
•
GO
65
70
75
Figure 2B-3b. Design values for the current Os standard in ppb (x-axis) versus 3-year
average W126 values in ppm-hrs (y-axis) based on ambient monitoring data
for 2009-2011 with a focus on monitors with 2009-2011 design values below 75
ppb.
2B-7
-------
CD
CN
CD
CD
CD
CO
CD
CD
• Central
• EastNorthCentral
NorthEast
* NorthWest
South
SouthEast
SouthWest
* West
* WestNorthCentral
t *
/\*./ •* *
<*> **
*<
40
QO
I
80
100
120
Figure 2B-4.Design values for the current Os standard in ppb based on ambient
monitoring data for 2001-2003 (x-axis) versus 2009-2011 (y-axis).
2B-8
-------
CD
r--
CD
CD
CD
rn
CD
CN
> Central
* EastNorthCentral
NorthEast
* NorthWest
• South
SouthEast
SouthWest
* West
* WestNorthCentral
I
10
20
I
30
40
I
50
60
I
70
Figure 2B-5. Three-year average W126 values in ppm-hrs based on ambient monitoring
data for 2001-2003 (x-axis) versus 2009-2011 (y-axis).
2B-9
-------
CD
<:-•>
CD
OD
•$BS&3 V •
Central
EastNorthCentral
NorthEast
NorthWest
South
SouthEast
SouthWest
West
WestNorthCentral
«.
* *
40
[^
60
I
80
1 00
1 20
Figure 2B-6. Design values for the current Os standard in ppb based on ambient
monitoring data for 2001-2003 (x-axis) versus unit (ppb) change in design
values from 2001-2003 to 2009-2011 (y-axis).
2B-10
-------
CD
OD
••
:•». •
*
*
Central
* EastNorthCentral
• NorthEast
* NorthWest
South
SouthEast
SouthWest
* West
* WestNorthCeritral
I
10
20
30
40
50
60
nr
70
Figure 2B-7. Three-year average W126 values in ppm-hrs based on ambient monitoring
data for 2001-2003 (x-axis) versus unit (ppm-hr) change in 3-year average
W126 values from 2001-2003 to 2009-2011 (y-axis).
2B-11
-------
CD
OD
•'•*>M3M*8"»^
-------
75-
£GO-
<§45-
12-
0)
_Q
-51
30-
0-
•-.
-
-
-
i
z_
-
-
r
z
-
-
"
-•
z
z
• 1
[N
"i
Regions
lohioValley
Rockies
| Southeast
1 South
I Southwest
| West
Years
Regions
• Rockies
Southwes'
-,
3
4
5
6
Figure 2B-9. Number of counties where the 8-hour design value is meeting the current standard and 3-year average W126
index value is greater than 15 ppm-hrs (left), and number of counties where the 8-hour design value is less than
or equal to 70 ppb and 3-year average W126 index value is greater thanlS ppm-hrs (right)3.
7 REFERENCES
8 Karl, T.R. and Koss, W.J., 1984: "Regional and National Monthly, Seasonal, and Annual Temperature Weighted by Area, 1895-1983." Historical Climatology
9 Series 4-3, National Climatic Data Center, Asheville, NC, 38 pp.
3 No counties in any of the studied 3-year periods were at or below a 3-year average of 4th highest daily maximum 8-hour averages of 65 ppb and also
above a 3-year W126 index value of 15 ppm-hrs.
January 2014
2B-13
Draft - Do Not Quote or Cite
-------
APPENDIX 2C
INTER-ANNUAL VARIABILITY IN W126 INDEX VALUES:
COMPARING ANNUAL AND 3-YEAR AVERAGE METRICS (2008-2010)
2C.1 OVERVIEW
This appendix describes an analysis comparing values for a single-year or annual W126
metric to a W126 metric averaged over three consecutive years. The purpose of this analysis is
to compare values based on a 3-year average of annual W126 indices to values based on a single
annual W126 index. The deviations of the annual W126 index values in 2008, 2009, and 2010
from the 2008-2010 average W126 index values are presented.
2C.2 GENERAL DATA PROCESSING
The air quality data for this analysis originated from EPA's Air Quality System (AQS)
data base, the official repository of ambient air measurements. The data used in this analysis
consisted of W126 index values calculated from hourly ozone concentrations measured at 1082
ozone monitors nationwide. Ozone monitors must have submitted data to AQS for at least 75%
days in their required ozone monitoring season in 2008, 2009, and 2010 to be included in the
analysis.
2C.3 RESULTS & CONCLUSION
The figure below shows a scatter plot of the deviations in the annual W126 index from
the 3-year average by monitor. The solid curves represent the average deviation in a moving
window along the x-axis for each year. From this figure, it is apparent that the highest annual
W126 index value occurred in 2008 for most monitoring locations, the lowest annual W126
index value occurred in 2009 for most monitoring locations, and the 2010 W126 index value was
generally somewhere in between. It is also apparent that the inter-annual variability in the W126
index increases along with the 3-year average. For monitors with 3-year average W126 values
near 15 ppm-hrs, the average deviation was +3.5 ppm-hrs in 2008 and -3.8 ppm-hrs in 2009.
This represents a 1-year swing of-7.3 ppm-hrs.
The model-based air quality adjustments in the 2nd draft of the Os Welfare REA show
that reducing NOx emissions is effective for reducing 3-year average W126 levels. In Appendix
2B, the analyses based on ambient monitoring data also show that large-scale reductions in NOx
emissions are associated with lower W126 levels. Finally, the data analysis presented in this
appendix shows that the inter-annual variability in the annual W126 index tends to decrease with
decreasing W126 levels. Thus, it is expected that reductions in NOx emissions will not only
2C-1
-------
result in lower 3-year average W126 levels, but also result in less inter-annual variability
associated with annual W126 levels.
The W126 index is based on a logistic weighting function that increases the weights
assigned to hourly ozone concentrations very rapidly. Hourly ozone concentrations of 50 parts
per billion are given a weight of about 10% while concentrations of 80 parts per billion are given
a weight of nearly 90%. The annual W126 index is calculated as a 3-month sum of weighted
ozone concentrations during daylight hours, which amounts to a sum of roughly 1100 weighted
hourly concentrations. Thus, even a modest change in the average daily ozone level may have a
significant impact upon the annual W126 index. Since ozone formation is heavily influenced by
meteorology, the inter-annual variability in meteorological conditions tends to cause a large
inter-annual variability in the W126 index.
In conclusion, this evaluation indicates the extent to which a form for the secondary
ozone standard that averages the annual W126 index values over three consecutive years might
be expected to account for the annual variability in this index since the 3-year period would be
expected to include year(s) below as well as above the 3-year average.
2C-2
-------
Annual W126 Index vs. 3-Year Average W126 Index
«J
I
Q
o 2008
n 2009
^ 2010
5 10 20
3-Year Average W126 Index, 2008-2010
50
Figure 2C-1. Deviation of the annual W126 index values in 2008, 2009, and 2010 (y-axis)
from the 3-year average W126 index value (x-axis).
2C-3
-------
APPENDIX 3A
RECENT STUDIES OF RESPIRATORY-RELATED
EMERGENCY DEPARTMENT VISITS AND HOSPITAL
ADMISSIONS
Hospital Admissions for All Respiratory Causes
The APHENA study (APHENA is for Air Pollution and Health: A European and North
American Approach) analyzed air pollution and health outcome data from existing Canadian,
European, and U.S. multi-city studies and examined the influence of varying model specification
to control for season and weather (Katsouyanni et al., 2009). The U.S.-based portion of the
APHENA study utilized the National Morbidity, Mortality, and Air Pollution Study (NMMAPS)
cohort which, for the Katsouyanni et al. (2009) analysis, comprised respiratory hospital
admissions among individuals 65 years of age and older from 14 US cities with Os data from
1985-1994 (7 cities had summer only Os data). For the year round analysis, Katsouyanni et al.
(2009) reported consistently positive, and statistically significant in models with 8 degrees of
freedom per year (U.S. EPA, 2013, section 6.2.7.2), associations between 1-hour Os
concentrations and respiratory hospital admissions across the datasets from the U.S., Canada, and
Europe (U.S. EPA 2013, Figure 6-15).l In co-pollutant models adjusting for PMio, Os effect
estimates remained positive, though effect estimates were somewhat attenuated in the U.S. and
European datasets, possibly due to the PM sampling schedule (U.S. EPA 2013, Figure 6-15).
Effect estimates for the warm season were larger than for the year-round analysis in the
Canadian dataset, but generally similar in magnitude to the year-round analysis in the U.S. and
European datasets.
Several additional multicity studies examined respiratory disease hospital admissions in
Canada and Europe. Cakmak et al. (2006) reported a statistically significant increase in
respiratory hospital admissions in 10 Canadian cities (4.4% increase per 20 ppb increase in 24-
hour average Os, 95% CI: 2.2, 6.5%). In analyses of potential effect modifiers of the Os-
respiratory hospital admission relationship, individuals with an education level less than the 9th
grade were found to be at greater risk. Dales et al. (2006) reported a 5.4% (95% CI: 2.9, 8.0%)
increase in neonatal respiratory hospital admissions for a 20 ppb increase in 24-hour average Os
lrrhe study by Katsouyanni et al. (2009) evaluated different statistical models. Although the investigators
did not identify the model they deemed to be the most appropriate for comparing the results across study locations,
they did specify that "overall effect estimates (i.e., estimates pooled over several cities) tended to stabilize at high
degrees of freedom" (Katsouyanni et al., 2009). In discussing of the results of this study, the ISA focused on models
with 8 degrees of freedom per year (US EPA, 2012a, section 6.2.7.2).
3A-1
-------
concentrations in 11 Canadian cities from 1986 to 2000. In contrast, Biggeri et al. (2005) did not
detect an association between short-term Os exposure and respiratory hospital admissions in four
Italian cities from 1990 to 1999.
In addition to the large multi-city studies discussed above, several smaller-scale studies
have also reported associations with total respiratory hospital admissions. Specifically, Lin et al.
(2008) reported a positive association between Os and pediatric (i.e., <18 years) respiratory
admissions in an analysis of 11 geographic regions in New York state from 1991 to 2001, though
results were not presented quantitatively. In co-pollutant models with PMio, the authors reported
that region-specific Os associations with respiratory hospital admissions remained relatively
robust.
Cause-Specific Hospital Admissions
With regard to cause-specific respiratory outcomes, the limited evidence available in the
last review indicated that the strongest findings were for ambient Os associated asthma and
chronic obstructive pulmonary disease (COPD) respiratory hospital admissions (U.S. EPA 2013,
6.2.7.2). Since the last review, a few additional studies have investigated cause-specific
respiratory admissions (i.e., COPD, asthma, pneumonia) in relation to Os exposure (Medina-
Ramon et al, 2006; Yang et al., 2005; Zanobetti and Schwartz, 2006; Silverman and Ito, 2010).
Medina-Ramon et al. (2006) examined the association between short-term ambient Os
concentrations and Medicare hospital admissions for COPD among individuals > 65 years of age
for COPD in 35 cities in the U.S. for the years 1986-1999. The authors reported an increase in
COPD admissions for lag 0-1 day in the warm season for a 30 ppb increase in 8-h max Os
concentrations. The authors found no evidence for such associations in cool season or in year
round analyses. In a co-pollutant model with PMio, the association between Os and COPD
hospital admissions remained robust. In Vancouver from 1994-1998, a location with low ambient
Os concentrations (U.S. EPA, 2013, Table 6-26), Yang et al. (2005) reported a statistically non-
significant increase in COPD admissions per 20 ppb increase in 24-hour average Os
concentrations. In two-pollutant models with every-day data for NO2, SO2, CO, and PMio, Os
risk estimates remained robust, though not statistically significant (U.S. EPA, 2013, Figure 6-20;
Table 6-29). In addition, Wong et al. (2009) reported increased Os-associated COPD admissions
during periods of increased influenza activity in Hong Kong.
The ISA assessed a study that evaluated asthma-related hospital admissions in New York
City (U.S. EPA, 2013, section 6.2.7.2) (Silverman and Ito, 2010). This study examined the
association of 8-hour max Os concentrations with severe acute asthma admissions (i.e., those
admitted to the Intensive Care Unit [ICU]) during the warm season in the years 1999 through
3A-2
-------
2006 (Silverman and Ito, 2010)). The investigators reported positive associations between Os and
ICU asthma admissions for the 6- to 18-year age group for a 30 ppb increase in max 8-hour
average Os concentrations, but little evidence of associations for the other age groups examined
(<6 years, 19-49, 50+, and all ages). However, positive associations were observed for each
age-stratified group and all ages for non-ICU asthma admissions, but again the strongest
association was reported for the 6- to 18-years age group. In two-pollutant models, Cb effect
estimates for both non-ICU and ICU hospital admissions remained robust to adjustment for
PM2.5. In an additional analysis, using a smooth function, the authors examined whether the
shape of the concentration-response curve for Os and asthma hospital admissions (i.e., both
general and ICU for all ages) is linear. When comparing the curve to a linear fit line, the authors
found that the linear fit was a reasonable approximation of the concentration-response
relationship between Os and asthma hospital admissions, but the limited data density at relatively
low Os concentrations contributes to uncertainty in the shape of the concentration-response
relationship at the low end of the distribution of Cb concentrations (U.S. EPA, 2013, Figure 6-
16).
In contrast to COPD and asthma, the evidence for pneumonia-related admissions was less
consistent. Medina-Ramon et al. (2006) examined the association between short-term ambient Os
concentrations and Medicare hospital admissions among individuals > 65 years of age for
pneumonia. The authors reported an increase in pneumonia hospital admissions in the warm
season for a 30 ppb increase in 8-hour max Os concentrations, with no evidence of an association
in the cool season or year round. In two-pollutant models restricted to days for which PMio data
was available, the association between Os exposure and pneumonia hospital admissions
remained robust. In contrast, Zanobetti and Schwartz (2006) reported a decrease in pneumonia
admissions for a 20 ppb increase in 24-hour average Os concentrations in Boston for the average
of lags 0 and 1 day.
The magnitude of associations with respiratory-related hospital admissions may be
underestimated due to behavioral modification in response to forecasted air quality (U.S. EPA,
2013, section 4.6.6). Recent studies (Neidell and Kinney, 2010; Neidell, 2009) conducted in
Southern California demonstrates that controlling for avoidance behavior increases Os effect
estimates for respiratory hospital admissions, specifically for children and older adults. This
study shows that on days where no public alert warning of high Os concentrations was issued,
there was an increase in asthma hospital admissions. Although only one study has examined
averting behavior and this study is limited to the outcome of asthma hospital admissions in one
location and time period (i.e., Los Angeles, CA for the years 1989-1997), it does provide
preliminary evidence indicating that some epidemiologic studies may underestimate associations
3A-3
-------
between Os and health effects by not accounting for behavioral modification when public health
alerts are issued.
Emergency Department Visits for All Respiratory Causes
A large single-city study conducted in Atlanta by Tolbert et al. (2007), and subsequently
reanalyzed by Darrow et al. (2011) using different air quality data and evaluating associations
with different metrics, provides evidence for associations between short-term exposures to
ambient Os concentrations and respiratory emergency department visits. Tolbert et al. (2007)
reported an increase in respiratory emergency department visits for a 30 ppb increase in 8-hour
max Os concentrations during the warm season. In copollutant models with CO, NCh, and PMio,
limited to days in which data for all pollutants were available, associations between Os and
respiratory emergency department visits remained positive, but were attenuated. Darrow et al.
(2011) reported the strongest associations with respiratory emergency department visits for 8-
hour daily max, 1-hour daily max, and day-time Os exposure metrics (all associations positive
and statistically significant), while positive, but statistically non-significant, associations were
reported with 24-hour average and commuting period exposure metrics. In addition, a negative
association was observed when using the night-time exposure metric (U.S. EPA, 2013, Figure 6-
17). The results of Darrow et al. (2011) suggest that averaging over nighttime hours may lead to
smaller Os effect estimates for respiratory emergency department visits due to dilution of
relevant Os concentrations (i.e., the higher concentrations that occur during the daytime); and
potential negative confounding by other pollutants (e.g., CO, NO2) during the nighttime hours
(U.S. EPA, 2013, section 6.2.7.3)
Cause-Specific Emergency Department Visits
In evaluating asthma emergency department visits in an all-year analysis, a Canadian
multi-city study (Stieb et al., 2009) reported that 24-hour Os concentrations were positively
associated with emergency department visits for asthma at lag 1 and lag 2. Though the authors
did not present seasonal analyses, they stated that no associations were observed with emergency
department visits in the winter season, suggesting that the positive associations reported in the
all-year analysis were due to the warm season (Stieb et al., 2009). In addition to asthma, the
authors reported that Os was positively associated with COPD emergency department visits in
all-year analyses, but that associations with COPD visits were statistically significant only for the
warm season (i.e., April-September).
5A-4
-------
Several single-city studies have also provided evidence for positive associations between
asthma emergency department visits and ambient Cb concentrations. Ito et al. (2007) reported
positive and statistically significant associations with asthma emergency department visits in
New York City during the warm season, and an inverse association in the cool season, for a 30
ppb increase in 8-hour max Cb concentrations. In two-pollutant models with PM2.5, NCh, SCh,
and CO, the authors found that Os risk estimates were not substantially changed during the warm
season (U.S. EPA, 2013, Figure 6-20; Table 6-29).
Strickland et al. (2010) examined the association between Os exposure and pediatric
asthma emergency department visits (ages 5-17 years) in Atlanta using air quality data over the
same years as Darrow et al. (2011) and Tolbert et al. (2007), but using population-weighting to
combine daily pollutant concentrations across monitors. Strickland et al. (2010) reported an
increase in emergency department visits for a 30 ppb increase in 8-hour max Os concentrations
in an all-year analysis. In seasonal analyses, stronger associations were observed during the
warm season (i.e., May-October) than the cold season. In co-pollutant analyses that included CO,
NO2, PM2.5 elemental carbon, or PM2.5 sulfate, Strickland et al. (2010) reported that Os risk
estimates were not substantially changed. The authors also examined the concentration-response
relationship between Os exposure and pediatric asthma emergency department visits and
reported that positive associations with Os persist at 8-hour ambient Os concentrations (3-day
average of 8-hour daily max concentrations) at least as low as 30 ppb.
In a single-city study conducted in Seattle, WA, Mar and Koenig (2009) examined the
association between Os exposure and asthma emergency department visits for children (< 18)
and adults (> 18). For children, positive and statistically significant associations were reported
across multiple lags, with the strongest associations observed at lag 0 and lag 3. Ozone was also
found to be positively associated with asthma emergency department visits for adults at all lags,
except at lag 0. The slightly different lag times for children and adults suggest that children may
be more immediately responsive to Os exposures than adults (Mar and Koenig, 2009).
In addition to the U.S. single-city studies discussed above, a single-city study conducted
in Alberta, Canada (Villeneuve et al., 2007) provides support for the findings from Stieb et al.
(2009), but also attempts to identify those lifestages at greatest risk for Os-associated asthma
emergency department visits. Villeneuve et al. reported an increase in asthma emergency
department visits in an all-year analysis across all ages with associations being stronger during
3A-5
-------
the warmer months. When stratified by age, the strongest associations were observed in the
warm season for individuals 5-14 and 15-44. These associations were not found to be
confounded by the inclusion of aeroallergens in age-specific models.
3.1 REFERENCES
Biggeri, A; Baccini, M; Bellini, P; Terracini, B. (2005). Meta-analysis of the Italian studies of short-term effects of
air pollution (MISA), 1990-1999. Int J Occup Environ Health 11: 107-122.
Cakmak, S; Dales, RE; Judek, S. (2006b). Respiratory health effects of air pollution gases: Modification by
education and income. Arch Environ Occup Health 61: 5-10. http://dx.doi.org/10.3200/AEOH.61.L5-10
Dales, RE; Cakmak, S; Doiron, MS. (2006). Gaseous air pollutants and hospitalization for respiratory disease in the
neonatal period. Environ Health Perspect 114: 1751-1754. http://dx.doi.org/10.1289/ehp.9044
Darrow, LA; Klein, M; Sarnat, JA; Mulholland, JA; Strickland, MJ; Sarnat, SE; Russell, AG; 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: 10-19.
.Ito, K; Thurston, GD; Silverman, RA. (2007b). Characterization of PM2.5, gaseous pollutants, and meteorological
interactions in the context of time-series health effects models. J Expo Sci Environ Epidemiol 17: S45-S60.
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; Zanobetti, A. (2009). Air
pollution and health: A European and North American approach (APHENA). (Research Report 142).
Boston, MA: Health Effects Institute. http://pubs.healtheffects.org/view.php?id=327
Lin, S; Bell, EM; Liu, W; Walker, RJ; Kim, NK; Hwang, SA. (2008a). Ambient ozone concentration and hospital
admissions due to childhood respiratory diseases in New York State, 1991-2001. Environ Res 108: 42-47.
http://dx.doi.0rg/10.1016/j.envres.2008.06.007
Mar, TF; Koenig, JQ. (2009). Relationship between visits to emergency departments for asthma and ozone exposure
in greater Seattle, Washington. Ann Allergy Asthma Immunol 103: 474-479.
Medina-Ramon, M; Zanobetti, A; Schwartz, J. (2006). The effect of ozone and PM10 on hospital admissions for
pneumonia and chronic obstructive pulmonary disease: A national multicity study. Am J Epidemiol 163:
579-588. http://dx.doi.org/10.1093/aje/kwj078
Neidell, M. (2009). Information, avoidance behavior, and health: The effect of ozone on asthma hospitalizations.
Journal of Human Resources 44: 450-478.
Neidell, M; Kinney, PL. (2010). Estimates of the association between ozone and asthma hospitalizations that
account for behavioral responses to air quality information. Environ Sci Pol 13: 97-103.
http://dx.doi.0rg/10.1016/j.envsci.2009.12.006
Silverman, RA; Ito, K. (2010). Age-related association of fine particles and ozone with severe acute asthma in New
York City. J Allergy Clin Immunol 125: 367-373. http://dx.doi.org/10.1016/jjaci.2009.10.061
Stieb, DM; Szyszkowicz, M; Rowe, BH; Leech, JA. (2009). Air pollution and emergency department visits for
cardiac and respiratory conditions: A multi-city time-series analysis. Environ Health Global Access Sci
Source 8: 25. http://dx.doi.org/10.1186/1476-069X-8-25
5A-6
-------
Strickland, MJ; Darrow, LA; Klein, M; Flanders, WD; Sarnat, JA; Waller, LA; Sarnat, SE; Mulholland, JA; Tolbert,
PE. (2010). Short-term associations between ambient air pollutants and pediatric asthma emergency
department visits. Am J Respir Crit Care Med 182: 307-316. http://dx.doi.org/10.1164/rccm.200908-
1201OC
Tolbert, PE; Klein, M; Peel, JL; Sarnat, SE; 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: S29-S35.
http://dx.doi.org/10.1038/sj.jes.7500625
U.S. EPA (2013). Integrated Science Assessment of Ozone and Related Photochemical Oxidants (Final Report).
U.S. Environmental Protection Agency, Washington, DC. EPA/600/R-10/076F. Available at:
http://www.epa.gov/ttn/naaqs/standards/ozone/s_o3_2008_isa.html
Villeneuve, PJ; Chen, L; Rowe, BH; 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
Global Access Sci Source 6: 40. http://dx.doi.org/10.1186/1476-069X-6-40
Wong, CM; Yang, L; Thach, TQ; Chau, PY; Chan, KP; Thomas, GN; Lam, TH; Wong, TW; Hedley, AJ; Peiris, JS.
(2009). Modification by influenza on health effects of air pollution in Hong Kong. Environ Health Perspect
117: 248-253. http://dx.doi.org/10.1289/ehp.11605
Yang, Q; Chen, Y; Krewski, D; Burnett, RT; Shi, Y; Mcgrail, KM. (2005). Effect of short-term exposure to low
levels of gaseous pollutants on chronic obstructive pulmonary disease hospitalizations. Environ Res 99: 99-
105. http://dx.doi.0rg/10.1016/j.envres.2004.09.014
Zanobetti, A; Schwartz, J. (2006). Air pollution and emergency admissions in Boston, MA. J Epidemiol Community
Health 60: 890-895. http://dx.doi.org/10.1136/iech.2005.039834
5A-7
-------
APPENDIX 3B: AMBIENT O3 CONCENTRATIONS IN LOCATIONS OF
HEALTH STUDIES
Annual 4th highest daily maximum Ch concentrations for all U.S. monitors operating
during the 1975 - 2010 period were retrieved from EPA's AQS database. These data were used
to calculate Os design values for the 2008 8-hour Os NAAQS of 0.075 parts per million (ppm)
according to 40 CFR part 50, Appendix P. Design values were calculated for each Os monitor
and each 3-year period between 1975-1977 and 2008-2010 whenever sufficient data were
available.
Ozone Design Values in Study Locations
Ozone monitors were matched to 200 health study locations on a case-by-case basis, using
the following guidelines:
1) Areas defined by a Metropolitan Statistical Area (MSA) were matched with Os monitors
by incorporating all of the monitors located in within the MSA boundaries.
2) Areas not represented by a MSA were matched to monitors by incorporating all of the
monitors in the county central to location of the health study area.
3) In some cases, EPA staff made judgment calls. For example, EPA staff matched the Los
Angeles, CA study area to the Los Angeles-Long Beach-Santa Ana, CA MSA defined by
Los Angeles County, CA and Orange County, CA, while the Long Beach, CA study area
was matched to Los Angeles County, CA and the Santa Ana, CA study area was matched
to Orange County, CA.
In some cases, EPA staff matched two or more study areas to the same county or MSA.
In other cases, a study area was matched to a MSA and another study area was matched to a
county within the same MSA. For each 3-year period, the area design value was determined by
the monitor reporting the highest design value in the county or MSA. This has two implications
for the design values:
1) Design values are sensitive to changes in the monitoring network. The addition or
discontinuation of Os monitors in an area may cause increases or decreases in the design
value trend.
2) Only valid design values are reported. According to 40 CFR Part 50, Appendix P, design
values greater than the level of the NAAQS (0.075 ppm) are always valid, while design
values less than or equal to 0.075 ppm must have 75% annual data completeness in order
to be valid. This may cause anomalies in the design value trend. For example, a monitor
may report a valid design value based on as few as 12 days of data, or a monitor with less
than 75% annual data completeness may have valid design values in some 3-year periods
and invalid design values in others.
3B-1
-------
We have identified design values for the U.S. Os epidemiologic studies identified in Sections
3.1.4.2 and 3.1.4.3 of the second draft Policy Assessment (see Tables 3D-1 to 3D-4). For each
study, design values were identified for the cities evaluated and for the years over which the
study was conducted. These design values are reported in tables A-l to A-22 of the Wells et al,
2012 memo "Analysis of Recent U.S. Ozone Air Quality Data to Support the Os NAAQS
Review and Quadratic Rollback Simulations to Support the First Draft of the Risk and Exposure
Assessment".
Table 3B-1. Number of Study Cities from Multicity Epidemiologic Studies of Hospital
Admissions and Emergency Department Visits Using Short-Term Os Metrics with 3-Year
Averages of Annual 4th Highest Daily Maximum 8-hour Oa Concentrations < 75 ppb1
Study
Location
Endpoint2
% Increase (95% CI)3
# Study Cities < 75
ppb over entire
study period
All-year
Medina-
Ramon et al.
(2006)
Katsouyanni
et al. (2009)
Dales et al.
(2006)
Stieb et al.
(2009)
Cakmak et al.
(2006)
36 U.S. cities
14 U.S. cities
12 Canadian
Cities
11 Canadian
cities
7 Canadian
cities
10 Canadian
cities
COPDHA
Pneumonia
HA
Respiratory
HA
Respiratory
HA
Respiratory
HA
Asthma ED
COPD ED
Respiratory
HA
All year: 0.24 (-0.78, 1.21)
Warm season: 1.63 (0.48,
2.85)
All year: 1.81 (-0.72, 4.52)
Warm Season: 2.49 (1.57,
3.47)
All Year: 2.38 (0.00, 4.89)
Warm Season: 2. 14 (-0.63,
4.97)
All year: 2.4 (0.5 1,4.40)
Warm Season: 4.1 (1.4,
6.8)
All year: 5.41(2.88,7.96)
All year: 3. 48 (0.33, 6.76)
All year: 4.03 (-0.54, 8.62)
Warm season: 6.76 (0.11,
13.9)
All year: 4.38 (2. 19, 6.46)
4
2
10
7
5
7
1 For U.S. study areas, we used EPA's Air Quality System (AQS) (http://www.epa.gov/ttn/airs/airsaqs/') to identify
8-hour Os concentrations. For Canadian study areas, we used publically available air quality data from the
Environment Canada National Air Pollution Surveillance Network (http: //www. etc -
cte.ec.gc.ca/napsdata/main/aspxX We followed the data handling protocols for calculating design values as detailed
in 40 CFR Part 50, Appendix P.
2HA stands for hospital admissions; ED stands for emergency department visits.
3Ozone effect estimates are taken from Table 6-28 in the ISA (U.S. EPA, 2013a).
3B-2
-------
Table 3B-2. Number of Study Cities from Multicity Epidemiologic Studies of Mortality
Using Short-Term Os Metrics with 3-Year Averages of Annual 4th Highest Daily Maximum
8-hour Oa Concentrations < 75 ppb
Study
Location
Endpoint
% Increase (95%
CI)4
# Study Cities < 75
ppb over entire
study period
All-year
Schwartz
(2005)
Bell et al.
(2007)
Bell and
Dominici
(2008)
Katsouyanni
et al. (2009)
Bell et al.
(2004)
Katsouyanni
et al. (2009)
14 U.S. cities
98 U.S.
communities
98 U.S.
communities
90 U.S. cities
95 U.S.
communities
12 Canadian
cities
Non-accidental
mortality
Non-accidental
mortality
Non-accidental
mortality
Non-accidental
mortality
Respiratory
mortality
Cardiovascular
mortality
Non-accidental
mortality
Non-accidental
mortality
Respiratory
mortality
Cardiovascular
mortality
0.76(0.13, 1.40)
0.64 (0.34, 0.92)
1.04(0.56, 1.55)
3.02(1.10,4.89)
2.54 (-3.32, 8.79) for
<75; 1.10 (-6.48, 9.21)
for 75+
3.83 (-0.16, 7.95) for
< 75; 2.30 (-1.33,
6.04) for 75+
1.04(0.54, 1.55)
0.73 (0.23, 1.20)
0.13 (-1.60, 1.90);
-0.60 (-2. 70, 1.60) for
75+
0.87 (-0.35, 2.10) for
<75; 1.1 (0.10,2.20)
for 75+
1
6
6
6
6
8
Warm Season
Schwartz
(2005)
Zanobetti
and
Schwartz
(2008a)
Zanobetti
and
Schwartz
(2008b)
14 U.S.
Cities
48 U.S. cities
48 U.S. cities
Non-accidental
mortality
Non-accidental
mortality
Non-accidental
mortality
Respiratory
mortality
Cardiovascular
mortality
1.00(0.30, 1.80)
1.51 (1.14, 1.87)
1.60(0.84,2.33)
2.51 (1.14,3.89)
2.42(1.45,3.43)
1
4
4
40zone effect estimates are taken from Tables 6-42 and 6-53 in the ISA (U.S. EPA, 2013a).
3B-3
-------
Medina-
Ramon and
Schwartz
(2008)
Franklin and
Schwartz
(2008)
Katsouyanni
et al. (2009)
Bell et al.
(2004)
Katsouyanni
et al. (2009)
48 U.S. cities
18 U.S.
communities
90 U.S. cities
95 U.S.
communities
12 Canadian
Cities
Non-accidental
mortality
Non-accidental
mortality
Non-accidental
mortality
Respiratory
mortality
Cardiovascular
mortality
Non-accidental
mortality
Non-accidental
mortality
1.96(1.14,2.82)
1.79(0.90,2.68)
3.83(1.90,5.79)
4.40 (-2.10, 11.3);
4.07 (-4.23, 13.0) for
75+
6.78(2.70, 11.0) for
<75; 3. 18 (-0.47, 6.95)
for 75+
0.78(0.26, 1.30)
0.42(0.16,0.67)
4
1
6
6
8
3B-4
-------
Table 3B-3. Number of Study Cities from Single-City Epidemiologic Studies Using Short-
Term Os Metrics with 3-Year Averages of Annual 4th Highest Daily Maximum 8-hour Os
Concentrations < 75 ppb
Study
Location
Age
Endpoint
% Increase (95%
CI)5
# Study Cities
< 75 ppb over
entire study
period
All-year
Strickland et al.
(2010)
Atlanta
Children
Asthma ED
visits
6.38(3.19,9.57)
0
Warm season
Ito et al. (2007)
Darrow et al.
(2011)
Tolbert et al.
(2007)
Strickland et al.
(2010)
Silverman and
Ito (2010)
Mar and Koenig
(2009)
New
York City
Atlanta
Atlanta
Atlanta
New
York City
Seattle,
WA
All
All
All
Children
6 to 18
years
All
18+
Asthma ED
visits
Respiratory
ED visits
Respiratory
ED visits
Asthma ED
visits
Asthma HA
Asthma HA
Asthma ED
visits
16.9(10.9,23.4)
2.08(1.25,2.91)
3.90(2.70,5.20)
8.43 (4.42, 12.7)
28.2(15.3,41.5)
12.5 (8.27, 16.7)
19.1 (3.00,40.5)
0
0
0
0
0
0
1
50zone effect estimates are taken from Table 6-28 in the ISA (U.S. EPA, 2013a).
3B-5
-------
Table 3B-4 Number of Study Cities from Epidemiologic Studies Using Long-Term Os
Metrics with 3-Year Averages of Annual 4th Highest Daily Maximum 8-hour
Os Concentrations > 75, 70, 65, or 60 ppb
Study
Islam etal. 2008, 2009
Jerrett et al. 2009
Lin et al. 2008
Mengetal. 2010
Moore etal. 2008
Salam etal. 2009
Zanobetti & Schwartz 2011
Number
of Cities
II6
947
26s
7
8
II9
105
Study
Period
1994-2003
1977-2000
1991-2001
1997-2002
1980-2000
1992-2005
1985-2006
Number
(Percent) of
Cities with
Maximum
cone >75
11 (100%)
91 (97%)
24 (92%)
7 (100%)
8 (100%)
12 (100%)
100 (95%)
Number
(Percent) of
Cities with
Maximum
cone >70
11 (100%)
92 (98%)
24 (92%)
7(100%)
8(100%)
12 (100%)
104 (99%)
Number
(Percent) of
Cities with
Maximum cone
>65
11 (100%)
93 (99%)
26 (100%)
7 (100%)
8 (100%)
12 (100%)
104 (99%)
Number
(Percent) of
Cities with
Maximum
cone >60
11 (100%)
94(100%)
26(100%)
7 (100%)
8 (100%)
12 (100%)
104 (99%)
6 Study authors included 12 cities in their analyses, air quality data that met completeness criteria described above
were available for 11 cities
7 Study authors included 96 cities in their analyses, air quality data that met completeness criteria described above
were available for 94 cities
8 Study authors included 27 cities in their analyses, air quality data that met completeness criteria described above
were available for 26 cities
9 Study authors included 12 cities in their analyses, air quality data that met completeness criteria described above
were available for 11 cities
3B-6
-------
Relationship between average 24-hour and highest 8-hour Os concentrations for cities
analyzed by Bell et al. (2006)
Bell et al. (2006) reported associations between mortality and 24-hour average Os
concentrations (i.e., averaged across monitors in cities with multiple monitors) in a multi-city
study of 98 U.S. cities. Positive associations persisted in a series of analyses that restricted Os
concentrations to those below various cut points (cut points ranged from 5 to 60 ppb in 5 ppb
increments). To facilitate consideration of these cut point analyses for the second draft of the Os
Policy Assessment, so as to match the form and averaging time of the existing primary standard,
we evaluated the relationship between 24-hour average Os concentrations, averaged across
monitors in cities with multiple monitors, and the highest 8-hour daily maximum Os
concentrations among the individual monitors in each city.
EPA staff retrieved daily 24-hour average and 8-hour maximum Os concentrations
reported to EPA by monitors in the 98 study areas defined in Bell et al. (2006) during the 1987-
2012 period from EPA's Air Quality System (AQS) database. Next, EPA staff obtained the
study area boundaries from the published study (Bell et al., 2006) and used them to determine
which Os monitoring sites were associated with each study area. The 24-hour average Os
concentrations were averaged spatially across all available monitors within each study area on
each day where monitoring data were collected. Next, days where the area-wide 24-hour average
concentration (i.e., averaged spatially across monitors in areas with multiple monitors) was
greater than 60 ppb were removed from the data. Based on the data remaining (i.e., with 24-hour
average concentrations of 60 ppb or below), the annual 4th highest 8-hour daily maximum
concentrations were identified for each study area and for each year from 1987-2012 (Table 3D-
3). This process was repeated by further removing days with area-wide 24-hour average
concentrations greater than 55 ppb, 50 ppb, etc., down to 5 ppb, and re-calculating the same
statistics after each removal. The resulting dataset consisted of the annual 4th highest 8-hour
daily maximum concentrations for all study areas.
3B-7
-------
Table 3B-5 Number of Study Cities with 4th Highest 8-hour Daily Maximum
Concentrations Greater Than the Level of the Current Standard and
Potential Alternative Standards For Various Cut-Point Analyses Presented
in Bell et al. (2006)10
Cut-point for 2-day moving average across monitors and cities (24-h avg)
20 25 30 35 40 45 50 55 60 All
Number (%) of Cities
with 4th highest >75 0(0%) 0(0%) 12(12%) 52(53%) 77(79%) 88(90%) 93(95%) 94(96%) 94(96%) 94(96%)
(any year; 1987-2000)
Number (%) of Cities
with 4th highest >70 0(0%) 3(3%) 31(32%) 77(79%) 86(88%) 93(95%) 94(96%) 94(96%) 95(97%) 95(97%)
(any year; 1987-2000)
Number (%) of Cities
with 4th highest >65 0(0%) 10(10%) 58(59%) 84(86%) 93(95%) 94(96%) 94(96%) 94(96%) 94(96%) 94(96%)
(any year; 1987-2000)
Number (%) of Cities
with 4th highest >60 1(1%) 36(37%) 74(76%) 93(95%) 96(8%) 97(99%) 97(99%) 97(99%) 97(99%) 97(99%)
(any year; 1987-2000)
10 Study authors included 98 cities in their analyses, air quality data only available for 95
3B-8
-------
Relationship between average and highest 8-hour daily maximum Os concentrations for
New York City, as analyzed by Silverman and Ito (2010)
EPA staff retrieved daily maximum 8-hour Os concentrations for the 13 monitors in the
New York City area used in the Silverman and Ito (2010) study for April-August of 1999-2006
from the AQS database. Next, EPA staff spatially averaged these concentrations across monitors
for each day during this period, and then paired them with the highest 8-hour daily maximum
value reported across the 13 monitors on each day.
Next, the range of observed average daily maximum 8-hour concentrations was broken
into 5 ppb increments. The number of days where the area-wide average daily maximum 8-hour
concentration fell within the increment and the number of days where one or more monitored 8-
hour daily maximum values were greater than 75, 70, 65 and 60 ppb were recorded for each 5
ppb increment. These numbers are summarized in Table 3D-4.
Table 3B-6 Summary statistics for Observed Os Concentrations in the New York City
Area, April - August 1999 - 2006
2-day moving average across monitors (ppb)
Days > 75 ppb
Days > 70 ppb
Days > 65 ppb
Days > 60 ppb
11 to 20
(62 days)
0
0
0
0
21 to 25 26 to 30 31 to 35 36 to 40 41 to 45 46 to 50 51 to 55 56 to 60
(92days) (178days) (206days) (236days) (196days) (153days) (111 days) (71days)
0 1 0 1 2 9 15 20
0 1 4 1 12 17 23 30
0 1 6 5 18 37 42 45
0 2 7 12 39 67 61 53
3B-9
-------
Relationship between average and highest 8-hour daily maximum Os concentrations for
Atlanta, as analyzed by Strickland et al. (2010)
For our assessment of the Strickland et al. (2010) study, based in the Atlanta metropolitan
area, we retrieved 8-hour daily maximum concentration data for 4 of the 5 monitors used in the
study during the study period (May-October, 1993-2004) from the AQS database. The 5th
monitor was a part of the Southeastern Aerosol Research and Characterization (SEARCH)
network, which does not report data to EPA. EPA staff calculated the area-wide average of the
8-hour daily maximum concentrations for each day, and compared to population-weighted
average concentrations obtained from the author. The correlation between the arithmetic average
values and the population-weighted average values was very high (R = 0.985), thus EPA staff
deemed the arithmetic average to be a suitable surrogate for the population-weighted average
used in the study. Finally, 3-day moving averages were calculated from the daily area-wide
average values (matching the air quality metric used in the study), and paired with the highest
monitored 8-hour daily maximum value occurring during each 3-day period.
Next, the range of observed average daily maximum 8-hour concentrations was broken
into 5 ppb increments. The number of days where the area-wide average daily maximum 8-hour
concentration fell within the increment and the number of days where one or more monitored 8-
hour daily maximum values were greater than 75, 70, 65 and 60 ppb were recorded for each 5
ppb increment. These numbers are summarized in Table 3D-5.
Table 3B-7 Summary statistics for Observed Os Concentrations in the Atlanta Area,
April - August 1999 - 2006
3-day moving average across monitors (ppb)
26-30 31-35 36-40 41-45 46-50 51-55 56-60 61-65 66-70 71 to 75 76 to 80
(75 days) (144 days) (165 days) (210 days) (235 days) (244 days) (272 days) (234 days) (169 days) (124 days) (106 days)
Days > 75
Days > 70
Days > 65
Days > 60
0
0
1
1
0
0
0
2
2
6
8
15
2
6
19
33
10
20
38
68
24
49
75
115
53
81
118
152
80
111
147
173
89
107
133
147
87
96
106
116
87
95
100
102
3B-10
-------
Relationship between annual and highest 1-hour daily maximum Os concentrations for 12
study areas, as analyzed by Jerrett et al. (2009)
The Jerrett et al. (2009) study used a long-term metric based on seasonal averages of 1-
hour daily maximum Os concentrations to evaluate associations between respiratory mortality
and long-term or repeated exposures to Os. Authors divided study cities into quartiles based on
these seasonal averages of 1-hour daily Os concentrations. Using AQS, we identified the 3-year
averages of annual 4th highest daily maximum 8-hour Os concentrations in study cities during the
study period. Table 3D-6 presents the means and maximums of these concentrations over the
study period.
In addition, for the 12 urban case study areas included in the epidemiology-based risk
assessment of the 2nd draft of the Health REA we identified the seasonal averages of 1-hour daily
maximum concentrations (i.e., the Os metric evaluated by Jerrett et al., 2009) for air quality
adjusted to the current and alternative standards. Specifically, for adjusted air quality "quarterly"
averages of 1-hour concentrations for April-June and July-August were calculated for each area
and year. The quarterly values were considered to be valid if valid daily maximum 1-hour
values were available for at least 75% of the days in the quarter. The two quarterly values were
then averaged, as was done by Jerrett et al. (2009) to generate the long-term metric used in the
study. This process was repeated for the various model-based adjustment scenarios in each of
the 12 study areas. Summary statistics based on this seasonal average of daily Os concentrations
are presented in Table 3D-7 for recent air quality and for air quality adjusted to just meet the
current and alternative standards.
3B-11
-------
Table 3B-8
Three-Year Averages of Annual 4th Highest Daily Maximum 8-hour
Concentrations in 9411 Study Areas Examined in Jerrett et al. (2009)
Cities in the lowest quartile of average exposure12
Cities in the highest three
quartiles of average
exposure13
City
Charleston, WV
Chicago, IL
Colorado Springs, CO
Corpus Christi, TX
Detroit, Ml
Flint, Ml
Ft. Lauderdale, FL
Kansas City, MO
Lansing, Ml
Madison, Wl
Minneapolis, MN
New Orleans, LA
Orlando, FL
Portland, OR
Providence, Rl
Salinas, CA
San Antonio, TX
San Francisco, CA
San Jose, CA
Seattle, WA
Tacoma, WA
Vallejo, CA
Wichita, KS
Charleston, SC
Charlotte, NC
Chattanooga, TN
Cincinnati, OH
Cleveland, OH
Columbia, SC
Columbus, OH
Dallas/Ft Worth, TX
Dayton, OH
Denver, CO
Mean over
study
period
81
103
62
82
95
83
74
87
81
82
74
86
79
81
110
68
85
88
91
78
78
74
75
79
97
90
101
98
85
93
106
95
83
Max over
study
period
99
114
66
89
103
91
79
97
90
102
80
99
82
91
124
74
92
96
103
88
88
82
81
90
112
97
119
108
109
103
118
122
91
11 Jerrett et al. (2009) examined 96 MSAs; this analysis included the 94 cities that met data completeness criteria
described above, after linking monitors to MSAs (see lines 10-28, above).
12 Based on visual inspection of Figure 1 in Jerrett et al. (2009)
13 Based on visual inspection of Figure 1 in Jerrett et al. (2009)
3B-12
-------
El Paso, TX
Evansville, IN
Fresno, CA
Gary, IN
Greely, CO
Greensboro, NC
Greenville, SC
Harrisburg, PA
Houston, TX
Huntington, WV
Indianapolis, IN
Jackson, MS
Jacksonville, FL
Jersey City, NJ
Johnstown, PA
Kenosha, Wl
Knoxville, TN
Lancaster, PA
Las Vegas, NV
Lexington, KY
Little Rock, AR
Los Angeles, CA
Memphis, TN
Milwaukee, Wl
Nashville, TN
Nassau, NY
New Haven, CT
New York City, NY
Newark, NJ
Norfolk, VA
Oklahoma City, OK
Philadelphia, PA
Phoenix, AZ
Pittsburgh, PA
Portland, ME
Portsmouth, NH
Racine, Wl
Raleigh, NC
Reading, PA
Richmond, VA
Riverside, CA
Roanoke, VA
85
93
112
91
69
89
86
94
121
94
93
79
81
106
90
101
91
94
80
88
86
193
94
103
94
NA14
116
118
90
91
86
117
86
101
106
92
102
90
99
94
196
83
96
100
123
105
75
100
94
103
140
103
103
98
87
118
107
114
97
101
85
99
107
248
103
117
106
NA
136
129
105
101
93
136
96
123
117
104
124
104
114
104
245
95
14 Air quality data did not meet completeness criteria described above
3B-13
-------
Rochester, NY
Sacramento, CA
San Diego, CA
Shreveport, LA
South Bend, IN
Springfield, MA
St Louis, MO
Steubenville, OH
Syracuse, NY
Tampa, FL
Toledo, OH
Trenton, NJ
Tucson, AZ
Ventura, CA
Washington, DC
Wilmington, DE
Worcester, MA
York, PA
Youngstown, OH
89
110
121
83
90
102
105
82
85
85
93
112
76
118
105
103
92
95
93
99
118
141
88
102
115
122
99
96
91
108
124
82
132
116
116
102
107
103
3B-14
-------
Table 3B-9 Long-Term Os Concentrations in 12 Urban Case Study Areas (Using the
Metric Evaluated by Jerrett et al., 2009) for Recent Air Quality and Air
Quality Adjusted to Meet Standard Levels of 75, 70, 65, and 60 ppb
Atlanta
Baltimore
Boston
Cleveland
Denver
Detroit
Houston
Los Angeles
New York City
Philadelphia
Sacramento
Saint Louis
Air Quality
Adjusted to:
Recent
75
70
65
60
Recent
75
70
65
60
Recent
75
70
65
60
Recent
75
70
65
60
Recent
75
70
65
60
Recent
75
70
65
60
Recent
75
70
65
60
Recent
75
70
65
60
Recent
75
70
65
60
Recent
75
70
65
60
Recent
75
70
65
60
Recent
75
70
65
60
2006
(AdjYrs 2006- 2008)
65
53
50
47
45
60
54
52
49
46
49
48
46
44
43
51
49
47
45
41
63
62
60
58
53
50
50
48
47
45
53
48
47
46
45
65
58
55
52
N/A
53
47
N/A
N/A
N/A
56
51
49
47
45
66
55
52
50
47
58
53
50
47
44
2007
(Adj Yrs 2006-2008)
63
52
49
46
44
59
54
51
49
46
50
49
47
45
43
52
50
48
45
41
63
61
59
58
53
54
52
50
49
46
48
46
45
44
43
61
59
56
53
N/A
54
47
N/A
N/A
N/A
59
52
50
48
46
59
50
48
46
44
58
53
51
48
45
2008
(Adj Yrs 2008-2010)
57
53
49
46
44
57
53
51
48
46
46
49
48
46
44
53
51
48
45
41
63
63
62
59
53
51
N/A
51
49
46
47
47
46
45
43
64
60
57
54
N/A
55
51
N/A
N/A
N/A
57
54
51
49
47
65
54
51
49
46
52
51
50
48
45
2009
(Adj Yrs 2008-2010)
50
47
44
42
40
52
49
48
46
44
45
45
44
43
41
49
47
45
43
40
58
58
58
56
51
48
N/A
49
47
45
47
48
47
46
44
62
60
58
54
N/A
48
47
N/A
N/A
N/A
51
49
47
45
43
61
51
49
47
44
51
50
48
46
43
2010
(Adj Yrs 2008- 2010)
56
52
49
46
44
60
55
53
50
48
49
48
48
46
44
54
51
48
45
42
60
60
58
55
50
52
N/A
52
50
47
46
46
46
45
44
57
58
56
53
N/A
55
51
N/A
N/A
N/A
58
54
52
49
47
55
48
46
44
42
55
54
52
49
46
3B-15
-------
REFERENCES
Bell, ML; Dominici, F. (2008). Effect modification by community characteristics on the short-term effects of ozone
exposure and mortality in 98 US communities. Am J Epidemiol 167: 986-997.
http://dx.doi.org/10.1093/aje/kwm396
Bell, ML; Kim, JY; Dominici, F. (2007). Potential confounding of paniculate matter on the short-term association
between ozone and mortality in multisite time-series studies. Environ Health Perspect 115: 1591-1595.
http://dx.doi.org/10.1289/ehp.10108
Bell, ML; Peng, RD; Dominici, F. (2006). The exposure-response curve for ozone and risk of mortality and the
adequacy of current ozone regulations. Environ Health Perspect 114: 532-536.
Cakmak, S; Dales, RE; Judek, S. (2006a). Do gender, education, and income modify the effect of air pollution gases
on cardiac disease? J Occup Environ Med 48: 89-94.
http://dx.doi.org/10.1097/01.jom.0000184878.11956.4b
Cakmak, S; Dales, RE; Judek, S. (2006b). Respiratory health effects of air pollution gases: Modification by
education and income. Arch Environ Occup Health 61: 5-10. http://dx.doi.org/10.3200/AEOH.61.L5-10
Dales, RE; Cakmak, S; Doiron, MS. (2006). Gaseous air pollutants and hospitalization for respiratory disease in the
neonatal period. Environ Health Perspect 114: 1751-1754. http://dx.doi.org/10.1289/ehp.9044
Franklin, M; Schwartz, J. (2008). The impact of secondary particles on the association between ambient ozone and
mortality. Environ Health Perspect 116: 453-458.
Islam, T; Berhane, K; McConnell, R; Gauderman, WJ; Avol, E; Peters, JM; Gilliland, FD. (2009). Glutathione-S-
transferase (GST) PI, GSTM1, exercise, ozone and asthma incidence in school children. Thorax 64: 197-
202. http://dx.doi.org/10.1136/thx.2008.099366
Islam, T; McConnell, R; Gauderman, WJ; Avol, E; Peters, JM; Gilliland, FD. (2008). Ozone, oxidant defense genes
and risk of asthma during adolescence. Am J Respir Crit Care Med 177: 388-395.
http://dx.doi.org/10.1164/rccm.200706-863OC
Jerrett, M; Burnett, RT; Pope, CA, III; Ito, K; Thurston, G; Krewski, D; Shi, Y; Calle, E; Thun, M. (2009). Long-
term ozone exposure and mortality. N Engl J Med 360: 1085-1095.
http://dx.doi.org/10.1056/NEJMoa0803894
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; Zanobetti, A. (2009). Air
pollution and health: A European and North American approach (APHENA). (Research Report 142).
Boston, MA: Health Effects Institute.
Lin, S; Liu, X; Le, LH; Hwang, SA. (2008b). Chronic exposure to ambient ozone and asthma hospital admissions
among children. Environ Health Perspect 116: 1725-1730. http://dx.doi.org/10.1289/ehp.11184
Medina-Ramon, M; Zanobetti, A; Schwartz, J. (2006). The effect of ozone and PM10 on hospital admissions for
pneumonia and chronic obstructive pulmonary disease: A national multicity study. Am J Epidemiol 163:
579-588. http://dx.doi.org/10.1093/aje/kwj078
Meng, YY; Rull, RP; Wilhelm, M; Lombardi, C; Balmes, J; Ritz, B. (2010). Outdoor air pollution and uncontrolled
asthma in the San Joaquin Valley, California. J Epidemiol Community Health 64: 142-147.
http://dx.doi.org/10.1136/jech.2008.083576
3B-16
-------
Moore, K; Neugebauer, R; Lurmann, F; Hall, J; Brajer, V; Alcorn, S; Tager, I. (2008). Ambient ozone
concentrations cause increased hospitalizations for asthma in children: An 18-year study in Southern
California. Environ Health Perspect 116: 1063-1070.
Salam, MT; Islam, T; Gauderman, WJ; Gilliland, FD. (2009). Roles of arginase variants, atopy, and ozone in
childhood asthma. J Allergy Clinlmmunol 123: 596-602. http://dx.doi.org/10.1016/jjaci.2008.12.020
Schwartz, J. (2005a). How sensitive is the association between ozone and daily deaths to control for temperature?
Am J Respir Crit Care Med 171: 627-631.
Schwartz, J. (2005b). Who is sensitive to extremes of temperature? A case-only analysis. Epidemiology 16: 67-72.
Stieb, DM; Szyszkowicz, M; Rowe, BH; Leech, JA. (2009). Air pollution and emergency department visits for
cardiac and respiratory conditions: A multi-city time-series analysis. Environ Health Global Access Sci
Source 8: 25. http://dx.doi.org/10.1186/1476-069X-8-25
Zanobetti, A; Schwartz, J. (2011). Ozone and survival in four cohorts with potentially predisposing diseases. Am J
Respir Crit Care Med 184: 836-841. http://dx.doi.org/10.1164/rccm.201102-0227OC
Zanobetti, A; Schwartz, J. (2008a). Is there adaptation in the ozone mortality relationship: A multi-city case-
crossover analysis. Environ Health 7: 22. http://dx.doi.org/10.1186/1476-069X-7-22
Zanobetti, A; Schwartz, J. (2008b). 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: 184-189.
http://dx.doi.org/10.1164/rccm.200706-823OC
3B-17
-------
Appendix 5A
Ozone-Sensitive Plant SpeciesA Used by Some Tribes*
*(Based on Feedback from 3 Tribes)
Common Name
(other common names)
Red alder (Oregon alder, Western alder)
Speckled alder (Tag alder, Gray alder, Hoary
alder)
Groundnut (Wild bean, American potato bean)
Spreading Dogbane (Common dogbane)
Common milkweed
New England Aster
Green ash
Twinberry
Bee-balm
Virginia creeper
Jack pine
Lodgepole pine
White pine
Black poplar (Balsam poplar)
Quaking aspen (Trembling aspen)
Black cherry
Choke cherry
Douglas fir
Allegheny blackberry (Common blackberry)
Thimbleberry
Cutleaf coneflower (Coneflower, Golden glow)
Pussy willow
Shinning willow
American elder (White elder)
Red elderberry
Sassafras
Goldenrod
Huckleberry
Wild grape
European wine grape
Scientific Name
Alnus rubra
Alnus rugosa (Alnus incana)
Apios americana
Apocynum androsamifolium
Asclepias syriaca
Aster novae-angliae
Symphyotrichum novae-angliae
Fraxinus pennsylvanica
Lonicera involucrate
Monarda didyma
Parthenocissus quinquefolia
Pinus banksiana
Pinus contorta
Pinus strobus
Populus balsamifera
trichocarpa
Populus tremuloides
Prunus serotina
Prunus virginiana
Pseudotsuga menziesii
Rubus allegheniensis
Rubus parviflorus
Rudbeckia laciniata
Salix discolor
Salix lucida
Sambucus canadensis
Sambucus racemosa
Sassafras albidum
Solidago altissima
Vaccinium membranaceum
Vitis spp.
Vitis vinifera
Confirmed bioindicator
species
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
ASpecies included in this list are identified in one or more of the following sources:
DSP 2007 (www.2.nature.nDS.20v/air/Pubs/Ddf/fla2/NPSozonesensDDFLAG06.Ddf)
2) NFS Os Bioindicators 2006 (www.nature.nDS.sov/air/Pubs/bioindicators/index.cfm)
3) Kline et al., 2008; 4) Davis, 2007/ 2009; 5) Flagler, et al., eds., 1998
6) USDA FS FHM/FIA: Ozone Bioindicator Sampling and Estimation
(www.nrs.fs.fed.us/fia/tODics/ozone/Dubs/Ddfs/ozone%20estimation%20document.Ddf) and
Ozone Injury in West Coast Forests: 6 Years of Monitoring (2007).
5A-1
-------
This page left intentionally blank.
5A-2
-------
APPENDIX 5B: CLASS I AREAS BELOW CURRENT STANDARD AND
ABOVE 15 PPM-HRS
This appendix identifies Class I areas that might have W126 index values above 15 ppm-
hrs allowed by the current standard based on an analysis of recent Os monitoring data. Table 5B-
1 provides all monitoring sites from 1998-2002 that were at or below 75 ppb (3-year average of
4th highest maximum 8-hour average), at or above 15 ppm-hrs (3-year average), and located in
counties with Class I areas. For each year that met these 3-year requirements, we also provide
the maximum annual 8-hour Os concentration (in ppb) and W126 index value (in ppm-hrs).
5B-1
-------
Table 5B-1 Examples of Counties Containing Class I Areas where Recent 3-Year
W126 Index Values were Above 15 ppm-hrs
Concentrations were Below 75 ppb and 3- Year Average
Monitor ID
#
400380011
400380011
400380011
400380011
400380011
400380011
400380011
400380011
400380011
400380011
400380011
400380011
400380011
400380011
400380011
400380011
400380011
400380011
400380011
400380011
400380011
400510081
400510081
400580011
400510081
400510081
400510081
400580011
400580011
400580011
400580011
400580011
Years
(3-year
average)
1998-2000
1998-2000
1998-2000
2002-2004
2002-2004
2002-2004
2003-2005
2003-2005
2003-2005
2004-2006
2004-2006
2004-2006
2005-2007
2005-2007
2005-2007
2006-2008
2006-2008
2006-2008
2010-2012
2010-2012
2010-2012
2008-2010
2008-2010
2008-2010
2006-2008
2007-2009
2007-2009
1998-2000
1998-2000
1998-2000
1999-2001
1999-2001
Year
(annual)
1998
1999
2000
2002
2003
2004
2003
2004
2005
2004
2005
2006
2005
2006
2007
2006
2007
2008
2010
2011
2012
2008
2009
2010
2008
2008
2009
1998
1999
2000
1999
2000
Max 8-hour
(Ppb)
(3-year
average)
70
70
70
71
71
71
71
71
71
72
72
72
71
71
71
69
69
69
73
73
73
69
69
69
70
68
68
73
73
73
72
72
Max 8-hour
(Ppb)
(annual)
67
72
71
74
71
70
71
70
72
70
72
74
72
74
67
74
67
68
71
75
74
74
66
69
74
74
66
72
76
71
76
71
W126
(3-year
average)
15.88
15.88
15.88
15.70
15.70
15.70
16.64
16.64
16.64
16.56
16.56
16.56
16.36
16.36
16.36
16.37
16.37
16.37
18.06
18.06
18.06
15.61
15.61
15.61
19.29
15.38
15.38
18.74
18.74
18.74
17.64
17.64
W126
(annual)
14.71
16.57
16.36
14.45
18.07
14.57
18.07
14.57
17.29
14.57
17.29
17.81
17.29
17.81
13.98
17.81
13.98
17.32
13.21
19.33
21.65
22.20
11.38
14.89
22.20
22.20
11.38
18.23
21.27
16.74
21.27
16.74
Monitor Site Name
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Flagstaff Middle School
Flagstaff Middle School
Flagstaff Middle School
Grand Canyon National Park, The Abyss
Grand Canyon National Park, The Abyss
Grand Canyon National Park, The Abyss
Grand Canyon National Park, The Abyss
Grand Canyon National Park, The Abyss
Grand Canyon National Park, The Abyss
Grand Canyon National Park, The Abyss
Grand Canyon National Park, The Abyss
State
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
County
Cochise
Cochise
Cochise
Cochise
Cochise
Cochise
Cochise
Cochise
Cochise
Cochise
Cochise
Cochise
Cochise
Cochise
Cochise
Cochise
Cochise
Cochise
Cochise
Cochise
Cochise
Coconino
Coconino
Coconino
Coconino
Coconino
Coconino
Coconino
Coconino
Coconino
Coconino
Coconino
Name of Class I Area Located in
County
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Chiricahua National Monument
Grand Canyon National Park
Grand Canyon National Park
Grand Canyon National Park
Grand Canyon National Park
Grand Canyon National Park
Grand Canyon National Park
Grand Canyon National Park
Grand Canyon National Park
Grand Canyon National Park
Grand Canyon National Park
Grand Canyon National Park
5B-2
-------
Monitor ID
#
400580011
400580011
400580011
400580011
400580011
400580011
400580011
400580011
400580011
400580011
400580011
400580011
400580011
400580011
400580011
400580011
400580011
400580011
400580011
400580011
400580011
400580011
400580011
400580011
400580011
400700101
400700101
400700101
400700101
400700101
400700101
400700101
400700101
400700101
400700101
Years
(3-year
average)
1999-2001
2000-2002
2000-2002
2000-2002
2001-2003
2001-2003
2001-2003
2002-2004
2002-2004
2002-2004
2003-2005
2003-2005
2003-2005
2004-2006
2004-2006
2004-2006
2005-2007
2005-2007
2005-2007
2006-2008
2006-2008
2007-2009
2010-2012
2010-2012
2010-2012
2007-2009
2007-2009
2007-2009
2008-2010
2008-2010
2008-2010
2009-2011
2009-2011
2009-2011
2010-2012
Year
(annual)
2001
2000
2001
2002
2001
2002
2003
2002
2003
2004
2003
2004
2005
2004
2005
2006
2005
2006
2007
2006
2007
2007
2010
2011
2012
2007
2008
2009
2008
2009
2010
2009
2010
2011
2010
Max 8-hour
0>pb)
(3-year
average)
72
73
73
73
74
74
74
74
74
74
74
74
74
73
73
73
72
72
72
70
70
68
72
72
72
75
75
75
73
73
73
72
72
72
74
Max 8-hour
(Ppb)
(annual)
70
71
70
79
70
79
73
79
73
72
73
72
79
72
79
70
79
70
69
70
69
69
69
74
73
76
78
72
78
72
70
72
70
76
70
W126
(3-year
average)
17.64
19.47
19.47
19.47
21.79
21.79
21.79
22.29
22.29
22.29
19.98
19.98
19.98
19.39
19.39
19.39
20.24
20.24
20.24
19.29
19.29
15.38
17.90
17.90
17.90
22.45
22.45
22.45
20.29
20.29
20.29
17.90
17.90
17.90
21.41
W126
(annual)
14.91
16.74
14.91
26.78
14.91
26.78
23.70
26.78
23.70
16.41
23.70
16.41
19.84
16.41
19.84
21.92
19.84
21.92
18.95
21.92
18.95
18.95
14.89
18.45
20.34
24.95
27.52
14.89
27.52
14.89
18.45
14.89
18.45
20.35
18.45
Monitor Site Name
Grand Canyon National Park, The Abyss
Grand Canyon National Park, The Abyss
Grand Canyon National Park, The Abyss
Grand Canyon National Park, The Abyss
Grand Canyon National Park, The Abyss
Grand Canyon National Park, The Abyss
Grand Canyon National Park, The Abyss
Grand Canyon National Park, The Abyss
Grand Canyon National Park, The Abyss
Grand Canyon National Park, The Abyss
Grand Canyon National Park, The Abyss
Grand Canyon National Park, The Abyss
Grand Canyon National Park, The Abyss
Grand Canyon National Park, The Abyss
Grand Canyon National Park, The Abyss
Grand Canyon National Park, The Abyss
Grand Canyon National Park, The Abyss
Grand Canyon National Park, The Abyss
Grand Canyon National Park, The Abyss
Grand Canyon National Park, The Abyss
Grand Canyon National Park, The Abyss
Grand Canyon National Park, The Abyss
Grand Canyon National Park, The Abyss
Grand Canyon National Park, The Abyss
Grand Canyon National Park, The Abyss
Tonto NM
Tonto NM
Tonto NM
Tonto NM
Tonto NM
Tonto NM
Tonto NM
Tonto NM
Tonto NM
Tonto NM
State
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
County
Coconino
Coconino
Coconino
Coconino
Coconino
Coconino
Coconino
Coconino
Coconino
Coconino
Coconino
Coconino
Coconino
Coconino
Coconino
Coconino
Coconino
Coconino
Coconino
Coconino
Coconino
Coconino
Coconino
Coconino
Coconino
Gila
Gila
Gila
Gila
Gila
Gila
Gila
Gila
Gila
Gila
Name of Class I Area Located in
County
Grand Canyon National Park
Grand Canyon National Park
Grand Canyon National Park
Grand Canyon National Park
Grand Canyon National Park
Grand Canyon National Park
Grand Canyon National Park
Grand Canyon National Park
Grand Canyon National Park
Grand Canyon National Park
Grand Canyon National Park
Grand Canyon National Park
Grand Canyon National Park
Grand Canyon National Park
Grand Canyon National Park
Grand Canyon National Park
Grand Canyon National Park
Grand Canyon National Park
Grand Canyon National Park
Grand Canyon National Park
Grand Canyon National Park
Grand Canyon National Park
Grand Canyon National Park
Grand Canyon National Park
Grand Canyon National Park
Sierra Ancha Wilderness Area
Sierra Ancha Wilderness Area
Sierra Ancha Wilderness Area
Sierra Ancha Wilderness Area
Sierra Ancha Wilderness Area
Sierra Ancha Wilderness Area
Sierra Ancha Wilderness Area
Sierra Ancha Wilderness Area
Sierra Ancha Wilderness Area
Sierra Ancha Wilderness Area
5B-3
-------
Monitor ID
#
400700101
400700101
401320051
401340081
401397061
401701191
401701191
401701191
401900211
401910181
401910281
401900211
401900211
401900211
401900211
401900211
401900211
401900211
401900211
401900211
401900211
401900211
401900211
401900211
401910111
401910181
402130011
402180011
402180011
402180011
402180011
402180011
402180011
402180011
402180011
Years
(3-year
average)
2010-2012
2010-2012
2007-2009
2007-2009
2007-2009
2010-2012
2010-2012
2010-2012
1998-2000
1998-2000
1998-2000
2001-2003
2001-2003
2006-2008
2006-2008
2007-2009
2007-2009
2007-2009
2008-2010
2008-2010
2008-2010
2010-2012
2010-2012
2010-2012
2001-2003
2006-2008
2007-2009
2007-2009
2007-2009
2008-2010
2008-2010
2008-2010
2009-2011
2009-2011
2009-2011
Year
(annual)
2011
2012
2009
2008
2007
2010
2011
2012
1998
1999
2000
2002
2003
2007
2008
2007
2008
2009
2008
2009
2010
2010
2011
2012
2001
2006
2007
2008
2009
2008
2009
2010
2009
2010
2011
Max 8-hour
0>pb)
(3-year
average)
74
74
75
75
75
70
70
70
73
73
73
73
73
74
74
71
71
71
69
69
69
71
71
71
73
74
75
75
75
74
74
74
73
73
73
Max 8-hour
(Ppb)
(annual)
76
78
70
78
79
68
69
73
77
73
77
77
78
73
74
73
74
67
74
67
68
68
75
71
69
76
77
80
70
80
70
72
70
72
78
W126
(3-year
average)
21.41
21.41
22.48
22.48
22.48
15.79
15.79
15.79
15.55
15.55
15.55
15.53
15.53
18.98
18.98
16.10
16.10
16.10
15.47
15.47
15.47
16.84
16.84
16.84
15.53
18.98
22.52
22.52
22.52
20.87
20.87
20.87
18.75
18.75
18.75
W126
(annual)
20.35
25.44
14.51
27.49
28.65
12.96
15.16
19.26
18.60
16.53
15.52
16.01
23.14
17.24
20.01
17.24
20.01
11.04
20.01
11.04
15.36
15.36
17.36
17.79
12.73
21.54
24.59
29.02
14.81
29.02
14.81
18.79
14.81
18.79
22.66
Monitor Site Name
Tonto NM
Tonto NM
Rio Verde
Rio Verde
Rio Verde
Petrified Forest National Park, South Entrance
Petrified Forest National Park, South Entrance
Petrified Forest National Park, South Entrance
22nd & Cray croft
22nd & Craycroft
22nd & Craycroft
Saguaro Park
Saguaro Park
Saguaro Park
Saguaro Park
Saguaro Park
Saguaro Park
Saguaro Park
Saguaro Park
Saguaro Park
Saguaro Park
Saguaro Park
Saguaro Park
Saguaro Park
Saguaro Park
Saguaro Park
Queen Valley
Queen Valley
Queen Valley
Queen Valley
Queen Valley
Queen Valley
Queen Valley
Queen Valley
Queen Valley
State
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
County
Gila
Gila
Maricopa
Maricopa
Maricopa
Navajo
Navajo
Navajo
Pima
Pima
Pima
Pima
Pima
Pima
Pima
Pima
Pima
Pima
Pima
Pima
Pima
Pima
Pima
Pima
Pima
Pima
Final
Final
Final
Final
Final
Final
Final
Final
Final
Name of Class I Area Located in
County
Sierra Ancha Wilderness Area
Sierra Ancha Wilderness Area
Superstition Wilderness Area
Superstition Wilderness Area
Superstition Wilderness Area
Petrified Forest National Park
Petrified Forest National Park
Petrified Forest National Park
Saguaro National Park
Saguaro National Park
Saguaro National Park
Saguaro National Park
Saguaro National Park
Saguaro National Park
Saguaro National Park
Saguaro National Park
Saguaro National Park
Saguaro National Park
Saguaro National Park
Saguaro National Park
Saguaro National Park
Saguaro National Park
Saguaro National Park
Saguaro National Park
Saguaro National Park
Saguaro National Park
Superstition Wilderness Area
Superstition Wilderness Area
Superstition Wilderness Area
Superstition Wilderness Area
Superstition Wilderness Area
Superstition Wilderness Area
Superstition Wilderness Area
Superstition Wilderness Area
Superstition Wilderness Area
5B-4
-------
Monitor ID
#
600500021
600500021
600500021
602701011
602701011
602701011
602701011
602701011
602701011
602701011
602701011
602701011
606900031
606900031
606900031
608900071
608900091
608930031
610900051
610900051
610900051
610900051
610900051
610900051
801300111
801300111
801300111
801300111
Years
(3-year
average)
2010-2012
2010-2012
2010-2012
2008-2010
2008-2010
2008-2010
2009-2011
2009-2011
2009-2011
2010-2012
2010-2012
2010-2012
2005-2007
2005-2007
2005-2007
2008-2010
2008-2010
2008-2010
2009-2011
2009-2011
2009-2011
2010-2012
2010-2012
2010-2012
2000-2002
2000-2002
2000-2002
2003-2005
Year
(annual)
2010
2011
2012
2008
2009
2010
2009
2010
2011
2010
2011
2012
2005
2006
2007
2009
2010
2008
2009
2010
2011
2010
2011
2012
2000
2001
2002
2003
Max 8-hour
0>pb)
(3-year
average)
74
74
74
72
72
72
71
71
71
72
72
72
74
74
74
75
75
75
74
74
74
73
73
73
73
73
73
75
Max 8-hour
(Ppb)
(annual)
75
70
78
77
70
69
70
69
75
69
75
73
71
78
75
74
74
83
77
72
74
72
74
75
72
71
78
82
W126
(3-year
average)
17.68
17.68
17.68
17.19
17.19
17.19
16.54
16.54
16.54
18.69
18.69
18.69
15.18
15.18
15.18
15.31
15.31
15.31
20.72
20.72
20.72
20.84
20.84
20.84
15.11
15.11
15.11
16.61
W126
(annual)
15.56
14.87
22.61
25.85
15.55
10.16
15.55
10.16
23.92
10.16
23.92
22.00
13.11
17.44
14.99
13.66
15.33
18.72
21.80
20.58
19.78
20.58
19.78
22.14
14.06
13.18
18.09
23.91
Monitor Site Name
201 Clinton Road, Jackson
201 Clinton Road, Jackson
201 Clinton Road, Jackson
Death Valley National Monument Near Nevares
Springs Access
Death Valley National Monument Near Nevares
Springs Access
Death Valley National Monument Near Nevares
Springs Access
Death Valley National Monument Near Nevares
Springs Access
Death Valley National Monument Near Nevares
Springs Access
Death Valley National Monument Near Nevares
Springs Access
Death Valley National Monument Near Nevares
Springs Access
Death Valley National Monument Near Nevares
Springs Access
Death Valley National Monument Near Nevares
Springs Access
Pinnacles National Monument, SW of East
Entrance Station
Pinnacles National Monument, SW of East
Entrance Station
Pinnacles National Monument, SW of East
Entrance Station
Anderson - North Street
Anderson - North Street
Anderson - North Street
251 S Barretta, Sonora, CA 95370
251 S Barretta, Sonora, CA 95370
251 S Barretta, Sonora, CA 95370
251 S Barretta, Sonora, CA 95370
251 S Barretta, Sonora, CA 95370
251 S Barretta, Sonora, CA 95370
South Boulder Creek
South Boulder Creek
South Boulder Creek
South Boulder Creek
State
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CO
CO
CO
CO
County
Amador
Amador
Amador
Inyo
Inyo
Inyo
Inyo
Inyo
Inyo
Inyo
Inyo
Inyo
San Benito
San Benito
San Benito
Shasta
Shasta
Shasta
Tuolumne
Tuolumne
Tuolumne
Tuolumne
Tuolumne
Tuolumne
Boulder
Boulder
Boulder
Boulder
Name of Class I Area Located in
County
Mokelumne Wilderness Area
Mokelumne Wilderness Area
Mokelumne Wilderness Area
John Muir Wilderness Area
John Muir Wilderness Area
John Muir Wilderness Area
John Muir Wilderness Area
John Muir Wilderness Area
John Muir Wilderness Area
John Muir Wilderness Area
John Muir Wilderness Area
John Muir Wilderness Area
Pinnacles National Monument
Pinnacles National Monument
Pinnacles National Monument
Lassen Volcanic National Park
Lassen Volcanic National Park
Lassen Volcanic National Park
Yosemite National Park
Yosemite National Park
Yosemite National Park
Yosemite National Park
Yosemite National Park
Yosemite National Park
Rocky Mountain National Park
Rocky Mountain National Park
Rocky Mountain National Park
Rocky Mountain National Park
5B-5
-------
Monitor ID
#
801300111
801300111
801300111
801300111
801300111
801300111
801300111
801300111
801300111
801300111
801300111
801300111
801300111
801300111
805199911
805199911
805199911
805199911
805199911
805199911
805199911
805199911
805199911
805199911
805199911
805199911
805199911
805199911
805199911
805199911
805199911
805199911
805199911
805199911
805199911
Years
(3-year
average)
2003-2005
2003-2005
2004-2006
2004-2006
2004-2006
2008-2010
2008-2010
2008-2010
2009-2011
2009-2011
2009-2011
2010-2012
2010-2012
2010-2012
1998-2000
1998-2000
1998-2000
1999-2001
1999-2001
1999-2001
2000-2002
2000-2002
2000-2002
2001-2003
2001-2003
2001-2003
2002-2004
2002-2004
2002-2004
2003-2005
2003-2005
2003-2005
2004-2006
2004-2006
2004-2006
Year
(annual)
2004
2005
2004
2005
2006
2008
2009
2010
2009
2010
2011
2010
2011
2012
1998
1999
2000
1999
2000
2001
2000
2001
2002
2001
2002
2003
2002
2003
2004
2003
2004
2005
2004
2005
2006
Max 8-hour
0>pb)
(3-year
average)
75
75
75
75
75
73
73
73
73
73
73
74
74
74
73
73
73
73
73
73
71
71
71
71
71
71
70
70
70
69
69
69
68
68
68
Max 8-hour
(Ppb)
(annual)
68
76
68
76
82
76
73
72
73
72
76
72
76
76
71
77
73
77
73
70
73
70
71
70
71
73
71
73
67
73
67
69
67
69
70
W126
(3-year
average)
16.61
16.61
17.01
17.01
17.01
16.11
16.11
16.11
16.13
16.13
16.13
19.34
19.34
19.34
20.18
20.18
20.18
18.40
18.40
18.40
18.01
18.01
18.01
18.90
18.90
18.90
17.95
17.95
17.95
15.82
15.82
15.82
15.60
15.60
15.60
W126
(annual)
9.57
16.35
9.57
16.35
25.11
20.77
12.57
14.98
12.57
14.98
20.82
14.98
20.82
22.20
21.13
23.98
15.43
23.98
15.43
15.80
15.43
15.80
22.82
15.80
22.82
18.07
22.82
18.07
12.96
18.07
12.96
16.42
12.96
16.42
17.40
Monitor Site Name
South Boulder Creek
South Boulder Creek
South Boulder Creek
South Boulder Creek
South Boulder Creek
South Boulder Creek
South Boulder Creek
South Boulder Creek
South Boulder Creek
South Boulder Creek
South Boulder Creek
South Boulder Creek
South Boulder Creek
South Boulder Creek
Gothic
Gothic
Gothic
Gothic
Gothic
Gothic
Gothic
Gothic
Gothic
Gothic
Gothic
Gothic
Gothic
Gothic
Gothic
Gothic
Gothic
Gothic
Gothic
Gothic
Gothic
State
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
County
Boulder
Boulder
Boulder
Boulder
Boulder
Boulder
Boulder
Boulder
Boulder
Boulder
Boulder
Boulder
Boulder
Boulder
Gunnison
Gunnison
Gunnison
Gunnison
Gunnison
Gunnison
Gunnison
Gunnison
Gunnison
Gunnison
Gunnison
Gunnison
Gunnison
Gunnison
Gunnison
Gunnison
Gunnison
Gunnison
Gunnison
Gunnison
Gunnison
Name of Class I Area Located in
County
Rocky Mountain National Park
Rocky Mountain National Park
Rocky Mountain National Park
Rocky Mountain National Park
Rocky Mountain National Park
Rocky Mountain National Park
Rocky Mountain National Park
Rocky Mountain National Park
Rocky Mountain National Park
Rocky Mountain National Park
Rocky Mountain National Park
Rocky Mountain National Park
Rocky Mountain National Park
Rocky Mountain National Park
West Elk Wilderness Area
West Elk Wilderness Area
West Elk Wilderness Area
West Elk Wilderness Area
West Elk Wilderness Area
West Elk Wilderness Area
West Elk Wilderness Area
West Elk Wilderness Area
West Elk Wilderness Area
West Elk Wilderness Area
West Elk Wilderness Area
West Elk Wilderness Area
West Elk Wilderness Area
West Elk Wilderness Area
West Elk Wilderness Area
West Elk Wilderness Area
West Elk Wilderness Area
West Elk Wilderness Area
West Elk Wilderness Area
West Elk Wilderness Area
West Elk Wilderness Area
5B-6
-------
Monitor ID
#
805199911
805199911
805199911
806710041
806710041
806710041
806710041
806710041
806710041
806710041
806710041
806710041
806710041
806710041
806710041
806710041
806710041
806710041
806900071
806900111
806900111
806900071
806900071
806900071
806900071
806900071
806999911
808301011
808301011
808301011
808301011
808301011
Years
(3-year
average)
2005-2007
2005-2007
2005-2007
2005-2007
2005-2007
2005-2007
2006-2008
2006-2008
2006-2008
2008-2010
2008-2010
2008-2010
2009-2011
2009-2011
2009-2011
2010-2012
2010-2012
2010-2012
2008-2010
2008-2010
2008-2010
1999-2001
1999-2001
1999-2001
2004-2006
2004-2006
2004-2006
1998-2000
1998-2000
1998-2000
1999-2001
1999-2001
Year
(annual)
2005
2006
2007
2005
2006
2007
2006
2007
2008
2008
2009
2010
2009
2010
2011
2010
2011
2012
2010
2008
2009
1999
2000
2001
2004
2006
2005
1998
1999
2000
1999
2000
Max 8-hour
0>pb)
(3-year
average)
68
68
68
72
72
72
70
70
70
71
71
71
74
74
74
73
73
73
74
74
74
74
74
74
74
74
74
70
70
70
69
69
Max 8-hour
(Ppb)
(annual)
69
70
65
75
74
69
74
69
69
69
71
74
71
74
77
74
77
69
77
76
73
74
78
70
73
76
78
68
69
73
69
73
W126
(3-year
average)
16.38
16.38
16.38
18.78
18.78
18.78
18.10
18.10
18.10
15.07
15.07
15.07
16.80
16.80
16.80
19.16
19.16
19.16
18.31
18.31
18.31
15.05
15.05
15.05
15.57
15.57
15.57
16.37
16.37
16.37
15.66
15.66
W126
(annual)
16.42
17.40
15.31
17.93
20.82
17.58
20.82
17.58
15.91
15.91
10.94
18.36
10.94
18.36
21.09
18.36
21.09
18.02
19.12
21.63
14.17
11.16
25.82
8.16
16.23
18.53
16.20
12.90
14.17
22.04
14.17
22.04
Monitor Site Name
Gothic
Gothic
Gothic
Fort Collins - West
Fort Collins - West
Fort Collins - West
Rocky Mountain National Park, Long's Peak
Rocky Mountain National Park, Long's Peak
Rocky Mountain National Park, Long's Peak
Rocky Mountain National Park, Long's Peak
Rocky Mountain National Park, Long's Peak
Rocky Mountain National Park, Long's Peak
Mesa Verde National Park, Resource
Management Area
Mesa Verde National Park, Resource
Management Area
Mesa Verde National Park, Resource
Management Area
Mesa Verde National Park, Resource
Management Area
Mesa Verde National Park, Resource
Management Area
State
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
County
Gunnison
Gunnison
Gunnison
La Plata
La Plata
La Plata
La Plata
La Plata
La Plata
La Plata
La Plata
La Plata
La Plata
La Plata
La Plata
La Plata
La Plata
La Plata
Larimer
Larimer
Larimer
Larimer
Larimer
Larimer
Larimer
Larimer
Larimer
Montezuma
Montezuma
Montezuma
Montezuma
Montezuma
Name of Class I Area Located in
County
West Elk Wilderness Area
West Elk Wilderness Area
West Elk Wilderness Area
Weminuche Wilderness Area
Weminuche Wilderness Area
Weminuche Wilderness Area
Weminuche Wilderness Area
Weminuche Wilderness Area
Weminuche Wilderness Area
Weminuche Wilderness Area
Weminuche Wilderness Area
Weminuche Wilderness Area
Weminuche Wilderness Area
Weminuche Wilderness Area
Weminuche Wilderness Area
Weminuche Wilderness Area
Weminuche Wilderness Area
Weminuche Wilderness Area
Rocky Mountain National Park
Rocky Mountain National Park
Rocky Mountain National Park
Rocky Mountain National Park
Rocky Mountain National Park
Rocky Mountain National Park
Rocky Mountain National Park
Rocky Mountain National Park
Rocky Mountain National Park
Mesa Verde National Park
Mesa Verde National Park
Mesa Verde National Park
Mesa Verde National Park
Mesa Verde National Park
5B-7
-------
Monitor ID
#
808301011
808301011
808301011
808301011
808301011
808301011
808301011
808301011
808301011
808301011
808301011
808301011
808301011
808301011
808301011
808301011
808301011
808301011
808301011
808301011
808301011
808301011
Years
(3-year
average)
1999-2001
2000-2002
2000-2002
2000-2002
2001-2003
2001-2003
2001-2003
2002-2004
2002-2004
2002-2004
2003-2005
2003-2005
2003-2005
2004-2006
2004-2006
2004-2006
2005-2007
2005-2007
2005-2007
2006-2008
2006-2008
2006-2008
Year
(annual)
2001
2000
2001
2002
2001
2002
2003
2002
2003
2004
2003
2004
2005
2004
2005
2006
2005
2006
2007
2006
2007
2008
Max 8-hour
0>pb)
(3-year
average)
69
69
69
69
67
67
67
68
68
68
70
70
70
73
73
73
73
73
73
71
71
71
Max 8-hour
(Ppb)
(annual)
65
73
65
70
65
70
67
70
67
69
67
69
76
69
76
74
76
74
70
74
70
69
W126
(3-year
average)
15.66
17.51
17.51
17.51
16.00
16.00
16.00
16.34
16.34
16.34
16.96
16.96
16.96
19.02
19.02
19.02
21.00
21.00
21.00
18.36
18.36
18.36
W126
(annual)
10.77
22.04
10.77
19.72
10.77
19.72
17.50
19.72
17.50
11.79
17.50
11.79
21.59
11.79
21.59
23.68
21.59
23.68
17.73
23.68
17.73
13.67
Monitor Site Name
Mesa Verde National Park, Resource
Management Area
Mesa Verde National Park, Resource
Management Area
Mesa Verde National Park, Resource
Management Area
Mesa Verde National Park, Resource
Management Area
Mesa Verde National Park, Resource
Management Area
Mesa Verde National Park, Resource
Management Area
Mesa Verde National Park, Resource
Management Area
Mesa Verde National Park, Resource
Management Area
Mesa Verde National Park, Resource
Management Area
Mesa Verde National Park, Resource
Management Area
Mesa Verde National Park, Resource
Management Area
Mesa Verde National Park, Resource
Management Area
Mesa Verde National Park, Resource
Management Area
Mesa Verde National Park, Resource
Management Area
Mesa Verde National Park, Resource
Management Area
Mesa Verde National Park, Resource
Management Area
Mesa Verde National Park, Resource
Management Area
Mesa Verde National Park, Resource
Management Area
Mesa Verde National Park, Resource
Management Area
Mesa Verde National Park, Resource
Management Area
Mesa Verde National Park, Resource
Management Area
Mesa Verde National Park, Resource
Management Area
State
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
County
Montezuma
Montezuma
Montezuma
Montezuma
Montezuma
Montezuma
Montezuma
Montezuma
Montezuma
Montezuma
Montezuma
Montezuma
Montezuma
Montezuma
Montezuma
Montezuma
Montezuma
Montezuma
Montezuma
Montezuma
Montezuma
Montezuma
Name of Class I Area Located in
County
Mesa Verde National Park
Mesa Verde National Park
Mesa Verde National Park
Mesa Verde National Park
Mesa Verde National Park
Mesa Verde National Park
Mesa Verde National Park
Mesa Verde National Park
Mesa Verde National Park
Mesa Verde National Park
Mesa Verde National Park
Mesa Verde National Park
Mesa Verde National Park
Mesa Verde National Park
Mesa Verde National Park
Mesa Verde National Park
Mesa Verde National Park
Mesa Verde National Park
Mesa Verde National Park
Mesa Verde National Park
Mesa Verde National Park
Mesa Verde National Park
5B-8
-------
Monitor ID
#
808301011
808301011
808301011
2106105011
2106105011
2106105011
3501510051
3501510051
3501510051
3501510051
3501510051
3501510051
3501510051
3501510051
3501510051
3504310011
3504310011
3504310011
3504310011
3504310031
3504310031
3504310011
3504310011
3504310011
Years
(3-year
average)
2007-2009
2007-2009
2007-2009
2006-2008
2006-2008
2006-2008
2004-2006
2004-2006
2004-2006
2005-2007
2005-2007
2005-2007
2006-2008
2006-2008
2006-2008
2000-2002
2000-2002
2001-2003
2001-2003
2000-2002
2001-2003
1999-2001
2002-2004
2002-2004
Year
(annual)
2007
2008
2009
2006
2007
2008
2004
2005
2006
2005
2006
2007
2006
2007
2008
2001
2002
2001
2002
2000
2003
2001
2002
2004
Max 8-hour
0>pb)
(3-year
average)
69
69
69
74
74
74
69
69
69
69
69
69
69
69
69
72
72
71
71
72
71
72
74
74
Max 8-hour
(Ppb)
(annual)
70
69
69
71
82
70
65
67
76
67
76
66
76
66
67
69
74
69
74
75
76
69
74
71
W126
(3-year
average)
15.58
15.58
15.58
15.99
15.99
15.99
15.30
15.30
15.30
15.33
15.33
15.33
15.07
15.07
15.07
17.19
17.19
17.37
17.37
17.19
17.37
17.87
20.86
20.86
W126
(annual)
17.73
13.67
15.35
12.55
22.58
12.83
8.65
10.55
26.71
10.55
26.71
8.72
26.71
8.72
9.78
12.17
19.62
12.17
19.62
23.51
25.38
12.17
19.62
17.73
Monitor Site Name
Mesa Verde National Park, Resource
Management Area
Mesa Verde National Park, Resource
Management Area
Mesa Verde National Park, Resource
Management Area
Mammoth Cave National Park, Houchin
Meadow
Mammoth Cave National Park, Houchin
Meadow
Mammoth Cave National Park, Houchin
Meadow
5ZR on BLM Land bordering residential area
outside Carlsbad
5ZR on BLM Land bordering residential area
outside Carlsbad
5ZR on BLM Land bordering residential area
outside Carlsbad
5ZR on BLM Land bordering residential area
outside Carlsba
5ZR on BLM Land bordering residential area
outside Carlsbad
5ZR on BLM Land bordering residential area
outside Carlsbad
5ZR on BLM Land bordering residential area
outside Carlsbad
5ZR on BLM Land bordering residential area
outside Carlsba
5ZR on BLM Land bordering residential area
outside Carlsbad
2ZR Site moved from Rio Rancho City Hall to
2ZR Site moved from Rio Rancho City Hall to
senior center
2ZR Site moved from Rio Rancho City Hall to
senior center
State
CO
CO
CO
KY
KY
KY
NM
NM
NM
NM
NM
NM
NM
NM
NM
NM
NM
NM
NM
NM
NM
NM
NM
NM
County
Montezuma
Montezuma
Montezuma
Edmonson
Edmonson
Edmonson
Eddy
Eddy
Eddy
Eddy
Eddy
Eddy
Eddy
Eddy
Eddy
Sandoval
Sandoval
Sandoval
Sandoval
Sandoval
Sandoval
Sandoval
Sandoval
Sandoval
Name of Class I Area Located in
County
Mesa Verde National Park
Mesa Verde National Park
Mesa Verde National Park
Mammoth Cave National Park
Mammoth Cave National Park
Mammoth Cave National Park
Carlsbad Caverns National Park
Carlsbad Caverns National Park
Carlsbad Caverns National Park
Carlsbad Caverns National Park
Carlsbad Caverns National Park
Carlsbad Caverns National Park
Carlsbad Caverns National Park
Carlsbad Caverns National Park
Carlsbad Caverns National Park
Bandelier Wilderness Area
Bandelier Wilderness Area
Bandelier Wilderness Area
Bandelier Wilderness Area
Bandelier Wilderness Area
Bandelier Wilderness Area
Bandelier Wilderness Area
Bandelier Wilderness Area
Bandelier Wilderness Area
5B-9
-------
Monitor ID
#
3504310011
3504310011
3504310031
3504310031
3504310031
3504310031
3504310031
3504310031
3504310031
3504310031
3504310031
3504310031
3504390041
3504390041
3504390041
4603301323
4603301323
4603301323
4903701011
4903701011
4903701011
4903701011
4903701011
4903701011
4903701011
4903701011
Years
(3-year
average)
2003-2005
2004-2006
1999-2001
1999-2001
2002-2004
2003-2005
2003-2005
2004-2006
2005-2007
2005-2007
2006-2008
2006-2008
2004-2006
2005-2007
2006-2008
2005-2007
2005-2007
2005-2007
1998-2000
1998-2000
1998-2000
2001-2003
2001-2003
2001-2003
2002-2004
2002-2004
Year
(annual)
2004
2004
1999
2000
2003
2003
2005
2005
2005
2007
2007
2008
2006
2006
2006
2005
2006
2007
1998
1999
2000
2001
2002
2003
2002
2003
Max 8-hour
0>pb)
(3-year
average)
74
73
72
72
74
74
74
73
73
73
70
70
73
73
70
70
70
70
73
73
73
70
70
70
72
72
Max 8-hour
(Ppb)
(annual)
71
71
76
75
76
76
75
75
75
71
71
65
72
72
72
70
73
69
71
73
76
66
72
74
72
74
W126
(3-year
average)
20.08
17.75
17.87
17.87
20.86
20.08
20.08
17.75
17.50
17.50
15.87
15.87
17.75
17.50
15.87
15.49
15.49
15.49
19.80
19.80
19.80
18.94
18.94
18.94
20.50
20.50
W126
(annual)
17.73
17.73
18.14
23.51
25.38
25.38
17.03
17.03
17.03
17.05
17.05
12.15
19.26
19.26
19.26
13.56
20.63
12.29
19.78
20.25
19.36
9.91
22.12
24.80
22.12
24.80
Monitor Site Name
2ZR Site moved from Rio Rancho City Hall to
senior center
2ZR Site moved from Rio Rancho City Hall to
2ZR Site moved from Rio Rancho City Hall to
senior center
2ZR Site moved from Rio Rancho City Hall to
senior center
2ZR Site moved from Rio Rancho City Hall to
senior center
2ZR Site moved from Rio Rancho City Hall to
senior center
2ZR Site moved from Rio Rancho City Hall to
senior center
2ZR Site moved from Rio Rancho City Hall to
senior center
2ZR Site moved from Rio Rancho City Hall to
senior center
2ZR Site moved from Rio Rancho City Hall to
senior center
2ZR Site moved from Rio Rancho City Hall to
senior center
2ZR Site moved from Rio Rancho City Hall to
senior center
2ZR Site moved from Rio Rancho City Hall to
senior center
2ZR Site moved from Rio Rancho City Hall to
senior center
2ZR Site moved from Rio Rancho City Hall to
senior center
Wind Cave National Park, Visitor Center
Wind Cave National Park, Visitor Center
Wind Cave National Park, Visitor Center
Canyonlands National Park, Island in the Sky
Canyonlands National Park, Island in the Sky
Canyonlands National Park, Island in the Sky
Canyonlands National Park, Island in the Sky
Canyonlands National Park, Island in the Sky
Canyonlands National Park, Island in the Sky
Canyonlands National Park, Island in the Sky
Canyonlands National Park, Island in the Sky
State
NM
NM
NM
NM
NM
NM
NM
NM
NM
NM
NM
NM
NM
NM
NM
SD
SD
SD
UT
UT
UT
UT
UT
UT
UT
UT
County
Sandoval
Sandoval
Sandoval
Sandoval
Sandoval
Sandoval
Sandoval
Sandoval
Sandoval
Sandoval
Sandoval
Sandoval
Sandoval
Sandoval
Sandoval
Custer
Custer
Custer
San Juan
San Juan
San Juan
San Juan
San Juan
San Juan
San Juan
San Juan
Name of Class I Area Located in
County
Bandelier Wilderness Area
Bandelier Wilderness Area
Bandelier Wilderness Area
Bandelier Wilderness Area
Bandelier Wilderness Area
Bandelier Wilderness Area
Bandelier Wilderness Area
Bandelier Wilderness Area
Bandelier Wilderness Area
Bandelier Wilderness Area
Bandelier Wilderness Area
Bandelier Wilderness Area
Bandelier Wilderness Area
Bandelier Wilderness Area
Bandelier Wilderness Area
Wind Cave National Park
Wind Cave National Park
Wind Cave National Park
Canyonlands National Park
Canyonlands National Park
Canyonlands National Park
Canyonlands National Park
Canyonlands National Park
Canyonlands National Park
Canyonlands National Park
Canyonlands National Park
5B-10
-------
Monitor ID
#
4903701011
4903701011
4903701011
4903701011
4903701011
4903701011
4903701011
4903701011
4903701011
4903701011
4903701011
4903701011
4903701011
4903701011
4903701011
4903701011
4903701011
4903701011
4903701011
4905301301
4905301301
4905301301
4905301301
4905301301
4905301301
4905301301
4905301301
4905301301
4905301301
4905301301
4905301301
4905301301
4905301301
4905301301
5603599911
Years
(3-year
average)
2002-2004
2003-2005
2003-2005
2003-2005
2004-2006
2004-2006
2004-2006
2005-2007
2005-2007
2005-2007
2006-2008
2006-2008
2006-2008
2007-2009
2007-2009
2007-2009
2010-2012
2010-2012
2010-2012
2006-2008
2006-2008
2006-2008
2007-2009
2007-2009
2007-2009
2008-2010
2008-2010
2008-2010
2009-2011
2009-2011
2009-2011
2010-2012
2010-2012
2010-2012
1998-2000
Year
(annual)
2004
2003
2004
2005
2004
2005
2006
2005
2006
2007
2006
2007
2008
2007
2008
2009
2010
2011
2012
2006
2007
2008
2007
2008
2009
2008
2009
2010
2009
2010
2011
2010
2011
2012
1998
Max 8-hour
0>pb)
(3-year
average)
72
71
71
71
70
70
70
70
70
70
71
71
71
70
70
70
69
69
69
71
71
71
70
70
70
70
70
70
70
70
70
73
73
73
72
Max 8-hour
(Ppb)
(annual)
72
74
72
69
72
69
70
69
70
72
70
72
71
72
71
68
68
69
72
72
71
72
71
72
68
72
68
72
68
72
72
72
72
75
71
W126
(3-year
average)
20.50
18.59
18.59
18.59
16.59
16.59
16.59
17.66
17.66
17.66
18.10
18.10
18.10
16.07
16.07
16.07
15.01
15.01
15.01
21.12
21.12
21.12
18.01
18.01
18.01
18.46
18.46
18.46
17.85
17.85
17.85
20.34
20.34
20.34
17.25
W126
(annual)
14.57
24.80
14.57
16.40
14.57
16.40
18.80
16.40
18.80
17.78
18.80
17.78
17.71
17.78
17.71
12.72
13.87
14.23
16.93
24.29
19.37
19.69
19.37
19.69
14.96
19.69
14.96
20.73
14.96
20.73
17.86
20.73
17.86
22.42
16.02
Monitor Site Name
Canyonlands National Park, Island in the Sky
Canyonlands National Park, Island in the Sky
Canyonlands National Park, Island in the Sky
Canyonlands National Park, Island in the Sky
Canyonlands National Park, Island in the Sky
Canyonlands National Park, Island in the Sky
Canyonlands National Park, Island in the Sky
Canyonlands National Park, Island in the Sky
Canyonlands National Park, Island in the Sky
Canyonlands National Park, Island in the Sky
Canyonlands National Park, Island in the Sky
Canyonlands National Park, Island in the Sky
Canyonlands National Park, Island in the Sky
Canyonlands National Park, Island in the Sky
Canyonlands National Park, Island in the Sky
Canyonlands National Park, Island in the Sky
Canyonlands National Park, Island in the Sky
Canyonlands National Park, Island in the Sky
Canyonlands National Park, Island in the Sky
Zion National Park, Dalton's Wash
Zion National Park, Dalton's Wash
Zion National Park, Dalton's Wash
Zion National Park, Dalton's Wash
Zion National Park, Dalton's Wash
Zion National Park, Dalton's Wash
Zion National Park, Dalton's Wash
Zion National Park, Dalton's Wash
Zion National Park, Dalton's Wash
Zion National Park, Dalton's Wash
Zion National Park, Dalton's Wash
Zion National Park, Dalton's Wash
Zion National Park, Dalton's Wash
Zion National Park, Dalton's Wash
Zion National Park, Dalton's Wash
Pinedale
State
UT
UT
UT
UT
UT
UT
UT
UT
UT
UT
UT
UT
UT
UT
UT
UT
UT
UT
UT
UT
UT
UT
UT
UT
UT
UT
UT
UT
UT
UT
UT
UT
UT
UT
WY
County
San Juan
San Juan
San Juan
San Juan
San Juan
San Juan
San Juan
San Juan
San Juan
San Juan
San Juan
San Juan
San Juan
San Juan
San Juan
San Juan
San Juan
San Juan
San Juan
Washington
Washington
Washington
Washington
Washington
Washington
Washington
Washington
Washington
Washington
Washington
Washington
Washington
Washington
Washington
Sublette
Name of Class I Area Located in
County
Canyonlands National Park
Canyonlands National Park
Canyonlands National Park
Canyonlands National Park
Canyonlands National Park
Canyonlands National Park
Canyonlands National Park
Canyonlands National Park
Canyonlands National Park
Canyonlands National Park
Canyonlands National Park
Canyonlands National Park
Canyonlands National Park
Canyonlands National Park
Canyonlands National Park
Canyonlands National Park
Canyonlands National Park
Canyonlands National Park
Canyonlands National Park
Zion National Park
Zion National Park
Zion National Park
Zion National Park
Zion National Park
Zion National Park
Zion National Park
Zion National Park
Zion National Park
Zion National Park
Zion National Park
Zion National Park
Zion National Park
Zion National Park
Zion National Park
Bridger Wilderness Area
5B-11
-------
Monitor ID
#
5603599911
5603599911
5603599911
5603599911
5603599911
5603599911
5603599911
5603599911
5603599911
5603599911
5603599911
5603599911
5603599911
5603599911
Years
(3-year
average)
1998-2000
1998-2000
1999-2001
1999-2001
1999-2001
2000-2002
2000-2002
2000-2002
2001-2003
2001-2003
2001-2003
2002-2004
2002-2004
2002-2004
Year
(annual)
1999
2000
1999
2000
2001
2000
2001
2002
2001
2002
2003
2002
2003
2004
Max 8-hour
0>pb)
(3-year
average)
72
72
71
71
71
71
71
71
70
70
70
69
69
69
Max 8-hour
(Ppb)
(annual)
72
73
72
73
69
73
69
72
69
72
70
72
70
65
W126
(3-year
average)
17.25
17.25
16.68
16.68
16.68
17.46
17.46
17.46
16.63
16.63
16.63
15.16
15.16
15.16
W126
(annual)
16.88
18.86
16.88
18.86
14.31
18.86
14.31
19.21
14.31
19.21
16.36
19.21
16.36
9.93
Monitor Site Name
Pinedale
Pinedale
Pinedale
Pinedale
Pinedale
Pinedale
Pinedale
Pinedale
Pinedale
Pinedale
Pinedale
Pinedale
Pinedale
Pinedale
State
WY
WY
WY
WY
WY
WY
WY
WY
WY
WY
WY
WY
WY
WY
County
Sublette
Sublette
Sublette
Sublette
Sublette
Sublette
Sublette
Sublette
Sublette
Sublette
Sublette
Sublette
Sublette
Sublette
Name of Class I Area Located in
County
Bridger Wilderness Area
Bridger Wilderness Area
Bridger Wilderness Area
Bridger Wilderness Area
Bridger Wilderness Area
Bridger Wilderness Area
Bridger Wilderness Area
Bridger Wilderness Area
Bridger Wilderness Area
Bridger Wilderness Area
Bridger Wilderness Area
Bridger Wilderness Area
Bridger Wilderness Area
Bridger Wilderness Area
5B-12
-------
APPENDIX 5C: EXPANDED EVALUATION OF RELATIVE BIOMASS
AND YIELD LOSS
This appendix expands to range W126 index values evaluated for relative biomass and
yield loss. Specifically, Tables 5C-1 and 5C-2 below provide estimates of the relative loss for
trees and crops 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 median
of the composite functions is calculated for all 11 tree species excluding cottonwood. These
tables also provide estimates of the number of species for trees and crops respectively that would
be below various benchmarks (e.g., 2% biomass loss for trees) at various W126 index values.
Table 5C-3 provides an expansion of Table 6-1 to reflect each integer W126 index value
between 7 ppm-hrs and 23 ppm-hrs.
5C-1
-------
Table 5C-1 Relative Biomass Loss for Eleven Individual Tree Seedlings and Median at Various W126 Index Values
W126
30
29
28
27
26
25
24
23
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
7
Douglas
Fir
0.1%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
Loblolly
0.8%
0.7%
0.7%
0.7%
0.7%
0.6%
0.6%
0.6%
0.6%
0.5%
0.5%
0.5%
0.5%
0.4%
0.4%
0.4%
0.4%
0.3%
0.3%
0.3%
0.3%
0.2%
0.2%
0.2%
Virginia
Pine
1.7%
1.7%
1.6%
1.6%
1.5%
1.4%
1.4%
1.3%
1.3%
1.2%
1.2%
1.1%
1.0%
1.0%
0.9%
0.9%
0.8%
0.8%
0.7%
0.6%
0.6%
0.5%
0.5%
0.4%
Red
maple
3.8%
3.6%
3.5%
3.3%
3.1%
3.0%
2.8%
2.7%
2.5%
2.4%
2.2%
2.1%
1.9%
1.8%
1.6%
1.5%
1.4%
1.2%
1.1%
1.0%
0.9%
0.7%
0.6%
0.5%
Sugar
maple
28.1%
23.7%
19.9%
16.4%
13.4%
10.9%
8.7%
6.9%
5.3%
4.1%
3.1%
2.3%
1.7%
1.2%
0.9%
0.6%
0.4%
0.3%
0.2%
0.1%
0.1%
0.0%
0.0%
0.0%
Red
Alder
10.4%
10.0%
9.6%
9.2%
8.8%
8.4%
8.0%
7.6%
7.2%
6.8%
6.4%
6.0%
5.7%
5.3%
4.9%
4.5%
4.2%
3.8%
3.5%
3.1%
2.8%
2.4%
2.1%
1.8%
Ponderosa
Pine
12.8%
12.3%
11.8%
11.4%
10.9%
10.4%
10.0%
9.5%
9.0%
8.6%
8.1%
7.6%
7.2%
6.7%
6.3%
5.8%
5.4%
4.9%
4.5%
4.1%
3.6%
3.2%
2.8%
2.4%
Aspen
18.6%
17.9%
17.2%
16.5%
15.8%
15.2%
14.5%
13.8%
13.1%
12.4%
11.8%
11.1%
10.4%
9.8%
9.1%
8.4%
7.8%
7.1%
6.5%
5.9%
5.2%
4.6%
4.0%
3.4%
Tulip
Poplar
27.7%
26.1%
24.5%
23.0%
21.4%
19.9%
18.4%
17.0%
15.6%
14.3%
13.0%
11.8%
10.6%
9.4%
8.4%
7.4%
6.4%
5.5%
4.7%
3.9%
3.2%
2.6%
2.0%
1.5%
Eastern
White
Pine
25.2%
24.0%
22.8%
21.6%
20.5%
19.3%
18.2%
17.1%
15.9%
14.9%
13.8%
12.7%
11.7%
10.7%
9.7%
8.8%
7.9%
7.0%
6.2%
5.4%
4.6%
3.9%
3.2%
2.6%
Black
Cherry
53.8%
52.6%
51.4%
50.1%
48.8%
47.5%
46.2%
44.8%
43.3%
41.9%
40.3%
38.8%
37.2%
35.6%
33.9%
32.2%
30.4%
28.6%
26.7%
24.8%
22.9%
20.9%
18.8%
16.7%
Median
(11
species)
12.8%
12.3%
11.8%
11.4%
10.9%
10.4%
8.7%
7.6%
7.2%
6.8%
6.4%
6.0%
5.7%
5.3%
4.9%
4.5%
4.2%
3.8%
3.5%
3.1%
2.8%
2.4%
2.0%
1.5%
Number
of Species
<2%
3
3
3
3
3
3
3
3
3
3
3
3
5
5
5
5
5
5
5
5
5
5
5
7
Number
of Species
<5%
4
4
4
4
4
4
4
4
4
5
5
5
5
5
6
6
6
7
8
8
9
10
10
10
Number
of Species
< 10%
4
5
5
5
5
5
7
7
7
7
7
7
7
9
10
10
10
10
10
10
10
10
10
10
Number
of Species
< 15%
6
6
6
6
7
7
8
8
8
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
5C-2
-------
Table 5C-2 Relative Yield Loss for Ten Individual Crop Species and Median at Various W126 Index Values
W126
30
29
28
27
26
25
24
23
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
7
Barley
0.1%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
Lettuce
5.1%
4.4%
3.7%
3.1%
2.6%
2.1%
1.7%
1.4%
1.1%
0.9%
0.7%
0.6%
0.4%
0.3%
0.2%
0.2%
0.1%
0.1%
0.1%
0.0%
0.0%
0.0%
0.0%
0.0%
Field
Corn
2.9%
2.7%
2.4%
2.2%
1.9%
1.7%
1.5%
1.3%
1.2%
1.0%
0.9%
0.8%
0.7%
0.6%
0.5%
0.4%
0.3%
0.2%
0.2%
0.2%
0.1%
0.1%
0.1%
0.0%
Grain
Sorghum
2.3%
2.1%
2.0%
1.9%
1.7%
1.6%
1.5%
1.4%
1.3%
1.1%
1.0%
0.9%
0.8%
0.8%
0.7%
0.6%
0.5%
0.5%
0.4%
0.3%
0.3%
0.2%
0.2%
0.1%
Peanut
10.4%
9.7%
9.1%
8.6%
8.0%
7.4%
6.9%
6.4%
5.9%
5.4%
5.0%
4.5%
4.1%
3.7%
3.3%
2.9%
2.6%
2.2%
1.9%
1.6%
1.4%
1.1%
0.9%
0.7%
Cotton
16.3%
15.6%
14.9%
14.1%
13.4%
12.7%
12.0%
11.3%
10.6%
10.0%
9.3%
8.7%
8.0%
7.4%
6.8%
6.2%
5.6%
5.0%
4.5%
3.9%
3.4%
2.9%
2.5%
2.0%
Soybean
15.7%
15.0%
14.4%
13.7%
13.1%
12.5%
11.8%
11.2%
10.6%
10.0%
9.4%
8.8%
8.2%
7.6%
7.0%
6.4%
5.9%
5.3%
4.8%
4.3%
3.8%
3.3%
2.8%
2.3%
Winter
Wheat
22.5%
21.0%
19.5%
18.0%
16.6%
15.3%
14.0%
12.7%
11.5%
10.4%
9.3%
8.3%
7.3%
6.4%
5.6%
4.8%
4.1%
3.5%
2.9%
2.3%
1.9%
1.5%
1.1%
0.8%
Potato
20.2%
19.4%
18.7%
18.0%
17.2%
16.5%
15.7%
15.0%
14.2%
13.5%
12.7%
12.0%
11.3%
10.5%
9.8%
9.1%
8.4%
7.7%
7.0%
6.3%
5.6%
4.9%
4.3%
3.6%
Kidney
Bean
36.1%
34.0%
31.9%
29.8%
27.8%
25.8%
23.9%
22.0%
20.1%
18.4%
16.6%
15.0%
13.4%
11.9%
10.5%
9.2%
7.9%
6.8%
5.7%
4.7%
3.8%
3.0%
2.4%
1.8%
Median
(10
species)
13.0%
12.4%
11.8%
11.2%
10.6%
10.0%
9.4%
8.8%
8.2%
7.7%
7.1%
6.4%
5.7%
5.1%
4.4%
3.9%
3.3%
2.8%
2.4%
2.0%
1.6%
1.3%
1.0%
0.8%
Number
of Species
<5%
3
4
4
4
4
4
4
4
4
4
5
5
5
5
5
6
6
6
8
9
9
10
10
10
Number
of Species
< 10%
4
5
5
5
5
5
5
5
5
7
8
8
8
8
9
10
10
10
10
10
10
10
10
10
Number
of Species
< 20%
7
8
9
9
9
9
9
9
9
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
Number of
Species > 5%
and < 10%
1
1
1
1
1
1
1
1
1
o
J
o
J
o
J
o
J
3
4
4
4
4
2
1
1
0
0
0
Number of
Species >
10% and <
20%
3
o
3
4
4
4
4
4
4
4
3
2
2
2
2
1
0
0
0
0
0
0
0
0
o
5C-3
-------
Table 5C-3 Tree Seedling Biomass Loss and Crop Yield Loss estimated for
over a Season.
exposure
W126 index
value
for exposure
period
23 ppm-hrs
22 ppm-hrs
21 ppm-hrs
20 ppm-hrs
19 ppm-hrs
18 ppm-hrs
17 ppm-hrs
16 ppm-hrs
15 ppm-hrs
Tree seedling biomass lossA
Median Value
Median species
w. 7.6% loss B
Median species
w. 7.2% loss B
Median species
w. 6.8% loss B
Median species
w. 6.4% loss B
Median species
w. 6.0% loss B
Median species
w. 5.7% loss B
Median species
w. 5.3% loss B
Median species
w. 4.9% loss B
Median species
w. 4.5% loss B
Individual Species
< 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
< 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
< 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
< 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
< 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
< 2% loss: 5/1 1 species
< 5% loss: 5/11 species
<10% loss: 7/11 species
<15%loss: 10/11 species
>30%loss: 1/11 species
< 2% loss: 5/1 1 species
<5%loss: 5/11 species
<10% loss: 9/11 species
<15%loss: 10/11 species
>30%loss: 1/11 species
< 2% loss: 5/1 1 species
< 5% loss: 6/11 species
<10%loss: 10/11 species
>30%loss: 1/11 species
< 2% loss: 5/1 1 species
<5%loss: 6/11 species
<10%loss: 10/11 species
>30%loss: 1/11 species
Crop yield lossc
Median Value
Median species w.
8O O/ 1~oo D
.0 /o 10SS
Median species w.
8.2 % loss D
Median species w.
7.7 % loss D
Median species w.
7.1%lossD
Median species w.
6.4 % loss D
Median species w.
5. 7% loss D
Median species w.
5.1 %lossD
Median species w.
4.4 % loss D
Median species w.
<5% loss D
Individual Species
< 5% loss: 4/10 species
>5,<10%loss: 1/10 species
>10,<20% loss: 4/10 species
>20: 1/10 species
< 5% loss: 4/10 species
>5,<10%loss: 1/10 species
>10,<20% loss: 4/10 species
>20: 1/10 species
< 5% loss: 4/10 species
>5,<10%loss: 3/10 species
>10,<20% loss: 3/10 species
< 5% loss: 5/10 species
>5,<10%loss: 3/10 species
>10,<20% loss: 2/10 species
< 5% loss: 5/10 species
>5, <10% loss: 3/10 species
>10,<20% loss: 2/10 species
< 5% loss: 5/10 species
>5,<10%loss: 3/10 species
>10,<20% loss: 2/10 species
< 5% loss: 5/10 species
>5, <10% loss: 3/10 species
>10,<20% loss: 2/10 species
< 5% loss: 5/10 species
>5,<10% loss: 4/10 species
>10,<20% loss: 1/10 species
< 5% loss: 6/10 species
>5, <10% loss: 4/10 species
5C-4
-------
14 ppm-hrs
13 ppm-hrs
12 ppm-hrs
1 1 ppm-hrs
10 ppm-hrs
9 ppm-hrs
8 ppm-hrs
7 ppm-hrs
Median species
w. 4.2% loss B
Median species
w. 3. 8% loss B
Median species
w. 3. 5% loss B
Median species
w. 3.1% loss B
Median species
w. 2.8% loss B
Median species
w. 2.4% loss B
Median species
w. 2.0% loss B
Median species
w. <2% loss B
< 2% loss: 5/1 1 species
< 5% loss: 6/11 species
<10%loss: 10/11 species
>30%loss: 1/11 species
< 2% loss: 5/1 1 species
<5%loss: 7/11 species
<10%loss: 10/11 species
>20%loss: 1/11 species
< 2% loss: 5/1 1 species
< 5% loss: 8/11 species
<10%loss: 10/11 species
>20%loss: 1/11 species
< 2% loss: 5/1 1 species
<5%loss: 8/11 species
<10%loss: 10/11 species
>20%loss: 1/11 species
< 2% loss: 5/1 1 species
< 5% loss: 9/11 species
<10%loss: 10/11 species
>20%loss: 1/11 species
< 2% loss: 5/1 1 species
^5% loss: 10/11 species
>20%loss: 1/11 species
< 2% loss: 5/1 1 species
< 5% loss: 10/11 species
>15%loss: 1/11 species
< 2% loss: 7/1 1 species
<5%loss: 10/11 species
>15%loss: 1/11 species
Median species w.
<5% loss D
Median species
w.<5% loss D
Median species w.
<5% loss D
Median species w.
<5% loss D
Median species w.
<5% loss D
Median species w.
<5% loss D
Median species w.
<5% loss D
Median species w.
<5% loss D
< 5% loss: 6/10 species
>5,<10% loss: 4/10 species
< 5% loss: 6/10 species
>5, <10% loss: 4/10 species
< 5% loss: 8/10 species
>5,<10% loss: 2/10 species
< 5% loss: 9/10 species
>5, <10% loss: 1/10 species
< 5% loss: 9/10 species
>5,<10%loss: 1/10 species
< 5% loss: all species
< 5% loss: all species
< 5% loss: all species
A Estimates here are based on the 1 1 E-R functions for tree seedlings described in WREA, Appendix 6F and
discussed in section 5.2.1, with the exclusion of cottonwood. See CASAC comments (Frey, 2014).
B This median value is the median of the composite E-R functions for 1 1 tree species in the WREA, Appendix 6F
(also discussed in section 5.2. 1).
C Estimates here are based on the 10 E-R functions for crops described in Appendix 6F and discussed in section
5.3.1.
D This median value is the median of the composite E-R functions for 10 crops from WREA, Appendix 6F (also
discussed in section 5.3.1).
5C-5
-------
United States Office of Air Quality Planning and Standards Publication No. EPA-452/R-14-006
Environmental Protection Health and Environmental Impacts Division August 2014
Agency Research Triangle Park, NC
------- |