EPA/600/R-20/012 | April 2020| www.epa.gov/isa
United States
Environmental Protection
Agency
&EPA
Integrated Science Assessment
for Ozone and Related
Photochemical Oxidants
Office of Research and Development
Center for Public Health & Environmental Assessment, Research Triangle Park, NC
-------
United States
Environmental Protection
^^Inal # % Agency
EP A/600/R-20/012
April 2020
www.epa.gov/isa
Integrated Science Assessment for
Ozone and Related Photochemical
Oxidants
April 2020
Center for Public Health and Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC
-------
Disclaimer
This document has been reviewed in accordance with the U.S. Environmental Protection Agency policy
and approved for publication. Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.
ii
-------
Contents
LIST OF TABLES x
LIST OF FIGURES xxiv
INTEGRATED SCIENCE ASSESSMENT TEAM FOR OZONE AND RELATED
PHOTOCHEMICAL OXIDANTS xxxiii
AUTHORS, CONTRIBUTORS, AND REVIEWERS xxxvi
CLEAN AIR SCIENTIFIC ADVISORY COMMITTEE xlii
ACRONYMS AND ABBREVIATIONS xliii
PREFACE Ixii
Legislative Requirements for the Review of the National Ambient Air Quality Standards Ixii
History of the Reviews of the Primary and Secondary National Ambient Air Quality Standard for
Ozone Ixiv
Purpose and Overview of the Integrated Science Assessment Ixix
Process for Developing Integrated Science Assessments Ixx
Scope of the ISA Ixxii
Evaluation of the Evidence Ixxiv
References for Preface Ixxvi
EXECUTIVE SUMMARY ES-1
ES.1 Purpose and Scope of the Integrated Science Assessment ES-1
ES.2 Ozone in Ambient Air ES-2
ES.3 Exposure to Ozone ES-4
ES.4 Health and Welfare Effects of Ozone Exposure ES-4
ES.4.1 Health Effects of Ozone Exposure ES-5
ES.4,2 Ozone Exposure and Welfare Effects ES-10
ES.5 Key Aspects of Health and Welfare Effects Evidence ES-16
ES.5.1 Health Effects Evidence: Key Findings ES-16
ES.5 2 Welfare Effects Evidence: Key Findings ES-17
ES.6 References ES-20
INTEGRATED SYNTHESIS IS-1
IS. 1 Introduction IS-2
IS 1.1 Purpose and Overview IS-2
IS.1.2 Process and Development IS-3
IS 1.3 New Evidence Evaluation and Causality Determinations IS-6
15.2 Atmospheric Chemistry, Ambient Air Ozone Concentrations, and Background Ozone IS-12
15.2.1 Ambient Air Ozone Anthropogenic Sources, Measurement, and Concentrations IS-12
15.2.2 Background Ozone IS-14
15.3 Exposure to Ambient Ozone IS-17
15.3.1 Human Exposure Assessment in Epidemiologic Studies IS-17
15.3.2 Ecological Exposure IS-19
15.4 Evaluation of the Health Effects of Ozone IS-20
IS.4.1 Connections among Health Effects IS-20
iii
-------
15.4.2 Biological Plausibility IS-21
15.4.3 Summary of Health Effects Evidence IS-23
15.4.4 At-Risk Populations IS-50
15.5 Evaluation of Welfare Effects of Ozone IS-62
15.5.1 Ecological Effects IS-67
15.5.2 Effects on Climate IS-80
15.6 Key Aspects of Health and Welfare Effects Evidence IS-82
15.6.1 Health Effects Evidence: Key Findings IS-83
15.6.2 Welfare Effects Evidence: Key Findings IS-88
15.7 References for Integrative Synthesis IS-94
APPENDIX 1 ATMOSPHERIC SOURCE, CHEMISTRY, METEOROLOGY, TRENDS, AND
BACKGROUND OZONE 1-1
1.1 Introduction 1-2
1.2 Metrics and Definitions 1-3
1.2.1 Ozone Metrics 1-3
1.2.2 Background Ozone Definitions 1-4
1.3 Sources of U.S. Ozone and Its Precursors 1-7
1.3.1 Precursor Sources 1-9
1.3.2 Stratosphere-Troposphere Exchange (STE) Processes 1-24
1.4 Ozone Photochemistry 1-26
1.4.1 Winter Ozone in Western Intermountain Basins 1-28
1.4.2 Halogen Chemistry 1-29
1.5 Interannual Variability and Longer-Term Trends in Meteorological Effects on
Anthropogenic and U.S. Background (USB) Ozone 1-31
1.5.1 Meteorological Effects on Ozone Concentrations at the Ground Level 1-32
1.5.2 Interannual and Multidecadal Climate Variability 1-33
1.5.3 Interactions between Meteorology and Topography 1-34
1.6 Measurements and Modeling 1-35
1.6.1 Advances in Ozone Measurement Methods 1-35
1.6.2 Advances in Regional Chemical Transport Modeling 1-38
1.7 Ambient Air Concentrations and Trends 1-40
1.8 U.S. Background (USB) Ozone Concentrations 1-52
1.8.1 Modeling Strategies Applied to Estimate U.S. Background (USB) Ozone 1-53
1.8.2 Concentrations and Trends of U.S. Background (USB) and Baseline Ozone 1-58
1.9 Summary 1-70
1.10 References 1-72
APPENDIX 2 EXPOSURE TO AMBIENT OZONE 2-1
2.1 Introduction 2-1
2.2 Exposure Concepts 2-2
2.3 Exposure Assessment Methods 2-4
2.3.1 Monitoring 2-4
2.3.2 Modeling 2-6
2.4 Personal Exposure 2-14
2.4.1 Time-Activity Data 2-14
2.4.2 Infiltration 2-21
2.4.3 Relationships between Personal Exposure and Ambient Concentration 2-30
2.5 Copollutant Correlations and Potential for Confounding 2-33
2.6 Interpreting Exposure Measurement Error for Use in Epidemiology Studies 2-37
2.6.1 Short-Term Exposure 2-38
2.6.2 Long-Term Exposure 2-42
iv
-------
2.7 Conclusions 2-55
2.8 Evidence Inventories—Data Tables to Provide Supporting Information 2-57
2.9 References 2-158
APPENDIX 3 HEALTH EFFECTS—RESPIRATORY 3-1
3.1 Short-Term Ozone Exposure 3-1
3.1.1 Introduction, Summary from the 2013 Ozone ISA, and Scope for Current Review 3-1
3.1.2 Population, Exposure, Comparison, Outcome, and Study Design (PECOS) Tool 3-3
3.1.3 Biological Plausibility 3-4
3.1.4 Respiratory Effects in Healthy Populations 3-10
3.1.5 Asthma Exacerbation and Associated Respiratory Effects in Populations with
Asthma 3-39
3.1.6 Respiratory Effects in Other Populations with Pre-existing Conditions 3-54
3.1.7 Respiratory Infection and other Associated Health Effects 3-62
3.1.8 Respiratory Related Hospital Admissions and Emergency Department (ED) Visits
for Aggregated Respiratory-Related Diseases 3-66
3.1.9 Respiratory Mortality 3-70
3.1.10 Relevant Issues for Interpreting Epidemiologic Evidence—Short-Term Ozone
Exposure and Respiratory Effects 3-70
3.1.11 Summary and Causality Determination 3-79
3.2 Long-Term Ozone Exposure 3-88
3.2.1 Introduction, Summary from the 2013 Ozone ISA, and Scope for Current Review 3-88
3.2.2 Population, Exposure, Comparison, Outcome, and Study Design (PECOS) Tool 3-89
3.2.3 Biological Plausibility 3-90
3.2.4 Respiratory Health Effects 3-93
3.2.5 Relevant Issues for Interpreting Epidemiologic Evidence—Long-Term Ozone
Exposure and Respiratory Effects 3-112
3.2.6 Summary and Causality Determination 3-113
3.3 Evidence Inventories—Data Tables to Summarize Study Details 3-120
3.3.1 Short-Term Exposure 3-120
3.3.2 Long-Term Exposure 3-178
Annex for Appendix 3: Evaluation of Studies on Health Effects of Ozone 3-194
3.4 References 3-201
APPENDIX 4 HEALTH EFFECTS—CARDIOVASCULAR 4-1
4.1 Short-Term Ozone Exposure and Cardiovascular Health Effects 4-1
4.1.1 Introduction, Summary from the 2013 Ozone ISA, and Scope for Current Review 4-1
4.1.2 Population, Exposure, Comparison, Outcome, and Study Design (PECOS) Tool 4-2
4.1.3 Biological Plausibility 4-3
4.1.4 Heart Failure, Impaired Heart Function, and Associated Cardiovascular Effects 4-7
4.1.5 Ischemic Heart Disease and Associated Cardiovascular Effects 4-9
4.1.6 Endothelial Dysfunction 4-12
4.1.7 Cardiac Depolarization, Repolarization, Arrhythmia, and Arrest 4-15
4.1.8 Blood Pressure Changes and Hypertension 4-19
4.1.9 Heart Rate (HR) and Heart Rate Variability (HRV) 4-22
4.1.10 Coagulation and Thrombosis 4-25
4.1.11 Systemic Inflammation and Oxidative Stress 4-27
4.1.12 Stroke and Associated Cardiovascular Effects 4-30
4.1.13 Nonspecific Cardiovascular Effects 4-33
4.1.14 Cardiovascular Mortality 4-35
4.1.15 Potential Copollutant Confounding of the Ozone-Cardiovascular Disease (CVD)
Relationship 4-35
v
-------
4.1.16 Effect Modification of the Ozone-Cardiovascular Health Effects Relationship 4-36
4.1.17 Summary and Causality Determination 4-38
4.2 Long-Term Ozone Exposure and Cardiovascular Health Effects 4-46
4.2.1 Introduction, Summary from the 2013 Ozone ISA, and Scope for Current Review 4-46
4.2.2 Biological Plausibility 4-47
4.2.3 Ischemic Heart Disease (IHD) and Associated Cardiovascular Effects 4-50
4.2.4 Atherosclerosis 4-50
4.2.5 Heart Failure and Impaired Heart Function 4-51
4.2.6 Vascular Function 4-52
4.2.7 Cardiac Depolarization, Repolarization, Arrhythmia, and Arrest 4-52
4.2.8 Blood Pressure Changes and Hypertension 4-52
4.2.9 Heart Rate and Heart Rate Variability 4-54
4.2.10 Coagulation 4-54
4.2.11 Systemic Inflammation and Oxidative Stress 4-55
4.2.12 Stroke and Associated Cardiovascular Effects 4-56
4.2.13 Other Cardiovascular Endpoints 4-56
4.2.14 Aggregate Cardiovascular Disease 4-57
4.2.15 Cardiovascular Mortality 4-57
4.2.16 Potential Copollutant Confounding of the Ozone-Cardiovascular Disease (CVD)
Relationship 4-59
4.2.17 Effect Modification of the Ozone-Cardiovascular Relationship 4-61
4.2.18 Summary and Causality Determination 4-63
4.3 Evidence Inventories—Data Tables to Summarize Study Details 4-66
4.3.1 Short-Term Ozone Exposure 4-66
4.3.2 Long-Term Ozone Exposure 4-127
Annex for Appendix 4: Evaluation of Studies on Health Effects of Ozone 4-145
4.4 References 4-152
APPENDIX 5 HEALTH EFFECTS—METABOLIC EFFECTS 5-1
5.1 Short-Term Ozone Exposure—Introduction, Summary from the 2013 Ozone ISA, and
Scope for Current Review 5-1
5.1.1 Population, Exposure, Comparison, Outcome, and Study Design (PECOS) Tool 5-2
5.1.2 Biological Plausibility 5-3
5.1.3 Metabolic Syndrome 5-7
5.1.4 Complications from Diabetes 5-21
5.1.5 Other Indicators of Metabolic Function 5-22
5.1.6 Potential Copollutant Confounding of the Ozone-Metabolic Effects Relationship 5-25
5.1.7 Effect Modification of the Ozone-Metabolic Effects Relationship 5-26
5.1.8 Summary and Causality Determination 5-27
5.2 Long-Term Ozone Exposure—Introduction, Summary from the 2013 Ozone ISA, and
Scope for Current Review 5-32
5.2.1 Population, Exposure, Comparison, Outcome, and Study Design (PECOS) Tool 5-32
5.2.2 Biological Plausibility 5-33
5.2.3 Metabolic Syndrome 5-35
5.2.4 Other Indicators of Metabolic Function 5-40
5.2.5 Development of Diabetes 5-41
5.2.6 Metabolic Disease Mortality 5-44
5.2.7 Potential Copollutant Confounding of the Ozone-Metabolic Effects Relationship 5-44
5.2.8 Effect Modification of the Ozone-Metabolic Effects Relationship 5-45
5.2.9 Summary and Causality Determination 5-46
5.3 Evidence Inventories—Data Tables to Summarize Study Details 5-50
Annex for Appendix 5: Evaluation of Studies on Health Effects of Ozone 5-63
vi
-------
5.4 References
5-70
APPENDIX 6 HEALTH EFFECTS—MORTALITY 6-1
6.1 Short-Term Ozone Exposure and Mortality 6-1
6.1.1 Introduction 6-1
6.1.2 Biological Plausibility 6-4
6.1.3 Total (Nonaccidental) Mortality 6-5
6.1.4 Cause-Specific Mortality 6-7
6.1.5 Effect Modification of the Ozone-Mortality Relationship 6-10
6.1.6 Potential Confounding of the Ozone-Mortality Relationship 6-16
6.1.7 Shape of the Concentration-Response (C-R) Relationship 6-18
6.1.8 Summary and Causality Determination 6-20
6.2 Long-Term Ozone Exposure and Mortality 6-25
6.2.1 Introduction 6-25
6.2.2 Biological Plausibility 6-26
6.2.3 Total (Nonaccidental) Mortality 6-27
6.2.4 Effect Modification of the Ozone-Mortality Relationship 6-34
6.2.5 Potential Copollutant Confounding of the Ozone-Mortality Relationship 6-35
6.2.6 Shape of the Concentration-Response Function 6-37
6.2.7 Summary and Causality Determination 6-40
6.3 Evidence Inventories—Data Tables to Summarize Study Details 6-45
6.3.1 Short-Term Ozone Exposure and Mortality: Data Tables 6-45
6.3.2 Long-Term Ozone Exposure and Mortality: Data Tables 6-58
Annex for Appendix 6: Evaluation of Studies on Health Effects of Ozone 6-69
6.4 References 6-76
APPENDIX 7 HEALTH EFFECTS—OTHER HEALTH ENDPOINTS 7-1
7.1 Reproductive and Developmental Effects 7-2
7.1.1 Introduction, Summary from the 2013 Ozone ISA, and Scope for Current Review 7-2
7.1.2 Male and Female Reproduction and Fertility 7-4
7.1.3 Pregnancy and Birth Outcomes 7-8
7.1.4 Effects of Exposure during Developmental Periods 7-16
7.1.5 Summary and Causality Determinations 7-18
7.2 Nervous System Effects 7-21
7.2.1 Short-Term Ozone Exposure 7-21
7.2.2 Long-Term Ozone Exposure 7-33
7.3 Cancer 7-45
7.3.1 Introduction, Summary from the 2013 Ozone ISA, and Scope for Current Review 7-45
7.3.2 Cancer and Related Health Effects 7-46
7.3.3 Summary and Causality Determination 7-48
7.4 Evidence Inventories—Data Tables to Summarize Reproductive and Developmental
Effects Study Details 7-50
7.4.1 Epidemiologic Studies 7-50
7.4.2 Toxicological Studies 7-96
7.5 Evidence Inventories—Data Tables to Summarize Nervous System Effects Study Details _ 7-101
7.5.1 Epidemiologic Studies 7-101
7.5.2 Toxicological Studies 7-112
7.6 Evidence Inventories—Data Tables to Summarize Cancer Study Details 7-118
7.6.1 Epidemiologic Studies 7-118
7.6.2 Toxicological Studies 7-124
Annex for Appendix 7: Evaluation of Studies on Health Effects of Ozone 7-125
7.7 References 7-132
vii
-------
APPENDIX 8 ECOLOGICAL EFFECTS
8-1
8.1 Introduction 8-1
8.1.1 Scope 8-3
8.1.2 Assessing Ecological Response to Ozone 8-5
8.1.3 Mechanisms Governing Vegetation Response to Ozone 8-11
8.2 Visible Foliar Injury and Biomonitoring 8-13
8.2.1 Summary and Causality Determination 8-24
8.3 Plant Growth 8-30
8.3.1 Declines in Growth Rates 8-30
8.3.2 Changes in Biomass Allocation 8-36
8.3.3 Connections with Community Composition and Water Cycling 8-36
8.3.4 Summary and Causality Determination 8-37
8.4 Plant Reproduction, Phenology, and Mortality 8-43
8.4.1 Plant Reproduction 8-44
8.4.2 Plant Phenology 8-46
8.4.3 Plant Mortality 8-47
8.4.4 Summary and Causality Determinations 8-47
8.5 Reduced Crop Yield and Quality 8-61
8.5.1 Field Studies and Meta-Analyses 8-61
8.5.2 Yield Loss at Regional and National Scales 8-63
8.5.3 Summary and Causality Determination 8-65
8.6 Herbivores: Growth, Reproduction, and Survival 8-72
8.6.1 Individual-Level Responses 8-73
8.6.2 Population- and Community-Level Responses 8-74
8.6.3 Summary and Causality Determination 8-75
8.7 Plant-Insect Signaling 8-88
8.7.1 Emission and Chemical Composition of Volatile Plant Signaling Compounds
(VPSCs) 8-89
8.7.2 Pollinator Attraction and Plant Host Detection 8-90
8.7.3 Plant Attraction of Natural Enemies of Herbivores 8-91
8.7.4 Summary and Causality Determination 8-91
8.8 Carbon Cycling in Terrestrial Ecosystems: Primary Productivity and Carbon Sequestration _ 8-99
8.8.1 Terrestrial Primary Productivity 8-99
8.8.2 Soil Carbon 8-102
8.8.3 Terrestrial Carbon Sequestration 8-103
8.8.4 Summary and Causality Determinations 8-104
8.9 Soil Biogeochemistry 8-115
8.9.1 Decomposition 8-116
8.9.2 Soil Carbon 8-117
8.9.3 Soil Nitrogen 8-119
8.9.4 Summary and Causality Determination 8-121
8.10 Terrestrial Community Composition 8-130
8.10.1 Plant Community 8-131
8.10.2 Microbes 8-136
8.10.3 Consumer Communities 8-139
8.10.4 Summary and Causality Determination 8-139
8.11 Water Cycling 8-161
8.11.1 Structural Changes in Plants 8-161
8.11.2 Impaired Stomatal Function 8-162
8.11.3 Models of Plant Water Use 8-163
8.11.4 Ecosystem Water Dynamics 8-164
8.11.5 Drought and Ozone 8-165
viii
-------
8.11.6 Summary and Causality Determination 8-165
8.12 Modifying Factors 8-175
8.12.1 Nitrogen 8-175
8.12.2 Carbon Dioxide 8-176
8.12.3 Weather and Climate 8-178
8.12.4 Summary 8-179
8.13 Exposure Indices/Exposure Response 8-180
8.13.1 Exposure Indices 8-181
8.13.2 Exposure Response 8-182
8.13.3 Summary 8-203
8.14 References 8-211
APPENDIX 9 THE ROLE OF TROPOSPHERIC OZONE ON CLIMATE EFFECTS 9-1
9.1 Introduction 9-1
9.1.1 Summary from the 2013 Ozone ISA 9-1
9.1.2 Scope for the Current Review 9-2
9.1.3 Introduction to Climate, Ozone Chemistry, and Radiative Forcing 9-3
9.2 Ozone Impacts on Radiative Forcing 9-8
9.2.1 Recent Evidence for Historical Period 9-8
9.2.2 Recent Evidence of Radiative Forcing Temporal and Spatial Trends 9-13
9.2.3 Summary and Causality Determination 9-15
9.3 Ozone Impacts on Temperature, Precipitation, and Related Climate Variables 9-17
9.3.1 Recent Evidence for Effects on Temperature 9-18
9.3.2 Recent Evidence for Other Climate Effects 9-20
9.3.3 Summary and Causality Determination 9-21
9.4 References 9-24
APPENDIX 10 DEVELOPMENT OF THE INTEGRATED SCIENCE
ASSESSMENT—PROCESS 10-1
10.1 Introduction 10-1
10.2 Literature Search and Initial Screen 10-1
10.2.1 Literature Search 10-5
10.2.2 Study Selection: Initial Screening (Level 1) of Studies from the Literature Search 10-6
10.2.3 Documentation 10-10
10.3 Study Selection: Full-Text Evaluation of Studies (Level 2) 10-10
10.3.1 Relevance 10-11
10.3.2 Individual Study Quality 10-21
10.3.3 Documentation 10-24
10.4 Peer Review and Public Participation 10-24
10.4.1 Call for Information 10-24
10.4.2 Integrated Review Plan 10-25
10.4.3 Peer Input 10-25
10.4.4 Internal Technical Review 10-26
10.4.5 Clean Air Scientific Advisory Committee (CASAC) Peer Review 10-27
10.5 Quality Assurance 10-28
10.6 Conclusion 10-29
10.7 References 10-30
ix
-------
LIST OF TABLES
Table I
Table II
Table 1-1
Table 1-2
Table 2-1
Table 2-2
Table 2-3
Table 2-4
Table 2-5
Table 2-6
Table 2-7
Table 2-8
Table 2-9
History of the National Ambient Air Quality Standards for Ozone,
1971-2015.
Weight of evidence for causality determinations.
Nationwide distributions of ozone concentrations (parts per billion
[ppb]) from the year-round data set 2015-2017.
Nationwide distributions of ozone concentrations (parts per billion
[ppb]) from the warm-season data set 2015-2017.
Summary of U.S. studies of ozone infiltration published after 2011.
lxvi
lxxi
1-41
1-43
Total and age-stratified percentage of hours spent in different locations
from the Consolidated Human Activity Database (Isaacs, 2014), warm
season for all hours and for afternoon hours (12:00 p.m.-8:00 p.m.). 2-18
Total and race/ethnicity-stratified percentage of hours spent in different
locations from the Consolidated Human Activity Database (Isaacs,
2014), warm season for all hours and for afternoon hours (12:00
p.m.-8:00 p.m.). 2-19
Total and sex-stratified percentage of hours spent in different locations
from the Consolidated Human Activity Database (Isaacs, 2014), warm
season for all hours and for afternoon hours (12:00 p.m.-8:00 p.m.). 2-20
2-22
Studies reporting relationships between personal ozone exposures and
ambient ozone concentrations. 2-31
Summary of the influence of exposure error on epidemiologic study
outcomes. 2-44
Summary of exposure estimation methods, their typical use in ozone
epidemiologic studies, and related errors and uncertainties. 2-45
Studies informing assessment of exposure measurement error when
concentrations measured by fixed-site monitors are used for exposure
surrogates. 2-59
Studies informing assessment of exposure measurement error when
concentrations measured by personal and microenvironmental monitors
are used for exposure surrogates. 2-64
X
-------
LIST OF TABLES (Continued)
Table 2-10 Studies informing assessment of exposure measurement error when
concentrations modeled by spatial interpolations methods are used for
exposure surrogates. 2-66
Table 2-11 Studies informing assessment of exposure measurement error when
concentrations modeled by land use regression or spatiotemporal
models are used for exposure surrogates. 2-73
Table 2-12 Studies informing assessment of exposure measurement error when
concentrations modeled by chemical transport modeling are used for
exposure surrogates. 2-81
Table 2-13 Studies informing assessment of exposure measurement error when
concentrations modeled by hybrid approaches are used for exposure
surrogates. 2-140
Table 2-14 Studies informing assessment of exposure measurement error when
concentrations modeled by microenvironmental modeling are used for
exposure surrogates. 2-156
Table 3-1 Heat map of daily lag associations between short-term exposure to
ozone and hospital admissions and emergency department (ED) visits
for asthma. 3-75
Table 3-2 Summary of evidence indicating a causal relationship between
short-term ozone exposure and respiratory effects.
3-83
Table 3-3 Summary of evidence for a likely to be causal relationship between
long-term ozone exposure and respiratory effects.
3-116
Table 3-4
Table 3-5
Study-specific details from controlled human exposure studies of lung
function in healthy populations. 3-120
Study-specific details from animal toxicological studies of short-term
ozone exposure and lung function in healthy animals. 3-122
Table 3-6 Epidemiologic studies of short-term exposure to ozone and lung
function in healthy populations.
3-126
Table 3-7 Epidemiologic studies of short-term exposure to ozone and lung
function, airway inflammation, and oxidative stress in general
populations. 3-128
Table 3-8 Study-specific details from controlled human exposure studies of
respiratory symptoms in healthy populations.
3-129
Table 3-9 Study-specific details from controlled human exposure studies of
inflammation, oxidative stress, and injury in healthy populations.
3-129
XI
-------
LIST OF TABLES (Continued)
Table 3-10
Table 3-11
Table 3-12
Table 3-13
Table 3-14
Table 3-15
Table 3-16
Study-specific details from animal toxicological studies of short-term
ozone exposure and allergic sensitization in healthy animals. 3-131
Study-specific details from animal toxicological studies of short-term
ozone exposure and inflammation, oxidative stress, and injury in
healthy animals. 3-132
Study-specific details from animal toxicological studies of short-term
ozone exposure and morphology in healthy animals. 3-142
Epidemiologic studies of short-term exposure to ozone and hospital
admission for asthma.
3-145
Epidemiologic studies of short-term exposure to ozone and emergency
department (ED) visits for asthma. 3-147
Epidemiologic studies of short-term exposure to ozone and respiratory
symptoms in children with asthma. 3-152
Study-specific details from controlled human exposure studies of lung
function in adults with asthma. 3-152
Table 3-17 Study-specific details from controlled human exposure studies of lung
function in healthy adults and adults with asthma. 3-153
Table 3-18 Study-specific details from animal toxicological studies of short-term
ozone exposure and lung function—allergy. 3-153
Table 3-19 Study-specific details from animal toxicological studies of short-term
ozone exposure and lung function—asthma. 3-154
Table 3-20 Study-specific details from controlled human exposure studies of
inflammation, oxidative stress, and injury in healthy adults and adults
with asthma. 3-155
Table 3-21
Table 3-22
Table 3-23
Table 3-24
Study-specific details from controlled human exposure studies of
allergic sensitization—atopy.
3-155
Study-specific details from animal toxicological studies of short-term
ozone exposure and inflammation, oxidative stress, and
injury—allergy. 3-156
Study-specific details from animal toxicological studies of short-term
ozone exposure and inflammation, oxidative stress, and
injury—asthma. 3-156
Study-specific details from animal toxicological studies of short-term
ozone exposure and morphology—asthma. 3-157
Xll
-------
LIST OF TABLES (Continued)
Table 3-25
Table 3-26
Table 3-27
Table 3-28
Table 3-29
Table 3-30
Table 3-31
Table 3-32
Table 3-33
Table 3-34
Table 3-35
Table 3-36
Table 3-37
Table 3-38
Epidemiologic studies of short-term exposure to ozone and
inflammation, oxidative stress, and injury in children with asthma.
3-157
Epidemiologic studies of short-term exposure to ozone and emergency
department (ED) visits for chronic obstructive pulmonary disease
(COPD). 3-158
Epidemiologic studies of short-term exposure to ozone and medication
use in adults with chronic obstructive pulmonary disease (COPD). 3-161
Study-specific details from animal toxicological studies of short-term
ozone exposure and lung function—chronic obstructive pulmonary
disease (COPD). 3-161
Study-specific details from animal toxicological studies of short-term
ozone exposure and inflammation, oxidative stress, and
injury—chronic obstructive pulmonary disease (COPD). 3-162
Study-specific details from animal toxicological studies of short-term
ozone exposure and morphology—chronic obstructive pulmonary
disease (COPD). 3-162
Study-specific details from controlled human exposure studies of
respiratory effects in obese adults.
3-163
Study-specific details from animal toxicological studies of short-term
ozone exposure and lung function—obesity. 3-163
Study-specific details from animal toxicological studies of short-term
ozone exposure and inflammation, oxidative stress, and
injury—obesity. 3-164
Study-specific details from animal toxicological studies of short-term
ozone exposure and morphology—obesity. 3-165
Epidemiologic studies of short-term exposure to ozone and pulmonary
inflammation in populations with metabolic syndrome. 3-165
Study-specific details from animal toxicological studies of short-term
ozone exposure and lung function—cardiovascular disease. 3-166
Study-specific details from animal toxicological studies of short-term
ozone exposure and inflammation, oxidative stress, and
injury—cardiovascular disease. 3-167
Study-specific details from animal toxicological studies of short-term
ozone exposure and morphology—cardiovascular disease. 3-168
Xlll
-------
LIST OF TABLES (Continued)
Table 3-39
Table 3-40
Table 3-41
Table 3-42
Table 3-43
Table 3-44
Table 3-45
Table 3-46
Table 3-47
Epidemiologic studies of short-term exposure to ozone and emergency
department (ED) visits for respiratory infection. 3-169
Study-specific details from animal toxicological studies of short-term
ozone exposure and host defense/infection—healthy. 3-173
Epidemiologic studies of short-term exposure to ozone and hospital
admissions for aggregate respiratory diseases.
Epidemiologic studies of long-term exposure to ozone and
development of asthma.
3-174
Epidemiologic studies of short-term exposure to ozone and emergency
department (ED) visits for aggregate respiratory diseases. 3-175
3-178
Study-specific details from animal toxicological studies of long-term
ozone exposure and inflammation, oxidative stress, and
injury—allergy. 3-180
Study-specific details from animal toxicological studies of long-term
ozone exposure and lung function—allergy. 3-181
Study-specific details from animal toxicological studies of long-term
ozone exposure and inflammation, oxidative stress, and
injury—healthy. 3-182
Study-specific details from animal toxicological studies of long-term
ozone exposure and morphology and other endpoints in healthy
animals. 3-184
Table 3-48
Table 3-49
Table 3-50
Table 3-51
Table 3-52
Epidemiologic studies of long-term exposure to ozone and lung
function and development.
Epidemiologic studies of long-term exposure to ozone and
development of chronic obstructive pulmonary disease (COPD).
3-185
Study-specific details from animal toxicological studies of long-term
ozone exposure and morphology—allergy. 3-187
Study-specific details from animal toxicological studies of long-term
ozone exposure and lung function in healthy animals. 3-188
3-189
Epidemiologic studies of long-term exposure to ozone and respiratory
infection. 3-190
Table 3-53 Epidemiologic studies of long-term exposure to ozone and severity of
respiratory disease. 3-191
XIV
-------
LIST OF TABLES (Continued)
Table 3-54 Epidemiologic studies of long-term exposure to ozone and allergic
sensitization. 3-192
Table 3-55 Study-specific details from animal toxicological studies of long-term
ozone exposure and allergic sensitization in healthy animals. 3-193
Table 4-1 Summary of evidence that is suggestive of, but not sufficient to infer, a
causal relationship between short-term ozone exposure and
cardiovascular effects. 4-43
Table 4-2 Summary of evidence that is suggestive of, but not sufficient to infer, a
causal relationship between long-term ozone exposure and
cardiovascular effects. 4-65
Table 4-3
Table 4-4
Epidemiologic studies of short-term exposure to ozone and heart
failure.
Study-specific details from controlled human exposure studies of
impaired heart function.
4-66
4-68
Table 4-5
Table 4-6
Study-specific details from short-term animal toxicological studies of
impaired heart function. 4-69
Epidemiologic studies of short-term exposure to ozone and ischemic
heart disease. 4-72
Table 4-7
Epidemiologic panel studies of short-term exposure to ozone and
ischemic heart disease.
4-79
Table 4-8
Study-specific details from controlled human exposure studies of
ST-segment depression.
4-80
Table 4-9
Study-specific details from short-term animal toxicological studies of
ST-segment depression. 4-80
Table 4-10
Table 4-11
Epidemiologic panel studies of short-term exposure to ozone and
endothelial function.
Study-specific details from controlled human exposure studies of
vascular function.
4-81
4-82
Table 4-12
Study-specific details from short-term animal toxicological studies of
vascular function. 4-83
Table 4-13 Epidemiologic studies of short-term exposure to ozone and emergency
department visits or hospital admissions for electrophysiological
changes, arrhythmia, and cardiac arrest. 4-84
XV
-------
LIST OF TABLES (Continued)
Table 4-14
Table 4-15
Table 4-16
Table 4-17
Table 4-18
Table 4-19
Table 4-20
Table 4-21
Table 4-22
Table 4-23
Table 4-24
Table 4-25
Table 4-26
Table 4-27
Table 4-28
Table 4-29
Table 4-30
Epidemiologic panel studies of short-term exposure to ozone and
electrophysiology, arrhythmia, and cardiac arrest.
4-88
Study-specific details from controlled human exposure studies of
electrophysiology, arrhythmia, cardiac arrest.
4-89
Study-specific details from short-term animal toxicological studies of
electrophysiology, arrhythmia, cardiac arrest. 4-90
Epidemiologic studies of short-term exposure to ozone and blood
pressure. 4-91
Epidemiologic panel studies of short-term exposure to ozone and blood
pressure. 4-93
Study-specific details from controlled human exposure studies of blood
pressure. 4-95
Study-specific details from short-term animal toxicological studies of
blood pressure. 4-96
Epidemiologic panel studies of short-term exposure to ozone and heart
rate variability (HRV), heart rate (HR). 4-98
Study-specific details from controlled human exposure studies of heart
rate variability (HRV), heart rate (HR). 4-100
Study-specific details from short-term animal toxicological studies of
heart rate variability (HRV), heart rate (HR). 4-101
Epidemiologic studies of short-term exposure to ozone and thrombosis. 4-103
Epidemiologic panel studies of short-term exposure to ozone and
coagulation. 4-105
Study-specific details from controlled human exposure studies of
coagulation. 4-106
Study-specific details from short-term animal toxicological studies of
coagulation. 4-107
Epidemiologic panel studies of short-term exposure to ozone and
inflammation. 4-108
Study-specific details from controlled human exposure studies of
systemic inflammation and oxidative stress. 4-110
Study-specific details from short-term animal toxicological studies of
systemic inflammation and oxidative stress. 4-112
XVI
-------
LIST OF TABLES (Continued)
Table 4-31
Table 4-32
Table 4-33
Table 4-34
Table 4-35
Table 4-36
Table 4-37
Table 4-38
Table 4-40
Table 4-41
Table 4-42
Table 4-43
Table 4-44
Table 4-45
Table 4-46
Epidemiologic studies of short-term exposure to ozone and
cerebrovascular disease.
4-115
Epidemiologic studies of short-term exposure to ozone and aggregate
cardiovascular disease. 4-123
Epidemiologic studies of short-term exposure to ozone and
cardiovascular mortality.
4-126
Epidemiologic studies of long-term exposure to ozone and ischemic
heart disease (IHD). 4-127
Epidemiologic studies of long-term exposure to ozone and
atherosclerosis.
4-128
Study-specific details from animal toxicological studies of
atherosclerosis.
4-129
Epidemiologic studies of long-term exposure to ozone and heart failure. 4-130
Study-specific details from animal toxicological studies of impaired
heart function. 4-131
Table 4-39 Study-specific details from animal toxicological studies of vascular
function. 4-131
Epidemiologic studies of long-term exposure to ozone and blood
pressure. 4-132
Study-specific details from animal toxicological studies of blood
pressure. 4-139
Study-specific details from animal toxicological studies of heart rate
variability (HRV), heart rate (HR). 4-139
Study-specific details from animal toxicological studies of coagulation. 4-140
Study-specific details from animal toxicological studies of
inflammation.
4-141
Epidemiologic studies of long-term exposure to ozone and
cerebrovascular disease.
4-142
Epidemiologic studies of long-term exposure to ozone and aggregate
cardiovascular disease. 4-144
Table 5-1 Criteria for clinical diagnosis of metabolic syndrome*.
5-7
XVII
-------
LIST OF TABLES (Continued)
Table 5-2
Table 5-3
Table 5-4
Table 5-5
Table 5-6
Table 5-7
Table 5-8
Table 5-9
Table 5-10
Table 5-11
Table 5-12
Table 5-13
Table 5-14
Table 5-15
Table 6-1
Summary of changes in serum lipids from experiments conducted in
male rats.
Summary of evidence for a likely to be causal relationship between
short-term ozone exposure and metabolic effects.
Criteria for clinical diagnosis of diabetes.
Study-specific details from animal toxicological studies of short-term
exposure to ozone and metabolic syndrome.
Study-specific details from epidemiologic studies of short-term
complications from diabetes.
Study-specific details from controlled human exposure studies of
short-term, other indicators of metabolic function.
Epidemiologic studies of long-term exposure to ozone and metabolic
syndrome.
Study-specific details from animal toxicological studies of long-term
exposure to ozone and metabolic syndrome.
Epidemiologic studies of long-term exposure to ozone and
development of diabetes.
5-17
5-30
5-42
Summary of evidence that is suggestive of, but not sufficient to infer, a
causal relationship between long-term ozone exposure and metabolic
effects. 5-48
Epidemiologic studies of short-term exposure to ozone and metabolic
syndrome. 5-50
Controlled human exposure study of short-term exposure to ozone and
metabolic syndrome. 5-52
5-52
5-55
Study-specific details from animal toxicological studies of short-term,
other indicators of metabolic function. 5-55
5-57
5-58
5-59
Study-specific details from animal toxicological studies of long-term
exposure to ozone and other indicators of metabolic function. 5-60
5-60
Summary of evidence that is suggestive of, but not sufficient to infer, a
causal relationship between short-term ozone exposure and total
mortality. 6-23
xviii
-------
LIST OF TABLES (Continued)
Table 6-2
Table 6-3
Table 6-4
Table 6-5
Table 6-6
Table 6-7
Table 6-8
Table 6-9
Table 7-1
Table 7-2
Table 7-3
Table 7-4
Table 7-5
Table 7-6
Table 7-7
Summary of evidence that is suggestive of, but not sufficient to infer, a
causal relationship between long-term ozone exposure and total
mortality. 6-43
Epidemiologic studies of short-term exposure to ozone and total
(nonaccidental) mortality.
Epidemiologic studies of short-term exposure to ozone and
cardiovascular mortality.
Epidemiologic studies of long-term exposure to ozone and total
(nonaccidental) mortality.
Epidemiologic studies of long-term exposure to ozone and
cardiovascular mortality.
Summary of evidence that is inadequate to determine if a causal
relationship exists between long-term ozone exposure and cancer.
6-45
6-54
Epidemiologic studies of short-term exposure to ozone and respiratory
mortality. 6-56
6-58
6-62
Epidemiologic studies of long-term exposure to ozone and respiratory
mortality. 6-65
Epidemiologic studies of long-term exposure to ozone and other
mortality. 6-67
Summary of evidence that is suggestive of, but not sufficient to infer, a
causal relationship between ozone exposure and male and female
reproduction. 7-19
Summary of evidence that is suggestive of, but not sufficient to infer, a
causal relationship between ozone exposure and pregnancy and birth
outcomes. 7-20
Summary of evidence for a relationship between short-term ozone
exposure and nervous system effects that is suggestive of, but not
sufficient to infer, a causal relationship. 7-32
Summary of evidence for a relationship between long-term ozone
exposure and nervous system effects that is suggestive, but not
sufficient to infer, a causal relationship. 7-43
7-49
Epidemiologic studies of exposure to ozone and reproduction—male. 7-50
Epidemiologic studies of exposure to ozone and reproduction—female. 7-51
XIX
-------
LIST OF TABLES (Continued)
Table 7-8
Table 7-9
Table 7-10
Table 7-11
Table 7-12
Table 7-13
Table 7-14
Table 7-15
Table 7-16
Table 7-17
Table 7-18
Table 7-19
Table 7-20
Table 7-21
Table 7-22
Table 7-23
Table 7-24
Epidemiologic studies of exposure to ozone and
pregnancy/birth—hypertension disorders.
7-53
Epidemiologic studies of exposure to ozone and
pregnancy/birth—diabetes.
7-56
Epidemiologic studies of exposure to ozone and pregnancy/birth—fetal
growth. 7-57
Epidemiologic studies of exposure to ozone and
pregnancy/birth—preterm birth.
7-65
Epidemiologic studies of exposure to ozone and pregnancy/birth—birth
defects. 7-75
Epidemiologic studies of exposure to ozone and
pregnancy/birth—infant and fetal mortality.
7-81
Epidemiologic studies of exposure to ozone and developmental effects. 7-84
Epidemiologic studies of exposure to ozone and other effects measured
during pregnancy. 7-92
Study-specific details from studies of ozone and pregnancy/birth
outcomes. 7-96
Study-specific details from studies of ozone and developmental effects. 7-98
Epidemiologic studies of short-term exposure to ozone and effects on
cognition, motor activity, and mood. 7-101
Epidemiologic studies of short-term exposure to ozone and hospital
admissions, emergency department, and outpatient visits. 7-102
Epidemiologic studies of long-term exposure to ozone and
cognitive/behavioral effects.
7-105
Epidemiologic studies of long-term exposure to ozone and
neurodegenerative diseases.
7-107
Epidemiologic studies of long-term exposure to ozone and
neurodevelopmental effects.
7-110
Study-specific details from short-term studies of brain inflammation
and morphol ogy. 7-112
Study-specific details from short-term studies of cognitive and
behavioral effects.
7-114
XX
-------
LIST OF TABLES (Continued)
Table 7-25
Table 7-26
Table 7-27
Table 7-28
Table 7-29
Table 7-30
Table 7-31
Table 7-32
Table 8-1
Table 8-2
Table 8-3
Table 8-4
Table 8-5
Table 8-6
Table 8-7
Table 8-8
Table 8-9
Table 8-10
Study-specific details from short-term studies of neuroendocrine
effects. 7-115
Study-specific details from long-term studies of brain inflammation and
morphology. 7-115
Study-specific details from long-term studies of cognitive and
behavioral effects.
7-117
Study-specific details of long-term exposures and neurodevelopmental
effects. 7-118
Epidemiologic studies of long-term exposure to ozone and cancer
incidence.
Epidemiologic studies of ozone exposure and lung cancer mortality.
7-
7-
Epidemiologic studies of long-term exposure to ozone and other cancer
endpoints. 7-
Study-specific details of ozone exposure and DNA damage.
7-
Summary of ozone causality determinations for effects on vegetation
and ecosystems in the 2013 Ozone ISA.
Population, Exposure, Comparison, Outcome, and Study Design
(PECOS) tool for ozone effects on vegetation and ecosystems. _
Plant species that have populations in the U.S. (USDA, 2015) that have
been tested for ozone foliar injury as documented in the references
listed with each in Bergmann et al. (2017).
Ozone exposure and foliar injury.
Plant species that have populations in the U.S. (USDA, 2015) that have
been tested for ozone growth reduction as documented in the references
listed with each species and synthesized in Bergmann et al. (2017).
Ozone exposure and plant growth and biomass.
Summary of evidence for causal relationship between ozone exposure
and plant reproduction.
Summary of evidence for likely to be causal relationship between
ozone exposure and tree mortality.
Ozone exposure and plant reproduction, phenology, and mortality.
Ozone and crop yield and quality.
119
120
122
124
_8-3
_8-5
8-15
8-25
8-32
8-38
8-49
8-50
8-51
8-66
xxi
-------
LIST OF TABLES (Continued)
Table 8-11
Table 8-12
Table 8-13
Table 8-14
Table 8-15
Table 8-16
Table 8-17
Table 8-18
Table 8-19
Table 8-20
Table 8-21
Table 8-22
Table 8-23
Summary of studies reporting altered growth in herbivores.
Table 8-24
Table 8-25
Table 8-26
Table 9-1
Summary of studies reporting altered reproduction in herbivores.
Summary of evidence for likely to be causal relationship between
ozone exposure and alteration of herbivore growth and reproduction.
Ozone exposure and effects on herbivores.
Summary of evidence for a likely to be causal relationship between
ozone exposure and alteration of plant-insect signaling.
Ozone exposure and plant insect signaling.
Ozone exposure effects on productivity and carbon sequestration.
Response of belowground processes and biogeochemical cycles to
ozone exposure.
Summary of evidence for a causal relationship between ozone exposure
and terrestrial community composition, based on Table 2 from the
Preamble.
Terrestrial community composition response to ozone exposure.
Ozone exposure and water cycling.
Ozone exposure-response functions for selected National Crop Loss
Assessment Network (NCLAN) crops.
Weibull exposure-response curves relating relative biomass loss as a
function of 12-hour W126 in ppm-hour for 51 seedling studies as
reported in Tables 12 and 13 of Lee and Hogsett (1996). Relative
Biomass Loss = 1 - exp[-(W126/B)c]
Median composite ozone exposure-response functions for tree
seedlings adjusted to 92-days exposure.
Grassland species that occur in the U.S. with biomass loss
exposure-response functions as a function of AOT40 calculated from
previously published open-top chamber (OTC) experiments by van
Goethem et al. (2013).
Exposure indices and exposure response.
Population, exposure, comparison, outcome, and study design
(PECOS) tool for radiative forcing and climate change.
MO
5-12
5-14
5-14
5-16
5-18
5-18
5-19
5-19
5-20
xxii
-------
LIST OF TABLES (Continued)
Table 9-2
Table 9-3
Table 9-4
Table 9-5
Table 10-1
Table 10-2
Table 10-3
Table 10-4
Contributions of tropospheric ozone changes to radiative forcing
(W/m2) from 1750 to 2011.
Confidence level for ozone forcing for the 1750-2011 period.
Summary of evidence for a causal relationship between tropospheric
ozone and radiative forcing.
Summary of evidence for a likely to be causal relationship between
ozone and temperature, precipitation, and related climate variables.
9-9
9-9
9-17
9-23
Population, Exposure, Comparison, Outcome, and Study Design
(PECOS) tool to define the parameters and provide a framework for
identifying relevant experimental studies. 10-14
Population, Exposure, Comparison, Outcome, and Study Design
(PECOS) tool to define the parameters and provide a framework for
identifying relevant epidemiologic studies. 10-15
Population, Exposure, Comparison, Outcome, and Study Design
(PECOS) tool to define the parameters and provide a framework for
identifying relevant ecological studies. 10-19
Population, Exposure, Comparison, Outcome, and Study Design
(PECOS) tool to define the parameters and provide a framework for
identifying relevant studies on the effects of tropospheric ozone on
climate. 10-20
xxiii
-------
LIST OF FIGURES
Figure ES-1 Individual monitor ozone concentrations in terms of design values
(i.e., 3-year avg of annual fourth-highest max daily 8-hour avg ozone
concentration) for 2015-2017. ES-3
Figure ES-2 Causality determinations for health effects of short- and long-term
exposure to ozone. ES-7
Figure ES-3 Cross-study comparisons of mean decrements in ozone-induced forced
expiratory volume in 1 second (FEVi) in young, healthy adults
following 6.6 hours of exposure to ozone. ES-9
Figure ES-4 Illustrative diagram of ozone effects cascading from the cellular level
to plants and ecosystems. ES-11
Figure ES-5 Causality determinations for ozone across biological scales of
organization and taxonomic groups. ES-12
Figure IS-1 Illustrative figure for potential biological pathways for health effects
following ozone exposure. IS-22
Figure IS-2 Cross-study comparisons of mean ozone-induced forced expiratory
volume in 1 second (FEVi) decrements in young healthy adults
following 6.6 hours of exposure to ozone. IS-28
Figure IS-3 Illustrative diagram of ozone effects cascading up through scales of
biological organization from the cellular level to plants and ecosystems._ IS-68
Figure IS-4 Representative ozone foliar injury in two common tree species in the
U.S. IS-70
Figure IS-5 Schematic illustrating the effects of tropospheric ozone on climate;
including the relationship between precursor emissions, tropospheric
ozone abundance, radiative forcing, climate response and climate
impacts. IS-81
Figure IS-6 Causality determinations for health effects of short- and long-term
exposure to ozone. IS-84
Figure IS-7 Causality determinations for ecological effects of ozone across
biological scales of organization and taxonomic groups. IS-90
Figure IS-8 Causality determinations for tropospheric ozone and climate change. IS-93
Figure 1-1 Major atmospheric processes and precursor sources contributing to
ambient ozone. 1-8
XXIV
-------
LIST OF FIGURES (Continued)
Figure 1-2 Relative ozone precursor emissions by U.S. sector: (A) nitrogen oxides
(NOx). (B) carbon monoxide (CO). (C) volatile organic compounds
(VOCs). Biogenic VOCs, which can be important in the production of
ozone in urban areas, are included for context. (D) methane (CH4). 1-11
Figure 1-3 U.S. anthropogenic ozone precursor emission trends. Sources shown
generate 90% or more of known emissions, excluding biogenic sources,
for the indicated precursor: (A) nitrogen oxides (NOx), (B) carbon
monoxide (CO), (C) volatile organic compounds (VOCs), (D) methane
(CH4). Not shown: "Other" NOx, CO, and VOC emissions categories
that, together, account for less than 10% of total emissions for each
precursor. 1-12
Figure 1-4 Asian anthropogenic ozone precursor emission trends. (A) The study
domain, indicating annual NOx flux rates by location, (B) Annual
nitrogen oxides (NOx) emissions from eight inventories over South
Korea, (C) Annual NOx emissions from eight inventories over China,
and (D) Temporal deviations among eight NOx emissions inventories,
when normalized with respect to 2008 emissions. 1-18
Figure 1-5 Anthropogenic ozone precursor emission trends derived using the
Multi-resolution Emissions Inventory for China emissions model. 1-19
Figure 1-6 Model-estimated April-May stratospheric ozone contributions and
observed surface ozone concentrations between 1990 and 2013 at
22 high-elevation sites in the western U.S. 1-34
Figure 1-7 Monitor locations for the warm-season and year-round data sets. 1-46
Figure 1-8 Individual monitor ozone concentrations in terms of design values for
2015-2017. 1-47
Figure 1-9 National 4th highest 8-hour daily max ozone trend and distribution
across 882 U.S. ozone monitors 2000-2017 (concentrations in ppb). 1-48
Figure 1-10 Trend in mean 4th highest 8-hour daily max ozone by U.S. region
2000-2017. 1-50
Figure 1-11 Individual monitor 3-year avg of the changes in ozone design values
from 2008-2010 to 2015-2017. 1-51
Figure 1-12 Individual monitor W126 exposure metric values for 2015-2017. 1-52
Figure 1-13 Seasonal mean daily max 8-hour avg values of North American
background (NAB) in the lowest model layer for the Geophysical Fluid
Dynamics Laboratory's AM3 (left) and Goddard Earth Observing
System (GEOS)-Chem (right) simulations for spring (March, April, and
May [MAM], top) and summer (June, July, and August [JJA], bottom)
XXV
-------
LIST OF FIGURES (Continued)
of 2006 estimated with North American anthropogenic emissions set to
zero. 1-60
Figure 1-14 Fourth highest daily max 8-hour avg (MDA8) North American
background (NAB) ozone concentration between March 1 and August
31, 2006 in the lowest model layer (top) and day of occurrence
(bottom) for the Geophysical Fluid Dynamics Laboratory's AM3 (left)
and Goddard Earth Observing System (GEOS)-Chem (right)
simulations. 1-61
Figure 1-15 The three annual 4th highest values (solid dots) used to calculate the
3-year avg of the 4th highest daily max 8-hour avg ozone (hollow
diamond) and the range of ozone concentrations for each year's 10
highest ozone concentration days (vertical bars) between March and
October in the (a) Southeast and (b) Mountains and Plains regions for
observed ozone concentration measurements (black), modeled total
ozone concentration estimates (blue), and modeled U.S. background
(USB) ozone concentration estimates (red). 1-62
Figure 1-16 Comparison of (a) Community Multiscale Air Quality (CMAQ)
zero-out and (b) Comprehensive Air Quality Model with Extensions
(CAMx) apportionment-based daily max 8-hour avg U.S. background
(USB and USBab) ozone estimates across eight model bias and two site
elevation bins. 1-64
Figure 1-17 Community Multiscale Air Quality (CMAQ) (a) and Comprehensive
Air Quality Model with Extensions (CAMx) (b) estimates of daily
distributions of bias-adjusted U.S. background (USB) daily max 8-hour
avg (MDA8) ozone concentration (parts per billion [ppb]) for the
period April-October 2007, binned by base model MDA8 ozone
concentration ranges. 1-67
Figure 1-18 Community Multiscale Air Quality (CMAQ) (a) and Comprehensive
Air Quality Model with Extensions (CAMx) (b) estimates of daily
distributions of bias-adjusted U.S. background (USB) ozone fraction at
monitoring locations across the western U.S. for the period
April-October 2007, binned by base model daily max 8-hour avg
(MDA8) ozone concentration ranges. 1-68
Figure 2-1 Year-round Pearson correlations of 8-hour daily max ozone
concentrations with copollutant concentrations measured in air quality
system 2015-2017. 2-34
Figure 2-2 Seasonal Pearson correlations of 8-hour daily max ozone
concentrations with copollutant concentrations measured in air quality
system, 2015-2017. 2-35
xxvi
-------
LIST OF FIGURES (Continued)
Figure 2-3 Year-round Pearson correlations by location setting of 8-hour daily
max ozone concentrations with copollutant concentrations measured in
air quality system, 2015-2017. 2-36
Figure 3-1 Potential biological pathways for respiratory effects following
short-term ozone exposure. 3-5
Figure 3-2 Inter-subject variability in forced expiratory volume in 1 second (FEVi)
decrements in young healthy adults following 6.6 hours of exposure to
ozone. 3-13
Figure 3-3 Cross-study comparisons of mean ozone-induced forced expiratory
volume in 1 second (FEVi) decrements in young healthy adults
following 6.6 hours of exposure to ozone. 3-14
Figure 3-4 Summary of associations from studies of short-term ozone exposures
and hospital admissions for asthma for a standardized increase in ozone
concentrations. 3-42
Figure 3-5 Summary of associations from studies of short-term ozone exposures
and asthma emergency department (ED) visits for a standardized
increase in ozone concentrations. 3-44
Figure 3-6 Summary of associations from studies of short-term ozone exposures
and respiratory infection emergency department (ED) visits for a
standardized increase in ozone concentrations. 3-64
Figure 3-7 Summary of associations from studies of short-term ozone exposures
and respiratory-related hospital admissions and emergency department
(ED) visits for a standardized increase in ozone concentrations. 3-69
Figure 3-8 Locally estimated scatterplot smoothing (LOESS) C-R estimates and
twice-standard-error estimates from generalized additive models for
associations between 8-hour max 3-day avg ozone concentrations and
emergency department (ED) visits for pediatric asthma. 3-76
Figure 3-9 Estimated relative risks (RRs) of asthma hospital admissions for 8-hour
daily max ozone concentrations at lag 0-1 allowing for possible
nonlinear relationships using natural splines. 3-77
Figure 3-10 Estimated percent change in asthma hospital admissions for 8-hour
daily avg ozone concentrations at lag 0-3 allowing for possible
nonlinear relationships using penalized splines. 3-78
Figure 3-11 LOESS C-R estimates and twice-standard-error estimates from
generalized additive models for associations between 3-day moving
avg 8-hour daily max ozone concentrations and emergency department
(ED) visits for pneumonia and upper respiratory infection. 3-79
xxvii
-------
LIST OF FIGURES (Continued)
Figure 3-12 Potential biological pathways for respiratory effects following
long-term ozone exposure. 3-91
Figure 3-13 Summary of associations from studies of long-term ozone exposure and
respiratory mortality for a standardized increase in ozone
concentrations. 3-111
Figure 4-1 Potential biological pathways for cardiovascular effects following
short-term exposure to ozone. 4-4
Figure 4-2 Associations between short-term exposure to ozone and ischemic heart
disease (IHD)-related emergency department visits and hospital
admissions. 4-11
Figure 4-3 Associations between short-term exposure to ozone and emergency
department (ED) visits and hospital admissions related to cardiac arrest,
arrhythmias, and dysrhythmias. 4-17
Figure 4-4 Associations between short-term exposure to ozone and
cerebrovascular-related emergency department visits and hospital
admissions. 4-32
Figure 4-5 Associations between short-term exposure to ozone and nonspecific
cardiovascular emergency department (ED) visits and hospital
admissions. 4-34
Figure 4-6 Potential biological pathways for cardiovascular effects following
long-term exposure to ozone. 4-48
Figure 4-7 Associations between long-term exposure to ozone and cardiovascular
mortality in recent cohort studies. 4-59
Figure 4-8 Associations between long-term exposure to ozone and cardiovascular
mortality with and without adjustment for PM2.5 concentrations in
recent cohort studies. 4-61
Figure 5-1 Potential biological pathways for metabolic outcomes following
short-term ozone exposure. 5-4
Figure 5-2 Potential biological pathways for metabolic outcomes following
long-term ozone exposure. 5-34
Figure 6-1 Summary of associations for short-term ozone exposure and total
(nonaccidental) mortality from recent multicity U.S. and Canadian
studies, and studies in previous ozone assessments. 6-7
xxviii
-------
LIST OF FIGURES (Continued)
Figure 6-2 Summary of associations for short-term ozone exposure and
cause-specific mortality from recent multicity U.S. and Canadian
studies, and studies evaluated in previous ozone assessments. 6-9
Figure 6-3 Results of a metaregression analysis in Liu et al. (2016) indicating
larger ozone-mortality risk estimates in cities with lower average
temperatures in the spring and summer seasons. 6-14
Figure 6-4 Mean log relative risk (RR) for mortality from 95 U.S. cities from the
National Morbidity, Mortality, and Air Pollution Study (NMMAPS) at
4 percentiles of temperature (50th, 75th, 95th, 99th). 6-14
Figure 6-5 Temperature-stratified ozone-mortality associations from 86 U.S. cities
within the National Morbidity, Mortality, and Air Pollution Study
(NMMAPS) using different approaches to control for nonlinearity in
temperature effects. 6-15
Figure 6-6 Flexible concentration-response relationship for short-term ozone
exposure and mortality at lag 1 for 24-hour avg ozone concentrations
adjusted by size of the bootstrap sample [size of the bootstrap (d) = 4], 6-19
Figure 6-7 Percentage increase in mortality for ozone in a two-pollutant model
with PM2.5 using penalized splines for both pollutants at lag 0-1 days
in the warm season (April-September). 6-20
Figure 6-8 Associations between long-term exposure to ozone and total
(nonaccidental) mortality in recent cohort studies. 6-29
Figure 6-9 Associations between long-term exposure to ozone and respiratory
mortality in recent cohort studies. 6-31
Figure 6-10 Associations between long-term exposure to ozone and cardiovascular
mortality in recent cohort studies. 6-33
Figure 6-11 Associations between long-term exposure to ozone and mortality with
and without adjustment for PM2.5 concentrations in recent cohort
studies. 6-37
Figure 6-12 The concentration-response relationship estimated with log-linear
model with a thin-plate spline (A; left panel) and the
concentration-response relationship estimated with threshold model (B;
right panel), indicating the potential for a threshold at 40 ppb (8-hour
daily max). 6-39
Figure 6-13 Concentration-response relationship between ozone concentrations
(parts per billion [ppb]) and total (nonaccidental) mortality in the
CanCHEC cohort (mean 39.6; knots: 30.0, 38.9, 50.7 ppb). 6-39
xxix
-------
LIST OF FIGURES (Continued)
Figure 6-14 Concentration-response curve for ozone associated with respiratory
mortality using a natural spline model with 3 degrees of freedom. 6-40
Figure 7-1 Potential biological pathways for male reproduction and fertility effects
following ozone exposure. 7-5
Figure 7-2 Potential biological pathways for pregnancy and birth outcomes
following ozone exposure. 7-9
Figure 7-3 Potential biological pathways for nervous system effects following
short-term exposure to ozone. 7-23
Figure 7-4 Results of studies of short-term ozone exposure and hospital
admissions or emergency department visits for diseases of the nervous
system or mental health. 7-30
Figure 7-5 Potential biological pathways for nervous system effects following
long-term exposure to ozone. 7-35
Figure 8-1 Illustrative diagram of ozone effects in plants and ecosystems adapted
from the 2013 Ozone ISA. 8-2
Figure 8-2 Schematic representation of the cellular and metabolic effects of ozone
on vegetation. 8-12
Figure 8-3 Meta-analysis of the effects of ozone exposure (relative to
charcoal-filtered air) on plant reproduction. 8-45
Figure 8-4 Meta-analysis of the effects of ozone exposure (relative to ambient air)
on plant reproduction. 8-45
Figure 8-5 Estimated percentage reduction of soybean and maize yield in the U.S.
from ozone for 1980-2011. 8-64
Figure 8-6 Conceptual model of ozone effects on herbivore growth, reproduction,
and survival. 8-72
Figure 8-7 Conceptual model of ozone effects on volatile plant signaling
compounds and plant-insect signaling. 8-89
Figure 8-8 Conceptual diagram of ozone effects on belowground processes and
biogeochemical cycles. 8-116
Figure 8-9 Conceptual diagram illustrating effects of elevated ozone exposure on
N-inputs and outputs. Elevated ozone increased soil N, increased
gaseous N loss, and reduced grain N. 8-120
Figure 8-10 Mechanisms by which ozone alters plant communities. 8-131
XXX
-------
LIST OF FIGURES (Continued)
Figure 8-11 Relationship between ozone concentrations and cover of grass
A. odoratum grown in competition with forb L. hispidus (top graph),
and cover of grass D. glomerata grown in competition with forb
L. hispidus. 8-135
Figure 8-12 Biological plausibility of ozone effects on soil microbial communities
and soil invertebrate communities. 8-142
Figure 8-13 Percentage change of modeled net carbon dioxide (CO2) assimilation,
transpiration, and water use efficiency in temperate deciduous forests in
the Northern Hemisphere in relation to daytime mean ozone
concentration or cumulative canopy ozone uptake (years 2006-2009).
(a) Net CO2 assimilation, (b) transpiration, and (c) water use efficiency
were simulated by the offline coupling simulation of SOLVEG-MRI-
CCM2. 8-164
Figure 8-14 Quantiles of predicted relative biomass loss for four tree species in
NHEERL-WED experiments. Quantiles of the predicted relative
aboveground biomass loss at seven exposure values of 12-hour W126
for Weibull curves estimated using nonlinear regression on data for
four tree species grown under well-watered conditions for 1 or 2 years. 8-189
Figure 8-15 Relative biomass loss predictions from Weibull exposure-response
functions that relate percent aboveground biomass loss to 12-hour
W126 exposures adjusted to 92 days. 8-191
Figure 8-16 Comparison of composite functions for the quartiles of 7 curves for
7 genotypes of soybean grown in the SoyFACE experiment, and for the
quartiles of 11 curves for 5 genotypes of soybean grown in the NCLAN
project. 8-192
Figure 8-17 Comparison between aboveground biomass observed in Aspen FACE
experiment in 6 years and biomass predicted by the median composite
function based on NHEERL-WED. 8-193
Figure 9-1 Schematic diagram of the effects of tropospheric ozone on climate. 9-6
Figure 9-2 Bar chart for radiative forcing (RF; hatched) and effective radiative
forcing (ERF; solid) for the period 1750-2011. 9-10
Figure 9-3 Radiative forcing (RF) over the industrial era associated with emitted
compounds, including ozone (green bars) and its precursors. 9-12
Figure 9-4 Time evolution of the radiative forcing (RF) from tropospheric ozone
from 1750 to 2010. 9-13
xxxi
-------
LIST OF FIGURES (Continued)
Figure 9-5 Radiative forcing (RF) spatial distribution of 1850 to 2000 ozone RF
among the atmospheric chemistry and Climate Model Intercomparison
Project models, mean values (left) and standard deviation (right). 9-14
Figure 9-6 Difference in annual average radiative forcing (W/m2) between
modeled (GISS-E2-R) and observed Tropospheric Emission
Spectrometer present-day (2005-2009) total natural plus anthropogenic
ozone throughout the atmosphere. 9-15
Figure 9-7 Mean annual change in surface temperature (°C) resulting from
tropospheric ozone concentration changes from 1850-2013. 9-19
Figure 9-8 Mean annual change in precipitation (mm/day) resulting from
tropospheric ozone concentration changes from 1850-2013. 9-21
Figure 10-1 General process for development of Integrated Science Assessments. 10-3
Figure 10-2 Literature flow diagram for the Ozone Integrated Science Assessment. 10-4
Figure 10-3 Summary of title/abstract screening in SWIFT-ActiveScreener. 10-7
Figure 10-4 Illustrative example of screening efficiency using
SWIFT-ActiveScreener. 10-7
xxxii
-------
INTEGRATED SCIENCE ASSESSMENT TEAM FOR OZONE
AND RELATED PHOTOCHEMICAL OXIDANTS
Executive Direction
Dr. John Vandenberg (Director)—Health and Environmental Effects Assessment Division,
Center for Public Health and Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Steven J. Dutton (Associate Director)—Health and Environmental Effects Assessment
Division, Center for Public Health and Environmental Assessment, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Jane Ellen Simmons (Branch Chief)—Health and Environmental Effects Assessment
Division, Center for Public Health and Environmental Assessment, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Christopher Weaver (Branch Chief)—Health and Environmental Effects Assessment
Division, Center for Public Health and Environmental Assessment, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Jennifer Nichols (Acting Branch Chief)—Health and Environmental Effects Assessment
Division, Center for Public Health and Environmental Assessment, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Tara Greaver (Acting Branch Chief)—Health and Environmental Effects Assessment
Division, Center for Public Health and Environmental Assessment, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Andrew Hotchkiss (Acting Branch Chief)—Health and Environmental Effects
Assessment Division, Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Technical Support Staff
Ms. Marieka Boyd—Health and Environmental Effects Assessment Division, Center for
Public Health and Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Ms. Annamarie Cory—Oak Ridge Associated Universities, Health and Environmental
Effects Assessment Division, Center for Public Health and Environmental Assessment,
Office of Research and Development, U.S. Environmental Protection Agency, Research
Triangle Park, NC
Ms. Madison Feshuk—Oak Ridge Associated Universities, Health and Environmental
Effects Assessment Division, Center for Public Health and Environmental Assessment,
Office of Research and Development, U.S. Environmental Protection Agency, Research
Triangle Park, NC
xxxiii
-------
INTEGRATED SCIENCE ASSESSMENT TEAM FOR OZONE AND RELATED
PHOTOCHEMICAL OXIDANTS (Continued)
Ms. Erin Gallagher—Oak Ridge Associated Universities, Health and Environmental Effects
Assessment Division, Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Ms. Beth Gatling—Oak Ridge Associated Universities, Health and Environmental Effects
Assessment Division, Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Ms. Hillary Hollinger—Oak Ridge Associated Universities, Health and Environmental
Effects Assessment Division, Center for Public Health and Environmental Assessment,
Office of Research and Development, U.S. Environmental Protection Agency, Research
Triangle Park, NC
Mr. Ryan Jones—Health and Environmental Effects Assessment Division, Center for Public
Health and Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Mr. Lukas Kerr—Oak Ridge Associated Universities, Health and Environmental Effects
Assessment Division, Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Ms. Emma Leath—Oak Ridge Associated Universities, Health and Environmental Effects
Assessment Division, Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Ms. Mckayla Lein—Oak Ridge Associated Universities, Health and Environmental Effects
Assessment Division, Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Ms. Danielle Moore—Senior Environmental Employment Program, Health and
Environmental Effects Assessment Division, Center for Public Health and Environmental
Assessment, Office of Research and Development, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Mr. Kevin Park—Oak Ridge Associated Universities, Health and Environmental Effects
Assessment Division, Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Mr. R. Byron Rice—Oak Ridge Associated Universities, Health and Environmental Effects
Assessment Division, Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Mr. Brayndon Stafford—Oak Ridge Associated Universities, Health and Environmental
Effects Assessment Division, Center for Public Health and Environmental Assessment,
Office of Research and Development, U.S. Environmental Protection Agency, Research
Triangle Park, NC
xxxiv
-------
INTEGRATED SCIENCE ASSESSMENT TEAM FOR OZONE AND RELATED
PHOTOCHEMICAL OXIDANTS (Continued)
Mr. S. Shane Thacker—Health and Environmental Effects Assessment Division, Center for
Public Health and Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Ms. Erin Vining—Oak Ridge Associated Universities, Health and Environmental Effects
Assessment Division, Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
XXXV
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS
Authors
Dr. Thomas Luben*' (Health Assessment Team Lead, Integrated Science Assessment for
Ozone)—Health and Environmental Effects Assessment Division, Center for Public
Health and Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Meredith Lassiter* (Welfare Assessment Team Lead, Integrated Science Assessment for
Ozone)—Health and Environmental Effects Assessment Division, Center for Public
Health and Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Ms. Rebecca Daniels' (Project Manager, Integrated Science Assessment for Ozone)—Center
for Computational Toxicology and Exposure, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Mr. Lance Avery—Region 7 U.S. Environmental Protection Agency, Kansas City, KS
Ms. Michelle Becker—Region 5, U.S. Environmental Protection Agency, Chicago, IL
Dr. James Brown*—Health and Environmental Effects Assessment Division, Center for
Public Health and Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Barbara Buckley*—Health and Environmental Effects Assessment Division, Center for
Public Health and Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Mr. Evan Coffman*'—Health and Environmental Effects Assessment Division, Center for
Public Health and Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Laura Dishaw*—Chemical and Pollutant Assessment Division, Center for Public Health
and Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Brian Eder—Center for Environmental Measurement and Modeling, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Emmi Felker-Quinn*—former U.S. Environmental Protection Agency, Research
Triangle Park, NC
Dr. Meridith Fry*—Health and Environmental Effects Assessment Division, Center for
Public Health and Environmental Assessment, Office of Research and Development, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Dr. Barbara Glenn*—Chemical and Pollutant Assessment Division, Center for Public Health
and Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Tara Greaver*—Health and Environmental Effects Assessment Division, Center for
Public Health and Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
xxxvi
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS (Continued)
Dr. Brooke Hemniing*'—Health and Environmental Effects Assessment Division, Center
for Public Health and Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Jeffrey D. Herrick*'—Health and Environmental Effects Assessment Division, Center
for Public Health and Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Erin Hines*'—Public Health and Integrated Toxicology Division, Center for Public
Health and Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. S. Douglas Kaylor*—Health and Environmental Effects Assessment Division, Center
for Public Health and Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Ellen Kirrane*'—Health and Environmental Effects Assessment Division, Center for
Public Health and Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Robert Kotchenruther—Region 10, U.S. Environmental Protection Agency, Seattle, WA
Dr. David Lehmann—Health and Environmental Effects Assessment Division, Center for
Public Health and Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Jennifer Liljegren—Region 5, U.S. Environmental Protection Agency, Chicago, IL
Dr. Rebecca Matichuk—Region 8, U.S. Environmental Protection Agency, Denver, CO
Dr. Steve McDow *'—Health and Environmental Effects Assessment Division, Center for
Public Health and Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Jennifer Nichols*—Health and Environmental Effects Assessment Division, Center for
Public Health and Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Robert Pinder—Office of Air Quality Planning and Standards, Office of Air and
Radiation, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Kristen Rappazzo—Public Health and Environmental Systems Division, Center for
Public Health and Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Jeanette Reyes—Office of Air Quality Planning and Standards, Office of Air and
Radiation, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Jennifer R i c h m o n d - B ry an t *'—former U.S. Environmental Protection Agency, Research
Triangle Park, NC
Dr. Caroline Ridley*—Health and Environmental Effects Assessment Division, Center for
Public Health and Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Mr. Jason Sacks*—Health and Environmental Effects Assessment Division, Center for
Public Health and Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
xxxvii
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS (Continued)
Dr. Michael Stewart*'—Health and Environmental Effects Assessment Division, Center for
Public Health and Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. James Szykman—Center for Environmental Measurement and Modeling, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Dr. Alan Talhelm—Department of Forest, Rangeland, and Fire Sciences, University of
Idaho, Moscow, ID
Dr. Gail Tonnesen—Region 8, U.S. Environmental Protection Agency, Denver, CO
Dr. Lukas Valin—Center for Environmental Measurement and Modeling, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Dr. Christopher Weaver* ^—Health and Environmental Effects Assessment Division, Center
for Public Health and Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Chelsea Weitekamp*—Health and Environmental Effects Assessment Division, Center
for Public Health and Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
'Health and Environmental Effects Assessment Division Scientific Staff.
tAppendix Lead.
Contributors
Dr. Jean Jacques Dubois—Independent Consultant, Raleigh, NC
Dr. Kristopher Novak—Independent Consultant, Harrisburg, PA
Mr. Benjamin Wells—Office of Air Quality, Planning, and Standards, Office of Air and
Radiation, U.S. Environmental Protection Agency, Research Triangle Park, NC
Reviewers
Dr. Elizabeth Ainsworth—U.S. Department of Agriculture, Urbana, IL; School of
Integrative Biology, University of Illinois, Urbana, IL
Dr. Chris Andersen—Pacific Ecological Systems Division, Center for Public Health and
Environmental Assessment, U.S. Environmental Protection Agency, Corvallis, OR
Dr. Michelle Angrish—Chemical and Pollutant Assessment Division, Center for Public
Health and Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Michael Bell—Air Resources Division, U.S. National Resource Stewardship and
Science, National Park Service, Lakewood, CO
Dr. Kent Burkey—U.S. Department of Agriculture, Raleigh, NC; Department of Crop and
Soil Sciences, College of Agriculture and Life Sciences, North Carolina State University,
Raleigh, NC
Dr. Alex Carll—Department of Physiology, School of Medicine, University of Louisville,
Louisville, KY
xxxviii
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS (Continued)
Dr. Laura Carlson—Health and Environmental Effects Assessment Division, Center for
Public Health and Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. John Couture—Department of Forestry and Natural Resources, College of Agriculture,
Purdue University, West Lafayette, IN
Dr. James Crooks—Division of Biostatistics and Bioinformatics, National Jewish Health,
Denver, CO
Dr. Stefanie Deflorio-Barker—Health and Environmental Effects Assessment Division,
Center for Public Health and Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. David DeMarini—Center for Computational Toxicology and Exposure, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Dr. Kathie Dionisio—Center for Computational Toxicology and Exposure, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Dr. Jan Dye—Public Health and Integrated Toxicology Division, Center for Public Health
and Environmental Assessment, U.S. Environmental Protection Agency, Research
Triangle Park, NC
Dr. Aimen Farraj—Public Health and Integrated Toxicology Division, Center for Public
Health and Environmental Assessment, Office of Research and Development, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Dr. Arlene Fiore—Department of Earth and Environmental Sciences, Columbia University,
Palisades, NY
Dr. William Gauderman—Keck School of Medicine, University of Southern California, Los
Angeles, CA
Dr. Ian Gilmour—Public Health and Integrated Toxicology Division, Center for Public
Health and Environmental Assessment, Office of Research and Development, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Dr. Mehdi Hazari—Public Health and Integrated Toxicology Division, Center for Public
Health and Environmental Assessment, Office of Research and Development, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Dr. Christian Hogrefe—Center for Environmental Measurement and Modeling, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Dr. Kazuhiko Ito—Bureau of Environmental Surveillance and Policy, New York City
Department of Health and Mental Hygiene, New York, NY
Dr. Daniel Jaffe—School of Science, Technology, Engineering, and Mathematics,
University of Washington Bothell, Bothell, WA
Dr. Annie Jarabek—Health and Environmental Effects Assessment Division, Center for
Public Health and Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
xxxix
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS (Continued)
Dr. Terry Keating—Office of Air and Radiation, U.S. Environmental Protection Agency,
Washington, DC
Dr. Travis Knuckles—Department of Occupational and Environmental Health Sciences,
School of Public Health, West Virginia University, Morgantown, WV
Dr. Urmila Kodavanti—Public Health and Integrated Toxicology Division, Center for Public
Health and Environmental Assessment, Office of Research and Development, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Dr. Andrew Langford—Chemical Sciences Division, Earth System Research Laboratory,
National Oceanic and Atmospheric Administration, Boulder, CO
Dr. Steve LeDuc—Health and Environmental Effects Assessment Division, Center for
Public Health and Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Danielle Lobdell—Public Health and Environmental Systems Division, Center for
Public Health and Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Tom Long—Center for Public Health and Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Dr. Elizabeth Mannshardt—Office of Air Quality Planning and Standards, Office of Air and
Radiation, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Loretta Mickley—School of Engineering and Applied Sciences, Harvard University,
Cambridge, MA
Dr. Colette Miller—Center for Computational Toxicology and Exposure, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Leigh Moorhead—Health and Environmental Effects Assessment Division, Center for
Public Health and Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Anu Mudipalli—Health and Environmental Effects Assessment Division, Center for
Public Health and Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Chris Nolte—Center for Environmental Measurement and Modeling, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. David Parrish—Air Quality Research Center, University of California Davis, Davis, CA
Dr. David Peden—Gillings School of Public Health, University of North Carolina at Chapel
Hill, Chapel Hill, NC
Dr. Melinda Power—School of Public Health, Department of Epidemiology and
Biostatistics, The George Washington University, Washington, DC
Dr. Armistead Russell—School of Civil and Environmental Engineering, College of
Engineering, Georgia Institute of Technology, Atlanta, GA
Dr. Edward Schelegle—Department of Anatomy, Physiology, and Cell Biology, School of
Veterinary Medicine, University of California Davis, Davis, CA
xl
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS (Continued)
Dr. John Stanek—Health and Environmental Effects Assessment Division, Center for Public
Health and Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Matthew Strickland—Department of Environmental Health, School of Community
Health Sciences, University of Nevada, Reno, Reno, NV
Dr. Jennifer Weuve—School of Public Health, Boston University, Boston, MA
Dr. Dongni Ye—Oak Ridge Institute for Science and Education, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
xli
-------
CLEAN AIR SCIENTIFIC ADVISORY COMMITTEE
Chair of the Clean Air Scientific Advisory Committee
Dr. Louis Anthony (Tony) Cox, Jr - Cox Associates, Denver, CO
Clean Air Scientific Advisory Committee Members
Dr. James Boylan - Georgia Department of Natural Resources, Atlanta, GA
Dr. Mark W. Frampton - University of Rochester Medical Center, Rochester, NY
Dr. Ronald J. Kendall - Texas Tech University, Lubbock, TX
Dr. Sabine Lange - Texas Commission on Environmental Quality, Austin, TX
Dr. Corey M. Masuca - Jefferson County Department of Health, Birmingham, AL
Dr. Steven C. Packham - Utah Department of Environmental Quality, Salt Lake City, UT
Scientific Advisory Board Staff
Mr. Aaron Yeow (Designated Federal Officer) - U.S. Environmental Protection Agency,
Science Advisory Board, Washington DC
xlii
-------
ACRONYMS AND ABBREVIATIONS
Acronym/
Abbreviation
13C
15N
3D
3-PG
5HT
5HT2aR
5HT4R
5HTT
Ale
ABA
AC
ACCMtP
ACE-1
ACM2
ACS
ACTH
ADA
ADAM
ADF
ADL
ADMS
Meaning
carbon-13
nitrogen-15, stable isotope of nitrogen
three-dimensional
Physiological Principles Predicting
Growth, a process-based stand level
model which is a transition model
between growth and carbon balance
models
5-hydroxytryptamine (serotonin)
5-hydroxytryptamine (serotonin)
receptor 2a
5-hydroxytryptamine (serotonin)
receptor 4
5-hydroxytryptamine (serotonin)
transporter
glycated hemoglobin
abscisic acid
air conditioning
Atmospheric Chemistry and Climate
Model Intercomparison Project
angiotensin converting enzyme
Asymmetric Convective Model
Version 2
American Cancer Society
adrenocorticotropic hormone
Americans with Disabilities Act
a disintegrin and metalloproteinase
acid digestible fiber
acid digestible lignin
Atmospheric Dispersion Modeling
System
Acronym/
Abbreviation Meaning
ADMS-Urban Atmospheric Dispersion Modeling
System-Urban
ADOS Autism Diagnostic Observation Schedule
ADREX, ADX adrenalectomy
AER air exchange rate
AERMtC AMS/EPA Regulatory Model
Improvement Committee
AERMOD AERMIC Model
AER05 fifth-generation modal CMAQ aerosol
model with extensions
AF atrial fibrillation
AHI apnea-hypopnea index
AHR airway hyperresponsiveness, aryl
hydrocarbon receptor
AHS Agricultural Health Study
AIC Akaike's information criterion
AIRDATA (U.S. EPA) air quality data collected at
outdoor monitors across the U.S.
AIRPACT-3 Air Indicator Report for Public
Awareness and Community Tracking
Version 3
AIRS Aerometric Information Retrieval
System
AK averaging kernel
AKT protein kinase B (PKB), also known as
Akt, is a serine/threonine-specific protein
kinase
ALRI acute lower respiratory infection
AM alveolar macrophage
AM3 Atmospheric Model 3 (chemical
transport model from Geophysical Fluid
Dynamics Laboratory)
xliii
-------
Acronym/
Abbreviation Meaning
AMI acute myocardial infarction
AMO Atlantic Multidecadal Oscillation
AMPK adenosine monophosphate-activated
protein kinase
AMS American Meteorological Society
AMX adrenal medullectomy
ANPP annual net primary productivity
AO Arctic Oscillation
AOD aerosol optical density
AOTO seasonal sum of the difference between
an hourly concentration at the threshold
value of 0 ppb, minus the threshold value
of 0 ppb
AOT40 seasonal sum of the difference between
an hourly concentration at the threshold
value of 40 ppb, minus the threshold
value of 40 ppb
AOT60 seasonal sum of the difference between
an hourly concentration at the threshold
value of 60 ppb, minus the threshold
value of 60 ppb
AOTx sum of the differences between hourly
concentrations greater than a specified
threshold (x) during a specified daily and
seasonal time window
APEX Air Pollutants Exposure Model
APHEA Air Pollution and Health: A European
Approach
APHEA2 Air Pollution and Health: A European
Approach 2
APHENA Air Pollution and Health: A Combined
European and North American Approach
API Air Pollution Index
APMoSPHERE Air Pollution Modeling for Support to
Policy on Health and Environmental
Risk in Europe
ApoAl apolipoprotein A-I
Acronym/
Abbreviation Meaning
ApoB apolipoprotein B
ApoE- apolipoprotein E-deficient
APP+ amyloid beta precursor protein
APT Advanced Plume Treatment
AQCD Air Quality Criteria Document
AQMEII Air Quality Model Evaluation
International Initiative
AQS U.S. EPA Air Quality System database
AR airway responsiveness
AR4 Fourth Assessment Report from the
IPCC
AR5 Fifth Assessment Report from the IPCC
ARDS adult respiratory distress syndrome
ARG arginase
ARI acute respiratory infection
ARS Air Resource Specialists
ASD autism spectrum disorder
ASHRAE American Society of Heating,
Refrigeration, and Air-Conditioning
Engineers
AT apparent temperature
ATA American Trucking Association
ATPase adenosine triphosphotase
AUC area under the curve
AUP unpaired predicted-to-observed peak
ozone ratio
AURAMS Unified Regional Air Quality Modeling
System
AV atrioventricular
AVCD atrioventricular conduction disorders
avg average
xliv
-------
Acronym/
Abbreviation Meaning
BAD bronchial artery diameter
BAL bronchoalveolar lavage
BALF bronchoalveolar lavage fluid
BAMSE Children, Allergy, Milieu, Stockholm,
Epidemiology (Swedish Abbreviation)
BASIC Brain Attack Surveillance in Corpus
Christi
BC black carbon
BCSC Breast Cancer Surveillance Consortium
BEIS Biogenic Emission Inventory System
BELD Biogenic Emissions Land Use Database
BME Bayesian maximum entropy
BMI body mass index
BN brown Norway rat strain
BP blood pressure
bpm beats per minute
BRAVO Big Bend Regional Aerosol and
Visibility Observational
BrdU bromodeoxyuridine
BSA body surface area
BUN blood urea nitrogen
BVAIT B-Vitamin Atherosclerosis Intervention
Trial
BVOC biogenic volatile organic compound
BW birth weight; body weight
BWHS Black Women's Health Study
C carbon; degrees Celsius; the product of
microenvironmental concentration
C3 plants that only use the Calvin cycle for
fixing the carbon dioxide from the air
Acronym/
Abbreviation Meaning
C4 plants that use the Hatch-Slack cycle for
fixing the carbon dioxide from the air
C57BL wild type c57 black mouse strain
Ca ambient ozone concentration
CAA Clean Air Act
CAD coronary artery disease
CALGRID California Grid Simulations
CalNEX California Research at the Nexus of Air
Quality and Climate Change
CAM Community Atmosphere Model; plants
that use crassulacean acid metabolism
for fixing the carbon dioxide from the air
CAMP Constant Air Quality Model
Performance
CAMS Continuous Monitoring Station
CAMx Comprehensive Air Quality Model with
extensions
CanCHEC Canadian Census Health and
Environment Cohort
CAP concentrated ambient particle
CAPPS Canadian Asthma Primary Prevention
Study
CARB California Air Resources Board
CASAC Clean Air Scientific Advisory
Committee
CASTNet Clean Air Status and Trends Network
CAT catalase
CATHGEN catheterization genetics
CB05 carbon bond mechanism developed in
2005
CBVD cerebrovascular disease
CCL chemokine ligand
CCR2 CC chemokine receptor type 2
xlv
-------
Acronym/
Abbreviation Meaning
CCSP U.S. Climate Change Science Program,
forerunner to the U.S. Global Change
Research Program; club cell secretory
protein
CD cluster of differentiation; confidence
distribution
CD36 cluster of differentiation 36
CD4 cluster of differentiation 4
cd56 neural cell adhesion molecule
CDC Centers for Disease Control and
Prevention
CDR Clinical Dementia Rating Sum of Boxes
CEM continuous emissions modeling
CF charcoal filter; carbon filtered
CFR Code of Federal Regulations
CH2O carbohydrate
CH4 methane
CHAD Consolidated Human Activity Database
CHARGE Childhood Autism Risks from Genetics
and the Environment
CHASER CHemical Atmospheric general
circulation model for Study of
atmospheric Environment and Radiative
forcing (chemical transport model)
CHD chronic heart disease; coronary heart
disease
CHE controlled human exposure
CHF congestive heart failure
Chil4 chitinase-like protein 4
CHIMERE a regional chemical transport model.
Multiscale Chemical Transport Model
for Atmospheric Composition Analysis
and Forecast
CHRONOS Canadian Hemispheric and Regional
Ozone and NOx System
Acronym/
Abbreviation Meaning
CHS Children's Health Study
CI confidence interval
Ci intracellular carbon dioxide; substrate
concentrations
CIMT carotid intima-medial thickness
CL critical level
CI2 chlorine (gas)
Clcal chloride channel accessory 1
CLM chemiluminescence method
CINO2 nitryl chloride
CLPP community-level physiological profile
cm centimeter
CM conditioned medium
CMAQ Community Multiscale Air Quality
modeling system
CMAQ-HBM Community Multiscale Air
Quality-Hierarchical Bayesian Model
CMIP6 Coupled Model Intercomparison Project
Phase 6
CMP central mean arterial pressure
C:N carbon nitrogen ratio
CNS central nervous system
CO carbon monoxide
CO2 carbon dioxide
COFI COhorte Fibrose
CONUS continental U.S.
COPD chronic obstructive pulmonary disease
CP coverage prediction interval
CPC condensation particle counter
xlvi
-------
Acronym/
Abbreviation Meaning
CPHEA Center for Public Health and
Environmental Assessment
cPOM course particulate organic matter
CPSII (ACS) Cancer Prevention Study II
CRH corticotropin-releasing hormone
CRP C-reactive protein
CSAPR Cross-State Air Pollution Rule
CSI critical success index
CSS calculated severity score (ADOS-CSS)
CSTR continuous stirred tank reactor
CTM chemical transport model
CV cardiovascular; coefficient of variation
CVD cardiovascular disease
CX3CL1 fractalkine
CX3CR1 fractalkine receptor
CXC cys-xxx-cys—(amino acid motif)
CXCL chemokine family of cytokines with
highly conserved motif: cys-xxx-cys
(CXC)
CXCR receptor for chemokine family of
receptors
Cypb5 cytochrome p450 b5
D dose
DA24 daily 24-hour avg concentration
db decibel
DBP diastolic blood pressure
DC3 Deep Convective Cloud and Chemistry
(field study)
DDM Decoupled Direct Method
DDS Department of Developmental Services
Acronym/
Abbreviation Meaning
DECSO daily emission estimates constrained by
satellite observations (algorithm)
DEHM Danish Eulerian Hemispheric Model
DEMED adrenal demedullation
df degrees of freedom
dg decigram
DISCOVER-AQ Deriving Information on Surface
Conditions from COlumn and
VERtically resolved observations
relevant to Air Quality
DL distributed lag
dL deciliter
DLEM Dynamic Land Ecosystem Model
DLNM distributed lag nonlinear model
DM dry moderate; dry matter
DM8H 8-hour daily max ozone
DNA deoxyribonucleic acid
DOC dissolved organic carbon
DOE Department of Energy
DOHaD Developmental Origins of Health and
Disease
dP change in pressure
DP dry polar
DSM-IV-R Diagnostic and Statistical Manual of
Mental Disorders, 4th edition Revised
DT dry tropical
Ea ambient ozone exposure
EBC exhaled breath condensate
EC elemental carbon
ECG electrocardiogram
ED emergency department
xlvii
-------
Acronym/
Abbreviation Meaning
EDGAR Emissions Database for Global
Atmospheric Research
EDMUS European Database for Multiple
Sclerosis
EGAS Economic Growth Analysis System
EGF epidermal growth factor
EGU electricity-generating unit
EH redox potential
EI emissions inventory
EIB emissions-influenced background
ELITE Early Versus Late Intervention Trial with
Estradiol
EMEP European Monitoring and Evaluation
Program
EMF ectomycorrhizal fungal
EMS emergency medical service
EnKF ensemble Kalman filter
eNO exhaled nitric oxide
eNOS endothelial nitric oxide synthase
ENSO El Nino Southern Oscillation
E-O3 elevated ozone
EPA U. S. Environmental Protection Agency
ER emergency room; estrogen receptor
ERF effective radiative forcing
ET-1 endothelin-1
EUgrow forest productivity model
F344 Fisher 344 rat strain
FA filtered air; adjusted forcings; fatty acid
FACE Free-Air Carbon Dioxide Enrichment
FACS florescence activated cell sorting
Acronym/
Abbreviation Meaning
FAR false alarm ratio
FB fractional bias
FBG fasting blood glucose
FDDA four-dimensional data assimilation
FE fractional error
FEF25-75 mean forced expiratory flow over the
middle half of the forced vital
FEM Federal Equivalent Method
FeNO fractional exhaled nitric oxide
FEPS Fire Emissions Production Simulator
FEVi forced expiratory volume in 1 second
FFA free fatty acid
FHH fawn-hooded hypertensive
FIA U.S. Department of Agriculture F orest
Service Forest Inventory and Analysis
Program
FLN flower number
FMD flow-mediated dilation
FN fruit number
FOR fecundability odds ratio
Fp percentage of cases where simulation
results were close to observations
FPG fasting plasma glucose
fPOM fine particulate organic matter
FR Federal Register; fecundity risk
FRM Federal Reference Method
FW fruit weight
g gram
GAM generalized additive model
GB gross bias
xlviii
-------
Acronym/
Abbreviation Meaning
GCM general circulation model
GD gestational day
GDAS Global Data Assimilation System
GDM gestational diabetes mellitus
GE gross error
GEM-MACH Global Environmental Multiscale
coupled with Model of Air quality and
CHemistry
GEMS Global and regional Earth-system
Monitoring using Satellite and in-situ
data
GEOS Goddard Earth Observing System
GEOS-Chem Goddard Earth Observing System with
global chemical transport model
GFAP glial fibrillatory acidic protein
GFDL Geophysical Fluid Dynamics Laboratory
GGT gamma glutamyl transferase
GE1G greenhouse gas
GINI German Infant Nutritional Intervention
study
GINIplus German Infant Nutritional Intervention
plus environmental and genetic
influences
GIS geographic information system
GISS Goddard Institute for Space Studies
GLM generalized linear model
GLP good laboratory practice
GLV green leaf volatile
GMRF Gaussian Markov random field
GOES Geostationary Operational
Environmental Satellite
Acronym/
Abbreviation Meaning
GOME2a Global Ozone Monitoring Experiment
2A (instrument system borne on the
European Remote Sensing Satellite)
GPP gross primary productivity
GPS global positioning system
GPx glutathione peroxidase
GRP gastrin-releasing peptide
GRPR gastrin-releasing peptide receptor
Gs stomatal conductance
GSH reduced glutathione
GSH/GSSG ratio of reduced to oxidized glutathione
GSS gas-sensitive semiconducting oxide;
glutathione synthetase
GSSG oxidized glutathione
GST glutathione S-transferase
GSTM1 glutathione S-transferase polymorphism
mu 1
GSTP1 glutathione S-transferase pi gene
GTT glucose tolerance test
GVRD Greater Vancouver Regional District
h hour
H2O water
H2O2 hydrogen peroxide
HA hospital admission
HAWC Health Assessment Workspace
Collaborative
HbA 1 c glycated hemoglobin
HBM hierarchical Bayesian model
HCHO formaldehyde
xlix
-------
Acronym/
Abbreviation Meaning
HDDM higher order decoupled direct method;
Hierarchical Bayesian Diffusion Drift
Model
HDL high-density lipoprotein
HDM house dust mite
HDMA house dust mite allergen
HEEAD Health and Environmental Effects
Assessment Division
HEI Health Effects Institute
HERO Health and Environmental Research
Online
HF high-frequency component of HRV;
heart failure; high fat
HFD high-fat diet
HFr right heart failure
Hg mercury
HGB Houston-Galveston-Brazoria
HIRA Health Insurance Review and
Assessment Service
HISA Highly Influential Scientific Assessment
HMOX heme oxygenase
HMS Hazard Mapping System; Hospital
Morbidity Survey
HNO3 nitric acid
HO2 hydroperoxyl radical
HOC1 hypochlorous acid
HOMA homeostatic model assessment
HOMA-IR Homeostatic Model Assessment of
Insulin Resistance
HOMOVA homogeneity of molecular variance
HONO nitrous acid
Acronym/
Abbreviation Meaning
HOx hydrogen and oxygen containing radicals
(sum of hydroxy 1 and hydroperoxyl
radicals)
HPA hypothalamic-pituitary-adrenal axis
HR heart rate; hazard ratio
HRV heart rate variability
HSP70 heat shock protein 70
HTAP Hemispheric Transport of Air Pollutants
HVAC heating, ventilation, and air conditioning
HYSPLIT Hybrid Single Particle Lagrangian
Integrated Trajectory
ICAM intercellular adhesion molecule
ICARTT International Consortium for
Atmospheric Research on Transport and
Transformation
ICC interclass correlation coefficient
ICD International Classification of Diseases;
implantable cardioverter defibrillator
ICD 10 International Classification of
Diseases—version 10
ICD9 International Classification of
Diseases—version 9
ICU intensive care unit
IDW inverse-distance weighting
IFN interferon
IGAC International Global Atmospheric
Chemistry
IgE immunoglobulin E
IHD ischemic heart disease
IL interleukin
ILC immune lymphoid cell
IMPROVE Interagency Monitoring of Protected
Visual Environments
1
-------
Acronym/
Abbreviation Meaning
iNOS inducible nitric oxide synthase
IO iodine monoxide
IOA index of agreement
IOM Institute of Medicine
IP inhalable particle
i.p. intraperitoneal injection
IPCC Intergovernmental Panel on Climate
Change
IQR interquartile range
IR infrared; incidence rate
IRB Institutional Review Board
IRP Integrated Review Plan
ISA Integrated Science Assessment
ISAM Integrated Source Apportionment
Method (in CMAQ)
1ST insulin sensitivity test
IT intratracheal instillation
ITT insulin tolerance test
IVDMD in vitro dry matter digestibility
IVF in vitro fertilization
IVND in vitro nitrogen digestibility
JA jasmonic acid
JCR A/JCr mouse strain
JNK c-Jun N-terminal kinase
K2SO4 potassium sulfate-measurement of
extractable soil carbon
KC a local neutrophil chemoattractant
protein
KCLurban King's College London urban
KEEP Kidney Early Evaluation Program
Acronym/
Abbreviation Meaning
kg kilogram
KK KK mouse strain
KKAy KKAy mouse strain
km kilometer
kPa kilopascal
KROFEX Kranzberg Ozone Fumigation
Experiment
LAEI large artery elasticity index
LAI leaf area index
LANDIS forest landscape model
lbs pounds
LC50 median lethal concentration
LDH lactate dehydrogenase
LDL low-density lipoprotein
LE Lake Elsinore; Long-Evans rat strain
LF low-frequency component of HRV
LHID2000- Longitudinal Health Insurance Database
NHIRD 2000—National Health Insurance
Research Database
LIF leukemia inhibitory factor
LIFE Longitudinal Investigation of Fertility
and the Environment
LISA Lifestyle Factors on the Development of
the Immune System and Asthma
LISAplus Lifestyle Factors on the Development of
the Immune System and Asthma plus Air
Pollution and Genetics on Allergy
Development
LNOx nitrogen oxides generated by lightning
LOOCV leave-one-out cross validation
LOESS Locally Estimated Scatterplot Smoothing
li
-------
Acronym/
Abbreviation Meaning
LOTOS-EUROS Long-Term Ozone Simulation-European
Operational Smog
LPS lipopolysaccharide
LT50 median lethal time
LTB4 leukotriene B4
LUR land use regression
LV left ventricle
LVDP left ventricular developed pressure
LVOS Las Vegas Ozone Study
LWRE longwave radiative effect
m meter
Ml month 1
M2 month 2
M3 month 3
M7 7-hour seasonal mean
MA moving average
MADRID Model of Aerosol Dynamics, Reaction,
Ionization, and Dissolution
MAE mean absolute error
MAGE mean absolute gross error
MAP mean arterial pressure
MAQSIP Multiscale Air Quality Simulation
Platform
MASAES Moderate and Severe Asthmatics and
their Environment Study
max maximum
MB mean bias
MBE mean bias error
MCh methacholine
MCM master chemical mechanism
Acronym/
Abbreviation Meaning
MCP monocyte chemoattractant protein;
monocyte chemotactic protein
MCP-1 monocyte chemotactic protein 1
MDA malondialdehyde
MDA1 daily max 1-hour avg
MDA8 daily max 8-hour avg
MDL method detection limit
ME microenvironmental exposure; mean
error
MEGAN Model of Emissions of Gases and
Aerosols from Nature
MEIC Multiresolution Emissions Inventory for
China
MERRA Modern-Era Retrospective analysis for
Research and Applications (a NASA
reanalysis of satellite ozone data)
MESA Multi-Ethnic Study of Atherosclerosis
MESA-Air Multi-Ethnic Study of Atherosclerosis
and Air Pollution
MFB mean fractional bias
MFE mean fractional error
mg milligram
MI myocardial infarction; myocardial
ischemia
min minute(s); minimum
MINAP Myocardial Ischaemia National Audit
Project
MIP macrophage inflammatory protein
MIROC Model for Interdisciplinary Research on
Climate
mL milliliter
ML Mira Loma
MLN mediastinal lymph node
lii
-------
Acronym/
Abbreviation Meaning
mm millimeter
MM5 Mesoscale Model Version 5
MNB mean normalized bias
MNE mean normalized error
MNGE mean normalized gross error
mo month
MODIS MODerate resolution Imaging
Spectroradiometer
MOSES Met Office Surface Exchange Scheme
MOVES U.S. EPA Motor Vehicle Emission
Simulator
MOZART Model for OZone and Related chemical
Tracers
MP mid polar; myelopeptide; moist polar
MP AN peroxymethacrylic nitrate
MPO myeloperoxidase
MRI-CCM2 Meteorological Research Institute
Chemistry-Climate Model, version 2
mRNA messenger ribonucleic acid
MSA metropolitan statistical area
MSE mean squared error
MSEL Mullen Scales of Early Learning
MT metric ton; moist tropical
mtDNA mitochondrial DNA
MT-nx total nonoxygenated terpenes
MUC5AC mucin 5AC glycoprotein
MUC5B mucin 5B
MUSCAT MUltiScale Chemistry Aerosol Transport
MW midwest
MX mean metric
Acronym/
Abbreviation Meaning
MYJ Mellor-Yamada-Janjic
N nitrogen
n sample size; number
N100 number of hours when the measured
ozone concentration is greater than or
equal to 0.100 ppm
N2 nitrogen (gas)
N2O nitrous oxide
NA not available
NAAQS National Ambient Air Quality Standards
NAB North American background (ozone)
NACC National Alzheimer's Coordinating
Center
NACRS National Ambulatory Care Reporting
System
NADP National Atmospheric Deposition
Program
NADPH reduced form of nicotinamide adenine
dinucleotide phosphate
NAEPP National Asthma Education and
Prevention Program
NAG W-acetyl-P-D-glucosaminidase
NAI net annual increment
NAM Northern Annular Mode; North
American mesoscale
NAMS National Air Monitoring Stations
NAPCA U. S. National Air Pollution Control
Administration
NAPS National Air Pollution Surveillance
NAQFC National Air Quality Forecast Capability
NASA U.S. National Aeronautics and Space
Administration
liii
-------
Acronym/
Abbreviation Meaning
NASEM U.S. National Academy of Science,
Engineering, and Medicine
NB normalized bias
NBDPS National Birth Defects Prevention Study
NCAR National Center for Atmospheric
Research
NCDENR North Carolina Department of
Environment and Natural Resources
NCEA U. S. EPA National Center for
Environmental Assessment
NCLAN National Crop Loss Assessment Network
NCore National Core network
ND nondetectable
nDer f 1 dermatophagoides farinae allergen
NE normalized error
NECSS National Enhanced Cancer Surveillance
System
NEI U.S. EPA National Emissions Inventory
NES2 Neurobehavioral Evaluation System-2
NEu Northern Europe
NF nonfiltered air
NFkB nuclear factor kappa light-chain-
enhancer of activated B cells
ng nanogram
NGE normalized gross error
NGF nerve growth factor
NH3 ammonia
NH4+ ammonium
NH4NO3 ammonium nitrate
NHANES National Health and Nutrition
Examination Survey
Acronym/
Abbreviation Meaning
NHEERL U. S. EPA National Health and
Environmental Effects Research
Laboratory
NHIRD National Health Insurance Research
Database
NHIS National Health Insurance Service
NHIS-NSC National Health Insurance
Service—National Sample Cohort
NHLBI National Heart, Lung, and Blood
Institute
NHS Nurses' Health Study
NK neurokinin
nL nanoliter
NLDN National Lightning Detection Network
Nlrp Nucleotide-binding oligomerization
domain, Leucine rich Repeat and Pyrin
domain containing
nm nanometer
NMB normalized mean bias
NME normalized mean error
NMMAPS U.S. National Morbidity, Mortality, and
Air Pollution Study
NNs the interval between normal beats
NO nitric oxide
NO2 nitrogen dioxide
MV nitrate
NOAA U.S. National Oceanic and Atmospheric
Administration
NOS nitric oxide synthase
Notch3 neurogenic locus notch homolog
protein 3
Notch4 neurogenic locus notch homolog
protein 4
liv
-------
Acronym/
Abbreviation Meaning
NOx oxides of nitrogen (NO + NO2)
NOy the sum of NOx with its related reservoir
forms (gas-phase HNO3, PAN, HONO,
NO3, N2O5, organic nitrates [RNO3], and
nitrate in particles [pNOs])
NOz abbreviation for the sum of NOy minus
NOx (i.e., NOx reservoir species, only)
NPP net primary production
NQOl NADPH-quinone oxidoreductase
(genotype)
NR not reported
NRC National Research Council
NRCS USDA National Resources Conservation
Service
NRF2 nuclear factor (erythroid-derived 2)-like
2
NS not statistically significant; natural spline
NTS neurotensin
NU NASA-Unified
NUR nuclear receptor subfamily
NW northwest
O outdoor ozone air concentration
OlD the oxygen "singlet D" radical (a high
energy, electronically excited form of the
monatomic oxygen radical)
02 oxygen
03 ozone
OA objective analysis
obs observed
OC organic carbon
ODS ozone-depleting substance
OE elevated ozone treatment
Acronym/
Abbreviation Meaning
OGG1 8-oxoguanine glycosylase
OGTT oral glucose tolerance test
OH hydroxide; hydroxyl radical
OHCA out-of-hospital cardiac arrest
OI optimal interpolation
Oil ozone injury index
OK ordinary kriging
OMB U.S. EPA Office of Management and
Budget
OMI Ozone Monitoring Instrument
ONPHEC Ontario Population Health and
Environment Cohort
OPEC Outdoor Plant Environment Chamber
OR odds ratio
ORD U.S. EPA Office of Research and
Development
OSAT Ozone Source Apportionment Tool (in
CAMx)
OTC open-top chamber
OTU operational taxonomic unit
OZOVEG Ozone Vegetation Database
P population; probability value
PA photoacoustic analyzer; physical activity;
plasminogen activator; Policy
Assessment
Pa pascal
PAH polycyclic aromatic hydrocarbon
PAI plasminogen activator inhibitor
(e.g., PAI-1)
PAMS Photochemical Assessment Monitoring
Stations
PAN peroxyacetyl nitrate; peroxyacyl nitrate
lv
-------
Acronym/
Abbreviation
PAR
PAT
PBL
PCA
PCI
PCR
PD
PDLR
PDO
PE
PECOS
PEF
Per
PFT
Pgam5
PGE2
PGF2a
PH
PI
PIAMA
PKS
PLANTS
PLFA
PLS
Meaning
photosynthetically active radiation
pulse amplitude tonometry; paroxysmal
planetary boundary layer
principal component analysis
percutaneous coronary intervention
polymerase chain reaction
provocative dose
partial derivative linear regression
Pacific Decadal Oscillation
post-exposure; post-exercise;
phenylephrine; pulmonary embolism
Population, Exposure, Comparison,
Outcome, and Study Design
peak expiratory flow
perylene
pulmonary function test
petagram, equal to 1015 grams or one
1 billion tonnes
phosphoglycerate mutase 5
prostaglandin E2
prostaglandin F2 alpha
measure of hydrogen ion concentration
prediction interval
Prevention and Incidence of Asthma and
Mite Allergy
polyketide synthases
USDA-National Resource Conservation
Services Plant List of Accepted
Nomenclature, Taxonomy, and Symbols
phospholipid fatty acid
partial least squares
Acronym/
Abbreviation Meaning
PM particulate matter
PMio particulate matter with an aerodynamic
diameter less than or equal to 10 |xm
PM2.5 particulate matter with an aerodynamic
diameter less than or equal to 2.5 |xm
PMAE predictive mean absolute error
PMN polymorphonuclear leukocyte;
polymorphic neutrophil
PMNs polymorphic neutrophil
PMSE predictive mean squared error
PN particle number
PND postnatal day
pNN50 proportion of pairs of successive normal
sinus intervals exceeds 50 milliseconds
divided by the total number of
successive pairs of normal sinus intervals
POD probability of detection; phytotoxic
ozone dose
POD6 phytotoxic ozone dose above a threshold
of 6 nmol/m2/s
ppb parts per billion
ppbv parts per billion by volume
PPL potential productivity loss
ppm parts per million
ppm-h parts per million per hours; weighted
concentration values based on hourly
concentrations, usually summed over a
certain number of hours, day(s), months,
and/or season
PPN peroxypropionyl nitrate
PPROM preterm premature rupture of membranes
ppt parts per trillion
PQAPP Program-level Quality Assurance Project
Plan
lvi
-------
Acronym/ Acronym/
Abbreviation Meaning Abbreviation
PR time interval between the beginning of R6MA1
the P wave to the peak of the R wave
PR protein pathogenesis related protein Rag
PRB policy-relevant background, typically RAMP
used in the phrase, "PRB ozone"
PROM premature rupture of membranes RCGC
Prxd peroxiredoxin
RCT
PTB preterm birth
REA
pts points
REAS
PTT partial thromboplastin time
redox
PVD peripheral vascular disease
RF
P VN paraventricular nucleus
RFLP
PWA population-weighted average
Q1 1 st quartile or quintile RH
Q2 2nd quartile or quintile RISCAT
Q3 3rd quartile or quintile
RMSE
Q4 4th quartile or quintile
rMSSD
QA quality assurance
QAPP Quality Assurance Project Plan RNA
QBME quantile-based Bayesian maximum
entropy
ROCK
QC quality control
QNSE quasi-normal scale elimination
qPCR quantitative polymerase chain reaction RP-N
QRS time interval between the beginning of
the Q wave and the peak of the S wave
QT interval time interval between the beginning of rRNA
the Q wave to end of the T wave
RuBisCO
QTc QT interval corrected for heart rate
r correlation coefficient
Meaning
running 6-month avg of the 1-hour daily
max
recombination activating gene
Regionalized Air Quality Model
Performance
Research and Development Center for
Global Change
randomized clinical trial
risk and exposure assessment
Regional Emissions Inventory in Asia
reduction-oxidation
radiative forcing
restriction fragment length
polymorphism
relative humidity
Cardiovascular Risk and Air Pollution in
Tuscany
root-mean-squared error
root-mean-square of successive
differences
ribonucleic acid
reactive nitrogen species
rho-associated coiled-coil-containing
protein kinase
reactive oxygen species
reducing power of protein-binding
compounds on nitrogen digestibility
risk ratio, relative risk
ribosomal ribonucleic acid
ribulo se-1,5 -bispho sphate
carboxylase/oxygenase
right ventricular
lvii
-------
Acronym/
Abbreviation Meaning
sec second
S07 Statewide Air Pollution Research Center
2007
S99 Statewide Air Pollution Research Center
1999
SAEI small artery elasticity index
SAGE Stratospheric Aerosol and Gas
Experiment
SAM S-adenosyl methionine
SAPRC07 Statewide Air Pollution Research Center
2007
SAPRC07T SAPRC07 toxics
SAPRC99 Statewide Air Pollution Research Center
1999
SAT satellite
SBP systolic blood pressure
SD standard deviation
S-D Sprague-Dawley rat strain
SD-Fire satellite-derived fire emissions
SDNN standard deviation normal-to-normal
(NN or RR) time interval
SEARCH Southeastern Aerosol Research and
Characterization
SEBAS Social Environment and Biomarkers of
Aging Study
SED sedimentation rate
SES socioeconomic status
SETIL population-based case-control study of
childhood cancer in Italy
SEu southern Europe
SF seeds per fruiting structure
sFlt-1 soluble fms-like tyrosine kinase 1
Acronym/
Abbreviation Meaning
sfpd surfactant protein D
sGAW specific airway conductance
SGDS Korean Geriatric Depression Scale
(short form)
SE1 spontaneously hypertensive
SE1AM sham surgery (placebo surgery)
SEEDS Stochastic Eluman Exposure and Dose
Simulation
SE1HF spontaneously hypertensive heart failure
SHIS Shanghai Elealth Insurance System
Si silicon
SILAM System for Integrated modeLing of
Atmospheric coMposition
SIP State Implementation Plan
SJV San Joaquin Valley
SLAMS State and Local Air Monitoring Stations
SMARTFIRE Satellite Mapping Automated Reanalysis
Tool for Fire Incident Reconciliation
SMBD Spanish Minimum Basic Data set
SMOKE Sparse Matrix Operator Kernel
Emissions system
SO2 sulfur dioxide
SOC semi-volatile organic compound
SOD superoxide dismutase
SOEP Socioeconomic Panel
SOLVEG atmosphere-soil-vegetation land surface
model
SOx sulfur oxides
SoyFACE Soybean Free-Air Concentration
Enrichment (facility)
sp. species
lviii
-------
Acronym/
Abbreviation Meaning
SP surfactant protein (e.g., SPA, SPD)
SP+ substance-P-positive
spp. several species
SPSH stroke-prone spontaneously hypertensive
sRaw specific airway resistance
ST spatiotemporal
STAT3 signal transducer and activator of
transcription 3
std standard
STE stratosphere-troposphere exchange
STEM! ST elevation myocardial infarction
STN Speciation Trends Network
STROBE Strengthening the Reporting of
Observational Studies in Epidemiology
SUMOO sum of all hourly average concentrations
SUM06 seasonal sum of all average hourly
concentrations > 0.06 ppm
SUM60 seasonal sum of all hourly average
concentrations > 60 ppb
SVT supraventricular tachycardia
SW southwest
SWAN Study of Women's Health Across
Nations
t time
T1D type 1 diabetes
T2D type 2 diabetes
T3 thyroid hormone triiodothyronine
T4 thyroid hormone thyroxine
TAC1 tachykinin, precursor 1
TAG Traffic, Asthma, and Genetics
Acronym/
Abbreviation Meaning
TB tracheobronchial
TBARS thiobarbituric acid reactive substances
TC total carbon
TCCON Total Carbon Column Observing
Network
TCEQ Texas Commission on Environmental
Quality
TCR T cell receptor
TES Tropospheric Emission Spectrometer;
Tropospheric Emissions System
TES L3 TES Level 3
TF tissue factor
Tg teragram (Tg), 1 x 1012 g, a unit of mass
TGF transforming growth factor
Th thorium
Th2 T-derived lymphocyte helper 2
TIA transient ischemic attack
TID ter in die, three times per day
TIMP tissue inhibitor of metalloproteinase
TLC total lung capacity
TLR toll-like receptor
TM5 Tracer Model version 5
TNC total nonstructural carbohydrates
TNF tumor necrosis factor
TNF-a tumor necrosis factor alpha
TNFR tumor necrosis factor receptor
TNHIP Taiwan National Health Insurance
Program
TNHIP-NHIRD Taiwan National Health Insurance
Program—National Health Insurance
Research Database
lix
-------
Acronym/
Abbreviation Meaning
TOA top of the atmosphere
TOAR Tropospheric Ozone Assessment Report
TOPP Tropospheric Ozone Pollution Project
tPA tissue plasminogen activator
TPWA "true" population-weighted average
TROY Testing Responses on Youth
TRPA1 transient receptor potential cation
channel, subfamily A, member 1
TRPV1 transient receptor potential vanilloid-1
receptor
TSH thyroid stimulating hormone
TSLP thymic stromal lymphopoietin
U zonal velocity of wind vector
UA unweighted average
UAM Urban Airshed Model
UB Uinta Basin
UCD-CIT University of California at Davis
California Institute of Technology
UFP ultrafme particle
UGRB Upper Green River Basin
UK universal kriging
UNEP United Nations Environment Programme
UNFCCC United Nations Framework Convention
on Climate Change
UPA unpaired normalized bias
URI upper respiratory infection
URTI upper respiratory tract infection
USB U.S. background
USBAB U.S. background apportionment-based
USDA U.S. Department of Agriculture
Acronym/
Abbreviation Meaning
UV ultraviolet radiation
UVAFME University of Virginia Forest Model
Enhanced
V meridional velocity of wind vector
VABS Vineland Adaptive Behavior Scales
VBS volatility basis set
VCAM vascular cell adhesion molecule
VEGF vascular endothelial growth factor
VIIc factor VII coagulant activity
VIS visible (spectrum)
VOC volatile organic compound
VPD vapor pressure deficit
VPSC volatile plant signaling compound
VSD Very Simple Dynamic (soil
biogeochemical process model)
Vt tidal volume
VTI velocity-time interval
VW Volkswagen
vWF von Willebrand factor
W126 cumulative integrated exposure index
with a sigmoidal weighting function
WBC white blood cell
WDCGG World Data Centre for Greenhouse
Gases
WED U. S. EPA National Health and
Environmental Effects Research
Laboratory Western Ecology Division
WHI-OS Women's Health Initiative Observational
Study
WHO World Health Organization
WISH Women's Isoflavone Soy Health
lx
-------
Acronym/
Abbreviation Meaning
WKY Wistar Kyoto rat strain
W/m2 watts per meters squared
WMO World Meteorological Organization
WP seed weight per plant
WP-3D Lockheed WP-3D Orion Aircraft
operated by NOAA
WRF Weather Research and F orecasting
WRF-ARW WRF-Advanced Research W eather
WRF-Chem WRF with Chemistry
WRF-NMM WRF-Nonhydrostatic Mesoscale Model
WS wood smoke; Wistar rat strain
WT wild type
WUE water use efficiency
XO radical containing a halogen atom and an
oxygen atom, X = I or Br
YIB Yale Interactive Terrestrial Biosphere
Model
yr year(s)
YSU Yonsei University
ZCTA zip-code tabulation area
lxi
-------
PREFACE
The Preface to the Integrated Science Assessment for Ozone and Related Photochemical Oxidants
(Ozone ISA) outlines the legislative requirements of a National Ambient Air Quality Standard (NAAQS)
review and the history of the Ozone NAAQS. This information details the general purpose and function
of the ISA. The Preface presents the basis for the decisions that supported the previous Ozone NAAQS
review. In addition, it details specific issues pertinent to the evaluation of the scientific evidence that takes
place within this ISA, including the scope of the ISA and discipline-specific decisions that governed parts
of the review.
Legislative Requirements for the Review of the National Ambient Air
Quality Standards
Two sections of the Clean Air Act (CAA) govern the establishment, review, and revision of the
National Ambient Air Quality Standards (NAAQS). Section 108 (42 U.S. Code [U.S.C.] 7408) directs the
Administrator to identify and list certain air pollutants and then to issue air quality criteria for those
pollutants. The Administrator is to list those air pollutants "emissions of which, in their 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 ..." [42 U.S.C. 7408(a)(1)]. Air
quality criteria are intended to "accurately reflect the latest scientific knowledge useful in indicating the
kind and extent of all identifiable effects on public health or welfare, which may be expected from the
presence of [a] pollutant in the ambient air ..." (42 U.S.C. 7408(a)(2). Section 109 [42 U.S.C. 7409; CAA
(1990)1 directs the Administrator to propose and promulgate "primary" and "secondary" NAAQS for
pollutants for which air quality criteria are issued. Section 109(b)(1) 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."1 Under
Section 109(b)(2), a secondary standard must "specify a level of air quality the attainment and
maintenance of which, in the judgment of the Administrator, based on such criteria, is requisite to protect
the public welfare from any known or anticipated adverse effects associated with the presence of [the] air
pollutant in the ambient air."2
1 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).
2 Section 302(h) of the Act [42 U.S.C. 7602(h)] provides that all language referring to effects on welfare includes,
but is 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..." (CAA. 20051.
lxii
-------
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 researchers have not yet identified.1 Both kinds of uncertainty 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 pollutant 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, but rather at a level that
reduces risk sufficiently so as to protect public health with an adequate margin of safety.2
In addressing the requirement for an adequate margin of safety, the U.S. Environmental
Protection Agency (U.S. EPA) considers such factors as the nature and severity of the health effects
involved, the size of the sensitive group(s), and the kind and degree of the uncertainties. The selection of
any particular approach to providing an adequate margin of safety is a policy choice left specifically to
the Administrator's judgment.3
In setting primary and secondary standards that are "requisite" to protect public health and
welfare, respectively, as provided in Section 109(b), the U.S. EPA's task is to establish standards that are
neither more nor less stringent than necessary for these purposes. In so doing, the U.S. EPA may not
consider the costs of implementing the standards.4 Likewise, "Attainability and technological feasibility
are not relevant considerations in the promulgation of national ambient air quality standards."5
Section 109(d)(1) 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 criteria and standards as may be appropriate...." Since the early
1 See Lead Industries Association v. EPA, 647 F.2d 1130, 1154 [U.S. Court of Appeals forthe District of Columbia
Circuit (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); Coalitionof Battery Recyclers v. EPA, 604 F.
3d 613, 617-18 (D.C. Cir. 2010).
2 See Lead Industries Association v. EPA, 647 F.2d at 1156 n.51; Mississippi v. EPA, 744 F. 3d 1334, 1339, 1351,
1353 (D.C. Cir. 2013).
3 See Lead Industries Association v. EPA, 647 F.2d at 1161-62; Mississippi v. EPA, 744 F. 3d at 1353.
4 See generally, Whitman v. American Trucking Associations, 531 U.S. 457, 465-472, 475-476 (2001).
5 See American Petroleum Institute v. Costle, 665 F. 2d at 1185. accord Murray Energy Corp. v. EPA, 936 F.3d 597,
623-24 (D.C. Cir. 2019).
lxiii
-------
1980s, this independent review function has been performed by the Clean Air Scientific Advisory
Committee (CASAC).1
History of the Reviews of the Primary and Secondary National
Ambient Air Quality Standard for Ozone
NAAQS are defined by four basic elements: indicator, averaging time, level, and form. The
indicator defines the chemical species or mixture to be measured in the ambient air for the purpose of
determining whether an area attains the standard. The averaging time defines the time period over which
air quality measurements are to be averaged or otherwise analyzed. The level of a standard defines the air
quality concentration used (i.e., a specific concentration of the indicator pollutant in ambient air) in
determining whether the standard is achieved. The form of the standard defines the air quality statistic
that is compared to the level of the standard in determining whether an area attains the standard. For
example, the form of the current primary and secondary Ozone NAAQS is the 3-year avg of the annual
fourth-highest daily max 8-hour concentration. The Administrator considers these four elements
collectively in evaluating the protection to public health or welfare provided by the primary and
secondary NAAQS, respectively.
Tropospheric ozone is produced near the earth's surface due to chemical interactions involving
solar radiation and specific ozone precursors, such as nitrogen oxides (NOx), volatile organic compounds
(VOCs), and carbon monoxide (CO), which can be emitted from both natural and anthropogenic sources.2
The chemistry that leads to ozone formation is complex and varies depending upon the relative
proportions of different types of precursor pollutants, as well as external conditions such as temperature
and sunlight. Over most areas of the U.S., summer daytime ozone production typically increases as NOx
concentrations increase (2013 ISA, Section 3.2.4). Formation of ozone in this regime is described as
"NOx limited." At other times and locations, where NOx concentrations are higher or when
meteorological conditions do not favor photochemical production, ozone formation may be only weakly
dependent on NOx emissions, or even inversely correlated (i.e., NOx emissions actually deplete ozone
locally3). Ozone formation in these regimes increases as concentrations of VOCs increase and is
described as "VOC-limited." Once formed, ozone near the Earth's surface can be transported by the
prevailing winds before eventually being removed from the atmosphere over the course of hours to weeks
via chemical reactions or deposition to surfaces.
1 The List of CASAC members for the current review is available at:
https://vosemite.epa.gov/sab/sabpeople.nsf/WebExternalCommitteeRosters?QpenView&committee=CASAC&seco
ndname=Clean%20Aii%20Scientific%20Advisorv%20Committee%20.
2 Methane (CH4) emissions can also contribute to ozone formation, but its effects are more frequently observed at
the global scale over longer time periods (e.g., decadal scale).
3 In these cases, NOx generally results in eventual net ozone production downwind of the emissions sources over
longer timescales.
lxiv
-------
The U.S. EPA initially set primary and secondary NAAQS for photochemical oxidants in 1971,
with a 1-hour avg time and a level of 0.08 ppm not to be exceeded more than 1 hour per year (36 FR
8186, April 30, 1971). These standards were based on scientific information contained in the 1970 Air
Quality Criteria Document (AQCD). The U.S. EPA initiated the first periodic review of the NAAQS for
photochemical oxidants in 1977. Based on the 1978 AQCD1 (U.S. EPA. 1978). the U.S. EPA published
proposed revisions to the original NAAQS in 1978 (43 FR 26962, June 22, 1978) and final revisions in
1979 (44 FR 8202, February 8, 1979). At that time, the U.S. EPA changed the indicator from
photochemical oxidants to ozone, revised the level of the primary and secondary standards from 0.08 to
0.12 ppm and revised the form of both standards from a deterministic (i.e., not to be exceeded more than
1 hour per year) to a statistical form. With these changes, attainment of the standards was defined to occur
when the average number of days per calendar year (across a 3-year period) with maximum hourly
average ozone concentration greater than 0.12 ppm equaled one or less (44 FR 8202, February 8, 1979;
43 FR 26962, June 22, 1978). Since then, the Agency has completed multiple reviews of the air quality
criteria standards, as summarized in Table I.
The next periodic reviews of the criteria and standards for ozone and related photochemical
oxidants began in 1982 and 1983, respectively (47 FR 11561, March 17, 1982; 48 FR 38009, August 22,
1983). The U.S. EPA subsequently published the 1986 AQCD (U.S. EPA. 1986a) and the 1989 Staff
Paper2 (U.S. EPA. 1989). Following publication of the 1986 AQCD, a number of scientific abstracts and
articles were published that appeared to be of sufficient importance concerning the potential health and
welfare effects of ozone to warrant preparation of a supplement to the 1986 AQCD. In August of 1992,
the U.S. EPA proposed to retain the existing primary and secondary standards based on the health and
welfare effects information contained in the 1986 AQCD and its 1992 Supplement (57 FR 35542, August
10, 1992). In March 1993, the U.S. EPA announced its decision to conclude this review by affirming its
proposed decision to retain the standards, without revision (58 FR 13008, March 9, 1993).
1 The AQCD served the same purpose as the ISA in the current review.
2 The Staff Paper served the same purpose as the Policy Assessment (PA) in the current review.
lxv
-------
Table I History of the National Ambient Air Quality Standards for Ozone,
1971-2015.
Final
Rule/Decision
Indicator
Averaging
Time (h)
Level (ppm)
Form
36 FR 8186
April 30, 1971
Total
photochemi
cal oxidants
1
0.08
Not to be exceeded more than 1 h per year
44 FR 8202
February 8, 1979
Ozone
1
0.12
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
58 FR 13008
March 9, 1993
U.S
. EPA decided revisions to the standards were not warranted at the time.
62 FR 38856
July 18, 1997
Ozone
8
0.08
Annual fourth-highest daily max 8-h
concentration averaged over 3 yr
73 FR 16483
March 27, 2008
Ozone
8
0.075
Annual fourth-highest daily max 8-h
concentration averaged over 3 yr
80 FR 65292
October 26, 2015
Ozone
8
0.070
Annual fourth-highest daily max 8-h
concentration averaged over 3 yr
ppm = parts per million.
Note: Primary and secondary standards are identical.
In the 1992 notice of its proposed decision in that review, the U.S. EPA announced its intention to
proceed as rapidly as possible with the next review of the air quality criteria and standards for ozone and
other photochemical oxidants in light of emerging evidence of health effects related to 6- to 8-hour ozone
exposures (57 FR 35542, August 10, 1992). The U.S. EPA subsequently published the AQCD and Staff
Paper for that next review (U.S. EPA. 1996a. b). In December 1996, the U.S. EPA proposed revisions to
both the primary and secondary standards (61 FR 65716, December 13, 1996). With regard to the primary
standard, the U.S. EPA proposed replacing the then-existing 1-hour primary standard with an 8-hour
standard to be set at a level of 0.08 ppm (equivalent to 0.084 ppm based on the proposed data handling
convention) as a 3-year avg of the annual third-highest daily max 8-hour concentration. The U.S. EPA
proposed to revise the secondary standard either by setting it identical to the proposed new primary
standard or by setting it as a new standard with a cumulative seasonal form. The U.S. EPA completed this
review in 1997 by setting both the primary and secondary standards at a level of 0.08 ppm, based on the
annual fourth-highest daily max 8-hour avg concentration, averaged over 3 years (62 FR 38856, July 18,
1997).
lxvi
-------
On May 14, 1999, in response to challenges by industry and others to the U.S. EPA's 1997
decision, the D.C. Circuit remanded the Ozone NAAQS to the U.S. EPA, finding that Section 109 of the
CAA, as interpreted by the U.S. EPA, effected an unconstitutional delegation of legislative authority
(American Trucking Assoc. v. EPA, 175 F.3d 1027, 1,034-1,040 [D.C. Cir. 1999]). In addition, the court
directed that, in responding to the remand, the U.S. EPA should consider the potential beneficial health
effects of ozone pollution in shielding the public from the effects of solar ultraviolet (UV) radiation, as
well as adverse health effects (id. at 1,051-53). In 1999, the U.S. EPA sought a panel rehearing and a
rehearing en banc on several issues related to that decision. The court granted the request for panel
rehearing in part and denied it in part, but declined to review its ruling with regard to the potential
beneficial effects of ozone pollution (American Trucking Assoc. v. EPA,195 F.3d 4, 10 [D.C. Cir., 1999]).
On January 27, 2000, the U.S. 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 ozone. 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 U.S.
EPA in contravention of the Constitution). The Court remanded the case to the D.C. Circuit to consider
challenges to the Ozone NAAQS that had not been addressed by that court's earlier decisions. On March
26, 2002, the D.C. Circuit issued its final decision on the remand, finding the 1997 Ozone NAAQS to be
"neither arbitrary nor capricious," and so denying the remaining petitions for review. See American
Trucking Associations, Inc. v. EPA, 283 F.3d 355, 379 (D.C. Cir. 2002).
Coincident with the continued litigation of the other issues, the U.S. EPA responded to the court's
1999 remand to consider the potential beneficial health effects of ozone pollution in shielding the public
from effects of UV radiation (66 FR 57268, November 14, 2001; 68 FR 614, January 6, 2003). The U.S.
EPA provisionally determined that the information linking changes in patterns of ground-level ozone
concentrations to changes in relevant patterns of exposures to UV radiation of concern to public health
was too uncertain, at that time, to warrant any relaxation of the 1997 Ozone NAAQS. The U.S. EPA also
expressed the view that any plausible changes in UV-B radiation exposures from changes in patterns of
ground-level ozone concentrations would likely be very small from a public health perspective. In view of
these findings, the U.S. EPA proposed to leave the 1997 primary standard unchanged (66 FR 57268,
November 14, 2001). After considering public comments on the proposed decision, the U.S. EPA
published its final response to this remand in 2003, reaffirming the 8-hour primary standard set in 1997
(68 FR 614, January 6, 2003).
The U.S. EPA initiated the fourth periodic review of the air quality criteria and standards for
ozone and other photochemical oxidants with a call for information in September 2000 (65 FR 57810,
September 26, 2000). In 2007, the U.S. EPA proposed to revise the level of the primary standard within a
range of 0.070 to 0.075 ppm (72 FR 37818, July 11, 2007). The U.S. EPA proposed to revise the
secondary standard either by setting it identical to the proposed new primary standard or by setting it as a
new seasonal standard using a cumulative form (e.g., seasonal average). The U.S. EPA completed the
lxvii
-------
review in March 2008 by revising the levels of both the primary and secondary standards from 0.08 to
0.075 ppm, while retaining the other elements of the prior standards (73 FR 16436, March 27, 2008). On
September 16, 2009, the U.S. EPA announced its intention to reconsider the 2008 Ozone NAAQS,1 and
initiated a rulemaking to do so.
In January 2010, the U.S. EPA issued a notice of proposed rulemaking to reconsider the 2008
final decision (75 FR 2938, January 19, 2010). In that notice, the U.S. EPA proposed that further
revisions to the primary and secondary standards were necessary to provide a requisite level of protection
to public health and welfare. The U.S. EPA proposed to revise the level of the primary standard from
0.075 ppm to a level within the range of 0.060 to 0.070 ppm, and to revise the secondary standard to one
with a cumulative, seasonal form. In view of the need for further consideration, and the fact that the
Agency's next periodic review of the Ozone NAAQS required under CAA Section 109 had already begun
(as announced on September 29, 2008) [see 73 FR 56581]), the U.S. EPA decided to consolidate the
reconsideration with its statutorily required periodic review.2
On July 23, 2013, the court upheld the U.S. EPA's 2008 primary ozone standard but remanded
the 2008 secondary standard to the U.S. EPA (Mississippi v. EPA, 744 F. 3d 1,334 [D.C. Cir. 2013]).
With respect to the secondary standard, the court held that the U.S. EPA's explanation for setting the
secondary standard identical to the revised 8-hour primary standard was inadequate under the CAA
because the U.S. EPA had not adequately explained how that standard provided the required public
welfare protection.
At the time of the court's decision, the U.S. EPA had already completed significant portions of its
next statutorily required periodic review of the Ozone NAAQS. This review had been formally initiated in
2008 with a call for information in the Federal Register (73 FR 56581, September 29, 2008). In late 2014,
based on the Integrated Science Assessment (ISA), Risk and Exposure Assessments (REAs) for health
and welfare, and Policy Assessment3 developed for this review, the U.S. EPA proposed to revise the 2008
primary and secondary standards by reducing the level of both standards to within the range of 0.065 to
0.070 ppm (79 FR 75234, December 17, 2014).
The U.S. EPA's final decision in this review was published in October 2015, establishing the
now-current standards (80 FR 65292, October 26, 2015). In this decision, based on consideration of the
health effects evidence on respiratory effects of ozone in at-risk populations, the U.S. EPA revised the
primary standard from a level of 0.075 ppm to a level of 0.070 ppm, while retaining all the other elements
of the standard (80 FR 65292, October 26, 2015). The level of the secondary standard was also revised
1 The press release of this announcement is available at:
https://archive.epa.gov/epapages/newsroom archive/newsreleases/85f90b77Ilacb0c88525763300617d0d.html.
2 This rulemaking, completed in 2015, concluded the reconsideration process.
3 The final versions of these documents, released in August 2014, were developed with consideration of the
comments and recommendations from the CASAC, as well as comments from the public on the draft documents
fFrev. 2014a. b; U.S. EPA. 2014a. b, c).
lxviii
-------
from 0.075 to 0.070 ppm based on the scientific evidence of ozone effects on welfare, particularly the
evidence of ozone effects on vegetation, and quantitative analyses available in the review.1 The other
elements of the standard were retained. This decision on the secondary standard also incorporated the
U.S. EPA's response to the D.C. Circuit's remand of the 2008 secondary standard in Mississippi v. EPA,
744 F.3d 1,344 (D.C. Cir. 2013).
After publication of the final rule, a number of industry groups, environmental and public health
organizations, and certain states filed petitions for judicial review in the D.C. Circuit. The industry and
state petitioners filed briefs arguing that the revised standards were too stringent, while the environmental
and health petitioners' brief argued that the revised standards were not stringent enough to protect public
health and welfare as the Act requires. On August 23, 2019, the court issued an opinion that denied all the
petitions for review with respect to the 2015 primary standard while also concluding that the U.S. EPA
had not provided a sufficient rationale for aspects of its decision on the 2015 secondary standard and
remanding that standard to the U.S. EPA {Murray Energy Corp. v. EPA, 936 F.3d 597 [D.C. Cir. 2019]).
Purpose and Overview of the Integrated Science Assessment
The ISA is a comprehensive evaluation and synthesis of the policy-relevant science "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," as described in Section 108 of the Clean
Air Act (CAA. 1990). This ISA communicates critical science judgments of the health and welfare
criteria for ozone and related photochemical oxidants, and serves as the scientific foundation for the
review of the current primary (health-based) and secondary (welfare-based) Ozone NAAQS.
As stated in the Ozone IRP (Section 4.1), the purpose of this ISA is to draw upon the existing
body of evidence to synthesize and provide a critical evaluation of the current state of scientific
knowledge on the most relevant issues pertinent to the review of the NAAQS for ozone and other
photochemical oxidants, to identify changes in the scientific evidence bases since the previous review,
and to describe remaining or newly identified uncertainties. The ISA identifies, critically evaluates, and
synthesizes the most policy-relevant current scientific literature (e.g., epidemiology, controlled human
exposure, animal toxicology, atmospheric science, exposure science, ecology, and climate-related
science), including key science judgments that are important to inform the development of risk and
exposure analyses (as warranted) and the Policy Assessment, as well as other aspects of the NAAQS
review process.
This ISA evaluates relevant scientific literature published since the 2013 Ozone ISA (U.S. EPA.
2013). integrating key information and judgments contained in the 2013 Ozone ISA and previous
assessments of ozone, specifically, the 2006 AQCD for Ozone and Related Photochemical Oxidants (U.S.
EPA. 2006). the 2007 Staff Paper ("U.S. EPA. 2007). the 1996 AQCD and Staff Paper for Ozone and
1 The current NAAQS for ozone are specified at 40 CFR 50.19.
lxix
-------
Related Photochemical Oxidants (U.S. EPA. 1996a. b). the 1986 AQCD for ozone (U.S. EPA. 1982) and
its Supplement (U.S. EPA. 1986b). and the 1978 AQCD for Ozone and Other Photochemical Oxidants
(NAPCA. 1969). Thus, this ISA updates the state of the science from that available for the 2013 Ozone
ISA, which informed decisions on the primary and secondary Ozone NAAQS in the review completed in
2015.
This new review of the primary and secondary Ozone NAAQS is guided by the policy-relevant
questions identified in the IRP for the National Ambient Air Quality Standards for Ozone.1 To address
these questions and update the scientific judgments in the 2013 Ozone ISA (U.S. EPA. 2013). this ISA
aims to:
• Assess whether new information (since the last Ozone NAAQS review) further informs the
relationship between exposure to ozone and specific health and welfare effects.
• Provide new information as to whether the NAAQS (comprised of indicator, averaging time,
form, and level) are appropriate.
In addressing policy-relevant questions, this ISA aims to characterize the health and welfare
effects of ozone independent from co-occurring air pollutants. In the characterization of whether there is
evidence of an independent health and welfare effect due to ozone, the ISA considers possible influences
of other atmospheric pollutants, including both gaseous (i.e., NO2, SO2, and CO) and various particulate
matter (PM) size fractions. The information summarized in this ISA will serve as the scientific foundation
for the review of the current primary and secondary Ozone NAAQS.
Process for Developing Integrated Science Assessments
The U.S. EPA uses a structured and transparent process for evaluating scientific information and
determining the causal nature of relationships between air pollution exposures and health effects [details
provided in the Preamble to the Integrated Science Assessments; U.S. EPA (2015a)l. The ISA
development process describes approaches for literature searches, criteria for selecting and evaluating
relevant studies, and a framework for evaluating the weight of evidence and forming causality
determinations. Table II provides a description of each of the five causality determinations and the types
of scientific evidence that is considered for each category for both health and welfare effects.
1 https://www.epa.gov/naaas/ozone-o3-standards-planning-documents-current-review.
lxx
-------
Table II
Weight of evidence for causality determinations.
Health Effects
Ecological and Other Welfare Effects
Causal Evidence is sufficient to conclude that there is a
relationship causal relationship with relevant pollutant
exposures (e.g., doses or exposures generally
within one to two orders of magnitude of recent
concentrations). That is, the pollutant has been
shown to result in health effects in studies in
which chance, confounding, and other biases
could be ruled out with reasonable confidence.
For example: (1) controlled human exposure
studies that demonstrate consistent effects, or
(2) observational studies that cannot be
explained by plausible alternatives or that are
supported by other lines of evidence (e.g., animal
studies or mode of action information). Generally,
the determination is based on multiple
high-quality studies conducted by multiple
research groups.
Evidence is sufficient to conclude that there is a
causal relationship with relevant pollutant
exposures. That is, the pollutant has been
shown to result in effects in studies in which
chance, confounding, and other biases could be
ruled out with reasonable confidence. Controlled
exposure studies (laboratory or small- to
medium-scale field studies) provide the
strongest evidence for causality, but the scope
of inference may be limited. Generally, the
determination is based on multiple studies
conducted by multiple research groups, and
evidence that is considered sufficient to infer a
causal relationship is usually obtained from the
joint consideration of many lines of evidence
that reinforce each other.
Likely to be a Evidence is sufficient to conclude that a causal
causal relationship is likely to exist with relevant
relationship pollutant exposures. That is, the pollutant has
been shown to result in health effects in studies
where results are not explained by chance,
confounding, and other biases, but uncertainties
remain in the evidence overall. For example:
(1) observational studies show an association,
but copollutant exposures are difficult to address
and/or other lines of evidence (controlled human
exposure, animal, or mode of action information)
are limited or inconsistent, or (2) animal
toxicological evidence from multiple studies from
different laboratories demonstrate effects, but
limited or no human data are available.
Generally, the determination is based on multiple
high-quality studies.
Evidence is sufficient to conclude that there is a
likely causal association with relevant pollutant
exposures. That is, an association has been
observed between the pollutant and the
outcome in studies in which chance,
confounding, and other biases are minimized
but uncertainties remain. For example, field
studies show a relationship, but suspected
interacting factors cannot be controlled, and
other lines of evidence are limited or
inconsistent. Generally, the determination is
based on multiple studies by multiple research
groups.
Suggestive of, Evidence is suggestive of a causal relationship
but not with relevant pollutant exposures but is limited,
sufficient to and chance, confounding, and other biases
infer, a causal cannot be ruled out. For example: (1) when the
relationship body of evidence is relatively small, at least one
high-quality epidemiologic study shows an
association with a given health outcome and/or at
least one high-quality toxicological study shows
effects relevant to humans in animal species, or
(2) when the body of evidence is relatively large,
evidence from studies of varying quality is
generally supportive but not entirely consistent,
and there may be coherence across lines of
evidence (e.g., animal studies or mode of action
information) to support the determination.
Evidence is suggestive of a causal relationship
with relevant pollutant exposures, but chance,
confounding, and other biases cannot be ruled
out. For example, at least one high-quality study
shows an effect, but the results of other studies
are inconsistent.
lxxi
-------
Table II (Continued): Weight of evidence for causality determinations.
Health Effects
Ecological and Other Welfare Effects
Inadequate to Evidence is inadequate to determine that a
infer a causal causal relationship exists with relevant pollutant
relationship exposures. The available studies are of
insufficient quantity, quality, consistency, or
statistical power to permit a conclusion regarding
the presence or absence of an effect.
Evidence is inadequate to determine that a
causal relationship exists with relevant pollutant
exposures. The available studies are of
insufficient quality, consistency, or statistical
power to permit a conclusion regarding the
presence or absence of an effect.
Not likely to be Evidence indicates there is no causal relationship
a causal with relevant pollutant exposures. Several
relationship adequate studies, covering the full range of
levels of exposure that human beings are known
to encounter and considering at-risk populations
and lifestages, are mutually consistent in not
showing an effect at any level of exposure.
Evidence indicates there is no causal
relationship with relevant pollutant exposures.
Several adequate studies examining
relationships with relevant exposures are
consistent in failing to show an effect at any
level of exposure.
Source: U.S. EPA (2015aV
As part of this process, the ISA is reviewed by the CASAC, which is a formal independent panel
of scientific experts appointed by the Administrator of the EPA, and by the public. Because this ISA
informs the review of the primary and secondary Ozone NAAQS, it integrates and synthesizes
information characterizing exposure to ozone and potential relationships with health and welfare effects.
Relevant studies include those examining atmospheric chemistry, spatial and temporal trends, and
exposure assessment, as well as U.S. EPA analyses of air quality and emissions data. Relevant health
research includes epidemiologic, controlled human exposure, and toxicological studies on health effects,
as well as the biological plausibility that ozone could cause such health effects. Additionally, relevant
welfare research includes studies examining effects on environmental biota and ecosystems, as well as
climate.
Scope of the ISA
This ISA updates the state of the science from that available for the 2013 Ozone ISA, which
informed decisions on the primary and secondary Ozone NAAQS in the review completed in 2015. The
previous ozone ISA was published in 2013 (U.S. EPA. 2013) and included peer-reviewed literature
published through July 2011. Search techniques for the current ISA identified and evaluated studies and
reports that have undergone scientific peer review and were published or accepted for publication
between January 1, 2011 (providing some overlap with the cutoff date from the last review) and March
30, 2018. Studies published after the literature cutoff date for this review were also considered if they
were submitted in response to the Call for Information (83 FR 29785, June 26, 2018) or identified in
subsequent phases of ISA development (e.g., peer-input consultation, CASAC review of draft Integrated
Review Plan), particularly to the extent that they provide new information that affects key scientific
lxxii
-------
conclusions. Section 10.2. "Literature Search and Initial Screen," details the study selection process in
further detail.
For human health effects, the U.S. EPA concluded in the 2013 Ozone ISA that the findings of
epidemiologic, controlled human exposure, and animal toxicological studies collectively provided
evidence of a "causal relationship" for short-term ozone exposures and respiratory effects. In evaluating a
broader range of health effects for ozone, the 2013 Ozone ISA concluded there was evidence of a "likely
to be causal relationship" for long-term ozone exposures and respiratory effects and for short-term ozone
exposures and cardiovascular effects and mortality. Additionally, there was evidence "suggestive of a
causal relationship" for ozone exposures and other health effects, including developmental and
reproductive effects (e.g., low birth weight, infant mortality) and central nervous system effects
(e.g., cognitive development).
For welfare effects, the evidence in the 2013 Ozone ISA indicated a "causal relationship"
between ozone exposure 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. The evidence indicated a "likely to be causal
relationship" for reduced carbon sequestration in terrestrial ecosystems, alteration of terrestrial ecosystem
water cycling, and alteration of terrestrial community composition. For climate, there was a causal
relationship between changes in tropospheric ozone concentration and radiative forcing and likely to be a
causal relationship between changes in tropospheric ozone concentration and effects on climate. For this
current review, specific science questions related to the causality determinations that were addressed
include:
• Does the evidence base from recent studies contain new information to support or call into
question the causality determinations made for relationships between ozone exposure and various
health and welfare effects in the 2013 Ozone ISA?
• Is there new information to extend causality determinations to other ecological endpoints?
• Does new evidence confirm or extend biological plausibility of ozone-related health effects?
• What is the strength of inference from epidemiologic studies based on the extent to which they
have:
o Examined exposure metrics that capture the spatial and/or temporal pattern of ozone in
the study area?
o Assessed potential confounding by other pollutants and factors?
• What does the available information indicate regarding changes in population health status that
may be associated with a decrease in ambient air ozone concentrations that might inform
causality determinations?
lxxiii
-------
Evaluation of the Evidence
The Preamble to the ISAs (U.S. EPA. 2015a) describes the general framework for evaluating
scientific information, including criteria for assessing study quality and developing scientific conclusions.
Aspects specific to evaluating studies of ozone are described in Appendix 10 of the ISA, which were
applied to studies that fit the overall scope of this Ozone ISA. Appendix 10 complements the Preamble by
providing additional details regarding methods used in the literature search, study quality evaluations, and
quality assurance. Categories of health and welfare effects were considered for evaluation in this ISA if
they were examined in previous U.S. EPA assessments for ozone or in multiple recent studies. Therefore,
in this ISA, the broad health effects categories evaluated include those considered in the 2013 Ozone ISA
(i.e., respiratory effects, cardiovascular effects, central nervous system effects, cancer, and mortality),
along with the addition of metabolic effects. Further, new research indicates it is appropriate to refine the
category of reproductive and developmental effects and provide separate conclusions on male/female
reproduction and fertility and pregnancy and birth outcomes.
In the 2013 Ozone ISA, the welfare effects evidence for ozone focused on the effects of ozone on
vegetation and ecosystems and the role of tropospheric ozone on climate change. In this ISA, the U.S.
EPA builds on the 2013 Ozone ISA by evaluating the newly available literature related to ozone
exposures and welfare effects, specifically ecological effects and effects on climate. With regards to
ecological effects, this ISA evaluates the literature related to ozone exposures at levels of biological
organization from the organism to the ecosystem level, including effects on biodiversity. Evidence from
experimental (e.g., laboratory, greenhouse, open-top chamber, free-air carbon dioxide enrichment), field,
gradient, and modeling studies that address effects of ozone on ecological endpoints are considered to
identify concentrations at which effects are observed.
Peer review is an important component of any scientific assessment. U.S. EPA has formal
guidance about peer review in the Peer Review Handbook (U.S. EPA. 2015b). and this ISA follows the
policies and procedures identified therein. Additionally, this ISA follows the U.S. EPA's Information
Quality Guidelines (U.S. EPA. 2002).
In forming the key science judgments for each of the health and welfare effects categories
evaluated, the Ozone ISA draws conclusions about relationships between ozone exposure and health or
welfare effects by integrating information across scientific disciplines and related health or welfare
outcomes, respectively, and synthesizing evidence from previous and recent studies. To impart
consistency in the evaluation of health effects evidence for epidemiologic studies, additional parameters
to those outlined in the scope were developed. To help compare results across epidemiologic studies, risk
estimates were standardized to a defined increment for both short- and long-term exposure to ozone,
unless otherwise noted in the text. All epidemiologic results are standardized to either a 15-ppb increase
in 24-hour avg, a 20-ppb increase in 8-hour daily max, a 25-ppb increase in 1-hour daily max ozone
concentrations, or a 10-ppb increase in seasonal/annual ozone concentrations. These increments are based
loosely on the 50th-95th percentile of concentrations observed for each averaging time and exposure
lxxiv
-------
duration. Additionally, while assessing copollutants or other variables in epidemiologic studies, high,
moderate, or low correlations are defined as the following: low correlation, r < 0.40; moderate
correlation, r > 0.40 and r < 0.70; and high correlation, r > 0.70. Consistency in interpreting the
epidemiologic evidence through approaches such as the standardization of risk estimates and the
evaluation of correlations, in combination with the integration of evidence across scientific disciplines,
supports a thorough evaluation of the current state of the science for ozone.
In evaluating the evidence, determinations are made about causation, not just association, and are
based on judgments of aspects such as the consistency of evidence within a discipline, coherence of
effects across disciplines, and biological plausibility of observed effects. Determinations account for
related uncertainties. The ISA uses a formal causal framework [Table II of the Preamble to the ISAs; U.S.
EPA (2015a) I to classify the weight of evidence according to the five-level hierarchy as described in
Table II above.
lxxv
-------
References for Preface
CAA (Clean Air Act). (1990). Clean Air Act, as amended by Pub. L. No. 101-549, section 108: Air
quality criteria and control techniques, 42 USC 7408.
http ://www .law. Cornell. edu/uscode/text/42/7408
CAA (Clean Air Act). (2005). Clean Air Act, section 302: Definitions, 42 USC 7602.
http://www.gpo.gov/fdsvs/pkg/USCODE-2005-title42/pdf/USCQDE-2005-title42-chap85-
subchapIII-sec7602 .pdf
Frev. HC. (2014a). Letter from Dr. H. Christopher Frey, Chair, Clean Air Scientific Advisory
Committee, to Administrator Gina McCarthy. CASAC Review of the EPAs Second Draft Policy
Assessment for the Review of the Ozone National Ambient Air Quality Standards. EPA-CASAC-
14-004. Available online at
https://vosemite.epa.gov/sab/sabproduct.nsf/264cbl227d55e02c85257402007446a4/5EFA32QCCA
D326E885257D030071531C/$File/EPA-CASAC-14-004+unsigned.pdf
Frev. HC. (2014b). Letter from Dr. H. Christopher Frey, Chair, Clean Air Scientific Advisory
Committee, to Administrator Gina McCarthy. CASAC Review of the EPAs Welfare Risk and
Exposure Assessment for Ozone (Second External Review Draft). EPACASAC-14-003. Available
online at
https://vosemite.epa.gov/sab/sabproduct.nsf/264cbl227d55e02c852574020Q7446a4/3D561C9413E
49E8E85257CFB0069E0DF/$File/EPA-CASAC-14-003+unsigned.pdf
NAPCA (National Air Pollution Control Administration). (1969). Air quality criteria for particulate
matter. Washington, DC.
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, DC.
http://nepis.epa.gov/exe/Z vPURL.cgi?Dockev=200089CW.txt
U.S. EPA (U.S. Environmental Protection Agency). (1982). Air quality criteria for particulate matter
and sulfur oxides (final, 1982) [EPA Report]. (EPA 600/8-82/029a). Washington, DC:
Environmental Criteria and Assessment Office.
http://cfbub.epa.gov/ncea/cfm/recordisplav.cfm?deid=46205
U.S. EPA (U.S. Environmental Protection Agency). (1986a). 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.
https://ntrl.ntis.gov/NTRL/dashboard/searchResults.xhtml?searchOuerv=PB87142949
U.S. EPA (U.S. Environmental Protection Agency). (1986b). Second addendum to air quality criteria
for particulate matter and sulfur oxides (1982): Assessment of newly available health effects
information [EPA Report]. (EPA/600/8-86/020F). Research Triangle Park, NC: Environmental
Criteria and Assessment Office.
U.S. EPA (U.S. Environmental Protection Agency). (1989). Review of the national ambient air quality
standards for ozone: Assessment of scientific and technical information: OAQPS staff report [EPA
Report]. (EPA-450/2-92-001). Research Triangle Park, NC.
http://nepis.epa.gov/Exe/ZvPURL.cgi?Dockev=2000LQW6.txt
U.S. EPA (U.S. Environmental Protection Agency). (1996a). Air quality criteria for ozone and related
photochemical oxidants, Vol. Ill of III [EPA Report]. (EPA/600/P-93/004cF). Research Triangle
Park, NC. https://ntrl.ntis.gov/NTRL/dashboard/searchResults,xhtml?searchQuery=PB96185608
lxxvi
-------
U.S. EPA (U.S. Environmental Protection Agency). (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.
https://ntrl.ntis. gov/NTRL/dashboard/searchResults.xhtml?searchQuery=PB96203435
U.S. EPA (U.S. Environmental Protection Agency). (2002). Guidelines for ensuring and maximizing
the quality, objectivity, utility, and integrity of information disseminated by the Environmental
Protection Agency. (EPA/260/R-02/008). Washington, DC: U.S. Environmental Protection
Agency, Office of Environmental Information. https://www.epa.gov/sites/production/files/2Q17-
03/documents/epa-info-qualitv-guidelines.pdf
U.S. EPA (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: U.S.
Environmental Protection Agency, Office of Research and Development, National Center for
Environmental Assessment-RTP Office.
http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=149923
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-07/007). Research Triangle Park, NC.
https://www3.epa.gov/ttn/naaqs/standards/ozone/data/2007 07 ozone staff paper.pdf
U.S. EPA (U.S. Environmental Protection Agency). (2013). Integrated science assessment for ozone
and related photochemical oxidants [EPA Report]. (EPA/600/R-10/076F). Research Triangle Park,
NC: U.S. Environmental Protection Agency, Office of Research and Development, National Center
for Environmental Assessment-RTP Division.
http://cfbub.epa.gov/ncea/isa/recordisplav.cfm?deid=247492
U.S. EPA (U.S. Environmental Protection Agency). (2014a). Health Risk and Exposure Assessment
for Ozone. Final Report. (EPA-452/P-14-004a). Research Triangle Park, NC: Office of Air Quality
Planning and Standards. http://nepis.epa.gov/exe/ZvPURL.cgi?Dockev=P100KBUF.txt
U.S. EPA (U.S. Environmental Protection Agency). (2014b). Policy assessment for the review of the
ozone national ambient air quality standards (pp. EPA-452/R-414-006). Research Triangle Park,
NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards.
https://www3 .epa.gov/ttn/naaa s/standards/ozone/data/20140829pa.pdf
U.S. EPA (U.S. Environmental Protection Agency). (2014c). Welfare Risk and Exposure Assessment
for Ozone, Final. (EPA-452/P-14-005a). Research Triangle Park, NC: Office of Air Quality
Planning and Standards.
https://www3.epa.gOv/ttn/naaqs/standards/ozone/data/20141021welfarerea.pdf
U.S. EPA (U.S. Environmental Protection Agency). (2015a). Preamble to the integrated science
assessments [EPA Report]. (EPA/600/R-15/067). Research Triangle Park, NC: U.S. Environmental
Protection Agency, Office of Research and Development, National Center for Environmental
Assessment, RTP Division. https://cfpub.epa.gov/ncea/isa/recordisplav.cfm?deid=310244
U.S. EPA (U.S. Environmental Protection Agency). (2015b). Science policy council peer review
handbook [EPA Report] (4th ed.). (EPA/100/B-15/001). Washington, DC: U.S. Environmental
Protection Agency, Science Policy Council, https://www.epa.gov/osa/peer-review-handbook-4th-
edition-2015
lxxvii
-------
EXECUTIVE SUMMARY
ES.1 Purpose and Scope of the Integrated Science Assessment
This Integrated Science Assessment (ISA)1 is a comprehensive evaluation and synthesis of the
policy-relevant science aimed at characterizing the health and welfare2 effects caused by ozone. It
communicates critical science judgments of the health-based and welfare-based criteria for ozone and
related photochemical oxidants in ambient air. In 2015, the U.S. EPA lowered the health- and
welfare-based National Ambient Air Quality Standards (NAAQS) for ozone to 0.070 ppm (annual
fourth-highest daily max 8-hour concentration averaged over 3 years3). The health-based ozone NAAQS
is meant to protect public health, including at-risk populations such as children and people with asthma,
with an adequate margin of safety. The welfare-based ozone standard is intended to protect the public
welfare from known or anticipated adverse effects associated with the presence of ozone in ambient air.
The ISA identifies and critically evaluates the most policy-relevant scientific literature across
scientific disciplines, including epidemiology, controlled human exposure studies, animal toxicology,
atmospheric science, exposure science, vegetation studies, agricultural science, ecology, and
climate-related science. Key scientific conclusions (e.g., causality determinations; Section ES.4) are
presented and explained. These conclusions provide the scientific basis for developing risk and exposure
analyses, policy evaluations, and policy decisions for the review. This ISA draws conclusions about the
causal nature of the relationships between ozone exposure and health and welfare effects by integrating
recent evidence across scientific disciplines with the evidence base evaluated in previous reviews. U.S.
EPA engages the Clean Air Scientific Advisory Committee (CASAC) as an independent federal advisory
committee to conduct peer reviews of draft ISA and other materials. Peer review comments provided by
the CASAC and public comments about the external review draft were considered in the development of
this ISA (Section 10.4). The ISA thus provides the policy-relevant scientific information necessary to
conduct a review of the NAAQS.
This Executive Summary provides an overview of the important conclusions drawn in the ISA
across scientific disciplines, beginning with information on sources, concentrations, estimated
background concentrations of ozone and ozone exposure, followed by health and welfare effects. A more
detailed summary of the evidence is presented in the Integrated Synthesis, and individual Appendices for
1 The general process for developing an ISA, including the framework for evaluating weight of evidence and
drawing scientific conclusions and causality determinations, is described in a companion document, Preamble to the
Integrated Science Assessments (U.S. EPA. 20151. www.epa.gov/isa.
2 Under Clean Air Act section 302(h), effects on welfare include, but are not limited to, "effects on soils, water,
crops, vegetation, manmade materials, animals, wildlife, weather, visibility, and climate, damage to and
deterioration of property, and hazards to transportation, as well as effects on economic values and on personal
comfort and well-being."
3 Final rule signed October 1, 2015, and effective December 28, 2015 (80 FR 65291).
ES-1
-------
each topic area include study-level information and an in-depth characterization of the weight-of-evidence
conclusions.
ES.2 Ozone in Ambient Air
The general photochemistry of tropospheric ozone is well-established. Ozone is produced in
urban areas and downwind of sources mainly by the reaction of volatile organic compounds (VOCs) with
oxides of nitrogen (NOx) in the presence of sunlight, and outside of polluted areas mainly by reactions of
carbon monoxide (CO) and methane (CH4) with NOx (Section 1.4). Recent developments in
understanding ozone chemistry include observations of higher ozone concentrations during the winter in
some western U.S. mountain basins (Section 1.4.1) and new research on the role of marine halogen
chemistry in suppressing coastal ozone concentrations (Section 1.4.2). Air monitoring data for the period
2015-2017 show that U.S. daily max 8-hour avg concentrations of ozone (MDA8) are higher in spring
and summer (median = 46 ppb) than in autumn (median = 38 ppb) and winter (median = 34 ppb).
Figure ES-1 shows the highest values of the 3-year avg of annual fourth-highest MDA8 ozone
concentrations (design values above 70 ppb) occur in central and southern California, Arizona, Colorado,
Utah, Texas, along the shore of Lake Michigan, and in the Northeast Corridor, typically during the ozone
season between May and September (Section 1.2.1.1).
ES-2
-------
2015-2017 Ozone Design Values
• 43-60 ppb (179 sites) O 66 - 70 ppb (334 sites) • 76 -112 ppb (110 sites)
O 61-65 ppb (378 sites) © 71-75 ppb (136 sites)
Figure ES-1 Individual monitor ozone concentrations in terms of design
values (i.e., 3-year avg of annual fourth-highest max daily 8-hour
avg ozone concentration) for 2015-2017.
A better understanding of the origins of ground-level U.S. background (USB) ozone and its
concentration trends has emerged since the 2013 Ozone ISA. USB ozone concentration is defined as the
ozone concentration that would occur if all U.S. anthropogenic ozone precursor emissions were removed
(Section IS.2.2). Major contributors to USB ozone concentrations are stratospheric exchange,
international transport, wildfires, lightning, global methane emissions, and natural biogenic and geogenic
precursor emissions. Ozone monitors cannot discern the portion of ambient ozone concentrations that
come from USB. Instead, USB concentrations are estimated using photochemistry and transport models.
The estimates of USB ozone concentrations include uncertainties of about 10 ppb for seasonal average
concentrations, with higher uncertainty for MDA8 concentrations. Models consistently estimate higher
USB ozone concentrations at higher elevations of the western U.S. than in the eastern U.S. or along the
Pacific coast. The estimated seasonal pattern in USB ozone concentrations tends to indicate lower USB in
the summer than during the rest of the year. Several modeling studies using different approaches indicate
that for MDA8 concentrations above 50-60 ppb, USB concentration estimates generally do not increase
with increasing total ozone concentration (i.e., USB ozone concentrations are no higher on high ozone
days than on low or moderate ozone days). The temporal trend in estimated USB ozone concentrations
indicates increasing concentrations at high elevation western U.S. sites through approximately 2010.
ES-3
-------
Recently, however, this trend has shown signs of slowing or even reversing, possibly due to decreasing
East Asian precursor emissions.
ES.3 Exposure to Ozone
Ambient air ozone concentrations, either measured at fixed-site monitors or estimated by models,
are often used as surrogates for personal exposure in epidemiologic studies. Exposure measurement error
can lead to reduced precision and an underestimation of the association between short-term ambient
ozone exposure and a health effect (Section 2.6.1). For studies of long-term exposure, the true effect of
exposure to ambient ozone may be underestimated or overestimated when the exposure model
respectively overestimates or underestimates ozone exposure. It is much more common for the effect to
be underestimated, and bias in the effect estimate is typically small in magnitude (Section 2.6.2). The
availability and sophistication of models to predict ambient ozone concentrations to estimate exposure
have increased substantially in recent years (Section 2.3.2). For effects elicited by ozone, the use of
exposure estimates that do not account for population behavior and mobility (e.g., via use of time-activity
data) may result in underestimation of the true effect and reduced precision (Section 2.4.1).
Tropospheric ozone can cause plant damage, which can then have negative impacts on terrestrial
ecosystems as shown in observational and controlled exposure studies and in models using experimental
data to extrapolate to effects at the community and ecosystem scale. Robust exposure indices that quantify
exposure as it relates to measured plant response (e.g., growth) have been in use for decades and are
derived from hourly ozone concentrations. Exposure duration influences the degree of plant response, and
ozone effects on plants are cumulative. Cumulative indices summarize ozone concentrations overtime
and provide a consistent metric for reviewing and comparing exposure-response effects obtained from
various studies. Cumulative indices of exposure that differentially weight hourly concentrations have
been found to be best suited to characterize vegetation exposure to ozone with regard to reductions in
vegetation growth and yield (Section 8.1.2.1).
ES.4 Health and Welfare Effects of Ozone Exposure
Broad health and welfare effect categories are evaluated independently in the Appendices of this
ISA. Determinations are made about causation by evaluating evidence across scientific disciplines and are
based on judgments of consistency, coherence, and biological plausibility of observed effects, as well as
related uncertainties. The ISA uses a formal causality framework to classify the weight of evidence using
a five-level hierarchy described in Table II of the Preamble (U.S. EPA. 2015). The subsequent sections
characterize the evidence that forms the basis of causality determinations for health and welfare effect
categories of a "causal relationship" or a "likely to be causal relationship," or describe instances where a
causality determination has changed (i.e., "likely to be causal" changed to "suggestive of, but not
sufficient to infer, a causal relationship"). Other relationships between ozone and health effects are
"suggestive of but not sufficient to infer" and "inadequate to infer" a causal relationship. These causality
ES-4
-------
determinations appear in Table ES-1. and are more fully discussed in the respective health effects
Appendices.
ES.4.1 Health Effects of Ozone Exposure
Ozone-induced effects can occur through a variety of complex pathways within the body. After
inhalation, ozone reacts with lipids, proteins, and antioxidants in the epithelial lining fluid of the
respiratory tract, creating secondary oxidation products (Section 5.2.3). Initial ozone exposure leads to
physiological reactions that may induce a host of autonomic, endocrine, immune, and inflammatory
responses throughout the body at the cellular, tissue, and organ level. Recent evidence continues to
support ozone-induced effects on the respiratory system. In addition, recent evidence indicates that short-
term exposure to ozone is likely to induce metabolic effects, as shown in Figure ES-2. There is also some
evidence that ozone exposure can affect the cardiovascular and nervous systems, reproduction and
development, and mortality, although there are more uncertainties associated with interpretation of the
evidence for these effects.
ES-5
-------
Table ES-1 Summary of causality determinations by exposure duration and
health outcome.
Health Outcome3
Conclusions from
2013 Ozone ISA
Conclusions in the 2020 ISA
Short-term exposure to ozone
Respiratory effects
Causal relationship
Causal relationship
Cardiovascular effects
Likely to be causal relationship
Suggestive of, but not sufficient to infer, a causal
relationship0
Metabolic effects
No determination made
Likely to be causal relationship15
Total mortality
Likely to be causal relationship
Suggestive of, but not sufficient to infer, a causal
relationship0
Central nervous system
effects
Suggestive of a causal relationship11
Suggestive of, but not sufficient to infer, a causal
relationship
Long-term exposure to ozone
Respiratory effects
Likely to be causal relationship
Likely to be causal relationship
Cardiovascular effects
Suggestive of a causal relationship11
Suggestive of, but not sufficient to infer, a causal
relationship
Metabolic effects
No determination made
Suggestive of, but not sufficient to infer, a causal
relationship15
Total mortality
Suggestive of a causal relationship11
Suggestive of, but not sufficient to infer, a causal
relationship
Reproductive effects
Suggestive of a causal relationship11
Effects on fertility and reproduction: suggestive of, but
not sufficient to infer, a causal relationship15
Effects on pregnancy and birth outcomes: suggestive of,
but not sufficient to infer, a causal relationship15
Central nervous system
effects
Suggestive of a causal relationship11
Suggestive of, but not sufficient to infer, a causal
relationship
Cancer
Inadequate to infer a causal
relationship
Inadequate to infer the presence or absence of a causal
relationship®
aHealth effects (e.g., respiratory effects, cardiovascular effects) include the spectrum of outcomes, from measurable subclinical effects
(e.g., decrements in lung function, blood pressure) to observable effects (e.g., medication use, hospital admissions) and cause-specific mortality.
Total mortality includes all-cause (nonaccidental) mortality, as well as cause-specific mortality.
bDenotes new causality determination.
°Denotes change in causality determination from 2013 Ozone ISA.
dSince the 2013 Ozone ISA, the causality determination language has been updated and this category is now stated as suggestive of, but not
sufficient to infer, a causal relationship.
eSince the 2013 Ozone ISA, the causality determination language has been updated and this category is now stated as inadequate to infer the
presence or absence of a causal relationship.
ES-6
-------
Causality Determinations for Health Effects of Ozone
2020 Ozone ISA
Short-term
exposure
Kespirarory
Long-term
exposure
Metabolic
Short-term
exposure
+
Long-term
exposure
+
Cardiovascular
Short-term
exposure
i
Long-term
exposure
0)
E
o
a
Short-term
exposure
3
o
¦E
CO
Long-term
exposure
Q)
X
>
=3
"O
Male/Female
Reproduction
and Fertility
Long-term
*
o
Q_
0)
a:
Pregnancy and
Birth Outcomes
exposure
*
Cancer
Long-term
exposure
Mortality
Short-term
exposure
i
Long-term
exposure
Causal g Likely causal | Suggestive Q lnadequate ~
+ new causality determination; ~ causality determination changed from likely
causal to suggestive; * change in scope of health outcome category from 2013
Ozone ISA
Figure ES-2 Causality determinations for health effects of short- and
long-term exposure to ozone.
ES-7
-------
The strongest evidence for health effects due to ozone exposure continues to come from studies
of short- and long-term ozone exposure and respiratory health, and this evidence is detailed in
Appendix 3. Consistent with conclusions from the 2013 Ozone ISA (Table ES-1). there is a "causal
relationship" between short-term ozone exposure and respiratory effects (Section 3.Lll). and there
is a "likely to be causal relationship" between long-term ozone exposure and respiratory effects
(Section 3.2.6).
For short-term ozone exposure, controlled human exposure studies conducted over many decades
provide experimental evidence for ozone-induced lung function decrements (Figure ES-3). airway
responsiveness, respiratory symptoms, and respiratory tract inflammation. Epidemiologic studies continue
to provide evidence that ozone concentrations in ambient air are associated with a range of respiratory
effects, including asthma exacerbation, chronic obstructive pulmonary disease (COPD) exacerbation,
respiratory infection, and hospital admissions and emergency department (ED) visits for combined
respiratory diseases.
A large body of animal toxicological studies demonstrate ozone-induced alterations in lung
function, inflammation, increased airway responsiveness, and impaired lung host defense. These animal
toxicological studies also aid in our understanding of potential mechanisms underlying respiratory effects
at the population level and the biological plausibility of epidemiologic associations between short-term
ozone exposure and respiratory-related ED visits and hospital admissions.
With respect to long-term ozone exposure, there is strong coherence between animal
toxicological studies of changes in lung morphology and epidemiologic studies reporting positive
associations between long-term ozone exposure and new-onset asthma, respiratory symptoms in children
with asthma, and respiratory mortality. Furthermore, the experimental evidence provides biologically
plausible pathways through which long-term ozone exposure could lead to respiratory effects reported in
epidemiologic studies.
Metabolic effects related to ozone exposure are evaluated as a separate health endpoint category
for the first time in this ISA (Appendix 5). Recent evidence from animal toxicological, controlled human
exposure, and epidemiologic studies indicate that there is a "likely to be causal relationship" between
short-term ozone exposure and metabolic effects (Section 5.1.8). The strongest evidence for this
determination is provided by animal toxicological studies that demonstrate impaired glucose tolerance,
increased serum triglycerides, fasting hyperglycemia, and increased hepatic gluconeogenesis in various
stocks/strains of animals across multiple laboratories. Biological plausibility is provided by results from
controlled human exposure and animal toxicological studies that demonstrate activation of sensory nerve
pathways following ozone exposure triggers the central neuroendocrine stress response, which includes
increased corticosterone, Cortisol, and epinephrine production. These findings are coherent with
epidemiologic studies that report associations between ozone exposure and perturbations in glucose and
insulin homeostasis. In addition, these pathophysiological changes are often accompanied by increased
inflammatory markers in peripheral tissues and by activation of the neuroendocrine system.
ES-8
-------
8
"§ +£ 6
O c
3 g
TS E
C O
¦ u 4
m U H
= £
O ^
N -
o >
LU 2
LL.
0
X Adams (2002)
A Adams (2003)
a Adams (2006)
O Horstman et al. (1990)
• Kim etal. (2011)*
O McDonnell et al. (1991)
~ Schelegle et al. (2009)
— McDonnell et al. (2013)
30
40
50 60 70
Ozone (ppb)
80
90
All responses at and above 70 ppb (targeted concentration) were statistically significant (p < 0.05). Adams (20061 found statistically
significant responses to square-wave chamber exposures at 60 ppb based on the analysis of Brown et al. (20081 and Kim et al.
(20111. During each hour of the exposures, subjects were engaged in moderate quasi-continuous exercise (20 L/minute per m2
BSA) for 50 minutes and rest for 10 minutes. Following the 3rd hour, subjects had an additional 35-minute rest period for lunch. The
data at 60 and 80 ppb have been offset on the x axis for illustrative purposes. The solid line illustrates the predicted FENA
decrements using Model 3 coefficients at 6.6 hours as a function of ozone concentration for a 23.8-year-old with a BMI of 23.1 kg/m2
from McDonnell et al. (20131.
*80 ppb data for 30 health subjects were collected as part of the Kim et al. (20111 study, but only published in Figure 5 of McDonnell
et al. (20121.
Adapted from Figure 6-1 of 2013 Ozone ISA (U.S. EPA. 20131. Studies appearing in the figure legend are: Adams (20061. Adams
(20031. Adams (20021. Horstman etal. (19901. Kim etal. (20111. McDonnell et al. (20131. McDonnell et al. (19911. and Schelegle et
al. (20091.
Figure ES-3
Cross-study comparisons of mean decrements in ozone-induced
forced expiratory volume in 1 second (FEVi) in young, healthy
adults following 6.6 hours of exposure to ozone.
Notably, there are changes in the causality determinations for short-term ozone exposure and
cardiovascular effects (Appendix 4). as well as for total mortality (Appendix 6). In both instances, the
evidence synthesized in the 2013 Ozone ISA was sufficient to conclude a "likely to be causal
relationship," but after integrating the previous evidence with recent data, the collective evidence is
"suggestive of, but not sufficient to infer, a causal relationship" between short-term ozone exposure
and cardiovascular effects (Section 4.1.17) or total mortality (Section 6.1.8) in this ISA. The evidence
that supports this change in the causality determinations includes: (1) a growing body of controlled
human exposure studies providing less consistent evidence for an effect of short-term ozone exposure on
ES-9
-------
cardiovascular health endpoints; (2) a paucity of positive evidence from epidemiologic studies for more
severe cardiovascular morbidity endpoints (i.e., heart failure, ischemic heart disease and myocardial
infarction, arrhythmia and cardiac arrest, and stroke); and (3) uncertainties due to a lack of control for
potential confounding by copollutants in epidemiologic studies. Although there is generally consistent
evidence for a limited number of ozone-induced cardiovascular endpoints in animal toxicological studies
and for cardiovascular mortality in epidemiologic studies, these results are not coherent with results from
controlled human exposure and epidemiologic studies examining cardiovascular morbidity endpoints.
There remains evidence for ozone-induced cardiovascular mortality from epidemiologic studies.
However, inconsistent results from a larger number of recent controlled human exposure studies that do
not provide evidence of cardiovascular effects in response to short-term ozone exposure introduce
additional uncertainties.
ES.4.2 Ozone Exposure and Welfare Effects
The scientific evidence for welfare effects of ozone consists mainly of effects on vegetation and
ecosystems (Appendix S) and effects on climate (Appendix 9). For ecological effects, damage to
terrestrial ecosystems as evaluated through controlled exposure studies, observational studies and
modeling based on experimental data, is largely a function of uptake of ozone into the leaf via stomata
(gas exchange openings on leaves). Subsequent reactions with plant tissues alter whole-plant responses
that cascade up to effects at higher levels of biological organization (i.e., from the cellular and subcellular
level to the individual organism up to ecosystem level processes and services; Figure ES-4). At the leaf
level, ozone uptake produces reactive oxygen species that affect cellular function (Section 8.1.3 and
Figure 8-2). Reduced photosynthesis, altered carbon allocation, and impaired stomatal function lead to
observable responses in plants. Observed vegetation responses to ozone include visible foliar injury
(Section IS.5.1.1). and whole-plant level responses (Section IS.5.1.2). which encompass reduction in
aboveground and belowground growth, reproduction and yield. Plant-fauna linkages affected by ozone
include herbivores that feed on ozone-damaged vegetation and interactions of ozone with compounds
emitted by plants that can alter attraction of pollinators to plants (Section IS. 5.1.3). A combination of
observational and experimental data, and modeling output provides evidence for broad changes in
ecosystems such as decreased productivity and carbon sequestration (Section IS.5.1.4). altered
belowground processes (Section IS.5.1.5). terrestrial community composition (Section IS.5.1.6). and
water cycling (Section IS.5.1.7).
ES-10
-------
03 exposure
IS55
0 * y
03 uptake & physiology
•Antioxidant metabolism upregulated
¦Decreased photosynthesis
•Decreased stomatal conductance
or sluggish stomatal response
Effects on leaves
•Visible foliar injury
•Altered leaf production
•Altered leaf chemical composition
Plant growth
•Decreased biomass accumulation
•Altered root growth
•Altered carbon allocation
•Altered reproduction
' -Altered crop quality
1
Belowground processes
•Altered litter production and decomposition
•Altered soil carbon and nutrient cycling
•Altered soil fauna and microbial communities
CD
—I
CD
=3
0
GJ
U)
CD
3
CO
<
1 >
Affected ecosystem services
•Decreased productivity
•Decreased C sequestration
• Decreased crop yield
•Altered water cycling
•Altered community composition
•Altered pollination
•Altered forest products
Source: Adapted from U.S. EPA (2013).
Figure ES-4
Illustrative diagram of ozone effects cascading from the cellular
level to plants and ecosystems.
There are 12 causality determinations for ecological effects of ozone generally organized from
the individual-organism scale to the ecosystem scale in Figure ES-5. Like the findings of the 2013 Ozone
ISA (Table ES-2). five are causal relationships (i.e., visible foliar injury, reduced vegetation growth,
reduced crop yield, reduced productivity, and altered belowground biogeochemical cycles) and two are
likely to be causal relationships (i.e., reduced carbon sequestration, altered ecosystem water cycling). One
of the endpoints, alteration of terrestrial community composition, is now concluded to be a causal
relationship whereas in the 2013 Ozone ISA this endpoint was classified as a likely to be causal
relationship. Three new endpoint categories (i.e., increased tree mortality, alteration of herbivore growth
and reproduction, alteration of plant-insect signaling) not evaluated in the 2013 Ozone ISA, are all
determined to have a likely to be causal relationship with ozone. Plant reproduction, previously
considered as part of the evidence for growth effects, is now a stand-alone causal relationship.
ES-11
-------
Causality Determinations for Ecological Effects of Ozone
Belowground
Biogeochemical Cycles
U
A
>*
Water Cycling
0)
L
0
J
u
Carbon Sequestration
vt
c
o
a
DC
~nj
u
4
>•
3
Biodiversity
Terrestrial Community Composition!
Out
o
z
o
u
kU
o
u
Species Interactions
Plant-Insect Signaling +
«*-
o
J)
Survival
Trees +
CD
U
in
c
o
ro
fO
=;
m
Growth
Plants Herbivores +
3
a-
o
Q.
'¦5
c
Reproduction
Plants + Herbivores +
Yield
Agricultural Crops
Individual
Visible Foliar Injury
1 Causal
Likely Causal |L
+ new causality determination;! causality determination changed from likely to be
causal to causal
Figure ES-5 Causality determinations for ozone across biological scales of
organization and taxonomic groups.
ES-12
-------
Table ES-2 Summary of causality determinations for ecological effects.
Endpoint
Conclusions from
2013 Ozone ISA
Conclusions in the 2020 ISA
Visible foliar injury
Causal relationship
Causal relationship
Reduced vegetation growth
Causal relationship
Causal relationship
Reduced plant reproduction
No separate causality
determination; included with plant
growth
Causal relationship3
Increased tree mortality
Causality not assessed
Likely to be causal relationship3
Reduced yield and quality of agricultural
crops
Causal relationship
Causal relationship
Alteration of herbivore growth and
reproduction
Causality not assessed
Likely to be causal relationship3
Alteration of plant-insect signaling
Causality not assessed
Likely to be causal relationship3
Reduced productivity in terrestrial
ecosystems
Causal relationship
Causal relationship
Reduced carbon sequestration in terrestrial
ecosystems
Likely to be causal relationship
Likely to be causal relationship
Alteration of belowground biogeochemical
cycles
Causal relationship
Causal relationship
Alteration of terrestrial community
composition
Likely to be causal relationship
Causal relationship15
Alteration of ecosystem water cycling
Likely to be causal relationship
Likely to be causal relationship
aDenotes new causality determination.
bDenotes change in causality determination from 2013 Ozone ISA.
Visible foliar injury resulting from exposure to ozone has been well characterized and
documented in over six decades of controlled experimental research involving many tree, shrub,
herbaceous, and crop species and using both long-term field studies and laboratory approaches. Recent
experimental evidence (Section 8.2) continues to show a consistent association between visible injury and
ozone exposure supporting the conclusion of the 2013 Ozone ISA that, there is a "causal relationship"
between ozone and visible foliar injury. Measured changes in photosynthesis and carbon allocation in
ozone-exposed plants scale up to reduced growth documented in natural and managed (e.g., agriculture,
forestry, landscaping) species (Section 8.3). as well as impaired reproduction in individual plants
(Section 8.4.1). Consistent with the conclusions in the 2013 Ozone ISA, there is a "causal relationship"
ES-13
-------
between ozone and reduced plant growth and a "causal relationship" between ozone and reduced
crop yield and quality. In the 2013 Ozone ISA, reproduction was considered in the same category with
plant growth. Increased information on metrics of plant reproduction (e.g., observed flower number, fruit
number, fruit weight, seed number, rate of seed germination) and evidence for direct negative effects on
reproductive tissues as well as for indirect negative effects (resulting from decreased photosynthesis and
other whole-plant physiological changes) warrants a separate causality determination of a "causal
relationship" between ozone exposure and reduced plant reproduction. Since the 2013 Ozone ISA, a
large-scale multivariate analysis of factors contributing to tree mortality (1971-2005) concluded that
county-level ozone concentrations averaged over the study period significantly increased tree mortality in
7 out of 10 plant functional types in the eastern and central U.S. (Section 8.4.3). This evidence, combined
with observations of long-term declines of conifer forests in several high ozone regions and new
experimental evidence that sensitive genotypes of aspen have increased mortality with ozone exposure,
supports a "likely to be causal relationship" between ozone exposure and tree mortality.
In addition to effects on plants, ozone can alter ecological interactions between plants and other
species including herbivores consuming ozone-exposed vegetation. Studies of insect herbivores in
previous ozone assessments and newer experimental studies covering a range of species at varying levels
of ozone exposure frequently show statistically significant effects; however, effects on growth and
reproduction are highly context- and species-specific, and not all species tested show a response
(Section 8.6). The collective evidence supports "a likely to be causal relationship" between ozone
exposure and altered herbivore growth and reproduction. Many plant-insect interactions are mediated
through volatile plant signaling compounds which plants use to signal other community members. In the
2013 Ozone ISA, a few experimental and modeling studies reported altered insect-plant interactions that
are mediated through chemical signaling. New evidence from multiple studies show altered/degraded
emissions of chemical signals from plants and reduced detection of volatile plant signaling compounds by
insects, including pollinators, in the presence of ozone (Section 8.7). The collective evidence supports
"a likely to be causal relationship" between ozone exposure and alteration of plant-insect signaling.
At the ecosystem scale, ozone-caused decreases in plant photosynthesis can lead to reduced
ecosystem carbon content. Changes in patterns of aboveground and belowground carbon allocation
associated with ozone effects on plants can alter ecosystem properties of storage (e.g., productivity,
carbon sequestration) and cycling (e.g., biogeochemistry) through both experimental and modeling
studies. Consistent with the conclusions of the 2013 Ozone ISA, there is a "causal relationship"
between ozone exposure and reduced productivity and a "likely to be causal relationship" between
ozone and reduced carbon sequestration (Section 8.8). As described in the 2013 Ozone ISA and new
experimental studies, processes such as carbon and nitrogen cycling and decomposition in soils are
indirectly affected via ozone effects on the quality and quantity of carbon supply from plants and leaf
litter (Section 8.9). Recent evidence continues to support a "causal relationship" between ozone
exposure and the alteration of belowground biogeochemical cycles. Ozone can affect water use in
plants through several mechanisms including damage to stomatal functioning, loss of leaf area, and
ES-14
-------
changes in wood anatomy (e.g., vessel size and density) that can affect plant and stand evapotranspiration
and may lead, in turn, to possible effects on hydrological cycling as shown through a combination of
experimental data and modeling (Section 8.11). Evidence continues to support the conclusion of the 2013
Ozone ISA that, there is a "likely to be causal relationship" between ozone and alteration of
ecosystem water cycling. In terrestrial ecosystems, ozone may alter community composition by uneven
effects on co-occurring species, decreasing the abundance of sensitive species, and giving tolerant species
a competitive advantage. Alteration of community composition of some ecosystems including conifer
forests, broadleaf forests, and grasslands and altered fungal and bacterial communities in soils reported in
the 2013 Ozone ISA is augmented by additional experimental and modeling evidence for effects in forest
and grassland communities (Section 8.10); collective evidence indicates a change in the causality
determination to a "causal relationship" between ozone exposure and altered terrestrial community
composition of some ecosystems.
For effects on climate, changes in the abundance of tropospheric ozone perturb the radiative
balance of the atmosphere by interacting with incoming solar radiation and outgoing longwave radiation.
This effect is quantified by radiative forcing.1 Through this effect on the Earth's radiation balance,
tropospheric ozone plays a major role in the climate system and increases in tropospheric ozone
abundance contribute to climate change. Recent evidence continues to support a "causal relationship"
between tropospheric ozone and radiative forcing and a 'likely to be causal relationship," via
radiative forcing, between tropospheric ozone and temperature, precipitation, and related climate
variables (referred to as "climate change" in the 2013 Ozone ISA; the revised title for this causality
determination provides a more accurate reflection of the available evidence; Table ES-3). The new
evidence comes from the Intergovernmental Panel on Climate Change (IPCC) Fifth Assessment Report
(AR5) and its supporting references, as well as a limited number of more recent studies, and builds on
evidence presented in the 2013 Ozone ISA. The new studies further support the causality determinations
included in the 2013 Ozone ISA.
Table ES-3 Summary of causality determinations for tropospheric ozone effects
on climate.
Conclusions in 2013 Ozone ISA
Conclusions in the 2020 ISA
Radiative forcing
Causal relationship
Causal relationship
Temperature, precipitation, and related Likely to be causal relationship Likely to be causal relationship
climate variables
1 Radiative forcing is the perturbation in net radiative flux at the tropopause (or top of the atmosphere) caused by a
change in radiatively active forcing agent(s) after stratospheric temperatures have readjusted to radiative equilibrium
[stratospherically adjusted radiative forcing; Mvhre et al. (20f3VI.
ES-15
-------
ES.5 Key Aspects of Health and Welfare Effects Evidence
There is extensive scientific evidence that demonstrates health and welfare effects from exposure
to ozone. As documented by the evaluation of evidence throughout the subsequent Appendices to this
ISA, the U.S. EPA carefully considers uncertainties in the evidence, the extent to which recent studies
have addressed or reduced uncertainties from previous assessments, and the strengths of the evidence.
Uncertainties do not necessarily change the fundamental conclusions of the literature base. In fact, some
conclusions are robust to such uncertainties. Where there is clear evidence linking ozone with health and
welfare effects—with or despite remaining uncertainties—the U.S. EPA makes a determination of a
causal or likely to be causal relationship. The identification of the strengths and limitations in the
evidence will help in the prioritization of research efforts to support future ozone NAAQS reviews.
ES.5.1 Health Effects Evidence: Key Findings
A large body of scientific evidence spanning many decades clearly demonstrates there are health
effects related to both short- and long-term exposure to ozone. The strongest evidence supports a
relationship between ozone exposure and respiratory health effects. The collective body of evidence for
each health outcome category evaluated in this ISA is considered systematically and assessed; this
assessment includes evaluation of the inherent strengths, limitations, and uncertainties in the overall body
of evidence for the health outcome, resulting in the causality determinations detailed in Table ES-1.
An inherent strength of the evidence integration in this ISA is the extensive amount (in both
breadth and depth) of available evidence resulting from decades of scientific research that describes the
relationship between both short- and long-term ozone exposure and health effects. The breadth of the
enormous database is illustrated by the different scientific disciplines that provide evidence
(e.g., controlled human exposure, epidemiologic, animal toxicological studies), the range of health
outcomes examined (e.g., respiratory, cardiovascular, metabolic, reproductive, and nervous system
effects, cancer and mortality), and the large number of studies within several of these outcome categories.
The depth of the literature base is exemplified by the examination of effects that range from biomarkers
of exposure, to subclinical effects, to overt clinical effects, and even mortality.
There is strong and consistent experimental evidence linking short- and long-term ozone exposure
with respiratory effects and short-term ozone exposure with metabolic health effects. However, several
uncertainties should be considered when evaluating and synthesizing evidence from these studies.
Experimental animal studies are often conducted at ozone concentrations higher than those observed in
ambient air (i.e., 250 to >1,000 ppb) to evoke a response within a short time period. These studies are
informative and the conduct of studies at these concentrations is commonly used for identifying potential
human hazards. There are also substantial differences in exposure concentrations and exposure durations
between animal toxicological and controlled human exposure studies. Additionally, a number of animal
toxicological studies are performed in rodent models of disease states, while controlled human exposure
studies generally are conducted in healthy individuals. Controlled human exposure studies do not
ES-16
-------
typically include unhealthy or diseased individuals for ethical reasons; therefore, this exclusion represents
an important uncertainty to consider in interpreting the results of these studies (i.e., that other individuals
may be more sensitive and at risk to ozone than those in the study groups). Additional factors that differ
between human and experimental animal exposures include: exposure concentration and disease status;
differences in physiology (e.g., rodents are obligate nose breathers); differences in the duration and timing
of exposure (e.g., rodents are exposed typically during the day, during their resting cycle, while humans
are exposed during the day when they are normally active); and differences in the temperature at which
the exposure was conducted. These factors may contribute to any lack of coherence between results of
experimental animal and human studies. Despite these factors, there is consistent and coherent evidence
that spans scientific disciplines for respiratory and metabolic health effects.
Epidemiologic studies contribute important evidence supporting the relationship between short-
and long-term ozone exposure with respiratory effects. Although susceptible to chance, bias, and other
potential confounding due to their observational nature, epidemiologic studies have the benefit of
evaluating real-world exposure scenarios and can include sensitive populations that cannot typically be
included in controlled human exposure studies. Innovations in epidemiologic study designs and methods
have substantially reduced the role of chance, bias, and other potential confounders in well-designed,
well-conducted epidemiologic studies. The most common source of uncertainty in epidemiologic studies
of ozone is exposure measurement error. The exposure assignment methods used in short- and long-term
ozone exposure epidemiologic studies have inherent strengths and limitations, and exposure measurement
errors associated with those methods contribute bias and uncertainty to health effect estimates. For
short-term exposure studies, exposure measurement error generally leads to underestimation and reduced
precision of the association, whereas in long-term exposure studies exposure measurement error has the
potential to bias effect estimates in either direction, although it is more common that they are
underestimated. Furthermore, disentangling the effects of short-term ozone exposure from those of
long-term ozone exposure (and vice versa) is an inherent uncertainty in the evidence base. When
combined with coherent evidence from animal toxicological and controlled human exposure studies, the
epidemiologic evidence can support and strengthen determinations of the causal nature of the relationship
between health effects and exposure to ozone at relevant ambient air concentrations.
ES.5.2 Welfare Effects Evidence: Key Findings
The collective body of evidence for each welfare endpoint evaluated in this ISA was carefully
considered and assessed, including the inherent strengths, limitations, and uncertainties in the overall
body of evidence, resulting in the causality determinations for ecological effects detailed in Table ES-2
and effects on climate in Table ES-3. A large body of scientific evidence spanning more than 60 years
clearly shows effects on vegetation due to ozone exposure. Decades of research on many plant species
confirm effects on visible foliar injury, plant growth, reproduction and yield. The use of visible foliar
injury to identify phytotoxic levels of ozone is an established and widely used methodology. There are
robust exposure-response functions for reduced growth and yield (i.e., from carefully controlled
ES-17
-------
experimental conditions, involving multiple concentrations and based on multiple studies) for about a
dozen important tree species and a dozen major commodity crop species. Newer evidence supports a role
for ozone in tree mortality and shifts in community composition of forest tree and grassland species.
While the effect of ozone on vegetation is well established in general, there are some knowledge gaps
regarding precisely which species are sensitive, what exposures elicit adverse responses for many species
and how plant response changes with age and size.
There is high certainty in ozone effects on impairment to leaf physiology as mechanisms for
effects at higher levels of biological organization (i.e., from the cellular level through individual
organisms to the level of communities and ecosystems) and how those can ultimately affect aboveground
and belowground processes such as productivity, carbon sequestration, biogeochemical cycling, and
hydrology. However, ecosystems are inherently complex, and it is difficult to partition observed
responses within a suite of multiple stressors. Scaling ozone effects to the ecosystem level remains a
challenge, but there is a large body of knowledge of how ecosystems work through ecological
observations and models. Interactive effects in natural ecosystems with multiple stressors (e.g., drought,
disease) are difficult to study, but some have been investigated using different statistical methods.
Although models and methods for characterizing ecosystem-level responses to ozone are accompanied by
inherent uncertainties, more research will strengthen understanding of scaling across different levels of
biological organization.
There are multiple pathways in which ozone can affect plant-insect interactions. Studies that
characterize volatile plant signaling compounds in ozone-enriched environments and assess insect
response to altered chemical signals suggest that ozone alters scent-mediated interactions in ecological
communities. A relatively small number of insect species and plant-insect associations have been
assessed, and there are knowledge gaps in the mechanisms and consequences of modulation of signaling
by ozone. There are multiple studies demonstrating ozone effects on fecundity and growth in insects that
feed on ozone-exposed vegetation. However, no consistent directionality of response is observed across
studies and uncertainties remain in regard to different plant consumption methods across species and the
exposure conditions associated with particular severities of effects.
Changes in the abundance of tropospheric ozone affect radiative forcing, and thus tropospheric
ozone is considered an important greenhouse gas. The recent IPCC AR5 estimates global tropospheric
ozone radiative forcing to be 0.40 (0.20 to 0.60) W/m2 and recent studies reinforce the AR5 estimates.
Consistent with previous estimates, the effect of global, total tropospheric ozone increases on global mean
surface temperature, through its impact on radiative forcing, continues to be estimated at roughly 0.1 to
0.3°C since preindustrial times with larger effects regionally. Some new research has explored certain
additional aspects of the climate response to ozone radiative forcing beyond global and regional
temperature change. Specifically, ozone changes are understood to have impacts on other climate metrics
such as precipitation and atmospheric circulation patterns, and new evidence has continued to support and
further quantify this understanding.
ES-18
-------
While the warming effect of tropospheric ozone in the climate system is well established in
general, precisely quantifying changes in surface temperature due to tropospheric ozone changes, along
with related climate effects, requires complex climate simulations that include all relevant feedbacks and
interactions. For example, trends in free tropospheric ozone and upper tropospheric ozone (where
radiative forcing is particularly sensitive to changes in ozone concentrations) are not captured well by
models. In addition, substantial variation exists across models. Such modeling uncertainties make it
especially difficult to provide precise quantitative estimates of the climate effects of regional-scale ozone
changes. Uncertainties in estimates of preindustrial ozone concentrations represent another important
source of uncertainty in climate effects resulting from long-term ozone concentration changes.
ES-19
-------
S.6 References
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/10.1080/0895837029008461Q
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.
http://dx.doi.org/10.1080/0895837039Q168283
Adams. WC. (2006). Comparison of chamber 6.6-h exposures to 0.04-0.08 ppm ozone via square-
wave and triangular profiles on pulmonary responses. Inhal Toxicol 18: 127-136.
http://dx.doi.org/10.1080/089583705003061Q7
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
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.
http://dx.doi.Org/10.l 164/airccm/142.5.1158
Kim. CS; Alexis. NE: Rappold. AG; Kehrl. H: Hazucha. MJ: Lav. 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.201011-1813QC
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. http://dx.doi.org/10.1080/00039896.1991.9937441
McDonnell. WF; Stewart. PW; Smith. MY. (2013). Ozone exposure-response model for lung function
changes: An alternate variability structure. Inhal Toxicol 25: 348-353.
http://dx.doi.org/10.3109/08958378.2013.79Q523
McDonnell. WF; Stewart. PW; Smith. MY; 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. http://dx.doi.org/10.3109/08958378.2012.7Q5919
Mvhre. G; Samset. BH; Schulz. M; Balkanski. Y; Bauer. S; Berntsen. TK; Bian. H; Bellouin. N: Chin.
M; Diehl. T; Easter. RC; Feichter. J; Ghan. SJ; Hauglustaine. D; Iversen. T; Kinne. S; Kirkevag. A;
Lamarque. JF; Lin. G; Liu. X; Lund. MT; Luo. G; Ma. X; van Noiie. T; Penner. JE; Rasch. PJ;
Ruiz. A; Seland. O; Skeie. RB; Stier. P; Takemura. T; Tsigaridis. K; Wang. P; Wang. Z; Xu. L;
Yu. H; Yu. F; Yoon. JH; Zhang. K; Zhang. H; Zhou. C. (2013). Radiative forcing of the direct
aerosol effect from AeroCom Phase II simulations. Atmos Chem Phys 13: 1853-1877.
http://dx.doi.org/10.5194/acp-13-1853-2013
Schelegle. ES; Morales. CA; Walbv. 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.200809-1484QC
U.S. EPA (U.S. Environmental Protection Agency). (2013). Integrated science assessment for ozone
and related photochemical oxidants [EPA Report]. (EPA/600/R-10/076F). Research Triangle Park,
NC: U.S. Environmental Protection Agency, Office of Research and Development, National Center
for Environmental Assessment-RTP Division.
http://cfpub.epa.gov/ncea/isa/recordisplav.cfm?deid=247492
ES-20
-------
U.S. EPA (U.S. Environmental Protection Agency). (2015). Preamble to the integrated science
assessments [EPA Report]. (EPA/600/R-15/067). Research Triangle Park, NC: U.S. Environmental
Protection Agency, Office of Research and Development, National Center for Environmental
Assessment, RTP Division. https://cfpub.epa.gov/ncea/isa/recordisplav.cfm?deid=310244
ES-21
-------
INTEGRATED SYNTHESIS
Overall ( uncltishms of the Ozone Integrated Science. \ssessnient (IS. 1)
Unman Health l:JJect\
• Reccul siudies support ;iud exp;uid upon ihe siroim hod> ofe\ ideuee. which h;is heeu
;ijiimill;iliii'-i o\er iii;iii\ decides. Ili;il shori-lcrui o/oue exposure c;iuses respircilorx
effects I lie stroimcsi e\ ideuee comes from controlled Ihiiikiii exposure studies
dcuiousir;iliim o/oiic-iuduced decreases mi hum function ;ind i11l l;iInni;iIkhi in he;illh\.
exercisum ;idulls ;ii coiicciiir;ilious ;is low ;is(i() pph ;ifler<> <> hours of exposure. In
iiddiliou. epidemiologic studies continue li» pro\ ide sirouu e\ ideuee lh;il o/one is
;issoci;iled \x ilh ivspir;iuii"\ effects. uicludiim ;isihm;i ;ind chronic ohsirucli\e
piilnioii;ir\ disease cx;iccrh;ilious. ;is well ;is hospii;il ;idniissious ;md cnicrucucs
dcp;irinicui \ isiis for respir;iior\ discuses The resiilis from loxicolomc;il siudies
further ch;ir;iclcri/e poleuli;il nieeh;inis|ie p;ilhw;i\s ;iud pix>\ ide continued support lor
I he hiolomc;il pl;iusihihl\ of o/ouc-iiiduced respimiorx effects
• There is enicrmim e\ idenee lh;ii shori-terni o/oue exposure coiiirihuies to niel;iholic
disense. iiicludum complications rehiled Io di;iheles. Speeilie;iM>. ;iniin;il loxicolomcnl
siudies demons!r;iie lh;ii o/oue e\posure impaired ulucose loler;niee. inere;ised
limKcerides in serum. fiisium h\perul\eenn;i. ;md mere;ised hep;ilie ulucoiicoucuesis
• The iiiicumiiou ol" reeeui e\ ideuee from eouirolled hiim;m exposure siudies show mu
iiieousisieul e\ ideuee of o/oue-iudiieed c;irdio\ ;iseul;ir el leels w ilh ihe o\ er;ill hods of
e\ ideuee lor ;iu ;issoei;iliou of shori-lerui o/one exposure w ilh e;irdio\ ;iseul;ir el leels
;iud lol;il (uou;ieeideul;il i niori;ihl\ ;i\;nl;ihle in I he 2d I i ()Ad \e;irs clc;irl\ deniousir;iics
lh;ii o/oue ;il fccls \euel;iliou ;iud ecosvsicnis The sirouuesi e\ ideuee comes from
\ euel;iliou-rel;iled eudpoiuis. folmr i11111rx. reduced mow ill. ;ind decreased \ icld
resu I li nu from o/oue exposure ;ire well ch;ir;icleri/ed in ;i x ;i nelx of crop mid iioucrop
species. I !colomc;il el leels of o/oue ;ire obser\ ed ;icross se\ er;il sc;iles of biolomc;il
oru;uii/.;iliou ne. from ihe celluhir le\ el ihrouuh niclix idu;il oru;iuisnis io ihe lex el of
coniuiiiuiiics ;md ecos\ sieuisi. iiliini;iicl> ;ilfccliim ;iho\ curouud ;ind hclow mound
processes iiicludiuu producln il>. c;irhou sct|uesir;iliou. hioueocheuiic;il c\chim;iud
h\drolou\ lu uiosi c;ises. new research sircimiheus ihe pre\ lousk re;iched
conclusions mi ihe 2D I i O/oue IS \ New eudpoiuis included in llns re\ lew resu 11 from
cnicrmuu ;ire;is of siud\ such ;is cheniic;il ecolouv (e.u . pkiiii-iusecl simi;ilumi or new
e\ ideuee eii;ihlum furl her refiuenieui of pre\ ioiisK uudeiMood o/oue elleels
ten. pl;ml reproduclioii. iree niori;ilii\. I le rh i x tire mow ill ;md reprodiicliou. lerresi ri;il
commuuiis composition)
I.ITccts on ( liuiiilo
• New resenreh hiulds tin ihe e\ ideuee in ihe 2d I ' ()/oue IS \ ;iud coiiiiiiues io supporl
ihe pre\ ious findings of iropospheric o/oue unp;icls on r;idi;ili\ e forcum ;iud chui;ile
\;iri;ihles. iiicludiuu iemper;iiure ;iud precipitation (referred lo ;is 'clini;iie ckiime" in
ihe 2d I ' ()/oue IS \ i
IS-1
-------
IS,!
Introduction
IS.1.1 Purpose and Overview
The Integrated Science Assessment (ISA) serves as the scientific foundation of the National
Ambient Air Quality Standard (NAAQS) review process.1 The ISA is a comprehensive evaluation and
synthesis of the policy-relevant science "useful in indicating the kind and extent of all identifiable effects
on public health or welfare2 which may be expected from the presence of [a] pollutant in the ambient air,"
as described in Section 108 of the Clean Air Act (42 U.S. Code [U.S.C.] 7408).3 This ISA reviews and
synthesizes the air quality criteria for the health and welfare effects of ozone and related photochemical
oxidants in ambient air. It draws on the existing body of evidence to evaluate and describe the current
state of scientific knowledge on the most relevant issues pertinent to the current review of the ozone
NAAQS,4 to identify changes in the scientific evidence since the previous review, and to describe
remaining or newly identified uncertainties and limitations in the evidence. In 2015, the U.S. EPA
lowered the level of the primary and secondary ozone standards to 0.070 ppm and maintained the form of
the standard as the annual fourth-highest daily max 8-hour concentration averaged over 3 years.5 The
ozone primary NAAQS is established to protect public health, including at-risk populations such as
children and people with asthma, with an adequate margin of safety. The ozone secondary standard is
intended to protect the public welfare from known or anticipated adverse effects associated with the
presence of ozone and related photochemical oxidants in the ambient air.
This ISA identifies and critically evaluates the most policy-relevant current scientific literature
published since the 2013 Ozone ISA across scientific disciplines including epidemiology, controlled
human exposure studies, experimental animal toxicology, atmospheric science, exposure science,
vegetation studies, agricultural science, ecology, and climate-related science. Key scientific conclusions
(e.g., causality determinations; Section IS. 1.2.4) are presented that provide the basis for developing risk
1 Section 109(d)(1) of the Clean Air Act requires periodic review and, if appropriate, revision of existing air quality
criteria to reflect advances in scientific knowledge on the effects of the pollutant on public health and welfare. Under
the same provision, U.S. EPA is also to periodically review and, if appropriate, revise the NAAQS, based on the
revised air quality criteria.
2 Under CAA section 302(h), effects on welfare include, but are not limited to, "effects on soils, water, crops,
vegetation, manmade materials, animals, wildlife, weather, visibility, and climate, damage to and deterioration of
property, and hazards to transportation, as well as effects on economic values and on personal comfort and
well-being."
3 The general process for developing an ISA, including the framework for evaluating weight of evidence and
drawing scientific conclusions and causal judgments, is described in a companion document, Preamble to the
Integrated Science Assessments (U.S. EPA. 20151.
4 The "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 of the current NAAQS for photochemical oxidants is ozone
5 Final rule signed October 1, 2015 and effective December 28, 2015 (80 FR 65291).
IS-2
-------
and exposure analyses, evaluating policy, and making environmental health and welfare decisions. In
characterizing the evidence for each of the health and welfare effects categories evaluated, this ISA draws
conclusions about the causal nature of the relationships between ozone exposure and outcomes by
integrating information across scientific disciplines and synthesizing evidence from previous and recent
studies. As in previous reviews, the ISA for this review focuses mainly on the assessment of health and
welfare effects resulting from exposure to concentrations of tropospheric ozone. Ozone is currently the
NAAQS indicator for photochemical oxidants, and the primary literature evaluating the health and
ecological effects of photochemical oxidants includes ozone almost exclusively as an indicator of
photochemical oxidants.1 This ISA thus provides the policy-relevant scientific information that supports
the review of the current ozone NAAQS.
IS.1.2 Process and Development
Through iterative NAAQS reviews, ISAs build on evidence and conclusions from previous
assessments. The previous ozone ISA was published in 2013 (U.S. EPA. 2013b) and included
peer-reviewed literature published through July 2011. Prior assessments include the 2006 Air Quality
Criteria Document (AQCD) for Ozone and Related Photochemical Oxidants (U.S. EPA. 2006a). the 1996
AQCD for Ozone (U.S. EPA. 1996a). the 1986 AQCD for Ozone (U.S. EPA. 1986). the 1978 Air Quality
Criteria for Ozone and Other Photochemical Oxidants (U.S. EPA. 1978). and the 1970 Criteria Document
(NAPCA. 1970). This ISA focuses on synthesizing and integrating the new evidence (i.e., studies
published between January 2011 and March 2018, as well as more recent studies identified during peer
review or by public comments) with the information and conclusions from previous assessments.
In the process of developing an ISA, systematic review methodologies are used to identify and
evaluate relevant scientific information, which is synthesized into text and figures for the purpose of
communicating the state of the science. The process begins with a "Call for Information" published in the
Federal Register that announces the start of the NAAQS review and invites the public to assist in this
process by identifying relevant research studies in the subject areas of concern. For this Ozone NAAQS
review, the Federal Register notice was published on June 26, 2018 (83 FR 29785). The subsequent ISA
development steps are described in greater detail in the Preamble to the Integrated Science Assessments
(U.S. EPA. 2015). which provides a general overview of the process. The Preamble describes the general
framework for evaluating scientific information, including criteria for assessing study quality and
developing scientific conclusions. The U.S. EPA uses a structured and transparent process to evaluate
scientific information and to determine the causal nature of relationships between air pollution and health
1 Ozone is the only photochemical oxidant other than nitrogen dioxide (NO2) that is routinely monitored in ambient
air (i.e., U.S. EPA's AQS database; https://www.epa.gov/aas'). Data for other photochemical oxidants (e.g., PAN,
H2O2, etc.) typically have been obtained only as part of special field studies. Consequently, no data on nationwide
patterns of ambient air concentrations are available for these other photochemical oxidants; nor are extensive data
available on the relationships of concentrations and patterns of these photochemical oxidants to those of ozone.
IS-3
-------
and welfare effects [see Preamble (U.S. EPA. 2015)1. Development of the ISA includes approaches for
literature searches, application of criteria for selecting and evaluating relevant studies, and application of
framework for evaluating the weight of evidence and forming causality determinations. As part of this
process, the ISA is reviewed by the public and by the Clean Air Scientific Advisory Committee
(CASAC), which is a formal, independent scientific committee (Section 10.4). The Preamble describes a
science and policy workshop that often occurs at the beginning of the NAAQS review process; such a
workshop was not convened for the current Ozone NAAQS review. Instead, the "Call for Information"
published in the Federal Register requested public input on science and polity issues pertinent to the
Ozone NAAQS review.
IS.1.2.1 Scope of the ISA and the Population, Exposure, Comparison, Outcome, and
Study Design (PECOS) Tools
The Ozone ISA includes research relevant to characterizing ozone in ambient air (hereafter
referred to as ambient ozone) and assessing the health and welfare effects of exposure to ambient ozone.
Health effects evidence evaluated in the ISA includes experimental controlled human exposure and
animal toxicological studies, and observational epidemiologic studies. Welfare-based evidence included
in the Ozone ISA focuses specifically on ecological effects and effects on climate. The evidence
connecting tropospheric ozone and UV-B (short-wave ultraviolet rays) shielding was evaluated in the
2013 Ozone ISA and determined to be inadequate to draw a causal conclusion. The current ISA concludes
there was no new evidence since the 2013 Ozone ISA relevant to the question of UV-B shielding by
tropospheric ozone, including the incremental effects of tropospheric ozone concentration changes on
UV-B. (Section 9.1.3.4). and this topic is not discussed further in this synthesis.
The scope of the health and welfare effects evidence evaluated in this ISA is further refined by
using the Population, Exposure, Comparison, Outcome, and Study Design (PECOS) tool. The PECOS
tools provide structured frameworks for defining the scope of the ISA. There are discipline-specific
PECOS tools for experimental and epidemiologic studies (Section 3.1.2. Section 3.2.2. Section 4.1.2.
Section 4.2.1.1. Section 5.1.1. Section 5.2.1. Section 6.1.1.1. Section 6.2.1.1. Section 7.1.1.1.
Section 7.2.1.1. Section 7.2.2.1. and Section 7.3.1.1). ecological studies (Table 8-2). and studies of the
effects of tropospheric ozone on climate (Table 9-1). These PECOS criteria were developed with
consideration of the evidence base available at the time of the last review (i.e., the causality
determinations from the 2013 Ozone ISA) and the uncertainties and limitations associated with that
evidence. The use of PECOS tools is a widely accepted and rapidly growing approach to systematic
review in risk assessment, and their use is consistent with recommendations by the National Academy of
Sciences for improving the design of risk assessment through planning, scoping, and problem formulation
to better meet the needs of decision makers (NRC. 2009). The PECOS tools serve as guides for the
inclusion or exclusion of studies in the ISA. Additional details on the development and use of these
PECOS tools can be found in Appendix 10 (Section 10.3.1).
IS-4
-------
IS.1.2.2 Organization of the ISA
The ISA consists of the Preface (legislative requirements and history of the primary and
secondary ozone NAAQS; and purpose and overview of the ISA along with the overall scope, and
process for evaluating evidence), Executive Summary, Integrated Synthesis, and 10 Appendices. This
Integrated Synthesis provides the key information for each topic area, encompassing a description of
ozone concentrations in the U.S. (including background sources), conclusions regarding the health and
welfare effects associated with ozone exposure (including causality determinations for relationships
between exposure to ozone and specific types of health and welfare effects), identification of the human
lifestages and populations at increased risk of the effects of ozone, and a discussion of the key strengths,
limitations, and uncertainties inherent in this evidence base. The purpose of this Integrated Synthesis is
not to summarize each of the Appendices; rather it is to synthesize the key findings on each topic
considered in characterizing ozone exposure and relationships with health and welfare effects. This
Integrated Synthesis also discusses additional policy-relevant issues. These include exposure durations,
metrics, and concentrations eliciting health and welfare effects and the concentration-response (C-R)
relationships for specific effects, including their overall shapes and the evidence with regard to
discernibility of threshold exposures below which effects are unlikely to occur. Subsequent
Appendix 1-Appendix 10 are organized by subject area, with the detailed assessment of atmospheric
science (Appendix 1). exposure (Appendix 2). health (Appendix 3-Appendix 7). and welfare evidence
(Appendix 8-Appendix 9). Each of the Appendices contain an evaluation of results from recent studies
integrated with evidence from previous reviews. Appendices for each broad health effect category
(e.g., respiratory effects) discuss potential biological pathways and conclude with a causality
determination describing the strength of the evidence between exposure to ozone and the outcome(s)
under consideration. Likewise, the Appendices devoted to ecological (Appendix S) and climate evidence
(Appendix 9) for welfare effects include causality determinations for multiple effects on ecosystems and
climate, respectively. Appendix 10 describes the process of developing the ozone ISA, including aspects
related to systematic review of the literature, evaluation of study quality, and quality assurance (QA) and
quality control (QC) documentation.
IS,1,2,3 Quality Assurance Summary
The use of QA and peer review helps ensure that the U.S. EPA conducts high-quality science
assessments that can be used to help policymakers, industry, and the public make informed decisions.
Quality assurance activities performed by the U.S. EPA ensure that environmental data are of sufficient
quantity and quality to support the Agency's intended use. The U.S. EPA has developed a detailed
Program-level QA Project Plan (PQAPP) for the ISA Program to describe the technical approach and
associated QA/QC procedures associated with the ISA Program. All QA objectives and measurement
criteria detailed in the PQAPP have been employed in developing this ISA. Furthermore, the Ozone ISA
is classified as a Highly Influential Scientific Assessment (HISA), which is defined by the Office of
IS-5
-------
Management and Budget (OMB) as a scientific assessment that is novel, controversial, or
precedent-setting, or has significant interagency interest (OMB. 2004). OMB requires a HIS A to be peer
reviewed before dissemination. To meet this requirement, the U.S. EPA engages the Clean Air Scientific
Advisory Committee (CASAC) as an independent federal advisory committee to conduct peer reviews.
Both peer-review comments provided by the CASAC panel and public comments submitted to the panel
during its deliberations about the external review draft were considered in the development of this ISA.
For a more detailed discussion of peer review and quality assurance, see Section 10.4 and Section 10.5.
respectively.
IS.1.2.4 Evaluation of the Evidence
This ISA draws conclusions about the causal nature of relationships between exposure to ozone
and categories of related health and welfare effects (e.g., respiratory effects) by integrating recent
evidence across scientific disciplines and building on the evidence from previous assessments.
Determinations are made about causation, not just association, and are based on judgments of
consistency, coherence, and biological plausibility of observed effects, and on related uncertainties. The
ISA uses a formal causal framework to classify the weight of evidence using a five-level hierarchy
[i.e., "causal relationship"; "likely to be causal relationship"; "suggestive of, but not sufficient to infer, a
causal relationship"; "inadequate to infer the presence or absence of a causal relationship"; or "not likely
to be a causal relationship" as described in Table II of the Preamble (U.S. EPA. 2015)1 that is based
largely on the aspects for causality proposed by Sir Austin Bradford Hill, as well as other frameworks to
assess causality developed by other organizations.
IS.1.3 New Evidence Evaluation and Causality Determinations
In the 2013 Ozone ISA, the causality determinations communicated the extent of the then current
knowledge of health and welfare effects. Updates to the causality determinations for ozone based on new
evidence in this review are summarized below and are described in greater detail in Section IS.4 (Health)
and Section IS.5 (Welfare).
IS.1.3.1 Human Health
The results from the health studies, supported by the evidence from atmospheric chemistry and
exposure assessment studies, contribute to the causality determinations made for the health outcomes. The
conclusions from the 2013 Ozone ISA and the causality determinations from this review are summarized
in Table IS-1.
IS-6
-------
Table IS-1 Summary of causality determinations by exposure duration and
health outcome.
Health Outcome3
Conclusions from 2013 Ozone
ISA
Conclusions in the 2020 ISA
Short-term exposure to ozone
Respiratory effects
Causal relationship
Causal relationship
Cardiovascular effects
Likely to be causal relationship
Suggestive of, but not sufficient to infer, a causal
relationship0
Metabolic effects
No determination made
Likely to be causal relationship15
Total mortality
Likely to be causal relationship
Suggestive of, but not sufficient to infer, a causal
relationship0
Central nervous system
effects
Suggestive of a causal relationship11
Suggestive of, but not sufficient to infer, a causal
relationship
Long-term exposure to ozone
Respiratory effects
Likely to be causal relationship
Likely to be causal relationship
Cardiovascular effects
Suggestive of a causal relationship11
Suggestive of, but not sufficient to infer, a causal
relationship
Metabolic effects
No determination made
Suggestive of, but not sufficient to infer, a causal
relationship15
Total mortality
Suggestive of a causal relationship11
Suggestive of, but not sufficient to infer, a causal
relationship
Reproductive effects
Suggestive of a causal relationship11
Effects on fertility and reproduction: suggestive of, but
not sufficient to infer, a causal relationship15
Effects on pregnancy and birth outcomes: suggestive of,
but not sufficient to infer, a causal relationship15
Central nervous system
effects
Suggestive of a causal relationship11
Suggestive of, but not sufficient to infer, a causal
relationship
Cancer
Inadequate to infer a causal
relationship®
Inadequate to infer the presence or absence of a causal
relationship
aHealth effects (e.g., respiratory effects, cardiovascular effects) include the spectrum of outcomes, from measurable subclinical effects
(e.g., decrements in lung function, blood pressure) to observable effects (e.g., medication use, hospital admissions) and cause-specific mortality.
Total mortality includes all-cause (nonaccidental) mortality, as well as cause-specific mortality.
bDenotes new causality determination.
°Denotes change in causality determination from 2013 Ozone ISA.
dSince the 2013 Ozone ISA, the causality determination language has been updated and this category is now stated as suggestive of, but not
sufficient to infer, a causal relationship.
eSince the 2013 Ozone ISA, the causality determination language has been updated and this category is now stated as inadequate to infer the
presence or absence of a causal relationship.
IS-7
-------
The strongest evidence for health effects due to ozone exposure continues to come from studies
of short- and long-term ozone exposure and respiratory health. Consistent with conclusions from the 2013
Ozone ISA, it is determined that there is a "causal relationship" between short-term ozone exposure and
respiratory effects, and there is a "likely to be causal relationship" between long-term ozone exposure and
respiratory effects. For short-term ozone exposure, controlled human exposure studies provide
experimental evidence for ozone-induced lung function decrements, respiratory symptoms, and
respiratory tract inflammation. Epidemiologic studies continue to provide evidence that increased ozone
concentrations are associated with a range of respiratory effects, including asthma exacerbation, chronic
obstructive pulmonary disease (COPD) exacerbation, respiratory infection, and hospital admissions and
ED visits for combined respiratory diseases. A large body of experimental animal toxicological studies
demonstrates ozone-induced changes in measures of lung function, inflammation, increased airway
responsiveness, and impaired lung host defense. These animal studies also inform the potential
mechanisms underlying downstream respiratory effects (e.g., respiratory tract inflammation) and thereby
provide strong support for the biological plausibility of epidemiologic associations between short-term
ozone exposure and respiratory-related ED visits and hospital admissions. With respect to long-term
ozone exposure, there is strong coherence between animal toxicological studies of changes in lung
morphology and epidemiologic studies reporting positive associations between long-term ozone exposure
and new-onset asthma, and respiratory symptoms in children with asthma. Furthermore, the experimental
evidence provides biologically plausible pathways through which long-term ozone exposure could lead to
the types of respiratory effects reported in epidemiologic studies.
Metabolic effects related to ozone exposure are evaluated as a separate health endpoint category
for the first time in this ISA. Recent evidence from animal toxicological, controlled human exposure, and
epidemiologic studies support a "likely to be causal relationship" between short-term ozone exposure and
metabolic effects. The strongest evidence for this determination is provided by animal toxicological
studies that demonstrate impaired glucose tolerance, fasting hyperglycemia, and increased serum
triglycerides and free fatty acids in various stocks/strains of animals across multiple laboratories.
Biological plausibility is provided by results from controlled human exposure and animal toxicological
studies that demonstrate activation of sensory nerve pathways following ozone exposure that trigger the
central neuroendocrine stress response, as indicated by increased corticosterone/cortisol and adrenaline
production. These findings are coherent with epidemiologic studies that report associations between
ozone exposure and perturbations in glucose and insulin homeostasis. In addition, these
pathophysiological changes are often accompanied by increased inflammatory markers in peripheral
tissues and by changes in liver biomarkers.
The strongest evidence for metabolic effects following long-term ozone exposure is provided by
epidemiologic studies. In prospective cohort studies in the U.S. and Europe, increased incidence of type 2
diabetes was observed with long-term ozone exposure. In a study conducted in China, long-term ozone
exposure was associated with the development and diagnosis of metabolic syndrome. Several long-term
ozone exposure studies in China, one in adults and one in children, observed increased odds of obesity (a
IS-8
-------
risk factor for type 2 diabetes) in both adults and children. Positive associations between long-term
exposure to ozone and diabetes-related mortality were observed in well-established cohorts in the U.S.
and Canada. The results of mortality studies are supported by epidemiologic and experimental studies
reporting effects on glucose homeostasis and serum lipids, as well as other indicators of metabolic
function (e.g., peripheral inflammation and neuroendocrine activation). This evidence is "suggestive of,
but not sufficient to infer, a causal relationship" between long-term ozone exposure and metabolic effects.
Notably, compared with the 2013 Ozone ISA, there are changes in the causality determinations
for short-term ozone exposure and cardiovascular effects and total mortality. In both instances, the 2013
Ozone ISA concluded that the evidence was sufficient to conclude a "likely to be causal relationship," but
after integrating the previous evidence with recent evidence, the collective evidence is "suggestive of, but
not sufficient to infer, a causal relationship" in this ISA. The evidence that supports this change in the
causality determination includes (1) a growing body of controlled human exposure studies providing less
consistent evidence for an effect of short-term ozone exposure on cardiovascular health endpoints; (2) a
paucity of positive evidence from epidemiologic studies for more severe cardiovascular morbidity
endpoints (i.e., heart failure [HF], ischemic heart disease [IHD] and myocardial infarction [MI],
arrhythmia and cardiac arrest, and stroke); and (3) uncertainties due to few studies evaluating the potential
for confounding by copollutants in epidemiologic studies. Although there is consistent or generally
consistent evidence for several ozone-induced cardiovascular endpoints in animal toxicological studies
and for cardiovascular mortality in epidemiologic studies, these results are not coherent with those in
controlled human exposure and epidemiologic studies examining cardiovascular morbidity endpoints.
IS.1.3.2 Welfare: Ecological Effects
The 2013 Ozone ISA (U.S. EPA. 2013b) concluded that the responses to ozone exposure occur
across multiple biological scales and a broad array of ecological endpoints, with the strongest evidence
for effects on vegetation. The focus of the current ISA and literature evaluated herein are those effects
observed at the individual-organism level of biological organization and higher (e.g., population,
community, ecosystem). New research largely strengthens the previous conclusions on the ecological
effects of ozone. The types of ecological effects studies conducted since the 2013 Ozone ISA mostly fall
into three categories: (1) empirical research that has refined/reinforced earlier studies, in some cases using
new approaches, new species, or larger-scale systems; (2) meta-analyses that have provided a more
statistically based understanding of patterns compiled from existing literature; and (3) modeling
approaches that have increased in complexity and enabled examination of ozone effects at larger spatial
scales (e.g., regional, national). There are 12 causality determinations for ecological effects of ozone
(Table IS-2). generally organized from the individual-organism scale to the ecosystem scale. Similar to
the findings of the 2013 Ozone ISA, five are causal relationships (i.e., visible foliar injury, reduced
vegetation growth, reduced crop yield, reduced productivity, and altered belowground biogeochemical
cycles) and two are likely to be causal relationships (i.e., reduced carbon sequestration, altered ecosystem
IS-9
-------
water cycling). One endpoint, alteration of terrestrial community composition, is now concluded to be a
causal relationship, whereas this endpoint was classified as likely to be causal in the 2013 Ozone ISA.
Three new endpoint categories (i.e., increased tree mortality, alteration of herbivore growth and
reproduction, and alteration of plant-insect signaling) not evaluated for causality in the 2013 Ozone ISA
all have a "likely to be causal relationship." Plant reproduction, previously considered as part of the
evidence for growth effects, is now a stand-alone causal relationship.
IS.1.3.3 Welfare: Effects on Climate
Recent evidence continues to support a causal relationship between tropospheric ozone and
radiative forcing and a likely to be causal relationship, via radiative forcing, between tropospheric ozone
and temperature, precipitation, and related climate variables (referred to as "climate change" in the 2013
Ozone ISA; the revised title for this causality determination provides a more accurate reflection of the
available evidence ITablc IS-31). The new evidence comes from the Intergovernmental Panel on Climate
Change (IPCC) Fifth Assessment Report [AR5; Myhre et al. (2013)1 and its supporting references—in
addition to a few more recent studies—and builds on evidence presented in the 2013 Ozone ISA. The new
studies further support the causality determinations included in the 2013 Ozone ISA.
IS-10
-------
Table IS-2 Summary of causality determinations for ecological effects.
Endpoint
Conclusions from 2013 Ozone
ISA
Conclusions in the 2020 ISA
Visible foliar injury
Causal relationship
Causal relationship
Reduced vegetation growth
Causal relationship
Causal relationship
Reduced plant reproduction
No separate causality
determination; included with plant
growth
Causal relationship3
Increased tree mortality
Causality not assessed
Likely to be causal relationship3
Reduced yield and quality of agricultural
crops
Causal relationship
Causal relationship
Alteration of herbivore growth and
reproduction
Causality not assessed
Likely to be causal relationship3
Alteration of plant-insect signaling
Causality not assessed
Likely to be causal relationship3
Reduced productivity in terrestrial
ecosystems
Causal relationship
Causal relationship
Reduced carbon sequestration in terrestrial
ecosystems
Likely to be causal relationship
Likely to be causal relationship
Alteration of belowground biogeochemical
cycles
Causal relationship
Causal relationship
Alteration of terrestrial community
composition
Likely to be causal relationship
Causal relationship15
Alteration of ecosystem water cycling
Likely to be causal relationship
Likely to be causal relationship
aDenotes new causality determination.
bDenotes change in causality determination from 2013 Ozone ISA.
IS-11
-------
Table IS-3 Summary of causality determinations for tropospheric ozone effects
on climate.
Conclusions in 2013 Ozone ISA
Conclusions in the 2020 ISA
Radiative forcing
Causal relationship
Causal relationship
Temperature, precipitation, and related
climate variables
Likely to be causal relationship
Likely to be causal relationship
IS.2 Atmospheric Chemistry, Ambient Air Ozone Concentrations,
and Background Ozone
Scientific advances in atmospheric ozone research relevant to the Ozone NAAQS are reviewed in
this section, with a primary focus on understanding the relative contribution of precursor emissions to
ambient ozone concentrations from natural processes and anthropogenic activities. The section
summarizes recent developments in measurement and modeling methods, atmospheric chemistry, and
ambient air concentration trends (Section IS.2. IV The U.S. background (USB) ozone concentration is
defined as the ozone concentration that would occur if all U.S. anthropogenic ozone precursor emissions
were removed, as described in Section IS.2.2. This definition facilitates separate consideration of ozone
that results from anthropogenic precursor emissions within the U.S. and ozone originating from natural
and foreign precursor sources. This discussion is followed by a summary of recent observations and
research related to USB ozone, with an emphasis on major sources (Section IS.2.2.1). estimation methods
(Section IS.2.2.2). and geographic, seasonal, and long-term ozone concentration trends (Section IS.2.2.3V
IS.2.1 Ambient Air Ozone Anthropogenic Sources, Measurement, and
Concentrations
The general photochemistry of tropospheric ozone is described in detail in previous assessments
(U.S. EPA. 2013b. 2006a). Anthropogenic ozone in urban settings is produced primarily by the reaction
of volatile organic compounds (VOCs) with oxides of nitrogen (NOx) in the presence of sunlight. Carbon
monoxide (CO) and methane (CH4) also react with NOx to form ozone in the absence of more reactive
organic compounds (Section 1.4). The most abundant national and global sources of VOCs are biogenic
(U.S. EPA. 2013bV and oxides of nitrogen are predominately emitted from a range of anthropogenic
sources, including automobile exhaust, off-road vehicles and engines, electric power generation,
industrial activities, and stationary fuel combustion (U.S. EPA. 2016). Recent developments in
understanding ozone chemistry include observations of high ozone concentrations during the winter in
some western U.S. mountain basins (Section 1.4.1V For example, wintertime ozone concentrations in the
IS-12
-------
Uintah Basin of Utah and Upper Green River Basin of Wyoming have been measured as high as 150 ppb
(1-hour avg), with episodes driven by local concentrations of ozone precursor emissions from oil and gas
extraction coinciding with strong mountain valley temperature inversions on cold winter days with snow
cover. In addition, there is new research on the role of marine halogen chemistry in suppressing coastal
ozone concentrations (Section 1.4.2). Incorporating marine halogen chemistry into atmospheric modeling
methods for predicting ozone concentrations has improved agreement between model results and
observed ozone near marine environments.
Extensive air monitoring data are obtained from the state and local air monitoring site (SLAMS)
network for ozone, consisting of more than 1,300 monitors throughout the U.S. (Section 1.7). In the
SLAMS network, ozone is measured by ultraviolet spectroscopy using a Federal Equivalency Method
(FEM) at most sites (Section 1.6.1.1). A new Federal Reference Method (FRM) for ozone measurement
was adopted in 2015 (Section 1.6.1.1) based on chemiluminescence resulting from the reaction of ozone
with nitric oxide, and is used at some sites. In addition to network monitoring, satellite-based remote
sensing methods are increasingly used to measure the total ozone column in the atmosphere, and satellite
data are used to constrain model estimates of ground-level tropospheric ozone concentrations
(Section 1.6.1.2). Because tropospheric concentration estimates based on satellite measurements can have
much greater uncertainty than total column ozone measurements, these technologies are most suitable for
investigating trends in total column ozone or in the upper troposphere. The 2013 Ozone ISA provided an
overview of chemical transport models (CTMs), including the relevant processes, numerical approaches,
relevant spatial scales, and methods for evaluation (U.S. EPA. 2013b). Since the 2013 Ozone ISA,
numerous improvements to these models have been made. These include: the addition of a halogen
chemistry mechanism; improvements in the representation of land cover and near surface meteorology;
the inclusion of dry deposition and stomatal uptake, stratosphere-troposphere exchange, and biogenic
emissions; and, the integration of meteorological models and CTMs (Section 1.6.2).
SLAMS network data for the period 2015-2017 show higher nationwide median "max daily
8-hour avg" (MDA8) ozone concentrations across all monitoring sites in spring (median = 46 ppb) and
summer (median = 46 ppb) than in autumn (median = 38 ppb) and winter (median = 34 ppb). The highest
values of annual 4th-highest MDA8 ozone concentration (>75 ppb) occur in central and southern
California, Arizona, Colorado, Utah, Texas, along the shore of Lake Michigan, and in the Northeast
Corridor, typically during the warm season between May and September (Section 1.2.1.1). These results
are similar to those reported in the 2013 Ozone ISA (U.S. EPA. 2013b). The highest values of W126, an
example of a cumulative index of plant exposure (Section IS. 3.2 and Section 1.2.1.2). occurred in
California and the southwestern U.S.
Several recent studies have documented a long-term decreasing trend in nationwide average
ambient air MDA8 ozone concentration over several decades and a faster decline in the magnitude and
frequency of high MDA8 ozone episodes (Section 1.7). Comparison of the difference between 5th and
95th percentile concentrations indicates a compression of the MDA8 ozone concentration distribution
IS-13
-------
occurring widely across the U.S. This compression results from a decrease in 95th percentile
concentrations together with a general increase in 5th percentile concentrations. This is consistent with
observed reductions in NOx emissions (Section 1.3.1). because there is less NO available to react with
ozone at low ozone concentrations, as well as less NO2 available to form ozone at high ozone
concentrations.
IS.2.2 Background Ozone
Use of the term "background ozone" varies within the air pollution research community. It has
generally been used to describe ozone levels that would exist in the absence of anthropogenic emissions
within a particular area and has been broadly applied to every geospatial scale: local, regional, national,
continental, or global. For instance, on a local scale, ozone that originates from precursor emissions
outside of a locality's municipal boundaries could be considered background ozone for that locality.
Similarly, on a national scale, background ozone could be defined as ozone that is not formed from
anthropogenic emissions within national boundaries.
The USB concentration is defined as the ozone concentration that would occur if all U.S.
anthropogenic ozone precursor emissions were removed. It is a hypothetical construct that cannot be
measured. The 2006 Ozone AQCD (U.S. EPA. 2006a) and 2013 Ozone ISA (U.S. EPA. 2013b)
concluded that background ozone concentrations could not be determined solely from ozone
measurements, even at the most remote monitoring sites, because of long-range transport of ozone
originating from U.S. anthropogenic precursors. Since then, chemical transport models have been used as
the primary tool for estimating USB ozone concentrations.
IS.2.2.1 Sources of U.S. Background Ozone
Major contributors to ground-level USB ozone concentrations are stratospheric exchange,
international transport, wildfires, lightning, global methane emissions, and natural biogenic and geogenic
precursor emissions. As the USB literature has evolved, much of the discussion has focused on the
relative importance of stratospheric ozone and intercontinental transport as major sources.
Tropospheric ozone derived from stratosphere-troposphere dynamics was described in detail in
the 2013 Ozone ISA (U.S. EPA. 2013b). Stratospheric air naturally rich in ozone can be transported into
the troposphere under certain meteorological circumstances, with maximum contributions observed at
midlatitudes during the late winter and early spring. This process, known as "tropopause folding," is
characterized by episodic events typically lasting a few days from late winter through spring when deep
stratospheric intrusions rich in ozone can quickly and directly well into the troposphere and, more rarely,
reach ground level (U.S. EPA. 2013b). The 2013 Ozone ISA (U.S. EPA. 2013b) also discussed the
potential importance of deep convection, another form of stratosphere-troposphere exchange that occurs
IS-14
-------
mainly in summer, as a mechanism for transporting stratospheric ozone into the upper troposphere.
Stratospheric ozone contributions from deep intrusion between 17 and 40 ppb have been estimated at
ground level for springtime model simulations in the western U.S. (Section 1.3.2.1). Stratospheric
intrusion events related to frontal passage and tropopause folding that reach the surface have less
influence on surface ozone during the summer months when total ground-level ozone concentrations tend
to be highest.
Intercontinental transport from Asia has also been identified as a major source of precursors that
can contribute 5 to 7 ppb to USB ozone concentrations over the western U.S. (U.S. EPA. 2013b. 2006a.
b). Ozone precursor emissions from China and other Asian countries have been estimated to have more
than doubled in the period 1990-2010 (Section 1.3.1.2). and an estimated increase of 0.3 to 0.5 ppb/year
of midtropospheric ozone USB in spring over the western U.S. in the two decades after 1990 was largely
attributed to a tripling of Asian NOx emissions (Section 1.3.1). However, after this period, trends in NOx
emissions from China, the largest ozone precursor source in Asia, have declined as confirmed by rapidly
decreasing satellite-derived tropospheric NO2 column measurements over China since 2012. Stringent air
quality standards implemented in 2013 within China have markedly reduced national emissions of NOx
(Section 1.3.1.2).
Other contributors to USB are either smaller or more uncertain than stratospheric and
intercontinental contributions. Wildfires have been estimated to contribute a few ppb to seasonal mean
ozone concentrations in the U.S., but episodic contributions may be as high as 30 ppb (Section 1.3.1.2).
However, estimates of the magnitude of ozone formation from wildfires is highly uncertain with some
work showing large overpredictions of modeled wildfire contributions (Section 1.3.1.3). Lightning was
estimated to contribute 2 to 3 ppb to ground level ozone concentrations in the southeastern U.S. in the
summer (U.S. EPA. 2013b). Eighty percent of the NOx present in the upper troposphere is generated by
lightning where it can have a longer atmospheric residence time than NOx derived from ground sources
(Section 1.3.1.3). There is an approximately linear relationship between anthropogenic methane emissions
and tropospheric ozone, which is consistent with the contribution of anthropogenic methane emissions to
global annual mean ozone concentration of -4-5 ppb reported in the 2013 Ozone ISA (U.S. EPA. 2013b).
Biogenic emissions of NOx are estimated to contribute 0.3 Tg N/year, or about 7.5% of total NOx
emissions (Section 1.3.1.3).
IS.2.2.2 Methods for Estimating U.S. Background Ozone
Large uncertainties are associated with estimating USB ozone concentrations. Approaches for
estimating USB ozone are described in Section 1.8.1. USB ozone is estimated using either zero-out
simulations or source apportionment simulations. The most widely used approach to measuring USB or
other measures of background ozone is the zero-out method, in which anthropogenic U.S. or other areas
emissions are set to zero in a model simulation to estimate these ozone measures (Section 1.8.1.1). As an
IS-15
-------
alternative to model sensitivity approaches, source apportionment techniques track source contributions to
ozone formation without perturbing emissions (Section 1.8.1.2V Tracking techniques use reactive tracer
species to tag specific emissions source categories or source regions and then track the ozone produced by
emissions from those source groups. Both approaches are essential and complementary for understanding
and estimating USB ozone. The zero out approach is suited for estimating what ozone levels would have
existed in recent modeled years in the absence of all U.S. emissions, while the source apportionment
approach is suited for estimating the fraction of current ozone originating from background sources in
recent modeled years. The difference between estimates from these approaches is small in remote areas
that are most strongly affected by USB sources. However, the differences in the estimates given by these
methods can be substantial in urban areas strongly affected by anthropogenic sources that influence both
production and destruction of ozone.
USB ozone concentrations vary daily and by location and are a function of season, meteorology,
and elevation. Quantification of USB ozone on days when MDA8 ozone concentrations exceed 70 ppb is
more relevant to understanding USB ozone contributions on those days than are seasonal mean USB
ozone estimates, but also more uncertain (Jaffc et al.. 2018). Jaffe et al. (2018) reviewed recent modeling
results and reported that USB ozone estimates contain uncertainties of about 10 ppb for seasonal average
concentrations and 15 ppb for MDA8 avg concentrations on individual days. Because of uncertainty in
model predictions of USB, model results are often adjusted using simple bias correction approaches.
Because such approaches might not be reliable if the model has diverging errors in USB ozone and
locally produced ozone, however, days with poor model performance have sometimes been excluded
when using model results to estimate USB or other measures of background ozone. There have been
continued efforts to improve model performance and better understand biases and uncertainties involved
in the application of CTMs to estimating USB or other measures of background ozone (Section 1.8.1.5).
IS.2.2.3 U.S. Background Concentrations and Trends
A greater variety of approaches for estimating USB concentrations and other measures of
background ozone used in recent years have led to a wider range of USB estimates than reported in the
2013 Ozone ISA (U.S. EPA. 2013b). although some of the basic patterns remain consistent. For example,
higher USB concentrations (and related measures of background ozone) were estimated in the western
U.S. than in the eastern U.S. in the 2013 Ozone ISA (U.S. EPA. 2013b). especially in the intermountain
West and Southwest. Higher USB concentrations were also estimated at elevations higher than 1,500 m
than at lower elevations (U.S. EPA. 2013b). New studies since the 2013 Ozone ISA confirm these
findings (Section 1.8.2.1).
USB concentrations are relatively constant with increasing total ozone concentration, indicating
that days with higher ozone concentrations generally occur because of higher U.S. anthropogenic
contributions (Section 1.8.2.3). In the eastern U.S. and in urban and low-elevation areas of the western
IS-16
-------
U.S., there is consistent evidence across several studies that daily USB ozone concentrations are similar to
or smaller than seasonal mean USB ozone concentrations on most high ozone concentration days
(i.e., days with MDA8 ozone greater than 60 ppb). In contrast, for high elevation locations in the western
U.S., USB concentration estimates have been consistently predicted to increase with total ozone
concentration, consistent with a larger background contribution. Lower USB contributions on days of
high ozone concentration can result from meteorological conditions that favor large ozone production
from U.S. anthropogenic sources relative to USB sources (Section 1.5.2). The highest ozone
concentrations observed in the U.S. have historically occurred during stagnant conditions when an air
mass remains stationary over a region abundant in anthropogenic ozone precursor sources (U.S. EPA.
2013b. 2006a. 1996a). while the largest USB contributions often occur under the opposite conditions,
when the atmosphere is well mixed and transport of USB ozone generated in the stratosphere or during
long-range transport of Asian or natural precursors in the upper troposphere more readily occurs
(Section 1.5.2).
Characterizing long-term trends in USB presents numerous challenges (Section 1.8.2.4). Research
has mainly focused on high elevation sites in the western U.S. or measurements made aloft, where, until
recently, increasing midtropospheric ozone was reported. The most recent analyses suggest that this trend
has now slowed or reversed, and there is no evidence to suggest that USB is still increasing, even in the
western U.S. (Section 1.8.2.4).
IS.3 Exposure to Ambient Ozone
IS.3.1 Human Exposure Assessment in Epidemiologic Studies
With regard to exposure assessment relevant to human health effects, the 2013 Ozone ISA (U.S.
EPA. 2013b) primarily discussed personal exposure to ozone and its relationship to ambient air
concentrations.
Its primary conclusions were that personal exposure to ozone is moderately correlated with
ambient air concentration (Pearson R = 0.3-0.8) and indoor ozone concentrations were roughly 10-30%
of ambient air concentrations. In addition, ozone exposure minimization efforts through public messaging
(e.g., ozone action days) were effective in reducing exposures for people younger than 20 years old but
did not make an appreciable difference in exposure among those ages 20-64 years old. The 2013 Ozone
ISA noted that urban scale ozone concentrations often have low spatial variability except in the vicinity of
roadways, where nitrogen oxides emitted from motor vehicles tend to scavenge ozone.
The 2013 Ozone ISA also found that exposure measurement error can bias epidemiologic
associations between ambient ozone concentrations and health outcomes and widen confidence intervals
around effect estimates. Recently published studies agree with these previous findings. Although ozone
IS-17
-------
concentrations measured at fixed-site ambient air monitors are still widely used as surrogates for ozone
exposure in epidemiologic studies (Section 2.3.LP. the availability and sophistication of models to
predict ambient ozone concentrations for this purpose have increased substantially in recent years
(Section 2.3.2). The greatest expansion in modeling capability has occurred in chemical transport
modeling (CTM; Section 2.3.2.3). especially when incorporated into a hybrid spatiotemporal framework
that integrates modeling output with monitoring and satellite data over time and space (Section 2.3.2.4V
Hybrid methods have produced lower error predictions of ozone concentration compared with
spatiotemporal models using land use and other geospatial data alone (Section 2.3.2.2) but may be subject
to overfitting given the many different sources of data incorporated into the hybrid framework.
Use of an exposure surrogate in epidemiologic studies generally leads to underestimation of any
association between short-term exposure to ozone and a health effect, with reduced precision. Although
the magnitude of an association between ambient ozone and a health effect is uncertain, the evidence
indicates that the true effect is typically larger than the effect estimate in these cases. Epidemiologic
studies evaluating short-term ozone exposure examine how short-term (e.g., hourly, daily, weekly)
changes in health effects are associated with short-term changes in exposure (Section 2.6.1). Accurate
characterization of temporal variability is more important than accurate characterization of spatial
variability for these studies. Use of an exposure surrogate may produce bias when temporal variability in
the concentration at the location of the measurement or model prediction differs from temporal variability
of the true exposure concentration. As a result, the correlation between the exposure surrogate and the
incidence of the effect would decrease due to the additional scatter in that relationship, and the reduced
correlation would also likely flatten the slope of the relationship between the effect and exposure
surrogate.
For effects elicited by ozone, the use of exposure estimates that do not account for population
behavior and mobility (e.g., via use of time-activity data) may underestimate the true effect and have
reduced precision. Although the magnitude of association between ozone and such health effects are
uncertain, the evidence suggests that the true effect of ambient ozone exposure is larger than the effect
estimate when time-activity data are not considered in the analysis. Uncharacterized exposure variability
due to omission of time-activity data for short-term studies (Section 2.4.1) creates uncertainty in the
exposure estimate that could reduce the correlation between the exposure estimate and the health effect.
Depending on the exposure model and scenario being modeled for application in epidemiology
studies, the true effect of long-term exposure to ambient ozone may be underestimated or overestimated
when the exposure model respectively overestimates or underestimates ozone exposure. It is much more
common for the effect to be underestimated, and the bias is typically small in magnitude. Long-term
epidemiologic studies examine the association between the health effect endpoint and long-term average
ambient ozone exposure (Section 2.6.2). For cohort studies of long-term ambient ozone exposure,
ambient ozone concentration measured at monitors or estimated by a model is often used as a surrogate
for ambient ozone exposure. These studies typically examine differences among cohorts in different
IS-18
-------
locations, at the scale of neighborhoods, cities, or states. Uncharacterized spatial variability in ozone
exposure across the study area could lead to bias in the effect estimate if modeled or measured ambient
concentration is not representative of ambient exposure. Bias can occur in either direction but more often
has been reported to be towards the null in exposure measurement error studies. Uncertainties in time
activity and residential patterns of exposed individuals and surface losses of ozone can reduce precision in
the effect estimates.
IS.3.2 Ecological Exposure
The key conclusions from the 1996 and 2006 Ozone AQCDs, and the 2013 Ozone ISA regarding
ozone exposure to vegetation, highlighted below, are still valid and most effects observed for
nonvegetation biota are mediated through ozone effects on vegetation. Absorption of ozone from the
atmosphere into leaves is controlled by the leaf boundary layer and stomatal conductance. Stomata
provide the principal pathway for ozone to enter and affect plants, with subsequent oxidative injury to leaf
tissue triggering a cascade of physical, biogeochemical, and physiological events that may scale up to
responses at the whole-plant scale.
As described in previous ozone assessments, ozone-related injury is a function of flux (i.e., the
amount of ozone taken up by the plant overtime). Ozone flux is affected by modifying factors such as
temperature, vapor pressure deficit, light, soil moisture, and plant growth stage (U.S. EPA. 2013b). Flux
is very difficult to measure directly, requiring quantification of stomatal or canopy conductance. While
some efforts have been made in the U.S. to calculate ozone flux into leaves and canopies, little
information has been published relating these fluxes to effects on vegetation. The scarcity of flux data in
the U.S. and lack of understanding of plant detoxification processes have made this technique less viable
for risk assessments in the U.S. (U.S. EPA. 2013b). An alternative to flux-based exposure estimates are
exposure indices. Exposure indices quantify exposure as it relates to measured plant response
(e.g., growth) and only require ambient air quality data rather than more complex indirect calculations of
dose to the plant. Cumulative indices summarize ozone concentrations over time to provide a consistent
metric for reviewing and comparing exposure-response effects obtained from various studies. For
ecological studies in this ISA, emphasis is placed on studies that characterize exposures at concentrations
occurring in the environment or experimental ozone concentrations within an order of magnitude of
recent concentrations observed in the U.S. (Appendix 1).
It is well established that exposure duration influences the degree of plant response and that
ozone effects on plants are cumulative. In previous ozone assessments, effects are clearly demonstrated to
be related to the cumulative exposure over the growing season for crops and herbaceous plant species. For
long-lived plants, such as trees, exposures occur over multiple seasons and years. Cumulative indices of
exposure are, therefore, best suited to assess exposure. Since the 1980s, cumulative-type indices such as
threshold weighted (e.g., SUM06, AOTx) and continuous weighted (e.g., W126) functions have been
IS-19
-------
applied to evaluate ozone exposure in plants (U.S. EPA. 2013b). The 2013 Ozone ISA primarily
discussed SUM06, AOTx, and W126 exposure metrics. Below are the definitions of the three cumulative
index forms:
• SUM06: Sum of all hourly ozone concentrations greater than or equal to 60 ppb observed during
a specified daily and seasonal time window (U.S. EPA. 2013b).
• AOTx: Sum of the differences between hourly ozone concentrations greater than a specified
threshold during a specified daily and seasonal time window. For example, AOT40 is the sum of
the differences between hourly concentrations above 40 ppb during a specified period (U.S. EPA.
2013b).
• W126: Sigmoidally weighted sum of all hourly ozone concentrations observed during a specified
daily and seasonal time window (Lefohn et al.. 1988; Lefohn and Runeckles. 1987).
IS.4 Evaluation of the Health Effects of Ozone
IS.4.1 Connections among Health Effects
Broad health effect categories are evaluated separately in the Appendices of this ISA, though the
mechanisms underlying disease progression may overlap and not be restricted to a single organ system.
This section provides a brief overview of how the relationship between ozone and a variety of health
outcomes may be related or affect one another.
Ozone-induced injuries can take place via complex pathways within the body. After inhalation,
ozone reacts with lipids, proteins, and antioxidants in the respiratory tract epithelial lining fluid to create
secondary oxidation products RJ.S. EPA (2013b): Section 5.2.31. The first steps (i.e., initial events) in the
cascade of physiological events includes activation of sensory nerves in the respiratory tract and
respiratory tract inflammation. These early physiological reactions to ozone may trigger a host of
autonomic, endocrine, immune, and inflammatory responses throughout the body at the cellular, tissue,
and organ level. Because the circulatory system is connected to all body systems, insults to multiple organ
systems may contribute to a single health effect. The 2006 Ozone AQCD RJ.S. EPA (2006a): Chapter 4]
and the 2013 Ozone ISA rU.S. EPA (2013b): Section 5.31 provide extensive background on dosimetry
and potential pathways underlying health effects for these responses.
Modulations of the autonomic nervous system, which consists of the sympathetic and
parasympathetic systems, provide inhibitory or excitatory inputs to tissues to generate organ responses.
Some examples of responses from alterations of the autonomic nervous system include changes to heart
rate, bronchodilation/bronchoconstriction, altered blood glucose, glycogenolysis/gluconeogenesis,
hormone release, and other organ functions (McCorrv. 2007). Endocrine, immune, and inflammatory
responses can send signals capable of altering multiple pathways and eliciting cardiovascular, respiratory,
and metabolic health effects.
IS-20
-------
While all systems of the body are connected intrinsically, most research presented in the field of
air quality examines specific health endpoints resulting from exposure to a pollutant. In an effort to bring
together the scientific body of evidence in an easily understandable and relatable way, this document has
separated the supporting Appendices into Respiratory (Appendix 3). Cardiovascular (Appendix 4).
Metabolic (Appendix 5). Mortality (Appendix 6). and Other Health Effects (Appendix 7).
IS.4.2 Biological Plausibility
New to this Ozone ISA are biological plausibility sections for the broad health outcome
categories that are included in the human health Appendices (Appendix 3-Appendix 7). These sections
outline potential pathways along the exposure to outcome continuum and provide plausible links between
inhalation of ozone and health outcomes at the population level. Biological plausibility can strengthen the
basis for causal inference (U.S. EPA. 2015). In this ISA, biological plausibility is part of the weight-of-
evidence analysis that considers the totality of the health effects evidence, including consistency and
coherence of effects described in experimental and observational studies. Although there is some overlap
in the potential pathways between the Appendices, each biological plausibility section is tailored to the
specific broad health outcome category and exposure duration for which causality determinations are
made.
Each of the biological plausibility sections includes a figure depicting potential biological
pathways that is accompanied by text. The figures illustrate possible pathways related to ozone exposure
that are based on evidence evaluated in previous assessments, both AQCDs and ISAs, as well as evidence
from more recent studies. The text characterizes the evidence upon which the figures are based, including
results of studies demonstrating specific effects related to ozone exposure and considerations of
physiology and pathophysiology. Together, the figure and text portray the available evidence that
supports the biological plausibility of ozone exposure leading to specific health outcomes. Gaps in the
evidence base (e.g., health endpoints for which studies have not been conducted) are represented by
corresponding gaps in the figures and are identified in the accompanying text.
In the model figure below (Figure IS-1). each box represents evidence that has been demonstrated
in a study or group of studies for a particular effect related to ozone exposure. While most of the studies
used to develop the figures are experimental studies (i.e., animal toxicological and controlled human
exposure studies), some observational epidemiologic studies also contribute to the pathways. These
epidemiologic studies are generally (1) panel studies that measure the same or similar effects as the
experimental studies (and thus provide supportive evidence) or (2) emergency department and hospital
admission studies or studies of mortality, which are effects observed at the population level. The boxes
are arranged horizontally, with boxes on the left side representing initial effects that reflect early
biological responses and boxes to the right representing intermediate (i.e., subclinical or clinical) effects
IS-21
-------
and effects at the population level. The boxes are color-coded according to their position in the exposure
to outcome continuum.
Key Clinical Effect
Effect at Population
Level
Intermediate
Effect 3
Note: The boxes above represent the effects for which there is experimental or epidemiologic evidence related to ozone exposure,
and the arrows indicate a proposed relationship between those effects. Solid arrows denote evidence of essentiality as provided, for
example, by an inhibitor of the pathway or a genetic knockout model used in an experimental study involving ozone exposure.
Shading around multiple boxes is used to denote a grouping of these effects. Arrows may connect individual boxes, groupings of
boxes, and individual boxes within groupings of boxes. Progression of effects is generally depicted from left to right and color coded
(white, exposure; green, initial effect; blue, intermediate effect; orange, effect at the population level or a key clinical effect). Here,
population-level effects generally reflect results of epidemiologic studies. When there are gaps in the evidence base, there are
complementary gaps in the figure and the accompanying text below.
Figure IS-1 Illustrative figure for potential biological pathways for health
effects following ozone exposure.
The arrows that connect the boxes indicate a progression of effects resulting from ozone
exposure. In most cases, arrows are dotted (arrow 1), denoting a possible relationship between the effects.
While most arrows point from left to right, some arrows point from right to left, reflecting progression of
effects in the opposite direction or a feedback loop (arrow 2). In a few cases, the arrows are solid
(arrow 2), indicating that progression from the upstream to downstream effect occurs as a direct result of
ozone exposure. This relationship between the boxes, where the upstream effect is necessary for
progression to the downstream effect, is termed essentiality (OECD. 2016). Evidence supporting
essentiality is generally provided by experimental studies using pharmacologic agents (i.e., inhibitors) or
animal models that are genetic knockouts. The use of solid lines, as opposed to dotted lines, reflects the
IS-22
-------
availability of specific experimental evidence that ozone exposure results in an upstream effect which is
necessary for progression to a downstream effect.
In the figures, upstream effects are sometimes linked to multiple downstream effects. In order to
illustrate this proposed relationship using a minimum number of arrows, downstream boxes are grouped
together within a larger shaded box and a single arrow (arrow 3) connects the upstream single box to the
outside of the downstream shaded box containing the multiple boxes. Multiple upstream effects may
similarly be linked to a single downstream effect using an arrow (arrow 4) that connects the outside of a
shaded box, which contains multiple boxes, to an individual box. In addition, arrows sometimes connect
one individual box to another individual box that is contained within a larger shaded box (arrow 2) or two
individual boxes both contained within larger shaded boxes (arrow 5). Thus, arrows may connect
individual boxes, groupings of boxes, and individual boxes within groupings of boxes depending on the
proposed relationships between effects represented by the boxes.
IS.4.3 Summary of Health Effects Evidence
This ISA evaluates the relationships between an array of health effects and short- and long-term
exposure to ozone in epidemiologic, controlled human exposure, and animal toxicological studies.
Short-term exposures are defined as those with durations of hours up to 1 month, with most studies
examining effects related to exposures in the range of several hours to 1 week. Long-term exposures are
defined as those with durations of more than 1 month, with many studies spanning a period of years. As
detailed in the Preface, the evaluation of the health effects evidence from animal toxicological studies
focuses on exposures conducted at concentrations of ozone that are relevant to the range of human
exposures associated with ambient air (up to 2 ppm, which is one to two orders of magnitude above recent
ambient air concentrations in the U.S.). Drawing from evidence related to the discussion of biological
plausibility of ozone-related health effects and the broader health effects evidence spanning scientific
disciplines described in detail in Appendix 3-Appendix 7. as well as issues regarding exposure
assessment and potential confounding described in Appendix 2. the subsequent sections characterize the
evidence that forms the basis of the causality determinations for health effect categories of a "causal
relationship" or a "likely to be causal relationship," or describe instances where a causality determination
has been changed (i.e., "likely to be causal" changed to "suggestive of, but not sufficient to infer a causal
relationship"). The evidence that supports these causality determinations builds upon the potential
biological pathways, which provide evidence of biological plausibility, as well as the broader health
effects evidence spanning scientific disciplines for each health effects category, as well as issues related
to dosimetry, exposure assessment, and potential confounding. Other relationships between ozone and
health effects where the causality determinations are "suggestive of but not sufficient to infer a causal
relationship " or "inadequate to infer the presence or absence of a causal relationship" are noted in
Table IS-1. and more fully discussed in the respective health effects Appendices.
IS-23
-------
IS.4.3,1 Short-Term Exposure and Respiratory Health Effects
The 2013 Ozone ISA concluded that there is a "causal relationship" between short-term ozone
exposure and respiratory health effects (U.S. EPA. 2013b). This conclusion was based largely on
controlled human exposure studies demonstrating ozone-related respiratory effects in healthy individuals
(Table IS-4). Specifically, statistically significant decreases in group mean pulmonary function in
response to 6.6-hour ozone exposures (which included six 50-minutes periods of moderate exertion) to
concentrations as low as 60 ppb1 were observed in young, healthy adults (Figure IS-2). Additionally,
controlled human exposure and experimental animal studies demonstrated ozone-induced increases in
respiratory symptoms, lung inflammation, airway permeability, and airway responsiveness. The
experimental evidence was supported by strong evidence from epidemiologic studies demonstrating
associations between ambient ozone concentrations and respiratory hospital admissions and ED visits
across the U.S., Europe, and Canada. This evidence was further supported by a large body of
individual-level epidemiologic panel studies that demonstrated associations of short-term ozone
concentrations with respiratory symptoms in children with asthma. Additional support for a causal
relationship was provided by epidemiologic studies that observed ozone-associated increases in indicators
of airway inflammation and oxidative stress in children with asthma. Additionally, several multicity
studies and a multicontinent study reported associations between short-term increases in ozone
concentrations and increases in respiratory mortality.
Table IS-4 Summary of evidence from epidemiologic, controlled human
exposure, and animal toxicological studies on the respiratory effects
of short-term exposure to ozone.
1 Concentrations from controlled human exposure studies are target concentrations, unadjusted for study-specific
measurement information.
Conclusions from 2013 Ozone ISA Results and Conclusions from 2020 ISAa
Respiratory effects Evidence integrated across controlled
Recent evidence from controlled human
exposure, epidemiologic, and animal
toxicological studies support and extend the
conclusions from the 2013 Ozone ISA that
there is a causal relationship between
short-term ozone exposure and respiratory
effects.
human exposure, epidemiologic, and
animal toxicological studies and across
the spectrum of respiratory health
endpoints demonstrated that there was a
causal relationship between
short-term ozone exposure and
respiratory health effects.
IS-24
-------
Table IS-4 (Continued): Summary of evidence from epidemiologic, controlled
human exposure, and animal toxicological studies on the
respiratory effects of short-term exposure to ozone.
Conclusions from 2013 Ozone ISA
Results and Conclusions from 2020 ISAa
Lung function
Controlled human exposure studies of
young, healthy adults demonstrate group
mean decreases in FEVi in the range of
2 to 3% with 6.6-h exposures, while
exercising, from concentrations as low as
60 ppb ozone. The collective body of
epidemiologic evidence demonstrate
associations between short-term ambient
ozone concentrations and decrements in
lung function, particularly in children with
asthma, children, and adults who work or
exercise outdoors.
Controlled human exposure studies of young,
healthy adults demonstrate ozone-induced
decreases in FEVi at concentrations as low as
60 ppb and the combination of FEVi
decrements and respiratory symptoms at
ozone concentrations 70 ppb or greater
following 6.6-h exposures while exercising.
Studies show interindividual variability with
some individuals being intrinsically more
responsive. Results from recent epidemiologic
studies are consistent with evidence from the
2013 Ozone ISA of an association with lung
function decrements as low as 33 ppb (mean
8-h avg ozone concentrations).
Airway responsiveness
A limited number of studies observe
increased airway responsiveness in
rodents and guinea pigs after being
exposed for 72 h to ozone concentrations
ranging from less than 300 ppb up to
1,000 ppb. As previously reported in the
2006 63 AQCD, increased airway
responsiveness demonstrated at 80 ppb
in young, healthy adults, and at 50 ppb in
certain strains of rats.
Controlled human exposure studies provide
evidence of increased airway responsiveness
with exposures as low as 80 ppb. Baseline
airway responsiveness does not appear
predictive of changes in lung function following
ozone exposure. Recent animal toxicological
studies demonstrate increases in airway
responsiveness following ozone exposures as
low as 800 ppb. A recent animal toxicological
study showed increased airway
responsiveness to a greater degree in allergic
mice than in naive mice at 1,000 ppb for 8 h.
Pulmonary
inflammation, injury,
and oxidative stress
Epidemiologic studies provide evidence
for associations of ambient ozone with
mediators of airway inflammation and
oxidative stress and indicated that higher
antioxidant levels may reduce pulmonary
inflammation associated with ozone
exposure. Generally, these studies had
mean 8-h daily max ozone concentrations
less than 66 ppb. Controlled human
exposure studies show ozone-induced
inflammatory responses at 60 ppb, the
lowest concentration evaluated.
Controlled human exposure studies
demonstrate ozone-induced increases in
pulmonary inflammation at concentrations as
low as 60 ppb after 6.6 h of exposure. Studies
show interindividual variability in inflammatory
responses with some individuals reproducibly
experiencing intrinsically greater responses
than average. Animal toxicological studies
demonstrate inflammation, injury, and
oxidative stress following ozone exposures as
low as 300 ppb for up to 72 h. Epidemiologic
studies observe associations with pulmonary
inflammation in studies of healthy children
(mean 8-h daily max ozone concentrations as
low as 53 ppb).
Respiratory symptoms
and medication use
The collective body of epidemiologic
evidence demonstrate positive
associations between short-term
exposure to ambient ozone and
respiratory symptoms (e.g., cough,
wheeze, and shortness of breath) in
children with asthma. Generally, these
studies had mean 8-h daily max ozone
concentrations less than 69 ppb.
Controlled human exposure studies provide
evidence of increased respiratory symptoms
following 6.6-h exposures to 70 ppb and
greater. Limited data suggests that lung
function responses to ozone in individuals with
asthma may depend on baseline lung function
and medication use. The large body of
epidemiologic evidence from the 2013 Ozone
ISA continues to provide the strongest support
for these outcomes.
IS-25
-------
Table IS-4 (Continued): Summary of evidence from epidemiologic, controlled
human exposure, and animal toxicological studies on the
respiratory effects of short-term exposure to ozone.
Conclusions from 2013 Ozone ISA
Results and Conclusions from 2020 ISAa
Lung host defenses
Controlled human exposure studies
demonstrate the increased expression of
cell surface markers and alterations in
sputum leukocyte markers related to
innate adaptive immunity with short-term
ozone exposures of 80-400 ppb. Animal
toxicological studies demonstrate
increased susceptibility to infectious
disease with short-term ozone exposures
as low as 80 ppb. Altered macrophage
function was reported with exposures as
low as 100 ppb. Other effects on the
immune system (i.e., adaptive immunity
and natural killer cells) are seen with
exposures as low as 500 ppb.
A limited number of recent controlled human
exposure studies report results that are
consistent with studies evaluated in the 2013
Ozone ISA that demonstrated impaired lung
host defense following acute ozone exposure.
A limited number of recent animal toxicological
studies demonstrate susceptibility to infectious
disease at 2,000 ppb ozone for 3 h. Recent
epidemiologic studies of ED visits for
respiratory infection provide the strong support
for these outcomes.
Allergic and Controlled human exposure studies in
asthma-related atopic individuals with asthma
responses demonstrate increased airway
eosinophils, enhanced allergic cytokine
production, increased IgE receptors, and
enhanced markers of innate immunity
and antigen presentation with short-term
exposure to 80-400 ppb ozone, all of
which may enhance allergy and/or
asthma. Increased airway
responsiveness is seen in atopic
individuals with asthma at 120-250 ppb
ozone. In allergic rodents, enhanced
goblet cell metaplasia is seen using
exposure concentrations as low as
100 ppb, and enhanced responses to
allergen challenge is seen with
short-term exposure to 1,000 ppm
ozone.
A limited number of recent controlled human
exposure and animal toxicological studies
demonstrate enhanced type 2 immune
responses following acute ozone exposures as
low as 200 ppb in atopic adults with asthma
and 800 ppb (8 h a day for 3 days) in healthy
rodents. Exacerbated bronchoconstriction
(airway resistance) and lung injury is seen in
allergic rodents at 1,000 ppb. These results
support and expand upon evidence from the
2013 Ozone ISA that ozone enhances allergic
and asthma related responses.
Respiratory hospital
admissions, ED visits
and physician visits
54 ppb. Additional epidemiologic evidence for
associations between ozone and hospital
admissions and ED visits for combinations of
respiratory diseases (31 to 50 ppb as the
study mean 8-h daily max), ED visits for COPD
(33 to 55 ppb as the study mean daily 1-h
max), and ED visits for respiratory infection
(33 to 55 ppb as the study mean daily 1-h
max).
Consistent, positive associations of
ambient ozone concentrations with
respiratory hospital admissions and ED
visits in the U.S., Europe, and Canada
are observed with supporting evidence
from single-city studies. Generally, these
studies had mean 8-h max ozone
concentrations less than 60 ppb.
Evidence from many recent, large multicity
epidemiologic studies provide further support
for an association between ozone and ED
visits and hospital admissions for asthma;
associations are generally strongest in
magnitude for children between the ages of 5
and 18 years in studies with mean 8-h daily
max ozone concentrations between 31 and
IS-26
-------
Table IS-4 (Continued): Summary of evidence from epidemiologic, controlled
human exposure, and animal toxicological studies on the
respiratory effects of short-term exposure to ozone.
Conclusions from 2013 Ozone ISA Results and Conclusions from 2020 ISAa
Respiratory mortality
Multicity time-series studies and a
multicontinent study consistently
demonstrated associations between
ambient ozone concentrations and
respiratory-related mortality across the
U.S., Europe, and Canada with
supporting evidence from single-city
studies. Generally, these studies had
mean 8-h max ozone concentrations less
than 63 ppb.
Recent epidemiologic evidence for respiratory
mortality is limited, but there remains evidence
of consistent, positive associations, specifically
in the summer months, with mean daily
8-h max ozone concentrations between
8.7 and 63 ppb. When recent evidence is
considered in the context of the larger number
of studies evaluated in the 2013 Ozone ISA,
there remains consistent evidence of an
association between short-term ozone
exposure and respiratory mortality.
Conclusions from the 2020 ISA include evidence from recent studies integrated with evidence included in previous Ozone ISAs
and AQCDs.
Evidence from recent controlled human exposure studies augment the evidence from previously
available studies. There are, however, no new 6.6-hour ozone exposure studies since the 2013 Ozone ISA.
Evidence in the 2013 Ozone ISA demonstrated increases in FEVi decrements, respiratory symptoms, and
inflammation following ozone exposures of 6.6 hours, with exercise, as low as 60 to 70 ppb
(Section 3.1.4). Evidence from recent epidemiologic studies of short-term ozone exposure and hospital
admission or emergency department visits observed associations at concentrations as low as 31 ppb.
Controlled human exposure studies also provide consistent evidence of ozone-induced increases in airway
responsiveness (Section 3.1.4.3 and Section 3.1.5.5) and inflammation in the respiratory tract
(Section 3.1.4.4 and Section 3.1.5.6). Recent animal toxicological studies are consistent with evidence
summarized in the 2013 Ozone ISA (U.S. EPA. 2013b); these studies support the evidence observed in
healthy humans.
Evidence from epidemiologic studies of healthy populations is generally coherent with
experimental evidence, with most of the evidence coming from panel studies that were previously
evaluated in the 2013 Ozone ISA (U.S. EPA. 2013b). Several panel studies of healthy children reported
decreases in FEVi and increases in markers of pulmonary inflammation associated with increases in
short-term ozone exposure. While there is coherence between epidemiologic and experimental evidence
of ozone-induced lung function decrements and pulmonary inflammation, respiratory symptoms were not
associated with ozone exposure in a limited number of epidemiologic studies. However, these studies
generally relied on parent-reported outcomes that may have resulted in under- or over-reporting of
respiratory symptoms.
IS-27
-------
~
A
X
~
O
o
~
~ (S)
I T I I I I t
30 40 50 60 70 80 90
Ozone (ppb)
Note: All studies used constant exposure concentrations in a chamber unless designated as stepwise (S) and/or facemask (m)
exposures. All responses at and above 70 ppb (targeted concentration) were statistically significant. Adams (20061 found statistically
significant responses to square-wave chamber exposures at 60 ppb based on the analysis of Brown et al. (20081 and Kim et al.
(20111. During each hour of the exposures, subjects were engaged in moderate quasi-continuous exercise (20 L/minute per m2
BSA) for 50 minutes and rest for 10 minutes. Following the 3rd hour, subjects had an additional 35-minute rest period for lunch. The
data at 60 and 80 ppb have been offset along the x-axis for illustrative purposes. The curved solid line from McDonnell et al. (20131
illustrates the predicted FEI^ decrements using Model 3 coefficients at 6.6 hours as a function of ozone concentration for a
23.8-year-old with a BMI of 23.1 kg/m2.
*80 ppb data for 30 health subjects were collected as part of the Kim et al. (20111 study, but only published in Figure 5 of McDonnell
et al. (20121.
Source: Adapted from Figure 6-1 of 2013 Ozone ISA (U.S. EPA. 2013b1. Studies appearing in the figure legend are: Adams (20061.
Adams (20031. Adams (20021. Horstman et al. (19901. Kim etal. (20111. McDonnell et al. (20131. McDonnell et al. (19911. and
Scheleale et al. (20091.
Figure IS-2 Cross-study comparisons of mean ozone-induced forced
expiratory volume in 1 second (FEVi) decrements in young
healthy adults following 6.6 hours of exposure to ozone.
= !
T3 E
= Si
i O 4
= £
O ^
N -
o >
LU 2
Evidence from numerous recent, large, multicity epidemiologic studies conducted in the U.S.
among people of all ages also expands upon evidence from the 2013 Ozone ISA (U.S. EPA. 2013b) to
further support an association between ozone exposure and ED visits and hospital admissions for asthma
(Section 3.1.5.1 and Section 3.1.5.2). Reported associations were generally highest for children between
the ages of 5 and 18 at mean daily 8-hour concentrations of 31-54 ppb. Additionally, consistent, positive
associations were reported across models implementing measured and modeled ozone concentrations. A
large body of evidence from the 2013 Ozone ISA (U.S. EPA. 2013b) reported ozone associations with
markers of asthma exacerbation (e.g., respiratory symptoms, medication use, lung function) that support
the ozone-related increases in asthma hospital admissions and ED visits observed in recent studies. Few
Adams (2006)
Adams (2003)
Adams (2002)
Horstman et al. (1990)
Kim etal. (2011)*
McDonnell et al. (1991)
Schelegle et al. (2009)
McDonnell et al. (2013)
IS-28
-------
recent epidemiologic studies in the U.S. or Canada have examined respiratory symptoms and medication
use, lung function, and subclinical effects in people with asthma. Recent experimental studies in animals,
along with similar studies summarized in the 2013 Ozone ISA (U.S. EPA. 2013b). provide coherence
with and biological plausibility for the epidemiologic evidence of asthma exacerbation, indicating
respiratory tract inflammation, oxidative stress, injury, allergic skewing, goblet cell metaplasia, and
upregulation of mucus synthesis and storage in allergic mice exposed to ozone (Section 3.1.5.4.
Section 3.1.5.5. and Section 3.1.5.6).
In addition to epidemiologic evidence of asthma exacerbation, a number of recent epidemiologic
studies continue to provide evidence of an association of ozone concentrations with hospital admissions
and ED visits for combined respiratory diseases (Section 3.1.8). ED visits for respiratory infection
(Section 3.1.7.1). and ED visits for COPD (Section 3.1.6.1.1). Recent epidemiologic evidence for
respiratory mortality is limited, but there remains evidence of consistent, positive associations,
specifically in the summer months (Section 3.1.9). A limited number of recent controlled human exposure
and animal toxicological studies are consistent with studies evaluated in the 2013 Ozone ISA (U.S. EPA.
2013b) that demonstrate altered immunity and impaired lung host defense following acute ozone
exposure (Section 3.1.7.3). These findings support the epidemiologic evidence of an association between
ozone concentrations and respiratory infection. Additionally, results from recent animal toxicological
studies provide new evidence that chronic inflammation enhances sensitivity to ozone exposure,
providing coherence for ozone-related increases in ED visits for COPD (Section 3.1.6.2.1.2).
Copollutant analyses were limited in epidemiologic studies evaluated in the 2013 Ozone ISA, but
they did not indicate that associations between ozone concentrations and respiratory effects were
confounded by copollutants or aeroallergens. Copollutant analyses have been more prevalent in recent
studies and continue to suggest that observed associations are independent of coexposures to correlated
pollutants or aeroallergens (Section 3.1.10.1 and Section 3.1.10.2). Despite expanded copollutant analyses
in recent studies, determining the independent effects of ozone in epidemiologic studies is complicated by
the high copollutant correlations observed in some studies and the possibility for effect estimates to be
overestimated for the better measured pollutant in copollutant models (Section 2.5). Nonetheless, the
consistency of associations observed across studies with different copollutant correlations, the generally
robust associations observed in copollutant models, and evidence from controlled human exposure studies
demonstrating respiratory effects in response to ozone exposure in the absence of other pollutants,
provide compelling evidence for the independent effect of short-term ozone exposure on respiratory
symptoms.
Several controlled human exposure studies provided evidence on the C-R relationship for FEVi
decrements in young healthy adults exposed during moderate exercise for 6.6. hours to ozone
concentrations between 40 and 120 ppb. The lack of any studies at lower ozone concentrations and the
small decrements observed at 40 ppb preclude characterization of the C-R relationship at lower
concentrations. A model-predicted C-R function is described in a recent study presenting a mechanistic
IS-29
-------
model based on these [and other controlled human exposure data; McDonnell et al. (2013); Figure IS-1;
Section 3.1.4.1.11.
Epidemiologic studies examining the shape of the relationship between ambient air
concentrations and the studied health outcome and/or the presence of a threshold in this relationship have
been inconsistent (Section 3.1.10.1.4V While most studies assume a no-threshold, log-linear C-R shape, a
limited number of studies have used more flexible models to test this assumption. Results from some of
these studies indicate approximately linear associations between ozone concentrations and hospital
admissions for asthma, while others indicate the presence of a threshold ranging from 20 to 40 ppb 8-hour
max ozone concentrations.
Most epidemiologic studies that examine the relationship between short-term concentrations of
ozone in ambient air and health effects rely primarily on a 1-hour max, 8-hour max, or 24-hour avg
averaging times. Epidemiologic time-series and panel studies evaluated in the 2013 Ozone ISA do not
provide any evidence to indicate that any one averaging time is more consistently or strongly associated
with respiratory-related health effects (U.S. EPA. 2013b). Recent epidemiologic studies examining
respiratory effects continue to show evidence of positive associations for each of these averaging times
(see Figure 3-4. Figure 3-5. Figure 3-6. and Figure 3-7). For example, Darrow et al. (2011). as detailed in
the 2013 Ozone ISA, demonstrated a similar pattern of associations between short-term ozone exposure
and respiratory-related ED visits for 1-hour max, 8-hour max, and 24-hour avg exposure metrics
(Section 3.1.10.3.2). Similarly, a recent panel study focusing on respiratory symptoms in children
reported positive associations when using both a 1-hour max and 8-hour max averaging time I Lewis et al.
(2013); Section 3.1.5.3.21. The combination of evidence from studies evaluated in the 2013 Ozone ISA,
along with the results across recent studies that demonstrate positive associations using either a 1-hour
max, 8-hour max, or 24-hour avg averaging time, further supports the conclusion that no one averaging
time is more consistently or strongly associated with respiratory effects and that each of these averaging
times could be surrogates for the exposure conditions that elicit respiratory health effects.
The evaluation of the lag structure of associations is an important consideration when examining
the relationship between short-term ozone exposure and respiratory effects. With respect to ozone
exposure, epidemiologic studies often examine associations between short-term exposure and health
effects over a series of single-day lags, multiday lags, or by selecting lags a priori (Section 3.1.10.3). For
respiratory health effects, when examining more overt effects, such as respiratory-related hospital
admissions and ED visits (i.e., asthma, COPD, and all respiratory outcomes), epidemiologic studies
reported strongest associations occurring within the 1st few days of exposure (i.e., in the range of 0 to
3 days). The effects of ozone exposure on subclinical respiratory endpoints, including lung function,
respiratory symptoms, and markers of airway inflammation, similarly occur at lags of 0 and 1 day. This
finding is consistent with the evidence from controlled human exposure and experimental animal studies
of respiratory effects occurring relatively soon after ozone exposures.
IS-30
-------
In summary, recent studies evaluated since the completion of the 2013 Ozone ISA (U.S. EPA.
2013b) support and expand upon the strong body of evidence indicating a "causal relationship" between
short-term ozone exposure and respiratory effects. Controlled human exposure studies demonstrate
ozone-induced FEVi decrements and respiratory tract inflammation at concentrations as low as 60 ppb
after 6.6 hours of exposure with exercise among young, healthy adults. The combination of lung function
decrements and respiratory symptoms has been observed following exposure to 70 ppb and greater ozone
concentrations over 6.6-hours and combined with exercise. Epidemiologic studies continue to provide
evidence that increased ozone concentrations are associated with a range of respiratory effects, including
asthma exacerbation, COPD exacerbation, respiratory infection, and hospital admissions and ED visits for
combined respiratory diseases. A large body of animal toxicological studies demonstrate ozone-induced
changes in lung function measures, inflammation, increased airway responsiveness, and impaired lung
host defense. Additionally, mouse models indicate enhanced ozone-induced inflammation, oxidative
stress, injury, allergic skewing, goblet cell metaplasia, and upregulation of mucus synthesis and storage in
allergic mice compared with naive mice. These toxicological results provide further information on the
potential mechanistic pathways that underlie downstream respiratory effects. They also provide continued
support for the biological plausibility of the observed epidemiologic results. Thus, the recent evidence
integrated across disciplines, along with the total body of evidence evaluated in previous assessments, is
sufficient to conclude that there is a "causal relationship" between short-term ozone exposure and
respiratory effects.
IS.4.3.2 Long-Term Exposure and Respiratory Effects
The 2013 Ozone ISA concluded that there was "likely to be causal relationship" between
long-term exposure to ozone and respiratory health effects (U.S. EPA. 2013b). The epidemiologic
evidence for a relationship between long-term ozone exposure and respiratory effects in the 2013 Ozone
ISA was provided by epidemiologic studies that typically evaluated the association between the annual
average of daily ozone concentrations and new-onset asthma, respiratory symptoms in children with
asthma, and respiratory mortality. Notably, associations of long-term ozone concentrations with
new-onset asthma in children and increased respiratory symptoms in individuals with asthma were
primarily observed in studies that examined interactions between ozone and exercise or different genetic
variants. The evidence relating new-onset asthma to long-term ozone exposure was supported by
toxicological studies of allergic airways disease in infant monkeys exposed to biweekly cycles of
alternating filtered air and ozone (i.e., 9 consecutive days of filtered air and 5 consecutive days of 0.5 ppm
ozone, 8 hours/day). This evidence from a nonhuman primate study of ozone-induced changes in the
airways provided biological plausibility for early-life exposure to ozone contributing to asthma
development in children. Generally, the consistent evidence from epidemiologic and animal toxicological
studies formed the basis of the conclusions that there is "likely to be causal relationship" between
long-term exposure to ambient ozone and respiratory effects. Uncertainties in the evidence base included
limited assessment of potential copollutant confounding and the potential for exposure measurement error
IS-31
-------
relating to exposure assignment from fixed site monitors in epidemiologic studies. Although potential
copolluant confounding was examined in a limited number of epidemiologic studies, results suggested
that the reported associations were robust to adjustment for other pollutants, including PM25 Building
upon the evidence from the 2013 Ozone ISA, more recent epidemiologic evidence, combined with
toxicological studies in rodents and nonhuman primates, provides coherence and biological plausibility to
support that there is a "likely to be causal relationship" between long-term exposure to ozone and
respiratory effects.
Recent studies continue to examine the relationship between long-term exposure to ozone and
respiratory effects. Key evidence supporting the causality determination is presented in Table IS-5. A
limited number of recent epidemiologic studies provide generally consistent evidence that long-term
ozone exposure is associated with the development of asthma in children (Section 3.2.4.1.1). In addition
to investigating the development of asthma, epidemiologic studies have evaluated the relationship
between ozone exposure and asthma severity (Section 3.2.4.5). Like the studies described in the 2013
Ozone ISA (U.S. EPA. 2013b). recent studies provide evidence of consistent positive associations
between long-term exposure to ozone and hospital admissions and ED visits for asthma and prevalence of
bronchitic symptoms in children with asthma. Notably, some uncertainty remains about the validity of the
results from studies examining long-term ozone exposure and hospital admissions and ED visits for
asthma, because most of these studies do not adjust for short-term ozone concentrations, despite the
causal relationship between short-term exposure and asthma exacerbation (Section 3.1.4.2).
Table IS-5 Summary of evidence from epidemiologic and animal toxicological
studies on the respiratory effects associated with long-term ozone
exposure.
Conclusions from 2013 Ozone ISA
Conclusions from 2020 ISA3
Respiratory effects Epidemiologic evidence, combined with
toxicological studies in rodents and nonhuman
primates, provided biologically plausible
evidence that there is likely to be causal
relationship between long-term exposure to
ozone and respiratory effects.
Epidemiologic evidence, combined with
toxicological studies in rodents and
nonhuman primates, continue to provide
biologically plausible evidence for respiratory
effects due to long-term ozone exposure.
Overall, the collective evidence is
sufficient to conclude that there is a likely
to be causal relationship between
long-term ozone exposure and
respiratory effects.
IS-32
-------
Table IS-5 (Continued): Summary of evidence from epidemiologic and animal
toxicological studies on the respiratory effects
associated with long-term ozone exposure.
Conclusions from 2013 Ozone ISA
Conclusions from 2020 ISA3
New onset asthma
Animal toxicological studies provided evidence
that perinatal exposure to ozone compromises
airway growth and development in infant
monkeys (500 ppb; 6 h a day, 5 days a week for
20 weeks). Animal toxicological studies also
demonstrate increased airway responsiveness,
allergic airways responses, and persistent
effects on the immune system, which may lead
to the development of asthma. There is
evidence that different genetic variants (HMOX,
GST, ARG), in combination with ozone
exposure, are related to new-onset asthma.
These associations were observed when
subjects living in areas where the mean annual
8-h daily max ozone concentration was
55.2 ppb, compared with those who lived in
areas with a mean of 38.4 ppb.
Recent epidemiologic studies provide
generally consistent evidence for
associations of long-term ozone exposure
with the development of asthma in children.
Associations observed in locations with
mean annual concentrations of 32.1 ppb in
one study that reported study mean
concentrations (community-specific annual
average concentrations ranged from 26 to
76 ppb). Recent animal toxicological studies
demonstrate effects on airway development
in rodents (500 ppb; 6 h a day for
3-22 weeks) and build on and expand the
evidence for long-term ozone
exposure-induced effects that may lead to
asthma development.
Asthma hospital Epidemiologic studies provided evidence that
admissions long-term ozone exposure is related to
increased hospital admissions in children and
adults, and first childhood asthma hospital
admissions in a linear concentration-response
relationship. Generally, these studies had mean
annual 8-h daily max ozone concentrations less
than 41 ppb.
Long-term exposure is associated with
hospital admissions and ED visits for asthma
in study locations with mean annual ozone
concentrations between 30.6 and 47.7 ppb,
although uncertainties remain because most
studies do not adjust for short-term ozone
concentrations.
Pulmonary Evidence for pulmonary function effects was
structure and inconsistent, with some epidemiologic studies
function observing positive associations (mean annual
8-h daily max ozone concentrations less than
65 ppb). Results from toxicological studies
demonstrated that long-term exposure of adult
monkeys and rodents (>120 ppb; 6 h a day,
5 days a week for 20 weeks) can result in
irreversible morphological changes in the lung,
which in turn can influence pulmonary function.
Recent animal toxicological studies provide
evidence that postnatal ozone exposure may
affect processes in the developing lung,
including impaired alveolar morphogenesis,
a key step in lung development, in infant
monkeys (500 ppb; 6 h a day for
3-22 weeks). Notably, the impairments in
alveolar morphogenesis were reversible
(reversibility of the other effects was not
studied). A limited number of recent
epidemiologic studies continue to provide
inconsistent support for an association
between long-term ozone exposure and lung
function development in children.
Pulmonary Several epidemiologic studies (mean 8-h max
inflammation, ozone concentrations less than 69 ppb) and
injury, and animal toxicological studies (as low as 500 ppb)
oxidative stress added to existing evidence of ozone-induced
inflammation and injury.
Recent experimental studies in animals
provide evidence that postnatal ozone
exposure may affect the developing lung
(500 ppb). Results from studies of neonatal
rodents demonstrate ozone-induced
changes in injury and inflammatory and
oxidative stress responses during lung
development (1,000 ppb).
Lung host
defenses
Evidence demonstrated a decreased ability to
respond to pathogenic signals in infant monkeys
exposed to 500 ppb ozone and an increase in
severity of post-influenza alveolitis in rodents
exposed to 500 ppb.
A recent study demonstrates decreased
ability to respond to pathogenic signals in
infant monkeys exposed to 500 ppb.
IS-33
-------
Table IS-5 (Continued): Summary of evidence from epidemiologic and animal
toxicological studies on the respiratory effects
associated with long-term ozone exposure.
Conclusions from 2013 Ozone ISA
Conclusions from 2020 ISA3
Allergic responses
Evidence demonstrated a positive association
between allergic response and ozone exposure,
but the magnitude of the association varied
across studies; exposure to ozone may increase
total IgE in adult asthmatics. Allergic indicators
in infant monkeys and adult rodents were
increased by exposure to ozone concentrations
of 500 ppb.
Cross-sectional epidemiologic studies
provide generally consistent evidence that
ozone concentrations (mean annual
concentration less than 51.5 ppb) are
associated with hay fever/rhinitis and
serum-markers of allergic response,
although uncertainties related to study
design and potential confounding by pollen
remain. A recent animal toxicological study
provides evidence of ozone-induced airway
eosinophilia in a mouse model of allergic
sensitization (100 ppb; 0.33 h per day for
5 days per week for 2 weeks and once
weekly for 12 weeks).
Development of Animal toxicological studies provided evidence
COPD that long-term ozone exposure could lead to
persistent inflammation and interstitial
remodeling in adult rodents and monkeys,
potentially contributing to the development of
chronic lung disease such as COPD. The 2013
Ozone ISA did not evaluate any epidemiologic
studies that examined the relationship between
long-term exposure to ozone and the
development of COPD.
One recent epidemiologic study provides
evidence of an association between
long-term ozone concentrations and incident
COPD hospitalizations (mean annual
concentrations 39.3 ppb). Recent animal
toxicological studies provide consistent
evidence that subchronic ozone exposure
(500-1,000 ppb) can lead to airway injury
and inflammation. In adult animals, these
changes may underlie the progression and
development of chronic lung disease and
provide biological plausibility for
ozone-induced development of COPD.
Respiratory A single study demonstrated that exposure to
mortality ozone (long-term mean ozone less than
104 ppb) elevated the risk of death from
respiratory causes. This effect was robust to the
inclusion of PM2.5 in a copollutant model.
Recent epidemiologic studies provide some
evidence of an association with respiratory
mortality, but the evidence is not consistent
(mean annual ozone concentrations
25.9-57.5 ppb). New evidence from one
study reports an association with COPD
mortality.
Conclusions from the 2020 ISA include evidence from recent studies integrated with evidence included in previous Ozone ISAs
and AQCDs.
In support of evidence from recent epidemiologic studies, a number of recent animal
toxicological studies expand the evidence base for long-term ozone exposure-induced effects leading to
asthma development (Section 3.2.4.1.2). Specifically, both older and more recent long-term ozone
exposure studies in nonhuman primates show that postnatal ozone exposure can compromise airway
growth and development, promote the development of an allergic phenotype, and cause persistent
alterations to the immune system (Section 3.2.4.6.2V In addition, findings that ozone exposure enhances
injury, inflammation, and allergic responses in allergic rodents provide biological plausibility for the
relationship between ozone exposure and the exacerbation of allergic asthma.
IS-34
-------
In addition to studies of asthma, several new or expanded lines of evidence from epidemiologic
and animal toxicological studies published since the completion of the 2013 Ozone ISA provide evidence
of associations between long-term ozone exposure and the development of COPD (Section 3.2.4.3) and
allergic responses (Section 3.2.4.6V A recently available epidemiologic study provides limited evidence
that long-term ozone exposure is associated with incident COPD hospitalizations in adults with asthma.
This finding is supported by recent animal toxicological studies that provide consistent evidence of
airway injury and inflammation resulting from subchronic ozone exposures. These results are coherent
with animal toxicological studies reviewed in the 2013 Ozone ISA, which demonstrated that chronic
ozone exposure damages distal airways and proximal alveoli, resulting in persistent inflammation and
lung tissue remodeling that leads to irreversible changes including fibrotic- and emphysematous-like
changes in the lung. Respiratory tract inflammation and morphologic and immune system-related changes
may underlie the progression and development of chronic lung disease like COPD.
A larger body of epidemiologic studies also supports an association between long-term ozone
exposure and allergic responses, including hay fever/rhinitis and serum allergen-specific IgE. While
recent studies demonstrate generally consistent results, potential confounding by pollen exposure remains
an uncertainty. However, there is supporting evidence from animal toxicological studies demonstrating
enhanced allergic responses in allergic rodents (Section 3.2.4.6.2). In addition, animal toxicological
studies reviewed in the short-term exposure section show type 2 immune responses in nasal airways of
rodents exposed repeatedly to ozone, indicating that ozone exposure can trigger allergic responses
(Section 3.1.4.4.2). These findings are characteristic of induced nonatopic asthma and rhinitis and provide
biological plausibility for the observed epidemiologic associations with hay fever/rhinitis.
Taken together, previous and more recent animal toxicological studies of long-term exposure to
ozone provide biological plausibility for the associations reported in the recent epidemiologic studies.
Specifically, there is strong evidence of ozone-induced inflammation, injury, and oxidative stress in adult
animals. These effects represent initial events through which ozone may lead to a number of downstream
respiratory effects, including altered morphology in the lower respiratory tract and the development of
COPD. Furthermore, there is evidence of a range of ozone-induced effects on lung development in
neonatal rodents and infant monkeys, including altered airway architecture, airway sensory nerve
innervation, airway cell death pathways, increased serotonin-positive airway cells, and
immunomodulation. An infant monkey model of allergic airway disease also demonstrated effects on lung
development, including compromised airway growth, impaired alveolar morphogenesis, airway smooth
muscle hyperreactivity, an enhanced allergic phenotype, priming of responses to oxidant stress, and
persistent effects on the immune system. These various upstream effects provide a plausible pathway
through which ozone may act on downstream events. These events include altered immune function
leading to altered host defense and allergic responses, as well as morphologic changes leading to the
development of asthma. A more thorough discussion of the biological pathways that potentially underlie
respiratory health effects resulting from long-term exposure to ozone can be found in Section 3.2.3.
IS-35
-------
Recent epidemiologic studies provide some evidence that long-term ozone exposure is associated
with respiratory mortality, but the evidence is not consistent across studies (Section 3.2.4.9V A recent
nationwide study in the U.S. reported associations between ozone and the underlying causes of respiratory
mortality, including COPD. This finding is supported by the new lines of evidence from animal
toxicological and epidemiologic studies on the development of COPD, as discussed previously. Results
from epidemiologic studies of ozone-related respiratory mortality in populations outside the U.S. are
inconsistent.
A notable source of uncertainty across the reviewed epidemiologic studies is the lack of
examination of potential copollutant confounding. A limited number of studies that include results from
copollutant models suggest that ozone associations may be attenuated but still positive after adjustment
for NO2 or PM2 5. However, the few studies that include copollutant models examine different outcomes,
making it difficult to draw strong conclusions about the nature of potential copollutant confounding for
any given outcome. Importantly, in addition to studies that explicitly address potential copollutant
confounding through modeling adjustments, many studies report modest copollutant correlations,
suggesting that strong confounding due to copollutants is unlikely. Another source of uncertainty
common to epidemiologic studies of air pollution is the potential for exposure measurement error. The
majority of recent epidemiologic studies of long-term ozone exposure use concentrations from fixed-site
monitors as exposure surrogates. Exposure measurement error relating to exposure assignment from
fixed-site monitors has the potential to bias effect estimates in either direction, although it is more
common that effect estimates are underestimated, and the magnitude of the bias is likely small relative to
the magnitude of the effect estimate, given that ozone concentrations do not vary over space as much as
other criteria pollutants, such as NO2 or SO2 (Section 2.3.1.1V
Strong coherence from animal toxicological studies supports the observed epidemiologic
associations related to respiratory morbidity. Experimental evidence also provides biologically plausible
pathways through which long-term ozone exposure may lead to respiratory effects. Overall, the
collective evidence supports a "likely to be causal relationship" between long-term ozone exposure
and respiratory effects.
IS.4.3,3 Short-Term Exposure and Metabolic Effects
The metabolic effects reviewed in this ISA include the risk factors and related endpoints for
metabolic syndrome, complications due to diabetes, and indicators of metabolic function. Metabolic
syndrome is a clinical diagnosis used to describe a collection of risk factors that include high blood
pressure (elevated systolic and/or diastolic blood pressure), dyslipidemia (elevated triglycerides and low
levels of high-density lipoprotein [HDL] cholesterol), obesity (central obesity), and increased fasting
blood glucose (Alberti et al.. 2009). Diagnosis of metabolic syndrome in humans is based on the presence
of three of these five risk factors (Alberti et al.. 2009). The presence of these risk factors may predispose
IS-36
-------
individuals to an increased risk of type 2 diabetes and cardiovascular disease. Diabetes is characterized by
hyperglycemia (i.e., elevated glucose level) resulting from defects in insulin signaling, secretion, or both.
Indicators of metabolic function include adipose tissue inflammation, altered liver function, and
alterations in adrenal hormones, among other endpoints.
The evidence was not sufficient to evaluate metabolic effects as a separate health effect category
in the 2013 Ozone ISA. As a result, there were no causality determinations for metabolic effects in the
2013 Ozone ISA (U.S. EPA. 2013b). Since the completion of the 2013 Ozone ISA, the number of studies
examining the relationship between short-term ozone exposure and metabolic effects has expanded
substantially (Table IS-6). This recent evidence, primarily from experimental animal studies,
demonstrates that short-term ozone exposure triggers a stress response that leads to a cascade of transient
metabolic effects. Consistent animal toxicological evidence from multiple laboratories demonstrates that
short-term ozone exposure increases circulating levels of adrenaline and corticosterone, released from the
adrenal medulla and adrenal cortex, respectively (Section 5.1.5.3.2). This evidence is coherent with
results of a controlled human exposure study demonstrating that short-term ozone exposure to 300 ppb
resulted in increased circulating Cortisol and corticosterone. In animals, the metabolic effects that follow
short-term ozone exposure are similar to those that are used in the clinical diagnosis of metabolic
syndrome in humans. The strongest and most consistent evidence is for glucose and insulin homeostasis.
Several high-quality animal toxicological studies from multiple laboratories demonstrate that short-term
ozone exposure impairs glucose tolerance and causes insulin resistance (Section 5.1.3). Some but not all
animal toxicological studies show ozone-induced fasting hyperglycemia, with inconsistencies between
studies potentially caused by differences in rodent stock, strain, sex, or diet. Multiple animal toxicological
studies in several rodent strains demonstrate that short-term ozone exposure increases serum levels of
triglycerides and free fatty acids, results that are consistent with the mobilization of energy stores and
increased glucose (Section 5.1.3.2V Coherent with results in animal models, the controlled human
exposure study reported increases in medium- and long-chain circulating free fatty acids following
short-term exposure to 300 ppb ozone. However, this study did not find ozone-induced changes in serum
insulin, nonfasting glucose, insulin resistance, or triglyceride levels. Some epidemiologic studies
examining changes in glucose and lipids provide support for effects associated with short-term ozone
exposure.
There is additional evidence for ozone-induced metabolic effects from experimental animal
toxicological studies that are the same effects used for the clinical diagnosis of metabolic syndrome in
humans. This generally consistent evidence demonstrates that short-term ozone exposure affects
obesity-relevant endpoints and causes adipose tissue inflammation. Some, but not all, animal
toxicological studies reported that short-term ozone exposure reduces body-weight gain1 in rodent models
of diabetes and of spontaneous hypertension (Section 5.1.5V In addition, multiple animal toxicological
studies from different laboratories consistently reported that short-term ozone exposure affected levels of
1 Reductions or increases in body-weight gain can indicate altered metabolic function in animal models of disease,
such as those used in these studies.
IS-37
-------
leptin, a hormone that regulates food intake. In coherence with these results, an epidemiologic study
reported trends for an association between short-term ozone exposure and changes in the obesity-related
hormones. In addition to changes in hormone levels, multiple animal toxicological studies in both healthy
and disease-prone rodent models showed that short-term ozone exposure can induce adipose tissue
inflammation (Section 5.1.5.1). Furthermore, while several studies reported null effects, others reported
that short-term ozone exposure can affect levels of HDL, LDL, and total cholesterol, with the
directionality of the effect varying by the rodent model and exposure duration (Section 5.1.5.1). Finally,
some animal toxicological studies provide evidence that short-term ozone exposure affects blood pressure
(Section 5.1.3.5).
Table IS-6 Summary of evidence from epidemiologic, controlled human
exposure, and animal toxicological studies on the metabolic effects
of short-term exposure to ozone.
Results and Conclusions from 2020 ISA
Metabolic effects
Recent evidence from controlled human exposure, epidemiologic, and animal
toxicological studies support a likely to be causal relationship between short-term
ozone exposure and metabolic effects.
Effects contributing to
the clinical diagnosis of
metabolic syndrome in
humans
Animal toxicological studies provide evidence for elevated triglycerides and fasting
hyperglycemia. Evidence is present, though less consistent, for low HDL cholesterol, high
blood pressure, and central adiposity.
Complications from
diabetes
An epidemiologic study provides evidence of associations between increases in
short-term ozone exposure and hospital admissions for diabetic ketoacidosis and diabetic
coma in older population subgroups.
Other indicators of Multiple metabolic indicators provide evidence that ozone exposure induces changes
metabolic function within the liver, affecting glucose homeostasis. Healthy volunteers who exercised with
ozone exposure in controlled human exposure studies had increased ketone body
formation. In animal toxicological studies, ozone exposure induced changes to the liver,
including hepatic gluconeogenesis, altered bile acid profile, alterations to (3-oxidation, and
alterations to proteins in hepatic metabolic pathways. In addition, elevated circulating
stress hormones were consistently observed in animal models and in a single controlled
human exposure study. Removal of the adrenal glands prevented the release of
adrenaline and corticosterone, and furthermore, prevented ozone-induced metabolic
effects in animal toxicological studies. Thus, neuroendocrine stress activation may be a
primary mechanism through which adverse metabolic outcomes develop from short-term
ozone exposure.
Recent studies of short-term ozone exposure and metabolic effects evaluated associations
between different age groups. One epidemiologic study observed increased risk among older adults
(e.g., 75-84 years and 85+ years) compared with other age groups (<65 years) for hospital admissions for
diabetic coma (Section 5.1.7.1) with a 24-hour avg ozone concentration across study areas of 64.4 ppb. In
IS-38
-------
addition, an animal toxicological study demonstrated increases in metabolic indicators (i.e., increased
triglycerides and serum insulin) in aged animals.
Despite limited controlled human exposure and epidemiologic evidence, the expanding body of
animal toxicological studies shows robust evidence for short-term ozone exposure contributing to an array
of metabolic effects. These outcomes follow a biologically plausible pathway whereby ozone exposure
results in release of adrenaline and cortisol/corticosterone from the adrenal glands. These hormones act on
multiple organs and tissues of the metabolic system to mobilize energy reserves, including glucose and
lipids. In summary, based on evidence from animal toxicological and epidemiologic studies, as well as
some support from one controlled human exposure study, short-term ozone exposure consistently impairs
glucose and insulin homeostasis and increases triglycerides and fatty acids. In line with this, animal
toxicological studies show that inhibiting adrenaline and/or corticosterone, through either removal of the
adrenal glands or adrenal medulla, or by blocking the synthesis of corticosterone, prevents ozone-induced
metabolic effects, including hyperglycemia, glucose intolerance, and elevated circulating triglycerides. In
addition, there are generally consistent effects from animal toxicological studies showing that short-term
ozone exposure affects obesity-relevant endpoints and causes inflammation in adipose tissue. Further
supporting evidence comes from a limited number of animal toxicological studies providing some
evidence for alterations in HDL, LDL, and total cholesterol and changes in blood pressure following
short-term ozone exposure. Overall, the collective evidence is sufficient to conclude that the
relationship between short-term ozone exposure and metabolic effects is likely to be causal.
IS.4.3.4 Short-Term Exposure and Cardiovascular Effects
The 2013 Ozone ISA concluded that there is a "likely to be causal" relationship between relevant
short-term exposures and cardiovascular effects, but it also identified important uncertainties (U.S. EPA.
2013b). The available animal toxicological studies demonstrated ozone-induced impaired vascular and
cardiac function, as well as changes in heart rate (HR) and heart rate variability (HRV). The controlled
human exposure studies provided additional evidence but had limited coherence with the evidence from
animal studies. The epidemiologic evidence, while reporting associations between short-term ozone
exposure and cardiovascular mortality, did not show associations between short-term ozone exposure and
cardiovascular morbidity. This lack of coherence between the results for studies investigating associations
of cardiovascular morbidity with cardiovascular mortality was recognized as a complication in
interpreting the overall evidence for ozone-induced cardiovascular effects.
More recent animal toxicological studies published since the 2013 Ozone ISA provide generally
consistent evidence for impaired heart function and endothelial dysfunction, but limited evidence for
indicators of arrhythmia, HRV, and markers of oxidative stress and inflammation in response to ozone
exposure. Additional controlled human exposure studies have been published in recent years, although
they show little evidence for ozone-induced effects on cardiovascular endpoints. Specifically, some recent
IS-39
-------
studies do not indicate an effect of ozone on cardiac function, ST segment, endothelial dysfunction, or
HR, while some evidence from a small number of controlled human exposure studies indicates ozone
exposure can result in changes in blood pressure, indicators of arrhythmia, HRV, markers of coagulation,
and inflammatory markers. The number of epidemiologic studies evaluating short-term ozone
concentrations and cardiovascular effects has grown somewhat, but overall, remains limited and continues
to provide little, if any, evidence for associations with heart failure, heart attack, arrhythmia and cardiac
arrest, or stroke. Recent epidemiologic evidence for short-term ozone exposure and cardiovascular
mortality is limited to one multicity study, but the collective body of evidence spanning multicity studies
evaluated in the 2013 Ozone ISA provides evidence of consistent positive associations. Overall, many of
the same limitations and uncertainties that existed in the body of evidence in the 2013 Ozone ISA
continue to exist. However, the number of controlled human exposure studies evaluating short-term ozone
exposure and cardiovascular endpoints has grown, and now includes studies at concentrations closer to
those likely to be encountered in U.S. ambient air. When evaluated in the context of the studies available
for the 2013 Ozone ISA, the controlled human exposure study evidence, overall, is less consistent and
less indicative of a relationship (Table IS-7).
Table IS-7 Summary of evidence from epidemiologic, controlled human
exposure, and animal toxicological studies on the cardiovascular
effects of short-term ozone exposure.
Conclusions from 2013 Ozone ISA
Conclusions from 2020 ISA
Cardiovascular effects
Evidence from animal toxicological
studies demonstrated ozone-induced
impaired vascular and cardiac function,
as well as changes in HR and HRV. This
evidence was supported from a limited
number of controlled human exposure
studies in healthy adults demonstrating
changes in HRV, as well as in blood
markers associated with an increase in
coagulation. There was limited or no
evidence from epidemiologic studies for
short-term ozone exposure and
cardiovascular morbidity, such as effects
related to HF, IHD, and Ml, arrhythmia
and cardiac arrest, or thromboembolic
disease. There was consistent evidence
from epidemiologic studies reporting
positive associations between short-term
ozone exposure and
cardiovascular-related mortality. Overall,
there is likely to be causal relationship
between long-term exposure to ozone
and cardiovascular effects.
Recent animal toxicological studies continue to
provide evidence for impaired heart function
and endothelial dysfunction, with limited
evidence from a small number of studies for
indicators of arrhythmia, HRV, and markers of
oxidative stress and inflammation in response
to ozone exposure. Recent controlled human
exposure studies provide little evidence for
ozone-induced effects on a number of
cardiovascular endpoints. No effect of ozone
was reported for indicators of cardiac function,
IHD, endothelial dysfunction, or changes in
HR. There is limited or inconsistent evidence
from a small number of studies for changes in
cardiac electrophysiology, HRV, blood
pressure, markers of coagulation, and
inflammatory markers. Epidemiologic studies
remain few and continue to provide little, if any,
evidence for associations with HF, IHD, and
Ml, arrhythmia and cardiac arrest, or stroke.
Overall, the evidence is suggestive of, but
not sufficient to infer, a causal relationship.
IS-40
-------
Table IS-7 (Continued): Summary of evidence from epidemiologic, controlled
human exposure, and animal toxicological studies on the
cardiovascular effects of short-term ozone exposure.
Conclusions from 2013 Ozone ISA
Conclusions from 2020 ISA
Heart failure, impaired
heart function
A limited number of animal toxicological
studies demonstrated ozone-induced
cardiovascular effects, including
decreased cardiac function.
Epidemiologic studies generally did not
observe associations between short-term
ozone exposure and cardiovascular
morbidity; studies of
cardiovascular-related hospital
admissions and ED visits did not find
consistent evidence of a relationship with
ozone exposure.
Multiple animal toxicological studies report
some indicators of impaired cardiac function
following short-term ozone exposure
(~200-300 ppb for 3-4 h). However, a recent
controlled human exposure study (100 and
200 ppb for 3 h) reported no changes in
measures of cardiac function. There is a
limited number of recent studies of hospital
admissions and ED visits that analyzed
associations with heart failure, and they
continue to report inconsistent associations
with short-term exposure to ozone.
Ischemic heart disease
Animal toxicological studies, although
few, demonstrated ozone-induced
cardiovascular effects, including
enhanced ischemia/reperfusion (l/R)
injury. Epidemiologic studies generally did
not observe associations between
short-term ozone exposure and
cardiovascular morbidity; studies of
cardiovascular-related hospital
admissions and ED visits did not find
consistent evidence of a relationship with
ozone exposure.
An animal toxicological study in SH rats
demonstrates ST segment depression
following an 800- but not 200-ppb exposure to
ozone for 4 h. However, no such changes are
observed in the single controlled human
exposure study (70 and 120 ppb for 3 h).
Recent epidemiologic studies consistently
report null or weak positive effect estimates in
analyses of Ml, including for STEMI and
NSTEMI.
Cardiac and endothelial
dysfunction
Animal toxicological studies, although
limited in number, demonstrated
ozone-induced cardiovascular effects,
including vascular disease and injury.
Recent animal toxicological studies
demonstrate generally consistent evidence for
impaired cardiac and endothelial function in
rodents following short-term ozone exposure of
400-1,000 ppb for 4 h. However, coherence
with controlled human exposure and
epidemiologic studies is lacking.
Cardiac
electrophysiology,
arrhythmia, cardiac
arrest
Animal toxicological studies, although
few, demonstrated ozone-induced
cardiovascular effects, including disrupted
nitric oxide-induced vascular reactivity.
Epidemiologic studies reported generally
positive associations for hospital
admissions or ED visits due to arrythmia
or dysrhythmia.
A small number of recent animal toxicological
studies demonstrate some evidence for
changes in indicators of conduction
abnormalities (800 but not 200 ppb for 3-4 h).
Multiple controlled human exposure studies
report little effect of short-term ozone exposure
on conduction abnormalities (70 and 120 ppb
for 2-3 h). Increases in out-of-hospital cardiac
arrests associated with 8-h max or 24-h avg
increases in ozone concentrations were
reported by a few case-crossover studies;
however, analyses of other endpoints
(e.g., dysrhythmia, arrhythmia, or atrial
fibrillation) generally report null results.
IS-41
-------
Table IS-7 (Continued): Summary of evidence from epidemiologic, controlled
human exposure, and animal toxicological studies on the
cardiovascular effects of short-term ozone exposure.
Conclusions from 2013 Ozone ISA
Conclusions from 2020 ISA
Blood pressure
changes and
hypertension
A limited number of epidemiologic studies
reported inconsistent associations with
measures of blood pressure. Two studies
observed increases in DBP associated
with ozone concentration, but the
association was attenuated to null after
adjusting for PM2.5 concentrations.
Recent animal toxicological studies
demonstrate inconsistent effects of
ozone-induced effects on changes in blood
pressure (300 and 500 ppb for 3-8 h).
Multiple controlled human exposure studies
report no evidence of an ozone-induced effect
on blood pressure (120-700 ppb for 1-3 h),
while a single controlled human exposure
study reported a decrease in DBP. Few
epidemiologic panel studies evaluated blood
pressure, and the results were inconsistent.
Heart rate and heart
rate variability
Animal toxicological studies, although
few, demonstrated ozone-induced
cardiovascular effects, including
increased HRV. Controlled human
exposure studies provided some
coherence with the evidence from animal
toxicological studies, by demonstrating
increases and decreases in HRV
following relatively low (120 ppb during
rest) and high (300 ppb with exercise)
ozone exposures, respectively.
Evidence is inconsistent for changes in HR in
animals (~200-800 ppb for 3-8 h) and lacking
for changes in HR in healthy adults from
multiple controlled human exposure studies
(70-300 ppb for 1-4 h). With respect to HRV,
there is limited evidence for changes in animal
toxicological (200-800 ppb for 3-4 h) and
controlled human exposure (70-300 ppb for
1-4 h) studies. Similarly, recent epidemiologic
panel studies have reported inconsistent
associations between short-term exposure to
ozone and both HR and HRV.
Coagulation and
thrombosis
A controlled human exposure study
demonstrated changes in markers of
coagulation following short-term ozone
exposure. Specifically, there were
decreases in PAI-1 and plasminogen
levels and a trend toward an increase in
tPA. There was very limited animal
toxicological evidence that short-term
exposure to ozone could result in an
increase in factors related to coagulation.
Epidemiologic studies observed
inconsistent results for coagulation
biomarkers such as PAI-1, fibrinogen,
and vWF.
Recent animal toxicological studies provide
limited evidence for changes in factors that
may promote coagulation (250-1,000 ppb for
4 h). Similarly, there is limited additional
evidence from recent controlled human
exposure studies that short-term ozone
exposure can result in changes to markers of
coagulation that may promote thrombosis
(100-300 ppb for 1-2 h). Epidemiologic
studies continue to observe inconsistent
associations with changes in biomarkers of
coagulation.
Systemic inflammation
and oxidative stress
Controlled human exposure studies
demonstrated ozone-induced effects on
blood biomarkers of systemic
inflammation and oxidative stress.
There is inconsistent evidence from recent
animal toxicological studies for an increase in
markers associated with systemic inflammation
and oxidative stress (300-800 ppb for 2-24 h)
and some evidence for increases in markers of
systemic inflammation from CHE studies
(100-300 ppb for 0.5-4 h). Additionally, the
newly available epidemiologic panel study did
not observe an association between short-term
ozone concentrations and myeloperoxidase.
Stroke
A limited number of epidemiologic studies
observed inconsistent associations with
stroke.
Inconsistent results were observed in several
recent epidemiologic studies that analyzed
hospital admissions and ED visits for stroke
and stroke subtypes.
IS-42
-------
Table IS-7 (Continued): Summary of evidence from epidemiologic, controlled
human exposure, and animal toxicological studies on the
cardiovascular effects of short-term ozone exposure.
Conclusions from 2013 Ozone ISA
Conclusions from 2020 ISA
Cardiovascular hospital
admissions and ED
visits
With few exceptions, studies of ozone
concentrations and cardiovascular
hospital admissions and ED visits for all
CVD diagnoses combined did not report
positive associations.
Recent studies that reported a risk ratio for
combined cardiovascular disease outcomes
show a similar inconsistent pattern to those
studies included in the 2013 Ozone ISA.
Cardiovascular Multicity epidemiologic studies observed
mortality positive associations for cardiovascular
mortality in all-year and summer/warm
season analyses. Lack of coherence with
epidemiologic studies of cardiovascular
morbidity remains an important
uncertainty.
A recent multicity study is consistent with the
evidence examining cardiovascular mortality
evaluated in the 2013 Ozone ISA.
Conclusions from the 2020 ISA include evidence from recent studies integrated with evidence included in previous Ozone ISAs
and AQCDs.
When considered as a whole, the evidence is "suggestive of, but not sufficient to infer, a
causal relationship" between short-term exposure to ozone and cardiovascular effects. This causality
determination represents a change from the conclusion in the 2013 Ozone ISA. This change is largely
because the number of controlled human exposure studies showing little evidence of ozone-induced
cardiovascular effects has grown substantially, while the epidemiologic evidence for ozone effects on
endpoints other than mortality continues to be limited. Consequently, the plausibility for a relationship
between short-term ozone exposure to cardiovascular health effects is weaker than it was in the previous
review, leading to the revised causality determination.
IS.4.3.5 Short-Term Exposure and Total Mortality
Recent multicity epidemiologic studies conducted in the U.S. and Canada continue to provide
evidence of consistent, positive associations between short-term ozone exposure and total mortality in
both all-year and summer/warm season analyses across different averaging times (i.e., max daily 1-hour
max, max daily 8-hour avg, 8-hour avg, and 24-hour avg; Table IS-8). Cause-specific mortality
(e.g., respiratory mortality, cardiovascular mortality) was assessed in a limited number of recent studies.
The evidence from these recent studies is consistent with the pattern of positive associations reported for
studies evaluated in the 2013 Ozone ISA. Lastly, most of the recent multicity studies examined
associations between short-term ozone exposure and mortality using ozone data collected before the year
2000, with only Di et al. (2017) including more recent ozone concentration data.
Recent studies continue to assess the influence of important potential confounders on the
ozone-mortality relationship, including copollutants, temporal/seasonal trends, and weather covariates.
Overall, these studies report that associations remain relatively unchanged across the different approaches
IS-43
-------
used to control for each confounder. The assessment of potential copollutant confounding in recent
studies demonstrates that associations between short-term ozone concentrations and mortality remain
positive in copollutant models with PMi0 or NO2. Importantly, the issues surrounding the assessment of
potential copollutant confounding that complicate interpretation of the ozone-mortality relationship (as
detailed in the 2013 Ozone ISA) persist, specifically within studies that relied on PM data collected using
every 3rd- and 6th-day sampling schedules (U.S. EPA. 2013b).
Building upon the 2013 Ozone ISA, there remains strong evidence for respiratory effects due to
short-term ozone exposure (Appendix 3) that is consistent within and across disciplines and which
provides coherence and biological plausibility for the positive respiratory mortality associations reported
across epidemiologic studies. Although there remains epidemiologic evidence for ozone-induced
cardiovascular mortality along with animal toxicological evidence of cardiovascular effects, recent
controlled human exposure studies do not provide evidence that is consistent with the controlled human
exposure studies presented in the 2013 Ozone ISA showing cardiovascular effects. Additionally, there is
limited evidence from epidemiologic studies of relationships between short-term ozone exposure and
more severe cardiovascular effects, such as emergency department visits and hospital admissions. The
limited experimental evidence, in combination with the lack of coherence between experimental and
epidemiologic studies of cardiovascular morbidity, does not allow for an understanding of potential
biological pathways leading to cardiovascular mortality (Appendix 4) or other causes of mortality.
Overall, the recent multicity studies conducted in the U.S. and Canada provide additional support
for the consistent, positive associations with total mortality reported across multicity studies evaluated in
the 2006 Ozone AQCD (U.S. EPA. 2006a) and 2013 Ozone ISA (U.S. EPA. 2013b). These results are
supported by studies that further examine uncertainties in the ozone-mortality relationship, such as
potential confounding by copollutants and other variables, modification by temperature, and the C-R
relationship and whether a threshold exists. Although there continues to be strong evidence from studies
of respiratory morbidity to support respiratory mortality, there remains relatively limited biological
plausibility and coherence within and across disciplines to support the epidemiologic evidence for
cardiovascular mortality, the largest contributor to total mortality. Collectively, evidence is "suggestive
of, but not sufficient to infer, a causal relationship" between short-term ozone exposure and total
mortality.
IS-44
-------
Table IS-8 Summary of evidence from epidemiologic studies on the association
of short-term ozone exposure with mortality.
Conclusions from 2013 Ozone ISA
Results and Conclusions from 2020 ISAa
Mortality
Consistent, positive associations were
reported across multicity and
multicontinent studies in combination with
strong evidence from studies of
respiratory morbidity. There was evidence
from a limited number of studies of
cardiovascular morbidity, providing
coherence and biological plausibility.
Evidence demonstrated that there was
a likely to be causal relationship
between short-term ozone exposure
and mortality.
Recent multicity studies continue to provide
evidence of consistent, positive associations,
which is supported by strong evidence from
studies of respiratory morbidity, providing
coherence and biological plausibility. Recent
studies of cardiovascular morbidity do not
provide coherence between experimental and
epidemiologic studies, and therefore, biological
plausibility for cardiovascular mortality is
limited. Evidence is suggestive of, but not
sufficient to infer, a causal relationship
between short-term ozone exposure and
mortality.
Epidemiologic evidence Multicity and multicontinent studies
provided evidence of consistent positive
associations for total (nonaccidental),
respiratory, and cardiovascular mortality.
Recent multicity studies continue to provide
evidence of consistent, positive associations
with total (nonaccidental), respiratory, and
cardiovascular mortality, but the cause-specific
mortality evidence is limited to one recent
multicity study.
Copollutant
confounding
Ozone-mortality associations remained
positive and relatively unchanged in
copollutant models with PM and PM2.5
components, but analyses of PM2.5
components are limited by the every-3rd
and 6th-day sampling schedule.
Recent multicity studies have conducted a
limited assessment of potential copollutant
confounding, but report that ozone-mortality
associations remain positive and relatively
unchanged in copollutant models with PM10
and NO2, the only pollutants assessed.
Biological plausibility
The strong and consistent evidence
within and across scientific disciplines for
respiratory morbidity provided coherence
and biological plausibility for respiratory
mortality. For cardiovascular mortality,
controlled human exposure and animal
toxicological studies provided initial
evidence supporting a biologically
plausible mechanism by which short-term
ozone exposure could lead to
cardiovascular mortality, but there was
inconsistency in results between
experimental and epidemiologic studies
of cardiovascular morbidity.
There continues to be strong and consistent
evidence within and across disciplines for
respiratory morbidity, which provides
coherence and biological plausibility for
respiratory mortality. Although there remains
evidence of cardiovascular mortality, recent
controlled human exposure studies do not
report evidence of cardiovascular effects in
response to short-term ozone exposure and
epidemiologic studies provide limited evidence
of associations with more sever cardiovascular
effects, such as emergency department visits
and hospital admissions. Collectively, there is
a lack of coherence between experimental and
epidemiologic studies providing limited
evidence of a biologically plausible pathway to
cardiovascular mortality or to other causes of
mortality.
Conclusions from the 2020 ISA include evidence from recent studies integrated with evidence included in previous Ozone ISAs
and AQCDs.
IS-45
-------
IS.4.3.6 Other Health Endpoints
The evidence for the other health endpoints not discussed in previous sections, including
long-term ozone exposure and cardiovascular and metabolic effects and mortality, and short- and
long-term ozone exposure and reproductive effects, nervous system effects, and cancer, is limited or
inconsistent, resulting in causality determinations of either "suggestive of, but not sufficient to infer, a
causal relationship" or "inadequate to infer the presence or absence of a causal relationship." The
evidence for these health effects is summarized here, with more details of the evidence that formed the
basis for these conclusions in Appendix 4. Appendix 5. Appendix 6. and Appendix 7.
15.4.3.6.1 Long-Term Ozone Exposure and Cardiovascular Effects
Collectively, the body of evidence for long-term ozone exposure and cardiovascular effects is
"suggestive of, but not sufficient to infer, a causal relationship." Recent animal toxicological and
epidemiologic studies add to the body of evidence that formed the basis of the conclusions in the 2013
Ozone ISA for cardiovascular health effects. This body of evidence is limited, however, with some
experimental and observational evidence for subclinical cardiovascular health effects and little evidence
for associations with outcomes such as IHD or MI, HF, or stroke. The strongest evidence for the
association between long-term ozone exposure and cardiovascular health outcomes continues to come
from animal toxicological studies of impaired cardiac contractility and epidemiologic studies of blood
pressure changes and hypertension and cardiovascular mortality. Recent epidemiologic studies observed
positive associations with changes in blood pressure or hypertension, but animal toxicological studies do
not report effects of ozone on blood pressure changes. In conclusion, the results observed across both
recent and older experimental and observational studies conducted in various locations provide limited
evidence for an association between long-term ozone exposure and cardiovascular health effects.
15.4.3.6.2 Long-Term Exposure and Metabolic Effects
In the 2013 Ozone ISA, evidence was insufficient to evaluate metabolic effects as a separate
health effect category. Therefore, no causality determinations for metabolic effects were made in that
document (U.S. EPA. 2013b). Since then, the epidemiologic and experimental literature investigating
long-term ozone exposure and outcomes related to metabolic effects has expanded substantially. Positive
associations between long-term exposure to ozone and diabetes-related mortality were observed in recent
evaluations of well-established cohorts in the U.S. and Canada. The mortality results are supported by
epidemiologic and experimental studies reporting effects on glucose homeostasis and serum lipids, as
well as other indicators of metabolic function (e.g., peripheral inflammation and neuroendocrine stress
response). Findings from an epidemiologic study of metabolic disease demonstrate increases in the
clinical diagnosis of metabolic syndrome. Additionally, in prospective cohort studies in the U.S. and
IS-46
-------
Europe, increased incidence of type 2 diabetes is observed in association with long-term ozone exposure.
Despite an increased number of studies, many uncertainties remain regarding the metabolic effects related
to long-term ozone exposure. Most studies from the epidemiologic literature did not evaluate potential
copollutant confounding. There were a very limited number of studies available for review from the
animal toxicological literature; these studies had few overlapping endpoints, and furthermore, they were
primarily conducted by the same set of authors. Overall, considering the positive epidemiologic studies
and limited support from animal toxicological studies, the collective evidence is "suggestive of, but not
sufficient to infer, a causal relationship" between short-term exposure to ozone and metabolic
effects.
IS.4.3.6.3 Ozone Exposure and Reproductive Effects
Overall, the evidence is "suggestive of, but not sufficient to infer, a causal relationship"
between ozone exposure and (1) male and female reproduction and fertility and (2) pregnancy and
birth outcomes. Separate conclusions are made for these groups of reproductive effects because they are
likely to have different etiologies and critical exposure windows over different lifestages. The 2013
Ozone ISA concluded that the evidence was "suggestive of a causal relationship"1 between ozone
exposure and the inclusive category for all reproductive and developmental outcomes.
The strongest evidence in the 2013 Ozone ISA for effects on reproduction and fertility came from
epidemiologic and animal toxicological studies of sperm. Recent studies of sperm quality are consistent
with this evidence but remain limited. Uncertainties that contribute to the determination include a lack of
evaluation of copollutant confounding or multiple potential sensitive windows of exposure, and the
generally small sample size of studies in human subjects.
The strongest evidence in the 2013 Ozone ISA for effects on pregnancy and reproduction came
from epidemiologic studies of birth weight. Recent studies of birth weight are consistent with this
evidence but remain limited. There are several well-designed, well-conducted studies that indicate an
association between ozone and poorer birth outcomes, particularly for outcomes of continuous birth
weight and preterm birth. In particular, studies of preterm birth that examine exposures in the first and
second trimesters show fairly consistent positive associations (increased ozone exposures associated with
increased odds of preterm birth). In addition, some animal toxicological studies demonstrate decreased
birth weight and changes in uterine blood flow. Epidemiologic studies of continuous birth weight and
preterm birth did not generally adjust for potential copollutant confounding, although studies that did
appeared to show limited impacts. There is also inconsistency across exposure windows for associations
with continuous birth weight. Also, the magnitude of effect estimates varies.
1 Since the 2013 Ozone ISA, the causality determination language has been updated and this category is now stated
as suggestive of, but not sufficient to infer, a causal relationship.
IS-47
-------
IS.4.3.6.4
Short-Term Ozone Exposure and Nervous System Effects
Overall, the evidence is "suggestive of, but not sufficient to infer, a causal relationship"
between short-term exposure to ozone and nervous system effects. The 2013 Ozone ISA concluded
that the evidence was "suggestive of a causal relationship"1 between short-term ozone exposure and
nervous system effects. The strongest evidence supporting this causality determination came from
experimental animal studies of CNS structure and function. Most of the recent experimental animal
studies demonstrate that short-term exposure to ozone induces oxidative stress and inflammation in the
central nervous system (Section 7.2.1.3). In some cases, these effects are associated with changes in brain
morphology and effects on neurotransmitters. In some instances, the effects of short-term ozone exposure
on the nervous system were exacerbated in aged animals. No epidemiologic studies of short-term ozone
exposure and nervous system effects were reviewed in the 2013 Ozone ISA, and the epidemiologic
evidence remains limited. Recent epidemiologic evidence consists only of a study reporting an association
between short-term ozone exposure and depressive symptoms, and several studies of hospital admissions
or ED visits for symptoms related to a range of nervous system diseases or mental disorders
(e.g., multiple sclerosis, Alzheimer's disease, Parkinson's disease, depression, psychiatric disorders).
These findings for depressive symptoms are coherent with experimental animal studies showing
depression-like behaviors in rodents. Biological plausibility of these effects is supported by multiple
toxicological studies in laboratory animals showing inflammation and morphological changes in the brain
following short-term ozone exposure (Section 7.2.1.2).
IS.4.3.6.5 Long-Term Ozone Exposure and Nervous System Effects
Overall, the evidence is "suggestive of, but not sufficient to infer, a causal relationship"
between long-term ozone exposure and nervous system effects. This conclusion is consistent with that
of the 2013 Ozone ISA. The strongest evidence supporting the causality determination for long-term
ozone exposure and nervous system effects from the 2013 Ozone ISA came from animal toxicological
studies demonstrating effects on CNS structure and function, with several studies indicating the potential
for neurodegenerative effects similar to Alzheimer's or Parkinson's diseases in a rat model. The body of
evidence has grown since the 2013 Ozone ISA. Recent epidemiologic studies have examined nervous
system effects, including cognitive effects, depression, neurodegenerative disease, and autism. Although
the epidemiologic evidence remains limited, the strongest evidence is for effects on cognition in adults.
Recent experimental animal studies continue to provide coherence for these effects. Several recent animal
toxicological studies report increased markers of oxidative stress and inflammation, including lipid
peroxidation, microglial activation, and cell death following long-term exposure to ozone. There was
some evidence to support that aged and young populations may have increased sensitivity to ozone
exposure. Uncertainties that contribute to the causality determination include the limited number of
epidemiologic studies, the lack of consistency across the available studies of Alzheimer's and Parkinson's
disease, and the limited evaluation of copollutant confounding in these studies. In addition, the evidence
IS-48
-------
supporting the biological plausibility of the associations with autism or ASD in epidemiologic studies is
limited.
15.4.3.6.6 Long-Term Ozone Exposure and Cancer
The evidence describing the relationship between exposure to ozone and cancer remains
"inadequate to infer the presence or absence of a causal relationship." In the 2013 Ozone ISA, very
few studies were available to assess the relationship between long-term ozone exposure and cancer. The
few available epidemiologic and animal toxicological studies indicated that ozone exposure may
contribute to DNA damage. However, given the overall lack of studies, the 2013 Ozone ISA concluded
that the evidence was inadequate to determine whether a causal relationship existed between long-term
ozone exposure and cancer. More recent studies provide some additional animal toxicological evidence of
DNA damage. In addition, several, but not all, recent cohort and case-control studies have observed
positive associations between long-term ozone exposure and lung cancer incidence or mortality. Several
of the studies evaluating lung cancer mortality were conducted in populations that had already been
diagnosed with cancer in a different organ system. Associations between ozone exposure and other types
of cancer were generally null. Given the limited evidence base, the lack of an evaluation of copollutant
confounding in epidemiologic studies reporting associations, and the evaluation of study populations that
had already been diagnosed with cancer in several of the epidemiologic studies, the evidence is not
sufficient to draw a conclusion regarding causality.
15.4.3.6.7 Long-Term Ozone Exposure and Mortality
Collectively, this body of evidence is "suggestive of, but not sufficient to infer, a causal
relationship" between long-term ozone exposure and total mortality. Recent epidemiologic studies
add to the limited body of evidence that formed the basis of the conclusions of in 2013 Ozone ISA for
total mortality. This body of evidence is generally inconsistent, with some U.S. and Canadian cohorts
reporting modest positive associations between long-term ozone exposure and total mortality, while other
recent studies conducted in the U.S, Europe, and Asia reporting null or negative associations. The
strongest evidence for the association between long-term ozone exposure and total (nonaccidental)
mortality continues to come from analyses of patients with pre-existing disease from the Medicare cohort
and from recent evidence demonstrating positive associations with cardiovascular mortality. The evidence
from the assessment of ozone-related respiratory disease, with more limited evidence from cardiovascular
and metabolic morbidity, provides some biological plausibility for mortality due to long-term ozone
exposures. In conclusion, the inconsistent associations observed across both recent and older cohort and
cross-sectional studies conducted in various locations provide limited evidence for an association between
long-term ozone exposure and mortality.
IS-49
-------
1S.4.4
At-Risk Populations
Interindividual variation in exposure to or human responses to ambient air pollution exposure can
result in some groups or lifestages being at increased risk for health effects. The NAAQS are intended to
protect public health with an adequate margin of safety. In so doing, protection is provided for both the
population as a whole and those potentially at increased risk for health effects in response to exposure to a
criteria air pollutant [e.g., ozone; see Preamble to the ISAs (U.S. EPA. 2015)1. There is interindividual
variation in both physiological responses, and exposure to ambient air pollution. The scientific literature
has used a variety of terms to identify factors and subsequently populations or lifestages that may be at
increased risk of an air pollutant-related health effect, including susceptible, vulnerable, sensitive, at risk,
and response-modifying factor rVinikoor-Imler et al. (2014); see Preamble to the ISAs (U.S. EPA. 2015)1.
Acknowledging the inconsistency in definitions for these terms across the scientific literature and the lack
of a consensus on terminology in the scientific community, "at-risk" is the all-encompassing term used in
ISAs for groups with specific factors that increase the risk of an air pollutant (e.g., ozone)-related health
effect in a population, as initially detailed in the 2013 Ozone ISA (U.S. EPA. 2013b). Therefore, this ISA
takes an inclusive and all-encompassing approach and focuses on identifying those populations or
lifestages potentially "at risk" of an ozone-related health effect.
As discussed in the Preamble to the ISAs (U.S. EPA. 2015). the risk of health effects from
exposure to ozone may be modified as a result of intrinsic (e.g., pre-existing disease, genetic factors) or
extrinsic factors (e.g., sociodemographic or behavioral factors), differences in internal dose (e.g., due to
variability in ventilation rates or exercise behaviors), or differences in exposure to air pollutant
concentrations (e.g., more time spent in areas with higher ambient concentrations). Some factors may lead
to a reduction in risk and are recognized as such during the evaluation. However, in order to inform
decisions on the NAAQS, this ISA focuses on identifying those populations or lifestages at greater risk.
While a combination of factors (e.g., residential location and socioeconomic status [SES]) may increase
the risk of ozone-related health effects in portions of the population, information on the interaction among
factors remains limited. Thus, this ISA characterizes the individual factors that potentially result in
increased risk for ozone-related health effects [see Preamble to the ISAs (U.S. EPA. 2015)1.
IS.4.4,1 Approach to Evaluating and Characterizing the Evidence for At-Risk Factors
The ISA takes a pragmatic approach to identifying and evaluating factors that may increase the
risk of a population or specific lifestage to an ambient air ozone-related health effect; this approach is
described in detail in the Preamble to the ISAs (U.S. EPA. 2015) and illustrated in Table IS-9. While
Appendix 3-Appendix 7 include a discussion of some populations and lifestages in order to explicitly
characterize the causal nature between ozone exposure and health effects based on the body of evidence
(e.g., children, individuals with asthma), this section focuses on summarizing evidence that can inform
the identification of such populations and lifestages. Those populations and lifestages explicitly
IS-50
-------
considered in this ISA include those with pre-existing asthma, children, older adults, and outdoor
workers, for which there was adequate evidence of increased risk in the 2013 Ozone ISA.
The evidence evaluated in this section includes relevant studies discussed in
Appendix 3-Appendix 7 of this ISA and builds on the evidence presented in the 2013 Ozone ISA (U.S.
EPA. 2013b). Based on the approach developed in previous ISAs (U.S. EPA. 2016. 2013a. b), recent
evidence is integrated across scientific disciplines and health effects, and where available, with
information on exposure and dosimetry. In evaluating factors and population groups, greater emphasis is
placed on the evidence for those health outcomes for which a "causal" or "likely to be causal"
relationship is concluded in Appendix 3-Appendix 7 of this ISA.
Table IS-9 Characterization of evidence for factors potentially increasing the
risk for ozone-related health effects.
Classification Health Effects
Adequate There is substantial, consistent evidence within a discipline to conclude that a factor results in a
evidence population or lifestage being at increased risk of air pollutant-related health effect(s) relative to
some reference population or lifestage. Where applicable, this evidence includes coherence
across disciplines. Evidence includes multiple high-quality studies.
The collective evidence suggests that a factor results in a population or lifestage being at
increased risk of air pollutant-related health effect(s) relative to some reference population or
lifestage, but the evidence is limited due to some inconsistency within a discipline or, where
applicable, a lack of coherence across disciplines.
The collective evidence is inadequate to determine whether a factor results in a population or
lifestage being at increased risk of air pollutant-related health effect(s) relative to some reference
population or lifestage. The available studies are of insufficient quantity, quality, consistency,
and/or statistical power to permit a conclusion to be drawn.
Evidence of no There is substantial, consistent evidence within a discipline to conclude that a factor does not
effect result in a population or lifestage being at increased risk of air pollutant-related health effect(s)
relative to some reference population or lifestage. Where applicable, the evidence includes
coherence across disciplines. Evidence includes multiple high-quality studies.
As discussed in the Preamble to the ISAs (U.S. EPA. 2015). consideration of at-risk populations
includes evidence from epidemiologic, controlled human exposure, and animal toxicological studies, in
addition to relevant exposure-related information. Regarding epidemiologic studies, the evaluation
focuses on those studies that include stratified analyses to compare populations or lifestages exposed to
similar air pollutant concentrations within the same study design along with consideration of the
strengths and limitations of each study. Other epidemiologic studies that do not stratify results but instead
examine a specific population or lifestage can provide supporting evidence for the pattern of associations
observed in studies that formally examine effect measure modification. Similar to the characterization of
Suggestive
evidence
Inadequate
evidence
IS-51
-------
evidence in Appendix 3-Appendix 7. the greatest emphasis is placed on patterns or trends in results
across studies. Experimental studies in human subjects or animal models that focus on factors, such as
genetic background or health status, are evaluated because they provide coherence and biological
plausibility of effects observed in epidemiologic studies. Also evaluated are studies examining whether
factors may result in differential exposure to ozone and subsequent increased risk of ozone-related health
effects. Conclusions are made with respect to whether a specific factor increases the risk of an
ozone-related health effect based on the characterization of evidence using the framework detailed in
Table III of the Preamble (U.S. EPA. 2015). and presented in Table IS-9.
IS.4.4.2 Summary of At-Risk Populations
The 2013 Ozone ISA (U.S. EPA. 2013b) concluded that there was adequate evidence to classify
individuals with pre-existing asthma, children and older adults, individuals with reduced intake of certain
nutrients (i.e., vitamins C and E), and outdoor workers as populations at increased risk to the health
effects of ozone. These conclusions were based on the consistency in findings across studies, as well as
on coherence of results from different scientific disciplines. Recent studies provide additional evidence
that individuals with pre-existing asthma and children are at increased risk of the effects of ozone. There
is relatively little recent evidence to add to the evidence presented in the 2013 Ozone ISA for older adults,
individuals with reduced intake of certain nutrients, and outdoor workers.
Recent, large multicity epidemiologic studies conducted in the U.S. expand upon evidence from
the 2013 Ozone ISA to provide further support the relationship between ozone and ED visits and hospital
admissions for asthma among individuals with pre-existing asthma (Table IS-10; Section IS.4.4.3. IV
Generally, studies comparing age groups also reported higher magnitude associations for
respiratory hospital admissions and ED visits for children (Section IS.4.4.4.1) than for adults. In addition,
recent evidence from studies of nonhuman primates and rodents demonstrate ozone-induced respiratory
effects and support the biological plausibility of associations observed in epidemiologic studies between
long-term exposure to ozone and the development of asthma in children. Specifically, these experimental
studies indicate that early-life ozone exposure can cause structural and functional changes that could
potentially contribute to airway obstruction and increased airway responsiveness. Also, children have
both higher exposure (due to increased time spent outdoors) and dose (due to their greater ventilation
rate). Childrens' respiratory systems are also still undergoing lung growth.
The majority of evidence for older adults being at increased risk of health effects related to ozone
exposure comes from studies of short-term ozone exposure and mortality evaluated in the 2013 Ozone
ISA (Section IS.4.4.4.2).
IS-52
-------
Table IS-10 Summary of evidence for populations at increased risk to the health
effects of ozone.
Conclusions from 2013 Ozone ISA Conclusions from 2020 ISA
Adequate evidence
Pre-existing
asthma
Collective evidence from controlled human
exposure studies is supported by animal
toxicological studies. Some, but not all,
epidemiologic studies report greater risk of
health effects among individuals with
asthma.
Evidence from controlled human exposure and
animal toxicological studies provide biological
plausibility for the associations observed in
epidemiologic studies of short-term ozone
exposure and asthma exacerbation. Results from
experimental studies in humans demonstrate that
ozone exposures lead to increased respiratory
symptoms, lung function decrements, increased
airway responsiveness, and increased lung
inflammation in individuals with asthma.
Children
Controlled human exposure and animal
toxicological studies provide evidence of
increased risk from ozone exposure for
younger ages, which is coherent with
findings from epidemiologic studies that
report larger associations for respiratory ED
visits and hospital admissions for children
than adults.
Recent, large multicity epidemiologic studies
conducted in the U.S. expand upon previous
evidence and support an association between
ozone and ED visits and hospital admissions for
asthma, which are strongest in children between
the ages of 5 and 18; animal toxicological studies
in infant monkeys and neonatal rats show that
early-life ozone exposure can cause structural
and functional changes that could potentially
contribute to airway obstruction and increased
airway responsiveness.
Older adults Epidemiologic studies report consistent
positive associations between short-term
ozone exposure and mortality in older
adults.
Controlled human exposure studies demonstrate
changes in FEVi and FVC among older adults at
a relatively light activity level and brief duration of
ozone exposure, though these responses are not
greater than in other age groups; evidence from
studies of metabolic effects is inconsistent.
Outdoor workers Strong evidence from 2006 Ozone AQCD, No recent information has been evaluated that
which demonstrated increased exposure, would inform or change prior conclusions,
dose, and ultimately risk of ozone-related
health effects in this population supports that
there is adequate evidence to indicate that
increased exposure to ozone through
outdoor work increases the risk of
ozone-related health effects.
Genetic factors Multiple genetic variants have been No recent information has been evaluated that
observed in epidemiologic and controlled would inform or change prior conclusions,
human exposure studies to affect the risk of
ozone-related respiratory outcomes and
support is provided by animal toxicological
studies of genetic factors.
Individuals with reduced intake of vitamins E No recent information has been evaluated that
and C are at risk for ozone-related would inform or change prior conclusions,
respiratory effects based on substantial,
consistent evidence both within and among
disciplines.
IS-53
-------
Table IS-10 (Continued): Summary of evidence for populations at increased risk
to the health effects of ozone.
Conclusions from 2013 Ozone ISA
Conclusions from 2020 ISA
Suggestive evidence
Sex
Evidence for increased risk for ozone-related
health effects present for females in some
studies and males in other studies; some
indication that females are increased risk of
ozone-related respiratory hospital
admissions and ED visits.
No recent information has been evaluated that
would inform or change prior conclusions.
Pre-existing
obesity
Multiple epidemiologic, controlled human
exposure, and animal toxicological studies
report increased ozone-related respiratory
health effects among obese individuals.
Recent animal toxicological studies expand upon
previous evidence and continue to indicate that,
compared to lean mice, obese mice exhibit
enhanced airway responsiveness and pulmonary
inflammation in response to acute ozone
exposures.
Most studies report that individuals with low
SES and those living in neighborhoods with
low SES are more at risk for ozone-related
respiratory hospital admissions and ED
visits; inconsistent results for mortality and
reproductive outcomes.
No recent information has been evaluated that
would inform or change prior conclusions.
Inadequate evidence
Race/ethnicity A small number of studies provide
inadequate evidence that there may be
race-related increase in risk of ozone-related
health effects for some outcomes.
No recent information has been evaluated that
would inform or change prior conclusions.
Pre-existing
COPD
Epidemiologic studies indicate that persons
with COPD may have increased risk of
ozone-related cardiovascular effects, but
little information is available on whether
COPD leads to an increased risk of
ozone-induced respiratory effects.
Small number of recent studies provided
inadequate evidence to determine whether COPD
results in an increased risk of ozone-related
health effects.
Pre-existing CVD
Most short-term exposure studies did not
report increased ozone-related
cardiovascular morbidity for individuals with
pre-existing CVD. Limited number of studies
examined whether CVD modifies the
association between ozone and respiratory
effects. Some evidence that CVD increases
risk of ozone-related total mortality.
Some studies provide evidence that
cardiovascular disease exacerbates the
respiratory effects of ozone exposure; a limited
number of recent epidemiologic cohort studies
observed increased risk estimates for incident
diabetes among those with pre-existing
hypertension or among subjects that had some
pre-existing condition (Ml, COPD, hypertension,
or hyperlipidemia) compared to those without
pre-existing disease.
Pre-existing
diabetes
There are a limited number of epidemiologic
studies and lack of controlled human
exposure studies or toxicological studies to
determine whether pre-existing diabetes
modifies ozone effects on health.
A small number of studies provide inadequate
evidence that individuals with pre-existing
metabolic disease may be at greater risk of
mortality associated with long-term ozone
exposure.
IS-54
-------
Table IS-10 (Continued): Summary of evidence for populations at increased risk
to the health effects of ozone.
Conclusions from 2013 Ozone ISA
Conclusions from 2020 ISA
Smoking
There are a limited number of studies and
insufficient coherence for differences in
ozone-related health effects by smoking
status.
No recent information has been evaluated that
would inform or change prior conclusions.
COPD = chronic obstructive pulmonary disease; CVD = cardiovascular disease; ED = emergency department; FENA = forced
expiratory volume in 1 second; FVC = forded vital capacity; Ml = myocardial infarction; SES = socioeconomic status.
IS.4.4.3 Pre-existing Disease
Individuals with some pre-existing diseases may be at greater risk of an air pollution-related
health effect because they may be in a compromised biological state that can vary depending on the
disease and severity. The 2013 Ozone ISA (U.S. EPA. 2013b) concluded that there was adequate
evidence that those with pre-existing respiratory disease, specifically asthma, were at greater risk for the
health effects associated with exposure to ozone, but that evidence was inadequate to determine whether
those with COPD, cardiovascular disease, or diabetes were at increased risk of ozone-related health
effects. Of the recent epidemiologic studies evaluating effect measure modification by pre-existing
disease or condition, most focused on asthma, COPD, or cardiovascular disease. Table IS-11 presents the
prevalence of these diseases according to the Centers for Disease Control and Prevention's (CDC's)
National Center for Health Statistics (Blackwcll et al.. 2014). including the proportion of adults with a
current diagnosis categorized by age and geographic region. The large proportions of the U.S. population
affected by many chronic diseases, including various respiratory and cardiovascular diseases, indicates
the potential public health impact, and thus, the importance of identifying populations that may be at
increased risk for ozone-related health effects.
IS-55
-------
Table IS-11 Prevalence of respiratory diseases, cardiovascular diseases,
diabetes, and obesity among adults by age and region in the U.S. in
2012.
Adults
(18+)
Age (%f
Region (%)b
Chronic
Disease/Condition
N (in
thousands)
18-44
45-64
65-74
75+
North-
east
Midwest
South
West
All (N, in
thousands)
234,921
111,034
82,038
23,760
18,089
42,760
53,378
85,578
53,205
Selected respiratory diseases
Asthma0
18,719
8.1
8.4
7.8
6.0
9.2
8.1
7.3
7.8
COPD—chronic
bronchitis
8,658
2.5
4.7
4.9
5.2
3.2
4.4
3.9
2.4
COPD—
emphysema
4,108
0.3
2.3
4.7
4.7
1.3
2.0
1.9
1.0
Selected cardiovascular diseases/conditions
All heart disease
26,561
3.8
12.1
24.4
36.9
10.0
11.6
11.6
9.3
Coronary heart
disease
15,281
0.9
7.1
16.2
25.8
5.3
6.5
7.0
5.1
Hypertension
59,830
8.3
33.7
52.3
59.2
21.4
24.1
26.6
21.5
Stroke
6,370
0.6
2.8
6.3
10.7
1.8
2.5
3.0
2.5
Metabolic disorders/conditions
Diabetes
21,391
2.4
12.7
21.1
19.8
7.6
8.4
10.0
7.3
Obesity (BMI
>30 kg/m2)
64,117
26
33.7
29.7
18
25.1
29.9
29.9
25.2
Overweight (BMI
25-30 kg/m2)
78,455
31.4
36.8
40.7
38.6
34.3
34.1
34.2
35.3
BMI = body mass index; COPD = chronic obstructive pulmonary disease.
Percentage of individual adults within each age group with disease, based on N (at the top of each age column).
Percentage of individual adults (18+) within each geographic region with disease, based on N (at the top of each region column).
°Asthma prevalence is reported for "still has asthma."
Source: Blackwell et al. (20141: National Center for Health Statistics: Data from Tables 1 -4, 7, 8, 28, and 29 of the Centers for
Disease Control and Prevention report.
IS-56
-------
IS.4.4.3.1
Pre-existing Asthma
Asthma is the leading chronic illness affecting children. Approximately 8% of adults and 9% of
children (age <18 years) in the U.S. currently have asthma (Blackwcll et al.. 2014; Bloom et al.. 2013).
Regarding consideration of those with asthma potentially being at increased risk for an ozone-related
health effect, it is important to note that individuals with asthma, and children in general, tend to have a
higher degree of oronasal breathing, which can result in greater penetration of ozone into the lower
respiratory tract.
The 2013 Ozone ISA concluded that there is adequate evidence that individuals with asthma are
at increased risk of health effects related to ozone exposure; this conclusion is based on a number of
controlled human exposure, epidemiologic, and animal toxicological studies. Consistent with this
evidence, recent, large multicity epidemiologic studies conducted in the U.S. expand upon evidence from
the 2013 Ozone ISA to provide further support for an association between ozone and ED visits and
hospital admissions for asthma. Hospital admission and ED visit studies that presented age-stratified
results reported the strongest associations in children between the ages of 5 and 18 years. Additionally,
associations were observed across a range of ambient ozone concentrations and were consistent in models
where exposure was assigned using either measured or modeled ozone concentrations. While there is a
lack of recent epidemiologic studies conducted in the U.S. or Canada that have examined respiratory
symptoms and medication use, lung function, and subclinical effects in people with asthma, a large body
of evidence from the 2013 Ozone ISA (U.S. EPA. 2013b) reported ozone associations with these less
severe indicators of asthma exacerbation that provide support for the ozone-related increases in asthma
hospital admissions and ED visits observed in recent studies.
Evidence from controlled human exposure and animal toxicological studies provide biological
plausibility for the associations observed in epidemiologic studies of short-term ozone exposure and
asthma exacerbation. Results from experimental studies in humans demonstrate that ozone exposures lead
to increased respiratory symptoms, decrements in lung function, increased airway responsiveness, and
increased lung inflammation in individuals with asthma. However, observed responses across the range of
endpoints did not generally differ due to the presence of asthma. Animal toxicological studies similarly
found that ozone exposures altered lung function measures, increased airway responsiveness, and
increased pulmonary inflammation and bronchoconstriction in allergic animals. In contrast to controlled
human exposure studies, there was some evidence from studies of rodents that the observed respiratory
effects were enhanced in allergic animals compared to naive animals.
Overall, recent evidence expands upon evidence available in the 2013 Ozone ISA and is adequate
to conclude that individuals with pre-existing asthma are at greater risk of ozone-related health effects
based on the substantial and consistent evidence within epidemiologic studies and the coherence with
toxicological studies.
IS-57
-------
IS.4.4.3.2
Pre-existing Chronic Obstructive Pulmonary Disease
Chronic obstructive pulmonary disease (COPD) comprises chronic bronchitis and emphysema
and affects approximately 8.6 million adults in the U.S. (Table IS-11). In the U.S., over 4% of adults
report having chronic bronchitis and almost 2% report having emphysema (Pleis et al.. 2009V Chronic
lower respiratory disease, including COPD, was ranked as the third leading cause of death in the U.S. in
2011 (Hoyert and Xu. 2012). Given that people with COPD have compromised respiratory function and
underlying respiratory tract inflammation, it is plausible that they could be at increased risk for an array of
ozone-related health effects.
Epidemiologic studies evaluated in the 2013 Ozone ISA indicate that individuals with COPD may
have increased risk of ozone-related cardiovascular effects, but little information was available on
whether COPD leads to an increased risk of ozone-induced respiratory effects. A limited number of recent
epidemiologic studies provide inconsistent evidence that individuals with pre-existing COPD could be at
greater risk for respiratory health effects associations with ozone exposure. Overall, a limited number of
recent studies add to the scarce evidence available in the 2013 Ozone ISA and, collectively, is inadequate
to conclude whether or not individuals with pre-existing COPD are at greater risk of ozone-related health
effects.
IS.4.4.3.3 Pre-existing Obesity
Obesity, defined as a BMI of 30 kg/m2 or greater, is an issue of increasing importance in the U.S.,
with self-reported obesity at 39.8% of the general population in 2016, up from 26.7% in 2009 (Hales et
al.. 2017V BMI may affect ozone-related health effects through multiple avenues, including systematic
inflammation, increased pre-existing disease, and poor diet. Increased risk of air pollution-related health
effects has been observed among obese individuals compared with nonobese individuals (U.S. EPA.
2009). The 2013 Ozone ISA concluded that there was suggestive evidence for increased ozone-related
respiratory health effects among obese individuals. This conclusion was based on evidence from
controlled human exposure studies and epidemiologic studies reporting greater lung function decrements
in obese compared with nonobese individuals, as well as enhanced pulmonary inflammation in genetically
and dietarily obese mice (U.S. EPA. 2013b).
Recent animal toxicological studies expand the body of evidence evaluated in the 2013 Ozone
ISA and continue to indicate that, compared with lean mice, obese mice exhibit enhanced airway
responsiveness and pulmonary inflammation in response to acute ozone exposures. In contrast, a recent
controlled human exposure study reported evidence of ozone-related increases in pulmonary
inflammation in both obese and normal-weight adult women during exercise, but inflammatory responses
did not differ between the groups. Overall, recent studies contribute some additional support to the
evidence available in the 2013 Ozone ISA and there is suggestive evidence indicating that individuals
with pre-existing obesity are at potentially increased risk of ozone-related health effects based on the
IS-58
-------
limited evidence within epidemiologic studies and some coherence from controlled human exposure and
animal toxicological studies.
15.4.4.3.4 Pre-existing Metabolic Syndrome
Metabolic syndrome is a clinical diagnosis used to describe a collection of risk factors that
include high blood pressure, dyslipidemia (elevated triglycerides and low levels of high-density
lipoprotein [HDL] cholesterol), obesity (particularly central obesity), and increased fasting blood glucose
(Albcrti et al.. 2009). The presence of these risk factors may predispose an individual to an increased risk
of type 2 diabetes and cardiovascular disease. In the 2013 Ozone ISA, a limited number of epidemiologic
studies provided inadequate evidence to indicate whether individuals with metabolic syndrome (generally
indicated by a diabetes diagnosis) were at an increased risk of ozone-related health effects compared with
those without diabetes.
In recent studies of a diabetes-prone mouse model, subacute ozone exposure increased airway
inflammation and proinflammatory genes in lung tissue (Section 3.1.6.2). In contrast, an epidemiologic
panel study observed a negative association between increased ozone exposure and pulmonary
inflammation in adults with type 2 diabetes mellitus. This inverse association may be explained by
negative correlations with copollutants that demonstrated strong positive associations with pulmonary
inflammation in the same population. Overall, a limited number of recent studies add to the small body of
evidence available in the 2013 Ozone ISA and, collectively, the evidence is inadequate to conclude that
individuals with pre-existing metabolic disease are at greater risk of ozone-related health effects.
15.4.4.3.5 Pre-existing Cardiovascular Disease
Cardiovascular disease has become increasingly prevalent in the U.S., with about 12% of adults
aged 45-64 years reporting a diagnosis of heart disease (Table IS-11). This number doubles to 24%
among adults aged 65-74 years and is even higher for adults aged 75 years and older. A high prevalence
of other cardiovascular-related conditions has also been observed, such as hypertension which is prevalent
among more than 50% of older adults. In the 2013 Ozone ISA, most epidemiologic studies evaluating
short-term ozone exposure did not report increased risk of cardiovascular morbidity for individuals with
pre-existing cardiovascular disease. There was some evidence from a limited number of epidemiologic
studies that those with pre-existing cardiovascular disease were at greater risk of ozone-related mortality
compared with those without pre-existing cardiovascular disease. Overall, the 2013 Ozone ISA concluded
that the evidence was inadequate to classify pre-existing cardiovascular disease as a potential at-risk
factor for ozone-related health effects.
Several recent studies evaluated respiratory effects of acute ozone exposure (0.2-1 ppm,
3-6 hours) in rodents with cardiovascular disease. Some of the studies provide evidence that
IS-59
-------
cardiovascular disease exacerbates the respiratory effects of ozone exposure. Injury, inflammation, and
oxidative stress measured in the respiratory system, lung function changes, and increased airway
responsiveness were documented in animals with cardiovascular disease in response to ozone exposure.
Acute ozone exposure in animal models of hypertension resulted in enhanced injury and inflammation
measured in the respiratory system, and airway responsiveness compared with healthy animals. A limited
number of recent epidemiologic cohort studies evaluated the potential for pre-existing cardiovascular
disease to modify associations between long-term ozone exposure and metabolic effects. These studies
observed increased risk estimates for incident diabetes among those with pre-existing hypertension or
among subjects that had some pre-existing condition (MI, COPD, hypertension, or hyperlipidemia)
compared with those without pre-existing disease. Overall, a limited number of recent studies add to the
evidence available in the 2013 Ozone ISA and, collectively, are inadequate to conclude whether
individuals with pre-existing metabolic disease are at greater risk of ozone-related health effects.
IS.4.4.4 Lifestage
Lifestage refers to a distinguishable time frame in an individual's life characterized by unique and
relatively stable behavioral and/or physiological characteristics that are associated with development and
growth (U.S. EPA. 2014). Differential health effects of ozone across lifestages could be due to several
factors. With regard to children, the human respiratory system is not fully developed until 18-20 years of
age; therefore, it is biologically plausible for children to have increased intrinsic risk for respiratory
effects if exposures are sufficient to contribute to potential perturbations in normal lung development.
Moreover, children in general may experience higher exposure to ozone than adults based on more time
spent outdoors while exercising during afternoon hours when ozone concentrations may be highest. The
ventilation rates also vary between children and adults, particularly during moderate/heavy activity.
Children have higher ventilation rates relative to their lung volume, which tends to increase the dose
normalized to lung surface area. Older adults, typically considered those 65 years of age or greater, have
weakened immune function, impaired healing, decrements in pulmonary and cardiovascular function, and
greater prevalence of chronic disease ITable IS-11; Blackwell et al. (2014)1. which may contribute to, or
worsen, health effects related to ozone exposure. Also, exposure or internal dose of ozone may differ
across lifestages due to varying ventilation rates, increased oronasal breathing at rest, and time-activity
patterns.
For decades, children, especially those with asthma, and older adults have been identified as
populations at increased risk of health effects related to ozone exposure (U.S. EPA. 2013b. 2006a.
1996a). Long-standing evidence from controlled human exposure studies demonstrated that children have
greater spirometric responses to ozone compared with middle-aged or older adults (U.S. EPA. 1996a). In
addition, epidemiologic studies reported larger associations for respiratory hospital admissions and ED
visits for children than for adults, and animal toxicological studies demonstrated ozone-induced health
effects in immature animals, including infant monkeys (U.S. EPA. 2013b). Compared with other age
IS-60
-------
groups, there was evidence for an increased risk of mortality associated with ozone exposure among older
adults (U.S. EPA. 2013b. 2006a'). The 2013 Ozone ISA concluded that there was adequate evidence that
children and older adults are at increased risk of ozone-related health effects.
15.4.4.4.1 Children
Recent, large multicity epidemiologic studies conducted in the U.S. expand on evidence from the
2013 Ozone ISA and provide further support for an association between short-term ozone exposure and
ED visits and hospital admissions for asthma. Hospital admission and ED visit studies that presented
age-stratified results reported the strongest associations in children between the ages of 5 and 18 years.
The evidence relating new-onset asthma to long-term ozone exposure is supported by toxicological
studies in infant monkeys, which indicate that postnatal ozone exposures can lead to the development of
asthma. This nonhuman primate evidence of ozone-induced respiratory effects supported the biological
plausibility of associations between long-term exposure to ozone and the development of asthma in
children observed in epidemiologic studies. Specifically, these experimental studies indicate that
early-life ozone exposure can cause structural and functional changes that could potentially contribute to
airway obstruction and increased airway responsiveness.
Overall, recent evidence expands upon evidence available in the 2013 Ozone ISA and is adequate
to conclude that children are at greater risk of ozone-related health effects based on the substantial and
consistent evidence within epidemiologic studies and the coherence with animal toxicological studies.
15.4.4.4.2 Older Adults
Collectively, the majority of evidence for older adults being at increased risk of health effects
related to ozone exposure comes from studies of short-term ozone exposure and mortality. Many of these
were evaluated in the 2013 Ozone ISA. As reported in the 1996 and 2006 Ozone AQCDs (U.S. EPA.
2006a. 1996a). decrements in lung function and increases in respiratory symptoms in response to ozone
exposure decreased with increasing age. However, whether inflammatory responses persisted with
increasing age remained unstudied at the time of the 2013 Ozone ISA (U.S. EPA. 2013b). Two recent
controlled human exposure studies demonstrate inflammatory responses in older adults, but it is not
possible to quantify inflammatory response as a function of age because of differences in experimental
protocols (i.e., duration of exposure to ozone, ozone concentration, activity level, and post-exposure time
of sputum collection). A recent controlled human exposure study also demonstrates changes in FEVi and
FVC among adults aged 55-70 years at a relatively light activity level and brief duration of exposure, but
a statistically significant interaction with age was not observed. This is generally consistent with studies
evaluated in previous assessments that showed ozone-associated lung function decrements declining with
age, but still being present in adults 50-60 years of age. This recent study was conducted at a lower ozone
delivery rate, which is more representative of that likely to occur in the ambient environment and shows
IS-61
-------
small lung function decrements occurring in groups of older adults ranging up to 70 years of age. These
recent studies demonstrate that inflammatory responses and lung function changes following ozone
exposure can occur in older adults, but do not indicate greater responses in older adults than other age
groups.
Overall, recent studies add little to the evidence available in the 2013 Ozone ISA. This evidence
is adequate to conclude that older adults are at greater risk of ozone-related health effects.
IS,5 Evaluation of Welfare Effects of Ozone
The scientific evidence for welfare effects of ozone is largely for effects on vegetation and
ecosystems and effects on climate. Appendix 8 presents the most policy-relevant information related to
this review of the NAAQS for ecological effects of ozone. Appendix 9 presents the most policy-relevant
information related to this review of the NAAQS for effects on climate. The framework for causal
determinations [see Preamble (U.S. EPA. 2015)1 has been applied to the body of scientific evidence to
examine effects attributed to ozone exposure. Conclusions from the 2013 Ozone ISA and key findings
that inform the current causality determinations for welfare effects of ozone are summarized in
Table IS-12.
Table IS-12 Summary of evidence for welfare effects of ozone.
Endpoint
Conclusions from 2013 Ozone ISA
Conclusions from 2020 ISA3
Visible foliar injury
Section 8.2
Causal relationship
Visible foliar injury from ozone exposure was
well characterized and documented over
several decades of research prior to the 2013
Ozone ISA on sensitive tree, shrub,
herbaceous, and crop species in the U.S. Some
sensitive species that show visible injury
identified in field surveys are verified in
controlled exposure settings. Ozone
concentrations are high enough to induce
visible symptoms in sensitive vegetation.
Causal relationship
Studies published since the 2013 Ozone ISA
strengthen previous conclusions that there is
strong evidence that ozone causes foliar
injury in a variety of plant species. The use of
bioindicators to detect phytotoxic levels of
ozone is a longstanding and effective
methodology and is supported by more
information on sensitive species.
IS-62
-------
Table IS-12 (Continued): Summary of evidence for welfare effects of ozone.
Endpoint
Conclusions from 2013 Ozone ISA
Conclusions from 2020 ISA3
Reduced
vegetation growth
Section 8.3
Causal relationship
Studies added to the evidence from the 2006
AQCD and earlier assessments and indicated
that ozone reduced growth of vegetation.
Studies from the Aspen FACE experiment
showed reduction in total biomass in aspen,
paper birch, and sugar maple, findings which
were overall consistent with OTC studies in
previous NAAQS reviews. Meta-analysis
showed ambient ozone concentrations (approx.
40 ppb avg across all hours of exposure)
decreased annual total biomass growth of forest
species by an avg of 7% with potentially greater
exposures with elevated ozone. Studies also
demonstrated that ozone alters biomass
allocation, generally reducing C allocated to
roots.
Causal relationship
New evidence from controlled exposure
experiments and illustration of potential
impacts using models built with empirical data
strengthen previous conclusions that ozone
reduces plant growth and biomass. Additional
studies find that ozone significantly changes
patterns of carbon allocation below and
aboveground.
Reduced plant No separate causality determination;
reproduction included with plant growth
Section 8.4 Evidence from studies that ozone alters
reproduction in herbaceous and woody plant
species adds to evidence from the 2006 AQCD
(primarily in herbaceous and crop species) for
ozone effects on metrics of plant reproduction.
Causal relationship
A new meta-analysis published since the
2013 Ozone ISA provides strong and
consistent evidence for negative effects of
ozone on plant reproduction. For all exposure
categories evaluated, including the lowest
exposure category of <40 ppb, between one
and eight metrics of reproduction significantly
decreased. In addition, more evidence is
available that plant reproductive tissues are
directly affected by ozone exposure.
Increased tree Causality not assessed
mortality Evidence built on observations from the 2006
Section 8.4.3 Ozone AQCD of decline of conifer forests over
time observed in several regions affected by
elevated ozone along with other factors (Valley
of Mexico, southern France, Carpathian
Mountains). At the Aspen FACE site, there was
reduced growth and increased mortality of a
sensitive aspen clone.
Likely to be causal relationship
In a new large-scale multivariate analysis
evaluating tree mortality over a 15-year
period ozone significantly increased tree
mortality in 7 out of 10 plant functional types
in the eastern and central U.S. An Aspen
FACE study shows that sensitive aspen
genotypes have increased mortality
compared to tolerant genotypes.
Reduced yield and
quality of
agricultural crops
Section 8.5
Causal relationship
Detrimental effects of ozone on crop production
were recognized since the 1960s. There are
well-documented yield losses in a variety of
agricultural crops with increasing ozone
concentration. Ozone also decreased crop
quality. Modeling studies at large geographic
scales showed ozone generally reduced crop
yield, but effects vary across regions and
species.
Causal relationship
Greenhouse, OTC, FACE, and modeling
studies published since the 2013 Ozone ISA
strengthen previous conclusions that ozone
reduces yield in major U.S. crops including
wheat, soybean, and other non-soy legumes.
Advances in characterization of ozone effects
on U.S. crop yield include further geographic
and temporal refinement of ozone sensitivity.
For soybean, there are updated
exposure-response curves.
IS-63
-------
Table IS-12 (Continued): Summary of evidence for welfare effects of ozone.
Endpoint
Conclusions from 2013 Ozone ISA
Conclusions from 2020 ISA3
Altered herbivore
growth and
reproduction
Section 8.6
Causality not assessed
A meta-analysis of 16 studies found that
elevated ozone decreased development time
and increased pupal mass in insect herbivores.
Other field and laboratory studies reported
species-level and community-level responses in
insects yet the directionality of response to
ozone was mixed. This is congruent with
findings from the 2006 AQCD and 1996 AQCD,
where statistically significant effects on
herbivorous insects were observed, but did not
provide any consistent pattern of response
across growth, reproduction, and mortality
endpoints.
Likely to be causal relationship
There is a large body of evidence showing
altered growth and reproduction in insect
herbivores. More research has since been
published on a range of species and at
varying levels of ozone exposure although
there is no clear trend in the directionality of
response for most metrics. The most
commonly measured responses are
fecundity, development time, and growth.
Alteration of
plant-insect
signaling
Section 8.7
Causality not assessed
A few experimental and modeling studies
reported altered chemical signaling in
insect-plant interactions due to ozone exposure.
The effect of ozone on chemical signaling is an
emerging area of study that may result in further
elucidation of effects with more empirical data.
Likely to be causal relationship
Laboratory, greenhouse, OTC, and Finnish
FACE experiments expand the evidence for
altered/degraded emissions of chemical
signals from plants and reduced detection of
volatile plant signaling compounds by insects,
including pollinators, in the presence of
ozone. Affected plant-insect interactions
include plant defense against herbivory and
insect attraction to plants. New evidence
includes consistent effects in multiple insect
species.
Reduced
productivity in
terrestrial
ecosystems
Section 8.8.1
Causal relationship
Studies from long-term FACE experiments
provided evidence of the association of ozone
exposure and reduced productivity at the
ecosystem scale. Results across different
ecosystem models were consistent with the
FACE experimental evidence. Models
consistently found that ozone exposure
negatively impacted indicators of ecosystem
productivity. Studies at the leaf and plant scales
show that ozone decreased photosynthesis and
plant growth, providing coherence and
plausibility for reported decreases in ecosystem
productivity. Magnitude of response varied
among plant communities.
Causal relationship
Modeling studies and controlled exposure
experiments (including Aspen FACE),
published since the 2013 Ozone ISA
strengthen previous conclusions. Much of the
research is confirmatory, with some work
providing new mechanistic insight into the
effects of ozone on productivity and creating
a more nuanced understanding of how these
effects vary among species, communities,
and environmental conditions.
Reduced carbon
sequestration in
terrestrial
ecosystems
Section 8.8
Likely to be causal relationship
Studies add to the strong and consistent
evidence in the 2006 AQCD that ozone
decreases plant photosynthesis. Most
assessments of the effects of ozone on
terrestrial C are from model simulations.
Likely to be causal relationship
Several new model simulations strengthen
previous conclusions from the 2013 Ozone
ISA by providing further support for regional
and global scale decreases in terrestrial C
sequestration from ozone pollution; however,
these relationships are spatially and
temporally dependent. One empirical study
from the Aspen FACE experiment adds to the
evidence base for reduced ecosystem C
content.
IS-64
-------
Table IS-12 (Continued): Summary of evidence for welfare effects of ozone.
Endpoint
Conclusions from 2013 Ozone ISA
Conclusions from 2020 ISA3
Alteration of Causal relationship
belowground ^ has |3een documented since the 2006 Ozone
biogeochemical AQCD that while belowground roots and soil
cycles organisms are not exposed directly to ozone,
Section 8.9 belowground processes could be affected by
ozone through alterations in the quality and
quantity of carbon supply to the soils from
photosynthates and litterfall. The 2013 Ozone
ISA presented evidence that ozone was found
to alter multiple belowground endpoints
including root growth, soil food web structure,
soil decomposer activities, soil respiration, soil
carbon turnover, soil water cycling, and soil
nutrient cycling.
Causal relationship
New evidence confirms conclusions from the
2013 Ozone ISA on effects on soil
decomposition, soil carbon, and soil nitrogen.
The direction and magnitude of these
changes often depends on the species, site,
and time of exposure.
Alteration of
terrestrial
community
composition
Section 8.10
Likely to be causal relationship
The body of evidence is for effects on
community composition shifts in terrestrial plant
communities. For broadleaf forests, the
ozone-tolerant aspen clone was the dominant
clone at the Aspen FACE site. In grasslands,
evidence generally showed shifts from
grass-legume mix to grass species. A shift in
community composition of bacteria and fungi
was observed in both natural and agricultural
systems, although no general pattern could be
discerned.
Causal relationship
Recent evidence builds upon the conclusions
of the 2013 Ozone ISA by strengthening the
understanding of effects of ozone on forest
and grassland communities and confirming
that effects upon soil microbial communities
are diverse. New observational and
experimental studies of ozone effects on tree
species extend to regional forest composition
in the eastern U.S. In grasslands, new studies
are consistent with previous research that
ozone shifts grassland community
composition.
Alteration of
ecosystem water
cycling
Section 8.11
Likely to be causal relationship
Ozone can affect water use in plants and
ecosystems through several mechanisms
including damage to stomatal functioning and
loss of leaf area. Several field and modeling
studies showed an association of ozone
exposure and the alteration of water use and
cycling in vegetation and ecosystems. Direction
of response varied among studies.
Likely to be causal relationship
New evidence is consistent with the findings
in the 2013 Ozone ISA. New evidence
identifies a relationship between ozone and
wood anatomy associated with water
transport. Additional studies add to the
evidence base for decreased root growth and
density. New empirical and modeling studies
continue to show reduced sensitivity of
stomatal closing in response to ozone. There
are a few studies that scale-up these changes
to effects on ecosystem scales including a
study linking ozone effects on tree growth and
water use to ecosystem stream flow in six
watersheds in eastern U.S. forests and from
Aspen FACE.
Radiative forcing
(RF)
Section 9.2
Causal relationship
The 2013 Ozone ISA reported an RF of
0.35 W/m2 from tropospheric ozone from
preindustrial times to the present (1750 to 2005)
based on multimodel studies as reported in the
AR4 IPCC assessment.
Causal relationship
New evidence is consistent with the findings
in the 2013 Ozone ISA. The most recent
IPCC assessment, AR5, reports tropospheric
ozone RF as 0.40 (0.20 to 0.60) W/m2, which
is within range of previous assessments
(i.e., AR4). There have also been a few
individual modeling studies of tropospheric
ozone RF since AR5 which reinforce the AR5
estimates and the causal relationship
between tropospheric ozone and RF.
IS-65
-------
Table IS-12 (Continued): Summary of evidence for welfare effects of ozone.
Endpoint Conclusions from 2013 Ozone ISA Conclusions from 2020 ISA3
Likely to be Causal Relationship
Consistent with previous estimates, the effect
of tropospheric ozone on global surface
temperature continues to be estimated at
roughly 0.1-0.3°C since preindustrial times,
with larger effects regionally. In addition to
temperature, ozone changes have impacts on
other climate metrics such as precipitation
and atmospheric circulation patterns. Current
limitations in climate modeling tools, variation
across models, and the need for more
comprehensive observational data on these
effects represent sources of uncertainty in
quantifying the precise magnitude of climate
responses to ozone changes, particularly at
regional scales.
AQCD = Air Quality Criteria Document; AR4 = IPCC Fourth Assessment Report; AR5 = IPCC Fifth Assessment Report;
FACE = free-air carbon dioxide enrichment; IPCC = Intergovernmental Panel on Climate Change; NAAQS = National Ambient Air
Quality Criteria; OTC = open-top chamber; RF = radiative forcing.
Conclusions from the 2020 ISA include evidence from recent studies integrated with evidence included in previous Ozone ISAs
and AQCDs.
Temperature,
precipitation and
related climate
variables
Section 9.3
Likely to be Causal Relationship
The increase of tropospheric ozone abundance
has contributed an estimated 0.1-0.3°C
warming to the global climate since 1750 based
on studies included in the AR4 IPCC
assessment.
IS-66
-------
IS.5.1 Ecological Effects
The evidence for ozone effects on vegetation and ecosystems is best understood in the context of
some general concepts within ecology. Ecosystems1 are inherently complex and inter-connected.
Ecosystem structure may be described by a variety of measurements used to assess ozone response at
different levels of biological organization [i.e., suborganismal, organism, population,2 community;3 Suter
et al. (2005)1. For example, ozone effects on sensitive species at the whole-plant scale of biological
organization (i.e., reduced growth and biomass, reduced plant reproduction, decreased yield) cascade up
to effects on population and community structure and ecosystem function (Figure IS-3). "Function" refers
to the suite of processes and interactions among the ecosystem components that involve energy or matter.
Examples include water dynamics and the flux of trace gases from processes such as photosynthesis,
decomposition, or carbon cycling. Ecosystem changes are often considered undesirable if important
structural or functional components of the ecosystems are altered following pollutant exposure (U.S.
EPA. 2013a. 1998). Methods to assess effects of ozone on ecological structure and function range from
indoor controlled environment laboratory and greenhouse studies to field observational studies where
biological changes are measured in uncontrolled situations with high natural variability (U.S. EPA. 2015).
Free-air carbon dioxide/ozone enrichment (FACE) systems are a more natural way of estimating ozone
effects on aboveground and belowground processes. Research conducted at the SoyFACE facility in
Illinois (to study responses in soybean fields) and the Aspen FACE (in operation from 1998 to 2011)
system in Wisconsin (to study responses in broadleaf forest) have contributed a substantial body of robust
evidence that supports the characterization of ozone effects at multiple scales. Experimental
methodologies and approaches are summarized in Section 8.1.2.
1 A functional unit consisting of living organisms (biota), their nonliving environment and the interactions within
and between them (IPCC. 201.4).
2 An ecological population consists of interbreeding groups of individuals of the same species that occupy a defined
geographic space. Metrics to assess response in ecological populations include changes over time in abundance or
density (number of individuals in a defined area), age or sex structure, and production or sustainable rates of harvest
(Barnthouse et al.. 2008).
3 Interacting populations of different species occupying a common spatial area form a community (Barnthouse et al..
2008). Community level attributes affected by pollutants include species richness, species abundance, composition,
evenness, dominance of one species over another, or size (area) of the community (U.S. EPA. 2013a).
IS-67
-------
03 exposure
IS55
0 * y
03 uptake & physiology
•Antioxidant metabolism upregulated
¦Decreased photosynthesis
•Decreased stomatal conductance
or sluggish stomatal response
Effects on leaves
•Visible foliar injury
•Altered leaf production
•Altered leaf chemical composition
Plant growth
•Decreased biomass accumulation
•Altered root growth
•Altered carbon allocation
•Altered reproduction
' -Altered crop quality
1
Belowground processes
•Altered litter production and decomposition
•Altered soil carbon and nutrient cycling
•Altered soil fauna and microbial communities
CD
—I
CD
=3
0
GJ
U)
CD
3
CO
<
1 >
Affected ecosystem services
•Decreased productivity
•Decreased C sequestration
• Decreased crop yield
•Altered water cycling
•Altered community composition
•Altered pollination
•Altered forest products
Source: Adapted from U.S. EPA (2013b).
Figure IS-3 Illustrative diagram of ozone effects cascading up through scales
of biological organization from the cellular level to plants and
ecosystems.
Ozone effects on ecosystems are also inter-connected to human health and well-being. The term
"ecosystem services" refers to a concept that ecosystems provide benefits to people, directly or indirectly
(Costanza et al.. 2017). and these benefits are socially and economically valuable goods and services
deserving of protection, restoration, and enhancement (Bovd and Banzhaf. 2007). The concept of
ecosystem services recognizes that human well-being and survival are not independent of the rest of
nature and that humans are an integral and inter-dependent part of the biosphere. Preservation of
ecosystem structure and function contributes to the sustainability of ecosystem services that benefit
human welfare and society. Ecosystem services affected by ozone include productivity, carbon
sequestration, crop yield, water cycling, pollination, and production of forest commodities (Figure IS-3).
Tropospheric ozone affects terrestrial ecosystems across the entire continuum of biological
organization from the cellular and subcellular level to the individual organism up to ecosystem level
IS-68
-------
processes and services (Figure IS-3). For ozone, the majority of evidence for ecological effects is for
vegetation. Damage to terrestrial ecosystems caused by ozone is largely a function of damage to plants,
which starts with uptake of ozone into the leaf via stomata (gas exchange openings on leaves).
Subsequent reactions with plant tissues produce reactive oxygen species that affect cellular function
(Section 8.1.3 and Figure 8-2). Reduced photosynthesis, altered carbon allocation, and impaired stomatal
function lead to observable responses in plants. Observed vegetation responses to ozone include visible
foliar injury (Section IS.5.1.1); and whole-plant level responses (Section IS.5.1.2) including reduction in
aboveground and belowground growth, altered reproduction, and decreased yield. Plant-fauna linkages
affected by ozone include herbivores that feed on ozone-damaged plants and interactions mediated by
volatile plant signaling compounds (Section IS.5.1.3). Ozone can result in broad changes in ecosystems
such as productivity and carbon sequestration (Section IS.5.1.4). belowground processes
(Section IS.5.1.5). terrestrial community composition (Section IS.5.1.6). and water cycling
(Section IS. 5.1.7). Effects of ozone exposure on aboveground and belowground ecosystem components,
across trophic levels, and on carbon allocation at multiple scales of biological organization are described
for forests (Section IS.5.1.8.1) and grasslands (Section IS.5.1.8.2).
IS.5.1.1 Visible Foliar Injury
In the 2013 Ozone ISA the evidence was sufficient to conclude a causal relationship between
ozone exposure and visible foliar injury on sensitive vegetation across the U.S. Visible foliar injury
(Figure IS-4) resulting from exposure to ozone has been well characterized and documented in over six
decades of research on many tree, shrub, herbaceous, and crop species using both long-term field studies
and laboratory approaches (U.S. EPA. 2013b. 2006a. 1996b. 1986. 1978; NAPCA. 1970; Richards et al..
1958). Recent experimental evidence continues to show a consistent association between visible injury
and ozone exposure (Section 8.2). In a recent global-scale synthesis documenting foliar injury from ozone
exposure in the field, across gradients, or in controlled ozone experiments, at least 179 of the identified
plant species have populations in the U.S. (Table 8-4). The use of sensitive species as biological
indicators to detect phytotoxic levels of ozone is a longstanding and effective methodology. More
recently, ozone-sensitive species planted in ozone gardens serve as a source of data on plant responses
and as an educational outreach tool. Although visible injury is a bioindicator of the presence of phytotoxic
concentrations of ozone in ambient air, it is not always a reliable predictor of other negative effects on
vegetation (e.g., growth, reproduction), and foliar injury can vary considerably between and within
taxonomic groups (U.S. EPA. 2013b). Since the 2013 Ozone ISA, new sensitive species showing visible
foliar injury continue to be identified and the role of modifying factors such as soil moisture and time of
day in visible foliar injury symptoms are further characterized (Section 8.2 and Section 8.12). New
information is consistent with the conclusions of the 2013 Ozone ISA that the body of evidence is
sufficient to infer a "causal relationship" between ozone exposure and visible foliar injury.
IS-69
-------
Note: Tulip poplar (Liriodendron tulipifera) on the left and black cherry (Prunus serotina) on the right.
Source: USDA Plants Database. Forest Service Forest Inventory and Analysis Program,
Figure IS-4 Representative ozone foliar injury in two common tree species in
the U.S.
IS.5.1.2 Whole-Plant Effects
The phytotoxicity of tropospheric ozone has been documented for over 50 years in a variety of
plant species (U.S. EPA. 2013b. 2006a. 1996b. 1986. 1978). Ozone-mduced oxidative damage at the
biochemical and leaf-level (Figure IS-3) lead to changes in photosynthesis and carbon allocation which
scale up to reduced growth and impaired reproduction in individual plants. Plant growth is assessed by
quantification of biomass, and analysis of patterns in carbon allocation to aboveground and belowground
plant parts. Direct exposure of reproductive tissues to ozone or indirect effects due to injury of vegetative
tissues results in fewer total available resources to invest in flowers or seeds. In plants cultivated for
agricultural production, damage due to ozone is assessed as reduced crop yield and quality. The evidence
supports causal relationships between ozone and plant growth, plant reproduction, and crop yield, and a
likely to be causal relationship between ozone and tree mortality. Such relationships indicate detrimental
effects of ozone at the individual-organism scale of biological organization.
In the 2013 Ozone ISA the evidence was sufficient to conclude a causal relationship between
ozone exposure and reduced growth of native woody and herbaceous vegetation. As reported in previous
assessments, ozone has long been known to cause decreases in growth which is documented in many
species including herbaceous plants, grasses, shrubs, and trees (U.S. EPA. 2013b. 2006a. 1996b. 1986.
1978). In an analysis conducted in the 2013 Ozone ISA, effects on growth from the Aspen FACE site
closely agreed with exposure-response functions based on data from earlier OTC experiments (U.S. EPA.
2013b). New controlled exposure experiments consistently demonstrate reduced plant growth, and models
IS-70
-------
built with empirical data illustrate potential larger-scale impacts (Section 8.3). In support of findings in
the 2013 Ozone ISA and prior AQCDs, a recent international synthesis of studies published over the past
five decades documents reductions in biomass due to ozone exposure. At least 69 plant species of those
documented in the study have populations in the U.S. (Table 8-7). In addition to reduced growth,
numerous studies from different ecosystems find ozone significantly changes patterns of carbon allocation
below- and aboveground. New evidence from Aspen FACE for effects on growth and biomass of
vegetation includes shifts in wood anatomy (e.g., vessel size and density) and altered distribution of roots
across the soil profile following long-term exposure to elevated ozone. Biomass allocation within an
individual plant is relevant to whole plant growth and function. New studies provide context for scaling
up long-known detrimental effects of ozone on photosynthesis and growth on numerous plant species to
changes at the community and ecosystem level (Section 8.3.3). New information is consistent with the
conclusions of the 2013 Ozone ISA that the evidence is sufficient to infer a "causal relationship"
between ozone exposure and reduced vegetation growth.
Ozone effects on metrics of plant reproduction (e.g., flower number, fruit number, fruit weight,
seed number, rate of seed germination) in multiple experimental settings (e.g., in vitro, whole plants in the
laboratory, whole plants and/or reproductive structures in the green house, and whole plant communities
in the field) reported in the 2006 Ozone AQCD, the 2013 Ozone ISA, and this ISA clearly show ozone
reduces plant reproduction rSection 8.4; U.S. EPA (2013b. 2006a)l. A qualitative review in the 2006
Ozone AQCD showed that plant reproductive organs may be particularly sensitive to ozone injury (Black
et al.. 2000). The biological mechanisms underlying ozone's effect on plant reproduction are twofold.
They include both direct negative effects on reproductive tissues and indirect negative effects that result
from decreased photosynthesis and other whole-plant physiological changes. Since the 2013 Ozone ISA,
a quantitative meta-analysis of >100 independent studies of crop and noncrop species (published from
1968 to 2010) showed statistically significant and sometimes large decreases in reproduction (Leisner and
Ainsworth. 2012). Two metrics of plant reproduction, fruit number and fruit weight, show greater
reductions under increased ozone when combined across species for ozone concentrations that span 40 to
>100 ppb; other metrics do not show such reductions or do so across a narrower range of ozone
concentrations. In addition, there is more recent evidence that plant reproductive tissues are directly
affected by ozone exposure. There are a few new studies on the effects of ozone on phenology
(i.e., timing of germination and flowering), and similar to previously reviewed studies, they have less
consistent results than the studies on plant reproduction. In the 2013 Ozone ISA, plant reproduction was
considered with plant growth. Increased research and synthesis on ozone effects on plant reproduction
(Table 8-9) warrants a separate causality category and evidence is now sufficient to infer a "causal
relationship" between ozone exposure and reduced plant reproduction.
Multiple studies from different research groups show the co-occurrence of ozone exposure and
increased mortality of trees (Section 8.4.3 and Table 8-10). Evidence for plants other than trees is
currently lacking. Studies linking ozone and tree mortality are consistent with known and well-established
individual plant-level mechanisms that explain ozone phytotoxicity, including variation in sensitivity and
IS-71
-------
tolerance based on age class, genotype, and species. Increased mortality is also consistent with effects at
higher levels of biological organization, including changes in vegetation cover and altered community
composition (Section 8.10). Since the 2013 Ozone ISA, a large-scale empirical analysis was conducted of
factors contributing to annual mortality of trees using over three decades of Forest Inventory and Analysis
data. This U.S. Forest Service data showed a significant positive correlation between 8-hour max ozone
concentration and tree mortality. Ozone significantly increased tree mortality in 7 out of 10 plant
functional types in the eastern and central U.S. (Dictzc and Moorcroft. 2011). Experimentally, elevated
ozone exposure has been shown to increase mortality in sensitive aspen genotypes (Moran and Kubiske.
2013). This evidence is considered with studies from the 2006 AQCD and 2013 Ozone ISA where decline
of conifer forests under ozone exposure was continually observed in several regions [Valley of Mexico,
southern France, Carpathian Mountains; U.S. EPA (2013b. 2006a)]. Previous evidence and new evidence
evaluated here is sufficient to infer a "likely to be causal relationship" between ozone exposure and
tree mortality.
In the 2013 Ozone ISA, the evidence was sufficient to conclude a causal relationship between
ozone exposure and reduced yield and quality of agricultural crops. The detrimental effect of ozone on
crop production has been recognized since the 1960s, and a large body of research has subsequently
characterized decreases in yield and quality of a variety of agricultural and forage crops (U.S. EPA.
2013b. 2006a. 1996b. 1986. 1978). The 1986 Ozone AQCD and 1996 Ozone AQCD reported new OTC
experiments on growth and yield, including U.S. EPA's National Crop Loss Assessment Network
(NCLAN), that served at the basis for exposure-response functions for agricultural crop species (U.S.
EPA. 1996b. 1986). As in noncrop plants, the concentrations at which damage is observed vary from
species to species and sometimes between genotypes of the same species.
There is a considerable amount of new research on major U.S. crops, especially soybean, wheat,
and other non-soy legumes at concentrations of ozone occurring in the environment (Section 8.5). For
soybean, further refinement of exposure-response curves and analysis of yield data identified a critical
level of 32 ppb (7-hour seasonal mean) at which a 5% loss can occur (Osborne et al.. 2016). At
SoyFACE, a linear decrease in yield at the rate of 37 to 39 kg per hectare per ppb ozone exposure over
40 ppb (AOT40) was observed across two growing seasons (Betzelberger et al.. 2012). Meta-analyses
published since the 2013 Ozone ISA provide further supporting evidence that current levels of ambient
ozone decrease wheat growth and yield and affect reproductive and developmental plant traits important
to agricultural and horticultural production (Section 8.5). Recent advances in characterizing ozone's
effects on U.S. crop yield include further geographic and temporal refinement of ozone sensitivity and
national-scale estimates of crop losses attributable to ozone. Previous research highlighted in the 2013
Ozone ISA and previous AQCDs show ozone effects on crop yield and crop quality (U.S. EPA. 2013b.
2006a. 1996a. 1986. 1978). New information is consistent with the conclusions of the 2013 Ozone ISA
that the body of evidence is sufficient to infer a "causal relationship" between ozone exposure and
reduced yield and quality of agricultural crops.
IS-72
-------
IS.5.1.3 Effects on Plant-Fauna Interactions
In addition to detrimental effects on plants, elevated ozone can alter ecological interactions
between plants and other species, including (1) herbivores consuming ozone-exposed vegetation,
(2) pollinators and seed dispersers, and (3) predators and parasitoids of insect herbivores. Many of these
interactions are mediated through volatile plant signaling compounds (VPSCs), which plants use to signal
to other community members (Section 8.7V Elevated tropospheric ozone has been shown to alter the
production, emission, dispersion, and lifespan of VPSCs thereby reducing the effectiveness of these
signals. VPSCs play an important role in attracting pollinators, and their alteration can affect the crucial
ecosystem service of pollination of wild plants and crops. Ozone exposure also modifies chemistry and
nutrient content of leaves (U.S. EPA. 2013b). which may affect the physiology and behavior of
herbivores (Section 8.6).
Previous ozone assessments have evaluated studies examining ozone-insect-plant interactions and
found information on a wide range of insect species studied in the orders Coleoptera (weevils, beetles),
Hemiptera (aphids), and Lepidoptera [moths, butterflies; U.S. EPA (2013b. 2006a. 1996bVI. The majority
of studies focused on growth and reproduction while fewer studies considered herbivore survival and
population- and community-level responses to ozone. Although statistically significant effects were
frequently observed, they did not provide any consistent pattern of response across growth, reproduction,
and mortality endpoints. Research has since been published on additional species and at varying levels of
ozone exposure, although there is no clear trend in the directionality of response for most effects
(Section 8.6). The most commonly measured responses are fecundity, development time, growth, and
feeding preferences (Table 8-14). The strongest evidence of ozone effects is from herbivorous insects
with limited evidence from vertebrate feeding studies. Changes in nutrient content and leaf chemistry
following ozone exposure likely account for observed effects in herbivores. The body of evidence is
sufficient to infer a "likely to be causal relationship" between ozone exposure and alteration of
herbivore growth and reproduction.
In the 2013 Ozone ISA, a few experimental and modeling studies reported altered insect-plant
interactions that are mediated through chemical signaling (U.S. EPA. 2013b). New empirical research
from laboratory, greenhouse, OTC, and FACE experiments expand the evidence for altered/degraded
emissions of chemical signals from plants and reduced detection of volatile plant signaling compounds by
insects, including pollinators, in the presence of ozone (Section 8.7 and Table 8-17). New evidence
includes consistent effects in multiple insect species, although this research has examined only a small
fraction of the total number of chemical-signaling responses potentially affected by ozone. Elevated
ozone (>50 ppb) degrades some plant VPSCs, changing the floral scent composition and reducing floral
scent dispersion. Preference studies in a few insect species show reduced pollinator attraction, decreased
plant host detection, and altered plant host preference in the presence of elevated, yet environmentally
relevant ozone concentrations. Exposure to elevated ozone had variable effects on VPSCs emissions and
on the stability of individual volatile compounds with potentially important ecological implications for
IS-73
-------
plant-insect signaling involved in defense against herbivory. To attract predators and parasitoids that
target phytophagous insects, plants emit more VPSCs. Parasitoid-host attraction was either reduced,
enhanced, or unaffected by elevated ozone. The body of evidence is sufficient to infer a "likely to be
causal relationship" between ozone exposure and alteration of plant-insect signaling.
IS.5.1.4 Reduced Productivity and Carbon Sequestration
The evidence in the 2013 Ozone ISA was sufficient to conclude a causal relationship between
ozone exposure and reduced plant productivity (U.S. EPA. 2013b). Studies at the leaf and plant scale
show that ozone decreases plant growth, providing biological plausibility for decreases in ecosystem
productivity. Evidence of decreased ecosystem productivity from ozone exposure comes from many
different experiments with different study designs in a variety of ecosystems: OTC experiments;
long-term, ecosystem-manipulation, chamberless exposure experiments (Aspen FACE, SoyFACE,
FinnishFACE); empirical models using eddy covariance measures; forest productivity models
parameterized with empirical physiological and tree life history data; and various well-studied ecosystem
models and scenario analysis (Section 8.8. IV New information is consistent with the conclusions of the
2013 Ozone ISA that the body of evidence is sufficient to infer a "causal relationship" between ozone
exposure and reduced productivity in terrestrial ecosystems.
The evidence in the 2013 Ozone ISA was sufficient to conclude a likely causal relationship
between ozone exposure and decreased terrestrial carbon sequestration (U.S. EPA. 2013b).
Ozone-mediated changes in plant carbon budgets result in less carbon available for allocation to various
pools: reproductive organs, leaves, stems, storage, and roots as well as maintenance, defense, and repair.
Changes in allocation (Section 8.8.3) can scale up to population- and ecosystem-level effects, including
changes in soil biogeochemical cycling (Section 8.9). increased tree mortality (Section 8.4.3). shifts in
community composition (Section 8.10). changes to species interactions (Section 8.6). declines in
ecosystem productivity and carbon sequestration (Section 8.8). and alteration of ecosystem water cycling
(Section 8.11). The relationship between ozone exposure and terrestrial C sequestration is difficult to
measure at the landscape scale. Most of the evidence regarding this relationship is from model
simulations, although this endpoint was also examined in a long-term manipulative chamberless
ecosystem experiment (Aspen FACE). For example, experiments at Aspen FACE found ozone exposure
caused a 10% decrease in cumulative (Net Primary Production) and an associated 9% decrease in
ecosystem C storage, although the effects of ozone gradually disappeared towards the end of the 10-year
exposure (Talhelm et al.. 2014; Zak et al.. 2011) possibly due to loss of ozone-sensitive individuals and
lower ozone exposures in the last 3 years. Additional studies at this research site suggests that the effects
of ozone on plant productivity will be paralleled by large and meaningful decrease in soil C, but the
experimental observations reviewed did not find a direct link between ozone, NPP, and soil C pools. It is
likely that stand age and development and disturbance regimes are complicating factors in the partitioning
of ecosystem-level effects of ozone exposure on carbon sequestration. Even with these limitations, the
IS-74
-------
results from the Aspen FACE experiment and the model simulations provide further evidence that is
consistent with the conclusions of the 2013 Ozone ISA that the body of evidence is sufficient to infer a
"likely to be causal relationship" between ozone exposure and reduced carbon sequestration in
ecosystems.
15.5.1.5 Belowground Processes/Biogeochemical Cycles
In the 2013 Ozone ISA, the evidence was sufficient to conclude that there is a causal relationship
between ozone exposure and the alteration of belowground biogeochemical cycles (U.S. EPA. 2013b). It
has been documented since the 2006 Ozone AQCD (U.S. EPA. 2006a') that while belowground roots and
soil organisms are not exposed directly to ozone, below-ground processes can be affected by ozone
through alterations in the quality and quantity of carbon supply to the soils from photosynthates and
litterfall (Andersen. 2003). The 2013 Ozone ISA presented evidence that ozone was found to alter
multiple belowground endpoints including root growth, soil food web structure, soil decomposer
activities, soil respiration, soil carbon turnover, soil water cycling, and soil nutrient cycling. The new
evidence since the 2013 Ozone ISA (U.S. EPA. 2013b) included in this assessment confirms ozone effects
on soil decomposition (Section 8.9.1). soil carbon (Section 8.9.2). and soil nitrogen (Section 8.9.3).
although the direction and magnitude of these changes often depends on the species, site, and length of
exposure. As in the 2013 Ozone ISA, the evidence is sufficient to conclude that there is a "causal
relationship" between ozone exposure and the alteration of belowground biogeochemical cycles.
15.5.1.6 Terrestrial Community Composition
In the 2013 Ozone ISA, the evidence was sufficient to conclude that there is a likely causal
relationship between ozone exposure and the alteration of community composition of some ecosystems,
including conifer forests, broadleaf forests, and grasslands, and altered fungal and bacterial communities
in the soil in both natural and agricultural systems (U.S. EPA. 2013b). Ozone effects on individual plants
can alter the larger plant community as well as the belowground community of microbes and
invertebrates, which depend on plants as carbon sources. Ozone may alter community composition by
having uneven effects on co-occurring species, decreasing the abundance of sensitive species and giving
tolerant species a competitive advantage. Key new studies (Wang et al.. 2016; Gustafson et al.. 2013)
model ozone effects on regional forest composition in the eastern U.S. Additionally, a global-scale
synthesis of decades of research on an array of ozone effects on plants confirms that some plant families
(e.g., Myrtaceae, Salicaceae, and Onagraceae) are more susceptible to ozone damage than others
(Bergmann et al.. 2017). This lends biological plausibility to a mechanism by which elevated ozone alters
terrestrial community composition by inhibiting or removing ozone-sensitive plant species or genotypes,
which alters competitive interaction to favor the growth or abundance of ozone-tolerant species or
genotypes. In grasslands, previous evidence included multiple studies from multiple research groups to
IS-75
-------
show that elevated ozone shifts the balance among grasses, forbs, and legumes (Section 8.10.1.2). There
are new studies with findings consistent with earlier research (Section 8.10). including new studies from
European grasslands that found exposure-response relationships between ozone and community
composition. The 2013 Ozone ISA presented multiple lines of evidence that elevated ozone alters
terrestrial community composition, and recent evidence strengthens our understanding of the effects of
ozone upon plant communities, while confirming that the effects of ozone on soil microbial communities
are diverse (Table 8-20). The evidence is sufficient to conclude that there is a "causal relationship"
between ozone exposure and the alteration of community composition of some ecosystems.
15.5.1.7 Ecosystem Water Cycling
In the 2013 Ozone ISA, the evidence was sufficient to conclude a likely causal relationship
between ozone exposure and the alteration of ecosystem water cycling (U.S. EPA. 2013b). Ozone can
affect water use in plants and ecosystems through several mechanisms, including damage to stomatal
functioning and loss of leaf area, which may affect plant and stand evapotranspiration and lead, in turn, to
possible effects on hydrological cycling. Although the 2013 Ozone ISA found no clear universal
consensus on leaf-level stomatal conductance response to ozone exposure, many studies reported
incomplete stomatal closure and loss of stomatal control in several plant species, which result in increased
plant water loss [Section 9.4.5; U.S. EPA (2013b)l. Additionally, ozone has been found to alter plant
water use through decreasing leaf area index, accelerating leaf senescence, and by causing changes in
branch architecture, which can significantly affect stand-level water cycling. There is mounting
biologically relevant, statistically significant, and coherent evidence from multiple studies of various
types about the mechanisms of ozone effects on plant water use and ecosystem water cycling (reduced
leaf area, reduced leaf longevity, changes in root and branch biomass and architecture, changes in vessel
anatomy, stomatal dysfunction, reduced sap flow; rSection 8.111). Additionally, there are a few strong
studies that scale up these changes to effects on ecosystem scales and show significant effects. The most
compelling evidence is from six watersheds in eastern forests and from Aspen FACE (Kostiainen et al..
2014; Sun et al.. 2012). This new information adds to the evidence base in the 2013 Ozone ISA and
supports the conclusion that the body of evidence is sufficient to infer a "likely to be causal
relationship" between ozone exposure and the alteration of ecosystem water cycling.
15.5.1.8 Integration of Ozone Effects in Ecosystems
IS.5.1.8.1 Forests
The effects of ozone exposure on U.S. forests have been an active area of research for over
50 years; evaluation of the role of ozone in forest health declines in the mixed conifer forest of the San
IS-76
-------
Bernardino Mountains began in the early 1960s (Miller and McBride. 1999). Since that time, studies have
confirmed variation in sensitivity to ozone exposure in trees and plants based on age class, genotype, and
species (U.S. EPA. 2013b. 2006a. 1996b). There has been strong and consistent evidence from multiple
studies that ozone-induced oxidative damage leads to declines in photosynthesis and carbon gain, which
scale up to reduced growth in individual plants [Section 8.3; U.S. EPA (2013b. 2006a. 1996bVI. For
example, studies from the Aspen FACE experiment have shown that ozone caused reduction in total
biomass in quaking aspen (Populus tremuloides), paper birch (Betula papyrifera), and sugar maple [Acer
saccharum; U.S. EPA (2013b)l. These findings were overall consistent with open-top chamber studies
that established ozone exposure-response relationships on growth in a number of native U.S. tree species
detailed in previous NAAQS reviews (U.S. EPA. 2013b); these species include aspen, black cherry
(Prunus serotina), tulip poplar (Liriodendron tulipifera), white pine (Pinus strobus), and ponderosa pine
(Pinusponderosa). In addition to overall reductions in growth, there is evidence that ozone changes plant
growth patterns by significantly reducing root growth in some tree species. New information reviewed in
the current document support earlier conclusions that ozone reduces photosynthesis, growth, and carbon
allocation in a number of plant species found in forest ecosystems.
In addition to declines in root carbon allocation, results from Aspen FACE and other
experimental studies reviewed in the 2013 Ozone ISA consistently found that ozone exposure reduced
litter production and altered leaf chemistry in trees (U.S. EPA. 2013b). These direct effects of ozone on
plants may lead to changes in soil properties and processes in forests, but these changes are dependent on
species and genotype of community members, and potentially on other factors like the stage of stand
development.
Ozone effects on tree water use can also scale up to significant and measurable effects on
ecosystem water cycling in forests. Ozone-mediated impairment of stomatal function in plants has been
documented for decades (Keller and Hasler. 1984). although impairment seems to be species specific.
Studies continue to show reduced sensitivity of stomatal closing in response to various factors (light,
vapor pressure deficit, temperature, soil moisture) when exposed to ozone ("sluggish stomata") in a
number of species. A recent meta-analysis of ozone effects on stomatal response in 68 species (including
trees, crops, and grassland) found that trees were the most adversely affected, with 73% showing an
altered stomatal response. In this synthesis, 4 tree species exhibited sluggish stomata and 13 showed
stomatal opening in response to ozone (Mills et al.. 2016; Mills et al.. 2013). Ozone exposure has also
been linked to decreased water use efficiency and changes in sap flow (Mclaughlin et al.. 2007a;
Mclaughlin et al.. 2007b) and to reduced late-season stream flow in eastern forest ecosystems (Sun et al..
2012).
Differences between species in ozone sensitivity leads to significant changes in forest community
composition, as ozone sensitive trees decline and are replaced by less sensitive ones (Section 8.10.1.1).
Species-specific responses to ozone in terms of plant growth reductions and biomass allocation are a
possible mechanism for these community shifts. In a model simulation of long-term effects of ozone on a
IS-77
-------
typical forest in the southeastern U.S. involving different tree species with varying ozone sensitivity,
Wang et al. (2016) found that ozone significantly altered forest community composition and decreased
plant biodiversity. Models using Aspen FACE data confirm that ozone effects on tree biomass and
productivity scale to affect community composition at the genotype and species level (Moran and
Kubiske. 2013). In simulations using Aspen FACE data of northern forests at the landscape level over
centuries, elevated ozone altered species abundance and the speed of replacement and succession
(Gustafson et al.. 2013). Multiple studies from different research groups show the co-occurrence of ozone
exposure and increased mortality of trees (Section 8.4.3). In a Bayesian empirical model built with field
measurement data from the U.S. Forest Service's Forest Inventory and Analysis program, ozone
significantly increased tree mortality in 7 out of 10 plant functional types in the eastern and central U.S.
(Dietze and Moorcroft. 2011).
IS.5.1.8.2 Grasslands
In grassland ecosystems, herbaceous plants and grasses in particular are the dominant vegetation
rather than shrubs or trees. There is a wide range of sensitivity to ozone in grassland plant communities.
For example, studies going back to the 1996 Ozone AQCD show varying ozone sensitivity within the
genus Trifolium (clover) and general shifts in community biomass that favors grass species (U.S. EPA.
1996a). Evidence reviewed in the 2013 Ozone ISA from a large-scale ozone fumigation experiment in
grasslands demonstrated ozone decreases gross primary productivity in these systems (Volketal.. 2011).
Experiments reviewed in the 2013 Ozone ISA and previous AQCDs and the current Ozone ISA generally
show ozone associated with biomass loss, and a decrease in nutritive quality of forage species. Further,
ozone responses differed across species of grassland plants (Yolk et al.. 2006). Ozone effects on seed
production, germination, and flower number and date of peak flowering have been demonstrated in
representative grassland species (Section 8.4).
In grasslands, ozone effects on biodiversity or species composition may result from competitive
interactions among plants in mixed-species communities. Studies from mesocosm, OTC, and FACE
experiments generally show a shift in the biomass from grass-legume mixtures over time, in favor of
grass species. There are also new studies from European grasslands that found exposure-response
relationships for community composition (Section IS.5.1.9) that included some species that also grow in
the U.S. In the 2013 Ozone ISA, a review of ozone sensitive plant communities [identified as sensitive if
they had six or more species that exhibited significant ozone-caused changes in biomass in peer-reviewed
controlled experiments; Mills et al. (2007)1 found that the largest number of these sensitive communities
were associated with grassland ecosystems (U.S. EPA. 2013b). Among grassland ecosystems, alpine
grassland, subalpine grassland, woodland fringe, and dry grassland were identified as the most
ozone-sensitive communities. Ozone effects on grassland ecosystems extend belowground to the
associated soil microbial communities (Section 8.10.2). which show changes in proportions of bacteria or
fungi in response to elevated ozone and to fauna that feed on grassland vegetation.
IS-78
-------
IS.5.1.9 Exposure-Response Relationships
For over 40 years, controlled ozone exposure experiments have yielded a wealth of information
on exposure-response relationships. Ozone exposure response has been demonstrated in many tree and
herbaceous species, including crops (U.S. EPA. 2013b. 2006a. 1996b. 1986. 1978). As described in
Section IS.3.2. various indices have been used to quantify ozone exposure in plants, including
threshold-weighted (e.g., SUM06) and continuous sigmoid-weighted (e.g., W126) functions. Weighting
of cumulative indices takes into account the greater effects of ozone on vegetation with elevated ozone
concentrations. As ozone concentrations increase, plant defense mechanisms are overwhelmed and the
capacity of the plant to detoxify reactive oxygen species is compromised (U.S. EPA. 2013b). For decades,
it has also been well characterized that plant sensitivity varies by time of day and development stage.
Growth responses vary depending on the growth stage of the plant. Furthermore, the time of highest
ozone concentrations may not occur at the time of maximum plant uptake. Weighted hourly
concentrations during the daylight hours and during the growing season are the most important variables
in a cumulative exposure index (U.S. EPA. 2013b). For vegetation, quantifying exposure with indices that
accumulate the ozone hourly concentrations and preferentially weight the higher concentrations improves
the explanatory power of exposure for effects on growth and yield, compared with using indices based on
mean and peak exposure values.
None of the information on the effects of ozone on vegetation published since the 2013 Ozone
ISA has modified conclusions on quantitative exposure-response relationships. Since the 2013 Ozone
ISA, there have been a few new experimental studies that add more exposure-response relationship
information to the large historical database available on U.S. plants (Section 8.13.2). In a new
experimental study, Betzelberger et al. (2012) studied seven cultivars of soybean at the SoyFACE
experiment in Illinois. They found that the cultivars showed similar responses in a range of ozone
exposures expressed as AOT40 (Section IS. 3.2). These results support conclusions of previous studies
(Betzelberger et al.. 2010) and the 2013 Ozone ISA that sensitivity of current soybean genotypes is not
different from early genotypes; therefore, soybean response functions developed in the NCLAN program
remain valid. A study by Neufeld et al. (2018) provided information on foliar injury response on two
varieties of cutleaf coneflower (Rudbeckia laciniata). For example, one variety had statistically detectable
foliar injury when the 24-hour W126 index reached 23 ppm hour (12-hour AOT40 = 12 ppm hour).
Although recent U.S. exposure-response studies in experimental systems are limited, U.S. and
international syntheses have highlighted response function information (e.g., biomass growth, foliar
injury, yield) for grassland and other plant species that occur in the U.S. (see Section 8.13.2). For
example, in a synthesis of previously published studies, linear relationships of biomass growth in
response to ozone were found using AOT40 for 87 grassland species that occur in Europe (van Goethem
et al.. 2013). Seventeen of these species are native to the U.S. and 65 additional species have been
introduced to the U.S. and may have significant ecological, horticultural, or agricultural value (USDA.
2015). This study has the most significant amount of new exposure response information for plants in the
U.S.
IS-79
-------
IS.5.2
Effects on Climate
Changes in the abundance of tropospheric ozone perturb the radiative balance of the atmosphere
by interacting with incoming solar radiation and outgoing longwave radiation. This effect is quantified by
the radiative forcing metric. Radiative forcing is the perturbation in net radiative flux at the tropopause (or
top of the atmosphere) caused by a change in radiatively active forcing agent(s) after stratospheric
temperatures have readjusted to radiative equilibrium (stratospherically adjusted radiative forcing).
Through this effect on the Earth's radiation balance, tropospheric ozone plays a major role in the climate
system, and increases in its ozone abundance contribute to climate change (Mvhre et al.. 2013).
For ozone effects on climate (Appendix 9). there are inter-connections to human health and
ecosystems. As discussed in the 2013 Ozone ISA, the Earth's atmosphere-ocean system responds to
changes in radiative forcing with a climate response, including a change in near-surface air temperature
with associated impacts on precipitation and atmospheric circulation patterns. This climate response
causes downstream climate-related health and ecosystem effects, such as the combined health effects of
both climate (e.g., heat waves) and ozone air quality or redistribution of diseases or ecosystem
characteristics. Feedbacks from both the direct climate response and such downstream effects can, in turn,
affect the abundance of tropospheric ozone and ozone precursors through multiple mechanisms
(Figure IS-5). Variations in climate can potentially alter the conditions that lead to the formation,
transport, and persistence of ozone in the troposphere (Appendix 1). as well as increase the vulnerability
of plants and ecosystems. The degree to which climate and weather alter the effects of ozone is context
and species specific because damage to terrestrial ecosystems caused by ozone is largely a function of
plant uptake. Factors that modify the effects of ozone in ecosystems, including carbon dioxide, weather,
and climate are discussed in Section 8.12.
IS-80
-------
Climate Effects
on Human Health r
and Ecosystems )
Changes in Tropospheric
O, Abundance
^ ™
Radiative Forcing
Due to O, Change
(W/m2)
Climate Response
(°C)
Precursor Emissions of
CO, VOCs, CH4, NOx
(Tg/y)
Source: U.S. EPA (2013bl
Figure IS-5 Schematic illustrating the effects of tropospheric ozone on
climate; including the relationship between precursor emissions,
tropospheric ozone abundance, radiative forcing, climate
response and climate impacts.
Characterization of ozone impacts on radiative forcing (Section 9.2) builds on the findings in the
2013 Ozone ISA and draws heavily on the IPCC Assessment Reports. In the 2013 Ozone ISA, the
evidence was sufficient to conclude a causal relationship between tropospheric ozone and radiative
forcing (U.S. EPA. 2013b). The 2013 Ozone ISA reported a radiative forcing (RF) of 0.35 W/m2 from the
change in global tropospheric ozone abundance from preindustrial times to the present (1750 to 2005)
based on multimodel studies (Forster et al.. 2007). The most recent IPCC assessment, AR5, reports global
tropospheric ozone RF as 0.40 (0.20 to 0.60) W/m2 (Mvhre et al.. 2013). which is within range of
previous assessments (i.e., AR4). There have also been a few individual studies of tropospheric ozone RF
(Section 9.2) since AR5, including the study of tropospheric ozone RF based on the Coupled Model
Intercomparison Project Phase 6 (CMIP6) data set, and the Atmospheric Chemistry and Climate Model
Intercomparison Project (ACCMIP) multimodel study of tropospheric chemistry, all of which reinforce
IS-81
-------
the AR5 estimates and continues to support a "causal relationship" between tropospheric ozone and
RF.
In the 2013 Ozone ISA, the evidence was sufficient to conclude a likely to be causal relationship,
via radiative forcing, between tropospheric ozone and climate change (now referred to as "temperature,
precipitation, and related climate variables"; the revised title for this causality statement provides a more
accurate reflection of the available evidence) (U.S. EPA. 2013b). New studies reviewed in Section 9.3 are
consistent with previous estimates and the effect of global, total tropospheric ozone increases on global
mean surface temperature continues to be estimated at roughly 0.1-0.3°C since preindustrial times (Xic et
al.. 2016; Mvhre et al.. 2013). with larger effects regionally. In addition to temperature, ozone changes
affect other climate metrics such as precipitation and atmospheric circulation patterns (Macintosh et al..
2016; Allen et al.. 2012; Shindell et al.. 2012). All of this evidence reinforces a "likely to be causal
relationship" between temperature, precipitation, and related climate variables.
1S.6 Key Aspects of Health and Welfare Effects Evidence
There is extensive scientific evidence that demonstrates health and welfare effects from exposure
to ozone. In assessing the older and more recent evidence, the U.S. EPA characterizes the key strengths
and remaining limitations of this evidence. In the process of assessing the evidence across studies and
scientific disciplines and ultimately forming causality determinations, the U.S. EPA takes into
consideration multiple aspects that build upon the Hill criteria (Hill. 1965) and include, but are not limited
to, consistency in findings, coherence of findings, and evidence of biological plausibility [see U.S. EPA
(2015)1. As documented by the extensive evaluation of evidence throughout the subsequent Appendices
to this ISA, the U.S. EPA carefully considers uncertainties in the evidence, and the extent to which recent
studies have addressed or reduced uncertainties from previous assessments, as well as the strengths of the
evidence. Uncertainties considered in the epidemiologic evidence, for example, include potential
confounding by copollutants or covarying factors and exposure error. The U.S. EPA evaluates many other
important considerations (not uncertainties) such as coherence of evidence from animal and human
studies, heterogeneity of risk estimates, and the shape of the concentration-response relationships. All
aspects are considered along with the degree to which chance, confounding, and other biases affect
interpretation of the scientific evidence in the process of drawing scientific conclusions and making
causality determinations. Uncertainties do not necessarily change the fundamental conclusions of the
literature base. In fact, some conclusions may be robust to such uncertainties. Where there is clear
evidence linking ozone with health and welfare effects with or despite minimal remaining uncertainties,
the U.S. EPA makes a determination of a causal or likely to be causal relationship.
IS-82
-------
IS.6.1
Health Effects Evidence: Key Findings
A large body of scientific evidence spanning many decades clearly demonstrates there are health
effects related to both short- and long-term ozone exposure (Figure IS-6). The strongest evidence supports
a relationship between ozone exposure and respiratory health effects. The collective body of evidence for
each health outcome category evaluated in this ISA is systematically considered and assessed, including
the inherent strengths, limitations, and uncertainties in the overall body of evidence, resulting in the
causality determinations detailed in Table IS-1. Through identification of the strengths and limitations in
the evidence, this ISA may help in the prioritization of research efforts to support future ozone NAAQS
reviews.
An inherent strength of the evidence integration in this ISA is the extensive amount (in both
breadth and depth) of available evidence resulting from decades of scientific research that describes the
relationship between both short- and long-term ozone exposure and health effects. The breadth of the
enormous database is illustrated by the different scientific disciplines that provide evidence
(e.g., controlled human exposure, epidemiologic, animal toxicological studies), the range of health
outcomes examined (e.g., respiratory, cardiovascular, metabolic, reproductive, and nervous system
effects, as well as cancer and mortality), and the large number of studies within several of these outcome
categories. The depth of the literature base is exemplified by the examination of effects that range from
biomarkers of exposure, to subclinical effects, to overt clinical effects, and even mortality. Depth is
further demonstrated through the variety of the study designs used across the scientific disciplines and
exposure duration periods.
In this ISA, systematic review methodologies are applied to identify and characterize this
expansive evidence base (see Appendix 10 for details). The evidence is integrated from (1) a variety of
study designs within the same scientific discipline, (2) different scientific disciplines, and (3) a span of
different health endpoints within a health effect category. Finally, a formal framework is applied
systematically to draw conclusions about the causal nature of the relationship between ozone exposure
and health effects (U.S. EPA. 2015).
A first step in integrating evidence for a health effect category is to consider the biological
plausibility of health responses observed in association with ozone exposure. The process for
characterizing biological plausibility is described in Section IS.4.2. Recent studies in humans and animals
expand on findings from prior assessments (U.S. EPA. 2013b. 2006a. 1996a) to further understand
plausible pathways that may underlie the observed respiratory health effects related to short-term
exposure to ozone (Figure 3-1). Consistent evidence for several respiratory endpoints within a large
number of animal toxicological, controlled human exposure, and epidemiologic studies, as well as
coherent evidence across these studies contribute to a large degree of certainty in assessing the
relationship between short-term ozone exposure and this health effect category. Furthermore, uncertainty
is addressed by epidemiologic studies that examine potential copollutant confounding, examine different
model specifications, and account for potential confounders.
IS-83
-------
Causality Determinations for Health Effects of Ozone
2020 Ozone ISA
Short-term
exposure
Kespiraiory
Long-term
exposure
Metabolic
Short-term
exposure
+
Long-term
exposure
+
Cardiovascular
Short-term
exposure
i
Long-term
exposure
0)
E
o
o
Nervous System
Short-term
exposure
O
.c
*¦*
CO
Long-term
exposure
~o
=5
"O
Male/Female
Reproduction
and Fertility
Long-term
*
o
Q.
<1)
Pregnancy and
Birth Outcomes
exposure
*
Cancer
Long-term
exposure
Mortality
Short-term
exposure
i
Long-term
exposure
Causal J Likely causal Q Suggestive Q Inadequate
+ new causality determination; 1 causality determination changed from likely
causal to suggestive; * change in scope of health outcome category from 2013
Ozone ISA
Figure IS-6 Causality determinations for health effects of short- and
long-term exposure to ozone.
IS-84
-------
Both older and more recent studies provide evidence for biologically plausible pathways that may
underlie respiratory effects related to long-term ozone exposure, and metabolic effects related to
short-term exposure. Epidemiologic studies of long-term ozone exposure and respiratory effects are
supported by numerous animal toxicological studies examining related endpoints. This coherence reduces
some of the uncertainty related to the independence of the ozone effect, though there are some remaining
uncertainties for these health effects. For example, there are still relatively few studies evaluating the
effect of ozone exposure on metabolic effects in human populations (i.e., controlled human exposure or
epidemiologic studies).
With regard to short-term ozone exposure and cardiovascular health effects, there is some
evidence for biologically plausible pathways for the worsening of IHD or HF, the development of heart
attack or stroke, and cardiovascular-related ED visits and hospital admissions (Figure 4-1). However, the
evidence comes mainly from animal toxicological studies, is generally not supported by controlled human
exposure studies, and is limited for epidemiologic studies. While there is some epidemiologic evidence
that short-term ozone concentrations are associated with total mortality, the evidence of plausible steps
that could lead to death (e.g., IHD, HF) are lacking in epidemiologic studies that examined these types of
endpoints (e.g., hospital admissions for IHD or HF). Furthermore, controlled human exposure studies in
healthy adults generally do not show that short-term ozone exposure leads to the types of intermediate
health effects (e.g., impaired vascular function, changes in ECG measures) that could lead to IHD or
stroke. Most of the studies supporting the biological plausibility of epidemiologic studies of mortality are
from animal studies that are not generally supported by studies in humans.
Older and recent studies examining short- or long-term ozone exposure and several other health
effects (i.e., nervous system effects, reproductive effects, cancer) are few or report inconsistent evidence
of an association with the health effect of interest. For these health effects, there is often limited
coherence across studies from different scientific disciplines, and limited evidence for biologically
plausible pathways by which effects could occur. Other sources of uncertainty, such as limited assessment
of potential copollutant confounding, are inherent in these evidence bases.
There is strong and consistent animal toxicological evidence linking short- and long-term ozone
exposure with respiratory, cardiovascular, and metabolic health effects. However, several uncertainties
should be considered when evaluating and synthesizing evidence from these studies. Experimental studies
are often conducted at ozone concentrations higher than those observed in ambient air (i.e., 250 to
>1,000 ppb) to evoke a response within a reasonable study length. These studies are informative and the
conduct of studies at these concentrations is commonly used for identifying potential human hazards.
There are also substantial differences in exposure concentrations and exposure durations between animal
toxicological and controlled human exposure studies. For example, animal toxicological studies generally
expose rodents to 250 to >1,000 ppb, while controlled human exposure studies generally expose humans
to 60 to 300 ppb. Additionally, a number of animal toxicological studies were performed in rodent disease
models, while controlled human exposure studies generally are conducted in healthy individuals. This
IS-85
-------
difference could explain some of the inconsistencies across studies between these scientific disciplines.
Controlled human exposure studies do not typically include unhealthy or diseased individuals for ethical
reasons; therefore, this represents an important uncertainty to consider in interpreting the results of these
studies. Additional animal toxicological studies conducted at lower concentrations could help to reduce
this uncertainty. Finally, in addition to exposure concentration and disease status differences in
physiology (e.g., rodents are obligate nose breathers), differences in the duration and timing of exposure
(e.g., rodents are exposed during the day, during their resting cycle, while humans are exposed during the
day when they are normally active), and differences in the temperature at which the exposure was
conducted may contribute to the lack of coherence between results of experimental animal and human
studies. Dosimetric studies of animals and humans might inform understanding of the potential role of
such differences.
Controlled human exposure studies provide the strongest evidence for the effects of short-term
ozone exposure on respiratory effects. There are, however, several limitations of controlled human
exposure studies. These include the study of generally healthy individuals and the measurement of
relatively minor health effects (or indices of health effects) for ethical reasons (unhealthy or very sick
people are studied rarely). Therefore, individuals that may be at greater risk are not included in controlled
human exposure studies. However, controlled human exposure studies offer several strengths for studying
human health effects from ozone exposure. The experimental nature of controlled human exposure studies
allows them to virtually eliminate the chance, bias, and other potential confounding factors inherent in
observational epidemiologic studies. In addition, controlled human exposure studies are not susceptible to
some of the uncertainties commonly attributed to animal toxicological studies, such as the need to
extrapolate between animal models and humans, and the use of relatively high ozone concentrations
compared with concentrations typically encountered in ambient air.
Though susceptible to chance, bias, and other potential confounding due to their observational
nature, epidemiologic studies have the benefit of evaluating real-world exposure scenarios and can
include populations that cannot typically be included in controlled human exposure studies, such as
children, pregnant women, and individuals with pre-existing disease. In addition, innovations in
epidemiologic study designs and methods have substantially reduced the role of chance, bias, and other
potential confounders in well-designed, well-conducted epidemiologic studies. Many epidemiologic
studies have been conducted in diverse geographic locations, encompassing different population
demographics, and using a variety of exposure assignment techniques. They continue to report consistent,
positive associations between short-term ozone exposure and health effects. When combined with
coherent evidence from experimental studies, the epidemiologic evidence can support and strengthen
determinations of the causal nature of the relationship between health effects and exposure to ozone at
relevant ambient air concentrations.
The most common source of uncertainty in epidemiologic studies of ozone is exposure
measurement error. The majority of recent epidemiologic studies of long-term ozone exposure use
IS-86
-------
concentrations from fixed-site monitors as exposure surrogates. Some recent epidemiologic studies
incorporate new ozone exposure assignment methods that integrate several sources of available data
(i.e., satellite observations, CTM predictions, and ambient monitors) into a spatiotemporal model. These
hybrid methods are well validated by ozone monitors in areas with moderate to high population density,
and they better allow for the inclusion of populations from less urban areas, where monitor density is
lower. Relatively low spatial variability of ozone (compared with UFP, CO, NO2, or SO2) in most
locations increases confidence in application of these methods for predicting ozone exposure.
Furthermore, disentangling the effects of short-term ozone exposure from those of long-term ozone
exposure (and vice-versa) is an inherent uncertainty in the evidence base.
Additionally, the populations included in epidemiologic studies have long-term, variable, and
uncharacterized exposures to ozone and other ambient pollutants. Epidemiologic studies evaluate the
relationship between health effects and specific ozone concentrations during a defined study period. The
generally consistent and coherent associations observed in these epidemiologic studies contribute to the
causality determinations and the conclusions regarding the causal nature of the effect of ozone exposure
on health effects, However, they do not provide information about which averaging times or exposure
metrics may be eliciting the health effects under study.
Each of the exposure assignment methods used in short- and long-term ozone exposure
epidemiologic studies have inherent strengths and limitations, and exposure measurement errors
associated with those methods contribute bias and uncertainty to health effect estimates. For short-term
exposure studies, exposure measurement error generally leads to underestimation and reduced precision
of the association between short-term ozone concentrations and health effects. For long-term exposure
studies, exposure measurement error can bias effect estimates in either direction, although it is more
common that effect estimates are underestimated. Underestimation of health effect associations in short-
and long-term ozone exposure studies implies that true health effect associations are even larger than
what is reported in epidemiologic studies. The magnitude of bias in the effect estimate is likely small for
ozone, because ozone concentrations do not vary over space as much as other criteria pollutants, such as
NOx or SO2 (Section 2.6V
Copollutant analyses were limited in epidemiologic studies evaluated in the 2013 Ozone ISA but
indicated that associations between ozone concentrations and health effects were not confounded by
copollutants or aeroallergens (U.S. EPA. 2013b). Copollutant analyses are more prevalent in recent
studies and continue to suggest that observed associations are independent of coexposures to correlated
pollutants or aeroallergens. Despite expanded copollutant analyses in recent studies, determining the
independent effects of ozone in epidemiologic studies is complicated by the high copollutant correlations
observed in some studies, and the possibility for effect estimates to be overestimated for the pollutant
measured with less error in copollutant models (Section 2.5). That said, some studies report modest
copollutant correlations, which suggests that strong confounding due to copollutants is unlikely. In
addition, evidence from copollutant models is available for a small subset of all the pollutants that
IS-87
-------
co-occur with ozone in the air. Nonetheless, the consistency of associations observed across studies with
different copollutant correlations, the generally robust associations observed in copollutant models, and
evidence from other scientific disciplines generally provide compelling evidence for an independent
effect of ozone exposure on human health and reduce the uncertainties associated with potential
copollutant confounding.
The 2013 Ozone ISA noted that multicity epidemiologic studies, particularly examining
short-term ozone exposure and mortality, reported evidence of heterogeneity in the magnitude and
precision of risk estimates across cities. There are few recent multicity studies of short-term ozone
exposure and health effects that could allow an evaluation of such heterogeneity; thus, the uncertainty
identified in the 2013 Ozone ISA remains.
Examination of the concentration-response (C-R) relationship has primarily been conducted in
studies of short-term ozone exposure and respiratory health effects or mortality, with some more recent
studies characterizing this relationship for long-term ozone exposure and mortality. Across recent studies
that used a variety of statistical methods to examine potential deviations from linearity, evidence
continues to support a linear C-R relationship, but with less certainty in the shape of the curve at lower
concentrations (i.e., below 30-40 ppb). In addition, some studies evaluate the potential for a
population-level threshold, below which health effects would unlikely be observed. Generally, these
studies conclude that if a population-level threshold exists, it would occur at lower concentrations
(i.e., below 30-40 ppb) where there is less certainty in the ozone-health effect relationship due to few
observations at these lower concentrations. Similar to the uncertainty mentioned previously, the
populations included in epidemiologic studies have long-term, variable, and uncharacterized exposures to
ozone and other ambient pollutants. Epidemiologic studies evaluate the C-R relationship between health
effects and specific ozone concentrations during a defined study period. The generally consistent C-R
relationships observed in these epidemiologic studies do not indicate which averaging times or exposure
metrics may be eliciting the health effects under study.
IS.6.2 Welfare Effects Evidence: Key Findings
The collective body of evidence for each welfare endpoint evaluated in this ISA was carefully
considered and assessed, including the inherent strengths, limitations, and uncertainties in the overall
body of evidence, resulting in the causality determinations for ecological effects detailed in Table IS-2
and effects on climate in Table IS-3.
IS.6.2.1 Ecological Effects
A large body of scientific evidence spanning more than 60 years clearly demonstrates there are
effects on vegetation and ecosystems attributed to ozone exposure resulting from anthropogenic activities
IS-88
-------
(U.S. EPA. 2013b. 2006a. 1996b. 1986. 1978; NAPCA. 1970; Richards et al.. 1958). There is high
certainty in ozone effects on impairment to leaf physiology as mechanisms for cascading effects at higher
levels of biological organization (e.g., plant growth, ecosystem productivity; Section 8.1.3; Figure IS-7).
The overwhelming strength of many of the studies is that they consist of controlled ozone exposure to
plants, plots of forests, and crop fields to eliminate any confounding factors (Section 8.12). For example,
for ozone effects on plants, there are robust exposure response functions (i.e., from carefully controlled
experimental conditions, involving multiple concentrations and based on multiple studies) for about a
dozen important tree species and ten major commodity crop species.
The use of visible foliar injury to identify phytotoxic levels of ozone is an established and widely
used methodology. However, foliar injury is not always a reliable indicator of other negative effects on
vegetation (e.g., growth, reproduction), and there is a lack of quantitative exposure-response information
that accounts for the important role of soil moisture in foliar injury. As documented in the 2013 Ozone
ISA (Table IS-12) and retained in the current Ozone ISA (Figure IS-7). there are causal relationships
between ozone exposure and visible foliar injury at the individual-organism level, and causal relationships
between ozone exposure and reduced plant growth and crop yield from the individual to population
levels. Since the 2013 Ozone ISA (U.S. EPA. 2013b). a meta-analysis of existing literature on plant
reproductive metrics and new research support a causal relationship between ozone exposure and reduced
plant reproduction. In the previous ISA, plant reproduction was considered within the broader category of
growth but the current body of evidence for this endpoint warrants a separate causality category.
While the effect of ozone on vegetation is well established in general, there are some knowledge
gaps regarding precisely which species are sensitive and what exposures elicit adverse responses for many
species. Currently there are over 40,000 plants and lichens occurring in the U.S. as documented by the
USDA PLANTS database (USDA. 2015). It not feasible to know what the effects are on all U.S. species
and what the ecological consequences of the differential sensitivities are of these species. However, there
have been many important trees, crops, and other plants studied to indicate the potential array of
ecological effects in the U.S. The exposure-response relationships for a subset of individual plants are
discussed in Section 8.13. Within and between these species there is a range of sensitivities, and it is
difficult to identify the representativeness of these relationships within the wider population of plants that
occur in the U.S. There are also uncertainties about how plant responses change with age and size. The
technique of meta-analysis is one approach that can be used to consolidate and extract a summary of
significant responses from a selection of previously published studies. These meta-analyses can show
patterns of cause and effect relationships between ecological endpoints and ozone exposure; they are
robust enough to overcome individual variation and are useful for looking at trends in plant response
across, for example, geographic locations, environmental conditions, plant functional groups, and
ecosystems.
IS-89
-------
Causality Determinations for Ecological Effects of Ozone
Scale of Ecological Response
Ecosystem
Belowground
Biogeochemical Cycles
Water Cycling
Carbon Sequestration
Productivity
Community
Biodiversity
Terrestrial Community Composition!
Species Interactions
Plant-Insect Signaling +
Population
Individual
Survival
Trees+
Growth
Plants
Herbivores +
Reproduction
Plants+
Herbivores +
Yield
Agricultural Crops
Individual
Visible Foliar Injury
Causal |^| Likely Causal |
+ new causality determination;!causality determination changed from likely to be
causal to causal
Figure IS-7 Causality determinations for ecological effects of ozone across
biological scales of organization and taxonomic groups.
IS-90
-------
The majority of evidence for ecological effects of ozone is for vegetation. Fewer studies examine
plant-ozone-insect interactions. There are multiple, statistically significant findings showing ozone effects
on fecundity and growth in insect herbivores. However, no consistent directionality of response is
observed across the literature, and uncertainties remain in regard to different plant consumption methods
across species and the exposure conditions associated with particular severities of effects. There is also
variation in study designs and endpoints used to assess ozone responses. Most responses observed in
insects appear to be indirect (i.e., mediated through ozone effects on vegetation, although direct effects of
ozone exposure on insects could also play a role). New research in chemical ecology has provided clear
evidence of ozone modification of VPSCs and behavioral responses of insects to these modified chemical
signatures; however, most of these studies have been carried out in laboratory conditions rather than in
natural environments. Characterization of airborne pollutant effects on chemical signaling in ecosystems
is an emerging area of research with information available on a relatively small number of insect species
and plant-insect associations and knowledge gaps in the mechanisms and consequences of modulation of
VPSCs by ozone.
There are some uncertainties in characterizing how ozone damage to leaves and individual plant
species scale up to ecological communities and ecosystem processes. Although estimating ozone effects
to the ecosystem level remains a challenge, there is a large body of knowledge of how ecosystems work
gained through ecological observations and models that simulate processes at multiple scales. The models
attempt to capture interactive effects of multiple stressors in ecosystems in the field. Studies of ozone
effects beyond the plant scale use a combination of empirical studies and statistical modeling, or large
controlled exposure ecosystem experiments, or field observations along ozone gradients. Interactive
effects in natural ecosystems with multiple stressors (e.g., drought, disease) are difficult to study, but can
be addressed through different statistical methods. For example, multivariate models and mechanistic
models have been used for studying ozone with other environmental factors [e.g., Dietze and Moorcroft
(2011)1 and for scaling up ozone effects on tree growth and water use to ecosystem stream flow [e.g., Sun
et al. (2012)1. Another approach is to use meta-analysis techniques to examine trends across large
geographic areas or at higher biological levels of organization (e.g., plant functional groups, forest types).
More research on ecosystem-level responses will strengthen understanding of how effects at lower levels
of biological organization influence higher level responses.
In general, the most promising approaches to estimating or characterizing ozone effects at the
ecosystem level include evaluation of ecological response using a suite of parameters and
exposure-response functions, both empirical and modeled. The quantitative uncertainty of empirically
observed variables in ecology is determined by the use of statistics. In general, ecological endpoints
affected by ozone were reported in the ISA if they were statistically significant. In addition, models of
chemical and ecological processes provide representations of biological interactions through
mathematical expressions. The models used can be complex, including many interacting variables. Each
of the input variables in a model has some uncertainty. Models can also be evaluated on the basis of the
IS-91
-------
mechanistic understanding of how ecological systems work and how ozone effects may propagate
through ecological systems.
IS.6.2.2 Effects on Climate
Ozone is an important greenhouse gas and increases in its abundance have affected the Earth's
climate. Over the last century, global average surface air temperature has increased by approximately
1.0°C, and emissions of greenhouse gases are the dominant cause (Wuebbles et al.. 2017; IPC'C. 2013).
There are many other aspects of the global climate system that are changing in addition to this warming,
including melting glaciers, reductions in snow cover and sea ice, sea level rise, ocean acidification, and
increases in the frequency or intensity of many types of extreme weather events ("Wuebbles et al.. 2017).
The magnitude of future climate change, globally and regionally, and in terms of both temperature
increases and these other types of associated impacts, will depend primarily on the amount of greenhouse
gases emitted globally (Wuebbles et al.. 2017; IPC'C. 2013). The most recent IPCC report, AR5, which is
a comprehensive assessment of the peer-reviewed literature, reported global tropospheric ozone RF as
0.40 (0.20 to 0.60) W/m2 (Mvhrc et al.. 2013). In the 2013 Ozone ISA, there was a causal relationship
between tropospheric ozone and RF and a likely to be causal relationship between tropospheric ozone and
climate change (U.S. EPA. 2013b). None of the new studies support a change to either causality
determination (Figure IS-8).
While the warming effect of tropospheric ozone in the climate system is well established in
general, various uncertainties render the precise magnitude of the overall effect of tropospheric ozone on
climate more uncertain than that of the well-mixed greenhouse gases (Mvhre et al.. 2013). These include
several uncertainties associated with estimating the magnitude of RF attributed to tropospheric ozone
increases, such as uncertainties in estimating preindustrial ozone concentrations. In addition, precisely
quantifying changes in surface temperature due to tropospheric ozone changes, along with related climate
effects, requires complex climate simulations that include all relevant feedbacks and interactions. For
example, trends in free tropospheric ozone and upper tropospheric ozone (where RF is particularly
sensitive to changes in ozone concentrations) are not captured well by models. Substantial variation also
exists across models. Such modeling uncertainties make it especially difficult to provide precise
quantitative estimates of the climate effects of regional-scale ozone changes.
IS-92
-------
Causality Determinations for Tropospheric Ozone arid
Climate Change
Radiative Forcing
Temperature, precipitation and
related climate variables
Causal B-l Likely Causal
Figure IS-8 Causality determinations for tropospheric ozone and climate
change.
IS-93
-------
IS.7 References for Integrative Synthesis
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/10.1080/0895837029008461Q
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.
http://dx.doi.org/10.1080/0895837039Q168283
Adams. WC. (2006). Comparison of chamber 6.6-h exposures to 0.04-0.08 ppm ozone via square-
wave and triangular profiles on pulmonary responses. Inhal Toxicol 18: 127-136.
http://dx.doi.org/10.1080/089583705003061Q7
Alberti. KG; Eckel. RH: Grundy. SM; Zimmet. PZ; Cleeman. JI; Donato. KA: Fruchart. JC; James.
WP; Loria. CM; Smith. SC. (2009). Harmonizing the metabolic syndrome: A joint interim
statement of the International Diabetes Federation Task Force on Epidemiology and Prevention;
National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation;
International Atherosclerosis Society; and International Association for the Study of Obesity.
Circulation 120: 1640-1645. http://dx.doi.org/10.1161/CIRCULATIQNAHA.109.192644
Allen. RJ; Sherwood. SC; Norris. JR; Zender. CS. (2012). Recent Northern Hemisphere tropical
expansion primarily driven by black carbon and tropospheric ozone. Nature 485: 350-354.
http://dx.doi.org/10.1038/naturell097
Andersen. CP. (2003). Source-sink balance and carbon allocation below ground in plants exposed to
ozone. New Phytol 157: 213-228. http://dx.doi.Org/10.1046/i.1469-8137.2003.00674.x
Barnthouse. LW; Munns. RM. Jr; Sorensen. MT. (2008). Population-level ecological risk assessment.
Pensacola, FL: Taylor & Francis, http://www.crcpress.com/product/isbn/9781420053326
Bergmann. E; Bender. J; Weigel. HJ. (2017). Impact of tropospheric ozone on terrestrial biodiversity:
A literature analysis to identify ozone sensitive taxa. J Appl Bot Food Qual 90: 83-105.
http://dx.doi.org/10.5073/JABFO.2017.090.Q12
Betzelberger. A; Yendrek. CR; Sun. J; Leisner. CP; Nelson. RL; Ort. PR: Ainsworth. EA. (2012).
Ozone exposure response for U.S. soybean cultivars: Linear reductions in photosynthetic potential,
biomass, and yield. Plant Physiol 160: 1827-1839. http://dx.doi.Org/10.1104/pp.l 12.205591
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.org/ 10.1111/i.l365-
3040.2010.02165.x
Black. VJ; Black. CR; Roberts. JA; Stewart. CA. (2000). Tansley Review No. 115: Impact of ozone
on the reproductive development of plants. New Phytol 147: 421-447.
http://dx.doi.Org/10.1046/i.1469-8137.2000.00721.x
Blackwell. PL; Lucas. JW; Clarke. TC. (2014). Summary health statistics for U.S. adults: National
Health Interview Survey, 2012. (Vital and Health Statistics: Series 10, No. 260). Hyattsville, MP:
National Center for Health Statistics, http://www.cdc.gov/nchs/data/series/sr 10/srlQ 260.pdf
Bloom. B; Jones. LI; Freeman. G. (2013). Summary health statistics for U.S. children: National Health
Interview Survey, 2012. (Vital and Health Statistics: Series 10, No. 258). Hyattsville, MP:
National Center for Health Statistics, http://www.cdc.gov/nchs/data/series/sr 10/srlQ 258.pdf
IS-94
-------
Bovd. J; Banzhaf. S. (2007). What are ecosystem services? The need for standardized environmental
accounting units. Ecol Econ 63: 616-626.
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
Costanza. R; De Groot. R; Braat. L: Kubiszewski. I: Fioramonti. L: Sutton. P: Farber. S; Grasso. M.
(2017). Twenty years of ecosystem services: How far have we come and how far do we still need
to go? Ecosyst Serv 28: 1-16. http://dx.doi.Org/10.1016/i.ecoser.2017.09.008
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/ies.2009.49
Di. O: Dai. L: Wang. Y: Zanobetti. A: Choirat. C: Schwartz. JD: Dominici. F (2017). Association of
short-term exposure to air pollution with mortality in older adults. JAMA 318: 2446-2456.
http://dx.doi.org/10.1001/iama.2017.17923
Dietze. MC: Moorcroft. PR. (2011). Tree mortality in the eastern and central United States: Patterns
and drivers. Global Change Biol 17: 3312-3326. http://dx.doi.Org/10.l 11 l/i.1365-
2486.2011.02477.x
Forster. P: Ramaswamv. V: Artaxo. P: Berntsen. T: Betts. R: Fahev. DW: Havwood. J: Lean. J: Lowe-
DC: Mvhre. G: Nganga. J: Prinn. R: Raga. G: Schultz. M: Van Dorland. R. (2007). Changes in
atmospheric constituents and in radiative forcing. In Climate change 2007: The physical science
basis Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental
Panel on Climate Change (pp. 129-234). Cambridge, United Kingdom: Cambridge University
Press, http://www.ipcc.ch/pdf/assessment-report/ar4/wgl/ar4-wgl-chapter2.pdf
Gustafson. EJ; Kubiske. ME: Sturtevant. BR: Miranda. BR. (2013). Scaling Aspen-FACE
experimental results to century and landscape scales. Landsc Ecol 28: 1785-1800.
http://dx.doi.org/10.1007/slQ980-013-9921-x
Hales. CM: Carroll. MP: Frvar. CD: Ogden. CL. (2017). Prevalence of Obesity Among Adults and
Youth: United States, 20152016. NCHS Data Brief 288.
Hill. AB. (1965). The environment and disease: Association or causation? Proc R Soc Med 58: 295-
300.
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.
http://dx.doi.Org/10.l 164/airccm/142.5.1158
Hovert. PL: Xu. J. (2012). Deaths: Preliminary data for 2011. Natl Vital Stat Rep 61: 1-51.
IPCC (Intergovernmental Panel on Climate Change). (2013). Climate change 2013: The physical
science basis. Contribution of working group I to the fifth assessment report of the
Intergovernmental Panel on Climate Change. In TF Stacker; P Qin; GK Plattner; MMB Tignor;
SK Allen; J Boschung; A Nauels; Y Xia; V Bex; PM Midgley (Eds.). Cambridge, UK: Cambridge
University Press, http://www.ipcc.ch/report/ar5/wg 1 /
Jaffe. PA: Cooper. OR: Fiore. AM: Henderson. BH; Tonnesen. GS: Russell. AG: Henze. PK;
Langford. AO: Lin. M: Moore. T. (2018). Scientific assessment of background ozone over the US:
Implications for air quality management. Elementa: Science of the Anthropocene 6.
http://dx.doi.org/10.1525/elementa.309
IS-95
-------
Keller. T; Hasler. R. (1984). The influence of a fall fumigation with ozone on the stomatal behavior of
spruce and fir. Oecologia 64: 284-286. http://dx.doi.org/10.1007/BF0Q376884
Kim. CS; Alexis. NE: Rappold. AG; Kehrl. H: Hazucha. MJ: Lav. 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.201011-1813QC
Kostiainen. K: Saranpaa. P; Lundqvist. SO; Kubiske. ME: Vapaavuori. E. (2014). Wood properties of
Populus and Betula in long-term exposure to elevated C02 and 0-3. Plant Cell Environ 37: 1452-
1463. http://dx.doi.org/10.Ill 1/pce. 12261
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/10.1016/00Q4-
6981(87)90038-2
Leisner. CP; Ainsworth. EA. (2012). Quantifying the effects of ozone on plant reproductive growth
and development. Global Change Biol 18: 606-616. http://dx.doi.org/10.1111/i. 1365-
2486.2011.02535.x
Lewis. TC; Robins. TG; Mentz. GB; Zhang. X; Mukheriee. B; Lin. X; Keeler. GJ; Dvonch. JT; Yip.
FY; O'Neill. MS; Parker. EA; Israel. BA; Max. PT; Reves. A; Committee. CAAACS. (2013). Air
pollution and respiratory symptoms among children with asthma: Vulnerability by corticosteroid
use and residence area. Sci Total Environ 448: 48-55.
http ://dx.doi .org/10.1016/i. scitotenv.2012.11.070
Macintosh. CR; Allan. RP; Baker. LH; Bellouin. N; Collins. W; Mousavi. 7, Shine. KP (2016).
Contrasting fast precipitation responses to tropospheric and stratospheric ozone forcing. Geophys
Res Lett 43: 1263-1271. http://dx.doi.org/10.1002/2015GLQ67231
McCorrv. LK. (2007). Physiology of the autonomic nervous system [Review]. Am J Pharm Educ 71:
78.
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. http://dx.doi.org/10.1080/00039896.1991.9937441
McDonnell. WF; Stewart. PW; Smith. MY. (2013). Ozone exposure-response model for lung function
changes: An alternate variability structure. Inhal Toxicol 25: 348-353.
http://dx.doi.org/10.3109/08958378.2013.79Q523
McDonnell. WF; Stewart. PW; Smith. MY; 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. http://dx.doi.org/10.3109/08958378.2012.7Q5919
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.Org/10.llll/i.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.org/10.1111/i. 1469-8137.2007.01970.X
IS-96
-------
Miller. PR; McBride. JR. (1999). Assessment of ecological risks and implications for policy and
management in the San Bernardino Mountains. In PR Miller; JR McBride (Eds.), Oxidant air
pollution impacts in the montane forests of southern California: A case study of the San Bernardino
Mountains (pp. 397-404). New York, NY: Springer. http://dx.doi.org/10.10Q7/978-l-4612-1436-
6 19
Mills. G: Harmens. H; Wagg. S: Sharps. K; Haves. F; Fowler. D; Sutton. M; Davies. B. (2016). Ozone
impacts on vegetation in a nitrogen enriched and changing climate. Environ Pollut 208: 898-908.
http://dx.doi.Org/10.1016/i.envpol.2015.09.038
Mills. G: Haves. F; Jones. MLM; Cinderbv. S. (2007). Identifying ozone-sensitive communities of
(semi-)natural vegetation suitable for mapping exceedance of critical levels. Environ Pollut 146:
736-743. http://dx.doi.Org/10.1016/i.envpol.2006.04.005
Mills. G; Wagg. S; Harmens. H. (2013). Ozone pollution: Impacts on ecosystem services and
biodiversity. Bangor, UK: NERC/Centre for Ecology & Hydrology.
http://nora.nerc.ac.uk/id/eprint/502675/
Moran. EV; Kubiske. ME. (2013). Can elevated C02 and ozone shift the genetic composition of aspen
(Populus tremuloides) stands? New Phytol 198: 466-475. http://dx.doi.org/10.Ill 1/nph. 12153
Mvhre. G: Shindell. D: Breon. FM; Collins. W: Fuglestvedt. J: Huang. J; Koch. D; Lamarque. JF; Lee.
D; Mendoza. B; Nakajima. T; Robock. A; Stephens. G; Takemura. T; Zhang. H. (2013).
Anthropogenic and natural radiative forcing. In TF Stacker; D Qin; GK Plattner; MMB Tignor; SK
Allen; J Boschung; A Nauels; Y Xia; V Bex; PM Midgley (Eds.), Climate change 2013: The
physical science basis Contribution of working group I to the fifth assessment report of the
Intergovernmental Panel on Climate Change (pp. 659-740). Cambridge, UK: Cambridge University
Press, http://www.ipcc.ch/report/ar5/wg 1/
NAPCA (National Air Pollution Control Administration). (1970). Air quality criteria for
photochemical oxidants [EPA Report]. (AP-63). Washington, DC: U.S. Department of Health,
Education, and Welfare, http://nepis.epa.gov/Exe/ZvPURL.cgi?Dockev=9100E2Z7.txt
Neufeld. HS: Johnson. J; Kohut. R. (2018). Comparative ozone responses of cutleaf coneflowers
(Rudbeckia laciniata var. digitata, var. ampla) from Rocky Mountain and Great Smoky Mountains
National Parks, USA. Sci Total Environ 610-611: 591-601.
http ://dx.doi .org/10.1016/i. scitotenv.2017.08.046
NRC (National Research Council). (2009). Science and decisions: Advancing risk assessment.
Washington, DC: The National Academies Press, http://dx.doi.org/10.17226/12209
OECD (Organisation for Economic Co-operation and Development). (2016). Users handbook
supplement to the guidance document for developing and assessing AOPs.
(ENV/JM/MONO(2016)12). http://dx.doi.org/10.1787/5ilvlm9dlg32-en
OMB (U.S. Office of Management and Budget). (2004). Final information quality bulletin for peer
review. (Memorandum M-05-03). Washington, DC: US Office of Management and Budget
(OMB). http://www.whitehouse.gov/sites/default/files/omb/assets/omb/memoranda/fV2005/m05-
03.pdf
Osborne. SA; Mills. G: Haves. F: Ainsworth. EA; Biiker. P; Emberson. L. (2016). Has the sensitivity
of soybean cultivars to ozone pollution increased with time? An analysis of published dose-
response data. Global Change Biol 22: 3097-3111. http://dx.doi.org/10.llll/gcb.13318
Pleis. JR: Lucas. JW: Ward. BW. (2009). Summary health statistics for U.S. adults: National Health
Interview Survey, 2008. (Vital and Health Statistics: Series 10, No. 242). Hyattsville, MD:
National Center for Health Statistics, http://www.cdc.gov/nchs/data/series/sr 10/srlQ 242.pdf
IS-97
-------
Richards. BL; Middleton. JT; Hewitt. WB. (1958). Air pollution with relation to agronomic crops: V:
Oxidant stipple of grape. J Am Soc Agron 50: 559-561.
http://dx.doi.org/10.2134/agronil958.00021962005000090Q19x
Schelegle. ES; Morales. CA; Walbv. 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.200809-1484QC
Shindell. D; Kuvlenstierna. JC: Vignati. E; van Dingenen. R; Amann. M: Klimont. Z; Anenberg. SC:
Muller. N; Janssens-Maenhout. G; Raes. F; Schwartz. J; Faluvegi. G; Pozzoli. L; Kupiainen. K;
Hoglund-Isaksson. L; Emberson. L; Streets. D: Ramanathan. V; Hicks. K; Oanh. NT; Millv. G:
Williams. M; Demkine. V: Fowler. D. (2012). Simultaneously mitigating near-term climate change
and improving human health and food security. Science 335: 183-189.
http ://dx.doi .org/10.1126/science. 1210026
Sun. G: Mclaughlin. SB; Porter. JH; Uddling. J; Mulholland. PJ: Adams. MB; Pederson. N. (2012).
Interactive influences of ozone and climate on streamflow of forested watersheds. Global Change
Biol 18: 3395-3409. http://dx.doi.org/10.1111/i. 1365-2486.2012.02787.X
Suter. GW; Norton. SB; Fairbrother. A. (2005). Individuals versus organisms versus populations in the
definition of ecological assessment endpoints. Integr Environ Assess Manag 1: 397-400.
http://dx.doi.org/10.1002/ieam.56300104Q9
Talhelm. AF; Pregitzer. KS; Kubiske. ME; Zak. PR; Campanv. CE; Burton. AJ; Dickson. RE;
Hendrev. GR; Isebrands. JG; Lewin. KF; Nagy. J; Karnoskv. DF. (2014). Elevated carbon dioxide
and ozone alter productivity and ecosystem carbon content in northern temperate forests. Global
Change Biol 20: 2492-2504. http://dx.doi.org/10. Ill 1/gcb. 12564
Team. CW; Pachauri. RK; Mever. LA. (2014). Annex II: Glossary [Mach, K.J., S. Planton and C. von
Stechow (eds.)]. In T Core Writing; RK Pachauri; LA Meyer (Eds.), Climate Change 2014:
Synthesis Report Contribution of Working Groups I, II and III to the Fifth Assessment Report of
the Intergovernmental Panel on Climate Change (pp. 117-130). Geneva, Switzerland: IPCC.
https://www.ipcc.ch/site/assets/uploads/2018/02/AR5 SYR FINAL Annexes.pdf
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, DC.
http://nepis.epa.gov/exe/Z vPURL.cgi?Dockev=200089CW.txt
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.
https://ntrl.ntis.gov/NTRL/dashboard/searchResults.xhtml?searchOuerv=PB87142949
U.S. EPA (U.S. Environmental Protection Agency). (1996a). Air quality criteria for ozone and related
photochemical oxidants [EPA Report]. (EPA/600/P-93/004AF). Research Triangle Park, NC.
U.S. EPA (U.S. Environmental Protection Agency). (1996b). Air quality criteria for ozone and related
photochemical oxidants, Vol. II of III [EPA Report]. (EPA/600/P-93/004BF). Research Triangle
Park, NC.
U.S. EPA (U.S. Environmental Protection Agency). (1998). Guidelines for ecological risk assessment
[EPA Report]. (EPA/630/R-95/002F). Washington, DC: U.S. Environmental Protection Agency,
Risk Assessment Forum, https://www.epa.gov/risk/guidelines-ecological-risk-assessment
IS-98
-------
U.S. EPA (U.S. Environmental Protection Agency). (2006a). Air quality criteria for ozone and related
photochemical oxidants [EPA Report]. (EPA/600/R-05/004AF). Research Triangle Park, NC: U.S.
Environmental Protection Agency, Office of Research and Development, National Center for
Environmental Assessment-RTP Office.
http://cfpub.cpa.gov/ncca/cfin/rccordisplav.cfin7dcidH49923
U.S. EPA (U.S. Environmental Protection Agency). (2006b). Air Quality Criteria for Ozone and
Related Photochemical Oxidants. Volume 1 of 3. (EPA/600/R-05/004AF; EPA600R05004AF).
Research Triangle Park, NC. National Center for Environmental Assessment.
:
Environmental Protection Agency General.
http://nepis.epa.gov/exe/ZvPURL.cgi?Dockev=P100CJZN.txt
U.S. EPA (U.S. Environmental Protection Agency). (2009). Integrated science assessment for
particulate matter [EPA Report]. (EPA/600/R-08/139F). Research Triangle Park, NC: U.S.
Environmental Protection Agency, Office of Research and Development, National Center for
Environmental Assessment- RTP Division.
http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=216546
U.S. EPA (U.S. Environmental Protection Agency). (2013a). Integrated science assessment for lead
[EPA Report]. (EPA/600/R-10/075F). Research Triangle Park, NC: U.S. Environmental Protection
Agency, National Center for Environmental Assessment.
http://cfpub.cpa.gov/ncca/cfm/rccordisplav.cfm?dcid=255721
U.S. EPA (U.S. Environmental Protection Agency). (2013b). Integrated science assessment for ozone
and related photochemical oxidants [EPA Report]. (EPA/600/R-10/076F). Research Triangle Park,
NC: U.S. Environmental Protection Agency, Office of Research and Development, National Center
for Environmental Assessment-RTP Division.
http://cfpub.epa.gov/ncea/isa/recordisplav.cfin?deid=247492
U.S. EPA (U.S. Environmental Protection Agency). (2014). Integrated review plan for the primary
national ambient air quality standards for nitrogen dioxide [EPA Report]. (EPA-452/R-14/003).
Research Triangle Park, NC: U.S. Environmental Protection Agency, National Center for
Environmental Assessment.
http://www.epa.gOv/ttn/naaqs/standards/nox/data/201406finalirpprimarvno2.pdf
U.S. EPA (U.S. Environmental Protection Agency). (2015). Preamble to the integrated science
assessments [EPA Report]. (EPA/600/R-15/067). Research Triangle Park, NC: U.S. Environmental
Protection Agency, Office of Research and Development, National Center for Environmental
Assessment, RTP Division. https://cfbub.epa.gov/ncea/isa/recordisplav.cfm?deid=310244
U.S. EPA (U.S. Environmental Protection Agency). (2016). Integrated science assessment for oxides
of nitrogen-health criteria (final report) [EPA Report]. (EPA/600/R-15/068). Research Triangle
Park, NC: U.S. Environmental Protection Agency, Office of Research and Development, National
Center for Environmental Assessment.
http://ofinpub.epa.gov/eims/eimscomm.getfile7p download id=526855
USD A (U.S. Department of Agriculture). (2015). Natural Resources Conservation Service PLANTS
database. Available online at http://plants.usda.gov/iava/
van Goethem. TM; Azevedo. LB; van Zelm. R; Haves. F; Ashmore. MR; Huiibregts. MA. (2013).
Plant species sensitivity distributions for ozone exposure. Environ Pollut 178: 1-6.
http://dx.doi.Org/10.1016/i.envpol.2013.02.023
IS-99
-------
Vinikoor-Imler. LC; Owens. EO; Nichols. JL; Ross. M; Brown. JS; Sacks. JD. (2014). Evaluating
potential response-modifying factors for associations between ozone and health outcomes: A
weight-of-evidence approach [Review]. Environ Health Perspect 122: 1166-1176.
http://dx.doi.org/10.1289/ehp.1307541
Yolk. M; Bungener. P; Contat. F; Montani. M; Fuhrer. J. (2006). Grassland yield declined by a quarter
in 5 years of free-air ozone fumigation. Global Change Biol 12: 74-83.
http://dx.doi.org/10.1111/i. 1365-2486.2005.01083.X
Yolk. M; Obrist. D; Novak. K; Giger. R; Bassin. S; Fuhrer. J. (2011). Subalpine grassland carbon
dioxide fluxes indicate substantial carbon losses under increased nitrogen deposition, but not at
elevated ozone concentration. Global Change Biol 17: 366-376. http://dx.doi.Org/10.l 111/i. 1365-
2486.2010.02228.x
Wang. B. in; Shugart. HH: Shuman. JK; Lerdau. MT. (2016). Forests and ozone: productivity, carbon
storage, and feedbacks. Sci Rep 6: 22133. http://dx.doi.org/10.1038/srep22133
Wuebbles. DJ; Fahev. DW; Hibbard. KA: Dokken. DJ; Stewart. BC; Mavcock. TK. (2017). USGCRP,
2017: Climate science special report: Fourth national climate assessment, Volume I. Washington,
DC: U.S. Global Change Research Program, https://science2017.globalchange.gov/
Xie. B; Zhang. H; Wang. Z; Zhao. S: Fu. O. (2016). A modeling study of effective radiative forcing
and climate response due to tropospheric ozone. Adv Atmos Sci 33: 819-828.
http://dx.doi.org/10.1007/s00376-Q16-5193-0
Zak. PR; Pregitzer. KS: Kubiske. ME; Burton. AJ. (2011). Forest productivity under elevated C02
and 03: positive feedbacks to soil N cycling sustain decade-long net primary productivity
enhancement by C02. Ecol Lett 14: 1220-1226. http://dx.doi.0rg/lO.l 111/i.1461-
0248.2011.01692.x
IS-100
-------
APPENDIX 1 ATMOSPHERIC SOURCE,
CHEMISTRY, METEOROLOGY,
TRENDS, AND BACKGROUND
OZONE
Stininuiry of Yt'ir ICviilcncc Related to (hone in . \mhient. 1 ir
• O/one formed m I lie troposphere is. pri m;iri l\. ilic product of photochemical rc;iclions
hem ecu iniroucii oxides (\C) i mid c;irhon coiilmiiiim compounds iiichidiim \ okililc
orumnc compounds (YO( \ i. cmhoii moiioMilc (CO). ;nnl niclkiiic Since I lie 2u I '
O/one IS V ;uldilioii;il d;ii;i collcclioii ;nul research h;i\ c ;ul\ ;inccd onr iiiulcrs|;iiidiim
of ilic sources ill uroiiiid-lc\ el ozone. iiichidiim emissions due lo Iiiiiikiii ;icli\ Hies
w ilInii 1 lie I S ;ind iiileriKiluiiKill\. Ii;il11i';iI ;nul hiolouic;il processes. ;nnl dsiimnics
w illiiii I Willi's ;iiiiKispheie While ikimesiie milhropoucmc emisskiiis ofo/one
precursors h;i\e l;irucl> deelnied o\cr llie p;isi 15 2o \e;irs. ii/iine le\els siill c\cccd
llie \ \ \OS mi Mime I S kie;ilkiiis
• While o/one is ordinmiK ;i \\;inn sc;ison polliil;nil. iiiiiisii;iII\ hiuli coiicciiir;ilioii
ozone c\ enls h;i\ e occurred mi llie w inler mi l\xn wesieni iikiiiiiI;iiii h;isins llie I inl;i
;iiul I pper (iieen ki\cr I5;isms keeenl c\ iilenee snuucsls ih;il kie;il w inler
nieleorolouv ;iinl 11iu11 emissions from oil ;nul u;is c\ir;iclioii oper;ilions ;ippe;ir lo he
llie principal dri\ers of w inler o/one foriiuilion. in iliese loe;ilioiis
• \ilil11hiii;i11>. coiiliiniiim rose;iiv11 on llie role of IkiIoucii ehemisirs in hoiind;ir\ l;i\er
o/one eoiieeiiir;ilioiis indie;iles lh;il llie process ni;i\ ser\ e ;is ;in o/one sink in eo;is|;il
iiiii;ni cn\ iroiimeiils. When ;iddcd lo model ehemie;il meehmiisiiis. kiloucn ehemisirs
;ippe;irs lo eorreel pre\ ions o\ erpredielion of o/one coiicciiir;ilioiis
• ()/one iiie;isiireineiil e;ip;ihiliiies h;i\ e iinpro\ ed since llie prc\ ions ozone ;issessnieiii.
iiichidiim esi;ihhshnieiii of;i new lk\l ;md eiih;iiiced use of snielhie h;ised reniole
seiisinu nielhods \i ihe smiie 11inc. llierc h;i\e hcen iioi;ihle ;ul\;iiiccs in reuion;il
("I'M nielhods. incliulinu inipro\ cniciii in ch;ir;iclcri/nm IkiIoucii chcniisin. kind
co\cr. ne;ir siirl';icc nicleorolous. dr\ dcposiiion. sir;iiospherc troposphere c\ch;iimc.
hioucmc emissions. mid iiiicumiioii w nil niclcorolomc;il models.
• IniproN eniciiis mi icuioikiI ;md hemispheric modeliiiu nielhods h;i\e led lo impro\ed
csiinuiles in I S h;ickuroiind o/one (I SI5) \cross ilic eiiscnihle of ;i\;nl;ihlc
modehiiu sindics in llie hlcr;iliirc. sc;ison;il me;iii I SI! coiicciitr;ilioiis ;irc cs|im;iled lo
in line from 2<> 5( > pph ' These model csiini;iles of sc;isoii;iI nie;ni I SI! o/one coiil;iin
iiiicei'Ciililies of ;ihonl In pph for se;ison;il ;i\ cr;mc concern r;il ions w nil liiuher
iiiiccri;iiiil\ for ni;i\d;nl> S hour ;i\ u coiicciiir;ilioiis I SI5 csiinuilcs ucner;ill\ ni;ike
np ;i dccrc;isiim l'r;iclioii of lol;il o/one coiicciiir;ilion w il 11 iiicrc;isinu lol;il ozone
coiicciiir;iiioiis in llie c;isicrn I S. ;nul ;il iiili;in locnlioiis m ilic wesiern I S Trends m
h;ischiic o/one lc\cls snuucsied ;i risinu coiiirihiilion from ii;ilnr;il ;ind iiHerii;iIiiiii;iI
sources ihroimh ;ippro\ini;ilel> 2o lo Keeenll\. howe\er. llns irend h;is show n smns of
slow iim or e\ en rc\crsiim. possihk due lo dccrc;isiim T;isi Asimi precursor emissions
1-1
-------
1.1 Introduction
This Appendix reviews scientific advances in atmospheric ozone research relevant to this review
of the NAAQS for ozone and other photochemical oxidants and the related air quality criteria. Strong
emphasis is placed on new evidence concerning the contributions of ozone from natural and non-U. S.
sources. Ozone is one of a group of photochemical oxidants formed by atmospheric photochemical
reactions of hydrocarbons with nitrogen oxides in the presence of sunlight. Photochemical oxidants were
defined in the 1970 Air Quality Criteria Document as compounds found in the atmosphere that oxidize a
reference material such as potassium iodide that is not oxidized by atmospheric oxygen (NAPCA. 1970).
Other photochemical oxidants formed by photochemical reactions of hydrocarbons and nitrogen oxides
include nitrogen dioxide (NO2), peroxyacetyl nitrate (PAN), hydrogen peroxide, nitrous acid, and organic
peroxides. Close agreement between ozone measurements and the photochemical oxidant measurements
upon which the early NAAQS was based indicated that the contribution of these other oxidant species
was very small (NAPCA. 1970). Shortly after, measurements of photochemical oxidants as a class of
pollutant species became increasingly rare, and in 1979 ozone became the NAAQS indicator for air
pollutant photochemical oxidants. Ozone is the only photochemical oxidant other than NO2 that is
routinely monitored. The current state of scientific understanding concerning NO2 is addressed in the
2016 ISA for the Oxides of Nitrogen—Health Based Criteria (U.S. EPA. 2016c). Data for other
photochemical oxidants are generally derived from a few special field studies. Extensive national scale
data on temporal or geospatial concentration patterns of these other oxidants are scarce. Moreover, few
studies of the health and welfare impacts of other photochemical oxidants beyond ozone have been
identified by literature searches conducted for other recent ozone assessments (U.S. EPA. 2013. 2006).
For these reasons, discussion of photochemical oxidants in this document focuses on ozone.
Material in this Appendix is based primarily on a systematic literature review of sources,
chemistry, estimation methods, and concentration trends of ozone from natural and non-U.S. sources
following procedures described in Appendix 10. For context, brief summaries of recent trends in U.S.
anthropogenic ozone precursor emissions and ozone concentration trends from available U.S. EPA
databases are also included. In addition, winter ozone, halogen chemistry, satellite measurements, and
chemical transport modeling are identified as research areas in which progress had been made since
publication of the 2013 Ozone ISA (U.S. EPA. 2013). For these topics, brief summaries of the most
relevant developments are also provided.
The most commonly used ozone metrics for assessing the impacts on human health and
ecosystems, and the performance of atmospheric models are evaluated (Section 1.2). followed by a
discussion of new advances in our understanding of ozone sources and emissions (Section 1.3).
atmospheric chemistry (Section 1.4). the influence of meteorology and climate change on ozone
concentrations (Section 1.5). and measurements and modeling (Section 1.6). The Appendix also includes
a summary of ambient ozone concentrations throughout the U.S. through 2017, as well as an assessment
of concentration trends (Section 1.7). Section 1.8 discusses the latest developments in understanding
1-2
-------
background ozone and how it contributes to ambient concentrations. It includes background estimation
methods, as well as estimates of the contribution of background ozone to ambient ozone concentrations.
This section is followed by an appendix summary (Section 1.9).
1.2 Metrics and Definitions
1.2.1 Ozone Metrics
Several different averaging times are commonly used in ozone metrics. Each has had a role in
NAAQS and been used in studies of human health, ecological effects, and atmospheric model evaluation.
The choice of metric in a study depends on the study purpose.
1.2.1.1 Ambient Air Concentration Metrics
Ozone concentration metrics are generally based on measurements or estimates expressed as a
volume mixing ratio, with units of parts per million (ppm) or parts per billion (ppb). Technically, ppm
and ppb are not concentration units, which are defined as moles per unit volume and depend on
temperature and pressure. This distinction is generally acknowledged in the atmospheric science
literature. In contrast, the term mixing ratio is rarely used in the literature on health and vegetation effects
but is instead usually substituted with the term concentration, understood to be more broadly interpreted
as the amount of a substance in a fluid without distinguishing units. For this reason, the term
concentration is generally used instead of mixing ratio in this document to maintain consistency with its
use in the health and ecological effects literature. Mixing ratio is still used in the more technical
discussions of atmospheric sources and chemistry in Section 1.3 and Section 1.4.
• The daily max 1-hour avg (MDA1), daily max 8-hour avg (MDA8), and daily 24-hour avg
concentrations (DA24) are among the most widely used short-term air quality metrics in
epidemiologic studies.
• Seasonal and monthly averages of MDA1, MDA8, and DA24 are used for long-term metrics in
epidemiologic studies. Hourly ozone concentrations and longer term averages of these metrics are
also used for atmospheric model evaluation (Dennis et al.. 2010).
• Design values are used by the U.S. EPA to designate and classify nonattainment areas, as well as
to assess progress towards meeting the NAAQS. A design value is a statistic that describes the air
quality status of a given location relative to a particular NAAQS. The design values for the ozone
NAAQS are the 3-year avg of the annual 4th highest MDA8 ozone concentrations.
1-3
-------
1.2.1.2 Ecosystem Exposure Metrics
For ecosystem exposure, cumulative exposure indicators are frequently used that extend over
longer time periods, such as growing season or year (U.S. EPA. 2013). The W126, SUM06, and AOTx
exposure indices are metrics used for ecosystem exposure. Further details on these exposure indices are
provided in the 2013 Ozone ISA (U.S. EPA. 2013) and Section 8.13.1.
• The W126 exposure metric is a sigmoidally weighted sum of all hourly ozone concentrations
observed during a specified daily and seasonal time window. The sigmoidal weighting of hourly
ozone concentration is given by Wc = 1/(1 + 4.403c l2"c). where C is the hourly ozone
concentration in ppm.
• SUM06 is the sum of all hourly concentrations greater than or equal to 60 ppb observed during a
specified daily and seasonal time window.
• AOTx is the sum of differences between hourly ozone concentrations greater than a specified
threshold during a specified daily and seasonal time window. For example, AOT40 is the sum of
differences between hourly concentrations above 40 ppb.
1.2.2 Background Ozone Definitions
Use of the term background ozone varies within the air pollution research community. The most
widely used definitions and applications are described in this section. The term has generally been used to
describe ozone levels that would exist in the absence of anthropogenic emissions within a particular area
and has been broadly applied to every geospatial scale: local, regional, national, continental, and global.
For instance, on a local scale, ozone that originates from precursor emissions outside of a locality's
municipal boundaries could be considered background ozone in that locality. Similarly, on a national
scale, background ozone could be defined as ozone that is not formed from anthropogenic emissions
within national boundaries.
1.2.2.1 U.S. Background (USB) Ozone
In this document, the term U.S. background (USB) is used to assess background ozone. The USB
concentration is defined as the ozone concentration that would occur if all U.S. anthropogenic ozone
precursor emissions were removed.
• This definition helps distinguish the ozone that can be controlled by precursor emissions
reductions within the U.S. from ozone originating from natural and foreign precursor sources that
cannot be controlled by U.S. regulations.
• The distinction between U.S. anthropogenic and USB sources is not always straightforward, with
ambiguities or debate regarding U.S. anthropogenic methane (Fiore et al.. 2014). U.S.
anthropogenic emissions that have recirculated globally (McDonald-Buller et al.. 2011).
1-4
-------
international shipping and aviation (U.S. EPA. 2015). prescribed fires (U.S. EPA. 2015). and soil
emissions (Rasool et al.. 2016); see Section 1.3.2.1.
• As defined here, USB is a model construct that cannot be measured using ambient monitoring
data. This approach is consistent with the 2006 Ozone Air Quality Criteria Document [AQCD;
U.S. EPA (2006)1 and the 2013 Ozone ISA (U.S. EPA. 2013). which also assessed modeled
estimates of USB. The 2006 Ozone AQCD (U.S. EPA. 2006) included an extensive discussion of
alternative methodologies for estimating background ozone in its Annex AX3.9 and concluded
that background ozone concentrations cannot be obtained solely by examining measurements of
ozone obtained at relatively remote monitoring sites in the U.S. because of long-range transport
of ozone originating from U.S. anthropogenic precursors even at the most remote monitoring
locations. Reasons for not using measurement-based estimates of USB are explained in further
detail in Section 1.2.2.5. Reliance on atmospheric modeling for USB concentrations estimates
continued in the 2013 Ozone ISA (U.S. EPA. 2013).
1.2.2.2 Apportionment-Based U.S. Background (USBab)
Modeling approaches for estimating USB can be classified as either source-sensitivity or
source-apportionment approaches (see Section 1.8.1). USB was originally estimated using
source-sensitivity approaches (e.g., "zero-out" modeling). Apportionment-based USB (USBab) has been
defined as the amount of ozone formed from sources other than U.S. anthropogenic sources as estimated
via an apportionment technique (Dolwick et al.. 2015).
• The distinction between USB and USBab is important because apportionment techniques for
estimating USBab are designed to realistically treat nonlinear and nonadditive interactions of
USB and U.S. anthropogenic emissions that affect both production and destruction of ozone. In
contrast, source-sensitivity modeling approaches originally used for estimating USB are not
designed to address these interactions.
• USB and USBab are not the same quantity estimated with different approaches but are actually
estimates of conceptually different quantities. While USB is an estimate of ozone concentrations
that could be achieved if U.S. anthropogenic sources were eliminated, USBab is an estimate of
how much ozone can be attributed to background sources when those anthropogenic sources are
present. Differences in modeling approaches used to estimate USB and USBab are described in
Section 1.8.1.
1.2.2.3 U.S. Background (USB) Averaging Time
The averaging time of a USB estimate is intended to match the averaging time of the total ozone
concentration measured. For example, it would be inappropriate to estimate the USB contribution to an
MDA8 ozone concentration (see Section 1.2.1.1) using a seasonal mean USB estimate. This is because
meteorological conditions under which high anthropogenic ozone concentrations are produced differ from
those under which high USB ozone concentrations are produced (see Section 1.3 and Section 1.5.1).
• Estimates of USB on days with high MDA8 concentrations are more relevant for understanding
USB contributions on those days than are seasonal mean USB estimates.
1-5
-------
• Seasonal mean USB is more relevant for understanding source contributions to long-term average
concentrations.
• As discussed by Jaffe et al. (2018) and in Section 1.8.1. USB MDA8 estimates on specific days
are more uncertain than USB seasonal mean estimates, because of considerable daily variation
influenced by season, meteorology, and elevation.
1.2.2.4 Other Background Ozone Definitions
Other definitions besides USB have been used in previous U.S. EPA science assessments.
Although USB is emphasized in this document, research results based on North American background
(NAB) and natural background are also included. These terms were also widely used in the 2013 Ozone
ISA (U.S. EPA. 2013) and in earlier ozone assessments.
• NAB has been defined as the ozone concentration that would occur in the U.S. in the absence of
anthropogenic emissions in continental North America (U.S. EPA. 2013). NAB has also been
referred to as policy-relevant background (PRB) in earlier publications (U.S. EPA. 2007).
• Emissions-influenced background (EIB) has been defined as another measure of background
ozone estimated from source apportionment modeling approaches while including chemical
interactions with anthropogenic emissions (Lefohn et al.. 2012).
• Natural background ozone is defined as the ozone concentrations that would occur if all
anthropogenic emissions were removed worldwide. Processes that contribute to natural
background ozone include ozone transport from the stratosphere and ozone formed from
precursor emissions originating from wildfires, lightning, natural methane sources, plants, and
other natural VOC and NOx emissions (see Section 1.3).
1.2.2.5 Baseline Ozone
Baseline ozone is an alternative metric to USB or NAB that has been defined as the measured
ozone concentration at rural or remote sites that have not been influenced by recent, local emissions (Jaffe
et al.. 2018). In contrast to USB, baseline ozone is by definition directly measured.
• Baseline measurements are typically from monitors in locations that are minimally influenced by
local anthropogenic sources, and samples used as baseline measurements are limited to those
monitored during meteorological conditions consistent with the relative absence of local
contamination.
• Baseline ozone can include the ozone produced from U.S. emissions that circle the globe and may
also include effects of same-state emissions. An example of the latter would be ozone from U.S.
emissions near the West Coast or Gulf Coast that is transported over the Pacific Ocean or Gulf of
Mexico, respectively, and then transported back onshore.
• In some cases, sources that impact baseline ozone may not similarly impact ozone in populated
locations. For instance, baseline ozone measured on a mountaintop may include stratospheric
influences that are not representative of contributions in nearby lower elevation locations.
1-6
-------
• There are several reasons why baseline ozone measurements cannot be used as a proxy to
estimate USB ozone levels in urban areas. As previously described, baseline ozone can include
contributions from U.S. emissions. Additionally, baseline ozone monitors can be very distant
from urban sites, and ozone measured at the baseline site can be destroyed through surface
deposition or chemical reactions during transport from the baseline site to a downwind monitor.
In addition, atmospheric conditions may not favor transport of baseline ozone from the monitor
location to populated areas at lower elevations.
• Another reason why baseline ozone measurements cannot be used as a proxy for USB ozone
levels in urban areas is that meteorological conditions that favor mixing from the free troposphere
to ground level have strong ventilation and are not conducive to photochemical ozone episodes
that produce the highest urban ozone concentrations (see Section 1.5.IV Stratospheric intrusion
events are an exception (see Section 1.3.2V
• While baseline ozone measurements cannot be used directly to estimate USB ozone, baseline
ozone data are useful for evaluating the CTMs that are used to provide model estimates of USB
ozone.
1.3 Sources of U.S. Ozone and Its Precursors
U.S. tropospheric ozone (i.e., ozone present in the atmosphere below the stratospheric ozone
layer) is classified in this assessment as either being derived from U.S. anthropogenic sources or
background (USB). Anthropogenic ozone within the U.S. is further defined as the product of
photochemical reactions involving precursors derived from human activities. USB ozone, as defined in
Section 1.2.2.1. has a broader, more complex array of sources. These include natural precursor sources as
well as precursors transported from across U.S. borders from both nearby and distant locations within the
Northern Hemisphere.
The atmosphere can be divided into several distinct vertical layers, based primarily on the major
mechanisms by which they are heated and cooled. The lowest major layer is the troposphere, which
extends from the earth's surface to about 8 km above polar regions and to about 16 km above tropical
regions. Lying above the troposphere is the stratosphere, which extends from the top of the troposphere to
an altitude of about 50 km. The boundary between the troposphere and stratosphere is known as the
tropopause.
As with the 2013 Ozone ISA, the emphasis in this assessment is on the effects of "tropospheric
ozone," which refers broadly to ozone occurring throughout the total depth of the troposphere, consistent
with the usage in the primary scientific literature on ozone and climate. Within the troposphere, the lowest
sublayer is the planetary boundary layer, extending from the surface to about 1-2 km, which is most
strongly affected by surface conditions, including local emissions of ozone precursors. The portion of the
troposphere lying above the planetary boundary layer, in which atmospheric transport processes occur
over much larger spatial scales, is often referred to as the "free troposphere," with the "upper
troposphere" referring to the high-altitude portion of the free troposphere nearest the tropopause.
1-7
-------
In the context of this ISA, high concentrations of anthropogenic ozone may be localized to within
the atmospheric boundary layer above an urban area characterized by substantial precursor emissions.
Depending upon meteorological conditions, both urban precursors and ozone may be lofted into the free
troposphere allowing for downwind transport to rural or other urban areas.
Ozone derived from the stratosphere and from the reaction of internationally-transported
precursors in the upper troposphere can be drawn down into the lower troposphere through atmospheric
dynamics (i.e., vertical movement of large air masses between the stratosphere and the troposphere).
Figure 1-1 illustrates the complexities associated with attributing measured ground-level ozone to
particular sources.
The focus of this section is recent scientific findings concerning the sources of USB ozone. To
provide context for this discussion, updated information on U.S. anthropogenic ozone precursor emissions
and trends in those emissions is included.
Tropopause Folding
Aircraft Emissions
STRATOSPHERE
TROPOSPHERE
NOx CO CH4
Photochemistry 03
PAN
Lightning
NOx
scherrfistry
\J 3
Deposition
NOx VOCs
CH4 CO
[Trans porta uon
Agriculture and
Animal Husbandry
Wild and
Prescribed
FREE
TROPOSPHERE
Forests and Other
Ecosystems
«..
I #
m
4l
BOUNDARY
LAYER
Power
Generation
Landfill Gas
.
W® - -
© ©
Fossil Fuel Extraction
Source: Based on CCSP (2003), modified to include ozone sources and chemistry.
Figure 1-1 Major atmospheric processes and precursor sources contributing
to ambient ozone.
1-8
-------
1.3.1
Precursor Sources
Ozone formed in the troposphere is, primarily, the product of photochemical reactions between
nitrogen oxides (NOx) and carbon-containing compounds including carbon monoxide (CO), methane
(CH4), and volatile organic compounds (VOCs). This section summarizes current estimates of U.S.
anthropogenic precursor emissions by source type. Following this summary is a discussion of recent
findings concerning global/international and natural precursor emissions sources.
1.3.1.1 Ozone Precursor Emissions: Anthropogenic Sources and Trends in the U.S.
Figure 1-2 provides a visual summary of annual emissions by the largest U.S. sources of
anthropogenic NOx, CO, VOCs, and methane. Estimates for NOx, CO and VOC emissions are taken from
the publicly available versions of the U.S. EPA National Emissions Inventory [2014 NEI, Version 2; U.S.
EPA (2017)1. Methane is not one of the pollutants or pollutant precursors included in the NEI. However,
methane is an important greenhouse gas and is, therefore, included in the U.S. Inventory of Greenhouse
Gases and Sinks (U.S. EPA. 2016d). The NEI and U.S. GHG inventory do not share the same source
classification scheme; therefore, direct sector comparisons are not possible. Emissions of each precursor
are shown in Figure 1-2 and Figure 1-3 as a function of sector, as defined in the inventory from which the
data were derived.
The U.S. EPA also maintains a Trends database, beginning with 1970, that provides information
about criteria pollutant (or precursor) emissions trends for a set of aggregate categories that account for
major, or "Tier l"source types (U.S. EPA. 2019b). National emissions estimates for these categories are
derived from the NEI and are included in the Trends data set when an updated version of the inventory
has been finalized for public release. The 2014 NEI is the most recent inventory available to the general
public, with the 2017 NEI currently in development (due for public release in 2020). However, annual
emissions estimates for some of the relevant ozone precursors are currently available for mobile sources
and the electric utility sector. Mobile source emissions are calculated for the NEI using the U.S. EPA
Motor Vehicle Emission Simulator (MOVES) model (U.S. EPA. 2011). These values are available for
inclusion in the Tier 1 Trends data set in advance of the release of the official 2017 NEI. Electric utilities
continuously monitor and provide quarterly reports of emissions of NOx and SOx to U.S. EPA's Clean
Air Markets Program, as required under the Clean Air Act. These data (U.S. EPA. 2019a') were used to
provide an estimate of 2017 emissions from the electric utility sector for the trends data set. Figure 1-3.
showing U.S. Tier 1 precursor emissions trends since 2002, includes estimated NOx emissions by electric
utilities, as described, for 2017. Emissions of NOx, CO, and VOCs, as estimated by the MOVES model,
for 2017 are likewise included. Emissions by all other source categories are given through 2014, as taken
from the 2014 NEI.
1-9
-------
The development of the NEI is, primarily, a "bottom-up" approach, relying on activity and
emissions factors for developing emissions estimates for specific sources. For some sources, such as
electricity-generating units (EGUs) included in U.S. EPA's Clean Air Markets trading programs (U.S.
EPA. 2019d). continuous stack measurements provide reliable emissions values. However, for thousands
of sources that must be included in the NEI, there may be no uncertainty information for either or both the
activity and emissions factors upon which the emissions estimates are based. For example, emissions
factors for residential wood combustion are based on values found in the scientific literature and activity
levels based on information collected from limited state and regional surveys. Emissions from wood
combustion are sensitive to burning conditions, fuel-type, and other variables. Some emissions factors
may have been derived from data gathered using equipment or methods that were imprecise or were
reported without information about the combustion conditions or fuel type. Others may have been
collected with appropriate equipment but published without an accounting of experimental uncertainties.
U.S. EPA uses the best available data while acknowledging any known limitations in the resulting
emissions estimates. The WebFIRE database serves as the repository for recommended emissions factors
for criteria pollutants, criteria pollutant precursors, and hazardous air pollutants (HAP) for industrial and
nonindustrial processes, as well as the individual data values used to develop the recommended factors
and other related data submitted to U.S. EPA by federal, state, tribal, and local agencies, consultants, and
industries. For each recommended emissions factor and individual data value, WebFIRE contains
descriptive information such as industry and source category type, control device information, the
pollutants emitted, and supporting documentation (U.S. EPA. 2016e). There are no commonly accepted
standards for the reporting of uncertainty in emissions and related activity data; therefore, a rigorous and
consistent treatment of uncertainty in the NEI is not currently possible. While top-down methods can help
to constrain total emissions and may provide information on uncertainty, it is challenging to draw
conclusions about specific source sectors from these types of analyses except in cases where a single
sector dominates (Simon et al.. 2018V Many of the issues related to the analysis of uncertainty in the NEI
are discussed in Day et al. (2019).
1-10
-------
A) NOx (14,366 kTon/yr)
Commercial
Marine
Vessels
Non-Road
Equipment-
Diesel
Biogenics -
Vegetation and Soil
6%
9%
8%
Electric
Generation -
Coal
Combustion
10%
On-Road
Diesel
Heavy Duty
Vehicles
15%
On-Road non-Diesel
Light Duty Vehicles
17%
5%
Oil & Gas
Production
5%
Industrial
Boilers, ICEs-
Natural Gas
4%
VS.
IVljV.
Non-Road
[Other
17%
Equipment
- Gasoline
Residential-Natural
Gas Combustion
2%
2%
B) CO (72,353 kTons/yr)
On-Road non-
Diesel Light
Duty Vehicles
31%
Non-Road
Equipment-
Gasoline
16%
Wildfires
Other
11% .
Waste
Disposal
3%
w
Prescribed
Fuel Comb-
Residential-Wood
3%
Biogenics -
Vegetation
and Soil
9%
C) VOCs (55,630 kTon/yr)
D) CH4 (26,298 kTon/yr)
Vegetation and Soil (Biogenics)
mm
Other
Wildfires
Prescribed
Oil & Gas
Production
Petroleum Systems
6%
Coal Mining
8%
Other
9%
Non-Road
Equipment
- Gasoline
3% Consumer &
Commercial
Solvent Use
3%
Landfil s
Agriculture -
Animal
Husbandry
36%
On-Road non-
Diesel Light
Duty Vehicles
3%
Natural Gas
Systems
25%
Sources: (A)-(C) 2014 U.S. EPA National Emissions Inventory, Version 2 fll.S. EPA. 2016a 1 and; (D) 2016 U.S. Inventory of
Greenhouse Gases (U.S. EPA. 2016d1.
Figure 1-2 Relative ozone precursor emissions by U.S. sector: (A) nitrogen
oxides (NOx). (B) carbon monoxide (CO). (C) volatile organic
compounds (VOCs). Biogenic VOCs, which can be important in
the production of ozone in urban areas, are included for context.
(D) methane (CH4).
1-11
-------
Highway Vehicles
Highway Vehicles
50000
10000
40000
Fuel
Combustion
EG Us
Off-Highway
Equipment
Off-Highway
Equipment
30000
20000
Wildfires
10000
2002 2005 2008 2011 2014 2017
Inventory Year
2002 2005 2008 2011 2014 2017
Inventory Year
VOCs
Petroleum
& Related
Industries
Agriculture - Animal Husbandry
Highway
Vehicles
10000
Natural Gas Systems
Solvent
Landfills
Wildfires
Other
Coal Mining
Petroleum
Off - Highway
Equipment
Agriculture - Animal Husbandry
Natural Gas Systems
Landfills
Coal Mining
Petroleum Systems
Other
2002 2005 2008 2011 2014 2017
Inventory Year
A) NOx
12000
a
£
LLi
X
O
B) CO
60000
C)
TS
QJ
u
O
>
Leeend: NOx. CO. VOCs
Petroleum & Related Industries
Fuel Combustion - EGUs
Other Industrial Processes
Miscellaneous (w/o Wildfires)
Storage and Transport
— Chemical & Allied Product MFG
Waste Disposal & Recycling
— Fuel Combustion - Industrial
Highway Vehicles
Solvent
— Off-Highway Equipment
Metals Processing
Wildfires
— Fuel Combustion - Other
Sources: (A)-(C) U.S. EPA National Emissions Trends (U.S. EPA, 2019b) and; (D) the 2016 U.S. Inventory of Greenhouse Gases
(U.S. EPA. 20166).
Figure 1-3 U.S. anthropogenic ozone precursor emission trends. Sources
shown generate 90% or more of known emissions, excluding
biogenic sources, for the indicated precursor: (A) nitrogen oxides
(NOx), (B) carbon monoxide (CO), (C) volatile organic compounds
(VOCs), (D) methane (CH4). Not shown: "Other" NOx, CO, and VOC
emissions categories that, together, account for less than 10% of
total emissions for each precursor.
1-12
-------
1.3.1.1.1 U.S. Anthropogenic Nitrogen Oxides (NOx)
Anthropogenic NOx sources at local and regional scales within the U.S. have been recently
discussed in detail in the ISAs devoted to ecological effects of NOx, SOx, and PM (U.S. EPA. 2018) and
to the health effects of NOx (U.S. EPA. 2016c).
• Emissions of NOx within the U.S. decreased by 47% between 2002 and 2014. Figure 1-2
summarizes the main NOx emissions source categories included in the 2014 NEI (U.S. EPA.
2017). Highway vehicles are the largest source category of NOx emissions nationwide,
contributing 1 Tg N/year to total NOx emissions nationwide. Off-highway vehicles, electricity
generating-units (EGUs), other forms of stationary fuel combustion, and industrial processes each
contribute between 0.4 and 0.7 Tg-N/year to nationwide NOx emissions. Figure 1-3 shows the
steep decline in U.S. NOx emissions, primarily due to on-road vehicle emissions changes,
between 2002 and 2014. Estimated Tier 1 emissions of NOx have decreased by 47% between
2002 and 2014 (Figure 1-3).
1.3.1.1.2 U.S. Anthropogenic Carbon Monoxide
The Integrated Science Assessment for Carbon Monoxide (U.S. EPA. 2010) describes the sources
of anthropogenic carbon monoxide (CO) as primarily on- and off-road mobile emissions, followed by
prescribed burning. Wildfires and soils emit much of the remaining total.
• Overall, emissions at the national scale, between 2002 and 2014, have declined by approximately
30%. The 2014 NEI reports on-road mobile emissions at 31% of total U.S. CO emissions;
off-road at 16%; wildfires at 15%; prescribed fires at 12%; and soil emissions at 9% (U.S. EPA.
2017). The values reported for wildfires and soils are uncertain to a much larger degree than for
the other sources. Estimated Tier 1 emissions of CO have declined by 36% between 2002 and
2014 (Figure 1-3).
1.3.1.1.3 U.S. Anthropogenic and Biogenic Volatile Organic Compounds (VOCs)
The NEI includes estimates of biogenic as well as anthropogenic VOC emissions. At the national
scale, biogenic sources contribute substantially more to VOC emissions in the U.S. than do anthropogenic
sources. Biogenic VOCs can play an important role in urban ozone formation and are, therefore, included
in this discussion for context in this section. As described in the 2013 Ozone ISA, VOCs that are
important for the photochemical formation of ozone include alkanes, alkenes, aromatic hydrocarbons,
carbonyl compounds, alcohols, organic peroxides, and halogenated organic compounds. These
compounds range widely in photochemical reactivity and, consequently, atmospheric lifetimes. For
example, isoprene has an atmospheric lifetime of approximately 1 hour, whereas ethane has an
atmospheric lifetime of about 6 weeks. In urban areas, compounds representing all classes of VOCs,
along with CO, are important for ozone formation. In nonurban vegetated areas, biogenic VOCs emitted
from vegetation tend to dominate the VOC budget.
1-13
-------
• U.S. industrial and related VOC emissions have increased by approximately -20% since 2012,
while other anthropogenic emissions have declined over the same period. At the national scale,
emissions by biogenic sources dominate the U.S. inventory at 71%. These emissions are spatially
heterogeneous, having a greater effect on VOC concentrations in certain U.S. locations. Wildfires
emit 4% with the remaining 25% attributed to anthropogenic sources in 2014 (U.S. EPA. 2017).
Figure 1-3 shows the trends in Tier 1 emissions (i.e., not including biogenic VOCs) between 2002
and 2014. Overall, VOC emissions by Tier 1 sources have declined by 17% over that period.
1.3.1.1.4 U.S. Anthropogenic Methane
Methane, a major precursor for ozone at the global scale, is not included in the U.S. NEI.
Methane emissions are, however, reported in U.S. Inventory of Greenhouse Gases and Sinks (U.S. EPA.
2016d). The U.S. GHG Inventory and the NEI are not directly comparable because of differences in
source classifications, methods, and underlying assumptions. However, Figure 1-2 provides methane
trends as reported in the U.S. Greenhouse Gas Emissions inventory for the 2002-2016 time frame.
• Overall, total U.S. anthropogenic methane emissions decreased between 1990 and 2015. Recent
studies indicate that total U.S. anthropogenic methane emissions decreased by 16% between 1990
and 2015 (NASEM. 2018). However, the methane trends differed between the individual source
categories. The U.S. GHG inventory indicates that agriculture and natural gas systems are the
largest U.S. sources of methane. Emissions from landfills and coal mining have trended
downwards since the 2005-2008 time period. The agricultural emissions trend varied between
2002 and 2014 but has shown a notable increase since 2014. Petroleum systems were constant
between 2002 and 2011, increased between 2011 and 2014, then remained constant between 2014
and 2016. From 2002 to 2016, the inventory showed little change (-5%), in overall annual
estimated emissions.
1.3.1.2 Global and International Sources of Anthropogenic Ozone Precursors
Quantifying the emissions from sources that contribute to USB ozone represents a substantial
scientific challenge. As mentioned in Section 1.2.2.1. in the case of emissions from international sources
(i.e., anthropogenic and wildfire emissions from other countries, international shipping and aviation), and
of long-lived chemical precursors such as methane, identifying the specific sources and quantifying their
contributions to USB ozone is difficult under most circumstances. In some cases, satellites can capture
images of intact or partially intact emissions plumes of some precursors making it possible, using
back-trajectory modeling tools, to track these plumes to their origins. But, in most cases atmospheric
mixing and transport processes obscure the origins of those international emissions that can be detected
by remote sensing. In the case of chemical species that are stabilized by the low temperatures in the upper
troposphere, such as PAN, recirculation within the global atmosphere further confuses emissions
accounting.
1-14
-------
1.3.1.2.1 Global Methane
The 2013 Ozone ISA (U.S. EPA. 2013) reported an estimate by Zhang etal. (2011) of the effect
of anthropogenic methane emissions on global annual mean ozone concentrations at ground level of
-4-5 ppb. North American emissions of methane were described as uncertain but were considered to be a
small fraction of total anthropogenic input. Before the last assessment, ozone production derived from
methane oxidation was shown to be most prominent in regions with frequent vertical mixing and in
locations with NOxsaturated chemistry, such as southern California and the New York-New Jersey
region (Fiore et al.. 2008). In the same study, surface ozone was close to twice as sensitive to methane in
the planetary boundary layer (i.e., below about 2.5 km) than to methane in the free troposphere (Fiore et
al.. 2008). Model studies also indicate that the sensitivity of global tropospheric ozone to methane is
about 0.11-0.16 Tg ozone per Tg CFU/year (Zhang et al.. 2016; Fiore et al.. 2008).
• The methane concentrations over the U.S. are influenced by global methane sources. The
atmospheric methane abundance over the U.S. is influenced by global methane sources because
of the residence time for methane (NASEM. 2018). The atmospheric residence time for methane
is about a decade, allowing methane to be relatively homogeneously distributed around the globe
(NASEM. 2018). Therefore, the U.S. methane budget cannot be considered in isolation from the
global methane budget.
• The U.S. contributes approximately 20% to total methane in the atmosphere of the Northern
Hemisphere, and about 10% of total global methane emissions in recent years. For the 2003 to
2012 period, half of the total global methane emissions were attributed to Africa, South America,
and Southeast Asia combined, while the U.S. accounted for about one-tenth of the total global
emissions (Saunois et al.. 2016).
• Main global anthropogenic methane sources include agriculture and waste, fossil fuels, and
biomass and biofuel burning. An ensemble of studies attribute about 34% of the global
anthropogenic methane to agriculture and waste, 19% to fossil fuels (coal mining and oil and gas
industry), and 6% to biomass and biofuel burning between 2003 and 2012 (Saunois et al.. 2016).
The remaining total global methane emissions (i.e., about 41% of the total global methane
emissions) are generated by natural sources. These studies also estimated that global
anthropogenic methane emissions are about 328 Tg CH4/year using top-down inventories
(Saunois et al.. 2016). Top-down inventories use atmospheric observations within an atmospheric
inverse-modeling framework. Model results indicate that the global anthropogenic emissions of
methane decreased by about 15% between 1980 and 2010 (Zhang et al.. 2016). However, it
should be noted that methane emission estimates are highly uncertain due to measurement and
model uncertainties and not fully understanding the methane sources and sinks.
• Recent studies show that global mean methane concentrations are well over twice that of the
preindustrial period. The global mean methane concentration has nearly tripled between
preindustrial time and December 2017. Studies show that methane concentrations rose sharply
throughout the 20th century, then leveled off for a period of time beginning around 2000
(NASEM. 2018). Studies also reveal a sustained increase in atmospheric methane levels in the
1980s (by an avg of 12 ± 6 ppb/year), a slowdown in growth in the 1990s (6 ± 8 ppb/year), and a
general stabilization from 1999 to 2006 (Kirschke et al.. 2013). Between 2007 and 2010, methane
levels resumed rising (Kirschke et al.. 2013).
• In recent years, the total global mean methane concentration has increased annually by about
3.5 ppb. Between 2003 and 2012, the global mean methane concentration is estimated to have
1-15
-------
increased at a rate of 3.5 + 0.2 ppb/year (Saunois et al.. 2016). Some recent studies suggest that
the methane increases were mainly due to increases in fossil fuel production (e.g., coal and oil
and gas industry) and agricultural emissions, while other studies point to large uncertainties in
natural emissions (Van Dingenen et al.. 2018). Modeling studies also suggest that natural sources
contribute to the interannual variability of methane, while anthropogenic emissions, mainly
emitted in the Northern Hemisphere, have played a major role in the increase of methane
observed since 2005 (Badcr et al.. 2017).
• Global tropospheric ozone levels are enhanced when methane increases. Studies suggest that
increases in global methane since the 1800s have yielded higher levels of global tropospheric
ozone (and ground-level ozone) worldwide (NASEM. 2018). Studies indicate that there is an
approximately linear relationship between anthropogenic methane emissions and tropospheric
ozone, such that for every teragram per year decrease in methane emissions, ozone could decrease
by 11 to 15 ppt (Fiore et al.. 2008).
• Global methane abundance contributes to rising U.S. surface ozone during all months. Based on a
set of transient chemistry-climate model simulations between 2005 and 2100, the global methane
abundance contributes to rising surface ozone during all months, with the largest influence during
cooler months when the ozone lifetime is longer (Rieder et al.. 2018; Clifton et al.. 2014). These
simulations indicate that the sensitivity of the ozone mixing ratio to potential changes in global
methane abundance will be about 7-16 ppb over the northeastern U.S. and by about 12-19 ppb
over the intermountain western U.S. at the end of the 21st century (Clifton et al.. 2014).
1.3.1.2.2 International Emissions of Ozone Precursors
Ozone precursor emissions by countries that are "upwind" of the U.S. can contribute to U.S.
ozone. As described in earlier assessments (U.S. EPA. 2013. 2006) under certain atmospheric conditions,
precursors emitted by large cities and other sources can be lofted above the boundary layer into the
high-altitude zone referred to as the "free troposphere" (see Figure 1-1) where transport of atmospheric
constituents typically occurs more rapidly, over greater distances, than near the surface. NOx and ozone
have significantly longer atmospheric residence times in this colder atmospheric zone due to slower rates
of reaction than they have near Earth's surface (Rastigcicv et al.. 2010). Furthermore, NOx can react to
form reservoir species (i.e., species that can remain stable over very long distances) at these altitudes.
These reservoir species include PAN and similar compounds that become unstable at the warmer
temperatures of the lower troposphere, regenerating reactive NOx.
Large-scale atmospheric flows in the free troposphere can transport these pollutants and their
reaction products (i.e., ozone precursors and ozone formed within the plume) across continents and
oceans. Plumes from these international sources experience shear processes and dilution during advection
downwind. However, distinctive, coherent plumes have been observed by aircraft, sondes, and satellites
for a week or more. Downward mixing from the upper troposphere by way of other meteorological
processes, such as convective mixing, can then bring ozone down into the boundary layer.
International sources of ozone precursors do vary in significance for USB ozone, depending on
their relationship with the continental U.S. with respect to atmospheric dynamics and long-range
1-16
-------
circulation patterns. Asia, as described in previous IS As (U.S. EPA. 2013. 2006). has been an important
source of ozone precursors. Emitted due west of the continental U.S., across the Pacific Ocean, Asian
precursors have been identified as contributing to USB ozone in the western states, and in the central and
eastern U.S. under particular atmospheric transport conditions. Ozone precursor emissions from China
and other Asian countries have been estimated to have consistently grown in the 1990-2010 period
(Hocslv et al.. 20 IS). However, within the past decade, trends in NOx and CO emissions from China, the
largest source in Asia, have begun to level off, then decline.
• Satellite-derived NOx inventories for China show a rapid decline in emissions beginning in 2012.
Inventories based on bottom-up accounting of emissions using activity values and emissions
factors can be time consuming to develop. Emissions estimates for Asia are not currently
available beyond 2012. However, inventories derived from inverse modeling constrained by
satellite observations can be produced in near real time and are available for assessing Asian NOx
emissions rates. Ding et al. (2017) compared emissions estimates from four conventional
bottom-up inventories (Emissions Database for Global Atmospheric Research [EDGAR],
Multiresolution Emissions Inventory for China [MEIC], Regional Emissions Inventory in Asia,
Versions 2.1 and 2.2 [REAS 2.1 and REAS 2.2]) to four satellite-derived inventories
(DECSO-OMI, DECSOGOME2a, EnKFMIROC, EnKFCHASER) for the domain shown in
Figure 1-4. Panel A. While differences are present in the time-series results among all of the
inventories, a clear trend in emissions from China is present across the ensemble (see Figure 1-4).
Deviations in the temporal behavior of the various satellite-derived emissions are shown in
Panel D of Figure 1-4. in which emissions estimates from all of the inventories have been
normalized to their 2008 values. Chinese NOx emissions climbed annually until approximately
2012 before leveling off and then declining. In contrast, there is very little agreement among the
conventional NOx inventories for South Korea. South Korean satellite-derived emissions
estimates also differ significantly but demonstrate the same increasing, then decreasing trend
between 2010 and 2015, as shown in Panel B of Figure 1-4.
• Stringent air quality standards implemented in 2013 within China have markedly reduced national
emissions. Zheng et al. (2018) applied the bottom-up inventory model underlying the
Multi-resolution Emission Inventory for China (MEIC) to estimate anthropogenic emissions for
31 Chinese provinces. Figure 1-5 shows these estimates, aggregated to provide annual national
emissions values. The results of this accounting indicate that China's emissions of NOx and CO
have declined by 17 and 27%, respectively, while nonmethane VOCs grew by approximately
5 Tg/year between 2010 and 2017. Zheng et al. (2018) analyzed this inventory using index
decomposition analysis to identify the drivers behind these changes. The results of this analysis
indicated that stringent controls on power plant emissions were responsible for declines in NOx.
Improvements in combustion efficiency and oxygen blast furnace gas recycling in the industrial
sector accounted for reductions in CO emissions.
1-17
-------
A)
C)
10
i 4
i
0.5S
0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Mg N km'2 yn'
Mootood Ch*na
Aoo
OCCSO-OMI
DCCSO-GOME2* |
MMMOC
CnKF-CMASER
RE AS v2.1
RE AS v2J
CAPSS
EOGAA
, 'm'J
x:
% 1
/
• - OCCSO-OMI
• - DCCSO-GOME2*
- - EfXCF^MROC
• - EnXT-CHASER
— WAS v2.l
REAS v2.2
— mbc
— EDGAR
2002 2004 2006 2008 2010
Y«*r
2012 2014
D)
1.2
1.0
0.8
0.6
Mainland China
I • • D€CSOOM|
-- OCCSO-GOME2*
-• EnKF-MIROC
1 -- EnlCF-CHASER
— REAS v2,l
— REAS v2.2
/>::•. \
/ ~ • %
— MEIC
»/ -»\ •
||— EOGAR
ft/
X/
—
2002 2004 2006 2008 2010 2012 2014
Yetr
2000 2002 2004 2006 2008 2010
Year
2012 2014
Source: Adapted from Ding et ah (201 /). Reprinted with permission from the publisher.
Figure 1-4 Asian anthropogenic ozone precursor emission trends. (A) The
study domain, indicating annual NOx flux rates by location,
(B) Annual nitrogen oxides (NOx) emissions from eight
inventories over South Korea, (C) Annual NOx emissions from
eight inventories over China, and (D) Temporal deviations among
eight NOx emissions inventories, when normalized with respect to
2008 emissions.
1-18
-------
NMVOC
2010 2012 2014 2016 2010 2012 2014 2016 2010 2012 2014 2016
Note; Red = power sector emissions; yeilow = industrial emissions; green = residential emissions; blue = transportation emissions;
purple = solvent use. Lines marked with inverted triangles show the projected emissions trajectory, assuming activity levels were
held constant at 2010 levels; upright triangles indicate projected trajectories assuming pollution controls were held constant at 2010
levels.
Source: Adapted from Zheng et al. (2018:. Reprinted with permission from the publisher.
Figure 1-5 Anthropogenic ozone precursor emission trends derived using
the Multi-resolution Emissions Inventory for China emissions
model.
1.3.1.3 Natural Ozone Precursor Emissions
Ozone attributed to natural sources is formed through photochemical reactions involving natural
emissions of ozone precursors from vegetation, microbes, animals, burning biomass (e.g., forest fires),
and lightning.
1.3.1.3.1 Biogenic Nitrogen Oxide Emissions: Fertilized Soils
Biogenic sources of NOx were not discussed in either the 2013 Ozone ISA orthe ISA forthe
Oxides of Nitrogen—Health Criteria (U.S. EPA. 2016c. 2013). The topic was briefly mentioned in the
ISA for Oxides of Nitrogen, Oxides of Sulfur, and Particulate Matter—Ecological Criteria [2nd external
review draft; U.S. EPA (2018)1. Microbial nitrification (NFL -> NO;, ) and denitrification (NO; N2)
processes in soils produce NO, contributing to local and regional atmospheric NOx concentrations. Soil
NO emissions rates can be high enough to affect local and regional ozone concentrations under certain
circumstances (Vinken et al.. 2014). However, these rates are highly uncertain, being sensitive to biotic,
abiotic, and anthropogenic factors and their interactions, such as climate, soil moisture and temperatures,
and soil N content that can be altered by the addition of ammonium or nitrate fertilizers (Hall et al.. 2018)
Short, intense NOx pulses following agricultural fertilization activities and precipitation events have been
detected by satellite (Vinken et al.. 2014). Hickman et al. (2017) found a nonlinear response in NO soil
1-19
-------
emissions, as a function of increasing fertilizer application and crop species, but high spatial variability
among flux-rates led to significant uncertainty in the nature of the functional relationship. Soil moisture,
conversely, substantially reduces NO emissions, leading to added uncertainty due to inhomogeneities in
moisture content at the field scale (Hall et al.. 2018V
Biogenic emissions of NOx are estimated to contribute only a small part to national NOx
emissions, about 7.5% or 0.3 Tg N/year of the national total for all NOx of 4.0 Tg N/year nationwide
based on values reported in the 2014 NEI (U.S. EPA. 2018. 2017). This estimate was computed based on
2014 meteorology data from the Weather Research and Forecasting (WRF) model Version 3.8
(WRF 3.8), using the Biogenic Emission Inventory System, Version 3.61 (BEIS 3.61) model, based on
land use and vegetation data (U.S. EPA. 2016b). However, the fraction of total soil NOx due to fertilizer
applications versus from natural soils are not reported separately. U.S. EPA (2018) estimated fertilizer
application contributes -10-20% of global NOx emissions. Further details on estimating biogenic NOx
emissions are given in the NEI Technical Support Document (U.S. EPA. 2016b).
1.3.1.3.2 Biogenic Volatile Organic Compounds (VOCs)
Vegetation emits substantial quantities of VOCs, such as terpenoid compounds (isoprene,
2-methyl-3-buten-2-ol, monoterpenes), compounds in the hexanal family, alkenes, aldehydes, organic
acids, alcohols, ketones, and alkanes. Biogenic VOCs contribute to the mix of reactive organic precursors
in polluted areas, such as urban settings with high concentrations of NOx. As described in the 2013 Ozone
ISA (U.S. EPA. 2013). vegetation is a major source of highly reactive, relatively low molecular weight
organic compounds that contribute to the production of tropospheric ozone. Biogenic VOCs are
particularly important precursors in the southeastern U.S. because of that region's warm climate and
diversity of vegetation.
As discussed in the 2013 Ozone ISA, satellite measurements of formaldehyde (HCHO), produced
by the oxidation of isoprene and other VOCs, have been used to estimate biogenic VOC emissions
attributed to isoprene. Satellite-based and model techniques capture the spatial variability of biogenic
isoprene emissions in the U.S. reasonably well, with -40% uncertainty in satellite-derived isoprene
emissions, which is similar to the -50% error associated with model-based techniques (U.S. EPA. 2013).
• Biogenic VOCs fall into three major classes, with the smallest (isoprene) comprising one-third of
all emissions, followed by increasingly large and complex compounds. According to the 2014
NEI (U.S. EPA. 2017). the major chemicals emitted by plants are isoprene (30%) and other
terpenoid and sesquiterpenoid compounds (25%), with the remainder consisting of assorted
oxygenated compounds and hydrocarbons. These specific estimates of biogenic emissions of
VOCs were provided by the Biogenic Emissions Inventory System (BEIS) model Version 3.61
with data from the Biogenic Emissions Land-use Database (BELD) Version 4.1 and annual
meteorological data. However, other emissions models are available, such as the Model of
Emissions of Gases and Aerosols from Nature (MEGAN), which can also be used to develop
emissions inputs for global and regional modeling efforts.
1-20
-------
• VOC emissions from biogenic sources are estimated to be substantially larger than from
anthropogenic sources at the global and national scales. The annual rate of VOC emissions from
biogenic sources reported in the 2014 NEI v2 is -39 MT/year. By comparison, VOC emissions
from anthropogenic sources in the 2014 NEI v2 were -17 MT/year. (Note: wildfire-derived VOC
emissions [-2 MT/year] are counted as anthropogenic in the NEI. The effects of wildfire
emissions on USB ozone are discussed in Section 1.3.1.3.3.) Anthropogenic VOCs make up a
larger fraction of VOCs in certain urban areas such as Los Angeles.
• Differences in vegetation-related biogenic VOC emissions as a function of species, meteorology,
and geographic location introduce significant uncertainty in emissions estimates. Insufficient
measurement data and modeling limitations, as summarized in the 2013 Ozone ISA (U.S. EPA.
2013). contribute to significant uncertainty in estimates of natural emissions. Uncertainty
estimates can range from about 50% for isoprene under midday summer conditions in some
locations to about a factor of ten for some compounds and landscapes (Guenther et al.. 2000).
Most biogenic VOC emissions occur during the warmer seasons because of their dependence on
temperature and incident sunlight, but sesquiterpene emissions occur year-round. The BEIS and
MEGAN models have been shown to predict spatially similar emissions, but modeling results can
differ between them by about a factor of two, specifically for isoprene (Carlton and Baker. 2011).
which is the most abundant biogenic VOC globally and nationally.
• Recent modeling studies provide a range of estimates of the contribution of biogenic VOCs to
ozone mixing ratios, including max daily 8-hour avg (MDA8) concentrations. For instance,
Huang et al. (2013b) ran the multiscale Sulfur Transport and Deposition Modeling system at a
60-km grid scale for a time period in the summer of 2008, turned off biogenic emissions relative
to a base case simulation, and found that biogenic emissions have slight negative impacts over
most regions in Nevada, Idaho, Washington, and Oregon, due to the NOx sensitive regime in
those areas, and an estimated contribution of up to 15 ppb to MDA8 ozone from biogenic
emissions over Northern California and the California Central Valley. In a study by Zare et al.
(2014). a Danish Eulerian Hemispheric Model (DEHM) simulation for the year 2006 indicated
that biogenic VOCs enhanced the average ozone mixing ratio by about 11% over the land areas of
the Northern Hemisphere relative to a base case simulation for which BVOC were not turned off.
In additional sensitivity simulations. Zare et al. (2014) turned off all natural emissions of VOCs,
NOx, NH3, SO2, CH4, PM, CO, and sea salt collectively and individually relative to a base case
simulation. Zare et al. (2014) acknowledged that the sum of the individual sensitivities can be
different from the results of the collective zero-out simulation for the various regions of the world
due to nonlinearity in the processes of ozone formation chemistry. The discrepancies are different
from region to region because of different atmospheric chemical regimes in each individual
region. In a modeling simulation over the continental U. S. (CONUS domain) from May to
September 2011, Zhang et al. (2017) used the Comprehensive Air Quality Model with Extensions
(CAMx) Ozone Source Apportionment Tool (OSAT) with BEIS. The CAMx OSAT algorithm
attributes ozone production to VOCs only when ozone forms under VOC-limited conditions.
Zhang et al. (2017) found that domain-wide, biogenic VOCs can contribute on average 10-19%
to regional ozone formation, with higher contributions in the western U.S. and lower
contributions in the southeastern U.S. Ozone formation in the southeast is typically NOx-limited
due to intense BVOC emissions in that region of the U.S. Hence, CAMx OSAT attributes most of
the ozone in this region to NOx rather than VOCs.
• Model emissions estimates of isoprene are sensitive to estimates of photosynthetically active
radiation (PAR) and details concerning land use and species mapping. Isoprene makes up the
largest fraction of vegetation emissions. Hu et al. (2015) found that the GEOS-Chem atmospheric
model with the MEGAN v2.1 biogenic inventory reproduced isoprene observations at a site in the
U.S. upper Midwest to within model uncertainty given improved land cover and temperature
1-21
-------
estimates. One of the key uncertainties for modeling biogenic VOC emissions comes from the
estimation of PAR reaching the vegetation canopy. Zhang et al. (2017) found that using satellite
retrievals instead of modeled PAR reduced BEIS and MEGAN estimates of isoprene by an avg of
3-4% and 9-12%, respectively, but the simulations still overestimate observed ground-level
isoprene concentrations by a factor of 1.1 for BEIS and 2.6 for MEGAN. The satellite retrievals
in this study covered most of the continental U.S. Another key limitation for modeling biogenic
VOC emissions comes from a lack of complete and up-to-date land use and species mapping
information. Bash et al. (2016). using BEIS 3.61 with an updated canopy model formulation and
improved land use and vegetation representation, found better agreement between CMAQ
isoprene and monoterpene estimates compared with observations in northern California.
• Confidence in estimates of the contribution of biogenic VOC emissions to ground-level ozone
remains low. Biogenic VOCs are well-understood contributors to ozone formation. However,
uncertainties in the emissions of biogenic VOCs, limitations in the tools employed by
photochemical models, complete and up-to-date land use and species mapping information, and a
lack of a complete and detailed scientific understanding of biogenic VOC oxidation chemistry
make it difficult to accurately apportion the fraction of ozone due to anthropogenic VOCs versus
biogenic VOCs.
1.3.1.3.3 Landscape Fires
Landscape fires, including prescribed burning and wildfires, are complex sources of VOCs,
methane, CO, and NOx. Emissions from wildfire, in particular, are episodic but can have a significant
downwind impact on ozone concentrations in populated areas. New observations and modeling study
results identifying wildfire effects on observed ozone concentrations corroborate and extend the evidence
presented in the 2013 Ozone ISA (U.S. EPA. 2013).
• Wildfires contribute a few parts per billion (ppb) to seasonal mean ozone values in the U.S., but
episodic contributions may be as high as 30ppb. Wildfire emissions and their subsequent
photochemistry have highly variable impacts on ozone. Jaffe et al. (2018) and Jaffe and Wigder
(2012) concluded, on the basis of their synthesis of the results of more than 100 recent scientific
studies related to USB or other measures of background ozone, that wildfires contribute up to a
few ppb of ozone to seasonal mean surface concentrations in the continental U.S. However, on an
episodic basis, numerous studies demonstrate that wildfires may contribute up to 30 ppb to
MDA8 at specific times and surface locations. They noted that ozone production generally
increases up to 5 days downwind of emissions following a wildfire event. Ozone production
measured by the ratio AO3/ACO was highly variable, and typically higher when emissions were
transported and mixed with air from NOx-rich urban areas. A global modeling study (Mao et al..
2013) further supports the estimate of biomass burning's contribution to mean tropospheric ozone
concentrations given by Jaffe et al. (2018). although as discussed below, modeled estimates of
ozone production from fires remain highly uncertain.
• In-plume photochemistry converts NOx to PAN, a reservoir species for NOx, increasing its
capacity for affecting ozone concentrations far downwind of a fire. Lagrangian plume models and
box models have also been employed to better understand wildfire smoke plume chemistry and
have generally corroborated observational studies showing rapid in-plume NOx sequestration into
PAN, which provides a reservoir of reactive nitrogen with a long lifetime in the free troposphere.
Additional observational evidence of significant PAN production in wildfire smoke has emerged
recently (Fischer et al.. 2018; Busilacchio et al.. 2016).
1-22
-------
• Eulerian photochemical modeling remains highly uncertain in its estimates of the impact offire
emissions on ozone due to insufficient information on fuels, meteorological conditions influencing
smoke production, as well as existing model grid scales and the sufficiency of the photochemical
mechanisms available. Jaffe et al. (2018) discussed current methodologies of isolating the effects
on ozone from biomass burning. Such modeling continues to have high uncertainty, arising from
wide variation in the production of NOx and VOC among different fires. These variations in
precursor emissions arise from physical and chemical differences between fuel types, moisture
content, and meteorological conditions. Capturing wildfire dynamical processes, such as plume
heights, in models is also an area of active model development and improvement. Jaffe et al.
(2018) observed that reductions in the uncertainty in the estimates of USB ozone from fire
emissions will require developing or improving models that integrate the results of intensive field
studies with evaluation and comparison of Eulerian, Lagrangian, and statistical models.
• Statistical models based on observational data have also been used to identify the effects of fire
emissions on downwindMDA8 ozone concentrations. Statistical models of observed ozone data,
combined with colocated particulate matter measurements and satellite data from the NOAA
Hazard Mapping system, have shown some capability of identifying days with surface smoke
impacts and in estimating the amount of added ozone from wildfire smoke above what would be
expected on atypical smoke-free day under similar conditions. Jaffe et al. (2018) described a few
instances of such statistical model applications in the western U.S. where they identified a few
days with MDA8 ozone greater than 70 ppb that were impacted by ozone from wildfire smoke.
Newer studies have used statistical models to attribute the amount of ozone from biomass
burning. Liu et al. (2017b) used 10 monitoring sites in Kansas to apportion the contribution of
MDA8 ozone from prescribed range/pasture burning on elevated ozone days in April between
2001 and 2016. On days exceeding 70 ppb, they found the average ozone attributed to biomass
burning was 21 ± 9 ppb. Additionally, Lindaas et al. (2017). using surface monitoring data in
Colorado, estimated a contribution of up to 10 ppb ozone on each day. In a more comprehensive
analysis, including colocated PM2 5 measurements and nearby temperature measurements, Gong
et al. (2017) estimated a fire emissions impact on mean MDA8 ozone of 3-36 ppb (88% of the
monitors, within a 95% confidence interval). They also looked at the frequency of
smoke-impacted days when ozone monitors exceeded 70 ppb and found that the percentage of
impacted days ranged widely, but sites with the highest number of 70 ppb exceedance days
generally had fewer than 20% of days that were smoke-impacted. Brev and Fischer (2016) looked
at all smoke impact days but did not separate wild and prescribed fire from anthropogenic
biomass burning sources. The more recent study by Gong et al. (2017) suggested that the Brev
and Fischer (2016) analysis overestimates the effect of fire emissions on ozone production,
especially in coastal areas because they did not include a number of additional meteorological
variables, such as air mass transport patterns, in their statistical model.
• Satellite-based detection of fire emissions continues to be a work in progress. Many studies
employing satellite data to estimate NOx and other trace gases emitted by wildfires have been
conducted (Schreier et al.. 2015; Tanimoto et al.. 2015; Mebust and Cohen. 2014; Schreier et al..
2014; Ross et al.. 2013; Worden et al.. 2013; Mebust etal.. 2011; Tereszchuk et al.. 2011). A
recent study assessed emission coefficients of NOx via OMINO2 tropospheric column densities
and MODIS fire radiative energy for three fuel land types and reported markedly lower estimates
then previous estimates. These lower estimates are potentially due to underlying satellite retrieval
inputs (Mebust et al.. 2011). Also, researchers assessed the potential of the TES instrument to
characterize fire-derived PAN in the free troposphere over North America in summer (Fischer et
al.. 2018). but validation of the data was incomplete.
1-23
-------
1.3.1.3.4
Lightning Nitrogen Oxides (NOx)
Nitrogen oxide is produced when lightning causes the dissociation of N2 into nitrogen radicals
that subsequently react with molecular oxygen. The 2013 Ozone ISA (U.S. EPA. 2013) discussed the
highly uncertain U.S. estimates provided by Fang et al. (2010) for lightning-generated NOx (LNOx) of
-0.6 MT for July 2004, or ~ 40% of the anthropogenic emissions for the same period. However, Fang et
al. (2010) also estimated that -98% is formed in the free troposphere, limiting the direct effect on local,
ground-level ozone. Contributions to the surface NOx burden are low because most of this NOx is
oxidized to NOz species, including nitric acid (HNO3), and nitrous acid (HONO), peroxyacetyl nitrate
(PAN), peroxymethacrylic nitrate (MPAN), and peroxypropionyl nitrate (PPN), during downward
transport into the planetary boundary layer. The remaining 2% of LNOx is formed within the planetary
boundary layer. The 2013 Ozone ISA (U.S. EPA. 2013) also described the indirect effect that lightning
has on USB or NAB ozone by initiating wildfires.
• Eighty percent of NOx is generated by lightning in the upper troposphere, where it can have a
longer atmospheric residence time than NOx derived from ground sources. Although the LNOx
source is significantly smaller than combustion-derived NOx, it is produced in the upper
troposphere where the atmospheric lifetimes of NOx and ozone are long (Murray et al.. 2012).
Monks et al. (2015). in their synthesis of several studies, reported that LNOx is responsible for
more than 80% of upper tropospheric NOx and can also affect surface ozone levels through its
role in the determining the OH/HO2 ratio.
• LNOx shortens the atmospheric lifetime of CEL. As previously discussed, methane is an
important ozone precursor. By producing sudden bursts of excess OH, methane is removed from
the atmosphere to form methyl-peroxy radical. The methyl-peroxy radical, once formed, reacts
immediately to form ozone (Monks et al.. 2015).
1.3.2 Stratosphere-Troposphere Exchange (STE) Processes
1.3.2.1 Tropopause Folding
Tropospheric ozone derived from stratosphere-troposphere dynamics was described in detail in
the 2013 Ozone ISA (U.S. EPA. 2013). Stratospheric air rich in ozone can be transported into the
troposphere under certain meteorological circumstances, with maximum contributions at midlatitudes
during the late winter and early spring. In a process known as "tropopause folding," deep stratospheric
intrusions of ozone-rich air can occur; they form only episodically but have the ability to quickly and
directly reach the surface (U.S. EPA. 2013). These intrusions are often observed as "filaments" or
"ribbons" in water vapor satellite imagery or identified by meteorological data (e.g., relative humidity,
potential vorticity) or chemical (e.g., beryllium 7) tracers (U.S. EPA. 2013).
1-24
-------
The 2013 Ozone ISA (U.S. EPA. 2013) discussed the potential role of deep convection, another
form of stratosphere-troposphere exchange, as a mechanism for transporting stratospheric ozone into the
upper troposphere. The 2013 Ozone ISA noted the study of Tang et al. (2011). which through modeling
estimated that deep convection penetrating the tropopause increases the stratospheric-to-troposphere
ozone flux by 19% annually in the Northern Hemisphere, with greatest impacts occurring in the summer
months (49% in June). While the 2013 Ozone ISA (U.S. EPA. 2013) highlighted studies showing the
influence of stratospheric-tropospheric exchange to surface ozone, it did not estimate STE's impact on
USB ozone.
• Deep stratospheric intrusions are common in the western U.S., impacting high elevation locations
during the springtime. The incidence of tropopause folds is greatest in the early part (late winter
and spring) of the year when synoptic-scale midlatitude cyclones are most active, occurring near
upper level frontal zones where Rossby wave breaking is prevalent (Langford et al.. 2017;
Skerlak et al.. 2015; U.S. EPA. 2013; Lin et al.. 2012a).
• Stratospheric intrusions can be observed with lidar and other tools. Lang ford et al. (2015) used
lidar measurements and modeling results to estimate stratospheric influence of up to 30 ppb
during high surface ozone events around the Las Vegas, NV area during the 2013 Las Vegas
Ozone Study (LVOS). Meteorological variables and ozone data from the high-resolution NASA
MERRA-2 reanalysis data set were used to identify stratospheric intrusion events over Colorado
that occurred during the spring of 2012 (Knowland et al.. 2017).
• Stratospheric intrusions can be simulated with global chemistry models, although uncertainties
remain. The GEOS-Chem and GFDL-AM3 global chemistry models (Zhang et al.. 2014; Lin et
al.. 2012a) have successfully simulated springtime deep stratospheric events affecting
high-elevation sites in the western U.S. The GEOS-Chem simulations showed consistent
springtime contributions of stratospheric ozone between 8.8 and 9.4 ppb, with contributions of up
to 15 ppb during intrusion events. The AM3 model estimated contributions from stratospheric
ozone ranging from 17 to 40 ppb during the high surface ozone events from a model simulation
of the spring of 2010. However, AM3 is thought to overestimate ozone contributions from the
stratosphere (Lin et al.. 2012b).
• Stratospheric intrusions can lead to spikes in hourly and daily ozone concentrations, or smaller
increases over several days. Deep stratospheric intrusions have been shown to directly reach the
ground surface, albeit infrequently. Intrusions often extend into the mid troposphere over longer
timescales (up to 2 weeks) and may mix downward and affect surface ozone concentrations
(Stohl et al.. 2000). For example, the influence of stratospheric intrusions has been seen in
populated areas like Boulder, CO (Langford et al.. 2009). which showed ozone concentrations as
high as 100 ppb in 1-minute data. Stratospheric intrusions can lead to ozone spikes seen in hourly
and daily data (Langford et al.. 2009) or to smaller ozone increases over several days (Lin et al..
2012a).
• Quantifying the contribution of STE to surface ozone remains challenging and is a source of
uncertainty in estimating USB ozone. Stratosphere-troposphere exchange of ozone has been
observed using ground measurements, in situ aircraft or balloon measurements; through remote
sensing (lidar, satellite); identified with reanalysis data; and modeled via chemical transport and
global chemistry models. However, STE's contribution to USB ozone remains hard to quantify.
As previously mentioned, the AM3 global chemistry model (Lin et al.. 2012b) has been used to
estimate the stratospheric ozone contribution from deep intrusion events to be between 17 and
40 ppb at high surface ozone sites during springtime in the western U.S. Stratospheric intrusion
1-25
-------
events reaching the surface have less influence on surface ozone during the summer months when
total ground-level ozone concentrations tend to be highest.
1.3.2.2 Deep Convective Mixing
Since the previous assessment, studies of the dynamics within thunderstorm anvil clouds has
revealed that deep convection can entrain stratospheric ozone and draw it down into the upper
troposphere. The Deep Convective Cloud and Chemistry (DC3) (Barth et al.. 2015) aircraft field
campaign over the central U.S. in May and June of 2012 identified this process using in situ
measurements (Huntricscr et al.. 2016; Pan et al.. 2014). Pan et al. (2014) observed in situ ozone mixing
ratios as high as 150 ppb in the upper troposphere adjacent to the storm cloud edge. They postulated that
these high concentrations could be the result of the dynamical response to tropospheric air overshooting
the tropopause, with stratospheric air being mixed down into the upper troposphere and wrapping around
the cloud edges of the thunderstorm outflow. The high ozone concentrations found at the storm edges
were anticorrelated with mixing ratios of measured CO, indicating the stratosphere as the source of the
ozone-enriched air. The study found ozone enhancement in the upper troposphere near storm cloud edges
on numerous flight sample cases that indicated the prevalence of the deep convection
stratospheric-tropospheric exchange (STE) mechanism during the 2012 field campaign. Although the
studies of Pan et al. (2014) and Huntrieser et al. (2016) provided observed data of deep convection
leading to the downward flux of stratospheric air into the troposphere, the authors did not estimate the
contribution deep convection made to USB or other measures of background ozone at the surface.
1.4 Ozone Photochemistry
The general photochemistry of tropospheric ozone is well understood and described in previous
U.S. EPA integrated science assessments, criteria documents, and textbooks (Seinfeld and Pandis. 2006;
Finlavson-Pitts and Pitts. 2000). The photochemical reactions responsible for producing tropospheric
ozone are described in substantial detail, including an explanation of the nonlinearity in ozone production
as a function of the ratio in the local concentrations of NOx and VOCs, in Annex AX2 (Physics and
Chemistry of Ozone in the Atmosphere) in Volume II of the 2006 Air Quality Criteria for Ozone and
Related Photochemical Oxidants (U.S. EPA. 2006). The 2013 ISA for Ozone and Related Photochemical
Oxidants provides a succinct update on information concerning some of the finer points of smog
chemistry (U.S. EPA. 2013). Those discussions will not be repeated, here.
To briefly summarize the extensive discussion found in the 2006 AQCD for Ozone (U.S. EPA.
2006). ozone and the related atmospheric oxidants, peroxyacetylnitrate (PAN) and hydrogen peroxide
(H2O2), are products of the complex set of reactions that comprise the oxidation of carbonaceous
precursor gases in the presence of NOx.
1-26
-------
Photochemical oxidation involving carbonaceous precursors (VOCs) and NOx, sometimes known
as "smog chemistry," is rapid in the presence of strong solar radiation, leading to high concentrations of
ozone in urban settings during summertime stagnation events. The first step in the production of the
hydroxyl radical (OH), primary oxidant for both NOx and VOCs, is photolysis of an ozone molecule to
produce highly reactive 0(' D). 0(' D) promptly reacts with a water molecule to form 2 OH radicals. Past
this step, the competition between VOC and NOx for reaction with OH becomes the defining process
determining local ozone concentrations.
The optimum ratio between VOC and NOx for efficient ozone production will depend on two
central properties of the local airshed: the reactivity of the local VOC mixture with OH, and the ratio of
the concentration of ambient VOC with the concentration of NOx. Some volatile compounds react more
slowly than others, and the local VOC composition will be a function of the local and upwind VOC
sources. Without detailed information about the composition of the local ambient VOC mixtures, ambient
ozone concentrations cannot be reliably calculated a priori. In the absence of adequate measurements of
the composition and concentration of local VOCs, air quality scientists have developed a number of
strategies for estimating local VOC reactivity (U.S. EPA. 2006). The maximum (optimal) rate of ozone
production for an airshed with a given mix of VOCs is associated with a specific ratio of ambient VOC
concentration to ambient NOx concentration. When ozone production is limited by the concentration of
either NOx or VOC (i.e., NOx-limited or VOC-limited), this status reflected as a deviation away from the
ideal ozone production ratio for the airshed. These indicators are discussed in greater detail in Seinfeld
and Pandis (2006); Finlavson-Pitts and Pitts (2000) and U.S. EPA (2006).
The smog photochemical mechanism differs greatly from the chemistry of stratospheric ozone
formation or of ozone formed by lightning. The former requires the high-frequency solar ultraviolet
radiation present above the troposphere to photolyze molecular oxygen into reactive oxygen atoms. These
atoms react with molecular oxygen to form ozone. In the case of lightning-produced ozone, the high
voltage electric discharge provides the energy sufficient to ionize molecular oxygen, leading to ozone. As
discussed previously, NOx is also produced by lightning by ionizing molecular nitrogen.
Since the 2013 ISA, developments in ground-level ozone chemistry include the publication of
studies that applied ambient ozone measurements and modeling to the examination of potential causes of
unexpectedly high ozone concentrations observed during winter in western U.S. mountain basins, and
new work concerning the role of marine halogen chemistry in depleting boundary-layer marine and
coastal ozone concentrations. Chemistry and emissions associated with these processes are not included in
all models, adding to uncertainty in the evaluation of USB ozone at sites impacted by ozone that has been
transported out of mountain basins and over marine environments.
1-27
-------
1.4.1
Winter Ozone in Western Intermountain Basins
Ordinarily, ozone is a spring/summer/fall pollutant with the highest annual MDA8 levels
typically occurring on hot, sunny, stagnant days associated with summer weather conditions. As first
described in the 2013 Ozone ISA (U.S. EPA. 2013). high ozone levels during winter conditions have been
observed in two western U.S. intermountain basins with relatively high levels of anthropogenic precursor
emissions from oil and gas activity: Utah's Uinta Basin (UB) and Wyoming's Upper Green River Basin
(UGRB). These high ozone episodes date back to at least the winter of 2005 in the UGRB and the winter
of 2009 in the UB (Hclmig et al.. 2014; Schnell et al.. 2009).
• High wintertime ozone events continue to occur in the Uinta and Upper Green River Basins.
Winter ozone levels in the UB and UGRB have been measured as high as 150 ppb (1-hour avg) or
greater (Hclmig et al.. 2014; Rappengliick et al.. 2014). For comparison, max 1- and 8-hour ozone
levels in the winter of 2013 in the UB exceeded that of summer ozone levels of the Los Angeles
basin (Helmig et al.. 2014). a location that has historically experienced some of the highest
summertime ozone episodes in the U.S. In the winter of 2008, the UGRB observed MDA8 values
above 75 ppb 14 times (Schnell et al.. 2009). and in the winter of 2013 the UB experienced
39 days with MDA8 values greater than 75 ppb at individual monitoring stations (Helmig et al..
2014).
• Wintertime mountain basin high ozone episodes occur on cold winter days with low wind speeds,
clear skies, substantial snow cover, extremely shallow boundary layers driven by strong
temperature inversions, and substantial ozone precursor emissions activity from the oil and gas
sector. Wintertime inversions with low winds are sometimes referred to as "cold pool events," or
more specifically, "valley cold pool events" which are defined as an inversion below the
maximum crest height of the surrounding mountains coupled with average wind speeds beneath
the inversion top that are less than 5 m/s (Ahmadov et al.. 2015). These inversions, which trap
and concentrate local anthropogenic precursor emissions, can last several days or longer until
advection or turbulent mixing breaks them up (Ahmadov et al.. 2015). During these events, the
strong inversion isolates the local air mass from overlying layers of the atmosphere (no mixing).
Therefore, there is little to no influence from upwind emissions sources. Large sources of local
precursor emissions drive the ozone episodes during these cold pool events. High ozone episodes
during valley cold pool events have not been observed in areas without oil and gas sector activity.
Snow cover enhances the strength and persistence of the surface inversion layer and contributes
to ozone formation photochemistry by enhanced photolysis rates [due to the high albedo of the
snow surface; Ahmadov et al. (2015); Field et al. (2015); Edwards et al. (2014); Rappengliick et
al. (2014); Warneke et al. (2014)1. The relatively snow-free conditions during the winter of 2012
in the UB were not accompanied by high ozone events, but the cold pool conditions during the
snow-covered winter of 2013 resulted in a number of days where MDA8 values measured above
75 ppb (Ahmadov et al.. 2015).
• The chemistry driving wintertime ozone episodes seems to be different from the chemistry
driving summertime ozone episodes in terms of radical sources, as seen in measurements and by
modeling studies. Ozone production involves the hydroxyl (OH) radical, which in summer is
primarily formed from the photolysis of pre-existing ozone and subsequent reaction of one of the
products of this reaction, the electronically excited state atomic oxygen (0|1D |). with water
vapor. There is typically less solar radiation and water vapor during the winter, which is why
Edwards et al. (2014) saw a 15- to 60-fold decrease in OH production from this pathway (relative
to summer) when modeling the UB high ozone episodes for the winter of 2013. Edwards et al.
1-28
-------
(2014) found that the dominant source of radicals was from the photolysis of carbonyl
compounds associated with the high VOC emissions from oil and gas activity in the basin during
these episodes. Like Edwards et al. (2014). Ahmadov et al. (2015) suggested that VOC
photochemistry is an important source of radicals, including those formed from primary and
secondary formaldehyde photolysis, as well as from photolysis of dicarbonyls and hydroxy
ketones. Sensitivity studies show the ozone formation regime during the 2013 episodes in the UB
was VOC-limited (Ahmadov et al.. 2015). In the UGRB, Rappengliick et al. (2014). found that
the dominant source of OH production for the winter of 2011 was nitrous acid (HONO)
photolysis with minor pathways of production from alkene ozonolysis and formaldehyde
photolysis. Rappengliick et al. (2014) suggested the HONO is formed through nitric acid (HNO3)
produced during the atmospheric oxidation of NOx deposited onto the snow surface where it
undergoes photo-enhanced heterogeneous conversion to HONO as well as combustion-related
emissions of HONO. However, Edwards et al. (2014) found that HONO was not present in high
concentrations and, therefore, could not be a major source of OH production during winter ozone
episodes in the UB. Oil and gas extraction is the only major source of anthropogenic emissions in
the remotely located UGRB. These emissions include NOx from compressors and drill rigs and
methane and nonmethane hydrocarbons (VOCs) from wellhead production equipment
(Rappengliick et al.. 2014).
• Oil and gas sector impacts on ambient ozone levels extend beyond wintertime ozone episodes.
Recent occurrences of high wintertime ozone episodes and initial investigations into the
anthropogenic emissions and the chemistry driving these events indicate the importance of future
research to accurately quantify the role of increasing oil and gas sector emissions on ambient
ozone in the western U.S. Modeling studies summarized by Ahmadov et al. (2015) indicated
enhancements of 5-10 ppb to summertime 8-hour ozone concentrations that are attributed to oil
and gas extraction activity in various locations across the U.S. Cheadle et al. (2017) analyzed
precursor species measurements and meteorology data including back trajectories in the northern
Front Range in Colorado to estimate ambient ozone enhancement attributable to nearby oil and
gas activity and found that on specific summer days oil- and gas-related precursor emissions
could contribute locally up to 30 ppb ozone.
1.4.2 Halogen Chemistry
Multiphase processes have been associated with the release of reactive halogen species from
marine aerosol particles. The atmospheric chemistry of halogens involves compounds containing
chlorine, bromine, or iodine which can react among themselves and with other species and can be
important for tropospheric ozone destruction (U.S. EPA. 2013).
• Additional studies have further resolved the influences of halogen chemistry on ozone mixing
ratios. Ozone mixing ratios and deposition velocities over the ocean vary with atmospheric
turbulence and seawater chemical composition. The sea-to-air movement of chemical species
containing halogens like bromine, iodine, and chlorine affects the ozone photochemistry in the
atmosphere above the oceans. For example, photolysis and oxidation of halogen-bearing species
can release iodine and bromine, which can catalytically react with ozone to reduce ozone levels
over the ocean (Sarwar et al.. 2015). Tuite et al. (2018) measured iodine monoxide (IO) during
periods when low ozone (<25 ppb) air masses originating over the Gulf of Mexico flowed
onshore near Galveston, TX. Tuite et al. (2018) compared these measurements to a CAMx model
simulation that incorporated halogen chemistry and concluded that iodine is the most influential
1-29
-------
halogen in the Texas gulf coast area. The analyses of their measurements and model simulations
indicate iodine chemistry played a role in keeping ozone mixing ratios low in the relatively clean
offshore air that flowed onshore during the study period (Tuite et al.. 2018).
• Ozone is sometimes overpredicted along marine coastlines in photochemical model simulations.
Incorporating marine halogen chemistry into modeling studies improved agreement with
observed ozone in some circumstances. Sarwar et al. (2015) incorporated enhanced ozone
deposition and marine halogen chemistry involving photolysis of higher iodine oxides into a
photochemical model (hemispheric CMAQ) simulation and found that including these reactions
improves ozone model performance by reducing ozone levels to better compare to observations
near marine environments in the Northern Hemisphere. Sarwar et al. (2015) found enhanced
deposition reduces mean summer-time surface ozone by -3% over marine regions in the Northern
Hemisphere. Halogen chemistry without the photochemical reactions of higher iodine oxides
reduces surface ozone by ~15% whereas simulations with the photochemical reactions of higher
iodine oxides indicate ozone reductions of -48%. Over most terrestrial regions near the coast,
ozone mixing ratios are reduced by 2-4 ppb due to halogen chemistry without the photolysis of
higher iodine oxides. Gantt et al. (2017) incorporated the same detailed iodide-mediated ozone
deposition and marine halogen chemistry as Sarwar etal. (2015) to a finer (CMAQ) domain over
the continental U.S. as well as a parameterized version of the marine halogen chemistry to
preserve computational time. The parameterized version was applied as a first-order ozone loss
rate over oceanic grid cells as a function of atmospheric pressure. Gantt et al. (2017) did this for
the lateral boundary conditions generated by the hemispheric model feeding the regional scale
model as well as for the regional model simulation over the continental U.S. domain and
compared the results to ambient air measurements. Including the marine halogen processes in the
model improved overpredictions of surface ozone along the coast and over the open ocean,
achieving reductions in bias of 2-3 ppb for the majority of the sites along the Gulf and Atlantic
coasts, but exacerbated underpredictions of high surface ozone in some near-coast areas like
California's Central Valley and the urban areas of Washington, DC and New York City (Gantt et
al.. 2017). Many previous modeling studies which characterize background ozone did not include
a complete treatment of marine halogen chemistry (Emery et al.. 2012; Zhang etal.. 2011) and
therefore may overestimate background ozone transported over marine environments.
• Halogen marine chemistry can play a role in coastal urban air quality. The ocean is a natural
source of halogenated compounds which when released to the atmosphere can undergo photolysis
and oxidation to release reactive chlorine, bromine, and iodine radicals. In marine environments
near coastal cities with polluted urban air, gas-phase chlorine emissions (Ch and HOC1) and
chloride from sea salt can increase ozone mixing ratios by releasing NO2 from photolysis of nitryl
(CINO2) as well as through the oxidation of VOCs by chlorine radicals. In a 4-km photochemical
model simulation incorporating marine halogen chemistry over Los Angeles, Muniz-Unamunzaga
et al. (2018) saw improved regional/coastal air quality predictions compared with measurements.
Some earlier CINO2 modeling papers showed that CINO2 can increase ozone formation in winter
by up to 13 ppb and in summer by up to 6.6 ppb, although typical ozone increases were generally
below 2 ppb (Sarwar etal.. 2012; Simon et al.. 2009). While photolysis of CINO2 can lead to the
formation of ozone, Muniz-Unamunzaga et al. (2018) found that the chemistry of chlorine-,
bromine-, and iodine-containing compounds together have a net reduction effect on surface ozone
concentrations, with the reduction being larger near the coast and smaller farther inland. In terms
of the impact of halogen chemistry on NOx, which is important as a precursor to ozone, the effect
of halogen chemistry on NO2 varies by emission source distribution; however, Muniz-
Unamunzaga et al. (2018) saw that NO2 concentrations generally increased over nonurbanized
areas and the ocean and decreased in downtown Los Angeles when halogen chemistry was
incorporated into the model.
1-30
-------
• Halogen chemistry depletes ground-level ozone directly by reaction with bromine and iodine
radicals and indirectly by changing the budget and balance of important atmospheric oxidants like
NOx and HOx (Muniz-Unamunzaga et al.. 2018; Stone et al.. 2018). The research of Muniz-
Unamunzaga et al. (2018) supports the finding that in polluted coastal areas like the megacity of
Los Angeles, halogen chemistry can shift the NOx partitioning to NO, while in nonpolluted areas
with high concentrations of halogens, the reaction between XO and NO (where X = I or Br) shifts
the balance to NO2. Muniz-Unamunzaga et al. (2018) saw a decrease in HO2 due to halogen
chemistry with a small increase in diurnal mean OH concentration for an overall decrease in HOx
radicals. Likewise, Stone et al. (2018) saw a substantial decrease in HO2 with the inclusion of
halogen chemistry and a marginal increase in OH concentrations at certain locations but an
overall decrease in OH and HOx globally.
1.5 Interannual Variability and Longer-Term Trends in
Meteorological Effects on Anthropogenic and U.S.
Background (USB) Ozone
In addition to the quantities of ozone precursors emitted into the atmosphere by human activities
and natural sources, temperature, wind patterns, cloud cover, and precipitation also very important
variables in the production of atmospheric ozone (Nolte et al.. 2018). The 2013 Ozone ISA highlighted
the importance of meteorology on ozone formation (i.e., temperature dependence, the magnitude of solar
radiation, and the mixing/transport of ozone and its precursors). Meteorology is, therefore, an important
factor in the formation and transport of USB ozone. The 2013 Ozone ISA explained that multiyear trends
in U.S. ozone concentrations are influenced by the number of synoptic-scale and mesoscale stagnation
events, which vary from year-to-year, often making it difficult to evaluate the progress and effectiveness
of emissions reduction programs. Since the 2013 Ozone ISA, additional studies have examined the role of
meteorological effects on ozone and the year-to-year trends in ozone concentrations. Large-, regional-,
and local-scale atmospheric circulation patterns have been shown to influence both observed U.S.
background ozone and the local production of ground-level ozone. More recently, Nolte et al. (2018)
described emerging, robust evidence that the effects of climate warming on meteorology are increasing
ground-level ozone concentrations.
Large-scale meteorology patterns influence USB ozone in several ways, including the likelihood
of the occurrence of deep stratospheric intrusions events in the western U.S., the transport of Asian
pollution to the U.S., and regional temperature and precipitation patterns which can influence the
frequency and distribution of wildfires emissions of VOC precursors from vegetation and
combustion-derived NOx. During localized stagnation events conducive to ozone production,
ground-level ozone concentrations can be further influenced by regional and largescale meteorology
patterns or by regional-scale background ozone aloft being mixed down to the surface at urban sites.
Large-scale meteorology patterns help create the local-scale conditions that are conducive to
photochemical production of the ground-level. Example conditions include stagnation events associated
1-31
-------
with high temperatures and high ozone concentrations versus cool, wet meteorological conditions
associated with lower ambient ozone concentrations.
Largescale atmospheric circulation patterns are also subject to variability on annual and decadal
scales, which is reflected in the patterns of regional- and local-scale U.S. ground-level ozone
concentrations.
1.5.1 Meteorological Effects on Ozone Concentrations at the Ground Level
Meteorology at the regional and local scales establishes the chemical conditions that govern the
formation of ozone. Meteorological variables of importance at these scales include temperature, relative
humidity, wind speed, and precipitation. Synoptic-scale circulation (i.e., meteorological processes at
scales on the order of 1,000 km) are particularly important in determining ozone formation at regional and
local scales.
• Ozone was found to be strongly correlated to meteorology over the Intermountain West. Reddv
and Pfister (2016) found that surface ozone in the western U.S. is well correlated with the
500 millibar (mb) pressure level height. The study showed that the July mean MDA8 ozone
increased when the mean July 500 mb height also increased. Over the western U.S., increases in
the 500 mb level are often associated with weather (clear skies, low wind) that is conducive to
ozone formation. By using the 500 mb height variable to detrend and correct for the influence of
meteorology, the study found that July MDA8 ozone has steadily decreased from 1995 to 2013 in
the Wasatch Front area surrounding Salt Lake City, UT. Over the same time period, a general
increase in July MDA8 ozone was found along the Front Range (Denver area) of Colorado. The
study hypothesizes that ozone increases in Denver areas may be the result of emissions associated
with population growth and/or emissions from the increased activity of nearby oil and gas
development.
• Summertime ozone in the eastern U.S. and Midwest are affected by synoptic-scale meteorology
patterns. Shen et al. (2015) quantified the sensitivity of MDA8 ozone in the eastern U.S. and
Midwest to regional meteorology patterns. They found that ozone over the eastern U.S. and
Midwest is sensitive to the location of the polar jet stream and the location of the Bermuda High
pressure system. Lower surface ozone concentrations occur more frequently when the polar jet
stream is situated over the Midwest and eastern U.S. When the location of the Bermuda High
shifts eastward, clean marine air masses, which are less conducive to ozone formation, are able to
reach the eastern U.S. A westward shift of the Bermuda High prevents marine air from reaching
the continent and promotes conditions conducive to ozone formation (e.g., stagnation, clear skies,
higher temperatures). Overall, the study finds observed ozone trend (from 1980-2012) in the
eastern U.S. is decreasing, supporting the work of (Cooper et al.. 2012).
• The effects of local precursor emissions controls can be masked by meteorological variability. By
using empirical methods to detrend and account for meteorological variability on ozone,
Henneman et al. (2015) showed that the overall mean MDA8 ozone had decreased by 4%
between 2000 and 2012 in Atlanta, GA.
1-32
-------
1.5.2
Interannual and Multidecadal Climate Variability
The 2013 Ozone ISA did not discuss trends in meteorology associated with periodic variations in
winds and sea-surface temperatures. Nevertheless, natural variability induced by large-scale
climatological cycles has the ability to influence synoptic-scale patterns important for surface ozone.
Interannual (e.g., the El Nino-Southern Oscillation [ENSO] cycle) and multidecadal (e.g., the Pacific
Decadal Oscillation [PDO]) climate variability is especially important for USB ozone or other measures
of background ozone in the U.S. because these climate patterns affect long-range transport of
international pollution, the frequency of deep stratospheric intrusion, stagnation events, and wildfire
activity.
• The frequency of stratospheric intrusion events is linked to natural climate variability. Lin et al.
(2015) showed the frequency and interannual variability of springtime stratospheric intrusion
events due to deep tropopause folds in the western U.S. are linked to the ENSO cycle and the
subsequent location of the polar jet stream. Under La Nina conditions, the polar jet is often set up
over the western U.S. and leads to more frequent springtime stratospheric intrusion events
reaching surface locations compared with the same season following an El Nino (see Figure 1-6).
Recent work by Albers et al. (2018) highlighted the importance of wintertime buildup of ozone in
the lower portion of the stratosphere on the interannual variability of stratospheric intrusions in
the western U.S.
• Long-range transport of Asian pollution is sensitive to ENSO and the PDO. Lin et al. (2014)
found that the influence of Asian pollution and ozone measurements at surface observations in
Hawaii are tied to interannual (via ENSO) and multidecadal (via the PDO) climate variability and
notes that these climate patterns are likely important for transporting international emissions to
the western U.S.
• Deep stratospheric intrusions are sensitive to the ENSO cycle and the Northern Annular Node
(NAM) interannual oscillation. An increase in upper tropospheric ozone is often associated with
El Nino events (Langford. 1999). However, this increase in upper level ozone rarely reaches the
surface. During La Nina, the upper level jet is typically positioned over the western U.S. where
deep stratospheric intrusions more frequently reach the surface (Lin et al.. 2015V Thus, in the
western U.S., La Nina years are more likely to see an increase in stratospheric-influenced
background ozone during springtime/early summer than El Nino years. Albers et al. (2018) tied
the strength of springtime ozone intrusion events to the NAM interannual oscillation.
• The Atlantic Multidecadal Oscillation (AMO) and ENSO influences ozone levels in the eastern
U.S. Shen et al. (2017) found that the warm phase of the AMO drives warmer, drier, and stagnant
weather in the eastern and midwestern U.S. and that a shift from the cold to warm phase of the
AMO can increase max daily 8-hour ozone between 1 and 5 ppb in this region.
• Variability in climate can influence the activity of wildfires in the western U.S. The frequency
and distribution of fire activity in the western U.S. is influenced by temperature and precipitation
patterns associated with climate variability (Abatzoglou et al.. 2016).
1-33
-------
1992 1996 2000 2004 2008 2012
Pinatubo El Nino La Nina
r2 (OBS AM3 )=0.56
r2 (OBS. OgStrat)=0.43
r2 (AM3, O3Strat)=0.74
Note; The black trace provides the observed median daily 8-hour ozone. The red trace is the Geophysical Fluid Dynamics
Laboratory's (GFDL) AM3 model estimate of ozone, and the blue trace is the AM3 model estimated stratospheric contribution.
Source: Reprinted with permission from the publisher; Lin et al. (2015).
Figure 1-6 Model-estimated April-May stratospheric ozone contributions and
observed surface ozone concentrations between 1990 and 2013 at
22 high-elevation sites in the western U.S.
1.5.3 Interactions between Meteorology and Topography
Topography influences ozone concentrations through terrain-influenced flows (mountain-valley
upslope/downslope flows) and by providing the setup for valley basin wintertime temperature inversion
(i.e., cold pools) that, when accompanied with snow cover, leads to high ozone concentration in oil and
gas development basins rSection 1.4.1; Oltmans et al. (2014); Rappengluck et al. (2014); Neemann et al.
(2015)1. Under stagnant conditions conducive to the formation and accumulation of ozone,
terrain-influenced flows transport ozone and its precursors within the complex mountain valley system.
Complex terrain wind systems including slope, valley, and land/lake breezes [i.e., Los Angeles Basin; Lu
and Turco (1995)1 drive the transport and exchange of ozone and ozone precursors between the urban
areas and nearby mountains (Blavlock et al.. 2017; Angevine et al.. 2012).
1-34
-------
1.6
Measurements and Modeling
1.6.1 Advances in Ozone Measurement Methods
This section provides a concise overview of methods used in monitoring networks and advances
in remote sensing using satellite-based technology for ozone and ozone precursor measurements. While
there is growing literature on low-cost sensors for ozone measurement, they have not been widely applied
in studies of atmospheric concentration distributions, human exposure, or health impacts, so studies about
them will not be reviewed here.
1.6.1.1 Network Monitoring Methods
A new Federal Reference Method (FRM) for ozone measurement was established in 2015
(40 CFR Part 50 Appendix D). The new ozone FRM is based on the detection of chemiluminescence
from the reaction of ozone with nitric oxide (NO). It was adopted because instruments based on
chemiluminescence from the reaction of ozone with ethylene were no longer commercially available.
Further discussion of chemiluminescence and UV measurements of ozone are presented in the 2013
Ozone ISA (U.S. EPA. 2013V Almost all State and Local Air Monitoring Stations (SLAMS) that report
data to the U.S. EPA's Air Quality System (AQS) database use the Federal Equivalence Method (FEM)
based on UV absorption.
1.6.1.2 Remote Sensing Methods Used for Investigating the Distribution and
Atmospheric Concentrations of Ozone
Both active (e.g., LIDAR) and passive (e.g., satellite-based spectrometer of reflected sunlight)
remote sensing techniques are capable of measuring ozone abundance from ground-based, airborne and
satellite platforms. LIDAR techniques for measuring ozone abundance are generally all based on the same
physical principles as measurement systems operate at similar UV wavelengths.
• Ground- and aircraft-based measurement methods used for determining vertical ozone and
precursor profiles. Light Detection and Ranging (LIDAR) is a laser-based method that is used for
measuring vertical or three-dimensional ozone concentration profiles from the ground or aircraft.
A laser source is used to illuminate a target, and light reflected from the target is measured with a
sensor to create a three-dimensional representation of higher concentrations of pollutants
(Sullivan et al.. 2014; Kuang et al.. 2013; Alvarez et al.. 2011). LIDAR or instruments based on
LIDAR, such as Differential Absorption Lidar (DIAL), have been used in several STE
(Section 1.3.2) and model evaluation studies (Section 1.6.2). A recent measurement
intercomparison between different ground-based ozone lidar systems with different transmitter
and receiver components showed good agreement and similar vertical profiles (Sullivan et al..
1-35
-------
2015a). The physical basis of ozone DIAL measurements is differential absorption and scattering
of laser pulses at two ultraviolet wavelengths, one at which ozone is much more strongly
absorbing than the other. Ozone DIAL is especially effective in measuring ozone from -150 m
above Earth's surface through the mid- to upper troposphere. The data quality at high altitude is a
function of both telescope size and solar background radiation, so the vertical extent of
high-quality measurements is shallower at midday. The accurate quantification of ozone in the
near field, typically from 0-150 m, is limited by the properties of the laser system and other
design elements." Ground-based tropospheric ozone lidars can reliably measure continuous
vertical profiles of ozone during both daytime and nighttime to more fully characterize pollution
events. To ensure continued support of these instruments, an interagency network initiated by
NASA, NOAA, and U.S. EPA, known as the Tropospheric Ozone Lidar Network (TOLNet),
began in 2011. Ozone lidars within TOLNet provide accurate (mostly within 5-10%)
observations to generate consistent, long-term data sets. Uncertainty in TOLNet lidar
measurements has been rigorously verified by direct comparisons with other lidars and
ozonesondes (Farris et al.. 2019; Leblanc et al.. 2018; Wang et al.. 2017; Sullivan et al.. 2015b;
Sullivan et al.. 2015a) and from first principles as recognized by the lidar community (Leblanc et
al.. 2016a; Leblanc et al.. 2016b).
• Measurement campaigns that have utilized LIDAR instrumentation. Most TOLNet instruments
are portable and have been deployed in air quality campaigns in coordination with state and local
agencies. Among these are: the Ozone Water Land Environmental Transition Study (OWLETS)
in coordination with the Maryland Department of the Environment; the Long Island Sound
Tropospheric Ozone Study (LISTOS) in coordination with the North East States for Coordinated
Air Use Management (NESCAUM); the Fires, Asian, and Stratospheric Transport—Las Vegas
Ozone Study (FAST—LVOS) in coordination with Clark County, Nevada; the Department of Air
Quality, and; the California Baseline Ozone Transport Study (CABOTS) in coordination with the
California Air Resources Board (CARB).
Passive remote sensing satellite-based techniques for measuring ozone are diverse, relying on
radiative transfer and chemical transport models to interpret ozone spectral fitting optimizations of
ultraviolet, visible, thermal infrared and microwave spectrometric data. These techniques can also be
applied to infer abundances of other ozone-relevant gases, including most notably nitrogen dioxide (NO2),
formaldehyde (HCHO) and carbon monoxide (CO). These satellite data do not however provide a
discreetly vertically resolved measurement of pollutant abundance given that light does not directly
distinguish molecules at various altitudes. Duncan et al. (2014) provides an excellent summary of satellite
measurement capabilities, their limitations and an overview of techniques, written in language meant to
inform non-expects.
Briefly, the quality of any space-based passive remote sensing trace gas retrieval product depends
on the molecular physics of the molecule of interest, the radiative properties of the atmosphere and the
spectral and spatial sampling characteristics of the instrument platform. Satellite retrievals of trace gas
abundances from passive remote sensing are intrinsically vertically integrated measurements, given that
light does not distinguish a molecular spectral feature at 10 km above the surface from one at 1 km above
the surface, but does have smoothly varying sensitivity to molecules at different altitudes. Given that
space-based measurements of trace gases are not vertically resolved to an extent necessary to quantify
surface abundances, chemical transport and radiative transfer models must be used as inputs to
1-36
-------
satellite-based retrievals, which adds a level of uncertainty and potential bias in using these products for
understanding surface ozone air quality.
• Total Column Ozone. Satellite-based spectrometers provide measurements of outgoing ultraviolet
(UV), visible (VIS), thermal infrared (IR) and microwave spectral ranges at various spectral
(wavelength) resolutions and spatial sampling rates. These radiance measurements can be used
either alone or in combination to retrieve the abundance of atmosphere ozone. Each spectral
channel has different retrieval characteristics because the molecular physics and the atmospheric
radiative transfer physics are different for each channel. The vast majority of the ozone vertical
column is located in the stratosphere above 10 km AMSL. Given that most ozone is located in the
stratosphere where the radiative transfer of ultraviolet light is relatively straightforward,
space-based UV spectrometers have provided valuable global maps and time series of total
column ozone abundance for the internationally-sanctioned Scientific Assessment of Ozone
Depletion over several decades (WMO Global Ozone Research and Monitoring Project, Report
No. 58). These datasets are also routinely published in the Bulletin of the American
Meteorological Society (Weber et al.. 2018). These total column ozone products have been
developed and evaluated and harmonized in the context of a global network of well-characterized
spectrometers (Network for the Detection of Atmospheric Composition Change [NDACC; De
Maziere etal. (2018)1). Observed variations of total column ozone have been interpreted in the
context of dynamical models and surface in situ observations to identify when and where STE
events influence surface air quality (Lin et al.. 2012a).
• Tropospheric ozone abundance. Providing satellite-based inferences of tropospheric ozone
abundances (below 10 km AMSL) have been of interest to researchers and space agencies for
decades (Fishman et al.. 1990V However, despite efforts to combine measurements from multiple
satellite-based spectrometer channels, including ultraviolet, visible, thermal infrared and
microwave, inference of the tropospheric abundance of ozone from satellite platforms is strongly
weighted towards the upper troposphere (and lower stratosphere), and is thus uncertain due to the
large impact of stratospheric ozone on outgoing radiances (Natrai et al.. 2011). Any inference
about impacts of upper tropospheric ozone on surface ozone concentrations would require use of
a dynamical model and would thus be subject to model uncertainties. The most promising
approach for understanding ozone in the lower layer of the atmosphere involves the combination
of visible measurements, which have near-uniform vertical sensitivity throughout the entire
atmosphere, with UV, thermal infrared or microwave measurements, which have sensitivity that
is heavily weighted to upper parts of the atmosphere. The Tropospheric Emissions: Monitoring of
Pollution (TEMPO) mission, scheduled for launch in 2022, will provide combined visible and
UV measurements, and has ozone retrieval algorithms in place that seek to provide some limited
and uncertain information on ozone abundance at 2 km AGL (Zoogman et al.. 2017V In short,
satellite-based retrievals of ozone have extremely limited information on surface ozone
abundance that must be interpreted with the aid of chemical transport models.
Satellite-based measurements of ozone precursors NO2 and formaldehyde. While satellite-based
inference of near-surface ozone is hampered by large stratospheric and upper tropospheric
concentrations of ozone, the UV and visible retrieval of formaldehyde and NO2, both ozone
precursors, do provide some information on boundary layer chemistry, as both pollutants are
short-lived and predominantly formed or emitted near the surface. While NO2 and formaldehyde
products are also subject to uncertainties and biases (Duncan et al.. 2014; Martin. 2008) there is
evidence that seasonal or annual averages and the spatial pattern of these satellite-based total
column products corresponds to mid—day surface measurements for both chemical species (Zhu
et al.. 2017; Lamsal et al.. 2015; Russell et al.. 2010). These relationships diverge, however, as
surface abundances become smaller as is occurring in the U.S. following emission controls, given
1-37
-------
that there is a nonzero contribution of global and regional tropospheric and stratospheric
abundances on the total vertical column abundance (Silvern et al.. 2019). With the recent late
2017 launch of TROPOMI, the improved spatial resolution relative to predecessor UV/Vis
spectrometers is providing researchers with opportunities to improve data characterization (Judd
et al.. 2019) and enhance applications (Beirle et al.. 2019; Goldberg et al.. 2019). The retrieval of
formaldehyde column data from predecessor instruments has been more uncertain than retrieval
of NO2 column data (Zhu et al.. 2016). but theoretical research has shown how this data can be
useful if appropriately characterized [e.g., Jin et al. (2017); Schroeder et al. (2017)1.
1.6.2 Advances in Regional Chemical Transport Modeling
The 2013 Ozone ISA provided an overview of chemical transport models (CTMs), including the
relevant processes, numerical approaches, relevant spatial scales, and methods for evaluation (U.S. EPA.
2013). Since the previous review, numerous improvements have been developed including (1) the
addition of a halogen chemistry mechanism (Gantt et al.. 2017); (2) better representation of land cover
and near-surface meteorology (Ran et al.. 2016). dry deposition and stomatal uptake (Rydsaa et al.. 2016).
stratosphere-troposphere exchange (Phoenix et al.. 2017). and biogenic emissions (Bash et al.. 2016); and
(3) better integration of meteorological models and CTMs (Xing et al.. 2017). The 2013 Ozone ISA
identified uncertainties in the fate of nitrogen oxides and oxidant chemistry in remote areas, which have
been reduced by advances in biogenic VOC chemistry (Lee et al.. 2014; Xie et al.. 2013) and new
analyses of nitrogen oxide lifetime in the atmosphere (Li et al.. 2018). The 2013 Ozone ISA (U.S. EPA.
2013) also identified errors introduced by the coupling of regional and global-scale models, which has
since been improved by the development of more systematic techniques (Henderson et al.. 2014).
development of hemispheric scale CMAQ (Mathur et al.. 2017). and improvement of horizontal
resolution in global models (Huang et al.. 2013a). This section summarizes recent efforts to evaluate the
performance of these more advanced models for simulating ozone over the U.S.
• Proper interpretation of regional chemical transport modeling requires comprehensive evaluation
with measurements and statistical metrics that are relevant and specific to the modeling
application (Simon et al.. 2012). To ensure that best practices are followed as it changes in
response to new air quality science, U.S. EPA has developed an extensive model evaluation
framework for CMAQ (U.S. EPA. 2019c).
• The accuracy of model estimates of ozone concentration, when compared to observations, varies
depending on location, time, and averaging metric. The most straightforward form of model
evaluation is to compare the simulated ozone concentrations from different models with the
ambient measurements. The Air Quality Model Evaluation International Initiative included
simulations over North America from four different research groups. The hourly ozone was
compared with 200 surface observation sites and the normalized mean bias was reported to range
from -22 to 2.4% (Im et al.. 2015). The most recently published evaluation of the CMAQ model
finds that hourly ozone concentrations in all seasons (Appel et al.. 2017) are underestimated, but
that the bias varies spatially. An evaluation of the WRF-Chem model using the 1-hour max ozone
concentrations reported normalized mean bias of-15% at urban locations (Yahvaetal.. 2015). A
meta-analysis examining six peer-reviewed journal articles published from 2006-2012 also found
that the average ozone concentration is usually simulated with lower mean bias than the 1-hour
1-38
-------
max ozone concentration. Simon et al. (2012) reported that the average ozone concentration is
usually simulated with mean bias between 1 and 7 ppb, while the 1-hour max ozone concentration
mean bias ranged between 4 and 12 ppb (25th-75th percentile of reported studies). The
normalized mean error for hourly ozone ranged between 21 and 47 ppb, while the normalized
mean error for the 1-hour max ozone concentration ranged between 19 and 22 ppb (25th-75th
percentile of reported studies). Because the estimated model error varies considerably, it is
important to evaluate the model results using observations and statistical metrics relevant to the
application of interest.
• Differences between model chemical parameterizations can introduce a variance in simulated
ozone concentrations of 5%. The accuracy of the ozone simulation depends on the accuracy of the
simulation of many interconnected physical, chemical, and biological systems. Many studies have
examined each of these processes to further improve chemical transport modeling. An
intercomparison of just the chemistry models that participated in AQMEII, using identical
meteorological conditions, chemical boundary conditions, photolysis rates, and biogenic and
anthropogenic emissions, found on avg 5% variability due to differences in the chemistry
parameterization, with larger differences for the modeled NOx:VOC ratio, suggesting greater
variability in the model's estimate of the sensitivity to emission changes (Knote et al.. 2015).
• Limitations in meteorological process simulations can introduce errors. A study by Rvu et al.
(2018) attributed up to 40% of the ozone bias to errors in the simulation of clouds, noting that
photolysis reactions and biogenic VOC emissions both depend on sunlight. The simulation of
atmospheric mixing near the surface, namely the planetary boundary layer, is also relevant to
estimating the daily peak ozone, and an intercomparison of different approaches did not yield a
single model that performed best (Cuchiara et al.. 2014).
• Uncertainty in emissions leads to uncertainty in simulated ozone concentrations. Ozone
simulations can be improved with more accurate estimates of the magnitude and timing of
biogenic and anthropogenic emissions (Travis et al.. 2016; Ahmadov et al.. 2015). although the
importance of errors in estimated emissions depends on the relative availability of NOx or VOCs
(Kota et al.. 2015).
• Models are able to capture the spatial and temporal features of ozone trends but tend to
underestimate the magnitude of the trend. Another important aspect to model evaluation is the
determination of whether the model can correctly simulate the trends in concentrations and
attribute those trends to changes in emissions and weather (Foley et al.. 2015). A 21-year
hemispheric CMAQ simulation (Xing et al.. 2015) captured the decline in ozone concentrations
over the U.S. due to precursor emission reductions overthe period 1990-2010, but
underestimated the magnitude (observed: -1.1% change per year, simulated: -0.64% change per
year), although the change in NO2 was more accurately simulated (observed: -2.3% change per
year, simulated: -2.2% change per year). During the 2000-2010 period, the model captured the
observed downward trend in the Southwest and Midwest but underestimated the trends in other
regions (Astitha et al.. 2017). A study using the CAMx model overthe South Coast Air Basin in
California (Karamchandani et al.. 2017) showed an improvement over previous results but still
generally underestimated the reduction in ozone in response to emission reductions overthe years
in that area. With more coarse spatial resolution, global-scale models have also been used to
examine trends overthe U.S. (Lin et al.. 2017; Strode et al.. 2015). The global simulations are
evaluated using more remote monitoring stations designed to capture regional trends, and the
evaluation demonstrates that the models are able to capture the spatial and seasonal differences in
the ozone trends, but underestimate the magnitude of the decrease in ozone attributed to emission
reductions overthe eastern U.S.
1-39
-------
1.7 Ambient Air Concentrations and Trends
This section investigates spatiotemporal variability in ambient ozone concentrations. Ambient
ozone data reported in this section were obtained from AQS using data obtained from the State and Local
Air Monitoring Stations (SLAMS) network for ozone. The SLAMS network was described in detail in the
2013 Ozone ISA (U.S. EPA. 2013). and there have been no major changes. The number of monitors has
increased slightly to more than 1,300, and subsets of monitors are also part of the Photochemical
Assessment Monitoring Stations (PAMS) network and the National Core (NCore) network for
multipollutant measurements, also described in the 2013 Ozone ISA (U.S. EPA. 2013). Most ozone
monitors report hourly average concentrations with a required precision of 1 ppb and minimum detection
limit of 5 ppb or less. Data are available as reported (1-hour avg), or further summarized as DA24,
MDA1, and MDA8 (see Section 1.2.1.1).
Analyses in this section are based on data from 2015-2017 using either (1) a year-round data set,
with data only from those monitors that report year-round or (2) a warm-season data set with data from all
monitors that report data from May to September. Table 1-1 and Table 1-2 provide summary statistics
generated from the year-round and warm-season data sets, respectively, using SLAMS network
monitoring data from 2015-2017. Monitoring site locations corresponding to the warm-season and
year-round data sets are shown in Figure 1-7. The year-round data set includes data from considerably
fewer monitors than the warm-season data set, and year-round monitors are more concentrated in the
southern half of the U.S. because of monitoring requirements in these areas. States are required to monitor
for ozone for varying lengths of time during the year depending on which months are likely to see
elevated ozone levels from at least May to September. The warm-season data set was used to examine the
majority of ozone season data while providing a consistent time frame for comparison across states. All
available monitoring data including data from year-round monitors were also included in the
warm-season data set after removing observations outside the 5-month window. The data in Table 1-1 and
Table 1-2 show a strong seasonal pattern of ozone concentration.
1-40
-------
Table 1-1 Nationwide distributions of ozone concentrations (parts per billion [ppb]) from the year-round data
set 2015-2017.
Time Period3
N Sites
N Obs
Mean
SD
Min
1
5
10
25
50
75
90
95
98
99
Max
Max Site ID
1-h max (MDA1)
Year-round
809
830,984
45
14
0
17
25
29
36
44
53
63
69
78
85
163
060710005
Winter
761
196,858
37
9
0
13
21
26
32
38
43
48
51
55
59
132
490472003
Spring
792
207,700
50
11
0
25
32
36
42
49
56
63
68
74
79
134
201730010
Summer
789
206,617
51
16
0
20
26
30
40
50
60
71
79
89
97
163
060710005
Autumn
792
204,603
43
13
0
17
25
29
35
42
50
59
66
75
82
152
060370016
8-h max (MDA8)
Year-round
804
819,452
41
13
0
14
22
26
32
40
49
57
62
69
74
136
060719004
Winter
756
194,106
34
9
0
10
18
22
28
34
40
44
47
51
54
121
490472003
Spring
784
203,990
46
10
0
22
29
33
39
46
53
59
63
68
71
109
060714003
Summer
782
203,088
46
14
0
18
23
27
35
46
55
64
70
77
83
136
060719004
Autumn
788
201,810
39
11
0
14
21
25
31
38
45
53
59
66
72
112
060370016
1-41
-------
Table-1-1 (Continued): Nationwide distributions of ozone concentrations (parts per billion [ppb]) from the year-
round data set 2015-2017.
Time Period3
N Sites
N Obs
Mean
SD
Min
1
5
10
25
50
75
90
95
98
99
Max
Max Site ID
24-h avg (DA24)
Year-round
809
830,984
30
11
0
7
13
17
23
30
38
44
48
53
56
96
490472003
Winter
761
196,858
25
10
0
4
9
12
18
26
32
38
41
44
47
96
490472003
Spring
792
207,700
36
9
0
15
20
24
29
36
42
47
51
55
57
83
090050005
Summer
789
206,617
33
11
0
12
16
19
25
33
41
48
52
57
61
95
060710005
Autumn
792
204,603
27
9
0
8
13
16
21
27
34
40
44
48
52
85
060570005
N sites = number of sites; N Obs = number of observations; SD = standard deviation; Min = minimum; 1,5, 10, 25, 50, 90, 95, 98, 99 = 1st, 5th, 10th, 25th, 50th, 90th, 95th, 98th,
99th percentiles; Max = maximum; Max Site ID = U.S. EPA Air Quality System identification number for monitoring site corresponding to observation in max column.
aWinter = December-February, spring = March-May, summer = June-August, autumn = September-November.
1-42
-------
Table 1-2 Nationwide distributions of ozone concentrations (parts per billion [ppb]) from the warm-season
data set 2015-2017.
U.S.
Region3
N Sites
N Obs
Mean
SD
Min
1
5
10
25
50
75
90
95
98
99
Max
Max Site ID
1-h max (MDA1)
U.S.
1,279
548,118
50
14
0
21
28
32
40
49
58
67
74
83
90
237
470370011
c
199
87,066
50
12
0
25
31
35
41
49
57
65
70
75
79
140
290190011
ENC
96
45,052
46
12
0
22
28
31
38
45
54
62
68
74
79
119
550790085
NE
195
84,892
48
14
0
22
28
31
38
47
57
67
73
81
87
126
090011123
NW
33
13,709
43
14
1
18
24
27
33
42
51
60
68
76
82
125
410290201
S
150
63,789
46
14
6
19
23
27
36
46
55
64
70
77
84
136
482010024
SE
216
90,523
46
13
0
20
25
29
36
45
54
62
66
72
77
237
470370011
SW
132
54,252
57
11
10
32
40
44
51
57
63
70
75
82
86
123
490495010
W
201
87,817
57
18
2
22
30
35
44
55
67
80
89
100
108
163
060710005
WNC
57
24,018
49
10
6
22
32
36
43
50
55
60
63
67
69
129
300630024
8-h max (MDA8)
U.S.
1,273
541,670
45
13
0
18
24
28
36
45
53
61
66
73
78
136
060719004
c
199
86,314
45
11
0
21
27
31
37
45
52
59
63
68
71
97
170310076
ENC
95
41,476
42
11
0
19
25
28
34
41
49
57
62
67
71
99
550790085
1-43
-------
Table 1-2 (Continued): Nationwide distributions of ozone concentrations (parts per billion [ppb]) from the warm-
season data set 2015-2017.
U.S.
Region3
N Sites
N Obs
Mean
SD
Min
1
5
10
25
50
75
90
95
98
99
Max
Max Site ID
NE
195
84,147
43
12
0
19
25
28
34
42
52
60
65
71
75
101
090070007
NW
33
13,633
39
12
1
15
20
24
30
38
46
55
60
66
71
116
410050004
S
150
63,241
41
13
4
17
20
24
31
41
50
58
62
68
72
109
482010024
SE
216
89,652
41
12
0
17
22
25
32
41
49
56
60
64
67
106
130670003
SW
132
53,842
53
9
8
29
37
41
47
53
58
64
67
71
74
93
080350004
W
198
85,923
51
15
1
20
27
32
40
50
61
71
77
85
91
136
060719004
WNC
56
23,442
46
9
3
19
29
33
40
47
52
57
59
62
64
78
460990008
24-h avg (DA24)
U.S.
1,279
548,118
32
10
0
12
16
19
25
32
39
46
50
55
58
95
060710005
c
199
87,066
31
8
0
14
18
21
25
31
37
42
46
50
52
70
390850007
ENC
96
45,052
31
9
0
13
18
20
25
31
37
43
47
51
54
68
260190003
NE
195
84,892
31
9
0
12
17
20
25
30
37
43
47
52
55
83
090050005
NW
33
13,709
27
9
1
9
13
16
21
27
34
40
44
48
51
80
410050004
S
150
63,789
29
10
2
11
14
16
21
28
36
42
46
50
52
70
481671034
SE
216
90,523
28
9
0
11
14
17
21
27
34
41
45
49
52
68
471550101
SW
132
54,252
41
8
7
20
26
30
35
41
46
51
54
57
59
75
040218001
W
201
87,817
37
12
0
15
20
23
29
36
45
53
58
64
68
95
060710005
1-44
-------
Table 1-2 (Continued): Nationwide distributions of ozone concentrations (parts per billion [ppb]) from the warm-
season data set 2015-2017.
U.S.
Region3
N Sites
N Obs
Mean
SD
Min
1
5
10
25
50
75
90
95
98
99
Max
Max Site ID
WNC
57
24,018
36
9
2
13
20
24
31
37
42
47
50
53
55
66
560130099
N sites = number of sites; N obs = number of observations; SD = standard deviation; Min = minimum; 1,5,10,25,50,90,95,98,99 = 1st, 5th, 10th, 25th, 50th, 90th, 95th, 98th, 99th
percentiles; Max = maximum; Max Site ID = U.S. EPA Air Quality System identification number for monitoring site corresponding to observation in max column.
aC = Central (Illinois, Indiana, Kentucky, Missouri, Ohio, Tennessee, West Virginia); ENC = East North Central (Iowa, Minnesota, Michigan, Wisconsin); NE = Northeast (Connecticut,
Delaware, Maine, Maryland, Massachusetts, New Hampshire, New Jersey, New York, Pennsylvania, Rhode Island, Vermont); NW = Northwest (Alaska, Idaho, Oregon, Washington);
S = South (Arkansas, Kansas, Louisiana, Mississippi, Oklahoma, Texas); SE = Southeast (Alabama, Florida, Georgia, North Carolina, South Carolina, Virginia); SW = Southwest
(Arizona, Colorado, New Mexico, Utah); W = West (California, Hawaii, Nevada), WNC = West North Central (Montana, Nebraska, North Dakota, South Dakota, Wyoming).
1-45
-------
Year-Round Only (11 sites) Warm Season Only (481 sites) • Both Datasets (798 sites)
Source; U.S. EPA 2018 analysis of Air Quality System network data 2015-2017.
Figure 1-7 Monitor locations for the warm-season and year-round data sets.
• The mean and upper percentiles of the nationwide ozone concentrations are slightly higher in the
warm-season data in Table I-2 than in the year-round data from Table I-I. and the standard
deviation (SD) is similar between the two data sets.
• A strong seasonal pattern in ozone concentrations is evident in the year-round data, with lower
MDA8 concentrations in autumn (median = 38 ppb) and winter (median = 34 ppb) and higher
concentrations in spring (median = 46 ppb) and summer (median = 46 ppb). Seasonal differences
are even greater for upper percentiles. A similar seasonal pattern was reported in the 2013 Ozone
ISA (U.S. EPA. 2013).
• For the warm-season data set. the 2015-2017 98th percentile MDA1, MDA8, and DA24
concentrations are 83, 73, and 55 ppb, respectively.
The median 2015-2017 MDA1, MDA8, and DA24 ozone concentrations for the warm-season
data set are 49, 45, and 32 ppb, respectively.
1-46
-------
• For the year-round data set, the 2015-2017 98th percentile MDA1, MDA8, and DA24
concentrations are 78, 69, and 53 ppb, respectively, and median 2015-2017 MDA1, MDA8, and
DA24 ozone concentrations are 44, 40, and 30 ppb, respectively.
Figure 1-8 through Figure 1-11 summarize ambient ozone concentration patterns and trends.
These figures contain data from both warm-season and year-round monitors for all monitors that met the
completeness criterion of 75% data capture. The data sets used in Table 1- and Table 1-2 are combined,
and the data in Figure 1-8 through Figure 1-11 reflect concentration metrics applied to the entire period of
monitor operation, rather than the same season across all monitors.
• Figure 1-8 shows the design values, or the 3-year avg of the annual 4th highest 8-hour daily max
(MDA8) ozone concentrations for 2015-2017 (see Section 1.2.1.1). The highest design values
(>76 ppb) occur in central and southern California, Arizona, Colorado, Utah, Texas, along the
shore of Lake Michigan, and in the Northeast Corridor.
,o o
ocA
o o
2015 - 2017 Ozone Design Values
• 43-60 ppb (179 sites) O 66-70 ppb (334 sites) • 76-112 ppb (110 sites)
© 61 - 65 ppb (378 sites) O 71-75 ppb (136 sites)
ppb = parts per billion.
Note; Values determined for the entire period of monitor operation for each monitor with 75% or greater data capture. Both
warm-season and year-round monitors included.
Source: U.S. EPA 2018 analysis of Air Quality System network data 2015-2017.
Figure 1-8 Individual monitor ozone concentrations in terms of design
values for 2015-2017.
1-47
-------
• Figure 1-9 shows the decreasing trend in the annual 4th highest MDA8 ozone concentration from
882 U.S. monitors. The median annual 4th highest MDA8 ozone concentration across those sites
decreased from more than 80 ppb in 2000 to less than 70 ppb in 2017. Other studies also reported
a decreasing trend over periods of 15 years or more for 4th highest MDA8 ozone concentration or
other ozone concentration metrics associated with higher concentrations (Lcfohn et al.. 2017;
Simon et al. 2015; Strode et al.. 2015).
100
80
N
70
.eve
60
National Trend Based on 882 Monitoring S
to
s
CO
§
o
o
CM
§
a
o
o
o
CM
ID
o
o
CM
o
CM
CM
CM
CM
CM
Year
ppb = parts per billion.
Note: Although the trend lines are annual values, the NAAQS level marked on the figure pertains to the 3-year avg of annual
4th highest daily max 8-hour concentrations over a consecutive 3-year period, and conclusions cannot be reached regarding
exceedances through its comparison to individual years. All monitors with 75% or greater data capture included. Both warm-season
and year-round monitors included.
Source: U.S. EPA, National Air Quality: Status and Trends of Key Air Pollutants, https://www.epa.gov/air-trends/ozone-trends.
accessed July 2018.
Figure 1-9 National 4th highest 8-hour daily max ozone trend and
distribution across 882 U.S. ozone monitors 2000-2017
(concentrations in ppb).
1-48
-------
This decline in annual 4th high MDA8 ozone concentrations across the U.S. was accompanied by
reductions in ozone precursor concentrations in many areas of the U.S. over the same period. A
substantial drop in annual average NO2 concentrations was highlighted in the 2016 ISA for
Oxides of Nitrogen—Health Based Criteria (U.S. EPA. 2016c). and a recent update indicated a
35% decrease between 2000 to 2018 (https://www.epa.gov/air-trends/nitrogen-dioxide-trends').
Over the same time period, CO concentration declined by 59% (https://www.epa.gov/air-
trends/carbon-monoxide-trends). National ambient concentration data of this type are not
regularly compiled for VOCs by U.S. EPA. However, as described in Section 1.3.1.1.3. an
estimated 17% decrease in national anthropogenic VOC emissions occurred from 2002-2014.
Figure 1-10 shows a regional breakdown of the trend in annual 4th highest MDA8 ozone
concentrations. Declines are observed in most regions, with the strongest declines in regions that
had the greatest concentrations.
In contrast to the decreasing trend in ozone metrics associated with higher concentrations,
5th percentile ozone concentrations at the lower end of the ozone concentration distribution have
exhibited both increasing and decreasing trends in summer, depending on individual monitors,
and a generally increasing trend in winter from 1998-2013 (Simon et al.. 2015). These
observations demonstrate that a compression of the ozone concentration distribution has occurred
over this period.
Figure 1-11 shows the geographic difference in the design values for all U.S. monitors between
the 2008-2010 period and the 2015-2017 period. Since the 2008-2010 period was used to
designate attainment and nonattainment areas for the 2008 ozone NAAQS, this comparison
indicates progress achieved since efforts to meet that standard began.
Note that Figure 1-11 compares concentrations between two time periods but the differences
should not be interpreted as a trend. As discussed in Section 1.5. natural interannual variability in
synoptic-scale meteorology patterns can influence ozone formation and transport in specific
years; therefore, trends must be derived from a longer time series rather than comparisons of two
discrete sets of years.
Some seasonal shifts in ozone concentration patterns have occurred as NOx emissions have
decreased. For example, in the Southeastern Aerosol Research Characterization (SEARCH)
network in Georgia, Alabama, and Mississippi from 1992 to 2014, MDA8 ozone concentrations
declined more in summer than in other seasons, resulting in a decrease in seasonal variability and
in some cases a shift in annual ozone maxima from summer to spring (Blanchard et al.. 2019).
Similarly, in the state of New York annual ozone maxima typically occurred in summer before
about 2010, and while this is still the case in metropolitan New York City, spring maxima are
now observed at rural sites like Whiteface Mountain, NY and Pinnacle State Park, NY (Blanchard
et al.. 2019).
National avg 99th percentile MDA8 ozone concentrations also exhibit less seasonal variability
than previously reported. 99th percentile summer MDA8 concentrations in Table 1-1 averaged
83 ppb for 2015-2017, compared with 90 ppb for the period from 2007-2009 (U.S. EPA. 2006).
or 7 ppb lower. In contrast, the avg 99th percentile winter MDA8 concentration is 2 ppb higher
than the winter avg of 52 ppb for 2007-2009.
1-49
-------
100
90
.Q
Q.
Q.
C
o
ts
c
c;
u
c
o
u
ai
c
o
N
o
80
70
60
o
x—
CM
CO
m
o
o
O
O
O
o
o
o
O
O
O
o
CM
C\j
CM
CM
CM
CM
CD
h-
od
CD
O
T—
o
o
o
O
—
T—
o
o
o
O
o
o
CM
CM
CM
CM
CM
CM
Year
CM
CO
m
CD
h-
¦*—
^—
r—
t—
T—
t—
O
o
o
o
o
o
CM
CM
CM
CM
CM
CM
NOM Climate Region (# sites)
Central (158) Southeast (170)
East North Central (63) Southwest (60)
Northeast (143) West (154)
Northwest (17) West North Central (13)
South (104)
C = Central (Illinois, Indiana, Kentucky, Missouri, Ohio, Tennessee, West Virginia); ENC = East North Central (Iowa, Minnesota,
Michigan, Wisconsin); NE = Northeast (Connecticut, Delaware, Maine, Maryland, Massachusetts, New Hampshire, New Jersey,
New York, Pennsylvania, Rhode Island, Vermont); NW = Northwest (Alaska, Idaho, Oregon, Washington); ppb = parts per billion.
S = South (Arkansas, Kansas, Louisiana, Mississippi, Oklahoma, Texas); SE = Southeast (Alabama, Florida, Georgia, North
Carolina, South Carolina, Virginia); SW = Southwest (Arizona, Colorado, New Mexico, Utah); W = West (California, Hawaii,
Nevada); WNC = West North Central (Montana, Nebraska, North Dakota, South Dakota, Wyoming).
Note: All monitors with 75% or greater data capture included, both warm-season and year-round monitors.
Source: U.S. EPA, National Air Quality: Status and Trends of Key Air Pollutants, https://www.epa.qov/air-trends/ozone-trends.
accessed July 2018.
Figure 1-10 Trend in mean 4th highest 8-hour daily max ozone by U.S. region
2000-2017.
1-50
-------
Change in Ozone Design Values from 2008 - 2010 to 2015 - 2017
• Decrease of 8 to 17 ppb (192 sites) o Decrease of 3 to 7 ppb (378 sites) ° Change of less than 3 ppb (253 sites)
• Increase of 8 to 12 ppb (8 sites) o Increase of 3 to 7 ppb (70 sites)
ppb = parts per billion.
Note: All monitors with 75% or greater data capture included, both warm-season and year-round monitors.
Source: U.S. EPA 2018 analysis of Air Quality System network data 2015-2017 and 2008-2010.
Figure 1-11 Individual monitor 3-year avg of the changes in ozone design
values from 2008-2010 to 2015-2017.
• Diel characteristics of ozone concentration were described in the 2013 Ozone ISA (U.S. EPA.
2013) and have changed little. In urban areas, 1-hour daily max concentrations typically occur in
the early afternoon, and the difference between highest and lowest concentration varies
considerably by city. There is little difference in diel profiles between weekdays and weekends.
In rural areas, there was considerable variability in diel patterns. Diel patterns are described in
more detail in the 2013 Ozone ISA (U.S. EPA. 2013).
• Figure 1-12 shows W126 exposure metric values (see Section 1.2.1.2) for network monitoring
sites averaged over 2015-2017. The highest W126 values occur in California, Nevada, Arizona,
Colorado, and Utah at sites with design values above 70 ppb (Figure 1-8).
1-51
-------
••
:v
• •
2015 - 2017 Average W126 Values
• 0 - 7 ppm-hrs (769 sites) O 14 -15 ppm-hrs (32 sites) • 18 - 57 ppm-hrs (89 sites)
o 8 -13 ppm-hrs (294 sites) o 16-17 ppm-hrs (18 sites) A. Design Value > 70 ppb
ppb = parts per billion.
Note: All monitors with 75% or greater data capture included, both warm-season and year-round monitors.
Source: U.S. EPA 2018 analysis of Air Quality System network data 2015-2017.
Figure 1-12 Individual monitor W126 exposure metric values for 2015-2017.
1.8 U.S. Background (USB) Ozone Concentrations
Broadly speaking, USB ozone is used in this document to mean ozone that cannot be reduced by
domestic emission controls or other domestic interventions within the U.S. More precise definitions of
USB and other definitions of background ozone are thoroughly discussed in Section 1.8.1. Major
contributors to USB are stratospheric transport (Section 1.3.2). wildfires (Section 1.3.1.3), natural
precursors (Section 1.3.1.3). and international sources [Section 1.3.1.2; Jaffe et al. (2018)1. Quantification
of USB ozone on days when MDA8 ozone concentrations exceed 70 ppb is more relevant to
understanding USB ozone contributions at the upper end of the distribution than are seasonal mean USB
ozone estimates because USB varies daily and is a function of season, meteorology, and elevation (Jaffe
et al.. 2018). Applications of CTMs to estimate USB ozone have found that USB concentrations are
relatively constant with increasing total ozone concentration, indicating that days with higher ozone
1-52
-------
concentrations generally occur because of higher U.S. anthropogenic contributions (Dolwick et al.. 2015).
Thus, estimates of average percentage USB contributions will generally be higher for seasonal averages
than for days at the upper end of the distribution because these longer periods include many days with
lower ozone concentrations. Lower USB contributions on days of high ozone concentration can result
from meteorological conditions that favor large ozone production from U.S. anthropogenic sources
relative to USB sources. The highest ozone concentrations observed in the U.S. have historically occurred
during stagnant conditions when an air mass remains stationary over a region abundant in anthropogenic
ozone precursor sources (U.S. EPA. 2013. 2006. 1996). Conversely, the largest USB contributions often
occur when the atmosphere is well mixed (see Section 1.5) and transport of USB ozone generated in the
stratosphere or during long-range transport of Asian or natural precursors in the upper troposphere more
readily occurs (Langford et al.. 2015).
Based on these considerations, this section emphasizes USB on days with high ozone
concentration as the most relevant for discussing USB ozone, and wherever possible, the focus is on
estimates of USB under these conditions because they are most relevant for evaluating the potential for a
role of USB ozone in contributing to the highest ozone concentrations. Discussion of seasonal and
monthly means of hourly data are also included because longer averaging times are relevant to
assessments of health and ecological effects. Seasonal and monthly mean USB ozone estimates are also
useful for comparing model data to monitoring data to get a first-order estimate of a model's ability to
simulate annual cycles, long-term trends, and interannual variability.
1.8.1 Modeling Strategies Applied to Estimate U.S. Background (USB) Ozone
As described in Section 1.2.2.1. the 2006 Ozone AQCD (U.S. EPA. 2006) reasoned that USB
ozone cannot be obtained solely by examining measurements of ozone from relatively remote monitoring
sites in the U.S. Instead, air quality model simulations are used to estimate USB ozone. The 2006 Ozone
AQCD (U.S. EPA. 2006) followed this approach after concluding that background ozone concentrations
could not be determined exclusively from ozone measurements because of long-range transport of ozone
originating from U.S. anthropogenic precursors even at the most remote monitoring locations. At the time
that the 2006 Ozone AQCD (U.S. EPA. 2006) was published, GEOS-Chem [v4.3.3; Fiore et al. (2003)1
was the only model documented in the literature for calculating background ozone concentrations, and it
was used for the 2006 Ozone AQCD estimates of background ozone. Global-scale simulations like those
obtained from GEOS-Chem for the 2006 Ozone AQCD had coarse spatial resolution, on the order of
100 km, and may not have adequately resolved topographic features in complex terrain or concentration
gradients of ozone and its precursors in areas with large emissions, including urban areas and large point
sources. A common approach to achieve finer scale spatial resolution is to use nested modeling systems,
in which a coarse resolution global scale CTM is used to provide the boundary condition data for a higher
resolution regional-scale model. This approach was described in the 2013 Ozone ISA (U.S. EPA. 2013)
using the regional CTMs CMAQ and CAMx with boundary conditions taken from the global-scale CTM
1-53
-------
GEOS-Chem. The 2013 Ozone ISA also reported background ozone estimates using just coarse
resolution global-scale models.
• CTMs still remain the preferred approach for estimating USB or other measures of background
ozone, but since publication of the 2013 Ozone ISA (U.S. EPA. 2013). coupled global/regional
models rather than global CTMs have become more widely applied.
• A major advance in methodology since the 2013 Ozone ISA is the capability for estimating USB
ozone using global CTMs coupled with higher resolution regional models (Jaffc et al.. 2018). and
regional models such as CMAQ (Bvun and Schere. 2006) and CAMx (Emery et al.. 2012) are
typically used to estimate USB or other measures of background ozone for air quality
management applications. Boundary conditions are set using output from a global CTM (Lefohn
et al.. 2014; Emery et al.. 2012) and U.S. anthropogenic emissions in the global CTM can also be
set to zero (Emery et al.. 2012).
• Background ozone is estimated using either zero-out simulations (USB) or source apportionment
simulations (USBab). The most widely used approach for measuring USB or other measures of
background ozone is the "zero-out" method, in which anthropogenic U.S. emissions are set to
zero in a model simulation to estimate USB ozone in the absence of U.S. anthropogenic
emissions (see Section 1.2.2.1). In the source apportionment approach, all emissions sources are
included in the model, and reactive tracer species are used to track the mass contributions of USB
sources and U.S. anthropogenic emissions to ozone (see Section 1.2.2.2). Both zero-out and
source apportionment modeling approaches require the use of global-scale CTMs to provide
boundary condition data for the finer resolution, regional-scale models.
1.8.1.1 Zero-Out and Other Source Sensitivity Approaches
In the model zero-out sensitivity approach to estimate USB, the model simulations include all
international emissions and U.S. natural emissions, but U.S. anthropogenic emissions are set to zero. This
approach provides a model estimate of the lowest ozone levels that would occur in the absence of U.S.
anthropogenic emissions. This is the most widely used approach for estimating USB. Other source
sensitivity approaches are described below.
• The zero-out method is part of a larger set of methods called model sensitivity analyses which can
be used to assess how ozone responds to changes in emissions (Huang et al.. 2013a; Reidmiller et
al.. 2009). There are several categories of sensitivity methods that have been the subject of recent
research and evaluation.
• Direct perturbation modeling is the simplest sensitivity method (Galmarini et al.. 2017; Wu et al..
2009). The model is run with emissions for each source of interest perturbed, typically with
reductions of 20 to 50%, and then the model outputs are compared to the base case run with full
emissions. Zero-out is a special case of perturbation modeling in which emissions for a source
category or source region are set to zero.
• Adjoint modeling is a variation of perturbation modeling that calculates the sensitivity of a
specified model parameter to individual components of the initial model state over the course of a
simulation (Zhang et al.. 2009; Sandu et al.. 2005). Adjoint techniques are well suited for
receptor-oriented applications.
1-54
-------
• Decoupled direct methods (DDM) are designed to calculate local linear sensitivities of ozone
responses to small emissions perturbations. Similar to adjoint, HDDM uses derivatives of the
underlying governing equations within the model to track sensitivity of ozone to emissions for
designated sources without actually perturbing the emissions imports. Unlike the direct
perturbation and adjoint methods, higher order DDM (HDDM) can be set up to track nonlinear
ozone responses to emissions changes (Hakami et al.. 2004; Dunker. 1981).
• Path-integral methods are applied to nonlinear ozone responses (Dunker et al.. 2017). In
path-integral methods, source contributions are determined by integrating sensitivity coefficients
over the range of emissions from the background case to the base case. This contrasts with other
source apportionment approaches that determine source contributions from the base case
chemistry. The disadvantage is that more computational effort is required.
• Results of sensitivity methods are strongly dependent on emission inventories used as input,
making evaluation of uncertainties in source estimates critical (Jaffe et al.. 2018V This
dependence also applies to brute force zero-out and emissions tagging techniques.
• In the model sensitivity approach, contributions to ozone are evaluated by scaling the model
response to an emissions change; for example, the contribution of Asian emissions to ozone in the
U.S. can be evaluated using a 20% reduction in Asian emissions and multiplying the modeled
ozone response by a factor of 5. A key limitation of sensitivity approaches is that ozone can have
a nonlinear response depending on the size of the emissions reduction, so scaling the model
response may not provide an accurate estimate of the source contribution (Huang et al.. 2013a).
While the HDDM method can be used to account for nonlinear ozone response, its accuracy
decreases when trying to estimate ozone response to very large emissions changes.
1.8.1.2 Source Apportionment Approaches
As an alternative to model sensitivity approaches, source apportionment techniques track source
contributions to ozone formation without perturbing emissions. Tracking techniques use reactive tracer
species to tag specific emissions source categories or source regions and then track the ozone produced by
emissions from those source groups (Cohan and Napelenok. 2011; Grewe et al.. 2010). A challenge in the
use of source apportionment techniques is that both VOC and NOx precursors contribute to the
production of ozone, so rules must be developed to assign ozone production to either the VOC or NOx
source groups.
• Tagging approaches include CAMx Ozone Source Apportionment Technology (OSAT;
http://www.camx.com/) and CMAQ Integrated Source Apportionment Method [ISAM; Kwok et
al. (2015)1. These approaches assign ozone production to either the tagged VOC or NOx
precursors depending on whether the ozone is produced in a VOC-sensitive or NOx-sensitive
chemical regime.
• Tagging can be applied to track contributions to ozone production based on source regions or
source types (Fiore et al.. 2002; Wang et al.. 1998) or to ozone transportation from the
stratosphere (Zhang et al.. 2014; Lin et al.. 2012a).
• Other tagging approaches that have been developed attribute source contributions to a single
precursor, either NOx or VOC. For rural or remote areas in which ozone is mostly produced in
NOx-sensitive chemical regimes, tracers can be used to track the source contributions from NOx
1-55
-------
emissions (Pfister et al.. 2013; Emmons et al.. 2012). For urban areas where ozone is mostly
produced in VOC-sensitive conditions, tracers can be used to track the source contributions from
VOC emissions (Butler et al.. 2011; Ying and Krishnan. 2010).
• Tagging of ozone source contributions is more complex when natural and anthropogenic
precursors react to produce ozone. The CAMx model source apportionment technique includes an
option for preferentially attributing ozone production to anthropogenic precursors (Jaffe et al..
2018) when anthropogenic precursors react with natural precursors.
• Tracking techniques have been used to define an emissions-influenced background (EIB) ozone
concentration (see Section 1.2.2.4) that addresses the reduced lifetime of ozone that is transported
from the stratosphere or produced from natural and international precursors due to reaction with
and is chemically destroyed by anthropogenic emissions (Lcfohn et al.. 2014).
1.8.1.3 Differences between Zero-Out and Source Apportionment Approaches
Due to the nonlinear character of ozone chemistry, removing emissions in model sensitivity or
model zero-out simulations will give a slightly different answer than tracking emissions contributions to
ozone production in a source apportionment approach. For this reason, USB estimated with a source
apportionment approach is identified in this document as apportionment-based USB (USBab) following
(Dolwick et al.. 2015). while USB without qualification (and without a subscript) generally refers to USB
based on zero-out or other source sensitivity-based modeling approaches (see Section 1.2.2.1). The
zero-out approach is more suited for answering the question "what ozone levels would exist in the
absence of all U.S. emissions"? while the source apportionment approach is more suited for answering the
question "what amount of current ozone comes from background sources"? The difference between USB
and USBab is small in remote areas most strongly affected by USB sources, but can be substantial in
urban areas strongly affected by anthropogenic sources that influence both production and destruction of
ozone (Dolwick et al.. 2015).
• Comparison of U.S. background estimates between the zero-out approach using CMAQ and a
tagged source apportionment method using CAMx gave similar April to October mean estimates
in rural areas, but the CAMx source apportionment approach produced lower estimates in urban
areas (Dolwick et al.. 2015).
• Differences in seasonal mean MDA8 U.S. background estimates from the zero-out and source
apportionment approaches were less than 2.5 ppb at 75% of locations after base case model bias
correction (Dolwick et al.. 2015).
• Differences between USB and USBab in urban areas indicate that ozone reductions resulting from
a reduction of U.S. anthropogenic ozone precursor emissions could be partially offset by the
absence of interactions with U.S. anthropogenic emissions that destroy USB ozone. However, this
offset may not apply to other photochemical oxidants that are produced along with U.S.
anthropogenic ozone (see Section 1.2.1).
1-56
-------
1.8.1.4 Other Approaches for Estimating Background Ozone
One additional recently developed approach to estimating background ozone involves fitting a
running average of ozone concentrations over a long period to an exponential decay function (Parrish et
al.. 2017b).
• This approach is difficult to compare with modeling studies that rely on a more rigorous
definition of background. The regression approach also requires numerous assumptions, including
that U.S. emissions asymptotically approach zero and that background estimates remain constant
overtime.
• In addition, results reported in Parrish et al. (2017b) suggest that estimates of background ozone
are sensitive to assumptions of the exponential decay rate and the years of data included in the
analysis.
• It has been suggested that estimates using this approach are more representative of baseline ozone
concentrations plus some additional unquantified amount of ozone produced from local U.S.
anthropogenic emissions, rather than background concentration as defined by various modeling
approaches (Jaffe et al.. 2018).
1.8.1.5 Uncertainties and Model Disagreement
Jaffe et al. (2018) reviewed recent modeling results and reported that USB ozone estimates
contain uncertainties of about 10 ppb for seasonal average concentrations, with higher uncertainty for
MDA8 average concentrations. Because of uncertainty in model predictions, simple bias correction
approaches are useful to adjust model results for bias and error. However, these approaches might not be
reliable if the model has large errors in the proportion of USB ozone and locally produced ozone.
Accordingly, days with poor model performance are typically excluded when using model results to
estimate USB or other measures of background ozone (Fiore et al.. 2014). There have been continued
efforts to improve model performance and better understand biases and uncertainties involved in the
application of CTMs to estimating USB or other measures of background ozone:
• While determining an overall uncertainty for USB ozone is challenging, confidence in estimates
of USB ozone or other measures of background ozone can be evaluated by comparing results
from multiple models and approaches. Several direct comparisons of results between models have
recently been reported. A complete table of model comparisons was recently published in Jaffe et
al. (2018).
• In many cases, discrepancies have been attributed to differences in model representations of
various processes. For example, higher seasonal mean values were estimated in both spring and
summer with the AM3 model compared with other models, most likely due to different model
representations of stratosphere-troposphere exchange, wildfires, lightning source and chemistry,
and isoprene oxidation chemistry (Fiore et al.. 2014). Differences in Asian transport have also
been observed (Huang et al.. 2013a). and differences in how convection is modeled have been
shown to have a large influence on transport (Orbe et al.. 2017).
1-57
-------
• Differences in seasonal mean ozone estimated with a regional model using four sets of boundary
conditions from different global models (AM3, MOZART, Hemispheric CMAQ, and
GEOS-Chem) exceeded 10 ppb and on individual days, differences as high as 15 ppb were
observed (Hogrcfc ct al.. 20IS).
• Multimodel approaches have been carried out to investigate the influence of intercontinental
transport on ground-level ozone concentrations throughout North America and Europe (Galmarini
et al.. 2017). This approach could help to estimate USB ozone in areas where large differences
between model results are observed (Jaffc et al.. 2018V
1.8.2 Concentrations and Trends of U.S. Background (USB) and Baseline
Ozone
The 2013 Ozone ISA (U.S. EPA. 2013) summarized estimates of USB, NAB, and natural
background ozone from the published literature using the CTMs GEOS-Chem, CAMx, and CMAQ.
Higher USB and NAB concentrations were estimated in the western U.S. than in the eastern U.S.,
especially in the intermountain West and Southwest. NAB was also found to constitute a larger fraction of
modeled ozone at the upper end of the concentration distribution in the intermountain West than in other
regions of the country. Higher USB and NAB concentrations were also estimated at elevations greater
than 1,500 m than at lower elevations. The east versus west and the high versus low elevation differences
were both similar in magnitude to the estimated uncertainty for CTM seasonal mean USB concentrations
of 10 ppb (Jaffc et al.. 2018) described in Section 1.8.1. As detailed in this section, more recent research
has confirmed these broad features of higher USB in the West than in the East and at higher elevations.
This research has also provided new evidence for both an inverse relationship between relative USB
contribution and total ozone concentration in most U.S. locations and a leveling off of baseline ozone
concentrations that have been increasing since monitoring began.
1.8.2.1 New U.S. Background (USB) and North American Background (NAB)
Estimates
A greater variety of approaches has led to a wider range of USB and NAB estimates than reported
in the 2013 Ozone ISA (U.S. EPA. 2013). Jaffe et al. (2018) summarized model results from
14 publications in a supplementary table that reported seasonal mean NAB concentrations or seasonal
mean concentrations based on alternative background metrics that ranged widely from 20-50 ppb.
• Geographic trends were generally similar to those described in the 2013 Ozone ISA. Additional
modeling supports the prediction of higher NAB and USB estimates at high-elevation sites in the
western U.S. than in the eastern U.S. or along the Pacific Coast. For example, Fiore et al. (2014)
estimated summer seasonal mean NAB MDA8 ozone concentrations ranging from 25 to 40 ppb at
high-elevation sites in the western U.S. compared with 20 to 30 ppb in the eastern U.S. Dolwick
et al. (2015) estimated April to October mean USB MDA8 ozone concentrations of 40 to 45 ppb
at intermountain west monitors, compared with 25 to 35 ppb along the Pacific Coast. Guo et al.
(2018) estimated seasonal means for spring of 41 ppb for U.S. EPA Region 8, the region most
1-58
-------
closely corresponding to the intermountain West, but seasonal means for all other U.S. EPA
regions were narrowly distributed from 34 to 37 ppb.
• Several new studies reported results for daily or 4th highest MDA8 USB ozone concentrations
(Guo et al.. 2018; Dunker etal.. 2017; Nopmongcol et al.. 2017; Dolwick et al.. 2015; Fiore et al..
2014). As indicated in Section 1.8.1.5. uncertainty is considerably greater for individual days than
for seasonal or monthly means. Moreover, metrics based on high-concentration days can be
driven by extreme episodes that can be identified as natural events and are not related to
anthropogenic pollution. For example, major wildfires can result in extremely high particulate
matter concentrations well outside the range observed during anthropogenic pollution events in
the U.S. (Laing and Jaffe. 2019).
• Air quality models agree reasonably well on seasonal mean USB concentrations and their
variation with region and altitude, but they are not capable of providing precise daily USB ozone
concentrations. This is illustrated by Figure 1-13 and Figure 1-14 (Fiore et al.. 2014). Figure 1-13
illustrates agreement between the AM3 and GEOS-Chem models that the highest seasonal
average NAB concentrations in the U.S. were 40-50 ppb, and that they occurred in spring in the
four corners region of Colorado, New Mexico, Utah, and Arizona. However, poor agreement
between air quality models is observed for 4th highest MDA8 USB ozone concentrations in
Figure 1-14 (Fiore et al.. 2014). The AM3 model predicted that the 4th highest MDA8 USB
ozone concentrations for a given site were highest in Colorado early in the year, while the
GEOS-Chem model predicted a maximum for 4th highest MDA8 USB ozone concentrations in
New Mexico, much later in the year. The differences in model timing of the 4th highest NAB
over many regions have been attributed to model treatment of driving processes, including
stratosphere-troposphere exchange, wildland fires, and lightning (Fiore et al.. 2014).
• Figure 1-15 compares modeled 4th highest MDA8 USB ozone concentrations with modeled and
measured 4th highest MDA8 ozone concentrations for monitors in the Southeast and the
Mountains and Plains regions (Guo et al.. 2018). In the Mountains and Plains region, the 4th
highest MDA8 USB ozone concentrations were only slightly less than 4th highest MDA8 total
ozone concentrations, exceeding 60 ppb in some years. In the Southeast region, 4th highest
MDA8 USB concentrations were much smaller than 4th highest MDA8 total ozone
concentrations, generally less than 40 ppb at most sites. Guo et al. (2018) concluded that
interannual variability of 4th highest MDA8 was correlated with USB in the western U.S., but
there was little correlation in the eastern U.S. These results are a further illustration of USB ozone
patterns described in the 2006 Ozone AQCD, that while high USB concentrations can frequently
occur on high ozone days at high-elevation western sites, they are not representative of USB
concentrations on high ozone days at locations where ozone is mainly formed from anthropogenic
precursors during stagnant meteorological conditions (U.S. EPA. 2006). This is further discussed
in Section 1.8.2.3.
• Similar results were reported in other studies. Nopmongcol et al. (2017) used CAMx to estimate
annual 4th highest MDA8 USB ozone concentrations over the past five decades and reported an
increase of about 5 ppb over this period, with current concentrations mostly in the range of 45 to
60 ppb in the western U.S. and below 45 ppb in the eastern U.S. Dunker et al. (2017) also
reported CAMx estimates of MDA8 USB ozone concentrations above 60 ppb in some locations
in the western U.S. averaged over the days of the 10 highest ozone concentrations from March to
September 2010. They also observed USB contributing a higher fraction of ozone in the western
U.S., but a lower fraction in the eastern U.S.
1-59
-------
50°N
45"N
40°N
35"N
30°N
25°N
120 W
100°W
80°W
120°W
100°W
80°W
NAB ozone at surface, maximum daily 8-hour average [ppbv]
15 23 32
Source: Reprinted with permission from the publisher; Fiore et al. (2014).
GEOS-Chem
120°W
100°W
80°W
50"N
45°N
40"N
< 35°N
^ 30°N
25°N
120°W
100°W
80°W
Figure 1-13 Seasonal mean daily max 8-hour avg values of North American
background (NAB) in the lowest model layer for the Geophysical
Fluid Dynamics Laboratory's AM3 (left) and Goddard Earth
Observing System (GEOS)-Chem (right) simulations for spring
(March, April, and May [MAM], top) and summer (June, July, and
August [J J A], bottom) of 2006 estimated with North American
anthropogenic emissions set to zero.
1-60
-------
AM3 GEOS-Chem
50°N
45DN
40DN
35°N
30"N
25°N
120°W
50°N
45°N
40°N
35°N
30"N
25°N
120°W
35 42 50 57 65
Fourth highest MDA8 NAB ozone in model surface layer [ppb]
59 84 110 135 161 186 212
Day of fourth highest MDA8 NAB ozone in model surface layer [ppb]
Source: Reprinted with permission from the publisher; Fiore et al. (2014V
Figure 1-14 Fourth highest daily max 8-hour avg (MDA8) North American
background (NAB) ozone concentration between March 1 and
August 31, 2006 in the lowest model layer (top) and day of
occurrence (bottom) for the Geophysical Fluid Dynamics
Laboratory's AM3 (left) and Goddard Earth Observing System
(GEOS)-Chem (right) simulations.
100°W
80°W
100°W
80°W
120°W 100°W 80'W
120°W 100W 80"W
1-61
-------
Southeast
Obs
Oi_B«se
0,USB
i I::
(b)
100
90
80
I 70
04-06 05-07 06-08 07-09 08-09 09-10 10-12
Year
60
SO
40
30
20
Mount. + Plains
Obs
0,_B»se
O, USB
r
04-06 05-07 06-08 07-09 08-09 09-10 10-12
Year
Source: Reprinted with permission from the publisher; Guo et al. (20181.
Figure 1-15 The three annual 4th highest values (solid dots) used to calculate
the 3-year avg of the 4th highest daily max 8-hour avg ozone
(hollow diamond) and the range of ozone concentrations for each
year's 10 highest ozone concentration days (vertical bars)
between March and October in the (a) Southeast and (b)
Mountains and Plains regions for observed ozone concentration
measurements (black), modeled total ozone concentration
estimates (blue), and modeled U.S. background (USB) ozone
concentration estimates (red).
• Figure 1-16 (Dolwick et al.. 2015) illustrates the relationship between model bias and predictions
of 4th highest MDA8 USB ozone concentration, showing a weak relationship between
CMAQ-estimated USB concentrations (based on zeroing out anthropogenic emissions) but not
for CAMx source apportionment modeled USB concentrations for monitors at higher than 1-km
elevation, where USB concentrations are generally highest. Median USB values for days that
MDA8 ozone was overestimated by more than 10 ppb were 4 to 5 ppb higher than on days with
the greatest underestimation. Dolwick et al. (2015) explained this as an implication that model
error observed at high elevation western sites may be related to overestimation of background
ozone in the CMAQ model. Figure 1-16 also shows that CMAQ predictions of 4th highest MDA8
USB ozone concentration at western high-elevation sites uncorrected for model bias range from
less than 20 ppb to more than 70 ppb. However, concentrations at the upper end of this range are
not representative of USB ozone concentrations on high ozone days in less remote areas, as
discussed in Section 1.8.2.3.
1-62
-------
• As described in Section 1.2.2.1. previous assessments concluded that estimates of USB or NAB
concentration cannot be obtained directly by examining ozone measurements because of
long-range transport (U.S. EPA. 2006). and that the definition of USB implied that only CTMs
can be used to estimate USB concentrations (U.S. EPA. 2013). However, as described in
Section 1.8.1.4. an alternative measurement-based method that uses an exponential decay
function to fit a time series of ozone concentrations has also been proposed. Estimates of USB
contribution to ozone design value (see Section 1.2.1.1) based on this approach were at the
extreme end of 2006 model predictions likely due to the shortcomings discussed in this section
and in Section 1.8.1.4.
• There are several potential issues associated with this approach. As described in Section 1.8.1.4.
it assumes convergence to a constant ozone concentration attributed entirely to USB (Parrish and
Ennis. 2019; Parrish et al.. 2017b). but relatively constant anthropogenic sources have also been
identified in some instances (Parrish et al.. 2017b). It also assumes a constant USB concentration
over time (Parrish and Ennis. 2019). whereas intercontinental transport, remote global
measurements, and fire emissions exhibit long-term trends (Section 1.3.1.2.2). Finally,
convergence to a constant value for USB ozone concentration implies convergence to zero for
anthropogenic precursor emissions, which is not observed in emissions data (Section 1.3.1.1).
• In this section, both modeling and observation-based USB estimates and sources of their
uncertainties have been described. Higher uncertainties are estimated or anticipated for both
modeling and observation-based estimates of daily USB concentrations than for model predicted
seasonal means. Further research and additional comparison are proceeding with the goal of
reducing uncertainties in both measurement and modeling-based approaches (Parrish and Ennis.
2019).
1-63
-------
0-1 km > 1 km
Sites in the western U.S. binned by elevation
Sites in the western U.S. binned by elevation
Source: Dolwick et al. (2015)
Figure 1-16
Binned Model Bias
m
10 ppb under
m
6-10 ppb under
$
3-6 ppb under
$
0-3 ppb under
0
0-3 ppb over
$
3-6 ppb over
s
6-10 ppb over
*
>10 ppb over
Binned Model Bias
*
10 ppb under
m
6-10 ppb under
3-6 ppb under
a
0-3 ppb under
B
0-3 ppb over
&
3-6 ppb over
m
6-10 ppb over
m
>10 ppb over
Comparison of (a) Community Muitiscaie Air Quality (CMAQ)
zero-out and (b) Comprehensive Air Quality Model with
Extensions (CAMx) apportionment-based daily max 8-hour avg
U.S. background (USB and USBab) ozone estimates across eight
model bias and two site elevation bins.
1-64
-------
1.8.2.2
Seasonal Distributions in U.S. Background (USB) and Baseline Ozone
The 2013 Ozone ISA (U.S. EPA. 2013) reported higher seasonal mean USB and NAB
concentration estimates in spring than in summer for most regions of the U.S, and these results are
consistent with earlier modeling estimates (Fiorc et al.. 2003). However, while some new results
consistent with this pattern have been reported, other results suggest that summer USB and baseline ozone
concentrations can be comparable to or greater than spring concentrations. This is significant because
numerous studies of USB and other measures of background ozone have focused on spring as the season
with the greatest USB concentrations, in part because major sources of USB have been reported to make
greater contributions to ozone concentrations in the spring (see Section 1.3.2.1 and Section 1.5.2).
• Recent publications reached conflicting conclusions about seasonal trends in USB. Higher
seasonal mean USB concentrations in spring than in winter were reported for intermountain
western sites (Fiorc et al.. 2014).
• Fiore et al. (2014) reported higher seasonal mean NAB concentrations in spring than in summer
at high-elevation western U.S. sites, consistent with the 2013 Ozone ISA (U.S. EPA. 2013).
• Region-wide seasonal mean USB concentrations greater in summer than spring were reported for
most U.S. regions (Guo et al.. 2018). Improvement of isoprene-NOx chemistry was proposed as
the reason for the difference in results compared with earlier modeling results like those of Fiore
et al. (2014).
• Jaffe et al. (2018) reported comparable median spring and summer baseline ozone concentrations
at elevations >1 km in the western U.S., while below 1-km baseline ozone concentrations were
higher in spring.
• These patterns in seasonal mean USB concentrations are important for identifying atmospheric
processes leading to high USB concentrations and for understanding total ozone exposures over
long periods but are less relevant for estimating USB concentrations on days with high MDA8
concentrations.
1.8.2.3 U.S. Background (USB) Contribution to Ambient Air Ozone as a Function of
Ozone Concentration
USB estimates generally make up a decreasing fraction of total ozone concentration, with
increasing total ozone concentrations in the eastern U.S. and at urban locations in the western U.S. Fiore
et al. (2002) first described model results of lower background concentrations under conditions favorable
for accumulation of high ozone concentrations. Before definitions of USB, NAB, or natural background
had been established (see Section 1.2.2.1). they defined background ozone as ozone produced outside of
the U.S. boundary layer, and they estimated average afternoon background ozone concentrations ranging
from 15-30 ppb in the eastern U.S. and 25-35 ppb in the western U.S., but only 15 ppb during stagnant
meteorological conditions.
• Figure 1-17 and Figure 1-18 based on CMAQ and CAMx model results of daily USB and total
ozone concentrations from 2007 do not show an inverse relationship between USB and total
1-65
-------
ozone concentrations across the U.S., but the results do show that median MDA8 USB
concentrations do not increase with total ozone concentration above 60 ppb total ozone
concentration, resulting in decreasing predicted relative contributions of USB to total ozone at
higher total ozone concentrations (Dolwick et al.. 2015).
Fiore et al. (2014) also described NAB and observed ozone concentrations as largely uncorrelated
in the eastern U.S., and Guo et al. (2018) reported little difference between average USB
concentration and USB concentrations on the 10 highest ozone days the eastern U.S. Lefohn et al.
(2014) described a decreasing trend of relative EIB contribution with increasing total ozone
concentration.
At low-elevation and urban sites in the western U.S., ozone concentrations estimated as USB,
NAB, or EIB (see Section 1.2.1.1) contributions were also reported to be independent of overall
ozone concentration, resulting in a decreasing relative background contribution with increasing
total ozone concentration (Guo et al.. 2018; Dolwick et al.. 2015; Lefohn et al.. 2014).
In contrast, model results have shown increasing USB and NAB concentrations with increasing
ozone concentration at high-elevation western U.S. sites (Fiore et al.. 2014; Lefohn et al.. 2014).
The absence of an inverse relationship between absolute USB concentration and total ozone
concentration like that described by Fiore et al. (2002) in the modeling results of Dolwick et al.
(2015). Guo et al. (2018). and others is consistent with observed meteorological influences on
ozone concentration (see Section 1.5). As the highest ozone concentrations have decreased, they
might now occur under a wider variety of meteorological conditions and might not be limited to
the stagnation conditions that suppress USB concentrations (see Section 1.5). However, it is still
the case that the relative USB contribution on days with the highest ozone concentrations is
usually predicted to be smaller than the seasonal mean USB contribution.
While the average USB fraction has been shown to decrease on high ozone days compared with
low ozone days, there are some instances of high USB fraction on high ozone days as shown by
outliers above 0.75 in Figure l-17a and Figure 1-18a between 70 and 90 ppb. These are most
often associated with model-predicted ozone events from wildfires.
There is consistent evidence across several studies using different background measurement
approaches that USB or other background concentration estimates on most days with high ozone
concentrations have been generally predicted to be similar to or smaller than seasonal mean USB
ozone estimates in the eastern U.S. and in urban and low-elevation areas of the western U.S., and
an inverse relationship between relative USB contribution and total ozone concentration in these
areas has been consistently predicted. This contrasts with high-elevation locations in the western
U.S., where USB and NAB have been consistently predicted to increase with total ozone
concentration.
1-66
-------
„iin ii iiiiii ii ii ii iii
ra 20 -
<25 25-30 30-35 35-40 40-45 45-50 50-55 55-50 60-65 65-70 70-75 75-80 80-85 85-90 90-95 95-100 >100
Bins of CMAQ Base Model MDA8 Ozone (ppb)
360
CO 40
D
i
i
^ gg EB EH E3 E3 Eil ES E3 EG E3 ¦ h m E3
<25 25-30 30-35 35-40 40-45 46-50 50-55 55«0 60-66 65-70 70-75 75-20 00-86 05-90 90-95 95-100 >100
Bins of CAMx Base Model MDA8 Ozone (ppb)
ppb = parts per billion.
Source: Dolwick et al. (2015).
Figure 1-17 Community Multiscale Air Quality (CMAQ) (a) and Comprehensive
Air Quality Model with Extensions (CAMx) (b) estimates of daily
distributions of bias-adjusted U.S. background (USB) daily max
8-hour avg (MDA8) ozone concentration (parts per billion [ppb])
for the period April-October 2007, binned by base model MDA8
ozone concentration ranges.
1-67
-------
2 0 25
<25 25-30 30-35 35-40 40-45 <5-50 50-55 55-50 60-55 55-70 70-75 75-80 80-85 85-90 90-95 95-100 >100
Bins of CMAQ Base Model MDA8 Ozone (ppb)
0 75"
ro 0 50 ~
^ 0.25
<25 25-30 30-35 35-«0 40-45 45-50 50-55 55-50 50-55 55-70 70-75 75-80 80-85 85-90 90-95 95-100 >100
Bins of CAMx Base Model MDA8 Ozone (ppb)
ppb = parts per billion.
Source: Dolwick et al. (2015).
Figure 1-18 Community Multiscale Air Quality (CMAQ) (a) and Comprehensive
Air Quality Model with Extensions (CAMx) (b) estimates of daily
distributions of bias-adjusted U.S. background (USB) ozone
fraction at monitoring locations across the western U.S. for the
period April-October 2007, binned by base model daily max
8-hour avg (MDA8) ozone concentration ranges.
1-68
-------
1.8.2.4
Long-Term Trends in U.S. Background (USB) and Baseline Ozone
Characterization of long-term trends in USB and baseline ozone presents numerous challenges
because of interannual variability in and complex interactions between precursor emissions,
meteorological events and synoptic patterns, surface deposition, atmospheric circulation, and
stratosphere-troposphere exchange (Young et al.. 2018; Lin et al.. 2015V Further complications arise from
unknown errors in emission inventories, limitations of coarse-resolution models in resolving baseline
conditions, and known and unknown weaknesses in model representation of chemical and physical
processes (Young et al.. 2018). Other challenges are introduced by the short observation period and sparse
geographic coverage of surface ozone monitoring efforts (Parrish et al.. 2017a; Lin et al.. 2015). Satellite
retrievals provide greater spatial coverage at mid tropospheric levels. However, the period of satellite data
collection of 10 years is too short for robust trend analysis (Lin et al.. 2015). and satellites are poorly
suited for detecting ground-level ozone. Despite these limitations, there have been some studies on
long-term trends in USB and baseline ozone. However, it is largely limited to high-elevation sites in the
western U.S. or measurements made aloft, where increasing USB trends were reported until recently. The
most recent analyses suggest that this trend has now slowed or reversed.
• Simulated USB ozone trends exhibit poor agreement with monitoring measurements, as well as
between different global models (Young et al.. 2018; Parrish et al.. 2017a). The relative
importance between weaknesses in model processes, inaccuracies in model inputs, and
inadequate representativeness of measurements as contributors to model disagreement are poorly
understood (Young et al.. 2018). Based on a series of modeling studies to investigate USB trends,
Lin et al. (2015) concluded that accurate quantification of USB requires greater spatial density
and temporal frequency in the observational data used in evaluating and improving models than
currently exist in the western U.S.
• For context, on average, both annual mean and annual 4th highest MDA8 ozone values exhibit a
lack of trend or a decreasing trend at most rural U.S. monitoring sites (Jaffe et al.. 2018; Simon et
al.. 2015). High-elevation western U.S. sites are the exceptions. Until recently, model results and
baseline measurements suggested a long-term increasing trend in both USB and baseline ozone at
high-elevation western U.S. sites in the troposphere in the spring (Lin et al.. 2017; Zhang and
Jaffe. 2017; Gratz et al.. 2015; Parrish et al.. 2014; Cooper et al.. 2012; Parrish et al.. 2012).
• An estimated increase of 0.3 to 0.5 ppb/year of USB in spring over the western U.S. in the two
decades after 1990 was largely attributed to a tripling of Asian NOx emissions (Section 1.3.1).
with a smaller contribution for increasing global methane concentrations rSection 1.3.1; Lin et al.
(2017)1.
• Although interannual variability makes it difficult to evaluate, there is evidence from baseline
monitoring, satellite retrievals, and chemical transport modeling that the ozone resulting from
transport from Asia (Section 1.3.1) reached a maximum before 2012 and has been decreasing
since then (Parrish et al.. 2017a; Oetien et al.. 2016). probably as a result of well-documented
decreasing Asian precursor emissions (Liu et al.. 2017a; Duncan et al.. 2016; Krotkov et al..
2016).
• The existing literature on USB trends has focused mainly on monthly or seasonal means. Model
uncertainties are higher, but have not been quantitatively estimated for metrics based on shorter
1-69
-------
averaging times like MDA8 (Jaffc et al.. 2018). and modeling capabilities for reproducing ozone
trends might be different between mean values and other percentiles (Young et al.. 2018).
• There is little evidence to suggest that USB is still increasing even in the western U.S. Analyses
have been largely limited to the western U.S. high elevations, and these appear to show signs of
slowing or even reversing, although this should be considered in the context of high interannual
variability, poor model agreement, and sparse monitoring coverage that present serious
challenges for USB trends analysis.
1.9 Summary
This Appendix reviews scientific advances in atmospheric ozone research relevant to this review
of the NAAQS for ozone and other photochemical oxidants and the related air quality criteria. Strong
emphasis is placed on new evidence concerning the contributions of ozone from natural and non-U. S.
sources.
• For this assessment, U.S. background (USB) ozone is defined as ambient ozone that would be
present at ground level within the U.S. in the absence of all U.S. anthropogenic ozone precursor
emissions. Major contributors to ground-level USB ozone concentrations are stratospheric
exchange, international transport, wildfires, lightning, global methane emissions, and natural
biogenic and geogenic precursor emissions (Section 1.2).
• Ozone formed in the troposphere is, primarily, the product of photochemical reactions between
NOx and carbon-containing compounds including VOCs, CO, and methane. Major source sectors
that emit these ozone precursors include: motor vehicles, EGUs, other industrial processes
involving fuel combustion, agricultural processes, wildfires, and vegetation. These emissions can
be emitted within or outside the U.S. Emissions trends vary by pollutant, source sector, and
source location. Domestic anthropogenic emissions of ozone precursors have largely declined
over the past 15-20 years (Section 1.3).
• While ozone is ordinarily a warm-season pollutant, unusually high concentration ozone events
have occurred in the winter in two western mountain basins, the Uinta and Upper Green River
Basins. Local winter meteorology and high emissions from oil and gas extraction operations
appear to be the principal drivers of winter ozone formation, in these locations.
• Continuing research on the role of halogen chemistry in boundary-layer ozone concentrations
indicates that the process may serve as an ozone sink in coastal urban environments. When added
to model chemical mechanisms, halogen chemistry appears to correct previous overprediction of
ozone concentrations in certain regions (Section 1.4).
• The effects of local precursor emissions controls can be masked by meteorological variability.
Interannual variability in climate has been shown to play a role in influencing total ozone
concentrations across the U.S., as well as USB levels. The El Nino-Southern Oscillation cycle
directly affects the frequency of springtime stratosphere-troposphere exchange events, as well as
the efficiency of international air pollutant transfer processes impacting the western U.S.
(Section 1.5).
• Ozone measurement capabilities have improved since the previous ozone assessment, including
establishment of a new FRM and enhanced use of satellite-based remote sensing methods. At the
same time, there have been notable advances in regional CTM methods, including improvement
in characterizing halogen chemistry, land cover, near surface meteorology, dry deposition,
1-70
-------
stratosphere-troposphere exchange, biogenic emissions, and integration with meteorological
models (Section 1.6).
• For the 2015-2017 time period, the 98th percentile MDA1, MDA8, and DA24 concentrations are
78, 69, and 53 ppb, respectively. Overthe same time period, median 2015-2017 MDA1, MDA8,
and DA24 ozone concentrations are 44, 40, and 30 ppb, respectively. Nationally, the ambient
ozone concentration distribution is compressing as 95th percentile concentrations are decreasing
at the same time 5th percentile concentrations are increasing. This change is consistent with
expected reductions in NOx, which destroys ozone at low ozone concentrations and produces
ozone at higher ozone concentrations (Section 1.7).
• Models consistently predict higher USB ozone concentrations at higher elevations in the western
U.S. than in the eastern U.S. or along the Pacific coast. Across the ensemble of available
modeling studies in the literature, seasonal mean USB concentrations are estimated to range from
20-50 ppb. These model estimates of seasonal mean USB ozone contain uncertainties of about
10 ppb for seasonal average concentrations with higher uncertainty for max daily 8-hour avg
concentrations. Uncertainties in emissions, transport processes, and chemistry contribute to model
result uncertainties. There have been continued efforts to improve model performance and better
understand biases and uncertainties involved in the application of CTMs to estimating USB. With
the exception of high-elevation locations in the western U.S., model simulations suggest that
domestic anthropogenic sources have a greater proportional contribution on the highest ozone
days. Trends in baseline ozone levels suggested a rising contribution from natural and
international sources through approximately 2010. Recently, however, this trend has shown signs
of slowing or even reversing, possibly due to decreasing East Asian precursor emissions
(Section 1.8).
1-71
-------
1.10 References
Abatzoglou. JT; Kolden. CA; Balch. JK; Bradley. BA. (2016). Controls on interannual variability in
lightning-caused fire activity in the western US [Letter], Environ Res Lett 11.
http://dx.doi.Org/10.1088/1748-9326/ll/4/045005
Ahmadov. R; Mckeen. S; Trainer. M; Banta. R; Brewer. A; Brown. S; Edwards. PM; de Gouw. JA;
Frost. GJ; Gilman. J; Helmig. D; Johnson. B; Karion. A; Koss. A; Langford. A; Lerner. B; Olson.
J; Oltmans. S; Peischl. J; Petron. G; Pichugina. Y; Roberts. JM; Rverson. T; Schnell. R; Senff. C;
Sweeney. C; Thompson. C; Veres. PR; Warneke. C; Wild. R; Williams. EJ; Yuan. B; Zamora. R.
(2015). Understanding high wintertime ozone pollution events in an oil- and natural gas-producing
region of the western US. Atmos Chem Phys 15: 411-429. http ://dx.doi .org/10.5194/acp-15-411-
2015
Albers. JR; Perlwitz. J; Butler. A. mvH; Birner. T; Kiladis. GN; Lawrence. ZD; Mannev. GL;
Langford. AO; Dias. J. (2018). Mechanisms governing interannual variability of stratosphere-to-
troposphere ozone transport. J Geophys Res Atmos 123: 234-260.
http://dx.doi.org/10.1002/2017JD02689Q
Alvarez. RJ: Senff. CJ; Langford. AO; Weickmann. AM; Law. DC; Machol. JL; Merritt. DA;
Marchbanks. RD; Sandberg. SP; Brewer. WA; Hardestv. RM; Banta. RM. (2011). Development
and Application of a Compact, Tunable, Solid-State Airborne Ozone Lidar System for Boundary
Layer Profiling. J Atmos Ocean Tech 28: 1258-1272. http://dx.doi.Org/10.l 175/JTECH-D-10-
05044.1
Angevine. WM; Eddington. L. ee; Durkee. K; Fairall. C; Bianco. L; Brioude. J. (2012).
Meteorological Model Evaluation for CalNex 2010. Mon Weather Rev 140: 3885-3906.
http://dx.doi.Org/10.1175/MWR-D-12-00042.l
Appel. KW; Napelenok. SL; Foley. KM; Pve. HOT; Hogrefe. C; Luecken. DJ; Bash. JO; Roselle. SJ;
Pleim. JE; Foroutan. H; Hutzell. WT; Pouliot. GA; Sarwar. G; Fahev. KM; Gantt. B; Gilliam. RC;
Heath. NK; Kang. D; Mathur. R; Schwede. DB; Spero. TL; Wong. DC; Young. JO. (2017).
Description and evaluation of the Community Multiscale Air Quality (CMAQ) modeling system
version 5.1. GMD 10: 1703-1732. http://dx.doi.org/10.5194/gmd-10-1703-2017
Astitha. M; Luo. H; Rao. ST; Hogrefe. C; Mathur. R; Kumar. N. (2017). Dynamic evaluation of two
decades ofWRF-CMAQ ozone simulations over the contiguous United States. Atmos Environ 164:
102-116. http://dx.doi.Org/10.1016/i.atmosenv.2017.05.020
Bader. W; Bow. B; Conway. S; Strong. K; Smale. D. an; Turner. AJ; Blumenstock. T; Boone. C;
Coen. MC; Coulon. A; Garcia. O; Griffith. DT; Hase. F; Hausmann. P; Jones. N; Krummel. P;
Murata. I; Morino. I; Nakaiima. H; O'Dohertv. S; Paton-Walsh. C; Robinson. J; Sandrin. R;
Schneider. M; Servais. C; Sussmann. R; Mahieu. E. (2017). The recent increase of atmospheric
methane from 10 years of ground-based NDACC FTIR observations since 2005. Atmos Chem
Phys 17: 2255-2277. http://dx.doi.org/10.5194/acp-17-2255-2017
1-72
-------
Barth. MC; Cantrell. CA; Brune. WH; Rutlcdgc. SA; Crawford. JH; Huntrieser. H; Carey. LP;
Macgorman. D; Weisman. M; Pickering. KE; Briming. E; Anderson. B; Apel. E; Biggcrstaff. M;
Campos. T; Campuzano-Jost. P; Cohen. R; Crounse. J; Day. DA; Diskin. G; Flocke. F; Fried. A;
Garland. C; Heikes. B; Honomichl. S; Hornbrook. R; Huev. LG; Jimenez. JL; Lang. T;
Lichtenstern. M; Mikovinv. T; Nault. B; O'Sullivan. D; Pan. LL; Peischl. J; Pollack. I; Richter. D;
Riemer. D; Rverson. T; Schlager. H; St Clair. J; Walega. J; Weibring. P; Weinheimer. A;
Wennberg. P; Wisthaler. A; Wooldridge. PJ; Ziegler. C. (2015). The deep convective clouds and
chemistry (DC3) field campaign. Bull Am Meteorol Soc 96: 1281-1309.
http://dx.doi.Org/10.1175/BAMS-D-13-00290.l
Bash. JO: Baker. KR: Beaver. MR. (2016). Evaluation of improved land use and canopy
representation in BEIS v3.61 with biogenic VOC measurements in California. GMD 9: 2191-2207.
http ://dx.doi .org/10.5194/gmd-9-2191-2016
Beirle. S: Borger. C: Dorner. S: Li. A: Hu. Z: Liu. F: Wang. Y: Wagner. T. (2019). Pinpointing
nitrogen oxide emissions from space. Science Advances 5: 1-6.
http://dx.doi.org/10.1126/sciadv.aax9800
Blanchard. CL: Shaw. SL: Edgerton. ES: Schwab. JJ. (2019). Emission influences on air pollutant
concentrations in New York State: I. ozone. Atmospheric Environment: X 3: 100033.
http://dx.doi.Org/10.1016/i.aeaoa.2019.100033
Blavlock. BK: Horel. JD: Crosman. ET. (2017). Impact of Lake Breezes on Summer Ozone
Concentrations in the Salt Lake Valley. J Appl Meteor Climatol 56: 353-370.
http://dx.doi.Org/10.1175/JAMC-D-16-0216.l
Brev. SJ: Fischer. EV. (2016). Smoke in the city: How often and where does smoke impact
summertime ozone in the United States? Environ Sci Technol 50: 1288-1294.
http://dx.doi.org/10.1021/acs.est.5b05218
Busilacchio. M; Di Carlo. P; Aruffo. E; Biancofiore. F; Salisburgo. CD: Giammaria. F; Bauguitte. S:
Lee. J: Moller. S: Hopkins. J: Punjabi. S: Andrews. S: Lewis. AC: Parrington. M: Palmer. PI: Hver.
E: Wolfe. GM. (2016). Production of peroxy nitrates in boreal biomass burning plumes over
Canada during the BORTAS campaign. Atmos Chem Phys 16: 3485-3497.
http://dx.doi.org/10.5194/acp-16-3485-2016
Butler. TJ: Vermevlen. FM: Rurv. M: Likens. GE: Lee. B: Bowker. GE: Mcclunev. L. (2011).
Response of ozone and nitrate to stationary source NO(x) emission reductions in the eastern USA.
Atmos Environ 45: 1084-1094. http://dx.doi.Org/10.1016/i.atmosenv.2010.l 1.040
Bvun. D: Schere. KL. (2006). Review of the governing equations, computational algorithms, and other
components of the models-3 community multiscale air quality (CMAQ) modeling system
[Review]. Appl Mech Rev 59: 51-77. http://dx.doi.org/10.1115/1.2128636
Carlton. AG: Baker. KR. (2011). Photochemical modeling of the Ozark isoprene volcano: MEGAN,
BEIS, and their impacts on air quality predictions. Environ Sci Technol 45: 4438-4445.
http://dx.doi.org/10.1021/es200Q50x
CCSP (U.S. Climate Change Science Program). (2003). Strategic plan for the U.S. climate change
science program: A report by the climate change science program and the subcommittee on global
change research. http://www.climatescience.gov/Librarv/stratplan2003/final/ccspstratplan20Q3-
all.pdf
Cheadle. LC: Oltmans. SJ: Petron. G: Schnell. RC: Mattson. EJ: Herndon. SC: Thompson. AM: Blake.
PR: Mcclure-Beglev. A. (2017). Surface ozone in the Colorado northern Front Range and the
influence of oil and gas development during FRAPPE/DISCOVER-AQ in summer 2014. 5.
http://dx.doi.org/10.1525/elementa.254
1-73
-------
Clifton. OE; Fiore. AM; Correa. G; Horowitz. LW; Naik. V. (2014). Twenty-first century reversal of
the surface ozone seasonal cycle over the northeastern United States. Geophys Res Lett 41: 7343-
7350. http://dx.doi.org/10.1002/2014GLQ61378
Cohan. DS; Napelenok. SL. (2011). Air quality response modeling for decision support. Atmosphere
(Basel) 2: 407-425. http://dx.doi.org/10.3390/atmos2030407
Cooper. OR: Gao. RS; Tarasick. D: Leblanc. T: Sweeney. C. (2012). Long-term ozone trends at rural
ozone monitoring sites across the United States, 1990-2010. J Geophys Res Atmos 117.
http://dx.doi.org/10.1029/2012JDQ18261
Cuchiara. GC: Li. X: Carvalho. J: Rappenglueck. B. (2014). Intercomparison of planetary boundary
layer parameterization and its impacts on surface ozone concentration in the WRF/Chem model for
a case study in Houston/Texas. Atmos Environ 96: 175-185.
http://dx.doi.Org/10.1016/i.atmosenv.2014.07.013
Day. M: Pouliot. G: Hunt. S: Baker. KR: Beardslev. M: Frost. G: Moblev. D: Simon. H: Henderson.
BB; Yelverton. T; Rao. V. (2019). Reflecting on progress since the 2005 NARSTO emissions
inventory report. J Air Waste Manag Assoc 69: 1023-1048.
http://dx.doi.org/10.1080/10962247.2Q19.1629363
De Maziere. M: Thompson. AM: Kurvlo. MJ: Wild. JD: Bernhard. G: Blumenstock. T: Braathen. GO:
Hannigan. JW; Lambert. JC: Leblanc. T; Mcgee. TJ; Nedoluha. G: Petropavlovskikh. I;
Seckmever. G: Simon. PC: Steinbrecht. W: Strahan. SE. (2018). The network for the detection of
atmospheric composition change (NDACC): history, status and perspectives. Atmos Chem Phys
18: 4935-4964. http://dx.doi.org/10.5194/acp-18-4935-2018
Dennis. R; Fox. T; Fuentes. M; Gilliland. A: Hanna. S: Hogrefe. C: Irwin. J: Rao. ST: Scheffe. R;
Schere. K: Stevn. D: Venkatram. A. (2010). A framework for evaluating regional-scale numerical
photochemical modeling systems. Environ Fluid Mech 10: 471-489.
http://dx.doi.org/10.1007/slQ652-009-9163-2
Ding. J: Mivazaki. K: Johannes van der A. R: Milling. B: Kurokawa. J: Cho. S: Janssens-Maenhout.
G: Zhang. O: Liu. F: Levelt. PF. (2017). Intercomparison of NOx emission inventories over East
Asia. Atmos Chem Phys 17: 10125-10141. http://dx.doi.org/10.5194/acp-17-10125-2017
Dolwick. P; Akhtar. F; Baker. KR: Possiel. N: Simon. H; Tonnesen. G. (2015). Comparison of
background ozone estimates over the western United States based on two separate model
methodologies. Atmos Environ 109: 282-296. http://dx.doi.Org/l0.1016/i.atmosenv.2015.01.005
Duncan. BN: Lamsal. LN: Thompson. AM: Yoshida. Y; Lu. Z; Streets. DG: Hurwitz. MM: Pickering.
KE. (2016). A space-based, high-resolution view of notable changes in urban NOx pollution
around the world (2005-2014). J Geophys Res Atmos 121: 976-996.
http://dx.doi.org/10.1002/2015JDQ24121
Duncan. BN: Prados. AI; Lamsal. LN: Liu. Y; Streets. DG: Gupta. P: Hilsenrath. E; Kahn. RA;
Nielsen. JE; Beversdorf. AJ; Burton. SP: Fiore. AM; Fishman. J; Henze. DK; Hostetler. CA;
Krotkov. NA; Lee. P; Lin. M; Pawson. S; Pfister. G; Pickering. KE; Pierce. RB; Yoshida. Y;
Ziemba. LP. (2014). Satellite data of atmospheric pollution for US air quality applications:
Examples of applications, summary of data end-user resources, answers to FAQs, and common
mistakes to avoid. Atmos Environ 94: 647-662. http://dx.doi.Org/10.1016/i.atmosenv.2014.05.061
Dunker. AM. (1981). Efficient calculation of sensitivity coefficients for complex atmospheric models.
Atmos Environ 15: 1155-1161. http://dx.doi.org/10.1016/0004-698K8n903Q5-X
1-74
-------
Dunker. AM; Koo. B; Yarwood. G. (2017). Contributions of foreign, domestic and natural emissions
to US ozone estimated using the path-integral method in CAMx nested within GEOS-Chem.
Atmos Chem Phys 17: 12553-12571. http://dx.doi.org/10.5194/acp-17-12553-2017
Edwards. PM; Brown. SS; Roberts. JM; Ahmadov. R; Banta. RM: Pegouw. JA; Dube. WP; Field.
RA; Flvnn. JH; Gilman. JB; Graus. M; Helmig. D; Koss. A; Langford. AO; Lefer. BL; Lerner.
BM; Li. R; Li. SM; Mckeen. SA; Murphy. SM; Parrish. DP; Senff. CJ; Soltis. J; Stutz. J; Sweeney.
C; Thompson. CR; Trainer. MK; Tsai. C; Veres. PR; Washenfelder. RA; Warneke. C; Wild. RJ;
Young. CJ; Yuan. B; Zamora. R. (2014). High winter ozone pollution from carbonyl photolysis in
an oil and gas basin. Nature 514: 351-354. http://dx.doi.org/10.1038/nature 13767
Emery. C; Jung. J; Downey. N: .Johnson. J; Jimenez. M; Yarvvood. 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.Org/10.1016/i.atmosenv.2011.11.012
Emmons. LK; Hess. PG; Lamarque. JF; Pfister. GG. (2012). Tagged ozone mechanism for MOZART-
4, CAM-chem and other chemical transport models. GMD 5: 1531-1542.
http ://dx.doi .org/10.5194/gmd-5 -1531-2012
Fang. Y; Fiore. AM; Horowitz. LW; Lew II. H; Hu. Y; Russell. AG. (2010). Sensitivity of the NOy
budget over the United States to anthropogenic and lightning NOx in summer. J Geophys Res 115.
http://dx.doi.org/10.1029/2010JD014Q79
Farris. BM; Gronoff. GP; Carrion. W; Knepp. T; Pippin. M; Berkoff. TA. (2019). Demonstration of an
off-axis parabolic receiver for near-range retrieval of lidar ozone profiles. Atmos Meas Tech 12:
363-370. http://dx.doi.org/10.5194/amt-12-363-2019
Field. RA; Soltis. J; McCarthy. MC; Murphy. S; Montague. DC. (2015). Influence of oil and gas field
operations on spatial and temporal distributions of atmospheric non-methane hydrocarbons and
their effect on ozone formation in winter. Atmos Chem Phys 15: 3527-3542.
http://dx.doi.org/10.5194/acp-15-3527-2015
Finlavson-Pitts. BJ; Pitts. JN. Jr. (2000). Chemistry of the upper and lower atmosphere: Theory,
experiments and applications. In Chemistry of the upper and lower atmosphere: Theory,
experiments and applications. San Diego, CA: Academic Press. http://dx.doi.org/10.1016/B978-Q-
12-257060-5 .X5000-X
Fiore. A; Jacob. DJ; Liu. H; Yantosca. RM; Fairlie. TP; Li. O. (2003). Variability in surface ozone
background over the United States: Implications for air quality policy. J Geophys Res 108: 4787.
http://dx.doi.org/10.1029/2003JD0Q3855
Fiore. AM; Jacob. DJ; Bev. I; Yantosca. RM; Field. BP; Fusco. AC; Wilkinson. JG (2002).
Background ozone over the United States in summer: origin, trend, and contribution to pollution
episodes. J Geophys Res 107: 4275. http://dx.doi.org/10.1029/2001JP000982
Fiore. AM; Oberman. JT; Lin. MY; Zhang. L; Clifton. OE; Jacob. PJ; Naik. V; Horowitz. LW; Pinto.
JP; Millv. GP. (2014). Estimating North American background ozone in U.S. surface air with two
independent global models: Variability, uncertainties, and recommendations. Atmos Environ 96:
284-300. http://dx.doi.Org/10.1016/i.atmosenv.2014.07.045
Fiore. AM; West. JJ; Horowitz. LW; Naik. V; Schwartzkopf. MP. (2008). Characterizing the
tropospheric ozone response to methane emission controls and the benefits to climate and air
quality. J Geophys Res 113: P08307. http://dx.doi.org/10.1029/2007JP0Q9162
1-75
-------
Fischer. EV; Zhu. L; Payne. VH; Worden. JR; Jiang. Z; Kulawik. SS; Brev. S; Hecobian. A; Gombos.
D; Cadv-Pereira. K; Flocke. F. (2018). Using TES retrievals to investigate PAN in North American
biomass burning plumes. Atmos Chem Phys 18: 5639-5653. http://dx.doi.org/10.5194/acp-18-
5639-2018
Fishman. J; Watson. CE; Larsen. JC; Logan. JA. (1990). Distribution of tropospheric ozone
determined from satellite data. J Geophys Res 95: 3599-3617.
http://dx.doi.org/10.1029/JD095iD04pQ3599
Foley. KM; Hogrefe. C; Pouliot. G; Possiel. N; Roselle. SJ; Simon. H; Timin. B. (2015). Dynamic
evaluation of CMAQ part I: Separating the effects of changing emissions and changing
meteorology on ozone levels between 2002 and 2005 in the eastern US. Atmos Environ 103: 247-
255. http://dx.doi.Org/10.1016/i.atmosenv.2014.12.038
Galmarini. S; Koffi. B; Solazzo. E; Keating. T; Hogrefe. C; Schulz. M; Benedictow. A; Griesfeller. J;
Janssens-Maenhout. G; Carmichael. G; Fu. J; Dentener. F. (2017). Technical note: Coordination
and harmonization of the multi-scale, multi-model activities HTAP2, AQMEII3, and MICS-Asia3:
simulations, emission inventories, boundary conditions, and model output formats. Atmos Chem
Phys 17: 1543-1555. http://dx.doi.org/10.5194/acp-17-1543-2017
Gantt. B: Sarwar. G; Xing. J: Simon. H: Schwede. D; Hutzell. WT: Mathur. R: Saiz-Lopez. A. (2017).
The impact of iodide-mediated ozone deposition and halogen chemistry on surface ozone
concentrations across the continental United States. Environ Sci Technol 51: 1458-1466.
http://dx.doi.org/10.1021/acs.est.6b03556
Goldberg. PL; Lu. Z; Streets. DG; de Foy. B; Griffin. D; Mclinden. CA; Lamsal. LN; Krotkov. NA;
Eskes. H. (2019). Enhanced capabilities of TROPOMIN02: estimating NOx from North American
cities and power plants. Environ Sci Technol 53: 12594-12601.
http://dx.doi.org/10.1021/acs.est.9b04488
Gong. X; Kaulfus. A; Nair. U; Jaffe. DA. (2017). Quantifying 03 impacts in urban areas due to
wildfires using a generalized additive model. Environ Sci Technol 51: 13216-13223.
http://dx.doi.org/10.1021/acs.est.7b03130
Gratz. LE; Jaffe. DA; Hee. JR. (2015). Causes of increasing ozone and decreasing carbon monoxide in
springtime at the Mt. Bachelor Observatory from 2004 to 2013. Atmos Environ 109: 323-330.
http://dx.doi.Org/10.1016/i.atmosenv.2014.05.076
Grewe. V; Tsati. E; Hoor. P. (2010). On the attribution of contributions of atmospheric trace gases to
emissions in atmospheric model applications. GMD 3: 487-499. http://dx.doi.Org/10.5194/gmd-3-
487-2010
Guenther. A; Geron. C; Pierce. T; Lamb. B; Harlev. P; Fall. R. (2000). Natural emissions of non-
methane volatile organic compounds, carbon monoxide, and oxides of nitrogen from North
America. Atmos Environ 34: 2205-2230. http://dx.doi.Org/10.1016/S 1352-2310(99)00465-3
Guo. JJ; Fiore. AM; Murray. LT; Jaffe. DA; Schnell. ,TL; Moore. CT; Millv. GP. (2018). Average
versus high surface ozone levels over the continental USA: Model bias, background influences, and
interannual variability. Atmos Chem Phys 18: 12123-12140. http://dx.doi.org/10.5194/acp-18-
12123-2018
Hakami. A; Odman. MT; Russell. AG. (2004). Nonlinearity in atmospheric response: A direct
sensitivity analysis approach. J Geophys Res Atmos 109: 1-12.
http://dx.doi.org/10.1029/2003JD0045Q2
1-76
-------
Hall. SJ; Reves. L; Huang. W; Homvak. PM. (2018). Wet spots as hotspots: Moisture responses of
nitric and nitrous oxide emissions from poorly drained agricultural soils. Jour Geo Res: Biog 123:
3589-3602. http://dx.doi.org/10.1029/2018JG0Q4629
Helmig. D; Thompson. CR; Evans. J; Boylan. P; Hueber. J; Park. JH. (2014). Highly elevated
atmospheric levels of volatile organic compounds in the Uintah Basin, Utah. Environ Sci Technol
48: 4707-4715. http://dx.doi.org/10.1021/es405Q46r
Henderson. BH; Akhtar. F; Pve. HOT; Napelenok. SL; Hutzell. WT. (2014). A database and tool for
boundary conditions for regional air quality modeling: description and evaluation. GMD 7: 339-
360. http://dx.doi.org/10.5194/gmd-7-339-2014
Henneman. LRF; Holmes. HA; Mulholland. JA; Russell. AG. (2015). Meteorological detrending of
primary and secondary pollutant concentrations: Method application and evaluation using long-
term (2000-2012) data in Atlanta. Atmos Environ 119: 201-210.
http://dx.doi.Org/10.1016/i.atmosenv.2015.08.007
Hickman. JE; Huang. Y; Wu. S; Diru. W; Groffman. PM; Tullv. KL; Palm. CA. (2017). Non-linear
response of nitric oxide fluxes to fertilizer inputs and the impacts of agricultural intensification on
tropospheric ozone pollution in Kenya. Global Change Biol 23: 3193-3204.
http://dx.doi.Org/10.l 111/gcb. 13644
Hoeslv. RM; Smith. SJ; Feng. L; Klimont. Z; Janssens-Maenhout. G; Pitkanen. T; Seibert. JJ; Vu. L;
Andres. RJ; Bolt. RM; Bond. TC; Dawidowski. L; Kholod. N; Kurokawa. JI; Li. M; Liu. L; Lu. Z;
Moura. MCP; O'Rourke. PR; Zhang. O. (2018). Historical (1750-2014) anthropogenic emissions of
reactive gases and aerosols from the Community Emissions Data System (CEDS). GMD 11: 369-
408. http://dx.doi.org/10.5194/gmd-ll-369-2018
Hogrefe. C; Liu. P; Pouliot. G; Mathur. R; Roselle. S; Flemming. J; Lin. M; Park. RJ. (2018). Impacts
of different characterizations of large-scale background on simulated regional-scale ozone over the
continental United States. Atmos Chem Phys 18: 3839-3864. http://dx.doi.org/10.5194/acp-18-
3839-2018
Hu. L; Millet. DB; Baasandori. M; Griffis. TJ; Turner. P; Helmig. D; Curtis. AJ; Hueber. J. (2015).
Isoprene emissions and impacts over an ecological transition region in the US Upper Midwest
inferred from tall tower measurements. J Geophys Res Atmos 120: 3553-3571.
http://dx.doi.org/10.1002/2014JDQ22732
Huang. M; Bowman. KW; Carmichael. GR; Pierce. RB; Worden. HM; Luo. M; Cooper. OR; Pollack.
IB; Rverson. TB; Brown. SS. (2013a). Impact of Southern California anthropogenic emissions on
ozone pollution in the mountain states: Model analysis and observational evidence from space. J
Geophys Res Atmos 118: 12784-12803. http://dx.doi.org/10.1002/2013JD0202Q5
Huang. M; Carmichael. GR; Chai. T; Pierce. RB; Oltmans. SJ; Jaffe. DA; Bowman. KW; Kaduwela.
A; Cai. C; Spak. SN; Weinheimer. AJ; Huev. LG; Diskin. GS. (2013b). Impacts of transported
background pollutants on summertime western US air quality: model evaluation, sensitivity
analysis and data assimilation. Atmos Chem Phys 13: 359-391. http://dx.doi.org/10.5194/acp-13-
359-2013
Huntrieser. H; Lichtenstern. M; Scheibe. M; Aufmhoff. H; Schlager. H; Pucik. T; Minikin. A:
Weinzierl. B; Heimerl. K; Fuetterer. D; Rappengluck. B; Ackermann. L; Pickering. KE;
Cummings. KA: Biggerstaff. MI; Betten. DP; Honomichl. S; Barth. MC. (2016). On the origin of
pronounced 0-3 gradients in the thunderstorm outflow region during DC3. J Geophys Res Atmos
121: 6600-6637. http://dx.doi.org/10.1002/2015JDQ24279
1-77
-------
Im. U; Bianconi. R; Solazzo. E; Kioutsioukis. I; Badia. A; Balzarini. A; Baro. R; Bellasio. R; Brunner.
D; Chemel. C; Curci. G; Flemming. J; Forkel. R; Giordano. L; Jimenez-Guerrero. P; Hirtl. M;
Hodzic. A; Honzak. L; Jorba. O; Knote. C; Kuenen. JJP; Makar. PA; Manders-Groot. A; Neal. L;
Perez. JL; Pirovano. G; Pouliot. G; San Jose. R; Savage. N; Schroder. W; Sokhi. RS; Svrakov. D;
Torian. A; Tuccella. P; Werhahn. J; Wolke. R; Yahva. K; Zabkar. R; Zhang. Y; Zhang. J; Hogrcfe.
C; Galmarini. S. (2015). Evaluation of operational on-line-coupled regional air quality models over
Europe and North America in the context of AQMEII phase 2. Part I: Ozone. Atmos Environ 115:
404-420. http://dx.doi.Org/10.1016/i.atmosenv.2014.09.042
Jaffe. DA: Cooper. OR: Fiore. AM: Henderson. BH: Tonnesen. GS; Russell. AG; Henze. DK:
Langford. AO; Lin. M; Moore. T. (2018). Scientific assessment of background ozone over the US:
Implications for air quality management. Elementa: Science of the Anthropocene 6.
http://dx.doi.org/10.1525/elementa.309
Jaffe. DA; Wigder. NL. (2012). Ozone production from wildfires: A critical review. Atmos Environ
51: 1-10. http://dx.doi.Org/10.1016/i.atmosenv.2011.l 1.063
Jin. X; Fiore. AM; Murray. L; Valin. LC; Lamsal. L; Duncan. B; Folkert Boersma. K; De Smedt. I;
Abad. GG; Chance. K; Tonnesen. GS. (2017). Evaluating a Space-Based Indicator of Surface
Ozone-NOx-VOC Sensitivity Over Midlatitude Source Regions and Application to Decadal
Trends. J Geophys Res Atmos 122: 10231-10253. http://dx.doi.org/10.1002/2017JD02672Q
Judd. LM; Al-Saadi. JA; Janz. SJ; Kowalewski. MG; Pierce. RB; Szvkman. JJ; Valin. LC; Swap. R;
Cede. A; Mueller. M; Tiefengraber. M; Abuhassan. N; Williams. D. (2019). Evaluating the impact
of spatial resolution on tropospheric N02 column comparisons within urban areas using high-
resolution airborne data. Atmos Meas Tech 12: 6091-6111. http://dx.doi.org/10.5194/amt-12-6Q91-
2019
Karamchandani. P; Morris. R; Wentland. A; Shah. T; Reid. S; Lester. J. (2017). Dynamic evaluation
of photochemical grid model response to emission changes in the south coast air basin in
California. Atmosphere (Basel) 8. http://dx.doi.org/10.3390/atmos808Q145
Kirschke. S; Bousquet. P; Ciais. P; Saunois. M; Canadell. JG; Dlugokenckv. EJ; Bergamaschi. P;
Bergmann. D; Blake. PR; Bruhwiler. L; Cameron-Smith. P; Castaldi. S; Chevallier. F; Feng. L;
Fraser. A; Heimann. M; Hodson. EL; Houweling. S; Josse. B; Fraser. PJ; Krummel. PB; Lamarque.
JF; Langenfelds. R, avL; Le Ouere. C; Naik. V; O'Dohertv. S; Palmer. PI; Pison. I; Plummer. D;
Poulter. B; Prinn. RG; Rigbv. M; Ringeval. B; Santini. M; Schmidt M; Shindell. DT; Simpson. IJ;
Spahni. R; Steele. LP; Strode. SA; Sudo. K; Szopa. S; van Per Werf. GR; Voulgarakis. A; van
Weele. M; Weiss. RF; Williams. JE; Zeng. G. (2013). Three decades of global methane sources
and sinks. Nat Geosci 6: 813-823. http://dx.doi.org/10.1038/NGEQ1955
Knote. C; Tuccella. P; Curci. G; Emmons. L; Orlando. JJ; Madronich. S; Baro. R; Jimenez-Guerrero.
P; Luecken. P; Hogrefe. C; Forkel. R; Werhahn. J: Hirtl. M; Perez. JL; San Jose. R; Giordano. L;
Brunner. P; Yahva. K; Zhang. Y. (2015). Influence of the choice of gas-phase mechanism on
predictions of key gaseous pollutants during the AQMEII phase-2 intercomparison. Atmos Environ
115: 553-568. http://dx.doi.Org/10.1016/i.atmosenv.2014.l 1.066
Knowland. KE; Ott. LE; Puncan. BN; Wargan. K. (2017). Stratospheric intrusion-influenced ozone air
quality exceedances investigated in the NASA MERRA-2 reanalysis. Geophys Res Lett 44: 10691-
10701. http://dx.doi.org/10.1002/2017GLQ74532
Kota. SH; Schade. G; Estes. M; Bover. P; Ying. O. (2015). Evaluation of MEGAN predicted biogenic
isoprene emissions at urban locations in Southeast Texas. Atmos Environ 110: 54-64.
http://dx.doi.Org/l 0.1016/j .atmosenv.2015.03.027
1-78
-------
Krotkov. NA; McLinden. CA; Li. C; Lamsal. LN; Celarier. EA; Marchenko. SV; Swartz. WH;
Bucsela. EJ; Joiner. J; Duncan. BN; Boersma. KF; Veefkind. JP; Levelt. PF; Fioletov. YE;
Dickerson. RR: He. H; Lu. Z; Streets. DG. (2016). Aura OMI observations of regional S02 and
N02 pollution changes from 2005 to 2015. Atmos Chem Phys 16: 4605-4629.
http://dx.doi.org/10.5194/acp-16-4605-2Q16
Kuang. S. hi: Newchurch. MJ: Burris. J; Liu. X. (2013). Ground-based lidar for atmospheric boundary
layer ozone measurements. Appl Opt 52: 3557-3566. http://dx.doi.org/10.1364/AQ.52.003557
Kwok. RHF; Baker. KR; Napelenok. SL; Tonnesen. GS. (2015). Photochemical grid model
implementation and application of VOC, NOx, and 0-3 source apportionment. GMD 8: 99-114.
http://dx.doi.org/10.5194/gmd-8-99-2015
Laing. JR; Jaffe. DA. (2019). Wildfires are causing extreme PM concentrations in the western U.S
[Magazine]. EM: Air And Waste Management Association's Magazine For Environmental
Managers, June 2019.
Lamsal. LN: Duncan. BN: Yoshida. Y; Krotkov. NA: Pickering. KE; Streets. DG: Lu. Z. (2015). U.S.
N02 trends (2005-2013): EPA Air Quality System (AQS) data versus improved observations from
the Ozone Monitoring Instrument (OMI). Atmos Environ 110: 130-143.
http://dx.doi.Org/10.1016/i.atmosenv.2015.03.055
Langford. AO. (1999). Stratosphere-troposphere exchange at the subtropical jet: contribution to the
tropospheric ozone budget at midlatitudes. Geophys Res Lett 26: 2449-2452.
Langford. AO: Aikin. KC: Eubank. CS: Williams. EJ. (2009). Stratospheric contribution to high
surface ozone in Colorado during springtime. Geophys Res Lett 36: L12801.
http://dx.doi.org/10.1029/2009glQ38367
Langford. AO: Alvarez. RJ: Brioude. J: Fine. R: Gustin. MS: Lin. MY: Marchbanks. RD: Pierce. RB:
Sandberg. SP: Senff. CJ: Weickmann. AM: Williams. EJ. (2017). Entrainment of stratospheric air
and Asian pollution by the convective boundary layer in the southwestern US. J Geophys Res
Atmos 122: 1312-1337. http://dx.doi.org/10.1002/2016JDQ25987
Langford. AO: Senff. CJ: Alvarez. RJ: Brioude. J: Cooper. OR: Hollowav. JS: Lin. MY: Marchbanks.
RD: Pierce. RB: Sandberg. SP: Weickmann. AM: Williams. EJ. (2015). An overview of the 2013
Las Vegas Ozone Study (LVOS): Impact of stratospheric intrusions and long-range transport on
surface air quality. Atmos Environ 109: 305-322. http://dx.doi.Org/10.1016/i.atmosenv.2014.08.040
Leblanc. T; Brewer. MA: Wang. PS: Jose Granados-Munoz. M; Strawbridge. KB: Travis. M; Firanski.
B: Sullivan. JT: Mcgee. TJ: Sumnicht. GK: Twigg. LW: Berkoff. TA: Carrion. W: Gronoff. G:
Aknan. A. li: Chen. G. ao: Alvarez. RJ: Langford. AO: Senff. CJ: Kirgis. G: Johnson. MS: Kuang.
S. hi: Newchurch. MJ. (2018). Validation of the TOLNet lidars: the Southern California Ozone
Observation Project (SCOOP). Atmos Meas Tech 11: 6137-6162. http://dx.doi.org/10.5194/amt-
11-6137-2018
Leblanc. T: Sica. RJ: van Gijsel. JAE: Godin-Beekmann. S: Haefele. A: Trickl. T: Paven. G: Gabarrot.
F. (2016a). Proposed standardized definitions for vertical resolution and uncertainty in the NDACC
lidar ozone and temperature algorithms - Part 1: Vertical resolution. Atmos Meas Tech 9: 4029-
4049. http://dx.doi.org/10.5194/amt-9-4029-2016
Leblanc. T; Sica. RJ: van Gijsel. JAE: Godin-Beekmann. S: Haefele. A: Trickl. T; Paven. G: Liberti.
G. (2016b). Proposed standardized definitions for vertical resolution and uncertainty in the
NDACC lidar ozone and temperature algorithms - Part 2: Ozone DIAL uncertainty budget. Atmos
Meas Tech 9: 4051-4078. http://dx.doi.org/10.5194/amt-9-4051-2016
1-79
-------
Lee. L; Teng. AP; Wennberg. PO; Crounse. JD; Cohen. RC. (2014). On rates and mechanisms of OH
and 03 reactions with isoprene-derived hydroxy nitrates. J Phys Chem A 118: 1622-1637.
http://dx.doi.org/10.1021/ip41076Q3
Lefohn. AS; Emery. C; Shadwick. D; Wernli. H; Jung. J; Oltmans. SJ. (2014). Estimates of
background surface ozone concentrations in the United States based on model-derived source
apportionment. Atmos Environ 84: 275-288. http://dx.doi.Org/10.1016/i.atmosenv.2013.l 1.033
Lefohn. AS; Mallev. CS; Simon. H: Wells. B: Xu. X: Zhang. L: Wang. T. (2017). Responses of
human health and vegetation exposure metrics to changes in ozone concentration distributions in
the European Union, United States, and China. Atmos Environ 152: 123-145.
http://dx.doi.Org/10.1016/i.atmosenv.2016.12.025
Lefohn. AS; Wernli. H; Shadwick. D; Oltmans. SJ; Shapiro. M. (2012). Quantifying the importance of
stratospheric-tropospheric transport on surface ozone concentrations at high- and low-elevation
monitoring sites in the United States. Atmos Environ 62: 646-656.
http://dx.doi.Org/10.1016/i.atmosenv.2012.09.004
Li. J; Mao. J; Fiore. AM; Cohen. RC; Crounse. JD; Teng. AP; Wennberg. PO; Lee. BH; Lopez-
Hilfiker. FD; Thornton. JA; Peischl. J; Pollack. IB; Rverson. TB; Veres. P; Roberts. JM; Neuman.
JA; Nowak. JB; Wolfe. GM; Hanisco. TF; Fried. A; Singh. HB; Dibb. J; Paulot. F; Horowitz. LW.
(2018). Decadal changes in summertime reactive oxidized nitrogen and surface ozone over the
Southeast United States. Atmos Chem Phys 18: 2341-2361. http://dx.doi.org/10.5194/acp-18-2341-
2018
Lin. M; Fiore. AM; Cooper. OR; Horowitz. LW; Langford. AO; Levy. H; Johnson. BJ; Naik. V;
Oltmans. SJ; Senff. CJ. (2012a). Springtime high surface ozone events over the western United
States: Quantifying the role of stratospheric intrusions. J Geophys Res 117: D00V22.
http://dx.doi.org/10.1029/2012JD018151
Lin. M; Fiore. AM; Horowitz. LW; Cooper. OR; Naik. V; Hollowav. J; Johnson. BJ; Middlebrook.
AM; Oltmans. SJ; Pollack. IB; Rverson. TB; Warner. JX; Wiedinmver. C; Wilson. J; Wvman. B.
(2012b). Transport of Asian ozone pollution into surface air over the western United States in
spring. J Geophys Res 117: D00V07. http://dx.doi.org/10.1029/2011JD016961
Lin. M; Fiore. AM; Horowitz. LW; Langford. AO; Oltmans. SJ; Tarasick. D; Rieder. HE. (2015).
Climate variability modulates western US ozone air quality in spring via deep stratospheric
intrusions. Nat Commun 6: 7105. http://dx.doi.org/10.1038/ncomms8105
Lin. M; Horowitz. LW; Oltmans. SJ; Fiore. AM; Fan. S. (2014). Tropospheric ozone trends at Mauna
Loa Observatory tied to decadal climate variability. Nat Geosci 7: 136-143.
http://dx.doi.org/10.1038/NGEQ2066
Lin. M; Horowitz. LW; Pavton. R; Fiore. AM; Tonnesen. G. (2017). US surface ozone trends and
extremes from 1980 to 2014: quantifying the roles of rising Asian emissions, domestic controls,
wildfires, and climate. Atmos Chem Phys 17: 2943-2970. http://dx.doi.org/10.5194/acp-17-2943-
2017
Lindaas. J; Farmer. DK; Pollack. IB; Abeleira. A; Flocke. F; Roscioli. R; Herndon. S; Fischer. EV.
(2017). Changes in ozone and precursors during two aged wildfire smoke events in the Colorado
Front Range in summer 2015. Atmos Chem Phys 17: 10691-10707. http://dx.doi.org/10.5194/acp-
17-10691-2017
Liu. F; Beirle. S; Zhang. 0; van der A. RJ; Zheng. B; Tong. D; He. K. (2017a). NOx emission trends
over Chinese cities estimated from OMI observations during 2005 to 2015. Atmos Chem Phys 17:
9261-9275. http://dx.doi.org/10.5194/acp-17-9261-2017
1-80
-------
Liu. Z; Liu. Y; Murphy. JP; Maghirang. R. (2017b). Contributions of Kansas rangeland burning to
ambient 03: Analysis of data from 2001 to 2016. Sci Total Environ 618: 1024-1031.
http://dx.doi.Org/10.1016/i.scitotenv.2017.09.075
Lu. R; Turco. RP. (1995). Air pollutant transport in a coastal environmentll. Three-dimensional
simulations over Los Angeles basin. Atmos Environ 29: 1499-1518.
http://dx.doi.org/10.1016/1352-2310(95)00015-0
Mao. J; Horowitz. LW; Naik. V: Fan. S; Liu. J; Fiore. AM. (2013). Sensitivity of tropospheric
oxidants to biomass burning emissions: implications for radiative forcing. Geophys Res Lett 40:
1241-1246. http://dx.doi.org/10.1002/grl.5021Q
Martin. RV. (2008). Satellite remote sensing of surface air quality. Atmos Environ 42: 7823-7843.
http://dx.doi.Org/10.1016/i.atmosenv.2008.07.018
Mathur. R; Xing. J; Gilliam. R; Sarwar. G; Hogrefe. C; Pleim. J; Pouliot. G; Roselle. S; Spero. TL;
Wong. DC; Young. J. (2017). Extending the Community Multiscale Air Quality (CMAQ)
modeling system to hemispheric scales: overview of process considerations and initial applications.
Atmos Chem Phys 17: 12449-12474. http://dx.doi.org/l0.5194/acp-17-12449-2017
McDonald-Buller. EC; Allen. DT: Brown. N; Jacob. DJ; Jaffe. D; Kolb. CE: Lefohn. AS; Oltmans. S;
Parrish. DP; Yarwood. G; Zhang. L. (2011). Establishing policy relevant background (PRB) ozone
concentrations in the United States [Review]. Environ Sci Technol 45: 9484-9497.
http://dx.doi.org/10.1021/es2022818
Mebust. AK; Cohen. RC. (2014). Space-based observations of fire NOx emission coefficients: a
global biome-scale comparison. Atmos Chem Phys 14: 2509-2524. http://dx.doi.Org/10.5194/acp-
14-2509-2014
Mebust. AK; Russell. AR; Hudman. RC; Valin. LC; Cohen. RC. (2011). Characterization of wildfire
NOx emissions using MODIS fire radiative power and OMI tropospheric N02 columns. Atmos
Chem Phys 11: 5839-5851. http://dx.doi.org/10.5194/acp-ll-5839-2011
Monks. PS; Archibald. AT; Colette. A; Cooper. O; Covle. M; Derwent. R; Fowler. D; Granier. C;
Law. KS; Mills. GE; Stevenson. DS; Tarasova. O; Thouret. V; von Schneidemesser. E;
Sommariva. R; Wild. O; Williams. ML. (2015). Tropospheric ozone and its precursors from the
urban to the global scale from air quality to short-lived climate forcer. Atmos Chem Phys 15: 8889-
8973. http://dx.doi.org/10.5194/acp-15-8889-2015
Muniz-Unamunzaga. M; Borge. R; Sarwar. G; Gantt. B; de la Paz. D; Cuevas. CA; Saiz-Lopez. A.
(2018). The influence of ocean halogen and sulfur emissions in the air quality of a coastal
megacity: The case of Los Angeles. Sci Total Environ 610-611: 1536-1545.
http://dx.doi.Org/10.1016/i.scitotenv.2017.06.098
Murray. LT; Jacob. DJ; Logan. JA; Hudman. RC; Koshak. WJ. (2012). Optimized regional and
interannual variability of lightning in a global chemical transport model constrained by LIS/OTD
satellite data. J Geophys Res 117: 1-14. http://dx.doi.org/10.1029/2012JD017934
NAPCA (National Air Pollution Control Administration). (1970). Air quality criteria for
photochemical oxidants [EPA Report]. (AP-63). Washington, DC: U.S. Department of Health,
Education, and Welfare, http://nepis.epa.gov/Exe/ZvPURL.cgi?Dockev=9100E2Z7.txt
NASEM (National Academies of Sciences, Engineering, and Medicine). (2018). Improving
characterization of anthropogenic methane emissions in the United States. Washington, D.C.:
National Academies Press, http://dx.doi.org/10.17226/24987
1-81
-------
Natrai. V; Liu. X; Kulawik. S; Chance. K; Chatfield. R; Edwards. DP; Eldcring. A; Francis. G;
Kurosu. T; Pickering. K; Spurr. R; Worden. H. (2011). Multi-spectral sensitivity studies for the
retrieval of tropospheric and lowermost tropospheric ozone from simulated clear-sky GEO-CAPE
measurements. Atmos Environ 45: 7151-7165. http://dx.doi.Org/10.1016/i.atmosenv.2011.09.014
Neemann. EM; Crosman. ET; Horel. JD; Avev. L. (2015). Simulations of a cold-air pool associated
with elevated wintertime ozone in the Uintah Basin, Utah. Atmos Chem Phys 15: 135-151.
http://dx.doi.org/10.5194/acp-15-135-2015
Nolte. CG; Dolwick. PD; Fann. N; Horowitz. LW; Naik. V; Pinder. RW; Spero. TL; Winner. DA;
Ziska. LH. (2018). Air quality. In DR Reidmiller; CW Avery; DR Easterling; KE Kunkel; KLM
Lewis; TK Maycock; BC Stewart (Eds.), Impacts, Risks, and Adaptation in the United States:
Fourth National Climate Assessment, Volume II (pp. 512538). Washington, DC: U.S. Global
Change Research Program, https://nca2018.globalchange.gov/chapter/13/
Nopmongcol. U; Alvarez. Y; Jung. J; Grant. J; Kumar. N; Yarwood. G. (2017). Source contributions
to United States ozone and particulate matter over five decades from 1970 to 2020. Atmos Environ
167: 116-128. http://dx.doi.Org/10.1016/i.atmosenv.2017.08.009
Oetien. H; Payne. VH; Neu. JL; Kulawik. SS; Edwards. DP; Eldering. A; Worden. HM; Worden. JR.
(2016). A joint data record of tropospheric ozone from Aura-TES and MetOp-IASI. Atmos Chem
Phys 16: 10229-10239. http://dx.doi.org/10.5194/acp-16-10229-2016
Oltmans. S: Schnell. R: Johnson. B; Petron. G; Mefford. T; Neelvii. R. (2014). Anatomy of wintertime
ozone associated with oil and natural gas extraction activity in Wyoming and Utah. Elementa:
Science of the Anthropocene 2: 1-15. http://dx.doi.org/10.12952/iournal.elementa.000024
Orbe. C; Waugh. DW; Yang. H; Lamarque. JF; Tilmes. S; Kinnison. DE. (2017). Tropospheric
transport differences between models using the same large-scale meteorological fields. Geophys
Res Lett 44: 1068-1078. http://dx.doi.org/10.1002/2016GLQ71339
Pan. LL; Homever. CR; Honomichl. S; Ridley. BA; Weisman. M; Barth. MC; Hair. JW; Fenn. MA;
Butler. C: Diskin. GS; Crawford. JH; Rverson. TB; Pollack. I; Peischl. J; Huntrieser. H. (2014).
Thunderstorms enhance tropospheric ozone by wrapping and shedding stratospheric air. Geophys
Res Lett 41: 7785-7790. http://dx.doi.org/10.1002/2014GLQ61921
Parrish. DP; Ennis. CA. (2019). Estimating background contributions and US anthropogenic
enhancements to maximum ozone concentrations in the northern US. Atmos Chem Phys 19:
12587-12605. http://dx.doi.org/10.5194/acp-19-12587-2019
Parrish. DP; Lamarque. JF; Naik. V; Horowitz. L; Shindell. DT; Staehelin. J; Derwent. R; Cooper.
OR: Tanimoto. H: Volz-Thomas. A; Gilge. S; Scheel. HE; Steinbacher. M: Froehlich. M. (2014).
Long-term changes in lower tropospheric baseline ozone concentrations: Comparing chemistry-
climate models and observations at northern midlatitudes. J Geophys Res Atmos 119: 5719-5736.
http://dx.doi.org/10.1002/2013JPQ21435
Parrish. PP; Law. KS; Staehelin. J; Perwent. R; Cooper. OR; Tanimoto. H; Volz-Thomas. A; Gilge.
S; Scheel. HE; Steinbacher. M; Chan. E. (2012). Long-term changes in lower tropospheric baseline
ozone concentrations at northern mid-latitudes. Atmos Chem Phys 12: 11485-11504.
http://dx.doi.org/10.5194/acp-12-11485-2012
Parrish. PP; Petropavlovskikh. I; Oltmans. SJ. (2017a). Reversal of long-term trend in baseline ozone
concentrations at the North American West Coast. Geophys Res Lett 44: 10675-10681.
http://dx.doi.org/10.1002/2017GL07496Q
1-82
-------
Parrish. DP; Young. LM; Newman. MH; Aikin. KC; Rverson. TB. (2017b). Ozone design values in
southern California's air basins: Temporal evolution and US background contribution. J Geophys
ResAtmos 122: 11166-11182. http://dx.doi.org/10.1002/2016JDQ26329
Pfister. GG; Walters. S; Emmons. LK; Edwards. DP; Avise. J. (2013). Quantifying the contribution of
inflow on surface ozone over California during summer 2008. J Geophys Res Atmos 118: 12282-
12299. http://dx.doi.org/10.1002/2013JD02Q336
Phoenix. DB; Homever. CR: Barth. MC. (2017). Sensitivity of simulated convection-driven
stratosphere-troposphere exchange in WRF-Chem to the choice of physical and chemical
parameterization. 4: 454-471. http://dx.doi.org/10.1002/2017EA00Q287
Ran. L; Pleim. J; Gilliam. R: Binkowski. FS: Hogrefe. C: Band. L. (2016). Improved meteorology
from an updated WRF/CMAQ modeling system with MODIS vegetation and albedo. J Geophys
Res Atmos 121: 2393-2415. http://dx.doi.org/10.1002/2015JD0244Q6
Rappengliick. B; Ackermann. L; Alvarez. S: Golovko. J; Buhr. M; Field. RA; Soltis. J: Montague-
DC; Hauze. B; Adamson. S; Risch. D; Wilkerson. G; Bush. D; Stoeckenius. T; Keslar. C. (2014).
Strong wintertime ozone events in the Upper Green River basin, Wyoming. Atmos Chem Phys 14:
4909-4934. http://dx.doi.org/10.5194/acp-14-4909-2014
Rasool. OZ; Zhang. R; Lash. B; Cohan. DS; Cooter. EJ; Bash. JO; Lamsal. LN. (2016). Enhanced
representation of soil NO emissions in the Community Multiscale Air Quality (CMAQ) model
version 5.0.2. GMD 9: 3177-3197. http://dx.doi.org/10.5194/gmd-9-3177-2016
Rastigeiev. Y; Park. R; Brenner. MP; Jacob. DJ. (2010). Resolving intercontinental pollution plumes
in global models of atmospheric transport. J Geophys Res 115: D02302.
http://dx.doi.org/10.1029/2009JDQ12568
Reddv. PJ; Pfister. GG. (2016). Meteorological factors contributing to the interannual variability of
midsummer surface ozone in Colorado, Utah, and other western US states. J Geophys Res Atmos
121: 2434-2456. http://dx.doi.org/10.1002/2015JD023840
Reidmiller. PR; Fiore. AM; Jaffe. DA; Bergmann. D: Cuvelier. C; Dentener. FJ; Duncan; Bryan. N;
Folberth. G; Gauss. M; Gong. S; Hess. P; Jonson. JE; Keating. T; Lupu. A; Marmer. E; Park. R;
Schultz. MG; Shindell. DT; Szopa. S; Vivanco. MG; Wild. O; Zuber. A. (2009). The influence of
foreign vs. North American emissions on surface ozone in the US. Atmos Chem Phys 9: 5027-
5042.
Rieder. HE; Fiore. AM; Clifton. OE; Correa. G; Horowitz. LW; Naik. V. (2018). Combining model
projections with site-level observations to estimate changes in distributions and seasonality of
ozone in surface air over the USA. Atmos Environ 193: 302-315.
http://dx.doi.Org/10.1016/i.atmosenv.2018.07.042
Ross. AN; Wooster. MJ; Boesch. H; Parker. R. (2013). First satellite measurements of carbon dioxide
and methane emission ratios in wildfire plumes. Geophys Res Lett 40: 4098-4102.
http://dx.doi.org/10.1002/grl.50733
Russell. AR; Valin. LC; Bucsela. EJ; Wenig. MO; Cohen. RC. (2010). Space-based constraints on
spatial and temporal patterns of NO(x) emissions in California, 2005-2008. Environ Sci Technol
44: 3608-3615. http://dx.doi.org/10.1021/es903451i
Rydsaa. JH; Stordal. F; Gerosa. G; Finco. A; Hodnebrog. O. (2016). Evaluating stomatal ozone fluxes
in WRF-Chern: Comparing ozone uptake in Mediterranean ecosystems. Atmos Environ 143: 237-
248. http://dx.doi.Org/10.1016/i.atmosenv.2016.08.057
1-83
-------
Rvu. YH; Hodzic. A; Barre. J; Descombes. G; Minnis. P. (2018). Quantifying errors in surface ozone
predictions associated with clouds over the CONUS: a WRF-Chem modeling study using satellite
cloud retrievals. Atmos Chem Phys 18: 7509-7525. http://dx.doi.org/10.5194/acp-18-7509-2Q18
Sandu. A; Daescu. DN; Carmichael. GR; Chai. TF. (2005). Adjoint sensitivity analysis of regional air
quality models. Journal of Computational Physics 204: 222-252.
http://dx.doi.Org/10.1016/i.icp.2004.10.011
Sarwar. G; Gantt. B: Schwede. D; Foley. K: Mathur. R: Saiz-Lopez. A. (2015). Impact of Enhanced
Ozone Deposition and Halogen Chemistry on Tropospheric Ozone over the Northern Hemisphere.
Environ Sci Technol 49: 9203-9211. http://dx.doi.org/10.1021/acs.est.5b01657
Sarwar. G; Simon. H: Bhave. P; Yarwood. G. (2012). Examining the impact of heterogeneous nitryl
chloride production on air quality across the United States. Atmos Chem Phys 12: 6455-6473.
http://dx.doi.org/10.5194/acp-12-6455-2012
Saunois. M: Bousauet. P; Poulter. B. en: Peregon. A: Ciais. P; Canadell. JG; Dlugokenckv. EJ: Etiope.
G: Bastviken. D; Houweling. S: Janssens-Maenhout. G: Tubiello. FN: Castaldi. S: Jackson. RB:
Alexe. M; Arora. VK; Beerling. DJ; Bergamaschi. P; Blake. PR: Brailsford. G: Brovkin. V:
Bruhwiler. L: Crevoisier. C: Crill. P: Covev. K: Curry. C: Frankenberg. C: Gednev. N: Hoeglund-
Isaksson. L: Ishizawa. M: Ito. A: Joos. F: Kim. HS: Kleinen. T: Krummel. P: Lamarque. JF:
Langenfelds. R. av: Locatelli. R: Machida. T; Maksvutov. S: Mcdonald. KC: Marshall. J: Melton.
JR: Morino. I: Naik. V: O'Dohertv. S: Parmentier. FJW: Patra. PK: Peng. C: Peng. S: Peters. GP:
Pison. I: Prigent C: Prinn. R: Ramonet. M: Rilev. WJ: Saito. M: Sa.nt.ini. M: Schroeder. R:
Simpson. IJ; Spahni. R: Steele. P; Takizawa. A: Thornton. BF; Tian. H: Tohjima. Y; Viovv. N:
Voulgarakis. A: van Weele. M; van Per Werf. GR: Weiss. R; Wiedinmver. C: Wilton. DJ:
Wiltshire. A: Worthy. D: Wunch. D: Xu. X: Yoshida. Y: Zhang. B: Zhu. O. (2016). The global
methane budget 2000-2012. Earth System Science Data 8: 697-751. http://dx.doi.org/10.5194/essd-
8-697-2016
Schnell. RC: Oltmans. SJ: Neelv. RR: Endres. MS: Molenar. JV: White. AB. (2009). Rapid
photochemical production of ozone at high concentrations in a rural site during winter. Nat Geosci
2: 120-122. http://dx.doi.org/10.1038/NGEQ415
Schreier. SF; Richter. A: Kaiser. JW: Burrows. JP. (2014). The empirical relationship between
satellite-derived tropospheric N02 and fire radiative power and possible implications for fire
emission rates of NOx. Atmos Chem Phys 14: 2447-2466. http://dx.doi.org/10.5194/acp-14-2447-
2014
Schreier. SF: Richter. A: Schepaschenko. D: Shvidenko. A: Hilboll. A: Burrows. JP. (2015).
Differences in satellite-derived NOx emission factors between Eurasian and North American boreal
forest fires. Atmos Environ 121: 55-65. http://dx.doi.Org/10.1016/i.atmosenv.2014.08.071
Schroeder. JR: Crawford. JH: Fried. A: Walega. J: Weinheimer. A: Wisthaler. A: Muller. M:
Mikovinv. T: Chen. G: Shook. M: Blake. PR: Tonnesen. GS (2017). New insights into the column
CH20/N02 ratio as an indicator of near-surface ozone sensitivity. J Geophys Res Atmos 122:
8885-8907. http://dx.doi.org/10.1002/2017JDQ26781
Seinfeld. JH: Pandis. SN. (2006). Atmospheric chemistry and physics: From air pollution to climate
change. In JH Seinfeld; SN Pandis (Eds.), Atmospheric chemistry and physics: From air pollution to
climate change (3rd ed.). Hoboken, NJ: John Wiley & Sons, https://www.wilev.com/en-us/
Atmospheric+Chemistrv+and+Phvsics%3A+From+Air+Pollution+to+Climate+Change%2C+3rd
+Edition-p-9781118947401
1-84
-------
Shen. L; Micklev. LJ; Tai. APK. (2015). Influence of synoptic patterns on surface ozone variability
over the eastern United States from 1980 to 2012. Atmos Chem Phys 15: 10925-10938.
http://dx.doi.org/10.5194/acp-15-10925-2Q15
Shen. L. u; Micklev. LJ; Leibensperger. EM; Li. M. (2017). Strong dependence of US summertime air
quality on the decadal variability of Atlantic Sea surface temperatures. Geophys Res Lett 44:
12527-12535. http://dx.doi.org/10.1002/2017GL0759Q5
Silvern. RF; Jacob. DJ; Micklev. LJ; Sulprizio. MP; Travis. KR; Marais. EA; Cohen. RC; Laughner.
JL; Choi. S; Joiner. J; Lamsal. LN. (2019). Using satellite observations of tropospheric N02
columns to infer long-term trends in US NOx emissions: the importance of accounting for the free
tropospheric N02 background. Atmos Chem Phys 19: 8863-8878. http://dx.doi.Org/10.5194/acp-
19-8863-2019
Simon. H; Baker. KR; Phillips. S. (2012). Compilation and interpretation of photochemical model
performance statistics published between 2006 and 2012. Atmos Environ 61: 124-139.
http://dx.doi.Org/10.1016/i.atmosenv.2012.07.012
Simon. H; Kimura. Y; McGaughev. G; Allen. DT; Brown. SS; Osthoff. HP; Roberts. JM; Bvun. D;
Lee. D. (2009). Modeling the impact of C1N02 on ozone formation in the Houston area. J Geophys
Res Atmos 114: D00F03. http://dx.doi.org/10.1029/2008id010732
Simon. H; Reff. A; Wells. B; Xing. J; Frank. N. (2015). Ozone trends across the United States over a
period of decreasing NOx and VOC emissions. Environ Sci Technol 49: 186-195.
http://dx.doi.org/10.1021/es504514z
Simon. H; Valin. LC; Baker. KR; Henderson. BH; Crawford. JH; Pusede. SE; Kelly. JT; Foley. KM;
Chris Owen. R; Cohen. RC; Timin. B; Weinheimer. AJ; Possiel. N; Misenis. C; Diskin. GS; Fried.
A. (2018). Characterizing CO and NOy sources and relative ambient ratios in the Baltimore area
using ambient measurements and source attribution modeling. J Geophys Res Atmos 123: 3304-
3320. http://dx.doi.org/10.1002/2017JDQ27688
Skerlak. B; Sprenger. M; Pfahl. S; Tvrlis. E; Wernli. H (2015). Tropopause folds in ERA-Interim:
Global climatology and relation to extreme weather events. J Geophys Res Atmos 120: 4860-4877.
http://dx.doi.org/10.1002/2014JDQ22787
Stohl. A; Spichtinger-Rakowskv. N; Bonasoni. P; Feldmann. H; Memmesheimer. M; Scheel. HE;
Tricklv. T; Hubener. S; Ringer. W; Mandl. M. (2000). The influence of stratospheric intrusions on
alpine ozone concentrations. Atmos Environ 34: 1323-1354. http://dx.doi.org/10.lQ16/S1352-
2310(99)00320-9
Stone. D; Sherwen. T; Evans. MJ; Vaughan. S; Ingham. T; Whallev. LK; Edwards. PM; Read. KA;
Lee. JD; Moller. SJ; Carpenter. LJ; Lewis. AC; Heard. DE. (2018). Impacts of bromine and iodine
chemistry on tropospheric OH and H02: comparing observations with box and global model
perspectives. Atmos Chem Phys 18: 3541-3561. http://dx.doi.org/10.5194/acp-18-3541-2018
Strode. SA; Rodriguez. JM; Logan. JA; Cooper. OR; Witte. JC; Lamsal. LN; Damon. M; Van Aartsen.
B; Steenrod. SD; Strahan. SE. (2015). Trends and variability in surface ozone over the United
States. J Geophys Res Atmos 120: 9020-9042. http://dx.doi.org/10.1002/2014JDQ22784
Sullivan. JT; Mcgee. TJ; Devoung. R; Twigg. LW; Sumnicht. GK; Pliutau. D; Knepp. T; Carrion. W.
(2015a). Results from the NASA GSFC and LaRC Ozone Lidar Intercomparison: New Mobile
Tools for Atmospheric Research. J Atmos Ocean Tech 32: 1779-1795.
http://dx.doi.Org/10.1175/JTECH-D-14-00193.l
1-85
-------
Sullivan. JT; Mcgee. TJ; Leblanc. T; Sumnicht. GK; Twigg. LW. (2015b). Optimization of the GSFC
TROPOZ DIAL retrieval using synthetic lidar returns and ozonesondes - Part 1: Algorithm
validation. Atmos Meas Tech 8: 4133-4143. http://dx.doi.org/10.5194/amt-8-4133-2015
Sullivan. JT; Mcgee. TJ; Sumnicht. GK; Twigg. LW; Hoff. RM. (2014). A mobile differential
absorption lidar to measure sub-hourly fluctuation of tropospheric ozone profiles in the Baltimore-
Washington, DC region. Atmos Meas Tech 7: 3529-3548. http://dx.doi.org/10.5194/amt-7-3529-
2014
Tang. Q; Prather. MJ; Hsu. J. (2011). Stratosphere-troposphere exchange ozone flux related to deep
convection. Geophys Res Lett 38: L03806. http://dx.doi.org/10.1029/2010GL046039
Tanimoto. H; Ikeda. K; Boersma. KF; Ronald. J. vanderA; Garivait. S. (2015). Interannual variability
of nitrogen oxides emissions from boreal fires in Siberia and Alaska during 1996-2011 as observed
from space. Environ Res Lett 10. http://dx.doi.Org/10.1088/1748-9326/10/6/065004
Tereszchuk. KA; Abad. GG; Clerbaux. C; Hurtmans. D; Coheur. PF; Bernath. PF. (2011). ACE-FTS
measurements of trace species in the characterization of biomass burning plumes. Atmos Chem
Phys 11: 12169-12179. http://dx.doi.org/10.5194/acp-ll-12169-2011
Travis. KR; Jacob. DJ; Fisher. JA; Kim. PS; Marais. EA; Zhu. L; Yu. K; Miller. CC; Yantosca. RM;
Sulprizio. MP; Thompson. AM; Wennberg. PO; Crounse. JD; St Clair. JM; Cohen. RC; Laughner.
JL; Dibb. JE; Hall. SR; Ullmann. K; Wolfe. GM; Pollack. IB; Peischl. J; Neuman. JA; Zhou. X.
(2016). Why do models overestimate surface ozone in the Southeast United States? Atmos Chem
Phys 16: 13561-13577. http://dx.doi.org/10.5194/acp-16-13561-2016
Tuite. K; Brockwav. N; Colosimo. SF; Grossmann. K; Tsai. C; Flynn. J; Alvarez. S; Erickson. M;
Yarwood. G; Nopmongcol. U; Stutz. J. (2018). Iodine catalyzed ozone destruction at the Texas
coast and Gulf of Mexico. Geophys Res Lett 45: 7800-7807.
http://dx.doi.org/10.1029/2018GL078267
U.S. EPA (U.S. Environmental Protection Agency). (1996). Air quality criteria for ozone and related
photochemical oxidants [EPA Report]. (EPA/600/P-93/004AF). Research Triangle Park, NC.
U.S. EPA (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: U.S.
Environmental Protection Agency, Office of Research and Development, National Center for
Environmental Assessment-RTP Office.
http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=149923
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-07/007). Research Triangle Park, NC.
https://www3.epa.gov/ttn/naaqs/standards/ozone/data/2007 07 ozone staff paper.pdf
U.S. EPA (U.S. Environmental Protection Agency). (2010). Integrated science assessment for carbon
monoxide (pp. 1-593). (EPA/600/R-09/019F). Research Triangle Park, NC: U.S. Environmental
Protection Agency, Office of Research and Development, National Center for Environmental
Assessment. http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=218686
U.S. EPA (U.S. Environmental Protection Agency). (2011). MOVES (Motor Vehicle Emission
Simulator). Available online at http://www.epa.gov/otaq/models/moves/index.htm (accessed
January 28, 2011).
1-86
-------
U.S. EPA (U.S. Environmental Protection Agency). (2013). Integrated science assessment for ozone
and related photochemical oxidants [EPA Report]. (EPA/600/R-10/076F). Research Triangle Park,
NC: U.S. Environmental Protection Agency, Office of Research and Development, National Center
for Environmental Assessment-RTP Division.
http://cfpub.cpa.gov/ncca/isa/rccordisplav.cfm?dcid=247492
U.S. EPA (U.S. Environmental Protection Agency). (2015). Implementation of the 2015 Primary
Ozone NAAQS: Issues associated with background ozone. White paper for discussion. Research
Triangle Park, NC: U.S. Environmental Protection Agency :: U.S. EPA.
U.S. EPA (U.S. Environmental Protection Agency). (2016a). 2014 National Emissions Inventory
(NEI) data (Version 2). Washington, DC. Retrieved from https://www.epa.gov/air-emissions-
inventories/2014-national-emissions-inventorv-nei-data
U.S. EPA (U.S. Environmental Protection Agency). (2016b). 2014 National Emissions Inventory,
version 1 technical support document (Draft). Research Triangle Park, NC.
U.S. EPA (U.S. Environmental Protection Agency). (2016c). Integrated science assessment for oxides
of nitrogen-health criteria (final report) [EPA Report]. (EPA/600/R-15/068). Research Triangle
Park, NC: U.S. Environmental Protection Agency, Office of Research and Development, National
Center for Environmental Assessment.
http://ofrnpub.epa.gov/eims/eimscomm.getfile7p download id=526855
U.S. EPA (U.S. Environmental Protection Agency). (2016d). Inventory of U.S. greenhouse gas
emissions and sinks: 1990-2016. (EPA 430-R-18-003). Washington, D.C.
https://www.epa.gov/ghgemissions/inventorv-us-greenhouse-gas-emissions-and-sinks
U.S. EPA (U.S. Environmental Protection Agency). (2016e). WebFIRE. Available online at
https://www3.epa.gov/ttn/chief/webfire/index.html
U.S. EPA (U.S. Environmental Protection Agency). (2017). Profile of version 1 of the 2014 National
Emissions Inventory. Washington, DC. https://www.epa.gov/sites/production/files/2Q17-
04/documents/2014neivl profile final aprill82017.pdf
U.S. EPA (U.S. Environmental Protection Agency). (2018). Integrated science assessment for oxides
of nitrogen, oxides of sulfur and particulate matterecological criteria (2nd external review draft).
(EPA/600/R-18/097). Research Triangle Park, NC: U.S. Environmental Protection Agency, Office
of Research and Development, National Center for Environmental Assessment.
https://cfpub.epa.gov/ncea/isa/recordisplay.cfm?deid=340671
U.S. EPA (U.S. Environmental Protection Agency). (2019a). Air markets program data. Available
online at https://ampd.epa.gov/ampd/
U.S. EPA (U.S. Environmental Protection Agency). (2019b). Air pollutant emissions trends data:
Criteria pollutants national tier 1 for 1970-2017. Retrieved from https://www.epa. gov/air-
emissions-inventories/air-pollutant-emissions-trends-data
U.S. EPA (U.S. Environmental Protection Agency). (2019c). CMAQ model evaluation framework.
Available online at https://www.epa.gov/cmaq/cmaq-model-evaluation-framework (accessed
February 4, 2020).
U.S. EPA (U.S. Environmental Protection Agency). (2019d). Overview: NOx budget trading program.
Available online at https://www.epa.gov/airmarkets/nox-budget-trading-program
Van Dingenen. R; Crippa. M; Janssens-Maenhout. G; Guizzardi. D; Dentener. F. (2018). Global trends
of methane emissions and their impacts on ozone concentrations. (JRC113210; EUR 29394 EN;
OP KJ-NA-29394-EN-N). Luxembourg: Office for Official Publications of the European Union.
http://dx.doi.org/10.2760/820175
1-87
-------
Vinken. GCM; Boersma. KF; Maasakkers. JD; Adon. M; Martin. RV. (2014). Worldwide biogenic
soil NOx emissions inferred from OMIN02 observations. Atmos Chem Phys 14: 10363-10381.
http://dx.doi.org/10.5194/acp-14-10363-2014
Wang. L; Newchurch. MJ; Alvarez. RJ: Berkoff. TA; Brown. SS; Carrion. W; De Young. RJ:
Johnson. BJ; Ganoe. R; Gronoff. G; Kirgis. G; Kuang. S; Langford. AO; Leblanc. T; Mcduffie. EE;
Mcgee. TJ; Pliutau. D; Senff. CJ; Sullivan. JT; Sumnicht. G; Twigg. LW; Weinheimer. AJ. (2017).
Quantifying TOLNet ozone lidar accuracy during the 2014 DISCOVER-AQ and FRAPPE
campaigns. Atmos Meas Tech 10: 3865-3876. http://dx.doi.org/10.5194/amt-10-3865-2017
Wang. YH; Jacob. DJ; Logan. JA. (1998). Global simulation of tropospheric O-3-NOx-hydrocarbon
chemistry 1. Model formulation. J Geophys Res Atmos 103: 10713-10725.
http://dx.doi.org/10.1029/98JD00158
Warneke. C; Geiger. F; Edwards. PM; Dube. W; Petron. G; Kofler. J; Zahn. A; Brown. SS; Graus. M;
Gilman. JB; Lerner. BM; Peischl. J; Rverson. TB; de Gouw. JA; Roberts. JM. (2014). Volatile
organic compound emissions from the oil and natural gas industry in the Uintah Basin, Utah: oil
and gas well pad emissions compared to ambient air composition. Atmos Chem Phys 14: 10977-
10988. http://dx.doi.org/10.5194/acp-14-10977-2014
Weber. M; Steinbrecht. W; Arosio. C; van der A. R; Frith. SM; Anderson. J; Coldewev-Egbers. M;
Davis. S; Degenstein. D; Fioletov. YE; Froidevaux. L; Hubert. D; Long. CS; Loyola. D; Rozanov.
A; Roth. C; Sofieva. V; Tourpali. K; Wang. R; Wild. JD. (2018). State of the climate in 2018:
Stratospheric ozone. Bull Am Meteorol Soc 100: S54-S56.
Worden. HM; Deeter. MN; Frankenberg. C; George. M; Nichitiu. F; Worden. J; Aben. I; Bowman.
KW; Clerbaux. C; Coheur. PF; de Laat. ATJ; Detweiler. R; Drummond. JR. JR; Edwards. DP;
Gille. JC; Hurtmans. D; Luo. M; Martinez-Alonso. S; Massie. S; Pfister. G; Warner. JX. (2013).
Decadal record of satellite carbon monoxide observations. Atmos Chem Phys 13: 837-850.
http://dx.doi.org/10.5194/acp-13-837-2013
Wu. SL; Duncan. BN; Jacob. DJ; Fiore. AM; Wild. O. (2009). Chemical nonlinearities in relating
intercontinental ozone pollution to anthropogenic emissions. Geophys Res Lett 36: L05806.
http://dx.doi.org/10.1029/2008glQ36607
Xie. Y; Paulot. F; Carter. WPL; Nolte. CG; Luecken. DJ; Hutzell. WT; Wennberg. PO; Cohen. RC;
Pinder. RW. (2013). Understanding the impact of recent advances in isoprene photooxidation on
simulations of regional air quality. Atmos Chem Phys 13: 8439-8455.
http://dx.doi.org/10.5194/acp-13-8439-2013
Xing. J. ia; Wang. J; Mathur. R; Wang. S; Sarwar. G; Pleim. J; Hogrefe. C; Zhang. Y; Jiang. J; Wong.
DC; Hao. J. (2017). Impacts of aerosol direct effects on tropospheric ozone through changes in
atmospheric dynamics and photolysis rates. Atmos Chem Phys 17: 9869-9883.
http://dx.doi.org/10.5194/acp-17-9869-2017
Xing. J: Mathur. R; Pleim. J; Hogrefe. C; Gan. CM; Wong. DC; Wei. C; Gilliam. R; Pouliot. G.
(2015). Observations and modeling of air quality trends over 1990-2010 across the Northern
Hemisphere: China, the United States and Europe. Atmos Chem Phys 15: 2723-2747.
http://dx.doi.org/10.5194/acp-15-2723-2015
Yahva. K; Wang. K; Zhang. Y; Kleindienst. TE. (2015). Application ofWRF/Chem over North
America under the AQMEII Phase 2-Part 2: Evaluation of 2010 application and responses of air
quality and meteorology-chemistry interactions to changes in emissions and meteorology from
2006 to 2010. GMD 8: 2095-2117. http://dx.doi.org/10.5194/gmd-8-2095-2015
1-88
-------
Ying. Q; Krishnan. A. (2010). Source contributions of volatile organic compounds to ozone formation
in southeast Texas. J Geophys Res Atmos 115: Article #D17306.
http://dx.doi.org/10.1029/2010JDQ13931
Young. PJ; Naik. V; Fiore. AM; Gaudel. A; Guo. J; Lin. MY; Neu. JL; Parrish. DP; Rieder. HE;
Schnell. JL; Tilmes. S; Wild. O; Zhang. L; Ziemke. J; Brandt. J; Delcloo. A; Dohertv. RM; Geels.
C; Hegglin. MI; Hu. L; Im. U; Kumar. R; Luhar. A; Murray. L; Plummer. D; Rodriguez. J; Saiz-
Lopez. A; Schultz. MG; Woodhouse. MT; Zeng. G. (2018). Tropospheric Ozone Assessment
Report: Assessment of global-scale model performance for global and regional ozone distributions,
variability, and trends. Elementa: Science of the Anthropocene 6.
http://dx.doi.org/10.1525/elementa.265
Zare. A; Christensen. JH; Gross. A; Iranneiad. P; Glasius. M; Brandt. J. (2014). Quantifying the
contributions of natural emissions to ozone and total fine PM concentrations in the Northern
Hemisphere. Atmos Chem Phys 14: 2735-2756. http://dx.doi.org/10.5194/acp-14-2735-2014
Zhang. L; Jacob. DJ; Downey. NY; 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. Atmos Environ 45: 6769-6776. http://dx.doi.Org/10.1016/i.atmosenv.2011.07.054
Zhang. L; Jacob. DJ; Kopacz. M; Henze. DK; Singh. K; Jaffe. DA. (2009). Intercontinental source
attribution of ozone pollution at western US sites using an adjoint method. Geophys Res Lett 36:
LI 1810. http://dx.doi.org/10.1029/2009glQ37950
Zhang. L; Jacob. DJ; Yue. X; Downey. NY; Wood. DA; Blewitt. D. (2014). Sources contributing to
background surface ozone in the US Intermountain West. Atmos Chem Phys 14: 5295-5309.
http://dx.doi.org/10.5194/acp-14-5295-2014
Zhang. L; Jaffe. DA. (2017). Trends and sources of ozone and sub-micron aerosols at the Mt. Bachelor
Observatory (MBO) during 2004-2015. Atmos Environ 165: 143-154.
http://dx.doi.Org/10.1016/i.atmosenv.2017.06.042
Zhang. R; Cohan. A; Biazar. AP; Cohan. DS. (2017). Source apportionment of biogenic contributions
to ozone formation over the United States. Atmos Environ 164: 8-19.
http://dx.doi.Org/10.1016/i.atmosenv.2017.05.044
Zhang. Y; Cooper. OR; Gaudel. A; Thompson. AM; Nedelec. P; Qgino. SY; West. JJ. (2016).
Tropospheric ozone change from 1980 to 2010 dominated by equatorward redistribution of
emissions. Nat Geosci 9: 875-+. http://dx.doi.org/10.1038/NGEQ2827
Zheng. B; Tong. D; Li. M; Liu. F; Hong. C; Geng. G; Li. H; Li. X; Peng. L; Oi. J; Yan. L; Zhang. Y;
Zhao. H; Zheng. Y; He. K; Zhang. O. (2018). Trends in China's anthropogenic emissions since
2010 as the consequence of clean air actions. Atmos Chem Phys 18: 14095-14111.
http://dx.doi.org/10.5194/acp-18-14095-2018
Zhu. L; Jacob. DJ; Keutsch. FN; Micklev. LJ; Scheffe. R; Strum. M; Gonzalez Abad. G; Chance. K;
Yang. K; Rappengliick. B; Millet. DB; Baasandorj. M; Jaegle. L; Shah. V. (2017). Formaldehyde
(HCHO) as a hazardous air pollutant: Mapping surface air concentrations from satellite and
inferring cancer risks in the United States. Environ Sci Technol 51: 5650-5657.
http://dx.doi.org/10.1021/acs.est.7b01356
1-89
-------
Zhu. L; Jacob. DJ; Kim. PS; Fisher. JA; Yu. K; Travis. KR; Micklev. LJ; Yantosca. RM: Sulprizio.
MP; De Smedt. I; Abad. GG; Chance. K; Li. C. an; Ferrare. R; Fried. A; Hair. JW; Hanisco. TF;
Richter. D; Scaring. A; Walesa. J; Weibring. P; Wolfe. GM. (2016). Observing atmospheric
formaldehyde (HCHO) from space: validation and intercomparison of six retrievals from four
satellites (OMI, GOME2A, GOME2B, OMPS) with SEAC(4)RS aircraft observations over the
southeast US. Atmos Chem Phys 16: 13477-13490. http://dx.doi.org/10.5194/acp-16-13477-2Q16
Zoogman. P; Liu. X; Suleiman. RM: Pennington. WF; Flittner. DE; Al-Saadi. JA; Hilton. BB; Nicks.
DK; Newchurch. MJ; Carr. JL; Janz. SJ; Andraschko. MR; Arola. A; Baker. BP; Canova. BP;
Miller. CC; Cohen. RC; Davis. JE; Dussault. ME; Edwards. DP; Fishman. J; Ghulam. A; Abad.
GG; Grutter. M; Herman. JR. JR; Houck. J; Jacob. DJ; Joiner. J; Kerridge. BJ; Kim. J; Krotkov.
NA; Lamsal. L; Li. C; Lindfors. A; Martin. RV; Mcelroy. CT; Mclinden. C; Natrai. V; Neil. DO;
Nowlan. CR; O'Sullivan. EJ; Palmer. PI; Pierce. RB; Pippin. MR; Saiz-Lopez. A; Spurr. RJD;
Szvkman. JJ; Torres. O; Veefkind. JP; Veihelmann. B; Wang. H; Wang. J; Chance. K. (2017).
Tropospheric emissions: Monitoring of pollution (TEMPO). J Quant Spectrosc Radiat Transf 186:
17-39. http://dx.doi.Org/10.1016/i.iasrt.2016.05.008
1-90
-------
APPENDIX 2 EXPOSURE TO AMBIENT OZONE
Overall ( 'onclnsions regarding ilstimates of Exposure to. I mhient Ozone for I se in
Epidemiologic Studies
• Since llie 2(i I i ( )a»iic IS A. ;id\ ;iiices h;i\ e heeu ni;ide in se\ ercil ;ippro;iclies lor
predicliim ;inihieui oa»iic coiicciiiriiiious ;is surrou;iies for exposure I !itoi\ ;issoci;iled
w nil exposure iissessnieui methods ;iiv olicu siiiiihir in or urh;iu sciiles hec;iuse
;inihieul oa»iic couceuinilioiis lend lo h;i\ e low sp;iii;il \ ;in;ihilil\
• I'or epidemiologic studies ol" shorl-lcrni exposure io iinihieul oa»iic. the ;issoci;iIioii
between exposure esiinuiles ;md he;illh effects m;i\ he uuderesiini;iled h\ llie
nic;isurcnieui or model used lo represenl exposure. ;md llie el'leel esiiin;iie m;i\ h;i\e
reduced preeiskiu I !\ eu w lieu l lie ni;miiiiude of llie ;issoei;iliou is iiiiccrl;uu. llie I rue
el'leel would likel> he l;iruer llinii ilie estimated ;issoei;iliou mi iliese e;ises I lie hi;is
;iud reduction mi precision ;ire l\ pic;il l\ s i u;i 11 mi iu;iuiiilude.
• for epidemiologic siudies of lonu-ierni exposure lo ;inihieui oa»iic. dcpcudiim on llie
model ;md sceu;irio heinu modeled, llie ;issoci:iliou between exposure esiinuiles ;iud
lie;illli effects ni;i\ he uuderesiim;iied oro\eresiuu;iied ll is much more common I'or
llie ;issoci;iiioii lo he uuderesiinuiied hecnuse ue;ir-ro;id o/oue sc;i\ eiiuiuu c;iu result m
ure;iler sp;iii;il \ ;iri;ihilii\ due lo ;i reducliou in o/oue coiiceulr;iiiou compared w illi
;inihieui oapic;iIIx siikiII iii ni;miiilude
• I !siiiikiiuiu exposure w iiIkmii iiccoiiiiium lor tinie-;icli\ il\ d;il;i in;i\ resull in
uiideresiim;iiioii ol' llie ;issoci;ilion ;iud reduced precision Mlliouuli llie m;iuiiiiude ol'
llie ;issoci;iiioii helweeu exposure esiini;iles ;md lie;illli ellecls is iiuceri;iiu. llie Hue
el'leel lends io he hiruer llinii llie esiini;iled ;issoci;ilious iu iliese c:ises
2.1 Introduction
This Appendix presents new developments in methodology for developing exposure estimates for
epidemiologic studies and interpreting the results, given the strengths and limitations of the exposure
assessment data. The Appendix describes concepts and terminology relating to exposure (Section 2.2),
methodologies used for exposure assessment (Section 2.3). factors influencing personal exposure to ozone
(Section 2.4). copollutant correlations and potential for confounding (Section 2.5). and interpreting
exposure measurement error for use in epidemiologic studies (Section 2.6). This Appendix focuses on the
ambient air component of personal exposure to ozone, because the National Ambient Air Quality
Standards (NAAQS) pertains to ambient ozone. Because there are very few indoor sources of ozone,
individuals are typically exposed to ozone from ambient air rather than ozone generated from indoor
sources. This Appendix focuses on studies of exposure among the general population. The information
provided in this Appendix will be used to help interpret the evidence for the health effects of ozone
exposure presented in the health Appendices that follow (Appendix 3-Appendix 7).
2-1
-------
2.2 Exposure Concepts
A conceptual model of personal exposure to ambient ozone is described in the 2013 Ozone ISA
(U.S. EPA. 2013). Ozone in ambient air is generally produced by photochemical oxidation of NO2, little
or no precursor gases occurring in ambient air are transformed to ozone indoors. Indoor ozone therefore
either infiltrates from outdoors or is generated by indoor sources, such as some types of air purifiers. Part
of ozone infiltrating indoors is lost to surface reactions. A variety of metrics and terms are used to
characterize air pollution exposure. They are described here to provide clarity for the subsequent
discussion.
The concentration of ozone is defined as the volume of the pollutant in a given volume of air
(e.g., ppb). Concentrations observed in outdoor locations accessible to the public are referred to as
ambient concentrations. The term exposure refers to contact at the interface of the breathing zone with the
concentration of a specific pollutant over a certain period of time (Zartarian et al.. 2005). in single or
multiple locations. For example, contact with a concentration of 10 ppb ozone for 1-hour would be
referred to as a 1-hour exposure to 10 ppb ozone, and 10 ppb is referred to as the exposure concentration.
As discussed in Appendix 3. dose incorporates the concept of intake into the body (via inhalation).
A location where exposure occurs is referred to as a microenvironment, and an individual's daily
exposure consists of the time-integrated concentrations in each of the microenvironments visited during
the day. Ambient air pollution may penetrate indoors, where it combines with air pollution from indoor
sources (indoor air pollution) to produce the total measured indoor concentration. Personal exposure to
ambient ozone is exposure to the ambient fraction of total indoor ozone concentration, together with
exposure to ambient ozone concentrations in outdoor microenvironments such as parks, yards, sidewalks,
and roads (e.g., while riding on bicycles or motorcycles), is referred to as ambient exposure (Wilson.
2000). This differs from overall total personal exposure, which may also include exposure to indoor air
pollution. Personal exposure to ambient ozone is influenced by several factors, including:
• time-activity in different microenvironments (e.g., vehicle, residence, workplace, outdoor);
• climate (e.g., weather, season);
• characteristics of indoor microenvironments (e.g., window openings, draftiness, air conditioning);
• ambient concentrations of NOx from incomplete combustion (e.g., mobile sources, construction
equipment) that are photolyzed to form ozone; and
• scavenging of ozone immediately near roads when NO reacts with ozone to produce NO2.
Exposure assessment studies are evaluated from the reference point of personal exposure to
ambient ozone, with epidemiologic studies of health effects from ambient ozone exposure generally
employing concentration as a surrogate for ozone exposure. Because personal exposures are not routinely
measured, the term exposure surrogate is used in this Appendix to describe a quantity meant to estimate
or represent exposure to ambient ozone, such as ozone concentration measured at an ambient monitor
(Foley et al.. 2003; Sarnat et al.. 2000). A fixed-site monitor (i.e., a monitor with a fixed position) is a
2-2
-------
type of ambient air monitor used to estimate population average ambient concentrations and their trends
over neighborhood and urban scales for epidemiologic studies.
When surrogates are used to estimate exposure in epidemiologic studies, exposure measurement
error or exposure misclassification can result. Exposure measurement error refers to the bias and
uncertainty associated with using concentration metrics to represent the true, but unknown, exposure of an
individual or population (Lipfert and Wyzga. 1996). Exposure misclassification refers to exposure
measurement error that occurs when exposure conditions such as location, timing, or population grouping
are assigned incorrectly or with uncertainty. Exposure misclassification and exposure measurement error
can result in bias and reduced precision of the effect estimate (i.e., the slope of the concentration-response
function in epidemiologic studies). Bias refers to the difference between the effect estimate derived from
a statistical model and the true effect (Armstrong et al.. 1992). Negative bias, or attenuation, of the effect
estimate indicates an underestimate of the magnitude of the effect and tends to occur when the exposure
surrogate and effect are not well correlated in time or space [temporal correlation is important for
short-term studies, while spatial correlation is important for long-term studies; Armstrong et al. (1992)1.
Low spatial correlation with negative bias of the effect estimate in a long-term study indicates an
underestimation of the effect and may occur when the exposure measurement is systematically higher
than the true population exposure. Positive bias of the effect estimate indicates an overestimate of the
magnitude of the effect and may occur when the magnitude of the exposure measurement is
systematically lower than the true population exposure (Armstrong etal.. 1992). Such an overestimate
may occur for an exposed group of people living near a road where O3 scavenging by NOx occurs far
from a fixed-site monitor measuring a higher concentration (Simon et al.. 2016; Cleveland and Graedel.
1979). Exposure measurement error can also lead to an incorrect estimation of standard errors around the
effect estimate.
Exposure measurement error has two components: (1) exposure measurement error derived from
uncertainty in the metric being used to represent exposure and (2) error due to use of a surrogate
parameter of interest in the epidemiologic study in lieu of the true exposure, which may be unobservable.
Classical exposure measurement error is defined as exposure measurement error scattered around the true
personal exposure and independent of the level of the measured exposure. Classical exposure
measurement error may occur when a fixed-site monitor measuring ambient concentration is imprecise,
even if it is accurate, and it is also independent of time and space (Szpiro et al.. 2011). Classical exposure
measurement error can result in bias of the epidemiologic effect estimate. When variation in the exposure
measurements is greater than variation in the true exposures, classical exposure measurement error
typically biases the effect estimate negatively (indicating no or lesser effect of the exposure relative to the
true effect). This would cause the effect to be underestimated. Classical exposure measurement error can
also cause inflation or reduction of the standard error of the effect estimate. Berkson exposure
measurement error is defined as error scattered around the measured exposure surrogate (in most cases,
the measured ambient concentration) and is independent of the true exposure (Goldman etal.. 2011;
Reeves et al.. 1998). Berkson exposure measurement error may occur when the time series of ambient
2-3
-------
ozone concentration measured at a monitor differs from the time series of a person's true exposure such
that the true variability in the person's ozone exposure goes unmeasured. Berkson exposure measurement
error is not expected to bias the effect estimate.
Definitions for classical-like and Berkson-like exposure measurement errors were developed for
exposures estimated using models (Section 2.3.2V These errors can depend on how exposure metrics are
averaged across space and time. Szpiro et al. (2011) defined classical-like and Berkson-like exposure
measurement errors as errors sharing some characteristics with classical and Berkson exposure
measurement errors, respectively, but with some differences. Specifically, classical-like exposure
measurement errors can add variability to predicted exposures and can bias effect estimates in a manner
similar to pure classical exposure measurement errors, but they differ from pure classical errors in that the
variability is around the predicted exposures. Berkson-like exposure measurement errors occur when the
modeled exposure does not capture all sources of variability in the true exposure. Berkson-like exposure
measurement errors increase the variability around the effect estimate in a manner similar to pure Berkson
exposure measurement error, but Berkson-like exposure measurement errors are not independent of
predicted exposures, unlike pure Berkson exposure measurement errors. Berkson-like exposure
measurement error can lead to bias of the effect estimate in either direction (Szpiro and Paciorek. 2013).
The influence of these types of exposure measurement errors on effect estimates for specific
short- and long-term exposure study designs is evaluated in Section 2.6. This review of the influence of
error on exposure estimates used in epidemiologic studies informs the evaluation of confounding and
other biases and uncertainties when considering the health effects evidence in Appendix 3-Appendix 7.
2.3 Exposure Assessment Methods
The 2013 Ozone ISA (U.S. EPA. 2013) reported on fixed-site monitors, passive and active
personal samplers, and microenvironmental models. Since that time, many new modeling methods have
become available to characterize ozone concentrations at neighborhood, urban, and regional scales for use
as exposure surrogates. These methods are described below. Their application in epidemiologic studies of
different averaging times is discussed in Section 2.6.1 for short-term exposure studies and in Section 2.6.2
for long-term exposure studies.
2.3.1 Monitoring
This section builds upon discussions from the 2013 Ozone ISA (U.S. EPA. 2013) about
fixed-site, area, and personal ozone monitoring. Section 2.3.1.1 describes recent studies of fixed-site
monitors, which largely agree with previous studies of fixed-site monitors reported in the 2013 Ozone
ISA. Section 2.3.1.2 presents new studies of microenvironmental and personal ozone monitors, the results
2-4
-------
of which mostly agree with studies presented in the 2013 Ozone ISA. A new development in using
data-processing algorithms to improve the quality of ozone concentration data obtained using low-cost
monitors is highlighted.
2.3.1.1 Fixed-Site Monitors
The 2013 Ozone ISA (U.S. EPA. 2013) described the use of two types of fixed-site Federal
Reference Method (FRM) ozone monitors to provide a surrogate for ozone exposure: ultraviolet (UV)
absorption photometric analyzers and chemiluminescence analyzers. Positive biases were noted for the
UV analyzers when concentrations of volatile organic compounds (VOCs), mercury, and humidity were
high, although the interference due to humidity was found to be small (Ollison et al.. 2013). More than
95% of monitors (including both UV and chemiluminescence analyzers) in the ozone monitoring
network, including the Photochemical Assessment Monitoring Stations (PAMS), met the U.S. EPA data
quality goal of less than 7% bias and precision for 2005 through 2009. Fixed-site monitors were noted to
provide reasonable approximations for area ambient air concentration when spatial variability of ozone
was low.
Recent studies have noted advantages and limitations of using fixed-site monitors to provide an
exposure surrogate (Table 2-8). Strengths include having high quality assurance of the monitors, ease of
assigning exposures to study participants, and availability of multiple years of data to ascertain trends.
Several of the studies described in Table 2-8 cite the lack of data on ozone spatial variability as an
uncertainty. However, Dionisio et al. (2014) noted that spatial variability of ozone tends to be lower than
that of NO2, SO2, or some particulate matter (PM) size fractions, so fixed-site monitors may provide an
acceptable representation of ozone concentration in many cases. Reported errors from using a fixed-site
monitor tended to be within 6 ppb for the 8-hour daily max, ranging from <1 to 7% of reported ozone
concentrations. On a short-term basis, considerable gradients in ozone concentrations can occur within
urban areas, largely resulting from ozone scavenging by NO emitted by transportation during the daytime
(Simon et al.. 2016). Modeling analyses suggest that these gradients are sensitive to changes in NO
concentrations and thus may not be constant overtime (Simon et al.. 2016).
2.3.1.2 Personal and Microenvironmental Monitors
The 2013 Ozone ISA (U.S. EPA. 2013) described methods useful for personal or area monitoring.
Passive badges are typically used for personal exposure monitoring. Passive badges integrate ozone
concentration over a period of 24 hours or longer by measuring the reaction of a nitrite filter coating to
nitrate. Badges have a method detection limit (MDL) of 5-10 ppb. Portable active monitors may be used
for personal exposure or local area ambient air monitoring. Variations include drawing air past a nitrite
coated glass tube for an integrated ozone sample or using UV photometry for continuous ozone
2-5
-------
monitoring. Studies in the 2013 Ozone ISA reported the MDL for the nitrite coated glass tube to be
10 ppb-hour, while studies reported in the 2013 Ozone ISA did not report on MDLs for portable UV
photometers. High MDLs can lead to uncertainties in exposure when ozone concentrations are low.
Because of their averaging times, passive badges can miss when peak exposure occurs. Coupled with
indoor/outdoor ratios (Section 2.4.2). can lead to passive badges being below the MDL (U.S. EPA. 2014).
Other biases and uncertainties in passive badges and portable active monitors were not reported.
Table 2-9 presents recent papers testing continuous or passive personal or microenvironmental
ozone monitors. Continuous microenvironmental ozone monitors have demonstrated low bias and
correlation >0.8 with Federal Equivalent Method (FEM) samplers. In the case of the low-cost continuous
monitors, reduction of bias was dependent on how data from the monitors were analyzed. Low-cost
monitors tend to experience drift when they are deployed in the field. Use of a random forest method, in
which concentrations measured at a specific location are predicted from a combination of temperature,
relative humidity, and concentrations measured at other locations in the network, allowed Zimmerman et
al. (2018) to correct for drift by analyzing trends at each monitoring site over time, thus decoupling
monitor drift from the true ozone concentration signal. Wheeler etal. (2011) measured ozone with a mean
concentration of 26 ppb using passive badges with bias and precision well below 1 ppb. With such low
bias and high precision, uncertainty is low because these concentrations are well above the MDL reported
in the 2013 Ozone ISA (U.S. EPA. 2013).
2.3.2 Modeling
The 2013 Ozone ISA (U.S. EPA. 2013) reviewed three types of models: models that estimate
ambient air ozone concentrations, microenvironmental models, and air exchange models. Air quality
estimates ambient air ozone concentrations at locations where monitoring data are not available. Different
modeling approaches are described in Section 2.3.2.1 through Section 2.3.2.4. Microenvironmental
models use population demographics, time-activity patterns, building characteristics, and air quality data
as inputs for stochastic exposure simulations. Air exchange models estimate air exchange rates (AERs)
for buildings based on building characteristics and meteorological variables and are used as inputs to
microenvironmental models. Air exchange models describe airflow. However, AERs are not specific to
ozone. AER modeling is discussed in the 2013 Ozone ISA.
2.3.2.1 Spatial Interpolation
Spatial interpolation approaches discussed in the 2013 Ozone ISA (U.S. EPA. 2013) include
inverse-distance weighting (IDW) models and kriging. These methods use a mathematical function to
estimate concentrations of ambient ozone in between outdoor locations where ozone is monitored. They
are not resource intensive and can represent concentration with high resolution. However, their limitations
2-6
-------
include the potential to skew the concentration surface by concentrations measured at one monitor
reporting data that are much lower or higher than the majority of monitors, and inaccuracies in capturing
true spatial heterogeneity of ozone concentration.
Recent spatial interpolation studies examine the strengths and limitations of three approaches:
data averaging, IDW, and kriging (Table 2-10). listed in order of increasing model complexity. Joseph et
al. (2013) compared all three interpolation approaches and found that exposure measurement errors were
lower for kriging than for IDW and data averaging. An identified strength of the data averaging and IDW
approaches is their simplicity. Because ozone typically has lower spatial variability compared with NOx
and SOx, greater model complexity (e.g., kriging) might not be needed. However, these methods could
lead to incorrect model fitting if the spatial averaging occurs over an area where variation would be
anticipated, such as next to a road (Simon et al.. 2016). A model evaluation study conducted for daily data
across the city of Montreal, Canada, Buteau et al. (2017) compared IDW with more complex approaches,
including land use regression (LUR) and Bayesian maximum entropy (BME)-LUR, and found that IDW
had the highest interclass correlation coefficient (ICC = 0.89) compared with fixed-site monitors. This
finding suggests that IDW can produce well-validated simulations in an urban area, which is consistent
with findings of low spatial variability of ozone concentrations across cities.
Kriging applies a distance-based covariance or variogram function (spatially or spatiotemporally)
to estimate the ambient the concentration field, which allows for calculation of uncertainties so that the
model provides more information about the range of ambient concentrations at a given location
(Kethireddv et al.. 2014). Like data averaging and IDW, kriging is sensitive to monitor locations.
Misspecification of the covariance or variogram function may not lead to substantial error if the spatial
domain has low variability, as is often the case for ozone (U.S. EPA. 2013). All three methods also have
the potential for preferential sampling that may lead to overestimation or underestimation of
concentrations in some cases (Gelfand et al.. 2012). Specifically, Gelfand et al. (2012) developed kriged
O3 concentration surfaces when fitting only to monitors capturing high concentrations (akin to urban
areas), only to monitors capturing low concentrations (akin to rural areas), and models that incorporate
sites of both type randomly. They found lower out-of-sample prediction errors for the random site
assignment compared with either the urban-focused or rural-focused preferential sampling model,
although the magnitude of the error was not substantially different (high concentration preferential
sampling: 22.7 ppb, low concentration preferential sampling: 23.9 ppb, random sampling: 18.0 ppb).
2.3.2.2 Land Use Regression and Spatiotemporal Modeling
The 2013 Ozone ISA (U.S. EPA. 2013) included some discussion of LUR models. LUR models
regress observed ozone concentrations on land use (and sometimes additional geographic) covariates and
then use the model to predict ambient concentrations where ozone is not measured. The 2013 Ozone ISA
cites high concentration resolution as a strength of the LUR approach. However, for ozone, a lack of
2-7
-------
monitors near roads prevents accurate characterization of ambient concentration in these locations where
scavenging by NO can increase the gradient of ozone concentration. A limited number of studies
employed LUR for ozone exposure assessment at the time of the 2013 Ozone ISA. Specifically,
validation of LUR model results for annual average ozone in urban areas was low (if = 0.06), but the
model performed better in rural areas (R2 = 0.62). The 2013 Ozone ISA did not review spatiotemporal
modeling.
Recent LUR studies have compared model results to fixed-site monitoring data for ozone
concentration and/or alternative models. Comparison with fixed-site monitoring data produced low to
moderate R2 values. In a study of 10 urban areas across the U.S., Clark etal. (2011) reported R2 = 0.34 for
a 5-month summertime average of ozone from hours 10:00 a.m.-6:00 p.m. A study of the entire province
of Quebec produced an R2 of 0.47 in a model of daily 8-hour avg ozone [9:00 a.m.-5:00 p.m.; Adam-
Poupart et al. (2014)1. and the authors discussed larger observed differences between measured and
predicted ozone concentrations in Montreal than in the remainder of the province. In contrast to the other
urban studies reported in the 2013 Ozone ISA (U.S. EPA. 2013) or more recently, Buteau et al. (2017)
found better agreement between LUR estimates and concentrations from fixed-site monitors (interclass
correlation coefficient, ICC = 0.67) in a model of daily 8-hour avg ozone (9:00 a.m.-5:00 p.m.) for
Montreal.
Like LUR, spatiotemporal models may use land use and geographic covariates to model ozone
concentration, but these models may also include a more flexible statistical model formulation and
additional modeling inputs, such as kriging or methods that incorporate an autoregressive structures
(Wang et al.. 2015). Spatiotemporal models were applied in several recent studies to improve predictions
of ozone concentrations for exposure assessment studies (Table 2-11). Several studies used BME
approaches, although the prior information varied across studies and included the kriging mean trend and
the covariance function. BME applies known information as a prior geostatistical distribution, maximizes
an entropy function of the prior distribution, and then applies a Bayes function to estimate a posterior
distribution [i.e., the predicted concentration field; He and Kolovos (2018); Christakos (1990)1. An
important advantage of BME is that it incorporates multiple sources of model inputs (e.g., land use
regression and monitoring data), allowing for the minimization of errors (Adam-Poupart et al.. 2014;
Warren et al.. 2012). BME can also provide a good representation of variability, with both spatial and
temporal variability well represented when autoregressive priors are used (Sahu and Bakar. 2012a. b).
However, predictions tend to be more accurate when there is a higher density of monitors. Partial least
squares (PLS) have also been used as a framework for spatiotemporal modeling. When a large number of
geographic covariates are included in a model, PLS constructs linear combinations of variables, similar to
principal component analysis, which are called "scores." The scores are designed to maximize the spatial
covariance structure of the concentration field while avoiding model overfitting from inclusion of
correlated covariates. PLS was thought to be appropriate for ozone because spatial variability of ozone is
low at most locations in these studies [except near roads; Wang et al. (2016); Xu et al. (2016a); Wang et
al. (2015)1. Like BME models, PLS approaches require sufficient input data to produce accurate models.
2-8
-------
2.3.2.3
Chemical Transport Modeling
In the 2013 Ozone ISA (U.S. EPA. 2013). chemical transport models (CTMs) were briefly
discussed in regard to exposure assessment. CTMs use first principles to characterize the processes that
influence ozone formation (EPA. 2018). CTMs require emissions and meteorological data as inputs. The
chemistry is specified in the model, and concentrations of air pollutants (e.g., ozone) are outputted to a
discrete grid. However, CTMs are limited by their grid cell resolution (e.g., the grid resolution may not
accurately reflect the true heterogeneity of concentration), may be resource intensive to run. The
Community Multiscale Air Quality (CMAQ) model (EPA. 2018) may better capture spatial
heterogeneity, especially in rural areas, compared to interpolation methods (Bell. 2006). However, a
coarse grid size may result in difficulty differentiating ozone concentrations near roadways due to
scavenging by NO (Marshall et al.. 2008).
The number of studies using CTMs has greatly expanded since the 2013 Ozone ISA ITJ.S. EPA
(2013); Table 2-121. A few studies directly compared different CTMs. Zhang et al. (2013) compared
CMAQ with the Comprehensive Air quality Model with extensions (CAMx) in January and July of 2002
in the southeastern U.S. Results were presented by monitoring network, month, and averaging time
(i.e., 1-hour daily max ozone and 8-hour daily max ozone). The configurations between CMAQ and
CAMx for this work varied in several ways, including the horizontal and vertical advection mechanism
and deposition. Both models mostly overpredicted ozone concentrations when compared with monitor
data in January, most likely due to neither model accounted for vertical mixing. In July, overprediction
occurred for CAMx while underprediction occurred for CMAQ when compared to monitor data, as noted
in Appendix 1 (Section 1.6.2). Underprediction was mostly due to underestimation of emissions
precursors and peak temperatures. Wang et al. (2016) modeled surface ozone in Los Angeles, CA over
the years 2000-2008 and found that ozone had a greater root mean squared error (RMSE) for the
University of California at Davis—California Institute of Technology (UCD-CIT) CTM in rural areas,
especially during the warm season. Wang et al. (2016) did not quantify the direction of error. Herwehe et
al. (2011) compared CMAQ with Weather Research and Forecasting coupled with Chemistry
(WRF/Chem) across the continental U.S. (CONUS) in August 2006. While CMAQ generally performed
better than WRF/Chem, both CTMs overpredicted the 8-hour daily max ozone concentration (e.g., mean
bias was 3.62 ppb for CMAQ). WRF/Chem likely predicted higher ozone concentrations due to vertical
mixing in the boundary layer, dry deposition, and convection schemes. Many studies covered
approximately all of the CONUS (Appel et al.. 2017; Appel et al.. 2013; Appel et al.. 2012) or covered
localized, urban areas of either the U.S. or Canada [e.g., Houston, TX; Seattle, WA; San Joaquin Valley,
CA; Los Angeles, CA; Vancouver, Canada; Wang et al. (2016); Stevn et al. (2013); Hu et al. (2012);
Tsimpidi et al. (2012); Ying and Li (2011)1.
The horizontal grid resolution of a CTM can influence the heterogeneity of an ambient
concentration surface and the associated exposure estimate, but those effects are generally only seen for
areas of peak concentrations. Some studies directly compared how ozone model performance statistics
2-9
-------
changed with the size of a given grid cell (Yu et al.. 2016; Schaap et al.. 2015; Thompson and Selin.
2012; Tsimpidi et al.. 2012). The majority of these studies found that the cumulative distribution function
and summary statistics did not vary greatly among the simulations using different spatial resolutions.
However, the upper and lower tails of the distributions had greater error for coarse resolution simulations
compared with fine resolution simulations (Yu et al.. 2016). Schaap et al. (2015) compared the Multiscale
Chemical Transport Model for Atmospheric Composition Analysis and Forecast (CHIMERE), CMAQ,
European Monitoring and Evaluation Program (EMEP), Research and Development Center for Global
Change (RCGC), and Long-Term Ozone Simulation-European Operational Smog (LOTOS-EUROS)
models across grid resolutions for rural, suburban, and urban areas. They found that the largest magnitude
biases between the model and observations were for the urban simulations, and the models more often
overestimated rather than underestimated measurements for all settings. These findings are consistent
with older studies not cited in the 2013 Ozone ISA (U.S. EPA. 2013). Cohan et al. (2006) compared
CMAQ simulations for 4-, 12-, and 36-km resolutions and found that the 4-km resolution simulation was
more sensitive to fluctuations in precursor emissions but observed little difference among the simulations
in daily 8-hour max concentrations. Similarly, Henderson et al. (2010) found that 1-km grid resolution
allowed detection of much larger ozone concentration peaks, but otherwise observed little difference
between the 1- and 4-km simulations.
Many short-term CTM studies relevant to short-term exposure assessment either characterized a
specific high ozone event (e.g., wildfire), tested a new mechanism of a model (e.g., planetary boundary
layer schemes, a new version of the master chemical mechanisms exploring two-way coupling), or both.
For example, Baker et al. (2016) explored how the Wallow wildfire and Flint Hills prescribed burn in
2011 influenced ozone concentration in those localized areas. In the case of a wildfire, bias increased by
approximately 2 ppb for every 1 ppb increase in estimated ozone contribution from the fire. For
prescribed burns, bias increased by approximately 1 ppb for every 1 ppb increase in estimated ozone
contribution from the fire. Wong et al. (2012) developed a two-way coupled system for CMAQ in which
the WRF and CMAQ components could consistently be executed in and around California for a week in
June 2008 during a wildfire event. For all data, comparison of the model with measurements showed little
bias (slope = 0.98) with observable scatter (R = 0.62). When data were limited to aerosol optical density
(AOD) >0.5, the model had a positive bias (slope = 1.2) but with less scatter (R = 0.75). Given that bias
was related to ozone concentration in these studies, the results suggest that bias is greater, and emissions
are more uncertain for wildfires than for prescribed burns.
Inaccurate characterization of cloudiness has been shown to lead to biased or uncertain ozone
concentrations due to the influence of photolysis on ozone formation, potentially leading to biased or
uncertain exposure estimates. In Ngan et al. (2012). the modeling scheme misrepresented cloud locations,
which also affected the modeling of PM due to the deposition and removal processes that occur by
precipitation, leading to an overestimation of ozone. Pan et al. (2015) overestimated ozone during part of
the modeling time period possibly due to uncertainties in the cloud fraction along with other
meteorological variables. Yahya et al. (2016) found that for a 10-year avg of certain cloud variables
2-10
-------
(i.e., droplet number concentration, cloud water path, and cloud optical thickness), ozone concentrations
were generally underpredicted for most regions of the U.S.
CTMs have been shown to underestimate high concentrations and overestimate low
concentrations, which could impact estimates of peak exposure. In Tsimpidi et al. (2012). the normalized
mean bias (NMB) was slightly negative (-7.9%) at a 4-km grid resolution in CMAQ when concentrations
less than 40 ppb were excluded from the bias calculations for the Pacific northwest in July 2006.
However, when all concentrations were included, the statistic became large and positive in magnitude
(42.7%). Tsimpidi et al. (2012) attributed overprediction of nighttime ozone concentrations to inaccurate
models of vertical diffusivity in CMAQ. Similarly, for northern California in July 2009, Bash et al.
(2016) observed overprediction of ozone concentrations in CMAQ by a median bias of 29-32%
(depending on the biogenic VOC model) when ozone concentrations were less than 60 ppb, while median
bias was -8 to -9% when ozone concentration was greater than 60 ppb. Garner et al. (2015) observed a
positive mean bias in the early morning hours that decreased throughout the morning for Baltimore, MD
in July 2011.
2.3.2.4 Hybrid Approaches
Hybrid models were not reviewed in the 2013 Ozone ISA (U.S. EPA. 2013). Like spatiotemporal
models, hybrid models use information from multiple data sources to estimates ambient concentrations.
Several hybrid models combine observed data from fixed-site regulatory monitors with CTMs that are
defined over a spatiotemporal grid (Table 2-13). These separate data sources are combined in such a
manner that the resulting exposure prediction is a "hybrid" of the input data sources. These approaches
are frequently referred to as "data fusion" methods. A CTM estimates ambient ozone concentration over a
grid and in the data fusion framework, is typically assigned at a point located at the centroid of the grid. In
the "downscaler" method, Berrocal et al. (2012) adjusted CTM estimates at any arbitrary location in the
domain based on a weighted average of the CTM estimates for surrounding grids paired with
observational data such that exposure estimates were predicted at spatial scales finer than the input CTM
(e.g., a census tract). This hybrid model had an improved performance (mean squared error, MSE:
45.4 ppb2, mean absolute error, MAE: 5.0 ppb) compared with either the observed data alone (MSE:
124 ppb2, MAE: 8.7 ppb) or the CTM estimates alone (MSE: 136 ppb2, MAE: 9.1 ppb) when predicting
ozone over the eastern CONUS in summer 2001 with CMAQ estimates. Its predictive power (as
quantified through the difference in the predictive mean squared error [PMSE]) was more pronounced
than previous iterations of the "downscaler" in areas far from monitoring locations. That is, the difference
in PMSE from an older version of the "downscaler" to a newer version increased with increasing average
distance from training sites.
Other studies found similarly improved performance with use of BME. Xu et al. (2016b) used a
BME approach to merge ambient ozone concentration data from the Air Quality System (AQS) database
2-11
-------
with CAMx simulations modeled at a 36-km scale and incorporated a regional correction factor to allow
for flexible selection of spatial points included in the model. The regionalized model decreased RMSE
from 6.7 to 5.5 ppb and increased R2 from 0.88 to 0.89 for 8-hour daily max ambient ozone concentration.
Xu et al. (2017) found that RMSE was larger when using hourly ambient ozone concentrations as model
input compared with 8-hour daily max and 24-hour avg input concentrations to make 8-hour daily max
and 24-hour avg predictions, respectively. RMSE was also slightly larger when a 36-km grid was used in
the CAMx model compared with a 12-km grid. For the U.S. and Canada, Robichaud and Menard (2014)
combined predictions from the CTM Canadian Hemispheric and Regional Ozone and NOx System
(CHRONOS, 2005) and Global Environmental Multiscale coupled with Model of Air quality and
Chemistry (GEM-MACH, 2012) with surface data from AQS and Canadian databases through an optimal
interpolation scheme in which the model and monitor data were linked through a Kalman filter
optimization matrix. This approach is known as Objective Analysis, and it was shown to produce
near-zero systematic errors and smaller random errors (of positive magnitude) compared with CTM alone.
Reich et al. (2014) employed a more flexible downscaler using spectral methods. When compared with a
linear downscaler, the spectral downscaler had a smaller bias and MSE in the ambient ozone
concentration estimate. Other hybrid methods are more straightforward and use weighting factors to
combine data sources (Friberg et al.. 2016). This weighting approach reduced error in ambient ozone
concentration and increased the spatial correlation compared with using CMAQ alone.
Hybrid models need not be restricted to only CTM and observed data. Di et al. (2017) calibrated
satellite-based MODIS vegetation data against GEOS-Chem CTM output. They predicted ground-level
8-hour daily max ambient ozone concentrations as a function of the calibrated satellite data along with
surface ambient ozone concentrations in the AQS database and land use variables in a neural network
model that can accommodate nonlinearity of the variables. For the years 2000-2012, 10-fold
cross-validation bias was reported to be 20% with R2 of 0.76. Tang et al. (2015b) adjusted CMAQ output
with both observed data and MODIS AOD data from Terra and Aqua satellites. Incorporating observed
and satellite data improved the correlation between surface observed data when compared to CMAQ
alone in the southeastern CONUS, with mixed results for mean bias in the prediction of ambient ozone
concentration. A recent study of spatial and temporal biases in satellite data informs our understanding of
the limitations of hybrid models using satellite data as inputs (Verstraeten et al.. 2013). Global column
data obtained using Tropospheric Emissions Spectrometer (TES) version 4 were compared with
ozonesonde balloon measurements obtained over 2005-2010. Negative biases in ozonesonde
concentration measurements up to 8 ppb were noted for June, July, and August in the midlatitudes
(coincident with the U.S.). Larger biases were observed for the midlatitudes compared with the subtropics
or tropics. These data were obtained at a single time during the early afternoon as the ozonesonde passed
each location, so nighttime ambient ozone concentrations were not accounted for.
2-12
-------
2.3.2.5 Microenvironmental Modeling
The 2013 Ozone ISA and the 2019 PM ISA (U.S. EPA. 2018. 2013) presented several studies that
evaluated integrated microenvironmental exposure (ME) and dose models, such as the Air Pollutants
Exposure (APEX) and the Stochastic Human Exposure and Dose Simulation (SHEDS) model (U.S. EPA.
2016. 2009). These ME models use pollutant concentrations, human time-activity patterns, population
demographics (e.g., age, sex, race/ethnicity), anthropometric attributes (e.g., height, weight),
physiological information (e.g., energy expenditure, breathing rate), and building ventilation to predict
human exposures that occur in different locations (e.g., at home, at work, in a vehicle). Both APEX and
SHEDS models are population-based, meaning that demographic data from the US Census are used to
appropriately weight the model simulated population to represent a larger population. APEX in particular
(and prior versions of that model) has been used to estimate exposure and risk for criteria pollutants over
the past three decades. Because these models are probabilistic, many of the key model input variables
(e.g., resting metabolic rate, ventilation rate, microenvironmental factors) are sampling for statistical
distributions to account for variability in these factors and the resulting exposures. In addition, both
models can be run using Monte Carlo simulations that allow for evaluating how uncertainty of input
parameters affects variability in exposures to simulated individuals. These two models different in that
SHEDS can estimate multiroute exposures, including inhalation, while APEX only simulates inhalation
exposure. Advantages of ME that were identified in the 2013 Ozone ISA include the ability of the user to
design analyses for specific populations (assuming that demographic and time-activity data are available)
and to include indoor air sources (which are uncommon for ozone). Limitations include resource
intensiveness of the ME models and that indoor exposures cannot be easily evaluated (Gcorgopoulos ct
al.. 2005). the latter of which is of lesser importance for estimating ambient air exposure.
Strengths and limitations identified in recent studies (Table 2-14) agree with ME model studies
presented in the 2013 Ozone ISA (U.S. EPA. 2013). Dionisio et al. (2014) found that exposure estimates
were 72% higher when using the ME model with an incorporated CTM compared with using fixed-site
measurements alone as exposure surrogates. The majority of that difference came from the inclusion of
time-activity data in the ME model. Reich et al. (2012) found that ME models produced lower estimates
of exposure than did fixed-site monitors. However, the ME models were not validated with personal
monitors, so the extent of this error was unknown. Dionisio et al. (2017) used APEX to simulate O3
exposure to individuals in 12 U.S. cities. They found for the simulated scenarios that climate and air
quality were more influential on exposure than simulated demographic changes. Ozkavnak et al. (2014)
compared SHEDS and APEX with other exposure models for use as input into an epidemiologic analysis
in Atlanta, GA and New Jersey.
2-13
-------
2.4 Personal Exposure
This section builds upon discussions from the 2013 Ozone ISA (U.S. EPA. 2013) about
relationships between indoor and outdoor ambient ozone concentrations and between personal exposure
to ambient ozone and ambient ozone concentrations. Section 2.4.1 describes recent advances in
characterizing time-activity data for exposed people, given advances in global positioning system (GPS)
technologies and the continued updating of the Consolidated Human Activity Database (CHAD).
Summaries of relevant discussions from the 2013 Ozone ISA are included in Section 2.4.2 and
Section 2.4.3. Findings in more recent studies are largely consistent with the findings reported in the 2013
Ozone ISA.
2.4.1 Time-Activity Data
The 2013 Ozone ISA (U.S. EPA. 2013) provided only limited discussion of time-activity
patterns. Ozone-averting behavior, or the tendency to stay indoors as much as possible to avoid exposure
on days with high ambient ozone concentrations, as reported by the news media, was described as one
factor that could change time-activity patterns. Recent technological advances in GPS technologies and
expansions to existing time-activity databases have expanded the information base regarding
time-activity. Such new tools have enabled an examination of factors that influence time-activity patterns
and errors in those relationships.
Data through 2010 are available from the CHAD database to compare time-activity data among
different population strata for 25,431 individuals who reported 54,373 days of data (Isaacs. 2014).
Summaries of the percentage and number of person-minutes in different locations based on all individuals
with diaries in this version of CHAD during the warm months (April-September) for all day
(12:00 a.m.-11:59 p.m.) and for the afternoon and early evening (12:00 p.m.-8:00 p.m.), assumed to be
the period when ozone concentration is at a maximum each day, are presented in Table 2-1 through
Table 2-3 based on an analysis of the data within the Office of Research and Development (ORD). These
tables indicate that across this population of individuals in CHAD, substantially more time was spent
indoors at home for children younger than 6 years and for adults older than 64 years, while teens ages
12-19 years and adults 20-64 years spent the least amount of time indoors at home. Similarly, young
children spent the least amount of time in transit, while adults 20-64 years spent the most time in transit.
Teens ages 12-19 years spent the largest proportion of the day outdoors, while older adults spent the least
amount of time outdoors. Time spent outdoors by young children ages 0-5 years was similar to that of
older adults. A prior analysis of CHAD data gauged the percentage of study participants who engaged in
outdoor activities (participation rate was defined as the percentage of person-days in spending at least
1 minute outdoors) and the number of minutes spent outdoors per day during the afternoon and early
evening (12:00 p.m.-8:00 p.m.), assumed to be the period when ozone concentration is at a maximum
each day (U.S. EPA. 2014). Children and teens ages 4-18 years had the largest participation rate among
2-14
-------
those spending more than 2 hours outdoors and the largest mean time outdoors per person spending at
least 1 minute outdoors, while younger adults (ages 19-35 years) had the highest participation rate among
those spending more than 1 minute outdoors, and adults ages 35-50 years had the largest mean outdoors
per person among those spending more than 2 hours outdoors. Moreover, U.S. EPA (2014) calculated that
79% of time spent by children ages 4-18 years and 63% of afternoon time outdoors spent by adults ages
19-95 years involved at least moderate exertion. This CHAD summary of time-activity data by
race/ethnicity from the CHAD database (Table 2-2) indicates that Hispanic study participants spent
slightly more time at a residential location indoors than the total population, while white study
participants spent the most time outdoors compared with Asian, black, and Hispanic participants
(Table 2-2). However, 11 % of participants had missing race/ethnicity data or did not to provide
information regarding race/ethnicity, so these results should be interpreted cautiously. Males spent more
time outdoors than females (Table 2-3). Such analyses of the CHAD data collectively suggest that older
children, males, and those of white race may spend the most time outdoors during warm weather, where
they could be exposed to elevated ozone concentrations.
The CHAD database provides a detailed picture of time-activity across population groups and has
a large number of days. Several caveats should be noted (Graham and Mccurdv. 2004). CHAD combines
data from several different studies conducted over several years. These studies collected data under
different circumstances, and in some cases, variables could not be combined. Validation techniques for
the data inputted into CHAD may have differed across studies, and it is possible that participants were not
precise in providing time increments or that missingness of data could have been handled differently
across studies. Moreover, when presenting the data by age, race/ethnicity, and sex, some studies may have
contributed a disproportionate amount of data to that group, because the objective of the individual study
may have been to characterize time-activity patterns for a segment of the population rather than for the
population as a whole.
Recent studies have focused on the use of GPS technologies, such as in smartphones, to develop
detailed time-activity pattern data. This technology has the potential to allow a time-activity study to
overcome limitations of time-activity diaries, such as imprecise estimation of time-location data. For
example, Glasgow et al. (2014) analyzed the frequency of Android-based smartphones in recording
positional data among a panel of study participants and found that on average 74% of the data was
collected over intervals shorter than 5 minutes, which is a marked improvement over many time-activity
studies using diaries. Positional errors are also a concern for GIS and GPS-based technologies. Several
studies found that typical location errors based on smartphones were less than 26 m (Ganguly et al.. 2015;
Lane et al.. 2013; Wu et al.. 2010). Glasgow et al. (2014) observed much larger errors, with an overall
median positional accuracy of 342 m and a range from 98 to 1,169 m using an Android-based
smartphone, while Wu et al. (2010) observed much smaller errors when comparing two smartphones with
three other GPS technologies. The magnitude of positional errors may be important, because positional
error has the potential to lead to location misclassification of time-activity patterns.
2-15
-------
Survey tools to assess time-activity patterns may be subject to recall error among the subjects.
Spalt et al. (2015) administered a retrospective survey to all participants in the Multi-Ethnic Study of
Atherosclerosis (MESA) Air Study to ascertain time spent indoors and outdoors at home, at
work/volunteer/school, in transit, or in other locations. A subset of the study population was asked to
complete a detailed time-activity diary in addition to the survey. Correlation between the MESA Air
surveys and the time-activity diaries for indoor locations was Spearman R = 0.63 for home, Spearman
R = 0.73 for work/volunteer/school, and Spearman R = 0.20 for other indoor locations. Correlation
between the MESA Air surveys and the time-activity diaries for outdoor locations was much lower, with
Spearman R = 0.14 at home, Spearman R = 0.20 for work/volunteer/school, and Spearman R = 0.10 for
other outdoor locations. Correlation between MESA Air surveys and time-activity diaries for individuals
in transit was Spearman R = 0.39. These results suggest that study participants have better recall of the
times spent inside their home or work/volunteer/school compared to other activities, because time spent at
home or at work/volunteer/school tends to occur at routine times.
Residential mobility is another source of exposure measurement error in long-term exposure
studies. Using a single address to represent exposure concentration over a period of several years may
result in either under- or overestimating exposure during the study period. For example, Brokamp et al.
(2015) analyzed residential mobility for a cohort of children over the first 7 years of life in Cincinnati,
OH and found that 54% of the children changed residential address during that time, resulting in a 4.4%
decrease in the cohort's average traffic-related air pollution concentration (defined as black carbon
estimates from an LUR model for this study). They also noted that if the birth address is used for
exposure estimation during the entire study period, exposure misclassification is increased for those that
move earlier (due to more years at the incorrect address) or are more highly exposed (due to a greater
likelihood of moving). The Brokamp et al. (2015) study showed that not accounting for residential
mobility resulted in bias toward the null. Exposure measurement error due to incorrect home address
would be expected to be lower for ozone compared with more spatially variable air pollutants, but it
would not necessarily be negligible.
Studies examining seasonal differences in time-activity data are limited. Spalt et al. (2015) in the
MESA Air Study found that participants spent more time outdoors in the summer than in winter. Using
the CHAD database, Baxter et al. (2013) found that time spent outdoors were similar across seasons for
participants. However, also using the CHAD database, Jiao et al. (2012) found for the cities investigated
that the elderly population in this study spent more time outside in the summer than in winter.
Updated time-activity data and tools for assessing time-activity data have improved the general
understanding of time-activity data and related uncertainties in recent years. Analysis of CHAD diaries
indicated that young children ages 0-5 years were found to spend less time outdoors than older children,
teens, and adults, and white respondents spent more time outdoors than their Asian, black, and Hispanic
counterparts (Isaacs. 2014). New technologies to assess study participant location, errors related to study
participant recall, and residential mobility have been used to determine that location-based errors are
2-16
-------
within 6% for short- and long-term exposure assessment, while omission of residential mobility can
produce bias in the exposure estimate, resulting in negatively biasing the effect estimate for a study of
long-term ozone exposure.
2-17
-------
Table 2-1 Total and age-stratified percentage of hours spent in different
locations from the Consolidated Human Activity Database (Isaacs.
2014). warm season for all hours and for afternoon hours
(12:00 p.m.-8:00 p.m.).
Location Type
All
0-5 yr
6-11 yr
12-19 yr
20-64 yr
65+ yr
N (number of
individuals)
(%)
12,673
(100)
2,253
(18)
2,010
(16)
1,080
(8.5)
5,785
(46)
1,403
(11)
Warm season, 12:00 a.m-
11:59 p.m. (person-minutes [%])
Indoor-residential
31,038,736
(75)
5,932,419
(81)
4,474,880
(74)
1,846,578
(71)
13,593,134
(71)
5,012,405
(83)
Transit
2,359,073
(5.7)
284,770
(3.9)
242,706
(4.0)
134,854
(5.2)
1,352,166
(7.1)
329,628
(5.4)
Indoor-
work/school/other
5,956,425
(14)
725,350
(9.9)
906,573
(15)
429,633
(17)
3,342,988
(17)
504,961
(8.3)
Outdoor
1,562,018
(3.8)
214,340
(2.9)
255,882
(4.2)
130,973
(5.1)
793,278
(4.1)
162,906
(2.7)
Uncertain or missing
521,188
(1.3)
171,281
(2.3)
179,479
(3.0)
47,082
(1.8)
74,754
(0.39)
48,180
(0.80)
Warm season, 12:00 p.m.-
-8:00 p.m. (person-minutes [%])
Indoor-residential
9,631,182
(63)
1,867,690
(68)
1,327,384
(59)
529,438
(55)
3,853,625
(55)
1,601,243
(72)
Transit
1,374,499
(9.0)
184,339
(6.7)
157,601
(7.0)
81,709
(8.5)
781,252
(11)
194,099
(8.8)
Indoor-
work/school/other
3,165,397
(21)
412,785
(15)
470,148
(21)
225,926
(24)
1,819,323
(26)
288,478
(13)
Outdoor
981,361
(6.4)
154,838
(5.6)
198,851
(8.8)
92,928
(9.7)
476,679
(6.7)
89,982
(4.1)
Uncertain or missing
190,037
(1.2)
53,793
(2.0)
56,847
(2.5)
14,958
(1.6)
37,363
(0.53)
26,851
(1.2)
Note: Data presented in this table for person-minutes are calculated as the sum of minutes across individuals and percentage is
calculated as the percentage of total person-minutes for a given category. Data are filtered by the criteria noted in the column
headings. Some participants had missing age data. Data were downloaded from
https://www.epa.aov/healthresearch/consolidated-human-activitv-database-chad-use-human-exposure-and-health-studies-and.
Source: CHAD database.
2-18
-------
Table 2-2 Total and race/ethnicity-stratified percentage of hours spent in
different locations from the Consolidated Human Activity Database
(Isaacs. 2014). warm season for all hours and for afternoon hours
(12:00 p.m.-8:00 p.m.).
Location
Type
All
Asian
Black
Hispanic
White
Other
N (number of
12,673
248
1,829
729
8,083
310
individuals)
(%)
(100)
(2.0)
(14)
(5.8)
(64)
(2.4)
Warm season, 12:00 a.m.-11:59 p.m. (person-minutes [%])
Indoor-
31,038,736
693,831
4,026,861
2,278,661
20,590,280
968,084
residential
(75)
(75)
(75)
(78)
(75)
(77)
Transit
2,359,073
45,730
309,221
153,744
1,576,269
69,455
(5.7)
(4.9)
(5.8)
(5.3)
(5.7)
(5.5)
Indoor-
5,956,425
153,540
768,617
376,251
3,957,782
170,660
work/school/
other
(14)
(17)
(14)
(13)
(14)
(14)
Outdoor
1,562,018
21,449
160,968
95,143
1,134,048
38,127
(3.8)
(2.3)
(3.0)
(3.2)
(4.1)
(3.0)
Uncertain or
521,188
12,810
69,533
23,721
357,941
13,674
missing
(1.3)
(1.4)
(1.3)
(0.81)
(1.3)
(1.1)
Warm season, 12:00 p.m.-8:00 p.m. (person-minutes [%])
Indoor-
9,631,182
201,679
1,166,578
693,833
6,157,342
297,935
residential
(63)
(58)
(60)
(64)
(60)
(63)
Transit
1,374,499
28,237
188,304
93,949
935,729
41,811
(9.0)
(8.2)
(9.6)
(8.7)
(9.1)
(8.9)
Indoor-
3,165,397
90,168
413,189
208,599
2,151,117
90,752
work/school/
other
(21)
(26)
(21)
(19)
(21)
(19)
Outdoor
981,361
15,146
113,455
62,517
727,574
25,541
(6.4)
(4.4)
(5.8)
(5.8)
(7.1)
(5.4)
Uncertain or
188,871
4,427
38,950
9,393
114,541
5,295
missing
(1.2)
(2.0)
(2.0)
(0.87)
(1.1)
(1.1)
Note: Data presented in this table for person-minutes are calculated as the sum of minutes across individuals and percentage is
calculated as the percentage of total person-minutes for a given category. Data are filtered by the criteria noted in the column
headings. Some participants had missing race/ethnicity data. Data were downloaded from
https://www.epa.aov/healthresearch/consolidated-human-activitv-database-chad-use-human-exposure-and-health-studies-and.
Source: CHAD database.
2-19
-------
Table 2-3 Total and sex-stratified percentage of hours spent in different
locations from the Consolidated Human Activity Database (Isaacs.
2014). warm season for all hours and for afternoon hours
(12:00 p.m.-8:00 p.m.).
Location Type
All
Female
Male
N (number of individuals)
12,673
6,821
5,849
(%)
(100)
(54)
(46)
Warm season, 12:00 a.m.-11:59 p.m. (person-minutes [%])
Indoor-residential
31,038,736
16,945,109
14,086,470
(75)
(77)
(73)
Transit
2,359,073
1,235,125
1,123,433
(5.7)
(5.6)
(5.8)
Indoor-work/school/other
5,956,425
2,996,029
2,959,731
(14)
(14)
(15)
Outdoor
1,562,018
657,845
903,889
(3.8)
(3.0)
(4.7)
Uncertain or missing
521,188
246,852
274,317
(1.3)
(1.1)
(1.4)
Warm season, 12:00 p.m.-8:00 p.m. (person-minutes [%])
Indoor-residential
9,631,182
5,069,614
4,162,147
(63)
(62)
(58)
Transit
1,374,499
752,611
654,873
(9.0)
(9.2)
(9.1)
Indoor-work/school/other
3,165,397
1,670,920
1,569,613
(21)
(20)
(22)
Outdoor
981,361
441,741
573,859
(6.4)
(5.4)
(8.0)
Uncertain or missing
188,871
96,730
93,307
(1.2)
(1.2)
(1.3)
Note: Data presented in this table for person-minutes are calculated as the sum of minutes across individuals and percentage is
calculated as the percentage of total person-minutes for a given category. Data are filtered by the criteria noted in the column
headings. Some participants had missing gender data. Data were downloaded from
https://www.epa.aov/healthresearch/consolidated-human-activitv-database-chad-use-human-exposure-and-health-studies-and
Source: CHAD database.
2-20
-------
2.4.2 Infiltration
The 2013 Ozone ISA (U.S. EPA. 2013) reviewed literature on indoor-outdoor (I/O) ratios to
describe infiltration of ambient ozone into homes and buildings. Ozone generation indoors is uncommon,
as described in the 2013 Ozone ISA. Assuming an absence of devices that generate ozone, such as certain
ozone-generating household air purifiers and some consumer products (Zhang and Jenkins. 2017). I/O
ratios generally ranged from 0.1-0.4.
Table 2-4 summarizes I/O ratios from ozone infiltration studies across the U.S. Several of the
studies report I/O ratios below 0.2 when the windows are closed and the air exchange rate (AER) is
0.5/hour or lower. Across studies, I/O tended to increase with higher values of AER from open windows
or mechanical ventilation. Studies where air exchange was reported to have been higher, primarily in
commercial areas or offices (Ben-David and Waring. 2018; Chan et al.. 2014) or where windows were
open (Dutton et al.. 2013). tended to report higher I/O ratios compared with studies of lower AER,
primarily in homes (Singer et al.. 2016; Sarnat et al.. 2013; Chen et al.. 2012). Johnson et al. (2014) also
examined ozone infiltration in vehicles, and the mean and range of I/O ratios were between the values
obtained for open versus closed windows or doors. Ozone vehicle infiltration was further explored in
Johnson et al. (2018) to evaluate parameter inputs into the APEX model for different microenvironments
including in-vehicle microenvironments. They found that I/O ratios for measured values were lower than
APEX up until the 75th percentile.
Sarnat et al. (2013) explored how AER can modify the effect of ozone related to asthma
emergency department (ED) visits in Atlanta neighborhoods. Parsing their data by low (<0.227/hour
threshold), medium (0.228-0.308/hour) and high AER (>0.309/hour threshold) did not appreciably
influence the risk of asthma ED visits, but the ozone level (low: <32 ppb, moderate: 33-53 ppb, high:
>54 ppb) was related to an increase in risk ratio from approximately 1 for low ozone to 1.02 for moderate
ozone to 1.08 for high ozone. In contrast, the risk ratios of asthma ED visits and PM2 5 and NOx
exposures showed some sensitivity to AER. High levels of poverty (8.5% threshold) were associated with
high AER. They attributed this observation to old, drafty housing being more prevalent among those in
poverty.
The literature on I/O ratios for ozone is limited (Table 2-4). Within the available literature,
sources of uncertainty include the sample size within the study (Chan et al.. 2014) and the use of windows
(Dutton et al.. 2013). Variation in I/O may be due to differences in experimental procedures from study to
study, building-to-building variation, day-to-day variation (Chan et al.. 2014). cooking events, seasonal
differences in ambient concentrations, the type of filter in a building (Singer et al.. 2016). AER, and home
ventilation (Sarnat et al.. 2013).
2-21
-------
Table 2-4 Summary of U.S. studies of ozone infiltration published after 2011.
Reference
Location
Ambient
Time Ozone
Period Population Microenvironment Concentration
I/O Ratio
Correlation
AER
Sarnatetal. (2013)
Atlanta, GA
January Residents Home Mean (SD):
1999- above and 41.9 ppb
December below (18.6 ppb)
2002 poverty
NR
AER: -0.19
Mean (SD):
0.265/h
(0.108/h)
Chen et al. (2012)
NMMAPS cities:
1987-2000 All Home (estimated NR
residents by model)
Calculated change in
indoor ozone per
unit change in
ambient ozone3
NR
Mean
Atlanta, GA
NR
0.14
NR
0.43/h
Birmingham, AL
NR
0.14
NR
0.43/h
Boston, MA
NR
0.20
NR
0.68/h
Buffalo, NY
NR
0.20
NR
0.70/h
Chicago, IL
NR
0.18
NR
0.61/h
Cincinnati, OH
NR
0.16
NR
0.52/h
Corpus Christi,
TX
NR
0.17
NR
0.48/h
Dallas/Ft. Worth,
TX
NR
0.16
NR
0.50/h
Denver, CO
NR
0.16
NR
0.49/h
Los Angeles, CA
NR
0.13
NR
0.42/h
2-22
-------
Table 2-4 (Continued): Summary of U.S. studies of ozone infiltration published after 2011.
Reference
Chen et al. (2012)
(continued)
Location
Time
Period
Population
Microenvironment
Ambient
Ozone
Concentration
I/O Ratio
Correlation
AER
Miami, FL
1987-2000
All
residents
(continued)
Home (estimated
by model)
(continued)
NR
0.15
NR
0.35/h
Nashville, TN
NR
0.16
NR
0.51/h
New York City,
NY
NR
0.20
NR
0.62/h
Phoenix, AZ
NR
0.14
NR
0.42/h
Seattle, WA
NR
0.17
NR
0.62/h
St. Louis, MO
NR
0.18
NR
0.58/h
Washington, DC
NR
0.16
NR
0.54/h
Worcester, MA
NR
0.18
NR
0.60/h
2-23
-------
Table 2-4 (Continued): Summary of U.S. studies of ozone infiltration published after 2011.
Reference
Location
Time
Period
Population
Microenvironment
Ambient
Ozone
Concentration
I/O Ratio
Correlation
AER
Dutton etal. (2013)
Alameda, CA
September
6—
December
4, 2011
Office
workers
Office 1 windows
closed
NR
Mean (SD): 0.18
(0.11)
Peak: 0.78
NR
NR
Office 1 windows
open
NR
Mean (SD): 0.37
(0.18)
Peak: 0.78
NR
NR
Oakland, CA
June 15-
July 1, 2012
Office 2 windows
closed
NR
Mean (SD): NR
Peak: 0.52
NR
NR
Office 2 windows
open
NR
Mean (SD): 0.24
(0.10)
Peak: 0.52
NR
NR
El Cerrito, CA
July 2-20,
2012
Office 3 windows
closed
NR
Mean (SD): 0.18
(0.07)
Peak: 0.54
NR
NR
Office 3 windows
open
NR
Mean (SD): 0.28
(0.14)
Peak: 0.54
NR
NR
Berkeley, CA
NR
Office 4 windows
closed
NR
Mean (SD): NR
Peak: 0.68
NR
NR
Office 4 windows
open
NR
NR
NR
NR
2-24
-------
Table 2-4 (Continued): Summary of U.S. studies of ozone infiltration published after 2011.
Ambient
Time
Ozone
Reference
Location
Period Population
Microenvironment
Concentration
I/O Ratio
Correlation
AER
Chan et al. (2014)
San Francisco
September Store
Grocery stores
Avg, 1-h daily
Mean (range): 0.40
NR
Across stores:
Bay Area,
2011-March occupants
max across
(0.18-0.59)
0.65-1.47/h
Sacramento Area,
2013
stores:
Fresno, Los
24.1-66.7 ppb,
Angeles Area, CA
30.0-79.4 ppb
Furniture/hardware
Avg, 1-h daily
Mean (range): 0.42
NR
Across stores:
stores
max across
(0.29-0.47)
0.39-2.38/h
stores:
30.1-62.1 ppb,
30.1-62.1 ppb
Apparel stores
Avg, 1-h daily
Mean (range): 0.33
NR
Across stores:
max across
(0.11-0.47)
0.52-2.33/h
stores:
12.1-51.5 ppb,
15.1-59.6 ppb
Johnson et al. (2014) Durham. NC
August-
September
2012
Various stores
1-h avg:
29-58 ppb
(by day)
Mean (range): 0.17
(-0.012-0.78)
NR
NR
Windows or doors
open
Mean (range): 0.44
(0.13-0.780)
Windows or doors
closed
Mean (range): 0.093
(0.00-0.30)
In-vehicle (driving,
parked, refueling,
roadside)
Mean (range): 0.33
(0.0063-0.70)
2-25
-------
Table 2-4 (Continued): Summary of U.S. studies of ozone infiltration published after 2011.
Reference
Location
Ambient
Time Ozone
Period Population Microenvironment Concentration
I/O Ratio
Correlation
AER
Gall et al. (2011) Houston, TX
Simulation Simulated No passive removal
homes materials
NR
Average: 0.16
NR
Simulated Gypsum, activated
homes carbon cloth, or
other removal
materials
NR
Average: 0.047-0.12
NR
0.5/h
0.5/h
Na et al. (2015)
Atlanta, GA
NR
Simulated
box store
ASHRAE
prescribed
ventilation
Volume-weighted
concentrations
Average of
daily average
(daily peak):
76 ppb
(93 ppb)
Calculated as mean
indoor/daily average
outdoor: daily
average (daily peak):
0.13 (0.13)
NR
Average of
daily average
(daily peak):
76 ppb
(93 ppb)
Calculated as mean
indoor/daily average
outdoor: daily
average (daily peak):
0.079 (0.075)
NR
1.2 L/s-m2
0.4 L/s-m2
2-26
-------
Table 2-4 (Continued): Summary of U.S. studies of ozone infiltration published after 2011.
Ambient
Time Ozone
Reference Location Period Population Microenvironment Concentration I/O Ratio Correlation AER
Lai etal. (2015) West Lafayette, NR Test Infiltration 24.02- 0.050-0.099 NR Median
IN chamber 53.5 ppb (10th—90th
percentile) 0.40
(0.15-0.85)
Median
(10th—90th
percentile) 0.98
(0.22-4.84)
Median
(10th—90th
percentile) 0.98
(0.22-4.84)
Median
(10th—90th
percentile) 3.67
(0.74-7.70)
Fagade natural 22.09- 0.18-0.33 NR Median
ventilation 27.68 ppb (10th—90th
percentile) 3.67
(0.74-7.70)
Simple mechanical 38.96- 0.57-0.63 NR
ventilation 39.69 ppb
HVAC 20.11- 0.15-0.43 NR
34.49 ppb
Window open 30.85- 0.23-0.42 NR
51.02 ppb
2-27
-------
Table 2-4 (Continued): Summary of U.S. studies of ozone infiltration published after 2011.
Ambient
Time Ozone
Reference Location Period Population Microenvironment Concentration I/O Ratio Correlation AER
Ben-David and
Waring (2016)
Miami, FL;
Houston, TX;
Phoenix, AZ;
Atlanta, GA;
El Paso, TX;
Los Angeles, CA;
Philadelphia, PA;
Albuquerque, NM;
Seattle, WA;
Boston, MA; Salt
Lake City, UT;
Milwaukee, Wl;
Billings, MT;
Fargo, ND
Data from
2013 or
earlier
Office
buildings
Mechanical
ventilation:
ASHRAE 62.1
Range of
means: 17 ppb
(Seattle, WA)-
35 ppb
(Albuquerque,
NM)
Mean (5th—95th
percentile): 0.121
(0.116, 0.127)
NR
0.39 (for all
locations)
Mechanical mixed
Range of
NR
NR
0.40 (Miami,
with added outdoor
means: 17 ppb
FL)—1.4 (Los
air when
(Seattle, WA)-
Angeles, CA)
thermodynamically
35 ppb
favorable
(Albuquerque,
NM)
Natural ventilation:
Range of
Mean (5th—95th
NR
0.33 (Miami,
ASHRAE 62.1
means: 17 ppb
percentile): 0.107
FL)-0.39 (Los
(Seattle, WA)-
(0.0926, 0.128)
Angeles, CA
35 ppb
and Seattle,
(Albuquerque,
WA)
NM)
Natural ventilation
Range of
NR
0.49 (Seattle,
with added outdoor
means: 17 ppb
WA and Boston,
air when
(Seattle, WA)-
MA)-1.6 (Los
thermodynamically
35 ppb
Angeles, CA)
favorable
(Albuquerque,
NM)
2-28
-------
Table 2-4 (Continued): Summary of U.S. studies of ozone infiltration published after 2011.
Reference
Location
Ambient
Time Ozone
Period Population Microenvironment Concentration
I/O Ratio
Correlation
AER
Singer et al. (2016) Sacramento, CA January-
February
2014
NR
Home
-h daily max
avg per
ventilation
condition:
44-72 ppb
3-h daily max
avg per
ventilation
condition:
37-60 ppb
0.03-0.12
NR
0.03-0.13
NR
Summer:
0.21-0.31,
Fall/Winter:
0.22-0.35
Summer:
0.21-0.31,
Fall/Winter:
0.22-0.35
Ben-David and
Waring (2018)
15 cities:
Miami, FL;
Houston, TX;
Phoenix, AZ;
Memphis, TN;
El Paso, TX; San
Francisco, CA;
Baltimore, MD;
Albuquerque, NM;
Salem, OR;
Chicago, IL;
Boise, ID;
Burlington, VT;
Helena, MT;
Duluth, MN;
Fairbanks, AK
1999-2015
data from
U.S. EPA
Office
buildings
Constant air
volume ventilation
Hourly avg:
17.5 ppb
(Miami, FL)-
34.0 ppb
(Albuquerque,
NM)
0.18-0.49
NR
Variable air volume
ventilation
Hourly avg:
17.5 ppb
(Miami, FL)-
34.0 ppb
(Albuquerque,
NM)
0.19-0.51
NR
Infiltration:
0.08/h
Ventilation:
0.8-3.2/h
Infiltration:
0.08/h
Ventilation:
0.8-3.2/h
AER = air exchange rate, ASHRAE = American Society of Heating, Refrigeration, and Air-Conditioning Engineers, HVAC = heating, ventilation, and air conditioning, I
air concentration, NR = not reported, O = outdoor ozone air concentration, SD = standard deviation.
al/0 was calculated from data provided in Table 1 of Chen et al. (2012) by normalizing to unit ozone ratherthan the I/O provided in the table per 10 ppb of ozone.
indoor ozone
2-29
-------
2.4.3
Relationships between Personal Exposure and Ambient Concentration
The 2013 Ozone ISA (U.S. EPA. 2013) reviewed literature on personal exposure-ambient
concentration (P/A) ratios where an individual is exposed. P/A ratios generally ranged from 0.1-0.3.
Correlations between personal exposure and ambient concentration were reported as 0.05-0.91 over
timescales of hours to days. Higher ratios (0.5-0.9) and correlations (R > 0.64) were reported in the 2013
Ozone ISA for personal exposure measurements when a greater proportion of time was spent outdoors for
studies incorporating timescales up to 14 hours, especially in the vicinity of roadways where ozone
titration by NOx occurs over a small spatial scale.
Results from recent studies of relationships between personal exposure and ambient concentration
(Table 2-5) are somewhat consistent with those described in the 2013 Ozone ISA (U.S. EPA. 2013).
although new studies are limited. P/A ratios calculated using data by Chen et al. (2012) for the National
Morbidity, Mortality, and Air Pollution Study (NMMAPS) study ranged from 0.25 to 0.30 and accounted
for both indoor and outdoor exposure. Jones et al. (2013) noted avg P/A of 0.48 with a 95th percentile of
0.83 and correlation of 0.98. With a substudy of the Moderate and Severe Asthmatics and Their
Environment Study (MASAES), Williams et al. (2012) observed no relationship between ozone exposure
and personal activities for 16 adults with mild to severe asthma (mean age 35.8 years), with a P/A ratio
below 0.1 and correlation between ambient ozone concentration and personal ozone exposure of 0.27 for
24-hour integrated sampling periods. Variation in P/A may be due to variability in infiltration
(Section 2.4.2) or differences in time-activity patterns (Section 2.4.1).
Ozone participates in surface reactions indoors to cause a reduction in concentrations and
exposures. For example, ozone has been shown to participate in surface reactions with VOCs such as
terpenes, a common ingredient of household cleaners and air fresheners (Waring and Wells. 2015;
Springs etal.. 2011). Gall etal. (2011) found that activated carbon and gypsum also reacted with ozone to
reduce indoor concentrations. Human presence has also been shown to lead to reduced ozone
concentrations, because squalene, a natural oil in skin or dust containing skin cells, reacts with ozone
(Rim et al.. 2018; Fadevi et al.. 2013). potentially reducing inhaled ozone concentrations.
2-30
-------
Table 2-5 Studies reporting relationships between personal ozone exposures and ambient ozone
concentrations.
Personal Ambient
Reference Location Time Period Population Concentration Concentration P/A Ratio Correlation
Williams etal. (2012) Detroit, Ml February U.S. EPA mild to Mean (SD): 3.4 ppb Mean (SD): 0.0665 (slope) P/A: 0.27
2008—April severe, (3.6 ppb) 29.7 ppb
2009 nonsmoking adult (15.0 ppb)
asthmatics and
their environment
study panel
(N = 16)
Chen et al. (2012) NMMAPS cities: 1987-2000 All residents NR NR Calculated NR
change in total
ozone exposure
per unit change
in ambient
ozone3
Atlanta, GA 0.25
Birmingham, AL 0.26
Boston, MA 0.30
Buffalo, NY 0.30
Chicago, IL 0.28
Cincinnati, OH 0.27
Corpus Christi, 0.28
TX
Dallas/Ft. Worth, 0.27
TX
2-31
-------
Table 2-5 (Continued): Studies reporting relationships between personal ozone exposures and ambient ozone
concentrations.
Reference
Location
Time Period Population
Personal
Concentration
Ambient
Concentration
P/A Ratio
Correlation
Chen et al. (2012)
(continued)
Denver, CO 1987-2000
(continued)
Los Angeles, CA
All residents
(continued)
NR (continued)
NR (continued)
0.27
0.25
Miami, FL
0.26
Nashville, TN
New York City,
NY
Phoenix, AZ
Seattle, WA
St. Louis, MO
Washington, DC
Worcester, MA
0.27
0.30
0.25
0.30
0.29
0.27
0.27
NR
(continued)
Jones et al. (20131
New York City,
June 1-
Hospital
Avg 8-h daily max
Avg 8-h daily max:
Mean (95th P/A: 0.979
NY
August 31,
admissions for
(95th percentile):
30.67 ppb
percentile): 0.48
2001-2005
respiratory
diagnoses
(N = 34,760
respiratory
hospitalizations)
12.78 ppb
(20.78 ppb)
(0.83)
NMMAPS = National Morbidity, Mortality, and Air Pollution Study; NR = not reported; P/A = f0 + ftpal(a + k) where P = personal exposure to ambient ozone, A = ambient ozone
concentration, f0 = fraction of time spent outdoors, f, = fraction of time spent indoors, p = penetration of ozone indoors (assumed 100% prior to losses), a = air exchange rate, k = loss
rate.
aP/A was calculated from Chen et al. (20121 by inputting the data provided in Table 1 into the equation for P/A presented in the 2013 Ozone ISA (U.S. EPA. 20131.
2-32
-------
2.5
Copollutant Correlations and Potential for Confounding
Confounding among copollutants can occur when the copollutants are correlated with each other
and with the incidence of the health effect being studied (Billionnct et al.. 2012). Potential confounding is
limited to copollutants in this section. However, some parameters are source-based and would not be
expected to correlate with ozone produced by atmospheric chemistry. Other confounders are addressed in
the health effects Appendices, as detailed in the study quality criteria Annex for Appendix 3.
Correlation of ozone with copollutants can lead to inflation of the effect estimates reported in
epidemiologic studies (Goldberg. 2007; Zeger etal.. 2000). Winquist et al. (2014) compared joint effects
calculated with single-pollutant models and with joint effects that account for copollutant correlation.
Consistent with older studies, they found that the effect estimates from single pollutant models including
ozone were inflated beyond the multipollutant joint effect estimates. Evaluation of copollutant correlation
is helpful to understand where there is potential for confounding by evaluating pollutants in the typical
forms used in an epidemiologic analysis.
Average copollutant correlations with 8-hour daily max ozone ranged from -0.17 for 8-hour daily
max CO to 0.26 for 24-hour avg PMio (Figure 2-1). CO provides a surrogate for traffic-related pollution.
While NO2 also is generally a traffic-related pollutant, it can also be a product of the reaction between NO
and O3 in the near-road environment. Outliers can have correlations as high as 0.7 or -0.7, but the bulk of
the data are clustered near zero. During the summer (Figure 2-2). average copollutant correlations are
higher and positive for all pollutants, ranging from 0.17 for 1-hour daily max SO2 to 0.33 for 24-hour avg
PM2 5. Copollutant correlations over the 25th percentile were generally positive for summer, while the
majority of copollutant correlation data were negative during the winter. For rural sites (Figure 2-3).
copollutant correlations only differed for CO compared with urban and suburban sites, where CO for
urban and suburban sites generally have low negative correlations centered around -0.2 and rural sites are
centered around zero. Given that the majority of the copollutant correlation data are low, confounding of
the relationship between ambient ozone exposure and a health effect by exposure to CO, SO2, NO2, PM10,
or PM2 5 is less of a concern for studies of the health effects of ambient ozone exposure compared with
studies of the health effects related to exposure of other criteria air pollutants. When copollutant
correlations are higher during the warm season, greater risk of copollutant confounding exists. However,
summertime correlations remain relatively low for the majority of ambient monitors.
2-33
-------
Year-Round
co -
N = 212
N = 375
802-
N = 289
PM10 -
N = 309
PM2.5 -
N = 514
1 1 1 i 1 1 1 1
-1 -0,8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
Note: Daily metrics based ori the form of the standards were used for all pollutants (ozone: 8-hour daily max, CO: 8-hour daily max,
N02: 1-hour daily max, PM2.5: 24-hour avg, PM10: 24-hour avg, S02: 1-hour daily max), "x" signifies the mean, while the vertical line
within each box represents the median. The box covers the interquartile range, and the whiskers cover the 5th to 95th percentiles of
the data.
Source: AQS database.
Figure 2-1 Year-round Pearson correlations of 8-hour daily max ozone
concentrations with copollutant concentrations measured in air
quality system 2015-2017.
2-34
-------
Winter
co -
N02 -
S02 -
PM10
PM 2.5
~i 1 1 r
N = 189
N = 321
N = 215
•N = 180
•N = 254
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.!
Summer
co
N02
S02
PM 10
PM 2.5
N = 191
N = 346
N = 272
N = 191
N = 327
"I
i 1 1 r
i 1 1 r
CO
N02
S02
PM 10
PM 2.5
CO
N02
S02
PM 10
PM 2.5
N = 193
N = 340
N = 270
N = 187
N = 317
uring
~i 1 1 r
*
-huh
~i 1 r
N = 195
N = 341
N = 265
N = 188
N = 308
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
-1 -0.8 -0.6 -0.4 -0.2 0 ~0~2 04 CL6 CUE
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
Autumn
Note: Daily metrics based on the form of the standards were used for all pollutants (ozone: 8-hour daily max, CO: 8-hour daily max,
N02: 1 -hour daily max, PM2.5: 24-hour avg, PM10: 24-hour avg, S02: 1 -hour daily max). The "x" signifies the mean, while the vertical
line within each box represents the median. The box covers the interquartile range, and the whiskers cover the 5th to 95th
percentiles of the data.
Source: AQS database.
Figure 2-2 Seasonal Pearson correlations of 8-hour daily max ozone
concentrations with copollutant concentrations measured in air
quality system, 2015-2017.
2-35
-------
Rural
CO
N02
S02
PM10
PM2.5
CO
N02
S02
PM10
PM2.5
CO
N02
S02
PM10
PM2.5
N = 36
N = 117
N = 91
N = 86
N= 149
I*
fCD—(-
EH
>4
-1 -0.8 -0.6 -0.4 -0.2 0.2 0.4 0.6 O.I
Suburban
-0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 O.I
Urban
N = 91
N = 108
N = 82
N = 86
N = 135
4
-1 -0.8 -0.6 -0.4 -0.2
H-
I
N = 147
N = 115
N = 136
N = 229
t
•
•
4-
|n*n
0.2 0.4 0.6 0.8
Note: Daily metrics based ori the form of the standards were used for all pollutants (ozone: 8-hour daily max, CO: 8-hour daily max,
N02: 1-hour daily max, PMzst 24-hour avg, PM10: 24-hour avg, S02: 1-hour daily max). The "x" signifies the mean, while the vertical
line within each box represents the median. The box covers the interquartile range, and the whiskers cover the 5th to 95th
percentiles of the data.
Source: AQS database.
Figure 2-3 Year-round Pearson correlations by location setting of 8-hour
daily max ozone concentrations with copollutant concentrations
measured in air quality system, 2015-2017.
2-36
-------
2.6 Interpreting Exposure Measurement Error for Use in
Epidemiology Studies
As described in the 2013 Ozone ISA (U.S. EPA. 2013). exposure measurement error, which
refers to the biases and uncertainties associated with using concentration metrics as surrogates for the
actual exposure of an individual or population (Section 2.2). can be an important contributor to error in
epidemiologic study results. Short-term exposure studies include time-series studies, case-crossover
studies, and panel studies. Time-series studies generally assess the association of the daily health status of
a population of thousands or millions of people over the course of multiple years (i.e., thousands of days)
across an urban area with estimates of human exposure using a short monitoring interval (hours to days).
In these studies, the community-averaged concentration of an air pollutant measured at ambient monitors
is typically used as a surrogate for individual or population ambient exposure. Case-crossover studies use
individuals as their own controls and compare exposures during a health event with exposures before
and/or after the event occurs. Case-crossover studies can be considered a subset of time-series studies,
albeit with different assumptions about baseline risk (Lu and Zeger. 2007). because the conditional
logistic regression function used in case-crossover studies is a form of the log-linear model utilized in
time-series studies. Therefore, the influence of exposure assessment method on effect estimates for
case-crossover studies is not considered separately from time-series studies. Panel studies, which consist
of a relatively small sample (typically tens) of study participants followed over a period of days to
months, have been used to examine the association of specific health effects with short-term exposure to
ambient concentrations of pollutants [e.g., Delfino et al. (1996)1. Long-term exposure studies usually are
longitudinal cohort studies. A longitudinal cohort epidemiologic study, such as the American Cancer
Society (ACS) cohort study, typically involves hundreds or thousands of subjects followed over several
years or decades [e.g., Jerrett et al. (2009)1. Ambient concentrations are generally aggregated overtime
and by community as exposure surrogates.
Exposure measurement error can bias epidemiologic associations between ambient pollutant
concentrations and health outcomes and tends to widen confidence intervals around those estimates
(Sheppard et al.. 2005; Zeger et al.. 2000). Some have developed techniques to correct for exposure
measurement error (Carroll et al.. 2006; Thomas etal.. 1993). The importance of exposure measurement
error depends on the spatial and temporal aspects of the study design. Other factors that could influence
exposure estimates include meteorology, instrument errors, unaccounted localized sources of precursor
species, use of ambient ozone concentration as a surrogate for exposure to ambient ozone, and the
presence of copollutants. This section will summarize information compiled in Section 2.3 about the
different methods used for ozone exposure assessment in epidemiologic studies and related strengths,
limitations, and errors along with how those errors would influence effect estimates for epidemiologic
studies of short-term and long-term ozone exposure (Table 2-6. Table 2-7).
2-37
-------
2.6.1
Short-Term Exposure
In most short-term exposure time-series epidemiologic studies, changes in the incidence of the
health effect are modeled as a function of changes in estimates of ambient exposure, Ea (Davalos et al..
2017V In the absence of indoor ozone sources, Ea can be thought of as a function of the product of
ambient concentration, Ca, and a term encompassing time-weighted averaging of microenvironmental
exposures and infiltration of ozone. This model is presented in the 2013 Ozone ISA (U.S. EPA. 2013).
2.6.1.1 Time-Series Studies
Time-series epidemiologic studies capturing the exposures and health outcomes of a large cohort
frequently use the ambient concentration at a fixed-site monitor or an average of ambient concentrations
across monitors as a surrogate for Ea in a statistical model, as detailed in the 2013 Ozone ISA (U.S. EPA.
2013). This is necessary because measuring personal exposures in studies involving thousands of
participants can be infeasible. Moreover, for time-series epidemiologic studies of short-term exposure, the
temporal variability in concentration is of primary importance to relate to variability in the effect estimate
(Zeger etal.. 2000). Ca can be an acceptable surrogate if the ambient monitor captures the temporal
variability of the true air pollutant exposure. Spatial variability in ozone concentrations across the study
area could attenuate an epidemiologic study's effect estimate if the exposures are not correlated in time
with Ca when the ambient monitoring is used to represent exposure in the statistical model. Differences
between personal exposure to ambient ozone and Ca due to unaccounted for time-activity patterns could
bias time-series studies. If exposure assessment methods that more accurately capture spatial variability in
the concentration distribution over a study area are employed, then the confidence intervals around the
effect estimate may decrease.
A summary of the methods-related studies evaluated in Section 2.3 showed that several methods
can be used in time-series studies, because they can provide an estimate for a geographical domain
containing a large number of individuals and because the data they provide are of a timescale less than
1 month (Table 2-6). These methods include fixed-site monitors, data averaging, LUR, spatiotemporal
modeling, chemical transport models, hybrid models, and microenvironmental models, which can
incorporate information like I/O ratios (Section 2.4.2). Among these methods, fixed-site monitors tend to
be used more frequently for short-term exposure studies. Short-term exposure studies examine how
short-term (i.e., hourly, daily, weekly) changes in health effects are related to short-term changes in
exposure, so accurate characterization of temporal variability by a fixed-site monitor is more important
than accurate characterization of spatial variability. This assumes temporal variability of the exposure is
constant over space.
For short-term exposure assessment methods, use of an exposure surrogate may produce
inaccuracies when temporal variability in the concentration at the location of measurement or model
2-38
-------
prediction differs from temporal variability of the true exposure concentration. As a result, the correlation
between the exposure surrogate and the incidence of the effect would decrease because the additional
scatter in that relationship would flatten the slope of the relationship between the effect and exposure
surrogate, causing the true effect of exposure on incidence of the health outcome to be underestimated
and imprecise. Darrow etal. (2011) examined spatial variability for ozone concentration measurement
timescales (1-hour daily max, 8-hour daily max, commuting hours [7:00 a.m.-10:00 a.m. and
4:00 p.m.-7:00 p.m.], workday hours [8:00 a.m.-7:00 p.m.], and night hours [12:00 a.m.-6:00 a.m.]) and
its impact on effect estimates. Over 60 km, inter-monitor correlation was greater than 0.75 for all but
nighttime ozone measurements, indicating low spatial variability during the day. Risk ratios were greater
than 1 for each case except for nighttime ozone. This finding implies that most ozone concentration
measurements (excluding nighttime) used as a surrogate for exposure to ambient ozone would produce a
small magnitude underestimation of the effect because spatial variability is low over an urban scale. This
analysis did not account for microscale ozone scavenging due to a high NOx gradient near roads. In a
recent study, Shmool et al. (2016) used 24-hour avg temporal and spatiotemporal ozone concentrations in
models of the risk of inpatient hospitalization or outpatient ED visits for asthma in a case-crossover
analysis in New York City. For both outcomes, no difference between models including only a temporal
model of ozone concentration or a spatiotemporal model of ozone concentration could be ascertained,
implying that spatial variability was not important for this time-series study of ambient ozone exposure.
Goldman et al. (2010) simulated the effect of spatial error with and without autocorrelation on risk ratio
and found that the risk was slightly underestimated when spatial error was added (with autocorrelation:
relative risk, RR = 1.0128 per ppm; without autocorrelation: RR = 1.0126 per ppm compared with a base
case RR = 1.0139 with no spatial error added).
Goldman et al. (2012) evaluated the effect of different types of spatial averaging on bias in the
risk ratio and the effect of correlation between measured and "true" ambient concentrations of ozone and
other air pollutant measures. Concentrations were simulated at alternate monitoring locations using the
geostatistical approach described above (Goldman et al.. 2010) for the 20-county Atlanta metropolitan
area for comparison with measurements obtained directly from monitors at those sites. Geostatistically
simulated concentrations using a semivariogram were considered by the authors to be "true" in this study,
and other exposure assessment methods were assumed to have some error. Five different exposure
assessment approaches were tested: (1) using a single fixed-site ambient monitor, (2) averaging the
simulated ambient concentrations across all monitoring sites, (3) performing a population-weighted
average across all monitoring sites, (4) performing an area-weighted average across all monitoring sites,
and (5) population-weighted averaging of the geostatistical simulation. Goldman et al. (2012) observed
that the exposure measurement error was somewhat correlated with both the measured and "true" values,
reflecting both Berkson and classical exposure measurement error components. For the single fixed-site
ambient monitor, the exposure measurement errors had a moderate positive correlation with the measured
value. For the other ambient concentration estimation methods, the exposure measurement errors were
moderately negatively correlated with the "true" value, while having positive but lower magnitude
correlation with the measured value. Additionally, the exposure bias, given by the ratio of the exposure
2-39
-------
measurement error to the measured value, was higher in magnitude at the single fixed-site monitor than
for the spatial averaging techniques for ozone. Hence, compared with other exposure assessment methods,
the effect estimate would likely have greater negative bias (i.e., underestimation of the true effect) with
reduced precision when a single fixed-site monitor is used to measure ozone concentration as a surrogate
for exposure. However, exposure measurement error is likely to cause some bias and decreased precision
for other exposure surrogate methods.
The role of classical and Berkson exposure measurement error on effect estimates has been
explored in recent time-series studies. For example, in a time-series study of ED visits for cardiovascular
disease, Goldman et al. (2011) simulated the effect of classical and Berkson exposure measurement errors
due to spatiotemporal variability among ambient (fixed-site) or simulated outdoor (i.e., an ambient
monitor situated outside the home) air pollutant concentrations over a large urban area, based on the
method used in Goldman et al. (2010). For 8-hour daily max ozone concentrations, the RR per unit mass
was negatively biased in the case of classical exposure measurement error (1.0114 compared to the base
case of 1.0139) and negligibly positively biased in the case of Berkson exposure measurement error
(1.0142). Negative bias means that the true effect was underestimated. The 95% confidence interval range
for RR per ppm of ozone was slightly wider for Berkson exposure measurement error (0.0133) compared
with classical exposure measurement error (0.0109). In addition to the effect of the correlations and ratios
themselves, spatial variation in their values across urban areas also affects time-series epidemiologic
results. The Goldman et al. (2010) and Goldman et al. (2012) findings suggest more Berkson exposure
measurement error in the spatially resolved ambient concentration metrics compared with the fixed-site
ambient monitors, and more classical exposure measurement error for the fixed-site ambient monitor
estimate compared with the other exposure assessment techniques. Hence, more bias would be anticipated
for the effect estimate calculated from the fixed-site ambient monitor, and more uncertainty would be
expected for the effect estimate calculated with the more spatially resolved methods.
A recent study by Strickland et al. (2013) added instrument error to concentrations estimated with
a fixed-site monitor, population-weighted average (PWA), unweighted average (UA), and "true"
population-weighted average (TPWA) concentration obtained from a grid with 1,054 receptor locations.
Berkson exposure measurement error, considered by Strickland et al. (2013) to be the difference in effect
estimates from using the TPWA and a 5-km-resolution simulated ambient concentration surface, was
2.21% per ppb ambient ozone. Positive Berkson exposure measurement error suggested that variability in
the true exposure concentration was uncharacterized but correlated with the ambient concentration.
Median biases for the ambient concentration measurement methods were -16.9% for the fixed-site
monitor, -1.6% for PWA, and -2.6% for UA. These biases reflected errors in capturing all components of
the ambient concentration (Berkson-like) and the imprecision in the ambient concentration estimate
(classical-like). Differences in the magnitude of exposure concentration estimates are not likely to cause
substantial bias, but they tend to widen confidence intervals and thus reduce the precision of the effect
estimate (Zeger et al.. 2000). The more spatially variable air pollutants studied in Goldman et al. (2012)
also had more bias in their effect estimates. This occurred across exposure assessment methods but was
2-40
-------
more pronounced for the fixed-site ambient monitoring data. It is important to note that the Goldman et
al. (2010). Goldman et al. (2011). Goldman et al. (2012). and Strickland et al. (2013) studies were
performed only in Atlanta, GA. These simulation studies are informative, but similar simulation studies in
additional cities would aid generalization of these results.
Introducing errors in the time-series of data rather than across space had a larger impact on effect
estimates. For example, Samoli et al. (2014) compared effects in four cities (Athens, Greece; London,
U.K.; Milano, Italy; Zurich, Switzerland) estimated using a complete daily time-series with effects where
a time-series with only 1 day in 6 was systematically included. For all cities and for results pooled across
cities, the percentage change in total mortality corresponding to a 10 (ig/m3 increase in ozone
concentration decreased from positive to negative (of equal or lesser magnitude) with larger confidence
intervals when the l-in-6 day data were used in lieu of the full data set.
Data for time-activity patterns and avoidance behaviors are often omitted from exposure
assessment studies. This omission has the potential to add negative bias to and decrease precision of the
effect estimate. Bias would result from a reduction in correlation between the exposure surrogate and the
incidence of the health effect, while decreased precision could occur when the lack of time-activity data
prevents characterization of the true variability in exposure. These errors can potentially occur for
fixed-site monitors, data averaging, LUR, spatiotemporal models, CTMs, and hybrid approaches
(Table 2-7). Jones et al. (2013) compared respiratory and asthma hospitalization estimates obtained from
use of a fixed-site monitor with those obtained from a microenvironmental model to ascertain the impact
of time-activity data on the results. Little differences between the mean and confidence interval of the
hazard ratios were observed for the entire population for respiratory hospitalizations (fixed-site: 1.013
confidence interval, CI 0.999-1.028; mean error, ME: 1.013, CI 0.998-1.029) and asthma
hospitalizations (fixed-site: 1.029, CI 1.010-1.047; ME: 1.029, CI 1.009-1.049). However, differences in
hazard ratios were noted for the 5-14 year (fixed-site: 1.056, ME: 1.013), 15-24 year (fixed-site: 1.051,
ME: 1.013), 25-64 year (fixed-site: 1.021, ME: 1.013), and >65 year (fixed-site: 0.993, ME: 1.015) age
groups. The >65-years-old group, which spends the most time indoors (Table 2-1). was the only group for
which the effect was underestimated by the fixed-site monitor concentrations.
2.6.1.2 Panel Studies
The 2013 Ozone ISA (U.S. EPA. 2013) did not comment on potential effects of exposure
measurement errors on results of panel studies. Panel studies representing short-term exposure to ozone
typically use active or passive microenvironmental monitors to represent exposure (Table 2-7). A strength
of the measurement methods is the ability to have a representation of the exposure at the location of the
individuals being studied, while a limitation is greater sensitivity to instrument errors. Active monitors are
subject to interference from humidity and copollutants, while passive monitors have diffusion-related
losses when ozone reacts with the instrument manifold. These instrument errors tend to be small but
2-41
-------
negative (i.e., the instrument reports a lower concentration than the true concentration). Because panel
studies take measurements at the exposed individual sites, correlation between change in effect with
change in exposure is less important than estimating the relationship between the ozone exposure and the
occurrence of the health effect. As a result, instrument error from use of microenvironmental monitors
could add a small amount of positive bias to effect estimates.
2.6.2 Long-Term Exposure
In most epidemiologic studies evaluating long-term exposures, the health effect endpoint is
modeled as a function of long-term average ambient ozone exposure, Ea (U.S. EPA. 2013V For cohort
epidemiologic studies of long-term exposure to ambient ozone, where the difference in the magnitude of
the concentration is of most interest, Ca is used as a surrogate for ambient exposure. Uncertainties in
time-activity patterns of exposed individuals and surface losses of ozone can reduce precision in the effect
estimates. Spatial variability in ozone concentrations across the study area could lead to bias in the effect
estimate if Ca is not representative of Ea. There are limited data regarding whether Ca is a biased exposure
surrogate in the near-road environment for epidemiologic studies of long-term exposure. However, ozone
is known to be fairly spatially homogeneous at the urban scale (Appendix 1). Spatial variability may be
greater in some locations, such as near roads where ozone scavenging occurs due to NOx chemistry
(Kimbrough et al.. 2017). Scavenging would result in ozone concentrations that are lower near the road
than at a fixed-site monitor located away from the road. It would therefore be anticipated that effects
would be underestimated by using fixed-site monitoring data to describe exposures for a population living
or working near a road or traveling on a road. Biases in effect estimates would be small but could occur in
either direction.
A summary of the methods-related studies evaluated in Section 2.3 showed that several methods
have the potential to be used in long-term exposure studies because they can provide an estimate for a
geographical domain containing a large number of individuals and because the data they provide are of a
timescale greater than 1 month (Table 2-7). These methods include fixed-site monitors, data averaging,
IDW, kriging, LUR, spatiotemporal modeling, CTMs, hybrid models, and microenvironmental models.
IDW, kriging, LUR, spatiotemporal modeling, CTMs, and hybrid models are spatial concentration
prediction models listed in order of increasing sophistication (i.e., producing increasing model fit;
Section 2.3.2). In recent studies, hybrid models incorporating CTM output with satellite data have
produced simulated ambient ozone concentration surfaces with low spatial model error at a national scale
[e.g., Robichaud and Menard (2014)1. Higher resolution exposure assessment models are intended to
minimize bias and uncertainty in the effect estimate due to spatial variability. Microenvironmental models
incorporate time-activity patterns, including indoor and outdoor microenvironments along with
information like AER, which influences I/O ratios (Section 2.4.2). with high spatial resolution of ambient
ozone concentration predictions to estimate ambient ozone exposures among the population.
2-42
-------
Nonspatial sources of exposure measurement error can also influence the effect estimate
produced from modeled exposure surrogates. Model misspecification, where an exposure assessment
model is not fit with the correct predictive variables, can also lead to bias in the effect estimate in either
direction. Omission of time-activity data in the spatiotemporal exposure assessment models can decrease
precision in the effect estimate because variability in the exposure is curtailed without time-activity
patterns. Likewise, omission of time-activity data can result in negative bias when it causes the spatial
correlation between the exposure estimate and the effect to decrease. Dionisio et al. (2014) recently
compared effect estimates derived from a microenvironmental model of exposure with effect estimates
derived from a CTM. Using personal exposure as a reference, the effect estimate was considered to be
negatively biased when the CTM alone was used. Omission of time-activity data was responsible for
87.6% [RMSE (SD) = -0.85 ppb (0.015 ppb)] of the total bias in the effect estimate.
As described in the 2013 Ozone ISA (U.S. EPA. 2013). spatial variability is typically low at the
urban scale, with the exception of near-road areas. Bias related to spatial variability is typically
anticipated to be low except where ozone scavenging takes place. Uncharacterized ambient ozone
scavenging near a road would mean that a population living or working near roads would have
overestimated exposure and a negatively biased effect estimate. When LUR or spatiotemporal models are
applied, then bias can occur in either direction if the model is applied in a location different from where it
was fit (Table 2-7). Since the 2013 Ozone ISA, Punger and West (2013) estimated exposure using a CTM
(CMAQ 4.7.1) and compared the effect estimates using a coarse (36-km) and medium (12-km) grid. Use
of a 36-km grid led to a 12% higher effect estimate compared with the medium grid. Dionisio et al.
(2014) recently evaluated bias by comparing an effect estimate considering exposure derived from a
dispersion model (AERMOD) with an effect estimate considering exposure derived from a fixed-site
monitor. Omission of spatial exposure measurement error accounted for 12.4% [RMSE (standard
deviation; SD) = -0.12 ppb (0.093 ppb)] of the total bias in the effect estimate, and biases were negative.
The SD was a larger proportion of bias in the effect estimate for spatial exposure measurement error
compared with bias in the effect estimate related to omission of time-activity data in the exposure
assessment. Lopiano et al. (2011) compared effect estimates computed with exposures from three
variations each of kriging and parametric bootstrapping. Several cases were evaluated for the set of
models. The health effect estimate was slightly overestimated (by <2.5%) for the cases where exposure
assessment occurred at the location of cases and where some case locations were omitted from the model
with prediction of the effect's variability close to the true 95% confidence intervals (within ±2.5%), so
negligibly small positive biases in the effect estimates were observed with good coverage. This suggests
that the spatial variability for ozone may not be a large source of error.
2-43
-------
Table 2-6 Summary of the influence of exposure error on epidemiologic study
outcomes.
Potential Influence on Effect
Estimates
Sources of Exposure
Measurement Errors
<
(5
+-»
GJ
Q
0)
o
(5
9.e
0) +-
£
Ji
g)
*
0)
i_
O)
0)
a:
o
z>
¦O
c
cu
-o
o
o
Q.
E
o
cu
Q.
r
o
Q.
(5
o
-o
o
= — -o
Eo
o
n
><
x
o
E
c
o
">
C
o _
o o
Time Series
Studies
Errors Mostly
Due to
Reduced
Correlation
between
Surrogate
and True
Exposure
Long-Term
Studies
Potential
Errors Mostly
Due to
Differences
between
Surrogate
and True
Exposure
Omission of
time-activity data
X
X
X
X
X
X X
X
((-))
((-))
Near road scavenging
X
X
X
X
X
((-))
((-))
Poorly characterized
spatiotemporal
variability
X
X
((-))
((-/+))
Over-smoothing
X
X
((-))
Exposure model
misspecification
X
X
X
-/+
Spatial misalignment
X
X
X
-/+
Distributions of input
data differ from true
population
distributions
X
(0)
((-/+))
"X" = where a source of exposure measurement error can occur for a given exposure methodology;
"+" = bias away from the null;"(())" = decreased precision.
: bias towards the null;
2-44
-------
Table 2-7 Summary of exposure estimation methods, their typical use in ozone epidemiologic studies, and
related errors and uncertainties.
Exposure
Concentration
Assignment
Method
Description
Epidemiologic
Application
Strengths
Limitations
Exposure Measurement
Errors
Influence on Effect
Estimates in
Epidemiologic Study
Results
Measurement methods
Fixed-site monitor
[Section 2.3.1.1
Table 2-8; U.S. EPA
(2013)1
An FRM or FEM
monitor located at a
fixed location to
measure ambient
ozone concentration
by
Short-term
exposure studies:
surrogate for
ambient ozone
exposure
concentration of a
Ambient ozone
concentration
measurements
undergo rigorous
quality assurance
chemiluminescence population within a
or UV absorption
city
Measurements of
ambient ozone
concentration
made at a fixed
location may differ
from an exposed
individual's true
exposure
concentration, and
no spatial variation
is assumed
Correlation between
outdoor ozone
concentrations proximal to
the receptors and ambient
ozone concentration
measurements decreases
with increasing distance
from the monitor,
especially in cities with a
lot of solar radiation and
roadways, where ozone
production is high but
scavenging occurs near
roads (in some cities,
correlations >0.80 over
distances of 50 km)
Potential for
simultaneous
decreased precision
and negative bias in the
effect estimate,
because decreased
correlation between the
exposure surrogate and
effect drives the slope
towards zero
Omission of time-activity
data
Negative bias and
decreased precision in
the effect estimate
2-45
-------
Table 2-7. (Continued): Summary of exposure estimation methods, their typical use in ozone epidemiologic
studies, and related errors and uncertainties.
Exposure
Concentration
Assignment
Method
Description
Epidemiologic
Application
Strengths
Limitations
Exposure Measurement
Errors
Influence on Effect
Estimates in
Epidemiologic Study
Results
Fixed-site monitor
TSection 2.3.1.1
Table 2-8; U.S. EPA
(2013)1 (continued)
An FRM or FEM
monitor located at a
fixed location to
measure ambient
ozone concentration
by
chemiluminescence
or UV absorption
(continued)
Long-term
exposure studies:
surrogate for
ambient ozone
exposure
concentration to
compare
populations within
a city or among
multiple cities
Ambient ozone
concentration
measurements
undergo rigorous
quality assurance
(continued)
Measurements of
ambient ozone
concentration
made at a fixed
location may differ
from an exposed
individual's true
exposure
concentration, and
no spatial variation
is assumed
(continued)
Ambient ozone
concentration at a receptor
location is higher or lower
than the ambient ozone
concentration measured at
the monitor
Potential for bias in the
effect estimate in either
direction, but likely
small in magnitude
Localized ozone loss
Potential for negative
processes near roads are
bias in the effect
not captured
estimate
Spatial variability of ozone
Potential for decreased
concentration is not
precision in the effect
characterized
estimate
Omission of time-activity
Negative bias and
data
decreased precision in
the effect estimate
Microenvironmental
exposure monitor
(non-FRM or FEM)
(Section 2.3.1.2
Table 2-9)
Typically, a
miniaturized UV
absorption sampler
for ozone, where air
is pulled through a
pump; may be an
FEM
Panel studies:
ozone exposure
(e.g., personal or
residential
samples) within a
geographic area
Ozone
concentrations may
be obtained at the
site of the exposed
person and
therefore
automatically
account for
time-activity
patterns; spatial
variability is better
captured by
deploying monitors
with higher spatial
density
Non-FEM UV
absorption
instruments subject
to interference from
humidity, mercury,
and VOCs
Instrument errors more
typically lead to positive
artifacts from interferences
Instrument errors are
typically small but
positive and so have
the potential to add
negative bias to the
effect estimate
2-46
-------
Table 2-7. (Continued): Summary of exposure estimation methods, their typical use in ozone epidemiologic
studies, and related errors and uncertainties.
Exposure
Concentration
Assignment
Method
Description
Epidemiologic
Application
Strengths
Limitations
Exposure Measurement
Errors
Influence on Effect
Estimates in
Epidemiologic Study
Results
Passive personal
exposure monitor
(Section 2.3.1.2
Table 2-9)
Ozone is captured
on a nitrite-treated
substrate via
passive exposure for
a time period to
measure a personal
or area sample;
oxidation by ozone
converts the nitrite to
nitrate, which is
analyzed by ion
chromatography
Panel studies:
ambient ozone
exposure within a
city or among
multiple cities
Ozone
concentrations are
obtained at the site
of the exposed
person and
therefore
automatically
account for
time-activity
patterns; low cost
Long duration
integrated sampling
time (e.g., 7 days)
does not allow for
time-series
analysis;
diffusion-related
losses to the
passive sampler
hardware
Diffusion-related losses to
the passive sampler
hardware have the
potential to bias the
concentration estimation
based both on reduced
ozone detection and
overestimation of flux to
the sampling substrate
Instrument errors are
typically small but
negative and so have
the potential to add
positive bias to the
effect estimate
2-47
-------
Table 2-7. (Continued): Summary of exposure estimation methods, their typical use in ozone epidemiologic
studies, and related errors and uncertainties.
Exposure
Concentration
Assignment
Method
Description
Epidemiologic
Application
Strengths
Limitations
Exposure Measurement
Errors
Influence on Effect
Estimates in
Epidemiologic Study
Results
Modeling methods
Data averaging
(Section 2.3.2.1
Table 2-10)
Averaging across
multiple monitors
during the same
time window and
within a
geographical area,
such as a city or
county, typically
using fixed-site
monitoring data
Short-term
exposure studies:
surrogate for
ambient ozone
exposure
concentration of a
population within a
city
Spatial averaging
(area averaging,
population-weighted
averaging), typically
using fixed-site
monitoring data
Long-term
exposure studies:
surrogate for
ambient ozone
exposure
concentration,
usually within a
city or geographic
region
Ambient ozone
concentration
measurements
undergo rigorous
quality assurance;
averaging scheme
designed for
population or trend
of interest; simple to
implement
Measurements of
ambient ozone
concentration
made at a fixed
location may differ
from an exposed
individual's true
exposure
concentration, and
either no spatial
variation is
assumed or spatial
variation is
assumed to be well
represented by the
averaging scheme;
errors in average
concentration can
be caused by one
errant monitor; in
areas where
different monitors
peak on different
days, this method
will mute overall
temporal variation
Correlation between
outdoor ozone
concentrations proximal to
the receptors and ambient
ozone concentration
measurement at a centrally
located fixed-site monitor
decreases with increasing
distance from the monitor,
especially in cities with a
lot of solar radiation and
roadways, where ozone
production is high but
scavenging occurs near
roads (in some cities,
correlations >0.80 over
distances of 50 km)
Low correlation
potentially leads
simultaneously to
decreased precision
and to negative bias in
the effect estimate due
to decreased
correlation between the
exposure surrogate and
effect
Omission of time-activity
Negative bias and
data
decreased precision in
the effect estimate
Localized ozone loss
Potential for negative
processes near roads are
bias in the effect
not captured
estimate
Assumption of constant
Potential for decreased
ozone concentration within
precision in the effect
some geographic area
estimate
Omission of time-activity
Negative bias and
data
decreased precision in
the effect estimate
2-48
-------
Table 2-7. (Continued): Summary of exposure estimation methods, their typical use in ozone epidemiologic
studies, and related errors and uncertainties.
Exposure
Concentration
Assignment
Method
Description
Epidemiologic
Application
Strengths
Limitations
Exposure Measurement
Errors
Influence on Effect
Estimates in
Epidemiologic Study
Results
Inverse-distance
weighting
(Section 2.3.2.1
Table 2-10)
Measured ambient
ozone
concentrations are
interpolated to
estimate ambient
ozone concentration
surfaces across
regions; IDW uses
an inverse function
of distance to
monitors
Long-term
exposure studies:
surrogate for
ambient ozone
exposure
concentration,
usually within a
city or geographic
region
High spatial
resolution
Does not account
for atmospheric
chemistry or
meteorology;
over-smoothing is
possible based on
the smoothing
function between
monitors
Ozone concentration is
overly smoothed
Potential for negative
bias with decreased
precision in the effect
estimate
Omission of time-activity
data
Negative bias and
decreased precision in
the effect estimate
Localized ozone loss
processes near roads are
not captured
Potential for negative
bias in the effect
estimate
2-49
-------
Table 2-7. (Continued): Summary of exposure estimation methods, their typical use in ozone epidemiologic
studies, and related errors and uncertainties.
Exposure
Concentration
Assignment
Method
Description
Epidemiologic
Application
Strengths
Limitations
Exposure Measurement
Errors
Influence on Effect
Estimates in
Epidemiologic Study
Results
Kriging
(Section 2.3.2.1
Table 2-10)
Measured ambient
ozone
concentrations are
interpolated to
estimate ambient
ozone concentration
surfaces across
regions
Long-term
exposure studies:
surrogate for
ambient ozone
exposure
concentration,
usually within a
city or geographic
region
High spatial
resolution
Does not account
for atmospheric
chemistry or
meteorology;
over-smoothing is
possible based on
smoothing function
between monitors
Ozone concentration is
overly smoothed
Potential for negative
bias with decreased
precision in the effect
estimate
Omission of time-activity
data
Negative bias and
decreased precision in
the effect estimate
Localized ozone loss
processes near roads are
not captured
Potential for negative
bias in the effect
estimate
Land use regression
(Section 2.3.2.2
Table 2-11)
Measured ambient
Short-term and
High spatial
Does not account
ozone
long-term
resolution
for precursor
concentrations are
exposure studies:
emission rates,
regressed on local
surrogate for
dispersion, or
variables (e.g., land
ambient ozone
atmospheric
use factors); the
exposure
chemistry and may
resulting model is
concentration,
account for
used to estimate
usually across a
meteorology only in
ambient ozone
city but sometimes
terms of wind
concentrations at
among multiple
speed and wind
specific locations
cities
direction,
depending on
model formulation;
has limited
generalizability to
other locations;
uncertainties are
highest where
training monitors
are sparse
Potential for model
misspecification
Model is applied to a
location different from
where it was fit
Short-term exposure
studies: potential for
negative bias with
decreased precision in
the effect estimate
Long-term exposure
studies: potential bias
in the effect estimate in
either direction
Short-term exposure
studies: potential for
negative bias with
decreased precision in
the effect estimate
Long-term exposure
studies: potential bias
in the effect estimate in
either direction
Omission of time-activity
data
Negative bias and
decreased precision in
the effect estimate
2-50
-------
Table 2-7. (Continued): Summary of exposure estimation methods, their typical use in ozone epidemiologic
studies, and related errors and uncertainties.
Exposure Influence on Effect
Concentration Estimates in
Assignment Epidemiologic Exposure Measurement Epidemiologic Study
Method Description Application Strengths Limitations Errors Results
Spatiotemporal
model
(Section 2.3.2.2
Table 2-11)
Measured ambient
ozone
concentrations are
modeled by a spatial
average, spatially
varying covariates,
and a
spatiotemporal
residual; the
resulting model is
used to estimate
ambient ozone
concentrations at
specific locations
Short-term and
long-term
exposure studies:
surrogate for
ambient ozone
exposure
concentration,
usually across a
city but sometimes
among multiple
cities
High spatial
resolution; flexible
modeling framework
allows for
minimization of
errors
Does not account
for precursor
emission rates,
dispersion, or
atmospheric
chemistry and may
account for
meteorology only in
terms of wind
speed and wind
direction,
depending on
model formulation;
has limited
generalizability to
other locations;
uncertainties are
highest where
training monitors
are sparse
Potential for model
misspecification
Model is applied to a
location different from
where it was fit
Short-term exposure
studies: potential for
negative bias with
decreased precision in
the effect estimate
Long-term exposure
studies: potential bias
in the effect estimate in
either direction
Short-term exposure
studies: potential for
negative bias with
decreased precision in
the effect estimate
Long-term exposure
studies: potential bias
in the effect estimate in
either direction
Omission of time-activity
data
Negative bias and
decreased precision in
the effect estimate
2-51
-------
Table 2-7. (Continued): Summary of exposure estimation methods, their typical use in ozone epidemiologic
studies, and related errors and uncertainties.
Exposure
Concentration
Assignment
Method
Description
Epidemiologic
Application
Strengths
Limitations
Exposure Measurement
Errors
Influence on Effect
Estimates in
Epidemiologic Study
Results
Chemical transport
model
(Section 2.3.2.3
Table 2-12)
Grid-based ambient
ozone
concentrations are
estimated from
precursor emissions,
meteorology, and
atmospheric
chemistry and
physics
Short-term and
long-term
exposure studies:
surrogate for
ambient ozone
exposure
concentration,
sometimes within
a city but more
typically across a
larger region
Accounting for
precursor emission
rates, mixing height,
atmospheric
stability,
meteorology,
atmospheric
chemistry, and
complex terrain
Limited grid cell
resolution (i.e., grid
cell length scale is
typically 4-36 km);
spatial smoothing
of local ozone
precursor
emissions
Localized ozone loss
processes near roads are
not captured because grid
cell scale is too large
Short-term exposure
studies: potential for
negative bias with
decreased precision in
the effect estimate
Long-term exposure
studies: potential for
negative bias in the
effect estimate
Omission of time-activity
data
Negative bias and
decreased precision in
the effect estimate
2-52
-------
Table 2-7. (Continued): Summary of exposure estimation methods, their typical use in ozone epidemiologic
studies, and related errors and uncertainties.
Exposure
Concentration
Assignment
Method
Description
Epidemiologic
Application
Strengths
Limitations
Exposure Measurement
Errors
Influence on Effect
Estimates in
Epidemiologic Study
Results
Hybrid approaches
(Section 2.3.2.4
Table 2-13)
Grid-based ambient
ozone
concentrations are
estimated from
precursor emissions,
meteorology, and
atmospheric
chemistry and
physics with bias
correction based on
monitoring data
Short-term and
long-term
exposure studies:
surrogate for
ambient ozone
exposure
concentration,
sometimes within
a city but more
typically across a
larger region
Accounting for
ozone precursor
emission rates,
mixing height,
atmospheric
stability,
meteorology,
atmospheric
chemistry, and
complex terrain;
bias correction
improves model
results, particularly
where biases are
large; fusing model
results with
monitoring, satellite,
and chemical
transport model
values helps to
minimize exposure
measurement errors
The modeling
process can be
resource intensive;
spatial smoothing
of local precursor
emissions sources;
has limited
generalizability to
other locations;
uncertainties are
highest where
training monitors
are sparse
Potential for model
misspecification
Model is applied to a
location different from
where it was fit
Short-term exposure
studies: potential for
negative bias with
decreased precision in
the effect estimate
Long-term exposure
studies: potential bias
in the effect estimate in
either direction
Short-term exposure
studies: potential for
negative bias with
decreased precision in
the effect estimate
Long-term exposure
studies: potential bias
in the effect estimate in
either direction
Omission of time-activity
data
Negative bias and
decreased precision in
the effect estimate
2-53
-------
Table 2-7. (Continued): Summary of exposure estimation methods, their typical use in ozone epidemiologic
studies, and related errors and uncertainties.
Exposure
Concentration
Assignment
Method
Description
Epidemiologic
Application
Strengths
Limitations
Exposure Measurement
Errors
Influence on Effect
Estimates in
Epidemiologic Study
Results
Microenvironmental Estimates
modeling
(e.g., APEX,
SHEDS
[Section 2.3.2.5
Table 2-141)
distributions of
microenvironmental
ozone
concentrations,
exposures, and
doses for
populations
(e.g., census tracts)
based on air quality
data, demographic
variables, and
activity patterns
Short-term and
long-term
exposure studies
Accounts for
variability of ozone
exposures across
large populations;
accounts for
different
concentrations in
different
microenvironments;
accounts for
location-activity
information
Models simulate
individuals and
their exposures;
they do not model
actual individuals
but simulated
representative
individuals based
on the population
being modeled
The modeled distributions
of ambient ozone
concentration,
indoonoutdoor pollutant
ratios, and time-activity
patterns may differ from
the true distributions,
depending on model inputs
Potential for decreased
precision in the effect
estimate
APEX = Air Pollutants Exposure model; BC = black carbon; CPC = condensation particle counter; FEM :
IDW = inverse-distance weighting; SHEDS = Stochastic Human Exposure and Dose Simulation.
Federal Equivalent Method; FRM = Federal Reference Method;
2-54
-------
2.7 Conclusions
The 2013 Ozone ISA (U.S. EPA. 2013) focused on personal exposure to ozone. Recently
published data on I/O ratios (Section 2.4.2). P/A ratios (Section 2.4.3). and associated correlations
(Section 2.5) are similar to those presented in the 2013 Ozone ISA. Likewise, ambient ozone
characteristics, including its spatial distribution over urban scales, high variability near roads, and
seasonal and diurnal variation, have not changed substantially since the 2013 Ozone ISA. More
information is now available on losses of ozone at surfaces, and those studies support the published ratios.
Personal measurements of ozone exposure (Section 2.3.1.2) from active or passive monitors are subject to
interference from humidity, mercury, and VOCs (active monitors) or from deposition to the sampling
manifold (passive monitors). This can lead to positively biased concentrations and negatively biased
effect estimates.
Fixed-site monitors are still widely in use as ozone exposure surrogates ITJ.S. EPA (2013).
Section 2.3.1.11. given the low spatial variability typical of ozone in many places. Biases tend to be small
in magnitude for this reason. Localized atmospheric chemistry nearNOx sources (i.e., ozone sinks) such
as highways may result in overestimation of exposure when ozone concentration monitored away from
the highway is used in an epidemiologic analysis. Data averaging techniques (Section 2.3.1.2). including
IDW and kriging, provide spatial interpolation where monitors are sparse, but they can produce a less
precise exposure estimate compared with hybrid or spatiotemporal models.
The largest development since the 2013 Ozone ISA (U.S. EPA. 2013) is the expanded availability
of models to predict ozone concentrations as surrogates for exposure assessment, thereby addressing a
key uncertainty for modeling ozone exposure where measurements are not available, such as in rural
areas. Both LUR and spatiotemporal modeling (Section 2.3.2.2) can be subject to model misspecification,
where the model is not fit with the optimal set of variables. Larger exposure measurement errors can
occur if the models are fit to one location and then applied in a different location. CTMs have greatly
expanded in usage and in number of available models (Section 2.3.2.3). Misspecification can occur
through inadequate characterization of emissions, meteorology, and chemistry. High magnitude errors
most typically happen around low and high ozone modeled outputs. Spatiotemporal models sometimes
become the framework for incorporating CTMs, satellite data (Section 2.3.2.4). and fixed-site monitoring
data (Section 2.3.2.1) into a single hybrid model. By combining so many sources of data, overfitting may
be a larger concern for the hybrid model than for other exposure estimation models.
For epidemiologic studies of short-term exposure to ozone, the effect estimates potentially have
decreased precision and negative bias if the correlation between the exposure surrogate and the health
effect is lower than the correlation between the true exposure and the health effect (Table 2-7). Negative
bias with decreased precision may occur for fixed-site monitors or any of the spatial interpolation
methods. Attenuation of the effect estimate may also occur for LUR, spatiotemporal, and hybrid models
2-55
-------
when they are misspecified or fit to a different geographical area than where they are applied. Fixed-site
monitors and CTMs may also produce negatively biased effect estimates if individuals are exposed to
localized areas of low ozone, such as near roads where ambient ozone is scavenged by NOx so that the
monitor or modeled estimate of ozone exposure is higher than the true exposure. In these cases, use of
the exposure surrogate generally leads to an underestimation of the association between short-term
exposure to ambient ozone and the health effect with reduced precision. Although the magnitude of the
association between ambient ozone and the health effect is uncertain, the evidence indicates that the
true effect is typically larger than the effect estimate in these cases.
Panel studies tend to use microenvironmental or personal monitors to measure exposure at the
locations of individuals in a study. The small instrument errors observed for active and passive monitors
can lead to small but positive biases in the effect estimates for short-term exposures to ozone.
For epidemiologic studies of long-term exposure to ozone, when concentrations measured at
fixed-site monitors are used as exposure surrogates, effect estimates have the potential to be biased in
either direction. However, it is more common that these methods contribute to an underestimation of the
effect, and the magnitude of bias is likely small given that ozone concentration does not vary over space
as much as other criteria pollutants, such as NOx or SO2 (Table 2-7). Localized ozone scavenging by NOx
creates potential for negative bias in the effect estimate, if people are exposed on or near a roadway with
traffic but have their exposures estimated by concentrations measured at a monitor positioned away from
that location. The assumption of a constant ozone concentration within some radius of the monitor or
model receptor location also may reduce precision for fixed-site monitors and CTMs. The coarse
horizontal grid resolution in CTMs can reduce the spatial heterogeneity of the true exposure. Smoothing
may also lead to reduced precision for the data averaging schemes presented. Model misspecification and
model fit in a location apart from the field study in LUR, spatiotemporal models, and hybrid models may
cause bias in either direction. Depending on the model and scenario being modeled, the true effect of
long-term exposure to ambient ozone may be underestimated or overestimated by the model It is much
more common for the effect estimate to be underestimated, and the bias is typically small in magnitude.
For most exposure estimation methods, omission of time-activity data may lead to negative bias
and decreased precision of the effect estimates, because exposure variability is largely uncharacterized
(Section 2.4.1). That was demonstrated by the comparison of exposure and effect estimates based on
monitored concentrations with exposure and health estimates based on microenvironmental models
(Section 2.3.2.5) that do use time-activity data from sample populations. Estimating exposure without
accounting for time-activity data may result in underestimation of the true effect and reduced
precision. Although the magnitude of the association between ozone and the health effect is uncertain,
the evidence suggests that the true effect of ambient ozone exposure is larger than the effect estimate
when time-activity data are not considered in the analysis.
2-56
-------
2.8 Evidence Inventories—Data Tables to Provide Supporting
Information
Validation measures are used to evaluate concentrations measured or modeled, as described in
Section 2.3. Method performance measures are listed below and are included in Table 2-8 through
Table 2-14:
Unpaired predicted-to-observed
peak ozone ratio (AUP)
Mean bias (MB)
Mean error (ME)
Mean-squared error (MSE)
Root-mean-squared error (RMSE)
p — n
1 i,peak wi,peak
pf
i,peak
N
i=1
N
- °
-------
Mean fractional error (MFE)
(0 to +200%)
N
2 V I Pt- Oi
JvZJPi + 0;
Coefficient of determination (R2) j^N ^q _ pjj2
Sf=1(o;-o) HUiPi-p)
Mean absolute error (MAE) ~~ Oil
N
Index of agreement (IOA) Z"=i(^i ~~ °i)2
Normalized bias (NB)
ZUUPi ~ o\ + \ot - o\y
Ot
Fractional bias (FB) 1 v-1™ (Pi ~ °i)
i
nL1=1~.
i=11/2 (Pi - Ot)
Fractional error (FE) |P£ — 0£|
1/2 (f; - Ot)
Normalized error (NE) |P; - 0
x 100%
Sf=1o;
Normalized gross error (NGE) f\Pt - 0,
, x 100%
"i
Mean normalized gross error 1 v™ /|P, — 0, |\
(MNGE) ivZ4^^)xl00%
Critical success index (CSI) { b \
—rm xl00%
\a + b + d)
False alarm ratio (FAR) t b \
v ' r x 100%
\a + bj
Unpaired peak accuracy (UPA) Ppeak Opeak
Opeak
Pi and Oi are prediction and observation at the /th monitoring site,
respectively; N is the number of monitoring sites; a, b, d and quadrants
of values as defined in Kanq et al. (2005).
2-58
-------
Table 2-8 Studies informing assessment of exposure measurement error when concentrations measured by
fixed-site monitors are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Monitor
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Blanchard et al.
Observed,
L: In and near
Observed
Long-term
Mean summer
Observed
Spatial
NR
(2011)
fixed-site
the Atlanta
fixed-site
exposure
quartiles of peak
data collected
variability is
ambient
metropolitan
ambient
8-h ozone in
directly;
limited to
monitors only
area;
monitors only,
ppb, 44.85,
speciation
monitoring
from
T: Between
protocols of
55.73, 64.93,
collected;
locations
SEARCH,
the yr 1999
observed data
80.23
several years
U.S. EPA
and 2007;
were referenced
of data
PAMS, U.S.
P: Entire
in previous
collected
EPA STN,
population
publications
IMPROVE
considered
monitoring
networks
Hackbarth et al.
Daily 8-h max
L: California;
Observed data
Short-term
Ozone daily avg
Observed
Sensitivity
Ozone daily
(2011)
ozone within
T: 2005-
used in
exposure
in California
data used
analysis of
avg across
20 miles of
2007;
inverse-distance
between 2005
directed with
exposure
California =
the zip-code
P: Those with
weighting
and 2007
ER data
method not
39.9 ppb,
centroid
ER visits for
validated from
because
explored
standard
weighted by
cardio, resp,
U.S. EPA
37.1-73.8 ppb
(e.g., buffer
error = 0.225
inverse
asthma
size of nearest
ppb
distance from neighbor)
U.S. EPA's
fixed-site
monitors
2-59
-------
Table 2-8 (Continued): Studies informing assessment of exposure measurement error when concentrations
measured by fixed-site monitors are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Monitor
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Joseph et al.
Nearest
L: Houston
Comparison
Short-term
Houston =
Simple to
Overfitting
N = 10
(2013)
monitor
and Los
with 10-20
exposure
31.0-112.5 ppb;
implement
may be
RMSE =
Angeles;
monitors in
Los Angeles =
caused by just
54-21.01 ppb,
T: Select
each
10.8-117.1 ppb
one incorrect
n =20
days in
metropolitan
parameter;
RMSE = 8.41-
2009-2011
area
does not
capture the
underlying
phenomena
19.06 ppb
Dionisio et al.
Ambient
L: Atlanta,
Comparison
Long-term
NR
Less spatial
Uncertainty in
Mean (SD)
(2014)
monitor
GA;
with dispersion
exposure
variability in
areas where
exposure
T: 1999-
model
ozone, so
there are
measurement
2002;
fixed-site
known sinks
error for
P: Entire
monitors do a
(e.g., near
omission of
population
better job
than for
spatially
variable
pollutants
roads)
spatial
variability:
-0.055 (0.037);
bias on effect
estimate for
omission of
spatial
variability:
-0.12 (0.093)
2-60
-------
Table 2-8 (Continued): Studies informing assessment of exposure measurement error when concentrations
measured by fixed-site monitors are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Monitor
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Ollison et al.
Comparison
L: Houston,
Comparison
Short-term
8-h daily max =
Average and
Positive bias
Differences
(2013)
between FRM
TX;
between
exposure
avg 22 ppb, max
maximum
of FRM due to
between
(Thermo-
T: August
ambient monitor
94 ppb,
data compare
water vapor,
monitors
scientific
26-
types using air
19 values and
well; frequent
gas-phase
presented
49C)and
November
spiked with
6 days with 8-h
calibrations
mercury, and
graphically but
FEM ambient
19, 2010;
known quantity
daily max above
and
VOCs
not reported
monitors (two
P: Entire
of ozone
75 ppb
zero/span
models:
population
improved
Teledyne
data quality
Model 211
and Teledyne
Model 265E).
The Teledyne
Model 265E
has since
been certified
as an FRM
Johnson et al.
FEM ambient
L: Durham,
Comparison
Short-term
10-min avg
Average
Potentially
In the vicinity
(2014)
monitor
NC;
with an FRM
exposure
value =
values are
high positive
of VOC
T: September
and a
33.0-55.2 ppb
only slightly
errors in
sources, ozone
2012;
microenviron-
higher than
vicinity of VOC
concentration
P: Entire
mental model
FRM
sources
from the FEM
population
was several
hundred
percentage
higher
(depending on
the source)
microenviron-
mental monitor
that compared
well with the
FRM
2-61
-------
Table 2-8 (Continued): Studies informing assessment of exposure measurement error when concentrations
measured by fixed-site monitors are used for exposure surrogates.
Reference
Monitor
Location,
Time Period,
and
Population
Measurement
Evaluation
Technique
Epidemiology
Applications
Concentrations
Measured
Strengths
Limitations
Exposure
Measurement
Errors
Buteau et al.
Fixed-site
L: Montreal,
Comparison
Short-term
8-h daily max;
Most
Lacks spatial
ICC mean
(2017)
monitors
Quebec,
with BME, IDW,
exposure
mean (SD) =
accurate
resolution
(95% CI) vs.
reporting data
Canada;
and back-
27.9 ppb
measure of
IDW 0.89
for 8-h daily
T: January 1,
extrapolation
(15.2 ppb)
ozone
(0.89, 0.89),
max
1991—
LUR
median =
vs. LUR
December
26.3 ppb
w/back-
31, 2002;
extrapolation
P: Entire 0.67 (0.47,
population 0.78), vs. BME
0.64 (0.41,
0.77)
2-62
-------
Table 2-8 (Continued): Studies informing assessment of exposure measurement error when concentrations
measured by fixed-site monitors are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Monitor
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Yu etal. (2018)
Centralized
L: Atlanta,
Comparison
Short-term
NR
Accurate
No spatial
Urban site:
monitor
GA;
with fixed-site
exposure
capture of
resolution,
MB = -2.54
reporting data
T: 2011;
monitors
temporal
autocorrelation
ppb,
for 8-h daily
P: Entire
variation
introduces
ME = 4.15 ppb,
max
population
bias
RMSE = 5.71
ppb,
MNB = -5%,
MNE = 10%,
NMB = -6%,
NME = 9%,
MFB = -6%,
MFE= 10%,
R2 = 0.91,
Slope = 1.02;
Rural site:
MB = -2.43
ppb,
ME = 7.30 ppb,
RMSE = 9.29
ppb,
MNB = -7%,
MNE = 18%,
NMB = -5%,
NME = 16%,
MFB = -9%,
MFE = 19%,
R2 = 0.70,
Slope = 0.67
AQS = Air Quality System; BME = Bayesian maximum entropy; ER = emergency room; ICC = interclass correlation coefficient; IDW = inverse-distance weighting;
IMPROVE = Interagency Monitoring of Protected Visual Environments; L = location; LUR = land use regression; MB = mean bias; ME = mean error; MFB = mean fractional bias;
MFE = mean fractional error; MNB = mean normalized bias; MNE = mean normalized error; NMB = normalized mean bias; NME = normalized mean error; NR = not reported;
OK= ordinary kriging; P = population; PAMS = Photochemical Assessment Monitoring System; RMSE = root-mean-squared error; SD = standard deviation;
SEARCH = Southeastern Aerosol Research and Characterization; STN = Speciation Trends Network; T = time; UK = universal kriging; U.S. EPA = Environmental Protection Agency.
2-63
-------
Table 2-9 Studies informing assessment of exposure measurement error when concentrations measured by
personal and microenvironmental monitors are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Monitor
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Wheeler et al.
Ogawa passive
L: Windsor,
Inter-sampler
Panel study
Mean (SD) =
Can be used
Integrated
Median
(2011)
badge
Ontario,
comparison;
26 ppb (9 ppb);
for personal,
measurements,
bias = -0.24 ppb
(diffusion-
CAN;
comparison to
median =
indoor, and
negative bias
precision = 0.09 ppb
based)
T: Winter
fixed-site
25 ppb
outdoor
and decreased
and summer,
monitors
sampling
precision
2005-2006;
P: Adults and
children with
asthma
Bart et al. (2014)
Gas sensitive
L: Lower
Collocate
Panel study
NR (shown on
Low cost, easy
Interferents:
MB = -1 ppb
semiconductor
Fraser
10 GSS
figure)
to deploy over
humidity, NO
SE = 6 ppb
monitors
Valley,
microsensors
many
British
around each of
locations;
Columbia,
10 fixed-site
provides
Canada;
monitors
real-time
T: May-
measurements
September
2012;
P: Entire
population
2-64
-------
Table 2-9 (Continued): Studies informing assessment of exposure measurement error when concentrations
measured by personal and microenvironmental monitors are used for exposure
surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Monitor
Population
Technique
Applications
Measured
Strengths
Limitations Errors
Zimmerman et al.
Low-cost
L: Pittsburgh,
Deployed a
Short-term
15-min avg
Allows for
Sensitivity to Multiple linear
(2018)
sensor
PA;
dense network of
exposure
time; NR
better spatial
model quality regression MAE avg
(Real-Time
T: August 3,
low-cost
coverage
and input data (SD) = 5.1ppb
Affordable
2016-
samplers then
quality (0.6 ppb) R = 0.81
Multipollutant
February 7,
applied one of
random forest MAE
sensors) data
2017;
two models
avg (SD) = 0.7 ppb
filtered through
P: Carnegie
(random forest
(0.1 ppb) R = 0.99
model to
Mellon
or multiple linear
improve data
University
regression) to
quality based
campus
smooth data for
on data across
population
ozone
geographical
area
Saqona et al.
Personal ozone
L:
Comparison of
Panel study
Outdoor test
Good
Measurable bias Intercomparison
(2018)
monitor
Piscataway,
personal
range (5-min
accuracy
observed during chamber:
operates by UV
NJ;
monitors with
avg time) = 0-
when
personal monitor R = 0.947,
light absorption
T: July 2014;
FEM;
65 ppb; indoor
compared
intercomparison; slope = 0.82
at 254 nm
P: Panel of
intercomparison
test range
with FEM
correlations outdoor:
wavelength (2B
volunteers
of personal
(1-min avg time)
between R = 0.991,
Technologies)
monitors
= 30-55 ppb;
chamber test
range (1-min
avg time) =
85-125 ppb
personal slope =1.08
monitors indoor;
dropped when comparison with
VOCswere FEM: outdoor
introduced to the R = 0.982,
test chamber slope = 0.92
indoor R = 0.867,
slope = 0.88
FEM = Federal Equivalent Method; L = location; MAE = mean absolute error; MB = mean bias; NR = not reported; P = population; Pearson R = correlation coefficient; SD = standard
deviation; SE = standard error; T = time; VOC = volatile organic compound.
2-65
-------
Table 2-10 Studies informing assessment of exposure measurement error when concentrations modeled by
spatial interpolations methods are used for exposure surrogates.
Reference
Model
Location,
Time Period,
and
Population
Measurement
Evaluation
Technique
Epidemiology
Applications
Concentrations
Measured
Strengths
Limitations
Exposure
Measurement
Errors
Data averaging
Joseph et Simple
al. (2013) averaging
L: Houston
and Los
Angeles;
T: Select days
in 2009-2011
Comparison
with 10-20
monitors in each
metropolitan
area
Short-term
exposure
Houston =
31.0-112.5 ppb;
Los Angeles =
10.8-117.1 ppb
Simple to
implement
Overfitting may be
caused by just one
incorrect
parameter; does
not capture the
underlying
phenomena
Houston: n = 10
RMSE =
11.30-15.35 ppb,
n = 20
RMSE =
10.77-15.07 ppb;
Los Angeles:
n = 10
RMSE =
15.16-25.13 ppb,
n = 20
RMSE =
12.96-24.35
2-66
-------
Table 2-10 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by spatial interpolations methods are used for exposure surrogates.
Reference
Model
Location,
Time Period,
and
Population
Measurement
Evaluation
Technique
Epidemiology
Applications
Concentrations
Measured
Strengths
Limitations
Exposure
Measurement
Errors
Yu et al.
Average of
L: Atlanta,
Comparison
Short-term
NR
No bias due to
Low spatial
Urban site:
(2018)
10 monitors
GA;
with fixed-site
exposure
autocorrelation
resolution
MB = -0.75 ppb,
located
T: 2011;
monitors
ME = 4.00 ppb,
across the
P: Entire
RMSE = 5.73 ppb,
city reporting
population
MNB = -1%,
data for 8-h
MNE = 10%,
daily max
NMB = -2%,
NME = 9%,
MFB = 0%,
MFE = 9%,
R2 = 0.91,
Slope = 1.14
Rural site:
MB = -0.53 ppb,
ME = 5.00 ppb,
RMSE = 6.52 ppb,
MNB = -1%,
MNE = 12%,
NMB = -1%,
NME = 11%,
MFB = -3%,
MFE = 12%,
R2 = 0.80,
Slope = 0.80
Inverse-distance weighting
Buteau et
IDW of data
L: Montreal,
Comparison Short-term
8-h daily max;
Low spatial
Quality of model
ICC mean
al. (2017)
from fixed-site
Quebec,
with BME, back- exposure
mean (SD) =
variability of ozone
depends on spatial
(95% CI) vs.
monitors
Canada;
extrapolation
28.1 ppb
may negate
density of monitors
fixed-site
T: January 1,
LUR, and
(13.0 ppb) median
limitation
monitor = 0.89
1991-
fixed-site
= 26.5 ppb
(0.89, 0.89), vs.
December 31,
monitors
LUR w/back-
2002;
extrapolation =
P: Entire
0.62 (0.59, 0.64),
population
vs. BME = 0.76
(0.72, 0.78)
2-67
-------
Table 2-10 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by spatial interpolations methods are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Yu et al.
IDW of data
L: Atlanta,
Comparison
Short-term
NR
Better spatial
Quality of model
Urban site:
(2018)
from
GA;
with fixed-site
exposure
resolution than
depends on spatial
MB = -1.27 ppb,
10 fixed-site
T: 2011;
monitors
monitor-based
density of the
ME = 2.94 ppb,
monitors
P: Entire
population
approaches
monitors
RMSE = 4.31 ppb,
MNB = -2%,
MNE = 7%,
NMB = -3%,
NME = 7%,
MFB = -3%,
MFE = 7%,
R2 = 0.95,
Slope = 1.05
Rural site:
MB = -2.43 ppb,
ME = 4.31 ppb,
RMSE = 5.44 ppb,
MNB = -6%,
MNE = 11%,
NMB = -5%,
NME = 10%,
MFB = -7%,
MFE = 11%,
R2 = 0.88,
Slope = 0.89
2-68
-------
Table 2-10 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by spatial interpolations methods are used for exposure surrogates.
Location,
Time Period
Measurement
Exposure
and
Evaluation Epidemiology Concentrations
Measurement
Reference
Model
Population
Technique Applications Measured
Strengths
Limitations
Errors
Kriging
Joseph et
Ordinary
L: Houston
Comparison Short-term
Houston =
Simple to
Overfitting may be
Houston:
al. (2013)
and
and Los
with 10-20 exposure
31.0-112.5 ppb;
implement
caused by just one
n = 10 valid pts
universal
Angeles;
monitors in
Los Angeles = 10.8—
incorrect
RMSE =
kriging
T: Select
each
117.1 ppb
parameter, does
8.53-13.12 ppb,
days in
metropolitan
not capture the
n = 20
2009-2011
area
underlying
phenomena
RMSE =
7.56-12.72 ppb;
Los Angeles:
n = 10 valid pts
RMSE =
12.55-19.30 ppb,
n = 20
RMSE =
11.04-17.84 ppb
Liu et al.
Universal
L: Eastern
Comparison of Long-term
NR
As a traditional
Assumes linearity
Kriging Model 1
(2011)
kriging
and
model points exposure
method, this is
or some simplified
Daily RMSE = 8.58-
midwestern
with
better established
function between
22.67 ppb; Kriging
U.S.;
concentrations
sampling points
Model 2 Daily
T: May 15-
from 375
RMSE = 8.65-
September
monitors
21.11 ppb
11, 1995
reporting to
10:00-17:00;
AQS
P: Entire
population
2-69
-------
Table 2-10 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by spatial interpolations methods are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Joseph et
al. (2013)
Ordinary and
universal
kriging
L: Houston
and Los
Angeles;
T: Select days
in 2009-2011
Comparison
with 10-20
monitors in each
metropolitan
area
Short-term
exposure
Houston =
31.0-112.5 ppb
Los Angeles =
10.8-117.1 ppb
Yields superior
validation compared
to other methods
Overfitting may be
caused by just one
incorrect
parameter
Houston:
OK
n = 10 valid pts
RMSE =
7.01-10.39 ppb,
n = 20
RMSE =
5.84-9.59 ppb
UK
n = 10
RMSE =
8.15-13.29 ppb,
n = 20
RMSE =
6.36-10.42 ppb
Los Angeles:
OK
n = 10 valid pts
RMSE =
12.21-16.79 ppb,
n = 20
RMSE =
10.20-19.22 ppb
UK
n = 10
RMSE =
12.43-18.96 ppb,
n = 20
RMSE =
10.85-19.70 ppb
2-70
-------
Table 2-10 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by spatial interpolations methods are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Kethireddv Ordinary
et al. kriging
(2014)
L: Texas
Comparison Short-term
Mean (SD) for
cities;
with monitors exposure
select hours
T: 2012;
reporting to the
3/25/2012
P: Entire
Texas Air
2:00 p.m. =
population
Monitoring
68.2 ppb
Information
(6.15 ppb)
System
4/24/2012
2:00 p.m. =
65 ppb (6.32 ppb)
5/17/2012
2:00 p.m. =
72.7 ppb (10 ppb)
6/28/2012
3:00 p.m. =
70 ppb (15.7 ppb)
7/21/2012
2:00 p.m. =
47 ppb (17 ppb)
8/20/2012
3:00 p.m. =
71 ppb (12 ppb)
Prediction
uncertainty can be
calculated
Accuracy depends
on input data,
proximity between
points used to fit
the model
For the same
select hours
ME = -0.000166
to 0.000407
RMSE = 0.004823
to 0.00956
standardized
mean = -0.02046
to 0.0270
RMSE
standardized =
0.714 to 1.099 avg
std
error = 0.00527 to
0.0105
Gelfand et Kriqinq
L: California;
Compare kriged
Long-term
NR
Comparison of
Preferential
RMSE high
al. (2012) (approach is
T: 2008;
ozone surface
exposure
monitor selection
sampling causes
monitors = 22.7
unspecified)
P: Entire
to monitors
allows for
overestimation or
ppb,
population
reporting to
evaluation of best
underestimation
low
AQS; kriged
practices (i.e., use
monitors = 23.9
surface is fit to
of randomly
ppb, randomly
high
selected monitors);
selected
concentration
otherwise, selection
monitors = 18.0
monitors, low
of high monitors
ppb,
concentration
causes
all monitors:
monitors,
overestimation of
18.0 ppb
randomly
concentrations and
selected
vice versa
monitors, or all
monitors
2-71
-------
Table 2-10 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by spatial interpolations methods are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Yu et al.
Kriging of
L: Atlanta,
Comparison
Short-term
NR
Better spatial
Quality of model
Urban site:
(2018)
data from
GA;
with fixed-site
exposure
resolution than
depends on spatial
MB = -2.32 ppb,
10 fixed-site
T: 2011;
monitors
monitor-based
density of the
ME = 4.85 ppb,
monitors
P: Entire
population
approaches
monitors
RMSE = 9.53 ppb,
MNB = -4%,
MNE = 11%,
NMB = -5%,
NME = 11%,
MFB = -6%,
MFE = 13%,
R2 = 0.74,
Slope = 0.87
Rural site:
MB = -4.34 ppb,
ME = 5.19 ppb,
RMSE = 8.35 ppb,
MNB = -10%,
MNE = 12%,
NMB = -10%,
NME = 12%,
MFB = -12%,
MFE = 14%,
R2 = 0.74,
Slope = 0.82
AQS = Air Quality System; BME = Bayesian maximum entropy; ICC = interclass correlation coefficient; IDW = inverse-distance weighting; L = location; LUR = land use regression;
MB = mean bias; ME = mean error; MFB = mean fractional bias; MFE = mean fractional error; MNB = mean normalized bias; MNE = mean normalized error; NMB = normalized
mean bias; NME = normalized mean error; NR = not reported; OK = ordinary kriging; P = population; RMSE = root-mean-squared error; SD = standard deviation; T = time;
UK = universal kriging.
2-72
-------
Table 2-11 Studies informing assessment of exposure measurement error when concentrations modeled by
land use regression or spatiotemporal models are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Land use regression
Clark et al.
(2011)
Land use
regression with
variables related
to the built
environment,
climate,
transportation,
and income
L: 100 U.S.
urban areas
across the
U.S.;
T: May-
September
1990;
P: Entire
population
Observed data Short-term 8-h daytime avg Observed data
used in model exposure during ozone used
was validated season
arithmetic mean
is 45 ppb
Observed data
were sparse
Model LUR
R2 = 0.34
Adam-PouDart
Land use
L: Montreal,
Cross-validation
Short- and
8-h daily max;
More accurate
Output quality
R2 = 0.466,
et al. (2014)
regression
Quebec,
against NAPS
long-term
NR
than BME
depends on
RMSE =
mixed-effects
Canada;
monitoring data
exposure
variation in some
quality of input
8.747 ppb,
model
T: May-
from 2005
cases
data
percentage
incorporating
September
change in
temperature,
1990-2009;
MSE = -19.9%
precipitation, day
P: Entire
(compared with a
of year, road
population
BME-LUR hybrid
density, and
model)
latitude
2-73
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by land use regression or spatiotemporal models are used for exposure
surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Buteau et al.
Land use
L: Montreal,
Comparison with
Short-term
8-h daily max;
Easy to
Depends on
ICC mean
(2017)
regression with
Quebec,
BME, IDW, and
exposure
mean (SD) =
implement
quality of
(95% CI) vs.
back-
Canada;
fixed-site
21.5 ppb
independent
fixed-site
extrapolation,
T: January
monitors
(15.8 ppb)
variables used
monitor = 0.67
LUR model built
1, 1991-
median =
to fit model
(0.47, 0.78), vs.
from 76 monitors
December
17.5 ppb
IDW = 0.62 (0.59,
and included
31, 2002;
0.64), vs.
variables for land
P: Entire
BME = 0.37
use and the built
population
(0.16, 0.52)
environment
Spatiotemporal models
Warren et al.
Two-stage model
L: 13
See prior
Long-term
NR
Bayesian
It is difficult to
NR
(2012)
with Stage 1 with
counties in
references by
exposure
framework pulls
distinguish how
Bayesian kriging
eastern
Fuentes and
information from
ozone exposure
of weather
Texas;
Rafterv (2005)
different sources
is dealt with in
patterns and
T: 2002-
and Fuentes
to minimize error
the model,
Stage 2 as an
2004;
(2009)
based on the
underlying
P: Preterm
data provided
process specific
births
to the pollutant
2-74
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by land use regression or spatiotemporal models are used for exposure
surrogates.
Location,
Time Period, Measurement Exposure
and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Chai et al.
(2013)
U.S. National Air
Quality Forecast
Capability
(NAQFC) 12-km
horizontal
resolution
L: CONUS;
T: 2010;
P: Entire
population
Model compared
with observed
data from AQS
Short- and
long-term
exposure
Figure 4:
between 15 and
50 ppb daily avg
across domain
Figure 5: daily
average across
urban, suburban,
rural between 15
and 50 ppb
Figure 6: across
six regions daily
avg between 10
and 60 ppb;
Figure 8:
average at
monitors for
August 2010
between 8 and
50 ppb
Figure 10: by
average hour by
region between 0
and 80 ppb
Validation
method
compared to
observed data;
multiple
timescales
explored;
regional
differences
explored
Future
predictions are
not of interest
MB = 5.6 ppb
RMSE =
15.4 ppb;
weekly MB =
9.2 ppb
Adam-Poupart
Bayesian
L: Montreal,
Cross-validation
Long-term
NR
Accurate when
When monitors
R2 = 0.414,
et al. (2014)
maximum entropy
Quebec,
against NAPS
exposure
monitoring
are not
RMSE =
model with priors
Canada;
monitoring data
stations
clustered where
9.164 ppb,
from land use
T: May-
from 2005
clustered in
people live,
percentage
regression
September
1990-2009;
P: Entire
population
study area
model quality is
reduced
change in MSE =
-23.0%
2-75
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by land use regression or spatiotemporal models are used for exposure
surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Wanq et al.
Spatiotemporal-
L: Los
10-fold cross-
Short- and
8-h daily max;
Improved
Modeling
ST-LUR:
(2016)
LUR incorporating
Angeles and
validation
long-term
range =
integration of
approach used
RMSE = 5.64 ppb
the UCD-CIT
Riverside
against
exposure
10-50 ppb
different models,
2-week data, did
R2 = 0.86;
chemical
Counties;
37 monitors for
(long-term avg)
takes advantage
not look at 8-h
ST-LUR + CTM:
transport model
T: 2000-
each variation of
of spatial
daily max
RMSE = 4.65
with meteorology
2008;
model
residuals
ppb, R2 = 0.87
modeled by WRF
P: Entire
v.3.1.1 on a 4-km
population
grid, ST-LUR
alone
Wana et al.
Spatiotemporal
L: Baltimore,
10-fold cross-
Short- and
2-week avg; NR
Low spatial
Missing ozone
Overall cross-
(2015)
model drawing
Chicago,
validation
long-term
variability of
data during cold
validation MSE
from universal
Los
against home
exposure
ozone
seasons may
R2'.
kriging,
Angeles,
and AQS
concentration
limit the
Baltimore = 0.90,
spatiotemporal
New York,
monitors in each
across space
applicability of
Chicago = 0.71,
trend, and
St. Paul,
city
allows this
the model
Los
spatiotemporal
Winston-
method to be
Angeles = 0.67,
residuals
Salem;
T: 1993-
2013;
P: MESA Air
study
participants
more applicable
New York = 0.60,
St. Paul = 0.91,
Winston-
Salem = 0.66
Overall cross-
validation R2\
Baltimore = 0.89,
Chicago = 0.72,
Los
Angeles = 0.78,
New York= 0.61,
St. Paul = 0.90,
Winston-
Salem = 0.76
2-76
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by land use regression or spatiotemporal models are used for exposure
surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Xu etal. (2016a)
Partial least
squares model
with or without
universal kriging
L: Baltimore,
LOOCV, using
MD;
comparison with
T: February
measurements
11-22, 2012
obtained on a
and June
mobile
18-27,
monitoring
2012;
platform
P:
Participants
in the
MESA-Air
study
Short-term
exposure
2-week avg;
range
Summer =
50.5-93.6 ppb,
Winter =
20.6-37.4 ppb
PLS+ UK model
accounts well for
spatial residuals
There would be
more
confidence in
the cross-
validation if the
monitoring
periods were
run for more
days; averaging
times for cross-
validation
monitors varied
by season,
meteorology
was not
accounted for
LOOCV R2 PLS
summer = 0.55
winter = 0.40,
PLS + UK
summer = 0.71
winter = 0.58
Sahu and Bakar
(2012b)
Bayesian
autoregressive
models
L: Eastern
U.S.;
T: 1997-
2006;
P: Entire
population
Comparison of
model variations
with
concentrations
from
69 fixed-site
monitors
(622 sites used
to fit the model)
Long-term
exposure
Range:
annual 4th
highest =
48.5-109 ppb,
3-yr avg = 50.6-
100.2 ppb
Accurate, good
representation of
both spatial and
temporal
variability, can
be used for
downscaling
CTMs
Because the
model takes
meteorological
inputs, it cannot
be validated
with
meteorological
data
Annual fourth
highest:
RMSE = 5.24 ppb
MAE = 4.17 ppb;
3-yr avg
RMSE = 4.21
ppb, MAE = 3.36
ppb
2-77
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by land use regression or spatiotemporal models are used for exposure
surrogates.
Location,
Time Period, Measurement Exposure
and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Sahu and Bakar
(2012a)
Dynamic linear
and Bayesian
autoregressive
models
L: New York
State;
T: July-
August
2006;
P: Entire
population
Comparison of
model variations
with
concentrations
from 4 fixed-site
monitors
(25 sites used to
fit the model)
Short-term
exposure
Range: median
of 8-h daily max
= 25-70 ppb
Autoregressive
model performs
well
Model
performance
depends on
parameter
selection; need
to run the model
for the same
domain for
which it is fit
MSE
dynamic linear
model = 58.42
ppb2
Autoregressive
model = 46.16
ppb2 (RMSE
dynamic linear
model = 7.64
ppb,
autoregressive
model =
6.79 ppb)
Buteau et al.
Bayesian
L: Montreal,
Comparison with Short-term
8-h daily max;
Captures spatial
Higher
ICC mean
(2017)
maximum entropy
Quebec,
IDW, back- exposure
mean (SD) =
30.0 ppb
(9.1 ppb)
median =
29.8 ppb
variability more
complexity
(95% CI) vs.
model drawing
Canada;
extrapolation
completely
compared with
fixed-site
output from a land
T: January
LUR, and
other models
monitor = 0.64
use regression as
1, 1991 —
fixed-site
(0.41, 0.77), vs.
its prior
December
monitors
IDW = 0.76 (0.72.
31, 2002;
P: Entire
population
0.78), vs. LUR
w/back-
extrapolation = 0.
37 (0.16, 0.52)
2-78
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by land use regression or spatiotemporal models are used for exposure
surrogates.
Location,
Time Period, Measurement Exposure
and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Gong et al.
(2017)
HYSPLIT v4.9
L: Eight
Model compared
with GDAS 1° x
western U.S.
(and built) with
1° data with a
cities (i.e.,
surface,
GAM, satellite
Houston,
fixed-site
data from HMS
Boise,
monitoring data
Denver, Fort
Collins,
Yellowstone,
Provo, Salt
Lake City,
Spokane);
T: May to
September
for 2008 to
2015;
P: Entire
population
Short- and
long-term
exposure
8-h daily max
Multiple models
ozone for
used together
Houston
(e.g., HYSPLIT
between 0 and
and HMS and
120 ppb;
obs data)
Table 3: obs 8-h
daily max ozone
mean =
39.29 ppb, no
smoke n = 1,082,
smoke n = 41,
no smoke
residuals =
-0.33,
smoke residuals
= 8.10; 8-h daily
max ozone for
Provo site
between 20 and
80 ppb
Limited cities
explored in
western U.S.
Model-obs
comparison
R2 = 0.816 for
Houston
2-79
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by land use regression or spatiotemporal models are used for exposure
surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Chanq et al.
Stochastic partial
L:
Cross-validated
Long-term
NR (graphed for
Model performs
Large domain
Cross validation
(2015)
differential
Worldwide;
model against
exposure
entire world but
better over
means local
SE December-
equation
T: 2000-
measurements
not reported)
longer time
problems with
February = 6.55-
2005;
across the world
periods
model fitting
11.95 ppb
P: Entire
March-
population
May = 5.86-9.96
ppb
June-
August = 6.13-8.
65 ppb
September-
November = 5.82
-11.23 ppb
AQS = air quality system; BME = Bayesian maximum entropy; CI = confidence interval; CTM = chemical transport model; GAM = generalized additive model; GDAS = Global Data
Assimilation System; HMS = Hazard Mapping System; ICC = interclass correlation coefficient; IDW = inverse-distance weighting; L = location; LOOCV = leave-one-out cross
validation; LUR = land use regression; MAE = mean absolute error; MB = mean bias; MESA = Multi-Ethnic Study of Atherosclerosis; MSE = mean squared error; NAPS = National
Air Pollution Surveillance; NAQFC = National Air Quality Forecast Capability; NR = not reported; P = population; PLS = partial least squares; RMSE = root-mean-squared error;
SD = standard deviation; SE = standard error; ST = spatiotemporal; T = time; UCD-CIT = University of California at Davis—California Institute of Technology model; UK = universal
kriging; WRF = Weather Research and Forecasting model.
2-80
-------
Table 2-12 Studies informing assessment of exposure measurement error when concentrations modeled by
chemical transport modeling are used for exposure surrogates.
Location,
Time Period, Measurement Exposure
and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Hutzell et al.
(2012)
CMAQ
modeled
output version
4.7.1 with a
36-km grid
and nested
12-km grid
L: Eastern
CONUS;
T: January
and July in
2002;
P: Entire
population
CMAQ with
SAPRC07T
mechanism
compared with
SAPRC-99
mechanism; both
mechanisms
compared with
observed fixed-site
monitoring data
Short-term
exposure
NR (shown
graphically)
Specific
mechanism in
CMAQ
investigated
during and not
during the ozone
season
Modeling time is
relatively short, so
an epi application
may be limited
Mechanisms and
observed data
compared
January
NMB = -16 to
16 ppb, July
NMB = -20 to
20 ppb; R2 = 0.7-
0.8, January
RMSE = 7.46-
7.59 ppb, July
RMSE = 12.4-
13.6 ppb,
January
MB = -1.11 to
-1.37 ppb, July
MB = 5.2-7.1
ppb, January
NMB = -4.1 to
-5.1 %, July
NMB = 9.2-
12.6%, January
ME = 5.77-
5.86 ppb, July
ME = 9.73-10.6
ppb, January
NME = 1.5-
21.8%, July
NME = 17.2-
18.8% between
the two
mechanism
2-81
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period, Measurement Exposure
and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
CMAQ 4.6
L: Southeast
CMAQ with
Short-term
8-h avg during
The scale of the
Only 3 weeks of
MFE by site for
with nested
Texas;
SAPRC07
exposure
ozone episode;
CMAQ was very
data investigated
S99 = 0.14-0.33
grids (36 and
T: 3 weeks
mechanism
hourly ozone
fine, and the
ppb and 0.25 ppb
12 km) going
of hourly
compared with
NR (shown
specific
overall, MFE by
down to 4-km
ozone
SAPRC99
graphically)
mechanisms
site for
grid size
between
mechanism; both
were directly
S07 = 0.17-32
August 16
mechanisms
compared
ppb and 0.25 ppb
and
compared with
overall, MFB by
September
observed fixed-site
site for
6, 2000;
monitoring data
S99 = -0.21
P: Entire
to-0.04 ppb and
population
-0.12 overall,
MFB by site for
S07 = -0.24 to
-0.07 ppb and
-0.16 overall,
accuracy of
paired peak by
site for
S99 = -0.29-
0.00 ppb and
-0.16 overall,
accuracy of
paired peak by
site for
S07 = -0.29 to
-0.05 ppb and
-0.20 overall,
n = 27-138 and
1,887 overall
comparing
mechanisms to
observed data
2-82
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Reference
Model
Location,
Time Period,
and
Population
Measurement
Evaluation
Technique
Epidemiology
Applications
Concentrations
Measured
Strengths
Limitations
Exposure
Measurement
Errors
Emerv et al.
GEOS-Chem
L: CONUS;
Comparison done
Short- and
Summer
Two different
The coarse grid of
R2 for
(2012)
8-03-01 on a
T: Hourly
between
long-term
average for
modeling
GEOS-Chem may
CAMx = 0.34-
2- x 2.5-
ozone data
GEOS-Chem and
exposure
2006; NR
methods are
be a weakness
0.61, R2 for
degree grid,
for all of
collocated
(shown
compared and
GEOS-Chem =
CAMx 5.30
2006;
observed data
graphically)
are both are
0.21-0.66
run on a
P: Entire
compared to
between
36-km and
population
observed data
observed and
then 12-km
grid size
modeled;
R2 = 0.50 of
fraction of days
>60, >65,
>70 ppb for
CAMx and
R2 = 0.42, 0.47,
0.47 for
GEOS-Chem and
R2 = 0.42, 0.30,
0.25 for
GEOS-Chem,
number of days
with certain
ozone ranges 0-
240
2-83
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
McDonald-Buller
et al. (2011)
GEOS-Chem
L: CONUS,
Comparison done
Short- and
(it is not clear
with global
with fixed-site
long-term
the version of
comparisons
CASTNet monitors,
exposure
GEOS-Chem
ofTES, OMI,
global scales
used with
GEOS-
compared with
resolution
Chem (TES
OMI, TES,
0.5 x 0.67
AK), and
GEOS-Chem
degree), on a
GEOS
(TES AK), and
4- x 5-degree
Chem
GEOS-Chem
global grid
(OMI AK);
(OMI AK)
T: 8-h daily
max from
March-May
2006
compared
with June-
August
2006, with
results
displayed
between
2006 and
2008;
P: Entire
population
NR (shown
graphically)
Modeled output
compared to
observed data
and other
modeled values
Assumption that
observed data is
representative
No exposure
measurement
error presented
in tables
2-84
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Yina and Li (2011)
CMAQ-MCM
L: Houston-
CMAQ-MCM
Short-term
Hourly ozone
Very fine scale
Only measured
Bias between the
(Master
Galveston
compared with
exposure
differences
spatial
during an ozone
two models
Chemical
Bay area;
CMAQ with
between the two
resolution, two
episode, so longer
between 4 and
Mechanism)
T: 3-week
SAPRC07 (version
models ranged
modeled
ozone values may
12 ppb
with three
ozone
not stated) with
from 4-12 ppb
compared, both
be less
nested
episode
fixed-site U.S. EPA
for averaged
modeled
represented
domains
period
monitors from the
across the
compared with
(36 km, 12 km,
between
AIRS database
ozone episode
observed data
4 km), CMAQ
August 16
4.6 with MCM
and
3.1
September
2000;
P: Entire
population
2-85
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Schere et al.
CTM ofCMAQ
L: CONUS,
CMAQ over
Short-term
NR (shown
Varying
Not clear the
Seasonal
(2011)
using the
Western
CONUS compared
exposure
graphically)
boundary
version of each
differences in
boundary
Europe;
with observed
conditions of
model
surface ozone
conditions of
T: Hourly
ozonesonde,
multiple models
MB = -20 to
GEOS-Chem
ozone data
CHIMERE with
compared over
20 ppb, MB for
8-03-01 (not
for the
AQMEII boundary
different parts of
four different
clear the
entirety of
conditions
the world also
3-month periods
CMAQ version
2006;
compared with
compared with
between = -3 to
or grid cell
P: Entire
typical CHIMERE
observed data,
3 ppb, difference
resolution),
population
boundary
vertical profiles
in SD of modeled
CHIMERE at a
conditions over
also explored
ozone = 0-
0.25-degree
CONUS, CMAQ
2.5 ppb
horizontal
with AQMEII
resolution
boundary
(note clear on
conditions
the CHIMERE
compared with
version)
GEOS-Chem
boundary
conditions over
Western Europe,
CHIMERE with 3-h
boundary
conditions
compared with
monthly
climatology over
Western Europe
Brauer et al.
TM5 CTM
L: global
TM5 data
Long-term
NR (shown
Multiple data
Only one data
No exposure
(2012)
model with
model;
evaluated
exposure
graphically)
sources merged
source available
measurement
0.1-degree
T: 1990,
elsewhere
together
for ozone
errors found in
resolution
2005;
P: Entire
including
observed data
tables or figures
population
2-86
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Simon et al.
CMAQ 4.7.1
L: Eastern
Modeled output
Short-term
NR (shown
A NAAQS
Greater
RMSE between
(2013)
with 12-km
U.S. from
compared with
exposure
graphically)
application was
uncertainties
HDDM and
horizontal
CONUS;
observed ozone
applied to the
associated with
brute-force
resolution
T: Hourly
data from AQS
method, method
introducing a new
method for
ozone in
coming from
allows for
process to the
selected stations
July and
fixed-site monitors;
specification
CMAQ model,
by ozone
August
CMAQ 4.7.1 using
with chemical
results only shown
concentration,
2005;
brute force
processes with
for selected
RMSE = 6 ppb at
P: Entire
emissions changes
HDDM, method
monitors; model
Charlotte site,
population
clearly
explained;
method
compared to
observed data
only run for 2 mo
RMSE = 4-7 at
Detroit site
2-87
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Reference Model
Location,
Time Period,
and
Population
Measurement
Evaluation Epidemiology
Technique Applications
Concentrations
Measured Strengths
Limitations
Exposure
Measurement
Errors
TsimDidi et al. CMAQ 4.7
L: Seattle,
CMAQ 4-km Short-term
NR (shown Method was
This study only
Comparing
(2012) with 12-km
WA;
resolution exposure
graphically) compared to
looks at a 12 days
modeled to
horizontal
T: Hourly
compared with
observed data;
study period, so
observed to
resolution with
data for July
12 and 36 km,
the paper
this study is not
hourly, max
nested 4-km
12-24,
compared with
explored the
appropriate for
hourly, max 8-h
grid in Seattle
2006;
observed data from
effect of grid
long-term
ozone to different
with a
P: Entire
fixed-site monitors
resolutions
exposure, the
grid resolutions
DDM-3D
population
from AQS
short time period
MB, GE, RMSE,
does not capture
NMB, NME,
low values well,
mean
the short time
obs = 30.2-
period would not
57.0 ppb for
be representative
4 km, 30.3-57.0
of typical
ppb for 12 km,
long-term
38.3-65.1 ppb for
exposures
36km, mean
mod = 39.1-43.1
ppb for 4 km,
39.1-58.2 ppb for
12 km, 48.6-68.9
ppb for 36 km, n
68-1,610 for
4 km, 138-3,283
for 12 km, 9,375-
226,597 for
36 km,
MB = -4.5 to
12.9 ppb for
4 km,
2-88
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period, Measurement Exposure
and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Tsimpidi et al.
(2012) (continued)
CMAQ 4.7
with 12-km
horizontal
resolution with
nested 4-km
grid in Seattle
with a
DDM-3D
(continued)
L: Seattle,
WA;
T: Hourly
data for July
12-24,
2006;
P: Entire
population
(continued)
CMAQ 4-km
resolution
compared with
12 and 36 km,
compared with
observed data from
fixed-site monitors
from AQS
(continued)
Short-term
exposure
(continued)
NR (shown
graphically)
(continued)
Method was
compared to
observed data;
the paper
explored the
effect of grid
resolutions
(continued)
This study only
looks at a 12 days
study period, so
this study is not
appropriate for
long-term
exposure, the
short time period
does not capture
low values well,
the short time
period would not
be representative
of typical
long-term
exposures
(continued)
-6.2 to 8.8 ppb
for 12 km, 1.5-
10.9 ppb for
36 km,
MAGE = 14.3-
18.5 ppb for
4 km, 14.3-15.8
ppb for 12 km,
12.4-16.1 ppb for
36 km,
RMSE = 19.4-
22.2 ppb for
4 km, 17.7-19.6
ppb for 12 km,
16.5-20.9 ppb for
36 km,
NMB = -7.9 to
42.7% for 4 km,
-11.0-28.9% for
12 km,
2.6-26.8% for
36 km,
NME = 25.1-
61.3% for 4 km,
25.1-52.2% for
12 km, 21.5-
42.0% for 36 km;
sensitivity of
ozone from NOx
and VOC is
percentage
change in
ppb = 0.000-
0.050%
2-89
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period, Measurement Exposure
and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Ferreira et al.
(2012)
CAMx (version
L: North
unclear, but its
America;
citation is from
T: Hourly
2010) on a
ozone data
24-km grid,
from all of
MM5 3.7 (the
2006;
met data) on a
P: Entire
1-degree grid,
population
emissions in a
12-km grid
based on NEI,
Canadian
emissions
inventory,
1999 Mexican
BRAVO
inventory,
biogenic
emissions
from BEIS
3.14, fire
emissions
from HMS and
SMARTFIRE
(2006), point
sources from
Continuous
Emissions
Monitoring
data
Comparison to Short- and
observed fixed-site long-term
monitors exposure
Summer 2006
daily ozone =
between 20 and
45 ppb
Modeled output
compared to
observed data;
the method
focuses on three
particular ozone
periods of
concern
Although an
MM5-CAMx
combination is
part of the
AQMEII initiative,
the method was
not compared to
any other
modeling method,
inputs of
emissions
inventory and met
data has poorer
performance
Correlation
between
modeled and
observed as a
function of RMSE
and normalized
SD, R = 0.5-0.6
2-90
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Hu etal. (2012)
UCD-CIT
L: San
Compare model
Short-term
Mapped
Study designed
4-km resolution
SAPRC-07: 1-h
model
Joaquin
results from two
exposure
concentrations
to interpret
not fine enough to
daily max SJV
Valley, CA
photochemistry
from different
different
detect local
bias = -15.3 ppb
and entire
modules with
emissions
emissions
gradients and
NB = -15.6%
state of
concentrations
sources but did
sources' impacts
small-scale
gross error
California;
from CARB
not tabulate
on
variation in
15.6 ppb
T: July 27-
observation sites
concentrations
sources
NGE = 16.0%,
August 2,
and to evaluate
domain
2000;
different model
bias = -12.7 ppb
P: Entire
variations
NB = -14.5%
population
gross
error = 13.6 ppb
NGE = 15.6%,
8-h daily max
SJV bias = -12.6
nnb
NB = -14.4%
gross
error = 12.8 ppb
NGE = 14.7%,
domain bias
-10.8 ppb
NB = -13.5%
gross
error = 11.5 ppb
NGE = 4.4%;
SAPRC-99 = 1-h
daily max SJV
bias = -0.2 ppb
NB = -0.2%
gross
error = 6.2 ppb
NGE = 6.3%,
domain
bias = -3.6 ppb
NB = -4.1%
2-91
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Hu etal. (2012)
UCD-CIT
L: San
Compare model
Short-term
Mapped
Study designed
4-km resolution
gross
(continued)
model
Joaquin
results from two
exposure
concentrations
to interpret
not fine enough to
error =11.5 ppb
(continued)
Valley, CA
photochemistry
(continued)
from different
different
detect local
NGE = 13.1%,
and entire
modules with
emissions
emissions
gradients and
8-h daily max
state of
concentrations
sources but did
sources' impacts
small-scale
SJV bias = -0.5
California;
from CARB
not tabulate
on
variation in
ppb NB = -0.5%
T: July 27-
observation sites
(continued)
concentrations
sources
gross
August 2,
(continued)
and to evaluate
(continued)
error = 5.7 ppb
2000;
different model
NGE = 6.6%,
P: Entire
variations
domain
population
(continued)
bias = -2.7 ppb
(continued)
NB = -3.3%
gross
error = 9.6 ppb
NGE = 12.0%
Liu and Zhana
CMAQ 4.4
L: CONUS;
Comparison of U.S.
Short-term
1-h mean obs =
Many evaluation
The horizontal
1-h R = 0.7-0.8,
(2011)
over CONUS
T: Hourly
EPA observed
exposure
53.0-60.3 ppb,
methods: both
resolution was
1-h NMB = 4.9-
at 32-km
ozone from
fixed-site monitors
1-h mean mod =
the horizontal
coarse; modeling
17.0%, 1-h
horizontal
June 12-28
62.0-67.4 ppb,
grids and
period was short
NME = 15.6-
resolution with
1999;
1 h n = 84-
vertical grids
and specific so epi
25.7%, 8-h daily
MM5 3.4 with
P: Entire
14,659; 8-h
(through flight
application will not
max R = 0.8-0.8,
NEI 3
population
daily max mean
data) were
be representative
8-h daily max
obs = 46.6-55.0
evaluated with
of longer-term
NMB = 8.5-
ppb, 8-h daily
observed data,
exposure
25.0%, 8-h daily
max mean mod
satellite data
max
= 58.3-62.2
used
NME = 17.0-
ppb, 8-h daily
30.1%
max n = 82-
14,619
2-92
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Reference
Model
Location,
Time Period,
and
Population
Measurement
Evaluation
Technique
Epidemiology
Applications
Concentrations
Measured
Strengths
Limitations
Exposure
Measurement
Errors
Wana et al. (2012)
CMAQ 4.7
L: Global
CMAQ-Dust
Short-term
In U.S. mean
The module
The short
In U.S. during
horizontal grid
model;
compared to
exposure
max 1 h
introduced is
modeling period
dust episode
resolution of
T: April
observed data from
modeled from
highly
may not be
between obs and
108 km with a
2001;
AQS; several
ozone =
specialized;
appropriate in a
modeled,
dust
P: Entire
versions of the
43.4-54.1 ppb;
global model
long-term epi
R = 0.48-0.54,
component
population
CMAQ model
AIRS 8-h daily
used
study and may not
NMB = -7.3-
called
compared with
max mean mod
be representative
2.8%, NME =
CMAQ-Dust
observed data from
= 45.4-51.1
of exposures
16.6-18.5%,
with
different U.S. EPA
ppb; AIRS max
outside of this
R = 0.46-0.53,
incorporation
fixed-site
1 h n =29,993,
time window;
NMB = -4.7 to
of
monitoring
mean obs =
coarse resolution
6.8%, NME =
ISORROPIA
networks
52.7 ppb, mean
is very coarse,
17.9-18.8%;
II, WRF 3.2,
mod = 48.7-
making exposure
R = -0.03 to
1999 NEI 1
54.1 ppb; in
assignment in an
0.06, NMB =
Beijing mean
epi study have
-9.36 to 17.3%,
1-h max ozone
potential
NME = 25.5-
is 86.8-112.4
misclassification
30.6%; ozone
ppb, max 1 h, n
difference
= 30, mean obs
spatially = -1.5 to
= 95.8 ppb,
1.5 ppb
mean mod =
86.8-109.9 ppb
2-93
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
CMAQ
L: Eastern
Two different
Short-term
1 h for model
Multiple
Version of CMAQ
1 h for model
(version NR)
U.S. with
modeling methods
exposure
ARW (model
comparison
never stated;
ARW (model
12 km where
flight data
were compared
NMM),
methods with
vertical validation
NMM) MB = 7.5
WRF-ARW
over east
with observed data
n = 51,532,
two different
not applicable for
ppb (8.1 ppb),
(Advanced
Texas;
from U.S. EPA's
mean obs =
types of
an epi setting;
RMSE = 13.4
Research
T: Hourly
fixed-site monitors
48.6 ppb, mean
modeled and
because the study
ppb (13.9 ppb),
WRF 3.0) and
data from
and observed data
mod = 56.2 ppb
multiple types of
was short term,
NMB = 15.5%
WRF-NMM
August 1-
from airplanes with
(56.7 ppb), 8-h
observed data
ambient
(16.7%),
are compared
October 15,
flight paths and
daily max mean
(e.g., fixed-site,
concentrations
NME = 22.3%
2006;
ship data in port
obs = 42.7,
flight data, ship
may not be
(22.8%), R =
P: Entire
mean
data)
representative of
0.76 (0.75), 8-h
population
mod = 50.4 ppb
a longer-term
daily max mean
(52.0 ppb);
exposure
MB = 7.7 ppb
mean ± SD for
(9.3 ppb), RMSE
obs = 36.38 ±
= 12.6 ppb (13.8
24.13
ppb), NMB =
18.0% (21.8%),
NME = 24.2%
(26.4%), R =
0.76 (0.74); NMB
by ozone
concentration =
-9.7 to 48.3 ppb;
time series
between the two
WRF models, MB
= 2-14 ppb,
RMSE = 8-16
ppb,
2-94
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Yu etal. (2012)
CMAQ
L: Eastern
Two different
Short-term
1 h for model
Multiple
Version of CMAQ
NMB = 0-0.5%,
(continued)
(version NR)
U.S. with
modeling methods
exposure
ARW (model
comparison
never stated;
NME = 0.1-
12 km where
flight data
were compared
(continued)
NMM),
methods with
vertical validation
0.5%, R = 0.3-
WRF-ARW
over east
with observed data
n = 51,532,
two different
not applicable for
0.9; mod NMM =
(Advanced
Texas;
from U.S. EPA's
mean obs =
types of
an epi setting;
40.07 ± 22.46,
Research
T: Hourly
fixed-site monitors
48.6 ppb, mean
modeled and
because the study
mod ARW =
WRF 3.0) and
data from
and observed data
mod = 56.2 ppb
multiple types of
was short term,
41.33 ± 20.36,
WRF-NMM
August 1-
from airplanes with
(56.7 ppb), 8-h
observed data
ambient
NMB NMM =
are compared
October 15,
flight paths and
daily max mean
(e.g., fixed-site,
concentrations
10.1%, NMB
(continued)
2006;
ship data in port
obs = 42.7,
flight data, ship
may not be
ARW = 13.6%
P: Entire
(continued)
mean
data)
representative of
population
mod = 50.4 ppb
(continued)
a longer-term
(continued)
(52.0 ppb);
mean ± SD for
obs = 36.38 ±
24.13
(continued)
exposure
(continued)
Godowitch et al.
CMAQ 4.7,
L: Eastern
Compared with
Short-term
Daily 8-h max
Methods clearly
Short-term
Time series
(2011)
12-km
U.S.;
observed data from
exposure
ozone during
explained,
exposure values
mean of 8-h daily
horizontal
T: June-
U.S. EPA's
each day of
multiple
not representative
max = 40-80
resolution,
August 2002
CASTNet,
study period =
observed data
of longer-term
ppb; 95% time
MM5 3.7.4,
hourly ozone
observed data from
40-120 ppb
for evaluation
exposures
series of 8-h daily
SMOKE 2.2
data;
P: Entire
population
flights taken in the
afternoon of July
2002
methods
max = 55-120
ppb
2-95
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Bravo et al. (2012)
CMAQ 4.5.1
L: Eastern
Comparing CMAQ
Long-term
County-level
Explicit
No new method
Monthly NMB
12-km
U.S.;
outputs to
exposure
seasonal
county-level
developed, only
between -2 to
horizontal
T: 2002;
observed data
average (April-
aggregations
an evaluation of
12%; annual
resolution
P: Entire
September) =
between two
CMAQ at a county
average NMB in
population
25.1-70.0 ppb
methods were
compared
level
southeastern
U.S. = 10-30%
R > 0.80 in upper
Southeast,
Northeast, Ohio
River Valley,
0.61-0.80 in
Florida, Gulf
Coast, Great
Lakes
Carlton and Baker CMAQ 4.7.1
L: Oark
Biogenic emission
Short-term Hourly ozone;
Comparison of
Given the short
1-1 line with
(2011) 12-km
region of the
from both BEIS
exposure NR (shown
biogenic
time period,
observed data
horizontal
U.S.
3.14 and MEGAN
graphically)
emissions has
concentrations
and difference
resolution,
covering
2.04 compared with
the specificity to
over a month and
modeled values
AER05,
Missouri,
fixed-site monitors
understand the
half may not be
but not
CB05, WRF
Illinois,
(AIRS, CASTNet/
crux of ozone
indicative of more
correlation
3.1, 2001 NEI
Indiana,
IMPROVE
differences in
long-term
calculated; no
2 and BEIS
Kentucky,
network), balloon
areas with
exposure; the
clear exposure
3.14
and northern
and aircraft
higher isoprene
area is relatively
measures related
compared with
Arkansas;
measurements
emissions;
rural so precursor
to ozone found in
MEGAN 2.04
T: Hourly
comparison
emissions may
text, figures, or
ozone data
methods were
vary in other parts
tables
from June
thorough with
of the U.S. where
15-July 31,
comparing two
more population
1998;
types of
may be affected
P: Entire
modeled values
population
with multiple
sources of
observed data
2-96
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Stevn etal. (2013)
WRF 3.1,
L: Metro
Models compared
Short- and
CMAQ (NRC)
Very few studies
Difficulties in
For CMAQ
SMOKE 2.5,
Vancouver,
with each other, all
long-term
for 2001
explore
validating
(NRC) for 2001
CMAQ 4.7.1
Canada;
model compared
exposure
n = 1,717
long-term
historical
MB = 5.7 ppb
with 4-km
T: July 19-
with historical
(5,948), mean
concentrations;
emissions
(2.6 ppb),
horizontal
21, 1985,
surface, fixed-site
mod = 26.8 ppb
multiple types of
NMB = 5.7%
resolution,
July 17-19,
monitors from
(21.8 ppb),
obs data;
(13.3%),
model
1995,
Canada's NAPS,
mean obs =
multiple models
ME = 11.5 ppb
compared with
August 10-
aloft observed data
21.2 ppb
compared
(9.8 ppb),
NRC
12, 2001,
from aircraft in
(19.2 ppb);
NME = 54.4%
MM5/CMAQ
June 24-26,
1995
mean mod
(51.2%); for
for 2001,
2006;
across years =
stations T12 in
model
P: Entire
21.8-27.6 ppb,
1985 CMAQ
compared with
population
mean obs
(CALGRID,
CALGRID and
across years =
UAM)MAE =
UAM for 1985
16.8-27.7 ppb
-20.0 ppb (25.0
ppb, 28.2 ppb),
RMSE = 23.7
ppb (29.5 ppb,
33.1 ppb),
IOA = 0.79 (0.57,
0.44)
2-97
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Reference Model
Location,
Time Period,
and
Population
Measurement
Evaluation
Technique
Epidemiology
Applications
Concentrations
Measured
Strengths
Limitations
Exposure
Measurement
Errors
Zhou etal. (2013) CMAQ 4.7
L: Eastern
Observed data
Long-term
Only ozone
There are
The paper
Relative
12-km
U.S.;
from surface
exposure
differences are
multiple
recognizes the
difference
horizontal
T: 2002 and
fixed-site monitors
explored, and
comparisons of
issues of
between
resolution,
2006;
were compared
direct ozone
this paper:
long-range
quantities of obs
MM5 3.6.3,
P: Entire
with CMAQ; 2002
concentrations
modeled to
transport of ozone
and mod, in SIP
SMOKE 2.2
population
compared with
are not explored
observed, data
call region (obs is
2006; NOxSIP Call
from 2002
reference):
region compared
compared with
average change
with data outside of
2006; ozone by
(2002-2006) in
NOxSIP Call
percentage;
8-h daily max =
region
inside NOx SIP
42.5%, average
Call area vs.
percentage
outside
change in 8-h
daily max =
38.9%; outside
SIP call region:
average change
(2002-2006) in
8-h daily max =
66.7%, average
percentage
change in 8-h
daily max =
69.7%
2-98
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Appel etal. (2012)
CMAQ 4.7.1
12-km
horizontal
resolution, met
data from both
GEOS-Chem
and GEMS
L: North
America and
Europe;
T: 2006;
P: Entire
population
CMAQ-GEMS
compared with
CMAQ-GEOS-
Chem; CMAQ in
North American
compared with U.S.
EPA data and
CMAQ in Europe
compared with
AirBase data
Long-term
exposure
NR (shown
graphically)
Several different
comparison
models and
observed data
used
This may be less
amenable to
shorter term
exposures
MB winter = -3.5
ppb spring
-1.8 ppb summer
= 4.4 ppb fall =
2.6 ppb; NMB
winter = -13.4%
spring = -4.1%
summer = 9.8%
fall = 8.4%; ME
winter = 9.0 ppb
spring = 9.3 ppb
summer =11.0
ppb fall =
8.8 ppb; NME
winter = 34.7%
spring = 29.4%
summer = 24.2%
fall = 28.0%
Choet al. (2012)
CMAQ
(version
unknown),
MM5 with
4-km
horizontal
resolution,
emissions
data from
2006
compared with
2002
L: East
Alberta,
Canada;
T: May-
August
2002:
P: Entire
population
Emissions data
compared to each
other, all modeled
outputs compared
to surface,
fixed-site observed
data (origin of
monitors unknown)
Short-term
exposure
NR (shown
graphically)
Fine scale
spatial
resolution;
complete spatial
coverage with
CMAQ
Version of some
of the model
components are
unclear; a 4-mo
exposure window
may not be
indicative of a
more long-term
exposure
Hourly ozone at
4 km resolution,
May-August
across all sites,
no threshold: FB
= 13%,
FE = 39%;
40 ppb threshold:
FB = 16%,
FE = 20%
2-99
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Herwehe et al.
CMAQ 4.7,
L: CONUS,
WRFCMAQ
Short-term
NR (shown
Modeled
Given the short
RMSE = 11.52
(2011)
WRF-ARW 2.2
observed
compared with
exposure
graphically)
differences are
time period of the
ppb (13.57 ppb),
compared with
aloft data
WRF/Chem,
specific enough
model run,
NME = 18.2%
WRF/Chem
from
modeled values
to pinpoint
concentrations
(21.5%),
3.0.1.1 with
Centreville
compared with
differences in
may not be
MB = 3.62 ppb
12-km
and
AQSand SEARCH
modeled ozone;
indicative of
(6.18 ppb), NMB
horizontal
Birmingham,
fixed-site surface
several
typical ozone
= 7.4% (12.7%),
resolution
AL;
T: August
2006;
P: Entire
population
monitors, observed
aloft ozone data
difference
sources of
ozone data
explored
(e.g., fixed site
and aloft data)
exposures
R = 0.72 (0.66)
Wona et al. (2012)
CMAQ 4.7.1
L: Portion of
Coupled
Short-term
NR (shown
A highly
The model run is
Comparison
12-km
California
WRF-CMAQ
exposure
graphically)
specialized
only for a few
between mod
horizontal
and
compared with
modeled
days during a
and obs: all data
resolution with
surrounding
offline WRF with
component
wildfire; therefore,
(daytime) slope =
two-way
states;
CMAQ, both
pinpoints the
short-term
0.98, R = 0.62;
coupled
T: June 20-
methods are
differences in
exposures may be
when AOD > 0.5
WRF-CMAQ
29, 2008;
compared with
modeled ozone;
higher than a
slope = 1.2,
model
P: Entire
population
fixed-site observed
monitoring data
from AQS
comparisons are
made with
observed data
typical
concentration
R = 0.75
2-100
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period, Measurement Exposure
and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Zhang et al.
(2013)
CMAQ 4.5.1
L: South-
with CAMx
eastern U.S.
4.42 both with
in North
MM5 3.7 and
Carolina,
a 4-km
northeastern
horizontal
Georgia,
resolution
South
Carolina,
Tennessee,
Virginia, and
Kentucky;
T: Hourly
ozone in
January and
July 2002;
P: Entire
population
CMAQ 4.5
L: CONUS;
36-km
T: Hourly
horizontal
ozone July
resolution
1-August
31,2004;
P: Entire
population
Both CMAQ and
Short-term
January 2002:
Fine-scale
Short-term
1-h max, n
CAMx are
exposure
Mean 1-h max =
horizontal
exposure during
between 62 and
compared to each
31.8-42.0 ppb,
resolution; two
only 2 mo may not
384, R = 0.5-0.7,
other; both
8-h max = 27.5-
different models
be indicative of
NMB = -7.6 to
modeled compared
39.0 ppb across
compared with
typical, long term
10.1%, NME =
to observed
CMAQ and
both models
exposures
15.4-24.5%, and
fixed-site monitors
CAMx at
compared to
8-h max, n = 61-
from AQS,
locations of
observed data;
384, R = 0.6-0.7,
CASTNet,
AQS, CASTNet,
errors with
NMB = 0.1-
SEARCH, and
and SEARCH
observed data
15.8%, NME =
NCDENR
monitors;
presented by
19.2-25.4%
January 2002:
monitoring
Mean 1-h max =
network
59.8-74.7 ppb,
8-h max =
55.7-67.3 ppb
Kavnak et al.
(2013)
Ground-level
CMAQ compared
with fixed-site U.S.
EPA monitors from
AIRS, SEARCH,
and CASTNet;
vertical profiles of
CMAQ compared
with ICARTT data
which includes
aircraft, ship, and
ozonesonde data
Short-term
exposure
NR (shown
graphically)
CMAQ is
compared both
to surface data
and aloft data
from multiple
sources
Short-term
exposure during
only 2 mo may not
be indicative of
typical, long term
exposures
CMAQ to
observed for the
whole U.S., n =
1,267, R = 0.15,
MB = 9.191 ppb,
RMSE = 12.181
ppb, MNB =
34.77%, MNE =
37.38%; CMAQ
vs. ICARTT R2 =
0.51; obs vs.
CMAQ R2 = 0.15;
CMAQ vs.
observed, n =
363, R2 = 0.43,
MB = 5.457 ppb,
RMSE = 20.856
ppb, MNB =
10.49%, MNE =
29.93%
2-101
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Naan etal. (2012) CMAQ 4.6 in
L: Greater
Comparison with
Short-term
NR (shown
Nudging of the
Limitations in
Regular
nested models
Houston
monitors reporting
exposure
graphically)
meteorological
prediction of
forecasting: U R
at 36-, 12-,
area, TX;
to AQS
variables
precipitation may
= 0.49, RMSE =
and 4-km
T: August
improves ozone
cause error
1.60; V R = 0.68,
resolution with
23-
predictions
because
RMSE = 1.97
nudging of the
September
(based on
photolysis is
Retrospective
meteorological
9, 2006;
comparison with
influenced by
forecasting: U R
fields, called
P: Entire
monitors)
cloudiness
= 0.75, RMSE =
"retrospective
population
1.04; V R = 0.82,
forecasting" in
RMSE = 1.20
the paper
Weir et al. (2013) CMAQ model
L: U.S.;
Inverse-distance
Long-term
Annual
Observed data
The coarseness of
R = 0.66
on a 36-km
T: 2005-
weightings
exposure
observed ozone
were compared
the exposure
between
horizontal
2006;
compared with
(1 yr prior to
with CMAQ;
assessments may
observed and
resolution
P: NHANES
CMAQ
study participant
there is an epi
lose spatial
CMAQ
(unclear which
participants
examination) =
application
heterogeneity;
CMAQ version
(considered
37.5-60.3 ppb
inverse-distance
was used and
representa-
mean = 51.5
weighting is a
how CMAQ
tive of the
ppb median =
crude exposure
exposure
entire
52.0 ppb;
assignment
assignment
population)
annual CMAQ
method
happened)
ozone =
compared with
45.6-70.8 ppb
annual
mean = 57.2
averaged
ppb median =
concentrations
57.0 ppb
from observed
data from
fixed-site
monitors from
U.S. EPA's
AQS using
inverse-
distance
weighting of all
monitors
within 20 miles
2-102
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Reference
Model
Location,
Time Period,
Measurement
and
Evaluation
Population
Technique
L: Salt Lake
CMAQ and
City, Boise,
WRF-Chem have
Reno;
been validated in i
T: June-
previous
September
publications;
2000-2012;
multilinear modeled
L: Western
not validated, but
1 U.S.;
assessed
T:June 10-
July 10,
2008;
L: CONUS;
T: summer
2012;
P: Entire
population
Epidemiology
Applications
Concentrations
Measured
Strengths
Limitations
Exposure
Measurement
Errors
Jaffe etal. (2013)
CMAQ 4.5.1
with 36-km
horizontal
resolution
MM5 3.7 in
summer 2012
WRF-Chem
3.2 with 24-km
horizontal
resolution
between June
10 and July
10, 2008,
multilinear
relationship
between 8-h
daily max from
June-
September
between 2000
and 2012 for
Salt Lake City,
Boise, and
Reno using
observed data
only
Short- and
long-term
exposures
Obs n =
1,449-1,586,
obs min =
17.0-24.8 ppb,
obs max =
82-101.5 ppb,
obs mean =
50.9-55.8 ppb,
and obs SD =
8.4-11.0
Several different
data sources
used
(e.g., monitoring
data, CMAQ,
and
WRF-Chem),
variety of
timescales
explored
Modeled values
never directly
compared with
obs data; for the
short time
window, exposure
may not be
indicative of
longer termed
exposure
NR
2-103
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Chen et al. (2013)
WRF-Chem
L: Los
WRF-Chem
Short-term
Average for
Fine horizontal
Short-term
(CalNEX
3.1.1 4 km
Angeles
compared with
exposure
May 15-June 8,
resolution;
exposure may not
supersite) MB =
horizontal
basin, CA;
Caltech supersite,
2010 =
multiple
be indicative of
-10.6 ppb,
resolution
T: May 15-
CARB sites, aloft
9.1-62.7 ppb
observed data
typically
RMSE = 12.2
June 15,
data from NOAA
set used; both
exposures
ppb, R2 = 0.63;
2010;
WP-3D from flights
surface and aloft
by site MB =
P: Entire
on May 4, 14, 19,
ozone data
-14.6 to -1.1
population
20, and June 20
collected;
exploration of
NEI emissions
ppb, RMSE =
10.5-12.6 ppb,
R2 = 0.12-0.74;
(CalNEX
supersite) RMSE
= 12.22 ppb, R2 =
0.63
Choi (2014)
CMAQ 4.7.1
L: CONUS;
Baseline and NOx
Short-term
NR (shown
Entire CONUS
Short-term
Comparison of
over CONUS
T: August
satellite-adjusted
exposure
graphically)
covered; specific
exposure may not
means across
with 12-km
2009;
ozone compared to
input change of
be indicative of
five cities
horizontal
P: Entire
each other; each
emissions
typically
CMAQ-AQS =
resolution over
population
method compared
inventory and
exposures
-31.3 to -13.1%;
August 2009
with surface,
pinpoint ozone
CMAQ (with
with baseline
fixed-sited U.S.
differences;
satellite-based
emissions
EPA AQS data
comparison to
emissions) -AQS
compared with
observed data
= 9.6-38.1%
NOx satellite-
adjusted
emissions
2-104
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
PonaDrueksa
CMAQ 4.7.1,
L: CONUS;
CMAQ methods
Short- and
NR (shown
Several different
Horizontal
Surface observed
(2013)
WRF 3.4,
T: 2009;
were compared
long-term
graphically)
method
resolution is
ozone compared
CONUS with
P: Entire
with observed data
exposure
comparisons
coarse
with CMAQ, n =
36-km
population
from U.S. EPA's
used with two
26,234, MB = 7
horizontal
AQS and
types of
ppb, ME = 11
resolution for
ozonesonde data
modeled values
ppb, NMB =
2009
collected across
CONUS from
WOUDC, NOAA,
and TOPP
and two types of
observed data;
long-term
exposure from
CONUS is more
indicative of a
typical exposure
17%, NME =
28%, R = 0.53;
model
performance by
region 1-21,
CMAQ, MB = 1-
12 ppb, ME = 6-
13 ppb, NMB =
1-38%, NME =
14-40%, R =
0.30-0.62
2-105
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Karamchandani et
al. (2014)
CMAQ 5.01
L: Eastern
CMAQ APT Short-term
Mean in July
Specific CMAQ
with APT with
U.S.;
compared with exposure
above 40 ppb
component
12-km
T: January
CMAQ base and
was 55.4 ppb,
update to see a
horizontal
1-15, July
both models are
above 60 ppb
pointed
resolution
1-15, 2005;
compared with
was 70.2;
difference; APT
P: Entire
fixed-site surface
around point
differences
population
monitors from U.S.
sources in July
explored around
EPA's AQS sites
above 40 ppb
was 55.1 ppb,
above 60 ppb
was 70.7 ppb;
around point
sources in July
above 40 ppb
was 56.1 ppb,
above 60 ppb
was 70.3 ppb
point sources
Short-term
exposure may not
be indicative of a
typical exposure
in epi studies
In July with
40 ppb cut off
CMAQ, n =
49,765, mean
obs = 55.4 ppb,
mean mod = 53.9
ppb, ratio of
means = 0.97,
GB = -1.5 ppb,
NB = -1.5%, FB
= -4.4%, GE =
0.4 ppb, NE =
17.4%, FE =
18.4%, NMB =
-2.7%, NME =
16.9%, R2 = 0.30,
CMAQ APT, n =
49,765, mean
obs = 55.4 ppb,
mean mod = 53.9
ppb, ratio of
means = 0.97,
GB = -1.5 ppb,
NB = -1.4%, FB
= -4.4%, GE =
0.4 ppb, NE =
17.4%, FE =
18.4%, NMB =
-2.7%, NME =
16.9%,
2-106
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Karamchandani et
al. (2014)
(continued)
CMAQ 5.01
with APT with
12-km
horizontal
resolution
(continued)
L: Eastern
CMAQ APT
Short-term
Mean in July
Specific CMAQ
R2 = 0.30; in July
U.S.;
compared with
exposure
above 40 ppb
component
with 40 ppb cut
T: January
CMAQ base and
(continued)
was 55.4 ppb,
update to see a
off with 5x5 grid
1-15, July
both models are
above 60 ppb
pointed
CMAQ, n =
1-15, 2005;
compared with
was 70.2;
difference; APT
2,791, mean obs
P: Entire
fixed-site surface
around point
differences
= 55.1 ppb, mean
population
monitors from U.S.
sources in July
explored around
mod = 55.4 ppb,
(continued)
EPA's AQS sites
above 40 ppb
point sources
ratio of means =
(continued)
was 55.1 ppb,
(continued)
1.01, GB = 0.3
above 60 ppb
ppb, NB =
was 70.7 ppb;
-2.1%, FB =
around point
-1.3%, GE =
sources in July
10.0 ppb, NE =
above 40 ppb
19.0%, FE =
was 56.1 ppb,
19.6%, NMB =
above 60 ppb
0.5%, NME =
was 70.3 ppb
18.2%, R2 = 0.22,
(continued)
CMAQ APT, n =
2,791, mean obs
= 55.1 ppb, mean
mod = 55.2 ppb,
ratio of means =
1.00, GB = 0.1
ppb, NB =
-1.8%, FB =
-1.6%, GE = 9.9
ppb, NE =
18.7%, FE =
19.4%, NMB =
0.1%, NME =
18.0%,
2-107
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Karamchandani et
CMAQ 5.01
L: Eastern
CMAQ APT
Short-term
Mean in July
Specific CMAQ
R2 = 0.22, in July
al. (2014)
with APT with
U.S.;
compared with
exposure
above 40 ppb
component
with 40 ppb cut
(continued)
12-km
T: January
CMAQ base and
(continued)
was 55.4 ppb,
update to see a
off with 9x9 grid
horizontal
1-15, July
both models are
above 60 ppb
pointed
CMAQ n = 7,197,
resolution
1-15, 2005;
compared with
was 70.2;
difference; APT
mean obs = 56.1
(continued)
P: Entire
fixed-site surface
around point
differences
ppb, mean mod =
population
monitors from U.S.
sources in July
explored around
55.7 ppb, ratio of
(continued)
EPA's AQS sites
(continued)
above 40 ppb
was 55.1 ppb,
above 60 ppb
was 70.7 ppb;
around point
sources in July
above 40 ppb
was 56.1 ppb,
above 60 ppb
was 70.3 ppb
(continued)
point sources
(continued)
means = 0.99,
GB = -0.4 ppb,
NB = 0.7%, FB =
-2.5%, GE = 9.7
ppb, NE =
18.2%, FE =
18.9%, NMB =
-0.8%, NME =
17.4%, R2 = 0.27,
CMAQ APT, n =
7,197, mean obs
mod = 55.6 ppb,
ratio of means =
0.99, GB = -0.5
ppb, NB = 0.5%,
FB = -2.6%, GE
= 9.7 ppb, NE =
18.1%, FE =
18.8%, NMB =
-0.9%, NME =
17.2%, R2 = 0.27;
Difference in
methods of 8-h
daily max ozone
for selected days
-10 to 10 ppb
2-108
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured Strengths Limitations
Errors
Zhana et al.
CMAQ 4.7.1
L: CONUS,
CMAQ output
Long-term
NR (shown This paper has Limited obs for
Monthly mean
(2014)
from 2000-
eastern
compared with
exposure
graphically) fine-scale performance
ozone for
2006, 36-km
U.S., seven
surface, fixed-site
(through the
resolution over a evaluation
selected cities
horizontal
U.S. cities in
monitoring data
MESA and
sizeable
MNB = -0.4 to
resolution over
eastern U.S.
from U.S. EPA's
WHI-OS
geographical
0.4 ppb, NGE =
CONUS, WRF
(NYC,
AQS
studies)
area for a long
0.1 to 0.35 ppb,
3.2.1, 12-km
Pittsburgh,
time period
AUP = -0.6 to
horizontal
Baltimore,
0.4 ppb
resolution over
Chicago,
the eastern
Detroit, St.
U.S., 4-km
Paul,
horizontal
Winston-
resolution over
Salem);
seven cities
T: Hourly
(NYC,
ozone
Pittsburgh,
between
Baltimore,
2000 and
Chicago,
2006;
Detroit, St.
P: Entire
Paul, Winston-
population
Salem)
2-109
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Reference
Model
Location,
Time Period,
and
Population
Measurement
Evaluation
Technique
Epidemiology
Applications
Concentrations
Measured Strengths
Limitations
Exposure
Measurement
Errors
Hoarefe et al.
Model
L: North
Comparison with
Short-term
NR This study
Not all data are
(Across synoptic
(2014)
comparison
America and
measurements by
exposure
compares
easily discernible,
patterns) RMSE
including
Europe;
calculating MBE
several different
as presented in
NE = 8.2-12.8
CHIMERE
T: May 1-
and RMSE in
models and
the paper
ppb MW= 9.3-
(36 km),
September
comparison with
presenting the
14.5 ppb SE =
DEHM
30, 2006;
monitors in eight
comparison by
10.5-13.1 ppb
(50 km),
P: Entire
synoptic regions
meteorological
NW= 8.7-11.9
CAMx (12, 15,
population
(Northeast,
zone
ppb CA =
and 24 km),
Midwest,
10.8-15.2 ppb
CMAQ (12,
Southeast,
SW= 10.2-10.9
18, 24 km),
Northwest,
ppb NEu =
AURAMS
California,
8.0-15.1 ppb
(45 km),
Southwest,
SEu = 9.7-12.4
Polair3D
Northern Europe,
ppb; MB NE =
(24 km),
Southern Europe)
-5.8 to -0.8 ppb
MUSCAT
MW = -6.3 to 2.0
(24 km),
ppb SE = -9.1 to
SILAM
-4.4 ppb NW =
(24 km),
-5.1 to -2.1 ppb
EMEP
CA = -3.9 to
(50 km),
-1.9 ppb SW =
LOTOS/
1.8 to 3.0 ppb
EUROS
NEu = 0.7 to 7.2
(25 km);
ppb SEu = 0.2 to
versions not
4.0 ppb
reported
2-110
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period, Measurement Exposure
and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Yahva et al.
(2014)
WRF/Chem-
MADRID using
WRF/Chem
3.0, CB05,
12-km
horizontal
resolution over
the
southeastern
U.S.
comparing
biogenic
emission from
MEGAN2,
satellite-
derived fire
emissions
(SD-Fire), and
MEGAN2+
SD-Fire
L: Eastern
U.S. states;
T: May to
September
2009-2011,
December to
February
2009-2012,
sensitivity
analysis
performed
July 2011;
P: Entire
population
WRF/Chem-
MADRID compared
with surface,
fixed-site monitors
from AQS,
CASTNet,
IMPROVE, and
SEARCH
Long-term
exposure
Avg 8-h daily
max during
ozone season =
41.4-48.6 ppb;
avg 8-h daily
max during
winter = 31.1—
36.0 ppb
Long term
exposure is
more indicative
of typical
exposes;
sensitivity
analysis was
extensive
Forecasting daily
ozone only 1-day
forward is not an
ideal exposure
methodology in an
epi setting
Between model
and obs during
ozone season
1-h max R = 0.3
to 0.7, NMB =
-6.0 to 15.5%,
NME = 17.6-
27.1%, 8-h max
R = 0.4 to 0.7,
NMB = -4.5 to
14.6% NME =
17.8-26.1%;
sensitivity
analysis for July
2011 for 1-h max
R = 0.5, NMB
-0.9 to 10.1%,
NME = 20.6-
24.2%, and 8-h
max R = 0.5-0.6,
NMB = 1.6-
12.8%, NME =
20.7-25.4%;
sensitivity
analysis for July
2011 for 1-h max
accuracy = 85.9-
91.5%, bias =
0.9-2.2, CSI =
15.6-19.1%,
FAR = 67.7-
78.3%, POD =
25.0-46.6%,
2-111
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Yahva et al.
WRF/Chem-
L: Eastern
WRF/Chem-
Long-term
Avg 8-h daily
Long term
Forecasting daily
8-h daily max
(2014) (continued)
MADRID using
U.S. states;
MADRID compared
exposure
max during
exposure is
ozone only 1-day
accuracy = 66.5-
WRF/Chem
T: May to
with surface,
(continued)
ozone season =
more indicative
forward is not an
74.0%, bias =
3.0, CB05,
September
fixed-site monitors
41.4-48.6 ppb;
of typical
ideal exposure
1.1-1.7, CSI =
12-km
2009-2011,
from AQS,
avg 8-h daily
exposes;
methodology in an
27.4-30.6%,
horizontal
December to
CASTNet,
max during
sensitivity
epi setting
FAR = 58.8-
resolution over
February
IMPROVE, and
winter = 31.1—
analysis was
(continued)
63.3%, POD =
the
2009-2012,
SEARCH
36.0 ppb
extensive
44.2-62.1%
southeastern
sensitivity
(continued)
(continued)
(continued)
U.S.
analysis
comparing
performed
biogenic
July 2011;
emission from
P: Entire
MEGAN2,
population
satellite-
(continued)
derived fire
emissions
(SD-Fire), and
MEGAN2+
SD-Fire
(continued)
Li et al. (2014a)
WRF-CHEM
L: California-
WRF-Chem
Short-term
NR (shown
High spatial
Only having a few
Obs and mod by
using CMAQ
Mexico
compared with
exposure
graphically)
resolution
days of modeled
station MB = -7.6
4.6 and
border
surface, fixed-site
concentration
to 13.5 ppb, R2 =
ISORROPIA
region;
monitors
during an ozone
0.20-0.79,
1.7 horizontal
T: May 15-
episode is not
RMSE =
resolution of
16, May 90-
indicative of more
17.1-22.0 ppb
2 km
30, June 4-
5, June 13-
14, 2010;
P: Entire
population
typical long-term
exposures
2-112
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Pan et al. (2014)
NAQFC-beta
L: CONUS;
Modeled output
Short-term
Average hourly
Complete cover
Ozone
Bias U.S.
(coupled
T: July 2011;
compared with
exposure
ozone by hour
of CONUS;
concentrations in
emissions urban
NAM-CMAQ
P: Entire
surface, fixed-site
NR (shown
incremental
1 summer mo is
= 7.08 ppb,
4.7.1) with
population
monitors from U.S.
graphically)
change of model
not indicative of
suburban =
mobile
EPA's AQS and
inputs pinpoints
more long-term
7.48 ppb, rural =
sources
CTM with
differences
exposures
7.80 ppb; U.S. +
defined from
base-case
between two
Canadian
2005
emissions
different models
emissions urban
MOBILE6 + 05
= 6.16 ppb,
to 12
suburban =
projections,
6.22 ppb, rural =
point sources
5.93 ppb
from 2010
CEM + DOE
Annual Energy
Outlook,
nonroad from
CSAPR,
Canadian
emissions
from 2006 El,
with 12-km
horizontal
resolution over
CONUS
2-113
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period, Measurement Exposure
and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Cuchiara et al.
(2014)
ARW-WRF, L: Houston,
WRF/Chem TX;
3.5 with 4-km T: October
horizontal 5,2006;
resolution, P: Entire
with four population
planetary
boundary layer
schemes from
YSU, MJY,
ACM2, QNSE.
YSU, and
ACM2 are
computed
based on the
bulk
Richardson
number, which
is the ratio of
buoyancy to
turbulence
caused by
shear
stresses. MJY
and QNSE are
computed
based on
eddy-
diffusivity, or
atmospheric
mixing
All four PBL
schemes compared
to each other and
to observed,
fixed-site monitors
from U.S. EPA's
CAMS and aloft
observed data from
ozonesonde and
aircrafts
Short-term
exposure
NR (shown
graphically)
Very fine-scale
resolution;
small,
incremental
changes in
model pinpoints
model
differences and
assumptions
Highly localized
space/time
modeling scenario
is not indicative of
more long-term
exposures
Statistics across
sites for four
boundary layer
schemes: YSU R
= 0.79-0.92, bias
= 0.59-0.99,
RMSE = 13.20-
21.02 ppbv; MYJ
R = 0.70-0.90,
bias 0.64-1.05,
RMSE = 12.17-
20.76 ppbv;
ACM2 R = 0.37-
0.77, bias =
0.75-1.26,
RMSE = 18.53-
25.77 ppbv;
QNSE R = 0.54-
0.71, bias =
0.72-1.09,
RMSE = 15.59-
24.85 ppbv
2-114
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Thompson and
CAMx 4.5.3
L: Houston,
All modeled output
Short-term
Population-
Very fine-scale
Limited temporal
Average MNGE
Selin (2012)
with 36-, 12-,
Galveston,
compared to each
exposure
weighted
horizontal
run and spatial
across sites
4-, and 2-km
Brazoria
other and to
maximum ozone
resolution of the
coverage of the
36-km resolution
horizontal
area, TX;
surface, fixed-site
across days; NR
model
model may not be
= 74%, 12-km
resolution
T: August
monitors for air
(shown
indicative of
resolution = 63%,
13-
quality monitors in
graphically)
long-term
4-km resolution =
September
the region
exposures
26%, 2-km
15, 2006;
resolution = 25%
P: Entire
population
Lu et al. (2014)
CMAQ 5.0,
L: Mid-
CMAQ compared
Short-term
Hourly ozone
Fine scale
2 mo of short-term
Hourly ozone in
SMOKE 3.0,
southern
against surface,
exposure
NR (shown
resolution;
exposure is not
July
WRF 3.3,
U.S.
observed fixed-site
graphically)
CMAQ has
necessarily
NMB = 48.2%,
CB05 with
(Mississippi,
monitors from U.S.
complete
indicative of
RMSE = 20.9
4-km
Arkansas,
EPA's AQS
coverage in
typical exposures
ppb, UPA = 28%,
horizontal
Tennessee,
spatial domain
R = 0.67
resolution
Alabama);
T: July 2011,
February
2012;
P: Entire
population
Wanq and Zhanq
CMAQ 4.7
L: Eastern
All modeled runs
Short-term
Mean in
There are
2 mo of short-term
January 2002:
(2014)
with 12-km
U.S.;
compared against
exposure
January 2002
several different
exposure is not
8-h daily max
horizontal
T: January
each other and
for 8-h daily
comparisons
necessarily
ozone NMB =
resolution with
and July
modeled values
max ozone =
modeled
indicative of
-1.6 to 3.9%,
offline dry
2002;
compared against
between 26.8
methods
typical exposures
NME = 19.4-
deposition,
P: entire
surface, observed
and 29.2 ppb
21.7%, R = 0.70-
inline dry
population
fixed-site monitors
0.74; July 2002:
deposition,
from CASTNet,
8-h daily max
four difference
IMPROVE, AQS,
ozone NMB =
sensitivity
SEARCH, NADP,
-2.2—4.3%, NME
analysis with
and NC DENR
= 15.4-16.8%, R
the inline dry
= 0.75-0.77
deposition
2-115
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period, Measurement Exposure
and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Xing etal. (2015) CMAQ 5.0
with 108-km
horizontal
resolution,
EDGAR 4.2,
EDGAR HTAP
1
Summer: R =
0.59 MB =
-14.3 |jg/m3
NMB = -8.1%
RMSE =
30.5 ug/m3 NME
= 14.5%;
Fall: R = 0.60 MB
= -3.9 |jg/m3
NMB = -2.5%
RMSE =
23.5 |jg/m3 NME
= 12.4%;
Winter R = 0.51
MB = -3.6 |jg/m3
NMB = -3.2%
RMSE =
10.1 |jg/m3 NME
= 7.6%
L: Northern
Hemisphere;
T: 1990-
2010;
P: Entire
population
CMAQ compared
to surface,
observed, fixed-site
monitors from
AQS, CASTNet,
IMPROVE (U.S.),
EMEP, AIRBASE
(Europe), API
(China), and
WDCGG (global)
Short- and
long-term
exposure
No figures or
tables showing
concentration of
ozone
Excellent spatial
coverage and
temporal
coverage
Coarse resolution
across the U.S.
Comparison of
model with
CASTNet
network spring: R
= 0.52 MB =
-22.8 |jg/m3
NMB = -13.6%
RMSE =
29.7 ug/m3 NME
= 16.1%;
2-116
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Herron-Thoroe et
CMAQ 4.6,
L: Pacific
CMAQ compared
Short-term
8-h daily max
Complete
Short time
8-h daily max
al. (2014)
EGAS,
Northwest,
with surface,
exposure
ozone before
coverage with
windows may not
ozone before
MOBILE 6.2,
U.S.;
fixed-site observed
(after) smoke
CTM; multiple
be indicative of
(after) smoke
GVRD,
T: July 3-
data from U.S.
event mean obs
data sources
typical exposures
event R = 0.7
BEIS-3,
August
EPA's AQS
= 45.8 ppb
used to model
(0.8), MB = -4.7
SMOKE 2.4,
2007, June
(42.3 ppb)
ozone
ppb (-0.7 ppb),
BlueSky 3.1
22-August
ME = 8.9 ppb
with 12-km
27, 2008;
(7.7 ppb), NMB =
horizontal
P: Entire
-7% (3%), NME
resolution with
population
= 20 ppb (21
AIRPACT-3
ppb), FB = -10%
FEPS Plume
(-1%), FE = 22%
Rise
(20%)
compared with
AIRPACT-3
SMOKE
Plume Rise
compared with
MOZART-4
Tana et al.
CAMx 5.3,
L: Eastern
Photolysis from
Short-term
Monthly avg 8-h
Incremental
Short time
Difference in
(2015a)
MM5 3.7.3,
TX;
GEOS is compared
exposure
daily max ozone
model changes
windows may not
modeled ozone
MOZART,
T: August
with Texas SIP and
NR (shown
demonstrate the
be indicative of
by day R2
emissions for
13-
both modeling
graphically)
specific
typical exposures
between = -0.06
HGB SIP from
September
methods are
influence of
to 0.07, NMB =
TCEQ, NLDN,
15, 2006;
compared with
photolysis;
between -0.1 to
comparison of
P: Entire
surface, fixed-site
multiple
0.04%, NME =
clouds with
population
monitors from U.S.
comparison
between -0.1 to
GEOS vs.
EPA's AQS
methods with
0.02%
Texas SIP
multiple model
with 12-km
comparison and
horizontal
observed data
resolution
comparisons
2-117
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period, Measurement Exposure
and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Tessum et al.
(2015)
WRF-Chem
3.4 with 12-km
horizontal
resolution
L: CONUS;
T: 2005;
P: Entire
population
Model compared
with surface,
fixed-site monitors
from U.S. EPA's
CASTNet and AQS
Short- and
long-term
exposure
Annual ozone,
annual peak
ozone, annual
daytime ozone
NR (shown
graphically)
Comparison to
observed data;
full spatial
coverage of
domain; having
a year's worth of
data make the
study applicable
to long-term
exposures
Short-term
exposures were
not explored
Annual average
MB = 7.92 ppb,
ME = 8.58 ppb,
MFB = 23%,
MFE = 26%, R2 =
0.37; errors in
average daytime
ozone per
season of
WRF-Chem
(CMAQ) winter
MB = 3.5 ppb
(-3.5 ppb), ME =
5.5 ppb (9.0
ppb), NMB =
12% (-13%),
NME = 19%
(35%), spring MB
= 1.5 ppb (-1.8
ppb), ME = 4.6
ppb (9.3 ppb),
NMB = 3%
(-4%), NME =
10% (29%),
summer MB =
9.2 ppb (4.4
ppb), ME = 10.1
ppb (11.0 ppb),
NMB = 21%
(10%),
NME = 23%
(24%), fall MB =
5.2 ppb (2.6
ppb), ME = 6.2
ppb (8.8 ppb),
NMB = 19%
(8%), NME =
23% (28%)
2-118
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Hoarefe et al.
(2015)
CMAQ 5.0.1,
WRF 3.4 with
12-km
horizontal
resolution
L: CONUS;
T: June to
August
2006, May to
September
2010;
P: Entire
population
Model compared
with surface,
fixed-site monitors
from U.S. EPA's
AQS
Short- and
long-term
exposure
Mean during
2006 of 8-h
daily max ozone
= 32.3-51.1
ppb; mean
during 2010 of
8-h daily max
ozone = 33.6-
47.5 ppb
CMAQ has full
coverage of
spatial domain
Number of months
and different
years explored
lends itself to both
short- and
long-term
exposures
Avg 8-h daily
max ozone
June-August
2006: MB = -0.9
to 6.6 ppb, ME =
5.6-10.9 ppb,
RMSE = 7.4-
14.4 ppb, NMB =
-1.9 to 20.5%,
NME = 13.8-
26.6%, R = 0.69-
0.78; May-
September 2010:
MB = -1.9 to 6.6
ppb, ME = 6.6-
9.7 ppb, RMSE =
8.7-12.2 ppb,
NMB = -5.6 to
14.9%, NME =
13.9-21.8 %, R =
0.58-0.78
Tana et al.
(2015b)
CMAQ alone
(base case)
L: CONUS;
T: July 2011;
P: Entire
population
Modeled outputs
compared with
AirNow observed
data and aircraft
measurements
from Discover-AQ
Short-term
exposure
Hourly ozone in
the first half of
July 2011 in the
northeastern
U.S.
Two different
observed data
sources used;
multiple models
compared to
each other
1 summer mo
may not be
indicative of more
long-term
exposures
Hourly ozone
from July 6-7,
2011 over
CONUS R =
0.53, MB = 2.54;
hourly ozone
from July 6-7,
2011 over
southeastern
U.S. R = 0.55,
MB = 0.22; R
between obs and
CMAQ alone for
aircraft data is
0.604
2-119
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Reference
Model
Location,
Time Period,
and
Population
Measurement
Evaluation
Technique
Epidemiology
Applications
Concentrations
Measured
Strengths
Limitations
Exposure
Measurement
Errors
Koo et al. (2015)
CAMx 5.4.1
L: Eastern
Both models
Long-term
Ozone
The shows
CTMs have
CMAQ: all NMB
compared with
U.S.;
compared to each
exposure
concentrations
models
inherent error
were within
CMAQ
T: 2005;
other and models
not displayed in
sensitivities in
±15%, all NME
5.0.1-VBS with
P: Entire
compared to
table or figure
two or the most
were within 25%;
a 12-km
population
surface, fixed-site
for 2005
common
CAMx: all NMB
horizontal
monitors from U.S.
models; multiple
were within
resolution
EPA's AQS,
comparison
±25%, NME were
CASTNet, and
method between
within 30%
SEARCH
models and
observed data
Yahva et al.
(2015b)
WRF/Chem
L: CONUS;
Model compared
3.4.1 with
T: January,
with surface,
36-km
February,
observed fixed-site
horizontal
December
monitors from U.S.
resolution for
2006 and
EPA's CASTNet
2006 and
2010 with
and AQS
2010
June, July,
August 2006
and 2010;
P: Entire
population
Short- and
long-term
exposure
Mean 1-h max
ozone =
between 33.2
and 48.4 ppb,
mean 8-h daily
max ozone =
between 32.7
and 43.8 ppb
Comparison
method to obs
data; time
window lends
itself to both
short-and
long-term
exposure
CTMs have
inherent errors
and horizontal
resolution is
coarse
R 1-h daily max
CASTNet = 0.40,
AQS = 0.34; R =
8-h daily max
CASTNet = 0.40,
AQS = 0.20;
NMB 1-h daily
max CASTNet =
-30.0%, AQS =
-15.8%; NMB =
8-h daily max
CASTNet =
-25.3%, AQS =
-17.0%; NME =
1-h daily max
CASTNet = 34.8
and AQS =
28.0%; NME =
8-h daily max
CASTNet =
32.0%, AQS =
29.2%
2-120
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Pan et al. (2015)
CMAQ 5.0.1,
L: Southeast
Models compared
Short-term
Mean obs =
Fine scale
Short-term
Hourly ozone, R
WRF 3.5 base
TX;
to each other and
exposure
24.4 ppb, mean
resolution;
exposure may not
= 0.73, IOA =
compared with
T:
both models
mod = 32.7 ppb,
sensitivity
be indicative of
0.80, MB = 8.3
sensitivity
September
compared to
mean sensitivity
analysis of the
typical ozone
ppb, R of model
analysis of
2013;
TCEQ's CAMS
analysis = 33.2
model
exposures
difference = 0,
adjusted
P: Entire
surface, fixed-site
ppb
IOA model
emissions with
population
monitors
different = 0, MB
4-km
model difference
horizontal
= 0.4 ppb
resolution
Barrett et al.
GEOS-Chem
L: U.S.,
Comparison with
Long-term
Average
Nationwide
Low spatial
Mean NMB =
(2015)
model (version
T: 2008-
data from fixed-site
exposure
addition of
model,
resolution
25.3%, SD of
not reported)
2015,
monitors reporting
study
2.6 ppbv across
1,200 monitoring
(50 km), model
NMB = 17.9%
with adjoint
P: Entire
to the AQS
(contribution
U.S. due to
sites used for
based on 2005
(NMB calculated
population
from VW
emissions)
excess NOx
emissions
validation
emissions
inventory
(emissions have
dropped over
time)
from 1-h daily
max)
2-121
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Fribera et al.
CMAQ 4.5
L: Georgia;
Cross-validation by
Long-term
Mean (IQR) =
Low mean bias,
Errors in
CMAQ MFE =
(2016)
alone and
T: 2002-
fixed-site monitors
exposure
47.6 ppb
low RMSE, and
measurements
0.18, MFB =
annual
2008,
study
(22.0 ppb) for
relatively high
used as input are
0.11, NME =
adjustment
P: Entire
8-h daily max
R2 (compared to
propagated into
16.5%, NMB =
population
application of
the model, limited
8.58%, MB =
the model for
spatial coverage
0.004, RMSE =
other pollutants),
of monitors
0.01, R2
and errors are
increases errors
(cross-validation)
minimized
(although this is
= 67.1%; CMAQ
through the
less of a limitation
with annual
model-weighting
for ozone and
adjustment MFE
approach
other secondary
= 0.17, MFB =
pollutants)
0.03, NME =
15.0%, NMB =
0.14%, MB =
6.9 x 10"5, RMSE
= 0.01, R2
(cross-validation)
= 67.2%
Tao et al. (2016)
NU-WRF
L:
Comparison with
Short- and
Mean = 30-
Enabled
27-km resolution
(Mean, range)
model,
Contiguous
data from fixed-site
long-term
50 ppbv for
analysis of the
may lead to bias
NB = -2.8%
focused on
U.S.;
monitors reporting
exposure
3-mo avg;
influence of
because not all
(-28.2, 28.0%),
impact of
T: March
to the AQS
impact of
meteorology and
cloud chemistry
NGE = 18.8%
trans-Pacific
21-June 30,
transpacific PM
Asian air
can be
(11.9, 29.7%)
aerosol
2010;
on ozone
pollution on U.S.
represented,
transport
P: Entire
concentrations
ozone
study did not
population
= -0.33 ppbv to
concentrations,
examine the
0.50 ppbv
low mean bias
model's internal
variability
2-122
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Li etal. (2016a) WRF-Chem
L: Central
Short- and NR
Inclusion of
Planetary
Irrigation not
3.6.1 with
Valley, CA;
long-term
irrigation and
boundary layer
included: MB =
nested 36-,
T: June 23-
exposure
cloud cover on
designation in the
-6.1 ppb, NMB =
12-, and 4-km
August 1,
model output;
model may be
-24.6%, NME =
domains
2005 (first
the nested
uncertain
28.9%, MNB =
few days
model design
(different
-23.9%, MNGE =
considered
includes 4-km
approaches have
28.3%, R = 0.63,
spin-up);
resolution, which
been used in
IOA: 0.81, RMSE
P: Entire
is sufficiently
different studies).
= 18.0 ppb;
population
fine to capture
irrigation
(6.5 million
dynamics in
inclusion: MB =
residents)
rural settings
-5.4 ppb, NMB =
-21.1%, NME =
26.0%, MNB =
-21.5%, MNGE =
26.1%, R = 0.70,
IOA: 0.83, RMSE
= 17.6 ppb
2-123
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Reference
Model
Location,
Time Period,
and
Population
Measurement
Evaluation Epidemiology
Technique Applications
Concentrations
Measured
Strengths
Limitations
Exposure
Measurement
Errors
Yahva et al.
WRF-Chem
L:
Comparison with Long-term
(Model values at
Extensive
Lower resolution
2001 CASTNet
(2015a)
3.4.1 at 36-km
Contiguous
data from fixed-site exposure
sites for
comparisons
(36 km)
8-h max: MB =
resolution with
U.S.;
monitors reporting
networks
made at
-4.8 ppb, NMB =
initial and
T: 2001,
to the AQS and
mentioned)
different time
-11.0%, NME =
boundary
2006, 2010;
CASTNet
2001 CASTNet
averages and
28.2%, CASTNet
conditions
P: Entire
8-h max =
validation data
1-h max = MB:
downscaled
population
39.0 ppb,
sets, validation
-7.9 ppb, NMB =
from global
CASTNet 1-h
on
-17.2%, NME =
models
max = 38.3 ppb,
meteorological
30.1%, AQS 8-h
AQS 8-h max =
variables as well
max = MB: -0.3
44.6 ppb, AQS
ppb, NMB =
1-h max =
-0.7%, NME =
49.4 ppb; 2006
29.9%, AQS 1-h
CASTNet 8-h
max MB = -1.7
max = 38.4 ppb,
ppb, NMB =
CASTNet 1-h
-3.3%, NME =
max = 39.3 ppb,
28.5%; 2006
AQS 8-h max =
CASTNet 8-h
43.2 ppb, AQS
max: MB =
1-h max = 48.1
-5.2 ppb, NMB =
ppb; 2010
-11.8%, NME =
CASTNet 8-h
27.1%, CASTNet
max = 38.2 ppb,
1-h max: MB =
CASTNet 1-h
-8.3 ppb, NMB =
max = 38.6 ppb,
-17.4%, NME =
AQS 8-h max =
28.7%, AQS 8-h
41.8 ppb, AQS
max:
1-h max =
47.3 ppb
2-124
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Reference
Model
Location,
Time Period,
and
Population
Measurement
Evaluation
Technique
Epidemiology
Applications
Concentrations
Measured
Strengths
Limitations
Exposure
Measurement
Errors
Yahva et al.
WRF-Chem
L:
Comparison with
Long-term
(Model values at
Extensive
Lower resolution
MB = -1.2 ppb,
(2015a)
3.4.1 at 36-km
Contiguous
data from fixed-site
exposure
sites for
comparisons
(36 km)
NMB = -2.8%,
(continued)
resolution with
U.S.;
monitors reporting
(continued)
networks
made at
(continued)
NME = 27.5%,
initial and
T: 2001,
to the AQS and
mentioned)
different time
AQS 1-h max MB
boundary
2006, 2010;
CASTNet
2001 CASTNet
averages and
= -2.2 ppb, NMB:
conditions
P: Entire
(continued)
8-h max =
validation data
= -4.5%, NME =
downscaled
population
39.0 ppb,
sets, validation
26.3%; 2010
from global
(continued)
CASTNet 1-h
on
CASTNet 8-h
models
max = 38.3 ppb,
meteorological
max: MB =
(continued)
AQS 8-h max =
variables as well
-5.7 ppb, NMB =
44.6 ppb, AQS
(continued)
-13.0%, NME =
1-h max =
26.9%, CASTNet
49.4 ppb; 2006
1-h max: MB =
CASTNet 8-h
-8.8 ppb, NMB =
max = 38.4 ppb,
-18.6%, NME =
CASTNet 1-h
28.7%, AQS 8-h
max = 39.3 ppb,
max: MB = -0.4
AQS 8-h max =
ppb, NMB =
43.2 ppb, AQS
-1.1%, NME =
1-h max = 48.1
26.1%, AQS 1-h
ppb; 2010
max MB = -1.1
CASTNet 8-h
ppb, NMB =
max = 38.2 ppb,
-2.3%, NME =
CASTNet 1-h
25.3%
max = 38.6 ppb,
AQS 8-h max =
41.8 ppb, AQS
1-h max =
47.3 ppb
(continued)
2-125
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Reference
Model
Location,
Time Period,
and
Population
Measurement
Evaluation
Technique
Epidemiology
Applications
Concentrations
Measured
Strengths
Limitations
Exposure
Measurement
Errors
Li et al. (2014b)
WRF-Chem
L: Phoenix,
Comparison with
Short-term
Hourly ozone
Late afternoon
Simulations
June 9, 2011: MB
3.5.1 with
AZ;
24 fixed-site
exposure
NR (shown
heat island
overestimated
= -1.69 ppb,
nested 36-,
T: June 9,
monitors
graphically)
captured well
wind speed
RMSE = 14.70
12-, 4-, and
2011 and
ppb, NMB =
1-km domains
May 14,
-6.32%, NME =
2012;
15.32%, MNB =
P: Entire
-5.59%, MNGE =
population 15.70%, IOA =
0.80, R = 0.75;
May 14, 2012:
MB = -1.50 ppb,
RMSE = 14.75
ppb, NMB =
-6.50%, NME =
14.43%, MNB =
-5.60%, MNGE =
15.76%, IOA =
0.81, R = 0.74
CMAQ: April
2006: RMSE =
9.51 ppb, MAE =
7.33 ppb, MB =
3.94 ppb; August
2006: RMSE =
12.80 ppb, MAE
= 9.70 ppb, MB =
4.84 ppb;
October 2006:
RMSE = 10.10
ppb, MAE = 8.20
ppb, MB = 5.34
ppb
Ran et al. (2016)
CMAQ
L:
5.0.2/WRF 3.4
Contiguous
with MODIS
U.S.,
leaf area index
southern
model
Canada,
included in
northern
some runs
Mexico;
T: April,
August,
October
2006;
P: Entire
population
Comparison with
data from fixed-site
monitors reporting
to the AQS
Long-term
exposure
CMAQ: April
2006 = 52.70
ppb, August
2006 = 55.00
ppb, October
2006 = 42.2.0
ppb; CMAQ +
MODIS: April
2006 = 55.40
ppb, August
2006 = 57.10
ppb, October
2006 = 44.60
ppb
Addition of leaf
area index
allows for
consideration of
role of
vegetation
12-km resolution
2-126
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Godowitch et al.
CMAQ 5.0.2
L: Eastern
Comparison with
Short- and
Eastern U.S.:
Use of
12-km resolution
Eastern U.S.:
(2015)
with 12-km
U.S.;
data from fixed-site
long-term
meteorology
continually
still inhibits urban
meteorology
resolution with
T: June 1-
monitors reporting
exposure
Model 1 = 50.1
updated data
studies
Model 1 MB =
WRF/FDDA
August 31,
to the AQS, tower
ppbv, Model 2 =
improves
9.8 ppbv, MAE
meteorology
2002;
sensors at one
48.2 ppbv;
accuracy of
13.5 = ppb, Fp
and boundary
P: Entire
location (Raleigh,
northeastern
model
42%, Model 2
conditions
population
NC), and
U.S.: Model 1 =
MB = 7.9 ppbv,
from a global
DISCOVER-AQ
49.6 ppbv,
MAE = 12.6 ppb,
GEOS-Chem
flight sensors
Model 2 = 47.5
Fp 58%;
simulation
ppbv; CASTNet:
Model 1 = 52.9
ppbv, Model 2 =
51.1 ppbv
northeastern
U.S.: Model 1:
MB = 8.2 ppbv,
MAE = 12.4 ppb,
Fp = 41%, Model
2: MB = 6.1
ppbv, MAE =
11.5 ppb, Fp =
59%; CASTNet:
Model 1: MB =
10.8 ppbv, MAE
= 13.3 ppb, Fp =
42%, Model 2:
MB = 9 ppbv,
MAE = 12.5 ppb,
Fp = 58%
Li et al. (2016b)
CMAQ 5.0.2
L: Southeast
Comparison with Short-term
Without
Data
4-km resolution
Without
with
TX,
data from fixed-site exposure
assimilation
assimilation
misses spatial
assimilation: IOA
WRF/FDDA
southwest
monitors reporting
mean = 33.7
improves
variation
= 0.78, RMSE =
meteorological
LA;
to the AQS
ppb, SD = 14.1
representation
14.9 ppb, MAE =
model and
T:
ppb; with
of short-term
12.3 ppb, MB =
assimilation of
September
assimilation
variability in
9.3 ppb; with
meteorological
2013;
mean = 30.6
concentration
assimilation: IOA
data
P: Entire
population
ppb, SD = 17.4
ppb
field, better
captures hot
spots
= 0.83, RMSE =
13.8 ppb, MAE =
11.0 ppb, MB =
6.1 ppb
2-127
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Garner et al.
National Air
L: Baltimore,
Comparison with
Short-term
NR
NAQFC-beta
Both versions of
NAQFC: Corr =
(2015)
Quality
MD;
data from fixed-site
exposure
provides more
the model
0.69-0.82,
Forecast
T: July 2011;
monitors reporting
mechanistic
overpredict at low
RMSE =
Capability
P: Entire
to the AQS
information,
concentrations
13.59-18.75 ppb,
combines
population
despite having
and underpredict
MB = -1.15 to
CMAQ 4.5
higher error
at high
8.96 ppb, NMB =
with
concentrations
-2.28 to 22.34%;
WRF-NMM,
NAQFC-beta:
also tested
Corr = 0.67-0.81,
beta version
RMSE = 15.81-
with full gas
20.92 ppb, MB =
and aerosol
3.40-13.84 ppb,
mechanism
NMB = 6.75-
34.49%
Wana et al. (2016)
UCD-CIT
L: Los
10-fold
Short- and
8-h daily max,
CTM accounts
Positive bias in
RMSE =
chemical
Angeles and
cross-validation
long-term
annual
for atmospheric
the concentration,
8.83 ppb, R2 =
transport
Riverside
against 37 monitors
exposure
averages
chemistry,
more variability
0.56
model with
counties,
for each variation
(annual average
long-range
compared with
meteorology
CA;
of model
used for
transport of
spatiotemporal
modeled by
T: 2000-
summary stats)
ozone and its
models
WRF 3.1.1 on
2008;
precursors, and
a 4-km grid
P: Entire
population
biogenic VOCs
2-128
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Seltzer et al.
CMAQ
L: CONUS;
Comparison of
Short- and
Median (range
Variability is
Model is positively
Mean (SD) of
(2016)
5.0.2/WRF
T: January
model predictions
long-term
estimated from
accurately
biased in the
median bias
3.4.1 with
1, 2000-
with data from
exposure
graph) March-
captured
summer and fall
across years
36-km domain
December
monitors reporting
May = 45-50
and negatively
March-May = 0.9
with met fields
31, 2010;
to the AQS
ppb, June-
biased in the
ppb (0.7 ppb),
downscaled
P: Entire
population
August = 55-65
ppb,
September-
November =
45-55 ppb,
December-
February = 35-
40 ppb
winter
June-August =
9.7 ppb (1.16
PPb),
September-
November = 10.9
ppb (0.96 ppb),
December-
February = 6.7
ppb (1.01 ppb)
2-129
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period, Measurement Exposure
and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Yahva et al. WRF/Chem L: CONUS; Model compared Long-term
(2016) 3.6.1 with T: 2001- with surface, exposure
36-km 2010; fixed-site monitors
horizontal P: Entire from U.S. EPA's
resolution population AQS and CASTNet
AQS Hourly
ozone mean
obs = 29.3 ppb,
mean sim =
32.1 ppb, 8-h
daily max mean
obs = 43.7 ppb,
mean sim =
45.9 ppb;
CASTNet hourly
mean obs =
35.0 ppb, mean
sim = 31.9 ppb,
1-h daily max
mean obs =
47.4 ppb, mean
sim = 38.5 ppb,
8-h daily max,
mean obs =
43.3 ppb, mean
sim = 37.9 ppb
Long-term
modeling well
suited for
long-term
exposures
Temperature
typically
overpredicted
during the winter,
overpredictions of
biogenic
emissions in rural
areas
Vs AQS hourly
ozone R = 0.6,
MB = 2.8 ppb,
NMB = 9.7%,
NME = 22.4%;
vs. AQS 1-h max
ozone mean obs
= 48.9 ppb, mean
sim = 49.7 ppb,
R = 0.6, MB = 0.8
ppb, NMB =
1.7%, NME =
7.9%; vs. AQS
8-h daily max R =
0.6, MB = 2.2
ppb, NMB =
5.0%, NME =
9.3%; vs.
CASTNet hourly
ozone R = 0.7,
MB = -3.1 ppb,
NMB = -8.8%,
NME = 19.8%;
CASTNet 1-h
max ozone R =
0.4, MB = -8.9
ppb, NMB =
-18.8 ppb, NME
= 31.4%; vs.
CASTNet 8-h
max ozone R =
0.5, MB = -5.4
ppb, NMB =
-12.5%, NME =
29.6%
2-130
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Baker et al. (2016)
CMAQ 5.0.2,
2011 NEI 2,
SMOKE 3.6.5,
WRF 3.4.1
with 12 km
horizontal
resolution
L: Wallow
fire (eastern
Arizona and
western New
Mexico,
U.S.) and
Flint Hills fire
(eastern
Kansas,
U.S.);
T: June 1-6,
2011
(Wallow),
April 1-5,
2011 (Flint
Hills);
P: Entire
population
Model compared
with surface,
fixed-site monitors
from U.S. EPA's
CASTNet
Short-term
exposure
Hourly ozone
NR (shown
graphically)
Localized
spatiotemporal
region allows for
measuring
ozone from a
specific event;
comparison to
observed data
appropriate
Short time period
may not be
indicative of
typical ozone
exposures
Bias presented
as a function of
ozone
concentration for
wildfire and
prescribed burn.
Wildfire increase
in bias of
approximately
2 ppb for every
1 ppb increase in
estimated ozone
contribution from
fire; prescribed
burn increase in
bias of
approximately
1 ppb for every
1 ppb increase in
estimated ozone
contribution from
fire
2-131
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique Applications
Measured
Strengths
Limitations
Errors
Bash etal. (2016)
CMAQ 5.0.2,
L: Central
Models compared Short-term
Avg hourly obs
This incremental
Localized
Biases and errors
WRF 3.3,
and northern
to each other and exposure
ozone greater
improvement in
spatiotemporal
when using
biogenic
California;
compared to
than 60 ppb =
modeled inputs
modeling domain
satellite
emission from
T: June 3-
observed, fixed-site
70.9 ppb, less
allows for seeing
may not be typical
parameterization
BEIS 3.61 with
July 31,
monitors from U.S.
than 60 =
pointed
of average ozone
of weather
4-km
2009;
EPA's AQS
32.0 ppb, avg
concentration
exposures
model: ozone
horizontal
P: Entire
mod hourly
changes; fine
greater than
resolution;
population
ozone greater
spatial
60 ppb: median
sensitivity
than 60 ppb =
resolution
bias = -8 to -9
analysis
between 62.1
ppb, median
includes BEIS
and 64.8, less
error = 13-14
3.14, BEIS
than 60 ppb =
ppb, MB = -6.2
3.61 WRF par,
between 40.7
to -5.5 ppb, ME
MEGAN 2.1
and 41.7 ppb
= 11-12 ppb, FB
WRF par
= -10.1 to
-9.5%, FE =
16.7-17.8%; less
than 60 ppb:
median bias =
29-32 ppb,
median error =
32-34 ppb, MB =
8.8-9.7 ppb, ME
= 11.1-11.8 ppb,
FB =
29.8-31.9%, FE
= 36.4-37.9%
Benchmark 12-km resolution NR
study of state-of-
the-art CTM
science and
evaluation
Appel etal. (2017) WRF 3.7 and L: CONUS;
CMAQ5.1 T: 2011
annual
simulation;
P: Entire
population
Annual, monthly, Long-term NR
seasonal and exposure
diurnal evaluations
provided against
AQS data
2-132
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Pleimet al. (2016) WRF 3.7 and
CMAQ 5.1
L: CONUS;
T: 3-week
simulation
August 10-
30, 2006;
P: Entire
population
Daily evaluation
against 1,144 sites
Short-term
exposure
NR
Improvements
made in
Land-Surface
Model and PBL
model provides
more accurate
ozone
simulations
12-km resolution,
3-week simulation
NR
Pan et al. (2017a)
WRF 3.4 and
CMAQ 5.0.1
L: Houston,
Hourly evaluation
Short-term
NR
4-km resolution
1-day study
With
TX;
with TCEQ
exposure
improvements in
T: 1-day
monitors
both
simulation
meteorological
September
and emissions
25, 2013;
input, model
P: Entire
under-prediction
population
of peak ozone
concentrations
(>100 ppb),
improves with
mean biases
decreasing from
50 to 9 ppb
L: Greater
Hourly and
Short- and
NR
4-km resolution
1-mo analysis
Inclusion of
Los Angeles
month-long
long-term
marine halogen
area, CA;
aggregated
exposure
and sulfur
T:
evaluation against
concentrations
September
eight AQS sites
reduced model
2006;
overprediction as
P: Entire
mean bias is
population
reduced from
13.5 to 4.9%
across the
domain and
month
Muniz-
Unamunzaqa et
al. (2018)
WRF 3.71,
CMAQ 5.1
2-133
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Reference
Model
Location,
Time Period,
and
Population
Measurement
Evaluation
Technique
Epidemiology
Applications
Concentrations
Measured Strengths
Limitations
Exposure
Measurement
Errors
U.S. EPA (2011)
MM5 3.7.4,
L: CONUS;
Hourly and 8-h
Short-term
NR CONUS annual
12-km resolution
For 8-h daily
CMAQ 4.7.1
T: 2005
daily max
exposure
simulation
max: bias ranged
annual
from -2 to -9
simulation;
ppb, error ranged
P: Entire
from 9 to 9 ppb,
population
fractional bias
ranged from -3
to -14% and
fractional error 12
to 15%. For
maximum daily
hourly, bias
ranged from -4
to -9 ppb. Error
10 to 11 ppb and
FB-6 to-13%
and FE 14 to
15%
Correlations
ranged from 0.63
-0.67 for hourly
to 0.67-0.72 for
8-h daily max;
NMB ranged
from -7.5 to
-13% (hourly)
1.6-4.3% (8-h
daily max); NME
ranged from 11-
23% (hourly) and
16-21% (8-h
daily max)
depending on
year
Henneman et al.
(2017b)
WRF 3.6.1
and CMAQ
5.0.2
L: Eastern
U.S.;
T: Two, 2-yr
periods,
2001-2002,
2011-2012;
P: Entire
population
Hourly and 8-h
daily max
evaluation against
more than 500
AQS sites for each
of the 4 yr
Short-term
exposure
NR
4 full yr of
simulation/
evaluation
12-km resolution
and only 13
vertical layers
2-134
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured Strengths
Limitations
Errors
Henneman et al.
WRF 3.6.1,
L: Atlanta
8-h daily max
Short-term
NR NR
12-km resolution,
For all 8-h daily
(2017a)
CMAQ 5.0.2
GA;
exposure
only one location
max: NMB =
T: 2001;
in downtown
-0.8%, NME =
P: Entire
Atlanta was used
27.3%, MB =
population
in the evaluation
-0.4%, r= 0.70.
for 8-h daily max
>60 ppb: NMB =
-16.8%, NME =
19.7%, MB =
-12.2%, r= 0.43.
Pan et al. (2017b)
WRF 3.7 with
L: Houston
Hourly
Short-term
NR 1 and 4 km
Only two
Evaluation
CMAQ 5.0.2
TX;
concentrations
exposure
simulations
observations sites
statistics were
T:
from the TCEQ
provided in
September
supplementary
2013;
material. R
P: Entire
ranged from
population
0.75-0.77; MB
from 10-13 ppb
NoDmonacol et al.
GEOS-Chem
L: CONUS;
8-h daily max from
Short-term
NR NR
36-km resolution
AQS sites:
(2017)
9.1.3, CAMx
T: 2005;
AQS and CASTNet
exposure
Annual NMB =
6.1
P: Entire
monitors
7.7%, NME =
population
15%, r= 0.76,
RMSE = 11.20;
CASTNet sites:
Annual NMB =
0.4%, NME =
14%, r= 0.52,
RMSE = 19.40
Matichuk et al.
WRF 3.4,
L: Utah
Hourly ozone at a
Short-term
NR 4-km resolution
10-day period in
Model bias
(2017)
CMAQ 5.0.2
(Uinta
dozen "field study"
exposure
February
ranged from 15
Basin);
locations
to 60 ppb
T: 10 days in
2013;
P: Entire
population
2-135
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured Strengths
Limitations
Errors
Seltzer et al.
GEOS-Chem
L: CONUS;
R6MA1 (running
Short-term
NR Annual
Grid resolutions
CONUS values
(2017)
T: Two, 2-yr
6-mo avg of the 1-h
exposure
simulations
were 2.0* 2.5°
of the NMB of 8-h
annual
daily max) and 8-h
and 0.5 x 0.666°
daily max ranged
simulations
daily max of over
from 2.0 to 6.6
(2004-2006,
1,000 AQS sites
depending on
2009-2011);
simulation year
P: Entire
population
Solazzo et al.
WRF, CMAQ
L: CONUS;
Hourly ozone
Short-term
NR CONUS
12-km resolution
Annual MSE
(2017)
T: 2010
concentrations
exposure
ranged from 28.6
annual
to 79.3 ppb2
simulation;
P: Entire
population
Hall et al. (2012)
WRF 3.1,
L: CONUS;
Hourly and 8-h
Short-term
NR Annual
12-km resolution
Evaluation was
CMAQ 4.7.1
T: 2008
daily max ozone at
exposure
simulation
segregated into
annual
1,176 AQS sites
seasons and
simulation;
eight CONUS
P: Entire
subregions. NMB
population
ranged from
-10.4 to 19.5%,
FB ranged from
-10.1 to 20.0%;
NME ranged
from 11.2-25.3%
and FE ranged
from 11.9-25.3%
Zhanq and Yinq
MM5, CMAQ
L: Houston
60 AQS sites,
Short-term
NR 4-km resolution
Only a 12-day
MNB ranged
(2011)
(version NR)
TX MSA;
hourly data
exposure
simulation
from -0.3 to
T: 12 days
+0.2%
only in
August
2000;
P: Entire
population
2-136
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured Strengths
Limitations
Errors
Castellanos et al.
MM5 3.6,
L: Eastern
612 AQS sites and
Short- and
NR NR
12-km resolution,
R2 for urban sites
(2011)
CMAQ 4.5.1
U.S.;
85 CASTNet sites,
long-term
short-term study
0.55, for rural
T: May 15-
hourly ozone
exposure
sites 0.49, hourly
September
biases were 6-12
15, 2000;
ppb during rural
P: Entire
nighttime and -1
population
to 3 ppb during
urban afternoons
Napelenok et al.
MM5 3.6.3,
L: Eastern
684 AQS sites,
Short-term
NR NR
12-km simulation,
NMB ranged
(2011)
CMAQ 4.7.1
U.S.;
DM8H ozone
exposure
paper focused on
from 0.8% in
T: June 1-
a dynamical
2002 to 2.6%in
August 31,
evaluation using
2005; NME
2002 and
two emission
ranged from
2005;
scenarios and
16.6% in 2002 to
P: Entire
less on actual
17.6% in 2005
population
evaluation with
observations
Tana et al. (2011)
MM5 3.6.1
L: Houston,
58 AQS sites in
Short-term
NR 4-km resolution
7-day simulation/
MNE = 15.4%,
and CMAQ 4.5
TX MSA;
southeastern
exposure
validation
MNB = -4.9%,
T: 7-day
Texas
R2 = 0.49
period
September
2006;
P: Entire
population
2-137
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Yu etal. (2018)
CMAQ 5.0.2
L: Atlanta,
Comparison with
Short-term
NR
Better spatial
Relatively low
Urban site: MB =
36- x 36-km
GA;
fixed-site monitors
exposure
resolution than
spatial resolution
-4.5 ppb, ME = 8
resolution
T: 2011;
P: Entire
population
monitor-based
approaches
(36 km)
ppb, RMSE = 12
ppb, MNB =
-9%, MNE =
21%, NMB =
-10%, NME =
18%, MFB =
-13%, MFE =
23%, R2 = 0.65,
Slope = 0.81;
Rural site: MB =
-0.73 ppb, ME =
5.92 ppb, RMSE
= 7.64 ppb, MNB
= 4%, MNE =
15%, NMB = 2%,
NME = 13%,
MFB = 2%, MFE
= 14%, R2 = 0.69,
Slope = 0.85
2-138
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Reference
Location,
Time Period, Measurement
and Evaluation Epidemiology Concentrations
Model Population Technique Applications Measured Strengths
Limitations
Exposure
Measurement
Errors
ACM2 = Asymmetric Convective Model Version 2; AIRBASE = European air quality database; AIRPACT-3 = Air Indicator Report for Public Awareness and Community Tracking
Version 3 ; AIRS = Aerometric Information Retrieval System; AOD = aerosol optical density; APT = Advanced Plume Treatment; AQMEII = Air Quality Model Evaluation International
Initiative; AQS = Air Quality System; ARW = Advanced Research Weather; AUP = unpaired predicted-to-observed peak ozone ratio; AURAMS = Unified Regional Air Quality Modeling
System; BEIS = Biogenic Emissions Inventory System; BME = Bayesian maximum entropy; BRAVO = Mexican Emissions Inventory System; CA = California; CALGRID = California
Grid Simulations; CAMS = Continuous Monitoring Station; CAMx = Comprehensive Air Quality Model with Extensions; CARB = California Air Resources Board; CASTNet = Clean Air
Status and Trends Network; CB05 = carbon bond mechanism of CMAQ; CEM = continuous emissions modeling; CMAQ = Community Multiscale Air Quality model;
CONUS = continental U.S.; CSAPR = Cross-State Air Pollution Rule; CSI = critical success index; CTM = chemical transport model; DDM-3D = decoupled direct method in three
dimensions; DEHM = Danish Eulerian Hemispheric Model; DOE = Department of Energy; EDGAR = Emissions Database for Global Atmospheric Research; EMEP = European
Monitoring and Evaluation Program; FAR = false alarm ratio; FB = fractional bias; FDDA = four-dimensional data assimilation; FE = fractional error; FEPS = Fire Emissions Production
Simulator; Fp = percentage of cases where simulation results were close to observations; GB = gross bias; GE = gross error; HDDM = Hierarchical Bayesian Diffusion Drift Model;
HTAP = Hemispheric Transport of Air Pollutants; ICARTT = International Consortium for Atmospheric Research on Transport and Transformation; ICC = interclass correlation
coefficient; IDW = inverse-distance weighting; IMPROVE = Interagency Monitoring of Protected Visual Environments; IOA = index of agreement; IQR = interquartile range; L = location;
LUR = land use regression; MADRID = Model of Aerosol Dynamics, Reaction, Ionization, and Dissolution; MAE = mean absolute error; MB = mean bias; MCM = master chemical
mechanism; ME = mean error; MEGAN2 = Model for Gases and Aerosols from Nature Version 2; MESA = Multi-Ethnic Study of Atherosclerosis; MFB = mean fractional bias;
MFE = mean fractional error; MM5 = Mesoscale Model Version 5; MNB = mean normalized bias; MNE = mean normalized error; MNGE = mean normalized gross error;
MOBILE6 = mobile emission model; MODIS = Moderate Resolution Imaging Spectroradiometer; MOZART = Model for Ozone and Related Chemical Tracers; MYJ = Mellor-Yamada-
Janjic; NADP = National Atmospheric Deposition Program; NAM = North American mesoscale; NAPS = National Air Pollution Surveillance; NAQFC = National Air Quality Forecast
Capability; NB = normalized bias; NC DENR = North Carolina Department of Environment and Natural Resources; NEI = National Emissions Inventory; NEu = Northern Europe;
NGE = normalized gross error; NMB = normalized mean bias; NME = normalized mean error; NMM = nonhydrostatic mesoscale model; NOAA = National Oceanographic and
Atmospheric Administration; NR = not reported; NRC = National Research Council; NU = NASA-Unified; NW = northwest; NYC = New York City; OK = ordinary kriging; OMI = Ozone
Monitoring Instrument; P = population; PBL = planetary boundary layer; POD = probability of detection; QNSE = quasi-normal scale elimination; R = Pearson correlation;
RMSE = root-mean-squared error; SD = standard deviation; SE = southeast; SEARCH = Southeastern Aerosol Research and Characterization; SEu = Southern Europe; SIP = State
Implementation Plan; SJV = San Joaquin Valley; SMOKE = Sparse Matrix Operator Kernel Emissions model; SW = southwest; T = time; TCEQ = Texas Commission on Air Quality;
TES = Tropospheric Emissions System; TOPP = Tropospheric Ozone Pollution Project; U = east-west component, UAM = Urban Airshed Model; UCD-CIT = UC Davis—California
Institute of Technology model; UK = universal kriging; UPA = unpaired normalized bias; V = north-south component, VOC = volatile organic compound; VW = Volkswagen;
WDCGG = World Data Centre for Greenhouse Gases; WHI-OS = Women's Health Initiative Observational Study; WOUDC = World Ozone and Ultraviolet Data Centre; WRF = Weather
Research Forecasting model; YSU = Yonsei University.
2-139
-------
Table 2-13 Studies informing assessment of exposure measurement error when concentrations modeled by
hybrid approaches are used for exposure surrogates.
Reference
Model
Location,
Time Period,
and
Population
Measurement
Evaluation
Technique
Epidemiology
Applications
Concentrations
Measured
Strengths
Limitations
Exposure
Measurement
Errors
Di etal. (2017)
Neural network
L: CONUS;
10-fold cross
Long-term
Annual 4th
Good spatial
Potential for model
Mean R2 = 0.76,
incorporating OMI
T: 8-h daily
validation
exposure
highest ozone
coverage, high
overfitting
RMSE = 7.36
column data
max ozone
against monitor
concentration by
spatial
ppb; spatial R2 =
calibrated to
for 2000-
data reporting
region:
resolution; high
0.80, RMSE =
ground level
2012;
to AQS for
Northeast =
R2 and low
2.91 ppb;
ozone
P: Medicare
annual
0.05-0.085
RMSE
temporal R2 =
concentration
population
fourth-highest
ppm, Southeast
0.75, RMSE =
predicted by
8-h daily max
= 0.055-0.075
6.79 ppb; bias =
GEOS-Chem, land
ppm, West =
1.20 ppb; slope
use variables, and
0.055-0.07
= 0.99
monitoring
ppm, National =
network
0.055-0.06 ppb
2-140
-------
Table 2-13 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by hybrid approaches are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Hao et al. (2012)
IDW ofCMAQ at
the block group
level (version of
CMAQ unclear) for
two different grid
resolutions: 36
and 12 km
L: CONUS;
T: 8-h daily
max ozone
for 2006;
P: Entire
population
CMAQ grid
resolutions of
36 and 12 km
compared
Short-term
exposure
98th percentile
of ozone for
2006 = between
39.8 and 100.7
ppb (unclear
which data
source
produced these
concentrations),
90th percentile
of ozone for
2006 = between
36.7 and 84.4
ppb (again,
unclear which
data source
produced these
concentrations)
CMAQ has
been well
validated
Methods are unclear
for many of the
figures in the paper
Number of
monitoring sites
between 790
and 897, n
between
195,035 and
232,081, mean
absolute
deviation
(12 km) between
3.51 and 4.53
ppb, mean
absolute
deviation
(36 km) between
2.98 and 3.23
ppb, R (12 km)
between 0.94
and 0.96, R (36
km) between
0.96 and 0.97
2-141
-------
Table 2-13 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by hybrid approaches are used for exposure surrogates.
Location,
Time Period, Measurement Exposure
and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Hvstad et al.
(2012)
Historically
calibrated hybrid
model, with a
separate model for
each historical
year, using a
regression by at a
21-km horizontal
resolution by
combining
Canadian
Hemispheric and
Regional Ozone
NOx System
(CHRONOS) with
ozone obs from
Canada and the
U.S. from 2004-
2006 with
historical
calibration through
NAPS monitors,
interpolation to
specific locations
were done with:
(1) IDW and
(2) regression with
ozone calibration
and population
density at 10-km
buffers around
NAPS stations
L: Canada;
T: Annual
ozone from
1975-1994;
P: Entire
population
Historical
model
compared with
surface,
fixed-site
monitors from
NAPS
Long-term
exposure
IDW exposure
estimates from
NAPS monitors
N =6,919,
mean = 23.2
ppb, SD = 3.7
ppb, min = 12.9
ppb, IQR = 4.6
ppb, max = 35.4
ppb, linear
model N =
6,919, mean =
26.4 ppb, SD =
3.4 ppb, min =
18.1 ppb, IQR =
4.7 ppb, max =
37.2
Very few
studies
examine
long-term
exposures to
ozone
Some aspects of
methodology were
not clear
Cross-validation
with 10% of
monitoring data
CHRONOS-IDW
R2 = 0.39, RMSE
= 5.29 ppb;
CHRONOS-
linear R2 = 0.56,
RMSE = 4.48
ppb
2-142
-------
Table 2-13 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by hybrid approaches are used for exposure surrogates.
Reference
Model
Location,
Time Period,
and
Population
Measurement
Evaluation
Technique
Epidemiology
Applications
Concentrations
Measured
Strengths
Limitations
Exposure
Measurement
Errors
Davis et al. (2011)
GLM method
L: 8-h max
CMAQ output
Long-term
8-h max ozone
GLM method
Because the GLM
R2 between
developed
ozone data
is compared to
exposure
= between 0
based on
models were
ozone and fitted
between CMAQ
from 74
observed data;
and 150 ppb
CMAQ was
developed for each
ozone, obs R2 =
output and
cities across
modeled met
directly
location, the GLM
0.74 and CMAQ
modeled met and
the eastern
data are
compared with
model may not have
R2 = 0.70;
observed data and
U.S.;
compared to
observed data;
predictive power
monitoring
observed met
T: May
observed met
all models
spatially
station-specific
through
data; both
developed
GLM model by
September
GLM models
were highly
R2 between 50.0
from 2002 to
are compared
localized
and 80.0
2005;
to each other
P: Entire
population
2-143
-------
Table 2-13 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by hybrid approaches are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique Applications
Measured
Strengths
Limitations
Errors
Berrocal et al.
Downscaled
L: East coast
Comparison Short-term
NR (shown
Data
GMRFs cause some
CMAQ: PMSE =
(2012)
CMAQ (version
U.S.;
with monitors exposure
graphically)
assimilation
over smoothing to
135.9, PMAE =
not stated) at
T: July 4,
reporting to
improves
blur predictive maps
9.1, 95% PI NR
12-km resolution
July 20,
AQS
model fit for
CP NR;
downscaled with
August 9,
different
regressor PMSE
either a Gaussian
2001;
variations of
= 124.2, PMAE =
Markov random
P: Entire
downscaling,
8.7, 95% PI NR
field or a
population
and fine-tune
CP NR; ordinary
univariate model
model
enhancements
improve fit
further
kriging: PMSE =
60.9, PMAE =
5.8, 95% PI =
30.6 CP =
94.8%;
downscaler:
PMSE = 53.1,
PMAE = 5.3,
95% PI = 30.4
CP = 94.9%;
GMRF
downscaler:
PMSE = 50.3,
PMAE = 5.2,
95% PI = 29.4
CP = 94.9%;
smoothed
downscaler:
PMSE = 45.4,
PMAE = 5.0,
95% PI = 27.7
CP = 95.0%
2-144
-------
Table 2-13 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by hybrid approaches are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Liu etal. (2011)
Multiscale Air
L: Eastern
Comparison of
Long-term
NR
Method
Computationally
Daily RMSPE =
Quality Simulation
and
model points
exposure
connects
intensive; this
6.98-18.55 ppb
Platform
midwestern
with
measurements
version did not
(MAQSIP) model,
U.S.;
concentrations
with model
include a space-time
6-km resolution
T: May 15-
from
results through
framework; model
with Bayesian
September
375 monitors
latent
assumed
downscaling to
11, 1995
reporting to
processes; the
measurements and
monitoring data
10:00 a.m-
5:00 p.m.;
P: Entire
population
AQS
model has
flexibility
model outputs are
Gaussian processes
2-145
-------
Table 2-13 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by hybrid approaches are used for exposure surrogates.
Location,
Time Period, Measurement Exposure
and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Pongprueksa
(2013)
CMAQ 4.7.1, WRF
L:
CONUS;
CMAQ using
3.4, CONUS with
T:
2009;
satellite data
36-km horizontal
P:
Entire
were
resolution for 2009
population
compared to
merged with
the typical
Tropospheric
conditions for
Emission
CMAQ, both
Spectrometer
CMAQ
(TES) L3 (using
methods were
TES data as is
compared with
and shifted by
observed data
10 ppb to better
from U.S.
account for
EPA's AQS
boundary
and
conditions)
ozonesonde
data collected
across CONUS
from WOUDC,
NOAA, and
TOPP
Short- and Ozonesonde; Addition of Satellite data
long-term annual 8-h daily satellite data overestimates
exposure max ozone in reduces error tropospheric ozone
southern states and uncertainty
for 2009; 8-h both in the
daily max from upper
Texas; annual atmosphere
8-h daily max and in the
ozone across troposphere
CONUS
Surface
observed ozone
compared
CMAQ-TES: n =
26,234, MB = 9
ppb, ME = 12
ppb, NMB =
23%, NME =
31%, R = 0.56,
compared with
CMAQ-TESadj:
n = 26,234, MB
= 4 ppb, ME =
10 ppb, NMB =
10%, NME =
24%, R = 0.58;
model
performance by
region n = 1-21
sites,
CMAQ-TES, MB
= 3-15 ppb, ME
= 7-15 ppb,
NMB = 6—45%,
NME = 14-46%,
R = 0.46-0.64,
and
CMAQ-TESadj,
MB = -3 to 9
ppb, ME = 6-11
ppb, NMB = -6
to 28%, NME =
14-30%, R =
0.48-0.66
2-146
-------
Table 2-13 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by hybrid approaches are used for exposure surrogates.
Location,
Time Period, Measurement Exposure
and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Reich etal. (2014)
Spectral
L: CONUS;
Spectral
Short-term
1-day avg
Method is
downscaling using
T: July 2005;
downscaling
exposure
ozone
explicitly
surface, fixed-site
P: Entire
compared with
concentration
stated; hybrid
monitors from U.S.
population
no CMAQ,
models allows
EPA's AQS and
linear
for strength of
CASTNet and
downscaler,
both CTMs and
CMAQ 5.0.1 with
and kernel
obs data
a 12-km horizontal
smoothed
resolution
downscaler, all
comparison
methods are
compared
against
observed data
Short-term exposure
is not indicative of
longer ozone
exposures;
collocation needed
for validation
preferentially selects
higher ozone areas
Spatial
prediction:
monitors only
MSE = 62.8
ppb2, bias =
-0.14 ppb,
variance = 66.3
ppb2 CP = 0.91,
linear
downscaler MSE
= 57.5 ppb2, bias
= -0.26 ppb,
variance = 56.2
ppb2 CP = 0.91,
spectral
downscaler MSE
= 53.7 ppb2, bias
= -0.23 ppb,
variance = 53.3
ppb2 CP = 0.91,
kernel
downscaler
12-km resolution
MSE = 54.9
ppb2, bias =
-0.23 ppb,
variance = 54.8
ppb2 CP = 0.91,
kernel
downscaler
60-km resolution
MSE = 58.7
ppb2, bias =
-0.17 ppb,
variance = 59.2
ppb2 CP
2-147
-------
Table 2-13 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by hybrid approaches are used for exposure surrogates.
Location,
Time Period, Measurement Exposure
and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Reich etal. (2014)
(continued)
Spectral
downscaling using
surface, fixed-site
monitors from U.S.
EPA's AQS and
CASTNet and
CMAQ 5.0.1 with
a 12-km horizontal
resolution
(continued)
L: CONOUS;
T: July 2005;
P: Entire
population
(continued)
Spectral
downscaling
compared with
no CMAQ,
linear
downscaler,
and kernel
smoothed
downscaler,
comparison
methods are
compared
against
observed data
(continued)
Short-term
exposure
(continued)
1-day avg
ozone
concentration
(continued)
all
Method is
explicitly
stated; hybrid
models allows
for strength of
both CTMs and
obs data
(continued)
Short-term exposure
is not indicative of
longer ozone
exposures;
collocation needed
for validation
preferentially selects
higher ozone areas
(continued)
= 0.91, kernel
downscaler
120-km
resolution MSE =
60.9 ppb2, bias =
-0.14 ppb,
variance = 62.9
ppb2 CP = 0.91;
nonspatial
prediction:
monitors only
MSE = 339.7
ppb2, bias =
-6.17 ppb,
variance = 302.0
ppb2 CP = 0.89,
linear
downscaler MSE
= 202.1 ppb2,
bias = -2.80
ppb, variance =
177.7 ppb2 CP =
0.89, spectral
downscaler MSE
= 145.7 ppb2
bias = 0.57 ppb,
variance = 129.1
ppb2 CP = 0.89,
kernel
downscaler
12-km resolution
MSE = 151.2
ppb2,
2-148
-------
Table 2-13 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by hybrid approaches are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Reich et al. (2014) Spectral
(continued)
downscaling using
surface, fixed-site
monitors from U.S.
EPA's AQS and
CASTNet and
CMAQ 5.0.1 with
a 12-km horizontal
resolution
(continued)
T: July 2005:
P: Entire
population
(continued)
; Spectral
Short-term
1-day avg
Method is
Short-term exposure
bias = -0.06
downscaling
exposure
ozone
explicitly
is not indicative of
ppb, variance =
compared with
(continued)
concentration
stated; hybrid
longer ozone
134.8 ppb2 CP =
no CMAQ,
(continued)
models allows
exposures;
0.89, kernel
linear
for strength of
collocation needed
downscaler
downscaler,
both CTMs and
for validation
60-km resolution
and kernel
obs data
preferentially selects
MSE = 157.6
smoothed
(continued)
higher ozone areas
ppb2, bias = 1.06
downscaler, all
(continued)
ppb, variance =
comparison
142.9 ppb2 CP =
methods are
0.89, kernel
compared
downscaler
against
120 km
observed data
resolution MSE =
(continued)
169.1 ppb2, bias
= 0.93 ppb,
variance = 151.7
ppb2 CP = 0.89
Full model MSE
24.97, ppb MAE
3.80 ppb, CP
85.7%, avg
length of PI
15.70 ppb;
reduced model
MSE 24.66 ppb,
MAE 3.79 ppm,
CP 85.5%, avg
length of PI
13.67 ppb
2-149
-------
Table 2-13 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by hybrid approaches are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Pad etal. (2013)
Eta-CMAQ at
L: Eastern
Compare
Short-term
NR
Works well for
Method is linear, so it
Full model MSE
12-km scale
and
predicted
exposure
forecasting
may not handle more
24.97 ppb2, MAE
(CMAQ version
midwestern
ozone
because the
complex temporal
3.80 ppb, CP
not specified),
U.S.;
concentration
downscaler is
models well
85.7%, avg
downscaled to 717
T: August 1-
to ozone
based on
length of PI
ozone monitors
14, 2011;
P: Entire
population
measured at
monitors set
aside from the
analysis for
validation
temporal
gradients
15.70 ppb;
reduced model
MSE 24.66 ppb2,
MAE 3.79 ppb,
coverage
probability
85.5%, avg
length of PI
13.67 ppb
Tana et al.
Optimal
L:
CONUS;
Modeled
Short-term
Hourly ozone in
Two different
1 summer mo may
Hourly ozone
(2015b)
interpolation (Ol)
T: July 2011;
outputs
exposure
the first half of
observed data
not be indicative of
from July 6-7,
hybrid method
P:
Entire
compared with
July 2011 in
sources used;
more long-term
2011 over
with AirNow data,
population
AirNow
northeastern
multiple
exposures
CONUS for
MODIS AOD from
observed data
U.S. NR (shown
models
optimal
Terra and Aqua
and aircraft
graphically)
compared to
interpolation (Ol)
satellites
measurements
each other
1-4, R =
incorporated into
from
0.52-0.58, MB =
CMAQ 5.0.2,
Discover-AQ
1.06-2.36;
WRF-ARW 3.4.1
hourly ozone
with relative
from July 6-7,
uncertainties of
2011 over
0.4, up to 0.6, up
southeastern
to 1.0,
U.S. for Ol 1-4,
respectively, with
R = 0.58-0.61,
12-km horizontal
MB = -1.40 to
resolution
0.43; R between
obs and Ol 4 for
aircraft data is
0.753
2-150
-------
Table 2-13 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by hybrid approaches are used for exposure surrogates.
Reference
Model
Location,
Time Period,
and
Population
Measurement
Evaluation
Technique
Epidemiology
Applications
Concentrations
Measured
Strengths
Limitations
Exposure
Measurement
Errors
Huana et al.
CAMx 6.10 with
L: Eastern
Models
Short-term
NR
Fine scale
Limited spatial and
Difference
(2015)
4-km horizontal
Texas;
compared with
exposure
resolution; this
temporal coverage
between obs and
resolution with
T: June
each other and
speaks to the
may not be indicative
simulation:
land cover
2006;
with ambient
importance of
of a typical exposure
CAMx with
compared
P: Entire
monitoring
the effect of
concentration
MODIS mean =
generated from
population
sites (unclear
model inputs to
2-6 ppb, max =
MODIS and
who is
measured
>20 ppb; CAMx
TCEQ; models
responsible for
concentration
with TCEQ data
compared with
these sites)
mean = 2 ppb,
each other and
max = 30 ppb
with ambient
monitoring sites
(unclear who is
responsible for
these sites)
2-151
-------
Table 2-13 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by hybrid approaches are used for exposure surrogates.
Location,
Time Period, Measurement Exposure
and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Fribera et al.
(2016)
CMAQ 4.5 fused
with data by three
methods: fusing
CMAQ with
interpolated data,
scaling CMAQ
fields to data that
are corrected for
seasonal bias, and
combining these
methods through a
weighted model
L: Georgia;
T: 2002-
2008;
P: Entire
population
Cross-
validation by
fixed-site
monitors
Long-term
exposure
Mean (IQR) =
47.6 ppb
(22.0 ppb) for
8-h daily max
Low mean
Errors in
Interpolated
bias, low
measurements used
observations
RMSE, and
as input are
MFE = 0.16,
relatively high
propagated into the
MFB = 0.02,
R2 (compared
model, limited spatial
NME = 14.7%,
to application
coverage of monitors
NMB = -0.57%,
of the model
increases errors
MB = -2.7 x
for other
(although this is less
10"4, RMSE =
pollutants), and
of a limitation for
0.01, R2 (cross-
errors are
ozone and other
validation):
minimized
secondary
68.7%; for
through the
pollutants)
optimized
model-
method MFE =
weighting
0.05, MFB =
approach
0.01, NME =
4.49%, NMB =
0.03%, MB =
1.5 x 10"5,
RMSE = 0.003,
R2 (cross-
validation) =
97.0%; weighted
combination of
methods MFE =
0.10, MFB =
0.02, NME =
8.57%, NMB =
0.03%, MB =
1.3 x 10"5,
RMSE = 0.006,
R2 (cross-
validation) =
87.1%
2-152
-------
Table 2-13 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by hybrid approaches are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Xu etal. (2016b)
CAMx 5.30 at
nested 36-km
domain over the
CONUS and
12-km domain
over the eastern
U.S., with BME
implemented by a
model with
parameters held
constant across
the CONUS
[CAMP] and with
regional
parameters
[RAMP]
L: CONUS.:
T: 2005;
P: Entire
population
Comparison
with data from
fixed-site
monitors
reporting to the
AQS, tower
sensors at one
location
(Raleigh, NC),
and
DISCOVER-
AQ flight
sensors
Long-term
exposure
NR
BME enables
estimation of
concentration
below scale of
CTM
simulation,
better spatial
and temporal
validation with
RAMP model,
computation-
ally efficient
and
straightforward
approach
Uncertainty in
concentration
estimates increases
with distance from
the monitors
CAMP: RMSE
0 km = 5.675
ppb, 36 km =
6.442 ppb, 72
km = 6.966 ppb,
108 km = 7.250
ppb R2 0 km =
0.884, 36 km =
0.853, 72 km =
0.831, 108 km =
0.819; RAMP:
RMSE 0 km =
5.445 ppb, 36
km = 6.109 ppb,
72 km = 6.531
ppb, 108 km =
6.732 ppb R2 0
km = 0.893, 36
km = 0.866, 72
km = 0.849, 108
km = 0.841
2-153
-------
Table 2-13 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by hybrid approaches are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Bash etal. (2016)
CMAQ 5.0.2, WRF
L: Central
Models
Short-term
Avg hourly obs
This
Localized
Biases and
3.3, biogenic
and northern
compared to
exposure
ozone greater
incremental
spatiotemporal
errors when
emission from
CA, U.S.;
each other and
than 60 ppb =
improvement in
modeling domain
using satellite
BEIS 3.61 with
T: June 3-
compared to
70.9 ppb, less
modeled inputs
may not be typical of
parameterization
4-km horizontal
July 31,
observed,
than 60 =
allow for
average ozone
of weather
resolution;
2009;
fixed-site
32.0 ppb, avg
seeing pointed
exposures
model: ozone
sensitivity analysis
P: Entire
monitors from
mod hourly
concentration
greater than 60
includes BEIS
population
U.S. EPA's
ozone greater
changes; fine
ppb = median
3.14, BEIS 3.61
AQS
than 60 ppb =
spatial
bias -9 to -12
SAT (from
between 62.1
resolution
ppb, median
MODIS) par,
and 64.8, less
error = 13-14
MEGAN 2.1 SAT
than 60 ppb =
ppb, MB = -6.6
(from MODIS) par
between 40.7
and 41.7 ppb
to -8.8 ppb, ME
= 11-11.9 ppb,
FB = -10.8 to
-14.1%, FE =
16.8-18.3%;
less than 60 ppb:
median bias =
29 ppb, median
error = 32 ppb,
MB = 8.7 ppb,
ME = 11 ppb, FB
= 29.4-30%, FE
= 36.2-36.4%
Xu etal. (2017)
CAMx with
L: CONUS;
Hourly, 8-h
Short-term
NR
CONUS
36- and 12-km
r ranges from
observations
T: 2005
daily max and
exposure
simulations
simulations
0.78-0.82;
integrated using
annual
24-h avg
using
RMSE =
BME
simulation;
P: Entire
population
observations to
improve ozone
simulation
5.2-6.3 ppb
2-154
-------
Table 2-13 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by hybrid approaches are used for exposure surrogates.
Reference
Model
Location,
Time Period,
and
Population
Measurement
Evaluation
Technique
Epidemiology
Applications
Concentrations
Measured Strengths
Limitations
Exposure
Measurement
Errors
Robichaud and
An objective
L: Canada
Models
Long-term
NR Very small or
Impact of NOx on
Frequency of
Menard (2014)
analysis (OA)
and U.S.;
compared to
exposure
no biases were
spatial variability of
model
scheme is
T: 2001-
each other and
observed;
ozone would not be
predictions
developed to
2012;
to observed,
automated
captured over coarse
within a factor of
integrate
P: Canadian
fixed-site
process
grid
two of the
predictions from
Census
monitors from
observations:
CHRONOS and
Health and
the Canadian
Canada =
GEM-MACH
Environment
Meteorological
0.654-0.927;
CTMs (with 21-km
Cohort
Centre and the
U.S. =
horizontal
AQS
0.641-0.969
resolution for
CHRONOS and
15-km resolution
for GEM-MACH)
with surface data
Yu etal. (2018) CMAQ-kriging
hybrid model
where CMAQ
5.0.2, 36- x 36-km
resolution was
run, ratios
calculated
between CMAQ
and observations
across the
surface, CMAQ
output was
adjusted by those
ratios, and then
the surface was
kriged to
interpolate
between grid
centroids (Friberq
etal.. 2016)
= 0.88, slope =
0.88
L: Atlanta,
GA;
T: 2011;
P: Entire
population
Comparison
with fixed-site
monitors
Short-term
exposure
NR
Improves
spatial
resolution over
CMAQ alone
and monitor-
based
approaches
Complex model
design is more
difficult to implement
compared with
CMAQ alone or the
CMAQ-kriging hybrid
model
Urban site: MB =
-5.8 ppb, ME =
6 ppb, RMSE = 7
ppb, MNB =
-12%, MNE =
14%, NMB =
-13%, NME =
14%, MFB =
-13%, MFE =
15%, R2 = 0.95,
slope = 1.20;
Rural site: MB =
-1.79 ppb, ME =
3.98 ppb, RMSE
= 5.16 ppb, MNB
= -4%, MNE =
10%, NMB =
-4%, NME =
9%, MFB = -5%,
MFE = 10%, R2
2-155
-------
Table 2-14 Studies informing assessment of exposure measurement error when concentrations modeled by
microenvironmental modeling are used for exposure surrogates.
Reference
Model
Location,
Time Period,
and
Population
Measurement
Evaluation
Technique
Epidemiology
Applications
Concentrations
Measured
Strengths
Limitations
Exposure
Measurement
Errors
Dionisio et al.
(2014)a
Stochastic
Human
Exposure
and Dose
Simulation
(SHEDS)
model
L: Atlanta,
GA;
T: 1999-
2002;
P: Entire
population
Comparison with Long-term 8-h daily max More precise
dispersion model exposure ozone, NR model
Computationally
intensive, but
might not be
needed
Mean (SD)
exposure
measurement
error:
population =
-0.66 (0.029)
spatial =
-0.055 (0.037),
total = -0.72
(0.010)
2-156
-------
Table 2-14 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by microenvironmental modeling are used for exposure surrogates.
Location,
Time Period,
Measurement
Exposure
and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Reich etal. (2012)
Air Pollutants
L:
Fivefold cross-
Short-term
Daily average
Predicts
The model
Comparison is
Exposure
Philadelphia;
validation against
exposure
ozone NR
exposure
includes many
shown
(APEX)
T: June-
monitors
rather than
assumptions;
graphically
model
August 2001;
P: Entire
population
reporting to AQS
providing a
surrogate for
exposure
accuracy is
limited to quality
of input data
linear
relationship
between
predictions and
observations,
but there are
many instances
where the
model is
positively
biased
AOD = aerosol optical density; APEX = the Air Pollutants Exposure; AQS = Air Quality System; BEIS = Biogenic Emissions Inventory System; CAMP = constant air quality model
performance; CAMx = Comprehensive Air Quality Model with Extensions; CASTNet = Clean Air Status and Trends Network; CHRONOS = Canadian Hemispheric and Regional
Ozone and NOx System; CMAQ = Community Multiscale Air Quality; CONUS = continental U.S.; CP = confidence prediction; CTM = chemical transport model; GEM-MACH = Global
Environmental Multiscale coupled with Model of Air quality and Chemistry; GEOS-Chem = Goddard Earth Observing System Chemistry; GLM = generalized linear model;
GMRF = Gaussian Markov Random Fields; IDW = inverse-distance weighting; IQR = interquartile range; L = location; MAQSIP = Multiscale Air Quality Simulation Platform
MB = mean bias; ME = mean error; MEGAN = Model for Gases and Aerosols from Nature; MFB = mean fractional bias; MFE = mean fractional error; MODIS = Moderate Resolution
Imaging Spectroradiometer; NAPS = National Air Pollution Surveillance; NMB = normalized mean bias; NME = normalized mean error; NOAA = National Oceanographic and
Atmospheric Administration; OA = objective analysis; OMI = Ozone Monitoring Instrument; P = population; PI = prediction interval; PMAE = prediction mean absolute error;
PMSE = prediction mean squared error; R = correlation coefficient; RAMP = regional air quality model performance; RMSE = root-mean-squared error; RMSPE = root-mean-squared
prediction error; SHEDS = Stochastic Human Exposure and Dose Simulation; T = time; TCEQ = Texas Commission on Air Quality; TES = Tropospheric Emissions System;
TOPP = Tropospheric Ozone Pollution Project; U.S. = United States; WOUDC = World Ozone and Ultraviolet Data Centre; WRF = Weather Research Forecasting model; WRF-
ARW = Weather Research and Forecasting—Advanced Research Weather.
aData were obtained from the study author.
2-157
-------
2.9 References
Adam-Poupart. A; Brand. A; Fournier. M; Jerrett. M; Smargiassi. A. (2014). Spatiotemporal modeling
of ozone levels in Quebec (Canada): a comparison of kriging, land-use regression (LUR), and
combined Bayesian maximum entropy-LUR approaches. Environ Health Perspect 122: 970-976.
http://dx.doi.org/10.1289/ehp.1306566
Appel. KW; Chemel. C; Roselle. SJ; Francis. XV; Hu. RM: Sokhi. RS; Rao. ST; Galmarini. S. (2012).
Examination of the Community Multiscale Air Quality (CMAQ) model performance over the
North American and European domains. Atmos Environ 53: 142-155.
http ://dx.doi .org/10.1016/i .atmosenv.2011.11.016
Appel. KW; Napelenok. SL; Foley. KM; Pve. HOT; Hogrefe. C; Luecken. DJ; Bash. JO; Roselle. SJ;
Pleim. JE; Foroutan. H; Hutzell. WT; Pouliot. GA; Sarwar. G; Fahev. KM; Gantt. B; Gilliam. RC;
Heath. NK; Kang. D; Mathur. R; Schwede. DB; Spero. TL; Wong. DC; Young. JO. (2017).
Description and evaluation of the Community Multiscale Air Quality (CMAQ) modeling system
version 5.1. GMD 10: 1703-1732. http://dx.doi.org/10.5194/gmd-10-1703-2017
Appel. KW; Pouliot. GA; Simon. H; Sarwar. G; Pve. HOT; Napelenok. SL; Akhtar. F; Roselle. SJ.
(2013). Evaluation of dust and trace metal estimates from the Community Multiscale Air Quality
(CMAQ) model version 5.0. GMD 6: 883-899. http://dx.doi.org/10.5194/gmd-6-883-2013
Armstrong. BK; White. E; Saracci. R. (1992V Principles of exposure measurement in epidemiology.
New York, NY: Oxford University Press.
Baker. KR; Woody. MC; Tonnesen. GS; Hutzell. W; Pve. HOT; Beaver. MR; Pouliot. G; Pierce. T.
(2016). Contribution of regional-scale fire events to ozone and PM2.5 air quality estimated by
photochemical modeling approaches. Atmos Environ 140: 539-554.
http://dx.doi.Org/10.1016/i.atmosenv.2016.06.032
Barrett. SRH; Speth. RL; Eastham. SD; Dedoussi. IC; Ashok. A; Malina. R; Keith. DW. (2015).
Impact of the Volkswagen emissions control defeat device on US public health. Environ Res Lett
10. http://dx.doi.org/10.1088/1748-9326/10/ll/1140Q5
Bart. M; Williams. DE: Ainslie. B; Mckendrv. I. an; Salmond. J; Grange. SK; Alavi-Shoshtari. M;
Stevn. D; Henshaw. GS. (2014). High density ozone monitoring using gas sensitive semi-conductor
sensors in the lower fraser valley, british Columbia. Environ Sci Technol 48: 3970-3977.
http://dx.doi.org/10.1021/es404610t
Bash. JO; Baker. KR; Beaver. MR. (2016). Evaluation of improved land use and canopy
representation in BEIS v3.61 with biogenic VOC measurements in California. GMD 9: 2191-2207.
http ://dx.doi .org/10.5194/gmd-9-2191-2016
Baxter. LK; Burke. J; Lunden. M; Turpin. BJ; Rich. DO; Thevenet-Morrison. K; Hodas. N;
Oezkavnak. H. (2013). Influence of human activity patterns, particle composition, and residential
air exchange rates on modeled distributions of PM2.5 exposure compared with central-site
monitoring data. J Expo Sci Environ Epidemiol 23: 241-247.
http://dx.doi.org/10.1038/ies.2012.118
Bell. ML. (2006). The use of ambient air quality modeling to estimate individual and population
exposure for human health research: A case study of ozone in the Northern Georgia region of the
United States. Environ Int 32: 586-593. http://dx.doi.Org/10.1016/i.envint.2006.01.005
2-158
-------
Ben-David. T. om; Waring. MS. (2016). Impact of natural versus mechanical ventilation on simulated
indoor air quality and energy consumption in offices in fourteen US cities. Build Environ 104: 320-
336. htto://dx.doi.org/10.1016/i .buildenv.2016.05.007
Ben-David. T; Waring. MS. (2018). Interplay of ventilation and filtration: Differential analysis of cost
function combining energy use and indoor exposure to PM2.5 and ozone. Build Environ 128: 320-
335. htto://dx.doi.org/10.1016/i .buildenv.2017.10.025
Berrocal. VJ; Gelfand. AE: Holland. DM. (2012). Space-time data fusion under error in computer
model output: An application to modeling air quality. Biometrics 68: 837-848.
http://dx.doi.org/10.1111/i. 1541-0420.2011.01725.x
Billionnet. C; Sherrill. D: Annesi-Maesano. I. (2012). Estimating the health effects of exposure to
multi-pollutant mixture [Review]. Ann Epidemiol 22: 126-141.
http://dx.doi.Org/10.1016/i.annepidem.2011.11.004
Blanchard. CL: Hidv. GM: Tanenbaum. S; Edgerton. ES. (2011). NMOC, ozone, and organic aerosol
in the southeastern United States, 19992007: 3. Origins of organic aerosol in Atlanta, Georgia, and
surrounding areas. Atmos Environ 45: 1291-1302.
http://dx.doi.Org/10.1016/i.atmosenv.2010.12.004
Brauer. M: Amann. M: Burnett. RT: Cohen. A: Dentener. F: Ezzati. M: Henderson. SB:
Krzvzanowski. M; Martin. RV; Van Dingenen. R; van Donkelaar. A: Thurston. GD. (2012).
Exposure assessment for estimation of the global burden of disease attributable to outdoor air
pollution. Environ Sci Technol 46: 652-660. http://dx.doi.org/10.102l/es2025752
Bravo. MA: Fuentes. M; Zhang. Y; Burr. MJ; Bell. ML. (2012). Comparison of exposure estimation
methods for air pollutants: ambient monitoring data and regional air quality simulation. Environ
Res 116: 1-10. http://dx.doi.Org/10.1016/i.envres.2012.04.008
Brokamp. C: Rao. MB: Fan. Z: Ryan. PH. (2015). Does the elemental composition of indoor and
outdoor PM2.5 accurately represent the elemental composition of personal PM2.5? Atmos Environ
101: 226-234. http://dx.doi.Org/10.1016/i.atmosenv.2014.l 1.022
Buteau. S: Hatzopoulou. M: Crouse. PL: Smargiassi. A: Burnett. RT: Logan. T: Cavellin. LP:
Goldberg. MS. (2017). Comparison of spatiotemporal prediction models of daily exposure of
individuals to ambient nitrogen dioxide and ozone in Montreal, Canada. Environ Res 156: 201-230.
http://dx.doi.Org/10.1016/i.envres.2017.03.017
Carlton. AG: Baker. KR. (2011). Photochemical modeling of the Ozark isoprene volcano: MEGAN,
BEIS, and their impacts on air quality predictions. Environ Sci Technol 45: 4438-4445.
http://dx.doi.org/10.1021/es200Q50x
Carroll. RJ: Ruppert. P: Stefanski. LA. (2006). Measurement error in nonlinear models: A modern
perspective. London, England: Chapman & Hall.
Castellanos. P: Marufu. LT: Poddridge. BG: Taubman. BF: Schwab. JJ; Hains. JC: Ehrman. SH:
Pickerson. RR. (2011). Ozone, oxides of nitrogen, and carbon monoxide during pollution events
over the eastern United States: An evaluation of emissions and vertical mixing. J Geophys Res 116:
P16307. http://dx.doi.org/10.1029/2010JP01454Q
Chai. T: Kim. HC: Lee. P: Tong. P: Pan. L: Tang. Y: Huang. J: Mcaueen. J: Tsidulko. M: Stainer. I.
(2013). Evaluation of the United States National Air Quality Forecast Capability experimental real-
time predictions in 2010 using Air Quality System ozone and N02 measurements. GMP 6: 1831-
1850. http://dx.doi.org/10.5194/gmd-6-1831-2013
2-159
-------
Chan. WR; Cohn. S; Sidheswaran. M; Sullivan. DP; Fisk. WJ. (2014). Contaminant levels, source
strengths, and ventilation rates in California retail stores. Indoor Air 25: 381-392.
http://dx.doi.Org/10.l 111/ina. 12152
Chang. KL; Guillas. S; Fioletov. YE. (2015). Spatial mapping of ground-based observations of total
ozone. Atmos Meas Tech 8: 4487-4505. http://dx.doi.org/10.5194/amt-8-4487-2015
Chen. C; Zhao. B: Weschler. CJ. (2012). Assessing the influence of indoor exposure to "outdoor
ozone" on the relationship between ozone and short-term mortality in us communities. Environ
Health Perspect 120: 235-240. http://dx.doi.Org/10.1289/ehp.l 103970
Chen. D: Li. O: Stutz. J: Mao. Y: Zhang. L. i: Pikelnava. O; Tsai. J; Haman. C; Lefer. B:
Rappenglueck. B: Alvarez. SL: Neuman. JA: Flvnn. J: Roberts. JM: Nowak. JB: de Gouw. J;
Hollowav. J; Wagner. NL; Veres. P; Brown. SS; Rverson. TB; Warneke. C; Pollack. LB. (2013).
WRF-Chem simulation of NOx and 0-3 in the LA basin during CalNex-2010. Atmos Environ 81:
421-432. http://dx.doi.Org/10.1016/i.atmosenv.2013.08.064
Cho. S: Mceachern. P; Morris. R; Shah. T; Johnson. J: Nopmongcol. U. (2012). Emission sources
sensitivity study for ground-level ozone and PM2.5 due to oil sands development using air quality
modeling system: Part I- model evaluation for current year base case simulation. Atmos Environ
55: 533-541. http://dx.doi.Org/10.1016/i.atmosenv.2012.02.026
Choi. Y. (2014). The impact of satellite-adjusted NOx emissions on simulated NOx and 0-3
discrepancies in the urban and outflow areas of the Pacific and Lower Middle US. Atmos Chem
Phys 14: 675-690. http://dx.doi.org/10.5194/acp-14-675-2014
Christakos. G. (1990). A bayesian maximum-entropy view to the spatial estimation problem. Math
Geol 22: 763-777. http://dx.doi.org/10.1007/BF0089Q661
Clark. LP: Millet. DB: Marshall. JD. (2011). Air quality and urban form in U.S. urban areas: Evidence
from regulatory monitors. Environ Sci Technol 45: 7028-7035.
http://dx.doi.org/10.1021/es2006786
Cleveland. WS: Graedel. TE. (1979). Photochemical air pollution in the northeast United States.
Science 204: 1273-1278. http://dx.doi.org/10.1126/science.204.4399.1273
Cohan. DS: Hu. Y; Russell. AG. (2006). Dependence of ozone sensitivity analysis on grid resolution.
Atmos Environ 40: 126-135. http://dx.doi.Org/10.1016/i.atmosenv.2005.09.031
Cuchiara. GC: Li. X: Carvalho. J: Rappenglueck. B. (2014). Intercomparison of planetary boundary
layer parameterization and its impacts on surface ozone concentration in the WRF/Chem model for
a case study in Houston/Texas. Atmos Environ 96: 175-185.
http://dx.doi.Org/10.1016/i.atmosenv.2014.07.013
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/ies.2009.49
Davalos. AD: Luben. TJ; Herring. AH: Sacks. JD. (2017). Current approaches used in epidemiologic
studies to examine short-term multipollutant air pollution exposures [Review]. Ann Epidemiol 27:
145-153 .e 141. http://dx.doi.org/10.1016/i .annenidem.2016.11.016
Davis. J: Cox. W: Reff. A: Dolwick. P. at. (2011). A comparison of CMAQ-based and observation-
based statistical models relating ozone to meteorological parameters. Atmos Environ 45: 3481-
3487. http://dx.doi.Org/10.1016/i.atmosenv.2010.12.060
2-160
-------
Delfino. RJ: Coate. BP; Zeiger. RS; Seltzer. JM; Street. DH; Koutrakis. P. (1996). Daily asthma
severity in relation to personal ozone exposure and outdoor fungal spores. Am J Respir Crit Care
Med 154: 633-641. http://dx.doi.Org/10.1164/airccm.154.3.8810598
Di. Q; Rowland. S; Koutrakis. P; Schwartz. J. (2017). A hybrid model for spatially and temporally
resolved ozone exposures in the continental United States. J Air Waste Manag Assoc 67: 39-52.
http://dx.doi.org/10.1080/10962247.2016.120Q159
Dionisio. KL: Baxter. LK: Chang. HH. (2014). An empirical assessment of exposure measurement
error and effect attenuation in bipollutant epidemiologic models. Environ Health Perspect 122:
1216-1224. http://dx.doi.org/10.1289/ehp.1307772
Dionisio. KL: Nolte. CG; Spero. TL: Graham. S; Caraway. N; Foley. KM: Isaacs. KK. (2017).
Characterizing the impact of projected changes in climate and air quality on human exposures to
ozone. J Expo Sci Environ Epidemiol 27: 260-270. http://dx.doi.org/10.1038/ies.2016.81
Dutton. SM: Banks. D: Brunswick. SL: Fisk. WJ. (2013). Health and economic implications of natural
ventilation in California offices. Build Environ 67: 34-45.
http://dx.doi.Org/10.1016/i.buildenv.2013.05.002
Emery. C: Jung. J: Downey. N: .Johnson. J: Jimenez. M: Yarvvood. 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.Org/10.1016/i.atmosenv.2011.11.012
EPA. US. (2018). CMAQ (Version 5.2.1) [Computer Program]: Zenodo. Retrieved from
https://zenodo.org/record/1212601# .XGFtoVxKi IV
Fadevi. MO: Weschler. CJ: Tham. KW: Wu. WY: Sultan. ZM. (2013). Impact of human presence on
secondary organic aerosols derived from ozone-initiated chemistry in a simulated office
environment. Environ Sci Technol 47: 3933-3941. http://dx.doi.org/10.1021/es3050828
Ferreira. J: Rodriguez. A: Monteiro. A: Miranda. AI: Dios. M: Souto. JA: Yarwood. G: Nopmongcol.
U: Borrego. C. (2012). Air quality simulations for North America-MM5-CAMx modelling
performance for main gaseous pollutants. Atmos Environ 53: 212-224.
http://dx.doi.Org/10.1016/i.atmosenv.2011.10.020
Foley. GJ: Georgopoulos. PG: Liov. PJ. (2003). Accountability within new ozone standards [Review].
Environ Sci Technol 37: 392A-399A.
Friberg. MP: Zhai. X: Holmes. HA: Chang. HH: Strickland. MJ: Sarnat. SE: Tolbert. PE: Russell-
AG: Mulholland. JA. (2016). Method for fusing observational data and chemical transport model
simulations to estimate spatiotemporally resolved ambient air pollution. Environ Sci Technol 50:
3695-3705. http://dx.doi.org/10.1021/acs.est.5b05134
Fuentes. M. (2009). Statistical issues in health impact assessment at the state and local levels. Air Qual
Atmos Health 2: 47-55. http://dx.doi.org/10.1007/sll869-009-0Q33-3
Fuentes. M: Rafterv. AE. (2005). Model evaluation and spatial interpolation by Bayesian combination
of observations with outputs from numerical models. Biometrics 61: 36-45.
http://dx.doi.org/10.1111/i. 0006-341X.2005.030821.x
Gall. ET; Corsi. RL: Siegel. JA. (2011). Barriers and opportunities for passive removal of indoor
ozone. Atmos Environ 45: 3338-3341. http://dx.doi.Org/10.1016/i.atmosenv.2011.03.032
Ganguly. R; Batterman. S: Isakov. V: Snyder. M; Breen. M; Brakefield-Caldwell. W. (2015). Effect of
geocoding errors on traffic-related air pollutant exposure and concentration estimates :
Supplemental materials [Supplemental Data]. J Expo Sci Environ Epidemiol 25.
2-161
-------
Garner. GG; Thompson. AM; Lee. P; Martins. DK. (2015). Evaluation ofNAQFC model performance
in forecasting surface ozone during the 2011 DISCOVER-AQ campaign. J Atmos Chem 72: 483-
501. http://dx.doi.org/10.1007/slQ874-013-9251-z
Gelfand. AE; Sahu. SK; Holland. DM. (2012). On the Effect of Preferential Sampling in Spatial
Prediction. Environmetrics 23: 565-578. http://dx.doi.org/10.1002/env.2169
Georgopoulos. PG; Wang. SW; Vvas. VM: Sun. O: Burke. J; Vedantham. R: Mccurdv. T: Ozkavnak.
H. (2005). A source-to-dose assessment of population exposures to fine PM and ozone in
Philadelphia, PA, during a summer 1999 episode. J Expo Anal Environ Epidemiol 15: 439-457.
http://dx.doi.org/10.1038/si.iea.7500422
Glasgow. ML: Rudra. CB: Yoo. EH: Demirbas. M: Merriman. J: Navak. P: Crabtree-Ide. C; Szpiro.
AA; Rudra. A: Wactawski-Wende. J: Mu. L. (2014). Using smartphones to collect time-activity
data for long-term personal-level air pollution exposure assessment. J Expo Sci Environ Epidemiol
26: 356-364. http://dx.doi.org/10.1038/ies.2014.78
Godowitch. JM; Gilliam. RC: Rao. ST. (2011). Diagnostic evaluation of ozone production and
horizontal transport in a regional photochemical air quality modeling system. Atmos Environ 45:
3977-3987. http://dx.doi.Org/10.1016/i.atmosenv.2011.04.062
Godowitch. JM: Gilliam. RC: Roselle. SJ. (2015). Investigating the impact on modeled ozone
concentrations using meteorological fields from WRF with an updated four-dimensional data
assimilation approach. Atmos Pollut Res 6: 305-311. http://dx.doi.org/10.5094/APR.2015.034
Goldberg. MS. (2007). On the interpretation of epidemiological studies of ambient air pollution. J
Expo Sci Environ Epidemiol 17: S66-S70. http://dx.doi.org/10.1038/si.ies.7500629
Goldman. GT; Mulholland. JA; Russell. AG: Gass. K; Strickland. MJ; Tolbert. PE. (2012).
Characterization of ambient air pollution measurement error in a time-series health study using a
geostatistical simulation approach. Atmos Environ 57: 101-108.
http://dx.doi.Org/10.1016/i.atmosenv.2012.04.045
Goldman. GT: Mulholland. JA: Russell. AG: Srivastava. A: Strickland. MJ: Klein. M: Waller. LA:
Tolbert. PE; Edgerton. ES. (2010). Ambient air pollutant measurement error: characterization and
impacts in a time-series epidemiologic study in Atlanta. Environ Sci Technol 44: 7692-7698.
http://dx.doi.org/10.1021/esl01386r
Goldman. GT; Mulholland. JA; Russell. AG; Strickland. MJ; Klein. M; Waller. LA; Tolbert. PE.
(2011). Impact of exposure measurement error in air pollution epidemiology: Effect of error type in
time-series studies. Environ Health 10: 61. http://dx.doi.Org/10.l 186/1476-069X-10-61
Gong. X; Kaulfus. A; Nair. U; Jaffe. DA. (2017). Quantifying 03 impacts in urban areas due to
wildfires using a generalized additive model. Environ Sci Technol 51: 13216-13223.
http://dx.doi.org/10.1021/acs.est.7b03130
Graham. SE; Mccurdv. T. (2004). Developing meaningful cohorts for human exposure models. J Expo
Anal Environ Epidemiol 14: 23-43. http://dx.doi.org/10.1038/si.iea.7500293
Hackbarth. AD; Romlev. JA; Goldman. DP. (2011). Racial and ethnic disparities in hospital care
resulting from air pollution in excess of federal standards. Soc Sci Med 73: 1163-1168.
http://dx.doi.Org/10.1016/i.socscimed.2011.08.008
Hall. ES; Evth. AM; Phillips. SB; Mason. R. (2012). Hierarchical Bayesian model (HBM): Derived
estimates of air quality for 2008: Annual report. (EPA/600/R-12/048). U.S. Environmental
Protection Agency.
2-162
-------
Hao. Y; Flowers. H; Monti. MM; Qualters. JR. (2012). U.S. census unit population exposures to
ambient air pollutants. Int J Health Geogr 11:3. http://dx.doi.org/10.1186/1476-072X-11-3
He. J; Kolovos. A. (2018). Bayesian maximum entropy approach and its applications: a review. Stoch
Environ Res Risk Assess 32: 859-877. http://dx.doi.org/10.1007/sQ0477-017-1419-7
Henderson. BH; Jeffries. HE; Kim. BU; Vizuete. WG. (2010). The influence of model resolution on
ozone in industrial volatile organic compound plumes. J Air Waste Manag Assoc 60: 1105-1117.
Henneman. LRF; Chang. HH; Liao. K; Lavoue. D; Mulholland. JA; Russell. AG. (2017a).
Accountability assessment of regulatory impacts on ozone and PM2.5 concentrations using
statistical and deterministic pollutant sensitivities. Air Qual Atmos Health 10: 695-711.
http://dx.doi.org/10.1007/sll869-017-Q463-2
Henneman. LRF; Liu. C; Hu. Y; Mulholland. JA; Russell. AG. (2017b). Air quality modeling for
accountability research: Operational, dynamic, and diagnostic evaluation. Atmos Environ 166: 551-
565. http://dx.doi.Org/10.1016/i.atmosenv.2017.07.049
Herron-Thorpe. FL; Mount. GH; Emmons. LK; Lamb. BK; Jaffe. DA; Wigder. NL; Chung. SH;
Zhang. R; Woelfle. MP; Vaughan. JK. (2014). Air quality simulations of wildfires in the Pacific
Northwest evaluated with surface and satellite observations during the summers of 2007 and 2008.
Atmos Chem Phys 14: 12533-12551. http://dx.doi.org/10.5194/acp-14-12533-2014
Herwehe. JA; Otte. TL; Mathur. R; Rao. ST. (2011). Diagnostic analysis of ozone concentrations
simulated by two regional-scale air quality models. Atmos Environ 45: 5957-5969.
http://dx.doi.Org/10.1016/i.atmosenv.2011.08.011
Hogrefe. C; Pouliot. G; Wong. D; Torian. A; Roselle. S; Pleim. J; Mathur. R. (2015). Annual
application and evaluation of the online coupled WRF-CMAQ system over North America under
AQMEII phase 2. Atmos Environ 115: 683-694. http://dx.doi.Org/10.1016/i.atmosenv.2014.12.034
Hogrefe. C; Roselle. S; Mathur. R; Rao. ST; Galmarini. S. (2014). Space-time analysis of the Air
Quality Model Evaluation International Initiative (AQMEII) Phase 1 air quality simulations. J Air
Waste Manag Assoc 64: 388-405. http://dx.doi.org/10.1080/10962247.2013.811127
Hu. J; Howard. CJ; Mitloehner. F; Green. PG; Kleeman. MJ. (2012). Mobile source and livestock feed
contributions to regional ozone formation in Central California. Environ Sci Technol 46: 2781 -
2789. http://dx.doi.org/10.1021/es2Q3369p
Huang. L; Mcdonald-Buller. E; Mcgaughev. G; Kimura. Y; Allen. DT. (2015). Comparison of
regional and global land cover products and the implications for biogenic emission modeling. J Air
Waste Manag Assoc 65: 1194-1205. http://dx.doi.org/10.1080/10962247.2015.10573Q2
Hutzell. WT; Luecken. DJ; Appel. KW; Carter. WPL. (2012). Interpreting predictions from the
SAPRC07 mechanism based on regional and continental simulations. Atmos Environ 46: 417-429.
http://dx.doi.Org/10.1016/i.atmosenv.2011.09.030
Hvstad. P; Demers. PA: Johnson. KC; Brook. J; van Donkelaar. A; Lamsal. L; Martin. R; Brauer. M.
(2012). Spatiotemporal air pollution exposure assessment for a Canadian population-based lung
cancer case-control study. Environ Health 11: 22. http://dx.doi.org/10.1186/1476-069X-l 1-22
Isaacs. K. (2014). The consolidated human activity database - master version (CHAD-Master)
technical memorandum. Washington, DC: U.S. Environmental Protection Agency, National
Exposure Research Laboratory, https://www.epa.gov/sites/production/files/2015-
02/documents/chadmaster 091814 l.pdf
2-163
-------
Jaffe. DA; Wigder. N; Downey. N; Pfister. G; Boynard. A; Reid. SB. (2013). Impact of wildfires on
ozone exceptional events in the Western U.S. Environ Sci Technol 47: 11065-11072.
http://dx.doi.org/10.1021/es4Q2164f
Jerrett. M; Burnett. RT; Pope. CA. Ill; 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
Jiao. W; Frev. HC; Cao. Y. (2012). Assessment of inter-individual, geographic, and seasonal
variability in estimated human exposure to fine particles. Environ Sci Technol 46: 12519-12526.
http://dx.doi.org/10.1021/es302803g
Johnson. T; Capel. J; Ollison. W. (2014). Measurement of microenvironmental ozone concentrations
in Durham, North Carolina, using a 2B Technologies 205 Federal Equivalent Method monitor and
an interference-free 2B Technologies 211 monitor. J Air Waste Manag Assoc 64: 360-371.
http://dx.doi.org/10.1080/10962247.2Q13.839968
Johnson. TR; Langstaff. JE; Graham. S; Fujita. EM; Campbell. DE. (2018). A multipollutant
evaluation of APEX using microenvironmental ozone, carbon monoxide, and particulate matter
(PM2.5) concentrations measured in Los Angeles by the exposure classification project. 4.
http://dx.doi.org/10.1080/23311843.2018.1453Q22
Jones. RR; Oezkavnak. H; Navak. SG; Garcia. V; Hwang. SA; Lin. S. (2013). Associations between
summertime ambient pollutants and respiratory morbidity in New York City: Comparison of
results using ambient concentrations versus predicted exposures. J Expo Sci Environ Epidemiol 23:
616-626. http://dx.doi.org/10.1038/ies.2013.44
Joseph. J; Sharif. HO; Sunil. T; Alamgir. H. (2013). Application of validation data for assessing
spatial interpolation methods for 8-h ozone or other sparsely monitored constituents. Environ Pollut
178: 411-418. http://dx.doi.Org/10.1016/i.envpol.2013.03.035
Kang. D; Eder. BK; Stein. AF; Grell. GA; Peckham. SE; Mchenry. J. (2005). The New England Air
Quality Forecasting Pilot Program: development of an evaluation protocol and performance
benchmark. J Air Waste Manag Assoc 55: 1782-1796.
Karamchandani. P; Johnson. J; Yarwood. G; Knipping. E. (2014). Implementation and application of
sub-grid-scale plume treatment in the latest version of EPA's third-generation air quality model,
CMAQ 5.01. J Air Waste Manag Assoc 64: 453-467.
http://dx.doi.org/10.1080/10962247.2Q13.855152
Kavnak. B; Hu. Y; Russell. AG. (2013). Analysis of NO, N02, and 03 between model simulations
and ground-based, aircraft, and satellite observations. Water Air Soil Pollut 224: 1674.
http://dx.doi.org/10.1007/sll270-Q13-1674-2
Kethireddv. SR; Tchounwou. PB; Ahmad. HA; Yerramilli. A; Young. JH. (2014). Geospatial
interpolation and mapping of tropospheric ozone pollution using geostatistics. Int J Environ Res
Public Health 11: 983-1000. http://dx.doi.org/10.3390/iierphll0100983
Kimbrough. S; Owen. RC; Snyder. M; Richmond-Bryant. J. (2017). NO to NO 2 conversion rate
analysis and implications for dispersion model chemistry methods using Las Vegas, Nevada near-
road field measurements. Atmos Environ 165: 23-34.
http://dx.doi.Org/10.1016/i.atmosenv.2017.06.027
Koo. B; Kumar. N; Knipping. E; Nopmongcol. U; Sakulvanontvittava. T; Odman. MT; Russell. AG;
Yarwood. G. (2015). Chemical transport model consistency in simulating regulatory outcomes and
the relationship to model performance. Atmos Environ 116: 159-171.
http://dx.doi.Org/10.1016/i.atmosenv.2015.06.036
2-164
-------
Lai. D; Karava. P; Chen. Q. (2015). Study of outdoor ozone penetration into buildings through
ventilation and infiltration. Build Environ 93: 112-118.
http://dx.doi.Org/10.1016/i.buildenv.2015.06.015
Lane. KJ; Scammell. MK; Levy. JI; Fuller. CH; Parambi. R; Zamore. W; Mwamburi. M; Brugge. D.
(2013). Positional error and time-activity patterns in near-highway proximity studies: an exposure
misclassification analysis. Environ Health 12: 75. htto://dx.doi.org/10.1186/1476-069X-12-75
Li. G; Bei. N; Zavala. M: Molina. LT. (2014a). Ozone formation along the California Mexican border
region during Cal-Mex 2010 field campaign. Atmos Environ 88: 370-389.
http://dx.doi.Org/10.1016/i.atmosenv.2013.ll.067
Li. J: Georgescu. M: Hyde. P; Mahalov. A: Moustaoui. M. (2014b). Achieving accurate simulations of
urban impacts on ozone at high resolution. Environ Res Lett 9. http://dx.doi.org/10.lQ88/1748-
9326/9/11/114019
Li. J: Mahalov. A: Hyde. P. (2016a). Impacts of agricultural irrigation on ozone concentrations in the
Central Valley of California and in the contiguous United States based on WRF-Chem simulations.
Agr Forest Meteorol 221: 34-49. http://dx.doi.Org/10.1016/i.agrformet.2016.02.004
Li. J: Zhang. H: Ying. O. (2012). Comparison of the SAPRC07 and SAPRC99 photochemical
mechanisms during a high ozone episode in Texas: Differences in concentrations, OH budget and
relative response factors. Atmos Environ 54: 25-35.
http://dx.doi.Org/10.1016/i.atmosenv.2012.02.034
Li. X: Choi. Y: Czader. B: Roy. A: Kim. H: Lefer. B: Pan. S. (2016b). The impact of observation
nudging on simulated meteorology and ozone concentrations during DISCOVER-AQ 2013 Texas
campaign. Atmos Chem Phys 16: 3127-3144. http://dx.doi.org/10.5194/acp-16-3127-2016
Lipfert. FW; Wvzga. RE. (1996). The effects of exposure error on environmental epidemiology. In RF
Phalen; RC Mannix; MC Tonini (Eds.), The second colloquium on particulate air pollution &
human mortality & morbidity (pp. 295-302). Sacramento, CA: California Air Resources Board.
https://ww3.arb.ca.gov/research/apr/past/95-323d.pdf
Liu. P; Zhang. Y. (2011). Use of a process analysis tool for diagnostic study on fine particulate matter
predictions in the U.S. - Part I: Model evaluation. Atmos Pollut Res 2: 49-60.
http://dx.doi.org/10.5094/APR.2011.0Q7
Liu. Z: Le. N; Zidek. JV. (2011). An empirical assessment of Bayesian melding for mapping ozone
pollution. Environmetrics 22: 340-353. http://dx.doi.org/10.1002/env.1054
Lopiano. KK; Young. LJ; Gotwav. CA. (2011). A comparison of errors in variables methods for use in
regression models with spatially misaligned data. Stat Methods Med Res 20: 29-47.
http://dx.doi.org/10.1177/096228021037Q266
Lu. D; Cizdziel. JV; Jiang. Y; White. L; Reddv. RS. (2014). Numerical simulation of atmospheric
mercury in mid-south USA. Air Qual Atmos Health 7: 525-540. http://dx.doi.org/10.1007/sl 1869-
014-0256-9
Lu. Y. un; Zeger. SL. (2007). On the equivalence of case-crossover and time series methods in
environmental epidemiology. Biostatistics 8: 337-344.
http://dx.doi.org/10.1093/biostatistics/kxl013
Marshall. JD; Nethery. E; Brauer. M. (2008). Within-urban variability in ambient air pollution:
Comparison of estimation methods. Atmos Environ 42: 1359-1369.
http://dx.doi.Org/10.1016/i.atmosenv.2007.08.012
2-165
-------
Matichuk. R; Tonnesen. G; Luecken. D; Gilliam. R. ob; Napelenok. SL; Baker. KR; Schwede. D;
Murphy. B. en; Helmig. D; Lyman. SN; Roselle. S. (2017). Evaluation of the community
multiscale air quality model for simulating winter ozone formation in the uinta basin. J Geophys
ResAtmos 122: 13545-13572. http://dx.doi.org/10.1002/2017JD027Q57
McDonald-Buller. EC; Allen. DT; Brown. N; Jacob. DJ; Jaffe. D; Kolb. CE; Lefohn. AS; Oltmans. S;
Parrish. DP; Yarwood. G; Zhang. L. (2011). Establishing policy relevant background (PRB) ozone
concentrations in the United States [Review]. Environ Sci Technol 45: 9484-9497.
http://dx.doi.org/10.1021/es2022818
Muniz-Unamunzaga. M; Borge. R; Sarwar. G; Gantt. B; de la Paz. D; Cuevas. CA; Saiz-Lopez. A.
(2018). The influence of ocean halogen and sulfur emissions in the air quality of a coastal
megacity: The case of Los Angeles. Sci Total Environ 610-611: 1536-1545.
http://dx.doi.Org/10.1016/i.scitotenv.2017.06.098
Napelenok. SL; Foley. KM; Kang. D; Mathur. R; Pierce. T; Rao. ST. (2011). Dynamic evaluation of
regional air quality models response to emission reductions in the presence of uncertain emission
inventories. Atmos Environ 45: 4091-4098. http://dx.doi.Org/10.1016/i.atmosenv.2011.03.030
Ng. LC; Persilv. AK; Emmerich. SJ. (2015). IAQ and energy impacts of ventilation strategies and
building envelope airtightness in a big box retail building. Build Environ 92: 627-634.
http://dx.doi.Org/10.1016/i.buildenv.2015.05.038
Ngan. F; Bvun. D; Kim. H; Lee. D; Rappenglueck. B; Pour-Biazar. A. (2012). Performance
assessment of retrospective meteorological inputs for use in air quality modeling during TexAQS
2006. Atmos Environ 54: 86-96. http://dx.doi.Org/10.1016/i.atmosenv.2012.01.035
Nopmongcol. U; Alvarez. Y; Jung. J; Grant. J; Kumar. N; Yarwood. G. (2017). Source contributions
to United States ozone and particulate matter over five decades from 1970 to 2020. Atmos Environ
167: 116-128. http://dx.doi.Org/10.1016/i.atmosenv.2017.08.009
Ollison. WM; Crow. W; Spicer. CW. (2013). Field testing of new-technology ambient air ozone
monitors. J Air Waste Manag Assoc 63: 855-863. http://dx.doi.org/10.1080/10962247.2Q13.796898
Ozkavnak. H; Isakov. V; Baxter. L; Graham. SE; Sarnat. SE; Sarnat. JA; Mulholland. J; Turpin. B;
Rich. DQ; Lunden. M. (2014). Evaluating alternative exposure metrics used for multipollutant air
quality and human health studies. In NATO Science for Peace and Security Series C-
Environmental Security. Switzerland: Springer. http://dx.doi.org/10.1007/978-94-0Q7-5577-2 11
Pari. L; Gelfand. AE; Holland. DM. (2013). Spatio-temporal modeling for real-time ozone
forecasting. Spat Stat 4: 79-93. http://dx.doi.Org/10.1016/i.spasta.2013.04.003
Pan. L; Tong. D; Lee. P; Kim. HC; Chai. T. (2014). Assessment of NOx and 0-3 forecasting
performances in the US National Air Quality Forecasting Capability before and after the 2012
major emissions updates. Atmos Environ 95: 610-619.
http://dx.doi.Org/10.1016/i.atmosenv.2014.06.020
Pan. S; Choi. Y; Jeon. W; Rov. A; Westenbarger. DA; Kim. HC. (2017a). Impact of high-resolution
sea surface temperature, emission spikes and wind on simulated surface ozone in Houston, Texas
during a high ozone episode. Atmos Environ 152: 362-376.
http://dx.doi.Org/10.1016/i.atmosenv.2016.12.030
Pan. S; Choi. Y; Rov. A; Jeon. W. (2017b). Allocating emissions to 4 km and 1 km horizontal spatial
resolutions and its impact on simulated NOx and 0-3 in Houston, TX. Atmos Environ 164: 398-
415. http://dx.doi.Org/10.1016/i.atmosenv.2017.06.026
2-166
-------
Pan. S; Choi. Y; Roy. A; Li. X; Jeon. W; Souri. AH. (2015). Modeling the uncertainty of several VOC
and its impact on simulated VOC and ozone in Houston, Texas. Atmos Environ 120: 404-416.
http://dx.doi.Org/10.1016/i.atmosenv.2015.09.029
Pleim. J; Gilliam. R; Appel. W; Ran. L. (2016). Recent advances in modeling of the atmospheric
boundary layer and land surface in the coupled WRF-CMAQ model. In DG Steyn; N Chaumerliac
(Eds.), Air pollution modeling and its application XXIV (pp. 391-396). Cham, Switzerland:
Springer, http://dx.doi.org/10.1007/978-3-319-24478-5 64
Pongprueksa. P. (2013). Application of satellite data in a regional model to improve long-term ozone
simulations. J Atmos Chem 70: 317-340. http://dx.doi.org/10.1007/slQ874-013-9270-9
Punger. EM: West. JJ. (2013). The effect of grid resolution on estimates of the burden of ozone and
fine particulate matter on premature mortality in the USA. Air Qual Atmos Health 6: 563-573.
http://dx.doi.org/10.1007/sll869-013-Q197-8
Ran. L: Pleim. J: Gilliam. R: Binkowski. FS: Hogrefe. C: Band. L. (2016). Improved meteorology
from an updated WRF/CMAQ modeling system with MODIS vegetation and albedo. J Geophys
Res Atmos 121: 2393-2415. http://dx.doi.org/10.1002/2015JD0244Q6
Reeves. GK: Cox. PR: Darby. SC: Whitley. E. (1998). Some aspects of measurement error in
explanatory variables for continuous and binary regression models. Stat Med 17: 2157-2177.
http://dx.doi.org/10.1002/(SICD 1097-0258(19981015) 17:19<2157: :AID-SIM916>3.0.CQ;2-F
Reich. BJ: Chang. HH: Foley. KM. (2014). A spectral method for spatial downscaling. Biometrics 70:
932-942. http://dx.doi.Org/10.l 11 l/biom.12196
Reich. BJ: Kalendra. E; Storlie. CB; Bondell. HP: Fuentes. M. (2012). Variable selection for high
dimensional Bayesian density estimation: Application to human exposure simulation. J R Stat Soc
Ser C Appl Stat 61: 47-66. http://dx.doi.org/10.1111/i .1467-9876.2011,00772.x
Rim. D: Gall. ET: Ananth. S: Won. Y. (2018). Ozone reaction with human surfaces: Influences of
surface reaction probability and indoor air flow condition. Build Environ 130: 40-48.
http://dx.doi.Org/10.1016/i.buildenv.2017.12.012
Robichaud. A: Menard. R. (2014). Multi-year objective analyses of warm season ground-level ozone
and PM2.5 over North America using real-time observations and Canadian operational air quality
models. Atmos Chem Phys 14: 1769-1800. http://dx.doi.org/10.5194/acp-14-1769-2014
Sagona. JA: Weisel. CP: Meng. O. (2018). Accuracy and practicality of a portable ozone monitor for
personal exposure estimates. Atmos Environ 175: 120-126.
http://dx.doi.Org/10.1016/i.atmosenv.2017.ll.036
Sahu. SK: Bakar. KS. (2012a). A comparison of Bayesian models for daily ozone concentration levels.
Stat Methodol 9: 144-157. http://dx.doi.Org/10.1016/i.stamet.2011.04.009
Sahu. SK: Bakar. KS. (2012b). Hierarchical Bayesian autoregressive models for large space-time data
with applications to ozone concentration modelling. Appl Stoch Models Bus Ind 28: 395-415.
http://dx.doi.org/10.1002/asmb.1951
Samoli. E; Peng. RD: Ramsay. T. im; Touloumi. G: Dominici. F; Atkinson. RW: Zanobetti. A: Le
Tertre. A: Anderson. HR; Schwartz. J: Cohen. A: Krewski. D; Samet. JM; Katsouvanni. K. (2014).
What is the impact of systematically missing exposure data on air pollution health effect estimates?
Air Qual Atmos Health 7: 415-420. http://dx.doi.org/10.1007/sll869-014-0250-2
Sarnat. JA: Koutrakis. P; Suh. HH. (2000). Assessing the relationship between personal particulate and
gaseous exposures of senior citizens living in Baltimore, MD. J Air Waste Manag Assoc 50: 1184-
1198. http://dx.doi.org/10.1080/10473289.200Q.10464165
2-167
-------
Sarnat. JA; Sarnat. SE; Flanders. WD; Chang. HH: Mulholland. J; Baxter. L; Isakov. V; Ozkavnak. H.
(2013). Spatiotemporally resolved air exchange rate as a modifier of acute air pollution-related
morbidity in Atlanta. J Expo Sci Environ Epidemiol 23: 606-615.
http://dx.doi.org/10.1038/ies.2Q13.32
Schaap. M; Cuvelier. C; Hendriks. C; Bessagnet. B; Baldasano. JM; Colette. A; Thunis. P; Karam. D;
Fagerli. H: Graff A: Kranenburg. R: Nviri. A: Pay. MT: Rouil. L: Schulz. M: Simpson. D: Stern.
R; Terrenoire. E: Wind. P. (2015). Performance of European chemistry transport models as
function of horizontal resolution. Atmos Environ 112: 90-105.
http://dx.doi.Org/10.1016/i.atmosenv.2015.04.003
Schere. K: Flemming. J: Vautard. R; Chemel. C; Colette. A: Hogrefe. C; Bessagnet. B: Meleux. F:
Mathur. R; Roselle. S; Hu. RM: Sokhi. RS; Rao. ST; Galmarini. S. (2011). Trace gas/aerosol
boundary concentrations and their impacts on continental-scale AQMEII modeling domains.
Atmos Environ 53: 38-50. http://dx.doi.Org/10.1016/i.atmosenv.2011.09.043
Seltzer. KM; Nolte. CG; Spero. TL; Appel. KW; Xing. J. ia. (2016). Evaluation of near surface ozone
and particulate matter in air quality simulations driven by dynamically downscaled historical
meteorological fields. Atmos Environ 138: 42-54.
http://dx.doi.Org/10.1016/i.atmosenv.2016.05.010
Seltzer. KM; Shindell. DT; Faluvegi. G; Murray. LT. (2017). Evaluating Modeled Impact Metrics for
Human Health, Agriculture Growth, and Near-Term Climate. J Geophys Res Atmos 122: 13506-
13524. http://dx.doi.org/10.1002/2017JD02678Q
Sheppard. L; Slaughter. JC; Schildcrout. J; Liu. JS; Lumlev. T. (2005). Exposure and measurement
contributions to estimates of acute air pollution effects. J Expo Anal Environ Epidemiol 15: 366-
376. http://dx.doi.org/10.1038/si.iea.7500413
Shmool. JLC; Kinnee. E; Sheffield. PE; Cloughertv. JE. (2016). Spatio-temporal ozone variation in a
case-crossover analysis of childhood asthma hospital visits in New York City. Environ Res 147:
108-114. http://dx.doi.Org/10.1016/i.envres.2016.01.020
Simon. H; Baker. KR; Akhtar. F; Napelenok. SL; Possiel. N; Wells. B; Timin. B (2013). A direct
sensitivity approach to predict hourly ozone resulting from compliance with the national ambient
air quality standard. Environ Sci Technol 47: 2304-2313. http://dx.doi.org/10.1021/es303674e
Simon. H; Wells. B; Baker. KR; Hubbell. B. (2016). Assessing temporal and spatial patterns of
observed and predicted ozone in multiple urban areas. Environ Health Perspect 124: 1443-1452.
http://dx.doi.org/10.1289/EHP190
Singer. BC; Delp. WW; Black. PR; Walker. IS. (2016). Measured performance of filtration and
ventilation systems for fine and ultrafine particles and ozone in an unoccupied modern California
house. Indoor Air 27: 780-790. http://dx.doi.org/10. Ill 1/ina. 12359
Solazzo. E; Hogrefe. C; Colette. A; Garcia-Vivanco. M; Galmarini. S. (2017). Advanced error
diagnostics of the CMAQ and Chimere modelling systems within the AQMEII3 model evaluation
framework. Atmos Chem Phys 17: 10435-10465. http://dx.doi.org/10.5194/acp-17-10435-2Q17
Spalt. EW; Curl. CL; Allen. RW; Cohen. M; Adar. SD; Stukovskv. KH; Avol. E; Castro-Diehl. C;
Nunn. C; Mancera-Cuevas. K; Kaufman. JD. (2015). Time-location patterns of a diverse
population of older adults: the Multi-Ethnic Study of Atherosclerosis and Air Pollution (MESA
Air). J Expo Sci Environ Epidemiol 26: 349-355. http://dx.doi.org/10.1038/ies.2Q15.29
Springs. M; Wells. J. R.; Morrison. GC. (2011). Reaction rates of ozone and terpenes adsorbed to
model indoor surfaces. Indoor Air 21: 319-327. http://dx.doi.Org/10.l 11 l/j.1600-
0668.2010.00707.x
2-168
-------
Stevn. DG; Ainslie. B; Reuten. C; Jackson. PL. (2013). A retrospective analysis of ozone formation in
the lower Fraser Valley, British Columbia, Canada. Part I: Dynamical model evaluation. Atmos
Ocean 51: 153-169. http://dx.doi.org/10.1080/07055900.2013.78194Q
Strickland. MJ; Gass. KM; Goldman. GT; Mulholland. JA. (2013). Effects of ambient air pollution
measurement error on health effect estimates in time-series studies: a simulation-based analysis. J
Expo Sci Environ Epidemiol 25: 160-166. http://dx.doi.org/10.1038/ies.2013.16
Szpiro. AA: Paciorek. CJ. (2013). Measurement error in two-stage analyses, with application to air
pollution epidemiology. Environmetrics 24: 501-517. http://dx.doi.org/10.1002/env.2233
Szpiro. AA: Paciorek. CJ; Sheppard. L. (2011). Does more accurate exposure prediction necessarily
improve health effect estimates? Epidemiology 22: 680-685.
http://dx.doi.org/10.1097/EDE.0b013e3182254cc6
Tang. W; Cohan. DS; Morris. GA; Bvun. DW; Luke. WT. (2011). Influence of vertical mixing
uncertainties on ozone simulation in CMAQ. Atmos Environ 45: 2898-2909.
http://dx.doi.Org/10.1016/i.atmosenv.2011.01.057
Tang. W; Cohan. DS; Pour-Biazar. A; Lamsal. LN; White. AT; Xiao. X; Zhou. W; Henderson. BH;
Lash. BF. (2015a). Influence of satellite-derived photolysis rates and NOx emissions on Texas
ozone modeling. Atmos Chem Phys 15: 1601-1619. http://dx.doi.org/10.5194/acp-15-1601-2015
Tang. Y; Chai. T; Pan. L; Lee. P; Tong. D; Kim. HC; Chen. W. (2015b). Using optimal interpolation
to assimilate surface measurements and satellite AOD for ozone and PM2.5: A case study for July
2011. J Air Waste Manag Assoc 65: 1206-1216. http://dx.doi.org/10.1080/10962247.2Q15.1062439
Tao. Z; Yu. H; Chin. M. (2016). Impact of transpacific aerosol on air quality over the United States: A
perspective from aerosol-cloud-radiation interactions. Atmos Environ 125: 48-60.
http://dx.doi.Org/10.1016/i.atmosenv.2015.10.083
Tessum. CW; Hill. JD; Marshall. JD. (2015). Twelve-month, 12 km resolution North American WRF-
Chem v3.4 air quality simulation: performance evaluation. GMD 8: 957-973.
http://dx.doi.org/10.5194/gmd-8-957-2015
Thomas. D; Stram. D; Dwver. J. (1993). Exposure measurement error: Influence on exposure-disease
relationships and methods of correction [Review]. Annu Rev Public Health 14: 69-93.
http://dx.doi.Org/10.l 146/annurev.pu. 14.050193.000441
Thompson. TM; Selin. NE. (2012). Influence of air quality model resolution on uncertainty associated
with health impacts. Atmos Chem Phys 12: 9753-9762. http://dx.doi.org/10.5194/acp-12-9753-
2012
Tsimpidi. AP; Trail. M; Hu. Y; Nenes. A; Russell. AG. (2012). Modeling an air pollution episode in
northwestern United States: identifying the effect of nitrogen oxide and volatile organic compound
emission changes on air pollutants formation using direct sensitivity analysis. J Air Waste Manag
Assoc 62: 1150-1165. http://dx.doi.org/10.1080/10962247.2012.697Q93
U.S. EPA (U.S. Environmental Protection Agency). (2009). Human exposure modeling: Air pollutants
exposure model (APEX/TRIM.Expo Inhalation). Available online at
https://www.epa.gov/fera/human-exposure-modeling-air-pollutants-exposure-model (accessed June
13,2012).
U.S. EPA (U.S. Environmental Protection Agency). (2011). Air quality modeling technical support
document: Final EGUNESHAP. (EPA-454/R-l 1-009). U.S. Environmental Protection Agency,
Office of Air Quality Planning and Standards, Air Quality Assessment Division.
2-169
-------
U.S. EPA (U.S. Environmental Protection Agency). (2013). Integrated science assessment for ozone
and related photochemical oxidants [EPA Report]. (EPA/600/R-10/076F). Research Triangle Park,
NC: U.S. Environmental Protection Agency, Office of Research and Development, National Center
for Environmental Assessment-RTP Division.
http://cfpub.cpa. gov/ncca/isa/rccordisplav.cfm?dcid=247492
U.S. EPA (U.S. Environmental Protection Agency). (2014). Health Risk and Exposure Assessment for
Ozone. Final Report. (EPA-452/P-14-004a). Research Triangle Park, NC: Office of Air Quality
Planning and Standards, http://nepis. epa.gov/exe/ZvPURL.cgi ?Dockev=P 100KBUF.txt
U.S. EPA (U.S. Environmental Protection Agency). (2016). Stochastic human exposure and dose
simulation (SHEDS) to estimate human exposure to chemicals. Available online at
https://www.epa.gov/chemical-research/stochastic-human-exposure-and-dose-simulation-sheds-
estimate-human-exposure (accessed September 4, 2019).
U.S. EPA (U.S. Environmental Protection Agency). (2018). U.S. EPA. Integrated Science Assessment
(ISA) for Particulate Matter (External Review Draft). (EPA/600/R-18/179). Washington, DC.
https://cfpub.epa.gov/ncea/isa/recordisplav.cfm?deid=341593
Verstraeten. WW: Boersma. KF; Zorner. J; Allaart. MAF; Bowman. KW: Worden. J. R. (2013).
Validation of six years of TES tropospheric ozone retrievals with ozonesonde measurements:
implications for spatial patterns and temporal stability in the bias. Atmos Meas Tech 6: 1413-1423.
http://dx.doi.org/10.5194/amt-6-1413-2013
Wang. K. ai: Zhang. Y. (2014). 3-D agricultural air quality modeling: Impacts of NH3/H2S gas-phase
reactions and bi-directional exchange ofNH3. Atmos Environ 98: 554-570.
http://dx.doi.Org/10.1016/i.atmosenv.2014.09.010
Wang. K: Zhang. Y: Nenes. A: Fountoukis. C. (2012). Implementation of dust emission and chemistry
into the Community Multiscale Air Quality modeling system and initial application to an Asian
dust storm episode. Atmos Chem Phys 12: 10209-10237. http://dx.doi.org/10.5194/acp-12-102Q9-
2012
Wang. M: Keller. JP: Adar. SD: Kim. S: Larson. TV: Olives. C: Sampson. PD: Sheppard. L: Szpiro.
AA; Vedal. S: Kaufman. JD. (2015). Development of long-term spatiotemporal models for ambient
ozone in six metropolitan regions of the United States: The MESA Air study. Atmos Environ 123:
79-87. http://dx.doi.org/10.1016/i.atmosenv.2015.10.042
Wang. M; Sampson. PD: Hu. J: Kleeman. M; Keller. JP: Olives. C: Szpiro. AA: Vedal. S: Kaufman.
JD. (2016). Combining land-use regression and chemical transport modeling in a spatiotemporal
geostatistical model for ozone and PM2.5. Environ Sci Technol 50: 5111-5118.
http://dx.doi.org/10.1021/acs.est.5b06Q01
Waring. MS: Wells. JR. (2015). Volatile organic compound conversion by ozone, hydroxyl radicals,
and nitrate radicals in residential indoor air: Magnitudes and impacts of oxidant sources. Atmos
Environ 106: 382-391. http://dx.doi.Org/10.1016/i.atmosenv.2014.06.062
Warren. J: Fuentes. M; Herring. A: Langlois. P. (2012). Spatial-temporal modeling of the association
between air pollution exposure and preterm birth: Identifying critical windows of exposure.
Biometrics 68: 1157-1167. http://dx.doi.org/10.1111/i. 1541-0420.2012.01774,x
Weir. CH; Yeatts. KB: Sarnat. JA; Vizuete. W: Salo. PM; Jaramillo. R; Cohn. RD: Chu. H; Zeldin.
DC: London. SJ. (2013). Nitrogen dioxide and allergic sensitization in the 2005-2006 National
Health and Nutrition Examination Survey. RespirMed 107: 1763-1772.
http://dx.doi.Org/10.1016/i.rmed.2013.08.010
2-170
-------
Wheeler. AJ; Xu. X; Kulka. R; You. H; Wallace. L; Mallach. G; Van Rvswvk. K; MacNeill. M;
Kearney. J; Rasmussen. PE; Dabek-Zlotorzvnska. E; Wang. D; Poon. R; Williams. R; Stocco. C;
Anastassopoulos. A; Miller. JD; Dales. R; Brook. JR. (2011). Windsor, Ontario Exposure
Assessment Study: Design and methods validation of personal, indoor, and outdoor air pollution
monitoring. J Air Waste Manag Assoc 61: 324-338. http://dx.doi.Org/10.3155/1047-3289.61.3.324
Williams. R: Rappold. A; Case. M; Schmitt. M; Stone. S: Jones. P: Thornburg. J; Devlin. RB. (2012).
Multi-pollutant exposures in an asthmatic cohort. Atmos Environ 61: 244-252.
http://dx.doi.Org/10.1016/i.atmosenv.2012.07.049
Wilson. WE. (2000). Size distributions of ambient and indoor particles: does the overlap of fine and
coarse particles cause problems in the interpretation of research results?
Winquist. A; Kirrane. E; Klein. M; Strickland. M; Darrow. LA; Sarnat. SE; Gass. K; Mulholland. J;
Russell. A; Tolbert. P. (2014). Joint effects of ambient air pollutants on pediatric asthma
emergency department visits in Atlanta, 1998-2004. Epidemiology 25: 666-673.
http://dx.doi.org/10.1097/EDE.000000000000Q146
Wong. DC; Pleim. J; Mathur. R; Binkowski. F; Otte. T; Gilliam. R; Pouliot. G; Xiu. A; Young. JO;
Kang. D. (2012). WRF-CMAQ two-way coupled system with aerosol feedback: Software
development and preliminary results. GMD 5: 299-312. http://dx.doi.org/10.5194/gmd-5-299-2012
Wu. J; Jiang. C; Liu. Z; Houston. D; Jaimes. G; McConnell. R. (2010). Performances of different
global positioning system devices for time-location tracking in air pollution epidemiological
studies. Environ Health Insights 4: 93-108. http://dx.doi.org/10.4137/EHI.S6246
Xing. J; Mathur. R; Pleim. J; Hogrefe. C; Gan. CM; Wong. DC; Wei. C; Gilliam. R; Pouliot G.
(2015). Observations and modeling of air quality trends over 1990-2010 across the Northern
Hemisphere: China, the United States and Europe. Atmos Chem Phys 15: 2723-2747.
http://dx.doi.org/10.5194/acp-15-2723-2015
Xu. W; Rilev. EA; Austin. E; Sasakura. M; Schaal. L; Gould. TR; Hartin. K; Simpson. CD; Sampson.
PD; Yost. MG; Larson. TV; Xiu. G; Vedal. S. (2016a). Use of mobile and passive badge air
monitoring data for NOX and ozone air pollution spatial exposure prediction models. J Expo Sci
Environ Epidemiol 27: 184-192. http://dx.doi.Org/10.1038/ies.2016.9
Xu. Y; Serre. ML; Reves. J; Vizuete. W. (2016b). Bayesian maximum entropy integration of ozone
observations and model predictions: a national application. Environ Sci Technol 50: 4393-4400.
http://dx.doi.org/10.1021/acs.est.6b00Q96
Xu. Y; Serre. ML; Reves. JM; Vizuete. W. (2017). Impact of temporal upscaling and chemical
transport model horizontal resolution on reducing ozone exposure misclassification. Atmos
Environ 166: 374-382. http://dx.doi.Org/10.1016/i.atmosenv.2017.07.033
Yahva. K; He. J; Zhang. Y. (2015a). Multiyear applications of WRF/Chem over continental US:
Model evaluation, variation trend, and impacts of boundary conditions. J Geophys Res Atmos 120:
12748-12777. http://dx.doi.org/10.1002/2015JD023819
Yahva. K; Wang. K. ai; Campbell. P; Glotfeltv. T; He. J; Zhang. Y. (2016). Decadal evaluation of
regional climate, air quality, and their interactions over the continental US and their interactions
using WRF/Chem version 3.6.1. GMD 9: 671-695. http://dx.doi.org/10.5194/gmd-9-671-2016
Yahva. K; Wang. K; Zhang. Y; Kleindienst. TE. (2015b). Application of WRF/Chem over North
America under the AQMEII Phase 2-Part 2: Evaluation of 2010 application and responses of air
quality and meteorology-chemistry interactions to changes in emissions and meteorology from
2006 to 2010. GMD 8: 2095-2117. http://dx.doi.org/10.5194/gmd-8-2095-2015
2-171
-------
Yahva. K; Zhang. Y; Vukovich. JM. (2014). Real-time air quality forecasting over the southeastern
United States using WRF/Chem-MADRID: Multiple-year assessment and sensitivity studies.
Atmos Environ 92: 318-338. htto://dx.doi.org/10.1016/i.atmosenv.2014.04.024
Ying. Q; Li. J. (2011). Implementation and initial application of the near-explicit Master Chemical
Mechanism in the 3D Community Multiscale Air Quality (CMAQ) model. Atmos Environ 45:
3244-3256. http://dx.doi.Org/10.1016/i.atmosenv.2011.03.043
Yu. H: Russell. A: Mulholland. J; Odman. T: Hu. Y: Chang. HH: Kumar. N. (2018). Cross-
comparison and evaluation of air pollution field estimation methods. Atmos Environ 179: 49-60.
http://dx.doi.Org/l 0.1016/i .atmosenv.2018.01.045
Yu. K: Jacob. DJ; Fisher. JA: Kim. PS; Marais. EA: Miller. CC; Travis. KR: Zhu. L: Yantosca. RM:
Sulprizio. MP; Cohen. R; Dibb. JE; Fried. A; Mikovinv. T; Ryerson. TB; Wennberg. PO;
Wisthaler. A. (2016). Sensitivity to grid resolution in the ability of a chemical transport model to
simulate observed oxidant chemistry under high-isoprene conditions. Atmos Chem Phys 16: 4369-
4378. http://dx.doi.org/10.5194/acp-16-4369-2016
Yu. S; Mathur. R; Pleim. J; Pouliot. G; Wong. D; Eder. B; Schere. K; Gilliam. R; Rao. ST. (2012).
Comparative evaluation of the impact of WRF-NMM and WRF-ARW meteorology on CMAQ
simulations for 0-3 and related species during the 2006 TexAQS/GoMACCS campaign. Atmos
Pollut Res 3: 149-162. http://dx.doi.org/10.5094/APR.2012.015
Zartarian. VG; Xue. J; Ozkavnak. H; Dang. W; Glen. G; Smith. L; Stallins. C. (2005). A probabilistic
exposure assessment for children who contact CCA-treated playsets and decks: Using the
stochastic human exposure and dose simulation model for the wood preservative exposure scenario
(SHEDS-Wood). U.S. Environmental Protection Agency.
http://www.epa.gov/oppadOQ 1/reregistration/cca/final cca factsheet.htm
Zeger. SL; Thomas. D; Dominici. F: Samet. JM; Schwartz. J; Dockerv. D; Cohen. A. (2000).
Exposure measurement error in time-series studies of air pollution: Concepts and consequences.
Environ Health Perspect 108: 419-426. http://dx.doi.org/10.1289/ehp.00108419
Zhang. H; Chen. G; Hu. J; Chen. SH: Wiedinmver. C; Kleeman. M; Ying. O. (2014). Evaluation of a
seven-year air quality simulation using the Weather Research and Forecasting (WRF)/Community
Multiscale Air Quality (CMAQ) models in the eastern United States. Sci Total Environ 473-474:
275-285. http://dx.doi.org/10.1016/i.scitotenv.2013.11.121
Zhang. HL; Ying. Q. (2011). Contributions of local and regional sources of NO(x) to ozone
concentrations in Southeast Texas. Atmos Environ 45: 2877-2887.
http://dx.doi.Org/10.1016/i.atmosenv.2011.02.047
Zhang. O; Jenkins. PL. (2017). Evaluation of ozone emissions and exposures from consumer products
and home appliances. Indoor Air 27: 386-397. http://dx.doi.Org/10.l 111/ina. 12307
Zhang. Y; Olsen. KM; Wang. K. (2013). Fine scale modeling of agricultural air quality over the
southeastern United States using two air quality models. Part I. application and evaluation. Aerosol
Air Qual Res 13: 1231-1252. http://dx.doi.org/10.4209/aaar.2012.12.0346
Zhou. W; Cohan. DS; Napelenok. SL. (2013). Reconciling NOx emissions reductions and ozone
trends in the US, 2002-2006. Atmos Environ 70: 236-244.
http://dx.doi.Org/10.1016/i.atmosenv.2012.12.038
Zimmerman. N; Presto. AA; Kumar. SPN; Gu. J; Hauryliuk. A; Robinson. ES; Robinson. AL;
Subramanian. R. (2018). A machine learning calibration model using random forests to improve
sensor performance for lower-cost air quality monitoring. Atmos Meas Tech 11: 291-313.
http://dx.doi.org/10.5194/amt-l 1-291-2018
2-172
-------
APPENDIX 3 HEALTH EFFECTS —RESPIRATORY
Stimniiiry ofCiiiisulity Dctcvminations fov Short- and Long-icrm Ozone
Exposure and Respiratory ilffccts
This \ppeiidi\ cli;ir;iclcri/es llie scientific c\ idciicc lli:il siippDils c;ius;ilil\
dckTlllilKIIIDIls I'dI'sIlDI'l-;illd iDim-kTlll D/DI1C CXpDslire ;illd l'Cspir;ilDI'\ I ic;i 1111 cITcels I lie
l\ pes dI" siiidics c\ ;ilu;ilcd willim llns \ppcndi\ ;irc cdiisisIcih Willi llic d\ cnill scDpcDl'lhc IS \
;is deluded in llic hd;icc In ;isscssnm llic d\ cnill e\ idciicc. sircimllis ;ind limil;iliDiis dI'
nidiN idiul siudics were c\;ilu;ilcd h;iscd dm scientific cDiisideniliDiis decided mi llic \niic\ I'pr
\nivndi\ ' \1pic dcl;nls dm llic c;ius;il fr;niic\u>rk used k> re;icli iliese conclusions ;irc included
mi llic Piviimhlc Id llic ISA il S. Ill' V 151. I lie c\ idciicc presented iIii'diiuIidiii this
\ppcudi\ support I lie follow nm c;ius;ili|\ eDiielusiDiis
l;.\|)nsiiiv Duniiiiin
( ;ius;ilil\ l)i-kTiiiin;ilii>n
Shorl-lerm exposure
Causal
1 .ong-lenn exposure
Likely 1
o be causal
3.1 Short-Term Ozone Exposure
3.1.1 Introduction, Summary from the 2013 Ozone ISA, and Scope for Current
Review
The 2013 Ozone ISA concluded that "short-term ozone exposure is causally associated with
respiratory health effects" [see Chapter 6 of U.S. EPA (2013a) I. This conclusion was based largely on
controlled human exposure studies demonstrating ozone-related respiratory effects in healthy individuals.
Specifically, statistically significant decreases in group mean pulmonary function relative to ozone
exposures as low as 60 ppb were observed in young, healthy adults. Additionally, controlled human
exposure and toxicological studies demonstrated ozone-induced increases in respiratory symptoms, lung
inflammation, airway permeability, and airway responsiveness. The experimental evidence was supported
by strong evidence from epidemiologic studies. Specifically, these studies demonstrated associations
between ozone concentrations and respiratory hospital admissions and emergency department (ED) visits
across the U.S., Europe, and Canada. Most effect estimates ranged from a 1.6 to 5.4% increase in daily
3-1
-------
respiratory-related ED visits or hospital admissions in all-year analyses for unit increases1 in ambient
ozone concentrations. This evidence was further supported by a large body of individual-level
epidemiologic panel studies demonstrating associations of ambient ozone with respiratory symptoms in
children with asthma. Additionally, several multicity studies and a multicontinent study reported
associations between short-term increases in ambient ozone concentrations and increases in respiratory
mortality. Additional support for a causal relationship was provided by epidemiologic panel studies that
observed ozone-associated increases in indicators of airway inflammation and oxidative stress in children
with asthma. Few epidemiologic studies examined the shape of the concentration-response relationship.
Two studies provided no evidence of a threshold, and observed generally linear relationships between
short-term ozone exposure and asthma hospital admissions and pediatric asthma ED visits, with greater
uncertainty in the shape of the C-R relationship at the lower end of the distributions of 8-hour daily max
concentrations measured in the studies.
Across respiratory endpoints, mechanistic evidence indicated that antioxidant capacity may
modify the risk of respiratory morbidity associated with ozone exposure. The potentially elevated risk of
populations with diminished antioxidant capacity and the reduced risk of populations with enhanced
antioxidant capacity identified in epidemiologic studies was strongly supported by similar findings from
controlled human exposure studies and by evidence that characterizes ozone-induced decreases in
intracellular antioxidant levels as a potential mechanistic pathway for downstream effects.
Along with this mechanistic evidence, animal toxicological and controlled human exposure
studies demonstrated ozone-induced increases in airway responsiveness, decreased pulmonary function,
allergic responses, lung injury, impaired host defense, and airway inflammation. These findings provided
biological plausibility for epidemiologic associations of ambient ozone concentrations with lung function
and respiratory symptoms, hospital admissions, ED visits, and mortality. Together, the evidence
integrated across controlled human exposure, epidemiologic, and toxicological studies and across the
spectrum of respiratory health endpoints support the determination of a causal relationship between
short-term ozone exposure and respiratory health effects.
The following section on short-term ozone exposure and respiratory effects begins with an
overview of study inclusion criteria (Section 3.1.2) that defines the scope of the literature that was
considered for inclusion in the section. The ensuing section presents a discussion of biological plausibility
(Section 3.1.3) that provides background for the subsequent sections in which groups of related endpoints
are presented in the context of relevant disease pathways. The respiratory effects subsections are
organized by outcome group and aim to clearly characterize the extent of coherence among related
endpoints (e.g., hospital admissions, symptoms, inflammation). These outcome groups include respiratory
effects in healthy populations (Section 3.1.4). respiratory effects in populations with asthma
(Section 3.1.5). respiratory effects in other populations with pre-existing conditions (Section 3.1.6).
1 Effect estimates were standardized to a 40-, 30-, and 20-ppb unit increase for 1-hour max, 8-hour max, and
24-hour avg ozone, respectively.
3-2
-------
including COPD (Section 3.1.6.1). obese populations or populations with metabolic syndrome
(Section 3.1.6.2). and populations with pre-existing cardiovascular disease (Section 3.1.6.3). respiratory
infection (Section 3.1.7). combinations of respiratory related disease hospital admissions and ED visits
(Section 3.1.8). and respiratory mortality (Section 3.1.9). Finally, Section 3.1.10 comprises a discussion
of relevant issues for interpreting the epidemiologic evidence discussed in the preceding sections.
Throughout the sections on respiratory health effects, results from recent studies are evaluated in the
context of evidence provided by studies that were previously evaluated in the 2013 Ozone ISA (U.S.
EPA. 2013a). Study-specific details, including exposure time periods and exposure concentrations in
experimental studies, and study design, exposure metrics, and select results in epidemiologic studies are
presented in evidence inventories in Section 3.3.
3.1.2 Population, Exposure, Comparison, Outcome, and Study Design
(PECOS) Tool
The scope of this section is defined by a scoping tool that generally describes the relevant
Population, Exposure, Comparison, Outcome, and Study Design (PECOS). The PECOS tool defines the
parameters and provides a framework to help identify the relevant literature to inform the Ozone ISA.
Because the 2013 Ozone ISA concluded there is a causal relationship between short-term ozone exposure
and respiratory health effects, the recent epidemiologic studies evaluated in this ISA are limited to study
locations in the U.S. and Canada to provide a focus on study populations and air quality characteristics
that are most relevant to circumstances in the U.S. The studies evaluated and subsequently discussed
within this section were included if they satisfied all of the components of the following PECOS tool:
Experimental Studies:
• Population: Study populations of any controlled human exposure or animal toxicological study of
mammals at any lifestage
• Exposure: Short-term (on the order of minutes to weeks) inhalation exposure to relevant ozone
concentrations (i.e., <0.4 ppm for humans, <2 ppm for other mammals); while ozone
concentrations in animal toxicological studies appear high, it should be noted that deposition of
ozone resulting from a 2-hour exposure to 2 ppm ozone in a resting rat is roughly equivalent to
deposition of ozone resulting from a 2-hour exposure to 0.4 ppm ozone in an exercising human
(Hatch et al.. 1994). More recently, Hatch et al. (2013) showed that resting rats and resting
humans receive similar alveolar ozone doses
• Comparison: Human subjects serve as their own controls with an appropriate washout period or
groups may be compared at the same or varied exposure concentrations; or, in toxicological
studies of mammals and some human studies, an appropriate comparison group is exposed to a
negative control (i.e., clean air or filtered-air control)
• Outcome: Respiratory effects
• Study Design: Controlled human exposure studies and animal studies meeting the above criteria
3-3
-------
Epidemiologic Studies:
• Population: Any U.S. or Canadian population, including populations or lifestages that might be at
increased risk
• Exposure: Short-term exposure (on the order of hours to several days) to ambient concentrations
of ozone
• Comparison: Per unit increase (in ppb), or humans exposed to lower levels of ozone compared
with humans exposed to higher levels
• Outcome: Change in risk (incidence/prevalence) of respiratory effects
• Study Design: Epidemiologic studies consisting of panel, case-crossover, time-series, and
case-control studies, as well as cross-sectional studies with appropriate timing of exposure for the
health endpoint of interest
3.1.3 Biological Plausibility
This section describes biological pathways that potentially underlie respiratory health effects
resulting from short-term exposure to ozone. Figure 3-1 graphically depicts the proposed pathways as a
continuum of upstream events, connected by arrows, that may lead to downstream events observed in
epidemiologic studies. This discussion of how short-term exposure to ozone may lead to respiratory
health effects contributes to an understanding of the biological plausibility of epidemiologic results
evaluated later in Section 3.1.4 to Section 3.1.9. Note that the structure of the biological plausibility
sections and the role of biological plausibility in contributing to the weight-of-evidence analysis used in
the 2020 Ozone ISA are discussed in Section IS.4.2.
Evidence that short-term exposure to ozone may affect the respiratory tract generally informs two
proposed pathways (Figure 3-1). The first pathway begins with the activation of sensory nerves in the
respiratory tract that can trigger local reflex responses and transmit signals to regions of the central
nervous system that regulate autonomic outflow. The second pathway begins with injury, inflammation,
and oxidative stress responses, which are difficult to disentangle. Inflammation generally occurs as a
consequence of injury and oxidative stress, but it can also lead to further oxidative stress and injury due to
secondary production of reactive oxygen species (ROS) by inflammatory cells.
3-4
-------
Activation
of Sensory
Nerves in
Respiratory
Tract
Sho rt-
T erm
Ozone
Exposure
Decreased
Inspiratory
Capacity/
Pain on
Inspiration
Allergic
Sensitization
Allergic
Responses
Decrements in Lung
Function (FVC,
FEV,) and
Increased
Respiratory
Symptoms
Emergency
Department Visits/
Hospital
Admissions
Asthma
Exace rbation
Emergency
Department Visits/
Hospital
Admissions
Respiratory
Infections
Note: The boxes above represent the effects for which there is experimental or epidemiologic evidence related to ozone exposure,
and the arrows indicate a proposed relationship between those effects. Solid arrows denote evidence of essentiality as provided, for
example, by an inhibitor of the pathway or a genetic knockout model used in an experimental study involving ozone exposure.
Shading around multiple boxes is used to denote a grouping of these effects. Arrows may connect individual boxes, groupings of
boxes, and individual boxes within groupings of boxes. Progression of effects is generally depicted from left to right and color-coded
(gray, exposure; green, initial effect; blue, intermediate effect; orange, effect at the population level or a key clinical effect). Here,
population level effects generally reflect results of epidemiologic studies. When there are gaps in the evidence, there are
complementary gaps in the figure and the accompanying text below. The structure of the biological plausibility sections and the role
of biological plausibility in contributing to the weight-of-evidence analysis used in the 2020 Ozone ISA are discussed in
Section IS.4.2.
Figure 3-1 Potential biological pathways for respiratory effects following
short-term ozone exposure.
Activation of Sensory Nerves in the Respiratory Tract
As discussed in the 2013 Ozone ISA [Section 5.3.2 of U.S. EPA (2013a)l. airway sensory nerves
in the lower respiratory tract are vagal afferents that carry signals to the nucleus tractus solitarius in the
brain. Signals are integrated in the brain and may result in altered autonomic activity that affects the lung
(e.g., airway obstruction) or other organs (e.g., altered heart rhythm). In addition, activation of some types
of sensory nerves (e.g., C-fibers) leads to local axon reflex responses in the airways that result in altered
3-5
-------
ventilatory parameters (e.g., altered breathing frequency and inspiratory capacity) and airway obstruction.
The release of substance P or other tachykinins from C-fibers and subsequent binding to neurokinin
receptors in the airway has been identified as a mechanism underlying local axon reflex responses.
Tachykinins, which mediate bronchoconstriction and neurogenic inflammation, can contribute to
increased airway responsiveness. These reflexes at the central-nervous-system or local axon level serve as
lung irritant responses—adaptive responses to noxious chemicals that help decrease exposure to that
chemical. Activation of vagal afferent pathways in the respiratory tract may also affect stress-responsive
regions of the brain and lead to neuroendocrine stress responses that have multiple systemic effects.
Controlled human exposure studies described in the 2013 Ozone ISA (U.S. EPA. 2013a)
demonstrated the involvement of sensory nerves and subsequent reflex responses in ozone-induced
changes in lung function (Figure 3-1). In studies using pharmacological tools, nociceptive sensory nerves,
presumably bronchial and pulmonary C-fibers, were identified as linking ozone exposure to a local axon
reflex response that resulted in pain-related respiratory symptoms and inhibition of maximal inspiration.
This mechanism underlies the observed decrease in forced vital capacity (FVC) and contributes to the
observed decrease in forced expiratory volume in 1 second (FEVi) in humans exposed to ozone. The
essentiality of this mechanism is depicted by the solid lines linking activation of sensory nerves to local
reflex responses to decreased inspiratory capacity and increased respiratory symptoms depicted in
Figure 3-1. Activation of airway sensory nerves also led to rapid shallow breathing in human subjects
exposed to ozone. Similarly, animal toxicological studies described in the 2013 Ozone ISA (U.S. EPA.
2013a) and more recent studies reviewed later in this Appendix demonstrated ozone-induced rapid
shallow breathing and other changes in lung ventilatory parameters. Supportive evidence for a role of
sensory C-fibers is provided by the association between airway responses to ozone and sensitivity to
capsaicin, which is a known activator of sensory C-fibers, found in a recent study in humans (Hoffmever
et al.. 2013). In addition, a recent study in animals demonstrates the involvement of TRPVi receptors,
which are a type of sensory nerve receptor, in ozone-mediated cough (Clay et al.. 2016).
Mild airway obstruction, measured as a change in specific airway resistance (sRaw), was
observed in humans exposed to ozone (U.S. EPA. 2013a). This response was inhibited by pretreatment
with atropine, an inhibitor of muscarinic cholinergic receptors of the parasympathetic nervous system.
This pathway is depicted by the solid lines linking activation of sensory nerves to modulation of the
autonomic nervous system and to airway obstruction in Figure 3-1. Airway obstruction may contribute to
decreases in FEVi. Studies in humans and animals indicate that airway obstruction resulting from
exposure to ozone is also mediated by a local axon reflex through the release of substance P from sensory
nerves. Thus, two mechanisms may contribute to ozone-induced airway obstruction—a parasympathetic
cholinergic pathway and a substance P-mediated pathway. Furthermore, the autonomic nervous system is
implicated in ozone-mediated effects on heart rhythm (Section 4.1.3).
Ozone exposure increased airway responsiveness in humans (U.S. EPA. 2013a). This effect was
blocked by atropine pretreatment, implicating a parasympathetic cholinergic process. This mechanism is
3-6
-------
depicted by the solid lines linking activation of sensory nerves to modulation of the autonomic nervous
system and to increased airway responsiveness in Figure 3-1. A correlation was found between airway
responsiveness to methacholine and increases in sRaw in ozone-exposed humans, pointing to an effect of
ozone on the parasympathetic nervous system that affects both responses. Animal toxicological studies
also provide evidence for ozone-induced increases in airway responsiveness and demonstrate the
involvement of the parasympathetic cholinergic pathway, the substance P-mediated pathway, and
contributions from arachidonic acid metabolites and cytokines/chemokines (U.S. EPA. 2013a). A recent
study showing enhanced bronchoconstriction in a model of vagal nerve electrical stimulation (Vcrhcin et
al.. 2013) provides additional evidence for ozone-induced increased airway responsiveness. Because
airway smooth muscle contraction to electrical field stimulation is a measure of post-ganglionic and
parasympathetic-mediated processes, this study provides support for the parasympathetic pathway.
Another recent study found that ozone exposure increased release of tachykinins (Barker et al.. 2015).
further supporting a role for a local axon reflex in mediating the effects of ozone.
Animal models of allergic airway disease share similar phenotypic features with asthma and are
used as a surrogate for human asthma. Airway responsiveness was enhanced by ozone exposure to a
larger degree in animals with allergic airway disease than in animals without allergic airway disease (U.S.
EPA. 2013a; Schclcglc and Walbv. 2012). This increased airway responsiveness occurred in response to
allergens (specific airway responsiveness) and nonallergens (nonspecific airway responsiveness).
Moreover, airway resistance in response to ozone exposure was increased to a greater degree in allergic
animals than in nonallergic animals (Schelegle and Walbv. 2012). This increase in airway resistance was
due to bronchoconstriction, and not to other mechanisms that could lead to airway obstruction. The role of
vagal afferents in mediating ozone-induced increased airway responsiveness and bronchoconstriction was
evaluated by vagotomy and by using pharmacologic tools (Schelegle and Walbv. 2012). Vagal lung
C-fibers were found to mediate reflex bronchoconstriction and enhance specific airway responsiveness
resulting from ozone exposure. Evidence from this study supports an essential role for activation of
sensory nerves and the parasympathetic nervous system in enhancing airway responses to ozone in an
allergy model. Vagal myelinated fibers mediated an opposing effect (i.e., reflex bronchodilation). A role
for neuropeptides such as substance P in mediating the bronchoconstrictive response was also suggested.
Other lung irritation effects, besides reflex bronchoconstriction and bronchodilation, were demonstrated
in recent studies of allergic airway disease in animals. Ozone exposure was associated with sensory
(i.e., upper airway) and pulmonary (i.e., lower airway) irritation in nonallergic animals, but only sensory
irritation in allergic animals (Hansen et al.. 2016). Ozone-induced rapid, shallow breathing was greater in
allergic animals than in nonallergic animals.
Taken together, mechanistic studies may provide biological plausibility for results of
epidemiologic panel studies in healthy children and in children with asthma, in which ozone
concentrations were associated with decrements in lung function and increased asthma symptoms.
Furthermore, they support results of epidemiologic studies showing associations between ozone exposure
and asthma-related emergency department visits and hospital admissions.
3-7
-------
Injury, Inflammation, and Oxidative Stress
Regarding the second pathway, a large body of evidence from controlled human exposure and
animal toxicological studies found injury, inflammation, and oxidative stress responses in healthy
individuals and animals exposed to ozone. As described in the 1996 and 2006 Ozone AQCDs (U.S. EPA.
2006. 1996a) and the 2013 Ozone ISA (U.S. EPA. 2013a). some studies in humans found increased
numbers of neutrophils, a marker of inflammation, and shed epithelial cells, a marker of injury, in
bronchoalveolar lavage fluid (BALF) or sputum. BALF neutrophils correlated in number with sRaw but
not with changes in lung volumes. In addition, BALF neutrophils and epithelial cells correlated with the
loss of substance P immunoreactivity from neurons in the bronchial mucosa. These findings suggest a
common mechanism underlying airway obstruction and inflammation, which possibly involves
substance P. In addition, studies from the 2013 Ozone ISA (U.S. EPA. 2013a). and one published more
recently (Alexis et al.. 2013). show that the glutathione S-transferase mu l(GSTMl) genotype may affect
the inflammatory response to ozone in young adults and suggest that greater antioxidant capacity may
mitigate the effects of ozone.
Animal toxicological studies described in the 2013 Ozone ISA (U.S. EPA. 2013a) found evidence
of oxidative stress resulting from exposure to ozone. This evidence includes decreased levels of ascorbate
in the BALF of rodents and decreased levels of glutathione in the respiratory bronchioles of monkeys. In
addition, ascorbate deficiency enhanced ozone-induced lung injury. Further support for a role of oxidative
stress is provided by a recent study in which Vitamin E supplementation dampened inflammation and
airway responsiveness in ozone-exposed animals (Zhu et al.. 2016). This relationship is depicted by the
solid lines linking respiratory tract oxidative stress to respiratory tract inflammation and to increased
airway responsiveness in Figure 3-1. Other studies described in the 2013 Ozone ISA (U.S. EPA. 2013a)
found evidence of injury, including increased flux of small solutes from the lung to the plasma and
increases in total protein, albumin, and shed epithelial cells in the BALF. In addition, markers of
inflammation such as BALF neutrophils and cytokines were observed. Recent studies provide additional
evidence for injury and inflammation.in animals exposed to ozone (Section 3.1.4.4.2). Taken together,
these studies may provide biological plausibility for results of epidemiologic panel studies in healthy
children and children with asthma, in which ozone was associated with markers of pulmonary
inflammation and oxidative stress.
Studies described in the 2013 Ozone ISA (U.S. EPA. 2013a) and more recent studies reviewed
later in this Appendix provide evidence for numerous cell signaling pathways underlying these oxidative
stress, injury, and inflammatory responses. One recent study implicated glucocorticoids in mediating
respiratory tract injury and inflammation resulting from ozone exposure (Miller et al.. 2016b). In this
study, adrenalectomy blocked the effects of ozone on the respiratory tract. This relationship is depicted by
the solid lines linking the neuroendocrine stress response to respiratory tract injury and inflammation in
Figure 3-1. Evidence for a neuroendocrine stress response initiated by ozone-mediated activation of
sensory nerves and propagated systemically has recently been described (see Appendix 7).
3-8
-------
Downstream effects of inflammation may result in morphologic changes. The 2013 Ozone ISA
(U.S. EPA. 2013a) described mild morphologic changes, such as hyperplasia of the bronchoalveolar duct
(rodents) and respiratory bronchioles (monkeys) and mucous cell metaplasia of the nasal epithelium
(rodents and monkeys). Recent studies also provide evidence of morphologic changes in the upper and
lower respiratory tracts following subacute or repeated exposure to ozone (Harkema et al.. 2017; Kumagai
et al.. 2017; Kumagai et al.. 2016; Ong et al.. 2016; Cho et al.. 2013V
In addition to activating the innate immune system, which is demonstrated by increases in airway
neutrophils, ozone exposure affects the adaptive immune system. Alterations in antigen presentation and
co-stimulation by innate immune cells such as macrophages and dendritic cells may lead to T-cell
activation, which may enhance host defense or allergic responses. The 2013 Ozone ISA (U.S. EPA.
2013a) describes altered antigen presentation in macrophages and dendritic cells in humans exposed to
ozone. Recent studies in animals exposed to ozone demonstrate dendritic cell activation (Brand et al..
2012) and a role for macrophage and T-lymphocyte subpopulations in the resolution of ozone-induced
inflammation (Mathews et al.. 2015). However, ozone exposure impairs, rather than enhances, host
defense. The 2013 Ozone ISA (U.S. EPA. 2013a) describes evidence for altered macrophage function,
decreased mucociliary clearance, and increased susceptibility to infectious disease in animals exposed to
ozone. Recent studies in animals provide further evidence for increased susceptibility to infection
following ozone exposure (Durrani et al.. 2012V This demonstration of impaired host defense provides
plausibility for epidemiologic findings indicating an association between short-term ozone concentrations
and respiratory infection.
However, ozone skews immune responses towards an allergic phenotype. The 2013 Ozone ISA
(U.S. EPA. 2013a) describes an animal study in which increased numbers of IgE-containing cells were
found in the lungs of mice exposed repeatedly to ozone. Recent studies found increased serum
immunoglobulin E (IgE), T helper 2 (Th2) cytokines, thymic stromal lymphopoietin (TSLP), eosinophilic
inflammation, and a role for immune lymphoid cells in the development of type 2 immune responses in
the upper and lower respiratory tract (Harkema et al.. 2017; Kumagai et al.. 2017; Kumagai et al.. 2016;
Ong et al.. 2016; Zhu et al.. 2016V Vitamin E supplementation was found to dampen allergic responses,
implicating ozone-mediated oxidative stress (Zhu et al.. 2016). This mechanism is depicted by the solid
line connecting respiratory tract oxidative stress and allergic responses in Figure 3-1.
In addition, ozone exposure enhances allergic responses in humans with asthma and in animal
models of allergic airway disease. Controlled human exposure studies described in the 2013 Ozone ISA
(U.S. EPA. 2013a) provide evidence of increased airway eosinophils and Th2 cytokines, increased
expression of IgE receptors on macrophages, and increased expression of CD86 in human subjects with
atopy and asthma. Enhanced nasal and airway eosinophilia were seen in human subjects with asthma who
were exposed first to allergen and then to ozone. In addition, there is increased uptake of particles by
airway macrophages of human subjects with asthma that may also enhance the processing of particulate
antigens and lead to greater progression of allergic airway disease. In animal models of allergic airway
3-9
-------
disease, ozone exposure leads to enhanced allergic inflammatory responses, goblet cell metaplasia, and
upregulated mucin expression, as described in the 2013 Ozone ISA (U.S. EPA. 2013a) and in several
recent studies (Harkcma et al.. 2017; Kumagai et al.. 2017; Kumagai et al.. 2016; Ong et al.. 2016; Baoet
al.. 2013).
Summary
As described here, there are two proposed pathways by which short-term exposure to ozone may
lead to respiratory health effects. One pathway involves the activation of sensory nerves in the respiratory
tract leading to lung function decrements, airway obstruction, and increased airway responsiveness. The
second pathway involves respiratory tract injury, inflammation, and oxidative stress that may lead to
morphologic changes and an allergic phenotype. Respiratory tract inflammation may also lead to altered
host defense, which is linked to increased respiratory infections. While experimental studies involving
animals or human subjects contribute most of the evidence of upstream effects, epidemiologic studies
found associations between exposure to ozone and markers of respiratory tract inflammation, lung
function decrements, and ED visits and hospital admissions for asthma and respiratory infection.
Together, these proposed pathways provide biological plausibility for epidemiologic evidence of
respiratory health effects and will be used to inform a causality determination, which is discussed later in
this Appendix (Section 3.1.11).
3.1.4 Respiratory Effects in Healthy Populations
3.1.4.1 Lung Function
3.1.4.1.1 Controlled Human Exposure Studies
Controlled human exposure studies have provided strong and quantifiable exposure response data
on the human health effects of ozone for decades (U.S. EPA. 1997). Respiratory responses to acute ozone
exposures in the range of ambient concentrations (i.e., <80 ppb) include decreased inspiratory capacity;
mild bronchoconstriction as demonstrated by increases in sRaw; rapid, shallow breathing patterns during
exercise; and symptoms of cough and pain on deep inspiration. Reflex inhibition of inspiration results in a
decrease in forced vital capacity (FVC) and total lung capacity (TLC) and, in combination with mild
bronchoconstriction (i.e., airway obstruction), contributes to a decrease in the forced expiratory volume in
1 second (FEVi). Reductions in FVC and increases in sRaw appear to be mediated by different
mechanisms.
3-10
-------
Most of the controlled human exposure studies described in this Appendix have exposed subjects
to a constant ozone concentration in a chamber. Cases where subjects were exposed to ozone via a
facemask are indicated in the text or figures. However, similar responses between facemask and chamber
exposures have been reported for exposures to 80 and 120 ppb O3 (6.6-hour, moderate quasi-continuous
exercise, 50 minute periods of exercise and 10 minutes of rest each hour, at 40 L/minute) and 300 ppb O3
[2 hours, heavy intermittent exercise, alternating 15 minute periods of exercise and rest, at 70 L/minute;
Adams (2003a. 2003b. 2002)1. Some studies [e.g., Adams (2006) and Schelegle et al. (2009)1 have
increased the ozone concentration in a chamber in a step-wise manner (e.g., rapid change from 70 to
80 ppb each hour over the first 3 to 4 hours of exposure) and then subsequently decreased ozone
concentration (e.g., from 80 to 50 ppb on an hourly basis) to achieve a targeted average ozone
concentration over a 6.6-hour exposure. Although greater peak responses have been observed in stepwise
and triangular (smooth increases and decreases in concentration) exposures versus constant concentration
exposure protocols, similar FEVi responses have been reported at 6.6 hours regardless of the exposure
protocol (i.e., constant vs. stepwise) for average ozone exposures to 60, 80, and 120 ppb (Adams. 2006.
2003a; Adams and Ollison. 1997).
The most salient observations from studies reviewed in the 1996 and 2006 ozone AQCDs (U.S.
EPA. 2006. 1996a) include: (1) young healthy adults exposed to >80 ppb ozone develop significant
reversible, transient decrements in pulmonary function and symptoms of breathing discomfort if minute
ventilation (Ve) or duration of exposure is increased sufficiently; (2) relative to young adults, children
experience similar spirometric responses but lower incidence of symptoms from ozone exposure;
(3) relative to young adults, ozone-induced spirometric responses are decreased in older individuals;
(4) there is a large degree of inter-subject variability in physiologic and symptomatic responses to ozone,
but responses tend to be reproducible within a given individual over a period of several months; and
(5) subjects exposed repeatedly to ozone for several days experience an attenuation of spirometric and
symptomatic responses on successive exposures, which is lost after about a week without exposure.
Mechanistic studies conducted in humans described in the 2013 Ozone ISA (U.S. EPA. 2013a)
contributed to the understanding of neurogenic mechanisms underlying lung function responses in
humans exposed to ozone while exercising. Controlled human exposure studies involving exposure to
400-420 ppb ozone provided evidence that nociceptive sensory nerves, presumably bronchial C-fibers,
were responsible for pain-related symptoms and inhibition of maximal inspiration that resulted in
decreased FVC. Eicosanoids, which are products of arachidonic acid metabolism, may also play a role in
this response. Mild airway obstruction, measured as changes in sRaw, in response to ozone exposure, is
inhibited by pretreatment with atropine, indicating the involvement of the parasympathetic nervous
system. Tachykinins may also contribute to increases in sRaw because ozone exposure (250 ppb)
increased substance P in BALF. Moreover, ozone exposure (200 ppb) resulted in a loss of substance P
immunoreactivity in the neurons of the bronchial mucosa. Substance P is released by sensory nerves and
mediates neurogenic edema and bronchoconstriction. Thus, increased sRaw may be attributed to vagally
mediated pathways and to local axon reflexes.
3-11
-------
The 2013 Ozone ISA (U.S. EPA. 2013a) included the FEVi responses of 150 young healthy
adults exposed to 60 ppb [targeted concentration; Kim et al. (2011); Schelegle et al. (2009); Adams
(2006)11 and 31 young healthy adults exposed to 70 ppb (targeted concentration) ozone (Schelegle et al..
2009) for 6.6 hours during quasi-continuous exercise (i.e., 50-minute exercise period and 10 minutes of
rest each hour). The moderate exercise level used in these studies is equivalent to walking at a pace of
17 to 18 minutes per mile at a grade of 4 to 5%. Although this is a relatively slow-paced walk, it does
account for an average of about 17 miles of walking over six 50-minute exercise periods. On average
across studies, the exposures to 60 ppb ozone resulted in a group mean FEVi decrement of 2.7%, with
10% of the exposed subjects experiencing greater than a 10% decrement in FEVi (see Figure 3-2).
Although not consistently statistically significant, these group mean changes in FEVi at 60 ppb are
consistent across studies (i.e., none observed an average improvement in lung function following a
6.6-hour exposure to 60 ppb ozone). There were no statistically significant effects in respiratory
symptoms reported in any of the studies at 60 ppb ozone.
1 Adams (2006) and Adams (2002) provide data for an additional group (30 of the 150) healthy subjects that were
exposed via facemask to 60 ppb (constant concentration or square-wave exposure profile) ozone for 6.6 hours with
moderate quasi-continuous exercise (Ve = 23 L/minute per m2 body surface area [BSA]). These subjects are
described on page 133 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/minute per m2 BSA relative to other studies that had a target Ve of
20 L/minute per m2 BSA.
3-12
-------
0.7 j
0.6
0.4 --
>. 0.5 :-
u
c
a;
ZJ
a- 0.3 3-
a;
LL 0.2
0.1
0
~ 0 ppb 040 ppb
0% 0%
60 ppb B 80 ppb
10% 30% with >10% FEV-l
-10 to -5
-5 toO 0 to 5 5 to 10 10 to 15
FEVi decrement (%)
>15
The illustrated data are for 30 subjects in the study conducted by Adams (20061. FEVi decrements following each exposure were
calculated as pre-exposure FEV-! minus post-exposure FEV-! then divided by the pre-exposure FEV^ The FEV-! decrements for
filtered air (0 ppb ozone) were subtracted from the FEVi decrements on ozone exposure days. The data for 60 and 80 ppb are the
average of a stepwise exposure day and constant exposure day. During each hour of the exposures, subjects were engaged in
moderate quasi-continuous exercise (20 L/minute per m2 body surface area) for 50 minutes and rest for 10 minutes. Following the
3rd hour, subjects had an additional 35-minute rest period for lunch.
Figure 3-2 Inter-subject variability in forced expiratory volume in 1 second
(FEVi) decrements in young healthy adults following 6.6 hours of
exposure to ozone.
Statistically significant effects on both lung function and respiratory symptoms were observed in
young healthy adults following 6.6 hours of exposure to 70 ppb (targeted concentration) ozone (Schelegle
et al.. 2009). Illustrated in Figure 3-3 are group mean FEVi responses for all 6.6-hour studies of healthy
young adults (age 18-35 years), conducted at average exposure concentrations of <120 ppb ozone with a
target exercise ventilation rate of -20 L/minute per m2 body surface area (BSA). During each hour of
exposure, subjects were engaged in moderate quasi-continuous exercise for 50 minutes and rest for
10 minutes. During chamber exposure studies, following the third hour, subjects had an additional
35-minute rest period within the chamber for lunch. During facemask exposure studies, following each
hour of exposure (50 minutes of exercise followed by 10 minutes of rest), subjects removed their
facemask (no ozone or filtered air delivery) for 2-3 minutes for measurement of pulmonary function.
Additionally, following measurement of pulmonary function after the third hour of exposure, the
facemask remained off (no ozone or filtered air delivery) for a 24-minute lunch period. Thus, for the
6.6-hour facemask studies, there was a total period of -36 minutes at rest during which there was no
delivery of ozone or filtered air exposure, whereas for the chamber studies there was 6.6 hours of
continuous ozone or filtered air exposure. Predicted FEVi responses are also illustrated by the solid line
3-13
-------
in Figure 3-3 based on the model described in McDonnell et al. (2013). Predicted FEVi responses
decrease with age, increase with BMI, and increase with ventilation rates. There are no more recent
6.6-hour ozone exposure studies.
16
14
12
10
8
6
4
2
0
30 40 50 60 70 80 90 100 110 120 130
Ozone (ppb)
All illustrated studies used a constant target quasi-continuous exercise ventilation rate of ~20 L/minute per m2 body surface area
(BSA). For studies using step-wise (s) or triangular (t) increases and decreases in ozone concentration, the FENA response is
plotted at the average ozone exposure concentration for the 6.6-hour exposure. Some exposures were conducted using a facemask
(m), all other studies were conducted within a chamber. All responses at and above 70 ppb (targeted concentration) were
statistically significant relative to filtered air exposure. At a constant exposure concentration of 60 ppb in a chamber, statistically
significant FENA responses were found by Kim et al. (20111 and in the Adams (20061 study based on the analysis of Brown et al.
(20081. With the exception of the Scheleale et al. (20091 data, the data at 60, 80, and 120 ppb have been offset along the x-axis for
illustrative purposes. The McDonnell et al. (20131 line illustrates the predicted FENA decrements at 6.6 hours as a function of ozone
concentration using Model 3 coefficients for a 23.5-year-old with a BMI of 23.1 kg/m2 having a ventilation rate during rest and
exercise of 6 and 20 L/minute per m2 BSA. 80 ppb data for 30 health subjects were collected as part of the Kim et al. (20111 study,
but only published in Figure 5 of McDonnell et al. (20121.
Adapted from Figure 6-1 of 2013 Ozone ISA (U.S. EPA. 2013a1. Studies appearing in the figure legend are: Adams (20061. Adams
(2003a1. Adams (20021. Adams (20001. Adams and Ollison (19971. Folinsbee et al. (19941. Folinsbee et al. (19881. Horstman et al.
(19901. Kim et al. (20111. McDonnell et al. (20131. McDonnell et al. (19911. and Scheleale et al. (20091.
Figure 3-3 Cross-study comparisons of mean ozone-induced forced
expiratory volume in 1 second (FEVi) decrements in young
healthy adults following 6.6 hours of exposure to ozone.
~o
a)
o
3
¦a
c
a>
c
c
a>
E
a>
o
a>
o -o
N -
o >
LU
Since the 2013 Ozone ISA, one new study has evaluated the effect of ozone on lung function at
concentrations below 80 ppb. The results of this study of older adults (55-70 years) exposed for 3 hours
3-14
X
Adams (2000)
X
Adams (2002)
A
Adams (2003)
A
Adams (2006)
-
Adams and Ollison (1997)
~
Folinsbee et al. (1988)
¦
Folinsbee et al. (1994)
o
Horstman et al. (1990)
•
Kimet al. (2011)*
~
McDonnell et al. (1991)
~
Schelegle et al. (2009)
¦McDonnell et al. (2013)
— m
— s,m
-------
to 0, 70, and 120 ppb ozone appear in both an HEI report (Frampton et al.. 2017) and in the scientific
literature (Ariomandi et al.. 2018) and are discussed in a subsection on lifestage (Section 3.1.4.1.1.1).
Several new studies have investigated the effects of 100-300 ppb ozone exposure on lung function
[e.g., Billeretal. (2011). Ghio et al. (2014). Frampton et al. (2015). Hoffmeveretal. (2013). Madden et
al. (2014). Stiegel et al. (2017). Tank etal. (2011)1. Given that lower ambient concentrations are more
common currently, any such studies are most relevant with regard to consideration of mechanistic
information or existence of associations between lung function and other indicators of respiratory health.
As discussed in Section 6.2.1.1 of the 2013 Ozone ISA, repeated consecutive days of ozone exposure
typically show that the FEVi response is enhanced on the second day of exposure. Consistent with older
studies, Madden et al. (2014) reported that 2 consecutive days of ozone exposure caused a statistically
greater decrement in FEVi (18.2 ± 4.5%) than the decrement immediately after the first day of ozone
exposure (i.e., 9.9 ± 2.5%;p < 0.05) and immediately after ozone exposure (i.e., 10.9 ± 2.6%) preceded
by an air exposure on the prior day. Although changes in lung function have generally been found to be
unrelated to inflammatory responses of the lung following ozone exposure, significant relationships have
been reported between FEVi decrements and plasma ferritin [r = -0.67,p = 0.003; i.e., larger FEVi
decrements in individuals with lower baseline plasma ferritin (Ghio et al.. 2014). and with the
inflammatory cytokine IFN-y in the blood (Stiegel et al.. 2017)1. Hoffmever et al. (2013) used 40 ppb as
their control exposure for comparisons against an exposure of 240 ppb. Relatively consistent with the
Adams (2002) and Adams (2006) studies of 6.6-hour exposures to 40 ppb with quasi-continuous exercise,
the 4-hour exposure to 40 ppb with two 20-minute periods of light exercise caused no statistically
significant changes in lung function in the study by Hoffmever et al. (2013). Study-specific details,
including exposure concentrations and durations, are summarized in Table 3-4 in Section 3.3.1.
3.1.4.1.1.1 Predicted Lung Function Response to Ozone Exposure in Healthy Adults
The similarities and differences in two models (McDonnell et al.. 2012; Schelegle et al.. 2012)
predicting FEVi responses to ozone exposure in healthy adults were described in the 2013 Ozone ISA
(U.S. EPA. 2013a). In brief, both are two compartment models that consider a dose of onset in response
or a threshold of response. The first compartment in the McDonnell et al. (2012) model considers the
level of oxidant stress in response to ozone exposure to increase over time as a function of dose rate
(concentration x minute ventilation) and decrease by clearance or metabolization over time according to
first order reaction kinetics. In the second compartment of the threshold model, once oxidant stress
reaches some threshold level, the decrement in FEVi increases as a sigmoid-shaped function of the
oxidant stress. In the Schelegle et al. (2012) model, the 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 conceptually the same as the first compartment in the McDonnell et al. (2012)
model; that is, oxidant stress increases as a function of dose rate (concentration x minute ventilation) and
oxidant stress decreases according to first order clearance kinetics. The oxidant levels in the second
compartment of the Schelegle et al. (2012) model are multiplied by a responsiveness coefficient to predict
3-15
-------
FEVi responses. Two new models (Hsich et al.. 2014; McDonnell et al.. 2013) have become available
since the 2013 Ozone ISA.
The McDonnell et al. (2012) and McDonnell et al. (2013) models were fit to a large data-set
consisting of the FEVi responses of 741 young healthy adults (104 F, 637 M; mean age 23.8 years) from
23 individual controlled exposure studies conducted in either Chapel Hill, NC or Davis, CA. The models
were fit using a SAS procedure specially designed for fitting nonlinear random-effects models. Statistical
estimates were obtained for the primary model parameter coefficients, a variance term for inter-subject
variability in response, and an error term representing intra-subject variation. McDonnell et al. (2013)
provides alternative variance structures relative to the McDonnell et al. (2012) model. McDonnell et al.
(2013) partitioned the intra-subject error term to include (1) random noise in measurement of FEVi and
(2) increasing variability with increasing FEVi response. The addition of random intra-subject noise in
the error term allows lower percentiles of the FEVi response distribution to have improvements in FEVi
during ozone exposure.
Hsieh et al. (2014) emulated the mechanistic model developed by Freiier et al. (2002). The Freiier
et al. (2002) model predicts changes in FEVi to occur as a function of the balance between respiratory
cells naive to ozone exposure and those previously exposed cells having developed some antioxidant
protection. An interesting aspect of this model is that it is capable of predicting the effects of consecutive
days of ozone exposure on FEVi responses. As discussed in Section 6.2.1.1 of the 2013 Ozone ISA,
repeated consecutive days of ozone exposure typically show that FEVi responses are enhanced on the
second day of exposure and become attenuated after 3 to 4 consecutive days of exposure relative to the
first ozone exposure day. Hsieh et al. (2014) fitted three parameters of the Freiier etal. (2002) model to
best match model-estimated FEVi decrements with the Schelegle et al. (2009) data. Overall, across all
exposure concentrations (targets of 60, 70, 80, and 87 ppb) and time points (0, 1, 2, 3, 4.6, 5.6, and
6.6 hours), there was an r2 of 0.73 between the predicted and observed FEVi responses. The 70-ppb target
exposure in the Schelegle et al. (2009) study is the lowest concentration at which both statistically
significant FEVi decrements and respiratory symptoms have been observed following 6.6 hours of
exposure with quasi-continuous exercise. Figure 3 of Hsieh et al. (2014) shows that had the Schelegle et
al. (2009) study been extended to 8 hours, a 6.14% FEVi decrement would be observed at 63 ppb after
8 hours of exposure, the same decrement observed following 6.6 hours of exposure to 70 ppb (targeted
concentration) by Schelegle et al. (2009).
3.1.4.1.1.2 Factors Affecting Lung Function Response to Ozone
Airway Responsiveness
Although ozone exposure has been shown to increase airway responsiveness, fewer studies have
assessed whether baseline airway responsiveness is associated with ozone-induced changes in lung
3-16
-------
function. In the 2006 ozone AQCD (U.S. EPA. 2006). there was limited discussion of Aris et al. (1995)
who exposed healthy adults (24 F, 42 M; 27 ± 4.5 years) to 0 and 200 ppb ozone for 4 hours during
quasi-continuous exercise (50 minutes at 25 L/minute per m2 BSA and 10 minutes rest). These authors
observed a weak correlation between pre-exposure methacholine responsiveness and ozone-induced
increases in sRaw, but not with ozone-induced decreases in FEVi and FVC. Recent studies expand upon
the previous evidence base, but provide no new evidence per se. Specifically:
• Hoffmeveretal. (2013) exposed healthy adults (7 F, 8 M; 26 years) to 40 ppb (control exposure)
and 240 ppb ozone for 4 hours with two 20-minute exercise (15 L/minute per m2 BSA) periods.
Five subjects having >5% decrement in FEVi following the 240-ppb exposure were characterized
as responders. There was a tendency towards a greater FEVi response to methacholine in the
5 responders as compared to the 10 nonresponders. Responsiveness to capsaicin as a predictor of
ozone responses was also examined. Across all subjects, capsaicin responsiveness was correlated
with ozone-induced changes in peak expiratory flow (r = 0.716,/? = 0.003) and maximal
expiratory flow at 50% of vital capacity (r = 0.589,/? = 0.021), but less so with FEVi (r = 0.417,
p = 0.122). The cumulative dose of capsaicin causing two or more coughs was also significantly
lower in the ozone responders than nonresponders. The association between ozone and capsaicin
responsiveness likely reflected the role of sensory C-fibers.
• Bennett et al. (2016) found statistically greater FVC decrements in obese (19 F; 27.7 ±5.2 years)
than normal weight (19F;24±3.7 years) individuals following a 2-hour exposure to 400 ppb
ozone. This difference was not associated with methacholine responsiveness on the training day,
which was similar between the groups.
• In a large study of individuals with asthma (34 F, 86 M; 32.9 ± 12.9 years), Bartoli et al. (2013)
also found the magnitude of ozone-induced FEVi response (based on 2-hour exposures to
300 ppb and filtered air) was unrelated to baseline methacholine responsiveness.
Ambient Temperature
Studies reviewed in Section 10.2.9.3 of the 1986 ozone AQCD (U.S. EPA. 1986) suggested an
additive effect of increased temperature with ozone exposure on lung function decrements. However, the
effect of temperature and humidity on respiratory responses was termed as equivocal in Section 7.2.1.3 of
the 1996 ozone AQCD (U.S. EPA. 1996a). In Section 6.5.5 of the 2006 ozone AQCD (U.S. EPA. 2006).
a single new study (Foster etal.. 2000) was discussed that suggested elevated temperature may partially
attenuate spirometric responses but enhance airway reactivity. Discussion of the effect of temperature on
responses in controlled human exposure studies was not included in the 2013 Ozone ISA (U.S. EPA.
2013a). Overall, recent studies are consistent with the equivocal findings related to effects of temperature
in prior reviews. Specifically:
• Recently, Kahle et al. (2015) exposed healthy volunteers (2 F, 14 M) to filtered air and 300 ppb
ozone for 2 hours with intermittent exercise (alternating 15-minute periods rest and exercise at
25 L/minute per m2 BSA) at both 22 and 32.5°C. FEVi and FVC were significantly reduced by
exposure to 300 ppb ozone relative to filtered air, but no significant effect of temperature or
ozone-temperature interaction was observed. There was a tendency for smaller ozone-induced
FEVi and FVC decrements at 32.5°C than at 22°C.
3-17
-------
• In another study, Gomes et al. (201 lb) exposed 10 male athletes to filtered air and 100 ppb ozone
while completing an 8-km time trial at either 20°C + 50% RH and 31°C + 70% HR. The elevated
temperature and humidity with and without ozone significantly decreased running speed. The
combination of heat and ozone also significantly increased the athletes perceived exertion level
relative to the lower temperature scenario with and without ozone. This study supports a trend for
an additive effect of ozone and temperature on decreased exercise performance and perceived
exertion level.
• In another high temperature (31°C + 70% HR) ozone study, Gomes et al. (201 la) showed a
tendency toward improved exercise performance in nine male athletes during exposure to
100 ppb ozone between vitamin and placebo trials (p = 0.075). This is generally consistent with
studies reviewed in the 2006 Ozone AQCD (AX6.5.6 Oxidant-Antioxidant Balance).
Cigarette Smoking
Studies reviewed in the 2006 ozone AQCD (U.S. EPA. 2006) and earlier reviews showed active
smokers experienced smaller lung function decrements than nonsmokers in response to ozone exposure.
A recent study found similar FEVi decrements between smokers and nonsmokers:
• In a recent study, Bates et al. (2014) exposed smokers (11 F, 19 M; 24 ± 4 years) and nonsmokers
(13 F, 17 M; 25 ± 6 years) to 0 and 300 ppb ozone for 1 hour during light exercise. Statistically
significant ozone-induced FEVi decrements (about 9-10%) were similar between smokers and
nonsmokers. The lack of diminished FEVi decrements in smokers relative to nonsmokers may
have been due to the relatively light smoking history of the smokers. For comparison, Frampton
et al. (1997) observed statistically significantly smaller FEVi decrements in smokers than
nonsmokers, but the smokers had an average smoking history of 13 pack-years, which is three
times greater than the mean smoking history of the smokers in the Bates et al. (2014) study.
Based on exhaled CO2 profiles, smokers, but not nonsmokers, showed a reduction in dead space
(-6.1 ± 1.2%) and an increase in the alveolar slope (9.1 ± 3.4%). This finding could be caused by
nonuniform bronchoconstriction, which would alter the pattern of filling and emptying lung units.
An effect on pulmonary ventilation was also reported in the 2006 ozone AQCD; a study by Foster
et al. (1997) showed a 24% reduction in the washout rate of the lungs of healthy males following
ozone exposure, which remained or developed in 50% of subjects a day after the ozone exposure.
Lifestage
Healthy older subjects (52 F, 35 M; 59.9 ± 4.5 years) were exposed to 0, 70, and 120 ppb ozone
for 3 hours during intermittent exercise [15-minute intervals of rest and exercise at 15-17 L/minute per
m2 BSA; Ariomandi et al. (2018)1. Lung function (FEVi, FVC, FEVi/FVC, and FEF25 75) was measured
10 minutes before exposure and at 0.25 and 22 hours post-exposure. As has been reported in prior studies
[see pg. 6-4 of U.S. EPA (2013a)l. FEVi increased after exercise during exposure to filtered air at both
15 minutes (85 mL; 95% CI: 64-106; paired Z-test: p < 0.001) and 22 hours (45 mL; 95% CI, 26-64;
p < 0.001) after exposure. The ozone exposures resulted in a smaller exercise-related increase in FEVi,
specifically 15- and 33-mL smaller increase at 70 ppb (p = 0.12) and 120 ppb (p = 0.001), respectively.
The observed FEVi and FVC changes following ozone exposure showed no interaction by sex (52 F,
3-18
-------
35 M), age (55-70 years), or GSTM1 genotype (57% null, 43% positive). Inflammatory responses
measured as part of this study are provided in another section of this document.
While the decrements in lung function observed by Ariomandi et al. (20 IS) are small, with a
group mean ozone-induced FEVi decrement of only 1.2% (based on group mean changes in lung function
provided in Table 2 of the paper) following the 120 ppb exposure, the decrement was not expected in
these older subjects at a relatively light activity level (intermittent 15-minute periods of rest and exercise
at 15-17 L/min-m2) and brief 3-hour duration of exposure as discussed below. For comparison, the
McDonnell et al. (2013) model predicts a 2% FEVi decrement in 23.8-year-olds (less than the 3% FEVi
decrement observed and predicted in 6.6 hours studies at 60 ppb) for this exposure protocol and no FEVi
decrement is predicted in individuals over 48.5 years of age. Results from Ariomandi et al. (2018) are
generally consistent with the prior work of Hazucha et al. (2003). who studied adults up to 60 years of age
exposed to 400 ppb for 1.5 hours with intermittent exercise (17.5 L/min-m2) and showed that lung
function decrements decline across the studied age range, but may still be present in older adults
50-60 years of age. However, this recent study was conducted at a lower ozone delivery rate than
Hazucha et al. (2003). which is more representative of that likely to occur in ambient air and shows small
lung function decrements may occur in older adults.
3.1.4.1.2 Animal Toxicological Studies
The 2013 Ozone ISA summarized the animal toxicological evidence of changes in lung function
resulting from exposure to ozone. Most of the studies involved acute exposures (U.S. EPA. 2013a).
Changes in frequency of breathing and tidal volume, reflecting a pattern of rapid, shallow breathing, were
commonly observed at ozone concentrations of about 0.2 ppm. Decreased lung volumes were observed in
rats exposed to 0.5 ppm, while changes in compliance and resistance were observed at ozone
concentrations of 1 ppm and above. Repeated acute exposures over several days led to attenuation of the
pulmonary function decrement response. A lung imaging study found that continuous or half-day
exposure to 0.5 ppm ozone for several days led to ventilatory abnormalities that suggested narrowing of
peripheral small airways and increased airway resistance. While ozone concentrations in animal
toxicological studies seem to be high, it should be noted that deposition of ozone resulting from a 2-hour
exposure to 2 ppm ozone in a resting rat is roughly equivalent to deposition of ozone resulting from a
2-hour exposure to 0.4 ppm ozone in an exercising human (Hatch et al.. 1994). More recently, Hatch et al.
(2013) showed that resting rats and resting humans receive similar alveolar ozone doses.
Studies described in the 2013 Ozone ISA provide evidence that neurogenic mechanisms underlie
the changes in lung function observed in animals exposed to ozone (U.S. EPA. 2013a). Activation of
sensory nerves in the airway epithelium occurs as a result of ozone exposure. Stimulation of bronchial
C-fibers leads to rapid, shallow breathing and other changes in respiratory mechanics in response to
ozone. TRPA1 ion channels, which are found on a subset of bronchial C-fibers, may be activated by
secondary oxidation products of ozone and components of the extracellular lining fluid in the respiratory
3-19
-------
tract, such as aldehydes. In addition, arachidonic acid metabolites, such as prostaglandins, may be
involved in activation or sensitization of the TRPA1 ion channels. As discussed previously, these airway
sensory nerves are vagal afferents that carry signals to the nucleus tractus solitarius neurons in the brain.
These pathways can be integrated in the brain, resulting in altered autonomic activity that affects the
airways (e.g., bronchoconstriction) or extrapulmonary responses such as changes in heart rhythm.
Stress-responsive regions of the brain may also be affected by these vagal afferent pathways from the
respiratory tract. In addition, activation of bronchial C-fibers may lead to local axon reflex responses in
the airways, such as the release of substance P or other tachykinins, which act through neurokinin
receptors to increase airway resistance (i.e., bronchoconstriction).
A large number of recent studies evaluated changes in lung function in response to short-term
ozone exposure. Study-specific details are summarized in Table 3-5 in Section 3.3.1. All of these studies
were conducted in rodent strains with varying degrees of sensitivity to ozone. Lung function was assessed
by changes in ventilatory parameters such as tidal volume and enhanced pause. Enhanced pause is a
measure of respiratory distress that may or may not be related to an increase in airway resistance. Airway
resistance can be examined by direct measures of lung mechanics in vivo such as the flexiVent and the
pneumotachometer/pressure transducer system, which are both invasive methods. These recent studies,
which are detailed below and grouped by concentration-time profile, demonstrated that exposure to
0.1-2 ppm ozone results in changes in lung function, as measured by altered ventilatory parameters. All
of the changes in lung function described below were statistically significant. Changes in enhanced pause
and evidence of sensory and pulmonary irritation were observed following acute exposure to 2 ppm
ozone. Changes in enhanced pause and tidal volume were observed with acute exposure to 0.5-1 ppm
ozone. Repeated exposure to ozone resulted in numerous effects, with decreased respiratory frequency
occurring at concentrations of 0.1 ppm ozone.
• Acute exposure of rodents to 2 ppm ozone for 3 hours resulted in increases in enhanced pause
(Ghio et al.. 2014; Bao et al.. 2013; Lee et al.. 2013). Evidence for sensory and pulmonary
irritation is provided by Hansen et al. (2016). Sensory irritation reflects changes in the upper
airways, while pulmonary irritation reflects changes in the lower airways.
• Acute exposure to 0.8-1 ppm ozone for 1-6 hours resulted in alterations in tidal volume and
enhanced pause (Gordon et al.. 2016b; Dve et al.. 2015; Schclcglc and Walbv. 2012). Alterations
were also found following exposure to 0.5 ppm ozone (Dve et al.. 2015).
• Repeated ozone exposures with differing concentration-duration profiles also resulted in altered
ventilatory parameters.
o Decreased respiratory frequency—0.1 ppm ozone for 1 hour/day for 10 days (Wolkoff et
al.. 2012).
o Increased minute volume and enhanced pause, and decreased relaxation time—1 ppm
ozone for 6 hours/day for 2 days (Snow et al.. 2016V
o Increased enhanced pause—1 ppm ozone for 4 hours/day for 1 and 2 days (Miller etal..
2016b) or 1 ppm ozone for 5 hours/day for 2 days (Gordon et al.. 2017b') or 0.8 ppm
ozone for 4 hours/day for 1 and 2 days (Gordon et al.. 2017a).
3-20
-------
o Increased peak expiratory flow and enhanced pause—0.8 ppm ozone for 4 hours/day for
1 and 2 days (Hcnriqucz et al.. 2017).
o No evidence of altered ventilatory parameters was seen in response to 0.25-0.5 ppm
ozone for 5-6 hours per day for 2 days (Gordon et al.. 2017b; Snow et al.. 2016).
• Two studies examined changes in lung function following acute ozone exposure in rodents of
varying life stages.
o In a study of rodents from adolescence to senescence (Snow et al.. 2016). ozone exposure
resulted in age-dependent changes in minute volume. Increases in minute volume were
observed in 1-month-old animals but not in 4-, 12-, and 24-month-old animals exposed to
1 ppm ozone for 6 hours per day for 2 days.
o In Groves et al. (2013). ozone exposure increased resistance in young adult mice and
increased resistance and elastance in older adult mice.
3.1.4.1.3 Epidemiologic Studies
A number of studies evaluated in the 2013 Ozone ISA provided consistent evidence for
ozone-related decreases in lung function in healthy children (U.S. EPA. 2013a). Noteworthy evidence of
the effect of short-term exposure to ozone on respiratory effects in healthy children came from panel
studies with daily assessment of lung function in children attending summer camps (Berry et al.. 1991;
Spektor and Lippmann. 1991; Avol et al.. 1990; Burnett et al.. 1990; Higgins et al.. 1990; Raizenne et al..
1989; Spektor et al.. 1988). Specifically, ozone exposure was consistently associated with decreases in
FEVi in 7- to 17-year-old children without asthma. Analyses were conducted during summer months in
the 1980s and included diverse locations across the U.S. and Canada. Additionally, ozone monitoring
generally occurred at the site of the camp, reducing potential exposure measurement error. While
associations for peak expiratory flow (PEF) were more variable than those for FEVi, increases in ambient
ozone concentration were generally associated with decreases in PEF. None of the referenced studies
examined copollutant models.
In addition to studies of children, the 2013 Ozone ISA evaluated a number of studies that
examined lung function in healthy adults. There was consistent evidence of ozone-related lung function
decrements in panel studies of adults participating in outdoor recreation, exercise, or work [see
Section 6.2.1.2 of the 2013 Ozone ISA (U.S. EPA. 2013a)l. Like the summer camp studies, these studies
had on-site ozone measurements during the time of the outdoor activity, resulting in higher personal
exposure and ambient concentration correlations. Cohort and cross-sectional studies that used the average
of fixed-site monitors, nearest monitor, or spatial interpolation to assign exposure across a larger study
area observed inconsistent evidence of an association between short-term ambient ozone concentrations
and lung function in adults and older adults. The inconsistent results relative to panel studies may have
been due to differences in study design, geographic location, and/or increased exposure measurement
error, among other factors.
3-21
-------
A recent study in Canada also reports a positive association between short-term ambient ozone
concentrations and lung function effects in a healthy population. Study-specific details, including air
quality characteristics and select effect estimates, are highlighted in Table 3-6 in Section 3.3.1. An
overview of the evidence is provided below.
• In a randomized crossover study of young adults in Sault Ste. Marie, Canada, Dales et al. (2013)
observed decreases in a range of lung function metrics, including FEVi, FVC, and FEVi/FVC,
associated with ozone concentrations. Participants alternated five consecutive 10-hour days
outdoors at two locations and ozone concentrations were measured on-site to reduce potential
exposure measurement error. SO2 concentrations were notably higher at one location near a steel
plant, but the lung function associations with ozone were independent of study site.
• Other recent studies of adults examined respiratory effects in the general population (Lcpculc et
al.. 2014; Rice et al.. 2013). These studies included both healthy participants and those with
pre-existing respiratory conditions, with asthma or COPD prevalence ranging from 6 to 20%.1
Because these studies do not directly inform the understanding of the relationship between
short-term ozone exposure and lung function in healthy populations or populations with asthma,
they are not discussed in either section. However, study specific details can be found in Table 3-7
in Section 3.3.1.
In summary, one recent study, along with studies evaluated in the previous ozone ISA, support
the presence of an association between short-term ozone exposure and decreased lung function in healthy
populations. Onsite exposure measurement at study site locations has reduced the potential for exposure
measurement error in these studies, but the independence of the observed associations relative to other
pollutants remains uncertain.
3.1.4.1.4 Integrated Summary for Lung Function
Controlled human exposure studies evaluated in the 1996 and 2006 Ozone AQCDs (U.S. EPA.
2006. 1996a) provided evidence for a number of lung function effects in healthy subjects exposed to
>80 ppb ozone. Young adults and children experience similar transient decrements in pulmonary function
when exposed to ozone, but spirometric responses become less pronounced with increasing age. Further
evidence from the 2013 Ozone ISA (U.S. EPA. 2013a) demonstrated decreases in group mean pulmonary
function following 6.6-hour ozone exposures as low as 60 to 70 ppb in young, healthy adults performing
moderate quasi-continuous exercise. One recent study observed a small, but not statistically significant,
group decrease in lung function in older adults following 3-hour exposures to 70 ppb with light to
moderate intermittent exercise.
Like studies in humans, experimental studies in animals also provide evidence of changes in lung
function resulting from exposure to ozone. Evidence summarized in the 2013 Ozone ISA indicated
1 All epidemiologic results standardized to a 15-ppb increase in 24-hour avg, 20-ppb increase in 8-hour daily max,
25-ppb increase in 1-hour daily max ozone concentrations, or a 10-ppb increase in seasonal/annual ozone
concentrations to facilitate comparability across studies.
3-22
-------
changes in the frequency of breathing and tidal volume, decreased lung volume, increased airway
resistance, and attenuation of the pulmonary function decrement response following repeated exposures
(U.S. EPA. 2013a). Recent evidence further demonstrates changes in ventilatory parameters resulting
from ozone exposure. Experimental studies in both humans and animals indicate that changes in lung
function, including FEVi and sRaw, may be attributed to activation of sensory nerves in the respiratory
tract that trigger local and autonomic reflex responses. Specifically, mechanistic studies provide evidence
that local reflex responses mediate the observed decreases in inspiratory capacity and pain on inspiration
that result in truncated inspiration. In addition, modest increases in airway resistance may occur due to
activation of parasympathetic pathways. These changes, along with observed alterations in breathing
frequency, are a type of irritant response. Results from recent animal toxicological studies are generally
consistent with those described in the 2013 Ozone ISA, reporting changes in ventilatory parameters
resulting from acute exposure to 0.5-2 ppm ozone and from repeated exposure to 0.1 ppm ozone.
Epidemiologic studies of lung function in healthy populations are coherent with experimental
studies that demonstrate evidence of ozone-related lung function decrements in a controlled environment
and provide mechanistic evidence for the plausibility of the observed changes. Most epidemiologic
evidence comes from panel studies of healthy children that were previously evaluated in the 2013 Ozone
ISA (U.S. EPA. 2013a).
3.1.4.2 Respiratory Symptoms
3.1.4.2.1 Controlled Human Exposure Studies
As described in Section 6.2.1.1 of the 2013 Ozone ISA (U.S. EPA. 2013a). in addition to lung
function decrements, controlled human exposure studies clearly indicate ozone-induced increases in
respiratory symptoms including pain on deep inspiration, shortness of breath, and cough. In brief, the
available evidence during the last review indicated that respiratory symptoms increase with increasing
ozone concentration, duration of ozone exposure, and activity level of exposed subjects. For exposures of
1-2 hours to >120 ppb, statistically significant respiratory symptoms and effects on FEVi were observed
when exercise sufficiently increased ventilation rates (McDonnell et al.. 1999b). During exposures at rest,
5% of young healthy adults exposed to 400 ppb ozone for 2 hours experienced pain on deep inspiration,
but not at 1 hour of exposure. Respiratory symptoms were also not observed following 1 to 2 hours of
resting exposure at lower concentrations of 120 to 300 ppb. However, when exposed during light to
moderate intermittent exercise (22-35 L/minute) to 120 ppb for 2 hours, 9% of individuals experienced
pain on deep inspiration, 5% experienced cough, and 4% experienced shortness of breath. For longer
duration, 6.6-hour exposures to 80 ppb with moderate quasi-continuous exercise, FEVi decrements and
total respiratory symptoms diverge from filtered air responses after 3 hours and become statistically
different by 6.6 hours (Adams. 2006). For the 6.6-hour exposures to ozone, 70 ppb is the lowest
3-23
-------
concentration where statistically significant ozone-induced lung function decrements and subjective
symptoms have been reported (Schclcglc et al.. 2009). Although several studies have investigated the
effects of 6.6-hour exposures during moderate exercise to 60 ppb ozone, none have observed a
statistically significant increase in respiratory symptoms following ozone relative to filtered air. There are
no new controlled human exposure studies conflicting with the above or contributing a better
characterization of ozone-induced respiratory symptoms.
3.1.4.2.2 Animal Toxicological Studies
There were no animal toxicological studies that examined respiratory symptoms in the 2013
Ozone ISA (U.S. EPA. 2013a). In fact, symptoms cannot be assessed in studies of rodents because
symptoms are, by definition, self-reported. However, cough can be elicited in rodents through the use of
an irritant. This reflex is termed a hypertussive response. A recent study in guinea pigs and rabbits found
that ozone acts through sensory nerves to enhance coughing that is elicited by citric acid (Clay et al..
2016). Details of this study are summarized in Table 3-5 in Section 3.3.1. Acute exposure to 2 ppm ozone
for 0.5-1 hour resulted in statistically significant increases in cough frequency and decreases in time to
cough in response to citric acid. Experiments with pharmacological agents implicated TRPV1 receptors, a
type of sensory nerve receptor often found on C-fibers, in mediating the hypertussive response to ozone.
3.1.4.2.3 Epidemiologic Studies
The 2013 Ozone ISA did not evaluate any studies that examined respiratory symptoms in study
populations consisting solely of healthy populations [i.e., respiratory disease-free; U.S. EPA (2013a) I.
Several panel studies of children that were not restricted to healthy individuals, in which asthma
prevalence was 50% or less, reported null or negative associations between ambient ozone concentrations
and respiratory symptoms, such as cough, wheeze, and phlegm (see Section 3.1.4 of the 2013 Ozone
ISA). Notably, the majority of these studies assessed respiratory symptoms through parental reported
outcomes, which may be differentially misreported based on asthma status.
3.1.4.2.4 Integrated Summary for Respiratory Symptoms
Controlled human exposure studies evaluated in the 2013 Ozone ISA reported symptoms of
cough and pain on deep inspiration corresponding to FEVi decrements in healthy young adults exposed to
70 ppb ozone for 6.6 hours with moderate quasi-continuous exercise (U.S. EPA. 2013a). A recent model
can be used to determine the ozone concentration that would lead to the same FEVi decrement following
an 8-hour exposure (McDonnell et al.. 2013). Under the assumption that respiratory symptoms might
accompany similar ozone-induced FEVi decrements, regardless of exposure duration, the model indicates
3-24
-------
that an 8-hour exposure to 64 ppb ozone concentration might reasonably be expected to cause an adverse
response in young healthy adults.
In coherence with evidence observed in controlled human exposure studies, a recent mechanistic
animal toxicological study observed that ozone acts through sensory nerves to induce coughing. In
contrast, ozone-induced respiratory symptoms observed in healthy subjects in controlled human exposure
and animal toxicological studies were not evident in epidemiologic studies in the general population.
However, these epidemiologic studies generally relied on parental reported outcomes that may result in
under- or over-reporting of respiratory symptoms, depending on a child's asthma status.
3.1.4.3 Airway Responsiveness
3.1.4.3.1 Controlled Human Exposure Studies
As reviewed in the 2006 Ozone AQCD (U.S. EPA. 2006) and in the 2013 Ozone ISA (U.S. EPA.
2013a). ozone has been shown to cause an increase in airway responsiveness in controlled human
exposure studies. In general, airway responsiveness is assessed by increasing inhaled concentrations of a
bronchoconstrictive drug and measuring the effect on lung mechanics (FEVi or sRaw). A dose-dependent
increase in airway responsiveness of young, healthy, nonsmoking males following exposures to 0, 80,
100, and 120 ppb ozone (6.6 hours, quasi-continuous moderate exercise at 39 L/minute) has been
demonstrated. Changes in airway responsiveness appear to persist longer than changes in pulmonary
function, although this has been studied only on a limited basis. Studies suggest that ozone-induced
increases in airway responsiveness usually resolve 18 to 24 hours after exposure but may persist in some
individuals for longer periods. Although FEVi decrements and respiratory symptoms become attenuated
following several consecutive days of ozone exposure, the ozone-induced increase in airway
responsiveness (measured by increase in sRaw upon methacholine challenge) over 5 consecutive days is
not attenuated. Increases in airway responsiveness following ozone exposure do not appear to be
associated with ozone-induced changes in lung function, respiratory symptoms, or changes in epithelial
permeability. First described in the 1986 ozone AQCD (U.S. EPA. 1986). a mechanistic study of subjects
exposed to 600 ppb ozone while exercising added to the understanding of mechanisms underlying
changes in airway responsiveness caused by ozone exposure. Atropine inhibited an ozone-induced
increase in airway responsiveness to histamine, indicating the involvement of the parasympathetic
nervous system in this response. A recent study of 38 healthy adult women (average age, 26 years)
exposed to 0 and 400 ppb ozone for 2 hours performing light intermittent exercise (15-minute periods of
exercise at 25 L/minute and seated rest) showed a tendency (statistical significance not assessed by
investigators) for increases in airway responsiveness due to ozone with 4 and 12 subjects being
responsive to methacholine after exposure to filtered air and ozone, respectively (Bennett et al.. 2016).
Study-specific details are summarized in Table 3-8 in Section 3.3.1.
3-25
-------
3.1.4.3.2
Animal Toxicological Studies
The 2013 Ozone ISA (U.S. EPA. 2013a) summarized the animal toxicological evidence of
increased airway responsiveness resulting from exposure to ozone. In general, airway responsiveness is
assessed by measuring the effects of challenge with increasing concentrations of a bronchoconstrictive
drug on lung mechanics (FEVi or sRaw). Methacholine is the most commonly used nonspecific challenge
agent, but histamine and other agents are also used. Most of the studies discussed in the 2013 Ozone ISA
(U.S. EPA. 2013a') found increased airway responsiveness in guinea pigs, rats, or mice exposed to 1 ppm
and higher concentrations of ozone, although increased airway responsiveness was, in a few cases,
demonstrated after exposure to less than 0.3 ppm ozone. While ozone concentrations in animal
toxicological studies seem to be high, it should be noted that deposition of ozone resulting from a 2-hour
exposure to 2 ppm ozone in a resting rat is roughly equivalent to deposition of ozone resulting from a
2-hour exposure to 0.4 ppm ozone in an exercising human (Hatch etal.. 1994). More recently, Hatch et al.
(2013) showed that resting rats and resting humans receive similar alveolar ozone doses. Studies
involving animal models of allergic airway disease are discussed in Section 3.1.5.5.2 because these
animal models share phenotypic features with asthma.
Studies described in the 2013 Ozone ISA (U.S. EPA. 2013a) provide evidence that neurogenic
mechanisms underlie the increased airway responsiveness observed in animals exposed to ozone (U.S.
EPA. 2013a). In one study, eosinophils promoted the activation of airway parasympathetic nerves by
releasing major basic protein, which blocked a muscarinic receptor-mediated pathway that attenuates
acetylcholine release from the nerves. Acetylcholine, like methacholine, acts on receptors in airway
smooth muscle to stimulate bronchoconstriction. In another study, substance P acted through the
neurokinin 1 receptor to cause vagally mediated bronchoconstriction. There is also evidence that
arachidonic acid metabolites and cytokines/chemokines such as tumor necrosis factor alpha (TNF-a),
C-S-C chemokine receptor type 2 (CXCR2), and macrophage inflammatory protein-1 (MIP-2) play a role
in increased airway responsiveness following exposure to ozone. Furthermore, activation of an innate
immune pathway involving natural killer cells, interleukin-17 (IL-17), and airway neutrophils was
reported to lead to the development of increased nonspecific airway responsiveness following repeated
exposure to ozone.
A large number of recent studies evaluated changes in airway responsiveness following acute and
repeated ozone exposure in rodent strains with varying degrees of sensitivity to ozone. Study-specific
details are summarized in Table 3-5 in Section 3.3.1. A subset of studies investigated the role of specific
cell signaling pathways in mediating responses by using genetic knockout models or pharmacologic
agents. Airway responsiveness to a challenge agent was often assessed using the flexiVent system to
assess respiratory system mechanics. Another invasive method to assess airway resistance—pulmonary
inflation pressure following electrical stimulation of the vagal nerve was used in a study by Verhein et al.
(2013). Taken together, these recent studies, which are detailed below and grouped by concentration-time
profile, demonstrate increases in airway responsiveness following exposure to 0.8-2 ppm ozone. All of
3-26
-------
the changes in airway responsiveness described below were statistically significant. Ozone exposure
enhanced the sensitivity of the airway to vagal nerve stimulation by decreasing muscarinic type 2 receptor
function in one study (Vcrhcin et al.. 2013) and increased BALF levels of the tachykinin substance P in
another study (Barker et al.. 2015). Enhanced vagal nerve sensitivity and substance P release due to
activation of a local axon reflex in the airways may explain the ability of ozone to act as a nonallergic
asthma trigger resulting in bronchoconstriction.
• Acute exposure to 2 ppm ozone for 3 hours resulted in increased airway responsiveness to
methacholine or acetylcholine (Cho et al.. 2018; Mathews et al.. 2018; Malik et al.. 2017; Stober
et al.. 2017; Elkhidir et al.. 2016; Kasahara et al.. 2015; Razvi et al.. 2015; Barreno et al.. 2013;
Cho et al.. 2013; Sunil et al.. 2013); however, no increases were observed in Mathews et al.
(2017b) or Cho et al. (2013).
o This response was persistent over time in Sunil et al. (2013).
o Several studies provide evidence for cell signaling and other pathways underlying
increases in airway responsiveness resulting from acute ozone exposure.
¦ TNF-stimulated gene 6 and hyaluronan-heavy chain complexes (Stober et al..
2017)
¦ Rho-associated coiled-coil-containing protein kinase [ROCK; Kasahara et al.
(2015)1
¦ Dietary short chain fatty acids/gut microbiome (Cho et al.. 2018)
¦ IL-17 (Mathews et al.. 2018)
¦ Osteopontin (Barreno et al.. 2013)
o Evidence for ozone exposure-induced release of tachykinins is provided by Barker et al.
(2015). Acute ozone exposure (2 ppm for 3 hours) increased levels of the tachykinin
substance P levels in the BALF through upstream effects on IL-1J3 and nerve growth
factor.
• Exposure to 2 ppm for 4 hours increased airway responsiveness measured as increased
bronchoconstriction in response to electrical stimulation of the vagal nerve (Verhein et al.. 2013).
Both decreased function of M2 muscarinic receptors and involvement of the p38/JNK pathway
were implicated in this response.
• Acute exposure to 0.8-1 ppm ozone for 1-6 hours resulted in increased airway responsiveness
(Zvchowski et al.. 2016; Groves et al.. 2012). Rho kinase was implicated in this response
(Zvchowski et al.. 2016).
• Repeated exposure to 1 ppm ozone (3 hours/day for 7 days) resulted in increased airway
responsiveness (Zhu et al.. 2016). This response was blocked by treatment with Vitamin E,
implicating oxidative stress in mediating ozone-induced increased airway responsiveness.
3.1.4.3.3 Integrated Summary for Airway Responsiveness
Controlled human exposure studies and animal toxicological studies evaluated in the 2006 Ozone
AQCD (U.S. EPA. 2006) and the 2013 Ozone ISA (U.S. EPA. 2013a) provide consistent evidence of
ozone-induced increases in airway responsiveness. In experimental studies in humans, changes in airway
3-27
-------
responsiveness were less transient than the observed ozone-related lung function changes discussed in
Section 3.1.4.1.1. One recent study of healthy adult women showed a tendency (statistical significance
not reported) for increased airway responsiveness following 400 ppb ozone exposure for 2 hours with
intermittent exercise. In recent experimental animal studies, increases in airway responsiveness resulted
from ozone exposures in the range of 0.8 to 2 ppm, but not in response to acute and repeated exposures of
0.25 and 0.5 ppm. Mechanistic studies provide evidence that local reflex responses and activation of
parasympathetic pathways mediate increases in airway responsiveness due to ozone exposure. This may
explain the ability of ozone to act as a nonallergic asthma trigger resulting in bronchoconstriction.
3.1.4.4 Respiratory Tract Inflammation, Injury, and Oxidative Stress
3.1.4.4.1 Controlled Human Exposure Studies
As reported in studies reviewed in the 1996 and 2006 ozone AQCDs (U.S. EPA. 2006. 1996a).
acute ozone exposure initiates an acute inflammatory response throughout the respiratory tract that has
been observed to persist for at least 18-24 hours post-exposure. A single acute exposure (1-4 hours) of
humans to moderate concentrations of ozone (200-600 ppb) while exercising at moderate to heavy
intensities results in a number of cellular and biochemical changes in the lung, including an inflammatory
response characterized by increased numbers of PMNs, increased permeability of the epithelial lining of
the respiratory tract, cell damage, and production of proinflammatory cytokines and prostaglandins. These
changes also occur in humans exposed to 80 and 200 ppb ozone for 6-8 hours.
The presence of PMNs in the lung has long been accepted as a hallmark of inflammation and is
an important indicator that ozone causes pulmonary inflammation. Studies reviewed in the 2006 ozone
AQCD showed that inflammatory responses to ozone did not appear to be correlated with changes in lung
function in either healthy subjects or those with asthma, but there was some indication of a correlation
with changes in sRaw (HEI. 1997: Balmes et al.. 1996). The number of PMN and shed epithelial cells (a
marker of injury) in the BALF also correlated with loss of substance P immunoreactivity from neurons in
the bronchial mucosa in humans following exposure to 200 ppb ozone. Taken together, these findings
suggest disparate mechanisms underlying changes in lung volume and inflammation, and a commonality
in the mechanisms underlying airway obstruction and inflammation, which possibly involves neurogenic
edema.
By the completion of the 2006 ozone AQCD, studies had shown that within-individual
inflammatory responses to ozone were reproducible and correlated between repeated exposures. Thus,
just as was observed for changes in lung function in response to ozone exposure, some individuals are
intrinsically predisposed to having increased PMN responses relative to others. In the 2013 Ozone ISA
(U.S. EPA. 2013a). significant (p = 0.002) increases in sputum PMN (16-18 hours post-exposure)
relative to filtered air responses had been reported for 60 ppb ozone which is the lowest exposure
3-28
-------
concentration that has been investigated in young healthy adults. There was also some new evidence that
GSTM1 genotype may affect inflammatory responses to ozone, with greater PMN levels observed in
GSTMl-null subjects 24 hours after ozone exposure [see Genetic Polymorphisms on pg. 6-80 of U.S.
EPA (2013a)l. Study-specific details, including exposure concentrations and durations, are summarized in
Table 3-9 in Section 3.3.1. Some studies only appear in Table 3-9 to document that they were considered
but are not discussed in the text.
• Alexis et al. (2013) conducted a post hoc analysis of sputum PMN collected by Kim etal. (2011)
from 24 healthy adults (20-33 years; 12 GSTM1-positive, 12 GSTMl-null) 18 hours after
exposure to 60 ppb ozone or clean air for 6.6 hours with quasi-continuous exercise. Individuals
were stratified as PMN responders (10% or greater ozone-induced PMN increase, n = 13) or
nonresponders (n = 11). Responders were 13 times more likely to be GSTMl-null than
GSTM1-positive. Sputum macrophage phagocytosis was also significantly increased after filtered
air in responders compared with nonresponders (51 ± 2% vs. 45 ± 3%, p < 0.05). This result is
consistent with that of a study in the 2013 Ozone ISA showing macrophage oxidative burst and
phagocytic activity was increased in GSTMl-null compared with GSTM1-positive subjects
(Alexis et al.. 2009). However, a larger study of healthy older adults (52 F, 35 M;
59.9 ± 4.5 years) by Arjomandi et al. (2018) reported a significant increase in PMN following
120 ppb ozone relative to filtered air, which was not dependent on GSTM1 genotype
(50 GSTMl-null, 37 GSTM1-positive).
• Bosson et al. (2013) investigated the time course of pulmonary and peripheral PMN following a
2-hour exposure of subjects to 0 and 200 ppb ozone in an exposure chamber with moderate
exercise and rest. Following exposures, bronchoscopy was performed at 1.5 hours (5 F, 8 M;
24.6 years), at 6 hours (9F, 6M; 25 years), and at 18 hours (16 F, 13 M; 24.5 years). PMNs were
not increased at 1.5 hours post-exposure in either bronchial wash (BW) fluid or BALF.
Significant PMN increases were apparent at 6 hours in both the BW (4 times, p < 0.01),
BAL-fluid sample (1.5 times,/? < 0.05), and in the bronchial epithelium and submucosa biopsies.
At 18 hours, ozone-induced increase in PMN persisted both in BW (2 times,/? = 0.01) and BALF
(1.5 times,/? < 0.05). However, PMN in biopsies at 18 hours tended to be slightly lower than after
air. Based on a metabolomics analysis of BALF samples, Cheng et al. (2018) concluded that the
responses at 1 hour reflected oxidative stress, while the responses at 24 hours were consistent
with tissue repair. Consistent with prior work, studies using ozone to test anti-inflammatory
agents continue to report reproducible inflammatory responses following repeated ozone
exposures [e.g., Holz et al. (2015)1.
• Emphasizing the need for air control exposures, recent studies show exercise itself increased
blood PMNs (Bosson et al.. 2013). increased the occurrence of micronuclei in blood PMNs
(Holland et al.. 2014). and increased the pH of exhaled breath condensate (Hoffmever et al..
2015). Studies also show that changes in the blood and lungs should not be viewed as
independent of one another. There were significant correlations between PMNs in the lungs with
PMNs in the blood, which suggested that peripheral PMNs were reflective of the magnitude of
pulmonary inflammation (Bosson et al.. 2013). Another study reported that airway inflammation
was paralleled by systemic inflammation, with the percentage of PMN increasing in the blood at
5 hours after the start of a 3-hour ozone exposure and returning to baseline by 21 hours
post-exposure (Tanketal.. 2011).
• There were several analyses of inflammatory responses following ozone exposure in healthy
adults (Fry et al.. 2014) and groups of individuals with and without asthma (Fry et al.. 2012;
Hernandez et al.. 2012). but included no filtered-air control arm. Without an air control, it is not
possible to assess potential effects of exercise and/or the laboratory procedures on results. One of
3-29
-------
these studies reported that percent predicted FEVi both before and after ozone exposure did not
differ between PMN responders (>10% increase) and nonresponders (Fry et al.. 2012). This is
consistent with studies reviewed in the 2006 Ozone AQCD showing that spirometric measures
lung function and inflammatory responses to ozone are unrelated.
3.1.4.4.1.1 Factors Affecting Pulmonary Inflammation, Injury, and Oxidative Stress
Ambient Temperature
Gomes et al. (201 lb) exposed nine male endurance runners (24 ± 6 years) to 0 and 100 ppb ozone
at 20°C and 50% RH and at 31°C and 70% RH while they completed an 8 km time trial (i.e., each subject
completed four exposures). Nasal lavage was conducted approximately 15 minutes post-exposure. There
were no differences in inflammatory markers among the exposures. Although there were no differences
between the heat only or ozone only compared to control, levels of nasal Club cells following the
high-temperature ozone exposure were significantly increased (p = 0.03) relative to the lower temperature
air control. Glutathione concentrations were also significantly increased (p = 0.001) following the
high-temperature ozone exposure relative to the lower temperature air control. The increases in Club cells
and glutathione appeared to be additive, but no trend analysis was reported.
Lifestage
As reported in the 1996 and 2006 Ozone AQCDs (U.S. EPA. 2006. 1996a). decrements in lung
function and increases in respiratory symptoms in response to ozone exposure decreased with increasing
age. However, whether inflammatory responses persist with increasing age remained unstudied at the
time of the 2013 Ozone ISA (U.S. EPA. 2013a). Two recent studies demonstrated inflammatory
responses in older adults.
• Ariomandi et al. (2018) investigated changes in sputum markers of inflammation and injury in
healthy older adults (52 F, 35 M; 59.9 ± 4.5 years) exposed to 0, 70, and 120 ppb ozone for
3 hours during light-to-moderate, intermittent exercise. Sputum samples were obtained 22.5 hours
post-exposure. A mixed effects model showed marginally significant (p = 0.012)
concentration-dependent increases in PMNs by 4.1% of total (n.s.; p = 0.134) and 8.2% of total
(0.003) following 70 and 120 ppb ozone exposures, respectively. Sputum PMN increases
following ozone exposure showed no interaction with sex (52 F, 35 M), age (55-70 years), or
GSTM1 genotype (57% null, 43% positive). Due to the activity level and duration of exposure,
the total delivered ozone dose (120 ppb exposure) was estimated by Ariomandi et al. (2018) to be
about 60% of the delivered dose in the Kim etal. (2011) study, which identified a significant
increase in sputum PMN in young healthy adults following exposure to 60 ppb ozone. Sputum
IL-6, IL-8, TNF-a, and total protein concentrations did not show any significant changes due to
ozone exposure.
• Kirsten et al. (2011) studied Bimosiamose effectiveness in mitigating PMN response in healthy
older subjects (3 F, 15 M; 43.9 ± 7.4 years) who were found to be responsive (>20% increase in
sputum PMN) following exposure to 250 ppb ozone (no air control) for 3 hours with intermittent
3-30
-------
exercise (alternating 15 minutes intervals of rest and exercise at 14 L/minute per m2 BSA).
Sputum was collected 3 hours post-exposure. Another nine individuals (age and sex not specified)
were also exposed to ozone, but did not experience a sufficient increase in sputum PMN for
inclusion in the drug trial. Bimosiamose pretreatment of the 18 PMN responders reduced PMN
after ozone exposure to approximately the pre-exposure baseline. This study shows that 2/3 of the
screened subjects, who were older than the 18-35 year-old subjects typically examined in studies
available in prior reviews, were characterized as PMN responders to ozone.
It is not possible to quantify PMN responses as a function of age due to differences in
experimental protocols (i.e., duration of exposure to ozone, ozone concentration, activity level, and
post-exposure time of sputum collection). The new studies, nonetheless, show that inflammatory
responses following ozone exposure occur in older subjects.
3.1.4.4.2 Animal Toxicological Studies
As discussed in the 2013 Ozone ISA, ozone exposure affects both innate and adaptive immunity.
Both tissue damage and foreign pathogens are triggers for activating the innate immune system. This
results in the influx of neutrophils, mast cells, basophils, eosinophils, monocytes and dendritic cells and
the generation of cytokines such as TNF-a, IL-1, IL-6, keratinocyte-derived chemokine (KC), and IL-17.
Innate immunity encompasses the actions of complement and collectins, and the phagocytic functions of
macrophages, neutrophils, and dendritic cells. Airway epithelium also contributes to innate immune
responses. Innate immunity is highly dependent on cell signaling networks involving the toll like receptor
family including toll like receptor 4 (TLR4). Adaptive immunity provides immunologic memory through
the actions of B and T cells. Important links between the two systems are provided by dendritic cells and
antigen presentation.
The 2013 Ozone ISA summarized the animal toxicological evidence of injury, inflammation, and
oxidative stress resulting from exposure to ozone. These responses are hard to disentangle because injury
leads to inflammation and inflammation leads to further injury, with oxidative stress mediating both
injury and inflammation. A large number of studies have documented injury and inflammation in dogs,
rabbits, guinea pigs, rats, mice, and nonhuman primates. Numerous studies evaluated injury by assessing
histological lesions. In the lower respiratory tract, airway ciliated epithelial cells and type 1 alveolar cells
are the initial targets of ozone exposure and ozone exposure-mediated injury leads to epithelial
hyperplasia. In rats, repeated exposure to 0.2 ppm ozone over 7 days resulted in lesions at the junction of
the small airways and gas exchange region and included necrotic type 1 cells, hyperplastic type 2 cells,
damage to ciliated and nonciliated Club cells, and the accumulation of macrophages. In nonhuman
primates (macaques and rhesus monkeys), inflammation and related morphometric changes in necrotic
cells, smooth muscle, fibroblasts, and nonciliated bronchiolar cells of the tracheobronchial region of the
respiratory tract have been shown after 8-hour exposure to 1 ppm ozone. Repeated exposure of monkeys
to 0.2 ppm for 8 hours/day over 7 days also resulted in lesions in the respiratory bronchioles. Repeated
exposure to 0.15 ppm ozone over 6 days led to morphometric changes in lung, nose, and vocal cords in
3-31
-------
monkeys. Mucous cell metaplasia of nasal epithelium has been observed in both rodents and monkeys
exposed to ozone over several days.
Impaired epithelial barrier function has also been assessed as an index of injury. Histologic
evidence of damage to tight junctions and increased flux of small solutes from lung to plasma have been
demonstrated. Other studies have focused on assessing markers in the BALF. For injury, these markers
often include total protein, albumin, and shed epithelial cells. For inflammation, they include neutrophils
and cytokines/chemokines. The pattern of response varies depending on concentration, duration of
exposure, species, and strain. In general, acute (up to 8 hours) exposure to 0.8-2 ppm ozone and subacute
exposure (24-72 hours) to 0.3 ppm ozone reproducibly result in increased markers of injury and
inflammation in rodents, while acute exposure to 1 ppm ozone produces similar changes in nonhuman
primates. Attenuation of inflammatory and injury responses has been observed following repeated
exposures for some markers but not others in both rodents and nonhuman primates.
Studies evaluating oxidative stress in animals are fewer in number. However, ozone exposure
resulted in decreased levels of ascorbate in BALF of rodents, suggesting that ascorbate reacted with
secondary oxidation products produced in the epithelial lining fluid. In addition, ozone exposure
decreased glutathione levels in the respiratory bronchioles of rhesus monkeys. Ascorbate deficiency
enhanced ozone-induced lung injury, indicating a role for oxidative stress in the response.
Studies described in the 2013 Ozone ISA provide evidence for cell signaling pathways that
potentially underlie the injury and inflammation observed in animals exposed to ozone. Key roles have
been demonstrated for platelet activating factor, inter-cellular adhesion molecule 1 (ICAM-1), MIP-2,
TNF receptor, nuclear factor kappa-light-chain-enhancer of activated B cells (NFkB), c-Jun kinase 1
(JNK1), CXCR2, and IL-6 in mediating inflammation. Tachykinins, TLR4, and heat shock protein 70
(HSP70) have been shown to mediate change in barrier function. Pathways that confer protection against
injury and inflammation have also been investigated, with protective roles demonstrated for matrix
metalloproteinase-9, IL-10, surfactant protein-A (SP-A; acollectin), club cell secretory protein (CCSP),
and metallothionein.
Furthermore, ozone exposure skews immune responses towards an allergic phenotype. For
example, increased numbers of IgE-containing cells were found in the lungs of mice exposed to
0.5-0.8 ppm for 4 days. Other studies evaluated the effects of ozone exposure in animal models of
allergic airway disease but are discussed in Section 3.1.5.6.2 because these animal models share
phenotypic features with asthma.
A large number of recent studies evaluated respiratory tract injury, inflammation, and oxidative
stress in response to short-term ozone exposure. Study-specific details are summarized in Table 3-10.
Table 3-11. and Table 3-12 in Section 3.3.1. All of these studies were conducted in rodent strains with
varying degrees of sensitivity to ozone. Acute exposures generally consisted of a single exposure to
2 ppm ozone for 3 hours. Subacute exposures generally consisted of exposure to 0.3 ppm ozone for
3-32
-------
24-72 hours. One study compared responses with acute and subacute exposures overtime (C'ho ct al..
2013). Other exposure concentration-durations, including repeated exposures over several days, have
been employed. Some studies investigated the role of specific cell signaling pathways in mediating
responses by using genetic knockout models or pharmacologic agents. Inflammation, injury, and
oxidative stress were commonly assessed by measurement of cells and biological markers in the BALF.
Some studies also employed histopathology and/or immunocytochemistry of lung tissue. Flow cytometry
was used to identify different inflammatory cell subsets and cell surface markers on other cells present in
the lavage fluid or lung tissue.
Recent studies, detailed below and grouped according to concentration-time profile, provide
additional evidence for ozone-induced inflammation, injury, and oxidative stress. All of these changes,
which are described below, were statistically significant. These effects are seen following acute exposure
to 0.5-2 ppm ozone, subacute exposure to 0.3-0.7 ppm ozone, and repeated exposure to 0.5-2 ppm
ozone. Evidence is provided mainly by BALF markers, but some studies also provide evidence of
histological lesions. Responses to 0.1-0.25 ppm ozone are variable, with some studies demonstrating
mild histological lesions but not increases in BALF markers. A recent time-course study shows that the
earliest measurable response is epithelial barrier injury, as indicated by increased BALF protein levels.
Another study shows early activation of NFkB, a transcription factor that upregulates proinflammatory
genes, which precedes oxidative stress and a cytokine response. A plausible sequence of events begins
with ozone reacting with respiratory tract components to generate an oxidized species that disrupts barrier
function and activates innate immunity. A cascade of inflammation, injury, and oxidative stress responses
ensues. The influx of airway neutrophils is the hallmark of ozone exposure-induced inflammation, but
other inflammatory cell types infiltrate the lung following exposure. Some recent studies focus on
macrophage subpopulations that are pro- and anti-inflammatory and that infiltrate to the lung from the
spleen. A shift towards anti-inflammatory macrophages is correlated with the resolution of inflammation
following an acute exposure. Other studies examine eosinophilic inflammation, an indicator of atopy and
T helper 2 immunity. These latter studies, conducted in the nasal and lower airways, demonstrate that
repeated exposure to 0.5-0.8 ppm ozone result in airway eosinophilia, increased Th 2 cytokines, and
increased mucosubstances, consistent with induced nonatopic asthma and rhinitis. Innate lymphoid cells
were found to mediate this response to ozone. Of the many cell-signaling components and other factors
tested for inhibitory effects in recent studies, four were found to impact both airway responsiveness and
inflammation (ROCK, IL-17, osteopontin, and vitamin E). This finding suggests that there may be
common upstream events that trigger both increases in airway responsiveness and inflammatory processes
in healthy populations.
• Acute exposure to ozone (2 ppm for 3 hours) resulted in inflammation, injury, or oxidative stress
(Cho et al.. 2018; Mathews et al.. 2018; Tighe et al.. 2018; Malik et al.. 2017; Mathews et al..
2017a; Stober et al.. 2017; Elkhidir et al.. 2016; Mishra et al.. 2016; Barker et al.. 2015; Cabello
et al.. 2015; Kasahara et al.. 2015; Razvi et al.. 2015; Williams et al.. 2015; Ghio et al.. 2014; Bao
et al.. 2013; Barreno et al.. 2013; Cho et al.. 2013; Lee et al.. 2013; Sunil et al.. 2013; Hulo et al..
2011; Shore etal.. 2011).
3-33
-------
o In one case (Mathews et al.. 2017b). the evidence for inflammation was minimal given
that ozone exposure increased BALF levels of IL-33, but not neutrophils.
o The presence of two different macrophage subpopulations in the lung was reported:
classically activated (Ml) and alternatively activated (M2), with pro- and
anti-inflammatory roles, respectively (Sunil et al.. 2013).
o Several studies provide evidence for cell signaling and other pathways underlying
inflammation, injury, or oxidative stress effects of acute ozone exposure.
¦ TNF receptor 1 (Shore etal.. 2011)
¦ ROCK (Kasahara et al.. 2015)
¦ Nuclear factor (erythroid-derived 2)-like 2 [NRF2; Cho et al. (2013)1
¦ Osteopontin (Barreno et al.. 2013)
¦ Dysregulated iron homeostasis (Ghio et al.. 2014)
o Evidence for ozone-induced release of tachykinins is provided by one study (Barker et
al.. 2015). Tachykinins mediate bronchoconstriction and neurogenic inflammation. Acute
ozone exposure increased levels of the tachykinin substance P levels in the BALF
through upstream effects on IL-1(3 and nerve growth factor. Two studies examined
inflammation following acute ozone exposure in rodents of varying lifestages or sex.
Less inflammation was found in older rodents compared with younger ones (Shore et al..
2011). Females had greater inflammatory responses compared with males (Mishra et al..
2016; Cabello et al.. 2015). Acute exposure to ozone (2 ppm, 4-6 hours) also resulted in
inflammation, injury, and oxidative stress (Verhein et al.. 2013; Yanagisawa et al.. 2012).
o Tighe et al. (2018) found that different methods employed for euthanasia, but not for
lavage, were a source of variability in the measured indices of inflammation and injury
parameters.
• Acute exposure to 0.8-1 ppm ozone for 1-6 hours resulted in inflammation, injury, or oxidative
stress (Michaudel et al.. 2018; Wong et al.. 2018; Francis et al.. 2017a; Francis et al.. 2017b;
Yonchuk et al.. 2017; Zvchowski et al.. 2016; Gabehart et al.. 2015; Hatch et al.. 2015; Kodavanti
et al.. 2015; Kumarathasan et al.. 2015; Paffett et al.. 2015; Ramot et al.. 2015; Sunil et al.. 2015;
Ward et al.. 2015; Ward and Kodavanti. 2015; Gonzalez-Guevara et al.. 2014; Bhoopalan et al..
2013; Groves et al.. 2013; Kummarapurugu et al.. 2013; Robertson et al.. 2013; Connor et al..
2012; Groves et al.. 2012).
o Some of these were time-course studies.
¦ Michaudel et al. (2018) followed changes for up to 48 hours post-exposure to a
1-hour exposure to ozone. The earliest measured response was the injury marker,
BALF protein, which was increased 1-hour post-exposure and reflects barrier
disruption. This response was followed by increases at 4-6 hours in BALF
cytokines and chemokines, lung tissue interstitial macrophages, and another
marker of epithelial cell injury. Later responses began at 18 hours and included
increases in BALF neutrophils, eosinophils, reactive oxygen-producing cells and
cell death markers. The time dependence of effects on tight junction proteins was
also reported.
¦ Gonzalez-Guevara et al. (2014) examined early responses and found increases in
lung tissue TNF-a immediately after 3 and 6 hours of exposure, but not after
1 hour of exposure.
3-34
-------
¦ NFkB activation, which is an early step in the induction of inflammation,
occurred prior to changes in oxidative stress and upregulation of the cytokine
TNF-a (Connor et al.. 2012).
¦ Resolution of inflammation and injury within 72 hours following ozone exposure
was demonstrated (Groves et al.. 2012). However, airway resistance remained
increased, indicating that the lung was functionally compromised.
o Several studies focused on the accumulation of macrophages in the lung in response to
ozone exposure (Francis et al.. 2017b; Sunil et al.. 2015; Groves et al.. 2013).
¦ Resident alveolar macrophages were not affected, but infiltrating monocytic and
granulocytic cells were increased.
¦ Increases in both classically activated macrophages (Ml, proinflammatory) and
alternatively activated macrophages (M2, anti-inflammatory) were found.
¦ A time-course study showed that Ml macrophages increased in number rapidly
and persisted for 72 hours post-exposure, while M2 macrophages were increased
beginning at 72 hours.
¦ The spleen was found to be a source for these Ml and M2 cells.
o Some studies provide evidence for cell signaling and other pathways underlying the
inflammatory, injury, or oxidative effects of acute ozone exposure.
¦ CD36 (Robertson et al.. 2013)
¦ IL-33 and ST2 (Michaudel et al.. 2018)
¦ Glucocorticoids (Thomson et al.. 2016)
¦ TLR4 (Connor et al.. 2012)
¦ Galectin (Sunil et al.. 2015)
¦ CC chemokine receptor type 2 (CCR2) (Francis et al.. 2017a)
o Other endpoints examined include mucus secretion (Gabehart et al.. 2015) and
upregulation of glucocorticoid-sensitive genes (Thomson et al.. 2016; Thomson et al..
2013).
¦ Mucus secretion was not seen in juvenile or adult mice in response to ozone.
¦ Transient changes in glucocorticoid-sensitive genes occurred immediately after
exposure to ozone.
o One study provides evidence for respiratory effects of acute ozone exposure in rodents of
varying lifestages. In Gabehart et al. (2015). inflammation and injury responses in 2-, 3-,
and 6-week-old mice, representing weanling, juvenile, and adult stages, respectively,
were examined. Results in 1-week-old mice (neonates) similarly exposed are discussed in
the long-term exposure section of this Appendix. In general, responses were smallest in
1-week-old mice and greatest in 6-week-old mice, with responses in the 2- and
3-week-old mice sometimes in between and sometimes as high as responses in the
6-week-old mice. The exception was mucus secretion, which was highest in 1-week-old
mice and minimal in 2-week-old mice.
• Acute exposure to 0.25-0.5 ppm ozone resulted in minimal or no changes in inflammation,
injury, or oxidative stress markers in BALF (Michaudel et al.. 2018; Kodavanti et al.. 2015;
Kumarathasan et al.. 2015; Kurhanewicz et al.. 2014; Mclntosh-Kastrinskv et al.. 2013; Thomson
3-35
-------
et al.. 2013). Histopathological lesions were seen in response to 0.25 and 0.5 ppm ozone (Ramot
et al.. 2015).
• Subacute exposures to 0.3-0.7 ppm ozone for up to 72 hours resulted in inflammation, injury, and
oxidative stress (Che et al.. 2016; Mathews et al.. 2015; Verhein et al.. 2015; Kasahara et al..
2014; Cho et al.. 2013; Kasahara et al.. 2013; Kasahara et al.. 2012).
o Several studies involving subacute exposure to 0.3 ppm ozone examined the time course
of changes in inflammatory cells (Mathews et al.. 2015; Kasahara et al.. 2014; Kasahara
et al.. 2013; Kasahara et al.. 2012).
¦ Increases in BALF neutrophils and protein (a marker of injury) occurred earlier
(24 hours) than changes in BALF macrophages (48 hours).
¦ These changes persisted for up to 72 hours.
¦ Macrophage subpopulations in lung tissue consisted of Ml proinflammatory and
M2 anti-inflammatory macrophages, as well as macrophages positive for IL-6
and apoptotic macrophages.
¦ Numbers of gamma delta T cells were also increased and contributed to
resolution of inflammation that occurred over several days post-exposure.
o Another study (Che et al.. 2016). which involved subacute exposure to 0.7 ppm ozone,
found increased IL-17A-producing gamma delta T cells.
o Dendritic cells were increased by subacute exposure to 0.4 ppm ozone, but there was no
impact on neutrophilic inflammation (Brand et al.. 2016).
o Evidence of mucus hypersecretion was found in Cho et al. (2013).
o Some studies provide evidence for cell signaling and other pathways underlying
inflammation, injury, or oxidative stress or effects on mucus secretion resulting from
subacute ozone exposure.
¦ Adiponectin (Kasahara et al.. 2012)
¦ T-cadherin (Kasahara et al.. 2013)
¦ IL-6 (Kasahara et al.. 2014)
¦ IL-17A and gamma delta T cells (Mathews et al.. 2015)
¦ NRF2 (Cho etal.. 2013)
¦ Notch (Verhein et al.. 2015)
¦ Mannose binding lectin (Ciencewicki et al.. 2016)
¦ IL-17A, IL-1R, and caspase (Che et al.. 2016)
• Repeated exposure to ozone (0.5-2 ppm), using many different concentration-duration profiles,
resulted in inflammation, injury, or oxidative stress in many studies (Snow et al.. 2018; Gordon et
al.. 2017b; Gordon et al.. 2017a; Harkema et al.. 2017; Henriquez et al.. 2017; Kumagai et al..
2017; Zhang et al.. 2017; Kumagai et al.. 2016; Liu et al.. 2016; Miller et al.. 2016b; Ong et al..
2016; Zhu et al.. 2016; Feng et al.. 2015; Gonzalez-Guevara et al.. 2014; Tankerslev et al.. 2013;
Wang et al.. 2013; Brand et al.. 2012; Xiang et al.. 2012).
o Effects were found in the upper (nasal) airways following repeated exposure to
0.5-0.8 ppm ozone for up to 9 days (Harkema et al.. 2017; Kumagai et al.. 2016; Ong et
al.. 2016).
3-36
-------
o Some studies of repeated ozone exposures provide evidence for cell signaling and other
pathways underlying inflammation, injury, or oxidative stress.
¦ Vitamin E (Zhu et al.. 2016)
¦ Epidermal growth factor (EGF) receptor (Feng et al.. 2015)
¦ Glucocorticoids and stress hormones (Miller et al.. 2016b)
o Ozone exposure induced inflammation, injury, or oxidative stress in rodents of varying
lifestage from adolescence to senescence (Snow et al.. 2016).
o Mucous cell metaplasia indicated by increased mucosubstances, eosinophilic
inflammation, Th2 cytokines, and other type 2 immune responses were seen in the upper
(nasal) and lower airways (Harkema et al.. 2017; Kumagai et al.. 2017; Kumagai et al..
2016; Ong et al.. 2016).
¦ These findings are characteristic of induced nonatopic asthma and rhinitis.
¦ A role for immune lymphoid cells in the development of type 2 immunity was
demonstrated.
o Zhu et al. (2016) found effects on Th2 cytokines, mast cell degranulation, serum IgE, and
TSLP in the lower respiratory tract that were attenuated by treatment with vitamin E.
o Brand et al. (2012) found dendritic cell activation and increased T cell number in specific
regions of the respiratory tract.
• No evidence of inflammation, injury, or oxidative stress was found in other studies involving
repeated exposure to 0.1-0.5 ppm ozone (Gordon et al.. 2017b; Snow et al.. 2016; Feng et al..
2015; WolkoffetaL 2012).
3.1.4.4.3 Epidemiologic Studies
A limited number of studies evaluated in the 2013 Ozone ISA (U.S. EPA. 2013a) reported
consistent evidence of ozone-related pulmonary inflammation in children without asthma (Berhane et al..
2011; Barraza-Villarreal et al.. 2008). There are no recent studies in the U.S. or Canada that examine the
relationship between short-term ozone exposure and pulmonary inflammation in healthy populations.
Recent studies have examined pulmonary inflammation in general population studies of children
(Patel et al.. 2013; Salam et al.. 2012). with asthma prevalence ranging from 14 to 47%. Because these
studies do not directly inform the understanding of the relationship between short-term ozone exposure
and pulmonary inflammation in healthy populations, they are not discussed here. However, study specific
details can be found in Table 3-7 in Section 3.3.1.
3.1.4.4.4 Integrated Summary for Respiratory Tract Inflammation, Injury, and Oxidative
Stress
Controlled human exposure studies evaluated in the 1996 and 2006 Ozone AQCDs (U.S. EPA.
2006. 1996a) established evidence of respiratory tract inflammation in response to acute ozone exposures.
3-37
-------
Notably, these inflammatory responses are not correlated with lung function changes but are at least
partially correlated with airway resistance. These results indicate that changes in pulmonary inflammation
and airway obstruction may share similar underlying mechanisms, while inflammation and lung function
as measured by FEVi may not. Additionally, the evidence suggested that there is interindividual
variability in inflammatory responses to ozone. This was expanded upon in the 2013 Ozone ISA (U.S.
EPA. 2013a) in studies that demonstrated GSTM1 genotype interaction with ozone exposure on
pulmonary inflammation. A recent study (Alexis et al.. 2013) provides some further evidence suggesting
that young healthy GSTMl-null adults may be more susceptible to ozone-related inflammatory responses,
although the evidence is not entirely consistent, with a relatively large study of older adults by Ariomandi
et al. (2018) reporting that inflammatory responses to ozone are not dependent on GSTM1 genotype
(50 GSTM1 null, 37 GSTM1 positive).
Consistent with experimental studies in humans, a large body of evidence from recent animal
toxicological studies and studies previously evaluated in the 2013 Ozone ISA (U.S. EPA. 2013a)
demonstrate inflammatory responses to acute, subacute, and repeated ozone exposures in various animal
models. Additionally, results from recent experimental animal studies are also consistent with previous
findings of ozone-related pulmonary injury (0.3-2 ppm ozone) and oxidative stress (0.15-2 ppm ozone).
Mechanistic studies present a plausible pathway by which ozone reacts with respiratory tract components,
produces oxidized species that injure barrier function and activates innate immunity, resulting in a cycle
of inflammation, injury, and oxidative stress.
A limited number of epidemiologic panel studies evaluated in the 2013 Ozone ISA (U.S. EPA.
2013a) observed evidence of pulmonary inflammation in children without asthma associated with
short-term ambient ozone exposure. These results are coherent with results from experimental studies in
humans and animals. No recent studies in the U.S. or Canada are available for review.
3.1.4.5 Overall Summary of Respiratory Effects in Healthy Populations
Evidence from recent controlled human exposure studies of respiratory effects in healthy
populations is generally consistent with findings from prior assessments (U.S. EPA. 2013a. 2006. 1996a).
Notably, there is consistent evidence demonstrating ozone-induced decreases in group mean pulmonary
function in young, healthy adults performing moderate exercise. Lung function decrements were observed
after ozone exposures as low as 60 to 70 ppb, for young adults, and 120 ppb in older adults. The 2013
Ozone ISA also evaluated studies that indicate symptoms of cough and pain on deep inspiration
corresponding to FEVi decrements in healthy young adults exposed to 70 ppb ozone for 6.6 hours (U.S.
EPA. 2013a).
Controlled human exposure studies evaluated in the 2006 Ozone AQCD (U.S. EPA. 2006) and
the 2013 Ozone ISA (U.S. EPA. 2013a) provide consistent evidence of ozone-induced increases in airway
responsiveness and inflammation in the respiratory tract and lungs. A recent study is consistent with
3-38
-------
previous findings and expand on observed interindividual variability in inflammatory responses,
providing some additional evidence suggesting that young healthy GSTMl-null adults may be more
susceptible to ozone-related inflammatory responses.
Like studies of humans, experimental studies of animals also provide evidence of respiratory
effects resulting from exposure to ozone. Evidence summarized in the 2013 Ozone ISA indicated changes
in the frequency of breathing and tidal volume, decreased lung volume, increased airway resistance, and
attenuation of the pulmonary function decrement response following repeated exposures to ozone (U.S.
EPA. 2013a). Additionally, previously evaluated studies indicate ozone-induced increases in airway
responsiveness, inflammation, injury, and oxidative stress. A large body of recent evidence further
demonstrates changes in each of the specified endpoints resulting from ozone exposure, providing
coherence with results from controlled human exposure studies.
Recent mechanistic studies in humans and animals expand on findings from previously reviewed
studies to provide plausible pathways that may underlie the observed respiratory health effects resulting
from short-term exposure to ozone. Experimental studies in both humans and animals indicate that
changes in lung function may be attributed to activation of sensory nerves in the respiratory tract that
trigger local and autonomic reflex responses. Specifically, mechanistic studies provide evidence that local
reflex responses mediate the observed decreases in inspiratory capacity and pain on inspiration that result
in truncated inspiration. In addition, modest increases in airway resistance may occur due to activation of
parasympathetic pathways. Mechanistic studies also present a plausible pathway by which ozone reacts
with respiratory tract components, produces oxidized species that injure barrier function, and activates
innate immunity, resulting in a cycle of inflammation, injury, and oxidative stress.
Evidence from epidemiologic studies is generally coherent with experimental evidence. Most of
the epidemiologic evidence comes from panel studies of healthy children that were previously evaluated
in the 2013 Ozone ISA (U.S. EPA. 2013a). Several of these panel studies examined children in summer
camps and demonstrated decreases in FEVi associated with short-term ozone exposure. A smaller body of
panel studies in children without asthma also consistently reported associations between ozone and
increases in markers of pulmonary inflammation. While there is coherence between epidemiologic and
experimental evidence of ozone-induced lung function decrements and pulmonary inflammation,
respiratory symptoms were not associated with ozone exposure in a limited number of epidemiologic
studies. However, these studies generally relied on parental reported outcomes that may result in under-
or over-reporting of respiratory symptoms.
3.1.5 Asthma Exacerbation and Associated Respiratory Effects in
Populations with Asthma
Asthma is a chronic inflammatory lung disease characterized by reversible airway obstruction and
increased airway responsiveness. Exacerbation of asthma is associated with symptoms such as wheeze,
3-39
-------
cough, chest tightness, and shortness of breath. Symptoms may be treated with asthma medication, while
uncontrollable symptoms may lead to medical treatment, including ED visits and, in extreme cases,
hospital admissions. In characterizing the relationship between ozone and asthma exacerbations, this
section sequentially considers the effects of short-term exposure to ozone on hospital admissions and ED
visits for asthma, respiratory symptoms and asthma medication use, lung function, and subclinical effects,
such as pulmonary inflammation and oxidative stress, in people with asthma. ED visits for asthma are
more common and often less serious than hospital admissions. Generally, only a small fraction of
respiratory ED visits result in a hospital admission. Accordingly, the two outcomes may reflect different
severities of asthma and are evaluated separately.
3.1.5.1 Hospital Admissions
A single study evaluated in the 2013 Ozone ISA (U.S. EPA. 2013a) examined the association
between short-term exposure to ozone and hospital admissions for asthma. In New York City, 8-hour max
ozone concentrations were associated with severe acute asthma admissions in the warm season
(Silverman and Ito. 2010). The authors reported positive associations with non-ICU asthma admissions
and severe (ICU) asthma admissions that were strongest (i.e., of greatest magnitude) for children ages 6 to
18 years, compared to the other age groups examined (ages <6 years, 19-49, 50+, and all ages).
Associations for non-ICU asthma admissions were observed for all age groups, while ICU admissions
were only associated with ozone in children ages 6-18 years old. The observed effect remained robust to
adjustment for PM2 5. The authors also performed an analysis examining the shape of the
concentration-response (C-R) relationship, which is discussed in more detail in Section 3.1.10.4.
Recent studies expand the existing evidence base and provide consistent evidence of an
association between ozone and hospital admissions for asthma in children (Figure 3-4). Recent evidence
was also generally consistent for an association in adults, although the relationship is consistently null for
older adults. Generally, the evaluated studies use an 8-hour daily max averaging time, although there are
studies that use daily 8-hour avg (Zu et al.. 2017) and 24-hour avg (Shmool et al.. 2016). The averaging
time used in each study, along with other study-specific details, including air quality characteristics and
select effect estimates, are highlighted in Table 3-13 in Section 3.3.1. An overview of the evidence is
provided below.
• Multicity studies in Texas (Goodman et al.. 2017b; Zu et al.. 2017) and single-city studies in New
York City (Goodman et al.. 2017a; Shmool et al.. 2016; Sheffield et al.. 2015) and St. Louis, MO
(Winquist et al.. 2012) reported evidence of an association between short-term ozone
concentrations and hospital admissions for asthma.
• Shmool et al. (2016) compared monitor-based ozone concentrations to ozone estimated at a
300-m spatial scale using a fusion of monitoring data and land-use regression (LUR). In short, the
authors used LUR with local monitoring inputs to estimate seasonal average concentrations
within 300 m radial buffers around geocoded participant addresses. The ratio of these spatially
resolved seasonal average concentrations and the citywide averages were multiplied by daily
3-40
-------
monitor averages to estimate spatially-refined daily exposures. The effect estimates derived from
the spatiotemporal model were similar to those estimated using monitored ozone concentrations.
These results indicate that the observed association of ozone concentrations with asthma hospital
admissions is robust to exposure assignment technique.
• Like previous findings from Silverman and Ito (2010). recent studies that reported age-stratified
results (Goodman et al.. 2017b; Goodman et al.. 2017a; Zu et al.. 2017; Sheffield et al.. 2015)
generally observed ozone-asthma hospital admission associations that were strongest (i.e., greater
in magnitude) in younger populations (5 to 18 years of age). Many studies exclude data for
children less than 5 years of age due to less reliable asthma diagnosis in young children.
Additionally, most studies that examined hospital admissions in adults older than 50 reported null
associations. While most studies observed associations in analyses of all ages combined, stratified
analyses suggest that these associations are likely being driven by hospital admissions among
children.
• In recent studies, there was some limited evaluation of the shape of the C-R relationship (Zu et
al.. 2017). potential copollutant confounding (Shmool et al.. 2016). and seasonal differences in
effect estimates (Goodman et al.. 2017a) across the evaluated studies. These topics are discussed
in more detail in the Relevant Issues for Interpreting Epidemiologic Evidence section
(Section 3.1.10).
3-41
-------
Mean
I
Study
Location
Ages
Concentrations3
Season
Lag
i
i
Silverman et al (2010)
New York, NY
<6
41
Warm
0-1
1
1 ^
1
tGoodman etal. (2017)
New York, NY
<6
30.7
Warm
0-1
4
1
tZu et al. (2017)
6 Texas Cities
5-14
32.2; 8-hr avg
Year-Round
0-3
1
1
1
tGoodman et al. (2017)
3 Texas Cities
5-14
41.8
Year-Round
0
|
Silverman et al. (2010)
New York, NY
6-18
41
Warm
0-1
1
1
tGoodman etal. (2017)
New York, NY
6-18
30.7
Warm
0-1
1
tZu et al. (2017)
6 Texas Cities
15-64
32.2; 8-hr avg
Year-Round
0-3
1
tGoodman etal. (2017)
3 Texas Cities
15-64
41.8
Year-Round
0
V
Silverman et al. (2010)
New York, NY
19-49
41
Warm
0-1
!
tGoodman etal. (2017)
New York, NY
19-49
30.7
Warm
0-1
1-#-
i
Silverman et al (2010)
New York, NY
50+
41
Warm
0-1
i
i
tGoodman etal. (2017)
New York, NY
50+
30.7
Warm
0-1
i
tZu etal. (2017)
6 Texas Cities
65+
32.2; 8-hr avg
Year-Round
0-3
-o-
tGoodman etal. (2017)
3 Texas Cities
65+
41.8
Year-Round
0
1
tWinquistet al. (2012)
St. Louis, MO
All
41
Year-Round
0-4 DL
1
' •
tZu etal. (2017)
6 Texas Cities
All (5+)
32.2; 8-hr avg
Year-Round
0-3
1
1
tGoodman etal, (2017)
3 Texas Cities
All (5+)
41.8
Year-Round
0
»
1
Silverman et al. (2010)
New York, NY
All
41
Warm
0-1
1
tGoodman etal. (2017)
New York, NY
All
30.7
Warm
0-1
~
, 1 ,
0.9 1 1.1 1.2 1.3
Effect Estimate {95% CI)
DL = distributed lag.
Note: tStudies published since the 2013 Ozone ISA. Black text = studies included in the 2013 Ozone ISA.
aMean concentrations reported in ppb and are for an 8-hour daily max averaging time, unless otherwise noted.
Results standardized to a 15-ppb increase in 24-hour avg, 20-ppb increase in 8-hour daily max, or 25-ppb increase in 1 -hour daily
max ozone concentrations. Corresponding quantitative results are reported in Table 3-5.
Figure 3-4 Summary of associations from studies of short-term ozone
exposures and hospital admissions for asthma for a standardized
increase in ozone concentrations.
3.1.5.2 Emergency Department (ED) Visits
A number of studies evaluated in the 2013 Ozone ISA (U.S. EPA. 2013a) examined the
association between short-term ozone exposure and ED visits for asthma. A multicity study in Canada
(Stieb et al.. 2009) as well as single-city studies in the U.S. and Canada (Ito et al.. 2007; Villeneuve et al..
2007) provided consistent evidence that increased ozone exposure is associated with increases in asthma
3-42
-------
ED visits. The observed associations were consistently stronger in magnitude in the warm season. Ito et
al. (2007) ran three separate models with varying degrees of adjustment for weather-related variables and
observed associations that varied in magnitude but were consistently positive for each of the models.
Additionally, the authors reported an association that was robust in copollutant models that adjusted for
PM2 5, N02, S02, and CO.
Recent studies continue to present consistent evidence of an association between ozone and ED
visits for asthma across a number of study locations, a range of mean ozone concentrations, and a variety
of study designs and exposure assignment techniques, including population-weighted monitor averages,
CMAQ modeling estimates, and fusions of modeled and monitored data (Figure 3-5). Generally, the
evaluated studies use an 8-hour daily max averaging time, although there are instances in which the
1-hour daily max (Malig et al.. 2016) and 24-hour avg (Szvszkowicz et al.. 2018; Sarnat et al.. 2013) are
used. The averaging time used in each study, along with other study-specific details, including air quality
characteristics and select effect estimates, are highlighted in Table 3-14 in Section 3.3.1. Additionally,
information on potential copollutant confounding and seasonal differences in effect estimates is presented
in the Relevant Issues for Interpreting Epidemiologic Evidence section (Section 3.1.10). An overview of
the recent evidence is provided below.
• The strongest evidence of an association between short-term exposure to ozone and ED visits for
asthma is presented in multicity studies, including statewide studies conducted in California
(Malig et al.. 2016). North Carolina (Sacks et al.. 2014). and Georgia (Xiao et al.. 2016) and in
other multicity studies in the U.S. (Barry et al.. 2018; Alhanti et al.. 2016; Gleason et al.. 2014)
and Canada (Szvszkowicz et al.. 2018).
• Supporting evidence is provided by single-city studies in New York (Shmool et al.. 2016;
Sheffield et al.. 2015). Atlanta (O'Lenick et al.. 2017; Strickland et al.. 2014; Winquist et al..
2014; Sarnat et al.. 2013). and elsewhere (Bvers et al.. 2015; Sarnat etal.. 2015). demonstrating
consistent increases in ED visits for asthma corresponding to short-term ozone exposure.
• Most recent studies of ED visits for asthma included all ages and/or focused more specifically on
children. The evidence is consistent for both study populations. A limited number of studies
examined ED visits for asthma in adults, and reported some evidence that associations exist
among these older age groups (Alhanti et al.. 2016; Bvers et al.. 2015).
3-43
-------
Study
Location
Ages
Mean
Concentration
Lag
Season
1
1
|
Stieb 2009
tMalig etal. (2016)
Multicity. Canada
California
All
All
10.3-22.1; 24-hr avg
33-55; 1-hr max
1
0-1
Year-Round
'(*¦
!
tSacks etal. (2014)
North Carolina
All
436
0-2
I*
tSarnat et al. (2013)
Atlanta, GA
All
41.9; 24-hr avg
0-2
l
1
tWinquistet al. (2012)
St. Louis, MO
All
NR
0-4 DL
tBarry etal. (2018)
Atlanta, GA
Birmingham, AL
Dallas, TX
Pittsburgh, PA
St, Louis, MO
All
37 5-42.2
0-2
!~
!•-
l
tSarnat etal. (2015)
St. Louis, MO
All
36.2
0-2 DL
tAlhantietal. (2016)
3 U.S. Cities
0-4
37.3-43.7
0-2
I*-
-(-Strickland et al. (2014)
Atlanta, GA
2-16
42.2
0-2
i
i
fXiao etal. (2016)
Georgia
2-1 a
42 1
0-3
;~
tO'Lenick et al. (2017)
Atlanta, GA
5-17
NR
0-2
! -©¦
tAlhanti etal. (2016)
3 U.S. Cities
5-18
37.3-43.7
0-2
i
i •
\
fSzyszkowiczetal, (2018)
Multicity, Canada
<20; Females
<20; Males
22.5-29.2; 24-hr avg
1
r»-
tAlhanti etal. (2016)
3 U.S. Cities
19-39
40-64
65+
37.3-43.7
0-2
~
¥
i
tMalig etal. (2016)
California
All
33-55; 1-hr max
0-1
Warm
i
;
tSacks etal. (2014)
North Carolina
All
50.1
0-2
y
tByers et al. (2015)
Indianapolis, IN
All (5+)
48 5
0-2
t
*
tGleason et al. (2014)
New Jersey
3-17
NR
0-2
• ®.
tWinquistet al. (2014)
Atlanta, GA
5-17
53.9
0-2
!
tSheffield et al. (2015)
New York. NY
5-17
NR
0-3
t
tByers et al (2015)
Indianapolis, IN
5-17
18-44
45+
48 5
0-2
-j—•—
i—•—
-•-I—
tWinquistet al. (2014)
Atlanta, GA
5-17
53.9
0-2
Cold
i 1 r
0.9 1 1.1 1.2
Effect Estimate (95% CI)
Note: jStudies published since the 2013 Ozone ISA. Black text = studies included in the 2013 Ozone ISA.
aMean concentrations reported in ppb and are for an 8-hour daily max averaging time unless otherwise noted.
Results standardized to a 15-ppb increase in 24-hour avg, 20-ppb increase in 8-hour daily max, or 25-ppb increase in 1-hour daily
max ozone concentrations. Corresponding quantitative results are reported in Table 3-6.
Figure 3-5 Summary of associations from studies of short-term ozone
exposures and asthma emergency department (ED) visits for a
standardized increase in ozone concentrations.
3-44
-------
3.1.5.3 Respiratory Symptoms
Respiratory symptoms in children with asthma, including cough, wheeze, sputum production,
shortness of breath, and chest tightness, may indicate an exacerbation of disease. Further, uncontrollable
symptoms may lead people with asthma to seek medical care. Thus, studies examining the association
between ozone and increases in asthma symptoms and medication use may provide biological plausibility
for ozone-induced acute respiratory effects, as supported by evidence of ozone-related increases in asthma
hospital admissions and ED visits in children.
3.1.5.3.1 Controlled Human Exposure Studies
As discussed in Section 3.1.4.2.1. controlled human exposure studies of healthy adults clearly
demonstrate ozone-induced increases in respiratory symptoms including pain on deep inspiration,
shortness of breath, and cough. In Section 7.5.1.2 of the 1996 Ozone AQCDs (U.S. EPA. 1996a).
individuals with and without asthma were reported to have similar respiratory symptom responses to
ozone exposure; however, the study by Horstman et al. (1995) showed that after 7.6 hours of exposure to
160 ppb ozone with light quasi-continuous exercise there was a statistically significant increase in the
incidence of wheeze in subjects with asthma relative to healthy controls, which did not experience
wheeze. There was also a statistically significant increase in the incidence of wheeze in the subjects with
asthma on their ozone exposure day relative to their filtered-air exposure day. These observations are not
changed by recently available studies or those in subsequent assessments (U.S. EPA. 2013a. 2006).
3.1.5.3.2 Epidemiologic Studies
A number of epidemiologic panel studies evaluated in the 2013 Ozone ISA (U.S. EPA. 2013a)
examined the relationship between short-term ozone exposure and incidence of respiratory symptoms and
increased symptom scores in children with asthma. Evidence from a limited number of multicity U.S.
studies was inconsistent, but many single-city studies provided evidence of an association (see
Section 6.2.4.1 of the 2013 Ozone ISA). Methodological distinctions in the evaluated multicity studies,
including lack of power and extended averaging times (e.g., 19-day avg), reduced the consideration given
to the observed results. Associations were observed in single-city panel studies across a diversity of
locations and ambient ozone concentrations.
One recent panel study of school-aged children in Detroit tracked respiratory symptoms in
children with asthma for periods of 14 consecutive days during 11 seasons (Lewis et al.. 2013). The
authors reported increases in a range of respiratory symptoms, including cough, wheeze, shortness of
breath, and chest tightness, associated with increases in 1- and 8-hour daily max ozone concentrations.
Associations were generally negative or null at lag 1, but positive at lags 2, 3-5, and 1-5. Consistent with
results from studies evaluated in the 2013 Ozone ISA, Lewis et al. (2013) observed associations that were
3-45
-------
larger in magnitude in children taking corticosteroids. However, these associations were much less
precise (i.e., wider 95% CIs) than the associations for children not taking steroids. See Table 3-15 in
Section 3.3.1 for complete study details.
3.1.5.3.3 Integrated Summary for Respiratory Symptoms
Controlled human exposure studies provide evidence of ozone-induced increases in respiratory
symptoms in individuals with asthma, with respiratory symptom responses that are generally comparable
to those from individuals without asthma. A number of epidemiologic panel studies evaluated in the 2013
Ozone ISA also provided evidence that ozone exposure is associated with increased respiratory symptoms
in children with asthma. A recent epidemiologic panel study provides additional evidence that ozone
concentrations are associated with a range of respiratory symptoms in children with asthma.
3.1.5.4 Lung Function
3.1.5.4.1 Controlled Human Exposure Studies
Based on studies reviewed in the 1996 and 2006 Ozone AQCDs (U.S. EPA. 2006. 1996a) and the
2013 Ozone ISA (U.S. EPA. 2013a). it was concluded that individuals with asthma were at least as
sensitive to acute effects of ozone as healthy individuals. In the 2013 Ozone ISA (U.S. EPA. 2013a). the
study by Horstman et al. (1995) was recognized as showing clearly larger FEVi responses in individuals
with asthma relative to those without (19 vs. 10% FEVi decrements, respectively,/? = 0.04) following
7.6-hour exposures to 160 ppb ozone with light quasi-continuous exercise. In asthmatics, ozone-induced
FEVi decrements were also well correlated with baseline percent predicted FEVi (r = 0.53,/? < 0.05); that
is, responses to ozone increased with severity of disease, and individuals using bronchodilators
experienced greater ozone-induced lung function decrements. Based on FEVi/FVC, this study also
showed that the obstructive response to ozone is greater in individuals with asthma than those without.
Kreit et al. (1989) also reported a large statistically significant difference in ozone-induced FEVi
decrements between individuals with asthma and those without (25 vs. 16%, respectively, p < 0.05)
exposed to 400 ppb ozone with heavy intermittent exercise for 2 hours. Overall, however, the majority of
controlled human exposure studies found little to no difference in ozone-induced lung function responses
between individuals with and without asthma.
• Since the 2013 Ozone ISA, four controlled human exposure studies examining ozone effects on
lung function in individuals with asthma have been published (Arjomandi et al.. 2015; Lerov et
al.. 2015; Bartoli et al.. 2013; Fry et al.. 2012). Study-specific details, including exposure
concentrations and durations, are summarized in Table 3-16 and EI3-13 in Section 3.3.1.
• Neither Arjomandi et al. (2015) nor Fry et al. (2012) reported FEVi responses to ozone
differentiated by the presence of asthma.
3-46
-------
• Consistent with Horstman et al. (1995). in a large study of individuals with asthma (34 F, 86 M;
32.9 ± 12.9 years), Bartoli et al. (2013) found that the magnitude of ozone-induced FEVi
response increased with decreasing baseline FEVi (p = 0.02). Bartoli et al. (2013) also found that
inhaled corticosteroid treatment was associated with a decrease in ozone-induced FEVi
decrements (p = 0.04). This study, however, did not include a healthy nonasthmatic control
group, limiting our understanding of differences between asthmatic and nonasthmatic individuals.
In a smaller study of healthy nonasthmatic individuals (5 F, 7 M; 31.8 ± 6.0 years) and subjects
with mild asthma (5 F, 3 M; 33.7 ± 10.1 years), although baseline FEVi and FEVi/FVC were
significantly lower in asthmatics than nonasthmatics, there was no significant association between
the presence of asthma and lung function response to ozone (Lcrov et al.. 2015). These new
studies do not contribute to our understanding of lung function responses to ozone in individuals
with asthma relative to those without.
3.1.5.4.2 Animal Toxicological Studies
Several recent studies provide evidence for ozone exposure-induced respiratory effects in animal
models of allergic airway disease. These effects include sensory and pulmonary irritation and changes in
lung function. Study-specific details are summarized in Table 3-18 and Table 3-19 in Section 3.3.1. All of
these changes, which are described below, were statistically significant. Allergic mice exhibited enhanced
responses compared with naive mice. One study provides insight into mechanisms underlying
ozone-induced bronchoconstriction. Recent studies, detailed below, are grouped according to
concentration-time profile.
• Sensory and pulmonary irritation to acute ozone exposure (2 ppm, 3 hours) were examined in
naive and allergic mice, which were sensitized with ovalbumin. Sensory irritation reflects
changes in the upper airways, while pulmonary irritation reflects changes in the lower airways.
Bao et al. (2013) found increased baseline enhanced pause, with a greater enhancement seen in
allergic mice. Hansen et al. (2016) found sensory irritation in naive and allergic mice and
pulmonary irritation in naive, but not in allergic, mice.
• Schelegle and Walbv (2012) investigated the role of vagal afferents in mediating
bronchoconstriction to acute ozone exposure (1 ppm, 8 hours). Direct measurements of airway
resistance were made in naive and allergic rats (sensitized and challenged with nDer f 1). Ozone
exposure induced rapid shallow breathing in all the rats, but the response was greatest in the
allergic rats. Ozone exposure also increased airway resistance (i.e., bronchoconstriction) in
allergic rats, but not in naive rats. The mechanisms underlying increased airway resistance were
explored using vagotomy and pharmacological agents and were found to involve vagal C-fibers,
vagal myelinated fibers, and possibly mediators released in the airway. Vagal lung C-fibers
mediated the reflex bronchoconstriction to ozone. The vagal myelinated fibers mediate a reflex
bronchodilation. Neuropeptides (e.g., substance P) may also be involved in the
bronchoconstrictive response. This new study provides evidence that ozone exposure exacerbates
bronchoconstriction in allergic animals. Sensory nerve pathways, specifically vagal afferents,
played an important role in increased airway resistance.
3-47
-------
3.1.5.4.3
Epidemiologic Studies
A large body of epidemiologic studies reviewed in the 2013 Ozone ISA (U.S. EPA. 2013a)
provides generally consistent evidence that increases in short-term ozone concentrations are associated
with decreased lung function in children with asthma. Associations were observed across a range of ozone
concentrations, daily averaging times (e.g., 24-hour avg, 8-hour avg, 8-hour max, and 1-hour max), and
diverse geographic locations, including multicity U.S. studies (O'Connor et al.. 2008; Mortimer et al..
2002; Mortimer et al.. 2000). In contrast to studies of lung function in healthy children (Section 3.1.4.1.3)
that generally had reduced potential for exposure measurement error due to the use of ozone monitors at
study sites, studies in children with asthma generally relied on central site monitors. The majority of the
observed ozone-related decrements in FEVi and PEF ranged from <1 to 2%, but results were more
variable for FEVi. Additionally, in studies that observed increases in ozone-related respiratory symptoms
and decreases in lung function, associations were generally reported at similar lags. No recent U.S. or
Canadian studies examined ozone associations with lung function in children with asthma.
In addition to studies of children with asthma, the 2013 Ozone ISA evaluated a limited number of
studies that examined lung function in adults with asthma (U.S. EPA. 2013a). In contrast to results from
studies of children, short-term ozone concentrations were not consistently associated with lung function
decrements in adults with asthma. Differences in exposure assignment techniques, including single
fixed-site monitors, on-site monitoring during outdoor activity, and personal exposure monitoring, did not
appear to explain the inconsistent results. No recent studies examined ozone associations with lung
function in adults with asthma.
3.1.5.4.4 Integrated Summary for Lung Function
Based on studies reviewed in the 1996 and 2006 Ozone AQCDs (U.S. EPA. 2006. 1996a) and the
2013 Ozone ISA (U.S. EPA. 2013a). there is evidence that individuals with asthma were at least as
sensitive to acute effects of ozone on lung function as healthy individuals. Several controlled human
exposure studies evaluated in the 2013 Ozone ISA (U.S. EPA. 2013a) demonstrated that acute ozone
exposures result in lung function decrements in individuals with asthma. However, the majority of these
studies observed similar ozone-induced lung function changes in individuals with and without asthma.
Consistent with prior findings, one recent study showed increasing ozone-induced decrements in lung
function with decreasing baseline lung function. That is, the effect of ozone on lung function increased
with increasing asthma severity. Beyond this, recent studies do little to inform potential differences in
lung function responses to ozone among individuals with and without asthma. While one recent study
observed similar lung function decrements in individuals with and without asthma, most recent studies do
not examine a healthy comparison group. However, despite limited evidence demonstrating increased
sensitivity to ozone in individuals with asthma compared to those without asthma, there is consistent
evidence that asthmatic individuals experience lung function decrements in response to acute ozone
3-48
-------
exposures. A recent animal toxicological study also provides additional evidence of ozone-induced lung
function changes. Changes in ventilator parameters (e.g., breathing frequencies) and increased airway
resistance were more pronounced in allergic rats. Additionally, the recently available study provides
mechanistic evidence that sensory nerve pathways play an important role in reflex bronchoconstriction to
ozone.
Similar to controlled human exposure studies, few epidemiologic panel studies evaluated in the
2013 Ozone ISA (U.S. EPA. 2013a) compare lung function responses with ozone in individuals with and
without asthma. Nonetheless, many studies have reported that increases in ambient ozone concentrations
are associated with decreases in lung function in children with asthma. This association is established
from studies evaluated in the 2013 Ozone ISA, as recent epidemiologic studies in the U.S. or Canada have
not examined ozone associations with lung function in study populations restricted to individuals with
asthma. In contrast, a limited number of studies discussed in the 2013 Ozone ISA did not report consistent
associations between short-term exposure to ozone and lung function decrements in adults with asthma.
3.1.5.5 Airway Responsiveness
3.1.5.5.1 Controlled Human Exposure Studies
As reviewed in the 2016 Oxides ofNitrogen ISA [see Section 5.2.2.1 of U.S. EPA (2016)1.
airway responsiveness is log-normally distributed in the general population, with individuals having
airway hyperresponsiveness tending to be those with asthma. Along with symptoms, variable airway
obstruction, and airway inflammation, airway hyperresponsiveness is a primary feature in the clinical
definition and characterization of asthma severity. Thus, individuals with asthma generally have greater
baseline airway responsiveness than those unaffected by asthma. Similar relative changes (i.e., percent
decrease in provocative dose caused by ozone exposure) in airway responsiveness are seen in subjects
with asthma and healthy control subjects exposed to ozone despite their markedly different baseline
airway responsiveness [see Section 6.2.2.1 of U.S. EPA (2013a)l. There are no new studies evaluating the
effect of ozone exposure on airway responsiveness in individuals with asthma. Increased airway
responsiveness can be an important consequence of exposure to ambient ozone in individuals with asthma
because their airways are potentially predisposed to narrowing on inhalation of a variety of stimuli. An
important aspect of ozone-induced increases in airway responsiveness is that this effect may provide
biological plausibility for associations observed between increases in ambient ozone concentrations and
increased respiratory symptoms in children with asthma and increased hospital admissions and ED visits
for asthma.
3-49
-------
3.1.5.5.2 Animal Toxicological Studies
The 2013 Ozone ISA summarized evidence of increased airway responsiveness in rodent models
of allergic airway disease. Repeated ozone exposure (0.1-0.5 ppm) over 10 days increased nonspecific
airway responsiveness in allergen-sensitized animals. Ozone exposure (1 ppm) increased airway
responsiveness to inhaled allergens in allergen-sensitized animals. A recent study, detailed below, also
found that allergic mice exhibited enhanced airway responsiveness compared with naive mice. This effect
was statistically significant. Sensory nerve pathways, specifically vagal afferents, were found to play an
important role in increased airway responsiveness. Additional study-specific details are summarized in
Table 3-18 in Section 3.3.1.
• Schelegle and Walbv (2012) evaluated the role of vagal afferents in mediating ozone-induced
increased airway responsiveness to allergen. Direct measurements of airway resistance were
made in naive and allergic rats (sensitized and challenged with nDer f 1) following ozone
exposure (1 ppm, 8 hours). Ozone exposure enhanced allergen-induced airway resistance
(i.e., increase in specific airway responsiveness) in allergic rats to a greater degree than in naive
rats. This was an early airway response; no late airway response was observed. The mechanisms
underlying this response were explored using vagotomy and pharmacological agents and
demonstrated the involvement of vagal C-fibers, vagal myelinated fibers, and possibly
neuropeptides released in the airway. Results indicated that vagal lung C-fibers mediated the
enhanced specific airway reactivity (to the allergen). Neuropeptides (e.g., substance P) may also
be involved in the bronchoconstrictive response to allergen.
3.1.5.5.3 Integrated Summary for Airway Responsiveness
Controlled human exposure studies previously evaluated in the 2013 Ozone ISA indicate that
individuals with and without asthma exhibit similar relative increases in ozone-induced airway
responsiveness. However, in general individuals with asthma have greater baseline airway responsiveness
than individuals without asthma. Increased airway responsiveness can result in the narrowing of airways
upon inhalation of a variety of stimuli, providing biological plausibility for ozone-induced asthma
exacerbation, as supported by the epidemiologic associations observed between increases in ozone and
HA and ED visits for asthma and prevalence of respiratory symptoms in children with asthma. No recent
controlled human exposure studies or epidemiologic studies were identified for review.
Consistent with previously reviewed experimental studies in humans, animal toxicological studies
of ozone exposure reviewed in the 2013 Ozone ISA observed increased airway responsiveness to inhaled
allergens in allergen-sensitized animal models. A recent study also found that ozone exposure resulted in
enhanced airway responsiveness in allergic mice compared to naive mice.
3-50
-------
3.1.5.6 Respiratory Tract Inflammation, Injury, and Oxidative Stress
3.1.5.6.1 Controlled Human Exposure Studies
Studies reviewed in Section AX6.9.3 of the 2006 Ozone AQCD (U.S. EPA. 2006) and carried
forward into Section 6.2.3.1 starting on pg. 6-77 of the 2013 Ozone ISA (U.S. EPA. 2013a) showed
greater ozone-induced neutrophilic responses in lavage samples collected at 18 hours post-exposure from
individuals with asthma than without asthma. Specifically, two studies showed that individuals with
asthma exposed to 200 ppb ozone for 4-6 hours with exercise exhibited significantly more neutrophils in
BALF (18 hours post-exposure) than similarly exposed healthy individuals. In another study, when lavage
samples were collected at 6 hours following a 2-hour exposure with exercise to 200 ppb ozone, there were
no observed differences in inflammatory responses between those with and without asthma. However, the
subjects with asthma were on average 5 years older than the healthy subjects in this study, and it is still
not yet known how age affects inflammatory responses. It is also possible that the time course of
neutrophil influx differs between healthy individuals and those with asthma.
Human studies described in the 2013 Ozone ISA (U.S. EPA. 2013a) contribute to the
understanding of mechanisms underlying respiratory effects in individuals with atopy or asthma exposed
to ozone. Indicating allergic skewing of responses, increases in airway eosinophils and the Th2 cytokine
IL-5 were observed in subjects with atopy and mild asthma exposed to 160-400 ppb ozone. In addition,
increased expression of high and low affinity IgE receptors on sputum macrophages, which may enhance
IgE-dependent inflammation, was observed. Studies of subjects with allergic asthma also found increased
expression of TLR4 and CD86. While TLR4 is an activator of innate immunity, CD86 is associated with
Th2 responses. Prior allergen challenge enhanced nasal and airway eosinophilia in subjects with mild
asthma exposed to ozone. Studies indicated that ozone exposure enhances components of allergic
inflammation. In addition, controlled human exposure studies have shown increased airway
responsiveness in subjects with mild allergic asthma and allergic rhinitis exposed to 120-250 ppb ozone.
Study-specific details from recent studies, including exposure concentrations and durations, are
summarized in Table 3-20 and Table 3-21 in Section 3.3.1. Some studies appear in tables only (not in the
text) to indicate that they were considered in this review.
• In a recent study, Arjomandi et al. (2015) exposed healthy adults (7 F, 9 M; 30.8 ± 6.9 years) and
asthmatic adults (6 F, 4 M; 33.5 ± 8.8 years) to 0, 100, and 200 ppb ozone for 4 hours with
intermittent exercise (30-minute intervals of rest and exercise at 20 L/minute per m2 BSA).
Sputum neutrophil and eosinophil concentrations increased significantly with the increasing
ozone concentrations. Eosinophil effects remained significant after adjustment for asthma and
atopy, suggesting the effect may be unrelated to the presence of asthma or atopy.
• Dokic and Traikovska-Dokic (2013) exposed subjects with allergic rhinitis (5 F, 5 M;
27.9 ±2.1 years) to 0 and 400 ppb ozone during and out of grass pollen season. Based on a
greater statistical significance of increases in nasal mucus total protein, albumin, PMNs, and
eosinophils following ozone exposures during pollen season than were observed out of pollen
season, the authors concluded that allergens exaggerate the response to ozone. However, the
3-51
-------
statistical tests the authors used did not support their conclusions: the tests appeared to be relative
to a baseline, were not adjusted to responses following air control, and were not performed across
seasons. The authors did not correctly test for differences across seasons, rather seasonal effects
were inferred based on significant changes during pollen season versus no change out of pollen
season.
• Hernandez et al. (2012) examined inflammatory responses of healthy volunteers (20 F, 14 M;
24.2 ±3.9 years) and atopic individuals with asthma (10 F, 7 M; 24.4 ±5.5 years) exposed to
400 ppb ozone for 2 hours with moderate intermittent exercise. Induced sputum samples were
collected 4 hours after exposure. This study is a continuation (i.e., an additional 15 subjects were
included) of the Hernandez et al. (2010) study discussed in the 2013 Ozone ISA (U.S. EPA.
2013a). Although there was no filtered air control, it is possible to make comparisons between the
healthy and asthmatic subjects. After ozone exposure, the proinflammatory cytokines IL6, IL8,
IL18, and TNF-a, were significantly increased in asthmatic compared to healthy volunteers
despite similar neutrophil and macrophages proportions between groups. The authors suggested
that unlike healthy subjects, those with atopic asthma cannot limit epithelial cell proliferative
responses due to oxidative stress from ozone exposures.
3.1.5.6.2 Animal Toxicological Studies
The 2013 Ozone ISA summarized evidence of increased injury, inflammation, and oxidative
stress following ozone exposure in rodent models of allergic airway disease. Repeated ozone exposure
(0.1-0.5 ppm) over 10 days enhanced goblet cell metaplasia in allergen-sensitized animals. In addition,
ozone exposure (1 ppm for 2 days) enhanced inflammation and allergic responses to allergen challenge in
allergen-sensitized animals. Further, treatment with the antioxidant y tocopherol (but not a tocopherol)
blunted ozone-induced inflammation in allergic rhinosinusitis and allergic inflammation of the lower
airways, indicating a role for oxidative stress mediating these effects. Recent studies, detailed below and
grouped by concentration-exposure profile, provide additional evidence for ozone exposure-induced
respiratory effects in animal models of allergic airway disease. This includes injury, inflammation, and
increased mucin/mucosubstance content. Allergic mice showed enhanced responses compared with naive
mice for some of these endpoints. These changes, which are described below, were statistically
significant. Study-specific details are summarized in Table 3-22. Table 3-23. and Table 3-24 in
Section 3.3.1.
• Allergic and inflammatory responses to acute ozone exposure (2 ppm, 3 hours) were evaluated in
naive and allergic mice, which were sensitized with ovalbumin. Hansen et al. (2016) found no
enhancement of allergic responses such as serum IgE, bronchoalveolar cells, and lung tissue
cytokines. Bao et al. (2013) found that allergic mice exhibited greater inflammatory responses to
ozone compared with naive mice, including enhancement of neutrophils, hyaluronan, and the Th2
cytokines 11-5 and IL-13 in BALF. Stored mucosubstance content and MUC5AC gene expression
were also enhanced to a greater degree in allergic mice exposed to ozone.
• Schclcglc and Walbv (2012) found increased BALF protein, an injury marker, but no increase in
BALF cells in allergic rats (sensitized and challenged with nDer f 1) following ozone exposure
(1 ppm, 8 hours).
3-52
-------
3.1.5.6.3
Epidemiologic Studies
The 2013 Ozone ISA described generally consistent epidemiologic evidence of an association
between short-term ozone exposure and subclinical effects in children with asthma (U.S. EPA. 2013a).
The most commonly studied respiratory biomarker was exhaled nitric oxide (eNO), an indicator of
pulmonary inflammation. A link between eNO and asthma exacerbation is well supported in the literature
(Jones et al.. 2001; Kharitonov and Barnes. 2000). Increases in 8-hour daily max ozone concentrations
were associated with increased eNO in a CHS study in southern California (Berhane et al.. 2011) and a
single-city panel study conducted in Mexico City that assigned exposure from monitors within 5 km of
children's homes or schools (Barraza-Villarrcal et al.. 2008). Ozone was also associated with other
subclinical markers of pulmonary inflammation and oxidative stress in children with asthma across a
number of other single-city panel studies. Biomarkers examined included IL-6, IL-8, eosinophils,
TBARS, 8-isoprostane, and malondialdehyde.
One recent epidemiologic study of ozone exposure examined subclinical effects in children with
asthma. In a panel study in southern California, 8-hour max ozone concentrations measured at fixed-site
monitors within 12 km of subjects' residences were not associated with increases in exhaled nitric oxide
[eNO; Delfino et al. (2013)1. See Table 3-25 in Section 3.3.1 for complete study details.
Recent studies have examined pulmonary inflammation in general population studies of children
(Patel et al.. 2013; Salam et al.. 2012). with asthma prevalence ranging from 14 to 47%. Because these
studies do not directly inform the understanding of the relationship between short-term ozone exposure
and pulmonary inflammation in populations with asthma, they are not discussed here. However, study
specific details can be found in Table 3-7 in Section 3.3.1.
3.1.5.6.4 Integrated Summary for Respiratory Tract Inflammation, Injury, and Oxidative
Stress
Controlled human exposure studies evaluated in the 2013 Ozone ISA (U.S. EPA. 2013a)
established evidence of enhanced allergic inflammation to ozone in individuals with asthma. Specifically,
markers of airway and lung inflammation, and innate immunity, were increased in response to short-term
ozone exposures. As with the findings for lung function, there is limited evidence that ozone-induced
inflammatory responses differ due to the presence of asthma. Results from experimental animal studies
are coherent with evidence from humans. Recent studies expand on findings summarized in the 2013
Ozone ISA (U.S. EPA. 2013a). indicating inflammation, oxidative stress, injury, allergic skewing, goblet
cell metaplasia, and upregulation of mucus synthesis and storage in allergic animals exposed to ozone.
Allergic mice generally exhibited enhanced responses compared to naive mice for these endpoints.
Epidemiologic studies evaluated in the 2013 Ozone ISA (U.S. EPA. 2013a) support findings from
experimental studies. Short-term ozone concentrations were associated with markers of oxidative stress
and pulmonary inflammation in panel studies of children with asthma. While the only recent study
3-53
-------
available for review reported a null association between ozone and FeNO (fraction exhaled nitric oxide)
in children with asthma, the results should be considered in the context of previously reviewed studies,
the majority of which observed positive associations.
3.1.5.7 Overall Summary of Respiratory Effects in Populations with Asthma
In summary, evidence from recent epidemiologic and experimental studies continues to support
an association between ozone and asthma exacerbation. Recent, large multicity epidemiologic studies
conducted in the U.S. build on evidence from the 2013 Ozone ISA and provide further support for an
association between ozone and ED visits and hospital admissions for asthma. Hospital admission and ED
visit studies that presented age-stratified results generally reported the strongest associations in children
between the ages of 5 and 18. Additionally, associations were observed across a range of ozone
concentrations, and were consistent in models with measured or modeled concentrations. A limited
number of recent epidemiologic studies in the U.S. or Canada have examined respiratory symptoms,
medication use, lung function, and subclinical effects in people with asthma. However, a large body of
evidence from the 2013 Ozone ISA (U.S. EPA. 2013a) demonstrates ozone associations with these less
severe indicators of asthma exacerbation, providing support for the ozone-related increases in asthma
hospital admissions and ED visits observed in recent studies.
Evidence from controlled human exposure and animal toxicological studies provide biological
plausibility for the associations observed in epidemiologic studies of short-term ozone exposure and
asthma exacerbation. Results from experimental studies in humans demonstrate that ozone exposures lead
to increased respiratory symptoms, lung function decrements, increased airway responsiveness, and
increased lung inflammation in individuals with asthma. However, observed responses across the range of
endpoints did not generally differ due to the presence of asthma. Animal toxicological studies similarly
found that ozone exposures altered ventilatory parameters, increased airway responsiveness, and
increased pulmonary inflammation and bronchoconstriction in allergic animals. In contrast to controlled
human exposure studies, there was some evidence from studies of rodents that the observed respiratory
effects were enhanced in allergic animals compared to naive animals.
3.1.6 Respiratory Effects in Other Populations with Pre-existing Conditions
3.1.6.1 Exacerbation of Chronic Obstructive Pulmonary Disease and Associated
Respiratory Effects in Populations with COPD
Chronic obstructive pulmonary disease (COPD) is a chronic lung disorder characterized by
destruction of alveolar tissue, airway remodeling, and minimally reversible airflow limitation (Vcstbo et
3-54
-------
al.. 2013). Reduced airflow is associated with decreased lung function, and clinical symptoms
demonstrating exacerbation of COPD include cough, sputum production, and shortness of breath (Vestbo
et al.. 2013). Severe exacerbation can lead to ED visits or hospital admissions.
3.1.6.1.1 Epidemiologic Studies
A limited number of epidemiologic studies evaluated in the 2013 Ozone ISA provided some
evidence that short-term exposure to ozone is associated with increased ED visits for COPD (U.S. EPA.
2013a). A large multicity study in Canada reported a year-round association that was driven largely by
ED visits in the warm season (Stieb et al.. 2009). In a single-city study in Sao Paulo, Brazil, Arbex et al.
(2009) also observed a positive association, but only in a stratified analysis of women. There was little
supporting evidence from studies of lung function or respiratory symptoms in adults with COPD.
Specifically, epidemiologic studies did not provide strong evidence that short-term increases in ozone
exposure result in lung function decrements in adults with COPD, and a single study of adults with COPD
found that ozone was both positively and inversely associated with a range of respiratory symptoms
(Peacock et al.. 2011).
Recent studies provide generally consistent evidence that short-term exposure to ozone is
associated with ED visits for COPD, with the strongest evidence coming from large multicity studies.
Supporting evidence from less severe manifestations of COPD is still lacking. The majority of the
evaluated studies use an 8-hour daily max averaging time, although there are instances in which the
1-hour daily max (Malig et al.. 2016) and the 24-hour avg (Szvszkowicz et al.. 2018) are used. The
averaging time used in each study, along with other study-specific details, including air quality
characteristics and select effect estimates, are highlighted in Table 3-26 and Table 3-27 in Section 3.3.1.
An overview of the evidence is provided below.
• Large case-crossover studies, including a statewide study in California (Malig et al.. 2016) and a
multicity study in Ontario, Canada (Szvszkowicz et al.. 2018). reported positive associations
between ozone and COPD ED visits. Malig et al. (2016) noted stronger associations in the warm
season than in year-round analysis, and associations that were robust to adjustment for NO2, SO2,
and CO in copollutant models. Further discussion of potential copollutant confounding and the
role of season and temperature on ozone associations with respiratory health effects can be found
in the Relevant Issues for Interpreting Epidemiologic Evidence section (Section 3.1.10). In an
effort to reduce potential exposure misclassification, both studies assigned ozone exposure using
the nearest monitor or the average of the nearest monitors within maximum distance buffers
(Szvszkowicz et al.. 2018; Malig et al.. 2016). A large time-series study in five U.S. cities also
observed ozone-related increases in COPD ED visits in three of the five cities (Barry et al.. 2018).
• Another large case-crossover study in the state of Georgia observed increased ED visits for
chronic bronchitis, a condition that can contribute to or occur independently of COPD,
corresponding to increases in fused-CMAQ and ground-based ozone concentrations (Xiao et al..
2016).
3-55
-------
• Notably smaller studies in Little Rock, AR (Rodopoulou et al.. 2015) and St. Louis, MO (Sarnat
et al.. 2015) examined ozone-related ED visits for COPD and reported a positive, but imprecise
association (i.e., wide 95% CIs) and a null association, respectively. Each study used one monitor
for the entire study area, which may have introduced exposure measurement error. Additionally,
the short length of the time-series (Sarnat et al.. 2015) and the small mean number of daily
hospital admissions (Rodopoulou et al.. 2015) likely reduced the statistical power to detect an
association.
• COPD exacerbation measured by frequency of short-term bronchodilator inhaler use was not
associated with short-term exposure to ozone (Magzamcn et al.. 2018). See Table 3-26 in
Section 3.3.1 for complete study details.
3.1.6.1.2 Animal Toxicological Studies
No animal toxicological studies evaluating respiratory effects in animal models of COPD were
described in the 2013 Ozone ISA (U.S. EPA. 2013a). Two recent studies employed an animal model of
progressive pulmonary inflammation as a surrogate for COPD. This model involves deficiency of
surfactant protein D (sfpd), a collectin protein synthesized by lung type 2 cells. Results suggest that
chronic inflammation enhanced sensitivity to short-term ozone exposure. This effect was statistically
significant. Study-specific details are summarized in Table 3-28. Table 3-29. and Table 3-30 in
Section 3.3.1.
• Groves et al. (2012) found that acute ozone exposure (0.8 ppm for 3 hours) leads to increased
indicators of injury, inflammation, oxidative stress in sfpd-deficient mice that do not resolve by
72 hours. In contrast, resolution of responses occurred by 72 hours in sfpd-sufficient mice. Ozone
exposure resulted in altered lung mechanics that is indicative of central airway and peripheral
tissue involvement in sfpd-deficient mice. In sfpd-sufficient mice, ozone exposure resulted in
altered lung mechanics that is indicative of only central airway involvement. These
determinations were made by analysis of resistance and elastance spectra obtained from
impedance data. In a second study, Groves et al. (2013) found age-related increases in enlarged
vacuolated macrophages, alveolar wall rupture, type 2 hyperplasia, BALF protein and cell
number, and changes in lung mechanics consistent with COPD are observed in sfpd-deficient
mice. Acute ozone exposure (0.8 ppm, 3 hours) resulted in greater alveolar hyperplasia and
classically activated macrophages in sfpd-deficient than in sfpd-sufficient mice. The effects of
ozone on lung mechanics were dampened in 27-week-old mice compared with 8-week-old mice.
3.1.6.1.3 Integrated Summary for Chronic Obstructive Pulmonary Disease (COPD)
In summary, recent large multicity epidemiologic studies of ED visits support an association
between short-term ozone exposure and COPD exacerbation. Associations are reported across a variety of
study locations, exposure levels, and exposure assignment methods, including nearest monitor
concentrations and CMAQ-fused models. In limited copollutant results, the observed association is robust
to adjustment for other gaseous pollutants (NO2, SO2, and CO). While none of the experimental animal
studies evaluated in the 2013 Ozone ISA examined acute ozone exposure in animals with chronic
3-56
-------
inflammation, results from recent studies suggest that chronic inflammation enhances sensitivity to ozone
exposure, providing coherence with ozone-related COPD exacerbation observed in epidemiologic studies.
3.1.6.2 Obese Populations or Populations with Metabolic Syndrome
Metabolic syndrome is a term used to describe a collection of risk factors that include high blood
pressure, dyslipidemia (elevated triglycerides and low levels of high-density lipoprotein [HDL]
cholesterol), obesity (particularly abdominal obesity), and increased fasting blood glucose
[hyperglycemia; Alberti et al. (2009)1. There is growing evidence that components of metabolic
syndrome, including obesity, may increase susceptibility to air pollution-related health effects (Jiu-Chiuan
and Schwartz. 2008). The following section evaluates studies examining the relationship between
short-term exposure to ozone and respiratory health effects in obese populations or populations with
metabolic syndrome. The scientific evidence that supports causality determinations for short-term ozone
exposure and metabolic effects is assessed in Appendix 5.
3.1.6.2.1 Lung Function
3.1.6.2.1.1 Controlled Human Exposure Studies
In the 2013 Ozone ISA (U.S. EPA. 2013a). two retrospective analyses of controlled human
exposure studies showed ozone-induced FEVi decrements increased with increasing BMI. Since the 2013
Ozone ISA, there is a new controlled human exposure study and a larger retrospective analysis
demonstrating an effect of BMI on lung function responses to ozone. Study-specific details, including
exposure concentrations and durations, are summarized in Table 3-4 and Table 3-31 in Section 3.3.1.
• Bennett et al. (2016) exposed obese (19 F; 27.7 ±5.2 years) and normal-weight (19 F;
24 ± 3.7 years) women to 0 and 400 ppb ozone for 2 hours during intermittent exercise
(15-minute periods of seated rest and exercise at 25 L/minute per m2 BSA). The ozone-induced
FVC decrement was significantly (p < 0.05) greater in the obese women (12.5%) than
normal-weight women (8.0%). The FVC decrement also tended (p = 0.08) to be greatest in the
obese African-Americans (15.7%) relative to other obese subjects (9.6%). There was also a
tendency (p = 0.11) for greater ozone-induced FEVi decrements in obese women (15.9%) relative
to the normal-weight women (11.7%). While respiratory function was diminished, respiratory
symptoms in response to ozone exposure did not differ between obese and normal-weight
women.
• The new retrospective analysis by McDonnell et al. (2013) includes data from prior studies of
young healthy adults (104 F, 637 M; 18-36 years) exposed one or more times to ozone and/or
filtered air. The prior analysis by McDonnell et al. (2010). discussed in the 2013 Ozone ISA in
relation to BMI effects, used data from 541 healthy nonsmoking white males (18-35 years). The
analysis based on a larger data set continues to show that the BMI effect is of the same order of
3-57
-------
magnitude but in the opposite direction of the age effect. Thus, the model predicts FEVi
responses increase with increasing BMI and diminish with increasing age.
3.1.6.2.1.2 Animal Toxicological Studies
No studies evaluating the effects of ozone exposure on lung function in obese animals or animal
models of metabolic syndrome were available in the 2013 Ozone ISA (U.S. EPA. 2013a). A recent study
(Gordon et al.. 2016b) involved male and female rats fed normal, high-fructose or high-fat diets prior to
acute and subacute ozone exposure (0.8 ppm x 5 hours). While there were some differences in effects
depending on duration of exposure, diet, and sex of rat, ozone exposure generally resulted in statistically
significant increases in enhanced pause and tidal volume, which are ventilatory parameters that reflect a
change in lung function. Study-specific details are summarized in Table 3-32 in Section 3.3.1. Findings
related to behavior and metabolism are found elsewhere in Appendix 7 and Appendix 5. respectively.
3.1.6.2.1.3 Epidemiologic Studies
In a study evaluated in the 2013 Ozone ISA (U.S. EPA. 2013a). short-term ozone concentrations
were associated with decreases in lung function in older adults with airway hyperresponsiveness
(Alexeeff et al.. 2007). The observed association was stronger among those who were obese. A recent
analysis of the Offspring and Third Generation Framingham Heart Study cohorts also found that obese
participants had significantly stronger associations between 8-hour max summertime ozone
concentrations and reduced lung function (Rice et al.. 2013). Study specific details, including effect
estimates, are summarized in Table 3-7 in Section 3.3.1.
3.1.6.2.2 Airway Responsiveness
3.1.6.2.2.1 Controlled Human Exposure Studies
No controlled human exposure studies were available for review in the 2013 Ozone ISA that
examined airway responsiveness in obese individuals or individuals with metabolic syndrome (U.S. EPA.
2013a).
• A recent study showed an increase in airway responsiveness after ozone exposure did not differ
between normal-weight and obese women (Bennett et al.. 2016). Study-specific details, including
exposure concentrations and durations, are summarized in Table 3-31 in Section 3.3.1.
3-58
-------
3.1.6.2.2.2
Animal Toxicological Studies
The 2013 Ozone ISA summarized the evidence of respiratory effects in obese animals resulting
from exposure to ozone (U.S. EPA. 2013a). In mouse models of obesity, airways were innately more
responsive and responded more vigorously to acute ozone exposure (2 ppm for 3 hours) than lean
controls. Newly available information confirms and extends these findings.
Several recent studies evaluated the respiratory effects of acute ozone exposure (2 ppm, 3 hours)
in mouse models of obesity (Mathews et al.. 2018; Mathews et al.. 2017a; Mathews et al.. 2017b;
Williams et al.. 2015). These studies compared responses in obese mice with those of lean mice. Changes
in airway responsiveness described below were statistically significant. Study-specific details are
summarized in Table 3-32 in Section 3.3.1.
• Pulmonary mechanics were assessed by using the flexiVent system. Baseline and nonspecific
(i.e., methacholine challenge) airway responsiveness were greater in obese mice than lean mice in
the absence of ozone exposure. Acute ozone exposure increased baseline and nonspecific airway
responsiveness in obese mice, but not in lean mice.
• Williams et al. (2015) probed the role of TNF-a and TNF-a receptor in the augmented responses
to ozone exposure in obese mice and found that deficiency in these factors enhanced the increase
in airway responsiveness.
3.1.6.2.3 Respiratory Tract Inflammation, Injury, and Oxidative Stress
3.1.6.2.3.1 Controlled Human Exposure Studies
No controlled human exposure studies were available for review in the 2013 Ozone ISA that
examined pulmonary inflammation, injury, or oxidative stress in obese individuals or individuals with
metabolic syndrome (U.S. EPA. 2013a). Study-specific details from recent studies, including exposure
concentrations and durations, are summarized in Table 3-21 and Table 3-31 in Section 3.3.1.
• Bennett et al. (2016) recently investigated PMN responses in obese (19 F; 27.7 ±5.2 years) and
normal-weight (19F;24±3.7 years) women exposed to 0 and 400 ppb for 2 hours during
intermittent exercise. Although PMN were significantly increased after ozone exposure relative to
air, the PMN response did not differ between groups.
• In their study of healthy adults (7 F, 9 M; 30.8 ± 6.9 years) and asthmatic adults (6 F, 4 M;
33.5 ± 8.8 years), Arjomandi et al. (2015) also found that adjustment for age, sex, and BMI did
not affect the association between PMN responses and ozone exposure.
3.1.6.2.3.2 Animal Toxicological Studies
The 2013 Ozone ISA summarized the evidence of respiratory effects in obese animals resulting
from exposure to ozone (U.S. EPA. 2013a). In mouse models of obesity, respiratory tract inflammation
3-59
-------
and injury responses to acute ozone exposure (2 ppm for 3 hours) were enhanced compared with lean
controls. However, the inflammatory response to subacute ozone exposure (0.3 ppm for 72 hours) was
dampened. Several recent studies have evaluated the respiratory effects of ozone exposure in animal
models of obesity, high fructose/fat diet, and diabetes. Enhanced inflammatory and injury responses were
found in obese compared with lean mice and in animals fed high-fat/high-fructose diets compared with
those fed a normal diet. These effects, which are described below, were statistically significant.
Study-specific details are summarized in Table 3-33 and Table 3-34 in Section 3.3.1.
• Four studies of acute exposure to ozone (2 ppm, 3 hours) were conducted in mouse models of
obesity (Mathews et al.. 2018; Mathews et al.. 2017a; Mathews et al.. 2017b; Williams et al..
2015). These studies compared responses in obese mice with those of lean mice. Taken together,
these studies shed new light on mechanisms underlying the augmentation of ozone-exposure-
induced effects in animal models of obesity. Study-specific details are summarized in Table 3-38
in Section 3.3.1. Acute ozone exposure increased BALF markers of injury and inflammation to a
greater extent in obese than in lean mice. Williams et al. (2015) probed the role of TNF-a and
TNF-a receptor in the augmented responses to ozone exposure in obese mice and found that
deficiency in these proteins attenuated the inflammatory effect. Mathews et al. (2017b) provided
evidence that IL-33 contributes to the augmented responses to ozone exposure in obese mice by
acting on immune lymphoid cells 2 (ILC2) and on gamma delta T cells, which express the Th2
cytokines IL-5 and IL-13. Mathews et al. (2018) showed a role for IL-17A and gastrin-releasing
peptide in the augmented responses to ozone exposure in obese mice. Mathews et al. (2018) noted
differences in metabolism, antioxidants, and microbiome in obese and lean mice exposed to
ozone. Levels of corticosterone were increased by ozone exposure in obese mice.
• Two studies of subacute ozone exposure (0.5 for 4 hour/day for 13 days) were conducted in a
diabetes-prone mouse model (Ying et al.. 2016; Zhong et al.. 2016). Ozone exposure increased
BALF inflammatory cells and upregulated proinflammatory genes in lung tissue. However, no
change in the T-cell profiles was found in the pulmonary lymph nodes. Study-specific details are
summarized in Table 3-40 in Section 3.3.1. Findings related to systemic inflammation and insulin
resistance are reported in Appendix 5.
• Gordon et al. (2016b) fed male and female rats normal, high-fructose, or high-fat diets prior to
acute and subacute ozone exposure (0.8 ppm x 5 hours). While there were some differences in
effects depending on duration of exposure, diet, and sex of rat, in general ozone exposure resulted
in increased eosinophils and albumin (a marker of injury) in BALF. The high-fructose and
high-fat diets did not enhance the effects of ozone on inflammatory or injury-related markers.
Findings related to metabolism and behavior are found in Appendix 5 and Appendix 7.
respectively.
3.1.6.2.3.3 Epidemiologic Studies
No epidemiologic studies in the 2013 Ozone ISA examined potential associations between
short-term exposure to ozone and respiratory health effects in people with pre-existing metabolic
syndrome (U.S. EPA. 2013a). A recent panel study of adults with type 2 diabetes mellitus reported
decreases in pulmonary inflammation corresponding to 6- (3:00 a.m. to 9:00 a.m.) and 24-hour avg ozone
concentrations prior to FeNO measurement (Peng et al.. 2016). The apparent protective association may
be explained by negative correlations between ozone and NOx, black carbon (BC), and particle number
3-60
-------
(PN), each of which demonstrated strong positive associations with pulmonary inflammation.
Study-specific details, including air quality characteristics and select effect estimates, are highlighted in
Table 3-35 in Section 3.3.1.
3.1.6.2.4 Overall Summary for Respiratory Effects in Obese Populations or Populations
with Metabolic Syndrome
A recent controlled human exposure study reported evidence of ozone-related increases in
pulmonary inflammation in both obese and normal weight adult women during exercise, but
inflammatory responses did not differ between the groups. In contrast, epidemiologic studies provide
some evidence that ozone-related lung function decrements are larger in obese individuals. Similarly,
recent animal toxicological studies expand the body of evidence evaluated in the 2013 Ozone ISA and
continue to indicate that, compared to lean mice, obese mice exhibit enhanced airway responsiveness and
pulmonary inflammation in response to acute ozone exposures.
In studies of a diabetes-prone mouse model, subacute ozone exposure increased airway
inflammation and proinflammatory genes in lung tissue. In contrast, an epidemiologic panel study
observed a protective association between ozone and pulmonary inflammation in adults with type 2
diabetes mellitus. This inverse association may be explained by negative correlations with copollutants
that demonstrated strong positive associations with pulmonary inflammation in the same population.
In summary, experimental animal studies provide evidence for enhanced respiratory tract
inflammation in obese and diabetic models, but evidence from a limited number of controlled human
exposure and epidemiologic studies do not demonstrate coherence.
3.1.6.3 Populations with Pre-existing Cardiovascular Disease
3.1.6.3.1 Animal Toxicological Studies
No animal toxicological studies evaluating respiratory effects in populations with cardiovascular
disease were described in the 2013 Ozone ISA (U.S. EPA. 2013a). Several recent studies evaluated
respiratory effects of acute ozone exposure (0.2-1 ppm, 3-6 hours) in rodent models of cardiovascular
disease. Some of the studies provide evidence that cardiovascular disease exacerbates the respiratory
effects of ozone exposure. Injury, inflammation, oxidative stress, lung function changes, and increased
airway responsiveness were seen in animals with cardiovascular disease in response to ozone exposure.
Acute ozone exposure in animal models of hypertension resulted in enhanced injury, inflammation, and
airway responsiveness compared with healthy animals. These effects, which are described below, were
statistically significant. Study-specific details are summarized in Table 3-36. Table 3-37. and Table 3-38
in Section 3.3.1.
3-61
-------
• A group of investigators from the same institution examined the effects of a 4-hour ozone
exposure in spontaneously hypertensive (SH), fawn-hooded hypertensive (FHH), stroke-prone
spontaneously hypertensive (SPSH), obese spontaneously hypertensive heart failure (SHHF), and
obese atherosclerosis prone JCR rats across a range of ozone exposure concentrations
[0.25-1 ppm ozone; Dye et al. (2015); Hatch et al. (2015); Kodavanti et al. (2015); Ramot et al.
(2015); Ward and Kodavanti (2015); Farrai et al. (2012)1. Histopathological lesions and
indicators of inflammation and injury were seen in all strains at 1 ppm, but rats with
cardiovascular disease were less sensitive than healthy rats to the lowest concentration tested
(0.25 ppm). Decreases in lung antioxidants were seen only in response to 1 ppm ozone. Some of
the rats with cardiovascular disease exposed to 0.25 ppm exhibited changes in ventilatory
parameters, while all of the strains were responsive to 1.0 ppm ozone. Another study from this
same group of investigators (Farrai et al.. 2016) did not find any increased indicators of
inflammation following a 3-hour exposure to 0.3 ppm ozone in SH rats.
• Histopathologic responses following a 6-hour exposure to 1 ppm ozone were evaluated in healthy
(Wistar Kyoto) and SH rats (Wong et al.. 2018). Ozone exposure induced lesions in terminal
bronchioles and alveolar ducts in both strains, with lesions also seen in large airways of the SH
rats. In addition, lesion scores were higher in the SH rats than healthy rats for edema, PMN
infiltrate, terminal bronchiole epithelial necrosis, exudate, and large airway cilia cell
loss/necrosis. In contrast, Ramot et al. (2015) observed similar histopathologic lesions in SH and
Wistar Kyoto rats following a 4-hour exposure to 1 ppm ozone. This apparent discrepancy may
be attributed to differences between the age and disease status of the animals studied.
Specifically, Wong et al. (2018) examined mature adult (-48 week old) SH rats that showed fully
developed cardiovascular disease while Ramot et al. (2015) evaluated young (12- to 14-week-old)
SH rats that were just beginning to develop hypertension.
• Respiratory effects of 4-hour exposure to 1 ppm ozone were evaluated in a mouse model of
pulmonary hypertension that had been induced using exposure to hypoxia (Zvchowski et al..
2016). Ozone exposure increased lung weight and lung water weight in mice with pulmonary
hypertension but not in control mice. Mice with pulmonary hypertension exhibited larger
increases in BALF cells and airway responsiveness to methacholine (measured in terms of airway
resistance) than control mice in response to ozone exposure.
3.1.7 Respiratory Infection and other Associated Health Effects
3.1.7.1 Epidemiologic Studies
A single study evaluated in the 2013 Ozone ISA (U.S. EPA. 2013a) examined the association
between ozone exposure and respiratory infection ED visits. Stieb et al. (2009) observed no evidence of
an association between ozone exposure and respiratory infection ED visits at any lag examined (i.e., 0, 1,
or 2 days) in an all-year analysis across seven Canadian cities. Several recent studies that have become
available since the 2013 Ozone ISA provide generally consistent evidence of an association between
short-term exposure to ozone and ED visits for a range of respiratory infection endpoints (Figure 3-6).
Generally, the evaluated studies use an 8-hour daily max averaging time, although there are instances in
which the 1-hour daily max (Malig et al.. 2016) and the 24-hour avg (Szvszkowicz et al.. 2018) are used.
3-62
-------
The averaging times used in each study, along with other study-specific details, including air quality
characteristics and select effect estimates, are highlighted in Table 3-39 in Section 3.3.1. The recently
available multicity and single-city studies provide evidence of associations despite the implementation of
various study designs, exposure assessment methods, and ozone averaging times. An overview of the
evidence is provided below.
• Recent large multicity studies in the U.S. and Canada reported associations between ozone and
ED visits for pneumonia (Malig et al.. 2016; Xiao et al.. 2016). acute respiratory infections
(Malig et al.. 2016). upper respiratory tract infections (Barry et al.. 2018; Szvszkowicz et al..
2018; Malig et al.. 2016; Xiao et al.. 2016). and ear infections (Xiao et al.. 2016). Increases in ED
visits ranged from about 2 to 6% per standardized increase in 24-hour avg (Szvszkowicz et al..
2018). 8-hour max (Barry et al.. 2018; Xiao et al.. 2016). and 1-hour max (Malig et al.. 2016)
ozone concentrations.
• Large single-city studies in Atlanta (Darrow et al.. 2014). Edmonton (Kousha and Rowe. 2014).
and St. Louis (Sarnat et al.. 2015; Winquist et al.. 2012) also provided generally consistent
evidence that ozone is associated with increases in ED visits for pneumonia (Sarnat et al.. 2015;
Darrow et al.. 2014). upper respiratory tract infection (Darrow et al.. 2014). and acute bronchitis
(Kousha and Rowe. 2014). In contrast to results from Sarnat et al. (2015). another time-series
study in St. Louis did not observe an association between ozone and ED visits for pneumonia
(Winquist et al.. 2012). The study periods overlapped, but Winquist et al. (2012) considered a
longer time frame. Each study assigned exposure using one monitor for the entire study area,
which may have introduced exposure measurement error.
• Notably smaller studies in Windsor, Canada (Kousha and Castner. 2016) and Little Rock, AR
(Rodopoulou et al.. 2015) did not observe associations between ozone and ED visits for acute
respiratory infections, pneumonia, or ear infections. The observed effect estimates were imprecise
(i.e., wide 95% CIs), likely due to the limited sample sizes.
• One recent multicity study that examined ED visits for respiratory infection evaluated copollutant
confounding (Malig et al.. 2016). These results are discussed in more detail in the Relevant Issues
for Interpreting Epidemiologic Evidence Section (Section 3.1.10).
3-63
-------
Study
Location
Ages
Mean
Concentration3
Season
Lag
t
t
Combined Respiratory Infection
1
l
Stieb et al. (2009)
Multicity, Canada
All Ages
10.3-22.1; 24-hr avg
Year-Round
1
l
Acute Respiratory Infection
t
1
fMalig et al. (2016)
California
All
33-55; 1-hr max
Year-Round
Warm
0-1
1
!•
|#
Upper Respiratory Infection
1
1
tDarrow et al. (2014)
Atlanta, GA
0-4
45.9
Year-Round
Lag 0-2
f Barry et al. (2018)
Atlanta, GA
Birmingham, AL
Dallas, TX
Pittsburgh, PA
St. Louis, MO
All
37.5-42 2
Year-Round
0-2
I#-
h®-
r*-
+-
fMalig et al. (2016)
California
All
33-55; 1-hr ma*
Year-Round
Warm
0-1
fXiao et al. (2016)
Georgia
2-18
42.1
Year-Round
0-3
\ •
fSzyszkcwicz et al. (2018)
Multicity, Canada
<20; Females
<20; Males
22.5-29 2; 24-hr avg
Year-Round
Year-Round
1
, <*-
«•-
Lower Resoiratory Infection
!
fKousha et al. (2014)
Edmonton, Canada
All
18.6
Year-Round
0
i
fSzyszkowicz et al (2018)
Multicity, Canada
<20; Females
<20; Males
22.5-29.2; 24-hr avg
Year-Round
Year-Round
1
r®-
Pneumonia
i
i
fWinquist et al. (.2012)
St. Louis, MO
All
NR
Year-Round
0-4 DL
tDarrow et al. (2014)
Atlanta, GA
0-4
45.9
Year-Round
Lag 0-2
fMalig et al. (2016)
California
All
33-55; 1-hr ma*
Year-Round
Warm
0-1
b
U-
fXiao et al. (2016)
Georgia
2-18
42.1
Year-Round
0-3
1
Bronchitis
1
1
fDarrow etal. (2014)
Atlanta, GA
0-4
45.9
Year-Round
Lag 0-2
,—1
0.95 1 1.05
Effect Estimate (95% CI)
DL = distributed lag.
Note: jStudies published since the 2013 Ozone ISA. Black text = studies included in the 2013 Ozone ISA.
aMean concentrations reported in ppb and are for an 8-hour daily max averaging time unless otherwise noted.
Results standardized to a 15-ppb increase in 24-hour avg, 20-ppb increase in 8-hour daily max, or 25-ppb increase in 1-hour daily
max ozone concentrations. Corresponding quantitative results are reported in Table 3-38.
Figure 3-6 Summary of associations from studies of short-term ozone
exposures and respiratory infection emergency department (ED)
visits for a standardized increase in ozone concentrations.
3-64
-------
3.1.7.2
Controlled Human Exposure
The inflammatory effects of ozone involve the innate immune system, as indicated by increases
in airway neutrophils. The adaptive immune system may also be involved via alterations in antigen
presentation and co-stimulation by innate immune cells such as macrophages and dendritic cells, which
may lead to T-cell activation. Controlled human exposure studies described in Section 6.2.5 of the 2013
Ozone ISA (U.S. EPA. 2013a) show that ozone exposure results in airway neutrophilia, reflecting
activation of the innate immune system, and altered antigen presentation in macrophages and dendritic
cells. Subjects involved in these studies were exposed to 80-400 ppb ozone with moderate intermittent
exercise. Enhanced adaptive immunity may bolster defenses against infection, as well as increase allergic
responses via T-cell activation. Other controlled human exposure studies showed minimal effects of
ozone exposure on macrophage phagocytosis or function. In asthmatics, there is increased uptake of
particles by airway macrophages that may also enhance the processing of particulate antigens and lead to
greater progression of allergic airway disease and contribute to an increased risk of asthma exacerbation.
There are no new controlled human exposure studies contributing to this evidence base.
3.1.7.3 Animal Toxicological Studies
The 2013 Ozone ISA (U.S. EPA. 2013a) summarized the animal toxicological evidence of
impaired host defense resulting from exposure to ozone. Increased susceptibility to challenge with
infectious agents was observed at ozone concentrations of 0.08-0.5 ppm. Decreases in mucociliary
clearance occurred following exposure to 1 ppm ozone and altered macrophage phagocytosis or function
following exposure to 0.1 ppm ozone. In addition, effects on adaptive immunity, such as altered T cell
subsets in the spleen (0.6 ppm), decreased antibody response following influenza virus infection
(0.5 ppm), and decreased mitogen activated T-cell proliferation (0.5 ppm), have been reported. Effects on
natural killer cells, which are effectors of innate and adaptive immunity, have also been reported with
decreased activity at concentrations of 0.6-1 ppm, and increased activity or no effect at lower
concentrations. Acute exposures to 2 ppm ozone resulted in SP-A oxidation and impairment of SP-A
dependent phagocytosis, which led to increased susceptibility to pneumonia.
Two recent studies provided evidence that acute ozone exposure (2 ppm, 3 hours) increased
susceptibility to infectious disease. Effects, described below, were statistically significant. These studies
build upon the investigators' previous work showing that the survival rate of mice infected with
pneumonia was decreased by previous exposure to ozone (2 ppm, 3 hours). Study-specific details are
summarized in Table 3-40 in Section 3.3.1.
• In one study Durrani et al. (2012) found that ozone exposure had different impacts on survival in
male and female mice. To investigate sex-related differences in survival, mice were subjected to
gonadectomy or gonadectomy plus hormone replacement. Survival was improved by
gonadectomy and worsened by hormone treatment of gonadectomized mice. Mikerov et al.
3-65
-------
(2011) found that lung and spleen inflammation were evaluated in mice acutely exposed to ozone
and later exposed to pneumonia. Ozone exposure increased the area and severity of lung
inflammation in both male and female mice, with a larger response observed in females. In
addition, spleen red pulp congestion, indicating compromised spleen immune function, occurred
in female mice.
3.1.7.4 Integrated Summary for Respiratory Infection and other Associated Health
Effects
In summary, a large number of recently available epidemiologic studies expand the evidence base
considerably, and provide consistent evidence of an association between short-term ozone exposure and
ED visits for a variety of respiratory infection endpoints (Figure 3-6). The strongest evidence comes from
large multicity studies, and the consistent associations observed across a variety of study designs and
exposure assessment methods. Additionally, there was some limited evidence that the observed
associations were robust, or attenuated, but still positive in copollutant models adjusting for gaseous
pollutants [NO2, SO2, and CO; Malig et al. (2016)1. Further discussion of potential copollutant
confounding can be found in Section 3.1.10. The epidemiologic evidence is supported by animal
toxicological studies evaluated in the 2013 Ozone ISA (U.S. EPA. 2013a) that demonstrate altered
immunity following acute ozone exposure. Additionally, results from a limited number of recent
experimental studies in mice were consistent with previous findings of ozone-induced infectious disease
susceptibility. In contrast, the results of controlled human exposure studies are inconsistent with these
findings and suggest that ozone exposure does not impair macrophage phagocytosis and may enhance
adaptive immunity.
3.1.8 Respiratory Related Hospital Admissions and Emergency Department
(ED) Visits for Aggregated Respiratory-Related Diseases
The 2013 Ozone ISA evaluated a limited number of studies conducted in the U.S., Canada, and
Europe that examined ozone exposure and hospital admissions and/or ED visits for aggregated respiratory
diseases (U.S. EPA. 2013a). These studies generally include hospital admissions or ED visits where the
ICD codes identify a respiratory health outcome (e.g., asthma, COPD, respiratory infection, etc.) as the
primary reason for the visit. In this sense, the measure is not specific to any respiratory outcome in
particular, but represents an aggregate of all respiratory-related diseases. The studies reviewed in the 2013
Ozone ISA added to existing evidence from the 2006 ozone AQCD, which concluded that there was
strong evidence that short-term ozone exposures are associated with increased ED visits and hospital
admissions in the warm season (U.S. EPA. 2006). The strongest evidence from the 2013 Ozone ISA came
from multicity studies of hospital admissions (Katsouvanni et al.. 2009; Cakmak et al.. 2006) and large
single-city studies examining ED visits (Darrow et al.. 2011; Tolbert et al.. 2007). Most studies examined
hospital admissions or ED studies for individuals of all ages (Darrow et al.. 2011; Tolbert et al.. 2007;
3-66
-------
Cakmak et al.. 2006). although Katsouvanni et al. (2009) restricted their analysis to older adults. While
there was limited evaluation of potential copollutant confounding, Tolbert et al. (2007) observed an
association between ozone and respiratory ED visits in Atlanta (March-October) that was robust to
adjustment for CO and NO2, and attenuated, but still positive, in a copollutant model adjusting for PM10.
In a study of respiratory ED visits in Atlanta, GA, Darrow etal. (2011) compared a range of daily
ozone averaging times, including 1-hour max, 8-hour max, 24-hour avg, 6-hour commute time avg
(7:00 a.m.-10:00 a.m., 4:00 p.m.-7:00 p.m.), 11-hour daytime avg (8:00 a.m.-7:00 p.m.), and 6-hour
overnight avg (12:00 a.m -6:00 a.m.) concentrations. Respiratory ED visits were most strongly associated
with 1-hour max, 8-hour max, and 11-hour daytime avg ozone metrics. Associations with 6-hour
commute time avg and 24-hour avg ozone were smaller in magnitude, but still positive, and 6-hour
overnight avg ozone was inversely associated with increased ED visits.
In the following summary of recent studies, hospital admissions and ED visits are evaluated
separately. ED visits for respiratory effects are more common and often less serious than hospital
admissions. Generally, only a small fraction of respiratory ED visits result in a hospital admission. As
such, the two outcomes may reflect different severities of respiratory effects and are best considered
independently.
3.1.8.1 Hospital Admissions
A single recent study provides further evidence of an association between respiratory-related
hospital admissions and ozone exposure (Figure 3-7). Study-specific details, including air quality
characteristics and select effect estimates, are highlighted in Table 3-41 in Section 3.3.1. An overview of
the evidence is provided below.
• A large time-series study in St. Louis, MO (Winquist et al.. 2012) reported that 8-hour daily max
ozone concentrations were associated with hospital admissions for respiratory disease in children
ages 2 to 18 years. Associations with other age groups were null. The study only used one
monitor for the entire study area, which likely contributed exposure measurement error.
3.1.8.2 Emergency Department (ED) Visits
A larger evidence base exists for recent studies of ED visits for aggregated respiratory diseases.
Generally, the evaluated studies use an 8-hour daily max averaging time, although there is one study in
which the 1-hour daily max (Malig et al.. 2016) is used. The averaging-times used in each study, along
with other study-specific details, including air quality characteristics and select effect estimates, are
highlighted in Table 3-42 in Section 3.3.1. An overview of the evidence is provided below.
• Multicity studies provide consistent evidence of an association between ozone and ED visits for
respiratory disease across diverse locations rBarrv et al. (2018): O' Lenick et al. (2017): Malig et
3-67
-------
al. (2016); Figure 3-71. Large single-city studies in St. Louis (Sarnat et al.. 2015; Winquist et al..
2012) and Atlanta (Darrow et al.. 2011) provide corroborating evidence.
• In addition to the diversity of locations examined in the above studies, the positive associations
were observed across a number of exposure assignment techniques, including single monitors for
an entire study area, nearest monitor within 20 km, and population-weighted citywide averages
from 12 km ozone concentration grids estimated using a fusion of observational data from
monitors and pollutant concentration simulations from the CMAQ emissions-based chemical
transport model.
• A limited number of studies evaluated lag structures (Malig et al.. 2016; Darrow et al.. 2011).
seasonal differences in associations (Malig et al.. 2016). and copollutant confounding (Malig et
al.. 2016). These results are discussed in more detail in the Relevant Issues for Interpreting
Epidemiologic Evidence section (Section 3.1.10).
In summary, studies conducted in diverse locations with a variety of exposure assignment
techniques continue to provide evidence of an association between ozone and both hospital admissions
and ED visits for combined respiratory diseases. Additionally, there is some evidence, previously
characterized in the 2013 Ozone ISA, that daily 8-hour max, 1-hour max, and daytime average ozone
concentrations may be most strongly associated with respiratory ED visits (Darrow et al.. 2011).
3-68
-------
Study
ED Visits
Darrow etal. (2011)
Location
Atlanta, GA
Ages Mean Concentration Season Lag
tMalig etal. (2016)
tBarry etal. (2018)
fO'Lenick et at. (2017)
California
Atlanta, GA
Birmingham, AL
Dallas, TX
Pittsburgh, PA
St. Louis, MO
Atlanta, GA
Dallas, TX
St. Louis, MO
Hospital Admissions
Katsouyanni et al. (2009) 90 U.S. Cities
All
fWinquistet at. (2012) St. Louis, MO All
tDarrow etal. (2011) Atlanta, GA All
All
All
5-18
65+
Katsouyanni et al. (2009) 12 Canadian Cities 65+
Cakmaketal. (2006)
10 Canadian Cities All
tWinquistet al. (2012) St. Louis, MO All
2-18
53
62; 1-hr max
20; 24-hr avg
NR
53
33-55; 1-hr max
37.5-42.2
40.0-42.4
34.9-60.0; 1-hr max
6.7-8.3; 1-hr max
17.4; 24-hr avg
NR
Warm
Warm
Warm
Year-Round 0-4 DL
Year-Round
Year-Round
Warm
1
0-1
0-1
Year-Round 0-2
Year-Round 0-2
Year-Round 0-1
Warm
Year-Round 0-1
Warm
Year-Round 1
Year-Round 0-4 DL
1.19 (1.11, 1.27) ~
-+-
0.97 1 1.05 1.1
Effect Estimate (95% CI)
DL = distributed lag.
Note: fStudies published since the 2013 Ozone ISA. Black text = studies included in the 2013 Ozone ISA.
aMean concentrations reported in ppb and are for an 8-hour daily max averaging time unless otherwise noted.
Results standardized to a 15-ppb increase in 24-hour avg, 20-ppb increase in 8-hour daily max, or 25-ppb increase in 1 -hour daily
max ozone concentrations. Corresponding quantitative results are reported in Table 3-40 and Table 3-41.
Figure 3-7 Summary of associations from studies of short-term ozone
exposures and respiratory-related hospital admissions and
emergency department (ED) visits for a standardized increase in
ozone concentrations.
3-69
-------
3.1.9
Respiratory Mortality
Recent multicity studies have not extensively examined the relationship between short-term
ozone exposure and respiratory mortality. The majority of evidence examining respiratory mortality
consists of studies evaluated in the 2013 Ozone ISA, which reported positive associations for respiratory
mortality in all-year and summer/warm season analyses. Of the recent multicity studies evaluated, only
Vanos et al. (2014) examined respiratory mortality and reported positive associations in all-year and
summer season analyses, which is consistent with the multicity studies previously evaluated. These
studies are further characterized in Table 6-5. An additional single-city study examined respiratory
mortality and reported results that are inconsistent with the large body of evidence from multicity studies:
• Klemm etal. (2011) conducted a study in Atlanta, GA that included 7.5 additional years of data
compared with Klemm and Mason (2000) and Klemm et al. (2004). In analyses that examined
respiratory mortality, the authors reported no evidence of an association with respiratory
mortality (-0.44% change in mortality [95% CI: -6.06, 5.51] for a 20-ppb increase in 8-hour max
ozone concentrations).
3.1.10 Relevant Issues for Interpreting Epidemiologic Evidence—Short-Term
Ozone Exposure and Respiratory Effects
3.1.10.1 Potential Copollutant Confounding of the Ozone-Respiratory Relationship
The 2013 Ozone ISA (U.S. EPA. 2013a) evaluated a limited number of studies that examined
potential copollutant confounding. The available studies observed ozone associations with respiratory
hospital admissions and ED visits that were generally robust to the inclusion of gaseous pollutants and
PM in copollutant models (Silverman and Ito. 2010; Ito et al.. 2007; Tolbert et al.. 2007). Along with
some limited evidence from studies of respiratory mortality (Stafoggia et al.. 2010; Katsouvanni et al..
2009). the 2013 Ozone ISA concluded that "copollutant-adjusted findings across respiratory endpoints
provide support for the independent effects of short-term exposures to ambient ozone" (U.S. EPA.
2013a). A number of recent studies are available that provide further evidence that ozone-related
respiratory effects persist in statistical models adjusting for copollutants. The following summary of the
recent evidence is organized by copollutant. Effect estimates from studies that provided quantitative
results from copollutant models are presented in the evidence inventories in Section 3.3.1.
3.1.10.1.1 Fine Particulate Matter (PM2.5)
• Studies evaluated in the 2013 Ozone ISA observed single-pollutant associations between ozone
and hospital admissions (Silverman and Ito. 2010) and ED visits (Ito et al.. 2007) for asthma that
3-70
-------
were robust to statistical adjustment for PM2 5 Tolbert et al. (2007) reported ozone-related
increases in ED visits for combined respiratory diseases that were attenuated, but still positive, in
copollutant models with PM10, but they did not evaluate copollutant models with PM2 5.
• In recent studies, ozone correlations with PM2 5 varied greatly across studies (r = -0.19 to 0.66;
see Section 3.3.1). PM2 5-adjusted ozone effect estimates for asthma- (Sarnat et al.. 2015; Sacks et
al.. 2014) and COPD-related (Rodopoulou et al.. 2015) ED visits were slightly attenuated, but
still positive, compared to single-pollutant estimates. Wendt et al. (2014) used Medicaid claims to
examine initial asthma diagnosis in children in Houston, TX. An association with warm-season
ozone concentrations was similar in magnitude and precision when PM2 5 was included in the
model.
• In one of the few studies on subclinical respiratory effects to examine potential copollutant
confounding, Peng et al. (2016) observed ozone-related increases in FeNO that were similar in
magnitude, but less precise, in a copollutant model adjusting for PM2 5
• One study examined short-term exposure to ozone and asthma- and wheeze-related ED visits in
copollutant models adjusting for various PM2 5 components (Sarnat et al.. 2015). In comparison to
a single-pollutant model, models adjusting for SO42 or NO, were slightly attenuated but still
positive. Effect estimates from copollutant models adjusting for OC, EC, «-Alkanes, hopanes,
PAHs, Si, K, Ca, Fe, Cu, Zn, or Pb were similar to, or slightly larger than the single-pollutant
estimate.
3.1.10.1.2 Sulfur Dioxide (SO2)
• In recently evaluated studies of short-term exposure to ozone and respiratory health effects, ozone
correlations with SO2 were generally weak (r = -0.06 to 0.42; see Section 3.3.1).
• As evaluated in the 2013 Ozone ISA, Ito et al. (2007) reported similar associations between
ozone and asthma ED visits in single pollutant models and copollutant models adjusting for SO2.
• Similar to Ito et al. (2007). a statewide study in California observed single-pollutant,
warm-season associations between ozone and a range of respiratory-related ED visits, including
asthma, ARI, pneumonia, COPD, URTI, and aggregated respiratory diseases, that were relatively
unchanged in copollutant models with SO2 (Malig et al.. 2016).
3.1.10.1.3 Nitrogen Dioxide (NO2)
• The associations between ozone and NO2 in recent studies range from moderate negative
correlations to moderate positive correlations (r = -0.52 to 0.54; see Section 3.3.1).
• Studies evaluated in the 2013 Ozone ISA observed single-pollutant associations between ozone
and ED visits for asthma (Ito et al.. 2007) and combined respiratory disease (Tolbert et al.. 2007)
that persisted in copollutant models adjusting for NO2.
• In a recent study, Malig et al. (2016) reported single-pollutant, warm-season ozone associations
for a variety of respiratory-related ED visit outcomes that were persistent, although sometimes
attenuated, in copollutant models adjusting for NO2. Wendt et al. (2014) similarly reported in a
Medicaid cohort an ozone association with childhood asthma incidence that was reduced in
magnitude, but still positive in a model with NO2.
3-71
-------
3.1.10.1.4
Carbon Monoxide (CO)
• Except for Wendt et al. (2014). the same studies that evaluated potential confounding by NO2
also examined models with CO. The within-study trends for NO2 adjusted models were similar
for CO.
3.1.10.1.5 Summary of Copollutant Confounding Evaluation
In summary, evidence from recent studies is consistent with the 2013 Ozone ISA in supporting an
association between ozone concentrations and respiratory health effects independent of coexposures to
correlated pollutants. Across pollutants, single-pollutant associations reported between ozone and a range
of respiratory-related hospital admissions and ED visits were persistent, although sometimes attenuated,
in copollutant models.
3.1.10.2 The Role of Season and Temperature on Ozone Associations with Respiratory
Health Effects
The 2013 Ozone ISA concluded that stratified seasonal analyses provided evidence of stronger
ozone-respiratory effect associations in the warm season or summer months than in the cold season (U.S.
EPA. 2013a). Seasonal differences were particularly evident for asthma (Strickland et al.. 2010; Ito et al..
2007; Villeneuve et al.. 2007) and COPD (Stieb et al.. 2009; Medina-Ramon et al.. 2006) hospital
admissions and ED visits. Recent studies are generally consistent with these findings. Seasonally
stratified analyses of asthma hospital admissions and ED visits found warm season associations with
ozone that were either similar to or stronger in magnitude than cold season or year-round associations
(Goodman et al.. 2017b; Goodman et al.. 2017a; Malig et al.. 2016; Byers et al.. 2015; Sacks et al.. 2014;
Winquist et al.. 2014). A few studies of COPD also reported a larger increase in ozone-related ED visits
during the warm season (Malig et al.. 2016; Rodopoulou et al.. 2015). In contrast, results from recent
studies of respiratory infection were reversed, with associations that were similar across seasons, or
slightly stronger in magnitude during the cold season (Malig et al.. 2016; Darrow et al.. 2014; Kousha and
Rowe. 2014).
The generally consistent epidemiologic findings throughout the short-term exposure section of
this Appendix include results from models that adjust for temperature using a variety of metrics, including
same-day and single- and multi-day lags of minimum, maximum, and mean temperature. While most
studies adjust for temperature to account for potential confounding related to daily morbidity trends and
time-activity patterns, no recent studies examined whether temperature modifies the relationship between
short-term ozone exposure and respiratory morbidity. However, a recent study of asthma hospitalizations
conducted in 10 Canadian cities assessed potential effect modification by synoptic weather type (Hebbern
and Cakmak. 2015). Individual days were grouped into six weather types based on a range of
meteorological variables, including temperature, dew point, wind speed, pressure, and cloud cover.
3-72
-------
Hebbern and Cakmak (2015) reported ozone-related increases in hospital admissions for asthma across all
six synoptic weather types, including those corresponding to low ozone concentrations. The associations
were heterogeneous across weather types, but no statistically significant differences were observed.
In addition to seasonal analyses, a number of recent studies examined the influence of
aeroallergens on the association between ozone and respiratory health. Like ozone, aeroallergens have
seasonal patterns and have been found to exacerbate asthma. Consequently, aeroallergens may act as a
potential confounder or modifier on the relationship between ozone and respiratory health effects. A few
studies evaluated in the 2013 Ozone ISA (U.S. EPA. 2013a) reported increases in respiratory symptoms
and asthma medication use in models adjusting for pollen (Just et al.. 2002; Ross et al.. 2002; Gielenet
al.. 1997). Several recent studies compared models with and without adjustment for pollen and provided
further evidence supporting an association between ozone and respiratory effects that is independent of
coexposure to pollen. Multicity studies of hospital admissions for asthma in Canada (Hebbern and
Cakmak. 2015) and Texas (Goodman et al.. 2017b; Goodman et al.. 2017a). and a single-city study of
respiratory infection in children in Atlanta (Darrow et al.. 2014) observed single-pollutant associations
with ozone that were at times attenuated (Goodman et al.. 2017a; Hebbern and Cakmak. 2015). but still
persisted in models adjusting for pollen. While most studies adjusted for pollen as a potential confounder,
Gleason et al. (2014) examined pollen as a potential effect modifier on the relationship between ozone
and pediatric asthma ED visits in New Jersey. The authors reported increases in ED visits that were only
associated with same day ozone concentrations on high 3-day avg weed pollen days, indicating that weed
pollen is a potential effect modifier in the relationship between ozone and pediatric asthma ED visits.
3.1.10.3 The Effect of Lag Structure on Associations of Short-Term Ozone Exposure
and Respiratory Effects
The evaluation of lag structure is an important aspect of epidemiologic research on short-term
exposure to air pollution. The examination of lags, along with experimental evidence, can help determine
whether ozone elicits an immediate (lags ranging from 0 to 1 days), delayed (lags ranging from 2 to
5 days), or prolonged (lags averaged from 0 to 5+ days) effect on respiratory health endpoints. Many
recent epidemiologic studies of short-term exposure to ozone and respiratory health effects use previous
evidence established in epidemiologic and experimental literature to define a temporal metric of interest
a priori. Most of these studies, particularly those examining the association between ozone and asthma,
have used 0-2 day avg ozone concentrations (Barry et al.. 2018; O'Lenick et al.. 2017; Alhanti et al..
2016; Xiao et al.. 2016; Sarnat et al.. 2015; Sacks et al.. 2014; Strickland et al.. 2014; Winquist et al..
2014; Sarnat et al.. 2013). Other recent studies have evaluated associations across a range of single-day
and multiday lags. Results from these studies are summarized below.
3-73
-------
3.1.10.3.1 Asthma
• Associations between short-term ozone exposure and hospital admissions and ED visits for
asthma are generally present across daily lags ranging from 0 to 6 days (Table 3-1). Although
precision (e.g., 95% CIs) is not specified in the table, within-study precision was generally
consistent across single-day lags.
• The strongest single-day associations were generally observed with ozone concentrations on the
same day as the outcome, or within the first 3 days prior to the outcome.
• Studies that examined multiday average lag associations generally reported stronger, but less
precise associations than single-day lags (Goodman et al.. 2017b; Zu et al.. 2017; Malig et al..
2016; Bvers et al.. 2015).
3.1.10.3.2 Other Respiratory Effects
• A limited number of studies examined the lag structure of associations between short-term
exposure to ozone and COPD ED visits. In a statewide study in California, Malig et al. (2016)
observed associations between ED visits for COPD and single-day lagged ozone on Days 0
through 3. The largest and most precise (i.e., smallest 95% CI) effect estimate was observed with
ozone concentrations on the day prior to ED visit. Similarly, in a multicity study in Canada,
Szvszkowicz et al. (2018) reported evidence of more immediate effects of ozone in males. The
authors observed associations of similar magnitude and precision on lag Days 0 through 2.
Results for females were more delayed, with associations between ozone and COPD ED visits
noted on lag Days 2 through 4.
• In a study of combined respiratory-related ED visits in Atlanta, warm season associations with
same-day ozone concentrations were strongest (i.e., of greatest magnitude), compared to 1-, 2-,
and 3-day lags (Darrow et al.. 2011). Malig et al. (2016) similarly observed consistent warm
season associations on single-day lags from 0 to 3 but reported the strongest associations with
1- and 2-day-lagged ozone. The authors additionally reported that moving average ozone
concentrations were associated with larger increases in respiratory ED visits, but the estimates
were less precise than single day lag estimates.
3.1.10.3.3 Summary of Evidence on Lag Structures
In summary, the largest evidence base for lag structure comes from studies examining the
association between ozone exposure and hospital admissions or ED visits for asthma. Associations were
generally observed across a range of lags, extending as far as 6 days prior to the health outcome of
interest. This range indicates that ozone may trigger HA or ED visits for asthma both immediately and
over a more extended period. Additionally, the strongest associations were observed with multiday
averages of ozone that were indicative of more immediate effects. Notably, effect estimates derived from
multiday average concentrations were less precise than effect estimates from single-day lag estimates.
Finally, it is important to note that different lag responses may be observed across different population
subgroups (e.g., age or sex groups), as seen in Szvszkowicz et al. (2018).
3-74
-------
Table 3-1 Heat map of daily lag associations between short-term exposure to
ozone and hospital admissions and emergency department (ED)
visits for asthma.
Asthma - Hospital Admissions
Reference
Age
Season
Daily Lag
0
1
2
3
4
5
6
Sheffield et al. (2015)
5-17
Warm
Shmool et al. (2016)
5-17
Warm
Goodman et al. (2017)
5-14
All Year
Zu et al. (2017)
5-14
All Year
Zu et al. (2017)
15-64
All Year
Goodman et al. (2017)
15-64
All Year
H
Asthma
-ED Visits
Reference
Age
Season
Daily Lag
0
i
2
3
4
5
6
Szyszkowiczet al. (2018)
<19 (Female)
Warm
Szyszkowiczet al. (2018)
<19 (Male)
Warm
Sheffield et al. (2015)
5-17
Warm
Shmool et al. (2016)
5-17
Warm
B
B
Qeason et al. (2014)
3-17
Warm
B
Byers et al. (2015)
5-17
All Year
B
Byers et al. (2015)
18-44
All Year
Malig et al. (2016)
AH
Warm
* = Lag at which the strongest association was observed (i.e., largest in magnitude).
Note: Dark blue = study reported statistically significant association (p < 0.05) between ozone and impaired respiratory health
outcome; light blue = study reported association between ozone and impaired respiratory health outcome regardless of width of
confidence intervals; light orange = study reported null or inverse association; red = study reported statistically significant
association between ozone and improved respiratory health outcome; gray = study did not examine individual lags.
3.1.10.4 Shape of the Concentration-Response Relationship
The 2013 Ozone ISA evaluated a large body of epidemiologic evidence that provided evidence of
an association between short-term exposure to ambient ozone and respiratory health effects. Of the
evaluated studies, a limited number attempted to characterize the shape of the C-R relationship or
determine the presence of a concentration threshold below which a positive association with health effects
does not occur. Studies examining asthma-related hospital admissions (Silverman and Ito. 2010) and ED
visits (Strickland et al.. 2010) used natural splines and locally weighted smoothing functions,
respectively, to examine the shape of the C-R relationship between ozone concentrations and
3-75
-------
asthma-related hospital admissions or ED visits. Visual inspections of the plots revealed approximately
linear associations and no evidence of a threshold with 8-hour daily max ozone concentrations as low as
30 ppb (Figure 3-8 and Figure 3-9). There is increased uncertainty in the shape of the C-R curve at the
lower end of the distribution of ozone concentrations, starting around 30 ppb, due to the low density of
data in this range.
Ozone Warm Season
Lfi
CM
O
+j
03
cm
0
-t—'
(0
o:
LO
o
ID
O)
d
30
40
50
60
70
80
Concentration (ppb)
Note: The reference for the rate ratio is the estimated rate at the 5th percentile of the 8-hour daily max ozone concentration.
Estimates are presented for the 5th percentile through the 95th percentile of pollutant concentrations due to instability in the C-R
estimates at the distribution tails. The solid lines are smoothed-fit data, with long broken lines indicating 95% confidence bands.
Source: Reprinted with permission of the American Thoracic Society. Copyright © 2020 American Thoracic Society. Strickland et al
(20101. The American Journal of Respiratory and Critical Care Medicine, Vol. 182, Issue 3, pg. 307-316. The American Journal of
Respiratory and Critical Care Medicine is an official journal of the American Thoracic Society.
Figure 3-8 Locally estimated scatterplot smoothing (LOESS) C-R estimates
and twice-standard-error estimates from generalized additive
models for associations between 8-hour max 3-day avg ozone
concentrations and emergency department (ED) visits for
pediatric asthma.
3-76
-------
Ozone: All
cc
cc
NAAQS
inp»:'i> mi mi i i 11
100
20
40
60
80
Ozone
Note: The average of 0 day and 1 day lagged 8-hour daily max ozone was used in a two-pollutant model with PM25 lag 0-1,
adjusting for temporal trends, day of the week, and immediate and delayed weather effects. The solid lines are smoothed-fit data,
with long broken lines indicating 95% confidence bands. The density of lines at the bottom of the figure indicates sample size. The
NAAQS line indicated in the figure is reflective of a previous standard set in 1997. The form of this NAAQS was the 3-year avg of
annual 4th highest daily max 8-hour concentrations.
Source: Reprinted with permission from the publisher, adapted from Silverman and Ito (20101.
Figure 3-9 Estimated relative risks (RRs) of asthma hospital admissions for
8-hour daily max ozone concentrations at lag 0-1 allowing for
possible nonlinear relationships using natural splines.
In addition, a small number of recent studies show evidence of C-Rnonlinearity and the presence
of a threshold. In contrast to evidence from the 2013 Ozone ISA, a multicity study in Texas estimated
C-R curves using penalized spline models and observed evidence of nonlinearity in the relationship
between 8-hour daily avg ozone and asthma hospital admissions (Zu et al.. 2017). The C-R curves
indicate the potential presence of a threshold between 30 and 40 ppb for children aged 5-14 years and
adults aged 15-64 years (see Figure 3-10). The presence of a threshold in this range is supported by a
recent statewide study in New Jersey that examined associations between pediatric asthma ED visits and
quintiles of 8-hour daily max ozone exposure (Gleason et al.. 2014). In comparison to the lowest quintile
of ozone exposure, only quintiles 3 through 5 were associated with increased odds of ED visits. The third
quintile exposure range started at 42.48 ppb.
3-77
-------
1 s
UD
in
b. children (aged 5-14 yrs)
5 S
o. youngef adults (aged 15-64 yrs}
1
CO
E w
"O *—
CD
%
z ?
J5
f 2-
a
jC
ID
E
g *-
ra O
c ™
S
1 °
0)
Ol
1°
at
p -
119
o
i
10
20 30 40 50
Ozone LagO-3 (ppb)
60
70
10
20 SO 40 50
Ozone LagO-3 (ppb)
60
70
Note: The solid lines are smoothed-fit data, with long broken lines indicating 95% confidence bands.
Source: Reprinted with permission from the publisher, adapted from Zu et al. (20171.
Figure 3-10 Estimated percent change in asthma hospital admissions for
8-hour daily avg ozone concentrations at lag 0-3 allowing for
possible nonlinear relationships using penalized splines.
In contrast to recent results from studies of asthma hospital admissions, Darrow et al. (2014)
reported evidence of approximately linear associations between ozone exposure and pneumonia and upper
respiratory infection. Loess C-R curves provide evidence of an association down to 20 ppb, indicating
that a threshold does not exist in the range of concentrations included in the study (Figure 3-11).
Additionally, in a study of numerous respiratory outcomes in five U.S. cities, Barry et al. (2018) used
maximum likelihood estimation to test for the presence of thresholds ranging from the minimum observed
value (i.e., no threshold) to 50 ppb in 1 ppb increments. The presence of 8-hour daily max thresholds
varied across cities and outcomes, but ranged from no threshold to 40 ppb, with the most commonly
identified thresholds in the 20 to 30 ppb range. The authors also tested linearity by fitting a number of
flexible models and comparing Akaike's Information Criteria values to determine the best fit. The best
model fit also varied by city and outcome, with linear models and cubic spline models providing the best
fit in an equal number of cases.
3-78
-------
30 40 50 60 70
Concentration (ppb)
30 «0 50 60 70
Concentration (ppb)
Ozone - Pneumonia
Ozone - Urtl
Note: The reference for the rate ratio is the estimated rate at the 5th percentile of the pollutant concentration. Estimates are
presented for the 5th percentile through the 95th percentile of pollutant concentrations due to instability in the C-R estimates at the
distribution tails. The solid lines are smoothed-fit data, with long broken lines indicating 95% confidence bands.
Source: Reprinted with permission from the publisher, adapted from Darrow et al. (20141.
Figure 3-11 LOESS C-R estimates and twice-standard-error estimates from
generalized additive models for associations between 3-day
moving avg 8-hour daily max ozone concentrations and
emergency department (ED) visits for pneumonia and upper
respiratory infection.
3.1.11 Summary and Causality Determination
In the 2013 Ozone ISA, it was concluded that "there is a causal relationship between short-term
ozone exposure and respiratory health effects" (U.S. EPA. 2013a). This causality determination was made
on the basis of a strong body of evidence integrated across controlled human exposure, animal
toxicological, and epidemiologic studies, in addition to established findings from previous AQCDs,
demonstrating respiratory effects due to short-term exposure to ozone (U.S. EPA. 2006. 1996a). In
particular, controlled human exposure studies provided evidence of lung function decrements, respiratory
symptoms, and increased inflammation in young healthy adults exposed to ozone concentrations as low as
60 to 70 ppb following 6.6-hour exposures with quasi-continuous exercise. Dose-dependent increases in
airway responsiveness were also noted after exposures to 0, 80, 100, and 120 ppb ozone. These studies
were supported by epidemiologic studies that not only reported ozone-related respiratory effects in
healthy populations, but also provided evidence of ozone associations with asthma exacerbation, COPD
exacerbation, and hospital admissions and ED visits for combined respiratory disease. Additionally, there
was consistent evidence of an association between short-term increases in ambient ozone concentrations
and increases in respiratory mortality. Results observed in controlled human exposure and epidemiologic
studies were supported by animal toxicological studies that indicated changes to ventilatory parameters,
3-79
-------
increased airway responsiveness, and lung injury and inflammatory responses resulting from ozone
exposures. Experimental studies also described the potential mechanistic pathways that underlie the
respiratory effects observed in epidemiologic studies. Taken together, the synthesized results provided
compelling evidence of a causal relationship between short-term exposure to ozone and respiratory
effects.
Recent studies further expand the body of evidence regarding the relationship between short-term
exposure to ozone and respiratory effects (Table 3-2). Evidence from a recent controlled human exposure
study of respiratory effects in healthy adults is consistent with findings from prior assessments
demonstrating post-exercise decrements in group mean pulmonary function after ozone exposures
(Section 3.1.4.1. IV There were no recent experimental studies in humans that examined respiratory
symptoms in relation to short-term ozone exposures nor were there new 6.6-hour exposure studies.
However, ozone-induced respiratory symptoms in combination with FEVi decrements in young healthy
adults at concentrations as low as 70 ppb (targeted concentration) were reported in the 2013 Ozone ISA
(U.S. EPA. 2013a).
Controlled human exposure studies evaluated in the 2006 Ozone AQCD (U.S. EPA. 2006) and
the 2013 Ozone ISA (U.S. EPA. 2013a) also provide consistent evidence of ozone-induced increases in
airway responsiveness and inflammation in the respiratory tract and lungs. A recent study expand on
observed interindividual variability in inflammatory responses, providing some additional evidence
suggesting that young healthy GSTMl-null adults may be more susceptible to ozone-related
inflammatory responses (Section 3.1.4.4.1). Recent animal toxicological studies are consistent with
evidence summarized in the 2013 Ozone ISA (U.S. EPA. 2013a) and support the evidence observed in
healthy humans. Specifically, recent studies demonstrated altered ventilatory parameters and increases in
airway responsiveness, inflammation, injury, and oxidative stress following ozone exposures.
Additionally, repeated exposure to ozone resulted in type 2 immune responses in the upper and lower
airways (Section 3.1.4.4.2).
Evidence from epidemiologic studies of healthy populations is generally coherent with
experimental evidence, although the majority of the evidence comes from panel studies that were
previously evaluated in the 2013 Ozone ISA (U.S. EPA. 2013a). including a number of studies of
children in summer camps that observed decreases in FEVi and increases in markers of pulmonary
inflammation associated with increases in short-term ozone exposure. In contrast to coherence of panel
studies with experimental evidence of ozone-induced lung function decrements and respiratory tract
inflammation, respiratory symptoms were not associated with ozone exposure in a limited number of
panel studies. However, these studies of children generally relied on parental reported outcomes that may
result in under- or over-reporting of respiratory symptoms.
Evidence from a large number of recent, large multicity epidemiologic studies conducted in the
U.S. also expand upon evidence from the 2013 Ozone ISA (U.S. EPA. 2013a) to provide further support
for an association between ozone and ED visits and hospital admissions for asthma (Section 3.1.5.1 and
3-80
-------
Section 3.1.5.2). Observed associations were generally of greatest magnitude for children between the
ages of 5 and 18 years. Additionally, associations were observed across models implementing measured
and modeled ozone concentrations. Although there are a limited number of recent epidemiologic studies
conducted in the U.S. or Canada that examine respiratory symptoms and medication use, lung function,
and subclinical effects in people with asthma, a large body of evidence from the 2013 Ozone ISA (U.S.
EPA. 2013a) reported ozone associations with these less severe markers of asthma exacerbation that
provide support for the ozone-related increases in asthma hospital admissions and ED visits observed in
recent studies. Recent experimental studies in animals, along with similar studies summarized in the 2013
Ozone ISA (U.S. EPA. 2013a). provide coherence with the epidemiologic evidence of asthma
exacerbation, indicating respiratory tract inflammation, oxidative stress, injury, allergic skewing, goblet
cell metaplasia, and upregulation of mucus synthesis and storage in allergic mice exposed to ozone
(Section 3.1.5.4. Section 3.1.5.5. and Section 3.1.5.6).
In addition to epidemiologic evidence of asthma exacerbation, and consistent with studies
reviewed in the 2013 Ozone ISA (U.S. EPA. 2013a). several recent epidemiologic studies provide
evidence of an association between ozone and ED visits for respiratory infection (Section 3.1.7.1) and ED
COPD (Section 3.1.6.1.1). A limited number of recent epidemiologic studies examining respiratory
mortality were inconsistent (Section 3.1.9). but should be considered in the context of studies evaluated in
the 2013 Ozone ISA that provided consistent evidence of an association between short-term ozone
exposure and respiratory mortality (U.S. EPA. 2013a). A limited number of recent controlled human
exposure and animal toxicological studies are consistent with studies evaluated in the 2013 Ozone ISA
(U.S. EPA. 2013a) that demonstrate altered immunity and impaired lung host defense following acute
ozone exposure (Section 3.1.7.3). These findings support the epidemiologic evidence of an association
between ozone concentrations and respiratory infection. Additionally, results from recent animal
toxicological studies provide new evidence that chronic inflammation enhances sensitivity to ozone
exposure, providing coherence for epidemiologic evidence of ozone-related exacerbation of COPD
(Section 3.1.6.1.2). Along with evidence from the 2013 Ozone ISA (U.S. EPA. 2013a). recent
epidemiologic studies continue to provide evidence of an association between ozone and hospital
admissions and ED visits for combined respiratory-related diseases. Because combined respiratory-related
diseases are non-specific and encompass a range of respiratory endpoints, these associations are
potentially driven by the observed associations with specific respiratory ED visits and hospital admissions
discussed previously.
Recent mechanistic studies in humans and animals have expanded on findings from previous
assessments (U.S. EPA. 2013a. 2006. 1996a) and improved the understanding of plausible pathways that
may underlie the observed respiratory health effects resulting from short-term exposure to ozone.
Notably, changes in lung function may be attributed to activation of sensory nerves in the respiratory tract
that trigger local and autonomic reflex responses. Modest increases in airway resistance may occur due to
activation of parasympathetic pathways. Mechanistic studies also present a plausible pathway by which
ozone reacts with respiratory tract components to produce oxidized species that injure barrier function and
3-81
-------
activate innate immunity, resulting in a cycle of inflammation, injury, and oxidative stress. A recent
animal toxicological study has also demonstrated that vagal C-fibers, vagal myelinated fibers, and
possibly neuropeptides released in the airway are involved in increased airway responsiveness and
bronchoconstriction in allergic animals. Together, results from mechanistic studies may provide
biological plausibility for evidence of ozone-related lung function decrements and increased asthma
symptoms from epidemiologic panel studies in healthy children and in children with asthma.
Furthermore, they support the results of epidemiologic studies showing associations between ozone
exposure and asthma-related ED visits and hospital admissions.
Copollutant analyses were limited in epidemiologic studies evaluated in the 2013 Ozone ISA, but
did not indicate that associations between ozone concentrations and respiratory effects were confounded
by copollutants or aeroallergens (U.S. EPA. 2013a). Copollutant analyses have been more prevalent in
recent studies and continue to suggest that observed associations are independent of coexposures to
correlated pollutants or aeroallergens (Section 3.1.10.1 and Section 3.1.10.2). Despite expanded
copollutant analyses in recent studies, determining the independent effects of ozone in epidemiologic
studies is complicated by the high copollutant correlations observed in some studies, and the possibility
for effect estimates to be overestimated for the better measured pollutant in copollutant models
(Section 2.5). Nonetheless, the consistency of associations observed across studies with different
copollutant correlations, the generally robust associations observed in copollutant models, and evidence
from controlled human exposure studies demonstrating respiratory effects in response to ozone exposure
in the absence of other pollutants, provide compelling evidence for the independent effect of short-term
ozone exposure on respiratory symptoms.
Recent epidemiologic studies have also attempted to further inform our understanding of the lag
structure (Section 3.1.10.3) and the shape of the C-R relationship (Section 3.1.10.4) for associations
between short-term exposure to ozone and respiratory effects. The largest evidence base for lag structure
comes from studies of ozone exposure and hospital admissions or ED visits for asthma. Single-day
associations were generally observed with ozone concentrations across a range of lags, extending as far as
6 days prior to the health outcome of interest. This range indicates that ozone may trigger immediate
respiratory effects and respiratory effects occuring over a more extended period. Studies examining the
shape of the C-R relationship and/or the presence of a threshold have been inconsistent. While most
studies assume a no-threshold, log-linear C-R shape, a limited number of studies have used more flexible
models to test this assumption. Results from some of these studies indicate approximately linear
associations between ozone concentrations and hospital admissions for asthma, while others indicate the
presence of a threshold ranging from 20 to 40 ppb 8-hour max ozone concentrations.
In summary, recent studies evaluated since the completion of the 2013 Ozone ISA support and
expand upon the strong body of evidence that indicated a causal relationship between short-term ozone
exposure and respiratory health effects. Controlled human exposure studies demonstrate ozone-induced
decreases in FEVi and pulmonary inflammation at concentrations as low as 60 ppb after 6.6 hours of
3-82
-------
exposure with quasi-continuous exercise. The combination of lung function decrements and respiratory
symptoms has been observed following 6.6 hour exposures with quasi-continuous exercise to 70 ppb
(targeted concentration) and greater ozone concentrations. Epidemiologic studies continue to provide
evidence that increased ozone concentrations are associated with a range of respiratory effects, including
asthma exacerbation, COPD exacerbation, respiratory infection, and hospital admissions and ED visits for
combined respiratory diseases. A large body of toxicological studies demonstrate ozone-induced changes
in ventilatory parameters, inflammation, increased airway responsiveness, and impaired lung host
defense. Additionally, mouse models indicate enchanced ozone-induced inflammation, oxidative stress,
injury, allergic skewing, goblet cell metaplasia, and upregulation of mucus synthesis and storage in
allergic mice compared to naive mice. These toxicological results further inform the potential mechanistic
pathways that underlie downstream respiratory effects, providing continued support for the biological
plausibility of the observed epidemiologic results. Thus, the recent evidence integrated across
disciplines, along with the total body of evidence evaluated in previous integrated reviews, is
sufficient to conclude that there is a causal relationship between short-term ozone exposure and
respiratory health effects.
Table 3-2 Summary of evidence indicating a causal relationship between
short-term ozone exposure and respiratory effects.
Ozone
Concentrations
Rationale for Causality Associated with
Determination3 Key Evidence13 Key References'3 Effects0
Respiratory Effects in Healthy Populations
Consistent evidence from controlled Studies show:
human exposure studies at relevant
concentrations Decrements in lung function Section 3.1.4.1.1 60-400 ppb
Increased respiratory symptoms Section 3.1.4.2.1 70-400 ppb
Increased airway responsiveness Section 3.1.4.3.1 80-1,000 ppb
Inflammation, injury, and oxidative Section 3.1.4.4.1 60-600 ppb
stress
3-83
-------
Table 3-2 (Continued): Summary of evidence indicating a causal relationship
between short-term ozone exposure and respiratory
effects.
Rationale for Causality
Determination3 Key Evidence13
Consistent evidence from Studies show:
toxicological studies at relevant
concentrations Altered ventilatory parameters
Cough response
Increased airway responsiveness
Inflammation, injury, and oxidative
stress
Type 2 immune responses—upper
and lower airways
Coherence in epidemiologic studies
of respiratory effects in healthy
children
Consistent epidemiologic evidence
of severe asthma exacerbation
from multiple, high-quality studies
at relevant concentrations
Panel studies provide support for
experimental studies with consistent
associations for lung function and
pulmonary inflammation in healthy
children
Increases in asthma-related hospital
admissions and ED visits in children,
and all ages combined in studies
conducted in the U.S. and Canada
Ozone
Concentrations
Associated with
Key References'3 Effects0
Section 3.1.4.1.2 0.1-2 ppm
Clay et al. (2016) 2 ppm
Section 3.1.4.3.2 0.3-2 ppm
Section 3.1.4.4.2 0.15-2 ppm
Harkema et al. 0.5-0.8 ppm
(2017): Kumaaai et
al. (2017): Ona et
al. (2016): Kumaqai
et al. (2016).
Section 3.1.4.1.3 32.6 ppb (8-h
moving avg)
53-123 ppb (1-h
avg)
Section 3.1.4.4.3 31.6 ppb 8-h
moving avg
400-420 ppb
2 ppm
Section 3.1.5.1 8-h max/avg:
Section 3.1.5.2 30.7-53.9 ppb
24-h avg:
22.5-41.9 ppb
Evidence supporting biological
plausibility
Controlled human exposure studies Section 3.1.4.1.1
provide evidence showing
involvement of vagal C-fibers in pain
on inspiration, decreased forced vital
capacity, and altered breathing
frequency. In addition, there is
involvement of parasympathetic
pathways leading to increased airway
resistance
Animal toxicological studies provide Section 3.1.4.1.2
evidence that changes in lung
function may be attributed to
activation of sensory nerves and
involvement of parasympathetic
pathways
Section 3.1.4.3.2
Clav et al. (2016)
Verhein et al.
(2013)
Asthma Exacerbation and Associated Respiratory Effects in Populations with Asthma
3-84
-------
Table 3-2 (Continued): Summary of evidence indicating a causal relationship
between short-term ozone exposure and respiratory
effects.
Ozone
Concentrations
Rationale for Causality Associated with
Determination3 Key Evidence13 Key References'3 Effects0
Consistent evidence from controlled
human exposure studies at relevant
concentrations
Studies show that individuals with
asthma experience all the
ozone-induced respiratory outcomes
(e.g., lung function decrements)
observed in individuals without
asthma. However, studies are not
available at concentrations below
125 ppb
Section 3.1.5.3.1 >125 ppb
Section 3.1.5.4.1
Section 3.1.5.5.1
Section 3.1.5.6.1
Consistent evidence from
toxicological studies at relevant
concentrations
Studies show enhanced allergic
responses, bronchoconstriction,
airway responsiveness, and altered
ventilatory parameters in animal
models of allergic airway disease
Section 3.1.5.4.2 0.1-2 ppm
Section 3.1.5.5.2
Section 3.1.5.6.2
Coherence in epidemiologic studies
across the continuum of effects
Panel studies in children with asthma
provide support for asthma
exacerbation in children, with
consistent associations for respiratory
symptoms, lung function decrements,
and pulmonary inflammation
Section 3.1.5.3.2 1-h max:
Section 3.1.5.4.3 43.0-65.8 ppb
Section 3.1.5.6.3 3.^ max-
31.6-52.9 ppb
Epidemiologic evidence from Potential copollutant confounding is Section 3.1.10.1
copollutant models provide some examined in a number of studies with
support for an independent ozone evidence that associations persist in
association models with gaseous pollutants and
PM2.5
Evidence supporting biological Evidence from animal toxicological Scheleale and 1 ppm
plausibility study demonstrates involvement of Walbv (2012)
vagal C-fibers in increased airway
resistance and airway
responsiveness in a model of allergic
airway disease, providing biological
plausibility for epidemiologic findings
for exacerbation of allergic asthma,
the most common asthma phenotype
in children
3-85
-------
Table 3-2 (Continued): Summary of evidence indicating a causal relationship
between short-term ozone exposure and respiratory
effects.
Rationale for Causality
Determination3
Key Evidence13
Key References'3
Ozone
Concentrations
Associated with
Effects0
Exacerbation of COPD and Associated Respiratory Effects in Populations with COPD
Consistent epidemiologic evidence
of severe exacerbation of COPD
from a limited number of
high-quality multicity studies at
relevant concentrations
Increases in ED visits for COPD in
studies conducted in the U.S. and
Canada
Stieb et al. (2009)
24-h avg:
18.4 ppb
Maliq et al. (2016)
1-h max:
33-55 ppb
Szvszkowicz et al.
(2018)
24-h avg:
22.5-29.2 ppb
Consistent evidence from a limited
number of toxicological studies at
relevant concentrations
Results show enhanced injury,
inflammation, oxidative stress, and
altered morphology and lung
mechanics in animal model of COPD
Groves et al. (2012)
Groves et al. (2013)
0.8 ppm
But, lack of coherence in
epidemiologic studies across the
continuum of effects
Panel studies in adults with COPD do
not observe ozone associations with
lung function or respiratory symptoms
in adults with COPD
Peacock et al.
(2011): Maqzamen
et al. (2018)
Also, limited evaluation of
confounding by copollutants
Potential copollutant confounding is
examined in a single study, with
evidence that associations remain
robust in copollutant models adjusted
for gaseous pollutants
Maliq et al. (2016)
Respiratory Infection
Generally consistent epidemiologic
evidence from multiple, high-quality
studies at relevant concentrations11
Increases in ED visits for:
Pneumonia
Maliq et al. (2016):
Xiao et al. (2016)
1-h max:
33-55 ppb
8-h max:
42.1 ppb
Acute respiratory infections
Maliq et al. (2016)
1-h max:
33-55 ppb
Upper respiratory tract infections
Maliq et al. (2016):
Xiao et al. (2016):
Szvszkowicz et al.
(2018): Barrvetal.
(2018).
1-h max:
33-55 ppb
8-h max:
37.5-42.2 ppb
24-h avg:
22.5-29.2 ppb
3-86
-------
Table 3-2 (Continued): Summary of evidence indicating a causal relationship
between short-term ozone exposure and respiratory
effects.
Rationale for Causality
Determination3
Key Evidence13
Key References'3
Ozone
Concentrations
Associated with
Effects0
Coherence in toxicological studies
at relevant concentrations
Increased susceptibility to infectious
disease
Section 3.1.7.3
0.08-2 ppm
Increased inflammatory response to Mikerov et al. 2 ppm
infectious disease (2011)
Evidence of biological plausibility
Animal toxicoloqical studies show Section 3.1.7.3
increased susceptibility to infections
0.08-2 ppm
Other Respiratory-Related Hospital Admissions and ED Visits
Epidemiologic studies provide
consistent evidence of positive
associations when examining
combined respiratory-related
diseasesd
Increases in hospital admissions and Section 3.1.8
ED visits for combined respiratory-
related diseases in multicity studies.
Combined respiratory-related
diseases are characterized by ICD
codes that identify any respiratory
health outcome (e.g., asthma, COPD,
bronchitis, etc.) as the primary reason
for a hospital admission or ED visit
1-h max:
33-55 ppb
8-h max:
30.7-50.3 ppb
But, limited evaluation of
confounding by copollutants
Potential copollutant confounding is
examined in a limited number of
studies, with evidence that
associations generally remain robust
in models with gaseous pollutants
Section 3.1.10.1
Additional evidence of positive
associations from epidemiologic
studies of hospital admissions and
ED visits for specific respiratory
endpoints are summarized in the
endpoint-specific sections of this
table
Asthma Exacerbation
Exacerbation of COPD
Respiratory Infection
Section 3.1.5.1
Section 3.1.5.2
Section 3.1.6.1
Section 3.1.7.1
Respiratory Mortality
Generally consistent epidemiologic
evidence from multiple, high-quality
studies at relevant concentrations11
Generally consistent evidence of
increases in mortality in response to
short-term ozone exposure in
multicity studies in the U.S. and
Canada. Evidence of effects within
the first 2 days of exposure (lag 0 to
2 days)
Section 6.1.4
1-h max:
6.7-38.4 ppb
8-h avg/max:
15.1-62.8 ppb
24-h avg:
19.3 ppb
3-87
-------
Table 3-2 (Continued): Summary of evidence indicating a causal relationship
between short-term ozone exposure and respiratory
effects.
Rationale for Causality
Determination3
But, limited evaluation of
confounding by copollutants
Key Evidence13
Potential copollutant confounding for
is examined in a single study, with
evidence that associations remain
robust in copollutant models adjusted
PM10
Ozone
Concentrations
Associated with
Key References'3 Effects0
Katsouvanni et al.
(2009)
aBased on aspects considered in judgments of causality and weight of evidence in causal framework in Table I and Table II of the
Preamble to the ISAs (U.S. EPA. 2015V
bDescribes the key evidence and references, supporting or contradicting, contributing most heavily to causality determination and,
where applicable, to uncertainties or inconsistencies. References to earlier sections indicate where full body of evidence is
described.
°Describes the ozone concentrations with which the evidence is substantiated. For epidemiologic studies, the study area mean or
median ozone concentrations from the relevant studies are reported. Study-specific ozone concentrations are presented in the
evidence inventories (Section 3.3.11.
dStudy populations were not stratified by, or restricted to, baseline underlying health status (e.g., people with asthma).
3.2 Long-Term Ozone Exposure
3.2.1 Introduction, Summary from the 2013 Ozone ISA, and Scope for Current
Review
The 2013 Ozone ISA concluded that "there is likely to be a causal relationship between
long-term exposure to ozone and respiratory health effects'" (U.S. EPA. 2013a). The epidemiologic
evidence for a relationship between long-term ozone exposure and respiratory health effects was provided
by studies of new-onset asthma, respiratory symptoms in children with asthma, and respiratory mortality.
Associations between long-term exposure to ozone and new-onset asthma in children and increased
respiratory symptoms in individuals with asthma were primarily observed in studies that examined
interactions between ozone and exercise or different genetic variants. The evidence relating new-onset
asthma to long-term ozone exposure was supported by toxicological studies of allergic airways disease in
infant monkeys. This nonhuman primate evidence that ozone exposure altered airway development
supported the biological plausibility of early-life exposure to ozone contributing to asthma development
in children. Generally, the epidemiologic and toxicological evidence provided a compelling case that
supported the causality determination for long-term exposure to ambient ozone and measures of
respiratory health effects. Results from a limited number of epidemiologic studies examining potential
copollutant confounding suggested that the observed associations were robust to adjustment for other
pollutants, including PM25 in a study of long-term ozone exposure and respiratory mortality.
Additionally, the evidence for short-term exposure to ozone and effects on respiratory endpoints provided
3-88
-------
support for the observed respiratory health associations with long-term exposure to ozone. Building upon
that evidence, more recent epidemiologic evidence, combined with toxicological studies in rodents and
nonhuman primates, provides biologically plausible evidence of a likely to be causal relationship between
long-term exposure to ozone and respiratory effects.
The following section on long-term ozone exposure and respiratory effects begins with an
overview of study inclusion criteria (Section 3.2.2) that defines the scope of the literature to be considered
for inclusion in the section. The ensuing section presents a discussion of biological plausibility
(Section 3.2.3) that provides background for the subsequent sections in which groups of related endpoints
are presented in the context of relevant disease pathways. The respiratory effects subsections are
organized by outcome group and aim to clearly characterize the extent of coherence among related
endpoints and biological plausibility of ozone effects. These outcome groups include development of
asthma (Section 3.2.4.1). lung function and development (Section 3.2.4.2). development of COPD
(Section 3.2.4.3). respiratory infection (Section 3.2.4.4). severity of respiratory disease (Section 3.2.4.5).
allergic responses (Section 3.2.4.6). respiratory effects in healthy pregnancy (Section 3.2.4.7). respiratory
effects in populations with metabolic syndrome (Section 3.2.4.8). and respiratory mortality
(Section 3.2.4.9). Finally, Section 3.2.5 comprises an integrated discussion of relevant issues for
interpreting the epidemiologic evidence discussed in Section 3.2.4. Throughout the sections on respiratory
health effects, results from recent studies are evaluated in the context of the evidence provided by
previous studies in the 2013 Ozone ISA (U.S. EPA. 2013a). Study-specific details, including exposure
time periods and exertion levels in experimental studies, and study design, exposure metrics, and select
results in epidemiologic studies are presented in evidence inventories in Section 3.3.
3.2.2 Population, Exposure, Comparison, Outcome, and Study Design
(PECOS) Tool
The scope of this section is defined by a scoping tool that generally describes the relevant
Population, Exposure, Comparison, Outcome, and Study Design (PECOS). The PECOS tool defines the
parameters and provides a framework to help identify the relevant literature to inform the Ozone ISA.
Because the 2013 Ozone ISA concluded there is a likely to be causal relationship between long-term
ozone exposure and respiratory health effects, the recent epidemiologic studies evaluated in this ISA are
limited to study locations in the U.S. and Canada to provide a focus on study populations and air quality
characteristics that are most relevant to circumstances in the U.S. The studies evaluated and subsequently
discussed within this section were included if they satisfied all of the components of the following
PECOS tool:
Experimental Studies:
• Population: Study population of any animal toxicological study of mammals at any lifestage
3-89
-------
• Exposure: Long-term (on the order of months to years) or perinatal inhalation exposure to
relevant ozone concentrations (i.e., <2 ppm)
• Comparison: Appropriate comparison group exposed to a negative control (i.e., clean air or
filtered air control)
• Outcome: Respiratory effects
• Study Design: Studies in mammals meeting the above criteria
Epidemiologic Studies:
• Population: Any U.S. or Canadian population, including populations or lifestages that might be at
increased risk
• Exposure: Long-term exposure (months to years) to ambient concentration of ozone
• Comparison: Per unit increase (in ppb), or humans exposed to lower levels of ozone compared
with humans exposed to higher levels
• Outcome: Change in risk (incidence/prevalence) of respiratory effects
• Study Design: Epidemiologic studies consisting of panel, case-crossover, time-series studies, and
case-control studies, as well as cross-sectional studies with appropriate timing of exposure for the
health endpoint of interest
3.2.3 Biological Plausibility
This section describes biological pathways that potentially underlie respiratory health effects
resulting from long-term exposure to ozone. Figure 3-12 graphically depicts the proposed pathways as a
continuum of upstream events, connected by arrows, that lead to downstream events observed in
epidemiologic studies. This discussion of how long-term exposure to ozone may lead to respiratory health
effects contributes to an understanding of the biological plausibility of epidemiologic results evaluated
later in Section 3.2.4. Note that the structure of the biological plausibility sections and the role of
biological plausibility in contributing to the weight-of-evidence analysis used in the 2020 Ozone ISA are
discussed in Section IS.4.2.
Evidence that long-term exposure to ozone may affect the respiratory tract generally informs one
proposed pathway (Figure 3-12). It begins with oxidative stress, inflammation, and injury in the
respiratory tract, as demonstrated by studies in rodents described in the 2013 Ozone ISA (U.S. EPA.
2013a) and more recently. These responses, which are difficult to disentangle, were also observed in some
studies of short-term exposure to ozone. Prolonged or intermittent exposure of adult rodents to ozone over
months to years led to persistent inflammation and morphologic alterations, including fibrotic- and
emphysematous-like changes (U.S. EPA. 2013a). Also discussed in the 2013 Ozone ISA was an increase
in the severity of post-influenza alveolitis and injury to nasal airways that resulted in altered structure and
function of the nose (U.S. EPA. 2013a). In an infant monkey model of allergic airway disease, postnatal
ozone exposure compromised airway growth and development and resulted in changes that favor allergic
3-90
-------
airways responses and persistent effects on the immune system (U.S. EPA. 2013a). Baseline airway
responsiveness and nonspecific airway responsiveness were increased, morphologic changes occurred
that were consistent with increased airway responsiveness, airway neural innervation was altered, and a
host defense response was diminished. These types of alterations in structure and function in the
developing lung may underlie the development of asthma.
Altered
Morphology in
Lower Airways and
Alveolar Region
Long-
Term
Ozone
Exposure
Respiratory
Tract
Inflammation
Increased Severity
of Influenza
Altered Morphology
in Nasal Airways
Allergic Responses
Serotonin
Upregulation/
Altered Neural
innervation
Altered Innate
Immune Function
Increased
Airway
Responsiveness
Fibrotic- or
Emphysema-
Like
Disease/COPD
~
Respiratory
Mortality
Altered Lung
Development
Development of
Asthma
Note; The boxes above represent the effects for which there is experimental or epidemiologic evidence related to ozone exposure,
and the arrows indicate a proposed relationship between those effects. Solid arrows denote evidence of essentiality as provided, for
example, by an inhibitor of the pathway or a genetic knockout model used in an experimental study involving ozone exposure.
Shading around multiple boxes is used to denote a grouping of these effects. Arrows may connect individual boxes, groupings of
boxes, and individual boxes within groupings of boxes. Progression of effects is generally depicted from left to right and color-coded
(gray, exposure; green, initial effect; blue, intermediate effect; orange, effect at the population level or a key clinical effect). Here,
population level effects generally reflect results of epidemiologic studies. When there are gaps in the evidence, there are
complementary gaps in the figure and the accompanying text below. The structure of the biological plausibility sections and the role
of biological plausibility in contributing to the weight-of-evidence analysis used in the 2020 Ozone ISA are discussed in
Section IS.4.2.
Figure 3-12 Potential biological pathways for respiratory effects following
long-term ozone exposure.
3-91
-------
Recent studies include those conducted in adult and neonatal rodents and those conducted in
infant monkeys. Studies in adult rodents found that long-term exposure to ozone results in respiratory
tract oxidative stress, inflammation, and injury (Gordon et al.. 2016b; Gordon et al.. 2016a; Miller et al..
2016a; Snow et al.. 2016; Gordon et al.. 2013). Epithelial hyperplasia of terminal bronchioles, alveolar
ducts, and adjacent alveoli was observed in adult rats exposed repeatedly to ozone (Gordon et al.. 2013).
Findings of oxidative stress, inflammation, and injury were reported in the developing lungs of rodents
exposed postnatally to ozone (Dve et al.. 2017; Gabehart et al.. 2015; Gabehart et al.. 2014). In addition,
secretion and upregulation of mucus expression, which can offer protection against injury, were
increased, while cell proliferation was decreased in the neonatal rodents.
Postnatal ozone exposure had morphological effects. For example, decreased sensory neuron
development (Zellner et al.. 2011) and altered airway architecture (Lee etal.. 2011) were demonstrated in
rodents. Studies in infant monkeys found altered components of a cell death pathway, altered expression
of serotonin, which is a neurotransmitter involved in airway smooth muscle contraction, and altered
innate immune function (Clay et al.. 201.4; Murphy et al.. 2014; Murphy et al.. 2013). Alterations in cell
growth and cell death pathways observed in these long-term studies may underlie changes in structure
(i.e., airway architecture) in the developing lung. Effects on serotonin could potentially underlie changes
in function in the developing lung (i.e., increased airway responsiveness), while effects on innate immune
function may lead to altered immune response. Studies in the infant model of allergic airway disease
model found impaired alveolar morphogenesis (Herring et al.. 2015; Avdalovic et al.. 2012). airway
smooth muscle hyperreactivity (Moore et al.. 2012a). an enhanced allergic phenotype (Crowley et al..
2017; Chouetal.. 2011). and priming of responses to oxidant stress (Murphy et al.. 2012).
As described here, there is one main pathway, with many branches, by which long-term exposure
to ozone could lead to respiratory health effects. It involves respiratory tract oxidative stress,
inflammation, and injury as early events resulting from prolonged exposure. Respiratory tract
inflammation may also lead to morphologic and immune system-related changes that may affect the
structure and function of the respiratory tract. In adult animals these changes may underlie the progression
and development of chronic lung disease. In developing lungs, these changes may underlie impaired lung
development or the development of asthma. The multibranched pathway described here may provide
biological plausibility for the epidemiologic evidence of COPD- and influenza-related mortality in adults.
Increased severity of infection-related lung disease may also underlie mortality. In addition, ozone-related
effects on the developing lung may provide biological plausibility for epidemiologic evidence for
new-onset asthma and increased respiratory symptoms in children with asthma. These pathways will be
used to inform a causality determination, which is discussed later in the Appendix (Section 3.2.6).
3-92
-------
3.2.4
Respiratory Health Effects
3.2.4.1 Development of Asthma
Asthma is a chronic inflammatory disease of the airways that develops overtime (NHLBI. 2017).
Pulmonary inflammation can increase airway responsiveness and induce airway remodeling, resulting in
bronchoconstriction (bronchial smooth muscle contraction), and in turn, episodes of shortness of breath,
coughing, wheezing, and chest tightness. When the pathophysiology of asthma advances to the stage at
which symptoms lead people to seek medical treatment, a diagnosis of asthma can result. Asthma
development is often assessed in epidemiologic studies by identifying asthma incidence, or new-onset
asthma cases, through parental-reported outcomes, physician-confirmed diagnoses, or administrative
databases of medical claims data.
3.2.4.1.1 Epidemiologic Studies
The 2013 Ozone ISA reported evidence of an association between long-term exposure to ozone
and asthma incidence in adults from two epidemiologic studies of the same cohort (U.S. EPA. 2013a).
Evidence was available from a cohort study and a subsequent extended follow-up of nonsmoking,
non-Hispanic, white, Seventh-Day Adventist adults in California (McDonnell et al.. 1999a; Greer et al..
1993). The association, which was only observed in stratified analyses of male participants, was robust to
the inclusion of PMio, SO42 , SO2, and NO2 in copollutant models (McDonnell et al.. 1999a). Notably, the
results provide limited generalizability, given the restricted cohort demographics.
Studies evaluated in the 2013 Ozone ISA did not provide evidence of a main effect of long-term
ozone exposure on asthma development in children, but they did indicate potential interactions between
ozone and exercise or different genetic variants on childhood asthma (U.S. EPA. 2013a). In analyses of
the Children's Health Study (CHS) cohort in southern California, Islam et al. (2008) and Salam et al.
(2009) observed evidence of interaction between ozone concentrations and functional polymorphisms of
the heme oxygenase-1 (HMOX-1) gene and variants in genes for arginase, respectively, related to the risk
of new-onset asthma in children. Specifically, Islam et al. (2008) found that the protective effect of a
functional promoter variant in the HMOX-1 gene that decreases oxidant burden was larger in children
living in low-ozone communities. In the same cohort, Mcconnell et al. (2002) reported increased asthma
incidence in children who played three or more sports in high-ozone communities, compared with those
who played no sports. In contrast, no such association was observed in low-ozone communities,
indicating that ozone concentrations may modify the effect of exercise and asthma development.
Alternatively, children who played three or more sports in high-ozone communities may have had
3-93
-------
increased asthma incidence as a result of higher ozone exposure from increased ventilation rates while
exercising.
A limited number of recent studies provide evidence of an association between long-term
exposure to ozone and asthma development in children. Only a few recent epidemiologic studies in the
U.S. or Canada examine asthma development or subclinical effects underlying asthma development in
children, while none focus on asthma development in adults. Study specific details, including air quality
characteristics and select effect estimates, are highlighted in Table 3-43 in Section 3.3.2. An overview of
the evidence is provided below.
• A recent CHS analysis examined asthma incidence in relation to improved air quality in nine
southern California communities (Garcia et al.. 2019). Decreases in baseline ozone concentrations
in three CHS cohorts, enrolled in 1993, 1996, and 2006, were associated with decreased asthma
incidence. The findings indicate that improved air quality is associated with lower asthma
incidence. The magnitude and precision of the observed association was comparable in a model
adjusting for local near-road pollution. Due to modeling constraints, the authors used 1-year
ozone concentrations at baseline.
• In analyses of a large administrative database birth cohort in Quebec, an increase in average
summertime ozone concentrations at participants' birth addresses were associated with a 19%
increase (95% CI: 16, 23%) in asthma onset in children of all ages (Tetreault et al.. 2016a). The
case definition used for new-onset asthma included a single hospital discharge with a diagnosis of
asthma or two or more physician claims for asthma, which does not necessarily distinguish
exacerbation of mild/undiagnosed asthma from asthma onset. In the event that the case definition
is not capturing some asthma cases, the observed effects could be partially reflective of an acute
outcome. In this case, the observed effects could be partially confounded by short-term increases
in air pollution or other temporally-varying covariates, such as temperature. Notably, the
associations were present at low concentrations, and were robust in sensitivity analyses that used
time-varying ozone concentrations or relied on a more stringent case definition for children under
five, an age group with less reliable asthma diagnoses.
• In contrast, a pooled retrospective case-control analysis of minority children in the U.S. reported
null associations between early-life ozone exposure and asthma incidence (Nishimura et al..
2013). The study was much smaller than Tetreault et al. (2016a) and consequently had less
precision (i.e., wider 95% CIs).
• Results from a previous CHS analysis (Bastain et al.. 2011) showed that elevated eNO was
associated with increased risk of asthma development in children. However, Berhane et al. (2014)
examined airway inflammation in response to long-term ozone exposure in the CHS cohort and
observed a null association between ozone and changes in eNO in children.
In summary, recent studies provide support for an association between long-term ozone exposure
and asthma development in children. While one study presented contrasting evidence, the authors focused
on a specific at-risk population and the study included fewer participants (Nishimura et al.. 2013).
3-94
-------
3.2.4.1.2
Animal Toxicological Studies
The 2013 Ozone ISA summarized the animal toxicological evidence of the development of
asthma resulting from ozone exposure during the early postnatal period (U.S. EPA. 2013a). Several
studies found that cyclic challenge of infant rhesus monkeys to allergen and ozone during the postnatal
period compromised airway growth and development and resulted in changes that favor allergic airways
responses, increased airway responsiveness, and persistent effects on the immune system. Rhesus
monkeys were chosen as a model because the branching pattern and distribution of airways in rhesus
monkeys are more similar to humans than those of rodents. In addition, a model of allergic airways
disease, which exhibits the main features of human asthma, had already been established in the adult
rhesus monkey. Studies in infant monkeys were designed to determine whether repeated exposure to
ozone altered postnatal lung growth and development, and if so, whether such effects were reversible. In
addition, exposure to ozone was evaluated for its potential to increase the development of allergic airways
disease. The animals were exposed episodically to ozone beginning at 1 month of age. The exposure
regimen involved biweekly cycles of alternating filtered air and ozone (i.e., 9 consecutive days of filtered
air and 5 consecutive days of 0.5 ppm ozone, 8 hour/day) and to house dust mite allergen (HDMA) for
2 hours per day for 3 days on the last 3 days of ozone exposure, followed by 11 days of filtered air. In
most of these studies, infant monkeys were sensitized to HDMA before the start of the cyclical exposures.
These animals exhibited the hallmarks of allergic asthma 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 was 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.
The infant monkey studies reported numerous key findings (U.S. EPA. 2013a). Baseline airway
resistance and airway responsiveness to inhaled histamine were dramatically increased by combined
exposure to ozone plus HDMA. This finding suggests that long-term ozone exposure may contribute to
the effects of asthma in children. A follow-up study assessing ex vivo airway responsiveness of the infant
monkeys found that ozone plus HDMA exposure resulted in increased airway responsiveness in the
respiratory bronchioles, where dosimetric models indicated that the dose would be higher. In another
study, the growth pattern of distal airways was changed to a large extent by exposure to ozone alone and
in combination with HDMA. More specifically, the airways became longer and narrower and the number
of conducting airway generations between the trachea and the gas exchange area was decreased. This
effect was not ameliorated by a recovery period of 6 months in filtered air. Other structural changes
included increases in mucus goblet cell mass and alterations in smooth muscle orientation in the
respiratory bronchioles, epithelial nerve fiber distribution, and basement membrane zone morphometry,
all of which could potentially contribute to airway obstruction and increased airway responsiveness.
Additional effects on neural innervation in the epithelium of the conducting airways were observed in
response to ozone alone or ozone plus HDMA, including decreased nerve fiber density and altered nerve
3-95
-------
bundle morphology. Six months of recovery in filtered air led to reversal of some, but not all, of these
structural and functional effects.
As described in the 2013 Ozone ISA (U.S. EPA. 2013a). exposure to ozone also resulted in
increased number and proportion of eosinophils and decreased number of neutrophils and lymphocytes in
BALF and the blood of infant monkeys. These effects were not evident after a 6-month recovery period in
filtered air. Challenge with LPS, which activates monocytes and other innate immune cells, elicited a
lower response in ozone-exposed animals. While increased airway eosinophilia suggests an increased
allergic profile, the decreased response to LPS suggests diminished host defenses. Other effects on the
developing immune system of exposure to ozone plus HDMA included increased CD4+ and CD8+
lymphocytes in blood and BALF and activated lymphocytes (CD25+ cells) in airway mucosa.
The effect of cyclic episodic ozone exposure on nasal airways was also studied in the infant
rhesus monkey (U.S. EPA. 2013a). The three-dimensional detail of the nasal passages was analyzed for
developing predictive dosimetry models and exposure dose-response relationships. The relative amounts
of the five epithelial cell types in the nasal airways remained consistent between infancy and adulthood.
Ozone exposure resulted in 50-80% decreases in epithelial thickness and epithelial cell volume of the
ciliated respiratory and transitional epithelium, confirming that these cell types in the nasal cavity were
the most sensitive to ozone exposure. The character and location of nasal lesions were similar in infant
and adult monkeys that were similarly exposed. However, the nasal epithelium of infant monkeys did not
undergo nasal airway epithelial remodeling or adaptation which occurs in adult animals following
ozone-mediated injury and which may protect against subsequent ozone challenge. This lack of
remodeling suggests the potential for developing persistent necrotizing rhinitis following longer term
exposure.
Recent studies have examined a wide range of effects of postnatal ozone exposure. Several
studies were conducted in infant monkey model of allergic airway in which monkeys were sensitized to
HDMA. Several other studies were conducted in nonallergic infant monkeys and neonatal rodents. A brief
discussion of their findings is found below, with studies grouped according to the animal model
employed. Study-specific details are summarized in Table 3-44. Table 3-45. Table 3-46. and Table 3-47
in Section 3.3.2. All of the effects described below were statistically significant. Recent studies add to
previous evidence that postnatal ozone exposure may lead to the development of asthma by
compromising airway growth and development, promoting the development of an allergic phenotype and
increased airway responsiveness, and causing persistent alterations of the immune system.
Recent studies in the infant monkey model of allergic airway disease demonstrated airway
smooth muscle hyperreactivity, an enhanced allergic phenotype, and priming of responses to oxidant
stress as a result of postnatal ozone exposure.
• Postnatal ozone exposure increased airway smooth muscle contraction mediated by serotonin
(Moore et al.. 2012a). Exposures were to episodic ozone beginning at 1 month of age. This
involved biweekly cycles of alternating filtered air and ozone (i.e., 9 consecutive days of filtered
3-96
-------
air and 5 consecutive days of 0.5 ppm ozone, 8 hour/day) and to house dust mite allergen
(HDMA) for 2 hours per day for 3 days on the last 3 days of ozone exposure, followed by 11 days
of filtered air. Eleven cycles of episodic ozone exposure (0.5 ppm) did not increase airway
responsiveness to histamine in vivo or airway responsiveness to acetylcholine in an in vitro study
using tracheal rings. In fact, responsiveness to acetylcholine was decreased by postnatal ozone
exposure and exogenous serotonin. When electrical field stimulation, which causes the release of
acetylcholine from airway nerves, was used to test the properties of the airway smooth muscle in
the tracheal rings model, the response was not increased in the ozone or ozone/HDMA groups.
But when exogenous serotonin was added, the electrical field stimulation response was enhanced
in the ozone and ozone/HDMA groups. Airway smooth muscle contraction to electrical field
stimulation is a measure of post-ganglionic and parasympathetic-mediated processes.
Experiments with receptor agonists and antagonists found that postnatal ozone exposure
enhanced the excitatory parasympathetic pathway in the presence of serotonin. Postnatal exposure
to HDMA also had this effect and it enhanced intrinsic airway smooth muscle contractility. The
result of postnatal exposure to ozone and HDMA was a hyperresponsive airway. This study
provides evidence of a functional change in the airways due to postnatal ozone exposure. The
presence of the neurotransmitter serotonin is required for the enhanced airway smooth muscle
contraction. Given that previous work from this laboratory showed increased serotonin positive
cells in the airway resulting from postnatal ozone exposure, the development of asthma in this
model could be due to increased serotonin acting on post-ganglionic parasympathetic fibers to
increase airway responsiveness.
• Two other recent studies found that postnatal exposure to ozone and HDMA altered immune
system development, including effects on eosinophils that are characteristic of an allergic
phenotype. In one study (Chou etal.. 2011). episodic ozone exposure (5 biweekly cycles,
0.5 ppm) resulted in decreased total white blood cells and blood eosinophils and increased BALF
eosinophils. However, ozone exposure did not lead to an increase in airway mucosa eosinophils.
Ozone/HDMA exposure increased BALF eosinophils and eotaxin, but no increase in airway
mucosa eosinophils was found. In the study by Crowley et al. (2017). episodic ozone exposure
(11 biweekly cycles, 0.5 ppm ozone) decreased numbers of monocytes and frequency of
eosinophils in peripheral blood and increased the frequency of eosinophils in BALF. Exposure to
ozone and HDMA had similar effects on blood monocytes, immune system development,
including effects on eosinophils which are characteristic of an allergic phenotype.
• Postnatal ozone exposure primed the airway for an enhanced response to oxidant stress (Murphy
et al.. 2012). In this study, episodic ozone exposure (11 biweekly cycles, 0.5 ppm) enhanced
responses of airway explant tissue to an exogenous oxidant, including increased expression of
IL-8, a neutrophil chemokine, and increased expression of neurokinin-1 receptor, which is
involved in a nonapoptotic pathway of cell death.
Recent studies in nonallergic infant monkeys demonstrated increased serotonin-positive airway
cells and immunomodulation as a result of postnatal ozone exposure.
• Acute exposure to ozone (0.5 ppm, 8 hours) altered serotonin expression in a specific region of
the developing lung (Murphy et al.. 2013). Serotonin-positive cells were increased in midlevel
airways in 2-month-old monkeys. Serotonin is a neurotransmitter involved in airway smooth
muscle contraction.
• Episodic exposure to ozone (11 biweekly cycles of 0.5 ppm) altered innate immune function in
airway epithelia (Clay et al.. 201.4). Ozone exposure attenuated the inflammatory response to
LPS, measured in an in vitro assay as gene expression of the cytokines IL-6 and IL-8. In contrast,
ozone-exposed infant monkeys subsequently challenged with LPS in vivo exhibited increased
3-97
-------
IL-6 and IL-8 gene expression in response to LPS in the in vitro assay. These results suggest that
early life exposure to ozone is immunomodulatory.
• Episodic ozone exposure decreased expression of components of a nonapoptotic cell death
pathway that had been increased by a single acute ozone exposure (Murphy et al.. 2014). In this
study, ozone exposure consisted of 1 or 11 biweekly cycles (0.5 ppm) plus or minus an acute
ozone exposure of 0.5 ppm for 8 hours. Results were compared with exposure to filtered air plus
or minus acute ozone exposure. Components that were upregulated by acute ozone exposure
included TAC1, which is the substance P precursor, neurokinin-1 receptor, and nuclear
receptor 77. These same components were downregulated by episodic/acute ozone exposure in
the same specific regions of the developing lung in which they were upregulated by a single acute
ozone exposure.
Recent studies in rats demonstrated impaired airway growth and altered airway sensory nerve
innervation as a result of postnatal ozone exposure.
• Persistent changes in airway architecture resulted from early postnatal exposure to ozone (Lee et
al.. 2011). Rats were exposed for 3 weeks to ozone (0.5 ppm x 6 hour/day) beginning on
Postnatal Day (PND) 7. After a 56 days recovery period, diameter, length, and branching angle of
conducting airways were examined. Decreased diameter and airway length were demonstrated for
airway generations 7-22 and 16-20, respectively. These changes occurred when ozone exposures
consisted of 5 days on and 2 days off per week, but not when ozone exposures consisted of 2 days
on and 5 days off per week. This study suggests that early postnatal ozone exposure may impact
airway resistance through effects on airway architecture.
• Two studies from the same laboratory examined effects of early postnatal ozone exposure on
airway innervation in rats. Zellner et al. (2011) provides evidence of altered sensory neuron
development. In rats, exposure to ozone (2 ppm for 3 hours) on Postnatal Day 5 resulted in
decreased total neuron number at Postnatal Day 21, but no effect on substance P-containing
neurons in the developing airway. Hunter etal. (2011) demonstrated that ozone exposure (2 ppm
for 3 hours) on Postnatal Day 6 enhances the production of nerve growth factor by airways in
response to a later ozone exposure on Postnatal Day 28. Postnatal ozone exposure also increased
numbers of BALF neutrophils. Nerve growth factor may link ozone exposure to an increase in
substance P-containing nerve fibers in the airways. Taken together, these two studies indicate that
early postnatal ozone exposure affects the number, and possibly the type, of neurons in the
developing airway.
3.2.4.1.3 Integrated Summary for Development of Asthma
The 2013 Ozone ISA (U.S. EPA. 2013a) presented evidence of ozone modified associations
between exercise and asthma incidence in children, and interactions between ozone and different genetic
variants on associations with childhood asthma. A recent large administrative cohort study provides
evidence of an association between long-term ozone exposure and asthma onset in children. This finding
is supported by a CHS analysis that reported a decrease in childhood asthma incidence associated with
decreases in ozone concentrations. A smaller cohort study focusing on minority children found a null
association, but the larger studies provide compelling evidence of a positive association. Recent animal
toxicological studies in rodents and monkeys support the epidemiologic results and findings from
previous toxicological studies that postnatal ozone exposure may lead to the development of asthma by
3-98
-------
compromising airway growth and development, promoting the development of an allergic phenotype and
increased airway responsiveness, and causing persistent alterations to the immune system.
3.2.4.2 Lung Function and Development
After organogenesis in the embryonic stage, the development of the human lung continues
throughout the fetal period and into early adulthood (Schittnv. 2017). This continued development
comprises an extended window of potential vulnerability to environmental stressors, such as ozone. To
characterize lung health, lung function metrics capture the cumulative effects of pulmonary growth,
damage, and repair (Wang et al.. 1993). Accordingly, measures of lung function are effective indicators of
pulmonary effects related to exposure to environmental stressors.
3.2.4.2.1 Epidemiologic Studies
Epidemiologic studies evaluated in the 2013 Ozone ISA provided inconsistent evidence of an
association between long-term exposure to ozone and lung development in children (U.S. EPA. 2013a). In
an 8-year follow-up of the CHS cohort, Gauderman et al. (2004) observed a null association between
mean annual 8-hour ozone concentrations and deficits in lung function growth (FEVi). In contrast, in a
subsequent CHS analysis, Breton etal. (2011) reported ozone-related deficits in 8-year lung function
growth among children without a particular GSS glutathione gene haplotype. Cross-sectional studies of
ozone and lung function in children or adults were similarly inconsistent.
A limited number of recent studies in the U.S. continue to provide inconsistent evidence of an
association between ozone and lung development or lung function. Study-specific details, including air
quality characteristics and select effect estimates, are highlighted in Table 3-48 in Section 3.3.2. An
overview of the evidence is provided below.
• An extended follow-up of the CHS combined data obtained from three separate cohorts to
examine the association between long-term reductions in air pollution and lung development in
children between the ages of 11 and 15 (Gilliland et al.. 2017; Gauderman et al.. 2015). The
authors did not observe a notable change in lung function growth or cross-sectional lung function
corresponding to decreasing ozone concentrations.
• Other cross-sectional studies reported modest decreases in lung function metrics associated with
ozone, including a pooled retrospective case-control analysis of minority children with asthma in
the U.S. (Neophvtou et al.. 2016) and another analysis of a recent CHS cohort that overlaps with
one of the cohorts included in the Gauderman et al. (2015) study (Urman et al.. 2014).
• While cross-sectional studies of adult lung function evaluated in the 2013 Ozone ISA provided
inconsistent evidence of an association with ozone (Forbes et al.. 2009; Qian et al.. 2005). a
recent longitudinal study of lung function in older adults in the U.S. reported decrements in FEVi
and FVC relative to ozone concentrations (Eckel et al.. 2012).
3-99
-------
In summary, a limited number of recent studies continue to provide inconsistent evidence of an
association between long-term ozone exposure and lung development or lung function in children. While
the only recent study that examined lung function in adults observed evidence of an association, this
result should be considered in the context of inconsistent evidence presented in the 2013 Ozone ISA (U.S.
EPA. 2013a).
3.2.4.2.2 Animal Toxicological Studies
The 2013 Ozone ISA summarized the animal toxicological evidence of altered lung function and
development resulting from ozone exposure during both the prenatal and early postnatal periods. These
studies are described above in Section 3.2.4.1.2 because they found evidence for compromised airway
development in the infant rhesus monkey exposed episodically to ozone. In addition, maternal exposure to
0.8-1.2 ppm ozone during gestation resulted in developmental health effects, mainly related to immune
function and allergic lung disease, in the respiratory tract of offspring mice. Recent studies include
several in the infant monkey model of allergic airway disease in which monkeys were sensitized to
HDMA and several in rodents of varying ages. These studies examined a wide range of effects of
postnatal ozone exposure. A brief discussion of their findings is found below, with studies grouped
according to the animal model employed. Postnatal ozone exposure resulted in altered lung development
in the infant monkeys and increased oxidative stress, inflammation, and injury in neonatal rodents. Effects
on lung function parameters were found in rodents of different ages following long-term exposure to
ozone.
Recent studies examined alveolar morphogenesis in a model of allergic airways disease using
infant monkeys that were sensitized to HDMA. This model shares many features with childhood asthma.
Alveolar morphogenesis is the process by which alveoli are formed de novo in the lower respiratory tract
during lung development. Results of these studies demonstrating statistically significant effects provide
evidence that postnatal ozone exposure leads to impairment of alveolar morphogenesis. Study-specific
details are summarized in Table 3-49 in Section 3.3.2.
• In Avdalovic et al. (2012). episodic ozone exposure resulted in altered alveolar morphogenesis.
Exposures to episodic ozone began at 1 month of age. This involved biweekly cycles of
alternating filtered air and ozone (i.e., 9 consecutive days of filtered air and 5 consecutive days of
0.5 ppm ozone, 8 hours/day) and to house dust mite allergen (HDMA) for 2 hours per day for
3 days on the last 3 days of ozone exposure, followed by 11 days of filtered air. Five cycles of
episodic ozone exposure (0.5 ppm) resulted in decreased alveolar number, increased alveolar
volume, decreased distribution of alveolar volume, and decreased capillary surface density in
infant monkeys that were sensitized to HDMA. These changes reflect reduced alveolarization,
which was also seen in infant monkeys exposed to 5 cycles of episodic ozone and HDMA
challenge (ozone/HDMA). However, these changes were not seen after 11 cycles of episodic
ozone exposure. Instead, increases in lobe volume were seen in the HDMA/ozone group,
suggesting that a "catch-up" phase of alveolarization had occurred. Changes in TGF-(3 gene
3-100
-------
expression in lung parenchyma occurred between 5 and 11 cycles of episodic ozone exposure in
ozone and ozone/HDMA groups, suggesting that TGF-(3 played a role in the later alveolarization.
• In a follow-up study, Herring et al. (2015) demonstrated additional effects on alveolar
morphogenesis. Eleven cycles of episodic ozone exposure (0.5 ppm) resulted in decreased
numbers of alveoli in the right middle lobe in ozone/HDMA group. After a 30-month recovery
period, the number of alveoli in the right middle lobe was increased in the ozone/HDMA group
compared with controls. The coefficient of variation of distribution of mean number-weighted
alveolar volumes and ratio of pulmonary capillary to inter-alveolar septal surface in the left
cranial lobe was also increased after a 30-month recovery period in the ozone/HDMA group
compared with controls. This indicates that alveoli that formed during the 30-month recovery
period were smaller and had a greater capillary-to-alveolar gas-exchange surface and suggests a
potentially greater susceptibility to obstructive lung disease.
In addition, recent studies found respiratory tract oxidative stress, inflammation and injury in
neonatal rodents exposed to 1 ppm ozone during the early postnatal period. The effects described below
were statistically significant. Study-specific details are summarized in Table 3-46 in Section 3.3.2.
• Dve et al. (2017) demonstrated injury and oxidative stress-related responses to ozone exposure
(1 ppm, 2 hours) in rats at Postnatal Days 14, 21, and 28. These changes included increased lung
wet weight:body weight ratio, altered levels of the antioxidants uric acid and glutathione, and
altered activities of the antioxidant enzymes superoxide dismutase, glutathione peroxidase, and
glucose-6-phosphate dehydrogenase. These changes were dependent on sex, age, and strain, with
activities of antioxidants and antioxidant enzymes decreasing in younger animals and increasing
in older animals in response to ozone exposure.
• Two studies from the same laboratory (Gabehart et al.. 2015; Gabehart et al.. 2014) found
inflammation, injury, and oxidative stress-related responses to ozone exposure (1 ppm, 3 hours)
in mice at Postnatal Day 1 and 7. This included increases in metallothionein I, heme oxygenase 1,
and chemokine gene expression and increases in BALF neutrophils. In addition, secretion and
expression of mucus, which can offer protection against injury, were increased. Neutrophil and
chemokine responses to ozone exposure were inhibited in toll receptor 4-deficient mice. Cell
proliferation was decreased by ozone exposure. Responses to ozone exposure in 2-, 3-, and
6-week-old mice (i.e., juvenile, weanling, and adult lifestages) are reported elsewhere in this
document. In general, responses were smallest in 1-week-old and greatest in 6-week-old mice,
except for effects on mucus, which were found only in the 1-week-old mice. Toll receptor 4
expression was found to increase with age, suggesting that toll receptor 4 pathway may underlie
responses to ozone that were more pronounced in adult compared with neonatal mice.
Three other recent studies examined the effects of long-term ozone exposure on lung function in
rodents of varying lifestages. Results of these studies demonstrate that subchronic exposure to
0.5-1.0 ppm ozone alters ventilatory parameters. The effects described below were statistically
significant. Study-specific details are summarized in Table 3-50 in Section 3.3.2.
• In Gordon et al. (2013). adult and senescent rats were exposed for 15 weeks to ozone (0.8 ppm
for 6 hours/day, once a week). Effects on ventilatory parameters included increased respiratory
frequency, decreased tidal volume, increased minute volume, and increased enhanced pause.
Several of these effects persisted 1 week after the exposure had ceased. In general, larger
responses to ozone exposure were seen in the adult rats compared with the senescent rats.
3-101
-------
• In Snow et al. (2016). adolescent, young adult, adult, and senescent rats were exposed for
13 weeks to ozone (0.25 and 1.0 ppm for 6 hours/day, twice a week). No effects on ventilatory
parameters were observed at 0.25 ppm. Relaxation time was decreased in senescent rats. Minute
volume and enhanced pause were increased in young adult rats; these effects were largely
resolved after 5 recovery days.
• In Gordon et al. (2016a). rats that were exercise trained or sedentary were exposed for 6 weeks to
ozone (0.25, 0.5, and 1.0 ppm for 5 hours/day, once a week). Exposure to 1.0 ppm ozone
increased enhanced pause in sedentary but not exercise-trained rats.
3.2.4.2.3 Integrated Summary for Lung Function and Development
The 2013 Ozone ISA (U.S. EPA. 2013a) described inconsistent epidemiologic evidence that
long-term exposure to ozone is associated with lung function development in children. Recent
epidemiologic studies continue to provide limited support for an association between long-term ozone
exposure and lung function development in children. A CHS cohort study in the 2013 Ozone ISA
reported ozone-related impairment in lung function development in children without a particular GSS
glutathione gene haplotype; however, recent studies have not examined similar genetic variants.
Additionally, while a limited number of recent epidemiologic studies of long-term ozone exposure and
lung function development in children are consistently null, cross-sectional studies of children and adults
have observed some evidence of an association between long-term ozone concentrations and lung
function.
In contrast to the limited and inconsistent evidence from epidemiologic studies, recent
experimental studies in animals provide evidence that postnatal ozone exposure may affect the developing
lung. Results from studies of neonatal rodents demonstrate ozone-induced injury and changes in
inflammatory and oxidative stress responses during lung development. In an infant monkey model with
similarities to childhood asthma, postnatal ozone exposure resulted in impaired alveolar morphogenesis, a
key step in lung development. Notably, these studies indicated some capacity for repair. Additional
studies in adult rats suggest that chronic ozone exposure may alter ventilatory parameters.
3.2.4.3 Development of Chronic Obstructive Pulmonary Disease and Other
Associated Respiratory Effects
Chronic obstructive pulmonary disease (COPD) is a lung disease characterized by persistent
respiratory symptoms and airflow limitation due to destruction of alveolar tissue and airway remodeling.
Reduced airflow is associated with decreased lung function, and clinical symptoms demonstrating
exacerbation of COPD include cough, sputum production, and shortness of breath.
3-102
-------
3.2.4.3.1
Epidemiologic Studies
There were no epidemiologic studies examining the association between long-term exposure to
ozone and COPD available for inclusion in the 2013 Ozone ISA (U.S. EPA. 2013a). One recent study
used the Ontario Asthma Surveillance Information System to identify adults with asthma and found that
ozone was associated with an increase in the odds of COPD incidence in this population (To et al.. 2016).
The association was attenuated and less precise (i.e., wider 95% CIs) but still positive in copollutant
models adjusted for PM2 5. Further discussion of potential copollutant confounding of the relationship
between respiratory effects and long-term exposure to ozone can be found in the "Relevant Issues for
Interpreting Epidemiologic Evidence" section (Section 3.2.5). Additional study-specific details from To et
al. (2016). including air quality characteristics, are highlighted in Table 3-51 in Section 3.3.2.
3.2.4.3.2 Animal Toxicological Studies
The 2013 Ozone ISA (U.S. EPA. 2013a) summarized the animal toxicological evidence of
respiratory effects in healthy populations resulting from exposure to ozone. While most of the studies
were conducted in rodents, a few involved chronic exposures to ozone in adult or young adult monkeys.
Chronic ozone exposure (0.12-1 ppm) resulted in damage to the distal airways and proximal alveoli,
resulting in persistent inflammation and lung tissue remodeling, leading to irreversible changes. Some
studies demonstrated increased collagen synthesis and deposition, inducing fibrotic-like changes in the
lung. Changes in other components of the extracellular matrix, such as glycosaminoglycans, were
reported. Some of these effects were found to be dependent on the TGF-|3 signaling pathway. Other
studies demonstrated emphysematous-like changes or attenuation of inflammation. Thus, chronic ozone
exposure may lead to persistent inflammation and interstitial remodeling that may contribute to the
progression and development of chronic lung disease, such as pulmonary fibrosis and COPD. In addition,
chronic ozone exposure (0.12-0.5 ppm) is capable of damaging nasal airways resulting in altered
structure and function, as demonstrated by increased mucus flow and goblet cell metaplasia.
Several recent studies examined the effects of repeated ozone exposure on airway inflammation
and injury in rodents of varying lifestages. These studies demonstrate that subchronic exposure to
0.5-1.0 ppm ozone resulted in airway injury and inflammation. All of the changes described below were
statistically significant. While most studies were conducted in male rats, one study found injury and
inflammatory effects in both male and female rats. Some of these effects were dependent on the age of the
animal or whether it was exercise-trained and some of these effects resolved following a 5-day recovery
period. Study-specific details are summarized in Table 3-46 in Section 3.3.2.
• In Gordon et al. (2013). adult and senescent rats were exposed for 17 weeks to ozone (0.8 ppm
for 6 hours/day, once a week). Ozone exposure increased a marker of injury (albumin) and
decreased the number of macrophages in BALF. Epithelial hyperplasia of terminal bronchioles,
alveolar ducts, and adjacent alveoli were observed to a comparable degree in the adult and
senescent rats.
3-103
-------
• In Gordon et al. (2016b). male and female rats were exposed for 4 weeks to ozone (0.8 ppm for
5 hours/day, once a week). Ozone exposure increased a marker of injury (albumin) and a marker
of inflammation (eosinophils) in BALF of male and female rats.
• In Miller et al. (2016a). rats were exposed for 13 weeks to ozone (0.25 and 1.0 ppm for
5 hours/day, three times/week). No effects on inflammation or injury were observed in response
to 0.25 ppm ozone. Markers of injury (protein, albumin, /v'-acctyl-glutaminidasc) and
inflammation (neutrophils and alveolar macrophages) were increased in the BALF in response to
1.0 ppm ozone. Most of these effects were lost after 5 recovery days.
• In Snow et al. (2016). adolescent, young adult, adult, and senescent rats were exposed for
13 weeks to ozone (0.25 and 1.0 ppm for 6 hours/day, twice a week). No effects on inflammation
or injury were observed at 0.25 ppm. A marker of inflammation (BALF total cell number) was
increased in young adult rats exposed to 1 ppm ozone.
• In Gordon et al. (2016a). rats that were exercise-trained or sedentary were exposed for 6 weeks to
ozone (0.25, 0.5, and 1.0 ppm for 5 hours/day, once a week). No effects on inflammation or
injury were observed at 0.25 ppm. Exposure to 0.5 ppm ozone increased markers of injury (BALF
protein and albumin) in exercise-trained rats. Exposure to 1.0 ppm ozone increased inflammatory
markers in sedentary and exercise-trained rats, with more pronounced effects in sedentary rats.
3.2.4.3.3 Integrated Summary for Development of Chronic Obstructive Pulmonary Disease
and Other Associated Respiratory Effects
The 2013 Ozone ISA did not evaluate any epidemiologic studies that examined the relationship
between long-term exposure to ozone and the development of COPD. One recent epidemiologic study
provides evidence of an association between long-term ozone concentrations and incident COPD
hospitalizations. Animal toxicological studies reviewed in the 2013 Ozone ISA found that chronic ozone
exposure can damage the distal airways and proximal alveoli, resulting in persistent inflammation and
lung tissue remodeling that leads to irreversible changes including fibrotic- and emphysematous-like
changes in the lung. Additionally, recent animal toxicological studies provide consistent evidence that
subchronic ozone exposure can lead to airway injury and inflammation. In adult animals these changes
may underlie the progression and development of chronic lung disease and provide biological plausibility
for ozone-induced development of COPD.
3.2.4.4 Respiratory Infection and Other Associated Respiratory Effects
3.2.4.4.1 Epidemiologic Studies
There were no epidemiologic studies examining the association between long-term exposure to
ozone and respiratory infection available for inclusion in the 2013 Ozone ISA (U.S. EPA. 2013a). Two
recent studies observed inverse associations between ozone and respiratory infection. Smith et al. (2016)
reported an inverse association between 2-year avg ozone concentrations and pulmonary tuberculosis in a
3-104
-------
nested case-control study of adults in northern California. The authors did observe a strong positive
association with NO2 and a negative correlation between ozone and NO2, which may explain the inverse
association. In a study of otitis media in the first 2 years of life, 2-month avg ozone concentrations were
associated with decreased risk of infection (Maclntvrc et al.. 201IV Study-specific details, including air
quality characteristics and select effect estimates, are highlighted in Table 3-52 in Section 3.3.2.
3.2.4.5 Severity of Respiratory Disease
3.2.4.5.1 Epidemiologic Studies
Respiratory symptoms and increased medication use are often indicators of disease severity.
Symptom frequency is used as a measure of asthma severity, in particular. Additionally, more severe
symptoms, potentially resulting in hospitalization or ED visits for asthma, are indicative of exacerbation
severity, but may also suggest greater underlying disease severity (NAEPP. 2008). While the 2013 Ozone
ISA did not have a delineated discussion of epidemiologic studies that examined severity of respiratory
diseases, a limited number of relevant studies were evaluated and provided evidence of associations
between long-term ozone concentrations and respiratory hospital admissions or symptoms (U.S. EPA.
2013a). Specifically, an ecological study of asthma hospital admissions in children (Moore et al.. 2008)
and a cross-sectional study of adults (Mcng et al.. 2010) in California observed associations with
long-term exposure to ozone. Notably, Meng et al. (2010) also reported an association between ozone
concentrations and self-reported asthma symptoms. Similarly, McConnell et al. (2003) observed that
bronchitic symptoms in children with asthma were associated with yearly variation in ozone within CHS
communities, but not with 4-year avg ozone across communities. The longitudinal nature of the
within-community estimate makes it more informative than the cross-sectional between-group estimate
because it establishes a temporal relationship between the exposure and outcome. Another cross-sectional
study, in the U.K., reported that ozone concentrations (annual accumulated ozone over 40 ppb per
daylight hour) were associated with more severe emphysema, as measured by a density mask analysis of a
CT Scan (Wood et al.. 2009).
Recent studies further support a relationship between long-term exposure to ozone and severity of
respiratory disease, although some uncertainties remain. Study-specific details, including air quality
characteristics and select effect estimates, are highlighted in Table 3-53 in Section 3.3.2. An overview of
the evidence is provided below.
• In a large administrative database cohort of adults with asthma in Quebec, summertime average
ozone was associated with aggregated hospital admissions and ED visits for asthma (Tetreault et
al.. 2016b). The hazard ratio (HR) was positive (1.17; 95% CI: 1.12, 1.22) when time-dependent
ozone concentrations were used to estimate exposure, but not when ozone exposure was assigned
at birth residences (0.99; 95% CI: 0.96, 1.11). This may indicate that asthma exacerbation is
related to continued exposure to ozone, rather than prenatal and 1st year of life exposure.
3-105
-------
However, as the authors do not adjust for short-term exposures to ozone, the time-dependent
model could be capturing acute responses to ozone.
• Like McConnell et al. (2003). Berhane et al. (2016) and Gilliland et al. (2017) observed
decreased prevalence of bronchitic symptoms in children with asthma associated with decreased
ozone exposure over two decades of follow-up of CHS cohorts. The association was attenuated
but still present in copollutant models with NO2 or PM2 5. Further discussion of potential
copollutant confounding of the relationship between respiratory effects and long-term exposure to
ozone can be found in the "Relevant Issues for Interpreting Epidemiologic Evidence" section
(Section 3.2.5).
• In the study of asthma hospital admissions discussed previously, Moore et al. (2008) used
marginal structural models with G-computation parameter estimation to estimate the causal effect
of quarterly ozone exposure on hospital discharges for asthma in children. In a recent
methodological follow-up study, Moore et al. (2012b) used causal models for realistic
individualized exposure rules (CMRIER) to account for bias resulting from improper control for
measured confounders due to a lack of observed exposures for subjects in some strata. In other
words, CMRIER attempts to account for violations of the experimental treatment assignment
assumption, whereby all exposure levels are assumed to be observable for subjects within all
strata (i.e., groups defined by potential confounders), such that proper within-group comparisons
can be made. The application of CMRIER resulted in a larger, but much less precise (i.e., wider
95% CIs) association between asthma-related hospital discharges and quarterly ozone
concentrations. The greatly reduced precision may reflect the new modeling approach or could be
the result of reduced power to detect an association due to a restricted sample size consisting of
41% of the geographic units used in the initial analysis and a less efficient parameter estimation
method.
In summary, recent studies continue to provide support for an association between ozone and
respiratory disease severity. However, notable uncertainties remain. Some studies examine hospital
admissions or ED visits as a measure of disease severity but do not control for short-term exposures to
ozone. Given the acute nature of the health endpoint, the observed effects could be confounded by
short-term increases in air pollution. There is also a lack of available studies that examine potential
copollutant confounding. However, one study observed an association between ozone and bronchitic
symptoms in children with asthma that is robust to adjustment for PM2 5 and NO2. Additionally, results
from causal inference studies suggest uncertainty in the estimated confidence in the observed
associations.
3.2.4.5.2 Animal Toxicological Studies
The 2013 Ozone ISA (U.S. EPA. 2013a) summarized the animal toxicological evidence related to
severity of disease resulting from exposure to ozone. A 4-week exposure to ozone (0.5 ppm for 5 hours,
once a week) enhanced injury, inflammation, and allergic responses in a rodent model of allergic airway
disease. In addition, 4 months exposure (0.5 ppm) resulted in increased severity of post-influenza
alveolitis and lung parenchymal changes. No additional studies have become available since then.
3-106
-------
3.2.4.5.3
Integrated Summary for Severity of Respiratory Disease
Results from recent epidemiologic studies are consistent with evidence evaluated in the 2013
Ozone ISA that provides support for an association between ozone and respiratory disease severity.
Specifically, there is consistent evidence that long-term exposure to ozone is associated with hospital
admissions and ED visits for asthma and prevalence of bronchitic symptoms in children with asthma.
There is some uncertainty due to the acute nature of some of these outcomes. Additionally, while there
are no recent animal toxicological studies available for review, a previously evaluated study provides
biological plausibility for enhanced respiratory effects in populations with pre-existing respiratory
conditions.
3.2.4.6 Allergic Responses
3.2.4.6.1 Epidemiologic Studies
The 2013 Ozone ISA reviewed a limited number of epidemiologic studies examining a range of
allergic indicators that found generally positive associations with long-term exposure to ozone (U.S. EPA.
2013a). Cross-sectional studies reported increases in prevalence of hay fever (Parker et al.. 2009) and
rhinitis (Hwang et al.. 2006; Penard-Morand et al.. 2005). and increased total serum IgE levels (Rage et
al.. 2009) associated with ozone concentrations. In copollutant models adjusting for NO2, the observed
association between ozone and rhinitis was persistent (Penard-Morand et al.. 2005). while the association
with IgE levels was attenuated, but still positive (Rage et al.. 2009). In contrast to generally consistent
evidence of an association, one study reported null associations between ozone and hay fever (Ramadour
et al.. 2000).
One recent cross-sectional study provides additional support for an association between long-term
exposure to ozone and allergic response. A 2005-2006 NHANES analysis, comprising a nationally
representative sample of the U.S. population, examined allergic sensitization measured by detectable
allergen-specific IgE levels (Weir et al.. 2013). Weir et al. (2013) found that annual average ozone
concentrations were associated with increased odds of sensitization to indoor allergens and inhalants. The
observed ORs were comparable for exposure assigned from monitors within 20 miles of the participants'
home address and using geocoded CMAQ ozone concentration estimates. The authors did not present
models adjusted for copollutants, and while limited evidence from the previous ozone ISA indicated that
associations were persistent to adjustment forN02, potential copollutant confounding remains an
uncertainty, specifically regarding potential confounding by pollen levels. Complete study details,
including air quality characteristics, are highlighted in Table 3-54 in Section 3.3.2.
3-107
-------
3.2.4.6.2
Animal Toxicological Studies
The 2013 Ozone ISA (U.S. EPA. 2013a) summarized the animal toxicological evidence of
allergic responses resulting from exposure to ozone. A 4-week exposure to ozone (0.5 ppm for 5 hours,
once a week) increased injury, inflammation, and allergic responses in a rodent model of allergic airway
disease. Newly available evidence shows that repeated subchronic exposure to 0.1 ppm ozone promoted
eosinophilic airway inflammation in a model of allergic sensitization.
A recent study (Hansen et al.. 2013) was conducted in mice exposed for 12 weeks to ozone
(0.1 ppm for 20 minutes/day for 5 days a week for 2 weeks and once weekly for 12 weeks). Mice were
also exposed to a low dose of ovalbumin which produced minimal sensitization because levels of serum
ovalbumin-specific IgE were minimally affected. After 14 weeks, mice were challenged with a high dose
of ovalbumin. As mentioned above in Section 3.2.4.1.2. no increases in ventilatory parameters or
indicators of bronchoconstriction were observed. In addition, ozone exposure did not increase
ovalbumin-specific IgE levels indicating that ozone did not act as an adjuvant. However, ozone exposure
resulted in a statistically significant increase in BALF eosinophils. Study-specific details are summarized
in Table 3-55 in Section 3.3.2.
3.2.4.6.3 Integrated Summary for Allergic Response
Cross-sectional epidemiologic studies provide generally consistent evidence that ozone
concentrations are associated with hay fever/rhinitis and serum-markers of allergic response. However, in
addition to uncertainties regarding cross-sectional associations, potential confounding by pollen
concentrations also remains a considerable uncertainty. There is supporting evidence from recent and
previously evaluated toxicological studies that provides biological plausibility for some of the observed
epidemiologic associations. Specifically, ozone exposure induced airway eosinophilia in a rodent model
of allergic sensitization and enhanced allergic responses in a rodent model of allergic airway disease. In
contrast to the epidemiologic evidence, one recent experimental study did not observe ozone-related
changes in allergen-specific IgE levels in mice.
3.2.4.7 Respiratory Effects in Pregnancy
3.2.4.7.1 Animal Toxicological Studies
No animal toxicological studies evaluating respiratory effects in pregnancy were described in the
2013 Ozone ISA (U.S. EPA. 2013a). Newly available evidence shows that pregnant rats responded to
ozone exposure with immediate effects on ventilatory parameters and later effects reflecting airway
injury.
3-108
-------
A recent study in pregnant rats (Miller et al.. 2017) demonstrated that exposure to 0.4 and
0.8 ppm ozone on Gestational Days 5 and 6 resulted in altered ventilatory parameters (decreased minute
volume and increased enhanced pause) immediately post-exposure and increased markers of injury
(gamma glutamyl transferase and /v'-accty 1 -g 1 utam i n idase) in BALF on Gestational Day 21. The observed
alterations in enhanced pause were dose dependent. These effects were statistically significant.
Study-specific details are summarized in Table 3-50 in Section 3.3.2. Nonrespiratory endpoints evaluated
in this study are discussed elsewhere in this document.
3.2.4.8 Respiratory Effects in Populations with Metabolic Syndrome
3.2.4.8.1 Animal Toxicological Studies
No animal toxicological studies evaluating respiratory effects in populations with diabetes or
metabolic syndrome were described in the 2013 Ozone ISA (U.S. EPA. 2013a). Newly available evidence
shows that male and female rats had different responses to subchronic ozone exposure that were
dependent on diet. These effects, described below, were statistically significant. Study-specific details are
summarized in Table 3-46 in Section 3.3.2. The scientific evidence that supports causality determinations
for long-term ozone exposure and metabolic effects is assessed in Appendix 5.
• A recent study (Gordon et al.. 2016b) was conducted in male and female rats fed high-fructose
and high-fat diets prior to and during 4 weeks of ozone exposure (0.8 ppm for 5 hours/day, once a
week). Ozone exposure increased a marker of injury (albumin) and a marker of inflammation
(eosinophils) in BALF of males on the high-fructose and high-fat diets. Females on the high-fat
diet had increased albumin and females on the high-fructose diet had increased eosinophils in
response to ozone exposure. The high-fructose and high-fat diets enhanced some of the effects of
ozone on inflammatory or injury-related markers.
3.2.4.9 Respiratory Mortality
When considering the entire body of evidence, there is limited support for an association with
long-term ozone exposure and respiratory mortality. Recent studies use a variety of both fixed-site
(i.e., monitors) and models (e.g., CMAQ, dispersion models) to measure or estimate ozone concentrations
for use in assigning long-term ozone exposure in epidemiologic studies (Section 2.6.2). The strongest
evidence comes from analyses of the ACS cohort data, including a study discussed in the 2013 Ozone
ISA that observed positive associations between long-term ozone exposure and respiratory mortality
(Jerrett et al.. 2009) and a recent analysis of respiratory, COPD, and pneumonia mortality (Turner et al..
2016). Results from other recent studies are less consistent, with analyses of U.S., Canadian, and
European cohorts reporting inconsistent associations between long-term ozone exposure and respiratory
mortality. The differences in how ozone exposure was assessed do not explain the heterogeneity in the
3-109
-------
observed associations. The results from studies evaluating long-term ozone exposure and respiratory
mortality are presented in Figure 3-13. Overall, there is some evidence that long-term ozone exposure is
associated with respiratory mortality, but the evidence is not consistent across studies. Specifically:
• The strongest evidence for an association between long-term ozone exposure and respiratory
mortality comes from nationwide analyses of the ACS cohort, demonstrating positive associations
with respiratory mortality (Turner et al.. 2016; Jerrett et al.. 2009) and COPD, and pneumonia/flu
(Turner etal.. 2016). The observed associations were relatively unchanged in copollutant models
adjusting for PM2 5 and NO2 (Turner et al.. 2016). In contrast, Jerrett et al. (2013) reported a null
association between long-term ozone exposure and respiratory mortality in an analysis of the
ACS cohort limited to participants from California.
• Several recent analyses of the CanCHEC cohort in Canada provide inconsistent evidence for an
association between long-term ozone exposure and respiratory mortality, with one reporting a
positive association (Weichenthal et al.. 2017) and the other reporting a negative association
(Crouse et al.. 2015). Cohort studies conducted in France (Bentaveb et al.. 2015) and the U.K.
(Carey et al.. 2013) also report negative associations between long-term ozone exposure and
respiratory mortality.
3-110
-------
Reference
Cohort
Years
Mean
Jerrett et al. 2009 ACS
tTurner etaL 2016 ACS
tJerrett et at 2013 ACS
tCrouse etaL 2015 CanCHEC
IWeichenthal etaL2017 CanCHEC
IBertayeb etaL 2015 Gaael
tCareyetal. 2013 English Medical Practice 2003-2007
1982-1999 57.5
1982-2004 38 2
1982-2000 50.35
1991-2006 39.6
1991-2011 3829
1989-2013 40.5
ITurner etal. 2016
ITurner etal. 2016
ACS
ACS
25.85 ¦*-
1982-2004 382
1982-2004 382
I-
-I-
Outcome
Respiratory Mortality
COPD Mortality
Pneumonia/Flu Mortality
-I
0.5 0.75 1 125 1.5 175
Hazard Ratio (95% CI)
ACS = American Cancer Society; CanCHEC = Canadian Census Health and Environment Cohort.
Note: fStudies published since the 2013 Ozone ISA. Black text = studies included in the 2013 Ozone ISA.
Associations are presented per 10 ppb increase in pollutant concentration. Circles represent point estimates; horizontal lines
represent 95% confidence intervals for ozone. Black text and circles represent evidence included in the 2013 Ozone ISA; red text
and circles represent recent evidence not considered in previous ISAs or AQCDs.
Figure 3-13 Summary of associations from studies of long-term ozone
exposure and respiratory mortality for a standardized increase in
ozone concentrations.
3-111
-------
3.2.5 Relevant Issues for Interpreting Epidemiologic Evidence—Long-Term
Ozone Exposure and Respiratory Effects
3.2.5.1 Potential Copollutant Confounding of the Ozone-Respiratory Relationship
Potential copollutant confounding is a recurrent issue in epidemiologic studies of the health
effects of air pollutants. Pollutant concentrations are often correlated, such that it can be difficult to
distinguish the effect of one pollutant from another. In the recently evaluated studies of long-term
exposure to ozone and respiratory health effects, ozone correlations with other pollutants have varied
greatly across studies (PM25: r = -0.21 to 0.66; NO2: r = -0.42 to 0.38; SO2: -0.24 to 0.15). Limited
evaluation of copollutant models in recent studies provides some evidence that ozone associations may be
attenuated, but still positive in copollutant models with NO2 and PM2 5 (Gilliland et al.. 2017; Berhane et
al.. 2016; To et al.. 2016). Additionally, because many studies report modest copollutant correlations,
strong copollutant confounding from any of the measured copollutants is unlikely. However, given the
limited amount of available evidence, including a lack of measurement of noncriteria pollutants, the
potential confounding effect of copollutants remains a notable source of uncertainty in the relationship
between long-term ozone exposure and respiratory health effects.
3.2.5.2 The Role of Season and Temperature on Ozone Associations with Respiratory
Health Effects
A number of epidemiologic studies of short-term ozone concentrations and respiratory health
effects have conducted seasonal analyses, comparing associations observed in the warm season to cold
season or year-round estimates (Section 3.1.10.2). Results from these studies have generally supported
stronger associations in the warm season. Despite this line of evidence, there have not been seasonal
comparisons of long-term exposures to ozone. While one study in Quebec, Canada used summertime
average ozone concentrations as a surrogate for annual exposure, the results are not informative to
seasonal differences given the evaluation of year-round health outcomes (Tctrcault et al.. 2016a).
3.2.5.3 Shape of the Concentration-Response Relationship
There are no recent epidemiologic studies or studies previously considered in the 2013 Ozone
ISA (U.S. EPA. 2013a) that examine the C-R relationship of the association between long-term exposure
to ozone and respiratory health effects. While most studies use linear or log-linear models to characterize
health effects, there are a lack of studies that empirically assess deviations from linearity or use
alternative models that allow for nonlinearity. Thus, there is some uncertainty regarding the shape of the
C-R relationship and existence of a threshold.
3-112
-------
3.2.6
Summary and Causality Determination
The 2013 Ozone ISA concluded that a "causal relationship is likely to exist" between long-term
ozone exposure and respiratory effects (U.S. EPA. 2013a). This conclusion was based on epidemiologic
evidence of associations between long-term ozone exposure and asthma development, respiratory
symptoms in children with asthma, and respiratory mortality. Notably, associations between long-term
exposure to ozone and new-onset asthma in children were primarily evident in longitudinal studies that
examined interactions between ozone and exercise or different genetic variants, including HMOX-1,
ARG, and GSTP1. The observed gene-environment interactions were supported by evidence that the
specific enzymes corresponding to these genetic variants have antioxidant and/or anti-inflammatory
properties, providing biological plausibility for the observed interactions. Additionally, the evidence
relating new-onset asthma to long-term ozone exposure was supported by toxicological studies in infant
monkeys, which indicate that postnatal ozone exposures can lead to the development of asthma. This
nonhuman primate evidence of ozone-induced respiratory effects supported the biological plausibility of
long-term exposure to ozone contributing to the development of asthma in children. Specifically, these
studies indicate that early-life ozone exposure can cause structural and functional changes that could
potentially contribute to airway obstruction and increased airway responsiveness. Some uncertainties were
acknowledged in the previous causality determination, specifically regarding the limited number of
epidemiologic studies examining potential copollutant confounding and the potential for exposure
measurement error in epidemiologic studies. However, in general, the epidemiologic and toxicological
evidence provided evidence of a likely to be causal relationship between long-term exposure to ozone and
respiratory effects.
Recent studies continue to examine the relationship between long-term exposure to ozone and
respiratory effects. Key studies that inform the causality determination are presented in Table 3-3. A
limited number of recent epidemiologic studies provide generally consistent evidence that long-term
ozone exposure is associated with the development of asthma in children (Section 3.2.4.1.1). A large
administrative cohort study, following over one million children from birth, observed an association
between long-term exposure to ozone and asthma onset. The case definition in this analysis does not
exclude potentially acute asthma exacerbations, which increases the potential for confounding by time-
varying covariates. Acknowledging this uncertainty, the finding was consistent with a recent CHS
analysis that reported a decrease in childhood asthma incidence associated with decreases in ozone
concentrations across nine southern California communities. Although a smaller study restricted to
minority children did not find evidence of an association between long-term ozone concentrations and
asthma development, the two larger studies provides noteworthy evidence of a positive association. In
addition to the development of asthma, epidemiologic studies have also evaluated the relationship
between ozone and asthma severity (Section 3.2.4.5). Consistent with results from the 2013 Ozone ISA,
recent studies have presented consistent evidence that long-term exposure to ozone is associated with
hospital admissions and ED visits for asthma and prevalence of bronchitic symptoms in children with
asthma. Notably, there is some uncertainty regarding the results from studies of hospital admissions and
3-113
-------
ED visits for asthma, which typically represent an acute outcome. Most of these studies do not adjust for
short-term ozone concentrations, despite there being an established association between short-term
exposure and asthma exacerbation (Section 3.1.4.2).
In support of the evidence from recent epidemiologic studies, there are a number of recent animal
toxicological studies that expand the evidence for long-term ozone exposure-induced effects that may
lead to asthma development (Section 3.2.4.1.2). Specifically, studies in nonhuman primates have shown
that postnatal ozone exposure can compromise airway growth and development, promote the
development of an allergic phenotype and increased airway responsiveness, and cause persistent
alterations to the immune system. In addition, findings that ozone exposure enhances injury,
inflammation, and allergic responses in allergic rodents provide biological plausibility for the relationship
between ozone exposure and the exacerbation of allergic asthma.
In addition to studies of asthma, there is new and/or expanded evidence from epidemiologic and
animal toxicological studies published since the completion of the 2013 Ozone ISA that provide evidence
of associations between long-term ozone exposure and the development of COPD (Section 3.2.4.3).
allergic responses (Section 3.2.4.6V A recently available epidemiologic study provides limited evidence
that long-term ozone exposure is associated with incident COPD hospitalizations in adults with asthma.
This finding is supported by recent animal toxicological studies that provide consistent evidence of
airway injury and inflammation resulting from subchronic ozone exposures. These results are coherent
with animal toxicological studies reviewed in the 2013 Ozone ISA, which demonstrated that chronic
ozone exposure damages distal airways and proximal alveoli, resulting in persistent inflammation and
lung tissue remodeling that leads to irreversible changes, including fibrotic- and emphysematous-like
changes in the lung. Respiratory tract inflammation and morphologic and immune system-related changes
may underlie the progression and development of chronic lung disease, such as COPD.
A larger body of epidemiologic studies also provides support for an association between
long-term ozone exposure and allergic responses, including hay fever/rhinitis and serum allergen-specific
IgE. While recent studies demonstrate generally consistent results, potential confounding by pollen
exposure remains an uncertainty. However, there is supporting evidence from animal toxicological studies
demonstrating enhanced responses in ozone-exposed allergic rodents (Section 3.2.4.6.2). In addition,
animal toxicological studies reviewed in the short-term exposure section show type 2 immune responses
in nasal airways of rodents exposed repeatedly to ozone (Section 3.1.4.4.2V These findings are
characteristic of induced nonatopic asthma and rhinitis and provide biological plausibility for the
observed epidemiologic associations with hay fever/rhinitis.
Taken together, previous and more recent animal toxicological studies of long-term exposure to
ozone demonstrate biological plausibility for many of the associations observed in recent epidemiologic
studies. Specifically, there is strong evidence of ozone-induced inflammation, injury, and oxidative stress
in adult animals. These effects represent initial events through which ozone may lead to a number of
downstream respiratory endpoints, including altered morphology in the lower respiratory tract and the
3-114
-------
development of COPD. Further, there is evidence of a range of ozone-induced effects on lung
development in neonatal rodents and infant monkeys, including altered airway architecture, airway
sensory nerve innervation, airway cell death pathways, increased serotonin-positive airway cells, and
immunomodulation. An infant monkey model of allergic airway disease also demonstrated effects on lung
development, including compromised airway growth, impaired alveolar morphogenesis, airway smooth
muscle hyperreactivity, an enhanced allergic phenotype, priming of responses to oxidant stress, and
persistent effects on the immune system. These various upstream effects provide a plausible pathway
through which ozone may act on downstream events, such as altered immune function leading to altered
host defense and allergic responses, as well as morphologic changes leading to the development of
asthma. A more thorough discussion of the biological pathways that potentially underlie respiratory health
effects resulting from long-term exposure to ozone can be found in Section 3.2.3.
Recent epidemiologic studies provide some evidence that long-term ozone exposure is associated
with respiratory mortality, but the evidence is not consistent across studies (Section 3.2.4.9). A recent
nationwide study in the U.S. observed associations between ozone and underlying causes of respiratory
mortality, including COPD. This finding is supported by the new lines of evidence from animal
toxicological and epidemiologic studies on the development of COPD, as discussed previously. Results
from epidemiologic studies of ozone-related respiratory mortality in populations outside the U.S. are
inconsistent.
A notable source of uncertainty across the reviewed epidemiologic studies is the lack of
examination of potential copollutant confounding. A limited number of studies that include results from
copollutant models suggest that ozone associations may be attenuated but still positive after adjustment
for NO2 or PM2 5. However, the few studies that include copollutant models examine different outcomes,
making it difficult to draw strong conclusions about the nature of potential copollutant confounding for
any given outcome. Importantly, in addition to studies that explicitly address potential copollutant
confounding through modeling adjustments, many studies report modest copollutant correlations, which
suggests that strong confounding due to copollutants is unlikely. Another source of uncertainty common
to epidemiologic studies of air pollution is the potential for exposure measurement error. Most of the
recent epidemiologic studies of long-term ozone exposure use concentrations from fixed-site monitors as
exposure surrogates. Exposure measurement error relating to exposure assignment from fixed-site
monitors has the potential to bias effect estimates in either direction, although it is more common that
effect estimates are underestimated, and the magnitude of the bias is likely small given that ozone
concentrations do not vary over space as much as other criteria pollutants, such as NOx or SO2
(Section 2.3.1.1V
Despite some uncertainties in the epidemiologic literature, there is coherence from animal
toxicological studies that provides support for the observed epidemiologic associations. Experimental
evidence also provides biologically plausible pathways through which long-term ozone exposure may
3-115
-------
lead to respiratory effects. Overall, the collective evidence is sufficient to conclude that a likely to be
causal relationship exists between long-term ozone exposure and respiratory effects.
Table 3-3 Summary of evidence for a likely to be causal relationship between
long-term ozone exposure and respiratory effects.
Ozone
Concentrations
Rationale for Causality Associated with
Determination3 Key Evidence13 Key References'3 Effects0
Development of Asthma
Consistent evidence from
toxicological studies at relevant
concentrations
Animal toxicological studies show
postnatal exposure results in
compromised airway growth and
development
Section 3.2.4.1.2 0.5 ppm
(Infant monkeys)
Lee et al. (2011) 0.5 ppm
(rats)
Animal toxicological studies show Section 3.2.4.1.2 0.5 ppm
postnatal exposure promotes an (Infant monkeys)
allergic phenotype in the developing
lung
Animal toxicological studies show Section 3.2.4.1.2 0.5 ppm
postnatal exposure alters sensory (Infant monkeys)
nerve innervation in the developing
lung
Animal toxicological studies show
postnatal exposure alters airway
responsiveness
Section 3.2.4.1.2 0.5 ppm
(Infant monkeys)
Moore et al. (2012a) 0.5 ppm
Generally consistent evidence
from a limited number of
epidemiologic studies of asthma
development in children
Cohort studies demonstrating an
association with asthma
development in children
Garcia et al. (2019) 32.1 ppb mean
summer ozone
concentration,
based on 8-h
midday avg
Longitudinal studies provide
evidence of associations with
asthma development in populations
with specific genetic variants
Islam etal. (2008);
Salam et al. (2009)
38.4 ppb mean
annual ozone
concentration in
low ozone
communities;
55.2 ppb in high
ozone
communities,
based on 8-h avg
(10:00 a.m.-6:00
p.m.)
3-116
-------
Table 3-3 (Continued): Summary of evidence for a likely to be causal relationship
between long-term ozone exposure and respiratory effects.
Ozone
Concentrations
Rationale for Causality Associated with
Determination3 Key Evidence13 Key References'3 Effects0
Uncertainty regarding confounding
by copollutants
No examination of copollutant
confounding in models of
gene-environment interaction.
Available studies report low to
moderate copollutant correlations
Severity of Respiratory Disease
Limited, but consistent
epidemiologic evidence from
studies of respiratory disease
severity
Longitudinal studies provide
consistent evidence of an
association between long-term
ozone concentrations and bronchitic
symptoms in children with asthma
McConnell et al.
(2003); Berhane et
al. (2016); Gilliland
et al. (2017)
44.8-47.7 ppb
annual average,
across cohorts,
based on 8-h avg
(10:00 a.m.-6:00
p.m.)
Consistent evidence of an
association between long-term
ozone concentrations and hospital
admissions and ED visits for asthma
Moore et al. (2008);
Menq et al. (2010);
Tetreault et al.
(2016b)
32.1 ppb mean
summer ozone
concentration,
based on 8-h
midday avg
(Tetreault et al..
2016b)
87.8 ppb quarterly
1 h daily max
(Moore et al..
2008)
Uncertainty regarding confounding Limited evidence that observed Berhane et al.
by copollutants associations were attenuated but still (2016): Gilliland et
positive in copollutant models al. (2017)
adjusting for NO2 or PM2.5
Other uncertainties Studies of hospital admissions and Section 3.2.4.5.1
ED visits for asthma do not account
for the potential effect of short-term
exposures leading to these acute
events
Biological plausibility Evidence that ozone exposure Section 3.2.4.5.2
enhances injury, inflammation, and
allergic responses in allergic rodents
provide biological plausibility for the
relationship between ozone
exposure and the exacerbation of
allergic asthma
3-117
-------
Table 3-3 (Continued): Summary of evidence for a likely to be causal relationship
between long-term ozone exposure and respiratory effects.
Limited epidemiologic evidence
from study of COPD incidence
Uncertainty regarding confounding
by copollutants
Allergic Response
Limited, but consistent
epidemiologic evidence from
studies of allergic response
Uncertainty regarding confounding
by copollutants
The only study evaluated indicates
an association between ozone
concentrations and COPD incidence
in adults with asthma
Limited evidence from a single study
reported an association between
ozone and rhinitis that was
persistent in a copollutant model
adjusting for NO2.
Ozone
Concentrations
Associated with
Key References'3 Effects0
Limited evidence from a single study To et al. (2016)
reported an association between
ozone and COPD incidence that was
attenuated, but still positive in a
copollutant model adjusting for PM2.5
Epidemiologic studies provide Section 3.2.4.6
consistent evidence of ozone
associations with hay fever/rhinitis
and allergen-specific IgE levels
Rationale for Causality
Determination3
Key Evidence13
Development of Chronic Obstructive Pulmonary Disease
Consistent evidence from
toxicological studies at relevant
concentrations
Animal toxicological evidence of
morphologic changes in distal
airways and proximal alveoli leading
to lung tissue remodeling and
fibrotic/emphysematous-like
changes in rodents and monkeys
Section 3.2.4.3.2
0.12-1 ppm
To et al. (2016)
38.4 ppb mean
ozone
concentration,
based on avg of
monthly 24-h max
from time of
enrollment
To et al. (2016)
Section 3.2.4.6.1 51.5 ppb annual
avg, based on 8-h
max
Penard-Morand et
al. (2005)
Section 3.2.4.6.1
Section 3.2.4.6.2 0.1-0.5 ppm
Section 3.1.4.4.2 0.5-0.8 ppm
Other uncertainties All available studies were
cross-sectional. Additionally,
potential confounding by pollen
concentrations also remains a
considerable uncertainty
Coherent evidence from Animal toxicological evidence for
toxicological studies at relevant enhanced allergic responses
concentrations
Animal toxicological evidence from
short-term studies show type 2
immune responses in nasal airways
of rodents repeatedly exposed
3-118
-------
Table 3-3 (Continued): Summary of evidence for a likely to be causal relationship
between long-term ozone exposure and respiratory effects.
Ozone
Concentrations
Rationale for Causality Associated with
Determination3 Key Evidence13 Key References'3 Effects0
Respiratory Mortality
Inconsistent epidemiologic
Recent epidemioloqic studies Section 3.2.4.9
evidence from multiple,
provide some evidence of an
high-quality studies
association with respiratory mortality,
but the evidence is not consistent.
New evidence from one study
demonstrating an association with
COPD mortality
Some coherence with underlying
Studies of COPD development Section 3.2.4.3
causes of mortality
provide coherence with COPD
mortality
Biological plausibility Animal toxicological studies show
the development of
emphysematous-like disease and
increased severity of
infection-related alveolitis
aBased on aspects considered in judgments of causality and weight of evidence in causal framework in Table I and Table II of the
Preamble to the ISAs (U.S. EPA. 2015).
bDescribes the key evidence and references, supporting or contradicting, contributing most heavily to causality determination and,
where applicable, to uncertainties or inconsistencies. References to earlier sections indicate where full body of evidence is
described.
°Describes the ozone concentrations with which the evidence is substantiated.
Section 3.2.4.3.2
Section 3.2.4.5.2
3-119
-------
3.3 Evidence Inventories—Data Tables to Summarize Study
Details
3.3.1 Short-Term Exposure
Table 3-4 Study-specific details from controlled human exposure studies of
lung function in healthy populations.
Study
Population n, Sex, Age
(Range or Mean ± SD)
Exposure Details
(Concentration,
Duration)
Endpoints Examined
Madden et al. (2014)
Healthy adults
n = 11 males, 4 females
Age: 27 ± 4 yr
0 (Day 1) ppb, 2 h
300 (Day 1) ppb, 2 h
300 (Day 2) ppb, 2 h
Spirometry (before, immediately
PE, and once per hour for 4 h
PE)
Ghio et al. (2014)
Healthy adults
n = 14 males, 5 females
Age: 25 ± 3 yr
0 ppb, 2 h
300 ppb, 2 h
FEVi (immediately before and
PE)
Hoffmever et al. (2013)
Healthy adults
n = 8 males, 7 females
Age: 26 yr
40 ppb, 4 h
240 ppb, 4 h
Spirometry (before and
immediately PE)
Plethysmograph (before and
immediately PE)
FramDton et al. (2015)
Healthy adults
n = 15 males, 9 females
Age: 26.4 yr
0 ppb, 3 h
100 ppb, 3 h
200 ppb, 3 h
FEVi, FVC (30 min before and
immediately PE and 4 h PE)
Kahleetal. (2015)
Healthy adults
n = 14 males, 2 females
Age: 27 yr
0 ppb at 22°C, 2 h
0 ppb at 32.5°C, 2 h
300 ppb at 22°C, 2 h
300 ppb at 32.5°C, 2 h
FEVi/FVC (before and
immediately PE)
Bates et al. (2014)
Healthy adult nonsmokers
n = 17 males, 13 females
Age: 25 ± 6 yr
Healthy adult smokers
n = 19 males, 11 females
Age: 24 ± 4 yr
300 ppb, 1 h
FEVi, FVC, dead space, alveolar
slope, spirometry (before and
PE)
3-120
-------
Table 3-4 (Continued): Study-specific details from controlled human exposure
studies of lung function in healthy populations.
Study
Population n, Sex, Age
(Range or Mean ± SD)
Exposure Details
(Concentration,
Duration)
Endpoints Examined
Bennett et al. (2016)
Healthy adults
n = 19 normal weight
females
Age: 24 ± 4 yr
n = 19 obese females
Age: 28 ± 5 yr
0 ppb, 2 h
400 ppb, 2 h
Airway responsiveness (3 h PE)
FEV-i, FVC, sGaw (before and
PE)
PFT, PMN, airway
responsiveness, symptoms (train
day, before and immediately and
3 h PE)
Stieqel et al. (2017)
Healthy adults
n = 11 males, 4 females
Age: 27 yr
0 ppb, 2 h
300 ppb, 2 h
FEV-i, FVC (before and
immediate PE)
Ariomandi et al. (2018)
FramDton et al. (2017)
Healthy adults
n = 35 males, 52 females
Age: 59.9 ± 4.5 yr
0 ppb, 3 h
70 ppb, 3 h
120 ppb, 3 h
Spirometry (30 min before, 5 min
PE, 22 h PE)
Biller et al. (2011)
Healthy adults
n = 11 males, 3 females
Age: 33.1 ± 9.5 yr
0 ppb, 3 h
250 ppb, 3 h
FEV-i, FVC (before and 0, 3, and
21 h PE)
Tanket al. (2011)
Healthy adults
n = 11 males, 3 females
Age: 34 ± 10 yr
0 ppb, 3 h
250 ppb, 3 h
FEV-i, FVC (before and
immediate PE)
FE\A| = forced expiratory volume in 1 second; FVC = forced vital capacity; PE = post-exposure; PFT = pulmonary function test;
PMN = polymorphonuclear leukocyte (e.g., neutrophils); sGaw = specific airways conductance.
3-121
-------
Table 3-5 Study-specific details from animal toxicological studies of short-term
ozone exposure and lung function in healthy animals.
Study
Species (Stock/Strain), n,
Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Scheleale and
Walbv (2012)
Rats (BN)
n = 5-11 males
Age: 8-10 weeks
1 ppm, 8 h
Lung function, breathing pattern,
(immediately PE)
Wolkoff et al. (2012)
Mice (BALB/cA)
n = 9-20 males
Age: NR but mean weight
was 24 g
0.1 ppm, 1 h/day for 10 days
Lung function (during exposure)
Lee et al. (2013)
Mice (BALB/c)
n = 6 females
Age: 5-6 weeks
2 ppm, 3 h
Enhanced pause (immediately
PE)
Sunil et al. (2013)
Rats (WS)
n = 3-6 females
Age: NR but weight was
200-225 g
2 ppm, 3 h
Pulmonary mechanics (48 and
96 h PE)
Groves et al. (2012)
Mice (C57BL/6J)
n = 4-9 males
Age: 8 weeks
0.8 ppm, 3 h
Pulmonary mechanics (72 h PE)
Choet al. (2013)
Mice (ICR) WT" and NRF2
deficient
n = 3-12
Sex and age: NR
2 ppm, 3 h
Pulmonary mechanics,
acetylcholine challenge (24 h PE)
Barreno et al.
(2013)
Mice (C57BL/6) WT and
osteopontin deficient
n = 6-10 females
Age: 8 weeks
2 ppm, 3 h
Pulmonary mechanics, MCh
challenge (24 h PE)
Groves et al. (2013)
Mice (C57BL/6J)
n = 3-10 males
Age: 8, 27, 80 weeks
0.8 ppm, 3 h
Pulmonary mechanics, resistance
and elastance spectra (72 h PE)
Ghio et al. (2014)
Mice (CD-1)
n = 6 females
Age: 4 weeks
2 ppm, 3 h
Enhanced pause, MCh challenge
(24 h PE)
3-122
-------
Table 3-5 (Continued): Study-specific details from animal toxicological studies of
short-term ozone exposure and lung function in healthy
animals.
Species (Stock/Strain), n,
Study Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Razvi et al. (2015)
Mice (C57BL/6J WT" and
resistin deficient)
n = 6-8 males and
females
Age: 4-8 weeks
2 ppm, 3 h
Pulmonary mechanics, MCh
challenge (24 h PE)
Dye et al. (2015)
Rats (WKY, WS, SD)
n = 4-8 males
Age: 12-14 weeks
0.25 ppm, 4 h
0.5 ppm, 4 h
1.0 ppm, 4 h
Whole body plethysmography (0
and 20 h PE)
Clav et al. (2016)
Guinea pigs
(Dunkin-Hartley)
n = 4-32 males
Age: NR but weight was
300-500 g
Rabbits (New Zealand
white)
n = 4-16 males
Age: NR but weight was
2.5-4 g
2 ppm, 1 h
2 ppm, 30 min
Cough response, pulmonary
mechanics, challenge with Mch
(4 h or 3 days PE)
Snow et al. (2016)
Rats (BN)
n = 6-8 males
Age: 1,4, 12, 24 mo
0.25 ppm, 6 h/day for 2 days
1 ppm, 6 h/day for 2 days
Ventilatory parameters (18 h PE)
Gordon et al.
(2016b)
Rats (BN)
n = 9-10 males and
females
Age: 20 weeks
0.8 ppm, 5 h
Ventilatory parameters (18 h PE)
Kasahara et al.
(2015)
Mice (C57BL/6 WT,
ROCK1 insufficient,
ROCK2 insufficient)
n = 4-12 males
Age:20-25 weeks
2 ppm, 3 h
Pulmonary mechanics, challenge
with MCh (24 h PE)
Verhein et al. (2013) Guinea pigs
(Dunkin-Hartley)
n = 3-7 females
Age: NR but weight was
300-470 g
2 ppm, 4 h
Pulmonary inflation pressure,
challenge with i.v. of acetylcholine
and electrical stimulation of the
vagal nerve (24 h PE)
Williams et al.
(2015)
Mice (C57BL/6, TNF-a
sufficient and deficient)
n = 5-9 females
Age: 10-12 weeks
2 ppm, 3 h
Pulmonary mechanics, challenge
with MCh (24 h PE)
3-123
-------
Table 3-5 (Continued): Study-specific details from animal toxicological studies of
short-term ozone exposure and lung function in healthy
animals.
Study
Species (Stock/Strain), n,
Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Hansen et al. (2016)
Mice (BALB/cJ)
n = 5 females
Age: 6 weeks
2 ppm, 1 h/day for 3 days
Breathing frequency, tidal volume,
time of brake, time of pause
(during exposure)
Zvchowski et al.
(2016)
Mice (C57BL/6)
n = 4-8 males
Age: 6-8 weeks
1 ppm, 4 h
Pulmonary mechanics, challenge
with MCh (18-20 h PE)
Miller etal. (2016b)
Rats (WKY)
n = 4-6 males
Age: 12-13 weeks
1 ppm, 4 h/day for 1-2 days
Ventilatory parameters
(immediately post-exposure Day 1
and about 12 h later)
Zhu et al. (2016)
Mice (BALB/c)
n = 3-5 males
Age: 5-6 weeks
0.1 ppm, 3 h/day for 7 days
0.5 ppm, 3 h/day for 7 days
1 ppm, 3 h/day for 7 days
Pulmonary mechanics, challenge
with MCh (24 h PE)
Gordon et al.
(2017b)
Rats (LE)
n = 10 females
Age: 13 weeks
0.25 ppm, 5 h/day for 2 days
0.5 ppm, 5 h/day for 2 days
1 ppm, 5 h/day for 2 days
Ventilatory parameters
(immediately PE)
Mathews et al.
(2017b)
Mice (C57BL/6J WT and
TCR gamma delta
deficient)
n = 6 males, 4-10 females
Age: 10 weeks
2 ppm, 3 h
Pulmonary mechanics, challenge
with MCh (24 h PE)
Henriauez et al.
(2017)
Rats (WKY)
n = 6-8 males
Age: 12 weeks
0.8 ppm, 4 h/day for
1-2 days
Ventilatory parameters
(immediately PE)
Malik etal. (2017)
Mice (C57BL/6J WT, Ccrl2
deficient)
n = 8-13 females
Age: 8 weeks
2 ppm, 3 h
Pulmonary mechanics, challenge
with MCh (4 and 24 h PE)
Michaudel et al.
(2018)
Mice (C57BL/6J WT, ST2
deficient)
n = 4-6 females
Age: 8-10 weeks
1 ppm, 1 h
Ventilatory parameters, challenge
with MCh (24 h PE)
Stober et al. (2017)
Mice (BALB/cByJ WT,
TSG-6 deficient)
n = 4-8
Sex and age: NR
2 ppm, 3 h
Pulmonary mechanics, challenge
with MCh (24 h PE)
3-124
-------
Table 3-5 (Continued): Study-specific details from animal toxicological studies of
short-term ozone exposure and lung function in healthy
animals.
Study
Species (Stock/Strain), n,
Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Mathews et al.
(2018)
Mice (C57BL/6J)
n = 4-14 females
Age: 10 weeks
2 ppm, 3 h
Pulmonary mechanics, challenge
with MCh (24 h PE)
Liu et al. (2016)
Mice (BALB/c)
n = 6
Sex and age: NR but
weight was 20 g
1.5 ppm, 0.5 h/day for 5 days
Ventilatory parameters, challenge
with MCh (immediately PE)
Choet al. (2018)
Mice (C57BL/6) specific
pathogen free and germ
free
n = 6-14 males
Age: 10 weeks
2 ppm, 3 h
Pulmonary mechanics, challenge
with MCh (24 h PE)
BN = brown Norway; LE = Long-Evans; MCh = methacholine; NR = not reported; PE = post-exposure; S-D = Sprague-Dawley;
WKY = Wistar Kyoto; WS = Wistar; WT = wild type.
3-125
-------
Table 3-6 Epidemiologic studies of short-term exposure to ozone and lung
function in healthy populations.
Study
Exposure
Mean
Copollutant
Effect Estimates
Study
Population
Assessment
ppb
Examination
(95% Cla)
Berrv et al. (1991)
n = 14
Regional monitor for
Mean: NR
Correlation
FEVi (mL):
Hamilton, NJ, U.S.
Campers
part of study
Maximum:
(r): NR
20.5 (4.3, 36.7)
July 1988
without asthma
(<8 miles from
204
Copollutant
PEF (mL/sec):
Panel study
Age: <14 yr
camps)
Mobile trailer monitor
models: NR
-25.3 (-82.6, 32.1)
onsite at one camp
1-h max
SDektorand LiDDmann
n = 46
On-site monitor
Mean: 69
Correlation
Percentage
(1991)
Campers
1-h avg
Maximum:
(r): NR
increase
Fairview Lake, NJ, U.S.
without asthma
137
Copollutant
FEVi:
July-August 1988
Age: 8-14 yr
models: NR
-1.4 (-1.9, -0.8)
Panel study
Avol et al. (1990)
n = 295
On-site monitoring
Mean: 94
Correlation
Percentage
Pine Springs, CA, U.S.
Campers
1-h avg
Maximum:
(r): NR
increase
June-August 1988
without asthma
161
Copollutant
FEVi:
Panel study
Age: 8-17 yr
models: NR
-0.4 (-0.6, -0.1)
PEF:
1.2 (0.4, 1.9)
Burnett et al. (1990)
n = 29
On-site monitoring
Mean: 59
Correlation
Percentage
Lake Couchiching,
Campers
1-h avg
Maximum:
(r): NR
increase
Ontario, Canada
without asthma
95
Copollutant
FEVi:
June-July 1983
Age: 7-15 yr
models: NR
-0.2 (-1.1, 0.7)
Panel study
PEF:
-1.19 (-2.38, -0.03)
Hiqains et al. (1990)
n = 43
On-site monitoring
Mean: 103
Correlation
Percentage
San Bernardino, CA,
Campers
1-h avg
Maximum:
(r): NR
increase
U.S.
without asthma
245
Copollutant
FEVi:
June-July 1987
Age: 7-13 yr
models: NR
-1.0 (-1.5, -0.5)
Panel study
PEF:
-0.5 (-1.3, 0.2)
Raizenne et al. (1989)
n = 112
On-site monitoring
Mean: 71
Correlation
Percentage
Lake Erie, Ontario,
Campers
1-h avg
Maximum:
(r): NR
increase
Canada
without asthma
143
Copollutant
FEVi:
June-August 1986
Age: mean
models: NR
-0.3 (-0.5, -0.1)
Panel study
11.6 yr
PEF:
-0.04 (-0.35, 0.26)
3-126
-------
Table 3-6 (Continued): Epidemiologic studies of short-term exposure to ozone
and lung function in healthy populations.
Study
Study
Population
Exposure
Assessment
Mean
ppb
Copollutant
Examination
Effect Estimates
(95% Cla)
Spektor et al. (1988)
Fairview Lake, NJ, U.S.
July-August 1984
Panel study
n = 91
Campers
without asthma
Age: 8-15 yr
On-site monitoring
1-h avg
Mean: 53
Maximum:
113
Correlation
(r): PM2.5:
0.78; SO42":
0.82
Copollutant
models: NR
Percentage
Increase
FEV1:
-0.6 (-0.9, -0.2)
PEF:
-1.1 (-2.1, -0.3)
tDales et al. (2013)
Sault Ste. Marie,
Ontario, Canada
Ozone: May-August,
2010
Follow-up: May-August,
2010
Panel study
n = 61
Age: 24 ± 6 yr
8 h near steel
plant or college
campus for
5 consecutive
days at each
site with 9-day
period between.
Outcomes 0 h
after exposure
period
Portable monitor at
site
8-h avg (8-h period
between 7:50
a.m.-5:50 p.m.)
Summer days
Mean:
college
campus:
32.6; steel
plant: 29.7
Correlation
(r): NR
Copollutant
models: NR
Percentage
increase
FEV1: -0.47 (-1.00,
0.06)
FEV1/FVC: -0.48
(-0.90, -0.05)
FVC: -0.56 (-1.39,
0.26)
TLC: -0.97 (-2.71,
0.76)
FEF25-75: -1.46
(-3.46, 0.53)
RV: -6.48 (-12.47,
0.50)
CI = confidence interval; FENA = forced expiratory volume in 1 second; FEF25-75 = forced expiratory flow at 25-75% of the
pulmonary volume; FVC = forced vital capacity; NR = not reported; PEF = peak expiratory flow; RV = residual volume;
S042" = sulfate; TLC = total lung capacity.
tStudies published since the 2013 Ozone ISA.
aResults standardized to a 15-ppb increase in 24-h avg, 20-ppb increase in 8-h daily max, or 25-ppb increase in 1-h daily max
ozone concentrations.
3-127
-------
Table 3-7 Epidemiologic studies of short-term exposure to ozone and lung
function, airway inflammation, and oxidative stress in general
populations.
Study
Study
Population
Exposure
Assessment
Mean
ppb
Copollutant
Examination
Effect Estimates
(95% Cla)
tLepeule et al. (2014)
Boston, MA, U.S.
1999-2009
n = 776
Adult men
Age: 72.3
(mean); 6.8
(SD)
5.9% asthma
prevalence,
6.8% chronic
bronchitis
prevalence
City wide monitor
average (median
monitor distance
from participant
homes: 22.3)
4-h avg (4:00 a.m.
to 8:00 a.m.)
24-h avg
Mean: 47
(24-h avg)
95th: 60
Correlation
(r):
CO: -0.29;
NO2: -0.31;
PM2.5: 0.04;
BC: -0.21
Copollutant
models: NR
Results presented
graphically. 1-day lag
ozone concentrations,
as well as longer
moving averages (3, 4,
5, 6, 7, 14, 21, and
28 days), were
associated with
decreased FEV1 and
FVC. DNA methylation
did not significantly
modify the effect of
ozone on lung function.
tRice et al. (2013)
Boston, MA, U.S.
1995-2011
n = 3,362
Adults
Age: 51.8
(mean); 12.7
(SD)
20.7% asthma
orCOPD
prevalence
City wide monitor
average
8-h max
Warm season
(April-September)
Mean: 28.7
75th: 35.3
Maximum:
59.6
Correlation
(r):
NO2: 0.01;
PM2.5: 0.33
Copollutant
models: NR
Lag 1
FEV1 (mL): -34.8
(-61.8, -8.0)
Obese participants
FEV1 (mL): -60.8
(-94.0, -27.4)
Nonobese participants
FEV1 (mL): -24.8
(-52.8, 3.4)
tPatel et al. (2013)
n = 36
Single monitor
Median:
Correlation
Lag 0
New York City, NY,
Schoolchildren
within 14 km of
38.8
(r): BC: 0.02
Unit change in exhaled
U.S.
ages 14-19
schools
Copollutant
breath condensate pH
2005
50% asthma
8-h max
models: NR
-0.14 (-0.33, 0.05)
Panel study
prevalence
May-June
Unit change in
8-isoprostane -0.41
(-0.72, -0.10)
tSalam et al. (2012)
Multicity, southern
California, U.S.
2004-2006
n = 940
Schoolchildren
ages 6-11
14.2% asthma
prevalence
Single monitor in
each study
community.
8-h avg
(10:00 a.m. to
6:00 p.m.)
Mean: 35.1
Maximum:
63.7
Correlation
(r):
PM2.5: 0.07;
PM10: 0.34;
NO2: -0.41
Copollutant
models: NR
7-day avg
Inducible nitric oxide
synthase (iNOS)
% methylation
-0.08 (-1.40, 1.28)
BC = black carbon; CI = confidence interval; FE\A| = forced expiratory volume in 1 second; FVC = forced vital capacity; NR = not
reported; SD = standard deviation.
tStudies published since the 2013 Ozone ISA.
aResults standardized to a 15-ppb increase in 24-h avg, 20-ppb increase in 8-h daily max, or 25-ppb increase in 1-h daily max
ozone concentrations.
3-128
-------
Table 3-8 Study-specific details from controlled human exposure studies of
respiratory symptoms in healthy populations.
Exposure Details
Population n, Sex, Age (Concentration,
Study (Range or Mean ± SD) Duration) Endpoints Examined
Bennett et al. (2016) Healthy adults 0 ppb, 2 h Symptoms (immediately PE)
n = 19 normal weight females 400 ppb, 2 h
Age: 24 ± 4 yr
n = 19 obese females
Age: 28 ± 5 yr
PE = post-exposure.
Table 3-9 Study-specific details from controlled human exposure studies of
inflammation, oxidative stress, and injury in healthy populations.
Study
Population n, Sex,
Age(Range or
Mean ± SD)
Exposure Details
(Concentration, Duration)
Endpoints Examined
Bosson et al. (2013)
Healthy adults
n = 8 males,
5 females
Age: 24.6 yr
0 ppb, 2 h
200 ppb, 2 h
BALF PMNs (1.5 h PE)
Healthy adults
n = 9 males,
6 females
Age: 24.5 yr
BALF PMNs (6 h PE)
Healthy adults
n = 10 males,
5 females
Age: 23 yr
BALF PMNs (18 h PE)
Holland et al. (2014)
Healthy adults
n = 10 males,
12 females
Age: 33.0 ± 7.4 yr
0 ppb, 4 h
100 ppb, 4 h
200 ppb, 4 h
BALF (20 h PE)
Gomes et al. (2011a)
Healthy adults
100 ppb with heat, 0.5 h
Nasal lavage (0 and 6 h PE)
n = 9 males,
0 females
Age: 30 ± 2.6 yr
3-129
-------
Table 3-9 (Continued): Study-specific details from controlled human exposure
studies of inflammation, oxidative stress, and injury in
healthy populations.
Study
Population n, Sex,
Age(Range or
Mean ± SD)
Exposure Details
(Concentration, Duration)
Endpoints Examined
Alexis et al. (2013)
Healthy adults
60 ppb, 6.6 h
Sputum PMN (18 h PE)
n = 24
Age: 20-33 yr
Hoffmever et al. (2015)
Healthy adults
n = 5 males,
5 females
Age: 25.6 ± 2.5 yr
40 ppb, 4 h
240 ppb, 4 h
EBC-pH (before and immediately
after and 16 h PE)
FeNO (before and immediately
after and 16 h PE)
Holz et al. (2015)
Healthy adults
n = 12 males,
12 females; only
18 subjects
completed study
Age: 35 yr (median)
250 ppb, 3 h
Sputum (3 h PE)
Speen et al. (2016)
Healthy adults
n = 9-11
Age: 18-35 yr
0 ppb, 2 h
300 ppb, 2 h
BALF Oxysterols (1 and 24 h
PE)
Bennett et al. (2016)
Healthy adults 0 ppm, 2 h
n = 19 normal weight 400 ppb, 2 h
females
Age: 24 ± 4 yr
n = 19 obese
females
Age: 28 ± 5 yr
Sputum PMN (4 h PE)
Cheng et al. (2018)
Devlin et al. (2012)
Healthy adults
n = 20 males,
3 females
Age: 28.8 yr
(median)
0 ppb, 2 h
300 ppb, 2 h
BALF samples (1 or 24 h PE)
Ariomandi et al. (2018)
Frampton et al. (2017)
Healthy adults
n = 35 males,
52 females
Age: 59.9 ± 4.5 yr
0 ppb, 3 h
70 ppb, 3 h
120 ppb, 3 h
Sputum protein and PMNs
(22.5 h PE)
Lazaar et al. (2011)
Healthy adults
n = 24 males,
0 females
Age: 35.5 yr
250 ppb, 3 h
Sputum PMN (3 h PE)
Biller et al. (2011)
Healthy adults
n = 11 males,
3 females
Age: 33.1 ± 9.5 yr
0 ppb, 3 h
250 ppb, 3 h
Sputum (screening and 3 h PE)
3-130
-------
Table 3-9 (Continued): Study-specific details from controlled human exposure
studies of inflammation, oxidative stress, and injury in
healthy populations.
Study
Population n, Sex,
Age(Range or
Mean ± SD)
Exposure Details
(Concentration, Duration)
Endpoints Examined
Tanket al. (2011)
Healthy adults
n = 11 males,
3 females
Age: 34 ± 10 yr
0 ppb, 3 h
250 ppb, 3 h
Sputum (3 h PE)
Kirsten et al. (2011)
Healthy adults
n = 15 males,
3 females
Age: 43.9 ± 7.4 yr
250 ppb, 3 h
Sputum PMN (3 h PE)
Gomes et al. (2011b)
Healthy adults
n = 9 males,
0 females
Age: 24 ± 6 yr
0 ppb with control, 0.5 h
0 ppb with heat, 0.5 h
100 ppb with control, 0.5 h
100 ppb with heat, 0.5 h
Sputum markers (15 min PE)
BALF = bronchoalveolar lavage fluid; EBC pH = exhaled breath condensate pH; FeNO = fraction exhaled nitric oxide;
PE = post-exposure; PMN = polymorphonuclear leukocyte (e.g., neutrophils).
Table 3-10 Study-specific details from animal toxicological studies of short-term
ozone exposure and allergic sensitization in healthy animals.
Study
Species
(Stock/Strain),
Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Brand et al. (2012)
Mice (C57BL/6)
n = NR males
Age: 8-12 weeks
0.8 ppm, 8 h/day for 3 days
Histopathology, BALF total cells
and differentials, dendritic cell
number and activation in specific
sites, T cell number in MLN
(immediately PE)
Zhu et al. (2016) Mice (BALB/c) 0.1 ppm, 3 h/day for 7 days IgE, Th2 cytokines, mast cell
n = 3-5 males 0.5 ppm, 3 h/day for 7 days degranulation (24 h PE)
Age: 5-6 weeks 1 ppm, 3 h/day for 7 days
BALF = bronchoalveolar lavage fluid; IgE = immunoglobulin E; MLN = mediastinal lymph node; NR = not reported;
PE = post-exposure; Th2 = T helper 2.
3-131
-------
Table 3-11 Study-specific details from animal toxicological studies of short-term
ozone exposure and inflammation, oxidative stress, and injury in
healthy animals.
Species (Stock/Strain), n, Exposure Details
Study Sex, Age (Concentration, Duration) Endpoints Examined
Hulo etal. (2011) Mice (C57BL6/SV129 WT
and AMPK-a deficient)
n = 3-10 males
Age: 20-24 weeks
ppm, 3 h Markers of oxidative stress,
inflammation, injury; AMPK
activation; Na/K-ATPase
abundance (24 h PE)
Connor etal. (2012) Mice (C3H/HeOuJ and TLR4 0.
mutant C3H/HeJ)
n = 3-18 males
Age: 11-12 weeks
ppm, 3 h Markers of oxidative stress,
injury, and inflammation;
surfactant protein D (0.5-48 h
PE)
Kasahara et al. Mice (C57BL/6J WT and 0.3 ppm, up to 72 h Markers of injury and
(2012) adiponectin deficient) inflammation (PE)
n = 3-10 (sex and age
matched) males, females
Age: 11-13 weeks
Shore etal. (2011)
Mice (C57BL/6 WT and
TNRF1 deficient)
n = 3-6 males or females
Age: 7 and 39 weeks
2 ppm, 3 h
BALF total and differential cell
count, cytokines, chemokines
and tissue mRNA MT, HO-1,
claudin-4 and amphiregulin (4 h
PE)
Scheleqle and Walbv Rats (BN) 1 ppm, 8 h BALF markers of injury and
(2012) n = 5-11 males inflammation (immediately PE)
Age: 8-10 weeks
Wolkoff et al. (2012) Mice (BALB/cA) 0.1 ppm, 1 h/day for BALF cells (immediately PE)
n = 9-20 males 10 days
Age: NR, but mean weight
was 24 g
Tankerslev et al. Mice (C57BL/6J WT and 0.5 ppm, 3 h BALF cell counts and cell
(2013) atrial natriuretic differentials, total protein
peptide-deficient) (8-10 h PE)
n = 5-6 males
Age: 11 weeks
Lee et al. (2013) Mice (BALB/c) 2 ppm, 3 h BALF cells, MDA, antioxidants,
n = 6 females RNS (immediately PE)
Age: 5-6 weeks
Sunil et al. (2013) Rats (WS) 2 ppm, 3 h BALF protein and CCSP (3-72 h
n = 3-6 females PE)
Age: NR but weight was
200-225 g
3-132
-------
Table 3-11 (Continued): Study-specific details from animal toxicological studies
of short-term ozone exposure and inflammation, oxidative
stress, and injury in healthy animals.
Study
Species (Stock/Strain), n,
Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Groves et al. (2012)
Mice (C57BL/6J WT)
n = 4-9 males
Age: 8 weeks
0.8 ppm, 3 h
BALF protein, RNS,
macrophage number,
chemotactic activity s (24-72 h
PE)
Sunil et al. (2012)
Rats (WS)
3-11 females
Age: NR but weight was
200-225 g
2 ppm, 3 h
BALF protein, cell number,
differentials,
immunohistochemistry—markers
of oxidative stress, apoptosis
and autophagy, BALF
macrophages—markers of
classical and alternative
activation pathways, (3-72 h
PE)
BhooDalan et al.
(2013)
Rats (S-D)
n = 6 males
Age: 8-9 weeks
0.8 ppm, 3 h
BALF total cells and
differentials, protein, albumin,
LDH, total antioxidant capacity,
lung tissue SODs, catalase,
p-actin (18-24 h PE)
Choet al. (2013)
Mice (ICR WT and NRF2
deficient)
n = 3-12
Sex and age: NR
0.3 ppm, 6-72
2 ppm, 3 h
BALF total protein and cell
differentials, mucin, glutathione,
lung tissue redox
measurements, histopathology
(immediately and 3-24 h PE)
Robertson et al.
(2013)
Mice (C57BL/6 WT and
CD36-deficient)
n = 3-8 females
Age: 8-10 weeks
1 ppm, 4 h
BALF protein, cell number and
cell differentials (24 h PE)
Thomson et al.
(2013)
Rats (F344)
n = 4-6 males
Weight NR but age was
200-250 g
0.4 ppm, 4 h
0.8 ppm, 4 h
mRNA expression in tissue
(immediately and 24 h PE)
Yanaqisawa et al.
(2012)
Mice (C57BL/6J WT and
peroxiredoxin-1 deficient)
n = 4-8 males
Age: 18 weeks
2 ppm, 6 h
BALF total cell count and cell
differentials, total protein,
mediators, Prxdl tissue HO-1
and GST mRNA, NRF2 protein,
histopathology (0, 4, 18 h PE)
Barreno et al. (2013)
Mice (C57BL/6 WT and
osteopontin deficient)
n = 6-10 females
Age: 8 weeks
2 ppm, 3 h
BALF and serum osteopontin,
cytokines, total cells and cell
differentials, total protein,
soluble collagen, epithelial cells
(6 and 24 h PE)
Mclntosh-Kastrinskv
et al. (2013)
Mice (C57BL/6)
n = 8 females
Age: 7 mo
0.245 ppm, 4 h
BALF cells (12 h PE)
3-133
-------
Table 3-11 (Continued): Study-specific details from animal toxicological studies
of short-term ozone exposure and inflammation, oxidative
stress, and injury in healthy animals.
Study
Species (Stock/Strain), n,
Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Groves et al. (2013)
Mice (C57BL/6J WT)
n = 3-10 males
Age: 8, 27, 80 weeks
0.8 ppm, 3 h
BALF total cells and differential
cell counts, protein (72 h PE)
Brand et al. (2012)
Mice (C57BL/6)
n = NR males
Age: 8-12 weeks
0.8 ppm, 8 h/day for 3 days
Histopathology, BALF total cells
and differentials, dendritic cell
number and activation in specific
sites, T cell number in MLN
(immediately PE)
Wana et al. (2013)
Rats (WS)
n = 6 males
Age NR but weight was
150-180 g
0.8 ppm, 4 h/day, 2 days
per week for 3 weeks
BALF total cells and cell
differentials, LDH, protein,
albumin, alkaline phosphatase;
histopathology; lung tissue
activity of GPx and SOD, MDA
levels, mRNA eNOS, iNOS,
ICAM-1 (24 h after
6th exposure)
Gonzalez-Guevara et
al. (2014)
Rats (WS)
n = 3-6 males
Age: NR but weight was
250-300 g
1 ppm, 1 or 3 h/day for
5 days
1 ppm, 1, 3, and 6 h
Tissue IL-6 and TNF-a
(immediately PE)
Kurhanewicz et al.
(2014)
Mice (C57BL/6)
n = 6 females
Age: 10-12 weeks
0.3 ppm, 4 h
BALF LDH, microalbumin, NAG,
total protein (24 h PE)
Ghio et al. (2014)
Mice (CD-1)
n = 6 females
Age: 4 weeks
2 ppm, 3 h
BALF and liver nonheme iron,
BALF ferritin; BALF injury
markers and neutrophils and
cytokines (24 h PE)
Paffett et al. (2015)
Rats (SD)
n = 3-5 males
Age: 8-12 weeks
1 ppm, 4 h
BALF total cell and cell
differential counts, total protein
(24 h PE)
Sunil et al. (2015)
Mice (C57BL/6J WT and
galectin-deficient)
n = 3-14 females
Age: 8-11 weeks
0.8 ppm, 3 h
BALF protein, tissue cytochrome
b5 as injury markers;
macrophage subpopulations in
tissue and BAL (24-72 h PE)
Gabehart et al.
(2015)
Mice (BALB/c)
n = 3-14 females
Age: 6 weeks
1 ppm, 3 h
BAFL total cell number and
differential cell counts, albumin,
MUC5AC; gene expression
chemokines, antioxidants, TLR4,
neuropeptides (6, 24, and 48 h
PE)
3-134
-------
Table 3-11 (Continued): Study-specific details from animal toxicological studies
of short-term ozone exposure and inflammation, oxidative
stress, and injury in healthy animals.
Study
Species (Stock/Strain), n,
Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Razvi et al. (2015)
Mice (C57BL/6J WT and
resistin deficient)
n = 6-8 males/females
Age: 4-8 weeks
2 ppm, 3 h
BALF cells, protein, and
mediators, tissue injury and
inflammation (24 h PE)
Kumarathasan et al.
(2015)
Rats (F344)
n = 8-17 males
Age: NR but weight was
200-250 g
0.4 ppm, 4 h
0.8 ppm, 4 h
BALF cells and cell differentials,
BALF markers of oxidative
stress and injury (24 h PE)
Verhein et al. (2015)
Mice (B6129SF1/J WT and
Notch3 and Notch4 deficient)
n = 4-10 males
Age: 7-13 weeks
0.3 ppm, 6-72 h
BALF total protein, total cell
count, and cell differentials;
tissue NFkB activation, mRNA
fortnf and Notch-related genes,
microarray analysis (immediately
PE)
Cabello et al. (2015)
Mice (C57BL/6J)
n = 6 males, 6 females
Age: 8 weeks
2 ppm, 3 h
BALF total cells and cell
differential counts, protein,
albumin (24 and 72 h PE); lung
tissue mRNA array of
84 inflammatory gene; lung
tissue PCR of proinflammatory
cytokines and chemokines,
pattern recognition receptors,
transcription factors, STAT3
phosphorylation (4 h PE)
Ward et al. (2015)
Rats (WKY)
n = 3-4 males
Age: 10-12 weeks
0.25 ppm, 4 h
0.5 ppm, 4 h
1 ppm, 4 h
BALF protein and neutrophils,
lung gene expression (0 and
20 h PE)
Kodavanti et al.
(2015)
Rats (WKY, WS, SD)
n = 4-8 males
Age: 12-14 weeks
0.25 ppm, 4 h
0.5 ppm, 4 h
1 ppm, 4 h
BALF total cell counts and cell
differentials, total protein,
albumin, LDH, NAG, GGT; lung
tissue mRNA for HO-1, MIP-2,
TNF-a, IL-6, IL-10 (0 and 20 h
PE)
Ramot et al. (2015)
Rats (WKY, WS, S-D)
n = 4-8 males
Age: 12-14 weeks
0.25 ppm, 4 h
0.5 ppm, 4 h
1 ppm, 4 h
Lung histopathology (0 and 20 h
PE)
Ward and Kodavanti
(2015)
Rats (WKY)
n = 3-4 males
Age: 12-14 weeks
1 ppm, 4 h
Lung gene expression profiling
(immediately PE)
Ona et al. (2016)
Mice (C57BL/6 WT and
lymphoid cell-deficient)
n = 6 males
Age: 6-8 weeks
0.5 ppm, 4 h for up to
9 days
Histopathology,
immunochemistry, mRNA
expression (2-24 h after
exposure)
3-135
-------
Table 3-11 (Continued): Study-specific details from animal toxicological studies
of short-term ozone exposure and inflammation, oxidative
stress, and injury in healthy animals.
Study
Species (Stock/Strain), n,
Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Hatch et al. (2015)
Rats (WKY, WS)
n = 8 males
Age: 12-14 weeks
1 ppm, 4 h
BALF and tissue antioxidants (0
and 20 h PE)
Mishra et al. (2016)
Mice (C57BL/6)
n = 6-8 males, 6-8 females
with females at different
stages of estrous cycle
Age: 8 weeks
2 ppm, 3 h
Lung tissue inflammatory
mediators and transcription
factors (4 h PE)
Cheet al. (2016)
Mice (C57BL/6 WT and
11-17a and 11-1 r1 deficient)
n = 6 females
Age: 6-8 weeks
0.7 ppm, 72 h
BALF total cells and cell
differentials, protein; lung tissue
cytokines and chemokines,
mRNA, flow cytometry of
lymphocyte subpopulations; flow
cytometry of lung macrophage
ROS; lung macrophage mtDNA
(24 h PE)
Snow et al. (2016)
Rats (BN)
n = 6-8 males
Age: 1, 4, 12, 24 mo
0.25 ppm, 6 h/day for
2 days
1 ppm, 6 h/day for 2 days
BALF total cells, cell
differentials, protein, albumin,
GGT, NAG (18 h PE)
Gordon et al. (2016b)
Rats (BN)
n = 9-10 males,
9-10 females
Age: 16 weeks
0.8 ppm, 5 h
BALF total cells, cell
differentials, albumin (18 h PE)
Kasahara et al.
(2015)
Mice (C57BL/6 WT, ROCK1
insufficient, ROCK2
insufficient)
n = 4-12 males
Age: 20-25 weeks
2 ppm, 3 h
BALF total cells, cell
differentials, albumin, epithelial
cells, protein, cytokines,
chemokines, hyaluronan; lung
tissue GRPR mRNA, ROCK,
and rho protein (24 h PE)
Verhein et al. (2013)
Guinea pigs (Dunkin-Hartley)
n = 3-7 females
Age: NR but weight was
300-470 g
2 ppm, 4 h
BALF total cells and cell
differentials (24 h PE)
Mathews et al. (2015) Mice (C57BL/6 WT, gamma
delta T cell deficient)
n = 4-14 males
Age: 10-13 weeks
0.3 ppm, 24-72 h
BALF total cells and cell
differentials, cytokines, protein;
lung tissue mRNA; lung tissue
macrophage subpopulations,
histopathology (0, 1, 3, 5 days
PE)
3-136
-------
Table 3-11 (Continued): Study-specific details from animal toxicological studies
of short-term ozone exposure and inflammation, oxidative
stress, and injury in healthy animals.
Study
Species (Stock/Strain), n,
Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Williams etal. (2015)
Mice (C57BL/6J WT or Cpe
fat, TNF-a sufficient and
deficient)
5-9 females
Age: 10-12 weeks
2 ppm, 3 h
BALF total cells and cell
differentials, MCP-1, G-CSF,
hyaluronan, osteopontin, IL-13,
protein carbonyls; lung tissue
mRNA for antioxidant proteins
and 1117a (24 h PE)
Zvchowski et al.
(2016)
Mice (C57BL/6)
n = 4-8 males
Age: 6-8 weeks
1 ppm, 4 h
Lung weightbody weight ratios,
BAL total cells (18-20 h PE)
Thomson et al.
(2016)
Rats (F344)
n = 5 males
Age: NR but weight was
200-250 g
0.8 ppm, 4 h
BALF total cells and cytokines;
lung tissue mRNA for cytokines
and antioxidant proteins and
glucocorticoid inducible proteins
(immediately PE)
Miller et al. (2016b)
Rats (WKY)
n = 4-6 males
Age: 12-13 weeks
1 ppm, 4 h/day for
1-2 days
BALF total cells and differential
cells, protein, albumin, LDH
(immediately PE Day 1 and
Day 2)
Zhu et al. (2016)
Mice (BALB/c)
n = 3-5 males
Age: 5-6 weeks
0.1 ppm, 3 h/day for 7 days
0.5 ppm, 3 h/day for 7 days
1 ppm, 3 h/day for 7 days
Oxidative stress and
upregulation of antioxidants
(24 h PE)
Kasahara et al.
(2013)
Mice (C57BL/6 WT,
adiponectin, and T-cadherin
deficient)
n = 3-16 (sex matched
males and females)
Age: 11-13 weeks
0.3 ppm, 72 h
BALF total cells and cell
differentials, cytokines,
chemokines; lung tissue mRNA
for IL-17A, serum amyloid A3
and Ki67 (immediately PE)
Kasahara et al.
(2014)
Mice (C57BL/6 WT,
adiponectin deficient, IL-6
deficient)
n = 3-13 (sex matched
males and females)
Age: 11-13 weeks
0.3 ppm, 24-72 h
BALF total cells and
differentials, cytokines,
adiponectin; lung tissue mRNA
for serum amyloid A3, TIMP1,
1117a, microarray analysis; flow
cytometry (immediately PE)
Brand etal. (2016)
Mice (C57BL/6 WT, IL-23
deficient, Flt3l deficient,
lacking conventional
dendritic cells)
n = 5-12 (sex matched
males and females)
Age: 8-12 weeks
0.3 ppm, 24-72 h
BALF total cells and cell
differentials, cytokines,
chemokines, protein, lung tissue
mRNA for 1117a and Il23a; flow
cytometry for cDC macrophages
(immediately PE)
3-137
-------
Table 3-11 (Continued): Study-specific details from animal toxicological studies
of short-term ozone exposure and inflammation, oxidative
stress, and injury in healthy animals.
Study
Species (Stock/Strain), n,
Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Kumaqai et al. (2016)
Mice (C57BL/6 WT, Rag2
deficient, Il2rg deficient)
n = 6 males
Age: 6-8 weeks
0.5 ppm, 4 h/day for 1 or
9 days
Quantitative immunochemistry
for markers of nasal epithelial
remodeling and eosinophilic
rhinitis; upregulation of
Th2-related genes (24 h after
last exposure)
Elkhidir et al. (2016)
Mice (C57BL/6 WT, PAI-1
deficient)
n = 6-10 females
Age: NR but WT and
deficient mice were age
matched
2 ppm, 3 h
BALF total cells and cell
differential, epithelial cells,
protein, II-6, KC, MIP-2, PAI-1 (4
and 24 h PE)
Gordon et al. (2017b)
Rats (LE)
n = 10 females
Age: 13 weeks
0.25 ppm, 5 h/day for
2 days
0.5 ppm, 5 h/day for 2 days
1 ppm, 5 h/day for 2 days
BALF total cells and cell
differentials, total protein,
albumin, NAG, GGT
(immediately PE)
Ciencewicki et al.
(2016)
Mice (C57BL/6 WT,
mannose binding lectin
deficient)
n = 6-12 males
Age: 6 weeks
0.3 ppm; 24, 48, 72 h
BALF total cell counts and cell
differentials, protein; lung tissue
mRNA II6, tnf, cxcl2, cxcl5,
microarray (immediately PE)
Fenq et al. (2015)
Mice (BALB/c)
n = 3-7 males, 3-7 females
Age: 6-8 weeks
0.25 ppm, 3 h/day for
7 days
0.5 ppm, 3 h/day for 7 days
1 ppm, 3 h/day for 7 days
BALF total cells and cell
differentials, protein, ROS, EGF,
TGF-a (20-24 h PE)
Francis et al. (2017b)
Mice (C57BL/6J)
n = 3-10 females
Age: 11-14 weeks
0.8 ppm, 3 h
BALF total cells and cell
differentials, total proteins; lung
tissue immunohistochemistry for
macrophage markers and
4-HNE, western blotting for
SP-D, mRNA for chemokines
and ligands; flow cytometry of
lung tissue and BALF cells for
monocyte subpopulations
(24-72 h PE)
Harkema et al. (2017)
Mice (C57BL/6NTac and
BALB/cNTac)
n = 6 males
Age: 6-8 weeks
0.8 ppm, 4 h/day for 9 days
BALF total cells and cell
differentials; lung tissue mRNA
for MUC5AC, MUC5B,
Clca1/Gob5, II33, II25, Tslp, II5,
1113, Chia, Chil4/Ymw (24-72 h
PE)
3-138
-------
Table 3-11 (Continued): Study-specific details from animal toxicological studies
of short-term ozone exposure and inflammation, oxidative
stress, and injury in healthy animals.
Study
Species (Stock/Strain), n,
Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Mathews et al.
(2017b)
Mice (C57BL/6J T, TCR
gamma delta deficient)
n = 4-10 females
Age: 10 weeks
2 ppm, 3 h
BALF total cells and cell
differentials, cytokines,
chemokines; flow cytometry of
isolated cells from lung tissue
(24 h PE)
Xianq et al. (2012)
Mice (BALB/c)
n = 5
Sex and age: NR
2 ppm, 0.5 h for 1-8 days
mRNA for transcription factor
Mathews et al.
(2017a)
Mice (C57BL/6J)
n = 5-8 females
Age: 10 weeks
2 ppm, 3 h
Markers of oxidative stress (24 h
PE)
Malik etal. (2017)
Mice (C57BL/6J WT, Ccrl2
deficient)
n = 8-13 females
Age: 8 weeks
2 ppm, 3 h
BALF total cell number and
differential cell counts, total
protein, epithelial cells,
chemerin, adiponectin, eotaxin,
hyaluronan, IL-6, KC, MIP-2,
MIP-3a; lung tissue mRNA for
Ccrl2 (4 and 24 h PE)
Tiaheetal. (2018)
Mice (C57BL/6J)
n = 5 males
Age: 8-10 weeks
2 ppm, 3 h
BALF total cells and cell
differentials, total protein,
albumin (24 h PE)
Holze etal. (2018)
Mice (C57BL/6 WT, Nlrp
deficient, caspase deficient,
Asc deficient, and Pgam5
deficient)
n = 5
Sex and age: NR
1 ppm, 1 h
BALF cells, protein, MPO,
cytokines, (4 or 24 h PE)
Michaudel et al.
(2018)
Mice (C57BL/6J WT, ST2
deficient, IL-33 deficient, and
IL-33 citrine reporter)
n = 8-12 males, 4-6 females
Age: 8-10 weeks
0.3 ppm, 1 h
1 ppm, 1 h
BALF total cells and cell
differentials, proinflammatory
cytokines, chemokines, and
remodeling parameters, ROS
producing cells, total protein,
vascular leak, epithelial
desquamation marker, tissue
IL-33, ST2, tight junction
proteins and mRNA, cell death
marker, FACS (1-48 h PE)
Snow et al. (2018)
Rats (WKY)
n = 6-8 males
Age: 12 weeks
0.8 ppm, 4 h/day for 2 days
BALF total cell counts and cell
differentials, injury markers,
cytokines; lung tissue mRNA for
cholesterol transporters,
cholesterol receptors, nuclear
receptors (2 h PE)
3-139
-------
Table 3-11 (Continued): Study-specific details from animal toxicological studies
of short-term ozone exposure and inflammation, oxidative
stress, and injury in healthy animals.
Species (Stock/Strain),
Study Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Mathews et al. (2018) Mice (C57BL/6J)
n = 4-14 females
Age: 10 weeks
2 ppm, 3 h
BALF total cells and cell
differentials, IL-17A, IL-23,
IL-33, CCL20, CXCL1, CXCL2,
IL-6, G-CSF, GRP; lung tissue
flow cytometry for IL-17A
producing cells, microarray
analysis, mRNAforGRPR,
NQOI (24 h PE)
Kumaaai et al. (2017) Mice (C57BL/6 WT, Rag2 0.8 ppm, 4 h/day for 1 or
deficient, and IL2rg deficient) 9 days
n = 3-6 males
Age: 7-9 weeks
BALF total cells and cell
differentials; lung tissue mRNA
for type 2 immunity-related
transcripts, flow cytometry for
ILCs. (24 h and 2 weeks PE)
Yonchuk et al. (2017)
Rats (Han Wistar)
n = 5
Sex and age: NR but weight
was 290-370 g
1 ppm, 3 h
BALF total cells and cell
differential counts; lung tissue
glutathione (immediately PE)
Zhang et al. (2017)
Rats (S-D)
n = 4-5
Sex and age: NR but weight
was 180-220 g
2 ppm, 0.5 h/day for up to
12 days
BALF TNF-a, TGF(31, IL10,
MPO; lung tissue 8-oxoguanine,
OGG1, NOS and arginase
activity/protein level, ROS
(immediately PE)
Francis et al. (2017a) Mice (C57BL/6J)WT and
CCR2 deficient
n = 3-10 females
Age: 8-11 weeks
0.8 ppm, 1 h
BALF inflammatory cell
subpopulations, BAL total
protein; lung tissue iNOS, MR-1,
ADAM 17, Cypb5, 4-HNE, HO-1,
CCR2, mRNA for TNF-a, IL-(3,
iNOS, CX3CR1, CX3CL1,
NUR77 (24-72 h PE)
Liu et al. (2016)
Mice (BALB/c)
n = 6
Sex and age: NR but weight
was 20 g
1.5 ppm, 0.5 h/day for BALF inflammatory cell
5 days subpopulations, IFN-y, IL-4,
IL-17, TGFp, PGE2
(immediately PE)
3-140
-------
Table 3-11 (Continued): Study-specific details from animal toxicological studies
of short-term ozone exposure and inflammation, oxidative
stress, and injury in healthy animals.
Species (Stock/Strain), n, Exposure Details
Study Sex, Age (Concentration, Duration) Endpoints Examined
Cho et al. (2018) Mice (C57BL/6) Specific 2 ppm, 3 h BALF total cell counts and cell
pathogen free and germ free differential counts, total protein,
n = 6-14 males IL-17A, osteopontin, IL-33, IL-5,
GRP, G-CSF, eotaxin, IL-6,
Age: 10 weeks |p_10| KC MIP-2,
MIP-1a, LTB4 (24 h PE)
4-HNE = 4-hydroxynonenol; ADAM = a disintegrin and metalloproteinase; AMPK = AMP-activated protein kinase;
BALF = bronchoalveolar lavage fluid; BN = brown Norway; CCL = chemokine ligand; CCR2 = CC chemokine receptor type 2;
CCSP = club cell secretory protein; CD36 = cluster of differentiation 36; CXCL = chemokine family of cytokines with highly
conserved motif: cys-xxx-cys (CXC) ligand; CXCR = receptor for chemokine family of receptors; EGF = epidermal growth factor;
eNOS = endothelial nitric oxide synthase; F344 = Fischer 344; FACS = fluorescence activated cell sorting; GGT = gamma
glutamyl transferase; G-CSF = granulocyte colony-stimulating factor; GPx = glutathione peroxidase; GRP = gastrin releasing
peptide; GRPR = gastrin-releasing peptide receptor; GST = glutathione S-transferase; HO-1 = heme oxygenase 1;
ICAM-1 = intercellular adhesion molecule 1; IFN-y = interferon gamma; IL = interleukin; ILC = immune lymphoid cell;
iNOS = inducible nitric oxide synthase; KC = keratinocyte-derived chemokine; LDH = lactate dehydrogenase; LE = Long-Evans;
LTB4 = leukotriene B4; MCP-1 = monocyte chemotactic protein 1; MIP = macrophage inflammatory protein; MLN = mediastinal
lymph node; MPO = myeloperoxidase; MR-1 = major histocompatibility complex class l-related gene protein; mRNA = messenger
ribonucleic acid; mtDNA = mitochondrial DNA; MUC = mucin; NAG = /V-acetyl-glucosaminidase;
Na/K-ATPase = sodium-potassium adenosine 5'-triphosphatase; NFkB = nuclear factor kappa-light-chain-enhancer of activated B
cells; NQ01 = NADPH quinone oxidoreductase 1; NR = not reported; NRF2 = nuclear factor (erythroid-derived 2)-like 2;
NUR = nuclear receptor subfamily; OGG1 = 8-oxoguanine glycosylase; PAI-1 = plasminogen activator inhibitor 1;
PE = post-exposure; PGE2 = prostaglandin E2; Prxdl = peroxiredoxin 1; RNS = reactive nitrogen species;
ROCK = rho-associated coiled-coil-containing protein kinase; ROS = reactive oxygen species; S-D = Sprague-Dawley;
SOD = superoxide dismutase; SP-D = surfactant protein D; STAT3 = signal transducer and activator of transcription 3;
TGF = transforming growth factor; Th2 = T helper 2; TIMP1 = TIMP metalloprotease inhibitor 1; TLR = toll-like receptor;
TNF = tumor necrosis factor; TSLP = thymic stromal lymphopoietin; WKY = Wistar Kyoto; WS = Wistar; WT = wild type.
3-141
-------
Table 3-12 Study-specific details from animal toxicological studies of short-term
ozone exposure and morphology in healthy animals.
Species (Stock/Strain),
Study n, Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Connor et al. (2012)
Mice (C3H/HeOuJ and
C3H/HeJ TLR4 mutant)
n = 3-18 males
Age:11-12 weeks
0.8 ppm, 3 h
Type 2 cell proliferation (12-72 h
PE)
Groves et al. (2012)
Mice (C57BL/6J)
n = 4-9 males
Age: 8 weeks
0.8 ppm, 3 h
Histopathology,
immunohistochemistry (24-72 h PE)
Choet al. (2013)
Mice (ICR WT and NRF2
deficient)
n = 12
Sex and age: NR
0.3 ppm, 6-72 h
2 ppm, 3 h
Histopathology (3-24 h PE—acute)
Histopathology (immediately
PE—subacute)
Groves et al. (2013)
Mice (C57BL/6J WT)
n = 3-10 males
Age: 8, 27, 80 weeks
0.8 ppm, 3 h
Markers of cell proliferation, radial
alveolar counts, lesion scores (72 h
PE)
Brand et al. (2012)
Mice (C57BL/6)
Males
n = NR
Age:8-12 weeks
0.8 ppm, 8 h/day for 3 days Histopathology and flow cytometry
(immediately PE)
Wang et al. (2013)
Rats (WS)
n = 6 males
Age NR but weight was
150-180 g
0.8 ppm, 4 h day for 2 days Histopathology (24 h PE)
per week for 3 weeks
Gabehart et al.
(2015)
Mice (BALB/c)WT and
TLR4 deficient
n = 3-14 females
Age: 6 weeks
1 ppm, 3 h
Tissue: quantitative
immunochemistry—MUC5AC (6, 24,
and 48 h PE)
Cabello et al. (2015)
Mice (C57BL/6J)
n = 4-6 males,
4-6 females
Age: 8 weeks
2 ppm, 3 h
Histopathology (24 and 72 h PE)
Ramot et al. (2015)
Rats (WKY, WS, S-D)
n = 4-8 males
Age:12-14 weeks
0.25 ppm, 4 h
0.5 ppm, 4 h
1 ppm, 4 h
Lung histopathology (0 and 20 h PE)
3-142
-------
Table 3-12 (Continued): Study-specific details from animal toxicological studies
of short-term ozone exposure and morphology in healthy
animals.
Species (Stock/Strain),
Study n, Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Onq etal. (2016)
Mice (C57BL/6J WT,
Rag2 deficient, and Il2rg
deficient)
n = 6 males
Age: 6-8 weeks
0.5 ppm, 4 h for 1-9 days
Histopathology, immunochemistry,
mRNA expression (2-24 h PE)
Zhu et al. (2016)
Mice (BALB/c)
n = 3-5 males
Age: 5-6 weeks
0.1 ppm, 3 h/day for 7 days Histopathology scores (24 h PE)
Kasahara et al.
(2013)
Mice (C57BL/6 WT,
adiponectin, and
T-cadherin deficient)
n = 3-16 males and
females (age and sex
matched)
Age:11-13 weeks
0.3 ppm, 72 h
Histopathology (immediately PE)
Kumaqai et al.
(2016)
Mice (C57BL/6 WT, Rag2
deficient, Il2rg deficient)
n = 6 males
Age: 6-8 weeks
0.5 ppm, 4 h/day for 1 or
9 days
Histopathology, quantitative
histochemistry and
immunochemistry for markers of
nasal epithelial remodeling and
eosinophilic rhinitis (24 h PE)
Feng etal. (2015)
Mice (BALB/cJ)
n = 3-7, half males and
half females
Age: 6-8 weeks
0.25 ppm, 3 h/day for Lung tissue immunohistochemistry,
7 days inflammation scores, mean linear
0.5 ppm, 3 h/day for 7 days intercept (20-24 h PE)
1 ppm, 3 h/day for 7 days
Harkema et al.
(2017)
Mice (C57BL/6NTac and
BALB/cNT ac)
n = 6 males
Age: 6-8 weeks
0.8 ppm, 4 h/day for 9 days Lung tissue histochemistry and
immunochemistry for
mucosubstances and myelin basic
protein (24 h PE)
Wong etal. (2018)
Rats (WKY)
n = 8-12 males
Age:44-48 weeks
1 ppm, 6 h
Histopathology lesion scores (8 h
PE)
Michaudel et al. Mice (C57BL/6J WT, ST2 0.3 ppm, 1 h
(2018) deficient, IL-33 deficient, <| ppm <| ^
and IL-33 citrine reporter)
n = 4-6 females
Age:8-10 weeks
Histopathological lesion scores,
immunofluorescence, and confocal
microscopy of specific proteins
(1-48 h PE)
Kumagai et al. Mice (C57BL/6 WT, Rag2 0.8 ppm, 4 h/day for 1 or
(2017) deficient, and IL2rg 9 days
deficient)
n = 3-6 males
Age: 7-9 weeks
Quantitative histopathology,
Histochemistry for mucosubstances,
immunochemistry for BrdU (24 h or
2 weeks PE)
3-143
-------
Table 3-12 (Continued): Study-specific details from animal toxicological studies
of short-term ozone exposure and morphology in healthy
animals.
Species (Stock/Strain), Exposure Details
Study n, Sex, Age (Concentration, Duration) Endpoints Examined
Liu et al. (2016) Mice (BALB/c) 1.5 ppm, 0.5 h/day for Lung histopathology scores
n = 6 5 days
Sex and age: NR but
weight was 20 g
BALF = bronchoalveolar lavage fluid; BrdU = bromodeoxyuridine; ICR = Institute of Cancer Research; Il2rg = interleukin 2 receptor
subunit gamma; MLN = mediastinal lymph node; mRNA = messenger ribonucleic acid; MUC5AC = mucin 5AC glycoprotein;
NR = not reported; NRF2 = nuclear factor (erythroid-derived 2)-like 2; PE = post-exposure; SD = Sprague-Dawley; TLR4 = toll-like
receptor 4; WKY = Wistar Kyoto; WS = Wistar; WT = wild type.
3-144
-------
Table 3-13 Epidemiologic studies of short-term exposure to ozone and hospital
admission for asthma.
Study
Study
Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
(95% Clf
Silverman and Ito
(2010)
New York, NY, U.S.
Ozone: 1999-2006
Follow-up:
1999-2006
Time-series study
n = 75,383
All ages
Average of
13 monitors within
20 miles of the
geographic city
center
8-h max
Warm season
(April-August)
Mean: 41.0
75th: 53
90th: 68
Correlation (r):
PM2.5: 0.59
Copollutant
models: PM2.5
RR
All ages:
1.08 (1.05,
+PM2.5:
1.05 (1.02,
<6 yr:
1.08 (1.03,
+PM2.5:
1.05 (1.00,
6-18 yr:
1.18 (1.10,
+PM2.5:
1.14 (1.06,
19-49 yr:
1.07 (1.04,
+PM2.5:
1.05 (1.01,
50+ yr:
1.05 (1.02,
+PM2.5:
1.05 (1.01,
1.11)
1.09)
1.13)
1.10)
1.26)
1.21)
1.11)
1.09)
1.09)
1.08)
tWinauist et al.
All ages
One monitor
Correlation (r):
RR
(2012)
8-h max
NR
0-4 DL: 1.05 (0.99,
St. Louis, MO, U.S.
Year-round
Copollutant
1.11)
Ozone: 2001-2007
models: NR
Follow-up:
2001-2007
Time-series study
tSheffield et al.
n = 8,009
Average of city
Maximum: 60 Correlation (r):
No quantitative
(2015)
Age: 5-17 yr
monitors
NR
results. Results
New York City, NY,
24-h avg
Copollutant
presented
U.S.
Warm season
models: NR
graphically
Ozone:
(May-September)
May-September
2005-2012
Follow-up:
May-September
2005-2011
Case-crossover study
3-145
-------
Table 3-13 (Continued): Epidemiologic studies of short-term exposure to ozone
and hospital admission for asthma.
Study
Study
Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
(95% Cl)a
tShmool etal. (2016)
New York City, NY,
U.S.
Ozone: June-August
2005-2011
Follow-up:
June-August
2005-2011
Case-crossover study
n = 2,353
Age: 5-17 yr
Temporal estimates: Mean:
Average of city
monitors
Spatiotemporal
estimates: Fusion of
monitors and LUR
24-h avg
Warm season
(May-September)
Temporal:
30.4
Spatio-
temporal: 29.0
Maximum:
Temporal:
60.0
Spatio-
temporal: 60.3
Correlation (r): No quantitative
NR results. Results
Copollutant presented
models: NR graphically
tGoodman et al.
(2017a)
New York City, NY,
U.S.
Ozone: 1999-2009
Follow-up:
1999-2009
Time-series study
n = 295,497
All ages
Average of monitors
within 20 miles of the
geographic city
center
8-h max
Seasonal: Warm
season
(April-August) and
year-round estimates
Mean: 30.7
Median: 28
75th: 39.9
Maximum:
105.4
Correlation (r): Lag 0-1 RRs
PM2.5: 0.2
Copollutant
models: NR
Warm season
All ages: 1.01
(0.99, 1.03)
<6 yr: 0.99 (0.96,
1.03)
6-18 yr: 1.05 (1.01,
1.10)
19-49 yr: 1.03
(1.00, 1.06)
50+ yr: 1.00 (0.97,
1.03)
tZu et al. (2017)
Six Texas cities, U.S.
Ozone: 2001-2013
Follow-up:
2001-2013
Time-series study
n = 1,552,432
Age: 5+ yr
Average of monitors
in each city
8-h avg
Year-round
Mean: 32.2
Median: 31
75th: 40.1
90th: 48.6
Maximum:
82.8
Correlation (r): Lag 0-3 RRs
NR
Copollutant
models: NR
All ages (5+ yr):
1.05 (1.03, 1.07)
5-14 yr: 1.10 (1.05,
1.14)
15-64 yr: 1.04
(1.01, 1.07)
65+ yr: 1.00 (0.96,
1.05)
tGoodman et al.
(2017b)
Houston, Dallas, and
Austin, TX, U.S.
Ozone: 2003-2011
Follow-up:
2003-2011
Time-series study
n = 74,824
All ages
Average of monitors
within each city
8-h max
Year-round
Mean: 41.8
Median: 39.7
75th: 51.3
90th: 62.3
Maximum:
107
Correlation (r):
NR
Copollutant
models: NR
Lag 0 RRs
All ages: 1.01
(1.00, 1.03)
5-14 yr: 1.05 (1.02,
1.10)
15-64 yr: 1.01
(0.99, 1.03)
65+ yr: 0.99 (0.95,
1.03)
CI = confidence interval; DL = distributed lag; NR = not reported; RR = relative risk.
tStudies published since the 2013 Ozone ISA.
aResults standardized to a 15-ppb increase in 24-h avg, 20-ppb increase in 8-h daily max, or 25-ppb increase in 1-h daily max
ozone concentrations.
3-146
-------
Table 3-14 Epidemiologic studies of short-term exposure to ozone and
emergency department (ED) visits for asthma.
Study
Copollutant
Effect Estimates
Study
Population
Exposure Assessment
Mean (ppb)
Examination
(95% CI)
Stieb et al. (2009)
All ages
Average of monitors in
Mean: 18.4
Correlation
Percentage
Seven Canadian
each city
75th:
(r):
increase
cities
24-h avg
19.3-28.6
Warm season
Lag 1: 2.6 (0.2,
Ozone: 1992-2003
Year-round
across cities
(across
5.0)
Follow-up:
cities):
PM2.5: -0.05,
1992-2003
0.62;
Time-series study
NO2: -0.17,
0.10;
SO2: -0.24,
0.21;
CO: -0.34,
0.17
Cold season:
PM2.5: -0.65,
0.06;
NO2: -0.57,
-0.35;
SO2: -0.52,
-0.18;
CO: -0.67,
-0.16
Copollutant
models: NR
Villeneuve et al.
n = 57,912
Average of three monitors.
Summer:
Correlation
Lag 1 OR
(2007)
All ages
8-h max
Mean: 38.0
(r): NR
All ages
Alberta, Canada
Year-round and seasonal
75th: 46.0
Copollutant
Year-round: 1.04
Ozone: 1992-2002
(April-September,
Winter:
models: NR
(1.02, 1.07)
Follow-up:
October-March)
Mean: 24.3
75th: 31.5
Winter: 1.02
1992-2002
(0.98, 1.06)
Time-series study
Summer: 1.07
(1.04, 1.10)
I to et al. (2007)
All ages
Average of 16 monitors
All year:
Correlation
Percentage
New York City, NY,
within 20 miles of the
Mean: 30.4
(r): NR
increase
U.S.
geographic city center
95th: 68.0
Copollutant
Lag 0-1
Ozone: 1999-2002
8-h max
Warm:
models:
Warm season:
Follow-up:
1999-2002
Year-round and seasonal
Mean: 42.7
PM2.5, NO2,
11.0 (7.1, 15.0)
(April-September,
95th: 77.0
SO2, CO
Time-series study
October-March)
Cold:
Mean: 18.0
95th: 33.0
3-147
-------
Table 3-14 (Continued): Epidemiologic studies of short-term exposure to ozone
and emergency department (ED) visits for asthma.
Study
Study
Population
Exposure Assessment
Copollutant Effect Estimates
Mean (ppb) Examination (95% CI)
tSarnat et al.
(2013)
Atlanta, GA, U.S.
Ozone: 1999-2002
Follow-up:
1999-2002
Time-series study
n = 270,816
All ages
Zip-code centroid
estimates from a hybrid
model fusing spatially
interpolated background
O3 concentrations with
local-scale AERMOD
output
24-h avg
Year-round
Mean: 41.9
Median: 39.3
75th: 53.8
95th: 76.2
Maximum:
132.7
Correlation
(r):
PM2.5: 0.51;
NOx: -0.03
Copollutant
models: NR
Lag 0-2 RRs
Overall: 1.03
(1.01, 1.04)
High AER: 1.02
(1.00, 1.04)
Low AER: 1.04
(1.02, 1.06)
tWinquist et al.
(2012)
St. Louis, MO, U.S.
Ozone: 2001-2007
Follow-up:
2001-2007
Time-series study
All ages
One monitor
8-h max
Year-round
Correlation
(r): NR
Copollutant
models: NR
RR
0-4 DL: 1.05
(1.02, 1.08)
tSacks et al. (2014)
North Carolina
(statewide), U.S.
Ozone: 2006-2008
Follow-up:
2006-2008
Case-crossover
study
n = 122,607
All ages
CMAQ model estimates
predicted to census tract
centroids and aggregated
to the county-level using
area-weighted average of
census tract centroids
8-h max
Seasonal: Warm season
(April-October) and
year-round estimates
Mean:
Correlation
All-year:
(r): PM2.5:
43.6; warm
0.54
season: 50.1
Copollutant
75th:
models:
All-year:
PM2.5
54.3; warm
season: 59.2
Maximum:
All-year:
108.1; warm
season:
108.1
Lag 0-2 ORs
All-year: 1.02
(1.00, 1.04)
+PM2.5:
1.01 (0.99, 1.04)
Warm season:
1.02 (1.00, 1.04)
+PM2.5:
1.02 (0.98, 1.05)
tWinquist et al.
(2014)
Atlanta, GA, U.S.
Ozone: 1998-2004
Follow-up:
1998-2004
Time-series study
Age: 5-17 yr
Population-weighted
monitor averages
8-h max
Seasonal:
Cold season
(November-April) and
warm season
(May-October) estimates
Mean: 53.9
Median: 53.3
75th: 67.7
Correlation
(r):
PM2.5: 0.66;
NO2: 0.54;
SO2: 0.27
Copollutant
models: NR
Lag 0-2 RRs
Cold season
(November-April):
1.05 (1.01, 1.09)
Warm season
(May-October):
1.05 (1.02, 1.09)
3-148
-------
Table 3-14 (Continued): Epidemiologic studies of short-term exposure to ozone
and emergency department (ED) visits for asthma.
Study
Copollutant
Effect Estimates
Study
Population
Exposure Assessment
Mean (ppb)
Examination
(95% CI)
tBarrvetal. (2018)
All ages
Fusion of CMAQ model
Mean:
Correlation
RR
Five U.S. cities
estimates and
37.5-42.2
(r): NR
Lag 0-2
Ozone: 2002-2008
Follow-up:
ground-based
measurements;
population-weighted
75th:
50.1-54.4
90th:
59.3-63.5
Copollutant
models: NR
Atlanta:
1.03 (1.01, 1.05)
2002-2008
average of 12-km grid
Birmingham:
Time-series study
cells for each city
1.03 (0.98, 1.08)
8-h max
Year-round
Maximum:
80.2-106.3
Dallas:
1.03 (1.00, 1.06)
Pittsburgh:
1.03 (1.00, 1.06)
St. Louis:
1.06 (1.02, 1.09)
tGleason et al.
n = 21,854
Fusion of monitor and
Correlation
OR
(2014)
Age: 3-17 yr
CMAQ modeling
(r): PM2.5:
Lag 0-2: 1.06
New Jersey
8-h max
0.56
(1.05, 1.08)
(statewide), U.S.
Warm season
Copollutant
Ozone:
(April-September)
models:
April-September,
PM2.5
2004-2007
Follow-up:
April-September,
2004-2007
Case-crossover
study
tStrickland et al.
(2014)
Atlanta, GA, U.S.
Ozone: 2002-2010
Follow-up:
2002-2010
Time-series study
tSarnat et al. n = 34,086
(2015) ah ages
St. Louis, MO, U.S.
Ozone: 2001-2004
Follow-up:
2001-2003
Time-series study
n = 109,758 Population-weighted Mean: 42.22 Correlation RR
Age: 2-16 yr monitor averages (r): NR Lag 0_2: 107
8-h max Copollutant (1.04,1.09)
Year-round models: NR
One monitor Mean:
8-h max
Year-round
i.2 Correlation RR
(r): 0-2 DL: 1.05
PM2.5: 0.23; m go 1 09)
NO2: 0.37;
SO2: -0.04;
SO42": 0.49;
NO3-: -0.57;
OC: 0.30;
EC: -0.09
Copollutant
models:
PM2.5, S042",
EC, OC, NO2
3-149
-------
Table 3-14 (Continued): Epidemiologic studies of short-term exposure to ozone
and emergency department (ED) visits for asthma.
Study
Study
Population
Exposure Assessment
Copollutant Effect Estimates
Mean (ppb) Examination (95% CI)
tBvers et al. (2015) n = 165,056
Indianapolis, IN, Age: >5 yr
U.S.
Ozone: 2007-2011
Follow-up:
2007-2011
Time-series study
Inverse-distance and
population-weighted
average of 11 monitors
8-h max
Seasonal:
Cold season
(October-March) and
warm season
(April-September)
estimates
Mean: 48.5
Correlation
(r): PM2.5:
0.54;
SO2: 0.42
Copollutant
models: NR
Lag 0-2 RRs
Warm season
All ages:
1.03 (0.99, 1.07)
5-17 yr: 1.05
(0.97, 1.14)
18-44 yr: 1.06
(1.00, 1.12)
45+ yr: 0.97 (0.91,
1.04)
tAlhanti et al.
(2016)
Three U.S. cities,
U.S.
Ozone: 1993-2009
Follow-up:
1993-2009
Time-series study
n = 611,970
All ages
Population weighted
monitor averages, using
all monitors in each city
8-h max
Year-round
Mean:
Range
across cities:
37.3 to 43.7
Correlation
(r): NR
Copollutant
models: NR
Lag 0-2 RRs
0-4 yr: 1.02 (1.01,
1.04)
5-18 yr: 1.05
(1.03, 1.07)
19-39 yr: 1.02
(1.00, 1.04)
40-64 yr: 1.01
(0.99, 1.03)
65+ yr: 1.02 (0.98,
1.06)
tSheffield et al.
(2015)
New York City, NY,
U.S.
Ozone:
May-September
2005-2012
Follow-up:
May-September
2005-2011
Case-crossover
study
n = 8,009
Age: 5-17 yr
Average of city monitors Maximum: 60 Correlation Percentage
24-h avg
Warm season
(May-September)
(r): NR
Copollutant
models: NR
increase
Lag 0-3: 10.81
(6.84, 15.03)
tMalia et al. (2016)
California
(statewide), U.S.
Ozone: 2005-2009
Follow-up:
2005-2008
Case-crossover
study
All ages Nearest monitor within
20 km of population
weighted zip-code centroid
1-h max
Seasonal:
Warm season
(May-October) and
year-round estimates
Mean: 33-55
across
climate zones
Correlation
(r):
NO2: -0.01
YR;
0.26 warm;
SO2: -0.06
YR;
0.02 warm;
CO: -0.28
YR;
0.02 warm
Copollutant
models: NO2,
CO, SO2
Percentage
increase
Lag 0-1
Year-round: 3.85
(2.17, 5.56)
Lag 0-3
Warm season:
6.79 (3.67, 10.01)
+NO2
3.89 (0.99, 6.87)
+SO2
4.49 (1.49, 7.57)
+CO
4.12 (1.36, 6.95)
3-150
-------
Table 3-14 (Continued): Epidemiologic studies of short-term exposure to ozone
and emergency department (ED) visits for asthma.
Study
Study
Population
Exposure Assessment
Copollutant Effect Estimates
Mean (ppb) Examination (95% CI)
tShmool et al.
(2016)
New York City, NY,
U.S.
Ozone:
June-August
2005-2011
Follow-up:
June-August
2005-2011
Case-crossover
study
n = 11,719
Age: 5-17 yr
Temporal estimates:
Average of city monitors;
spatiotemporal estimates:
Fusion of monitors and
LUR
24-h avg
Warm season
(May-September)
Mean:
Temporal:
30.4;
spatio-
temporal:
29.0
Maximum:
Temporal:
60.0;
spatio-
temporal:
60.3
Correlation
(r): NR
Copollutant
models: NR
No quantitative
results. Results
presented
graphically
tO'Lenick et al.
(2017)
Atlanta, GA, U.S.
Ozone: 2002-2008
Follow-up:
2002-2008
Case-crossover
study
n = 128,758
Age: 5-17 yr
Fusion of CMAQ model
estimates and
ground-based
measurements; 12-kmgrid
cells
8-h max
Year-round
Correlation
(r): NR
Copollutant
models: NR
OR
Lag 0-2: 1.06
(1.03, 1.08)
tXiao et al. (2016)
Georgia (statewide),
U.S.
Ozone: 2002-2008
Follow-up:
2002-2008
Case-crossover
study
n = 148,256
Age: 2-18 yr
Fusion of CMAQ model
estimates and
ground-based
measurements; 12-kmgrid
cells
8-h max
Year-round
Mean: 42.1
75th: 50.9
Maximum:
106.1
Correlation
(r):
PM2.5: 0.61;
NO2: -0.12;
SO2: -0.03;
SO42": 0.61;
NOs": -0.39;
OC: 0.35;
EC: 0.01
Copollutant
models: NR
OR
Lag 0-3: 1.03
(1.01, 1.05)
tSzvszkowicz et al.
(2018)
Multicity, Canada
Ozone: 2004-2011
Follow-up:
2004-2011
Case-crossover
study
n = 223,845
Age: 0-19 yr
Average of all monitors
within 35 km
24-h avg
Year-round
Mean:
22.5-29.2
across cities
Maximum: 80
Correlation
(r): NR
Copollutant
models: NR
OR
Lag 1:
Males: 1.03 (1.00,
1.06)
Females: 1.04
(1.00, 1.08)
AERMOD = American Meteorological Society/EPA Regulatory Model; CI = confidence interval; CMAQ = Community Multiscale Air
Quality; DL = distributed lag; KM = kilometers; LUR = land use regression; NR = not reported; OR = odds ratio; RR = relative risk.
tStudies published since the 2013 Ozone ISA.
Results standardized to a 15-ppb increase in 24-h avg, 20-ppb increase in 8-h daily max, or 25-ppb increase in 1 -h daily max
ozone concentrations.
3-151
-------
Table 3-15 Epidemiologic studies of short-term exposure to ozone and
respiratory symptoms in children with asthma.
Study
Study
Population
Exposure Assessment
Mean
ppb
Copollutant
Examination
Effect Estimates
(95% CI)
tLewis et al. (2013)
Detroit, Ml, U.S.
Ozone: 1999-2002
Follow-up: 1999-2002
Panel study
n = 298
Children with
asthma,
primarily
African
American and
Latino, living in
low-income
communities
Age: 5-12 yr
One rooftop school
monitor for each of
two communities.
95% of participants
lived within 5 km of
one of the monitors
8-h max
Year-round
Mean:
41.8
Correlation (r):
PM25: 0.55
Copollutant
models: NR
No quantitative
results. Results
presented
graphically.
CI = confidence interval; NR = not reported.
tStudies published since the 2013 Ozone ISA.
Table 3-16 Study-specific details from controlled human exposure studies of
lung function in adults with asthma.
Study
Population
n, Sex, Age (Range or
Mean ± SD)
Exposure Details
(Concentration,
Duration)
Endpoints Examined
Bartoli et al. (2013)
Adults with asthma
0 ppb, 2 h
FEVi (2 days before exposure)
n = 86 males, 34 females
300 ppb, 2 h
FEVi (before and PE)
Age: 32.9 ± 12.9 yr
FE\A| = forced expiratory volume in 1 second; PE = post-exposure.
3-152
-------
Table 3-17 Study-specific details from controlled human exposure studies of
lung function in healthy adults and adults with asthma.
Study
Population n, Sex, Age
(Range or Mean ± SD)
Exposure Details
(Concentration,
Duration)
Endpoints Examined
Frvetal. (2012)
Healthy adults and adults with
asthma
n = 12 males, 15 females
Age: 21-35 yr
400 ppb, 2 h FEVi (immediately before and
PE)
Ariomandi et al. (2015)
Healthy adults and adults with
asthma
n = 13 males, 13 females
Age: 31.8 ± 7.6 yr
0 ppb, 4 h
100 ppb, 4 h
200 ppb, 4 h
FEVi, FVC, FEVi/FVC (before,
after, 20 h PE)
Lerov et al. (2015)
Asthmatic adults
n = 3 males, 4 females
Age: 33.7 ± 10.1 yr
Nonasthmatic adults
n = 7 males, 5 females
Age: 31.8 ± 6.0 yr
0 ppb, 4 h
100 ppb, 4 h
200 ppb, 4 h
FEVi, FVC (before, immediately
after, and 20 h PE)
FE\A| = forced expiratory volume in 1 second; FVC = forced vital capacity; PE = post-exposure.
Table 3-18 Study-specific details from animal toxicological studies of short-term
ozone exposure and lung function—allergy.
Study
Species (Stock/Strain), n,
Sex, Age
Exposure Details
(Concentration,
Duration)
Endpoints Examined
Scheleale and Walbv
(2012)
Rats (BN) naive and
sensitized/challenged with
allergen
n = 5-11 males
Age: 8-10 weeks
1 ppm,
Lung function, breathing pattern
(immediately PE)
BN = brown Norway; PE = post-exposure.
3-153
-------
Table 3-19 Study-specific details from animal toxicological studies of short-term
ozone exposure and lung function—asthma.
Species (Stock/Strain), n, Exposure Details
Study Sex, Age (Concentration, Duration) Endpoints Examined
Bao et al. (2013) Mice (BALB/c naive and 2 ppm, 3 h Enhanced pause, MCh
ovalbumin challenge (24 h PE)
sensitized/challenged)
n = 6-7 females
Age: 6-8 weeks
Hansen et al. (2016) Mice (BALB/cJ) naive and 2 ppm, 1 h/day for 3 days Breathing frequency, tidal
ovalbumin sensitized volume, time of brake, time
n = 5 females of Pause (during exposure)
Age: 6 weeks
MCh = methacholine. PE = post-exposure.
3-154
-------
Table 3-20 Study-specific details from controlled human exposure studies of
inflammation, oxidative stress, and injury in healthy adults and
adults with asthma.
Study
Population n, Sex, Age
(Range or Mean ± SD)
Exposure Details
(Concentration,
Duration)
Endpoints Examined
Frvetal. (2012)
Healthy adults and adults
with asthma
n = 12 males, 15 females
Age: 21-35 yr
400 ppb, 2 h
Induced sputum PMN (48 h
before and 5 h PE)
Hernandez et al. (2012) Healthy adults
N = 14 males, 20 females
Age: 24.2 ± 3.9 yr
Atopic adults with asthma
N = 7 males, 10 females
Age: 24.4 ± 5.5 yr
400 ppb, 2 h
Induced sputum (screening visit
and 4 h PE)
Ariomandi et al. (2015)
Healthy adults and adults
with asthma
n = 13 males, 13 females
Age: 31.8 ± 7.6 yr
0 ppb, 4 h
100 ppb, 4 h
200 ppb, 4 h
BALF protein, PMNs, and
eosinophils (20 h PE)
Lerov et al. (2015)
Healthy adults and adults
with asthma
n = 14 males, 2 females
Age: 32.5 ± 7.6 yr
0 ppb, 4 h
100 ppb, 4 h
200 ppb, 4 h
BALF (20 h PE)
BALF = bronchoalveolar lavage fluid; PE = post-exposure; PMN = polymorphonuclear leukocyte (e.g., neutrophils).
Table 3-21 Study-specific details from controlled human exposure studies of
allergic sensitization—atopy.
Study
Population n, Sex, Age
(Range or Mean ± SD)
Exposure Details
(Concentration,
Duration)
Endpoints Examined
Dokic and Traikovska-Dokic Adults with atopy
(2013) n = 5 males, 5 females
Age: 27.9 ± 6.6 yr
0 ppb before pollen
season, 2 h
0 ppb pollen season, 2 h
400 ppb before pollen
season, 2 h
400 ppb pollen season, 2 h
Nasal lavage (2 h before,
immediately after, and 6 h
PE)
PE = post-exposure.
3-155
-------
Table 3-22 Study-specific details from animal toxicological studies of short-term
ozone exposure and inflammation, oxidative stress, and
injury—allergy.
Exposure Details
Species (Stock/Strain), n,
(Concentration,
Study Sex, Age
Duration) Endpoints Examined
Scheleqle and Walbv Rats (BN) naive and 1 ppm, 8 h BALF markers of injury,
(2012) sensitized/ inflammation (immediately PE)
challenged with allergen
n = 5-11 males
Age: 8-10 weeks
BALF = bronchoalveolar lavage fluid; BN = brown Norway; PE = post-exposure.
Table 3-23 Study-specific details from animal toxicological studies of short-term
ozone exposure and inflammation, oxidative stress, and
injury—asthma.
Study
Species (Stock/Strain), n,
Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Bao et al. (2013)
Mice (BALB/c naive and
ovalbumin
sensitized/challenged)
n = 6-7 females
Age: 6-8 weeks
2 ppm, 3 h
BALF total cells and cell
differentials, mediators
(24 h PE)
BALF = bronchoalveolar lavage fluid; PE = post-exposure.
3-156
-------
Table 3-24 Study-specific details from animal toxicological studies of short-term
ozone exposure and morphology—asthma.
Study
Species (Stock/Strain), n, Exposure Details
Sex, Age (Concentration, Duration) Endpoints Examined
Bao et al. (2013)
Mice (BALB/c naive and
ovalbumin
sensitized/challenged)
n = 6-7 females
Age: 6-8 weeks
2 ppm, 3 h
Mucosubstance secretion
and MUC5AC, epithelial cell
density (24 h PE)
MUC5AC = mucin 5AC glycoprotein; PE = post-exposure.
Table 3-25 Epidemiologic studies of short-term exposure to ozone and
inflammation, oxidative stress, and injury in children with asthma.
Study
Study
Population
Exposure Assessment
Mean
ppb
Copollutant
Examination
Effect Estimates
(95% Cl)a
tDelfino et al. (2013)
Riverside and Whittier,
CA, U.S.
Ozone:
August-December
2003 (Riverside);
July-November 2004
(Whittier)
Follow-up:
August-December
2003 (Riverside);
July-November 2004
(Whittier)
Panel study
n = 45
Children with
asthma
Age: 9-18 yr
One monitor
(Riverside);
average of two
monitors (Whittier)
8-h max
Year-round
Mean:
52.9
Median:
46.8
Maximum:
120.8
Correlation (r): Lag 0: 0.42
PM25: 0.39;
NO2: 0.07;
EC: 0.55;
OC: 0.71
Copollutant
models: NR
(-1.33, 2.19)
Lag 1: 0.63
(-1.05, 2.35)
Lag 0-2: 1.24
(-1.04, 3.57)
CI = confidence interval; EC = elemental carbon; OC = organic carbon; NR = not reported; RR = relative risk.
tStudies published since the 2013 Ozone ISA.
aResults standardized to a 15-ppb increase in 24-h avg, 20-ppb increase in 8-h daily max, or 25-ppb increase in 1-h daily max
ozone concentrations.
3-157
-------
Table 3-26 Epidemiologic studies of short-term exposure to ozone and
emergency department (ED) visits for chronic obstructive pulmonary
disease (COPD).
Study
Study Population
Exposure
Assessment
Mean
ppb
Copollutant
Examination
Effect
Estimates
(95% Cla)
Arbex et al. (2009)
Sao Paulo, Brazil
Ozone: 2001-2003
Follow-up:
2001-2003
Time-series study
n = 1,769
All ages
Average of four
monitors
1-h max
Year-round
Mean: 48.8
75th: 61.0
Maximum:
143.8
Correlation (r): Percent
NR increase
Copollutant
models: NR
Women
Lag 0: 1.0
(0.0, 2.0)
tRodopoulou et al.
n = 12,511
One monitor
Mean: 40
Correlation (r):
Percent
(2015)
Ages 15+
8-h max
Median: 39
PM25: 0.33
increase
Little Rock, AR, U.S.
Seasonal:
75th: 50
Copollutant
Warm
Ozone: 2002-2012
cold season
models: PM2.5
season
Follow-up:
(October-March)
Lag 2:
2002-2012
and warm season
3.4 (-2.3,
Time-series study
(April-September)
estimates
9.3)
+PM2.5:
2.9 (-3.3,
9.5)
tSarnat et al. (2015)
n = 34,086
One monitor
Mean: 36.2
Correlation (r):
RR
St. Louis, MO, U.S.
All ages
8-h max
PM2.5: 0.23;
Lag 0-2
Ozone: 2001-2004
Follow-up:
Year-round
NO2: 0.37;
SO2: -0.04;
SO42": 0.49;
0.98 (0.92,
1.06)
2001-2003
NOs": -0.57;
Time-series study
OC: 0.30;
EC: -0.09
Copollutant
models: PM2.5,
SO42", EC,
OC, NO2
3-158
-------
Table 3-26 (Continued): Epidemiologic studies of short-term exposure to ozone
and emergency department (ED) visits for chronic
obstructive pulmonary disease (COPD).
Study
Study Population
Exposure
Assessment
Mean
ppb
Copollutant
Examination
Effect
Estimates
(95% Cla)
tMalia et al. (2016)
California
(statewide), U.S.
Ozone: 2005-2009
Follow-up:
2005-2008
Case-crossover
study
All ages
Nearest monitor Mean: 33-55
within 20 km of across climate
population-weighted zones
zip-code centroid
1-h max
Seasonal:
warm season
(May-October) and
year-round
estimates
Correlation (r):
N02: -0.01
YR;
0.26 warm;
SO2:
-0.06 YR;
0.02 warm;
CO:
-0.28 YR;
0.02 warm
Copollutant
models: NO2,
CO, SO2
Percent
increase
Lag 0-1
Year-round:
0.89 (-0.26,
2.06)
Lag 3
Warm
season:
2.11 (0.08,
4.17)
+NO2 1.37
(-1.05,
3.85)
+S02 2.54
(-0.29,
5.45)
+CO 1.46
(-1.15,
5.45)
tXiao et al. (2016)
n = 84,597
Fusion of CMAQ
Mean: 42.1
Correlation (r):
OR
Georgia (statewide),
Ages: 2-18 yr
model estimates
75th: 50.9
PM2.5: 0.61;
Lag 0-3:
U.S.
and ground-based
measurements;
Maximum:
NO2: -0.12;
SO2: -0.03;
1.03 (1.00,
1.06)
Ozone: 2002-2008
12-km grid cells
106.1
SO42": 0.61;
Follow-up:
8-h max
NOs": -0.39;
2002-2008
OC: 0.35;
Case-crossover
Year-round
EC: 0.01
study
Copollutant
models: NR
tBarrvetal. (2018)
All ages
Fusion of CMAQ
Mean:
Correlation (r):
RR
Five U.S. cities
model estimates
37.5-42.2
NR
Lag 0-2
Ozone: 2002-2008
and ground-based
measurements;
75th:
50.1-54.4
Copollutant
models: NR
Atlanta:
1.00 (0.97,
1.03)
Follow-up:
population-weighted
90th:
59.3-63.5
2002-2008
average of 12-km
Time-series study
grid cells for each
Birmingham:
city
8-h max
Maximum:
80.2-106.3
0.99 (0.93,
1.05)
Year-round
Dallas:
1.04 (0.99,
1.09)
Pittsburgh:
1.02 (0.98,
1.07)
St. Louis:
1.03 (0.99,
1.08)
3-159
-------
Table 3-26 (Continued): Epidemiologic studies of short-term exposure to ozone
and emergency department (ED) visits for chronic
obstructive pulmonary disease (COPD).
Effect
Exposure Mean Copollutant Estimates
Study Study Population Assessment ppb Examination (95% Cla)
tSzvszkowicz et al.
(2018)
Multicity, Canada
Ozone: 2004-2011
Follow-up:
2004-2011
Case-crossover
Study
n = 183,544
Age: 55+ yr
Average of all
monitors within
35 km
24-h avg
Year-round
Mean:
22.5-29.2
across cities
Maximum: 80
Correlation (r):
NR
Copollutant
models: NR
OR
Lag 1;
females:
1.01 (0.99,
1.03)
Lag 0;
males: 1.01
(0.99, 1.03)
CI = confidence interval; CMAQ = Community Multiscale Air Quality; EC = elemental carbon; Km = kilometers; LUR = land use
regression; N03"= nitrate; NR = not reported; OC = organic carbon; OR = odds ratio; RR = relative risk; S042" = sulfate.
fStudies published since the 2013 Ozone ISA.
aResults standardized to a 15-ppb increase in 24-h avg, 20-ppb increase in 8-h daily max, or 25-ppb increase in 1-h daily max
ozone concentrations.
3-160
-------
Table 3-27 Epidemiologic studies of short-term exposure to ozone and
medication use in adults with chronic obstructive pulmonary disease
(COPD).
Study
Study
Population
Exposure Assessment
Mean
ppb
Copollutant
Examination
Effect Estimates
(95% Cla)
tMaqzamen et al.
(2018)
Seattle and Tacoma,
WA, U.S.
Ozone: December
2011 to October 2012
Follow-up: December
2011 to October 2012
Panel study
n = 35
Age: 40+ yr
Former
smokers with
COPD but not
asthma
Monitors
8-h max
Year-round
Median:
17.21
75th:
24.37
Maximum:
40.86
Correlation (r):
NR
Copollutant
models: NR
RR
Lag 0: 0.98 (0.93,
1.45)
CI = confidence interval; NR = not reported; RR = relative risk.
fStudies published since the 2013 Ozone ISA.
aResults standardized to a 15-ppb increase in 24-h avg, 20-ppb increase in 8-h daily max, or 25-ppb increase in 1-h daily max
ozone concentrations.
Table 3-28 Study-specific details from animal toxicological studies of short-term
ozone exposure and lung function—chronic obstructive pulmonary
disease (COPD).
Species (Stock/Strain), n,
Study Sex, Age
Exposure Details
(Concentration,
Duration)
Endpoints Examined
Groves et al. (2012)
Mice (C57BL/6J WT" and
surfactant protein D
deficient)
n = 4-9 males
Age: 8 weeks
0.8 ppm, 3 h
Pulmonary mechanics (72 h PE)
Groves et al. (2013)
Mice (C57BL/6J WT and
surfactant protein D
deficient)
n = 3-10 males
Age: 8, 27, 80 weeks
0.8 ppm, 3 h
Resistance and elastance
spectra (72 h PE)
PE = post-exposure; WT = wild type.
3-161
-------
Table 3-29 Study-specific details from animal toxicological studies of short-term
ozone exposure and inflammation, oxidative stress, and
injury—chronic obstructive pulmonary disease (COPD).
Study
Species
(Stock/Strain), n,
Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Groves et al. (2012) Mice (C57BL/6J WT
and surfactant
protein D deficient)
n = 4-9 males
Age: 8 weeks
0.8 ppm, 3 h
BALF protein, RNS, macrophage
number, chemotactic activity
(24-72 h PE)
Groves et al. (2013) Mice (C57BL/6J WT
and surfactant
protein D deficient)
n = 3-10 males
Age: 8, 27, 80 weeks
0.8 ppm, 3 h
BALF total cells and differential cell
counts, protein (72 h PE)
BALF = bronchoalveolar lavage fluid; PE = post-exposure; RNS = reactive nitrogen species; WT = wild type.
Table 3-30 Study-specific details from animal toxicological studies of short-term
ozone exposure and morphology—chronic obstructive pulmonary
disease (COPD).
Species (Stock/Strain), n,
Study Sex, Age
Exposure Details
(Concentration,
Duration)
Endpoints Examined
Groves et al. (2012)
Mice (C57BL/6J WT and
surfactant protein D
deficient)
n = 4-9 males
Age: 8 weeks
0.8 ppm, 3 h
Histopathology,
immunohistochemistry (72 h PE)
Groves et al. (2013)
Mice (C57BL/6J WT and
surfactant protein D
deficient)
n = 3-10 males
Age: 8, 27, 80 weeks
0.8 ppm, 3 h
Markers of cell proliferation,
radial alveolar counts, lesion
scores (72 h PE)
PE = post-exposure, WT = wild type.
3-162
-------
Table 3-31 Study-specific details from controlled human exposure studies of
respiratory effects in obese adults.
Study
Population n, Sex, Age
(Range or Mean ± SD)
Exposure Details
(Concentration, Duration)
Endpoints Examined
Bennett et al. (2016)
Healthy adult women
n = 19 obese
Age: 28 ± 5 yr
n = 19 normal weight
Age: 24 ± 4 yr
0 ppb, 2 h
400 ppb, 2 h
FEV-i, FVC, sGaw (before and
PE); airway responsiveness (3 h
PE)
PFT, PMN, airway
responsiveness, symptoms (train
day, before, immediately after,
and 3 h PE); symptoms
(immediately PE)
FE\A| = forced expiratory volume in 1 second; FVC = forced vital capacity; PE = post exposure; PFT = pulmonary function test;
PMN = polymorphonuclear leukocyte (e.g., neutrophils); sGaw = specific airways conductance.
Table 3-32 Study-specific details from animal toxicological studies of short-term
ozone exposure and lung function—obesity.
Study
Species (Stock/Strain), n,
Sex, Age
Exposure Details
(Concentration,
Duration)
Endpoints Examined
Williams et al. (2015) Mice (C57BL/6) WT and Cpe 2 ppm, 3 h
fat, TNF-a sufficient and
deficient
n = 5-9 females
Age:10-12 weeks
Pulmonary mechanics,
challenge with MCh (24 h PE)
Mathews et al.
(2017b)
Mice (C57BL/6J) WT and
db/db, TCR gamma delta
deficient mice on high-fat diet
for 24 weeks
n = 4-10 females
Age: 10 weeks and greater
than 24 weeks
2 ppm, 3 h
Pulmonary mechanics,
challenge with MCh (24 h PE)
Mathews et al. (2018) Mice (C57BL/6J) WT and 2 ppm, 3 h Pulmonary mechanics,
db/db, Cpe fat/TNFR2 challenge with MCh (24 h PE)
deficient mice; some animals
on high-fat diet for 24 weeks
n = 4-14 females
Age: 10 weeks and older than
24 weeks
MCh = methacholine; PE = post-exposure; TCR = T cell receptor; TNF = tumor necrosis factor; TNFR2 = tumor necrosis factor
receptor 2.
3-163
-------
Table 3-33 Study-specific details from animal toxicological studies of short-term
ozone exposure and inflammation, oxidative stress, and
injury—obesity.
Species
(Stock/Strain), n, Exposure Details
Study Sex, Age (Concentration, Duration) Endpoints Examined
Yinq et al. (2016) Mice (KKAy) 0.5 ppm, 4 h/day for 13 days Lung tissue mRNA for
n = 8 males proinflammatory genes; T cell
„ , subpopulations in lymph nodes (about
Age. 6 ( weeks 2 |-j
Zhona et al. (2016) Mice (KKAy) 0.5 ppm, 4 h/day for 13 days BALF total cells and cell differentials
n = 8 (22 h PE)
Sex and age: NR
BALF total cells and cell differentials,
cytokines, chemokines; flow
cytometry of isolated cells from lung
tissue (24 h PE)
n = 4-10 females
Age:10 weeks and
greater than 24 weeks
Mathews et al. Mice (C57BL/6J) WT 2 ppm, 3 h
(2017b) and db/db, WT and
TCR gamma delta
deficient mice on high
fat diet for 24 weeks
Mathews et al. Mice (C57BL/6J) WT 2 ppm, 3 h Markers of oxidative stress (24 h PE)
(2017a) and db/db
n = 5-8 females
Age: 10 weeks
Mathews et al. Mice (C57BL/6J) WT 2 ppm, 3 h
(2018) and db/db, and Cpe
fat/TNFR2 deficient
mice, WT and TCR
gamma delta-deficient
mice; some animals
on high fat diet for
24 weeks
n = 4-14 females
Age: 10-12 weeks
and older than
24 weeks
BALF total cells and cell differentials,
IL-17A, IL-23, IL-33, CCL20, CXCL1,
CXCL2, IL-6, G-CSF, GRP; lung
tissue flow cytometry for IL-17A
producing cells (24 h PE)
BALF = bronchoalveolar lavage fluid; CCL20 = CC motif chemokine ligand 20; CXCL = chemokine family of cytokines with highly
conserved motif: cys-xxx-cys (CXC) ligand; G-CSF = granulocyte colony-stimulating factor (receptor); GRP = gastrin-releasing
peptide; IL = interleukin; mRNA = messenger ribonucleic acid; NR = not reported; PE = post-exposure; TCR = T cell receptor;
WT = wild type.
3-164
-------
Table 3-34 Study-specific details from animal toxicological studies of short-term
ozone exposure and morphology—obesity.
Study
Species
(Stock/Strain),
n, Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Zhona et al. (2016)
Mice (KKAy) 0.5 ppm, 4 h/day for 13 days
n = 8
Sex and age:
NR
Qualitative histopathology (22 h
PE)
NR = not reported; PE = post-exposure.
Table 3-35 Epidemiologic studies of short-term exposure to ozone and
pulmonary inflammation in populations with metabolic syndrome.
Study
Study Population
Exposure
Assessment
Mean
ppb
Copollutant
Examination
Effect Estimates
(95% Cl)a
tPena et al. (2016)
Boston, MA, U.S.
Ozone: 2006-2010
Follow-up: 2006-2010
Panel study
Adults with type 2
diabetes mellitus.
Mostly white
population (83%)
n = 69
Age: 44-85 yr
One monitor
24-h avg
Year-round
Median:
26.76
75th:
32.57
Correlation (r): Estimated percent
PM25: 0.22;
NOx: -0.35;
BC: 0.28;
OC: 0.24;
sulfate: 0.32;
PN: -0.78
Copollutant
models: PM2.5
change in FeNO
(ppb);
Lag 1: -5.99
(-10.67, -0.91)
+PM2.5: -7.14
(-12.45, -1.51)
BC = black carbon; CI = confidence interval; FeNO = fraction exhaled nitric oxide; OC = organic carbon; PN = particle number.
tStudies published since the 2013 Ozone ISA.
aResults standardized to a 15-ppb increase in 24-h avg, 20-ppb increase in 8-h daily max, or 25-ppb increase in 1-h daily max
ozone concentrations.
3-165
-------
Table 3-36 Study-specific details from animal toxicological studies of short-term
ozone exposure and lung function—cardiovascular disease.
Study
Species
(Stock/Strain), n,
Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Dve et al. (2015)
Rats (WKY, WS, S-D,
SH, FHH, SPSH,
obese SHHF, obese
atherosclerosis prone
JCR rats)
n = 4-8 males
Age: 12-14 weeks
0.25 ppm, 4 h
0.5 ppm, 4 h
1 ppm, 4 h
Whole body plethysmography (0 and
20 h PE)
Zvchowski et al.
(2016)
Mice (C57BL/6)
control and mice with
induced pulmonary
hypertension
n = 4-8 males
Age: 6-8 weeks
1 ppm, 4 h
Airway responsiveness to MCh
(18-20 h PE)
FHH = fawn-hooded hypertensive; MCh = methacholine; PE = post-exposure; S-D = Sprague-Dawley; SH = spontaneously
hypertensive; SHHF = spontaneously hypertensive heart failure; SPSH = stroke-prone spontaneously hypertensive; WKY = Wistar
Kyoto; WS = Wistar.
3-166
-------
Table 3-37 Study-specific details from animal toxicological studies of short-term
ozone exposure and inflammation, oxidative stress, and
injury—cardiovascular disease.
Species (Stock/Strain), n,
Study Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Farrai et al. (2012)
Rats (SH)
n = 6males
Age: 12 weeks
0.2 ppm, 4 h
0.8 ppm, 4 h
BALF total cells and differential
cell counts, total protein, LDH,
NAG, SOD, GPx, GST (1 and
18 h PE)
Kodavanti et al.
(2015)
Rats (WKY, WS, SD, SH,
FHH, SPSH, obese SHHF,
obese atherosclerosis prone
JCR rats)
n = 4-8 males
Age:12-14 weeks
0.25 ppm, 4 h
0.5 ppm, 4 h
1 ppm, 4 h
BALF total cell counts and cell
differentials, BALF total protein,
albumin, LDH, NAG, GGT; lung
tissue mRNAforHO-1, MIP-2,
TNF-a, IL-6, IL-10 (BALF 0 and
20 h PE, tissue 0 h PE)
Ramot et al. (2015)
Rats (WKY, WS, SD, SH,
FHH, SPSH, obese SHHF,
obese atherosclerosis prone
JCR rats)
n = 4-8 males
Age:12-14 weeks
0.25 ppm, 4 h
0.5 ppm, 4 h
1 ppm, 4 h
Lung histopathology (0 and 20 h
PE)
Ward and Kodavanti Rats (WKY, SH, SPSH, 1 ppm, 4 h Lung gene expression profiling
(2015) obese SHHF, obese (immediately PE)
atherosclerosis prone JCR)
n = 3-4 males
Age:10-12 weeks
Hatch et al. (2015) Rats (WKY, WS, SD, SH, 1 ppm, 4 h BALF and tissue antioxidants (0
FHH, SPSH, SHHF, obese and 20 h PE)
atherosclerosis prone JCR)
n = 8 males
Age:12-14 weeks
Zvchowski et al. Mice (C57BL/6) control and 1 ppm, 4 h BALF cells and lung tissue
(2016) mice with induced indicators of injury (18-20 h PE)
pulmonary hypertension
n = 4-8 males
Age: 6-8 weeks
BALF = bronchoalveolar lavage fluid; FHH = fawn-hooded hypertensive; GGT = gamma glutamyl transferase; GPx = glutathione
peroxidase; GST = glutathione S-transferase; HO-1 = heme oxygenase 1; IL = interleukin; LDH = lactate dehydrogenase;
MIP-2 = macrophage inflammatory protein 2; mRNA = messenger ribonucleic acid; NAG = /V-acetyl-glucosaminidase;
PE = post-exposure; S-D = Sprague-Dawley; SH = spontaneously hypertensive; SHHF = spontaneously hypertensive heart failure;
SPSH = stroke-prone spontaneously hypertensive; SOD = superoxide dismutase; TNF-a = tumor necrosis factor alpha;
WKY = Wistar Kyoto, WS = Wstar.
3-167
-------
Table 3-38 Study-specific details from animal toxicological studies of short-term
ozone exposure and morphology—cardiovascular disease.
Species (Stock/Strain), n,
Study Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Ramotetal. (2015)
Rats (WKY, WS, SD, SH,
FHH, SPSH, obese SHHF,
obese atherosclerosis prone
JCR rats)
n = 4-8 males
Age: 12-14 weeks
0.25 ppm, 4 h
0.5 ppm, 4 h
1 ppm, 4 h
Lung histopathology (0 and
20 h PE)
Wong etal. (2018)
Rats (WKY, SH)
n = 8-12 males
Age:44-48 weeks
1 ppm, 6 h
Histopathology scores (8 h
PE)
FHH = fawn-hooded hypertensive; PE = post-exposure; S-D = Sprague-Dawley; SH = spontaneously hypertensive;
SHHF = spontaneously hypertensive heart failure; SPSH = stroke-prone spontaneously hypertensive; WKY = Wistar Kyoto;
WS = Wstar.
3-168
-------
Table 3-39 Epidemiologic studies of short-term exposure to ozone and
emergency department (ED) visits for respiratory infection.
Study
Study
Population
Exposure
Assessment
Copollutant
Mean ppb Examination
Effect Estimates
(95% CI)
Stieb et al. (2009)
Seven Canadian cities
Ozone: 1992-2003
Follow-up: 1992-2003
Time-series study
All ages Average of monitors
in each city
24-h avg
Year-round
18.4
Mean
75th:
19.3-28.6
across
cities
Correlation
(r):
Warm
season
(across
cities):
PM2.5:
-0.05, 0.62;
NO2: -0.17,
0.10;
SO2: -0.24,
0.21;
CO: -0.34,
0.17
Cold
season:
PM2.5:
-0.65, 0.06;
NO2: -0.57,
-0.35;
SO2: -0.52,
-0.18;
CO: -0.67,
-0.16
Copollutant
models: NR
Percent increase
Lag 1: 1.00 (0.98,
1.02)
tWinquist et al. (2012)
St. Louis, MO, U.S.
Ozone: 2001-2007
Follow-up: 2001-2007
Time-series study
All Ages
One monitor
8-h max
Year-round
Correlation
(r): NR
Copollutant
models: NR
RR
Pneumonia
0-4 DL: 1.01
(0.98, 1.04)
tKousha and Rowe (2014)
Edmonton, Canada
Ozone: 1999-2002
Follow-up: 1999-2002
Case-crossover study
n = 48,252 Three monitors
All ages 8-h max
Seasonal:
cold season
(October-March) and
warm season
(April-September)
estimates
Mean: 18.6 Correlation OR
Median:
17.8
(r): NR
Copollutant
models: NR
Lower respiratory
disease
Lag 0;
Year-round: 1.07
(1.03, 1.10)
3-169
-------
Table 3-39 (Continued): Epidemiologic studies of short-term exposure to ozone
and emergency department (ED) visits for respiratory
infection.
Study
Study
Population
Exposure
Assessment
Copollutant
Mean ppb Examination
Effect Estimates
(95% CI)
tDarrow et al. (2014)
Atlanta, GA, U.S.
Ozone: 1993-2010
Follow-up: 1993-2010
Time-series study
n = 80,399
Age: 0-4 yr
Population-weighted
monitor averages
8-h max
Seasonal:
cold season
(November-February)
and warm season
(March-October)
estimates
Mean: 45.9 Correlation
Median:
43.8
75th: 58.7
95th: 80.6
Maximum:
127.1
(r):
PM25: 0.3;
NO2: 0.37;
CO: 0.21
Copollutant
models: NR
Lag 0-2 RRs
Year-round
Bronchitis: 1.02
(0.99, 1.05)
URI: 1.03 (1.01,
1.05)
Pneumonia: 1.06
(1.03, 1.09)
tRodopoulou et al. (2015)
Little Rock, AR U.S.
Ozone: 2002-2012
Follow-up: 2002-2012
Time-series study
n = 13,650
Age: 15+ yr
One monitor
8-h max
Seasonal:
cold season
(October-March) and
warm season
(April-September)
estimates
Mean: 40
Median: 39
75th: 50
Correlation
(r):
PM2.5: 0.33
Copollutant
models: NR
Percent increase
Acute Rl; Lag 2:
-1.49 (-5.79,
3.00)
Pneumonia;
Lag 2: -8.19
(-16.64, 1.16)
tBarrv et al. (2018)
Five U.S. cities
Ozone: 2002-2008
Follow-up: 2002-2008
Time-series study
All ages Fusion of CMAQ
model estimates and
ground-based
measurements;
population-weighted
average of 12-km grid
cells for each city
8-h max
Year-round
Mean:
37.5-42.2
75th:
50.1-54.4
90th:
59.3-63.5
Maximum:
80.2-106.3
Correlation
(r): NR
Copollutant
models: NR
RR
URI—Lag 0-2
Atlanta: 1.02
(1.01, 1.04)
Birmingham: 1.02
(1.00, 1.05)
Dallas: 1.05
(1.02, 1.07)
Pittsburgh: 1.02
(1.00, 1.05)
St. Louis: 1.01
(0.99, 1.03)
tKousha and Castner (2016)
n = 4,815 Monitors in city
Mean: 25.3 Correlation
OR
Windsor, Canada
Age: 0-3 yr 8-h max
(r): NR
Otitis Media
Ozone: 2004-2010
Follow-up: 2004-2010
Case-crossover study
Year-round
Copollutant
models: NR
Lag 0;
Year-round: 1.04
(0.86, 1.21)
3-170
-------
Table 3-39 (Continued): Epidemiologic studies of short-term exposure to ozone
and emergency department (ED) visits for respiratory
infection.
Study
Study
Population
Exposure
Assessment
Copollutant
Mean ppb Examination
Effect Estimates
(95% CI)
tMaliq etal. (2016)
California (statewide), U.S.
Ozone: 2005-2009
Follow-up: 2005-2008
Case-crossover study
All ages Nearest monitor
within 20 km of
population-weighted
zip-code centroid
1-h max
Seasonal:
warm season
(May-October) and
year-round estimates
Mean:
33-55
across
climate
zones
Correlation
(r): NO2:
-0.01 YR;
0.26 warm;
SO2:
-0.06 YR;
0.02 warm;
CO:
-0.28 YR;
0.02 warm
Copollutant
models:
NO2, CO,
SO2
Percent increase
Pneumonia
Year-round
Lag 0-1: 1.32
(0.20, 2.46)
Warm season
Lag 0-3: 3.37
(0.80, 6.01)
+NO2 1.91
(-0.96, 4.87)
+SO2 4.69
(0.58, 8.96)
+C0 2.14
(-0.82, 5.20)
ARI
Year-round
Lag 0-1: 2.15
(1.45, 2.86)
Warm season
Lag 0-3: 3.41
(2.04, 4.81)
+N02 2.34
(0.68, 4.02)
+SO2 4.05
(1.54, 6.62)
+CO 2.44 (0.55,
4.36)
URTI
Year-round
Lag 0-1: 3.77
(0.40, 7.26)
Warm season
Lag 3:
4.23 (0.39, 8.21)
+NO2 2.83
(-0.96, 6.77)
+S023.21
(-1.24, 7.85)
+CO 3.04
(-0.97, 7.21)
3-171
-------
Table 3-39 (Continued): Epidemiologic studies of short-term exposure to ozone
and emergency department (ED) visits for respiratory
infection.
Study
Study
Population
Exposure
Assessment
Copollutant
Mean ppb Examination
Effect Estimates
(95% CI)
tXiao etal. (2016)
Georgia (statewide), U.S.
Ozone: 2002-2008
Follow-up: 2002-2008
Case-crossover study
n = 90,063
Age:
2-18 yr
Fusion of CMAQ
model estimates and
ground-based
measurements;
12-km grid cells
8-h max
Year-round
Mean: 42.1
75th: 50.9
Maximum:
106.1
Correlation
(r):
PM2.5: 0.61;
NO2: -0.12;
SO2: -0.03;
S042": 0.61;
NOs":
-0.39; OC:
0.35; EC:
0.01
Copollutant
models: NR
Lag 0-3 ORs
Otitis Media: 1.02
(1.01, 1.03)
Pneumonia: 1.04
(1.02, 1.07)
URI: 1.04 (1.03,
1.05)
tSzvszkowicz et al. (2018)
n = 717,676 Average of all
Mean:
Correlation
OR
Multicity, Canada
All ages monitors within 35 km
22.5-29.2
(r): NR
URI—Lag 0
Ozone: 2004-2011
24-h avg
across
cities
Maximum:
Copollutant
Females: 1.03
Follow-up: 2004-2011
Year-round
models: NR
(1.02, 1.05)
Males: 1.02 (1.01,
Case-crossover study
80
1.04)
ALRI—Lag 0
Females: 1.05
(1.02, 1.07)
Males: 1.02 (1.00,
1.05)
ALRI = acute lower respiratory infection; CI = confidence interval; CMAQ = Community Multiscale Air Quality; OC = organic
carbon; OR = odds ratio; Rl = respiratory infection; RR = relative risk; URI = upper respiratory infection; URTI = upper respiratory
tract infection.
tStudies published since the 2013 Ozone ISA.
aResults standardized to a 15-ppb increase in 24-h avg, 20-ppb increase in 8-h daily max, or 25-ppb increase in 1-h daily max
ozone concentrations.
3-172
-------
Table 3-40 Study-specific details from animal toxicological studies of short-term
ozone exposure and host defense/infection—healthy.
Study
Species (Stock/Strain),
n, Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Durrani et al. (2012)
Mice (C57BL/6)
2 ppm, 3 h
Survival after infection (14 days
n = 5 males, 5 females
PE)
Age: 10 weeks
Mikerov et al. (2011)
Mice (C57BL/6J)
n = 14 males,
11-14 females
Age: 8-12 weeks
2 ppm, 3 h
Lung, liver, and spleen
histopathology (48 h PE)
PE = post-exposure.
3-173
-------
Table 3-41 Epidemiologic studies of short-term exposure to ozone and hospital
admissions for aggregate respiratory diseases.
Study
Study
Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
(95% Cl)a
Katsouvanni et al.
NMMAPS
Average of monitors
NMMAPS:
Correlation (r):
Percent Increase
(2009)
90 U.S. cities
32 European cities
12 Canadian cities
APHEA
All ages
in each city.
1-h max
Year-round and
warm season
(April-September)
50th:
34.9-60.0
75th:
46.8-68.8
APHEA:
50th:
11.0-38.1
75th:
No
quantitative
results.
Results
presented
graphically.
Copollutant
models: NR
Lag 0-1
Year-round
U.S.:
1.5 (0.0, 3.0)
Canada:
5.0 (0.1, 10.1)
Warm season:
U.S.:
1.3 (-0.4, 3.1)
Canada:
18.9 (11.1, 27.4)
15.3-49.4
12 Canadian
cities:
50th: 6.7 8.3
75th: 8.4-12.4
Cakmak et al. (2006)
All ages
Average of monitors
Mean: 17.4
Correlation (r):
Percent increase
10 Canadian cities
in each city.
Max (across
NR
Lag 1
24-h avg
Year-round
cities):
38.0-79.0
Copollutant
models: NR
3.3 (1.7, 4.9)
tWinauist et al.
(2012)
St. Louis, MO, U.S.
Ozone: 2001-2007
Follow-up:
2001-2007
Time-series study
All ages
One monitor
8-h max
Year-round
Correlation (r):
NR
Copollutant
models: NR
RR
All ages
0-4 DL: 1.00 (0.98,
1.03)
2-18 yr
0-4 DL: 1.07 (1.00,
1.16)
APHEA = Air Pollution and Health: A European Approach; CI = confidence interval; DL = distributed lag; NMMAPS = National
Morbidity, Mortality, and Air Pollution Study; NR = not reported; RR = relative risk.
tStudies published since the 2013 Ozone ISA.
aResults standardized to a 15-ppb increase in 24-h avg, 20-ppb increase in 8-h daily max, or 25-ppb increase in 1-h daily max
ozone concentrations.
3-174
-------
Table 3-42 Epidemiologic studies of short-term exposure to ozone and
emergency department (ED) visits for aggregate respiratory
diseases.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
(95% Cl)a
Tolbert et al. (2007)
Atlanta, GA, U.S.
Ozone: 1993-2004
Follow-up:
1993-2004
Time-series study
n = 1,072,429
All ages
Average of monitors Mean: 53.0 Correlation (r): RR
in city
8-h max
Warm season
(March-October)
75th: 67.0
90th: 82.1
Maximum:
147.5
PM2.5: 0.62;
NO2: 0.44;
SO2: 0.21;
CO: 0.27
S042": 0.56;
TC: 0.52;
OC: 0.54;
EC: 0.40
Copollutant
models: CO,
NO2, PM2.5
Lag 0-1:
1.03 (1.02, 1.03)
Darrow et al. (2011)
Atlanta, GA, U.S.
Ozone: 1993-2004
Follow-up:
1993-2004
Time-series study
All ages
One monitor
1-h max, 24-h avg,
8-h max
Warm season
(March-October)
1-h max:
Mean: 62
75th: 76
Maximum:
180
24-h avg:
Mean: 30
75th: 37
Maximum: 81
8-h max:
Mean: 53
75th: 67
Maximum:
148
Correlation (r): Lag 1 RR
1-h max O3:
PM2.5: 0.49;
NO2: 0.33;
CO: 0.21;
24-h avg O3:
PM2.5: 0.25;
NO2: -0.15;
CO: -0.17;
8-h max O3:
PM2.5: 0.46;
NO2: 0.24;
CO: 0.15
Copollutant
models: NR
1-h max:
1.01 (1.01, 1.02)
24-h avg:
1.01 (1.00, 1.01)
8-h max:
1.01 (1.01, 1.02)
tWinauist et al.
(2012)
St. Louis, MO, U.S.
Ozone: 2001-2007
Follow-up:
2001-2007
Time-series study
All ages
One monitor
8-h max
Year-round
Correlation (r): RR
NR
Copollutant
models: NR
0-4 DL: 1.01
(1.00, 1.03)
3-175
-------
Table 3-42 (Continued): Epidemiologic studies of short-term exposure to ozone
and emergency department (ED) visits for aggregate
respiratory diseases.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
(95% Clf
tDarrow et al. (2011)
Atlanta, GA, U.S.
Ozone:
March-October,
1993-2004
Follow-up:
March-October,
1993-2004
Time-series study
n = 1,068,525
All ages
One monitor
8-h max
Year-round
Mean: 53
Median: 51
75th: 67
Maximum:
148
Correlation (r): Lag 1 RRs
PM2.5: 0.46;
NO2: 0.24
Copollutant
models: NR
Commute (7:00
a.m.-10:00 a.m.;
4:00 p.m.-7:00
p.m.): 1.01
(1.00, 1.02) per
25 ppb increase
Daytime (8:00
a.m.-7:00 p.m.):
1.01 (1.01, 1.02)
per 20 ppb
increase
Nighttime (12:00
a.m.-6:00 a.m.):
0.99 ( 0.98,
1.00) per 25 ppb
1-h max: 1.01
(1.01, 1.02)
24-h avg: 1.01
(1.00, 1.01)
8-h max: 1.01
(1.01, 1.02)
tMaliq et al. (2016)
California
(statewide), U.S.
Ozone: 2005-2009
Follow-up:
2005-2008
Case-crossover
study
All ages
Nearest monitor
within 20 km of
population weighted
zip-code centroid
1-h max
Seasonal:
warm season
(May-October) and
year-round
estimates
Mean: 33-55
across
climate zones
Correlation (r):
NO2:
-0.01 YR;
0.26 warm;
SO2:
-0.06 YR;
0.02 warm;
CO: -0.28 YR;
0.02 warm
Copollutant
models: NO2,
CO, SO2
Percent increase
Lag 0-1
Year-round: 2.07
(1.63, 2.52)
Lag 0-3
Warm season:
3.75 (2.55, 4.98)
+NO2
2.13 (0.63, 3.66)
+SO2
3.61 (1.44, 5.82)
+CO
2.39 (0.75, 4.06)
tBarrv et al. (2018) All aaes
Fusion of CMAQ
Mean:
Correlation (r):
RR
Five U.S. cities
model estimates and
37.5-42.2
NR
Lag 0-2
Ozone: 2002-2008
Follow-up:
ground-based
measurements;
population weighted
75th:
50.1-54.4
90th:
59.3-63.5
Copollutant
models: NR
Atlanta:
1.02 (1.01, 1.04)
2002-2008
avg of 12-km grid
Birmingham:
Time-series study
cells for each city
1.02 (1.00, 1.05)
8-h max
Maximum:
Dallas:
Year-round
80.2-106.3
1.04 (1.02, 1.06)
Pittsburgh
1.02 (1.01, 1.04)
St. Louis
1.02 (1.00, 1.03)
3-176
-------
Table 3-42 (Continued): Epidemiologic studies of short-term exposure to ozone
and emergency department (ED) visits for aggregate
respiratory diseases.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
(95% Clf
tO' Lenick et al.
(2017)
Atlanta, GA; Dallas,
TX; and St. Louis,
MO, U.S.
Ozone: 2002-2008
Follow-up:
2002-2008
Case-crossover
study
n = 421,798
Age: 5-18 yr
Fusion of CMAQ
model estimates and
ground-based
measurements;
12-km grid cells
area weighted to
ZCTAs
8-h max
Year-round
Mean:
40.0-42.2
across cities
Maximum:
125
Correlation (r): Lag 0-2 ORs
NR
Copollutant
models: NR
St. Louis:
1.02 (0.99, 1.06)
Dallas:
1.03 (1.01, 1.06)
Atlanta:
1.06 (1.05, 1.09)
CI = confidence interval; CMAQ = Community Multiscale Air Quality; DL = distributed lag; EC = elemental carbon; NR = not
reported; OC = organic carbon; OR = odds ratio; RR = relative risk; S042" = sulfate; TC = total carbon.
tStudies published since the 2013 Ozone ISA.
aResults standardized to a 15-ppb increase in 24-h avg, 20-ppb increase in 8-h daily max, or 25-ppb increase in 1-h daily max
ozone concentrations.
3-177
-------
3.3.2
Long-Term Exposure
Table 3-43 Epidemiologic studies of long-term exposure to ozone and
development of asthma.
Study
Study Population
Exposure Copollutant Effect Estimates
Assessment Mean ppb Examination (95% Cl)a
tGarcia et al.
(2019)
Multicity, southern
California
Ozone: 1993, 1996,
and 2006
Follow-up:
1993-2001;
1996-2004;
2006-2014
Cohort study
Children's Health
Study
n = 4,140
Age: 4th grade at
enrollment to
12th grade at end of
follow up.
8 yr follow-up for
three different
cohorts spanning
21 yr
Ozone measure at
one monitor in each
of nine
communities.
Community-specific
mean annual
concentration
(10:00 a.m. to
6:00 p.m. average)
measured at
baseline in each
community.
Mean: NR;
Ozone
concentrations
depicted
graphically.
Annual average
ozone
concentrations
ranged from
about 26 ppb to
76 ppb. Median
decrease across
communities
from baseline to
end of follow-up
was 8.9 ppb
Correlation
(r):
N02: 0.54;
PM2.5: 0.62
Copollutant
models: NR
Asthma
Incidence (per
10 ppb
decrease)
Fully adjusted
model: 0.83
(0.68, 1.02)
Adjusted for
traffic with local
near roadway
pollution term:
0.84 (0.68, 1.05)
tTetreault et al.
(2016a)
Quebec, Canada
Ozone: 1999-2011
Follow-up:
1999-2011
Cohort study
Quebec Integrated Average summer Mean: 32.07 Correlation
Chronic Disease
Surveillance System
n = 1,183,865
Children born in
Quebec
(June-August)
concentrations of
8-h midday O3
estimated using a
BME-LUR model.
Median: 32.19
75th: 33.76
Maximum:
43.12
(r): NR
Copollutant
models: NR
Asthma onset
HRs
Birth address:
1.20 (1.16, 1.23)
Time-varying
exposure:
1.23 (1.20, 1.27)
3-178
-------
Table 3-43 (Continued): Epidemiologic studies of long term exposure to ozone
and development of asthma.
Study
Study Population
Exposure
Assessment
Mean ppb
Copollutant
Examination
Effect Estimates
(95% Cl)a
tNishimura et al.
(2013)
Multicity, U.S.
Ozone: NR
Follow-up:
Case-control study
Gala II and Sage II
n = 1,968
African American
and Latino
American children
and young adults.
Case subjects had
physician-diagnosed
asthma, while
control subjects,
matched 1:1 by age,
had no history of
asthma or other
respiratory disease.
Age: 8-21 yr
IDW from up to four
monitors within
50 km of residence.
First year of life and
first 3 yr of life
exposures
estimated.
1-h max; 8-h max
Mean: 27.6
Median: 27.3
75th: 30.9
Correlation
(r): NR
Copollutant
models: NR
ORs
First 3 yr of life,
8-h max: 0.90
(0.66, 1.23)
First year of life,
1-h max: 0.94
(0.81, 1.12)
First 3 yr of life,
1-h max: 0.96
(0.71, 1.28)
BME-LUR = Bayesian maximum entropy-land use regression; CI = confidence interval; HR = hazard ratio; IDW = inverse-distance
weighting; Km = kilometer; NR = not reported; OR = odds ratio.
tStudies published since the 2013 Ozone ISA.
aResults standardized to a 10-ppb increase in long-term ozone concentrations.
3-179
-------
Table 3-44 Study-specific details from animal toxicological studies of long-term
ozone exposure and inflammation, oxidative stress, and
injury—allergy.
Species
(Stock/Strain), Exposure Details
Study n, Sex, Age (Concentration, Duration) Endpoints Examined
Chou et al. (2011)
Rhesus
macaque
(Macaca
mulatta)
Sensitized and
challenged
with house
dust mite
n = 6 males
Age: 1 mo
0.5 ppm, 8 h/day for 5 days followed by
9 days of FA—5 cycles
BALF differential cell counts,
eotaxins/airway mucosa
eosinophil number; lung tissue
immunofluorescence of major
basic protein, eotaxin and
eotaxin receptor; lung tissue
mRNA for eotaxin (4-5 days PE,
at 90 days of age)
Zellner et al. (2011) Rats (F344) 2 ppm, 3 h
n = 3-14
Sex: NR
Age: PND 5
Numbers of airway neurons
(PNDs 10-28)
BALF total cell and differential
cell counts and immune cell
phenotypes; mRNA for
T-lymphocyte markers,
cytokines, and CCR3 (3-5 h PE,
at 25 weeks of age)
dust mite
n = 5-6 males
Age: 1 mo
Crowley et al. (2017) Rhesus 0.5 ppm, 8 h/day for 5 days followed by
macaque 9 days of FA—11 cycles
(Macaca
mulatta)
Sensitized and
challenged
with house
Murphy et al. (2012) Rhesus 0.5 ppm, 8 h/day for 5 days followed by Neurokinin pathway components
macaque 9 days of FA—11 cycles (at 12 mo of age)
(Macaca
mulatta)
Sensitized and
challenged
with house
dust mite
n = 4-6 males
Age: 6 mo
BALF = bronchoalveolar lavage fluid; CCR3 = CC motif chemokine receptor 3; FA = filtered air, mRNA = messenger ribonucleic
acid; NR = not reported; PE = post-exposure; PND = post-natal day.
3-180
-------
Table 3-45 Study-specific details from animal toxicological studies of long-term
ozone exposure and lung function—allergy.
Species
(Stock/Strain), Exposure Details
Study n, Sex, Age (Concentration Duration) Endpoints Examined
Pulmonary mechanics, challenge
with histamine; airway smooth
muscle contraction to electrical
field stimulation (PE, at 6 mo of
age)
challenged
with house
dust mite
n = 6-9 males
Age: 1 mo
Moore et al. (2012a) Rhesus 0.5 ppm, 8 h/day for 5 days followed by
macaque 9 days of FA—11 cycles
(Macaca
mulatta)
Sensitized and
FA = filtered air; PE = post-exposure.
3-181
-------
Table 3-46 Study-specific details from animal toxicological studies of long-term
ozone exposure and inflammation, oxidative stress, and
injury—healthy.
Species (Stock/Strain),
Study Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Hunter et al. (2011)
Rats
n = 4-6
Strain and sex: NR
Age: PND 6-28
2 ppm, 3 h
BALF neutrophils and NGF; NGF
mRNA in tracheal epithelial cells,
SP+ airway neurons in vagal
ganglia, airway SP-nerve fiber
density (12-24 h PE)
Gordon et al. (2013)
Rats (BN)
n = 7-10
Age: 4 and 20 mo
0.8 ppm, 6 h/day, once a
week for 16 weeks
BALF total and differential cell
counts, protein, albumin, LDH,
NAG, GGT, histology (1 day PE)
Gabehart et al. (2014) Mice (BALB/c)
n = NR
Sex: NR
Age: 3 days
1 ppm, 3 h
BALF total and differential cell
counts, albumin; lung tissue
mRNA for chemokine and
antioxidant genes, transcriptome
analysis, histology, cell
proliferation (6 and 24 h PE)
Clay et al. (2014)
Rhesus macaque (Macaca
mulatta)
n = 3-5 males
Age: 1 mo
0.5 ppm, 8 h/day for 5 days
followed by 9 days of
FA—11 cycles followed by
FA until 12 mo
IL-6 and IL-8 mRNA and protein
in airway epithelial cells in vitro
and cell culture apical
supernatant following LPS
challenge; micro RNAgene
expression in airway epithelial
cells in vitro following LPS
challenge (12 mo of age)
Gabehart et al. (2015) Mice (BALB/c), wild type
and TLR4 deficient
n = 3-14 females
Age: 1, 2, 3 weeks
1 ppm, 3 h
BALF total cell number and
differential cell counts, albumin,
MUC5AC; lung tissue mRNA for
chemokines, antioxidants, TLR4,
neuropeptides (6, 24, 48 h PE)
Snow et al. (2016)
Rats (BN)
n = 8-10 males
Age: 1,4, 12, 24 mo
0.25 ppm, 6 h/day,
2 days/week for 13 weeks
1 ppm, 6 h/day,
2 days/week for 13 weeks
BALF total cells, cell differentials,
protein, albumin, GGT, NAG
(18 h PE)
Gordon et al. (2016b)
Rats (BN)
n = 9-10 males,
9-10 females
Age: 20 weeks
0.8 ppm, 5 h/day for
1 day/week for 4 weeks
BALF total cells, cell differentials,
albumin (18 h PE)
Gordon et al. (2016a) Rats (S-D)
0.25 ppm, 5 h/day for
BALF total cells, cell differentials,
n = 10 females
1 day/week for 6 weeks
albumin, NAG, GGT (24 h PE)
Age: 20 weeks
0.5 ppm, 5 h/day for
1 day/week for 6 weeks
1 ppm, 5 h/day for
1 day/week for 6 weeks
3-182
-------
Table 3-46 (Continued): Study-specific details from animal toxicological studies
of long-term ozone exposure and inflammation, oxidative
stress, and injury—healthy.
Species (Stock/Strain), n,
Study Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Miller etal. (2016a)
Rats (WKY)
n = 8-10 males
Age: 10 weeks
0.25 ppm, 5 h/day for
3 days/week for 13 weeks
0.5 ppm, 5 h/day for
3 days/week for 13 weeks
1 ppm, 5 h/day for
3 days/week for 13 weeks
BALF total cells, cell differentials,
albumin, NAG, GGT
(immediately PE or following
1-week recovery)
Dye et al. (2017)
Rats (F344)
n = 3-4 males,
3-4 females
Age: PND 14, 21, 28
Rats (S-D)
n = 7-8 males,
7-8 females
Age: PND 14, 21, 28
Rats (WS)
n = 7-8 males,
7-8 females
Age: PND 14, 21, 28
1 ppm, 2 h
Lung tissue antioxidants
(immediately PE)
Miller et al. (2017) Rats (LE) 0.4 ppm, 4 h/day for 2 days; BALF total protein, albumin,
n _ g_-|Q females ®Ds LDH, NAG, GGT, total cell, and
x ¦ ,x n ft nnm a h/Hau fnr 9 Have- differential cell count (GD 21)
Age : NR but weight was u.B ppm, 4 n/aay Tor days, v i
200 g, pregnant GDs 5-6
BALF = bronchoalveolar lavage fluid; BN = brown Norway; F344 = Fischer 344; FA = filtered air; GD = gestational day;
GGT = gamma glutamyl transferase; IL = interleukin; LDH = lactate dehydrogenase; LE = Long-Evans; LPS = lipopolysaccharide;
mRNA = messenger ribonucleic acid; MUC5AC = mucin 5AC glycoprotein; NAG = /V-acetyl-glucosaminidase; NGF = nerve growth
factor; NR = not reported; PE = post-exposure; PND = postnatal day; S-D = Sprague-Dawley; SP = substance P; TLR4 = toll-like
receptor 4; WKY = Wistar Kyoto; WS = Wstar.
3-183
-------
Table 3-47 Study-specific details from animal toxicological studies of long-term
ozone exposure and morphology and other endpoints in healthy
animals.
Study
Species
(Stock/Strain),
n, Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Lee etal. (2011)
Rats (S-D)
n = 7-9 males
Age: 7 days
0.5 ppm, 6 h/day for 3 weekly cycles
either 5 days ozone and 2 days
recovery or 2 days ozone and 5 days
recovery
Airway architectural
parameters—diameter, length,
branching angles of conducting
airways (56 days PE)
Murphy etal. (2013)
Rhesus
macaque
(Macaca
mulatta)
n = 3-4 males
Age: 1 mo
0.5 ppm, 8 h
0.5 ppm, 8 h/day for 5 days followed
by 9 days of FA—1 or 11 cycles
Serotonin pathway components
(PE, at 2 and 6 mo of age)
Murphy etal. (2014)
Rhesus
macaque
(Macaca
mulatta)
n = 3-4 males
Age: 1 mo
0.5 ppm, 8 h/day for 5 days followed
by 9 days of FA—1 or 11 cycles
Lung tissue mRNA and
immunostaining for NK-1R,
TAC1/SP, Nur77 (PE, up to
25 weeks)
FA = filtered air; mRNA = messenger ribonucleic acid; NK-1 R = neurokinin-1 receptor; Nur77 = nuclear receptor 77;
PE = post-exposure; S-D = Sprague-Dawley; SP = substance P; TAC1 = tachykinin precursor 1.
3-184
-------
Table 3-48 Epidemiologic studies of long-term exposure to ozone and lung
function and development.
Effect
Study Exposure
Copollutant
Estimates
Study Population Assessment Mean (ppb)
Examination
(95% Clf
tEckel etal. (2012)
Multicity, U.S.
Ozone: 1990-1997
Follow-up: 1990-1997
Cohort study
Cardiovascular
Health Study
n = 3,382
Age: >65 yr
Cumulative sum of
monthly averages
calculated using
IDW from up to
three monitors
within 50 km of
residences.
Cumulative
exposures
estimated for time
on study.
8-h avg
Mean: 39.7
Median: 39.7
75th: 51.3
95th: 64.1
Maximum: 79.6
Correlation (r):
PM10: 0.96
Copollutant
models: NR
Difference in
FEVi (mL)
Women: -0.17
(-0.24, -0.10)
Men: -0.34
(-0.47, -0.21)
Difference in
FVC (mL)
Women: -0.76
(-0.86, -0.66)
Men: -1.24
(-1.40, -1.09)
tUrman et al. (2014)
Multicity, southern
California, U.S.
Ozone: 2002-2007
Follow-up: 2007-2008
Cross-sectional study
Children's
Health Study
n = 1,811
Age: 11-12 yr
6-yr avg of
10:00 a.m. to
6:00 p.m. Ozone
measured at one
monitor in each of
the eight
communities
Other
Mean: 22.7
Correlation (r):
PM2.5: 0.66;
NO2: 0.12
Copollutant
models: NR
FVC (percent
increase):
-0.14 (-1.38,
1.12)
FEV1 (percent
increase):
-1.38 (-2.34,
-0.40)
tGauderman et al.
(2015)
Multicity, southern
California, U.S.
Ozone: 1994-1997;
1997-2000; 2007-2010
Follow-up: 1994-1997;
1997-2000; 2007-2010
Cohort study
Children's
Health Study
n = 2,120
Age: 11-15 yr
4 yr follow-up
for three
different
cohorts
spanning 19 yr
4-yr avg of
10:00 a.m. to
6:00 p.m. Ozone
measured at one
monitor in each of
the five
communities
Other
Mean: Range
across
communities:
28.6 to 61.9
Correlation (r):
PM2.5: 0.39;
NO2: 0.02
Copollutant
models: NR
4-yr FEV1
growth (mL)
per decrease
in Os: -12.18
(-92.73,
68.18)
4-yr FVC
growth (mL)
per decrease
in Os: -13.27
(-144.18,
117.45)
3-185
-------
Table 3-48 (Continued): Epidemiologic studies of long-term exposure to ozone
and lung function and development.
Effect
Study Exposure Copollutant Estimates
Study Population Assessment Mean (ppb) Examination (95% Cl)a
FEVi (percent
increase)
First year of
life exposure:
-1.12 (-2.60,
0.40)
Average
lifetime
exposure:
-1.30 (-3.88,
1.36)
CI = confidence interval; FENA = forced expiratory volume in 1 second; FVC = forced vital capacity; IDW = inverse-distance
weighting; Km = kilometer; NR = not reported.
tStudies published since the 2013 Ozone ISA.
aResults standardized to a 10-ppb increase in long-term ozone concentrations.
tNeophvtou et al.
(2016)
Multicity, U.S.
Ozone: NR
Follow-up:
Cross-sectional study
Gala II and
Sage II
n = 1,968
Age: 8-21 yr
African
American and
Latino
American
children and
young adults
with asthma
IDW from up to
four monitors
within 50 km of
residence. First
year of life and
lifetime exposures
estimated.
8-h max
Mean: NR
Median: Results
presented
graphically.
Median average
lifetime
concentrations
range from
approximately
20 to 37 across
study sites
Correlation (r):
PM2.5:
First year: 0.19;
Lifetime: 0.73;
NO2:
First year: 0.02;
Lifetime: 0.49;
SO2:
First year: 0.15;
Lifetime: 0.04
Copollutant
models: NR
3-186
-------
Table 3-49 Study-specific details from animal toxicological studies of long-term
ozone exposure and morphology—allergy.
Species
(Stock/Strain), Exposure Details
Study n, Sex, Age (Concentration, Duration) Endpoints Examined
Alveolar volume and number,
distribution of alveolar size, and
capillary surface density per
alveolar septa; mRNA of
candidate genes (PE, at 3 or
6 mo of age)
mite
n = 12 males
Age: 30 days
Avdalovic et al. (2012) Rhesus 0.5 ppm, 8 h/day for 5 days followed by
macaque 9 days of FA—5 or 11 cycles
(Macaca
mulatta),
sensitized and
challenged with
house dust
Alveolar number and size,
alveolar capillary surface density,
length and volume of terminal
and respiratory bronchioles (PE
at 6 and 36 mo)
mite
n = 6 males
Age: 30 days
Herring et al. (2015) Rhesus 0.5 ppm, 8 h/day for 5 days followed by
macaque 9 days of FA—11 cycles followed by
(Macaca 30 mo recovery in FA
mulatta),
sensitized and
challenged with
house dust
FA = filtered air; mRNA = messenger ribonucleic acid; PE = post-exposure.
3-187
-------
Table 3-50 Study-specific details from animal toxicological studies of long-term
ozone exposure and lung function in healthy animals.
Species
(Stock/Strain), n, Exposure Details
Study Sex, Age (Concentration, Duration) Endpoints Examined
Gordon et al. (2013)
Rats (BN)
0.8 ppm, 6 h/day, once a week, for
Ventilatory parameters (1 and
n = 7-8 males
16 weeks
7 days PE each week)
Age: 4 and 20 mo
Snow et al. (2016)
Rats (BN)
0.25 ppm, 6 h/day, 2 days/week for
Ventilatory parameters
n = 6-10 males
13 weeks
(1-5 days PE each week)
Age: 1,4, 12, 24 mo
1 ppm, 6 h/day, 2 days/week for
13 weeks
Gordon et al. (2016a)
Rats (S-D)
0.25 ppm, 5 h/day for 1 day/week for
Ventilatory parameters (24 h
n = 10 females
6 weeks
post 5th week of exposure)
Age: 20 weeks
0.5 ppm, 5 h/day for 1 day/week for
6 weeks
1 ppm, 5 h/day for 1 day/week for
6 weeks
Miller et al. (2017)
Rats (LE)
0.4 ppm, 4 h/day for 2 days; GDs 5-6
Ventilatory parameters
n = 9-10 females
0.8 ppm, 4 h/day for 2 days; GDs 5-6
(immediately PE)
Age: NR but weight
was 200 g, pregnant
BN = brown Norway; GD = gestational day; LE = Long-Evans; NR = not reported; PE = post-exposure, S-D = Sprague-Dawley.
3-188
-------
Table 3-51 Epidemiologic studies of long-term exposure to ozone and
development of chronic obstructive pulmonary disease (COPD).
Study
Copollutant
Effect Estimates
Study
Population
Exposure Assessment
Mean ppb
Examination
(95% Cla)
tTo etal. (2016)
Ontario Asthma
Interpolated surface
Mean:
Correlation (r):
COPD in adults with
Ontario, Canada
Surveillance
using IDW of
39.3
PM25: NR
asthma:
Ozone: 1996-2013
Information
System and the
49 monitors across
the province. Average
Median:
39.2
Copollutant
models: PM2.5
HR 2.05 (1.17, 3.60)
Follow-up:
Canadian
of monthly 24-h max
75th: 40.4
1996-2014
Community
from time of asthma
Cohort study
Health Survey
incidence to time of
n = 6,040
COPD incidence or
Age: >18 yr
end of follow-up.
Other
Adults with
asthma
CI = confidence interval; HR = hazard ratio; IDW = inverse-distance weighting; NR = not reported.
tStudies published since the 2013 Ozone ISA.
aResults standardized to a 10-ppb increase in long-term ozone concentrations.
3-189
-------
Table 3-52 Epidemiologic studies of long-term exposure to ozone and
respiratory infection.
Study
Study
Population
Exposure Assessment
Mean
ppb
Copollutant
Examination
Effect Estimates
HR (95% CI)a
tMaclntvre et al.
(2011)
Georgia Basin Airshed
(including Vancouver
and Victoria, British
Columbia), Canada
Ozone: 1999-2002
Follow-up: 1999-2002
Cohort study
n = 45,513
All singleton
live births in the
Georgia Basin
Airshed,
followed for the
first 2 yr of life
IDW average of three
closest monitors within
50 km
Mean:
28.2
Median:
26.1
Maximum:
71.8
Correlation (r):
NR
Copollutant
models: NR
Otitis media
HR 0.96 (0.95, 0.97)
tSmith et al. (2016)
Multicity, northern
California, U.S.
Ozone: 1996-2010
Follow-up: 1996-2010
Case-control study
n = 6,913
Cases are adult
members of
Kaiser
Permanente
Northern
California with a
clinical
diagnosis of TB
and a
corresponding
anti-TB
prescription or
a positive TB
culture.
Controls were
matched 2-1
on age, sex,
and
race/ethnicity.
Age: >21 yr
2-yr avg from the
nearest monitor
8-h avg
Median:
31.5
Maximum:
67
Correlation (r):
PM25: 0.25;
NO2: -0.33;
SO2: -0.24;
CO: -0.28
Copollutant
models: NR
Pulmonary
tuberculosis ORs
1st quintile: Ref
2nd quintile: 0.92
(0.78, 1.10)
3rd quintile: 0.95
(0.80, 1.14)
4th quintile: 0.71
(0.59, 0.85)
5th quintile: 0.66
(0.55, 0.79)
CI = confidence interval; HR :
ratio.
hazard ratio; IDW = inverse-distance weighting; Km = kilometer; NR = not reported; OR = odds
tStudies published since the 2013 Ozone ISA.
aResults standardized to a 10-ppb increase in long-term ozone concentrations.
3-190
-------
Table 3-53 Epidemiologic studies of long-term exposure to ozone and severity
of respiratory disease.
Study
Study
Population
Exposure
Assessment
Mean(ppb)
Copollutant
Examination
Effect
Estimates
(95% Cla)
tTetreault et al.
(2016b)
Quebec, Canada
Ozone: 1990-2006
Follow-up: 1996-2011
Cohort study
Quebec
Integrated
Chronic
Disease
Surveillance
System
n = 1,183,865
Children born
in Quebec
Average summer Mean: 30.57
(June-August)
concentrations of
8-h midday O3
estimated using a
BME-LUR model.
Other
Median: 30.8
75th: 32.42
Maximum: 38.92
Correlation (r): Hospital/ED
NR
Copollutant
models: NR
visits HRs
Birth residence:
0.99 (0.96, 1.11)
Time-
Dependent: 1.17
(1.12, 1.22)
tTo et al. (2016)
Ontario, Canada
Ozone: 1996-2013
Follow-up: 1996-2014
Cohort study
Ontario
Asthma
Surveillance
Information
System and
the Canadian
Community
Health Survey
n = 6,040
Age: >18 yr
Adults with
asthma
Interpolated surface
using IDW of
49 monitors across
the province.
Average of monthly
24-h max from time
of asthma incidence
to time of COPD
incidence or end of
follow-up.
Other
Mean: 39.3
Median: 39.2
75th: 40.4
Correlation (r):
PM2.5: NR
Copollutant
models: PM2.5
COPD in adults
with asthma:
2.05 (1.17, 3.60)
tBerhane et al. (2016) Children's
Multicity, southern
California, U.S.
Os: 1992-2011
Follow-up: 1992-
1995-2003;
2002-2011
Cohort study
¦2000;
Health Study
n = 4,602
Age: 10 and
15-yr-olds
9-or 10-yr avg of Mean: Range Correlation (r): Absolute
10:00 a.m. to
6:00 p.m. Ozone
measured at one
monitor in each of
the eight
communities
Other
across cohorts:
44.8-47.7
PM2.5: 0.54;
NO2: 0.38
Copollutant
models: NO2,
PM2.5
(percent)
changes in
bronchitis
symptoms
15-yr-olds with
asthma: -29.22
(-40.80, -12.77)
10-yr-olds with
asthma: -39.00
(-53.34, -17.52)
BME-LUR = Bayesian maximum entropy-land use regression; CI = confidence interval; HR :
weighting; Km = kilometer; NR = not reported.
tStudies published since the 2013 Ozone ISA.
aResults standardized to a 10-ppb increase in long-term ozone concentrations.
hazard ratio; IDW = inverse-distance
3-191
-------
Table 3-54 Epidemiologic studies of long-term exposure to ozone and allergic
sensitization.
Study
Study
Population
Exposure
Assessment
Mean
ppb
Copollutant
Examination
Effect Estimates
(95% Cla)
tWeiretal. (2013)
Multicity, U.S.
Ozone: 2005-2006
Follow-up: 2005-2006
Cross-sectional study
NHANES
n = 6,227
(CMAQ); 5,201
(IDW monitors)
Age: >6 yr
Annual average
CMAQ estimates
and estimates
derived from
inverse-distance
weighting for
participants living
within 20 miles of a
monitor
8-h max
Mean:
51.5 (IDW);
57.2 (CMAQ)
Median:
52 (IDW);
57.0 (CMAQ)
75th:
55.3 (IDW);
61.2 (CMAQ)
95th:
60.3 (IDW);
70.8 (CMAQ)
Correlation (r):
PM2.5: 0.08
(IDW); -0.21
(CMAQ);
NO2: -0.25
(IDW); -0.42
(CMAQ)
Copollutant
models: NR
ORs
IDW
Food allergens: 0.80
(0.54, 1.19)
Indoor allergens:
0.91 (0.78, 1.06)
Outdoor allergens:
1.17 (0.99, 1.38)
Inhalant allergens:
1.06 (0.93, 1.20)
Any allergens: 1.07
(0.94, 1.21)
CMAQ
Food allergens: 1.01
(0.77, 1.32)
Indoor allergens:
1.02 (0.86, 1.22)
Outdoor allergens:
1.14 (0.90, 1.43)
Inhalant allergens:
1.11 (0.93, 1.32)
Any allergens: 1.10
(0.93, 1.29)
CI = confidence interval; CMAQ = Community Multiscale Air Quality; IDW = inverse-distance weighting; NR = not reported;
OR = odds ratio.
tStudies published since the 2013 Ozone ISA.
aResults standardized to a 10-ppb increase in long-term ozone concentrations.
3-192
-------
Table 3-55 Study-specific details from animal toxicological studies of long-term
ozone exposure and allergic sensitization in healthy animals.
Study
Species
(Stock/Strain),
n, Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Hansen et al. (2013)
Mice
0.1 ppm, 0.33 h/day for 5 days/week
Development of allergy (IgE),
(BALB/cJ),
for 2 weeks and once weekly for
BALF total and differential cell
minimally
12 weeks
counts, ventilatory parameters
sensitized by
low dose
ovalbumin
n = 8-10
females
Age:
6-7 weeks
BALF = bronchoalveolar lavage fluid; IgE = immunoglobulin E.
3-193
-------
Annex for Appendix 3: Evaluation of Studies on Health Effects of
Ozone
This annex describes the approach used in the Integrated Science Assessment (ISA) for Ozone
and Related Photochemical Oxidants to evaluate study quality in the available health effects literature. As
described in the Preamble to the ISA (U.S. EPA. 2015). causality determinations were informed by the
integration of evidence across scientific disciplines (e.g., exposure, animal toxicology, epidemiology) and
related outcomes and by judgments of the strength of inference in individual studies. Table Annex 3-1
describes aspects considered in evaluating study quality of controlled human exposure, animal
toxicological, and epidemiologic studies. The aspects found in Table Annex 3-1 are consistent with
current best practices for reporting or evaluating health science data.1 Additionally, the aspects are
compatible with published U.S. EPA guidelines related to cancer, neurotoxicity, reproductive toxicity,
and developmental toxicity (U.S. EPA. 2005. 1998. 1996b. 1991V
These aspects were not used as a checklist, and judgments were made without considering the
results of a study. The presence or absence of particular features in a study did not necessarily lead to the
conclusion that a study was less informative or should be excluded from consideration in the ISA.
Further, these aspects were not used as criteria for determining causality in the five-level hierarchy. As
described in the Preamble, causality determinations were based on judgments of the overall strengths and
limitations of the collective body of available studies and the coherence of evidence across scientific
disciplines and related outcomes. Table Annex 3-1 is not intended to be a complete list of aspects that
define a study's ability to inform the relationship between ozone and health effects, but it describes the
major aspects considered in this ISA to evaluate studies. Where possible, study elements, such as
exposure assessment and confounding (i.e., bias due to a relationship with the outcome and correlation
with exposures to ozone), are considered specifically for ozone. Thus, judgments on the ability of a study
to inform the relationship between an air pollutant and health can vary depending on the specific pollutant
being assessed.
1 For example, NTP OHAT approach (Roonev et al.. 20141. IRIS Preamble (U.S. EPA. 2013b). ToxRTool
(Klimisch et al.. 19971. STROBE guidelines (von Elm et al.. 20071. and ARRIVE guidelines (Kilkenny et al.. 20101.
3-194
-------
Table Annex 3-1 Scientific considerations for evaluating the strength of inference
from studies on the health effects of ozone.
Study Design
Controlled Human Exposure:
Studies should describe clearly the primary and any secondary objectives of the study, or specific hypotheses being
tested. Study subjects should be randomly assigned to treatment groups and exposed, to the extent possible,
without knowledge of the exposure condition. Preference is given to balanced crossover (repeated measures) or
parallel design studies which include control exposures (e.g., to clean filtered air). In crossover studies, a sufficient
and specified time between exposure days should be provided to avoid carry over effects from prior exposure days.
In parallel design studies, all arms should be matched for individual characteristics, such as age, sex, race,
anthropometric properties, and health status. In studies evaluating effects of disease, appropriately matched healthy
controls are desired for interpretative purposes.
Animal Toxicology:
Studies should describe clearly the primary and any secondary objectives of the study, or specific hypotheses being
tested. Studies should include appropriately matched control exposures (e.g., to clean filtered air, time matched).
Studies should use methods to limit differences in baseline characteristics of control and exposure groups. Studies
should randomize assignment to exposure groups and where possible conceal allocation to research personnel.
Groups should be subjected to identical experimental procedures and conditions to the extent possible; animal care
including housing, husbandry, etc. should be identical between groups. Blinding of research personnel to study
group may not be possible due to animal welfare and experimental considerations; however, differences in the
monitoring or handling of animals in all groups by research personnel should be minimized.
Epidemiology:
Inference is stronger for studies that describe clearly the primary and any secondary aims of the study, or specific
hypotheses being tested.
For short-term exposure, time-series, case-crossover, and panel studies are emphasized over cross-sectional
studies because they examine temporal correlations and are less prone to confounding by factors that differ
between individuals (e.g., SES, age). Panel studies with scripted exposures, in particular, can contribute to
inference because they have consistent, well-defined exposure durations across subjects, measure personal
ambient pollutant exposures, and measure outcomes at consistent, well-defined lags after exposures. Studies with
large sample sizes and conducted over multiple years are considered to produce more reliable results. Additionally,
multicity studies are preferred over single-city studies because they examine associations for large diverse
geographic areas using a consistent statistical methodology, avoiding the publication bias often associated with
single-city studies.3 If other quality parameters are equal, multicity studies carry more weight than single-city studies
because they tend to have larger sample sizes and lower potential for publication bias.
For long-term exposure, inference is considered to be stronger for prospective cohort studies and case-control
studies nested within a cohort (e.g., for rare diseases) than cross-sectional, other case-control, or ecologic studies.
Cohort studies can better inform the temporality of exposure and effect. Other designs can have uncertainty related
to the appropriateness of the control group or validity of inference about individuals from group-level data. Study
design limitations can bias health effect associations in either direction.
3-195
-------
Table Annex 3-1 (Continued): Scientific considerations for evaluating the strength
of inference from studies on the health effects of
ozone.
Study Population/Test Model
Controlled Human Exposure:
In general, the subjects recruited into study groups should be similarly matched for age, sex, race, anthropometric
properties, and health status. In studies evaluating effects of specific subject characteristics (e.g., disease, genetic
polymorphism, etc.), appropriately matched healthy controls are preferred. Relevant characteristics and health
status should be reported for each experimental group. Criteria for including and excluding subjects should be
indicated clearly. For the examination of populations with an underlying health condition (e.g., asthma),
independent, clinical assessment of the health condition is ideal, but self-report of physician diagnosis generally is
considered to be reliable for respiratory and cardiovascular disease outcomes.15 The loss or withdrawal of recruited
subjects during the course of a study should be reported. Specific rationale for excluding subject(s) from any portion
of a protocol should be explained.
Animal Toxicology:
Ideally, studies should report species, strain, substrain, genetic background, age, sex, and weight. Unless data
indicate otherwise, all animal species, stocks and strains are considered appropriate for evaluating effects of ozone
exposure. It is preferred that the authors test for effects in both sexes and multiple lifestages, and report the result
for each group separately. All animals used in a study should be accounted for, and rationale for exclusion of
animals or data should be specified.
Epidemiology:
There is greater confidence in results for study populations that are recruited from and representative of the target
population. Studies with high participation and low dropout over time that is not dependent on exposure or health
status are considered to have low potential for selection bias. Clearly specified criteria for including and excluding
subjects can aid assessment of selection bias. For populations with an underlying health condition, independent,
clinical assessment of the health condition is valuable, but self-report of physician diagnosis generally is considered
to be reliable for respiratory and cardiovascular diseases.15 Comparisons of groups with and without an underlying
health condition are more informative if groups are from the same source population. Selection bias can influence
results in either direction or may not affect the validity of results but rather reduce the generalizability of findings to
the target population.
Pollutant
Controlled Human Exposure:
The focus is on studies testing ozone exposure.
Animal Toxicology:
The focus is on studies testing ozone exposure.
Epidemiology:
The focus is on studies evaluating ozone exposure.
3-196
-------
Table Annex 3-1 (Continued): Scientific considerations for evaluating the strength
of inference from studies on the health effects of
ozone.
Exposure Assessment or Assignment
Controlled Human Exposure:
For this assessment, the focus is on studies that use ozone concentrations <0.4 ppm. Studies that use higher
exposure concentrations may provide information relevant to biological plausibility, dosimetry, or inter-species
variation. Studies should have well-characterized pollutant concentration, temperature, and relative humidity and/or
have measures in place to adequately control the exposure conditions. Preference is given to balanced crossover or
parallel design studies that include control exposures (e.g., to clean filtered air). Study subjects should be randomly
exposed without knowledge of the exposure condition. Method of exposure (e.g., chamber, facemask, etc.) should
be specified and activity level of subjects during exposures should be well characterized.
Animal Toxicology:
For this assessment, the focus is on studies that use ozone concentrations <2 ppm. Studies that use higher
exposure concentrations may provide information relevant to biological plausibility, dosimetry, or inter-species
variation. Studies should characterize pollutant concentration, temperature, and relative humidity and/or have
measures in place to adequately control the exposure conditions. The focus is on inhalation exposure.
Noninhalation exposure experiments (i.e., intratracheal instillation [IT]) are informative and may provide information
relevant to biological plausibility and dosimetry. In vitro studies may be included if they provide mechanistic insight
or examine similar effects as in vivo studies but are generally not included. All studies should include exposure
control groups (e.g., clean filtered air).
Epidemiology:
Of primary relevance are relationships of health effects with the ambient component of ozone exposure. However,
information about ambient exposure rarely is available for individual subjects; most often, inference is based on
ambient concentrations. Studies that compare exposure assessment methods are considered to be particularly
informative. Inference is stronger when the duration or lag of the exposure metric corresponds with the time course
for physiological changes in the outcome (e.g., up to a few days for symptoms) or latency of disease (e.g., several
years for cancer).
Ambient ozone concentration tends to have low spatial heterogeneity at the urban scale, except near roads where
ozone concentration is lower because ozone reacts with nitric oxide emitted from vehicles. For studies involving
individuals with near-road or on-road exposures to ozone, in which ambient ozone concentrations are more spatially
heterogeneous and relationships between personal exposures and ambient concentrations are potentially more
variable, validated methods that capture the extent of variability for the epidemiologic study design (temporal vs.
spatial contrasts) and location carry greater weight.
Fixed-site measurements, whether averaged across multiple monitors or assigned from the nearest or single
available monitor, typically have smaller biases and smaller reductions in precision compared with spatially
heterogeneous air pollutants. Concentrations reported from fixed-site measurements can be informative if correlated
with personal exposures, closely located to study subjects, highly correlated across monitors within a location, or
combined with time-activity information.
Atmospheric models may be used for exposure assessment in place of or to supplement ozone measurements in
epidemiologic analyses. For example, grid-scale models (e.g., CMAQ) that represent ozone exposure over relatively
large spatial scales (e.g., typically greater than 4- * 4-km grid size) often do provide adequate spatial resolution to
capture acute ozone peaks that influence short-term health outcomes. Uncertainty in exposure predictions from
these models is largely influenced by model formulations and the quality of model input data pertaining to precursor
emissions or meteorology, which tends to vary on a study-by-study basis.
In studies of short-term exposure, temporal variability of the exposure metric is of primary interest. For long-term
exposures, models that capture within-community spatial variation in individual exposure may be given more weight
for spatially variable ambient ozone. Given the low spatial variability of ozone at the urban scale, exposure
measurement error typically causes health effect estimates to be underestimated for studies of either short-term or
long-term exposure. Biases and decreases in the precision of the association (i.e., wider 95% CIs) tend to be small.
Even when spatial variability is higher near roads, the reduction in ozone exposure would cause the exposure to be
overestimated at a monitor distant from the road or when averaged across a model grid cell, so that health effects
would likely be underestimated.
3-197
-------
Table Annex 3-1 (Continued): Scientific considerations for evaluating the strength
of inference from studies on the health effects of
ozone.
Outcome Assessment/Evaluation
Controlled Human Exposure:
Endpoints should be assessed in the same manner for control and ozone exposure groups (e.g., time after
exposure, methods, endpoint evaluator) using valid, reliable methods. Blinding of endpoint evaluators is ideal,
especially for qualitative endpoints (e.g., histopathology). For each experiment and each experimental group,
including controls, precise details of all procedures carried out should be provided including how, when, and where.
Time of the endpoint evaluations is a key consideration that will vary depending on endpoint evaluated. Endpoints
should be assessed at time points that are appropriate for the research questions.
Animal Toxicology:
Endpoints should be assessed in the same manner for control and exposure groups (e.g., time after exposure,
methods, endpoint evaluator) using valid, reliable methods. Blinding of endpoint evaluators is ideal, especially for
qualitative endpoints (e.g., histopathology). For each experiment and each experimental group, including controls,
precise details of all procedures carried out should be provided including how, when, and where. Time of the
endpoint evaluations is a key consideration that will vary depending on endpoint evaluated. Endpoints should be
assessed at time points that are appropriate for the research questions.
Epidemiology:
Inference is stronger when outcomes are assessed or reported without knowledge of exposure status. Knowledge
of exposure status could produce artifactual associations. Confidence is greater when outcomes assessed by
interview, self-report, clinical examination, or analysis of biological indicators are defined by consistent criteria and
collected by validated, reliable methods. Independent, clinical assessment is valuable for outcomes like lung
function or incidence of disease, but report of physician diagnosis has shown good reliability.15 When examining
short-term exposures, evaluation of the evidence focuses on specific lags based on the evidence presented in
individual studies. Specifically, the following hierarchy is used in the process of selecting results from individual
studies to assess in the context of results across all studies for a specific health effect or outcome:
i. Distributed lag models;
ii. Average of multiple days (e.g., 0-2);
iii. If a priori lag days were used by the study authors these are the effect estimates presented; or
iv. If a study focuses on only a series of individual lag days, expert judgment is applied to select the
appropriate result to focus on considering the time course for physiologic changes for the health effect or
outcome being evaluated.
When health effects of long-term exposure are assessed by acute events such as symptoms or hospital
admissions, inference is strengthened when results are adjusted for short-term exposure. Validated questionnaires
for subjective outcomes such as symptoms are regarded to be reliable,0 particularly when collected frequently and
not subject to long recall. For biological samples, the stability of the compound of interest and the sensitivity and
precision of the analytical method is considered. If not based on knowledge of exposure status, errors in outcome
assessment tend to bias results toward the null.
Potential Copollutant Confounding
Controlled Human Exposure:
Exposure should be well characterized to evaluate independent effects of ozone.
Animal Toxicology:
Exposure should be well characterized to evaluate independent effects of ozone.
3-198
-------
Table Annex 3-1 (Continued): Scientific considerations for evaluating the strength
of inference from studies on the health effects of
ozone.
Epidemiology:
Not accounting for potential copollutant confounding can produce artifactual associations; thus, studies that
examine copollutant confounding carry greater weight. The predominant method is copollutant modeling
(i.e., two-pollutant models), which is especially informative when correlations are not high. However, when
correlations are high (r> 0.7), such as those often encountered for UFP and other traffic-related copollutants,
copollutant modeling is less informative. Although the use of single-pollutant models to examine the association
between ozone and a health effect or outcome are informative, ideally studies should also include copollutant
analyses. Copollutant confounding is evaluated on an individual study basis considering the extent of correlations
observed between the copollutant and ozone, and relationships observed with ozone and health effects in
copollutant models.
Other Potential Confounding Factorsd
Controlled Human Exposure:
Preference is given to studies using experimental and control groups that are matched for individual level
characteristics (e.g., race/ethnicity, sex, body weight, smoking history, age) and time-varying factors (e.g., seasonal
and diurnal patterns).
Animal Toxicology:
Preference is given to studies using experimental and control groups that are matched for individual level
characteristics (e.g., strain, sex, body weight, litter size, feed and water consumption) and time-varying factors
(e.g., seasonal and diurnal patterns).
Epidemiology:
Factors are considered to be potential confounders if demonstrated in the scientific literature to be related to health
effects and correlated with ozone. Not accounting for confounders can produce artifactual associations; thus,
studies that statistically adjust for multiple factors or control for them in the study design are emphasized. Less
weight is placed on studies that adjust for factors that mediate the relationship between ozone and health effects,
which can bias results toward the null. Confounders vary according to study design, exposure duration, and health
effect and may include, but are not limited to the following:
Short-term exposure studies: Meteorology, day of week, season, medication use, allergen exposure, and long-term
temporal trends.
Long-term exposure studies: Socioeconomic status, race, age, medication use, smoking status, stress, noise, and
occupational exposures.
Statistical Methodology
Controlled Human Exposure:
Statistical methods should be described clearly and appropriate for the study design and research question
(e.g., correction for multiple comparisons). Generally, statistical significance is used to evaluate the findings of
controlled human exposure studies. However, consistent trends are also informative. Detection of statistical
significance is influenced by a variety of factors including, but not limited to, the size of the study, exposure and
outcome measurement error, and statistical model specifications. Sample size is not a criterion for exclusion;
ideally, the sample size should provide adequate power to detect hypothesized effects (e.g., sample sizes less than
three are considered less informative). Because statistical tests have limitations, consideration is given to both
trends in data and reproducibility of results.
3-199
-------
Table Annex 3-1 (Continued): Scientific considerations for evaluating the strength
of inference from studies on the health effects of
ozone.
Animal Toxicology:
Statistical methods should be described clearly and appropriate for the study design and research question
(e.g., correction for multiple comparisons). Generally, statistical significance is used to evaluate the findings of
animal toxicology studies. However, consistent trends are also informative. Detection of statistical significance is
influenced by a variety of factors including, but not limited to, the size of the study, exposure and outcome
measurement error, and statistical model specifications. Sample size is not a criterion for exclusion; ideally, the
sample size should provide adequate power to detect hypothesized effects (e.g., sample sizes less than three are
considered less informative). Because statistical tests have limitations, consideration is given to both trends in data
and reproducibility of results.
Epidemiology:
Multivariable regression models that include potential confounding factors are emphasized. However, multipollutant
models (more than two pollutants) are considered to produce too much uncertainty due to copollutant collinearity to
be informative. Models with interaction terms aid in the evaluation of potential confounding as well as effect
modification. Sensitivity analyses with alternate specifications for potential confounding inform the stability of
findings and aid in judgments of the strength of inference from results. In the case of multiple comparisons,
consistency in the pattern of association can increase confidence that associations were not found by chance alone.
Statistical methods that are appropriate for the power of the study carry greater weight. For example, categorical
analyses with small sample sizes can be prone to bias results toward or away from the null. Statistical tests such as
f-tests and chi-squared tests are not considered sensitive enough for adequate inferences regarding ozone-health
effect associations. For all methods, the effect estimate and precision of the estimate (i.e., width of 95% CI) are
important considerations rather than statistical significance.
aU.S. EPA (2008V
"Muraia etal. (2014V Weakley et al. (2013V Yang et al. (2011V Heckbert et al. (2004V Barr et al. (2002V Muhaiarine et al. (1997V
Toren et al. (1993V
cBurnev et al. (1989V
dMany factors evaluated as potential confounders can be effect measure modifiers (e.g., season, comorbid health condition) or
mediators of health effects related to ozone (comorbid health condition).
3-200
-------
3.4 References
Adams. WC. (2000). Ozone dose-response effects of varied equivalent minute ventilation rates. J
Expo Anal Environ Epidemiol 10: 217-226. http://dx.doi.org/10.1038/si.iea.7500Q86
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/10.1080/0895837029008461Q
Adams. WC. (2003a). 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.
http://dx.doi.org/10.1080/0895837039Q168283
Adams. WC. (2003b). Relation of pulmonary responses induced by 66-h exposures to 0.08 ppm ozone
and 2-h exposures to 0.30 ppm ozone via chamber and face-mask inhalation. Inhal Toxicol 15:
745-759. http://dx.doi.org/10.1080/0895837039Q217828
Adams. WC. (2006). Comparison of chamber 6.6-h exposures to 0.04-0.08 ppm ozone via square-
wave and triangular profiles on pulmonary responses. Inhal Toxicol 18: 127-136.
http://dx.doi.org/10.1080/089583705003061Q7
Adams. WC; Ollison. WM. (1997). Effects of prolonged simulated ambient ozone dosing patterns on
human pulmonary function and symptomatology. Pittsburgh, PA: Air & Waste Management
Association.
Alberti. KG; Eckel. RH: Grundy. SM: Zimmet. PZ: Cleeman. JI; Donato. KA; Fruchart. JC; James.
WP; Loria. CM; Smith. SC. (2009). Harmonizing the metabolic syndrome: A joint interim
statement of the International Diabetes Federation Task Force on Epidemiology and Prevention;
National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation;
International Atherosclerosis Society; and International Association for the Study of Obesity.
Circulation 120: 1640-1645. http://dx.doi.org/10.1161/CIRCULATIQNAHA.109.192644
Alexeeff. SE; Litonjua. AA; Suh. H; Sparrow. D; Yokonas. 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; Lav. JC; Zhou. H; Kim. CS; Hernandez. ML; Kehrl. H; Hazucha. MJ; Devlin. RB; Diaz-
Sanchez. D; Peden. DB. (2013). The glutathione-S-transferase mu 1 (GSTM1) null genotype and
increased neutrophil response to low-level ozone (0.06 ppm) [Letter], J Allergy Clin Immunol 131:
610-612. http://dx.doi.Org/10.1016/i.iaci.2012.07.005
Alexis. NE; Zhou. H; Lav. JC; Harris. B; Hernandez. ML; Lu. TS; Bromberg. PA; Diaz-Sanchez. D;
Devlin. RB; Kleeberger. SR; Peden. DB. (2009). The glutathione-S-transferase Mu 1 null genotype
modulates ozone-induced airway inflammation in human subjects. J Allergy Clin Immunol 124:
1222-1228. http://dx.doi.Org/10.1016/i.iaci.2009.07.036
Alhanti. BA; Chang. HH; Winquist. A; Mulholland. JA; Darrow. LA; Sarnat. SE. (2016). Ambient air
pollution and emergency department visits for asthma: A multi-city assessment of effect
modification by age. J Expo Sci Environ Epidemiol 26: 180-188.
http://dx.doi.org/10.1038/ies.2015.57
3-201
-------
Arbex. MA; de Souza Conceicao. GM; Cendon. SP; Arbex. FF; Lopes. AC; Movses. EP; Santiago.
SL; Saldiva. PHN; Pereira. LAA; Brag a. ALF. (2009). Urban air pollution and chronic obstructive
pulmonary disease-related emergency department visits. J Epidemiol Community Health 63: 777-
783. http://dx.doi.org/10.1136/iech.2008.07836Q
Aris. RM; Tager. I; Christian. D; Kelly. T; Balmes. JR. (1995). Methacholine responsiveness is not
associated with 03-induced decreases in FEV1. Chest 107: 621-628.
Ariomandi. M; Balmes. JR; Frampton. MW; Bromberg. P; Rich. DO; Stark. P; Alexis. NE; Costantini.
M; Hollenbeck-Pringle. D; Dagincourt. N; Hazucha. MJ. (2018). Respiratory responses to ozone
exposure: the multicenter ozone study in older subjects (moses). Am J Respir Crit Care Med 197:
13191327. http://dx.doi.org/10.1164/rccm.201708-1613QC
Ariomandi. M; Wong. H; Donde. A; Frelinger. J; Dalton. S; Ching. W; Power. K; Balmes. JR. (2015).
Exposure to medium and high ambient levels of ozone causes adverse systemic inflammatory and
cardiac autonomic effects. Am J Physiol Heart Circ Physiol 308: H1499-1509.
http://dx.doi.org/10.1152/aipheart.00849.2014
Avdalovic. MY; Tyler. NK; Putney. L; Nishio. SJ; Quesenberrv. S; Singh. PJ; Miller. LA; Schelegle.
ES; Plopper. CG; Vu. T; Hyde. DM. (2012). Ozone exposure during the early postnatal period
alters the timing and pattern of alveolar growth and development in nonhuman primates. Anat Rec
295: 1707-1716. http://dx.doi.org/10.1002/ar.22545
Avol. EL; Trim. SC; Little. DE; Spier. CE; Smith. MN; Peng. RC: Linn. WS; Hackney. JD; Gross.
KB; DArcv. JB; Gibbons. D; Higgins. ITT. (1990). Ozone exposure and lung function in children
attending a southern 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. PO; Kelly. T; Aris. RM. (1996).
Ozone-induced decrements in FEV1 and FVC do not correlate with measures of inflammation. Am
J Respir Crit Care Med 153: 904-909. http://dx.doi.Org/10.l 164/airccm. 153.3.8630571
Bao. A; Liang. L; Li. F; Zhang. M; Zhou. X. (2013). Effects of acute ozone exposure on lung peak
allergic inflammation of mice. Front Biosci 18: 838-851. http://dx.doi.org/10.2741/4147
Barker. JS; Wu. Z; Hunter. DP; Dev. RD. (2015). Ozone exposure initiates a sequential signaling
cascade in airways involving interleukin-lbeta release, nerve growth factor secretion, and
substance P upregulation. J Toxicol Environ Health A 78: 397-407.
http://dx.doi.org/10.1080/15287394.2014.971924
Barr. RG; Herbstman. J; Speizer. FE; Camargo. CA. Jr. (2002). Validation of self-reported chronic
obstructive pulmonary disease in a cohort study of nurses. Am J Epidemiol 155: 965-971.
http://dx.doi.org/10.1093/aie/155.10.965
Barraza-Villarreal. A; Sunver. 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
Barreno. RX; Richards. JB; Schneider. DJ; Cromar. KR; Nadas. AJ; Hernandez. CB; Hallberg. LM;
Price. RE; Hashmi. SS; Blackburn. MR; Hague. IU; Johnston. RA. (2013). Endogenous
osteopontin promotes ozone-induced neutrophil recruitment to the lungs and airway
hyperresponsiveness to methacholine. Am J Physiol Lung Cell Mol Physiol 305: LI 18-L129.
http://dx.doi.org/10.1152/aiplung.00080.2Q13
3-202
-------
Barry. V; Klein. M; Winquist. A; Chang. HH: Mulholland. JA; Talbott. EO; Rager. JR; Tolbert. PE;
Sarnat. SE. (2018). Characterization of the concentration-response curve for ambient ozone and
acute respiratory morbidity in 5 US cities. J Expo Sci Environ Epidemiol.
http://dx.doi.org/10.1038/s41370-018-0Q48-7
Bartoli. ML; Vagaggini. B; Malagrino. L; Bacci. E; Cianchetti. S; Dente. FL; Novelli. F; Costa. F;
Paggiaro. P. (2013). Baseline airway inflammation may be a determinant of the response to ozone
exposure in asthmatic patients. Inhal Toxicol 25: 127-133.
http://dx.doi.org/10.3109/08958378.2013.763313
Bastain. T; Islam. T; Berhane. K; McConnell. R; Rappaport. E; Salam. M: Linn. W: Avol. E; Zhang.
Y; Gilliland. F. (2011). Exhaled nitric oxide, susceptibility and new-onset asthma in the Children's
Health Study. Eur Respir J 37: 523-531. http://dx.doi.org/10.1183/09031936.0002121Q
Bates. ML; Brenza. TM; Ben-Jebria. A; Bascom. R; Eldridge. MW: Ultman. JS. (2014). Pulmonary
function responses to ozone in smokers with a limited smoking history. Toxicol Appl Pharmacol
278: 85-90. http://dx.doi.org/10.1016/i.taap.2014.04.011
Bennett. WD; Ivins. S; Alexis. NE; Wu. J; Bromberg. PA; Brar. SS; Travlos. G; London. SJ. (2016).
Effect of obesity on acute ozone-induced changes in airway function, reactivity, and inflammation
in adult females. PLoS ONE 11: e0160030. http://dx.doi.org/10.1371/iournal.pone.016003Q
Bentaveb. M; Wagner. V; Stempfelet. M; Zins. M; Goldberg. M; Pascal. M; Larrieu. S; Beaudeau. P;
Cassadou. S; Eilstein. D; Filleul. L; Le Tertre. A; Medina. S; Pascal. L; Prouvost. H; Ouenel. P;
Zeghnoun. A; Lefranc. A. (2015). Association between long-term exposure to air pollution and
mortality in France: A 25-year follow-up study. Environ Int 85: 5-14.
http://dx.doi.Org/10.1016/i.envint.2015.08.006
Berhane. K; Chang. CC; McConnell. R; Gauderman. WJ; Avol. E; Rapapport. E; Urman. R: Lurmann.
F; Gilliland. F. (2016). Association of changes in air quality with bronchitic symptoms in children
in California, 1993-2012. JAMA 315: 1491-1501. http://dx.doi.org/10.1001/iama.2016.3444
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/10.1183/09031936.0008141Q
Berhane. K; Zhang. Y; Salam. MT; Eckel. SP; Linn. WS; Rappaport. EB; Bastain. TM; Lurmann. F;
Gilliland. FD. (2014). Longitudinal effects of air pollution on exhaled nitric oxide: the Children's
Health Study. Occup Environ Med 71: 507-513. http://dx.doi.Org/10.l 136/oemed-2013-101874
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.
Bhoopalan. V; Han. SG; Shah. MM; Thomas. DM; Bhalla. DK. (2013). Tobacco smoke modulates
ozone-induced toxicity in rat lungs and central nervous system. Inhal Toxicol 25: 21-28.
http://dx.doi.org/10.3109/08958378.2012.751143
Biller. H; Holz. O; Windt. H; Koch. W; Miiller. M; Jorres. RA; Krug. N; Hohlfeld. JM. (2011). Breath
profiles by electronic nose correlate with systemic markers but not ozone response. Respir Med
105: 1352-1363. http://dx.doi.Org/10.1016/i.rmed.2011.03.002
Bosson. JA; Blomberg. A; Stenfors. N; Helledav. R; Kelly. FJ; Behndig. AF; Mudwav. I. (2013).
Peripheral blood neutrophilia as a biomarker of ozone-induced pulmonary inflammation. PLoS
ONE 8: e81816. http://dx.doi.org/10.1371/iournal.pone.0Q81816
3-203
-------
Brand. JD; Ballingcr. CA; Tugglc. KL; Fanucchi. MY; Schwiebert. LM; Postlethwait. EM. (2012).
Site-specific dynamics of CD1 lb(+) and CD103(+) dendritic cell accumulations following ozone
exposure. Am J Physiol Lung Cell Mol Physiol 303: L1079-L1086.
http://dx.doi.org/10.1152/aiplung.00185.2Q12
Brand. JD; Mathews. JA; Kasahara. DI; Wurmbrand. AP; Shore. SA. (2016). Regulation of IL-17A
expression in mice following subacute ozone exposure. J Immunotoxicol 13: 428-438.
http://dx.doi.org/10.3109/1547691X.2015.1120829
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.201006-0849QC
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
Burnett. R; Raizenne. M; Krewski. D. (1990). Acute health effects of transported air pollution: A
study of children attending a residential summer camp. Can J Stat 18: 367-373.
http://dx.doi.org/10.2307/3315843
Burnev. PG; Laitinen. LA: Perdrizet. S; Huckauf. H: Tattersfield. AE: Chinn. S; Poisson. N; Heeren.
A: Britton. JR; Jones. T. (1989). Validity and repeatability of the IUATLD (1984) Bronchial
Symptoms Questionnaire: an international comparison. Eur Respir J 2: 940-945.
Bvers. N: Ritchev. M: Vaidvanathan. A: Brandt. AJ: Yip. F. (2015). Short-term effects of ambient air
pollutants on asthma-related emergency department visits in Indianapolis, Indiana, 2007-2011. J
Asthma 53: 1-8. http://dx.doi.org/10.3109/02770903.2015.10910Q6
Cabello. N: Mishra. V: Sinha. U: Diangelo. SL: Chroneos. ZC: Ekpa. NA: Cooper. TK: Caruso. CR:
Silvevra. P. (2015). Sex differences in the expression of lung inflammatory mediators in response
to ozone. Am J Physiol Lung Cell Mol Physiol 309: ajplung.00018.02015.
http://dx.doi.org/10.1152/aiplung.00018.2Q15
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/AEQH.61.l.5-10
Carey. IM: Atkinson. RW: Kent. A J: van Staa. T: Cook. DG: Anderson. HR. (2013). Mortality
associations with long-term exposure to outdoor air pollution in a national English cohort. Am J
Respir Crit Care Med 187: 1226-1233. http://dx.doi.Org/10.l 164/rccm.201210-1758QC
Che. L: Jin. Y: Zhang. C: Lai. T: Zhou. H: Xia. L: Tian. B: Zhao. Y: Liu. J: Wu. Y: Wu. Y: Du. J: Li.
W: Ying. S: Chen. Z: Shen. H. (2016). Ozone-induced IL-17A and neutrophilic airway
inflammation is orchestrated by the caspase-l-IL-1 cascade. Sci Rep 6: 18680.
http://dx.doi.org/10.1038/srepl8680
Cheng. W: Duncan. KE; Ghio. AJ: Ward-Caviness. C: Karolv. ED: Diaz-Sanchez. D: Conollv. RB:
Devlin. RB. (2018). Changes in metabolites present in lung lining fluid following exposure of
humans to ozone. Toxicol Sci 163: 430439. http://dx.doi.org/10.1093/toxsci/kfV043
Cho. H: Gladwell. W: Yamamoto. M: Kleeberger. SR. (2013). Exacerbated airway toxicity of
environmental oxidant ozone in mice deficient in Nrf2. Oxid Med Cell Longev 2013: Article
#254069. http://dx.doi.org/10.1155/2013/254069
Cho. Y: Abu-Ali. G: Tashiro. H: Kasahara. DI: Brown. TA: Brand. JD: Mathews. JA: Huttenhower.
C: Shore. SA. (2018). The microbiome regulates pulmonary responses to ozone in mice. Am J
Respir Cell Mol Biol 59: 346-354. http://dx.doi.Org/10.l 165/rcmb.2017-0404QC
3-204
-------
Chou. PL; Gerriets. JE; Schclcglc. ES; Hyde. DM; Miller. LA. (2011). Increased CCL24/eotaxin-2
with postnatal ozone exposure in allergen-sensitized infant monkeys is not associated with
recruitment of eosinophils to airway mucosa. Toxicol Appl Pharmacol 257: 309-318.
http://dx.doi.Org/10.1016/i.taap.2011.09.001
Ciencewicki. JM; Verhein. KC; Gerrish. K; Mccaw. ZR; Li. J; Bushel. PR; Kleeberger. SR. (2016).
Effects of mannose-binding lectin on pulmonary gene expression and innate immune inflammatory
response to ozone. Am J Physiol Lung Cell Mol Physiol 311: L280-L291.
http://dx.doi.org/10.1152/aiplung.00205.2Q15
Clay. CC; Maniar-Hew. K; Gerriets. JE; Wang. TT; Postlethwait. EM; Evans. MJ; Fontaine. JH;
Miller. LA. (2014). Early life ozone exposure results in dysregulated innate immune function and
altered microRNA expression in airway epithelium. PLoS ONE 9: e90401.
http://dx.doi.org/10.1371/iournal.pone.009Q401
Clay. E; Patacchini. R; Trevisani. M; Preti. D; Brana. M; Spina. D; Page. C. (2016). Ozone-induced
hypertussive responses in rabbits and guinea pigs. J Pharmacol Exp Ther 357: 73-83.
http://dx.doi.Org/10.l 124/jpet.l 15.230227
Connor. AJ; Laskin. JD; Laskin. PL. (2012). Ozone-induced lung injury and sterile inflammation.
Role of toll-like receptor 4. Exp Mol Pathol 92: 229-235.
http://dx.doi.Org/10.1016/i.vexmp.2012.01.004
Crouse. PL; Peters. PA; Hvstad. P; Brook. JR; van Ponkelaar. A; Martin. RV; Villeneuve. PJ; Jerrett.
M; Goldberg. MS; Pope. CA; Brauer. M; Brook. RP; Robichaud. A; Menard. R; Burnett. RT.
(2015). Ambient PM 2.5, O 3, and NO 2 exposures and associations with mortality over 16 years of
follow-up in the Canadian Census Health and Environment Cohort (CanCHEC). Environ Health
Perspect 123: 1180-1186. http://dx.doi.org/10.1289/ehp.1409276
Crowley. CM; Fontaine. JH; Gerriets. JE; Schelegle. ES; Hyde. PM; Miller. LA. (2017). Early life
allergen and air pollutant exposures alter longitudinal blood immune profiles in infant rhesus
monkeys. Toxicol Appl Pharmacol 328: 60-69. http://dx.doi.Org/10.1016/i.taap.2017.05.006
Pales. R; Kauri. LM; Cakmak. S; Mahmud. M; Weichenthal. SA; Van Rvswvk. K; Kumarathasan. P;
Thomson. E; Vincent. R; Broad. G; Liu. L. (2013). Acute changes in lung function associated with
proximity to a steel plant: A randomized study. Environ Int 55: 15-19.
http ://dx.doi .org/10.1016/i .envint.2013.01.014
Parrow. LA; Klein. M; Flanders. WP; Mulholland. JA; Tolbert. PE; Strickland. MJ. (2014). Air
pollution and acute respiratory infections among children 0-4 years of age: an 18-year time-series
study. Am J Epidemiol 180: 968-977. http://dx.doi.org/10.1093/aie/kwu234
Parrow. 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/ies.2009.49
Pelfino. RJ: Staimer. N; Tjoa. T; Gillen. PL; Schauer. JJ; Shafer. MM. (2013). Airway inflammation
and oxidative potential of air pollutant particles in a pediatric asthma panel. J Expo Sci Environ
Epidemiol 23: 466-473. http://dx.doi.org/10.1038/ies.2013.25
Pevlin. RB; Puncan. KE; Jardim. M; Schmitt. MT; Rappold. AG; Piaz-Sanchez. P. (2012).
Controlled exposure of healthy young volunteers to ozone causes cardiovascular effects.
Circulation 126: 104-111. http://dx.doi.Org/10.l 161/CIRCULATIONAHA.l 12.094359
Pokic. P; Traikovska-Pokic. E. (2013). Ozone exaggerates nasal allergic inflammation. Prilozi 34:
131-141.
3-205
-------
Durrani. F; Phelps. DS; Weisz. J; Silvevra. P; Hu. SM; Mikerov. AN; Floros. J. (2012). Gonadal
hormones and oxidative stress interaction differentially affects survival of male and female mice
after lung Klebsiella Pneumoniae infection. Exp Lung Res 38: 165-172.
http://dx.doi.org/10.3109/01902148.2011.654Q45
Dye. JA; Gibbs-Flournov. EA; Richards. JH; Norwood. J; Kraft. K; Hatch. GE. (2017). Neonatal rat
age, sex and strain modify acute antioxidant response to ozone. Inhal Toxicol 29: 291-303.
http://dx.doi.org/10.1080/08958378.2017.13696Q2
Dye. JA; Ledbetter. AD; Schladweiler. MC; Costa. PL; Kodavanti. UP. (2015). Whole body
plethysmography reveals differential ventilatory responses to ozone in rat models of cardiovascular
disease. Inhal Toxicol 27: 14-25. http://dx.doi.org/10.3109/08958378.2014.954167
Eckel. SP; Louis. TA; Chaves. PH; Fried. LP; Margolis. AH. (2012). Modification of the association
between ambient air pollution and lung function by frailty status among older adults in the
cardiovascular health study. Am J Epidemiol 176: 214-223. http://dx.doi.org/10.1093/aie/kws001
Elkhidir. HS; Richards. JB; Cromar. KR; Bell. CS; Price. RE; Atkins. CL; Spencer. CY; Malik. F;
Alexander. AL; Cockerill. KJ; Hague. IU; Johnston. RA. (2016). Plasminogen activator inhibitor-1
does not contribute to the pulmonary pathology induced by acute exposure to ozone. 4.
http ://dx.doi .org/10.14814/phv2.12983
Farrai. AK; Hazari. MS; Winsett. DW; Kulukulualani. A; Carll. AP; Havkal-Coates. N; Lamb. CM;
Lappi. E; Terrell. D; Cascio. WE; Costa. PL. (2012). Overt and latent cardiac effects of ozone
inhalation in rats: evidence for autonomic modulation and increased myocardial vulnerability.
Environ Health Perspect 120: 348-354. http://dx.doi.org/10.1289/ehp.1104244
Farrai. AK; Malik. F; Havkal-Coates. N; Walsh. L; Winsett. D; Terrell. D; Thompson. LC; Cascio.
WE; Hazari. MS. (2016). Morning N02 exposure sensitizes hypertensive rats to the cardiovascular
effects of same day 03 exposure in the afternoon. Inhal Toxicol 28: 170-179.
http://dx.doi.org/10.3109/08958378.2016.1148Q88
Feng. F; Jin. Y; Duan. L; Yan. Z; Wang. S; Li. F; Liu. Y; Samet. JM; Wu. W. (2015). Regulation of
ozone-induced lung inflammation by the epidermal growth factor receptor in mice. Environ
Toxicol 31: 2016-2027. http://dx.doi.org/10.1002/tox.22202
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. http://dx.doi.Org/10.l 164/airccm. 149.1.8111607
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. http://dx.doi.org/10.1080/0894063Q.1988.10466349
Forbes. LJL; Kapetanakis. V; Rudnicka. AR; Cook. DG; Bush. T; Stedman. JR; Whincup. PH;
Strachan. DP; Anderson. HR. (2009). Chronic exposure to outdoor air pollution and lung function
in adults. Thorax 64: 657-663. http://dx.doi.Org/10.l 136/thx.2008.109389
Foster. WM; Brown. RH; Macri. K; Mitchell. CS. (2000). Bronchial reactivity of healthy subjects: 18-
20 h postexposure to ozone. J Appl Physiol (1985) 89: 1804-1810.
http://dx.doi.Org/10.l 152/iappl.2000.89.5.1804
Foster. WM; Weinmann. GG; Menkes. E; Macri. K. (1997). Acute exposure of humans to ozone
impairs small airway function. Ann Occup Hyg 1: 659-666.
http://dx.doi.org/10.1093/annhyg/41.inhaled particles VIII.659
3-206
-------
Frampton. MW; Balmes. JR; Bromberg. PA; Stark. P; Arjomandi. M; Hazucha. MJ; Rich. DQ;
Hollcnbcck-Pringlc. D; Dagincourt. N; Alexis. N; Ganz. P; Zareba. W; Costantini. MG. (2017).
Multicenter Ozone Study in oldEr Subjects (MOSES: Part 1. Effects of exposure to low
concentrations of ozone on respiratory and cardiovascular outcomes) [HEI], (Research Report 192,
Pt 1). Boston, MA: Health Effects Institute.
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. http://dx.doi.Org/10.l 164/airccm. 155.1.9001299
Frampton. MW; Pietropaoli. A; Dentler. M; Chalupa. D; Little. EL; Stewart. J; Frasier. L; Oakes. D;
Wiltshire. J; Vora. R; Utell. MJ. (2015). Cardiovascular effects of ozone in healthy subjects with
and without deletion of glutathione-S-transferase Ml. Inhal Toxicol 27: 113-119.
http://dx.doi.org/10.3109/08958378.2014.996272
Francis. M; Groves. AM; Sun. R; Cervelli. JA; Choi. H; Laskin. JD; Laskin. PL. (2017a). Editor's
highlight: CCR2 regulates inflammatory cell accumulation in the lung and tissue injury following
ozone exposure. Toxicol Sci 155: 474-484. http://dx.doi.org/10.1093/toxsci/kfw226
Francis. M; Sun. R; Cervelli. JA; Choi. H; Mandal. M; Abramova. EV; Gow. AJ; Laskin. JD; Laskin.
PL. (2017b). Editor's highlight: role of spleen-derived macrophages in ozone-induced lung
inflammation and injury. Toxicol Sci 155: 182-195. http://dx.doi.org/10.1093/toxsci/kfw 192
Freiier. JI; Van Eiikeren. JCH; Van Bree. L. (2002). A model for the effect on health of repeated
exposure to ozone. Environ Modell Softw 17: 553-562.
Fry. RC; Rager. JE; Bauer. R; Sebastian. E; Peden. DB; Jaspers. I; Alexis. NE. (2014). Air toxics and
epigenetic effects: ozone altered microRNAs in the sputum of human subjects. Am J Physiol Lung
Cell Mol Physiol 306: LI 129-L1137. http://dx.doi.org/10.1152/aiplung.00348.2013
Fry. RC; Rager. JE; Zhou. H; Zou. B; Brickev. JW; Ting. J; Lav. JC; Peden. DB; Alexis. NE. (2012).
Individuals with increased inflammatory response to ozone demonstrate muted signaling of
immune cell trafficking pathways. Respir Res 13: 89. http://dx.doi.Org/10.l 186/1465-9921-13-89
Gabehart. K; Correll. KA; Loader. JE; White. CW; Dakhama. A. (2015). The lung response to ozone
is determined by age and is partially dependent on toll-Like receptor 4. Respir Res 16: 117.
http://dx.doi.org/10.1186/sl2931-015-0279-2
Gabehart. K; Correll. KA; Yang. J; Collins. ML; Loader. JE; Leach. S; White. CW; Dakhama. A.
(2014). Transcriptome profiling of the newborn mouse lung response to acute ozone exposure.
Toxicol Sci 138: 175-190. http://dx.doi.org/10.1093/toxsci/kft276
Garcia. E; Berhane. KT; Islam. T; Mcconnell. R; Urman. R; Chen. Z; Gilliland. FD. (2019).
Association of changes in air quality with incident asthma in children in California, 1993-2014.
JAMA 321: 1906-1915. http://dx.doi.org/10.1001/iama.2Q19.5357
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.
http://dx.doi.org/10.1056/NEJMoa04061Q
Gauderman. WJ; Urman. R; Avol. E; Berhane. K; McConnell. R; Rappaport. E; Chang. R; Lurmann.
F; Gilliland. F. (2015). Association of improved air quality with lung development in children. N
Engl J Med 372: 905-913. http://dx.doi.org/10.1056/NEJMoa 1414123
Ghio. AJ; Soukup. JM; Dailev. LA; Richards. JH; Duncan. KE; Lehmann. J. (2014). Iron decreases
biological effects of ozone exposure. Inhal Toxicol 26: 391-399.
http://dx.doi.org/10.3109/08958378.2014.90833Q
3-207
-------
Gielen. MH; Van Per 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. http://dx.doi.Org/10.1164/airccm.155.6.9196122
Gilliland. F; Avol. E; McConnell. R; K. B; Gauderman. WJ; Lurmann. FW: Urnam. R; Change. R;
Rappaport. EB; Howland. S. (2017). The effects of policy-driven air quality improvements on
childrens respiratory health [HEI], (Research Report 190). Boston, MA: Health Effects Institute.
https://www.healtheffects.org/svstem/files/GillilandRR190.pdf
Gleason. JA; Bielorv. L; Fagliano. JA. (2014). Associations between ozone, PM2.5, and four pollen
types on emergency department pediatric asthma events during the warm season in New Jersey: a
case-crossover study. Environ Res 132: 421-429. http://dx.doi.Org/10.1016/i.envres.2014.03.035
Gomes. EC; Allgrove. JE; Florida-James. G; Stone. V. (201 la). Effect of vitamin supplementation on
lung injury and running performance in a hot, humid, and ozone-polluted environment. Scand J
Med Sci Sports 21: e452-e460. http://dx.doi.Org/10.llll/i.1600-0838.2011.01366.x
Gomes. EC: Stone. V: Florida-James. G. (201 lb). Impact of heat and pollution on oxidative stress and
CC16 secretion after 8km run. Eur J Appl Physiol 111: 2089-2097.
http://dx.doi.org/10.1007/sQ0421-011-1839-x
Gonzalez-Guevara. E: Carlos Martinez-Lazcano. J: Custodio. V: Hernandez-Ceron. M: Rubio. C: Paz.
C. (2014). Exposure to ozone induces a systemic inflammatory response: possible source of the
neurological alterations induced by this gas. Inhal Toxicol 26: 485-491.
http://dx.doi.org/10.3109/08958378.2014.922648
Goodman. JE: Loftus. CT; Liu. X: Zu. K. (2017a). Impact of respiratory infections, outdoor pollen,
and socioeconomic status on associations between air pollutants and pediatric asthma hospital
admissions. PLoS ONE 12: eO 180522. http://dx.doi.org/10.1371/iournal.pone.018Q522
Goodman. JE: Zu. K: Loftus. CT: Tao. G: Liu. X: Lange. S. (2017b). Ambient ozone and asthma
hospital admissions in Texas: a time-series analysis. 3: 6. http://dx.doi.org/10.1186/s40733-017-
0034-1
Gordon. CJ: Jarema. KA; Lehmann. J. R.; Ledbetter. AD: Schladweiler. MC: Schmid. JE: Ward. WO:
Kodavanti. UP: Nvska. A: Macphail. RC. (2013). Susceptibility of adult and senescent Brown
Norway rats to repeated ozone exposure: an assessment of behavior, serum biochemistry and
cardiopulmonary function. Inhal Toxicol 25: 141-159.
http://dx.doi.org/10.3109/08958378.2013.764946
Gordon. CJ: Phillips. PM; Beaslev. TE; Ledbetter. A: Avdin. C: Snow. SJ: Kodavanti. UP: Johnstone.
AF. (2016a). Pulmonary sensitivity to ozone exposure in sedentary versus chronically trained,
female rats. Inhal Toxicol 28: 293-302. http://dx.doi.org/10.3109/08958378.2016.1163441
Gordon. CJ: Phillips. PM: Johnstone. AFM: Beaslev. TE: Ledbetter. AD: Schladweiler. MC: Snow.
SJ; Kodavanti. UP. (2016b). Effect of high-fructose and high-fat diets on pulmonary sensitivity,
motor activity, and body composition of brown Norway rats exposed to ozone. Inhal Toxicol 28:
203-215. http://dx.doi.org/10.3109/08958378.2015.113473Q
Gordon. CJ; Phillips. PM; Johnstone. AFM; Schmid. J; Schladweiler. MC; Ledbetter. A; Snow. SJ;
Kodavanti. UP. (2017a). Effects of maternal high-fat diet and sedentary lifestyle on susceptibility
of adult offspring to ozone exposure in rats. Inhal Toxicol 29: 239-254.
http://dx.doi.org/10.1080/08958378.2Q17.1342719
3-208
-------
Gordon. CJ; Phillips. PM; Ledbetter. A; Snow. SJ; Schladweiler. MC; Johnstone. AF; Kodavanti. UP.
(2017b). Active vs. sedentary lifestyle from weaning to adulthood and susceptibility to ozone in
rats. Am J Physiol Lung Cell Mol Physiol 312: L100-L109.
http://dx.doi.org/10.1152/aiplung.00415.2Q16
Greer. JR; Abbey. DE; Burchette. RJ. (1993). Asthma related to occupational and ambient air
pollutants in nonsmokers. J Occup Environ Med 35: 909-915. http://dx.doi.org/10.1097/00Q43764-
199309000-00014
Groves. AM; Gow. AJ; Massa. CB; Hall. L; Laskin. JD; Laskin. PL. (2013). Age-related increases in
ozone-induced injury and altered pulmonary mechanics in mice with progressive lung
inflammation. Am J Physiol Lung Cell Mol Physiol 305: L555-L568.
http://dx.doi.org/10.1152/aiplung.00027.2Q13
Groves. AM; Gow. AJ; Massa. CB; Laskin. JD; Laskin. PL. (2012). Prolonged injury and altered lung
function after ozone inhalation in mice with chronic lung inflammation. Am J Respir Cell Mol Biol
47: 776-783. http://dx.doi.org/10.1165/rcmb.2011-0433QC
Hansen. JS; Nielsen. GD; Sorli. JB; Clausen. P; Wolkoff. P; Larsen. ST. (2013). Adjuvant and
inflammatory effects in mice after subchronic inhalation of allergen and ozone-initiated limonene
reaction products. J Toxicol Environ Health A 76: 1085-1095.
http://dx.doi.org/10.1080/15287394.2013.838915
Hansen. JS; Norgaard. AW; Koponen. IK; Sorli. JB; Paidi. MP; Hansen. SW; Clausen. PA; Nielsen.
GP; Wolkoff. P; Larsen. ST. (2016). Limonene and its ozone-initiated reaction products attenuate
allergic lung inflammation in mice. J Immunotoxicol 13: 793-803.
http://dx.doi.org/10.1080/1547691X.2016.1195462
Harkema. JR; Hotchkiss. LA; Vetter. NA; Jackson-Humbles. PN; Lewandowski. RP; Wagner. JG.
(2017). Strain differences in a murine model of air pollutant-induced nonatopic asthma and rhinitis.
Toxicol Pathol 45: 161-171. http://dx.doi.org/10.1177/0192623316674274
Hatch. GE; Crissman. K; Schmid. J; Richards. JE; Ward. WO; Schladweiler. MC; Ledbetter. AP;
Kodavanti. UP. (2015). Strain differences in antioxidants in rat models of cardiovascular disease
exposed to ozone. Inhal Toxicol 27: 54-62. http://dx.doi.org/10.3109/08958378.2014.95417Q
Hatch. GE; Mckee. J; Brown. J; Mcdonnell. W; Seal. E; Soukup. J; Slade. R; Crissman. K; Pevlin. R.
(2013). Biomarkers of dose and effect of inhaled ozone in resting versus exercising human
subjects: Comparison with resting rats. Biomarker Insights 8: 53-67.
http://dx.doi.org/10.4137/BMI.Sl 1102
Hatch. GE; Slade. R; Harris. LP; Mcdonnell. WF; Pevlin. RB; Koren. HS; Costa. PL; 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.
http://dx.doi.Org/10.l 164/airccm. 150.3.8087337
Hazucha. MJ; Folinsbee. LJ; Bromberg. PA. (2003). Pistribution and reproducibility of spirometric
response to ozone by gender and age. J Appl Physiol (1985) 95: 1917-1925.
http ://dx.doi .org/10.1152/iapplphysiol. 00490.2003
Hebbern. C; Cakmak. S. (2015). Synoptic weather types and aeroallergens modify the effect of air
pollution on hospitalisations for asthma hospitalisations in Canadian cities. Environ Pollut 204: 9-
16. http://dx.doi.Org/10.1016/i.envpol.2015.04.010
3-209
-------
Heckbert. SR; Kooperberg. C; Safford. MM; Psatv. BM; Hsia. J; McTiernan. A; Gaziano. JM;
Frishman. WH: Curb. JD. (2004). Comparison of self-report, hospital discharge codes, and
adjudication of cardiovascular events in the Women's Health Initiative. Am J Epidemiol 160: 1152-
1158. http://dx.doi.org/10.1093/aie/kwh314
HEI (Health Effects Institute). (1997). Effects of ozone on normal and potentially sensitive human
subjects.
Henriquez. AR: Snow. SJ; Schladweiler. MC; Miller. CN; Dve. JA: Ledbetter. AD: Richards. JE:
Mauge-Lewis. K; Mcgee. MA: Kodavanti. UP. (2017). Adrenergic and glucocorticoid receptor
antagonists reduce ozone-induced lung injury and inflammation. Toxicol Appl Pharmacol 339:
161-171. http://dx.doi.org/10.1016/i.taap.2017.12.006
Hernandez. M; Brickev. WJ: Alexis. NE; Fry. RC: Rager. JE: Zhou. B; Ting. JP; Zhou. H; Peden. DB.
(2012). Airway cells from atopic asthmatic patients exposed to ozone display an enhanced innate
immune gene profile [Letter], J Allergy Clin Immunol 129: 259-261.e251-252.
http://dx.doi.org/10.1016/i .iaci.2011.11.007
Hernandez. ML: Lav. JC: Harris. B; Esther. CR; Brickev. 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.Org/10.1016/i.iaci.2010.06.043
Herring. MJ; Putney. LF; St George. JA; Avdalovic. MY; Schelegle. ES; Miller. LA; Hyde. DM.
(2015). Early life exposure to allergen and ozone results in altered development in adolescent
rhesus macaque lungs. Toxicol Appl Pharmacol 283: 35-41.
http://dx.doi.Org/10.1016/i.taap.2014.12.006
Higgins. ITT; D'Arcv. JB; Gibbons. DI; Avol. EL; Gross. KB. (1990). Effect of exposures to ambient
ozone on ventilatory lung function in children. Am J Respir Crit Care Med 141: 1136-1146.
http://dx.doi.Org/10.1164/airccm/141.5 Pt 1.1136
Hoffmever. F; Sucker. K; Monse. C; Berresheim. H; Jettkant. B; Rosenkranz. N: Briining. T; Biinger.
L (2015). Different patterns in changes of exhaled breath condensate pH and exhaled nitric oxide
after ozone exposure. In M Pokorski (Ed.), Environment Exposure to Pollutants (pp. 39-47).
Switzerland: Springer International Publishing, http://dx.doi.org/10.1007/5584 2014 63
Hoffmever. F; Sucker. K; Monse. C; Berresheim. H; Rosenkranz. N; Jettkant. B; Beine. A; Briining.
T; Biinger. J. (2013). Relationship of pulmonary function response to ozone exposure and capsaicin
cough sensitivity. Inhal Toxicol 25: 569-576. http://dx.doi.org/10.3109/08958378.2013.812699
Holland. N; Dave. V; Venkat. S; Wong. H; Donde. A; Balmes. JR; Ariomandi. M. (2014). Ozone
inhalation leads to a dose-dependent increase of cytogenetic damage in human lymphocytes.
Environ Mol Mutagen 56: 378-387. http://dx.doi.org/10.1002/em.21921
Holz. O; Biller. H; Mueller. M; Kane. K; Rosano. M; Hanrahan. J; Hava. PL; Hohlfeld. JM. (2015).
Efficacy and safety of inhaled calcium lactate PUR118 in the ozone challenge model - a clinical
trial. 16: 21. http://dx.doi.org/10.1186/s40360-015-0Q21-l
Holze. C; Michaudel. C; Mackowiak. C; Haas. DA; Benda. C; Hubel. P; Pennemann. FL; Schnepf. D;
Wettmarshausen. J; Braun. M; Leung. DW; Amarasinghe. GK; Perocchi. F; Staeheli. P; Rvffel. B;
Pichlmair. A. (2018). Oxeiptosis, a ROS-induced caspase-independent apoptosis-like cell-death
pathway. Nat Immunol 19: 130-140. http://dx.doi.org/10.1038/s41590-017-0013-v
Horstman. DH; Ball. BA; Brown. J; Gerritv. 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. http://dx.doi.org/10.1177/074823379501100401
3-210
-------
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.
http://dx.doi.Org/10.l 164/airccm/142.5.1158
Hsieh. N; Cheng. Y. iH; Liao. C. (2014). Changing variance and skewness as leading indicators for
detecting ozone exposure-associated lung function decrement. Stoch Environ Res Risk Assess 28:
2205-2216. http://dx.doi.org/10.1007/s00477-014-Q887-2
Hulo. S; Tiesset. H; Lancel. S; Edme. JL; Viollet. B; Sobaszek. A; Neviere. R. (2011). AMP-activated
protein kinase deficiency reduces ozone-induced lung injury and oxidative stress in mice. Respir
Res 12: 64. http://dx.doi.org/10.1186/1465-9921-12-64
Hunter. DP; Carrell-Jacks. LA; Batchelor. TP; Dev. RD. (2011). Role of nerve growth factor in
ozone-induced neural responses in early postnatal airway development. Am J Respir Cell Mol Biol
45: 359-365. http://dx.doi.org/10.1165/rcmb.2010-0345QC
Hwang. BF; Jaakkola. JJK; Lee. YL; Lin. YC; Guo. YLL. (2006). Relation between air pollution and
allergic rhinitis in Taiwananese schoolchildren.
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.l 164/rccm.200706-863QC
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. http://dx.doi.org/10.1038/si.ies.7500627
Jerrett. M; Burnett. RT; Beckerman. BS; Turner. MC; Krewski. D; Thurston. G; Martin. RV; van
Donkelaar. A; Hughes. E; Shi. Y; Gapstur. SM; Thun. MJ; Pope. CA. III. (2013). Spatial analysis
of air pollution and mortality in California. Am J Respir Crit Care Med 188: 593-599.
http://dx.doi.Org/10.l 164/rccm.201303-0609QC
Jerrett. M; Burnett. RT; Pope. CA. Ill; 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
Jiu-Chiuan. C; Schwartz. J. (2008). Metabolic syndrome and inflammatory responses to long-term
particulate air pollutants. Environ Health Perspect 116: 612-617.
Jones. SL; Kittelson. J; Cowan. JO; Flannerv. EM; Hancox. RJ: McLachlan. CR; Taylor. DR. (2001).
The predictive value of exhaled nitric oxide measurements in assessing changes in asthma control.
Am J Respir Crit Care Med 164: 738-743. http://dx.doi.Org/10.l 164/airccm. 164.5.2012125
Just. J; Segala. C; Sahraoui. F; Priol. G; Grimfeld. A; Neukirch. F. (2002). Short-term health effects of
particulate and photochemical air pollution in asthmatic children. Eur Respir J 20: 899-906.
http://dx.doi.org/10.1183/09031936.02.002369Q2
Kahle. JJ; Neas. LM; Devlin. RB; Case. MW; Schmitt. MT; Madden. MC; Diaz-Sanchez. D. (2015).
Interaction effects of temperature and ozone on lung function and markers of systemic
inflammation, coagulation, and fibrinolysis: a crossover study of healthy young volunteers.
Environ Health Perspect 123: 310-316. http://dx.doi.org/10.1289/ehp.1307986
Kasahara. DI; Kim. H; Mathews. JA; Verbout. NG; Williams. AS; Wurmbrand. AP; Ninin. FMC;
Neto. FL; Benedito. LAP; Hug. C; Umetsu. DT; Shore. SA. (2014). Pivotal role of IL-6 in the
hyperinflammatory responses to subacute ozone in adiponectin-deficient mice. Am J Physiol Lung
Cell Mol Physiol 306: L508-L520. http://dx.doi.org/10.1152/aiplung.00235.2013
3-211
-------
Kasahara. DI; Kim. HY: Williams. AS; Verbout. NG; Tran. J; Si. H; Wurmbrand. AP; Jastrab. J; Hug.
C; Umetsu. DT; Shore. SA. (2012). Pulmonary inflammation induced by subacute ozone is
augmented in adiponectin-deficient mice: role of IL-17A. J Immunol 188: 4558-4567.
http://dx.doi.Org/10.4049/iimmunol.l 102363
Kasahara. DI; Mathews. JA; Park. CY; Cho. Y; Hunt. G; Wurmbrand. AP; Liao. JK; Shore. SA.
(2015). ROCK insufficiency attenuates ozone-induced airway hyperresponsiveness in mice. Am J
Physiol Lung Cell Mol Physiol 309: L736-L746. http://dx.doi.org/10.1152/aiplung.00372.2014
Kasahara. DI; Williams. AS; Benedito. LA; Ranscht. B; Kobzik. L; Hug. C; Shore. SA. (2013). Role
of the adiponectin binding protein, T-cadherin (cdhl3), in pulmonary responses to subacute ozone.
PLoS ONE 8: e65829. http://dx.doi.org/10.1371/iournal.pone.0Q65829
Katsouvanni. 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) (pp. 5-90). (Research Report 142). Boston, MA: Health Effects
Institute, https://www.healtheffects.org/publication/air-pollution-and-health-european-
and-north-american-approach
Kharitonov. SA; Barnes. PJ. (2000). Clinical aspects of exhaled nitric oxide [Review]. Eur Respir J
16: 781-792.
Kilkenny. C; Browne. WJ; Cuthill. IC; Emerson. M; Altman. DG. (2010). Improving bioscience
research reporting: The ARRIVE guidelines for reporting animal research [Review]. PLoS Biol 8:
el000412. http://dx.doi.org/10.1371/iournal.pbio.1000412
Kim. CS; Alexis. NE; Rappold. AG; Kehrl. H; Hazucha. MJ; Lav. 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.201011-1813QC
Kirsten. A; Watz. H; Kretschmar. G; Pedersen. F; Bock. D; Mever-Sabellek. W; Magnussen. H.
(2011). Efficacy of the pan-selectin antagonist Bimosiamose on ozone-induced airway
inflammation in healthy subjects - A double blind, randomized, placebo-controlled, cross-over
clinical trial. Pulm Pharmacol Ther 24: 555-558. http://dx.doi.org/10.1016/i .pupt.2011.04.029
Klemm. RJ; Lipfert. FW; Wvzga. RE; Gust. C. (2004). Daily mortality and air pollution in Atlanta:
two years of data from ARIES. Inhal Toxicol 16 Suppl 1: 131-141.
http://dx.doi.org/10.1080/0895837049Q443213
Klemm. RJ; Mason. RM. Jr. (2000). Aerosol Research and Inhalation Epidemiological Study
(ARIES): air quality and daily mortality statistical modeling—interim results. J Air Waste Manag
Assoc 50: 1433-1439.
Klemm. RJ; Thomas. EL; Wvzga. RE. (2011). The impact of frequency and duration of air quality
monitoring: Atlanta, GA, data modeling of air pollution and mortality. J Air Waste Manag Assoc
61: 1281-1291. http://dx.doi.org/10.108Q/10473289.2011.617648
Klimisch. HJ; Andreae. M; Tillmann. U. (1997). A systematic approach for evaluating the quality of
experimental toxicological and ecotoxicological data. Regul Toxicol Pharmacol 25: 1-5.
http://dx.doi.org/10.1006/rtph.1996.1076
Kodavanti. UP; Ledbetter. AD; Thomas. RF; Richards. JE; Ward. WO; Schladweiler. MC; Costa. PL.
(2015). Variability in ozone-induced pulmonary injury and inflammation in healthy and
cardiovascular-compromised rat models. Inhal Toxicol 27: 39-53.
http://dx.doi.org/10.3109/08958378.2014.954169
3-212
-------
Kousha. T; Castner. J. (2016). The air quality health index and emergency department visits for otitis
media. JNurs Scholarsh 48: 163-171. http://dx.doi.org/10. Ill 1/jnu. 12195
Kousha. T: Rowe. BH. (2014). Ambient ozone and emergency department visits due to lower
respiratory condition. Int J Occup Med Environ Health 27: 50-59.
http://dx.doi.org/10.2478/sl3382-014-Q229-0
Kreit. JW; Gross. KB: Moore. TB: Lorenzen. TJ: D'Arcv. J; Eschenbacher. WL. (1989). Ozone-
induced changes in pulmonary function and bronchial responsiveness in asthmatics. J Appl Physiol
(1985)66:217-222.
Kumagai. K: Lewandowski. R: Jackson-Humbles. DN: Li. N: Van Dvken. SJ: Wagner. JG: Harkema.
JR. (2016). Ozone-induced nasal type 2 immunity in mice is dependent on innate lymphoid cells.
Am J Respir Cell Mol Biol 54: 782-791. http://dx.doi.Org/10.l 165/rcmb.2015-0118QC
Kumagai. K; Lewandowski. RP: Jackson-Humbles. DN: Buglak. N: Li. N: White. K; Van Dvken. SJ:
Wagner. JG: Harkema. JR. (2017). Innate lymphoid cells mediate pulmonary eosinophilic
inflammation, airway mucous cell metaplasia, and type 2 immunity in mice exposed to ozone.
Toxicol Pathol 45: 692-704. http://dx.doi.org/10.1177/0192623317728135
Kumarathasan. P: Blais. E: Saravanamuthu. A: Bielecki. A: Mukheriee. B: Biarnason. S: Guenette. J:
Goegan. P: Vincent. R. (2015). Nitrative stress, oxidative stress and plasma endothelin levels after
inhalation of particulate matter and ozone. Part Fibre Toxicol 12: 28.
http://dx.doi.org/10.1186/sl2989-015-0103-7
Kummarapurugu. AB: Fischer. BM: Zheng. S: Milne. GL: Ghio. AJ: Potts-Kant. EN: Foster. WM:
Soderblom. EJ; Dubois. LG: Moselev. MA: Thompson. JW: Vovnow. JA. (2013). NADPH:
quinone oxidoreductase 1 regulates host susceptibility to ozone via isoprostane generation. J Biol
Chem 288: 4681-4691. http://dx.doi.org/10.1074/ibc.M112.438440
Kurhanewicz. N: Mcintosh-Kastrinskv. R: Tong. H: Walsh. L: Farrai. A: Hazari. MS. (2014). Ozone
co-exposure modifies cardiac responses to fine and ultrafine ambient particulate matter in mice:
Concordance of electrocardiogram and mechanical responses. Part Fibre Toxicol 11: 54.
http://dx.doi.org/10.1186/sl2989-014-0054-4
Lazaar. AL; Sweeney. LE; Macdonald. AJ: Alexis. NE; Chen. C: Tal-Singer. R. (2011). SB-656933, a
novel CXCR2 selective antagonist, inhibits ex-vivo neutrophil activation and ozone-induced
airway inflammation in humans. Br J Clin Pharmacol 72: 282-293.
http://dx.doi.Org/10.l 111/i. 1365-2125.2011.03968.x
Lee. D; Wallis. C: Van Winkle. LS: Wexler. AS. (2011). Disruption of tracheobronchial airway
growth following postnatal exposure to ozone and ultrafine particles. Inhal Toxicol 23: 520-531.
http://dx.doi.org/10.3109/08958378.2011.591447
Lee. MS: Moon. KY; Bae. DJ: Park. MK: Jang. AS. (2013). The effects of pycnogenol on antioxidant
enzymes in a mouse model of ozone exposure. 28: 216-223.
http://dx.doi.Org/10.3904/kiim.2013.28.2.216
Lepeule. J: Bind. MA: Baccarelli. AA; Koutrakis. P; Tarantini. L; Litoniua. A: Sparrow. D; Vokonas.
P: Schwartz. JD. (2014). Epigenetic influences on associations between air pollutants and lung
function in elderly men: the normative aging study. Environ Health Perspect 122: 566-572.
http://dx.doi.org/10.1289/ehp.1206458
Leroy. P: Tham. A: Wong. H: Tennev. R: Chen. C: Stiner. R: Balmes. JR: Paquet. AC: Ariomandi. M.
(2015). Inflammatory and repair pathways induced in human bronchoalveolar lavage cells with
ozone inhalation. PLoS ONE 10: eOl27283. http://dx.doi.org/10.1371/iournal.pone.0127283
3-213
-------
Lewis. TC; Robins. TG; Mentz. GB; Zhang. X; Mukheriee. B; Lin. X; Keeler. GJ; Dvonch. JT; Yip.
FY; O'Neill. MS; Parker. EA; Israel. BA; Max. PT; Reves. A; Committee. CAAACS. (2013). Air
pollution and respiratory symptoms among children with asthma: Vulnerability by corticosteroid
use and residence area. Sci Total Environ 448: 48-55.
http ://dx.doi .org/10.1016/i. scitotenv.2012.11.070
Liu. J; Mao. F; Mao. Z; Wang. F; Zhu. Y; Wang. H. (2016). Emodin inhibited ozone stress-induced
airway hyperresponsiveness by regulating epithelial protection mechanisms against injury.
International Journal of Clinical and Experimental Medicine 9: 17404-17411.
Maclntvre. EA; Karr. CJ; Koehoorn. M; Demers. PA; Tamburic. L; Lencar. C; Brauer. M. (2011).
Residential air pollution and otitis media during the first two years of life. Epidemiology 22: 81-89.
http://dx.doi.org/10.1097/EDE.0b013e3181fdb6Qf
Madden. MC; Stevens. T; Case. M; Schmitt. M; Diaz-Sanchez. D; Bassett. M; Montilla. TS; Berntsen.
J; Devlin. RB. (2014). Diesel exhaust modulates ozone-induced lung function decrements in
healthy human volunteers. Part Fibre Toxicol 11: 37. http://dx.doi.Org/10.l 186/sl2989-014-0037-5
Magzamen. S; Oron. AP; Locke. ER; Fan. VS. (2018). Association of ambient pollution with inhaler
use among patients with COPD: a panel study. Occup Environ Med.
http://dx.doi.Org/10.l 136/oemed-2017-104808
Malig. BJ; Pearson. PL; Chang. Y; Broadwin. R; Basu. R; Green. RS; Ostro. B. (2016). A time-
stratified case-crossover study of ambient ozone exposure and emergency department visits for
specific respiratory diagnoses in California (2005-2008). Environ Health Perspect 124: 745-753.
http://dx.doi.org/10.1289/ehp. 1409495
Malik. F; Cromar. KR; Atkins. CL; Price. RE; Jackson. WT; Siddiqui. SR; Spencer. CY; Mitchell.
NC; Hague. IU; Johnston. RA. (2017). Chemokine (C-C Motif) Receptor-Like 2 is not essential for
lung injury, lung inflammation, or airway hyperresponsiveness induced by acute exposure to
ozone. Physiol Rep 5: el3545. http://dx.doi.org/10.14814/phv2.13545
Mathews. JA; Kasahara. DI; Cho. Y; Bell. LN; Gunst. PR; Karolv. ED; Shore. SA. (2017a). Effect of
acute ozone exposure on the lung metabolomes of obese and lean mice. PLoS ONE 12: e0181017.
http://dx.doi.org/10.1371/iournal.pone.0181Q17
Mathews. JA; Kasahara. DI; Ribeiro. L; Wurmbrand. AP; Ninin. FMC; Shore. SA. (2015). gamma
delta T cells are required for M2 macrophage polarization and resolution of ozone-induced
pulmonary inflammation in mice. PLoS ONE 10: e0131236.
http://dx.doi.org/10.1371/iournal.pone.0131236
Mathews. JA; Krishnamoorthv. N; Kasahara. DI; Cho. Y; Wurmbrand. AP; Ribeiro. L; Smith. D;
Umetsu. D; Lew. BP; Shore. SA. (2017b). IL-33 drives augmented responses to ozone in obese
mice. Environ Health Perspect 125: 246-253. http://dx.doi.org/10.1289/EHP272
Mathews. JA; Krishnamoorthv. N; Kasahara. PI; Hutchinson. J; Cho. Y; Brand. JP; Williams. AS;
Wurmbrand. AP; Ribeiro. L; Cuttitta. F; Sunday. ME; Lew. BP; Shore. SA. (2018). Augmented
responses to ozone in obese mice require IL-17A and gastrin-releasing peptide. Am J Respir Cell
Mol Biol 58: 341-351. http://dx.doi.org/10.1165/rcmb.2017-0071QC
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. http://dx.doi.org/10.1016/S0140-6736(02)07597-9
3-214
-------
McConnell. R; Berhane. K; Gilliland. F; Molitor. J; Thomas. D; Lurmann. F; Avol. E; Gauderman.
WJ; Peters. JM. (2003). Prospective study of air pollution and bronchitic symptoms in children
with asthma. Am J Respir Crit Care Med 168: 790-797. http://dx.doi.Org/10.l 164/rccm.200304-
466QC
McDonnell. WF; Abbey. DE; Nishino. N; Lebowitz. MP. (1999a). Long-term ambient ozone
concentration and the incidence of asthma in nonsmoking adults: the Ahsmog study. Environ Res
80: 110-121. http://dx.doi.org/10.1006/enrs.1998.3894
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. http://dx.doi.org/10.1080/00Q39896.1991.9937441
McDonnell. WF; Stewart. PW; Smith. MY. (2010). Prediction of ozone-induced lung function
responses in humans. Inhal Toxicol 22: 160-168. http://dx.doi.org/10.3109/08958370903Q89557
McDonnell. WF; Stewart. PW; Smith. MY. (2013). Ozone exposure-response model for lung function
changes: An alternate variability structure. Inhal Toxicol 25: 348-353.
http://dx.doi.org/10.3109/08958378.2013.79Q523
McDonnell. WF; Stewart. PW; Smith. MY; 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. http://dx.doi.org/10.3109/08958378.2012.7Q5919
McDonnell. WF; Stewart. PW; Smith. MY; Pan. WK; Pan. J. (1999b). Ozone-induced respiratory
symptoms: Exposure-response models and association with lung function. Eur Respir J 14: 845-
853. http://dx.doi.Org/10.1034/i.1399-3003.1999.14d21.x
Mclntosh-Kastrinsky. R; Diaz-Sanchez. D; Sexton. KG; Jania. CM; Zavala. J; Tillev. SL; Jaspers. I;
Gilmour. MI; Devlin. RB: Cascio. WE; Tong. H. (2013). Photochemically altered air pollution
mixtures and contractile parameters in isolated murine hearts before and after ischemia. Environ
Health Perspect 121: 1344-1348. http://dx.doi.org/10.1289/ehp.1306609
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/aie/kwi078
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/iech.2008.083576
Michaudel. C; Mackowiak. C; Maillet. I; Fauconnier. L; Akdis. CA; Sokolowska. M; Dreher. A; Tan.
HT; Ouesniaux. VF; Rvffel. B; Togbe. D. (2018). Ozone exposure induces respiratory barrier
biphasic injury and inflammation controlled by IL-33. J Allergy Clin Immunol.
http://dx.doi.Org/10.1016/i.iaci.2017.ll.044
Mikerov. AN; Cooper. TK; Wang. G; Hu. S; Umstead. TM; Phelps. DS; Floros. J. (2011).
Histopathologic evaluation of lung and extrapulmonary tissues show sex differences in Klebsiella
pneumoniae - infected mice under different exposure conditions. International Journal of
Physiology, Pathophysiology and Pharmacology 3: 176-190.
Miller. CN; Dve. JA; Ledbetter. AD; Schladweiler. MC; Richards. JH; Snow. SJ; Wood. CE;
Henriquez. AR; Thompson. LC; Farrai. AK; Hazari. MS; Kodavanti. UP. (2017). Uterine artery
flow and offspring growth in long-evans rats following maternal exposure to ozone during
implantation. Environ Health Perspect 125: 127005. http://dx.doi.org/10.1289/EHP2019
3-215
-------
Miller. DB; Snow. SJ; Henriquez. A; Schladweiler. MC; Ledbetter. AD; Richards. JE; Andrews. PL;
Kodavanti. UP. (2016a). Systemic metabolic derangement, pulmonary effects, and insulin
insufficiency following subchronic ozone exposure in rats. Toxicol Appl Pharmacol 306: 47-57.
http://dx.doi.Org/10.1016/i.taap.2016.06.027
Miller. DB; Snow. SJ; Schladweiler. MC; Richards. JE; Ghio. AJ; Ledbetter. AD; Kodavanti. UP.
(2016b). Acute ozone-induced pulmonary and systemic metabolic effects are diminished in
adrenalectomized rats. Toxicol Sci 150: 312-322. http://dx.doi.org/10.1093/toxsci/kfV331
Mishra. V; Diangelo. SL; Silvevra. P. (2016). Sex-specific IL-6-associated signaling activation in
ozone-induced lung inflammation. 7: 16. http://dx.doi.Org/10.l 186/sl3293-016-0069-7
Moore. BP; Hyde. D; Miller. L; Wong. E; Frelinger. J; Schelegle. ES. (2012a). Allergen and ozone
exacerbate serotonin-induced increases in airway smooth muscle contraction in a model of
childhood asthma. Respiration 83: 529-542. http://dx.doi.Org/10.l 159/000336835
Moore. K; Neugebauer. R; l.nrmann. F; Hall. J; Braier. 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
Moore. KL; Neugebauer. R; van Per Laan. MJ; Tager. I. (2012b). Causal inference in epidemiological
studies with strong confounding. Stat Med 31: 1380-1404. http ://dx. doi .org/10.1002/sim.4469
Mortimer. KM; Neas. LM; Pockery. PW; Redline. S; Tager. IB. (2002). The effect of air pollution on
inner-city children with asthma. Eur Respir J 19: 699-705.
http://dx.doi.org/10.1183/09031936.02.002471Q2
Mortimer. KM; Tager. IB; Pockery. PW; Neas. LM; Redline. S. (2000). The effect of ozone on inner-
city children with asthma: Identification of susceptible subgroups. Am J Respir Crit Care Med 162:
1838-1845. http://dx.doi.Org/10.1164/airccm.162.5.9908113
Muhaiarine. N; Mustard. C; Roos. LL; Young. TK; Gelskev. PE. (1997). Comparison of survey and
physician claims data for detecting hypertension. J Clin Epidemiol 50: 711-718.
http://dx.doi.org/10.1016/S0895-4356(97')00019-X
Murgia. N; Brisman. J; Claesson. A; Muzi. G; Olin. AC; Toren. K. (2014). Validity of a questionnaire-
based diagnosis of chronic obstructive pulmonary disease in a general population-based study.
BMC Pulm Med 14: 49. http://dx.doi.org/10.1186/1471-2466-14-49
Murphy. SR; Oslund. KL; Hyde. PM; Miller. LA; Van Winkle. LS; Schelegle. ES. (2014). Ozone-
induced airway epithelial cell death, the neurokinin-1 receptor pathway, and the postnatal
developing lung. Am J Physiol Lung Cell Mol Physiol 307: L471-L481.
http://dx.doi.org/10.1152/aiplung.00324.2013
Murphy. SR; Schelegle. ES; Edwards. PC; Miller. LA; Hyde. PM; Van Winkle. LS. (2012). Postnatal
exposure history and airways oxidant stress responses in airway explants. Am J Respir Cell Mol
Biol 47: 815-823. http://dx.doi.Org/10.l 165/rcmb.2012-0110OC
Murphy. SR; Schelegle. ES; Miller. LA; Hyde. PM; Van Winkle. LS. (2013). Ozone exposure alters
serotonin and serotonin receptor expression in the developing lung. Toxicol Sci 134: 168-179.
http://dx.doi.org/10.1093/toxsci/kft090
NAEPP (National Asthma Education and Prevention Program). (2008). Expert panel report 3 (EPR-3):
Guidelines for the diagnosis and management of asthmasummary report 2007 - Updates [Erratum],
J Allergy Clin Immunol 120: 1330. http://dx.doi.Org/10.1016/i.iaci.2008.04.033
3-216
-------
Neophvtou. AM; White. MJ; Oh. SS; Thakur. N; Galanter. JM; Nishimura. KK; Pino-Yanes. M;
Torgerson. DG; Gignoux. CR; Eng. C; Nguyen. EA; Hu. D; Mak. AC; Kumar. R; Seibold. MA;
Davis. A; Farber. HJ; Meade. K; Avila. PC; Serebriskv. D; Lenoir. MA; Brigino-Buenaventura. E;
Rodriguez-Cintron. W; Bibbins-Domingo. K; Thyne. SM; Williams. LK; Sen. S; Gilliland. FD;
Gauderman. WJ; Rodriguez-Santana. JR; Lurmann. F; Balmes. JR; Eisen. EA; Burchard. EG.
(2016). Air pollution and lung function in minority youth with asthma in the GALA II
(Genesenvironments and admixture in Latino Americans) and SAGE II (Study of African
Americans, asthma, genes, and environments) studies. Am J Respir Crit Care Med 193: 1271-1280.
http://dx.doi.Org/10.l 164/rccm.201508-1706QC
NHLBI (National Institutes of Health, National Heart Lung and Blood Institute). (2017). NHLBI fact
book, fiscal year 2012: Disease statistics. Available online at
https://www.nhlbi.nih.gov/about/documents/factbook/2012/chapter4 (accessed August 23, 2017).
Nishimura. KK; Galanter. JM; Roth. LA; Oh. SS; Thakur. N; Nguyen. EA; Thyne. S; Farber. HJ;
Serebriskv. D; Kumar. R; Brigino-Buenaventura. E; Davis. A; Lenoir. MA; Meade. K; Rodriguez-
Cintron. W; Avila. PC; Borrell. LN; Bibbins-Domingo. K; Rodriguez-Santana. JR; Sen. S;
Lurmann. F; Balmes. JR; Burchard. EG. (2013). Early-life air pollution and asthma risk in minority
children: the GALA II and SAGE II studies. Am J Respir Crit Care Med 188: 309-318.
http://dx.doi.Org/10.l 164/rccm.2013 02-0264QC
O' Lenick. CR; Chang. HH; Kramer. MR; Winquist. A; Mulholland. JA; Friberg. MP; Sarnat. SE.
(2017). Ozone and childhood respiratory disease in three US cities: Evaluation of effect measure
modification by neighborhood socioeconomic status using a Bayesian hierarchical approach.
Environ Health 16: #36. http://dx.doi.Org/10.l 186/s 12940-017-0244-2
O'Connor. GT; Neas. L; Vaughn. B: Kattan. M; Mitchell. H; Crain. EF; Evans. R. Ill; Gruchalla. R;
Morgan. W; Stout. J; Adams. GK; Lippmann. M. (2008). Acute respiratory health effects of air
pollution on children with asthma in US inner cities. J Allergy Clin Immunol 121: 1133-
1139.el 131. http://dx.doi.Org/10.1016/i.iaci.2008.02.020
O'Lenick. CR; Winquist. A; Mulholland. JA; Friberg. MP; Chang. HH; Kramer. MR; Darrow. LA;
Sarnat. SE. (2017). Assessment of neighbourhood-level socioeconomic status as a modifier of air
pollution-asthma associations among children in Atlanta. J Epidemiol Community Health 71: 129-
136. http://dx.doi.org/10.1136/iech-2015-20653Q
Ong. CB; Kumagai. K; Brooks. PT; Brandenberger. C; Lewandowski. RP; Jackson-Humbles. PN;
Nault. R; Zacharewski. TR; Wagner. JG; Harkema. JR. (2016). Ozone-induced type 2 immunity in
nasal airways. Pevelopment and lymphoid cell dependence in mice. Am J Respir Cell Mol Biol 54:
331-340. http://dx.doi.org/10.1165/rcmb.2015-0165QC
Paffett. ML; Zvchowski. KE; Sheppard. L; Robertson. S; Weaver. JM; Lucas. SN; Campen. MJ.
(2015). Ozone inhalation impairs coronary artery dilation via intracellular oxidative stress:
Evidence for serum-borne factors as drivers of systemic toxicity. Toxicol Sci 146: 244-253.
http://dx.doi.org/10.1093/toxsci/kfVQ93
Parker. JP; Akinbami. LJ; Woodruff. TJ. (2009). Air pollution and childhood respiratory allergies in
the United States. Environ Health Perspect 117: 140-147. http://dx.doi.org/10.1289/ehp.11497
Patel. MM; Chillrud. SN; Peepti. KC; Ross. JM; Kinney. PL. (2013). Traffic-related air pollutants and
exhaled markers of airway inflammation and oxidative stress in New York City adolescents.
Environ Res 121: 71-78. http://dx.doi.Org/10.1016/i.envres.2012.10.012
Peacock. JL; Anderson. HR; Bremner. SA; Marston. L; Seemungal. TA; Strachan. PP; Wedzicha. JA.
(2011). Outdoor air pollution and respiratory health in patients with COPP. Thorax 66: 591-596.
http://dx.doi.org/10.1136/thx.201Q.155358
3-217
-------
Penard-Morand. C; Charpin. D; Raherison. C; Kopferschmitt. C; Caillaud. D; Lavaud. F; Annesi-
Maesano. I. (2005). Long-term exposure to background air pollution related to respiratory and
allergic health in schoolchildren. Clin Exp Allergy 35: 1279-1287.
http://dx.doi.org/10.1111/i. 1365-2222.2005.02336.x
Peng. C; Luttmann-Gibson. H; Zanobetti. A; Cohen. A; De Souza. C; Coull. BA; Horton. ES;
Schwartz. J: Koutrakis. P; Gold. DR. (2016). Air pollution influences on exhaled nitric oxide
among people with type II diabetes. Air Qual Atmos Health 9: 265-273.
http://dx.doi.org/10.1007/sll869-015-Q336-5
Oian. Z: Liao. D; Lin. HM: Whitsel. EA: Rose. KM: Duan. Y. (2005). Lung function and long-term
exposure to air pollutants in middle-aged American adults. Arch Environ Occup Health 60: 156-
163. http://dx.doi.Org/10.3200/AEQH.60.3.156-163
Rage. E; Jacquemin. B; Nadif. R; Orvszczvn. MP: Siroux. V: Aguilera. I: Kauffmann. F; Kiinzli. N.
(2009). Total serum IgE levels are associated with ambient ozone concentration in asthmatic adults.
Allergy 64: 40-46. http://dx.doi.Org/10.llll/i.1398-9995.2008.01800.x
Raizenne. ME: Burnett. RT; Stern. B; Franklin. CA; Spengler. JD. (1989). Acute lung function
responses to ambient acid aerosol exposures in children. Environ Health Perspect 79: 179-185.
Ramadour. M: Burel. C: Lanteaume. A: Vervloet. D: Charpin. D: Brisse. F: Dutau. H: Charpin. D.
(2000). Prevalence of asthma and rhinitis in relation to long-term exposure to gaseous air
pollutants. Allergy 55: 1163-1169. http://dx.doi.Org/10.1034/i.1398-9995.2000.00637.x
Ramot. Y: Kodavanti. UP: Kissling. GE: Ledbetter. AD: Nvska. A. (2015). Clinical and pathological
manifestations of cardiovascular disease in rat models: the influence of acute ozone exposure. Inhal
Toxicol 27: 26-38. http://dx.doi.org/10.3109/08958378.2014.954168
Razvi. SS: Richards. JB: Malik. F: Cromar. KR: Price. RE: Bell. CS: Weng. T: Atkins. CL: Spencer.
CY: Cockerill. KJ: Alexander. AL: Blackburn. MR: Alcorn. JL: Hague. IT J: Johnston. RA. (2015).
Resistin deficiency in mice has no effect on pulmonary responses induced by acute ozone
exposure. Am J Physiol Lung Cell Mol Physiol 309: LI 174 LI 185.
http://dx.doi.org/10.1152/aiplung.00270.2Q15
Rice. MB: Liungman. PL: Wilker. EH: Gold. PR: Schwartz. JD: Koutrakis. P; Washko. GR;
O'Connor. GT: Mittleman. MA. (2013). Short-term exposure to air pollution and lung function in
the Framingham heart study. Am J Respir Crit Care Med 188: 1351-1357.
http://dx.doi.Org/10.l 164/rccm.201308-1414QC
Robertson. S: Colombo. ES: Lucas. SN: Hall. PR: Febbraio. M; Paffett. ML: Campen. MJ. (2013).
CD36 mediates endothelial dysfunction downstream of circulating factors induced by 03 exposure.
Toxicol Sci 134: 304-311. http://dx.doi.org/10.1093/toxsci/kftl07
Rodopoulou. S: Samoli. E: Chalbot. MG: Kavouras. IG. (2015). Air pollution and cardiovascular and
respiratory emergency visits in Central Arkansas: A time-series analysis. Sci Total Environ 536:
872-879. http://dx.doi.Org/10.1016/i.scitotenv.2015.06.056
Roonev. AA: Bovles. AL: Wolfe. MS: Bucher. JR: Thayer. KA. (2014). Systematic review and
evidence integration for literature-based environmental health science assessments. Environ Health
Perspect 122: 711-718. http://dx.doi.org/10.1289/ehp.1307972
Ross. MA: Perskv. VW: Scheff. PA: Chung. J: Curtis. L: Ramakrishnan. V: Wadden. RA:
Hrvhorczuk. DO. (2002). Effect of ozone and aeroallergens on the respiratory health of asthmatics.
Arch Environ Occup Health 57: 568-578. http://dx.doi.org/10.1080/0003989020960209Q
3-218
-------
Sacks. JD; Rappold. AG; Davis. JA. Jr; Richardson. DB; Waller. AE; Luben. TJ. (2014). Influence of
urbanicity and county characteristics on the association between ozone and asthma emergency
department visits in North Carolina. Environ Health Perspect 122: 506-512.
http://dx.doi.org/10.1289/ehp.130694Q
Salam. MT; Byun. HM; Lurmann. F; Breton. CV; Wang. X; Eckel. SP; Gilliland. FD. (2012). Genetic
and epigenetic variations in inducible nitric oxide synthase promoter, particulate pollution, and
exhaled nitric oxide levels in children. J Allergy Clin Immunol 129: 232-239.e237.
http://dx.doi.Org/10.1016/i.iaci.2011.09.037
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.Org/10.1016/i.iaci.2008.12.020
Sarnat. JA: Sarnat. SE; Flanders. WD: Chang. HH: Mulholland. J: Baxter. L; Isakov. V: Ozkavnak. H.
(2013). Spatiotemporally resolved air exchange rate as a modifier of acute air pollution-related
morbidity in Atlanta. J Expo Sci Environ Epidemiol 23: 606-615.
http://dx.doi.org/10.1038/ies.2013.32
Sarnat. SE: Winquist. A: Schauer. JJ; Turner. JR: Sarnat. JA. (2015). Fine particulate matter
components and emergency department visits for cardiovascular and respiratory diseases in the St.
Louis, Missouri-Illinois, metropolitan area. Environ Health Perspect 123: 437-444.
http://dx.doi.org/10.1289/ehp.1307776
Schelegle. ES: Adams. WC: Walbv. 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: Morales. CA: Walbv. 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.200809-1484QC
Schelegle. ES: Walbv. WF. (2012). Vagal afferents contribute to exacerbated airway responses
following ozone and allergen challenge. Respir Physiol Neurobiol 181: 277-285.
http://dx.doi.Org/10.1016/i.resp.2012.04.003
Schittnv. JC. (2017). Development of the lung [Review]. Cell Tissue Res 367: 427-444.
http://dx.doi.org/10.1007/s00441-Q16-2545-0
Sheffield. PE; Zhou. J: Shmool. JL; Cloughertv. JE. (2015). Ambient ozone exposure and children's
acute asthma in New York City: a case-crossover analysis. Environ Health 14: 25.
http://dx.doi.org/10.1186/sl2940-015-001Q-2
Shmool. JLC: Kinnee. E; Sheffield. PE; Cloughertv. JE. (2016). Spatio-temporal ozone variation in a
case-crossover analysis of childhood asthma hospital visits in New York City. Environ Res 147:
108-114. http://dx.doi.Org/10.1016/i.envres.2016.01.020
Shore. SA; Williams. ES; Chen. L; Benedito. LA; Kasahara. DI; Zhu. M. (2011). Impact of aging on
pulmonary responses to acute ozone exposure in mice: role of TNFR1. Inhal Toxicol 23: 878-888.
http://dx.doi.org/10.3109/08958378.2011.622316
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/i.iaci.2009.10.061
3-219
-------
Smith. GS; Van Den Eeden. SK; Garcia. C; Shan. J; Baxter. R; Herring. AH; Richardson. DB; Van
Rie. A; Emch. M; Gammon. MP. (2016). Air pollution and pulmonary tuberculosis: A nested case-
control study among members of a northern California health plan. Environ Health Perspect 124:
761-768. http://dx.doi.org/10.1289/ehp.1408166
Snow. SJ; Cheng. WY; Henriquez. A; Hodge. M; Bass. V; Nelson. GM; Carswell. G; Richards. JE;
Schladweiler. MC: Ledbetter. AD; Chorlev. B; Gowdv. KM; Tong. H; Kodavanti. UP. (2018).
Ozone-induced vascular contractility and pulmonary injury are differentially impacted by diets
enriched with coconut oil, fish oil, and olive oil. Toxicol Sci 163: 5769.
http://dx.doi.org/10.1093/toxsci/kfV003
Snow. SJ; Gordon. CJ; Bass. VL; Schladweiler. MC; Ledbetter. AD; Jarema. KA; Phillips. PM;
Johnstone. AF; Kodavanti. UP. (2016). Age-related differences in pulmonary effects of acute and
subchronic episodic ozone exposures in Brown Norway rats. Inhal Toxicol 28: 313-323.
http://dx.doi.org/10.3109/08958378.2016.117091Q
Speen. AM; Kim. HH; Bauer. RN; Meyer. M; Gowdv. KM; Fessler. MB; Duncan. KE; Liu. W; Porter.
NA; Jaspers. I. (2016). Ozone-derived oxysterols affect liver X receptor (LXR) signaling: a
potential role for lipid-protein adducts. J Biol Chem 291: 25192-25206.
http://dx.doi.org/10.1074/ibc.M116.732362
Spektor. DM; Lippmann. M. (1991). Health effects of ambient ozone on healthy children at a summer
camp. In RL Berglund; DR Lawson; DJ McKee (Eds.), Tropospheric Ozone and the Environment:
Papers from an International Conference; March 1990; Los Angeles, CA (pp. 83-89). Pittsburgh,
PA: Air & Waste Management Association.
Spektor. DM; Lippmann. M; Liov. PJ; Thurston. GD; Citak. K; James. DJ; Bock. N; Speizer. FE;
Haves. C. (1988). Effects of ambient ozone on respiratory function in active, normal children. Am
RevRespirDis 137: 313-320. http://dx.doi.org/10.1164/airccm/137.2.313
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.200908-1269QC
Stieb. DM; Szvszkowicz. 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 8:
25. http://dx.doi.org/10.1186/1476-069X-8-25
Stiegel. MA; Pleil. JD; Sobus. JR; Stevens. T; Madden. MC. (2017). Linking physiological parameters
to perturbations in the human exposome: Environmental exposures modify blood pressure and lung
function via inflammatory cytokine pathway. J Toxicol Environ Health A 80: 485-501.
http://dx.doi.org/10.1080/15287394.2017.133Q578
Stober. VP; Johnson. CG; Majors. A; Lauer. ME; Cali. V; Midura. RJ; Wisniewski. HG; Aronica.
MA; Garantziotis. S. (2017). TNF-stimulated gene 6 promotes formation of hyaluronan-inter-a-
inhibitor heavy chain complexes necessary for ozone-induced airway hyperresponsiveness. J Biol
Chem 292: 20845-20858. http://dx.doi.org/10.1074/ibc.M116.756627
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.l 164/rccm.200908-120 IOC
3-220
-------
Strickland. MJ; Klein. M; Flanders. WD; Chang. HH: Mulholland. JA; Tolbert. PE; Darrow. LA.
(2014). Modification of the effect of ambient air pollution on pediatric asthma emergency visits:
susceptible subpopulations. Epidemiology 25: 843-850.
http://dx.doi.org/10.1097/EDE.000000000000017Q
Sunil. VR; Francis. M; Vavas. KN; Cervelli. JA; Choi. H; Laskin. JD; Laskin. PL. (2015). Regulation
of ozone-induced lung inflammation and injury by the beta-galactoside-binding lectin galectin-3.
Toxicol Appl Pharmacol 284: 236-245. http://dx.doi.Org/10.1016/i.taap.2015.02.002
Sunil. VR; Patel-Vavas. K; Shen. J; Laskin. JD; Laskin. PL. (2012). Classical and alternative
macrophage activation in the lung following ozone-induced oxidative stress. Toxicol Appl
Pharmacol 263: 195-202. http://dx.doi.Org/10.1016/i.taap.2012.06.009
Sunil. VR; Vavas. KJ; Massa. CB; Gow. AJ; Laskin. JD; Laskin. PL. (2013). Ozone-induced injury
and oxidative stress in bronchiolar epithelium is associated with altered pulmonary mechanics.
Toxicol Sci 133: 309-319. htto://dx.doi.org/10.1093/toxsci/kft071
Szyszkowicz. M; Kousha. T; Castner. J; Pales. R. (2018). Air pollution and emergency department
visits for respiratory diseases: A multi-city case crossover study. Environ Res 163: 263-269.
http://dx.doi.org/10.1016/i .envres.2018.01.043
Tank. J; Biller. H; Heusser. K; Holz. O; Piedrich. A; Framke. T; Koch. A; Grosshennig. A: Koch. W;
Krug. N; Jordan. J; Hohlfeld. JM. (2011). Effect of acute ozone induced airway inflammation on
human sympathetic nerve traffic: a randomized, placebo controlled, crossover study. PLoS ONE 6:
el8737. http://dx.doi.org/10.1371/iournal.pone.0Q18737
Tankerslev. CG; Georgakopoulos. P; Tang. WY; Sborz. N. (2013). Effects of ozone and particulate
matter on cardiac mechanics: role of the atrial natriuretic peptide gene. Toxicol Sci 131: 95-107.
http://dx.doi.org/10.1093/toxsci/kfs273
Tetreault. LF; Poucet. M; Gamache. P; Fournier. M; Brand. A; Kosatskv. T; Smargiassi. A. (2016a).
Childhood exposure to ambient air pollutants and the onset of asthma: an administrative cohort
study in Quebec. Environ Health Perspect 124: 1276-1282. http://dx.doi.org/10.1289/ehp.1509838
Tetreault. LF; Poucet. M; Gamache. P; Fournier. M; Brand. A; Kosatskv. T; Smargiassi. A. (2016b).
Severe and moderate asthma exacerbations in asthmatic children and exposure to ambient air
pollutants. Int J Environ Res Public Health 13: 771. http://dx.doi.org/10.3390/iierphl3080771
Thomson. EM; Pal. S; Guenette. J; Wade. MG; Atlas. E; Hollowav. AC; Williams. A; Vincent. R.
(2016). Ozone inhalation provokes glucocorticoid-dependent and -independent effects on
inflammatory and metabolic pathways. Toxicol Sci 152: 17-28.
http://dx.doi.org/10.1093/toxsci/kfw061
Thomson. EM; Vladisavlievic. P; Mohottalage. S; Kumarathasan. P; Vincent. R. (2013). Mapping
acute systemic effects of inhaled particulate matter and ozone: multiorgan gene expression and
glucocorticoid activity. Toxicol Sci 135: 169-181. http://dx.doi.org/10.1093/toxsci/kftl37
Tighe. RM; Birukova. A; Yeager. MJ; Reece. SW; Gowdv. KM. (2018). Euthanasia and lavage
mediated effects on bronchoalveolar measures of lung injury and inflammation. Am J Respir Cell
Mol Biol. http://dx.doi.Org/10.l 165/rcmb.2017-0357QC
To. T; Zhu. J; Larsen. K; Simatovic. J; Feldman. L; Rvckman. K; Gershon. A; Lougheed. MP;
Licskai. C; Chen. H; Villeneuve. PJ; Crighton. E; Su. Y; Sadatsafavi. M; Williams. P; Carlsten. C.
(2016). Progression from asthma to chronic obstructive pulmonary disease (COPP): is air pollution
a risk factor? Am J Respir Crit Care Med 194: 429-438. http://dx.doi.Org/10.l 164/rccm.201510-
1932QC
3-221
-------
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/si.ies.750Q625
Toren. K; Brisman. J; Jarvholm. B. (1993). Asthma and asthma-like symptoms in adults assessed by
questionnaires: A literature review [Review]. Chest 104: 600-608.
http://dx.doi.Org/10.1378/chest.104.2.600
Turner. MC; Jerrett. M: Pope. A. TIT: Krewski. D: Gapstur. SM: Diver. WR; Beckerman. BS;
Marshall. JD; Su. J; Crouse. PL; Burnett. RT. (2016). Long-term ozone exposure and mortality in a
large prospective study. Am J Respir Crit Care Med 193: 1134-1142.
http://dx.doi.Org/10.l 164/rccm.201508-1633QC
U.S. EPA (U.S. Environmental Protection Agency). (1986). Air quality criteria for ozone and other
photochemical oxidants; Volume V [EPA Report]. (EPA-600/8-84-020eF). Research Triangle
Park, NC.
U.S. EPA (U.S. Environmental Protection Agency). (1991). Guidelines for developmental toxicity risk
assessment (pp. 1-71). (EPA/600/FR-91/001). Washington, DC: U.S. Environmental Protection
Agency, Risk Assessment Forum, http://cfbub.epa.gov/ncea/cfm/recordisplav.cfm?deid=23162
U.S. EPA (U.S. Environmental Protection Agency). (1996a). Air quality criteria for ozone and related
photochemical oxidants [EPA Report]. (EPA/600/P-93/004AF). Research Triangle Park, NC.
U.S. EPA (U.S. Environmental Protection Agency). (1996b). Guidelines for reproductive toxicity risk
assessment (pp. 1-143). (EPA/630/R-96/009). Washington, DC: U.S. Environmental Protection
Agency, Risk Assessment Forum, https://www.epa.gov/sites/production/files/2014-
11/documents/guidelines repro toxicitv.pdf
U.S. EPA (U.S. Environmental Protection Agency). (1997). National ambient air quality standards for
ozone; final rule. Fed Reg 62: 38856-38896.
U.S. EPA (U.S. Environmental Protection Agency). (1998). Guidelines for neurotoxicity risk
assessment [EPA Report] (pp. 1-89). (EPA/630/R-95/00IF). Washington, DC: U.S. Environmental
Protection Agency, Risk Assessment Forum, http://www.epa.gov/risk/guidelines-neurotoxicitv-
risk-assessment
U.S. EPA (U.S. Environmental Protection Agency). (2005). Guidelines for carcinogen risk assessment
[EPA Report]. (EPA/630/P-03/00IB). Washington, DC: U.S. Environmental Protection Agency,
Risk Assessment Forum, https://www.epa.gov/sites/production/files/2013-
09/documents/cancer guidelines final 3-25-05.pdf
U.S. EPA (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: U.S.
Environmental Protection Agency, Office of Research and Development, National Center for
Environmental Assessment-RTP Office.
http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=149923
U.S. EPA (U.S. Environmental Protection Agency). (2008). Integrated science assessment for sulfur
oxides: Health criteria [EPA Report]. (EPA/600/R-08/047F). Research Triangle Park, NC: U.S.
Environmental Protection Agency, Office of Research and Development, National Center for
Environmental Assessment- RTP. http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=198843
3-222
-------
U.S. EPA (U.S. Environmental Protection Agency). (2013a). Integrated science assessment for ozone
and related photochemical oxidants [EPA Report]. (EPA/600/R-10/076F). Research Triangle Park,
NC: U.S. Environmental Protection Agency, Office of Research and Development, National Center
for Environmental Assessment-RTP Division.
http://cfpub.cpa.gov/ncca/isa/rccordisplav.cfm?dcid=247492
U.S. EPA (U.S. Environmental Protection Agency). (2013b). Toxicological review of
trimethylbenzenes (CASRN 25551-13-7, 95-63-6, 526-73-8, and 108-67-8) in support of summary
information on the Integrated Risk Information System (IRIS): revised external review draft [EPA
Report]. (EPA/635/R13/171a). Washington, D.C.: U.S. Environmental Protection Agency,
National Center for Environmental Assessment.
http://vosemite.epa.gov/sab/SABPRODUCT.NSF/b5d8alce9b07293485257375007012b7/eele28Q
e77586de985257b65005d37e7!OpenDocument
U.S. EPA (U.S. Environmental Protection Agency). (2015). Preamble to the integrated science
assessments [EPA Report]. (EPA/600/R-15/067). Research Triangle Park, NC: U.S. Environmental
Protection Agency, Office of Research and Development, National Center for Environmental
Assessment, RTP Division. https://cfbub.epa.gov/ncea/isa/recordisplav.cfm?deid=310244
U.S. EPA (U.S. Environmental Protection Agency). (2016). Integrated science assessment for oxides
of nitrogen-health criteria (final report) [EPA Report]. (EPA/600/R-15/068). Research Triangle
Park, NC: U.S. Environmental Protection Agency, Office of Research and Development, National
Center for Environmental Assessment.
http://ofrnpub.epa.gov/eims/eimscomm.getfile7p download id=526855
Urman. R; McConnell. R; Islam. T; Avol. EL; Lurmann. FW: Vora. H; Linn. WS; Rappaport. EB;
Gilliland. FD; Gauderman. WJ. (2014). Associations of children's lung function with ambient air
pollution: Joint effects of regional and near-roadway pollutants. Thorax 69: 540-547.
http://dx.doi.Org/10.l 136/thoraxjnl-2012-203159
Vanos. JK; Hebbern. C: Cakmak. S. (2014). Risk assessment for cardiovascular and respiratory
mortality due to air pollution and synoptic meteorology in 10 Canadian cities. Environ Pollut 185:
322-332. http://dx.doi.Org/10.1016/i.envpol.2013.ll.007
Verhein. KC; Mccaw. Z; Gladwell. W; Trivedi. S; Bushel. PR; Kleeberger. SR. (2015). Novel roles
for Notch3 and Notch4 receptors in gene expression and susceptibility to ozone-induced lung
inflammation in mice. Environ Health Perspect 123: 799-805.
http://dx.doi.org/10.1289/ehp.1408852
Verhein. KC; Salituro. FG; Ledeboer. MW; Fryer. AD; Jacobv. DB. (2013). Dual p38/JNK mitogen
activated protein kinase inhibitors prevent ozone-induced airway hyperreactivity in guinea pigs.
PLoS ONE 8: e75351. http://dx.doi.org/10.1371/iournal.pone.0075351
Vestbo. J; Hurd. SS; Agusti. AG; Jones. PW; Vogelmeier. C; Anzueto. A; Barnes. PJ; Fabbri. LM;
Martinez. F.T: Nishimura. M; Stocklev. RA; Sin. D. onD; Rodriguez-Roisin. R. (2013). Global
Strategy for the Diagnosis, Management, and Prevention of Chronic Obstructive Pulmonary
Disease GOLD Executive Summary. Am J Respir Crit Care Med 187: 347-365.
http://dx.doi.Org/10.l 164/rccm.201204-0596PP
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 6: 40. http://dx.doi.Org/10.l 186/1476-069X-6-40
3-223
-------
von Elm. E; Altman. DG; Egger. M; Pocock. SJ; Gotzschc. PC; Vandenbroucke. JP. (2007). The
strengthening the reporting of observational studies in epidemiology (strobe) statement: guidelines
for reporting observational studies [Review]. PLoS Med 4: e296.
http://dx.doi.org/10.1371/iournal.pmed.004Q296
Wang. G; Zhao. J; Jiang. R; Song. W. (2013). Rat lung response to ozone and fine particulate matter
(PM2.5 ) exposures. Environ Toxicol 30: 343-356. http://dx.doi.org/10.1002/tox.21912
Wang. X; Dockerv. DW: Wvpii. D: Gold. PR: Speizer. FE; Ware. JH; Ferris. BG. Jr. (1993).
Pulmonary function growth velocity in children 6 to 18 years of age. Am J Respir Crit Care Med
148: 1502-1508. http://dx.doi.Org/10.1164/airccm/148.6 Pt 1.1502
Ward. WO: Kodavanti. UP. (2015). Pulmonary transcriptional response to ozone in healthy and
cardiovascular compromised rat models. Inhal Toxicol 27: 93-104.
http://dx.doi.org/10.3109/08958378.2014.954173
Ward. WO: Ledbetter. AD: Schladweiler. MC: Kodavanti. UP. (2015). Lung transcriptional profiling:
insights into the mechanisms of ozone-induced pulmonary injury in Wistar Kyoto rats. Inhal
Toxicol 27: 80-92. http://dx.doi.org/10.3109/08958378.2014.954172
Weakley. J: Webber. MP: Ye. F: Zeig-Owens. R: Cohen. HW: Hall. CB: Kelly. K: Prezant. DJ.
(2013). Agreement between obstructive airways disease diagnoses from self-report questionnaires
and medical records. Prev Med 57: 38-42. http://dx.doi.Org/10.1016/i.ypmed.2013.04.001
Weichenthal. S: Pinault. LL: Burnett. RT. (2017). Impact of oxidant gases on the relationship between
outdoor fine particulate air pollution and nonaccidental, cardiovascular, and respiratory mortality.
Sci Rep 7: 16401. http://dx.doi.org/10.1038/s41598-017-16770-v
Weir. CH; Yeatts. KB: Sarnat. JA; Vizuete. W: Salo. PM; Jaramillo. R; Cohn. RD: Chu. H; Zeldin.
DC: London. SJ. (2013). Nitrogen dioxide and allergic sensitization in the 2005-2006 National
Health and Nutrition Examination Survey. Respir Med 107: 1763-1772.
http://dx.doi.Org/10.1016/i.rmed.2013.08.010
Wendt. JK: Svmanski. E: Stock. TH: Chan. W: Du. XL. (2014). Association of short-term increases in
ambient air pollution and timing of initial asthma diagnosis among Medicaid-enrolled children in a
metropolitan area. Environ Res 131: 50-58. http://dx.doi.Org/10.1016/i.envres.2014.02.013
Williams. AS: Mathews. JA: Kasahara. DI: Wurmbrand. AP; Chen. L; Shore. S. (2015). Innate and
ozone-induced airway hyperresponsiveness in obese mice: role of TNF-alpha. Am J Physiol Lung
Cell Mol Physiol 308: L1168-L1177. http://dx.doi.org/10.1152/aiplung.00393.2014
Winquist. A: Kirrane. E; Klein. M; Strickland. M; Darrow. LA: Sarnat SE; Gass. K; Mulholland. J:
Russell. A; Tolbert. P. (2014). Joint effects of ambient air pollutants on pediatric asthma
emergency department visits in Atlanta, 1998-2004. Epidemiology 25: 666-673.
http://dx.doi.org/10.1097/EDE.000000000000Q146
Winquist. A; Klein. M; Tolbert. P; Flanders. WD; Hess. J; Sarnat. SE. (2012). Comparison of
emergency department and hospital admissions data for air pollution time-series studies. Environ
Health 11: 70. http://dx.doi.org/10.1186/1476-069X-11-70
Wolkoff. P; Clausen. PA; Larsen. ST; Hammer. M; Nielsen. GD. (2012). Airway effects of repeated
exposures to ozone-initiated limonene oxidation products as model of indoor air mixtures. Toxicol
Lett 209: 166-172. http://dx.doi.Org/10.1016/i.toxlet.2011.12.008
Wong. EM; Walbv. WF; Wilson. DW; Tablin. F; Schelegle. ES. (2018). Ultrafine particulate matter
combined with ozone exacerbates lung injury in mature adult rats with cardiovascular disease.
Toxicol Sci 163: 140-151. http://dx.doi.org/10.1093/toxsci/kfVO 18
3-224
-------
Wood. AM; Harrison. RM: Semple. S; Avres. JG; Stocklev. RA. (2009). Outdoor air pollution is
associated with disease severity in al-antitrypsin deficiency. Eur Respir J 34: 346-353.
http://dx.doi.org/10.1183/09031936.000879Q8
Xiang. Y; Qin. XQ; Liu. HJ; Tan. YR; Liu. C; Liu. CX. (2012). Identification of transcription factors
regulating CTNNAL1 expression in human bronchial epithelial cells. PLoS ONE 7: e31158.
http://dx.doi.org/10.1371/iournal.pone.0Q31158
Xiao. O; Liu. Y: Mulholland. JA: Russell. AG; Darrow. LA: Tolbert. PE: Strickland. MJ. (2016).
Pediatric emergency department visits and ambient Air pollution in the U.S. State of Georgia: a
case-crossover study. Environ Health 15: 115. http://dx.doi.org/10.1186/sl2940-016-0196-y
Yanagisawa. R; Warabi. E; Inoue. KI; Yanagawa. T; Koike. E; Ichinose. T; Takano. H; Ishii. T.
(2012). Peroxiredoxin I null mice exhibits reduced acute lung inflammation following ozone
exposure. J Biochem 152: 595-601. http://dx.doi.org/10.1093/ib/mvsl 13
Yang. CL; To. T; Fotv. RG; Stieb. DM; Dell. SD. (2011). Verifying a questionnaire diagnosis of
asthma in children using health claims data. BMC Pulm Med 11. http://dx.doi.org/10.1186/1471-
2466-11-52
Ying. Z; Allen. K; Zhong. J; Chen. M; Williams. KM; Wagner. JG; Lewandowski. R; Sun. O;
Raiagopalan. S; Harkema. JR. (2016). Subacute inhalation exposure to ozone induces systemic
inflammation but not insulin resistance in a diabetic mouse model. Inhal Toxicol 28: 155-163.
http://dx.doi.org/10.3109/08958378.2016.11468Q8
Yonchuk. JG; Foley. JP; Bolognese. BJ; Logan. G; Wixted. WE; Kou. JP; Chalupowicz. DG; Feldser.
HG; Sanchez. Y; Nie. H; Callahan. JF; Kerns. JK; Podolin. PL. (2017). Characterization of the
potent, selective Nrf2 activator, 3-(pyridin-3-ylsulfonyl)-5-(trifluoromethyl)-2h-chromen-2-one, in
cellular and in vivo models of pulmonary oxidative stress. J Pharmacol Exp Ther 363: 114-125.
http://dx.doi.org/10.1124/ipet.117.241794
Zellner. LC; Brundage. KM; Hunter. DP; Dev. RD. (2011). Early postnatal ozone exposure alters rat
nodose and jugular sensory neuron development. Toxicol Environ Chem 93: 2055-2071.
http://dx.doi.org/10.1080/02772248.2011.61Q882
Zhang. S; Li. J; Li. Y; Liu. Y; Guo. H; Xu. X. (2017). Nitric oxide synthase activity correlates with
OGG1 in ozone-induced lung injury animal models. Front Physiol 8: 249.
http://dx.doi.org/10.3389/fbhvs.2017.00249
Zhong. J; Allen. K; Rao. X; Ying. Z; Braunstein. Z; Kankanala. SR; Xia. C; Wang. X; Bramble. LA;
Wagner. JG; Lewandowski. R; Sun. Q; Harkema. JR; Raiagopalan. S. (2016). Repeated ozone
exposure exacerbates insulin resistance and activates innate immune response in genetically
susceptible mice. Inhal Toxicol 28: 383-392. http://dx.doi.Org/10.1080/08958378.2016.l 179373
Zhu. Y; Li. J; Wu. Z; Lu. Y; You. H; Li. R; Li. B; Yang. X; Duan. L. (2016). Acute exposure of ozone
induced pulmonary injury and the protective role of vitamin E through the Nrf2 pathway in Balb/c
mice. Toxicology Research 5: 268-277. http://dx.doi.org/10.1039/c5tx00259a
Zu. K; Liu. X; Shi. L; Tao. G; Loftus. CT; Lange. S; Goodman. JE. (2017). Concentration-response of
short-term ozone exposure and hospital admissions for asthma in Texas. Environ Int 104: 139-145.
http://dx.doi.Org/10.1016/i.envint.2017.04.006
Zvchowski. KE; Lucas. SN; Sanchez. B; Herbert. G; Campen. MJ. (2016). Hypoxia-induced
pulmonary arterial hypertension augments lung injury and airway reactivity caused by ozone
exposure. Toxicol Appl Pharmacol 305: 40-45. http://dx.doi.Org/10.1016/i.taap.2016.06.003
3-225
-------
APPENDIX 4 HEALTH EFFECTS-
CARDIOVASCULAR
Summary of Causality Determinations for Short- and Long-Term Ozone
Exposure and Cardiovascular Effects
This Appendix characterizes the scientific evidence that supports causality
determinations for short- and long-term ozone exposure and cardiovascular effects. The types
of studies evaluated within this Appendix are consistent with the overall scope of the ISA as
detailed in the Preface. In assessing the overall evidence, the strengths and limitations of
individual studies were evaluated based on scientific considerations detailed in the Annex for
Appendix 4. More details on the causal framework used to reach these conclusions are included
in the Preamble to the ISA (U.S. EPA. 2015).
Exposure Duration
Causality Determination
Short-term exposure
Suggestive of, but not sufficient to infer, a causal
relationship
Long-term exposure
Suggestive of, but not sufficient to infer, a causal
relationship
4.1 Short-Term Ozone Exposure and Cardiovascular Health
Effects
4.1.1 Introduction, Summary from the 2013 Ozone ISA, and Scope for Current
Review
The 2013 Ozone ISA concluded that "a likely causal relationship exists between short-term
exposure to ozone and cardiovascular effects/' This conclusion was based on multiple lines of evidence,
including animal toxicological studies demonstrating ozone-induced impaired vascular and cardiac
function, as well as changes in time domains of heart rate variability [HRV; U.S. EPA (2013a)l. There
was also evidence from animal toxicological studies for changes in heart rate, although the ISA noted
inconsistencies in that both bradycardia and tachycardia were reported. Controlled human exposure
(CHE) studies also reported cardiovascular effects in response to short-term ozone exposure. More
specifically, CHE studies reported both increases and decreases in measures in the high-frequency domain
of HRV (U.S. EPA. 2013a). Changes in HRV observed in both animal and human studies provided
putative evidence for ozone-induced modulation of the autonomic nervous system potentially through the
4-1
-------
activation of neural reflexes in the lung. In addition, CHE studies from the last review demonstrated some
evidence of ozone-induced effects on blood biomarkers of systemic inflammation and oxidative stress, as
well as changes in biomarkers associated with increased coagulation and/or decreased fibrinolysis (U.S.
EPA. 2013a). Taken together, this experimental evidence was coherent with the consistently positive
associations reported in epidemiologic studies between short-term ozone exposure and cardiovascular
mortality.
Key uncertainties from the last review included a lack of coherence between epidemiologic
mortality and morbidity studies. Although multicity studies and a multicontinent study reported positive
associations between short-term ozone exposure and cardiovascular mortality, with few exceptions, the
findings from epidemiologic studies on short-term ozone exposure and cardiovascular-related morbidity
outcomes, specifically hospital admissions and emergency department (ED) visits, were generally null. In
addition, given that relatively few epidemiologic studies in the 2013 Ozone ISA examined the potential
for copollutant confounding, some uncertainty remains regarding the extent to which ozone is driving the
positive associations reported in studies of mortality. However, the few studies that did examine the
potential for copollutant confounding suggested an independent effect of ozone exposure (U.S. EPA.
2013a). The subsections below provide an evaluation of the most policy-relevant scientific evidence
relating short-term ozone exposure to cardiovascular health effects. These sections focus on studies
published since the 2013 Ozone ISA, and particular emphasis is placed on those studies that address
uncertainties identified in that review. Importantly, when considered as a whole, these newer studies call
into question that a likely causal relationship exists between short-term exposure to ozone and
cardiovascular effects.
4.1.2 Population, Exposure, Comparison, Outcome, and Study Design
(PECOS) Tool
The scope of this section is defined by a scoping tool that generally defines the relevant
Population, Exposure, Comparison, Outcome, and Study Design (PECOS). The PECOS tool defines the
parameters and provides a framework to help identify the relevant evidence in the literature to inform the
ISA. Because the 2013 Ozone ISA concluded a likely to be a causal relationship between short-term
ozone exposure and cardiovascular health effects, the epidemiologic studies evaluated are more limited in
scope and targeted toward study locations, as reflected in the PECOS tool, that are most informative to
address the policy-relevant considerations forming the basis of this section. The studies evaluated and
subsequently discussed within this section were included because they satisfied all of the components of
the following PECOS tool:
Experimental Studies:
• Population: Study populations of any controlled human exposure or animal toxicological study of
mammals at any lifestage
4-2
-------
• Exposure: Short-term (on the order of minutes to weeks) inhalation exposure to relevant ozone
concentrations (i.e., <0.4 ppm for humans; <2 ppm for other mammals)
• Comparison: Human subjects that serve as their own controls with an appropriate washout period
or subjects compared to a reference population exposed to lower levels (when available), or, in
toxicological studies of mammals, an appropriate comparison group that is exposed to a negative
control (e.g., filtered air)
• Outcome: Cardiovascular effects
• Study Design: Controlled human exposure (i.e., chamber) studies; in vivo acute, subacute, or
repeated-dose toxicity studies in mammals, immunotoxicity studies
Epidemiologic Studies:
• Population: Any U.S., Canadian, European, or Australian population1, including populations or
lifestages that might be at increased risk
• Exposure: Short-term ambient concentration of ozone
• Comparison: Per unit increase (in ppb)
• Outcome: Change in risk (incidence/prevalence) of cardiovascular effects
• Study Design: Epidemiologic studies consisting of panel, case-crossover, time-series studies,
case-control studies, and cross-sectional studies with appropriate timing of exposure for the
health endpoint of interest
4.1.3 Biological Plausibility
This subsection describes the biological pathways that potentially underlie cardiovascular health
effects resulting from short-term inhalation exposure to ozone. Figure 4-1 graphically depicts these
proposed pathways as a continuum of pathophysiological responses—connected by arrows—that may
ultimately lead to the apical cardiovascular events associated with short-term exposures to ozone at
concentrations observed in epidemiologic studies (e.g., ED visits and hospital admissions). This
discussion of how short-term exposure to ozone may lead to these cardiovascular events also provides
biological plausibility for the epidemiologic results reported later in this Appendix. In addition, most
studies cited in this subsection are discussed in greater detail throughout this Appendix. Note that the
structure of the biological plausibility sections and the role of biological plausibility in contributing to the
weight-of-evidence analysis used in the 2020 Ozone ISA are discussed in Section IS.4.2.
1 A list of considered studies conducted in other geographic locations is available via the HERO database.
4-3
-------
Short-term
Ozone
Exposure
Exacerbation
of Conduction
Abnormalities
or Arrhythmia
Exacerbation
of Ischemic
Heart Disease/
Potential
Myocardial
infarction or
Stroke
Emergency
Department
Visits/
Hospital
Admissions
: : and/or Mortality
Note: The boxes above represent the effects for which there is experimental or epidemiologic evidence related to ozone exposure,
and the arrows indicate a proposed relationship between those effects. Solid arrows denote evidence of essentiality as provided, for
example, by an inhibitor of the pathway or a genetic knockout model used in an experimental study involving ozone exposure.
Shading around multiple boxes is used to denote a grouping of these effects. Arrows may connect individual boxes, groupings of
boxes, and individual boxes within groupings of boxes. Progression of effects is generally depicted from left to right and color-coded
(gray, exposure; green, initial effect; blue, intermediate effect; orange, effect at the population level or a key clinical effect). Here,
population level effects generally reflect results of epidemiologic studies. When there are gaps in the evidence, there are
complementary gaps in the figure and the accompanying text below. The structure of the biological plausibility sections and the role
of biological plausibility in contributing to the weight-of-evidence analysis used in the 2020 Ozone ISA are discussed in
Section IS.4.2.
Figure 4-1 Potential biological pathways for cardiovascular effects following
short-term exposure to ozone.
When considering the available health evidence, plausible pathways connecting short-term
exposure to ozone with the apical events reported in epidemiologic studies are proposed in Figure 4-1.
The first pathway begins as respiratory tract inflammation leading to systemic inflammation. Hie second
pathway involves activation of sensory nerve pathways in the respiratory tract that lead to modulation of
the autonomic nervous system. Once these pathways are initiated, there is evidence from experimental
and observational studies that short-term exposure to ozone may result in a series of pathophysiological
responses that could lead to cardiovascular events such as ED visits and hospital admissions for ischemic
heart disease (IHD), heart failure (HF), and possible mortality.
There are plausible pathways through which respiratory tract inflammation and oxidative stress
could exacerbate existing IHD and HF and contribute to the development of a myocardial infarction or
stroke. Inflammatory mediators, such as cytokines produced in the respiratory tract (Appendix 3). have
the potential to enter into the circulatory system where they may amplify the initial inflammatory
4-4
-------
response and/or cause distal pathophysiological events that can contribute to overt cardiovascular disease.
Thus, it is important to note that there is evidence from epidemiologic panel studies for an increase in the
cytokines IL-6 and TNF-a (Mirowskv et al.. 2017). as well as for the TNF-a receptor (Li et al.. 2016)
following short-term exposure to ozone. Similarly, there is also evidence for increases in circulating
inflammatory cells (e.g., monocytes, neutrophils) from CHE studies (Stiegel et al.. 2016; Billeretal..
2011). an epidemiologic panel study (Mirowskv et al.. 2017). and animal toxicological studies (Zhong et
al.. 2016; Paffett et al.. 2015). Generally, increases in cytokines like interleukin 6 (IL-6) have been
correlated with increases in liver-derived inflammatory markers such as C-reactive protein (CRP) and the
promotion of hemostasis (i.e., the stopping of blood flow), which is characterized by increases in markers
of coagulation (i.e., clot promoting factors) and/or decreases in markers of fibrinolysis [i.e., clot
dissolving factors; Tanaka et al. (2014)1. With respect to short-term ozone exposure, a CHE (Kahle et al..
2015) and an epidemiologic panel (Mirowskv et al.. 2017) study reported changes in protein levels
associated with fibrinolysis. In addition, CHE (Ariomandi et al.. 2015; Billeretal.. 2011) and an
epidemiologic panel study (Bind et al.. 2012) reported increases in serum levels of CRP following
short-term exposure. In agreement with these studies, an animal toxicological study (Snow et al.. 2018)
reported that in rats fed a normal, coconut oil, or fish oil supplemented diet, short-term exposure to ozone
resulted in increases in platelet circulation. Platelets typically form a plug when the endothelium is
damaged to prevent bleeding. However, platelets can also lead to clot formation when present in the
endothelium in the absence of a wound. Taken together, enhanced inflammation, hemostasis, and
increases in the circulation of platelets likely enhances the potential for thrombosis, which could
exacerbate existing IHD and HF.
In addition to affecting hemostasis, systemic inflammation and/or oxidative stress may result in
impaired vascular function. Impaired vascular function stems from impaired functioning of the
endothelium, which maintains the normal balance of mediators that promote vasorelaxation (e.g., nitric
oxide) and vasoconstriction (e.g., endothelin-1). In endothelial dysfunction, the balance is tipped towards
greater production of vasoconstrictors causing increased vascular resistance which could lead to rupture
of existing plaques (Halvorsen et al.. 2008). Dislodged plaques might then obstruct blood flow to the
heart or stimulate intra-vascular clotting (Karolv et al.. 2007). both of which could result in acute
myocardial ischemia, and set the stage for HF. If the dislodged plaque obstructs blood flow to the brain,
there is potential for stroke. Impaired vascular function has been reported following short-term ozone
exposure in epidemiologic panel studies (Mirowskv et al.. 2017; Lanzingcr et al.. 2014) and animal
toxicological studies (Paffett et al.. 2015; Robertson et al.. 2013; Chuang et al.. 2009). With respect to
impaired vascular function, Paffett et al. (2015) reported that following in vivo ozone exposure,
cotreatment of rat coronary artery segments with acetylcholine and an NAPDH oxidase inhibitor
improved the vasodilatory response compared to acetylcholine and control, indicating that one likely
mechanism of impaired vasodilation was through oxidative stress-related pathways. These results are in
agreement with animal toxicological studies reporting increases in markers of oxidative stress following
short-term exposure to ozone (Kumarathasan et al.. 2015; Martinez-Campos et al.. 2012). Moreover,
Robertson et al. (2013) used knockout mice to determine that the presence of CD36 was required for
4-5
-------
ozone-induced impaired vascular function. We also note that clinical indicators of potential ischemia
(e.g., ST segment depression on an electrocardiogram) following short-term exposure to ozone have been
shown in an animal toxicological study (Farrai et al.. 2012). and increased odds of STEMI (ST-elevation
myocardial infarction) have been reported in an epidemiologic panel study (Evans et al.. 2016V
Impaired vascular function can also lead to increases in blood pressure (BP) through
vasoconstriction. Increases in BP may then exacerbate IHD or HF through altered hemostasis and/or
impaired vascular function. For example, in patients with high blood pressure, changes in arterial shear
stress due to changes in blood flow (i.e., laminar vs. turbulent) are associated with impaired vascular
function (Khdcr et al.. 1998). which as noted above, could lead to a worsening of IHD or HF (Figure 4-IV
Thus, it is notable that following short-term ozone exposure, there is evidence for increases in BP from
epidemiologic panel (Cakmak et al.. 2011) and animal toxicological studies (Farrai et al.. 2016;
Tankerslev et al.. 2013). Taken together, there are plausible pathways through which respiratory tract
inflammation could exacerbate existing IHD and HF, contribute to the development of a myocardial
infarction or stroke, and lead to ED visits and hospital admissions.
There is also evidence that exposure to ozone could lead to these outcomes potentially through
activation of sensory nerves in the respiratory tract. Once activated, these nerves send sensory input to
autonomic centers in the brain, which in turn relay reflex motor outputs that modulate autonomic tone
(e.g., increased sympathetic tone) to the heart and vasculature. Shifts toward increased sympathetic
nervous system tone may result in increases in BP and decreases in vascular function, which as mentioned
above, could exacerbate IHD and/or HF. It is therefore important to note the evidence from CHE
(Ariomandi et al.. 2015). epidemiologic panel (Bartell et al.. 2013; Cakmak et al.. 2011) and animal
toxicological (Wagner et al.. 2014; Mclntosh-Kastrinskv et al.. 2013; Wang et al.. 2013; Farrai et al..
2012) studies of autonomic nervous system modulation—including some evidence from a few studies for
a shift toward increased sympathetic tone (as evidenced by changes in HRV)—following short-term
ozone exposure. Similarly, there is evidence from epidemiologic panel (Cakmak et al.. 2014; Bartell et
al.. 2013; Sarnat et al.. 2006) and animal toxicological (Farrai et al.. 2016; Farrai et al.. 2012) studies that
short-term exposure to ozone can result in conduction abnormalities or arrhythmia. Conduction
abnormalities or arrhythmia could then potentially exacerbate IHD and/or HF. Taken together, there are
multiple potential pathways by which activation of sensory nerves in the respiratory tract may lead to
worsening of IHD orHF.
Overall, the evidence suggests plausible pathways through which short-term exposure to ozone
may worsen IHD or HF as well as contribute to the development of MI or stroke (Figure 4-IV These
proposed pathways also provide some biological plausibility for ED visits and hospital admissions
following short-term ozone exposure. However, considerable uncertainties remain as the evidence
supporting some of the individual events in these pathways is limited and not supported by CHE studies.
This information will be used to inform causality, which is discussed later in the Appendix
(Section 4.1.16).
4-6
-------
4.1.4 Heart Failure, Impaired Heart Function, and Associated Cardiovascular
Effects
Heart failure refers to a set of conditions in which the heart's pumping action is compromised. In
congestive heart failure (CHF), the flow of blood from the heart slows and fails to meet the body's
oxygen demand. Edema from heart failure frequently occurs from increased sodium reabsorption
resulting in an increase in blood volume (hypervolemia) and fluid retention, which often causes swelling
in the lungs or other tissues (typically in the legs and ankles). The effect of short-term ozone exposure on
people with CHF, which is a chronic condition, is generally evaluated using ICD codes recorded when a
patient is admitted or discharged from the hospital or ED. The relevant diagnostic codes for heart failure
are ICD9 428 and ICD10 150. These codes encompass left, systolic, diastolic, and combined heart failure.
In experimental studies, indicators of heart failure include decreased contractility and/or relaxation in
response to pharmacological challenge, reduced ejection fraction (i.e., the percentage of blood pumped
from the ventricles during each contraction), reduced stroke volume (i.e., the volume of blood pumped
per contraction) and reduced cardiac output (stroke volume multiplied by heart rate), as well as decreases
in left ventricular developed pressure (LVDP). Of note, the most prevalent form of heart failure is
diastolic heart failure (i.e., heart failure where cardiac filling is impaired, but ejection fraction is not).
4.1.4.1 Epidemiologic Studies of Emergency Department Visits and Hospital
Admissions
The 2013 Ozone ISA reported the results of several studies in the U.S., Canada, and the U.K., all
of which observed null results for the association between CHF-related emergency department or hospital
visits and ozone exposure averaged over either 8 or 24 hours. A few additional studies have been
conducted since the 2013 Ozone ISA with mixed results (Table 4-3). Specifically:
• While studies conducted in the U.K. and U.S. did not observe positive associations between CHF
(alone or combined with hypertensive heart disease) and 8-hour max ozone concentrations
(Rodopoulou et al.. 2015; Miloievic et al.. 2014). a study in St. Louis, MO reported a 5% increase
in ED visits (RR: 1.05; 95% CI: 1.01, 1.09)1 and hospital admissions (95% CI: 1.02, 1.09) for
CHF (Winquist et al.. 2012) associated with 8-hour max ozone. Similarly, an additional study in
St. Louis observed a 4% (RR: 1.04; 95% CI: 0.99, 1.10) increase in ED visits for CHF, which
increased to 6% (RR: 1.06; 95% CI: 1.00, 1.12) when CO was included in the model (Sarnat et
al.. 2015V Copollutant models with either PM2 5 or NO2 did not change the predicted risk for
ozone.
• Studies evaluating the role of lifestage in ozone's effects on heart failure reported no notable
differences for older adults (>65 or 70 years) compared with other adult age groups [19-64 or
<70 years; Miloievic et al. (2014); Winquist et al. (2012)1.
1 All epidemiologic results are standardized to a 15 ppb increase in 24-hour avg, 20 ppb increase in 8-hour daily
max, 25-ppb increase in 1-hour daily max ozone concentrations, or a 10-ppb increase in seasonal/annual ozone
concentrations to facilitate comparability across studies.
4-7
-------
4.1.4.2
Controlled Human Exposure Studies
In the 2013 Ozone ISA, there were no CHE studies examining the relationship between
short-term ozone exposure and impaired cardiac function. In a recent study in healthy subjects with or
without deletion of GSTM1, Frampton et al. (2015) reported that short-term exposure (3 hours) to ozone
(0.1, 0.2 ppm) did not result in statistically significant changes in stroke volume or left ventricular
ejection time relative to FA. Results were independent of the GSTM1 phenotype. Additional information
on this study can be found in Table 4-4.
4.1.4.3 Animal Toxicological Studies
In the 2013 Ozone ISA, an animal toxicological study demonstrated that ozone exposure resulted
in decreased LVDP, rate of change of pressure development, and rate of change of pressure decay (Pcrcpu
et al.. 2010). Another study demonstrated that ozone exposure resulted in an increase in left ventricular
chamber dimensions at end-diastole in young and old mice, as well as a decrease in left ventricular
posterior wall thickness at end-systole in older mice (Tankerslev et al.. 2010). Moreover, these authors
also reported a decrease in fractional shortening—an indicator of impaired cardiac contraction
characterized by the percentage change in left ventricular diameter from end-diastole to end-systole
following short-term ozone.
Since the publication of the 2013 Ozone ISA, there is additional evidence from animal
toxicological studies that short-term exposure (3-4 hours, some studies with multiple day exposures) to
ozone can result in impaired cardiac function. With respect to this evidence, we note the following key
points:
• Tankerslev et al. (2013) exposed wild-type mice to ozone (0.5 or 0.8 ppm) and then FA, and
demonstrated that this short-term exposure resulted in a decrease in LVDP that was not
statistically significant, as well as a decrease in left ventricular stroke volume (p < 0.05), and an
increase in right ventricular pressure (p < 0.05) relative to an exposure of FA followed by a
second FA exposure. Moreover, an approximately 33% decrease in left ventricular cardiac output
relative to FA exposure was reported (p < 0.05). Tankerslev et al. (2013) also demonstrated that
short-term exposure to ozone resulted in a significant decrease in left ventricular minimum and
maximum volumes, (p < 0.05) as well as an increase in total peripheral resistance. Finally, they
also demonstrated through the use of knockout mice that many of these effects may be mediated
by the atrial natriuretic peptide gene.
• In mice, Kurhanewicz et al. (2014) reported a decrease in LVDP and other measures of
contractility 24 hours post-exposure (0.3 ppm) relative to FA. However, the authors did not
denote these results as having statistical significance relative to FA in their figure.
• Mclntosh-Kastrinskv et al. (2013) reported that short-term ozone exposure (0.245 ppm) reduced
diastolic function (i.e., cardiac filling) as indicated by impaired cardiac relaxation rate
(dP!dtminimum) relative to FA exposure in isolated, perfused murine hearts (p < 0.05).
4-8
-------
• Wang et al. (2013) reported at least some evidence of dissolved myofilaments (a potential
indicator of cardiac damage) in right ventricles by microscopy following short-term ozone
(0.8 ppm) exposure.
Although results from the studies mentioned above demonstrated an effect of short-term ozone
exposure on changes in heart function, other results from these studies showed no effect. That is:
• Mclntosh-Kastrinskv et al. (2013) reported that short-term ozone exposure (0.245 ppm) did not
result in changes in LVDP, dP/dtmaKimum, or coronary flow rate relative to FA exposure in isolated,
perfused murine hearts prior to ischemia. Moreover, following ischemia/reperfusion there was no
difference between ozone and FA exposure with respect to time to ischemic contracture, recovery
of LVDP, or ischemia-induced infarct size. Similarly, Kurhanewicz et al. (2014) reported that in
mice, there were no differences in time to ischemic contracture, or coronary flow rate prior or
after ischemia with ozone exposure. They also reported no differences in the recovery of left
ventricular developed pressure or pressure development over time post-ischemia.
• Tankerslev et al. (2013) did not report changes in left ventricular pressure over time (dP/dtmmimum
or t/PMmaximum) following short-term ozone (0.5,0.8 ppm) exposure relative to FA in wild-type
mice. Similarly, Zvchowski et al. (2016) reported that short-term ozone (1.0 ppm) exposure did
not result in appreciable right ventricular hypertrophy in mice kept in normal oxygen conditions,
nor did ozone exacerbate right ventricular hypertrophy in hypoxia-induced mice.
• Ramot et al. (2015) reported no effect of ozone effect on heart pathology in rats (0.25, 0.5,
1.0 ppm).
Although not every heart function-related endpoint tested in these studies was found to be
sensitive to ozone exposure, most of the studies presented above report some indicator of impaired
cardiac function following short-term ozone exposure (Table 4-5). In addition, there is evidence
suggesting that the atrial natriuretic peptide gene may mediate some of these ozone-induced
cardiovascular effects Tankerslev et al. (2013).
4.1.5 Ischemic Heart Disease and Associated Cardiovascular Effects
IHD is a chronic condition characterized by atherosclerosis and reduced blood flow to the heart.
Myocardial infarction (MI), more commonly known as a heart attack, occurs when heart tissue death
occurs secondary to prolonged ischemia due to occlusion of the coronary artery. The effect of short-term
ozone exposure on acute MI, complications from recent MI, and other acute or chronic IHD are generally
evaluated using ICD codes recorded when a patient is admitted or discharged from the hospital or
emergency department (ICD9: 410-414 or ICD10: 120-125). In experimental or epidemiologic panel
studies, indicators of MI include ST-segment depression as measured by an electrocardiograph (ECG).
The ST segment of an electrocardiogram recorded by surface electrodes corresponds to the electrical
activity of the heart registered between ventricular depolarization and repolarization, and is normally
isoelectric.
4-9
-------
4.1.5.1 Epidemiologic Studies of Emergency Department and Hospital Admission
Studies
In the 2013 Ozone ISA, all of the studies involving U.S. or European populations reported null
effect estimates for IHD and MI, but mixed findings for angina. A multicity study in Europe reported no
association for MI, but the same study observed a positive association (OR: 1.19; 95% CI: 1.05, 1.35) for
angina pectoris during the warm season [April-September; von Klot et al. (2005)1. In contrast, a study in
London, reported null results for angina [OR: 0.98; 95% CI: 0.94, 1.03; Poloniecki et al. (1997)1.
• Recent studies from Europe, Canada, and the U.S. published since the 2013 ISA, several of which
analyzed a large number of MI, IHD or angina events per day in multiple cities, confirm the
pattern indicated by the earlier studies. These studies also consistently reported null or small
positive effect estimates (i.e., OR < 1.02) in analyses of MI, including ST-elevation myocardial
infarction (STEMI) and non-ST-elevation myocardial infarction (NSTEMI; Table 4-6:
Figure 4-2). A study of five urban areas in Tuscany, Italy reported a 5% increase in incident MI
associated with an increase in ozone concentrations during the warm season using a 0-1-day
distributed lag [95% CI: -4, 16%; Nuvolone et al. (2013)1.
• A study in Iceland that analyzed associations with air pollutants, including ozone, and dispensing
glyceryl trinitrate against angina pectoris did not observe increases in odds ratios in single
pollutant models (Finnbiornsdottir et al.. 2013).
4-10
-------
Reference
Poloniecki et al. 1997
IVidale et al. 2017
lBhaskanm et al. 2011
IFinnbjomsdottir etal. 2013
lClaeys etal. 2015
IRasche et al. 2018
IWang ct al. 2015
IMilojevic etal. 2014
IButland et al. 2016
IBard etal. 2014
ISamat et al. 2015
INuvoIone et al. 2013
Von Klot et al. 2005
Location Notes
London, England
Como, Italy
UK MINAP
Reykjavik, Iceland
Belgium
Jena, Germany
Canada
with PCI
Calgary
Edmonton
England, Wales MINAP
HES
England, Wales MINAP 0-2
Strasbourg, France
St. Louis, MO
Italy. 6 cities
Europe, 5 cities
Lag
0
0
0-3
0
0-5
2 0.29 {0.11, 0.86)
0
0
0
0
0
0
0-4
0-1
0-2
0-1
1
I
~r*~
-T*-
t
¦+"
Typc
Angina pectoris
AMI
MI
Angina pectoris
STEMT
STFMI
Incident MI
STEMI
NSTEMI
Incident MI
STEMI
NSTEMI
MI
STEMI
NSTF.MI
IHD
MI
MI
STEMI
NSTF.MI
MI
IHD
Incident MI
¦ Angina pectoris
0.8 0.85
0,9 0.95 1 1.05 1.1
Odds Ratio (95% CI)
1.2
AMI = acute myocardial infarction; HES = Hospital Episode Statistics; IHD = ischemic heart disease; Ml = myocardial infarction;
MINAP = Myocardial Ischemia National Audit Project; NSTEMI = non-ST-elevation myocardial infarction; PCI = percutaneous
coronary intervention; STEMI = ST-elevation myocardial infarction.
Note: tStudies published since the 2013 Ozone ISA. Studies are listed from the top in order of increasing mean or median ozone
concentration reported in the publication. Associations are presented per 25 ppb increase in pollutant concentration for 1 -hour max
averaging times, 20 ppb increase for 8-hour avg times, and 15 ppb increase for 24-hour avg times. Symbols represent point
estimates, circles, triangles and diamonds represent the entire year, warm season and cold season, respectively; horizontal lines
represent 95% confidence intervals for ozone. Black text and symbols represent evidence included in the 2013 Ozone ISA; red text
and symbols represent recent evidence not considered in previous ISAs or AQCDs.
Figure 4-2 Associations between short-term exposure to ozone and
ischemic heart disease (IHD)-related emergency department visits
and hospital admissions.
4.1.5.2 Epidemiologic Panel Studies
The 2013 Ozone ISA reported inconsistent results with respect to an association between
short-term ozone exposure and MI. One study reported elevated risks for recurrent Mis (Henrotin et al..
2010). while another observed no associations between short-term ozone exposure and ST-segment
depression in elderly men with a history of coronary artery disease (Delfino et al.. 2011).
4-11
-------
Since the 2013 Ozone ISA, one additional study examined the potential for STEMI following
short-term ozone exposure (Table 4-7). This study provides evidence of increased incidence of STEMI
resulting from increased concentrations of short-term ozone exposure. Specifically, in a cohort of
362 subjects in Rochester, NY with acute coronary syndrome identified as STEMI, NSTEMI, or unstable
angina, Evans et al. (2016) reported a 35% (OR = 1.35; 95% CI: 1.00, 1.84) increase in the odds of
STEMI for exposure over the previous hour; these associations were attenuated but remained positive at
12, 24, 48, and 72 hours prior to the event. Larger increases in the odds of STEMI were observed for
increases in ozone concentrations measured over the previous hour for patients with previous MI
(OR = 2.06; 95% CI: 0.96, 4.44), CVD (OR = 1.98; 95% CI: 1.02, 3.81), and hypertension (OR = 1.44;
95% CI: 1.00, 2.24). When evaluated by season, the authors observed elevated odds of STEMI in the
cooler months (November to April), and decreased odds in the warmer months (May to October) for
ozone exposure estimated over the 24, 48, and 72 hours preceding the event.
4.1.5.3 Controlled Human Exposure Studies
In the 2013 Ozone ISA, there were no controlled human exposure studies examining indicators of
IHD. That said, a study from the previous AQCD indicated that exposure to ozone did not result in
ST-segment depression (Gong et al.. 1998). Recently, Rich et al. (2018) reported no appreciable change in
the ST segment as a result of ozone (0.07, 0.12 ppm) exposure (3 hours) in older adults (Table 4-8).
4.1.5.4 Animal Toxicological Studies
The 2013 Ozone ISA did not have any animal toxicological studies examining the relationship
between short-term exposure to ozone and the ST segment (U.S. EPA. 2013a). Since the publication of
that document, Farrai et al. (2012) reported that in spontaneously hypertensive (SH) rats, short-term
(4 hours) exposure to 0.8 but not 0.2 ppm ozone resulted in ST-segment depression during exposure when
compared to pre-exposure baseline conditions (p < 0.05). However, there were no statistically significant
post-exposure effects when compared to baseline. Thus, evidence from animal toxicological studies that
short-term exposure to ozone can result in potential indicators of ischemic heart disease is limited. Details
from this study can be found in Table 4-9.
4.1.6 Endothelial Dysfunction
Endothelial dysfunction is the physiological impairment of the inner lining of blood vessels that is
characterized by an imbalance between vasodilators such as nitric oxide and vasoconstrictors such as
endothelin-1 (ET-1) that favors vasoconstrictors. Endothelial dysfunction is typically measured by
flow-mediated dilation percentage (FMD%). It is a noninvasive technique involving measurement of the
4-12
-------
percentage change in brachial artery diameter (BAD) after reactive hyperemia (increased blood flow
following removal of an artery occluding blood pressure cuff) or pharmacological challenge. In addition
to measuring FMD or BAD, experimental studies also examine arterial stiffness as indicated by pulse
wave velocity and levels of biomarkers such as ET-1.
4.1.6.1 Epidemiologic Panel Studies
In the 2013 Ozone ISA, endothelial biomarkers indicated the potential for cardiovascular disease
and injury. However, no epidemiologic studies had evaluated short-term ozone exposure and endothelial
function. Recent panel studies have specifically evaluated short-term ozone exposure and the effects on
endothelial function (e.g., FMD, BAD) and biomarkers. Considering a number of endpoints in
epidemiologic panel studies, there is some evidence from a small number of these studies of endothelial
dysfunction following short-term ozone exposures (Table 4-10). However, this could be due to
differences in study size, demographics, exposure, time lags, and the health endpoints examined across
studies. With respect to this evidence, we note:
• Lanzingeretal. (2014) reported FMD decreases in 22 individuals between the ages of
48-78 years with type 2 diabetes at lag 0 (-29.2; 95% CI: -52.6, -5.80) and lag 1 (-27.0; 95%
CI: -54.0, -0.08). However, Mirowskv et al. (2017) saw no change in FMD at any lag in 13 men
with a previous diagnosis of coronary artery disease.
• In one study of 64 patients with type 2 diabetes, null associations were observed between
short-term ozone effects and BAD rZanobetti et al. (2014); qualitative results only]. However,
Mirowskv et al. (2017) observed opposing effects in BAD in a small cohort of 13 men with
coronary artery disease. They reported a decrease in BAD at lag 2 (-2.68; 95% CI: -5.36, 0.10)
followed by an increase at lag 4 (3.75; 95% CI: 1.29, 6.32).
• Mirowskv et al. (2017) also evaluated several markers of endothelial dysfunction: I-CAM,
V-CAM, LAEI, SAEI, and observed a decrease in V-CAM of 10.3% (95% CI: -18.43, -1.29) at
a lag 2. Conversely, Bind et al. (2012) used the Normative Aging Study Cohort with 704 men in
the greater Boston area who were free from chronic medical conditions and observed no change
in either V-CAM or I-CAM (qualitative results only).
4.1.6.2 Controlled Human Exposure Studies
A publication available at the time of the last review, Brook et al. (2009) found no effect of ozone
exposure alone on clinical indicators of endothelial dysfunction, such as FMD. More recent CHE studies
(1-3 hours in duration) have also shown no evidence of an ozone effect. Specifically:
• Barath et al. (2013) reported no decreases in measures of blood flow relative to FA in response to
acetylcholine, sodium nitroprusside, verapamil, or bradykinin following ozone exposure
(0.3 ppm) in healthy young men. In fact, the study authors reported an increase in blood flow with
ozone relative to FA exposure following acetylcholine or nitroprusside challenge.
4-13
-------
• Frampton et al. (2015) and Rich et al. (2018) also reported no changes in measures of vascular
function (via peripheral arterial tonometry, FMD) due to short-term exposure to ozone (0.07, 0.1,
0.12, 0.2 ppm) in healthy subjects with or without a GSTM1 deletion or in older adults,
respectively.
Thus, there is no evidence from CHE studies that short-term ozone exposure results in
vasoconstriction. Additional information on these studies can be found in Table 4-11.
4.1.6.3 Animal Toxicological Studies
In the 2013 Ozone ISA, the Chuang et al. (2009) study reported that short-term ozone exposure
inhibited acetylcholine-induced vasorelaxation. In addition, a few studies demonstrated that short-term
exposure (4 hours, some studies with multiple day exposures) to ozone was associated with an increase in
the vasoconstrictor ET-1. Since the publication of the 2013 ISA, additional studies have reported similar
effects following short-term exposure to ozone (Table 4-12). With respect to this evidence, we note the
following key points:
• In rats, Paffett et al. (2015) demonstrated that short-term ozone (1.0 ppm) exposure resulted in
increased vasoconstriction and reduced vasodilation relative to control animals following ex-vivo
treatment of coronary artery segments with serotonin and acetylcholine respectively (p < 0.05).
The authors demonstrated that cotreatment of coronary artery segments with acetylcholine and an
NAPDH oxidase inhibitor improved the vasodilatory response, suggesting one likely mechanism
of impaired vasodilation was through oxidative-stress-related pathways (Paffett et al.. 2015).
• Impaired vasodilation relative to control animals in response to acetylcholine was also reported in
wild-type, but not CD 36 null, mouse abdominal and thoracic aortic segments following ozone
(1.0 ppm) exposure [Robertson et al. (2013);p < 0.05], These authors also provided some
evidence that decreased vasodilation in wild-type mice was due to impaired endothelial release of
NO.
• In a dietary intervention study, relative to control rats, Snow et al. (2018) reported significant
phenylephrine-induced vasoconstriction in aortic rings from ozone-exposed rats fed either a
normal diet (p < 0.05) or a diet supplemented with coconut oil, or olive oil (p < 0.05), but not in
rats supplemented with fish oil prior to ozone exposure. However, neither ozone nor diet resulted
in an impaired vasodilation response to acetylcholine or sodium nitroprusside (Snow et al.. 2018).
With respect to blood markers of vasodilation, vasoconstriction, or vascular damage:
• Paffett et al. (2015) demonstrated decreased serum levels of NO2/NO3 in ozone
(1.0 ppm)-exposed animals relative to FA (p < 0.05).
• In rats, Kumarathasan et al. (2015) found an increase in plasma ET-1 and BET-1 (i.e., the
precursor to ET-1) following exposure to 0.8, but not 0.4 ppm ozone immediately after and
24 hours post-exposure (p < 0.05). Similarly, Thomson et al. (2013) reported increased ET-1
mRNA expression in rat heart tissue following short-term ozone (0.4, 0.8 ppm) exposure relative
to control exposure.
• In contrast to these results, Wang et al. (2013) reported no difference in plasma levels of ET-1 or
VEGF when comparing rats exposed to ozone (0.8 ppm) to animals exposed to FA.
4-14
-------
Overall, the animal toxicological evidence is generally consistent. Some studies demonstrated
increased vasoconstriction while others showed impaired vasodilation. This evidence is further supported
by studies reporting increased blood markers associated with vasoconstriction and/or endothelial injury.
4.1.7 Cardiac Depolarization, Repolarization, Arrhythmia, and Arrest
In epidemiologic studies, the effect of short-term ozone exposure on arrhythmia is generally
evaluated using ICD codes for ED visits, hospital admissions, and out-of-hospital cardiac arrests
(OHCA). In addition, there is a body of epidemiologic studies that examine arrhythmias recorded on
implantable cardio-defibrillators.
Experimental and epidemiologic panel studies typically use surface ECGs to measure electrical
activity in the heart resulting from depolarization and repolarization of the atria and ventricles. The
P wave of the ECG represents atrial depolarization, while the QRS represents ventricular depolarization
and the T wave, ventricular repolarization. Because the ventricles account for the largest proportion of
heart mass overall and thus are the primary determinants of the electrical activity recorded in the ECG,
ECG changes indicating abnormal electrical activity in the ventricles are of greatest concern. Changes in
QT, ST, as well as changes in T-wave shape, duration or amplitude may indicate abnormal impulse
propagation in the ventricles. Cardiac arrhythmias can vary in severity from the benign to the potentially
lethal, such as in cardiac arrest when an electrical disturbance disrupts the heart's pumping action causing
loss of heart function.
4.1.7.1 Epidemiologic Studies of Emergency Department Visits and Hospital
Admissions
Few studies evaluating short-term ozone exposure and cardiac arrest, arrhythmias, or
dysrhythmias were discussed in the 2013 Ozone ISA. Two studies in the U.S. and Australia reported no
positive associations for out-of-hospital cardiac arrest (Dcnnckamp et al.. 2010; Silverman et al.. 2010). A
modest elevation in risk of arrhythmia was associated with 8-hour max ozone concentrations during the
warm season in Helsinki, Finland [OR: 1.04; 95% CI: 0.8, 1.35; Halonen et al. (2009)1. but null results
were reported in Atlanta, London, and a multicity study in Canada (Sticb et al.. 2009; Peel et al.. 2007;
Poloniecki et al.. 1997). Several recent studies in the U.S., Europe, and Australia have analyzed the
association between ozone concentration and cardiac arrest, arrhythmias, or dysrhythmias. Findings from
these studies indicate increases in out-of-hospital cardiac arrests associated with 8-hour max or 24-hour
avg increases in ozone concentrations; however, null associations are reported for other endpoints
(e.g., dysrhythmia, arrhythmia, or atrial fibrillation; Table 4-13; Figure 4-3). Specifically:
• In Europe, odds ratios for out-of-hospital cardiac arrest associated with 24-hour avg ozone
concentration were 1.18 (95% CI: 1.00, 1.41) in Helsinki, Finland, 1.13 (95% CI: 1.03, 1.24) in
4-15
-------
Gironde Department in France, and 1.16 (95% CI: 1.03, 1.29) in Stockholm County, Sweden
(Pradcau et al.. 2015; Raza etal.. 2014; Rosenthal et al.. 2013). These associations were observed
in models using the average of the previous 3 days, a 1 day constrained lag, or the concentration
on the same day as the hospitalization, respectively. Rosenthal et al. (2013) presented the results
of models of cardiac arrest risk stratified by the underlying cause of the event, either acute
myocardial infarction or other cardiac causes. The model results indicated that the elevated risk
for cardiac arrest was primarily due to causes other than acute myocardial infarction. The odds
ratios increased and remained statistically significant in copollutant models with PM2 5, PM10,
other particulate size classes, NO, NO2, SO2, or CO. Raza et al. (2014) analyzed a high number of
events per day and confirmed the independent effect of ozone on cardiac arrest in a copollutant
model with NO2. The authors also observed an exposure response pattern in a categorical analysis
using 10.2 ppb increments from 11.7 ppb to >66 ppb.
In contrast to the associations observed in Finland, France, and Sweden, a study in Perth,
Australia analyzed hourly lags and cumulative hourly lags over a 48-hour period, and observed
no association with out-of-hospital cardiac arrest and 1-hour max ozone concentrations (Strancy
etal.. 2014).
For a study in Houston, TX, an OR for cardiac arrest of 1.04 (95% CI: 1.00, 1.07) was associated
with an increase in 8-hour max ozone concentration, and the risk was higher during the warm
season (Ensor et al.. 2013).
A few other studies assessed whether risk ratios varied by season, but no clear trend was
apparent. In contrast to the findings by Ensor et al. (2013). no notable seasonal differences in the
risk of either OHCA or onset of atrial fibrillation were observed by other studies (Pradcau et al..
2015; Sade et al.. 2015; Rosenthal et al.. 2013).
A number of studies evaluating the onset of dysrhythmia or atrial fibrillation, identified via ED or
hospital admission records, did not observe increased risk ratios associated with ozone
concentrations using single- or multiple-day lags (Sade et al.. 2015; Sarnat et al.. 2015; Miloievic
et al.. 2014; Winquist et al.. 2012). However, a study in Little Rock, AR observed a moderately
increased risk ratio for conduction disorders and dysrhythmias associated with an 8-hour max
ozone concentration using a 1-day lag [OR: 1.05; 95% CI: 0.99, 1.12; Rodopoulou et al. (2015)1.
Risk comparisons of OHCA by sex did not consistently indicate greater susceptibility for either
men or women. Rosenthal et al. (2013) found the risk of out-of-hospital cardiac arrest from
causes other than acute MI to be larger in women (OR = 1.76; 95% CI: 1.33-2.33, lag 1 day) than
in men (/?-value for difference by sex = 0.003). Another study reported increased odds ratios for
OHCA with presumptive cardiac etiology for both women and men, although the odds ratios
were higher among women (Pradeau et al.. 2015). An opposite pattern was observed by Ensor et
al. (2013) in their Texas study; the increased RR of OHCA associated with the avg 1-3-hour
lagged ozone concentration was statistically significant for men and higher than the RR for
women.
The risk of OHCA associated with 24-hour avg ozone concentrations were reported to be higher
by two studies for individuals older than 64 years (Pradeau et al.. 2015; Ensor et al.. 2013).
4-16
-------
Reference
Location
l.ag
Dennekamp et al. 2010
Australia
0-1
tStranev et al. 2014
Perth, Australia
0
tRosenthal et al. 2013
Helsinki, Finland
0-3
lEnsor ct al. 2013
Houston, TX
0-1
IPradcau ct al. 2015
Girondc, France
1
!Ra?a etal. 2014
Stockholm, Sweden
0
IMilojcvic el al. 2014
UK
0-4
iSarnat etal. 2015
St. Louis, MO
0-2
IWinquist et al. 2012
St. Louis, MO
0-4
IRodopoulou etal. 2015
Little Rock, AR
1
ISade et al. 2015
Negev, Israel
0
-fr-
Type
OHCA
OHCA
01ICA
OHCA
OHCA
CA
Arrythmias
AF
AVCD
I )vM'hvthm]a
I )ysrhvlhmia - ED
Dy srhythmia - I1A
Dysrhythmias
AF
0.9 1 11
Risk Ratio (95% CI)
AF = atrial fibrillation; AVCD = atrioventricular conduction disorders; CA = cardiac arrest; ED = emergency department;
HA = hospital admissions; OHCA = out of hospital cardiac arrest.
Note: fStudies published since the 2013 Ozone ISA. Studies are listed from the top in order of increasing mean or median ozone
concentration reported in the publication. Associations are presented per 25-ppb increase in pollutant concentration for 1 -hour max
avg times, 20-ppb increase for 8-hour avg times, and 15-ppb increase for 24-hour avg times. Circles represent point estimates;
horizontal lines represent 95% confidence intervals for ozone. Black text and circles represent evidence included in the 2013 Ozone
ISA; red text and circles represent recent evidence not considered in previous ISAs or AQCDs.
Figure 4-3 Associations between short-term exposure to ozone and
emergency department (ED) visits and hospital admissions
related to cardiac arrest, arrhythmias, and dysrhythmias.
4.1.7.2 Epidemiologic Panel Studies
The 2013 Ozone ISA stated that many studies reported positive associations for
arrhythmia-associated endpoints, yet collectively, results were inconsistent. In a population of subjects
with an implantable cardioverter defibrillator [ICD; Mctzgcr et al. (2007)1 observed no evidence of an
association for tachyarrhythmic events with an increase in ozone concentrations. In contrast, in a study of
nonsmoking adults, increased odds were observed for supraventricular ectopy (Sarnat et al.. 2006). In the
4-17
-------
few studies published since the 2013 Ozone ISA, none have the same endpoints so the results remain
inconsistent (Table 4-14). Specifically:
• In a cohort of 50 elderly nonsmokers with previous coronary artery disease, Bartell et al. (2013)
observed relatively strong associations between short-term ozone exposure and daily ventricular
tachycardia (VT) events, specifically for the 24-hour avg lag period (RR = 1.50; 95% CI: 1.10,
2.05) and at the 3-day avg lag period (RR = 2.54; 95% CI: 1.25, 5.18). A secondary analysis,
which adjusted for daytime and evening hours, found opposing associations for VT events:
positive associations in the nighttime hourly exposure and negative associations for daytime
hourly exposure. Specifically, at 24 hours after the increase in ozone exposure, the daytime odds
ratio was 0.69 (95% CI: 0.40, 1.21) and the evening odds ratio was 2.34 (95% CI: 1.27, 4.32).
• In a study conducted in Boston, MA using 2,369 participants in the Framingham Heart Study
Third Generation and Offspring Cohorts, a positive association between ozone and pulse
amplitude was identified for the 2-day moving avg lag period (7.63%; 95% CI: 0.87, 14.40%).
However, there was no change in pulse amplitude for the l-,3-,5-, or 7-day moving avg lag
periods (Liungman et al.. 2014).
• Cakmak et al. (2014) looked at eight endpoints of cardiac rhythm in 8,662 patients in Ottawa,
Ontario and Gatineau, Quebec, Canada referred for 24-hour ambulatory cardiac monitoring. An
increase in AV block (1.13%; 95% CI: 1.01, 1.26%) was observed for an increase of 15.67 ppb
ozone calculated as a 3-hour max. When stratified by season, AV block was still elevated by
1.23% (95% CI: 1.07, 1.42%) in the warm season from April to September. Additionally, in the
cold season, an increased number of supraventricular ectopic runs (8.15%; 95% CI: 0.34,
16.57%) was observed.
4.1.7.3 Controlled Human Exposure Studies
In the 2013 Ozone ISA Devlin et al. (2012) reported that the QTc interval significantly increased
immediately after ozone exposure and that the QRS complex significantly decreased immediately after
ozone exposure. However, an additional study from the previous review noted that ventricular
repolarization was most affected by the combined pollutant exposure of ozone and PM rather than to
ozone alone in healthy adults (Sivagangabalan et al.. 2011). Similarly, a study from the 2006 AQCD
noted that short-term ozone exposure alone did not result in ECG abnormalities (Gong et al.. 1998). The
few CHE studies published since the 2013 Ozone ISA provide little evidence that short-term ozone
exposure (2-3 hours) can appreciably affect cardiac electrophysiology. That is:
• In older adults, Rich et al. (2018) reported that short-term exposure to ozone (0.07, 0.12 ppm) did
not result in changes in a variety of cardiac electrophysiological endpoints, including the QTc
interval, QRS complexity, or T-wave amplitude. Moreover, there was no ozone effect on
ventricular or supraventricular arrhythmia. However, the authors did report a trend toward an
increase in the probability of ventricular but not supraventricular ectopy couplets or runs at
0.070 ppm (but not 0.12 ppm).
• In healthy adults, Kusha et al. (2012) reported a significant change in T-wave alternans during the
first 5 minutes of exposure (0.12 ppm; p < 0.05) relative to FA, but no change relative to FA later
in the exposure. The authors speculated that the significant effect reported during the first
4-18
-------
5 minutes of exposure was likely an artifact. Thus, they concluded that there was little evidence
of an ozone-induced effect on T-wave alternans.
Altogether, there is little evidence from a small number of CHE studies indicating that ozone
exposure may result in conduction abnormalities or arrhythmia. Additional information about these
studies can be found in Table 4-15.
4.1.7.4 Animal Toxicological Studies
The 2013 Ozone ISA describes studies from the 2006 Ozone AQCD reporting an effect of
short-term ozone exposure on cardiac electrophysiology and indicators of arrhythmia. For example, the
2013 ISA notes that short-term ozone exposure in rats induced premature atrial contraction, indicators of
atrial block, and arrhythmia (U.S. EPA. 2013a). Recent studies demonstrate similar effects resulting from
short-term exposure (3-4 hours, some studies with multiple day exposures) to ozone.
• Farrai et al. (2012) found that in spontaneously hypertensive (SH) rats, short-term exposure to a
higher (0.8 ppm), but not a lower (0.2 ppm) ozone concentration resulted in a decrease in the QTc
interval and an increase in the PR interval (indicative of atrial block,/? < 0.05) during exposure.
No post-exposure effects were reported. In addition, the authors demonstrated that during
exposure, 0.8 ppm ozone resulted in an increase in atrial premature beats, atrioventricular block,
and sinoatrial block. Importantly, this study also found that after 18-hour ozone exposure, both
levels of ozone increased sensitivity to aconitine-induced arrhythmia (p < 0.05). Similar results
were also found in an another study by this group (Farrai et al.. 2016).
• In contrast, Wang et al. (2013) noted that short-term ozone (0.8 ppm) exposure in normotensive
rats resulted in ECGs that were similar to control exposures. Similarly, in normotensive mice,
Kurhanewicz et al. (2014) reported no significant effect of ozone (0.3 ppm) exposure on ECG
readings, including QRS, PR, and QTc, relative to FA exposure.
Overall, the results of these studies provide evidence that in SH rats, short-term exposure to
ozone can result in conduction abnormalities and indicators of arrhythmia (Table 4-16). Importantly,
these results also suggest that even at ozone exposure concentrations that do not result in overt symptoms,
these ozone exposures could "prime" SH rats for arrhythmic responses to an arrhythmogenic agent
(e.g., aconitine) at lower concentrations than would normally be expected. That said, results in
normotensive rats indicated that short-term ozone exposure did not significantly alter ECG measures.
4.1.8 Blood Pressure Changes and Hypertension
High blood pressure is typically defined as a systolic blood pressure above 130 mm Hg or a
diastolic blood pressure above 80 mm Hg. Systolic blood pressure (SBP) represents the pressure in the
arteries as the heart contracts, while diastolic blood pressure (DBP) represents the pressure in the arteries
as the heart is relaxed and is filling with blood. Prolonged high blood pressure is known as hypertension
and can lead to a thickening of the ventricular wall resulting in diminished filling during diastole. This
4-19
-------
can ultimately contribute to the development of arrythmia and heart failure. Pulse pressure (PP) or the
difference between SBP and DBP, as well as mean arterial pressure (MAP), which is a function of cardiac
output, systemic vascular resistance, and central venous pressure, are additional outcome metrics used in
studies of air pollution on blood pressure. Moreover, hypertension is one of an array of conditions,
including high blood sugar, excess body fat around the waist, and abnormal triglyceride levels, that
comprise metabolic syndrome (see Appendix 5). which is a risk factor for heart disease, stroke, and
diabetes.
4.1.8.1 Epidemiologic Studies of Emergency Department Visits and Hospital
Admissions
No time-series or case-crossover studies analyzing ED visits or hospital admissions for
hypertension were reported in the 2013 Ozone ISA. Recent evidence is limited in number and generally
inconsistent.
• A study of ED visits for hypertension in two Canadian cities, Calgary and Edmonton, reported an
increased OR of 1.15 among women (95% CI: 1, 1.31), but not men, during the warm season
(Brook and Kousha. 2015). No association was observed for women or men during the cold
season. A study in Lithuania analyzed emergency medical service records of emergency calls for
exacerbations of essential hypertension with elevated arterial blood pressure and found
associations with 8-hour max ozone concentrations primarily during the warm season
(Vcnclovicnc et al.. 2017). While median ozone concentrations in the two study areas were
similar (approximately 20 ppb), the maximum concentration in Kaunas, Lithuania (102 ppb) was
twice that in the two Canadian cities (50 ppb). No association with ED visits for hypertension and
8-hour max ozone concentration was reported in a time-series study in Arkansas, an area with a
higher median ozone concentration (39 ppb) compared with the two other studies that analyzed
associations with hypertension rRodopoulou et al. (2015); Table 4-171.
4.1.8.2 Epidemiologic Panel Studies
Few studies were available in the 2013 Ozone ISA regarding the association between blood
pressure endpoints and short-term ozone exposure. One study found a positive association between
subjects with CVD and higher DBP associated with a 5-day avg; however, the effect estimate was not
sustained when the model was adjusted for PM2 5 (Zanobetti et al.. 2004). The evidence from recent
studies remains inconsistent and is characterized in Table 4-18. Specifically:
• Cakmak et al. (2011) observed positive associations between ozone concentration and resting
SBP (1.17 mm Hg; 95% CI: 0.29, 2.05) and resting DBP (0.65 mm Hg; 95% CI: 0.06, 1.23) in a
nationwide Canadian cohort with 5,011 subjects. However, in 70 subjects with pre-existing type 2
diabetes an opposing effect was observed over a 5-day mean of ozone exposure with decreases in
MAP (-3.15; 95% CI: -5.86, -0.34), SBP (-4.51; 95% CI: -7.44, -1.58), and DBP [-2.26; 95%
CI: -4.74, 0.02; Hoffmann et al. (2012)1. Additionally, in a cohort of Canadian children ages
6-17 years of age, Dales and Cakmak (2016) observed increases in SBP (4.41; 95% CI: 1.91,
4-20
-------
6.93) and DBP (3.55; 95% CI: 1.01,6.08) in children clinically diagnosed with a mood disorder
and no change in SBP (-0.52; 95% CI: -1.18, 0.14) or DBP (-0.24; 95% CI: -0.85, 0.36) in
children without a diagnosed mood disorder. Yet, several additional studies reported no changes
in blood pressure measures (Cole-Hunter et al.. 2018; Mirowskv et al.. 2017; Cakmak et al..
2014).
4.1.8.3 Controlled Human Exposure Studies
CHE studies available at the time of the last review indicated that short-term ozone exposure
alone did not have an effect on diastolic blood pressure (Sivagangabalan et al.. 2011; Brook etal.. 2009;
Fakhri et al.. 2009). Since the publication of the 2013 Ozone ISA, CHE studies continue to report little
evidence of an effect of short-term (1-3 hours) ozone exposure on measures of blood pressure.
Specifically:
• Frampton et al. (2015) reported a blunting of an exercise-induced increase in SBP (p < 0.05)
(0.2 ppm), with no change in DBP following ozone exposure (0.1, 0.2 ppm) in healthy subjects
with or without GSTM1 deletion. However, the authors were unclear of the clinical significance
of the effect. Other CHE studies reported no effect of short-term ozone exposure (0.3, 0.7,
0.12 ppm) on SBP, DBP, or angiotensin converting enzyme (ACE) levels in healthy or older
adults (Rich et al.. 2018; Arjomandi et al.. 2015; Barath et al.. 2013). Additional information with
respect to these studies can be found in Table 4-19.
• Stiegel et al. (2017) reported a significant decrease in DBP but not SBP in response to ozone
exposure (0.3 ppm) in healthy adults when comparing post versus pre-exposure.
4.1.8.4 Animal Toxicological Studies
The 2013 Ozone ISA cited a study by Chuang et al. (2009) that reported an increase in BP in
mice following short-term ozone exposure when compared with control animals. Since the publication of
the 2013 ISA, there is additional evidence from some, but not all studies to suggest that short-term
exposure (3-8 hours, some studies with multiple day exposures) to ozone can result in changes in blood
pressure in animals (Table 4-20). Moreover, some results also suggest that changes in diet may mediate
these effects. With respect to this evidence we note the following key points:
• Farrai et al. (2016) reported that relative to FA exposed animals, SH rats exposed to ozone
(0.3 ppm) alone experienced an increase in pulse pressure and a decrease in DBP (p < 0.05). No
change in SBP was reported.
• In rats fed a high fructose diet, Wagner et al. (2014) reported a decrease in SBP, DBP, and MAP
(p < 0.05) with ozone (0.5 ppm). In contrast, ozone-exposed rats fed a normal diet displayed an
increase in DBP (p < 0.05). Furthermore, Tankerslev et al. (2013) demonstrated an increase in
right ventricular systolic pressure and total peripheral resistance in ozone (-0.5 ppm)-exposed
mice compared with control mice (p < 0.05).
• No differences in SBP were found in studies of rats following short-term exposure to ozone
(0.8 ppm) 24 hours after six exposures (Wang et al.. 2013). Also, Ramot et al. (2015) reported no
4-21
-------
change in ACE activity following short-term ozone exposure (0.25, 0.5, 1 ppm) in several
different mouse strains.
4.1.9 Heart Rate (HR) and Heart Rate Variability (HRV)
Heart rate (HR) is a key indicator of autonomic function. It is modulated at the sinoatrial node of
the heart by both parasympathetic and sympathetic branches of the autonomic nervous system and
represents the number of times the heart beats in a given time frame (e.g., per minute). In general,
increased sympathetic activation increases HR, while enhanced activation of parasympathetic, vagal tone,
decreases HR (Lahiri et al.. 2008). Heart rate variability (HRV) represents the degree of difference in the
inter-beat intervals of successive heartbeats. Given that both arms of the autonomic nervous system
contribute, changes in HRV are an indicator of the relative balance of sympathetic and parasympathetic
tone to the heart and their interaction (Rowan III et al.. 2007). Low HRV is associated with an increased
risk of cardiac arrhythmia and an increased risk of mortality in patients with congestive heart failure
awaiting a heart or lung transplant (Fauchicr et al.. 2004; Bigger etal.. 1992V Low HRV has also been
shown to be predictive of coronary artery disease (Kotecha et al.. 2012). Notably, increases in HRV have
also been associated with increases in mortality (Carll et al.. 2018). In general, the two most common
ways to measure HRV are time-domain measures of variability and frequency-domain analysis of the
power spectrum. With respect to time-domain measures, the standard deviation ofNN intervals
(i.e., normal to normal or the interval between consecutive normal beats; SDNN) reflects overall heart
rate variability, and root-mean-square of successive differences in NN intervals (rMSSD) reflect
parasympathetic influence on the heart. In terms of frequency domain, high frequency (HF) domain is
widely thought to reflect cardiac parasympathetic activity while the low frequency (LF) domain has been
posited as an indicator of the interaction of the sympathetic and parasympathetic nervous systems
(Billman. 2013). although its linkage with sympathetic tone is controversial and uncertain (Notarius et al..
1999).
4.1.9.1 Epidemiologic Panel Studies
The 2013 Ozone ISA noted inconsistent results in studies for HRV. It specifically noted that
studies showing positive associations were in the same geographic area and that ozone may have been a
proxy for other pollutants ITJ.S. EPA (2013a) pgs. 6-172 to 6-75], Since the last ISA, studies evaluating
heart rate and HRV have continued to have inconsistent results (Table 4-21). The inconclusive evidence
may result from the variations in studies, including, but not limited to, sample size, demographics,
exposure, and time lags evaluated for these endpoints. For example:
• Several studies that evaluated resting heart rate observed inconsistent results. One study of
5,011 subjects aged 6-79 years in the Canadian Health Measures Survey showed an increase of
0.90 BPM (95% CI: 0.18, 1.63) with short-term exposure to ozone (Cakmak et al.. 2011).
4-22
-------
However, Cole-Hunter etal. (2018). who used 227 subjects from the TAPAS and EXPOsOMICS
cohorts in Barcelona, Spain, did not observe changes in heart rate when assigning spatially
weighted ozone exposure according to residential address or when using a mixed model to assign
exposure based on home and work address. Cakmak et al. (2014) used a population of
8,662 Ottawa and Gatineau patients referred for 24-hour ambulatory cardiac monitoring with
exposure linked to the monitor closest to their home address and observed no differences in
resting heart rate due to short-term exposure to ozone. Finally, in a cohort of Canadian children
ages 6-17 years, Dales and Cakmak (2016) observed no change in heart rate (bpm) in children
clinically diagnosed with a mood disorder (2.47; 95% CI: -1.52, 6.47) or in children without a
diagnosed mood disorder (-0.42; 95% CI: -1.36, 0.52). However, the heart rate was higher in the
clinically diagnosed population.
• Two studies evaluated the HRV measures SDNN and rMSSD in elderly populations with
previously diagnosed coronary artery disease [CAD; Mirowsky et al. (2017); Bartell et al.
(2013)1. Bartell etal. (2013) found decreases of-9.21% (95% CI: -15.80, -2.63%) for SDNN
and -9.03% (95% CI: -19.23, 1.15%) for rMSSD in a pool of 50 elderly nonsmokers in the Los
Angeles area (mean 24-hour avg ozone concentration 27.1 ppb). Conversely, Mirowsky et al.
(2017) observed no change in these variables in 13 elderly men in the vicinity of Chapel Hill, NC
(mean 24-hour avg ozone concentration 26.0 ppb).
4.1.9.2 Controlled Human Exposure Studies
In the 2013 Ozone ISA, a couple of controlled human exposure studies demonstrated some
evidence of changes in HRV following short-term ozone exposure. More specifically, both studies
reported changes in HF following short-term ozone exposure. However, one study demonstrated an
increase in HF, while the other reported a decrease (Devlin et al.. 2012; Fakhri et al.. 2009). In addition,
there was some evidence of a trend for an increase in SDNN (Fakhri et al.. 2009). One CHE study
reported an increase in HR following ozone exposure in a combined group of hypertensive and healthy
controls (Gong et al.. 1998).
Since the publication of the 2013 Ozone ISA, additional CHE studies have examined the
relationship between short-term exposure (1 to 4 hours) to ozone and HRV-related measures, but
evidence of an ozone-mediated effect remains limited. There is also no evidence from more recent CHE
studies for an ozone effect on HR. With respect to this evidence we note the following:
• In healthy men (Barath et al.. 2013) and older adults (Rich et al.. 2018). no changes in time and
frequency measures of HRV following short-term ozone (0.07, 0.12, 0.3 ppm) exposure were
reported.
• However, Ariomandi et al. (2015) reported that decreases in normalized HF and increases in
normalized LF were statistically significantly caused by increasing ozone (0.1, 0.2 ppm)
concentrations from 0 to 24 hours, but not from 0 to 4 hours in a group of asthmatics and
nonasthmatics (measurements were taken at 0, 4, and 24 hours). However, no changes were
reported in time-domain measures of HRV. Additional information about these studies can be
found in Table 4-22.
4-23
-------
• No CHE study reported a statistically significant effect of ozone (0.07, 0.1,0.12, 0.2, 0.3 ppm) on
changes in HR (Rich et al.. 2018; Arjomandi et al.. 2015; Frampton et al.. 2015; Barath et al..
2013; Kusha et al.. 2012).
4.1.9.3 Animal Toxicological Studies
The 2013 Ozone ISA presented some evidence that short-term exposure to ozone could result in
changes in HR and HRV ITJ.S. EPA (2013a). pgs. 6-203 to 6-204], With respect to HR, subsequent
studies in animals have reported inconsistent results following short-term ozone exposure (3-8 hours,
some studies with multiple-day exposures):
• Mclntosh-Kastrinskv et al. (2013) reported a decrease in HR in mice following short-term ozone
exposure (0.245 ppm) relative to FA, but not 40 minutes after reperfusion following ischemia.
Note, however, that the results following reperfusion could have been due to the ex vivo nature of
the experiment.
• Farrai et al. (2012) found that in rats, short-term exposure to a higher (0.8 ppm), but not a lower
(0.2 ppm), ozone concentration resulted in a 22.1% decrease in HR relative to pre-exposure
baseline levels (p < 0.05). In an additional study by this group (Farrai et al.. 2016). no change in
rat HR was reported following a FA exposure in the morning of Day 1 and a 0.3 ppm ozone
exposure in the afternoon of Day 2 (relative to FA exposures on both days). Si mi larlv. Wang et al.
(2013) reported no change in HR following short-term exposure to ozone in rats.
• Kurhanewicz et al. (2014) reported, no change in HR following short-term exposure to ozone
(0.3 ppm) in mice before ischemia. However, a significant decrease in HR was found relative to
FA 20 minutes after reperfusion in the ozone group.
• Wagner et al. (2014) reported that rats fed either a normal or high fructose diet had a significantly
decreased HR during a multiday ozone exposure (0.5 ppm) relative to FA. More information
about these studies can be found in the Table 4-23.
With respect to HRV, there is some recent evidence in studies of rodents that short-term exposure
(3-4 hours) to ozone can result in changes in HRV. It also appears from the limited available evidence
that the direction of this change may be dependent upon the exposure concentration, duration, and time
point examined. More specifically:
• In rats, Farrai et al. (2012) reported that exposure to a higher (0.8 ppm), but not a lower (0.2 ppm)
ozone concentration resulted in an increase in both time and frequency measures of HRV during
exposure, but not post-exposure relative to baseline. In an additional study by this group (Farrai et
al.. 2016). a decrease in time and frequency domains of HRV and no change in the LF:HF ratio
were reported in rats 24 hours after a FA exposure in the morning of Day 1 and a 0.3 ppm ozone
exposure in the afternoon of Day 2 (relative to FA exposures on both days). Together, these
results suggest that ozone exposure may initially result in a parasympathetic response, but some
hours later result in a transition to a more sympathetic response. The extent to which this
phenomenon may apply to humans, however, remains unclear.
• In rats, Wang et al. (2013) reported an increase in LF after six but not three exposures to ozone
(0.8 ppm; see Table 4-23) and no change in HF or the LF:HF ratio at eithertime point.
4-24
-------
• During a multiday ozone exposure (0.5 ppm) relative to FA, Wagner et al. (2014) reported a
significant increase in SDNN and RMSDD in SD rats fed a normal diet and in RMSDD, but not
SDNN in rats fed a high fructose diet.
• In addition, Kurhanewicz et al. (2014) reported no changes in time or frequency domains of HRV
during or 1 hour post ozone exposure (0.3 ppm) in mice. More information about these studies
can be found in the Table 4-23.
4.1.10 Coagulation and Thrombosis
Coagulation refers to the process by which blood changes from a liquid to a semisolid state to
form a clot. Increases in coagulation factors (e.g., fibrinogen, thrombin) or decreases in factors that
promote fibrinolysis like tissue plasminogen activator (tPA) can promote clot formation, and thus,
increase the potential for MI.
4.1.10.1 Epidemiologic Studies of Emergency Department Visits and Hospital
Admissions
In a case-crossover study of cases identified from discharge data in Spain from 2001-2013, an
increased risk of pulmonary embolism was reported for ozone concentrations averaged over the 3 days
around the time of diagnosis as compared to the average concentration for a similar period 3 weeks prior
(dc Miguel-Diez et al.. 2016). No associations were observed when control periods closer to the time of
diagnosis were analyzed. No associations with first diagnosis for pulmonary embolism and average
monthly ozone concentration were reported by a case-control study in Italy (Spiezia et al.. 2014) or in a
case-crossover study in the U.K. that analyzed 8-hour max ozone concentrations and lags of 0-4 days
rMiloievic et al. (2014); Table 4-241.
4.1.10.2 Epidemiologic Panel Studies
Previously, short-term exposure to ozone showed inconsistent results for coagulation biomarkers
such as PAI-1, fibrinogen, and vWF. These studies varied in location and study design, making
conclusions difficult ITJ.S. EPA (2013a). pgs. 6-178 to 6-180], Studies since the last ISA continued to be
inconsistent with respect to changes in biomarkers of coagulation (Table 4-25). That is:
• A panel study conducted in six U.S. cities evaluated 2,086 women with an average age of
46.3 years reported no change in PAI-1 for lags of 1 or 30 days for short-term increases in ozone
exposure (Green et al.. 2015). Conversely, in a small sample size of men with pre-existing CAD
(n = 13), Mirowskv et al. (2017) found positive associations of short-term ozone exposure and
PAI of 21.43% (95% CI: 0.86, 45.86%) at lag 2 and 43.39% (95% CI: 9.32, 87.43%) for a 5-day
moving avg.
4-25
-------
• No studies observed changes in fibrinogen levels resulting from increases in short-term ozone
exposure in large study populations (Li et al.. 2017; Green et al.. 2015; Bind et al.. 2012).
4.1.10.3 Controlled Human Exposure Studies
In the 2013 Ozone ISA, a controlled human exposure study demonstrated changes in markers of
coagulation following short-term ozone exposure. More specifically, Devlin et al. (2012) reported a
statistically significant decrease in PAI-1 both immediately following and 24 hours post-exposure, as well
as a decrease in plasminogen levels and a trend toward an increase in tPA. Given these results, the authors
suggested that ozone exposure may activate the fibrinolysis system ITJ.S. EPA (2013a). pg. 6-166], Since
the publication of the 2013 Ozone ISA, several CHE studies have examined the potential for short-term
ozone exposure (1-2 hours) to result in changes to markers of coagulation or fibrinolysis, but evidence of
an effect on these endpoints remains limited. Specifically:
• A study on the effect of temperature on ozone exposure (0.3 ppm) in healthy young volunteers
reported a statistically significant decrease in PAI-1 and plasminogen levels 24 hours
post-exposure (p < 0.05) when the experiment was carried out at 22°C, but a significant increase
in these coagulation markers when the exposure was conducted at 32.5°C (p < 0.05) (Kahle et al..
2015). This study also reported no changes in D-dimer, tPA, or vWF at either temperature.
• However, other CHE studies (Frampton et al.. 2017; Arjomandi et al.. 2015; Frampton et al..
2015; Barath et al.. 2013) have reported that were no measurable changes in markers of
coagulation or fibrinolysis (e.g., D-dimer, platelet activation, PAI-1, plasminogen) following
short-term ozone (0.1, 0.2, 0.3 ppm) exposure. Additional information about these studies can be
found in Table 4-26.
4.1.10.4 Animal Toxicological Studies
The 2013 Ozone ISA contained very limited animal toxicological evidence that short-term
exposure (4 hours, some studies with multiple-day exposures) to ozone could result in changes to factors
related to coagulation or fibrinolysis (U.S. EPA. 2013a). This remains the case in the current review
(Table 4-27):
• Snow et al. (2018) demonstrated that in rats fed a normal or coconut oil or fish oil supplemented
diet, short-term exposure to ozone resulted in an increase in circulating platelets relative to FA
exposure given the same diet (p < 0.05).
• In a study comparing the susceptibility of six different strains of mice to ozone (0.25, 0.5,
1.0 ppm, see Table 4-27). Ramot et al. (2015) reported that short-term ozone exposure did not
increase blood D-dimer levels in any mouse strain and decreased fibrinogen levels in just one of
these strains (FHH mice, which are characterized as developing hypertension and proteinuria at a
young age).
4-26
-------
4.1.11
Systemic Inflammation and Oxidative Stress
Systemic inflammation has been linked to a number of CVD-related outcomes. For example,
circulating cytokines such as IL-6 can stimulate the liver to release inflammatory proteins (e.g., CRP) and
coagulation factors that can ultimately increase the risk of thrombosis and embolism. Other indicators of
systemic inflammation include an increase in inflammatory cells such as neutrophils and monocytes and
other cytokines such as TNF. Similarly, oxidative stress can result in damage to healthy cells and blood
vessels and further increase the inflammatory response. Thus, this section discusses the evidence for
changes in markers of systemic inflammation and oxidative stress following short-term ozone exposures.
4.1.11.1 Epidemiologic Panel Studies
Studies in the 2013 Ozone ISA showed inconsistent results for inflammatory and oxidative stress
biomarkers. Specifically, a positive association was observed in IL-6 (Thompson et al.. 2010). while CRP
studies reported either no association (Rudcz et al.. 2009; Steinvil et al.. 2008) or increases (Chuang et al..
2007) following short-term ozone exposure. In addition, oxidative stress markers had mixed results, with
no studies evaluating the same biomarkers ITJ.S. EPA (2013a). pg. 6-180],
There are few studies that demonstrate short-term exposure to ozone results in changes in
inflammatory biomarker levels. Studies reviewed for these endpoints are summarized in Table 4-28.
Altogether, the recent and older epidemiologic panel studies provide evidence that short-term ozone
exposure is associated with increased inflammatory responses.
• Most commonly, studies examined changes in C-reactive protein (CRP) as a biomarker to
identify inflammation. Across these studies, a single study reported changes in CRP after
short-term exposure to ozone. Bind et al. (2012). looking 24 hours post-exposure, reported a
10.8% (95% CI: 2.2, 20.5%) increase in CRP in more than 700 elderly men free of chronic
medical conditions, living in the greater Boston area. Among the remaining studies, consisting of
cohorts of middle-aged women, men with previously diagnosed CVD, and noncurrent smokers,
there were no differences in CRP reported over several different lag times (Li et al.. 2017;
Mirowskv et al.. 2017; Green et al.. 2015).
• Mirowskv et al. (2017) found increases in IL-6 at lag 4 (17.04%; 95% CI: 3.86, 33.71%),
neutrophils at lag 1 (9.32%; 95% CI: 1.61, 17.57%) and lag 2 (9.00%; 95% CI: 1.07, 17.46%),
monocytes at lag 1 (10.92%; 95% CI: 1.07, 21.54%), and TNF-a at lag 2 (6.32; 95% CI: -0.96,
14.14) in 13 men with previously diagnosed CAD. However, these results occur at various
time-lapses and have wide confidence intervals with a small sample size. These increases
changed less than 10% when adjusted for PM2 5, suggesting that they may be related to ozone
exposure.
• TNFR2 increased in a cohort of over 3,000 subjects when evaluated over 1-7 days moving avg
exposure to ozone. Additionally, when these results were stratified by age, CVD or no CVD,
statin use, and season, the associations remained positive (Li et al.. 2017).
4-27
-------
• A single study in a cohort of over 3,000 subjects evaluated 1-7 days moving avg exposure to
ozone reported no change in the oxidative stress biomarkers myeloperoxidase and indexed
8-epi-prostaglandin F2alpha (Li et al.. 2016).
4.1.11.2 Controlled Human Exposure Studies
In the 2013 Ozone ISA, a controlled human exposure study reported significant increases in CRP,
IL-1, and IL-8, but not TNF-a following exposure to ozone (Devlin et al.. 2012). In addition, Brook et al.
(2009) found a decrease in total white blood cell count, but not in TNF, or neutrophil percentage. Since
the 2013 Ozone ISA, CHE studies have provided limited additional evidence for changes in inflammatory
markers following short-term ozone exposure (0.5-4 hours). For example:
• Biller etal. (2011) reported an increase in percentage blood neutrophils (p < 0.05) relative to FA
exposure at 5, 7, but not 24 hours post-exposure (0.25 ppm) in healthy volunteers. These authors
also reported increased neutrophil activation at 5 and 7, but not 24 hours post-exposure. With
respect to total leukocytes, there was a significant increase at 5 and 7 hours, but not at 24 hours
(p < 0.05).
• In a time course study, Bosson et al. (2013) reported a decrease in blood neutrophils (p < 0.05) in
healthy volunteers at 1.5 hours post-exposure (0.2 ppm) when compared to FA exposure. These
levels rebounded above FA levels when measured at 6 hours (p < 0.05), and at 18 hours
post-exposure, there was no difference in neutrophil levels when compared to FA. Similar results
were found with respect to total leukocytes. In addition, the authors also describe a correlation
between neutrophil levels in the blood and the lung. No impact of ozone was found on blood
monocytes or lymphocytes.
• In healthy volunteers, Stiegel et al. (2016) reported an increase in percentage neutrophils
following ozone exposure (0.3 ppm) immediately after (p > 0.05), but not 24 hours post-exposure
when compared to pre-exposure. However, similar results were reported following clean air
exposure, calling into question the significance of the ozone exposure on these changes. These
authors also reported a decrease in the total percentage of lymphocytes (p < 0.05), but no change
in the percentage of monocytes. Again however, similar results were reported following clean air
exposure. No appreciable changes in a number of cytokines, including IL-8 and TNF-a, were
reported following ozone exposure.
• Ariomandi et al. (2015) reported a decline in eosinophil levels from 0-4 (p < 0.05), but not
0-24 hours associated with increasing ozone concentrations from 0.1 to 0.2 ppm in adults with or
without asthma. However, asthma status of the volunteers had no impact on these changes. No
significant changes in total leukocytes or monocytes, or neutrophils were reported. No significant
changes in a number of cytokines were reported, including IL-1 and TNF-a.
• In addition, a study reported statistically significant increases in blood CRP levels across
exposures ranging from 0 to 200 ppb, while another reported a significant increase in CRP when
comparing post-exposure to pre-exposure levels (Ariomandi et al.. 2015; Biller etal.. 2011).
However, Frampton et al. (2017) found no statistically significant changes in CRP, IL-6,
P-selectin, or 8-isoprostane (an oxidative stress marker) levels in older adults following 70 or
120 ppb ozone exposure.
• Ramanathan et al. (2016) also demonstrated that ozone exposure (0.12 ppm) did not alter HDL
antioxidant or anti-inflammatory capacity in healthy adults.
4-28
-------
Taken together, there is limited additional evidence that short-term exposure to ozone may result
in changes to some inflammatory cells and cytokines in a manner that is concentration and timepoint
dependent (Table 4-29). This may particularly the case with neutrophils. That is, exposure to ozone may
first cause a decrease in neutrophils in the blood (perhaps as these cells migrate into the lung), followed
by an increase later post-exposure.
4.1.11.3 Animal Toxicological Studies
In the 2013 Ozone ISA, animal toxicological studies demonstrated that short-term exposure to
ozone resulted in an increase in inflammatory markers (U.S. EPA. 2013a). In addition, studies in mice
and monkeys demonstrated that short-term exposure to ozone resulted in an increase in markers of
oxidative stress. Although not entirely consistent within and across studies, more recent animal
toxicological studies provide some evidence that short-term exposure (2-24 hours, some studies with
multiple-day exposures) to ozone results in an increase in markers of inflammation and oxidative stress.
With respect to this evidence, we note the following key points:
• Zhong et al. (2016) reported that in obese-prone mice, short-term exposure to ozone (0.5 ppm)
resulted in an increase in inflammatory monocytes and CD4 T cells in blood (p < 0.05). Similarly,
in rats Paffett et al. (2015) reported an increase in neutrophils and macrophages in blood as a
result of short-term ozone exposure (p < 0.05).
• Studies also demonstrated that lymphocytes, T cells, or WBC counts decreased (Snow et al..
2018; Ramot et al.. 2015; Thomson et al.. 2013) following short-term ozone (0.25, 0.4, 0.5, 0.8,
1.0 ppm) exposure (p < 0.05). However, some of these studies also found no appreciable effect of
short-term ozone exposure on other cell populations. For example, in rats Paffett et al. (2015)
reported no change in lymphocytes or eosinophils in blood following short-term ozone (1 ppm)
exposure.
• With respect to markers of inflammation, evidence was inconsistent across studies. For example,
Thomson et al. (2013) reported a decrease in TNF-a mRNA and IL-1 (p < 0.05) in rat heart tissue
following short-term ozone exposure of 0.8 ppm, but not 0.4 ppm. However, there was little
effect of ozone (0.8 ppm) exposure on a panel of 24 cytokines in the blood of rats (Thomson et
al.. 2016).
With respect to markers of oxidative stress:
• Kumarathasan et al. (2015) reported that exposure to 0.8 but not 0.4 ppm ozone can increase
o-tyrosine, but not m-tyrosinc or the lipid peroxidation marker 8-isoPGF2a in rats.
• Martinez-Campos et al. (2012) also reported in rats that short-term ozone (0.5 ppm) exposure
could lead to an increase in plasma levels of MDA and 8-IP. Moreover, Farrai et al. (2016)
reported decreased SOD activity following short-term ozone (0.3 ppm) exposure in rats.
• However, Cestonaro et al. (2017) found that short-term exposure to ozone (0.05 ppm) resulted in
no evidence of lipid peroxidation in rats. Similarly, Thomson et al. (2013) reported that in rats,
short-term exposure to ozone (0.4, 0.8 ppm) did not cause an increase in MDA mRNA in heart
tissue exposure. Furthermore, Wang et al. (2013) found that short-term exposure to ozone alone
did not affect SOD 1 or MDA levels in rat hearts.
4-29
-------
Although not demonstrated in all studies, the recent studies presented above provide some
evidence that short-term exposure to ozone can result in changes in markers of inflammation and
oxidative stress (Table 4-30).
4.1.12 Stroke and Associated Cardiovascular Effects
4.1.12.1 Epidemiologic Studies of Emergency Department Visits and Hospital
Admissions
A few studies of cerebrovascular disease and stroke were discussed in the 2013 Ozone ISA,
including one Finnish study (Haloncn et al.. 2009). a multicity French study (Larrieu et al.. 2007). and an
analysis of stroke subtypes in Edmonton, Canada (Villcncuvc et al.. 2006). The Canadian study reported a
weak elevated risk of ischemic and hemorrhagic stroke for 24-hour avg ozone concentrations during the
warm season, but not in other seasons; however, confidence intervals were wide (Villeneuve et al.. 2006).
In contrast, an inverse association with transient ischemic stroke during the warm season was observed.
Several studies have been published since the 2013 Ozone ISA, and results have been inconsistent.
Confidence intervals around the risk ratios tended to be wide, indicating the relative imprecision in the
reported associations.
• A more recent study in Edmonton that evaluated acute ischemic stroke using lags of different
time periods between 0 and 72 hours found an inverse association with 1-hour max ozone
concentration during the warm season and a positive association during the cold season, although
effect estimates were not precise (Chen et al.. 2014). Several recent studies in Europe found no
association with ischemic stroke (Table 4-31; Figure 4-4).
• While a study in Nice, France also found no associations with ischemic stroke overall with 8-hour
max ozone concentrations in the preceding 3 days, the study reported a 43% increase in recurrent
stroke (OR: 1.43; 95% CI: 1.03, 1.99) per 20 ppb 8-hour max ozone concentration at a lag of
1 day, and a 8% increase in large artery stroke risk (OR: 1.08; 95% CI: 1.01, 1.16) per 15 ppb
24-hour avg ozone concentration at lag Day 3 (Suissa et al.. 2013). For recurrent stroke, larger
odds ratios were observed with 8-hour max ozone concentrations concurrent with and the day
prior to the event, and for large artery stroke, an elevated odds ratio was observed only for
24-hour avg ozone concentrations on lag Day 3. A dose-response trend was observed with
increasing quartiles of ozone concentration for both stroke groups. The results of this study were
consistent with those of a study in Dijon, (Henrotin et al.. 2010; Henrotin et al.. 2007). Two other
studies that analyzed associations with ischemic stroke overall or for stroke subtypes primarily
found null or weakly positive associations with 8-hour max or 24-hour avg ozone concentrations
(Maheswaran et al.. 2016; Corea et al.. 2012).
• A comparison of risk of transient ischemic attacks (TIA) and minor stroke in the NORTHSTAR
cohort in England found associations in opposite directions in two communities (Bedada et al..
2012). In Manchester, an increased TIA and stroke risk with increasing ozone concentration was
found at lag Day 0 and an inverse association was found at lag Day 1. In Liverpool, an opposite
pattern was observed: an inverse association at lag Day 0 and an increased OR at lag Day 1
4-30
-------
(Table 4-31). The number of cases accrued over the 5-year study was low (N = 374 from
Liverpool, N = 335 from Manchester) resulting in imprecise effect estimates.
• In the U.S., a small elevated risk was found for stroke hospitalizations in Allegheny County, PA,
with 24-hour avg ozone concentrations on the day of hospitalization (Xu et al.. 2013). One study
in Nueces County, TX evaluated associations with incident stroke and stroke severity with cases
identified in the Brain Attack Surveillance in Corpus Christi project between 2000 and 2012
(Wing et al.. 2017b; Wing et al.. 2015). The investigators reported a small elevated increase in
risk of incident stroke with a 20 ppb increase in 8-hour max ozone concentrations on the 4 days
concurrent with and preceding the event record, with the highest increase on lag Day 2 (OR: 1.05;
95% CI: 0.97, 1.12). Effect measure estimates were not changed in a model that included PM2 5.
This study also reported an elevated risk among adults with severe incident stroke (OR: 1.27;
95% CI: 1.12, 1.41). Severe stroke was defined as the upper quartile (score >7) of the score
obtained using the National Institutes of Health Stroke Scale (NIHSS). An analysis of first
recurrent stroke also was conducted in the Texas population. A total of 317 recurrent ischemic
strokes were identified between 2000 and 2012, and in contrast to the findings for incident stroke,
no associations were observed with increases in 8-hour max ozone concentration (Wing et al..
2017a).
• Two other studies in the U.S. reported inverse associations with ED visits or hospital admissions
for cerebrovascular disease [ICD-9 430-438; Montresor-Lopez et al. (2015); Rodopoulou et al.
(2015)1. Rodopoulou et al. (2015) found an inverse association for both the cold and warm
seasons in Little Rock, AR, which was not altered in a copollutant model with PM2 5. Montresor-
Lopez et al. (2015) also conducted separate analyses for ischemic and hemorrhagic stroke in their
study in South Carolina, and found no associations for these subgroups generally, other than a
small increase at lag Day 2 (OR: 1.02; 95% CI: 0.9, 1.17).
• Few studies of cerebrovascular disease have examined differences by age, sex, or ethnicity.
Studies conducted in the U.K. did not find notable differences between men and women or for
individuals 75 years and older for ischemic stroke diagnoses (Maheswaran et al.. 2016; Milojevic
et al.. 2014). In the U.S., an increase in risk of stroke hospitalization was strongest among men
and individuals between the ages of 65 and 79 years compared with those 80 years or older (Xu et
al.. 2013). However, no difference in risk by sex was found in another study among hospitalized
residents of South Carolina with a first diagnosis of stroke (Montresor-Lopez et al.. 2015). Wing
et al. (2015) observed a higher risk among non-Hispanic whites compared to no elevated risk
among Mexican-Americans associated with 8-hour max ozone concentrations at lag Days 2 and
3.
• The risk of ischemic stroke associated with a 0-6 days mean 24-hour avg ozone concentration
was higher among stroke cases from the South London Stroke register with pre-existing
hypertension or atrial fibrillation (Maheswaran et al.. 2016).
4-31
-------
Reference Location
Villeoeuve et al. 2006 Canada
tCheil el al. 2014
Canada
IMontresor-L opez et a I South Caroliiia
201?
DCuetal. 2013 Pennsylvania
Haloaca ct a!. 2009 Finland
I, an ieu et al. 2007 8 French cities
Lag Type
0-2 Ischemic
0-2 Hemorrhagic
0-2 T1A
0-2
IMalieswara et al. 2016 London, UK 0-6
IButlaild et al. 2017 London. UK 0
IBedada et al. 2012 Manchester, UK 0
Liverpool. UK 0
IVidale et al. 2017 Como, Italy 0
IMcchtouffet al, 2012 Rhone, France NR
IMilojevic et al. 2014 UK 0-4
IWitig et al, 2015 Texas 0
I Wing et al. 2017 Texas
IRodopoulou et al. 2015 Arkansas
Henrotin et al. 2007 Dijon, France
ISuissa ct al. 2013 Nice. France
0
0
0
0-3
0-5
0-1
All
Ischemic
First-ever
Ischemic
Hemorrhagic
TIA. stroke
TIA. stroke
Ischemic
Ischemic ¦*-
First-ever
First-ever, severe
Recurrent
All
Ischemic
All ischemic
Recurrent
Large artery
All
Ischemic
Hemorrhagic
First-ever
All
All
t—
0.7
1 h
-1 h
0.8 0.9 1 1.1 1.2 1.3 1.4
15
Risk Ratio (95% CI)
TIA = transient ischemic attack.
Note: "("Studies published since the 2013 Ozone ISA. Studies are listed from the top in order of increasing mean or median ozone
concentration reported in the publication. Associations are presented per 25-ppb increase in pollutant concentration for 1-hour max
avg times, 20-ppb increase for 8-hour avg times, and 15-ppb increase for 24-hour avg times. Symbols represent point estimates,
circles, triangles and squares represent the entire year, warm season, and cold season, respectively; horizontal lines represent 95%
confidence intervals for ozone. Black text and symbols represent evidence included in the 2013 Ozone ISA; red text and symbols
represent recent evidence not considered in previous ISAs or AQCDs.
Figure 4-4 Associations between short-term exposure to ozone and
cerebrovascular-related emergency department visits and
hospital admissions.
4-32
-------
4.1.13 Nonspecific Cardiovascular Effects
4.1.13.1 Epidemiologic Studies of Emergency Department Visits and Hospital
Admissions
Several studies of ozone concentrations and cardiovascular hospital admissions and ED visits for
all CVD diagnoses combined were discussed in the 2013 Ozone ISA. Generally, these studies did not
report an association between ozone concentrations and an increased risk of aggregated CVD in
populations in the U.S., Canada, Europe, and Australia.
• Recent studies that reported a risk ratio for combined cardiovascular disease outcomes show a
similar pattern to those studies included in the 2013 Ozone ISA (Table 4-32; Figure 4-5).
Although changes were small (<1%), associations were positive during the cold season and
negative during the warm season.
• Studies that evaluated effect modification by sex or age did not find notable differences
(Miloicvic et al.. 2014; Rodopoulou et al.. 2014). Winquist et al. (2012) observed a higher
relative risk per 8-hour max ozone concentration among individuals residing in a poverty area.
4-33
-------
Reference
IVidale et al. 2017
Location
Como, Italy
Lag Notes
IMilojevic et al. 2014 UK
0-4
IRodopoulou et al. 2014 Dona Ana County, NM 1 ED
1 HA
ISarnat et al. 2015
St. Louis, MO
0-2
IWinquist et al. 2012 St. Louis MSA, MO 0-4 ED
0-4 HA
IRodopoulou et al. 2015 Little Rock, AR 1
IHunova et al. 2013 Prague, Czech Republic 1
IHunova et al. 2017 Prague, Czech Republic 1
IChoi et al. 2011
Maryland
0-4
I 1 1 h
i 1 1 f
0.8 0.85 0.9 0.95 1 1.05 1.1 1.15 1.2
Risk Ratio (95% CI)
ED = emergency department; HA = hospital admissions.
Note: fStudies published since the 2013 Ozone ISA. Studies are listed from the top in order of increasing mean or median ozone
concentration reported in the publication. Associations are presented per 25-ppb increase in pollutant concentration for 1 -hour max
avg times, 20-ppb increase for 8-hour avg times, and 15-ppb increase for 24-hour avg times. Circles represent point estimates;
horizontal lines represent 95% confidence intervals for ozone. Black text and circles represent evidence included in the 2013 Ozone
ISA; red text and circles represent recent evidence not considered in previous ISAs or AQCDs.
Figure 4-5 Associations between short-term exposure to ozone and
nonspecific cardiovascular emergency department (ED) visits and
hospital admissions.
4-34
-------
4.1.14
Cardiovascular Mortality
No recent multicity study has extensively examined the relationship between short-term ozone
exposure and cardiovascular mortality. The majority of evidence examining cardiovascular mortality
consists of studies evaluated in the 2013 Ozone ISA, which reported positive associations for
cardiovascular mortality in all-year and summer/warm season analyses. Of the recent multicity studies
evaluated, only Vanos et al. (2014) examined cardiovascular mortality and reported positive associations
in all-year and summer season analyses, which is consistent with the multicity studies previously
evaluated. These studies are further characterized in Table 4-33. Additional single-city studies examined
cardiovascular mortality and reported the following:
• In a study conducted in Philadelphia, PA, with the aim of examining the influence of model
specification (i.e., control for seasonal/temporal trends, and weather covariates) on associations
between air pollution and cardiovascular mortality using statistical models from recent multicity
studies, Sacks et al. (2012) reported evidence of positive associations ranging from 1.30 to 2.20%
for those studies that more aggressively controlled for temperature in the statistical model
(i.e., either multiple temperature terms or a term for apparent temperature), while those studies
that only included one temperature term did not, with associations ranging from -1.60 to 0.50%
at lag 0-1 day for a 20-ppb increase in 8-hour max ozone concentrations.
• Klemm etal. (2011) conducted a study in Atlanta, GA that included 7.5 additional years of data
than Klemm and Mason (2000) and Klemm et al. (2004). In analyses that examined
cardiovascular mortality, the authors reported positive, but imprecise associations at lag 0-1 day
in all-year analyses (0.69% [95% CI: -2.28, 3.75%] for a 20-ppb increase in 8-hour max ozone
concentrations).
4.1.15 Potential Copollutant Confounding of the Ozone-Cardiovascular
Disease (CVD) Relationship
Recent studies that examined potential copollutant confounding focused on either PM2 5 or PM10,
or gaseous copollutants. The results of these studies extend the evidence from a small number of studies
in the 2013 Ozone ISA that demonstrated that ozone-cardiovascular health endpoint associations are
relatively unchanged in copollutant models as detailed below:
• Associations between short-term ozone exposure and cardiovascular health endpoints are
relatively unchanged in copollutant models that include PM. Specifically, Rosenthal et al. (2013)
observed that the elevated risks for cardiac arrest were either unchanged or increased in
copollutant models with PM2 5, PM10, or other particulate size classes. A study in St. Louis
observed a 4% (RR: 1.04; 95% CI: 0.99, 1.10) elevation in ED visits for CHF, which was
unchanged in copollutant models with PM2 5 (Sarnat et al.. 2015). Mirowskv et al. (2017) found
increases in IL-6, neutrophils, monocytes, and TNF-a in men with previously diagnosed CVD.
These increases were relatively unchanged when adjusted for PM2 5, suggesting that these
increases are directly related to ozone exposure alone.
• Associations between short-term ozone exposure and cardiovascular health endpoints are
relatively unchanged in copollutant models that include other gaseous pollutants. Rosenthal et al.
4-35
-------
(2013) observed that the elevated risks of cardiac arrest were either unchanged or increased in
copollutant models with NO, NO2, SO2, or CO. Similar results were found for the effect of ozone
on cardiac arrest (Raza etal.. 2014) or CHF (Sarnat et al.. 2015) after evaluating a copollutant
model with NO2. A study in St. Louis observed a 4% (RR: 1.04; 95% CI: 0.99, 1.10) elevation in
ED visits for CHF, which was increased to a 6% increased risk (RR: 1.06, 95% CI: 1.00, 1.12)
when CO was included in the model (Sarnat et al.. 2015).
4.1.16 Effect Modification of the Ozone-Cardiovascular Health Effects
Relationship
4.1.16.1 Lifestage
The 1996 and 2006 Ozone AQCDs identified children, especially those with asthma, and older
adults as at-risk populations (U.S. EPA. 2006. 1996a). The 2013 Ozone ISA confirmed these earlier
findings and concluded that there was adequate evidence that children and older adults are at increased
risk of ozone-related health effects (U.S. EPA. 2013a). Collectively, the majority of evidence for older
adults has come from studies of short-term ozone exposure and mortality. No recent studies contribute
evidence for whether children are at a greater risk of cardiovascular health effects due to short-term ozone
exposure. A limited number of recent studies of short-term ozone exposure and cardiovascular health
effects have compared associations between different age groups, but the studies do not report consistent
evidence that older adults are at increased risk.
• Among the studies that evaluated the modification of the effect of exposure to ozone on heart
failure by lifestage, no notable differences were reported for older adults compared with other
adult age groups (Miloievic et al.. 2014; Winquist et al.. 2012).
• Pradeau et al. (2015) and Ensoretal. (2013) observed higher risks among persons older than
64 years for out-of-hospital cardiac arrest with a cardiac etiology (Ensor et al.. 2013). No
differences for age were reported by other studies of out-of-hospital cardiac arrest (Miloievic et
al.. 2014; Raza etal.. 2014).
• Cakmak et al. (2014) used a population of 8,662 Ottawa and Gatineau patients referred for
24-hour ambulatory cardiac monitoring with exposure linked to the monitor closest to their home
address. In subjects over the age of 50 (n = 6,009) cardiac rhythm was disrupted by an increased
presence of heart block (i.e., first-, second-, or third-degree atrioventricular blocks) frequency
(1.13; 95% CI: 1.01, 1.27).
• Increases in TNFR2 were associated with ozone exposure in a cohort of over 3,000 subjects.
When stratified by above or below age of 53 years, the results persisted, however, there was no
difference between the age groups (Li et al.. 2017).
4-36
-------
4.1.16.2 Pre-existing Disease
Individuals with certain pre-existing diseases may be considered at greater risk of an air
pollution-related health effect because they are likely in a compromised biological state that can vary
depending on the disease and severity. The 2013 Ozone ISA concluded that there was adequate evidence
for increased ozone-related health effects among individuals with asthma (U.S. EPA. 2013a). The results
of controlled human exposure studies, as well as epidemiologic and animal toxicological studies,
contributed to this evidence. For example, a few studies of short-term ozone exposure and mortality
provided some evidence for stronger associations among individuals with pre-existing cardiovascular
disease or diabetes.
A limited number of recent studies provides some evidence that individuals with pre-existing
diseases may be at greater risk of cardiovascular health effects associated with short-term ozone exposure.
These studies focus on specific diseases of varying severity (e.g., previous CVD events, type 2 diabetes).
Specifically:
• Larger increases in the odds of STEMI were observed for patients with previous MI (OR = 1.78;
95% CI: 0.97,3.28), CVD (OR = 1.72; 95% CI: 1.02, 2.90), and hypertension [OR= 1.34; 95%
CI: 1.00, 1.90: Evans etal. (2016)1
• Lanzingeretal. (2014) reported FMD decreases in individuals with type 2 diabetes. However,
Mirowskv et al. (2017) saw no change in FMD in men with a previous diagnosis of coronary
artery disease.
• Increases in TNFR2 were associated with short-term ozone exposure in a cohort of over
3,000 subjects. When these results were stratified by pre-existing CVD or no pre-existing CVD
the associations remained positive and relatively unchanged (Li et al.. 2017).
• A single study provided emerging evidence that children ages 6-17 years clinically diagnosed
with mood disorders showed increases in SBP (mm Hg; 4.41; 95% CI: 1.91, 6.93), DBP (mm Hg;
3.55; 95% CI: 1.01, 6.08), and HR (bpm; 2.47; 95% CI: -1.52, 6.47) relative to children without
a clinically diagnosed mood disorders for SBP (-0.52; 95% CI: -1.18, 0.14), DBP (-0.24; 95%
CI: -0.85, 0.36), and HR [-0.42; 95% CI: -1.36, 0.52: Dales and Cakmak (2016)1.
4.1.16.3 The Role of Season on Ozone Associations with Cardiovascular Health
Effects
As detailed in Appendix 1. Section 1.7. ozone concentrations are generally higher in the summer
or warm months due to the atmospheric conditions that lead to ozone formation. Therefore, many
locations, particularly within the U.S., only monitor ozone during the summer or warm months. Thus,
many of the epidemiologic studies tend to focus on summer or warm season analyses. However, some
studies conduct all-year analyses based on areas that monitor ozone year-round, with a subset of these
studies then examining whether the magnitude of the ozone-cardiovascular health association varies either
across seasons or in the summer/warm season compared with the entire year. Studies evaluated in the
2013 Ozone ISA reported evidence of positive ozone-cardiovascular health associations in all-year
4-37
-------
analyses that tended to be larger in magnitude during the warm or summer months. A limited number of
recent studies that conducted seasonal analyses reported associations that were similar for both warm
season and cool season analyses, Specifically, recent studies indicate:
• Evans et al. (2016) observed increased odds of STEMI in the cooler months (November to April)
for ozone exposure at 12 hours (OR= 1.43; 95% CI: 1.03, 1.98), 24 hours (OR= 1.45; 95% CI:
1.04, 2.03), and 72 hours (OR = 1.60; 95% CI: 1.05, 2.46), and no increased associations during
the warmer months.
• Increases in TNFR2 were associated with short-term ozone exposure in a cohort of over
3,000 subjects. When stratified by warm and cool seasons, the associations remained positive in
both season (Li et al.. 2017).
• Seasonality altered cardiovascular electrophysiology in a population of 8,662 Ottawa and
Gatineau patients referred for 24-hour ambulatory cardiac monitoring with exposure linked to the
3-hour max exposure for the 24 hours prior to the visit, based on the monitor closest to their home
address. During the warm season (April-September), Cakmak et al. (2014) reported increases in
the presence of heart block (1.23; 95% CI: 1.07, 1.42). However, in the cold season, the same
study reported increases in the number of supraventricular ectopic runs (defined as more than
three consecutive beats; 8.15; 95% CI: 0.34, 16.57) and the length of the longest ventricular
ectopic runs (20.68; 95% CI: 5.3, 38.31).
4.1.17 Summary and Causality Determination
The 2013 Ozone ISA concluded that the strongest evidence for an effect of short-term ozone
exposure on cardiovascular health was from animal toxicological studies demonstrating ozone-induced
impaired vascular and cardiac function, as well as changes in HR and HRV (U.S. EPA. 2013a). This
evidence was supported by a limited number of controlled human exposure studies in healthy adults
demonstrating changes in HRV, as well as in blood markers associated with an increase in coagulation,
systemic inflammation, and oxidative stress. Evidence of these effects in animals and humans was cited
as providing biological plausibility for the evidence from epidemiologic studies reporting positive
associations between short-term ozone exposure and cardiovascular-related mortality. However, there was
limited or no evidence from controlled human exposure or epidemiologic studies for short-term ozone
exposure and cardiovascular morbidity, such as effects related to HF, IHD and MI, arrhythmia and
cardiac arrest, or thromboembolic disease. The lack of evidence connecting the effects observed on
impaired vascular and cardiac function in animal toxicological studies and the association between
short-term ozone exposure and cardiovascular mortality observed in epidemiologic studies was a major
source of uncertainty in the 2013 Ozone ISA.
Animal toxicological studies published since the 2013 Ozone ISA provide generally consistent
evidence for impaired cardiac function and endothelial dysfunction, but limited or inconsistent evidence
for endpoints including indicators of arrhythmia and markers of oxidative stress and inflammation.
Additional controlled human exposure studies have been published in recent years, however the evidence
for an ozone-induced effect on cardiovascular endpoints is inconsistent. That is, no effect of ozone was
4-38
-------
reported from CHE studies of cardiac function, indicators of IHD (i.e., ST segment), endothelial
dysfunction, or HR, although some evidence from a small number of CHE studies indicates that ozone
exposure can result in changes in blood pressure, HRV, indicators of arrhythmia, markers of coagulation,
and inflammatory markers. In addition, the number of epidemiologic studies evaluating short-term ozone
exposure and cardiovascular health effects has grown somewhat, but overall remains limited and
continues to provide little, if any, evidence for associations with HF, IHD and MI, arrhythmia and cardiac
arrest, or thromboembolic disease. Recent epidemiologic evidence for associations between short-term
ozone exposure and cardiovascular mortality is limited, and the studies included in the 2013 Ozone ISA
continue to provide the strongest evidence for this association. Overall, many of the same limitations and
uncertainties that existed in the body of evidence in the 2013 Ozone ISA continue to exist. However, the
body of controlled human exposure studies evaluating short-term ozone exposure and cardiovascular
endpoints has grown, and when evaluated in the context of the controlled human exposure studies
available for the 2013 Ozone ISA, the evidence is less consistent and weaker, overall. Evidence published
since the completion of the 2013 IS A and its effect on judgments regarding the extent to which short-term
exposure to ozone causes cardiovascular effects is discussed in greater detail below.
Similar to the evidence in the 2013 Ozone ISA, there is evidence from some, but not all recent
animal toxicological studies for an increase in markers associated with systemic inflammation and
oxidative stress (Section 4.1.11.3) following short-term ozone exposure. The systemic inflammation
results are coherent with generally consistent evidence from epidemiologic panel studies demonstrating
increases in markers of systemic inflammation such as CRP following short-term ozone exposure
(Section 4.1.11.2 and Section 4.1.11.1. respectively). However, there is some evidence from a small
number of controlled human exposure studies examining the potential for increased markers of
inflammation and oxidative stress following short-term ozone exposure (Section 4.1.11.2). Additionally,
the newly available epidemiologic panel study did not observe an association between short-term ozone
concentrations and myeloperoxidase.
The 2013 Ozone ISA included evidence from animal toxicological studies for changes in cardiac
and endothelial function following short-term exposure to ozone. There is generally consistent evidence
from recent animal toxicological studies published since the last review demonstrating impaired cardiac
and endothelial function in rodents following short-term ozone exposure (Section 4.1.4.3 and
Section 4.1.6.3). However, coherence with studies in humans is lacking. A controlled human exposure
study in healthy individuals did not report ozone-induced changes in stroke volume or left ventricular
ejection time relative to FA. Moreover, multiple controlled human exposure studies in healthy subjects
found no evidence of an ozone-induced effect on measures of endothelial function such as FMD
following reactive hyperemia or pharmacological challenge (Section 4.1.6.2). In addition, results from
recent epidemiologic panel studies were inconsistent, with a limited number of studies reporting either
positive, negative, or null associations with short-term ozone concentrations (Section 4.1.6.1).
4-39
-------
In the last review, there was also a limited number of animal toxicological and controlled human
exposure studies demonstrating changes in HR and HRV. In the current review, there is inconsistent
evidence for changes in HR in animals (Section 4.1.9.3). and no additional evidence for changes in HR in
healthy adults from multiple controlled human exposure studies (Section 4.1.9.2'). With respect to HRV,
there is inconsistent evidence in both animal toxicological and controlled human exposure studies of
healthy adults (Section 4.1.9.2 and Section 4.1.9.3'). Similarly, recent epidemiologic panel studies
reported inconsistent associations between short-term exposure to ozone and both HR and HRV
(Section 4.1.9.1'). Moreover, although some but not all recent animal toxicological studies demonstrate
ozone-induced changes in blood pressure (Section 4.1.8.4') and changes in indicators of conduction
abnormalities in SH rats (Section 4.1.7.2'). there is again a lack of coherence with evidence in humans.
Multiple controlled human exposure studies reported little effect of short-term ozone exposure on
conduction abnormalities, and little evidence of an ozone-induced effect on blood pressure. Few
epidemiologic panel studies evaluated blood pressure, and the results were inconsistent.
In addition, a limited number of epidemiologic time-series and case-crossover studies conducted
in North America, Europe, or Australia and published since the last review report inconsistent results1.
With respect to the limited number of recent studies of hospital admissions and ED visits that analyzed
associations with heart failure, associations continued to be inconsistent. Studies conducted in the U.K.
and Arkansas did not observe associations for CHF alone or combined with hypertensive heart disease
and increases in ozone concentrations, but a pair of studies in St. Louis, MO reported a 5% increase in
either ED visits or hospital admissions associated with short-term exposure to ozone (Section 4.1.4.1').
Studies from Europe, Canada, and the U.S., several of which analyzed a large number of events per day in
multiple cities, consistently reported null or only small positive effect estimates (i.e., OR < 1.02) in
analyses of MI, including for STEMI and NSTEMI (Section 4.1.5.1'). One multicity study in Italy
reported a 5% increase in incident MI associated with an increase in 8-hour max ozone concentrations
during the warm season using a 0-1 day distributed lag. Similarly, inconsistent results were observed in
several studies that analyzed hospital admissions and ED visits for stroke and stroke subtypes in the U.S.,
Canada, and Europe (Section 4.1.12.1). Increases in out-of-hospital cardiac arrests associated with 8-hour
max or 24-hour avg increases in ozone concentrations were reported by a few case-crossover studies,
however analyses of other endpoints (e.g., dysrhythmia, arrhythmia, or atrial fibrillation) generally
reported null results (Section 4.1.7.1). In addition, increases in ED visits for hypertension of 11 to 15%
were observed among females in a study conducted in two Canadian cities during the warm season, and in
a study in Kaunas, Lithuania (Section 4.1.8.1). However, no association between ED visits for
hypertension and ozone concentration was observed in a time-series study in Arkansas.
The 2013 Ozone ISA concluded that there was adequate evidence that children and older adults
are at increased risk of ozone-related health effects (U.S. EPA. 2013a). No recent studies of short-term
ozone exposure and cardiovascular health effects contribute evidence to determine if children are at a
1 A list of considered studies conducted in other geographic locations is available via the HERO database.
4-40
-------
greater risk compared to other lifestages. A limited number of recent studies of short-term ozone exposure
and cardiovascular health effects have compared associations between different age groups, but do not
report consistent evidence that older adults are at increased risk compared to other lifestages
(Section 4.1.16.1V When considering pre-existing disease as a modifying factor, the 2013 Ozone ISA
concluded that there was adequate evidence for increased ozone-related health effects among individuals
with asthma (U.S. EPA. 2013a). A limited number of recent studies provides some evidence that
individuals with pre-existing diseases may be at greater risk of cardiovascular health effects associated
with short-term ozone exposure. These studies focus on specific cardiovascular and metabolic diseases
(e.g., previous CVD events, type 2 diabetes; Section 4.1.16.2).
Notably, there is a lack of coherence between cardiovascular effects when they are observed in
animals and corresponding effects in humans, particularly when examining the results of controlled
human exposure studies. This could be because a number of the animal toxicological studies were
performed in rodent disease models, while controlled human exposure studies generally include healthy
individuals. For example, evidence for changes in blood pressure were found in SH rats and in rats fed a
high fructose diet, and in a panel study that included individuals with pre-existing type 2 diabetes
(Hoffmann et al.. 2012). but practically no evidence of an effect on blood pressure was reported in
multiple controlled human exposure studies in generally healthy subjects. Thus, it is possible that if those
with underlying cardiovascular or metabolic disease were included in controlled human exposure studies,
results may have been different. That being said, controlled human exposure studies do not typically
include unhealthy or diseased individuals for ethical reasons, and therefore this represents an important
uncertainty to consider in interpreting the results of controlled human exposure studies.
In addition to underlying disease status, there are also substantial differences in exposure
concentrations between animal toxicological and controlled human exposure studies. Animal
toxicological studies generally expose rodents to 0.3 to 1 ppm, while CHE studies generally expose
humans to 0.07 and 0.3 ppm. Thus, additional animal toxicological studies conducted at lower
concentrations could help to reduce this uncertainty. In fact, there is evidence in SH rats that exposure to
0.2 ppm ozone results in no statistically significant effects on measures of cardiac electrophysiology,
while exposure to 0.8 ppm exposure results in statistically significant effects on these endpoints (Farrai et
al.. 2012). A caveat to this study, however, is that both concentrations increased sensitivity to the
arrhythmia-inducing drug aconitine (Farrai et al.. 2012). Nevertheless, additional studies in wild-type and
disease-model mice exposed to lower ozone concentrations would be greatly beneficial for future review.
Finally, in addition to disease status and exposure concentration, the lack of coherence between some
animal and human studies could be due to differences in physiology (e.g., rodents are obligate nose
breathers), differences in the duration and timing of exposure (e.g., rodents are exposed during the day,
during their resting cycle, while humans are exposed during the day when they are normally active), and
the temperature at which the exposure was conducted (Kahle et al.. 2015).
4-41
-------
When considered as a whole, the evidence is suggestive of, but not sufficient to infer, a
causal relationship between short-term exposure to ozone and cardiovascular effects. This
determination is different from the conclusion in the 2013 Ozone ISA. The evidence that supports this
change in the causality determination includes: (1) a growing body of controlled human exposure studies
providing less evidence for an effect of short-term ozone exposure and cardiovascular health endpoints;
(2) a paucity of evidence for more severe cardiovascular morbidity endpoints (i.e., HF, IHD and MI,
arrhythmia and cardiac arrest, and thromboembolic disease) to connect the evidence for impaired vascular
and cardiac function from animal toxicological studies with the evidence from epidemiologic studies of
cardiovascular mortality; and (3) uncertainties and limitations acknowledged in the 2013 Ozone ISA
(e.g., lack of control for potential confounding by copollutants in epidemiologic studies) remain in recent
evidence (Table 4-IV Although there exists some generally consistent evidence for a limited number of
ozone-induced cardiovascular endpoints in animal toxicological studies, there is a general lack of
coherence between these results and those in controlled human exposure and epidemiologic studies. Thus,
while some consistent results in animals and limited positive results in humans provide biological
plausibility for more serious endpoints such as mortality (Section 4.1.14). the underlying evidence
supporting biological plausibility is limited and thus, important uncertainties remain. Additional animal
toxicological studies at lower exposure concentrations in animal models of disease and epidemiologic
studies in populations with underlying disease would be useful to address these uncertainties.
4-42
-------
Table 4-1 Summary of evidence that is suggestive of, but not sufficient to infer,
a causal relationship between short-term ozone exposure and
cardiovascular effects.
Ozone
Concentrations
Rationale for Causality Associated with
Determination3 Key Evidence13 Key References'3 Effects0
Generally consistent
evidence from animal
toxicological studies at
relevant ozone
concentrations
Indicators of impaired heart function,
endothelial dysfunction
Tankerslev et al.
(2013)
Mclntosh-
Kastrinskv et al.
(20131
Section 4.1.4.3.
Section 4.1.6.3
-0.2 to 0.3 ppm
Limited or inconsistent
evidence from animal
toxicological studies at
relevant ozone
concentrations
ST-segment depression, changes in
indicators of cardiac electrophysiology or
potential arrhythmia in SH rats, changes in
changes in BP and HR or HRV, markers of
systemic inflammation and oxidative stress
Farrai et al. (2012).
Section 4.1.7.4,
Section 4.1.8.4,
Section 4.1.9.3
Section 4.1.11.3
0.8 but not at
0.2 ppm
Limited or inconsistent
evidence from controlled
human exposure studies
at relevant ozone
concentrations
No changes in a number of electrophysiology Rich et al. (2018) 0.07 but not
measures by ECG, but there was increased 0.12 ppm
probability of ventricular but not g 12 ppm
supraventricular ectopy couplets or runs
Change in T-wave alternans during the first Kusha et al. (2012) 0.12 ppm
5 min of exposure, but no change relative to
FA later in the exposure. The effect observed
in first 5-minutes is likely not meaningful
No meaningful changes in SBP and/or
Changes in HRV
Frampton et al. 0.07-0.3 ppm
(2015)
Barath et al. (2013)
Ariomandi et al.
(2015)
Rich et al. (2018)
Barath et al. (2013) 0.07-0.3 ppm
Rich etal. (2018)
Ariomandi et al.
(2015)
Markers of coagulation, systemic
inflammation and oxidative stress
Kahle etal. (2015) 0.1-0.3 ppm
Barath et al. (2013)
Ariomandi et al.
(2015)
Frampton et al.
(2015)
Section 4.1.11.2
4-43
-------
Table 4-1 (Continued): Summary of evidence that is suggestive of, but not
sufficient to infer, a causal relationship between short-term
ozone exposure and cardiovascular effects.
Rationale for Causality
Determination3
Key Evidence13
Key References'3
Ozone
Concentrations
Associated with
Effects0
Evidence of no effect
from controlled human
exposure studies at
relevant ozone
concentrations
Changes in stroke volume or left ventricular
ejection time
Frampton et al.
(2015)
0.1, 0.2 ppm
Changes in ST segment
Rich et al. (2018)
0.07, 0.12 ppm
Clinical indicators of endothelial dysfunction
Section 4.1.6.2
0.07-0.3 ppm
Changes in HR
Frampton et al.
(2015)
Barath et al. (2013)
Ariomandi et al.
(2015)
Rich et al. (2018)
Kusha et al. (2012)
0.07-0.3 ppm
Consistent evidence from
high-quality,
epidemiologic studies of
cardiovascular mortality
Sacks et al. (2012) 19.3-36.0 ppb
A number of studies evaluated in the 2013
Ozone ISA reported positive associations for Klemm et al (2011
cardiovascular mortality in all-year and
seasonal analyses. A more limited number of
recent studies continue to report positive
associations.
Section 4.1.14
Limited epidemiologic
evidence from multiple
studies of CVD hospital
admissions or ED visits
Generally null or inconsistent associations
(both negative and positive direction)
observed in studies of CVD hospital
admissions or ED visits limited by low ozone
concentrations (averaging <40 ppb), low
number of daily events in many studies, and
few multicity studies to allow for evaluation of
geographic heterogeneity. Although there
were a few exceptions, among studies
averaging more events per day (>1),
associations with heart failure, hypertension,
stroke, ischemic heart disease, and
dysrhythmia/atrial fibrillation were primarily
null or in the negative direction. More
consistent associations were reported by a
few studies for out-of-hospital cardiac arrest.
Pradeau et al.
(2015)
Raza et al.
(2014)
Mean: 20-40 ppb
75th: 27-50 ppb
Rosenthal et al.
(2013)
Section 4.1.4,
Section 4.1.5,
Section 4.1.7,
Section 4.1.8,
Section 4.1.12
Limited epidemiologic
evidence from panel
studies of CVD endpoints
Limited number of studies with generally
positive associations (ventricular
tachycardia, pulse amplitude, and myocardial
infarction) observed among populations with
or without pre-existing disease and without
any repeated endpoints evaluated.
Section 4.1.5.2,
Section 4.1.7.2
Mean:
23-40.56 ppb
Inconsistent
epidemiologic evidence
from multiple panel
studies of CVD endpoints
Generally null or inconsistent associations Section 4.1.6.1.
(e.g., heart rate variability, endothelial Section 4.1.8.2,
outcomes, coagulation markers, BP) Section 4.1.9.1.
observed among populations with or without Section 4.1.10.2
pre-existing disease; limited number of
studies evaluating the same endpoint; limited
number of subjects in some studies.
Mean: 22-41 ppb
4-44
-------
Table 4-1 (Continued): Summary of evidence that is suggestive of, but not
sufficient to infer, a causal relationship between short-term
ozone exposure and cardiovascular effects.
Ozone
Concentrations
Rationale for Causality Associated with
Determination3 Key Evidence13 Key References'3 Effects0
Limited epidemiologic
The maqnitude of ozone associations Section 4.1.15
evidence from
remains relatively unchanged, but in some
copollutant models
cases with wider confidence intervals in a
provides some support
limited number of studies evaluating
for an independent ozone
copollutant models, including PM2.5 and
association
other gaseous pollutants.
When reported, correlations with PM2.5 or
gaseous copollutants were primarily in the
low to moderate range (r< 0.7).
Consistent positive associations observed in
studies of short-term ozone exposure and
mortality, although there is only some
evidence from a small number of studies of a
relationship between short-term ozone
exposure and CVD morbidity (e.g., HF, IHD
and Ml, arrhythmia and cardiac arrest, and
stroke) in epidemiologic and controlled
human exposure studies.
Uncertainty regarding Multicity U.S. studies demonstrate city-to-city
geographic heterogeneity and regional heterogeneity in ozone-CVD ED
in ozone associations visit and hospital admission associations.
Evidence supports that a combination of
factors, including composition and exposure
factors may contribute to the observed
heterogeneity.
aBased on aspects considered in judgments of causality and weight of evidence in causal framework in Table I and Table II of the
Preamble to the ISAs (U.S. EPA. 2015V
bDescribes the key evidence and references, supporting or contradicting, contributing most heavily to causality determination and,
where applicable, to uncertainties or inconsistencies. References to earlier sections indicate where the full body of evidence is
described.
°Describes the ozone concentrations with which the evidence is substantiated.
Uncertainty due to limited
coherence between CVD
morbidity and CVD
mortality
Section 4.1.5.1.
Section 4.1.7.1.
Section 4.1.12.1
Section 4.1.13.1
4-45
-------
4.2 Long-Term Ozone Exposure and Cardiovascular Health
Effects
4.2.1 Introduction, Summary from the 2013 Ozone ISA, and Scope for Current
Review
The 2013 Ozone ISA concluded that evidence was suggestive of a causal relationship between
long-term exposures to ozone and cardiovascular effects. In the last review, a small number of
well-conducted animal toxicological studies provided evidence for ozone-enhanced atherosclerosis or
ischemic/reperfusion injury. There was also evidence that long-term exposure to ozone resulted in
systemic inflammation and oxidative stress. This evidence was in addition to a small number of
epidemiologic studies reporting an association between long-term exposure to ozone and cardiovascular
disease-related biomarkers. Of note, the only epidemiologic study to investigate the relationship between
long-term ozone exposure and cardiovascular mortality did not observe a positive association. A key
uncertainly from the last review was the mechanism by which ozone inhalation may result in systemic
effects. However, there was some evidence from a small number of studies in the 2013 Ozone ISA that
activation of LOX-1 by ozone-oxidized lipids and proteins could result in changes in genes involved in
proteolysis, thrombosis, and vasoconstriction.
The subsections below provide an evaluation of the most policy-relevant scientific evidence
relating long-term ozone exposure to cardiovascular health effects. These sections focus on studies
published since the completion of the 2013 Ozone ISA, and emphasis is placed on those studies that
address uncertainties remaining from the last review. Overall, a limited number of animal toxicological
and epidemiologic studies contribute some new evidence characterizing the relationship between
long-term ozone exposure and cardiovascular health effects. There is some emerging epidemiologic
evidence that long-term ozone exposure may be associated with blood pressure changes or hypertension
among different lifestages or those with pre-existing disease. With respect to the toxicological evidence,
there was some evidence for inflammation, oxidative stress, and impaired cardiac contractility in rodents
following long-term ozone exposure. Overall, however, many of the uncertainties identified in the
previous review remain.
4.2.1.1 Population, Exposure, Comparison, Outcome, and Study Design (PECOS)
Tool
The scope of this section is defined by a scoping tool that generally defines the relevant
Population, Exposure, Comparison, Outcome, and Study Design (PECOS). The PECOS tool defines the
4-46
-------
parameters and provides a framework to help identify the relevant evidence in the literature to inform the
ISA. Because the 2013 Ozone ISA concluded that there is evidence to suggest a causal relationship
between long-term ozone exposure and cardiovascular health effects, the epidemiologic studies evaluated
are less limited in scope and not targeted towards specific study locations, as reflected in the PECOS tool.
The studies evaluated and subsequently discussed within this section were identified using the following
PECOS tool:
Experimental Studies:
• Population: Study population of any animal toxicological study of mammals at any lifestage
• Exposure: Long-term (on the order of months to years) inhalation exposure to relevant ozone
concentrations (i.e., <2 ppm)
• Comparison: Appropriate comparison group exposed to a negative control (i.e., clean air or
filtered air control)
• Outcome: Cardiovascular effects
• Study Design: In vivo chronic-duration, subchronic-duration, or repeated-dose toxicity studies in
mammals or immunotoxicity studies
Epidemiologic Studies:
• Population: Any population, including populations or lifestages that might be at increased risk
• Exposure: Long-term ambient concentration of ozone
• Comparison: Per unit increase (in ppb)
• Outcome: Change in risk (incidence/prevalence) of a cardiovascular effect
• Study Design: Epidemiologic studies consisting of cohort, case-control studies, and
cross-sectional studies with appropriate timing of exposure for the health endpoint of interest
4.2.2 Biological Plausibility
This subsection describes the biological pathways that potentially underlie cardiovascular health
effects resulting from long-term inhalation exposure to ozone. Figure 4-6 graphically depicts these
proposed pathways as a continuum of pathophysiological responses—connected by arrows—that may
ultimately lead to the apical cardiovascular events observed in epidemiologic studies associated with
long-term exposure. This discussion of "how" long-term exposure to ozone may lead to these
cardiovascular events also provides biological plausibility for the epidemiologic results reported later in
this Appendix. In addition, most studies cited in this subsection are discussed in greater detail throughout
this Appendix. Note that the structure of the biological plausibility sections and the role of biological
plausibility in contributing to the weight-of-evidence analysis used in the 2020 Ozone ISA are discussed
in Section IS.4.2.
4-47
-------
Modulation of
the Autonomic
Nervous System
(e.g., HRV, HR)
Long-term
Ozone
Exposure
Respiratory
Tract
inflammation/
Oxidative
Stress
Mortality
Systemic
Inflammation/
Oxidative Stress
Impaired
Vascular
Function
Note; The boxes above represent the effects for which there is experimental or epidemiologic evidence related to ozone exposure,
and the arrows indicate a proposed relationship between those effects. Solid arrows denote evidence of essentiality as provided, for
example, by an inhibitor of the pathway or a genetic knockout model used in an experimental study involving ozone exposure.
Shading around multiple boxes is used to denote a grouping of these effects. Arrows may connect individual boxes, groupings of
boxes, and individual boxes within groupings of boxes. Progression of effects is generally depicted from left to right and color-coded
(gray, exposure; green, initial effect; blue, intermediate effect; orange, effect at the population level or a key clinical effect). Here,
population level effects generally reflect results of epidemiologic studies. When there are gaps in the evidence, there are
complementary gaps in the figure and the accompanying text below. The structure of the biological plausibility sections and the role
of biological plausibility in contributing to the weight-of-evidence analysis used in the 2020 Ozone ISA are discussed in
Section IS.'4.2.
Figure 4-6 Potential biological pathways for cardiovascular effects following
long-term exposure to ozone.
There is evidence from epidemiologic studies for cardiovascular-related mortality following
long-term exposure to ozone. However, when attempting to construct a biologically plausible pathway
that could result in cardiovascular-related mortality following long-term exposure to ozone, there are
important gaps in the health evidence (Figure 4-6). Specifically, there is no evidence from epidemiologic
studies of an association between long-term exposure to ozone and IHD or MI, HF, arrhythmia, or
thromboembolic disease. More information on this pathway and the important gaps that exist are
described below.
Long-term inhalation exposure to ozone may result in respiratory tract inflammation and
oxidative stress (Appendix 3). In general, inflammatory mediators, such as cytokines produced in the
respiratory tract, have the potential to enter the circulatory system where they may cause distal
pathophysiological responses that could lead to overt cardiovascular disease. In addition, release of
inflammatory mediators into the circulation, such as monocyte chemoattractant protein-1 (MCP-1), can
result in the recruitment of additional inflammatory cells, and thus amplify the initial inflammatory
4-48
-------
response. Thus, it is important to note that there is evidence from long-term experimental studies in
animals (Miller et al.. 2016; Perepu et al.. 2012; Sethi et al.. 2012) demonstrating an increase in cytokines,
and/or oxidative stress markers in the circulatory system following long-term ozone exposure. The release
of cytokines like IL-6 into the circulation can stimulate the liver to release inflammatory proteins and
coagulation factors that can alter hemostasis and increase the potential for thrombosis (Tanaka et al..
2014). Thus, it is important to note animal toxicological studies demonstrating changes in these types of
coagulation factors following long-term ozone exposure (Gordon et al.. 2014; U.S. EPA. 2013a). These
changes may alter the balance between pro- and anticoagulation proteins, and therefore, increase the
potential for thrombosis, which may then promote IHD, stroke, or thromboembolic disease elsewhere in
the body. However, there is no epidemiologic evidence of an association between long-term exposure to
ozone and IHD, stroke, or thromboembolic disease, and thus, considerable uncertainty in the potential
pathway leading to mortality.
Systemic inflammation has the potential to result in impaired vascular function, a systemic
pathological condition characterized by the altered production of vasoconstrictors and vasodilators, which
over time promotes plaque formation leading to atherosclerosis. Specifically, vascular dysfunction is
often accompanied by endothelial cell expression of adhesion molecules and release of chemoattractants
that recruit inflammatory cells. Macrophages may then internalize circulating lipids, leading to the
formation of foam cells: a hallmark of atherosclerotic lesions. Overtime, these atherosclerotic lesions
may become calcified, and this often leads to arteriole stiffening and promotion of IHD or stroke.
Importantly, evidence for changes in molecular markers associated with impaired vascular function
following ozone exposure are found in an experimental study in animals from the 2013 Ozone ISA ITJ.S.
EPA (2013a). pg. 7-38], This is in agreement with results from an epidemiologic study reporting a
positive association between long-term exposure to ozone and increased CIMT (Breton et al.. 2012). as
well as with animal toxicological results indicating changes in caveolin 1 and caveolin 3 (Sethi et al..
2012). two molecular markers possibly associated with the development of atherosclerosis. Moreover,
one study in the 2013 Ozone ISA reported enhanced aortic atherosclerotic lesions in mice following
long-term ozone exposure ITJ.S. EPA (2013a). pg. 7-38], However, considerable uncertainty remains in
how long-term ozone exposure may lead to mortality given that there are few epidemiologic studies to
provide evidence of an association between long-term exposure to ozone and other cardiovascular
endpoints such as IHD, stroke, or thromboembolic disease, and these studies generally report null effects.
Thus, how these earlier events could lead to mortality remains unclear.
In addition to long-term ozone exposure leading to cardiovascular disease through inflammatory
pathways, there is also evidence that long-term exposure to ozone could lead to cardiovascular disease
through modulation of the autonomic nervous system. Studies in animals showed modulation of
autonomic function (as evidenced by changes in HR) following long-term ozone exposure (Gordon et al..
2013). Moreover, there is epidemiologic evidence of positive associations between long-term exposure to
ozone and increases in BP and hypertension (Cole-Hunter et al.. 2018; Coogan et al.. 2017; Yang et al..
2017; Dong et al.. 2014).
4-49
-------
When considering the available evidence, important uncertainties remain in potential pathways
connecting long-term exposure to ozone to cardiovascular-related mortality. That is, while there is some
evidence for a number of early and intermediate cardiovascular-related effects from animal toxicological
and epidemiologic studies, there is no epidemiologic evidence of associations between long-term
exposure to ozone and outcomes that could directly result in death, such as IHD, stroke, or arrhythmia.
4.2.3 Ischemic Heart Disease (IHD) and Associated Cardiovascular Effects
4.2.3.1 Epidemiologic Studies
No studies examining long-term ozone exposure and IHD were included in the 2013 Ozone ISA.
A recent national cohort study conducted in England observed a null association between long-term ozone
exposure and Mis (Atkinson ct al.. 2013) and a cohort study conducted in South Korea observed an
inverse association HCim et al. (2017); Table 4-341.
4.2.4 Atherosclerosis
4.2.4.1 Epidemiologic Studies
No studies examining long-term ozone exposure and atherosclerosis were included in the 2013
Ozone ISA. A recent U.S. cohort study evaluated long-term ozone exposure, averaged in early life (ages
0-5 years), during elementary school (ages 6-12 years) and during the first 20 years of life (Breton et al..
2012). These authors observed positive associations between ozone averaged over all three exposure
windows and increases in CIMT measured in southern California college students (Table 4-35). These
results were robust to the inclusion of PM2 5, PM10, and NO2 in copollutant models.
4.2.4.2 Animal Toxicological Studies
The 2013 Ozone ISA presented evidence that long-term exposure to ozone in ApoE-/- mice
resulted in enhanced aortic atherosclerotic lesions when compared with filtered air exposure ITJ.S. EPA
(2013a). pg. 7-38], The last review also noted that activation of lectin-like oxidized-low density
lipoprotein receptor-1 (LOX-1) could have a role in vascular pathology associated with atherosclerosis
rU.S. EPA (2013a). pg. 7-38], Since the 2013 Ozone ISA, there is inconsistent evidence of an effect of
4-50
-------
long-term ozone exposure (4-17 weeks) on potential markers of atherosclerosis (Table 4-36).
Specifically:
• Sethi et al. (2012) reported that long-term exposure to ozone (0.8 ppm) decreased caveolin 1 and
increased caveolin 3 expression at 28 days relative to control animals. At 56 days, caveolin 1
expression and caveolin 3 expression decreased, setting up the potential for a proapoptotic and
atherosclerotic environment (p < 0.05). However, Gordon et al. (2013) reported that 17-week
ozone exposure (0.8 ppm) had no effect on LOX-1, caveolin 1, or RAGE gene expression in the
aortas of younger or older rats.
4.2.5 Heart Failure and Impaired Heart Function
4.2.5.1 Epidemiologic Studies
No studies examining long-term ozone exposure and heart failure were included in the 2013
Ozone ISA. A recent national cohort study conducted in England observed an inverse association between
long-term ozone exposure and heart failure (Atkinson et al.. 2013). and a cohort study conducted in South
Korea observed an inverse association rKim et al. (2017); Table 4-371.
4.2.5.2 Animal Toxicological Studies
In the 2013 Ozone ISA, there was evidence from an animal toxicological study that long-term
ozone exposure decreased LVDP, rate of pressure development, and rate of change of pressure in isolated
perfused rat hearts ITJ.S. EPA (2013a). pg. 7-39], Similarly, two recent studies from the same laboratory
that contributed evidence to the 2013 Ozone ISA reported a decrease in LVDP following long-term
exposure (4-8 weeks) to ozone (0.8 ppm) in isolated perfused rat hearts \p < 0.05; Perepu et al. (2012);
Sethi et al. (2012); Table 4-381. Moreover, Perepu et al. (2012) also reported a decrease in the rate of
pressure development and a decrease in pressure decay in these hearts (p < 0.05). This decrease in
pressure decay is consistent with impaired diastolic function (i.e., cardiac filling) and is consistent with
additional results from this study indicating an increase in left ventricular end diastolic pressure. Thus,
both studies demonstrate that long-term ozone exposure can result in abnormal cardiac function.
However, these studies were all conducted by the same laboratory and therefore, there is uncertainty with
respect to the broad applicability of the results.
4-51
-------
4.2.6
Vascular Function
4.2.6.1 Animal Toxicological Studies
The 2013 Ozone ISA presented evidence from an animal toxicological study of an increase in
ET-1, ET-1 receptor, and eNOS mRNA in rat aortas following long-term exposure to ozone ITJ.S. EPA
(2013a). pg. 7-38], However, a more recent study reported no change in eNOS/iNOS or ET-1 mRNA
expression in adult or senescent rat aorta tissue following long-term ozone (0.8 ppm) exposure [17 weeks;
Gordon et al. (2013); Table 4-391. Thus, there remains limited evidence from animal toxicological studies
that long-term exposure to ozone may result in an increase in markers that promote vasoconstriction.
4.2.7 Cardiac Depolarization, Repolarization, Arrhythmia, and Arrest
4.2.7.1 Epidemiologic Studies
No studies examining long-term ozone exposure and arrhythmia were included in the 2013 Ozone
ISA. A recent national cohort study conducted in England observed a null association between long-term
ozone exposure and arrhythmia (Atkinson et al.. 2013).
4.2.8 Blood Pressure Changes and Hypertension
4.2.8.1 Epidemiologic Studies
At the time of the 2013 Ozone ISA, one study was available that investigated the relationship
between long-term ozone exposure and blood pressure. Chuang etal. (2011) observed increases in both
systolic and diastolic blood pressure associated with ozone concentrations among older adults in Taiwan,
although these increases were attenuated in models that included copollutants. A number of recent
studies, conducted mainly in Asia, observed inconsistent results between long-term ozone exposure and
blood pressure or hypertension among healthy adults (Table 4-40). There is some emerging evidence that
long-term ozone exposure may be associated with changes in blood pressure or hypertension among
different lifestages or those with pre-existing disease. Specifically:
4-52
-------
• A U.S. cohort study observed positive associations between long-term ozone exposure and
incident hypertension among black women (Coogan et al.. 2017). These associations were robust
to copollutant adjustment with PM2 5 and somewhat attenuated, though still positive, with
adjustment for NO2. Similarly, cross-sectional studies conducted in China observed positive
associations between long-term ozone concentrations and prevalent prehypertension (Yang et al..
2017) and hypertension (Dong et al.. 2015; Dong et al.. 2014; Dong et al.. 2013b; Zhao et al..
2013).
• A cohort study conducted in Spain (Cole-Hunter et al.. 2018) observed positive associations
between both systolic and diastolic blood pressure and long-term ozone exposure; the
associations were robust to the inclusion of PM10 in a copollutant model. Similarly,
cross-sectional studies conducted in China observed positive associations with both systolic and
diastolic blood pressure (Yang etal.. 2017; Liu et al.. 2016; Dong et al.. 2013b; Zhao et al.. 2013;
Chuang etal.. 2011). In some instances, this effect was larger for systolic, compared to diastolic,
blood pressure (Yang etal.. 2017; Liu et al.. 2016).
• A cross-sectional study conducted in China (Yang et al.. 2017) observed stronger associations
between long-term ozone exposure and prevalent prehypertension and blood pressure among
women compared with the entire population. In contrast, a separate cross-sectional study
conducted in China reported stronger associations between long-term ozone exposure and
hypertension and blood pressure among men compared to women (Dong et al.. 2013b).
• A cross-sectional study conducted in China (Yang et al.. 2017) observed stronger associations
between long-term ozone exposure and prevalent prehypertension among older adults (>55 years)
compared with younger adults (<35 years). In an additional cross-sectional study conducted in
China, Dong et al. (2013b) observed stronger associations between long-term ozone exposure and
hypertension in both older adults (>65 years) and younger adults (<55 years) compared to adults
aged 55-64 years. Similarly, Yang et al. (2017) observed stronger associations between
long-term ozone exposure and blood pressure in younger (<35 years) compared to older
(>55 years) adults.
• Zhao et al. (2013) reported stronger associations between long-term ozone exposure and
hypertension and blood pressure among overweight and obese adults, compared to normal weight
adults. This trend was especially strong among men, and less apparent in women. Similarly, Dong
et al. (2015) observed a higher magnitude of effect for both hypertension and blood pressure
among overweight and obese children compared with normal-weight children, although no
difference was observed between boys and girls.
• Dong et al. (2014) reported positive associations between long-term ozone exposure and
hypertension and blood pressure in children and observed stronger associations among children
that had never been breastfed. In a related analysis (Dong et al.. 2015). they observed stronger
associations in overweight and obese children compared to normal-weight children.
4.2.8.2 Animal Toxicological Studies
In the 2013 Ozone ISA, no studies examined the relationship between long-term exposure to
ozone and changes in BP. Recently, Gordon et al. (2013) reported that long-term exposure (17 weeks) to
ozone (0.8 ppm) did not result in changes in SBP or DBP in adult or senescent rats (Table 4-41). Thus,
there continues to be no evidence from animal toxicological studies that long-term exposure to ozone can
result in changes in BP.
4-53
-------
4.2.9
Heart Rate and Heart Rate Variability
4.2.9.1 Epidemiologic Studies
No studies examining long-term ozone exposure and heart rate were included in the 2013 Ozone
ISA. A recent cohort study observed positive associations between annual average ozone concentrations
and increases in heart rate in a Spanish population (Cole-Hunter et al.. 2018). These associations were
robust to the inclusion of PMio in a copollutant model.
4.2.9.2 Animal Toxicological Studies
No animal toxicological studies examining the relationship between long-term exposure to ozone
and HR or HRV were included in the 2013 Ozone ISA. Recently, Gordon et al. (2013) reported that
long-term exposure (17 weeks) to ozone (0.8 ppm) did not result in changes in HR in adult or senescent
rats. However, in an additional study using a different exposure protocol (Table 4-42). this group did find
an increase in HR following long-term episodic exposure (13 weeks) to ozone (1.0 ppm; p < 0.05) in
adult or senescent rats (Gordon et al.. 2014). Overall, the evidence for an effect of long-term exposure to
ozone on HR remains limited. There were no studies examining the relationship between long-term
exposure to ozone and HRV.
4.2.10 Coagulation
4.2.10.1 Animal Toxicological Studies
The 2013 Ozone ISA presented some evidence that long-term exposure to ozone resulted in
changes in factors involved in coagulation, such as tissue plasminogen activator, plasminogen activator
inhibitor-1, and von Willebrand factor ITJ.S. EPA (2013a). pg. 7-8], Since the 2013 Ozone ISA was
published, Gordon et al. (2013) has reported that long-term exposure (17 weeks) to ozone (0.8 ppm)
results in small changes in aortic mRNA levels of TF (p < 0.05), but not tPA, vWF, thrombomodulin, and
other mRNA markers of coagulation in adult or senescent rats. These authors also report no effect of
long-term ozone exposure on platelet levels in blood in adult or senescent rats. Overall, there is limited
evidence from a small number of animal toxicological studies that long-term exposure to ozone can result
in changes in mRNA levels of coagulation factors (Table 4-43).
4-54
-------
4.2.11
Systemic Inflammation and Oxidative Stress
4.2.11.1 Epidemiologic Studies
The majority of studies evaluating long-term ozone exposure and cardiovascular outcomes
included in the 2013 Ozone ISA assessed cardiovascular disease-related biomarkers. The studies used
annual or multiyear averages of air monitoring data for exposure assessment and reported generally null
effects with common biomarkers, including CRP, fibrinogen, and IL-6. A limited number of recent
studies provide evidence that is generally consistent with the evidence included in the 2013 Ozone ISA.
Specifically:
• A cohort study of midlife, multiethnic women conducted in the U.S. (Green et al.. 2015) observed
positive associations with factor VIIc and hs-CRP.
• Cross-sectional studies conducted in Germany (Pilz et al.. 2018) and Taiwan (Chuang et al..
2011) reported null or negative associations with CRP and IL-6, respectively. Chuang et al.
(2011) observed positive associations between long-term ozone exposure and increases in
neutrophils and small changes in hemoglobin Ale.
4.2.11.2 Animal Toxicological Studies
In the 2013 Ozone ISA, there was evidence that long-term exposure to ozone resulted in
increased levels of TNF-a while decreasing the anti-inflammatory cytokine IL-10 ITJ.S. EPA (2013a).
pg. 7-39], In addition, there was evidence that long-term exposure to ozone decreased SOD enzyme
activity and increased levels of malondialdehyde. Recent studies provide some evidence that long-term
exposure (4-17 weeks) to ozone can result in an increase in markers of inflammation and oxidative stress
(Table 4-44). Specifically:
• The same laboratory cited in the 2013 Ozone ISA reported that, long-term exposure of S-D rats to
ozone resulted in an increase in myocardial production of TNF-a \p < 0.05; Perepu et al. (2012);
Sethi et al. (2012)1. This laboratory (Perepu et al.. 2012) also reported a decrease in the
anti-inflammatory cytokine IL-10 following long-term exposure to ozone (0.8 ppm).
• In rats, Miller et al. (2016) also reported an increase in serum levels of IL-4, IL-10, and IFN-y,
but no change in IL-1 or TNF-a following long-term ozone (1 ppm) exposure.
• Notably, some studies also found that long-term exposure to ozone (0.8, 1.0 ppm) resulted in no
appreciable changes in other inflammatory markers, including TNF-a, IL-1 (Miller et al.. 2016).
and total lymphocytes (Gordon et al.. 2013).
With respect to markers of oxidative stress, there is some evidence from a small number of
studies that long-term exposure to ozone can result in markers of oxidative stress. That is:
4-55
-------
• In rats, Sethi et al. (2012) and Perepu et al. (2012) reported a decrease in SOD activity (p < 0.05)
following long-term ozone (0.8 ppm) exposure. Perepu et al. (2012) also reported an increase in
lipid peroxidation.
• However, Gordon et al. (2013) reported no appreciable change in HO-1 levels following
long-term ozone (0.8 ppm) exposure in rats.
4.2.12 Stroke and Associated Cardiovascular Effects
4.2.12.1 Epidemiologic Studies
No studies examining long-term ozone exposure and stroke or other cerebrovascular outcomes
were included in the 2013 Ozone ISA. A recent national cohort study conducted in England observed null
associations between long-term ozone exposure and both stroke and cerebrovascular disease (Atkinson et
al.. 2013). In addition, several recent publications report results from a cross-sectional study conducted in
33 Chinese communities, noting positive associations between long-term ozone exposure and stroke
(Dong et al.. 2013a). When stratified by obesity status, positive associations were observed between
long-term ozone exposure and stroke for adults that were overweight or obese, and null associations for
adults with normal weight (Oin et al.. 2015). These studies are characterized in Table 4-45.
4.2.13 Other Cardiovascular Endpoints
4.2.13.1 Pulmonary Embolism
No studies examining long-term ozone exposure and heart rate were included in the 2013 Ozone
ISA. A recent case-control study conducted in Italy (Spiezia etal.. 2014) reported negative associations
between monthly average ozone concentrations and unprovoked acute isolated pulmonary embolism.
4.2.13.2 Erectile Dysfunction Incidence
No studies examining long-term ozone exposure and erectile dysfunction were included in the
2013 Ozone ISA. A recent U.S. nationwide study in a cohort of older men (Tallon et al.. 2017) observed
positive (though imprecise) associations between self-reported incident erectile dysfunction and long-term
warm-season ozone exposure averaged over 1 to 7 years.
4-56
-------
4.2.14 Aggregate Cardiovascular Disease
4.2.14.1 Epidemiologic Studies
No studies examining long-term ozone exposure and aggregate endpoints related to
cardiovascular disease (i.e., different cardiovascular endpoints grouped into one category and considered
together) were included in the 2013 Ozone ISA. A recent national cohort study conducted in England
observed null associations between long-term ozone exposure and cardiovascular disease (Atkinson et al..
2013). In addition, several recent publications report results from a cross-sectional study conducted in
33 Chinese communities, noting positive associations between long-term ozone exposure and
cardiovascular disease, although when stratified by sex, a positive association was only observed for
males (Dong et al.. 2013a). When stratified by obesity status, positive associations were observed
between long-term ozone exposure and cardiovascular disease for adults that were obese, and null
associations were observed for normal-weight or overweight adults (Qin et al.. 2015). In females, the
association was positive among those with higher BMIs (i.e., >25 kg/m2) and negative among those with
lower BMIs. These studies are characterized in Table 4-46.
4.2.15 Cardiovascular Mortality
Recent cohort studies extend the body of evidence for the relationship between long-term ozone
exposure and cardiovascular-related mortality. The 2013 Ozone ISA noted inconsistent evidence for
cardiopulmonary mortality, and there was some evidence from an analysis of the ACS cohort for the
association between long-term ozone exposure and cardiovascular mortality (Jerrett et al.. 2009). Recent
analyses from the ACS cohort in the U.S. and the CanCHEC cohort in Canada provide consistent
evidence for positive associations between long-term ozone exposure and cardiovascular and IHD
mortality, as well as mortality due to diabetes or cardiometabolic diseases. Associations with mortality
due to cerebrovascular disease (e.g., stroke) were less consistent, and generally yielded closer-to-the-null
values. Other recent studies conducted in Europe and Asia report null or negative associations. Recent
studies used a variety of fixed-site (i.e., monitors), models (e.g., CMAQ, dispersion models) and hybrid
methods (combining fixed-site and model techniques) to measure or estimate ozone concentrations for
use in assigning long-term ozone exposure in epidemiologic studies (Appendix 2. Section 2.3). The
differences in the way exposure to ozone was assessed do not explain the heterogeneity in the observed
associations. The results from studies evaluating long-term ozone exposure and cardiovascular mortality
are presented in Figure 4-7. Overall, there is increased evidence that long-term ozone exposure is
associated with cardiovascular mortality compared to the evidence included in the 2013 Ozone ISA.
Specifically:
4-57
-------
The strongest evidence for an association between long-term ozone exposure and cardiovascular
mortality comes from nationwide analyses of the ACS cohort, demonstrating positive associations
with cardiovascular mortality (Turner et al.. 2016; Jerrett et al.. 2013; Jerrett et al.. 2009). IHD
mortality (Jerrett et al.. 2013). cerebrovascular disease mortality (Turner et al.. 2016). and
mortality due to dysrhythmia and heart failure (Turner et al.. 2016).
Several recent analyses of the CanCHEC cohort in Canada provide consistent evidence for a
positive association between long-term ozone exposure and cardiovascular and IHD mortality
(Cakmak et al.. 2018; Weichenthal et al.. 2017; Cakmak et al.. 2016; Crouse et al.. 2015).
Cohort studies conducted in France (Bentaveb et al.. 2015). the U.K. (Carey et al.. 2013). and
South Korea (Kim et al.. 2017) report negative associations between long-term ozone exposure
and cardiovascular mortality.
Several recent studies conducted in the U.S. and Canada provide limited and inconsistent
evidence for an association between long-term ozone exposure and mortality due to
cerebrovascular disease (Figure 4-7).
A limited body of evidence demonstrates positive associations between long-term ozone exposure
and mortality from diabetes and cardiometabolic diseases (Turner et al.. 2016; Crouse et al..
2015).
4-58
-------
Reference Cohort
Jerrett et al. 2009 ACS
TTurner etal. 2016 ACS
TJerrett et al. 2013 ACS
TCrouse etal. 2015 CanCHEC
TCakmak et al. 2016 CanCHEC
TWeichenthal etal. 2017
TBentayeb et a I. 2015
TCarey et al. 2013
TKimetal. 2017
Jerrett et al. 2009
TTurner etal. 2016
IJerrett et al. 2013
TCrouse etal. 2015
TCakmak et al. 2016
TCakmak et al. 2018
TTurner etal. 2016
TJerrett et al. 2013
TCrouse etal. 2015
TCakmak et al. 2016
TTurner etal. 2016
TCrouse etal. 2015
Notes
Circulatory Disease
Cardiovascular Disease
Base Model
Adj. for Climate Zone
CanCHEC
Gazel
English Medical Practice
NHIS-NSC
ACS
ACS
ACS
CanCHEC
CanCHEC
CanCHEC
ACS
ACS
CanCHEC
CanCHEC
ACS
CanCHEC
Base Model
Adj. for Climate Zone
Base Model
Adj. for Climate Zone
Diabetes
Cardiometabolic Disease
Years
1982-2000
1982-2004
1982-2000
1991-2006
1991-2006
1991-2011
1989-2013
2003-2007
2007-2013
1982-2000
1982-2004
1982-2000
1991-2006
1991-2006
Mean
57.5
38.2
50.35
39.6
14.3-40.9
38.29
40.5
25.85
19.93
1982-2004
1982-2000
1991-2006
1991-2006
1982-2004
1991-2006
57.5
38.2
50.35
39.6
14.3-40.9
1991-2011 15.0-43.0
38.2
50.35
39.6
14.3-40.9
38.2
39.6
Outcome
Cardiovascular Mortality
IHD Mortality
TTurner etal. 2016 ACS
1982-2004 38.2
CBVD Mortality
Diabetes Mortality
Other CV Mortality
-f
1 1.25 1.5 1.75
Hazard Ratio (95% CI)
ACS = American Cancer Society; CanCHEC = Canadian Census Health and Environment Cohort; CBVD = cerebrovascular
disease; CV = cardiovascular; IHD = ischemic heart disease; NHIS-NSC = National Health Insurance Service—National Sample
Cohort.
Note: tStudies published since the 2013 Ozone ISA. Associations are presented per 10-ppb increase in pollutant concentration.
Circles represent point estimates; horizontal lines represent 95% confidence intervals for ozone. Black text and circles represent
evidence included in the 2013 Ozone ISA; red text and circles represent recent evidence not considered in previous ISAs or
AQCDs.
Figure 4-7 Associations between long-term exposure to ozone and
cardiovascular mortality in recent cohort studies.
4.2.16 Potential Copollutant Confounding of the Ozone-Cardiovascular
Disease (CVD) Relationship
The evaluation of potential confounding effects of copollutants on the relationship between
long-term ozone exposure and cardiovascular effects allows for examination of whether ozone risks are
changed in copollutant models. In the 2013 Ozone ISA, Jerrett et al. (2009) reported associations with
cardiovascular mortality that were attenuated, changing from positive to negative, after adjustment for
PM2 5 concentrations. Recent studies examined the potential for copollutant confounding by evaluating
4-59
-------
copollutant models that included PM2 5, PM10, and NO2. These recent studies address a previously
identified data gap by informing the extent to which effects associated with long-term ozone exposure are
independent of coexposure to correlated copollutants in long-term analyses.
• Several recent studies of cardiovascular mortality observe that the association between long-term
ozone exposure and cardiovascular mortality is attenuated in models that also include PM2 5
(Figure 4-8). consistent with the results presented in Jerrett et al. (2009). Whereas the associations
were attenuated and changed from positive to negative in Jerrett et al. (2009). the associations
between long-term ozone exposure and cardiovascular mortality are attenuated but remain
positive after adjusting for PM25 in several recent studies. Similarly, the inclusion of PM2 5 in
copollutant models had little effect on the association between long-term ozone exposure and
markers of inflammation in a cohort of multiethnic women (Green et al.. 2015).
• When examining other cardiovascular endpoints, several recent studies report that the
associations with long-term ozone exposure were robust to the inclusion of PM2 5 or PM10 in
copollutant models. The association between long-term ozone exposure and incident hypertension
in a cohort of black women was relatively unchanged when PM2 5 was included in copollutant
models (Coogan et al.. 2017). Adding PM2 5 to the model had little impact on the association with
cardiovascular health effects across studies. When PM10 was included in copollutant models, it
had little effect on the association between long-term ozone exposure and MI, stroke, arrhythmia
or heart failure (Atkinson et al.. 2013). measures of blood pressure or heart rate (Cole-Hunter et
al.. 2018). or changes in CIMT (Breton et al.. 2012).
• When NO2 was included in copollutant models, it had little effect on the association between
long-term ozone exposure and MI, stroke, arrhythmia or heart failure (Atkinson et al.. 2013). or
changes in CIMT (Breton et al.. 2012). The association between long-term ozone exposure and
incident hypertension in a cohort of black women was attenuated, but remained positive, when
NO2 was included in copollutant models (Coogan et al.. 2017).
4-60
-------
Jerrett eta I. 2009
ITurner eta I. 2016
IJerrett et al. 2013
ICakmaketal. 2016
Jerrett eta I. 2009
IJerrett et al. 2013
ICakmaketal. 2018 CanCHEC
IJerrett et al. 2013
Cardiovascular Mortality
IHD Mortality
Stroke Mortality
-1
0.9 1 1.1 1.2
Hazard Ratio (95% CI)
ACS = American Cancer Society; CanCHEC = Canadian Census Health and Environment Cohort; IHD = ischemic heart disease.
Note: Studies published since the 2013 Ozone ISA. Associations are presented per 10-ppb increase in pollutant concentration.
Circles represent point estimates; horizontal lines represent 95% confidence intervals for ozone. Black text and circles represent
evidence included in the 2013 Ozone ISA; red text and circles represent recent evidence not considered in previous ISAs or
AQCDs. Closed circles represent effect of ozone in single pollutant models, open circles represent effect of ozone adjusted for
PM2.5.
Figure 4-8 Associations between long-term exposure to ozone and
cardiovascular mortality with and without adjustment for PM2.5
concentrations in recent cohort studies.
4.2.17 Effect Modification of the Ozone-Cardiovascular Relationship
4.2.17.1 Pre-existing Disease
Individuals with certain pre-existing diseases may be considered at greater risk of an air
pollution-related health effect because they are likely in a compromised biological state that can vary
depending on the disease and severity. The 2013 Ozone ISA concluded that there was adequate evidence
for increased ozone-related health effects among individuals with asthma (U.S. EPA. 2013a). The results
4-61
-------
of controlled human exposure studies, as well as epidemiologic and animal toxicological studies,
contributed to this evidence. No studies evaluated in the 2013 Ozone ISA evaluated the potential of
pre-existing disease to modify the relationship between long-term ozone exposure and cardiovascular
health effects. Several recent studies conducted in China have evaluated the potential for pre-exiting
disease (e.g., hypertension, obesity) to modify the associations between long-term ozone exposure and
stroke, hypertension, or measures of blood pressure.
• Yang et al. (2017) observed increases in SBP and DBP associated with long-term ozone
exposure; these associations were stronger among those with prehypertension compared to
"normotensive" adults, although the associations were attenuated to near-null when hypertensive
adults are compared with normotensive adults, regardless of medication use.
• Positive associations were observed between long-term ozone exposure and stroke among
overweight and obese adults but not for normal-weight adults (Qin et al.. 2015). Zhao et al.
(2013) reported positive associations between long-term ozone exposure and hypertension, SBP,
and DBP among adults, which increased in magnitude when restricted to overweight and obese
adults. This trend was especially strong among men and was less apparent in analyses restricted
to women. In evaluations of children, the associations between long-term ozone exposure and
hypertension, SBP, and DBP were stronger among overweight children compared to
normal-weight children (Dong et al.. 2015).
• Qin et al. (2015) provided evidence of an interaction between sex and obesity status on the effect
of long-term ozone of cardiovascular health effects. Among all adults, there were positive
associations between ozone exposure and cardiovascular health effects, and these associations
were positive and higher in magnitude for those with higher BMI (i.e., >25 kg/m2). When
stratifying by BMI and sex, positive associations were observed between long-term ozone
exposure and cardiovascular health effects in men, with stronger associations in men with higher
BMIs. In females, the association was positive among those with higher BMIs (i.e., >25 kg/m2)
and negative among those with lower BMIs. Positive associations were observed between
long-term ozone exposure and CVD effects among obese adults; these associations remained
positive but were attenuated and near null for normal-weight and overweight adults. Among all
adults, there were positive associations between ozone exposure and CVD effects, and these
associations were similar after stratifying by BMI among all adults and males. In females, the
association was positive among those with higher BMIs (i.e., >25 kg/m2) and negative among
those with lower BMIs.
4.2.17.2 Lifestage
The 1996 and the 2006 Ozone AQCDs identified children, especially those with asthma, and
older adults as at-risk populations (U.S. EPA. 2006. 1996a). In addition, the 2013 Ozone ISA concluded
that there was adequate evidence to conclude that children and older adults are at increased risk of
ozone-related health effects (U.S. EPA. 2013a). Collectively, the majority of evidence for older adults has
come from studies of short-term ozone exposure and mortality, with little evidence contributed by studies
of long-term ozone exposure. No recent studies contribute evidence to determine whether children are at a
greater risk of cardiovascular health effects due to long-term ozone exposure compared to adults. A
limited number of recent studies of long-term ozone exposure and cardiovascular effects have compared
4-62
-------
associations between different age groups, but do not report consistent evidence that older adults are at
increased risk.
• In an English cohort, Atkinson et al. (2013) observed no difference in the association between
long-term ozone exposure and heart failure for participants aged 40-64 years compared with
those aged 65-89 years.
• In a cross-sectional study of 33 Chinese communities (Dong et al.. 2013a'). the association
between long-term ozone exposure and prevalent prehypertension was stronger among older
women (>55 years) compared with younger women (<35 years), whereas the association for
increases in blood pressure were stronger among younger adults (<35 years) compared with older
adults (>55 years). In an additional cross-sectional analysis of a Chinese population, stronger
associations were observed between long-term ozone exposure and hypertensions in both younger
(<55 years) and older (>65 years) adults, compared with adults that were between 55 and 64 years
old.
4.2.18 Summary and Causality Determination
This section evaluates evidence for cardiovascular health effects, with respect to the causality
determination for long-term exposures to ozone using the framework described in the Preamble to the
ISA (U.S. EPA. 2015). The key evidence, as it relates to the causal framework, is summarized in
Table 4-2. A small number of toxicological studies reviewed in the 2013 Ozone ISA provided some
evidence for enhanced atherosclerosis and impaired cardiac contraction in isolated perfused rat hearts
following long-term ozone exposure ITJ.S. EPA (2013a). see pg. 7-40], In addition, an animal
toxicological study reported increases in markers associated with inflammation, oxidative stress,
thrombosis, and vasoconstriction following long-term exposure ITJ.S. EPA (2013a). see pg. 7-40], The
limited body of epidemiologic evidence included in the 2013 Ozone ISA included studies of long-term
ozone exposure and circulating biomarkers, as well as a study evaluating cardiovascular mortality. Recent
epidemiologic evidence remains limited, although several recent studies provide some evidence for
changes in measures of blood pressure or increases in hypertension outcomes. Further, the number of
studies of cardiovascular mortality has increased, and these studies generally report positive associations.
Overall, the limited number of recent studies are consistent with, and in some cases extend, the
conclusions in the 2013 Ozone ISA. This evidence is discussed in greater detail below.
Overall, the evidence base describing the relationship between long-term ozone exposure and
cardiovascular effects remains limited. A couple of recent animal toxicological studies continue to
demonstrate impaired cardiac function following long-term ozone exposure. Note that these studies were
conducted by the same laboratory and show similar effects to those studies included in the 2013 Ozone
ISA (Section 4.2.5.2). In addition, a limited number of recent animal toxicological studies show
inconsistent evidence with respect to increases in markers of inflammation, oxidative stress, and a
proatherosclerotic environment.
4-63
-------
There continues to be a limited number of epidemiologic studies evaluating the association
between long-term ozone exposure and cardiovascular effects. In the 2013 Ozone ISA, some studies
considered the relationship between long-term ozone exposure and circulating biomarkers in the blood,
observing generally null associations. Few recent studies evaluated circulating biomarkers, but instead
focused on changes in blood pressure or hypertension, with relatively few studies evaluating outcomes
such as IHD or MI, HF, or stroke. In addition, a number of recent epidemiologic studies of cardiovascular
mortality provide evidence of positive associations with long-term ozone exposure. Compared to the 2013
Ozone ISA, a greater number of recent epidemiologic studies of cardiovascular morbidity and mortality
evaluate the potential for copollutant confounding, especially with PMi0 and NO2 (Section 4.2.16). One
study (Coogan et al.. 2017) evaluated PM2 5 in copollutant models. Generally, these studies report that the
ozone association is relatively unchanged or slightly attenuated in copollutant models (Section 4.2.16;
Figure 4-8). Potential copollutant confounding continues to be a source of uncertainty when
characterizing the relationship between long-term ozone exposure and cardiovascular health effects.
Consistent with previous evidence, recent studies continue to demonstrate associations between
long-term ozone exposure and cardiovascular health effects among older adults, although the limited
number of studies that evaluated effect modification by age do not provide evidence that older adults are
at increased risk of cardiovascular health effects related to long-term ozone exposure compared with other
adults. Similarly, there is some emerging evidence that long-term ozone exposure may be associated with
changes in blood pressure among children, but there are no studies that evaluate whether children are at
increased risk of ozone-related cardiovascular health effects compared with adults. With regard to
pre-existing disease, there is limited recent evidence that BMI or obesity status may modify the risk of
long-term ozone exposure on changes in blood pressure, but this evidence base is small and not entirely
consistent.
Overall, recent animal toxicological and epidemiologic studies add to the body of evidence that
formed the basis of the conclusions in the 2013 Ozone ISA for cardiovascular health effects. This body of
evidence is limited, however, with some experimental and observational evidence for subclinical
cardiovascular health effects and little evidence for associations with outcomes such as IHD or MI, HF, or
stroke. The strongest evidence for the association between long-term ozone exposure and cardiovascular
health outcomes continues to come from animal toxicological studies of impaired cardiac contractility and
epidemiologic studies of blood pressure changes and hypertension and cardiovascular mortality. Recent
epidemiologic studies observed positive associations with changes in blood pressure or hypertension, but
animal toxicological studies report no effect of ozone on blood pressure changes. In conclusion, the
results observed across both recent and older experimental and observational studies conducted in various
locations provide some evidence from a limited number of studies for an association between long-term
ozone exposure and cardiovascular health effects. Collectively, the body of evidence for long-term
ozone exposure and cardiovascular effects is suggestive of, but not sufficient to infer, a causal
relationship.
4-64
-------
Table 4-2 Summary of evidence that is suggestive of, but not sufficient to infer,
a causal relationship between long-term ozone exposure and
cardiovascular effects.
Rationale for Causality
Determination3
Key Evidence13
Key References'3
Ozone
Concentrations
Associated with
Effects0
Limited or inconsistent
evidence from animal
toxicological studies at
relevant ozone
concentrations
Impaired cardiac contractility, increased
markers associated with systemic
inflammation/oxidative stress, and a
proatherosclerotic environment
Sethi et al. (2012) O.i
Gordon et al. (2013)
Miller et al. (2016)
Section 4.2.11.2
-1.0 ppm
Consistent evidence from
epidemiologic studies of
cardiovascular mortality
at relevant ozone
concentrations
Nationwide analyses of the ACS cohort,
demonstrating positive associations with
cardiovascular mortality; CanCHEC cohort in
Canada provides consistent evidence for a
positive association with IHD mortality
Turner et al.
Jerrett et al.
(2016)
(2013)
14.3-57.5 ppb
Jerrett et al. (2009)
Cakmak et al.
(2018)
Weichenthal et al.
(2017)
Cakmak et al.
(2016)
Crouse et al. (2015)
Section 4.2.15
Generally null evidence
from epidemiologic
cohort studies of IHD,
HF, and stroke
A limited number of studies evaluated these
cardiovascular morbidity endpoints and
generally report null or inverse associations
with ozone exposure
Atkinson et al.
(2013)
Section 4.2.3.1
Section 4.2.5.1
Section 4.2.12.1
19.9-24.7 ppb
Evidence of no effect
from a limited number of
animal toxicological
studies
Changes in blood pressure
Gordon et al. (2013)
(Gordon et al..
2014)
0.8 ppm
aBased on aspects considered in judgments of causality and weight of evidence in causal framework in Table I and Table II of the
Preamble to the ISAs U.S. EPA (2015).
bDescribes the key evidence and references, supporting or contradicting, contributing most heavily to causality determination and,
where applicable, to uncertainties or inconsistencies. References to earlier sections indicate where full body of evidence is
described.
°Describes the ozone concentrations with which the evidence is substantiated.
4-65
-------
4.3
Evidence Inventories—Data Tables to Summarize Study Details
4.3.1
Short-Term Ozone Exposure
Table 4-3 Epidemiologic studies of short-term exposure to ozone and heart failure.
Study
Study Population
Exposure Assessment
Mean(ppb)
Copollutant
Examination
Effect Estimates 95% Cla
IWinauist et al. (2012)
St. Louis MSA, U.S.
Ozone: January 1, 2001-June
27, 2007
Follow-up: January 1,
2001-June 27, 2007
Time-series study
n = 22.4
Counts of daily ED
visits and HA for CHF
among people
residing in the St.
Louis MSA
Concentrations from U.S. EPA Mean: 36.3
AQS at Tudor Street stationary Maximum'
monitor; data missing 1.9% of m 8
days
8-h max
Correlation (r):PM2.s: ED visits, lag 0-4: 1.05 (1.01,
0.25
Copollutant models:
NR
1.09)
HA, lag 0-4: 1.05 (1.02,
1.09)
IMiloievic et al. (2014)
England and Wales, U.K.
Ozone: 2003-2009
Follow-up: 2003-2009
Study
Hospital Episode
Statistics (HES) study
n = 312,332
Emergency hospital
admissions for heart
failure to NHS
hospitals, 2003-2008,
in HES database
using centroid of
census ward; median
age (IQR) 73 yr
(60-82), 54% male
HES
Data from nearest monitoring
station to residence on event
day. Control exposure days
defined using time-stratified
design using other days of the
month when case occurred
8-h max
Mean: NR
Median: 30.96
75th: 38.58
Correlation (r): PM2.5: Heart failure, lag 0-4: 0.99
-0.096; (0.98, 1.01)
NO2: -0.3489;
SO2: -0.0849; Other:
PM10 0.0302,
CO -0.2973
Copollutant models:
NA
4-66
-------
Table 4-3 (Continued): Epidemiologic studies of short-term exposure to ozone and heart failure.
Study
Study Population
Exposure Assessment
Mean(ppb)
Copollutant
Examination
Effect Estimates 95% Cla
ISarnat et al. (2015)
St. Louis, MO, U.S.
Ozone: June 1, 2001 -May 30,
2003
Follow-up: June 1, 2001-May
30, 2003
Time-series study
n = 69,679
ED visit records of
patients with CHF
residing in St. Louis
MSA (eight counties
each in Missouri and
Illinois) from 36 out of
43 acute care
hospitals
Averaged hourly concentrations Mean: 36.2
in St. Louis from U.S. EPA AQS
8-h max
Correlation (r):
PM2.5: 0.23;
NO2: 0.37;
SO2: -0.04;CO -0.01
Copollutant models:
NR
0-2 day distributed lag: 1.04
(0.99, 1.10)
Copollutant model with NO2,
2 day distributed lag: 1.02
(0.96, 1.08)
Copollutant model with PM2.5,
2 day distributed lag: 1.02
(0.97, 1.08)
Copollutant model with CO,
2 day distributed lag: 1.06
(1.00, 1.12)
IRodopoulou et al. (2015)
Little Rock, AR, U.S.
Ozone: 2002-2012
Follow-up: 2002-2012
Time-series study
n = 84,269
Daily emergency room
visits among persons 8-h max
15 yr and older, 19%
65 yr and older, 42.5%
male
U.S. AQS data from stationary
monitor in Little Rock
Mean: 40
Median: 39
75th: 50
Correlation (r): NR
Copollutant models:
NR
Hypertensive heart disease
and heart failure, lag 1: 0.97
(0.91, 1.05)
aAII epidemiologic results standardized to a 15 ppb increase in 24 hour avg, 20 ppb increase in 8 hour daily max, 25 ppb increase in 1 hour daily max ozone concentrations, or a
10-ppb increase in seasonal/annual ozone concentrations to facilitate comparability across studies.
4-67
-------
Table 4-4 Study-specific details from controlled human exposure studies of impaired heart function.
Study
Population
n, Sex, Age (Range or Mean ± SD)
Exposure Details Endpoints
(Concentration, Duration) Examined
Frampton et al. (2015)
Healthy adults (GSTM+/-)
n = GSTM+ 8, GSTM- 7 males, GSTM+ 4, GSTM-
5 females
Age: GSTM+: 25.4 ± 2.8 yr, GSTM-: 27.3 ± 4.2 yr
0.1, 0.2 ppm, 3 h
(alternating 15 min periods
of rest and exercise)
LV ejection time
1.5 h the day
before and 2.5 h
post-exposure
GSTM = glutathione S-transferase M1, LV = left ventricular; LVDP = left ventricular developed pressure.
4-68
-------
Table 4-5 Study-specific details from short-term animal toxicological studies of impaired heart function.
Study
Species (Stock/Strain),
n, Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Tankerslev et al. (2013)
Mice (C57BL/6J)
n = 10-16/group males, by 3 h of FA
0 females
Age: NR
Nppa null mice
n = 10-16/group males,
0 females
Age: NR
Approximately 0.5 ppm, 3 h of ozone followed Measures of cardiac function (8-10 h PE)
Mclntosh-Kastrinskv et al. Mice (C57BL/6) 0.245 ppm, 4 h (aged, FA, or ozone) on
(2013) n = o males, 14-15/group 3 separate days outdoors
females
Age: NR
LVDP, dP/dt, coronary flow in isolated perfused hearts
(8-11 h PE hearts were isolated and post-induced
ischemia)
Kurhanewicz et al. (2014) Mice (C57BL/6)
n = 5-8/group males,
0 females
Age:10-12 weeks
0.3 ppm, 4 h
LVDP, contractility (24 h PE)
4-69
-------
Table 4-5 (Continued): Study-specific details from short-term animal toxicological studies of impaired heart
function.
Study
Species (Stock/Strain),
n, Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Ramotetal. (2015)
Rats (FHH)
n = NR males,
NR females
Age:10-12 weeks
Rats (S-D)
n = NR males,
NR females
Age:10-12 weeks
Rats (SH)
n = NR males,
NR females
Age:10-12 weeks
Rats (SHHF)
n = NR males,
NR females
Age:10-12 weeks
Rats (SHSP)
n = NR males,
NR females
Age:10-12 weeks
Rats (WKY)
n = NR males,
NR females
Age:10-12 weeks
Rats (Wistar)
n = NR males,
NR females
Age: 10-12 weeks
0.25 ppm, 4 h
0.5 ppm, 4 h
1 ppm, 4 h
Cardiac pathology (immediately after and 24 h PE)
Wang et al. (2013) Rats (Wistar) 0.8 ppm, 4 h of ozone followed by Cardiac microscopy after 6th exposure sacrifice
n = 6/group males intra-tracheal instillation of saline or PM2.5
0 females ' twice/week for 3 weeks
Age: NR
4-70
-------
Table 4-5 (Continued): Study-specific details from short-term animal toxicological studies of impaired heart
function.
Study
Species (Stock/Strain),
n, Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Zvchowski et al. (2016)
Mice (C57BL/6J)
n = 4-8/group males,
0 females
Age: 6-8 weeks
1 ppm, 4 h of ozone (acute hypoxia [10.0%
O2] or normoxia [20.9% O2] 24 h/day for
3 weeks prior to exposure)
RV hypertrophy (18-20 h PE)
FA = filtered air; FHH = fawn-hooded hypertensive; LVDP = left ventricular developed pressure; PE = post-exposure; SH = spontaneously hypertensive; SHHF = spontaneously
hypertensive heart failure; S-D = Sprague-Dawley, WKY = Wistar Kyoto.
4-71
-------
Table 4-6 Epidemiologic studies of short-term exposure to ozone and ischemic heart disease.
Copollutant
Study Study Population Exposure Assessment Mean (ppb) Examination Effect Estimates 95% Cla
IFinnbiornsdottir et al. (2013)
Reykjavik, Iceland
Ozone: January 1,
2005-December 31, 2009
Follow-up: January 1,
2005-December 31, 2009
Case-crossover study
Icelandic Medicines
Registry
n = 5,246
Adults 18 yr or older
living in Reykjavik
capital area to whom
glyceryl trinitrates
were dispensed at
least once, mean age
74 yr, 57.9% male
Averaged hourly concentrations
at busy intersection; calculated
24-h avg and running avg of
3-day means including the day
of dispensing and 2 days prior.
Control exposure days selected
using symmetric bidirectional
design, 7 days before and after
the index day of event
24-h avg
Mean: 20.66
Maximum:
72.25
Correlation (r): NO2:
-0.62; PM-ioO.13
Copollutant models:
NA
24-h avg, lag 0: 0.96 (0.90,
1.04)
24-h avg, lag 1: 1.05 (0.81,
1.13)
24-h avg, lag 2: 1.11 (0.99,
1.26)
24-h avg, lag 3: 1.06 (0.94,
1.20)
Multipollutant model with NO2
and PM10
24-h avg, lag 0: 1.11 (0.99,
1.25)
24-h avg, lag 1: 1.28 (1.14,
1.37)
3-day mean, lag 0: 1.29 (1.11,
1.50)
INuvolone et al. (2013)
Tuscany region, five urban
areas, Italy
Ozone: January
2002-December 2005
Follow-up: January
2002-December 2005
Time-series study
Cardiovascular Risk
and Air Pollution in
Tuscany (RISCAT)
study
n = 4,555
All hospitalized Ml
cases in the study
region and period;
49.1 <75 yr, mean age
72.5 yr, 60.2% male
Daily 8-h max moving average
concentrations for each of
29 sites were combined into
5 areas with homogenous
concentration levels
8-h max
Mean: 47.51
Correlation (r):
NO2: -0.08;
CO -0.15,
PM10 0.21
Copollutant models:
NR
0-1 distributed lag: 1.05 (0.96,
1.16)
4-72
-------
Table 4-6 (Continued): Epidemiologic studies of short-term exposure to ozone and ischemic heart disease.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates 95% Cla
IMiloievic et al. (2014)
England and Wales, U.K.
Ozone: 2003-2009
Follow-up: 2003-2009
Case-crossover study
MINAP register, HES
n = 410 341 Ml
All Ml events,
2003-2009, in MINAP
registry located in
enumeration district of
residence (100-m
resolution) and
emergency hospital
admissions to NHS
hospitals, 2003-2008,
in HES database
using centroid of
census ward; median
age (IQR) 71 yr
(60-81) MINAP, 73 yr
(60-82), 65% male
MINAP, 54% male
HES
Data from nearest monitoring
station to residence on event
day. Control exposure days
defined using time-stratified
design using other days of the
month when case occurred
8-h max
Mean: NR
Median: 30.96
75th: 38.58
Correlation (r):
PM25: -0.096;
NO2: -0.3489;
SO2: -0.0849;
PM10 0.0302, CO
-0.2973
Copollutant models:
NA
All Ml, MINAP, lag 0-4: 0.99
(0.98, 1.00)
STEMI, MINAP, lag 0-4: 0.98
(0.96, 1.00)
nonSTEMI, MINAP, lag 0-4:
1.00 (0.98, 1.01)
IHD, HES, lag 0-4: 0.99 (0.98,
1.00)
Ml, HES, lag 0-4: 0.99 (0.98,
1.01)
IBard et al. (2014)
Strasbourg metropolitan area,
France
Ozone: 2000-2007
Follow-up: 2000-2007
Case-crossover study
Bas-Rhin Coronary Modeled hourly concentrations Mean: 32.13 Correlation (r):
Heart Disease
Register, a WHO
MONICA center
n = 2,134
Fatal and nonfatal Ml
cases, aged 35-74 yr,
76.9% male
at census block level using
ADMS-Urban air dispersion
model. Control days selected
using a monthly time-stratified
design
8-h avg
Median: 30.16
75th: 43.15
Maximum:
114.15
NO2: -0.34;
PM-io-0.16,
CO -0.34,
benzene -0.51
Copollutant models:
NA
Lag 0: 0.95 (0.86, 1.05)
Lag 1: 0.88 (0.79, 0.98)
Lag 0-1: 0.90 (0.80, 1.01)
ISarnat et al. (2015)
St. Louis, MO, U.S.
Ozone: June 1, 2001 -May 30,
2003
Follow-up: June 1, 2001-May
30, 2003
Time-series study
n = 69,679
ED visit records of
patients residing in St.
Louis MSA (eight
counties each in
Missouri and Illinois)
from 36 out of 43
acute care hospitals
Averaged hourly concentrations Mean: 36.2
in St. Louis from U.S. EPA AQS
8-h max
Correlation (r):
PM2.5: 0.23;
NO2: 0.37;
SO2: -0.04;
CO -0.01
Copollutant models:
NR
Ischemic heart disease,
0-2 day distributed lag: 0.99
(0.95, 1.04)
4-73
-------
Table 4-6 (Continued): Epidemiologic studies of short-term exposure to ozone and ischemic heart disease.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates 95% Cla
IWana et al. (2015a)
Calgary, Canada
Ozone: April 1, 1999-March 31,
2010
Follow-up: April 1, 1999-March
31, 2010
Case-crossover study
n = 25,894
Cases who were
residents of Alberta,
20 yr or older, 67.5%
male, living within
15 km to closest
stationary pollution
monitor and 50 km to
closest meteorological
monitor
Hourly concentrations from
41 monitor locations used to
calculate 24-h avg, 6-h avg for
morning and afternoon, 12-h
avg, daily 1-h max and daily 1-h
min. Cases linked to pollution
data by postal code, missing
records were imputed using
linear interpolation. Control
exposure days selected using
time-stratified design matching
on weekday stratified on month
and year
Mean: NR
Correlation (r): NR
Copollutant models:
NR
Analytical results were not
reported for main effects for
ozone, only statistically
significant results reported
4-74
-------
Table 4-6 (Continued): Epidemiologic studies of short-term exposure to ozone and ischemic heart disease.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates 95% Cla
IWana et al. (2015b)
Calgary, Edmonton, Canada
Ozone: April 1, 1999-March 31,
2010
Follow-up: April 1, 1999-March
31, 2010
Case-crossover study
n = 12,066
AMI cases aged 20 or
older living in urban
Calgary and
Edmonton
Averaged hourly concentrations Mean: NR
from four monitor locations in
each city. Control exposure
days selected using
time-stratified design matching
by weekday of event stratified
on month and year
24-h avg
Correlation (r): NR
Copollutant models:
NR
Calgary-all Mis, lag 0: 1.00
(0.96, 1.05)
Calgary-all Mis, lag 1: 0.97
(0.93, 1.01)
Calgary-all Mis, lag 2: 0.98
(0.94, 1.02)
Calgary-STEMI, lag 0: 1.00
(0.93, 1.07)
Calgary-STEMI, lag 1: 0.94
(0.87, 1.01)
Calgary-STEMI, lag 2: 0.97
(0.90, 1.04)
Calgary-NSTEMI, lag 0: 1.02
(0.95, 1.09)
Calgary-NSTEMI, lag 1: 0.98
(0.92, 1.05)
Calgary-NSTEMI, lag 2: 1.00
(0.93, 1.06)
Edmonton-all Mis, lag 0: 1.00
(0.96, 1.05)
Edmonton-all Mis, lag 1: 1.00
(0.95, 1.04)
Edmonton-all Mis, lag 2: 1.01
(0.97, 1.06)
Edmonton-STEMI, lag 0: 0.98
(0.91, 1.06)
Edmonton-STEMI, lag 1: 0.98
(0.91, 1.06)
Edmonton-STEMI, lag 2: 0.99
(0.92, 1.07)
Edmonton-NSTEMI, lag 0:
1.01 (0.95, 1.08)
Edmonton-NSTEMI, lag 1:
1.01 (0.95, 1.08)
Edmonton-NSTEMI, lag 2:
1.02 (0.96, 1.09)
4-75
-------
Table 4-6 (Continued): Epidemiologic studies of short-term exposure to ozone and ischemic heart disease.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates 95% Cla
ICIaevs etal. (2015)
National, Belgium
Ozone: 2006-2009
Follow-up: 2006-2009
Time-series study
National percutaneous
coronary intervention
(PCI) database
n = 15,964
All cases receiving
PCI procedures within
24 h of symptom
onset, 2006-2009, at
32 PCI centers in
Belgium, mean age
63 yr, 75% male
Averaged hourly concentrations
measured across all
73 monitors in Belgium, daily
average and 5-day avg
Mean: 21.68
Maximum: 65.5
Correlation (r):
PM25: -0.35;
PM-io-0.24
Copollutant models:
NR
Lag 5: 1.03 (0.97, 1.06)
IButland etal. (2016)
National, U.K.
Ozone: 2003-2010
Follow-up: 2003-2010
Case-crossover study
MINAP
n = 626,239
Acute coronary cases
from the MINAP
registry covering
National Health
Service hospitals in
England and Wales
excluding missing
geocodes, missing
data on date of event,
discharge diagnosis or
not residing in
England and Wales
and missing exposure
or covariate data,
median age 70.6 yr,
65% male
Daily concentrations (using
hourly data) with 5- * 5-km
resolution from EMEP4 U.K.
atmospheric chemistry
transport model (ACTM);
calculated daily max 8-h
running mean for ozone. Ml
events linked to concentrations
in closest 5-km grid. Control
concentrations selected using
time-stratified analysis using
event day stratified on month
8-h max
Mean: NR Correlation (r): NR All Ml, lag 0-2: 1.00 (0.99,
Copollutant models: 1-01)
NR STEMI, lag 0-2: 0.99 (0.98,
1.01)
nonSTEMI, lag 0-2: 1.00
(0.99, 1.01)
4-76
-------
Table 4-6 (Continued): Epidemiologic studies of short-term exposure to ozone and ischemic heart disease.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates 95% Cla
lAraacha et al. (2016)
National, Belgium
Ozone: 2009-2013
Follow-up: 2009-2013
Case-crossover study
Belgian STEMI
Registry
n = 11,420
STEMI cases included
in registry,
2009-2013, mean age
62.8 yr and 75.4%
male
National daily average
estimated using measurements
from 41 monitors, interpolation,
and adjustment for population
density. Control exposure days
selected using a time-stratified
design, stratifying by month and
year with a 4-day exclusion
period around the event day
24-h avg
Mean: 5.38
Median: 21.32
75th: 27.51
95th: 36.19
Correlation (r):
PM2.5: -0.388;
NO2: -0.6;
PM10-0.287
Copollutant models:
NR
Not statistically significant,
results in figure
tCollart et al. (2017)
Wallonia, Belgium
Ozone: January 1, 2008-
December31, 2011
Follow-up: January 1, 2008-
December31, 2011
Time-series study
n = 21,491
Daily counts of
hospital admissions at
42 hospitals in study
region, ages 25 yr and
older, mean age
66.9 yr, 66.9% male
Averaged daily concentrations
from 6-16 stationary monitors
24-h avg
Correlation (r): NR
Copollutant models:
NR
Analytic results displayed in
Figure 4. No associations
using any lag
tVidale et al. (2017)
n = 4,110
Average daily concentrations
Correlation (r): NR
Lag 0:
1.00 (0.99,
1.00)
Como, Italy
All residents of Como
from two stationary monitors
Copollutant models:
Lag 1:
0.98 (0.97,
1.01)
Ozone: January
2005-December 2014
Follow-up: January
2005-December 2014
with hospital
admission for acute Ml
between January
2005 and December
2014, mean age 71 yr,
24-h avg
NR
Time-series study
65% male
IRasche et al. (2018)
Jena, Germany
Ozone: January 1,
2003-December 31, 2010
Follow-up: January 1,
2003-December 31, 2010
Case-crossover study
n = 693
STEMI cases admitted
to university hospital
within 72 h of
symptom onset and
residing within 10 km
around the hospital,
median age 69 yr,
67.2% male
Daily average concentration
from monitor. Control exposure
days selected using
bidirectional design, previous
and following week
24-h avg
Median: 22.71 Correlation (r): NR Lag 2: 0.29 (0.11, 0.86)
Maximum: 23.5 Copollutant models:
NR
4-77
-------
Table 4-6 (Continued): Epidemiologic studies of short-term exposure to ozone and ischemic heart disease.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates 95% Cla
IHannaetal. (2011)
North Carolina, five cities, U.S.
Ozone: January 1,
1996-December 31, 2004
Follow-up: January 1,
1996-December 31, 2005
Time-series study
All hospital
admissions in North
Carolina
Daily concentrations from U.S.
EPA AQS in five cities in North
Carolina
1-h max
NR
Correlation (r): NR
Copollutant models:
NR
Data in figures by air mass
type and city; only one
statistically significant
association for extreme moist
tropical air mass at 5 day lag
IBhaskaran et al. (2011)
National, U.K.
Ozone: 2003-2006
Follow-up: 2003-2006
Case-crossover study
Ml NAP
n = 79,288
Ml cases, 64% male,
aged 59-80 yr, with
time of event recorded
in MINAP within
15 conurbations
during 2003-2006
Averaged hourly concentrations
for each conurbation from
stationary monitors, average
concentration for the hour of
the event. Referent exposures
selected using time-stratified
approach using day of week
within each month
Median: 19.29
75th: 28.43
Correlation (r):
N02: -0.58;
S02: -0.14; CO
-0.24
Copollutant models:
NR
1-h avg, lag 1-6 h: 0.99 (0.96,
1.02)
1-h avg, lag 7-12 h: 1.02
(0.99, 1.06)
1-h avg, lag 13-18 h: 0.97
(0.94, 1.00)
1-h avg, lag 19-24 h: 1.00
(0.97, 1.02)
1-h avg, lag 1-72 h: 0.97
(0.94, 1.00)
1-h avg, lag 25-72 h: 0.99
(0.96, 1.02)
aAII epidemiologic results standardized to a 15 ppb increase in 24 hour avg, 20 ppb increase in 8 hour daily max, 25 ppb increase in 1 hour daily max ozone concentrations, or a
10-ppb increase in seasonal/annual ozone concentrations to facilitate comparability across studies.
4-78
-------
Table 4-7 Epidemiologic panel studies of short-term exposure to ozone and ischemic heart disease.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates (95%
Cl)a
lEvans et al.
(2016)
Rochester,
NY, U.S.
Ozone:
2007-2012
Panel study
n = 362
Treated forSTEMI, NSTEMI, or
unstable angina
Mean concentrations from
NYDEC monitor
1-h max, 12, 24, 48, and
72-h avg
Mean: 27.4 Median: 27 75th:
36.9 Maximum: 104
Correlation (r): NR
Copollutant
models: NR
Increased odds of
STEMI
1 h prior to event: 1.35
(1.00, 1.85)
12 h prior to event: 1.26
(0.94, 1.69)
24 h prior to event: 1.16
(0.90, 1.50)
48 h prior to event: 1.11
(0.81, 1.51)
72 h prior to event: 1.21
(0.84, 1.74)
aAII epidemiologic results standardized to a 15 ppb increase in 24 hour avg, 20 ppb increase in 8 hour daily max, 25 ppb increase in 1 hour daily max ozone concentrations, or a 10
ppb increase in seasonal/annual ozone concentrations to facilitate comparability across studies.
4-79
-------
Table 4-8 Study-specific details from controlled human exposure studies of ST segment depression.
Study
Population, n, Sex, Age
(Range or Mean ± SD)
Exposure Details
(Concentration, Duration)
Endpoints Examined
Rich etal. (2018)
Older adults
0, 0.07, 0.120 ppm, 3 h (alternating 15-min
ST-segment depression 15 min, 4 and 24 h
n = 35 males, 52 females
periods of rest and exercise)
PE
Age: 55-70 yr
Ml = myocardial infarction; NSTEMI = non-ST-elevation myocardial infarction; PE = post-exposure; STEMI = ST-elevation myocardial infarction; NYDEC = New York State
Department of Environmental Conservation.
Table 4-9 Study-specific details from short-term animal toxicological studies of ST-segment depression.
Study
Species (Stock/Strain), n, Sex, Age
Exposure Details
(Concentration, Duration) Endpoints Examined
Farrai etal. (2012)
Rats (SH)
n = 6/group males, 0 females
Age: 12 weeks
0.2 ppm, 4 h ST-segment depression during exposure
0.8 ppm, 4 h
SH = spontaneously hypertensive; ST = beginning of the S wave to the end of the T wave.
4-80
-------
Table 4-10 Epidemiologic panel studies of short-term exposure to ozone and endothelial function.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates (95% Clf
tZanobetti et al. (2014)
Boston, MA, U.S.
Ozone: 2006-2009
Panel study
n = 64
T2D
Averaged hourly
concentrations from local sites
24-h avg
Mean: 10
Median: 28
75th: 33
Maximum: 47
Correlation (r):
NR
Copollutant
models: NR
No change in BAD at 5-day avg
exposure to ozone—qualitative
result
tLiunaman et al. (2014)
Boston, MA, U.S.
Ozone: 2003-2008
Panel study
Lags examined: 1-7 day
moving avg
Framingham
Offspring/Third
Generation
n =2,369
Hourly concentrations from
Boston area monitors were
averaged to create moving
averages
24-h avg
Mean: 23
Maximum: 64
Correlation (r):
NR
Copollutant
models: NR
Percentage increase PAT ratio:
1-day moving avg: -3.43
(-6.33, -0.53)
2-day moving avg: -4.42
(-7.90, -0.93)
5-day moving avg: -0.32
(-5.01, 4.37)
tLanzinqer et al. (2014)
Chapel Hill, NC, U.S.
Ozone: 2004-2005
Panel study
n = 22
Subjects with T2D aged
48-78 yr
Monitor data
8-h max
Mean: 41
Median: 39
75th: 52
Maximum: 82
Correlation (r):
NR
Copollutant
models: NR
Percentage increase FMD
Lag 0: -29.2 (-52.6, -5.80)
Lag 1: -27.0 (-54.0, -0.08)
tMirowskv et al. (2017)
CATHGEN
AQS monitor
Mean: 26
Correlation (r):
Percentage increase FMD
Chapel Hill, NC, U.S.
n = 13
24-h avg
Median: 25
NR
Lag 0: -17.14 (-40.82, 15.11)
Ozone: 2012-2014
Panel study
Have undergone cardiac
catheterization
75th: 33
Maximum: 63
Copollutant
models: PM2.5
Lag 1: 4.82 (-27.21, 49.82)
5-day avg: -19.93
(-53.46, 34.39)
Lags examined: 0-4, 5-day
Age 53-68
Percentage increase BAD
avg
Lag 0: -2.25 (-5.46, 1.07)
Additional endpoints
Lag 1: -2.04 (-5.25, 1.29)
reported: LAEI, SAEI
5-day avg: 1.82 (-3.11, 7.07)
BAD = brachial artery diameter; PAT = pulse amplitude tonometry; FMD = flow-mediated dilation; LAEI = large artery elasticity index; SAEI = small artery elasticity index;
T2D = type 2 diabetes.
aAII epidemiologic results standardized to a 15 ppb increase in 24 hour avg, 20 ppb increase in 8 hour daily max, 25 ppb increase in 1 hour daily max ozone concentrations, or a
10-ppb increase in seasonal/annual ozone concentrations to facilitate comparability across studies.
4-81
-------
Table 4-11 Study-specific details from controlled human exposure studies of vascular function.
Study
Population n, Sex, Age
(Range or Mean ± SD)
Exposure Details
(Concentration, Duration)
Endpoints Examined
Rich etal. (2018)
Older adults
n = 35 males, 52 females
Age: 55-70 yr
0, 0.07, 0.120 ppm, 3 h
(alternating 15-min periods of
rest and exercise)
BAD, FMD, VTI (day before exposure and at the end of each
of the three exposures)
Barath etal. (2013)
Healthy adults
n = 36 males, 0 females
Age: 26 ± 1 yr
0.3 ppm, 75 min (alternating
15-min periods of exercise and
rest)
Forearm blood flow in response to acetylcholine, sodium
nitroprusside, verapamil, or bradykinin. 2 and 6 h
post-exposure
Frampton et al. (2015)
Healthy adults (GSTM +/-)
n = GSTM+ 8, GSTM- 7 males,
GSTM+ 4, GSTM- 5 females
Age: GSTM+: 25.4 ± 2.8 yr,
GSTM-: 27.3 ± 4.2 yr
0.1, 0.2 ppm, 3 h (alternating
15-min periods of rest and
exercise)
Indicators of endothelial dysfunction including flow in response
to reactive hyperemia measured by arterial tonometry 1.5 h
the day before and 2.5 h post-exposure
BAD = brachial artery diameter; FMD = flow-mediated dilation; GSTM = glutathione S-transferase M1; VTI = velocity-time interval.
4-82
-------
Table 4-12 Study-specific details from short-term animal toxicological studies of vascular function.
Study
Species (Stock/Strain), n, Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Wana et al. (2013)
Rats (Wistar)
n = 6/group males, 0 females
Age: NR
0.8 ppm, 4 h of ozone followed by
intra-tracheal instillation of saline or
PM2.5 twice/week for 3 weeks
Markers of endothelial
dysfunction in blood (after 6th
exposure animals sacrificed
blood drawn)
Robertson et al. (2013)
Mice (C57BL/6)
n = NR males, NR females
Age:8-10 weeks
Mice (CD 36-/-)
n = NR males, NR females
Age:8-10 weeks
1 ppm, 4 h
Relaxation of aortic rings in
response to acetylcholine (aortic
rings isolated 24 h PE)
Paffett et al. (2015)
Rats (S-D)
n = 65 males, 0 females
Age:8-12 weeks
1 ppm, 4 h
Serum-induced vascular
dysfunction (serum collected
immediately before sacrifice)
Vascular function (24 h PE)
Kumarathasan et al. (2015)
Rats (F344)
n = 8/exposure group, 17/control group
males; 0 females
Age: NA
0.8 ppm, 4 h
Markers of endothelial
dysfunction in blood
(immediately and 24 h PE)
Markers of oxidative stress in
blood (immediately and 24 h PE)
Snow et al. (2018)
Rats (WKY)
n = 6-8/group males, 0 females
Age: -12 weeks
0.8 ppm, 4 h/day for 2 consecutive
days (diets enriched with coconut,
olive, or fish oil for 8 weeks prior)
Endothelial function (2 h PE)
Thomson et al. (2013)
Rats (F344)
n = 4-6/group males, 0 females
Age: NR
0.4 ppm, 4 h
0.8 ppm, 4 h
mRNA markers of vascular
function (tissue collected
immediately PE)
PE = post-exposure, S-D = Sprague-Dawley, WKY = Wistar Kyoto.
4-83
-------
Table 4-13 Epidemiologic studies of short-term exposure to ozone and emergency department visits or hospital
admissions for electrophysiological changes, arrhythmia, and cardiac arrest.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% Cla
lEnsor et al. (2013)
Houston, TX, U.S.
Ozone: 2004-2011
Follow-up:
2004-2011
Case-crossover study
n = 11,677
All adults in EMS
database, aged 18 yr and
over, mean 64 yr,
59% male, 46% black
TCEQ monitoring data, hourly
concentration from 47 monitors,
calculated daily max 8-h running
mean. Control days selected using
time-stratified design matching on day
(or hour) of event for the same month.
8-h max
Mean: NR Correlation (r):
PM2.5: 0.4;
NO2: -0.33;
SO2: 0.11; CO -0.32
Copollutant models:
NA
8-h max, lag 0: 1.04 (1.00,
1.07)
8-h max, lag 1: 1.02 (0.99,
1.05)
8-h max, lag 2: 1.03 (0.99,
1.06)
8-h max, lag 0-1: 1.04
(1.00, 1.07)
8-h max, lag 1-2: 1.03
(0.99, 1.07)
1-h max, lag 0: 1.05 (1.00,
1.10)
1-h max, lag 1: 1.05 (1.01,
1.10)
1-h max, lag 2: 1.06 (1.01,
1.11)
1-h max, lag 3: 1.05 (1.00,
1.10)
1-h max, 1-3 h distributed
lag: 1.06 (1.01, 1.11)
IRosenthal et al.
(2013)
Helsinki, Finland
Ozone: 1998-2006
Follow-up:
1998-2006
Case-crossover study
n = 2,134
Out-of-hospital cardiac
arrests due to cardiac,
mean age 67.7 yr, 66.2%
male
Hourly concentrations from four
stationary monitors. Control exposure
days selected using time-stratified
design matching on day of week
stratified on month and year
24-h avg
Mean: 23.76
Correlation (r): NR
Copollutant models:
PM coarse, PM2.5,
PM10, UFP, CO, NO,
NO2, SO2
Lag 0-7 h: 1.04 (0.95, 1.16)
Lag 0-24 h: 1.08 (0.96,
1.21)
Lag 24-48 h: 1.11 (0.99,
1.26)
Lag 48-72 h: 1.16 (1.03,
1.31)
Lag 0-3 days: 1.18 (1.00,
1.41)
4-84
-------
Table 4-13 (Continued): Epidemiologic studies of short-term exposure to ozone and emergency department visits
or hospital admissions for electrophysiological changes, arrhythmia, and cardiac arrest.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% Cla
IWinauist et al. (2012) n = 18.3
St. Louis MSA, U.S
Ozone: January 1,
2001-June 27, 2007
Follow-up: January 1,
2001-June 27, 2007
Time-series study
Counts of daily ED visits
and HA for dysrhythmia
among people residing in
the St. Louis MSA
Concentrations from U.S. EPA AQS Mean: 36.3
at Tudor Street stationary monitor, Maximum'
data missing 1.9% of days 1118
8-h max
Correlation (r):
PM2.5: 0.2;
Copollutant models:
NR
ED visits, lag 0-4: 1.00
(0.97, 1.04)
HA, lag 0-4: 1.00 (0.95,
1.04)
IRaza et al. (2014)
Stockholm County,
Sweden
Ozone: 2000-2010
Follow-up:
2000-2010
Case-crossover study
Swedish Cardiac Arrest
Register
n = 55,973
All cases that occurred in
Stockholm County between
2000 and 2010, excluding
those classified as
noncardiac, dead on arrival
of EMS or missing time
data, mean age 74 yr in
women and 70 yr in men,
67% male
Hourly concentrations from central
monitors in Stockholm and one
monitor in a rural location. Control
exposure days selected using
time-stratified design matching on
weekday stratified on month and year
24-h avg
Mean: 31.57
Maximum: 28.7
Correlation (r):
PM2.5: 0.22;
NO2: -0.32
Copollutant models:
NR
OR remained elevated (not
significant) in two-pollutant
model with NO2 (3-day
mean). Independent
association observed for
lag 0 and nonsignificant
associations for lag 1, lag 2
and lag 4 using 24-h
distributed lags up to 168 h
Lag 0: 1.16 (1.03, 1.29)
4-85
-------
Table 4-13 (Continued): Epidemiologic studies of short-term exposure to ozone and emergency department visits
or hospital admissions for electrophysiological changes, arrhythmia, and cardiac arrest.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% Cla
IStranev et al. (2014)
Perth, WA, Australia
Ozone: 2000-2010
Follow-up:
2000-2010
Case-crossover study
n = 8,551
Adult cases over 35 yr old
attended by a paramedic.
Time of event defined by
date and time of
emergency call. Referent
exposures selected using
time-stratified approach
using day of the week
within each month
Averaged hourly concentrations from
monitor with closest distance to case
based on postal code for the day and
hour of cardiac arrest
1-h max
Median: 20
75th: 27.3
95th: 35
Correlation (r):
PM2.5: -0.1346;
NO2: -0.5612;
SO2: 0.1301; PM10
0.0067, CO -0.405
Copollutant models:
NR
Lag 0-1 h
Lag 0-2 h
Lag 0-3 h
Lag 0-4 h
Lag 0-8 h
1.00 (0.99, 1.01)
1.00 (0.99, 1.01)
1.00 (0.99, 1.01)
1.00 (0.99, 1.01)
1.00 (0.99, 1.01)
Lag 0-12 h: 1.00 (0.99,
1.01)
Lag 0-24 h: 1.00 (0.99,
1.01)
Lag 0-48 h.: 1.00 (0.99,
1.01)
No associations observed in
multipollutant models, or
effect modification by sex or
age category 35-65, >65,
>75 yr
TMiloievic et al. (2014) HES
England and Wales,
U.K.
Ozone: 2003-2008
Follow-up:
2003-2009
n = 352,775
Emergency hospital
admissions for arrythmias,
atrial fibrillation and
conduction disorders to
NHS hospitals, in HES
database using centroid of
census ward; median age
(IQR) 73 yr (60-82 yr),
54% male HES
Data from nearest monitoring station Median: 30.96 Correlation (r):
to residence on event day. Control
exposure days defined using
time-stratified design using other days
of the month when case occurred
8-h max
75th: 38.58
PM2.5: -0.096;
NO2: -0.3489;
SO2: -0.0849;
PM10 0.0302,
CO -0.2973
Copollutant models:
NA
Arrythmias, lag 0-4: 0.99
(0.98, 1.00)
Atrial fibrillation, lag 0-4:
0.99 (0.97, 1.00)
AVCD, lag 0-4: 1.00 (0.96,
1.03)
ISarnat et al. (2015)
St. Louis, MO, U.S.
Ozone: June 1,
2001-May 30, 2003
Follow-up: June 1,
2001-May 30, 2003
Time-series study
n = 69,679
ED visit records of patients
with dysrhythmia residing
in St. Louis MSA (eight
counties each in Missouri
and Illinois) from 36 out of
43 acute care hospitals
Averaged hourly concentrations in St.
Louis from U.S. EPA AQS
8-h max
Mean: 36.2
Correlation (r):
PM2.5: 0.23;
NO2: 0.37;
SO2: -0.04;
CO -0.01
Copollutant models:
NR
0-2 day distributed lag:
1.00 (0.94, 1.06)
4-86
-------
Table 4-13 (Continued): Epidemiologic studies of short-term exposure to ozone and emergency department visits
or hospital admissions for electrophysiological changes, arrhythmia, and cardiac arrest.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% Cla
tRodoDoulou et al.
(2015)
Little Rock, AR, U.S.
Ozone: 2002-2012
Follow-up:
2002-2012
Time-series study
n = 84,269
Daily emergency room
visits among persons 15 yr
and older, 19% were 65 yr
and older, 42.5% male
U.S. AQS data from stationary
monitor in Little Rock
8-h max
Mean: 40
Median: 39
75th: 50
Correlation (r): NR
Copollutant models:
PM2.5
Conduction disorders and
cardiac dysrhythmias, lag 1:
1.05 (0.99, 1.12)
Copollutant model with
PM2.5, lag 1: 1.06 (0.99,
1.13)
ISadeetal. (2015) n = 1,458
Negev, Israel
Ozone: 2006-2010
Follow-up:
2006-2010
Case-crossover study
All medical center patients
with first episode of atrial
fibrillation, living within
20 km of the monitoring
site, mean age 69 yr,
45.5% male
Averaged concentrations over 24 h.
Control exposure days selected using
time-stratified design matching on day
of week stratifying on month and year
24-h avg
Mean:
60.6-85.2
Correlation (r): NR
Copollutant models:
NA
Lag 0: 0.97 (0.89, 1.05)
Similar results for analyses
stratified by season
IPradeau et al. (2015) n = 4,558
Gironde Department,
France
Ozone: 2007-2012
Follow-up:
2007-2012
Case-crossover study
OHCA events among
adults aged 18 yr or older
recorded in the EMS
database, mean age 70 yr,
64% male
Averaged hourly concentrations from Mean: 27.26 Correlation (r): NR Lag 1: 1.14 (1.03, 1.24)
eight stationary monitors located in
Gironde. Control exposure days were
selected using time-stratified design
matching by day of week stratifying on
month
24-h avg
Median: 27.51
Maximum: 57.0
Copollutant models:
NR
aAII epidemiologic results standardized to a 15 ppb increase in 24 hour avg, 20 ppb increase in 8 hour daily max, 25 ppb increase in 1 hour daily max ozone concentrations, or a
10-ppb increase in seasonal/annual ozone concentrations to facilitate comparability across studies.
4-87
-------
Table 4-14 Epidemiologic panel studies of short-term exposure to ozone and electrophysiology, arrhythmia,
and cardiac arrest.
Study
Study Population
Exposure Assessment
Copollutant
Mean (ppb) Examination
Effect Estimates (95% Cl)a
TBartell et al. (2013)
n = 55
Hourly monitor values
Mean: 27.1 Correlation (r): NR
Increased risk for SVT, 24-h:
Los Angeles, CA, U.S.
Elderly nonsmokers
24-h avg
Maximum: 60.7 Copollutant models:
1.13 (0.83, 1.53)
Ozone: 2005-2007
(Age 71+ years)
NR
3-day avg: 0.51 (0.27, 0.96)
Panel study
5-day avg: 0.80 (0.28, 2.31)
Lags reported: 4-, 8-, 24-h, or
Increased risk for VT
3-, and 5-day avg
24-h: 1.50 (1.10, 2.05)
3-day avg: 2.54 (1.25, 5.18)
5-day avg: 0.94(0.14, 6.11)
ILiunaman et al. (2014)
Boston, MA, U.S.
Ozone: 2003-2008
Panel study
Lags reported: 1-7-day moving
avg
Framingham
Offspring/Third
Generation
n = 2,369
Hourly concentrations from
Boston area monitors were
averaged to create moving
averages
24-h avg
Mean: 23
Maximum: 64
Correlation (r): NR
Copollutant models:
NR
Percentage increase pulse
amplitude
1-day avg: 4.45 (-1.18,
10.08)
2-day avg: 7.64 (0.87, 14.40)
5-day avg: 3.87 (-5.22,
12.96)
ICakmak et al. (2014)
Ottawa and Gatineau, Canada
Ozone: 2004-2009
Panel study
Additional endpoints: SVT
ectopic runs, VT ectopic runs
n = 8,595
Referred for cardiac
monitoring ages
12-99 yr
Gatineau residents were
assigned levels at single
monitor serving the area;
Ottawa residents had three
monitors averaged to create
exposure
3-h max concentration for
preceding 24-h period
Mean: 34.89
Correlation (r):NR
Copollutant models:
NR
Nonstandardized data due to
unique exposure assessment
Percentage increase atrial
fibrillation
1.58 (-0.95, 4.17)
Percentage increase heart
block
1.13 (1.01, 1.26)
SVT = supraventricular tachycardia; VT = ventricular tachycardia.
aAII epidemiologic results standardized to a 15 ppb increase in 24 hour avg, 20 ppb increase in 8 hour daily max, 25 ppb increase in 1 hour daily max ozone concentrations, or a
10-ppb increase in seasonal/annual ozone concentrations to facilitate comparability across studies.
4-88
-------
Table 4-15 Study-specific details from controlled human exposure studies of electrophysiology, arrhythmia,
cardiac arrest.
Study
Population
n, Sex, Age
(Range or Mean ± SD)
Exposure Details
(Concentration, Duration)
Endpoints Examined
Kushaetal. (2012)
Healthy adults
n = 8 males, 9 females
Age: 18-38 yr
0.12 ppm, 2 h at rest
ECG endpoints, e.g., T-wave
alternans (continuously during
exposure)
Rich etal. (2018)
Older adults
n = 35 males, 52 females
Age: 55-70 yr
0, 0.07, 0.120 ppm, 3 h (alternating
15-min periods of rest and exercise)
Arrhythmia (over24-h recording
period including during exposure
3 h after exposure ECG recordings
made)
ECG endpoints (over 24-h
recording period including during
exposure)
ECG = electrocardiography.
4-89
-------
Table 4-16 Study-specific details from short-term animal toxicological studies of electrophysiology, arrhythmia,
cardiac arrest.
Study
Species (Stock/Strain), n, Sex,
Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Farrai et al. (2012)
Rats (SH)
n = 6/group males, 0 females
Age: 12 weeks
0.2 ppm, 4 h
0.8 ppm, 4 h
Arrhythmia induced by aconitine
(PE)
QRS QT PR ST intervals and
R-amplitude T-wave amplitude
(before, during, and after
exposure)
Wana et al. (2013)
Rats (Wistar)
n = 6/group males, 0 females
Age: NR
0.8 ppm, 4 h of ozone followed by
intra-tracheal instillation of saline or
PM2.5 twice/week for 3 weeks
ECG measures (24 h after 3rd
and 6th exposure)
Kurhanewicz et al. (2014)
Mice (C57BL/6)
n = 5-8/group males, 0 females
Age:10-12 weeks
0.3 ppm, 4 h
ECG (before, during and after
exposure)
Farrai et al. (2016)
Rats (SH)
n = 6/group males, 0 females
Age: 12 weeks
0.3 ppm
Day 1: 3 h of FA in the morning, 3 h of
FA in the afternoon;
Day 2: 3 h 0.5 ppm NO2 or FA exposure
in the morning, 0.3 ppm ozone or FA in
the afternoon
Cardiac sensitivity to aconitine
challenge (24 h after Day 2
exposure animals sacrificed)
PR interval (during exposure)
QT interval (during exposure)
ECG = electrocardiography; FA = filtered air; PE = post-exposure; PR = time interval between the beginning of the P wave to the peak of the R wave; QRS = time interval between
the beginning of the Q wave and the peak of the S wave; QT = time interval between the beginning of the Q wave to end of the T wave; SH = spontaneously hypertensive.
4-90
-------
Table 4-17 Epidemiologic studies of short-term exposure to ozone and blood pressure.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates (95% Cl)a
IBrook and Kousha (2015)
Edmonton and Calgary, Alberta,
Canada
Ozone: January
2010-December 2011
Follow-up: January
2010-December 2011
Case-crossover study
NACRS
n = males 2,688,
females 3,844
All ED visits for
hypertension in the
NACRS with
residence within
35 km from an air
monitor; included all
ages (97% >30 yr),
41% male. Controls
days were selected
using time-stratified
design matching on
day of week for case
and stratifying on
month and year
Averaged hourly concentrations
from monitors within 35 km of
residential postal code centroid
24-h avg
Mean: NR
Median: 22
Maximum: 50.1
Correlation (r): NR
Copollutant models:
NA
Females, cold season, lag 0;
pooled results for two cities:
0.98 (0.84, 1.12)
Females, cold season, lag 1;
pooled results for two cities:
0.98 (0.84, 1.12)
Females, cold season, lag 2;
pooled results for two cities:
0.96 (0.82, 1.10)
Females, cold season, lag 3;
pooled results for two cities:
0.98 (0.84, 1.12)
Females, warm season, lag
3, pooled results for two
cities: 1.15 (1.00, 1.31)
tRodopoulou et al. (2015)
n = 84,269
U.S. AQS data from stationary
Mean: 40
Correlation (r): NR Lag 1: 0.98 (0.96, 1.00)
Little Rock, AR, U.S.
Daily emergency room
monitor in Little Rock
Median: 39
Copollutant models:
Ozone: 2002-2012
visits for hypertension
8-h max
75th: 50
NR
Follow-up: 2002-2012
among persons 15 yr
and older, 19% 65 yr
Time-series study
and older, 42.5% male
4-91
-------
Table 4-17 (Continued): Epidemiologic studies of short-term exposure to ozone and blood pressure.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates (95% Cl)a
IVencloviene et al. (2017)
Kaunas, Lithuania
Ozone: January 1, 2009-June
30, 2011
Follow-Up: January 1,
2009-June 30, 2011
Time-series study
n = 17,114 calls
Individuals residing in
Kaunas and recorded
in the emergency calls
database, ages
17-104, 60.2% >65 yr,
21.6% male
Averaged hourly concentrations
from one stationary monitor
8-h max
Mean: 21.17
Median: 20.91
75th: 27.36
Maximum:
101.76
Correlation (r):
PM10 -0.028, CO
-0.298
Copollutant models:
NR
All year, lag 0: 0.94 (0.88,
0.98)
All year, lag 2-4: 1.06 (0.98,
1.14)
Autumn-winter, lag 0: 0.94
(0.87, 1.03)
Autumn-winter, lag 2-4: 0.96
(0.84, 1.08)
Spring-summer, lag 0: 0.93
(0.85, 1.00)
All year, lag 0, low ozone:
1.08 (1.00, 1.23)
All year, lag 0, high ozone:
0.93 (0.85, 1.03)
All year, lag 2-4, high ozone:
1.08 (0.97, 1.22)
Autumn-winter, lag 0, low
ozone: 1.08 (0.97, 1.23)
Autumn-winter, lag 0, high
ozone: 0.97 (0.81, 1.16)
Autumn-winter, lag 2-4, high
ozone: 0.93 (0.70, 1.25)
Spring-summer, lag 2-4:
1.11 (1.00, 1.23)
Spring-summer, lag 0, low
ozone: 1.17 (1.00, 1.43)
Spring-summer, lag 0, high
ozone: 0.93 (0.83, 1.03)
Spring-summer, lag 2-4,
high ozone: 1.16 (1.03, 1.34)
aAII epidemiologic results standardized to a 15 ppb increase in 24 hour avg, 20 ppb increase in 8 hour daily max, 25 ppb increase in 1 hour daily max ozone concentrations, or a
10-ppb increase in seasonal/annual ozone concentrations to facilitate comparability across studies.
4-92
-------
Table 4-18 Epidemiologic panel studies of short-term exposure to ozone and blood pressure.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates 95% CI
ICakmak et al. (2011)
Canada
Ozone: 2007-2009
Panel study
Canadian Health Measures Monitor
Survey 14l max
n = 5,604
Mean: 34.1
95th: 59.6
Correlation (r): NR
Copollutant
models: none
Absolute change DBP
(mm Hg), lag 0: 0.65 (0.06,
1.23)
Absolute change SBP
(mm Hg), lag 0: 1.17 (0.29,
2.05)
IHoffmann et al. (2012)
Boston, MA, U.S.
Ozone: 2006-2010
Panel study
n = 70
T2D; 40-85 yr
Monitor
24-h avg
Mean: 25
Median: 24
75th: 32
Correlation (r):
PM2.5: 0.09;
Copollutant
models: PM2.5
(presented
graphically; no
quantitative results
available)
Percentage Increase CMP
2-day mean: -0.33 (-2.30,
1.64)
5-day mean: -3.16 (-5.86,
-0.34)
Percentage Increase SBP
2-day mean: -0.66 (-2.74,
1.53)
5-day mean: -4.51 (-7.44,
-1.58)
Percentage Increase DBP
2-day mean: 0.11 (-1.64,
1.86)
5-day mean: -2.26 (-4.74,
0.02)
4-93
-------
Table 4-18 (Continued): Epidemiologic panel studies of short-term exposure to ozone and blood pressure.
Copollutant
Study
Study Population
Exposure Assessment
Mean (ppb)
Examination
Effect Estimates 95% CI
tDales and Cakmak (2016)
Canada Health Measures
Concentration on the day of
Mean: 29.5
Correlation (r):
Percentage increase SBP
National, Canada
Survey
testing from monitors located
Maximum: 83
PM2.5: NR;
Absence of mood disorder:
Ozone: 2007-2009
n = 1,883 (n = 1,693 absence
closest to clinic site
NO2: NR;
SO2: NR;
-0.52 (-1.18, 0.14)
Follow-up: 2007-2009
of mood disorder, n = 190
8-h max
Presence of mood disorder
presence of mood disorder)
Copollutant
4.41(1.91, 6.93)
Cross-sectional study
Population-based national
sample, aged 6-17 yr,
stratified by the presence or
absence of clinically
diagnosed mood disorder
models: NA
Percentage Increase DBP
Absence of mood disorder
-0.24(-0.85, 0.36)
Presence of mood disorder
3.55(1.01, 6.08)
IMirowskv et al. (2017)
Chapel Hill, NC, U.S.
Ozone: 2012-2014
Panel study
Lags reported 0-4, 5 day avg
CATHGEN
n = 13
Have undergone cardiac
catheterization
AQS Monitor
24-h avg
Mean: 26
Median: 25
75th: 33
Maximum: 63
Correlation (r): NR
Copollutant
models: PM2.5
Percentage increase SBP
Lag 0: 2.46 (-2.14, 7.18)
Lag 1: -0.11 (-3.54, 3.64)
5-day avg: 1.50 (-3.75,
6.96)
Percentage increase DBP
Lag 0: 2.46 (-2.25, 7.39)
Lag 1: -1.93 (-5.57, 1.82)
5-day avg: -0.43 (-5.89,
5.36)
ICole-Hunter et al. (2018)
Barcelona, Spain
Ozone: 2011-2014
Panel study
T APAS/EXPOsOM ICS
n = 57
Healthy, nonsmokers
Monitored values used to
model daily time weighted
based on location
(home/work)
24-h avg
Mean: 22
Maximum: 32.9
Correlation (r): NR
Copollutant
models: NR
BP = blood pressure, DBP = diastolic blood pressure, SBP = systolic blood pressure, T2D = type 2 diabetes.
Percentage increase SBP
home, exposure 3-days
prior: -0.52 (-1.75, 0.71)
Percentage increase home
DBP, exposure 3-days
prior: -0.20 (-1.04, 0.64)
4-94
-------
Table 4-19 Study-specific details from controlled human exposure studies of blood pressure.
Study
Population n, Sex, Age (Mean ± SD)
Exposure Details
(Concentration, Duration)
Endpoints Examined
Barath et al. (2013)
Healthy adults
n = 36 males, 0 females
Age: 26 ± 1 yr
0.3 ppm, 75 min (alternating 15 min
periods of exercise and rest)
SBP, DBP, ACE levels (2 and
6 h PE)
FramDton et al. (2015)
Healthy adults
n = GSTM+ 8 males, GSTM- 7 males; GSTM+
4 females, GSTM- 5 females
Age: GSTM+: 25.4 ± 2.8 yr, GSTM-: 27.3 ± 4.2 yr
0.1, 0.2 ppm; 3 h (alternating 15 min
periods of rest and exercise)
SBP, DBP (during exposure and
immediately and 2.5 h PE)
Rich et al. (2018)
Older adults
n = 35 males, 52 females
Age: 55-70 yr
0, 0.07, 0.120 ppm, 3 h (alternating
15 min periods of rest and exercise)
SBP, DBP 15 min, 4 and 22 h
PE
Stieqel et al. (2017)
Healthy adults
n = 11 males, 4 females
Age: 23 to 31 yr
0.3 ppm, 2 h (four 15 min periods of
exercise)
SBP (pre- and immediately
post-exposure)
Ariomandi et al. (2015)
Adults with asthma (n = 10) and adults without
asthma (n = 16)
n = 13 males, 13 females
Age: asthma: 33.5 ± 8.8 yr, healthy: 30.8 ± 6.9 yr
0.1, 0.2 ppm, 4 h (alternating 30 min
periods of exercise and rest)
SBP, DBP, (before, immediately
after and 20 h PE)
ACE activity (before, immediately
after and 20 h PE)
ACE = angiotensin-converting enzyme; DBP = diastolic blood pressure; GSTM = glutathione S-transferase M1; PE = post-exposure; SBP = systolic blood pressure.
4-95
-------
Table 4-20 Study-specific details from short-term animal toxicological studies of blood pressure.
Study
Species (Stock/Strain), n,
Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Wana et al. (2013)
Rats (Wistar)
n = 6/group males,
0 females
Age: NR
0.8 ppm, 4 h of ozone
followed by intra-tracheal
instillation of saline or PM2.5
twice/week for 3 weeks
Blood pressure (24 h after 3rd and 6th exposure)
Waqner et al. (2014)
Rats (S-D), fed high
fructose or normal diet
n = 4/group males,
0 females
Age: 8 weeks
0.5 ppm, 8 h/day for
9 consecutive weekdays
Blood pressure (during 9-day exposure)
Farrai et al. (2016)
Rats (SH)
0.3 ppm
Blood pressure (during exposure)
n = 6/group males,
0 females
Age: 12 weeks
Day 1: 3 h of FA in the
morning, 3 h of FA in the
afternoon
Day 2: 3 h 0.5 ppm NO2 or
FA exposure in the
morning, 0.3 ppm ozone or
FA in the afternoon
Pulse pressure (during exposure)
Tankerslev et al. (2013)
Mice (C57BL/6J)
n = 10-16/group males,
0 females
Age: NR
Nppa null mice
n = 10-16/group males,
0 females
Age: NR
Approximately 0.5 ppm, 3 h
of ozone followed by 3 h of
FA
Right ventricular systolic pressure and total peripheral resistance
(8-10 h PE)
4-96
-------
Table 4-20 (Continued): Study-specific details from short-term animal toxicological studies of blood pressure.
Study
Species (Stock/Strain), n,
Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Ramotetal. (2015)
Rats (FHH)
n = NR males, NR females
Age:10-12 weeks
Rats (S-D)
n = NR males, NR females
Age:10-12 weeks
Rats (SH)
n = NR males, NR females
Age:10-12 weeks
Rats (SHHF)
n = NR males, NR females
Age:10-12 weeks
Rats (SHSP)
n = NR males, NR females
Age:10-12 weeks
Rats (WKY)
n = NR males, NR females
Age:10-12 weeks
Rats (Wistar)
n = NR males, NR females
Age: 10-12 weeks
0.25 ppm, 4 h
0.5 ppm, 4 h
1 ppm, 4 h
ACE levels in blood (immediately after and 24 h PE)
ACE = angiotensin-converting enzyme; FHH = fawn-hooded hypertensive; PE = post-exposure; S-D :
hypertensive heart failure; WKY = Wistar Kyoto.
Sprague-Dawley; SH = spontaneously hypertensive; SHHF = spontaneously
4-97
-------
Table 4-21 Epidemiologic panel studies of short-term exposure to ozone and heart rate variability (HRV), heart
rate (HR).
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates 95% Cla
ICakmak et al. (2011)
Canada
Ozone: 2007-2009
Panel study
Canadian Health Measures Monitor
Survey 14l max
n = 5,604
Mean: 34.1
95th: 59.6
Correlation (r): NR
Copollutant
models: None
Absolute change resting
heart rate (bpm), lag 0: 0.90
(0.18, 1.63)
TBartell et al. (2013)
Los Angeles, CA, U.S.
Ozone: 2005-2007
Panel study
Lags reported: 4, 8, 24 h, or
3-, and 5-day avg
Additional endpoints reported:
pNN50
n = 55
Elderly nonsmokers
Hourly monitor values
24-h avg
Mean: 27.1
Maximum:
60.7
Copollutant
models: NR
Correlation (r): NR Percentage increase
rMSSD, 24-h: 0.54 (-3.04,
4.13)
3-day avg: -1.68 (-7.71,
4.34)
5-day avg: -9.03 (-19.23,
1.16)
Percentage increase
SDNN, 24-h: 2.09 (-0.28,
4.45)
3-day avg: -0.04 (-3.91,
3.84)
5-day avg: -9.21 (-15.79,
-2.63)
ICakmak et al. (2014) n = 8,595
Ottawa and Gatineau, Canada Referred for cardiac
Ozone: 2004-2009 monitoring ages 12-99 yr
Panel study
Gatineau residents were
assigned levels at single
monitor serving the area,
Ottawa residents had three
monitors averaged to create
exposure
3-h max concentration for
preceding 24-h period
Mean: 34.89
Correlation (r): NR
Copollutant
models: NR
Nonstandardized data due
to unique exposure metric
Percentage increase
maximum HR
0.54 (-0.09, 1.16)
Percentage increase
average HR
0.11 (-0.46, 0.67)
4-98
-------
Table 4-21 (Continued): Epidemiologic panel studies of short-term exposure to ozone and heart rate variability
(HRV), heart rate (HR).
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates 95% Cla
IDales and Cakmak (2016)
Canada
Ozone: 2007-2009
Cross-sectional study
Canada Health Measures
Survey
n = 1,883 (n = 1,693 absence
of mood disorder, n = 190
presence of mood disorder)
Population-based national
sample, aged 6-17 yr,
stratified by the presence or
absence of clinically
diagnosed mood disorder
Concentration on the day of
testing from monitors located
closest to clinic site
8-h max
Mean: 29.5
Maximum: 83
Correlation (r):
PM25: NR;
NO2: NR;
SO2: NR;
Copollutant
models: NA
Percentage increase HR
Absence of mood disorder:
-0.42 (-1.36, 0.52)
Presence of mood disorder
2.47 (-1.52, 6.47)
IMirowskv et al. (2017)
Chapel Hill, NC, U.S.
Ozone: 2012-2014
Panel study
Lags reported: 0-4, 5-day avg
CATHGEN
n = 13
Have undergone cardiac
catheterization
AQS monitor
24-h avg
Mean: 26
Median: 25
75th: 33
Maximum: 63
Correlation (r): NR
Copollutant
models: PM2.5
Percentage increase
SDNN, lag 0: 0.21 (-11.79,
13.71)
Lag 1: -2.89 (-12.96, 8.25)
Lag 2: 1.07 (-9.21, 12.32)
5-day avg: -6.64 (-20.25,
9.11)
Percentage increase
rMSSD, lag 0: 6.11 (-13.18,
29.25)
Lag 1: 2.14 (-13.61, 20.46)
Lag 2: 4.29 (-11.14, 22.29)
5-day avg: -5.25 (-25.29,
19.71)
ICole-Hunter et al. (2018)
Barcelona, Spain
Ozone: 2011-2014
Panel study
T APAS/EXPOsOM ICS
n = 62
Healthy nonsmokers
Monitored values used to
model daily time weighted
based on location
(home/work)
24-h avg
Mean: 22
Maximum:
32.9
Correlation (r):
Copollutant
models: NR
NR Percentage increase HR,
3-days prior: 0.41 (-0.66,
1.49)
CATHGEN = catheterization genetics; HR = heart rate; pNN50 = the proportion of NN50 divided by the total number of NN (R-R) intervals, rMSSD = root-mean-square of the
successive differences between adjacent NNs, SDNN = standard deviation of NN intervals.
aAII epidemiologic results standardized to a 15 ppb increase in 24 hour avg, 20 ppb increase in 8 hour daily max, 25 ppb increase in 1 hour daily max ozone concentrations, or a
10-ppb increase in seasonal/annual ozone concentrations to facilitate comparability across studies.
4-99
-------
Table 4-22 Study-specific details from controlled human exposure studies of heart rate variability (HRV), heart
rate (HR).
Study
Population n, Sex, Age (Range or Mean ± SD)
Exposure Details
(Concentration, Duration)
Endpoints Examined
Kushaetal. (2012)
Healthy adults
n = 8 males, 9 females
Age: 18-38 yr
0.12 ppm, 2 h at rest
Heart rate (first and last 5 min of
exposure)
Barath et al. (2013)
Healthy adults
n = 36 males, 0 females
Age: 26 ± 1 yr
0.3 ppm, 75 min (alternating 15-min
periods of exercise and rest)
HRV time and frequency
parameters (2 and 6 h PE)
Heart rate (2 and 6 h PE)
Frampton et al. (2015)
Healthy adults (GSTM+/-)
n = GSTM+ 8, GSTM- 7 males, GSTM+ 4,
GSTM- 5 females
Age: GSTM+: 25.4 ± 2.8 yr, GSTM-:
27.3 ± 4.2 yr
0.1, 0.2 ppm, 3 h (alternating 15-min
periods of rest and exercise)
Heart rate (during exposure and
immediately and 2.5 h PE)
Ariomandi et al. (2015)
Adults with asthma (n = 10) and adults without
asthma (n = 16)
n = 13 males, 13 females
Age: asthma: 33.5 ± 8.8 yr, healthy: 30.8 ± 6.9 yr
0.1, 0.2 ppm, 4 h (alternating 30-min
periods of exercise and rest)
HRV time and frequency
parameters (before, immediately
after and 20 h PE)
Rich et al. (2018)
Older adults
n = 35 males, 52 females
Age: 55-70 yr
0, 0.07, 0.120 ppm, 3 h (alternating
15-min periods of rest and exercise)
HR (over24-h recording period
including during exposure)
HRV time and frequency
parameters (over24-h recording
period including during
exposure)
GSTM = glutathione S-transferase M1, HR = heart rate; HRV = heart rate variability; PE = post-exposure.
4-100
-------
Table 4-23 Study-specific details from short-term animal toxicological studies of heart rate variability (HRV),
heart rate (HR).
Study
Species (Stock/Strain), n, Sex,
Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Farrai et al. (2012)
Rats (SH)
0.2 ppm, 4 h
HRV (before, during, and after
n = 6/group males, 0 females
Age: 12 weeks
0.8 ppm, 4 h
exposure)
HRV: time and frequency
domains (before, during, and
after exposure)
Heart rate (before, during, and
after exposure)
Wanq et al. (2013)
Rats (Wistar)
n = 6/group males, 0 females
Age: NR
0.8 ppm, 4 h of ozone followed by
intra-tracheal instillation of saline or
PM2.5 twice/week for 3 weeks
Heart rate (24 h after 3rd and
6th exposure)
Measures of HRV (24 h after 3rd
and 6th exposure)
Mclntosh-Kastrinskv et al. (2013)
Mice (C57BL/6)
n = 0 males, 14-15/group females
Age: NR
0.245 ppm, 4 h (aged, FA, or ozone)
3 separate days outdoors
on Heart rate in isolated perfused
hearts (8-11 h PE hearts were
isolated and post-induced
ischemia)
Waqner et al. (2014)
Rats (S-D), fed high-fructose or
normal diet
n = 4/group males, 0 females
Age: 8 weeks
0.5 ppm, 8 h/day for 9 consecutive
weekdays
HR (during 9-day exposure)
Time domains of HRV (during
9-day exposure)
Kurhanewicz et al. (2014)
Mice (C57BL/6)
n = 5-8/group males, 0 females
Age: 10-12 weeks
0.3 ppm, 4 h
HR (before, during, and after
exposure)
4-101
-------
Table 4-23 (Continued): Study-specific details from short-term animal toxicological studies of heart rate
variability (HRV), heart rate (HR).
Species (Stock/Strain), n, Sex, Exposure Details
Study Age (Concentration, Duration) Endpoints Examined
Farrai et al. (2016) Rats (SH) 0.3 ppm Heart rate (during exposure)
n = 6/group males, 0 females Day 1: 3 h of FA in the morning, 3 h of Time and frequency domains of
Age-12 weeks FA in the afternoon HRV (during exposure)
Day 2: 3 h 0.5 ppm NO2 or FA
exposure in the morning, 0.3 ppm
ozone or FA in the afternoon
FA = filtered air; HR = heart rate; HRV = heart rate variability; S-D = Sprague-Dawley; SH = spontaneously hypertensive.
4-102
-------
Table 4-24 Epidemiologic studies of short-term exposure to ozone and thrombosis.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% Cla
ISpiezia et al. (2014)
Padua, Italy
Ozone: January 2008 and
October 2012
Follow-up: January 2008 and
October 2012
Case-control study
n = 33 cases and
72 controls
All consecutive
hospital admissions to
thrombosis unit at
university hospital with
objective identification
of acute first episode
of isolated pulmonary
embolism between
January 2008 and
October 2012. Cases
defined as having no
predisposition and
controls defined as
having permanent or
transient risk factors;
mean age of cases
and controls, 67 yr
and 68 yr,
respectively. Patients
excluded if under
18 yr, being treated
with anticoagulants,
had previous episode
of pulmonary
embolism, or did not
reside in Padua
Averaged concentration using
two stationary monitors in the
city; data from the closest
monitor to the patient's address
was used. Mean concentration
over month preceding date of
diagnosis
Correlation (r): NR
Copollutant models:
NR
No associations with monthly
average ozone >37 ppb
4-103
-------
Table 4-24 (Continued): Epidemiologic studies of short-term exposure to ozone and pulmonary vascular disease
(PVD), thrombosis.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% Cla
IMiloievic et al. (2014)
England and Wales, U.K.
Ozone: 2003-2009
Follow-up: 2003-2009
Case-crossover study
HES
n = 82,231
Emergency hospital
admissions for
pulmonary embolism
to NHS hospitals,
2003-2008, in HES
database using
centroid of census
ward; median age
(IQR) 73 (60-82),
54% male HES
Data from the nearest
monitoring station to residence
on event day. Control exposure
days defined using
time-stratified design using
other days of the month when
case occurred
8-h max
Mean: NR
Median: 30.96
75th: 38.58
Correlation (r):
PM2.5: -0.096
NO2: -0.3489
SO2: -0.0849
Lag 0-4: 0.99 (0.96, 1.02)
PM10
0.0302, CO -0.2973
Copollutant models:
NA
tde Miquel-Diez et al. (2016)
National, Spain
Ozone: January 1,
2000-December 31, 2013
Follow-up: January 1,
2001-December 31, 2013
Case-crossover study
Spanish Minimum
Basic Data Set, covers
97.7% of all
admissions to public
hospitals
n = 105,117
Cases recorded in
SMBD database
during the study
period, mean age
70.73 yr, 45.8% male
Concentration from stationary
monitor nearest to postal code,
calculated 3-day avg including
day of embolism and Days 1
and 2 prior. Control exposures
were 3-day avg at 1 week,
1.5 weeks, 2 weeks, and
3 weeks before the event
Mean: NR
Correlation (r): NR
Copollutant models:
NR
Control period 3 weeks prior
to event: 1.03 (1.01, 1.06);
effect estimate not
standardized, increment not
reported
PVD = peripheral vascular disease.
aAII epidemiologic results standardized to a 15 ppb increase in 24 hour avg, 20 ppb increase in 8 hour daily max, 25 ppb increase in 1 hour daily max ozone concentrations, or a
10-ppb increase in seasonal/annual ozone concentrations to facilitate comparability across studies.
4-104
-------
Table 4-25 Epidemiologic panel studies of short-term exposure to ozone and coagulation.
Copollutant
Study Study Population Exposure Assessment Mean (ppb) Examination Effect Estimates (95% CI)a
tGreen et al. (2015)
SWAN
Monthly averages of AQS
Mean: 35.2
Correlation (r):NR
Percentage increase factor lie,
Multicity; Chicago, Detroit, Los
n = 2,086
data from one monitor
Maximum: 122
Copollutant models:
1-day: -0.80 (-2.00, 0.20)
Angeles, Newark, Oakland, and
Midlife women
within 20 km of residence
NR
Percentage increase fibrinogen,
Pittsburgh, U.S.
(42-52 yr)
8-h max
1-day: -0.80 (-3.20, 1.60)
Ozone: 1999-2004
Percentage increase PAI-1, 1-day:
Cohort study
-2.20 (-5.00, 0.80)
Percentage increase tPA, 1-day:
0.20 (-1.20, 1.80)
TBind et al. (2012)
Normative Aging
Monitors in the Boston
Mean: 24
Correlation (r):NR
No change in fibrinogen,
Boston, MA, U.S.
Study
area
95th: 49
Copollutant models:
qualitative results only
Ozone: 2000-2009
n = 704
24-h avg
NR
Panel study
Lags reported: 4 and 24 h, or 3-,
7-, 14-, 21- or 28-day avg
tMirowskv et al. (2017)
CATHGEN
AQS monitor
Mean: 26
Correlation (r):NR
Percentage increase tPA
Chapel Hill, NC, U.S.
n = 13
24-h avg
Median: 25
Copollutant models:
Lag 0: 5.79 (-3.32, 15.75)
Ozone: 2012-2014
Have undergone
75th: 33
NR
Lag 1: -0.96 (-7.71, 6.32)
Lag 2: 2.89 (-4.29, 10.71)
5-day avg: 9.43 (-2.14, 22.18)
Panel study
cardiac catheterization
Maximum: 63
Lags reported: 0-4, 5-day avg
Percentage increase PAI-1
Lag 0: 8.79 (-13.71, 36.75)
Lag 1: 11.36 (-7.82, 34.29)
Lag 2: 21.43 (0.86, 45.86)
5-day avg: 43.39 (9.32, 87.43)
CATHGEN = catheterization genetics; PAI-1 = plasminogen activator inhibitor 1; SWAN = study of women's health across nations; tPA = tissue plasminogen activator.
aAII epidemiologic results standardized to a 15 ppb increase in 24 hour avg, 20 ppb increase in 8 hour daily max, 25 ppb increase in 1 hour daily max ozone concentrations, or a
10-ppb increase in seasonal/annual ozone concentrations to facilitate comparability across studies.
4-105
-------
Table 4-26 Study-specific details from controlled human exposure studies of coagulation.
Study
Population n, Sex, Age (Range or Mean ± SD)
Exposure Details
(Concentration, Duration)
Endpoints Examined
Barath et al. (2013)
Healthy adults
n = 36 males, 0 females
Age: 26 ± 1 yr
0.3 ppm, 75 min (alternating 15-min
periods of exercise and rest)
Markers of coagulation in blood
(2 and 6 h PE)
Frampton et al. (2015)
Healthy adults (GSTM+/-)
n = GSTM+ 8, GSTM- 7 males, GSTM+ 4,
GSTM- 5 females
Age: GSTM+: 25.4 ± 2.8 yr, GSTM-:
27.3 ± 4.2 yr
0.1, 0.2 ppm; 3 h (alternating 15-min
periods of rest and exercise)
Platelet activation and
microparticle circulation 1.5 h the
day before and 2.5 h PE
Kahleetal. (2015)
Healthy adults
n = 14 males, 2 females
Age: 20-36
0.3 ppm, 2 h, 15 min of exercise
alternating with 15 min of rest one
exposure at 22°C other at 32.5°C
Markers of coagulation (24 h PE)
Ariomandi et al. (2015)
Adults with asthma (n = 10) and adults without
asthma (n = 16)
n = 13 males, 13 females
Age: Asthma: 33.5 ± 8.8 yr, healthy:
30.8 ± 6.9 yr
0.1, 0.2 ppm, 4 h (alternating 30-min
periods of exercise and rest)
Markers of coagulation in blood
(before, immediately after and
20 h PE)
FramDton et al. (2017)
Older adults
n = 35 males, 52 females
Age: 55-70 yr
0, 0.07, 0.120 ppm, 3 h (alternating
15-min periods of rest and exercise)
Markers of coagulation in blood
(day before, day of, and up to
22 h PE)
GSTM = glutathione S-transferase M1; PE = post-exposure.
4-106
-------
Table 4-27 Study-specific details from short-term animal toxicological studies of coagulation.
Study
Species (Stock/Strain), n,
Sex, Age
Exposure Details (Concentration, Duration)
Endpoints Examined
Snow et al. (2018)
Rats (WKY)
n = 6-8/group males,
0 females
Age: -12 weeks
0.8 ppm, 4 h/day for 2 consecutive days (diets
enriched with coconut, olive, or fish oil for 8 weeks
prior)
Circulating platelets
WKY = Wistar Kyoto.
4-107
-------
Table 4-28 Epidemiologic panel studies of short-term exposure to ozone and inflammation.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates 95% Cla
TBind et al. (2012)
Normative aging
Monitors in the Boston area
Mean: 24
Correlation (r): NR
Percentage increase CRP
Boston, MA, U.S.
study
24-h avg
95th: 49
Copollutant models:
24-h: 0.87 (0.81, 0.95)
Ozone: 2000-2009
n = 704
NR
Panel study
tGandhi et al. (2014)
Rutgers, NJ, U.S.
Ozone: 2006-2009
Panel study
Lags reported: 0-6
n = 49
Healthy, nonsmoking
young adults
Hourly concentrations from
East Brunswick from AQS
24-h avg
Mean: 25.3
Median: 24.8
75th: 33.2
Maximum:
67.7
Correlation (r):
PM25: -0.05, S04:
-0.05, NOx: -0.52
Copollutant models:
NR
Percentage increase plasma
nitrite
Lag 0: -5.61 (-20.61, 9.47)
Lag 1: -4.91 (-18.33, 8.42)
IGreen et al. (2015)
Multicity; Chicago, Detroit, Los
Angeles, Newark, Oakland,
Pittsburgh, U.S.
Ozone: 1999-2004
Cohort study
SWAN
n = 2,086
Midlife women
(42-52 yr)
Monthly averages of AQS data
from one monitor within 20 km
of residence
8-h max
Mean: 35.2
Maximum: 122
Correlation (r): NR
Copollutant models:
NR
Percentage increase CRP
1-day: 0.80 (-2.00, 3.60)
TLi et al. (2016)
Boston, MA, U.S.
Ozone: 1998-2008
Panel study
Framingham
Offspring
n = 2,035
Nonsmokers
Mean concentrations from
Harvard supersite
24-h avg
Mean: 20
Correlation (r): NR
Copollutant models:
NR
Qualitative results for
myeloperoxidase and
indexed 8-epi-prostaglandin
F2alpha show no change
4-108
-------
Table 4-28 (Continued): Epidemiologic panel studies of short-term exposure to ozone and inflammation.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates 95% Cla
IMirowskv et al. (2017)
Chapel Hill, NC, U.S.
Ozone: 2012-2014
Panel study
Lags reported: 0-4, 5-day avg
Additional endpoints reported:
VCAM, monocytes, neutrophils
CATHGEN
n = 13
Have undergone
cardiac
catheterization
AQS monitor
24-h avg
Mean: 26
Median: 25
75th: 33
Maximum: 63
Correlation (r): NR
Copollutant models:
NR
Percentage increase CRP
Lag 0: -1.61 (-45.32, 73.07)
Lag 1: 3.43 (-34.93, 62.36)
5-day avg: 2.68 (-52.50,
113.57)
Percentage increase ICAM
Lag 0: 4.39 (-6.21, 15.96)
Lag 1: 0.32 (-7.71, 8.89)
5-day avg: 4.82 (-8.04,
19.39)
Percentage increase IL—6,
Lag 0: 14.46 (-3.36, 35.46)
Lag 1: 7.5 (-6.00, 22.82)
5-day avg: 18.86 (-3.64,
46.18)
Percentage increase TNF-a
Lag 0: 6.75 (-2.25, 16.50)
Lag 1: 2.25 (-4.82, 9.64)
5-day avg: 4.61 (-6.11,
16.50)
TLi et al. (2017)
Boston, MA, U.S.
Ozone: 2005-2008
Panel study
Lags reported: 1-7 day moving
avg
Framingham
Offspring Cohort
n = 3,396
Averaged ozone monitors in
the area and made moving
averages per lag
24-h avg
Mean: 23.7
Correlation (r): NR
Copollutant models:
NR
Percentage increase TNFR2
1-day moving avg: 1.69
(0.45, 2.93)
2-day moving avg: 2.34
(0.84, 3.83)
7-day avg: 5.40 (2.99, 7.81)
CATHGEN = catheterization genetics; CRP = high sensitivity c-reactive protein; CVD = cardiovascular disease; ICAM = inter-cellular adhesion model; IL6 = interleukin 6;
MA = moving average; SWAN = study of women's health across the nation; TNF-a = tumor necrosis factor alpha; TNFR2 = tumor necrosis factor receptor 2; VCAM = vascular cell
adhesion model.
aAII epidemiologic results standardized to a 15 ppb increase in 24 hour avg, 20 ppb increase in 8 hour daily max, 25 ppb increase in 1 hour daily max ozone concentrations, or a
10-ppb increase in seasonal/annual ozone concentrations to facilitate comparability across studies.
4-109
-------
Table 4-29 Study-specific details from controlled human exposure studies of systemic inflammation and
oxidative stress.
Study
Population n, Sex, Age
(Range or Mean ± SD)
Exposure Details
(Concentration, Duration)
Endpoints Examined
Billeretal. (2011)
Healthy adults
n = 11 males, 3 females
Age: 33.1 ± 9.5
0.25 ppm, 15 min of exercise
alternating with 15 min of rest
Markers of systemic
inflammation (before exposure
and 5, 7, and 24 h PE)
Barath et al. (2013)
Healthy adults
n = 36 males, 0 females
Age: 26 ± 1 yr
0.3 ppm, 75 min (alternating 15-min
periods of exercise and rest)
Markers of systemic
inflammation in blood (2 and 6 h
PE)
Kahleetal. (2015)
Healthy adults
n = 14 males, 2 females
Age: 20-36
0.3 ppm, 15 min of exercise alternating
with 15 min of rest one exposure at
22°C other at 32.5°C
Markers of systemic
inflammation (24 h PE)
Ariomandi et al. (2015)
Adults with asthma (n = 10) and adults
without asthma (n = 16)
n = 13 males, 13 females
Age: asthma: 33.5 ± 8.8 yr, healthy:
30.8 ± 6.9 yr
0.1, 0.2 ppm, 4 h (alternating 30 min
periods of exercise and rest)
Markers of systemic
inflammation in blood (before,
immediately after and 20 h PE)
Stieael et al. (2016)
Healthy adults
n = 11 males, 4 females
Age: 23-31 yr
0.3 ppm, 2 h (four 15-min periods of
exercise)
Markers of systemic
inflammation in blood (before,
immediately after, and next day)
Ramanathan et al. (2016)
Healthy adults
n = 13 males, 17 females
Age: 23 ± 4 yr
0.12 ppm, 2 h
HDL antioxidant and
anti-inflammatory capacity
(before exposure and 1 and 20 h
PE)
4-110
-------
Table 4-29 (Continued): Study-specific details from controlled human exposure studies of systemic inflammation
and oxidative stress.
Study
Population n, Sex, Age
(Range or Mean ± SD)
Exposure Details
(Concentration, Duration)
Endpoints Examined
Stieael et al. (2017)
Healthy adults
n = 11 males, 4 females
Age: 23-31 yr
0.3 ppm, 2 h (four 15-min periods of
exercise)
Markers of systemic
inflammation in blood (before,
immediately after, and next day)
Bosson et al. (2013)
Healthy adults
N = 24 males, 19 females
Age: 19-32 yr
0.2 ppm, 2 h (moderate exercise and
rest)
Markers of systemic
inflammation in blood (before,
1.5, 6, and 18 h PE)
Frampton et al. (2017)
Older adults
n = 35 males, 52 females
Age: 55-70 yr
0, 0.07, 0.120 ppm, 3 h (alternating
15-min periods of rest and exercise)
Markers of inflammation in blood
(day before, day of, and up to
22 h PE)
HDL = high-density lipoproteins; PE = post-exposure.
4-111
-------
Table 4-30 Study-specific details from short-term animal toxicological studies of systemic inflammation and
oxidative stress.
Study
Species (Stock/Strain), n, Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Farrai et al. (2012)
Rats (SH)
n = 6/group males, 0 females
Age: 12 weeks
0.2 ppm, 4 h
0.8 ppm, 4 h
Markers of systemic
inflammation in blood (animals
sacrificed 1 h PE)
Martinez-Campos et al. (2012)
Rats (Wistar)
n = 6/group males, 0 females
Age: 10 weeks
0.5 ppm, 4 h/day for 2 weeks (with or
without exercise)
Markers of oxidative stress in
blood (at the end of 2-week
exposure)
Wanq et al. (2013)
Rats (Wistar)
n = 6/group males, 0 females
Age: NR
0.8 ppm, 4 h of ozone followed by
intra-tracheal instillation of saline or
PM2.5 twice/week for 3 weeks
Markers of antioxidants in heart
tissue (heart tissue collected
after 6th exposure)
Markers of systemic
inflammation in blood (blood
drawn after 6th exposure)
Thomson et al. (2013)
Rats (Fischer)
n = 4-6/group males, 0 females
Age: NR
0.4 ppm, 4 h
0.8 ppm, 4 h
mRNA markers of oxidative
stress (tissue collected
immediately PE)
mRNA markers of systemic
inflammation in tissue (tissue
collected immediately PE)
Mclntosh-Kastrinskv et al. (2013)
Mice (C57BL/6)
n = 0 males, 14-15/group females
Age: NR
0.245 ppm, 4 h (aged, FA, or ozone) on
3 separate days outdoors
Heart rate in isolated perfused
hearts (8-11 h PE hearts were
isolated and post-induced
ischemia)
Kurhanewicz et al. (2014)
Mice (C57BL/6)
n = 5-8/group males, 0 females
Age: 10-12 weeks
0.3 ppm, 4 h
Markers of systemic
inflammation in blood (24 h PE)
4-112
-------
Table 4-30 (Continued): Study-specific details from short-term animal toxicological studies of systemic
inflammation and oxidative stress.
Study
Species (Stock/Strain), n, Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Paffett et al. (2015)
Rats (S-D)
n = 65 males, 0 females
Age: 8-12 weeks
1 ppm, 4 h
Markers of systemic
inflammation in blood (serum
collected immediately before
sacrifice)
Kumarathasan et al. (2015)
Rats (Fischer)
n = 8/exposure group 17/control group
males, 0 females
Age: NA
0.4 ppm, 4 h
0.8 ppm, 4 h
Markers of oxidative stress in
blood (immediately and 24 h PE)
Markers of systemic
inflammation in blood
(immediately and 24 h PE)
Ramot et al. (2015)
Rats (FHH)
n = NR males, NR females
Age: 10-12 weeks
Rats (S-D)
n = NR males, NR females
Age: 10-12 weeks
Rats (SH)
n = NR males, NR females
Age: 10-12 weeks
Rats (SHHF)
n = NR males, NR females
Age: 10-12 weeks
Rats (SHSP)
n = NR males, NR females
Age: 10-12 weeks
Rats (WKY)
n = NR males, NR females
Age: 10-12 weeks
Rats (Wistar)
n = NR males, NR females
Age: 10-12 weeks
0.25 ppm, 4 h
0.5 ppm, 4 h
1 ppm, 4 h
Markers of systemic
inflammation in blood
(immediately after and 24 h PE)
Hatch et al. (2015)
Rats (multiple strains)
n = NR males, NR females
Age: 10-12 weeks
0.25, 0.5, or 1 ppm, 4 h
Markers of oxidative stress in
blood (24 h after Day 2 exposure
animals sacrificed)
4-113
-------
Table 4-30 (Continued): Study-specific details from short-term animal toxicological studies of systemic
inflammation and oxidative stress.
Study
Species (Stock/Strain), n, Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Yina etal. (2016)
Mice (KKAy)
n = NR males, 0 females
Age: 7 weeks
0.8 ppm, 4 h/day for 13 consecutive
weekdays
Markers of systemic
inflammation in blood (1 or
3 days PE)
Zhonq etal. (2016)
Mice (Japanese KK)
n = 8/group males, 0 females
Age: NA
0.5 ppm, 4 h/day for 13 consecutive
weekdays
Inflammatory cell populations in
blood (24 h PE)
Markers of systemic
inflammation in blood (24 h PE)
Thomson et al. (2016)
Rats (F344)
n = 5/group males, 0 females
Age: NR
0.8 ppm, 4 h of ozone exposure
(treated with metyrapone,
corticosterone, or vehicle for 1-h prior)
Markers of systemic
inflammation in blood PE
Henriquez et al. (2017)
Rats (WKY)
n = 8/group males, 0 females
Age: 10 weeks
0.8 ppm, 1-2 days of ozone (with or
without pretreatment with propranolol,
mifepristone, or propranolol followed by
mifepristone)
Markers of systemic
inflammation in blood
(immediately PE Day 1 or Day 2)
Cestonaro et al. (2017)
Rats (Wistar)
n = 12/group males, 0 females
Age: 9-10 weeks
0.05 ppm, 24 h/day for 14 or 28 days
0.05 ppm, 3 h/day for 14 and 28 days
Markers of oxidative stress (at
the end of a given exposure)
Snow et al. (2018)
Rats (WKY)
n = 6-8/group males, 0 females
Age: -12 weeks
0.8 ppm, 4 h/day for 2 consecutive
days (diets enriched with coconut,
olive, or fish oil for 8 weeks prior)
Markers of oxidative stress in
blood (2 h PE)
Markers of systemic
inflammation in blood (2 h PE)
Francis et al. (2017)
Mice (C57BL/6J WT and CCR2 null)
n = 0 males, 3-4/group females
Age: 8-11 weeks
0.8 ppm, 3 h
Markers of systemic
inflammation in blood (24-72 h
PE)
CCR2 = C-C chemokine receptor type 2; FHH = fawn-hooded hypertensive; PE = post-exposure; S-D = Sprague-Dawley; SH = spontaneously hypertensive; SHHF = spontaneously
hypertensive heart failure; SHSP = spontaneously hypertensive stroke-prone; WKY = Wistar Kyoto; WT = wild type.
4-114
-------
Table 4-31 Epidemiologic studies of short-term exposure to ozone and cerebrovascular disease.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% Cla
IMechtouff et al. (2012)
Rhone Department, France
Ozone: November 6,
2006-June 6, 2007
Follow-up: November 6,
2006-June 6, 2007
Case-crossover study
AVC69 study
n = 376
Consecutive patients
18 yr or older enrolled
in AVC69 study living
within study area,
excluding nonstroke,
TIA, ICH and
undetermined stroke,
mean age 76.6 yr,
46.3% male
Averaged hourly concentrations
from two to five stationary
monitors, calculated max of 8 h
moving avg. Control exposure
days selected using
time-stratified design matching
on day of week stratifying by
month
8-h max
Mean: 28.02
Median: 29.44
75th: 39.09
Maximum:
65.99
Correlation (r):
PM2.5: -0.2 to 0.53;
NO2
SO2
-0.2 to 0.53;
-0.2 to 0.53;
Copollutant models:
NA
Lag NR: 0.95 (0.68, 1.31)
IBedada et al. (2012)
NORTHSTAR
Manchester and Liverpool, U.K. n = 335 Manchester
Ozone: 2003-2007
Follow-up: 2003-2007
Case-crossover study
709 patients with
incident TIA or minor
stroke confirmed by
stroke physician or
neurologist with
symptom onset within
preceding 6 weeks,
recruited from TIA
clinics, ER or hospital
stroke units in
Northwest England,
age >18 yr with no
comorbidity or
disability, mean age
66.8 yr, 58.7% male.
Averaged hourly concentrations
from eight monitors; separate
estimates for Manchester and
Liverpool. Control exposure
days selected using
time-stratified design matched
on day of week for the event
date in the same month.
8-h avg
Mean: 18.98
Median: 19.29
75th: 24.37
Correlation (r):
NO2: -0.68;
SO2: -0.38; CO
-0.54, PM-io-0.23
Copollutant models:
NA
Liverpool, lag 0: 0.73 (0.52,
1.02)
Liverpool, lag 1: 1.10 (0.79,
1.57)
Liverpool, lag 2: 1.16 (0.82,
1.63)
Liverpool, lag 3: 1.31 (0.92,
1.87)
Manchester, lag 0: 1.29
(0.89, 1.86) Manchester, lag
1: 0.82 (0.57, 1.19)
Manchester, lag 2: 1.11
(0.77, 1.62)
Manchester, lag 3: 0.77
(0.53, 1.13)
4-115
-------
Table 4-31 (Continued): Epidemiologic studies of short-term exposure to ozone and cerebrovascular disease.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% Cla
ICorea et al. (2012)
Mantua County, Italy
Ozone: 2006-2008
Follow-up: 2006-2008
Case-crossover study
Lombardia Stroke Unit
Network registry
n = 781
781 of 1,680
consecutive cases
admitted to stroke unit
over 3 yr between
2006-2008; lived in
urban area within
10 km from a
stationary monitor,
mean age 71.2 yr,
46.8% male
Averaged hourly concentrations
from seven stationary monitors.
Control days were selected
using a bidirectional symmetric
design, Days 1,3,5,7 before
and after the admission for
stroke.
8-h avg
Mean: NR
Correlation (r):
PM2.5: NA
Copollutant models:
PM10, SO2, NO2,
NO, CO, benzene
No association at lag 0 for
any CV event, cardioembolic
disease or ischemic stroke or
stroke subtypes; increment
per 8-h avg ozone not
reported
IXu et al. (2013)
Allegheny County, PA, U.S.
Ozone: September
1994-December 2000
Follow-up: 1994-2000
Case-crossover study
n = 26,210
Stroke cases aged
65 yr and older who
lived in Allegheny
County between 1994
and 2000; 41.2% male
Daily concentrations from U.S.
EPA AQS. Control exposure
days selected using a
time-stratified design matching
on day of week stratified on
month and year.
24-h avg
Mean: NR Correlation (r): NR Lag 0
Copollutant models: Lag 1
Lag 2
Lag 3
NR
1.00 (1.00, 1.00)
1.00 (1.00, 1.00)
1.00 (1.00, 1.00)
1.00 (1.00, 1.00)
Lag 0-3: 1.00 (1.00, 1.01)
4-116
-------
Table 4-31 (Continued): Epidemiologic studies of short-term exposure to ozone and cerebrovascular disease.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% Cla
Suissa et al. (2013)
Nice, France
Ozone: 2007-2011
Follow-up: 2007-2011
Case-crossover study
n = 1,729
2,067 consecutive
patients admitted to
university hospital for
stroke, 1,729 with
diagnosis confirmed
by neurologists using
medical records,
residents of Nice,
mean age 76 yr,
46.7% male. Referent
exposures selected
using time-stratified
approach using day of
week within each
month
Averaged hourly concentrations
from urban stationary monitor
8-h avg
Mean: 40.98
Median: 42.83
75th: 53.49
Maximum:
79.83
Correlation (r):
NO2: -0.54; minimal
temperature 0.67
Copollutant models:
NR
Recurrent stroke, 6
lag 1: 1.43 (1.03, 1
Large artery stroke
lag 1: 0.96 (0.77, 1
All ischemic stroke,
lag 0: 0.97 (0.85, 1
All ischemic stroke
lag 1: 0.99 (0.87, 1
All ischemic stroke,
lag 2: 0.99 (0.88, 1
All ischemic stroke,
lag 3: 1.03 (0.91, 1
All ischemic stroke
lag 0: 0.98 (0.84, 1
All ischemic stroke,
lag 1: 1.02 (0.88, 1
All ischemic stroke
lag 2: 0.98 (0.85, 1
All ischemic stroke
lag 3: 0.99 (0.85, 1
All ischemic stroke,
lag 0: 1.01 (0.93, 1
All ischemic stroke
lag 1: 1.00 (0.89, 1
All ischemic stroke,
lag 2: 1.01 (0.90, 1
All ischemic stroke,
lag 3: 1.03 (0.92, 1
-h avg,
.99)
, 8-h avg,
.21)
8-h avg,
.11)
8-h avg,
13)
8-h avg,
13)
8-h avg,
16)
1-h avg,
15)
1-h avg,
19)
1-h avg,
14)
1-h avg,
15)
24-h avg,
16)
24-h avg,
14)
24-h avg,
14)
24-h avg,
17)
4-117
-------
Table 4-31 (Continued): Epidemiologic studies of short-term exposure to ozone and cerebrovascular disease.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% Cla
IChen etal. (2014)
Edmonton, Alberta, Canada
Ozone: April 17, 1998-March
31, 2002
Follow-up: April 17,
1998-March 31, 2002
Case-crossover study
n = 5,229
Acute ischemic stroke
cases aged 25 yr or
older presenting to
EDs, 50.7% male
Average of hourly mean
concentrations from three
stationary monitors in National
Air Pollution Surveillance
network. Control exposure days
selected using time-stratified
design matching on day of
week stratified on month and
year
1-h max
Mean: 17.22
Median: 15
95th: 41.17
Correlation (r):
PM2.5: -0.15;
NO2: -0.59;
SO2: -0.02; CO
-0.47
Copollutant models:
NR
All seasons, lag 1-8 h: 0.97
(0.86, 1.10)
All seasons, lag 9-16 h: 0.96
(0.86, 1.08)
All seasons, lag 1-24 h: 0.96
(0.84, 1.12)
All seasons, lag 25-48 h:
0.92 (0.80, 1.06)
All seasons, 1-72 h: 0.96
(0.79, 1.18)
Warm season, lag 1-8 h:
0.87 (0.72, 1.02)
Warm season, lag 9-16 h:
0.85 (0.73, 1.01)
Warm season, lag 1-24 h:
0.82 (0.67, 1.01)
Warm season, lag 25-48 h:
0.82 (0.67, 1.00)
Warm season, 1-72 h: 0.78
(0.59, 1.04)
Cold season, lag 1-8 h: 1.16
(0.91, 1.48)
Cold season, lag 9-16 h:
1.14 (0.91, 1.41)
Cold season, lag 1-24 h:
1.22 (0.91, 1.60)
Cold season, lag 25-48 h:
1.09 (0.83, 1.43)
Cold season, 1-72 h: 1.39
(0.93, 2.08)
4-118
-------
Table 4-31 (Continued): Epidemiologic studies of short-term exposure to ozone and cerebrovascular disease.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% Cla
IMiloievic et al. (2014)
England and Wales, U.K.
Ozone: 2003-2009
Follow-up: 2003-2009
Study
HES
n = 426,940
Emergency hospital
admissions for stroke
to NHS hospitals,
2003-2008, in HES
database using
centroid of census
ward; median age
(IQR) 73 yr (60-82 yr),
54% male HES
Data from nearest monitoring
station to residence on event
day. Control exposure days
defined using time-stratified
design using other days of the
month when case occurred
8-h max
Mean: NR
Median: 30.96
75th: 38.58
Correlation (r):
PM2.5: -0.096
NO2: -0.3489
SO2: -0.0849;
0.0302, CO
PM10
0.2973
Copollutant models:
NA
All stroke, lag 0-4: 1.01
(0.99, 1.02)
<70 yr, lag 0-4: 0.99 (0.97,
1.01)
70+ yr, lag 0-4: 1.01 (1.00,
1.03)
Females, lag 0-4: 1.00 (0.98,
1.01)
Males, lag 0-4: 1.01 (1.00,
1.03)
tRodopoulou et al. (2015)
n = 84,269
U.S. AQS data from stationary
Mean: 40
Correlation (r): NR
Cerebrovascular disease, lag
Little Rock, AR, U.S.
Daily emergency room
monitor in Little Rock
Median: 39
Copollutant models:
1: 0.88 (0.77, 1.00)
Ozone: 2002-2012
visits among persons
8-h max
75th: 50
NR
15 yr and older,
19% 65 yr and older,
Follow-up: 2002-2012
Time-series study
42.5% male
IWina et al. (2015)
Nueces County, TX, U.S.
Ozone: January 1, 2000-June
30, 2012
Follow-up: January 1,
2000-June 30, 2012
Case-crossover study
Brain Attack
Surveillance in Corpus
Christi register; active
and passive
surveillance
n = 2,948
Incident ischemic
stroke cases with
exposure data over
45 yr old living in
Nueces County,
median age 71 yr,
56% Mexican
American, 48.7% male
Daily maximal 8-h concentration
from one central monitor in
TCEQ TAMIS. Control
exposure days selected using
time-stratified design matching
on weekday stratifying on
month and year
8-h max
Mean: NR
Median: 35.7
75th: 46.3
Correlation (r): NR
Copollutant models:
NA
Lag 0: 1.02 (0.95, 1.10)
Lag 1: 1.04 (0.97, 1.12)
Lag 2: 1.05 (0.97, 1.12)
Lag 3: 1.02 (0.95, 1.09)
Copollutant model PM2.5,
lag 0: 1.02 (0.95, 1.10)
Copollutant model PM2.5,
lag 1: 1.05 (0.98, 1.13)
Copollutant model PM2.5,
lag 2: 1.05 (0.98, 1.12)
Copollutant model PM2.5,
lag 3: 1.02 (0.95, 1.10)
4-119
-------
Table 4-31 (Continued): Epidemiologic studies of short-term exposure to ozone and cerebrovascular disease.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% Cla
IMontresor-Lopez et al. (2015)
South Carolina, U.S.
Ozone: 2002-2006
Follow-up: 2002-2006
Case-crossover study
n = 21,301
Hospitalized cases
with no prior stroke in
the last 24 mo, 18 yr
old or older and
residents of South
Carolina, mean age
68.7 yr, 47.4% male
Hourly concentrations modeled
using U.S. EPA's Hierarchical
Bayesian Model combining
measurements and CMAQ
outputs at 12-km grid cell
resolution
8-h max
Mean: 46
Median: 46.2
Correlation (r): NR
Copollutant models:
NR
All stroke, lag 0: 0.96 (0.92,
1.00)
All stroke, lag 1: 0.94 (0.90,
1.00)
All stroke, lag 2: 0.96 (0.90,
1.00)
Ischemic stroke, lag 0: 0.96
(0.92, 1.02)
Ischemic stroke, lag 1: 0.94
(0.90, 1.02)
Ischemic stroke, lag 2: 0.94
(0.90, 1.00)
Hemorrhagic stroke, lag 0:
0.90 (0.79, 1.04)
Hemorrhagic stroke, lag 1:
0.96 (0.85, 1.08)
Hemorrhagic stroke, lag 2:
1.02 (0.90, 1.17)
IMaheswaran et al. (2016)
London, U.K.
Ozone: 1995-2006
Follow-up: 1995-2006
Case-crossover study
South London Stroke Averaged hourly concentrations Mean: 15.3
Register
n = 2,590
First-ever ischemic
stroke cases recorded
on stroke register
between 1995 and
2006, mean age
71.7 yr, 50.3% male
from monitors nearest to
residential postal code centroid.
Control exposure days selected
using time-stratified design
matching on weekday stratified
by season
24-h avg
Correlation (r): NR
Copollutant models:
NR
Lag 0
Lag 1
Lag 2
Lag 3
0.99 (0.89, 1.07)
1.00 (0.92, 1.09)
1.04 (0.94, 1.13)
1.05 (0.96, 1.15)
Lag 0-6: 1.09 (0.95, 1.25)
4-120
-------
Table 4-31 (Continued): Epidemiologic studies of short-term exposure to ozone and cerebrovascular disease.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% Cla
IWina et al. (2017a)
Nueces County, TX, U.S.
Ozone: 2000-2012
Follow-up: 2000-2012
Case-crossover study
Brain Attack
Surveillance in Corpus
Christi register
n = 317
First recurrent stroke
on a different day after
incident event
recorded in BASIC,
cases were 45 yr or
older and lived in
Nueces County, and
had air pollution data
available, mean age
71 yr, 47% male,
64% Mexican
American
Daily maximal 8-h concentration
from one central monitor in
TCEQ TAMIS. Control
exposure days selected using
time-stratified design matching
on day of week stratifying on
month and year
8-h max
Median: 35.2
75th: 46.1
Correlation (r): NR
Copollutant models:
NR
Lag 1: recurrent stroke: 0.94
(0.76, 1.14)
Lag 1: severe incident stroke:
1.27 (1.12, 1.41)
IButland et al. (2017)
London, U.K.
Ozone: 2005-2012
Follow-up: 2005-2012
Case-crossover study
South London Stroke
Register
n = 1,799
Stroke cases (and
subtypes) included in
the register; 63% with
ages over 64 yr and
52.4% male
Annual mean concentration with
a 20 m by 20 m spatial
resolution modeled using
measurements, emissions data,
and dispersion modeling, linked
at postal-code level and year,
and then modified to daily mean
concentrations using
time-series scaling factors for
the years 2005-2012. Control
exposure days selected using
time-stratified design matching
on weekday and stratifying on
month
24-h avg
Mean: 18.68
Median: 18.48
75th: 25.03
Correlation (r):
PM2.5: -0.4;
NO2: -0.59; NOx
-0.72, PMio-0.33
Copollutant models:
NR
All stroke, 8-h avg, lag 0:
0.93 (0.74, 1.11)
All stroke, 24-h avg, lag 0:
0.96 (0.85, 1.09)
Ischemic stroke, 8-h avg,
lag 0: 0.96 (0.75, 1.19)
Ischemic stroke, 24-h avg,
lag 0: 0.98 (0.85, 1.13)
Hemorrhagic stroke, 8-h avg,
lag 0: 1.07 (0.65, 1.85)
Hemorrhagic stroke, 24-h
avg, lag 0: 1.09 (0.78, 1.52)
4-121
-------
Table 4-31 (Continued): Epidemiologic studies of short-term exposure to ozone and cerebrovascular disease.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% Cla
tVidale et al. (2017)
n = 4,110
Average daily concentrations
Mean: NR
Correlation (r): NR
Ischemic stroke, lag 0: 0.99
Como, Italy
All residents of Como
from 2 stationary monitors
Copollutant models:
(0.98, 1.02)
Ozone: January
2005-December 2014
Follow-up: January
2005-December 2014
with hospital
admission for acute Ml
or ischemic stroke
between January 2005
and December 2014,
24-h avg
NR
Ischemic stroke, lag 1: 1.00
(0.99, 1.01)
Time-series study
mean age 71 yr,
65% male
IWina et al. (2017b)
Nueces County, TX, U.S.
Ozone: 2000-2012
Follow-up: 2000-2012
Time-series study
Brain Attack
Surveillance in Corpus
Christi register
n = 3,035
Cases recorded in
registry in Nueces
County, TX, mean age
70 yr, 48.7% male,
53% Mexican
American
Daily maximal 8-h concentration
from one central monitor in
TCEQ TAMIS
24-h avg
Mean: NR
Correlation (r): NR
Copollutant models:
NR
Severe incident stroke risk,
lag 1: 1.27 (1.12, 1.41)
Severe incident stroke risk,
lag 1, with neighborhood
disadvantage: 1.27 (1.12,
1.41)
aAII epidemiologic results standardized to a 15 ppb increase in 24 hour avg, 20 ppb increase in 8 hour daily max, 25 ppb increase in 1 hour daily max ozone concentrations, or a
10-ppb increase in seasonal/annual ozone concentrations to facilitate comparability across studies.
4-122
-------
Table 4-32 Epidemiologic studies of short-term exposure to ozone and aggregate cardiovascular disease.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% Cla
IKalantzi et al. (2011)
Magnesia Prefecture, Greece
Ozone: January 1,
2001-December 31, 2007
Follow-up: January 1,
2001-December 31, 2007
Time-series study
n = 4.88/day
Emergency hospital
admissions, counts
over 3 days, during
the study period
among patients over
14 yr of age with a
respiratory or
cardiovascular
disease diagnosis
(ICD-10)
Averaged concentrations
measured continuously from
three stationary monitors within
5 km of the hospital
24-h avg
Mean: 25.53
Correlation (r): NR
Copollutant models:
NR
Lag 0: 1.02 (1.01, 1.03)
Lag 1: 1.02 (1.01, 1.03)
IHunova et al. (2013)
Prague, Czech Republic
Ozone: April-September
2002-2006
Follow-up: April-September
2002-2006
Time-series study
Daily counts of
hospital admissions
among all permanent
residents in Prague
Averaged hourly concentrations
from three stationary monitors;
24-h mean and max daily
running 8-h mean
8-h avg
Mean: 47.463
Median: 45.84
75th: 56.24
Maximum:
83.45
Correlation (r): NR
Copollutant models:
PM10
24-h mean, lag 1: 0.97 (0.95,
1.00)
24-h mean, lag 2: 0.99 (0.97,
1.01)
8-h max, lag 1: 0.99 (0.96,
1.01)
8-h max, lag 2: 1.00 (0.97,
1.02)
IWinquist et al. (2012)
St. Louis MSA, U.S.
Ozone: January 1, 2001-June
27, 2007
Follow-up: January 1,
2001-June 27, 2007
Time-series study
n = 88.8
Counts of daily ED
visits and HA among
people residing in the
St. Louis MSA
Concentrations from U.S. EPA Mean: 36.3
AQS at Tudor Street stationary Maximum'
monitor, data missing 1.9% of m 8
days
8-h max
Correlation (r):
PM25: 0.25;
Copollutant models:
NR
HA, lag 0-4: 0.99 (0.95, 1.00)
ED, lag 0-4: 1.00 (0.98, 1.02)
4-123
-------
Table 4-32 (Continued): Epidemiologic studies of short-term exposure to ozone and aggregate cardiovascular
disease.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% Cla
IRodopoulou et al. (2014)
Dona Ana County, NM, U.S.
Ozone: 2007-2010
Follow-up: 2007-2010
Time-series study
n = ED visits 2,031,
HA 5,161
Daily ED visits and
hospital admissions
for the adult
population (18 yr and
older)
Averaged hourly concentrations Mean: 43.2
from three sites in the county, Median' 43
U.S. EPA AQS data r'
75th: 51
8-h max ...
Maximum: 70
Correlation (r):
PM2.5: -0.05; PM10
0.18
Copollutant models:
NR
HA, lag 1: 1.03 (0.94, 1.13)
ED visits, lag 1: 1.06(0.91,
1.23)
IMiloievic et al. (2014)
England and Wales, U.K.
Ozone: 2003-2008
Follow-up: 2003-2008
Case-crossover study
HES
n = 2,663,067
Emergency hospital
admissions to NHS
hospitals, in HES
database using
centroid of census
ward; median age
(IQR) 73 (60-82),
54% male HES
Data from nearest monitoring
station to residence on event
day. Control exposure days
defined using time-stratified
design using other days of the
month when case occurred
8-h max
Mean: NR
Median: 30.96
75th: 38.58
Correlation (r):
PM2.5: -0.096
NO2
SO2
-0.3489
-0.0849
PM10
All CVD,
1.00)
lag 0-4: 0.99 (0.99,
0.0302, CO -0.2973
Copollutant models:
NA
ISarnat et al. (2015)
St. Louis, MO, U.S.
Ozone: June 1, 2001 -May 30,
2003
Follow-up: June 1, 2001-May
30, 2003
Time-series study
n = 69,679
ED visit records of
patients residing in St.
Louis MSA (eight
counties each in
Missouri and Illinois)
from 36 out of
43 acute care
hospitals
Averaged hourly concentrations Mean: 36.2
in St. Louis from U.S. EPA AQS
8-h max
Correlation (r):
PM2.5: 0.23;
NO2: 0.37;
SO2: -0.04; CO
-0.01
Copollutant models:
NR
Lag 0-2: 0.99 (0.97, 1.02)
tRodoDoulou et al. (2015)
n = 84,269
U.S. AQS data from stationary
Mean: 40
Correlation (r): NR All, lag 1: 0.99 (0.97, 1.01)
Little Rock, AR, U.S.
Daily emergency room
monitor in Little Rock
Median: 39
Copollutant models:
Ozone: 2002-2012
visits among persons
8-h max
75th: 50
NR
15 yr and older,
19% 65 yr and older,
Follow-up: 2002-2012
Time-series study
42.5% male
4-124
-------
Table 4-32 (Continued): Epidemiologic studies of short-term exposure to ozone and aggregate cardiovascular
disease.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% Cla
tVidale et al. (2017)
n = 4,110
Average daily concentrations
Mean: NR
Correlation (r): NR
CVD, lag 0: 1.08 (1.03, 1
.14)
Como, Italy
All residents of Como
from two stationary monitors
Copollutant models:
CVD, lag 1: 1.07 (1.03, 1
.12)
Ozone: January
2005-December 2014
Follow-up: January
2005-December 2014
with hospital
admission for acute Ml
or ischemic stroke
between January 2005
and December 2014,
24-h avg
NR
Time-series study
mean age 71 yr, 65%
male
IHunova et al. (2017)
Prague, Czech Republic
Ozone: April-September
2002-2006
Follow-up: April-September
2002-2006
Time-series study
Daily counts of
hospital admissions
among all permanent
residents in Prague
Averaged hourly mean
concentration from up to three
stationary monitors in Prague
8-h max
Mean: NR Correlation (r): PM10 0.98 (0.95,1.01)
lag-1 0.457
Copollutant models:
NR
IChoi et al. (2011)
Maryland, U.S.
Ozone: June-August 2002
Follow-up: June-August 2002
Time-series study
n = 19,752 total,
214.7 visits/day
All ED visits for CVD
in Maryland
Daily mean concentrations
during June-August 2002 for
each zip-code tabulation area
using block kriging and
monitoring data in Maryland
(16 sites) and sites near border
zip-codes in adjoining states
8-h max
Mean: 76.68
Maximum:
119.42
Correlation (r): NR
Copollutant models:
NR
Lag 0-4: 1.07 (1.03, 1.11)
aAII epidemiologic results standardized to a 15 ppb increase in 24 hour avg, 20 ppb increase in 8 hour daily max, 25 ppb increase in 1 hour daily max ozone concentrations, or a
10-ppb increase in seasonal/annual ozone concentrations to facilitate comparability across studies.
4-125
-------
Table 4-33 Epidemiologic studies of short-term exposure to ozone and cardiovascular mortality.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
(95% Cl)a
IKIemm et al. (2011)
Atlanta, GA, U.S.
August 1998-December 2007
Time-series study
65+
Data from several monitors
8-h max
Mean: 35.54
75th: 47.82
Maximum:
109.07
Correlation (r): NR
Copollutant models:
NR
Lag 0-1: 0.69 (-2.28, 3.75)
ISacks et al. (2012)
Philadelphia, PA, U.S.
May 12, 1992-September 30,
1995
Time-series study
All ages
Single monitor ~6 km
west/southwest of City Hall
8-h max
Mean: 36
Median: 33
Maximum: 110
Correlation (r):
PM25: 0.43;
NO2: 0.18;
SO2: -0.19; CO:
-0.35
Copollutant models:
NR
Harvard (lag 0-1): -1.60
(-5.10, 2.10)
California (lag 0-1): 0.20
(-3.40, 3.90)
Canada (lag 0-1): 0.50
(-3.10, 4.30)
Harvard AT (lag 0-1): 1.30
(-2.10, 4.90)
APHEA2 (lag 0-1): 1.70
(-1.80, 5.30)
NMMAPS (lag 0-1): 2.20
(-1.80, 6.40)
IVanos et al. (2014)
10 Canadian cities
1981-1999
Time-series study
All ages
Monitor located downtown or at Mean: 19.3
city airports within 27 km of
downtown in each city
24-h avg
Correlation (r):
NR
Copollutant models:
NR
All-year (lag 0): 4.65 (1.86,
7.43)
Spring (lag 0): 3.16 (0.25,
6.08)
Summer (lag 0): 5.58 (1.94,
9.21)
Fall (lag 0): 1.96 (0.13, 3.78)
Winter (lag 0): 4.46 (1.55,
7.37)
aAII epidemiologic results standardized to a 15 ppb increase in 24 hour avg, 20 ppb increase in 8 hour daily max, 25 ppb increase in 1 hour daily max ozone concentrations, or a
10-ppb increase in seasonal/annual ozone concentrations to facilitate comparability across studies.
4-126
-------
4.3.2
Long-Term Ozone Exposure
Table 4-34 Epidemiologic studies of long-term exposure to ozone and ischemic heart disease (IHD).
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% Cla
lAtkinson et al. (2013)
Nationwide, U.K.
Ozone: 2002-2007
Follow-up: 2003-2007
Cohort study
English cohort
n = 836,557
Age: 40-89 yr
Annual average from
emission-based model with
1- x 1-km resolution
Mean: NR
Correlation (r):
PM2.5: -0.43
Copollutant models:
NR
Ml; 2003-2007 exposure
period; NO2 copollutant:
0.71 (0.57, 0.87)
Ml; 2003-2007 exposure
period; PM10 copollutant:
0.71 (0.57, 0.87)
Ml; 2002 exposure period:
0.76 (0.62, 1.00)
Ml; 2003-2007 exposure
period: 0.77 (0.62, 0.96)
Ml; 2003-2007 exposure
period; SO2 copollutant:
0.87 (0.71, 1.14)
Kimet al. (2017)
NHIS-NSC
Average from monitors
Mean: 19.93
Correlation (r):
HR for Acute Ml: 0.81
Seoul, South Korea
n = 136,094
linked to participants'
Median: 18.75
PM2.5: 0.67;
(0.75, 0.88)
zip-codes
NO2: 0.68;
Ozone: NR
Healthy adults
75th: 27.08
SO2: 0.84;
Follow-up: 2007-2013
Maximum:
CO: 0.55; PM-io-2.5:
Cohort study
71.12
0.37
Copollutant models:
NR
aAII epidemiologic results standardized to a 15 ppb increase in 24 hour avg, 20 ppb increase in 8 hour daily max, 25 ppb increase in 1 hour daily max ozone concentrations, or a
10-ppb increase in seasonal/annual ozone concentrations to facilitate comparability across studies.
4-127
-------
Table 4-35 Epidemiologic studies of long-term exposure to ozone and atherosclerosis.
Study
Study
Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% Cla
IBreton et al. (2012)
Southern California, U.S.
Ozone: 1980-2009
Follow-up: 2007-2009
Cohort study
TROY
n = 768
College
students
Monthly AQS data from up
to four monitors within
50 km spatially interpolated
to residence using IDW
averaged for ages 6-12
Mean: 23.2
Maximum:
41.8
Correlation (r):
PM2.5: -0.15;
NO2: 0.35; PM10:
-0.05
Copollutant
models: NO2,
PM10, PM2.5
Change in CIMT for exposure averaged during ages
0-5; NO2 copollutant: 10.00 (1.40, 18.60)
Change in CIMT for exposure averaged over
lifetime; PM10 copollutant: 10.13 (-0.51, 20.63)
Change in CIMT for exposure averaged during ages
6-12; PM2.5 copollutant: 10.22 (1.18, 19.35)
Change in CIMT for exposure averaged during ages
6-12: 10.86 (1.94, 19.89)
Change in CIMT for exposure averaged during ages
6-12; PM10 copollutant: 10.86 (1.83, 19.89)
Change in CIMT for exposure averaged during ages
0-5: 7.80 (-0.30, 15.90)
Change in CIMT for exposure averaged during ages
0-5; PM10 copollutant: 8.50 (0.20, 16.90)
Change in CIMT for exposure averaged over
lifetime; NO2 copollutant: 8.86 (-2.03, 19.75)
Change in CIMT for exposure averaged over
lifetime; PM2.5 copollutant: 8.86 (-1.65, 19.37)
Change in CIMT for exposure averaged during ages
0-5; PM2.5 copollutant: 9.10 (0.90, 17.40)
Change in CIMT for exposure averaged during ages
6-12; NO2 copollutant: 9.46 (-0.11, 19.03)
Change in CIMT for exposure averaged over
lifetime: 9.49 (-1.01, 20.00)
4-128
-------
Table 4-35 (Continued): Epidemiologic studies of long-term exposure to ozone and atherosclerosis.
Study
Study
Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% Cla
IBreton et al. (2016)
Southern California, U.S.
Ozone: NR
Follow-up: 2002-2003
Case-control study
Children's
Health Study
n = 459
Public school
children
enrolled in
kindergarten or
first grade; CV
measures at
age 11
IDW from up to four
monitors averaged over
prenatal trimesters; based
on residential history at
birth and age 6-7 yr
Mean: about
40, presented
in box plot
only
Correlation (r):
PM2.5: 0.21-0.41;
NO2: -0.63;
PM10: 0.21-0.66
Copollutant
models: NR
Left CIMT (mm); first trimester: -0.00 (-0.00, 0.00)
Left CIMT (mm); third trimester: -0.00 (-0.00, 0.00)
Right CIMT (mm); third trimester: -0.00 (-0.00,
0.00)
Right CIMT (mm); first trimester: -0.00 (-0.00, 0.00)
Left CIMT (mm); second trimester: 0.00 (-0.00,
0.00)
Right CIMT (mm); second trimester: 0.00 (-0.00,
0.00)
aAII epidemiologic results standardized to a 15 ppb increase in 24 hour avg, 20 ppb increase in 8 hour daily max, 25 ppb increase in 1 hour daily max ozone concentrations, or a
10-ppb increase in seasonal/annual ozone concentrations to facilitate comparability across studies.
Table 4-36 Study-specific details from animal toxicological studies of atherosclerosis.
Study
Species (Stock/Strain), n, Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Gordon et al. (2013)
Rats (BN)
n = 12/treatment group males, 0 females
Age: 4 and 20 mo
0.8 ppm, 6 h/day, 1 day/week for
17 weeks
Potential markers of
atherosclerosis at the end of the
given exposure (28 or 56 days)
Sethi et al. (2012)
Rats (S-D)
n = 6/treatment group males, 0 females
Age: adult
0.8 ppm, 8 h/day for 28 or 56 days
Potential markers of
atherosclerosis at the end of the
given exposure (28 or 56 days)
BN = brown Norway; S-D = Sprague-Dawley.
4-129
-------
Table 4-37 Epidemiologic studies of long-term exposure to ozone and heart failure.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates 95% Cla
tAtkinson et al. (2013)
Nationwide, U.K.
Ozone: 2002-2007
Follow-up: 2003-2007
Cohort study
English cohort
n = 836,557
Age: 40-89 yr
Annual average from
emission-based model with
1- x 1-km resolution
Mean: NR
Correlation (r):
PM2.5: -0.43
Copollutant models:
NR
Heart failure; 2002 exposure
period: 0.66 (0.50, 0.87)
Heart failure; 2003-2007
exposure period: 0.66 (0.49,
0.85)
Heart failure; 2003-2007
exposure period; NO2
copollutant: 0.71 (0.53, 0.94)
Heart failure; 2003-2007
exposure period; PM10
copollutant: 0.71 (0.53, 0.94)
Heart failure; 2003-2007
exposure period; SO2
copollutant: 0.71 (0.53, 0.94)
tKimetal. (2017)
NHIS-NSC
Average from monitors linked to
Mean: 19.93
Correlation (r): HR forCHF: 0.76 (0.71, 0.81)
Seoul, South Korea
n = 136,094
participants' zip-codes
Median: 18.75
PM2.5: 0.67;
NO2: 0.68;
SO2: 0.84; CO: 0.55;
Ozone: NR
Healthy adults
75th: 27.08
Follow-up: 2007-2013
Maximum:
PM10-2.5: 0.37
Cohort study
71.12
Copollutant models:
NR
aAII epidemiologic results standardized to a 15 ppb increase in 24 hour avg, 20 ppb increase in 8 hour daily max, 25 ppb increase in 1 hour daily max ozone concentrations, or a
10-ppb increase in seasonal/annual ozone concentrations to facilitate comparability across studies.
4-130
-------
Table 4-38 Study-specific details from animal toxicological studies of impaired heart function.
Study
Exposure Details
Species (Stock/Strain), n, Sex, Age (Concentration, Duration)
Endpoints Examined
PereDU et al. (2012)
Rats (S-D) 0.8 ppm, 8 h/day for 28 or 56 days
n = 6/treatment group males, 0 females
Age: adult
LVDP (28 and 56 days PE)
Sethi et al. (2012)
Rats (S-D) 0.8 ppm, 8 h/day for 28 or 56 days
n = 6/treatment group males, 0 females
Age: adult
LVDP (28 and 56 days PE)
LVDP = left ventricular developed pressure; PE
= post-exposure; S-D = Sprague-Dawley.
Table 4-39 Study-specific details from animal toxicological studies of vascular function.
Study
Exposure Details
Species (Stock/Strain), n, Sex, Age (Concentration, Duration)
Endpoints Examined
Gordon et al. (2013)
Rats (BN) 0.8 ppm, 6 h/day, 1 day/week for
n = 12/treatment group males, ^ weeks
0 females
Age: 4 and 20 mo
Markers of endothelial function
blood drawn day after final
exposure
BN = brown Norway.
4-131
-------
Table 4-40 Epidemiologic studies of long-term exposure to ozone and blood pressure.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates (95% Clf
IDona et al.
(2013b)
Three northeastern
cities, China
Ozone: 2006-2008
Follow-up:
2009-2010
Cross-sectional
study
33 Communities
Chinese Health Study
n = 24,845
Age: 18-74 yr
3-yr avg
concentration
from single
monitor within
1 mile of
residence
8-h avg
Mean: 24.7
Median: 25
Maximum:
35.5
Correlation
(r): NR
Copollutant
models: NR
Absolute increase in
Absolute increase in
Absolute increase in
Absolute increase in
Absolute increase in
Absolute increase in
OR for hypertension;
OR for hypertension;
OR for hypertension;
OR for hypertension;
OR for hypertension;
OR for hypertension;
DBP; women: 0.02 (-0.26, 0.31)
SBP; women: 0.04 (-0.45, 0.53)
DBP; all: 0.34 (0.13, 0.55)
DBPI; men: 0.53 (0.22, 0.83)
SBP; all: 0.66 (0.32, 1.01)
SBP; men: 0.95 (0.47, 1.44)
55-64 yr: 1.02 (0.93, 1.13)
women: 1.06 (0.92, 1.16)
<55 yr: 1.12 (1.06, 1.18)
all: 1.12 (1.05, 1.18)
65+ yr: 1.14 (0.96, 1.35)
men: 1.19 (1.04, 1.34)
4-132
-------
Table 4-40 (Continued): Epidemiologic studies of long-term exposure to ozone and blood pressure.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates (95% Cl)a
IZhao et al. (2013)
Three northeastern
cities, China
Ozone: 2006-2008
Follow-up:
2009-2010
Cross-sectional
study
33 Communities
Chinese Health Study
n = 24,845
Age: 18-74 yr
3-yr avg
concentration
from single
monitor within
1 mile of
residence
8-h avg
Mean: 24.7 Correlation Absolute difference for SBP; normal weight—female: -0.12
Median: 25 (r): NR (-0.66, 0.44)
Maximum' Copollutant Absolute difference for DBP; normal weight—female: -0.13
35 5 models: NR (-0.45, 0.20)
Absolute difference for DBP; overweight—female: -0.15 (-0.66,
0.35)
Absolute difference for SBP; overweight—female: 0.14 (-0.75,
1.04)
Absolute difference for DBP; obese—female: 0.21 (-0.95, 1.37)
Absolute difference for DBP; normal weight—all: 0.26 (0.02, 0.51)
Absolute difference for SBP; normal weight—all: 0.31 (-0.10, 0.72)
Absolute difference for DBP; overweight—all: 0.32 (-0.02, 0.65)
Absolute difference for SBP; normal weight—male: 0.39 (-0.20,
0.99)
Absolute difference for SBP; obese—female: 0.50 (-1.74, 2.74)
Absolute difference for DBP; overweight—male: 0.59 (0.14, 1.04)
Absolute difference for DBP; normal weight—male: 0.61 (0.25,
0.97)
OR for hypertension; obese—female: 0.90 (0.69, 1.16)
OR for hypertension; normal weight—female: 0.95 (0.86, 1.04)
OR for hypertension; normal weight—all: 1.05 (0.99, 1.12)
OR for hypertension; overweight—female: 1.06 (0.95, 1.19)
OR for hypertension; normal weight—male: 1.09(1.01, 1.19)
OR for hypertension; overweight—all: 1.17 (1.09, 1.25)
Absolute difference for DBP; obese—all: 1.19 (0.33, 2.06)
OR for hypertension; obese—all: 1.22(1.03, 1.44)
OR for hypertension; overweight—male: 1.22(1.12, 1.33)
OR for hypertension; obese—male: 1.44(1.14, 1.82)
Absolute difference for SBP; overweight—all: 1.56 (0.98, 2.12)
Absolute difference for DBP; obese—male: 1.76 (0.58, 2.93)
Absolute difference for SBP; overweight—male: 2.36 (1.65, 3.08)
Absolute difference for SBP; obese—all: 3.15(1.61, 4.67)
Absolute difference for SBP; Obese—male: 4.04 (2.20, 5.86)
4-133
-------
Table 4-40 (Continued): Epidemiologic studies of long-term exposure to ozone and blood pressure.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates (95% Cl)a
IDona etal. (2014)
Seven northeastern
cities, China
Ozone: 2009-2012
Follow-up:
2012-2013
Cross-sectional
study
n = 9,354
School-aged children,
5-17 yr
4-yr avg
concentration
from monitor
within 1 km of
school
8-h avg
Mean: 54
Median: 21.9
Correlation
(r): NR
Copollutant
models: NR
Absolute increase in SBP; girls: 0.20 (0.16, 0.24)
Absolute Increase in SBP; breastfeeding only: 0.22 (0.18, 0.25)
Absolute increase in SBP; all: 0.22 (0.19, 0.25)
Absolute increase in SBP; no breastfeeding: 0.22 (0.16, 0.28)
Absolute Increase in DBP; breastfeeding only: 0.23 (0.21, 0.27)
Absolute increase in SBP; boys: 0.23 (0.19, 0.28)
Absolute increase in DBP; all: 0.25 (0.22, 0.27)
Absolute increase in DBP; girls: 0.25 (0.22, 0.29)
Absolute increase in DBP; boys: 0.25 (0.21, 0.28)
Absolute increase in DBP; no breastfeeding: 0.27 (0.22, 0.32)
OR for hypertension; breastfeeding only: 1.04 (1.03, 1.05)
OR for hypertension; boys: 1.05 (1.04, 1.06)
OR for hypertension; girls: 1.05 (1.04, 1.06)
OR for hypertension; no breastfeeding: 1.07 (1.05, 1.08)
Ivan Rossem et al.
(2015)
Boston, MA, U.S.
Ozone: 1999-2002
Follow-up:
1999-2002
Cohort study
Project Viva
n = 1,131
Newborn infants
Area-wide Median: 23.4
average of AQS 75^- 29 2
monitors (n = 4)
averaged over
trimesters
24-h avg
Correlation
(r):
PM2.5: -0.13;
NO2: -0.69;
BC: -0.35;
NOx: -0.92
Copollutant
models: NR
Increase in SBP for 3rd trimester exposure: -1.84 (-3.31, -0.29)
Increase in SBP for 1st trimester exposure: 0.92 (-0.77, 2.69)
Increase in SBP for 2nd trimester exposure: 1.33 (0.23, 2.34)
4-134
-------
Table 4-40 (Continued): Epidemiologic studies of long-term exposure to ozone and blood pressure.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates (95% Cl)a
IDona etal. (2015)
Seven northeastern
cities, China
Ozone: 2009-2012
Follow-up:
2012-2013
Cross-sectional
study
n = 9,354
School-aged children,
5-17 yr
4-yr avg
concentration
from monitor
within 1 km of
school
8-h avg
Mean: 54
Median: 21.9
Maximum:
287
Correlation
(r):
N02: 0.33;
SO2: 0.6;
PM10: 0.85
Copollutant
models: NR
Absolute increase in SBP; normal weight children: 0.13 (0.10,
0.17)
SBP; normal weight girls: 0.13 (0.09, 0.18)
SBP; normal weight boys: 0.14 (0.08, 0.19)
SBP; overweight boys: 0.14 (0.03, 0.25)
DBP; normal weight boys: 0.17 (0.13, 0.22)
SBP; overweight children: 0.17 (0.09, 0.25)
DBP; normal weight children: 0.19 (0.16,
Absolute increase in
Absolute increase in
Absolute increase in
Absolute increase in
Absolute increase in
Absolute increase in
0.22)
OR for hypertension
OR for hypertension
OR for hypertension
OR for hypertension
OR for hypertension
OR for hypertension
OR for hypertension
OR for hypertension
OR for hypertension
normal weight girls: 0.19 (0.16, 0.23)
overweight girls: 0.20 (0.08, 0.32)
overweight girls: 0.24 (0.13, 0.35)
overweight children: 0.25 (0.18, 0.33)
obese boys: 0.25 (0.13, 0.36)
obese children: 0.25 (0.16, 0.34)
obese boys: 0.26 (0.17, 0.35)
obese girls: 0.26 (0.10, 0.42)
obese children: 0.27 (0.20, 0.35)
overweight boys: 0.29 (0.19, 0.38)
obese girls: 0.30 (0.17, 0.44)
normal weight boys: 1.03 (1.02, 1.04)
normal weight children: 1.03 (1.03, 1.04)
normal weight girls: 1.03 (1.02, 1.05)
overweight boys: 1.05 (1.03, 1.08)
overweight children: 1.05 (1.03, 1.07)
overweight girls: 1.05 (1.03, 1.07)
obese boys: 1.06 (1.04, 1.08)
obese children: 1.07 (1.05, 1.08)
obese girls: 1.07 (1.04, 1.10)
in
DBP
in
SBP
in
DBP
in
DBP
in
SBP
in
SBP
in
DBP
in
SBP
in
DBP
in
DBP
in
DBP
4-135
-------
Table 4-40 (Continued): Epidemiologic studies of long-term exposure to ozone and blood pressure.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates (95% Cl)a
TLiu et al. (2016)
n = 3,762
Annual average
Mean: 27
Correlation
Absolute
difference
for DBP
AH I 0-4: -0.10 (-0.85, 0.65)
Taipei, Taiwan
Age: 20-80 yr
of nearest
monitor
Maximum:
(r): NR
Absolute
difference
for SBP
AH I 5-29: -1.20 (-2.12, -0.27)
Ozone: 2005-2012
28.7
Copollutant
models: NR
Absolute
difference
for SBP
AH I 30+: -1.54 (-2.48, -0.61)
Follow-up:
Absolute
difference
for SBP
all: -1.54 (-2.11, -0.98)
2005-2012
Absolute
difference
for SBP
AH I 0-4: -1.92 (-2.96, -0.87)
Cohort study
Absolute
Absolute
Absolute
difference
difference
difference
for DBP
for DBP
for DBP
AH I 30+: 0.19 (-0.46, 0.84)
all: 0.27 (-0.12, 0.66)
AH I 5-29: 0.70 (0.10, 1.30)
IBreton et al.
(2016)
Southern CA, U.S.
Ozone: NR
Follow-up:
2002-2003
Case-control study
Children's Health
Study
n = 459
Public school children
enrolled in
kindergarten or first
grade; CV measures
at age 11 yr
IDW from up to
four monitors
averaged over
prenatal
trimesters;
based on
residential
history at birth
and age 6-7 yr
Mean: about
40, presented
in box plot
only
Correlation
(r): PM2.5:
0.21-0.41;
NO2: -0.63;
PM10:
0.21-0.66
Copollutant
models: NR
DBP (mm Hg); second trimester: -0.04 (-0.32, 0.24)
SBP (mm Hg); first trimester: -0.14 (-0.53, 0.25)
DBP (mm Hg); first trimester: -0.15 (-0.43, 0.13)
SBP (mm Hg); second trimester: 0.05 (-0.33, 0.43)
SBP (mm Hg); third trimester: 0.05 (-0.39, 0.48)
DBP (mm Hg); third trimester: 0.07 (-0.25, 0.39)
ICooaan et al.
(2017)
Nationwide, U.S.
Ozone: 2007-2008
Follow-up:
1995-2011
Cohort study
BWHS
n = 33,771
African American
women
CMAQ
downscaler
8-h max
Mean: 37.4
Maximum:
56.4
Correlation
(r):
PM2.5: 0.14;
NO2: -0.54;
Copollutant
models: NO2;
PM2.5
Hypertension incidence copollutant—NO2: 1.06 (0.91, 1.23)
Hypertension incidence copollutant—PM2.5: 1.12 (0.99, 1.30)
Hypertension incidence: 1.14 (1.00, 1.28)
4-136
-------
Table 4-40 (Continued): Epidemiologic studies of long-term exposure to ozone and blood pressure.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates (95% Cl)a
lYana et al. (2017)
Three northeastern
cities, China
Ozone: 2006-2008
Follow-up: 2009
Cross-sectional
study
33 Communities Data from
Chinese Health Study monitoring
n = 24,845
Age: 18-74 yr
stations
8-h avg
Mean: 24.7
Maximum:
35.5
Correlation
(r): NR
Copollutant
models: NR
Increase in DBP
Increase in SBP
Increase in SBP
(-0.55, 0.61)
Increase in DBP
Increase in DBP
(-0.20, 0.49)
Increase in DBP
Increase in DBP
Increase in DBP
Increase in DBP
Increase in DBP
Increase in SBP
Increase in DBP
Increase in DBP
Increase in SBP
Increase in SBP
Increase in SBP
Increase in SBP
Increase in SBP
Increase in SBP
Prehypertension;
Prehypertension;
Prehypertension;
Prehypertension;
Prehypertension;
Increase in SBP
Prehypertension;
(mm Hg
(mm Hg
(mm Hg
(mm Hg
(mm Hg
hypertensive: -0.11 (-0.42, 0.20)
hypertensive: 0.03 (-0.49, 0.55)
hypertensive without medication: 0.04
55+ yr: 0.15 (-0.23, 0.55)
hypertensive without medication: 0.15
(mm Hg), prehypertensive: 0.18 (0.03, 0.33)
(mm Hg), men: 0.35 (0.12, 0.49)
(mm Hg), normotensive: 0.35 (0.16, 0.55)
(mm Hg), <35 yr: 0.45 (0.16, 0.73)
(mm Hg), all: 0.45 (0.29, 0.60)
(mm Hg), normotensive: 0.48 (0.22, 0.75)
(mm Hg), women: 0.49 (0.28, 0.71)
(mm Hg), 35-55 yr: 0.50 (0.29, 0.71)
(mm Hg), men: 0.61 (0.29, 1.22)
(mm Hg), 55+ yr: 0.85 (0.25, 1.46)
(mm Hg), prehypertensive: 0.87 (0.64, 1.10)
(mm Hg), men 35-55 yr: 0.96 (0.65, 1.28)
(mm Hg), <35 yr: 1.03 (0.64, 1.41)
(mm Hg), all: 1.03 (0.79, 1.25)
men: 1.03 (0.91, 1.17)
<35 yr: 1.08 (1.03, 1.15)
35-55 yr: 1.10(1.04, 1.14)
all: 1.12 (0.99, 1.25)
women: 1.18 (1.05, 1.33)
(mm Hg), women: 1.22 (0.89, 1.55)
55+ yr: 1.24 (1.14, 1.33)
4-137
-------
Table 4-40 (Continued): Epidemiologic studies of long-term exposure to ozone and blood pressure.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates (95% Cl)a
ICole-Hunter et al. TAPAS/EXPOsOMICS Annual average Mean: 22
(2018)
Barcelona, Spain
Ozone: 2011-2014
Follow-up:
2011-2014
Cohort study
n = 57
Healthy adults
Age: 18-60 yr
assigned to
participant
address from
closest
reference
station
Maximum:
32.9
Correlation
(r):
PM25: -0.4;
NO2: -0.21;
PM10: -0.56;
NOx: -0.37
Copollutant
models:
PM10
Increase in SBP (mm Hg): 4.13 (-1.13, 9.38)
Increase in SBP (mm Hg) copollutant PM10: 4.87 (-1.36, 11.10)
Increase in DBP (mm Hg): 6.42 (2.15, 10.69)
Increase in DBP (mm Hg) copollutant PM10: 7.60 (2.64, 12.55)
IChuana et al.
(2011)
Taiwan, Taiwan
Ozone: 2000
Follow-up: 2000
Case-crossover
study
SEBAS
n = 1,023
Age: 54+ yr
City- or
countywide
annual average
from monitoring
stations
Mean: 22.95
Maximum:
42.3
Correlation
(r): NR
Copollutant
models: NR
Change in DBP (mm Hg): 22.97 (20.27, 25.66)
Change in SBP (mm Hg): 24.03 (18.88, 29.20)
aAII epidemiologic results standardized to a 15 ppb increase in 24 hour avg, 20 ppb increase in 8 hour daily max, 25 ppb increase in 1 hour daily max ozone concentrations, or a
10-ppb increase in seasonal/annual ozone concentrations to facilitate comparability across studies.
4-138
-------
Table 4-41 Study-specific details from animal toxicological studies of blood pressure.
Study
Species (Stock/Strain), n, Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Gordon et al. (2013)
Rats (BN)
n = 12/treatment group males, 0 females
Age: 4 and 20 mo
0.8 ppm, 6 h/day, 1 day/week for
17 weeks
Blood pressure (biweekly
through week 15)
BN = brown Norway.
Table 4-42 Study-specific details from animal toxicological studies of heart rate variability (HRV), heart rate
(HR).
Study
Species (Stock/Strain), n, Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Gordon et al. (2014)
Rats (BN)
n = 12/treatment group males, 0 females
Age: 4 and 20 mo
1 ppm, 6 h/day, 2 day/week for
13 weeks
Heart rate (rats implanted with
telemeter)
Gordon et al. (2013)
Rats (BN)
n = 12/treatment group males, 0 females
Age: 4 and 20 mo
0.8 ppm, 6 h/day, 1 day/week for
17 weeks
Heart rate (biweekly through
week 15)
BN = brown Norway; HR = heart rate; HRV = heart rate variability.
4-139
-------
Table 4-43 Study-specific details from animal toxicological studies of coagulation.
Study
Species (Stock/Strain), n, Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Gordon et al. (2013)
Rats (BN)
n = 12/treatment group males, 0 females
Age: 4 and 20 mo
0.8 ppm, 6 h/day, 1 day/week for
17 weeks
mRNA levels of coagulation
factors in aorta tissue collected a
day after final exposure
BN = brown Norway.
4-140
-------
Table 4-44 Study-specific details from animal toxicological studies of inflammation.
Study
Species (Stock/Strain), n, Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
PereDU et al. (2012)
Rats (S-D)
n = 6/treatment group males, 0 females
Age: adult
0.J
3 ppm, 8 h/day for 28 or 56 days
Markers of oxidative stress (28
and 56 days PE)
Markers of systemic
inflammation in heart tissue (28
and 56 days PE)
Sethi et al. (2012)
Rats (S-D)
n = 6/treatment group males, 0 females
Age: adult
0.J
3 ppm, 8 h/day for 28 or 56 days
Markers of oxidative stress (28
and 56 days PE)
Markers of systemic
inflammation in heart tissue (28
and 56 days PE)
Gordon et al. (2013)
Rats (BN)
n = 12/treatment group males, 0 females
Age: 4 and 20 mo
0.8 ppm, 6 h/day, 1 day/week for
17 weeks
Histology (17 weeks PE)
Markers of oxidative stress
(17 weeks PE)
Markers of systemic
inflammation (17 weeks PE)
Miller et al. (2016) Rats (WKY) 1 ppm, 5 h/day, 3 consecutive Markers of systemic
n = fourto five/treatment group males, days/week for 13 weeks inflammation (13 weeks PE)
0 females
Age: 10 weeks
BN = brown Norway; PE = post-exposure; S-D = Sprague-Dawley; WKY = Wistar Kyoto.
4-141
-------
Table 4-45 Epidemiologic studies of long-term exposure to ozone and cerebrovascular disease.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% Cla
lAtkinson et al. (2013)
Nationwide, U.K.
Ozone: 2002-2007
Follow-up: 2003-2007
Cohort study
English Cohort
n = 836,557
Ages 40-89
Annual average from
emission-based model with
1- x 1-km resolution
Mean: NR
Correlation (r):
PM2.5: -0.43
Copollutant models:
NR
Stroke; 2003-2007 exposure
period; PM10 copollutant: HR:
0.94 (0.76, 1.22)
CBVD; 2003-2007 exposure
period: HR: 0.96 (0.90, 1.02)
Stroke; 2002 exposure
period: HR: 1.00 (0.82, 1.30)
Stroke; 2003-2007 exposure
period; NO2 copollutant: HR:
1.00 (0.76, 1.22)
Stroke; 2003-2007 exposure
period: HR: 1.02 (0.81, 1.28)
Stroke; 2003-2007 exposure
period; SO2 copollutant: HR:
1.07 (0.87, 1.38)
KDonq et al., 2013a)
33 Communities
3-yr avg concentration from
Mean: 24.7
Correlation (r):
OR for stroke;
female: 1.13
Three northeastern cities, China
Chinese Health Study
single monitor within 1 mile of
Median: 25
NO2: 0.45
(0.92, 1.39)
Ozone: 2006-2008
n = 24,845
residence
Maximum: 35.5
;S02: 0.87; PM10:
0.80
Copollutant models:
OR for stroke;
all: 1.14 (0.99,
Follow-up: 2009-2010
Age: 18-74 yr
8-h avg
1.30)
OR for stroke;
male: 1.14
Cross-sectional study
NR
(0.95, 1.37)
ISpiezia et al. (2014)
Padua, Italy
Ozone: NR
Follow-up: 2008-2012
Case-control study
n = 105 (33 cases)
Patients with "high
probability" of a PE
Average monthly mean
concentrations from nearest
monitoring site
75th: 37 Correlation (r): NR
Copollutant models:
NR
OR compares exposures
>37 ppb to those lower than
37 ppb: 0.83 (0.26, 2.70)
4-142
-------
Table 4-45 (Continued): Epidemiologic studies of long-term exposure to ozone and cerebrovascular disease.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% Cla
TQin et al. (2015)
Three northeastern cities, China
Ozone: 2006-2008
Follow-up: 2009-2010
Cross-sectional study
33 Communities
Chinese Health Study
n = 24,845
Age: 18-74 yr
3-yr avg concentration from
single monitor within 1 mile of
residence
8-h avg
Mean: 24.7
Median: 25
Maximum: 35.5
Correlation (r): NR
Copollutant models:
NR
OR for stroke; BMI
<25 kg/m2—female: 0.89
(0.66, 1.19)
OR for stroke; normal weight:
0.98 (0.83, 1.16)
OR for stroke; BMI
<25 kg/m2—all: 1.03 (0.87,
1.21)
OR for stroke; BMI
<25 kg/m2—male: 1.14 (0.94,
1.39)
OR for stroke; overweight:
1.26 (1.05, 1.52)
OR for stroke; BMI
>25 kg/m2—male: 1.27 (0.99,
1.63)
OR for stroke; BMI
>25 kg/m2—all: 1.29 (1.08,
1.54)
OR for stroke; BMI
>25 kg/m2—female: 1.32
(1.02, 1.71)
OR for stroke; obese: 1.42
(0.84, 2.38)
IKimetal. (2017)
Seoul, South Korea
Ozone: NR
Follow-up: 2007-2013
Cohort study
NHIS-NSC
n = 136,094
Healthy adults
Average from monitors linked to
participants' zip-codes
Mean: 19.93
Median: 18.75
75th: 27.08
Maximum:
71.12
Correlation (r):
PM2.5: 0.67;
NO2: 0.68;
SO2: 0.84; CO: 0.55;
PMio-2.5: 0.37
Copollutant models:
NR
HR for ischemic stroke: 0.73
(0.68, 0.77)
HR for stroke: 0.73 (0.69,
0.76)
HR for hemorrhagic stroke:
0.74 (0.67, 0.81)
aAII epidemiologic results standardized to a 15 ppb increase in 24 hour avg, 20 ppb increase in 8 hour daily max, 25 ppb increase in 1 hour daily max ozone concentrations, or a
10-ppb increase in seasonal/annual ozone concentrations to facilitate comparability across studies.
4-143
-------
Table 4-46 Epidemiologic studies of long-term exposure to ozone and aggregate cardiovascular disease.
Study
Study
Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% Cla
IDonq et al. (2013a) 33 Communities 3-yr avg concentration
Three northeastern cities, China Chinese Health from single monitor within
_ Study 1 mile of residence
Ozone: 2006-2008 3
„„„„ n =24,845 8-h avg
Follow-up: 2009-2010
, x , Age: 18-74 yr
Cross-sectional study
Mean: 24.7
Median: 25
Maximum: 35.5
Correlation (r):
N02: 0.45; S02: 0.87;
PM10: 0.80
Copollutant models:
NR
OR for CVDs; female: 0.98 (0.64, 1.43)
OR for CVDs; all: 1.08 (0.85, 1.37)
OR for CVD; male: 1.10 (0.80, 1.52)
TQin et al. (2015)
Three northeastern cities, China
Ozone: 2006-2008
Follow-up: 2009-2010
Cross-sectional study
33 Communities
Chinese Health
Study
n = 24,845
Age: 18-74 yr
3-yr avg concentration
from single monitor within
1 mile of residence
8-h avg
Mean: 24.7
Median: 25
Maximum: 35.5
Correlation (r): NR
Copollutant models:
NR
OR for CVDs
0.71 (0.42, 1
OR for CVDs
(0.88, 1.31)
OR for CVDs
1.31)
OR for CVDs
1.09 (0.89, 1.
OR for CVDs
1.15 (0.94, 1.
OR for CVDs
1.16 (0.96, 1
OR for CVDs
1.17 (0.94, 1.
OR for CVDs
1.23 (0.82, 1
OR for CVDs
BMI <25 kg/m2—female:
.22)
normal weight: 1.07
overweight: 1.07 (0.87,
BMI <25 kg/m2—all:
.33)
BMI >25 kg/m2—male:
41)
BMI >25 kg/m2—all:
39)
BMI <25 kg/m2—male:
.45)
BMI >25 kg/m2—female:
86)
;; obese: 1.50 (1.02, 2.21)
aAII epidemiologic results standardized to a 15 ppb increase in 24 hour avg, 20 ppb increase in 8 hour daily max, 25 ppb increase in 1 hour daily max ozone concentrations, or a
10-ppb increase in seasonal/annual ozone concentrations to facilitate comparability across studies.
4-144
-------
Annex for Appendix 4: Evaluation of Studies on Health Effects of
Ozone
This annex describes the approach used in the Integrated Science Assessment (ISA) for Ozone
and Related Photochemical Oxidants to evaluate study quality in the available health effects literature. As
described in the Preamble to the ISA (U.S. EPA. 2015). causality determinations were informed by the
integration of evidence across scientific disciplines (e.g., exposure, animal toxicology, epidemiology) and
related outcomes and by judgments of the strength of inference in individual studies. Table Annex 4-1
describes aspects considered in evaluating study quality of controlled human exposure, animal
toxicological, and epidemiologic studies. The aspects found in Table Annex 4-1 are consistent with
current best practices for reporting or evaluating health science data.1 Additionally, the aspects are
compatible with published U.S. EPA guidelines related to cancer, neurotoxicity, reproductive toxicity,
and developmental toxicity (U.S. EPA. 2005. 1998. 1996b. 1991).
These aspects were not used as a checklist, and judgments were made without considering the
results of a study. The presence or absence of particular features in a study did not necessarily lead to the
conclusion that a study was less informative or to exclude it from consideration in the ISA. Further, these
aspects were not used as criteria for determining causality in the five-level hierarchy. As described in the
Preamble, causality determinations were based on judgments of the overall strengths and limitations of
the collective body of available studies and the coherence of evidence across scientific disciplines and
related outcomes. Table Annex 4-1 is not intended to be a complete list of aspects that define a study's
ability to inform the relationship between ozone and health effects, but it describes the major aspects
considered in this ISA to evaluate studies. Where possible, study elements, such as exposure assessment
and confounding (i.e., bias due to a relationship with the outcome and correlation with exposures to
ozone), are considered specifically for ozone. Thus, judgments on the ability of a study to inform the
relationship between an air pollutant and health can vary depending on the specific pollutant being
assessed.
1 For example, NTP OHAT approach (Roonev et al.. 20141. IRIS Preamble (U.S. EPA. 2013b). ToxRTool
(Klimisch et al.. 19971. STROBE guidelines (von Elm et al.. 20071. and ARRIVE guidelines (Kilkenny et al.. 20101.
4-145
-------
Table Annex 4-1 Scientific considerations for evaluating the strength of
inference from studies on the health effects of ozone.
Study Design
Controlled Human Exposure:
Studies should clearly describe the primary and any secondary objectives of the study or specific hypotheses being
tested. Study subjects should be randomly exposed without knowledge of the exposure condition. Preference is given
to balanced crossover (repeated measures) or parallel design studies which include controlled exposures (e.g., to
clean filtered air). In crossover studies, a sufficient and specified time between exposure days should be provided to
avoid carry over effects from prior exposure days. In parallel design studies, all arms should be matched for individual
characteristics such as age, sex, race, anthropometric properties, and health status. In studies evaluating effects of
disease, appropriately matched healthy controls are desired for interpretative purposes.
Animal Toxicology:
Studies should clearly describe the primary and any secondary objectives of the study or specific hypotheses being
tested. Studies should include appropriately matched controlled exposures (e.g., to clean filtered air, time matched)
and use methods to limit differences in baseline characteristics of control and exposure groups. Studies should
randomize assignment to exposure groups and where possible conceal allocation to research personnel. Groups
should be subjected to identical experimental procedures and conditions; animal care including housing, husbandry,
etc. should be identical between groups. Blinding of research personnel to study group may not be possible due to
animal welfare and experimental considerations; however, differences in the monitoring or handling of animals in all
groups by research personnel should be minimized.
Epidemiology:
Inference is stronger for studies that clearly describe the primary and any secondary aims of the study or specific
hypotheses being tested.
For short-term exposure, time-series, case-crossover, and panel studies are emphasized over cross-sectional studies
because they examine temporal correlations and are less prone to confounding by factors that differ between
individuals (e.g., SES, age). Panel studies with scripted exposures, in particular, can contribute to inference because
they have consistent, well-defined exposure durations across subjects, measure personal ambient pollutant
exposures, and measure outcomes at consistent, well-defined lags after exposures. Studies with large sample sizes
and those conducted over multiple years are considered to produce more reliable results. Additionally, multicity
studies are preferred over single-city studies because they examine associations for large diverse geographic areas
using a consistent statistical methodology, avoiding the publication bias often associated with single-city studies.3 If
other quality parameters are equal, multicity studies carry more weight than single-city studies because they tend to
have larger sample sizes and lower potential for publication bias.
For long-term exposure, inference is considered to be stronger for prospective cohort studies and case-control studies
nested within a cohort (e.g., for rare diseases) than cross-sectional, other case-control, or ecological studies. Cohort
studies can better inform the temporality of exposure and effect. Other designs can have uncertainty related to the
appropriateness of the control group or validity of inference about individuals from group-level data. Study design
limitations can bias health effect associations in either direction.
4-146
-------
Table Annex 4-1 (Continued): Scientific considerations for evaluating the strength
of inference from studies on the health effects of
ozone.
Study Population/Test Model
Controlled Human Exposure:
In general, the subjects recruited into study groups should be similarly matched for age, sex, race, anthropometric
properties, and health status. In studies evaluating effects of specific subject characteristics (e.g., disease, genetic
polymorphism, etc.), appropriately matched healthy controls are preferred. Relevant characteristics and health status
should be reported for each experimental group. Criteria for including and excluding subjects should be clearly
indicated. For the examination of populations with an underlying health condition (e.g., asthma), independent, clinical
assessment of the health condition is ideal, but self-report of physician diagnosis generally is considered to be reliable
for respiratory and cardiovascular disease outcomes.15 The loss or withdrawal of recruited subjects during the course
of a study should be reported. Specific rationale for excluding subject(s) from any portion of a protocol should be
explained.
Animal Toxicology:
Ideally, studies should report species, strain, substrain, genetic background, age, sex, and weight. Unless data
indicate otherwise, all animal species and strains are considered appropriate for evaluating effects of ozone exposure.
It is preferred that the authors test for effects in both sexes and multiple lifestages and report the result for each group
separately. All animals used in a study should be accounted for, and rationale for exclusion of animals or data should
be specified.
Epidemiology:
There is greater confidence in results for study populations that are recruited from and representative of the target
population. Studies that have high participation, have low drop-out over time, and are not dependent on exposure or
health status are considered to have low potential for selection bias. Clearly specified criteria for including and
excluding subjects can aid assessment of selection bias. For populations with an underlying health condition,
independent, clinical assessment of the health condition is valuable, but self-report of physician diagnosis generally is
considered to be reliable for respiratory and cardiovascular diseases.15 Comparisons of groups with and without an
underlying health condition are more informative if groups are from the same source population. Selection bias can
influence results in either direction or may not affect the validity of results but rather reduce the generalizability of
findings to the target population.
Pollutant
Controlled Human Exposure:
The focus is on studies testing ozone exposure.
Animal Toxicology:
The focus is on studies testing ozone exposure.
Epidemiology:
The focus is on studies testing ozone exposure.
4-147
-------
Table Annex 4-1 (Continued): Scientific considerations for evaluating the strength
of inference from studies on the health effects of
ozone.
Exposure Assessment or Assignment
Controlled Human Exposure:
For this assessment, the focus is on studies that use ozone concentrations <0.4 ppm. Studies that use higher
exposure concentrations may provide information relevant to biological plausibility, dosimetry, or inter-species
variation. Studies should have well-characterized pollutant concentration, temperature, and relative humidity and/or
have measures in place to adequately control the exposure conditions. Preference is given to balanced crossover or
parallel design studies which include control exposures (e.g., to clean filtered air). Study subjects should be randomly
exposed without knowledge of the exposure condition. Method of exposure (e.g., chamber, facemask, etc.) should be
specified and activity level of subjects during exposures should be well characterized.
Animal Toxicology:
For this assessment, the focus is on studies that use ozone concentrations <2 ppm. Studies that use higher exposure
concentrations may provide information relevant to biological plausibility, dosimetry, or inter-species variation. Studies
should characterize pollutant concentration, temperature, and relative humidity and/or have measures in place to
adequately control the exposure conditions. The focus is on inhalation exposure. Noninhalation exposure experiments
(i.e., intra-tracheal instillation [IT]) are informative for size fractions that cannot penetrate the airway of a study animal
and may provide information relevant to biological plausibility and dosimetry. In vitro studies may be included if they
provide mechanistic insight or examine similar effects as in vivo studies but are generally not included. All studies
should include exposure control groups (e.g., clean filtered air).
Epidemiology:
Of primary relevance are relationships of health effects with the ambient component of ozone exposure. However,
information about ambient exposure rarely is available for individual subjects; most often, inference is based on
ambient concentrations. Studies that compare exposure assessment methods are considered to be particularly
informative. Inference is stronger when the duration or lag of the exposure metric corresponds with the time course for
physiological changes in the outcome (e.g., up to a few days for symptoms) or latency of disease (e.g., several years
for cancer).
Ambient ozone concentration tends to have low spatial heterogeneity at the urban scale, except near roads where
ozone concentration is lower because ozone reacts with emitted nitric oxide. For studies involving individuals with
near-road or on-road exposures to ozone in which ambient ozone concentrations are more spatially heterogeneous
and relationships between personal exposures and ambient concentrations are potentially more variable, validated
methods that capture the extent of variability for the epidemiologic study design (temporal vs. spatial contrasts) and
location carry greater weight.
Fixed-site measurements, whether averaged across multiple monitors or assigned from the nearest or single available
monitor, typically have smaller biases and smaller reductions in precision compared with spatially heterogeneous air
pollutants. Concentrations reported from fixed-site measurements can be informative if correlated with personal
exposures, closely located to study subjects, highly correlated across monitors within a location, or combined with
time-activity information.
Atmospheric models may be used for exposure assessment in place of or to supplement ozone measurements in
epidemiologic analyses. For example, grid-scale models (e.g., CMAQ) that represent ozone exposure over relatively
large spatial scales (e.g., typically greater than 4- * 4-km grid size) often do provide adequate spatial resolution to
capture acute ozone peaks that influence short-term health outcomes. Uncertainty in exposure predictions from these
models is largely influenced by model formulations and the quality of model input data pertaining to precursor
emissions or meteorology, which tends to vary on a study-by-study basis.
In studies of short-term exposure, temporal variability of the exposure metric is of primary interest. For long-term
exposures, models that capture within-community spatial variation in individual exposure may be given more weight
for spatially variable ambient ozone. Given the low spatial variability of ozone at the urban scale, exposure
measurement error typically causes health effect estimates to be underestimated for studies of either short-term or
long-term exposure. Biases and decreases in the precision of the association (i.e., wider 95% CIs) tend to be small.
Even when spatial variability is higher near roads, the reduction in ozone exposure would cause the exposure to be
overestimated at a monitor distant from the road or when averaged across a model grid cell, so that health effects
would likely be underestimated.
4-148
-------
Table Annex 4-1 (Continued): Scientific considerations for evaluating the strength
of inference from studies on the health effects of
ozone.
Outcome Assessment/Evaluation
Controlled Human Exposure:
Endpoints should be assessed in the same manner for control and exposure groups (e.g., time after exposure,
methods, endpoint evaluator) using valid, reliable methods. Blinding of endpoint evaluators is ideal, especially for
qualitative endpoints (e.g., histopathology). For each experiment and each experimental group, including controls,
precise details of all procedures carried out should be provided including how, when, and where. Time of the endpoint
evaluations is a key consideration that will vary depending on endpoint evaluated. Endpoints should be assessed at
time points that are appropriate for the research questions.
Animal Toxicology:
Endpoints should be assessed in the same manner for control and exposure groups (e.g., time after exposure,
methods, endpoint evaluator) using valid, reliable methods. Blinding of endpoint evaluators is ideal, especially for
qualitative endpoints (e.g., histopathology). For each experiment and each experimental group, including controls,
precise details of all procedures carried out should be provided including how, when, and where. Time of the endpoint
evaluations is a key consideration that will vary depending on endpoint evaluated. Endpoints should be assessed at
time points that are appropriate for the research questions.
Epidemiology:
Inference is stronger when outcomes are assessed or reported without knowledge of exposure status. Knowledge of
exposure status could produce artefactual associations. Confidence is greater when outcomes assessed by interview,
self-report, clinical examination, or analysis of biological indicators are defined by consistent criteria and collected by
validated, reliable methods. Independent, clinical assessment is valuable for outcomes such as lung function or
incidence of disease, but report of physician diagnosis has shown good reliability.15 When examining short-term
exposures, evaluation of the evidence focuses on specific lags based on the evidence presented in individual studies.
Specifically, the following hierarchy is used in the process of selecting results from individual studies to assess in the
context of results across all studies for a specific health effect or outcome:
• Distributed lag models;
• Multiple days (e.g., 0-2) are averaged;
• Effect estimates are presented for lag days selected a priori by the study authors; or
• If a study focuses on only a series of individual lag days, expert judgment is applied to select the appropriate
result to focus on considering the time course for physiologic changes for the health effect or outcome being
evaluated.
When health effects of long-term exposure are assessed by acute events such as symptoms or hospital admissions,
inference is strengthened when results are adjusted for short-term exposure. Validated questionnaires for subjective
outcomes such as symptoms are regarded to be reliable,0 particularly when collected frequently and not subject to
long recall. For biological samples, the stability of the compound of interest and the sensitivity and precision of the
analytical method is considered. If not based on knowledge of exposure status, errors in outcome assessment tend to
bias results toward the null.
Potential Copollutant Confounding
Controlled Human Exposure:
Exposure should be well characterized to evaluate independent effects of ozone.
Animal Toxicology:
Exposure should be well characterized to evaluate independent effects of ozone.
4-149
-------
Table Annex 4-1 (Continued): Scientific considerations for evaluating the strength
of inference from studies on the health effects of
ozone.
Epidemiology:
Not accounting for potential copollutant confounding can produce artifactual associations; thus, studies that examine
copollutant confounding carry greater weight. The predominant method is copollutant modeling (i.e., two-pollutant
models), which is especially informative when correlations are not high. However, when correlations are high (r> 0.7),
such as those often encountered for UFP and other traffic-related copollutants, copollutant modeling is less
informative. Although the use of single-pollutant models to examine the association between ozone and a health effect
or outcome are informative, ideally studies should also include copollutant analyses. Copollutant confounding is
evaluated on an individual study basis considering the extent of correlations observed between the copollutant and
ozone, and relationships observed with ozone and health effects in copollutant models.
Other Potential Confounding Factors'1
Controlled Human Exposure:
Preference is given to studies using experimental and control groups that are matched for individual level
characteristics (e.g., race/ethnicity, sex, body weight, smoking history, age) and time varying factors (e.g., seasonal
and diurnal patterns).
Animal Toxicology:
Preference is given to studies using experimental and control groups that are matched for individual level
characteristics (e.g., strain, sex, body weight, litter size, food and water consumption) and time varying factors
(e.g., seasonal and diurnal patterns).
Epidemiology:
Factors are considered to be potential confounders if demonstrated in the scientific literature to be related to health
effects and correlated with ozone. Not accounting for confounders can produce artifactual associations; thus, studies
that statistically adjust for multiple factors or control for them in the study design are emphasized. Less weight is
placed on studies that adjust for factors that mediate the relationship between ozone and health effects, which can
bias results toward the null. Confounders vary according to study design, exposure duration, and health effect and
may include, but are not limited to the following:
• Short-term exposure studies: Meteorology, day of week, season, medication use, allergen exposure, and
long-term temporal trends.
• Long-term exposure studies: Socioeconomic status, race, age, medication use, smoking status, stress,
noise, and occupational exposures.
Statistical Methodology
Controlled Human Exposure:
Statistical methods should be clearly described and appropriate for the study design and research question
(e.g., correction for multiple comparisons). Generally, statistical significance is used to evaluate the findings of
controlled human exposure studies. However, consistent trends are also informative. Detection of statistical
significance is influenced by a variety of factors including, but not limited to, the size of the study, exposure and
outcome measurement error, and statistical model specifications. Sample size is not a criterion for exclusion; ideally,
the sample size should provide adequate power to detect hypothesized effects (e.g., sample sizes less than three are
considered less informative). Because statistical tests have limitations, consideration is given to both trends in data
and reproducibility of results.
4-150
-------
Table Annex 4-1 (Continued): Scientific considerations for evaluating the strength
of inference from studies on the health effects of
ozone.
Animal Toxicology:
Statistical methods should be clearly described and appropriate for the study design and research question
(e.g., correction for multiple comparisons). Generally, statistical significance is used to evaluate the findings of animal
toxicological studies. However, consistent trends are also informative. Detection of statistical significance is influenced
by a variety of factors including, but not limited to, the size of the study, exposure and outcome measurement error,
and statistical model specifications. Sample size is not a criterion for exclusion; ideally, the sample size should provide
adequate power to detect hypothesized effects (e.g., sample sizes less than three are considered less informative).
Because statistical tests have limitations, consideration is given to both trends in data and reproducibility of results.
Epidemiology:
Multivariable regression models that include potential confounding factors are emphasized. However, multipollutant
models (more than two pollutants) are considered to produce too much uncertainty due to copollutant collinearity to be
informative. Models with interaction terms aid in the evaluation of potential confounding as well as effect modification.
Sensitivity analyses with alternate specifications for potential confounding inform the stability of findings and aid in
judgments of the strength of inference from results. In the case of multiple comparisons, consistency in the pattern of
association can increase confidence that associations were not found by chance alone. Statistical methods that are
appropriate for the power of the study carry greater weight. For example, categorical analyses with small sample sizes
can be prone to bias results toward or away from the null. Statistical tests such as f-tests and chi-squared tests are not
considered sensitive enough for adequate inferences regarding ozone-health effect associations. For all methods, the
effect estimate and precision of the estimate (i.e., width of 95% CI) are important considerations rather than statistical
significance.
aU.S. EPA (2008V
"Muraia etal. (2014V Weakley et al. (2013V Yang et al. (2011V Heckbert et al. (2004V Barr et al. (2002V Muhaiarine et al. (1997V
Toren et al. (1993V
cBurnev et al. (1989V
4-151
-------
4.4 References
Argacha. JF; Collart. P; Wauters. A; Kavaert. P; Lochy. S; Schoors. D; Sonck. J; de Vos. T; Forton.
M; Brasseur. O; Beaulove. C; Gevaert. S; Evrard. P; Coppieters. Y; Sinnaeve. P; Claevs. MJ.
(2016). Air pollution and ST-elevation myocardial infarction: A case-crossover study of the
Belgian STEMI registry 2009-2013. Int J Cardiol 223: 300-305.
http://dx.doi.Org/10.1016/i.iicard.2016.07.191
Ariomandi. M: Wong. H: Donde. A: Frelinger. J: Dalton. S: Ching. W; Power. K: Balmes. JR. (2015).
Exposure to medium and high ambient levels of ozone causes adverse systemic inflammatory and
cardiac autonomic effects. Am J Physiol Heart Circ Physiol 308: H1499-1509.
http://dx.doi.org/10.1152/aipheart.00849.2014
Atkinson. RW; Carey. IM: Kent. AJ: van Staa. TP: Anderson. HR: Cook. DG. (2013). Long-term
exposure to outdoor air pollution and incidence of cardiovascular diseases. Epidemiology 24: 44-
53. http://dx.doi.org/10.1097/EDE.0b013e318276ccb8
Barath. S: Langrish. JP: Lundback. M: Bosson. JA: Goudie. C: Newbv. DE: Sandstrom. T: Mills. NL:
Blomberg. A. (2013). Short-term exposure to ozone does not impair vascular function or affect
heart rate variability in healthy young men. Toxicol Sci 135: 292-299.
http://dx.doi.org/10.1093/toxsci/kftl57
Bard. D: Kihal. W: Schillinger. C: Fermanian. C: Segala. C: Glorion. S: Arveiler. D: Weber. C.
(2014). Traffic-related air pollution and the onset of myocardial infarction: disclosing benzene as a
trigger? A small-area case-crossover study. PLoS ONE 9: e 100307.
http://dx.doi.org/10.1371/iournal.pone.01003Q7
Barr. RG: Herbstman. J: Speizer. FE; Camargo. CA. Jr. (2002). Validation of self-reported chronic
obstructive pulmonary disease in a cohort study of nurses. Am J Epidemiol 155: 965-971.
http://dx.doi.org/10.1093/aie/155.10.965
Bartell. SM: Longhurst. J: Tioa. T: Sioutas. C: Delfino. RJ. (2013). Particulate air pollution,
ambulatory heart rate variability, and cardiac arrhythmia in retirement community residents with
coronary artery disease. Environ Health Perspect 121: 1135-1141.
http://dx.doi.org/10.1289/ehp.1205914
Bedada. GB; Smith. CJ: Tyrrell. PJ: Hirst AA; Agius. R. (2012). Short-term effects of ambient
particulates and gaseous pollutants on the incidence of transient ischaemic attack and minor stroke:
a case-crossover study. Environ Health 11: 77. http://dx.doi.org/10.1186/1476-069X-11-77
Bentaveb. M; Wagner. V: Stempfelet. M; Zins. M; Goldberg. M; Pascal. M; Larrieu. S: Beaudeau. P;
Cassadou. S: Eilstein. D; Filleul. L; Le Tertre. A: Medina. S: Pascal. L; Prouvost. H; Quenel. P;
Zeghnoun. A: Lefranc. A. (2015). Association between long-term exposure to air pollution and
mortality in France: A 25-year follow-up study. Environ Int 85: 5-14.
http://dx.doi.Org/10.1016/i.envint.2015.08.006
Bhaskaran. K: Haiat. S: Armstrong. B: Haines. A: Herrett. E: Wilkinson. P; Smeeth. L. (2011). The
effects of hourly differences in air pollution on the risk of myocardial infarction: Case crossover
analysis of the MINAP database. BMJ 343: d5531. http://dx.doi.org/10.1136/bmi.d5531
Bigger. JT. Jr; Fleiss. JL; Steinman. RC: Rolnitzkv. LM; Kleiger. RE: Rottman. JN. (1992). Frequency
domain measures of heart period variability and mortality after myocardial infarction. Circulation
85: 164-171.
4-152
-------
Biller. H; Holz. O; Windt. H; Koch. W; Miiller. M; Jorres. RA: Krug. N; Hohlfeld. JM. (2011). Breath
profiles by electronic nose correlate with systemic markers but not ozone response. Respir Med
105: 1352-1363. http://dx.doi.Org/10.1016/i.rmed.2011.03.002
Billman. GE. (2013). The LF/HF ratio does not accurately measure cardiac sympatho-vagal balance.
Front Physiol 4: 26. http://dx.doi.org/10.3389/fbhvs.2013.00026
Bind. MA: Baccarelli. A: Zanobetti. A: Tarantini. L: Suh. H: Yokonas. P; Schwartz. J. (2012). Air
pollution and markers of coagulation, inflammation, and endothelial function: Associations and
epigene-environment interactions in an elderly cohort. Epidemiology 23: 332-340.
http://dx.doi.org/10.1097/EDE.0b013e31824523fD
Bosson. JA: Blomberg. A: Stenfors. N: Helledav. R: Kelly. FJ: Behndig. AF: Mudwav. I. (2013).
Peripheral blood neutrophilia as a biomarker of ozone-induced pulmonary inflammation. PLoS
ONE 8: e81816. http://dx.doi.org/10.1371/iournal.pone.0Q81816
Breton. CV: Wang. X: Mack. WJ: Berhane. K: Lopez. M: Islam. TS: Feng. M: Lurmann. F:
McConnell. R; Hodis. HN: Kiinzli. N: Avol. E. (2012). Childhood air pollutant exposure and
carotid artery intima-media thickness in young adults. Circulation 126: 1614-1620.
http ://dx.doi .org/10.1161/CIRCULATIONAHA. 112.096164
Breton. CV: Yao. J: Millstein. J: Gao. L: Siegmund. KD: Mack. W: Whitfield-Maxwell. L: Lurmann.
F; Hodis. H; Avol. E; Gilliland. FD. (2016). Prenatal Air Pollution Exposures, DNA Methyl
Transferase Genotypes, and Associations with Newborn LINE1 and Alu Methylation and
Childhood Blood Pressure and Carotid Intima-Media Thickness in the Children's Health Study.
Environ Health Perspect 124: 1905-1912. http://dx.doi.org/10.1289/EHP181
Brook. RD: Kousha. T. (2015). Air pollution and emergency department visits for hypertension in
Edmonton and Calgary, Canada: A case-crossover study. Am J Hypertens 28: 1121-1126.
http://dx.doi.org/10.1093/aih/hpu302
Brook. RD: Urch. B; Dvonch. JT; Bard. RL: Speck. M; Keeler. G: Morishita. M; Marsik. FJ: Kamal.
AS: Kaciroti. N: Harkema. J: Corey. P: Silverman. F: Gold. PR: Wellenius. G: Mittleman. MA:
Raiagopalan. S: Brook. JR. (2009). Insights into the mechanisms and mediators of the effects of air
pollution exposure on blood pressure and vascular function in healthy humans. Hypertension 54:
659-667. http://dx.doi.org/10.1161 /hypertensionaha. 109.130237
Burnev. PG: Laitinen. LA: Perdrizet. S: Huckauf. H: Tattersfield. AE: Chinn. S: Poisson. N: Heeren.
A: Britton. JR; Jones. T. (1989). Validity and repeatability of the IUATLD (1984) Bronchial
Symptoms Questionnaire: an international comparison. Eur Respir J 2: 940-945.
Butland. BK: Atkinson. RW: Crichton. S: Barratt. B: Beevers. S: Spiridou. A: Hoang. U: Kelly. FJ:
Wolfe. CD. (2017). Air pollution and the incidence of ischaemic and haemorrhagic stroke in the
South London Stroke Register: a case-cross-over analysis. J Epidemiol Community Health 71: 707-
712. http://dx.doi.org/10.1136/iech-2016-208Q25
Butland. BK: Atkinson. RW: Miloievic. A: Heal. MR: Dohertv. RM: Armstrong. BG: Mackenzie. TA:
Vieno. M; Lin. C: Wilkinson. P. (2016). Myocardial infarction, ST-elevation and non-ST-elevation
myocardial infarction and modelled daily pollution concentrations: a case-crossover analysis of
MINAP data. 3: e000429. http://dx.doi.Org/10.l 136/openhrt-2016-000429
Cakmak. S: Dales. R; Leech. J: Liu. L. (2011). The influence of air pollution on cardiovascular and
pulmonary function and exercise capacity: Canadian Health Measures Survey (CHMS). Environ
Res 111: 1309-1312. http://dx.doi.Org/10.1016/i.envres.2011.09.016
4-153
-------
Cakmak. S; Hebbern. C; Pinault. L; Lavignc. E; Vanos. J; Crouse. PL; Tjepkema. M. (2018).
Associations between long-term PM2.5 and ozone exposure and mortality in the Canadian Census
Health and Environment Cohort (CANCHEC), by spatial synoptic classification zone. Environ Int
111: 200-211. http://dx.doi.Org/10.1016/i.envint.2017.ll.030
Cakmak. S; Hebbern. C; Vanos. J; Crouse. PL; Burnett. R. (2016). Ozone exposure and
cardiovascular-related mortality in the Canadian Census Health and Environment Cohort
(CANCHEC) by spatial synoptic classification zone. Environ Pollut 214: 589-599.
http://dx.doi.Org/10.1016/i.envpol.2016.04.067
Cakmak. S; Kauri. L: Shutt. R; Liu. L: Green. MS; Mulholland. M: Stieb. P; Pales. R. (2014). The
association between ambient air quality and cardiac rate and rhythm in ambulatory subjects.
Environ Int 73: 365-371. http://dx.doi.Org/10.1016/i.envint.2014.08.015
Carey. IM; Atkinson. RW; Kent. AJ; van Staa. T; Cook. PG; Anderson. HR. (2013). Mortality
associations with long-term exposure to outdoor air pollution in a national English cohort. Am J
Respir Crit Care Med 187: 1226-1233. http://dx.doi.Org/10.l 164/rccm.201210-1758QC
Carll. AP; Farrai. AK; Roberts. AM. (2018). Role of the autonomic nervous system in cardiovascular
toxicity. In C McQueen (Ed.), Comprehensive toxicology (3rd ed., pp. 61-114). Amsterdam,
Netherlands: Elsevier. http://dx.doi.org/10.1016/B978-0-12-801238-3.64259-9
Cestonaro. LV; Marcolan. AM; Rossato-Grando. LG; Anzolin. AP; Goethel. G; Vilani. A; Garcia. SC;
Bertol. CP. (2017). Ozone generated by air purifier in low concentrations: friend or foe? Environ
Sci Pollut Res Int 24: 22673-22678. http://dx.doi.org/10.1007/sll356-017-9887-3
Chen. L; Villeneuve. PJ; Rowe. BH; Liu. L; Stieb. PM. (2014). The Air Quality Health Index as a
predictor of emergency department visits for ischemic stroke in Edmonton, Canada. J Expo Sci
Environ Epidemiol 24: 358-364. http://dx.doi.org/10.1038/ies.2013.82
Choi. M; Curriero. FC; Johantgen. M; Mills. ME; Sattler. B; Lipscomb. J. (2011). Association
between ozone and emergency department visits: An ecological study. Int J Environ Health Res 21:
201-221. http://dx.doi.org/10.1080/09603123.201Q.533366
Chuang. GC; Yang. Z; Westbrook. PG; Pompilius. M; Ballinger. CA; White. RC; Krzvwanski. PM;
Postlethwait. EM; Ballinger. SW. (2009). Pulmonary ozone exposure induces vascular dysfunction,
mitochondrial damage, and atherogenesis. Am J Physiol Lung Cell Mol Physiol 297: L209-L216.
http://dx.doi.org/10.1152/aiplung.00102.20Q9
Chuang. KJ; Chan. CC; Su. TC; Lee. CT; Tang. CS. (2007). The effect of urban air pollution on
inflammation, oxidative stress, coagulation, and autonomic dysfunction in young adults. Am J
Respir Crit Care Med 176: 370-376. http://dx.doi.Org/10.l 164/rccm.200611-1627QC
Chuang. KJ; Yan. YH; Chiu. SY; Cheng. TJ. (2011). Long-term air pollution exposure and risk factors
for cardiovascular diseases among the elderly in Taiwan. Occup Environ Med 68: 64-68.
http://dx.doi.Org/10.l 136/oem.2009.052704
Claevs. MJ; Coenen. S; Colpaert. C; Bilcke. J; Beutels. P; Wouters. K; Legrand. V; Van Pamme. P;
Vrints. C. (2015). Environmental triggers of acute myocardial infarction: Results of a nationwide
multiple-factorial population study. Acta Cardiol 70: 693-701.
http://dx.doi.org/10.2143/AC.70.6.3120182
Cole-Hunter. T; de Nazelle. A; Ponaire-Gonzalez. P; Kubesch. N; Carrasco-Turigas. G; Matt. F;
Foraster. M; Martinez. T; Ambros. A; Cirach. M; Martinez. P; Belmonte. J; Nieuwenhuiisen. M.
(2018). Estimated effects of air pollution and space-time-activity on cardiopulmonary outcomes in
healthy adults: A repeated measures study. Environ Int 111: 247-259.
http://dx.doi.Org/10.1016/i.envint.2017.l 1.024
4-154
-------
Collart. P; Dramaix. M; Leveque. A; Coppieters. Y. (2017). Short-term effects of air pollution on
hospitalization for acute myocardial infarction: Age effect on lag pattern. Int J Environ Health Res
27: 68-81. http://dx.doi.org/10.1080/09603123.2Q16.1268678
Coogan. PF; White. LF; Yu. J; Brook. RD: Burnett. RT; Marshall. JD; Bethea. TN; Rosenberg. L;
Jerrett. M. (2017). Long-term exposure to N02 and ozone and hypertension incidence in the black
women's health study. Am J Hypertens 30: 367-372. http://dx.doi.org/10.1093/aih/hpwl68
Corea. F: Silvestrelli. G; Baccarelli. A: Giua. A: Previdi. P; Siliprandi. G; Murgia. N. (2012). Airborne
pollutants and lacunar stroke: a case cross-over analysis on stroke unit admissions. Neurology
International 4: ell. http://dx.doi.org/10.408l/ni.2012.e 11
Crouse. PL: Peters. PA: Hvstad. P; Brook. JR: van Donkelaar. A: Martin. RV: Villeneuve. PJ; Jerrett.
M; Goldberg. MS: Pope. CA; Brauer. M; Brook. RD: Robichaud. A: Menard. R; Burnett. RT.
(2015). Ambient PM 2.5, O 3, and NO 2 exposures and associations with mortality over 16 years of
follow-up in the Canadian Census Health and Environment Cohort (CanCHEC). Environ Health
Perspect 123: 1180-1186. http://dx.doi.org/10.1289/ehp.1409276
Dales. RE: Cakmak. S. (2016). Does mental health status influence susceptibility to the physiologic
effects of air pollution? A population based study of Canadian children. PLoS ONE 11: eO 168931.
http://dx.doi.org/10.1371/iournal.pone.0168931
de Miguel-Diez. J: Jimenez-Garcia. R; de Andres. A: Hernandez-Barrera. V: Carrasco-Garrido. P;
Monreal. M: Jimenez. D: Jara-Palomares. L: Alvaro-Meca. A. (2016). Analysis of environmental
risk factors for pulmonary embolism: A case-crossover study (2001-2013). Eur J Intern Med 31:
55-61. http://dx.doi.Org/10.1016/i.eiim.2016.03.001
Delfino. RJ: Gillen. PL: Tioa. T; Staimer. N: Polidori. A: Arhami. M; Sioutas. C: Longhurst. J.
(2011). Electrocardiographic ST-segment depression and exposure to trafficrelated aerosols in
elderly subjects with coronary artery disease. Environ Health Perspect 119: 196-202.
http://dx.doi.org/10.1289/ehp.1002372
Dennekamp. M: Akram. M: Abramson. MJ: Tonkin. A: Sim. MR: Fridman. M: Erbas. B. (2010).
Outdoor air pollution as a trigger for out-of-hospital cardiac arrests. Epidemiology 21: 494-500.
http://dx.doi.org/10.lQ97/EDE.0b013e3181e093db
Devlin. RB: Duncan. KE; Jardim. M; Schmitt. MT; Rappold. AG: Diaz-Sanchez. D. (2012).
Controlled exposure of healthy young volunteers to ozone causes cardiovascular effects.
Circulation 126: 104-111. http://dx.doi.Org/10.l 161/CIRCULATIONAHA.l 12.094359
Dong. G: Qian. Z; Wang. J: Chen. W: Ma. W: Trevathan. E; Xaverius. PK; DeClue. R; Wiese. A:
Langston. M: Liu. MM: Wang. D: Ren. W. (2013a). Associations between ambient air pollution
and prevalence of stroke and cardiovascular diseases in 33 Chinese communities. Atmos Environ
77: 968-973. http://dx.doi.Org/10.1016/i.atmosenv.2013.06.034
Dong. GH: Qian. ZM: Trevathan. E: Zeng. XW: Vaughn. MG: Wang. J: Zhao. Y: Liu. YO: Ren. WH:
Oin. XD. (2014). Air pollution associated hypertension and increased blood pressure may be
reduced by breastfeeding in Chinese children: the Seven Northeastern Cities Chinese Children's
Study. Int J Cardiol 176: 956-961. http://dx.doi.Org/10.1016/i.iicard.2014.08.099
Dong. GH: Qian. ZM: Xaverius. PK: Trevathan. E: Maalouf. S: Parker. J: Yang. L: Liu. MM: Wang.
D: Ren. WH: Ma. W: Wang. J: Zelicoff. A: Fu. Q: Simckes. M. (2013b). Association between
long-term air pollution and increased blood pressure and hypertension in China. Hypertension 61:
578-584. http://dx.doi.org/10.1161 /HYPERTENSIONAHA. 111.00003
4-155
-------
Dong. GH; Wang. J; Zeng. XW; Chen. L; Oin. XD; Zhou. Y; Li. M; Yang. M; Zhao. Y; Ren. WH:
Hu. OS. (2015). Interactions between air pollution and obesity on blood pressure and hypertension
in Chinese children. Epidemiology 26: 740-747.
http://dx.doi.org/10.1097/EDE.000000000000Q336
Ensor. KB; Raun. LH; Persse. D. (2013). A case-crossover analysis of out-of-hospital cardiac arrest
and air pollution. Circulation 127: 1192-1199.
http://dx.doi.org/10.1161/CIRCULATIONAHA.113.000Q27
Evans. KA: Hopke. PK; Utell. MJ; Kane. C; Thurston. SW; Ling. FS; Chalupa. D; Rich. DQ. (2016).
Triggering of ST-elevation myocardial infarction by ambient wood smoke and other particulate and
gaseous pollutants. J Expo Sci Environ Epidemiol 27: 198-206.
http://dx.doi.org/10.1038/ies.2016.15
Fakhri. AA; Ilic. LM; Wellenius. GA; Urch. B; Silverman. F; Gold. PR; Mittleman. MA. (2009).
Autonomic effects of controlled fine particulate exposure in young healthy adults: Effect
modification by ozone. Environ Health Perspect 117: 1287-1292.
http://dx.doi.org/10.1289/ehp.0900541
Farrai. AK: Hazari. MS; Winsett. DW; Kulukulualani. A: Carll. AP: Havkal-Coates. N; Lamb. CM:
Lappi. E; Terrell. D; Cascio. WE; Costa. PL. (2012). Overt and latent cardiac effects of ozone
inhalation in rats: evidence for autonomic modulation and increased myocardial vulnerability.
Environ Health Perspect 120: 348-354. http://dx.doi.org/10.1289/ehp.1104244
Farrai. AK; Malik. F; Havkal-Coates. N; Walsh. L; Winsett. D; Terrell. D; Thompson. LC; Cascio.
WE; Hazari. MS. (2016). Morning N02 exposure sensitizes hypertensive rats to the cardiovascular
effects of same day 03 exposure in the afternoon. Inhal Toxicol 28: 170-179.
http://dx.doi.org/10.3109/08958378.2016.1148Q88
Fauchier. L; Babutv. D; Melin. A; Bonnet. P: Cosnav. P; Paul Fauchier. J. (2004). Heart rate
variability in severe right or left heart failure: the role of pulmonary hypertension and resistances.
Eur J Heart Fail 6: 181-185. http://dx.doi.Org/10.1016/i.eiheart.2003.09.007
Finnbiornsdottir. RG; Zoega. H; Olafsson. O; Thorsteinsson. T; Rafnsson. V. (2013). Association of
air pollution and use of glyceryl trinitrate against angina pectoris: a population-based case-
crossover study. Environ Health 12: 38. http://dx.doi.Org/10.l 186/1476-069X-12-38
Frampton. MW; Balmes. JR; Bromberg. PA; Stark. P; Ariomandi. M; Hazucha. MJ; Rich. DO;
Hollenbeck-Pringle. D; Dagincourt. N; Alexis. N; Ganz. P; Zareba. W; Costantini. MG. (2017).
Multicenter Ozone Study in oldEr Subjects (MOSES: Part 1. Effects of exposure to low
concentrations of ozone on respiratory and cardiovascular outcomes) [HEI], (Research Report 192,
Pt 1). Boston, MA: Health Effects Institute.
Frampton. MW; Pietropaoli. A; Dentler. M; Chalupa. D; Little. EL; Stewart. J; Frasier. L; Oakes. D;
Wiltshire. J; Vora. R; Utell. MJ. (2015). Cardiovascular effects of ozone in healthy subjects with
and without deletion of glutathione-S-transferase Ml. Inhal Toxicol 27: 113-119.
http://dx.doi.org/10.3109/08958378.2014.996272
Francis. M; Groves. AM; Sun. R; Cervelli. JA; Choi. H; Laskin. JD; Laskin. PL. (2017). Editor's
highlight: CCR2 regulates inflammatory cell accumulation in the lung and tissue injury following
ozone exposure. Toxicol Sci 155: 474-484. http://dx.doi.org/10.1093/toxsci/kfw226
Gandhi. SK; Rich. DQ; Ohman-Strickland. PA; Kipen. HM; Gow. A. (2014). Plasma nitrite is an
indicator of acute changes in ambient air pollutant concentrations. Inhal Toxicol 26: 426-434.
http://dx.doi.org/10.3109/08958378.2Q14.913216
4-156
-------
Gong. H. Jr; Wong. R; Sarma. RJ: Linn. WS; Sullivan. ED; Shamoo. DA; Anderson. KR; Prasad. SB.
(1998). Cardiovascular effects of ozone exposure in human volunteers. Am J Respir Crit Care Med
158: 538-546. http://dx.doi.Org/10.1164/airccm.158.2.9709034
Gordon. CJ; Jarema. KA; Lehmann. J. R.; Ledbetter. AD; Schladweiler. MC; Schmid. JE; Ward. WO;
Kodavanti. UP; Nvska. A; Macphail. RC. (2013). Susceptibility of adult and senescent Brown
Norway rats to repeated ozone exposure: an assessment of behavior, serum biochemistry and
cardiopulmonary function. Inhal Toxicol 25: 141-159.
http://dx.doi.org/10.3109/08958378.2013.764946
Gordon. CJ; Johnstone. AF; Avdin. C; Phillips. PM; Macphail. RC; Kodavanti. UP; Ledbetter. AD;
Jarema. KA. (2014). Episodic ozone exposure in adult and senescent Brown Norway rats: acute
and delayed effect on heart rate, core temperature and motor activity. Inhal Toxicol 26: 380-390.
http://dx.doi.org/10.3109/08958378.2014.9Q5659
Green. R; Broadwin. R; Malig. B; Basu. R; Gold. EB; Oi. L; Sternfeld. B; Bromberger. JT; Greendale.
GA; Kravitz. HM; Tomev. K; Matthews. K; Derby. C; Jackson. EA; Green. R; Ostro. B. (2015).
Long-and short-term exposure to air pollution and inflammatory/hemostatic markers in midlife
women. Epidemiology 27: 211-220. http://dx.doi.org/10.1097/EDE.000000000000Q421
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/iech.2009.0871Q6
Halvorsen. B; Otterdal. K; Dahl. TB; Skielland. M; Gullestad. L; Oie. E; Aukrust. P. (2008).
Atherosclerotic plaque stability - what determines the fate of a plaque? Prog Cardiovasc Dis 51:
183-194. http://dx.doi.Org/10.1016/i.pcad.2008.09.001
Hanna.. AF; Yeatts. KB; Xiu. A; Zhu. Z; Smith. RL; Davis. NN; Talgo. KD; Arora. G; Robinson. PJ;
Meng. O; Pinto. JP. (2011). Associations between ozone and morbidity using the spatial synoptic
classification system. Environ Health 10: 49. http://dx.doi.Org/10.l 186/1476-069X-10-49
Hatch. GE; Crissman. K; Schmid. J; Richards. JE; Ward. WO; Schladweiler. MC; Ledbetter. AD;
Kodavanti. UP. (2015). Strain differences in antioxidants in rat models of cardiovascular disease
exposed to ozone. Inhal Toxicol 27: 54-62. http://dx.doi.org/10.3109/08958378.2014.95417Q
Heckbert. SR; Kooperberg. C; Safford. MM; Psatv. BM; Hsia. J; McTiernan. A; Gaziano. JM;
Frishman. WH; Curb. JD. (2004). Comparison of self-report, hospital discharge codes, and
adjudication of cardiovascular events in the Women's Health Initiative. Am J Epidemiol 160: 1152-
1158. http://dx.doi.org/10.1093/aie/kwh314
Henriquez. AR; Snow. SJ; Schladweiler. MC; Miller. CN; Dve. JA; Ledbetter. AD; Richards. JE;
Mauge-Lewis. K; Mcgee. MA; Kodavanti. UP. (2017). Adrenergic and glucocorticoid receptor
antagonists reduce ozone-induced lung injury and inflammation. Toxicol Appl Pharmacol 339:
161-171. http://dx.doi.Org/10.1016/i.taap.2017.12.006
Henrotin. JB; Besancenot. JP; Beiot. Y; Giroud. M. (2007). Short-term effects of ozone air pollution
on ischaemic stroke occurrence: A case-crossover analysis from a 10-year population-based study
in Dijon, France. Occup Environ Med 64: 439-445. http://dx.doi.Org/10.l 136/oem.2006.029306
Henrotin. JB; Zeller. M; Lorgis. L; Cottin. Y: Giroud. M; Beiot. Y. (2010). Evidence of the role of
short-term exposure to ozone on ischaemic cerebral and cardiac events: The Dijon Vascular Project
(DIVA). Heart 96: 1990-1996. http://dx.doi.org/10.1136/hrt.2010.200337
4-157
-------
Hoffmann. B; Luttmann-Gibson. H; Cohen. A; Zanobetti. A; de Souza. C; Foley. C; Suh. HH: Coull.
BA; Schwartz. J; Mittleman. M; Stone. P; Horton. E; Gold. DR. (2012). Opposing effects of
particle pollution, ozone, and ambient temperature on arterial blood pressure. Environ Health
Perspect 120: 241-246. http://dx.doi.Org/10.1289/ehp.l 103647
Hunova. I; Malv. M; Rezacova. J; Branis. M. (2013). Association between ambient ozone and health
outcomes in Prague. Int Arch Occup Environ Health 86: 89-97. http://dx.doi.org/10.1007/s00420-
012-0751-v
Hunova. I. va; Brabec. M; Malv. M; Knobova. V; Branis. M. (2017). Major heat waves of 2003 and
2006 and health outcomes in Prague. Air Qual Atmos Health 10: 183-194.
http://dx.doi.org/10.1007/sll869-016-Q419-v
Jerrett. M; Burnett. RT; Beckerman. BS; Turner. MC; Krewski. D; Thurston. G; Martin. RV; van
Donkelaar. A; Hughes. E; Shi. Y; Gapstur. SM; Thun. MJ; Pope. CA. III. (2013). Spatial analysis
of air pollution and mortality in California. Am J Respir Crit Care Med 188: 593-599.
http://dx.doi.Org/10.l 164/rccm.201303-0609QC
Jerrett. M; Burnett. RT; Pope. CA. Ill; 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
Kahle. JJ; Neas. LM; Devlin. RB; Case. MW; Schmitt. MT; Madden. MC; Diaz-Sanchez. D. (2015).
Interaction effects of temperature and ozone on lung function and markers of systemic
inflammation, coagulation, and fibrinolysis: a crossover study of healthy young volunteers.
Environ Health Perspect 123: 310-316. http://dx.doi.org/10.1289/ehp.1307986
Kalantzi. EG; Makris. D; Duquenne. MN; Kaklamani. S; Stapountzis. H; Gourgoulianis. KI. (2011).
Air pollutants and morbidity of cardiopulmonary diseases in a semi-urban Greek peninsula. Atmos
Environ 45: 7121-7126. http://dx.doi.Org/10.1016/i.atmosenv.2011.09.032
Karolv. ED; Li. Z; Dailev. LA; Hvseni. X; Huang. YCT. (2007). Up-regulation of tissue factor in
human pulmonary artery endothelial cells after ultrafine particle exposure. Environ Health Perspect
115: 535-540. http://dx.doi.org/10.1289/ehp.9556
Khder. Y; Briancon. S; Petermann. R; Quilliot. D; Stoltz. JF; Drouin. P; Zannad. F. (1998). Shear
stress abnormalities contribute to endothelial dysfunction in hypertension but not in type II
diabetes. J Hypertens 16: 1619-1625. http://dx.doi.org/10.1097/00004872-199816110-000Q8
Kilkenny. C; Browne. WJ; Cuthill. IC; Emerson. M; Altman. DG. (2010). Improving bioscience
research reporting: The ARRIVE guidelines for reporting animal research [Review]. PLoS Biol 8:
el000412. http://dx.doi.org/10.1371/iournal.pbio.100Q412
Kim. H; Kim. J; Kim. S; Kang. SH; Kim. HJ; Kim. H; Heo. J; Yi. SM; Kim. K; Youn. TJ; Chae. IH.
(2017). Cardiovascular effects of long-term exposure to air pollution: a population-based study
with 900845person-years of follow-up. J Am Heart Assoc 6.
http://dx.doi.org/10.1161/JAHA.117.00717Q
Klemm. RJ: Lipfert. FW; Wvzga. RE; Gust. C. (2004). Daily mortality and air pollution in Atlanta:
two years of data from ARIES. Inhal Toxicol 16 Suppl 1: 131-141.
http://dx.doi.org/10.1080/0895837049Q443213
Klemm. RJ; Mason. RM. Jr. (2000). Aerosol Research and Inhalation Epidemiological Study
(ARIES): air quality and daily mortality statistical modeling—interim results. J Air Waste Manag
Assoc 50: 1433-1439.
4-158
-------
Klemm. RJ: Thomas. EL; Wvzga. RE. (2011). The impact of frequency and duration of air quality
monitoring: Atlanta, GA, data modeling of air pollution and mortality. J Air Waste Manag Assoc
61: 1281-1291. http://dx.doi.org/10.108Q/10473289.2011.617648
Klimisch. HJ; Andreae. M; Tillmann. U. (1997). A systematic approach for evaluating the quality of
experimental toxicological and ecotoxicological data. Regul Toxicol Pharmacol 25: 1-5.
http://dx.doi.org/10.1006/rtph.1996.1076
Kotecha. D; New. G; Flather. MP: Eccleston. D: Pepper. J: Krum. H. (2012). Five-minute heart rate
variability can predict obstructive angiographic coronary disease. Heart 98: 395-401.
http://dx.doi.Org/10.l 136/heartinl-2011-300033
Kumarathasan. P: Blais. E: Saravanamuthu. A: Bielecki. A: Mukheriee. B: Biarnason. S; Guenette. J;
Goegan. P; Vincent R. (2015). Nitrative stress, oxidative stress and plasma endothelin levels after
inhalation of particulate matter and ozone. Part Fibre Toxicol 12: 28.
http://dx.doi.org/10.1186/sl2989-015-0103-7
Kurhanewicz. N; Mcintosh-Kastrinskv. R; Tong. H; Walsh. L; Farrai. A; Hazari. MS. (2014). Ozone
co-exposure modifies cardiac responses to fine and ultrafine ambient particulate matter in mice:
Concordance of electrocardiogram and mechanical responses. Part Fibre Toxicol 11: 54.
http://dx.doi.org/10.1186/sl2989-014-0054-4
Kusha. M; Masse. S; Farid. T; Urch. B; Silverman. FS; Brook. RD: Gold. PR; Mangat. I; Speck. M;
Nair. K: Poku. K: Mever. C; Mittleman. MA: Wellenius. GA: Nanthakumar. K. (2012). Controlled
exposure study of air pollution and t wave alternans in volunteers without cardiovascular disease.
Environ Health Perspect 120: 1157-1161. http://dx.doi.org/10.1289/ehp. 1104171
Lahiri. MK; Kannankeril. PJ; Goldberger. JJ. (2008). Assessment of autonomic function in
cardiovascular disease. J Am Coll Cardiol 51: 1725-1733.
http://dx.doi.Org/10.1016/i.iacc.2008.01.038
Lanzinger. S: Breitner. S: Neas. L; Cascio. W: Piaz-Sanchez. P; Hinderliter. A: Peters. A: Pevlin.
RB: Schneider. A. (2014). The impact of decreases in air temperature and increases in ozone on
markers of endothelial function in individuals having type-2 diabetes. Environ Res 134C: 331-338.
http://dx.doi.Org/10.1016/i.envres.2014.08.003
Larrieu. S: Jusot. JF; Blanchard. M; Prouvost. H; Peclercq. C: Fabre. P; Pascal. L; Le Tertre. A:
Wagner. V: Riviere. S: Chardon. B: Borelli. P: Cassadou. S: Eilstein. P: Lefranc. A. (2007). Short
term effects of air pollution on hospitalizations for cardiovascular diseases in eight French cities:
The PSAS program. Sci Total Environ 387: 105-112.
http://dx.doi.Org/10.1016/i.scitotenv.2007.07.025
Li. W: Porans. KS: Wilker. EH: Rice. MB: Liungman. PL: Schwartz. JP: Coull. BA: Koutrakis. P:
Gold. PR: Keanev. JF: Vasan. RS: Benjamin. EJ; Mittleman. MA. (2017). Short-term exposure to
ambient air pollution and biomarkers of systemic inflammation: the Framingham Heart Study.
Arterioscler Thromb Vase Biol 37: 1793-1800. http://dx.doi.org/10.1161/ATVBAHA.117.309799
Li. W: Wilker. EH: Porans. KS: Rice. MB: Schwartz. J: Coull. BA: Koutrakis. P; Gold. PR: Keanev.
JF: Lin. H: Vasan. RS: Benjamin. EJ: Mittleman. MA. (2016). Short-term exposure to air pollution
and biomarkers of oxidative stress: the framingham heart study. J Am Heart Assoc 5.
http://dx.doi.org/10.1161/JAHA.115.002742
Liu. WT; Lee. KY; Lee. HC; Chuang. HC; Wu. P; Juang. JN; Chuang. KJ. (2016). The association of
annual air pollution exposure with blood pressure among patients with sleep-disordered breathing.
Sci Total Environ 543: 61-66. http://dx.doi.org/10.1016/i.scitotenv.2015.10.135
4-159
-------
Ljungman. PL; Wilker. EH; Rice. MB; Schwartz. J; Gold. PR; Koutrakis. P; Vita. JA; Mitchell. GF;
Vasan. RS; Benjamin. EJ; Mittleman. MA; Hamburg. NM. (2014). Short-term exposure to air
pollution and digital vascular function. Am J Epidemiol 180: 482-489.
http://dx.doi.org/10.1093/aie/kwul61
Maheswaran. R; Pearson. T; Beevers. SD; Campbell. MJ; Wolfe. CD. (2016). Air pollution and
subtypes, severity and vulnerability to ischemic stroke-a population based case-crossover study.
PLoS ONE 11: eO 158556. http://dx.doi.org/10.1371/iournal.pone.0158556
Martinez-Campos. C; Lara-Padilla. E; Bobadilla-Lugo. RA; Kross. RD; Villanueva. C. (2012). Effects
of exercise on oxidative stress in rats induced by ozone. ScientificWorld Journal 2012: 1-5.
http://dx.doi.org/10.1100/2012/135921
Mclntosh-Kastrinsky. R; Diaz-Sanchez. D; Sexton. KG; Jania. CM; Zavala. J; Tillev. SL; Jaspers. I;
Gilmour. MI; Devlin. RB; Cascio. WE; Tong. H. (2013). Photochemically altered air pollution
mixtures and contractile parameters in isolated murine hearts before and after ischemia. Environ
Health Perspect 121: 1344-1348. http://dx.doi.org/10.1289/ehp.1306609
Mechtouff. L; Canoui-Poitrine. F; Schott. AM; Nighoghossian. N; Trouillas. P; Termoz. A; Porthault-
Chatard. S; David. JS; Chasles. V; Derex. L. (2012). Lack of association between air pollutant
exposure and short-term risk of ischaemic stroke in Lyon, France. Int J Stroke 7: 669-674.
http://dx.doi.org/10.1111/i. 1747-4949.2011.00737.x
Metzger. KB; Klein. M; Flanders. WD; Peel. JL; Mulholland. JA; Langberg. JJ; Tolbert. PE. (2007).
Ambient air pollution and cardiac arrhythmias in patients with implantable defibrillators.
Epidemiology 18: 585-592. http://dx.doi.org/10.1097/EDE.Qb013e318124ffDe
Miller. DB; Snow. SJ; Henriquez. A; Schladweiler. MC; Ledbetter. AD; Richards. JE; Andrews. PL;
Kodavanti. UP. (2016). Systemic metabolic derangement, pulmonary effects, and insulin
insufficiency following subchronic ozone exposure in rats. Toxicol Appl Pharmacol 306: 47-57.
http://dx.doi.Org/10.1016/i.taap.2016.06.027
Miloievic. A; Wilkinson. P; Armstrong. B; Bhaskaran. K; Smeeth. L; Haiat. S. (2014). Short-term
effects of air pollution on a range of cardiovascular events in England and Wales: Case-crossover
analysis of the MINAP database, hospital admissions and mortality. Heart 100: 1093-1098.
http://dx.doi.Org/10.l 136/heartinl-2013-304963
Mirowskv. JE; Carrawav. MS; Dhingra. R; Tong. H; Neas. L; Diaz-Sanchez. D; Cascio. W; Case. M;
Crooks. J; Hauser. ER; Elaine Dowdy. Z; Kraus. WE; Devlin. RB. (2017). Ozone exposure is
associated with acute changes in inflammation, fibrinolysis, and endothelial cell function in
coronary artery disease patients. Environ Health 16: 126. http://dx.doi.org/10.1186/sl2940-Q17-
0335-0
Montresor-Lopez. JA; Yanoskv. JD; Mittleman. MA; Sapkota. A; He. X; Hibbert. JD; Wirth. MP;
Puett. RC. (2015). Short-term exposure to ambient ozone and stroke hospital admission: A case-
crossover analysis. J Expo Sci Environ Epidemiol 26: 162-166.
http://dx.doi.org/10.1038/ies.2015.48
Muhaiarine. N; Mustard. C; Roos. LL; Young. TK; Gelskev. PE. (1997). Comparison of survey and
physician claims data for detecting hypertension. J Clin Epidemiol 50: 711-718.
http://dx.doi.org/10.1016/S0895-4356(97')00019-X
Murgia. N; Brisman. J; Claesson. A; Muzi. G; Olin. AC; Toren. K. (2014). Validity of a questionnaire-
based diagnosis of chronic obstructive pulmonary disease in a general population-based study.
BMC Pulm Med 14: 49. http://dx.doi.org/10.1186/1471-2466-14-49
4-160
-------
Notarius. CF; Butler. GC; Ando. S; Pollard. MJ; Senn. BL; Floras. JS. (1999). Dissociation between
microneurographic and heart rate variability estimates of sympathetic tone in normal subjects and
patients with heart failure. Clin Sci (Lond) 96: 557-565.
Nuvolone. D; Balzi. D; Pepe. P; Chini. M; Scala. D; Giovannini. F; Cipriani. F; Barchielli. A. (2013).
Ozone short-term exposure and acute coronary events: A multicities study in Tuscany (Italy).
Environ Res 126: 17-23. http://dx.doi.Org/10.1016/i.envres.2013.08.002
Paffett. ML: Zvchowski. KE: Sheppard. L: Robertson. S; Weaver. JM: Lucas. SN; Campen. MJ.
(2015). Ozone inhalation impairs coronary artery dilation via intracellular oxidative stress:
Evidence for serum-borne factors as drivers of systemic toxicity. Toxicol Sci 146: 244-253.
http://dx.doi.org/10.1093/toxsci/kfV093
Peel. JL; Metzger. KB: Klein. M; Flanders. WD: Mulholland. JA; Tolbert. PE. (2007). Ambient air
pollution and cardiovascular emergency department visits in potentially sensitive groups. Am J
Epidemiol 165: 625-633. http://dx.doi.org/10.1093/aie/kwk051
Perepu. RS: Garcia. C: Postal. D; Sethi. R. (2010). Enhanced death signaling in ozone-exposed
ischemic-reperfused hearts. Mol Cell Biochem 336: 55-64. http://dx.doi.org/10.1007/sl 1010-009-
0265-4
Perepu. RSP: Postal. DE: Garcia. C: Kennedy. RH: Sethi. R. (2012). Cardiac dysfunction subsequent
to chronic ozone exposure in rats. Mol Cell Biochem 360: 339-345.
http://dx.doi.org/10.1007/sll010-Qll-1073-l
Pilz. V: Wolf. K: Breitner. S: Ruckerl. R: Koenig. W: Rathmann. W: Cvrvs. J: Peters. A: Schneider.
A: group. K-S. (2018). C-reactive protein (CRP) and long-term air pollution with a focus on
ultrafine particles. Int J Hyg Environ Health. http://dx.doi.Org/10.1016/i.iiheh.2018.01.016
Poloniecki. JD: Atkinson. RW: Ponce de Leon. A: Anderson. HR. (1997). Daily time series for
cardiovascular hospital admissions and previous day's air pollution in London, UK. Occup Environ
Med 54: 535-540. http://dx.doi.Org/10.1136/oem.54.8.535
Pradeau. C: Rondeau. V: Leveaue. E: Guernion. PY: Tentillier. E: Thicoipe. M: Brochard. P. (2015).
Air pollution and activation of mobile medical team for out-of-hospital cardiac arrest. Am J Emerg
Med 33: 367-372. http://dx.doi.Org/10.1016/i.aiem.2014.12.007
Qin. XD; Qian. Z; Vaughn. MG: Trevathan. E; Emo. B; Paul. G: Ren. WH: Hao. YT; Dong. GH.
(2015). Gender-specific differences of interaction between obesity and air pollution on stroke and
cardiovascular diseases in Chinese adults from a high pollution range area: A large population
based cross sectional study. Sci Total Environ 529: 243-248.
http ://dx.doi .org/10.1016/i. scitotenv.2015.05.041
Ramanathan. G: Yin. F: Speck. M: Tseng. CH: Brook. JR: Silverman. F: Urch. B: Brook. RD: Arauio.
JA. (2016). Effects of urban fine particulate matter and ozone on HDL functionality. Part Fibre
Toxicol 13: 26. http://dx.doi.org/10.1186/sl2989-016-0139-3
Ramot. Y: Kodavanti. UP: Kissling. GE: Ledbetter. AD: Nvska. A. (2015). Clinical and pathological
manifestations of cardiovascular disease in rat models: the influence of acute ozone exposure. Inhal
Toxicol 27: 26-38. http://dx.doi.org/10.3109/08958378.2014.954168
Rasche. M: Walther. M: Schiffner. R: Kroegel. N: Rupprecht. S: Schlattmann. P: Schulze. PC:
Franzke. P; Witte. OW: Schwab. M; Rakers. F. (2018). Rapid increases in nitrogen oxides are
associated with acute myocardial infarction: A case-crossover study. Eur J Prev
Cardiol2047487318755804. http://dx.doi.org/10.1177/2047487318755804
4-161
-------
Raza. A; Bellander. T; Bero-Bedada. G; Dahlquist. M; Hollcnbcrg. J; Jonsson. M; Lind. T;
Rosenqvist. M; Svensson. L; Ljungman. PL. (2014). Short-term effects of air pollution on out-of-
hospital cardiac arrest in Stockholm. Eur Heart J 35: 861-867.
http://dx.doi.org/10.1093/eurhearti/eht489
Rich. DQ; Balmes. JR; Frampton. MW: Zareba. W; Stark. P; Ariomandi. M; Hazucha. MJ; Costantini.
MG: Ganz. P; Hollenbeck-Pringle. D; Dagincourt. N; Bromberg. PA. (2018). Cardiovascular
function and ozone exposure: The Multicenter Ozone Study in oldEr Subjects (MOSES). Environ
Int 119: 193-202. http://dx.doi.org/10.1016/i.envint.2018.06.014
Robertson. S; Colombo. ES; Lucas. SN; Hall. PR; Febbraio. M; Paffett. ML; Campen. MJ. (2013).
CD36 mediates endothelial dysfunction downstream of circulating factors induced by 03 exposure.
Toxicol Sci 134: 304-311. http://dx.doi.org/10.1093/toxsci/kftl07
Rodopoulou. S; Chalbot. MC; Samoli. E; Dubois. DW; Filippo. B; Kavouras. IG. (2014). Air pollution
and hospital emergency room and admissions for cardiovascular and respiratory diseases in Doria
Ana County, New Mexico. Environ Res 129: 39-46. http://dx.doi.Org/10.1016/i.envres.2013.12.006
Rodopoulou. S; Samoli. E; Chalbot. MG; Kavouras. IG. (2015). Air pollution and cardiovascular and
respiratory emergency visits in Central Arkansas: A time-series analysis. Sci Total Environ 536:
872-879. http://dx.doi.Org/10.1016/i.scitotenv.2015.06.056
Roonev. AA; Bovles. AL; Wolfe. MS; Bucher. JR; Thayer. KA. (2014). Systematic review and
evidence integration for literature-based environmental health science assessments. Environ Health
Perspect 122: 711-718. http://dx.doi.org/10.1289/ehp.1307972
Rosenthal. FS; Kuisma. M; Lanki. T; Hussein. T; Bovd. J; Halonen. JI; Pekkanen. J. (2013).
Association of ozone and particulate air pollution with out-of-hospital cardiac arrest in Helsinki,
Finland: evidence for two different etiologies. J Expo Sci Environ Epidemiol 23: 281-288.
http://dx.doi.org/10.1038/ies.2Q12.121
Rowan III. WH; Campen. MJ; Wichers. LB; Watkinson. WP. (2007). Heart rate variability in rodents:
uses and caveats in toxicological studies. Cardiovasc Toxicol 7: 28-51.
Rudez. G; Janssen. NA; Kilinc. E; Leebeek. FW; Gerlofs-Niiland. ME; Spronk. HM; ten Cate. H;
Cassee. FR; de Maat. MP. (2009). Effects of ambient air pollution on hemostasis and inflammation.
Environ Health Perspect 117: 995-1001. http://dx.doi.org/10.1289/ehp.080Q437
Sacks. JD; Ito. K; Wilson. WE; Neas. LM. (2012). Impact of covariate models on the assessment of
the air pollution-mortality association in a single- and multipollutant context. Am J Epidemiol 176:
622-634. http://dx.doi.org/10.1093/aie/kwsl35
Sade. MY; Vodonos. A; Novack. V; Friger. M; Amit. G. uv; Katra. I; Schwartz. J; Novack. L. (2015).
Can air pollution trigger an onset of atrial fibrillation: a population-based study. Air Qual Atmos
Health 8: 413-420. http://dx.doi.org/10.1007/sll869-014-0295-2
Sarnat. SE; Suh. HH; Coull. BA; Schwartz. J; Stone. PH; Gold. DR. (2006). Ambient particulate air
pollution and cardiac arrhythmia in a panel of older adults in Steubenville, Ohio. Occup Environ
Med 63: 700-706. http://dx.doi.org/10.1136/oem.2006.027292
Sarnat. SE; Winquist. A; Schauer. JJ; Turner. JR; Sarnat. JA. (2015). Fine particulate matter
components and emergency department visits for cardiovascular and respiratory diseases in the St.
Louis, Missouri-Illinois, metropolitan area. Environ Health Perspect 123: 437-444.
http://dx.doi.org/10.1289/ehp.1307776
4-162
-------
Sethi. R; Manchanda. S; Perepu. RS; Kumar. A; Garcia. C; Kennedy. RH: Palakurthi. S; Postal. D.
(2012). Differential expression of caveolin-1 and caveolin-3: potential marker for cardiac toxicity
subsequent to chronic ozone inhalation. Mol Cell Biochem 369: 9-15.
http://dx.doi.org/10.1007/sllQ10-012-1363-2
Silverman. RA; Ito. K; Freese. J; Kaufman. BJ; De Claro. D; Braun. J; Prezant. DJ. (2010).
Association of ambient fine particles with out-of-hospital cardiac arrests in New York City. Am J
Epidemiol 172: 917-923. http://dx.doi.org/10.1093/aie/kwq217
Sivagangabalan. G; Spears. D; Masse. S; Urch. B; Brook. RD: Silverman. F; Gold. PR; Lukic. KZ:
Speck. M: Kusha. M: Farid. T: Poku. K: Shi. E: Floras. J; Nanthakumar. K. (2011). The effect of
air pollution on spatial dispersion of myocardial repolarization in healthy human volunteers. J Am
Coll Cardiol 57: 198-206. http://dx.doi.Org/10.1016/i.iacc.2010.08.625
Snow. SJ; Cheng. WY; Henriquez. A; Hodge. M; Bass. V; Nelson. GM; Carswell. G; Richards. JE;
Schladweiler. MC; Ledbetter. AD: Chorlev. B: Gowdv. KM: Tong. H: Kodavanti. UP. (2018).
Ozone-induced vascular contractility and pulmonary injury are differentially impacted by diets
enriched with coconut oil, fish oil, and olive oil. Toxicol Sci 163: 5769.
http://dx.doi.org/10.1093/toxsci/kfV003
Spiezia. L: Campello. E: Bon. M: Maggiolo. S: Pelizzaro. E: Simioni. P. (2014). Short-term exposure
to high levels of air pollution as a risk factor for acute isolated pulmonary embolism. Thromb Res
Suppl 134: 259-263. http://dx.doi.Org/10.1016/i.thromres.2014.05.011
Steinvil. A: Kordova-Biezuner. L: Shapira. I: Berliner. S: Rogowski. O. (2008). Short-term exposure
to air pollution and inflammation-sensitive biomarkers. Environ Res 106: 51-61.
http://dx.doi.Org/10.1016/i.envres.2007.08.006
Stieb. DM: Szvszkowicz. 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 8:
25. http://dx.doi.org/10.1186/1476-069X-8-25
Stiegel. MA: Pleil. JD: Sobus. JR: Madden. MC. (2016). Inflammatory cytokines and white blood cell
counts response to environmental levels of diesel exhaust and ozone inhalation exposures. PLoS
ONE 11: e0152458. http://dx.doi.org/10.1371/iournal.pone.0152458
Stiegel. MA: Pleil. JD: Sobus. JR: Stevens. T; Madden. MC. (2017). Linking physiological parameters
to perturbations in the human exposome: Environmental exposures modify blood pressure and lung
function via inflammatory cytokine pathway. J Toxicol Environ Health A 80: 485-501.
http://dx.doi.org/10.1080/15287394.2017.133Q578
Stranev. I,: Finn. J: Dennekamp. M: Bremner. A: Tonkin. A: Jacobs. I. (2014). Evaluating the impact
of air pollution on the incidence of out-of-hospital cardiac arrest in the Perth Metropolitan Region:
2000-2010. J Epidemiol Community Health 68: 6-12. http://dx.doi.Org/10.l 136/iech-2013-202955
Suissa. L: Fortier. M: Lachaud. S: Staccini. P: Mahagne. MH. (2013). Ozone air pollution and
ischaemic stroke occurrence: a case-crossover study in Nice, France. BMJ Open 3: e004060.
http://dx.doi.Org/10.l 136/bmjopen-2013-004060
Tallon. LA: Maniourides. J: Pun. VC: Mittleman. MA: Kioumourtzoglou. MA: Coull. B; Suh. H.
(2017). Erectile dysfunction and exposure to ambient Air pollution in a nationally representative
cohort of older Men. Environ Health 16: 12. http://dx.doi.Org/10.l 186/s 12940-017-0216-6
Tanaka. T; Narazaki. M; Kishimoto. T. (2014). IL-6 in inflammation, immunity, and disease
[Review]. Cold Spring Harbor Perspectives in Biology 6: aO 16295.
http://dx.doi.org/10.1101/cshperspect.a016295
4-163
-------
Tankerslev. CG; Gcorgakopoulos. D; Tang. WY: Sborz. N. (2013). Effects of ozone and particulate
matter on cardiac mechanics: role of the atrial natriuretic peptide gene. Toxicol Sci 131: 95-107.
http://dx.doi.org/10.lQ93/toxsci/kfs273
Tankerslev. CG; Peng. RD: Bedga. D; Gabrielson. K; Champion. HC. (2010). Variation in
echocardiographic and cardiac hemodynamic effects of PM and ozone inhalation exposure in
strains related to Nppa and Nprl gene knock-out mice. Inhal Toxicol 22: 695-707.
http://dx.doi.org/10.3109/08958378.201Q.487549
Thompson. AM; Zanobetti. A; Silverman. F; Schwartz. J; Coull. B; Urch. B; Speck. M; Brook. JR;
Manno. M; Gold. DR. (2010). Baseline repeated measures from controlled human exposure
studies: Associations between ambient air pollution exposure and the systemic inflammatory
biomarkers IL-6 and fibrinogen. Environ Health Perspect 118: 120-124.
http://dx.doi.org/10.1289/ehp.090055Q
Thomson. EM; Pal. S; Guenette. J; Wade. MG; Atlas. E; Hollowav. AC; Williams. A; Vincent. R.
(2016). Ozone inhalation provokes glucocorticoid-dependent and -independent effects on
inflammatory and metabolic pathways. Toxicol Sci 152: 17-28.
http://dx.doi.org/10.1093/toxsci/kfw061
Thomson. EM; Vladisavlievic. D; Mohottalage. S; Kumarathasan. P; Vincent. R. (2013). Mapping
acute systemic effects of inhaled particulate matter and ozone: multiorgan gene expression and
glucocorticoid activity. Toxicol Sci 135: 169-181. http://dx.doi.org/10.1093/toxsci/kftl37
Toren. K; Brisman. J; Jarvholm. B. (1993). Asthma and asthma-like symptoms in adults assessed by
questionnaires: A literature review [Review]. Chest 104: 600-608.
http://dx.doi.Org/10.1378/chest.104.2.600
Turner. MC; Jerrett. M; Pope. A. Ill; Krewski. D; Gapstur. SM; Diver. WR; Beckerman. BS;
Marshall. JD; Su. J; Crouse. PL; Burnett. RT. (2016). Long-term ozone exposure and mortality in a
large prospective study. Am J Respir Crit Care Med 193: 1134-1142.
http://dx.doi.Org/10.l 164/rccm.201508-1633QC
U.S. EPA (U.S. Environmental Protection Agency). (1991). Guidelines for developmental toxicity risk
assessment (pp. 1-71). (EPA/600/FR-91/001). Washington, DC: U.S. Environmental Protection
Agency, Risk Assessment Forum, http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=23162
U.S. EPA (U.S. Environmental Protection Agency). (1996a). Air quality criteria for ozone and related
photochemical oxidants [EPA Report]. (EPA/600/P-93/004AF). Research Triangle Park, NC.
U.S. EPA (U.S. Environmental Protection Agency). (1996b). Guidelines for reproductive toxicity risk
assessment (pp. 1-143). (EPA/630/R-96/009). Washington, DC: U.S. Environmental Protection
Agency, Risk Assessment Forum, https://www.epa.gov/sites/production/files/2014-
11/documents/guidelines repro toxicitv.pdf
U.S. EPA (U.S. Environmental Protection Agency). (1998). Guidelines for neurotoxicity risk
assessment [EPA Report] (pp. 1-89). (EPA/630/R-95/00IF). Washington, DC: U.S. Environmental
Protection Agency, Risk Assessment Forum, http://www.epa.gov/risk/guidelines-neurotoxicitv-
risk-assessment
U.S. EPA (U.S. Environmental Protection Agency). (2005). Guidelines for carcinogen risk assessment
[EPA Report]. (EPA/630/P-03/00IB). Washington, DC: U.S. Environmental Protection Agency,
Risk Assessment Forum, https://www.epa.gov/sites/production/files/2013-
09/documents/cancer guidelines final 3-25-05.pdf
4-164
-------
U.S. EPA (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: U.S.
Environmental Protection Agency, Office of Research and Development, National Center for
Environmental Assessment-RTP Office.
http://cfpub.cpa.gov/ncca/cfin/rccordisplav.cfin7dcidH49923
U.S. EPA (U.S. Environmental Protection Agency). (2008). Integrated science assessment for sulfur
oxides: Health criteria [EPA Report]. (EPA/600/R-08/047F). Research Triangle Park, NC: U.S.
Environmental Protection Agency, Office of Research and Development, National Center for
Environmental Assessment- RTP. http://cfpub.cpa.gov/ncca/cfm/rccordisplav.cfm?dcid= 198843
U.S. EPA (U.S. Environmental Protection Agency). (2013a). Integrated science assessment for ozone
and related photochemical oxidants [EPA Report]. (EPA/600/R-10/076F). Research Triangle Park,
NC: U.S. Environmental Protection Agency, Office of Research and Development, National Center
for Environmental Assessment-RTP Division.
http://cfpub.epa.gov/ncea/isa/recordisplav.cfm?deid=247492
U.S. EPA (U.S. Environmental Protection Agency). (2013b). Toxicological review of
trimethylbenzenes (CASRN 25551-13-7, 95-63-6, 526-73-8, and 108-67-8) in support of summary
information on the Integrated Risk Information System (IRIS): revised external review draft [EPA
Report]. (EPA/635/R13/171a). Washington, D.C.: U.S. Environmental Protection Agency,
National Center for Environmental Assessment.
http://vosemite.epa.gov/sab/SABPRODUCT.NSF/b5d8alce9b07293485257375007012b7/eele28Q
e77586de985257b65005d37e7!OpenDocument
U.S. EPA (U.S. Environmental Protection Agency). (2015). Preamble to the integrated science
assessments [EPA Report]. (EPA/600/R-15/067). Research Triangle Park, NC: U.S. Environmental
Protection Agency, Office of Research and Development, National Center for Environmental
Assessment, RTP Division. https://cfpub.epa.gov/ncea/isa/recordisplav.cfm?deid=310244
van Rossem. L; Rifas-Shiman. SL; Mellv. SJ: Kloog. I; Luttmann-Gibson. H; Zanobetti. A; Coull.
BA; Schwartz. JD; Mittleman. MA; Oken. E; Gillman. MW: Koutrakis. P: Gold. DR. (2015).
Prenatal air pollution exposure and newborn blood pressure. Environ Health Perspect 123: 353-
359. http://dx.doi.org/10.1289/ehp.1307419
Vanos. JK; Hebbern. C: Cakmak. S. (2014). Risk assessment for cardiovascular and respiratory
mortality due to air pollution and synoptic meteorology in 10 Canadian cities. Environ Pollut 185:
322-332. http://dx.doi.Org/10.1016/i.envpol.2013.ll.007
Vencloviene. J: Braziene. A; Dedele. A; Lopatiene. K; Dobozinskas. P. (2017). Associations of short-
term exposure to ambient air pollutants with emergency ambulance calls for the exacerbation of
essential arterial hypertension. Int J Environ Health Res 27: 509-524.
http://dx.doi.org/10.1080/09603123.2017.14Q5246
Vidale. S: Arnaboldi. M; Bosio. V: Corrado. G: Guidotti. M; Sterzi. R; Campana. C. (2017). Short-
term air pollution exposure and cardiovascular events: A 10-year study in the urban area of Como,
Italy. Int J Cardiol 248: 389-393. http://dx.doi.Org/10.1016/i.iicard.2017.06.037
Villeneuve. PJ: Chen. L; Stieb. D: Rowe. BH. (2006). Associations between outdoor air pollution and
emergency department visits for stroke in Edmonton, Canada. Eur J Epidemiol 21: 689-700.
http://dx.doi.org/10.1007/sl0654-006-905Q-9
von Elm. E; Altman. DG: Egger. M; Pocock. SJ: Gotzschc. PC: Vandenbroucke. JP. (2007). The
strengthening the reporting of observational studies in epidemiology (strobe) statement: guidelines
for reporting observational studies [Review]. PLoS Med 4: e296.
http://dx.doi.org/10.1371/iournal.pmed.0040296
4-165
-------
von Klot. S; Peters. A; Aalto. P; Bellander. T; Bcrglind. N; P'Ippoliti. D; Elosua. R; Hormann. A;
Kulmala. M; Lanki. T; Lowel. H; Pekkanen. J; Picciotto. S; Sunver. J; Forastiere. F. (2005).
Ambient air pollution is associated with increased risk of hospital cardiac readmissions of
myocardial infarction survivors in five European cities. Circulation 112: 3073-3079.
http://dx.doi.org/10.1161/CIRCULATIQNAHA.105.548743
Wagner. JG; Allen. K: Yang. HY: Nan. B: Morishita. M: Mukheriee. B: Dvonch. JT: Spino. C: Fink.
GD; Raiagopalan. S; Sun. O: Brook. RD: Harkema. JR. (2014). Cardiovascular depression in rats
exposed to inhaled particulate matter and ozone: effects of diet-induced metabolic syndrome.
Environ Health Perspect 122: 27-33. http://dx.doi.org/10.1289/ehp.1307085
Wang. G; Jiang. R: Zhao. Z: Song. W. (2013). Effects of ozone and fine particulate matter (PM[2.5])
on rat system inflammation and cardiac function. Toxicol Lett 217: 23-33.
http://dx.doi.Org/10.1016/i.toxlet.2012.ll.009
Wang. X: Kindzierski. W; Kaul. P. (2015a). Air pollution and acute myocardial infarction hospital
admission in Alberta, Canada: a three-step procedure case-crossover study. PLoS ONE 10:
e0132769. http://dx.doi.org/10.1371/iournal.pone.0132769
Wang. X: Kindzierski. W; Kaul. P. (2015b). Comparison of transient associations of air pollution and
AMI hospitalisation in two cities of Alberta, Canada, using a case-crossover design. BMJ Open 5:
e009169. http://dx.doi.Org/10.l 136/bmiopen-2015-009169
Weakley. J; Webber. MP: Ye. F: Zeig-Owens. R; Cohen. HW: Hall. CB: Kelly. K: Prezant. DJ.
(2013). Agreement between obstructive airways disease diagnoses from self-report questionnaires
and medical records. Prev Med 57: 38-42. http://dx.doi.Org/10.1016/i.ypmed.2013.04.001
Weichenthal. S: Pinault. LL; Burnett. RT. (2017). Impact of oxidant gases on the relationship between
outdoor fine particulate air pollution and nonaccidental, cardiovascular, and respiratory mortality.
Sci Rep 7: 16401. http://dx.doi.org/10.1038/s41598-017-16770-v
Wing. JJ; Adar. SD; Sanchez. BN: Morgenstern. LB: Smith. MA: Lisabeth. LP. (2015). Ethnic
differences in ambient air pollution and risk of acute ischemic stroke. Environ Res 143: 62-67.
http://dx.doi.Org/10.1016/i.envres.2015.09.031
Wing. JJ: Adar. SD: Sanchez. BN: Morgenstern. LB: Smith. MA: Lisabeth. LP. (2017a). Short-term
exposures to ambient air pollution and risk of recurrent ischemic stroke. Environ Res 152: 304-307.
http://dx.doi.Org/10.1016/i.envres.2016.l 1.001
Wing. JJ: Sanchez. BN: Adar. SD: Meurer. WJ: Morgenstern. LB: Smith. MA: Lisabeth. LP. (2017b).
Synergism of short-term air pollution exposures and neighborhood disadvantage on initial stroke
severity. Stroke 48: 3126-3129. http://dx.doi.org/10.1161/STRQKEAHA.117.018816
Winquist. A: Klein. M: Tolbert. P: Flanders. WP: Hess. J: Sarnat. SE. (2012). Comparison of
emergency department and hospital admissions data for air pollution time-series studies. Environ
Health 11: 70. http://dx.doi.org/10.1186/1476-069X-11-70
Xu. X: Sun. Y: Ha. S: Talbott. EO: Lissaker. CTK. (2013). Association between ozone exposure and
onset of stroke in Allegheny County, Pennsylvania, USA, 1994-2000. Neuroepidemiology 41: 2-6.
http://dx.doi.org/10.1159/00Q345138
Yang. BY: Oian. ZM: Vaughn. MG: Nelson. EJ: Pharmage. SC: Heinrich. J: Lin. S: Lawrence. WR:
Ma. H: Chen. PH: Hu. LW: Zeng. XW: Xu. SL: Zhang. C: Pong. GH. (2017). Is prehypertension
more strongly associated with long-term ambient air pollution exposure than hypertension?
Findings from the 33 Communities Chinese Health Study. Environ Pollut 229: 696-704.
http://dx.doi.Org/10.1016/i.envpol.2017.07.016
4-166
-------
Yang. CL; To. T; Fotv. RG; Stieb. DM; Dell. SD. (2011). Verifying a questionnaire diagnosis of
asthma in children using health claims data. BMC Pulm Med 11. http://dx.doi.org/10.1186/1471-
2466-11-52
Ying. Z; Allen. K; Zhong. J; Chen. M; Williams. KM; Wagner. JG; Lewandowski. R; Sun. Q;
Raiagopalan. S; Harkema. JR. (2016). Subacute inhalation exposure to ozone induces systemic
inflammation but not insulin resistance in a diabetic mouse model. Inhal Toxicol 28: 155-163.
http://dx.doi.org/10.3109/08958378.2016.11468Q8
Zanobetti. A; Canner. MJ; Stone. PH; Schwartz. J; Sher. D; Eagan-Bengston. E; Gates. KA; Hartley.
LH; Suh. H; Gold. DR. (2004). Ambient pollution and blood pressure in cardiac rehabilitation
patients. Circulation 110: 2184-2189. http://dx.doi.Org/10.l 161/01.cir.OOOO 143831.33243,d8
Zanobetti. A; Luttmann-Gibson. H; Horton. ES; Cohen. A; Coull. BA; Hoffmann. B; Schwartz. JD;
Mittleman. MA; Li. Y; Stone. PH; de Souza. C; Lamparello. B; Koutrakis. P; Gold. DR. (2014).
Brachial artery responses to ambient pollution, temperature, and humidity in people with type 2
diabetes: a repeated-measures study. Environ Health Perspect 122: 242-248.
http://dx.doi.org/10.1289/ehp.1206136
Zhao. Y; Oian. Z; Wang. J; Vaughn. MG; Liu. Y; Ren. W; Dong. G. (2013). Does obesity amplify the
association between ambient air pollution and increased blood pressure and hypertension in adults?
Findings from the 33 Communities Chinese Health Study [Letter], Int J Cardiol 168: E148-E150.
http://dx.doi.Org/10.1016/i.iicard.2013.08.071
Zhong. J; Allen. K; Rao. X; Ying. Z; Braunstein. Z; Kankanala. SR; Xia. C; Wang. X; Bramble. LA;
Wagner. JG; Lewandowski. R; Sun. Q; Harkema. JR; Raiagopalan. S. (2016). Repeated ozone
exposure exacerbates insulin resistance and activates innate immune response in genetically
susceptible mice. Inhal Toxicol 28: 383-392. http://dx.doi.Org/10.1080/08958378.2016.l 179373
Zvchowski. KE; Lucas. SN; Sanchez. B; Herbert. G; Campen. MJ. (2016). Hypoxia-induced
pulmonary arterial hypertension augments lung injury and airway reactivity caused by ozone
exposure. Toxicol Appl Pharmacol 305: 40-45. http://dx.doi.Org/10.1016/i.taap.2016.06.003
4-167
-------
APPENDIX 5 HEALTH EFFECTS —METABOLIC
EFFECTS
Summary of ( tinsal Determinations for Short- and l.onx-ierm Ozone
Kvpoxnre
-------
Diabetes is characterized by a continuum of hyperglycemia (i.e., elevated glucose level) resulting
from defects in insulin signaling, secretion, or both. Several types of diabetes have been classified by the
American Diabetes Association (ADA. 2014). Type 1 diabetes is caused by |3-cell dysfunction or
destruction that leads to insulin deficiency, while type 2 diabetes is characterized by defects in insulin
secretion in an insulin-resistant environment. Gestational diabetes mellitus is generally diagnosed during
the second or third trimester of pregnancy.
The subsections below provide an evaluation of the most policy-relevant scientific evidence
relating short-term ozone exposure to metabolic health effects. These sections focus on studies published
since the completion of the 2013 Ozone ISA. There are a limited number of recent epidemiologic studies
examining the effects of short-term ozone exposure on glucose tolerance, insulin sensitivity, and diabetes
control. In addition, multiple animal toxicological studies evaluate ozone-mediated effects, and these
studies indicate that short-term exposure to ozone affects glucose homeostasis and other factors that
contribute to metabolic syndrome.
5.1.1 Population, Exposure, Comparison, Outcome, and Study Design
(PECOS) Tool
The scope of this section is defined by a scoping tool that generally defines the relevant
Population, Exposure, Comparison, Outcome, and Study Design (PECOS). The PECOS tool defines the
parameters and provides a framework to help identify the relevant evidence in the literature to inform the
ISA. Because the 2013 Ozone ISA did not make a causality determination for short-term ozone exposure
and metabolic health effects, the epidemiologic studies evaluated are less limited in scope and not
targeted towards specific study locations, as reflected in the PECOS tool. The studies evaluated and
subsequently discussed within this section were identified using the following PECOS tool:
Experimental Studies:
• Population: Study populations of any controlled human exposure or animal toxicological study of
mammals at any lifestage
• Exposure: Short-term (in the order of minutes to weeks) inhalation exposure to relevant ozone
concentrations (i.e., 0.4 ppm or below for humans, 2 ppm or below for other mammals)
• Comparison: Human subjects that serve as their own controls with an appropriate washout period
or when comparison to a reference population exposed to lower levels is available, or, in
toxicological studies of mammals, an appropriate comparison group that is exposed to a negative
control (i.e., clean air or filtered air control)
• Outcome: Metabolic effects (e.g., diabetes, metabolic syndrome, dyslipidemia, glucose
intolerance, insulin resistance, overweight, obesity)
• Study Design: Controlled human exposure (e.g., chamber) studies; in vivo acute, subacute, or
repeated-dose toxicity studies in mammals; and immunotoxicity studies
5-2
-------
Epidemiologic Studies:
• Population: Any population, including populations or lifestages that might be at increased risk
• Exposure: Ambient ozone from any source measured as short term (hours to days)
• Comparison: Per unit increase (in ppb)
• Outcome: Change in risk (incidence/prevalence) of metabolic effects
• Study Design: Epidemiologic studies consisting of panel, case-crossover, time-series studies, and
case-control studies; cross-sectional studies with appropriate timing of exposure for the health
endpoint of interest
5.1.2 Biological Plausibility
This section describes biological pathways that potentially underlie metabolic effects resulting
from short-term exposure to ozone. Figure 5-1 graphically depicts the proposed pathways as a continuum
of upstream events, connected by arrows, that may lead to downstream events observed in epidemiologic
studies. This discussion of "how" exposure to ozone may lead to metabolic effects contributes to an
understanding of the biological plausibility of epidemiologic results evaluated later. Additionally, most
studies cited in this subsection are discussed in greater detail throughout this Appendix. Note that the
structure of the biological plausibility sections and the role of biological plausibility in contributing to the
weight-of-evidence analysis used in the 2020 Ozone ISA are discussed in Section IS.4.2.
Ozone inhalation can contribute to metabolic syndrome or complications related to diabetes via
two primary pathways. The first pathway is initiated by activation of sensory neurons in the respiratory
tract, and the second pathway by respiratory tract injury, inflammation, and oxidative stress.
With respect to the first pathway, signals from the airways are integrated in the brainstem and
hypothalamus and lead to a modulation of the autonomic nervous system, subsequent activation of
neuroendocrine stress axes, and a depression of the hypothalamic-pituitary-thyroid axis. Specifically,
ozone inhalation stimulates nasopharyngeal and pulmonary nerves, as well as trigeminal and vagal nerve
receptors. Vagal sensory neurons reside in two ganglia, the jugular and nodose ganglia. In rats, short-term
ozone exposure in early life reduces the total number of jugular and nodose ganglia neurons (Zellner et
al.. 2011). From the nodose ganglion, vagal input passes to the nucleus tractus solitarius of the brainstem
or the parabrachial nucleus of the pons, two relay stations between the respiratory tract and the
hypothalamus (Gackicrc et al.. 201IV Within the nucleus tractus solitarius, neuronal activation following
ozone exposure was found specifically in areas with lung afferent fiber innervation (Gackicrc et al..
2011). In addition, the downstream paraventricular nucleus of the hypothalamus and regions of the
amygdala also showed increased neuronal activation with ozone exposure (Gackicrc et al.. 201IV Overall,
ozone exposure induces a signal that starts with stimulation of vagal afferents in the respiratory tract,
moves through relay centers in the brainstem, and ultimately impacts the more terminal stress-responsive
regions of the brain, including the amygdala and the hypothalamus, areas critically important to regulation
5-3
-------
of the metabolic system. Finally, ozone exposure activates neuroendocrine axes that regulate both
metabolism and the stress response, including the hypothalamic-pituitary-thyroid axis, the
sympathetic-adrenal-medullary axis, and the hypothalamic-pituitary-adrenal axis.
Short-
Term
Ozone
Exposure
Complications
Related to
Diabetes
Changes in
Contributors to
Metabolic
Syndrome
Note: The boxes above represent the effects for which there is experimental or epidemiologic evidence related to ozone exposure,
and the arrows indicate a proposed relationship between those effects. Solid arrows denote evidence of essentiality as provided, for
example, by an inhibitor of the pathway or a genetic knockout model used in an experimental study involving ozone exposure.
Shading around multiple boxes is used to denote a grouping of these effects. Arrows may connect individual boxes, groupings of
boxes, and individual boxes within groupings of boxes. Progression of effects is generally depicted from left to right and color coded
(white, exposure; green, initial effect; blue, intermediate effect; orange, effect at the population level or a key clinical effect). Here,
population-level effects generally reflect results of epidemiologic studies. When there are gaps in the evidence, there are
complementary gaps in the figure and the accompanying text below. The structure of the biological plausibility sections and the role
of biological plausibility in contributing to the weight-of-evidence analysis used in the 2020 Ozone ISA are discussed in
Section IS.4.2.
Figure 5-1 Potential biological pathways for metabolic outcomes following
short-term ozone exposure.
In the first of these axes, short-term ozone exposure leads to an inhibition of the
hypothal amic-pitui tary-thyroid axis, resulting in decreases in thyroid-stimulating hormone (TSH) and in
thyroid hormones T3 and T4 (Clemons and Garcia. 1980). Generally speaking, independent of ozone
exposure, clinically low T4 levels are associated with an increased risk of incident diabetes, likely
mediated by changes in insulin sensitivity and glucose tolerance (Chaker et al.. 2016). Short-term ozone
exposure also results in changes to core body temperature, which is regulated by thyroid hormones
5-4
-------
(Mautz and Bufalino. 1989; Clemons and Garcia. 1980). Alterations in metabolism and thermoregulation
that are modulated by thyroid function have been associated with contributors to metabolic syndrome
(Roos et al.. 2007).
In the second of these axes, ozone-induced stimulation of the hypothalamus activates the
sympathetic-adrenal-medullary (SAM) axis, leading to a release of adrenaline (epinephrine) from the
adrenal medulla (Miller et al.. 2016c). Similarly, following ozone-induced modulation of the autonomic
nervous system, the hypothalamic-pituitary-adrenal (HPA) axis is activated, inducing a release of
corticotropin-releasing hormone from the hypothalamus. This leads to a release of adrenocorticotropic
hormone (ACTH) from the anterior pituitary, and ultimately release of corticosterone (Cortisol in humans)
from the adrenal cortex (Thomson et al.. 2013; Ulrich-Lai and Herman. 2009). Notably, bilateral removal
of the adrenal medulla or of the entire adrenal glands (as well as blocking corticosterone synthesis)
prevents the ozone-induced release of adrenaline or corticosterone, respectively (Miller etal.. 2016c;
Thomson et al.. 2016). In addition, adrenaline increases ACTH secretion by the pituitary gland, thus
resulting in increased release of corticosterone (Thomson et al.. 2013; Ulrich-Lai and Herman. 2009).
Once released following ozone exposure, increased levels of adrenaline and corticosterone are
part of a short-term stress response that has been shown to affect multiple organs and tissues associated
with the metabolic system. Metabolic adaptation during stress events typically involves a transient
increase in circulating glucose (i.e., hyperglycemia) and fatty acid levels (i.e., hyperlipidemia), which
serve as the primary substrates for energy production (Kuo et al.. 2015). Following activation of the SAM
and HPA axes, the rapid surge of adrenaline and release of corticosterone trigger the release of glucose
and fats into the bloodstream from temporary storage sites in the body. Specifically, the binding of the
catecholamine adrenaline to adrenergic receptors increases blood glucose by inhibiting insulin release
from the pancreas, resulting in decreased insulin levels following ozone exposure (Ying et al.. 2016;
Zhong et al.. 2016; Miller et al.. 2015). The ozone-induced release of the glucocorticoid corticosterone
from activation of the HPA axis also contributes to increased blood glucose, for example, by stimulating
the synthesis of glucose from noncarbohydrate substances, primarily in the liver [hepatic
gluconeogenesis; Miller et al. (2016bVI. as well as by reducing glucose uptake and utilization in skeletal
muscle (Thomson et al.. 2016). In summary, because hyperglycemia can be prevented by removal of
either the adrenal glands or the adrenal medulla (Miller et al.. 2016c). ozone-induced elevated blood
glucose (hyperglycemia) likely results from the effects of adrenaline and corticosterone on insulin release,
hepatic gluconeogenesis, and glucose uptake.
In addition to increasing blood glucose levels, ozone-induced activation of the SAM and HPA
axes results in increased free fatty acids. Both adrenaline and corticosterone stimulate the breakdown of
fat in adipose tissue (lipolysis) resulting in the release of fatty acids in the circulation. In concordance,
short-term ozone exposure has been shown to cause adrenal hormone-derived hyperlipidemia, which is an
elevated amount of lipids in the blood (Miller et al.. 2016c; Miller et al.. 2016a).
5-5
-------
Following the short-term stress response outlined above, typically negative feedback mechanisms
are in place, which function to keep blood glucose and fatty acid levels stable. The hormones glucagon
and insulin, secreted by a- and P-cells within the pancreas, respectively, interact with the adrenal
hormones and function to maintain glucose homeostasis. When glucose levels are high, insulin is secreted
into the circulation, resulting in increased glucose uptake, utilization, and storage. When glucose levels
are low, insulin secretion is inhibited and glucagon is secreted instead, stimulating gluconeogenesis.
However, ozone exposure can disrupt these feedback mechanisms, leading to both insulin resistance
(Vella et al.. 2014) and glucose intolerance (Thomson et al.. 2018; Gordon et al.. 2017b; Miller et al..
2016b; Miller et al.. 2015; Bass et al.. 2013).
The second pathway through which short-term ozone exposure may affect the metabolic system
begins with the direct effect of ozone on respiratory tract injury, inflammation, and oxidative stress (see
Appendix 3 for details). Briefly, in the lung, ozone reacts with the respiratory epithelial cell lining fluid
leading to local and systemic inflammatory responses and oxidative stress (see also Appendix 7). When
circulating inflammatory cytokines and reactive oxygen species present in the bloodstream reach target
tissues and organs, such as the pancreas, liver, and adipose tissue, signaling mechanisms are initiated that
trigger local and systemic inflammation (Hotamisligil. 2017). Short-term ozone exposure has been shown
to cause inflammation in adipose tissue (Ying et al.. 2016; Zhong et al.. 2016; Sun et al.. 2013). Local and
systemic inflammation can contribute to insulin resistance (Arkan et al.. 2005). glucose intolerance (van
Beek et al.. 2014). hyperglycemia, and hyperlipidemia (Wellen and Hotamisligil. 2005). Indeed, in rats,
ozone-induced oxidative stress led to insulin resistance in skeletal muscle (Vella et al.. 2014). Notably,
adrenal-derived stress hormones can exacerbate systemic inflammation (Wellen and Hotamisligil. 2005).
Furthermore, blocking the release of these hormones through either adrenalectomy or pharmacological
inhibition results in a significant reduction in ozone-induced lung injury and inflammation (Henriquez et
al.. 2018; Henriquez et al.. 2017; Miller et al.. 2016c).
Acute activation of the neuroendocrine stress axes is typically an adaptive mechanism that allows
for rapid access to energy reserves. There are numerous negative feedback mechanisms in place that
promote homeostasis following this activation. Indeed, evidence shows that the ozone-induced increase in
adrenal-derived stress hormones is transient (Bass et al.. 2013). However, with repeated or continuous
exposure, or in those with pre-existing metabolic disease, short-term exposures to ozone could plausibly
cause the allostatic load to become too high, resulting in hormonal and metabolic dysregulation,
ultimately leading to diabetes or metabolic syndrome. Furthermore, short-term ozone exposure can
perturb the hypothalamic-pituitary-thyroid axis (Clemons and Garcia. 1980). resulting in further
dysregulation of metabolism and thermoregulation (Chaker et al.. 2016). Thus, the proposed pathways
discussed above provide biological plausibility for changes in contributors to metabolic syndrome and/or
complications related to diabetes following short-term ozone exposure, and will be used to inform a
causality determination, which is discussed later in this Appendix.
5-6
-------
5.1.3
Metabolic Syndrome
Individuals with metabolic syndrome are at a fivefold increased risk for developing type 2
diabetes and a twofold increased risk for developing cardiovascular disease within 5-10 years (Albcrti et
al.. 2009). Criteria for metabolic syndrome include elevated fasting blood glucose (hyperglycemia),
elevated triglycerides, low levels of high-density lipoprotein (HDL), obesity (particularly abdominal
obesity), and high blood pressure (Table 5-IV The presence of three out of the five criteria meets the
clinical diagnosis for metabolic syndrome. Below, the available studies that investigate the effects of
short-term ozone exposure on these and related endpoints are characterized. Evidence that short-term
ozone exposure may affect these important markers of metabolic function suggest that exposure to ozone
may contribute to or exacerbate metabolic syndrome.
Table 5-1 Criteria for clinical diagnosis of metabolic syndrome*.
Risk Factor
Threshold
Fasting glucose3
>100 mg/dL (5.6 mmol/L)
Triglycerides'5
>150 mg/dL (1.7 mmol/L)
HDL-Cb
<40 mg/dL (1.0 mmol/L in males); <50 mg/dL (1.3 mmol/L) in females
Waist circumference
>89 cm in women and >102 cm in males
Blood pressure0
Systolic >130 and/or diastolic >85 mm Hg
HDL-C = HDL cholesterol; mg/dL = milligrams per deciliter; mm Hg = millimeters of mercury; mmol/L = millimoles per liter.
'Presence of three out of the five criteria necessary to meet clinical diagnosis for metabolic syndrome.
aA person taking glucose-regulating medication is considered to exceed the threshold.
bA person taking drugs used to lower triglycerides or raise HDL-C is considered to exceed the threshold.
°A person taking blood pressure medication is considered to exceed the threshold.
Source: Modified with permission from the publisher, Alberti et al. (20091.
5-7
-------
5.1.3.1
Elevated Fasting Glucose
Elevated fasting blood glucose levels, or hyperglycemia, can occur when the body has too little
insulin, or if the insulin present cannot be used properly. Under normal conditions, insulin is secreted by
(3-cells within the pancreas in response to glucose levels. When glucose levels rise, depolarization of the
pancreatic (3-cells or modulation by other hormones stimulate insulin secretion (Nadal et al.. 2009). Thus,
during food intake, blood insulin levels rise, stimulating glucose uptake and replenishing body fuel
reserves in the form of triglycerides and glycogen. When glucose levels decrease (e.g., during fasting),
insulin secretion is inhibited and glucagon is secreted from the pancreas, which stimulates fuels, such as
lipids from adipose tissue and amino acids from muscle, to be mobilized to the bloodstream where they
are used by the liver to synthesize glucose.
The effects of short-term ozone exposure on blood glucose levels are characterized below. Given
the critical role of insulin on glucose homeostasis, the effects of short-term ozone exposure on glucose
tolerance and insulin secretion and signaling are also characterized. To measure how quickly glucose is
cleared from the blood, the glucose tolerance test (GTT) samples blood glucose levels at multiple time
points after glucose injection or ingestion. It is useful in diagnosing or monitoring diabetes or gestational
diabetes. The insulin tolerance test (ITT) induces an episode of hypoglycemia through the injection of
insulin. With a functioning HPA axis, hypoglycemia triggers a subsequent release of
cortisol/corticosterone from the adrenal glands and an increase in blood glucose levels. The Homeostatic
Model Assessment (HOMA) is a method for assessing (3-cell function (HOMA-(3%) and insulin resistance
(HOMA-IR). The model uses fasting blood glucose and insulin levels to calculate indices of (3-cell
function and insulin sensitivity, thus relating the level of feedback from glucose on (3-cells to increase
insulin secretion.
5.1.3.1.1 Epidemiologic Studies
One epidemiologic study of short-term ozone exposure and glucose or insulin homeostasis was
reviewed in the 2013 Ozone ISA. Chuang et al. (2010) observed increases in fasting glucose associated
with increased exposure to ozone concentrations averaged over 5 days. Recent epidemiologic studies
provide some evidence of associations between short-term ozone exposure and changes in glucose and
insulin homeostasis (Table 5-6). Specifically:
• Kim and Hong (2012) found increases in fasting glucose (0.19%; 95% CI: 0.09, 0.28%).1 insulin
(0.71%; 95% CI: 0.02, 1.38%), and HOMA (0.30%; 95% CI: 0.06, 0.53%) in the Korean Elderly
Environmental Panel (KEEP) cohort. The KEEP cohort consisted of 560 Koreans over 60 years
old. The authors observed a threefold larger percent increase in glucose, insulin, and HOMA-IR
1 All epidemiologic results standardized to a 15-ppb increase in 24-hour avg, 20-ppb increase in 8-hour daily max,
25-ppb increase in 1-hour daily max ozone concentrations, or a 10-ppb increase in seasonal/annual ozone
concentrations to facilitate comparability across studies.
5-8
-------
levels associated with short-term increases in ozone concentration in subjects with a previous
diagnosis of type 2 diabetes compared to those without type 2 diabetes (glucose [0.68%; 95% CI:
0.28, 1.07%], insulin [2.76%; 95% CI: 0.78, 4.75%], HOMA [1.21%; 95% CI: 0.44, 1.99%]). In
subjects without type 2 diabetes, a modest, positive association with glucose was observed
(0.09%; 95% CI: 0.02, 0.16%), while associations with insulin and HOMA were generally null.
Copollutant models with NO2 and PM10 were also evaluated. The associations with glucose
remained after adjustment for NO2 (0.16%; 95% CI: 0.06, 0.25%) and PM10 (0.15%; 95% CI:
0.01, 0.14%) in models that contained all subjects (i.e., with and without type 2 diabetes).
• Using 5,958 subjects from the Framingham Offspring Cohort and Third Generation Cohort, Li et
al. (2017) conducted a panel study evaluating the association between short-term ozone
concentrations and changes in fasting glucose, insulin, HOMA-IR, and other metabolic endpoints.
Decreases in fasting glucose, but not insulin or HOMA-IR, were observed with decreases in
3-, 5-, and 7-day moving avg ozone concentrations.
5.1.3.1.2 Controlled Human Exposure Studies
Controlled human exposure studies of short-term ozone exposure and glucose or insulin
homeostasis were not available for review in the 2013 Ozone ISA. A recent study showed no evidence
that short-term ozone exposure affected these endpoints (Table 5-7). In a study of 24 healthy volunteers
aged 22-30 years, Miller etal. (2016a') randomly exposed subjects to ozone (0.3 ppm) or filtered air for
2 hours while alternating 15 minutes of exercise and 15 minutes of rest. After a 2-week washout period,
the subjects had the alternate exposure. There were no changes in HOMA-IR or in nonfasting serum
insulin or glucose levels immediately following ozone exposure when compared with filtered air
exposure.
5.1.3.1.3 Animal Toxicological Studies
No animal toxicological studies of short-term ozone exposure and fasting blood glucose were
available for review in the 2013 Ozone ISA. A number of recent animal toxicological studies assessed
changes in both glucose and insulin homeostasis following short-term ozone exposure. As detailed below
and in the evidence inventory table that follows (Table 5-7). some studies reported ozone-induced
hyperglycemia, while others reported null effects. However, short-term ozone exposure consistently
impaired glucose tolerance. With respect to insulin, studies consistently demonstrated insulin resistance,
while changes in serum insulin were less consistent.
Some animal toxicological studies demonstrated that short-term ozone exposure resulted in an
increase in fasting glucose levels:
• Male Wistar Kyoto rats showed elevated fasting blood glucose following exposure to 0.5 or
1.0 ppm ozone but not to 0.25 ppm [6 hours/day for 2 days; Miller et al. (2015)1. Similarly, these
rats had elevated fasting blood glucose after 1 or 3 days of exposure to 1.0 ppm ozone
(5 hours/day). No effects were observed following exposure to 0.25 ppm ozone (Miller et al..
5-9
-------
2016b). Furthermore, Wistar rats exposed to 0.8 ppm ozone for 16 hours showed a statistically
significant increase in fasting blood glucose (Vella et al.. 2014).
• Male brown Norway rats, ages 1, 12, or 24 months, but not 4 months, showed elevated fasting
blood glucose following acute (6 hours/day for 2 days) exposure to 1.0 ppm ozone (Bass et al..
2013).
• Changes in fasting glucose levels may arise through ozone-mediated effects on the
hypothalamic-pituitary-adrenal axis. In male Wistar Kyoto rats, bilateral removal of the adrenal
medulla or of the entire adrenal gland prevented ozone-induced elevated blood glucose following
a 4-hour exposure to 1.0 ppm ozone (Miller et al.. 2016c).
However, other animal toxicological studies did not find a relationship between short-term ozone
exposure and fasting glucose levels:
• A 4-hour exposure to 0.8 ppm ozone in male Fischer 344 rats did not affect fasting blood glucose
levels (Thomson et al.. 2018). Similarly, a 5-hour exposure to 0.25, 0.5, or 1.0 ppm ozone did not
result in elevated blood glucose levels in either active or sedentary female Long-Evans rats
(Gordon et al.. 2017b). Moreover, in both male KKAy mice, a model for genetic diabetes, and in
KK mice, a diabetes-prone model, ozone exposure (0.5 ppm, 4 hours/day for 13 consecutive
weekdays) did not affect fasting blood glucose levels (Ying et al.. 2016; Zhong et al.. 2016). Male
and female brown Norway rats fed a normal, high-fat, or high-fructose diet for 12 weeks before
exposure to ozone or filtered air also showed no effects from acute (5 hours) or subacute
(1 day/week for 4 weeks) exposure to 0.8 ppm ozone on blood glucose levels in any group
(Gordon et al.. 2016). Likewise, there was no elevation in fasting glucose levels following a
4-hour ozone exposure (0.8 ppm) in male or female Long-Evans rats (Gordon et al.. 2017a).
Animal toxicology studies demonstrated a consistent relationship between short-term ozone
exposure and impaired glucose tolerance. Specifically:
• Male brown Norway rats of multiple ages (1-, 4-, 12-, and 24-months) showed impaired glucose
tolerance following acute (6 hours/day for 2 days) exposure to 1.0 ppm but not 0.25 ppm ozone
(Bass et al.. 2013). Time-course analysis in 4-month-old brown Norway rats showed that glucose
intolerance began on Day 1 of acute exposure to 1.0 ppm ozone and increased on Day 2,
returning to baseline 18 hours after exposure (Bass et al.. 2013). A similar result was found in
male Wistar Kyoto rats exposed to 1.0 ppm ozone for 1, 2, or 3 days (Miller et al.. 2016b; Miller
et al.. 2015). No effects were observed following exposure to 0.25 ppm ozone (Miller et al..
2016b). However, a 5-hour exposure to 0.25, 0.5, or 1.0 ppm ozone resulted in impaired glucose
tolerance in both active and sedentary female Long-Evans rats (Gordon et al.. 2017b). In male
Fischer 344 rats, a 4-hour exposure to 0.8 ppm ozone resulted in impaired glucose tolerance [at
30-minute post-injection only; Thomson et al. (2018)1. Finally, regardless of maternal diet or
exercise in Long-Evans rats, adult male, but not female offspring, showed glucose intolerance
after a 4-hour exposure to 0.8 ppm ozone (Gordon et al.. 2017a).
• These impairments in glucose tolerance are likely mediated by the adrenal hormones adrenaline
and corticosterone. Bilateral removal of the adrenal medulla (to diminish adrenaline) or of the
entire adrenal gland (to diminish both adrenaline and corticosterone) in male Wistar Kyoto rats
prevented ozone-induced glucose intolerance following a 4-hour exposure to 1.0 ppm ozone
(Miller et al.. 2016c).
As noted above, some animal toxicological studies suggest a relationship between short-term
ozone exposure and insulin resistance:
5-10
-------
• Male Wistar Kyoto rats exposed to 1.0 ppm ozone showed decreased insulin levels after 2 days;
levels returned to baseline 18 hours post-exposure (Miller et al.. 2015). In addition, Wistar rats
exposed to 0.8 ppm ozone for 16 hours showed a statistically significant increase in fasting
insulin levels and in HOMA-IR, suggesting impairment in insulin sensitivity (Vclla et al.. 2014).
The glucose infusion rate measured by euglycemic-hyperinsulinemic (EH) clamps decreased
following ozone exposure, suggesting peripheral insulin resistance. Further, insulin signaling was
impaired in skeletal muscle but not in liver or white adipose tissue following ozone exposure.
• In male KKAy mice, a model for genetic diabetes, ozone exposure (0.5 ppm, 4 hours/day for
13 consecutive weekdays) decreased fasting levels of insulin and HOMA-IR but did not affect
glucose levels in the insulin tolerance test. Ozone exposure increased insulin signaling in skeletal
muscle and liver, but not in adipose tissue (Ying et al.. 2016). In a separate study with the same
exposure paradigm, male KK mice, a diabetes-prone model, also showed decreased fasting
insulin and P-cell insulin secretory function (Zhong et al.. 2016). Additionally, ozone exposure
induced insulin resistance (Zhong et al.. 2016).
• Male Fischer 344 rats exposed to 0.8 ppm ozone for 4 hours had no change in insulin levels but
showed a decrease in plasma glucagon (Thomson et al.. 2016). When administered the
1 ^-hydroxylase inhibitor metyrapone, which blocks glucocorticoid synthesis, the ozone-induced
decrease in glucagon was prevented. A high dose of metyrapone also resulted in increased insulin
levels following ozone exposure. These results further support the well-characterized effects of
corticosterone that result in impaired glucose uptake and insulin resistance.
However, additional studies reported little to no effect when examining whether short-term ozone
exposure altered insulin homeostasis:
• Male Wistar Kyoto rats showed no impairment in peripheral insulin-mediated glucose clearance
and no differences in serum insulin levels after 5 hours of exposure to 0.25 or 1.0 ppm ozone
(Miller et al.. 2016b).
• Male and female brown Norway rats fed a normal, high-fat, or high-fructose diet for 12 weeks
before exposure to ozone or filtered air showed no effects from acute (5 hours) or subacute
(1 day/week for 4 weeks) exposure to 0.8 ppm ozone on serum insulin levels in any group
(Gordon et al.. 2016).
5.1.3.1.4 Summary
Recent evidence from the epidemiologic literature shows mostly null associations between
short-term ozone exposure and glucose and insulin homeostasis. However, one study reported
associations between ozone and increases in fasting glucose in healthy individuals and between ozone and
hyperglycemia and insulin resistance in individuals with type 2 diabetes. A controlled human exposure
study showed no evidence of insulin resistance in healthy volunteers exposed to 0.3 ppm ozone over
2 hours with intermittent exercise. In animal toxicological studies, effects of short-term ozone exposure
on fasting glucose levels were inconsistent across concentration levels and exposure durations, potentially
reflecting sensitivity in glucose levels to other variables, such as diet or strain. There was consistent
evidence, however, from animal toxicological literature showing that short-term ozone exposure results in
impaired glucose tolerance and that this effect is transient. Studies examining the relationship between
5-11
-------
short-term ozone exposure and insulin homeostasis report inconsistent effects on fasting insulin levels,
but suggest impaired insulin resistance, potentially mediated through alterations in (3-cell and a-cell
functioning in the pancreas. Finally, the effects from short-term ozone exposure on glucose and insulin
homeostasis may be mediated by adrenal hormones because removing the adrenal glands or blocking
glucocorticoid-dependent processes resulted in an amelioration of ozone-induced effects.
5.1.3.2 Elevated Triglycerides
High levels of triglycerides, one criterion for metabolic syndrome, have been linked to
cardiovascular disease. Triglycerides are stored in adipose tissue and are also present in the blood where
they bidirectionally transfer fat and glucose from the liver. Lipolysis is the pathway through which
triglycerides are broken down into glycerol and free fatty acids. Lipolysis is induced by several hormones,
including glucagon, adrenaline, and cortisol/corticosterone (Fruhbcck et al.. 2014V
5.1.3.2.1 Epidemiologic Studies
In the 2013 Ozone ISA, one epidemiologic study provided evidence of an association between
ozone exposure and altered blood lipids. Chuang et al. (2010) conducted a population-based
cross-sectional analysis of data collected on 7,578 participants during the Taiwanese Survey on
Prevalence of Hyperglycemia, Hyperlipidemia, and Hypertension in 2001. Levels of Apolipoprotein B
(ApoB), a lipid carrier that transports triglycerides and cholesterol around the body, was associated with
3-day avg ozone concentration. Increases in 5-day mean ozone concentration were associated with
increased levels of triglycerides. In addition, increases in the 1-, 3-, and 5-day mean ozone concentrations
were associated with increased HbAlc levels (a marker used to monitor the degree of control of glucose
metabolism). No association was observed between ozone concentration and ApoAl. Recent studies
support the findings that ozone exposure is associated with changes to serum lipids in animal and human
studies.
5.1.3.2.2 Controlled Human Exposure Studies
In the 2013 Ozone ISA, no controlled human exposure studies evaluated short-term ozone and
serum lipids. As indicated in Table 5-7. there is one study of healthy adult human volunteers (n = 24) who
exercised intermittently and were exposed separately to ozone and fresh air during two visits to the clinic
(2 hours, 0.3 ppm ozone or fresh air exposure with 15 minutes on/off exercise in a controlled chamber).
Ozone exposure was associated with increased serum levels of medium- and long-chain free fatty acids
and glycerol consistent with enhanced lipolysis (Miller et al.. 2016a). However, there were no observed
differences in triglyceride levels.
5-12
-------
5.1.3.2.3
Animal Toxicological Studies
Several animal toxicological studies suggest that short-term ozone exposure increases serum
triglyceride levels in male rodents. Fewer studies investigated changes in free fatty acids, with only one
showing ozone-induced increases.
• A 3-hour ozone exposure increased serum triglycerides in male SH rats [0.3 ppm ozone; Farrai et
al. (2016)1. One study suggests adrenal hormones may mediate ozone-induced changes in
triglycerides. In male Wistar Kyoto rats, a 1-day, but not 2-day, (4 hours/day) exposure to
1.0 ppm ozone increased serum triglycerides, an effect that was prevented by bilateral removal of
the adrenal medulla or entire adrenal gland (Miller etal.. 2016c). In contrast, no effects on
triglycerides were found following ozone exposure (0.25 and 1.0 ppm, 6 hour/day for 2 days) in
1-, 4-, 12-, and 24-month old male Norway rats that were examined 18 hours post-exposure (Bass
et al.. 2013). Similarly, there were no changes following a 2-day ozone exposure (0.8 ppm,
4 hours/day) in male or female Long-Evans rats (Gordon et al.. 2017a).
• Females may show decreased susceptibility to ozone-induced changes in triglycerides. In one
experiment, male and female brown Norway rats were fed a normal, high-fat or high-fructose diet
for 12 weeks before exposure to ozone or filtered air [0.8 ppm, 5 hours or 1 day/week for
4 weeks; Gordon et al. (2016)1. For serum triglycerides, a 5-hour ozone exposure increased levels
for males on both the normal and high-fat diets relative to control animals. The 5-hour ozone
exposure did not affect serum lipids in females. Intermittent ozone exposure for 4 weeks did not
result in any significant effects on triglycerides in either sex. In a second experiment, the effect of
exercise on ozone-dependent changes in serum lipids was examined in female Long-Evans rats. A
5-hour exposure to 0.25, 0.5, or 1.0 ppm ozone did not result in any changes in triglycerides in
either active or sedentary female rats (Gordon et al.. 2017b).
• In Wistar rats, neither plasma triglycerides nor nonesterified fatty acid levels were affected by
ozone exposure when assessed immediately post-exposure [0.8 ppm ozone for 16 hours; Vella et
al. (2014)1. Similarly, a separate study found no changes in free fatty acids following 1 or 2 days
of ozone exposure [1.0 ppm, 4/day; Miller et al. (2016c)l. However, in contrast, metabolomic
analysis showed that ozone exposure increased short- and long-chain free fatty acids [1.0 ppm,
6 hours/day for 1 or 2 days; Miller et al. (2015)1.
5.1.3.2.4 Summary
Limited evidence from the epidemiologic and controlled human exposure studies, as well as from
multiple animal toxicological studies, suggests that short-term ozone exposure increases levels of
triglycerides and/or free fatty acids. Results from animal toxicological experiments conducted in male rats
are further summarized in Table 5-8. In animal toxicological studies, diminishing adrenal hormones by
adrenalectomy ameliorates the ozone-induced increase in triglycerides. In the controlled human exposure
study, ozone exposure increased circulating levels of free fatty acids and glycerol, consistent with
increased lipolysis.
5-13
-------
5.1.3.3
Low High-Density Lipoprotein (HDL) Cholesterol
Low levels of high-density lipoprotein (HDL) have been associated with negative health
outcomes, including metabolic syndrome. HDL particles remove fat molecules from cells and transport
lipids, such as cholesterol, to the liver either for secretion or hormone synthesis (Yu et al.. 2019). In
contrast, low-density lipoprotein (LDL) particles deliver fat molecules to cells. LDL can also enter the
endothelium, become oxidized, and contribute to atherosclerosis (Yu et al.. 2019). Because direct
measurements of HDL and LDL are costly, blood tests are typically used to measure HDL-C or LDL-C,
the cholesterol carried within the associated particles.
5.1.3.3.1 Epidemiologic Studies
As noted above, Chuang et al. (2010) reported associations between altered serum lipids and
short-term ozone exposure. Since then, one epidemiologic study (Table 5-10) evaluated the effects of
short-term ozone exposure on blood lipids. Chen et al. (2016) analyzed the (3-gene cohort of
1,023 Mexican Americans living in southern California. The study considered LDL levels and
HDL-to-LDL ratios and used cumulative averages of daily concentrations from 0-90 days prior to testing.
Null results were reported for metabolic endpoints evaluated with short-term increases of ozone exposure.
The results were presented qualitatively.
5.1.3.3.2 Controlled Human Exposure Studies
In the 2013 Ozone ISA, no controlled human exposure studies evaluated short-term ozone and
serum lipids. As indicated in Table 5-7. there is one study of healthy adult human volunteers (n = 24) who
exercised intermittently and were exposed separately to ozone and fresh air during two visits to the clinic
(2 hours, 0.3 ppm ozone or fresh air exposure with repeated 15 minutes of exercise and 15 minutes of rest
in a controlled chamber). There were no observed differences in HDL-C, LDL-C, or total cholesterol.
5.1.3.3.3 Animal Toxicological Studies
Animal toxicological studies demonstrate that short-term ozone exposure can induce changes in
serum HDL-C, LDL-C, and total cholesterol, although the results are inconsistent. This inconsistency
may be a result of differences in time at assessment, strain, sex, and dose. Most of the studies mentioned
below use male rodents, and some evaluate these outcomes in rodents that are obese, diabetic, or have
cardiovascular disease. More specific information follows below, and detailed study design can be found
in the table that follows (Table 5-8).
5-14
-------
• In male Wistar Kyoto rats, serum total cholesterol, LDL, and HDL were elevated at 18 hours
post-ozone exposure [1.0 ppm, 6 hours/day for 2 days; Miller et al. (2015)1. For LDL only, this
increase was also significant immediately following 2 days of ozone exposure. There were no
changes after 1 day of exposure. In an additional study, the effect of ozone exposure (0.25 and
1.0 ppm, 6 hours/day for 2 days) on 1-, 4-, 12-, and 24-month-old male Norway rats was
examined 18 hours after exposure. In 12-month-old rats, 1.0 ppm ozone resulted in increased
HDL cholesterol levels. There were no other trends or significant effects from ozone exposure
(Bass et al.. 2013). A 4-hour ozone exposure at 0.8 ppm but not at 0.2 ppm resulted in decreased
HDL cholesterol in male SH rats (Farrai et al.. 2012). There were no changes in LDL cholesterol.
• Several studies reported no changes in LDL, HDL, or total cholesterol following short-term
ozone exposure. A 3-hour ozone exposure did not affect HDL or LDL cholesterol levels in male
SH rats [0.3 ppm ozone; Farrai et al. (2016)1. Further, in Wistar rats, plasma cholesterol levels
were not affected by ozone exposure when assessed immediately post-exposure [0.8 ppm ozone
for 16 hours; Vella et al. (2014)1. No changes were observed in total cholesterol in male Wistar
Kyoto rats exposed for 1 or 2 days (4 hours/day) to 1.0 ppm ozone (Miller et al.. 2016c). Finally,
there were no changes in HDL, LDL, or total cholesterol following a 2-day ozone exposure
(0.8 ppm, 4 hours/day) in male or female Long-Evans rats (Gordon et al.. 2017a).
• The effects of short-term ozone exposure on serum lipid profiles may continue after exposure
cessation. Immediately following exposure and after a period of recovery, the effect of a 4-hour
exposure to 0.25, 0.5, and 1.0 ppm ozone on eight different rat strains was investigated (Ramot et
al.. 2015). The healthy strains examined included Wistar Kyoto, Wistar, and Sprague-Dawley
rats. Models of cardiovascular disease/metabolic disorder examined included spontaneously
hypertensive (SH), fawn-hooded hypertensive (FHH), SH stroke prone (SHSP), SH heart failure
(SHHF), and JCR rats. Immediately following ozone exposure, there were no changes in total
cholesterol in any strain. However, in FHH rats, HDL levels decreased (1.0 ppm) and in SH,
FHH, and SHSP rats, LDL levels decreased (0.5 ppm) immediately following exposure. After a
20-hour recovery period, total serum cholesterol levels decreased in FHH (0.5 ppm) and JCR
(0.5, 1.0 ppm) rats and increased in SH rats. In addition, HDL levels decreased in JCR (0.5,
1.0 ppm) and in FHH (0.25, 0.5, 1.0 ppm) rats. Moreover, after the recovery period, short-term
ozone exposure increased LDL levels in Wistar Kyoto rats (1.0 ppm), but decreased LDL levels
in Wistar (1.0 ppm) and SHSP (0.25 ppm) rats. This study shows that short-term ozone exposure
induces changes in serum cholesterol in a strain-specific manner, and further, that changes in
cholesterol continue after the cessation of exposure, particularly for healthy rat strains.
• Sex may be a mediating factor for ozone-induced changes in serum lipids, with females showing
decreased susceptibility. In one experiment, male and female brown Norway rats were fed a
normal, high-fat, or high-fructose diet for 12 weeks before exposure to ozone or filtered air
[0.8 ppm, 5 hours or 1 day/week for 4 weeks; Gordon et al. (2016)1. For males on the normal diet,
5-hour ozone exposure resulted in a significant reduction in serum cholesterol. In contrast, males
on the high-fat diet showed a significant increase in serum cholesterol following the 5-hour ozone
exposure. The 5-hour ozone exposure did not affect serum lipids in females. Intermittent ozone
exposure for 4 weeks did not result in any significant effects on cholesterol in either sex. In a
second experiment, the effect of exercise on ozone-dependent changes in serum lipids was
examined in female Long-Evans rats. A 5-hour exposure to 0.25, 0.5, or 1.0 ppm ozone did not
result in any changes in cholesterol in either active or sedentary female rats (Gordon et al..
2017b).
5-15
-------
5.1.3.3.4 Summary
Neither the epidemiologic study nor the controlled human exposure study reported changes in
HDL-C, LDL-C, or total cholesterol from short-term ozone exposure. However, while there were some
inconsistencies, there is evidence from animal toxicological studies that short-term ozone exposure alters
levels of HDL-C, LDL-C, and total cholesterol. Several studies reported null results for these endpoints,
while others showed differences in the direction of effect. Results from experiments conducted in male
rats are further summarized in Table 5-2. In general, short-term ozone exposure resulted in increased
HDL and LDL in healthy male rats and decreased HDL and LDL in disease-prone rats. These findings are
not supported by the available experimental human study and epidemiologic study; thus, uncertainties
remain regarding the role of ozone exposure and these endpoints.
5-16
-------
Table 5-2 Summary of changes in serum lipids from experiments conducted in
male rats.
Study
Species
(Stock/
Strain),
Age
Exposure
Details
(Concentration,
Duration)
Time at
Measure-
ment
Free
Fatty
Triglycerides3 Acids3
HDL3
LDL3
Total
Cholesterol3
Miller et
al.
(2015)
Rats
(WKY)
Age:
10 weeks
1.0 ppm, 6 h
Immediately
PE
*
NS
NS
NS
1.0 ppm, 6 h/day
for 2 days
Immediately
PE
*
NS
*
NS
1.0 ppm, 6 h/day
for 2 days
18 h PE
-
*
*
*
Bass et
al.
(2013)
Rats
(BN)
Age: 1, 4,
12, and
24 mo
1.0 ppm, 2 days 18 h PE
*
(12 mo
old rats
only)
NS
NS
Farrai et
Rats 0.8 ppm, 4 h
24 h PE
*
NS
NS
al.
(SH)
(2012)
Age:
12 weeks
Farrai et
Rats
0.8 ppm, 3 h
24 h PE
*
NS
NS
NS
al.
(SH)
(2016)
Age:
12 weeks
Vella et
Rats
0.8 ppm, 16 h
Immediately
NS
NS
_
NS
al.
(Wistar)
PE
(2014)
Age:
adult
(400-
450 g)
Ramot
WKY,
0.5-1.0 ppm, 4 h
Immediately
_
sl/FHH
sl/SH,
NS
et al.
Wistar,
PE
FHH,
(2015)
S-D, SH,
SHSP
FHH,
SHSP,
1.0 ppm, 4 h
20 h PE
-
skJCR,
^WKY>
sl/FHH,
SHHF,
FHH
WIS
JCR.^SH
JCR
Gordon
Rats
0.8 ppm, 5 h
18 h PE
*
NS
NS
*
et al.
(BN)
(2016)
Age:
30 days
Rats 1.0 ppm, 4 h Immediately
(WKY) PE
NS
NS
5-17
-------
Table 5-2 (Continued): Summary of changes in serum lipids from experiments
conducted in male rats.
Study
Species
(Stock/
Strain),
Age
Exposure
Details
(Concentration,
Duration)
Time at
Measure-
ment
Triglycerides3
Free
Fatty
Acids3 HDL3
Total
LDL3 Cholesterol3
Miller et
al.
(2016c)
Age:
adult
1.0 ppm, 4 h/day
for 2 days
Immediately
PE
NS
NS
NS
Gordon
et al.
(2017a)
Rats (LE)
0.8 ppm, 4 h/day
for 2 days
Immediately
PE
NS
NS NS
NS NS
BN = brown Norway; FHH = fawn-hooded hypertensive; LE = Long-Evans; NS = not significant; PE = post-exposure; ppm = parts
per million; S-D = Sprague-Dawley; SH = spontaneously hypertensive; SHHF = spontaneously hypertensive heart failure;
SHSP = spontaneously hypertensive stroke prone; WKY = Wistar Kyoto.
aArrows indicate statistical significance at a <0.05; significance levels as reported in original studies.
5.1.3.4 Central Adiposity
Central adiposity or enlarged waist circumference is an additional criterion of metabolic
syndrome. While centrally accumulated body fat is associated with insulin resistance, peripherally
distributed body fat is less relevant to metabolism. Animal toxicological studies examine changes in
weight gain and body composition following short-term ozone exposure, which can be mediated through
dysregulated feed and water intake. The hormones leptin and ghrelin indirectly mediate glucose and
insulin homeostasis (Meier and Gressner. 2004V Leptin is secreted by adipocytes and regulates food
intake. High levels of leptin suppress feeding, while low levels increase food intake (Meier and Gressner.
2004). Leptin resistance can lead to hyperphagia and weight gain (Meier and Gressner. 2004). Ghrelin
also regulates food intake and is produced by enteroendocrine cells of the gastrointestinal tract (Alamri et
al.. 2016). Ghrelin levels are highest before food intake and fall strongly in response to carbohydrates and
protein. In addition, both leptin and ghrelin modulate insulin secretion and signaling (Alamri et al.. 2016).
Also involved in regulating energy homeostasis are the hormones adiponectin and resistin, both of which
have been linked to obesity (Meier and Gressner. 2004).
5.1.3.4.1 Epidemiologic Studies
No epidemiologic studies in the 2013 Ozone ISA assessed the association between short-term
ozone exposure and central adiposity. As noted in Table 5-9. one recent study evaluated the association
between short-term ozone exposure and the hormones adiponectin, leptin, and resistance. These hormones
are produced by adipose tissue and have been linked to obesity (Meier and Gressner. 2004). Using
5,958 participants from the Framingham Offspring Cohort and Third Generation Cohort, Li et al. (2017)
completed a panel study evaluating associations of glucose, insulin, HOMA-IR, adiponectin, leptin, and
5-18
-------
resistin with 1- to 7-day moving avg of ozone concentration. Based on the published qualitative results,
there were positive associations between adiponectin and short-term increases in ozone concentration and
there were negative associations between glucose, insulin, HOMA-IR, and resistin and short-term
increases in ozone concentration.
5.1.3.4.2 Controlled Human Exposure Studies
Controlled human exposure studies of the effects of short-term ozone exposure on central
adiposity were not available for review in the 2013 Ozone ISA. In a study of 24 healthy volunteers aged
22-30 years, Miller et al. (2016a) randomly exposed subjects to ozone (0.3 ppm) or filtered air for
2 hours while alternating 15 minutes of exercise and 15 minutes of rest (Table 5-7). After a 2-week
washout period, the subjects had the alternate exposure. There was no observed change in serum leptin
immediately following ozone exposure when compared to filtered air exposure.
5.1.3.4.3 Animal Toxicological Studies
The 2013 Ozone ISA did not include studies that examined the effects of ozone exposure on
endpoints relevant to central adiposity (Section 5.1.7.2). Recent toxicological studies provide some
evidence that ozone exposure affects food and water intake, body composition, and body weight gain,
potentially through changes in hormones that regulate food intake, such as leptin and ghrelin. More
detailed information on these studies is contained in the evidence inventory (Table 5-8).
Animal toxicological studies demonstrate that short-term ozone exposure affected food and water
intake, body weight, body composition, and blood levels of leptin and ghrelin:
• Male and female brown Norway rats were fed a normal, high-fat, or high-fructose diet for
12 weeks before exposure to ozone or filtered air [0.8 ppm ozone, 1 day/week for 4 weeks;
Gordon et al. (2016)1. Ozone exposure resulted in significantly more food consumed for males on
the control and high-fat diets, while significantly more calories were consumed by males only on
the high-fat diet. Ozone exposure resulted in an increase in water consumption for males on the
control and high-fructose diets. In females, food and water intake were not affected by ozone
exposure. For body composition, ozone exposure resulted in a decrease in percentage of fat and
an increase in percentage of lean for males on the control diet, and a decrease in percentage of
fluid for males on the high-fructose diet. For females on the high-fructose diet, ozone exposure
resulted in a decrease in percentage of lean. There were no other statistically significant changes
in percentage of fat, lean, or fluid (Gordon et al.. 2016).
• In spontaneously hypertensive rats, body weight was lower following a 2-day but not 1-day
exposure to ozone (0.8 ppm, 4 hours/day), likely reflecting loss in body fluid (Henriquez et al..
2018). In male KKAy mice, a model for genetic diabetes, ozone exposure (0.5 ppm, 4 hours/day
for 13 consecutive weekdays) reduced body-weight gain (Ying et al.. 2016). In contrast, male
Sprague-Dawley rats fed a normal or high-fructose diet showed no changes in body weight
following ozone exposure [0.5 ppm, 8 hours/day for 9 days over 2 weeks; Sun et al. (2013)1.
5-19
-------
• In healthy male rats, leptin levels increased after 6 hours of 1.0 ppm ozone exposure (Miller et al..
2015; Bass et al.. 2013; Sun et al.. 2013). In both male KKAy and KK mice, models for diabetes,
ozone exposure (0.5 ppm, 4 hours/day for 13 consecutive weekdays) decreased leptin (Ying et al..
2016; Zhong et al.. 2016). Further, male Fischer 344 rats exposed to 0.8 ppm ozone for 4 hours
showed a decrease in ghrelin levels (Thomson et al.. 2018; Thomson et al.. 2016). When
administered the 1 l(3-hydroxylase inhibitor metyrapone, which blocks glucocorticoid synthesis,
the ozone-induced decrease in ghrelin was prevented.
5.1.3.4.4 Summary
In summary, a limited number of studies have examined the effects of short-term ozone exposure
on endpoints relevant to central adiposity. Neither the epidemiologic nor the controlled human exposure
study reported significant changes in levels of obesity-relevant hormones. Animal toxicological studies
suggest that short-term ozone exposure may reduce body-weight gain, though results were inconsistent. In
addition, in animal toxicological studies, ozone exposure consistently affected serum leptin, with the
direction of effect differing between healthy and diabetic animal models (notably, the duration of
exposure also differed between studies). One animal study suggests that the ozone-induced decrease in
ghrelin may be a result of corticosterone signaling. See Section 5.1.7.2 for the role of obesity in
susceptibility to short-term ozone exposure.
5.1.3.5 Elevated Blood Pressure
Short-term ozone exposure-mediated effects on blood pressure are discussed in detail in the
Cardiovascular Appendix [see Appendix 4; Akcilar et al. (2015); Wagner et al. (2014); Gordon et al.
(2013); Uchivama and Yokovama (1989); Uchivama et al. (1986)1. Hypertension is a clinically relevant
consequence of chronically high blood pressure, which typically develops over years. High blood
pressure, dyslipidemia, increased fasting blood glucose, and obesity are criteria for metabolic syndrome,
which is a risk factor for heart disease, stroke, and diabetes (Alberti et al.. 2009).
Recent epidemiologic and controlled human exposure studies of ozone and blood pressure are
limited in number and generally inconsistent. Animal toxicological studies show some evidence to
suggest that short-term exposure to ozone can result in changes in blood pressure in animals. However,
some results also suggest that changes in diet may mediate these effects (see Section 4.1.8).
5.1.3.6 Summary of Metabolic Syndrome
Metabolic syndrome in humans is defined by the presence of at least three of the following five
risk factors: hyperglycemia (elevated blood glucose), elevated triglycerides, low HDL cholesterol, high
blood pressure, and central adiposity [enlarged waist circumference; Alberti et al. (2009)1. The presence
5-20
-------
of these risk factors may predispose individuals to an increased risk of type 2 diabetes and cardiovascular
disease. Evidence in rodent models that short-term ozone exposure may affect individual risk factors for
metabolic syndrome suggests that exposure to ozone has the potential to contribute to or exacerbate
metabolic syndrome in humans.
Most available animal toxicological evidence shows short-term ozone exposure results in
hyperglycemia (Section 5.1.3.1.3) and elevated triglycerides (Section 5.1.3.2.3). with inconsistencies
between studies potentially arising from differences in rodent strain, sex, or diet. In the controlled human
exposure study, while circulating free fatty acids increased, there were no significant changes in
triglycerides following short-term ozone exposure (Miller et al.. 2016a). Some epidemiologic studies
examining changes in glucose and lipids provide support for effects associated with short-term ozone
exposure, reporting, for example, a positive association between the 5-day mean ozone concentration and
fasting glucose and triglyceride levels (Kim and Hong. 2012).
There is less consistent support for short-term ozone exposure-induced low HDL cholesterol,
high blood pressure, and central adiposity, with changes in animal models often not supported by human
evidence. In general, in animal toxicological studies, short-term ozone exposure resulted in a reduction in
HDL cholesterol in disease-prone rat strains, but not in healthy rodent models or in females
(Section 5.1.3.3.3). For blood pressure, there is some animal toxicological evidence that short-term ozone
exposure results in effects (Section 5.1.3.5). For central adiposity, studies were reviewed that examined
changes in obesity-related hormones following short-term ozone exposure. In animal toxicological
studies, ozone exposure consistently affected serum leptin, a weight-regulating hormone, with the
direction of the effect differing between healthy and diabetic animal models (Section 5.1.3.4.3).
potentially due to leptin resistance in obese or diabetic individuals (Myers et al.. 2008). In the controlled
human exposure study, no changes in HDL cholesterol, blood pressure, or levels of obesity-relevant
hormones were observed (Miller et al.. 2016a). Finally, the epidemiologic evidence provides some
support for changes in obesity-relevant hormones associated with short-term ozone exposure (Li et al..
2017). but little support for changes in HDL cholesterol (Section 5.1.3.3.1) or blood pressure
(Section 5.1.3.5).
5.1.4 Complications from Diabetes
Diabetic ketoacidosis, which is usually seen in individuals with type 1 diabetes, can result in
unconsciousness from a combination of a severely increased blood sugar level, dehydration, and
accumulation of ketones or acids that were formed as the diabetic body used fat for fuel instead of sugar.
Diabetic coma is a reversible form of coma found in people with diabetes, which involves extremely low
blood sugar. Dales et al. (2012) evaluated associations of short-term ozone exposure and hospital
admissions for diabetic ketoacidosis and diabetic coma in the Santiago region of Chile. Using a 6-day
distributed lag, a null association was observed for the relationships of hospital admissions for diabetic
5-21
-------
ketoacidosis or diabetic coma (1.02; 95% CI: 1.00, 1.04). However, the effect increased in populations
aged 75-84 years (1.08; 95% CI: 1.01, 1.15) and over 85 years (1.08; 95% CI: 1.01, 1.1). While increases
were noted in the higher age brackets, the risks were not higher in other age groups (<64 or 65-74 years).
5.1.5 Other Indicators of Metabolic Function
5.1.5.1 Adipose Tissue Inflammation
Inflammation has been implicated in the development of type 2 diabetes and atherosclerosis
leading to coronary heart disease (Rav et al.. 2009). The role of systemic inflammation in ozone exposure
is covered in Section 4.1.11. of the Cardiovascular Effects Appendix. As outlined in Section 5.1.3.
systemic inflammation may promote a peripheral inflammatory response in organs and tissues, including
adipose tissue. New evidence for peripheral inflammation in adipose tissue following short-term ozone
exposure is presented below.
5.1.5.1.1 Animal Toxicological Studies
A limited number of animal toxicological studies provide evidence that short-term exposure to
ozone may result in inflammation of the visceral or perirenal adipose tissue. Specific examples are
detailed below and in the evidence inventory tables that follow (Table 5-10).
• In the visceral adipose tissue of male KK mice, a diabetes-prone model, ozone exposure
(0.5 ppm, 4 hours/day for 13 consecutive weekdays) promoted activation of macrophages and
CD4+T cells, increased the expression of the chemokine CXCL-11 (no change in CCL-5,
CSCL-9, CXCL-12, or MCP-1), and increased the expression of inflammatory genes IFN-y,
IL-12, iNOS, TNF-a, and CD-56 [no change in IL-6, RORC, or TBX21; Zhong etal. (2016)1.
• Inflammatory and oxidative stress biomarkers (TNF-a, MCP-1) were upregulated and
anti-inflammatory genes were downregulated (IL-10) in epicardial and perirenal adipose tissue in
rats (8-week-old male Sprague-Dawley rats were fed a normal high-fructose diet for 8 weeks)
exposed to ozone [0.5 ppm, 8 hours/day, 5 days/week, for 9 days over 2 weeks; Sun et al. (2013)1.
There was significantly increased infiltration of macrophages that was associated with increased
expression of TNF-a and iNOS.
• Exposure to ozone increased proinflammatory macrophages in adipose tissue of a diabetic mouse
model [male KKAy mice, 0.5 ppm ozone for 13 consecutive weekdays; Ying et al. (2016)1.
5-22
-------
5.1.5.2 Liver Biomarkers
The liver, which is between the portal and systemic circulation, is the site for primary energy and
xenobiotic metabolism (Boron and Boulpaep. 2017). The liver is a crucial organ for maintaining glucose
homeostasis. It can be stimulated to increase blood glucose by inducing gluconeogenesis (production of
glucose from noncarbohydrates) during fasting or to store glucose after feeding. With intense
gluconeogenesis, the liver produces ketone bodies that are released together with the newly produced
glucose. The liver can also synthesize and degrade proteins, carbohydrates, and lipids for distribution to
extrahepatic tissues depending on energy needs. Finally, the liver regulates whole-body cholesterol
balance via biliary excretion of cholesterol, cholesterol conversion to bile acids, and by regulating
cholesterol synthesis (Boron and Boulpaep. 2017). Consequently, the liver is an essential regulator of
whole-body metabolism and energy homeostasis.
Acute-phase liver proteins are discussed in more detail in Appendix 4. Section 4.1.11. An
epidemiologic study found associations between ozone exposure and C-reactive protein, an acute-phase
liver protein that serves as a marker of systemic inflammation. These proteins, in combination with other
liver enzymes, can provide information about overall health.
5.1.5.2.1 Controlled Human Exposure Studies
No controlled human exposure studies of liver biomarkers were evaluated in the 2013 Ozone
ISA. One liver outcome measured in humans is ketone body formation; more information is available in
the evidence inventory (Table 5-11). In one recent controlled human exposure study, healthy adult human
volunteers exercised intermittently and were exposed separately to ozone and fresh air during two visits to
the clinic (2 hours at 0.3 ppm ozone or fresh air exposure with 15 minute on/off exercise in a controlled
chamber). Ozone exposure caused increased carnitine conjugates of long-chain FFA and acetyl carnitine,
both suggestive of accelerated P-oxidation and increased ketone body generation (Miller et al.. 2016a).
5.1.5.2.2 Animal Toxicological Studies
Ozone exposure in animal models affects various pathways that are mediated through the liver,
including increasing hepatic glucose production through gluconeogenesis (generation of glucose from
noncarbohydrates), decreasing bile acid production, impairing glycolytic pathways, altering P-oxidation,
and altering expression of hepatic metabolism-related genes in the liver. More detailed information on
how ozone exposure affects metabolic outcomes in the liver follows below and in Table 5-10. but like
other pathways, the liver contributes to increased blood glucose with ozone exposure.
• To determine whether ozone-dependent hyperglycemia in male Wistar Kyoto rats was controlled
by hepatic gluconeogenesis, Miller et al. (2016b) performed a pyruvate tolerance test in which
pyruvate (a substance for gluconeogenesis) was injected and blood glucose measured over time.
5-23
-------
The test showed increased blood glucose levels following 1 week of ozone exposure (1.0 ppm but
not 0.25 ppm; 5 hours/day) compared with filtered air controls, confirming the stimulation of
gluconeogenesis with ozone exposure.
• To further delineate the role of the liver in ozone-induced metabolic impairment, global gene
expression was examined in male Wistar Kyoto rats following short-term ozone exposure
[1.0 ppm, 6 hours/day; Miller et al. (2015)1. Differential gene expression in the liver was highest
after 1 day of exposure, decreasing dramatically after 2 days of exposure and after an 18-hour
post-exposure recovery period. Ozone exposure significantly altered genes involved in steroid
biosynthesis, tricarboxylic acid cycle, and glyoxylate and decarboxylate metabolism pathways.
Ozone also affected genes involved in glycolysis and gluconeogenesis. In addition, ozone
exposure altered bile acid profiles (Miller et al.. 2015V Similarly, short-term exposure to ozone
(8 hours/day for 5 days) in male Sprague-Dawley rats was associated with altered expression of
certain proteins in the liver that can modulate hepatic metabolic function, including
glucose-regulated protein 78 or GRP-78 (post-translationally modified GRP-78 is a novel
autoantigen in human type 1 diabetes), protein disulfide isomerase, and glutathione S-transferase
Ml (Theis et al.. 2014).
5.1.5.2.3 Summary
Multiple metabolic indicators from the liver provide evidence that ozone exposure induces
hepatic changes affecting glucose homeostasis. Healthy volunteers who exercised with ozone exposure in
controlled human exposure studies had increased ketone body formation. In animal toxicological studies,
ozone exposure induced changes to the liver including hepatic gluconeogenesis, altered bile acid profile,
alterations to (3-oxidation, and alterations to proteins in hepatic metabolic pathways.
5.1.5.3 Adrenal Hormones
Activation of the autonomic nervous system can trigger the sympathetic-adrenal-medullary
(SAM) and hypothalamic-pituitary-adrenal (HPA) axes to release adrenaline and cortisol/corticosterone,
respectively, from the adrenal glands fNicolaides et al.. 2015). These hormones increase circulating levels
of blood glucose and fatty acids and have widespread effects on metabolic processes (see Section 5.1.2
for details).
5.1.5.3.1 Controlled Human Exposure Studies
No controlled human exposure studies of short-term ozone and endocrine hormones were
reviewed in the 2013 Ozone ISA. One recent study (Table 5-11) used healthy adult human volunteers
who experienced intermittent exercise and were exposed separately to ozone and fresh air during two
visits to the clinic (2 hours, 0.3 ppm ozone or fresh air exposure with 15 minutes of on/off exercise in a
controlled chamber). Acute ozone exposure increased circulating stress hormones (Cortisol and
corticosterone) in these volunteers (Miller et al.. 2016a).
5-24
-------
5.1.5.3.2
Animal Toxicological Studies
Ozone exposure activates both the SAM and HPA axes, resulting in the release of adrenaline
(epinephrine) and cortisol/corticosterone, respectively. Specific details of these studies are included in the
evidence inventory (Table 5-10).
• Ozone exposure activates the HPA axis and triggers glucocorticoid release from the adrenal
cortex. In male Fischer 344 rats, a 4-hour exposure to 0.8 ppm ozone transiently increased plasma
levels of adrenocorticotropic hormone (ACTH) and corticosterone (Thomson et al.. 2013). This
ozone-induced increase in plasma corticosterone was prevented by administration of metyrapone,
an inhibitor of glucocorticoid synthesis (Thomson et al.. 2016). However, in male Wistar Kyoto
rats, a 1- or 2-day (4 hours/day) exposure to 1.0 ppm ozone did not significantly alter
corticosterone levels compared with filtered air exposure (Miller etal.. 2016c). Similarly, there
were no changes in corticosterone following a 2-day ozone exposure (0.8 ppm, 4 hours/day) in
male or female Long-Evans rats (Gordon et al.. 2017a).
• Ozone exposure activates the SAM axis and triggers the release of adrenaline from the adrenal
medulla. Adrenaline levels transiently increased in male Wistar Kyoto rats following exposure for
1 or 2 days (6 hours/day) to 1.0 ppm ozone (Miller et al.. 2015). Similarly, in male brown
Norway rats, a 2-day exposure of 1.0 ppm ozone increased adrenaline levels, which remained
elevated 18 hours post-exposure (Bass et al.. 2013). In male Wistar Kyoto rats, a 1- or 2-day
(4 hours/day) exposure to 1.0 ppm ozone increased adrenaline, but not noradrenaline (Miller et
al.. 2016c). In this study, bilateral removal of the adrenal medulla or of the entire adrenal gland
prevented the ozone-induced increase in adrenaline (Miller etal.. 2016c). However, in male
Fischer 344 rats, a 4-hour exposure to 0.8 ppm ozone did not affect plasma adrenaline levels
(Thomson et al.. 2016). potentially due to the hyperresponsive HPA axis in this strain. Similarly,
there were no changes in adrenaline or noradrenaline following a 2-day ozone exposure (0.8 ppm,
4 hours/day) in male or female Long-Evans rats (Gordon et al.. 2017a).
5.1.5.3.3 Summary
Following short-term ozone exposure, elevated circulating stress hormones were consistently
observed in animal models and in a single controlled human exposure study. Removal of the adrenal
glands prevented the release of adrenaline and corticosterone, and further, prevented ozone-induced
metabolic effects. Thus, neuroendocrine stress activation may be a primary mechanism through which
adverse metabolic outcomes develop from short-term ozone exposure.
5.1.6 Potential Copollutant Confounding of the Ozone-Metabolic Effects
Relationship
Few studies have examined potential for copollutants (i.e., PM2 5 or PMi0, or gaseous
copollutants) to confound the association between short-term ozone exposure and metabolic effects.
When positive associations between short-term ozone exposure and metabolic effects were observed in
single-pollutant models the association with ozone remained in copollutant models. Similarly, when null
5-25
-------
associations were observed in single pollutant models, the null effect also remained in copollutant models.
This suggests that the associations of ozone exposure with metabolic effects is not substantially affected
by copollutant confounding.
• Kim and Hong (2012) analyzed the KEEP cohort consisting of 560 Koreans over 60 years old and
observed increases in fasting glucose (0.19%; 95% CI: 0.09, 0.28%). Copollutant models ofNC>2
and PMio were also evaluated. The associations with glucose remained after adjustment for NO2
(0.16%; 95% CI: 0.06, 0.25%) and PM10 (0.15%; 95% CI: 0.01, 0.14%).
• One study evaluated hospital admissions for diabetic ketoacidosis and diabetic coma in the
Santiago region of Chile (Dales et al.. 2012). Insulin resistance may lead to hospitalizations for
these endpoints in individuals with both type 1 and type 2 diabetes. Using a 6-day distributed lag,
the study authors observed a null association for the relationships of hospital admissions for
diabetic ketoacidosis or diabetic coma (RR: 1.02; 95% CI: 0.996, 1.04). The effect remained null
when divided into subregions of Santiago and when CO, PM10, PM2 5, or SO2 were individually
added to evaluate two-pollutant models.
5.1.7 Effect Modification of the Ozone-Metabolic Effects Relationship
5.1.7.1 Lifestage
A limited number of recent studies of short-term ozone exposure and metabolic effects compared
associations between different age groups. One epidemiologic study reported inconsistent evidence as to
whether older adults are at increased risk for metabolic effects; however, one animal toxicological study
did demonstrate greater effects in aged animals.
• Dales et al. (2012) evaluated associations of short-term ozone exposure and hospital admissions
for diabetic ketoacidosis and diabetic coma in the Santiago region of Chile. Using a 6-day
distributed lag, they observed a null association for the relationships of hospital admissions for
diabetic ketoacidosis or diabetic coma (RR: 1.02; 95% CI: 1.00, 1.04). However, the effect
increased in populations aged 75-84 years (RR: 1.08; 95% CI: 1.01, 1.15) and over 85 years
(RR: 1.08; 95% CI: 1.01, 1.1). While increases were noted in the higher age brackets, the risks
were not higher in other age groups (<64 or 65-74 years).
• Gordon et al. (2013) compared young brown Norway rats (4 months of age) to aged or senescent
rats (20 months of age). With ozone exposure, both age groups had significant metabolic
responses, including increased triglycerides and serum insulin, but the response was greater in the
aged animals. Ozone induces glucose intolerance in young and aged brown Norway rats (Bass et
al.. 2013). Ozone-induced glucose intolerance increased in an age-dependent manner in rats with
the greatest impairment in glucose clearance observed in the oldest animals (ages 1,4, 12, and
24 months).
5-26
-------
5.1.7.2
Pre-existing Disease
A limited number of recent studies provides some evidence that individuals with pre-existing
diseases may be at greater risk of metabolic health effects associated with short-term ozone exposure.
These studies focus on specific diseases of varying severity (e.g., previous CVD events, type 2 diabetes).
Specifically:
• Kim and Hong (2012) found increases in fasting glucose (0.19%; 95% CI: 0.09, 0.28%), insulin
(0.71%; 95% CI: 0.02, 1.38%), and HOMA (0.30%; 95% CI: 0.06, 0.53%) in the Korean Elderly
Environmental Panel (KEEP). The KEEP cohort consisted of 560 Koreans over 60 years old. The
association of increases in short-term ozone concentration with increases in glucose, insulin, and
HOMA-IR levels was approximately threefold larger in people with a previous diagnosis of
type 2 diabetes (glucose [0.68%; 95% CI: 0.28, 1.07%], insulin [2.76%; 95% CI: 0.78, 4.75%],
and HOMA [1.21%; 95% CI: 0.44, 1.99%]) compared to those without type 2 diabetes.
• Dales et al. (2012) evaluated associations of short-term ozone exposure and hospital admissions
for diabetic ketoacidosis and diabetic coma in the Santiago region of Chile. Using a 6-day
distributed lag, they observed a null association for the relationships of hospital admissions for
diabetic ketoacidosis or diabetic coma (RR: 1.02; 95% CI: 1.00, 1.04). However, the effect
increased in populations aged 75-84 years (RR: 1.08; 95% CI: 1.01, 1.15) and over 85 years
(RR: 1.08; 95% CI: 1.01, 1.1). While increases were noted in the higher age brackets, the risks
were not higher in other age groups (<64 or 65-74 years).
• An animal toxicological study using eight different rat strains showed that disease-prone rats had
different responses to short-term ozone exposure than healthy rats for levels of HDL, LDL, and
total cholesterol (Ramot et al.. 2015).
5.1.8 Summary and Causality Determination
There were no causality conclusions for metabolic effects in the 2013 Ozone ISA (U.S. EPA.
2013a). The evidence pertaining to outcomes from short-term exposure to ozone and metabolic effects
has expanded substantially since the 2013 Ozone ISA (U.S. EPA. 2013a'). with multiple epidemiologic
and animal toxicological studies and one controlled human exposure study currently available for review
(Table 5-3).
Recent studies show that short-term ozone exposure triggers a stress response, with increased
levels of adrenaline and corticosterone in rodent models (Section 5.1.5.3.2) and increased levels of
Cortisol and corticosterone in humans (Miller et al.. 2016a). Changes in these hormones can result in a
cascade of transient metabolic effects which function to mobilize stored energy. Some of these metabolic
effects overlap with those associated with metabolic syndrome in humans, which is characterized by the
presence of three of the following five risk factors: hyperglycemia (elevated glucose levels), elevated
triglycerides, low HDL, high blood pressure, and an enlarged waist circumference.
The strongest and most consistent evidence for metabolic effects is from animal toxicological
studies demonstrating that short-term ozone exposure impairs glucose and insulin homeostasis
5-27
-------
(Section 5.1.3) and increases levels of triglycerides and free fatty acids (Section 5.1.3.2). Coherent with
results in animal models, the controlled human exposure study reported increases in medium- and
long-chain circulating free fatty acids following short-term ozone exposure. However, this study did not
find ozone-induced changes in serum insulin, nonfasting glucose, insulin resistance, or triglyceride levels
(Miller et al.. 2016a). Some epidemiologic studies examining changes in glucose and lipids provide
support for effects associated with short-term ozone exposure. A study in Korea of the KEEP cohort
reported a positive association between the 5-day avg ozone concentration and glucose levels, and for
those with pre-existing type 2 diabetes, a threefold higher association for glucose, insulin, and insulin
resistance (Kim and Hong. 2012). In a separate epidemiologic study, apolipoprotein B, a lipid carrier
which transports triglycerides and cholesterol around the body, was positively associated with the 3-day
avg ozone concentration (Chuang et al.. 2010). In addition, the 5-day mean ozone concentration was
associated with increased fasting glucose and triglyceride levels.
Adding to the weight of evidence for ozone-induced metabolic effects that are consistent with
metabolic syndrome, there are generally consistent results demonstrating that short-term ozone exposure
affects obesity-relevant endpoints. For example, some, but not all, animal toxicological studies reported
that short-term ozone exposure reduces body-weight gain1 in rodent models of diabetes, obesity, and of
spontaneous hypertension (Section 5.1.5). In addition, multiple animal toxicological studies from
different laboratories consistently reported that short-term ozone exposure affected levels of leptin, a
hormone that regulates food intake and is often affected in diabetes (Meier and Gressner. 2004). Coherent
with these results, an epidemiologic study reported trends for an association between short-term ozone
exposure and changes in the obesity-related hormones adiponectin and resistin. Furthermore, there is
some evidence that short-term ozone exposure can affect levels of HDL, LDL, and total cholesterol
(Section 5.1.3.2). as well as adipose tissue inflammation (Section 5.1.5.1). and blood pressure
(Section 5.1.3.5). However, the controlled human exposure study did not find ozone-induced changes in
serum leptin, HDL, LDL, total cholesterol, or blood pressure (Miller et al.. 2016a). Similarly, as noted in
the Cardiovascular Disease Appendix, there is little evidence in humans for changes in blood pressure
following short-term ozone exposure (Section 4.1.8).
There is some evidence from a limited number of animal studies that short-term ozone exposure
alters liver function involved in metabolism (Section 5.1.5.2). Additional animal toxicological studies of
liver biomarkers reported differential gene expression and protein levels related to glycolysis and
gluconeogenesis in the liver following ozone exposure, results consistent with glucose and free fatty acid
liberation after ozone exposure.
The observed ozone-induced metabolic responses described above are consistent with the
induction of a short-term stress response. Importantly, environmental triggers that induce a stress
response have been shown to increase susceptibility toward development of or complications related to
1 Reductions or increases in body-weight gain can indicate altered metabolic function in animal models of disease,
such as those used in these studies.
5-28
-------
metabolic disorders, such as diabetes (Pasquali. 2012). Indeed, a recent epidemiologic study that
evaluated associations between short-term ozone exposure and hospital admissions for diabetic
ketoacidosis and diabetic coma in the Santiago region of Chile reported an ozone effect in higher age
brackets (75-84, 85+ years) when using a 6-day distributed lag (Section 5.1.7.1). This suggests that ozone
exposure may exacerbate existing diabetes, leading to diabetic ketoacidosis and diabetic coma, and
further, that older adults may be at increased risk. An additional effect modifier of short-term ozone
exposure may be sex. Males and females can have different basal levels of the adrenal hormones
adrenaline and cortisol/corticosterone, and further, sex hormones such as androgens and estradiol can
interact with the adrenal axes to mediate the physiological response to stress (Pasquali. 2012). However,
given that adrenalectomy was only conducted in males, sex differences cannot be fully evaluated based on
the available studies.
While there is strong animal toxicological evidence from multiple laboratories showing that
short-term ozone exposure induces an adrenal-mediated stress response that results in a cascade of
metabolic effects including metabolic syndrome-like effects in animals and complications related to
diabetes in humans, several uncertainties remain. The controlled human exposure study is generally
consistent with the animal toxicological evidence, suggesting activation of the HPA axis and an
associated mobilization of energy stores, but additional controlled human exposure studies were not
available for review. In addition, as described above, there were several metabolic-disease-related
endpoints that did not change following short-term ozone exposure in the controlled human exposure
study. Note, however, that this study was conducted in nonfasting, healthy volunteers, thus differences in
glucose or insulin levels may not be expected. Furthermore, there were a limited number of epidemiologic
studies available for review, and many of these did not control for copollutant confounding. Thus, the
overall paucity of human studies contributes to uncertainties in the evidence. There is also some added
uncertainty in extrapolating dosimetry from rodent models to humans. However, this issue was addressed
considerably in the 2013 Ozone ISA (U.S. EPA. 2013a). as well as in Appendix 3 (Section 3.1.4.1.2).
Finally, an additional source of uncertainty is some heterogenicity in the animal toxicological evidence.
This may arise from differences in study design, such as varying exposure concentrations (0.25-1.0 ppm)
and durations, differences in rodent strain, sex, or diet, or from the measurement of effects at different
time points post-exposure.
Nonetheless, despite limited controlled human exposure and epidemiologic evidence, the
expanding animal toxicological studies show robust evidence for short-term ozone exposure contributing
to an array of metabolic effects. These outcomes follow a biologically plausible pathway whereby
adrenaline and cortisol/corticosterone are released from the adrenal glands. These hormones act on
multiple organs and tissues of the metabolic system to mobilize energy reserves, including glucose and
lipids. In line with this, animal toxicological studies show that inhibiting adrenaline and/or corticosterone,
through either removal of the adrenal glands or adrenal medulla, or by blocking the synthesis of
corticosterone, prevents ozone-induced metabolic effects, including hyperglycemia, glucose intolerance,
and elevated circulating triglycerides. In summary, based on evidence from animal toxicological and
5-29
-------
epidemiologic studies, as well as some support from one controlled human exposure study, short-term
ozone exposure consistently impairs glucose and insulin homeostasis and increases triglycerides and fatty
acids. In addition, there are generally consistent effects from animal toxicological studies showing that
short-term ozone exposure affects obesity-relevant endpoints and causes inflammation in adipose tissue.
Further supporting evidence comes from a limited number of animal toxicological studies providing some
evidence for alterations in HDL, LDL, and total cholesterol and changes in blood pressure following
short-term ozone exposure. Overall, the collective evidence is sufficient to conclude that the
relationship between short-term ozone exposure and metabolic effects is likely to be causal.
Table 5-3 Summary of evidence for a likely to be causal relationship between
short-term ozone exposure and metabolic effects.
Ozone
Rationale for Concentrations
Causality Associated with
Determination Key Evidence Key References Effects
Consistent animal
toxicological
evidence from
multiple,
high-quality studies
at relevant ozone
concentrations
Glucose intolerance and insulin
resistance
Thomson et al. (2018); 0.25-1 ppm
Gordon et al. (2017b);
Gordon et al. (2017a);
Miller et al. (2016c): Miller
et al. (2016b): Miller et al.
(2015): Bass et al. (2013)
Increased triglycerides and free fatty
acids
Farrai et al. (2016): 0.8-1.0 ppm
Gordon et al. (2016):
Miller etal. (2016c)
Increase in corticosterone and/or
adrenaline
Miller et al. (2016c): 0.8-1 ppm
Thomson et al. (2016):
Miller etal. (2015): Bass
et al. (2013): Thomson et
al. (2013)
Consistent but Increased adipose tissue inflammation Yina et al. (2016): Zhona 0.5 ppm
limited animal et al. (2016): Sun et al.
toxicological (2013)
evidence at relevant
ozone Liver biomarkers Miller et al. (2016b): Miller 0.5-1 ppm but not
concentrations et al. (2015): Theis et al. 0.25 ppm
(2014)
5-30
-------
Table 5-3 (Continued): Summary of evidence for a likely to be causal relationship
between short-term ozone exposure and metabolic effects.
Ozone
Rationale for
Concentrations
Causality
Associated with
Determination
Key Evidence
Key References
Effects
Inconsistent animal
Hyperglycemia
Thomson et al. (2018):
0.25-1 ppm
toxicological studies
Gordon etal. (2017b):
at relevant ozone
Gordon et al. (2017a):
concentrations
Miller etal. (2016b): Yinq
et al. (2016): Zhonq etal.
(2016): Miller etal. (2015):
Vella etal. (2014): Bass et
al. (2013)
Changes in body weight and/or
Henriquez et al. (2018):
0.5-1.0 ppm
weight-regulating hormones
Thomson et al. (2018):
Gordon et al. (2016):
Thomson et al. (2016):
Yinq et al. (2016): Zhonq
et al. (2016): Miller etal.
(2015): Bass et al. (2013):
Sun et al. (2013)
Altered HDL, LDL, total cholesterol
Gordon et al. (2017b);
Gordon et al. (2017a):
Farrai et al. (2016):
Gordon et al. (2016):
Miller et al. (2016c): Miller
et al. (2015): Ramot et al.
(2015): Vella et al. (2014):
Bass et al. (2013): Farrai
et al. (2012)
0.8-1.0 ppm
Epidemiologic
Epidemioloqic evidence for positive Chuanq et al. (2010)
26.8 ppb
evidence of
associations between short-term ozone
increased risk of
exposure and increased indicators of
diabetes or
impaired glucose and insulin
metabolic syndrome
homeostasis, including HOMA-IR,
dyslipidemia, elevated HbA1c, and
increased fasting glucose
Limited
epidemiologic
evidence from
case-crossover and
panel studies of
metabolic endpoints
Limited number of studies with generally
null associations (glucose, HOMA-IR,
insulin) observed among populations
with or without pre-existing disease
Kim and Hong (2012)
Chen et al. (2016)
Li et al. (2017)
Dales et al. (2012)
19.4-64.4 ppb
Controlled human
exposure evidence
of increased
metabolic changes
with ozone
exposure at
relevant
concentration
One study observed ketone body
formation, increased fatty acids, and
increased Cortisol and corticosterone
Miller etal. (2016a)
0.3 ppm
5-31
-------
Table 5-3 (Continued): Summary of evidence for a likely to be causal relationship
between short-term ozone exposure and metabolic effects.
Rationale for
Causality
Determination
Key Evidence
Key References
Ozone
Concentrations
Associated with
Effects
Limited
epidemiologic
evidence from
copollutant models
provides some
support for an
independent ozone
association
The magnitude of ozone associations
remains relatively unchanged in a limited
number of studies evaluating copollutant
models, including PM2.5 and other
gaseous pollutants
Kim and Hong (2012)
Dales etal. (2012)
Section 5.1.6
Biological
plausibility
Experimental studies show removal of
the adrenal glands diminishes
ozone-induced metabolic effects
Miller et al. (2016c)
Thomson et al. (2016)
0.8-1.0 ppm
5.2 Long-Term Ozone Exposure—Introduction, Summary from
the 2013 Ozone ISA, and Scope for Current Review
Metabolic effects were not included in the 2013 Ozone ISA as a distinct section because there
were few studies evaluating the effects of long-term ozone exposure on these outcomes. One study
presented in the Cardiovascular Disease Appendix (Section 4.2.8.1) evaluated the association between
long-term ozone exposure and effects on blood lipids and glucose homeostasis (Chuang etal.. 2011). This
study reported increases in total cholesterol, fasting glucose, and hemoglobin Ale. Multiple experimental
animal studies have evaluated ozone-mediated effects, and these studies indicate that long-term exposure
to ozone may affect glucose homeostasis and other factors that may contribute to metabolic syndrome.
The effects from long-term ozone exposure reviewed here include indicators of metabolic
function that underlie metabolic and cardiovascular diseases. Outcomes evaluated include glucose and
insulin homeostasis, weight gain, metabolic syndrome, type 1 and type 2 diabetes, and mortality from
diabetes or cardiometabolic diseases. The subsections below evaluate the most policy-relevant scientific
evidence relating long-term ozone exposure to metabolic effects. These sections focus on studies
published since the completion of the 2013 Ozone ISA.
5.2.1 Population, Exposure, Comparison, Outcome, and Study Design
(PECOS) Tool
The scope of this section is defined by a scoping tool that generally defines the relevant
Population, Exposure, Comparison, Outcome, and Study Design (PECOS). The PECOS tool defines the
parameters and provides a framework to help identify the relevant evidence in the literature to inform the
5-32
-------
ISA. Because the 2013 Ozone ISA did not make a causality determination for long-term ozone exposure
and metabolic health effects, the epidemiologic studies evaluated are less limited in scope and not
targeted towards specific study locations, as reflected in the PECOS tool. The studies evaluated and
subsequently discussed within this section were identified using the following PECOS tool:
Experimental Studies:
• Population: Study populations of any controlled human exposure or animal toxicological study of
mammals at any lifestage
• Exposure: Long-term (over 30 days) inhalation exposure to relevant ozone concentrations
(i.e., <2 ppm for mammals)
• Comparison: In toxicological studies of mammals and in controlled human exposures, an
appropriate comparison group that is exposed to a negative control (i.e., clean air or filtered air
control)
• Outcome: Metabolic effects
• Study Design: In vivo chronic-duration, subchronic-duration, or repeated-dose toxicity studies in
mammals or immunotoxicity studies
Epidemiologic Studies:
• Population: Any population, including populations or lifestages that might be at increased risk
• Exposure: Long-term (months to years) exposure to ambient concentrations of ozone
• Comparison: Per unit increase (in ppb)
• Outcome: Change in risk (incidence/prevalence) of metabolic effects
• Study Design: Epidemiologic studies consisting of panel, case-crossover, time-series studies, and
case-control studies; cross-sectional studies with appropriate timing of exposure for the health
endpoint of interest
5.2.2 Biological Plausibility
This section describes biological pathways that potentially underlie metabolic effects resulting
from long-term exposure to ozone. Figure 5-2 graphically depicts the proposed pathways as a continuum
of upstream events, connected by arrows that may lead to downstream events observed in epidemiologic
studies. This discussion of "how" exposure to ozone may lead to metabolic health effects contributes to
an understanding of the biological plausibility of epidemiologic results evaluated later in Section 5.2.4.
Note that the structure of the biological plausibility sections and the role of biological plausibility in
contributing to the weight-of-evidence analysis used in the 2020 Ozone ISA are discussed in
Section IS.4.2.
5-33
-------
Activation of
Sensory
Nerves in
Respiratory
Tract
Gestational
Diabetes
Long-Term
Ozone
Exposure
Metabolic
Syndrome
Mortality
Diabetes
Note: The boxes above represent the effects for which there is experimental or epidemiologic evidence related to ozone exposure,
and the arrows indicate a proposed relationship between those effects. Solid arrows denote evidence of essentiality as provided, for
example, by an inhibitor of the pathway or a genetic knockout model used in an experimental study involving ozone exposure.
Shading around multiple boxes is used to denote a grouping of these effects. Arrows may connect individual boxes, groupings of
boxes, and individual boxes within groupings of boxes. Progression of effects is generally depicted from left to right and color coded
(gray, exposure; green, initial effect; blue, intermediate effect; orange, effect at the population level or a key clinical effect). Here,
population-level effects generally reflect results of epidemiologic studies. When there are gaps in the evidence, there are
complementary gaps in the figure and the accompanying text below. The structure of the biological plausibility sections and the role
of biological plausibility in contributing to the weight-of-evidence analysis used in the 2020 Ozone ISA are discussed in
Section IS.4.2.
Figure 5-2 Potential biological pathways for metabolic outcomes following
long-term ozone exposure.
Ozone inhalation can contribute to gestational diabetes or to metabolic syndrome or diabetes,
potentially leading to increased mortality via two primary pathways. The first pathway is initiated by
activation of sensory neurons in the respiratory tract, and the second pathway by respiratory tract injury,
inflammation, and oxidative stress.
With respect to the first pathway, signals from the airways are integrated in the brainstem and
hypothalamus and lead to a modulation of the autonomic nervous system and subsequent activation of the
sympathetic-adrenal-medullary (SAM) axis. Specifically, ozone inhalation activates vagal afferents in the
respiratory tract, resulting in signaling from catecholaminergic neurons in the nucleus tractus solitarius to
the paraventricular nucleus of the hypothalamus, a brain region that integrates limbic and autonomic
functions (Gackiere et aL 2011). Following long-term ozone exposure, stimulation of the hypothalamus
5-34
-------
activates the SAM axis, leading to a release of adrenaline (epinephrine) from the adrenal medulla (Miller
et al.. 2016c). Long-term ozone exposure also results in increased adrenaline levels (Miller et al.. 2016b).
likely mediated through the same mechanism, activation of the SAM axis. Chronic activation of the
adrenergic system affects metabolism (Ciccarclli et al.. 2013) and can potentially result in observed
ozone-induced hyperglycemia (Miller et al.. 2016b). glucose intolerance (Bass et al.. 2013). altered
insulin signaling (Miller et al.. 2016b). altered serum lipids (Bass et al.. 2013). and increased risk of being
overweight or obese (Li et al.. 2015; Dong et al.. 2014).
The second pathway through which long-term ozone exposure may affect the metabolic system
begins with the direct effect of ozone on respiratory tract injury, inflammation, and oxidative stress (see
Appendix 3 for detail). Briefly, in the lung, ozone reacts with the respiratory epithelial cell lining fluid
leading to local and systemic inflammatory responses and oxidative stress (see also Appendix 7). When
circulating inflammatory cytokines and reactive oxygen species present in the bloodstream reach target
tissues and organs, such as the pancreas, liver, and adipose tissue, signaling mechanisms are initiated that
trigger local and systemic inflammation (Hotamisligil. 2017). Following long-term exposure to ozone,
serum markers of inflammation have been detected (Miller et al.. 2016b). Systemic inflammation can
contribute to altered insulin signaling (Arkan et al.. 2005). glucose intolerance (van Beek et al.. 2014).
hyperglycemia, altered serum lipids (Wellen and Hotamisligil. 2005). and obesity (O'Rourke. 2009).
Activation of the neuroendocrine stress axes, including the SAM axis, is typically an adaptive
mechanism that allows for rapid access to energy reserves. There are numerous negative feedback
mechanisms in place that promote homeostasis following this activation. Indeed, evidence shows that the
ozone-induced increase in adrenaline is transient even following long-term exposure to ozone (Miller et
al.. 2016b). However, with repeated or continuous exposure, or in those with pre-existing metabolic
disease, exposure to ozone could plausibly cause the allostatic load to become too high, resulting in
hormonal and metabolic dysregulation, ultimately leading to diabetes or metabolic syndrome. Thus, the
proposed pathways discussed above provide biological plausibility for metabolic syndrome and/or
diabetes following long-term ozone exposure, and will be used to inform a causality determination, which
is discussed later in this Appendix (Section 5.2.9).
5.2.3 Metabolic Syndrome
Individuals with metabolic syndrome are at a fivefold increased risk for developing type 2
diabetes and a twofold increased risk for developing cardiovascular disease within 5-10 years (Alberti et
al.. 2009). Criteria for metabolic syndrome include elevated fasting blood glucose (hyperglycemia),
elevated triglycerides, low levels of high-density lipoprotein (HDL), obesity (particularly abdominal
obesity), and high blood pressure (Table 5-1). The presence of three out of the five criteria meets the
clinical diagnosis for metabolic syndrome. Below, the available studies that investigate the effects of
long-term ozone exposure on these and related endpoints are characterized.
5-35
-------
5.2.3.1
Elevated Fasting Glucose
Elevated fasting blood glucose levels, or hyperglycemia, can occur when the body has too little
insulin, or if the insulin present cannot be used properly. Under normal conditions, insulin is secreted by
(3-cells within the pancreas in response to glucose levels. When glucose levels rise, depolarization of the
pancreatic (3-cells or modulation by other hormones stimulate insulin secretion (Nadal et al.. 2009). Thus,
during food intake, blood insulin levels rise, stimulating glucose uptake and replenishing body fuel
reserves in the form of triglycerides and glycogen. When glucose levels decrease (e.g., during fasting),
insulin secretion is inhibited and glucagon is secreted from the pancreas, which stimulates fuels, such as
lipids from adipose tissue and amino acids from muscle, to be mobilized to the bloodstream where they
are used by the liver to synthesize glucose.
The effects of long-term ozone exposure on blood glucose levels are characterized below. Given
the critical role of insulin on glucose homeostasis, we also review effects of long-term ozone exposure on
glucose tolerance and insulin secretion and signaling. To measure how quickly glucose is cleared from the
blood, the glucose tolerance test (GTT) samples blood glucose levels at multiple time points after glucose
injection or ingestion. It is useful in diagnosing or monitoring diabetes or gestational diabetes. The insulin
tolerance test (ITT) induces an episode of hypoglycemia through the injection of insulin. With a
functioning HPA axis, hypoglycemia triggers a subsequent release of cortisol/corticosterone from the
adrenal glands and an increase in blood glucose levels.
5.2.3.1.1 Animal Toxicological Studies
The 2013 Ozone ISA did not contain information on long-term ozone exposure and metabolic
effects. Since then, several new animal toxicological studies have been published examining the effects of
long-term ozone exposure on glucose levels and insulin homeostasis. Effects on fasting glucose levels
were inconsistent, possibly due to the transience of ozone-induced effects. However, long-term ozone
exposure resulted in consistently impaired glucose tolerance. Results from one study suggest ozone
impairs insulin secretion, and in another, that senescent animals are more sensitive to ozone-induced
effects on insulin. Specific information is detailed below and in the evidence inventory tables that follow
(Table 5-13).
• Young and old male brown Norway rats (4, 12, and 24 months of age) exposed 2 days/week for
13 weeks (6 hours/day, 0.25 or 1.0 ppm ozone) did not show ozone-induced hyperglycemia when
evaluated 18 hours post-exposure (Bass et al.. 2013). A separate study found that male Wistar
Kyoto rats were hyperglycemic after 12 weeks of intermittent exposure to 1.0 ppm ozone
(5 hours/day, 3 days/week), an effect that was absent following a 1-week recovery period (Miller
et al.. 2016b). No effects were observed following exposure to 0.25 ppm ozone (Miller et al..
2016b).
• Long-term ozone-induced glucose intolerance was evaluated in young and old male brown
Norway rats (4, 12, and 24 months of age) exposed 2 days/week for 13 weeks [6 hours/day,
5-36
-------
0.25 or 1.0 ppm ozone; Bass et al. (2013)1. Glucose tolerance was significantly impaired in all
age groups exposed to 1.0 ppm. There were no effects from 0.25 ppm ozone exposure. Notably,
the magnitude of change in blood glucose levels following long-term exposure at 1.0 ppm was
lower than the magnitude of change following short-term exposure to the same dose
(Section 5.1.2). Similarly, in male Wistar Kyoto rats, 13 weeks of intermittent exposure
(5 hours/day, 3 days/week) to 1.0 ppm ozone impaired glucose tolerance, but this effect was
absent following a 1-week recovery period of no exposure (Miller et al.. 2016b). No effects were
observed following exposure to 0.25 ppm ozone (Miller et al.. 2016b).
• Male Wistar Kyoto rats showed no impairment in peripheral insulin-mediated glucose clearance
after 13 weeks of exposure to 0.25 or 1.0 ppm ozone, nor did they have elevated serum insulin
levels [5 hours/day, 3 days/week; Miller et al. (2016b)l. However, rats exposed to 1.0 ppm
(5 hours or 5 hour/day, 3 days/week for 13 weeks), but not 0.25 ppm, showed no glucose-induced
increase in insulin levels at 30 minutes, suggesting impairments in (3-cell insulin secretion (Miller
et al.. 2016b).
• The effect of long-term ozone exposure on insulin levels differed by age. When exposed to
0.8 ppm ozone for 1 day/week for 17 weeks (6 hours/day), male brown Norway rats that were
senescent (24 months old) had higher serum insulin levels than did filtered air controls, while
there were no ozone-induced effects in young rats [8 months old; Gordon et al. (2013)1.
5.2.3.2 Elevated Triglycerides
High levels of triglycerides are one criterion for metabolic syndrome. Triglycerides are stored in
adipose tissue and are also present in the blood where they bidirectionally transfer fat and glucose from
the liver. Lipolysis is the pathway through which triglycerides are broken down into glycerol and free
fatty acids. Lipolysis is induced by several hormones, including glucagon, adrenaline, and
cortisol/corticosterone (Friihbeck et al.. 2014).
5.2.3.2.1 Animal Toxicological Studies
No studies examining effects from long-term ozone exposure on serum lipids were included in
the 2013 Ozone ISA. Two recent animal toxicological studies suggest long-term ozone exposure may
result in decreased serum triglycerides (Table 5-13). These studies further suggest that ozone-induced
changes in lipid profiles are sensitive to animal age and differ based on the amount of elapsed time since
ozone exposure. Specifically:
• A recent animal toxicological study examined effects from long-term ozone exposure (0.25 and
1.0 ppm, 6 hours/day, 2 days/week for 13 weeks) on serum lipids in male brown Norway rats
aged 4, 12, and 24 months. In 24-month-old rats, exposure to 0.25 ppm ozone decreased levels of
triglycerides (Bass et al.. 2013). There were no other significant effects on triglycerides. In a
separate study, ozone-induced effects on serum lipids were investigated immediately after
13 weeks of exposure (0.25 and 1.0 ppm, 5 hours/day, 3 days/week) as well as following a
1-week recovery period post-exposure. Immediately following exposure, there were no changes
in triglycerides, free fatty acids, or branched chain amino acids. After the 1-week recovery period,
5-37
-------
levels of triglycerides (1.0 ppm) were significantly lower compared with filtered air controls
(Miller et al.. 2016b).
5.2.3.3 Low High-Density Lipoprotein Cholesterol (HDL-C)
Low levels of high-density lipoprotein (HDL) have been associated with negative health
outcomes, including metabolic syndrome. HDL particles remove fat molecules from cells and transport
lipids, such as cholesterol, to the liver either for secretion or hormone synthesis (Yu et al.. 2019). In
contrast, low-density lipoprotein (LDL) particles deliver fat molecules to cells. LDL can also enter the
endothelium, become oxidized, and contribute to atherosclerosis (Yu et al.. 2019). Because direct
measurements of HDL and LDL are costly, blood tests are typically used to measure HDL-C or LDL-C,
the cholesterol carried within the associated particles.
5.2.3.3.1 Animal Toxicological Studies
No studies examining effects from long-term ozone exposure on serum lipids were included in
the 2013 Ozone ISA. Two recent animal toxicological studies suggest long-term ozone exposure may
result in decreased HDL-C in some groups, with differences potentially mediated by age and the amount
of time elapsed since ozone exposure (Table 5-13). Specifically:
• A recent animal toxicological study examined effects from long-term ozone exposure (0.25 and
1.0 ppm, 6 hours/day, 2 days/week for 13 weeks) on serum lipids in male brown Norway rats
aged 4 , 12, and 24 months. In 12-month-old rats, 1.0 ppm ozone increased HDL levels. In
24-month-old rats, exposure to 0.25 ppm ozone decreased levels of HDL (Bass et al.. 2013).
There were no other significant effects on total cholesterol, HDL, or LDL. In a separate study,
ozone-induced effects on serum lipids were investigated immediately after 13 weeks of exposure
(0.25 and 1.0 ppm, 5 hours/day, 3 days/week) as well as following a 1-week recovery period.
Measured immediately, exposure to 1.0 ppm ozone resulted in increased total cholesterol and
increased HDL; however, when measured 1 week later, exposure to both 0.25 and 1.0 ppm
resulted in decreased total cholesterol and decreased HDL compared with filtered air controls.
After the 1-week recovery period, levels of LDL (0.25 and 1.0 ppm) were significantly lower than
in filtered air controls (Miller et al.. 2016b).
5.2.3.4 Central Adiposity
Central adiposity or enlarged waist circumference is an additional criterion of metabolic
syndrome. While centrally accumulated body fat is associated with insulin resistance, peripherally
distributed body fat is less relevant to metabolism. Animal toxicological studies have examined changes
in weight gain and body composition following long-term ozone exposure, which can be mediated
through dysregulated food and water intake. The hormones leptin and ghrelin indirectly mediate glucose
and insulin homeostasis (Meier and Gressner. 2004). Leptin is secreted by adipocytes and regulates food
5-38
-------
intake. High levels of leptin suppress feeding, while low levels increase food intake. Leptin also
modulates insulin secretion and signaling, and leptin resistance can lead to hyperphagia and weight gain.
Also involved in regulating energy homeostasis are the hormones adiponectin and resistin, both of which
have been linked to obesity (Meier and Gressner. 2004).
5.2.3.4.1 Epidemiologic Studies
No epidemiologic studies in the 2013 Ozone ISA evaluated the relationship between long-term
ozone exposure and adiposity, weight gain, or obesity. Recent evidence is limited but provides some
evidence that long-term exposure to ozone is associated with increased weight gain and obesity, although
these studies do not specifically address waist circumference (Table 5-12). Specifically:
• White et al. (2016) analyzed data from 38,374 women from the Black Women's Health Study
Cohort in a prospective study of weight gain. The women lived within 56 metropolitan areas,
weighed between 80-300 pounds, were under 55 years old, had not had gastric bypass surgery,
and had not given birth in the previous 2 years. Ozone exposure was estimated using the CMAQ
model 8-hour max concentration for the centroid of the census tract of residence. The study used
a 16-year follow-up and found no weight change associated with an increased exposure to ozone
(0.23 kg; 95% CI: -0.16, 0.64).
• Two studies evaluated the prevalence of being overweight or obese related to ambient
concentrations of ozone. In a study by Dong et al. (2014). 30,056 children were recruited from
seven cities in northeast China. Body mass index (BMI) was calculated according to World
Health Organization (WHO) protocol, and the Centers for Disease Control and Prevention's
(CDC) definition of overweight and obese were used to categorize status. Ozone exposure was
determined using the 3-year avg of the 8-hour max concentration of the monitor closest to the
school children attended. Increased odds for children being overweight (OR: 1.16; 95% CI: 1.05,
1.27) or obese (OR: 1.26; 95% CI: 1.07, 1.45) were observed. However, the study did not report
correlations for copollutants, making it difficult to assess the ozone-specific outcomes.
• The second study evaluated adults from the 33 Communities Chinese Health Study Cohort in
participants that were 18-74 years of age and had lived in the same location for more than 5 years
(Li et al.. 2015). The sample included 24,845 participants and used a 3-year avg of the daily
8-hour avg exposure recorded at the monitor nearest to their home. There were increased odds of
being overweight (OR: 1.08; 95% CI: 1.04, 1.12) and obese (OR: 1.09; 95% CI: 1.01, 1.18)
associated with long-term exposure to ozone. Both males (OR: 1.09; 95% CI: 1.03, 1.15) and
females (OR: 1.05; 95% CI: 1.00, 1.12) had increased odds of becoming overweight, while only
females had increased odds of becoming obese (OR: 1.12; 95% CI: 1.01, 1.26). As in Dong et al.
(2014). copollutant correlations were not reported, and both PMi0 and SO2 observations were
high in the 33 communities, so it is difficult to estimate the level of confounding from other
ambient pollutant exposures.
5.2.3.4.2 Animal Toxicological Studies
The 2013 Ozone ISA contained no animal toxicological studies on overweight and obesity-related
endpoints with ozone exposure. The effect of long-term ozone exposure on body composition, body
5-39
-------
weight, and serum leptin and adiponectin levels was examined in recent studies (Table 5-13). Only one of
these studies reported a change in leptin levels, which was time-point specific:
• Long-term ozone exposure had no effect on body composition or body weight in male brown
Norway rats at age 8 or 24 months [0.8 ppm, 6 hours/day, 1 day/week for 17 weeks; Gordon et al.
(2013)1. Furthermore, there was no ozone-induced effect on serum leptin or adiponectin.
Similarly, an additional study found no effects from 1.0 ppm ozone exposure (6 hours/day,
2 days/week for 13 weeks) on adiponectin levels (Bass et al.. 2013). Lastly, a third study found
no changes in serum leptin immediately following 13 weeks of ozone exposure (0.25 and
1.0 ppm, 5 hours/day, 3 days/week), but reported that exposure to 1.0 ppm ozone resulted in
decreased leptin levels at 1 week post-exposure (Miller etal.. 2016b).
5.2.3.5 Elevated Blood Pressure
High blood pressure is one of the criteria for metabolic syndrome. Hypertension is a clinically
relevant consequence of chronically high blood pressure, which typically develops over years.
The effects of long-term ozone exposure on blood pressure and hypertension are discussed in
detail in Appendix 4. In summary, recent epidemiologic evidence and human exposure evidence is limited
in number and generally inconsistent. However, there is some emerging evidence that long-term ozone
exposure may be associated with changes in blood pressure or hypertension among different lifestages or
in those with pre-existing disease. In the 2013 Ozone ISA, no animal toxicological studies examined the
relationship between long-term exposure to ozone and changes in blood pressure. Recently, Gordon et al.
(2013) reported that long-term exposure (17 weeks) to ozone (0.8 ppm) did not result in changes in blood
pressure in adult or senescent rats (Table 4-41). Thus, there continues to be no evidence from animal
toxicological studies that long-term exposure to ozone can result in changes in blood pressure.
5.2.4 Other Indicators of Metabolic Function
5.2.4.1 Liver Biomarkers
The liver, which connects the portal and systemic circulation, is the site for primary energy and
xenobiotic metabolism (Boron and Boulpaep. 2017). The liver is a crucial organ for maintaining glucose
homeostasis. It can be stimulated to increase blood glucose by inducing gluconeogenesis (production of
glucose from noncarbohydrates) during fasting or to store glucose after feeding. With intense
gluconeogenesis, the liver produces ketone bodies which are released together with the newly produced
glucose. The liver can also synthesize and degrade proteins, carbohydrates, and lipids for distribution to
extrahepatic tissues depending on energy needs. Finally, the liver regulates whole-body cholesterol
balance via biliary excretion of cholesterol, cholesterol conversion to bile acids, and by regulating
5-40
-------
cholesterol synthesis (Boron and Boulpaep. 2017). Consequently, the liver is an essential regulator of
whole-body metabolism and energy homeostasis.
5.2.4.1.1 Animal Toxicological Studies
The 2013 Ozone ISA contained no animal toxicological studies on liver biomarkers. One recent
animal toxicological study examined changes in the liver following long-term ozone exposure. To
determine whether ozone dependent hyperglycemia in male Wistar Kyoto rats was controlled by hepatic
gluconeogenesis, a pyruvate tolerance test was performed where pyruvate (a substance for
gluconeogenesis) was injected and blood glucose measured overtime rMiller et al. (2016b); Table 5-141.
There was no significant increase in gluconeogenesis following 13 weeks of ozone exposure [0.25 and
1.0 ppm, 5 hours/day, 3 days/week; Miller et al. (2016b)l.
5.2.4.2 Adrenal Hormones
Ozone exposure modulates the autonomic nervous system, which can trigger the
sympathetic-adrenal-medullary (SAM) and hypothalamic-pituitary-adrenal (HPA) axes, resulting in a
release of adrenaline and cortisol/corticosterone, respectively, from the adrenal glands (Nicolaides et al..
2015). These hormones increase circulating levels of blood glucose and fatty acids and have widespread
effects on metabolic processes (Section 5.2.2).
5.2.4.2.1 Animal Toxicological Studies
One animal toxicological study examined changes in adrenal hormones following long-term
ozone exposure. In male Wistar Kyoto rats, after 13 weeks of ozone exposure (0.25 and 1.0 ppm,
5 hours/day, 3 days/week), plasma levels of adrenaline were higher in animals exposed to 1.0 ppm ozone.
However, when measured 1 week post-exposure, the ozone-induced increase in adrenaline was no longer
present. There were no ozone-induced changes in noradrenaline or corticosterone rMiller et al. (2016b);
Table 5-141.
5.2.5 Development of Diabetes
The diagnostic testing criteria for diabetes are listed in Table 5-4. The Ale, which is also known
as the hemoglobin Ale, HbAlc, or glycohemoglobin, is a blood test that provides information about a
person's average blood glucose over the past 3 months by measuring the percentage of hemoglobin (i.e., a
blood protein with a 3-month lifespan) modified by glucose. In animal toxicological and epidemiologic
studies, the homeostasis model assessment (HOMA) has been widely used to quantify insulin resistance
5-41
-------
(HOMA-IR) and pancreatic p-cell (HOMA-J3) function and used to infer diabetes risk. The HOMA-IR
index is given by the product of basal insulin and glucose levels divided by 22.5, whereas the HOMA-|3
index is derived from the product of 20 and basal insulin levels divided by glucose concentration minus
3.5 (Wallace et al.. 2004; Matthews et al.. 1985).
Table 5-4 Criteria for clinical diagnosis of diabetes.
Test
Criteria
A1c
>6.5%a
OR
Fasting plasma glucose (FPG)
>126 mg/dL (7 mmol/L). Fasting is defined as no caloric intake for at least 8 h.a
OR
Oral glucose tolerance test
2-h plasma glucose >200 mg/dL (11.1 mmol/L) during OGTT. The test should
(OGTT)
be performed as described by the World Health Organization using a glucose
load containing the equivalent of 75 g of anhydrous glucose dissolved in water.3
OR
Random glucose test
>200 mg/dL (11.1 mmol/L) in a person with classical symptoms of
hyperglycemia or hyperglycemic crisis
mg/dL = milligrams per deciliter; mmol/L = millimoles per liter.
aln the absence of unequivocal hyperglycemia, Criteria 1—3 should be confirmed by repeat testing.
Source: Test criteria extracted from ADA (20141.
5.2.5.1 Type 2 Diabetes
5.2.5.1.1 Epidemiologic Studies
No long-term epidemiologic studies of metabolic syndrome or type 2 diabetes were evaluated in
the 2013 Ozone ISA. Recent studies, which are listed in Table 5-15. include large cohort studies around
the world; they provide evidence for increased incidence for type 2 diabetes and metabolic syndrome.
Specifically:
• Jerrett et al. (2017) analyzed data from the Black Women's Health Study Cohort in a prospective
study of type 2 diabetes. The 43,003 women were greater than 30 years old, resided in
56 metropolitan areas, and had BMI information at baseline. Ozone was estimated using the
CMAQ model 8-hour max concentration for the centroid of the census tract of residence between
5-42
-------
2007-2008 to approximate long-term averages. The study observed increased hazard ratios for
incident diabetes (1.28; 95% CI: 1.06, 1.55); when adjusted for NO2, this relationship was slightly
weaker and had wider confidence intervals (1.20; 95% CI: 0.96, 1.50).
• Using the Rome Longitudinal Study Cohort, Renzi et al. (2017) evaluated the effects of ozone
exposure in over one million subjects over 35 years old without diabetes at baseline. The study
used the Flexible Air Quality Regional Model (FARM) with a 1-km grid dispersion and 2005
seasonal ozone (May-September 8-hour avg) to predict the spatial distribution of ozone in Rome
between 2008-2013. The study observed modest positive hazard ratios for incidence of diabetes
for those living in Rome (1.01; 95% CI: 1.00, 1.02). Additionally, when the ozone model was
adjusted for NOx, the increased incidence remained relatively unchanged (1.02; 95% CI: 1.00,
1.03).
• Yang et al. (2018) looked at the odds of developing metabolic syndrome due to exposure to ozone
in adults from the 33 Communities Chinese Health Study Cohort in participants that were
18-74 years of age and had lived in the same location for more than 5 years. Ozone exposure was
measured at municipal air monitoring stations, and the 8-hour daily mean concentrations were
aggregated into a 3-year avg. In a study population of 15,477, the odds of metabolic syndrome
increased (1.16; 95% CI: 1.12, 1.23) according to the American Heart Association's definition.
The study reported high correlations of ozone with PM10 (r = 0.81) and SO2 (r = 0.84); these high
correlations may indicate copollutant confounding and are a source of uncertainty in estimating
the direct effect of ozone on metabolic syndrome.
5.2.5.2 Type 1 Diabetes
Type 1 diabetes mellitus (T1D), which typically affects children and young adults, is a chronic
condition that results when the pancreas fails to produce the insulin needed for glucose homeostasis.
There were no epidemiologic studies of T1D in the 2013 Ozone ISA. The evidence relating to the effect
of long-term exposure to ozone on T1D is limited to a prospective study in Scania, Sweden rMalmqvist et
al. (2015); Table 5-151. The study evaluated prenatal exposure during first, second, and third trimesters of
pregnancies for children born between 1999-2005. Ozone exposure was measured by the nearest
monitoring station, averaging the 24-hour ozone concentrations, and aggregating them into trimester
averages. The levels were categorized in quartiles with the reference exposure being set at a level less
than 22 ppb and the highest quartile exposure over 30.6 ppb. There were elevated ORs for T1D in the
highest quartile of ozone concentrations in the first (1.52; 95% CI: 0.88, 2.61) and second trimester
(1.62; 95% CI: 0.99, 2.65), although confidence intervals were wide. There was no evidence of
association with third-trimester exposure levels.
5.2.5.3 Gestational Diabetes
Several studies of gestational diabetes were conducted. Generally, the results of the studies were
inconsistent, although several reported positive associations with gestational diabetes or impaired glucose
tolerance with ozone exposures during the second trimester. While the evidence base for gestational
5-43
-------
diabetes is growing, it is still limited to a relatively small number of studies which report generally
inconsistent results (see the "Pregnancy and Birth Outcomes" section for more details rSection 7.1.31V
5.2.6 Metabolic Disease Mortality
Studies that examine the association between long-term ozone exposure and cause-specific
mortality outcomes, such as diabetes or other metabolic disease mortality, provide additional evidence for
ozone-related metabolic effects, and contribute to the evidence for an overall continuum of effects.
There were no studies that evaluated the relationship between long-term ozone exposure and
mortality due to diabetes or cardiometabolic disease in the 2013 Ozone ISA. However, recent analyses
from the ACS cohort in the U.S. and the CanCHEC cohort in Canada provide consistent evidence for
positive associations between long-term ozone exposure and mortality due to diabetes or cardiometabolic
diseases ITurner et al. (2016); Crouse et al. (2015); see Section 6.2.3.2. Figure 6-10 for more details].
5.2.7 Potential Copollutant Confounding of the Ozone-Metabolic Effects
Relationship
The evaluation of potential confounding effects of copollutants on the relationship between
long-term ozone exposure and metabolic effects allows for examination of whether ozone risks are
changed in copollutant models. Recent studies examined the potential for copollutant confounding by
evaluating copollutant models that included PM2 5, PM10, and NO2. These recent studies help inform the
extent to which effects associated with long-term ozone exposure are independent of coexposure to
correlated copollutants in long-term analyses.
• Using the Rome Longitudinal Study Cohort, Renzi et al. (2017) evaluated the effects of long-term
ozone exposure in over one million subjects over 35 years old without diabetes at baseline. The
study showed modest positive hazard ratios for incidence of diabetes for those living in Rome
(1.01; 95% CI: 1.00, 1.02). Additionally, when the ozone model was adjusted for NOx, the
positive association remained (1.02; 95% CI: 1.00, 1.03).
• Jerrett et al. (2017) analyzed data from the Black Women's Health Study Cohort in a prospective
study of type 2 diabetes. The authors observed increased hazard ratios for incident diabetes (1.28;
95% CI: 1.06, 1.55). When adjusted for PM2 5, the hazard ratios further increased (1.31; 95% CI:
1.08, 1.60), but when adjusted for NO2, the estimate was slightly attenuated and less precise
(1.20; 95% CI: 0.96, 1.50).
5-44
-------
5.2.8
Effect Modification of the Ozone-Metabolic Effects Relationship
5.2.8.1 Lifestage
No studies evaluated in the 2013 Ozone ISA compared associations between different age groups
for long-term ozone exposure and metabolic health effects. A limited number of recent epidemiologic
studies provide some evidence that long-term ozone exposure increases the risk for diabetes in younger
adults, while older adults may have an increased risk of being overweight or obese. A recent animal
toxicological study suggests older adult rodents show changes in insulin signaling with long-term ozone
exposure.
• Using the Rome Longitudinal Study Cohort, Renzi et al. (2017) evaluated the effects of ozone
exposure in over one million subjects over 35 years old without diabetes at baseline. When
stratified by age, the study showed increased hazard ratios for incidence of diabetes for those
under 50 years (1.05; 95% CI: 1.02, 1.08) but not those from 50-60 years (1.02; 95% CI: 0.99,
1.04), or over 60 years (1.00; 95%: 0.98, 1.02).
• Jerrett et al. (2017) analyzed data from the Black Women's Health Study Cohort in a prospective
study of type 2 diabetes. When the population was further analyzed by age, increased hazard
ratios for incident diabetes was seen in women aged 40-54 (1.33; 95% CI: 1.03, 1.72), was
higher, although less precise, for those under 40 (1.43; 95% CI: 0.90, 2.25), and lower for those
over 55 (1.24; 95% CI: 0.90, 1.72).
• In a study by Dong et al. (2014). 30,056 children were recruited from seven cities in northeast
China. BMI was calculated according to WHO protocol, and the CDC definitions of overweight
and obese were used to categorize status. Ozone exposure was determined using the 3-year avg of
the 8-hour max concentration of the monitor closest to the school children attended. Increased
odds for children being overweight (1.16; 95% CI: 1.05, 1.27) or obese (1.26; 95% CI: 1.07, 1.45)
were observed. Li et al. (2015) used the 33 Communities Chinese Health Study Cohort in
participants that were 18-74 years of age and had lived in the same location for more than
5 years. When the population was stratified for age (over or under 50 years), the population over
50 years had increased odds of being overweight (1.12; 95% CI: 1.05, 1.19), and females over
50 years also had increased odds of obesity (1.23; 95% CI: 1.04, 1.44). There were no differences
found in the under 50-year age group for increased odds of being overweight or obese.
• A recent study examined the effect of age on health outcomes in rodents. Senescent or aged
animals were more sensitive to ozone-dependent serum insulin changes. Ozone-exposed
senescent males had significantly increased serum insulin than did aged filtered air controls [male
brown Norway rats ozone, 6 hours/day, 1 day/week for 17 weeks; 4-month-olds or aged animals
20 months old; Gordon et al. (2013)1. In the same study, 4-month-old adult rodents exposed to
ozone did not have significant changes in serum insulin with ozone exposure (Gordon et al..
2013). Thus, age contributes to the insulin response to ozone, with aged animals producing
statistically significantly increased levels of insulin with ozone exposure, an effect not present in
younger adult rodents.
5-45
-------
5.2.8.2
Pre-existing Disease
No studies evaluated in the 2013 Ozone ISA evaluated the potential of pre-existing disease to
modify the relationship between long-term ozone exposure and metabolic health effects. Recent
epidemiologic studies evaluated the potential for pre-existing diseases to modify associations between
long-term ozone exposure and metabolic effects.
• Using the Rome Longitudinal Study Cohort, Renzi et al. (2017) evaluated the effects of ozone
exposure in over one million subjects over 35 years old without diabetes at baseline. When
stratified by subjects that had comorbidities (myocardial infarction, COPD, hypertension, or
hyperlipidemia), the authors observed an increased incidence of diabetes (1.02; 95% CI: 1.00,
1.03) that was similar to the increased incidence observed among those without comorbidities
(1.01; 95% CI: 1.00, 1.05).
• Jerrett et al. (2017) analyzed data from the Black Women's Health Study Cohort in a prospective
study of type 2 diabetes. The study found increased hazard ratios for incident diabetes (1.28; 95%
CI: 1.06, 1.55); with pre-existing hypertension, the effect increased (1.35, 95% CI: 1.03, 1.76) but
was attenuated without the presence of hypertension (1.15; 95% CI: 0.85, 1.53).
5.2.9 Summary and Causality Determination
There were no causality conclusions for metabolic effects in the 2013 Ozone ISA (U.S. EPA.
2013a). The evidence pertaining to outcomes from long-term exposure to ozone and metabolic effects has
expanded since the 2013 Ozone ISA (U.S. EPA. 2013a'). with multiple epidemiologic and animal
toxicological studies currently available for review (Table 5-5).
The strongest evidence for metabolic effects following long-term ozone exposure is provided by
epidemiologic studies. Positive associations between long-term ozone concentrations and diabetes-related
mortality were observed in well-established cohorts in the U.S. and Canada (Turner et al.. 2016; Crouse et
al.. 2015V Two studies also reported an association between long-term ozone concentrations and
increased hazard ratios for incident diabetes (Jerrett et al.. 2017; Renzi et al.. 2017). Furthermore, some
studies reported a relationship between long-term ozone concentrations and increased odds of being
overweight or obese. Finally, another study found increased odds of metabolic syndrome (Yang et al..
2018). which is characterized by hyperglycemia (elevated glucose levels), elevated triglycerides, low
HDL-C, high blood pressure, and an enlarged waist circumference. In the few animal toxicological
studies that were available, there is evidence that long-term ozone exposure may result in hyperglycemia
(Section 5.2.3.1) and decreased HDL-C (Section 5.2.3.3). as well as induce glucose intolerance, cause
(3-cell dysfunction, and alter insulin secretion (Section 5.2.3.1V However, while there were a limited
number of animal toxicological studies available for review, these provided little support for long-term
ozone-induced changes in other markers of metabolic syndrome, including triglycerides (Section 5.2.3.2).
body composition (Section 5.2.3.4). or blood pressure (Section 5.2.3.5). In addition, there was no
5-46
-------
evidence for increased hepatic gluconeogenesis or increased corticosterone levels, although one study
reported a transient increase in adrenaline following long-term ozone exposure (Section 5.2.4V
Despite an increased number of studies, there are many remaining uncertainties regarding the
metabolic effects from long-term ozone exposure. Most studies from the epidemiologic literature did not
evaluate potential copollutant confounding. There were a very limited number of studies available for
review from the animal toxicological literature and these studies had few overlapping endpoints. Overall,
considering the positive epidemiologic studies and limited support from animal toxicological
studies, the collective evidence is suggestive of, but not sufficient to infer, a causal relationship
between long-term exposure to ozone and metabolic effects.
5-47
-------
Table 5-5 Summary of evidence that is suggestive of, but not sufficient to infer,
a causal relationship between long-term ozone exposure and
metabolic effects.
Rationale for
Causality
Determination3
Key Evidence13
Key References'3
Ozone
Concentrations
Associated with
Effects0
Consistent
epidemiologic
evidence of
increased risk for
diabetes or
metabolic
syndrome
Increased odds of metabolic
syndrome, increased hazard ratio for
incidence of diabetes, including
evidence from a U.S. cohort
Yana et al. (2018)
Renzi et al. (2017)
Jerrett et al. (2017)
24.7-48.7 ppb
Increased odds of developing
gestational diabetes with ozone
exposure in the second trimester.
Elevated ORs for type 1 diabetes,
with higher ozone concentrations in
first and second trimester.
Malmqvist et al. (2015)
See Section 7.1.3
Epidemiologic
evidence of
increased
diabetes-
associated
mortality
A limited number of studies
observed positive associations
between long-term ozone exposure
and mortality from diabetes and
cardiometabolic diseases
Turner et al. (2016)
Crouse et al. (2015)
34.3-38.2 ppb
Consistent animal
toxicological
evidence but
limited number of
available studies
Impaired glucose tolerance
Bass et al. (2013)
Miller et al. (2016b)
1.0 ppm but not
0.25 ppm
Altered insulin
Gordon et al. (2013)
Miller et al. (2016b)
0.8, 1.0 ppm but not
0.25 ppm
Altered serum lipids
Bass et al. (2013)
Miller etal. (2016b)
0.25, 1.0 ppm
Increased inflammation
Miller etal. (2016b)
1.0 ppm
Transient increase in adrenaline
Miller etal. (2016b)
1.0 ppm but not
0.25 ppm
Inconsistent
and/or null results
in animal
toxicology studies
Hyperglycemia
Miller et al. (2016b): Bass et al.
(2013)
1.0 ppm
No change in body composition or
body weight
Gordon et al. (2013)
Bass etal. (2013)
0.8, 1.0 ppm
No hepatic gluconeogenesis
Miller etal. (2016b)
1.0 ppm
No increase in corticosterone
Miller etal. (2016b)
0.25, 1.0 ppm
5-48
-------
Table 5-5 (Continued): Summary of evidence that is suggestive of, but not
sufficient to infer, a causal relationship between long-term
ozone exposure and metabolic effects.
Rationale for
Causality
Determination3
Key Evidence13
Key References'3
Ozone
Concentrations
Associated with
Effects0
Uncertainty due to Overlapping authorship across all
lack of animal toxicological studies
independent
laboratories
conducting animal
toxicology studies
Bass et al. (2013) 0.25-1.0 ppm
Gordon et al. (2013)
Miller et al. (2016b)
Uncertainty Limited number of epidemiologic Renzi et al. (2017) Section 5.2.3
regarding studies evaluate potential Jerrett et al. (2017)
confounding by copollutant cofounding for PM or
copollutants NOx
Biological Experimental studies provide Section 5.2.2
plausibility evidence of metabolic syndrome
mediated by adrenergic activation
OR = odds ratio; ppm = parts per million.
aBased on aspects considered in judgments of causality and weight of evidence in causal framework in Table I and Table II of the
Preamble to the ISAs (U.S. EPA. 2015).
bDescribes the key evidence and references, supporting or contradicting, contributing most heavily to causality determination and,
where applicable, to uncertainties or inconsistencies. References to earlier sections indicate where full body of evidence is
described.
°Describes the ozone concentrations with which the evidence is substantiated.
5-49
-------
5.3 Evidence Inventories—Data Tables to Summarize Study
Details
Table 5-6 Epidemiologic studies of short-term exposure to ozone and
metabolic syndrome.
Study
Study Population
Exposure
Assessment
Mean(ppb)
Copollutant
Examination
Effect Estimates
HR (95% CI)
tKim and Hong
(2012)
Seongbuk-Gu,
Seoul, South Korea
Ozone: 2008-2010
Panel study
Lags examined:
0-10
KEEP
n = 560
Participants over
60 yr old in the
Seongbuk-Gu area
of Seoul, South
Korea
Daily mean
Mean:
concentration
19.38
of monitor
Median:
nearest
19.34
residence
75th: 26.67
24-h avg
90th: 29.56
95th: 31.33
Correlation
(r):
N02: -0.35;
SO2: -0.3;
PM10: -0.12
Copollutant
models: NO2,
PM10
Using O3
lag 5,
NO2 lag
Day 7, and
PM10 lag
Day 4
Lag 5:
Percentage increase in
glucose
0.19 (0.09, 0.28)
Without pre-existing T2D:
0.09 (0.02, 0.16)
With pre-existing T2D: 0.68
(0.28, 1.07)
Adjusted for PM10: 0.15
(0.05, 0.25)
Adjusted for NO2: 0.16
(0.06, 0.25)
Percentage increase in
HOMA:
0.30 (0.06, 0.53)
Without pre-existing T2D:
0.12 (-0.11, 0.35)
With pre-existing T2D: 1.21
(0.44, 1.99)
Adjusted for PM10: 0.25
(-0.001, 0.49)
Adjusted for NO2: 0.21
(-0.02, 0.45)
Percentage increase in
insulin:
0.71 (0.02, 1.38)
Without pre-existing T2D:
0.32 (-0.39, 1.02)
Pre-existing T2D: 2.76
(0.78, 4.75)
Adjusted for PM10: 0.67
(-0.06, 1.39)
Adjusted for NO2: 0.49
(-0.20, 1.19)
5-50
-------
Table 5-6 (Continued): Epidemiologic studies of short-term exposure to ozone
and metabolic syndrome.
Study
Study Population
Exposure
Assessment
Mean(ppb)
Copollutant
Examination
Effect Estimates
HR (95% CI)
tChen etal. (2016)
California, U.S.
Ozone: 2002-2008
Panel study
p-gene
n = 1,023
Mexican American
women with a
previous diagnosis
of GDM within
previous 5 yr,
siblings and
cousins (both
sexes) all with
fasting glucose
levels <7 mmol/L
Daily average
Mean:
Correlation
of monitored
30-day
(r): PM2.5:
air quality data
cumulative:
30 day:
spatially
43.4 ppb;
-0.02;
mapped to
Annual
annual: 0.04;
residence
average:
NO2: 30 day:
locations using
40.8 ppb
-0.37;
inverse
annual:
distance
-0.31
squared
interpolation
Copollutant
models: NR
with a
maximum
radius of 50 km
24-h avg
Daily averages
Mean: 23.7
Correlation
of two ozone
(r):
monitors in the
PM2.5: 0.01;
greater Boston,
NO2: -0.54;
MA area
SO2: 0.13;
24-h avg
BC: -0.26
Copollutant
models: NR
Qualitative results only, no
change in fractional
disappearance rate
Glucose
HOMA-IR
Insulin
Metabolic clearance
Insulin sensitivity
No change in: HDL-to-LDL
ratio
LDL
tLi etal. (2017)
Northeastern U.S.
Ozone: 2002-2005
and 2008-2011
Panel study
Framingham
Offspring Cohort
and Third
Generation Cohort
n = 4,116
Residents within
50 km of Harvard
Supersite
excluding patients
with diabetes at
the time of
examination visits
(fasting glucose
>126 mg/dL)
Qualitative results only:
Decrease in percentage
glucose at 24-h, 3- and
7-day avg
Negative trend for
percentage change in
Insulin HOMA-IR at 24-h,
3- and 7-day avg
Negative trend
nonsignificant; qualitative
results only: resistin
No trend; qualitative
results: leptin
Positive trend
nonsignificant; qualitative
results only: adiponectin
BC = black carbon; GDM = gestational diabetes mellitus; HDL = high-density lipoproteins; HOMA-IR = Homeostatic Model
Assessment of Insulin Resistance; LDL = low-density lipoproteins; mg/dL = milligrams per deciliter; mmol/L = millimoles per liter;
N02 = nitrogen dioxide; NR = not reported; 03 = ozone; PM25 = particulate matter with a nominal aerodynamic diameter less than
or equal to 2.5 |jm; PM10 = particulate matter with a nominal aerodynamic diameter less than or equal to 10 |jm; ppb = parts per
billion; S02 = sulfur dioxide; T2D = type 2 diabetes.
tStudies published since the 2013 Ozone ISA.
5-51
-------
Table 5-7 Controlled human exposure study of short-term exposure to ozone
and metabolic syndrome.
Study
Population N, Sex,
Age(Range or
Mean ± SD)
Exposure Details
(Concentration, Duration)
Endpoints Examined
Miller etal. (2016a)
Healthy young adults
n = 20 males,
4 females
Age: 25.6 ± 3.8 yr
0.3 ppm, 2 h (15 min of exercise
alternating with 15 min of rest)
Insulin, insulin resistance,
nonfasting glucose, leptin,
marker of ketone bodies, free
fatty acids, triglycerides, glycerol,
HDL-C, LDL-C, total cholesterol
HDL-C = high-density lipoprotein cholesterol; LDL-C = low-density lipoprotein cholesterol; ppm = parts per million.
Table 5-8 Study-specific details from animal toxicological studies of short-term
exposure to ozone and metabolic syndrome.
Species (Stock/Strain), N,
Study Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Bass etal. (2013)
Rats (BN)
n = 4-21/group males,
0 females
Age: 1,4, 12, and 24 mo
old
Rats (BN)
n = 4-21/group males,
0 females
Age: 1, 9, and 21 mo
0.25 ppm, 6 h/day for 2 days
0.25 ppm, 6 h/day, 2 days/week for
13 weeks
1 ppm, 6 h/day for 2 days
1 ppm, 6 h/day, 2 days/week for
13 weeks
Fasting glucose, glucose
tolerance, adrenaline,
cholesterol (total, HDL,
LDL), triglycerides, serum
leptin, IL-6, insulin, mRNA
biomarkers in liver and
adipose (NR)
Vella etal. (2014)
Rats (Wistar) 0.8 ppm, 16 h (with and without
n = 4-10 males, 0 females pretreatment of /V-acetylcysteine)
Age: adult (400-450 g)
Fasting glucose, serum
insulin, HOMA-IR, glucose
infusion rate, nonesterified
fatty acid, total cholesterol,
triglycerides, triacylglycerol
Zhona etal. (2016)
Mice (KK; obesity-prone
develops moderate
degrees of obesity, insulin
resistance, and diabetes)
n = 8/group males,
0 females
Age: adult
0.5 ppm, 4 h/day for 3 consecutive
days
Fasting glucose, (3-cell
insulin secretory function,
serum insulin, insulin
resistance, insulin, leptin,
adiponectin
5-52
-------
Table 5-8 (Continued): Study-specific details from animal toxicological studies of
short-term exposure to ozone and metabolic syndrome.
Species (Stock/Strain), N,
Study Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Thomson et al.
(2016)
Rats (F344)
n = NR males, 0 females
Age: adult (200-250 g)
0.8 ppm (with or without
metyrapone or corticosterone), 4 h
Serum insulin, glucagon,
ghrelin levels
Miller et al. (2016c)
Rats (WKY)
n = 5/group males,
0 females
Age: adult
1 ppm, 4 h/day for 1 or 2 days (rats
underwent bilateral adrenal
demedullation, total bilateral
adrenalectomy, or sham surgery)
Fasting glucose, glucose
tolerance, triglycerides, free
fatty acids, cholesterol
(total)
Immediately PE
Gordon et al.
(2017b)
Rats (LE)
n = 0 males, 10/group
females
Age: 22 days at start of
exercise regimen
0.25 ppm, 5 h/day for 2 days
0.5 ppm, 5 h/day for 2 days
1 ppm, 5 h/day for 2 days
(with or without exercise)
Fasting glucose, glucose
tolerance, serum insulin,
serum triglycerides
Body composition (lean, fat,
fluid percentage), BALF,
EMKA plethysmography
Immediately PE Day 1
Miller etal. (2016b)
Rats (WKY)
n = 8-10/group males,
0 females
Age: adult (250-300 g)
0.25 ppm, 5 h/day for
3 consecutive days/week for
1 week
1 ppm, 5 h/day for 3 consecutive
days/week for 1 week
Fasting glucose, glucose
tolerance, serum insulin,
free fatty acids, (3-cell insulin
secretory function, hepatic
gluconeogenesis,
cholesterol, adrenaline,
noradrenaline,
corticosterone, insulin
resistance-AKT (NR)
Gordon et al.
(2017a)
Rats (LE)
n = 8 offspring total,
4 males, 4 females when
possible.
10 dams/treatment group
0.8 ppm, 4 h/day for 2 consecutive
days
Age: adult (30 days) offspring
ozone challenge, PND
161-162 ozone exposure
Pregnant females and offspring
(control diet [CD]-sedentary [SED];
CD-run wheel [RW]; high fat
diet-SED; HFD-RW); begin diet
6 weeks prior to mating/conception
Fasting glucose, glucose
tolerance, triglycerides,
cholesterol, insulin, free
fatty acids
Miller etal. (2015)
Rats (WKY)
Male
n = 6-8/group males,
0 females
Age: 10 weeks
(250-300 g)
0.25, 0.50, or 1.0 ppm ozone,
6 h/day for 2 days
Fasting glucose, glucose
tolerance, serum insulin
GTT, insulin, leptin,
cholesterol (total, LDL,
HDL), metabolomics, free
fatty acids
Thomson et al.
(2018)
Rats (F344)
n = 6-8/group males,
0 females
Age: adult
0.8 ppm (with or without
metyrapone), 4 h, whole body
exposure
Fasting blood glucose,
serum insulin and glucagon,
glucose tolerance, insulin
release, ghrelin
5-53
-------
Table 5-8 (Continued): Study-specific details from animal toxicological studies of
short-term exposure to ozone and metabolic syndrome.
Study
Species (Stock/Strain), N,
Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Yina et al. (2016)
Mice (KKAy, diabetic
prone)
n = 8/group males,
0 females
Age: 4-5 weeks
0.5 ppm, 4 h/day for
13 consecutive days
Fasting glucose, insulin
signaling and secretion,
inflammatory cytokines,
body weight, leptin
Sun et al. (2013)
Rats (S-D)
n = 4-8 males, 0 females
Age: 8 weeks
0.5 ppm, 2 weeks (O3 and
O3 + CAPs; high fructose or
normal diet for 8 weeks prior)
Body-weight changes with
ozone ± diet modification,
characterization of fat
depots (brown adipose vs.
white adipose), histology of
fat tissue (morphology
changes), tissue
adiponectin concentration,
leptin
24 h PE
Farrai et al. (2012) Rats (SH)
Male
Age: 12 weeks
0.8 ppm ozone, 4 h, whole body
exposure
Cholesterol (total, LDL,
HDL)
Farrai et al. (2016) Rats (SH)
Male
Age: 12 weeks
0.3 ppm ozone, 3 h, whole body
exposure
Cholesterol (total, LDL,
HDL), triglycerides
Ramot et al. (2015) Rats (WKY)
Male
1.0 ppm ozone, 4 h, whole body
exposure
Cholesterol (total, LDL,
HDL)
Gordon et al. (2016) Rats (BN)
n = 10/group males,
10/group females
Age: 30 days
0.8 ppm, 5 h (high fructose or high Cholesterol (total, LDL,
fat diet for 12 weeks prior)
0.8 ppm, 5 h/day, 1 day/week for
4 weeks (high-fructose or high-fat
diet for 12 weeks prior)
HDL), blood glucose levels,
triglycerides, body weight,
effect of diet and exercise
on endpoints, changes in
body composition (fat, lean,
liquid mass)
18 h PE
Henriquez et al.
(2018)
Rats (SH)
Male
Age: 12 weeks
0.8 ppm, 4 h/day for 1 or 2 days Body weight
AKT = protein kinase B; BN = brown Norway; CAP = criteria air pollutants; F344 = Fischer 344; GTT = glucose tolerance test;
HDL = high-density lipoproteins; HOMA-IR = Homeostatic Model Assessment of Insulin Resistance; JNK = c-Jun N-terminal
kinase; LDL = low-density lipoproteins; LE = Long-Evans; NR = not reported; 03 = ozone; PE = post-exposure; PND = postnatal
day; ppm = parts per million; S-D = Sprague-Dawley; SH = spontaneously hypertensive; WKY = Wistar Kyoto.
5-54
-------
Table 5-9
Study-specific details from epidemiologic studies of short-term
complications from diabetes.
Exposure
Copollutant
Effect Estimates
Study
Study Population Assessment Mean (ppb)
Examination
HR (95% CI)
tDales et al. (2012) n = general
Santiago Province,
Chile
Ozone: 2001-2008
Cross-sectional
study
population of
five sectors was
5 million
Daily hospital
admissions where
diabetes was the
principal diagnosis
(insulin dependent
and noninsulin-
dependent) with
coma or
ketoacidosis
Daily
averaged
monitor(s) in
the sector of
residence
24-h avg
Mean: 64.41
Correlation (r):
PM2.5: -0.31;
NO2: -0.31; SO2:
-0.08
Copollutant
models: NR
Increased risk for
hospital
admission for
diabetic coma or
diabetic
ketoacidosis: 1.02
(1.00, 1.04)
N02 = nitrogen dioxide; PM2 5 = particulate matter with a nominal aerodynamic diameter less than or equal to 2.5 |jm; ppb = parts
per billion; S02 = sulfur dioxide.
tStudies published since the 2013 Ozone ISA.
Table 5-10 Study-specific details from animal toxicological studies of
short-term, other indicators of metabolic function.
Species
(Stock/Strain), N, Exposure Details
Study Sex, Age (Concentration, Duration) Endpoints Examined
Multiorgan gene expression
(mRNA pathway analysis
antioxidant response,
xenobiotic metabolism,
inflammatory signaling, and
endothelial dysfunction) and
glucocorticoid activity (plasma
levels of adrenocorticotropic
hormone and the
glucocorticoid corticosterone)
Immediately PE and after 24 h
FA recovery
Sun et al. (2013)
Rats (S-D)
0.5 ppm, 2 weeks (± CAPs; high
Adipose tissue gene
n = 4-8 males,
fructose or normal diet for 8 weeks
expression
0 females
prior)
24 h PE
Age: 8 weeks
Thomson et al. (2013) Rats (F344) 0.4 ppm, 4 h (nose only)
n = 4-6/group 0 8 PPm. 4 h (nose on|y)
males, 0 females
Age: adult
(200-250 g)
5-55
-------
Table 5-10 (Continued): Study-specific details from animal toxicological studies
of short-term, other indicators of metabolic function.
Study
Species
(Stock/Strain), N,
Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Bass etal. (2013)
Rats (BN)
n = 4-21/group
males, 0 females
Age: 1, 4, 12, and
24 mo old
Rats (BN)
n = 4-21/group
males, 0 females
Age: 1, 9, and 21 mo
0.25 ppm, 6 h/day for 2 days
0.25 ppm, 6 h/day, 2 days/week for
13 weeks
1 ppm, 6 h/day for 2 days
1 ppm, 6 h/day, 2 days/week for
13 weeks
Adrenaline, IL-6, mRNA
biomarkers in liver and
adipose (NR)
Miller etal. (2015)
Rats (WKY)
n = 6-8/group
males, 0 females
Age: 10 weeks
(250-300 g)
0.25 ppm, 6 h/day for 2 days
1 ppm, 6 h/day for 2 days
IL-6, livertranscriptomics,
adrenaline, bile acid profiles
Theis etal. (2014)
Rats (S-D)
n = 6/group males,
0 females
Age: adult
0.5 ppm, 8 h/day for 5 days
Liver endpoints (liver
enzymes, liver proteomics
[stress responsive proteins,
glucose-regulated protein 78,
and protein disulfide
isomerase, glutathione
s-transferase M1,
hemeoxygenase-1]; NR)
Yinq etal. (2016)
Mice (KKAy, diabetic 0.5 ppm, 4 h/day for 13 consecutive Adipose tissue inflammation
prone) days
n = 8/group males,
0 females
Age: 4-5 weeks
Zhonq etal. (2016)
Mice (KK;
obesity-prone
develops moderate
degrees of obesity,
insulin resistance,
and diabetes)
n = 8/group males,
no females
Age: adult
0.5 ppm, 4 h/day for 13 consecutive
weekdays
Visceral adipose
characterization (oil-red-o
stain), inflammatory genes in
adipose (CXCL-11, IFN-y,
TNF-a, IL-12, and iNOS)
22 h PE
Miller etal. (2016c)
Rats (WKY)
n = 5/group males,
0 females
Age: adult
1 ppm, 4 h/day for 1 or 2 days (rats
underwent bilateral adrenal
demedullation [DEMED], total
bilateral adrenalectomy; ADX), or
sham surgery (SHAM)
Corticosterone, adrenaline,
noradrenaline
Thomson et al. (2016)
Rats (F344)
0.8 ppm (with or without
Corticosterone, adrenal levels
n = NR males,
metyrapone or corticosterone), 4 h
0 females
Age: adult
(200-250 g)
5-56
-------
Table 5-10 (Continued): Study-specific details from animal toxicological studies
of short-term, other indicators of metabolic function.
Study
Species
(Stock/Strain), N,
Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Miller etal. (2016b)
Male Rats (WKY)
n = 8-10/group
Age: adult
(250-300 g)
0.25 ppm, 5 h/day for 1 week
Pyruvate tolerance
Gordon et al. (2017a)
Rats (LE)
n = 8 offspring total,
4 males, 4 females
when possible.
10 dams/treatment
group.
0.8 ppm, 4 h/day for 2 consecutive
days, Age: adult (30 days) offspring
ozone challenge, PND 161-162
ozone exposure
Pregnant females and offspring
(control diet [CD]-sedentary [SED];
CD-run wheel [RW]; high fat
diet-SED; HFD-RW); begin diet
6 weeks prior to mating/conception
Corticosterone,
noradrenaline
adrenaline,
BN = brown Norway; CAP :
PE = post-exposure; PND :
criteria air pollutants; F344 = Fischer 344; FA = filtered air; NR = not reported; 03 = ozone;
postnatal days; ppm = parts per million; S-D = Sprague-Dawley; WKY = Wistar Kyoto.
Table 5-11 Study-specific details from controlled human exposure studies of
short-term, other indicators of metabolic function.
Study
Population N, Sex,
Age(Range or
Mean ± SD)
Exposure Details
(Concentration, Duration)
Endpoints Examined
Miller etal. (2016a)
Healthy young
adults
n = 20 males,
4 females
Age: 25.6 ± 3.8 yr
0.3 ppm, 2 h (15 min of exercise
alternating with 15 min of rest)
Cortisol, corticosterone,
cortisone, ketone bodies, free
fatty acids
ppm = parts per million.
5-57
-------
Table 5-12 Epidemiologic studies of long-term exposure to ozone and metabolic
syndrome.
Study
Study
Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
HR (95% CI)
tDong etal. (2014) n = 30,056
Shenyang, Dalian,
Anshan, Fushun,
Benxi, Liaoyang, and
Yingkou, China
Ozone: 2006-2008
Follow-up:
cross-sectional study
Children aged
2-14 yr living in
seven cities in
northeast China
attending
schools within
1 km of a
monitoring site
Monitors
within 1 km
of school
8-h max
Mean: 27.4
Maximum: 44.5
Correlation (r): Increased odds of obesity
NR
Copollutant
models: NR
in children: 1.26 (1.07,
1.45)
Increased odds of
overweight children: 1.16
(1.05, 1.28)
Increased odds of obese or
overweight children: 1.26
(1.11, 1.41)
tLi etal. (2015)
Shenyang, Anshan,
and Jinzhou, China
Ozone: 2006-2008
Follow-up: 2009
Cross-sectional
study
33CCHS
n = 24,845
Participants
18-74 yr of age
in 11 districts in
three Chinese
cities
(Shenyang,
Anshan,
Jinzhou)
Monitor
within 1 km
of the
household,
using the
daily 8-h avg
to create a
3-yr avg
concentration
8-h avg
Mean: 25.1
Maximum: 36.0
Correlation (r):
NR
Copollutant
models: NR
Increased odds of being
overweight: 1.08 (1.04,
1.12)
Increased odds of males
being overweight: 1.09
(1.03 1.15)
Increased odds of females
being overweight: 1.05
(1.00, 1.12)
Increased odds of obesity:
1.09 (1.01, 1.18)
Increased odds of male
obesity: 0.99 (0.88, 1.12)
Increased odds of female
obesity: 1.12 (1.01, 1.26)
tWhite et al. (2016)
56 metropolitan
areas, U.S.
Ozone: 2007-2008
Follow-up:
1995-2011
Cohort study
Black Women's
Health Study
n = 38,374
Black women
living in
56 metropolitan
areas in the
U.S., under
55 yr of age,
without history
of cancer or
gastric bypass
surgery, and
who had not
given birth
within the past
2 yr
CMAQ model
with a
resolution of
12 km.
Estimates
were made at
the centroid
of each
census tract.
8-h max
Mean: 37.5
Correlation (r):
NR
Copollutant
models: NR
Absolute change in weight
(kg): 0.24 (-0.16, 0.64)
33CCHS = 33 Communities Chinese Health Study; CMAQ = Community Multiscale Air Quality; HR = hazard ratio; NR = not
reported.
tStudies published since the 2013 Ozone ISA.
5-58
-------
Table 5-13 Study-specific details from animal toxicological studies of long-term
exposure to ozone and metabolic syndrome.
Species
(Stock/Strain), N, Exposure Details
Study Sex, Age (Concentration, Duration) Endpoints Examined
Bass etal. (2013)
Rats (BN)
n = 4-21/group males,
0 females
Age: 1, 4, 12, and
24 mo old
Rats (BN)
n = 4-21/group males,
0 females
Age: 1, 9, and 21 mo
0.25 ppm, 6 h/day for 2 days
0.25 ppm, 6 h/day, 2 days/week
for 13 weeks
1 ppm, 6 h/day for 2 days
1 ppm, 6 h/day, 2 days/week for
13 weeks
Fasting glucose, glucose
tolerance, cholesterol (total,
HDL, LDL), triglycerides, serum
leptin, insulin, adiponectin
Gordon et al. (2013) Rats (BN) 0, 0.8 ppm, 6 h/day, 1 day/week Body weight, serum insulin,
n = 7-10/group males, for 17 weeks leptin, adiponectin, blood
0 females pressure
Age: adult (300 g)
Rats (LE, S-D, and
WKY)
n = 7-16 pups/group
males,
7-16 pups/group
females
Age: PND 14, PND21,
or PND 28
Miller etal. (2016b)
Rats (WKY)
n = 8-10/group males,
0 females
Age: adult (250-300 g)
0.25 ppm, 5 h/day for
3 consecutive days/week for
13 weeks
1 ppm, 5 h/day for
3 consecutive days/week for
13 weeks
Fasting glucose, glucose
tolerance, insulin, insulin-
mediated glucose clearance,
cholesterol (total, LDL, HDL),
triglycerides, free fatty acids,
branched chain amino acids,
leptin
BN = brown Norway; ER = estrogen receptor; HDL = high-density lipoproteins; LDL = low-density lipoproteins; LE = Long-Evans;
PND = postnatal days; ppm = parts per million; S-D = Sprague-Dawley; WKY = Wistar Kyoto.
5-59
-------
Table 5-14 Study-specific details from animal toxicological studies of long-term
exposure to ozone and other indicators of metabolic function.
Study
Species
(Stock/Strain), N,
Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Miller etal. (2016b)
Rats (WKY)
n = 8-10/group males,
0 females
Age: adult (250-300 g)
0.25 ppm, 5 h/day for
3 consecutive days/week for
13 weeks
1 ppm, 5 h/day for
3 consecutive days/week for
13 weeks
Adrenaline and noradrenaline
proinflammatory cytokines
(IFN-y, IL-4, and IL-10, IL-6,
TNF-a, IL-1 (3)
ppm = parts per million; WKY = Wistar Kyoto.
Table 5-15 Epidemiologic studies of long-term exposure to ozone and
development of diabetes.
Study
Study
Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
HR (95% CI)
tYana etal. (2018)
Shenyang, Anshan, and
Jinzhou, China
Ozone: 2006-2008
Follow-up: 2009
Cross-sectional study
33CCHS
n = 15,477
Participants
18-74 yr of age
in 11 districts in
three Chinese
cities
(Shenyang,
Anshan,
Jinzhou)
Monitor within
1 km of the
household,
using the daily
8-h avg to
create a 3-yr
avg
concentration
8-h avg
Mean: 25.1
Maximum:
36.0
Correlation (r):
PM2.5: 0.45;
NO2: 0.45;
SO2: 0.84;
PM10 0.81
Copollutant
models: NR
Increased odds of
metabolic syndrome
diagnosis American
Heart Association
criteria: 1.16 (1.12,
1.23)
Joint international
criteria: 1.21 (1.02,
1.39)
5-60
-------
Table 5-15 (Continued): Epidemiologic studies of long-term exposure to ozone
and development of diabetes.
Study
Study
Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
HR (95% CI)
tJerrett et al. (2017)
56 metropolitan areas,
U.S.
Ozone: 2007-2008
Follow-up: 1995-2011
Cohort study
Black Women's
Health Study
n = 43,003
Black women
living in
56 metropolitan
areas in the
U.S., aged 30 yr
and over at the
time of follow-up
without
prevalent
diabetes at
baseline
CMAQ model
with a
resolution of
12 km.
Estimates were
made at the
centroid of
each census
tract.
8-h max
Mean: 37.5
Correlation (r):
PM2.5: -0.29;
NO2: -0.57
Copollutant
models: NR
Increased HR for T2D
diagnosis: 1.28 (1.06,
1.55)
Increased HR for T2D
diagnosis adjusted for
NO2: 1.20 (0.96, 1.50)
Increased HR for T2D
diagnosis adjusted for
PM2.5: 1.31 (1.08, 1.60)
Increased HR for T2D
diagnosis under age
40 yr: 1.43 (0.90, 2.25)
Increased HR for T2D
diagnosis age
40-54 yr: 1.33 (1.03,
1.72)
Increased HR for T2D
diagnosis over age
55 yr: 1.25 (0.90, 1.72)
Increased HR for T2D
diagnosis without
presence of
hypertension: 1.15
(0.85, 1.53)
Increased HR for T2D
diagnosis with
presence of
hypertension: 1.35
(1.03, 1.76)
5-61
-------
Table 5-15 (Continued): Epidemiologic studies of long-term exposure to ozone
and development of diabetes.
Study
Study
Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
HR (95% CI)
tRenzi et al. (2017)
Rome, Italy
Ozone: 2005
Follow-up: 2008-2013
Cohort study
Rome
Longitudinal
Study
n = 1,319,193
Individuals over
35 yr of age
without
prevalent
diabetes at
baseline
Flexible Air
Quality
Regional
Model (FARM)
using a 1-km
grid dispersion
model
8-h avg
Mean: 49.4
Maximum:
57.3
Correlation (r):
PM2.5: -0.01;
NO2: -0.16
Copollutant
models: NR
Increased prevalence
of diabetes at baseline:
1.001 (0.991, 1.012)
Increased incidence of
diabetes: 1.012 (1,
1.024)
Increased HR of
incident diabetes
adjusted for NOx: 1.02
(1.00, 1.03)
Increased HR for
incident diabetes
female: 1.03 (1.01,
1.05)
Increased HR for
incident diabetes male:
0.99 (0.98, 1.01)
Increased HR for
incident diabetes under
age 50 yr: 1.05 (1.02,
1.08)
Increased HR for
incident diabetes age
50-60 yr: 1.02 (0.99,
1.04)
Increased HR for
incident diabetes over
age 60 yr: 1.00 (0.98,
1.02)
Increased HR for
incident diabetes with
comorbidities: 1.02
(1.00, 1.03)
Increased HR for
incident diabetes
without comorbidities:
1.02 (1.00, 1.05)
tMalmqvist et al. (2015) n = 930
Scania, Sweden
Ozone: 1999-2005
Follow-up: case-control
study
Gene-matched,
case-control
study of children
with or without
T1D born within
2 yr of each
other
Monitor within
32 km of
residence,
average
distance
8.5 km
24-h avg
Correlation (r):
NR
Copollutant
models: NR
Quartile 4 vs.
reference exposure
33CCHS = 33 Communities Chinese Health Study; CMAQ = Community Multiscale Air; HR = hazard ratio; N02 = nitrogen dioxide;
NR = not reported; PM2.5 = particulate matter with a nominal aerodynamic diameter less than or equal to 2.5 |jm; PM10 = particulate
matter with a nominal aerodynamic diameter less than or equal to 10 |jm; ppb = parts per billion; S02 = sulfur dioxide; T2D = type 2
diabetes.
tStudies published since the 2013 Ozone ISA.
5-62
-------
Annex for Appendix 5: Evaluation of Studies on Health Effects of
Ozone
This annex describes the approach used in the Integrated Science Assessment (ISA) for Ozone
and Related Photochemical Oxidants to evaluate study quality in the available health effects literature. As
described in the Preamble to the ISA (U.S. EPA. 2015). causality determinations were informed by
integrating evidence across scientific disciplines (e.g., exposure, animal toxicology, epidemiology) and
related outcomes and by judging the strength of inference in individual studies. Table Annex 5-1
describes aspects considered in evaluating study quality of controlled human exposure, animal
toxicological, and epidemiologic studies. The aspects found in Table Annex 5-1 are consistent with
current best practices for reporting or evaluating health science data.1 Additionally, the aspects are
compatible with published U.S. EPA guidelines related to cancer, neurotoxicity, reproductive toxicity,
and developmental toxicity (U.S. EPA. 2005. 1998. 1996. 1991).
These aspects were not used as a checklist, and judgments were made without considering the
results of a study. The presence or absence of particular features in a study did not necessarily lead to the
conclusion that a study was less informative or should be excluded from consideration in the ISA.
Further, these aspects were not used as criteria for determining causality in the five-level hierarchy. As
described in the Preamble, causality determinations were based on judgments of the overall strengths and
limitations of the collective body of available studies and the coherence of evidence across scientific
disciplines and related outcomes. Table Annex 5-1 is not intended to be a complete list of aspects that
define a study's ability to inform the relationship between ozone and health effects, but it describes the
major aspects considered in this ISA to evaluate studies. Where possible, study elements, such as
exposure assessment and confounding (i.e., bias due to a relationship with the outcome and correlation
with exposures to ozone), are considered specifically for ozone. Thus, judgments on the ability of a study
to inform the relationship between an air pollutant and health can vary depending on the specific pollutant
being assessed.
1 For example, NTP OHAT approach (Roonev et al.. 20141. IRIS Preamble (U.S. EPA. 2013b). ToxRTool
(Klimisch et al.. 19971. STROBE guidelines (von Elm et al.. 20071. and ARRIVE guidelines (Kilkenny et al.. 20101.
5-63
-------
Table Annex 5-1 Scientific considerations for evaluating the strength of
inference from studies on the health effects of ozone.
Study Design
Controlled Human Exposure:
Studies should describe clearly the primary and any secondary objectives of the study or specific hypotheses being
tested. Study subjects should be randomly assigned to treatment groups and exposed, to the extent possible
without knowledge of the exposure condition. Preference is given to balanced crossover (repeated measures) or
parallel design studies which include controlled exposures (e.g., to clean filtered air). In crossover studies, a
sufficient and specified time between exposure days should be provided to avoid carry-over effects from prior
exposure days. In parallel design studies, all arms should be matched for individual characteristics such as age,
sex, race, anthropometric properties, and health status. In studies evaluating effects of disease, appropriately
matched healthy controls are desired for interpretative purposes.
Animal Toxicology:
Studies should describe clearly the primary and any secondary objectives of the study or specific hypotheses being
tested. Studies should include appropriately matched controlled exposures (e.g., to clean filtered air, time matched)
and use methods to limit differences in baseline characteristics of control and exposure groups. Studies should
randomize assignment to exposure groups and where possible conceal allocation to research personnel. Groups
should be subjected to identical experimental procedures and conditions to the extent possible; animal care
including housing, husbandry, etc. should be identical between groups. Blinding of research personnel to study
group may not be possible due to animal welfare and experimental considerations; however, differences in the
monitoring or handling of animals in all groups by research personnel should be minimized.
Epidemiology:
Inference is stronger for studies that describe clearly the primary and any secondary aims of the study or specific
hypotheses being tested.
For short-term exposure, time-series, case-crossover, and panel studies are emphasized over cross-sectional
studies because they examine temporal correlations and are less prone to confounding by factors that differ
between individuals (e.g., SES, age). Panel studies with scripted exposures, in particular, can contribute to
inference because they have consistent, well-defined exposure durations across subjects, measure personal
ambient pollutant exposures, and measure outcomes at consistent, well-defined lags after exposures. Studies with
large sample sizes and those conducted over multiple years are considered to produce more reliable results.
Additionally, multicity studies are preferred over single-city studies because they examine associations for large
diverse geographic areas using a consistent statistical methodology, avoiding the publication bias often associated
with single-city studies.3 If other quality parameters are equal, multicity studies carry more weight than single-city
studies because they tend to have larger sample sizes and lower potential for publication bias.
For long-term exposure, inference is considered to be stronger for prospective cohort studies and case-control
studies nested within a cohort (e.g., for rare diseases) than cross-sectional, other case-control, or ecological
studies. Cohort studies can better inform the temporality of exposure and effect. Other designs can have uncertainty
related to the appropriateness of the control group or validity of inference about individuals from group-level data.
Study design limitations can bias health effect associations in either direction.
5-64
-------
Table Annex 5-1 (Continued): Scientific considerations for evaluating the strength
of inference from studies on the health effects of
ozone.
Study Population/Test Model
Controlled Human Exposure:
In general, the subjects recruited into study groups should be similarly matched for age, sex, race, anthropometric
properties, and health status. In studies evaluating effects of specific subject characteristics (e.g., disease, genetic
polymorphism, etc.), appropriately matched healthy controls are preferred. Relevant characteristics and health
status should be reported for each experimental group. Criteria for including and excluding subjects should be
indicated clearly. For the examination of populations with an underlying health condition (e.g., asthma),
independent, clinical assessment of the health condition is ideal, but self-report of physician diagnosis generally is
considered to be reliable for respiratory and cardiovascular disease outcomes.15 The loss or withdrawal of recruited
subjects during the course of a study should be reported. Specific rationale for excluding subject(s) from any portion
of a protocol should be explained.
Animal Toxicology:
Ideally, studies should report species, strain, substrain, genetic background, age, sex, and weight. Unless data
indicate otherwise, all animal species, stocks, and strains are considered appropriate for evaluating effects of ozone
exposure. It is preferred that the authors test for effects in both sexes and multiple lifestages and report the result
for each group separately. All animals used in a study should be accounted for, and rationale for exclusion of
animals or data should be specified.
Epidemiology:
There is greater confidence in results for study populations that are recruited from and representative of the target
population. Studies that have high participation, have low dropout over time, and are not dependent on exposure or
health status are considered to have low potential for selection bias. Clearly specified criteria for including and
excluding subjects can aid assessment of selection bias. For populations with an underlying health condition,
independent, clinical assessment of the health condition is valuable, but self-report of physician diagnosis generally
is considered to be reliable for respiratory and cardiovascular diseases.15 Comparisons of groups with and without an
underlying health condition are more informative if groups are from the same source population. Selection bias can
influence results in either direction or may not affect the validity of results but rather reduce the generalizability of
findings to the target population.
Pollutant
Controlled Human Exposure:
The focus is on studies testing ozone exposure.
Animal Toxicology:
The focus is on studies testing ozone exposure.
Epidemiology:
The focus is on studies testing ozone exposure.
5-65
-------
Table Annex 5-1 (Continued): Scientific considerations for evaluating the strength
of inference from studies on the health effects of
ozone.
Exposure Assessment or Assignment
Controlled Human Exposure:
For this assessment, the focus is on studies that use ozone concentrations <0.4 ppm. Studies that use higher
exposure concentrations may provide information relevant to biological plausibility, dosimetry, or inter-species
variation. Studies should have well-characterized pollutant concentration, temperature, and relative humidity and/or
have measures in place to adequately control the exposure conditions. Preference is given to balanced crossover or
parallel design studies which include control exposures (e.g., to clean filtered air). Study subjects should be
randomly exposed without knowledge of the exposure condition. Method of exposure (e.g., chamber, facemask,
etc.) should be specified and activity level of subjects during exposures should be well characterized.
Animal Toxicology:
For this assessment, the focus is on studies that use ozone concentrations <2 ppm. Studies that use higher
exposure concentrations may provide information relevant to biological plausibility, dosimetry, or inter-species
variation. Studies should characterize pollutant concentration, temperature, and relative humidity and/or have
measures in place to adequately control the exposure conditions. The focus is on inhalation exposure.
Noninhalation exposure experiments (i.e., intratracheal instillation [IT]) are informative for size fractions that cannot
penetrate the airway of a study animal and may provide information relevant to biological plausibility and dosimetry.
In vitro studies may be included if they provide mechanistic insight or examine similar effects as in vivo studies but
are generally not included. All studies should include exposure control groups (e.g., clean filtered air).
Epidemiology:
Of primary relevance are relationships of health effects with the ambient component of ozone exposure. However,
information about ambient exposure rarely is available for individual subjects; most often, inference is based on
ambient concentrations. Studies that compare exposure assessment methods are considered to be particularly
informative. Inference is stronger when the duration or lag of the exposure metric corresponds with the time course
for physiological changes in the outcome (e.g., up to a few days for symptoms) or latency of disease (e.g., several
years for cancer).
Ambient ozone concentration tends to have low spatial heterogeneity at the urban scale, except near roads where
ozone concentration is lower because ozone reacts with emitted nitric oxide. For studies involving individuals with
near-road or on-road exposures to ozone in which ambient ozone concentrations are more spatially heterogeneous
and relationships between personal exposures and ambient concentrations are potentially more variable, validated
methods that capture the extent of variability for the epidemiologic study design (temporal vs. spatial contrasts) and
location carry greater weight.
Fixed-site measurements, whether averaged across multiple monitors or assigned from the nearest or single
available monitor, typically have smaller biases and smaller reductions in precision compared with spatially
heterogeneous air pollutants. Concentrations reported from fixed-site measurements can be informative if correlated
with personal exposures, closely located to study subjects, highly correlated across monitors within a location, or
combined with time-activity information.
Atmospheric models may be used for exposure assessment in place of or to supplement ozone measurements in
epidemiologic analyses. For example, grid-scale models (e.g., CMAQ) that represent ozone exposure over relatively
large spatial scales (e.g., typically greater than 4- * 4-km grid size) often do provide adequate spatial resolution to
capture acute ozone peaks that influence short-term health outcomes. Uncertainty in exposure predictions from
these models is largely influenced by model formulations and the quality of model input data pertaining to precursor
emissions or meteorology, which tends to vary on a study-by-study basis.
In studies of short-term exposure, temporal variability of the exposure metric is of primary interest. For long-term
exposures, models that capture within-community spatial variation in individual exposure may be given more weight
for spatially variable ambient ozone. Given the low spatial variability of ozone at the urban scale, exposure
measurement error typically causes health effect estimates to be underestimated for studies of either short-term or
long-term exposure. Biases and decreases in the precision of the association (i.e., wider 95% CIs) tend to be small.
Even when spatial variability is higher near roads, the reduction in ozone exposure would cause the exposure to be
overestimated at a monitor distant from the road or when averaged across a model grid cell, so that health effects
would likely be underestimated.
5-66
-------
Table Annex 5-1 (Continued): Scientific considerations for evaluating the strength
of inference from studies on the health effects of
ozone.
Outcome Assessment/Evaluation
Controlled Human Exposure:
Endpoints should be assessed in the same manner for control and ozone exposure groups (e.g., time after
exposure, methods, endpoint evaluator) using valid, reliable methods. Blinding of endpoint evaluators is ideal,
especially for qualitative endpoints (e.g., histopathology). For each experiment and each experimental group,
including controls, precise details of all procedures carried out should be provided including how, when, and where.
Time of the endpoint evaluations is a key consideration that will vary depending on endpoint evaluated. Endpoints
should be assessed at time points that are appropriate for the research questions.
Animal Toxicology:
Endpoints should be assessed in the same manner for control and exposure groups (e.g., time after exposure,
methods, endpoint evaluator) using valid, reliable methods. Blinding of endpoint evaluators is ideal, especially for
qualitative endpoints (e.g., histopathology). For each experiment and each experimental group, including controls,
precise details of all procedures carried out should be provided including how, when, and where. Time of the
endpoint evaluations is a key consideration that will vary depending on endpoint evaluated. Endpoints should be
assessed at time points that are appropriate for the research questions.
Epidemiology:
Inference is stronger when outcomes are assessed or reported without knowledge of exposure status. Knowledge
of exposure status could produce artefactual associations. Confidence is greater when outcomes assessed by
interview, self-report, clinical examination, or analysis of biological indicators are defined by consistent criteria and
collected by validated, reliable methods. Independent, clinical assessment is valuable for outcomes such as lung
function or incidence of disease, but report of physician diagnosis has shown good reliability.15 When examining
short-term exposures, evaluation of the evidence focuses on specific lags based on the evidence presented in
individual studies. Specifically, the following hierarchy is used in the process of selecting results from individual
studies to assess in the context of results across all studies for a specific health effect or outcome:
• Distributed lag models;
• Multiple days (e.g., 0-2) are averaged;
• Effect estimates are presented for lag days selected a priori by the study authors; or
• If a study focuses on only a series of individual lag days, expert judgment is applied to select the
appropriate result to focus on considering the time course for physiologic changes for the health effect or
outcome being evaluated.
When health effects of long-term exposure are assessed by acute events such as symptoms or hospital
admissions, inference is strengthened when results are adjusted for short-term exposure. Validated questionnaires
for subjective outcomes such as symptoms are regarded to be reliable,0 particularly when collected frequently and
not subject to long recall. For biological samples, the stability of the compound of interest and the sensitivity and
precision of the analytical method is considered. If not based on knowledge of exposure status, errors in outcome
assessment tend to bias results toward the null.
Potential Copollutant Confounding
Controlled Human Exposure:
Exposure should be well characterized to evaluate independent effects of ozone.
Animal Toxicology:
Exposure should be well characterized to evaluate independent effects of ozone.
5-67
-------
Table Annex 5-1 (Continued): Scientific considerations for evaluating the strength
of inference from studies on the health effects of
ozone.
Epidemiology:
Not accounting for potential copollutant confounding can produce artifactual associations; thus, studies that
examine copollutant confounding carry greater weight. The predominant method is copollutant modeling
(i.e., two-pollutant models), which is especially informative when correlations are not high. However, when
correlations are high (r> 0.7), such as those often encountered for UFP and other traffic-related copollutants,
copollutant modeling is less informative. Although the use of single-pollutant models to examine the association
between ozone and a health effect or outcome are informative, studies ideally should also include copollutant
analyses. Copollutant confounding is evaluated on an individual study basis considering the extent of correlations
observed between the copollutant and ozone, and relationships observed with ozone and health effects in
copollutant models.
Other Potential Confounding Factors'1
Controlled Human Exposure:
Preference is given to studies using experimental and control groups that are matched for individual level
characteristics (e.g., race/ethnicity, sex, body weight, smoking history, age) and time varying factors (e.g., seasonal
and diurnal patterns).
Animal Toxicology:
Preference is given to studies using experimental and control groups that are matched for individual-level
characteristics (e.g., strain, sex, body weight, litter size, feed and water consumption) and time varying factors
(e.g., seasonal and diurnal patterns).
Epidemiology:
Factors are considered to be potential confounders if demonstrated in the scientific literature to be related to health
effects and correlated with ozone. Not accounting for confounders can produce artifactual associations; thus,
studies that statistically adjust for multiple factors or control for them in the study design are emphasized. Less
weight is placed on studies that adjust for factors that mediate the relationship between ozone and health effects,
which can bias results toward the null. Confounders vary according to study design, exposure duration, and health
effect and may include, but are not limited to the following:
• Short-term exposure studies: Meteorology, day of week, season, medication use, allergen exposure, and
long-term temporal trends.
• Long-term exposure studies: Socioeconomic status, race, age, medication use, smoking status, stress,
noise, and occupational exposures.
Statistical Methodology
Controlled Human Exposure:
Statistical methods should be described clearly and appropriate for the study design and research question
(e.g., correction for multiple comparisons). Generally, statistical significance is used to evaluate the findings of
controlled human exposure studies. However, consistent trends are also informative. Detection of statistical
significance is influenced by a variety of factors including, but not limited to, the size of the study, exposure and
outcome measurement error, and statistical model specifications. Sample size is not a criterion for exclusion;
ideally, the sample size should provide adequate power to detect hypothesized effects (e.g., sample sizes less than
three are considered less informative). Because statistical tests have limitations, consideration is given to both
trends in data and reproducibility of results.
5-68
-------
Table Annex 5-1 (Continued): Scientific considerations for evaluating the strength
of inference from studies on the health effects of
ozone.
Animal Toxicology:
Statistical methods should be described clearly and appropriate for the study design and research question
(e.g., correction for multiple comparisons). Generally, statistical significance is used to evaluate the findings of
animal toxicological studies. However, consistent trends are also informative. Detection of statistical significance is
influenced by a variety of factors including, but not limited to, the size of the study, exposure and outcome
measurement error, and statistical model specifications. Sample size is not a criterion for exclusion; ideally, the
sample size should provide adequate power to detect hypothesized effects (e.g., sample sizes less than three are
considered less informative). Because statistical tests have limitations, consideration is given to both trends in data
and reproducibility of results.
Epidemiology:
Multivariable regression models that include potential confounding factors are emphasized. However, multipollutant
models (more than two pollutants) are considered to produce too much uncertainty due to copollutant collinearity to
be informative. Models with interaction terms aid in evaluating potential confounding as well as effect modification.
Sensitivity analyses with alternate specifications for potential confounding inform the stability of findings and aid in
judgments of the strength of inference from results. In the case of multiple comparisons, consistency in the pattern
of association can increase confidence that associations were not found by chance alone. Statistical methods that
are appropriate for the power of the study carry greater weight. For example, categorical analyses with small
sample sizes can be prone to bias results toward or away from the null. Statistical tests such as f-tests and
chi-squared tests are not considered sensitive enough for adequate inferences regarding ozone-health effect
associations. For all methods, the effect estimate and precision of the estimate (i.e., width of 95% CI) are more
important considerations rather than statistical significance.
aU.S. EPA (2008).
bMuraia et al. (2014): Weakley et al. (2013): Yang et al. (2011): Heckbert et al. (2004): Barr et al. (2002): Muhaiarine
et al. (1997): Toren et al. (1993).
cBurnev et al. (1989).
5-69
-------
5.4 References
ADA (American Diabetes Association). (2014). Diagnosis and classification of diabetes mellitus.
Diabetes Care 37 Suppl 1: S81-S90. http://dx.doi.org/10.2337/dcl4-SQ81
Akcilar. R; Akcer. S; Simsek. H; Akcilar. A; Bavat. Z; Gene. O. (2015). The effect of ozone on blood
pressure in DOCA-salt-induced hypertensive rats. International Journal of Clinical and
Experimental Medicine 8: 12783-12791.
Alamri. BN; Shin. K; Chappe. V; Anini. Y. (2016). The role of ghrelin in the regulation of glucose
homeostasis [Review]. Hormone Molecular Biology and Clinical Investigation 26: 3-11.
http://dx.doi.org/10.1515/hmbci-2016-0018
Alberti. KG: Eckel. RH: Grundy. SM; Zimmet. PZ; Cleeman. JI; Donato. KA; Fruchart. JC: James.
WP: Loria. CM: Smith. SC. (2009). Harmonizing the metabolic syndrome: A joint interim
statement of the International Diabetes Federation Task Force on Epidemiology and Prevention;
National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation;
International Atherosclerosis Society; and International Association for the Study of Obesity.
Circulation 120: 1640-1645. http://dx.doi.org/10.1161/CIRCULATIONAHA.109.192644
Arkan. MC: Hevener. AL; Greten. FR; Maeda. S: Li. ZW: Long. JM; Wvnshaw-Boris. A: Poli. G:
Olefskv. J: Karin. M. (2005). IKK-beta links inflammation to obesity-induced insulin resistance.
Nat Med 11: 191-198. http://dx.doi.org/10.1038/nml 185
Barr. RG: Herbstman. J: Speizer. FE; Camargo. CA. Jr. (2002). Validation of self-reported chronic
obstructive pulmonary disease in a cohort study of nurses. Am J Epidemiol 155: 965-971.
http://dx.doi.org/10.1093/aie/155.lQ.965
Bass. V: Gordon. CJ: Jarema. KA: Macphail. RC: Cascio. WE: Phillips. PM; Ledbetter. AD:
Schladweiler. MC: Andrews. D; Miller. D; Doerfler. PL: Kodavanti. UP. (2013). Ozone induces
glucose intolerance and systemic metabolic effects in young and aged brown Norway rats. Toxicol
Appl Pharmacol 273: 551-560. http://dx.doi.Org/10.1016/i.taap.2013.09.029
Boron. W: Boulpaep. E. (2017). Medical physiology. In Medical Physiology (3rd ed.). Philadelphia,
PA: Elsevier.
Burnev. PG: Laitinen. LA: Perdrizet. S: Huckauf. H: Tattersfield. AE: Chinn. S: Poisson. N: Heeren.
A: Britton. JR; Jones. T. (1989). Validity and repeatability of the IUATLD (1984) Bronchial
Symptoms Questionnaire: an international comparison. Eur Respir J 2: 940-945.
Chaker. L: Ligthart. S: Korevaar. TI: Hofman. A: Franco. OH: Peeters. RP: Dehghan. A. (2016).
Thyroid function and risk of type 2 diabetes: a population-based prospective cohort study. BMC
Med 14(150): 1-8. http://dx.doi.org/10.1186/sl2916-016-0693-4
Chen. Z: Salam. MT: Toledo-Corral. C: Watanabe. RM: Xiang. AH: Buchanan. TA: Habre. R:
Bastain. TM: Lurmann. F: Wilson. JP: Trigo. E: Gilliland. FD. (2016). Ambient air pollutants have
adverse effects on insulin and glucose homeostasis in Mexican Americans. Diabetes Care 39: 547-
554. http://dx.doi.org/10.2337/dcl5-1795
Chuang. KJ: Yan. YH: Cheng. TJ. (2010). Effect of air pollution on blood pressure, blood lipids, and
blood sugar: A population-based approach. J Occup Environ Med 52: 258-262.
http://dx.doi.org/10.1097/JQM.0b013e3181ceff7a
5-70
-------
Chuang. KJ; Yan. YH; Chiu. SY; Cheng. TJ. (2011). Long-term air pollution exposure and risk factors
for cardiovascular diseases among the elderly in Taiwan. Occup Environ Med 68: 64-68.
http://dx.doi.Org/10.l 136/oem.2009.052704
Ciccarelli. M; Santulli. G; Pascale. V; Trimarco. B; Iaccarino. G. (2013). Adrenergic receptors and
metabolism: role in development of cardiovascular disease [Review]. Front Physiol 4: 265.
http://dx.doi.org/10.3389/fbhvs.2013.00265
Clemons. GK: Garcia. JF. (1980). Changes in thyroid function after short-term ozone exposure in rats.
J Environ Pathol Toxicol 4: 359-369.
Crouse. PL: Peters. PA: Hvstad. P; Brook. JR: van Donkelaar. A: Martin. RV: Villeneuve. PJ; Jerrett.
M: Goldberg. MS: Pope. CA: Brauer. M: Brook. RD: Robichaud. A: Menard. R: Burnett. RT.
(2015). Ambient PM 2.5, O 3, and NO 2 exposures and associations with mortality over 16 years of
follow-up in the Canadian Census Health and Environment Cohort (CanCHEC). Environ Health
Perspect 123: 1180-1186. http://dx.doi.org/10.1289/ehp.1409276
Dales. RE: Cakmak. S: Vidal. CB; Rubio. MA. (2012). Air pollution and hospitalization for acute
complications of diabetes in Chile. Environ Int 46: 1-5.
http://dx.doi.Org/10.1016/i.envint.2012.05.002
Dong. G: Oian. Z: Liu. MM: Wang. D: Ren. W: Flick. LH: Fu. J: Wang. J: Chen. W: Simckes. M:
Trevathan. E. (2014). Ambient air pollution and the prevalence of obesity in Chinese children: The
seven northeastern cities study. Obesity (Silver Spring) 22: 795-800.
http://dx.doi.org/10.1002/obv.20198
Farrai. AK; Hazari. MS: Winsett. DW: Kulukulualani. A: Carll. AP; Havkal-Coates. N: Lamb. CM:
Lappi. E; Terrell. D; Cascio. WE: Costa. PL. (2012). Overt and latent cardiac effects of ozone
inhalation in rats: evidence for autonomic modulation and increased myocardial vulnerability.
Environ Health Perspect 120: 348-354. http://dx.doi.org/10.1289/ehp.1104244
Farrai. AK: Malik. F; Havkal-Coates. N: Walsh. L; Winsett. D; Terrell. D; Thompson. LC: Cascio.
WE: Hazari. MS. (2016). Morning N02 exposure sensitizes hypertensive rats to the cardiovascular
effects of same day 03 exposure in the afternoon. Inhal Toxicol 28: 170-179.
http://dx.doi.org/10.3109/08958378.2016.1148Q88
Friihbeck. G: Mendez-Gimenez. L; Fernandez-Formoso. JA; Fernandez. S: Rodriguez. A. (2014).
Regulation of adipocyte lipolysis [Review]. Nutr Res Rev 27: 63-93.
http://dx.doi.org/10.1017/S0954422414000Q2X
Gackiere. F; Saliba. L; Baude. A: Bosler. O; Strube. C. (2011). Ozone inhalation activates stress-
responsive regions of the CNS. JNeurochem 117: 961-972. http://dx.doi.0rg/lO.l 111/i. 1471-
4159.2011.07267.x
Gordon. CJ: Jarema. KA: Lehmann. J. R.; Ledbetter. AD: Schladweiler. MC: Schmid. JE; Ward. WO:
Kodavanti. UP: Nvska. A: Macphail. RC. (2013). Susceptibility of adult and senescent Brown
Norway rats to repeated ozone exposure: an assessment of behavior, serum biochemistry and
cardiopulmonary function. Inhal Toxicol 25: 141-159.
http://dx.doi.org/10.3109/08958378.2013.764946
Gordon. CJ: Phillips. PM: Johnstone. AFM: Beaslev. TE: Ledbetter. AD: Schladweiler. MC: Snow.
SJ: Kodavanti. UP. (2016). Effect of high-fructose and high-fat diets on pulmonary sensitivity,
motor activity, and body composition of brown Norway rats exposed to ozone. Inhal Toxicol 28:
203-215. http://dx.doi.org/10.3109/08958378.2015.113473Q
5-71
-------
Gordon. CJ; Phillips. PM; Johnstone. AFM; Schmid. J; Schladweiler. MC; Ledbetter. A; Snow. SJ;
Kodavanti. UP. (2017a). Effects of maternal high-fat diet and sedentary lifestyle on susceptibility
of adult offspring to ozone exposure in rats. Inhal Toxicol 29: 239-254.
http://dx.doi.org/10.1080/08958378.2Q17.1342719
Gordon. CJ; Phillips. PM; Ledbetter. A; Snow. SJ; Schladweiler. MC; Johnstone. AF; Kodavanti. UP.
(2017b). Active vs. sedentary lifestyle from weaning to adulthood and susceptibility to ozone in
rats. Am J Physiol Lung Cell Mol Physiol 312: L100-L109.
http://dx.doi.org/10.1152/aiplung.00415.2016
Heckbert. SR; Kooperberg. C; Safford. MM; Psatv. BM; Hsia. J; McTiernan. A; Gaziano. JM;
Frishman. WH; Curb. JD. (2004). Comparison of self-report, hospital discharge codes, and
adjudication of cardiovascular events in the Women's Health Initiative. Am J Epidemiol 160: 1152-
1158. http://dx.doi.org/10.1093/aie/kwh314
Henriquez. AR; Snow. SJ; Schladweiler. MC; Miller. CN; Dve. JA; Ledbetter. AD; Richards. JE;
Hargrove. MM; Williams. WC; Kodavanti. UP. (2018). Beta-2 adrenergic and glucocorticoid
receptor agonists modulate ozone-induced pulmonary protein leakage and inflammation in healthy
and adrenalectomized rats. Toxicol Sci 166: 288-305. http://dx.doi.org/10.1093/toxsci/kfV198
Henriquez. AR; Snow. SJ; Schladweiler. MC; Miller. CN; Dve. JA; Ledbetter. AD; Richards. JE;
Mauge-Lewis. K; Mcgee. MA; Kodavanti. UP. (2017). Adrenergic and glucocorticoid receptor
antagonists reduce ozone-induced lung injury and inflammation. Toxicol Appl Pharmacol 339:
161-171. http://dx.doi.org/10.1016/i.taap.2017.12.006
Hotamisligil. GS. (2017). Inflammation, metaflammation and immunometabolic disorders [Review].
Nature 542: 177-185. http://dx.doi.org/10.1038/nature21363
Jerrett. M; Brook. R; White. LF; Burnett. RT; Yu. J; Su. J; Seto. E; Marshall. J; Palmer. JR;
Rosenberg. L; Coogan. PF. (2017). Ambient ozone and incident diabetes: A prospective analysis in
a large cohort of African American women. Environ Int 102: 42-47.
http://dx.doi.Org/10.1016/i.envint.2016.12.011
Kilkenny. C; Browne. WJ; Cuthill. TC; Emerson. M; Altman. DG. (2010). Improving bioscience
research reporting: The ARRIVE guidelines for reporting animal research [Review]. PLoS Biol 8:
el000412. http://dx.doi.org/10.1371/iournal.pbio.100Q412
Kim. JH; Hong. YC. (2012). GSTM1, GSTT1, and GSTP1 polymorphisms and associations between
air pollutants and markers of insulin resistance in elderly Koreans. Environ Health Perspect 120:
1378-1384. http://dx.doi.org/10.1289/ehp. 1104406
Klimisch. HJ; Andreae. M; Tillmann. U. (1997). A systematic approach for evaluating the quality of
experimental toxicological and ecotoxicological data. Regul Toxicol Pharmacol 25: 1-5.
http://dx.doi.org/10.1006/rtph.1996.1076
Kuo. T; McQueen. A; Chen. TC; Wang. JC. (2015). Regulation of glucose homeostasis by
glucocorticoids. Adv Exp Med Biol 872: 99-126. http://dx.doi.org/10.10Q7/978-l-4939-2895-8 5
Li. M; Qian. Z; Vaughn. M; Boutwell. B; Ward. P; Lu. T. ao; Lin. S; Zhao. Y; Zeng. XW; Liu. RQ;
Qin. XD; Zhu. Y; Chen. W; Dong. GH. (2015). Sex-specific difference of the association between
ambient air pollution and the prevalence of obesity in Chinese adults from a high pollution range
area: 33 Communities Chinese Health Study. Atmos Environ 117: 227-233.
http://dx.doi.Org/10.1016/i.atmosenv.2015.07.029
5-72
-------
Li. W; Dorans. KS; Wilker. EH; Rice. MB; Kloog. I; Schwartz. JD; Koutrakis. P; Coull. BA; Gold.
PR; Meigs. JB; Fox. CS; Mittleman. MA. (2017). Ambient air pollution, adipokines, and glucose
homeostasis: The Framingham Heart Study. Environ Int 111: 14-22.
http://dx.doi.Org/10.1016/i.envint.2017.l 1.010
Malmqvist. E; Elding Larsson. H; Jonsson. I; Rignell-Hvdbom. A; Ivarsson. SA; Tinnerberg. H; Stroh.
E; Rittner. R; Jakobsson. K; Swietlicki. E; Rvlander. L. (2015). Maternal exposure to air pollution
and type 1 diabetes - Accounting for genetic factors. Environ Res 140: 268-274.
http://dx.doi.Org/10.1016/i.envres.2015.03.024
Matthews. PR; Hosker. JP; Rudenski. AS; Navlor. BA; Treacher. DF; Turner. RC. (1985).
Homeostasis model assessment: Insulin resistance and beta-cell function from fasting plasma
glucose and insulin concentrations in man. Diabetologia 28: 412-419.
http://dx.doi.org/10.1007/BF0028Q883
Mautz. WJ; Bufalino. C. (1989). Breathing pattern and metabolic rate responses of rats exposed to
ozone. Respir Physiol Neurobiol 76: 69-77.
Meier. U; Gressner. AM. (2004). Endocrine regulation of energy metabolism: review of
pathobiochemical and clinical chemical aspects of leptin, ghrelin, adiponectin, and resistin
[Review]. Clin Chem 50: 1511-1525. http://dx.doi.org/10.1373/clinchem.2004.032482
Miller. DB; Ghio. AJ; Karolv. ED; Bell. LN; Snow. SJ; Madden. MC; Soukup. J; Cascio. WE;
Gilmour. MI; Kodavanti. UP. (2016a). Ozone exposure increases circulating stress hormones and
lipid metabolites in humans. Am J Respir Crit Care Med 193: 1382-1391.
http://dx.doi.Org/10.l 164/rccm.201508-1599QC
Miller. DB; Karolv. ED; Jones. JC; Ward. WO; Vallanat. BP; Andrews. PL; Schladweiler. MC;
Snow. SJ; Bass. VL; Richards. JE; Ghio. AJ; Cascio. WE; Ledbetter. AP; Kodavanti. UP. (2015).
Inhaled ozone (03)-induces changes in serum metabolomic and liver transcriptomic profiles in rats.
Toxicol Appl Pharmacol 286: 65-79. http://dx.doi.Org/10.1016/i.taap.2015.03.025
Miller. PB; Snow. SJ; Henriquez. A; Schladweiler. MC; Ledbetter. AP; Richards. JE; Andrews. PL;
Kodavanti. UP. (2016b). Systemic metabolic derangement, pulmonary effects, and insulin
insufficiency following subchronic ozone exposure in rats. Toxicol Appl Pharmacol 306: 47-57.
http://dx.doi.Org/10.1016/i.taap.2016.06.027
Miller. PB; Snow. SJ; Schladweiler. MC; Richards. JE; Ghio. AJ; Ledbetter. AP; Kodavanti. UP.
(2016c). Acute ozone-induced pulmonary and systemic metabolic effects are diminished in
adrenalectomized rats. Toxicol Sci 150: 312-322. http://dx.doi.org/10.1093/toxsci/kfV331
Muhaiarine. N; Mustard. C; Roos. LL; Young. TK; Gelskev. PE. (1997). Comparison of survey and
physician claims data for detecting hypertension. J Clin Epidemiol 50: 711-718.
http://dx.doi.org/10.1016/50895-4356(97)00019-X
Murgia. N; Brisman. J; Claesson. A; Muzi. G; Olin. AC; Toren. K. (2014). Validity of a questionnaire-
based diagnosis of chronic obstructive pulmonary disease in a general population-based study.
BMC Pulm Med 14: 49. http://dx.doi.org/10.1186/1471-2466-14-49
Myers. MG; Cowley. MA; Miinzberg. H. (2008). Mechanisms of leptin action and leptin resistance
[Review]. Annu Rev Physiol 70: 537-556.
http://dx.doi.Org/10.l 146/annurev.phvsiol.70.113006.100707
Nadal. A; Alonso-Magdalena. P; Soriano. S; Quesada. I; Ropero. AB. (2009). The pancreatic beta-cell
as a target of estrogens and xenoestrogens: Implications for blood glucose homeostasis and
diabetes [Review]. Mol Cell Endocrinol 304: 63-68. http://dx.doi.Org/10.1016/i.mce.2009.02.016
5-73
-------
Nicolaides. NC; Kvratzi. E; Lamprokostopoulou. A; Chrousos. GP; Charmandari. E. (2015). Stress,
the stress system and the role of glucocorticoids [Review]. Neuroimmunomodulation 22: 6-19.
http://dx.doi.org/10.1159/00Q362736
O'Rourke. RW. (2009). Inflammation in obesity-related diseases [Review]. Surgery 145: 255-259.
http://dx.doi.Org/10.1016/i.surg.2008.08.038
Pasauali. R. (2012). The hypothalamicpituitaryadrenal axis and sex hormones in chronic stress and
obesity: pathophysiological and clinical aspects. Ann N Y Acad Sci 1264.
http://dx.doi.org/10.1111/i. 1749-6632.2012.06569.X
Ramot. Y: Kodavanti. UP; Kissling. GE: Ledbetter. AD: Nvska. A. (2015). Clinical and pathological
manifestations of cardiovascular disease in rat models: the influence of acute ozone exposure. Inhal
Toxicol 27: 26-38. http://dx.doi.org/10.3109/08958378.2014.954168
Rav. A: Huisman. MY: Tamsma. JT; Writing-group. Ra: van Asten. J: Bingen. BO: Breeders. EA;
Hoogeveen. ES: van Hout. F: Kwee. VA: Laman. B: Malgo. F: Mohammadi. M: Niienhuis. M:
Rijkee. M; van Tellingen. MM: Tromp. M; Tummers. O; de Vries. L. (2009). The role of
inflammation on atherosclerosis, intermediate and clinical cardiovascular endpoints in type 2
diabetes mellitus [Review]. Eur J Intern Med 20: 253-260.
http://dx.doi.Org/10.1016/i.eiim.2008.07.008
Renzi. M; Cerza. F; Gariazzo. C: Agabiti. N: Cascini. S: Di Domenicantonio. R; Davoli. M;
Forastiere. F: Cesaroni. G. (2017). Air pollution and occurrence of type 2 diabetes in a large cohort
study. Environ Int 112: 68-76. http://dx.doi.Org/10.1016/i.envint.2017.12.007
Roonev. AA; Boyles. AL; Wolfe. MS: Bucher. JR; Thaver. KA. (2014). Systematic review and
evidence integration for literature-based environmental health science assessments. Environ Health
Perspect 122: 711-718. http://dx.doi.org/10.1289/ehp.1307972
Roos. A: Bakker. STL: Links. TP: Gans. ROB: Wolffenbutte. BHR. (2007). Thyroid function is
associated with components of the metabolic syndrome in euthyroid subjects. 92: 491-496.
http://dx.doi.org/10.1210/ic.2006-1718
Sun. L: Liu. C: Xu. X: Ying. Z; Maiseveu. A: Wang. A: Allen. K: Lewandowski. RP: Bramble. LA:
Morishita. M; Wagner. JG: Dvonch. JT: Sun. Z; Yan. X: Brook. RD: Raiagopalan. S: Harkema. JR:
Sun. Q: Fan. Z. (2013). Ambient fine particulate matter and ozone exposures induce inflammation
in epicardial and perirenal adipose tissues in rats fed a high fructose diet. Part Fibre Toxicol 10: 43.
http://dx.doi.org/10.1186/1743-8977-10-43
Theis. WS: Andringa. KK; Millender-Swain. T; Dickinson. DA: Postlethwait. EM: Bailey. SM.
(2014). Ozone inhalation modifies the rat liver proteome. 2: 52-60.
http://dx.doi.Org/doi:10.1016/i.redox.2013.l 1.006
Thomson. EM; Pal. S; Guenette. J; Wade. MG; Atlas. E; Hollowav. AC; Williams. A; Vincent. R.
(2016). Ozone inhalation provokes glucocorticoid-dependent and -independent effects on
inflammatory and metabolic pathways. Toxicol Sci 152: 17-28.
http://dx.doi.org/10.1093/toxsci/kfw061
Thomson. EM; Pilon. S; Guenette. J; Williams. A; Hollowav. AC. (2018). Ozone modifies the
metabolic and endocrine response to glucose: Reproduction of effects with the stress hormone
corticosterone. Toxicol Appl Pharmacol 342: 31-38. http://dx.doi.Org/10.1016/i.taap.2018.01.020
Thomson. EM; Vladisavlievic. D; Mohottalage. S; Kumarathasan. P; Vincent. R. (2013). Mapping
acute systemic effects of inhaled particulate matter and ozone: multiorgan gene expression and
glucocorticoid activity. Toxicol Sci 135: 169-181. http://dx.doi.org/10.lQ93/toxsci/kftl37
5-74
-------
Toren. K; Brisman. J; Jarvholm. B. (1993). Asthma and asthma-like symptoms in adults assessed by
questionnaires: A literature review [Review]. Chest 104: 600-608.
http://dx.doi.Org/10.1378/chest.104.2.600
Turner. MC; Jerrett. M; Pope. A. Ill; Krewski. D; Gapstur. SM; Diver. WR; Beckerman. BS;
Marshall. JD; Su. J; Crouse. PL; Burnett. RT. (2016). Long-term ozone exposure and mortality in a
large prospective study. Am J Respir Crit Care Med 193: 1134-1142.
http://dx.doi.Org/10.l 164/rccm.201508-1633QC
U.S. EPA (U.S. Environmental Protection Agency). (1991). Guidelines for developmental toxicity risk
assessment (pp. 1-71). (EPA/600/FR-91/001). Washington, DC: U.S. Environmental Protection
Agency, Risk Assessment Forum, http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=23162
U.S. EPA (U.S. Environmental Protection Agency). (1996). Guidelines for reproductive toxicity risk
assessment (pp. 1-143). (EPA/630/R-96/009). Washington, DC: U.S. Environmental Protection
Agency, Risk Assessment Forum, https://www.epa.gov/sites/production/files/2014-
11/documents/guidelines repro toxicitv.pdf
U.S. EPA (U.S. Environmental Protection Agency). (1998). Guidelines for neurotoxicity risk
assessment [EPA Report] (pp. 1-89). (ISSN 0097-6326 EISSN 2167-2520 EPA/630/R-95/00IF).
Washington, DC: U.S. Environmental Protection Agency, Risk Assessment Forum.
http://www.epa.gov/risk/guidelines-neurotoxicitv-risk-assessment
U.S. EPA (U.S. Environmental Protection Agency). (2005). Guidelines for carcinogen risk assessment
[EPA Report]. (EPA/630/P-03/00IB). Washington, DC: U.S. Environmental Protection Agency,
Risk Assessment Forum, https://www.epa.gov/sites/production/files/2013-
09/documents/cancer guidelines final 3-25-05.pdf
U.S. EPA (U.S. Environmental Protection Agency). (2008). Integrated science assessment for sulfur
oxides: Health criteria [EPA Report]. (EPA/600/R-08/047F). Research Triangle Park, NC: U.S.
Environmental Protection Agency, Office of Research and Development, National Center for
Environmental Assessment- RTP. http://cfbub.epa.gov/ncea/cfm/recordisplav.cfm?deid=198843
U.S. EPA (U.S. Environmental Protection Agency). (2013a). Integrated science assessment for ozone
and related photochemical oxidants [EPA Report]. (EPA/600/R-10/076F). Research Triangle Park,
NC: U.S. Environmental Protection Agency, Office of Research and Development, National Center
for Environmental Assessment-RTP Division.
http://cfpub.epa.gov/ncea/isa/recordisplav.cfm?deid=247492
U.S. EPA (U.S. Environmental Protection Agency). (2013b). Toxicological review of
trimethylbenzenes (CASRN 25551-13-7, 95-63-6, 526-73-8, and 108-67-8) in support of summary
information on the Integrated Risk Information System (IRIS): revised external review draft [EPA
Report]. (EPA/635/R13/171a). Washington, D.C.: U.S. Environmental Protection Agency,
National Center for Environmental Assessment.
http://vosemite.epa.gov/sab/SABPRODUCT.NSF/b5d8alce9b07293485257375007012b7/eele28Q
e77586de985257b65005d37e7!OpenDocument
U.S. EPA (U.S. Environmental Protection Agency). (2015). Preamble to the integrated science
assessments [EPA Report]. (EPA/600/R-15/067). Research Triangle Park, NC: U.S. Environmental
Protection Agency, Office of Research and Development, National Center for Environmental
Assessment, RTP Division. https://cfpub.epa.gov/ncea/isa/recordisplav.cfm?deid=310244
Uchivama. I: Simomura. Y; Yokovama. E. (1986). Effects of acute exposure to ozone on heart rate
and blood pressure of the conscious rat. Environ Res 41: 529-537.
5-75
-------
Uchiyama. I; Yokovama. E. (1989). Effects of short- and long-term exposure to ozone on heart rate
and blood pressure of emphysematous rats. Environ Res 48: 76-86.
http://dx.doi.org/10.1016/S0Q 13-9351(89)80087-8
Ulrich-Lai. YM; Herman. JP. (2009). Neural regulation of endocrine and autonomic stress responses.
Nat Rev Neurosci 10: 397-409. http://dx.doi.org/10.1038/nrn2647
van Beek. L: Lips. MA: Visser. A: Piil. H: Ioan-Facsinav. A: Toes. R; Berends. FJ: Willems van Diik.
K: Koning. F: van Harmelen. V. (2014). Increased systemic and adipose tissue inflammation
differentiates obese women with T2DM from obese women with normal glucose tolerance.
Metabolism 63: 492-501. http://dx.doi.Org/10.1016/i.metabol.2013.12.002
Vella. RE: Pillon. NJ: Zarrouki. B: Croze. ML: Koppe. L: Guichardant. M: Pesenti. S: Chauvin. MA:
Rieusset. J: Geloen. A: Soulage. CO. (2014). Ozone exposure triggers insulin resistance through
muscle c-Jun N-terminal Kinases (JNKs) activation. Diabetes 64: 1011-1024.
http://dx.doi.org/10.2337/dbl3-1181
von Elm. E; Altman. DG: Egger. M; Pocock. SJ: Gotzschc. PC: Vandenbroucke. JP. (2007). The
strengthening the reporting of observational studies in epidemiology (strobe) statement: guidelines
for reporting observational studies [Review]. PLoS Med 4: e296.
http://dx.doi.org/10.1371/iournal.pmed.0040296
Wagner. JG: Allen. K; Yang. HY: Nan. B; Morishita. M; Mukheriee. B; Dvonch. JT; Spino. C: Fink.
GD: Raiagopalan. S: Sun. O: Brook. RD: Harkema. JR. (2014). Cardiovascular depression in rats
exposed to inhaled particulate matter and ozone: effects of diet-induced metabolic syndrome.
Environ Health Perspect 122: 27-33. http://dx.doi.org/10.1289/ehp.1307085
Wallace. TM; Lew. JC: Matthews. DR. (2004). Use and abuse of HOMA modeling [Review].
Diabetes Care 27: 1487-1495. http://dx.doi.Org/10.2337/diacare.27.6.1487
Weakley. J: Webber. MP: Ye. F: Zeig-Owens. R: Cohen. HW: Hall. CB: Kelly. K: Prezant. DJ.
(2013). Agreement between obstructive airways disease diagnoses from self-report questionnaires
and medical records. Prev Med 57: 38-42. http://dx.doi.Org/10.1016/i.ypmed.2013.04.001
Wellen. KE: Hotamisligil. GS. (2005). Inflammation, stress, and diabetes [Review]. J Clin Invest 115:
1111-1119. http://dx.doi.org/10.1172/JCI25102
White. LF; Jerrett. M; Yu. J: Marshall. JD; Rosenberg. L; Coogan. PF. (2016). Ambient air pollution
and 16-year weight change in African-American women. Am J Prev Med 51: e99el05.
http://dx.doi.Org/10.1016/i.amepre.2016.03.007
Yang. BY: Qian. ZM; Li. S: Fan. S: Chen. G: Svberg. KM: Xian. H; Wang. SO: Ma. H; Chen. DH;
Yang. M: Liu. KK: Zeng. XW: Hu. LW: Guo. Y: Dong. GH. (2018). Long-term exposure to
ambient air pollution (including PM1) and metabolic syndrome: The 33 Communities Chinese
Health Study (33CCHS). Environ Res 164: 204-211.
http://dx.doi.Org/10.1016/i.envres.2018.02.029
Yang. CL: To. T: Fotv. RG: Stieb. DM: Dell. SD. (2011). Verifying a questionnaire diagnosis of
asthma in children using health claims data. BMC Pulm Med 11. http://dx.doi.org/10.1186/1471-
2466-11-52
Ying. Z: Allen. K: Zhong. J: Chen. M: Williams. KM: Wagner. JG: Lewandowski. R: Sun. O:
Raiagopalan. S: Harkema. JR. (2016). Subacute inhalation exposure to ozone induces systemic
inflammation but not insulin resistance in a diabetic mouse model. Inhal Toxicol 28: 155-163.
http://dx.doi.org/10.3109/08958378.2016.11468Q8
5-76
-------
Yu. XH; Zhang. DW; Zheng. XL; Tang. CK. (2019). Cholesterol transport system: An integrated
cholesterol transport model involved in atherosclerosis [Review]. Prog Lipid Res 73: 65-91.
http://dx.doi.Org/10.1016/i.plipres.2018.12.002
Zellner. LC; Brundage. KM; Hunter. DP; Dev. RD. (2011). Early postnatal ozone exposure alters rat
nodose and jugular sensory neuron development. Toxicol Environ Chem 93: 2055-2071.
http://dx.doi.org/10.1080/02772248.2011.61Q882
Zhong. J; Allen. K; Rao. X; Ying. Z; Braunstein. Z; Kankanala. SR; Xia. C; Wang. X; Bramble. LA;
Wagner. JG; Lewandowski. R; Sun. Q; Harkema. JR; Raiagopalan. S. (2016). Repeated ozone
exposure exacerbates insulin resistance and activates innate immune response in genetically
susceptible mice. Inhal Toxicol 28: 383-392. http://dx.doi.Org/10.1080/08958378.2016.l 179373
5-77
-------
APPENDIX 6 HEALTH EFFECTS —MORTALITY
Summary of Causality Determinations for Short- and l.on^-l erm Ozone
Exposure ami iota! pc of I lie IS \ ;is
del;iiled in 1 lie IVI;ice In ;issessnm ilie n\ ei~ilI e\ idenee. sirenullis ;md limil;ilKins nl' mdi\ idii;il
siudies were e\ ;ilu;iied h;ised mi seienl i l"ie anisideniliniis del;nled in I lie \mie\ I'm- \nnendi\ (¦
Mine del;nls on I lie e;ius;il ri;ime\\iii"k used Id re;ieli iliese an
Shorl-lemi exposure Su
ggeslive o
f. but not sufficient to infer
. ;i c;ius;il relationship
I .ong-lerm exposure Su
ggeslive o
f, but nol sufficient lo infer
. a causal relationship
6.1 Short-Term Ozone Exposure and Mortality
6.1.1 Introduction
The 2013 Integrated Science Assessment for Ozone and Related Photochemical Oxidants (2013
Ozone ISA) concluded there is likely to be a causal relationship between short-term ozone exposure and
total mortality (U.S. EPA. 2013a). which built upon the evidence presented in the 2006 Ozone Air
Quality Criteria Document [AQCD; U.S. EPA (2006)1. This conclusion was supported by a number of
multicity and multicontinent epidemiologic studies that provided evidence of consistent, positive
associations between short-term ozone exposure and mortality in all-year and summer/warm season
analyses and across different averaging times (i.e., 1-hour max, 8-hour max, and 24-hour avg), which
further confirmed the positive associations reported in multicity studies, single-city studies, and
meta-analyses evaluated in previous assessments. The multicity and multicontinent studies evaluated in
the 2013 Ozone ISA also addressed key uncertainties and limitations that remained in the evidence base
for short-term ozone exposure and mortality upon the completion of the 2006 Ozone AQCD. As
summarized below, these studies further informed the relationship between short-term ozone exposure
and cause-specific mortality, the potential confounding effects of copollutants and season, spatial
6-1
-------
heterogeneity in ozone-mortality risk estimates, the timing of mortality effects, and the shape of the
concentration-response (C-R) relationship.
Epidemiologic studies evaluated in the 2013 Ozone ISA expanded upon the evaluation of
associations between short-term ozone exposure and cause-specific mortality through multicity studies,
which previously was limited to primarily single-city studies. These studies provided evidence of
generally consistent, positive associations with both cardiovascular and respiratory mortality in all-year
and summer/warm season analyses. The strong and consistent evidence within and across scientific
disciplines for respiratory morbidity provided coherence and biological plausibility for respiratory
mortality. However, the cardiovascular morbidity evidence supporting cardiovascular mortality was
limited. A limited number of controlled human exposure studies and numerous animal toxicological
studies provided initial evidence supporting a biologically plausible mechanism by which short-term
ozone exposure could lead to cardiovascular endpoints, but there was inconsistency in results between
experimental and epidemiologic studies. Specifically, epidemiologic studies did not consistently
demonstrate positive associations with clinical cardiovascular effects, such as hospital admissions and
emergency department visits.
In the previous ISA, the evaluation of potential confounding of the ozone-mortality relationship
in epidemiologic studies focused on assessing both model specification (e.g., control for
temporal/seasonal trends) and the influence of copollutants on ozone-mortality associations. An
examination of modeling methods indicated that the extent of smoothing used to control for
temporal/seasonal trends (i.e., numbers of degrees of freedom used in time splines) can influence the
magnitude of associations observed. More detailed analyses of potential copollutant confounding focused
on not only particulate matter (PM) size fractions, but also fine particulate matter (PM2 5) components,
and reported that although associations were attenuated in copollutant models with PM in some instances,
overall associations remained positive. However, the assessment of potential copollutant confounding was
complicated by the variability in the correlation between PM and ozone across regions and the small
number of days with both ozone and PM data due to the PM sampling schedule (i.e., every 3rd or 6th
day).
Multicity studies also provided evidence of the geographic pattern of spatial heterogeneity
(i.e., regional and city-to-city) in ozone-mortality risk estimates, with associations largest in magnitude in
the northeastern U.S. A few studies examined whether specific factors, both time-invariant and
time-variant, explained this observed heterogeneity. Examination of the time-invariant factors showed
some evidence that individual- and community-level factors may contribute to spatial heterogeneity of
ozone-mortality associations, including but not limited to, unemployment rate, prevalence of air
conditioning, and indicators of socioeconomic status (SES). Additionally, there was initial evidence that
the time-variant factor of daily temperature modifies ozone effects on mortality, specifically high
temperatures, may increase the risk of ozone-related mortality.
6-2
-------
Lastly, the multicity and multicontinent studies evaluated in the 2013 Ozone ISA provided a more
thorough assessment of the timing of mortality effects after ozone exposure and the C-R relationship.
Across studies there was evidence that the strongest ozone-mortality association, in terms of magnitude
and precision, occurs within the first few days after exposure, within the range of 0-3 days. Additionally,
examination of the C-R relationship between short-term ozone exposure and mortality supported a linear
relationship with no evidence of a threshold below which effects do not occur.
Building off the evidence detailed in the 2013 Ozone ISA, the following sections provide a brief,
integrated evaluation of recent evidence for short-term ozone exposure and mortality. Specifically, the
sections focus on assessing the degree to which newly available studies further characterize the
relationship between short-term ozone exposure and mortality, and the continued evaluation of previously
identified uncertainties and limitations in the evidence base. The studies evaluated in the 2013 Ozone ISA
further expanded upon uncertainties and limitations in the evidence base identified at the completion of
the 2006 Ozone AQCD, specifically the limited evaluation of: the relationship between short-term ozone
exposure and cause-specific mortality, the potential confounding effects of copollutants and season,
heterogeneity in ozone-mortality risk estimates, the timing of mortality effects, and the shape of the C-R
relationship. Recent epidemiologic studies provide additional support to the evidence evaluated in the
2013 Ozone ISA and, in some cases, further address the uncertainties and limitations previously
identified. While this section focuses on epidemiologic studies examining short-term ozone exposure and
mortality, the causality determination draws on the morbidity evidence presented for respiratory (see
Appendix 3) and cardiovascular (see Appendix 4) effects spanning evidence across scientific disciplines
(i.e., animal toxicological, epidemiologic, and controlled human exposure studies) to assess coherence
between the morbidity and mortality evidence and inform biological plausibility for ozone-related
mortality.
6.1.1.1 Population, Exposure, Comparison, Outcome, and Study Design (PECOS)
Tool
The scope of this section is defined by a scoping tool that generally describes the relevant
Population, Exposure, Comparison, Outcome, and Study Design (PECOS). The PECOS tool defines the
parameters and provides a framework to help identify the relevant literature to inform the Ozone ISA.
Because the 2013 Ozone ISA concluded there is likely to be a causal relationship between short-term
ozone exposure and total mortality, the studies evaluated are more limited in scope and targeted towards
study locations that are most informative in addressing the policy-relevant considerations forming the
basis of this section. Therefore, the studies evaluated and subsequently discussed within this section were
included if they satisfied all the components of the following PECOS tool:
6-3
-------
• Population: Any U.S. or Canadian population, including populations or lifestages that might be at
increased risk1
• Exposure: Short-term exposure (on the order of 1 to several days) to ambient concentrations of
ozone
• Comparison: Per unit increase (in ppb)
• Outcome: Mortality
• Study Design: Epidemiologic studies consisting of case-crossover or time-series studies
6.1.2 Biological Plausibility
The preceding appendices characterized evidence to evaluate the biological plausibility by which
short-term ozone exposure may lead to the morbidity effects that are the most common causes of total
(nonaccidental) mortality, specifically respiratory (Appendix 3) and cardiovascular (Appendix 4)
morbidity, which comprise ~9 and -33%, respectively, of total mortality (NHLBI. 2017). This evidence is
derived from animal toxicological, controlled human exposure, and epidemiologic studies. Appendix 3.
Section 3.1.3 characterizes the strong evidence by which ozone exposure could plausibly progress from
initial events to endpoints relevant to the respiratory system, including increases in respiratory emergency
department (ED) visits and hospital admissions for chronic obstructive pulmonary disease (COPD) and
asthma. Together, experimental and epidemiologic studies indicate that short-term ozone exposure can
lead to a progression of respiratory effects that provide support for potential biological pathways by
which short-term ozone exposures could lead to more severe respiratory morbidity outcomes, such as a
respiratory-related ED visits or hospital admissions, and ultimately mortality. Appendix 4. Section 4.1.3
characterizes the available evidence for plausible mechanisms by which ozone exposure could progress
from initial events to endpoints relevant to the cardiovascular system and to population outcomes, such as
ED visits and hospital admissions from cardiovascular disease, particularly ischemic heart disease and
congestive heart failure. However, as depicted in Appendix 4. Section 4.1.3. although some experimental
evidence supports subclinical cardiovascular effects, there is limited evidence of relationships between
short-term ozone exposure and more severe cardiovascular effects, such as ED visits and hospital
admissions (e.g., Section 4.1.4.1 and Section 4.1.5.1). This lack of coherence in results between
experimental and epidemiologic studies contributes to uncertainty in the observed associations between
short-term ozone exposure and mortality because the contribution of cardiovascular mortality to total
mortality is relatively large.
1 A list of considered studies conducted in other geographic locations is available via the HERO database.
6-4
-------
6.1.3
Total (Nonaccidental) Mortality
Relatively few recent studies have been conducted within the U.S. and Canada that examined the
relationship between short-term ozone exposure and total (nonaccidental) mortality since the completion
of the 2013 Ozone ISA. Although these recent multicity studies present new analyses that further
characterize the association between short-term ozone exposure and mortality, most relied on population
and air quality data from previously conducted studies (e.g., the National Morbidity, Mortality, and Air
Pollution Study [NMMAPS]), with only Di et al. (2017a) using more recent air quality data (i.e., since
2000). Additionally, most recent studies continue to use the traditional approach of assigning ozone
exposures using ozone concentrations measured at a single monitor or the average of ozone
concentrations from multiple monitors within some defined geographic location. Of the studies evaluated,
Madrigano et al. (2015) and Di et al. (2017a) used novel exposure assignment methods that allowed for
the inclusion of populations residing in more diverse geographic locations (i.e., not limited to major urban
centers) through kriging (i.e., spatial interpolation) or the use of multiple sources of air quality data
(i.e., land use, chemical transport modeling, and satellite observations). All the studies evaluated continue
to show evidence of consistent, positive associations between short-term ozone exposure and mortality,
primarily within the first few days after exposure (i.e., lag 0-2 days), as well as evidence of spatial
heterogeneity in risk estimates (Liu et al.. 2016V Liu et al. (2016) as depicted in Figure 6-1. Additional
study details can be found in Table 6-3. Specifically, recent studies focusing on total (nonaccidental)
mortality indicate the following:
• The strongest recent evidence comes from a study by Di et al. (2017a'). who evaluated more
recent air quality data (i.e., 2000-2012) and analyzed the largest study population, with over
22 million case days included in the case-crossover analysis. Using a well-validated hybrid
exposure model, the authors reported a 1.1% increase in all-cause mortality (95% CI: 0.96, 1.24)1
at lag 0-1 for a 20-ppb increase in 8-hour max ozone concentrations in a single-pollutant model.
When limiting ozone data to days where 8-hour max ozone concentrations were less than 60 ppb,
there continued to be evidence of a positive association with all-cause mortality in copollutant
models with PM2 5 (1.16% [95% CI: 0.92, 1.40]; lag 0-1).
• A recent study by Madrigano et al. (2015) provides additional evidence for a positive association
between short-term ozone exposure and total mortality and characterizes the variation in the
association across urban and nonurban areas. The authors examined older air quality data
(i.e., 1988-1999) and used kriging to spatially interpolate ozone concentrations using available
monitoring data in 12 counties to examine associations between short-term ozone exposure and
total mortality across 91 northeastern U.S. counties. The authors examined associations in both
urban (>1,000 persons/mile2) and nonurban (<1,000 persons/mile2) counties. The authors reported
positive associations when using both observed and interpolated ozone concentrations
(Figure 6-1); they reported evidence of associations that are larger in magnitude for nonurban
counties (1.47% [95% CI: 0.38, 2.54], lag 0 for 20-ppb increase in 8-hour max ozone
concentrations) than for urban counties (0.90% [95% CI: 0.16, 1.67], lag 0). Although the
magnitude of the association is larger for nonurban areas, confidence intervals are larger as well
1 Results standardized to a 15-ppb increase in 24-hour avg, 20-ppb increase in 8-hour daily max, or 25-ppb increase
in 1-hour daily max ozone concentrations.
6-5
-------
because of the larger uncertainty in interpolating ozone concentrations from the fewer monitors
and smaller sample sizes in nonurban areas (Appendix 2—Section 2.3.2.1).
• Multiple recent studies that relied on data from NMMAPS spanning the years 1987-2000 also
provide evidence of positive associations between short-term ozone exposure and mortality, but
the studies vary by the number of cities, lags, exposure metrics, and seasons examined (Liu et al..
2016; Jhun et al.. 2014; Moolgavkar et al.. 2013). Additionally, the study by Liu et al. (2016)
provided evidence of spatial heterogeneity in ozone-mortality associations in an analysis focusing
on 10 northern and 10 southern U.S. cities (see Section 6.1.5.4).
• Peng et al. (2013) provides additional evidence of positive associations for short-term ozone
exposure and mortality using 1987-1996 air quality data from NMMAPS (50 U.S. cities all-year
data; 36 U.S. cities summer only) as well as data from 12 Canadian cities as part of the Air
Pollution and Health: A European and North American Approach (APHENA) study. Using a
conservative modeling approach that consisted of penalized splines and 8 degrees of freedom per
year (df/yr) to account for temporal trends, positive associations were observed in both all-year
and summer season analyses, with evidence of associations that are larger in magnitude in the
summer in the U.S. when using the NMMAPS data (Figure 6-1).
6-6
-------
Study
Bell et al (2004)a
Levy et al (2005)a
Bell et al (2005)a
Ito etal. (2005)a
Schwartz (2005)b
BeHetal (2007)b
Bell and Dominici (2008)b
Katsouyanni et al (2009)b
Katsouyanni et al (2009)b
Katsouyanni et al. (2009)b,c
jMoolgavkar etal (2013)
fVanos etal. (2013)
jPeng et al. (2013)
fPeng et al. (2013)
fPeng et al (2013)c
BeH etal (2004)a
Levy et al (2005)a
BeH etal (2005)a
Ito etal. (2005)a
Schwartz (2005)b
Franklin and Schwartz (2008)b
Zanobetti and Schwartz (2008)b
Zanobetti and Schwartz (2008)b
Medina-Ramon and Schwartz (2008)b
Katsouyanni et al (2009)b
Katsouyanni et al (2009)b
Katsouyanni et al. (2009)b,c
|Liu et al (2016)
|Liu et al (2016)
fVanos et al. (2014)
"Peng et al. (2013)
"Peng et al. (2013)
Peng et al (2013)c
Jhun etal. (2014)
"Madrigano et al (2015)d
jDietal (2017)e
Location
95 U. S. communities
U.S. and Non-U. S.
U.S. andNon-U.S.
U.S. and Non-U.S.
14U.S. cities
98 U. S. communities
98 U. S. communities
APHENA-U.S.
APHENA-Canada
APHENA-Canada
NMMAP S (98 U. S. cities)
10 Canadian cities
APHENA-U.S.
APHENA-Canada
APHENA-Canada
95 U. S. communities
U.S. and Non-U.S.
U.S. and Non-U.S.
U.S. and Non-U.S.
14U.S. cities
18 U. S. communities
48 U.S. cities
48 U.S. cities
48 U.S. cities
APHENA-U.S.
APHENA-Canada
APHENA-Canada
NMMAPS(10S. U.S. cities)
NMMAPS (10N. U.S. cities)
10 Canadian cities
APHENA-U.S.
APHENA-Canada
APHENA-Canada
NMMAP S (97 U. S. cities)
91 Northeast U.S. counties
12 Northeast U.S. counties
U.S. - National
Lag
0-6 DL
0
0-1
0-6
0-2 DL
0-2 DL
0-2 DL
1
0
0-2 DL
0-2 DL
0-2 DL
0-6 DL
0
0
0
0-3
0-2
0-2 DL
0-2 DL
0-2 DL
0-2
0-2
0
0-2 DL
0-2 DL
0-2 DL
0
0
0
0-1
Warm/Summe r
% Increase (95% Confidence Interval)
DL = distributed lag.
Note: t and red text = recent muiticity studies, black = U.S. and Canadian multicity studies and meta-analyses evaluated in the 2006
Ozone AQCD and 2013 Ozone ISA. Results standardized to a 15-ppb increase in 24-hour avg, 20-ppb increase in 8-hour max, or
25-ppb increase in 1-hour max ozone concentrations.
aMulticity studies and meta-analyses from the 2006 Ozone AQCD. Bell et al. f20051. Ito et al. f20051. and Levy et al. f20051 used a
range of lag days in the meta-analysis: Lag 0,1,2, or avg 0-1 or 1 -2; single-day lags from 0-3; and lag 0 and 1 -2.
bMulticity studies from the 2013 Ozone ISA.
°Risk estimates from APHENA-Canada standardized to an approximate IQR of 5.1 ppb for a 1-hour max increase in ozone
concentrations as detailed in the 2013 Ozone ISA.
dThe 91 counties analysis used interpolated ozone concentrations, while the 12 counties analysis used observed ozone
concentrations;
eExamined ages 65 and older and all-cause mortality.
Figure 6-1 Summary of associations for short-term ozone exposure and total
(nonaccidental) mortality from recent multicity U.S. and Canadian
studies, and studies in previous ozone assessments.
6.1.4
Cause-Specific Mortality
The majority of evidence examining cause-specific mortality consists of studies evaluated in the
2013 Ozone ISA, which reported primarily consistent, positive associations for both cardiovascular and
respiratory mortality in all-year and summer/warm season analyses. Recent studies have not extensively
examined the relationship between short-term ozone exposure and cause-specific mortality, but both
6-7
-------
multi- and single-city studies continue to support positive associations, particularly with cardiovascular
mortality. Epidemiologic studies evaluated in the 2013 Ozone ISA and recent multicity and single-city
studies that examined cause-specific mortality are characterized in Table 6-4 and Table 6-5. In summary:
• Of the recent multicity studies evaluated, only Vanos et al. (2014) in a study of 10 Canadian cities
examined cause-specific mortality and reported positive associations with both cardiovascular
and respiratory mortality in all-year and summer season analyses. These results are consistent
with the multicity studies and meta-analyses evaluated in the 2013 Ozone ISA (Figure 6-2).
• A few single-city studies also examined short-term ozone exposure and cause-specific mortality,
with Klemm etal. (2011) examining both respiratory and cardiovascular mortality and Sacks et
al. (2012) focusing on cardiovascular mortality in the context of examining the influence of
model specification (i.e., control for seasonal/temporal trends, and weather covariates). Sacks et
al. (2012) reported evidence of positive associations for cardiovascular mortality ranging from
1.30% (95% CI: -2.1, 4.9) to 2.20% (95% CI: -1.8, 6.4) at lag 0-1 days for a 20-ppb increase in
8-hour max ozone concentrations, specifically for those statistical models that more aggressively
controlled for temperature (i.e., using multiple temperature terms or a term for apparent
temperature versus including only one temperature term). In a study conducted in Atlanta, GA,
Klemm et al. (2011) included 7.5 more years of data than Klemm and Mason (2000) and Klemm
et al. (2004). and reported evidence of a positive association with cardiovascular mortality (0.69%
[95% CI: -2.28, 3.75]) at lag 0-1 days for a 20-ppb increase in 8-hour max ozone concentrations
in all-year analyses. However, the authors found no evidence of an association with respiratory
mortality (-0.44% [95% CI: -6.06, 5.51]), which differs from the consistent primarily positive
associations reported in multicity studies (Figure 6-2).
6-8
-------
Study
Location
Ages
Lag
Bell et al. (2005)a
U.S. and non-U.S.
All
...
¦fVanos et al. (2014)
10 Canadian cities
All
0
Katsouyamii et al. (2009)b
APHENA-US
>75
0-2
Katsouyamii et al. (2009)b
APHENA-CAN
>75
0-2
Katsouyamii et al. (2009)b
APHENA-US
<75
0-2
Katsouyamii et al. (2009)b
APHENA-CAN
<75
0-2
Zanobetti and Schwartz (2008)b
48 U.S. cities
All
0-3
¦fVanos et al. (2014)
10 Canadian cities
All
0
Katsouyamii et al. (2009)b
APHENA-US
>75
0-2
Katsouyamii et al. (2009)b
APHENA-CAN
>75
0-2
Katsouyamii et al. (2009)b
APHENA-US
<75
0-2
Katsouyamii et al. (2009)b
APHENA-CAN
<75
0-2
Bell et al. (2005)a
U.S. and non-U.S.
All
...
Katsouyamii et al. (2009)b
APHENA-US
All
0-2
Katsouyamii et al. (2009)b
APHENA-CAN
All
0-2
fVanos et al. (2014)
10 Canadian cities
All
0
Katsouyamii et al. (2009)b
APHENA-US
>75
0-2
Katsouyamii et al. (2009)b
APHENA-CAN
>75
0-2
Zanobetti and Schwartz (2008)b
48 U.S. cities
All
0-3
Katsouyamii et al. (2009)b
APHENA-US
All
0-2
Katsouyamii et al. (2009)b
APHENA-CAN
All
0-2
fVanos et al. (2014)
10 Canadian cities
All
0
Katsouyamii et al. (2009)b
APHENA-US
>75
0-2
Katsouyamii et al. (2009)b
APHENA-CAN
>75
0-2
Cardiovascular
Respiratory
<
All-ye ar
Summer
All-ye ar
Summer
~—
-+-
-5.0
0.0 5.0 10.0 15.0
% Increase (95% Confidence Interval)
20.0
Note: t and red text = recent multicity studies, black = U.S. and Canadian muiticity studies and meta-analyses evaluated in the 2006 Ozone AQCD and 2013 Ozone ISA. Results
standardized to a 15-ppb increase in 24-hour age, 20-ppb increase in 8-hour max, or 25-ppb increase in 1-hour max ozone concentrations.
aMulticity studies and meta-analyses from the 2006 Ozone AQCD used a range of lag days in the meta-analysis: Lag 0, 1, 2, or avg 0-1 or 1 -2.
bMulticity studies from the 2013 Ozone ISA.
Figure 6-2 Summary of associations for short-term ozone exposure and cause-specific mortality from recent
multicity U.S. and Canadian studies, and studies evaluated in previous ozone assessments.
6-9
-------
6.1.5 Effect Modification of the Ozone-Mortality Relationship
Multicity studies evaluated in the 2013 Ozone ISA reported evidence of spatial heterogeneity,
both regional as well as city-to-city, in the magnitude of ozone-mortality risk estimates. To assess what
may account for this heterogeneity, studies often examined factors that may modify the ozone-mortality
relationship. Studies that conducted such analyses in the 2013 Ozone ISA provided initial evidence that a
number of individual- and population-level factors (e.g., race, unemployment rate) may explain this
heterogeneity. In addition to examining individual- and population-level factors recent studies also
explored specific weather conditions (i.e., season, temperature, and weather patterns) that may modify the
ozone-mortality relationship and potentially contribute to these observed differences in risk estimates.
6.1.5.1 Lifestage
Few recent studies have conducted extensive analyses to examine whether specific
individual- and population-level factors modify the ozone-mortality relationship. Across studies, there is
some evidence of increased risk of ozone-related mortality in older adults (i.e., >65 years of age),
particularly as age increases, with more limited evidence for other factors, which is reflected in the
following studies:
• In analyses of potential modifiers of the ozone-mortality relationship, Di et al. (2017a') reported
slightly elevated risks in females compared to males, as well as for Medicaid eligible versus
noneligible participants. However, the most pronounced difference across the factors examined
was for increasing age, with the risk of mortality attributed to short-term ozone exposure almost
double for people 75-84 and >85 years of age compared to people <69 years of age.
• Vanos et al. (2013) also reported some evidence of increased risk in older individuals, with the
risk being greater in individuals 75-84 years of age compared to the other age ranges examined
(i.e., <64, 65-74, and >85; quantitative results not presented).
• Madrigano et al. (2015) examined modification of the ozone-mortality association by county
characteristics. The authors reported evidence of increased risk in counties with a large
percentage of the population over the age of 65 years, which provides some support for the results
of Di et al. (2017a) and Vanos et al. (2013). Additionally, Madrigano et al. (2015) reported no
evidence of increased risk as the percent of families in poverty or population density increased.
6.1.5.2 Pre-existing Diseases
A limited number of studies evaluated in the 2013 Ozone ISA provided some evidence that
pre-existing cardiovascular diseases, such as atrial fibrillation and atherosclerosis, may increase the risk
of ozone-related mortality. Recent single-city studies conducted in Montreal, Canada by Goldberg et al.
(2013) and Buteau et al. (2018) further examined the role of pre-existing cardiovascular diseases in the
relationship between short-term ozone exposure and mortality in individuals >65 years of age. Consistent
6-10
-------
with the few studies evaluated in the 2013 Ozone ISA, recent studies indicate that some pre-existing
cardiovascular disease may increase the risk of ozone-related mortality:
• When examining a distributed lag nonlinear model (DLNM) for 0-2 days, Goldberg et al. (2013)
reported little evidence of an association (i.e., positive, but with wide confidence intervals) when
focusing on individuals with a diagnosis of any cardiovascular disease 1 year prior to death.
Specifically, positive associations were reported in all-year analyses for pre-existing congestive
heart failure (2.99% [95% CI: -1.95, 8.17]) and any type of cancer (3.57% [95% CI: 0.16, 7.10]),
with associations that are larger in magnitude and more precise (i.e., smaller confidence intervals)
in warm seasons analyses for acute coronary artery disease (7.78% [95% CI: 2.43, 13.41]),
hypertension (3.70% [95% CI: -0.08, 7.63]), and cerebrovascular disease (4.93% [95% CI:
-0.04, 10.16]) for a 15-ppb increase in 24-hour avg ozone concentrations. The authors found no
evidence of an association for a number of other pre-existing cardiovascular diseases including
diabetes and atrial fibrillation, which was previously found to be associated with an increased risk
of ozone-related mortality (Medina-Ramon and Schwartz. 2008).
• While Goldberg et al. (2013) examined a number of pre-existing cardiovascular diseases, Buteau
et al. (2018) only focused on individuals with pre-existing congestive heart failure with an
emphasis on examining associations across five different exposure assignment approaches
(i.e., inverse-distance weighting [IDW], back extrapolation method based on land use regression
[LUR], Bayesian maximum entropy [BME] model, nearest monitor, and average across all
monitors) using two distinct study designs (i.e., case-crossover and nested case-control). In the
case-crossover analysis, which focuses on examining why a person died on a particular day rather
than on other days within the same month, the authors reported no evidence of an ozone-mortality
association across each of the exposure assignment approaches used. However, in the nested
case-control analysis, where the authors examined why a person died on this day while others did
not, there was evidence of positive associations at lag 0-3 DLNM for 20-ppb in 8-hour avg ozone
concentrations for the nearest station (6.84% [95% CI: 0.31, 13.79]), IDW (22.69% [95% CI:
-3.12, 55.30]), and back extrapolated LURmethods (8.97% [95% CI: 3.67, 14.70]).
6.1.5.3 Season
As detailed in Appendix 1 Section 1.5. ozone concentrations are generally higher in the summer
or warm months due to the atmospheric conditions that lead to ozone formation. Therefore, because of the
seasonal patterns in ozone concentrations, as well as many locations, particularly within the U.S., only
monitoring ozone during the summer or warm months, many of the epidemiologic studies tend to focus
on summer or warm season analyses. However, some studies conduct all-year analyses based on areas
that monitor ozone all year, with a subset of these studies then examining whether the magnitude of the
ozone-mortality association varies either across seasons or in the summer/warm season compared to the
entire year. Studies evaluated in the 2013 Ozone ISA, reported evidence of positive ozone-mortality
associations in all-year analyses that tended to be larger in magnitude during the warm or summer
months. Recent studies that conducted all-year as well as seasonal analyses reported associations in the
warm or summer months that were similar or larger in magnitude to those in all-year analyses.
Specifically, recent studies indicate:
6-11
-------
• In analyses examining ozone-mortality associations across the four seasons, associations were
largest in magnitude during the spring and/or summer depending on the mortality outcome
examined [i.e., total or cause-specific; Liu et al. (2016); Vanos et al. (2014)1. However, in Liu et
al. (2016). this pattern of associations differed between northern and southern U.S. communities,
with only northern U.S. communities having larger associations during the spring and summer.
This pattern of associations could be due to the differences in long-term mean temperatures
between locations or air conditioning (AC) prevalence (see Section 6.1.5.4).
• Peng et al. (2013) and Goldberg et al. (2013) compared ozone-mortality associations in all-year
and broad seasonal analyses (i.e., warm/summer and cold season). In the U.S., Peng et al. (2013)
reported a 3.23% (95% CI: 1.63, 4.85) increase in mortality in warm season analyses for a
distributed lag (DL) of 0-2 days for a 25-ppb increase in 1-hour max ozone concentrations
compared to a 2.13% (95% CI: 0.54, 3.73) increase in an all-year analysis. The pattern of
associations observed in Peng et al. (2013) was consistent with Goldberg et al. (2013). although
confidence intervals were wide, in Montreal, Canada across some of the pre-existing
cardiovascular disease outcomes examined (i.e., all cardiovascular diseases, acute coronary artery
disease, atrial fibrillation, and hypertension). However, when examining the Canadian cohort,
Peng et al. (2013) did not report associations that are larger in magnitude in the summer season
(2.08%) compared with the all-year (3.73%) analysis.
6.1.5.4 Temperature
Ozone formation is tangentially linked to temperature because the periods of greatest solar
radiation, a prerequisite for producing ozone (see Appendix 1). occur during months of the year when
temperatures are highest. However, in unique circumstances when VOC concentrations are extremely
high, high ozone concentrations can occur with less sunlight (see Appendix 1). Given this tangential
linkage, it has been hypothesized that temperature may modify the relationship between short-term ozone
exposure and mortality. Recent studies conducted analyses either focusing on long-term average
temperature, daily temperature, or the joint effect of ozone and temperature with the aim of elucidating
the role of temperature on the ozone-mortality relationship. These studies indicate that ozone-mortality
associations are larger in magnitude in locations with lower long-term average temperature, there is
evidence of increased risk of ozone-related mortality at higher daily temperatures, and there is evidence
the risk of mortality increases at higher ozone concentrations and higher temperatures (i.e., a synergistic
effect), which is more extensively detailed below:
• Both Peng et al. (2013) and Liu et al. (2016) examined whether ozone-mortality associations are
modified by long-term average temperature. When examining the distribution of mean
temperature across cities, Peng et al. (2013) reported no evidence of ozone-mortality risk
estimates increasing as temperatures increased from the 25th to the 75th percentile in the U.S.
data set. The results of Peng et al. (2013) in the U.S. cities analysis are supported by Liu et al.
(2016). As depicted in Figure 6-3. when examining average temperature, positive associations
were only observed in those cities with lower average temperatures.
• When examining the Canadian data set Peng et al. (2013) found that ozone-mortality risk
estimates were slightly elevated when moving from the 25th percentile (1.75% increase) to the
75th percentile (2.20% increase) of the mean temperature distribution for a 25-ppb increase in
6-12
-------
1-hour max ozone concentrations at lag 0-1. The results of the U.S. analyses by Peng et al.
(2013) and Liu et al. (2016). which showed no evidence that mortality risk increased across the
temperature distribution, could help explain the positive associations across the temperature
distribution in the Canadian analysis because mean temperatures are lower across Canadian cities.
Additionally, the pattern of associations observed in the Canadian analysis could be a reflection
of long-term average temperature being a surrogate for AC prevalence as detailed in the 2013
Ozone ISA (U.S. EPA. 2013a).
• Instead of examining long-term average temperatures, Jhun et al. (2014) examined whether mean
daily (i.e., 24-hour avg) temperature modified the ozone-mortality relationship depending on
where the mean temperature for an individual day fell along the distribution of mean daily
temperatures across the study duration. In three separate analyses, where low and high
temperatures were defined using cutoffs of the 25th and 75th percentile of mean temperatures,
10th and 90th percentile, and 5th and 95th percentile, the authors reported evidence of a U-shaped
curve with ozone-mortality risks being larger in magnitude at the lowest and highest temperature
categories, compared to the moderate category, for each temperature cutoff examined. Upon
closer examination, the ozone-mortality risk was largest at low mean temperatures when using the
25th percentile cutoff, and for high mean temperatures largest above the 95th percentile.
However, in a sensitivity analysis focusing on air conditioning prevalence, ozone-mortality risk
estimates at mean temperatures above the 95th percentile were attenuated when examining risks
across the distribution of air conditioning prevalence, indicating greater air conditioning use at
mean temperatures at the high end of the temperature distribution.
• While previous studies focused on whether ozone-mortality risk estimates were modified by
long-term average temperature or mean daily temperature, Wilson et al. (2014) conducted an
analysis examining the joint effects of ozone and temperature on mortality. The authors used a
spatial monotone surface model to examine the ozone-temperature interaction for the same
95 U.S. cities from NMMAPS detailed in Bell et al. (2004). This approach allows for the
examination of the interaction between ozone and temperature by evaluating mortality risk at
different temperature ranges for the same ozone concentration. In analyses focusing on
April-October using 1-hour max ozone concentrations at lag 0 and mean daily temperature, the
authors reported evidence of a larger ozone-mortality risk (i.e., a synergistic effect) at higher
mean daily temperatures and higher ozone concentrations (Figure 6-4). When examining the
national estimate results across the three different models (i.e., additive linear, additive nonlinear,
and the monotone spatial risk surface model) the percent increase in mortality, which represented
an increase from the median of ozone concentrations and temperature to the 95th percentile, was
found to vary with the models reporting a 3.06, 3.54, and 3.98% increase, respectively. These
results provide evidence of nonlinearity in the relationship between ozone concentrations and
temperature on mortality risk.
• In regional analyses, Wilson et al. (2014) reported results similar to Liu et al. (2016) by observing
a larger degree of modification of ozone-mortality risk estimates in northern U.S. cities where
there is a larger difference between temperatures on high temperature and moderate temperature
days.
6-13
-------
Spring
1 o:
51.01 H
j 1 00 -
C3
O
K 0 99 -
0 98
I
10
15
I
20
Average temperature (*C)
1.02
1.01 -
-tL
k.
-------
• Similar to Wilson et al. (2014). Chen et al. (2018) examined a bivariate response surface of ozone
(i.e., 24-hour avg) and temperature (i.e., mean daily or 24-hour avg) to capture the joint effect on
daily mortality in 86 U.S. cities from NMMAPS. The authors observed that mean daily
temperature positively modified the ozone-mortality relationship. In addition to examining a
bivariate response surface, Chen et al. (2018) also examined temperature-stratified
ozone-mortality associations across the distribution of mean daily temperatures within each city.
Overall, the authors reported evidence of modification of the ozone-mortality association at high
daily mean temperatures (i.e., >75th percentile; Figure 6-5). However, as noted in Section 6.1.6.2.
the authors also reported evidence of potential residual confounding when examining
temperature-stratified ozone-mortality associations.
^ CD
~ o
-s 8
O JZ
_>-csl
<5 .£
.5
& 75th percentile. Gray and circle = categorical term without
adjustment for smooth terms of temperature; purple and triangle = categorical term, plus distributed lag nonlinear model (DLNM) for
two different B-splines; red and square = categorical term, plus separate natural splines for low and high temperatures; yellow and
dash = categorical term, plus natural spline for high temperatures; blue and square with an x = categorical term, plus natural spline
for low temperatures.
Source: Reprinted with permission from the publisher; Chen et al. (20181.
Figure 6-5 Temperature-stratified ozone-mortality associations from 86 U.S.
cities within the National Morbidity, Mortality, and Air Pollution
Study (NMMAPS) using different approaches to control for
nonlinearity in temperature effects.
6-15
-------
6.1.5.5 Weather Patterns
While the majority of studies to date focus on whether season or temperature modify the
ozone-mortality relationship, a series of recent studies (Vanos et al.. 2015; Vanos et al.. 2014; Vanos et
al.. 2013) conducted in multiple cities in Canada examined the role of specific weather patterns
(i.e., synoptic weather categories). The weather categories examined included dry moderate (DM), dry
polar (DP), dry tropical (DT), moist moderate (MM), moist polar (MP), moist tropical (MT), and a
transitional category (TR), representing the shift from one weather category to another. Although each of
the aforementioned studies conducted analyses that differed by lag days, mortality outcome, and years
examined, they all showed some evidence of positive associations for short-term ozone exposure and
mortality, as well as cause-specific mortality, across each of the synoptic weather categories examined.
When examining individual weather categories, the highest risk was reported with the DT and MT
weather categories, which were the weather categories found to encompass the most extreme pollution
episodes, as detailed in Vanos et al. (2015).
6.1.6 Potential Confounding of the Ozone-Mortality Relationship
The assessment of potential copollutant confounding in the 2013 Ozone ISA revolved around
evaluating studies that focused primarily on PM2 5 and PM10. These studies reported that ozone-mortality
risk estimates were relatively unchanged in copollutant models, but they had difficulty assessing this
evidence due to the regional variability in ozone-PM correlations and the every-3rd or 6th-day PM
sampling schedule. Studies evaluated in the 2013 Ozone ISA also examined the impact of controlling for
seasonality, with the most extensive analysis conducted within the APHENA study (Katsouvanni et al..
2009). which demonstrated that model misspecification can occur when not enough degrees of freedom
(df) are applied to control for the opposing seasonal trends between ozone and mortality. Recent studies
that examined whether copollutant exposures and temporal/seasonal trends confound the ozone-mortality
relationship provide evidence that continues to support that ozone-mortality associations are relatively
unchanged in copollutant models and relatively consistent across a range of df that properly account for
temporal/seasonal trends. Additionally, a few recent studies conducted analyses aimed at informing
whether the potential confounding effects of temperature have been adequately controlled for when
examining ozone-mortality associations.
6.1.6.1 Potential Copollutant Confounding
Recent studies that examined potential copollutant confounding focused primarily on particulate
matter, either PM2 5 or PM10, with an additional study examining NO2. The results of these studies are
consistent with studies evaluated in the 2013 Ozone ISA that demonstrated that ozone-mortality
associations are relatively unchanged in copollutant models with PM as detailed below:
6-16
-------
• Using more recent air quality data, Di et al. (2017a) reported that ozone-mortality risk estimates
were attenuated, but remained positive in copollutant models with PM2 5 (ozone =1.10 [95% CI:
0.96, 1.24], ozone + PM2 5 = 1.02% [95% CI: 0.82, 1.22]; lag 0-1 for a 20-ppb increase in 8-hour
max ozone concentrations).
• The remaining multicity U.S. studies that examined potential copollutant confounding by
particulate matter focused on PMi0. Both Moolgavkar et al. (2013) and Peng et al. (2013)
examined potential copollutant confounding using NMMAPS data and provided evidence that
associations were slightly attenuated, but remained positive with wider confidence intervals in
copollutant models with PM10 rMoolgavkar et al. (2013). lag 0-1: ozone = 0.60% (95% CI: 0.44,
0.80), ozone + PM10 = 0.33% (95% CI: -0.7, 0.72) for a 15-ppb increase in 24-hour avg ozone
concentrations; Peng et al. (2013). lag 1: ozone = 0.89% (95% CI: 0.00, 1.73),
ozone + PM10 = 0.64% (95% CI: -0.88, 2.18) for a 25-ppb increase in 1-hour max ozone
concentrations]. The increase in the widths of the confidence intervals observed in these studies is
consistent with a decrease in precision due to the limited data available to conduct copollutant
analyses due to the PM sampling schedule.
• Chen et al. (2018) also used NMMAPS data for 86 U.S. cities to examine the relationship
between short-term ozone exposure and mortality and evaluated copollutant models in analyses
that were stratified by temperature within each city (i.e., <25th, 25th-75th, >75th percentiles). In
copollutant models with PM10 and NO2, the authors reported evidence of associations being
attenuated, but remaining positive in the high-temperature category (quantitative results not
presented). There was limited to no evidence of positive associations in single or copollutant
analyses in the low and medium temperature ranges.
• The results of copollutant models from the aforementioned U.S.-based multicity studies are
consistent with the single-city analysis conducted in Madrigano et al. (2015) based on the one city
(New Haven, CT) that had both PM10 and ozone data during the study duration (lag 0:
ozone = 5.14% [95% CI: 1.57, 8.85], ozone + PM10 = 5.04% [95% CI: 1.38, 8.81] for a 20-ppb
increase in 8-hour max ozone concentrations).
• Peng et al. (2013) also examined potential copollutant confounding using data from Canada and
reported that results were relatively unchanged when adjusting for PM10 (ozone = 2.78% [95%
CI: 1.38, 4.14], ozone + PM10 = 2.38% [95% CI: -0.88, 6.02]), which is consistent with the U.S.
analysis.
6.1.6.2 Potential Confounding by Temporal/Seasonal Trends and Weather
Recent studies examined the influence of alternative approaches to control for the potential
confounding effects of temporal/seasonal trends and weather on the association between short-term ozone
exposure and mortality through their systematic evaluations of various statistical models or by varying the
parameters of specific covariates included in statistical models (e.g., weather covariates examined,
degrees of freedom per year to control for temporal/seasonal trends). Analyses of temporal/seasonal
trends support the conclusions of the 2013 Ozone ISA, which demonstrated that not properly accounting
for temporal/seasonal trends can result in model misspecification. Additionally, recent studies provide
new information indicating that not properly accounting for the potential confounding effects of
temperature may result in residual confounding of the ozone-mortality relationship. Specifically, recent
studies found the following:
6-17
-------
• In all-year analyses, Peng et al. (2013) reported relatively consistent positive associations when
examining 3, 8, and 12 df/year using both natural and penalized splines in Canada; however, in
the U.S. there was evidence that less than 3 df/year does not properly account for
temporal/seasonal trends. While Peng et al. (2013) focused on all-year analyses, Liu et al. (2016)
examined the use of 5-9 df/year to account for temporal trends in seasonal analyses for both
southern and northern U.S. communities. Results were relatively unchanged for all seasons,
except the winter where there was some evidence that the magnitude of the association increased
at df/year greater than 7. This observation was more prominent in the southern communities, but
a similar pattern was also observed in the northern communities. This indicates a potential
subseasonal trend not present for other seasons that requires additional control when focusing on
season-specific analyses.
• The summer season results of Liu et al. (2016) are consistent with the warm season analysis
conducted by Madrigano et al. (2015) in the 12-county analysis using observed ozone
concentrations. Associations of similar magnitude and precision were observed when using either
4 or 7 df/year.
• While the aforementioned studies focused on assessing the control for temporal/seasonal trends,
Di et al. (2017a) examined whether the appropriate df were instituted to control for
meteorological factors in the statistical model. The authors did not report any evidence that the
magnitude of ozone-mortality risk estimates changed when increasing the df for meteorological
variables (i.e., temperature and dew point temperature) from 6 to 9.
While the studies detailed above focused primarily on examining how changing the df for
temporal/seasonal trends or temperature influenced ozone-mortality associations, additional studies
conducted systematic evaluations of the relationship between alternative model specifications and
ozone-mortality risk estimates and provided new information on the potential for residual confounding:
• Sacks et al. (2012) examined whether similar results were observed across the different statistical
models used in multicity studies using a common data set. The authors observed variability in the
ozone-cardiovascular mortality relationship corresponding to differing levels of adjustment for
temperature. Specifically, those statistical models that more thoroughly controlled for
temperature, such as by including multiple temperature terms or a term for apparent temperature,
were found to have larger risk estimates (1.3 to 2.2% for a 20-ppb increase in 8-hour max ozone
concentrations) compared with those models that included only one temperature term for a single
lag day (-1.6 to 0.5%).
• In examining the ozone-mortality relationship at different temperatures, Chen et al. (2018)
employed multiple methods to explore whether hot or cold temperatures confounded
temperature-stratified ozone-mortality associations. There was evidence of residual confounding
and the overestimation of ozone-mortality risk estimates, specifically at high temperatures, which
was attributed to not adequately controlling for heat effects.
6.1.7 Shape of the Concentration-Response (C-R) Relationship
Studies included in the 2013 Ozone ISA conducted a variety of statistical analyses to characterize
the shape of the concentration-response (C-R) relationship between short-term ozone exposure and
mortality and did not observe any evidence of a threshold or deviations from linearity within the range of
ozone concentrations observed within the U.S. However, it is important to note that the examination of
6-18
-------
the ozone-mortality C-R relationship was complicated by previously identified city-to-city and regional
heterogeneity in ozone-mortality risk estimates (U.S. EPA. 2013a). Recent studies continue to provide
evidence of a linear C-R relationship with no evidence of a threshold below which mortality effects do
not occur along the distribution of ozone concentrations observed within the U.S. Additionally, consistent
with studies evaluated in the 2013 Ozone ISA, recent studies do not conduct city-specific analyses of the
C-R relationship. These results are described below:
• Moolgavkar et al. (2013) reported evidence of a linear relationship down to concentrations of
60 ppb, with less certainty in the shape of the curve below 60 ppb when examining lag 1, 24-hour
avg ozone concentrations in a flexible model using 6 df (Figure 6-6).
106
a
O.Oi
0.02
"7
-0.02
-r~
20
—r~
10
Lag-1 0,
-r~
60
Source: Reprinted with permission from the publisher; Moolgavkar et al. (20131.
Figure 6-6 Flexible concentration-response relationship for short-term ozone
exposure and mortality at lag 1 for 24-hour avg ozone
concentrations adjusted by size of the bootstrap sample [size of
the bootstrap (d) = 4].
• While Moolgavkar et al. (2013) focused on 24-hour avg ozone concentrations, Di et al. (2017a)
examined the ozone-mortality C-R relationship in a copollutant model with PM2 5 using 8-hour
max ozone concentrations. In models using penalized splines for both ozone and PM2 5, the
authors reported evidence of a linear, no-threshold relationship with less certainty at
concentrations below approximately 30 ppb (Figure 6-7).
6-19
-------
0
20
40
60
80
100
Ozone, ppb
Source: Reprinted with permission from the publisher; Pi et al. (2017a1.
Figure 6-7 Percentage increase in mortality for ozone in a two-pollutant
model with PM2.5 using penalized splines for both pollutants at lag
0-1 days in the warm season (April-September).
• While Moolgavkar et al. (2013) and Di et al. (2017a) focused specifically on the shape of the C-R
relationship, Peng et al. (2013) examined whether there was evidence of a threshold below which
mortality effects are not observed. In a threshold analysis using 1-hour max ozone concentrations
where threshold values were set at 5 ppb increments from 0-75 ppb, there was no evidence of a
threshold in any of the data sets examined in the APHENA study, including data from the U.S.
and Canada.
6.1.8 Summary and Causality Determination
This section describes the evaluation of evidence for total (nonaccidental) mortality based on the
scientific considerations detailed in the Annex for Appendix 6. with respect to the causality determination
for short-term ozone exposures using the framework described in Table II of the Preamble to the ISAs
(U.S. EPA. 2015). The key evidence, as it relates to the causal framework, is summarized in Table 6-1.
Recent multicity studies conducted in the U.S. and Canada continue to provide evidence of consistent,
positive associations between short-term ozone exposure and total mortality in both all-year and
summer/warm season analyses across different averaging times (i.e., 1-hour max, 8-hour max, 8-hour
avg, and 24-hour avg; Figure 6-1).1 However, most of the recent studies examined associations between
1 A list of considered studies conducted in other geographic locations is available via the HERO database.
6-20
-------
short-term ozone exposure and mortality using ozone data from before the year 2000, with only Di et al.
(2017a) focusing on more recent (i.e., since 2000) ozone concentrations.
The assessment of cause-specific mortality is limited to a study conducted by Vanos et al. (2014).
which presented results that are consistent with the pattern of positive associations reported for studies
evaluated in the 2013 Ozone ISA (Figure 6-2). However, the experimental evidence for cardiovascular
morbidity, specifically from controlled human exposure studies, is not consistent with the studies
evaluated in the 2013 Ozone ISA. The combination of limited evidence of cardiovascular effects in
controlled human exposure studies, along with epidemiologic studies of cardiovascular-related ED visits
and hospital admissions that report limited evidence of a relationship with short-term ozone exposure,
lead to substantial gaps in the biologically plausible pathways by which short-term ozone exposure could
lead to cardiovascular mortality.
Recent studies continue to assess the influence of potential confounders on the ozone-mortality
relationship, including copollutants, temporal/seasonal trends, and weather covariates. The assessment of
potential copollutant confounding was examined within the NMMAPS data set by multiple studies (Chen
et al.. 2018; Moolgavkar et al.. 2013; Peng et al.. 2013). all of which reported that ozone-mortality
associations remained positive, although attenuated in some instances, in copollutant models with PMio or
NO2. These results were further supported in a large national analysis of Medicare participants
(i.e., >65 years of age) in which ozone-mortality associations were similar in magnitude in single
pollutant models and copollutant models with PM10 (Di et al.. 2017a). as well as for an individual city
within the study conducted by Madrigano et al. (2015). Importantly, the issues surrounding the
assessment of potential copollutant confounding detailed in the 2013 Ozone ISA persists, specifically
within studies that relied on NMMAPS data, due to the every-3rd and 6th-day PM sampling.
Additional analyses building off the extensive examination of model specification by
Katsouvanni et al. (2009) within APHENA demonstrate that not instituting enough df per year to account
for temporal/seasonal trends may underestimate ozone-mortality risk estimates (Peng et al.. 2013).
However, there is also preliminary indication that some seasons may require additional df when
examining seasonal associations [i.e., winter; Liu et al. (2016)1. but additional exploration is warranted
because many locations do not monitor ozone outside of the warm/summer season. A limited assessment
of model specification with respect to individual weather covariates indicates that increasing the df does
not affect ozone-mortality associations (Peng et al.. 2013). but not properly accounting for the effect of
temperature on mortality may overestimate (Chen et al.. 2018) or underestimate (Sacks et al.. 2012)
ozone-mortality associations.
Effect modification of the ozone-mortality relationship was assessed through studies focusing on
both individual and population-level factors, as well as weather-related conditions. There is some
evidence of increased risk of ozone-related mortality in older individuals (i.e., >65 years of age),
particularly in individuals 75-84 years of age (Di et al.. 2017a; Vanos et al.. 2013). An assessment of
pre-existing disease, which was limited to studies conducted in Montreal, Canada, reported evidence of
6-21
-------
increased risk in individuals with CHF, and more limited evidence for other cardiovascular-related
diseases, including acute coronary artery disease, hypertension, and cerebrovascular disease (Butcau et
al.. 2018; Goldberg et al.. 2013).
While there continues to be some evidence of differential ozone-mortality associations by season
(Section 6.1.5.3). the most extensive analyses conducted by recent studies examined whether temperature
(i.e., long-term average temperatures or the distribution of mean daily temperatures) modifies the
ozone-mortality association. Analyses focusing on temperature, indicate that locations with lower
long-term average temperature have higher ozone-mortality risk estimates (Liu et al.. 2016; Peng et al..
2013). which is also reflected by the observed difference in risk estimates between northern and southern
U.S. cities in Liu et al. (2016). However, long-term average temperature may be a surrogate for air
conditioning prevalence. Additionally, studies that examined either the joint effects of ozone and mean
daily temperature on mortality (Chen et al.. 2018; Wilson et al.. 2014) or temperature-stratified
ozone-mortality associations (Chen et al.. 2018; Jhun et al.. 2014) provided evidence of ozone-mortality
associations that are larger in magnitude at temperature extremes (i.e., low and/or high mean daily
temperatures).
Recent multicity studies continue to support a linear a C-R relationship with no evidence of a
threshold between short-term ozone exposure and mortality over the range of ozone concentrations
typically observed in the U.S. Studies that used different statistical approaches and ozone averaging times
(i.e., 24-hour avg and 8-hour max) provide evidence of a linear C-R relationship, with less certainty in the
shape of the curve at lower concentrations [i.e., 40 ppb for 24-hour avg Moolgavkar et al. (2013) and
30 ppb for 8-hour max Di et al. (2017a)l. An examination of whether a threshold exists in the
ozone-mortality C-R relationship provided no evidence of a concentration below which mortality effects
do not occur when examining 5 |ig/m3 (-2.55 ppb) increments across the range of 1-hour max
concentrations reported in the U.S. and Canadian cities included in APHENA (Peng et al.. 2013).
Collectively, these results continue to support the conclusion of the 2006 Ozone AQCD that "if a
population threshold level exists in ozone health effects, it is likely near the lower limit of ambient ozone
concentrations in the U.S."
Building on the 2013 Ozone ISA, there remains strong evidence for respiratory effects due to
short-term ozone exposure (Appendix 3) that is consistent within and across disciplines and provides
coherence and biological plausibility for the positive respiratory mortality associations reported across
epidemiologic studies. Although there remains evidence for ozone-induced cardiovascular mortality the
preliminary evidence presented in the 2013 Ozone ISA from controlled human exposure and animal
toxicological studies that provided a biologically plausible mechanism for ozone-induced cardiovascular
mortality is inconsistent with a larger number of recent controlled human exposure studies that do not
provide evidence of cardiovascular effects in response to short-term ozone exposure. The recent evidence
from controlled human exposure studies, in combination with the lack of coherence between animal
toxicological and epidemiologic studies of cardiovascular morbidity, specifically the lack of
6-22
-------
epidemiologic evidence for cardiovascular-related ED visits and hospital admissions, leads to substantial
gaps in the biologically plausible pathways by which short-term ozone exposure could lead to
cardiovascular mortality (Appendix 4).
Overall, the recent multicity studies conducted in the U.S. and Canada provide additional support
for the consistent, positive associations reported across multicity studies evaluated in the 2006 Ozone
AQCD and 2013 Ozone ISA. These results are supported by studies that further examined uncertainties in
the ozone-mortality relationship, such as potential confounding by copollutants and other variables,
modification by temperature, and the C-R relationship and whether a threshold exists. Although there
continues to be strong evidence from studies of respiratory morbidity to support respiratory mortality,
there remains relatively limited biological plausibility and coherence within and across disciplines to
support the relatively strong evidence for cardiovascular mortality, which comprises a large percentage of
total (nonaccidental) mortality. Collectively, evidence is suggestive of, but not sufficient to infer, a
causal relationship exists between short-term ozone exposure and total mortality.
Table 6-1 Summary of evidence that is suggestive of, but not sufficient to infer,
a causal relationship between short-term ozone exposure and total
mortality.
Rationale for
Causality
Determination3
Key Evidence13
Key References'3
Ozone
Concentrations
Associated
with Effects
(ppb)c
Consistent
epidemiologic evidence
from multiple,
high-quality studies at
relevant ozone
concentrations
Recent multicity studies
conducted in the U.S. and
Canada continue to support
the consistent positive
associations between
short-term ozone exposure
and total mortality in both
all-year and warm/summer
season analyses.
Pi et al. (2017a)
Figure 6-1
Section 6.1.3
Mean
concentrations
across studies:
24-h avg:
14.5-48.7
8-h max/avg:
15.1-62.8
1-h max:
6.7-60.0
6-23
-------
Table 6-1 (Continued): Summary of evidence that is suggestive of, but not
sufficient to infer, a causal relationship between short-term
ozone exposure and total mortality.
Ozone
Concentrations
Rationale for
Associated
Causality
with Effects
Determination3
Key Evidence13
Key References'3
(ppb)c
Epidemiologic
The few recent multicity
Moolaavkar et al. (2013)
evidence from
studies that examined
Penq et al. (2013)
copollutant models
potential copollutant
Chen et al. (2018)
provides some support
confounding provide
for an independent
evidence supporting that
Di et al. (2017a)
ozone association
ozone-mortality risk
estimates are relatively
unchanged or slightly
attenuated, but remain
positive, in copollutant
models with PM2.5, PM10,
and NO2.
Studies that reported
correlations between ozone
and PM2.5 or PM10 were
generally low (<0.40).
Section 6.1.6.1
Epidemiologic
Studies continue to provide
Moolaavkar et al. (2013)
24-h avg
evidence continues to
evidence of a linear C-R
Di et al. (2017a)
>40
support a linear C-R
relationship with no
relationship with no evidence
of a threshold. There is less
Pena et al. (2013)
8-h max
>30
evidence of a threshold
certainty in the shape of the
C-R relationship at the lower
end of concentrations
observed in the U.S.
Section 6.1.7
Limited biological
While animal toxicological
Appendix 4
plausibility from studies
studies provide evidence of
of cardiovascular
cardiovascular effects,
morbidity
recent controlled human
exposure studies do not
provide evidence to support
potential biological
pathways. Additionally, there
is a lack of coherence with
epidemiologic studies of
cardiovascular morbidity,
specifically, cardiovascular-
related ED visits and hospital
admissions, to support
cardiovascular mortality.
Biological plausibility
Strong evidence for
Appendix 3
from studies of
respiratory effects due to
respiratory morbidity
short-term ozone exposure,
such as asthma
exacerbation, are consistent
across disciplines and
support potential biological
pathways for respiratory
mortality.
6-24
-------
Table 6-1 (Continued): Summary of evidence that is suggestive of, but not
sufficient to infer, a causal relationship between short-term
ozone exposure and total mortality.
Rationale for
Causality
Determination3
Key Evidence13
Key References'3
Ozone
Concentrations
Associated
with Effects
(ppb)c
Uncertainty regarding
geographic
heterogeneity in
ozone-mortality
associations
Recent studies indicate
latitude and temperature
may account for some of the
observed heterogeneity, but
more extensive evaluations
have not been conducted.
Section 6.1.3
Section 6.1.5.4
C-R = concentration-response; N02 = nitrogen dioxide; PM2 5 = particulate matter with a nominal aerodynamic diameter less than
or equal to 2.5 |jm; PM10 = particulate matter with a nominal aerodynamic diameter less than or equal to 10 |jm; ppb = parts per
billion.
aBased on aspects considered in judgments of causality and weight of evidence in causal framework in Table I and Table II of the
Preamble.
bDescribes the key evidence and references contributing most heavily to the causality determination and, where applicable, to
uncertainties or inconsistencies. References to earlier sections indicate where the full body of evidence is described.
°Describes the ozone concentrations with which the evidence is substantiated.
6.2 Long-Term Ozone Exposure and Mortality
6.2.1 Introduction
A limited number of epidemiologic studies have assessed the relationship between long-term
ozone exposure and mortality in adults. The 2006 Ozone AQCD concluded that an insufficient amount of
evidence existed "to suggest a causal relationship between chronic ozone exposure and increased risk for
mortality in humans" (U.S. EPA. 2006). Noting limited support for an association with long-term ozone
exposure and total mortality, and inconsistent associations for cardiopulmonary mortality from the ACS
and Harvard Six Cities studies, the 2013 Ozone ISA concluded that the evidence was suggestive of a
causal relationship between long-term ozone exposure and total mortality (U.S. EPA. 2013a). The
strongest evidence for an association between long-term ozone exposure and mortality was derived from
associations with respiratory mortality reported by Jerrett et al. (2009) that remained robust after adjusting
for PM2 5 concentrations and an analysis that reported associations of ambient ozone concentrations and
total mortality among populations with pre-existing disease in the Medicare Cohort (Zanobetti and
Schwartz. 2011).
The following section provides a brief, integrated evaluation of evidence for long-term ozone
exposure and mortality presented in the previous NAAQS review with evidence that is newly available
6-25
-------
for this review. This section focuses on assessing the degree to which newly available studies further
characterize the relationship between long-term ozone exposure and mortality. For example, areas of
research that inform differences in the exposure window used to evaluate long-term exposures and
mortality or comparisons of statistical techniques are highlighted. Studies that address the variability in
the associations observed across ozone epidemiologic studies due to exposure error and the use of
different exposure assessment techniques are emphasized. Another important consideration is
characterizing the shape of the C-R relationship across the full concentration range observed in
epidemiologic studies. The evidence in this section focuses on epidemiologic studies because
experimental studies of long-term exposure and mortality are generally not conducted. However, this
section draws from the morbidity evidence presented for different health endpoints across the scientific
disciplines (i.e., animal toxicological, epidemiologic, and controlled human exposure studies) to support
the associations observed for cause-specific mortality.
6.2.1.1 Population, Exposure, Comparison, Outcome, and Study Design (PECOS)
Tool
The scope of this section is defined by a scoping tool that generally defines the relevant
Population, Exposure, Comparison, Outcome, and Study Design (PECOS). The PECOS tool defines the
parameters and provides a framework to help identify the relevant evidence in the literature to inform the
ISA. Because the 2013 Ozone ISA concluded that evidence existed to suggest a causal relationship
between long-term ozone exposure and total mortality the studies evaluated are less limited in scope and
not targeted towards specific study locations, as reflected in the PECOS tool. The studies evaluated and
subsequently discussed within this section were identified using the following PECOS tool:
• Population: Any population, including populations or lifestages that might be at increased risk
• Exposure: Long-term ambient concentration of ozone
• Comparison: Per unit increase (in ppb)
• Outcome: Change in risk (incidence/prevalence) of mortality
• Study Design: Epidemiologic cohort studies; time-series, case-crossover, and cross-sectional
studies with appropriate timing of exposure for the health endpoint of interest
6.2.2 Biological Plausibility
The preceding appendices characterized evidence related to evaluating the biological plausibility
by which long-term ozone exposure may lead to the morbidity effects that are the most common causes of
total (nonaccidental) mortality, specifically cardiovascular and respiratory morbidity and metabolic
disease (Appendix 4. Appendix 3. and Appendix 5. respectively). Respiratory and cardiovascular
morbidity comprise ~9 and -33%, respectively, of total mortality (NHLBI. 2017). This evidence is
6-26
-------
derived from animal toxicological, controlled human exposure, and epidemiologic studies. Appendix 3
characterizes the available evidence by which inhalation exposure to ozone could progress from initial
events to endpoints relevant to the respiratory system and to population outcomes such as exacerbation of
COPD. Appendix 4 outlines the available evidence for plausible mechanisms by which inhalation
exposure to ozone could progress from initial events to endpoints relevant to the cardiovascular system
and to population outcomes such as IHD, stroke, and atherosclerosis. Appendix 5 outlines the available
evidence for plausible mechanisms by which inhalation exposure to ozone could progress from initial
events (e.g., pulmonary inflammation, autonomic nervous system activation) to intermediate endpoints
(e.g., insulin resistance, increased blood glucose and lipids) and result in population outcomes such as
metabolic disease and diabetes. Collectively, the progression demonstrated in the available evidence for
respiratory morbidity and metabolic disease supports potential biological pathways by which long-term
ozone exposures could result in mortality; however, for cardiovascular morbidity, the evidence is more
limited due to the few studies that provide generally inconsistent results between experimental and
epidemiologic studies.
6.2.3 Total (Nonaccidental) Mortality
When considering the entire body of evidence, there is limited support for an association with
long-term ozone exposure and total mortality. Recent studies use fixed-site monitors and models
(e.g., CMAQ, dispersion models) measure, estimate, or predict ozone concentrations for use in assigning
long-term ozone exposure in epidemiologic studies. There are also hybrid methods that combine two or
more fixed-site, mode, and/or satellite-based techniques (Appendix 2 Section 2.3). Generally,
epidemiologic studies of long-term ozone exposure and total mortality use a long-term average
(e.g., annual or seasonal average) of the 8-hour daily max ozone metric, though there are some studies
that use the 24-hour avg [e.g., Sese et al. (2017)1. or the 1-hour daily max [e.g., Jerrett et al. (2009)1
metric when calculating long-term average concentrations. The exposure metric used in each study is
recorded in the Evidence Inventory tables (Section 6.3.2) for each study when that information was
reported by study authors. The strongest evidence comes from analyses of the Medicare cohort data,
including a study observing positive associations among different cohorts with pre-existing disease
(Zanobetti and Schwartz. 2011) included in the 2013 Ozone ISA, and a recent analysis of more than
61 million individuals in the Medicare cohort (Di et al.. 2017b). Results from other recent studies are less
consistent, with some U.S. and Canadian cohorts reporting modest positive associations between
long-term ozone exposure and total mortality, while other recent studies conducted in the U.S, Europe,
and Asia report null or negative associations. The differences in the way exposure to ozone was assessed
do not explain the heterogeneity in the observed associations. The results from studies evaluating
long-term ozone exposure and total mortality are presented in Figure 6-8. These studies are characterized
in Table 6-6. Overall, there is some evidence that long-term ozone exposure is associated with total
6-27
-------
mortality, especially among individuals with pre-existing disease, but the evidence is not consistent across
studies. Specifically:
• The strongest evidence for an association between long-term ozone exposure and total mortality
comes from an analysis among four subcohorts with pre-existing disease from the Medicare
cohort (Zanobctti and Schwartz. 2011). demonstrating positive associations among those with
pre-existing heart failure, MI, diabetes, or COPD. A recent analysis of the entire Medicare cohort,
including over 61 million older adults, observed positive associations between long-term ozone
exposure and total mortality, even when limited to areas in the U.S. where the predicted annual
average ozone concentrations were less than 50 ppb (Pi et al.. 2017b).
• Several recent analyses of the CanCHEC cohort in Canada provide additional evidence of a
modest positive association [consistent in magnitude with the association reported by Di et al.
(2017bVI between long-term ozone exposure and total mortality (Cakmak et al.. 2018;
Weichenthal et al.. 2017; Cakmak et al.. 2016; Crouse et al.. 2015V
• A recent study conducted in California among a cohort of individuals with cancer observed a
positive association between long-term ozone exposure and total mortality (Eckel et al.. 2016).
• Results from the ACS cohort provide little evidence for an association between long-term ozone
exposure and total mortality (Turner et al.. 2016; Jerrett et al.. 2013; Jerrett et al.. 2009).
• Several studies conducted outside of North America report negative associations between
long-term ozone exposure and total mortality, specifically in France (Bentaveb et al.. 2015). the
U.K. (Carey et al.. 2013). and South Korea (Kim et al.. 2017).
6-28
-------
Reference
Cohort
Notes
Years Mean (ppb)
Jerrett et al. 2009 ACS
Zanobetti & Schwartz, 2011 Medicare
tTurner etal. 2016
tJerrett et al. 2013
tDietal. 2017
tEckel et al. 2016
tCrouse etal. 2015
tCakmak et al. 2018
tWeichenthal et al. 2017
tBentayeb etal. 2015
tCarey et al. 2013
tKimetal. 2017
Heart failure cohort
Ml cohort
Diabetes cohort
COPD cohort
1982-2000 57.5
1985-2006 15.6-71.4
ACS
ACS
Medicare
Full cohort
1982-2004 38.2
1982-2000 50.35
2000-2012 46.3
CA Cancer Cohort
CanCHEC
CanCHEC
CanCHEC
Gaze I
English Medical Practice
NHIS-NSC
Low exposure (<50 ppb) cohort
1988-2011 40.2
1991-2006 39.6
1991-2011 15.0-43.0
1991-2011 38.29
1989-2013 40.5 -
2003-2007 25.85 —
2007-2013 19.93
I-
-+-
!•
-t
0.6 0.8 1 1.2 1.4
Hazard Ratio (95% CI)
ACS = American Cancer Society; CanCHEC = Canadian Census Health and Environment Cohort; NHIS-NSC = National Health
Insurance Service—National Sample Cohort.
Note: tStudies published since the 2013 Ozone ISA. Associations are presented per 10 ppb increase in pollutant concentration.
Circles represent point estimates; horizontal lines represent 95% confidence intervals for ozone. Black text and circles represent
evidence included in the 2013 Ozone ISA; red text and circles represent recent evidence not considered in previous ISAs or
AQCDs.
Figure 6-8 Associations between long-term exposure to ozone and total
(nonaccidental) mortality in recent cohort studies.
6-29
-------
6.2.3.1
Respiratory Mortality
When considering the entire body of evidence, there is limited support for an association with
long-term ozone exposure and respiratory mortality. Recent studies use both fixed-site monitors and
models (e.g., CMAQ, dispersion models) to measure or estimate ozone concentrations for use in
assigning long-term ozone exposure in epidemiologic studies (Appendix 2. Section 2.3). The strongest
evidence comes from analyses of the ACS cohort data, including studies observing positive associations
between long-term ozone exposure and respiratory mortality (Jerrett et al.. 2009). and a recent analysis of
respiratory, COPD, and pneumonia mortality (Turner et al.. 2016). Results from other recent studies are
less consistent, with analyses of U.S., Canadian, and European cohorts reporting inconsistent associations
between long-term ozone exposure and respiratory mortality. The differences in the way exposure to
ozone was assessed do not explain the heterogeneity in the observed associations. The results from studies
evaluating long-term ozone exposure and respiratory mortality are presented in Figure 6-9. These studies
are characterized in Table 6-7. Overall, there is some evidence that long-term ozone exposure is
associated with respiratory mortality, but the evidence is not consistent across studies. Specifically:
• The strongest evidence for an association between long-term ozone exposure and respiratory
mortality comes from nationwide analyses of the ACS cohort, demonstrating positive associations
with respiratory mortality (Turner et al.. 2016; Jerrett et al.. 2009) and COPD, and pneumonia/flu
(Turner etal.. 2016). In contrast, Jerrett et al. (2013) reported a null association between
long-term ozone exposure and respiratory mortality in an analysis of the ACS cohort limited to
participants from California.
• Several recent analyses of the CanCHEC cohort in Canada provide inconsistent evidence for an
association between long-term ozone exposure and respiratory mortality, with one reporting a
positive association (Weichenthal et al.. 2017) and the other reporting a negative association
(Crouse et al.. 2015). Cohort studies conducted in France (Bentaveb et al.. 2015) and the U.K.
(Carey et al.. 2013) also report negative associations between long-term ozone exposure and
respiratory mortality.
6-30
-------
Reference
Jerrett et al. 2009
tTurner et al. 2016
tJerrett et al. 2013
Cohort
ACS
ACS
ACS
tCrouse etal. 2015 CanCHEC
tWeicherithal et al. 2017 CanCHEC
tBentayeb etal. 2015 Gazel
Years
1982-2000
1982-2004
1982-2000
1991-2006
1991-2011
1989-2013
tCarey etal. 2013 English Medical Practice 2003-2007
tTurner et al. 2016
tTurner et al. 2016
ACS
ACS
Mean
57.5
38.2
50.35
39.6
38.29
40.5
25.85 < •
1982-2004 38.2
1982-2004 38.2
I-
Outcome
Respiratory Mortality
COPD Mortality
Pneumonia/Flu Mortality
-1
0.5 0.75 1 1.25 1.5 1.75
Hazard Ratio (95% CI)
ACS = American Cancer Society; CanCHEC = Canadian Census Health and Environment Cohort.
Note: tStudies published since the 2013 Ozone ISA. Associations are presented per 10 ppb increase in pollutant concentration.
Circles represent point estimates; horizontal lines represent 95% confidence intervals for ozone. Black text and circles represent
evidence included in the 2013 Ozone ISA; red text and circles represent recent evidence not considered in previous ISAs or
AQCDs.
Figure 6-9 Associations between long-term exposure to ozone and
respiratory mortality in recent cohort studies.
6.2.3.2 Cardiovascular Mortality
Recent cohort studies extend the body of evidence for the relationship between long-term ozone
exposure and cardiovascular mortality. The 2013 Ozone ISA noted inconsistent evidence for
cardiopulmonary mortality, and there was limited evidence for the association between long-term ozone
exposure and cardiovascular mortality based on an analysis of the ACS cohort (Jerrett et al.. 2009).
Recent analyses from the ACS cohort in the U.S. and the CanCHEC cohort in Canada provide consistent
evidence for positive associations between long-term ozone exposure and cardiovascular and IHD
mortality, as well as mortality due to diabetes or cardiometabolic diseases. Associations with mortality
6-31
-------
due to cerebrovascular disease (e.g., stroke) were less consistent, and generally closer to the null value.
Other recent studies conducted in the Europe and Asia report null or negative associations. Similar to total
mortality, the differences in the way exposure to ozone was assessed do not explain the heterogeneity in
the observed associations for cardiovascular mortality. The results from studies evaluating long-term
ozone exposure and cardiovascular mortality are presented in Figure 6-10. These studies are characterized
in Table 6-8 and Table 6-9. Overall, there is increased evidence that long-term ozone exposure is
associated with cardiovascular mortality compared to the evidence included in the 2013 Ozone ISA.
Specifically:
• The strongest evidence for an association between long-term ozone exposure and cardiovascular
mortality comes from nationwide analyses of the ACS cohort, demonstrating positive associations
with cardiovascular mortality (Turner et al.. 2016; Jerrett et al.. 2013; Jerrett et al.. 2009). IHD
mortality (Jerrett et al.. 2013). cerebrovascular disease mortality (Turner et al.. 2016). and
mortality due to dysrhythmia and heart failure (Turner et al.. 2016).
• Several recent analyses of the CanCHEC cohort in Canada provide consistent evidence for a
positive association between long-term ozone exposure and cardiovascular and IHD mortality
(Cakmak et al.. 2018; Cakmak et al.. 2016; Crouse et al.. 2015).
• Cohort studies conducted in France (Bentaveb et al.. 2015). the U.K. (Carey et al.. 2013). and
South Korea (Kim et al.. 2017) report negative associations between long-term ozone exposure
and respiratory mortality.
• Several recent studies conducted in the U.S. and Canada provide limited and inconsistent
evidence for an association between long-term ozone exposure and mortality due to
cerebrovascular disease (Figure 6-10).
• A limited body of evidence demonstrates positive associations between long-term ozone exposure
and mortality from diabetes and cardiometabolic diseases (Turner et al.. 2016; Crouse et al..
2015V
6-32
-------
Reference Cohort
Jerrett et al. 2009 ACS
TTurner etal. 2016 ACS
TJerrett et al. 2013 ACS
TCrause et al. 2015 CanCHEC
TCakmak et al. 2016 CariCHEC
TWeichenthal et al. 2017
TBeritayeb et a I. 2015
TCarey et al. 2013
TKimetal. 2017
Jerrett et aI. 2009
TTurner et al. 2016
TJerrett et al. 2013
TC rouse et al. 2015
TCakmak et al. 2016
TCakmak et al. 2018
TTurner et al. 2016
TJerrett et al. 2013
TC rouse et al. 2015
TCakmak et al. 2016
TTurner etal. 2016
TC rouse et al. 2015
Notes
Circulatory D isease
Cardiovascular Disease
Base Model
Adj. for C limate Zone
CanCHEC
Gaze I
English Medical Practice
N HIS-NSC
ACS
ACS
ACS
CanCHEC
CanCHEC
CanCHEC
ACS
ACS
CanCHEC
CanCHEC
ACS
CanCHEC
TTurner etal. 2016 ACS
Base Model
Adj. for C limate Zone
Base Model
Adj. for C limate Zone
Diabetes
Cardiometabolic D isease
Years
1982-2000
1982-2004
1982-2000
1991-2006
1991-2006
1991-2011
1989-2013
2003-2007
2007-2013
1982-2000
1982-2004
1982-2000
1991-2006
1991-2006
Mean
57.5
38.2
50.35
39.6
14.3-40.9
38.29
40.5 «
25.85
19.93
1982-2004
1982-2000
1991-2006
1991-2006
1982-2004
1991-2006
1982-2004 38.2
f—
0.5
57.5
38.2
50.35
39.6
14.3-40.9
1991-2011 15.0-43.0
Outcome
Cardiovascular Mortality
IHD Mortality
38.2
50.35
39.6
14.3-40.9
38.2
39.6
CBVD Mortality
Diabetes Mortality
Other CV Mortality
0.75 1 125 1.5 1.75
Hazard Ratio (95% CI)
ACS = American Cancer Society; CanCHEC = Canadian Census Health and Environment Cohort; CBVD = cerebrovascular
disease; CV = cardiovascular; IHD = ischemic heart disease; NHIS-NSC = National Health Insurance Service—National Sample
Cohort.
Note: tStudies published since the 2013 Ozone ISA. Associations are presented per 10 ppb increase in pollutant concentration.
Circles represent point estimates; horizontal lines represent 95% confidence intervals for ozone. Black text and circles represent
evidence included in the 2013 Ozone ISA; red text and circles represent recent evidence not considered in previous ISAs or
AQCDs.
Figure 6-10 Associations between long-term exposure to ozone and
cardiovascular mortality in recent cohort studies.
6.2.3.3 Studies of Life Expectancy
A recent study adds to the body of evidence on the relationship between long-term ozone
exposure and mortality by examining temporal trends in ozone concentrations and changes in life
expectancy, testing the hypothesis that populations living in areas with higher ozone concentrations have
lower life expectancies. Li et al. (2016) reported the mean life expectancy for males and females in the
U.S. from 2002 to 2008 at the county level, separating counties into three classes based on average ozone
concentrations: Class 1: 36.4 ppb (28.1-39.8); Class 2: 43.3 ppb (39.4, 46.2); and Class 3: 48.8 ppb
6-33
-------
(45.7-54.5). Nationwide, ozone concentrations reduced an avg of 0.15 ppb during the study period. After
adjustment for PM2 5 concentrations, life expectancy decreased by 0.2 and 0.6 years for males in Classes 2
and 3 counties (respectively, compared to counties in Class 1) and by 0.3 and 0.6 years for females in
Classes 2 and 3 counties (respectively, compared to counties in Class 1). When the study authors
evaluated the association for all counties on a continuous scale, they observed a 0.25 (0.19, 0.30) year
decrease in life expectancy for males and 0.21 (0.17, 0.25) year decrease in life expectancy for females
for every 5-ppb increase in average ozone concentration.
6.2.4 Effect Modification of the Ozone-Mortality Relationship
6.2.4.1 Pre-existing Disease
Individuals with certain pre-existing diseases may be considered at greater risk of an air
pollution-related health effect because they are likely in a compromised biological state that can vary
depending on the disease and severity. The 2013 Ozone ISA concluded that there was adequate evidence
for increased ozone-related health effects among individuals with asthma (U.S. EPA. 2013a). The results
of controlled human exposure studies, as well as epidemiologic and animal toxicological studies,
contributed to this evidence. Studies of short-term ozone exposure and mortality provided limited
evidence for stronger associations among individuals with pre-existing cardiovascular disease or diabetes.
When evaluating long-term ozone exposure and mortality, Zanobetti and Schwartz (2011) observed
positive associations with total mortality among individuals in the Medicare cohort with a recent hospital
admission for heart failure, MI, diabetes, or COPD, although the authors did not provide quantitative
results for a comparison population with no recent hospital admissions in their analysis.
A limited number of recent studies provides some evidence that individuals with pre-existing
diseases may be at greater risk of mortality associated with long-term ozone exposure. These studies
focus on specific diseases of varying severity (e.g., acute respiratory distress syndrome, pulmonary
fibrosis, ovarian cancer). In contrast, an analysis of the ACS cohort observed stronger associations
between long-term ozone exposure and respiratory or cardiovascular mortality among individuals with no
pre-existing disease. Specifically:
• The strongest evidence that individuals with pre-existing disease might be at greater risk of total
mortality associated with long-term ozone exposure continues to come from a study of four
disease cohorts (i.e., individuals with a recent hospital admission related to heart failure, MI,
diabetes, or COPD) among members of the Medicare cohort that observed positive and
statistically significant associations for each of the disease cohorts rZanobetti and Schwartz
(2011); Table 6-61. A recent study of the ACS cohort reported contrasting results, with stronger
associations among populations with no pre-existing respiratory or cardiovascular disease and
respiratory or cardiovascular mortality, respectively.
6-34
-------
• In addition, several studies reported positive associations between long-term ozone exposure and
total or cancer-specific mortality among those already diagnosed with ovarian cancer (Vieira et
al.. 2017) or respiratory cancer [i.e., cancers of the nose, nasal cavity and middle ear, larynx, lung
and bronchus, pleura and trachea, mediastinum, and other organs; Xu et al. (2013)1.
• Positive associations were observed between long-term ozone exposure and in-hospital mortality
among patients with acute respiratory distress syndrome (Rush et al.. 2017). but not with total
mortality among individuals with idiopathic pulmonary fibrosis (Sese et al.. 2017).
6.2.4.2 Lifestage
The 1996 and 2006 Ozone AQCDs identified children, especially those with asthma, and older
adults as at-risk populations (U.S. EPA. 2006. 1996a). In addition, the 2013 Ozone ISA concluded that
there was adequate evidence to conclude that children and older adults are at increased risk of
ozone-related health effects (U.S. EPA. 2013a). Collectively, the majority of evidence for older adults has
come from studies of short-term ozone exposure and mortality, with little evidence contributed by studies
of long-term ozone exposure. A limited number of recent studies of long-term exposure to ozone and
mortality have compared associations between different age groups, but do not report consistent evidence
that older adults are at increased risk:
• Turner et al. (2016) observed stronger associations among those less than 65 years old compared
to those 65 years and older.
• Results of the CanCHEC cohort observed positive associations between long-term ozone
exposure and total mortality that were similar among women aged less than 60, 60-69, and
70-79 years (Crouse et al.. 2015). This association was attenuated to null among women aged
80-89 years. For men in CanCHEC cohort, Crouse et al. (2015) observed positive associations
between long-term ozone exposure and total mortality among men aged less than 60 and
60-69 years, and these associations were attenuated, but remained positive for men aged 70-79
and 80-89 years.
6.2.5 Potential Copollutant Confounding of the Ozone-Mortality Relationship
The evaluation of potential confounding effects of copollutants on the relationship between
long-term ozone exposure and mortality allows for examination of whether ozone risk estimates are
changed in copollutant models. Year-round correlations of ozone concentrations with copollutant
concentrations can be found in Section 2.5; generally, the strongest positive correlations are with PMio
and PM2 5, while the strongest negative correlations are observed with CO. Recent studies examined the
potential for copollutant confounding by evaluating copollutant models that include PM2 5 (Figure 6-11)
and NO2. These recent studies address a previously identified data gap by informing the extent to which
effects associated with long-term ozone exposure are independent of coexposure to correlated
copollutants in long-term analyses:
6-35
-------
• The 2013 Ozone ISA included the study by Jerrett et al. (2009) that reported associations with
respiratory mortality that remained robust after adjustment for PM2 5, and associations with
cardiovascular mortality that were attenuated, changing from positive to negative, after
adjustment for PM2 5 concentrations. Recent studies (Figure 6-11) provide generally consistent
evidence for associations with ozone that are robust (i.e., relatively unchanged) to adjustment for
PM2 5 concentrations for total mortality, respiratory mortality, and cardiovascular mortality.
• The correlations between ozone and PM2 5 exposures in studies that conducted copollutant
analyses were highly variable, ranging from -0.705 to 0.73, and included low (e.g., <0.4),
moderate (e.g., 0.4-0.7), and high (e.g., >0.7) correlations (Table 6-6).
• Jerrett et al. (2013) reported copollutant models with ozone and NO2. The correlation between
ozone and NO2 concentrations was weak (r = -0.071), and associations with ozone were robust to
inclusion of NO2 in the model for total, respiratory, and cardiovascular mortality.
6-36
-------
Reference
Jerrett et al. 2009
ITurner etal. 2016
IJerrett et al. 2013
IDietal. 2017
ICakmaketal. 2018
Jerrett et al. 2009
ITurner etal. 2016
IJerrett et al. 2013
IJerrett et al. 2013
ICakmaketal. 2018
Jerrett et al. 2009
ITurner etal. 2016
IJerrett et al. 2013
ICakmaketal. 2016
Jerrett et al. 2009
IJerrett et al. 2013
ICakmaketal. 2018
IJerrett et al. 2013
IJerrett et al. 2013
Cohort
ACS
ACS
ACS
Medicare
CanCHEC
ACS
ACS
ACS
ACS
CanCHEC
ACS
ACS
ACS
CANCHEC
ACS
ACS
CanCHEC
ACS
ACS
Years
1982-2000
1982-2004
1982-2000
2000-2012
1991-2011
1982-2000
1982-2004
1982-2000
1982-2000
1991-2011
1982-2000
1982-2004
1982-2000
1991-2006
1982-2000
1982-2000
1991-2011
1982-2000
1982-2000
Means
57.5
38.2
50.35
46.3
15.0-43.0
57.5
38.2
50.35
50.35
15.0-43.0
57.5
38.2
50.35
14.3-40.9
57.5
50.35
15.0-43.0
50.35
50.35
Notes
+PM2.5
+PM2.5
+PM2.5
+PM2.5
+PM2.5
+PM2.5
+PM2.5
+PM2.5
+PM2.5
+PM2.5
+PM2.5
+PM2.5
+PM2.5
+PM2.5
+PM2.5
+PM2.5
+PM2.5
+PM2.5
+PM2.5
Total Mortality
Respiratory Mortality
Lung Cancer Mortality
Cardiovascular Mortality
IHD Mortality
Stroke Mortality
Other Mortality
-I
0.9 1 1.1 1.2
Hazard Ratio (95% CI)
ACS = American Cancer Society; CanCHEC = Canadian Census Health and Environment Cohort; IHD = ischemic heart disease.
Note: tStudies published since the 2013 Ozone ISA. Associations are presented per 10 ppb increase in pollutant concentration.
Circles represent point estimates; horizontal lines represent 95% confidence intervals for ozone. Black text and circles represent
evidence included in the 2013 Ozone ISA; red text and circles represent recent evidence not considered in previous ISAs or
AQCDs. Closed circles represent effect of ozone in single pollutant models, open circles represent effect of ozone adjusted for
PM2.5.
Figure 6-11 Associations between long-term exposure to ozone and mortality
with and without adjustment for PM2.5 concentrations in recent
cohort studies.
6.2.6 Shape of the Concentration-Response Function
An important consideration in characterizing the ozone-mortality association is whether the
concentration-response (C-R) relationship is linear across the full concentration range that is encountered
or there are concentration ranges that depart from linearity. The epidemiologic studies included in the
2013 Ozone ISA indicated a "generally linear C-R function with no indication of a threshold" (U.S. EPA.
2013a). With regard to studies of long-term ozone exposure and mortality, a threshold analysis indicated
that the linear model was not a better fit to the data (p > 0.05) than a threshold representation of the
overall ozone-mortality association (Jerrett et al.. 2009); however, the authors reported "limited evidence"
for an effect threshold at an ozone concentration (seasonal avg of 1-hour max) of 56 ppb (p = 0.06).
6-37
-------
Visual inspection of this concentration-response function suggests an inflection point just below 60 ppb,
which is close to the median concentration across cities (i.e., 57 ppb).
A number of recent studies examined the C-R function between long-term ozone exposure and
mortality and observed somewhat inconsistent results. While some studies provide evidence of a
generally linear C-R function, others observed a sublinear relationship, indicating larger changes in risk
for higher ozone concentrations compared to lower ozone concentrations. Several studies also include
threshold analyses and support the possibility of a threshold near 35 to 40 ppb (8-hour max). Specifically:
• In the U.S. Medicare cohort, Di et al. (2017b) used thin-plate spline regression to evaluate the
C-R relationship between long-term ozone exposure and total mortality and observed a generally
linear function with no signal of a threshold down to 30 ppb (8-hour max; Figure 6-12. Panel A).
When Crouse et al. (2015) used restricted cubic spline functions to evaluate the C-R function for
long-term ozone exposure and total mortality in the CanCHEC cohort, they observed a sublinear
relationship, indicating larger changes in risk for higher ozone concentrations compared with
lower ozone concentrations (Figure 6-13).
• Among studies conducting threshold analyses, Di et al. (2017b) reported evidence for a threshold
at around 40 ppb (8-hour daily max) based on the minimum AIC value and visual inspection of
the C-R function (Figure 6-12. Panel B).
• The C-R relationship between long-term ozone exposure and mortality may differ by the cause of
mortality and/or by ozone season. For example, Turner et al. (2016) reported improved model fit
for a threshold model compared to a linear model for the association between long-term,
year-round ozone exposure and respiratory mortality in the ACS cohort, with evidence of a
threshold near 35 ppb (8-hour daily max; Figure 6-14). However, when the data were restricted to
warm-season ozone only, the linear model was a better fit than the threshold model.
6-38
-------
s
§
2
¦
QC
¦n
I
—
X
§
§
1 020-
at
"5
i
a
z
1000-
40
45
SO
55
60
Ozone (ppb) Ozone (ppb)
Note: The solid line represents the estimate and the dotted lines represent the 95% confidence interval for the estimate.
Reprinted with permission from the publisher; Pi et al. (2017b).
Figure 6-12 The concentration-response relationship estimated with log-linear
model with a thin-plate spline (A; left panel) and the
concentration-response relationship estimated with threshold
model (B; right panel), indicating the potential for a threshold at
40 ppb (8-hour daily max).
CO
03
N
03
X
O
CD
O)
O
30
35
40
45
50
03 - ppb
Note: The solid blue line represents the estimate and the gray shaded areas represent the 95% confidence interval for the estimate.
Reprinted with permission from the publisher; Crouse et al. (20151.
Figure 6-13 Concentration-response relationship between ozone
concentrations (parts per billion [ppb]) and total (nonaccidental)
mortality in the CanCHEC cohort (mean 39.6; knots: 30.0, 38.9,
50.7 ppb).
6-39
-------
CM
O
o
30
35
40
45
50
Note: Mean annual 8-hour daily max ozone concentration (ppb), hierarchical Bayesian space-time Model (HBM), U.S., 2002-2004,
truncated at 99th percentile. Gray line along abscissa indicates data density.
Note: The solid line represents the estimate and the dotted lines represent the 95% confidence interval for the estimate.
Reprinted with permission from the publisher; Turner et al. (2016).
Figure 6-14 Concentration-response curve for ozone associated with
respiratory mortality using a natural spline model with 3 degrees
of freedom.
6.2.7 Summary and Causality Determination
This section describes the evaluation of evidence for total (nonaccidental) mortality based on the
scientific considerations detailed in the Annex for Appendix 6. with respect to the causality determination
for long-term exposures to ozone using the framework described in the Preamble to the ISAs (U.S. EPA.
2015). The key evidence, as it relates to the causal framework, is summarized in Table 6-2. Recent cohort
studies provide limited support for the association between long-term ozone exposure and total mortality,
with some U.S. and Canadian cohorts reporting modest positive associations between long-term ozone
exposure and total mortality, while other recent studies conducted in the U.S, Europe, and Asia report null
or negative associations. The strongest evidence for an association between long-term ozone exposure and
total mortality continues to come from analyses of the Medicare cohort data included in the 2013 Ozone
6-40
-------
ISA. Specifically, Zanobetti and Schwartz (2011) reported associations of ambient ozone concentrations
and total mortality among populations with pre-existing disease that remained robust after adjusting for
PM2 5 concentrations.
Additionally, Jerrett et al. (2009) reported positive associations between long-term ozone
exposure and respiratory mortality after adjusting for PM2 5 in copollutant models. This evidence is
supported by a recent analysis of respiratory, COPD, and pneumonia mortality (Turner et al.. 2016).
Results from other recent studies are less consistent, with analyses of U.S., Canadian, and European
cohorts reporting inconsistent associations between long-term ozone exposure and respiratory mortality.
Whereas the 2013 Ozone ISA noted inconsistent evidence for cardiopulmonary mortality (and
limited evidence for cardiovascular mortality, specifically), recent cohort studies extend the body of
evidence for the relationship between long-term ozone exposure and cardiovascular mortality. Analysis of
the ACS cohort provided limited evidence for the association between long-term ozone exposure and
cardiovascular mortality (Jerrett et al.. 2009) in the 2013 Ozone ISA. Recent analyses from the ACS
cohort in the U.S. and the CanCHEC cohort in Canada provide consistent evidence for positive
associations between long-term ozone exposure and cardiovascular and IHD mortality, as well as
mortality due to diabetes or cardiometabolic diseases. Associations with mortality due to cerebrovascular
disease (e.g., stroke) are less consistent, and generally closer to the null value. Other recent studies
conducted in the Europe and Asia report null or negative associations.
Additionally, recent studies that have evaluated copollutant confounding for a limited number of
pollutants reduce uncertainties related to potential copollutant confounding by PM2 5 and NO2
(Section 6.2.5) and contribute to the previously limited evidence characterizing the shape of the
concentration-response relationship (Section 6.2.6). Recent evidence helps to reduce uncertainties related
to potential copollutant confounding by two pollutants of the relationship between long-term ozone
exposure and mortality. Multiple studies evaluated PM2 5 (Figure 6-9). while fewer evaluated NO2 in
copollutant models, and observed similar hazard ratios for ozone regardless of whether PM2 5 or NO2 were
included in the model. This helps to reduce the uncertainty for an independent effect of long-term ozone
exposure on mortality.
The body of evidence for total mortality is supported by generally consistent positive associations
with cardiovascular mortality, and less so by the somewhat inconsistent evidence for respiratory
mortality. There is coherence across the scientific disciplines (i.e., animal toxicology and epidemiology)
and biological plausibility for ozone-related respiratory (Appendix 3) endpoints, which lend some
additional support to the ozone-mortality relationship.
Recent studies use a variety of both fixed-site monitors and models (e.g., CMAQ, dispersion
models) to measure or estimate ozone concentrations for use in assigning long-term ozone exposure in
epidemiologic studies (Appendix 2. Section 2.3). Overall, the exposure assessment techniques used in
these studies do not help to explain the inconsistent associations observed across studies, although they
6-41
-------
indicate that the observed effects of long-term ozone exposure on mortality are not overtly influenced by,
or a residual of, the exposure assessment technique used in the study.
The number of studies examining the shape of the C-R function for long-term ozone exposure
and mortality has substantially increased since the 2013 Ozone ISA. These studies used a number of
different statistical techniques to evaluate the shape of the C-R function, including linear models and
restricted cubic splines, and generally observed linear, no-threshold relationships down to 35-40 ppb,
although the results are not entirely consistent. Some studies observed a sublinear relationship, indicating
larger changes in risk for higher ozone concentrations compared with lower ozone concentrations. Several
studies also included threshold analyses and support the possibility of a threshold near 35 to 40 ppb.
Overall, recent epidemiologic studies add to the limited body of evidence that formed the basis of
the conclusions of in 2013 Ozone ISA for total mortality. This body of evidence is generally inconsistent,
with some U.S. and Canadian cohorts reporting modest positive associations between long-term ozone
exposure and total mortality, while other recent studies conducted in the U.S, Europe, and Asia report null
or negative associations. The strongest evidence for the association between long-term ozone exposure
and total (nonaccidental) mortality continues to come from analyses of patients with pre-existing disease
from the Medicare cohort, and recent evidence demonstrating positive associations with cardiovascular
mortality. The evidence from the assessment of ozone-related respiratory disease, with more limited
evidence from cardiovascular and metabolic morbidity, provides biological plausibility for mortality due
to long-term ozone exposures. In conclusion, the inconsistent associations observed across both recent
and older cohort and cross-sectional studies conducted in various locations provide limited evidence for
an association between long-term ozone exposure and mortality. Collectively, this body of evidence is
suggestive of, but not sufficient to infer, a causal relationship between long-term ozone exposure
and total mortality.
6-42
-------
Table 6-2 Summary of evidence that is suggestive of, but not sufficient to infer,
a causal relationship between long-term ozone exposure and total
mortality.
Rationale for
Causality
Determination3
Key Evidence13
Key References'3
Ozone
Concentrations
Associated
with Effects0
Limited and sometimes
inconsistent
epidemiologic evidence
from multiple,
high-quality studies at
relevant ozone
concentrations
Positive associations
between long-term ozone
exposure and total mortality
among those with
pre-existing disease in the
ACS cohorts, with some
additional evidence from
recent studies stratifying by
disease status.
Zanobetti and Schwartz (20111
Section 6.2.4.1
Mean
concentrations
across studies:
15.6-71.4 ppb
Recent analyses from the
ACS cohort in the U.S. and
the CanCHEC cohort in
Canada provide consistent
evidence for positive
associations between
long-term ozone exposure
and cardiovascular and IHD
mortality, as well as mortality
due to diabetes or
cardiometabolic diseases.
Jerrett et al. (2009); Turner et al. (2016);
Jerrett et al. (2013)
Crouse et al. (2015): Cakmak et al.
(2016): Cakmak et al.
Section 6.2.3.2
(2018)
Mean
concentrations
across studies:
14.3-57.5 ppb
Strona evidence from the Jerrett et al. (2009): Turner et al. (2016)
Mean
ACS cohort demonstrating Section 6 2 3 1
concentrations
positive associations
across studies:
between long-term ozone
15.0-57.5
exposure and respiratory
mortality. Results from other
recent studies are less
consistent, with analyses of
U.S., Canadian, and
European cohorts reporting
inconsistent associations.
Some recent U.S. and
Canadian cohorts report
modest positive associations
with total mortality, while
other recent studies
conducted in the U.S.,
Europe, and Asia report null
or negative associations.
Section 6.2.3
Mean
concentrations
across studies:
15-71.4 ppb
6-43
-------
Table 6-2 (Continued): Summary of evidence that is suggestive of, but not
sufficient to infer, a causal relationship between long-term
ozone exposure and total mortality.
Rationale for
Causality
Determination3
Key Evidence13
Key References'3
Ozone
Concentrations
Associated
with Effects0
Epidemiologic
evidence from
copollutant models
provides some support
for an independent
ozone association
Positive associations
observed between long-term
ozone exposure and total
mortality remain relatively
unchanged after adjustment
for PM2.5 and NO2.
When reported, correlations
with copollutants were highly
variable (low to high).
Section 6.2.5
Figure 6-11
Limited epidemiologic
evidence supports a
linear C-R relationship;
some evidence for a
sublinear C-R
relationship
Some studies provide
evidence of a generally
linear C-R relationship;
others observed a sublinear
relationship, indicating larger
changes in risk for higher
compared with lower ozone
concentrations.
Section 6.2.6
Biological plausibility
from studies of
cardiovascular and
respiratory morbidity
and metabolic disease
Evidence for respiratory
morbidity supports potential
biological pathways by which
long-term ozone exposures
could result in mortality;
limited evidence from
cardiovascular morbidity and
metabolic disease.
Appendix 3
Appendix 4
Appendix 5
ACS = American Cancer Society; N02 = nitrogen dioxide; PM2.5 = particulate matter with a nominal aerodynamic diameter less
than or equal to 2.5 |jm; ppb = parts per billion.
aBased on aspects considered in judgments of causality and weight of evidence in causal framework in Table I and Table II of the
Preamble.
bDescribes the key evidence and references contributing most heavily to the causality determination and, where applicable, to
uncertainties or inconsistencies. References to earlier sections indicate where the full body of evidence is described.
°Describes the ozone concentrations with which the evidence is substantiated.
6-44
-------
6.3
Evidence Inventories—Data Tables to Summarize Study Details
6.3.1 Short-Term Ozone Exposure and Mortality: Data Tables
Table 6-3 Epidemiologic studies of short-term exposure to ozone and total (nonaccidental) mortality.
Mean and
Upper
Percentiles
Copollutant
Effect Estimates
Study
Study Population
Exposure Assessment
(PPb)
Examination
Percentage Increase (95% CI)
Bell et al. (2004)
NMMAPS
Average across all
Mean: 26.0
Correlation (r): PM10:
All-year (lag 0-6 DL)
95 U.S. cities
All ages
monitors in each city, 10%
-0.38 to 0.63
0.78 (0.40, 1.16)
1987-2000
trimmed mean to correct
Copollutant models:
Warm/summer (lag 0-6 DL)
for yearly averages of
PM10
0.58 (0.19, 0.97)
Times-series study
each monitor
24-h avg
Lew et al. (2005)
U.S. and non-U.S.
24-h avg
NR
Correlation (r): NR
All-year
U.S. and non-U.S.
Copollutant models:
1.23 (0.94, 1.52)
Meta-analysis
NR
Warm/summer
2.52 (1.70, 3.30)
Bell et al. (2005)
U.S. and non-U.S.
24-h avg
NR
Correlation (r): NR
All-year
U.S. and non-U.S.
Copollutant models:
1.31 (0.82, 1.77)
Meta-analysis
NR
Warm/summer
2.26 (1.09, 3.45)
6-45
-------
Table 6-3 (Continued): Epidemiologic studies of short-term exposure to ozone and total (nonaccidental) mortality.
Study
Study Population Exposure Assessment
Mean and
Upper
Percentiles
(PPb)
Copollutant
Examination
Effect Estimates
Percentage Increase (95% CI)
I to et al. (2005)
U.S. and non-U.S.
Meta-analysis
U.S. and non-U.S. 24-h avg
NR
Correlation (r): NR
Copollutant models:
NR
All-year
1.65 (0.60, 2.69)
Warm/summer
2.61 (1.57, 3.65)
Schwartz (2005)
14 U.S. cities
1986-1993
Case-crossover study
All ages
Average of all monitors in Mean:
each county 35.1-60.0
1-h max 75th:
46.3-69.0
Correlation (r): NR
Copollutant models:
PM10
All-year (lag 0)
0.47 (0.08, 0.87)
Warm/summer (lag 0)
0.63 (0.19, 1.12)
Bell et al. (2007)
98 U.S. communities
1987-2000
Time-series study
NMMAPS
All ages
Average across all
monitors in each city, 10%
trimmed mean to correct
for yearly averages of
each monitor
24-h avg
Mean: 26.0a
Correlation (r): PM2.5:
-0.17 to 0.25; PM10:
<0.00 to 0.22
Copollutant models:
PM2.5, PM10
All-year (lag 0-1)
0.48 (0.26, 0.69)
Bell and Dominici (2008)
98 U.S. communities
1987-2000
Time-series study
NMMAPS
All ages
Average across all
monitors in each city; 10%
trimmed mean to correct
for yearly averages of
each monitor
24-h avg
Mean:
All-year: 26.8
May-
September:
30.0
Maximum:
All-year: 37.3
May-
September:
47.2
Correlation (r): NR
Copollutant models:
PM2.5, PM10
All-year (lag 0-6)
0.78 (0.42, 1.16)
6-46
-------
Table 6-3 (Continued): Epidemiologic studies of short-term exposure to ozone and total (nonaccidental) mortality.
Study
Study Population Exposure Assessment
Mean and
Upper
Percentiles
(PPb)
Copollutant
Examination
Effect Estimates
Percentage Increase (95% CI)
Katsouvanni et al. (2009) NMMAPS
APHENA 12 Canadi;
1987-1996 All ages
Time-series study
Exposure assignment
Mean:
s approach detailed in
U.S.:
original studies and based
13.0-38.0
on all available monitoring
Canada:
data
6.7-8.4
1-h max
75th:
U.S.:
21.0-52.0
Canada:
8.7-12.5
Correlation (r): NR
Copollutant models:
PM10
All-year (lag 0-2 DL)
U.S.: 1.88 (0.69, 3.03)
Canada: 3.63 (1.13, 6.02)
Warm/summer (lag 0-2 DL)
U.S.: 2.38 (1.18, 3.58)
Canada: 2.08 (0.79, 3.33)
Franklin and Schwartz
All ages
Average of all monitors in
Mean:
Correlation (r): PM2.5:
Warm/summer (lag 0)
(2008)
each county based on the
21.4-48.7
0.43; SO42": 0.34; OC:
1.34 (0.68, 2.00)
18 U.S. communities
method detailed in
0.50; NOs": 0.24
Time-series study
Schwartz (2000)
Copollutant models:
24-h avg
PM2.5, SO42", OC,
NO3-
Zanobetti and Schwartz
All ages
Average of all monitors in
Mean (across
Correlation (r): NR
Warm/summer (lag 0)
(2008a)
each city
seasons):
Copollutant models:
1.00 (0.76, 1.24)
48 U.S. cities
8-h avg
16.5-47.8
NR
Case-crossover study
Maximum
(across
seasons):
40.6-103.0
Zanobetti and Schwartz
All ages
Average of all monitors in
Mean:
Correlation (r): NR
Warm/summer (lag 0-3)
(2008b)
each city
15.1-62.8
Copollutant models:
1.06 (0.56, 1.55)
48 U.S. Cities
8-h avg
75th:
NR
Time-series study
19.8-75.4
Maximum:
34.3-146.2
6-47
-------
Table 6-3 (Continued): Epidemiologic studies of short-term exposure to ozone and total (nonaccidental) mortality.
Study
Study Population Exposure Assessment
Mean and
Upper
Percentiles
(PPb)
Copollutant
Examination
Effect Estimates
Percentage Increase (95% CI)
Medina-Ramon and
Schwartz (2008)
48 U.S. cities
Case-only study
All ages
Average of all monitors in Median:
each county based on the 16.1-58.8
method detailed in
Schwartz (2000)
8-h avg
Correlation (r): NR
Copollutant models:
NR
Warm/summer (lag 0-2)
1.30 (0.76, 1.87)
tKlemm et al. (2011)
Atlanta, GA, U.S.
8/1998-12/2007
Time-series study
65+
Data from several
monitors.
8-h max
Mean: 35.54
75th: 47.82
Maximum: NR
109.07
Correlation (r): NR
Copollutant models:
Lag 0-1: 1.36 (-0.50, 3.25)
tMoolaavkar et al. (2013) NMMAPS
98 U.S. cities All ages
1987-2000
Time-series study
Average across all
monitors in each city; 10%
trimmed mean to correct
for yearly averages of
each monitor
24-h avg
Mean: NR
Correlation (r): NR
Copollutant models:
PM10
100 df temporal trends (lag 1): 0.60 (0.44,
0.80)
100 df temporal trends with PM10 (lag 1):
0.33 (-0.07, 0.72)
6-48
-------
Table 6-3 (Continued): Epidemiologic studies of short-term exposure to ozone and total (nonaccidental) mortality.
Study
Study Population Exposure Assessment
Mean and
Upper
Percentiles
(PPb)
Copollutant
Examination
Effect Estimates
Percentage Increase (95% CI)
tGoldbera et al. (2013)
Montreal, Canada
1990-2003
Time-series study
65+
Daily average of each
monitor, then average
across all monitors
24-h avg
Mean: 16.47
Median: 15.2
75th: 21.4
Maximum:
69.65
Correlation (r):
PM2.5: -0.02;
NO2: -0.23;
SO2: -0.31;
Copollutant models:
NR
All-year (0-2 DLNM): -0.37 (-2.30, 1.60)
Warm (April-September; 0-2 DLNM): 1.35
(-1.10, 3.87)
All-year with CHF (0-2 DLNM): 0.39 (-3.23,
4.15)
Warm (April-September) with CHF (0-2
DLNM): 2.99 (-1.95, 8.17)
All-year with hypertension (0-2 DLNM): 0.93
(-2.54, 4.53)
Warm (April-September) with hypertension
(0-2 DLNM): 3.70 (-0.08, 7.63)
All-year with cancer (0-2 DLNM): 1.34
(-1.60, 4.35)
Warm (April-September) with cancer
(0-2 DLNM): 3.57 (0.16, 7.10)
All-year with acute CAD (0-2 DLNM): 2.55
(-1.90, 7.19)
Warm (April-September) with acute CAD
(0-2 DLNM): 7.78 (2.43, 13.41)
All-year with cerebrovascular disease
(0-2 DLNM): 3.77 (-0.93, 8.70)
Warm (April-September) with
cerebrovascular disease (0-2 DLNM): 4.93
(-0.04, 10.16)
tPena et al. (2013)
APHENA
Average of all monitors in
Mean: NR
Correlation (r): NR
50 U.S. cities
50 U.S. cities
All ages
each city
Median:
Copollutant models:
All-year (lag 0-2 DL): 2.13 (0.54, 3.73)
12 Canadian cities
1-h max
6.6-19.4
PM10
All-year with PM10 (lag 1): 0.64 (-0.88, 2.18)
Summer (lag 0-2 DL): 3.23 (1.63, 4.85)
1987-1996
12 Canadian cities
Time-series study
All-year (lag 0-2 DL): 3.73 (1.23, 6.54)
All-year with PM10 (lag 1): 2.38 (-0.88, 6.02)
Summer (lag 0-2 DL): 2.08 (0.79, 3.33)
6-49
-------
Table 6-3 (Continued): Epidemiologic studies of short-term exposure to ozone and total (nonaccidental) mortality.
Study
Study Population
Exposure Assessment
Mean and
Upper
Percentiles
(PPb)
Copollutant
Examination
Effect Estimates
Percentage Increase (95% CI)
tVanos etal. (2014)
10 Canadian cities
1981-1999
Time-series study
All ages
Monitor in each city
located downtown or at
city airports within 27 km
of downtown
24-h avg
Mean: 19.3
Correlation (r): NR
Copollutant models:
NR
All-year (lag 0): NA (NA, NA)
Winter (lag 0): 2.70 (-0.04, 5.43)
Spring (lag 0): 3.54 (-0.84, 7.90)
Summer (lag 0): 3.07 (1.61, 4.53)
Fall (lag 0): 1.40 (0.30, 2.49)
tVanos etal. (2013)
10 Canadian cities
1981-1999
Time-series study
All ages
Air pollution data from
NAPS network
24-h avg
Mean:
14.5-23.2
Correlation (r): NR
Copollutant models:
NR
Overall (lag 0): 2.25 (1.17, 3.33)
DM (lag 0): 2.02 (1.48, 2.56)
DP (lag 0): 1.32 (0.70, 1.94)
DT (lag 0): 4.26 (2.02, 6.56)
MM (lag 0): 1.55 (0.86, 2.25)
MP (lag 0): 1.94 (0.93, 2.95)
MT (lag 0): 3.02 (1.48, 4.64)
Transition (lag 0): 1.40 (0.39, 2.33)
tJhun etal. (2014)
97 U.S. cities
1987-2000
Time-series study
NMMAPS
All ages
Average across all
monitors in each city; 10%
trimmed mean to correct
for yearly averages of
each monitor
24-h avg
Mean: NR
Correlation (r): NR
Copollutant models:
NR
Warm season (May-September): linear
temp term (lag 0): 0.71 (0.29, 1.14)
Warm season (May-September):
Categorical temp term (lag 0): 0.81 (0.38,
1.25)
Temperature range
Warm season (May-September): low temp
days (25th percentile; lag 0): 1.08 (0.27,
1.90)
Warm season (May-September): moderate
temp days (lag 0): 0.59 (-0.04, 1.22)
Warm season (May-September): high temp
days (75th percentile; lag 0): 0.98 (0.30,
1.64)
Warm season (May-September): high temp
days (90th percentile; lag 0): 1.25 (0.26,
2.23)
Warm season (May-September): high temp
days (95th percentile; lag 0): 2.03 (0.66,
3.42)
6-50
-------
Table 6-3 (Continued): Epidemiologic studies of short-term exposure to ozone and total (nonaccidental) mortality.
Study
Study Population Exposure Assessment
Mean and
Upper
Percentiles
(PPb)
Copollutant
Examination
Effect Estimates
Percentage Increase (95% CI)
tVanos etal. (2015)
12 Canadian cities
1981-2008
Time-series study
All ages
Data from National Air
Pollution Surveillance
Network database
24-h avg
Mean: 23.04 Correlation (r):
PM2.5: 0.38;
NO2: 0.1;
SO2: 0.05;
Copollutant models:
NR
DM (0-6 DLNM): 3.83 (2.75, 4.91)
DT (0-6 DLNM): 7.97 (4.09, 11.99)
MM (0-6 DLNM): 4.52 (3.26, 5.66)
MT (0-6 DLNM): 4.44 (2.80, 6.19)
MT+ (0-6 DLNM): 7.84 (2.88, 13.07)
tLiu etal. (2016)
20 U.S. cities
(10 northern;
10 southern);
1987-2000
Time-series study
NMMAPS Average across all Mean: 39.7
All ages monitors in each city; 10% 75th: 412
trimmed mean to correct
r_ „ 1 IVI3X mum.
for yearly averages of
each monitor
8-h max
44.7
Correlation (r): NR
Copollutant models:
NR
Southern communities
Spring (lag 0-2): -0.20 (-1.00, 0.80)
Summer (lag 0-2): -0.40 (-1.00, 0.20)
Autumn (lag 0-2): 0.60 (-0.60, 2.01)
Winter (lag 0-2): 0.60 (-0.60, 1.61)
Northern communities
Spring (lag 0-2): 1.40 (0.60, 2.41)
Summer (lag 0-2): 2.41 (1.40, 3.43)
Autumn (lag 0-2): 1.00 (0.20, 2.01)
Winter (lag 0-2): -1.99 (-3.17, -1.00)
tDietal. (2017a)
39,182 zip-codes, U.S.
2000-2012
Case-crossover study
Medicare cohort
n = 22,433,862
65+
Validated prediction
models based on land
use, chemical transport
modeling, and satellite
remote sensing data.
8-h max
Mean: 37.8
Correlation (r): NR
Copollutant models:
PM2.5
Main analysis with PM2.5 (lag 0-1): 1.02
(0.82, 1.22)
Nearest monitor with PM2.5 from 1-km grid
model (lag 0-1): 0.70 (0.56, 0.82)
Single pollutant (lag 0-1): 1.10 (0.96, 1.24)
Limited to days where O3 <60 ppb W/PM2.5
(lag 0-1): 1.16 (0.92, 1.40)
6-51
-------
Table 6-3 (Continued): Epidemiologic studies of short-term exposure to ozone and total (nonaccidental) mortality.
Mean and
Upper
Percentiles
Copollutant
Effect Estimates
Study
Study Population
Exposure Assessment
(PPb)
Examination
Percentage Increase (95% CI)
tButeau et al. (2018)
n = 63,534
Nearest monitoring station
Mean: 28.9
Correlation (r): NR
Case-crossover
Montreal, Canada
65+ with CHF
8-h avg
Median: 27.3
Copollutant models:
Nearest monitoring station (0-3 DLNM):
1991-2002
75th: 37.5
NR
-2.24 (-9.38, 5.31)
Case-crossover study
95th: 57.5
BME (0-3 DLNM): -5.12 (-16.61, 7.88)
Case-control study
Maximum:
108.8
BME (0-3 DLNM): 1.38 (-12.25, 16.94)
Inverse-distance weighting (0-3 DLNM):
2.90 (-5.87, 12.54)
Case-control
Inverse-distance weighting (0-3 DLNM):
22.69 (-3.12, 55.30)
Back extrapolation from LUR (0-3 DLNM):
4.28 (-5.46, 14.95)
Nearest monitoring station (0-3 DLNM):
6.84 (0.31, 13.79)
Back extrapolation from LUR (0-3 DLNM):
8.97 (3.67, 14.70)
April-October (lag 0)
Additive linear
National: 3.06 (SE = 0.30)
Additive nonlinear
National: 3.54 (SE = 0.75)
Surface
National: 3.98 (SE = 0.24)
twilson etal. (2014)
95 U.S. cities
1987-2000
Time-series study
NMMAPS
All ages
Average across all
monitors in each city; 10%
trimmed mean to correct
for yearly averages of
each monitor
1-h max
NR
Correlation (r): NR
Copollutant models:
NR
6-52
-------
Table 6-3 (Continued): Epidemiologic studies of short-term exposure to ozone and total (nonaccidental) mortality.
Study
Study Population Exposure Assessment
Mean and
Upper
Percentiles
(PPb)
Copollutant
Examination
Effect Estimates
Percentage Increase (95% CI)
tMadriaano et al. (2015) New York:
91 northeast U.S. counties
1988-1999
Time-series study
62 counties
New Jersey:
21 counties
Connecticut: eight
counties
All ages
Analysis of the average
Mean:
across all monitors in
12 counties
each county for
(observed
12 counties; kriging for all
data): 45.6
91 counties analysis
91 counties
8-h max
(kriging data):
45.7
23 urban
counties:
45.6
68 nonurban
counties:
45.7
Maximum:
133.5-149
Correlation (r): NR
Copollutant models:
PMiob
April-October (lag 0)
91 U.S. counties (kriging data):
1.10 (0.50, 1.73)
23 U.S. urban counties (kriging data):
0.90 (0.16, 1.67)
68 U.S. nonurban counties (kriging data):
1.47 (0.38, 2.54)
12 U.S. counties (observed data):
1.61 (0.62, 2.62)
New Haven, CT (without PM10):
5.14 (1.57, 8.85)
New Haven, CT (with PM10):
5.04 (1.38, 8.81)
tChen et al. (2018)
86 U.S. cities
1987-2000
Time-series study
NMMAPS
All ages
Average across all
monitors in each city; 10%
trimmed mean to correct
for yearly averages of
each monitor
24-h avg
NR
Correlation (r): NR
Copollutant models:
PM10 and NO2
Temperature stratification (sTemp:DLNM;
lag 0-1)
Low temperature (<25th percentile)
0.17 (-0.46, 0.81)
Medium temperature (25th-75th percentile)
0.26 (-0.10, 0.62)
High temperature (>75th percentile)
0.89 (0.48, 1.28)
APHENA = Air Pollution and Health: A European and North American Approach; CHF = Congestive Heart Failure; CT = Connecticut; NMMAPS = National Morbidity, Mortality, and
Air Pollution Study; DL = distributed lag; DLNM = distributed lag nonlinear model; LUR = land use regression; NJ = New Jersey; NR = not reported; SE = standard error;
sTemp = smooth temperature term.
t = U.S. and Canadian studies published since the 2013 Ozone ISA.
aStudy examined all-cause mortality (including accidental deaths).
bCopollutant analysis with PM10 only conducted in New Haven, CT due to it being the only city that collected PM10 data during the study period.
6-53
-------
Table 6-4 Epidemiologic studies of short-term exposure to ozone and cardiovascular mortality.
Study
Study Population
Exposure Assessment
Mean and
Upper
Percentiles
PPb
Copollutant
Examination
Effect Estimates
Percentage Increase (95% CI)
Bell et al. (2005)
U.S. and non-U.S.
24-h avg
NR
Correlation (r): NR
All-year:
Meta-analysis
Copollutant models: NR
1.67 (1.02, 2.30)
Katsouvanni et al. (2009)
APHENA
1987-1996
Time-series study
NMMAPS
12 Canadian cities
All ages
Exposure assignment
Mean:
approach detailed in
U.S.:
original studies and based
13.-38.4
on all available monitoring
Canada:
data
6.7-8.4
1-h max
75th:
U.S.:
21.0-52.0
Canada:
8.7-12.5
Correlation (r): NR
Copollutant models:
PM10
All-year (lag 0-2; 8 df/yr-NS):
>75 yr
U.S.: 1.43 (-0.83, 3.73)
Canada: 5.51 (0.47, 11.26)
<75 yr
U.S.: 2.38 (-0.10, 4.90)
Canada: 4.34 (-1.70, 10.72)
Summer (lag 0-2; 8 df/yr-NS):
>75 yr
U.S.: 1.98 (-0.29, 4.29)
Canada: 0.93 (-1.75, 3.68)
<75 yr
U.S.: 4.19 (1.68, 6.74)
Canada: -0.64 (-2.67, 1.43)
Zanobetti and Schwartz All ages
(2008b)
48 U.S. cities
Time-series study
Average of all monitors in
each city
8-h avg
Mean:
15.1-62.8
75th:
19.8-75.4
Maximum:
34.3-146.2
Correlation (r): NR
Copollutant models: NR
Summer (lag 0-3)
1.61 (0.96, 2.27)
tKlemm et al. (2011)
Atlanta, GA, U.S.
8/1998-12/2007
Time-series study
65+
Data from several
monitors.
8-h max
Mean: 35.54 Correlation (r): NR Lag 0-1: 0.69 (-2.28, 3.75)
75th: 47.82 Copollutant models: NR
Maximum:
109.07
6-54
-------
Table 6-4 (Continued): Epidemiologic studies of short-term exposure to ozone and cardiovascular mortality.
Study
Study Population Exposure Assessment
Mean and
Upper
Percentiles
PPb
Copollutant
Examination
Effect Estimates
Percentage Increase (95% CI)
tSacks etal. (2012)
Philadelphia, PA, U.S.
5/12/1992-9/30/1995
Time-series study
All ages
Single monitor ~6 km Mean: 36
west/southwest of city hall Median 33
8-h max Maximum:
110
Correlation (r):
PM25: 0.43;
NO2: 0.18;
SO2: -0.19;
CO: -0.35
Copollutant models: NR
Harvard (lag 0-1): -1.60 (-5.10, 2.10)
California (lag 0-1): 0.20 (-3.40, 3.90)
Canada (lag 0-1): 0.50 (-3.10, 4.30)
Harvard AT (lag 0-1): 1.30 (-2.10, 4.90)
APHEA2 (lag 0-1): 1.70 (-1.80, 5.30)
NMMAPS (lag 0-1): 2.20 (-1.80, 6.40)
tVanos etal. (2014)
10 Canadian cities
1981-1999
Time-series study
All ages
Monitor in each city
located downtown or at
city airports within 27 km
of downtown
24-h avg
Mean: 19.3
Correlation (r): NR
Copollutant models: NR
All-year (lag 0): 4.65 (1.86, 7.43)
Spring (lag 0): 3.16 (0.25, 6.08)
Summer (lag 0): 5.58 (1.94, 9.21)
Fall (lag 0): 1.96 (0.13, 3.78)
Winter (lag 0): 4.46 (1.55, 7.37)
df = degrees of freedom; NS = natural spline,
t = U.S. and Canadian studies published since the 2013 Ozone ISA.
6-55
-------
Table 6-5 Epidemiologic studies of short-term exposure to ozone and respiratory mortality.
Study
Study Population Exposure Assessment
Mean and
Upper
Percentiles
(PPb)
Copollutant
Examination
Effect Estimates
Percentage Increase (95% CI)
Bell et al. (2005)
Meta-analysis
Meta-analysis
24-h avg
NR
Correlation (r): NR
Copollutant models: NR
All-year:0.70 (-0.77, 2.21)
Katsouvanni et al. (2009) NMMAPS
APHENA 12Canadi;
1987-1996 All ages
Time-series study
Exposure assignment
Mean:
s approach detailed in
U.S.:
original studies and based
13.-38.4
on all available monitoring
Canada:
data
6.7-8.4
1-h max
75th:
U.S.:
21.0-52.0
Canada:
8.7-12.5
Correlation (r): NR
Copollutant models:
PM10
All-year (lag 0-2; 8 df/yr-NS):
All
U.S.: 1.58 (-2.09, 5.41)
Canada: 0.64 (-7.60, 9.67)
>75 yr
U.S.: 0.69 (-4.10, 5.66)
Canada: -2.91 (-12.56, 8.09)
Summer (lag 0-2; 8 df/yr-NS):
All
U.S.: 2.73 (-1.32, 6.90)
Canada: 15.59 (8.09, 24.08)
>75 yr
U.S.: 2.52 (-2.67, 7.94)
Canada: 11.79 (1.38, 23.50)
Zanobetti and Schwartz All ages
(2008b)
48 U.S. cities
Time-series study
Average of all monitors in
each city
8-h avg
Mean:
15.1-62.8
75th:
19.8-75.4
Maximum:
34.3-146.2
Correlation (r): NR
Copollutant models: NR
Summer (lag 0-3):
1.67 (0.76, 2.58)
tKlemm et al. (2011)
Atlanta, GA, U.S.
8/1998-12/2007
Time-series study
65+
Data from several
monitors.
8-h max
Mean: 35.54 Correlation (r): NR Lag 0-1:-0.44 (-6.06, 5.51)
75th: 47.82 Copollutant models: NR
Maximum:
109.07
6-56
-------
Table 6-5 (Continued): Epidemiologic studies of short-term exposure to ozone and respiratory mortality.
Study
Study Population Exposure Assessment
Mean and
Upper
Percentiles
(PPb)
Copollutant
Examination
Effect Estimates
Percentage Increase (95% CI)
tVanos etal. (2014)
10 Canadian cities,
Canada
1981-1999
Time-series study
All ages
Monitor in each city
located downtown or at
city airports within 27 km
of downtown
24-h avg
Mean: 19.3
Correlation (r): NR
Copollutant models: NR
Fall (lag 0): 6.04 (3.31, 8.77)
Winter (lag 0): 6.23 (1.49, 10.95)
Summer (lag 0): 7.71 (4.26, 11.16)
All-year (lag 0): 8.36 (3.72, 12.98)
Spring (lag 0): 8.64 (2.45, 14.80)
df = degrees of freedom; NS = natural spline,
t = U.S. and Canadian studies published since the 2013 Ozone ISA.
6-57
-------
6.3.2
Long-Term Ozone Exposure and Mortality: Data Tables
Table 6-6 Epidemiologic studies of long-term exposure to ozone and total (nonaccidental) mortality.
Study
Study Population Exposure Assessment Mean (ppb)
Copollutant
Examination
Effect Estimates
Hazard Ratio (95% CI)
Jerrett et al. (2009)
ACS
Daily maximum of AIRS
Median: 57.4
Correlation (r):
Total mortality (96 MSAs): 1.00 (1.00, 1.01)
Nationwide, U.S.
n = 448,850
monitors averaged over
PM2.5 0.64
Total mortality (86 MSAs): 1.00 (1.00, 1.01)
Ozone: 1977-2000
30+ yr
each quarter; second
and third quarters
Copollutant models:
PM2.5
Total mortality (86 MSAs + PM2.5): 0.99 (0.98, 1.00)
Follow-up: 1982-2000
(April-September)
Cohort study
averaged together for
each year
1-h max
Zanobetti and Schwartz
Medicare
City wide average from
Mean:
Correlation (r): NR
Total mortality (pre-existing heart failure): 1.12
(2011)
n = 3,210,511
AQS
15.6-71.4
Copollutant models:
(1.06, 1.17)
105 cities, U.S.
65+ yr with
8-h avg
NR
Total mortality (pre-existing COPD): 1.14 (1.08,
Ozone: NR
pre-existing
1.19)
Follow-up: 1985-2006
disease
Total mortality (pre-existing diabetes): 1.14 (1.10,
1.21)
Cohort study
Total mortality (pre-existing Ml): 1.19(1.12, 1.25)
tSpencer-Hwanq et al.
n = 32,239
Monthly average of
NR
Correlation (r): NR
Total mortality: 1.1 (0.99, 1.21)
(2011)
Kidney transplant
AQS monitors within
Copollutant models:
Total mortality + PM10: 1.09 (0.99, 1.21)
Nationwide, U.S.
recipients
50 km of residence and
PM10
Ozone: 1997-2003
downscaled to zip-code
using IDW
Follow-up: 1997-2003
Cohort study
6-58
-------
Table 6-6 (Continued): Epidemiologic studies of long-term exposure to ozone and total (nonaccidental) mortality.
Study
Study Population Exposure Assessment Mean (ppb)
Copollutant
Examination
Effect Estimates
Hazard Ratio (95% CI)
tCarev et al. (2013)
Nationwide, U.K.
Ozone: 2002
Follow-up: 2003-2007
Cohort study
English Medical
Practice
n = 835,607
Adults, 40-89 yr,
from English
medical practices
Annual mean estimates
from dispersion model
for 1-km grid cells linked
to nearest residential
postal code centroid
Mean: 25.85
Maximum:
31.5
Correlation (r):
PM2.5: -0.39;
NO2: -0.46;
SO2: -0.41;
PM10: -0.40
Copollutant models:
NR
Total mortality: 0.76 (0.62, 0.87)
tJerrett et al. (2013)
California, U.S.
Ozone: 1988-2002
Follow-up: 1982-2000
Cohort study
ACS
n = 73,711
California
Monthly averages
calculated from IDW
from up to four monitors
within 50 km of
residence
Mean: 50.35
Median: 50.8
75th: 61
90th: 68.56
95th: 74.18
Maximum:
89.33
Correlation (r):
PM2.5: 0.56;
NO2: -0.0071;
Copollutant models:
PM2.5; NO2
Total mortality: 1.00 (0.98, 1.01)
Total mortality (+ PM2.5): 0.99 (0.98, 1.01)
Total mortality (+ NO2): 1.00 (0.99, 1.02)
tBentaveb et al. (2015)
Nationwide, France
Ozone: 1989-2008
Follow-up: 1989-2013
Cohort study
Gazel
n = 20,327
Adults working at
French national
electricity and gas
company
CHIMERE chemical
transport model
8-h max
Mean: 40.5
Median: 48
Correlation (r):
PM2.5: -0.38;
NO2: -0.34;
PM10: -0.21
Copollutant models:
NR
Total mortality: 0.81 (0.64, 1.03)
tCrouse et al. (2015)
Nationwide, Canada
Ozone: 2002-2009
Follow-up: 1991-2006
Cohort study
CanCHEC
n = 2,521,525
25+ yr
Model of warm season
concentration at 21-km
horizontal resolution
assigned at postal code
8-h max
Mean: 39.6
Median: 39
75th: 44.2
Maximum:
60
Correlation (r):
PM2.5: 0.73;
NO2: 0.19;
Copollutant models:
NR
Total mortality: 1.03 (1.03, 1.04)
6-59
-------
Table 6-6 (Continued): Epidemiologic studies of long-term exposure to ozone and total (nonaccidental) mortality.
Study
Study Population Exposure Assessment Mean (ppb)
Copollutant
Examination
Effect Estimates
Hazard Ratio (95% CI)
tTurner et al. (2016)
Nationwide, U.S.
Ozone: 2002-2004
Follow-up: 1982-2004
Cohort study
ACS
n = 669,046
35+
HBM with inputs from
NAMS/SLAMS and
CMAQ; downscaler for
eastern U.S.
8-h max
Mean: 38.2
Median: 38.1
75th: 40.1
95th: 45
Maximum:
59.3
Correlation (r):
PM2.5: 0.18;
NO2: -0.08;
Copollutant models:
PM2.5
Total mortality (year-round): 1.02 (1.01, 1.04)
Total mortality (year-round; + PM2.5): 1.02(1.01,
1.04)
tEckel et al. (2016)
n = 352,053
Monthly averages Mean: 40.2
Correlation (r): Total mortality: 1.02 (1.02, 1.03)
California, U.S.
California residents
calculated from IDW
PM2.5: -0.02;
Ozone: 1988-2011
Follow-up: 1988-2011
with newly
diagnosed cancer
from up to four monitors
within 50 km of
residence
NO2: -0.01;
PM10: 0.36
Copollutant models:
Cohort study
8-h max
NR
tDietal. (2017b)
Nationwide, U.S.
Ozone: 2000-2012
Follow-up: 2000-2012
Cohort study
Medicare
n = 61 million
Older adults
Neural network that
includes satellite-based
measurements,
chemical transport
model outputs, land-use
terms, meteorological
data, and observations
from 1,877 ozone
monitoring stations
Mean: 46.3
95th: 55.86
Correlation (r):
PM2.5: 0.24;
Copollutant models:
PM2.5
Total mortality [main analysis i
1.01)
1-PM2.5)]: 1.01 (1.01,
Total mortality (single pollutant): 1.02 (1.02, 1.02)
Total mortality (+PM2.5; low ozone exposure,
<50 ppb): 1.01 (1.01, 1.01)
tWeichenthal et al.
(2017)
Nationwide, Canada
Ozone: 2002-2009
Follow-up: 1991-2011
Cohort study
CanCHEC
n = 2,448,500
25+ yr
Model of warm season
concentration at 21 km
horizontal resolution
assigned at postal code
8-h max
Mean: 38.29
Median:
38.11
75th: 42.63
95th: 50.51
Maximum:
60.46
Correlation (r): NR
Copollutant models:
NR
Total mortality: 1.06 (1.05, 1.07)
6-60
-------
Table 6-6 (Continued): Epidemiologic studies of long-term exposure to ozone and total (nonaccidental) mortality.
Study
Study Population Exposure Assessment Mean (ppb)
Copollutant
Examination
Effect Estimates
Hazard Ratio (95% CI)
tCakmak et al. (2018)
Nationwide, Canada
Ozone: 2002-2009
Follow-up: 1991-2011
Cohort study
CanCHEC
n = 2,291,250
25+ yr
Model of warm season
concentration at 21-km
horizontal resolution
assigned at postal code
8-h max
Mean:
15.0-43.0
Maximum:
46.6-60.6
Correlation (r):
PM2.5: -0.705
Copollutant models:
PM2.5
Total mortality: 1.08 (1.02, 1.14)
Total mortality (+ PM2.5): 1.05 (0.99, 1.11)
tKimetal. (2017)
Seoul, Korea
Ozone: 2007-2013
Follow-up: 2007-2013
Cohort study
NHIS-NSC
n = 136,094
18+ yr, no previous
history of CVD
27 monitors in Seoul
linked to zip-code of
participant residence
Mean: 19.93
Median:
18.75
75th: 27.08
Maximum:
71.12
Correlation (r):
PM2.5: 0.67;
NO2: 0.68;
SO2: 0.84;
CO: 0.55
Copollutant models:
NR
Total mortality: 0.78 (0.75, 0.82)
tSese et al. (2017)
Nationwide, France
Ozone: 2007-2014
Follow-up: 2007-2014
Cohort study
COFI
n = 192
Patients with
idiopathic
pulmonary fibrosis
Measurements from
nearest monitor
24-h avg
NR
Correlation (r): NR Total mortality: 0.79 (0.44, 1.39)
Copollutant models:
NR
6-61
-------
Table 6-7 Epidemiologic studies of long-term exposure to ozone and cardiovascular mortality.
Study
Study Population Exposure Assessment Mean (ppb)
Copollutant
Examination
Effect Estimates
Hazard Ratio (95% CI)
Jerrett et al. (2009)
Nationwide, U.S.
Ozone: 1977-2000
Follow-up: 1982-2000
Cohort study
ACS
n = 448,850
30+ yr
Daily maximum of AIRS
monitors averaged over
each quarter; second
and third quarters
(April-September)
averaged together for
each year
1-h max
Median: 57.4
Correlation (r):
PM2.5 0.64
Copollutant models:
PM2.5
CVD mortality (96 MSAs): 1.01 (1.00, 1.02)
CVD mortality (86 MSAs): 1.01 (1.01, 1.02)
CVD mortality (86 MSAs + PM2.5): 0.98 (0.97,
0.99)
IHD mortality (96 MSAs): 1.02 (1.00, 1.03)
IHD mortality (86 MSAs): 1.02 (1.01, 1.03)
IHD mortality (86 MSAs + PM2.5): 0.97 (0.96, 0.99)
tSpencer-Hwanq et al.
(2011)
Nationwide, U.S.
Ozone: 1997-2003
Follow-up: 1997-2003
Cohort study
n = 32,239
Kidney transplant
recipients
Monthly average of
AQS monitors within
50 km of residence and
downscaled to zip-code
using IDW
NR
Correlation (r): NR CHD mortality: 1.35 (1.04, 1.77)
Copollutant models: CHD mortality (+ PM10): 1.34 (1.03, 1.76)
PM10
tCarev et al. (2013)
Nationwide, U.K.
Ozone: 2002
Follow-up: 2003-2007
Cohort study
English medical
practice
n = 835,607
Adults, 40-89 yr,
from English
medical practices
Annual mean estimates
from dispersion model
for 1-km grid cells linked
to nearest residential
postal code centroid
Mean: 25.85
Maximum:
31.5
Correlation (r):
PM2.5: -0.39;
NO2: -0.46;
SO2: -0.41;
PM10: -0.40
Copollutant models:
NR
Circulatory mortality: 0.76 (0.66, 0.87)
6-62
-------
Table 6-7 (Continued): Epidemiologic studies of long-term exposure to ozone and cardiovascular mortality.
Study
Study Population Exposure Assessment Mean (ppb)
Copollutant
Examination
Effect Estimates
Hazard Ratio (95% CI)
tJerrett et al. (2013)
California, U.S.
Ozone: 1988-2002
Follow-up: 1982-2000
Cohort study
ACS
n = 73,711
California
Monthly averages
calculated from IDW
from up to four monitors
within 50 km of
residence
Mean: 50.35
Median: 50.8
75th: 61
90th: 68.56
95th: 74.18
Maximum:
89.33
Correlation (r):
PM2.5: 0.56;
NO2: -0.0071;
Copollutant models:
PM2.5; NO2
Cardiovascular: 1.02 (0.99, 1.04)
Cardiovascular (+ PM2.5): 1.01 (0.98, 1.04)
Cardiovascular (+ NO2): 1.03 (1.00, 1.05)
IHD: 1.04 (1.01, 1.08)
IHD (+ PM2.5): 1.03 (0.99, 1.06)
IHD (+ NO2): 1.05 (1.02, 1.09)
Stroke: 1.00 (0.97, 1.04)
Stroke (+ PM2.5): 1.00 (0.95, 1.04)
Stroke (+ NO2): 1.01 (0.97, 1.06)
tBentaveb et al. (2015)
Nationwide, France
Ozone: 1989-2008
Follow-up: 1989-2013
Cohort study
Gazel
n = 20,327
Adults working at
French national
electricity and gas
company
CHIMERE chemical
transport model
8-h max
Mean: 40.5
Median: 48
Correlation (r):
PM2.5: -0.38;
NO2: -0.34;
PM10: -0.21
Copollutant models:
NR
CVD mortality: 0.83 (0.39, 1.75)
tCrouse et al. (2015)
Nationwide, Canada
Ozone: 2002-2009
Follow-up: 1991-2006
Cohort study
CanCHEC
n = 2,521,525
25+ yr
Model of warm season
concentration at21-km
horizontal resolution
assigned at postal code
8-h max
Mean: 39.6
Median: 39
75th: 44.2
Maximum: 60
Correlation (r):
PM2.5: 0.73;
NO2: 0.19;
Copollutant models:
NR
CVD: 1.04 (1.03, 1.05)
Cardiometabolic: 1.05 (1.04, 1.06)
IHD: 1.09 (1.08, 1.11)
CBVD: 0.98 (0.96, 1.00)
Diabetes: 1.16 (1.13, 1.20)
tTurner et al. (2016)
Nationwide, U.S.
Ozone: 2002-2004
Follow-up: 1982-2004
Cohort study
ACS
n = 669,046
35+ yr
HBM with inputs from
NAMS/SLAMS and
CMAQ; downscaler for
eastern U.S.
8-h max
Mean: 38.2
Median: 38.1
75th: 40.1
95th: 45
Maximum:
59.3
Correlation (r):
PM2.5: 0.18;
NO2: -0.08;
Copollutant models:
PM2.5
CVD: 1.03 (1.01, 1.05)
IHD: 0.98 (0.95, 1.00)
CBVD: 1.03 (0.98, 1.07)
Circulatory: 1.03 (1.02, 1.05)
Circulatory (+ PM2.5): 1.03 (1.01, 1.05)
Dysrhythmias, HF, cardiac arrest: 1.15 (1.10,
1.20)
Diabetes: 1.16 (1.07, 1.26)
6-63
-------
Table 6-7 (Continued): Epidemiologic studies of long-term exposure to ozone and cardiovascular mortality.
Study
Study Population Exposure Assessment Mean (ppb)
Copollutant
Examination
Effect Estimates
Hazard Ratio (95% CI)
tCakmak et al. (2016)
Nationwide, Canada
Ozone: 2002-2009
Follow-up: 1991-2006
Cohort study
CANCHEC
n = 2,415,505
25+ yr
Model of warm season
concentration at21-km
horizontal resolution
assigned at postal code
8-h max
Mean:
14.3-40.9
Maximum:
20.1-53
Correlation (r):
PM2.5: -0.67;
Copollutant models:
PM2.5
CVD (base model): 1.05 (1.03, 1.06)
CVD (adjustment for climate zone): 1.06 (1.04,
1.07)
CVD (+ PM2.5): 1.03 (1.02, 1.05)
IHD (base model): 1.09 (1.08, 1.11)
IHD (adjustment for climate zone): 1.07 (1.05,
1.09)
CBVD (base model): 1.00 (0.97, 1.02)
CBVD (adjustment for climate zone): 1.04 (1.01,
1.08)
tWeichenthal et al.
(2017)
Nationwide, Canada
Ozone: 2002-2009
Follow-up: 1991-2011
Cohort study
CanCHEC
n = 2,448,500
25+ yr
Model of warm season
concentration at21-km
horizontal resolution
assigned at postal code
8-h max
Mean: 38.29
Median: 38.11
75th: 42.63
95th: 50.51
Maximum:
60.46
Correlation (r): NR
Copollutant models:
NR
CVD mortality: 1.16 (1.14, 1.18)
tCakmak etal. (2018)
Nationwide, Canada
Ozone: 2002-2009
Follow-up: 1991-2011
Cohort study
CanCHEC
n = 2,291,250
25+ yr
Model of warm season
concentration at 21-km
horizontal resolution
assigned at postal code
8-h max
Mean:
15.0-43.0
Maximum:
46.6-60.6
Correlation (r):
PM2.5: -0.705;
Copollutant models:
PM2.5
IHD: 1.13 (1.12, 1.15)
IHD (+ PM2.5): 1.08 (1.07, 1.10)
tKimetal. (2017)
Seoul, South Korea
Ozone: 2007-2013
Follow-up: 2007-2013
Cohort study
NHIS-NSC
n = 136,094
18+ yr, no
previous history of
CVD
27 monitors in Seoul
linked to zip-code of
participant residence
NR
Mean: 19.93
Median: 18.75
75th: 27.08
Maximum:
71.12
Correlation (r):
PM2.5: 0.67;
NO2: 0.68;
SO2: 0.84;
CO: 0.55
Copollutant models:
NR
Cardiovascular mortality: 0.72 (0.64, 0.81)
6-64
-------
Table 6-8 Epidemiologic studies of long-term exposure to ozone and respiratory mortality.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
Hazard Ratio (95% CI)
Jerrett et al. (2009)
Nationwide, U.S.
Ozone: 1977-2000
Follow-up: 1982-2000
Cohort study
ACS
n = 448,850
30+ yr
Daily maximum of
AIRS monitors
averaged over each
quarter; second and
third quarters
(April-September)
averaged together for
each year
1-h max
Median: 57.4 Correlation (r):
PM2.5 0.64
Copollutant models:
PM2.5
Resp mortality (96 MSAs): 1.03 (1.01, 1.05)
Resp mortality (86 MSAs): 1.03 (1.01, 1.05)
Resp mortality (86 MSAs + PM2.5): 1.04 (1.01,
1.07)
tCarev et al. (2013)
Nationwide, U.K.
Ozone: 2002
Follow-up: 2003-2007
Cohort study
English medical
practice
n = 835,607
Adults, 40-89 yr,
from English
medical practices
Annual mean
estimates from
dispersion model for
1-km grid cells linked
to nearest residential
postal code centroid
Mean: 25.85
Maximum:
31.5
Correlation (r):
PM2.5: -0.39;
NO2: -0.46;
SO2: -0.41;
PM10: -0.40
Copollutant models:
NR
Respiratory: 0.62 (0.50, 0.76)
tJerrett et al. (2013)
California, U.S.
Ozone: 1988-2002
Follow-up: 1982-2000
Cohort study
ACS
n = 73,711
California
Monthly averages
calculated from IDW
from up to four
monitors within 50 km
of residence
Mean: 50.35
Median: 50.8
75th: 61
90th: 68.56
95th: 74.18
Maximum:
89.33
Correlation (r):
PM2.5: 0.56;
NO2: -0.0071;
Copollutant models:
PM2.5; NO2
Respiratory: 1.01 (0.96, 1.06)
Respiratory (+ PM2.5): 1.00 (0.95, 1.05)
Respiratory (+ NO2): 1.01 (0.96, 1.06)
6-65
-------
Table 6-8 (Continued): Epidemiologic studies of long-term exposure to ozone and respiratory mortality.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
Hazard Ratio (95% CI)
tBentaveb et al. (2015)
Nationwide, France
Ozone: 1989-2008
Follow-up: 1989-2013
Cohort study
Gazel
n = 20,327
Adults working at
French national
electricity and gas
company
CHIMERE chemical
transport model
8-h max
Mean: 40.5
Median: 48
Correlation (r):
PM2.5: -0.38;
NO2: -0.34;
PM10: -0.21
Copollutant models:
NR
Respiratory mortality: 0.95 (0.55, 1.69)
tCrouse et al. (2015)
CanCHEC
Model of warm season
Mean: 39.6
Correlation (r):
Respiratory: 0.97 (0.95, 0.99)
Nationwide, Canada
n = 2,521,525
concentration at21-km
Median: 39
PM2.5: 0.73;
COPD: 0.97 (0.95, 1.00)
Ozone: 2002-2009
25+ yr
horizontal resolution
75th: 44.2
NO2: 0.19;
assigned at postal
Copollutant models:
NR
Follow-up: 1991-2006
code
Maximum: 60
Cohort study
8-h max
tTurner et al. (2016)
Nationwide, U.S.
Ozone: 2002-2004
Follow-up: 1982-2004
Cohort study
ACS
n = 669,046
35+
HBM with inputs from
NAMS/SLAMS and
CMAQ; downscalerfor
eastern U.S.
8-h max
Mean: 38.2
Median: 38.1
75th: 40.1
95th: 45
Maximum:
59.3
Correlation (r):
PM2.5: 0.18;
NO2: -0.08;
Copollutant models:
PM2.5
Respiratory: 1.14(1.10, 1.18)
Respiratory (+ PM2.5): 1.12(1.08, 1.16)
COPD: 1.14 (1.08, 1.21)
Pneumonia and flu: 1.10 (1.03, 1.18)
tWeichenthal et al.
(2017)
Nationwide, Canada
Ozone: 2002-2009
Follow-up: 1991-2011
Cohort study
CanCHEC
n = 2,448,500
25+ yr
Model of warm season
concentration at21-km
horizontal resolution
assigned at postal
code
8-h max
Mean: 38.29
Median: 38.11
75th: 42.63
95th: 50.51
Maximum:
60.46
Correlation (r)\ NR
Copollutant models:
NR
Respiratory mortality: 1.04 (1.01, 1.07)
6-66
-------
Table 6-9 Epidemiologic studies of long-term exposure to ozone and other mortality.
Study
Study Population
Exposure
Assessment
Averaging Time
Mean (ppb)
Copollutant
Examination
Effect Estimates
HR (95% CI)
tXu et al. (2013)
Los Angeles, CA and
Honolulu, HI, U.S.
Ozone: 1992-2008
Follow-up: 1992-2008
Cross-sectional study
White respiratory
cancer patients
County-level monthly
means from U.S. EPA
AQS monitors
Mean: NR
Correlation (r): NR
Copollutant models:
NR
Cancer-specific death: 1.06 (1.04, 1.07)
Other cause of death: 1.04 (1.03, 1.06)
tJerrett et al. (2013)
California, U.S.
Ozone: 1988-2002
Follow-up: 1982-2000
Cohort study
ACS
n = 73,711
California
Monthly averages
calculated from IDW
from up to four
monitors within 50 km
of residence
Mean: 50.35
Median: 50.8
75th
90th
95th
61
68.56
74.18
Correlation (r):
PM2.5: 0.56;
NO2: -0.0071;
Copollutant models:
PM2.5; NO2
Other mortality: 0.99 (0.96, 1.01)
Other mortality (+ NO2): 0.99 (0.96, 1.01)
Other mortality (+ PM2.5): 0.99 (0.96, 1.01)
Maximum:
89.33
+Li et al. (2016)
48 states, U.S.
Ozone: 2002-2008
Follow-up: 2002-2008
Cross-sectional study
n = 3,109 counties
in CONUS
County-level rates
Rates of change of
county-level ozone
concentrations from
downscaler CMAQ
model
8-h max
Mean:
29.3-64.5
Correlation (r): NR
Copollutant models:
PM2.5
Reduction in life expectancy (males): -0.42
(-0.50, -0.34)
Reduction in life expectancy (males): -0.50
(-0.60, -0.38)
tRush et al. (2017)
STROBE
County-level average
NR
Correlation (r): NR
In-hospital mortality (continuous exposure model):
30 states, U.S.
n = 93,950
Copollutant models:
1.07 (1.06, 1.08)
Ozone: NR
Patients in hospital
NR
In-hospital mortality (15 high ozone cities): 1.11
Follow-up: 2011
for ARDS
(1.08, 1.15)
Cohort study
6-67
-------
Table 6-9 (Continued): Epidemiologic studies of long-term exposure to ozone and other mortality.
Study
Study Population
Exposure
Assessment
Averaging Time
Mean (ppb)
Copollutant
Examination
Effect Estimates
HR (95% CI)
tVieira et al. (2017)
California, U.S.
Ozone: 2009-2011
Follow-up: 1996-2006
Cohort study
n = 11,765
Women with ovarian
cancer
Daily exceedances
over 70 ppb averaged
over 3 yr to create
value for census tract
8-h max
Median: 3-29 Correlation (r): NR NA
95th: Copollutant models:
265-727
NR
6-68
-------
Annex for Appendix 6: Evaluation of Studies on Health Effects of
Ozone
This annex describes the approach used in the Integrated Science Assessment (ISA) for Ozone
and Related Photochemical Oxidants to evaluate study quality in the available health effects literature. As
described in the Preamble to the ISA (U.S. EPA. 2015). causality determinations were informed by the
integration of evidence across scientific disciplines (e.g., exposure, animal toxicology, epidemiology) and
related outcomes and by judgments of the strength of inference in individual studies. Table Annex 6-1
describes aspects considered in evaluating study quality of controlled human exposure, animal
toxicological, and epidemiologic studies. The aspects found in Table Annex 6-1 are consistent with
current best practices for reporting or evaluating health science data.1 Additionally, the aspects are
compatible with published U.S. EPA guidelines related to cancer, neurotoxicity, reproductive toxicity,
and developmental toxicity (U.S. EPA. 2005. 1998. 1996b. 1991V
These aspects were not used as a checklist, and judgments were made without considering the
results of a study. The presence or absence of particular features in a study did not necessarily lead to the
conclusion that a study was less informative or should be excluded from consideration in the ISA.
Further, these aspects were not used as criteria for determining causality in the five-level hierarchy. As
described in the Preamble, causality determinations were based on judgments of the overall strengths and
limitations of the collective body of available studies and the coherence of evidence across scientific
disciplines and related outcomes. Table Annex 6-1 is not intended to be a complete list of aspects that
define a study's ability to inform the relationship between ozone and health effects, but it describes the
major aspects considered in this ISA to evaluate studies. Where possible, study elements, such as
exposure assessment and confounding (i.e., bias due to a relationship with the outcome and correlation
with exposures to ozone), are considered specifically for ozone. Thus, judgments on the ability of a study
to inform the relationship between an air pollutant and health can vary depending on the specific pollutant
being assessed.
1 For example, NTP OHAT approach (Roonev et al.. 20141. IRIS Preamble (U.S. EPA. 2013b). ToxRTool
(Klimisch et al.. 19971. STROBE guidelines (von Elm et al.. 20071. and ARRIVE guidelines (Kilkenny et al.. 20101.
6-69
-------
Table Annex 6-1 Scientific considerations for evaluating the strength of inference
from studies on the health effects of ozone.
Study Design
Controlled Human Exposure:
Studies should describe clearly the primary and any secondary objectives of the study, or specific hypotheses being
tested. Study subjects should be randomly assigned to treatment groups and exposed, to the extent possible,
without knowledge of the exposure condition. Preference is given to balanced crossover (repeated measures) or
parallel design studies which include control exposures (e.g., to clean filtered air). In crossover studies, a sufficient
and specified time between exposure days should be provided to avoid carry over effects from prior exposure days.
In parallel design studies, all arms should be matched for individual characteristics, such as age, sex, race,
anthropometric properties, and health status. In studies evaluating effects of disease, appropriately matched healthy
controls are desired for interpretative purposes.
Animal Toxicology:
Studies should describe clearly the primary and any secondary objectives of the study, or specific hypotheses being
tested. Studies should include appropriately matched control exposures (e.g., to clean filtered air, time matched).
Studies should use methods to limit differences in baseline characteristics of control and exposure groups. Studies
should randomize assignment to exposure groups and where possible conceal allocation to research personnel.
Groups should be subjected to identical experimental procedures and conditions to the extent possible; animal care
including housing, husbandry, etc. should be identical between groups. Blinding of research personnel to study
group may not be possible due to animal welfare and experimental considerations; however, differences in the
monitoring or handling of animals in all groups by research personnel should be minimized.
Epidemiology:
Inference is stronger for studies that describe clearly the primary and any secondary aims of the study, or specific
hypotheses being tested.
For short-term exposure, time-series, case-crossover, and panel studies are emphasized over cross-sectional
studies because they examine temporal correlations and are less prone to confounding by factors that differ
between individuals (e.g., SES, age). Panel studies with scripted exposures, in particular, can contribute to
inference because they have consistent, well-defined exposure durations across subjects, measure personal
ambient pollutant exposures, and measure outcomes at consistent, well-defined lags after exposures. Studies with
large sample sizes and conducted over multiple years are considered to produce more reliable results. Additionally,
multicity studies are preferred over single-city studies because they examine associations for large diverse
geographic areas using a consistent statistical methodology, avoiding the publication bias often associated with
single-city studies.3 If other quality parameters are equal, multicity studies carry more weight than single-city studies
because they tend to have larger sample sizes and lower potential for publication bias.
For long-term exposure, inference is considered to be stronger for prospective cohort studies and case-control
studies nested within a cohort (e.g., for rare diseases) than cross-sectional, other case-control, or ecologic studies.
Cohort studies can better inform the temporality of exposure and effect. Other designs can have uncertainty related
to the appropriateness of the control group or validity of inference about individuals from group-level data. Study
design limitations can bias health effect associations in either direction.
6-70
-------
Table Annex 6-1 (Continued): Scientific considerations for evaluating the strength
of inference from studies on the health effects of
ozone.
Study Population/Test Model
Controlled Human Exposure:
In general, the subjects recruited into study groups should be similarly matched for age, sex, race, anthropometric
properties, and health status. In studies evaluating effects of specific subject characteristics (e.g., disease, genetic
polymorphism, etc.), appropriately matched healthy controls are preferred. Relevant characteristics and health
status should be reported for each experimental group. Criteria for including and excluding subjects should be
indicated clearly. For the examination of populations with an underlying health condition (e.g., asthma),
independent, clinical assessment of the health condition is ideal, but self-report of physician diagnosis generally is
considered to be reliable for respiratory and cardiovascular disease outcomes.15 The loss or withdrawal of recruited
subjects during the course of a study should be reported. Specific rationale for excluding subject(s) from any portion
of a protocol should be explained.
Animal Toxicology:
Ideally, studies should report species, strain, substrain, genetic background, age, sex, and weight. Unless data
indicate otherwise, all animal species, stocks and strains are considered appropriate for evaluating effects of ozone
exposure. It is preferred that the authors test for effects in both sexes and multiple lifestages and report the result
for each group separately. All animals used in a study should be accounted for, and rationale for exclusion of
animals or data should be specified.
Epidemiology:
There is greater confidence in results for study populations that are recruited from and representative of the target
population. Studies with high participation and low dropout over time that is not dependent on exposure or health
status are considered to have low potential for selection bias. Clearly specified criteria for including and excluding
subjects can aid assessment of selection bias. For populations with an underlying health condition, independent,
clinical assessment of the health condition is valuable, but self-report of physician diagnosis generally is considered
to be reliable for respiratory and cardiovascular diseases.15 Comparisons of groups with and without an underlying
health condition are more informative if groups are from the same source population. Selection bias can influence
results in either direction or may not affect the validity of results but rather reduce the generalizability of findings to
the target population.
Pollutant
Controlled Human Exposure:
The focus is on studies testing ozone exposure.
Animal Toxicology:
The focus is on studies testing ozone exposure.
Epidemiology:
The focus is on studies evaluating ozone exposure.
6-71
-------
Table Annex 6-1 (Continued): Scientific considerations for evaluating the strength
of inference from studies on the health effects of
ozone.
Exposure Assessment or Assignment
Controlled Human Exposure:
For this assessment, the focus is on studies that use ozone concentrations <0.4 ppm. Studies that use higher
exposure concentrations may provide information relevant to biological plausibility, dosimetry, or inter-species
variation. Studies should have well-characterized pollutant concentration, temperature, and relative humidity and/or
have measures in place to adequately control the exposure conditions. Preference is given to balanced crossover or
parallel design studies that include control exposures (e.g., to clean filtered air). Study subjects should be randomly
exposed without knowledge of the exposure condition. Method of exposure (e.g., chamber, facemask, etc.) should
be specified and activity level of subjects during exposures should be well characterized.
Animal Toxicology:
For this assessment, the focus is on studies that use ozone concentrations <2 ppm. Studies that use higher
exposure concentrations may provide information relevant to biological plausibility, dosimetry, or inter-species
variation. Studies should characterize pollutant concentration, temperature, and relative humidity and/or have
measures in place to adequately control the exposure conditions. The focus is on inhalation exposure.
Noninhalation exposure experiments (i.e., intra-tracheal instillation [IT]) are informative and may provide information
relevant to biological plausibility and dosimetry. In vitro studies may be included if they provide mechanistic insight
or examine similar effects as in vivo studies but are generally not included. All studies should include exposure
control groups (e.g., clean filtered air).
Epidemiology:
Of primary relevance are relationships of health effects with the ambient component of ozone exposure. However,
information about ambient exposure rarely is available for individual subjects; most often, inference is based on
ambient concentrations. Studies that compare exposure assessment methods are considered to be particularly
informative. Inference is stronger when the duration or lag of the exposure metric corresponds with the time course
for physiological changes in the outcome (e.g., up to a few days for symptoms) or latency of disease (e.g., several
years for cancer).
Ambient ozone concentration tends to have low spatial heterogeneity at the urban scale, except near roads where
ozone concentration is lower because ozone reacts with nitric oxide emitted from vehicles. For studies involving
individuals with near-road or on-road exposures to ozone, in which ambient ozone concentrations are more spatially
heterogeneous and relationships between personal exposures and ambient concentrations are potentially more
variable, validated methods that capture the extent of variability for the epidemiologic study design (temporal vs.
spatial contrasts) and location carry greater weight.
Fixed-site measurements, whether averaged across multiple monitors or assigned from the nearest or single
available monitor, typically have smaller biases and smaller reductions in precision compared with spatially
heterogeneous air pollutants. Concentrations reported from fixed-site measurements can be informative if correlated
with personal exposures, closely located to study subjects, highly correlated across monitors within a location, or
combined with time-activity information.
Atmospheric models may be used for exposure assessment in place of or to supplement ozone measurements in
epidemiologic analyses. For example, grid-scale models (e.g., CMAQ) that represent ozone exposure over relatively
large spatial scales (e.g., typically greater than 4- * 4-km grid size) often do provide adequate spatial resolution to
capture acute ozone peaks that influence short-term health outcomes. Uncertainty in exposure predictions from
these models is largely influenced by model formulations and the quality of model input data pertaining to precursor
emissions or meteorology, which tends to vary on a study-by-study basis.
In studies of short-term exposure, temporal variability of the exposure metric is of primary interest. For long-term
exposures, models that capture within-community spatial variation in individual exposure may be given more weight
for spatially variable ambient ozone. Given the low spatial variability of ozone at the urban scale, exposure
measurement error typically causes health effect estimates to be underestimated for studies of either short-term or
long-term exposure. Biases and decreases in the precision of the association (i.e., wider 95% CIs) tend to be small.
Even when spatial variability is higher near roads, the reduction in ozone exposure would cause the exposure to be
overestimated at a monitor distant from the road or when averaged across a model grid cell, so that health effects
would likely be underestimated.
6-72
-------
Table Annex 6-1 (Continued): Scientific considerations for evaluating the strength
of inference from studies on the health effects of
ozone.
Outcome Assessment/Evaluation
Controlled Human Exposure:
Endpoints should be assessed in the same manner for control and ozone exposure groups (e.g., time after
exposure, methods, endpoint evaluator) using valid, reliable methods. Blinding of endpoint evaluators is ideal,
especially for qualitative endpoints (e.g., histopathology). For each experiment and each experimental group,
including controls, precise details of all procedures carried out should be provided including how, when, and where.
Time of the endpoint evaluations is a key consideration that will vary depending on endpoint evaluated. Endpoints
should be assessed at time points that are appropriate for the research questions.
Animal Toxicology:
Endpoints should be assessed in the same manner for control and exposure groups (e.g., time after exposure,
methods, endpoint evaluator) using valid, reliable methods. Blinding of endpoint evaluators is ideal, especially for
qualitative endpoints (e.g., histopathology). For each experiment and each experimental group, including controls,
precise details of all procedures carried out should be provided including how, when, and where. Time of the
endpoint evaluations is a key consideration that will vary depending on endpoint evaluated. Endpoints should be
assessed at time points that are appropriate for the research questions.
Epidemiology:
Inference is stronger when outcomes are assessed or reported without knowledge of exposure status. Knowledge
of exposure status could produce artifactual associations. Confidence is greater when outcomes assessed by
interview, self-report, clinical examination, or analysis of biological indicators are defined by consistent criteria and
collected by validated, reliable methods. Independent, clinical assessment is valuable for outcomes like lung
function or incidence of disease, but report of physician diagnosis has shown good reliability.15 When examining
short-term exposures, evaluation of the evidence focuses on specific lags based on the evidence presented in
individual studies. Specifically, the following hierarchy is used in the process of selecting results from individual
studies to assess in the context of results across all studies for a specific health effect or outcome:
v. Distributed lag models;
vi. Average of multiple days (e.g., 0-2);
vii. If a priori lag days were used by the study authors these are the effect estimates presented; or
viii. If a study focuses on only a series of individual lag days, expert judgment is applied to select the
appropriate result to focus on considering the time course for physiologic changes for the health effect or
outcome being evaluated.
When health effects of long-term exposure are assessed by acute events such as symptoms or hospital
admissions, inference is strengthened when results are adjusted for short-term exposure. Validated questionnaires
for subjective outcomes such as symptoms are regarded to be reliable,0 particularly when collected frequently and
not subject to long recall. For biological samples, the stability of the compound of interest and the sensitivity and
precision of the analytical method is considered. If not based on knowledge of exposure status, errors in outcome
assessment tend to bias results toward the null.
Potential Copollutant Confounding
Controlled Human Exposure:
Exposure should be well characterized to evaluate independent effects of ozone.
Animal Toxicology:
Exposure should be well characterized to evaluate independent effects of ozone.
6-73
-------
Table Annex 6-1 (Continued): Scientific considerations for evaluating the strength
of inference from studies on the health effects of
ozone.
Epidemiology:
Not accounting for potential copollutant confounding can produce artifactual associations; thus, studies that
examine copollutant confounding carry greater weight. The predominant method is copollutant modeling
(i.e., two-pollutant models), which is especially informative when correlations are not high. However, when
correlations are high (r> 0.7), such as those often encountered for UFP and other traffic-related copollutants,
copollutant modeling is less informative. Although the use of single-pollutant models to examine the association
between ozone and a health effect or outcome are informative, ideally studies should also include copollutant
analyses. Copollutant confounding is evaluated on an individual study basis considering the extent of correlations
observed between the copollutant and ozone, and relationships observed with ozone and health effects in
copollutant models.
Other Potential Confounding Factorsd
Controlled Human Exposure:
Preference is given to studies using experimental and control groups that are matched for individual level
characteristics (e.g., race/ethnicity, sex, body weight, smoking history, age) and time varying factors (e.g., seasonal
and diurnal patterns).
Animal Toxicology:
Preference is given to studies using experimental and control groups that are matched for individual level
characteristics (e.g., strain, sex, body weight, litter size, feed and water consumption) and time varying factors
(e.g., seasonal and diurnal patterns).
Epidemiology:
Factors are considered to be potential confounders if demonstrated in the scientific literature to be related to health
effects and correlated with ozone. Not accounting for confounders can produce artifactual associations; thus,
studies that statistically adjust for multiple factors or control for them in the study design are emphasized. Less
weight is placed on studies that adjust for factors that mediate the relationship between ozone and health effects,
which can bias results toward the null. Confounders vary according to study design, exposure duration, and health
effect and may include, but are not limited to the following:
Short-term exposure studies: Meteorology, day of week, season, medication use, allergen exposure, and long-term
temporal trends.
Long-term exposure studies: Socioeconomic status, race, age, medication use, smoking status, stress, noise, and
occupational exposures.
Statistical Methodology
Controlled Human Exposure:
Statistical methods should be described clearly and appropriate for the study design and research question
(e.g., correction for multiple comparisons). Generally, statistical significance is used to evaluate the findings of
controlled human exposure studies. However, consistent trends are also informative. Detection of statistical
significance is influenced by a variety of factors including, but not limited to, the size of the study, exposure and
outcome measurement error, and statistical model specifications. Sample size is not a criterion for exclusion;
ideally, the sample size should provide adequate power to detect hypothesized effects (e.g., sample sizes less than
three are considered less informative). Because statistical tests have limitations, consideration is given to both
trends in data and reproducibility of results.
6-74
-------
Table Annex 6-1 (Continued): Scientific considerations for evaluating the strength
of inference from studies on the health effects of
ozone.
Animal Toxicology:
Statistical methods should be described clearly and appropriate for the study design and research question
(e.g., correction for multiple comparisons). Generally, statistical significance is used to evaluate the findings of
animal toxicology studies. However, consistent trends are also informative. Detection of statistical significance is
influenced by a variety of factors including, but not limited to, the size of the study, exposure and outcome
measurement error, and statistical model specifications. Sample size is not a criterion for exclusion; ideally, the
sample size should provide adequate power to detect hypothesized effects (e.g., sample sizes less than three are
considered less informative). Because statistical tests have limitations, consideration is given to both trends in data
and reproducibility of results.
Epidemiology:
Multivariable regression models that include potential confounding factors are emphasized. However, multipollutant
models (more than two pollutants) are considered to produce too much uncertainty due to copollutant collinearity to
be informative. Models with interaction terms aid in the evaluation of potential confounding as well as effect
modification. Sensitivity analyses with alternate specifications for potential confounding inform the stability of
findings and aid in judgments of the strength of inference from results. In the case of multiple comparisons,
consistency in the pattern of association can increase confidence that associations were not found by chance alone.
Statistical methods that are appropriate for the power of the study carry greater weight. For example, categorical
analyses with small sample sizes can be prone to bias results toward or away from the null. Statistical tests such as
f-tests and chi-squared tests are not considered sensitive enough for adequate inferences regarding ozone-health
effect associations. For all methods, the effect estimate and precision of the estimate (i.e., width of 95% CI) are
important considerations rather than statistical significance.
aU.S. EPA (2008V
"Muraia etal. (2014V Weakley et al. (2013V Yang et al. (2011V Heckbert et al. (2004V Barr et al. (2002V Muhaiarine et al. (1997V
Toren et al. (1993V
cBurnev et al. (1989V
dMany factors evaluated as potential confounders can be effect measure modifiers (e.g., season, comorbid health condition) or
mediators of health effects related to ozone (comorbid health condition).
6-75
-------
6.4 References
Barr. RG; Herbstman. J; Speizer. FE; Camargo. CA. Jr. (2002). Validation of self-reported chronic
obstructive pulmonary disease in a cohort study of nurses. Am J Epidemiol 155: 965-971.
http://dx.doi.org/10.1093/aie/155.lQ.965
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/aie/kwm396
Bell. ML; Dominici. F; Samet. JM. (2005). A meta-analysis of time-series studies of ozone and
mortality with comparison to the national morbidity, mortality, and air pollution study.
Epidemiology 16: 436-445. http://dx.doi.org/10.1097/01.ede.0000165817.4Q152.85
Bell. ML; Kim. JY; Dominici. F. (2007). Potential confounding of particulate 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; Mcdermott. A; Zeger. SL; Samet. JM; Dominici. F. (2004). Ozone and short-term mortality
in 95 US urban communities, 1987-2000. JAMA 292: 2372-2378.
http://dx.doi.org/10.1001/iama.292.19.2372
Bentaveb. M; Wagner. V; Stempfelet. M; Zins. M; Goldberg. M; Pascal. M; Larrieu. S; Beaudeau. P;
Cassadou. S; Eilstein. D; Filleul. L; Le Tertre. A; Medina. S; Pascal. L; Prouvost. H; Quenel. P;
Zeghnoun. A; Lefranc. A. (2015). Association between long-term exposure to air pollution and
mortality in France: A 25-year follow-up study. Environ Int 85: 5-14.
http://dx.doi.Org/10.1016/i.envint.2015.08.006
Burnev. PG; Laitinen. LA; Perdrizet. S; Huckauf. H; Tattersfield. AE: Chinn. S; Poisson. N; Heeren.
A; Britton. JR; Jones. T. (1989). Validity and repeatability of the IUATLD (1984) Bronchial
Symptoms Questionnaire: an international comparison. Eur Respir J 2: 940-945.
Buteau. S; Goldberg. MS; Burnett. RT; Gasparrini. A; Valois. MF; Brophv. JM; Crouse. PL;
Hatzopoulou. M. (2018). Associations between ambient air pollution and daily mortality in a
cohort of congestive heart failure: Case-crossover and nested case-control analyses using a
distributed lag nonlinear model. Environ Int. http://dx.doi.Org/10.1016/i.envint.2018.01.003
Cakmak. S; Hebbern. C; Pinault. L; Lavigne. E; Vanos. J; Crouse. PL; Tiepkema. M. (2018).
Associations between long-term PM2.5 and ozone exposure and mortality in the Canadian Census
Health and Environment Cohort (CANCHEC), by spatial synoptic classification zone. Environ Int
111: 200-211. http://dx.doi.org/10.1016/i.envint.2017.11.030
Cakmak. S; Hebbern. C; Vanos. J; Crouse. PL; Burnett. R. (2016). Ozone exposure and
cardiovascular-related mortality in the Canadian Census Health and Environment Cohort
(CANCHEC) by spatial synoptic classification zone. Environ Pollut 214: 589-599.
http://dx.doi.Org/10.1016/i.envpol.2016.04.067
Carey. IM; Atkinson. RW; Kent. AJ; van Staa. T; Cook. PG; Anderson. HR. (2013). Mortality
associations with long-term exposure to outdoor air pollution in a national English cohort. Am J
Respir Crit Care Med 187: 1226-1233. http://dx.doi.Org/10.l 164/rccm.201210-1758QC
6-76
-------
Chen. K; Wolf. K; Hampel. R; Stafoggia. M; Breitner. S; Cvrys. J; Samoli. E; Andersen. ZJ; Bero-
Bedada. G; Bellander. T; Hennig. F; Jacquemin. B; Pekkanen. J; Peters. A; Schneider. A. (2018).
Does temperature-confounding control influence the modifying effect of air temperature in ozone-
mortality associations? Environmental Epidemiology 2: 1-7.
http://dx.doi.org/10.1097/EE9.00000000000000Q8
Crouse. PL; Peters. PA; Hvstad. P; Brook. JR; van Donkelaar. A; Martin. RV; Villeneuve. PJ; Jerrett.
M; Goldberg. MS; Pope. CA; Brauer. M; Brook. RD: Robichaud. A; Menard. R; Burnett. RT.
(2015). Ambient PM 2.5, O 3, and NO 2 exposures and associations with mortality over 16 years of
follow-up in the Canadian Census Health and Environment Cohort (CanCHEC). Environ Health
Perspect 123: 1180-1186. http://dx.doi.org/10.1289/ehp.1409276
Di. Q; Dai. L; Wang. Y; Zanobetti. A; Choirat. C; Schwartz. JD; Dominici. F. (2017a). Association of
short-term exposure to air pollution with mortality in older adults. JAMA 318: 2446-2456.
http://dx.doi.org/10.1001/iama.2017.17923
Di. Q; Wang. Y; Zanobetti. A; Wang. Y; Koutrakis. P; Choirat. C; Dominici. F; Schwartz. JD.
(2017b). Air pollution and mortality in the Medicare population. N Engl J Med 376: 2513-2522.
http://dx.doi.org/10.1056/NEJMoal702747
Eckel. SP; Cockburn. M; Shu. YH; Deng. H; l.nrman n, FW; Liu. L; Gilliland. FD. (2016). Air
pollution affects lung cancer survival. Thorax 71: 891-898. http://dx.doi.org/10.1136/thoraxjnl-
2015-207927
Franklin. M; Schwartz. J. (2008). The impact of secondary particles on the association between
ambient ozone and mortality. Environ Health Perspect 116: 453-458.
http://dx.doi.org/10.1289/ehp.10777
Goldberg. MS; Burnett. RT; Stieb. DM; Brophv. JM; Daskalopoulou. SS; Valois. MF; Brook. JR.
(2013). Associations between ambient air pollution and daily mortality among elderly persons in
Montreal, Quebec. Sci Total Environ 463-464: 931942.
http://dx.doi.Org/10.1016/i.scitotenv.2013.06.095
Heckbert. SR; Kooperberg. C; Safford. MM; Psatv. BM; Hsia. J; McTiernan. A; Gaziano. JM;
Frishman. WH: Curb. JD. (2004). Comparison of self-report, hospital discharge codes, and
adjudication of cardiovascular events in the Women's Health Initiative. Am J Epidemiol 160: 1152-
1158. http://dx.doi.org/10.1093/aie/kwh314
Ito. K; De Leon. SF; Lippmann. M. (2005). Associations between ozone and daily mortality, analysis
and meta-analysis. Epidemiology 16: 446-457.
http://dx.doi.org/10.1097/01.ede.0000165821.9Q114.7f
Jerrett. M; Burnett. RT; Beckerman. BS; Turner. MC; Krewski. D; Thurston. G; Martin. RV: van
Donkelaar. A; Hughes. E; Shi. Y; Gapstur. SM; Thun. MJ; Pope. CA. III. (2013). Spatial analysis
of air pollution and mortality in California. Am J Respir Crit Care Med 188: 593-599.
http://dx.doi.Org/10.l 164/rccm.201303-0609QC
Jerrett. M; Burnett. RT; Pope. CA. Ill; 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
Jhun. I; Fann. N; Zanobetti. A; Hubbell. B. (2014). Effect modification of ozone-related mortality
risks by temperature in 97 US cities. Environ Int 73: 128-134.
http://dx.doi.Org/10.1016/i.envint.2014.07.009
6-77
-------
Katsouvanni. 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) (pp. 5-90). (Research Report 142). Boston, MA: Health Effects
Institute, https://www.healtheffects.org/publication/air-pollution-and-health-european-
and-north-american-approach
Kilkenny. C; Browne. WJ; Cuthill. TC; Emerson. M: Altman. DG. (2010). Improving bioscience
research reporting: The ARRIVE guidelines for reporting animal research [Review]. PLoS Biol 8:
el000412. http://dx.doi.org/10.1371/iournal.pbio.100Q412
Kim. H: Kim. J: Kim. S: Kang. SH: Kim. HJ: Kim. H: Heo. J: Yi. SM: Kim. K: Youn. TJ: Chae. IH.
(2017). Cardiovascular effects of long-term exposure to air pollution: a population-based study
with 900845person-years of follow-up. J Am Heart Assoc 6.
http://dx.doi.org/10.1161/JAHA.117.00717Q
Klemm. RJ: Lipfert. FW; Wvzga. RE: Gust. C. (2004). Daily mortality and air pollution in Atlanta:
two years of data from ARIES. Inhal Toxicol 16 Suppl 1: 131-141.
http://dx.doi.org/10.1080/0895837049Q443213
Klemm. RJ: Mason. RM. Jr. (2000). Aerosol Research and Inhalation Epidemiological Study
(ARIES): air quality and daily mortality statistical modeling—interim results. J Air Waste Manag
Assoc 50: 1433-1439.
Klemm. RJ: Thomas. EL: Wvzga. RE. (2011). The impact of frequency and duration of air quality
monitoring: Atlanta, GA, data modeling of air pollution and mortality. J Air Waste Manag Assoc
61: 1281-1291. http://dx.doi.org/10.1080/10473289.2011.617648
Klimisch. HJ: Andreae. M: Tillmann. U. (1997). A systematic approach for evaluating the quality of
experimental toxicological and ecotoxicological data. Regul Toxicol Pharmacol 25: 1-5.
http://dx.doi.org/10.1006/rtph.1996.1076
Lew. JI; Chemervnski. SM: Sarnat. JA. (2005). Ozone exposure and mortality, an empiric Bayes
metaregression analysis. Epidemiology 16: 458-468.
http://dx.doi.org/10.1097/01.ede.0000165820.Q8301.b3
Li. C: Balluz. LS: Vaidvanathan. A: Wen. X: Hao. Y; Qualters. J. R. (2016). Long-term exposure to
ozone and life expectancy in the United States, 2002 to 2008. Medicine (Baltimore) 95: e2474.
http://dx.doi.org/10.1097/MD.00000000000Q2474
Liu. T; Zeng. W: Lin. H; Rutherford. S: Xiao. J: Li. X: Li. Z; Qian. Z; Feng. B; Ma. W. (2016).
Tempo-spatial variations of ambient ozone-mortality associations in the usa: results from the
nmmaps data. Int J Environ Res Public Health 13. http://dx.doi.org/10.3390/iierphl3090851
Madrigano. J: Jack. D; Anderson. GB; Bell. ML: Kinney. PL. (2015). Temperature, ozone, and
mortality in urban and non-urban counties in the northeastern United States. Environ Health 14:3.
http://dx.doi.org/10.1186/1476-069X-14-3
Medina-Ramon. M; Schwartz. J. (2008). Who is more vulnerable to die from ozone air pollution?
Epidemiology 19: 672-679. http://dx.doi.org/10.1097/EDE.0b013e3181773476
Moolgavkar. SH: Mcclellan. RO: Dewanii. A: Turim. J: Luebeck. EG: Edwards. M. (2013). Time-
series analyses of air pollution and mortality in the United States: A subsampling approach.
Environ Health Perspect 121: 73-78. http://dx.doi.Org/10.1289/ehp.l 104507
Muhaiarine. N: Mustard. C: Roos. LL: Young. TK: Gelskev. DE. (1997). Comparison of survey and
physician claims data for detecting hypertension. J Clin Epidemiol 50: 711-718.
http://dx.doi.org/10.1016/S0895-4356(97')00019-X
6-78
-------
Murgia. N; Brisman. J; Claesson. A; Muzi. G; Olin. AC; Toren. K. (2014). Validity of a questionnaire-
based diagnosis of chronic obstructive pulmonary disease in a general population-based study.
BMC Pulm Med 14: 49. http://dx.doi.org/10.1186/1471-2466-14-49
NHLBI (National Institutes of Health, National Heart Lung and Blood Institute). (2017). NHLBI fact
book, fiscal year 2012: Disease statistics. Available online at
https://www.nhlbi.nih.gov/about/documents/factbook/2012/chapter4 (accessed August 23, 2017).
Peng. RD: Samoli. E: Pham. L: Dominici. F: Touloumi. G; Ramsay. T: Burnett. RT: Krewski. D: Le
Tertre. A; Cohen. A; Atkinson. RW; Anderson. HR; Katsouvanni. K; Samet. JM. (2013). Acute
effects of ambient ozone on mortality in Europe and North America: results from the APHENA
study. Air Qual Atmos Health 6: 445-453. http://dx.doi.org/10.1007/s 11869-012-0180-9
Roonev. AA; Bovles. AL; Wolfe. MS; Bucher. JR; Thaver. KA. (2014). Systematic review and
evidence integration for literature-based environmental health science assessments. Environ Health
Perspect 122: 711-718. http://dx.doi.org/10.1289/ehp.1307972
Rush. B; Mcdermid. RC; Celi. LA; Wallev. KR; Russell. JA; Bovd. JH. (2017). Association between
chronic exposure to air pollution and mortality in the acute respiratory distress syndrome. Environ
Pollut 224: 352-356. http://dx.doi.Org/10.1016/i.envpol.2017.02.014
Sacks. JD; Ito. K; Wilson. WE; Neas. LM. (2012). Impact of covariate models on the assessment of
the air pollution-mortality association in a single- and multipollutant context. Am J Epidemiol 176:
622-634. http://dx.doi.org/10.1093/aie/kwsl35
Schwartz. J. (2000). Assessing confounding, effect modification, and thresholds in the association
between ambient particles and daily deaths. Environ Health Perspect 108: 563-568.
Schwartz. J. (2005). How sensitive is the association between ozone and daily deaths to control for
temperature? Am J Respir Crit Care Med 171: 627-631. http://dx.doi.Org/10.l 164/rccm.200407-
933QC
Sese. L; Nunes. H; Cottin. V; Sanval. S; Didier. M; Carton. Z; Israel-Biet. D; Crestani. B; Cadranel. J;
Wallaert. B; Tazi. A; Maitre. B; Prevot. G; Marchand-Adam. S; Guillot-Dudoret. S; Nardi. A;
Durv. S; Giraud. V; Gondouin. A; Juvin. K; Borie. R; Wislez. M; Valevre. D; Annesi-Maesano. I.
(2017). Role of atmospheric pollution on the natural history of idiopathic pulmonary fibrosis.
Thorax 73: 145-150. http://dx.doi.Org/10.l 136/thoraxjnl-2017-209967
Spencer-Hwang. R; Knutsen. SF; Soret. S; Ghamsarv. M; Beeson. WL; Oda. K; Shavlik. D; Jaipaul.
N. (2011). Ambient air pollutants and risk of fatal coronary heart disease among kidney transplant
recipients. Am J Kidney Dis 58: 608-616. http://dx.doi.Org/10.1053/i.aikd.2011.05.017
Toren. K; Brisman. J; Jarvholm. B. (1993). Asthma and asthma-like symptoms in adults assessed by
questionnaires: A literature review [Review]. Chest 104: 600-608.
http://dx.doi.Org/10.1378/chest.104.2.600
Turner. MC; Jerrett. M; Pope. A. Ill; Krewski. D; Gapstur. SM; Diver. WR; Beckerman. BS;
Marshall. JD; Su. J; Crouse. PL; Burnett. RT. (2016). Long-term ozone exposure and mortality in a
large prospective study. Am J Respir Crit Care Med 193: 1134-1142.
http://dx.doi.Org/10.l 164/rccm.201508-1633QC
U.S. EPA (U.S. Environmental Protection Agency). (1991). Guidelines for developmental toxicity risk
assessment (pp. 1-71). (EPA/600/FR-91/001). Washington, DC: U.S. Environmental Protection
Agency, Risk Assessment Forum, http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=23162
U.S. EPA (U.S. Environmental Protection Agency). (1996a). Air quality criteria for ozone and related
photochemical oxidants [EPA Report]. (EPA/600/P-93/004AF). Research Triangle Park, NC.
6-79
-------
U.S. EPA (U.S. Environmental Protection Agency). (1996b). Guidelines for reproductive toxicity risk
assessment (pp. 1-143). (EPA/630/R-96/009). Washington, DC: U.S. Environmental Protection
Agency, Risk Assessment Forum, https://www.epa.gov/sites/production/files/2014-
11/documents/guidelines repro toxicitv.pdf
U.S. EPA (U.S. Environmental Protection Agency). (1998). Guidelines for neurotoxicity risk
assessment [EPA Report] (pp. 1-89). (EPA/630/R-95/00IF). Washington, DC: U.S. Environmental
Protection Agency, Risk Assessment Forum, http://www.epa.gov/risk/guidelines-neurotoxicitv-
risk-assessment
U.S. EPA (U.S. Environmental Protection Agency). (2005). Guidelines for carcinogen risk assessment
[EPA Report]. (EPA/630/P-03/00IB). Washington, DC: U.S. Environmental Protection Agency,
Risk Assessment Forum, https://www. epa.gov/sites/production/files/2013-
09/documents/cancer guidelines final 3-25-05.pdf
U.S. EPA (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: U.S.
Environmental Protection Agency, Office of Research and Development, National Center for
Environmental Assessment-RTP Office.
http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=149923
U.S. EPA (U.S. Environmental Protection Agency). (2008). Integrated science assessment for sulfur
oxides: Health criteria [EPA Report]. (EPA/600/R-08/047F). Research Triangle Park, NC: U.S.
Environmental Protection Agency, Office of Research and Development, National Center for
Environmental Assessment- RTP. http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=198843
U.S. EPA (U.S. Environmental Protection Agency). (2013a). Integrated science assessment for ozone
and related photochemical oxidants [EPA Report]. (EPA/600/R-10/076F). Research Triangle Park,
NC: U.S. Environmental Protection Agency, Office of Research and Development, National Center
for Environmental Assessment-RTP Division.
http://cfbub.epa.gov/ncea/isa/recordisplav.cfm?deid=247492
U.S. EPA (U.S. Environmental Protection Agency). (2013b). Toxicological review of
trimethylbenzenes (CASRN 25551-13-7, 95-63-6, 526-73-8, and 108-67-8) in support of summary
information on the Integrated Risk Information System (IRIS): revised external review draft [EPA
Report]. (EPA/635/R13/171a). Washington, D.C.: U.S. Environmental Protection Agency,
National Center for Environmental Assessment.
http://vosemite.epa.gov/sab/SABPRODUCT.NSF/b5d8alce9b07293485257375007012b7/eele28Q
e77586de985257b65005d37e7!QpenDocument
U.S. EPA (U.S. Environmental Protection Agency). (2015). Preamble to the integrated science
assessments [EPA Report]. (EPA/600/R-15/067). Research Triangle Park, NC: U.S. Environmental
Protection Agency, Office of Research and Development, National Center for Environmental
Assessment, RTP Division. https://cfpub.epa.gov/ncea/isa/recordisplav.cfm?deid=310244
Vanos. JK; Cakmak. S; Bristow. C; Brion. V; Tremblav. N; Martin. SL; Sheridan. SS. (2013).
Synoptic weather typing applied to air pollution mortality among the elderly in 10 Canadian cities.
Environ Res 126: 66-75. http://dx.doi.Org/10.1016/i.envres.2013.08.003
Vanos. JK; Cakmak. S; Kalkstein. LS; Yagouti. A. (2015). Association of weather and air pollution
interactions on daily mortality in 12 Canadian cities. Air Qual Atmos Health 8: 307-320.
http://dx.doi.org/10.1007/sll869-014-Q266-7
Vanos. JK; Hebbern. C: Cakmak. S. (2014). Risk assessment for cardiovascular and respiratory
mortality due to air pollution and synoptic meteorology in 10 Canadian cities. Environ Pollut 185:
322-332. http://dx.doi.Org/10.1016/i.envpol.2013.ll.007
6-80
-------
Vieira. VM; Villanueva. C; Chang. J; Ziogas. A; Bristow. RE. (2017). Impact of community
disadvantage and air pollution burden on geographic disparities of ovarian cancer survival in
California. Environ Res 156: 388-393. http://dx.doi.Org/10.1016/i.envres.2017.03.057
von Elm. E; Altman. DG; Egger. M; Pocock. SJ; Gotzschc. PC; Vandenbroucke. JP. (2007). The
strengthening the reporting of observational studies in epidemiology (strobe) statement: guidelines
for reporting observational studies [Review]. PLoS Med 4: e296.
http://dx.doi.org/10.1371/iournal.pmed.004Q296
Weakley. J; Webber. MP; Ye. F; Zeig-Owens. R; Cohen. HW: Hall. CB; Kelly. K; Prezant. DJ.
(2013). Agreement between obstructive airways disease diagnoses from self-report questionnaires
and medical records. Prev Med 57: 38-42. http://dx.doi.Org/10.1016/i.vpmed.2013.04.001
Weichenthal. S; Pinault. LL; Burnett. RT. (2017). Impact of oxidant gases on the relationship between
outdoor fine particulate air pollution and nonaccidental, cardiovascular, and respiratory mortality.
Sci Rep 7: 16401. http://dx.doi.org/10.1038/s41598-017-16770-v
Wilson. A; Rappold. AG; Neas. LM; Reich. BJ. (2014). Modeling the effect of temperature on ozone-
related mortality. Ann Appl Stat 8: 1728-1749. http://dx.doi.org/10.1214/14-AOAS754
Xu. X; Ha. S; Kan. H; Hu. H; Curbow. BA; Lissaker. CTK. (2013). Health effects of air pollution on
length of respiratory cancer survival. BMC Public Health 13: 800. http://dx.doi.org/10.1186/1471-
2458-13-800
Yang. CL; To. T; Fotv. RG; Stieb. DM; Dell. SD. (2011). Verifying a questionnaire diagnosis of
asthma in children using health claims data. BMC Pulm Med 11. http://dx.doi.org/10.1186/1471-
2466-11-52
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.l 164/rccm.200706-823 OC
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-
0227QC
6-81
-------
APPENDIX 7 HEALTH EFFECTS —OTHER
HEALTH ENDPOINTS
Summary of I ansality Determinations fitr Other Health Eff ects
This \ppeiidi\ eh;ir;ielen/es llie seieiilifie e\ uleiiee lh;il suppuris e;ius;ihl\
deleriiiiiKiiimis I'm' slkiii- ;nul li'im-ieriii ii/mie exposure mid he;illh elleels. iiiehidiiiu
Repri'diieli\ e ;md I )e\ eli'pineiil;il I ill eels (see Seelimi ~ 11. \er\ diis S\ siem I! ITeels (see
Seel k'li ~.21. ;md Cmieerisee Seelimi ~ '). The l> pes iil'siudies e\ ;ilu;iled williin llns \ppeiuhx
:ire ei'iisisienl \x illi llie in er;ill sei'pe i'l llie IS \ ;is del;nled in I he IVTuee In ;issessnm ihe
in er;ill e\ idenee. llie sireimlhs ;nnl h 1111 i;iik
-------
7.1
Reproductive and Developmental Effects
7.1.1 Introduction, Summary from the 2013 Ozone ISA, and Scope for Current
Review
This section evaluates the scientific evidence related to the potential effects of ozone on
reproductive outcomes, including (1) male and female reproduction and fertility (Section 7.1.2) and
(2) pregnancy and birth outcomes (Section 7.1.3). The body of literature characterizing reproductive and
developmental effects has expanded since the 2013 Ozone ISA (U.S. EPA. 2013). with the addition of
epidemiologic studies and short- and long-term animal toxicological studies. Well-designed studies that
consider sources of bias, including potential confounding by copollutant exposures, are emphasized in
this section.
Because the average length of gestation in rodents is 18-24 days, animal toxicological studies
investigating the effects of ozone are typically considered short-term studies. By contrast, epidemiologic
studies that examine the entire pregnancy are considered to have a long-term exposure period (about
40 weeks, on average). In order to evaluate and characterize the evidence for the effects of PM on
reproductive and developmental effects in a consistent, cohesive, and integrated manner, results from
studies of both short- and long-term exposure periods are included in Section 7.1 and are identified
accordingly in the text and tables throughout this section. A related issue in studying environmental
exposures and reproductive and developmental effects is selecting the relevant exposure period, since the
biologically plausible pathways leading to these outcomes and the critical periods of exposure are often
not completely understood. Multiple exposure periods are evaluated in many epidemiologic studies,
including long-term (multiple months) exposure periods, such as entire pregnancy, individual trimesters
or months of pregnancy, and short-term (days to weeks) exposure periods, such as the days and weeks
immediately preceding birth. Thus, the evaluation of biological plausibility for the effects of ozone on
reproductive and developmental outcomes will combine short- and long-term exposures. Further, infants
and fetal development processes may be particularly sensitive to ozone exposure, and although the
physical mechanisms are not always fully understood, the effects from exposure to ozone exposure during
these critical windows of development may have permanent, lifelong effects.
The 2013 Ozone ISA (U.S. EPA. 2013) determined that the evidence was suggestive of a causal
relationship between exposures to ozone and reproductive and developmental effects. Epidemiologic and
toxicological studies provided evidence for an effect of prenatal exposure to ozone on pulmonary
structure and function in infants, as well as alterations in placental and pup cytokines, and increased pup
airway hyper-reactivity. Also, there was limited toxicological evidence for an effect of prenatal and early
life exposure on central nervous system effects, including laterality, brain morphology, neurobehavioral
abnormalities, and sleep aberration. Epidemiologic studies examining the effects of ozone on sperm
7-2
-------
quality provided limited evidence for decrements in sperm concentration, which was supported by limited
toxicological evidence for testicular degeneration associated with ozone exposure. While the collective
evidence for many of the birth outcomes examined in the 2013 Ozone ISA was generally inconsistent
(including birth defects), there were several well-designed, well-conducted studies that indicated an
association between ozone and adverse outcomes. For example, as part of the southern California
Children's Health Study, Salam et al. (2005) observed a concentration-response association of decreasing
birth weight with increasing ozone concentrations averaged over the entire pregnancy, especially evident
at levels above 30-ppb. Similarly, Hansen et al. (2008). using fetal ultrasonic measurements, found a
decrease in average fetal size associated with ozone during days 31-60 of gestation for women living
within 2 km of a monitoring site.
The current ISA builds upon findings from the 2013 Ozone ISA but new evidence since the 2013
Ozone ISA has expanded such that separate causality determinations can be made for male and female
fertility and reproduction (see Section 7.1.2). and pregnancy and birth outcomes (see Section 7.1.3).
which are likely to have different etiologies and critical exposure windows over different lifestages.
Summaries of the effects of ozone exposure during developmental periods are included in Section 7.1.4;
however, full descriptions and causality determinations are found in the designated appendix for
individual outcomes (i.e., respiratory [see Appendix 31. cardiovascular [see Appendix 41. metabolic [see
Appendix 51. and nervous system effects [see Section 7.21.)
7.1.1.1 Population, Exposure, Comparison, Outcome, and Study Design (PECOS)
Tool
The scope of this section is defined by a scoping tool that generally describes the relevant
Population, Exposure, Comparison, Outcome, and Study Design (PECOS). The PECOS tool defines the
parameters and provides a framework to help identify the relevant literature to inform the ISA. Because
the 2013 Ozone ISA concluded that there was evidence to suggest a causal relationship between
long-term ozone exposure and reproductive and developmental effects, studies are evaluated regardless of
study location. The studies evaluated and subsequently discussed within this section were identified using
the following PECOS tools:
Experimental Studies:
• Population: Study population of any animal toxicological study of mammals at any lifestage
• Exposure: Long-term (in the order of months to years) or short-term (hours to less than
4 complete weeks) inhalation exposure to relevant ozone concentrations (i.e., <2 ppm)
• Comparison: Appropriate comparison group exposed to a negative control (i.e., clean air or
filtered-air control)
• Outcome: Reproductive or developmental effects
7-3
-------
• Study Design: In vivo chronic, subchronic or repeated-dose toxicity studies in mammals;
reproductive toxicity or immunotoxicity studies; genotoxicity/mutagenicity studies (studies that
examine the effects of exposure during developmental periods contribute the causality
determinations in Appendix 3 Appendix 4. and Appendix 5. and Section 7.2.2.7 and are
summarized in Section 7.1.4.)
Epidemiologic Studies:
• Population: Any population, including populations or lifestages that might be at increased risk
• Exposure: Long-term (in the order of months to years) or short-term (hours to less than
4 complete weeks)
• Comparison: Per unit increase (in ppb)
• Outcome: Change in risk (incidence/prevalence) of a reproductive or developmental effect
• Study Design: Epidemiologic studies consisting of cohort and case-control studies; time-series,
case-crossover, and cross-sectional studies with appropriate timing of exposure for the health
endpoint of interest
7.1.2 Male and Female Reproduction and Fertility
Reproductive health issues, commonly identified through impaired fecundity (ability to conceive)
and fertility (ability to have live born children), affect up to 15% of couples attempting to conceive, with
both male and female factors contributing (Thoma etal.. 2013). In the U.S., approximately 9% of men
aged 18-44 years and 11% of women aged 15-44 years are infertile (Agarwal et al.. 2015: Chandra et al..
2013). Reproductive health issues can have negative effects on quality of life and may signal poorer
physiological health and increased risk of adverse health outcomes during pregnancy and birth.
7.1.2.1 Biological Plausibility
When considering the available health evidence, there are plausible pathways connecting
inhalation of ozone to the apical reproductive and developmental events reported in epidemiologic
studies. This section describes biological pathways that potentially underlie reproductive and
developmental health effects specific to male and female reproduction and fertility resulting from
exposure to ozone. Biological plausibility is graphically depicted via the proposed pathways as a
continuum of upstream events, connected by arrows, that may lead to downstream events observed in
epidemiologic studies (Figure 7-1). This discussion of "how" exposure to ozone may lead to effects on
male and female reproduction and fertility contributes to an understanding of the biological plausibility of
epidemiologic results evaluated later. Note that the structure of the biological plausibility sections and the
role of biological plausibility in contributing to the weight-of-evidence analysis used in the 2020 Ozone
ISA are discussed in Section IS.4.2.
7-4
-------
When considering the available health evidence, there are plausible pathways connecting
inhalation of ozone to the apical reproductive effects reported in epidemiologic studies. The biological
plausibility for ozone-induced effects on reproduction and fertility is supported by evidence from the
2013 Ozone ISA (U.S. EPA. 2013) and by new evidence. Evidence from experimental and epidemiologic
studies indicates that ozone inhalation may initiate these pathways resulting in a series of physiological
responses that could lead to male and female reproductive effects and altered fertility (e.g., fertility,
fecundity, reproduction). The evidence for the initial events (Figure 7-1) that could result in effects on
fertility and reproduction includes respiratory tract inflammation following the inhalation of ozone.
Respiratory tract inflammation can be followed by systemic inflammation [e.g., C-reactive protein (CRP);
Lee et al. (2011); see Section 4.2.111. Ozone exposure may induce inflammatory or other processes in
extrapulmonary compartments. Beyond these events, there is also evidence from experimental and
epidemiologic studies demonstrating that exposure to ozone could result in a coherent series of
physiological responses that provide biological plausibility for the associations reported in epidemiologic
and laboratory animal studies, including altered fertility, fecundity, and reproduction.
Ozone
Exposure
Altered Fertility, Fecundity,
Reproduction
Note: The boxes above represent the effects for which there is experimental or epidemiologic evidence related to ozone exposure,
and the arrows indicate a proposed relationship between those effects. Solid arrows denote evidence of essentiality as provided, for
example, by an inhibitor of the pathway or a genetic knockout model used in an experimental study involving ozone exposure.
Shading around multiple boxes is used to denote a grouping of these effects. Arrows may connect individual boxes, groupings of
boxes, and individual boxes within groupings of boxes. Progression of effects is generally depicted from left to right and color-coded
(white, exposure; green, initial effect; blue, intermediate effect; orange, effect at the population level or a key clinical effect). Here,
population level effects generally reflect results of epidemiologic studies. When there are gaps in the evidence, there are
complementary gaps in the figure and the accompanying text below. The structure of the biological plausibility sections and the role
of biological plausibility in contributing to the weight-of-evidence analysis used in the 2020 Ozone ISA are discussed in
Section IS.4.2.
Figure 7-1 Potential biological pathways for male reproduction and fertility
effects following ozone exposure.
As depicted in Figure 7-1. these initial events can give rise to intermediate events, including
systemic inflammation from epidemiologic evidence of increased CRP during pregnancy, animal studies
of altered sperm quality, altered testicular morphology, aberrant testicular histology, including a depletion
of testicular germ cells and a decreased seminiferous tubule epithelial layer, impaired spermatogenesis
with focal epithelial cell desquamation to the basement membrane, and presence of giant spermatid cells
7-5
-------
(Jcdlinska-Krakowska et al.. 2006b). The 2013 Ozone ISA documented epidemiologic studies that
showed decreased sperm quality, and there is recent evidence for sperm-related effects, including
decreased sperm concentration and decreased sperm count with ozone exposure to males with lupus. In
studies of female reproduction, epidemiologic studies have shown altered reproductive success with
ozone exposure, with the effect differing by timing of ozone exposure. In the animal literature, female
reproductive outcomes from the 2013 Ozone ISA showed decreased reproductive success with ozone
exposure over much of pregnancy (gestation Days [GDs] 9-18); ozone exposure can induce a temporary
anorexigenic effect in pregnant dams. Together, these proposed pathways provide biological plausibility
for epidemiologic results of reproductive and developmental health effects and will be used to inform a
causality determination, which is discussed later in this Appendix.
7.1.2.2 Male Reproduction
7.1.2.2.1 Epidemiologic Evidence of Effects on Male Reproductive Function
Associations between male reproductive health and ozone exposure have been examined through
effects on sperm. In the 2013 Ozone ISA, there was evidence from a limited number of epidemiologic
studies that observed associations between ozone concentration and sperm quality. Associations between
reductions in sperm concentration and both short- and long-term ozone exposures were also observed.
Recent evidence includes a small panel study in Brazilian men with systematic lupus erythematosus that
reported decreases in sperm concentration and count with long-term (0-90 days before collection) ozone
exposure (Farhat et al.. 2016). and a Chinese cohort that observed no evidence of association (Liu et al..
2017). Data from current studies of male reproductive function are extracted and summarized in the
evidence inventories (see Table 7-6).
7.1.2.2.2 Toxicological Evidence of Effects on Male Reproductive Function
There are no recent animal toxicological studies on male reproduction. Evidence from the 2013
Ozone ISA showed decremental effects on testicular morphology demonstrated in a toxicological study
with histological evidence of ozone-induced depletion of germ cells in testicular tissue and decreased
seminiferous tubule epithelial layer (Jedlinska-Krakowska et al.. 2006a). In summary, this study provided
toxicological evidence of impaired spermatogenesis with ozone exposure that was attenuated by
antioxidant supplements.
7-6
-------
7.1.2.2.3 Summary
Overall, a limited number of epidemiologic studies provides evidence of an association between
ozone concentration and impaired spermatogenesis and decreased sperm count. Detrimental effects on
testicular morphology and impaired spermatogenesis were also demonstrated in a toxicological study.
7.1.2.3 Female Reproduction
7.1.2.3.1 Epidemiologic Evidence of Effects on Female Reproductive Function
A single study in the 2013 Ozone ISA showed some evidence for increased in vitro fertilization
(IVF) success with short-term ozone exposure during ovulation, but long-term exposure during gestation
reduced the likelihood of a live birth (Lcgro et al.. 2010). In recent studies, the overall findings are mixed.
In a French population undergoing IVF, Carre et al. (2016) observed an increased number of top embryos
(i.e., those considered of the best quality) with at least 1 day of high ozone exposure in the 30 day period
before ovulation. Another study found no evidence of association with exposures that occurred up to
2 months before conception, but did show an improvement in fecundity with ozone exposure
post-conception (Slama et al.. 2013). However, a longitudinal study in 500 U.S. couples reported
decreased fecundity with short-term ozone exposure near time of ovulation (Nobles et al.. 2018). Data
from current studies of female reproductive function are extracted and summarized in Table 7-7.
7.1.2.3.2 Toxicological Evidence of Effects on Female Reproduction
Evidence from the 2013 Ozone ISA showed that, in most toxicological studies, reproductive
success appears to be unaffected by ozone exposure. Nonetheless, one study reported that 25% of the
BALB/c mouse dams in the highest ozone exposure group (1.2 ppm, short-term exposure GDs 9-18),
compared to 55% in the filtered air group, did not complete a successful pregnancy (Sharkhuu et al..
2011). Ozone administration (continuous 0.4, 0.8 or 1.2 ppm ozone) to CD-I mouse dams throughout
most of the pregnancy (short-term exposure, PNDs 7-17, which excludes the preimplantation period) led
to no adverse effects on reproductive success [proportion of successful pregnancies, litter size, sex ratio,
frequency of still birth, or neonatal mortality; Bignami et al. (1994)1. There was a statistically
nonsignificant increase in pregnancy duration (0.8 and 1.2 ppm ozone). Initially, dam body weight (0.8
and 1.2 ppm ozone), water consumption (0.4, 0.8 and 1.2 ppm ozone), and feed consumption (0.4, 0.8 and
1.2 ppm ozone) during pregnancy were decreased with ozone exposure, but these deficits dissipated a
week or two after the initial exposure (Bignami et al.. 1994). This anorexigenic effect of ozone exposure
on the pregnant dam appeared to subside with time; the dams seemed to adapt to the ozone exposure.
Some evidence suggests that ozone may affect reproductive success when combined with other
chemicals. Kavlock et al. (1979) reported that ozone acted synergistically with sodium salicylate to
7-7
-------
increase the rate of pup resorptions after midgestational exposure (1.0 ppm ozone, short-term exposure,
GDs 9-12). With ozone exposure, toxicological studies showed reproductive effects to include a transient
anorexigenic effect of ozone on gestational weight gain, and a synergistic effect of ozone on
salicylate-induced pup resorptions; other fecundity, pregnancy- and gestation-related outcomes were
unaffected by ozone exposure.
7.1.2.3.3 Summary
In conclusion, results from epidemiologic studies are mixed, with benefits and detriments to
female reproductive function observed with ozone exposures, while a limited number of toxicological
studies provides some evidence of effects on successful completion of pregnancy.
7.1.3 Pregnancy and Birth Outcomes
7.1.3.1 Biological Plausibility
This section describes biological pathways that potentially underlie reproductive and
developmental health effects of pregnancy, birth weight, and birth outcomes resulting from exposure to
ozone. Figure 7-2 graphically depicts the proposed pathways as a continuum of upstream events,
connected by arrows, that may lead to the downstream events. This discussion of "how" exposure to
ozone may lead to reproductive and developmental health effects contributes to an understanding of the
biological plausibility of effects evaluated in Section 7.1.3.2 through Section 7.1.3.5. Note that the
structure of the biological plausibility sections and the role of biological plausibility in contributing to the
weight-of-evidence analysis used in the 2020 Ozone ISA are discussed in Section IS.4.2.
7-8
-------
Activation of Sensory
Nerves in
Respiratory Tract
Respiratory
Inflammation and
Oxidative Stress
Ozone
Exposure
—m
Elevated sFlt-1, a
biomarker of pre-
eclampsia
Impaired vascular
remodeling during
pregnancy (uterine
artery resistance
changes)
Decreased dam body
weight gain during
pregnancy
Gestational Diabetes
Decreased pup growth
in utero
Altered Anthropometric
Measurements at birth
(head circumference,
body length)
Altered Litter Size
Altered Fetal Growth
and/or Birth Weight
Note: The boxes above represent the effects for which there is experimental or epidemiologic evidence related to ozone exposure,
and the arrows indicate a proposed relationship between those effects. Solid arrows denote evidence of essentiality as provided, for
example, by an inhibitor of the pathway or a genetic knockout model used in an experimental study involving ozone exposure.
Shading around multiple boxes is used to denote a grouping of these effects. Arrows may connect individual boxes, groupings of
boxes, and individual boxes within groupings of boxes. Progression of effects is generally depicted from left to right and color-coded
(white, exposure; green, initial effect; blue, intermediate effect; orange, effect at the population level or a key clinical effect). Here,
population level effects generally reflect results of epidemiologic studies. When there are gaps in the evidence, there are
complementary gaps in the figure and the accompanying text below. The structure of the biological plausibility sections and the role
of biological plausibility in contributing to the weight-of-evidence analysis used in the 2020 Ozone ISA are discussed in
Section IS.4.2.
Figure 7-2 Potential biological pathways for pregnancy and birth outcomes
following ozone exposure.
Evidence is accumulating that ozone exposure may affect pregnancy and birth outcomes. The
evidence from the 2013 Ozone ISA (U.S. EPA. 2013) and recent evidence indicates multiple initial events
after ozone inhalation contribute to effects on pregnancy and birth outcomes. Beyond these initial events,
there is also evidence from experimental and epidemiologic studies demonstrating that ozone inhalation
could result in a coherent series of physiological responses, including systemic inflammation or oxidative
stress, that provide biological plausibility for the associations reported in epidemiologic studies and
animal toxicological studies of pregnancy-induced hypertension, altered development, preterm birth, and
altered fetal growth or birth weight (Geer et al.. 2012; Morello-Frosch et al.. 2010; Salam et al.. 2005).
Oxidative stress is reported in the epidemiologic literature with ozone-dependent increased odds of
elevated C-reactive protein (CRP) levels in nonpregnant individuals but CRP was unchanged at GD 5 in
ozone-exposed pregnant rodents (Miller et al.. 2019). Other initial events include activation of sensory
nerves in the respiratory tract. In pregnant rodents exposed to ozone peri-implantation at GD 5,
circulating serum cytokines are altered including statistically significantly decreased IL-6, IFN-y, and
IL-13 (Miller et al.. 2019) at a point when these cytokines may be critical for proper implantation. Serum
from these ozone-exposed dams added to trophoblasts in vitro led to impaired trophoblast invasion and
migration as well as impaired trophoblast metabolic capacity (Miller et al.. 2019). Ozone exposure using
this in vitro trophoblast model also caused the trophoblasts to produce increased levels of soluble fms-like
tyrosine kinase 1 (sFlt-1), a biomarker of pre-eclampsia (Miller et al.. 2019). Further, ozone-dependent
7-9
-------
reproductive organ-specific effects, included altered uterine artery vascularity of increased resistance
during periods of pregnancy when resistance should be decreasing to accommodate the physiological
changes of pregnancy rMiller et al. (2017); Section 7.1.4 and Section 7.1.51. At a certain point in a normal
pregnancy, vascular resistance decreases in the uterine artery, which enhances perfusion of the fetus and
placenta. But this pathway is altered substantially in ozone-exposed animals. Evidence from the 2013
Ozone ISA showed ozone-exposed virgin rodents manifest with impaired thyroid hormone status,
decreased T3, T4, and TSH, hormones that are important for pregnancy; thyroid hormone status has not
been monitored in gravid animals. Ozone-exposed pregnant dams eat less food and gain less weight than
control animals (Miller et al.. 2019; Miller et al.. 2017; Bignami et al.. 1994). an effect that may dissipate
with time (Bignami et al.. 1994) or when ozone exposure ceases (Miller et al.. 2017).
7.1.3.2 Maternal Health during Pregnancy
7.1.3.2.1 Epidemiologic Evidence of Effects on Maternal Health during Pregnancy
Studies of maternal health during pregnancy focus on hypertensive disorders of pregnancy, such
as preeclampsia and gestational hypertension, and gestational diabetes. Pregnancy-associated
hypertension is a leading cause of perinatal and maternal mortality and morbidity. Gestational diabetes
may increase the risk of high blood pressure during pregnancy and the occurrence of cesarean delivery; it
is also frequently related to later development of type 2 diabetes. Epidemiologic studies related to
maternal health during pregnancy were not identified for inclusion in the 2013 Ozone ISA (U.S. EPA.
2013). Most recent studies in this area investigated associations between ozone exposure and hypertensive
disorders of pregnancy, while a limited number examined the development of gestational diabetes.
For hypertensive disorders of pregnancy, study results were mixed, with positive (increased
hypertensive disorders associated with increased ozone concentrations) and null associations reported.
Studies for gestational diabetes are few, and they reported both null and positive associations depending
on timing of exposure.
• There are differences in studies by specific definition of outcomes; some studies examined
preeclampsia, some hypertension, and others hypertensive disorders of pregnancy, which may or
may not have included preeclampsia.
• Results for studies of preeclampsia were mixed, with some showing positive associations for first
trimester exposures (Lee et al.. 2013; Olsson et al.. 2013) and others reporting either no evidence
of association across different exposure time periods (Mendola et al.. 2016b) or positive effects
only in some study areas (Wuetal.. 2011).
• Studies of hypertensive disorders of pregnancy generally reported positive associations (Hu et al..
2016; Michikawa etal.. 2015; Mobasher et al.. 2013).
7-10
-------
• Two studies examining hypertension reported mixed associations with first trimester exposure,
with (Lee et al.. 2013) showing positive associations and (Xu et al.. 2014) showing no evidence
of association.
• Increased odds of gestational diabetes were observed for higher ozone exposures during the first
and second trimesters in a Florida population compared to lower ozone exposures (Hu et al..
2015) and for weekly exposures during the second trimester in a national study (Roblcdo et al..
2015).
• No evidence of association with gestational diabetes was observed in the national study for ozone
exposures 90 days before conception and in the first trimester (Roblcdo et al.. 2015).
• The single study of hypertensive disorders of pregnancy examined the potential for copollutant
confounding and showed an odds ratio increase (from 1.05 to 1.11) with adjustment for NO2
(Olsson et al.. 2013). In both studies of gestational diabetes, adjustment for copollutants did not
change effect estimates (Hu et al.. 2015; Robledo et al.. 2015). reducing uncertainties that the
associations observed with ozone are due to copollutant confounding.
Data from current studies of maternal health during pregnancy are extracted and summarized in
the evidence inventories (see Table 7-8. Table 7-9. and Table 7-15).
7.1.3.2.2 Toxicological Evidence of Effects on Pregnancy
Studies from the 2013 Ozone ISA demonstrated a transient anorexigenic effect of ozone on
pregnant rodent dam weight gain during pregnancy. Initially, dam body weight (0.8 and 1.2 ppm ozone),
water consumption (0.4, 0.8, and 1.2 ppm ozone), and feed consumption (0.4, 0.8, and 1.2 ppm ozone)
during pregnancy were decreased with ozone exposure but these deficits dissipated a week or two after
the initial exposure (Bignami et al.. 1994). The anorexigenic effect of ozone exposure on the pregnant
dam appears to dissipate with time; the dams seem to adapt to the ozone exposure. Studies from the 2013
Ozone ISA also demonstrated enhanced pulmonary inflammatory response in BALF of pregnant and
lactating rodents to ozone exposure (1.0 ppm, 6 hours); there was significantly enhanced sensitivity to
ozone-induced pulmonary inflammation during pregnancy, which was maintained during lactation, and
disappeared after lactation ceased at weaning (Gunnison et al.. 1992). Research since the 2013 Ozone ISA
also shows that ozone affects weight gain during pregnancy. Pregnant rats exposed to ozone (0.8 ppm
ozone) during the period of implantation (GDs 5-6) showed significantly lower body-weight gain during
this period (Miller et al.. 2017). demonstrating a similar anorexigenic effect as documented in the 2013
Ozone ISA. Exposure to 0.4 ppm ozone during implantation did not affect dam body weight gain during
pregnancy. Miller et al. (2017) also assessed dam blood pressure (GD 15, GD 19, GD 21) and kidney
histopathology in near-term ozone exposed dams to evaluate whether ozone exposure might contribute to
gestational hypertension/preeclampsia, with data showing null findings. Peri-implantation ozone exposure
(1.2 ppm, GD 5) caused increased homeostatic model assessment for insulin resistance (HOMA-IR) and
increased area under the curve with the glucose tolerance test in dams immediately after ozone exposure;
exposure to 0.4 or 0.8 ppm ozone did not induce these metabolic changes in the dam (Miller et al.. 2019).
7-11
-------
Data from current studies of maternal health during pregnancy are extracted and summarized in the
evidence inventories (see Section 7.6.2. Table 7-16).
7.1.3.2.3 Summary
Evidence for effects on maternal health during pregnancy is mixed, with a limited number of
epidemiologic studies providing some evidence for effects on hypertensive disorders of pregnancy and
gestational diabetes, and toxicological studies showing changes in maternal weight during pregnancy.
7.1.3.3 Fetal Growth, Birth Weight, and Body Length at Birth
Fetal growth is a marker of fetal well-being during pregnancy and an important indicator of future
infant and child health. Fetal growth can be difficult to quantify, and growth standards vary by
race/ethnicity, infant sex, parity, and maternal size (Zhang et al.. 2010). Birth weight is often used as a
proxy for fetal growth, either as a continuous measure or below a cutoff (typically 2,500 g) as low birth
weight. However, birth weight is determined through a mix of factors, including intrauterine growth and
gestational age, among others, so studies of these outcomes will often restrict to term births. Exposures
that may affect birth weight could potentially occur throughout pregnancy, as growth may be affected by
structural changes in the placenta or the placentation process or through inflammatory processes that
restrict nutritional flow to the fetus (Figure 7-2).
7.1.3.3.1 Epidemiologic Evidence for Fetal Growth, Birth Weight, and Body Length at Birth
In the current review, fetal growth is quantified through small-for-gestational-age measures
(typically an infant below the 10th percentile of weight for gestational age accounting for race and sex),
continuous birth weight in grams, and dichotomized low birth weight (less than 2,500 g or 5 lbs., 8 oz). In
the 2013 Ozone ISA, studies were exclusively of birth weight with only a limited number supporting an
association between ozone exposure and lower birth weight. Since then, the number of recent studies has
more than doubled, but findings remain largely inconsistent, with studies reporting either lower birth
weight or no evidence of association of lower birth weight with ozone concentrations across exposure
windows, study areas, study designs, and exposure assessment methods. Data from current studies of fetal
growth are extracted and summarized in the evidence inventories (see Table 7-10).
• Studies that examined continuous birth weight, including well-designed studies [e.g., Vinikoor-
Imler et al. (2014); Laurent et al. (2013)1. reported primarily that increases in ozone
7-12
-------
concentrations were associated with decrements in birth weight, although the magnitude of the
decrement varied, ranging from -4.61 to -27.27 g (per 10 ppb increase in ozone).1
• One study using geographically weighted regression indicated variation by spatial characteristics,
with lower birth weight associated with higher ozone concentration in less urbanized
communities (Tu et al.. 2016).
• Some studies of odds of low birth weight (<2,500 g), including well designed studies [e.g., Chen
et al. (2017b); Laurent et al. (2016a); Vinikoor-Imler et al. (2014); Laurent et al. (2013)1. reported
increased odds of low birth weight with increased ozone concentrations; however, those
associations are inconsistent across exposure windows. Studies reporting associations with ozone
concentrations (i.e., odds ratios [ORs]) during the entire pregnancy ranged from 1.03 [95% CI:
1.02, 1.05; Laurent et al. (2016a)I to 2.20 [95% CI: 1.74, 2.75; Chen etal. (2017b)l.
• In studies with copollutant adjusted models, effect estimates were largely similar to those
reported for single-pollutant models for ozone (Smith et al.. 2017; Ha et al.. 2014; Olsson et al..
2013V
7.1.3.3.2 Toxicological Evidence for Fetal Growth, Birth Weight, and Body Length at Birth
Evidence from the 2013 Ozone ISA showed decreased birth weight in pups whose pregnant dams
were exposed to ozone during pregnancy (Sharkhuu et al.. 2011; Haro and Paz. 1993). but no effects on
the number of pups born. A few studies reported that mice or rats exposed developmentally (gestationally
and/or lactationally) to ozone had deficits in postnatal body-weight gain (Bignami et al.. 1994; Haro and
Paz. 1993; Kavlock et al.. 1980). Recent animal toxicological evidence also shows that fetuses whose
dams were exposed to ozone (0.8 ppm for both sexes, 0.4 ppm ozone for male fetuses) during the period
of implantation (GDs 5-6) weighed significantly less than the air-exposed control pups at GD 21, near the
end of pregnancy (Miller et al.. 2017). In addition, male pups more sensitive to ozone exposure than
female pubs. Further examination showed that dams exposed to 0.8 ppm ozone had male and female
fetuses with significantly lower lean mass and fat mass compared with control-air dams at GD 21 (Miller
et al.. 2017V In summary, animal toxicological models show ozone exposure caused decreased birth
weight and decreased postnatal body-weight gain but did not affect litter number.
7.1.3.3.3 Summary
Overall, there is some epidemiologic evidence for the effects of ozone on fetal growth, especially
for continuous-term birth weight, a conclusion supported by toxicological evidence in rodents.
1 All epidemiologic results standardized to a 15-ppb increase in 24-hour avg, 20-ppb increase in 8-hour daily max,
25-ppb increase in 1-hour daily max ozone concentrations, or a 10-ppb increase in seasonal/annual ozone
concentrations to facilitate comparability across studies.
7-13
-------
7.1.3.4
Preterm Birth
Preterm birth (PTB), delivery that occurs before 37 weeks of completed gestation, is a marker for
fetal underdevelopment and is related to subsequent adverse health outcomes (Saigal and Dovle. 2008;
IOM. 2007; MacDorman et al.. 2007; Gilbert et al.. 2003). PTB is characterized by multiple etiologies
(spontaneous, premature rupture of membranes [PROM], or medically induced), which may have either
individual or shared mechanistic pathways (Figure 7-2).
7.1.3.4.1 Epidemiologic Evidence of Preterm Birth
In the 2013 Ozone ISA, short-term exposure to ozone during late pregnancy was consistently not
associated with preterm birth. However, associations with long-term exposures were inconsistent across
studies, particularly across study locations. Since then, the number of studies examining ozone exposure
and preterm birth has doubled. All recent studies that examined ozone exposures during the first or
second trimesters reported associations elevated from the null. Effects are more mixed with third trimester
and entire-pregnancy exposure, with both positive and null associations present. As in the 2013 Ozone
ISA, studies of short-term, near-birth exposures generally reported no evidence of association. Data from
current studies of preterm birth are extracted and summarized in the evidence inventories (see
Table 7-11).
• One study divided PTB into three categories (severe, moderate, late) and examined the
association with ozone exposure at 4-week intervals. The study authors observed elevated ORs
for late and moderate PTB (but not severe/very PTB; 20-28 weeks) with exposures during
gestation weeks 9-12. For second trimester exposures, they observed elevated ORs across
preterm birth groups for ozone exposures during gestation weeks 17-20, 21-24, and 25-28
(Svmanski et al.. 2016).
• A single study was conducted on premature rupture of the membrane (PROM), including both
preterm and term births and examining exposures at 0 to 4 hours before delivery and across the
entire pregnancy. The association with entire pregnancy exposure was null, however, the
associations with short-term, near-birth hourly exposures were all elevated form the null [OR
range: 1.03 to 1.07; Wallace etal. (2016)1.
• Adjustment for copollutants generally moved effect estimates slightly away from the null (Ha et
al.. 2014; Olsson et al.. 2013; Olsson et al.. 2012).
• There were no apparent differences in effect estimates based on study location or exposure
assessment method used for recent studies.
7.1.3.4.2 Summary
Overall, there is consistent evidence for an association between ozone exposures during early to
midpregnancy with preterm birth in epidemiologic studies. There are no toxicological studies examining
preterm birth in animals or endpoints that are analogous to preterm birth.
7-14
-------
7.1.3.5
Birth Defects
Birth defects are structural and functional abnormalities that can cause physical and intellectual
disability and other health problems; they are a leading cause of infant mortality and developmental
disability in the U.S. Critical periods for birth defect development are generally known, reducing
uncertainty related to timing of exposure, which is an uncertainty common to other birth and pregnancy
outcomes.
7.1.3.5.1 Epidemiologic Evidence of Birth Defects
In the 2013 Ozone ISA, studies of birth defects focused on cardiac and oral defects, showing
inconsistent results, perhaps due to variation in study location, study design, and/or analytic methods. In
this current review, cardiac defects are the only defect phenotype examined by multiple recent studies.
Individual recent studies also report on neurological and limb defects. Data from these studies are
extracted and summarized in the evidence inventories (see Table 7-12).
• For cardiac defects, which are a group of disparate defect phenotypes, associations are mixed,
with both positive and null associations reported across both studies and birth defect types.
• Using the U.S.-based National Birth Defects Prevention Study data, one study reported inverse
ORs with higher levels of ozone exposure for neurological defects [neural tube defects,
anencephaly, and spina bifida; Padulaet al. (2013)1.
• A Taiwan-based study of birth defects of the limbs—including Polydactyly, syndactyly, and limb
reduction—observed mixed effect estimates across exposure windows with increasing ozone
levels (Lin et al.. 2014b).
• In general, when studies look at single and copollutants models, effect estimates were generally
similar (Zhang et al.. 2016).
7.1.3.5.2 Toxicological Evidence of Birth Defects
No animal toxicological studies have been conducted on ozone exposure and birth defects since
the 2013 Ozone ISA.
7.1.3.5.3 Summary
Findings for ozone-associated birth defects are generally inconsistent across epidemiologic
studies, and there are no experimental animal studies on birth defects.
7-15
-------
7.1.3.6
Fetal and Infant Mortality
Fetal mortality encompasses spontaneous abortion (fetal deaths occurring before 20 weeks of
gestation) and miscarriage/stillbirth (after 20 weeks of completed gestation). Infant mortality is a death
occurring in the 1st year of life. In the 2013 Ozone ISA studies of infant mortality provided no evidence
for an association between ozone exposure and infant mortality. In the current review, studies are
primarily of stillbirth, with a U.S.-based study (Ha et al.. 2017b) and an Iranian (Dastoorpoor et al.. 2017)
study including both spontaneous abortion and stillbirth, and another Iranian study examining only
spontaneous abortion (Moridi et al.. 2014). No studies examined infant mortality, but one examined "late
fetal death," that is, less than 24 hours after birth (Arroyo et al.. 2016). Findings are inconsistent across
both short- and long-term exposure periods. Findings for ozone associated fetal and infant mortality are
generally inconsistent across exposure windows in epidemiologic studies, and there are no animal studies.
Data from current studies of fetal and infant mortality are extracted and summarized in the evidence
inventories (see Table 7-13).
7.1.4 Effects of Exposure during Developmental Periods
Pregnancy and infancy are periods of rapid development, and exposures occurring during these
times may have the potential to have long-lasting effects that do not manifest immediately; the
Developmental Origins of Health and Disease (DOHaD) is a theory that early life stressors or
environmental exposures can affect later life health outcomes (Heindel et al.. 2017). There are sensitive
windows of development early in life that have the potential to be reprogrammed and put an individual at
increased risk for future health outcomes across lifestages. This theory, otherwise known as Barker's
hypothesis (Barker and Osmond. 1986). detailed a mismatch between fetal environment (famine) and
adult environment (no famine) that was associated with low birth-weight infants that became adults at
greater risk for heart disease and cardiovascular mortality. The evidence from the ozone literature
indicates that ozone could be an exposure associated with DOHaD.
Researchers have examined several health outcomes in association with ozone exposure during
the periods of development, which are summarized below. Studies on the effects of ozone exposure
during developmental periods are evaluated with their respective causality determinations in the sections
of the ISA for the particular organ system or health effect grouping in which the health effect occurs
(e.g., respiratory, nervous system, and cardiovascular effects), but are also summarized here with a focus
on exposure during developmental periods.
7-16
-------
7.1.4.1
Respiratory Development
Several epidemiologic studies conducted in the U.S., Europe, and Asia report no evidence of an
association between long-term exposure to ozone during developmental periods {in utero or early life)
and asthma or allergy (Appendix 3. Section 3.2.4.1). A notable exception is Tetreault et al. (2016) who
reported an increase in asthma incidence among children with increasing summertime average ozone
concentrations. Experimental animal studies provide support for the effect of long-term exposure to ozone
on the development of asthma (Section 3.2.4.1.2) and on lung function development (Section 3.2.4.2.2).
Briefly, studies reviewed in the 2013 Ozone ISA demonstrated that cyclic challenge of infant rhesus
monkeys to an allergen and ozone during the postnatal period compromised airway growth and
development and resulted in changes that favor allergic airway responses and persistent effects on the
immune system. Ozone-exposure-induced nasal lesions were demonstrated in infant monkeys, and
maternal exposure to ozone during gestation resulted in changes related to immune function and allergic
lung disease in the respiratory tract of offspring mice. Recent studies in infant monkeys demonstrated
airway smooth-muscle hyperreactivity, an enhanced allergic phenotype, priming of responses to oxidant
stress, increased serotonin-positive airway cells, and immunomodulation. Recent studies in rodents
demonstrated impaired airway growth and altered airway sensory nerve innervation as a result of
postnatal ozone exposure. Another set of recent studies in infant monkeys demonstrated impaired alveolar
morphogenesis resulting from postnatal ozone exposure. Injury, inflammation, and oxidative stress were
also reported in ozone-exposed neonatal rodents.
7.1.4.2 Neurodevelopment
Effects on laterality, brain morphology, neurobehavioral abnormalities, and sleep aberration were
reported in the 2013 Ozone ISA. Evidence relating to neurodevelopmental effects contributes to the
causality determination in this ISA as detailed in Section 7.2.2 under long-term exposures. Briefly, there
is some epidemiologic evidence to suggest that prenatal or early life exposure to ozone may be associated
with autism (see Section 7.2.2.5.IV The current toxicological data were focused on effects in the
peripheral nervous system, showing decreased neuroproliferation (see Section 7.2.2.5.2).
7.1.4.3 Cardiovascular Development
Evidence from studies of ozone exposure and effects on the cardiovascular system that contribute
to the causality determination is characterized in Appendix 4. Briefly, three studies of exposures during
developmental periods, all based in the U.S., reported mixed effects across outcomes studied. One study
reported changes in newborn blood pressure with ozone exposure during pregnancy (van Rossem et al..
2015). while another reported no associations with blood pressure in kindergarten or first-grade students
(Breton et al.. 2016). The kindergarten or first-grade students also showed no changes in carotid artery
7-17
-------
intima-media thickness with ozone exposure during pregnancy, while a study of college-aged students
reported increased carotid artery intima-media thickness with exposures at 0-5 years of age (Breton ct al..
2012). An animal toxicological study in pregnant dams showed altered uterine artery vascularity and
resistance during pregnancy (Miller et al.. 2017) and is covered in more detail in Section 7.1.3.3.2.
7.1.5 Summary and Causality Determinations
The 2013 Ozone ISA (U.S. EPA. 2013) concluded that the evidence was suggestive of a causal
association between ozone exposure and reproductive and developmental outcomes. The strongest
evidence supporting the causality determination from the 2013 Ozone ISA came from studies of sperm
quality and birth weight (U.S. EPA. 2013). Current evidence continues to support these conclusions.
There is also new evidence supporting effects on preterm birth with exposures to ozone, particularly in
the first and second trimesters.
In the current ISA, separate conclusions are made for (1) male and female reproduction and
fertility and (2) pregnancy and birth outcomes because they are likely to have different etiologies and
critical exposure windows over different lifestages. All available evidence examining the relationship
between exposure to ozone and both of these groups of reproductive effects was evaluated using the
framework described in the Preamble to the ISAs (U.S. EPA. 2015). As noted previously, studies
examining the effect of exposure during developmental periods are summarized in Section 7.1.4
(Table 7-14 and Table 7-17) but contribute to organ-system-specific causality determinations in
Appendix 3. Appendix 4. Appendix 5. and Section 7.1.4.
Overall the evidence is suggestive of, but not sufficient to infer, a causal relationship
between ozone exposure and male and female reproduction and fertility. The key evidence as it
relates to the causal framework is summarized in Table 7-1. This conclusion is supported by consistent
evidence across epidemiologic studies, including studies in the 2013 Ozone ISA, showing decrements in
sperm count and concentration (Farhat et al.. 2016; Hansen et al.. 2010; Sokol et al.. 2006). and by a study
showing changes in rodent testicular morphology and spermatogenesis (Jedlinska-Krakowska et al..
2006a). Uncertainties that contribute to the determination include lack of evaluation of copollutant
confounding or multiple potential sensitive windows of exposure, and the generally small sample size of
studies in human subjects.
7-18
-------
Table 7-1 Summary of evidence that is suggestive of, but not sufficient to infer,
a causal relationship between ozone exposure and male and female
reproduction.
Rationale for Causality
Determination3
Key Evidence13
Key References'3
Ozone
Concentrations
Associated with
Effects0
Consistent evidence from
a limited number of
studies of male
reproduction
Decrements on sperm count and
concentration
Farhatetal. (2016)
-42 ppb
Sokol et al. (2006)
-21.68 ppb
Hansen et al. (2010)
-30.8 ppb
Coherence provided by
study in rodents
Changes to testicular morphology and
spermatogenesis demonstrated in
toxicological study
Jedlinska-Krakowska et
al. (2006a)
0.5 ppm
Lack of copollutant
models contributes to
uncertainty
No epidemiologic studies evaluate
potential copollutant confounding using
copollutant models
Limited study sizes
Observed effects are from smaller
studies on limited number of
individuals
Farhatetal. (2016)
Sokol et al. (2006)
Lack of information of
specific timing of
exposures
All studies use 0-90 days before
sampling exposure window, only one
examines shorter periods within this
window
Hansen et al. (2010)
aBased on aspects considered in judgments of causality and weight of evidence in causal framework in Table I and Table II of the
Preamble to the ISAs (U.S. EPA. 2015V
bDescribes the key evidence and references supporting or contradicting and contributing most heavily to causality determination
and, where applicable, to uncertainties or inconsistencies. References to earlier sections indicate where the full body of evidence
is described.
°Describes the ozone concentrations with which the evidence is substantiated.
Overall the evidence is suggestive of, but not sufficient to infer, a causal relationship
between ozone exposure and pregnancy and birth outcomes. The key evidence as it relates to the
causal framework is summarized in Table 7-2. There are several well-designed, well-conducted studies
that indicate an association between ozone and poorer birth outcomes, particularly for outcomes of birth
weight and preterm birth. In particular, studies of preterm birth that examine exposures in the first and
second trimesters show fairly consistent positive associations (increased ozone exposures associated with
increased odds of preterm birth). In addition, some animal toxicological studies demonstrate decreased
birth weight and changes in uterine blood flow. Studies of birth weight and preterm birth did not
generally adjust for potential copollutant confounding, although studies that did showed limited impacts.
There is also inconsistency across exposure windows for associations with continuous birth weight, and
the magnitude of effect estimates varies.
7-19
-------
Table 7-2 Summary of evidence that is suggestive of, but not sufficient to infer,
a causal relationship between ozone exposure and pregnancy and
birth outcomes.
Rationale for Causality
Determination3
Key Evidence13
Key References'3
Ozone
Concentrations
Associated with
Effects0
Multiple epidemiologic
studies of birth weight
contribute to the
evidence
Positive associations observed in
multiple studies, but there is variability
in timing of exposures and magnitude
of effects
Section 7.4.1
Mean
concentrations
across studies:
4-43 ppb
Multiple epidemiologic
studies and preterm birth
contribute to the
evidence
Positive associations from many
studies that examine exposure
windows in the first and second
trimesters, but magnitude of effects
differ across studies
Section 7.4.1
Mean
concentrations
across studies:
16-51 ppb
Coherence provided by
limited toxicologic
evidence of ozone on
fetal growth and birth
weight
Decreased pup birth and fetal weights
Increased uterine artery blood flow
resistance
Haro and Paz (1993)
Sharkhuu et al. (20111
Miller etal. (2017)
Uncertainty remains due
to limited assessment of
copollutant confounding
Few studies adjust for potential
confounding by NO2 and PM2.5.
Copollutant adjustment generally did
not change observed effect estimates
in studies of preterm birth
Section 7.4.1
Uncertainty remains
regarding information on
the specific timing of
exposures for continuous
birth weight
Several potentially sensitive windows
are examined, including entire
pregnancy, and each trimester, along
with others. However, decrements in
birth weight are not consistently
associated across exposure windows
Section 7.4.1
aBased on aspects considered in judgments of causality and weight of evidence in causal framework in Table I and Table II of the
Preamble to the ISAs (U.S. EPA. 2015).
bDescribes the key evidence and references, supporting or contradicting or contributing most heavily to causality determination
and, where applicable, to uncertainties or inconsistencies. References to earlier sections indicate where full body of evidence is
described.
°Describes the ozone concentrations with which the evidence is substantiated.
7-20
-------
7.2 Nervous System Effects
7.2.1 Short-Term Ozone Exposure
7.2.1.1 Introduction, Summary from the 2013 Ozone ISA, and Scope for Current
Review
This section evaluates the scientific evidence related to the potential effects of short-term
exposure (i.e., on the order of minutes to weeks) to ozone on the nervous system. The 2013 Ozone ISA
(U.S. EPA. 2013) determined that the available evidence was suggestive of a causal relationship between
short-term exposure to ozone and effects on the central nervous system (CNS). This conclusion was based
on the "strong" toxicological evidence linking short-term exposure to ozone to effects on the brain and
behavior of experimental animals. Specifically, short-term exposure was associated with several effects
on CNS structure and function, with several studies indicating the potential for neurodegenerative effects
similar to Alzheimer's or Parkinson's diseases in a rat model. Functional deficits in tasks of learning and
memory, and decreased motor activity were correlated with biochemical and morphological changes in
regions that are known to be affected by these diseases, including the hippocampus, striatum, and
substantia nigra. A study also reported perturbation of sleeping patterns in rodents. Other CNS regions
affected included the olfactory bulb and the frontal/prefrontal cortex. Effects of ozone in the CNS were
strongly correlated with increased markers of oxidative stress and inflammation, including lipid
peroxidation and microglial activation. There was also limited evidence indicating a role of ozone in
modulating neuroendocrine function. Short-term ozone exposure had mixed effects on thyroid hormones,
with one study reporting increased serum T3 and another reporting decreases in both T3 and T4.
Corticosterone levels were also increased in one study, suggesting a stress response. Epidemiologic
studies of short-term exposure to ozone and nervous system effects were lacking in the 2013 Ozone ISA.
The nervous system effects reviewed in this Appendix include brain inflammation and
morphology (Section 7.2.1.3): cognitive and behavioral effects, including mood disorders,
(Section 7.2.1.4): neuroendocrine effects (Section 7.2.1.5): and hospital admission and emergency
department visits (Section 7.2.1.6) for diseases of the nervous system, which are generally defined by
International Classification of Diseases (ICD) codes (i.e., ICD-9 codes 290-319 or 320-359 and ICD-10
codes F1-F99 or G00-G99). The subsections below evaluate the scientific evidence relating short-term
ozone exposure to nervous system effects. These sections focus on studies published since the completion
of the 2013 Ozone ISA. There are a limited number of recent epidemiologic studies examining the effects
of short-term ozone exposure on the nervous system. Multiple recent animal toxicological studies support
conclusions from the 2013 Ozone ISA (U.S. EPA. 2013).
7-21
-------
7.2.1.1.1 Population, Exposure, Comparison, Outcome, and Study Design (PECOS) Tool
The scope of this section is defined by a scoping tool that generally describes the relevant
Population, Exposure, Comparison, Outcome, and Study Design (PECOS). The PECOS tool defines the
parameters and provides a framework to help identify the relevant literature to inform the ISA. Because
the 2013 Ozone ISA concluded that there was evidence to suggest a causal relationship between
short-term ozone exposure and nervous system effects, studies are evaluated regardless of specific study
location. The studies evaluated and subsequently discussed within this section were identified using the
following PECOS tools:
Experimental Studies:
• Population: Study populations of any controlled human exposure or animal toxicological study of
mammals at any lifestage
• Exposure: Short-term (in the order of minutes to weeks) inhalation exposure to relevant ozone
concentrations (i.e., <0.4 ppm for humans, <2 ppm for other mammals)
• Comparison: Human subjects that serve as their own controls with an appropriate washout period
or when comparison to a reference population exposed to lower levels is available, or, in
toxicological studies of mammals, an appropriate comparison group that is exposed to a negative
control (i.e., clean air or filtered air control)
• Outcome: Nervous system effects
Study Design: Controlled human exposure (i.e., chamber) studies; in vivo acute, subacute or
repeated-dose toxicity studies in mammals
Epidemiologic Studies:
• Population: Any population, including populations or lifestages that might be at increased risk
• Exposure: Short-term ambient concentration of ozone
• Comparison: Per unit increase (in ppb)
• Outcome: Change in risk (incidence/prevalence) of a nervous system effect
• Study Design: Epidemiologic studies consisting of panel, case-crossover, time-series studies, and
case-control studies; cross-sectional studies with appropriate timing of exposure for the health
endpoint of interest
7.2.1.2 Biological Plausibility
This section describes biological pathways that potentially underlie nervous system effects
resulting from short-term exposure to ozone. Biological plausibility is depicted via the proposed pathways
as a continuum of upstream events, connected by arrows, that may lead to downstream events
(Figure 7-3). This discussion of "how" exposure to ozone may lead to effects on the nervous system
contributes to an understanding of the biological plausibility of the health effects evaluated in this
7-22
-------
Appendix. The biological plausibility for ozone-induced effects on the nervous system is supported by
evidence from the 2013 Ozone ISA and by recent evidence. Note that the structure of the biological
plausibility sections and the role of biological plausibility in contributing to the weight-of-evidence
analysis used in the 2020 Ozone ISA are discussed in Section IS.4.2.
Short-Term
1
r i
Modulation of the
Autonomic Nervous
System
1
¦
N euro inflammation:
Whole Brain
Olfactory Bulb
Cerebral Cortex
Cerebellum
Hippocampus
Altered
Neurotransmitter Levels
(serotonin, dopamine,
acetylcholine)
Sleep Disturbances
Symptoms
Mental Health
Hospitalizations
Note: The boxes above represent the effects for which there is experimental or epidemiologic evidence related to ozone exposure,
and the arrows indicate a proposed relationship between those effects. Solid arrows denote evidence of essentiality as provided, for
example, by an inhibitor of the pathway or a genetic knockout model used in an experimental study involving ozone exposure.
Shading around multiple boxes is used to denote a grouping of these effects. Arrows may connect individual boxes, groupings of
boxes, and individual boxes within groupings of boxes. Progression of effects is generally depicted from left to right and color-coded
(white, exposure; green, initial effect; blue, intermediate effect; orange, effect at the population level or a key clinical effect). Here,
population level effects generally reflect results of epidemiologic studies. When there are gaps in the evidence, there are
complementary gaps in the figure and the accompanying text below. The structure of the biological plausibility sections and the role
of biological plausibility in contributing to the weight-of-evidence analysis used in the 2020 Ozone ISA are discussed in
Section IS.4.2.
Figure 7-3 Potential biological pathways for nervous system effects
following short-term exposure to ozone.
Two primary pathways have been identified by which short-term exposure to ozone is thought to
affect the nervous system. In the first pathway, pulmonary inflammation is the initial event that leads to
downstream effects on the nervous system, whereas activation of sensory neurons in the lung are the
initiating event in the second pathway (Figure 7-3; solid lines). The majority of currently available studies
support the first pathway in which proinflammatory responses in the central nervous system are initiated
indirectly through respiratory and systemic inflammation. In the lung, ozone reacts with the respiratory
epithelial cell lining fluid leading to local and systemic inflammatory responses. The central nervous
7-23
-------
system is affected when circulating inflammatory cytokines and reactive oxygen species (ROS) present in
the bloodstream reach the brain. These can infiltrate the blood-brain barrier or initiate signaling
mechanisms that trigger neuroinflammation.
Numerous proinflammatory and oxidative stress responses have been reported in the brain
following short-term ozone exposure (see Section 7.2.1.3; Table 7-23). These responses include changes
in gene expression, microglial activation, lipid/protein oxidation, and mitochondrial dysfunction in animal
models. Although these effects have been observed throughout the brain, the region's most commonly
reported to be affected include the hippocampus and cerebral cortex, striatum, and olfactory bulb.
Inflammation and oxidative stress in these brain regions are associated with downstream effects,
including altered neurotransmitter levels, structural changes to the blood-brain barrier, cognitive and
behavioral changes and sleep disturbances. These effects may be drivers of depressive symptoms and
mental health hospitalizations (see Section 7.2.1.4 and Section 7.2.1.5; Table 7-24 and Table 7-25). A
single study reported accumulation of (3-amyloid proteins, a strong predictor of Alzheimer's disease in
humans, in aged mice after a short-term exposure (Tyler etal.. 20 IS); these results are also relevant to
long-term exposure and, therefore, will be discussed further in Section 7.2.2.4.
In addition to the inflammation pathway, some data suggest that activation of sensory nerves in
the lung is another mechanism by which ozone can elicit nervous system effects. Irritant effects of ozone
can modulate autonomic nervous system function and are associated with several cardiovascular and
respiratory effects (see Appendix 3). In the lung, vagal nerve stimulation by irritants, including ozone,
stimulates the release of acetylcholine which binds to both the M2 and M3 acetylcholine receptors. These
two receptors have opposing functions in the airways: M3 receptors stimulate smooth muscle contraction,
while M2 receptors inhibit contraction by limiting further release of acetylcholine. M2 receptor activation
is also affected by p38 and Jnk/MAPk, which suppress M2 activation and signaling. In guinea pigs, ozone
exposure increased airway responsiveness, but these effects were abolished when animals were
administered p38 and Jnk/MAPk inhibitors (Vcrhcin et al.. 2013). providing direct evidence of the role of
ozone in autonomic nervous system modulation (Figure 7-3; solid lines). Activation of pulmonary
sensory nerves has also been shown to modulate the sympathetic nervous system triggering the
neuroendocrine stress response and wide-ranging effects on the body, including systemic and
neuroinflammation rFigure 7-3; solid lines; Snow et al. (2018); Kodavanti (2016)1. Much of the recent
research has focused on outcomes related to metabolic function; therefore, this pathway and the potential
impacts are discussed in greater detail in Appendix 5. Note that there is some evidence to indicate that
these pathways may not be entirely independent of one another.
The proposed pathways described here provide biological plausibility for evidence of cognitive
and behavioral effects and sleep disturbances in association with short-term exposure to ozone. These
pathways will be used to inform a causality determination.
7-24
-------
7.2.1.3
Brain Inflammation and Morphology
7.2.1.3.1 Toxicological Studies
In the 2013 Ozone ISA, short-term ozone exposure resulted in increases in markers of oxidative
stress and inflammatory responses (U.S. EPA. 2013). These effects were observed in many regions of the
brain, including the olfactory bulbs, striatum, cortex, substantia nigra, and cerebellum and were associated
with changes in neuronal morphology, increased apoptosis, and decreased numbers of dopaminergic
neurons in the substantia nigra.
Recent studies (see Table 7-23) support the results summarized in the 2013 Ozone ISA, showing
increases in inflammatory responses and markers of oxidative stress in various regions of the brain. Most
studies evaluated a single concentration of ozone with exposure durations ranging from hours (single
exposure) to >15 days depending on the study. While the exposure concentration used in most studies was
0.25 ppm, some studies used higher concentrations [i.e., 0.3 ppm Mokoena et al. (2011). 1 ppm Tyler et
al. (2018); Mumaw et al. (2016); Gonzalez-Guevara et al. (2014). and 2 ppm Chounlamountrv et al.
(2015)1. In studies with multiple time points, the magnitude or severity of effects generally increased with
exposure duration. Several studies evaluated both short- and long-term exposures. Cellular markers of
oxidative stress were generally seen at the earlier (i.e., short-term) time points; effects on apoptosis/cell
counts were primarily observed at the later time points (i.e., long term).
• Increased brain inflammation and oxidative stress was commonly reported following short-term
ozone exposure in rodents (Tyler etal.. 2018; Mumaw et al.. 2016; Mokoena et al.. 2015; Rivas-
Arancibia et al.. 2015; Gomez-Crisostomo et al.. 2014; Gonzalez-Guevara et al.. 2014; Pinto-
Almazan et al.. 2014; Rodriguez-Martinez et al.. 2013; Mokoena et al.. 2011).
• Inflammation and oxidative stress were associated with increased mitochondrial damage (Rivas-
Arancibia et al.. 2015; Gomez-Crisostomo et al.. 2014; Rodriguez-Martinez et al.. 2013).
• Most of these data were generated in adult male rats (Mokoena et al.. 2015; Rivas-Arancibia et
al.. 2015; Gomez-Crisostomo et al.. 2014; Gonzalez-Guevara et al.. 2014; Pinto-Almazan et al..
2014; Rodriguez-Martinez et al.. 2013; Mokoena et al.. 2011). although Mumaw et al. (2016) did
report similar effects in both male and female CD-/- mice (pulmonary immune function
impaired).
• Some evidence suggests that aged populations may be more susceptible to ozone-induced
inflammation in the brain. One study evaluated the effects of ozone in both adult mice and those
termed aged by the investigators, and although there was a clear main effect of ozone exposure,
inflammatory outcomes were more pronounced in the aged animals (Tyler et al.. 2018).
• Brain inflammation and oxidative stress were largely observed in the hippocampus (Tyler et al..
2018; Mokoena et al.. 2015; Gomez-Crisostomo et al.. 2014; Pinto-Almazan et al.. 2014;
Rodriguez-Martinez et al.. 2013) and cerebral cortex (Tyler etal.. 2018; Mumaw et al.. 2016;
Mokoena et al.. 2015; Gonzalez-Guevara et al.. 2014; Mokoena et al.. 2011). with more limited
data for other regions of the brain (Tyler et al.. 2018; Rivas-Arancibia et al.. 2015).
7-25
-------
• There is some evidence in rats to suggest that ozone exposure may affect glial morphology and
blood-brain barrier permeability. Changes in glial morphology in the nucleus tractus solitarius
were reported following a 24-hour continuous ozone exposure, with treated animals showing
increased glial wrapping of synapses. The overall increase in glial coverage was driven by a
decrease in the proportion of synapses with no glial coverage. There were no changes in
expression of proteins associated with astrocyte activation (Chounlamountrv et al.. 2015). In
contrast, adult and aged mice exposed to ozone showed effects on blood brain barrier
permeability, resulting in increased infiltration of circulatory inflammatory cells and structural
changes in the microglia. Notably, this effect was only statistically significant in aged animals.
These effects were observed in the cortex, dentate gyrus, hippocampus, and hypothalamus: brain
regions that are known to have increased permeability of the blood-brain barrier or high
sensitivity of the cells to toxic insult (Tyler et al.. 2018).
• The effect of ozone exposure on (3-amyloid accumulation and structure was investigated in
several studies. Tyler etal. (2018) found that short-term ozone exposure increased (3-amyloid
formation in aged mice, but several other studies reported no effect in adult rats at the 7 or 15 day
time points (Rivas-Arancibia et al.. 2017; Fernando Hernandez-Zimbron and Rivas-Arancibia.
2016; Hernandez-Zimbron and Rivas-Arancibia. 2015). (3-Amyloid accumulation is strongly
associated with Alzheimer's disease in humans; therefore, these data are discussed further in the
long-term exposure section (see Section 7.2.2).
7.2.1.3.2 Summary
Increased brain inflammation, oxidative stress, and associated mitochondrial damage was
consistently reported following short-term ozone exposure in experimental animal studies. These effects
were largely observed in the hippocampus, with more limited evidence of effects in other brain regions.
Results of studies examining (3-amyloid accumulation and structure were mixed, and there was some
evidence of effects on blood-brain barrier permeability in aged mice.
7.2.1.4 Cognitive and Behavioral Effects
7.2.1.4.1 Epidemiologic Studies
No epidemiologic studies of short-term ozone exposure and its effects on cognitive and
behavioral effects were reviewed in the 2013 Ozone ISA. In a recent study, Lim et al. (2012) examined
older adults in South Korea during a 3-year, follow-up study using the Korean Geriatric Depression
Scale-Short Form (SGDS-K). An increase in SGDS-K score, indicating increased depressive symptoms,
largely driven by the emotional component of the test, was associated with 3-day moving avg ozone
concentration (see Table 7-18).
7-26
-------
7.2.1.4.2
Toxicological Studies
In the 2013 Ozone ISA, short-term exposure to ozone was associated with changes in behavior in
rodents, including decreased motor activity, impaired performance on learning and memory tasks, and
altered sleep-wake cycles. In general, these effects were more pronounced with increasing exposure
durations. Effects on sleep-wake cycles were associated with decreases in acetylcholine levels in the
medial preoptic area, a region of the brain that regulates sleep.
Several recent studies (see Table 7-24) have reported cognitive and behavioral changes in rodent
models following short-term exposure to ozone. Observed effects on cognition and behavior included
increases in depressive-like behaviors in a rodent model of depression and anxiety, decreases in
performance on learning and memory tasks, and declines in motor activity. Effects on neurotransmitter
levels were also reported.
• One study reported increases in depression and anxiety behaviors in Flinders sensitive line (FSL)
rats, a rodent model of depression (Mokocna et al.. 2015). The animals were exposed to 0.3 ppm
(4 hours/day for 15 days) in this study. In the forced swimming test, ozone enhanced
depressive-like behavior as indicated by significantly more time spent immobile and less time
attempting to climb to escape the water. Ozone exposure was also found to significantly decrease
the time spent in the open arms of the elevated plus maze and in the time spent interacting with a
peer in social interaction tests, indicators or increased anxiety.
• Two studies reported deficits in learning and memory following short-term ozone exposure. In
the passive avoidance task, male Wistar rats exposed to ozone concentrations of 0.25 ppm for
>15 days showed decreased latency in both the short-term (10 minutes) and long-term (24 hour)
tests; these results are indicative of impaired learning/memory function. No effects were observed
at the 7-day time point (Pinto-Almazan et al.. 2014). Similar results were reported in FSL rats
exposed to ozone concentrations of 0.3 ppm. Animals and spent significantly less time exploring
a novel object when presented alongside a familiar object. These results suggest that the animals
failed to recognize the familiar object (Mokocna et al.. 2015).
• Three studies evaluated motor activity following short-term ozone exposure. Of these, two
reported statistically significant decreases in motor activity associated with ozone concentrations
of 0.25 ppm (Pinto-Almazan et al.. 2014) and 0.8 ppm (Gordon et al.. 2016). A similar pattern of
behavior was reported by Mokoena et al. (2015): rats exposed to 0.3 ppm ozone showed a slight
decrease in total locomotor activity compared to untreated controls; however, this effect was not
statistically significant. Notably, ozone-related effects on motor activity were observed in male
and female rats exposed to ozone and fed a control diet; however, when animals were fed high-fat
or high-fructose diets, the effects of ozone on motor activity were not detected. Diet alone had no
effect on motor activity (Gordon et al.. 2016) The mechanisms underlying the potentially
ozone-mitigating effects of diet remain unclear.
• Changes in cognitive and behavioral function were supported by associated changes in
neurotransmitter levels. Exposure ozone for 15 days altered neurotransmitter levels in the brains
of FSL rats relative to unexposed controls. Specifically, serotonin levels were reduced in the
frontal cortex and hippocampus and norepinephrine levels were reduced in the hippocampus
(Mokoena et al.. 2015). The neurotransmitter serotonin is believed to play an important role in the
pathophysiology of depression, so these data support the increases in depressive-like behaviors
7-27
-------
described above. Bhoopalan et al. (2013) found decreases in dopamine levels in the striatum after
a single ozone exposure, but these effects were not statistically significant.
7.2.1.4.3 Summary
Experimental animal studies reported decreased motor activity and impaired learning and
memory following short-term exposure to ozone (Gordon et al.. 2016; Mokoena et al.. 2015; Pinto-
Almazan et al.. 2014). There were no epidemiologic studies of cognition or motor function-related effects.
Some of the behavioral effects in animals are supported by data showing effects on neurotransmitter
levels that are associated with these outcomes (see Section 7.2.1.4). A single epidemiologic study
reported an association of short-term ozone exposure with depressive symptoms (Lim et al.. 2012). This
finding was supported by a toxicological study of FSL rats (Mokoena et al.. 2015). Overall, the number of
available studies pertaining to cognitive and behavioral effects is limited.
7.2.1.5 Neuroendocrine Effects
7.2.1.5.1 Toxicological Evidence
In the 2013 Ozone ISA, two studies provided evidence that ozone alters neuroendocrine function,
affecting levels of thyroid hormones and corticosterone following short-term exposure. Since then,
several studies have been published investigating the potential effects of ozone on the HPA axis;
however, most of the data examine outcomes related to metabolic function and are therefore discussed in
detail in Appendix 5.
A recent study (see Table 7-25) evaluated potential neuroendocrine effects of ozone in the
nervous system following a short-term ozone exposure in rats (Thomson et al.. 2013). A 4-hour exposure
to ozone concentrations of 0.4 or 0.8 ppm induced a transient effect on a wide array of genes involved in
antioxidant response, xenobiotic metabolism, inflammation, and endothelial dysfunction. The pattern of
gene responses was largely consistent across several organs, including the brain and pituitary, supporting
systemic effects of neuroendocrine changes. Notably, the effects observed in the present study were
transient, largely disappearing by 24-hour post-exposure; however, chronic exposure could result in
prolonged neuroendocrine modulation. As described previously (see Section 7.2.1.2). ozone likely
modulates HPA axis function by activating the sensory nerves in the lung and thereby altering autonomic
nervous system activity.
7-28
-------
7.2.1.6
Hospital Admissions and Emergency Department Visits
There were no studies of hospital admissions, emergency department (ED), or outpatient visits for
diseases of the nervous system in the 2012 Ozone ISA. Recent studies (see Table 7-19) examining the
association of short-term ozone exposure with hospital admissions, ED visits, or outpatient visits for
diseases of the nervous system or mental health are presented in Figure 7-4. Outcomes that are presented
on the plot and included in this section generally include hospitalizations for International Classification
of Diseases version 9 (ICD-9) codes 290-319 or 320-359 and version 10 (ICD-10) codes F1-F99 or
G00-G99. Several of the studies shown in Figure 7-4 are stratified by season, reporting separate
associations for the warm and cold seasons.
• Some positive associations with hospitalizations for migraine, dementia, and multiple sclerosis
were observed in single studies. Several studies also reported associations of short-term ozone
exposure with mental health hospital admissions or ED visits for conditions such as depression
and panic attack, but the results were not entirely consistent (Figure 7-4).
• Because hospitalizations or ED visits among those with chronic diseases may be related to
comorbid conditions, the extent to which these studies are informative regarding the effect of
short-term ozone exposure on nervous system health is uncertain.
7-29
-------
Study
Cohort
Outcome
Lag
Mean
I Warm Season
I
f Jeanjean etal. 2018
Strasbourg, France
HA: multiple sclerosis relapse
0-3
44.3
'~
I
"|"Guo etal. 2018 1
Guangzhou, China
ED visit: neurologic disease
0-2
49.8
I
4
i
t Chiu al. 2015
Taipei, Taiwan
Outpatient visit: migrane
0-2
24.6
i
i
t Szyszkowicz et al. 2016
9 cities, Canada
ED visit: depression
0, Male
22.5-29.2
i
*
i
ED visit: depression
0, Female
22.5-29.2
i
i
f Oudin et al. 2018
Gothenberg, Sweden
ED visit: pychiatric emergencies
0
29.5
i
%-
i
i
1 Cold Season
i
f Jeanjean etal. 2018
Strasbourg, France
HA: multiple sclerosis relapse
0-3
19.0
i
i
f Guo etal.2018*
Guangzhou, China
ED visit: neurologic disease
0-2
49.8
i
¦
i
j-Chiu al. 2015
Taipei, Taiwan
Outpatient visit: migrane
0-2
24.6
i
i
tOudin etal. 2018
Gothenberg, Sweden
ED visit: pychiatric emergencies
0
21.1
i
*
i
l
i All Year
I
t Linares et al. 2017
Madrid, Spain
HA: dementia related
5d
18.2
i
1 o
i
fXu etal. 2016
Xifan, China
Outpatient visit: epilepsy
0
51.0
i
0
1
tChen etal. 2018*
Shanghai, China
HA: mental disorder
0-1
51.0
i
i
tCho et al. 2015
Seoul, South Korea
ED visit: panic attack
0-3
18.0
l
1
I
06 08 1 1.2 1.4 1.6
Relative Risk (95% CI)
Note: Relative risks are standardized to a 15 or 20 ppb increase ozone for 24-hour avg and 8-hour max* metrics, respectively. Lag
times reported in days, unless otherwise noted. Diamonds indicate effect estimates for the warm season, squares indicate effect
estimates for the cold season and circles indicate year-round effect estimates.
tStudies in red are recent studies.
Figure 7-4 Results of studies of short-term ozone exposure and hospital
admissions or emergency department visits for diseases of the
nervous system or mental health.
7.2.1.7 Relevant Issues for Interpreting the Epidemiologic Evidence
Evaluations of copollutant confounding and the effect of season were limited to studies of
hospital admission, ED visits, or outpatient visits. As discussed in Section 7.2.1.6. the extent to which
such studies inform the effect of short-term ozone exposure on nervous system effects is uncertain.
Further, the limited evidence did not reveal a clear pattern of association. For example, Chiu and Yang
(2015) reported an association with hospitalization for migraine that was larger in the cold season and
7-30
-------
persisted after adjustment for PMio, SO2, NO2, or CO. The inverse association between short-term
exposure to ozone and epilepsy outpatient visits observed by Xu et al. (2016) remained after adjustment
for NO2. The small increase in psychiatric ED visits observed by Oudin et al. (2018) was diminished after
adjustment for PM10 and NO2. Associations with hospitalization or ED visits for multiple sclerosis or
mental health were observed in the warm season (Jcanjcan et al.. 2018; Oudin et al.. 2018; Szvszkowicz
et al.. 2016) when ozone concentrations are higher.
7.2.1.8 Summary and Causality Determination
The 2013 Ozone ISA (U.S. EPA. 2013) concluded that the evidence was suggestive of a causal
relationship between short-term ozone exposure and nervous system effects. The strongest evidence
supporting this causality determination came from experimental animal studies of CNS structure and
function. No epidemiologic studies of short-term ozone exposure and nervous system effects were
reviewed in the 2013 Ozone ISA, and the epidemiologic evidence remains limited. Current evidence
continues to support conclusions for related endpoints, including brain inflammation and changes in brain
morphology, oxidative stress, and neurotransmitter levels.
All available evidence examining the relationship between exposure to ozone and nervous system
effects was evaluated using the framework described in the Preamble to the ISAs (U.S. EPA. 2015) and
summarized in Table 7-3. Most of the recent experimental animal studies indicate that short-term
exposure to ozone induces oxidative stress and inflammation in the central nervous system (see
Section 7.2.1.3 and Table 7-23). In some cases, these effects are associated with changes in brain
morphology and effects on neurotransmitters. In some instances, the effects of short-term ozone exposure
on the nervous system were exacerbated in aged animals. Adolescent and aged animals showed
differences in the patterns of oxidative stress, with young animals showing greater magnitude of effect in
the striatum and aged animals showing higher levels in the hippocampus (Tyler et al.. 2018).
Epidemiologic studies of effects from short-term ozone exposure were lacking in the previous
review. Recent evidence is limited to an association of short-term ozone exposure with depressive
symptoms (Lim et al.. 2012) and several studies of hospital admissions or ED visits for a range of
conditions coded according the International Classification of Disease system as nervous system diseases
or mental disorders (e.g., multiple sclerosis, Alzheimer's disease, Parkinson's disease, depression,
psychiatric disorders). The findings of Lim et al. (2012) are coherent with experimental animal data
showing depression-like behaviors in rodents (Mokoena et al.. 2015). Biological plausibility of these
effects is supported by multiple toxicological studies showing inflammation and morphological changes
in the brain following short-term ozone exposure (see Section 7.2.1.2). As noted in Section 7.2.1.6. these
hospital admission and ED visit studies provide limited information regarding the effect of short-term
ozone exposures on the nervous system because the extent to which people are treated for comorbid
conditions may not be discernable.
7-31
-------
Overall, the evidence is suggestive of, but not sufficient to infer, a causal relationship
between short-term exposure to ozone and nervous system effects. This conclusion remains based
largely on multiple toxicological studies demonstrating the effect of short-term exposure to ozone on the
brain.
Table 7-3 Summary of evidence for a relationship between short-term ozone
exposure and nervous system effects that is suggestive of, but not
sufficient to infer, a causal relationship.
Ozone
Concentrations
Rationale for Causality Associated with
Determination3 Key Evidence13 Key References'3 Effects0
Limited epidemiologic An increase in depressive Lim et al. (2012) Mean: 48.1 ppb
evidence symptoms reported in single study section 7 2 16
Relevance of studies of hospital
admissions, ED visits, and
outpatient visits to nervous system
effects is uncertain
Coherence with experimental Study of FSL rats demonstrates Mokoena et al. (2015) 0.3 ppm
animal study enhanced depressive-like Section 7 2 14
symptoms
Single toxicological studies 0.25-0.8 ppm
demonstrate effects on motor
activity and cognition
Multiple toxicological studies Multiple studies show brain Section 7.2.1.3 0.25-2 ppm
generally support effects on inflammation, oxidative stress, and
the brain and provide morphological changes following
biological plausibility short-term ozone exposure
Epidemiologic evidence from Evaluation of copollutant Section 7.2.1.7
copollutant models is lacking confounding limited to studies of
hospital admissions, ED visits, and
outpatient visits, which are subject
to limitations
C-R = concentration-response; N02 = nitrogen dioxide; PM2 5 = particulate matter with a nominal aerodynamic diameter less than
or equal to 2.5 |jm; PM10 = particulate matter with a nominal aerodynamic diameter less than or equal to 10 |jm; ppb = parts per
billion.
aBased on aspects considered in judgments of causality and weight of evidence in causal framework in Table I and Table II of the
Preamble.
bDescribes the key evidence and references contributing most heavily to the causality determination and, where applicable, to
uncertainties or inconsistencies. References to earlier sections indicate where the full body of evidence is described.
°Describes the ozone concentrations with which the evidence is substantiated.
7-32
-------
7.2.2
Long-Term Ozone Exposure
7.2.2.1 Introduction, Summary from the 2013 Ozone ISA, and Scope for Current
Review
This section evaluates the scientific evidence related to the potential effects of long-term
exposure to ozone (i.e., on the order of months to years) on the nervous system. The 2013 Ozone ISA
(U.S. EPA. 2013) determined that the evidence was suggestive of, but not sufficient to infer, a causal
relationship between exposures to ozone and effects on the central nervous system. The evidence is built
on findings from the 2006 Ozone AQCD demonstrating alterations in neurotransmitters, motor activity,
memory, and sleep patterns following short-term exposure to ozone, with the addition of studies that
demonstrated progressive damage in various regions of the brains of rodents in conjunction with altered
behavior following long-term ozone exposure. Specifically, several studies indicating the potential for
neurodegenerative effects similar to Alzheimer's or Parkinson's diseases in a rat model were conducted.
The evidence from epidemiologic studies of long-term exposure to ozone was limited to a single study
reporting cognitive decline in older adults (Chen and Schwartz. 2009).
The nervous system effects reviewed in this Appendix include brain inflammation and
morphology (Section 7.2.2.3); effects on cognition, motor activity, and mood (Section 7.2.2.4); and
neurodevelopmental effects (Section 1.2.2.5). In addition, issues relevant for interpreting the
epidemiologic studies are described in Section 7.2.2.6. The subsections below evaluate the scientific
evidence relating long-term ozone exposure to nervous system effects. These sections focus on studies
published since the completion of the 2013 Ozone ISA. The body of evidence has grown since the 2013
Ozone ISA. A limited number of recent epidemiologic studies examining nervous system effects are
available, with the strongest line of evidence supporting an effect on cognition in adults. Recent
experimental animal studies continue to provide coherence for these effects.
7.2.2.1.1 Population, Exposure, Comparison, Outcome, and Study Design (PECOS)
The scope of this section is defined by scoping statements that generally define the relevant
PECOS. The PECOS statements define the parameters and provide a framework to help identify the
relevant evidence in the literature to inform the ISA. Because the 2013 Ozone ISA concluded that there
was evidence to suggest a causal relationship between long-term ozone exposure and nervous system
effects, the studies are evaluated regardless of specific study location. The studies evaluated and
subsequently discussed within this section were identified using the PECOS statements below:
Experimental Studies:
• Population: Study population from any animal toxicological study of mammals at any lifestage
7-33
-------
• Exposure: Long-term (in the order of months to years) inhalation exposure to relevant ozone
concentrations (i.e., <2 ppm)
• Comparison: Appropriate comparison group exposed to a negative control (i.e., clean air or
filtered-air control)
• Outcome: Nervous system effects
• Study Design: In vivo chronic, subchronic, or repeated-dose toxicity studies in mammals
Epidemiologic Studies:
• Population: Any population, including populations or lifestages that might be at increased risk
• Exposure: Long-term ambient concentration of ozone
• Comparison: Per unit increase (in ppb)
• Outcome: Change in risk (incidence/prevalence) of a nervous system effect
• Study Design: Epidemiologic studies consisting of cohort and case-control studies, time-series,
case-crossover, and cross-sectional studies with appropriate timing of exposure for the health
endpoint of interest
7.2.2.2 Biological Plausibility
This section describes biological pathways that potentially underlie nervous system effects
resulting from long-term and developmental exposure to ozone. Studies that include exposure during the
perinatal period are discussed in the long-term exposure section, regardless of the duration of the
exposure because of the sensitivity of this lifestage to nervous system effects and potential for long-term
health impacts. Biological plausibility is depicted via the proposed pathways as a continuum of upstream
events, connected by arrows, that may lead to downstream events observed in epidemiologic studies
(Figure 7-5). This discussion of "how" exposure to ozone may lead to effects on the nervous system
contributes to an understanding of the biological plausibility of epidemiologic results evaluated later. The
biological plausibility for ozone-induced effects on the nervous system is supported by evidence from the
2013 Ozone ISA and by new evidence. Note that the structure of the biological plausibility sections and
the role of biological plausibility in contributing to the weight-of-evidence analysis used in the 2020
Ozone ISA are discussed in Section IS.4.2.
As discussed in the short-term exposure section (see Section 7.2.1.2). inflammation is also
expected to be an important mechanism driving nervous system effects following long-term ozone
exposure. The first proposed pathway (Figure 7-5) is largely conserved across the short- and long-term
exposure durations, however, there is a stronger link to neurodegenerative outcomes in humans following
long-term exposures. Briefly, inhaled ozone elicits inflammation releasing inflammatory cytokines and
ROS into the bloodstream that trigger systemic inflammation. Proinflammatory markers interact with, and
in some cases infiltrate, the blood-brain barrier initiating neuroinflammation, as indicated by altered gene
expression, increased apoptosis, lipid/protein oxidation, and microglial activation (see Section 7.2.2.3;
7-34
-------
Table 7-26). These effects are associated with changes to nervous system function
(e.g., behavior/cognition, sleep disturbances, neurotransmitter levels) and structure (e.g., blood brain
barrier, |3-amyloid accumulation, morphology) that are associated with neurodegenerative diseases such
as Alzheimer's and Parkinson's diseases, and mood disorders.
Long-Term
Ozone
Exposure
Neuroinflammation in
Developing Animals
Neuro inflammation
in Adult Animals:
Whole Brain
Olfactory Bulb
Cerebral Cortex
Cerebellum
Hippocampus
Altered
Neurotransmitter
Levels
Altered
Neurodevelopmental
Processes
Altered
Neurotransmitter Levels
(Seratonin, Dopamine,
Acetylcholine)
Cognitive and Behavioral
Changes (Learning,
Memory, Motor Activity)
Neurodevelo pmenta I
Disorders
Neurodegenerative
Diseases
Mood Disorders
J
J
Note: The boxes above represent the effects for which there is experimental or epidemiologic evidence related to ozone exposure,
and the arrows indicate a proposed relationship between those effects. Solid arrows denote evidence of essentiality as provided, for
example, by an inhibitor of the pathway or a genetic knockout model used in an experimental study involving ozone exposure.
Shading around multiple boxes is used to denote a grouping of these effects. Arrows may connect individual boxes, groupings of
boxes, and individual boxes within groupings of boxes. Progression of effects is generally depicted from left to right and color-coded
(white, exposure; green, initial effect; blue, intermediate effect; orange, effect at the population level or a key clinical effect). Here,
population level effects generally reflect results of epidemiologic studies. When there are gaps in the evidence, there are
complementary gaps in the figure and the accompanying text below. The structure of the biological plausibility sections and the role
of biological plausibility in contributing to the weight-of-evidence analysis used in the 2020 Ozone ISA are discussed in
Section IS.'4.2.
Figure 7-5 Potential biological pathways for nervous system effects
following long-term exposure to ozone.
In the second pathway (Figure 7-5). adverse nervous system effects have also been reported when
exposure occurs during development. Inflammation is expected to be a critical pathway for
neurodevelopmental effects of ozone on developing offspring, just as it is in adults. In animal models,
respiratory and systemic inflammation, either from direct (i.e., inhalation) or indirect (i.e.. via the dam)
exposure, are expected to elicit neuroinflammation that is associated with altered neurotransmitter levels,
7-35
-------
cognitive and behavioral changes, and altered development of the peripheral nervous system. Together,
these effects may contribute to neurodevelopmental disorders. Neuroinflammation in developing animals
may also be triggered by activation of sensory nerves in the lung, leading to altered neurodevelopment of
the nodose and jugular ganglia that transmit sensory information from the lung to the brain (see
Section 7.2.2.5; Table 7-28).
The pathway(s) described here provide biological plausibility for evidence of neurodegenerative
diseases, mood disorders, and sleep disturbances in adults (see Section 7.2.2.4) and neurodevelopmental
disorders in children (see Section 7.2.2.5) in association with long-term exposure to ozone. These
pathways will be used to inform a causality determination, which is discussed later in the Appendix.
7.2.2.3 Brain Inflammation and Morphology
7.2.2.3.1 Toxicological Studies
In the 2013 Ozone ISA, long-term ozone exposure elicited similar effects on the brain versus
short-term exposure (see Section 7.2.1.3.1). with many studies showing increases of inflammatory and
oxidative-stress responses, elevated cell death, and changes in neuronal morphology in various regions of
the brain. In general, the magnitude and severity of the effects was generally increased with longer
exposure durations; however, some studies found these effects could be mitigated by coexposure with
antioxidants.
As discussed below, the effects of long-term exposure on brain inflammation and morphology
were similar to those described in the short-term exposure section (see Section 7.2.1.1); however, the
magnitude and severity of the effects were generally increased with longer exposure durations. These
effects were observed in multiple brain regions. There is also some evidence to suggest that males may be
more susceptible than females to inflammation and oxidative damage. Study details are provided in
Table 7-26.
• Multiple studies measured elevated levels of oxidative stress and inflammation in the brains of
rats and mice following long-term exposure to ozone (Rodriguez-Martinez et al.. 2016; Akhter et
al.. 2015; Rivas-Arancibia et al.. 2015; Gomez-Crisostomo et al.. 2014; Pinto-Almazan et al..
2014; Rodriguez-Martinez et al.. 2013; Mokoena et al.. 2011). Histological analyses revealed
reduced cell counts and increased apoptosis and oxidative damage in several regions of the brain,
including the hippocampus (Rodriguez-Martinez et al.. 2016; Gomez-Crisostomo et al.. 2014;
Pinto-Almazan et al.. 2014; Rodriguez-Martinez et al.. 2013). frontal cortex (Mokoena et al..
2011). and substantia nigra (Rivas-Arancibia et al.. 2015). These regions were also found to have
damage to the mitochondria and endoplasmic reticulum (Rivas-Arancibia et al.. 2015; Gomez-
Crisostomo et al.. 2014; Rodriguez-Martinez et al.. 2013). Ozone concentrations ranged from
0.25 to 0.8 ppm in these studies.
• Although the majority of the data were generated in male Wistar rats, the study by Akhter etal.
(2015) evaluated the oxidative effects of ozone exposure in both males and females using a
7-36
-------
mouse model of Alzheimer's disease. Ozone-exposed Alzheimer's disease model males exhibited
significantly greater apoptosis in the hippocampus relative to the other experimental groups
(i.e., wild-type males and females + ozone, Alzheimer's disease model males and
females + filtered air, and Alzheimer's disease females + ozone). Male Alzheimer's disease
model mice were found to have significantly lower baseline antioxidant levels than wild-type
animals or Alzheimer's disease model females which may make them more susceptible to
oxidative stressors (Akhter et al.. 2015V
• A series of studies from the same research group reported that long term exposure to ozone
affected (3-amyloid accumulation and regulation in the brain. These effects are strongly associated
with development of Alzheimer's disease. Accumulation of |3-amyloid proteins was increased,
and alterations in several proteins and genes that regulate (3-amyloid formation and degradation
were observed in the hippocampus and cortex of male Wistar rats following exposure to 0.25 ppm
ozone (Rivas-Arancibia et al.. 2017; Fernando Hernandez-Zimbron and Rivas-Arancibia. 2016;
Hernandez-Zimbron and Rivas-Arancibia. 2015). In general, these results showed an
exposure-dependent trend, with the magnitude of effect increasing with exposure duration. Rivas-
Arancibia et al. (2017) also found that ozone exposure induced exposure-dependent changes in
the folding of (3-amyloid proteins in a manner consistent with those observed in (3-amyloid
plaques associated with Alzheimer's disease. In some cases, (3-amyloid was found to be
colocalized with mitochondria (Hernandez-Zimbron and Rivas-Arancibia. 2015) and the
endoplasmic reticulum (Fernando Hernandez-Zimbron and Rivas-Arancibia. 2016). In contrast, a
single study from a different laboratory did not find an effect of ozone exposure on (3-amyloid
accumulation in a transgenic Alzheimer's disease mouse model. These animals were
intermittently exposed to 0.8 ppm ozone for 4 months and while all Alzheimer's disease model
animals showed (3-amyloid accumulation in the hippocampus and cortex, there was no effect of
ozone exposure (Akhter et al.. 2015).
7.2.2.4 Effects on Cognition, Motor Activity, and Mood
The cognitive and behavioral effects measured in the epidemiologic studies reviewed in this
section include scores on the Mini-Mental State Examination (MMSE), which is a questionnaire used to
screen for dementia, and performance on neurobehavioral tests of cognitive function. Depression was
evaluated using self-reported information on depression diagnosis and use of antidepressant medication.
Clinically-diagnosed dementia, including Alzheimer's disease and vascular dementia, and Parkinson's
disease were also examined in a small number of studies. In a few animal toxicological studies, effects on
learning and memory, motor activity, and anxiety were evaluated.
7.2.2.4.1 Epidemiologic Studies
Cognition and Dementia-Related Effects
The 2013 Ozone ISA reported declines on tests of cognitive function measured using
Neurobehavioral Evaluation System-2 (NES2), in a cross-sectional analysis of NHANES III (1988-1991)
data (Chen and Schwartz. 2009). A small number of recent studies examine the effect of long-term
7-37
-------
exposure to ozone with performance on neurobehavioral tests (see Table 7-20). Alzheimer's disease, and
other forms of dementia (see Table 7-21). Overall, the limited number of epidemiologic studies support
an effect of long-term exposure to ozone on reduced cognitive function, but effect estimates reported in
studies of dementia are inconsistent. Examination of copollutants confounding was limited.
• Domain-specific (i.e., executive function) decrements were observed in association with
long-term exposure to ozone in a cross-sectional analysis of older adult women in Los Angeles,
CA (Gatto et al.. 2014). Study participants completed a battery of 14 neurobehavioral tests
designed to measure cognitive decline in middle-aged and older adults. Cleary et al. (2018)
examined the rate of cognitive decline using the MMSE among subjects followed through U.S.
Alzheimer's Disease Centers, reporting an effect of ozone among those who had normal
cognition at baseline.
• A small number of studies of Alzheimer's disease or dementia have reported results that vary in
direction, magnitude, and precision. Chen et al. (2017c) reported a small (relative to the width of
the confidence interval) inverse association with dementia in a population-based cohort study in
Ontario, Canada (HR: 0.97; 95% CI: 0.94, 1.00). In this study, information about residential
history was linked to modeled ozone concentrations and to registry information on
physician-diagnosed dementia (dementia-related ICD codes for hospital admission or three
physician claims). A positive association with confirmed Alzheimer's disease, which remained in
copollutants models adjusted for CO, NO2, and SO2, was observed in a study in Taiwan using the
National Health Insurance Research Database [NHIRD; HR: 1.06; 95% CI: 1.00, 1.12; Jung et al.
(2014)1. In a smaller case-control study conducted in Taiwan, relatively large, imprecise
associations of long-term exposure to ozone with Alzheimer's disease and vascular dementia
were reported [OR: 2.00; 95% CI: 1.14, 3.50 and OR: 2.09; 95% CI: 1.01, 4.33; Wu et al.
(2015)1.
Motor Function-Related Effects
Parkinson's disease is a nervous system disease that affects movement as well as nonmotor
function. It is characterized by loss of dopaminergic neurons in the substantia nigra. There were no
epidemiologic studies of Parkinson's disease reviewed in the 2013 Ozone ISA. Recent epidemiologic
studies (see Table 7-21) conducted in the U.S. and Taiwan report some positive, although imprecise
(i.e., wide confidence intervals), associations. Examination of copollutants confounding was limited.
• Large registry-based prospective studies conducted in Canada and Europe examined the
relationship of ozone exposure with Parkinson's disease. Shin et al. (2018) and Cerza et al. (2018)
reported positive associations (HR: 1.06; 95% CI: 1.02, 1.11 andHR: 1.04; 95% CI: 1.00, 1.11),
respectively, with summer average ozone concentrations. The association reported by Cerza et al.
(2018) remained after adjustment for NO2.
• Kirrane et al. (2015) reported an association between prevalent, self-reported doctor-diagnosed
Parkinson's disease in farmers in North Carolina (OR: 2.60; 95% CI: 0.94, 7.24, 4-year warm
avg) but not in Iowa (OR: 0.46; 95% CI: 0.11, 1.84, 4-year warm-season avg).
• In a nested case-control study of the National Health Insurance Research Database (NHIRD) of
Taiwan, Chen et al. (2017a) reported a positive yet imprecise (i.e., wide confidence intervals)
association between Parkinson's disease and long-term exposure to ozone estimated from
monitors located in areas where the subjects resided (OR: 1.10; 95% CI: 0.74, 1.48). In contrast,
7-38
-------
other researchers using the same database but a quantile-based Bayesian maximum entropy
spatiotemporal model to characterize long-term exposure, Lee et al. (2016) reported a null
association (OR: 1.00; 95% CI: 0.97, 1.03) comparing the highest quartile of exposure to the
lowest quartile (<23.93 ppb).
Mood and Mood Disorders
There were no epidemiologic studies of long-term exposure to ozone and mood disorders in the
2013 Ozone ISA. A prospective cohort study of depression onset among older women enrolled in the
Nurses' Health Study (NHS) is currently available for review (Kioumourtzoglou et al.. 2017). This study
reports an association of long-term exposure to ozone with use of antidepressant medication (HR: 1.08;
95% CI: 1.02, 1.14) but not with self-reported doctor-diagnosed depression (HR: 1.00; 95% CI: 0.92,
1.08; rsee Table 7-201).
7.2.2.4.2 Toxicological Studies
In the 2013 Ozone ISA, toxicological studies showed declines in learning and memory that
increased with the exposure duration. Coexposure with antioxidants was found to have a protective effect,
suggesting that oxidative damage particularly in regions of the brain that play a role in cognition, may
contribute to the observed cognitive decrements. Several recent studies investigated the role of ozone
exposure on cognitive and behavioral effects, including changes in learning and memory, motor activity,
and anxiety, that are associated with neurodegenerative diseases (see Table 7-27). Neurodegenerative
effects of ozone may be driven by increased oxidative stress in inflammatory responses in the central
nervous system leading to changes in brain morphology (e.g., increased apoptosis and reduced neuronal
cell counts) in regions of the brain associated with Alzheimer's and Parkinson's disease.
• Some evidence in animal models suggests that long-term exposure to ozone impairs learning and
memory formation, an important characteristic of Alzheimer's disease. Male Wistar exposed to
0.25 ppm ozone for 30, 60, or 90 days showed decreased latency in both short- (10 minutes) and
long-term (24 hour) passive avoidance tests (Pinto-Almazan et al.. 2014).
• The effects of ozone exposure on motor activity data were also evaluated, but the results were not
entirely consistent. Most of the studies reported decreased activity in rats (Gordon et al.. 2016;
Pinto-Almazan et al.. 2014; Gordon et al.. 2013). which builds on evidence reported in the 2013
Ozone ISA. In contrast, Gordon et al. (2014) reported a statistically significant increase in motor
activity, and Akhter et al. (2015) found no effect in a transgenic model of Alzheimer's disease
following ozone exposure. The variability in these results may be attributable to differences in the
study designs. Akhter et al. (2015) evaluated effects in mice, including a transgenic model of
Alzheimer's disease whereas Gordon et al. (2014) continuously monitored animals' motor
activity in the home cage via a subcutaneous radio transmitter.
• Akhter et al. (2015) also found no ozone-mediated effects on behavior in the elevated plus maze,
a measure of anxiety, in a mouse model of Alzheimer's disease. No other toxicological data
related to mood or mood disorders were available following long-term exposure.
7-39
-------
7.2.2.4.3
Summary
The current section describes and characterizes the epidemiologic and toxicological evidence
relating to the effect of long-term ozone exposure on cognition, motor activity, and mood. Biological
plausibility for the long-term effect of ozone on the nervous system is derived from multiple studies
demonstrating that long-term exposure to ozone can to lead to inflammation and oxidative stress
responses in the brain. Limited epidemiologic evidence reports associations with decrements on tests of
cognitive function that may be associated with neurodegenerative diseases. Toxicological studies provide
coherence for these findings, but the results of epidemiologic studies of Alzheimer's and Parkinson's
disease are not consistent. The animal data do not support an association between long-term ozone
exposure and mood disorders. The epidemiologic evidence is limited to one study reporting an association
with self-reported depression.
7.2.2.5 Neurodevelopmental Effects
In the 2013 Ozone ISA, discussion of the data on neurodevelopmental effects was split across the
short- and long-term exposure sections; however, in the current ISA, these data are only reviewed in the
long-term exposure section. The 2013 Ozone ISA reviewed toxicological evidence for
neurodevelopmental effects of prenatal and early life ozone exposure. Exposure that was limited the
prenatal period resulted in altered gene expression of nerve growth factors, affected regulation of
neurotransmitter levels and altered neuroadaptive responses to stress. Social interaction,
defensive/submissive behavior, and turning preferences were also affected in animals exposed either
prenatally or during both gestation and lactation. Notably, some of these outcomes persisted into
adulthood, suggesting early life exposure can have long lasting impacts on neurological function. There
were no epidemiologic studies of long-term exposure to ozone and neurodevelopmental outcomes. A
recent study by Lin et al. (2014a) examined the effect of prenatal ozone exposure and neurobehavioral
outcomes but reported no evidence of an association. The current evidence base also includes several
epidemiologic studies of autism spectrum disorder (ASD) and toxicological studies that focus on the
peripheral nervous system.
7.2.2.5.1 Epidemiologic Studies
There were no studies of long-term exposure to ozone and autism reviewed in the 2013 Ozone
ISA. Several recent studies conducted in the U.S. and Taiwan are currently available (see Table 7-14.
Table 7-22.). Overall, these studies report positive associations, but associations are imprecise (i.e., wide
confidence intervals) and are not consistently observed across pregnancy periods. In addition, outcome
definitions for autism, which is a heterogenous condition with potentially different etiologies, varied
7-40
-------
across studies. For example, Becerraetal. (2013) and Yolk et al. (2013) included cases of autistic
disorder or full syndrome autism, which are the most severe among the autism spectrum disorders (ASD).
• Becerraetal. (2013) conducted a case-control study of autistic disorder, diagnosed between 3 and
5 years of age, in Los Angeles, CA. Ozone exposure during the entire pregnancy but not
trimester-specific exposures was associated with autistic disorder (OR: 1.05; 95% CI: 1.01, 1.10).
The effect of ozone remained in copollutant models adjusted for NO2 estimated using land use
regression (LUR), PM2 5, or PM10.
• Also in California, among children enrolled in the Childhood Autism Risks from Genetics and the
Environment (CHARGE) study, Yolk et al. (2013) reported small imprecise (relative the width of
the confidence interval) associations of full syndrome autism with ozone concentrations during
the 1st year of life, during the entire pregnancy and with trimester-specific ozone concentrations
(e.g., OR: 1.05; 95% CI: 0.84, 1.31; entire pregnancy). Scores on cognitive and adaptive scales
were not associated with prenatal exposure to ozone among children with ASD among subjects
enrolled in the CHARGE cohort (Kerin et al.. 2017).
• Additional analyses of the CHARGE cohort reported an interaction between ozone exposure and
chromosome copy number variation, indicating a larger risk for the joint effect compared to the
effect of ozone or duplication burden alone (Kim et al.. 2017). but not between ozone exposure
and folic acid (Goodrich et al.. 2017).
• A cohort study in Taiwan reported an association between long-term ozone exposure and ASD
[HR: 1.59; 95% CI: 1.42, 1.78; Jung et al. (2013)1. This association remained after adjusting for
CO, NO, and S02.
7.2.2.5.2 Toxicological Studies
Two studies from the same research group evaluated neurodevelopmental effects in animal
models. Zellneretal. (2011) and Hunter etal. (2011) evaluated the effects of a single 3-hour exposure to
2 ppm ozone during the early postnatal window on lung innervation. As discussed previously, these
studies would normally be considered short-term, but due to the sensitivity of the developmental window
and increased potential for long-term outcomes, they are discussed below.
• In the first study, ozone exposure on PND 5 resulted in a statistically significant decrease in the
total number of neurons and change in the overall pattern of neuroproliferation in the nodose and
jugular ganglia. Whereas controls showed a large increase in the average total and substance
P-reactive neuron counts on PNDs 15 and 21, neuronal counts generally remained consistent in
ozone-treated animals across the four time points (PNDs 10, 15, 21, or 28). Notably, there was
high variability among the controls so that the difference for the total neuron count was only
statistically significant on PND 21. Although neurons from these ganglia innervate the lung to
provide sensory feedback, Zellneretal. (2011) reported no effect of ozone exposure on
pulmonary innervation.
• Hunter etal. (2011) studied the effects of ozone on lung nerve growth factor (NGF) levels and
sensory innervation. Here, coexposure to NGF on PND 6 and ozone on PND 28 elicited a
statistically significant increase in substance P-reactive neurons in the lung and extrapulmonary
smooth muscle. Ozone exposure was also found to increase NFG levels in BAL fluid. PND 6 is
believed to be a critical window for neuronal development in the lung. Statistically significant
7-41
-------
increases in NGF in the short term (24-hour PE) were elicited, potentiating the effects of
subsequent ozone exposures.
7.2.2.5.3 Summary
There is some epidemiologic evidence to suggest that prenatal or early life exposure to ozone
may increase the risk for autism or autism spectrum disorder. There were no experimental animal studies
showing effects in the brain that support the epidemiologic findings on autism. The toxicological
evidence was limited to two studies showing effects on the peripheral NS that indicate potential effects on
development of sensory neurons in the lung.
7.2.2.6 Relevant Issues for Interpreting the Epidemiologic Evidence
7.2.2.6.1 Potential Copollutant Confounding
Overall, only a few studies considered copollutant confounding in the analysis. For instance,
associations observed with autism or ASD persisted after adjustment for CO, NO2 SO2 (Jung et al.. 2013).
PM2 5, and PM10 (Becerra et al.. 2013V The association of ozone with Alzheimer's disease observed by
Jung et al. (2014) persisted in copollutant models adjusted for CO, NO2, and SO2.
7.2.2.7 Summary and Causality Determination
The strongest evidence supporting the causality determination from the 2013 Ozone ISA (U.S.
EPA. 2013) came from animal toxicological studies demonstrating effects on CNS structure and function,
with several studies indicating the potential for neurodegenerative effects similar to Alzheimer's or
Parkinson's diseases in a rat model. The body of evidence has grown since the 2013 Ozone ISA. Recent
epidemiologic studies examining nervous system effects, including cognitive effects, depression,
neurodegenerative disease, and autism, are currently available. Although the epidemiologic evidence
remains limited, the strongest line of evidence supports an effect on cognition in adults and recent
experimental animal studies continue to provide coherence for these effects. All available evidence
examining the relationship between exposure to ozone and nervous system effects was evaluated using
the framework described in the Preamble to the ISAs (U.S. EPA. 2015) and summarized in Table 7-4.
Multiple animal recent toxicological studies report increased markers of oxidative stress and
inflammation, including lipid peroxidation, microglial activation, and cell death following long-term
exposure to ozone. There was some evidence to indicate that both young and mature mice termed aged by
study authors (Tyler et al.. 2018) may have increased sensitivity to ozone exposure. Functional deficits in
tasks of learning and memory and decreased motor activity were correlated with biochemical and
7-42
-------
morphological changes in regions that are known to be affected by neurodegenerative diseases, including
the hippocampus, striatum, and substantia nigra. Other CNS regions affected include the olfactory bulb
and frontal/prefrontal cortex. Epidemiologic studies have reported cognitive decline in association with
long-term ozone exposure. Associations with neurodegenerative disease are not entirely consistent, but
some positive associations are reported. Epidemiologic studies of Parkinson's disease do not consistently
support an association with long-term exposure to ozone, and the findings from toxicological studies of
motor function-related effects are mixed, although loss of dopaminergic neurons in the substantia nigra is
observed in animals.
Effects on neurotransmitter levels, behavior, and cell signaling were identified in animals that
were exposed only during the prenatal period. In some cases, these effects persisted into adulthood.
Adolescent and mice termed aged showed differences in the patterns of oxidative stress, with young
animals showing higher levels in the striatum and older animals showing higher levels in the
hippocampus. Some epidemiologic studies of autism or ASD reported positive associations, but the
biological plausibility of these effects is limited because the toxicological data focused on effects in the
peripheral nervous system. A limited number of studies reported copollutant model results. Data derived
from toxicological studies provide coherence for the observed effects.
Overall, the evidence is suggestive of, but not sufficient to infer, a causal relationship
between long-term ozone exposure and nervous system effects. This is consistent with the conclusion
of the 2013 Ozone ISA (U.S. EPA. 2013). Uncertainties that contribute to the causality determination
include the limited number of epidemiologic studies, the lack of consistency across the available studies
of Alzheimer's and Parkinson's disease, and the limited evaluation of copollutant confounding. In
addition, the evidence supporting the biological plausibility of the associations with autism or ASD in
epidemiologic studies is limited.
Table 7-4 Summary of evidence for a relationship between long-term ozone
exposure and nervous system effects that is suggestive, but not
sufficient to infer, a causal relationship.
Ozone
Rationale for
Concentrations
Causality
Associated
Determination3
Key Evidence13
Key References'3 with Effects0
Limited epidemiologic
evidence is generally
consistent for cognitive
effects but not for
Associations with reduced
cognitive function
consistently observed in a
small number of
epidemiologic studies
Chen and Schwartz (2009)
Gatto etal. (2014)
Clearvetal. (2018)
7-43
-------
Table 7-4 (Continued): Summary of evidence for a relationship between long-term
ozone exposure and nervous system effects that is
suggestive, but not sufficient to infer, a causal
relationship.
Ozone
Rationale for
Concentrations
Causality
Associated
Determination3
Key Evidence13
Key References'3
with Effects0
neurodegenerative
Effect estimates reported in
Chen et al. (2017c)
NR
disease
studies of Alzheimer's
Juna et al. (2014)
92.6 ppb
disease or dementia are
Wu et al. (2015)
NR
imprecise and vary in
magnitude
Associations with
Kirrane et al. (2015)
40.6
Parkinson's disease are not
Chen et al. (2017a)
NR
consistently observed and
Lee et al. (2016)
26.1
generally lack precision
Shin et al. (2018)
49.8
(i.e., wide confidence
intervals)
Coherence across lines
Toxicological studies provide
Section 7.2.2.4.2
0.25-1 ppm
of evidence
coherence for
neurodegenerative disease
in humans, including
Alzheimer's and Parkinson's
disease
Biological plausibility
Multiple studies show brain
Section 7.2.2.3
0.25-0.8 ppm
provided by multiple
inflammation and
toxicological studies
morphological changes
demonstrating effects
following short- and
on the brain
long-term ozone exposure
Limited number of
Associations are imprecise
Section 7.2.2.5.1
epidemiologic studies
(i.e., wide confidence
generally report
intervals) and are not
consistent positive
consistently observed across
associations with
pregnancy periods
autism or ASD
Limited evidence of
Available studies focused on
Section 7.2.2.5.2
2 ppm
coherence and
effects in the peripheral
biological plausibility
nervous system
Uncertainty regarding
Few studies consider
the independent effect
copollutant confounding
of ozone exposure
aBased on aspects considered in judgments of causality and weight of evidence in causal framework in Table I and Table II of the
Preamble.
bDescribes the key evidence and references contributing most heavily to the causality determination and, where applicable, to
uncertainties or inconsistencies. References to earlier sections indicate where the full body of evidence is described.
°Describes the ozone concentrations with which the evidence is substantiated.
7-44
-------
7.3 Cancer
7.3.1 Introduction, Summary from the 2013 Ozone ISA, and Scope for Current
Review
In the 2013 Ozone ISA (U.S. EPA. 2013). the available evidence was inadequate to determine
whether there was a causal relationship between exposure to ambient ozone and cancer. That review
noted that very few epidemiologic and toxicological studies had been published examining ozone as a
carcinogen, but that collectively the results of these studies indicated that ozone may contribute to DNA
damage. The same conclusions are reached in this review: there continue to be relatively few studies
examining the association between ozone and cancer, although some animal toxicological studies have
shown indicators of DNA damage in animals. The evidence published since the 2013 Ozone ISA is
discussed in greater detail below.
7.3.1.1 Population, Exposure, Comparison, Outcome, and Study Design (PECOS)
Tool
The scope of this section is defined by a scoping tool that generally describes the relevant
Population, Exposure, Comparison, Outcome, and Study Design (PECOS). The PECOS tool defines the
parameters and provides a framework to help identify the relevant literature to inform the ISA. Because
the 2013 Ozone ISA concluded that evidence was inadequate to determine whether there is a causal
relationship between long-term ozone exposure and cancer, studies are evaluated regardless of specific
study locations. The studies evaluated and subsequently discussed in this section were identified using the
following PECOS tools:
Experimental Studies:
• Population: Study population of any animal toxicological study of mammals at any lifestage
• Exposure: Long-term (in the order of months to years) inhalation exposure to relevant ozone
concentrations (i.e., <2 ppm)
• Comparison: Appropriate comparison group exposed to a negative control (i.e., clean air or
filtered-air control)
• Outcome: Mutagenic, genotoxic, or carcinogenic effects
Study Design: In vivo chronic, subchronic or repeated-dose toxicity studies in mammals;
genotoxicity/mutagenicity studies
7-45
-------
Epidemiologic Studies:
• Population: Any population, including populations or lifestages that might be at increased risk
• Exposure: Long-term ambient concentration of ozone
• Comparison: Per unit increase (in ppb)
• Outcome: Change in risk (incidence/prevalence) of a cancer effect
• Study Design: Epidemiologic studies consisting of cohort, case-control and cross-sectional
studies with appropriate timing of exposure for the health endpoint of interest
7.3.2 Cancer and Related Health Effects
7.3.2.1 Genotoxicity
As noted in the 2013 Ozone ISA, the potential for genotoxic effects relating to ozone exposure
was predicted from the radiomimetic properties of ozone. The decomposition of ozone in water produces
OH and HO2 radicals, the same species that are generally considered to be the biologically active products
of ionizing radiation. Ozone has been observed to cause degradation of DNA in a number of different
models and bacterial strains. The toxic effects of ozone have been generally assumed to be confined to the
tissues directly in contact with the gas, such as the respiratory epithelium.
Several epidemiologic studies evaluated in the 2013 Ozone ISA observed positive associations
between long-term ozone exposure and DNA damage (i.e., DNA adduct levels, oxidative DNA damage,
DNA strand breaks). In addition, there was some evidence of cytogenetic damage (i.e., micronuclei
frequency among lymphocytes and buccal cells) after long-term, but not short-term ozone exposure. Such
DNA and cytogenic damage may be relevant to mechanisms leading to cancer development and serve as
early indicators of an elevated risk of mutagenicity.
Since the 2013 Ozone ISA, a few additional studies have looked at the relationship between
ozone exposure and the potential for DNA damage and found inconsistent results (see Table 7-32):
• Holland et al. (2014) exposed healthy volunteers to filtered air (FA), 100, and 200 ppb ozone and
collected blood lymphocytes 24-hours post-exposure. A statistically significant increase in the
frequency of micronuclei in binucleated cells was reported with increasing ozone concentrations
(p < 0.05). However, these authors also reported no appreciable changes in neoplasmic bridges
(an indicator of radiation and other types of genotoxic exposure) and no difference in cell
proliferation following ozone exposure.
• Finkenwirth et al. (2014) exposed healthy volunteers to FA and ozone, collected lymphocytes and
analyzed them for single stranded breaks. No appreciable difference in single stranded breaks
were observed at either 30 minutes or 4.5 hours post-exposure.
7-46
-------
• In rats, Cestonaro et al. (2017) evaluated exposure to 0.05 ppm ozone from an air purifier for
3 hours or 24 hours per day for 14 or 28 days. The results indicated no significant effects on
indicators of DNA damage such as the frequency of micronuclei in the 3-hour-exposed group (14
or 28 days). In the 24-hour exposure group, there was a statistically significant increase in DNA
damage relative to other groups. However, DNA in the tail was less than 1% and not different
from control exposure.
• In lung tissue from rats, Zhang et al. (2017) reported a significant increase in the most common
base lesion 8-oxoG following ozone exposure (p < 0.05) and that treatment with the
NO-precursor L-arginine reduced the presence of these lesions. Moreover, levels of the base
excision repair component OGG1 were significantly decreased following ozone exposure
(p < 0.05) but treatment with the NO-precursor L-arginine restored these levels.
7.3.2.2 Cancer Incidence, Mortality, and Survival
The 2013 Ozone ISA concluded that the evidence was inadequate to determine whether a causal
relationship exists between ambient ozone exposures and cancer. A limited number of epidemiologic and
animal toxicological studies of lung cancer mortality among humans and lung tumor incidence among
rodents contributed to the evidence informing this conclusion. The reanalysis of the full American Cancer
Society (ACS) Cancer Prevention Study II (CPS II) cohort reported no association between lung cancer
mortality and ozone concentration [HR: 1.00; 95% CI: 0.96, 1.04; Krewski et al. (2009)1. Additionally, no
association was observed when the analysis was restricted to the summer months. There was also no
association between ozone concentration and lung cancer mortality present in a subanalysis of the cohort
in the Los Angeles area. Animal toxicological studies did not demonstrate enhanced lung tumor incidence
in male or female rodents. However, there was an increase in the incidence of oviductal carcinoma in
mice exposed to 0.5 ppm ozone for 16 weeks (U.S. EPA. 2013) The implications of this result are unclear
because the report lacked statistical information. It was noteworthy that there was no mention of oviductal
carcinoma after 32 weeks of exposure, and no oviductal carcinoma was observed after 1 year of exposure.
In contrast, several recent cohort and case-control studies have observed positive associations
between long-term ozone exposure and lung cancer incidence or mortality (see Table 7-29 and
Table 7-30). A single study reported null associations between short-term ozone exposure and
lung-cancer mortality. Associations between ozone exposure and other types of cancer were generally
null. Specifically:
• A case-control study conducted in Canada (Hvstad et al.. 2013) and a cohort study conducted in
China (Guo et al.. 2016) observed positive associations between long-term ozone exposure and
lung cancer incidence. Hvstad et al. (2013) evaluated both modeled and measured ozone
concentrations, while Guo et al. (2016) relied on the exposure assessment hybrid model
developed for the Global Burden of Disease study (Brauer et al.. 2012). Hvstad et al. (2013)
observed ORs that were two times higher for squamous cell lung cancer compared with all lung
cancers. Guo et al. (2016) reported no differences in effects between men and women, but higher
risks for adults aged 65+ years (compared with adults between 30 and 65 years).
7-47
-------
• Two U.S.-based cohort studies (Eckel et al.. 2016; Xu et al.. 2013) reported positive associations
between long-term ozone exposure and lung cancer mortality or respiratory cancer mortality
among individuals that had already been diagnosed with cancer.
• A number of recent studies conducted in the U.S., Canada, and Europe provided limited and
inconsistent evidence for an association between long-term ozone exposure and lung cancer
mortality rCakmak et al. (2018); Turner et al. (2016); Crouse et al. (2015); Carey et al. (2013);
Jerrett et al. (2013); Table 7-301.
• A case-crossover study conducted in Shenyang, China observed null associations between
short-term ozone exposure and lung cancer mortality (Xue et al.. 2018).
• Studies of childhood leukemia (Badaloni et al.. 2013) and breast tissue density, an indicator of
breast cancer (Yaghivan etal.. 2017). observed null associations with long-term ozone exposure
(see Table 7-31).
7.3.3 Summary and Causality Determination
In the 2013 Ozone ISA, very few studies were available to assess the relationship between
long-term exposure to ozone and carcinogenesis (U.S. EPA. 2013). The few available toxicological and
epidemiologic studies suggested that ozone exposure may contribute to DNA damage. However, given
the overall lack of studies, the 2013 Ozone ISA concluded that the evidence was inadequate to determine
whether a causal relationship existed between ambient ozone exposures and cancer.
The number of studies examining the relationship between ozone exposure and the potential for
carcinogenesis reman few. Studies published since the 2013 Ozone ISA provide some additional animal
toxicological evidence that ozone exposure can lead to DNA damage. In addition, several but not all
recent cohort and case-control studies have observed positive associations between long-term ozone
exposure and lung cancer incidence or mortality. Several of the studies evaluating lung cancer mortality
were conducted among populations that had already been diagnosed with cancer in a different organ
system. Associations between ozone exposure and other types of cancer were generally null. Given the
limited evidence base, the lack of an evaluation of copollutant confounding in epidemiologic studies
reporting associations, and the evaluation of study populations that had already been diagnosed with
cancer in several of the epidemiologic studies, the evidence is not sufficient to draw a conclusion
regarding causality (Table 7-5). Thus, the evidence describing the relationship between exposure to
ozone and carcinogenesis remains inadequate to determine if a causal relationship exists.
7-48
-------
Table 7-5 Summary of evidence that is inadequate to determine if a causal
relationship exists between long-term ozone exposure and cancer.
Ozone
Concentrations
Rationale for Causality Associated with
Determination3 Key Evidence13 Key References'3 Effects0
Inconsistent evidence for A limited number of controlled human Holland et al. (2014) 100 ppb, 200 ppb
DNA damage in exposure studies report inconsistent Finkenwirth et al (2014)
experimental studies evidence for DNA damage measured
in lymphocytes
A limited number of animal Cestonaro et al. (2017) 50 ppb
toxicological studies report inconsistent zhana et al (2017)
evidence for DNA damage measured
in lymphocytes
Some epidemiologic
evidence for lung cancer
incidence or mortality
A limited number of recent studies
observed positive associations
between long-term ozone exposure
and lung cancer incidence
Hvstad et al. (2013)
20.3 ppb
Guoetal. (2016)
56.9 ppb
A limited number of recent studies Eckel et al. (2016) 28.5 ppb
observed positive associations
between long-term ozone exposure Xu et al. (2013) 40.2 ppb
and lung cancer or respiratory mortality
in study populations already diagnosed
with cancer
No epidemiologic A limited number of recent studies Badaloni et al. (2013) 24.2 ppb
evidence for other observed null associations between
cancers long-term ozone exposure and Yaqhivan et al. (2017) 36.0 ppb
childhood leukemia and breast cancer
Lack of copollutant No epidemiologic studies evaluated
models contributes to potential copollutant confounding using
uncertainty copollutant models
Limited evidence for Experimental studies provide
biological plausibility inconsistent evidence for DNA damage
in humans and laboratory animals
aBased on aspects considered in judgments of causality and weight of evidence in causal framework in Table I and Table II of the
Preamble to the ISAs (U.S. EPA. 2015).
bDescribes the key evidence and references, supporting or contradicting and contributing most heavily to causality determination
and, where applicable, to uncertainties or inconsistencies. References to earlier sections indicate where full body of evidence is
described.
°Describes the ozone concentrations with which the evidence is substantiated.
7-49
-------
7.4 Evidence Inventories—Data Tables to Summarize Reproductive and Developmental
Effects Study Details
7.4.1 Epidemiologic Studies
Table 7-6 Epidemiologic studies of exposure to ozone and reproduction—male.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% CI
tFarhat et al. (2016)
Sao Paulo, Brazil
Ozone: 1999-2006
Follow-up: January
2000-January 2006
Panel study
n = 35
Male patients with
systematic lupus
erythematosus
Monitor
1-h max
90 days before sample
collection
Mean: 42
75th: 48
Maximum: 54
Correlation (r):
NO2: 0.6;
CO: 0.28
Copollutant models:
NR
Total sperm count (A million per ejaculate)
-74.78 (-136.51, -13.04)
Sperm concentration (A million per mL)
-24.29 (-42.43, -6.15)
tLiu et al. (2017)
n = 1,759 men
Monitor Mean: 25-64
Correlation (r): NR
Sperm concentration (A million/mL): 0.082 (-0.077,
Wuhan, China
0-90 days before sample
Copollutant models:
0.240)
Ozone: NR
collection
NR
Sperm count (A million): 0.018 (-0.145, 0.181)
Follow-up: 2013-2015
Total motility (A percentage): 0.082 (-0.068, 0.236)
Cohort study
Progressive motility (A percentage): 0.068 (-0.086,
0.217)
Total motile sperm count (A million): 0.041 (-0.113,
0.199)
A = change; CI = confidence interval; CO = carbon monoxide; h = hour; max = maximum; mL = milliliters; n = sample size; N02 = nitrogen dioxide; NR = not reported.
tRecent studies evaluated since the 2013 Ozone ISA.
7-50
-------
Table 7-7 Epidemiologic studies of exposure to ozone and reproduction—female.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% CI
tSlama et al. (2013)
Teplice district, Czech Republic
Ozone: NR
Follow-up: November 25,
1993-July 31, 1996
Cohort study
n = 1,916 couples
Monitor
Mean: NR
Correlation (r) (1st mo Fecundity (FR)
unprotected
intercourse):
PM2.5: -0.41;
NO2
SO2
-0.49;
-0.69
Copollutant models:
NO2, PM2.5
30 days before unprotected intercourse:
0.98 (0.81, 1.19)
1st mo unprotected intercourse: 1.12
(0.94, 1.32)
Average of 1st mo unprotected
intercourse and 30 days previous: 1.08
(0.86, 1.37)
30 days after the end of the 1st mo of
unprotected intercourse exposure
window (post-outcome): 1.30 (1.08, 1.56)
tCarre et al. (2016)
Region NR, France
Ozone: 2012-2015
Follow-up: April
2012-December 2015
Cohort study
n = 292 couples
Couples undergoing
IVF attempts
Monitors
NR
Correlation (r): NR
Copollutant models:
NR
Ovarian response to stimulation and
number of top embryos were increased
with short- or long-term exposures to
high levels of ozone
7-51
-------
Table 7-7 (Continued): Epidemiologic studies of exposure to ozone and reproduction—female.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% CI
tNobles etal. (2018)
Michigan and Texas, U.S.
Ozone: 2005-2009
Follow-up: 2005-2009
Cohort study
n = 500 couples
LIFE Study
Couples had
presumed exposure to
persistent organic
pollutants
Model, modified Median (cycle Correlation (r): NR Fecundability (FOR)
CMAQ
avg for first
observed
cycle): 27.85
75th: 34.2
(5 days prior to
first observed
cycle)
Maximum:
40.54 (cycle
avg for first
observed cycle)
Copollutant models:
PM2.5, NOx, SO2, CO
Cycle prior to observed cycle: 0.93 (0.73,
1.21)
Days 1-10 (proliferative phase) of
observed cycle: 0.86 (0.62, 1.17)
FORs for exposures 5 days before to
10 days after ovulation are generally null
or slightly below null (reduced fecundity).
Effects for 5 and 1 days before ovulation
and day of ovulation are below the null.
CI = confidence interval; CMAQ = Community Multiscale Air Quality (model); CO = carbon monoxide; FR = fecundability ratio; fecundability odds ratio; h = hour; max = maximum;
IVF = in vitro fertilization; LIFE = Longitudinal Investigation of Fertility and the Environment (Study); mo = month; n = sample size; N02 = nitrogen dioxide; NOx = oxides of nitrogen;
NR = not reported; S02 = sulfur dioxide.
tRecent studies evaluated since the 2013 Ozone ISA.
7-52
-------
Table 7-8 Epidemiologic studies of exposure to ozone and pregnancy/birth—hypertension disorders.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% CI
tLee etal. (2013)
Pittsburgh, PA, U.S.
Ozone: 1996-2002
Follow-up: 1997-2002
Cohort study
n = 34,702
Magee Obstetric
Medical and Infant
(MOMI) database
Model, space-time Median: 21.7
ordinary kriging 75th: 30.2
95th: 36.2
Maximum: 46.i
Correlation (r):
PM2.5: 0.5;
PM10: 0.7
Copollutant models: NR
Gestational hypertension (OR): 1.07
(0.98, 1.16)
Preeclampsia (OR): 1.07 (0.93, 1.23)
tMobasher et al. (2013) n = 298
Monitor
Los Angeles, CA, U.S.
Ozone: NR
Follow-up: 1996-2008
Case-control study
Attended the Los
Angeles County +
University of
Southern California
Women's and
Children's Hospital
Mean:
1st trimester: 21.5
2nd trimester: 18.2
3rd trimester: 18.2
Correlation (r):
1st trimester PM2.5:
-0.45;
NO2: -0.72;
CO: -0.64
Copollutant models: NR
2nd trimester
PM2.5: -0.53;
NO2: -0.74;
CO: -0.57
Copollutant models: NR
3rd trimester
PM2.5: -0.55;
NO2: -0.78;
CO: -0.66
Copollutant models: NR
Hypertensive disorders of pregnancy
(OR)
1st trimester:
0.94 (0.66, 1.32)
BMI <30: 0.77 (0.51, 1.19)
BMI >30: 1.26 (0.58, 2.75)
2nd trimester:
1.61 (1.14, 2.29)
BMI <30: 1.67 (1.08, 2.59)
BMI >30: 1.23 (0.57, 2.63)
3rd trimester:
1.12 (0.80, 1.58)
BMI <30: 1.28 (0.84, 1.94)
BMI >30: 1.02 (0.52, 2.04)
tOlsson etal. (2013)
Stockholm, Sweden
Ozone: 1997-2006
Follow-up: 1998-2006
Cohort study
n = 120,755
Swedish medical
birth registry
Monitor
Mean: 35
Correlation (r):
NO2: -0.48
Copollutant models:
NO2
Preeclampsia (OR)
1.10 (1.04, 1.17)
Adjusted for NO2: 1.23(1.06,
1.44)
7-53
-------
Table 7-8 (Continued): Epidemiologic studies of exposure to ozone and pregnancy/birth—hypertension
disorders.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% CI
tXu et al. (2014)
Florida, U.S.
Ozone: 2003-2005
Follow-up: 2004-2005
n = 22,041
birth records
Monitor
24-h avg
Mean:
1st trimester: 40
2nd trimester: 41
3rd trimester: 40
Correlation (r): NR
Copollutant models: NR
Hypertension (OR)
1st trimester: 1.00 (0.84, 1.19)
2nd trimester: 0.93 (0.80, 1.08)
Entire pregnancy: 0.93 (0.63, 1.42)
Cohort study
tNahidi et al. (2014)
Tehran, Iran
Ozone: 2010-2011
Follow-up: September
2010-March 2011
n = 65 cases;
130 controls
admitted to hospitals
in Tehran
Monitor
Mean: NR
Correlation (r): NR
Copollutant models: NR
Preeclampsia (OR)
High vs. low exposure: 1.00 (0.49,
2.03)
Case-control study
tMichikawa et al.
(2015)
Kyushu-Okinawa
District, Japan
Ozone: NR
Follow-up: 2005-2010
Cohort study
n = 36,620
Japan Perinatal
Registry Network
Monitor
Mean: 41.3
Median: 40.1
75th: 48
Correlation (r):
PM2.5: 0.12
NO2: -0.18
SO2: -0.17
Copollutant models: NR
Hypertensive disorders of pregnancy
(OR)
1st quintile: reference
2nd quintile: 1.24 (1.07, 1.43)
3rd quintile: 1.35 (1.17, 1.56)
4th quintile: 1.26 (1.08, 1.47)
5th quintile: 1.20 (1.01, 1.42)
tMendola et al. (2016b) n = 192,687 women Model, modified
U.S.
Ozone: NR
Follow-up: 2002-2008
Cohort study
recruited from
12 centers
(19 hospitals) across
U.S.
Consortium on Safe
Labor
CMAQ
Median:
Preconception: 29.7
1st trimester: 29.2
2nd trimester: 29.4
Entire pregnancy: 28.5
Correlation (r): NR
Copollutant models:
NR
Preeclampsia (OR)
Asthma
Preconception: 1.02 (0.94, 1.11)
1st trimester: 0.99 (0.91, 1.07)
2nd trimester: 1.00 (0.92, 1.08)
Entire pregnancy: 0.97 (0.85, 1.12)
No asthma
Preconception: 1.01 (0.98, 1.04)
1st trimester: 1.02 (0.99, 1.05)
2nd trimester: 0.97 (0.94, 1.00)
Entire pregnancy: 0.94 (0.89, 1.01)
7-54
-------
Table 7-8 (Continued): Epidemiologic studies of exposure to ozone and pregnancy/birth—hypertension
disorders.
Exposure
Copollutant
Effect Estimates
Study
Study Population
Assessment
Mean (ppb)
Examination
95% CI
+Hu et al. (2016)
n = 655,529 birth
Model, CMAQ
Mean:
Correlation (r): NR
Hypertensive disorders of pregnancy
Florida, U.S.
records
hierarchical
1st trimester: 38.65
Copollutant models: NR
(each week of 1st two trimesters)
Bayesian
2nd trimester:
1st trimester: 1.08 (1.06, 1.12)
^jzone. ink
38.59Trimesters 1 and 2:
2nd trimester: 1.06 (1.04, 1.08)
Follow-up: 2005-2007
38.63
1st and 2nd trimesters: 1.14(1.10,
Cohort study
Median:
1.17)
1st trimester: 37.91
ORs elevated with ozone exposure at
2nd trimester: 37.84
each week of pregnancy (1-26)
Trimesters 1 and 2: 38.42
tWu et al. (2011)
n = 81,186
Monitor
Mean: 35.6
Correlation (r):
Pre-eclampsia (OR)
Los Angeles and
hospital-based birth
PM2.5: -0.61;
Los Angeles: 1.00 (0.74, 1.35)
Orange counties, CA,
database
NO2: -0.81;
Orange: 1.46 (1.12, 1.90)
U.S.
CO: -0.74
Ozone: 1997-2006
Copollutant models: NR
Follow-up: NR
Cohort study
BMI = body mass index; CI = confidence interval; CMAQ = Community Multiscale Air Quality (model); h = hour; max = maximum; n = sample size; N02 = nitrogen dioxide; NR = not
reported; OR = odds ratio; PM25 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm; S02 = sulfur dioxide.
tRecent studies evaluated since the 2013 Ozone ISA.
7-55
-------
Table 7-9
Epidemiologic studies of exposure to ozone and pregnancy/birth—diabetes.
Study
Study Population Exposure Assessment Mean (ppb) Copollutant Examination
Effect Estimates
95% CI
Gestational diabetes (RR)
Preconception, 91 days prior to
last menstrual period: 0.94
(0.92, 0.97)
1st trimester: 1.00 (0.98, 1.02)
tHu et al. (2015) n = 410,267 birth Model, CMAQ
Florida U S records hierarchical Bayesian
Ozone: NR
Follow-up: 2004-2005
Cohort study
CI = confidence interval; CMAQ = Community Multiscale Air Quality (model); h = hour; max = maximum; n = sample size; N02 = nitrogen dioxide; NR = not reported; OR = odds ratio;
RR = relative risk; S02 = sulfur dioxide.
tRecent studies evaluated since the 2013 Ozone ISA.
tRobledo et al. (2015) n = 220,264
U.S. Consortium on Safe
Ozone: 2001-2008 Labor
Follow-up: 2002-2008 12 clinical
„ , x , centers/19 hospitals
Cohort study
Model, modified
CMAQ
Median:
Preconception: 29.71
1st trimester: 29.21
75th:
Preconception: 35.82
1st trimester: 35.17
Correlation (r):
PM25: -0.38;
NO2: -0.39;
SO2: -0.42
Copollutant models: NR
Mean:
1st trimester: 37.22
2nd trimester: 37.54
Entire pregnancy: 37.40
Median:
1st trimester: 36.48
2nd trimester: 36.95
Entire pregnancy: 37.84
Correlation (r):
PM2.5: 0.39
Copollutant models: NR
Gestational diabetes (OR)
1st trimester: 1.19 (1.14, 1.23)
2nd trimester: 1.25 (1.21, 1.30)
Entire pregnancy: 1.39 (1.32,
1.46)
7-56
-------
Table 7-10 Epidemiologic studies of exposure to ozone and pregnancy/birth—fetal growth.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% CI
tConeus and Spiess
(2012)
Germany
Ozone: 2002-2007
Follow-up: 2002-2007
Cohort study
SOEP
Monitor
NR
Correlation (r): NR
Copollutant models:
NR
Reports only betas and statistical
significance. No evidence of
association between ozone
exposures during pregnancy or 1 mo
before birth and birth length, fetal
growth, or birth weight
tLee etal. (2013)
Pittsburgh, PA, U.S.
Ozone: 1996-2002
Follow-up: 1997-2002
Cohort study
n = 34,702
MOMI database
Model, space-time
ordinary kriging
Median: 21.7
75th: 30.2
95th: 36.2
Maximum: 46.i
Correlation (r):
PM25: 0.5;
PM10: 0.7
Copollutant models:
NR
Small for gestational age (OR):
0.99 (0.91, 1.06)
tEbisu and Bell (2012)
Connecticut, Delaware,
Maryland,
Massachusetts, New
Hampshire, New Jersey,
New York, Pennsylvania,
Rhode Island, Vermont,
Virginia, Washington,
DC, and West Virginia,
U.S.
Ozone: 1999-2007
Follow-up: 2000-2007
Cohort study
n = 1,207,800 Monitor
Birth records, term births
Mean: 23
Correlation (r):
PM2.5: -0.12;
NO2: -0.77;
CO: -0.28
Copollutant models:
NR
Low birth weight (percentage
change per 7 ppb):
-6.3 (-11, -1.3)
7-57
-------
Table 7-10 (Continued): Epidemiologic studies of exposure to ozone and pregnancy/birth—fetal growth.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% CI
tLaurent et al. (2013) n = 70,000 births
Los Angeles and Orange Hospital-based obstetric
counties CA, U.S. database
Ozone: NR
Follow-up: 1996-2006
Cohort study
Monitor
Mean: 35.66
Correlation (r):
PM2.5: -0.61;
NO2: -0.81;
CO: -0.74
Copollutant models:
NR
Term birth weight (A g):
-27.27 (-32.02, -22.51)
Low birth weight (OR):
1.11 (1.02, 1.21)
tOlsson et al. (2013)
Stockholm, Sweden
Ozone: 1997-2006
Follow-up: 1998-2006
Cohort study
n = 120,755
Swedish medical birth
registry
Monitor
Mean: 35
Correlation (r):
NO2: -0.48
Copollutant models:
NO2
Small for gestational age (OR):
0.98 (0.96, 1.02)
Adjusted for NO2: 0.98 (0.90, 1.06)
tGeer et al. (2012)
Texas, U.S.
Ozone: NR
Follow-up: 1998-2004
Cohort study
n = 1,548,904
Birth records, 40 Texas
counties
Monitor
Mean: 25.4
Correlation (r): NR
Copollutant models:
NR
Term birth weight (A g):
-4.61 (-7.34, -1.88)
tRitz et al. (2014)
n = 688 Monitor
Mean: 40.2
Correlation (r): NR
Biparietal diameter (A mm)
New York City, NY, U.S.
Behavior in pregnancy
Maximum: 96.1
Copollutant models:
Estimated date of conception to
Ozone: 1993-1996
study
NR
1st ultrasound date -0-19 weeks
Follow-up: 1993-1996
gestation: 0.026 (-0.153, 0.199)
1st to 2nd ultrasound date
Cohort study
(-19-29 weeks gestation): 0.041
(-0.104, 0.186)
2nd to 3rd ultrasound date
(-29-37 weeks gestation): 0.012
(-0.149, 0.169)
7-58
-------
Table 7-10 (Continued): Epidemiologic studies of exposure to ozone and pregnancy/birth—fetal growth.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% CI
tLaurent et al. (2014)
Los Angeles County CA,
U.S.
Ozone: 2000-2008
Follow-up: 2001-2008
Cohort study
n = 960,945
Birth records, term births
Model, empirical
Bayesian kriging based
on monitor
Mean: 38.95
Correlation (r): NR
Copollutant models:
NR
Low birth weight (OR): 0.99 (0.98,
1.00)
tGrav et al. (2014)
North Carolina, U.S.
Ozone: 2001-2006
Follow-up: 2002-2006
Cohort study
n = 457,642
Birth records
CMAQ downscaler
Mean: 43.2
Correlation (r): NR
Copollutant models:
NR
Birth weight (A g): -12.54 (-16.10,
-8.81)
Small for gestational age (OR): 1.76
(1.75, 1.80)
Low birth weight (OR): 1.80 (1.75,
1.85)
tVinikoor-lmler et al.
(2014)
North Carolina, U.S.
Ozone: 2002-2005
Follow-up: 2003-2005
Cohort study
n = 322,981
Birth registry
Model, hierarchical
Bayesian model CMAQ
and monitor
Mean: 40.85
Median: 41.86
75th: 48.86
Maximum: 70.35
Correlation (r): NR
Copollutant models:
NR
Birth weight (A g)
1st trimester: 1.56 (-2.52, 5.64)
2nd trimester: -7.24 (-12.35, -2.13)
3rd trimester: -22.84 (-28.05,
-17.95)
Low birth weight (RR)
1st trimester: 0.90 (0.85, 0.96)
2nd trimester: 0.87 (0.81, 0.94)
3rd trimester: 1.54(1.43, 1.66)
Small for gestational age (RR)
1st trimester: 1.02 (0.99, 1.04)
2nd trimester: 0.99 (0.96, 1.02)
3rd trimester: 1.09 (1.07, 1.13)
tHaet al. (2014)
Florida, U.S.
Ozone: 2003-2005
Follow-up: 2004-2005
Case-control study
n = 423,719
Birth records, singleton
live births
Model, CMAQ
hierarchical Bayesian
Mean: 37.2
Median: 36.5
75th: 41
Maximum: 56.2
Correlation (r): R
Copollutant models:
PM2.5
Low birth weight (OR)
1st trimester: 0.99 (0.96, 1.03)
Adjusted for PM2.5: 0.96 (0.89, 1.03)
2nd trimester: 0.97 (0.94, 1.01)
Adjusted for PM2.5: 1.02 (0.95, 1.09)
3rd trimester: 0.93 (0.89, 0.96)
Adjusted for PM2.5: 0.95 (0.89, 1.02)
Entire pregnancy: 0.91 (0.86, 0.97)
Adjusted for PM2.5: 0.96 (0.89, 1.04)
7-59
-------
Table 7-10 (Continued): Epidemiologic studies of exposure to ozone and pregnancy/birth—fetal growth.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% CI
tSmith et al. (2015)
Texas, U.S.
Ozone: NR
Follow-up: 2002-2004
Cohort study
n = 565,703
Birth records
Model, CMAQ
downscaler
Mean: NR
Correlation (r): R
Copollutant models:
NR
"Did not find a statistically significant
relationship between 1st-trimester
ozone and birth weight"
"Negative association between fetal
growth and large levels of ozone in
the 2nd trimester"
tBrown et al. (2015)
New York City, NY, U.S.
Ozone: 2001-2006
Follow-up: 2001-2006
Cohort study
n = 421,763
Birth records
Model, CMAQ
hierarchical Bayesian
Median: 38.77
75th: 42.03
Maximum: 60.35
Correlation (r): R
Copollutant models:
NR
Term low birth weight
1st trimester
10.60-29.51: ref
29.51-39.28: 1.02 (0
39.29-47.35: 1.04 (0
47.36-66.33: 1.07 (0.
2nd trimester
11.31-30.11: ref
31.12-40.12: 0.98 (0
40.13-47.43: 0.97 (0
47.44-65.73: 0.98 (0.
3rd Trimester
5.08-30.35: ref
(OR)
96, 1.09)
98, 1.11)
99, 1.14)
92, 1.05)
91, 1.03)
92, 1.05)
30.36-39.57
39.58-46.74
46.75-99.69
0.95 (0.
0.90 (0
0.95 (0.
90, 1.02)
84, 0.96)
90, 1.02)
Entire pregnancy
15.52-35.61
35.62-38.77
38.78-42.03
42.04-60.35
ref
0.88 (0.
0.86 (0
0.98 (0.
83, 0.94)
81, 0.92)
92, 1.04)
tCapobussi et al. (2016) n = 27,128
Como, Italy Birth records
Ozone: NR
Follow-up: 2005-2012
Cohort study
Monitor, within 5 km
Correlation (r): R
Copollutant models:
NR
Low birth weight (OR): 0.96 (0.85,
1.08)
Small for gestational age (OR): 1.00
(0.95, 1.06)
7-60
-------
Table 7-10 (Continued): Epidemiologic studies of exposure to ozone and pregnancy/birth—fetal growth.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% CI
tLaviane et al. (2016)
Ontario, Canada
Ozone: 2002-2009
Follow-up: January 1,
2005-March 31, 2012
n = 362,800
Better Outcomes
Registry & Network
Ontario
Singleton live births
Model
Mean: 27.8
Median: 28
95th: 33.05
Correlation (r):
PM2.5: -0.14;
NO2: -0.53
Copollutant models:
NR
Term low birth weight (OR): 1.17
(1.09, 1.24)
Small for gestational age (OR): 1.24
(1.21, 1.28)
Cohort study
tYitshak-Sade et al.
(2016)
Negev, Israel
Ozone: NR
Follow-up: December
2011-April 2013
n = 959
Bedouin-Arab population
in southern Israel,
seminomadic
Monitor, IDW
Mean: 11
Correlation (r): NR
Copollutant models:
NR
Birth weight (A g)
Entire pregnancy: -0.01 (-0.01,
0.00)
Last month: -0.03 (-0.06, -0.01)
3rd trimester: -0.11 (-0.18, -0.03)
Cohort study
tLaurent et al. (2016a)
California, U.S.
Ozone: 2000-2008
n = 72,632 cases; * five
controls
Birth records
Model, empirical
Bayesian kriging based
on monitor
Mean: 39.55
Correlation (r): NR
Copollutant models:
NR
Low birth weight (OR): 1.03 (1.02,
1.05)
Follow-up: 2001-2008
Case-control study
+Tu et al. (2016)
Georgia, U.S.
Ozone: 2001
Follow-up: 2000
n = 105,818
Term births, birth records
Model, CMAQ
downscaler
2001 annual average
Mean: 42.76
Maximum: 48.99
Correlation (r): NR
Copollutant models:
NR
Ozone was a significant risk factor
only in small parts of the state, and
variations depend on different
socioeconomic status and urbanicity
of communities.
Cohort study
7-61
-------
Table 7-10 (Continued): Epidemiologic studies of exposure to ozone and pregnancy/birth—fetal growth.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% CI
tCarvalho et al. (2016)
Sao Paulo, Brazil
Ozone: NR
Follow-up: April
2011 -December 2013
Cohort study
n = 453
Procriar
Pregnant women from
three prenatal care units
Personal sampler
24-h avg
Mean: 4
Median: 4
Maximum: 7
Correlation (r): NR
Copollutant models:
NO2
All outcomes are log-transformed,
and all ozone and NO2 trimester
exposures are in the same model
Pulsatility index umbilical) (A log
(index)
1st trimester: 0.07 (-0.13, 0.27)
2nd trimester: 1.56 (1.42, 1.70)
3rd trimester: -1.62 (-1.78, -1.46)
Fetal weight (A log [g])
1st trimester: 0.06 (-0.08, 0.18)
2nd trimester: 0.04 (-0.04, 0.12)
3rd trimester: 0.67 (0.57, 0.77)
Birth weight (A log [g])
1st trimester: -0.65 (-0.85, -0.45)
2nd trimester: 0.26 (0.14, 0.38)
3rd trimester: 0.25 (0.11, 0.39)
tArrovo et al. (2016)
Madrid, Spain
Ozone: 2001-2009
Follow-up: 2001-2009
Time-series study
n = 470 weeks Monitors
All live singleton births in
Madrid
Mean: 18
Median: NR
75th: NR
Maximum: 38
Correlation (r): NR
Copollutant models:
NR
Low birth weight (RR)
Only reported statistically significant
results
Week 12 of gestation: 1.01 (1.00,
1.02)
tDiaz et al. (2016)
Madrid, Spain
Ozone: NR
Follow-up: 2001-2009
Time-series study
Term births
Monitor
Mean: 17
Maximum: 38
Correlation (r): NR
Copollutant models:
NR
Low birth weight (weekly)
Reported only statistically significant
effects, no associations reported for
ozone
7-62
-------
Table 7-10 (Continued): Epidemiologic studies of exposure to ozone and pregnancy/birth—fetal growth.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% CI
tMichikawa et al.
(2017a)
Kyushu-Okinawa District,
Japan
Ozone: NR
Follow-up: 2005-2010
Study
n = 29,177
Japan Perinatal Registry
Network
Monitor
Mean: 41.2
Median: 40
75th: 47.8
Correlation (r): NR
Copollutant models:
NR
Adverse birth weight (small for
gestational age and low birth weight)
(OR)
Entire pregnancy: 1.06 (0.94, 1.19)
1st trimester: 1.07 (1.01, 1.14)
Small for gestational age (OR)
Entire pregnancy: 1.06 (0.96, 1.16)
1st trimester: 1.07 (1.01, 1.12)
tSmith et al. (2017)
London, U.K.
Ozone: NR
Follow-up: 2006-2010
Cohort study
n = 540,365
Term births
Model, KCLurban
Mean: 16
Correlation (r): NR
Copollutant models:
NO2, PM2.5
Low birth weight
0.92 (0.86, 0.98)
Adjusted NO2: 0.94 (0.86, 1.02)
Adjusted for PM2.5: 0.94 (0.88, 1.02)
Small for gestational age (OR)
0.98 (0.96, 1.02)
tFernando Costa
Nascimento et al. (2017)
Sao Jose do Rio Preto,
Brazil
Ozone: 2011-2013
Follow-up: 2012-2013
Cohort study
n = 8,948 births Monitor
Birth records, term births
singletons no birth
defects
Mean: 28
Median: 28
75th: 30
Maximum: 36
Correlation (r): NR
Copollutant models:
NR
Low birth weight (OR)
Reported elevated ORs for
exposures 30, 60, and 90 days
before delivery; exposure increment
is unclear
tChen et al. (2017b)
Brisbane, Australia
Ozone: 2002-2013
Follow-up: July 1,
2003-December 31,
2013
Cohort study
173,720 birth records Monitors
24-h avg
Mean: 16.82
Median: 16.78
75th: 17.58
Maximum: 22.34
Correlation (r):
PM2.5: 0.27;
NO2: -0.04;
SO2: -0.04;
CO: -0.16
Copollutant models:
NR
Low birth weight
Entire pregnancy: 2.20 (1.74, 2.75)
Adjusted for any copollutant: 1.21
(1.14, 1.29)
Trimester effect estimates reported
as figures
7-63
-------
Table 7-10 (Continued): Epidemiologic studies of exposure to ozone and pregnancy/birth—fetal growth.
Study
Study Population Exposure Assessment Mean (ppb)
Copollutant
Examination
Effect Estimates
95% CI
tReis et al. (2017) n = 13,660 birth records Monitor
Volta Redonda and Rio
de Janeiro, Brazil
Ozone: 2002-2006
Mean: 30
Correlation (r): NR
Copollutant models:
NR
Low birth weight
Exposure increment not reported,
ORs elevated from null reported for
2nd and 3rd trimester exposures but
not 1st
Follow-up: 2003-2006
Cohort study
A = change; CI = confidence interval; CMAQ = Community Multiscale Air Quality (model); g = gram; h = hour; IDW = inverse-distance weighting; max = maximum; MOMI = Magee
Obstetric Medical and Infant (database); n = sample size; N02 = nitrogen dioxide; NR = not reported; OR = odds ratio; PM25 = particulate matter with a nominal mean aerodynamic
diameter less than or equal to 2.5 |jm; ppb = parts per billion; RR = relative risk; S02 = sulfur dioxide; SOEP = Socioeconomic Panel (study).
tRecent studies evaluated since the 2013 Ozone ISA.
7-64
-------
Table 7-11 Epidemiologic studies of exposure to ozone and pregnancy/birth—preterm birth.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant Examination
Effect Estimates
95% CI
tWu etal. (2011)
Los Angeles and
Orange counties, CA,
U.S.
Ozone: 1997-2006
Follow-up: NR
Cohort study
n = 81,186
Hospital-based birth
database
Monitor
Mean: 36.5
Correlation (r):
PM2.5: -0.61;
NO2: -0.81;
CO: -0.74
Copollutant models: NR
Entire pregnancy:
Very PTB (OR)
Los Angeles: 0.94 (0.59, 1.54)
Orange: 1.44 (0.72, 2.86)
PTB (OR)
Los Angeles: 1.00(0.81, 1.21)
Orange: 1.08 (0.86, 1.35)
tOlsson etal. (2012)
Stockholm, Sweden
Ozone: 1986-1995
Follow-up:
1987-1995
Cohort study
n = 115,588
Swedish medical birth
registry
Monitor
Mean:
1st trimester:
29
2nd trimester:
29
Last week of
gestation: 30
Median:
1st trimester:
28
2nd trimester:
28
Last week of
gestation: 30
Correlation (r):
NO2:
1st trimester: -0.43;
2nd trimester: -0.39;
Last week of gestation: -0.26
Copollutant models: NO2
PTB (OR)
1st trimester
1.17 (1.06, 1.28)
Adjusted for NO2: 1.12 (1.00, 1.28)
2nd trimester
1.04 (0.94, 1.14)
Adjusted for NO2: 1.10 (0.96, 1.25)
Last week of gestation
1.03 (0.94, 1.12)
Adjusted for NO2: 1.06 (0.94, 1.16)
Gestational age (A weeks)
1st trimester:
-0.12 (-0.16, -0.08)
Adjusted for NO2: -0.10 (-0.16,
-0.04)
2nd trimester:
-0.06 (-0.10, -0.02)
Adjusted for NO2: -0.14 (-0.20,
-0.08)
Last week of gestation: -0.06
(-0.09, -0.03)
Adjusted for NO2: -0.06 (-0.09,
-0.03)
7-65
-------
Table 7-11 (Continued): Epidemiologic studies of exposure to ozone and pregnancy/birth—preterm birth.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant Examination
Effect Estimates
95% CI
tLee etal. (2013)
Pittsburgh, PA, U.S.
Ozone: 1996-2002
Follow-up:
1997-2002
Cohort study
n = 34,702
Magee Obstetric Medical
and Infant (MOMI)
database
Model, space-time
ordinary kriging
1st trimester
Median: 21.7
75th: 30.2
95th: 36.2
Maximum: 46.:
Correlation (r):
PM2.5: 0.5;
PM10: 0.7;
Copollutant models: NR
PTB (OR), 1st trimester
1.13 (1.01, 1.27)
tOlsson etal. (2013)
Stockholm, Sweden
Ozone: 1997-2006
Follow-up:
1998-2006
Cohort study
n = 120,755
Swedish medical birth
registry
Monitor
Mean: 35 Correlation (r):
NO2: -0.48
Copollutant models: NO2
PTB (OR), 1st trimester:
1.08 (1.02, 1.17)
Mother with asthma: 1.17 (1.02,
1.32)
Mother without asthma: 1.08 (1.00,
1.17)
Adjusted for NO2:
1.10 (0.98, 1.23)
Mother with asthma: 1.17 (1.00,
1.37)
Mother without asthma: 1.10(0.98,
1.23)
live birth
tWarren et al. (2012) n = 32,170 observations Monitor and Model
Texas, U.S. Birth records, singleton (fused CMAQ)
Ozone: 2002-2004
Follow-up:
2002-2004
Cohort study
Correlation (r): R
Copollutant models: NR
PTB, results presented as figures
Effect estimates elevated from null
with exposures in early weeks of
pregnancy, and for 1st and 2nd
trimester exposures.
tSchifano et al.
(2013)
Rome, Italy
Ozone: 2001-2007
Follow-up:
2001-2010
Cohort study
n = 132,691 births Monitor
Birth records 8-h max
Median: 19
75th: 29
Maximum: 66
Correlation (r): NR
Copollutant models: NR
PTB (percentage increase), lag 1-2:
1.01 (0.94, 1.09)
7-66
-------
Table 7-11 (Continued): Epidemiologic studies of exposure to ozone and pregnancy/birth—preterm birth.
Exposure Effect Estimates
Study Study Population Assessment Mean (ppb) Copollutant Examination 95% CI
tGrav etal. (2014) n = 457,642 CMAQ downscaler Mean: 43.2 Correlation (r): R PTB (OR): 1.03 (0.98, 1.07)
North Carolina, U.S. Birth records Entire pregnancy Copollutant models: NR
Ozone: 2001-2006
Follow-up:
2002-2006
Cohort study
7-67
-------
Table 7-11 (Continued): Epidemiologic studies of exposure to ozone and pregnancy/birth—preterm birth.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant Examination
Effect Estimates
95% CI
tHaet al. (2014)
Florida, U.S.
Ozone: 2003-2005
Follow-up:
2004-2005
Case-control study
n =423,719
Birth records, singleton
live births
Model, CMAQ
hierarchical
Bayesian
Mean:
1st trimester:
37.2
2nd trimester:
37.6
3rd trimester:
37.4
Entire
pregnancy:
37.4
Median:
1st trimester:
36.5
2nd trimester:
37
3rd trimester:
37
Entire
pregnancy:
37.9
75th:
1st trimester:
41
2nd trimester:
41.3
3rd trimester:
41.1
Entire
pregnancy: 41
Maximum:
1st trimester:
56.2
2nd trimester:
57.3
3rd trimester:
69.2
Entire
pregnancy:
51.3
Correlation (r): R
Copollutant models: PM2.5
Very PTB (OR)
1st trimester: 1.07 (1.02, 1.12)
2nd trimester: 1.09 (1.04, 1.14)
3rd trimester: 0.98 (0.93, 1.04)
Entire pregnancy: 1.18 (1.10, 1.27)
Adjusted for PM2.5
1st trimester: 1.16 (1.05, 1.28)
2nd trimester: 1.15 (1.05, 1.27)
3rd trimester: 0.97 (0.86, 1.08)
Entire pregnancy: 1.25(1.11, 1.40)
PTB (OR)
1st trimester: 1.02 (1.00, 1.03)
2nd trimester: 1.03 (1.01, 1.05)
3rd trimester: 0.99 (0.97, 1.01)
Entire pregnancy: 1.04 (1.01, 1.07)
Adjusted for PM2.5
1st trimester: 1.06 (1.02, 1.10)
2nd trimester: 1.07 (1.03, 1.11)
3rd trimester: 0.98 (0.95, 1.02)
Entire pregnancy: 1.08 (1.03, 1.13)
7-68
-------
Table 7-11 (Continued): Epidemiologic studies of exposure to ozone and pregnancy/birth—preterm birth.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant Examination
Effect Estimates
95% CI
tHao et al. (2016)
Georgia, U.S.
Ozone: 2002-2006
Follow-up: January 1,
2002-February 28,
2006
Cohort study
n = 511,658 births
Birth records
Model, CMAQ
fused with
monitors
8-h max
Median: 40.88
75th: 51.14
Maximum:
75.06
Correlation (r):
PM2.5: 0.47;
NO2
SO2
-0.19;
-0.18;
PM10: 0.70
Copollutant models: NR
PTB (27-36 weeks, OR)
1st trimester: 1.006 (0.995, 1.017)
2nd trimester: 1.008 (0.997, 1.019)
3rd trimester: 0.992 (0.983, 1.002)
Entire pregnancy: 1.012 (0.991,
1.036)
tLin et al. (2015)
Region NR, Taiwan
Ozone: 2000-2007
Follow-up:
2001-2007
Case-control study
n = 86,224 cases;
344,896 controls
Birth registry
Monitor
8-h max
Mean: 1st
trimester: 42.74
2nd trimester:
42.94
3rd trimester:
43.30
75th:
1st trimester:
48.22
2nd trimester:
48.43
3rd trimester:
48.98
Maximum:
1st trimester:
77.68
2nd trimester:
77.68
3rd trimester:
86.55
Correlation (r):
NO2: -0.05;
SO2: 0.17;
PM10: 0.53
Copollutant models: NR
PTB (OR)
1st trimester: 1.03 (1.02, 1.04)
2nd trimester: 1.02 (1.01, 1.02)
3rd trimester: 1.02 (1.01, 1.03)
tQian et al. (2015)
Wuhan, China
Ozone: 2010-2013
Follow-up:
August 19, 2010-
September 9, 2013
Cohort study
n = 95,911
Monitor
Mean: 38
Maximum: 74
Correlation (r):
PM2.5: -0.16;
NO2: -0.12;
SO2: -0.13;
CO: -0.12
Copollutant models: NR
PTB (OR)
2nd trimester: 1.08 (1.04, 1.12)
Entire pregnancy: 1.10 (1.04, 1.14)
Adjusted for PM2.5: 1.08 (1.04, 1.14)
Adjusted for NO2: 1.10(1.02, 1.17)
Adjusted for SO2: 1.08 (1.02, 1.14)
7-69
-------
Table 7-11 (Continued): Epidemiologic studies of exposure to ozone and pregnancy/birth—preterm birth.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant Examination
Effect Estimates
95% CI
tSchifano et al.
n = Rome: 78,633
Monitor
Mean: Rome:
Correlation (r):
Length of gestation (HR)
(2016)
Rome, Italy and
Barcelona, Spain
n = Barcelona: 27,255
8-h avg
~51
Barcelona: -25
NO2:
Rome: -0.1;
Barcelona: -0.36
Lag 0-2
Rome: 1.01 (1.00, 1.02)
Barcelona: 1.01 (1.00, 1.02)
Ozone: NR
Follow-up: Rome:
2001-2010
Barcelona:
PM10:
Rome: 0.17;
Barcelona: -0.19
Copollutant models: NR
2007-2012
tCapobussi et al.
n =27,128
Monitor, within
Correlation (r): R
PTB (OR): 0.99 (0.92, 1.06)
(2016)
Birth records
5 km
Copollutant models: NR
Como, Italy
Entire pregnancy
Ozone: NR
Follow-up:
2005-2012
Cohort study
tLaviqne et al. (2016) n = 362,800
Ontario, Canada
Ozone: 2002-2009
Follow-up: January 1,
2005-March 31,
2012
Cohort study
Singleton live births
Better Outcomes
Registry & Network
Ontario
Model
Entire pregnancy
Mean: 27.8
Median: 28
95th: 33.05
Correlation (r):
PM2.5: -0.14;
NO2: -0.53
Copollutant models: NR
PTB (OR)
1.04 (1.01, 1.08)
Asthma: 1.25 (1.07, 1.47)
No asthma: 1.04 (1.00, 1.07)
Diabetes: 0.89 (0.73, 1.08)
No diabetes: 1.07 (1.00, 1.12)
Preeclampsia: 1.00 (0.86, 1.16)
No preeclampsia: 1.05 (1.01, 1.09)
tWallace et al.
(2016)
U.S.
Ozone: NR
Follow-up:
2002-2008
Cohort study
n =223,375
Singleton
Consortium on Safe
Labor, Air Quality, and
Reproductive Health
Study
Model, modified
CMAQ
Entire pregnancy
Mean:
PROM:
28.5
0 h
1 h
2 h
3 h
4 h
29.3
28.9
28.8
28.8
29
Correlation (r): NR
Copollutant models: NR
PROM (OR): 1.01 (0.94, 1.09)
Lag before delivery
0 h: 1.03 (1.01, 1.05)
1 h: 1.05 (1.03, 1.07)
2 h: 1.05 (1.03, 1.07)
3 h: 1.05 (1.03, 1.07)
4 h: 1.04 (1.02, 1.05)
PPROM (OR): 1.08 (0.94, 1.22)
7-70
-------
Table 7-11 (Continued): Epidemiologic studies of exposure to ozone and pregnancy/birth—preterm birth.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant Examination
Effect Estimates
95% CI
tMendola et al.
n = 223,502 singleton
Model, modified
Mean: 29.65 Correlation (r): NR
PTB
(2016a)
pregnancies
CMAQ
Median: 37.3 Copollutant models: NR
Elevated ORs with ozone exposure
U.S.
Recruited from
for early PTB (<34 weeks) in
Ozone: NR
12 centers (19 hospitals)
mothers without asthma for
Follow-up:
2002-2008
across the U.S.
Consortium on Safe
exposures during Weeks 8-14 and
15-21
Cohort study
Labor, Air Quality, and
Reproductive Health
Study
tSvmanski et al.
(2016)
Houston, Harris
County, TX, U.S.
Ozone: NR
Follow-up:
2005-2007
Case-control study
n = 10,459 cases; Monitor
152,214 singleton births
Mean: 37.6
Median: 33.7
75th: 47.5
90th: 62.4
Correlation (r): NR
Copollutant models: NR
Severe PTB (20-28 weeks) (OR)
Weeks 1-4: 0.83 (0.74, 0.94)
Weeks 5-8: 0.88 (0.78, 1.01)
Weeks 9-12: 0.95 (0.84, 1.09)
Weeks 13-16: 1.05 (0.92, 1.19)
Weeks 17-20: 1.21 (1.08, 1.36)
Entire pregnancy: 1.16 (0.82, 1.62)
Moderate PTB (29-32 weeks) (OR)
Weeks 1-4: 0.93 (0.85, 1.03)
Weeks 5-8: 1.00 (0.89, 1.13)
Weeks 9-12: 1.09 (0.98, 1.22)
Weeks 13-16: 1.03 (0.93, 1.15)
Weeks 17-20: 1.13 (1.02, 1.25)
Weeks 21-24: 1.03(0.92, 1.16)
Weeks 25-28: 1.15 (1.04, 1.27)
Entire pregnancy: 1.31 (0.99, 1.74)
Late PTB (33-36 weeks) (OR)
Weeks 1-4: 0.99 (0.95, 1.02)
Weeks 5-8: 1.01 (0.97, 1.05)
Weeks 9-12: 1.04 (1.00, 1.08)
Weeks 13-16
Weeks 17-20
Weeks 21-24
Weeks 25-28
Weeks 29-32
1.03 (0.99, 1.07)
1.08 (1.04, 1.12)
1.05 (1.01, 1.09)
1.07 (1.03, 1.10)
1.02 (0.98, 1.05)
Entire pregnancy: 1.21 (1.10, 1.32)
7-71
-------
Table 7-11 (Continued): Epidemiologic studies of exposure to ozone and pregnancy/birth—preterm birth.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant Examination
Effect Estimates
95% CI
tLaurent et al.
(2016b)
California, U.S.
Ozone: 2000-2008
Follow-up:
2001-2008
Case-control study
n = 442,314 cases; * two Model, empirical
controls Bayesian kriging
Birth records based on monitor
Entire pregnancy
Mean: 39.71
Correlation (r):
PM2.5: -0.24;
NO2: -0.07
Copollutant models: NR
PTB (OR)
1.08 (1.07, 1.09)
Adjusted for PM2.5: 1.09 (1.08, 1.10)
Adjusted for NO2: 1.08 (1.07, 1.09)
tArrovo et al. (2016)
Madrid, Spain
Ozone: 2001-2009
Follow-up:
2001-2009
Time-series study
All live singleton births in Monitors
Madrid
Mean: 18
Median: NA
75th: NA
Maximum: 38
Correlation (r): NR
Copollutant models: NR
PTB (OR)
Week 12: 1.02 (1.01, 1.03)
Only significant results reported
7-72
-------
Table 7-11 (Continued): Epidemiologic studies of exposure to ozone and pregnancy/birth—preterm birth.
tChen etal. (2017b)
Brisbane, Australia
Ozone: 2002-2013
Follow-up: July 1,
2003-Dec 31, 2013
Cohort study
Median:
Entire
pregnancy:
16.78
1st trimester:
16.09
2nd trimester:
16.21
3rd trimester:
16.12
75th:
Entire
pregnancy:
17.58
1st trimester:
19.02
2nd trimester:
18.67
3rd trimester:
19.1
Maximum:
Entire
pregnancy:
22.34
1st trimester:
23.55
2nd trimester:
24.35
3rd trimester:
28.21
173,720 birth records Monitors Mean: Correlation (r): PM2 5: 0.27
24-h avg Entire NO2: -0.04;
pregnancy: SO2: -0.04;
16.82 CO:-0.16
1st trimester. Copollutant models: NR
16.82
2nd trimester:
16.76
3rd trimester:
16.91
PTB (HR)
Entire pregnancy: 2.20 (1.85, 2.61)
Adjusted for PM2.5, NO2, or SO2:
1.21 (1.14, 1.29)
Elevated HRs for exposures during
each trimester reported as figures
7-73
-------
Table 7-11 (Continued): Epidemiologic studies of exposure to ozone and pregnancy/birth—preterm birth.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant Examination
Effect Estimates
95% CI
tDastoomoor et al.
(2017)
Khuzestan Province,
Iran
Ozone: March
2008-March 2015
n =49,173 births
Ahvaz Imam Khomeini
Hospital
Monitor
24-h avg
Mean: 32
Median: 24
Maximum:
3,316
Correlation (r): NR
Copollutant models: NR
PTB (RR)
Cumulative lag 0-14: 0.98 (0.93,
1.03)
Lag 0: 1.00 (0.99, 1.01)
Lag 1: 1.00 (0.99, 1.01)
Lag 2: 1.00 (1.00, 1.01)
Follow-up: March
2008-March 2015
Time-series study
tSmith et al. (2015)
Texas, U.S
Ozone: NR
Follow-up:
2002-2004
Cohort study
n = 565,703
Birth records
Model, CMAQ
downscaler
Mean: NR
Correlation (r): NR
Copollutant models: NR
2nd-trimester ozone exposure
increases (low to middle or middle
to high) were negatively associated
with gestational age in south and
east Texas. 1st-trimester ozone was
negatively associated with
gestational age in southeast Texas
avg = average; CI = confidence interval; CMAQ = Community Multiscale Air Quality (model); g = gram; h = hour; HR = hazard ratio; IDW= inverse-distance weighting;
max = maximum; n = sample size; N02 = nitrogen dioxide; NR = not reported; OR = odds ratio; PM25 = particulate matter with a nominal mean aerodynamic diameter less than or
equal to 2.5 |jm; ppb = parts per billion; PPROM = preterm premature rupture of membranes; PROM = premature rupture of membranes; PTB = preterm birth; RR = relative risk;
S02 = sulfur dioxide.
tRecent studies evaluated since the 2013 Ozone ISA.
7-74
-------
Table 7-12 Epidemiologic studies of exposure to ozone and pregnancy/birth—birth defects.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Outcomes Examined
tDadvand et al. (20111
Northeast England
Ozone: 1993-2003
Follow-up:
Case-control study
n pooled Monitor
cases = 2,140; controls Weeks 3-8
= 14,256
Northern Congenital
Abnormality Survey
Mean: NR
Correlation (r): R
Copollutant
models: NR
Congenital malformations of cardiac chambers
and connections
Congenital malformations of cardiac septa
Congenital malformations of pulmonary and
tricuspid valves
Congenital malformations of aortic and mitral
valves
Congenital malformations of great arteries and
veins
Ventricular septal defect
Atrial septal defect
Congenital pulmonary valve stenosis
Tetralogy of Fallot
Coarctation of aorta
tPadula et al. (2013)
San Joaquin Valley,
CA, U.S.
Ozone: 1997-2006
Follow-up: October
1997-December 2006
Case-control study
n = 1,651 subjects
NBDPS, CA
Monitor
8-h max
1 st 2 mo of
pregnancy
Median: 46.95
75th: 62.65
Maximum: 91.92
Correlation (r):
PM2.5: -0.61; NO2:
-0.35;
CO: -0.57
Copollutant
models: NR
Neural tube defects
Anencephaly
Spina bifida
Cleft lip with or without cleft palate
Cleft palate only
Gastroschisis
tAqav-Shav et al.
(2013)
Israel
Ozone: 1999-2006
Follow-up: 2000-2006
Case-control study
Israel National Birth
and Birth Defect
Registry
Monitor, within Mean: 25.1
10 km, inverse- Median: 26.5
distance weighing 75th: 39.1
Weeks 3-8 Maximum: 128
Correlation (r): NR Multiple congenital heart defects
Copollutant Atrial and atrial septal defects
Isolated ventricular septal defects
Patent ductus arteriosus (BW > 2,500 g)
models: NR
7-75
-------
Table 7-12 (Continued): Epidemiologic studies of exposure to ozone and pregnancy/birth—birth defects.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Outcomes Examined
tVinikoor-lmler et al.
(2013)
North Carolina, U.S.
Ozone: 2002-2005
Follow-up: 2003-2005
Cohort study
n = 322,969
Birth defect
surveillance, birth
registry
Model, hierarchical
Bayesian model
CMAQ and
monitor
Weeks 3-8
Mean: 40.74
Median: 42.15
Maximum: 74.99
Correlation (r): NR
Copollutant
models: NR
Spina bifida
Hydrocephalus
Anophthalmia/microphthalmia
Congenital cataract
Microtia/anotia
Transposition of great vessels
Tetralogy of Fallot
Ventricular septal defect
Atrial septal defect
Endocardial cushion defect/atrioventricular
septal defect
Pulmonary valve atresia/stenosis
Tricuspid valve atresia/stenosis
Aortic valve stenosis
Hyperplastic left heart syndrome
Coarctation of aorta
Cleft palate
Cleft lip with or without cleft palate
Esophageal atresia/tracheoesophageal fistula
Anorectal atresia/stenosis
Pyloric stenosis
Renal agenesis
Obstructive genitourinary defect
Hypospadias
Deficiency defect—upper limbs
Deficiency defect—lower limbs
Gastroschisis
Omphalocele
Diaphragmatic hernia
7-76
-------
Table 7-12 (Continued): Epidemiologic studies of exposure to ozone and pregnancy/birth—birth defects.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Outcomes Examined
tStinaone et al. (2014)
Arkansas, Iowa,
Massachusetts,
California, Georgia,
New York, North
Carolina, Texas, and
Utah, U.S.
Ozone: 1997-2006
Follow-up: October 1,
1997-December 31,
2006
Case-control study
n = 3,328 cases;
4,632 controls
NBDPS
Monitor
Weeks 2-8
Mean: 42.9
90th: 51.8
Correlation (r): NR
Copollutant
models: NR
Cardiovascular birth defects
Left ventricular outflow tract obstructions
Aortic stenosis
Coarctation of the aorta
Hypoplastic left heart syndrome
D-transposition of the great arteries
Tetralogy of Fallot
Other conotruncals
Common truncus
Other double-outlet right ventricle with
transposition of the great arteries or not (other)
Interrupted aortic arch type B or not otherwise
specified
Conoventricular septal defects
Anomalous pulmonary venous return
Total anomalous pulmonary venous return
Atrioventricular septal defect
Right ventricular outflow tract obstructions
Pulmonary/tricuspid atresia
Pulmonary atresia
Tricuspid atresia
Pulmonary valve stenosis
Ebstein's anomaly
Septal defects
Perimembranous ventricular septal defects
Muscular-muscular ventricular septal defects
Atrial septal defect
7-77
-------
Table 7-12 (Continued): Epidemiologic studies of exposure to ozone and pregnancy/birth—birth defects.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Outcomes Examined
tLin etal. (2014b)
Taiwan
Ozone: 2000-2007
Follow-up: 2001-2007
Case-control study
n = 1,687 cases, 10x
controls
Birth registry, isolated
cases
Monitor
Weeks 1-4
Weeks 5-8
Weeks 9-12
Mean: 41
Correlation (r):
PM10: 0.52; N02:
-0.07; SO2: 0.18;
CO: -0.17
Copollutant
models: NR
Limb defects
Syndactyly
Polydactyly
Reduction deformities of limb
tJurewicz et al. (2014)
Poland
Ozone: NR
Follow-up: NR
Cohort study
n = 212 men
Men attending an
infertility clinic for
diagnostic purposes
Environmental factors
and male infertility
Monitor
90 days before
sample collection
Mean: 23
Median: 23
Maximum: 41
Correlation (r):
PM2.5: -0.41;
NO2: -0.44;
SO2: -0.26;
CO: -0.43
Copollutant
models: NR
Sperm chromosomal disomy
tFarhietal. (2014)
Israel
Ozone: NR
Follow-up: 1997-2004
Cohort study
n = 216,730 infants;
Monitor, kriging
Mean:
207,825 spontaneous
1st trimester
1st trimester: 32.4
conceptions,
2nd trimester: 32.7
8,905 assistive
Entire pregnancy:
reproductive
32.1
technology
Median:
conceptions
1st trimester: 32.3
Birth records
2nd trimester: 32.6
Entire pregnancy:
31.3
75th:
1st trimester: 36.1
2nd trimester: 36.5
Entire pregnancy:
34.6
Maximum:
1st trimester: 54.4
2nd trimester: 54.9
Entire pregnancy: 50
Correlation (r): NR
Copollutant
models: NR
Total birth defects
7-78
-------
Table 7-12 (Continued): Epidemiologic studies of exposure to ozone and pregnancy/birth—birth defects.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Outcomes Examined
tVinikoor-lmler et al.
21,060 cases;
Model, hierarchical
Mean:
Correlation (r): NR
Total birth defects
(2015)
1,401,611 controls
Bayesian model
Texas: 40.3
Copollutant
Texas, U.S.
Birth defect registry
CMAQ and
NBDPS: 37.2
models: NR
Ozone: NR
monitor
Median:
Follow-up: 2002-2006
1st trimester
Texas: 40.5
NBDPS: 34.9
Case-control study
75th:
Texas: 46.8
NBDPS: 43.4
Maximum:
Texas: 65.1
NBDPS: 62.3
tHwana et al. (2015)
n = 1,087 cases;
Monitor
Mean: 44.53
Correlation (r): NR
Ventricular septal defects
Taiwan
10,870 controls
1st trimester
Median: 44.14
Copollutant
Atrial septal defects
Ozone: NR
Birth records
models: NR
Patent ductus arteriosus
Follow-up: 2001-2007
Pulmonary artery and valve
Case-control study
Tetralogy of Fallot
Transposition of the great arteries
Conotruncal defects
tZhanq et al. (2016)
n = 105,988 births
Central site
Mean: 37
Correlation (r):
Congenital heart defects
Wuhan, China
monitor, nearest
75th: 54
N02: -0.12;
Ventricular septal defect
Ozone: 2010-2012
8-h avg
S02: -0.16;
CO: -0.20
Tetralogy of Fallot
Follow-up: June 10,
2011-June 9, 2013
1st trimester
Copollutant
models: NO2, SO2,
Cohort study
CO
7-79
-------
Table 7-12 (Continued): Epidemiologic studies of exposure to ozone and pregnancy/birth—birth defects.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Outcomes Examined
tZhou etal. (2016)
Arizona, Florida, New
York (excluding NYC),
and Texas, U.S.
Ozone: 2001-2007
Follow-up: 2001-2007
Case-control study
n = 7,035 cases;
4,697,523 total live
births
NBDPS
Model, CMAQ
downscaler
8-h max
5-10 weeks
Mean:
All oral clefts: 40.9
Cleft lip with/without
palate: 40.9
Cleft palate: 40.7
Median:
All oral clefts: 40.3
Cleft lip with/without
palate: 40.5
Cleft palate: 40.1
75th:
All oral clefts: 48.1
Cleft lip with/without
palate: 48
Cleft palate: 48.1
Maximum:
All oral clefts: 69
Cleft lip with/without
palate: 69
Cleft palate: 64.9
Correlation (r): NR All oral clefts
Copollutant
models: NR
Cleft palate
Cleft lip with/without palate
CI = confidence interval; CMAQ = Community Multiscale Air Quality (model); g = gram; h = hour; HR = hazard ratio; IDW = inverse-distance weighting; max = maximum; n = sample
size; NBDPS = National Birth Defects Prevention Study; N02 = nitrogen dioxide; NR = not reported; PM25 = particulate matter with a nominal mean aerodynamic diameter less than
or equal to 2.5 |jm; PM10 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 10 |jm; ppb=parts per billion; S02 = sulfur dioxide.
tRecent studies evaluated since the 2013 Ozone ISA.
7-80
-------
Table 7-13 Epidemiologic studies of exposure to ozone and pregnancy/birth—infant and fetal mortality.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% CI
tHwanq et al. (2011)
n = 9,325 stillbirths; Monitor
Mean: 35.93
Correlation (r):
Stillbirth (OR)
Taiwan
93,250 controls
Maximum: 61.27
N02: -0.31;
1st trimester: 1.01 (0.96, 1.06)
Ozone: NR
Birth records
S02: 0.13; PM10:
2nd trimester: 0.96 (0.91, 1.01)
0.62
3rd trimester: 0.98 (0.93, 1.04)
Follow-up: 2001-2007
Copollutant models:
Entire pregnancy: 0.97 (0.91, 1.04)
Case-control study
NR
tMoridi etal. (2014)
Tehran, Iran
Ozone: NR
Follow-up: June
2010-February 2011
Case-control study
n = 148 cases;
148 controls
Monitor
Mean: 22.29-28.88
Correlation (r): NR
Copollutant models:
NR
Spontaneous abortion before
14 weeks of pregnancy (OR): 2.43
(1.72, 3.42)
tGreen etal. (2015)
California, U.S.
Ozone: NR
Follow-up: 1999-2009
Cohort study
n = 13,999 stillbirths; Monitor
3,012,270 livebirths
Birth records
Mean: 48.48
Median: 47.27
75th: 55.52
Correlation (r): NR
Copollutant models:
NR
Stillbirth (OR)
1st trimester: 1.00 (0.98, 1.02)
2nd trimester: 1.01 (0.99, 1.03)
3rd pregnancy: 1.03 (1.01, 1.05)
Entire pregnancy: 1.01 (0.99, 1.04)
tArrovo et al. (2016)
Madrid, Spain
Ozone: 2001-2009
Follow-up: 2001-2009
Time-series study
n = 470 weeks
All live singleton
births in Madrid
Monitors
Mean: 18
Median: NR
75th: NR
Maximum: 38
Correlation (r):
NR
Copollutant models:
NR
Late fetal death (<24 h)
Only reports statistically significant
results, examined exposure during
each week of pregnancy
Week 24: 1.33 (1.32, 1.35)
7-81
-------
Table 7-13 (Continued): Epidemiologic studies of exposure to ozone and pregnancy/birth—infant and fetal
mortality.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% CI
Mean
Correlation (r): NR
Stillbirth (OR)
entire pregnancy:
Copollutant models:
Entire pregnancy: 1.53 (1.06, 2.19)
29.3
NR
Asthma: 1.29 (0.76, 2.20)
1st trimester: 29.0
No asthma: 1.33 (0.89, 1.97)
Lag-0
29.9
1st trimester: 1.14 (1.00, 1.31)
Lag-1
30.0
Lag 0
1.07 (0.97, 1.20)
Lag-2
30.1
Lag 1
1.07 (0.96, 1.19)
Lag-3
30.1
Lag 2
1.13 (1.01, 1.26)
Lag-4
30.1
Lag 3
1.11 (1.00, 1.24)
Lag-5
30.1
Lag 4
1.10 (0.99, 1.22)
Lag-6
30.0
Lag 5
1.18 (1.06, 1.31)
Lag-7
29.9
Lag 6
1.15 (1.03, 1.27)
Median:
Lag 7
1.12 (1.01, 1.25)
entire pregnancy:
28.5
1st trimester: 29.2
75th:
entire pregnancy:
32.7
1st trimester: 35.1
IQR:
entire
pregnancy:7.8
1st trimester: 12.3
Lag-0
17.9
Lag-1
17.8
Lag-2
17.7
Lag-3
17.7
Lag-4
17.7
Lag-5
17.7
Lag-6
17.8
Lag-7
17.8
Maximum:
entire pregnancy:
46.4
1st trimester: 48.7
tMendola et al. (2017)
U.S.
Ozone: NR
Follow-up: 2002-2008
Cohort study
n = 223,375 singleton
pregnancies
Recruited from
12 centers
(19 hospitals) across
the U.S.
Consortium on Safe
Labor
Model, modified
CMAQ
7-82
-------
Table 7-13 (Continued): Epidemiologic studies of exposure to ozone and pregnancy/birth—infant and fetal
mortality.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% CI
tDastoorpoor et al. (2017)
Khuzestan Province, Iran
Ozone: March 2008-March
2015
Follow-up: March 2008-March
2015
Time-series study
n = 49,173 births
Ahvaz Imam
Khomeini Hospital
Monitor
24-h avg
Mean: 32
Median: 24
Maximum: 3,316
Correlation (r): NR
Copollutant models:
NR
Spontaneous abortion (OR)
Lag 0-14: 0.99 (0.95, 1.03)
Lag 0: 1.00 (1.00, 1.01)
Lag 1: 1.00 (0.99, 1.00)
Lag 2: 1.00 (0.99, 1.00)
Stillbirth (OR)
Lag 0-14: 0.89 (0.83, 0.96)
Lag 0
Lag 1
Lag 2
0.99 (0.97, 1.01)
0.99 (0.97, 1.00)
0.99 (0.98, 1.00)
tHaet al. (2017b)
Michigan and Texas, U.S.
Ozone: 2005-2009
Follow-up: 2005-2009
Cohort study
n = 343
Couples trying to get
pregnant followed
through pregnancy
Longitudinal
Investigation of
Fertility and the
Environment
Model, modified
CMAQ
24-h avg
Median: 25
75th: 29.5
Maximum: 42.6
Correlation (r):
PM2 5: -0.25;
NO2
SO2
-0.42;
-0.04
Copollutant models:
NR
Pregnancy loss, 1st observed (HR)
gestational week of loss (e.g., 5, 6,
etc.): 1.00 (0.95, 1.05)
Week before loss: 1.05 (0.98, 1.11)
Entire pregnancy: 1.15 (1.08, 1.21)
tYanq et al. (2018)
Wuhan, China
Ozone: NR
Follow-up: June 10,
2011-June 9, 2013
Cohort study
n = 95,354
Monitor
Entire pregnancy
Mean: 38
Median: 38
75th: 73
Maximum: 76
Correlation (r):
PM2.5: -0.126;
NO2: -0.698;
SO2: -0.468;
CO: -0.499
Copollutant models:
NR
Stillbirth (OR):
Entire pregnancy: 0.85 (0.71, 1.
1st mo: 0.88 (0.35, 2.25)
2nd mo: 1.00 (0.90, 1.10)
3rd mo: 1.14 (0.42, 3.13)
1st trimester: 1.00 (0.88, 1.12)
2nd trimester: 1.02 (0.92, 1.14)
3rd trimester: 0.72 (0.64, 0.81)
02)
avg = average; CI = confidence interval; CMAQ = Community Multiscale Air Quality (model); g = gram; h = hour; HR = hazard ratio; max = maximum; n = sample size; N02 = nitrogen
dioxide; NR = not reported; OR = odds ratio; PM2.5 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm; PM10 = particulate matter with a
nominal mean aerodynamic diameter less than or equal to 10 |jm; ppb = parts per billion; S02 = sulfur dioxide.
tRecent studies evaluated since the 2013 Ozone ISA.
7-83
-------
Table 7-14 Epidemiologic studies of exposure to ozone and developmental effects.
Study
Study Population
Exposure Assessment Mean (ppb) Copollutant Examination
Effect Estimates
95% CI
tPeel etal. (2011)
Atlanta, GA, U.S.
Ozone: August 1,
1998-December 31,
2002
Follow-up: August 1,
1998-December 31,
2002
Cohort study
n = 4,277 infants
Apnea Center of
Children's Healthcare
of Atlanta at Egleston,
children at high risk for
cardiorespiratory
events
Monitor
8-h max
Mean: 43.9
Median: 39.6
90th: 78
Maximum:
130.8
Correlation (r):
PM2.5: 0.42;
NO2: 0.45;
SO2: -0.11;
PM10: 0.48
Copollutant models: NR
Apnea (OR)
Lag 0-1: 1.03 (0.99, 1.07)
Bradycardia (OR)
Lag 0-1: 1.04 (1.02, 1.06)
tConeus and Spiess
(2012)
Nationally representative
sample, Germany
Ozone: 2002-2007
Follow-up: 2002-2007
Cohort study
German SOEP
Monitor
NR
Correlation (r): NR
Copollutant models: NR
No correlation between mean ozone
exposure early life and bronchitis, croup
syndrome, respiratory disease or other
disorders at 2-3 yr of age.
tBreton et al. (2012) n = 768 Monitors, inverse
Southern California, U.S. College students TROY distance squared
Ozone: 1980-2009
Follow-up: 2007-2009
Cohort study
weighing, 4 within
50 km
24-h avg
0-5 yr:
Mean: 23.1
Maximum:
41.8
Correlation (r):
PM2.5: 0.09;
NO2: -0.09;
PM10: 0.18
Copollutant models: NR
Carotid artery intima-media thickness
(A |jm), 0-5 yr of age exposure
7.8 (-0.3, 15.9)
NO2 adjusted: 10.0 (1.4, 18.6)
PM10 adjusted: 8.5 (0.2, 16.9)
PM2.5 adjusted: 9.1 (0.9, 17.4)
tVolket al. (2013)
n = 524 mother-child Monitor
Correlation (r): NR
Autism (OR)
California, U.S.
Pa'rs 8-h max
Copollutant models: NR
1st trimester: 1.05 (0.97, 1.20)
Ozone: 1997-2008
CHARGE
2nd trimester: 1.02 (0.90, 1.17)
3rd trimester: 1.02 (0.89, 1.15)
Follow-up: 2002-2011
Entire pregnancy: 1.05 (0.84, 1.31)
Case-control study
1 st yr of life: 1.09 (0.82, 1.47)
7-84
-------
Table 7-14 (Continued): Epidemiologic studies of exposure to ozone and developmental effects.
Study
Study Population
Exposure Assessment Mean (ppb) Copollutant Examination
Effect Estimates
95% CI
tBecerra et al. (2013)
Los Angeles, CA, U.S.
Ozone: 1995-2006
Follow-up: 1998-2009
Case-control study
n = 7,603 cases;
75,782 controls
Mothers who gave birth
in Los Angeles, CA to
children diagnosed with
ASD at 3-5 yr old
during 1998-2009
Monitors
8-h avg
Entire pregnancy
Mean: 36.8
Median: NA
75th: NA
Correlation (r):
PM2.5: -0.47;
NO2: -0.50;
CO: -0.55
Copollutant models: NR
Autism (OR)
1.05 (1.01, 1.10)
Adjusted for PM10: 1.05(1.01, 1.10)
Adjusted for PM2.5: 1.10(1.05, 1.16)
Adjusted for NO: 1.07 (1.03, 1.12)
Adjusted for NO2: 1.07 (1.03, 1.12)
>High school: 1.03 (0.99, 1.08)
High school: 1.06 (1.01, 1.12)
-------
Table 7-14 (Continued): Epidemiologic studies of exposure to ozone and developmental effects.
Study
Study Population
Exposure Assessment Mean (ppb) Copollutant Examination
Effect Estimates
95% CI
tMaclntvre et al. (2014)
Europe and Canada
Ozone: NR
Follow-up: recruitment
1994-1997
Other study
n = 5,115
TAG study
Metacohort:
combination of several
cohorts: CAPPS,
SAGE, the BAMSE
survey, GINI plus
environmental and
genetic influences on
allergy, LISA, and the
PIAMA study
Model for European
populations,
APMoSPHERE; monitor
for Canadian
populations using IDW
of the closest three
ambient monitors
(within 50 km)
1st yr of life
Mean: 19
Maximum: 28
Correlation (r):
N02: -0.19
Copollutant models: NR
Current asthma (OR): 0.66 (0.26, 1.61)
Ever asthma (OR): 0.74 (0.44, 1.28)
Ever wheeze (OR): 0.81 (0.55, 1.21)
Ever asthma and current wheeze (OR):
1.02 (0.38, 2.79)
Current wheeze (OR): 1.66 (0.77, 3.53)
tFuertes et al. (2013b)
Germany
Ozone: NR
Follow-up: recruitment
1995-1998
Cohort study
n = 6,604 children
Birth cohort full term
normal weight
GINI plus and LISAplus
APMoSPHERE models,
only 2001
Mean: 22
Median: 22
75th: 23
Maximum: 30
Correlation (r): NR
Copollutant models: NR
Eyes and nose symptoms (OR): 0.90
(0.63, 1.29)
Aeroallergen sensitization (OR): 0.97
(0.68, 1.33)
Allergic rhinitis (OR): 1.07 (0.70, 1.64)
Asthma (OR): 1.84 (0.93, 3.69)
tVolket al. (2014)
California, U.S.
Ozone: 1997-2009
Follow-up:
Case-control study
n = 252 cases;
156 controls
Childhood Autism Risk
from Genetics and the
Environment Study
Monitor
Entire pregnancy
Mean: NR Correlation (r): NR
Copollutant models: NR
ASD (OR)
<41.8 ppb ozone, CG/GG SNP:
reference
<41.8 ppb ozone, CC SNP: 1.0 (0.59,
1.9)
>41.8 ppb ozone, CG/GG SNP: 1.2
(0.67, 2.2)
>41.8 ppb ozone, CC SNP: 0.95 (0.45,
2.2)
tLin et al. (2014a)
Taiwan
Ozone: NR
Follow-up: October
2003-January 2004,
recruitment
Cohort study
n = 511 mother-child
pairs
Taiwan Birth Cohort
Pilot Study
Monitor
Mean: 31-39
Correlation (r): NR
Copollutant models: NR
No associations between ozone
exposure at any time period
(1st trimester, 2nd and 3rd trimesters,
birth—12 mo, or 13-18 mo) and
neurodevelopmental scores (gross
motor, fine motor, language, and
social-personal)
7-86
-------
Table 7-14 (Continued): Epidemiologic studies of exposure to ozone and developmental effects.
Effect Estimates
Study Study Population Exposure Assessment Mean (ppb) Copollutant Examination 95% CI
tOrione et al. (2014)
n = 20 cases; Monitor
Mean: NR Correlation (r): NR
Juvenile dermatomyositis (OR)
Sao Paulo, Brazil
56 controls
Copollutant models: NR
No association with ozone for entire
Ozone: NR
Cases from Pediatric
pregnancy exposures, some
Follow-up: August
2011-August 2012
Rheumatology Unit of
inconsistent associations with trimester
the Children's Institute
specific exposures (e.g., elevated OR
for middle fertile exposure in
Case-control study
1st trimester)
tvan Rossem et al.
n = 1,131 mother-infant Monitor
Median: 15-24 Correlation (r): NR
Increases in newborn systolic blood
(2015)
pairs recruited from
Copollutant models: NR
pressure with 1st and 2nd trimester
Boston, MA, U.S.
eight urban and
ozone increases (13.0 and 12.8 ppb,
Ozone: NR
suburban observation
respectively), and decreases with 3rd
offices
trimester exposure (13.6 ppb).
Follow-up: recruited April
Project Viva
Decreases in blood pressure with
1999—July 2002
exposure lagged from birth up to
Cohort study 90 days.
tMalmavist et al. (2015)
Scania, Sweden
Ozone: NR
Follow-up: 1999-2013
Cohort study
n = 262 cases;
682 controls
Skane study
(1999-2005), Better
Diabetes Diagnosis
Monitor
Median: 26.5
75th: 30.6
Correlation (r): NR
Copollutant models: NR
Type 1 diabetes (OR)
1st trimester
<22 ppb: reference
22-26.5 ppb: 1.18 (0.69, 2.04)
26.6-30.6 ppb: 1.26 (0.71, 2.24)
>30.6 ppb: 1.52 (0.88, 2.61)
2nd trimester
<22 ppb: ref
22-26.5 ppb: 1.36 (0.82, 2.26)
26.6-30.6 ppb: 1.48 (0.86, 2.54)
>30.6 ppb: 1.62 (0.99, 2.65)
3rd trimester
<22 ppb: ref
22-26.5 ppb: 0.87 (0.52, 1.79)
26.6-30.6 ppb: 1.1 (0.67, 1.79)
>30.6 ppb: 0.79 (0.49, 1.27)
7-87
-------
Table 7-14 (Continued): Epidemiologic studies of exposure to ozone and developmental effects.
Study
Study Population
Exposure Assessment Mean (ppb) Copollutant Examination
Effect Estimates
95% CI
tHuana et al. (2015)
Taiwan
Ozone: 2004-2006
Follow-up: 2005
Cohort study
n = 16,686
Taiwan Birth Cohort
Study
Monitor, kriging
Mean: 27.9
Median: 27.5
75th: 32.5
Correlation (r): NR
Copollutant models: NR
Atopic dermatitis at 6 mo (OR)
1st trimester: 1.05 (0.90, 1.22)
2nd trimester: 1.16 (0.98, 1.37)
3rd trimester: 1.13 (0.94, 1.35)
Entire pregnancy: 1.09 (0.92, 1.28)
3 mo post-birth: 1.00 (0.84, 1.20)
Entire pregnancy + 3 mo post-birth:
1.13 (0.86, 1.47)
7-88
-------
Table 7-14 (Continued): Epidemiologic studies of exposure to ozone and developmental effects.
Study
Study Population
Exposure Assessment Mean (ppb) Copollutant Examination
Effect Estimates
95% CI
tBreton et al. (2016)
California, U.S.
Ozone: NR
Follow-up: Kindergarten
(2002-2003) through
age 11
Cohort study
n =459
Children enrolled in
kindergarten or
first grade from public
schools
CHS
Monitors, IDW2, four
within 50 km
Mean: NR
Median: NA
75th: NA
Correlation (r):
PM2.5: 0.41;
NO2: 0.01;
PM10: 0.7
Copollutant models: NR
1st trimester
Left CIMT (A mm): -0.00 (-0.00, 0.00)
Right CIMT (A mm): -0.00 (-0.00, 0.00)
Systolic BP (A mm Hg): -0.14 (-0.53,
0.25)
Diastolic BP (A mm Hg): -0.15 (-0.43,
0.13)
LINE 1 (A methylation): -0.20 (-0.32,
-0.07)
Percentage AluYb8 methylation (OR):
0.94 (0.82, 1.08)
2nd trimester
Left CIMT (A mm): 0.00 (-0.00, 0.00)
Right CIMT (A mm): 0.00 (-0.00, 0.00)
Systolic BP (A mm Hg): 0.05 (-0.33,
0.43)
Diastolic BP (A mm Hg): -0.04 (-0.32,
0.24)
LINE 1 (A methylation): 0.05 (-0.08,
0.18)
Percentage AluYb8 methylation (OR):
0.95 (0.83, 1.10)
3rd Trimester
Left CIMT (A mm): -0.00 (-0.00, 0.00)
Right CIMT (A mm): -0.00 (-0.00, 0.00)
Systolic BP (A mm Hg): 0.05 (-0.39,
0.48)
Diastolic BP (A mm Hg): 0.07 (-0.25,
0.39)
LINE 1 (A methylation): 0.15 (0.00,
0.31)
Percentage AluYb8 methylation (OR):
1.02 (0.87, 1.19)
7-89
-------
Table 7-14 (Continued): Epidemiologic studies of exposure to ozone and developmental effects.
Study
Study Population
Exposure Assessment Mean (ppb) Copollutant Examination
Effect Estimates
95% CI
tNishimura et al. (2016)
Chicago, IL; Bronx, NY;
Houston, TX; San
Francisco Bay Area, CA;
and Puerto Rico, U.S.
Ozone: NR
Follow-up: 2006-2011
Cohort study
n = 1,032 asthma
cases
GALA II
Monitor
8-h max
Correlation (r): NR
Copollutant models: NR
Atopic status (OR)
1st year of life: 1.74(1.23, 2.46)
tKimetal. (2017)
California, U.S.
Ozone: 1999-2010
Follow-up: 2002-2011
Case-control study
n = 158 cases;
147 controls
CHARGE
Monitor
8-h max
Correlation (r): NR
Copollutant models: NR
Autism (OR)
Entire pregnancy + 1st two yr of life:
1.34 (0.89, 2.01)
tConde et al. (2018)
Sao Paulo, Brazil
Ozone: NR
Follow-up: June
2014-July 2015
Case-control study
n = 30 cSLE patients;
86 healthy controls
1-h max
Correlation (r): NR
Copollutant models: NR
Childhood lupus gestational period, and
each year to 9 yr
Reports significant ORs only. OR
elevated from null for 2nd fertile of
ozone exposure in 3rd yr of life, but CIs
are wide
tGoodrich et al. (2017)
California, U.S.
Ozone: 1997-2011
Follow-up: 2002-2011
Case-control study
n = 346 ASD cases; Monitor
260 typical 8_h max
development controls
CHARGE
Median: 17
Correlation (r):
PM25: -0.463;
NO2: -0.425;
PM10: -0.022
Copollutant models: NR
ASD (OR)
1st trimester
Low folic acid intake: 1.06 (0.71, 1.58)
High folic acid intake: 1.12 (0.75, 1.65)
2nd trimester
Low folic acid intake: 1.15(0.77, 1.71)
High folic acid intake: 1.00 (0.67, 1.49)
3rd trimester
Low folic acid intake: 1.20 (0.80, 1.82)
High folic acid intake: 0.94 (0.64, 1.38)
7-90
-------
Table 7-14 (Continued): Epidemiologic studies of exposure to ozone and developmental effects.
Study
Study Population
Exposure Assessment Mean (ppb) Copollutant Examination
Effect Estimates
95% CI
tFranca et al. (2018)
Sao Paulo, Brazil
Ozone: NR
Follow-up: 2013-2014
Case-control study
n = 66 cases;
124 controls
Hospital recruits
Monitor
Mean: 44 Correlation (r): NR
Copollutant models: NR
Juvenile idiopathic arthritis (OR)
Examined exposures during each
trimester and each year of life to
diagnosis.
Reported only significant effects, with
2nd fertile (41-44 ppb) of ozone
exposure in the 2nd yr of life
tKerin et al. (2017) n = 325 ASD cases Monitor
California, U.S. CHARGE 8-h max
Ozone: 1998-2008
Follow-up:
Cohort study
A = change; ADOS = Autism Diagnostic Observation Schedule; APMoSPHERE = Air Pollution Modeling for Support to Policy on Health and Environmental Risks in Europe;
ASD = Autism Spectrum Disorder; avg = average; BAMSE = Children, Allergy, Milieu, Stockholm, Epidemiological (survey); CAPPS = Canadian Asthma Primary Prevention Study;
CHARGE = Childhood Autism Risks from Genetics and the Environment; CHS = Children's Health Study; CI = confidence interval; CIMT = carotid intima-media thickness;
CMAQ = Community Multiscale Air Quality (model); cSLE = Childhood-onset systemic lupus erythematosus; CSS = Calibrated Severity Score; g = gram;
GALA = Genes-Environments and Admixture in Latino Americans; GINIplus = German Infant Nutritional Intervention plus environmental and genetic influences on allergy
development; h = hour; HR = hazard ratio; km = kilometer; LISAplus = Influence of Life-style factors on Development of the Immune System and Allergies in East and West Germany
plus Air Pollution and Genetics on Allergy Development; max = maximum; |jm = micron; mm Hg = millimeter mercury; mo(s) = month(s); MSEL = Mullen Scales of Early Learning;
n = sample size; N02 = nitrogen dioxide; NR = not reported; OR = odds ratio; PIAMA = Prevention and Incidence of Asthma and Mite Allergy; PM2.5 = particulate matter with a
nominal mean aerodynamic diameter less than or equal to 2.5 |jm; PM10 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 10 |jm; ppb = parts per
billion; SAGE = Study of Asthma, Genes, and the Environment; S02 = sulfur dioxide; SOEP = Socio-Economic Panel; TAG = Traffic, Asthma, and Genetics; TROY = Testing
Responses on Youth; VABS = Vineland Adaptive Behavior Scales; yr = year.
fRecent studies evaluated since the 2013 Ozone ISA.
Mean: 37.3 Correlation (r): NR
Copollutant models: NR
Neurodevelopmental (A score)
No evidence of association between
prenatal or yr 1 ozone exposure and
any neurodevelopmental score (VABS,
MSEL, ADOS CSS)
7-91
-------
Table 7-15 Epidemiologic studies of exposure to ozone and other effects measured during pregnancy.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% CI
tLee etal. (2011)
Pittsburgh, PA, U.S.
Ozone: 1996-2001
Follow-up: 1997-2001
Cohort study
n = 1,696 women
Enrolled in clinics and
private practices early in
pregnancy
Prenatal Exposures and
Preeclampsia Prevention
Study
Model, space-time
ordinary kriging
Mean: 29.9
Median: 30.3
75th: 34.3
95th: 41.6
Maximum: 51.4
Correlation (r):
PM25: 0.5;
NO2: 0;
SO2: 0.1;
PM10: 0.5
Copollutant
models: NR
C-reactive protein (OR, <8 vs.
>8 ng/mL)
Lag Days 0-7 before draw: 1.32 (0.85,
2.04)
Lag Days 0-21 before draw: 1.33
(0.74, 2.37)
Lag Days 0-28 before draw: 1.07
(0.57, 2.01)
tLee etal. (2012)
Pittsburgh, PA, U.S.
Ozone: 1996-2001
Follow-up: 1997-2001
Cohort study
n = 1,684 women
Enrolled in clinics and
private practices early in
pregnancy
Prenatal Exposures and
Preeclampsia Prevention
Study
Model, space-time Mean: 22.7
ordinary kriging
1st trimester
Median: 22.9
75th: 30.1
95th: 35.6
Maximum: 42.7
Correlation (r):
PM2.5: 0.5;
NO2
SO2
-0.5
-0.6
PM10: 0.7
Copollutant
models: NR
Diastolic BP (A mm Hg):
0.48 (-0.31, 1.27)
Nonsmokers: 0.74 (-0.30, 1.77)
Systolic BP (A mm Hg):
0.96 (-0.07, 1.99)
Nonsmokers: 1.20 (0.69, 3.03)
7-92
-------
Table 7-15 (Continued): Epidemiologic studies of exposure to ozone and other effects measured during
pregnancy.
tMannisto et al. (2015b) n = 500
U.S.
Ozone: 2006
Follow-up: 2006
Cohort study
Random sample from
cohort
Consortium on Safe
Labor
Model, modified
CMAQ
Mean: 41.4
Median: 42.2
Maximum: 68.3
Correlation (r): NR Per 10% increase in Ozone
Copollutant
models: NR
Diastolic BP (A mm Hg)
Hourly mean, at BP measurement hour
Normotensive: 0.41 (0.07, 0.74)
Chronic hypertension: -0.45 (-2.06,
1.16)
Pregnancy related hypertension: 0.09
(-0.37, 0.54)
Hourly mean, 1 h before BP
measurement
Normotensive: 0.34 (0.01, 0.66)
Chronic hypertension: -0.64 (-2.28,
1.00)
Pregnancy related hypertension: 0.05
(-0.35, 0.44)
Daily average
Normotensive: 0.19 (-0.06, 0.44)
Chronic hypertension: -1.02 (-3.41,
1.36)
Pregnancy related hypertension: 0.37
(-0.25, 0.99)
Systolic BP (A mm Hg)
Hourly mean, at BP measurement hour
Normotensive: 0.27 (-0.15, 0.70)
Chronic hypertension: 0.46 (-2.10,
3.02)
Pregnancy related hypertension: 0.05
(-0.45, 0.54)
Hourly mean, 1 h before BP
measurement
Normotensive: 0.17 (-0.24, 0.57)
Chronic hypertension: -0.01 (-2.65,
2.62)
Pregnancy related hypertension: 0.00
(-0.43, 0.43)
Daily average
Normotensive: 0.15 (-0.16, 0.46)
Chronic hypertension: -0.76 (-4.58,
3.07)
Pregnancy related hypertension: 0.35
(-0.32, 1.03)
7-93
-------
Table 7-15 (Continued): Epidemiologic studies of exposure to ozone and other effects measured during
pregnancy.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% CI
tMannisto et al. (2015a)
n = 223,502 singleton
Model, modified
Mean: 29.9
Correlation (r):
Cardiovascular events during labor and
U.S.
pregnancies
CMAQ
Maximum: 79.8
PM2.5: -0.1;
delivery.
Ozone: NR
Recruited from
24-h avg
NO2: -0.35;
SO2: -0.18;
CO: -0.3
OR below the null for exposures at lag
Follow-up: 2002-2008
12 centers (19 hospitals)
across U.S.
Days 5, 6, and 7. No association with
exposures for lag Days 0 to 4
Cohort study
Consortium on Safe
Labor
Copollutant
models: NR
tMichikawa et al. (2016)
Kyushu-Okinawa District,
Japan
Ozone: NR
Follow-up: 2005-2010
Cohort study
n = 40,573
Japan Perinatal Registry
Network
Monitor
Mean: 41.3
Median: 39.9
Correlation (r):
PM2.5: 0.17;
NO2
SO2
-0.16;
-0.09
Copollutant
models: NR
Placenta previa (OR):
0-4 weeks of gestation: 1.08 (1.00,
1.16)
5-12 weeks of gestation: 1.07 (1.00,
1.15)
13-28 weeks of gestation: 0.97 (0.88,
1.08)
tHettfleisch et al. (2016)
Sao Paulo, Brazil
Ozone: NR
Follow-up: October
2011-January 2014
Cross-sectional study
n = 229
Personal monitor Mean: 4
Correlation (r):
NO2: 0.088
Copollutant
models: NR
No association between placental flow
index, placental vascularization index,
or placental vascularization flow index
with 1 week of ozone exposures
tMichikawa et al.
(2017b)
Kyushu-Okinawa District,
Japan
Ozone: NR
Follow-up: 2005-2010
Other study
n = 821 cases
Japan Perinatal Registry
Network
Monitor
Mean: 41.1
Median: 40.2
75th: 51.2
90th: 62.6
Correlation (r):
Copollutant
models: NR
NR No association with placental abruption
with daily ozone exposure, lags of 1 to
5 before event
7-94
-------
Table 7-15 (Continued): Epidemiologic studies of exposure to ozone and other effects measured during
pregnancy.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% CI
tHaet al. (2017a)
U.S.
Ozone: NR
Follow-up: 2002-2008
Other study
n = 680
12 U.S. clinical sites
Consortium on Safe
Labor
Model, modified
CMAQ
24-h avg
Mean: 30.32
75th: 37.85
Maximum: 54.47
Copollutant
models: NR
Correlation (r): NR Average of lag Days 0-7
Cardiovascular events at labor and
delivery, any (OR): 0.89 (0.72, 1.12)
Cardiac arrest (OR): 0.63 (0.45, 0.86)
Heart failure (OR): 0.86 (0.47, 1.57)
Unspecified event (OR): 1.17 (0.72,
1.93)
Stroke (OR): 1.44 (0.80, 2.61)
Ischemic heart disease (OR): 2.41
(1.12, 5.23)
tMorokuma et al. (2017)
Kyushu-Okinawa District,
Japan
Ozone: NR
Follow-up: 2005-2010
Cohort study
n = 23,782
Japan Perinatal Registry
Network
Monitor
Mean:
1st trimester: 41.3
2nd trimester: 42.0
3rd trimester: 41.6
Median:
1st trimester: 40.2
2nd trimester: 41.0
3rd trimester: 40.2
75th:
1st trimester: 47.9
2nd trimester: 48.3
3rd trimester: 50.4
Correlation (r): NR Fetal heart rate false positives (OR):
Copollutant 0.98 (0.92,1.05)
models' NR Fetal heart rate false positives (OR):
1.01 (0.95,1.08)
Fetal heart rate false positives (OR):
1.04 (0.99, 1.10)
A = change; avg = average; BP = blood pressure; CI = confidence interval; CMAQ = Community Multiscale Air Quality (model); h = hour; IDW = inverse-distance weighting;
m = millimeter; max = maximum; mL = milliliter; mm Hg-millimeter mercury; n = sample size; N02 = nitrogen dioxide; NR = not reported; OR = odds ratio; PM25 = particulate matter
with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm; ppb = parts per billion; RR = relative risk; S02 =sulfur dioxide.
tRecent studies evaluated since the 2013 Ozone ISA.
7-95
-------
7.4.2
Toxicological Studies
Table 7-16 Study-specific details from studies of ozone and pregnancy/birth outcomes.
Study
Species (Strain), N, Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
tAvdalovic et al. (2012)
Rhesus monkeys (healthy and HDM sensitized)
n = 12/group males, 0 females
Age: 1 mo
0.5 ppm
Ozone: 8 h/day for 5 days followed by
9 days FA
HDM: Days 3-5 of exposure and
sensitization at Day 14 or Day 28 of
exposure
Lung volume, alveolar volume
and number, alveolar capillary
surface density, mRNA for genes
in alveolar growth and
angiogenesis pathways (3 or
6 mo of age)
tGordon et al. (2017a)
Rats (LE); dams fed high fat or control diet for
6 weeks before breeding
n = 4 male and 4 female offspring;
10 dams/treatment group
Age: PND 161-162 ozone exposure
Offspring 0.8 ppm, 4 h/day for 2 days,
offspring exposed at PND 161 and 162,
exercise and/or diet challenges
Dam body weight, ventilation,
BALF counts, glucose tolerance
test (GD 7, dam body weight;
PND 162 offspring
measurements including BALF,
body weight and body
composition)
tMiller et al. (2019)
In vitro trophoblast cell culture treated with
serum from ozone-exposed LE pregnant dams;
gestation Days 5 and 6 (window of implantation)
LE Dams exposed to 0.4-1.2 ppm for
4 h on gestation Days 5 and 6 (window
of implantation)
Serum from control or
ozone-exposed dams added to
trophoblasts in vitro to evaluate
effect on trophoblast invasion
and migration as well as
trophoblast metabolic capacity
7-96
-------
Table 7-16 (Continued): Study-specific details from studies of ozone and pregnancy/birth outcomes.
Exposure Details
Study Species (Strain), N, Sex, Age (Concentration, Duration) Endpoints Examined
tMiller et al. (2017) Rats (LE) 0, 0.4, or 0.8 ppm, 4 h/day for 2 days at BALF (GD 21)
n = 10/group adult pregnant females implantation (GD 5, GD 6) Qam blood glucose and serum
n = 3 females, n = 3 males per treatment group ^ree acids (GD 21)
Age' adult Fetal 9rowth parameters (body
weight, length, percentage lean
mass, percentage fat mass)
(GD 21)
Dam body weight gain during
pregnancy (GDs 5-7)
Dam uterine blood flow and
resistance (GDs 15, 19, 21)
Dam serum inflammatory marker
(GD 21)
BALF = bronchoalveolar lavage fluid; FA = filtered air; GD = gestation day; h = hour; HDM = house dust mite; LE = Long-Evans; n = sample size; PND = postnatal day; ppm = parts
per million.
tRecent studies evaluated since the 2013 Ozone ISA.
7-97
-------
Table 7-17 Study-specific details from studies of ozone and developmental effects.
Study
Species (Stock, Strain), n, Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
tCarev et al. (20111
Rhesus monkeys
n = 4-14
0.5 ppm, acute or chronic ozone
exposure; one cycle = 5 consecutive
days of ozone for 8 h/day and 9 days of
FA (14-day cycle); one group FA, one
group 1 cycle of ozone; one group
11 cycles of ozone beginning at
32-37 days of age, ending at 6 mo of
age; animals with one cycle ozone
exposure ended at 6 mo of age
Morphological sections of nasal
tract (anatomical development);
histology of nasal tract (epithelial
height, cilia volume, nuclear
volume, histological images for
necrosis) (24 h PE)
tChou et al. (2011)
Rhesus monkeys
*n = 6/group * 4 groups (n = 24), FA control, HDM
(house dust mite), O3, O3 + HDM males, 0 females
Age: 4-14 weeks of age
0.5 ppm, ozone: one cycle of
14 days = 8 h/day for 5 days + 9 days
filtered air; HDM: Days 3-5 of
exposure and sensitization at Day 14
or Day 28 of exposure; five cycles of
ozone exposure over the time period
Blood (WBC, eosinophils), BAL
(total cell number, eosinophils,
macrophages, PMNs,
lymphocytes; chemotaxis
proteins [CCL 11, 24, 26],
histology [eosinophil count in
airway walls]); CCL in histology
staining and CCL mRNA
quantified (3 mo of age)
tHunter et al. (20111
Rats (F344)
n = 4-6*/group males, 0 females
Age: PND 6, PND10, PND15, PND21, or PND28
2 ppm, FA or ozone, 3 h, 1 day
NGF protein and mRNA (lung
lavage, lung tissue, SP-IR nerve
fiber density in extrapulmonary
smooth muscle) (24 h PE)
tMurphv et al. (2012)
Rhesus monkeys
n = *4-6/group males, 0 females
Age: 6-12 mo of age
0.5 ppm, 11 cycles. Ozone cycle:
8 h/day for 5 days + 9 days filtered air;
HDM: Days 3-5 of exposure
Lung morphology (HE), NK1R
protein expression, IL-8 mRNA,
oxidative stress challenge to
explanted airways (12 mo of
age, after end of exposures)
7-98
-------
Table 7-17 (Continued): Study-specific details from studies of ozone and developmental effects.
Study
Species (Stock, Strain), n, Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
tGabehart et al. (2014)
Mice (BALB/cJ)
n = NR
Age: PND 3
1 ppm, 3 h/day, 1 day
BAL (cell numbers, neutrophils,
albumin), lung histology, electron
microscopy of lung, lung gene
microarray and pathways
analysis (6 or 24 h PE)
tMurphv et al. (2013)
Rhesus monkeys
n = *12/group males, 0 females
Age: 1, 2, or 6 mo
0.5 ppm
Episodic ozone (repeating cycles of
ozone for 8 h/day for 5 days followed
by 9 days of FA) from 1 to 2 and 6 mo
Acute ozone (single 8-h exposure) at
2 or 6 mo
Serotonin patterns in distal and
midlevel lung (5HT, 5HTT, and
receptors 5HT2aR or 5Ht4R);
changes in epithelial thickness (2
or 6 mo of age)
tGabehart et al. (2015)
Mice (BALB/cJ WT and TLR4-/-)
n = 0 males, 3-7/group females
Age: neonatal (PND 3), juveniles (2 weeks old),
weanlings (3 weeks old), adults (6 weeks old)
1 ppm, 3 h
Whole lung TLR4 expression by
age, BAL (neutrophils,
antioxidants, albumin leakage,
chemokines, mucus production)
(Immediately PE)
tGordon et al. (2017b)
Rats (LE)
n = 8 female pups per treatment group.
Age: Sedentary and active rats were housed in
cages with or without running wheels from PND 22
to PND 100. During the last week of the study, rats
were subjected to the ozone or filtered air.
0.25, 0.5, 1.0 ppm, 5 h/day, 2 days
Glucose tolerance testing, BALF
immune cells, metabolic function
indicators
tDveetal. (2017)
Rats (LE, S-D, Wistar)
n = 7-16 pups
Age: exposure on PNDs 14, 21, 28
1 ppm x 2 h, 1 day
Respiratory outcomes, lung
antioxidants, redox enzymes
7-99
-------
Table 7-17 (Continued): Study-specific details from studies of ozone and developmental effects.
Exposure Details
Study Species (Stock, Strain), n, Sex, Age (Concentration, Duration) Endpoints Examined
tMiller et al. (2017) Rats (LE) 0, 0.4, or 0.8 ppm, 4 h/day for 2 days BALF (GD 21)
n = 0 males, 10/group females (GD 5, GD 6) Dam blood glucose and serum
Age: adult free fattY acids (GD 21)
Dam penh, whole body
plethysmography (GDs 5 and 6)
Dam blood pressure during
pregnancy (GDs 15, 19, 21)
Dam minute volume (GD 21)
Fetal growth parameters (body
weight, length, percentage lean
mass, percentage fat mass)
(GD 21)
Dam body weight gain during
pregnancy (GDs 5-7)
Dam kidney histopathology
(GD 21)
Dam uterine blood flow and
resistance (GDs 15, 19, 21)
Dam serum inflammatory marker
(GD 21)
BALF = bronchoalveolar lavage fluid; CCL = chemokine ligands; FA = filtered air; GD = gestation day; h = hour; HDM = house dust mite; HE = hematoxylin and eosin staining;
LE = Long-Evans; mRNA = messenger ribonucleic acid; NGF = nerve growth factor; PE = post-exposure; PMNs = polymorphonuclear neutrophils; PND = postnatal day; ppm = parts
per million; S-D = Sprague Dawley; SP IR = substance P immunoreactive; WBC = white blood cells; WT = wild type.
tRecent studies evaluated since the 2013 Ozone ISA.
7-100
-------
7.5
Evidence Inventories—Data Tables to Summarize Nervous System Effects Study Details
7.5.1 Epidemiologic Studies
Table 7-18 Epidemiologic studies of short-term exposure to ozone and effects on cognition, motor activity, and
mood.
Copollutant
Effect Estimates
Study
Study Population
Exposure Assessment
Mean (ppb)
Examination
95% CI
tLimetal. (2012)
n = 560
Nearest monitor
Mean: 48.1
Correlation (r):
Factor 3—affective
Seongbuk-Gu, South Korea
Older adults
8-h max
Median: 44
PM2.5: NR;
symptoms: OR 1.07 (0.83,
Maximum: 140
NO2: -0.15;
1.38)
Ozone: 2008-2010
SO2: -0.18;
Factor 2—somatic
Follow-up: August 2008-August
2010
CO: -0.30
symptoms: OR 1.25 (0.90,
Copollutant models:
1.74)
Other study
NR
Factor 1—emotional
symptoms: OR 1.58 (1.16,
2.14)
CI = confidence interval; CO = carbon monoxide; h = hour; max = maximum; n = sample size; N02 = nitrogen dioxide; NR = not reported; OR = odds ratio; PM25 = particulate matter
with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm; S02 = sulfur dioxide.
tRecent studies evaluated since the 2013 Ozone ISA.
7-101
-------
Table 7-19 Epidemiologic studies of short-term exposure to ozone and hospital admissions, emergency
department, and outpatient visits.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
RR (95% CI)
tChiu and Yanq (2015)
Taipei, Taiwan
Ozone: 2006-2011
Follow-up: 2006-2011
Case-crossover study
n = 13,676
Random sample of
enrollees
TNHIP
6-monitor avg
24-h avg
Mean: 24.6 (all
years)
Median: 23.77
Maximum:
70.98
Correlation (r):
N02: -0.06;
SO2: 0.07;
CO: -0.22
Copollutant models:
yes
Outpatient visit (migraine),
>23°C: 1.08 (1.02, 1.15)
Outpatient visit (migraine),
<23°C: 1.28 (1.19, 1.38)
tXu et al. (2016)
Xi'an, China
Ozone: 2013-2014
Follow-up: 2013-2014
n = 20,368
outpatient visits
13-monitor avg
24-h avg
Mean: 51
Median: 44.37
Maximum:
158.61
Correlation (r):
PM2.5: -0.226;
NO2: -0.165;
SO2: -0.414; CO:
-0.454
Epilepsy outpatient visit:
0.98 (0.95, 1.00)
Time-series study
Copollutant models:
yes—NO2, SO2
tLinares et al. (2017)
Madrid, Spain
n = 1,175
HMS
27-monitor avg
24-h avg
Mean: 18.21
Maximum:
45.59
Correlation (r): NR
Copollutant models:
Dementia-related hospital
admission: 1.29 (1.12,
1.51)
Ozone: 2001-2009
NR
Follow-up: 2001-2009
Time-series study
tCulqui et al. (2017)
Madrid, Spain
Ozone: 2001-2009
n = 1,183
HMS
27-monitor avg
24-h avg
Mean: 18.21
Median: 18.21
Correlation (r): NR
Copollutant models:
NR
Alzheimer's-related hospital
admission: (NR, not
statistically significant)
Follow-up: 2001-2009
Time-series study
7-102
-------
Table 7-19 (Continued): Epidemiologic studies of short-term exposure to ozone and hospital admissions,
emergency department, and outpatient visits.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
RR (95% CI)
tGuoet al. (2018)
Guangzhou, China
Ozone: 2013-2015
Follow-up: 2013-2015
Time-series study
n=162,771
ED visits
Daily avg of 36 monitors
8-h max
Mean: 49.81
Maximum:
125.89
Correlation (r): NR
Copollutant models:
NR
ED visit for diseases of
neurological system (cold):
0.99 (0.98, 1.00)
ED visit for diseases of
neurological system
(warm): 0.99 (0.97, 1.01)
tCarmona et al. (2018)
Madrid, Spain
Ozone: 2001-2009
n = 2,224
Citywide average all monitors
24-h avg
Mean: 18.21
Maximum:
45.59
Correlation (r): NR
Copollutant models:
NR
ED visit for multiple
sclerosis: (NR, not
statistically significant)
Follow-up: 2001-2009
Time-series study
tJeaniean et al. (2018)
Strasbourg, France
Ozone: 2000-2009
Follow-up: 2000-2009
Case-crossover study
n = 1,783
Relapse occurrence in
registry
EDMUS
ADMS (Urban)
24-h avg
Mean: 44.29
Median: 42.55
Maximum:
112.71
Correlation (r):
N02: -0.06;
CO: -0.21
Copollutant models:
NR
Multiple sclerosis relapse
(cold, October-March):
0.96 (0.90, 1.03)
Multiple sclerosis relapse
(hot, April-September):
1.06 (1.03, 1.09)
+Lee et al. (2017)
Seoul, South Korea
Ozone: 2002-2013
Follow-up: 2002-2013
n = 314
NHIS-NSC
27-monitor avg
8-h avg
Mean: 24.2
Correlation (r): NR
Copollutant models:
NR
Parkinson's disease
hospital admission: (NR,
figure only: no statistically
significant associations)
Case-crossover study
tChoet al. (2015)
Seoul, South Korea
Ozone: 2005-2009
n = 2,320
HIRA
27-monitor avg
24-h avg
Mean: 18
Median: 16
90th: 31
Correlation (r): PM2.5:
NR
Copollutant models:
NR
ED visit for panic attack:
1.11 (1.05, 1.17)
Follow-up: 2005-2009
Time-series study
7-103
-------
Table 7-19 (Continued): Epidemiologic studies of short-term exposure to ozone and hospital admissions,
emergency department, and outpatient visits.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
RR (95% CI)
tChen etal. (2018)
Shanghai, China
Ozone: 2013-2015
Follow-up: 2013-2015
Time-series study
n = 39,143
SHIS
10 monitoring stations
8-h max
Mean: 51
Median: 48.96
Maximum:
135.66
Correlation (r):
PM2.5: -0.03;
NO2: -0.21;
SO2: -0.2; CO:
-0.22
Copollutant models:
NR
Mental disorder hospital
admission: 1.01 (0.96,
1.07)
Mental disorder hospital
admission: 1.03 (0.95,
1.12)
tOudin etal. (2018)
Gothenburg, Sweden
Ozone: NR
Follow-up: July 2012-November
2016
Case-crossover study
n = NR
One monitor
24-h avg
Mean: 25.65
Median: 25.6
Maximum:
79.05
Correlation (r): NR
Copollutant models:
yes
ED visit for psychiatric
emergencies (cold,
October-March): 0.99
(0.95, 1.03)
ED visit for psychiatric
emergencies (all year):
1.00 (0.98, 1.03)
ED visit for psychiatric
emergencies (warm,
April-September): 1.02
(0.99, 1.05)
tSzvszkowicz et al. (2016)
Nine urban areas, Canada
Ozone: April 2004-December
2011
Follow-up: April 2004-December
2011
Case-crossover study
n = 118,602
NACRS
Daily average of monitors
within 3 km of postal code
24-h avg
Mean:
22.5-29.2
Maximum:
60.7-80.0
Correlation (r): NR
Copollutant models:
NR
ED visit for depression (all
year, male): 1.00 (0.98,
1.03)
ED visit for depression
(warm, April-September,
male): 1.01 (0.98, 1.04)
ED visit for depression (all
year, female): 1.01 (0.99,
1.03)
ED visit for depression
(warm, April-September,
female): 1.03 (1.00, 1.05)
ADMS = Atmospheric Dispersion Modelling System; Avg = average; C = Celsius; CO = carbon monoxide; CI = confidence interval; h = hour; ED = emergency department;
EDMUS = European Database for Multiple Sclerosis; HIRA = Health Insurance Review and Assessment Service; HMS = Hospital Morbidity Survey; max = maximum; km = kilometer;
OR = odds ratio; max = maximum; n = sample size; NACRS = National Ambulatory Care Reporting System; NHIS NSC = National Health Insurance Service National Sample Cohort;
N02 = nitrogen dioxide; NR = not reported; PM25 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm; SHIS = Shanghai Health Insurance
System; S02 = sulfur dioxide; TNHIP = Taiwan National Health Insurance Program.
tRecent studies evaluated since the 2013 Ozone ISA.
7-104
-------
Table 7-20 Epidemiologic studies of long-term exposure to ozone and cognitive/behavioral effects.
Study Copollutant Effect Estimates
Study Population Exposure Assessment Mean (ppb) Examination 95% CI
tGatto et al. (2014)
Los Angeles, U.S.
Ozone: 2000-2006
Follow-up: 2000-2006
Cross-sectional study
n = 1,496
women
3 RCTs
BVAIT, WISH,
ELITE
Monitor within 5 km of Mean: NR—geographic
residence, average of variability in
monitors within 100 km concentration shown in
(IDW)
8-h max
Figure 2
Maximum: >25
Correlation (r):
PM2.5: -0.62;
NO2: -0.77;
Copollutant models:
NR
Global cognition (>49 ppb vs. <34
[reference]): (3 = -0.08 (-0.45, 0.28)
Semantic memory (>49 ppb vs. <34
[reference]): (3 = -0.12 (-0.5, 0.26)
Verbal learning (34-49 ppb vs. <34
[reference]): (3 = -0.13 (-0.41, 0.16)
Visual processing (34-49 ppb vs. <34
[reference]): (3 = -0.18 (-0.43, 0.07)
Verbal learning (>49 ppb vs. <34
[reference]): (3 = -0.2 (-0.63, 0.23)
Visual processing (>49 ppb vs. <34
[reference]): (3 = -0.2 (-0.59, 0.18)
Executive function (34-49 ppb vs. <34
[reference]): (3 = -0.23 (-0.68, 0.22)
Executive function (>49 ppb vs. <34
[reference]): (3 = -0.66 (-1.35, 0.03)
Visual memory (>49 ppb vs. <34
[reference]): (3 = 0.01 (-0.42, 0.44)
Global cognition (34-49 ppb vs. <34
[reference]): (3 = 0.05 (-0.19, 0.29)
Semantic memory (34-49 ppb vs. <34
[reference]): (3 = 0.08 (-0.17, 0.33)
Visual memory (34-49 ppb vs. <34
[reference]): (3 = 0.12 (-0.16, 0.4)
Logical memory (>49 ppb vs. <34
[reference]): (3 = 0.24 (-0.21, 0.68)
Logical memory (34-49 ppb vs. <34
[reference]): (3 = 0.31 (0.01, 0.6)
7-105
-------
Table 7-20 (Continued): Epidemiologic studies of long-term exposure to ozone and cognitive/behavioral effects.
Study
Study
Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% CI
tClearv et al. (2018)
Nation-wide, U.S.
Ozone: 2001-2008
Follow-up: 2004-2008
N = 5,116
NACC
Alzheimer's
Disease Center
participants
(male and
female)
HBM to combine
monitored and
predicted (CMAQ)
concentrations
8-h max
Mean: NR (figure only)
Correlation (r): NR
Copollutant models:
NR
Cognitive decline on MMSE and
CDR-SB associated with ozone
exposure among those with no baseline
impairment
tKioumourtzoqlou et al.
(2017)
48 continental states, U.S.
Ozone: 1996-2008
Follow-up: 1996-2008
Cohort study
n = 41,844
Women
NHS
Summer average
(May-Sept) of up to
five monitors (at least
one monitor within
50 km (IDW), at
residential address
Mean: 31.9
Correlation (r): NR
Copollutant models:
NR
Depression onset (depression
diagnosis): HR 1.00 (0.92, 1.08)
Depression onset (antidepressant or
depression): HR 1.06 (1.00, 1.12)
Depression onset (use of
antidepressant medication): HR 1.08
(1.02, 1.14)
avg = average; BVAIT = B-Vitamin Atherosclerosis Intervention Trial; CDR-SB = Clinical Dementia Rating Scale-Sum of Boxes; CI = confidence interval; CMAQ = Community
Multiscale Air Quality; ELITE = Early Versus Late Intervention Trial with Estradiol; h = hour; HBM = hierarchical Bayesian model; HR = hazard ratio; IDW = inverse-distance
weighting; km = kilometer; max = maximum; MMSE = Mini-Mental State Exam; n = sample size; NACC = National Alzheimer's Coordinating Center; NHS = Nurses' Health study;
N02 = nitrogen dioxide; NR = not reported; PM25 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm; RCT = randomized controlled trial;
WISH = Women's Isoflavone Soy Health.
tRecent studies evaluated since the 2013 Ozone ISA.
7-106
-------
Table 7-21 Epidemiologic studies of long-term exposure to ozone and neurodegenerative diseases.
Study
Study Population
Exposure Assessment Mean (ppb)
Copollutant
Examination
Effect Estimates
95% CI
tJuna et al. (2014)
National, Taiwan
Ozone: 2000-2010
Follow-up: 2000-2010
Case-control study
n = 97,627
LHID2000-NHIRD
Annual avg of three
nearest monitors within
25 km of grid cell (IDW),
assigned to postal code
of residence
8-h max
Mean: 92.64 Correlation (r):
Maximum:
137.65
N02: -0.05;
SO2: 0.01;
CO: 0.10;
PM10: -0.26
Copollutant models:
yes
Alzheimer's disease
Baseline ozone
concentration:
1.06 (1.00, 1.12)
1.10 (1.03, 1.18) + CO
1.06 (0.99, 1.14) + NO2
1.04 (0.98, 1.11) + SO2
Change in ozone
concentration at follow-up
minus concentration at
baseline: 2.84 (2.67, 3.01)
3.07 (2.87, 3.28) + CO
3.09 (2.90, 3.28) + NO2
3.17 (2.98, 3.37) + SO2
2.93 (2.75, 3.12) + PM10
tKirrane et al. (2015)
North Carolina: n = 104 cases; Annual, seasonal
North Carolina Correlation (r):
North Carolina and Iowa, U.S. 29,612 controls
Ozone: 2002-2006
Enrollment: 1993-2005
Follow-up: 1997-2010
Case-control study
Iowa: n = 195 cases;
53,024 controls
Farmer pesticide applicators
and their spouses
AHS
(April-October) 4-yr avg Mean: 40.6
of daily predictions |owa
using measured
concentrations and
CMAQ
8-h max
Mean: 39.0
PM2.5: -0.15 to 0.06,
depending on metric
Copollutant models:
NR
Parkinson's disease (Iowa
4-yr avg): OR 0.46 (0.13,
1.69)
Parkinson's disease (Iowa
warm season average): OR
0.46 (0.11, 1.84)
Parkinson's disease (North
Carolina 4-yr avg): OR 1.49
(0.43, 5.16)
Parkinson's disease (North
Carolina warm season
average): OR 2.60 (0.94,
7.24)
7-107
-------
Table 7-21 (Continued): Epidemiologic studies of long-term exposure to ozone and neurodegenerative diseases.
Study
Study Population
Exposure Assessment Mean (ppb)
Copollutant
Examination
Effect Estimates
95% CI
tChen etal. (2017a)
National, Taiwan
Ozone: 2000-2013
Follow-up: 2000-2013
Case-control study
n = 249 cases; 497 controls
<40 yr
TNHIP-NHIRD
Monthly average during
follow-up in areas
where participants
reside
Mean: NR
Correlation (r): NR
Copollutant models:
NR
Parkinson's disease
HR: 1.10 (0.74, 1.48)
tChen etal. (2017c)
Ontario, Canada
Ozone: 1994-2013
Follow-up: April 2001-March
2013
Cohort study
n = 2,066,639
55-85 yr at enrollment
ONPHEC
5-yr avg estimated
using monitor
concentrations with
physically based air
quality prediction model
(2002-2009) calibrated
using data from
1995-2013
Mean: 45.8
Correlation (r): PM2.5:
0.38; NO2: -0.22
Copollutant models:
NR (multipollutant
only)
Dementia
HR: 0.97 (0.94, 1.00)
tWu etal. (2015)
Multicity, Taiwan
Ozone: 2007-2010
Follow-up: 2007-2010
Case-control study
n = 1,060 cases; 4,240 controls
>60 yr
Hospitals and clinics
Spatiotemporal model,
cumulative annual
average
Mean: NR
Correlation (r): NR
Copollutant models:
NR
Alzheimer's disease
(20.20-21.56 vs. <20.20
[reference])
OR: 0.60 (0.33, 1.09)
Vascular dementia
(>20.20-21.56 vs. <20.20
[reference])
OR: 0.62 (0.28, 1.38)
Alzheimer's disease
(>21.56 vs. <20.20 ppb
[reference])
OR: 2.00 (1.14, 3.50)
Vascular dementia (>21.56
vs. <20.20 ppb [reference])
OR: 2.09 (1.01, 4.33)
7-108
-------
Table 7-21 (Continued): Epidemiologic studies of long-term exposure to ozone and neurodegenerative diseases.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% CI
+Lee etal. (2016)
National, Taiwan
Ozone: 1998-2009
Follow-up: 2007-2009
Case-control study
n = 11,117 cases; 4 to 1 match
NHIRD
First clinic visit for PD (patients
>35 yr)
QBME spatio-temporal
model
Mean 26.1
Correlation (r):
SO2: 0.01;
CO: -0.60
Copollutant models:
NR
Parkinson's disease
OR: 1.00 (0.93, 1.07),
comparing the highest of
exposure to the lowest
quartile
tShin etal. (2018)
Ontario, Canada
Follow-up: 2001-2013
1994-2013
n = 38,745 cases (-2.2 million
followed)
55+ yr old
Registry record for Parkinson's
disease healthcare or
medication
ONPHEC
Summer average,
fused-based optimal
interpolation of
measured and predicted
ozone, 21 * 21 km grid
8 h max
Mean: 49.8
Correlation (r): NR
Copollutant models:
NR
Parkinson's disease
HR: 1.06 (1.02, 1.11)
tCerza etal. (2018)
Rome Italy
Ozone: 2008
Follow-up: 2008-2013
Cohort study
n = 1,008,253
50+ yr
Insurance registry claim for
Parkinson's disease
Regional Health Information
System
Summer average,
chemical dispersion
model with grid
resolution of 1 * 1 km
8-h avg
Mean: 45.5
Correlation (r): NR
Copollutant models:
yes
Parkinson's disease
HR: 1.04 (1.00, 1.11)
HR: 1.03 (0.99,
1.06) + NO2
AHS = Agricultural Health Study; avg = average; CI = confidence interval; CMAQ = Community Multiscale Air Quality; h = hour; HR = hazard ratio; IDW = inverse-distance weighting;
km = kilometer; LHID2000 = Longitudinal Health Insurance Database 2000; n = sample size; NHIRD = National Health Insurance Research Database; N02 = nitrogen dioxide;
NR = not reported; ONPHEC = Ontario Population Health and Environment Cohort; OR = odds ratio; PD = Parkinson's disease; PM25 = particulate matter with a nominal mean
aerodynamic diameter less than or equal to 2.5 |jm; PM10 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 10 |jm; QBME = quantile-based
Bayesian maximum entropy; S02 = sulfur dioxide; TNHIP = Taiwan National Health Insurance Program; yr = year.
tRecent studies evaluated since the 2013 Ozone ISA.
7-109
-------
Table 7-22 Epidemiologic studies of long-term exposure to ozone and neurodevelopmental effects.
Study
Study Population
Exposure Assessment Mean (ppb)
Copollutant
Examination
Effect Estimates
95% CI
tBecerra et al. (2013)
Los Angeles County, U.S
Ozone: 1995-2006
Follow-up: 1998-2009
Case-control study
n = 7,594 cases;
75,635 controls
Diagnosis between 36 and
71 mo
DDS registry
Nearest monitor,
trimester and whole
pregnancy averages
8-h avg
Mean: 36.8
Correlation (r):
PM2.5: -0.47;
NO2: -0.50; CO:
-0.55, PM10:
-0.17, NO: -0.73
Copollutant models:
yes
Autism disorder (ORs):
1.05 (1.01, 1.10)
1.07 (1.03, 1.12) + NO2
1.05 (1.01, 1.10) + PM10
1.10 (1.05, 1.16) + PM2.5
tJunq et al. (2013)
National, Taiwan
Ozone: 2000-2010
Follow-up: 2000-2010
Cohort study
n = 49,073
ASD
LHID2000-NHIRD
Annual average of three
nearest monitors within
25 km of grid cell (IDW),
assigned to postal code
of residence
8-h max
Mean: 90-120
depending on
season
Correlation (r):
NO2: 0;
SO2: 0.22;
PM10: 0.65
Copollutant models:
yes
ASD (HRs):
1.59 (1.42, 1.78)
1.57 (1.40, 1.77) + CO
1.53 (1.35, 1.73) + NO2
1.54 (1.37, 1.73) + SO2
tKerin et al. (2017)
California, U.S.
Ozone: NR
Follow-up:
Case-control study
n = 325 cases born
1999-2007
Diagnosed with ASD
between 24 and 60 mo
CHARGE
Pregnancy, yr 1
average of up to four
monitors within 50 km
(IDW) or one monitor
within 5 km
8-h max
Mean: 37.3
Correlation (r):
PM2.5: -0.21;
NO2: -0.45;
PM10: 0.04
Copollutant models:
NR
ADOS-CSS:
0.99 (0.94, 1.05)
MSEL: 0.99 (0.88, 1.10)
VABS: 1.00 (0.96, 1.04)
tVolket al. (2013)
California, U.S.
Ozone: 1997-2009
Follow-up: 1997-2008
Case-control study
n = 534 cases and controls 1st yr, entire pregnancy, Mean: NR
Diagnosed with autism
between 24 and 60 mo
Full-syndrome autism
(ADOS and autism
diagnostic
intervi ew—revi sed.)
CHARGE
1st trimester, 2nd
trimester, 3rd trimester
Average of four closest
monitors within 50 km
(IDW) or one monitor
within 5 km
8-h max
Correlation (r): NR
Copollutant models:
NR
Autism:
1.05 (0.84, 1.31), entire
pregnancy
7-110
-------
Table 7-22 (Continued): Epidemiologic studies of long-term exposure to ozone and neurodevelopmental effects.
Study
Study Population
Exposure Assessment Mean (ppb)
Copollutant
Examination
Effect Estimates
95% CI
tKimetal. (2017)
California, U.S.
Follow-up: 1999-2008
Ozone: 1997-2009
Case-control study
n = 158 cases; 147 controls Pregnancy, 1st yr, 2nd Mean: NR
Diagnosed with autism
between 24 and 60 mo
CHARGE
Confirmed autism
yr
Average of four closest
monitors within 50 km
(IDW) or one monitor
within 5 km
8-h max
Correlation (r): NR
Copollutant models:
NR
Joint effect of copy
number and ozone
greater than effect of
each ozone or
duplication burden alone
tGoodrich et al. (2017)
California, U.S.
Follow-up: 1999-2008
Ozone: 1997-2009
Case-control study
n = 297 confirmed autism; 1st trimester
143 ASD; 326 controls
Diagnosed with autism
between 24 and 60 mo
CHARGE
Mean: NR
Average of four closest
monitors within 50 km
(IDW) or one monitor
within 5 km
8-h max
Correlation (r): NR
Copollutant models:
NR
No interaction between
ozone exposure and folic
acid intake
tLin et al. (2014a)
11 towns, Taiwan
Follow-up: October 2003-January
2004
Cohort study
n = 511 mother-infant pairs,
neurodevelopment
assessed by parent report
1st, 2nd, 3rd trimester
monitor average
Mean: NR
Correlation (r): NR
Copollutant models:
NR
No associations with
ozone reported
ADOS-CSS = Autism Diagnostic Observation Schedule derived-Calculated Severity Score; AHS = Agricultural Health Study; ASD = autism spectrum disorder; avg = average;
BVAIT = B-Vitamin Atherosclerosis Intervention Trial; CHARGE = Childhood Autism Risks from Genetics and the Environment; CMAQ = Community Multiscale Air Quality;
DDS = Department of Developmental Services; DSM-IV-R = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition, Text Revision; EDMUS = European Database for
Multiple Sclerosis; ELITE = Early Versus Late Intervention Trial with Estradiol; h = hour; HBM = hierarchical Bayesian model; HIRA = Health Insurance Review and Assessment
Service; HMS = Hospital Morbidity Survey; IDW = inverse-distance weighting; LHID2000 = Longitudinal Health Insurance Database 2000; MSEL = Mullen Scales of Early Learning;
n = sample size; NACRS = National Ambulatory Care Reporting System; NHIRD = National Health Insurance Research Database; NHIS-NSC = National Health Insurance Service-
National Sample Cohort; NHS = Nurses' Health Study; N02 = nitrogen dioxide; ONPHEC = Ontario Population Health and Environment Cohort; PM2.5 = particulate matter with a
nominal mean aerodynamic diameter less than or equal to 2.5 |jm; SHIS = Shanghai Health Insurance System; S02 = sulfur dioxide; TNHIP = Taiwan National Health Insurance
Program; VABS = Vineland Adaptive Behavior Scales; WISH = Women's Isoflavone Soy Health.
tRecent studies evaluated since the 2013 Ozone ISA.
7-111
-------
7.5.2
Toxicological Studies
Table 7-23 Study-specific details from short-term studies of brain inflammation and morphology.
Study
Population
Species (Strain), N, Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Chounlamountrv et al. (2015)
Rats (Wistar)
n = 3-5 males, 0 females
Age: 6-7 weeks
2 ppm, 24 h, single exposure
Glial remodeling; markers of astrocyte
activation (GFAP, S100b, GLT1, Glyn Syn,
ezrin) (immediately PE)
Gomez-Crisostomo et al. (2014)
Rats (Wistar)
n = 12 males, 0 females
Age: NR (250-300 g)
0.25 ppm, 4 h/day for 7, 15, 30,
60, or 90 days
Markers of oxidative stress and apoptosis
(pFoxO 3a/1a, Mn SOD, Cyclin D2,
caspase 3) (24 h PE)
Gonzalez-Guevara et al. (2014)
Rats (Wistar)
n = 3 males, 0 females
Age: NR (250-300 g)
1 ppm, 1, 3, 6 h (single exposure);
1 h/day or 3 h/day for 5 days
Markers of inflammation (TNF-a, IL-6, NF-kB,
GFAP) (immediately PE)
Fernando Hernandez-Zimbron and
Rivas-Arancibia (2016)
Rats (Wistar)
n = 3-6 males, 0 females
Age: NR (250-300 g)
0.25 ppm, 4 h/day for 7, 15, 30,
60, or 90 days
p-Amyloid accumulation in the endoplasmic
reticulum (2 h PE)
Hernandez-Zimbron and Rivas-
Arancibia (2015)
Rats (Wistar)
n = 6 males, 0 females
Age: NR (250-300 g)
0.25 ppm, 4 h/day for 7, 15, 30,
60, or 90 days
p-Amyloid accumulation in the endoplasmic
reticulum (2 h PE)
Pinto-Almazan et al. (2014)
Rats (Wistar)
n = 10 males, 0 females
Age: NR (250-300 g)
0.25 ppm, 4 h/day for 7, 15, 30, or
60 days
Markers of oxidative stress (NT, 4-HNE); loss
of pyramidal neurons in the hippocampus
(PE)
Rivas-Arancibia et al. (2015)
Rats (Wistar)
n = 6 males, 0 females
Age: NR (250-300 g)
0.25 ppm, 4 h/day for 7, 15, 30,
60, or 90 days
Markers of oxidative stress and inflammation
(lba-1, NF-kB, GFAP, COX-2); mitochondrial
dysfunction and cell loss in substantia nigra
7-112
-------
Table 7-23 (Continued): Study-specific details from short-term studies of brain inflammation and morphology.
Study
Population
Species (Strain), N, Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Rivas-Arancibia et al. (2017)
Rats (Wistar)
n = 6 males, 0 females
Age: NR (250-300 g)
0.25 ppm, 4 h/day for 7, 15, 30,
60, or 90 days
p-Amyloid structure (2 h PE)
Rodriquez-Martinez et al. (2013)
Rats (Wistar)
n = 6 males, 0 females
Age: NR (250-300 g)
0.25 ppm, 4 h/day for 7, 15, 30, or
60 days
Protein oxidation; antioxidant activity (SOD,
GPx); mitochondrial dysfunction and cell
damage in hippocampus
Mokoena et al. (2011)
Rats (S-D)
n = 7-9 males, 0 females
Age: NR (270-310 g)
0.25 or 0.7 ppm, 4 h
Or 0.25 ppm, 4 h/day for 30 days
Lipid peroxidation/superoxide formation in
frontal cortex (PE)
Mokoena et al. (2015)
Rats (FRL or FSL)
n = 8-12 males, 0 females
Age: NR (230-250 g)
0.3 ppm, 4 h/day for 15 days
Lipid peroxidation; antioxidant (SOD, CAT)
activity (PE)
Mumaw et al. (2016)
Rats (S-D)
n = 7-9 males, 0 females
Age: 8 weeks
1 ppm, 4 h
Microglial activation; markers of inflammation
(TNF-a, IL-1|3) (24 h PE)
Tvler et al. (2018)
Mice (C57BL/6)
n = 3-11 males, 0 females
Age: adult (8-10 weeks); aged
(12-18 mo)
1 ppm, 4 h
Blood-brain barrier permeability/infiltration;
microglial activation; B-amyloid accumulation
(20 h PE)
Rodiiquez-Martfnez et al. (2016)
Rats (Wistar)
n = 6 males, 0 females
Age: NR (250-300 g)
0.25 ppm, 4 h/day for 7, 15, 30, or
60 days
Endoplasmic reticulum dysfunction and cell
death in the hippocampus
7-113
-------
Table 7-24 Study-specific details from short-term studies of cognitive and behavioral effects.
Study
Population
Species (Strain), N, Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Bhoopalan et al. (2013)
Rats (S-D)
n = 6 males, 0 females
Age: 9-10 weeks
0.8 ppm, 3 h, single exposure
Dopamine levels in the striatum (-24 h PE)
Pinto-Almazan et al. (2014)
Rats (Wistar)
n = 10 males, 0 females
Age: NR (250-300 g)
0.25 ppm, 4 h/day for 7, 15, 30, or 60 days
Passive avoidance, motor activity (2 or 24 h
PE)
Mokoena et al. (2015)
Rats (FRL or FSL)
n = 8-12 males, 0 females
Age: NR (230-250 g)
0.3 ppm, 4 h/day, 15 days
Novel object recognition, motor activity,
forced swim test, elevated plus maze
(immediately PE)
Gordon et al. (2016)
Rats (BN)
n = 9-10 males, 9-10 females
Age: adult (-20 weeks)
0.8 ppm, 5 h/day, 1 day/week, 4 weeks
Motor activity (measured after 1-week
exposure)
4-HNE = 4-hydroxynoneal; BN = brown Norway; CAT = catalase; COX-2 = cyclooxygenase-2; FRL = Flinders resistant line; FSL = Flinders sensitive line; g = gram; GFAP = glial
fibrillary acidic protein; GLT-1 = glutamate transporter-1; GPx = glutathione peroxidase; h = hour; Iba1 = Ionized calcium binding adaptor molecule 1; IL-1 (3 = interleukin 1 beta;
Mn = manganese; n = sample size; NF-kB = nuclear factor kappa-light-chain-enhancer of activated B cells; NR = not reported; NT = nitrotyrosine; PE = post-exposure;
pFoxO = phosphorylated form of forkhead box O; ppm = parts per million; S100b = S100 calcium-binding protein B; S-D = Sprague Dawley; SOD = superoxide dismutase;
TNF-a = tumor necrosis factor alpha.
7-114
-------
Table 7-25 Study-specific details from short-term studies of neuroendocrine effects.
Population Exposure Details
Study Species (Strain), N, Sex, Age (Concentration, Duration) Endpoints Examined
Thomson et al. (2013)
Rats (F344)
n = 4-6 males, 0 females
Age: NR (200-250 g)
0.4 or 0.8 ppm, 4 h
Expression of genes related to
antioxidant response, xenobiotic
metabolism, inflammation, and
endothelial dysfunction; adrenal
hormones levels in serum
g = gram; h = hour; n = sample size; NR =
not reported.
Table 7-26 Study-specific details from long-term studies of brain inflammation and morphology.
Study
Population
Species (Strain), N, Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Akhteretal. (2015)
Mice (C57BL/6 APP+/PS1+ or WT)
n = 5-7 males, 5-7 females
Age: 6 weeks
0.8 ppm, 16 weeks (eight cycles:
7 h/day for 5 days, 9 days FA only)
Lipid/protein oxidation, antioxidant levels,
cell death in the hippocampus
Fernando Hernandez-Zimbron and
Rivas-Arancibia (2016)
Rats (Wistar)
n = 3-6 males, 0 females
Age: NR (250-300 g)
0.25 ppm, 4 h/day for 7, 15, 30, 60,
90 days
or
p-Amyloid in the endoplasmic reticulum
(2 h PE); p-amyloid accumulation
Gomez-Crisostomo et al. (2014)
Rats (Wistar)
n = 12 males, 0 females
Age: NR (250-300 g)
0.25 ppm, 4 h/day for 7, 15, 30, 60,
90 days
or
Markers of oxidative stress and apoptosis
(pFoxO 3a/1a, Mn SOD, Cyclin D2,
caspase 3) (24 h PE)
Hernandez-Zimbron and Rivas-
Arancibia (2015)
Rats (Wistar)
n = 6 males, 0 females
Age: NR (250-300 g)
0.25 ppm, 4 h/day for 7, 15, 30, 60,
90 days
or
p-amyloid accumulation in the endoplasmic
reticulum (2 h PE)
7-115
-------
Table 7-26 (Continued): Study-specific details from long-term studies of brain inflammation and morphology.
Study
Population
Species (Strain), N, Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Pinto-Almazan et al. (2014)
Rats (Wistar)
n = 10 males, 0 females
Age: NR (250-300 g)
0.25 ppm, 4 h/day for 7, 15, 30, or
60 days
Lipid/protein oxidation; loss of pyramidal
neurons in the hippocampus (PE)
Mokoena et al. (2011)
Rats (S-D)
n = 7-9 males, 0 females
Age: NR (270-310 g)
0.25 or 0.7 ppm, 4 h or 0.25 ppm,
4 h/day for 30 days
Lipid peroxidation/superoxide formation in
frontal cortex (PE)
Rivas-Arancibia et al. (2015)
Rats (Wistar)
n = 6 males, 0 females
Age: NR (250-300 g)
0.25 ppm, 4 h/day for 7, 15, 30, 60, or
90 days
Markers of oxidative stress and
inflammation (lba-1, NF-kB, GFAP,
COX-2); mitochondrial dysfunction and cell
loss in substantia nigra
Rivas-Arancibia et al. (2017)
Rats (Wistar)
n = 6 males, 0 females
Age: NR (250-300 g)
0.25 ppm, 4 h/day for 7, 15, 30, 60, or
90 days
p-Amyloid structure (2 h PE)
Rodriauez-Martinez et al. (2013)
Rats (Wistar)
n = 6 males, 0 females
Age: NR (250-300 g)
0.25 ppm, 4 h/day for 7, 15, 30, or
60 days
Protein oxidation; antioxidant activity
(SOD, GPx); mitochondrial dysfunction and
cell damage in hippocampus
Rodriauez-Martinez et al. (2016)
Rats (Wistar)
n = 6 males, 0 females
Age: NR (250-300 g)
0.25 ppm, 4 h/day for 7, 15, 30, or
60 days
Endoplasmic reticulum dysfunction and cell
death in the hippocampus
COX-2 = cyclooxygenase-2; FA = filtered air; g = gram; GFAP = glial fibrillary acidic protein; GPx = glutathione peroxidase; h = hour; Iba1 = Ionized calcium binding adaptor molecule
1; n = sample size; NR = not reported; NF-kB = nuclear factor kappa-light-chain-enhancer of activated B cells; PE = post-exposure; pFoxO = phosphorylated form of forkhead box O;
ppm = parts per million; S-D = Sprague-Dawley; SOD = superoxide dismutase; WT = wild type.
7-116
-------
Table 7-27 Study-specific details from long-term studies of cognitive and behavioral effects.
Study
Population
Species (Strain), N, Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Akhteretal. (2015)
Mice (C57BL/6 APP+/PS1+ or WT)
n = 5-7 males, 5-7 females
Age: 6 weeks
0.8 ppm, 16 w/4 mo (eight cycles:
7 h/day for 5 days, 9 days FA only)
Swim maze, elevated plus maze, motor activity
Pinto-Almazan et al. (2014)
Rats (Wistar)
n = 10 males, 0 females
Age: NR (250-300 g)
0.25 ppm, 4 h/day for 7, 15, 30, or
60 days
Passive avoidance, motor activity (2 or 24 h PE)
Gordon et al. (2014)
Rats (BN)
n = 5-6 males, 0 females
Age: Adult (4 mo), senescent (20 mo)
1 ppm, 6 h/day, 2 days/week,
13 weeks
Motor activity
Gordon et al. (2013)
Rats (BN)
n = 7-8 males, 0 females
Age: Adult (4 mo), senescent (20 mo)
0.8 ppm, 6 h/day, 1 day/week,
17 weeks
Motor activity
BN = brown Norway; FA = filtered air; g = gram; h = hour; n = sample size; NR = not reported; PE = post-exposure; ppm = parts per million; WT = wild type.
7-117
-------
Table 7-28 Study-specific details of long-term exposures and neurodevelopmental effects.
Population Exposure Details
Study Species (Strain), N, Sex, Age (Concentration, Duration)
Endpoints Examined
Hunter et al. (2011) Rats (NR) 2 dditi. 3 h
n = 4-5
Sex NR
Age: PNDs 6, 10, 15, 21, 28
Lung innervation, NGF
production
Zellneretal. (2011) Rats (F344) 2 dditi. 3 h
n = 3, ganglia weight; n = 6-13, nerve cell count,
males and females combined
Age: PNDs 10, 15, 21, 28
Neurodevelopment (nodose and
jugular sensory ganglion)
(5-23 days PE); lung innervation
h = hour; n = sample size; NR = not reported; NGF = nerve growth factor; PE = post-exposure; PND = postnatal day; ppm = parts per million.
7.6 Evidence Inventories—Data Tables to Summarize Cancer Study Details
7.6.1 Epidemiologic Studies
7-118
-------
Table 7-29 Epidemiologic studies of long-term exposure to ozone and cancer incidence.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
95% CI
tHvstad etal. (2013)
Nationwide, Canada
Ozone: 1975-1994
Follow-up: 1994-1997
Case-control study
n = 2,390 cases
NECSS
Spatio-temporal model;
25 x 25 km;
May-September; includes
residential history
Mean: 20.3
Maximum: 33.8
Correlation (r):
PM2.5: 0.25;
NO2: 0.11
Copollutant models:
NR
OR for all lung cancers: 1.09
(0.85, 1.37)
OR for large cell lung cancer:
0.89 (0.57, 1.38)
OR for adenocarcinoma: 1.04
(0.74, 1.44)
OR for small cell lung cancer:
1.07 (0.65, 1.75)
OR for squamous cell lung
cancer: 1.19 (0.82, 1.71)
tGuoet al. (2016)
Nationwide, China
Ozone: 1990-2005
Follow-up: 1990-2009
Cohort study
n = 368,762 lung
cancer cases
30+ yr old
Hybrid model from Global
Burden of Disease
Mean: 56.9
Median: 56.8
75th: 60.5
Maximum: 76.i
Correlation (r): NR
Copollutant models:
NR
Lung cancer incidence—all: 1.09
(1.08, 1.1)
Lung cancer incidence—female:
1.08 (1.07, 1.09)
Lung cancer incidence—males:
1.09 (1.08, 1.1)
Lung cancer incidence—ages
30-65: 1.08 (1.07, 1.09)
Lung cancer incidence—ages
65-75: 1.12 (1.11, 1.13)
Lung cancer incidence—ages
75+: 1.1 (1.08, 1.12)
CI = confidence interval; n = sample size; NECSS = National Enhanced Cancer Surveillance System; N02 = nitrogen dioxide; NR = not reported; OR = odds ratio; PM2.5 = particulate
matter with a nominal mean aerodynamic diameter less than or equal to 2.5 |jm; ppb = parts per billion; yr = year.
tRecent studies evaluated since the 2013 Ozone ISA.
7-119
-------
Table 7-30 Epidemiologic studies of ozone exposure and lung cancer mortality.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
HR (95% CI)
tCarev et al. (2013)
Nationwide, U.K.
Ozone: 2002
Follow-up: 2003-2007
Cohort study
n = 835,607
Age: adults, 40-89 yr,
from English medical
practices
English Medical
Practice
Annual mean estimates from
dispersion model for 1-km grid
cells linked to nearest
residential postal code
centroid
Mean: 25.85
Maximum: 31.
Correlation (r):
PM25: -0.39;
NO2
SO2
-0.46;
-0.41;
PM10: -0.40
Copollutant models:
NR
Lung cancer: 0.66 (0.50,
0.87)
tJerrett et al. (2013)
California, U.S.
Ozone: 1988-2002
Follow-up: 1982-2000
Cohort study
n = 73,711
ACS
Monthly averages calculated Mean: 50.35
from IDW from up to four
monitors within 50 km of
residence
Median: 50.8
75th: 61
90th: 68.56
95th: 74.18
Maximum: 89.33
Correlation (r):
PM2.5: 0.56;
NO2: -0.0071
Copollutant models:
PM2.5; NO2
Lung cancer: 0.94 (0.89,
1.00)
Lung cancer (+ PM2.5):
0.93 (0.87, 0.98)
Lung cancer (+ NO2):
0.95 (0.89, 1.01)
tCrouse et al. (2015)
Nationwide, Canada
Ozone: 2002-2009
Follow-up: 1991-2006
Cohort study
n = 2,521,525
Age: 25+ yr
CanCHEC
Model of warm season Mean: 39.6
concentration at21-km Median: 39
horizontal resolution assigned 75th: 44.2
at postal code Maximum: 60
8-h max
Correlation (r):
PM2.5: 0.73;
NO2: 0.19
Copollutant models:
NR
Lung cancer: 1.01 (0.99,
1.02)
tTurner et al. (2016)
Nationwide, U.S.
Ozone: 2002-2004
Follow-Up: 1982-2004
Cohort study
n = 669,046
Age: 35+ yr
ACS
HBM with inputs from
NAMS/SLAMS and CMAQ;
downscaler for the eastern
U.S.
8-h max
Mean: 38.2
Median: 38.1
75th: 40.1
95th: 45
Maximum: 59.3
Correlation (r):
PM2.5: 0.18;
NO2: -0.08
Copollutant models:
PM2.5
Lung cancer: 0.96 (0.91,
1.00)
tCakmak et al. (2018)
Nationwide, Canada
Ozone: 2002-2009
Follow-up: 1991-2011
Cohort study
n = 2,291,250
Age: 25+ yr
CanCHEC
Model of warm season Mean: 15.0-
concentration at 21-km Maximum:
horizontal resolution assigned 46.6-60.6
at postal code
8-h max
¦43.0 Correlation (r):
PM2.5: -0.705
Copollutant models:
PM2.5
Lung cancer: 1.05 (0.97,
1.13)
Lung cancer (+ PM2.5):
1.01 (0.93, 1.09)
7-120
-------
Table 7-30 (Continued): Epidemiologic studies of ozone exposure and lung cancer mortality.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
HR (95% CI)
tXue etal. (2018)
Shenyang, China
Ozone: 2013-2015
Follow-up: 2013-2015
Case-crossover study
n = 29,112 lung cancer Average of six monitors
deaths
24-h avg
Mean: 28.5
Median: 26.9
75th: 39.1
Maximum: 90.4
Correlation (r):
PM2.5: -0.25;
NO2: -0.51;
SO2: -0.55;
PM10: -0.21;
CO: -0.25
Copollutant models:
NR
Lung cancer mortality
(lag 0-1): 0.98 (0.81,
1.21)
Lung cancer mortality
(lag 0-2): 0.98 (0.83,
1.19)
tEckel etal. (2016)
California, U.S.
Ozone: 1988-2011
Follow-up: 1988-2011
Cohort study
n = 352,053
California residents
with new diagnosis of
cancer
Monthly averages calculated
from IDW from up to four
monitors within 50 km of
residence
8-h max
Mean: 40.2
Correlation (r):
PM2.5: -0.02;
NO2: -0.01;
PM10: 0.36
Copollutant models:
NR
Lung cancer mortality:
1.03 (1.02, 1.03)
tRecent studies evaluated since the 2013 Ozone ISA.
ACS = American Cancer Society; avg = average; CanCHEC = Canadian Census Health and Environment Cohort; CI = confidence interval; CMAQ = Community Multiscale Air
Quality; CO = carbon monoxide; h = hour; HBM = hierarchical Bayesian model; HR = hazard ratio; IDW = inverse-distance weighting; km = kilometer; n = sample size;
NAMS = National Air Monitoring Stations; N02 = nitrogen dioxide; NR = not reported; OR = odds ratio; PM25 = particulate matter with a nominal mean aerodynamic diameter less
than or equal to 2.5 |jm; PM10 = particulate matter with a nominal mean aerodynamic diameter less than or equal to 10 |jm; SLAMS = State and Local Air Monitoring Stations;
S02 = sulfur dioxide; yr = year.
7-121
-------
Table 7-31 Epidemiologic studies of long-term exposure to ozone and other cancer endpoints.
Study Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
HR (95% CI)
tBadaloni et al. (2013) n = 620 cases
Nationwide, Italy Age: children <10 yr
Ozone: NR SETIL
Follow-up: children born between
1998-2001
Case-control study
LUR—6-yr mean
Mean: 24.2
Maximum: 50.1
Correlation (r): NR
Copollutant models:
NR
Q2 vs. Q1 ozone
exposure—incident
leukemia: 0.88 (0.65, 1.19)
Q3 vs. Q1 ozone
exposure—incident
leukemia: 1.2 (0.87, 1.65)
Q4 vs. Q1 ozone
exposure—incident
leukemia: 1.1 (0.76, 1.59)
tTurneretal. (2017) n = 623.048
Nationwide, U.S. Age: 30+ yrold
Ozone: 2002-2004 ACS
Follow-up: 1982-2004
Cohort study
HBM with inputs from
NAMS/SLAMS and CMAQ
8-h max
Mean: 38.2
Median: 38.1
75th: 40.1
95th: 44.9
Maximum: 59.3
Correlation (r): NO2:
-0.09
Copollutant models:
NR
(Selected results; highest
and lowest magnitude
results presented)
Salivary gland cancer
(n = 58): 1.70 (0.87, 3.34)
Pharynx cancer (n = 243):
1.16 (0.80, 1.68)
Eye cancer (n = 26): 0.67
(0.25, 1.85)
Connective tissue cancer
(n = 377): 0.84 (0.65, 1.12)
7-122
-------
Table 7-31 (Continued): Epidemiologic studies of long-term exposure to ozone and other cancer endpoints.
Study
Study Population Exposure Assessment Mean (ppb)
Copollutant
Examination
Effect Estimates
HR (95% CI)
tYaahivan et al. (2017)
Nationwide, U.S.
Ozone: 2001-2008
Follow-up: 2001-2009
Cohort study
n = 279,967
Age: 40+ yr old
women with no history
of breast cancer
BCSC
CMAQ-HBM
8-h max
Median: 36.1
75th: 37.9
Correlation (r): NR
Copollutant models:
NR
Q4 vs. Q1 ozone exposure;
breast tissue density: 0.8
(0.73, 0.87)
tRecent studies evaluated since the 2013 Ozone ISA.
ACS = American Cancer Society; avg = average; BCSC = Breast Cancer Surveillance Consortium; CI = confidence interval; CMAQ = Community Multiscale Air Quality; CO = carbon
monoxide; h = hour; HBM = hierarchical Bayesian model; HR = hazard ratio; IDW = inverse-distance weighting; km = kilometer; LUR = land use regression; n = sample size;
NAMS = National Air Monitoring Stations; N02 = nitrogen dioxide; NR = not reported; OR = odds ratio; PM25 = particulate matter with a nominal mean aerodynamic diameter less
than or equal to 2.5 |jm; Q = quartile; SETIL = Italian Multicentric Epidemiological Study on Risk Factors for Childhood Leukemia and Non-Hodgkin Lymphoma; SLAMS = State and
Local Air Monitoring Stations; yr = year.
7-123
-------
7.6.2
Toxicological Studies
Table 7-32 Study-specific details of ozone exposure and DNA damage.
Study
Population
Species (Strain), N, Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Holland et al. (2014)
Healthy adults
n = 11 males, 11 females
Age: 18-50 yr
0.10 and 0.20 ppm, 4 h with
intermittent exercise
Cell viability and proliferation in
culture (blood drawn at 0 and
24 h)
Frequencies of micronuclei and
nucleoplasms bridges in blood
lymphocytes (blood drawn at 0
and 24 h)
Finkenwirth et al. (2014)
Healthy adults
n = 19 in placebo group 18 in ozone group
males, 0 females
Age: 18-50 yr
0.21 ppm, 2 h
DNA damage in isolated
lymphocytes (30 min and 4.5 h
PE)
Cestonaro et al. (2017)
Rats (Wistar)
n = 12/group males, 0 females
Age: 9-10 weeks
0.05 ppm, 24 h/day for 14 or 28 days
0.05 ppm, 3 h/day for 14 and 28 days
DNA in tail/olive tail moment
(PE)
Micronuclei induction
(post-exposure)
Zhanq et al. (2017)
Rats (S-D)
n = 4 males/group for each time point,
0 females
Age: NR
2 ppm, 30 min and room air for
12 days. Groups also treated with
l-arginine and L-NAME.
Production of 8-oxoG/OGG1
during lung injury (baseline 4, 8,
and 12 days PE)
8-oxoG = 8-oxoguanine; h = hour; L-NAME = N(w)-nitro-L-arginine methyl ester; min = minute; n = sample size; NR = not reported; OGG1 = 8-oxoguanine DNA glycosylase-1;
PE = post-exposure; ppm = parts per million; S-D = Sprague-Dawley; yr = year.
7-124
-------
Annex for Appendix 7: Evaluation of Studies on Health Effects of
Ozone
This annex describes the approach used in the Integrated Science Assessment (ISA) for Ozone
and Related Photochemical Oxidants to evaluate study quality in the available health effects literature. As
described in the Preamble to the ISA (U.S. EPA. 2015). causality determinations were informed by the
integration of evidence across scientific disciplines (e.g., exposure, animal toxicology, epidemiology) and
related outcomes and by judgments of the strength of inference in individual studies. Table Annex 6-1
describes aspects considered in evaluating study quality of controlled human exposure, animal
toxicological, and epidemiologic studies. The aspects found in Table Annex 6-1 are consistent with
current best practices for reporting or evaluating health science data.1 Additionally, the aspects are
compatible with published U.S. EPA guidelines related to cancer, neurotoxicity, reproductive toxicity,
and developmental toxicity (U.S. EPA. 2005. 1998. 1996. 1991).
These aspects were not used as a checklist, and judgments were made without considering the
results of a study. The presence or absence of particular features in a study did not necessarily lead to the
conclusion that a study was less informative or should be excluded from consideration in the ISA.
Further, these aspects were not used as criteria for determining causality in the five-level hierarchy. As
described in the Preamble, causality determinations were based on judgments of the overall strengths and
limitations of the collective body of available studies and the coherence of evidence across scientific
disciplines and related outcomes. Table Annex 6-1 is not intended to be a complete list of aspects that
define a study's ability to inform the relationship between ozone and health effects, but it describes the
major aspects considered in this ISA to evaluate studies. Where possible, study elements, such as
exposure assessment and confounding (i.e., bias due to a relationship with the outcome and correlation
with exposures to ozone), are considered specifically for ozone. Thus, judgments on the ability of a study
to inform the relationship between an air pollutant and health can vary depending on the specific pollutant
being assessed.
1 For example, NTP OHAT approach (Roonev et al.. 20141. IRIS Preamble (U.S. EPA. 20151. ToxRTool (Klimisch
el al.. 1997). STROBE guidelines (von Elm et al.. 20071. and ARRIVE guidelines (Kilkenny et al.. 20101.
7-125
-------
Table Annex 7-1 Scientific considerations for evaluating the strength of
inference from studies on the health effects of ozone.
Study Design
Controlled Human Exposure:
Studies should clearly describe the primary and any secondary objectives of the study, or specific hypotheses being
tested. Study subjects should be randomly exposed without knowledge of the exposure condition. Preference is given
to balanced crossover (repeated measures) or parallel design studies which include control exposures (e.g., to clean
filtered air). In crossover studies, a sufficient and specified time between exposure days should be provided to avoid
carry over effects from prior exposure days. In parallel design studies, all arms should be matched for individual
characteristics, such as age, sex, race, anthropometric properties, and health status. In studies evaluating effects of
disease, appropriately matched healthy controls are desired for interpretative purposes.
Animal Toxicology:
Studies should clearly describe the primary and any secondary objectives of the study, or specific hypotheses being
tested. Studies should include appropriately matched control exposures (e.g., to clean filtered air, time matched).
Studies should use methods to limit differences in baseline characteristics of control and exposure groups. Studies
should randomize assignment to exposure groups and where possible conceal allocation to research personnel.
Groups should be subjected to identical experimental procedures and conditions; animal care including housing,
husbandry, etc. should be identical between groups. Blinding of research personnel to study group may not be
possible due to animal welfare and experimental considerations; however, differences in the monitoring or handling of
animals in all groups by research personnel should be minimized.
Epidemiology:
Inference is stronger for studies that clearly describe the primary and any secondary aims of the study, or specific
hypotheses being tested.
For short-term exposure, time-series, case-crossover, and panel studies are emphasized over cross-sectional studies
because they examine temporal correlations and are less prone to confounding by factors that differ between
individuals (e.g., SES, age). Panel studies with scripted exposures, in particular, can contribute to inference because
they have consistent, well-defined exposure durations across subjects, measure personal ambient pollutant
exposures, and measure outcomes at consistent, well-defined lags after exposures. Studies with large sample sizes
and conducted over multiple years are considered to produce more reliable results. Additionally, multicity studies are
preferred over single-city studies because they examine associations for large diverse geographic areas using a
consistent statistical methodology, avoiding the publication bias often associated with single-city studies.3 If other
quality parameters are equal, multicity studies carry more weight than single-city studies because they tend to have
larger sample sizes and lower potential for publication bias.
For long-term exposure, inference is considered to be stronger for prospective cohort studies and case-control studies
nested within a cohort (e.g., for rare diseases) than cross-sectional, other case-control, or ecologic studies. Cohort
studies can better inform the temporality of exposure and effect. Other designs can have uncertainty related to the
appropriateness of the control group or validity of inference about individuals from group-level data. Study design
limitations can bias health effect associations in either direction.
7-126
-------
Table Annex 7-1 (Continued): Scientific considerations for evaluating the strength
of inference from studies on the health effects of
ozone.
Study Population/Test Model
Controlled Human Exposure:
In general, the subjects recruited into study groups should be similarly matched for age, sex, race, anthropometric
properties, and health status. In studies evaluating effects of specific subject characteristics (e.g., disease, genetic
polymorphism, etc.), appropriately matched healthy controls are preferred. Relevant characteristics and health status
should be reported for each experimental group. Criteria for including and excluding subjects should be clearly
indicated. For the examination of populations with an underlying health condition (e.g., asthma), independent, clinical
assessment of the health condition is ideal, but self-report of physician diagnosis generally is considered to be reliable
for respiratory and cardiovascular disease outcomes.15 The loss or withdrawal of recruited subjects during the course
of a study should be reported. Specific rationale for excluding subject(s) from any portion of a protocol should be
explained.
Animal Toxicology:
Ideally, studies should report species, strain, substrain, genetic background, age, sex, and weight. Unless data
indicate otherwise, all animal species and strains are considered appropriate for evaluating effects of ozone exposure.
It is preferred that the authors test for effects in both sexes and multiple lifestages and report the result for each group
separately. All animals used in a study should be accounted for, and rationale for exclusion of animals or data should
be specified.
Epidemiology:
There is greater confidence in results for study populations that are recruited from and representative of the target
population. Studies with high participation and low dropout over time that is not dependent on exposure or health
status are considered to have low potential for selection bias. Clearly specified criteria for including and excluding
subjects can aid assessment of selection bias. For populations with an underlying health condition, independent,
clinical assessment of the health condition is valuable, but self-report of physician diagnosis generally is considered to
be reliable for respiratory and cardiovascular diseases.15 Comparisons of groups with and without an underlying health
condition are more informative if groups are from the same source population. Selection bias can influence results in
either direction or may not affect the validity of results but rather reduce the generalizability of findings to the target
population.
Pollutant
Controlled Human Exposure:
The focus is on studies testing ozone exposure.
Animal Toxicology:
The focus is on studies testing ozone exposure.
Epidemiology:
The focus is on studies evaluating ozone exposure.
7-127
-------
Table Annex 7-1 (Continued): Scientific considerations for evaluating the strength
of inference from studies on the health effects of
ozone.
Exposure Assessment or Assignment
Controlled Human Exposure:
For this assessment, the focus is on studies that use ozone concentrations <0.4 ppm. Studies that use higher
exposure concentrations may provide information relevant to biological plausibility, dosimetry, or inter-species
variation. Studies should have well-characterized pollutant concentration, temperature, and relative humidity and/or
have measures in place to adequately control the exposure conditions. Preference is given to balanced crossover or
parallel design studies that include control exposures (e.g., to clean filtered air). Study subjects should be randomly
exposed without knowledge of the exposure condition. Method of exposure (e.g., chamber, facemask, etc.) should be
specified and activity level of subjects during exposures should be well characterized.
Animal Toxicology:
For this assessment, the focus is on studies that use ozone concentrations <2 ppm. Studies that use higher exposure
concentrations may provide information relevant to biological plausibility, dosimetry, or inter-species variation. Studies
should characterize pollutant concentration, temperature, and relative humidity and/or have measures in place to
adequately control the exposure conditions. The focus is on inhalation exposure. Noninhalation exposure experiments
(i.e., intratracheal instillation [IT]) are informative for size fractions that cannot penetrate the airway of a study animal
and may provide information relevant to biological plausibility and dosimetry. In vitro studies may be included if they
provide mechanistic insight or examine similar effects as in vivo studies, but are generally not included. All studies
should include exposure control groups (e.g., clean filtered air).
Epidemiology:
Of primary relevance are relationships of health effects with the ambient component of ozone exposure. However,
information about ambient exposure rarely is available for individual subjects; most often, inference is based on
ambient concentrations. Studies that compare exposure assessment methods are considered to be particularly
informative. Inference is stronger when the duration or lag of the exposure metric corresponds with the time course for
physiological changes in the outcome (e.g., up to a few days for symptoms) or latency of disease (e.g., several years
for cancer).
Ambient ozone concentration tends to have low spatial heterogeneity at the urban scale, except near roads where
ozone concentration is lower because ozone reacts with nitric oxide emitted from vehicles. For studies involving
individuals with near-road or on-road exposures to ozone, in which ambient ozone concentrations are more spatially
heterogeneous and relationships between personal exposures and ambient concentrations are potentially more
variable, validated methods that capture the extent of variability for the epidemiologic study design (temporal vs.
spatial contrasts) and location carry greater weight.
Fixed-site measurements, whether averaged across multiple monitors or assigned from the nearest or single available
monitor, typically have smaller biases and smaller reductions in precision compared with spatially heterogeneous air
pollutants. Concentrations reported from fixed-site measurements can be informative if correlated with personal
exposures, closely located to study subjects, highly correlated across monitors within a location, or combined with
time-activity information.
Atmospheric models may be used for exposure assessment in place of or to supplement ozone measurements in
epidemiologic analyses. For example, grid-scale models (e.g., CMAQ) that represent ozone exposure over relatively
large spatial scales (e.g., typically greater than 4 * 4-km grid size) often do provide adequate spatial resolution to
capture acute ozone peaks that influence short-term health outcomes. Uncertainty in exposure predictions from these
models is largely influenced by model formulations and the quality of model input data pertaining to precursor
emissions or meteorology, which tends to vary on a study-by-study basis.
In studies of short-term exposure, temporal variability of the exposure metric is of primary interest. For long-term
exposures, models that capture within-community spatial variation in individual exposure may be given more weight
for spatially variable ambient ozone. Given the low spatial variability of ozone at the urban scale, exposure
measurement error typically causes health effect estimates to be underestimated for studies of either short-term or
long-term exposure. Biases and decreases in the precision of the association (i.e., wider 95% CIs) tend to be small.
Even when spatial variability is higher near roads, the reduction in ozone exposure would cause the exposure to be
overestimated at a monitor distant from the road or when averaged across a model grid cell, so that health effects
would likely be underestimated.
7-128
-------
Table Annex 7-1 (Continued): Scientific considerations for evaluating the strength
of inference from studies on the health effects of
ozone.
Outcome Assessment/Evaluation
Controlled Human Exposure:
Endpoints should be assessed in the same manner for control and exposure groups (e.g., time after exposure,
methods, endpoint evaluator) using valid, reliable methods. Blinding of endpoint evaluators is ideal, especially for
qualitative endpoints (e.g., histopathology). For each experiment and each experimental group, including controls,
precise details of all procedures carried out should be provided including how, when, and where. Time of the endpoint
evaluations is a key consideration that will vary depending on endpoint evaluated. Endpoints should be assessed at
time points that are appropriate for the research questions.
Animal Toxicology:
Endpoints should be assessed in the same manner for control and exposure groups (e.g., time after exposure,
methods, endpoint evaluator) using valid, reliable methods. Blinding of endpoint evaluators is ideal, especially for
qualitative endpoints (e.g., histopathology). For each experiment and each experimental group, including controls,
precise details of all procedures carried out should be provided including how, when, and where. Time of the endpoint
evaluations is a key consideration that will vary depending on endpoint evaluated. Endpoints should be assessed at
time points that are appropriate for the research questions.
Epidemiology:
Inference is stronger when outcomes are assessed or reported without knowledge of exposure status. Knowledge of
exposure status could produce artifactual associations. Confidence is greater when outcomes assessed by interview,
self-report, clinical examination, or analysis of biological indicators are defined by consistent criteria and collected by
validated, reliable methods. Independent, clinical assessment is valuable for outcomes like lung function or incidence
of disease, but report of physician diagnosis has shown good reliability.15 When examining short-term exposures,
evaluation of the evidence focuses on specific lags based on the evidence presented in individual studies. Specifically,
the following hierarchy is used in the process of selecting results from individual studies to assess in the context of
results across all studies for a specific health effect or outcome:
ix. Distributed lag models;
x. Average of multiple days (e.g., 0-2);
xi. If a priori lag days were used by the study authors these are the effect estimates presented; or
xii. If a study focuses on only a series of individual lag days, expert judgment is applied to select the appropriate
result to focus on considering the time course for physiologic changes for the health effect or outcome being
evaluated.
When health effects of long-term exposure are assessed by acute events such as symptoms or hospital admissions,
inference is strengthened when results are adjusted for short-term exposure. Validated questionnaires for subjective
outcomes such as symptoms are regarded to be reliable,0 particularly when collected frequently and not subject to
long recall. For biological samples, the stability of the compound of interest and the sensitivity and precision of the
analytical method is considered. If not based on knowledge of exposure status, errors in outcome assessment tend to
bias results toward the null.
Potential Copollutant Confounding
Controlled Human Exposure:
Exposure should be well characterized to evaluate independent effects of ozone.
Animal Toxicology:
Exposure should be well characterized to evaluate independent effects of ozone.
7-129
-------
Table Annex 7-1 (Continued): Scientific considerations for evaluating the strength
of inference from studies on the health effects of
ozone.
Epidemiology:
Not accounting for potential copollutant confounding can produce artifactual associations; thus, studies that examine
copollutant confounding carry greater weight. The predominant method is copollutant modeling (i.e., two-pollutant
models), which is especially informative when correlations are not high. However, when correlations are high (r> 0.7),
such as those often encountered for UFP and other traffic-related copollutants, copollutant modeling is less
informative. Although the use of single-pollutant models to examine the association between ozone and a health effect
or outcome are informative, ideally studies should also include copollutant analyses. Copollutant confounding is
evaluated on an individual study basis considering the extent of correlations observed between the copollutant and
ozone, and relationships observed with ozone and health effects in copollutant models.
Other Potential Confounding Factorsd
Controlled Human Exposure:
Preference is given to studies using experimental and control groups that are matched for individual level
characteristics (e.g., race/ethnicity, sex, body weight, smoking history, age) and time varying factors (e.g., seasonal
and diurnal patterns).
Animal Toxicology:
Preference is given to studies using experimental and control groups that are matched for individual level
characteristics (e.g., strain, sex, body weight, litter size, food and water consumption) and time varying factors
(e.g., seasonal and diurnal patterns).
Epidemiology:
Factors are considered to be potential confounders if demonstrated in the scientific literature to be related to health
effects and correlated with ozone. Not accounting for confounders can produce artifactual associations; thus, studies
that statistically adjust for multiple factors or control for them in the study design are emphasized. Less weight is
placed on studies that adjust for factors that mediate the relationship between ozone and health effects, which can
bias results toward the null. Confounders vary according to study design, exposure duration, and health effect and
may include, but are not limited to the following:
Short-term exposure studies: Meteorology, day of week, season, medication use, allergen exposure, and long-term
temporal trends.
Long-term exposure studies: Socioeconomic status, race, age, medication use, smoking status, stress, noise, and
occupational exposures.
Statistical Methodology
Controlled Human Exposure:
Statistical methods should be clearly described and appropriate for the study design and research question
(e.g., correction for multiple comparisons). Generally, statistical significance is used to evaluate the findings of
controlled human exposure studies. However, consistent trends are also informative. Detection of statistical
significance is influenced by a variety of factors including, but not limited to, the size of the study, exposure and
outcome measurement error, and statistical model specifications. Sample size is not a criterion for exclusion; ideally,
the sample size should provide adequate power to detect hypothesized effects (e.g., sample sizes less than three are
considered less informative). Because statistical tests have limitations, consideration is given to both trends in data
and reproducibility of results.
7-130
-------
Table Annex 7-1 (Continued): Scientific considerations for evaluating the strength
of inference from studies on the health effects of
ozone.
Animal Toxicology:
Statistical methods should be clearly described and appropriate for the study design and research question
(e.g., correction for multiple comparisons). Generally, statistical significance is used to evaluate the findings of animal
toxicology studies. However, consistent trends are also informative. Detection of statistical significance is influenced
by a variety of factors including, but not limited to, the size of the study, exposure and outcome measurement error,
and statistical model specifications. Sample size is not a criterion for exclusion; ideally, the sample size should provide
adequate power to detect hypothesized effects (e.g., sample sizes less than three are considered less informative).
Because statistical tests have limitations, consideration is given to both trends in data and reproducibility of results.
Epidemiology:
Multivariable regression models that include potential confounding factors are emphasized. However, multipollutant
models (more than two pollutants) are considered to produce too much uncertainty due to copollutant collinearity to be
informative. Models with interaction terms aid in the evaluation of potential confounding as well as effect modification.
Sensitivity analyses with alternate specifications for potential confounding inform the stability of findings and aid in
judgments of the strength of inference from results. In the case of multiple comparisons, consistency in the pattern of
association can increase confidence that associations were not found by chance alone. Statistical methods that are
appropriate for the power of the study carry greater weight. For example, categorical analyses with small sample sizes
can be prone to bias results toward or away from the null. Statistical tests such as f-tests and chi-squared tests are not
considered sensitive enough for adequate inferences regarding ozone-health effect associations. For all methods, the
effect estimate and precision of the estimate (i.e., width of 95% CI) are important considerations rather than statistical
significance.
aU.S. EPA (2008V
"Muraia etal. (2014V Weakley et al. (2013V Yang et al. (2011V Heckbert et al. (2004V Barr et al. (2002V Muhaiarine et al. (1997V
Toren et al. (1993V
cBurnev et al. (1989V
dMany factors evaluated as potential confounders can be effect measure modifiers (e.g., season, comorbid health condition) or
mediators of health effects related to ozone (comorbid health condition).
7-131
-------
7.7 References
Agarwal. A; Mulgund. A; Hamada. A; Chvatte. MR. (2015). A unique view on male infertility around
the globe. Reprod Biol Endocrinol 13: 37. http://dx.doi.Org/10.l 186/s 12958-015-0032-1
Agav-Shav. K; Friger. M; Linn. S; Peled. A; Amitai. Y; Peretz. C. (2013). Air pollution and congenital
heart defects. Environ Res 124: 28-34. http://dx.doi.Org/10.1016/i.envres.2013.03.005
Akhter. H; Ballinger. C; Liu. N; van Groen. T; Postlethwait. EM; Liu. R. (2015). Cyclic ozone
exposure induces gender-dependent neuropathology and memory decline in an animal model of
alzheimer's disease. Toxicol Sci 147: 222-234. http://dx.doi.org/10.1093/toxsci/kfV124
Arroyo. V: Diaz. J: Carmona. R; Ortiz. C; Linares. C. (2016). Impact of air pollution and temperature
on adverse birth outcomes: Madrid, 2001-2009. Environ Pollut 218: 1154-1161.
http://dx.doi.Org/10.1016/i.envpol.2016.08.069
Avdalovic. MY: Tyler. NK: Putney. L: Nishio. SJ; Ouesenberrv. S; Singh. PJ: Miller. LA: Schelegle.
ES: Plopper. CG: Vu. T; Hyde. DM. (2012). Ozone exposure during the early postnatal period
alters the timing and pattern of alveolar growth and development in nonhuman primates. Anat Rec
295: 1707-1716. http://dx.doi.org/10.10Q2/ar.22545
Badaloni. C: Ranucci. A: Cesaroni. G: Zanini. G: Vienneau. D: Al-Aidrous. F: De Hoogh. K:
Magnani. C: Forastiere. F; Group. SS. (2013). Air pollution and childhood leukaemia: a nationwide
case-control study in Italy. Occup Environ Med 70: 876-883. http://dx.doi.org/10.1136/oemed-
2013-101604
Barker. DJ; Osmond. C. (1986). Infant mortality, childhood nutrition, and ischaemic heart disease in
England and Wales. Lancet 1: 1077-1081.
Barr. RG: Herbstman. J: Speizer. FE: Camargo. CA. Jr. (2002). Validation of self-reported chronic
obstructive pulmonary disease in a cohort study of nurses. Am J Epidemiol 155: 965-971.
http://dx.doi.org/10.1093/aie/155.10.965
Becerra. TA: Wilhelm. M: Olsen. J: Cockburn. M: Ritz. B. (2013). Ambient air pollution and autism
in Los Angeles County, California. Environ Health Perspect 121: 380-386.
http://dx.doi.org/10.1289/ehp.1205827
Bhoopalan. V: Han. SG: Shah. MM: Thomas. DM: Bhalla. DK. (2013). Tobacco smoke modulates
ozone-induced toxicity in rat lungs and central nervous system. Inhal Toxicol 25: 21-28.
http://dx.doi.org/10.3109/08958378.2012.751143
Bignami. G: Musi. B; DeH'Omo. G: Laviola. G: Alleva. E. (1994). Limited effects of ozone exposure
during pregnancy on physical and neurobehavioral development of CD-I mice. Toxicol Appl
Pharmacol 129: 264-271. http://dx.doi.org/10.1006/taap.1994.1251
Brauer. M; Amann. M; Burnett. RT; Cohen. A: Dentener. F; Ezzati. M; Henderson. SB:
Krzvzanowski. M: Martin. RV: Van Dingenen. R: van Donkelaar. A: Thurston. GD. (2012).
Exposure assessment for estimation of the global burden of disease attributable to outdoor air
pollution. Environ Sci Technol 46: 652-660. http://dx.doi.org/10.102l/es2025752
Breton. CV: Wang. X: Mack. WJ: Berhane. K: Lopez. M: Islam. TS: Feng. M: Lurmann. F:
McConnell. R: Hodis. HN: Kiinzli. N: Avol. E. (2012). Childhood air pollutant exposure and
carotid artery intima-media thickness in young adults. Circulation 126: 1614-1620.
http ://dx.doi .org/10.1161/CIRCULATIONAHA. 112.096164
7-132
-------
Breton. CV; Yao. J; Millstein. J; Gao. L; Sicgmund. KD; Mack. W; Whitfield-Maxwell. L; Lurmann.
F; Hodis. H; Avol. E; Gilliland. FD. (2016). Prenatal Air Pollution Exposures, DNA Methyl
Transferase Genotypes, and Associations with Newborn LINE1 and Alu Methylation and
Childhood Blood Pressure and Carotid Intima-Media Thickness in the Children's Health Study.
Environ Health Perspect 124: 1905-1912. http://dx.doi.org/10.1289/EHP181
Brown. JM: Harris. G; Pantea. C; Hwang. SA: Talbot. TO. (2015). Linking air pollution data and
adverse birth outcomes: Environmental public health tracking in New York State. J Public Health
Manag Pract 21: S68-S74. http://dx.doi.org/10.1097/PHH.000000000000Q171
Burnev. PG; Laitinen. LA: Perdrizet. S; Huckauf. H: Tattersfield. AE: Chinn. S; Poisson. N; Heeren.
A: Britton. JR: Jones. T. (1989). Validity and repeatability of the IUATLD (1984) Bronchial
Symptoms Questionnaire: an international comparison. Eur Respir J 2: 940-945.
Cakmak. S: Hebbern. C: Pinault. L; Lavigne. E; Vanos. J: Crouse. PL: Tjepkema. M. (2018).
Associations between long-term PM2.5 and ozone exposure and mortality in the Canadian Census
Health and Environment Cohort (CANCHEC), by spatial synoptic classification zone. Environ Int
111: 200-211. http://dx.doi.org/10.1016/i.envint.2017.11.030
Capobussi. M: Tettamanti. R: Marcolin. L: Piovesan. L: Bronzin. S: Gattoni. ME: Polloni. I: Sabatino.
G: Tersalvi. CA: Auxilia. F: Castaldi. S. (2016). Air pollution impact on pregnancy outcomes in
Como, Italy. J Occup Environ Med 58: 47-52. http://dx.doi.org/10.1097/JOM.000000000000063Q
Carey. IM: Atkinson. RW: Kent. AJ: van Staa. T: Cook. DG: Anderson. HR. (2013). Mortality
associations with long-term exposure to outdoor air pollution in a national English cohort. Am J
Respir Crit Care Med 187: 1226-1233. http://dx.doi.Org/10.l 164/rccm.201210-1758QC
Carey. SA: Ballinger. CA: Plopper. CG: Mcdonald. RJ: Bartolucci. AA; Postlethwait. EM: Harkema.
J. R. (2011). Persistent rhinitis and epithelial remodeling induced by cyclic ozone exposure in the
nasal airways of infant monkeys. Am J Physiol Lung Cell Mol Physiol 300: L242-L254.
http://dx.doi.org/10.1152/aiplung.00177.201Q
Carmona. R: Linares. C: Recio. A: Ortiz. C: Diaz. J. (2018). Emergency multiple sclerosis hospital
admissions attributable to chemical and acoustic pollution: Madrid (Spain), 2001-2009. Sci Total
Environ 612: 111-118. http://dx.doi.Org/10.1016/i.scitotenv.2017.08.243
Carre. J: Gatimel. N: Moreau. J: Parinaud. J: Leandri. R. (2016). Influence of air quality on the results
of in vitro fertilization attempts: A retrospective study. Eur J Obstet Gynecol Reprod Biol 210:
116-122. http://dx.doi.Org/10.1016/i.eiogrb.2016.12.012
Carvalho. MA: Bernardes. LS: Hettfleisch. K; Pastro. LP: Vieira. SE; Saldiva. SR; Saldiva. PH;
Francisco. RP. (2016). Associations of maternal personal exposure to air pollution on fetal weight
and fetoplacental Doppler: A prospective cohort study. Reprod Toxicol 62: 9-17.
http://dx.doi.Org/10.1016/i.reprotox.2016.04.013
Cerza. F: Renzi. M: Agabiti. N: Marino. C: Gariazzo. C: Davoli. M: Michelozzi. P; Forastiere. F:
Cesaroni. G. (2018). Residential exposure to air pollution and incidence of Parkinsons disease in a
large metropolitan cohort. Environmental Epidemiology 2: e023.
http://dx.doi.org/10.1097/EE9.0000000000000Q23
Cestonaro. LV: Marcolan. AM: Rossato-Grando. LG: Anzolin. AP: Goethel. G: Vilani. A: Garcia. SC:
Bertol. CD. (2017). Ozone generated by air purifier in low concentrations: friend or foe? Environ
Sci Pollut Res Int 24: 22673-22678. http://dx.doi.org/10.1007/sll356-017-9887-3
Chandra. A: Copen. CE: Stephen. EH. (2013). Infertility and impaired fecundity in the United States,
1982-2010: Data from the National Survey of Family Growth. Atlanta, GA: Centers for Disease
Control, http://www.cdc.gov/nchs/data/nhsr/nhsr067.pdf
7-133
-------
Chen. C; Liu. C; Chen. R; Wang. W; Li. W; Kan. H; Fu. C. (2018). Ambient air pollution and daily
hospital admissions for mental disorders in Shanghai, China. Sci Total Environ 613-614: 324-330.
http://dx.doi.Org/10.1016/i.scitotenv.2017.09.098
Chen. CY; Hung. HJ; Chang. KH: Hsu. CY; Muo. CH; Tsai. CH; Wu. TN. (2017a). Long-term
exposure to air pollution and the incidence of Parkinson's disease: A nested case-control study.
PLoS ONE 12: e0182834. http://dx.doi.org/10.1371/iournal.pone.0182834
Chen. G; Guo. Y: Abramson. MJ: Williams. G; Li. S. (2017b). Exposure to low concentrations of air
pollutants and adverse birth outcomes in Brisbane, Australia, 2003-2013. Sci Total Environ 622-
623: 721-726. http://dx.doi.Org/10.1016/i.scitotenv.2017.12.050
Chen. H: Kwong. JC; Copes. R; Hvstad. P; van Donkelaar. A: Tu. K: Brook. JR: Goldberg. MS;
Martin. RV; Murray. BJ; Wilton. AS; Kopp. A; Burnett. RT. (2017c). Exposure to ambient air
pollution and the incidence of dementia: A population-based cohort study. Environ Int 108: 271-
277. http://dx.doi.Org/10.1016/i.envint.2017.08.020
Chen. JC; Schwartz. J. (2009). Neurobehavioral effects of ambient air pollution on cognitive
performance in US adults. Neurotoxicology 30: 231-239.
http://dx.doi.Org/10.1016/i.neuro.2008.12.011
Chiu. H; Yang. C. (2015). Air pollution and daily clinic visits for migraine in a subtropical city:
Taipei, Taiwan. J Toxicol Environ Health A 78: 549-558.
http://dx.doi.org/10.1080/15287394.2015.983218
Cho. J; Choi. YJ; Sohn. J: Suh. M; Cho. SK; Ha. KH; Kim. C; Shin. DC (2015). Ambient ozone
concentration and emergency department visits for panic attacks. J Psychiatr Res 62: 130-135.
http://dx.doi.Org/10.1016/i.ipsvchires.2015.01.010
Chou. PL; Gerriets. JE; Schelegle. ES; Hyde. DM; Miller. LA. (2011). Increased CCL24/eotaxin-2
with postnatal ozone exposure in allergen-sensitized infant monkeys is not associated with
recruitment of eosinophils to airway mucosa. Toxicol Appl Pharmacol 257: 309-318.
http://dx.doi.Org/10.1016/i.taap.2011.09.001
Chounlamountrv. K; Bover. B; Penalba. V; Francois-Bellan. AM; Bosler. O; Kessler. JP; Strube. C.
(2015). Remodeling of glial coverage of glutamatergic synapses in the rat nucleus tractus solitarii
after ozone inhalation. JNeurochem 134: 857-864. http://dx.doi.org/10. Ill l/jnc.13193
Clearv. EG; Cifuentes. M; Grinstein. G; Brugge. D; Shea. TB. (2018). Association of low-level ozone
with cognitive decline in older adults. J Alzheimers Dis 61: 67-78. http://dx.doi.org/10.3233/JAD-
170658
Conde. PG; Farhat. LC; Braga. ALF; Sallum. AEM; Farhat. SCL; Silva. CA. (2018). Are prematurity
and environmental factors determinants for developing childhood-onset systemic lupus
erythematosus? Mod Rheumatol 28: 156-160. http://dx.doi.org/10.1080/14397595.2017.13325Q8
Coneus. K; Spiess. CK. (2012). Pollution exposure and child health: evidence for infants and toddlers
in Germany. J Health Econ 31: 180-196. http://dx.doi.Org/10.1016/i.ihealeco.2011.09.006
Crouse. PL; Peters. PA; Hvstad. P; Brook. JR; van Donkelaar. A; Martin. RV; Villeneuve. PJ; Jerrett.
M; Goldberg. MS; Pope. CA; Brauer. M; Brook. RD; Robichaud. A; Menard. R; Burnett. RT.
(2015). Ambient PM 2.5, O 3, and NO 2 exposures and associations with mortality over 16 years of
follow-up in the Canadian Census Health and Environment Cohort (CanCHEC). Environ Health
Perspect 123: 1180-1186. http://dx.doi.org/10.1289/ehp.1409276
Culaui. PR; Linares. C; Ortiz. C; Carmona. R; Piaz. J. (2017). Association between environmental
factors and emergency hospital admissions due to Alzheimer's disease in Madrid. Sci Total Environ
592: 451-457. http://dx.doi.Org/10.1016/i.scitotenv.2017.03.089
7-134
-------
Dadvand. P; Rankin. J; Rushton. S; Pless-Mulloli. T. (2011). Ambient air pollution and congenital
heart disease: A register-based study. Environ Res 111: 435-441.
http://dx.doi.Org/10.1016/i.envres.2011.01.022
Dastoorpoor. M; Idani. E; Goudarzi. G; Khanjani. N. (2017). Acute effects of air pollution on
spontaneous abortion, premature delivery, and stillbirth in Ahvaz, Iran: a time-series study.
Environ Sci Pollut Res Int 25: 5447-5458. http://dx.doi.org/10.1007/sll356-017-0692-9
Diaz. J; Arroyo. V; Ortiz. C: Carmona. R: Linares. C. (2016). Effect of Environmental Factors on Low
Weight in Non-Premature Births: A Time Series Analysis. PLoS ONE 11: eO 164741.
http://dx.doi.org/10.1371/iournal.pone.0164741
Dye. JA; Gibbs-Flournov. EA; Richards. JH; Norwood. J: Kraft. K; Hatch. GE. (2017). Neonatal rat
age, sex and strain modify acute antioxidant response to ozone. Inhal Toxicol 29: 291-303.
http://dx.doi.org/10.1080/08958378.2017.13696Q2
Ebisu. K; Bell. ML. (2012). Airborne PM2.5 chemical components and low birth weight in the
Northeastern and Mid-Atlantic regions of the United States. Environ Health Perspect 120: 1746-
1752. http://dx.doi.org/10.1289/ehp. 1104763
Eckel. SP: Cockburn. M; Shu. YH; Deng. H; Lurmann. FW: Liu. L; Gilliland. FD. (2016). Air
pollution affects lung cancer survival. Thorax 71: 891-898. http://dx.doi.org/10.1136/thoraxinl-
2015-207927
Farhat. J: Farhat. SC: Braga. AL; Cocuzza. M; Borba. EF; Bonfa. E; Silva. CA. (2016). Ozone
decreases sperm quality in systemic lupus erythematosus patients. Revista Brasileira de
Reumatologia 56: 212-219. http://dx.doi.Org/10.1016/i.rbre.2015.08.005
Farhi. A; Bovko. V; Almagor. J; Benenson. I; Segre. E; Rudich. Y; Stern. E; Lerner-Geva. L. (2014).
The possible association between exposure to air pollution and the risk for congenital
malformations. Environ Res 135C: 173-180. http://dx.doi.Org/10.1016/i.envres.2014.08.024
Fernando Costa Nascimento. L; Blanco Machin. A; Antonio Almeida Dos Santos. D. (2017). Are
there differences in birth weight according to sex and associations with maternal exposure to air
pollutants? A cohort study. 135: 347-354. http://dx.doi.org/10.1590/1516-3180.2016.0262100317
Fernando Hernandez-Zimbron. L; Rivas-Arancibia. S. (2016). Syntaxin 5 overexpression and beta-
amyloid 1-42 accumulation in endoplasmic reticulum of hippocampal cells in rat brain induced by
ozone exposure. BioMed Res Int 2016: 2125643. http://dx.doi.Org/10.l 155/2016/2125643
Finkenwirth. C; Rossbach. B; Schroeder. HC; Muttrav. A. (2014). In vivo ozone exposure does not
increase DNA single-strand breaks in human peripheral lymphocytes. Hum Exp Toxicol 33: 517-
521. http://dx.doi.org/10.1177/0960327113499164
Franca. CMP; Sallum. AME; Braga. ALF; Strufaldi. FL; Silva. CAA; Farhat. SCL. (2018). Risk
factors associated with juvenile idiopathic arthritis: Exposure to cigarette smoke and air pollution
from pregnancy to disease diagnosis. J Rheumatol 45: 248-256.
http://dx.doi.org/10.3899/irheum.161500
Fuertes. E; Brauer. M; Maclntyre. E; Bauer. M; Bellander. T; von Berg. A; Berdel. D; Brunekreef. B;
Chan-Yeung. M; Gehring. U; Herbarth. O; Hoffmann. B; Kerkhof. M; Kliimper. C; Koletzko. S;
Kozvrskvi. A; Kull. I: Heinrich. I; Melen. E; Pershagen. G: Postma. D: Tiesler. CM; Carlsten. C.
(2013a). Childhood allergic rhinitis, traffic-related air pollution, and variability in the GSTP1,
TNF, TLR2, and TLR4 genes: Results from the TAG Study. J Allergy Clin Immunol 132: 342-
352.e342. http://dx.doi.Org/10.1016/i.iaci.2013.03.007
7-135
-------
Fuertes. E; Standi. M; Cyrvs. J; Berdel. D; von Berg. A; Bauer. CP; Kramer. U; Sugiri. D; Lehmann.
I; Koletzko. S; Carlsten. C; Brauer. M; Heinrich. J. (2013b). A longitudinal analysis of associations
between traffic-related air pollution with asthma, allergies and sensitization in the GINIplus and
LISAplus birth cohorts. Peer J 1: el93. http://dx.doi.org/10.7717/peeri.193
Gabehart. K; Correll. KA: Loader. JE; White. CW; Dakhama. A. (2015). The lung response to ozone
is determined by age and is partially dependent on toll-Like receptor 4. Respir Res 16: 117.
http://dx.doi.org/10.1186/sl2931-015-Q279-2
Gabehart. K; Correll. KA; Yang. J; Collins. ML; Loader. JE; Leach. S; White. CW; Dakhama. A.
(2014). Transcriptome profiling of the newborn mouse lung response to acute ozone exposure.
Toxicol Sci 138: 175-190. http://dx.doi.org/10.1093/toxsci/kft276
Gatto. NM; Henderson. VW; Hodis. HN; St John. JA; Lurmann. F; Chen. JC; Mack. WJ. (2014).
Components of air pollution and cognitive function in middle-aged and older adults in Los
Angeles. Neurotoxicology 40: 1 -7. http://dx.doi.org/10.1016/i .neuro.2013.09.004
Geer. LA; Weedon. J; Bell. ML. (2012). Ambient air pollution and term birth weight in Texas from
1998 to 2004. J Air Waste Manag Assoc 62: 1285-1295.
http://dx.doi.org/10.1080/10962247.2012.7Q7632
Gilbert. WM; Nesbitt. TS; Danielsen. B. (2003). The cost of prematurity: Quantification by gestational
age and birth weight. Obstet Gynecol 102: 488-492. http://dx.doi.org/10.1016/S0Q29-
7844(03)00617-3
Gomez-Crisostomo. NP; Rodriguez Martinez. E; Rivas-Arancibia. S. (2014). Oxidative stress
activates the transcription factors FoxO la and FoxO 3a in the hippocampus of rats exposed to low
doses of ozone. Oxid Med Cell Longev 2014: 805764. http://dx.doi.Org/10.l 155/2014/805764
Gonzalez-Guevara. E; Carlos Martinez-Lazcano. J; Custodio. V; Hernandez-Ceron. M; Rubio. C; Paz.
C (2014). Exposure to ozone induces a systemic inflammatory response: possible source of the
neurological alterations induced by this gas. Inhal Toxicol 26: 485-491.
http://dx.doi.org/10.3109/08958378.2014.922648
Goodrich. AJ; Yolk. HE; Tancredi. DJ; Mcconnell. R: Lurmann. FW; Hansen. RL; Schmidt. RJ.
(2017). Joint effects of prenatal air pollutant exposure and maternal folic acid supplementation on
risk of autism spectrum disorder. Autism Res 11: 69-80. http://dx.doi.org/10.10Q2/aur.1885
Gordon. CJ; Jarema. KA; Lehmann. J. R.; Ledbetter. AD; Schladweiler. MC; Schmid. JE; Ward. WO;
Kodavanti. UP; Nvska. A; Macphail. RC. (2013). Susceptibility of adult and senescent Brown
Norway rats to repeated ozone exposure: an assessment of behavior, serum biochemistry and
cardiopulmonary function. Inhal Toxicol 25: 141-159.
http://dx.doi.org/10.3109/08958378.2Q13.764946
Gordon. CJ; Johnstone. AF; Avdin. C; Phillips. PM; Macphail. RC; Kodavanti. UP; Ledbetter. AD;
Jarema. KA. (2014). Episodic ozone exposure in adult and senescent Brown Norway rats: acute
and delayed effect on heart rate, core temperature and motor activity. Inhal Toxicol 26: 380-390.
http://dx.doi.org/10.3109/08958378.2014.9Q5659
Gordon. CJ; Phillips. PM; Johnstone. AFM; Beaslev. TE; Ledbetter. AD; Schladweiler. MC; Snow.
SJ; Kodavanti. UP. (2016). Effect of high-fructose and high-fat diets on pulmonary sensitivity,
motor activity, and body composition of brown Norway rats exposed to ozone. Inhal Toxicol 28:
203-215. http://dx.doi.org/10.3109/08958378.2015.113473Q
7-136
-------
Gordon. CJ; Phillips. PM; Johnstone. AFM; Schmid. J; Schladweiler. MC; Ledbetter. A; Snow. SJ;
Kodavanti. UP. (2017a). Effects of maternal high-fat diet and sedentary lifestyle on susceptibility
of adult offspring to ozone exposure in rats. Inhal Toxicol 29: 239-254.
http://dx.doi.org/10.1080/08958378.2Q17.1342719
Gordon. CJ; Phillips. PM; Ledbetter. A; Snow. SJ; Schladweiler. MC; Johnstone. AF; Kodavanti. UP.
(2017b). Active vs. sedentary lifestyle from weaning to adulthood and susceptibility to ozone in
rats. Am J Physiol Lung Cell Mol Physiol 312: L100-L109.
http://dx.doi.org/10.1152/aiplung.00415.2016
Gray. SC; Edwards. SE; Schultz. BP; Miranda. ML. (2014). Assessing the impact of race, social
factors and air pollution on birth outcomes: a population-based study. Environ Health 13: 4.
http://dx.doi.org/10.1186/1476-069X-13-4
Green. R; Sarovar. V; Malig. B; Basu. R. (2015). Association of stillbirth with ambient air pollution in
a California cohort study. Am J Epidemiol 181: 874-882. http://dx.doi.org/10.1093/aie/kwu460
Gunnison. AF; Weideman. PA; Sobo. M. (1992). Enhanced inflammatory response to acute ozone
exposure in rats during pregnancy and lactation. Toxicol Sci 19: 607-612.
http://dx.doi.org/10.1016/0272-0590(92)90100-V
Guo. P; Feng. W; Zheng. M; Lv. J; Wang. L; Liu. J; Zhang. Y; Luo. G; Zhang. Y; Deng. C; Shi. T;
Liu. P; Zhang. L. (2018). Short-term associations of ambient air pollution and cause-specific
emergency department visits in Guangzhou, China. Sci Total Environ 613-614: 306-313.
http ://dx.doi .org/10.1016/i. scitotenv.2017.09.102
Guo. Y; Zeng. H; Zheng. R; Li. S; Barnett. AG; Zhang. S; Zou. X; Huxley. R; Chen. W; Williams. G.
(2016). The association between lung cancer incidence and ambient air pollution in China: A
spatiotemporal analysis. Environ Res 144: 60-65. http://dx.doi.Org/10.1016/i.envres.2015.l 1.004
Ha. S; Hu. H; Roussos-Ross. D; Haidong. K; Roth. J; Xu. X. (2014). The effects of air pollution on
adverse birth outcomes. Environ Res 134C: 198-204.
http://dx.doi.Org/10.1016/i.envres.2014.08.002
Ha. S; Mannisto. T; Liu. D; Sherman. S; Ying. O; Mendola. P. (2017a). Air pollution and
cardiovascular events at labor and delivery: a case-crossover analysis. Ann Epidemiol 27: 377-383.
http://dx.doi.Org/10.1016/i.annepidem.2017.05.007
Ha. S; Sundaram. R; Buck Louis. GM; Nobles. C; Seeni. I; Sherman. S; Mendola. P. (2017b).
Ambient air pollution and the risk ofpregnancy loss: a prospective cohort study. Fertil Steril 109:
148-153. http://dx.doi.Org/10.1016/i.fertnstert.2017.09.037
Hansen. C; Luben. TJ; Sacks. JD; Olshan. A; Jeffav. S; Strader. L; Perreault. SD. (2010). The effect of
ambient air pollution on sperm quality. Environ Health Perspect 118: 203-209.
http://dx.doi.org/10.1289/ehp.0901022
Hansen. CA; Barnett. AG; Pritchard. G. (2008). The effect of ambient air pollution during early
pregnancy on fetal ultrasonic measurements during mid-pregnancy. Environ Health Perspect 116:
362-369. http://dx.doi.org/10.1289/ehp.10720
Hao. H; Chang. HH; Holmes. HA; Mulholland. JA; Klein. M; Darrow. LA; Strickland. MJ. (2016).
Air pollution and preterm birth in the U.S. state of Georgia (2002-2006): Associations with
concentrations of 11 ambient air pollutants estimated by combining community multiscale air
quality model (CMAQ) simulations with stationary monitor measurements. Environ Health
Perspect 124: 875-880. http://dx.doi.org/10.1289/ehp. 1409651
Haro. R; Paz. C. (1993). Effects of ozone exposure during pregnancy on ontogeny of sleep in rats.
Neurosci Lett 164: 67-70. http://dx.doi.org/10.1016/0304-3940(93)90859-J
7-137
-------
Heckbert. SR; Kooperberg. C; Safford. MM; Psatv. BM; Hsia. J; McTiernan. A; Gaziano. JM;
Frishman. WH: Curb. JD. (2004). Comparison of self-report, hospital discharge codes, and
adjudication of cardiovascular events in the Women's Health Initiative. Am J Epidemiol 160: 1152-
1158. http://dx.doi.org/10.1093/aie/kwh314
Heindel. JJ; Skalla. LA; Joubert. BR; Dilworth. CH; Gray. KA. (2017). Review of developmental
origins of health and disease publications in environmental epidemiology. Reprod Toxicol 68: 34-
48. http://dx.doi.Org/10.1016/i.reprotox.2016.l 1.011
Hernandez-Zimbron. LF; Rivas-Arancibia. S. (2015). Oxidative stress caused by ozone exposure
induces beta-amyloid 1-42 overproduction and mitochondrial accumulation by activating the
amyloidogenic pathway. Neuroscience 304: 340-348.
http ://dx.doi .org/10.1016/i .neuroscience .2015.07.011
Hettfleisch. K; Bernardes. LS; Carvalho. MA; Pastro. LP; Vieira. SE; Saldiva. SR; Saldiva. P;
Francisco. RP. (2016). Short-Term Exposure to Urban Air Pollution and Influences on Placental
Vascularization Indexes. Environ Health Perspect 125: 753-759. http://dx.doi.org/10.1289/EHP300
Holland. N; Dave. V; Venkat. S; Wong. H; Donde. A; Balmes. JR; Ariomandi. M. (2014). Ozone
inhalation leads to a dose-dependent increase of cytogenetic damage in human lymphocytes.
Environ Mol Mutagen 56: 378-387. http://dx.doi.org/10.1002/em.21921
Hu. H; Ha. S; Henderson. BH; Warner. TP; Roth. J; Kan. H; Xu. X. (2015). Association of
atmospheric particulate matter and ozone with gestational diabetes mellitus. Environ Health
Perspect 123: 853-859. http://dx.doi.org/10.1289/ehp.1408456
Hu. H; Ha. S; Xu. X. (2016). Ozone and hypertensive disorders of pregnancy in Florida: Identifying
critical windows of exposure. Environ Res 153: 120-125.
http://dx.doi.Org/10.1016/i.envres.2016.12.002
Huang. CC; Wen. HJ; Chen. PC; Chiang. TL; Lin. SJ; Guo. YL. (2015). Prenatal air pollutant
exposure and occurrence of atopic dermatitis. Br J Dermatol 173: 981-988.
http://dx.doi.org/10.1111/bid. 14039
Hunter. DP; Carrell-Jacks. LA; Batchelor. TP; Dev. RD. (2011). Role of nerve growth factor in
ozone-induced neural responses in early postnatal airway development. Am J Respir Cell Mol Biol
45: 359-365. http://dx.doi.org/10.1165/rcmb.2010-0345QC
Hwang. BF; Lee. YL; Jaakkola. JJ. (2011). Air pollution and stillbirth: A population-based case-
control study in Taiwan. Environ Health Perspect 119: 1345-1349.
http://dx.doi.org/10.1289/ehp.1003056
Hwang. BF; Lee. YL; Jaakkola. JJ. (2015). Air Pollution and the Risk of Cardiac Defects: A
Population-Based Case-Control Study. Medicine (Baltimore) 94: el883.
http://dx.doi.org/10.1097/MD.00000000000Q1883
Hvstad. P; Demers. PA: Johnson. KC; Carpiano. RM; Brauer. M. (2013). Long-term residential
exposure to air pollution and lung cancer risk. Epidemiology 24: 762-772.
http://dx.doi.org/10.1097/EDE.0b013e3182949ae7
IOM (Institute of Medicine). (2007). Preterm birth: Causes, consequences, and prevention. In Preterm
birth: Causes, consequences, and prevention. Washington, DC: The National Academies Press.
http://dx.doi.org/10.17226/11622
Jeanjean. M; Bind. MA; Roux. J; Ongagna. JC; de Seze. J; Bard. D; Lerav. E. (2018). Ozone, N02and
PMlOare associated with the occurrence of multiple sclerosis relapses. Evidence from seasonal
multi-pollutant analyses. Environ Res 163: 43-52. http://dx.doi.Org/10.1016/i.envres.2018.01.040
7-138
-------
Jedlinska-Krakowska. M; Bomba. G; Jakubowski. K; Rotkiewicz. T; Jana. B; Penkowskii. A. (2006a).
Impact of oxidative stress and supplementation with vitamins E and C on testes morphology in rats.
J Reprod Dev 52: 203-209.
Jedlinska-Krakowska. M; Gizeiewski. Z; Dietrich. GJ; Jakubowski. K; Glogowski. J; Penkowski. A.
(2006b). The effect of increased ozone concentrations in the air on selected aspects of rat
reproduction. Pol J Vet Sci 9: 11-16.
Jerrett. M; Burnett. RT; Beckerman. BS; Turner. MC; Krewski. D; Thurston. G; Martin. RV: van
Donkelaar. A; Hughes. E; Shi. Y; Gapstur. SM; Thun. MJ; Pope. CA. III. (2013). Spatial analysis
of air pollution and mortality in California. Am J Respir Crit Care Med 188: 593-599.
http://dx.doi.Org/10.l 164/rccm.201303-0609QC
Jung. CR; Lin. YT; Hwang. BF. (2013). Air pollution and newly diagnostic autism spectrum
disorders: a population-based cohort study intaiwan. PLoS ONE 8: e75510.
http://dx.doi.org/10.1371/iournal.pone.007551Q
Jung. CR; Lin. YT; Hwang. BF. (2014). Ozone, particulate matter, and newly diagnosed alzheimer's
disease: A population-based cohort study in Taiwan. J Alzheimers Dis 44: 573-584.
http://dx.doi.org/10.3233/JAD-140855
Jurewicz. J; Radwan. M; Sobala. W; Polanska. K; Radwan. P; Jakubowski. L; Ulanska. A; Hanke. W.
(2014). The relationship between exposure to air pollution and sperm disomy. Environ Mol
Mutagen 56: 50-59. http://dx.doi.org/10.1002/em.21883
Kavlock. R; Daston. G; Grabowski. CT. (1979). Studies on the developmental toxicity of ozone. I.
Prenatal effects. Toxicol Appl Pharmacol 48: 19-28. http://dx.doi.org/10.1016/S0Q41-
008X(79')80004-6
Kavlock. RJ: Mever. E; Grabowski. CT. (1980). Studies on the developmental toxicity of ozone:
Postnatal effects. Toxicol Lett 5: 3-9. http://dx.doi.org/10.1016/0378-4274(80)90141-1
Kerin. T; Yolk. H; Li. W; Lurmann. F; Eckel. S; Mcconnell. R; Hertz-Picciotto. I. (2017). Association
Between Air Pollution Exposure, Cognitive and Adaptive Function, and ASD Severity Among
Children with Autism Spectrum Disorder. J Autism Dev Disord. http://dx.doi.org/10.1007/slQ803-
017-3304-0
Kilkenny. C; Browne. WJ; Cuthill. IC; Emerson. M; Altman. DG. (2010). Improving bioscience
research reporting: The ARRIVE guidelines for reporting animal research [Review]. PLoS Biol 8:
el000412. http://dx.doi.org/10.1371/iournal.pbio.1000412
Kim. D; Yolk. H; Giriraian. S; Pendergrass. S; Hall. MA; Verma. SS; Schmidt. RJ; Hansen. RL:
Ghosh. D; Ludena-Rodriguez. Y; Kim. K; Ritchie. MP; Hertz-Picciotto. I; Selleck. SB. (2017).
The joint effect of air pollution exposure and copy number variation on risk for autism. Autism Res
10: 1470-1480. http://dx.doi.org/10.1002/aur.1799
Kioumourtzoglou. MA; Power. MC; Hart. JE; Okereke. OI; Coull. BA; Laden. F; Weisskopf. MG.
(2017). The association between air pollution and onset of depression among middle-aged and
older women. Am J Epidemiol 185: 801-809. http://dx.doi.org/10.1093/aie/kwwl63
Kirrane. EF; Bowman. C; Davis. JA; Hoppin. JA; Blair. A; Chen. H; Patel. MM; Sandler. DP; Tanner.
CM; Vinikoor-Imler. L; Ward. MH; Luben. TJ; Kamel. F. (2015). Associations of ozone and
PM2.5 concentrations with Parkinson's disease among participants in the agricultural health study.
J Occup Environ Med 57: 509-517. http://dx.doi.org/10.1097/JOM.000000000000Q451
Klimisch. HJ; Andreae. M; Tillmann. U. (1997). A systematic approach for evaluating the quality of
experimental toxicological and ecotoxicological data. Regul Toxicol Pharmacol 25: 1-5.
http://dx.doi.org/10.1006/rtph.1996.1076
7-139
-------
Kodavanti. UP. (2016). Stretching the stress boundary: Linking air pollution health effects to a
neurohormonal stress response [Review]. Biochim Biophys Acta 1860: 2880-2890.
http://dx.doi.Org/10.1016/i.bbagen.2016.05.010
Krewski. D; Jerrett. M; Burnett. RT; Ma. R; Hughes. E; Shi. Y; Turner. MC; Pope. CA. Ill; Thurston.
G; Calle. EE; Thun. MJ; Beckerman. B; Deluca. P; Finkelstein. N; Ito. K; Moore. DK; Newbold.
KB; Ramsay. T; Ross. Z; Shin. H; Tempalski. B. (2009). Extended follow-up and spatial analysis
of the American Cancer Society study linking particulate air pollution and mortality [HEI]. (HEI
Research Report 140). Boston, MA: Health Effects Institute.
https://www.healtheffects.org/svstem/files/Krewskil40Statement.pdf
Laurent. O; Hu. J; Li. L; Cockburn. M; Escobedo. L; Kleeman. MJ; Wu. J. (2014). Sources and
contents of air pollution affecting term low birth weight in Los Angeles County, California,
20012008. Environ Res 134: 488-495. http://dx.doi.Org/10.1016/i.envres.2014.05.003
Laurent. O; Hu. J; Li. L; Kleeman. MJ; Bartell. SM; Cockburn. M; Escobedo. L; Wu. J. (2016a). Low
birth weight and air pollution in California: Which sources and components drive the risk? Environ
Int 92-93: 471-477. http://dx.doi.Org/10.1016/i.envint.2016.04.034
Laurent. O; Hu. J; Li. L; Kleeman. MJ; Bartell. SM; Cockburn. M; Escobedo. L; Wu. J. (2016b). A
statewide nested case-control study of preterm birth and air pollution by source and composition:
California, 2001-2008. Environ Health Perspect 124: 1479-1486.
http ://dx.doi .org/10.1289/ehp .1510133
Laurent. O; Wu. J; Li. L; Chung. J; Bartell. S. (2013). Investigating the association between birth
weight and complementary air pollution metrics: a cohort study. Environ Health 12: 18.
http://dx.doi.org/10.1186/1476-069X-12-18
Lavigne. E; Yasseen. AS; Stieb. DM; Hvstad. P; van Donkelaar. A; Martin. RV; Brook. JR; Crouse.
PL; Burnett. RT; Chen. H; Weichenthal. S; Johnson. M; Villeneuve. PJ; Walker. M. (2016).
Ambient air pollution and adverse birth outcomes: Differences by maternal comorbidities. Environ
Res 148: 457-466. http://dx.doi.Org/10.1016/i.envres.2016.04.026
Lee. H; Mvung. W; Kim. DK; Kim. SE; Kim. CT; Kim. H. (2017). Short-term air pollution exposure
aggravates Parkinson's disease in a population-based cohort. Sci Rep 7: 44741.
http://dx.doi.org/10.1038/srep44741
Lee. PC; Liu. LL; Sun. Y; Chen. YA; Liu. CC; Li. CY; Yu. HL; Ritz. B. (2016). Traffic-related air
pollution increased the risk of Parkinson's disease in Taiwan: A nationwide study. Environ Int 96:
75-81. http://dx.doi.Org/10.1016/i.envint.2016.08.017
Lee. PC; Roberts. JM; Catov. JM; Talbott. EO; Ritz. B. (2013). First Trimester Exposure To Ambient
Air Pollution, Pregnancy Complications And Adverse Birth Outcomes In Allegheny County, PA.
Matern Child Health J 17: 545-555. http://dx.doi.org/10.1007/slQ995-012-1028-5
Lee. PC; Talbott. EO; Roberts. JM; Catov. JM; Bilonick. RA; Stone. RA; Sharma. RK; Ritz. B.
(2012). Ambient air pollution exposure and blood pressure changes during pregnancy. Environ Res
117: 46-53. http://dx.doi.Org/10.1016/i.envres.2012.05.011
Lee. PC; Talbott. EO; Roberts. JM; Catov. JM; Sharma. RK; Ritz. B. (2011). Particulate air pollution
exposure and C-reactive protein during early pregnancy. Epidemiology 22: 524-531.
http://dx.doi.org/10.1097/EDE.0b013e31821c6c58
Legro. RS; Sauer. MY; Mottla. GL; Richter. KS; Li. X; Dodson. WC; Liao. D. (2010). Effect of air
quality on assisted human reproduction. Hum Reprod 25: 1317-1324.
http://dx.doi.org/10.1093/humrep/deq021
7-140
-------
Lim. YH; Kim. H; Kim. JH; Bae. S; Park. HY: Hong. YC. (2012). Air pollution and symptoms of
depression in elderly adults. Environ Health Perspect 120: 1023-1028.
http ://dx.doi .org/10.1289/ehp. 1104100
Lin. CC; Yang. SK; Lin. KC; Ho. WC; Hsieh. WS; Shu. BC; Chen. PC. (2014a). Multilevel analysis
of air pollution and early childhood neurobehavioral development. Int J Environ Res Public Health
11: 6827-6841. http://dx.doi.org/10.3390/iierphll0706827
Lin. Y. uT: Jung. C; Lee. Y: Hwang. BF. (2015). Associations Between Ozone and Preterm Birth in
Women Who Develop Gestational Diabetes. Am J Epidemiol 181: 280-287.
http://dx.doi.org/10.1093/aie/kwu264
Lin. YT: Lee. YL: Jung. CR; Jaakkola. JJ; Hwang. BF. (2014b). Air pollution and limb defects: A
matched-pairs case-control study in Taiwan. Environ Res 132: 273-280.
http://dx.doi.Org/10.1016/i.envres.2014.04.028
Linares. C; Culaui. D; Carmona. R; Ortiz. C; Diaz. J. (2017). Short-term association between
environmental factors and hospital admissions due to dementia in Madrid. Environ Res 152: 214-
220. http://dx.doi.Org/10.1016/i.envres.2016.10.020
Liu. Y: Zhou. Y: Ma. J; Bao. W; Li. J: Zhou. T: Cui. X: Peng. Z: Zhang. H: Feng. M: Yuan. Y: Chen.
Y: Huang. X: Li. Y: Duan. Y: Shi. T: Jin. L: Wu. L. (2017). Inverse Association between Ambient
Sulfur Dioxide Exposure and Semen Quality in Wuhan, China. Environ Sci Technol 51: 12806-
12814. http://dx.doi.org/10.1021/acs.est.7b03289
MacDorman. MF: Callaghan. WM: Mathews. TJ: Hovert. PL: Kochanek. KD. (2007). Trends in
preterm-related infant mortality by race and ethnicity, United States 1999-2004. Int J Health Serv
37: 635-641. http://dx.doi.Org/10.2190/HS.37.4.c
Maclntvre. EA: Brauer. M: Melen. E: Bauer. CP; Bauer. M: Berdel. D; Bergstrom. A: Brunekreef. B:
Chan-Yeung. M: Kliimper. C; Fuertes. E: Gehring. U: Gref. A: Heinrich. J: Herbarth. O: Kerkhof.
M; Koppelman. GH; Kozvrskvi. AL; Pershagen. G; Postma. DS; Thiering. E; Tiesler. CM;
Carlsten. C. (2014). GSTP1 and TNF gene variants and associations between air pollution and
incident childhood asthma: The traffic, asthma and genetics (TAG) study. Environ Health Perspect
122: 418-424. http://dx.doi.org/10.1289/ehp. 1307459
Malmqvist. E; Elding Larsson. H; Jonsson. I; Rignell-Hvdbom. A; Ivarsson. SA; Tinnerberg. H; Stroh.
E; Rittner. R; Jakobsson. K; Swietlicki. E; Rvlander. L. (2015). Maternal exposure to air pollution
and type 1 diabetes - Accounting for genetic factors. Environ Res 140: 268-274.
http://dx.doi.Org/10.1016/i.envres.2015.03.024
Mannisto. T; Mendola. P; Grantz. KL; Leishear. K; Sundaram. R; Sherman. S; Ying. O; Liu. D.
(2015a). Acute and recent air pollution exposure and cardiovascular events at labour and delivery.
Heart 101: 1491-1498. http://dx.doi.org/10.1136/heartinl-2014-307366
Mannisto. T; Mendola. P; Liu. D; Leishear. K; Ying. O. i; Sundaram. R. (2015b). Temporal variation
in the acute effects of air pollution on blood pressure measured at admission to labor/delivery. Air
Qual Atmos Health 8: 13-28. http://dx.doi.org/10.10Q7/sl 1869-014-0268-5
Mendola. P; Ha. S; Pollack. AZ; Zhu. Y; Seeni. I; Kim. SS; Sherman. S; Liu. D. (2017). Chronic and
Acute Ozone Exposure in the Week Prior to Delivery Is Associated with the Risk of Stillbirth. Int J
Environ Res Public Health 14. http://dx.doi.org/10.3390/iierphl407Q731
Mendola. P; Wallace. M; Hwang. BS; Liu. D; Robledo. C; Mannisto. T; Sundaram. R; Sherman. S;
Ying. O; Grantz. KL. (2016a). Preterm birth and air pollution: Critical windows of exposure for
women with asthma. J Allergy Clin Immunol 138: 432-440.e435.
http://dx.doi.Org/10.1016/i.iaci.2015.12.1309
7-141
-------
Mendola. P; Wallace. M; Liu. D; Robledo. C; Mannisto. T; Grantz. KL. (2016b). Air pollution
exposure and preeclampsia among US women with and without asthma. Environ Res 148: 248-
255. http://dx.doi.Org/10.1016/i.envres.2016.04.004
Michikawa. T; Morokuma. S; Fukushima. K; Kato. K; Nitta. H; Yamazaki. S. (2017a). Maternal
exposure to air pollutants during the first trimester and foetal growth in Japanese term infants.
Environ Pollut 230: 387-393. http://dx.doi.Org/10.1016/i.envpol.2017.06.069
Michikawa. T: Morokuma. S: Fukushima. K: Ueda. K: Takeuchi. A: Kato. K: Nitta. H. (2015). A
register-based study of the association between air pollutants and hypertensive disorders in
pregnancy among the Japanese population. Environ Res 142: 644-650.
http://dx.doi.Org/10.1016/i.envres.2015.08.024
Michikawa. T; Morokuma. S; Yamazaki. S; Fukushima. K; Kato. K; Nitta. H. (2016). Exposure to air
pollutants during the early weeks of pregnancy, and placenta praevia and placenta accreta in the
western part of Japan. Environ Int 92-93: 464-470. http://dx.doi.Org/10.1016/i.envint.2016.04.037
Michikawa. T; Morokuma. S; Yamazaki. S; Fukushima. K; Kato. K; Nitta. H. (2017b). Air pollutant
exposure within a few days of delivery and placental abruption in Japan. Epidemiology 28: 190-
196. http://dx.doi.org/10.1097/EDE.0000000000000605
Miller. CN; Dve. JA: Ledbetter. AD: Schladweiler. MC; Richards. JH: Snow. SJ; Wood. CE:
Henriquez. AR; Thompson. LC: Farrai. AK; Hazari. MS: Kodavanti. UP. (2017). Uterine artery
flow and offspring growth in long-evans rats following maternal exposure to ozone during
implantation. Environ Health Perspect 125: 127005. http://dx.doi.org/10.1289/EHP2019
Miller. CN: Stewart. EJ; Snow. SJ: Williams. WC: Richards. JH: Thompson. LC: Schladweiler. MC:
Farrai. AK: Kodavanti. UP: Dve. JA. (2019). Ozone Exposure During Implantation Increases
Serum Bioactivity in HTR-8/SVneo Trophoblasts. Toxicol Sci 168: 535-550.
http://dx.doi.org/10.1093/toxsci/kfzQ03
Mobasher. Z; Salam. MT; Goodwin. TM; Lurmann. F; Ingles. SA; Wilson. ML. (2013). Associations
between ambient air pollution and hypertensive disorders of pregnancy. Environ Res 123: 9-16.
http://dx.doi.Org/10.1016/i.envres.2013.01.006
Mokoena. ML: Brink. CB; Harvey. BH; Oliver. DW. (2011). Appraisal of ozone as biologically active
molecule and experimental tool in biomedical sciences. Med Chem Res 20: 1687-1695.
http://dx.doi.org/10.1007/s00044-010-9493-Q
Mokoena. ML: Harvey. BH: Vilioen. F: Ellis. SM: Brink. CB. (2015). Ozone exposure of Flinders
Sensitive Line rats is a rodent translational model of neurobiological oxidative stress with
relevance for depression and antidepressant response. Psychopharmacology 232: 2921-2938.
http://dx.doi.org/10.1007/sQ0213-015-3928-8
Morello-Frosch. R: Jesdale. BM: Sadd. JL: Pastor. M. (2010). Ambient air pollution exposure and full-
term birth weight in California. Environ Health 9: 44. http://dx.doi.Org/10.l 186/1476-069X-9-44
Moridi. M: Ziaei. S: Kazemneiad. A. (2014). Exposure to ambient air pollutants and spontaneous
abortion. J Obstet Gynaecol Res 40: 743-748. http://dx.doi.org/10.1111/iog. 12231
Morokuma. S: Michikawa. T; Yamazaki. S: Nitta. H: Kato. K. (2017). Association between exposure
to air pollution during pregnancy and false positives in fetal heart rate monitoring. Sci Rep 7:
12421. http://dx.doi.org/10.1038/s41598-017-12663-2
Muhaiarine. N: Mustard. C: Roos. LL: Young. TK: Gelskev. DE. (1997). Comparison of survey and
physician claims data for detecting hypertension. J Clin Epidemiol 50: 711-718.
http://dx.doi.org/10.1016/S0895-4356(97')00019-X
7-142
-------
Mumaw. CL; Levesque. S; McGraw. C; Robertson. S; Lucas. S; Stafflingcr. JE; Campen. MJ; Hall. P;
Norenberg. JP; Anderson. T; Lund. AK; McDonald. JD; Ottens. AK; Block. ML. (2016).
Microglial priming through the lung-brain axis: The role of air pollution-induced circulating
factors. FASEB J 30: 1880-1891. http://dx.doi.org/10.1096/fi.201500047
Murgia. N; Brisman. J; Claesson. A; Muzi. G; Olin. AC; Toren. K. (2014). Validity of a questionnaire-
based diagnosis of chronic obstructive pulmonary disease in a general population-based study.
BMC Pulm Med 14: 49. http://dx.doi.org/10.1186/1471-2466-14-49
Murphy. SR; Schelegle. ES; Edwards. PC; Miller. LA; Hyde. DM; Van Winkle. LS. (2012). Postnatal
exposure history and airways oxidant stress responses in airway explants. Am J Respir Cell Mol
Biol 47: 815-823. http://dx.doi.org/10.1165/rcmb.2012-0110OC
Murphy. SR; Schelegle. ES; Miller. LA; Hyde. DM; Van Winkle. LS. (2013). Ozone exposure alters
serotonin and serotonin receptor expression in the developing lung. Toxicol Sci 134: 168-179.
http://dx.doi.org/10.1093/toxsci/kft090
Nahidi. F; Gholami. R; Rashidi. Y; Maid. HA. (2014). Relationship between air pollution and pre-
eclampsia in pregnant women: a case-control study. East Mediterr Health J 19: S60-S66.
Nishimura. KK; Galanter. JM; Roth. LA; Oh. SS; Thakur. N; Nguyen. EA; Thvne. S; Farber. HJ;
Serebriskv. D; Kumar. R; Brigino-Buenaventura. E; Davis. A; Lenoir. MA; Meade. K; Rodriguez-
Cintron. W; Avila. PC; Borrell. LN; Bibbins-Domingo. K; Rodriguez-Santana. JR; Sen. S;
Lurmann. F; Balmes. JR; Burchard. EG. (2013). Early-life air pollution and asthma risk in minority
children: the GALA II and SAGE II studies. Am J Respir Crit Care Med 188: 309-318.
http://dx.doi.Org/10.l 164/rccm.2013 02-0264QC
Nishimura. KK; Iwanaga. K; Oh. SS; Pino-Yanes. M; Eng. C; Keswani. A; Roth. LA; Nguyen. EA;
Thvne. SM; Farber. HJ; Serebriskv. D; Meade. K; Lenoir. MA; Rodriguez-Cintron. W; Borrell.
LN; Bibbins-Domingo. K; Lurmann. F; Sen. S; Rodriguez-Santana. J. R.; Brigino-Buenaventura.
E; Avila. PC; Balmes. J. R.; Kumar. R; Burchard. EG. (2016). Early-life ozone exposure associated
with asthma without sensitization in Latino children [Letter], J Allergy Clin Immunol 138: 1703-
1706.el701. http://dx.doi.Org/10.1016/i.iaci.2016.03.058
Nobles. CJ; Schisterman. EF; Ha. S; Buck Louis. GM; Sherman. S; Mendola. P. (2018). Time-varying
cycle average and daily variation in ambient air pollution and fecundability. Hum Reprod 33: 166-
176. http://dx.doi.org/10.1093/humrep/dex341
Olsson. D; Ekstrom. M; Forsberg. B. (2012). Temporal variation in air pollution concentrations and
preterm birth-a population based epidemiological study. Int J Environ Res Public Health 9: 272-
285. http://dx.doi.org/10.3390/iierph9010272
Olsson. D; Mogren. I; Forsberg. B. (2013). Air pollution exposure in early pregnancy and adverse
pregnancy outcomes: a register-based cohort study. BMJ Open 3.
http://dx.doi.Org/10.l 136/bmjopen-2012-001955
Orione. MA; Silva. CA; Sallum. AM; Campos. LM; Omori. CH; Braga. AL; Farhat. SC. (2014). Risk
factors for juvenile dermatomyositis: Exposure to tobacco and air pollutants during pregnancy.
Arthritis Care and Research 66: 1571-1575. http://dx.doi.org/10.10Q2/acr.22358
Oudin. A; Astrom. DO; Asplund. P; Steingrimsson. S; Szabo. Z; Carlsen. HK. (2018). The association
between daily concentrations of air pollution and visits to a psychiatric emergency unit: a case-
crossover study. Environ Health 17: 4. http://dx.doi.Org/10.l 186/sl2940-017-0348-8
7-143
-------
Padula. AM; Tager. IB; Carmichael. SL; Hammond. SK; Lurmann. F; Shaw. GM. (2013). The
association of ambient air pollution and traffic exposures with selected congenital anomalies in the
San Joaquin Valley of California. Am J Epidemiol 177: 1074-1085.
http://dx.doi.org/10.1093/aie/kws367
Peel. JL; Klein. M; Flanders. WD; Mulholland. JA; Freed. G; Tolbert. PE. (2011). Ambient air
pollution and apnea and bradycardia in high-risk infants on home monitors. Environ Health
Perspect 119: 1321-1327. http://dx.doi.org/10.1289/ehp.1002739
Pinto-Almazan. R; Rivas-Arancibia. S; Farfan-Garcia. ED; Rodriguez-Martinez. E; Guerra-Araiza. C.
(2014). Neuroprotective effects of tibolone against oxidative stress induced by ozone exposure.
Rev Neurol 58: 441-449.
Qian. Z; Liang. S; Yang. S; Trevathan. E; Huang. Z; Yang. R; Wang. J; Hu. K; Zhang. Y; Vaughn. M;
Shen. L; Liu. W; Li. P; Ward. P; Yang. L; Zhang. W; Chen. W; Dong. G; Zheng. T; Xu. S; Zhang.
B. (2015). Ambient air pollution and preterm birth: A prospective birth cohort study in Wuhan,
China. Int J Hyg Environ Health 219: 195-203. http://dx.doi.Org/10.1016/i.iiheh.2015.l 1.003
Reis. MMD; Guimaraes. MT; Braga. ALF; Martins. LC; Pereira. LAA. (2017). Air pollution and low
birth weight in an industrialized city in southeastern Brazil, 2003-2006. Brazilian Journal of
Epidemiology 20: 189-199. http://dx.doi.org/10.1590/1980-5497201700020Q01
Ritz. B; Qiu. J; Lee. PC; Lurmann. F; Penfold. B; Erin Weiss. R; Mcconnell. R; Arora. C; Hobel. C;
Wilhelm. M. (2014). Prenatal air pollution exposure and ultrasound measures of fetal growth in
Los Angeles, California. Environ Res 130: 7-13. http://dx.doi.Org/10.1016/i.envres.2014.01.006
Rivas-Arancibia. S; Hernandez Zimbron. LF; Rodriguez-Martinez. E; Maldonado. PD; Borgonio
Perez. G; Sepulveda-Parada. M. (2015). Oxidative stress-dependent changes in immune responses
and cell death in the substantia nigra after ozone exposure in rat. Front Aging Neurosci 7: 65.
http://dx.doi.org/10.3389/fhagi.7015.00Q65
Rivas-Arancibia. S; Rodriguez-Martinez. E; Badillo-Ramirez. I; Lopez-Gonzalez. U; Saniger. JM.
(2017). Structural changes of amyloid beta in hippocampus of rats exposed to ozone: A Raman
spectroscopy study. Frontiers in Molecular Neuroscience 10: 137.
http://dx.doi.org/10.3389/fiimol.2017.00137
Robledo. CA; Mendola. P; Yeung. E; Mannisto. T; Sundaram. R; Liu. D; Ying. Q; Sherman. S;
Grantz. KL. (2015). Preconception and early pregnancy air pollution exposures and risk of
gestational diabetes mellitus. Environ Res 137: 316-322.
http://dx.doi.Org/10.1016/i.envres.2014.12.020
Rodriguez-Martinez. E; Martinez. F; Espinosa-Garcia. MT; Maldonado. P; Rivas-Arancibia. S.
(2013). Mitochondrial dysfunction in the hippocampus of rats caused by chronic oxidative stress.
Neuroscience 252: 384-395. http://dx.doi.org/10.1016/i.neuroscience.2013.08.018
Rodriguez-Martinez. E; Nava-Ruiz. C; Escamilla-Chimal. E; Borgonio-Perez. G; Rivas-Arancibia. S.
(2016). The Effect of Chronic Ozone Exposure on the Activation of Endoplasmic Reticulum Stress
and Apoptosis in Rat Hippocampus. Front Aging Neurosci 8: 245.
http://dx.doi.org/10.3389/fhagi.2016.00245
Roonev. AA; Bovles. AL; Wolfe. MS; Bucher. JR; Thaver. KA. (2014). Systematic review and
evidence integration for literature-based environmental health science assessments. Environ Health
Perspect 122: 711-718. http://dx.doi.org/10.1289/ehp.1307972
Saigal. S; Dovle. LW. (2008). An overview of mortality and sequelae of preterm birth from infancy to
adulthood. Lancet 371: 261-269. http://dx.doi.org/10.1016/S0140-6736(08)60136-l
7-144
-------
Salam. MT; Millstein. J; Li. YF; Lurmann. FW; Margolis. HG; Gilliland. FD. (2005). Birth outcomes
and prenatal exposure to ozone, carbon monoxide, and particulate matter: Results from the
Children's Health Study. Environ Health Perspect 113: 1638-1644.
http://dx.doi.org/10.1289/ehp.8111
Schifano. P; Asta. F; Dadvand. P; Davoli. M; Basagana. X; Michelozzi. P. (2016). Heat and air
pollution exposure as triggers of delivery: A survival analysis of population-based pregnancy
cohorts in Rome and Barcelona. Environ Int 88: 153-159.
http ://dx.doi .org/10.1016/i .envint.2015.12.013
Schifano. P: Lallo. A: Asta. F: De Sario. M: Davoli. M: Michelozzi. P. (2013). Effect of ambient
temperature and air pollutants on the risk of preterm birth, Rome 2001-2010. Environ Int 61: 77-
87. http://dx.doi.Org/10.1016/i.envint.2013.09.005
Sharkhuu. T; Doerfler. PL; Copeland. C; Luebke. RW; Gilmour. MI. (2011). Effect of maternal
exposure to ozone on reproductive outcome and immune, inflammatory, and allergic responses in
the offspring. J Immunotoxicol 8: 183-194. http://dx.doi.org/10.3109/1547691X.2011.568978
Shin. S; Burnett. RT; Kwong. JC; Hystad. P; van Donkelaar. A; Brook. JR; Copes. R; Tu. K;
Goldberg. MS; Villeneuve. PJ; Martin. RV: Murray. BJ: Wilton. AS; Kopp. A: Chen. H. (2018).
Effects of ambient air pollution on incident Parkinson's disease in Ontario, 2001 to 2013: a
population-based cohort study. Int J Epidemiol 47: 2038-2048.
http://dx.doi.org/10.1093/iie/dvvl72
Slama. R; Bottagisi. S; Solanskv. I; Lepeule. J; Giorgis-Allemand. L; Sram. R. (2013). Short-term
impact of atmospheric pollution on fecundability. Epidemiology 24: 871-879.
http://dx.doi.org/10.1097/EDE.0b013e3182a7Q2c5
Smith. LB; Reich. BJ; Herring. AH; Langlois. PH; Fuentes. M. (2015). Multilevel quantile function
modeling with application to birth outcomes. Biometrics 71: 508-519.
http://dx.doi.Org/10.l 111/biom. 12294
Smith. RB; Fecht. D; Gulliver. J; Beevers. SD; Dainak. D; Blangiardo. M; Ghosh. RE; Hansell. AL;
Kelly. FJ; Anderson. HR; Toledano. MB. (2017). Impact of London's road traffic air and noise
pollution on birth weight: retrospective population based cohort study. BMJ 359: j5299.
http://dx.doi.org/10.1136/bmi.i5299
Snow. SJ; Henriquez. AR; Costa. PL; Kodavanti. UP. (2018). Neuroendocrine regulation of air
pollution health effects: emerging insights [Review]. Toxicol Sci 164: 9-20.
http://dx.doi.org/10.1093/toxsci/kfV129
Sokol. RZ; Kraft. P; Fowler. IM; Mamet. R; Kim. E; Berhane. KT. (2006). Exposure to environmental
ozone alters semen quality. Environ Health Perspect 114: 360-365.
http://dx.doi.org/10.1289/ehp.8232
Stingone. JA; Luben. TJ; Daniels. JL; Fuentes. M; Richardson. DB; Avlsworth. AS; Herring. AH;
Anderka. M; Botto. L; Correa. A; Gilboa. SM; Langlois. PH; Moslev. B; Shaw. GM; Siffel. C;
Olshan. AF. (2014). Maternal exposure to criteria air pollutants and congenital heart defects in
offspring: Results from the national birth defects prevention study. Environ Health Perspect 122:
863-872. http://dx.doi.org/10.1289/ehp.1307289
Svmanski. E; Mchugh. MK; Zhang. X; Craft. ES; Lai. D. (2016). Evaluating narrow windows of
maternal exposure to ozone and preterm birth in a large urban area in Southeast Texas. J Expo Sci
Environ Epidemiol 26: 167-172. http://dx.doi.org/10.1038/ies.2015.32
7-145
-------
Szvszkowicz. M; Kousha. T; Kingsbury. M; Colman. I. (2016). Air Pollution and Emergency
Department Visits for Depression: A Multicity Case-Crossover Study. Environ Health Insights 10:
155-161. http://dx.doi.org/10.4137/EHI.S40493
Tetreault. LF; Doucet. M; Gamache. P; Fournier. M; Brand. A; Kosatskv. T; Smargiassi. A. (2016).
Childhood exposure to ambient air pollutants and the onset of asthma: an administrative cohort
study in Quebec. Environ Health Perspect 124: 1276-1282. http://dx.doi.org/10.1289/ehp.1509838
Thoma. ME: Mclain. AC; Louis. JF: King. RB; Trumble. AC; Sundaram. R: Buck Louis. GM. (2013).
Prevalence of infertility in the United States as estimated by the current duration approach and a
traditional constructed approach. Fertil Steril 99: 1324-1331 .e 1321.
http://dx.doi.Org/10.1016/i.fertnstert.2012.ll.037
Thomson. EM; Vladisavlievic. D; Mohottalage. S; Kumarathasan. P; Vincent R. (2013). Mapping
acute systemic effects of inhaled particulate matter and ozone: multiorgan gene expression and
glucocorticoid activity. Toxicol Sci 135: 169-181. http://dx.doi.org/10.1093/toxsci/kftl37
Toren. K; Brisman. J; Jarvholm. B. (1993). Asthma and asthma-like symptoms in adults assessed by
questionnaires: A literature review [Review]. Chest 104: 600-608.
http://dx.doi.Org/10.1378/chest.104.2.600
Tu. J; Tu. W; Tedders. SH. (2016). Spatial variations in the associations of term birth weight with
ambient air pollution in Georgia, USA. Environ Int 92-93: 146-156.
http://dx.doi.Org/10.1016/i.envint.2016.04.005
Turner. MC; Jerrett. M; Pope. A. TIT; Krewski. D; Gapstur. SM; Diver. WR; Beckerman. BS;
Marshall. JD; Su. J; Crouse. PL; Burnett. RT. (2016). Long-term ozone exposure and mortality in a
large prospective study. Am J Respir Crit Care Med 193: 1134-1142.
http://dx.doi.Org/10.l 164/rccm.201508-1633QC
Turner. MC; Krewski. D; Diver. WR; Pope. CA; Burnett. RT; Jerrett. M; Marshall. JD; Gapstur. SM.
(2017). Ambient air pollution and cancer mortality in the Cancer Prevention Study II. Environ
Health Perspect 125: 087013. http://dx.doi.org/10.1289/EHP1249
Tyler. CR; Noor. S; Young. T; Rivero. V; Sanchez. B; Lucas. S; Caldwell. KK; Milligan. ED;
Campen. MJ. (2018). Aging Exacerbates Neuroinflammatory Outcomes Induced by Acute Ozone
Exposure. Toxicol Sci. http://dx.doi.org/10.1093/toxsci/kfV014
U.S. EPA (U.S. Environmental Protection Agency). (1991). Guidelines for developmental toxicity risk
assessment (pp. 1-71). (EPA/600/FR-91/001). Washington, DC: U.S. Environmental Protection
Agency, Risk Assessment Forum, http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=23162
U.S. EPA (U.S. Environmental Protection Agency). (1996). Guidelines for reproductive toxicity risk
assessment (pp. 1-143). (EPA/630/R-96/009). Washington, DC: U.S. Environmental Protection
Agency, Risk Assessment Forum, https://www.epa.gov/sites/production/files/2014-
11/documents/guidelines repro toxicitv.pdf
U.S. EPA (U.S. Environmental Protection Agency). (1998). Guidelines for neurotoxicity risk
assessment [EPA Report] (pp. 1-89). (EPA/630/R-95/00IF). Washington, DC: U.S. Environmental
Protection Agency, Risk Assessment Forum, http://www.epa.gov/risk/guidelines-neurotoxicitv-
risk-assessment
U.S. EPA (U.S. Environmental Protection Agency). (2005). Guidelines for carcinogen risk assessment
[EPA Report]. (EPA/630/P-03/00IB). Washington, DC: U.S. Environmental Protection Agency,
Risk Assessment Forum, https://www.epa.gov/sites/production/files/2013-
09/documents/cancer guidelines final 3-25-05.pdf
7-146
-------
U.S. EPA (U.S. Environmental Protection Agency). (2008). Integrated science assessment for sulfur
oxides: Health criteria [EPA Report]. (EPA/600/R-08/047F). Research Triangle Park, NC: U.S.
Environmental Protection Agency, Office of Research and Development, National Center for
Environmental Assessment- RTP. http://cfpub.cpa.gov/ncca/cfin/rccordisplav.cfin?dcid=1 98843
U.S. EPA (U.S. Environmental Protection Agency). (2013). Integrated science assessment for ozone
and related photochemical oxidants [EPA Report]. (EPA/600/R-10/076F). Research Triangle Park,
NC: U.S. Environmental Protection Agency, Office of Research and Development, National Center
for Environmental Assessment-RTP Division.
http://cfpub.cpa.gov/ncca/isa/rccordisplav.cfm?dcid=247492
U.S. EPA (U.S. Environmental Protection Agency). (2015). Preamble to the integrated science
assessments [EPA Report]. (EPA/600/R-15/067). Research Triangle Park, NC: U.S. Environmental
Protection Agency, Office of Research and Development, National Center for Environmental
Assessment, RTP Division. https://cfbub.epa.gov/ncea/isa/recordisplav.cfm?deid=310244
van Rossem. L; Rifas-Shiman. SL; Melly. SJ; Kloog. I; Luttmann-Gibson. H; Zanobetti. A; Coull.
BA; Schwartz. JD; Mittleman. MA; Oken. E; Gillman. MW: Koutrakis. P; Gold. DR. (2015).
Prenatal air pollution exposure and newborn blood pressure. Environ Health Perspect 123: 353-
359. http://dx.doi.org/10.1289/ehp.1307419
Verhein. KC; Salituro. FG; Ledeboer. MW; Fryer. AD; Jacobv. DB. (2013). Dual p38/JNK mitogen
activated protein kinase inhibitors prevent ozone-induced airway hyperreactivity in guinea pigs.
PLoS ONE 8: e75351. http://dx.doi.org/10.1371/iournal.pone.0075351
Vinikoor-Imler. LC; Davis. JA; Meyer. RE; Luben. TJ. (2013). Early prenatal exposure to air pollution
and its associations with birth defects in a state-wide birth cohort from North Carolina. Birth
Defects Res A Clin Mol Teratol 97: 696-701. http://dx.doi.org/10.1002/bdra.23159
Vinikoor-Imler. LC; Davis. JA; Mever. RE; Messer. LC; Luben. TJ. (2014). Associations between
prenatal exposure to air pollution, small for gestational age, and term low birthweight in a state-
wide birth cohort. Environ Res 132: 132-139. http://dx.doi.Org/10.1016/i.envres.2014.03.040
Vinikoor-Imler. LC; Stewart. TG; Luben. TJ; Davis. JA; Langlois. PH. (2015). An exploratory
analysis of the relationship between ambient ozone and particulate matter concentrations during
early pregnancy and selected birth defects in Texas. Environ Pollut 202: 1-6.
http://dx.doi.Org/10.1016/i.envpol.2015.03.001
Volk. HE; Kerin. T; Lurmann. F; Hertz-Picciotto. I; McConnell. R; Campbell. DB. (2014). Autism
spectrum disorder: Interaction of air pollution with the MET receptor tyrosine kinase gene.
Epidemiology 25: 44-47. http://dx.doi.org/10.1097/EDE.000000000000003Q
Volk. HE: Lurmann. F: Penfold. B; Hertz-Picciotto. I; McConnell. R (2013). Traffic-related air
pollution, particulate matter, and autism. Arch Gen Psychiatry 70: 71-77.
http://dx.doi.org/10.1001/iamapsvchiatrv.2013.266
von Elm. E; Altman. DG; Egger. M; Pocock. SJ; Gotzschc. PC; Vandenbroucke. JP. (2007). The
strengthening the reporting of observational studies in epidemiology (strobe) statement: guidelines
for reporting observational studies [Review]. PLoS Med 4: e296.
http://dx.doi.org/10.1371/iournal.pmed.0040296
Wallace. ME; Grantz. KL; Liu. D; Zhu. Y; Kim. SS; Mendola. P. (2016). Exposure to ambient air
pollution and premature rupture of membranes. Am J Epidemiol 183: 1114-1121.
http://dx.doi.org/10.1093/aie/kwv284
7-147
-------
Warren. J; Fuentes. M; Herring. A; Langlois. P. (2012). Spatial-temporal modeling of the association
between air pollution exposure and preterm birth: Identifying critical windows of exposure.
Biometrics 68: 1157-1167. http://dx.doi.org/10.1111/i. 1541-0420.2012.01774,x
Weakley. J; Webber. MP; Ye. F; Zeig-Owens. R; Cohen. HW: Hall. CB; Kelly. K; Prezant. DJ.
(2013). Agreement between obstructive airways disease diagnoses from self-report questionnaires
and medical records. Prev Med 57: 38-42. http://dx.doi.Org/10.1016/i.ypmed.2013.04.001
Wu. J; Wilhelm. M: Chung. J: Ritz. B. (2011). Comparing exposure assessment methods for traffic-
related air pollution in an adverse pregnancy outcome study. Environ Res 111: 685-692.
http://dx.doi.Org/10.1016/i.envres.2011.03.008
Wu. YC: Lin. YC: Yu. HL: Chen. JH: Chen. TF: Sun. Y: Wen. LL: Yip. PK: Chu. YM: Chen. YC.
(2015). Association between air pollutants and dementia risk in the elderly. 1: 220-228.
http://dx.doi.Org/10.1016/i.dadm.2014.ll.015
Xu. C: Fan. YN: Kan. HP: Chen. RJ: Liu. JH: Li. YF: Zhang. Y: Ji. AL: Cai. TJ. (2016). The Novel
Relationship between Urban Air Pollution and Epilepsy: A Time Series Study. PLoS ONE 11:
e0161992. http://dx.doi.org/10.1371 hournal.pone.0161992
Xu. X: Ha. S: Kan. H: Hu. H: Curbow. BA: Lissaker. CTK. (2013). Health effects of air pollution on
length of respiratory cancer survival. BMC Public Health 13: 800. http://dx.doi.org/10.1186/1471-
2458-13-800
Xu. X: Hu. H: Ha. S: Roth. J. (2014). Ambient air pollution and hypertensive disorder of pregnancy. J
Epidemiol Community Health 68: 13-20. http://dx.doi.org/10.1136/iech-2013-202902
Xue. X: Chen. J: Sun. B; Zhou. B; Li. X. (2018). Temporal trends in respiratory mortality and short-
term effects of air pollutants in Shenyang, China. Environ Sci Pollut Res Int 25: 11468-11479.
http://dx.doi.org/10.1007/sll356-018-127Q-5
Yaghivan. L: Arao. R: Brokamp. C: O'Meara. ES: Sprague. BL: Ghita. G: Ryan. P. (2017).
Association between air pollution and mammographic breast density in the Breast Cancer
Surveilance Consortium. Breast Cancer Research 19: 36. http://dx.doi.org/10.1186/sl3058-Q17-
0828-3
Yang. CL; To. T; Fotv. RG: Stieb. DM: Dell. SD. (2011). Verifying a questionnaire diagnosis of
asthma in children using health claims data. BMC Pulm Med 11. http://dx.doi.org/10.1186/1471-
2466-11-52
Yang. S: Tan. Y; Mei. H; Wang. F; Li. N: Zhao. J: Zhang. Y; Qian. Z; Chang. JJ; Svberg. KM: Peng.
A; Mei. H; Zhang. D; Zhang. Y; Xu. S: Li. Y; Zheng. T; Zhang. B. (2018). Ambient air pollution
the risk of stillbirth: A prospective birth cohort study in Wuhan, China. Int J Hyg Environ Health
221: 502-509. http://dx.doi.Org/10.1016/i.iiheh.2018.01.014
Yitshak-Sade. M; Novack. L; Landau. D; Kloog. I; Sarov. B; Hershkovitz. R; Karakis. I. (2016).
Relationship of ambient air pollutants and hazardous household factors with birth weight among
Bedouin-Arabs. Chemosphere 160: 314-322. http://dx.doi.Org/10.1016/i.chemosphere.2016.06.104
Zellner. LC: Brundage. KM: Hunter. DP: Dev. RD. (2011). Early postnatal ozone exposure alters rat
nodose and jugular sensory neuron development. Toxicol Environ Chem 93: 2055-2071.
http://dx.doi.org/10.1080/02772248.2011.61Q882
Zhang. B; Zhao. J: Yang. R; Qian. Z; Liang. S: Bassig. BA: Zhang. Y; Hu. K; Xu. S: Dong. G: Zheng.
T; Yang. S. (2016). Ozone and other air pollutants and the risk of congenital heart defects. Sci Rep
6: 34852. http://dx.doi.org/10.1038/srep34852
7-148
-------
Zhang. J; Merialdi. M; Piatt. LP; Kramer. MS. (2010). Defining normal and abnormal fetal growth:
Promises and challenges [Review]. Am J Obstet Gynecol 202: 522-528.
http://dx.doi.Org/10.1016/i.aiog.2009.10.889
Zhang. S; Li. J; Li. Y; Liu. Y; Guo. H; Xu. X. (2017). Nitric oxide synthase activity correlates with
OGG1 in ozone-induced lung injury animal models. Front Physiol 8: 249.
http://dx.doi.org/10.3389/fbhvs.2017.00249
Zhou. Y: Gilboa. SM: Herdt. ML: Lupo. PJ; Flanders. WD; Liu. Y: Shin. M: Canfield. MA: Kirbv.
RS. (2016). Maternal exposure to ozone and PM2.5 and the prevalence of orofacial clefts in four
U.S. states. Environ Res 153: 35-40. http://dx.doi.Org/10.1016/i.envres.2016.l 1.007
7-149
-------
APPENDIX 8 ECOLOGICAL EFFECTS
Summary of Causality Determinations for Ideological Effects
This \ppendi\ di;ii;iclen/es ilie scientific e\ idence lh;il supports c;ius;ihl>
delei'iiiiiKilioiis I'm' ozone exposure ;md ecolomc;il elTecIs More del;uls mi ilie c;ws;il
I'niiieuork used In re;ich lliese conclusions ;ne included in I lie hviimhle In I lie IS \ 11 S I P \.
211151 ( ;nis;ilil> delei'iiiiiKilimis I h;il ;nv new m' iv\ ised since I he l;isi ie\ lew ;ne indie;iled w illi
;iii ;isiei'isk.
Vegetation and Ixosyslom I ¦ fleets
Causality 1 )elerminalion
Visible foliar injury
Causal
Reduced vegetation grow 111
Causal
Reduced plant reproduction
Causal*
Increased tree mortality
I.ikely to be causal*
Reduced yield and quality of agricultural crops
Causal
Alteration of herbivore growth and reproduction
Likely to be causal*
Alteration of plant-insect signaling
Likely to be causal*
Reduced productivity in terrestrial ecosystems
Causal
Reduced carbon sequestration in terrestrial ecosystems
Likely to be causal
Alteration of below ground biogeochemical cycles
Causal
Alteration of terrestrial community composition
Causal*
Alteration of ecosystem water cycling
1 .ikely to be causal
8.1 Introduction
This Appendix evaluates the relevant scientific information on ecological effects as part of the
review of the air quality criteria for ozone and other photochemical oxidants and to help form the
scientific foundation for the review of the secondary National Ambient Air Quality Standard (NAAQS)
for ozone. It serves as a concise update to Chapter 9 of the 2013 Ozone ISA (U.S. EPA. 2013) and
Chapter 9 of the 2006 Ozone Air Quality Criteria Document [AQCD; U.S. EPA (2006)1. Numerous
studies on the effects of ozone on vegetation and ecosystems were reviewed in the 2013 Ozone ISA. The
8-1
-------
document concluded that responses to ozone exposure occur across a broad array of spatial scales and
ecological endpoints (summarized here in Figure 8-1 and Table 8-1). The majority of evidence for
ecological effects has been for vegetation. Effects at the individual plant level can result in broad
ecosystem-level changes, such as productivity, carbon storage, water cycling, nutrient cycling, and
community composition. Figure 8-1 shows ozone's major ecological effects at multiple levels of
biological organization from the biochemical and subleaf level up to its effects on ecosystem services,
which are the benefits that ecosystems provide people, either directly or indirectly (Costanza etal.. 2017).
The focus of the current ISA and literature evaluated therein are those effects observed at the individual,
organism level of biological organization and higher (e.g., population, community, ecosystem, etc.).
03 exposure
03 uptake & physiology
5|| -Antioxidant metabolism upregulated
Decreased photosynthesis
Decreased stomatal conductance
or sluggish stomatal response
Effects on leaves
•Visible foliar injury
•Altered leaf production
•Altered leaf chemical composition
Plant growth
•Decreased biomass accumulation
•Altered root growth
•Altered carbon allocation
•Altered reproduction
•Altered crop quality
Belowground processes
•Altered litter production and decomposition
•Altered soil carbon and nutrient cycling
•Altered soil fauna and microbial communities
CD
O
u>
(/>
CD
=3
(/>
<;
11
<
Affected ecosystem services
•Decreased productivity
•Decreased C sequestration
• Decreased crop yield
•Altered water cycling
•Altered community composition
•Altered pollination
•Altered forest products
Figure 8-1 Illustrative diagram of ozone effects in plants and ecosystems
adapted from the 2013 Ozone ISA.
8-2
-------
8.1.1 Scope
The causality determinations for ecological effects of ozone from the 2013 Ozone ISA
(Table 8-1) inform the scope of the current review. The causality determinations are generally organized
according to biological scale of organization ranging from the individual organism-level to
ecosystem-level processes. As described in the Preamble to the ISA (U.S. EPA. 2015). the U.S. EPA uses
a structured causality framework to provide a consistent and transparent basis for classifying the weight
of available evidence for health and welfare1 effects according to a five-level hierarchy: (1) causal
relationship; (2) likely to be a causal relationship; (3) suggestive, but not sufficient, to infer a causal
relationship; (4) inadequate to infer a causal relationship; and (5) not likely to be a causal relationship.
Table 8-1 Summary of ozone causality determinations for effects on vegetation
and ecosystems in the 2013 Ozone ISA.
Vegetation and Ecosystem Effects
Causality Determination from 2013 Ozone ISA
Visible foliar injury
Causal
Reduced vegetation growth
Causal
Reduced productivity in terrestrial ecosystems
Causal
Reduced carbon sequestration in terrestrial ecosystems
Likely to be causal
Reduced yield and quality of agricultural crops
Causal
Alteration of terrestrial ecosystem water cycling
Likely to be causal
Alteration of belowground biogeochemical cycles
Causal
Alteration of terrestrial community composition Likely to be causal
The current ISA has adopted the use of the Population, Exposure, Comparison, Outcome, and
Study Design (PECOS) tool to further define the scope of the current review by conveying the criteria for
inclusion or exclusion of studies (Table 8-2). The units of study as defined in the PECOS for ecological
effects of ozone are the individual organism, species, population (in the sense of a group of individuals of
the same species), community, or ecosystem. All studies included in the ISA were conducted at
concentrations occurring in the environment or experimental ozone concentrations within an order of
1 Under section 302(h) of the Clean Air Act (42 U.S.C. § 7602[h]) effects on welfare include, but are not limited to,
"effects on soils, water, crops, vegetation, manmade materials, animals, wildlife, weather, visibility, and climate,
damage to and deterioration of property, and hazards to transportation, as well as effects on economic values and on
personal comfort and well-being.".
8-3
-------
magnitude of recent concentrations observed in the U.S. (as described in Appendix 1). The level of
causality determination (from the five-tier framework) in the 2013 Ozone ISA informed how the PECOS
was designed to scope the review. For ecological endpoints for which the 2013 Ozone ISA concluded that
the evidence was sufficient to infer a causal relationship (i.e., foliar injury, vegetation growth, ecosystem
productivity, yield and quality of agricultural crops, belowground biogeochemical cycling), the current
review only evaluates studies conducted in North America (Table 8-2). There were no geographic
constraints for all the other endpoints evaluated (terrestrial water cycling; carbon sequestration; terrestrial
community composition; plant reproduction, phenology, and survival; insects and other wildlife; and
plant-animal signaling). In the PECOS for ecological effects, relevant study designs include laboratory,
greenhouse, field, gradient, open-top chamber (OTC), free-air carbon dioxide enrichment (FACE), and
modeling studies.
Exposure methodologies included in the PECOS and used to evaluate the ecological effects of
ozone are discussed in Section 8.1.2. This discussion is followed by a description of the biological
pathways and mechanisms by which ozone exposure may lead to effects at higher levels of biological
organization (Section 8.1.3). Effects of ozone exposure on major endpoints are discussed in separate
sections and include the following: visible foliar injury (Section 8.2). plant growth and biomass
(Section 8.3); plant reproduction, phenology, and mortality (Section 8.4); and reduced crop yield and
quality (Section 8.5V Ecological effects of ozone extend to plant-associated fauna, primarily insect
herbivores (Section 8.6). Plant-insect interactions can be altered via ozone's effect on volatile plant
signaling compounds (Section 8.7V This is followed by a discussion of changes in ecosystem structure
and function in response to ozone, including reduced primary production and carbon sequestration
(Section 8.8). altered belowground processes (Section 8.9). shifts in terrestrial community composition
(Section 8.10). and altered water cycling (Section 8.11). Modifying factors that may exacerbate or negate
the effects of ozone are reviewed in Section 8.12. Finally, indices of ozone exposure and dose modeling
are discussed in Section 8.13. For each of the endpoint categories, key findings and conclusions from the
2013 Ozone ISA are briefly summarized followed by new evidence. Important older studies may be
discussed to reinforce key concepts and conclusions.
8-4
-------
Table 8-2 Population, Exposure, Comparison, Outcome, and Study Design
(PECOS) tool for ozone effects on vegetation and ecosystems.
Population, Exposure, Comparison, Outcome, and Study Design
Ecological Endpoint (PECOS) Tool
Visible foliar injury, vegetation growth, Population: For any species, an individual, population (in the sense
yield/quality of agricultural crops, of a group of individuals of the same species), community, or
productivity, belowground biogeochemical ecosystem in North America
cyc''n9 Exposure: Concentrations occurring in the environment or
experimental ozone concentrations within an order of magnitude of
recent concentrations (as described in Appendix 1)
Comparison: Relevant control sites, treatments, or parameters
Outcome: Visible foliar injury, alteration of vegetative growth,
yield/quality of agricultural crops, productivity, belowground
biogeochemical cycles
Study Design: Laboratory, greenhouse, OTC, FACE, field, gradient,
or modeling studies
Terrestrial water cycling; carbon
sequestration; terrestrial community
composition; plant reproduction, phenology,
or mortality; insects, other wildlife,
plant-animal signaling
Comparison: Relevant control sites, treatments, or parameters
Outcome: Alteration of: terrestrial water cycling; carbon
sequestration; terrestrial community composition; plant reproduction,
phenology, mortality; growth reproduction and survival of insects and
other wildlife; plant-animal signaling
Study Design: Laboratory, greenhouse, OTC, FACE, field, gradient,
or modeling studies
Population: For any species, an individual, population (in the sense
of a group of individuals of the same species), community, or
ecosystem in any continent3
Exposure: Concentrations occurring in the environment or
experimental ozone concentrations within an order of magnitude of
recent concentrations (as described in Appendix 1)
Population = Unit of study.
Exposure = Environmental variable to which population is exposed.
Comparison = Change in endpoint observed by unit increase in concentration of ozone in the same or in a control
population.
Outcome = Measurable endpoint resulting from exposure.
Study Design = Laboratory, field, gradient, open-top chamber (OTC), free-air carbon dioxide enrichment (FACE),
greenhouse, and modeling studies.
Notes: This definition of population is for the purpose of applying PECOS to ecology. Ecological populations are defined as a
group of individuals of the same species.
aln cases where a comprehensive list of affected species was available, nonagricultural North American species were separated
out from the larger data sets and the evidence on the North American species was evaluated (e.g., foliar injury, biomass).
8.1.2 Assessing Ecological Response to Ozone
This section reviews the methodologies of experimental studies and definitions of ecologically
relevant ozone exposure metrics that are discussed in the rest of the Appendix.
8-5
-------
8.1.2.1
Experimental Exposure Methodologies
A variety of methods for studying plant response to ozone exposures have been developed over
the last several decades. Most of the current methodologies were discussed in detail in the 1996 Ozone
AQCD (U.S. EPA. 1996). 2006 Ozone AQCD (U.S. EPA. 2006). and the 2013 Ozone ISA (U.S. EPA.
2013V Exposure methodologies such as greenhouse studies, continuous stirred tank reactors (CSTRs),
OTCs, and free-air fumigation are applied for assessing ozone effects on individual plants and
ecosystems. Free-air carbon dioxide/ozone enrichment (FACE) systems are a more natural way of
estimating ozone effects on aboveground and belowground processes. Other methods include the use of
ambient ozone gradients across the landscape and of multivariate statistical methods to control for other
environmental variables across space and time.
8.1.2.1.1 Indoor, Controlled Environment, and Greenhouse Chambers
The earliest experimental investigations of the effects of ozone on plants used simple glass or
plastic-covered chambers, often located within greenhouses, into which a flow of ozone-enriched air or
oxygen could be passed to provide the exposure. The types, shapes, styles, materials of construction, and
locations of these chambers have been numerous. Hogsett et al. (1987a) summarized the construction and
performance of the more elaborate and better instrumented chambers since the 1960s, including those
installed in greenhouses (with or without some control of temperature and light intensity).
One greenhouse chamber approach that continues to yield useful information on the relationships
of ozone uptake to both physiological and growth effects employs continuous stirred tank reactors
(CSTRs) first described by Heck et al. (1978V Although originally developed to permit mass-balance
studies of ozone flux to plants, these use of these reactors has recently expanded to include short-term
physiological and growth studies of ozone x CO2 interactions (Grantz et al.. 2016; Grantz et al.. 2012;
Loats and Rebbeck. 1999; Reinert et al.. 1997; Rao et al.. 1995; Reinert and Ho. 1995; Heagle et al..
1994). and validation of visible foliar injury on a variety of plant species (Kline et al.. 2009; Orendovici et
al.. 2003). In many cases, supplementary lighting and temperature control of the surrounding structure
have been used to control or modify the environmental conditions (Heagle etal.. 1994).
Many investigations have used commercially available controlled-environment chambers and
walk-in rooms adapted to permit the introduction of a flow of ozone into the controlled air volume (also
called phytotrons). Like greenhouse chambers, these chambers have temperature and light control and can
be used to study interactions with other pollutants.
8-6
-------
8.1.2.1.2 Field Chambers
In general, field chamber studies are dominated by the use of various versions of the open-top
chamber (OTC) design, first described by Heagle et al. (1973) and Mandl et al. (1973). The OTC method
continues to be widely used in the U.S. and Europe for exposing plants to varying levels of ozone.
Chambers are generally ~3 m in diameter with 2.5-m-high walls. Hogsctt et al. (1987b) described in detail
many of the various modifications to the original OTC designs that have appeared subsequently; these
have included the use of larger chambers for exposing small trees (Kats et al.. 1985) or grapevines (Mandl
etal.. 1989) and the addition of a conical baffle at the top to improve ventilation (Kats et al.. 1976). a
frustum at the top to reduce ambient air incursions, and a plastic rain-cap to exclude precipitation
(Hogsett et al.. 1985). All versions of OTCs discharge air via ports in annular ducting or interiorly
perforated double-layered walls at the base of the chambers to provide turbulent mixing and the upward
mass flow of air.
Chambered systems, including OTCs, have several advantages. For instance, they can provide a
range of treatment levels including clean air (charcoal-filtered [CF]) control, ambient air control, and
several above ambient concentrations for ozone experiments. Depending on experimental intent, a
replicated, clean-air control treatment is an essential component in many experimental designs. The OTC
can provide a consistent, definable exposure because of the constant wind speed and delivery systems.
Statistically robust concentration-response (C-R) functions can be developed using such systems to
evaluate the implications of various alternative air quality scenarios on vegetation response. Nonetheless,
there are several characteristics of the OTC design and operation that can lead to exposures that might
differ from those experienced by plants in the field. First, the OTC plants are subjected to constant air
flow turbulence, which, by lowering the boundary layer resistance to diffusion, may result in increased
uptake. This may lead to an overestimation of effects relative to areas with less turbulence (Krupa et al..
1995; Legge et al.. 1995). Research has also found that OTCs may slightly change the vapor-pressure
deficit (VPD) in a way that may decrease the uptake of ozone into leaves (Piikki et al.. 2008). As with all
methods that expose vegetation to modified ozone concentrations in chambers, OTCs create internal
environments that differ from ambient air. This so-called "chamber effect" refers to the modification of
microclimatic variables, including reduced and uneven light intensity, uneven rainfall, constant wind
speed, reduced dew formation, and increased air temperatures (Fuhrer. 1994; Manning and Krupa. 1992).
However, in at least one case where canopy resistance was quantified in OTCs and in the field, it was
determined that gaseous pollutant exposure to crops in OTCs was similar to that which would have
occurred at the same concentration in the field (Unsworth et al.. 1984a. b). Because of the standardized
methodology and protocols used in the National Crop Loss Assessment Network (NCLAN)
(Section 8.13.2) and similar programs, the databases are generally assumed to be internally consistent.
While it is clear that OTCs can alter some aspects of the microenvironment and plant growth, it is
important to establish whether these differences affect the relative response of a plant to ozone. As noted
in the 1996 Ozone AQCD, evidence from several comparative studies of OTCs and other exposure
8-7
-------
systems suggested that responses were essentially the same regardless of the exposure system used and
that chamber effects did not significantly affect response. In studies that included exposure to ambient
concentrations of ozone in both OTCs, and open-air, chamberless control plots, responses in the OTCs
were the same as in open-air plots (see Section 9.26 of the 2013 Ozone ISA).
Other types of field chambers such as the "terracosm" (Lee et al.. 2009) and the recirculating
Outdoor Plant Environment Chambers [OPECs; Flowers et al. (2007)1 have been used less frequently in
recent years. See the 2013 Ozone ISA for more details (U.S. EPA. 2013).
8.1.2.1.3 Free-Air Carbon Dioxide/Ozone Enrichment (FACE) and Plume-Type Systems
Plume systems are chamberless exposure facilities in which the atmosphere surrounding plants in
the field is modified by the injection of pollutant gas into the air above or surrounding them. This is
typically accomplished by releasing the pollutant gas from tubing with multiple orifices spaced to permit
diffusion and turbulence, so as to establish relatively homogeneous conditions as the individual plumes
disperse and mix with the ambient air (Ormrod et al.. 1988).
The most common plume system used in the U.S. is a modification of the free-air carbon
dioxide/ozone enrichment (FACE) system (Miglietta et al.. 2001; Hendrev et al.. 1999; Hendrev and
Kimball. 1994). Although originally designed to provide chamberless field facilities for studying the
effects of CO2, FACE systems have been adapted to include the dispensing of ozone (Morgan et al.. 2004;
Karnoskv et al.. 1999). This method has been employed in Illinois (SoyFACE) to study soybeans
[Glycine max;Morgan et al. (2004); Rogers et al. (2004)1 and in Wisconsin (Aspen FACE) to study
quaking aspen (Populus tremuloides), birch (Betula papyrifera), and maple [Acer saccharum; Karnoskv
et al. (1999)1. Yolk et al. (2003) described a similar system for exposing grasslands that uses
7-m-diameter plots. Other FACE systems have been used in Finland (Saviranta et al.. 2010; Oksanen.
2003). China (Feng et al.. 2015). and Japan (Hoshika et al.. 2012b).
The Aspen FACE system in the U.S. discharges the pollutant gas (ozone and/or CO2) through
orifices spaced along an annular ring (or torus) or at different heights on a ring of vertical pipes. In
general, these systems allow for two ozone levels (local ambient and elevated). Computer-controlled
feedback from the monitoring of gas concentration regulates the feed rate of enriched air to the dispersion
pipes. Feedback of wind speed and directional information ensures that the discharges only occur upwind
of the treatment plots, and that discharge is restricted or closed down during periods of low wind speed or
calm conditions. The diameter of the arrays and their height (25-30 m) in some FACE systems require
large throughputs of enriched air per plot, particularly in forest tree systems. The cost of the throughputs
tends to limit the number of enrichment treatments, although Hendrev et al. (1999) argued that the cost on
an enriched volume basis is comparable to that of chamber systems. In a similar system, the SoyFACE
uses an octagon (21 m in diameter) of horizontal pipes that releases ozone to provide a constant elevated
ozone concentration above the concurrent local ozone concentration. Ozone release is maintained at
8-8
-------
approximately 10 cm above the top of the crop canopy throughout the growing season by raising the
horizontal pipes as the crop grows taller (Morgan et al.. 2004; Miglietta et al.. 2001). Research conducted
at the SoyFACE facility in Illinois (to study soybeans) and the Aspen FACE system in Wisconsin (to
study responses in broadleaf forest), have contributed a substantial body of evidence in characterizing
ozone effects at multiple scales. Aspen FACE (in operation from 1998 to 2011) enabled long-term
characterization of ozone effects in mixed forest communities.
A different FACE-type facility has been developed for the Kranzberg Ozone Fumigation
Experiment (KROFEX) in Germany beginning in 2000 (Nunn et al.. 2002; Werner and Fabian. 2002).
The experiment was designed to study the effects of ozone on mature stands of beech (Fagus sylvatica)
and spruce (Picea abies) trees in a system that functions independently of wind direction. The enrichment
of a large volume of the ambient air within the canopy takes place via orifices in vertical tubes suspended
from a horizontal grid supported above the canopy.
Although plume systems make virtually none of the modifications to the physical environment
that are inevitable with chambers, their successful use depends on selecting the appropriate numbers,
sizes, and orientations of the discharge orifices to avoid "hot-spots" resulting from the direct impingement
of jets of pollutant-enriched air on plant foliage (Werner and Fabian. 2002). Because mixing is unassisted
and completely dependent on wind turbulence and diffusion, local gradients are inevitable, especially in
large-scale systems. FACE systems have provisions for shutting down under low wind speed or calm
conditions and for an experimental area that is usually defined within a generous border in order to strive
for homogeneity of the exposure concentrations within the treatment area. They are also dependent on
continuous computer-controlled feedback of the ozone concentrations in the mixed treated air and of the
meteorological conditions. FACE and other plume systems are also unable to reduce ozone levels below
local ambient conditions.
8.1.2.1.4 Ambient Gradients
The occurrence of ambient ozone concentration gradients in the U.S. hold potential for the
examination of plant responses over multiple levels of exposure. However, few such gradients can be
found that meet the rigorous statistical requirements for comparable site characteristics such as soil type,
temperature, rainfall, radiation, and aspect (Manning and Krupa. 1992); although with small plants, soil
variability can be avoided by placing plants in large pots. The use of soil monoliths transported to various
locations along natural ozone gradients is another possible approach to overcome differences in soils;
however, this approach is also limited to small plants.
Studies in the 1970s used the natural gradients occurring in southern California to assess yield
losses of alfalfa and tomato (Oshimaetal.. 1977; Oshimaetal.. 1976). A transect study of the impact of
ozone on the growth of white clover and barley in the U.K. was confounded by differences in the
concurrent gradients of SO2 and NO2 pollution (Ashmore etal.. 1988). Studies of forest tree species in
8-9
-------
national parks in the eastern U.S. (Winner et al.. 1989) revealed increasing gradients of ozone and visible
foliar injury with increased elevation.
Several studies have used the San Bernardino Mountains Gradient Study in southern California to
study the effects of ozone and N deposition on forests dominated by ponderosa and Jeffrey pine (Jones
and Paine. 2006; Arbaugh et al.. 2003; Grulke. 1999; U.S. EPA. 1977). However, it is difficult to separate
the effects of N and ozone in some instances in these studies (Arbaugh et al.. 2003). An ozone gradient in
Wisconsin has been used to study foliar injury in a series of quaking aspen clones (Populus tremuloides)
differing in ozone sensitivity (Mankovska et al.. 2005; Karnoskv et al.. 1999). Also in the Midwest, an
east-west ozone gradient around southern Lake Michigan was used to look at growth and visible foliar
injury in black cherry (P. serotina) and common milkweed [Asclepias syriaca; Bennett et al. (2006)1.
Studies have been published that have used natural gradients to study a variety of endpoints and
species. For example, Gregg et al. (2003) studied Cottonwood {Populus deltoides) saplings grown in an
urban to rural gradient of ozone by using seven locations in the New York City area. The secondary
nature of the reactions of ozone formation and NOx titration reactions within the city center resulted in
significantly higher cumulative ozone exposures in more rural sites. Potential modifying factors such as
other pollutants, soil composition, moisture, or temperature were either controlled or accounted for in the
analysis.
8.1.2.2 Definitions of Exposure Metrics and Indices
Exposure indices are metrics that quantify exposure as it relates to measured plant damage
(e.g., reduced growth). The details of these metrics are discussed in Section 8.13.1. In the over 60 years of
research, many forms of exposure metrics have been used, including 7-, 12-, and 24-hour avg. The current
secondary standard form is identical to the primary form of the 4th highest 8-hour max avg over 3 years.
This metric is rarely reported in the vegetation research.
The metrics that best describe vegetation responses have been differentially weighted hourly
concentrations that are cumulative during the growth of plants. This is because ozone effects on plants are
cumulative, higher concentrations elicit more response, and these indices capture exposures over
timescales that align with key processes related to vegetation growth (see Section 8.1.3). The 2013 Ozone
ISA primarily discussed SUM06, AOTx, and W126 exposure metrics. Below are the definitions of the
three cumulative index forms:
• SUM06: Sum of all hourly ozone concentrations greater than or equal to 0.06 ppm observed
during a specified daily and seasonal time window.
• AOTx: Sum of the differences between hourly ozone concentrations greater than a specified
threshold during a specified daily and seasonal time window. For example, AOT40 is sum of the
differences between hourly concentrations above 0.04 ppm during a specified period.
8-10
-------
• W126: Sigmoidally weighted sum of all hourly ozone concentrations observed during a specified
daily and seasonal time window (Lefohn et al.. 1988; Lefohn and Runeckles. 1987). The
sigmoidal weighting of hourly ozone concentration is given in the equation below, where C is the
hourly ozone concentration in ppm:
1
Wc ~ 1 + 4403e~126C
Equation 8-1
8.1.3 Mechanisms Governing Vegetation Response to Ozone
The ecological effects of ozone are observed across multiple levels of biological organization,
starting at the subcellular and cellular level, then to individual organisms, and finally to ecosystem-level
processes. The 2013 Ozone ISA summarized in detail the mechanisms for ozone's effects at the leaf level
(Section 9.3 of 2013 Ozone ISA). Figure 8-2 summarizes current scientific understanding of effects of
ozone on plant physiology at the biochemical and leaf level. These effects lead to changes in
photosynthesis and carbon allocation to different plant carbon pools. Carbon allocation links ozone effects
at the subleaf and leaf level to changes at larger scales.
As seen in Figure 8-2. ozone ("O3"; represented in gray) enters the plant through leaf stomatal
openings ("gsto") during gas exchange, although some reproductive tissues are also directly affected by
ozone exposure (see Section 8.4). Ozone and its derivatives, referred to as reactive oxygen species
("ROS"), are phytotoxic sources of oxidative stress in plants [Section 9.3.2 in U.S. EPA (2013)1. They
may be partially detoxified by "antioxidants" [Section 9.3.4 in U.S. EPA (2013)1; however, any
remaining effective ozone flux causes damage to photosynthetic machinery and results in declines in
"gross photosynthesis." Ozone flux into the leaf may also cause "downregulation" of RuBisCO (the
enzyme responsible for carbon fixation), which also results in declines in gross photosynthesis
[Section 9.3.5 in U.S. EPA (2013)1. Ozone exposure may cause elevated "ethylene" production, a
multifunctional plant hormone [Section 9.3.3 in U.S. EPA (2013)1. This ozone-induced elevated ethylene
production can lead to a dampening of the abscisic acid ("ABA") signal responsible for stomatal closure,
resulting in damages to stomatal function. Less responsive stomata can increase water loss to the
atmosphere ("H20"), reducing the plant's water use efficiency. Additionally, ROS may trigger leaf
senescence and abscission (detachment from the plant) through oxidative stress to leaf biochemistry.
8-11
-------
Ethylene
CO
H,0
Down regulation
gsto
[OJ & ROS
° V
u c
[C02], Ci
-De
3 o
DC u
Antioxidants
>c Gross photosynthesis
Repair
mechanisms
Growth respiration
Q.
Biomass C:N
ratio
C allocation
Leaves/LAI
Reproductive organs
(flowers/seeds/tubers)
Roots
Storage Organs
Stems
Water uptake
N uptake
Soil
reactive N
Soil
water
ABA = abscisic acid; C = carbon; CH20 = carbohydrate; Ci = intracellular carbon dioxide; C02 = carbon dioxide; gsto = leaf stomatal
openings; H20 = water; LAI = leaf area index; N = nitrogen; ROS = reactive oxygen species.
Note: Flow of matter (carbon, water, ozone, and nitrogen) are indicated by solid arrows, relationships between processes are
indicated by broken lines. Ozone is in gray and water is in blue. Circles indicate compounds outside the plant tissue.
Source: Modified with permission from the publisher, Emberson et al. (20181.
Figure 8-2 Schematic representation of the cellular and metabolic effects of
ozone on vegetation.
Plant carbon ("[CH2O]") is affected by ozone in two major ways: through (1) decreases to gross
photosynthesis via the mechanisms outlined above and (2) increases to carbon demands as more
carbohydrates are used in "dark respiration" to support maintenance and repair processes and to produce
antioxidants and secondary metabolites. Ozone-mediated changes in plant carbon budgets result in less
carbon available for allocation to various pools: "reproductive organs," "leaves," "stems," "storage," and
"roots," as well as maintenance, defense, and repair mechanisms. Changes in these organs can affect their
function (e.g., diminished pollen production by flowers, diminished N uptake by roots), as well as affect
dependent consumer organisms (e.g., altered detection of plant flowers and leaves by herbivores, altered
abundance of belowground organisms). These changes can in turn alter ecosystem properties of storage
(productivity, C sequestration) and cycling (biogeochemistry). Thus, changes in allocation can scale up to
8-12
-------
the population, community and ecosystem-level effects assessed in this document, including changes in
soil biogeochemical cycling (Section 8.9). increased tree mortality (Section 8.4). shifts in community
composition (Section 8.10). changes to species interactions (Section 8.6. Section 8.7. Section 8.10).
declines in ecosystem productivity and carbon sequestration (Section 8.8). and alteration of ecosystem
water cycling (Section 8.11).
8.2 Visible Foliar Injury and Biomonitoring
In the 2013 Ozone ISA the evidence was sufficient to conclude that there is a causal relationship
between ambient ozone exposure and the occurrence of ozone-induced visible foliar injury on sensitive
plant species across the U.S. (U.S. EPA. 2013). Visible foliar injury resulting from exposure to ozone has
been well characterized and documented on many tree, shrub, herbaceous, and crop species through
research beginning in 1958 (U.S. EPA. 2013. 2006. 1996. 1986. 1978; NAPCA. 1970; Richards et al..
1958). Ozone-induced visible foliar injury on certain plant species is considered diagnostic because such
injuries have been verified experimentally in exposure-response studies (see Section 8.1.2.1) and are
considered bioindicators for ozone exposure. Typical types of visible injury to broadleaf plants include
stippling, flecking, surface bleaching, bifacial necrosis, pigmentation (e.g., bronzing), and chlorosis or
premature senescence. Typical visible injury symptoms for conifers include chlorotic banding, tip burn,
flecking, chlorotic mottling, and premature senescence of needles. At the time of the 2013 Ozone ISA, it
was well understood that, although common patterns of injury develop within a species, these foliar
lesions can vary considerably within and among taxonomic groups. A triad of conditions is necessary for
visible foliar injury to occur. These conditions include the presence of ozone pollution, genetic
susceptibility, and sufficient soil moisture to promote open stomata. In general, plants with higher
stomatal conductance, allowing more ozone into the leaf, are more susceptible to injury. A lack of soil
moisture generally decreases stomatal conductance. Other factors, such as leaf age and light level, have
also been shown to influence the amount of foliar injury.
As described in the PECOS tool (Table 8-2). the scope for new evidence reviewed in this section
limits studies to those conducted in North America at concentrations occurring in the environment or
experimental ozone concentrations within an order of magnitude of recent concentrations. Recent
experimental evidence continues to show a consistent association between visible injury and ozone
exposure (see Table 8-4). Studies reviewed in the current ISA provide further support for earlier
observations that sensitivity to ozone varies within and between species. Since the 2013 Ozone ISA,
several studies have further characterized modifying factors:
• Additional field studies have shown that dry periods in local areas tend to decrease the incidence
and severity of ozone-induced visible foliar injury (Kohut et al.. 2012; Smith. 2012).
• Data using additional species from greenhouse studies add to the evidence that sensitivity to
ozone varies by time of day. Nighttime ozone exposure (<78 ppb) did not cause foliar injury in
snap beans, (P. vulgaris), while daytime exposure (>62 ppb) resulted in injury (Llovd et al..
8-13
-------
2018). Exposing Pima cotton (Gossypium barbadense) to pulses of ozone at different times
throughout the day showed that sensitivity measured by foliar injury was lowest early in the
photoperiod and reached a maximum in midafternoon (Grantz et al.. 2013).
• Phenotypic variation in foliar sensitivity to ozone has been observed among genotypes for
soybean. A comparison between two cultivars in a greenhouse study reported a mean foliar injury
score of 16% for the ozone-tolerant Fiskeby III and a score of 81% for the ozone-sensitive
Mandarin cultivar (Burton et al.. 2016).
• In OTC exposure (mean 12-hour ozone concentration of 37 ppb for 118 days) foliar injury to
loblolly pine seedlings (Pinus taeda) was not related to seedling inoculation with root-infecting
fungi (Chicppa etal.. 2015).
The use of bio indicator species to detect phytotoxic levels of ozone is a longstanding and
effective methodology (Chappelka and Samuelson. 1998). To be considered a good bioindicator species,
plants must (1) exhibit a distinct, verified response; (2) have few or no confounding disease or pest
problems; and (3) exhibit genetic stability (U.S. EPA. 2013). Bioindicators are also currently being grown
in ozone gardens in several places, including the St Louis Science Center, St. Louis, MO and the
Appalachian Highlands Science Learning Center at Great Smoky Mountains National Park (Fishman et
al.. 2014). These gardens serve as a source of data on the effects of ambient ozone exposure on plants as
well as an important educational outreach tool. The U.S. Department of Agriculture (USDA) Forest
Service historically has assessed data on the incidence and severity of visible foliar injury on a variety of
ozone-sensitive plant species throughout the U.S. (Smith. 2012). Biological indicators are especially
useful in areas without ozone monitors; however, the approach requires expertise in recognizing signs and
symptoms uniquely attributable to ozone exposure. Since the 2013 Ozone ISA, several additional studies
have been conducted on bioindicator species:
• Cutleaf coneflower (Rudbeckia laciniata L. var. digitata) is an ozone bioindicator species native
to Great Smoky Mountain National Park. It was recently shown that variety ampla, native to
Rocky Mountain National Park, displayed similar visible injury and may also serve as a
bioindicator (Neufeld et al.. 2018).
• Tree of heaven (Ailanthus altissima), an established invasive species found widely across the
U.S., has been identified as an effective ozone bioindicator species by the National Park Service
and Forest Service (Smith et al.. 2008; Kohut. 2007). In greenhouse exposures, foliar injury
occurred at 8-hour avg ozone exposure levels of 60 to 120 ppb, with greater injury corresponding
to higher exposures (Seiler et al.. 2014). In the field, an ambient ozone 3-month, 12-hour W126
value of 11.6 ppm-hour induced foliar injury (Seiler et al.. 2014).
In addition to these studies, a recent global-scale synthesis of published ozone exposure studies
documents foliar injury from ozone exposure in the field, across gradients, or in controlled ozone
experiments in hundreds of species (Bergmann et al.. 2017). In field and gradient studies involving ozone
concentrations in ambient air, 245 plant species from 28 plant genera experienced ozone foliar injury
(Bergmann et al.. 2017). Many of the species that experience ozone foliar injury have populations native
to the U.S. (see Table 8-3).
8-14
-------
Table 8-3 Plant species that have populations in the U.S.3 (USDA. 2015) that
have been tested for ozone foliar injury as documented in the
references listed with each in Berqmann et al. (2017).
Species
Ozone Causes
Foliar Injury
References
Abies concolor
Y
Williams et al. (1977): Williams and Macareaor (1975)
Acer macrophyllum
Y
Temple et al. (2005)
Acer rubrum
Y
Davis and Skellv (1992): Simini et al. (1992): Findlev et
al. (1996)
Acer saccharum
Y
Gaucher et al. (2005): Pell et al. (1999): Rebbeck
(1996a): Laurence et al. (1996): Kress and Skellv
(1982): Noble et al. (1992): Tioelker et al. (1993)
Achillea millefolium
Y
Bender et al. (2002): Bunaeneret al. (1999a): Bunaener
et al. (1999b)
Agrostis vinealis
Y
Haves et al. (2006)
Alchemilla sp.
Y
Mannina et al. (2002)
Alnus incana
Y
Mortensen and Skre (1990): Lorenz et al. (2005):
Bussotti and Gerosa (2002): De Vries et al. (2003):
Mannina et al. (2002): Ozolincius and Serafinaviciute
(2003)
Alnus viridis or Alnus alnobetula
Y
Vanderhevden et al. (2001): Skellv et al. (1999): Lorenz
et al. (2005): De Vries et al. (2003)
Amorpha californica
Y
U.S. EPA (1980): Temple (1999)
Apocynum androsaemifolium
Y
Berqweiler and Mannina (1999): Davis (2007a): Davis
(2007b)
Apocynum cannabinum
Y
Kline et al. (2009)
Armeria maritima
N
Haves et al. (2006)
Artemisia campestris
Y
Lorenz et al. (2005): De Vries et al. (2003)
Artemisia douglasiana
Y
Temple (1999): U.S. EPA (1980)
Artemisia dracunculus
Y
Temple (1999)
Aruncus dioicus
Y
Bussotti et al. (2003a)
Asclepias californica
Y
Temple (1999)
Asclepias exaltata
Y
Chappelka et al. (2007): Souza et al. (2006)
Asclepias fascicularis
Y
Temple (1999)
8-15
-------
Table 8-3 (Continued): Plant species that have populations in the U.S.a (USDA,
2015) that have been tested for ozone foliar injury as
documented in the references listed with each in
Bergmann et al. (2017).
Species
Ozone Causes
Foliar Injury
References
Asclepias incarnata
Y
Orendovici et al. (2003)
Asclepias syriaca
Y
Kline et al. (2009): Chappelka et al. (1997): Davis and
Orendovici (2006): Davis (2007b): Davis (2011): Yuska
et al. (2003): Lorenz et al. (2005)
Big no ni a sp.
Y
Lorenz et al. (2005)
Bromus orcuttianus
Y
U.S. EPA (1980)
Calocedrus decurrens
Y
Williams et al. (1977)
Camissonia californica
Y
Thompson et al. (1984)
Camissonia claviformis
Y
Bvtnerowicz et al. (1988): Thompson et al. (1984)
Camissonia hirtella
Y
Bvtnerowicz et al. (1988)
Campanula rotundifolia
N
Ashmore et al. (1995): Haves et al. (2006): Mortensen
and Nilsen (1992): Ashmore et al. (1996)
Carex arenaria
Y
Jones et al. (2010)
Carex atrofusca
Y
Mortensen (1994b)
Carex echinata
Y
Haves et al. (2006)
Carex nigra
Y
Franzarina et al. (2000)
Centaurea spp.
Y
Bussotti et al. (2006)
Cephaianthus occidentaiis
Y
Kline et al. (2008)
Cercis canadensis
Y
Kline et al. (2008)
Chamerion angustifoiium
Y
Skellv et al. (1999): Mortensen (1993)
Chenopodium album
Y
Bender et al. (2006): Beramann et al. (1995): Beramann
et al. (1999): Pleiiel and Danielsson (1997): Reilina and
Davison (1992): Romaneckiene et al. (2008)
Circaea lutetiana
Y
Lorenz et al. (2005)
Cirsium acaule
N
Warwick and Tavlor (1995)
Clarkia rhomboidea
Y
Wahid et al. (2011)
Collomia grandiflora
Y
Temple (1999): U.S. EPA (1980)
Comarum palustre
Y
Battvetal. (2001): Mortensen (1994b)
8-16
-------
Table 8-3 (Continued): Plant species that have populations in the U.S.a (USDA,
2015) that have been tested for ozone foliar injury as
documented in the references listed with each in
Bergmann et al. (2017).
Species
Ozone Causes
Foliar Injury
References
Conocarpus erectus
Y
Ceron-Breton et al. (2009)
Conyza canadensis
Y
Grantz et al. (2008)
Cordylanthus rigidus
Y
TemDle (1999)
Cornus alba
Y
Skellv et al. (1999): Novak et al. (2003)
Cornus florida
Y
Davis (2011)
Cornus nuttallii
Y
Temple (1999)
Cornus spp.
Y
Bussotti and Gerosa (2002)
Cornus stolonifera
Y
Skellv et al. (1999)
Corylus cornuta
Y
Davis (2007a)
Crataegus spp.
Y
Bussotti and Gerosa (2002)
Cryptantha nevadensis
Y
Thompson et al. (1984)
Echinacea purpurea
Y
Szantoi et al. (2007)
Elymus glaucus
Y
U.S. EPA (1980)
Erigeron breweri
Y
U.S. EPA (1980)
Eriophorum angustifolium
Y
Haves et al. (2006): Mortensen (1994b)
Eschscholzia parishii
Y
Thompson et al. (1984)
Eupatorium perfoliatum
Y
Orendovici et al. (2003)
Eupatorium sp.
Y
Fenn et al. (2002)
Festuca ovina
Y
Ashmore et al. (1995): Haves et al. (2006): Pleiiel and
Danielsson (1997): Reilinq and Davison (1992):
Warwick and Tavlor (1995): Ashmore et al. (1996)
Festuca rubra
Y
Ashmore et al. (1995): Bunaener et al. (1999b):
Bunaener et al. (1999a): Haves et al. (2006): Mortensen
(1992): Ashmore et al. (1996):
Fraxinus americana
Y
Kress and Skellv (1982): Hildebrand et al. (1996)
Fraxinus pennsylvanica
Y
Kress and Skellv (1982): Lorenz et al. (2005)
Fraxinus spp.
Y
Davis (2007a)
8-17
-------
Table 8-3 (Continued): Plant species that have populations in the U.S.a (USDA,
2015) that have been tested for ozone foliar injury as
documented in the references listed with each in
Bergmann et al. (2017).
Species
Ozone Causes
Foliar Injury
References
Galium aparine
Y
U.S. EPA (1980)
Gayophytum diffusum
Y
Wahid et al. (2011)
Geum rivale
N
Battv et al. (2001)
Helianthus hirsutus
Y
Orendovici et al. (2003)
Humulus lupulus
Y
Mannina and Godzik (2004): Mannina et al. (2002):
Blum et al. (1997)
Ipomoea nil
Y
Wan et al. (2014)
Juncus effusus
N
Haves et al. (2006)
Lagunularia racemosa
Y
Ceron-Breton et al. (2009)
Lamium spp.
Y
Lorenz et al. (2005): Bussotti et al. (2006): De Vries et
al. (2003)
Lepidium virginicum
Y
Wahid et al. (2011)
Liquidambar styraciflua
Y
Kress and Skellv (1982): Davis (2011)
Liriodendron tulipifera
Y
Rebbeck (1996a): Rebbeck (1996b): Cannon Jr et al.
(1993): Simini et al. (1992): Kress and Skellv (1982):
Davis and Skellv (1992): Chappelka et al. (1999b):
Hildebrand et al. (1996)
Malacothrix glabrata
Y
ThomDson et al. (1984)
Melica nitens
Y
Mortensen (1994b)
Mentha sp.
Y
Orendovici et al. (2003)
Mentzelia albicaulis
Y
Thompson et al. (1984)
Morus spp.
Y
Bussotti and Gerosa (2002)
Oenothera biennis
Y
Lorenz et al. (2005): De Vries et al. (2003)
Oenothera el at a
Y
Wahid et al. (2011)
Oenothera rosea
Y
Skellv et al. (1999)
Oenothera sp.
Y
Skellv et al. (1999)
Oxalis acetosella
Y
Haves et al. (2006)
8-18
-------
Table 8-3 (Continued): Plant species that have populations in the U.S.a (USDA,
2015) that have been tested for ozone foliar injury as
documented in the references listed with each in
Bergmann et al. (2017).
Species
Ozone Causes
Foliar Injury
References
Oxyria digyna
N
Haves et al. (2006): Mortensen and Nilsen (1992):
Mortensen (1993)
Parthenocissus quinquefolia
Y
Bussotti et al. (2003a): Gerosa and Ballarin-Denti
(2003): Bussotti et al. (2003b): Davis and Orendovici
(2006): Mannina et al. (2002)
Pectocarya heterocarpa
Y
Thompson et al. (1984)
Pectocarya platycarpa
Y
Thompson et al. (1984)
Phleum alpinum/commutatum
Y
Pleiiel and Danielsson (1997): Danielsson et al. (1999):
Mortensen (1993)
Picea glauca
Y
Mortensen (1994a)
Picea sitchensis
Y
Lucas et al. (1988): Lucas et al. (1993): Mortensen
(1994a)
Pinus contorta
Y
Mortensen (1994a): Williams et al. (1977)
Pinus ellioti
Y
Evans and Fitzqerald (1993): Dean and Johnson
(1992): Bvres et al. (1992)
Pinus jeffrey
Y
Temple et al. (2005): Miller et al. (1998): Williams and
Macareaor (1975)
Pinus lambertiana
Y
Williams and Macareaor (1975): Williams et al. (1977)
Pinus leiophyila
Y
Fenn et al. (2002)
Pinus ponderosa
Y
Takemoto et al. (1997): Temple and Miller (1994):
Temple et al. (1993): Bevers et al. (1992): Temple et al.
(1992): Fenn et al. (2002): Jones and Paine (2006):
Williams and Macareaor (1975): Williams et al. (1977)
Pinus pungens
Y
Neufeld et al. (2000)
Pinus rigida
N
Kress and Skellv (1982)
Pinus taeda
Y
Kress and Skellv (1982): Edwards et al. (1992): Qiu et
al. (1992): Adams and O'Neill (1991): Adams et al.
(1990): Shafer et al. (1993): Spence et al. (1990):
Wiseloael et al. (1991): Chappelka et al. (1990)
Pinus virginiana
Y
Neufeld et al. (2000): Kress and Skellv (1982)
Piantago sp.
Y
Orendovici et al. (2003)
Piatanus occidentaiis
Y
Kress and Skellv (1982): Kline et al. (2008)
8-19
-------
Table 8-3 (Continued): Plant species that have populations in the U.S.a (USDA,
2015) that have been tested for ozone foliar injury as
documented in the references listed with each in
Bergmann et al. (2017).
Species
Ozone Causes
Foliar Injury
References
Platanus racemosa
Y
TemDle et al. (2005)
Poa pratensis
Y
Bender et al. (2002): Bender et al. (2006): Bunqener et
al. (1999b): Bunqener et al. (1999a): Mortensen (1992):
Ashmore et al. (1996)
Polygonatum sp.
Y
Bussotti et al. (2006)
Populus spp.
Y
Davis (2007a): Bussotti and Gerosa (2002): De Vries et
al. (2003)
Populus tremuloides
Y
Volin et al. (1998): Karnoskv et al. (1999): Karnoskv et
al. (1996): Coleman et al. (1996): Yun and Laurence
(1999)
Potentilla glandulosa
Y
Wahid et al. (2011): U.S. EPA (1980)
Prunus emerginata
Y
Temple (1999)
Prunus pensylvanica
Y
Davis (2007a)
Prunus serotina
Y
Skellv et al. (1999): Vanderhevden et al. (2001):
Gunthardt-Goerq et al. (1999): Pell et al. (1999):
Rebbeck (1996a): Rebbeck (1996b): Neufeld et al.
(1995): Simini et al. (1992): Davis and Skellv (1992):
Chappelka et al. (1997): Chappelka et al. (1999b):
Chappelka et al. (1999a): Davis and Orendovici (2006):
Davis (2007a): Davis (2007b): Davis (2011): Hildebrand
et al. (1996): De Bauer et al. (2000): Bussotti and
Gerosa (2002): Yuska et al. (2003)
Pseudotsuga menziesii
N
Runeckles and Wriqht (1996): Mortensen (1994a)
Quercus kelloggii
Y
Handlev and Grulke (2008): U.S. EPA (1980): Temple
et al. (2005)
Quercus phellos
N
Kress and Skellv (1982)
Quercus rubra
Y
Volin et al. (1998): Pell et al. (1999): Samuelson et al.
(1996): Keltinq et al. (1995): Edwards et al. (1994):
Simini et al. (1992): Davis and Skellv (1992)
Ranunculus acris
Y
Haves et al. (2006): Wvness et al. (2011): Mortensen
(1993)
Rhamnus spp.
Y
Bussotti and Gerosa (2002)
Rhizophora mangle
Y
Ceron-Breton et al. (2009)
Rhus aromatica
Y
Kline et al. (2008)
8-20
-------
Table 8-3 (Continued): Plant species that have populations in the U.S.a (USDA,
2015) that have been tested for ozone foliar injury as
documented in the references listed with each in
Bergmann et al. (2017).
Species
Ozone Causes
Foliar Injury
References
Rhus copallina
Y
Davis and Orendovici (2006): Davis (2009)
Rhus typhina
Y
Wan et al. (2013): Wan et al. (2014)
Ribes spp.
N
TemDle (1999)
Robinia pseudoacacia
Y
Skellv et al. (1999): Wanq et al. (1986): Lorenz et al.
(2005): Bussotti et al. (2003a): Bussotti and Gerosa
(2002): Bussotti et al. (2006): Bussotti et al. (2003b): De
Vries et al. (2003)
Rosa spp.
Y
Lorenz et al. (2005): Bussotti et al. (2003a): Gerosa and
Ballarin-Denti (2003)
Rubus idaeus
Y
Hunova et al. (2011): Lorenz et al. (2005): Bussotti et al.
(2003a): De Vries et al. (2003): Gerosa and Ballarin-
Denti (2003): Ozolincius and Serafinaviciute (2003)
Rubus parviflorus
Y
Temple (1999)
Rubus spp.
Y
Lorenz et al. (2005): Bussotti et al. (2003a): Bussotti et
al. (2006): Bussotti and Gerosa (2002): De Vries et al.
(2003): Bussotti et al. (2003b)
Rudbeckia laciniata
Y
Szantoi et al. (2009): Chappelka et al. (2003): Davison
et al. (2003)
Rumex acetosa
Y
Battv et al. (2001): Bender et al. (2002): Bender et al.
(2006): Berqmann et al. (1999): Pleiiel and Danielsson
(1997): Manninq and Godzik (2004): Reilinq and
Davison (1992): Ashmore et al. (1996): Mortensen
(1993): Haves et al. (2006)
Rumex sanguineus
Y
Bussotti et al. (2003a)
Rumex sp.
Y
Orendovici et al. (2003)
Salix amygdaloides
Y
Kline et al. (2008)
Salix eriocephala
Y
Kline et al. (2008)
Salix exigua
Y
Kline et al. (2008)
Salix lucida
Y
Kline et al. (2008)
Salix nigra
Y
Kline et al. (2008)
Salix sericea
Y
Kline et al. (2008)
8-21
-------
Table 8-3 (Continued): Plant species that have populations in the U.S.a (USDA,
2015) that have been tested for ozone foliar injury as
documented in the references listed with each in
Bergmann et al. (2017).
Species
Ozone Causes
Foliar Injury
References
Salix spp.
Y
Lorenz et al. (2005): Bussotti and Gerosa (2002): De
Vries et al. (2003)
Sambucus canadensis
Y
Kline et al. (2008): Davis (2007b)
Sambucus nigra
Y
Cano et al. (2007): Kline et al. (2008): Lorenz et al.
(2005): De Vries et al. (2003)
Sambucus racemosa
Y
Skellv et al. (1999): Cano et al. (2007): Vanderhevden
et al. (2001): Lorenz et al. (2005): Bussotti et al.
(2003a): De Vries et al. (2003): Godzik and Grodzinska
(2002): Mannina et al. (2002): Blum et al. (1997)
Sambucus spp.
Y
Bussotti and Gerosa (2002)
Sassafras albidum
Y
ChaDDelka et al. (1999b): Davis and Orendovici (2006):
Davis (2011)
Scirpus cespitosus
Y
Haves et al. (2006)
Scrophularia nodosa
Y
Bussotti et al. (2003a)
Sequoiadendron giganteum
Y
Williams et al. (1977)
Sicyos sp.
Y
Fenn et al. (2002)
Silene verecunda
Y
U.S. EPA (1980)
Silphium perfoliatum
Y
Orendovici et al. (2003)
Solanum nigrum
N
Bender et al. (2006): Beramann et al. (1995): Beramann
et al. (1999): Beramann et al. (1996a)
Solanum spp.
Y
Fenn et al. (2002)
Solidago albopilosa
N
Mavitv and Berrana (1994)
Soiidago canadensis
Y
Lorenz et al. (2005)
Solidago gigantea
Y
Lorenz et al. (2005)
Solidago spp.
Y
U.S. EPA (1980)
Solidago virgaurea
Y
Mortensen and Nilsen (1992): Mortensen (1993)
Spartina alterniflora
Y
Tavlor (2002)
Stachys palustris
N
Battv et al. (2001)
St achys spp.
Y
Bussotti et al. (2006)
8-22
-------
Table 8-3 (Continued): Plant species that have populations in the U.S.a (USDA,
2015) that have been tested for ozone foliar injury as
documented in the references listed with each in
Bergmann et al. (2017).
Species
Ozone Causes
Foliar Injury
References
Symphoricarpos albus
Y
Kline et al. (2008)
Symphoricarpos orbiculatus
Y
Kline et al. (2008)
Symphoricarpos spp.
Y
Kline et al. (2008): Lorenz et al. (2005)
Thalictrum minus
Y
Bussotti et al. (2003a): Gerosa and Ballarin-Denti
(2003)
THia spp.
Y
Bussotti and Gerosa (2002)
Urtica dioica
N
Bender et al. (2006): Berqmann et al. (1999): Reilinq
and Davison (1992): Berqmann et al. (1996a): Bussotti
et al. (2003a)
Vaccinium myrtillus
Y
Lorenz et al. (2005): Bussotti et al. (2006): De Vries et
al. (2003): Godzikand Grodzinska (2002): Manninq et
al. (2002)
Vaccinium uliginosum
Y
De Vries et al. (2003): Gerosa and Ballarin-Denti (2003)
Verbesina occidentalis
Y
Chappelka et al. (2003)
Vernonia noveboracensis
Y
Orendovici et al. (2003)
Viburnum nudum
Y
Berqmann et al. (2017): Davis (2007a): Davis (2007b)
Viburnum spp.
Y
Bussotti and Gerosa (2002): Manninq et al. (2002)
Vicia caiifornica
Y
U.S. EPA (1980)
Vincetoxicum sp.
Y
Blum et al. (1997)
Vitis spp.
Y
Davis and Orendovici (2006): Davis (2009): Davis
(2011)
In ozone-response categories, Y = ozone induces effect at tested exposures, N = ozone has no effect at tested exposures.
Sixty-nine out of the 125 studies above have been cited in previous ISAs or AQCDs.
aBoth native and introduced/naturalized plant species documented to occur in the U.S. are included.
As noted in the 2013 ISA (U.S. EPA. 2013). visible foliar injury usually occurs when sensitive
plants are exposed to elevated ozone concentrations in a predisposing environment. A major modifying
factor for ozone-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. This is because lack of soil moisture generally
decreases stomatal conductance of plants and, therefore, limits the amount of ozone entering the leaf that
can cause injury (Matvssck et al.. 2006; Panek. 2004; Grulke et al.. 2003; Panek and Goldstein. 2001;
8-23
-------
Temple et al.. 1992; Temple et al.. 1988). Consequently, many studies have shown that dry periods in
local areas tend to decrease the incidence and severity of ozone-induced visible foliar injury; therefore,
the incidence of visible foliar injury is not always higher in years and areas with higher ozone, especially
with co-occurring drought (Smith. 2012; Smith et al.. 2003). In a series of recent studies, researchers have
found their spatial models of ozone injury in California improved significantly when GIS variables of
plant water status were included (Kefauver et al.. 2013; Kefauver et al.. 2012b; Kefauver et al.. 2012a). In
a statistical modeling study, Wang et al. (2012) reported that incorporating ecological factors with ozone
exposure and soil moisture improved model predictions of foliar injury in field plots
(http://www.fia.fs.fed.us/. 1997-2007) from 24 states in Northeast and North Central U.S.
Although visible injury is a valuable indicator of the presence of phytotoxic concentrations of
ozone in ambient air, it is not always a reliable indicator of other negative effects on vegetation
[e.g., growth, reproduction; U.S. EPA (2013)1. The significance of ozone injury at the leaf and
whole-plant levels depends on how much of the total leaf area of the plant has been affected, as well as
the plant's age, size, developmental stage, and degree of functional redundancy among the existing leaf
area (U.S. EPA. 2013). Previous ozone AQCDs have noted the difficulty in relating visible foliar injury
symptoms to other vegetation effects, such as individual plant growth, stand growth, or ecosystem
characteristics (U.S. EPA. 2006. 1996). Thus, it is not presently possible to determine, with consistency
across species and environments, what degree of injury at the leaf level has significance to the vigor of
the whole plant. However, in some cases, visible foliar symptoms have been correlated with decreased
vegetative growth (Somers et al.. 1998; Karnoskv et al.. 1996; Peterson et al.. 1987; Benoitetal.. 1982)
and impaired reproductive function (Chappelka. 2002; Black et al.. 2000). Conversely, the lack of visible
injury does not always indicate a lack of phytotoxic concentrations of ozone or a lack of ozone effects
(Gregg et al.. 2006. 2003).
8.2.1 Summary and Causality Determination
Visible foliar injury from ozone exposure has been well characterized for over several decades,
using both long-term field studies and laboratory approaches. Since the 2013 Ozone ISA, new research on
bioindicator species and the further characterization of modifying factors have provided further support
for these effects. Modifying factors for ozone-induced foliar injury include soil moisture, leaf age, and
light level, genotype, and time of day of exposure. The use of biological indicators to detect phytotoxic
levels of ozone is a longstanding and effective methodology, and recent evidence is supported by more
information on sensitive species, such as the native cutleaf coneflower and the invasive tree of heaven as
useful bioindicators. New information is consistent with the conclusions of the 2013 Ozone ISA that
the body of evidence is sufficient to infer a causal relationship between ozone exposure and visible
foliar injury.
8-24
-------
Table 8-4 Ozone exposure and foliar injury.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Foliar Injury
Grantzet al. (2013)
Greenhouse; Kearney
Research and Extension
Center, Parlier, CA
(36.598°N, 119.503°W)
Gossypium Each plant was exposed to a single
barbadense. (Pima 15-min pulse of O3 (0.0, 0.5, 1.0, 1.5,
cotton) 2.0 pmol/mol). Pulses were done at
2-h intervals throughout the daylight
period. After a single pulse, plants
were returned to greenhouse bench
and left undisturbed for 6 days
Plant sensitivity (chlorophyll content, stomatal
conductance, and noninjured leaf area) showed
a clear diel trend with greatest sensitivity
occurring in midafternoon. This trend was not
closely related to gas exchange, whole-leaf
ascorbate, or total antioxidant capacity.
Yendrek et al.
(2015)
Greenhouse; Urbana, IL Pisum sativum (garden Six growth chambers—three
pea), Glycine max
(soybean), and
Phaseolus vulgaris
(common bean)
fumigated with 151 ppb O3 for 8 h for
45 days; three maintained at
ambient with avg of 12.5 ppb
The garden pea displayed no visual signs of O3
damage, in contrast to soybean and common
bean, both of which had signs of chlorosis. More
extensive O3 damage was observed in the
common bean, including leaf bronzing and
necrosis.
Burton et al. (2016)
Greenhouse; North
Carolina State
University in Raleigh,
NC (36.3°N, 78.683°W)
Glycine max (soybean, 5 days of exposure in greenhouse
two genotypes: tolerant chambers, 7 h/day, at 68-72 ppb O3
Fiskeby III and
sensitive Mandarin
[Ottawa])
Mean injury score in the mid canopy was 16%
for Fiskeby III, and 81% for Mandarin (Ottawa).
Injury scores were lower in younger leaves.
Lloyd et al. (2018) Greenhouse; University Phaseolus vulgaris O3 treatments were a combination of Nighttime O3 exposure alone, at 62 ppb, did not
Park campus of The
Pennsylvania State
University (40.806°N,
77.852°W)
(snap bean, two
genotypes; O3 resistant
[R123]and O3
sensitive [S156])
O3 concentration and treatment time
as follows: (1) 45 ppb O3, day-only;
(2) 75 ppb O3, day-only; (3) 45 ppb
O3 day + night; (4) 75 ppb O3
day + night; (5) 30 ppb night-only
treatment; and (6) 60 ppb night-only
treatment
cause foliar injury. When data were pooled
across the day and day + night exposures times,
mean daytime O3 levels at 62 ppb caused foliar
injury decreases in all three trials.
8-25
-------
Table 8-4 (Continued): Ozone exposure and foliar injury.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Foliar Injury
Smith (2012)
Field; northeastern and
north central U.S.
Most commonly
sampled bioindicator
species: Rubus
allegheniensis
(Allegheny blackberry),
Asclepias syriaca
(common milkweed),
Asclepias exaltata (tall
milkweed), Prunus
serotina (black cherry),
Fraxinus americana
(white ash), and
Apocynum
androsaemifolium
(spreading dogbane).
Ambient levels across the CONUS,
values were grouped into three
categories to represent low
(SUM06 < 10 ppm-h; N100 < 5 h),
moderate (SUM06 > 10 to 30 ppm-h;
N100 5 to 30 h), and high
(SUM06 > 30 ppm-h; N100 > 30 h)
ozone exposure conditions
Foliar injury is significantly different between
low- and high-ozone exposure sites. In all cases,
injury was greater in high-ozone sites than
low-ozone sites. The foliar injury response is
also significantly different between low ozone
exposure and high peak ozone exposure sites.
When SUM06 and N100 are relatively low, the
percentage of uninjured sites is much greater
than the percentage of injured sites; and at all
SUM06 and N100 exposures, when site
moisture is limiting, the percentage of uninjured
sites is much greater than the percentage of
injured sites.
Kefauver et al.
(2012b)
Gradient; Yosemite
National Park (YOSE)
and Sequoia and Kings
Canyon National Park
(SEKI), CA; Catalonia,
Spain
California: Pinus
ponderosa (ponderosa
pine) and Pinus jeffreyi
(Jeffrey pine)
Spain: Pinus uncinata
(mountain pine)
Ozone data were collected using
passive monitors in both YOSE and
SEKI. One U.S. EPA-certified active
monitor was colocated at YOSE and
SEKI. Average yearly O3 mixing ratio
in 2002 ranged from 35-65 ppb for
all YOSE and SEKI sites. Yearly
averages within sites were 49 ppb
for YOSE and 46 ppb for SEKI
The ozone injury index (Oil) was transferable to
other conifer species and geographic regions
(i.e., P. uncinata in Catalonia, Spain).
Species-level image classifications produced
75% accuracy for YOSE yellow pines
(i.e., Jeffrey and ponderosa pines combined)
and 82% for SEKI yellow pines. Combining
remote sensing indices with landscape GIS
variables related to plant water status resulted in
an improved regression for California sites.
Kohut et al. (2012)
Field; Rocky Mountain
National Park, CO
Rudbeckia iaciniata
var. ampia (cutleaf
coneflower),
Apocynum
androsaemifolium
(spreading dogbane),
Populus tremuloides
(quaking aspen)
Monitoring of ambient levels
2006-2010. SUM06 = 26, 28, 24,
13, 27 ppm-h. W126 = 29.6, 33.2,
28.9, 19.9, 18.9 ppm-h.
W126-3 mo = 19, 20, 18, 11,
19 ppm-h
Foliar injury in the form of ozone stipple was
found on coneflower each year. The incidence of
injured plants on sites with injury ranged from 5
to 100%. The severity of injury on affected
foliage was generally <4% but occurred on some
leaves at a level greater than 12% in 3 yr and in
1 yr on one plant at a level >75%. No foliar
ozone injury was found on spreading dogbane or
quaking aspen in any year of the survey.
8-26
-------
Table 8-4 (Continued): Ozone exposure and foliar injury.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Foliar Injury
Kefauver et al.
(2013)
Gradient; Yosemite
National Park (YOSE)
and Sequoia and Kings
Canyon National Park
(SEKI), CA; Catalonia,
Spain
California: Pinus
ponderosa (ponderosa
pine) and Pinus jeffreyi
(Jeffrey pine)
Spain: Pinus uncinata
(mountain pine)
Ozone data were collected using
passive monitors in both YOSE and
SEKI. One U.S. EPA-certified active
monitor was colocated at YOSE and
SEKI. Average yearly O3 mixing ratio
in 2002 ranged from 35-65 ppb for
all YOSE and SEKI sites. Yearly
averages within sites were 49 ppb
for YOSE and 46 ppb for SEKI
The Ozone Injury Index (Oil) was transferable to
other conifer species and geographic regions
(i.e., P. uncinata in Catalonia, Spain).
Species-level image classifications produced
75% accuracy for YOSE yellow pines
(i.e., Jeffrey and ponderosa pines combined)
and 82% for SEKI yellow pines. The stepwise
regression model of ozone injury at California
sites using remote sensing vegetation indices
combined with GIS landscape variables was
significant.
Kefauver et al.
(2012a)
Gradient; Yosemite
National Park (YOSE)
and Sequoia and Kings
Canyon National Park
(SEKI), CA; Catalonia,
Spain
California: Pinus
ponderosa (ponderosa
pine) and Pinus jeffreyi
(Jeffrey pine)
Spain: Pinus uncinata
(mountain pine)
Ozone data were collected using
passive monitors in both YOSE and
SEKI. One U.S. EPA-certified active
monitor was colocated at YOSE and
SEKI. Average yearly O3 mixing ratio
in 2002 ranged from 35-65 ppb for
all YOSE and SEKI sites. Yearly
averages within sites were 49 ppb
for YOSE and 46 ppb for SEKI
Results show that the Ozone Injury Index (Oil)
was transferable to other conifer species and
geographic regions (i.e., P. uncinata in
Catalonia, Spain). Oil by itself was poorly
correlated to ambient ozone across all sites.
Models were improved when GIS variables
related to plant-water relations were included.
Chieppa et al.
(2015)
OTC; research site
located ~5 km north of
Auburn University
campus
Pinus taeda (loblolly
pine) inoculated with
either Leptographium
terebrantis or
Grosmannia huntii
(fungal species
associated with
Southern Pine Decline)
Three ozone treatments in OTCs
(12 h/day): charcoal filtered (-0.5%
ambient air), nonfiltered air (ambient)
and 2x ambient. The first 41 days
were acclimatization then exposure,
which continued 77 more days once
seedlings were inoculated with
fungus. Mean 12 h O3 over the
118 days was 14 (CF), 23 (NF), and
37 (2x) ppb in the treatments. 12 h
AOT40 values were 0.027 (CF),
1.631 (NF) and 31.2 (2*) ppm.
Seasonal W126 values were 0.033
(CF), 0.423 (NF) and 21.913 (2x)
In elevated O3, seedlings had 9.9x more needle
injury and lower needle greenness (13.7%) than
NF chambers. The two families selected for
sensitivity to ophiostomatoid fungi had 3* more
ozone injury compared with tolerant families;
however, visible foliar injury was not affected by
inoculation status of seedlings.
8-27
-------
Table 8-4 (Continued): Ozone exposure and foliar injury.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Foliar Injury
Davis (2011)
Field; Mingo National
Wildlife Refuge in
southeastern Missouri
Asclepias syriaca
(common milkweed),
Cercis canadensis
(redbud), Cornus
florida (flowering
dogwood), Fraxinus sp.
(ash), Liquidambar
styraciflua (sweetgum),
Prunus serotina (black
cherry), Rhus
aromatica (fragrant
sumac), Rhus copallina
var. latifolia (winged
sumac), Sambucus
canadensis (black
elderberry), Sassafras
albidum (sassafras),
and Vitis sp. (wild
grape)
Cumulative ambient ozone levels
(SUM60, ppb-h) monitored at the
closest U.S. EPA monitor (Bonne
Terre, MO) at time of survey were
1998 (44,886), 2000 (39,611), 2003
(38,465), 2004 (15,147)
Across all 4 survey yr and the seven species,
102 individuals out of 1,241 (8.22%) exhibited
stipple. Percentage of bioindicator plants
exhibiting stipple were wild grape (16.1%),
common milkweed (16.0%), ash (7.5%), black
cherry (6.7%), flowering dogwood (4.9%),
sassafras (2.3%), and sweetgum (1.2%). By
year, the incidence of symptomatic plants was
1998 (22.8%), 2003 (3.9%), 2000 (3.4%), and
2004 (2.5%). The cumulative SUM60 threshold
value of ozone needed to cause foliar symptoms
on ozone-sensitive plants within the refuge
appears to be -10,000 ppb-h.
Neufeld et al. (2018)
OTC; experiments
conducted in Boone,
NC. Rhizomes collected
from Great Smoky
Mountains National Park
and Rocky Mountains
National Park.
Rudbeckia laciniata
var. ampla and var.
digitata (cutleaf
coneflower)
Three treatment groups:
charcoal-filtered air (CF), nonfiltered
air (NF), and nonfiltered air + 50 ppb
O3 (2012) or + 0 ppb/+ 50 ppb
(2013) (EO). In 2012, 24-h W126
was 0.1 ppm-h in the CF treatment,
2.0 ppm-h in the NF treatment, and
74.2 ppb in the EO treatment. 12-h
AOT40 were 0.0, 2.0, and
24.1 ppm-h, respectively. In 2013,
24-h W126 were 0.1, 1.8, and
80.5 ppm-h, respectively. 12-h
AOT40 were 1.0, 2.0, and
53.8 ppm-h, respectively. Plants
were exposed for 47 days in 2012
and for 77 days in 2013.
In 2012 and 2013, injury levels in both varieties
were higher in the EO treatment than in either
the CF or NF treatments, which did not differ, but
there were no statistically significant differences
between the varieties. Stippling increased with
time.
8-28
-------
Table 8-4 (Continued): Ozone exposure and foliar injury.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Foliar Injury
Seileretal. (2014)
Greenhouse, field; Penn
State Russell E. Larson
Agricultural Research
Center, Rock Springs,
PA
Ailanthus altissima
(tree of heaven)
2010 Greenhouse Study:
Charcoal-filtered control (<8 ppb
daily hourly avg), exposures of 75
and 120 ppb 8 h/day, 5 days/week
for 3 weeks 2011
Greenhouse Study: Charcoal-filtered
control (<8 ppb daily hourly avg),
exposures of 40, 60, 80,100,
120 ppb, 5 days/week for 4 weeks
2011 Field Study: Seasonal
SUM06 = 15.2 ppm-h., 3-mo 12-h
W126 = 11.6 ppm-h. 3-yr mean of
the 4th-highest daily max 8-h
avg = 71.5 ppb for 2009-2011 and
1-yr 4th-highest daily max 8-h avg
ozone concentration of 76 ppb.
Calculated from U.S. EPA
Aerometric Information Retrieval
System located about 1.1 km
northeast of field site
In greenhouse exposures, foliar injury occurred
at 8-h avg O3 exposure levels of 60 to 120 ppb,
with greater injury corresponding to higher
exposures. In the field, an ambient O3 3-mo,
12-h W126 of 11.6 ppm-h induced foliar injury.
CF = charcoal-filtered air; EO = air + ozone treatment; |jmol/mol = micromoles/mole; N100 = the number of hours of ozone > 100 ppb; NF = nonfiltered air; 03 = ozone;
OTC = open-top chamber; ppb = parts per billion; ppm = parts per million; SUM06 = seasonal sum of all hourly average concentrations > 0.06 ppm; SUM60 = sum of hourly ozone
concentrations equal to or greater than 60 ppb; W126 = cumulative integrated exposure index with a sigmoidal weighting function; W126-3 mo = the running max 3-month,
cumulative 12-h W126 weighted value.
8-29
-------
8.3 Plant Growth
In the 2013 Ozone ISA, the evidence was sufficient to conclude that there is a causal relationship
between ambient ozone exposure and reduced growth of native woody and herbaceous vegetation (U.S.
EPA. 2013). The 2013 Ozone ISA and previous ozone AQCDs concluded that there is strong and
consistent evidence that exposure to ozone decreases photosynthesis and growth in numerous plant
species (U.S. EPA. 2013. 2006. 1996. 1986). The evidence available at that time and discussed here found
that ambient ozone concentrations cause decreased growth (measured as biomass accumulation) in
annual, perennial, and woody plants, inclusive of crops, annuals, grasses, shrubs, and trees. A
meta-analysis by Wittig et al. (2009) found that average ozone exposures of 40 ppb significantly
decreased annual total biomass by 7% across 263 studies. Annual biomass declines were reported to be
greater (11 to 17%) with elevated ozone exposures [avg of 97 ppb; Wittig et al. (2009)1. Biomass declines
were linked to reductions in photosynthesis [Section 9.3.5.1 in U.S. EPA (2013)1. which are consistent
with cumulative uptake of ozone into the leaf I Wittig et al. (2007); Figure 8-21. Further, there is evidence
ozone may change plant growth patterns by significantly reducing carbon allocated to roots in some
species (Jones et al.. 2010; Wittig et al.. 2009; Grantz et al.. 2006; Andersen. 2003; King et al.. 2001).
As described in the PECOS tool (Table 8-2). the scope for new evidence reviewed in this section
is limited to studies conducted in North America at ozone concentrations occurring in the environment or
experimental ozone concentrations within an order of magnitude of recent concentrations. Plant growth,
which is the increase in biomass over a period of time, may be determined using a number of metrics
(e.g., the change in height, stem volume, leaf area, canopy density, or weight of plant material), and
studies that examined these metrics were selected for review. The 2013 ISA broadly defined plant growth
to include effects on plant reproduction. However, due to increased research and synthesis of ozone
effects on plant reproduction, these endpoints are reviewed separately in Section 8.4.
8.3.1 Declines in Growth Rates
Since the 2013 Ozone ISA, there is more evidence from manipulative experiments that supports
known detrimental effects of ozone on plant growth as well as models built with empirical data that scale
up ozone exposure-response relationships of growth reductions to put these losses in context [e.g., Capps
et al. (2016); Lapina et al. (2016); Table 8-61.
• Results from the aspen-only (Populus tremuloides) stands at the Aspen FACE ("free air" ozone
and CO2 exposure) experiment in Wisconsin show a decrease of 12-19% in the relative growth
rate of three of five genotypes of aspen studied. Trees were exposed to hourly ozone levels
1.5 times ambient over the period from 1998-2008. The annual cumulative ambient W126
exposures ranged from 2.1 to 8.8 ppm-hour, and elevated treatments ranged from 12.7 to
35.1 ppm-hour (Moran and Kubiske. 2013).
8-30
-------
• When site-level results from the Aspen FACE experiment were scaled up using the forest
landscape model (LANDIS II), ozone was found to significantly reduce landscape biomass
(Gustafson et al.. 2013).
• A meta-analysis of nine studies (inclusive of the Aspen FACE experiments) examining
intraspecific variation in juvenile tree growth under elevated ozone found that elevated ozone
generally reduced photosynthetic rate as well as height growth and stem volume (metrics used for
biomass calculations and tracking growth rates) in multiple genotypes of silver birch (Betula
pendula), quaking aspen (Populus tremuloides), and poplar hybrids (Resco de Dios et al.. 2016).
• A study using the invasive Chinese tallow tree (Triadica sebifera) suggests ozone response may
be genotype-specific; elevated ozone decreased both root and total biomass from U.S. seed
sources, but had no effect on the biomass of Chinese seed sources (Wang et al.. 2018).
• Using simulations of the GEOS-Chem model for 2010 data coupled with established U.S. EPA
ozone exposure-response functions in seedlings, Lapina et al. (2016) estimated relative biomass
loss at 2.5% for ponderosa pine (Pinusponderosa) and 2.9% for aspen (Populus tremuloides)
across the continental U.S.
• In another estimation of biomass loss of adult tree species across the U.S. for modeled spatially
explicit ozone values, eastern cottonwood (Populus deltoides) and black cherry (Prunus serotina)
show large annual losses in biomass under the authors' reference scenario (ambient ozone levels,
W126 range 0-56 ppm-hour), 32, and 10%, respectively. Black cherry exhibits the greatest
annual loss (2,210 tons of biomass/ha) of the 11 tree species studied, with twice the biomass loss
potential of either eastern cottonwood or ponderosa pine. Biomass of quaking aspen (Populus
tremuloides), tulip poplar (Liriodendron tulipifera), and various pine species also respond
negatively to ozone, with losses ranging from 0.3-1.9% (Capps et al.. 2016).
In addition to these studies, there is a recent global-scale synthesis of published ozone exposure
studies that documents reductions in biomass due to ozone exposure in over a hundred species (Berg man n
et al.. 2017V Many of these species have populations native to the U.S., and a comprehensive list of U.S.
species identified by Bcrgmann et al. (2017) as sensitive to ozone are presented below in Table 8-5.
8-31
-------
Table 8-5 Plant species that have populations in the U.S. (USDA. 2015) that
have been tested for ozone growth reduction as documented in the
references listed with each species and synthesized in Berqmann et
al. (2017).ab
Species
Ozone
Reduces
Growth
References
Acer rubrum
Y
Davis and Skellv (1992): Simini et al. (1992): Findlev et al. (1996)
Acer saccharum
Y
Gaucher et al. (2005): Pell et al. (1999): Rebbeck (1996a): Laurence
et al. (1996): Kress and Skellv (1982): Noble et al. (1992): Tioelker et
al. (1993)
Achillea millefolium
N
Bender et al. (2002): Bunaener et al. (1999a): Bunaener et al.
(1999b)
Agropyron smithii
Y
Volin et al. (1998)
Agrostis vinealis
N
Haves et al. (2006)
Alnus incana
Y
Mortensen and Skre (1990): Lorenz et al. (2005): Bussotti and
Gerosa (2002): De Vries et al. (2003): Manninq et al. (2002):
Ozolincius and Serafinaviciute (2003)
Andropogon gerardii
Y
Lewis et al. (2006)
Angelica archangelica
Y
Mortensen (1993)
Antennaria dioica
Y
Mortensen (1993)
Apocynum androsaemifolium
N
Berqweiler and Manninq (1999): Davis (2007a): Davis (2007b)
Armeria maritima
Y
Haves et al. (2006)
Campanula rotundifolia
Y
Ashmore et al. (1995): Haves et al. (2006): Mortensen and Nilsen
(1992): Ashmore et al. (1996)
Carex arenaria
Y
Jones et al. (2010)
Carex atrofusca
Y
Mortensen (1994b)
Carex echinata
N
Haves et al. (2006)
Carex nigra
N
Franzarinq et al. (2000)
Chamerion angustifolium
Y
Skellv et al. (1999): Mortensen (1993)
Chenopodium album
Y
Bender et al. (2006): Berqmann et al. (1995): Berqmann et al.
(1999): Pleiiel and Danielsson (1997): Reilinq and Davison (1992):
Romaneckiene et al. (2008)
Cirsium acaule
Y
Warwick and Tavlor (1995)
8-32
-------
Table 8-5 (Continued): Plant species that have populations in the U.S. (USDA,
2015) that have been tested for ozone growth reduction as
documented in the references listed with each species and
synthesized in Bergmann et al. (2017).ab
Species
Ozone
Reduces
Growth
References
Comarum palustre
Y
Battv et al. (2001): Mortensen (1994b)
Echinacea purpurea
Y
Szantoi et al. (2007)
Eriophorum angustifolium
N
Haves et al. (2006): Mortensen (1994b)
Festuca ovina
Y
Ashmore et al. (1995): Haves et al. (2006): Pleiiel and Danielsson
(1997): Reilina and Davison (1992): Warwick and Tavlor (1995):
Ashmore et al. (1996)
Festuca rubra
Y
Ashmore et al. (1995): Bunqener et al. (1999b): Bunqener et al.
(1999a): Haves et al. (2006): Mortensen (1992): Ashmore et al.
(1996):
Fragaria vesca
Y
Mortensen (1993)
Fraxinus americana
Y
Kress and Skellv (1982): Hildebrand et al. (1996)
Fraxinus pennsylvanica
Y
Kress and Skellv (1982): Lorenz et al. (2005)
Geum rivale
Y
Battv etal. (2001)
Juncus effusus
Y
Haves et al. (2006)
Liquidambar styraciflua
Y
Kress and Skellv (1982): Davis (2011)
Liriodendron tulipifera
Y
Rebbeck (1996a): Rebbeck (1996b): Cannon Jr et al. (1993): Simini
et al. (1992): Kress and Skellv (1982): Davis and Skellv (1992):
ChaDDelka et al. (1999b): Hildebrand et al. (1996)
Melica nitens
N
Mortensen (1994b)
Oxalis acetosella
N
Haves et al. (2006)
Oxyria digyna
Y
Haves et al. (2006): Mortensen and Nilsen (1992): Mortensen (1993)
Paspalum notatum
Y
Muntiferinq et al. (2000)
Phleum alpinum/commutatum
Y
Pleiiel and Danielsson (1997): Danielsson et al. (1999): Mortensen
(1993)
Picea glauca
N
Mortensen (1994a)
Picea rubens
Y
Finnveden (2000): Amundson et al. (1991): Laurence et al. (1997)
Picea sitchensis
Y
Lucas et al. (1988): Lucas et al. (1993): Mortensen (1994a)
Pinus contorta
N
Mortensen (1994a): Williams et al. (1977)
8-33
-------
Table 8-5 (Continued): Plant species that have populations in the U.S. (USDA,
2015) that have been tested for ozone growth reduction as
documented in the references listed with each species and
synthesized in Bergmann et al. (2017).ab
Species
Ozone
Reduces
Growth
References
Pinus echinata
Y
Shelburne et al. (1993)
Pinus ellioti
Y
Evans and Fitzaerald (1993): Dean and Johnson (1992): Bvres et al.
(1992)
Pinus ponderosa
Y
Takemoto et al. (1997): Temple and Miller (1994): Temple et al.
(1993): Bevers et al. (1992): Temple et al. (1992): Fenn et al. (2002):
Jones and Paine (2006): Williams and Macareaor (1975): Williams et
al. (1977)
Pinus pungens
N
Neufeld et al. (2000)
Pinus rigida
Y
Kress and Skellv (1982)
Pinus taeda
Y
Kress and Skellv (1982): Edwards et al. (1992): Qiu et al. (1992):
Adams and O'Neill (1991): Adams et al. (1990): Shafer et al. (1993):
Spence et al. (1990): Wiseloqel et al. (1991): Chappelka et al. (1990)
Pinus virginiana
N
Neufeld et al. (2000): Kress and Skellv (1982)
Piatanus occidentaiis
Y
Kress and Skellv (1982): Kline et al. (2008)
Poa pratensis
Y
Bender et al. (2002): Bender et al. (2006): Bunaener et al. (1999b):
Bunaener et al. (1999a): Mortensen (1992): Ashmore et al. (1996)
Polygonum viviparum
Y
Mortensen and Nilsen (1992)
Popuius deitoides
Y
Wana et al. (1986)
Popuius tremuioides
Y
Volin et al. (1998): Karnoskv et al. (1999): Karnoskv et al. (1996):
Coleman et al. (1996): Yun and Laurence (1999)
Prunus serotina
Y
Skellv et al. (1999): Vanderhevden et al. (2001): Gunthardt-Goera et
al. (1999): Pell et al. (1999): Rebbeck (1996a): Rebbeck (1996b):
Neufeld et al. (1995): Simini et al. (1992): Davis and Skellv (1992):
Chappelka et al. (1997): Chappelka et al. (1999b): Chappelka et al.
(1999a): Davis and Orendovici (2006): Davis (2007a): Davis (2007b):
Davis (2011): Hildebrand et al. (1996): De Bauer et al. (2000):
Bussotti and Gerosa (2002): Yuska et al. (2003)
Pseudotsuga menziesii
Y
Runeckles and Wriaht (1996): Mortensen (1994a)
Quercus pheiios
Y
Kress and Skellv (1982)
Quercus rubra
Y
Volin et al. (1998): Pell et al. (1999): Samuelson et al. (1996): Keltina
et al. (1995): Edwards et al. (1994): Simini et al. (1992): Davis and
Skellv (1992)
Ranunculus acris
Y
Haves et al. (2006): Wvness et al. (2011): Mortensen (1993)
8-34
-------
Table 8-5 (Continued): Plant species that have populations in the U.S. (USDA,
2015) that have been tested for ozone growth reduction as
documented in the references listed with each species and
synthesized in Bergmann et al. (2017).ab
Species
Ozone
Reduces
Growth
References
Robinia pseudoacacia
Y
Skellv et al. (1999): Wanq et al. (1986): Lorenz et al. (2005): Bussotti
et al. (2003a): Bussotti and Gerosa (2002): Bussotti et al. (2006):
Bussotti et al. (2003b): De Vries et al. (2003)
Rudbeckia laciniata
Y
Szantoi et al. (2009): ChaDDelka et al. (2003): Davison et al. (2003)
Rumex acetosa
Y
Battv et al. (2001): Bender et al. (2002): Bender et al. (2006):
Beramann et al. (1999): Pleiiel and Danielsson (1997): Mannina and
Godzik (2004): Reilina and Davison (1992): Ashmore et al. (1996):
Mortensen (1993): Haves et al. (2006)
Schizachyrium scoparium
Y
Powell et al. (2003): Volin et al. (1998)
Scirpus cespitosus
Y
Haves et al. (2006)
Silene acaulis
Y
Mortensen and Nilsen (1992)
Solanum nigrum
Y
Bender et al. (2006): Beramann et al. (1995): Beramann et al.
(1999): Beramann et al. (1996a)
Solidago albopilosa
Y
Mavitvand Berrana (1994)
Solidago virgaurea
Y
Mortensen and Nilsen (1992): Mortensen (1993)
Spartina alterniflora
Y
Tavlor (2002)
Stachys palustris
Y
Battv etal. (2001)
Urtica dioica
Y
Bender et al. (2006): Beramann et al. (1999): Reilina and Davison
(1992): Beramann et al. (1996a): Bussotti et al. (2003a)
In ozone-response categories, Y = ozone induces effect at tested exposures, N = ozone has no effect at tested exposures
Sixty-five out of the 108 studies above have been cited in previous ISAs or AQCDs.
aBoth native and introduced/naturalized plant species documented to occur in the U.S. are included.
bData are found in the Supplemental Information in this publication.
8-35
-------
8.3.2
Changes in Biomass Allocation
In addition to declines in plant growth rates, ozone alters patterns of carbon allocation, both
belowground and aboveground (the portion of energy expended by the plant toward roots, stems, or
leaves; see Figure 8-2). Changes in biomass allocation alter plant nutrient uptake, plant water use, and
carbon fixation.
• Over the course of the Aspen FACE experiment (1998-2008), the effects of ozone on plant
carbon allocation were dynamic through time and varied among the forest communities (Talhclm
et al.. 2014; Pregitzer and Talhelm. 2013; Rhea and King. 2012). Elevated ozone consistently
suppressed leaf production in each of the three communities. There were effects on root biomass
in 2006 consistent with Aspen FACE studies of previous years, with elevated ozone increasing
small root (0-2 mm diameter) biomass in the aspen-only rings and decreasing small root biomass
in the aspen-birch rings (Rhea and King. 2012). There were also effects of ozone on the
distribution of roots across the soil profile, which are discussed in more detail in Section 8.9.2.
• Shifts in wood anatomy (change in growth, cell size, vessel density, and proportion) also occurred
with elevated ozone at the Aspen FACE site (Kostiaincn et al.. 2014). Elevated ozone
significantly decreased radial growth and diameters of wood fibers and vessels in quaking aspen.
Most treatment responses were observed in the early phase of the experiment, indicating
ontogenetic changes during wood maturation that are consistent with shifts in the trees' metabolic
priority from growth to hydraulic transport in response to ozone.
• A study of the effects of short-term ozone exposure on loblolly pine seedlings found positive
effects on aboveground growth, but the study authors attribute this finding to reduction in
photosynthate transport to roots, which contributed to declines in seedling vigor (Chieppa et al..
2015). Even with the increased aboveground growth observed, ozone alterations to carbon
transport and subsequent declines in seedling vigor and longevity may have negative impacts on
forest establishment and regeneration.
8.3.3 Connections with Community Composition and Water Cycling
Studies published since the 2013 Ozone ISA have provided insight on ozone-mediated alterations
to biomass allocations within an individual plant that are relevant to whole-plant growth and function.
Additionally, the studies provide context for scaling up the long-known detrimental effects of ozone on
photosynthesis and growth in numerous plant species to changes at the community and ecosystem level.
Decreases in photosynthesis due to ozone are well studied and quantified and are directly related to
declines in plant biomass discussed here. Ozone-caused declines in canopy density and leaf area index, an
important component of plant biomass, have similarly been well studied, (see Section 8.1.1.). These
effects were, however, thoroughly reviewed in the 2013 Ozone ISA (U.S. EPA. 2013). and studies
continue to be published in this area (U.S. EPA. 2008).
• Species-specific responses to ozone in terms of plant growth reductions and biomass allocation
are a possible mechanism for community shifts. In a model simulation of long-term effects of
8-36
-------
ozone on a typical forest in the southeastern U.S. involving different tree species with varying
ozone sensitivity, Wang et al. (2016) found that ozone significantly altered forest community
composition and decreased plant biodiversity. Using published peer-reviewed data to place tree
species into three sensitivity classes, Wang et al. (2016) applied either a 0, 10, or 20% growth
reduction to species in the University of Virginia Forest Model Enhanced (UVAFME), a gap
model which tracks the growth and survival of individual trees and species within a stand. Over
the 500-year simulation, ozone-resistant species like white oak (Quercus alba) and American
beech (Fagus grandifolia) dominate, and sensitive species like tulip poplar (Liriodendron
tulipifera) and red maple (Acer rubrum) decline. Although the communities changed
significantly, overall forest biomass and forest carbon storage did not decrease under high ozone
conditions because tolerant species growth was enhanced after these species were freed from
competition by the loss of ozone-sensitive species. The terrestrial community composition section
(see Section 8.10) contains more information about scaling up biomass response of individual
species and examining the ensuing compositional changes.
• Variable growth response between species may be a concern in agricultural systems as well. In a
study of ozone's effects on the noxious weed Palmer's pigweed (Amaranthus palmeri), elevated
ozone exposure and water stress had no effect on the daytime stomatal conductance, shoot
growth, and root growth of this plant. The authors suggest that this weed species may have much
higher tolerance to elevated ozone and moisture stress compared with crops, and therefore may
become a more serious pest in the future because of this competitive advantage (Paudcl et al..
2016).
• Changes in forest biomass may also affect ecosystem water use (see Section 8.11). Statistical
models examining climate and ozone effects on late-season streamflow in several Appalachian
forest watersheds were also found to accurately predict measurements of annual tree ring growth
over 20 years in five native species in these forests—an important mechanistic step in
understanding ecosystem-level effects of ozone exposure. The findings highlight the negative
effects of ozone on tree growth and explicitly connect these declines to tree water use and
seasonal watershed dynamics (Sun et al.. 2012).
8.3.4 Summary and Causality Determination
Previous studies showed strong and consistent evidence that ambient ozone concentrations cause
decreased growth and biomass accumulation in annual, perennial, and woody plants, inclusive of crops,
annuals, grasses, shrubs, and trees. Since the 2013 Ozone ISA, more evidence that supports this causal
relationship has been published. In addition to reductions in plant growth rates, numerous studies from
different ecosystems have found that ozone significantly changes patterns of carbon allocation
belowground and aboveground which also supports previous knowledge. This evidence contributes to the
understanding of ozone's effects on plant growth, biomass allocation, and development. Previous
evidence and new evidence reviewed here is sufficient to infer a causal relationship between ozone
exposure and reduced plant growth.
8-37
-------
Table 8-6 Ozone exposure and plant growth and biomass.
Study Study Type and Location Study Species Ozone Exposure Effects on Plant Growth and Biomass
Paudel et al. (2016) Greenhouse; Parlier, CA Amaranthus palmeri Two runs of exposure 30 Elevated O3 exposure and water stress had no effect
(Palmer amaranth) and 35 days. 12-h means of on daytime stomatal conductance, shoot growth, and
4, 59, and 114 ppb root growth. This weed species may have much more
tolerance to elevated O3 and moisture stress compared
to crops that it competes with.
Sun et al. (2012)
Gradient; six watersheds in
Appalachian mixed
deciduous forests: Walker
Branch and Little River
(eastern Tennessee),
Cataloochee Creek
(western North Carolina),
James River and New River
(Virginia), and Fernow
Experimental Forest (West
Virginia)
Mixed tree species in
eastern forests
AOT60 at each watershed:
1.72 (WBWS), 2.6 (LR),
1.72 (CC), 0.82 (NR), 0.83
(JR), 0.74 (FEW); max
hourly (in ppb): 68.2
(WBWS), 67.8 (LR), 68.2
(CC), 59.4 (NR), 58.7 (JR),
58.8 (FEW)
Empirical statistical models from data collected in six
watersheds in Tennessee, North Carolina, Virginia,
and West Virginia found that O3 and climate are both
significant predictors of late-season stream flow in
forests, and models incorporating these environmental
parameters also fit measurements of annual tree ring
growth, which is an important mechanistic step in
ozone effects on forested watersheds.
Rhea and King
(2012)
FACE; Aspen FACE, near
Rhinelander, Wl (45.7°N,
89.5° W)
Populus tremuloides
(quaking aspen),
Betula papyrifera
(paper birch)
Treatments up until the
2005 (when root samples
were taken): ambient
average W126 was
5.2 ppm-h and elevated O3
was 27.3 ppm-h. For hourly
ozone concentrations
during experimental ozone
treatment, see Kubiske and
Foss (2015).
Fine-root biomass in all-aspen (AA) and aspen-birch
(AB) plots fumigated with ozone differed by community
and soil depth. Biomass increased with depth in the AA
(aspen clones) community by 10.2, 36.4, and 34.8% in
the upper, middle, and lower soil layer relative to the
control. In the AB community, root biomass decreased
16% in the shallow layer with a small increase at the
middle soil layer, resulting in a total decrease of 11 %
across all layers. Total root length increased in the AA
community to a greater extent than the AB community
where smaller increases and some decreases were
observed. A 33% decrease in root tissue density was
observed across all soil layers in trees exposed to O3.
Specific root length increased with soil depth and O3,
with the greatest increases in the AA community.
8-38
-------
Table 8-6 (Continued): Ozone exposure and plant growth and biomass.
Study
Study Type and Location
Study Species
Ozone Exposure
Effects on Plant Growth and Biomass
Kostiainen et al.
(2014)
FACE; Aspen FACE, near
Rhinelander, Wl (45.7°N,
89.5° W)
Populus tremuloides
(quaking aspen),
Betula papyrifera
(paper birch)
Growing seasons
1998-2008. Ambient W126
2.1-8.8 ppm-h and
elevated 12.7-35.1 ppm-h.
For hourly ozone
concentrations during
experimental ozone
treatment, see Kubiske and
Foss (2015).
Elevated CO2 increased radial growth and cell
diameters in aspen, while vessel density and
proportion decreased. Elevated O3 decreased growth
and cell diameters, but increased vessel density and
proportion. Neither CO2 nor O3 responses were
consistent during the experiment. 63 exposed trees
had more and narrower vessels, which were packed
more densely per unit wood area. In birch, the
treatments had no major effects on wood anatomy or
wood density.
Talhelm et al. (2014)
FACE; Aspen FACE, near
Rhinelander, Wl (45.7°N,
89.5°W).
Populus tremuloides
(quaking aspen),
Betula papyrifera
(paper birch), Acer
saccharum (sugar
maple)
12 rings, factorial CO2 x O3
with three chamber reps.
Ambient O3
W126 = 2.1-8.8 ppm-h,
Elevated = 12.7-35.1
ppm-h. For hourly ozone
concentrations during
experimental ozone
treatment, see Kubiske and
Foss (2015).
Over the 11 yr of the experiment, O3 significantly
reduced C content of stems and branches by 17%; and
NPP by 10% however O3 effects on NPP disappeared
during final 7 yr of study; O3 shifted fine roots toward
soil surface.
Moran and Kubiske
(2013)
FACE; Aspen FACE, near
Rhinelander, Wl (45.7°N,
89.5°W).
Clones of five
genotypes of Populus
tremuloides (quaking
aspen), from the
aspen-only sections of
the experiment,
1997-2008
Full factorial: O3 and CO2,
1998-2008. Ozone:
ambient W126 = 2.1-8.8
ppm-h, elevated
W126 = 12.7-35.1 ppm-h.
CO2: ambient (360 ppm) or
elevated (560 ppm). For
hourly ozone
concentrations during
experimental ozone
treatment, see Kubiske and
Foss (2015).
Elevated O3 decreases relative growth rate by 12-19%
in three of the five genotypes.
8-39
-------
Table 8-6 (Continued): Ozone exposure and plant growth and biomass.
Study
Study Type and Location
Study Species
Ozone Exposure
Effects on Plant Growth and Biomass
Gustafson et al.
(2013)
Model; Rhinelander, Wl
Betula papyrifera
(paper birch), Acer
saccharum (sugar
maple), and four
clones of Populus
tremuloides (quaking
aspen)
Ambient O3
W126 = 2.1-8.8 ppm-h and
elevated = 12.7-35.1
ppm-h. Three chamber reps
for each treatment, control,
+CO2, +O3, and +CO2+O3.
For hourly ozone
concentrations during
experimental ozone
treatment, see Kubiske and
Foss (2015).
Site-level results from the Aspen FACE experiment
were scaled up using the forest landscape model
(LANDIS II). +O3 reduced landscape biomass and the
+CO2+O3 treatment was similar to the control; Total
biomass was always lowest under the O3 treatment.
Chieppa et al. (2015)
OTC; research site located
~5 km north of Auburn
University campus
Pinus taeda (loblolly
pine) inoculated with
either Leptographium
terebrantis or
Grosmannia huntii
(fungal species
associated with
Southern Pine
Decline)
Three ozone treatments in
OTCs (12 h/day):
CF(~0.5% ambient air), NF,
and 2x ambient, first
41 days were
acclimatization then
exposure continued
77 more days once
seedlings were inoculated
with fungus. Mean 12-h O3
over the 118 days was
14 (CF), 23 (NF), and
37 (2x) ppb. 12-h AOT40
values were 0.027 (CF),
1.631 (NF), and 31.2
(2x) ppm-h. Seasonal
W126 values were
0.03 (CF), 0.423 (NF) and
21.9 (2x) ppm-h.
Four loblolly pine families (two tolerant and two
susceptible) were inoculated with root-infecting
ophiostomatoid fungi following preacclimation to ozone
(41 days). Seedling growth was not affected by
inoculation but was affected by O3. Seedling volume
under 2x O3 increased significantly compared with CF
and NF seedlings and had greater aboveground and
total dry matter yield than CF seedlings.
8-40
-------
Table 8-6 (Continued): Ozone exposure and plant growth and biomass.
Study
Study Type and Location
Study Species
Ozone Exposure
Effects on Plant Growth and Biomass
Neufeld et al. (2018)
OTC; experiments
conducted in Boone, NC.
Rhizomes collected from
Great Smoky Mountains
National Park and Rocky
Mountains National Park
Rudbeckia laciniata
var. am pia and var.
digitata (cutleaf
coneflower)
Three treatment groups:
charcoal-filtered air (CF),
nonfiltered air(NF), and
nonfiltered air + 50 ppb O3
(2012) or +30 ppb/+ 50 ppb
(2013) (EO). In 2012, 24-h
W126 was 0.1 ppm-h in the
CF treatment, 2.0 ppm-h in
the NF treatment, and
74.2 ppb in the EO
treatment. 12-h AOT40
were 0.0, 2.0, and
24.1 ppm-h, respectively. In
2013, 24-h W126 were 0.1,
1.8, and 80.5 ppm-h,
respectively. 12-h AOT40
were 1.0, 2.0, and
53.8 ppm-h, respectively.
Plants were exposed for
47 days in 2012 and for
77 days in 2013.
In 2012 and 2013, injury levels in both varieties were
higher in the EO treatment than in either the CF or NF
treatments, which did not differ, but there were no
statistically significant differences between the
varieties. Stippling increased with time. Effects of O3
on biomass accumulation were not significant.
Wang et al. (2018)
Greenhouse; Rice
University, TX, with seeds
collected from trees in
China and North America
Triadica sebifera
(tallow tree)
seedlings, grown from
seeds collected in
native Chinese range
and invasive
American range
Factorial O3 by CO2
experiment for 78 days:
ambient O3 and ambient
CO2; elevated O3
(100 ppb); elevated CO2
(800 ppm); elevated
O3 + elevated CO2
Elevated O3 decreases U.S.-sourced T. sebifera root
and total biomass but does not affect the biomass of
plants grown from seed sourced from China.
8-41
-------
Table 8-6 (Continued): Ozone exposure and plant growth and biomass.
Study
Study Type and Location
Study Species
Ozone Exposure
Effects on Plant Growth and Biomass
Lapina et al. (2016) Model; continental U.S.
Pinus ponderosa
(ponderosa pine) and
Populus tremuloides
(quaking aspen)
Exposure response
functions for W126, AOT40,
and mean metric (MX) for
total ozone exposure were
used to model relative loss
estimates. The
GEOS-Chem adjoint model
was applied to estimate
source-receptor
relationships between tree
biomass loss and emission
sources.
This analysis of year 2010 data coupled with
established U.S. EPA ozone exposure-response
functions in seedlings estimates a nationwide biomass
loss of 2.5% for ponderosa pine and 2.9% for aspen.
Capps et al. (2016)
Model; continental U.S.
(CONUS)
Populus deltoides
(eastern cottonwood),
Prunus serotina (black
cherry), Populus
tremuloides (quaking
aspen), Pinus
ponderosa
(ponderosa pine),
Liriodendron tulipifera
(tulip poplar), Pinus
strobus (eastern white
pine), Pinus virginiana
(Virginia pine), Acer
rubrum (red maple),
Alnus rubra (red
alder)
Uses U.S. EPA-developed
CMAQ model to model
exposure values of W126
under three regulatory
scenarios of maximum local
decreases in W126: 1.3, 4,
and 5.3%, as well as a
reference (ambient, W126
range 0-56 ppm-h) over
CONUS. See Figure 3 in
paper.
Eastern cottonwood and black cherry show noticeable
potential productivity losses of 32 and 10%,
respectively, at ambient O3 concentrations. Black
cherry shows the greatest potential productivity losses
of 2,210 tons of biomass per hectare with twice the
biomass loss potential of either eastern cottonwood or
ponderosa pine. The quaking aspen, tulip poplar, and
various pine species also respond to ozone with
potential productivity losses ranging from 0.3 to 1.9%.
AOT40 = seasonal sum of the difference between an hourly concentration at the threshold value of 40 ppb, minus the threshold value of 40 ppb; AOT60 = seasonal sum of the
difference between an hourly concentration at the threshold value of 60 ppb, minus the threshold value of 60 ppb; C = carbon; C02 = carbon dioxide; FACE = free-air C02
enrichment; NPP = net primary production; ppb = parts per billion; ppm = parts per million; W126 = cumulative integrated exposure index with a sigmoidal weighting function.
8-42
-------
8.4
Plant Reproduction, Phenology, and Mortality
In the 2013 Ozone ISA, there was no separate determination of causality for plant reproduction.
Rather, evidence was sufficient to conclude that a causal relationship exists between ozone exposure and
reduced plant growth, which at the time was broadly defined to encompass plant reproduction (U.S. EPA.
2013). In this ISA, due to increased research and synthesis of ozone effects on plant reproduction,
evidence was evaluated for a separate causal statement for this endpoint. Studies in the 2013 Ozone ISA
were in agreement with previous research reviewed in the 2006 Ozone AQCD and by Black et al. (2000).
which included research going back to the 1970s, showing that ozone can affect plant reproductive
function (U.S. EPA. 2013. 2006). For instance, paper birches (Betulapapyrifera) at Aspen FACE that
were exposed to years-long elevated ozone produced male flowers more frequently but produced seeds of
lower weight that germinated less often than seeds from trees in ambient conditions (Darbah et al.. 2008;
Darbah et al.. 2007). Additional research reviewed here strengthens the evidence that ozone affects plant
reproduction, including newly summarized ozone response across species for a suite of reproductive
metrics (Leisner and Ainsworth. 2012). New experiments have also isolated the effects of ozone on
specific reproductive tissues and relative to reproductive developmental events. This evidence reinforces
the previous understanding of the biological mechanisms for effects which are classified as either "direct"
from exposure of reproductive tissues to ozone or "indirect" from reduction in photosynthesis that results
in fewer total available resources to invest in flowers or seeds (Figure 8-2).
In the 2013 Ozone ISA, there was no determination of causality for plant phenology (i.e., timing
of flowering or germination) or mortality (U.S. EPA. 2013). Black et al. (2000) reviewed the effects of
ozone on the timing of flowering and germination, noting that responses vary considerably across species.
Since then, new studies on the effects of ozone on phenology have been published, and those studies
continue to show less consistent results than the studies on plant reproduction. With respect to plant
mortality, the 2006 Ozone AQCD and 2013 Ozone ISA included studies identifying ozone as a
contributor to tree mortality, however causality was not assessed (U.S. EPA. 2013. 2006). Additional
studies are reviewed here (Table 8-8). including one large multivariate analysis that showed ozone
significantly increases mortality of trees in 7 out of 10 plant functional types that make up eastern and
central U.S. forests (Dietze and Moorcroft. 2011).
As described in the PECOS tool (Table 8-2). the scope for this section includes studies on any
continent in which alterations in plant reproduction (e.g., flower number, fruit number, fruit weight, seed
number, rate of seed germination), phenology (e.g., timing of flowering or germination), and mortality
(i.e., the fraction of individuals in a population that die over a given interval) were measured on the scale
of individual plants in response to concentrations occurring in the environment or experimental ozone
concentrations within an order of magnitude of recent concentrations (as described in Appendix 1).
8-43
-------
8.4.1
Plant Reproduction
The recent literature shows that across most plant reproduction metrics (e.g., flower number, fruit
number, fruit weight, seed number, rate of seed germination) and exposure concentrations that ozone has
significant negative effects on plant reproduction. Although crop yield is sometimes considered a measure
of reproduction, it is discussed separately in Section 8.5. Additional studies contribute to an increasingly
refined understanding of how ozone affects plant reproduction in agricultural and horticultural species
[e.g., snap bean (P. vulgaris), cowpea (Vigna unguiculata), pepper (Capsicum annuum), and petunia
{Petunia hybrid); Yang et al. (2017); Tetteh et al. (2015); Taiaetal. (2013); Burkev et al. (2012)1.
agricultural weeds (Li et al.. 2013a). and pasture/grassland species rGundel et al. (2015); Wang et al.
(2015); Table 8-91.
• In a first of its kind study, Leisner and Ainsworth (2012) conducted a quantitative meta-analysis
to understand the general magnitude and direction of the effects of ozone exposure on plant
reproduction. Their conclusions were based on data from 128 studies and many plant species.
They categorized ozone exposure concentration using daytime means (4-, 7-, 8-, or 12-hour
daytime mean depending on data reported). Compared with charcoal-filtered air, most metrics of
plant reproduction were reduced under elevated ozone (Figure 8-3). Furthermore, compared with
ambient air (an avg of 33 ppb across all studies), all metrics of plant reproduction were reduced
under elevated ozone (Figure 8-4). For instance, in experiments that used charcoal-filtered air as
the control, seed number decreased 16% (at an avg exposure of 70 ppb), and fruit number
decreased 9% (at an avg exposure of 90 ppb). In experiments that used ambient air as the control,
average fruit weight decreased 51% (at an avg exposure of 98 ppb), which was the largest effect
observed in this part of the meta-analysis, and seed number decreased approximately 10% (at an
avg exposure of 68 ppb). Some metrics significantly decreased even under the lowest exposure
category (<40 ppb). A trend in larger negative responses under higher exposure levels existed for
some metrics of plant reproduction, including fruit number and average fruit weight when
ambient air was used as the control.
• Leisner and Ainsworth (2012) also analyzed the response in reproduction to ozone exposure of
contrasting plant types (i.e., annual vs. perennial, monocot vs. dicot, C3 vs. C4, and indeterminate
vs. determinate growth form) and found few significant differences in response magnitude or
direction. One exception was that indeterminate plants had much greater reductions in fruit
weight and fruit number than determinate plants (53.9% decrease vs. 4.4% increase, and
44.9% decrease vs. 7.1% decrease, respectively).
• Sanz et al. (2016) developed linear exposure (AOT40)-response and Phytotoxic Ozone Dose
(PODl)-response curves for reproductive biomass in ozone-sensitive species of clover (Trifolium
spp.) found in Europe. Reproduction was reduced significantly with increasing ozone exposure
(r2 = 0.45) and ozone dose (r2 = 0.69).
8-44
-------
Inflorescence number (IFN)
Flower weight (FLW)
Flower number (FLN)
Yield (Y)
Harvest index (HI)
Fruit number (FN)
Fruit weight (FW)
Seeds per fruiting structures (SF)
Total seed number (SN)
Seed weight per plant (WP)
Individual seed weight (SW)
Pollen tube (PT)
Pollen germination (PG)
-100
>
~H ^
h*H
m
m
m
m
(-•—i
i • i
(21, 103 ppb)
(28, 43 ppb)
(30, 82 ppb)
(112, 60 ppb)
(59, 63 ppb)
(170, 90 ppb)
(142, 78 ppb)
(79, 84 ppb)
(96, 70 ppb)
(111,71 ppb)
(296, 68 ppb)
(24, 339 ppb)
(45, 409 ppb)
-50 0 50 100
Percentage change from charcoal-filtered air (%)
150 300
ppb = parts per billion. Note: The parentheticals on the right of the panel show degrees of freedom for each data point in the panel
and average exposure concentration represented by the effect. Source: Reprinted with permission from the publisher, adapted from
Leisnerand Ainsworth (20121.
Figure 8-3 Meta-analysis of the effects of ozone exposure (relative to
charcoal-filtered air) on plant reproduction.
FLN
I
a i
(28, 56 ppb)
1
•
Y
¦
(164, 79 ppb)
HI
¦
(39. 58 ppb)
FN
1
|
(51, 85 ppb)
1
W
1
FW
(38, 98 ppb)
SF
1—#-H
(29, 74 ppb)
SN
h«H
(21,68 ppb)
WP
1—•—1
(68, 77 ppb)
sw
i
*
(158, 73 ppb)
-80 -60 -40 -20 0 20
Percentage change from ambient air (%)
FLN = flower number; FN = fruit number; FW = fruit weight; HI = harvest index; ppb = parts per billion; SF = seeds per fruiting
structure; SN = total seed number; SW = seed weight; WP = seed weight per plant; Y = yield. Note: The parentheticals on the right
of the panel show degrees of freedom for each data point in the panel and average exposure concentration represented by the
effect. Source: Reprinted with permission from the publisher, adapted from Leisner and Ainsworth (20121.
Figure 8-4 Meta-analysis of the effects of ozone exposure (relative to
ambient air) on plant reproduction.
8-45
-------
• Gillespie et al. (2015) isolated the effects of ozone on particular reproductive tissues of tomato
(Lycopersicon esculentum). Pollen grains exposed to ozone have significantly reduced
germination and pollen tube growth in vitro. Reductions in pollen viability and pollen tube
development in vivo tended to be greatest with exposure of pollen and the pollen recipient's
stigma surface. Reduction in ovule fertilization, on the other hand, seemed to occur at
approximately the same magnitude whether the pollen, stigma, or both were exposed to ozone.
Finally, when developing fruits were exposed to ozone (but the rest of the plant was in
charcoal-filtered air), fruit fresh weight and the number of seeds per fruit were reduced by 19 and
14%, respectively, relative to fruit developing in charcoal-filtered air, thus showing direct effects
of ozone on developing fruit tissue.
• The timing of ozone exposure relative to reproductive development stages can affect reproductive
outcomes in some cases. Flowers exposed to ozone early in their development tended to produce
shorter fruits than flowers exposed later in their development. On the other hand, ozone exposure
seemed to decrease the number of mature seeds per fruit by about the same amount regardless of
flower developmental stage (Black et al.. 2012).
• For noncultivated plants, authors have long hypothesized that air pollution could be a selective
force driving the evolution of plant populations over generations, with potential consequences for
community interactions (Bell et al.. 1991). One recent study evaluated traits from three lines of
the agricultural weed Spergula arvensis that were selected over generations under three different
ozone concentrations (Landesmann et al.. 2013). Selected lines appeared to vary in their seed
germination rate under a range of laboratory conditions, but differences were not as clear for seed
viability under more field-like conditions. Further evidence would be necessary to evaluate
whether ozone-driven selection resulted in measurable changes in these lines that are relevant to
field settings and interactions with other community members, including agricultural crops.
8.4.2 Plant Phenology
Several new studies of European grassland species have been published since the 2013 Ozone
ISA that measure the effects of ozone on flowering phenology (e.g., time of first flowering, time of
maximum flowering).
• No effect on timing of flowering was recorded for Leontodon hispidus, Dactylis glomerate, or
Anthoxanthum odoratum over a range of ozone seasonal 24-hour means (21-103 ppb), either
during greenhouse exposure or 2 months after exposure (Haves et al.. 2011). Similarly, no effects
of two consecutive seasons of a range of ozone seasonal 24-hour means (19-73 ppb) were
observed on timing of flowering in Briza media, Sanguisorba minor, or Scabiosa columbaria
(Haves et al.. 2012b).
• There was no effect of ozone on flowering in Briza maxima during an open-top chamber
experiment, but 30 days after exposure the number of plants starting to flower was 66% lower in
the 45-65 ppb mean ozone treatment group compared with the charcoal-filtered air treatment
group (Sanz et al.. 2011).
• Flowering sped up for Lotus corniculatus under increasing ozone, especially for well-watered
plants (r2 = 0.64 mean ozone vs. date to 20% max flower number). The date of maximum
flowering was also earlier: an increase in the mean ozone concentration from 30 to 70 ppb
corresponded with maximum flowering occurring 6 days earlier in both well-watered and drought
conditions (Haves et al.. 2012b).
8-46
-------
The effect of ozone on the timing of seed germination has also been recently studied.
Germination is delayed in some species, sped up in others, or remains unaffected by ozone exposure
(Abeli et al.. 2017; Black et al.. 2012). Leaf senescence can be considered a phenological event, although
for the purposes of this review, it was considered a visible foliar injury (see Section 8.2).
8.4.3 Plant Mortality
Several new studies examined the impact of ozone exposure on plant mortality (i.e., the fraction
of individuals in population that die over a given interval). All were focused on tree species, and the study
results are consistent with previous evidence showing that ozone can affect tree mortality (Table 8-8).
• In the Aspen FACE experiment in Rhinelander, WI, the survival of sensitive aspen (Populus
tremuloides) genotypes 271 and 259 declined significantly between 1997 and 2008 under
elevated ozone concentrations relative to ambient conditions (Moran and Kubiske. 2013). In
contrast, the survival of tolerant genotype 8L increased 14.8% under elevated ozone. Genetically
based differences in ozone sensitivity could have implications for intraspecific diversity and
evolution of wild populations (see Section 8.4.1).
• Dietze and Moorcroft (2011) conducted a large-scale analysis of factors contributing to annual
mortality of trees in functional types (each characterized by different species) in the forests of the
eastern and central U.S. The U.S. Forest Service's FIA database (http://www.fia.fs.fed.us/.
version 2.1) was queried for data on tree mortality, and the analysis only included trees that were
measured in consecutive censuses. Overall, ozone was ranked 9th on the list of 13 factors that
forests were sensitive to, and ozone's effect was similar in magnitude to that of precipitation.
Mortality in eight out of ten plant functional types was significantly correlated with ozone 8-hour
max: seven experienced increasing mortality with increasing ozone exposure, while late
successional conifers showed a slight decrease in mortality with increasing ozone exposure.
Evergreen hardwoods were the functional type most sensitive to increasing ozone; they showed
annual mortality ranging from 1% in areas of the country with relatively low ozone to 3% in areas
of relatively high ozone. Assuming no replacement, a change in mortality rate from 1 to 3%
would shift the time to 50% loss of a species from 69 to 24 years. Such changes in mortality are
consistent with documented changes in community composition (Section 8.10).
8.4.4 Summary and Causality Determinations
Ozone exposure can affect plant reproduction. Over 100 studies of cultivated and noncultivated
species have now been synthesized qualitatively and quantitatively. They show that diverse metrics of
plant reproduction decline under ozone concentrations occurring either in the environment or under
experimental conditions within an order of magnitude of recent concentrations. The biological
mechanisms underlying ozone's effect on plant reproduction are twofold. They include both direct
negative effects on reproductive tissues and indirect negative effects that result from decreased
photosynthesis and other whole-plant physiological changes. Two metrics of plant reproduction, fruit
number and fruit weight, show greater reductions under increased ozone when combined across species
8-47
-------
for ozone concentrations that span 40 to >100 ppb; other metrics do not show such reductions or do so
across a narrower range of ozone concentrations. An exposure-response and a dose-response curve
developed for legume species in Europe both show significant negative effects of ozone on plant
reproductive biomass. Finally, experimental ozone exposure at multiple experimental settings
(e.g., in vitro, whole plants in the laboratory, whole plants and/or reproductive structures in the
greenhouse, and whole plant communities in field settings) convincingly show ozone independently
reduces plant reproduction. Therefore, previous evidence and new evidence reviewed here is
sufficient to infer a causal relationship between ozone exposure and reduced plant reproduction
(Table 8-7).
Studies of tree mortality indicate that ozone affects this endpoint. Multiple studies from different
research groups show the co-occurrence of ozone exposure and increased mortality of trees. Evidence for
plants other than trees is currently lacking. Studies linking ozone and tree mortality are consistent with
known and well-established individual plant-level mechanisms that explain ozone phytotoxicity,
including variation in sensitivity and tolerance based on age class, genotype, and species. Increased
mortality is also consistent with effects at higher levels of biological organization, including changes in
vegetation cover and decreased terrestrial biodiversity (Section 8.10). Relationships between ozone and
mortality have been modeled for eastern and central U.S. forests; 7 out of 10 plant functional types show
a significant positive correlation between 8-hour max ozone concentration and tree mortality.
Experimentally, elevated ozone exposure has been shown to increase mortality in sensitive aspen
genotypes. Therefore, previous evidence and new evidence reviewed here is sufficient to infer a
likely to be causal relationship between ozone exposure and tree mortality (Table 8-8).
8-48
-------
Table 8-7 Summary of evidence for causal relationship between ozone exposure and plant reproduction.
Rationale for Causality
Determination
Key Evidence
Key References
Consistent evidence from multiple
research groups under ozone
concentrations occurring in the
environment or experimental ozone
concentrations within an order of
magnitude of recent concentrations
Quantitative meta-analysis of >100 independent
studies using different experimental approaches
show statistically significant and sometime large
(up to 50%) decreases in reproduction across
crop and noncrop species with exposure to
<40 ppb ozone.
Independent studies synthesized using
qualitative review have also shown consistent
reduction in most measures of reproduction.
Leisner and Ainsworth (2012),
U.S. EPA (2006). Sections AX9.5.4.1,
AX9.6.2.4, AX9.6.2.5, AX9.6.4.2;
U.S. EPA (2013). Section 9.4.3.3
AX9.5.4.4, AX9.5.5.1,
Biologically plausible mechanisms are
well established and show evidence for
direct and indirect effects on
reproductive tissues and function
Experimental exposure of whole plants and
reproductive tissues in isolation demonstrate
that the effect of ozone on plant reproduction
can be indirect (via decreased available
photosynthate) or direct (via changes in male or
female function).
Gillespie et al. (2015),
U.S. EPA (2006). Sections AX9.2, AX9.6.4.2;
U.S. EPA (2013). Sections 9.3, 9.4.3.3
Exposure-response relationship is clear Fruit number and fruit weight show a clear
Leisner and Ainsworth (2012),
for some metrics of reproduction and
not well resolved for others; exposure-
response and dose-response curve
exists for a set of legume species
exposure-response relationship with exposure to Sanz et al. (2016)
ozone at a range of concentrations from 40 ppb
to >100 ppb. Other measures show an
exposure-response relationship over a narrower
range of concentrations.
Exposure-response and dose-response curves
for Trifolium spp. show a significant negative
relationship between ozone and reproductive
biomass.
Abundant experimental evidence
isolates and demonstrates the effect of
ozone on plant reproduction
Studies that compare plants grown in
charcoal-filtered air or ambient air as a control
with plants experimentally exposed to ozone
demonstrate that exposure to ozone causes
reduction in reproductive output.
Leisner and Ainsworth (2012),
U.S. EPA (2006). Sections AX9.5.4.1. AX9.5.4.4. AX9.5.5.1,
AX9.6.2.5, AX9.6.4.2;
U.S. EPA (2013). Section 9.4.3.3
8-49
-------
Table 8-8 Summary of evidence for likely to be causal relationship between ozone exposure and tree mortality.
Rationale for Causality
Determination
Key Evidence
Key References
Consistent evidence from multiple
research groups under ozone
concentrations occurring in the
environment or experimental ozone
concentrations within an order of
magnitude of recent concentrations
Independent studies show co-occurrence of
increasing mortality and exposure to ozone in
tree species from different forest types in the
U.S. and in specific sensitive tree species in
Mexico and Europe.
Dietze and Moorcroft (2011),
Diaz-De-Quiiano et al. (2016),
U.S. EPA (2006). Sections AX9.6.2.1,
U.S. EPA (2013). Section 9.4.7.1
AX9.6.2.2, AX9.6.2.3;
Biologically plausible mechanisms are
well established and support observed
effects at higher levels of biological
organization
Differences in mortality are consistent with
known physiological mechanisms of ozone
sensitivity and tolerance in age classes,
genotypes, and species.
Mortality due to ozone exposure is also
consistent with observed changes in vegetation
cover and terrestrial community composition.
Moran and Kubiske (2013),
U.S. EPA (2006). Sections AX9.2 AX9.6.2.2, AX9.6.2.3, AX9.6.4.1;
U.S. EPA (2013). Section 9.4.7.1
Independent effect of ozone modeled
in one large-scale study but
confounded in most observational
studies
One empirical model of eastern and central U.S.
forests shows a significant effect of ozone
independent of other factors, most gradient
studies cannot rule out other factors that
contribute to mortality.
Dietze and Moorcroft (2011).
U.S. EPA (2006). Sections AX9.6.2.1. AX9.6.2.2. AX9.6.2.3;
U.S. EPA (2013). Section 9.4.7.1
Limited evidence from experimental The Aspen FACE study shows sensitive Moran and Kubiske (2013).
studies that isolate the effect of ozone genotypes have increased mortality with ozone U.S. EPA (2013). Section 9.4.7.1
on tree mortality exposure compared with a tolerant genotype.
This table is provided as an example of a causal table using the modified Hill criteria (U.S. EPA. 2015). Tables were only used if there was a change in causality from the 2013
Ozone ISA or if a new causality determination was warranted based on evaluation of the evidence.
8-50
-------
Table 8-9 Ozone exposure and plant reproduction, phenology, and mortality.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Plant Reproduction and Mortality
Burkev et al. (2012)
FACE; Champaign, IL
(40.033°N, 88.233°W)
Phaseolus vulgaris
(snap bean) three
genotypes
(S156-03 sensitive;
R123, R331-Os
tolerant)
O3 exposure from May-August
2006; exposure hours uncertain,
perhaps 9:00 a.m. to 5:00 p.m.
Ambient (control)—8-h
mean = 43 ppb, 1-h
max = 29-78 ppb,
AOT40 = 5.3 ppm-h,
SUM60 = 5.3 ppm-h
+O3—8-h
mean = 59 ppb, 1-h
max = 32-114 ppb,
AOT40 = 16.3 ppm-h,
SUM60 = 27 ppm-h
+O3+CO2—8-h
mean = 59 ppb, 1-h
max = 33-112 ppb,
AOT40 = 16.2 ppm-h,
SUM60 = 26.7 ppm-h
Plant reproduction—sensitive genotype had 63%
decline in pod weight per plant and similar result for
seed weight per plant under elevated O3; no significant
difference for tolerant genotypes under elevated O3.
Sensitive genotype had 57% reduction in seed number
per plant and 19% reduction in weight per seed under
elevated O3.
Taia et al. (2013)
OTC; Al Montazah
botanical garden near
Alexandria, Egypt
Capsicum annuum
(pepper)
Three chambers exposed to
ambient air or charcoal-filtered air
8 h/day 9:00 a.m.-5:00 p.m. for
75 days.
Ambient: 8-h seasonal daily
average = 78 (±8) ppb, AOT40
29,600 (±42);
charcoal-filtered air: 15 (±3) ppb,
AOT40 0
Plant reproduction—fruit length, fruit weight, and
number of fruits per plant were all significantly lower in
ambient chambers. Number of fruits reduced by 28%.
Percentage pollen germination and pollen tube length
were also lower in ambient chambers.
8-51
-------
Table 8-9 (Continued): Ozone exposure and plant reproduction, phenology, and mortality.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Plant Reproduction and Mortality
Landesmann et al.
(2013)
Other; University of
Buenos Aires,
Argentina (34.58°S,
58.58°W)
Spergula arvensis
(annual agricultural
weed corn spurry);
populations grown
under O3 (0, 90,
120 ppb) for four
generations in
Corvallis, OR,
before being sent
to Argentina and
grown for a
generation in the
field
(1) populations exposed to O3 of 0,
90, or 120 ppb over four
generations before the beginning of
this study, and (2) maternal plants
exposed to O3 only for reproductive
stage: ambient range 0-20 ppb,
and elevated range 40-70 ppb O3
Plant reproduction—generational O3 exposure affects
germination under hot and wet conditions, with the
highest germination rate in the 90-ppb population and
the lowest germination in the 120-ppb population
(p = 0.022). Generational O3 exposure affects
germination under cold and dry conditions, with the
highest germination rate in the 90-ppb population, and
the lowest germination rate in the 0-ppb population
(p = 0.16). Forthe 120-ppb population, germination
rates were highest under cold and dry conditions.
Exposure of mother plants to O3 (40-70 ppb) as seeds
were developing resulted in higher seed viability than
in plants under ambient O3 conditions.
Li et al. (2013a)
OTC; wheat fields in
northern China
Agricultural weed
Descurainia sophia
(flixweed) grown
alone or in
competition with
Triticum aestivum
(winter wheat)
Three O3 treatments: ambient (less
than 40 ppb O3), elevated
(80 ± 5 ppb for 7 h/day for
30 days), highly elevated
(120 ±10 ppb for 7 h/day for
30 days)
Plant reproduction—elevated O3 had no statistically
significant effect on flixweed seed production, while
highly elevated O3 decreased flixweed seed production
24-29%.
8-52
-------
Table 8-9 (Continued): Ozone exposure and plant reproduction, phenology, and mortality.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Plant Reproduction and Mortality
Gillespie et al. Greenhouse; Close Lycopersicon Pollen and ovule experiments:
(2015) House, esculentum control = charcoal-filtered air
Northumberland, U.K. (tomato) (<5 nmol/mol O3),
treatment = CFA + 75 nmol/mol O3
for 7 h/day, also combinations of
these exposures (e.g., plant grown
in CFA, then pollen exposed to
75 nmol/mol in separate chamber).
Fruit experiment:
control = charcoal-filtered air
(<5 nmol/mol O3),
treatment = 100 nmol/mol O3 for
8 h/day
Plant reproduction—pollen germination (on
agar)—significant reduction in O3/O3 VS.O3/CFA and
CFA/O3 vs. CFA/CFA which indicates direct effect of
O3 on pollen; pollen tube growth (on
agar)—significantly reduced in O3/O3 vs. CFA/CFA and
CFA/O3 vs. CFA/CFA, which also suggests direct
effect on pollen; pollen viability index (germinated vs.
nongerminated pollen on stigma surface)—pollen from
O3 exposed plants used to pollinate an O3 exposed
stigma has 25% lower germination than other
treatments; pollen tube development index (pollen
tubes at base of the papillia vs. germinated pollen on
stigman surface)—pollen from O3 exposed plants used
to pollinate an O3 exposed stigma had lower pollen
tube development than other treatments; ovule
fertilization—percentage fertilized, viable ovules was
reduced by 26% in O3/O3 crosses vs. CFA/CFA
crosses; percentage nonviable, fertilized ovules, and
nonfertilized ovules was lowest in CFA/CFA crosses;
fruit fresh weight—19% lower in O3; fruit dry
weight—18% lower in O3; number of seeds per
fruit—14% lower in O3 (all significant reductions).
Tetteh et al. (2015)
OTC; Fuchu, Tokyo,
Japan
Vigna unguiculata
(cowpea); two
cultivars Blackeye
and Asontem
88 days of exposure, 5 h O3
addition per day
(11:00 a.m.-4:00 p.m.);
Filtered: 24-h avg = 13 ppb (daily
min/max = 1-55),
AOTO = 9.2 ppm-h, AOT40 = 0.2;
Ambient: 24-h avg = 26 ppm (daily
min/max = 1-110),
AOTO = 18.6 ppm-h, AOT40 = 2.7;
Ambient +O3: 24-h avg = 39 ppb
(daily min/max = 1-175),
AOTO = 27.3 ppm-h, AOT40 = 10.4
Plant reproduction—pod length per plant: significant
effect of O3 exposure (filtered = 14.05 cm,
ambient = 13.8, amb + 03 = 11.9). Number seeds per
pod: significant effect of O3 exposure
(filtered = 10 seeds/pod, ambient = 9, amb + 03 = 7.5).
Number pods per plant: no effect of O3 exposure.
8-53
-------
Table 8-9 (Continued): Ozone exposure and plant reproduction, phenology, and mortality.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Plant Reproduction and Mortality
Black et al. (2012)
Lab; unable to tell Brassica
where growth campestris (field
chambers were located mustard)
(potentially U.K.)
Control = charcoal-filtered air, O3
below detection limit.
Treatment = 100.4 ±1.16 ppb O3
for 6 h (10:00 a.m.-4:00 p.m.) on
4 consecutive days between
17-20 days after sowing
Plant reproduction—number reproductive sites—ozone
had no significant effect on the number of reproductive
sites or in the relative proportion of sites that were
aborted vs. produced a pod. Exception was
reproductive sites exposed as buds, where ozone
exposure increased development into pods from 10.7
to 15.7%; pod length decreased significantly for those
from flowers that bloomed on Days 3, 4, or that
remained as buds during ozone exposure (down by 16,
29, 25%, respectively); pod weight (minus seeds) not
significantly affected, except for an increase in those
from flowers that bloomed Day 2 of exposure; pod
number not affected by ozone exposure; number of
seeds per pod, number of seeds per plant decreased
by 33% in ozone-exposed plants; number of aborted
seeds, prematurely germinated seeds significantly
higher (129, 851%) in ozone-exposed plants; individual
seed weight 11% higher in ozone-exposed plants;
seed weight per plant 23% lower in ozone-exposed
plants. Plant phenology: seeds exposed to ozone
germinated more quickly but at the same final
percentage as control seeds.
Dietze and
Moorcroft (2011)
Gradient; eastern and
central U.S., bounded
to west by 98°W
longitude
10 plant functional
types each
characterized by
different species
Range of 0.050-0.114 ppm 8-h
max
Plant mortality—8 of 10 plant functional types had a
significant correlation with O3; evergreen hardwoods
plant functional type is most sensitive to O3 increase;
overall, eastern and central forests are 9th most
sensitive (in terms of tree mortality) to O3 (out of 13
variables).
Moran and Kubiske
(2013)
FACE; Aspen FACE,
near Rhinelander, Wl
(45.7°N, 89.5°W)
Clones of five
genotypes of
Populus
tremuloides from
the aspen-only
sections of the
experiment,
1997-2008
Full factorial: O3 and CO2,
1998-2008. Ozone: ambient
(W126 2.1-8.8 ppm-h) or elevated
(W126 12.7-35.1 ppm-h). CO2:
ambient (360 ppm) or elevated
(560 ppm); for hourly ozone
concentrations during experimental
ozone treatment, see Kubiske and
Foss (2015)
Plant mortality—survival of two genotypes (271 and
259, which had respectively the highest and lowest
survival under ambient conditions) declined
significantly under elevated O3. Survival of genotype
8L increased 14.8% under elevated O3.
8-54
-------
Table 8-9 (Continued): Ozone exposure and plant reproduction, phenology, and mortality.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Plant Reproduction and Mortality
Diaz-De-Quiiano et
al. (2016)
Gradient; central
Catalan Pyrenees
northeastern Spain
Pinus uncinata Along an altitudinal gradient with
(mountain pine) an annual avg of 35 ppb at 1,040 m
stands asl to 56 ppb at 2,300 m asl (2004
to 2007 but reached 38 and 74 ppb
from April to September)
Plant mortality—along two transects (Guils, Meranges)
in the Pyrenees, mortality of P. unicata was positively
correlated with accumulated O3 exposure over a period
of 5 yr. Tree mortality increased with altitude (1 to 30%
on Guils transect, 1 to 7.5% on Meranges transect).
Explanatory variables for these observations included
mean fortnightly O3 levels, water availability, and stand
characteristics. Higher percentage tree mortality was
observed above a threshold of sum of fortnightly O3
levels of 2,900 ppb. Authors note that O3 exposure not
established as the main cause of tree mortality due to
other environmental and anthropogenic stressors.
Haves et al. (2012b) Mesocosm; U.K.
Briza media,
Festuca ovina (data
not collected),
Campanula
rotundifolia,
Sanguisorba minor,
Scabiosa
columbaria,
Helianthemum
nummularium,
Lotus corniculatus
Eight treatments varying in
seasonal 24-h mean ozone, but
with the same weekly profile
(4 days of +10 to +25 ppb peaks
followed by 3 days of +5 ppb
peaks); exposure of same plants in
two seasons (2009 and 2010) over
12 weeks each season, but
flowering measurements taken
weekly only in Season 2; 2010
seasonal 24-h mean ozone levels
were: 19.0, 25.5, 34.8, 40.8, 51.2,
60.3, 66.2, 73.3 ppb
Plant phenology—no effect of O3 on timing of flowering
for B. media, S. minor, or S. columbaria. Flowering
sped up for L. corniculatus under increasing O3 levels,
especially for the well-watered plants—r2 = 0.64 mean
ozone vs. date to 20% max flower number (drought
somewhat dampened the effect). Date of maximum
flowering was also earlier—an increase in the mean
ozone concentration from 30 to 70 ppb corresponded
with maximum flowering occurring 6 days earlier in
both the well-watered and drought treatments. Plant
reproduction: No effect of O3 on maximum number of
flowers for L. corniculatus, B media, or S. minor, C.
rotundifolia showed a significantly lower number of
maximum flowers under higher O3 (but the range was
small 2-10 flowers); S. columbaria also showed a main
effect of O3 with lower maximum bud number under
high O3, but the range was small (4-7 buds).
8-55
-------
Table 8-9 (Continued): Ozone exposure and plant reproduction, phenology, and mortality.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Plant Reproduction and Mortality
Haves et al. (2011)
Greenhouse; near
Marchlyn Mawr, U.K.
Two communities:
four plants of forb
Leontodon hispidus
and three plants of
grass Dactylis
glomerata', four
plants of forb
Leontodon hispidus
and three plants of
Anthoxanthum
odoratum
Eight treatments: (1) Seasonal
24-h mean = 21.4 ppb (12-h
mean = 21.1 ppb, daylight
[7:00 a.m.-6:00 p.m.]
AOT40 = 0.07 ppm-h, 24-h
AOT40 = 0.07 ppm-h);
(2) Seasonal mean = 39.9 ppb
(12-h = 39.2 ppb, daylight
AOT40 = 4.93 ppm-h, 24-h
AOT40 = 10.91 ppm-h);
(3) Seasonal mean = 50.2 ppb
(12-h = 49.6 ppb, daylight
AOT40 = 21.44 ppm-h, 24-h
AOT40 = 41.29 ppm-h);
(4) Seasonal mean = 59.4 ppb
(12-h = 58.7 ppb, daylight
AOT40 = 38.04 ppm-h, 24-h
AOT40 = 72.19 ppm-h);
(5) Seasonal mean = 74.9 ppb
(12-h = 73.3 ppb, daylight
AOT40 = 62.49 ppm-h, 24-h
AOT40 = 119.82 ppm-h);
(6) Seasonal mean = 83.3 ppb
(12-h = 81.6 ppb, daylight
AOT40 = 77.13 ppm-h, 24-h
AOT40 = 147.42 ppm-h);
(7) Seasonal mean = 101.3 ppb
(12-h = 99.0 ppb, daylight
AOT40 = 108.43 ppm-h, 24-h
AOT40 = 206.70 ppm-h);
(8) Seasonal mean = 102.5 ppb
(12-h = 100.5, daylight
AOT40 = 112.47 ppm-h, 24-h
AOT40 = 214.34 ppm-h)
Plant phenology—there was no effect on timing of
flowering or number of flowers during O3 exposure or
2 mo after O3 exposure. 2 mo after O3 exposure
ended, the proportion of living mature leaves on L.
hispidus increased linearly with seasonal mean O3
concentration. In the next growing season, there was
no effect of the previous season's O3 exposure on the
number of flowers or seeds for L. hispidus or D.
glomerata. The ratio of L. hispidus flowers to
seed-heads in the second season decreased linearly
with increasing first season mean O3 concentration.
8-56
-------
Table 8-9 (Continued): Ozone exposure and plant reproduction, phenology, and mortality.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Plant Reproduction and Mortality
Sanz etal. (2016)
OTC; experimental
field located in a rural
area in the
northeastern Iberian
Peninsula, Tarragona
(40.41°N, 0.47°E)
Dehasa-type
pasture species,
Leguminoseae
(three species)
Data analyzed from independent
experiments, 45-day avg O3
exposure length
Plant reproduction—an O3 critical level for reproductive
capacity AOT40 = 2.0 (1.5, 2.8) ppm-h and Phytotoxic
Ozone Dose (POD)1 =7.2(1.1, 13.3) mmol/m2was
developed from linear exposure response functions
based on seed and flower production. Reproductive
capacity had the lowest critical level of the endpoints
evaluated.
Gundel et al. (2015) OTC; Buenos Aires,
Argentina
Lolium multiflorum
Low ozone = <10 ppb; High
ozone = -120 ppm for 2 h/day
noon-2:00 p.m. for 5 consecutive
days preanthesis
Plant reproduction—trend towards O3 increasing seed
viability under high temperature and humidity was not
significant.
Sanz et al. (2011) OTC; Mediterranean Briza maxima
coast, Spain, (40.68°N,
0.78°E)
O3 as AOT40 index—ozone:
charcoal-filtered (mean
O3 < 10 ppb, AOT40 = 0); Ambient
(mean O3 < 40 ppb,
AOT40 = 1,367 ppb-h); Addition of
40 ppb O3 from 7:00 a.m. to
5:00 p.m. for 5 days/week (mean
O3 = 40-65 ppb,
AOT40 = 10,841 ppb-h) NH4NO3
addition to mimic 10, 30, or 60 kg
N/ha
Plant phenology—while O3 exposure ran, there was no
significant effect of O3 on phenology. 30 days after O3
exposure ended, phenology differed by O3 treatment:
the number of plants starting to flower were 66% lower
in added O3 than in charcoal-filtered air, as the number
of plants with mature seeds were 280-340% higher in
the ambient and added O3 treatments respectively,
than in the filtered treatment. Nitrogen had no effects
on phenology.
8-57
-------
Table 8-9 (Continued): Ozone exposure and plant reproduction, phenology, and mortality.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Plant Reproduction and Mortality
Leisner and
Ainsworth (2012)
Other; all locations
included (but no
summary of geographic
distribution of studies
included)
All species included
(data set includes
monocots and
dicots, perennials
and annuals,
determinate and
indeterminate
growth habits, C3
and C4 plants)
Grouped and analyzed exposures
in several ways: (1) Charcoal-
filtered air vs. elevated
O3—elevated O3 in multiple
categories (4-, 7-, 8-, or 12-h
daytime means depending on data
reported): >100 ppb, 70-100 ppb,
40-70 ppb, <40 ppb; (2) Ambient
air vs. elevated 63—elevated O3 in
multiple categories (4-, 7-, 8-, or
12- h daytime means depending on
data reported): >100 ppb,
70-100 ppb, 40-70 ppb, <40 ppb
Plant reproduction—(1) compared with
charcoal-filtered air, most measurements of plant
reproduction were reduced under elevated O3 (but not
flower measurements); for example, seed number, fruit
number, and yield decreased by 16, 9, and 19%,
respectively, all at slightly different average exposure
levels. Some endpoints significantly decrease even
under the lowest exposure category (<40 ppm). Some
trends in larger negative responses under higher
exposure levels, but overall there was no clear
exposure-response across experiments and species.
Yield was not significantly affected below 70 ppb, but it
decreased 45% at highest exposure level.
(2) Compared with ambient air, all measurements of
plant reproduction were reduced under elevated
O3—for example, yield, fruit weight, and seed number
decreased by 25, 51% (the largest effect observed),
and -10%, respectively. Effects occurred even at the
lowest exposure level (<40 ppb). There was a clear
exposure-response with respect to yield and a clear
trend for fruit number and fruit weight. Yield was down
52% at the highest exposure level. The response to O3
by different types of plants was not significantly
different in many cases. One exception was that
indeterminate plants had much greater reductions in
fruit weight and fruit number than did determinate
plants.
Wang et al. (2015) Global meta-analysis
98 herbaceous
species tested for
CO2 effects on
reproductive
allocation in papers
previously
published in
1977-2011
Not specified
Without specific stressors, there is no effect of
elevated CO2 on plant allocation to reproductive
biomass (n = 508). With ozone exposure, plant
allocation to reproduction is 4% lower at elevated CO2
(+CO2+O3) than at ambient CO2 (+03).
8-58
-------
Table 8-9 (Continued): Ozone exposure and plant reproduction, phenology, and mortality.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Plant Reproduction and Mortality
Ferreira et al. (2016) Lab; Porto, Portugal
Plantago
lanceolata, Salix
atrocinerea
All exposures for 6 h;
Plantago—control = 0 ppm;
Treatment level + (approx. equal to
European standard for
health) = 0.065 ppm; Treatment
level ++ = 0.124 ppm;
Salix—control = 0 ppm; treatment
level + (approx. equal to European
standard for health) = 0.061 ppm;
treatment level ++ = 0.118 ppm
Plant reproduction—Plantago and Salix pollen viability
significantly declined with increasing O3 exposure
(e.g., -56 and 23% lower, respectively, in ++O3
treatment compared to control); Pollen germination
was only significantly reduced at ++O3 treatment; Salix
pollen germination was significantly reduced at + and
++O3 treatment, but O3 treatments did not differ.
Abeli et al. (2017)
Lab; Alpine seeds
collected on Mt.
Cimone, Mt.
Prado-Cusna and in
the Dolomites in Italy;
O3 exposure inside
incubators
Achillea clavennae,
Aster alpinus,
Festuca rubra
subsp. commutata,
Festuca violacea
subsp. puccinellii,
Plantago alpina,
Silene acaulis,
Silene nutans,
Silene suecica,
Vaccinium myrtillus
Control: Ambient air (0-1 ppb)
"125_5" treatment: 125 ppb O3
24 h/day for 5 days; "125_10"
treatment: 125 ppb O3 24 h/day for
10 days; "185_5" treatment:
185 ppb O3 24 h/day for 5 days
Plant reproduction—significant differences in seed
mortality for some species between all four
germination conditions, but not in a consistent way.
Combining all species, each treatment (compared with
control) significantly delayed germination
(125_5 = 0.71, 185_5 = 0.87, 125_10 = 1.17 day
delay). Six of nine individual species had reduction in
germination percentage for one or more of O3
treatment at the end of O3 exposure. Seven of nine
species showed a significant effect of at least one O3
treatment at 28 days after sowing, and effects ranged
from increasing to decreasing germination percentage.
Plant phenology—combining all species, 125_5 and
185_5 treatments did not affect mean germination time
either at end of O3 exposure or at end of the
experiment. The 125_10 treatment significantly
increased mean germination time by 1.25 days after O3
exposure, but by the end of the experiment that
difference did not exist. Individual species responded
in different ways to treatments.
8-59
-------
Table 8-9 (Continued): Ozone exposure and plant reproduction, phenology, and mortality.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Plant Reproduction and Mortality
Yang et al. (2017)
OTC; Zhangtou,
Changping District,
Beijing, China
(40.20°N, 116.13°E)
Tagetes erecta
(marigold) and four
varieties of Petunia
hybrid with pink,
red, rose-red, or
white flowers
Ozone exposure during growth
period of marigold: Ambient avg
37.1 ppb Os (AOT40 = 1.6 ppm-h);
elevated avg 99.2 ppb
(AOT40 = 20.4 ppm-h); highly
elevated avg 145.2 ppb
(AOT40 = 36.4 ppm-h); Ozone
exposure during growth period of
petunia: Ambient avg 40.0 ppb O3
(AOT40 = 4.0 ppm-h); elevated avg
96.0 ppb (AOT40 = 25.0 ppm-h);
highly elevated avg 153.3 ppb
(AOT40 = 47.6 ppm-h)
Plant reproduction—elevated O3 (96.0 ppb) reduced
flower diameter 7% and flower biomass 44% in white
petunias, and reduced flower biomass 25% for pink
petunias. Highly elevated O3 (153.3 ppb) reduced
flower diameter 7% in white petunias, 11% in rose
petunias, 9% in red petunias, and 12% in pink
petunias, and reduced floral biomass across all petunia
varieties 20-40%. There were no effects of O3 on
marigold flower biomass or flower diameter.
AOTO = seasonal sum of the difference between an hourly concentration at the threshold value of 0 ppb, minus the threshold value of 0 ppb; AOT40 = seasonal sum of the difference
between an hourly concentration at the threshold value of 40 ppb, minus the threshold value of 40 ppb; asl = above sea level; C3 = plants that use only the Calvin cycle for fixing the
carbon dioxide from the air; C4 = plants that use the Hatch-Slack cycle for fixing the carbon dioxide from the air; C02 = carbon dioxide; FACE = free-air C02 enrichment; kg
N/ha = kilograms of nitrogen/hectare; n = sample size; NH4NO3 = ammonium/nitrate solution; nmol/m2 = nanomole per meter squared; nmol/mol = nanomoles per mole; 03 = ozone;
OTC = open-top chamber; ppb = parts per billion; ppm = parts per million; SUM60 = sum of hourly ozone concentrations equal to or greater than 60 ppb; W126 = cumulative
integrated exposure index with a sigmoidal weighting function.
8-60
-------
8.5
Reduced Crop Yield and Quality
In the 2013 Ozone ISA, the evidence was sufficient to conclude there is a causal relationship
between ozone exposure and reduced yield and quality of agricultural crops (U.S. EPA. 2013). The
detrimental effect of ozone on crop production has been recognized since the 1960s, and a large body of
research has subsequently characterized decreases in yield and quality of a variety of agricultural and
forage crops. Ozone effects on cellular processes, plant metabolism, altered C allocation to vegetation and
roots, and leaf-level physiology (Section 8.1.3 and Figure 8-2) underlie agricultural crop damage, which
is measured as reduced crop yield and quality. The actual concentration and duration threshold for ozone
damage varies from species to species, and sometimes even among genotypes of the same species
[Section 9.4.4.1 in U.S. EPA (2013)1. Numerous experimental analyses have also demonstrated that the
effects of ozone exposure vary depending on the growth stage of the plant. In studies reviewed in the
2013 Ozone ISA, increasing ozone concentration decreased nutritive quality of grasses, decreased
macronutrient and micronutrient concentrations in fruits and vegetable crops, and decreased cotton fiber
quality [Section 9.4.4.2 in U.S. EPA (2013)1.
As described in the PECOS tool (Table 8-2). the scope for new evidence reviewed in this section
limits studies to those conducted in North America at ozone concentrations occurring in the environment
or experimental ozone concentrations within an order of magnitude of recent concentrations (as described
in Appendix 1). If data from other countries were included in meta-analyses with U.S. data (or
incorporated into exposure-response functions for crops), these studies were also considered.
8.5.1 Field Studies and Meta-Analyses
Greenhouse, OTC, FACE, field, and gradient studies reviewed in the 2013 Ozone ISA clearly
show negative impacts of ozone on crop yield at concentrations relevant to the then current conditions
(U.S. EPA. 2013). In the 2013 Ozone ISA, the results of several meta analytic studies for soybean
(Betzelberger et al.. 2010; Morgan et al.. 2006; Morgan et al.. 2003). wheat (Feng et al.. 2008). rice
(Ainsworth. 2008). and potato, bean, and barley (Feng and Kobavashi. 2009) provided evidence that
current levels of ozone decrease crop growth and yield. New field studies and meta-analyses of U.S. crop
data and global analyses that include U.S. crop data further characterize effects on crop species, improve
estimates of yield loss, and refine concentration response (Table 8-10).
• Ozone's effects on reproductive and developmental stages of the plant life cycle in a variety of
crop and noncrop species were evaluated in a meta-analysis by Leisner and Ainsworth (2012).
Grain or seed yield per unit area was decreased by 19% at an average ozone concentration of
60 ppb in ambient air compared with charcoal-filtered air (Figure 8-3). Compared with ambient
air, yield decreased by 25%. Seed and fruit number were also frequently affected by elevated
8-61
-------
ozone levels (Figure 8-4). Additional reproductive and developmental traits in the meta-analysis
affected by ozone exposure discussed in Section 8.4 and Table 8-9 have relevance for crop yield.
• For soybean, additional studies at SoyFACE in Illinois report decreased seed/crop yield (Leisner
et al.. 2017; Ainsworth et al.. 2014; Vanloocke et al.. 2012) as well as timing of ozone damage
and canopy senescence (Sun et al.. 2014). Betzelberger et al. (2012) refined an exposure-response
curve for soybean that previously relied on single concentrations over multiple years. A linear
decrease in yield was observed across two growing seasons at SoyFACE at the rate of 37 to 39 kg
per hectare per ppb cumulative ozone exposure over 40 ppb (AOT40); summed from the
beginning of the growing season up to the date of measurement. All seven cultivars showed
similar responses to ozone. Osborne et al. (2016) updated the exposure-response function for
soybean by pooling relative yield data from 28 experimental studies after 1998 from the U.S.,
China, and India. This analysis identified a critical level of 32.3 ppb (7-hour seasonal mean) at
which a statistically significant (5%) loss of yield can occur. Soybean cultivars varied in
sensitivity to ozone with a yield loss at a 7-hour mean concentration of 55 ppb (used in the study
to represent present day background levels) ranging from 13.3 to 37.9%.
• For wheat, meta-analyses using data from the U.S. and other countries provide further supporting
evidence that current levels of ambient ozone decrease growth, quality, and yield (Pleiiel et al..
2018; Broberg et al.. 2015). A meta-analysis of ozone's effects on wheat grain quality based on
42 studies (OTC and FACE conducted in the U.S. Europe and Asia) indicated that ozone
significantly and strongly reduced 1,000-grain weight. Volume weight and starch content were
also significantly lower with higher ozone exposure (Broberg et al.. 2015). Ozone OTC
experiments with field-grown wheat from 33 studies in 9 countries, including the U.S., showed an
average wheat yield loss per ppb ozone of 0.38% (Pleiiel et al.. 2018). Grain yield, grain mass,
total aboveground biomass, starch concentration, starch yield, and protein yield were significantly
decreased in nonfiltered air compared with charcoal-filtered air, with starch yield being the most
strongly affected.
• New studies in nonsoybean legumes include evaluation of biomass and seed yield in
ozone-exposed snap bean (P. vulgaris) under high and low vapor pressure deficit conditions
(Fiscus et al.. 2012). In elevated ozone treatments at high humidity (low vapor pressure
deficit = VPD), snap bean yield was decreased by 55-72% with no significant yield loss at high
VPD (Fiscus et al.. 2012). Both mass per seed and number of seeds per plant were reduced. Llovd
et al. (2018) assessed the effect of nighttime ozone exposure on yield of snap bean. Nighttime
ozone exposure alone, at 62 ppb, had no effect on yield. In combination with daytime ozone
exposure, nighttime ozone concentrations up to 78 ppb did not affect yields or show a consistent
effect on nocturnal stomatal conductance. When data were pooled across the day and day + night
exposure times, mean daytime ozone levels >62 ppb caused significant yield losses. Burkev et al.
(2012) considered the use of pod weight and seed weight per plant of a sensitive snap bean
genotype as a bioindicator of ozone pollution. Under elevated ozone, the sensitive genotype
showed a 63% decline in pod weight per plant and a similar decline for seed weight per plant. No
significant differences were observed for tolerant genotypes under elevated ozone.
• A few recent studies conducted on U.S. southern piedmont grassland species have added to the
evidence base for ozone effects on nutritive quality of forage (Gilliland et al.. 2016; Gilliland et
al.. 2012).
• A study by Grantz and Vu (2009) reviewed in the 2013 Ozone ISA showed that a hybrid of
sugarcane (Saccharum sp.j exhibited high sensitivity to ozone. However, in a follow-up
comparative study of five hybrids of sugarcane, Grantz et al. (2012) found a wide range of
sensitivities.
8-62
-------
8.5.2
Yield Loss at Regional and National Scales
Global and U.S. modeling studies in the 2013 Ozone ISA found that ozone generally reduced
crop yield and that different crops showed different sensitivity to ozone (Avncrv et al.. 2011; Van
Dingenen et al.. 2009; Tong et al.. 2007). Newly available regional- and national-scale analyses of
ozone's effects on major crops in the U.S., including soybean, wheat (Triticum sp.), and maize (Zea
mays), have enabled further characterization and quantification of yield losses. These advances include
estimates of yield loss based on field data, additional geographic refinement of crop ozone sensitivity, and
for wheat and soybean, analyses of state-by-state sources and contribution of ozone precursors affecting
crop loss.
• Regression analysis of historical ambient ozone concentrations (W126 calculated from hourly
ozone data from U.S. EPA monitors), climate, and yield data for maize and soybean (rainfed
only, irrigated fields excluded from analysis) in the U.S. from 1980 to 2011 was used to estimate
yield losses (Megrath et al.. 2015). Yield losses in the field averaged over the time period were
approximately 10% for maize and 5% for soybean. The authors attribute a temporal improvement
in crop loss to the more stringent ozone air quality standards implemented in 1997 (Figure 8-5).
An unexpected observation from this analysis was that production losses for maize, a C4 plant
thought to be less sensitive to ozone, were greater than for soybean, a C3 plant.
• For wheat and soybean, ozone exposure-response relationships of yield reductions were scaled up
to the continental U.S. to put these losses in context. Relative yield losses were estimated to be
4.9% for wheat and 6.7% for soybean based on 2010 data using the GEOS-Chem model (Lapina
et al.. 2016). State-by-state percentage influence maps were generated for ozone damage. On a
regional basis, the highest relative losses for wheat (12%) and soybean (25%) were in the eastern
U.S. Kansas produces the most wheat but also experiences the greatest percentage of wheat loss
due to ozone. The majority of NOx emissions responsible for ozone-related wheat loss originate
in Texas. For soybean, the highest loss occurs in Illinois which is most affected by NOx
emissions from Missouri. Twenty-seven percent of current soybean losses are attributed to
combined NOx emissions from Illinois, Missouri, and Indiana.
• Tai and Martin (2017) developed an empirical model (partial derivative linear regression [PDLR]
model) from multidecadal data sets to estimate geographical variations across the U.S. in
sensitivity to ozone of wheat, maize, and soybean. This approach takes into consideration strong
ozone-temperature covariation and does not rely on pooled concentration-response functions. For
all three crops, the revised sensitivities (calculated in latitude-longitude grid cells to account for
regional differences in temperature, water, and nutrient availability) are generally higher than
previously indicated by concentration-response functions derived from experimental studies.
Wheat yield sensitivities to ozone were statistically significant spatially along the northern U.S.
border, maize sensitivity was spatially statistically significant at various locations across the U.S.,
and soybean sensitivity was spatially statistically significant in a band from the Great Plains to
the south-central U.S. Crops in regions of elevated ozone and high water use, were more tolerant
to ozone. The PDLR model coupled with ozone and temperature projections by the Community
Earth System model from 2000-2050 have predicted average declines of U.S. wheat, maize, and
soybean of 13, 43, and 28%, respectively.
8-63
-------
X3
O
>*
O -10
-t—'
0
ZJ
"8
L_
0)
O)
1
§ -20
L_
CD
Q.
-30
Soybean
Maize
rai
1980
1990
2000
2010
Year
Note: Each point is a weighted mean of percentage reduction for all counties, where the value of a county was weighted by the
harvested acreage of soybean or maize in that county. Percentage reduction was estimated by using a linear model incorporating
climatic variables and ozone cumulative indices to predict yield using historical values of W126 or a value of 0 W126. The lines are a
local regression analysis fit to the points. The black, dashed, horizontal line marks 0 change for reference. The gray, vertical, dotted
line indicates when the U.S. EPA implemented more stringent standards for ozone emissions. Bars are 95% confidence intervals of
yield reduction for that year.
Source: Reprinted with permission of publisher, Mcqrath et al. (20151.
Figure 8-5 Estimated percentage reduction of soybean and maize yield in the
U.S. from ozone for 1980-2011.
• A modeling study considering the cobenefits associated with decreases of NOx under the U.S.
EPA Clean Power Plan (to regulate emissions of CO2) estimated the effects on total production
and biomass loss of four U.S. crops (potatoes, soybean, cotton, maize) under three policy
scenarios and a reference (ambient air) scenario for the year 2020 (Capps et al.. 2016). In this
analysis, the CMAQ model was used to model exposure values ofW126 and then apply these to
crop distribution maps using published data to estimate biomass loss and potential productivity
loss (PPL). At ambient ozone concentrations, modeled production loss is greatest for potatoes,
soybean, and cotton, with these losses ranging from 1.5 to 1.9%. Scenario 1 (which is closest to
current levels) results in an ozone impact reduction of <2% for each crop. Reductions in PPL of
8.4% (soybean) and 6.7% (cotton) in Scenario 2 (which is most similar to the final Clean Power
Plan) and 6.6 and 3.8% in Scenario 3 (most stringent policy option) suggest that reduction in NOx
with CO2 regulation will decrease agricultural yield losses associated with ozone.
8-64
-------
8.5.3
Summary and Causality Determination
The relationship between ozone exposure and reduced crop yield is well established in the
scientific literature and continues to be an area of active research with hundreds of papers on this topic
published since the 2013 Ozone ISA in the U.S. and other countries (U.S. EPA. 2009). There is a
considerable amount of new research on major U.S. crops, especially soybean, wheat, and nonsoybean
legumes, including updated soybean exposure-response curves. Meta-analyses published since the 2013
Ozone ISA provide further supporting evidence that ozone decreases growth and yield of wheat and
affects reproductive and developmental plant traits important to crop yield. Recent advances in
characterizing ozone's effects on U.S. crop yield include further geographic and temporal refinement of
ozone sensitivity and national-scale estimates of maize and soybean losses from ozone based on actual
yield data. A few studies on grassland species add to the existing body of evidence in the 2013 Ozone ISA
for ozone effects on nutritive quality. New information is consistent with the conclusions of the 2013
Ozone ISA that the body of evidence is sufficient to infer a causal relationship between ozone
exposure and reduced yield and quality of agricultural crops.
8-65
-------
Table 8-10
Ozone and crop yield and quality.
Study Type
Study
and Location Study Species
Ozone Exposure
Effects on Crop Yield
Burkev et al. (2012)
FACE;
SoyFACE,
Champaign,
(40.033°N,
88.233°W)
IL
Phaseolus vulgaris
(snap bean)
Three genotypes
S156—O3 sensitive;
R123, R331—O3
tolerant
O3 exposure from May-August 2006
Ambient (control)—8-h mean = 43 ppb 1-h
max = 29-78 ppb AOT40 = 5.3 ppm-h
SUM60 = 5.3 ppm-h; +O3—8-h
mean = 59 ppb 1-h max = 32-114 ppb
AOT40 = 16.3 ppm-h SUM60 = 27 ppm-h;
+O3+CO2—8-h mean = 59 ppb 1-h
max = 33-112 ppb AOT40 = 16.2 ppm-h
SUM60 = 26.7 ppm-h
Sensitive genotype declined 63% in pod weight per
plant with similar result for seed weight per plant under
elevated O3 compared with control. No significant
difference for tolerant genotypes under elevated O3
compared with control.
Betzelberaer et al.
(2012)
FACE;
SoyFACE,
Champaign,
(40.033°N,
88.233°W)
Glycine max
(soybean)
Seven cultivars
Eight 20-m-diameter SoyFACE plots with
different O3 concentrations were exposed
~8 h/day in two growing seasons (2009,
2010). Target concentrations were
ambient, 40, 55, 70, 85, 110, 130, 160,
200 ppb in 2009, and ambient, 55, 70, 85,
110, 130, 150, 170, 190 ppb in 2010
An exposure-response for soybean was refined from
previous estimates using target concentrations from
ambient to 200 ppb/8 h. As ozone increased, a linear
decrease in yield was observed at the rate of 37 to
39 kg/ha per ppb cumulative exposure >40 ppb. All
seven cultivars showed similar responses to O3 with
the range of responses between 18 to 30 kg ha per
ppb cumulative exposure >40 ppb. At the highest
target concentration of 200 ppb (AOT40 of
67.4 ppm-h) yields declined 64%.
Grantz et al. (2012) Other; Saccharum sp.
California (sugarcane)
(36.6°N, Four hybrids
119.5° W)
Ozone exposure conditions in the
continuously stirred tank reactors were the
same each day, with nominal 12-h mean
exposures of 4, 59, and 114 ppb and 8-h
mean exposures of 3, 76, and 147 ppb.
Leaf-level responses were measured
following exposure to ozone for 7 weeks,
then plants were excised and allowed to
regrow before exposing shoots to ozone
again for another 7 weeks.
The four hybrid clones exhibited a wide range of
sensitivity to O3 measured by net carbon assimilation.
Hybrids containing a greater percentage of Saccharum
spontaneum were less sensitive to ozone.
8-66
-------
Table 8-10 (Continued): Ozone and crop yield and quality.
Study
Study Type
and Location
Study Species
Ozone Exposure
Effects on Crop Yield
Ainsworth et al.
(2014)
FACE;
SoyFACE,
Glycine max
(soybean)
Champaign, IL n genotypes
(40.033°N,
88.233°W)
Eight ambient and eight +O3
20-m-diameter plots exposed 8 h daily for
the growing season. Ambient 8-h ozone
concentration was 44 ppb, and for FACE
plots ranged from 79 to 82 ppb.
Exposure to elevated O3 resulted in approximately
30% avg decrease in seed yield.
Fiscus et al. (2012)
Other;
USDA-ARS
Plant Science
Research
Unit, 5 km
south of
Raleigh, NC
Phaseolus vulgaris
(snap bean)
Two genotypes
S156—O3 sensitive;
R123—O3 tolerant
Two ozone concentrations in
charcoal-filtered air (12-h mean of 0 and
60 ppb) dispensed into outdoor plant
environment chambers. Exposure started
18 days after planting at 1/3 of target
concentrations and were gradually
increased to reach full exposure levels at
21 days after planting. Experiment was
62 days in duration. For +O3 concentration
12-h mean target resulted in daily
AOT40 = 245, SUM06 = 534,
W126 = 295 ppb-h. Two vapor pressure
deficit levels tested (1.26 and 1.96 kPa).
In elevated O3 treatments at high humidity (low vapor
pressure deficit), yield was decreased by 55-72% with
no significant yield loss under low humidity. Both mass
per seed and number of seeds per plant were reduced.
There was a difference in sensitivity in the two
genotypes.
Broberq et al. (2015) Other: OTC, Triticum sp. (wheat) Elevated O3 was at least 30 ppb, but no Meta-analysis of 42 studies showed O3 significantly
FACE in more than 100 ppb daily. reduces 1,000-grain weight (strongly), volume weight,
Europe, Asia, and starch concentration of wheat,
and U.S.
Osborne et al. (2016) Other: OTC or
FACE in U.S.,
Asia, and
China
1982-2014
critical level at which statistically significant (5%) loss
of yield can occur is 32.3 ppb M7. Cultivars varied in
sensitivity to O3 with a yield loss of 13.3 to 37.9% at
55 ppb M7. Sensitivity to O3 increased by an avg of
32.5% between 1960 and 2000.
Glycine max
(soybean)
48 cultivars
Ozone exposure data were all converted to A exposure-response function was calculated by
seasonal 7-h mean from studies that
reported concentration as 8-, 12-, or24-h
mean or 3-mo AOT40. Duration of O3
exposure was at least 60% of growing
season.
pooling relative yield data and plotting against the 7-h
seasonal mean (M7) for 28 experimental studies. All
data were scaled to theoretical yield at 0 ppb. 55 ppb
was used to represent current background. Relative
yield reduction at current background was 17.3%. A
8-67
-------
Table 8-10 (Continued): Ozone and crop yield and quality.
Study
Study Type
and Location
Study Species
Ozone Exposure
Effects on Crop Yield
Meg rath et al. (2015)
Other; Glycine max Regression analysis and statistical model
county-level (soybean), Zea of county-level data of maize and soybean
data from the mays (maize) production, and hourly O3 data from the
U.S. U.S. EPA AQS. Hourly O3 at 2,700 sites
department of from 1980 to 2011 obtained from U.S. EPA
Agriculture AQS, AOT40, SUM06, and W126 were
National calculated. Indices were summed over the
Agricultural growing season (J, J, A); only W126 data
Statistics was reported because it was the most
Service from linear.
1980 to 2011
Average yield loss from 1980 to 2011 in rainfed fields
was 9.8% for maize and 5.5% for soybean.
Percentage losses of crop yield showed a temporal
improvement in crop loss corresponding to
implementation of 63 standards in 1997. Current loss
of production is estimated at 4-7%.
Vanloocke et al.
(2012)
FACE;
SoyFACE,
Champaign,
(40.04°N;
88.24°W)
IL
Glycine max 12-h mean O3 in the experimental plots of
(soybean) 40, 46, 54, 58, 71, 88, 94, 116 ppb.
With increasing O3, harvested seed yield decreased
linearly; 64% reduction in yield at highest O3 treatment
compared with lowest.
Tai and Martin (2017)
Other;
multidecadal
U.S. crop yield
and climate
data to
estimate
geographical
variation
across the
U.S.
Glycine max
(soybean), Triticum
(wheat), Zea mays
(maize)
Three cumulative O3 annual exposure
metrics, AOT40, SUM-06, and W126,
calculated from hourly ozone observations
from the AQS and CASTNET networks
averaged over 1993-2010.
An empirical (partial derivative linear regression)
model incorporating the strong ozone-temperature
covariation was used to calculate crop sensitivity to O3.
For all three crops, modeled sensitivities (calculated in
latitude-longitude grid cells to account for regional
differences in temperature, water, and nutrient
availability) are generally higher than previously
indicated by concentration-response functions derived
from experimental studies.
Leisner and
Ainsworth (2012)
Other; all
locations
included (but
no summary
of geographic
distribution of
studies
included)
All species included
(data set includes
monocots and
dicots, perennials
and annuals,
determinate and
indeterminate
growth habits, C3
and C4 plants)
Grouped/analyzed exposures in several
ways:
(1) Charcoal-filtered air vs. elevated
O3—elevated O3 in multiple categories:
>100 ppb, 70-100 ppb, 40-70 ppb,
<40 ppb
(2) Ambient air vs. elevated O3—elevated
O3 in multiple categories: >100 ppb,
70-100 ppb, 40-70 ppb, <40 ppb
Grain or seed yield per unit area declined 19% at
average O3 concentration of 60 ppb compared with
charcoal-filtered air. Compared with ambient air, yield
decreased by 25% at an average O3 concentration of
79 ppb.
8-68
-------
Table 8-10 (Continued): Ozone and crop yield and quality.
Study
Study Type
and Location
Study Species
Ozone Exposure
Effects on Crop Yield
Leisner et al. (2017)
FACE;
SoyFACE,
Champaign,
(40.04°N,
88.24°W)
Glycine max Elevated O3 fumigation system increased
(soybean) O3 to 100 ppb from 10:00 a.m. to 5:00 p.m.
IL except when leaves were wet.
Season-long 8-h avg ambient O3 was
50.6 ppb, and the 8-h season-long
elevated O3 was 69.7 ± 1.3 ppb.
Number of seed pods per node was significantly
reduced in O3 treated soybeans. Avg 3.4 pods per
node (at 50.6 ppb ambient) decreasing to 2.8 pods per
node in the elevated O3 treatment (69.7 ppb).
Llovd et al. (2018)
Greenhouse;
Pennsylvania
State
University
(40.80°N,
77.85°W)
Phaseolus vulgaris
(snap bean)
Two genotypes
S156—O3 sensitive;
R123—O3 tolerant
O3 treatments were a combination of O3
concentration and treatment time as
follows: (1) 45 ppb O3 day-only, (2) 75 ppb
O3 day-only, (3) 45 ppb O3 day + night,
(4) 75 ppb O3 day + night, (5) 30 ppb
night-only, (6) 60 ppb night-only.
Nighttime O3 exposure alone, at 62 ppb, had no effect
on the yield of either genotype. In combination with
daytime O3 exposure, nighttime concentrations up to
78 ppb did not impact yields. When data were pooled
across the day and day + night exposures times, mean
daytime O3 levels (62 ppb) caused significant yield
decreases. Under control conditions, R123 and S156
produced similar pod masses in two of the three trials.
In all three trials, R123 produced significantly greater
yields by mass than S156 with elevated O3.
Sun et al. (2014)
FACE;
SoyFACE,
Champaign,
(40.04°N,
88.24°W)
IL
Glycine max
(soybean)
Two cultivars
Dwight and IA3010
Nine plots fumigated with various O3
concentrations from early vegetative stage
to maturity. Daily 9-h avg concentrations
over the 105 days were 37 (ambient), 40,
46, 54, 58, 71, 89, 95, and 116 ppb.
O3 caused greater damage at later reproductive stages
and in older leaves. Soybeans grown under O3 levels
of 116 ppb were senescent 1 week earlier than plants
grown under ambient control (37 ppb). Average
decrease of photosynthesis, total nonstructural
carbohydrate levels, and many metabolites and amino
acids (correlated to seed yield) was 7% for a 10-ppb
increase in O3. Loss of seed yield mainly due to loss of
photosynthetic capacity and canopy senescence
resulting in shorter growing season.
Gilliland et al. (2012)
OTC; Auburn
University,
Auburn, AL
Lolium arundinacae
(tall fescue),
Paspalum dilatatum
(dallisgrass),
Cynodon dactylon
(common Bermuda
grass), Trifolium
repens (white
clover)
Grasses in six OTC chambers (three
chambers per treatment) exposed for
8 weeks. Mean monthly 12-h ambient NF
was 21-32 ppb (avg peak 49 ppb). Mean
monthly 2x ambient was 37 to 56 ppb (avg
peak 102 ppb). Rabbits were fed (in the
form of 50 g dried forage blocks) a mixture
of the four plant species grown in OTCs at
either ambient or 2x ambient O3.
Neutral detergent fiber and acid detergent fiber
digestibility was significantly lower for grasses grown
under elevated O3.
8-69
-------
Table 8-10 (Continued): Ozone and crop yield and quality.
Study
Study Type
and Location
Study Species
Ozone Exposure
Effects on Crop Yield
Gilliland et al. (2016)
OTC; 5 km
north of
Auburn
University
Auburn, AL
Lolium arundinacea
(tall fescue),
Paspalum dilatatum
(dallisgrass),
Cynodon dactylon
(common Bermuda
grass), Trifolium
repens (white
clover)
12 OTC chambers (6 chambers ambient,
6 chambers 2x ambient, then treated to
3 levels of precipitation). Grasses were
exposed for 4 mo with the mean 12-h O3
concentration of 31 ppb (NF) and 56 ppb
(2x ambient). Average peak O3 = 39 (NF)
and 77 ppb (2*). Peak avg 1-h O3 = 73
(NF) and 155 ppb (2*).
Three grass species pooled into one "grass" sample.
Under elevated O3, primary growth of grasses (dry
matter yield) increased 19% compared with ambient,
while clover decreased in nutritive quality (increase in
acid detergent lignin). In regrowth, clover in 2x O3 had
a 60% decrease in DM yield, while grasses had lower
concentrations of neutral detergent fiber, higher
relative food value, and increased crude protein.
Clover was sensitive to O3 with decreased nutritive
quality and higher response in regrowth harvests.
Lapina et al. (2016)
Other;
continental
U.S.
Glycine max
(soybean), Triticum
(wheat)
Exposure response functions for W126,
AOT40, and mean metric for total O3
exposure were used to model relative loss
estimates. The GEOS-Chem adjoint model
was applied to estimate source-receptor
relationships between crop yield reduction
and emission sources.
Analysis of year 2010 O3 losses in wheat, soybean,
and two tree species showed sources of O3 in the U.S.
and how individual states' emissions contribute to O3
damage; the study suggests that most vegetation
damage is attributable to local, not international
sources. U.S. anthropogenic NOx contributions were
the highest total contributors (75-77%). State-by-state
maps provide information on sources associated with
vegetative damage. Relative yield loss: wheat 4.9%
and soybean 6.7%.
Capps et al. (2016) Other; U.S.
Zea mays (maize),
Gossypium (cotton),
Solanum tuberosum
(potato), Glycine
max (soybean)
Uses U.S. EPA-developed CMAQ model to
model exposure values of W126 under
three regulatory scenarios as well as a
reference (ambient) over CONUS. Three
C02-reduction scenarios were modeled.
Scenario 1—closest to current levels,
Scenario 2—most similar to the final Clean
Power Plan, and Scenario 3—most
stringent policy option.
At ambient O3 concentrations, modeled production
loss is greatest for potatoes, soybean, and cotton, with
these losses ranging from 1.5 to 1.9%. Although
potatoes show greatest impacts currently, the CO2
mitigation strategies improve yields the least. These
small changes are attributable to much of the potato
potential productivity loss (PPL) arising from high
W126 in southern California, which is not substantially
altered in the policy scenarios. The PPLs for soybeans
and cotton are reduced substantially in Scenario 2 and
Scenario 3, resulting in 8.4% (6.6%) and 6.7% (3.8%)
PPL reductions, respectively, for each crop. Scenario 1
reduces the O3 impact by <2% for each crop.
8-70
-------
Table 8-10 (Continued): Ozone and crop yield and quality.
Study
Study Type
and Location
Study Species
Ozone Exposure
Effects on Crop Yield
Pleiiel etal. (2018)
Other; nine
countries
Triticum sp. (wheat)
CF and NF in OTC treatments with
daytime O3 concentration converted to 7-h
mean.
Average yield loss per ppb ozone of 0.38% across
33 studies. Grain yield, grain mass, total aboveground
biomass, starch concentration, starch yield, and
protein yield significantly declined in nonfiltered air
compared with charcoal-filtered air in these studies,
with starch yield (10.9%) being the most strongly
affected.
AOT40 = seasonal sum of the difference between an hourly concentration at the threshold value of 40 ppb, minus the threshold value of 40 ppb; AQS = (U.S. EPA) Air Quality
System (database); C3 = plants that use only the Calvin cycle for fixing the carbon dioxide from the air; C4 = plants that use the Hatch-Slack cycle for fixing the carbon dioxide from
the air; CASTNET = Clean Air Status and Trends Network; CF = charcoal-filtered air; C02 = carbon dioxide; FACE = free-air C02 enrichment; kg/ha = kilograms per hectare;
kPa = kilopascal; NF = nonfiltered air; NOx = nitrogen oxides; 03 = ozone; OTC = open-top chamber; ppb = parts per billion; ppm = parts per million; SUM06 = seasonal sum of all
hourly average concentrations > 0.06 ppm; SUM60 = sum of hourly ozone concentrations equal to or greater than 60 ppb; W126 = cumulative integrated exposure index with a
sigmoidal weighting function.
8-71
-------
8.6 Herbivores: Growth, Reproduction, and Survival
In the 2013 Ozone ISA there was no causality determination between ozone exposure and effects
on herbivores. Reviewed studies included species-level responses (i.e., growth, reproduction, survival)
and population- and community-level responses of herbivorous insects due to ozone-induced changes in
plants. Ozone exposure can lead to changes in plant physiology (Figure 8-2). such as by modifying the
chemistry and nutrient content of leaves (U.S. EPA. 2013; Menendez et al.. 2009). These changes can
have significant effects on herbivore physiology and behavior by affecting plant-herbivore interactions.
In the 1996 Ozone AQCD, the 2006 Ozone AQCD, and the 2013 Ozone ISA, multiple studies showed
statistically significant effects on insect growth, fecundity, and development. The effects, however, were
highly context- and species-specific, there was no clear trend in directionality of response for most
effects, and not all species tested showed a response (U.S. EPA. 2013. 2006. 1996). Studies on insect
herbivores in previous ozone assessments included species from the orders Coleoptera (weevils, beetles),
Hemiptera (aphids), and Lepidoptera (moths). There was no consensus in the 2013 Ozone ISA on how
insects and other wildlife respond to elevated ozone.
Since that review, additional research has been published for more herbivorous insects, as well as
for a few mammalian herbivores, at various levels of ozone exposure (see Table 8-13). As described in
the PECOS tool, the scope for this section includes studies on any continent in which alterations in
invertebrate and vertebrate responses were measured in individual species or at the population and
community level to concentrations of ozone occurring in the environment or experimental ozone
concentrations within an order of magnitude of recent concentrations (as described in Appendix 1).
Ozone
Growth
Survival
Reproduction
Herbivore
populations
and
communities
Altered leaf
chemistry and
nutrient
content
Herbivores:
Figure 8-6 Conceptual model of ozone effects on herbivore growth,
reproduction, and survival.
8-72
-------
8.6.1 Individual-Level Responses
Consistently measured individual-level responses to ozone exposure include measures of growth,
(including development time and adult and pupal mass) reproduction (e.g., fecundity, oviposition
preference), and survival. A conceptual model of ozone effects on herbivores (Figure 8-6) illustrates
cascading effects from individual-scale responses to populations and communities. Both the 1996 Ozone
AQCD and 2006 Ozone AQCD stressed the variability in reported effects, including the lack of a
consistent pattern in directionality and degree of response (U.S. EPA. 2006. 1996). In the 2013 Ozone
ISA, a meta-analysis that included 16 studies published on insect herbivore species between 1996 and
2005 found that elevated ozone decreased development time and increased pupal mass in insect
herbivores, with more pronounced effects occurring with longer durations of exposure (Valkama ct al..
2007). In addition, for chewing insects, the meta-analysis found that relative growth rate increased under
elevated ozone. There were no effects found on consumption, survival, or number of eggs laid (Valkama
et al.. 2007). In an assessment of five herbivore species (three moths and two weevils) only the growth of
larvae of one moth species was affected (Peltonen et al.. 2010).
Since the 2013 Ozone ISA, there is new evidence for endpoints related to growth, reproduction,
and survival, as summarized below and in Table 8-14. As in the 2013 Ozone ISA, the insects encompass
the orders Coleoptera, Hemiptera, and Lepidoptera. In addition to studies of insects, there are recent
studies on a few mammalian herbivores. The new evidence describes ozone effects in a few additional
species:
Growth:
• In the gypsy moth (Lymantria dispar) and tent caterpillar (Malacosoma disstria), exposure to
1.5x ambient ozone led to decreased growth (Couture et al.. 2012). Female voles (Microtus
ochrogaster) that consumed ozone-exposed plants showed reduced growth [1.5x ambient vs.
control; Habeck and Lindroth (2013)1.
• In the whitefly (Bemisia tabaci) andZ. dispar, exposure to elevated ozone (72, 1.5/ ambient,
respectively) increased development time (Couture and Lindroth. 2012; Cui et al.. 2012).
However, at higher levels (238 vs. 50 ppb), development time decreased in B. tabaci (Hong et al..
2016V In the cabbage moth (Pieris brassicae), elevated ozone decreased development time
[120 ppb vs. 15-20 ppb; Khaling et al. (2015)1.
• In the leaf beetle (Agelastica coerulea) adults showed a preference to feed on leaves treated with
elevated ozone [ambient vs. 60 ppb; Agathokleous et al. (2017)1.
• Rabbits fed a mixture of common southern piedmont grassland species grown under
concentrations of ozone 2 times (mean monthly 12 hour 37 to 56 ppb) ambient ozone had
decreased digestible dry matter intake due to significantly lower neutral detergent fiber and acid
detergent fiber digestibility compared to ambient [mean monthly 12 hour 21-32 ppb; Gilliland et
al. (2012)1.
• A loss in liveweight gain of 3.6 to 4.4% was predicted for lambs in the U.K. from 2007 to 2020
due to ozone effects on grasslands. With an ozone concentration increase from 20 to 30 ppb,
liveweight gain was predicted to decrease by 12%. (Haves et al.. 2016).
8-73
-------
Reproduction:
• In B. tabaci, exposure to elevated ozone (72 vs. 37 ppb) decreased fecundity (Cui et al.. 2016b;
Cui et al.. 2012). However, at higher levels (238 vs. 50 ppb), fecundity increased (Hong et al..
2016). In L. dispar, exposure to 1.5/ ambient levels decreased fecundity (Couture and Lindroth.
2012).
• Egg laying in the diamondback moth (Plutella zylostella) was significantly higher in the absence
of ozone (when given a choice between artificial leaves fumigated with plant volatiles mixed with
clean air or elevated ozone [80 ppb; Li and Blande (2015); see Section 8.71). In the same lab
study, plants exposed to herbivore-damaged neighbor plants had more eggs deposited on them at
ambient ozone (10 ppb) than plants exposed to undamaged control plants. In presence of 80 ppb
ozone, the preference for egg-laying on damaged plants was lost. Under field conditions,
P. zylostella laid more eggs on plants exposed to control levels (10 ppb) compared with elevated
ozone [30-80 ppb; Giron-Calvaetal. (2016)1.
• In B. tabaci, adults preferred control plants for oviposition [37 vs. 72 ppb; Cui et al. (2014)1.
Survival:
• In P. brassicae, there was a nonsignificant trend whereby larval mortality tended to increase with
increasing ozone levels [15-20 ppb, 70, 120 ppb; Khaling et al. (2015)1. Elevated ozone (50 and
150 ppb vs. 0.5 ppb) increased mortality inMetopolophium dirhodum aphids (Telesnicki et al..
2015).
• In L. dispar, survival of early instars decreased in response to feeding on leaves exposed to
elevated ozone [1.5 x ambient; Couture and Lindroth (2012)1.
• At higher ozone levels, lifespan was prolonged in B. tabaci [238 vs. 50 ppb; Hong et al. (2016)1.
8.6.2 Population- and Community-Level Responses
Changes in host plant quality resulting from elevated ozone can alter the population density and
structure of associated insect herbivore communities, ultimately affecting ecosystem processes
(Cornelissen. 2011). In the 2013 Ozone ISA, these population- and community-level responses included
altered population growth rates in aphids (Menendez et al.. 2010; Awmack et al.. 2004). reduced total
arthropod abundance at the Aspen FACE site (Hillstrom and Lindroth. 2008). and changes in genotypic
frequencies of aphids over multiple generations (Mondor et al.. 2005). Recent studies report metrics of
altered population and community structure (e.g., population size, relative species abundance) adding to
the evidence base for herbivore responses to ozone at higher levels of biological organization. New
studies include:
• In a study from Aspen FACE, elevated ozone did not consistently influence arthropod community
composition (Hillstrom et al.. 2014).
• In a mesocosm study, past ozone exposure had no effect on the richness, diversity, or evenness of
the arthropod community associated with the descendant plant community but did increase the
relative abundance of carnivore arthropods while decreasing the relative abundance of herbivore
arthropods (Martinez-Ghersa et al.. 2017).
8-74
-------
• Under low ozone conditions (1.5 ppb), the population size of Rhopalosiphum padi aphids was
dependent on the symbiotic status of the host. However, under high ozone conditions (120 ppb),
this difference disappeared (Ueno et al.. 2016). InM dirhodum aphids, ozone exposure did not
affect population size, but did affect the proportion of dispersing aphids, with reduced dispersion
in the ozone treatments [0.5 vs. 50, 150 ppb; Telesnicki et al. (2015)1.
8.6.3 Summary and Causality Determination
Previous ozone assessments have summarized herbivorous insect-plant interactions and found
information on a range of insect species in the orders Coleoptera, Hemiptera, and Lepidoptera (U.S. EPA.
2013. 2006. 1996). The majority of studies focused on growth and reproduction, while fewer studies
considered herbivore survival and population and community-level responses to ozone. Although
statistically significant effects were observed frequently for growth (Table 8-11) and reproduction
(Table 8-12) responses were highly context- and species-specific, and not all species tested showed a
response. Recent studies reviewed here, including multiple experimental studies conducted by multiple
research groups, expand the evidence base for the effects of elevated ozone on growth and reproduction in
herbivores. Further, while effects were observed, there remains a more limited number of studies on the
effects of ozone on survival and population/community-level responses. The effects of ozone exposure on
plant biomass and biochemistry likely account, at least partially, for the observed changes across all
endpoints that were assessed. It is also possible that variation in the herbivore responses to ozone stem
from differences in study design, whereby ozone exposure was sometimes direct and other times indirect
via effects on vegetation. Further uncertainties relate to differences in the plant consumption methods
across species, for example chewing versus phloem-feeding in insects. Considering the large body of
available evidence (Table 8-13) on growth and reproduction (i.e., 1996, 2006 AQCD, 2013 Ozone
ISA, and more recent research efforts) and recognizing the above uncertainties, this ISA makes a
new causality determination that the body of evidence is sufficient to infer a likely to be causal
relationship between ozone exposure and alteration of herbivore growth and reproduction.
8-75
-------
Table 8-11 Summary of studies reporting altered growth in herbivores.
Herbivore
Plant
Exposure
PPb
Growth
Adult
Mass
Pupal
Mass
Development
Time
Reference
Gypsy moth
(Lymantria
dispar); tent
caterpillar
(Malacosoma
disstria)
Trembling
aspen
(Popuius
tremuioides);
paper birch
(Betuia
papyrifera)
50-100
Couture et al.
(2012)
Voles (Microtus
ochrogaster)
Soiidago
canadensis;
Taraxacum
officinale
50-100
Habeck and
Lindroth (2013)
Pieris brassicae
Brassica nigra
120
Khalina et al. (2015)
Lymantria
dispar
Trembling
aspen
(Popuius
tremuioides);
paper birch
(Betuia
papyrifera)
50-100
T
Couture and
Lindroth (2012)
Whitefly
(Bemisia tabaci)
Tomato plant
72.2
T
Cui et al. (2012)
Whitefly
(Bemisia tabaci)
Tomato
(Lycoperiscon
esculentum)
238
Hona et al. (2016)
Aphid (Rhopaio-
siphum padi)
Lolium
multiflorum
120
Uenoetal. (2016)
Aspen leaf
beetle
(Chrysomeia
crotchi)
Trembling
aspen
(Popuius
tremuioides)
50-100
T
Vique and Lindroth
(2010)c
Aphid
(Capegillettea
betulaefoliae)
Paper birch
(Betuia
papyrifera)
50-60
No effect
No
effect
Awmack et al.
(2004)°
Epirrita
autumnata
Silver birch
(Betuia
pendula)
2x
ambient
Peltonen et al.
(2010)c
Aphid
Broad bean
85
Dohmen (1988)a
Aphid
Broad bean
100
(>24 h)
1
Brown et al. (1992)a
8-76
-------
Table 8-11 (Continued): Summary of studies reporting altered growth in
herbivores.
Herbivore
Plant
Exposure
PPb
Adult
Growth Mass
Pupal
Mass
Development
Time
Reference
Monarch
butterfly
Milkweed
150-178
T
Bolsinaer et al.
(1991): Bolsinaer et
al. (1992)a
Gypsy moth
Hybrid poplar
(P. trisitis x P.
balsamifera)
Lindroth et al.
(1993)b
Gypsy moth
Sugar maple
(Acer
saccharum)
No effect
Lindroth et al.
(1993)b
Bug (Lygus
rugulipennis)
Scots pine
Manninen et al.
(2000)b
Sawfly (Gilpinia
pallida)
Scots pine
T
Manninen et al.
(2000)b
Colorado potato
beetle
(Leptinotarsa
decemlineata)
Potato
(Solatium
tuberosum)
No effect
Costa et al. (2001 )b
M. disstria
Aspen
T
Percv et al. (2002)b
Tobacco
horn worm
(Manduca
sexta)
Tobacco
(Nicotiana
tabacum)
T
Jackson et al.
(2000)b
a1996 Ozone AQCD.
"2006 AQCD.
°2013 ISA
8-77
-------
Table 8-12 Summary of studies reporting altered reproduction in herbivores.
Herbivore
Plant
Exposure (ppb)
Fecundity
Ovi position
Preference
Reference
Whitefly (Bemisia
tabaci)
Tomato
(Lycopersicon
escuientum)
72
Cui et al. (2016a)
Whitefly (Bemisia
tabaci)
Tomato
(Lycopersicon
escuientum)
72
Cui et al. (2012)
Whitefly (Bemisia
tabaci)
Tomato
(Lycopersicon
escuientum)
238
T
Hona et al. (2016)
Lymantria dispar
Trembling aspen
(Populus
tremuloides);
Paper birch
(Betula
papyrifera)
50-100
Couture and Lindroth (2012)
Diamondback
moth (Piuteiia
zyiosteiia)
Brassica
oleracea
80
Li and Blande (2015)
Diamondback
moth (Piuteiia
zyiosteiia)
Brassica
oleracea
30-80
Giron-Calva et al. (2016)
Colorado potato
beetle
(Leptinotarsa
decemlincata)
Potato (Solanum
tuberosum)
No effect
Costa et al. (2001 )b
Beetle
Cottonwood
200
1
Coleman and Jones (1988)a
Hornworm moth
Ambient +70%
T
Jackson et al. (1999)b
a1996 Ozone AQCD.
"2006 AQCD.
8-78
-------
Table 8-13 Summary of evidence for likely to be causal relationship between
ozone exposure and alteration of herbivore growth and reproduction.
Rationale for
Causality
Determination
Key Evidence
Key References
Multiple experimental
studies by multiple
research groups show
effects on growth in
herbivorous insects,
limited evidence in
mammalian herbivores
Increased or decreased,
herbivore growth, change in
pupal or adult mass, altered
development time
(Table 8-11)
Hong etal. (2016). Ueno et al. (2016). Khalinq et al. (2015).
Habeckand Lindroth (2013), Couture et al. (2012), Couture
and Lindroth (2012), Cui et al. (2012), Vique and Lindroth
(2010). Peltonen et al. (2010). Awmack et al. (2004)
Section AX-9.3.3.1 U.S. EPA (2006)
Multiple experimental Increased or decreased
studies by multiple fecundity, altered oviposition
research groups show preference (Table 8-12)
effects on reproduction
in herbivorous insects
Giron-Calva et al. (2016), Hong et al. (2016), Cui et al.
(2016b), Li and Blande (2015), Couture and Lindroth
(2012). Cui etal. (2012). Section AX-9.3.3.1 U.S. EPA
(2006)
8-79
-------
Table 8-14 Ozone exposure and effects on herbivores.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Insects and Other Wildlife
Cuietal. (2012)
OTC; Beijing, China
(40.183°N,116.4°E)
Three genotypes
(wild type, 35S,
spr2) of tomato
(Lycopersicon
esculentum) and
insect pest whitefly
(Bemisia tabaci)
Two ozone treatments: current
ambient O3 (37.3 ppb) and twice
the current ambient (72.2 ppb).
All values are averages from
9:00 a.m. to 5:00 p.m.
In general, (varied by tomato genotype) whiteflies on
ozone-treated plants had longer development time and
reduced fecundity, which was correlated to phytochemical,
enzymatic, and genetic alterations in the tomato plants. In
response to O3, there was an increased duration of larval
stage with: 12.22% on wild type, 11.72% for 35S genotype,
4.64% for spr2 genotype; total length of larval and pupal
stages: 9.65% for wild type, 9.71% for 35S genotype, and
5.41% for spr genotype; decreased fecundity: 19.33%
reduction for whiteflies on wild-type tomato, 34.76% for 35S
genotype, not significant for spr2; intrinsic rate of increase:
12.2% for wild type, 13.97% for 35S, not significant for
spr2.
Li and Blande
(2015)
Lab; Kuopio, Finland
Insect: Plutella
zylostella
Plant: Brassica
oleracea var.
italica
O3: ambient (10 ppb), 80 ppb
Plants exposed to herbivore-damaged neighbor plants had
more eggs deposited on them at ambient Osthan plants
exposed to undamaged control plants. At 10 ppb, there
were significantly more eggs deposited on plants previously
exposed to herbivore-damaged plants than those exposed
to undamaged plants. In the presence of 80 ppb O3, the
preference for damaged plants was lost. When given a
choice between artificial leaves fumigated with VPSCs
mixed with clean air or elevated O3, P. xylostella laid
significantly more eggs in the absence of O3.
Telesnicki et al.
(2015)
OTC; (34.035°S,
58.029° W)
Insect:
Metopolophium
dirhodum
Plant: Triticum
aestivum cultivar
Cronox
O3: filtered air (0.5 ± 0.3 ppb),
50 ± 5 and 150 ± 10 ppb.
Aphids received a single
exposure to ozone during the
first 6 h of daylight
The proportion of dead aphids was 0.054, 0.238, and 0.139
for control, 50 ppb, and 150 ppb O3, respectively. The
proportion of dispersed aphids was 0.654, 0.319, and 0.405
for control, 50 ppb, and 150 ppb O3, respectively. The
population of surviving aphids increased similarly for all
treatments. The proportion of aphids dispersing from the
diet cages was reduced by O3 treatment.
8-80
-------
Table 8-14 (Continued): Ozone exposure and effects on herbivores.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Insects and Other Wildlife
Cuietal. (2014)
OTC; Xiaotangshan
County, Beijing, China
(40.183°N, 116.4°E)
Insect: Bemisia
tabaci, Encarsia
formosa
Plant: Solanum
esculentum
cultivar
Castlemare,
Jasmonic acid (JA)
defense-enhanced
genotype (JA-OE
35S)
O3: ambient air
(average = 37.3 ppm) and
72.2 ppm (~2x ambient),
8 h/day for 24 days
O3 level, whitefly herbivory and tomato genotypes
significantly affected the feeding and oviposition
preferences of B. tabaci. Adult whiteflies preferred control
plants over other treatments (i.e., O3, herbivory and
O3 + herbivory treated) for feeding and oviposition.
Compared with S35 plants, adult whiteflies preferred
wild-type plants for feeding under control and herbivory
treatments and for oviposition under control, O3, herbivory,
and O3 + herbivory treatments. In a behavioral assay,
parasitoids preferred O3 + herbivory plants. Using an
olfactometer, it was determined that the 35S plants were
preferred by adult parasitoids under O3, herbivory, and
O3 + herbivory treatments. Adult parasitoids showed no
preference for either genotype under control conditions.
Hong et al. (2016) Greenhouse; China
Insect: Bemisia
tabaci
Plant: tomato
Fungi: Beauveria
bassiana
Ambient (50 ± 10 ppb) and
elevated (280 ± 20 ppb) 8 h/day
for 40 days
Elevated O3 shortened development time, prolonged adult
lifespan, increased fecundity, increased the female ratio of
offspring, and decreased the weight of newly enclosed
adults. In the presence of elevated O3 and fungal
challenge, whitefly (adult and pupae) mortality increased,
LC50 decreased, and the LT50 was shortened.
Cuietal. (2016b)
OTC; Observation
Station for Global
Change Biology,
Beijing China
(40.183°N, 116.4°E)
Insect: Bemisia.
tabaci biotype B
Plant: wild-type
tomato (L
esculentum
cultivar
Castlemart), 35S:
prosytemin
transgenic tomato
plants
Virus: tomato
yellow leaf curl
virus
Ambient (37.3) and elevated
(72.2 ppb). The OTCs were
ventilated with air daily from
8:00 a.m. to 6:00 p.m. The
experiment was terminated after
6 weeks
O3 and tomato yellow leaf curl virus infection decreased B.
tabaci fecundity and abundance, the greatest effect was
observed for the combination of O3 and infection.
8-81
-------
Table 8-14 (Continued): Ozone exposure and effects on herbivores.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Insects and Other Wildlife
Couture and
Lindroth (2012)
FACE; Aspen FACE,
near Rhinelander, Wl
(45.7°N, 89.5°W)
Gypsy moth
(Lymantria dispar)
fed ozone exposed
quaking aspen
(Populus
tremuloides) or
paper birch (Betula
papyrifera) leaves
Treatments for 1998-2008 were
ambient Os W126 = 2.1
-8.8 ppm-h and elevated
O3 = 12.7-35.1 ppm-h. Ambient
air CO2 and elevated (560 ppm)
CO2. For hourly ozone
concentrations during
experimental ozone treatment,
see Kubiske and Foss (2015).
For insect bioassays, insects
were fed leaves from 11-yr-old
ozone exposed trees for 7 days
Survivorship of early instars decreased by 16%,
development time increased (5%, small but significant
increase) across both tree species. Female pupal weight
decreased 8%, while the effect on males was not
significant. Development time was more influenced by tree
species than by treatment. With O3 exposure, insect egg
production decreased by 28% in both tree species. The
authors used statistical relationships to relate observed
changes to alterations in foliar chemistry. With O3-CO2
interactions, effects on mortality and development time
ameliorated.
Couture et al.
(2012)
FACE; Aspen FACE,
near Rhinelander, Wl
(45.7°N, 89.5°W)
Gypsy moth
(Lymantria dispar)
and forest tent
caterpillar
(Malacosoma
disstria) fed
ozone-exposed
quaking aspen
(Populus
tremuloides) or
paper birch (Betula
papyrifera) leaves
Treatments for 1998-2007 were
ambient O3 W126 = 2.9-8.8
ppm-h and elevated
O3 = 13.1-35.1 ppm-h. Ambient
air CO2 and elevated (560 ppm)
CO2. For hourly ozone
concentrations during
experimental ozone treatment,
see Kubiske and Foss (2015).
For insect bioassays, in 2007,
insects were fed leaves from
11-yr-old ozone exposed trees
for 7 days
Gypsy moth under elevated O3: growth decreased with both
aspen and birch. Consumption of both tree species
increased by 10% and produced more frass (11% with
aspen, 3% with birch). Finally, conversion of foliage to
biomass decreased by 20 and 8% with aspen and birch,
respectively. Tent caterpillar under elevated O3: growth
decreased with both aspen (32% response) and birch
(7% response). Increased consumption of both aspen
(37%) and birch (15%), produced more frass (23% increase
with aspen), conversion of foliage to biomass decreased by
31 and 7% with aspen and birch, respectively. O3-CO2
interactions: Negative effects of O3 on herbivores were
offset to some degree by CO2 but more so for gypsy moths
than for tent caterpillars.
8-82
-------
Table 8-14 (Continued): Ozone exposure and effects on herbivores.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Insects and Other Wildlife
Couture et al.
(2015)
FACE research facility;
Rhinelander, Wl
(45.7°N, 89.5°W)
Insect: specific
insects not
monitored, insect
frass
collected/analyzed
Plants: two aspen
genotypes (42 and
271, Populus
tremuloides) and
paper birch (Betula
papyrifera)
Treatments for 2006-2008 were
ambient Os W126 = 5.6, 4.9,
2.1 ppm-h and elevated
Os= 14.6, 13.1, 12.7 ppm-h.
Ambient air CO2 and elevated
(560 ppm) CO2. For hourly
ozone concentrations during
experimental ozone treatment,
see Kubiske and Foss (2015).
Leaves were collected from the
lower and upper thirds of the
canopies of 16 trees from each
of the 12 rings, in June, July,
and August of 2006, 2007, and
2008
Elevated O3 elicited a "modest" decrease in canopy
damage. Although organic deposition by insects was not
affected by elevated O3; N flux from the canopy to the soil
decreased by 19% and the ratio of foliar C:N increased.
Elevated CO2 and O3 (alone, not simultaneously) increased
total abundance of herbivorous insects at the canopy level.
Elevated O3 decreased the negative effects of herbivory on
ANPP.
Meehan et al.
(2014)
FACE research facility;
Rhinelander, Wl
(45.7°N, 89.5°W)
Aspen (Populus
tremuloides) and
paper birch (Betula
papyrifera) and
associated insect
herbivore
community
Treatments for 2006-2008 were
ambient O3 W126 = 5.6, 4.9,
2.1 ppm-h and elevated
Os= 14.6, 13.1, 12.7 ppm-h.
Ambient air CO2 and elevated
(560 ppm) CO2. For hourly
ozone concentrations during
experimental ozone treatment,
see Kubiske and Foss (2015)
Dry matter C concentration was 6% higher under elevated
O3, N concentrations 6% lower under O3, and C:N ratios
were 10% higher under elevated O3. Total C flux and
tannins by herbivores were not affected by O3. Elevated O3
did not affect dry matter (13% reduction but NS), C, or
tannin input, but had a small negative effect on N flux by
herbivores.
Hillstrom et al.
(2014)
FACE research facility;
Rhinelander, Wl
Insect: stratified
sampling of
canopy
Plant: aspen
genotypes (216,
217, 42E; Populus
tremuloides) and
paper birch (Betula
papyrifera)
Treatments for 2005-2007 were
ambient O3 W126 = 7.3, 5.6,
4.9 ppm-h and elevated
O3 = 29.6, 14.6, 13.1 ppm-h.
Ambient air CO2 and elevated
(560 ppm) CO2: For hourly
ozone concentrations during
experimental ozone treatment,
see Kubiske and Foss (2015)
The effects of elevated CO2 and O3 on arthropod
abundance were species-specific and temporally variable.
Unlike the results for aspen and birch trees exposed to
elevated CO2, the 10 most responsive arthropod species
exhibited similar differences among species and years
sampled in the elevated O3 exposure group. Overall, the
abundance of phloem-feeding increased and leaf chewing
and galling guilds decreased under elevated CO2. Elevated
O3 had the opposite effect. Effects on arthropod species
richness were small and not believed to be biologically
meaningful. While elevated CO2 and elevated O3 did not
consistently influence arthropod community composition,
tree genotype and the time of sample collection did.
8-83
-------
Table 8-14 (Continued): Ozone exposure and effects on herbivores.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Insects and Other Wildlife
Giron-Calva et al. Field; O3 FACE facility; Insect; Pieris
Laboratory: control (10 ppb) and
(2016)
Laboratory; Kuopio,
Finland (62.013°N,
27.035°E)
brassicae, Plutella elevated (30-80 ppb). Field:
zylostella
Plant: Brassica
oleracea var.
capitata, Brassica
oleracea var.
italica cultivar
Lucky
ambient and 1.5x ambient O3;
for hourly ozone concentrations
during experimental ozone
treatment, see Kubiske and
Foss (2015)
Under field conditions, female P. brassicae laid significantly
more eggs on undamaged plants near other undamaged
plants (cr-VOC) than on undamaged plants exposed to
volatiles from nearby damaged plants (ir-VOC). In
laboratory choice tests, there were no significant
differences between oviposition on amb-cr-VOC plants and
amb-ir-VOC plants. Similarly, there were no differences in
oviposition between ozo-cr-VOC and ozo-ir-VOC plants.
Unlike in the laboratory, more eggs were laid on amb-cr-
VOC plants than on amb-ir-VOC plants. Elevated O3 had
no effect on oviposition preference. In laboratory tests
conducted in large cages, P. brassicae laid "marginally"
more eggs on amb-cr-VOC than on amb-ir-VOC plants.
Significantly more eggs were laid on amb-cr-VOC plants
than on ozo-cr-VOC plants and ozo-ir-VOC plants. P.
brassicae laid significantly more eggs on amb-cr-VOC than
on amb-ir-VOC plants. P. brassicae laid significantly more
eggs on ozo-cr-VOC than on ozo-ir-VOC plants.
Aqathokleous et al.
(2017)
Sapporo Experimental
Forest of Hokkaido
University (43.1°N,
141,333°E)
Insect:
Coleopteran leaf
beetle (Ageiastica
coeruiea)
Plant: Japanese
white birch (Betuia
piatyphyiia var.
japonica)
Ambient (27.5 ±11.6 ppb) and
elevated (61.5 ± 13 ppb)
In the "no-choice assay," there were no statistical
differences in the grazing behavior of adult beetles on
leaves collected from ambient and elevated O3. However,
in the "choice assay," adults grazed 6 times more on leaves
collected from elevated O3 than on leaves collected from
ambient O3. Second instar grazed leaf area (no-choice vs.
choice)—there were no statistical differences in the grazing
behavior of larvae on leaves collected from ambient or
elevated O3 in either assay.
8-84
-------
Table 8-14 (Continued): Ozone exposure and effects on herbivores.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Insects and Other Wildlife
Gilliland et al.
(2012)
OTC/mammal feeding
study Atmospheric
deposition site of the
school of forestry and
wildlife Sciences,
Auburn University,
Auburn, AL
Tall fescue (Lolium
a run din a),
dallisgrass and
(Paspalum
dilatatum),
common Bermuda
grass (Cynodon
dactylon), and
white clover
(Trifolium repens)
fed to New
Zealand white
rabbits
(Oryctolagus
cunicuius)
Six OTC chambers (three
chambers per treatment).
Grasses were exposed for
8 weeks. Mean monthly 12-h
ambient was 21-32 ppb (avg
peak conc. 49 ppb). Mean
monthly 2x ambient was 37 to
56 ppb (avg peak concentration
102 ppb)
Neutral detergent fiber and acid detergent fiber digestibility
was significantly lower for grasses grown under elevated
ozone. Digestible dry matter intake (DM intake
g/day * coefficient of apparent dry matter digestibility
[percentage]) was 5.5 g/day greater in rabbits fed forage
grown under NF (ambient) conditions compared with
rabbits offered forage grown under 2x O3. Decreased
digestibility of O3 forage was associated with increased
concentrations of phenolics and lower neutral detergent
fiber and acid detergent fiber digestibility.
Haves et al. (2016)
Grassland ozone
exposure experiments
across four upland
grassland types in
locations throughout
England
U.K. grassland Grasses collected from a range
species of O3 exposure experiments in
the U.K., with ozone
concentrations ranging from 17
to 93 ppb
The authors predicted a loss in liveweight gain of 3.6 to
4.4% for lambs in the U.K. from 2007 to 2020. With O3
conc. increase from 20 to 30 ppb, liveweight gain predicted
to decrease by 12%.
Uenoet al. (2016)
OTC; Inland Pampa
subregion, Buenos
Aires, Argentina
(34.583°S, 58.583°W)
Insect:
Rhopalosiphon
padi
Plant: Lolium
multiflorum
Fungal endophyte:
Epichloe occultans
120 ppb O3
Aphid population size was significantly higher in the I0W-O3
group in the presence of fungal endophyte, but not in the
high-03 group. The proportion of nymphs to adults was
significantly higher on fungal endophyte-free plants in the
I0W-O3 group where there were more adults on endophyte
symbiotic plants compared with endophyte-free plants.
Compared with I0W-O3, the proportion of nymphs was
increased in endophyte symbiotic plants, but the proportion
of nymphs was lower in the absence of endophyte. The
proportion of adult aphids was higher in endophyte-free
plants under high Osthan in symbiotic plants. The average
body weight of insects at both instars (adult and nymph),
between symbiotic and nonsymbiotic plants, was higher
under low O3.
8-85
-------
Table 8-14 (Continued): Ozone exposure and effects on herbivores.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Insects and Other Wildlife
Habeck and
Lindroth (2013)
Lab; plants: FACE site,
Rhinelander, Wl;
wild-caught voles
Mammal: Microtus
ochrogaster
Plant: Solidago
canadensis and
Taraxacum
officinale
Treatments for 1998-2007 were
ambient Os W126 = 2.9-8.8
ppm-h and elevated
O3 = 13.1-35.1 ppm-h. Ambient
air CO2 and elevated (560 ppm)
CO2 For hourly ozone
concentrations during
experimental ozone treatment,
see Kubiske and Foss (2015).
Fumigation with elevated CO2 or O3 had no effect on the
total amount of treatment diet consumed or the relative
proportion of plant species consumed. Weanling male voles
were unaffected by fumigation treatments, but female voles
grew 36% less when fed plants harvested from the
understory of O3 rings. Total plant consumption by male
voles was not related to any of the plant traits measured,
but the growth rate of males was negatively associated with
levels of ADF, ADL, and N. Total plant consumption by
female voles was negatively associated with ADL and
IVDMD, and positively associated with IVND and RP-N.
The growth rate of female voles was positively associated
with N, and negatively associated with CN, TNC, ADF,
ADF:N, and ADL:N, although all of these associations were
small.
Martinez-Ghersa et
al. (2017)
Mesocosm; University
of Buenos Aires,
Argentina (34.58°S,
58.48°W)
Populations of
agricultural weeds
(mostly Eurasian
annuals) from the
seed bank in
Corvallis, OR;
planted and grown
in Argentina and
interacting with the
Argentinian insect
community
Plants are descended from
populations exposed to 0
(charcoal-filtered), 90, or
120 ppb for 4 yr in OR. At the
end of the fourth season, 5 cm
top soil containing seed bank of
the community resulting from
4-yr exposure to episodic ozone
was removed from each
chamber
There was a plant species richness and arthropod diversity
linear relationship at 0 ppb historical O3, but no relationship
between plant species richness and arthropod diversity at
90 or 120 ppb historical O3 exposure. Insects: Historical O3
did not affect the richness, diversity, or evenness of the
arthropod community associated with descendant plant
community but did increase the relative abundance of
carnivore arthropods, while decreasing the relative
abundance of herbivore arthropods.
8-86
-------
Table 8-14 (Continued): Ozone exposure and effects on herbivores.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Insects and Other Wildlife
Khalina et al. (2015)
Greenhouse; plant
seeds: near
Wageningen
University, the
Netherlands;
Greenhouse:
University of Eastern
Finland, Kuipio,
Finland. Insect eggs:
Wageningen
University, the
Netherlands
Insect: Pieris
brassicae
Plant: Brassica
nigra
Ambient, 70 and 120 ppb O3.
Ambient ozone concentrations
fluctuated between 15 and
20 ppb; the other chambers had
elevated concentrations from
4:00 a.m. to 8:00 p.m. and a
basal concentration of 30 ppb
from 8:00 p.m. to 4:00 a.m.
Plants exposed for 5 days
Compared with ambient O3, mean larval mass was
significantly lower at 70 and 120 ppb O3 after 3 and 6 days
of treatment, but larval masses were not significantly
different when comparing 70 and 120 ppb O3. Compared
with plants grown in ambient O3, mean larval mass was
significantly lower in larvae developing on plants pretreated
with 70 and 120 ppb O3 after 3 and 6 days of feeding, but
larval masses were not significantly different when
comparing the masses of larvae reared on plants
pretreated with 70 or 120 ppb O3. Compared with pupae
feeding on plants exposed to 120 ppb 63, pupae
developing on plants pre-exposed to ambient O3 had a
significantly shorter larval period and achieved a
significantly greater larval mass. There were no significant
effects observed with 70 ppb O3. Larval mortality tended to
increase with increasing O3 concentration, but the effect
was not statistically significant for either concentration of O3
tested. In dual choice assays, larvae consumed
significantly more leaf material when feeding on plants
pretreated with 120 ppb Osthan plants pretreated with
ambient O3. There were no significant differences between
ambient O3 and 70 ppb O3 or between 70 and 120 ppb O3.
ADF = acid digestible fiber; ADL = acid digestible lignin; C = carbon; C:N = carbon-nitrogen ratio; C02 = carbon dioxide; FACE = free-air C02 enrichment; IVDMD = in vitro dry matter
digestibility; IVND = in vitro nitrogen digestibility; LC50 = median lethal concentration; LT50 = median lethal time; N = nitrogen; 03 = ozone; OTC = open-top chamber; ppb = parts per
billion; ppm = parts per million; RP-N = reducing power of protein-binding compounds on nitrogen digestibility; TNC = total nonstructural carbohydrates; VOC(s) = volatile organic
compound(s); VPSC(s) = volatile plant signaling compound(s); W126 = cumulative integrated exposure index with a sigmoidal weighting function.
8-87
-------
8.7
Plant-Insect Signaling
In the 2013 ISA there was no causality determination between ozone exposure and alteration of
plant-insect signaling. Plants signal to other community members through the emission of volatile plant
signaling compounds [VPSCs; Blande et al. (2014)1. Each signal emitted by plants has an atmospheric
lifetime and a unique chemical signature comprised of different ratios of individual hydrocarbons that are
susceptible to atmospheric oxidants like ozone (Yuan et al.. 2009; Wright et al.. 2005). Insects and other
fauna discriminate between chemical signals of different plants. Scent-mediated ecological interactions
include (1) host plant detection by herbivores, (2) attraction of pollinators and seed dispersers, and
(3) plant attraction of natural enemies of insect herbivores (Figure 8-7). Evidence for ozone-mediated
effects on plant-insect signaling are from studies that characterize scent plume emission/composition and
studies that assess insect response to altered signals in ozone-enriched environments (Table 8-15). Ozone
also interferes with VPSCs important in plant-plant interactions, such as emission of airborne signals to
alert neighboring plants of insect attack and attraction of predators and parasitoids of herbivores (Giron-
Calva et al.. 2016; Li et al.. 2016b; Li and Blande. 2015).
As described in the PECOS tool (Table 8-2). the scope for this section includes studies on any
continent that assess altered plant insect signaling in response to concentrations of ozone occurring in the
environment or experimental ozone concentrations within an order of magnitude of recent concentrations
(as described in Appendix 1). Ozone effects on plant volatile chemical emissions were not specifically
reviewed, rather identification of recent literature focused on plant-insect interactions, including plant
signaling in response to herbivory. The effect of elevated ozone on plant insect signaling involves
interactions with other biotic (species identity, lifestage) and abiotic (e.g., copollutants, elevated
temperature, temporal) factors (Jamieson et al.. 2017).
-------
Production of volatile piant signaling compounds
Emission of volatile plant signaling compounds
Plant
Herbivory
Ozone
Pollination
Insect pollinators
nsect herbivores
Insect parasitoids
Parasitized
insect herbivores
Volatile plant signaling compounds in air
nsect detection of volatile plantsignaling compounds
Figure 8-7 Conceptual model of ozone effects on volatile plant signaling
compounds and plant-insect signaling.
8.7.1 Emission and Chemical Composition of Volatile Plant Signaling
Compounds (VPSCs)
Studies in the 2013 Ozone ISA reported ozone alters the emission and chemical composition of
VPSCs (Blande et al.. 2010; Pinto et al.. 2007; Vuorinen et al.. 2004). Olfactory cues may travel shorter
distances in ozone-enriched environments, reducing the effectiveness of chemical communication
(Blande et al.. 2010; Mcfrederick et al.. 2008). Although not comprehensively reviewed for this ISA,
elevated tropospheric ozone has been shown to alter plant production and emission of VPSCs and well as
the atmospheric dispersion and lifespan of these compounds, thereby reducing the effectiveness of these
signals (Jucrgcns and Bischoff. 2017). Recent studies consistently show ozone effects on VPSCs, and that
the emission and degradation of individual chemical signal components vary.
• Ozone-induced VPSCs degradation: Elevated ozone (>50 ppb) degraded some plant VPSCs,
changing the scent composition and reducing scent dispersion, potentially affecting (1) size of the
scent plume, (2) ability of insects to detect the scent plume, and (3) time required to find the
source of the scent plume (Mofikova et al.. 2017; Fuentes et al.. 2016; Li et al.. 2016b; Farre-
Armengol et al.. 2015; Li and Blande. 2015). In an enclosed ozone reaction system, Farre-
Armengol et al. (2015) quantified degradation of several floral scent volatiles emitted by flowers
(Brassica nigra) along a distance gradient. Degraded volatiles were first detected at 1.5 m in
8-89
-------
80 ppb ozone, and the highest degradation levels (25-30%) were observed at 120 ppb ozone up to
4.5 m from the source (the farthest distance tested). Results of large eddy simulations show that
ozone levels greater than 60 ppb degrade VPSCs, thus altering the chemical composition of the
floral scent while increasing insect foraging times (Fucntcs et al.. 2016).
Plants emit VPSCs in response to herbivore feeding, and these signals are altered in combination
with elevated ozone. Depending on the plant species studied, elevated ozone either increased or had no
effect on VPSCs emissions in the presence of herbivory:
• Herbivory and ozone-induced VPSCs emissions: Tomato VPSC emissions increased 4.78-fold
and 5.66-fold following exposure to elevated ozone (72.2 ppb) and whitefly herbivory stress,
respectively. The combined effect of elevated ozone and whitefly herbivory further enhanced
VPSC emissions (Cui et al.. 2016b; Cui et al.. 2014). Elevated ozone (up to 120 ppb) did not alter
plant VPSC emissions by Brassica nigra (Khaling et al.. 2016V However, simultaneous exposure
to 120 ppb ozone and insect herbivore feeding stress for 24-hour increased emissions of several
VPSCs beyond levels detected following insect feeding alone. The effect was not detectable
72 hours post exposure or in the presence of 70 ppb ozone (Khaling et al.. 2016V Emissions of
VPSCs varied by month in Scots pine (Pinus sylvestris) subjected to herbivory stress and elevated
ozone (Ghimire et al.. 2017).
8.7.2 Pollinator Attraction and Plant Host Detection
Ozone effects on chemical signaling are evaluated in insect preference studies. Reduced detection
of VPSCs may decrease the efficacy of insect pollination of native plants and crops, an important
ecosystem service. This effect was demonstrated through a Lagrangian diffusion modeling study in the
2013 Ozone ISA in which the ability of pollinators to locate highly reactive VPSCs may have decreased
from kilometers during preindustrial times to <200 m at current ambient concentrations (Mcfrederick et
al.. 2008). One new empirical study tested detection response to elevated ozone in a pollinator species:
• Pollinator attraction: Under conditions of elevated ozone in experimental chambers, the
degradation of VPSCs resulted in bumble bees (Bombus terrestris) orienting significantly less
towards floral scent cues and exhibiting preference for artificial flowers closer to the ozone
source [120 vs. 0 ppb; Farre-Armengol et al. (2015)1.
As reported in the 2013 Ozone ISA, herbivorous insects use VPSCs to locate suitable host plants,
and ozone can alter these interactions (Blande et al.. 2010; Iriti and Faoro. 2009; Vuorinen et al.. 2004;
Jackson et al.. 1999; Cannon. 1990). In an early study on VPSCs, ozone-induced emissions from red
spruce pine needles were chosen less often than control needles by spruce budworm larvae
(Choristoneura fumiferana), resulting in reduced plant host detection (Cannon. 1990V Subsequent studies
showed that ozone can make a plant either more attractive or repellant to herbivores (Pinto et al.. 2010;
Jackson et al.. 1999). Decreased detection of VPSCs by plant-eating insects may interrupt the ability of
herbivores to locate plant hosts.
• Plant host detection by insect herbivores: In chamber studies, elevated ozone reduced the ability
of insect herbivores to find their plant hosts (Li et al.. 2016b; Fuentes et al.. 2013). Striped
8-90
-------
cucumber beetles (Acalymma vittatum) could not distinguish between clean air and air containing
floral volatiles when the ozone concentration exceeded 80 ppb (Fuentes et al.. 2013).
Diamondback moth (Plutella xylostella) larvae oriented significantly more towards teflon filters
exposed to nonozonated plant volatiles over filters exposed to plant volatiles mixed with elevated
ozone (Li et al.. 2016b). In addition, the larvae spent less time searching when placed on filters
exposed to plant volatiles mixed with ozone [0 vs. 100, but not 50 ppb; Li et al. (2016b) I. In
OTCs, both ozone and herbivory by whiteflies (Bemisia tabaci) increased emissions of tomato
plant (Solarium esculentum) VPSCs. Adult whiteflies preferred tomato plants exposed to ambient
ozone levels over tomato plants exposed to elevated ozone for feeding [37.3 vs. 72.2 ppb; Cui et
al. (2014)1.
8.7.3 Plant Attraction of Natural Enemies of Herbivores
Plant defense responses include emission of VPSCs to attract predators and parasitoids that target
the herbivores feeding on the plant. In studies reviewed in the 2013 Ozone ISA and new studies
parasitoid-host attraction is either reduced, enhanced, or unaffected by elevated ozone (Cui et al.. 2016b;
Khaling et al.. 2016; Cui et al.. 2014; Pinto et al.. 2008; Pinto et al.. 2007; Gate et al.. 1995). Altered plant
signaling to natural enemies of herbivores disrupts predator-prey trophic interactions.
• Effect of elevated ozone on parasitoid-host interactions: The parasitoid Cotesia glomerata did not
exhibit orientational bias for Brassica nigra plants exposed to elevated ozone [15-20 vs. 70 and
120 ppb; Khaling et al. (2016)1. The parasitoid Encarsia formosa preferred plants exposed to
elevated ozone over control plants [37.3 vs. 72.2 ppb; Cui et al. (2014)1. Searching efficiency and
the proportion of host larval fruit flies parasitized by Asobara tabida were reduced in the
presence of 100 ppb ozone (Gate et al.. 1995).
• Combined effects of elevated ozone and insect herbivory on parasitoid-host interactions: In field
plots of potted cabbage plants, the behavior of the parasitoid Cotesia plutella was unaffected by
elevated ozone [2x ambient; Pinto et al. (2008)1. In the absence of herbivory, the parasitoid C.
glomerata did not exhibit preference for control or ozone exposed plants. (Khaling et al.. 2016).
When compared with plants only exposed to ozone, the parasitoid oriented toward plants exposed
to 70 ppb ozone followed by herbivore feeding. The parasitoid oriented more towards plants
exposed to a combination of 120 ppb ozone and herbivory more than herbivore-stressed plants
that had not been exposed to ozone. However, in a wind tunnel assay, the strength of orientation
toward insect-damaged plants was significantly reduced by 70 ppb ozone, but not by 120 ppb
ozone (Khaling et al.. 2016). The parasitoid E. formosa preferred insect-damaged plants exposed
to elevated ozone over control plants [37.3 vs. 72.2 ppb; Cui et al. (2016b); Cui et al. (2014)1.
8.7.4 Summary and Causality Determination
In the 2013 Ozone ISA experimental and modeling studies reported altered insect-plant
interactions mediated through chemical signaling. New empirical research from laboratory, greenhouse,
OTC and FACE experiments expand the evidence for altered/degraded emissions of chemical signals
from plants and reduced detection of VPSCs by insects, including pollinators, in the presence of ozone
(Table 8-16). The interaction of ozone (>50 ppb) with VPSCs disrupts the production, emission,
8-91
-------
dispersion, and lifespan of these compounds. Numerous preference studies in insects show altered plant
host detection, reduced pollinator attraction, and shifts in plant host preference in the presence of
elevated, yet environmentally relevant, ozone concentrations. Plant defense mechanisms (i.e., attraction of
predators and parasitoids that target phytophagous insects) were either reduced, enhanced, or unaffected
by elevated ozone. Considering the available evidence (i.e., the 2013 Ozone ISA and more recent
research efforts) and recognizing uncertainties around how chemical signaling responses observed
in the laboratory translate to natural environments (Table 8-13). this ISA makes a new causality
determination that the body of evidence is sufficient to infer a likely to be causal relationship
between ozone exposure and alteration of plant insect signaling.
8-92
-------
Table 8-15 Summary of evidence for a likely to be causal relationship between ozone exposure and alteration of
plant-insect signaling.
Rationale for Causality Determination
Key Evidence
Key References
Multiple experimental and modeling studies Ozone disrupts the production,
by multiple research groups show direct emission, dispersion, and lifespan of
effects of ozone on VPSCs VPSCs
Vuorinen et al. (2004). Pinto et al. (2007). Mcfrederick et al. (2008)
(model), Blande et al. (2010). Cui et al. (2014). Li and Blande (2015).
Farre-Armenqol et al. (2015), Cui et al. (2016b), Fuentes et al. (2016)
(model), Khalinq et al. (2016). Li et al. (2016b). Mofikova et al. (2017).
Ghimire et al. (2017)
Multiple experimental studies by multiple
research groups show altered insect
response to VPSCs in presence of ozone
Altered plant host detection and insect
herbivory; reduced pollinator attraction
and altered parasitoid attraction by
plants
Cannon (1990). Gate et al. (1995). Fuentes et al. (2013). Cui et al. (2014).
Farre-Armenqol et al. (2015), Li et al. (2016b), Khalinq et al. (2016), Cui et
al. (2016b)
VPSC(s) = volatile plant signaling compound(s).
8-93
-------
Table 8-16
Ozone exposure and plant insect signaling.
Study
Study Type and
Location Study Species Ozone Exposure
Effects on Plant Insect Signaling
Li and Blande (2015)
Lab; Kuopio, Insect: Plutella zylostella Ambient (-10 ppb, 24 h/day),
Finland (diamondback moth) 80 ppb 5 days (reduced to
Plant: Brassica oleracea 30 PPb at ni9ht)for 5 daVs
(broccoli)
Mixing VPSCs emitted by herbivore-damaged plants with O3
resulted in complete degradation of some compounds and
partial degradation of others. However, in a few instances,
quantities of some VPSCs increased, suggesting that VPSCs
degrade into other VOCs.
Khalina et al. (2016)
Greenhouse; plant
seeds and insect
eggs—near
Wageningen
University, the
Netherlands
Greenhouse;
University of
Eastern Finland,
Kuipio, Finland
Insect: Pieris brassicae
(large white moth),
Cotesia glomerata
Plant: Brassica nigra
(black mustard)
Experiment #1: ambient
(15-20 ppb), 70, 120 ppb
from 4:00 a.m. to 8:00 p.m.
and a basal concentration of
30 ppb for the remaining 8 h
each day.
Experiment #2: VOC
emissions were sampled
from ambient, 70, 120 ppb
and herbivore feeding in
combination with ambient,
70, 120 ppb for 24 and 72 h
Plant emissions were not significantly altered by O3 exposure
alone, but herbivore-feeding stress (24 and 72 h) induced
emission of VPSCs. Simultaneous exposure to 120 ppb O3
and insect herbivore feeding stress for 24 h increased
emissions of several VPSCs beyond levels detected
following insect feeding alone. The effect was not detectable
72 h post exposure or in the presence of 70 ppb O3. The
parasitoid did not show a preference for control or O3
exposed plants. However, when compared to plants only
exposed to O3, the parasitoid oriented toward plants exposed
to 70 ppb O3 followed by herbivore feeding, but the same
effect was not observed at 120 ppm O3. In the absence of
herbivory, the parasitoid did not show a preference between
elevated O3 and clean-air-exposed plants. Finally, the
parasitoid oriented more towards plants exposed to 120 ppm
O3 and herbivory more than herbivore-stressed plants. The
strength of parasitoid orientation toward herbivore-damaged
plants over nondamaged plants was significantly affected by
70 ppb O3, but not 120 ppb O3.
Cui et al. (2014)
OTC;
Xiaotangshan
County, Beijing,
China (40.183°N,
116.4°E)
Insect: Bemisia tabaci
(whitefly), Encarsia
formosa
Plant: Solarium
esculentum (wild-type
tomato plants)
Ambient (37.3 ppb; avg value
from 9:00 a.m.-5:00 p.m.)
2x ambient (72.2 ppb; avg
value from 9:00 a.m. to
5:00 p.m.)
Exposure duration was
8 h/day for 24 days,
excluding 2 days due to rain
Elevated O3 levels increased VPSC emissions 4.85-fold in
the wild-type tomato plants. Whitefly herbivory increased the
total amount of plant VPSC emissions 5.12-fold. VPSC
emissions were greatest for the O3 + herbivory treatment. In
a behavioral assay, adult parasitoids preferred
insect-damaged plants exposed to elevated O3 to elevated
O3 over control plants.
8-94
-------
Table 8-16 (Continued): Ozone exposure and plant insect signaling.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Plant Insect Signaling
Li etal. (2016b)
Greenhouse;
Kuopio, Finland
Insect: Plutella xylostella
(diamondback moth)
Plant: Brassica oleracea
(cabbage) Brassica
oleracea (broccoli)
Experiment #1: In a Y
chamber bioassay, insect
herbivore was given the
choice between VPSCs from
healthy plants vs. clean air
(250 mL/min flow rate)
Experiment #2: In a Y
chamber bioassay, insect
herbivore was given the
choice between
VPSCs + clean air vs.
VPSCs + Os (50, 100 ppb for
5 min, flow rate 300 mL/min)
Experiment #3: Insect
herbivore preference was
evaluated through four
choices between herbivore-
induced cabbage VPSCs
mixed with 50 ppb O3 vs.
clean air, herbivore induced
cabbage VPSCs mixed with
100 ppb O3 vs. clean air,
constitutive cabbage VPSCs
mixed with 100 ppb O3 vs.
clean air, and herbivore-
induced broccoli VPSCs
mixed with 100 ppb O3 vs.
clean air
Herbivore-induced VPSCs at both 50 and 100 ppb O3 were
significantly degraded, rendering the relative proportions of
03-treated VPSC components different that the original
VPSCs. Significantly more insect larvae oriented towards
VPSCs from undamaged plants than charcoal-filtered air.
Significantly more insect larvae oriented towards
insect-infested plants than undamaged plants. Elevated O3
degraded some VOCs emitted from herbivore-damaged
plants; the effect appeared to be dependent on the
concentration of O3 present. Insect larvae preferred VPSCs
exposed to filtered air over those mixed with 50 and 100 ppm
O3; preference was lost when the air supply was passed
through an O3 scrubber prior to mixing with VPSCs. Insect
larvae preferred filters that had not been exposed to O3 over
filters exposed to VOCs that were mixed with O3. Larvae
spent significantly less time searching when placed on filters
exposed to VOCs mixed with O3 than on filters exposed to
VOCs mixed with clean air. In both instances, effects were
observed for constitutive and O3 induced VOC blends and
only in the 100 ppb O3 treatment.
Mofikova et al.
(2017)
FACE; Kuopio,
Finland
(62.895°N,
27.625°E)
Insect: Pieris xylostella
Plant: Brassica oleracea
(white cabbage)
Ambient: (avg 29 ppb),
elevated: (avg 42 ppb) from
8:00 a.m. to 10:00 p.m.
Compared to ambient O3, plants at 0.5 m distance from the
myrcene dispensers and exposed to elevated O3 had lower
myrcene emission rates. The effect was lost at the 1.5- and
3-m distances. Levels of myrcene were significantly lower at
the 0.5-m distance from the dispenser in plots exposed to
elevated O3. Myrcene was not detectable at the 1.5- and 3-m
distances in ambient or 42 ppb O3 plots.
8-95
-------
Table 8-16 (Continued): Ozone exposure and plant insect signaling.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Plant Insect Signaling
Station for Global (whitefly)
Cui et al. (2016b) OTC; Observation Insect: Bemisia tabaci Ambient (37.3 ppb; avg value Elevated O3 increased the total amount of plant VOC
from 9:00 a.m.-5:00 p.m.) for emissions 4.78-fold in the wild-type tomato plants. Whitefly
herbivory increased the total amount of plant VOC emissions
5.66-fold in the wild-type tomato plants. Production of VPSCs
was highest in the treatment of O3 + herbivory. In a
dual-choice Y-tube assay, adult parasitoids preferred
O3 + herbivory plants over all other treatments.
Change Biology,
Beijing China
(40.183°N,
116.4°E)
Plant: Solanum
lycopersicum (wild-type
tomato)
3 weeks, in both 2010 and
2011
2x ambient (72.2 ppb; avg
value from
9:00 a.m.-5:00 p.m.)
Exposure duration was
8 h/day for 24 days,
excluding 2 days due to rain
Ghimire et al. (2017)
Other; Kuopio,
Finland
(62.217°N,
27.583°E)
Insect: Acantholyda
posticalis (great
web-spinning pine
sawfly)
Plant: Pinus sylvestris
(Scots pine)
O3 exposure was 14 h/day,
7 days/week and calculated
as daily average from
8:00 a.m.-10:00 p.m.
2011 and 2012: 1.48 ambient
O3 concentration
2013: 1.56x ambient O3
concentration
Daily average ozone
concentrations (ppb)
computed from daytime
(8:00 a.m.-10:00 p.m.);
hourly mean values for both
ambient and elevated O3
exposures graphed in
Figure 1
Herbivore feeding alone significantly increased emissions of
some VPSCs. BVOC emissions were exponentially related to
the proportion of needles damaged by insect herbivores on
the 7th feeding day in June. In July, herbivory increased
emission rates of some VOCs from nondamaged branches,
in combination with elevated O3, emissions of MT-nx was
further elevated. In August, herbivory stress did not alter
emission of most VOC (post-feeding effect), except GLVs in
the treatment with ambient O3 and lower nitrogen. In
September, post-feeding effects of herbivory significantly
increased emission rates of MT-nx at elevated N level.
8-96
-------
Table 8-16 (Continued): Ozone exposure and plant insect signaling.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Plant Insect Signaling
Fuentes et al. (2013) Lab
Insect: Acalymma
vittatum (striped
cucumber beetle)
Plant: Cucurbita
foetidissima
Beetles exposed to two
choices in a Y chamber, air
flow 1 L/min. Duration of
exposure no longer than
5 min. Three types of choice
trials conducted: Filtered air
vs. filtered + O3 where
filtered was 0 ppb and
filtered + O3 was 20, 40, 60,
80, 100, or 120 ppb. Filtered
air vs. flower + 63 where
filtered air was 0 ppb and
flower + O3 was 0, 40, 80, or
120 ppb. Flower vs.
flower + O3 where flower was
0 ppb and flower + O3 was
20, 40, 60, 80, 100, or
120 ppb
Under all choice conditions, beetles were equally likely to
choose filtered air or filtered air + elevated O3. At 0 ppb O3,
beetles chose flower over filtered air 83% of the time. At
40 ppb O3, beetles chose flower + O3 over filtered air 70% of
the time. At 80 ppb O3, beetles chose flower + O3 63% of the
time (not significantly different from no preference). At
120 ppb O3, beetles were equally likely to pick filtered air as
flower + O3. At 20, 40, and 60 ppb O3, beetles were
statistically equally as likely to choose flower as flower + O3.
At 80, 100, and 120 ppb O3, beetles chose flowers over
flowers + O3 between 75-80% of the time.
Farre-Armenqol et
al. (2015)
Lab
Insect: Bombus
terrestris (buff-tail
bumble bee)
Plant: Brassica nigra
(black mustard)
Flower VOC emission
exposed to 0, 80, and
120 ppb. Bumble bees
exposed to three choices in a
cylindrical chamber, airflow
1 L/min. Duration of
exposure 10 min. Three
types of choice trials
conducted: Floral scent from
distance 0 at 0 ppb O3 vs.
clean air; floral scent from
distance 3 at 120 ppb O3 vs.
clean air; floral scent from
distance 0 at 120 ppb O3 vs.
floral scent from distance 3
at 120 ppb O3
The concentration of floral scent volatiles decreased with
increasing O3 concentration and distance from the floral
scent source. Degraded volatiles were first detected at 1.5 m
in 80 ppb O3 and the highest degradation levels (25-30%)
were observed at 120 ppb O3 up to 4.5 m from the source
(the farthest distance tested). Because not all floral scents
were degraded equally, O3 altered the ratio of compounds in
the scent blend.
Bumble bees preferred floral scent at distance 0 m over
scent-free, filtered air. Bumble bees showed no clear bias
when presented with floral scent at a distance of 3 m with
120 ppb O3 and filtered air. Bumble bees preferred floral
scent at distance 0 and 120 ppb O3 over floral scent at a
distance of 3 m. More bumble bees landed more on artificial
flowers associated with floral scent at distance 0 m with
0 ppb O3 than artificial flowers associated with filtered air.
More bumble bees landed on artificial flowers associated with
floral scent from distance 0 at 120 ppb O3 than artificial
flowers associated with floral scent from distance 3 m at
120 ppb O3.
8-97
-------
Table 8-16 (Continued): Ozone exposure and plant insect signaling.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Plant Insect Signaling
Fuentes et al. (2016) Other
Generic insects
Using large eddy
simulations, modeled
exposure to a range of O3
concentrations (0 to 120 ppb
vs. O3 [ppb on a per volume
basis])
Reactivity with >60 ppb O3 reduces the lifetime of individual
VPSCs by varying degrees. Degradation ofVPSCs by
elevated O3 reduces the distance of floral scent dispersion.
The composition of floral scent plumes changed following
exposure to elevated O3, changing the proportion of
individual VOCs, and in some instances, reducing levels of
some VOCs below the limit of analytical detection.
BVOC(s) = biogenic volatile organic compound(s); FACE = free-air C02 enrichment; GLV(s) = green leaf volatile(s); L/min = liters per minute; mL/min = milliliters per minute;
MT-nx = total nonoxygenated monoterpenes; N = nitrogen; 03 = ozone; ppb = parts per billion; ppm = parts per million; VOC(s) = volatile organic compound(s); VPSC(s) = volatile
plant signaling compound(s).
8-98
-------
8.8 Carbon Cycling in Terrestrial Ecosystems: Primary
Productivity and Carbon Sequestration
In the 2013 Ozone ISA, the evidence was sufficient to conclude there is a causal relationship
between ozone exposure and reduced plant productivity, and a likely to be causal relationship between
ozone exposure and decreased terrestrial carbon sequestration (U.S. EPA. 2013). At the time of the 2013
Ozone ISA, there was evidence from long-term ecosystem-scale ozone fumigation experiments in forests,
grasslands, and agricultural systems that ozone exposure could decrease plant productivity. Similarly, the
2013 Ozone ISA included findings from five models that incorporated the negative effects of ozone on
leaf-level photosynthesis and plant growth into carbon-cycling models, which were then used to create
regional-, national-, or global-scale estimates of the effect of ozone pollution on terrestrial plant
productivity and carbon sequestration. Although the models and experiments varied widely in scale and
scope, there were consistent observations of decreased plant productivity and a smaller flux of CO2 into
ecosystems.
For the current review, the scope is based on causality determinations in the 2013 Ozone ISA. As
described in the PECOS (Table 8-2). the body of reviewed literature is different for the ecosystem
productivity and for carbon sequestration. For ecosystem productivity, which was causal in the 2013
Ozone ISA, this synthesis will only include studies conducted in North America at ozone concentrations
occurring in the environment or experimental ozone concentrations within an order of magnitude of
recent concentrations (as described in Appendix 1). recognizing that there is a substantial body of
research conducted in other countries that is out of scope of the current review. For studies of ozone
effects on carbon sequestration, given the relatively small number of studies and the likely to be causal
determination from the 2013 Ozone ISA, research from other countries examining ecosystems that were
analogous to those in the U.S. were considered and included. Studies that were reviewed for this ISA
include research that integrates ozone into carbon-cycling models at ecosystem to global scales and
experiments using OTC and FACE systems to expose plants to ozone.
8.8.1 Terrestrial Primary Productivity
The terrestrial carbon cycle integrates processes at a variety of scales, ranging from organelles to
individuals to biomes (Chapin et al.. 2002). Gross primary productivity (GPP), which is the influx CO2
from the atmosphere via photosynthesis at the ecosystem scale, is fundamental to global carbon cycling.
However, because photosynthesis occurs simultaneously with autotrophic respiration, GPP is rarely
measured directly, and estimates are most often derived from whole-ecosystem carbon flux measurements
(eddy covariance) or models that scale measurements of leaf-level processes to ecosystems (Chapin et al..
2002). Researchers have used process models to quantify at the ecosystem scale the change in GPP
8-99
-------
resulting from the widely documented decreases in photosynthesis at the leaf-scale caused by ozone
exposure.
Since the 2013 Ozone ISA, two new studies have reported on the effects of ozone on gross
primary productivity (Table 8-17):
• Working in three ecosystems with Mediterranean-type climates, Fares et al. (2013) conducted a
statistical analysis of data from eddy covariance flux towers to quantify the effect of ozone on
carbon assimilation (GPP). In California, ozone decreased carbon assimilation by 12% in a Pinus
ponderosa forest in the Sierra Nevada and by 19% in an orange (Citrus sinensis) grove in the
Central Valley; damage was greater in the orange grove because of higher average ozone
concentrations and because irrigation supported higher stomatal conductance during periods of
the day when ozone concentrations peaked. At the third site in Italy, ozone concentrations were
lower and there was no detectable effect on GPP.
• Yue and Unger (2014) adopted the same ozone-damage thresholds and sensitivity coefficients
that were used in the calibration of the MOSES model (described in the 2013 Ozone ISA) for use
in the Yale Interactive Terrestrial Biosphere Model (YIBs), a biophysical vegetation model.
Overall, inclusion of ozone damage improved the ability of the YIBs model to predict site-level
GPP measured at eddy covariance flux towers at 40 sites in the U.S. and Canada. Decreases in
GPP as a result of ozone ranged from 1 to 14% and were greatest at sites showing both high
stomatal conductance and high growing season ozone concentrations. Modeled across the U.S.,
ozone decreased GPP by 2-5%, with stronger effects in the eastern U.S. (east of 95°W longitude,
4-8% decrease, with localized decreases of 11-17%) because of both higher stomatal
conductance and higher ozone concentrations (Yue and Unger. 2014).
Carbon assimilated into plant tissue via photosynthesis is either respired or contributes to net
primary productivity (NPP), which is often measured as the rate of plant biomass accumulation. Estimates
of the effects of ozone on NPP have been derived from field experiments, such as Aspen FACE or
SoyFACE, as well as from C-cycling models. While much of the research published since the 2013
Ozone ISA is confirmatory, other work has provided new mechanistic insight into the effects of ozone on
NPP.
• Zak etal. (2011) and Talhelm et al. (2014) quantified NPP at the Aspen FACE experiment, which
exposed tree communities composed of aspen (Populus tremuloides), aspen-birch (Betula
papyrifera), or aspen-maple (Acer saccharum) planted 1 year before the experiment to elevated
ozone (ambient W126 2.1-8.8 ppm-hour, elevated 12.7-35.1 ppm-hour) from 1998 to 2008.
Although elevated ozone decreased cumulative NPP during the experiment by 10%, the effect of
ozone on annual NPP gradually disappeared over the last 7 years of the experiment such that
ozone had no significant effect on NPP during the last several years of the experiment (Talhelm et
al.. 2014; Zak etal.. 2011). Zak etal. (2011) attributed the disappearance of the ozone effect on
NPP to compensatory growth by ozone-tolerant individuals and species. Elevated ozone
exposures were also much lower in the last 3 years of the experiment (W126 avg of
13.5 ppm-hour) compared to the first 8 years of the experiment [W126 avg of27.4 ppm-hour;
Kubiske and Foss (2015)1. Further, Talhelm et al. (2014) used an empirical model to attribute the
disappearing ozone effect on NPP to canopy dynamics: ozone had a persistent negative effect on
leaf biomass and canopy N (Talhelm et al.. 2014) that created initial declines in NPP, but the
marginal effect of the decrease in canopy N on NPP declined as canopy leaf area increased
through time in the developing stand. Based on analysis of foliar insect herbivory patterns over
8-100
-------
the last 3 years of the experiment, Couture et al. (2015) suggested reduced canopy damage (16%
lower under elevated ozone) as another mechanism that contributed to the limited effect of ozone
on NPP. Under ambient conditions, foliar herbivory decreased NPP by approximately 2%, but
this effect was 23% smaller under elevated ozone.
• At the SoyFACE experiment in Illinois, Oikawa and Ainsworth (2016) studied soybean (Glycine
max) plant growth, photosynthesis, canopy N (g per plant), and canopy development across
treatments ranging from 37 to 116 ppb of ozone (9-hour means). Increasing exposure to ozone
decreased canopy leaf area, canopy N, and aboveground plant mass. Each 10 ppb increase in
ozone decreased whole-canopy net photosynthesis by 10%.
• In Finland, Kasurinen et al. (2012) conducted a free-air ozone fumigation (elevated: 30 ppb;
ambient: 24 ppb avg for about 5 months) experiment using four genotypes of birch (Betula
pendula) grown in pots for two growing seasons. There were no significant ozone effects
(p > 0.2) on tree biomass in a midsummer harvest, but in sampling in autumn after the onset of
leaf senescence, elevated ozone decreased overall leaf biomass and decreased stem biomass in
two of the four birch genotypes included in the experiment.
• The FACE site in Kranzberg Forest, Germany examined ozone and CO2 effects on Fagus
sylvatica (European beech) and Picea abies (Norway spruce). In beech trees under elevated ozone
treatments there was a significant decrease in the allocation of 13CC>2-labeled C to the stems
(60%) and marginally significant increase in coarse root respiration. In spruce, flux of
photosynthates to stem and coarse root respiration was slightly stimulated under elevated ozone.
Together, these new observations provide further evidence that ozone can decrease plant growth
and NPP, but also help give a more nuanced understanding of how these effects vary among genotypes,
species, communities, and environmental conditions. Models pair experimental observations of
dose-response relationships for photosynthesis or growth with estimates of ozone exposure to estimate the
effects on NPP at ecosystem or regional scales (see Table 9-2 in the 2013 Ozone ISA for a review).
Because resources to conduct ecosystem-scale experiments are limited, these models provide important
estimates of the consequences of ozone exposure for productivity and carbon cycling at larger temporal
and spatial scales. New studies at larger scales provide the following insights:
• Similar to the Aspen FACE experiment, the effects of ozone on forest productivity were also
dynamic in an analysis of long-term (500 years) forest productivity created by Wang et al. (2016)
using the University of Virginia Forest Model Enhanced (UVAFME). In contrast to the
physiological and ecosystem process models that have been widely used to model the effects of
ozone on plant productivity, UVAFME is a gap model, which tracks the growth and survival of
individual trees and species within a stand. Wang et al. (2016) applied UVAFME to a diverse
southeastern U.S. broadleaf forest (32 species) using a scenario that assigned individual species
sensitivities of 0, 10, and 20% growth reductions in response to ozone exposure. Ozone decreased
forest productivity and biomass during the first 100 years, but then had a neutral or positive effect
as more ozone-tolerant species grew more rapidly in response to a decrease in competition from
more ozone-sensitive species.
• Gustafson et al. (2013) integrated the tree growth results from the Aspen FACE experiment into
the LANDIS-II model, first in a scenario that modeled growth of the experimental stands for
180 years, then in a scenario that applied the results of Aspen FACE to a landscape simulation of
forest cover and productivity. In both simulations, ozone favored both birch and maple relative to
aspen, which was a function of both species' differences in ozone sensitivity, as well as greater
longevity of these species.
8-101
-------
• Ozone decreased stand-level NPP for three tree species (Pinus sylvestris, Picea abies, Betula
pendula) growing in stands modeled with the 3-PG model representing six different geographic
zones of Sweden, with effects ranging from 1.4 to 15.5% among species and climate scenarios
(Subramanian et al.. 2015).
• Based on empirical relationships between ozone exposure and tree growth derived primarily from
the results of OTC experiments, de Vries et al. (2017) used a forest productivity model (EUgrow)
that was coupled with the soil biogeochemical process model VSD to estimate that ozone
decreased forest biomass across Europe by 4% over the simulation period of 1900 to 2005.
• Application of the landscape ecosystem process model, the Dynamic Land Ecosystem Model
(DLEM), to the southeastern U.S. (Texas to Virginia, 1987-2007) by Tian et al. (2012) and to
agricultural ecosystems across China (1980-2005) by Ren et al. (2012) produced estimates that
ozone exposure decreased NPP by 3 and 10.5%, respectively. Consistent with DLEM simulations
reviewed in the 2013 Ozone ISA, Tian et al. (2012) estimated larger effects on broadleaf trees
(-3%) and crops (-7%) than on conifers (-0.5%).
• Using exposure-response relationships between W126 and plant growth published in the Welfare
Risk and Exposure Assessment for Ozone (U.S. EPA. 2014) for 4 crops and 11 trees, Capps et al.
(2016) examined the potential increases in crop and tree productivity that might result from
regulations intended to limit greenhouse gas emissions from power plants in the U.S. Increases in
productivity resulting from emissions controls were a function of the geographic distribution of
both the plant species and the electric generating stations, as well as the physiological sensitivity
of the plant species.
Results from the Aspen FACE experiment and the model simulation conducted by Wang et al.
(2016) both suggest that the effects on ozone on NPP could be dynamic and temporary. However, the
results of these experiments contrast with results from an 8-year FACE ozone experiment conducted in a
60-year-old beech (Fagus sylvatica)—spruce (Picea abies) forest in Germany (Matvssek et al.. 2010).
Although spruce growth increased under elevated ozone in this experiment, this increase amounted to
only 5% of the lost stem volume of beech under elevated ozone. Thus, there are apparent limits in some
systems to the extent that increased growth of ozone-insensitive species can compensate for decreased
productivity of ozone-sensitive species. More broadly, the extent to which ozone affects terrestrial
productivity will depend on more than just community composition, but other factors, which both directly
influence NPP (i.e., availability of N and water) and modify the effect of ozone on plant growth (see
Section 8.1.2: Modifying Factors).
8.8.2 Soil Carbon
Carbon in the soil can be bound in organisms (plant roots, microbial biomass, invertebrates) or
bound in organic compounds within soil particles or aggregates. In some terrestrial ecosystems, including
Aspen FACE's soils contain more carbon than is contained in the total plant biomass (Talhelm et al..
2014; Pregitzer and Euskirchen. 2004). Different forms of C within the soil have residence times ranging
from decades to centuries (Schmidt et al.. 2011). and soil C pools tend to respond more slowly than plant
pools to environmental change (Tian et al.. 2012). Ozone can alter terrestrial C storage through its effects
8-102
-------
on plant biomass and NPP (Section 8.3 and Section 8.8.3). as well as through its effects on C in soils
(Section 8.9.2V The experimental observations reviewed in Section 8.9.2 and in the 2013 Ozone ISA did
not find a direct link between ozone, NPP, and soil C pools. Thus, although Talhelm et al. (2014)
observed that ozone decreased soil C, the link between soil C and ozone may yet turn out to be as
complex as that between soil C and elevated CO2 (Terrer etal.. 2018; van Groenigen et al.. 2014).
8.8.3 Terrestrial Carbon Sequestration
Terrestrial carbon sequestration is the sum of C contained within biomass and soils within a
defined ecosystem, typically quantified on a multiyear scale (Koerner. 2006; Chapin et al.. 2002). As in
the 2013 Ozone ISA, most assessments of the effects of ozone on terrestrial C sequestration are from
model simulations. However, an assessment of the effect of ozone on ecosystem C content at the Aspen
FACE experiment was published in 2014 (Table 8-17).
• At the conclusion of the Aspen FACE experiment after 11 years of fumigation, Talhelm et al.
(2014) observed that elevated ozone decreased ecosystem C content (plant biomass, litter, and
soil C to 1 m in depth) by 9%. Total tree biomass C was 15% lower under elevated ozone, with
decreased woody biomass accounting for nearly all (98%) of the effect on tree biomass. With the
exception of surface soil C, no other individual pool of C was significantly affected by ozone.
The total pool of plant and litter C was closely related to cumulative NPP, but under elevated
ozone, the pool of plant and litter carbon within the aspen-only forest community was
significantly smaller than expected based on NPP, meaning that the biomass C produced under
elevated ozone was more quickly returned to the atmosphere.
Several new model simulations provide further support for regional- and global-scale decreases in
terrestrial C sequestration as a result of ozone pollution.
• Kvalcvag and Myhre (2013) used the Community Land Model (CLM), a terrestrial earth systems
model, to understand the effect of tropospheric ozone pollution on global terrestrial C
sequestration from 1900-2004. Here, a model scenario that included coupling between C and N
cycling produced a lower estimate of the negative effect of ozone on global terrestrial C
sequestration (8-26 Pg C/year) than a method comparable to a previous assessment [31-83 Pg
C/year; Sitch et al. (2007)1. However, this decrease in terrestrial C sequestration was still
estimated to have contributed up to 10% of the total increase in atmospheric CO2 that occurred
between 1900 and 2004 (Kvalcvag and Myhre. 2013).
• Tian et al. (2012) applied DLEM to the southeastern U.S. (Texas to Virginia) for the period of
1895-2007. As a single factor, ozone decreased overall C storage 2%, with larger effects on
broadleaf forests (-5%) and croplands (-5%) than on conifer forests (-0.3%), paralleling changes
in plant growth.
• de Vries et al. (2017) created a forest productivity model (EUgrow) that was coupled with the soil
biogeochemical process model VSD to predict the effects of ozone pollution and other
environmental factors on forest C pools in Europe. Ozone decreased forest carbon sequestration
by approximately 6%.
8-103
-------
• Ren et al. (2012) applied the agricultural module of DLEM to understand changes in soil C
storage in Chinese agricultural land caused by ozone and other environmental factors from 1980
to 2005. In this study, ozone decreased the rate of soil C sequestration by 12.6%.
The results from the Aspen FACE experiment and the model simulations provide further
evidence that ozone can decrease ecosystem C sequestration. Although the decreases in NPP were
temporary in the Aspen FACE experiment and UVAFME simulation, the 10% decrease in cumulative
NPP at Aspen FACE was associated with a 9% decrease in ecosystem C storage (Talhclm et al.. 2014).
The observed changes in NPP and ecosystem C storage at Aspen FACE are in part a demonstration of the
influence of stand- or ecosystem-development processes on C cycling. As stands age and develop, they
acquire structural features that increase ecosystem carbon storage, such as larger pools of coarse woody
debris and larger soil organic horizons (Prcgitzcr and Euskirchen. 2004). At Aspen FACE, elevated ozone
slowed stand development (Talhclm et al.. 2014; Talhelm et al.. 2012). At the landscape orbiome scale, C
storage is controlled by the demography of individuals and stands, with landscapes comprised of stands at
varying points of development following natural and anthropogenic disturbances (Koerner. 2006). Thus,
without a concomitant slowing of disturbances rates and landscape stand turnover, even temporary
decreases in NPP caused by ozone may be meaningful for biome-scale carbon sequestration because
stands at any given time since disturbance will contain less carbon.
8.8.4 Summary and Causality Determinations
Evidence on the effect of ozone exposure on ecosystem productivity comes from many different
experiments with different study designs (open-top chamber experiments, long-term ecosystem
manipulation chamberless exposure experiments such as Aspen FACE, SoyFACE, FinnishFace) in a
variety of ecosystems and models (including empirical models using eddy covariance measures, forest
productivity models parameterized with empirical physiological and tree life history data, and various
well-studied ecosystem models and scenario analysis). New information is consistent with the
conclusions of the 2013 Ozone ISA that the body of evidence is sufficient to infer a causal
relationship between ozone exposure and reduced ecosystem productivity.
The relationship between ozone exposure and terrestrial C sequestration is difficult to measure at
the landscape scale. Most of the evidence regarding this relationship is from model simulations, although
this endpoint was also examined in a long-term manipulative chamberless ecosystem experiment (Aspen
FACE). Experiments at Aspen FACE found ozone exposure caused a 10% decrease in cumulative NPP
and an associated 9% decrease in ecosystem C storage. Additional studies at this research site suggests
that the effects of ozone on plant productivity will be paralleled by large and meaningful decreases in soil
C, but the experimental observations reviewed did not find a direct link between ozone, NPP, and soil C
pools. It is likely that stand age and development and disturbance regimes are complicating factors in the
partitioning of ecosystem level effects of ozone exposure on carbon sequestration. Even with these
limitations, the results from the Aspen FACE experiment and the model simulations provide
8-104
-------
further evidence that is consistent with the conclusions of the 2013 Ozone ISA that the body of
evidence is sufficient to conclude that there is a likely to be causal relationship between ozone
exposure and reduced carbon sequestration in ecosystems.
8-105
-------
Table 8-17 Ozone exposure effects on productivity and carbon sequestration.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Productivity
Oikawa and
Ainsworth (2016)
FACE;
SoyFACE,
Champaign, IL
(40.04°N,
88.24°W)
Glycine max (soybean)
Plots were fumigated from
-10:00 a.m. to 7:00 p.m. daily.
Average [O3] from 10:00 a.m. to
7:00 p.m. during the growth
season was 36.9 ppb in the
control (ambient [O3]) plot, while
it was 39.8, 46.3, 53.9, 58.4,
71.0, 88.3, 94.2, and 115.7 ppb
in the eight elevated [O3] plots.
AOT40 (the hourly accumulated
exposure over a threshold of
40 ppb) of these plots were 3.3,
3.8, 9.0, 16.8, 21.0, 31.4, 47.2,
52.9, and 67.4 ppm-h,
respectively
This study scales decreased photosynthesis due to O3 from
the leaf to the canopy using a model dividing leaf canopy into
horizontal layers and within each layer estimates light
interception by the leaves. Leaf area and leaf nitrogen (N)
per plant decreased with increasing (O3; Figure 1a and b,
respectively), as did leaf canopy-level photosynthesis
(Figure 4a).
Talhelmetal. (2012)
FACE; Aspen
FACE,
Rhinelander,
(45.675°N,
89.625°W)
Betula papyrifera
(paper birch), Acer
Wl saccharum (sugar
maple), Populus
tremuloides (five
genotypes of quaking
aspen)
Fumigation 1998-2008 during
daylight hours of the growing
season. Ambient O3 W126
2.1-8.8 ppm-h and elevated
12.7-35.1 ppm-h; elevated CO2
515-540 ppm, ambient avg
374. For hourly ozone
concentrations during
experimental ozone treatment,
see Kubiske and Foss (2015)
Exposure to O3 (+O3 and +O3+CO2 vs. ambient and +CO2)
decreased leaf mass by 13%, decreased leaf area by 18%,
and decreased N mass by 16%.
8-106
-------
Table 8-17 (Continued): Ozone exposure effects on productivity and carbon sequestration.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Productivity
Talhelm et al. (2014)
FACE; Betula papyrifera
Rhinelander, Wl (paper birch), Acer
saccharum (sugar
maple), Populus
tremuloides (five
genotypes of quaking
aspen)
Fumigation 1998-2008 during
daylight hours of the growing
season. Ambient O3 W126
2.1-8.8 ppm-h and elevated
12.7-35.1 ppm-h; elevated CO2
515-540 ppm, ambient avg
374. For hourly ozone
concentrations during
experimental ozone treatment,
see Kubiske and Foss (2015)
O3 significantly affected ecosystem C (-9%); C in stems and
branches (-17%); C in 0-10 cm mineral soil (-11%); NPP
(-10%), although O3 effects on NPP disappeared during final
7 yr of study; NPPtree (-11%); canopy N (-21%).
O3 shifted fine roots toward soil surface.
Fares et al. (2013)
Gradient; Italy
and California
Pinus ponderosa,
Quercus spp. and P.
pinea, Citrus sinensis
Ambient data study over
several years at three sites.
Exposure duration varied based
on site from 1 to 7 yr. Ambient
concentrations were grouped as
follows: low (<50 ppb), medium
(>50 and <75 ppb), and high
(>75 ppb)
As much as 12-19% of GPP reduction was explained by
stomatal O3 deposition in ponderosa pines and citrus trees.
Stomatal O3 deposition was not found to limit GPP at the oak
site, likely due to higher stomatal resistance and low
exposure to ozone. Reduction in GPP was more related to
stomatal O3 deposition than to O3 concentration.
Gustafson et al.
(2013)
FACE; Aspen
FACE,
Rhinelander,
Wl; model
simulations into
the future
Betula papyrifera
(paper birch), Acer
saccharum (sugar
maple), Populus
tremuloides (four
genotypes of quaking
aspen)
Fumigation 1998-2008 during
daylight hours of the growing
season. Ambient O3 W126
2.1-8.8 ppm-h and elevated
12.7-35.1 ppm-h; elevated CO2
515-540 ppm, ambient avg
374. For hourly ozone
concentrations during
experimental ozone treatment,
see Kubiske and Foss (2015)
Measured total biomass was always highest under the
elevated CO2 treatment, lowest under the O3 treatment, and
the +CO2+O3 treatment was similar to the control. The O3
treatment significantly affected abundance of all taxa except
one aspen clone. By year 180, +CO2 doubled aboveground
productivity and +O3 decreased productivity by half. +O3
reduced forest biomass at the landscape scale.
8-107
-------
Table 8-17 (Continued): Ozone exposure effects on productivity and carbon sequestration.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Productivity
Couture et al. (2015)
FACE; Aspen
FACE,
Rhinelander,
(45.70°N,
89.50°W)
(1) Foliar herbivore
insects: individual
Wl species not monitored,
insect frass analyzed
(2) Plants: Populus
tremuloides (aspen,
genotypes 42 and 271)
and Betula papyrifera
(paper birch)
Treatments for 2006-2008
were ambient O3 W126 = 5.6,
4.9, 2.1 ppm-h and elevated
O3 = 14.6, 13.1, 12.7 ppm-h.
Ambient air CO2 and elevated
(560 ppm) CO2. For hourly
ozone concentrations during
experimental ozone treatment,
see Kubiske and Foss (2015)
Elevated O3 elicited a modest decrease in canopy damage.
Insect-mediated litter inputs (insect frass and greenfall) were
not impacted by elevated O3, but N flux from the canopy to
the soil decreased by 19%, and the ratio of foliar C:N
increased. Elevated O3 decreased the negative effects of
herbivory on ANPP.
Zaket al. (2011)
FACE; Aspen
FACE, near
Rhinelander, Wl
(45.70°N,
89.50°W)
Betula papyrifera
(paper birch), Acer
saccharum (sugar
maple), Populus
tremuloides (various
genotypes of quaking
aspen
Treatments for 2005-2008
were ambient O3 W126 = 7.3,
5.6, 4.9, 2.1 ppm-h and
elevated Os = 29.6, 14.6, 13.1,
12.7 ppm-h. Ambient air CO2
and elevated (560 ppm) CO2.
For hourly ozone
concentrations during
experimental ozone treatment,
see Kubiske and Foss (2015)
No effect of O3 on NPP was observed in the 10th through
12th yr of exposure. The authors speculate this was due to
compensatory growth of O3 tolerant genotypes and species.
Elevated ozone had no effect on forest floor mass, N content,
or 15N content.
Wang et al. (2016)
Model of
species-specific
biomass in a
small area with
"typical
temperate
deciduous
forest in the
southeast U.S."
32 tree species.
Dominant tree species
are Liriodendron
tulipifera (tulip poplar),
Acer rubrum (red
maple), Acer
saccharum (sugar
maple), Quercus alba
(white oak), Fagus
grandifolia (American
beech), and Quercus
muehlenbergii
(chinkapin oak). Other
species mostly pioneer
species
Each of the 32 species was
ranked based on O3
sensitivity-resistant,
intermediate, or sensitive.
Species-specific biomass
reductions due to O3 exposure
were 0, 10, and 20% for each of
the three categories,
respectively
As expected, O3 resistant species (white oak, beech)
dominate and sensitive species (tulip poplar, red maple)
decline over the 500-yr simulation. Overall forest biomass
and forest carbon storage do not decrease overtime under
high O3 conditions because growth of tolerant species is
enhanced due to decreased competition by the loss of O3
sensitive species.
8-108
-------
Table 8-17 (Continued): Ozone exposure effects on productivity and carbon sequestration.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Productivity
Tian etal. (2012)
Model;
southeastern
U.S. (30-37°N,
75-100°W;
includes
13 states)
10 different plant
functional types were
mapped across the
13-state region
AOT40 simulated from
1895-2007 using the data set
bv Felzer et al. (2004)
Terrestrial ecosystems were a C source from 1895-1950,
and a C sink from 1951-2007. Largest contributor to
increased sink was CO2, followed by N deposition. O3
reduced C storage by 0.58 Pg C and NPP by 2.5% during the
study period. O3 x climate and O3 x CO2 interactions
contributed to C gains in the study even though their
interactions reduced NPP; the increase O3 x CO2 interaction
was due to increased litter quantity and increasing soil C
storage. O3 greatest impact was in the NE portion of the
study area due to increased emissions from affected
broadleaf forest and cropland areas.
Yue and Unqer
(2014)
Model; U.S.
Eight primary
functional vegetation
types are identified,
seven natural
ecosystem types and
cropland. Model uses
either C3 or C4
photosynthesis
Hourly and daily max 8 h taken
from 2005 data set, NASA
Model-E2. Validated with
CASTNET and AIRDATA; plant
photosynthesis model used two
levels—high O3 sensitivity and
low O3 sensitivity
Total carbon uptake is estimated to be 4.43 Pg C during the
growing season across the U.S. Simulated summertime GPP
was 9.5 g C/m2-day in the eastern U.S., and 3.9 g C/m2-day
in the western U.S. when the models included the high O3
damage effect. Average GPP for the U.S. was 6.1 g
C/m2-day. O3 reduces GPP 4-8% in east, with 11-17%
decreases in "hot spots," and very small reductions in the
western U.S. due to stomatal limitations. Over all of U.S.,
total summer GPP reduced by 2-5% due to O3. A 25%
reduction in O3 is estimated to reduce GPP by only 2-4% in
the eastern U.S.
8-109
-------
Table 8-17 (Continued): Ozone exposure effects on productivity and carbon sequestration.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Productivity
Capps et al. (2016)
Model;
continental U.S.
(CONUS)
Zea mays (maize),
Gossypium sp.
(cotton), Solanum
tuberosum (potato),
Glycine max
(soybean), Populus
deltoides (eastern
cottonwood), Prunus
serotina (black cherry),
Populus tremuloides
(quaking aspen), Pinus
ponderosa (ponderosa
pine), Liriodendron
tulipifera (tulip poplar),
Pinus strobus (eastern
white pine), Pinus
virginiana (Virginia
pine), Acer rubrum
(red maple), Alnus
rubra (red alder)
Uses U.S. EPA-developed
CMAQ model to model
exposure values of W126 under
three regulatory scenarios
(summarized in Table 1) as well
as a reference (ambient) over
CONUS. Maximum W126 value
for the reference case with no
change in W126 = 56 ppm-h.
Study proposes three scenarios
in which maximum local
decrease in W126 is 1.3, 4, and
5.3%
At ambient ozone concentrations, production loss is greatest
for potatoes, soybean, and cotton (losses of 1.5 to 1.9),
eastern cottonwood and black cherry demonstrate noticeable
losses at ambient O3 concentrations, 32, and 10%,
respectively. Black cherry shows the greatest potential
productivity losses of 2,2101 of biomass per hectare with
twice the biomass loss potential of either eastern cottonwood
or ponderosa pine. The quaking aspen, tulip poplar, and
various pine species also respond to ozone with potential
productivity losses ranging from 0.3 to 1.9%.
Ren et al. (2012)
Model; five
different regions
throughout
China
Not specified
Two indices of O3 were
employed in each regional
simulation, both expressed as
AOT40 obtained from Felzer et
al. (2005) The "control" was a
constant level of O3, the
treatment simulation was based
on historical O3 levels in each
region for 1980 through 2005,
which showed dramatic
increases in O3 starting in 1995
(see Figure 2C)
O3 decreased NPP and SOC by 10.7 and 12.6%,
respectively, across all regions. O3 had an increasingly
negative impact over the study period. Land use change was
the dominant factor controlling temporal and spatial
variations in NPP and SOC. The combined contributions of
climate variability, CO2, N deposition, and O3 accounted for
less than 20% of changes in NPP and SOC. However,
sensitivity analyses indicated that simulated effects of O3
were doubled when combined with other climate factors.
8-110
-------
Table 8-17 (Continued): Ozone exposure effects on productivity and carbon sequestration.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Productivity
Betzelberaer et al.
(2012)
FACE;
SoyFACE,
Champaign, IL
(40.04°N,
88.24°W)
Seven cultivars of
Glycine max (soybean)
Soybeans in eight
20-m-diameter SoyFACE plots
with O3 concentrations for
8 h/day in two growing seasons
(2009, 2010): ambient, 40, 55,
70, 85, 110, 130, 160, 200 in
2009; and ambient, 55, 70, 85,
110, 130, 150, 170, 190 in
2010. 8-h, 24-h, and 1-h max
mean as well as AOT40 and
SUM06 for each plot in Table 2
An exposure-response for soybean was refined from
previous estimates using seven cultivars and a range of
target concentrations from ambient to 200 ppb/8 h. Harvest
index (partitioning of carbon into seeds) was reduced by 12%
over the range (for 2009 growing season only).
Pleiiel et al. (2014)
Secondary
analysis of
previously
published data;
dose-response
data from eight
countries and
three
continents;
analysis not
fully described
Triticum aestivum
(wheat)
Phytotoxic O3 Dose (POD6)
metric which is the stomatal O3
uptake above a threshold of
6 nmol/m2-s. Stomatal
conductance was estimated
from VPD, temperature, solar
irradiance, and phenology.
POD6 ranked on a relative
scale, with zero POD6 being set
to 1, meaning no effect; higher
POD6 ranked <1. Full details
not provided in this manuscript
Although O3 effects (vs. charcoal-filtered air) on aboveground
biomass and yield were correlated, the effect on yield was
larger than on aboveground biomass. Using the EMEP model
for the year 2000 to model POD6 O3 over Europe, O3 caused
a 9% reduction in biomass but a 14% reduction in yield.
Analysis suggests that O3 accounted for over 22.2 million
tonnes of lost biomass in 2000, while a similar analysis using
yield would result in a loss of 38.1 million tonnes,
overestimating the loss by 15.9 million tonnes.
8-111
-------
Table 8-17 (Continued): Ozone exposure effects on productivity and carbon sequestration.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Productivity
Cheng et al. (2011)
OTC; Lake
Wheeler
Experimental
Station, NC
(35.72°N,
78.67°W)
Triticum aestivum
(wheat)—Glycine max
(soybean) rotation: in
November to June, O3
tolerant soft red winter
wheat (Coker 9486),
and in June to
November, soybean
(multiple cultivars over
4-yr experiment)
Full factorial O3 and CO2
fumigation for 4 yr:
(1) charcoal-filtered control
(canopy height seasonal daily
12-h avg for June-November is
19.9 ppb O3; canopy height
seasonal daily 12 h avg for
November-June is 20.7 ppb
O3) with ambient CO2 (376 ppm
June-November and 388 ppm
November-June);
(2) elevated O3 (canopy height
seasonal daily 12-h avg for
June-November is 65.7 ppb
O3; canopy height seasonal
daily 12-h avg for
November-June is 49.8 ppb
Os);
(3) elevated CO2 (555 ppm
June-November and 547 ppm
November-June);
(4) elevated O3 and elevated
CO2, as described previously
Elevated O3 reduces C and N input to soils from senesced
soybean biomass by 12%. Elevated O3 had no effect on soil
C, soil N, or fungal and bacterial soil abundances or ratio
assessed by PLFA.
Ritteret al. (2011)
FACE;
Kranzberg
Forest,
Germany
(48.417°N,
11.65°E)
Fagus sylvatica
(European beech) and
Picea abies (Norway
spruce)
Ambient and 2x ambient O3
(maximum O3 concentrations
restricted to <150 ppb). Trees
exposed for 7 yr
In the 2x O3 treatment in beech trees, there was a significant
decrease in the allocation of 13CC>2-labeled C to the stems
(60%) and marginally significant increase in coarse root
respiration. In spruce, flux of photosynthates to stem and
coarse root respiration was slightly stimulated under elevated
O3.
8-112
-------
Table 8-17 (Continued): Ozone exposure effects on productivity and carbon sequestration.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Productivity
Kasurinen et al.
(2012)
FACE;
Ruohoniemi
FACE, Kuopio,
University of
Eastern Finland,
Finland
(62.88°N,
27.62°E)
Clones of four
genotypes from the
wild population of
Betula pendula (silver
birch), as well as their
associated mycorrhizal
community and fruiting
bodies of mycorrhizal
fungi Laccaria laccata
Factorial O3 by temperature
treatment: mean ambient O3 is
23.4 in 2007, 23.8 ppb in 2008
(AOT40 0.14 ppm-h in
2007,1.6 ppm-h in 2008); mean
elevated O3 is 28.1 ppb in 2007,
32.0 ppb in 2008 (AOT40
4.9 ppm-h in 2007, 9.0 ppm-h in
2008), fumigation is 800 to
2,200 daily in growing season;
temperature treatment is
ambient or elevated by infrared
rings
O3 marginally increases concentration of carbon fixed by
trees in the soil (based on 13C pulse-tracer, p < 0.1).
Subramanian et al.
(2015)
Model; six
zones in
Sweden
grouped
according to
similar O3 levels
within each
zone
Picea abies (Norway
spruce), Pinus
sylvestris (Scots pine),
and Betula sp. (birch)
AOT40 values were obtained
from European Monitoring and
Evaluation Program. Prehistoric
treatment = AOT40 of 0;
ambient treatment = AOT40
based on 4-yr avg per county,
Counties were grouped into six
regions with AOT40 values of
13.8, 13.0, 11.8, 10.5, 8.2, 3.7,
and 7.1 ppm-h; increased
treatment = AOT40 2x ambient
Ambient O3 reduced modeled Norway spruce NPP
4.3-14.8% compared to prehistoric treatment. At increased
O3 (2x ambient), reductions were 8.5-30%. Ambient O3
reduced modeled Scots pine NPP 4.5-15.5% compared with
prehistoric treatment and increased O3 reduced NPP
8.8-31.4%. Ambient O3 reduced modeled birch NPP
1.4-4.3%, and increased O3 reduced NPP 2.9-9.8%. When
all species combined, modeled NPP decreased 3.7-12.5% in
the ambient vs. prehistoric treatment. Total reductions in C
sequestration of 3.7-14.9% in Swedish forests are estimated
if current O3 levels double.
de Vries et al. (2017) Model; Europe
Main species include
Picea abies (Norway
spruce), Pinus
sylvestris (Scots pine),
other conifers (divided
into northern and
southern Europe),
Quercus sp. (oak),
Fagus sylvestris
(beech), Betula sp.
(birch), other
broadleaves (northern
and southern Europe)
O3 uptake estimated using
phytotoxic O3 dose (POD),
calculated using EMEP model.
Two O3 exposure relationships,
linear with total biomass, and
net annual increment (NAI). For
1900-2050, simulations,
comparison used a scaling
factor for O3 relative to the
reference O3 exposure in 2005.
Source of O3 data unclear
Simulated European average total C sequestration in both
forests and forest soils increased by 41% between 1950 and
2000 (mainly due to increased N and CO2), with an additional
17% increased C sequestration expected between 2000 and
2050 (due to increased CO2 and temperature). Effect of O3
on tree C sequestration was -4% over 150 yr simulation and
from 1900-2050 was -4.5 to -5% using linear O3 biomass
function and multiplicative model. Using net annual increment
function (NAI) for O3, the effect of O3 was about -8.5% on C
sequestration, regardless of model.
8-113
-------
Table 8-17 (Continued): Ozone exposure effects on productivity and carbon sequestration.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Productivity
Hofmockel et al.
(2011)
FACE; Aspen
FACE,
Rhinelander, Wl
(49.675°N,
89.625°E)
Populus tremuloides
(quaking aspen), Acer
saccharum (sugar
maple), and Betula
papyrifera (paper
birch)
Samples taken 2003, 2004 and
2007. Treatments for
1998-2007 were ambient O3
W126 = 2.9-8.8 ppm-h and
elevated O3 = 13.1-35.1 ppm-h.
Ambient air CO2 and elevated
(560 ppm) CCh.For hourly
ozone concentrations during
experimental ozone treatment,
see Kubiske and Foss (2015)
Elevated O3 reduced nitrogen in coarse particulate organic
matter (cPOM) and fine particulate organic matter (fPOM) by
14%, which increased C:N ratios in all soil size fractions
(coarse, fine, and mineral-associated organic matter) by
2-7%. Under elevated CO2, elevated O3 decreased storage
of newly fixed C in whole soil by 22%, while increasing the
storage of older C by 21% in cPOM and by 13% in fPOM.
Kvalevaq and Mvhre
(2013)
Model; global
Vegetation categorized
into broadleaf,
needle-leaf, shrub, C3
and C4 grasses
O3 profiles were obtained from
1900-2004 from Oslo-CTM2
chem transport model. Also
tested monthly average O3 from
MOZART chem transport
model. Plants grouped into O3
sensitivity groups (broadleaf
trees, needle-leaf trees, shrubs,
C3 and C4 grasses). Within
sensitivity groups, a low and
high simulation was run to
estimate lower and upper limits
of expected O3 impacts on
photosynthesis
Total ecosystem C was continuously decreased by O3
reductions in photosynthesis between 1900 and 2004. In
2004, O3 reduced total ecosystem C by 30 to 83 Pg C/yr in
the C only case, and by 8 to 26 Pg C/yr in the C-N coupling
example (which simulated N limitation). In 2004, the model
estimates that 3-8 ppm of CO2 is in the atmosphere due to
O3 effects on C sequestration.
13C = carbon-13 isotope; 15N = nitrogen-15, stable isotope of nitrogen; AIRDATA = (U.S. EPA) air quality data collected at outdoor monitors across the U.S.; ANPP = annual net
primary productivity; AOT40 = seasonal sum of the difference between an hourly concentration at the threshold value of 40 ppb, minus the threshold value of 40 ppb; C = carbon;
C3 = plants that use only the Calvin cycle for fixing the carbon dioxide from the air; C4 = plants that use the Hatch-Slack cycle for fixing the carbon dioxide from the air;
CASTNET = Clean Air Status and Trends Network; C02 = carbon dioxide; EMEP = European Monitoring and Evaluation Programme; FACE = free-air C02 enrichment; GPP = gross
primary productivity; N = nitrogen; nmol/m2 = nanomoles per meter squared; NPP = net primary productivity; 03 = ozone; PLFA = phospholipid fatty acid; Pg C = petagrams
(gigatonne) carbon; POD6 = phytotoxic ozone dose above a threshold of 6 nmol/m2/s; ppm = parts per million; SOC = soil organic carbon; SUM06 = seasonal sum of all hourly
average concentrations > 0.06 ppm; VPD = vapor pressure deficit; W126 = cumulative integrated exposure index with a sigmoidal weighting function.
8-114
-------
8.9 Soil Biogeochemistry
The 2013 Ozone ISA (U.S. EPA. 2013) concluded there is a causal relationship between O3
exposure and the alteration of belowground biogeochemical cycles (U.S. EPA. 2013). This causality
determination was based on the body of evidence known at that time. It has been documented since the
2006 Ozone AQCD (U.S. EPA. 2006) that while belowground roots and soil organisms are not exposed
directly to O3, belowground processes could be affected by O3 through alterations in the quality and
quantity of carbon (C) supply to the soils from photosynthates and litterfall (Andersen. 2003). although
few studies had been conducted at that time. The 2013 Ozone ISA (U.S. EPA. 2013) presented evidence
that O3 alters multiple belowground endpoints including root growth, soil food web structure, soil
decomposer activities, soil respiration, soil C turnover, soil water cycling, and soil nutrient cycling.
The scope for new evidence reviewed in this section limits studies to those conducted in North
America (Table 8-18). while recognizing that a substantial body of research has been conducted in other
countries, as described in the PECOS tool (Table 8-2). The endpoints reviewed in the 2013 Ozone ISA
(U.S. EPA. 2013) are not systematically reviewed in the current ISA, however, some new studies are
identified. The new evidence since the 2013 Ozone ISA (U.S. EPA. 2013) included in this assessment
confirms O3 affects soil decomposition (Section 8.9.1). soil carbon (Section 8.9.2). and soil nitrogen
(Section 8.9.3) and is summarized in the following section (Figure 8-8).
Some mechanisms by which O3 alters soil biogeochemistry are discussed in other sections of this
ISA. For example, soil biogeochemistry can be altered by ozone-induced changes in plant productivity
(Section 8.8.1) because changes in productivity often lead to decreases on C from leaves that form soil
litter. Ozone can also alter root biomass and/or distribution across the soil profile (Section 8.3). which can
alter the soil organisms that depend on roots as a primary source of carbon causing changes in soil
organism (1) abundance, (2) activity, or (3) community composition. Ozone-induced changes to soil
microbial communities and other root-associated communities of biota (Section 8.10.2) can alter carbon
cycling and nitrogen cycling belowground, which may alter emission rates of C and N from the soil to the
atmosphere, as well as C and N pools within the soil. Persistent or cumulative effects of O3 on soil C and
N pools can alter ecosystem C sequestration (Section 8.8.3) and soil fertility.
8-115
-------
COj. HP
C02, HjO
Altered stonatal function
Allocation of C
Litter production
and chemistry
retention
Altered species competition
Soil moisture
Run off
R bot sj'sym b i o n ts
Root, leaf litter
exudation
Nutrients
Soil foodweb
¦Bacteria
¦Fur>g
Micro & riarco hvenebrales
Organic matter
Soil physical &
chemical properties
Figure 8-8 Conceptual diagram of ozone effects on belowground processes
and biogeochemical cycles.
8.9.1 Decomposition
Soil decomposition is the breakdown and chemical transformation of senesced plant or animal
matter by consumers (e.g., bacteria, fungus, archaea, or invertebrates). Within the soil profile,
decomposition occurs most frequently in the leaf litter layer. Ozone-induced alteration of leaf chemistry
can affect the rate at which a soil organism decomposes the leaf. Leaf litter chemistry was not within the
scope of this review; however, it was reviewed in the 2006 AQCD (U.S. EPA. 2006) and 2013 Ozone
ISA (U.S. EPA. 2013). which documented that although the responses are often species- and
site-dependent, O3 tends to alter litter chemistry. Most of the studies in the 2013 Ozone ISA evaluated
forest trees and soils, with evidence often indicating mixed results. Some studies showed that ozone
8-116
-------
exposure decreases leaf litter nutrients (Liu et al.. 2007; Kasurinen et al.. 2006). increases leaf litter
nutrients (Rodcnkirchcn et al.. 2009; Parsons et al.. 2008; Kozovits et al.. 2005). or has no effect
(Baldantoni et al.. 2011; Rodenkirchen et al.. 2009). Similarly, one study showed ozone-exposure
increased leaf sugars, soluble phenolics, and fiber (Parsons et al.. 2008). while another showed no effect
(Kasurinen et al.. 2006). The 2013 Ozone ISA (U.S. EPA. 2013) also documented that O3 exposure via
litter nutrient alteration could be related to changes in decomposition rates, with some studies showing
slight decreases and others showing no effect. Likewise, O3 had mixed effects on the activity of
cellulose-degrading enzyme that is associated with decomposer organisms, with some studies showing
ozone-induced decreases and some studies showing no effect. Moreover, the 2013 Ozone ISA (U.S. EPA.
2013) documented mixed results on decomposition rates and associated metrics from O3 exposure, with
some studies showing slight reduction or increase and others showing no effect. Responses varied among
species, sites, and exposure lengths.
New research from U.S. ecosystems also shows mixed results. The endpoints evaluated include
the effects of O3 on soil decomposition rates relative to labile litter (insect frass and agricultural plant
residues) and recalcitrant litter (leaf litter and wood).
• Labile litter: Frass and crop residues are labile sources of nutrients for decomposer organisms in
soils. Two studies report on O3 effects on labile litter. A study at Aspen FACE found that
elevated O3 altered the N content, C:N, and condensed tannins of insect frass in trials of four
insect species that fed on aspen leaves (Couture and Lindroth. 2014). In the OTC soy-wheat
rotation at Lake Wheeler Experimental Station, NC, elevated O3 reduced C and N inputs from
soybean residues to soil by 12% (Cheng et al.. 2011). These studies document that ozone-induces
changes in labile litter N and C content.
• Recalcitrant litter: Elevated O3 had no effect on woody litter chemistry or initial decomposition
rates of Populus tremuloides logs or Betulapapyrifera logs (Ebanvenle et al.. 2016).
8.9.2 Soil Carbon
Soil carbon (C) is often a mix of inorganic and organic forms of C, the latter may be from living
and/or dead plant, animal, fungal, archaeal, and bacterial organisms. The effects of O3 on several aspects
of soil C have been investigated. This section includes soil respiration, roots, C formation, methane
emission, and perchlorate.
The 2006 Ozone AQCD (U.S. EPA. 2006) documented there was no consistent effect on soil
respiration. Ozone could increase or decrease soil respiration, depending on the approach and timing of
the measurements. The 2013 Ozone ISA (U.S. EPA. 2013) showed mixed results of O3 exposure on roots
(which contribute to soil respiration), and documented that long-term fumigation experiments, such as the
Aspen FACE, suggested that ecosystem response to O3 exposure can change overtime. Observations
made during the late exposure years can be inconsistent with those during the early years, highlighting the
8-117
-------
need for caution when assessing O3 effects based on short-term studies. New studies since the 2013
Ozone ISA (U.S. EPA. 2013) show no effect of elevated O3.
• Soil respiration in crops: New studies of CO2 emissions from agricultural soils at SoyFACE in
Illinois found no effect of elevated O3 on CO2 emissions, either at the site or measured in lab
incubations of collected soils (Dccock and Six. 2012; Decock et al.. 2012).
In the 2013 ISA (U.S. EPA. 2013) it was known that O3 could reduce the availability of
photosynthates for export to roots, and thus, indirectly increase root mortality and turnover rates. The
2013 Ozone ISA (U.S. EPA. 2013) found mixed effects of O3 on fine root biomass, with some studies
finding increases (Grebenc and Kraigher. 2007; Pregitzer et al.. 2006). and others finding no effect (King
et al.. 2001). New studies since the 2013 ISA (U.S. EPA. 2013) indicate there are ozone-induced effects
on root distribution.
• Root distribution in forests: In Aspen FACE, 9 years of O3 fumigation altered the distribution of
tree roots across the top 1 m of the soil profile in the aspen-birch community (Rhea and King.
2012). while 11 years of O3 fumigation shifted the distribution of tree fine roots within the
mineral soil profile towards the soil surface across all tree communities (Talhelm et al.. 2014).
There are effects of elevated O3 on tree root biomass in Aspen FACE, as well as in other tree and
herb species.
The 2006 Ozone AQCD (U.S. EPA. 2006) documented that O3 had the potential to alter soil C
formation; however, very few experiments directly measured changes in soil organic matter content under
O3 fumigation. Studies documented in the 2013 Ozone ISA (U.S. EPA. 2013) found O3 exposure resulted
in mixed effects, either reducing or having no effect on soil C formation. Several studies have been
published since the 2013 Ozone ISA (U.S. EPA. 2013) that indicate O3 decreases soil carbon in shallow
soils in some years and can differentially affect new and old carbon storage.
• Soil carbon in forests: In Aspen FACE, 11 years of O3 fumigation decreased soil carbon in the
top 10 cm of mineral soil by 11% (Talhelm et al.. 2014). Soils sampled at Aspen FACE in the
5th, 6th, and 10th years of fumigation found no changes in total soil C pools in the top 20 cm of
soil, but elevated O3 decreased soil storage of new carbon while increasing storage of older
carbon in particulate organic matter (Hofmockel et al.. 2011). indicating a slowing of
belowground C cycling which also affected N cycling (see Section 8.9.3).
• Soil carbon in crops: In an OTC soy-wheat rotation at Lake Wheeler Experimental Station, NC,
elevated O3 had no measurable effect on total soil organic C or extractable (K2SO4) soil C (Cheng
etal.. 2011). O3 effects on soil C can alter long-term carbon storage in soils and plant biomass;
for more on this topic, see Section 8.8.3 on terrestrial C sequestration.
No information was published in the 2006 O3 AQCD (U.S. EPA. 2006) or the 2013 Ozone ISA
(U.S. EPA. 2013) about root symbiont carbon endpoints (biomass or respiration). New information
published since 2013 indicates O3 had no effects.
• Root symbionts in forests: At Aspen FACE, elevated O3 had no effect on hyphal biomass
production, hyphal respiration, or sporocarp respiration by the mycorrhizal fungal symbionts
associated with tree roots (Andrew et al.. 2014). Effects of O3 on the community composition of
root-associated organisms are addressed in Section 8.10.2.
8-118
-------
The 2013 Ozone ISA (U.S. EPA. 2013) found O3 alters CH4 emissions (Toetetal.. 2011; Zheng
etal.. 2011; Morskv et al.. 2008). and dissolved organic carbon (DOC) in U.K. peatland fen pore water
(Jones et al.. 2009). There are no new studies conducted in U.S. ecosystems that addressed the effects of
O3 on soil CH4 emissions or DOC.
The 2013 Ozone ISA (U.S. EPA. 2013) did not address how O3 reacts directly with soil particles
and solutions. A recent study proposed that O3 would react with soil to form perchlorate, which can be
taken up by crop plants and could affect human consumers. In a greenhouse experiment, O3 exposure of
204 ppb increased soil perchlorate concentration, while 102 ppb O3 had no effect (Grantz et al.. 2014).
The study looked at leaf content and found no evidence of perchlorate in leaves.
8.9.3 Soil Nitrogen
O3 can alter the cycling of nitrogen in the soil via its direct effect on plants. Nitrogen is an
important element to plant life as it is often the limiting nutrient for most temperate ecosystems. Nitrogen
cycling may be measured by many different specific N pools or processes. The 2013 Ozone ISA (U.S.
EPA. 2013) documented mixed results of O3 effects on soil N pools and processes, with results indicating
no effect in meadow N biomass or potential nitrification and denitrification (Kanerva et al.. 2006). While
ozone was shown to increase N release from litter in a forest (Stoelken et al.. 2010). ozone decreased
gross N mineralization (Holmes et al.. 2006) at Aspen FACE and N release from soil litter (Liu et al..
2007). While in crops, O3 decreased soil mineral N content (Puiol Pereira et al.. 2011). In addition to
empirical results, a model simulation for O3 effects on N soil retention/stream flow showed that O3
exposure decreased N retention, increasing stream export (Hong et al.. 2006).
More recent studies from Aspen FACE and SoyFACE continue to find effects of O3 on N cycling
in soils by measuring endpoints of soil N, root N uptake, N transformations, and N emissions. These
results support the findings in the 2013 Ozone ISA (U.S. EPA. 2013). showing that in forests, O3 may
decrease soil N content in some studies but have no effect in others and that O3 did not affect forest root
uptake of N. In crops, ozone did not affect soil N in field studies and showed mixed results, depending on
the N chemical species, in lab studies.
• Soil N in forests: In Aspen FACE, elevated O3 increased C:N ratios of soil organic matter by
decreasing the N content of particulate organic matter (Hofmockel et al.. 2011). At the stand
level, elevated O3 in Aspen FACE did not change the amount of N stored in the litter layer on the
forest floor (Zak et al.. 2011).
• Soil N in crops: In the OTC soy-wheat rotation at Lake Wheeler Experimental Station, NC,
elevated O3 had no measurable effect on soil N (Cheng etal.. 2011). A set of studies conducted at
SoyFACE found no effects of elevated O3 on soil N or soil 15N (Decock et al.. 2012). although
elevated O3 significantly decreased surface soil ammonium in the field, and significantly
increased soil nitrate in both field and lab incubation studies (He et al.. 2014; Decock and Six.
2012).
8-119
-------
• Root N uptake in forests: Elevated O3 did not affect Aspen FACE stand uptake of a 15N tracer and
the incorporation of this 15N into leaves (Zak et al.. 2011). although there were differences
between aspen genotypes in N uptake (see terrestrial community).
• N transformations in crops: Elevated O3 in SoyFACE did not affect soil N transformation rates
measured by 15N tracers in incubations (Dccock and Six. 2012). but decreased the abundance of
multiple microbial N cycling genes in surface soils (He et al.. 2014). A model constructed using
SoyFACE natural abundance 15N values suggests that elevated O3 accelerated N cycling by
increasing both soybean belowground N allocation and N2 emissions from soil I Dccock et al.
(2012); see Figure 8-9 below].
• Nemissions from meadow: The 2013 ISA (U.S. EPA. 2013) found that elevated O3 emissions
decreased daily N2O emissions in a Finnish meadow (Kanerva et al.. 2007). At SoyFACE in
Illinois, elevated O3 did not alter N2O emissions, but a model suggested that O3 may affect N2
emissions (Decock et al.. 2012).
! Grain N
Biological
N-fixation
Plant N
Soil N
Total
gaseous N
loss
Note: Dashed lines indicate decreases, thin solid lines indicate no major change, and thick solid lines represent increases.
Source: Reprinted with permission from the publisher, adapted from Decock et al. (20121.
Figure 8-9 Conceptual diagram illustrating effects of elevated ozone
exposure on N-inputs and outputs. Elevated ozone increased soil
N, increased gaseous N loss, and reduced grain N.
8-120
-------
8.9.4
Summary and Causality Determination
The 2013 Ozone ISA (U.S. EPA. 2013) presented evidence that O3 was found to alter multiple
belowground endpoints including root growth, soil food web structure, soil decomposer activities, soil
respiration, soil C turnover, soil water cycling, and soil nutrient cycling. New evidence since the 2013
Ozone ISA (U.S. EPA. 2013) included in this assessment confirms O3 effects on soil decomposition
(Section 8.9.1). soil carbon (Section 8.9.2). and soil nitrogen (Section 8.9.3).
Decomposition of leaf litter in the soil may be altered by ozone-induced alteration of leaf
chemistry. Leaf litter chemistry was not within the scope of this review; however, it was reviewed in the
2006 AQCD (U.S. EPA. 2006) and 2013 Ozone ISA (U.S. EPA. 2013). The 2013 Ozone ISA (U.S. EPA.
2013) documented mixed results on ozone-exposure effects on leaf litter decomposition with some studies
showing slight reduction others showing no effect. Responses varied among species, sites, and exposure
lengths. New studies in the current review do not change these observations.
There are new studies on the effects of ozone on several endpoints associated with soil C: soil
respiration, root mortality, root symbionts and soil C formation. The 2006 Ozone AQCD (U.S. EPA.
2006) and the 2013 Ozone ISA (U.S. EPA. 2013) documented no consistent effect on soil respiration.
New studies since the 2013 ISA show no effect of elevated O3 on soil respiration. The 2013 ISA (U.S.
EPA. 2013) documented that ozone could increase root mortality and turnover rates by reducing the
availability of photosynthates for export to roots, while studies showed mixed effects of ozone on fine
root biomass, with some studies finding increases and others finding no effect. New studies since the
2013 ISA indicate an ozone-induced effect on a new endpoint: root distribution. Studies documented in
the 2013 Ozone ISA found ozone exposure resulted in mixed effects, either reducing or having no effect
on soil C formation. Several new studies indicate O3 decreases soil carbon in shallow forest soils. No
information was published in the 2006 Ozone AQCD (U.S. EPA. 2006) or the 2013 Ozone ISA (U.S.
EPA. 2013) about root symbiont carbon endpoints (biomass or respiration). New information published
since 2013 indicates ozone had no effects on carbon in root symbionts. Overall, these new findings
support the conclusions from the 2013 Ozone ISA (U.S. EPA. 2013) that there is no consistent effect of
ozone on soil respiration and soil carbon formation. New evidence indicates ozone has effects on root
distribution within the soil profile and no effect on carbon in root symbionts.
Ozone can alter the cycling of nitrogen in the soil via its direct effect on plants. Nitrogen is an
important element to plant life as it is often the limiting nutrient for most temperate ecosystems. Nitrogen
cycling may be measured by many different specific N pools or processes. The 2013 Ozone ISA (U.S.
EPA. 2013) documented mixed results of ozone effects on soil N pools and processes, with results
indicating no effect (NH4+ immobilization, gross nitrification, microbial biomass N and soil organic N),
increasing rates (N release from litter), or decreasing rates/concentrations (gross N mineralization, soil
mineral N content, and N release from soil litter). More recent studies from Aspen FACE and SoyFACE
continue to find effects of ozone on N cycling in soils by measuring endpoints of soil N, root N uptake,
N transformations, and N emissions. These results support the findings in the 2013 Ozone ISA (U.S.
8-121
-------
EPA. 2013) showing that in forests ozone may decrease soil N content in some studies, but have no effect
in others. Also, ozone did not affect forest root uptake of N. In crops, ozone did not affect soil N in field
studies and showed mixed results, depending on the N chemical species, in laboratory studies.
Overall, the evidence does not change the conclusions from the 2013 Ozone ISA (U.S. EPA.
2013). and therefore, suggests that ozone can alter soil biogeochemical cycling of carbon and nitrogen,
although the direction and magnitude of these changes often depends on the species, site, and time of
exposure. Currently, there does not appear to be a consistent exposure-response relationship. The body of
evidence is sufficient to conclude that there is a causal relationship between ozone exposure and the
alteration of belowground biogeochemical cycles.
8-122
-------
Table 8-18 Response of belowground processes and biogeochemical cycles to ozone exposure.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Soil Biogeochemistry Endpoints
Decock et al. (2012)
FACE; SoyFACE,
Illinois (40.056°N,
88.201°W)
(1) Soil collected from FACE
for lab incubations in 2008,
(2) field measurements of N2O
and CO2 emissions in 2005
and 2006 soybean seasons,
(3) soils collected from FACE
for natural abundance 15N
study in July 2006
Ambient and elevated O3 (target
concentration of 1.23x ambient
2002-2006, and 2x ambient
2007-2008), in factorial
combination with elevated CO2.
From 2001 to 2008, annual
ambient average tropospheric O3
concentrations ranged from 37.3
to 62.8 ppb
In lab soil incubations, six growing seasons
of elevated O3 did not affect soil N2O
emissions. Field measurements under 1.5x
O3 in 2005 and 2006 soybean seasons
found no effect of elevated O3 on CO2 or
N2O emissions. Soils sampled in 2006
showed that four growing seasons of
elevated O3 had no effect on soil N or soil
15N in soybean rhizosphere or bulk soil.
Models of soil 15N suggest that under
elevated O3, long-term soybean inputs of N
to rhizosphere and bulk soil increase, while
N losses from bulk soil increase (Figure 8-9
for conceptual diagram).
Decock and Six (2012)
Lab; SoyFACE,
Illinois (40.056°N,
88.201°W)
Study in soybean (Glycine
max) agroecosystem
Soils for the study collected from
soybean plots at the SoyFACE
agroecosystem with ambient and
elevated O3 (target concentration
of 1.23* ambient)
Elevated O3 significantly increased soil
NO3", and briefly increased soil NhU"1" early
in the incubation. Elevated O3 did not affect
mineral N transformation rates as
determined by 15N tracers and did not
affect potential CO2 or N2O emission rates.
8-123
-------
Table 8-18 (Continued): Response of belowground processes and biogeochemical cycles to ozone exposure.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Soil Biogeochemistry Endpoints
Heetal. (2014)
FACE; SoyFACE,
Illinois (40.056°N,
88.201°W)
Soil microbial community
under soybean (Glycine max)
Ambient (-37.9 ppb), elevated O3
(-61.3 ppb), soybeans also grown
under elevated CO2 (-550 ppm)
and elevated CO2 + elevated O3
In elevated O3 soybean crop surface soils,
abundance of niFH, narG, norB, and ureC
N-cycling genes were significantly
decreased. There were no significant
differences in the subsoil. For C cycling
genes in elevated O3 soil samples, most
were unchanged while fungal
arabinofuranosidase and endoglucanse
increased significantly and xylanase,
cellobiase and exochitinase decreased
significantly. Soil N was quantified and
NH4+ was significantly lower in the surface
soil and NO3 significantly higher in subsoil
of elevated O3 plots compared to ambient.
Paudel et al. (2016) Greenhouse; Parlier, Palmer amaranth (Amaranthus
CA palmeri)
Two runs of exposure 30 and Elevated O3 exposure and water stress had
35 days. 12-h means of 4, 59, and no effect on root growth. This weed species
114 ppb may have much more tolerance to elevated
O3 and moisture stress compared to crops
with which it competes.
8-124
-------
Table 8-18 (Continued): Response of belowground processes and biogeochemical cycles to ozone exposure.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Soil Biogeochemistry Endpoints
Cheng et al. (2011)
OTC; Lake Wheeler
Experimental
Station, NC
(35.72°N, 78.67°W)
Wheat-soybean rotation: in
November to June, O3 tolerant
soft red winter wheat (Coker
9486), and in June to
November, soybean (multiple
cultivars over 4-yr experiment)
Full factorial ozone and carbon
dioxide fumigation for 4 yr:
(1) charcoal-filtered control
(canopy height seasonal daily
12-h avg for June-November is
19.9 nL/L O3; canopy height
seasonal daily 12-h avg for
November-June is 20.7 nL/L O3)
with ambient CO2 (376 pL/L
June-November and 388 pL/L
November-June);
(2) elevated O3 (canopy height
seasonal daily 12-h avg for
June-November is 65.7 nL/L O3;
canopy height seasonal daily 12-h
avg for November-June is
49.8 nL/L O3);
(3) elevated CO2 (555 pL/L
June-November and 547 pL/L
November-June);
(4) elevated O3 and elevated CO2,
as described previously
Elevated O3 reduces C and N input to soils
from senesced soybean biomass by 12%.
Elevated O3 had no effect on soil C, soil N,
or fungal and bacterial soil abundances or
ratio assessed by PLFA.
8-125
-------
Table 8-18 (Continued): Response of belowground processes and biogeochemical cycles to ozone exposure.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Soil Biogeochemistry Endpoints
Rhea and King (2012)
FACE; Aspen FACE,
Rhinelander, Wl
(49.675°N,
89.625°E)
Quaking aspen (Populus
tremuloides), paper birch
(Betula papyrifera)
Treatments up until the 2005
(when root samples were taken):
ambient average W126 was
5.2 ppm-h and elevated O3 was
27.3 ppm-h. For hourly ozone
concentrations during
experimental ozone treatment,
see Kubiske and Foss (2015)
This study assessed fine root responses to
O3 at deeper soil depths than typically
studied. Fine root biomass in all-aspen
(AA) and aspen-birch (AB) plots fumigated
with ozone differed by community and soil
depth. Biomass increased with depth in the
AA (aspen clones) community by 10.2,
36.4, and 34.8% in the upper, middle, and
lower soil layer relative to the control. In the
AB (aspen-birch) community root biomass
decreased 16% in the shallow layer with a
small increase at the middle soil layer
resulting in a total decrease of 11% across
all layers. Total root length increased in the
AA community to a greater extent than the
AB community where smaller increases
and some decreases were observed. The
authors suggested compensatory root
growth occurred in the AB community
where a decrease in length in the
shallowest layer was mitigated by
increased growth at the middle layer. A
33% decrease in root tissue density was
observed across all soil layers in trees
exposed to O3. Specific root length
increased with soil depth and O3 with the
greatest increases in the AA community.
Talhelmetal. (2014)
FACE; Aspen FACE,
Rhinelander, Wl
(49.675°N,
89.625°E)
Aspen (Populus sp.), paper
birch (Betula papyrifera), sugar
maple (Acer saccharum)
Fumigation 1998-2008 during
daylight hours of the growing
season. Ambient O3 W126
2.1-8.8 ppm-h and elevated
12.7-35.1 ppm-h; elevated CO2
515-540 ppm, ambient avg 374.
For hourly ozone concentrations
during experimental ozone
treatment, see Kubiske and Foss
(2015)
Significant O3 effects: Ecosystem C (-9%);
mineral soil C 0-10 cm (-11%); canopy N
(-21%); O3 shifted fine roots toward soil
surface.
8-126
-------
Table 8-18 (Continued): Response of belowground processes and biogeochemical cycles to ozone exposure.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Soil Biogeochemistry Endpoints
Couture and Lindroth
(2014)
FACE; Aspen FACE,
Rhinelander, Wl
(49.675°N,
89.625°E)
(1) Insect: gypsy moths
(Lymantria dispar), forest tent
caterpillars (Malacosoma
disstria), white-marked tussock
moths (Orgyia leucostigma),
cecropia moths (Hyalophora
cecropia)
(2) Plants: single aspen
genotype (42E, Populus
tremuloides) and paper birch
(Betula papyrifera)
Treatments for 1998-2007 were
ambient Os W126 = 2.9-8.8
ppm-h and elevated
O3 = 13.1-35.1 ppm-h. Ambient
airCChand elevated (560 ppm)
CO2. For hourly ozone
concentrations during
experimental ozone treatment,
see Kubiske and Foss (2015)
Aspen grown under elevated O3
(1) decreased frass N 26% from forest tent
caterpillar, 16% from white marked tussock
moths, and 12% from gypsy moths;
(2) increased frass C:N 37% from forest
tent caterpillar, 18% from white marked
tussock moths, and 16% from gypsy moths;
and (3) increased frass condensed tannins
37% from forest tent caterpillar, 17% from
white marked tussock moths, and 17%
from gypsy moths. Frass and aspen leaf
litter chemistry were correlated, but
elevated O3 decreased the relative C:N of
frass to leaf litter by 35% and increased the
relative tannin concentration of frass to leaf
litter by 20%.
Andrew et al. (2014)
FACE; Aspen FACE,
Rhinelander, Wl
(49.675°N,
89.625°E)
Mycorrhizal fungi associated
with aspen
Fumigation 1998-2008 during
daylight hours of the growing
season. Ambient O3 W126
2.1-8.8 ppm-h and elevated
12.7-35.1 ppm-h; elevated CO2
515-540 ppm, ambient avg
374 ppm. For hourly ozone
concentrations during
experimental ozone treatment,
see Kubiske and Foss (2015)
No significant difference with treatment on
net hyphal biomass production. No
significant effects of treatment on hyphal
respiration per unit or hyphal and sporocarp
respiration on a mass-specific basis.
Zaket al. (2012)
FACE; Aspen FACE,
Rhinelander, Wl
(49.675°N,
89.625°E)
Five genotypes of quaking
aspen (Populus tremuloides)
together; P. tremuloides
genotype (216) with paper
birch (Betula papyrifera); and
P. tremuloides genotype (216)
with sugar maple (Acer
saccharum)
Fumigation 1998-2008 during
daylight hours of the growing
season. Ambient O3 W126
2.1-8.8 ppm-h and elevated
12.7-35.1 ppm-h; elevated CO2
515-540 ppm, ambient avg
374 ppm. For hourly ozone
concentrations during
experimental ozone treatment,
see Kubiske and Foss (2015)
Elevated O3 altered inter-specific
competition for N in aspen, increasing
genotype 8 plant N 93% and plant 15N
171%, and decreasing genotype 271 plant
N 40% and plant 15N 27%, with no effect on
plant competitiveness for N for the
remaining three genotypes. Elevated O3 did
not alter species competition for soil N
(measured as plant N and plant 15N).
8-127
-------
Table 8-18 (Continued): Response of belowground processes and biogeochemical cycles to ozone exposure.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Soil Biogeochemistry Endpoints
Zaket al. (2011)
FACE; Aspen FACE, Paper birch (Betula
Treatments for 2005-2008 were Elevated ozone had no effect on forest floor
Rhinelander, Wl
(49.675°N,
89.625°E)
papyrifera), sugar maple (Acer ambient O3 W126 = 7.3, 5.6, 4.9, mass, N content, or 15N content in
saccharum), various
genotypes of quaking aspen
(Populus tremuloides)
2.1 ppm-h and elevated O3 = 29.6,
14.6, 13.1, 12.7 ppm-h. Elevated
CO2 treatment (560 ppm), ambient
CO2. For hourly ozone
concentrations during
experimental ozone treatment,
see Kubiske and Foss (2015)
years 10-12 of the experiment.
Chieppa et al. (2015)
OTC; research field
5 km north of
Auburn, AL
Loblolly pine (Pinus taeda)
inoculated with root-infecting
ophiostomatoid fungi (either
Leptographium terebrantis or
Grosmannia huntii, fungal
species associated with
Southern Pine Decline)
Three ozone treatments in OTCs
(12 h/day): charcoal filtered
(-0.5% ambient air), nonfiltered
air (ambient), and 2x ambient.
The first 41 days were
acclimatization then exposure
continued 77 more days once
seedlings were inoculated with
fungus. Mean 12-h O3 over the
118 days was 14 (CF), 23 (NF),
and 37 (2x) ppb in the treatments.
12-h AOT40 values were 027
(CF), 1.631 (NF), and 31.2
(2*) ppm. Seasonal W126 were
0.033 (CF), 0.423 (NF), and
21.913 (2x)
Seedlings under 2* O3 had greater
belowground dry matter yield than CF
seedlings. Fungal lesion length was greater
on 2x O3 exposed seedlings but was not
specific to either fungal species.
Ebanvenle et al. (2016) FACE; Aspen FACE, Wood-decaying basidiomycete Fumigation 1998-2008 during
Rhinelander, Wl
(49.675°N,
89.625°E)
fungi (identified by sequencing
of primer pair ITS1f/ITS4),
growing on logs cut in 2009
from quaking aspen (Populus
tremuloides) and paper birch
(Betula papyrifera). Logs were
grown and placed in each of
the four FACE treatments in a
full factorial design
daylight hours of the growing
season. Ambient O3 W126
2.1-8.8 ppm-h and elevated
12.7-35.1 ppm-h; elevated CO2
515-540 ppm, ambient avg
374 ppm. For hourly ozone
concentrations during
experimental ozone treatment,
see Kubiske and Foss (2015)
O3 had no effect on 1 yr of decomposition
of logs, through effects on wood quality or
effects on soil decomposition conditions.
There was no statistically significant effect
of O3 on basidiomycete community
composition, although there were fewer
fungal species from logs in the O3 plots
than in the ambient O3 plots.
8-128
-------
Table 8-18 (Continued): Response of belowground processes and biogeochemical cycles to ozone exposure.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Soil Biogeochemistry Endpoints
Hofmockel et al. (20111
FACE; Aspen FACE,
Rhinelander, Wl
(49.675°N,
89.625°E)
Quaking aspen (Populus
tremuloides), sugar maple
(Acer saccharum), and paper
birch (Betula papyrifera)
Samples taken 2003, 2004 and
2007. Treatments for 1998-2007
were ambient O3
W126 = 2.9-8.8 ppm-h and
elevated O3 = 13.1-35.1 ppm-h.
Ambient air CO2 and elevated
(560 ppm) CO2. For hourly ozone
concentrations during
experimental ozone treatment,
see Kubiske and Foss (2015)
Elevated O3 reduced nitrogen in coarse
particulate organic matter (cPOM) and fine
particulate organic matter (fPOM) by 14%,
which increased C:N ratios in all soil size
fractions (from largest to smallest: coarse,
fine, and mineral-associated organic
matter) by 2-7%. Under elevated CO2,
elevated O3 decreased storage of newly
fixed C in whole soil by 22%, while
increasing the storage of older C by 21 % in
cPOM and by 13% in fPOM.
Grantzet al. (2014)
Greenhouse;
UC-Riverside, CA
Soybean, Pima cotton, bush
bean, sorghum, maize
As 12-h means of 4 nL/L,
102 nL/L, 204 nL/L
The highest O3 exposure (204 nL/L)
increases perchlorate concentration in the
potting soil 39% over soil perchlorate under
the unexposed potting mix.
Tian et al. (2012)
Model; southeastern 10 different plant functional AOT40 simulated from 1895-2007 Southeast terrestrial ecosystems were a C
U.S., 75-100° west
longitude, 30-37°
north latitude.
Includes 13 states
types were mapped across the
13-state region
using data set by Felzer et al.
(2004)
source from 1895-1950, and a C sink from
1951-2007. Largest contributor to
increased sink was CO2, followed by N
dep. O3 reduced C storage by 0.58 Pg C
during the period. The greatest impact by
O3 was in the northeast region of the study
area due to increased emissions from
impacted broadleaf forest and cropland
areas.
15N = nitrogen-15, stable isotope of nitrogen; AOT40 = seasonal sum of the difference between an hourly concentration at the threshold value of 40 ppb, minus the threshold value of
40 ppb; CF = charcoal-filtered air; C02 = carbon dioxide; FACE = free-air C02 enrichment; ITS1f/ITS4 = fungal specific primer pair; N20 = nitrous oxide; NH4+ = ammonium;
nL/L = nanoliters per liter; N03" = nitrate; 03 = ozone; OTC = open-top chamber; PLFA = phospholipid fatty acid; Pg C = petagrams (gigaton) carbon; ppb = parts per billion;
ppm = parts per million; |jl/L = microliters per liter; W126 = cumulative integrated exposure index with a sigmoidal weighting function.
8-129
-------
8.10
Terrestrial Community Composition
In the 2013 Ozone ISA the evidence was sufficient to conclude there is a likely to be causal
relationship between ozone exposure and alteration of terrestrial community composition of some
ecosystems (U.S. EPA. 2013). Ozone altered aboveground plant communities such as conifer forests,
broadleaf forests, and grasslands, and altered fungal and bacterial communities in the soil in both natural
and agricultural systems (U.S. EPA. 2013V Ozone effects on individual plants can alter the larger plant
community as well as the belowground community of microbes and invertebrates, which depend on
plants as carbon sources. Ozone may alter community composition by uneven effects on co-occurring
species, decreasing the abundance of sensitive species and giving tolerant species a competitive advantage
(Figure 8-10). Field studies linked ozone exposure to conifer decline in the San Bernardino Mountains, in
the Valley of Mexico, in alpine regions of southern France, and in the Carpathian Mountains. Evidence
suggesting ozone-induced changes to competitive interactions among trees in broadleaf forests was drawn
from the experiment at Aspen FACE and from European phytotron studies. Grassland studies suggested
that ozone alters the ratios of grasses, forbs, and legumes in these communities to favor grasses, and that
annual plant species are more sensitive to ozone than perennial species. In terms of belowground
communities, sampling from FACE and mesocosm studies suggested that ozone altered fungal
communities, including mycorrhizal species on which plants depend for water and nutrient acquisition, as
well as bacterial communities.
As described in the PECOS tool (Table 8-2). studies on community composition will be
considered for inclusion regardless of geography. Data of ozone alterations to suborganismal processes in
eukaryotes (e.g., effects on gene expression, molecular or chemical composition) will not be reviewed
from these studies. However, many microbial species in soil cannot be cultured in the laboratory, and
sampling and analysis of biologically derived chemicals and genes within the soil represent standard
methods for assessing microbial communities; these observations will be included. This section focuses
on effects of ozone on groups of different species including, for the first time, effects on plants grouped
by their shared phylogeny and on communities of animal consumers. The role of ozone in altering
community composition continues to be an area of active research in the U.S. and in other regions of the
world (Table 8-20).
8-130
-------
Historical Plant Community
Ozone exposure
overtime
Altered Plant Community
1) Decrease in legume abundance, 2) Loss of sensitive genotypes 3) Loss of sensitive species
increase in grass or herb abundance with no effect on productivity affects productivity
Figure 8-10 Mechanisms by which ozone alters plant communities.
8.10.1 Plant Community
8.10.1.1 Forest
In the 2013 Ozone ISA, evidence of ozone effects on forest composition was drawn from
observational studies of conifer decline correlated with ozone exposure (Allen et al.. 2007; de Lourdes de
Bauer and Hernandez-Teieda. 2007; Wieser et al.. 2006; Fenn et al.. 2002; McBride and Laven. 1999;
Miller. 1973) and from controlled exposure studies of broadleaf tree species in which ozone altered the
growth or mortality of sensitive genotypes or species when sensitive and tolerant trees were grown
together (Kubiskc et al.. 2007; Kozovits et al.. 2005). New evidence suggests that ozone alters tree
competitive interactions for nutrients, which partially determine forest community composition:
• Consistent with previous research on altered tree community composition at Aspen FACE, an
empirical 15N tracer study there showed that elevated ozone altered the relative competition for
nutrients among aspen genotypes (Zak et al.. 2012).
8-131
-------
New studies extend the scope of evidence regarding forest community composition beyond the
observational and controlled exposure studies summarized in the 2013 Ozone ISA to include synthesis
and models (see also Section 8.4.3):
• Models using Aspen FACE data illustrate how ozone effects on tree biomass and productivity can
scale to affect community composition at the genotype and species level. In models of aspen
genotype survival and mortality, elevated ozone altered genotype abundance and exerted a
selective pressure on aspens (Moran and Kubiske. 2013V In simulations using Aspen FACE data
of northern forests at the landscape level over centuries, elevated ozone altered species abundance
and the speed of replacement and succession (Gustafson ct al.. 2013).
• A model of forest community development through time that included ozone effects on biomass
without including ozone effects on competitive interactions showed that ozone effects on biomass
alone can alter community succession within a century (Wang et al.. 2016). This study used
published peer-reviewed data to place tree species into three sensitivity classes, applied either a 0,
10 or 20% growth reduction to species in the University of Virginia Forest Model Enhanced
(UVAFME), a gap model which tracks the growth and survival of individual trees and species
within a stand. This provides additional biological plausibility to the finding of the 2013 Ozone
ISA that differences between species in ozone sensitivity leads to decline of ozone-sensitive trees
in terrestrial communities.
8.10.1.2 Grassland and Agricultural Land
In the 2013 Ozone ISA, there was evidence of ozone effects on grassland community
composition in controlled experimental exposure studies, in models, and in reviews. Experimental
exposure or model studies found ozone shifted grass:legume dominance (Haves et al.. 2009; Yolk et al..
2006) or grass:forb dominance (Haves et al.. 2010) in grassland communities. High environmental
heterogeneity made it difficult to ascribe causality solely to ozone for changes in plant community
composition in several European experiments (Stampfli and Fuhrer. 2010; Bassin et al.. 2007; Yolk et al..
2006). while high annual variation in a U.S. study of agricultural weeds had a stronger impact on plant
community composition than did detected effects on ozone-sensitive species (Pfleeger et al.. 2010).
Key new studies include experimental ozone exposures that allow evaluation of ozone effects on
grassland community composition in analyses that explicitly include environmental or annual
heterogeneity. In a seeded pasture of three legume, two grass, and one forb species in Spain, ozone was a
more powerful explanatory factor than N for plant community biomass variation and explained 8-11% of
community biomass variation (Calvete-Sogo et al.. 2016). In a restored English grassland of 47 species,
ozone accounted for 10 and 40% of variation in herb and legume species composition in the 1st and
2nd years, respectively, of fumigation (Wedlich et al.. 2012).
The 2013 Ozone ISA included a review that identified grasslands as the most sensitive European
plant communities to ozone effects (Mills et al.. 2007) and another that identified annual plants or plants
with high leaf N concentration as particularly sensitive to ozone concentration in European species
(Haves et al.. 2007).
8-132
-------
New studies include a synthesis and a large-scale gradient exposure observational study that
confirm these findings while setting critical levels to protect European grasslands. Across an ambient
gradient of 64 grasslands in the U.K., ozone is the strongest predictor of plant species cover in single
factor models, with a change in species composition at an AOT40 of 3.1 ppm hour; and there are species
associated with low ozone sites and species associated with high ozone sites [see Table 8-18 for full
description; Payne etal. (2011)1.
Another set of new studies synthesized ozone response information from an array grassland and
herbaceous species from experiments around the world (Bcrgmann et al.. 2017; van Gocthcm ct al..
2013). Many of the species in these synthesis studies occur in the U.S. While these syntheses did not
specifically measure community composition, they do demonstrate that different species differ in ozone
sensitivity and is a mechanism for affecting community composition in grassland and agricultural lands.
In an analysis of ozone exposure-biomass loss studies of 25 annual grassland species and 62 perennial
grassland species that occur in northwestern Europe, annuals were significantly more sensitive to ozone
than were perennial species, with a projected 10% reduction in biomass across the community of
grassland annual plants at an AOT40 of 0.8 ppm hour and a 10% reduction in perennial grassland
community biomass at an AOT40 of 1.1 ppm hour (van Goethem et al.. 2013). There are 17 species from
this analysis that are native to the U.S. according to the USDA PLANTS database (USDA. 2015). 12 of
which experience biomass reduction in response to ozone (see also Section 8.13.2). In a global synthesis
of ozone effects on plants (Bcrgmann et al.. 2017). 47.5% of the 223 herbaceous plant species
experimentally exposed to ozone experienced effects in growth, productivity, C allocation, or
reproduction. This synthesis of all tested plant species to ozone exposure suggests that in considering how
ozone may shift composition from a mix of tolerant and sensitive species to a community of only tolerant
species, the sensitive species affected include roughly half of tested herbaceous species. The two
syntheses described have the strength of combining the results of many researchers, but there are some
limitations to this approach that include variation in experimental design, intraspecies variation in
response, and potential differences in species response when grown together in competition.
New studies confirm the 2006 Ozone AQCD and the 2013 Ozone ISA finding that ozone can
shift community composition towards ozone-tolerant grass species:
• In a seeded pasture of leguminous clover and three grass species in Alabama, experimentally
elevated ozone (56 ppb ozone) increased the biomass of grass species but had no effect on clover
biomass (Gilliland et al.. 2016). effectively increasing the relative biomass of grass to legumes in
the community.
• In a greenhouse study of grass and forbs in competition, grass cover increased linearly with
elevated ozone in a 12-hour mean range of 21 to 103 ppb IHaves etal. (2011); Figure 8-111.
• In an experimental fumigation in a Swiss high-elevation pasture reported in Bassin et al. (2007).
there was no effect of 7 years of elevated ozone on relative abundance of plants grouped as forbs,
grasses, or sedges, but elevated ozone did increase the abundance of a dominant grass species in
the community (Bassin et al.. 2013).
8-133
-------
A new study that directly tests the ozone response of an agricultural weed along with its crop
competitor suggests that Pfleeger et al. (2010) and newer studies should be used to infer the responses of
weeds within the context of crop production, where elevated ozone may favor weeds over crops:
• In a Chinese competition experiment between wheat and the cosmopolitan agricultural aggressive
weed species flixweed (Descurainia sophia, introduced in the U.S., state-listed as a noxious
weed, see USDA-NRCS), wheat biomass declined at 30-day exposures of 80 and 120 ppb ozone,
but flixweed biomass was not affected by either level of elevated ozone (Li et al.. 2013a).
• The community of agricultural weeds described by Pfleeger et al. (2010) was transported to
Argentina and planted under ambient ozone exposures to test whether historical experimentally
elevated ozone exposure had altered community composition. Historical exposure of ozone over
four generations had no effects on richness, diversity, or evenness of cosmopolitan agricultural
weeds, although 90 or 120 ppb ozone did increase the seedling density of the two dominant
weeds (Martinez-Ghersa et al.. 2017).
8-134
-------
160
tfl
'E
® 120
_ro
0>
w
T3
0
0)
>
o
o
ns
q
_ro
0}
l.
T3
CD
V
>
o
u
CO
0
80
40
r2 = 0.58
p=0.027
¦ ¦
^ - -A- T -A_
A
-A. -
A
&
r2 = 0.31
A L. hispidus
p=0.154
¦ A. odoratum
1
B
160
t/i
"E
a> 120
80
40
0
50 100 150
Seasonal mean 03 conc. (24 h, ppb)
r2 = 0.84
p=0.001
A—
r2 = 0.33
AL. hispidus
p=0.139
¦ D. glomerata
0
150
50 100
Seasonal mean 03 conc. (24 h, ppb)
Source: Reprinted with permission from the publisher, adapted from Haves et al. (2011).
Figure 8-11 Relationship between ozone concentrations and cover of grass
A. odoratum grown in competition with forb L. hispidus (top
graph), and cover of grass D. glomerata grown in competition
with forb L. hispidus.
8-135
-------
8.10.2
Microbes
The 2013 Ozone ISA documented effects of ozone on soil microbial communities, with changes
in proportions of bacteria or fungi as a result of experimental ozone exposure in grassland mesocosms
(Kancrva et al.. 2008; Dohrmann and Tebbe. 2005). peatland mesocosms (Morskv et al.. 2008). and forest
mesocosms (Pritsch et al.. 2009; Kasurinen et al.. 2005); as well as changes in soil microbial communities
in an agricultural ecosystem (Chen et al.. 2010) and changes specifically in fungal communities in forest
ecosystems at Aspen FACE (Edwards and Zak. 2011; Chung et al.. 2006). Soil communities can be
assessed based on the sequencing of genetic material within a sample; these analyses can target broad
phylogenetic groups (e.g., fungi, bacteria, archaea) or a subset of the community with a common
metabolic ability (e.g., nitrification, methane generation). Phospholipid fatty acid (PLFA) analysis allows
coarse characterization of soil communities based on the abundance of fatty acids in a soil sample;
different fatty acids are associated with the cell membranes of different groups of fungi or bacteria.
Ordination analysis of multiple microbial taxa allows assessment of broader microbial community
responses to ozone. Many new studies consider taxon-specific effects of ozone on bacteria
(Section 8.10.2.1). fungi (Section 8.10.2.2). and archaea (Section 8.10.2.3). and several new studies
consider metrics of soil microbial community change across taxa, as well as:
• At SoyFACE, elevated ozone altered the soil microbial community (bacteria, fungi, archaea, and
unidentified prokaryotes) based on sequencing of functional genes related to carbon, nitrogen,
phosphorus, and sulfur cycling (He et al.. 2014). Effects were detected for individual functional
genes involved in C fixation, in the breakdown of C substrates including forms of hemicellulose
and chitin, in N fixation, in denitrification, and in ammonification. These changes in microbial
community were consistent with effects of elevated ozone on N cycling at SoyFACE (see
Section 8.9).
• In a greenhouse exposure of snap beans to ozone, elevated ozone altered microbial community
structure based on PLFA data (Wang et al.. 2014). In a wheat-rice rotation at Shuangpiao Farm,
China, elevated ozone reduced functional diversity of rhizosphere microbial communities
determined by incubation on different carbon substrates (Chen et al.. 2015).
• Elevated ozone had no effect on bacterial:fungal abundance quantified by PLFA in a wheat-soy
rotation in Lake Wheeler, NC (Cheng et al.. 2011). Elevated ozone increased the archaea:bacteria
and decreased the fungi:bacteria ratios in soil associated with wheat (Li et al.. 2013b).
8.10.2.1 Bacteria
The 2013 Ozone ISA found decreases in bacterial abundance (measured as PLFA biomass) in
response to elevated ozone in meadow (Kanerva et al.. 2008) and forest (Pritsch et al.. 2009) mesocosms,
as well as increases in Gram-positive bacteria in peatland mesocosms (Morskv et al.. 2008). Many new
studies assess the effect of elevated ozone on soil bacteria:
• Bacterial abundance in agriculture: In a rice-wheat rotation at Lake Wheeler Farm, NC,
experimentally elevated ozone had no effect on bacterial abundance when assessed by PLFA
8-136
-------
analysis (Cheng etal.. 2011). In contrast, in a rice-wheat rotation at Shuangpiao Farm, China,
elevated ozone increased bacterial abundance in both rhizosphere and bulk soil when assessed by
PLFA (Chen etal.. 2015).
• Bacterial communities in natural and agricultural FACE studies: At Aspen FACE, elevated
ozone increased soil bacterial richness assessed by 16S rRNA sequencing, but did not affect the
functional composition of the bacterial communities (Dunbar et al.. 2014). At Ruohoniemi FACE
in Finland, elevated ozone increased bacterial abundance on senesced silver birch leaves, but
affected bacterial abundance during leaf decomposition only at particular stages of decomposition
on particular birch genotypes (Kasurincn et al.. 2017). A greenhouse study using rice plants
assessed the effects of 30 days of elevated ozone on bacterial communities on leaf surfaces and
root surfaces (Ucda etal.. 2016). There were no effects on broad community metrics (diversity,
richness, or evenness of bacteria assessed by 16S sequencing), but elevated ozone increased the
genetic variance of leaf-surface bacteria and decreased the relative abundance on root surfaces of
the numerically dominant operational taxonomic units [OTUs; Ueda et al. (2016)1. In the ozone
FACE rice-wheat rotation at Jiangdu City, China, elevated ozone decreased the abundance of
dominant bacterial groups determined by 16S sequencing and altered the phylogenetic diversity
of the bacterial communities (Feng et al.. 2015). In maize cultivated at SoyFACE in Illinois,
elevated ozone altered bacterial community composition in maize endosphere (i.e., plant
microbiome) and soil for one of two tested maize genotypes (Wang et al.. 2017).
• Evaluation of specific bacterial taxa: Experimentally elevated ozone affected the diversity and
evenness of methanotrophic bacteria assessed by qPCR and T-RFLPs of the gene pmoA in soil
communities associated with wheat at the Changping Seed Management Station in China, and
effects varied by ozone exposure and by soil depth (Huang and Zhong. 2015). In the ozone FACE
rice-wheat rotation at Jiangdu City, China, elevated ozone did not affect the abundance of
methanogenic bacteria assessed by 16S rRNA primers specific to methanogens [but see section
on archaea; Zhang et al. (2016)1. but did decrease the abundance of anoxygenic phototrophic
purple bacteria in soil, which are important for carbon cycling in flooded soils (Feng etal.. 2011).
In German mesocosms of European beech, elevated ozone altered root-associated actinobacterial
community composition seasonally without affecting functional diversity (Haesler et al.. 2014).
Similarly, at Shuangpiao Farm, China, elevated ozone decreased actinomycete abundance in soil
based on PLFA abundances (Chen et al.. 2015). In an Argentinian OTC pasture experiment,
elevated ozone reduced the number of Rhizobium nodules on clover roots (Menendez et al..
2017). These results can inform the assessment of ozone effects on belowground processes and
biogeochemistry in Section 8.9.
8.10.2.2 Fungi
The 2013 Ozone ISA found effects of ozone exposure on soil fungi. Studies found that ozone
exposure decreased fungal biomass in meadow mesocosms (Kanerva et al.. 2008). marginally increased
fungal abundance (quantified by PLFA profiling) in peatland mesocosms (Morskv et al.. 2008). and
altered fungal community composition in some studies of forest soils (Edwards and Zak. 2011; Chung et
al.. 2006; Kasurinen et al.. 2005). although some forest studies found no effects of ozone on fungi (Pritsch
et al.. 2009). Previous reviews have also found that ozone interactions with fungi that cause disease (U.S.
EPA. 2006). A number of new studies have evaluated the effects of ozone on fungi; as in the 2013 Ozone
ISA, evidence of effects is mixed:
8-137
-------
• Studies show effects of elevated ozone on mycorrhizal fungi in some but not all ecosystems. In
2013, a study from Aspen FACE found effects of ozone exposure on community composition of
mycorrhizal fungi (Edwards and Zak. 2011). and a German lysimeter study observed visible
differences in root mycorrhizal communities using microscopy (Kasurincn et al.. 2005V In more
recent Aspen FACE studies, there was no effect of elevated ozone on respiration by mycorrhizal
hyphae in the soil or by mycorrhizal mushrooms (Andrew et al.. 2014). Elevated ozone had no
effect on ectomycorrhizal community composition in aspen root tips assessed by sequencing
using ITS IF and ITS4 primers, but increased the abundance of four ectomycorrhizal taxa
(Andrew and Lilleskov. 2014). Experimentally elevated ozone at Ruohoniemi FACE in Finland
increased mycorrhizal colonization in silver birch roots (Kasurincn et al.. 2012). but increased
root fungal colonization in Scots pine roots (quantified as ergosterol concentration) without
increasing mycorrhizal colonization (Rasheed et al.. 2017). Elevated ozone had no effect on
arbuscular mycorrhizal communities sequenced in soy roots at SoyFACE (Cotton et al.. 2015).
but did reduce mycorrhizal colonization in a greenhouse experiment with snap beans in which
mycorrhizal inoculation was a controlled treatment (Wang et al.. 2014).
• There are effects of ozone on some but not all fungi involved in decomposition. In an Aspen
FACE study of mushrooms produced by saprophytic basidiomycete fungal communities in logs,
elevated ozone had no effect on basidiomycete community composition (Ebanvenle et al.. 2016).
Experimentally elevated ozone at Ruohoniemi FACE in Finland altered fungi in roots and litter,,
increasing fungal abundance on senesced leaf litter and altering fungal abundance in
decomposing leaf litter with leaf genotype-specific effects, as quantified by qPCR (Kasurinen et
al.. 2017).
• There are effects of ozone on some mushroom-forming fungal taxa (basidiomycetes and
ascomycetes). There were qualitative decreases under elevated ozone in species richness of
basidiomycete mushrooms on logs at Aspen FACE (Ebanvenle et al.. 2016). while qPCR of
Aspen FACE soil samples showed that elevated ozone increased the relative abundance of
basidiomycete to ascomycete fungal biomass in the soil without altering broader fungal
community composition (Dunbar et al.. 2014). At Ruohoniemi FACE in Finland, elevated ozone
increased mushroom production in a year when mushroom production was high enough to
quantify in all rings (Kasurinen et al.. 2012).
• Consistent with the 2013 Ozone ISA, new evidence of ozone effects across all fungal species is
mixed. Elevated ozone had no effect on fungal abundance broadly quantified by PLFA in a
wheat-soy rotation in Lake Wheeler, NC (Cheng et al.. 2011). Elevated ozone reduced fungal
PLFA abundance in rhizosphere and bulk soil in a wheat-rice rotation at Shuangpiao Farm, China
(Chen et al.. 2015).
• Some new studies have reported that elevated ozone may interact with fungi that cause disease in
plants. Elevated ozone may interact with plant pathogens to affect plant survival. In an exposure
experiment in Alabama in which loblolly pines were inoculated with two fungal pathogens
associated with Southern Pine Decline, elevated ozone increased the length of fungal lesions on
plant roots (Chieppa et al.. 2015). In an OTC experiment in India involving wheat and the fungal
disease Bipolaris sorokiniana, elevated O3 increased the frequency of leaf lesions, the production
of disease spores, and decreased by half the latency stage of the disease (Mina et al.. 2016).
8-138
-------
8.10.2.3 Archaea
The 2013 Ozone ISA did not address the effects of ozone on archaeal community composition.
The effects of elevated ozone on archaea have been assessed by three studies conducted at the ozone
FACE soy-wheat rotation in Jiangdu City, China. Elevated ozone increased the archaea:bacteria ratio in
soil associated with wheat (Li et al.. 2013b). In rice-associated soil, elevated ozone decreased the
richness, diversity, and abundance of the methanogenic archaeal community and decreased abundance of
the dominant archaeal methanogen Methanosaeta (Zhang et al.. 2016; Feng et al.. 2013).
8.10.3 Consumer Communities
This section addresses the effects of ozone on communities of animal consumers via effects on
vegetation and plant roots. The 2013 Ozone ISA did not address this topic within the context of altered
terrestrial community composition. The effects of ozone on aboveground herbivores or the ecosystem
service of pollination are addressed in more detail in Section 8.6 and Section 8.7.
• Ozone affects aboveground communities of invertebrates, which include both herbivores and
insectivores. In Aspen FACE, elevated ozone altered the abundance of some arthropod species
with trends towards effects on particular feeding guilds and decreased cumulative arthropod
species richness in aspen canopy (Hillstrom et al.. 2014). In a community of cosmopolitan
agricultural weeds grown under four generations of elevated ozone, there was a strong linear
relationship between plant species richness and aboveground arthropod diversity in a community
that had grown for four generations at 0 ppb historical ozone, but no relationship between plant
and arthropod diversity in communities grown for four generations at 90 or 120 ppb historical
ozone (Martinez-Ghersa et al.. 2017).
• Ozone affects belowground communities of invertebrates, including herbivores, detritovores, and
higher level consumers. In soils under elevated ozone (110 ppb), the diversity index of nematodes
decreased and the dominance index increased (Bao et al.. 2014). while in a different FACE
experiment, elevated ozone (60 ppb) changed the proportion of fungivorous nematodes in soil (Li
et al.. 2016a). In contrast, a separate study found no change in the nematode populations in soils
exposed to elevated levels [50, 60 ppb; Payne et al. (2017)1. However, in these soils, there was an
increase in abundance and loss of diversity of testate amoeba (Payne et al.. 2017).
8.10.4 Summary and Causality Determination
The 2013 Ozone ISA found the evidence sufficient to conclude that there is a likely to be causal
relationship between ozone exposure and the alteration of community composition of some ecosystems.
Evidence of this relationship was presented for forest communities of trees; grassland communities of
grasses, herbs, and legumes; and soil microbial communities of bacteria and fungi. Recently published
papers extend the evidence for each of these topics (Table 8-19 and Table 8-20).
8-139
-------
Table 8-19 Summary of evidence for a causal relationship between ozone
exposure and terrestrial community composition, based on Table 2
from the Preamble.
Aspect of Ecological
Weight of Evidence
Key Evidence
Key References
Relevant pollutant
exposures
Defined in PECOS tool
Table 8-2
Studies at relevant O3 exposures
2006 Ozone AQCD (U.S. EPA.
2006); 2013 Ozone ISA (U.S. EPA.
2013)
Studies in which
chance, confounding,
and other biases are
ruled out with Grassland studies
reasonable confidence
Models of forest tree community composition in the
U.S.
Gustafson et al. (2013): Wang et al.
(2016)
Calvete-Soqo et al. (2016); Wedlich
et al. (2012); Payne et al. (2011)
Controlled exposure Forest: Aspen FACE
studies (lab or small- to
medium-scale field
study) Grassland plants
2013 Ozone ISA (U.S. EPA. 2013):
Zak et al. (2012).
2006 AQCD (U.S. EPA. 2006):
2013 Ozone ISA (U.S. EPA. 2013):
Gilliland et al. (2016): Calvete-Soao
et al. (2016): Wedlich et al. (2012)
Studies with large scale Models of regional forest composition in the U.S. Gustafson et al. (2013): Wang et al.
of inference (2016)
Global synthesis of woody and herbaceous plant Beramann et al. (2017)
responses to controlled exposure of O3, grouped by
plant family (relevant to natural plant communities)
Grassland plant studies at national or European Pavne et al. (2011): van Goethem
scale et al. (2013): U.S. EPA (2013)
Previous U.S. EPA syntheses 2006 Ozone AQCD (U.S. EPA.
2006): 2013 Ozone ISA (U.S. EPA.
2013)
Forest (studies in the U.S., Europe) Section 8.10.1.1
Grassland (studies in the U.S., Argentina, China, Section 8.10.1.2
U.K., Spain, Switzerland, Europe)
Multiple studies by
multiple research
groups
8-140
-------
Table 8-19 (Continued): Summary of evidence for a causal relationship between
ozone exposure and terrestrial community composition,
based on Table 2 from the Preamble.
Aspect of Ecological
Weight of Evidence
Key Evidence
Key References
Many lines of evidence Forest (experimental exposure, observations at
ambient exposures, synthesis, multiple models)
2006 AQCD (U.S. EPA. 2006);
2013 Ozone ISA (U.S. EPA. 2013);
Section 8.10.1.1
Grassland and agricultural land (experimental
exposure, observations at ambient exposures,
synthesis, multiple models)
2006 Ozone AQCD (U.S. EPA.
2006); 2013 Ozone ISA (U.S. EPA.
2013); Section 8.10.1.2
Grassland: exposure-response relationships
van Goethem et al. (2013); Haves
etal. (2011); Pavne et al. (2011)
Soil microbial communities
Section 8.10.3
Aboveground and belowground invertebrate
communities
Section 8.10.4
FACE = free-air C02 enrichment; 03 = ozone.
In forests, previous evidence included correlational studies across ambient gradients of ozone
exposure that found effects of ozone on conifer species, and studies with controlled experimental
exposure of trees that found effects of ozone on deciduous trees. Key new studies (Wang et al.. 2016;
Gustafson et al.. 2013) show that observational and experimental observations of ozone effects on tree
species extend to affect regional forest composition in the eastern U.S. Additionally, a global-scale
synthesis of research on ozone effects on plants confirms that some plant families are more susceptible to
ozone damage than others (Bcrgmann et al.. 2017). which is consistent with studies reviewed in previous
ISA and AQCDs (U.S. EPA. 2013. 2006. 1996. 1986). This lends biological plausibility to a mechanism
by which elevated ozone alters terrestrial community composition by inhibiting or removing
ozone-sensitive plant species or genotypes, and thereby altering competitive interaction to favor the
growth or abundance of ozone-tolerant species or genotypes.
In grasslands, previous evidence included multiple studies from multiple research groups to show
that elevated ozone shifts the balance among grasses, forbs, and legumes in European grassland
communities. There are new studies with findings consistent with earlier research, including a study in the
U.S. that found elevated ozone affected the ratio of grass-to-legume biomass (Gilliland et al.. 2016).
There are also new studies from European grasslands that found exposure-response relationships between
ozone and community composition (Haves etal.. 2011; Pavne etal.. 2011). including a study that
calculated AOT40 values for 10% reduction in biomass for 87 grassland species (van Goethem et al..
2013). some of which also grow in the U.S.
8-141
-------
In soil microbial communities, previous evidence included studies that found effects on the ratio
of bacteria to fungi in soil communities, as well as effects on community composition of mycorrhizal
fungi. New studies confirm that elevated ozone alters soil microbial taxa, although as with previous
evidence, the strength and direction of effects are not consistent across ecosystems. This may be due to
the proposed mechanism of ozone effects on soil microbial communities, namely, that ozone indirectly
affects soil communities via effects on plant chemistry and plant carbon allocation, which alter the
substrates on which soil microbial communities subsist (Figure 8-12). This mechanism also explains an
aspect of altered community composition not directly addressed in the 2013 Ozone ISA: the alteration of
invertebrate community composition from effects that elevated ozone has on plants, as documented in
several recent studies.
Plant
Soil Microbial Community
Roots
Wood
Leaves
Mycorrhizae
Fungi:Bacteria
Rhizosymbionts
Ozone
Archaea
Fungi
Bacteria
Soil Invertebrates
Figure 8-12 Biological plausibility of ozone effects on soil microbial
communities and soil invertebrate communities.
The 2013 Ozone ISA presented multiple lines of evidence that elevated ozone alters terrestrial
community composition, and recent evidence strengthens our understanding of the effects of ozone on
plant communities, while confirming that the effects of ozone on soil microbial communities are diverse.
The body of evidence is sufficient to conclude that there is a causal relationship between ozone
exposure and the alteration of community composition of some ecosystems.
8-142
-------
Table 8-20
Terrestrial community composition response to ozone exposure.
Study
Study Type and
Location Study Species Ozone Exposure
Effects
Aqathokleous et al.
(2015)
Meta-analysis
473 wild plant species
tested for O3 effects in
195 previously published
papers
Multiple
Within the published literature testing O3 effects
on plants, 80% of wild plant species experience
negative effects of O3 exposure, representing
210 genera and 85 families.
Li et al. (2013b)
O3 FACE site;
Jiangdu City,
Jiangsu, China
(32.58°N, 119.70°
Two cultivars of Triticum
aestivum (wheat), grown
November-June in annual
wheat-rice rotation: O3
sensitive Yannong 19 and
Yangmai 16
Ambient: 40 ppb, elevated: 60 ppb Elevated O3 had no effect on number of
9:00 a.m.-6:00 p.m. from March 3 to
May 31, 2010
detected genes. Elevated O3 reduced
Simpson's evenness of soil microbial functional
genes, but only altered the abundance of gene
fhs, which decreased under elevated O3.
Elevated O3 changed the soil community
associated with Yannong 19 cultivar,
decreasing the fungi:bacteria ratio and
increasing the archaea:bacteria ratio.
Feng et al. (2013)
O3 FACE site;
Jiangdu City,
Jiangsu, China
(32.585°N,
119.70°E)
Surface soil samples pulled
from Oryza sativa (rice) in
vegetative (July) and
flowering (September)
stages in 2010, cultivated in
annual rice-wheat rotation
FACE: Three ambient O3 rings,
three elevated rings with mean
60 ppb O3 during rice season,
fumigated 9:00 a.m. to sunset
(target was 1.5* ambient O3, not to
exceed 250 ppb O3)
Elevated O3 decreased the diversity 18% and
the richness 39% of the methanogenic archaeal
soil community under vegetative rice. Elevated
O3 decreased total abundance of the dominant
archaeal methanogen Methanosaeta 35% in
soils under vegetative rice and 44% in soils
under flowering rice and decreased its relative
abundance within the methanogenic archaea as
well. Elevated O3 had a stronger influence on
methanogenic archaeal community composition
than did rice lifestage.
8-143
-------
Table 8-20 (Continued): Terrestrial community composition response to ozone exposure.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects
Heetal. (2014)
FACE; SoyFACE,
Illinois (40.056°N,
88.201 °W)
Soil microbial community
under Glycine max
(soybean)
Ambient (-37.9 ppb), elevated O3
(-61.3 ppb) soybeans also grown
under elevated CO2 (-550 ppm) and
elevated CO2 + elevated O3
In elevated O3 soybean crop surface soils,
abundance of niFH, narG, norB, and ureC
N-cycling genes were significantly decreased.
There were no significant differences in the
subsoil. For C-cycling genes in elevated O3 soil
samples, most were unchanged while fungal
arabinofuranosidase and endoglucanase
increased significantly and xylanase, cellobiase
and exochitinase decreased significantly. Soil N
was quantified; NH4-N was significantly lower in
the surface soil and NO3-N significantly higher
in subsoil of elevated O3 plots compared to
ambient.
Li etal. (2013a)
OTC; wheat fields,
north China
Agricultural weed
Descurainia sophia
(flixweed), grown alone or in
competition with Triticum
aestivum cultivar Liangxing
99 (winter wheat), planted
October., fumigated April,
harvested May
Three O3 treatments: ambient
(<40 ppb O3), elevated (80 ± 5 ppb
for 7 h/day for 30 days), highly
elevated (120 ± 10 ppb for 7 h/day
for 30 days)
Wheat biomass and yield decline in response to
competition under elevated and high ozone,
whereas competition does not affect flixweed
biomass or yield at any ozone exposure.
8-144
-------
Table 8-20 (Continued): Terrestrial community composition response to ozone exposure.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects
Chen et al. (2015)
OTC; winter
wheat-rice rotation
at Shuangpiao
Farm, Jianxing City,
Zhejiang Province,
China (31,88°N,
121,30°E)
Soil microbial communities
under wheat grown
November 2006-May 2007,
assessed as functional
diversity by incubations with
31 different C substrates for
community-level
physiological profiles
(CLPPs), and assessed as
microbial community
structure by phospholipid
fatty acid analysis (PLFAs)
Three O3 treatments March-May
2007: control, AOT40 = 0
(charcoal-filtered air); elevated O3,
AOT40 = 1,585 ppb-h (2 h daily of
50 ppb, 4 h daily of 100 ppb, and
2 h daily of 150 ppb O3); highly
elevated O3, AOT40 = 9,172 ppb-h
(2 h daily of 100 ppb, 4 h daily of
150 ppb, and 2 h daily of 200 ppb
O3)
Principal component analysis (PCA) of CLPPs
shows that elevated and highly elevated O3
affect microbial functional diversity in
rhizosphere (root-associated) soil. PCA of
PLFAs shows that elevated and highly elevated
O3 affect microbial structure and abundance in
rhizosphere and nonrhizosphere soil. Highly
elevated O3 reduced Shannon-Weaver diversity
10% in rhizosphere and 4% in nonrhizosphere
soils relative to control soil functional diversity,
and reduced rhizosphere soil functional
richness 24% relative to control. In
nonrhizosphere soils, both elevated and highly
elevated 63 increased bacterial abundance
(elevated O3: 4% increase over control relative
abundance, highly elevated O3: 5% increase
over control relative abundance) and decreased
fungal abundance (22 and 28%). In rhizosphere
soils, highly elevated O3 increased bacterial
abundance 1%, decreased actinomycete
abundance 23%, and decreased fungal
abundance 11%.
Huang and Zhona
(2015)
OTC; Seed
Management
Station of
Changping, Beijing,
China (40.20°N,
116.12°E)
Methanotrophic bacteria in
soil under winter wheat; soil
methanotrophs assessed by
qPCR of pmoA gene and
16S rRNA primers specific
to type 1 and type 2
methanotrophs, and by
T-RFLP of pmoA
Fumigation for 9 h/day, April-June
2010, with four different treatments:
control (charcoal-filtered), low O3
(nominally 40 ppb O3), moderate O3
(nominally 80 ppb), and high O3
(nominally 120 ppb)
O3 exposure affects soil methanotroph
communities at different soil depths. In 0-10 cm
soil, low O3 increases Shannon diversity of
methanotrophs 30% and increases evenness
32% relative to diversity in control soils, while
high O3 decreases diversity 13%. In 10-20 cm
depth soil, low O3 decreases diversity 18% and
evenness 18%, moderate O3 decreases
diversity 13%, and high O3 increases diversity
31% and increases evenness 22%. In
20-40 cm depth soil, moderate O3 increases
diversity 19%, and high O3 decreases diversity
23% and decreases evenness 23%.
8-145
-------
Table 8-20 (Continued): Terrestrial community composition response to ozone exposure.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects
Feng et al. (2015)
O3 FACE site;
Jiangdu City,
Jiangsu, China
(32.585°N,
119.70°E)
Soil bacteria community
assessed by bacterial 16S
rRNA in surface soil
samples pulled from Oryza
sativa (rice) in vegetative
(July) and flowering
(September) stages in
2012, cultivated in annual
rice-wheat rotation
Ambient O3, seasonal 7-h
(9:00 a.m.-4:00 p.m.) mean for
2012 was 33.7 ppb
(AOT40 = 5.2 ppm-h); elevated O3,
seasonal 7-h mean for 2012 was
42.6 ppb (AOT40 = 13.4 ppm-h)
Elevated O3 altered soil bacterial community
composition associated with two different rice
cultivars, decreasing the relative abundance of
dominant bacteria Acidobacteria, Bacteroidetes,
and Chloroflexi, and increasing the relative
abundance of Proteobacteria.
Cotton et al. (2015)
FACE; SoyFACE,
Illinois (40.056°N,
88.201 °W)
Sequencing of arbuscular
mycorrhizal fungi in Glycine
max cultivar Pioneer 93B15
(soybean) roots,
54-62 days after planting
Factorial CO2 and O3 treatment:
(1) ambient CO2 and ambient O3;
(2) ambient CO2 and elevated O3
(1.2* ambient in 2004 and 2006,
1.6* ambient in 2008); (3) elevated
CO2 (550 ppm) and ambient O3;
(4) elevated CO2 and elevated O3
No effects of elevated O3 on arbuscular
mycorrhizal fungi, richness, evenness, or
community composition.
Zhang et al. (2016)
O3 FACE site;
Jiangdu City,
Jiangsu, China
(32.585°N,
119.70°E)
Methanogenic archaea and
methanotrophic bacteria,
sequenced by 16S rRNA
primers specific in surface
soil samples pulled from
Oryza sativa (rice) in
vegetative (July) and
flowering (September)
stages in 2012, cultivated in
annual rice-wheat rotation
Ambient O3, seasonal 7-h
(9:00 a.m.-2:00 p.m.) mean for
2012 was 33.7 ppb
(AOT40 = 5.2 ppm-h); Elevated O3,
seasonal 7-h mean for 2012 was
42.6 ppb (AOT40 = 13.4 ppm-h)
O3 decreased methanogenic archaeal
abundance 20-21% under both rice cultivars in
their vegetative growth phase. O3 did not affect
methanotrophic bacterial abundance. O3
decreased methanogenic archaeal diversity:
22-41% for phylogenetic diversity under both
rice cultivars in their vegetative phase, and
25-59% by the Chao index for both rice
cultivars in their vegetative phase. Elevated O3
affected soil methanogenic archaeal diversity
more strongly under vegetative rice than under
flowering rice.
8-146
-------
Table 8-20 (Continued): Terrestrial community composition response to ozone exposure.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects
Feng et al. (2011)
O3 FACE site;
Jiangdu City,
Jiangsu, China
(32.585°N,
119.70°E)
Soils sampled from rice in
2009: anoxygenic
phototrophic purple bacteria
assessed by pufM and
bacterial 16S rRNA genes
and by counts of
representative purple
nonsulfur bacteria
Rhodopseudomonas
palustris
Ambient O3 mean for 2009 -40 ppb;
elevated O3 -60 ppb (nominally,
50% higher than ambient) from
9:00 a.m. to sunset, daily, unless
leaves were wet or ambient O3 was
lower than 20 ppb. When the target
O3 was higher than 250 ppb, the set
point was fixed at 250 ppb to
prevent the plants from being
exposed to extraordinarily high O3
Abundance of anoxygenic phototrophic purple
bacteria and R. palustris were significantly
lower in elevated O3 when quantified from rice
in vegetative growth and seed set stages (no
effect of O3 when rice was flowering). There
was less R. palustris diversity (number of
genotypes) in elevated O3 soils than in ambient
O3 soils.
Bao et al. (2014)
OTC; Shenyang
Experimental
Station of Ecology,
Chinese Academy
of Sciences,
(41.517°N,
123.367°E)
Glycine max cultivar Tiefeng
29 (soybean), nematodes
Ambient (-45 ppb) and elevated
(110 ± 10 ppb). Exposed to elevated
ozone or/and UV-B radiations for
8 h (9:00 a.m.-5:00 p.m.) per day in
the middle of the photoperiod from
June 20 to September 7
Soybean growth stage-dependent effects on
the abundance of bacterivores and fungivores
were reported for elevated O3. The ratios of
bacterivores and fungivores:plant parasites and
omnivores-carnivores:plant parasites were
significantly affected by elevated O3. The
observed effect was soybean growth
stage-dependent for omnivore-carnivore:plant
parasites, but not for bacterivores and
fungivores:plant parasites. Indices of nematode
diversity, enrichment and community structure
decreased under elevated O3. The nematode
dominance index increased under elevated O3.
Li et al. (2016a)
FACE; Jiangsu
Province, China
(32.583°N,
119.70°E)
Arthropod: nematode
population;
FACE soil: collected from
rice-wheat rotation system
(Oryza sativa, Triticum
aestivum);
Greenhouse plants:
Yangfumai 2 (Y2), Yannong
19 (Y19), Yangmai 15
(Y15), Yangmai 16 (Y16),
rice cultivar (ll-you)
FACE site: ambient (40 ppb) and
elevated (60 ppb)
Other than the ratio between fungal and
bacterial PLFAs, no other cultivar or O3
exposure effects were detected. Although the
total number of nematodes and the number of
bacterivorous and plant parasitic nematodes
were not affected by previous O3 exposure or
wheat cultivar, the number of fungivorous
nematodes increased (except for Y2 cultivar) in
soils previously exposed to O3. The number of
omnivorous-predatory nematodes differed
between cultivars.
8-147
-------
Table 8-20 (Continued): Terrestrial community composition response to ozone exposure.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects
Martinez-Ghersa et
al. (2017)
Mesocosm;
University of
Buenos Aires,
Argentina (34.58°S,
58.58°W)
Populations of agricultural
weeds (mostly Eurasian
annuals) from the seed
bank in Corvallis, OR;
planted and grown in
Argentina and interacting
with the Argentinian insect
community
Plants are descended from
populations from U.S. EPA
mesocosm experiments in Oregon.
O3 exposures for the 4 yr of
experiment were: charcoal-filtered
air with low O3 and two elevated
treatments (targets of 90 and
120 ppb). The ozone profile was
developed based on the regional air
quality data from the Midwest (U.S.)
and consisted of an episodic pattern
of varying daily peak concentration.
Each chamber received the same
episodic ozone exposure profile
each year. Hourly requested peaks
ranged from 1 to 155 ppb for the
90 ppb treatments and 1 to 219 ppb
for the 120 ppb ozone treatments.
The high peaks lasted for 1 h
There was no effect of historical O3 exposure
on descendant plant community richness,
diversity, or evenness. Historical O3 exposure
increased the seedling density of the two
dominant plants in descendant communities: 90
and 120 ppb increased Spergula arvensis
density 30-50%, and Calandrinia ciliata density
increased 109 and 217%, respectively. The
relative abundance of the other plant species
declined in response to O3. There was linear
relationship between plant species richness and
arthropod diversity at 0 ppb historical O3
(Spearman's r= 0.71), but no relationship at 90
or 120 ppb historical O3 exposure. Historical O3
does not affect the richness, diversity, or
evenness of the arthropod community
associated with descendant plant community
but does increase the relative abundance of
carnivore arthropods while decreasing the
relative abundance of herbivore arthropods
(p < 0.05).
Menendez et al.
(2017)
OTC; University of
Buenos Aires,
Argentina (34.59°S,
58.58°W)
6-week-old Trifolium repens
(white clover) and
Rhizobium spp. (N fixing
bacteria in root nodules on
clover)
Charcoal-filtered ambient O3 in
2010-2011 (concentrations not
reported), elevated O3 for 4 h/day
over 5 days at max 90-120 ppb O3
O3 exposure reduced the number of Rhizobium
nodules on clover roots by 35%.
Wang et al. (2017)
FACE; SoyFACE,
Illinois (40.056°N,
88.201 °W)
Soil, rhizosphere, and root
endosphere-associated
microbial communities in
maize crop grown under
elevated 63
O3 plots were enriched with O3 to a
target concentration of 100 ppb by
fumigation (summer 2014)
No change in a-biodiversity was observed in
endosphere, rhizosphere or soil with elevated
O3. There were significant differences in
p-biodiversity of microbial communities in the
endosphere and soil. Microbial community
composition shifted by maize genotype,
specifically in the endosphere samples of
inbred B73 and the soil where both hybrids
were grown under elevated O3.
8-148
-------
Table 8-20 (Continued): Terrestrial community composition response to ozone exposure.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects
Minaetal. (2016)
OTC; Indian
Agriculture
Research Institute,
(28.583°N, 77.20°E)
Plant: wheat (Triticum
aestivum cultivar PBW343);
Pathogen: Bipolaris
sorokiniana
Ambient (12-72 ppb) and elevated
(ambient O3 + 25-30 ppb). Wheat
plants were pretreated with different
O3 levels for 8 h/day from seedling
to 65-day-old stage
The maximum number of lesions/leaf and
spores/25mm2 were detected in elevated O3.
Elevated O3 also shortened the Bipolaris
latency period from 20 to 10 days or less.
Antioxidants interfered with the growth of
Bipolaris on wheat plants. Compared to
charcoal-filtered air, elevated O3 reduced PR
protein content and chitinase activity alone and
in combination with Bipolaris. The effect of
elevated O3 alone and O3 + Bipolaris on PR
protein content and chitinase activity were
reversed by the addition of antioxidants.
Cheng et al. (2011)
OTC; Lake Wheeler
Experimental
Station, NC
(35.72°N, 78.67°W)
Wheat-soybean rotation: in
November to June, O3
tolerant Triticum aestivum
cultivar Coker 9486 (soft red
winter wheat), and in June
to November, soybean
(multiple cultivars over 4 yr
experiment)
Full factorial O3 and CO2 fumigation
for 4 yr: (1) charcoal-filtered control
(canopy height seasonal daily 12-h
avg for June-November is 19.9 ppb
O3; November-June is 20.7 ppb 63)
with ambient CO2 (376 ppm
June-November and 388 ppm
November-June); (2) elevated O3
(canopy height seasonal daily 12-h
avg for June-November 65.7 ppb
O3; November-June 49.8 ppb 63);
(3) elevated CO2 (555 ppm
June-November and 547 ppm
November-June); (4) elevated O3
and elevated CO2, as above
Elevated O3 had no effect on soil C, soil N, or
fungal and bacterial soil abundances or ratio
assessed by PLFA.
8-149
-------
Table 8-20 (Continued): Terrestrial community composition response to ozone exposure.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects
Kasurinen et al.
(2012)
FACE; Ruohoniemi Clones of four genotypes
FACE, Kuopio,
University of
Eastern Finland,
Finland (62.88°N,
27.62°E)
from the wild population of
Betula pendula (silver birch)
as well as their associated
mycorrhizal community and
fruiting bodies of Laccaria
laccata (mycorrhizal fungi)
Factorial O3 by temperature
treatment: mean ambient O3 is
23.4 ppb in 2007, 23.8 ppb in 2008
(AOT40 = 0.14 ppm-h in
2007,1.6 ppm-h in 2008); mean
elevated O3 is 28.1 ppb in 2007,
32.0 ppb in 2008
(AOT40 = 4.9 ppm-h in 2007,
9.0 ppm-h in 2008), fumigation 800
to 2,200 daily, from spring leaf out to
autumn; temperature treatment is
ambient or elevated by infrared
rings above the canopy
Elevated O3 increases mycorrhizal infection in
roots by 9% at ambient temperatures (T) and
5% at elevated T. Elevated O3 increases
mushroom count 660% in ambient T and 230%
in elevated T above respective ambient O3
treatments.
Andrew et al. (2014)
FACE; Aspen
FACE, Rhinelander,
Wl (49.675°N,
89.625°E)
Mycorrhizal fungi
Fumigation 1998-2008 during
daylight hours of the growing
season. Ambient O3 W126
2.1-8.8 ppm-h and elevated
12.7-35.1 ppm-h; elevated CO2
515-540 ppm, ambient avg
374 ppm. For hourly ozone
concentrations during experimental
ozone treatment, see Kubiske and
Foss (2015)
No significant difference with treatment on net
hyphal biomass production. No significant
effects of treatment on hyphal respiration per
unit or hyphal and sporocarp respiration on a
mass-specific basis.
Andrew and Lilleskov
(2014)
FACE; Aspen
FACE, Rhinelander,
Wl (49.675°N,
89.625°E)
Plot areas planted with
Populus tremuloides.
ectomycorrhizal fungal root
tip communities
Fumigation 1998-2006 during
daylight hours of the growing
season. Ambient O3 W126
2.9-8.8 ppm-h and elevated
14.6-35.1 ppm-h; elevated CO2
(560 ppm), ambient CCh.For hourly
ozone concentrations during
experimental ozone treatment, see
Kubiske and Foss (2015)
Soil properties were a stronger determinant of
EMF root tip community structure than O3
treatment. The relative abundances of fourtaxa
(Tomentella sp. "A," Tomentella sp. "C,"
Sebacina [Serendipita] sp., and Hebeloma
crustuliniforme species group) were increased
under elevated O3.
8-150
-------
Table 8-20 (Continued): Terrestrial community composition response to ozone exposure.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects
Moran and Kubiske
(2013)
FACE; Aspen
FACE, Rhinelander,
Wl
Clones of five genotypes of
Populus tremuloides
(aspen), from the
aspen-only sections of the
experiment, 1997-2008
Full factorial: O3 and CO2,
1998-2008. Ozone: ambient (W126
2.1-8.8 ppm-h) or elevated (W126
12.7-35.1 ppm-h). CO2: ambient
(360 ppm) or elevated (560 ppm);
for hourly ozone concentrations
during experimental ozone
treatment, see Kubiske and Foss
(2015)
In model runs of genotype abundance after
12 yr of O3 exposure, 63 increased the relative
abundance of genotypes 216 and 8L by 6 and
26%, respectively, and decreased the relative
abundance of genotypes 42E and 259 by 9 and
44% respectively.
Gustafson et al.
(2013)
LANDIS Model
using Aspen FACE
data; Rhinelander,
Wl
Acer saccharum (sugar
maple), Betula papyrifera
(paper birch), four clones of
Populus tremuloides
(aspen)
Target over the course of the
Rhinelander experiment was
30-50 ppb for control, and
60-80 ppb for elevated
Overall, total biomass was lowest under the O3
treatment. The O3 treatment significantly
affected abundance of all taxa except one
clone. Simulations suggest that O3 will affect
forest composition at the landscape scale.
Simulations suggest that O3 will cause an
increase in birch at the expense of aspen. By
yr 180, elevated O3 decreased productivity by
half. Elevated O3 reduced landscape biomass.
Zaket al. (2012)
FACE; Aspen
FACE, Rhinelander,
Wl (49.675°N,
89.625°E)
Five genotypes of Populus
tremuloides (quaking
aspen) together;
P. tremuloides genotype
(216) with Betula papyrifera
(paper birch); and
P. tremuloides genotype
(216) with Acer saccharum
(sugar maple)
Fumigation 1998-2008 during
daylight hours of the growing
season. Ambient O3 W126
2.1-8.8 ppm-h and elevated
12.7-35.1 ppm-h; elevated CO2
515-540 ppm, ambient avg
374 ppm. For hourly ozone
concentrations during experimental
ozone treatment, see Kubiske and
Foss (2015)
Elevated O3 altered inter-specific competition
for N in aspen, increasing genotype 8 plant N
by 93% and plant 15N by 171%, and decreasing
genotype 271 plant N by 40% and plant 15N by
27%, with no effect on plant competitiveness for
N for the remaining three genotypes. Elevated
O3 did not alter species competition for soil N
(measured as plant N and plant 15N).
8-151
-------
Table 8-20 (Continued): Terrestrial community composition response to ozone exposure.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects
Wang et al. (2016)
Gap Model
(UVAFME)
representing "a
typical temperate
deciduous forest in
the southeast U.S."
32 total species were used
in the model; 10 dominant
tree species: Acer rubrum
(red maple), Acer
saccharum (sugar maple),
Carya cordiformis (bitternut
hickory), Fagus grandifolia
(American beech),
Liriodendron tulipifera (tulip
poplar), Quercus alba (white
oak), Quercus montana
(chestnut oak), Quercus
rubra (red oak), Quercus
velutina (black oak), Prunus
serotine (black cherry)
Each of the 32 species was ranked
based on O3 sensitivity (resistant,
intermediate, or sensitive). Growth
reduction parameters were 0, 10,
and 20% for each of the three
categories, respectively
O3 resistant species dominate and sensitive
species decline over the 500-yr simulation.
Overall forest biomass and forest carbon
storage do not decrease overtime under high
O3 conditions because tolerant species growth
is enhanced as they are released from
competition by the loss of O3 sensitive species.
O3 reduced biodiversity overtime.
Chieppa et al. (2015)
OTC; research field
5 km north of
Auburn, AL
Pinus taeda (loblolly pine)
inoculated with either
Leptographium terebrantis
or Grosmannia huntii (root
infecting ophiostomatoid
fungal species associated
with Southern Pine Decline)
O312 h/day. The first 41 days were
acclimatization; O3 exposure
continued 77 more days once
seedlings were inoculated with
fungus. Mean 12 h O3 over the
118 days was 14 (charcoal-filtered),
23 (ambient), and 37 (2x ambient)
ppb. 12-h AOT40 values were 0.027
(CF), 1.631 (ambient) and 31.2
(2*) ppm-h. Seasonal W126 were
0.033 (CF), 0.423 (ambient) and
21.913 (2x)
Seedlings under 2* O3 had greater
belowground dry matter yield than CF
seedlings. Fungal lesion length was greater on
2x O3 exposed seedlings but was not specific to
either fungal species.
8-152
-------
Table 8-20 (Continued): Terrestrial community composition response to ozone exposure.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects
Hillstrom et al. (2014)
FACE research
facility; Rhinelander,
Wl
(1) Tree canopy arthropods
sampled 3x each summer
2005-2007
(2) Plants: Populus
tremuloides genotypes
(216, 217, 42E) and Betula
papyrifera (paper birch)
(1) 03: ~1.5x ambient. (2) C02:
-560 ppm; for hourly ozone
concentrations during experimental
ozone treatment, see Kubiske and
Foss (2015)
Elevated O3 did not affect total arthropod
abundance on aspen or birch, but did
significantly alter the abundance of certain
species, tending to decrease the abundance of
phloem-feeders and increase the abundance of
leaf chewing and galling feeding guilds.
Elevated O3 did not affect arthropod species
richness in any single summer for either aspen
or birch, but elevated O3 significantly decreased
arthropod richness 15% cumulatively sampled
across all 3 yr in aspen canopies (p = 0.03).
Elevated O3 did not affect arthropod community
composition in aspen canopies in any single
year or across all years. Elevated O3 altered
community arthropod community composition in
birch canopies only in 2007 but had no effect
across all years, genotype was an important
determinant of community composition in aspen
canopies.
Dunbar et al. (2014)
FACE; Aspen
FACE, Rhinelander,
Wl (49.675°N,
89.625°E)
Fungal and bacterial
communities in top 5 cm of
mineral soil in Populus
tremuloides FACE soils,
July 2007; quantified by
qPCR, clone library
surveys, and gene pyrotag
surveys of bacterial 16S
rRNA, fungal 18S rRNA,
and fungal LSU rRNA
Treatments for 1998-2007 were
ambient O3 W126 = 2.9-8.8 ppm-h
and elevated O3 = 13.1-35.1 ppm-h.
Ambient air CO2 and elevated
(560 ppm) CO2. For hourly ozone
concentrations during experimental
ozone treatment, see Kubiske and
Foss (2015)
Bacterial communities had higher richness
under elevated O3 than under ambient O3.
Elevated O3 increased the ratio of
Basidiomycota:Ascomycota abundance (fungal
mushroom-producing taxa) based on fungal
rRNA clone libraries. O3 did not affect the
functional composition of the bacterial or fungal
soil communities.
Haesler et al. (2014) Mesocosm; soil
from a mixed
beech/spruce stand,
Eurasburger forest,
Augsburg, Germany
(48.30°N, 11.08°E)
Soil actinobacterial
communities associated
with Fagus sylvatica
(European beech)
characterized byT-RFLP
and clone libraries of
actinobacteria-specific 16s
rRNA primers and type 2
polyketide synthases (PKS)
Ambient O3 (range 20-80 ppb) and Elevated O3 altered actinobacterial community
elevated O3 (twice ambient
concentrations not exceeding
150 ppb)
composition as measured by T-RFLP in
summer but not in spring or fall. Elevated O3 did
not affect actinobacterial evenness or functional
diversity (PKS).
8-153
-------
Table 8-20 (Continued): Terrestrial community composition response to ozone exposure.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects
Kasurinen et al.
(2017)
FACE; Ruohoniemi
FACE, Kuopio,
University of
Eastern Finland,
Finland (62.88°N,
27.62°E)
Bacterial and fungal
communities subsisting
Betula pendula (silver birch)
leaf litter from clones of two
distinct birch genotypes
(gt14 and gt15). Microbial
abundance assessed by
qPCR (bacterial primers pE
and pF', fungal primers ITS3
and ITS4) at leaf drop,
217 days of decomposition,
and 257 days
decomposition
Factorial O3 by temperature
treatment: mean summer ambient
O3 is 24.2 ppb in 2009, 30.7 ppb in
2010 (AOT40 = 0.14 ppm-h in 2009,
1.1 ppm-h in 2010); mean elevated
O3 is 33.6 ppb in 2009, 43.7 ppb in
2010 (AOT40 = 4.7 ppm-h in 2009,
10.7 ppm-h in 2010); temperature
treatment is ambient or elevated by
infrared rings above the canopy
Elevated O3 increases bacterial abundance
196% (p = 0.002) and fungal abundance 61%
(p = 0.095) in freshly fallen litter. Effects of
elevated O3 on fungal abundance via changes
to soil microbial conditions varied with birch
genotype: in birch gt14 leaves, elevated O3
reduced fungal abundance 24%, and in birch
gt15 leaves, elevated O3 increased fungal
abundance 44%. Ozone Interactions: effects of
elevated O3 on microbial abundance during
decomposition via changes in litter quality only
occurred in interactions with birch genotype,
warming, and decomposition stage (p <0.1).
Effects of elevated O3 on bacterial abundance
via changes to soil microbial conditions only
occurred in interactions with birch genotype and
decomposition stage (p <0.1).
Rasheed et al.
(2017)
FACE; Ruohoniemi
FACE, Kuopio,
University of
Eastern Finland,
Finland (62.88°N,
27.62°E)
Pinus sylvestris (Scots pine)
seedlings and their
root-colonizing fungi
(quantified as root
ergosterol) and rhizosphere
soil microbial community
(quantified by PLFA)
O3 in full factorial design with air
warming treatment (+1°C), N
fertilization (+120 kg N/ha-yr), and
insect herbivore treatment (+4 larval
sawfly, Acantholyda posticalis). O3
exposure 2011 -2013: ambient O3
monthly averages during the
growing seasons 17.7-38.3 ppb O3
(AOT40 = 0.25 ppm-h in 2011,
0.52 ppm-h in 2012, and 1.27 ppm-h
in 2013); elevated O3 monthly
averages during the growing
seasons 26.7-55.0 ppb O3
(AOT40 = 6.29 ppm-h in 2011,
9.58 ppm-h in 2012, and
29.28 ppm-h in 2013)
In 2013, elevated O3 increased the extent of
fungal colonization in pine roots (measured as
ergosterol) with no main effect on mycorrhizal
colonization or soil fungi:bacteria.
8-154
-------
Table 8-20 (Continued): Terrestrial community composition response to ozone exposure.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects
Uedaet al. (2016)
Greenhouse; soil
from Meckenheim,
Germany (50.6°N,
7.0°E)
Bacterial communities
assessed by 16S rRNA
sequencing, bacteria on rice
leaf surfaces (phyllosphere)
and rice root surfaces
(rhizoplane)
Ozone fumigation for 30 days,
900-1,600, beginning with
10-week-old rice plants: control
(5 ± 4 ppb O3), elevated
(85 ± 34 ppb O3)
Elevated O3 did not affect diversity (inverse
Simpson index), richness, evenness, or
functional diversity of phyllosphere bacteria or
rhizoplane bacteria. The genetic variance of
phyllosphere bacteria was higher in elevated O3
than control O3 (HOMOVA, p = 0.021);
although, O3 did not affect relative abundance
of the most abundant phyllosphere bacterial
OTUs. Elevated O3 decreased the relative
abundance of two of the most abundant
rhizoplane bacterial OTUs, Rhodospirillaceae
(nonsulfur photosynthetic bacteria) and
Clostridiales (obligate anaerobe).
Ebanvenle et al.
(2016)
FACE; Aspen
FACE, Rhinelander,
Wl (49.675°N,
89.625°E)
Wood-decaying
basidiomycete fungi
(identified by sequencing of
primer pair ITS1 f/ITS4),
growing on logs cut in 2009
from Populus tremuloides
(quaking aspen) and Betula
papyrifera (paper birch).
Logs were grown and
placed in each of the four
FACE treatments in a full
factorial design
Fumigation 1998-2008 during
daylight hours of the growing
season. Ambient O3 W126
2.1-8.8 ppm-h and elevated
12.7-35.1 ppm-h; elevated CO2
515-540 ppm, ambient avg
374 ppm. For hourly ozone
concentrations during experimental
ozone treatment, see Kubiske and
Foss (2015)
There was no statistically significant effect of O3
on basidiomycete community composition,
although there were fewer fungal species from
logs in the O3 plots than in the ambient O3 plots.
8-155
-------
Table 8-20 (Continued): Terrestrial community composition response to ozone exposure.
Study Type and
Study Location Study Species Ozone Exposure Effects
Haves et al. (2011)
Greenhouses near
Two communities: four
Eight treatments: (1) seasonal 24-h Grass cover increases linearly with increasing
Marchlyn Mawr,
plants of forb Leontodon
mean 21.4 ppb (12-h mean seasonal O3 mean.
U.K.
hispidus and three plants of
21.1 ppb, daylight
grass Dactylis glomerata,
[7:00 a.m.-6:00 p.m.]
four plants of forb
AOT40 = 0.07 ppm-h, 24-h
Leontodon hispidus and
AOT40 = 0.07 ppm-h); (2) seasonal
three plants of grass
mean 39.9 ppb (12 h = 39.2 ppb,
Anthoxanthum odoratum
daylight AOT40 = 4.93 ppm-h, 24-h
AOT40 = 10.91 ppm-h);
(3) seasonal mean 50.2 ppb
(12 h = 49.6 ppb, daylight
AOT40 = 21.44 ppm-h, 24-h
AOT40 = 41.29 ppm-h);
(4) seasonal mean 59.4 ppb
(12 h = 58.7 ppb, daylight
AOT40 = 38.04 ppm-h, 24-h
AOT40 = 72.19 ppm-h);
(5) seasonal mean 74.9 ppb
(12 h = 73.3 ppb, daylight
AOT40 = 62.49 ppm-h, 24-h
AOT40 = 119.82 ppm-h);
(6) seasonal mean 83.3 ppb
(12 h = 81.6 ppb, daylight
AOT40 = 77.13 ppm-h, 24-h
AOT40 = 147.42 ppm-h);
(7) seasonal mean 101.3 ppb
(12 h = 99.0 ppb, daylight
AOT40 = 108.43 ppm-h, 24-h
AOT40 = 206.70 ppm-h);
(8) seasonal mean 102.5 ppb
(12 h = 100.5, daylight
AOT40 = 112.47 ppm-h, 24-h
AOT40 = 214.34 ppm-h)
8-156
-------
Table 8-20 (Continued): Terrestrial community composition response to ozone exposure.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects
Pavne et al. (2011)
Gradient; 64 acidic
grassland sites
stratified by N
deposition and
climate, U.K.
Entire plant community,
acidic grasslands
Site-specific O3 exposures from the
U.K. Air Pollution Information
System
In terms of single factor models, O3 is the
strongest predictor of species cover. Within the
multiple-factor model, only current total
inorganic N deposition and mean annual total
evapotranspiration are stronger predictors of
species cover than O3. Cluster analysis
identifies a change in species composition at an
AOT40 of 3,150 ppb-h. Nardus stricta (grass),
Deschampsia flexuosa (grass), and Juncus
effusus(sedge) are indicator species for I0W-O3
sites; Pseudoscleropodium purum (moss),
Festuca rubra (grass), and Dicranum scoparium
(moss) are indicator species for high-C>3 sites.
O3 affects species composition but not species
richness across the gradient.
Bassin et al. (2013)
FACE; Alp Flix, Sur,
Switzerland (9.65°N,
46.53°E)
Pasture turf: 107 vascular
plant species: 84 forbs,
11 grasses, 6 legumes,
6 sedges. Initial and control
community dominated by
Nardus stricta, Carex
sempervirens, and Festuca
spp., which together on
average comprise 35%
cover in plots
Ambient (mean during growing
season 45-47 ppb O3); elevated
(120% ambient O3), high elevated
(160% ambient O3) fumigated 24 h
April to October, 2004-2010.
Crossed with an N addition
experiment: ambient N deposition
4 kg N/h/yr, +5 kg, +10 kg, +25 kg,
+50 kg N/ha/yr
Plant diversity of mesocosms was high and
varied among mesocosms before the
experiment started, and analyses do not
account for initial conditions. Elevated and
highly elevated O3 had no effect on biomass of
functional groups (i.e., relative abundance of
forbs, sedges, or grasses; legumes not
included) across all years. Elevated O3 (120%
ambient) increased N. stricta abundance 22%,
and highly elevated O3 (160% ambient)
increased N. stricta abundance 40%, in the last
3 yr of the study (2008-2010).
Wedlich et al. (2012)
High Keenley Fell;
northern England,
U.K. (approximately
54.9°N, 2.3°W)
Restored and managed
mesotrophic grassland with
47 plant species (grasses,
herbs, legumes), dominated
by Festuca rubra, Holcus
lanatus, and Anthoxanthum
odoratum
Ambient: annual maximum monthly
mean was 45 ppb; moderately
elevated: annual maximum monthly
mean was 50 ppb (June-August
ambient +4 ppb in 2008 and
ambient +3 ppb in 2009); elevated:
annual maximum monthly mean
was 65 ppb (June-August ambient
+ 14 ppb in 2008 and ambient
+8 ppb in 2009)
In 2008, O3 explained 9.5% of the variation in
herb and legume species biomass composition
(p = 0.01), and in 2009, O3 explained 40.3% of
the variation in herb and legume species
biomass composition (p = 0.002).
8-157
-------
Table 8-20 (Continued): Terrestrial community composition response to ozone exposure.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects
Calvete-Soao et al.
(2016)
OTC; La Higueruela
agricultural research
farm, Toledo, Spain
(40.05°N, 4.43°W)
Pasture comprised of six
annual plants: three
legumes (Trifolium striatum,
Trifolium cherleri, and
Ornithopus compressus);
two grasses (Briza maxima
and Cynosurus echinatus);
and the herb Silene gallica
Fumigation for 39 days starting in
early April: charcoal-filtered air with
maximum of daily mean 29 ppb O3
(AOT40 = 3 ppb-h); ambient with
maximum of daily mean O3 42 ppb
(AOT40 = 760 ppb-h); moderate O3
with maximum of daily mean 61 ppb
(AOT40 = 5,771 ppb-h); and
elevated O3 with maximum of daily
mean 73 ppb
(AOT40 = 10.316 ppb-h)
O3 was a more powerful explanatory factor than
N addition in redundancy analysis of
aboveground biomass (O3 explained 8.4% of
total variability, p = 0.027), total living biomass
(O3 explained 11.1% of variability, p = 0.007),
and senesced biomass (O3 explained 10.6% of
variability, p = 0.012) of the pasture plant
community. O3 interactions: O3 and N
interactions did not have significant effects on
whole-community metrics.
Gilliland et al. (2016)
OTC; research site
located ~5 km north
of Auburn University
campus
Trifolium repens (white
clover) and three grass
species pooled into
"grasses" (Loiium
arundinacea, Paspaium
diiatatum, Cynodon
dactyl on)
Exposure for 4 mo with the mean
12-h O3 concentration of 31 ppb
(NF) and 56 ppb (2x ambient),
average peak O3 = 39 ppb (NF) and
77 (2x), peak avg 1-h O3 = 73 (NF)
and 155 (2*), 12-h AOT40
1.8 ppm-h (NF) and 29.8 ppm-h
(2x), seasonal 12-h W126
1.6 ppm-h (NF) and 42.5 ppm-h
(2x ambient)
Elevated O3 increased primary growth of
grasses (dry matter yield) 19%. In mowed
pasture, elevated O3 decreased clover yield
60% and increased grass yield 40%. There was
no effect of ozone on cover of clover or grass.
van Goethem et al.
(2013)
Meta-analysis;
northwestern
Europe (mapping of
sensitivity is for a
square area of 50 to
61°N, and 11°Eto
11°W)
25 annual grassland
species, 62 perennial
grassland species, 9 tree
species
OTC, FACE, or solardomes. All
experimental treatments were at
>40 ppb for at least 21 days, with
mean hourly O3 never exceeding
100 ppb. Control treatments were
charcoal-filtered air or ambient air
Annual grassland species were significantly
more sensitive to O3 >40 ppb than were
perennial grassland species. Mean 10%
reduction in biomass occurred at 0.84 ppm-h for
annual species and 1.14 ppm-h for perennial
grassland species. Exposure-response
relationships for 96 European plants (biomass
reduction vs. AOT40) are listed.
8-158
-------
Table 8-20 (Continued): Terrestrial community composition response to ozone exposure.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects
Wang et al. (2014)
Plant growth
chambers with soil
from a wheat-snap
bean agricultural
rotation; Haidian
District, Beijing City,
China
Soil microbial communities
assessed by PLFA on soil in
mesocosms of Phaseolus
vulgaris (snap bean, one
sensitive and one tolerant
genotype), either inoculated
or not with Glomus
aggregatum (an arbuscular
mycorrhizal fungi)
60-day O3 exposure: ambient,
20 ± 5 ppb (AOT40 = 0); elevated,
70 ± 10 ppb (AOT40 = 19.6)
Elevated O3 decreased root mycorrhizal
colonization 44% in sensitive genotype and
24% in tolerant genotype. Elevated 63 changed
soil microbial community structure of both bean
genotypes based on PCA of PLFA data.
Berqmann et al.
(2017)
Meta-analysis;
global;
peer-reviewed
papers, book
chapters, reports,
and conference
proceedings
published 1980 to
unspecified mid
2010s
Seed-bearing plants:
Grouped into herbaceous
plants (298 species in
47 plant families: wild native
or pasture species), woody
plants (165 species,
39 families, 69 genera),
crops (agricultural or
horticultural). Also assessed
ferns, mosses, lichens,
vertebrates
Multiple study designs, grouped into
experimental O3 exposures (growth
chambers where O3 did not exceed
100 ppb, greenhouse, solardome,
OTC, FACE) or ambient gradient O3
exposures for vascular plants
Among herbaceous plant families with at least
10 species tested, O3 sensitivity to foliar injury
is Onagraceae > Fabaceae > Cyperaceae >
Lamiaceae > Asteraceae > Poaceae. Among
135 woody plant species tested, ozone causes
foliar injury in 86% of broadleaf and 72% of
conifer species. In field and gradient O3
observations, 245 plant species and
28 genus-level plant groups experience O3
foliar injury. 47.5% of the 223 herbaceous plant
species experimentally exposed to O3
experience effects in growth, productivity, C
allocation, or reproduction. These effects are
more common across annuals/biennials than
perennials. 70% of Fabaceae species tested
are sensitive to these effects of O3. Among
woody plant species, 53% of the conifer and
67% of the broadleaf species experimentally
exposed to O3 experience effects in growth,
productivity, C allocation, or reproduction. The
woody plant families Myrtaceae, Oleaceae,
Salicaceae, and Betulaceae are particularly
sensitive to foliar injury and growth effects of
O3. O3 effects have been tested on 2 fern
species, 10 moss species, and 31 lichen
species; there are 63 effects at physiological
scales.
8-159
-------
Table 8-20 (Continued): Terrestrial community composition response to ozone exposure.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects
Pavne et al. (2017) Mesocosm; peat
sampled from wet,
heath peatland, U.K.
Microscopic algae
(desmids, diatoms),
protozoa (ciliates,
flagellates, testate
amoebae), and microscopic
animal consumers
(nematodes, rotifers)
sampled from Sphagnum
papillosum stems
Experimental O3 for 3.5 yr: ambient Testate amoeba community structure was
air (25 ppb O3), low O3
(ambient + 10 ppb for 24 h/day),
moderate O3 (ambient + 25 ppb O3
24 h/day), elevated O3
(ambient + 35 ppb 8 h/day in
summer, +10 ppb rest of year)
significantly affected by ozone. Moderate and
elevated O3 decreased testate amoeba species
richness 31%. Low and elevated O3 increased
grouped flagellate and ciliate abundance.
Exposure indices: authors indicated that O3
effects on microscopic food web in peat
generally start at moderate O3 exposures.
15N = nitrogen-15, stable isotope of nitrogen; 16s rRNA = bacteria-specific primer; 18s rRNA = fungi-specific primer; AOT40 = seasonal sum of the difference between an hourly
concentration at the threshold value of 40 ppb, minus the threshold value of 40 ppb; C = carbon; C02 = carbon dioxide; EMF = ectomycorrhizal fungal; FACE = free-air C02
enrichment; kg/ha = kilograms per hectare; Lsu rRNA = fungi-specific primer; N = nitrogen; N/h/yr = kg nitrogen per hectare per year; NF = nonfiltered air; 03 = ozone;
OTC = open-top chamber; OTU(s) = operational taxonomic unit(s); ppb = parts per billion; ppm = parts per million; PR protein = pathogenesis related protein; pufM = primer for
nonsulfur bacteria; qPCR = quantitative polymerase chain reaction; T-RFLP = terminal restriction fragment length polymorphism; W126 = cumulative integrated exposure index with a
sigmoidal weighting function.
8-160
-------
8.11 Water Cycling
In the 2013 Ozone ISA, the evidence was sufficient to conclude there is a likely to be causal
relationship between ozone exposure and the alteration of ecosystem water cycling (U.S. EPA. 2013).
Plants are responsible for a part of ecosystem water cycling through root uptake of soil moisture and
groundwater, as well as transpiration through leaf stomatato the atmosphere; changes to this part of the
water cycle may, in turn, affect the amount of water moving through the soil, running off overland or
through groundwater, and flowing through streams.
Ozone can affect water use in plants and ecosystems through several mechanisms, including
damage to stomatal functioning and loss of leaf area (Figure 8-2). which may affect plant and stand
transpiration. Although the 2013 Ozone ISA found no clear universal consensus on leaf-level stomatal
conductance response to ozone exposure, many studies reported incomplete stomatal closure and loss of
stomatal control in several plant species, which result in increased plant water loss [Section 9.4.5; U.S.
EPA (2013)1. Additionally, ozone has been found to alter plant water use through decreasing leaf area
index, accelerating leaf senescence and causing changes in branch architecture, which can significantly
impact stand-level water cycling. Some key studies attempted to scale up these effects of ozone on leaf
physiological measurements to ecosystems, connecting increased plant water loss to changes in soil
moisture, runoff, and streamflow, both through empirical study and modeling (Paoletti and Grulke. 2010;
Felzer et al.. 2009; Mclaughlin et al.. 2007a; Mclaughlin et al.. 2007b; Hanson et al.. 2005).
As described in the PECOS tool (Table 8-2). the scope for evidence reviewed and assessed
includes studies on any continent in which alterations in water acquisition and hydraulic transport (based
on plant structural changes), stomatal response, and plant water use were measured on the scale of
individual plants in response to ambient exposures and experimentally elevated ozone exposures within
an order of magnitude of recent concentrations. Many metrics are used to evaluate effects of ozone on
ecosystem water cycling including stomatal conductance, sap flow, vessel size and density, soil moisture,
and stream flow. The evidence presented here also includes studies from any continent where models
(both empirical statistical models and mechanistic models) were developed and assessed to examine
ozone effects on plant water use and ecosystem water cycling.
8.11.1 Structural Changes in Plants
In addition to well-documented ozone-mediated declines in leaf area and longevity, new evidence
identifies a relationship between ozone and changes in wood anatomy associated with water transport.
Additional studies find ozone alters plant biomass allocation by decreasing root growth and density,
which may result in lower drought tolerance and changes to soil moisture and runoff (Table 8-19). Both
alterations are important mechanisms for ozone effects on ecosystem water cycling.
8-161
-------
• Recent results from the long-term Aspen FACE experiment show ozone causes significant
changes in wood anatomy (along with changes in leaf area index and longevity reviewed in the
2013 Ozone ISA) and vessel architecture. Ozone-exposed trees had more and narrower vessels
which were packed more densely per unit wood area, indicating that trees prioritized hydraulic
safety over water transport efficiency. These developmental shifts in wood anatomy are one
mechanism for changes in tree water use efficiency, and thus, ecosystem water cycling
(Kostiainen et al.. 2014).
• Because plants rely on their root systems for water uptake, shifts in carbon allocation away from
roots can significantly alter water cycling. In a study by Haves et al. (2012a). Dactylis glomerata,
previously thought to be a species insensitive to ozone, showed increasing sensitivity with
increasing ozone concentration as seen by large reductions in root biomass, with a 50% reduction
between highest and lowest ozone treatments. Ozone was found to shift biomass allocation away
from roots in several other studies (Grantz et al.. 2016; Fiscus et al.. 2012; Rhea and King. 2012;
Calatavud et al.. 2011). and changes in root biomass in response to ozone exposure seems to be
species specific.
8.11.2 Impaired Stomatal Function
Ozone-mediated impairment of stomatal function has been documented for decades (Keller and
Haslcr. 1984). although impairment seems to be species specific, and the extent of its prevalence is not
clear. Studies continue to show reduced sensitivity of stomatal closing in response to various factors
(light, vapor pressure deficit, temperature, soil moisture) when exposed to ozone ("sluggish stomata") in a
number of species (Table 8-2IV
• A meta-analysis synthesized studies of ozone effects on stomatal response in 68 species
(including trees, crops and grassland); 10% showed a sluggish stomatal response to elevated
ozone, 24% showed an increased stomatal opening under elevated ozone, and 44% displayed
stomatal closure in response to ozone. Trees were the most adversely affected, with 73% showing
an altered stomatal response. Four tree species exhibited sluggish stomata and 13 showed
stomatal opening in response to ozone (Mills et al.. 2016; Mills et al.. 2013). This study provides
a comprehensive look at prevalence of ozone impairment to stomatal functioning across multiple
plant species and growth forms.
• Under increased ozone, a sluggish response of stomata was observed in a growth chamber
experiment using poplar trees (three genotypes of a Populus deltoides / Populus nigra hybrid) in
reaction to changes in light intensity, CO2 concentration, and vapor pressure deficit. The speed of
the responses varied by genotype and appeared to explain some of the genotype-related
sensitivity seen in poplar trees (Dumont et al.. 2013).
• An ozone-FACE experiment in Japan shows leaves of Siebold's beech (Fagus crenata) grown in
elevated ozone took a significantly longer time to close stomata (+27 and +73%, in August and
September, respectively) and slower rate of decrease in stomatal conductance ( -26 and -64%)
than leaves of trees grown in ambient conditions in response to decreasing light (Hoshika et al..
2012b). Models of transpiration created for this data set were found to better fit the data when
stomatal conductance was adjusted for ozone exposure (Hoshika et al.. 2012a). Reduced stomatal
sensitivity was also reported for Betula platyphylla var. japonica at this FACE site (Hoshika et
al.. 2018). as well as Ailanthus altissima, Fraxinus chinensis, and Platanus orientalis in OTC
experiments in China (Hoshika et al.. 2014).
8-162
-------
• A study of two grassland species shows that ozone causes a lack of stomatal sensitivity to
changes in environmental conditions. The widely distributed grassland species Dactylis
glomerate consistently exhibited reduced sensitivity of the stomatal closing response to all the
environmental parameters studied (light, vapor pressure deficit, temperature, soil moisture) in
elevated ozone treatments. In a co-occurring species, Ranunculus acris, stomatal conductance
was a found to be less responsive to light, vapor pressure deficit, and temperature under high
ozone (Wagg et al.. 2013). In another study of Dactylis glomerate, elevated ozone caused
stomatal insensitivity to drought conditions between 3 and 9 weeks. Lack of stomatal control was
shown on leaves that developed after a midseason harvest, implying that the results seen are not
due to long-term exposure damage to leaf architecture, but develop on the plant under adverse
conditions (Haves et al.. 2012a).
• Sluggish stomatal response have also been reported in ozone sensitive Phaseolus vulgaris
[snapbean; Hoshikaetal. (2016)1. but not in Glycine max (Bernacchi et al.. 2011).
• Finally, numerous studies outside of the scope of this review have considered ozone effects on
leaf-level physiological processes, particularly photosynthetic and gas exchange measurements,
and the biochemical mechanisms of ozone response (U.S. EPA. 2009).
8.11.3 Models of Plant Water Use
A few studies have attempted to assess ozone effects on plant water demands across large regions
using various modeling methods; significant differences exist between estimations generated by models
that assume stomatal closure with ozone exposure and those which account for stomatal impairment from
ozone.
• A conceptual model of leaf atmospheric boundaries developed using data collected in a pasture
from C3 grasses to assess changes in evapotranspiration estimated that ozone reduced maximum
evapotranspiration by 7.5% (Super etal.. 2015). Grasses are relatively ozone tolerant, and
stomatal closure may help limit ozone effects (Section 8.10).
• Using the Community Land Model, Lombardozzi et al. (2015) estimated that present-day ozone
exposure reduces transpiration globally by 2-2.4%. Larger reductions in GPP compared to
transpiration decreased water-use efficiency 5-10% in the eastern U.S., and increased surface
runoff more than 15% in eastern North America. However, uncertainties arise when trying to
estimate transpiration over large geographical regions consisting of different species with a
simple transpiration function. Additionally, this study did not consider stomatal sluggishness
which could modify the transpiration results.
• Models that do not incorporate sluggish stomatal response may significantly underestimate plant
water loss. When accounting for it, transpiration decreases until 30 ppb ozone and then increases
with increasing ozone exposure. A significant part (10%) of the water use efficiency (WUE) at
North American sites may be explained by ozone exposure with ozone-induced stomatal
sluggishness. The contribution of ozone to declines in WUE is estimated at 4.5 to 8.8 % in
different regions of the Northern Hemisphere fHoshika et al. (2015); Figure 8-131.
8-163
-------
0
-*
-s
-12
I -1ft
-g 10
a
Si
1 0
u
_3
2 -ia
«>
¦g
q -20
&
e
J 0
-10
-10
•30
-40
10 20 SO 40 50 0 30 60 90 120
Daytime O, cnnKeWralion (mnoJ mo) I Canopy cumulative 0-. uptake (Cmmot m ;)
Note; Effects of ozone-induced stomatai sluggishness were included (black open circles and red lines) or excluded (gray circles and
gray lines). The percentage of change of each parameter was calculated relative to "control run" (no ozone effect).
Source: Reprinted with permission from the publisher, adapted from Hoshika et al. (2015).
Figure 8-13 Percentage change of modeled net carbon dioxide (CO2)
assimilation, transpiration, and water use efficiency in temperate
deciduous forests in the Northern Hemisphere in relation to
daytime mean ozone concentration or cumulative canopy ozone
uptake (years 2006-2009). (a) Net CO2 assimilation,
(b) transpiration, and (c) water use efficiency were simulated by
the offline coupling simulation of SOLVEG-MRI-CCM2.
8.11.4 Ecosystem Water Dynamics
New work examines the influence of environmental measures, inclusive of ozone exposure and
climate, on late-season stream flow in forests in the eastern U.S. and shows that ozone effects scale up
from leaf level through to ecosystem level. The 2013 Ozone ISA reviewed the work of Mclaughlin et al.
(2007a); Mclaughlin et al. (2007b). which used field measurements to link ozone to changes in tree sap
flow and scale up to the ecosystem level. Building on this, Sun et al. (2012 built empirical statistical
models from data collected in six watersheds in Tennessee, North Carolina, Virginia, and West Virginia
a CO, assimilation
y--ftW**|01 JP-n42 0
*» >--a.23x*l -MS 0
V
v- 17
JP~0 »
m yfl 15* ' 1 -1* R- O JJi
_ O a
c Water use efficiency «
>**-0.1 Ix- 1.44 iT'-Ck 0
°0 0
O
• JJMI. W
8-164
-------
and found that ozone and climate are both significant predictors of late season stream flow in forests;
these predictor variables were also significant when applied to measurements of tree radial growth.
Findings from this study support the assertion that ambient ozone concentrations in Appalachian forests
decrease efficiency of tree water use through lowered stomatal control, which in turn, reduces streamflow
in forested watersheds. When statistical models were partitioned to examine the contribution of ozone and
climate variables to predictions of streamflow, Sun et al. (2012) also found statistically significant
negative interaction effects between climate and ambient ozone levels that resulted in a net decrease in
late season streamflow.
8.11.5 Drought and Ozone
Several studies have tested the interactive effects of ozone and drought on plant stomatal
function, water use, growth, and performance; these are discussed in the section on modifying factors
(Section 8.12).
8.11.6 Summary and Causality Determination
During the review for the 2013 Ozone ISA, the widely held assumption that ozone exposure
consistently reduced stomatal conductance in plants was being challenged. Several studies found
increased conductance, suggesting stomatal dysfunction in response to ozone exposure; other studies
found ozone caused a loss of stomatal control, incomplete stomatal closure at night, and a decoupling of
photosynthesis and stomatal conductance. The relationship of stomatal response to ozone exposure
continues to be an active area of research. There is mounting biologically relevant, statistically
significant, coherent, and cohesive evidence from multiple studies of various types about the mechanisms
of ozone effects on plant water use and ecosystem water cycling (reduced leaf area, reduced leaf
longevity, changes in root and branch biomass and architecture, changes in vessel anatomy, stomatal
dysfunction, reduced sap flow). Additionally, there are a few strong studies which scale up these changes
to effects at the ecosystem level and show significant effects. The most compelling evidence is from six
watersheds in eastern forests and from Aspen FACE. This new information supports and strengthens the
conclusions of the 2013 Ozone ISA. The body of evidence is sufficient to conclude that there is a
likely to be causal relationship between ozone exposure and the alteration of ecosystem water
cycling.
8-165
-------
Table 8-21 Ozone exposure and water cycling.
Study
Study Type and Location Study Species
Ozone Exposure
Effects on Water Cycling
Hoshika et al. (20111
Lab study; location not
stated
Phaseolus vulgaris,
S156 (snap bean)
Short-term 1 h, low (48 ppb),
middle (87 ppb), high (150 ppb),
and control (0 ppb)
Steady-state stomatal conductance decreased by
27% in low O3 and 75% in high O3 under
well-watered conditions. There was no effect of O3
treatment in water-stressed conditions. High O3
exposure caused higher nocturnal stomatal
conductance than control plants under well-watered
conditions.
Fiscus et al. (2012)
USDA-ARS Plant
Science Research Unit
field site 5 km south of
Raleigh, NC
Two genotypes of
Phaseolus vulgaris
(snap bean)
Two O3 concentrations
(charcoal-filtered air) dispensed
into outdoor chambers (12-h
mean of 0 and 60 ppb).
Exposures started 18 days after
planting at 1/3 target
concentrations and increased to
full exposure at 21 days after
planting. Experiment ran
62 days. For elevated O3 daily
AOT40 = 245, SUM06 = 534,
W126 = 295 ppb-h. There were
also two vapor pressure deficit
(VPD) levels tested (1.26 and
1.96 kPa)
In low VPD treatment with elevated O3, daily water
use significantly increased 23 to 38%.
Grantzet al. (2016)
Greenhouse; Parlier,
(36.60°N, 119.50° W)
CA Gossypium
barbadense (Pima
cotton)—O3 sensitive
cultivar
Approximate 12-h mean O3
concentrations were 4, 59, and
114 ppb, with peak
concentrations at solar noon
Midday stomatal responses at are not representative
of the morning or evening. Lowest responsivity was
observed during periods of rapid stomatal movement
in the morning and evening. Maximum responsivity
corresponded to previously determined maximum
plant sensitivity to short-term pulse exposures in the
cotton plants, not with maximum gas exchange active
regulation. A clear diel pattern emerged with stomatal
responsivity increasing in the early morning through
midafternoon then decreasing in the early evening.
8-166
-------
Table 8-21 (Continued): Ozone exposure and water cycling.
Study
Study Type and Location Study Species
Ozone Exposure
Effects on Water Cycling
Paudel et al. (2016) Greenhouse; Parlier, CA Amaranthus palmeri Two runs of exposure 30 and
(36.60°N, 119.50°W) (Palmer amaranth) 35 days. 12-h means of 4, 59,
and 114 ppb
Elevated O3 exposure and water stress had no effect
on the daytime stomatal conductance, shoot growth,
and root growth. This agricultural weed species may
be more tolerant to elevated O3 and moisture stress
than crop species.
Vanloocke et al.
(2012)
FACE; SoyFACE,
Champaign, IL
Glycine max 12-h means of 40, 46, 54, 58,
(soybean) 71, 88, 94, 116 ppb
With increasing ozone treatment, yield (-64%),
canopy evapotranspiration (-26%), and water use
efficiency (-50%) decreased. The sensible heat flux,
water use efficiency, and canopy temperature
increased linearly. These results indicate that O3
could alter meteorological conditions through warmer
surface temperatures and perturb the hydrologic
cycle via decreased water vapor release to the
atmosphere.
Hoshika et al. (2016)
Greenhouse; location not
stated
Phaseolus vulgaris
(snap bean) ozone
sensitive genotype
S156
Elevated O3 exposure level
149 ± 3 ppb, control 3 ± 1 ppb
Elevated O3 induced stomatal sluggishness only
under high light intensity (1,500 pmol/m-s); stomata
needed 53% more time to half Gs under high
light x elevated O3.
Bernacchi et al.
(2011)
FACE; SoyFACE,
Champaign, IL
(40.056°N, 88.201°W)
Glycine max
(soybean)
2002-2006; 8-h max (ppb):
ambient = 35-55,
elevated = 46-68; AOT40
(ppm-h): ambient = 3-35,
elevated = 25-65; SUM06
(ppm-h): ambient = 4-21,
elevated = 15-39
Elevated O3 reduced evapotranspiration for four of
five growing seasons. O3 decreased seasonal water
use by 12% in 2002, 14% in 2003, 13% in 2005, and
11% in 2006. In 2004, there was no significant effect
of O3. Under elevated O3, canopy temperatures were
consistently warmer. The results suggest that future
increased O3 exposure could lead to alterations in the
local and regional hydrologic cycles in areas of high
intensity soybean cultivation.
8-167
-------
Table 8-21 (Continued): Ozone exposure and water cycling.
Study
Study Type and Location Study Species
Ozone Exposure
Effects on Water Cycling
Sun et al. (2012)
Gradient; six watersheds:
Walker Branch and Littler
River (eastern
Tennessee),
Cataloochee Creek
(western North Carolina),
James River and New
River (Virginia), and
Fernow (West Virginia)
Appalachian mixed
deciduous forests
AOT60 at each watershed: 1.72
(WBWS), 2.6 (LR), 1.72 (CC),
0.82 (NR), 0.83 (JR), 0.74
(FEW); max hourly (in ppb):
68.2 (WBWS), 67.8 (LR), 68.2
(CC), 59.4 (NR), 58.7 (JR), 58.8
(FEW)
O3 and climate are both significant predictors of late
season stream flow, regardless of the seasonal
timescale used for these parameters. Models
incorporating O3 and climate capture the variation
and magnitude of stream flow, and also fit annual tree
ring growth (an important mechanistic step in O3
effects on forested watersheds). Models generated
from data from southern watersheds in Tennessee
(where O3 levels are higher) have better predictive
power throughout the study area than those in the
north. Ambient O3 concentrations in Appalachian
forests decrease efficiency of tree water use through
lowered stomatal control and that reduces streamflow
in forested watersheds.
Kostiainen et al.
(2014)
FACE; Aspen FACE,
Rhinelander, Wl
Populus tremuloides
(quaking aspen)
clones, Betula
papyrifera (paper
birch)
Fumigation 1998-2008 during
daylight hours of the growing
season. Ambient O3 W126
2.1-8.8 ppm-h and elevated
12.7-35.1 ppm-h; elevated CO2
515-540 ppm, ambient avg 374.
For hourly ozone concentrations
during experimental ozone
treatment, see Kubiske and
Foss (2015)
Elevated CO2 increased radial growth and cell
diameters in aspen, while vessel density and
proportion decreased. Elevated O3 decreased growth
and cell diameters, but increased vessel density and
proportion. Neither CO2 nor O3 responses were
consistent across years. O3 exposed trees had more
and narrower vessels, which were packed more
densely per unit wood area.
Kefauver et al.
(2012a)
Gradient; Yosemite
National Park (YOSE)
and Sequoia and Kings
Canyon National Park
(SEKI), CA; Catalonia,
Spain
California: Pinus
ponderosa
(ponderosa pine)
and Pinus jeffreyi
(Jeffrey pine)
Spain: Pinus
uncinata (mountain
pine)
Passive monitors in YOSE and
SEKI colocated with one U.S.
EPA-certified active monitor per
park. Average yearly O3 mixing
ratio in 2002 ranged from 35 to
65 ppb for all YOSE and SEKI
sites. Yearly averages within
sites were 49 ppb for YOSE and
46 ppb for SEKI
Ozone Injury Index by itself was poorly correlated to
ambient O3 across all sites. Models improved when
GIS variables related to plant water status were
included (YOSE, R2 = 0.36, p < 0.001; SEKI,
R2 = 0.33, p = 0.007).
8-168
-------
Table 8-21 (Continued): Ozone exposure and water cycling.
Study
Study Type and Location Study Species
Ozone Exposure
Effects on Water Cycling
Hoshika et al. (2014) OTC; China
Ailanthus altissima
(tree of heaven),
Fraxinus chinensis
(Chinese ash), and
Platanus orientalis
(Oriental planetree)
42, 69, 100 ppb avg from
9:00 a.m. to 6:00 p.m. for 3 mo
O3 did not affect stomatal density. Elevated O3
exposure slowed stomatal dynamics in these tree
species. Time for 50% decrease of stomatal
conductance increased with increasing stomatal O3
flux.
Holmes (2014)
Gradient; U.S. and
Europe
Many trees species
O3 concentrations not given.
Trends over 1995-2010
As a result of O3 AOT40 decreasing by approximately
half (-20 to 10 in the Midwest) over the period
1995-2010, forest WUE likely increased by -0.33%
per year in the midwestern U.S. and slightly less in
the northeastern U.S.
Dumont et al. (2013) Lab; France
Three Euramerican
Populus
deltoides * Populus
nigra (poplar)
genotypes
(Carpaccio, Cima,
and Robusta)
Elevated O3 at 120 ppb for
13 h/day and charcoal-filtered
air. Treatments run for 18 days
O3 significantly decreased stomatal conductance and
photosynthesis for the three genotypes. Under
increased O3, a sluggish response of stomata was
observed in reaction to blue light intensity, CO2
concentration and VPD, and lower amplitude of the
response to variations in light intensity. Speed of
responses varied by genotype and appeared to
explain some of the genotype-related sensitivity.
Hoshika et al.
(2012b)
FACE; Sapporo
Experimental Forest,
Hakkaido University,
northern Japan
(13.067°N, 141,333°E)
Fagus crenata
(Siebold's beech)
Daytime O3: ambient = 26 ppb,
elevated = 54 ppb; AOT40:
ambient = 0.3,
elevated = 11.9 ppm-h
Leaves under elevated O3 had lower stomatal
conductance (25 and 31% in September and
October, respectively) than control leaves and had a
steeper decline in photosynthesis after September.
Leaves in elevated O3 had significantly longer time to
close stomata (+27 and +73%, in August and
September, respectively) and slower rate of decrease
of stomatal conductance ( -26 and -64%) than
leaves of trees grown in ambient conditions in
response to decreasing light.
8-169
-------
Table 8-21 (Continued): Ozone exposure and water cycling.
Study
Study Type and Location Study Species
Ozone Exposure
Effects on Water Cycling
Hoshika et al.
(2012a)
FACE; Sapporo
Experimental Forest,
Hakkaido University,
northern Japan
(13.067°N, 141,333°E)
Fagus crenata
(Siebold's beech)
The target O3 concentration was
60 ppb during daylight hours.
Mean daytime O3
concentrations were
25.7 ± 11.4 ppb (ambient) and
56.7 ± 10.5 ppb (elevated).
Fumigation was August
6-November 11, 2011
Gs under elevated O3 was lower than under ambient
conditions after 2 weeks of exposure. The ratio of
daily maximum Gs (elevated 03:ambient O3)
decreased linearly with both cumulative O3 uptake
and AOT40, although the determination of coefficient
was substantially higher with cumulative exposure
calculation (r2 = 0.67 vs.0.44). Jarvis algorithm better
fits data when Gs is adjusted for O3 exposure.
Hoshika et al. (2015) Not site-specific
(although model
parameters were taken
from FACE); Sapporo
Experimental Forest,
Hakkaido University,
northern Japan
(13.067°N, 141,333°E)
sluggishness was included, transpiration decreased
until 30 nmol/mol of O3 concentration or 37 mmol/m2
of canopy cumulative O3 uptake, and then increased
with increasing O3 exposure or uptake. O3 decreased
WUE as compared to control, with higher exposure
causing greater declines in WUE the contribution of
O3 to the decline in WUE ranged from 4.5 ± 1.9 to
8.8 ± 3.0% in different regions of the Northern
Hemisphere. When taking sluggishness into account,
authors estimated that a 8-10 ppb decrease in O3
concentrations would yield an increase of 2-3% in
WUE of temperate forests, while only a -1% increase
of WUE at the same change in O3 was found without
sluggish stomatal response.
Northern Modeled response covers a
Hemisphere range of O3
temperate forests exposures—daytime
concentrations of 15 to 55 ppb;
canopy cumulative O3 uptake
from 0 to 115 ppb
The O3 induced decline of net CO2 assimilation at the
average daytime O3 concentrations of
37.2 ± 6.2 nmol/mol was 6.6 ± 2.1% and 6.0 ± 1.8%
(incorporating sluggish stomatal response and
without). O3 further reduced CO2 assimilation at
higher concentrations. Without the inclusion of
stomatal sluggishness parameters, transpiration
showed a linear decline as O3 increased. When
8-170
-------
Table 8-21 (Continued): Ozone exposure and water cycling.
Study
Study Type and Location Study Species
Ozone Exposure
Effects on Water Cycling
Gao etal. (2017)
OTC; Seed Station Field
ofChangping, northwest
Beijing, China
O3 sensitive clone
(546) of Populus
deltoides (eastern
cottonwood)
Three O3 treatments: charcoal-
filtered (CF), ambient (ambient),
and elevated (E-O3) for 96 days
(June-September). Mean O3
was 33.5 ± 2.4 ppb in the CF
treatment, 51.1 ±4.1 ppb in the
ambient treatment, and
78.2 ± 5.5 ppb in the E-O3
treatment. Accumulated
exposures in the CF, ambient,
and E-O3 treatments expressed
as AOT40 were 4.3, 16.0, and
38.7 ppm-h, respectively. Two
irrigation treatments: drought
treatment had 51.9% less water,
while well-watered plants were
watered to capacity. Average
soil water content of
well-watered and drought
treatments was 24.8 ± 0.38%
(95% CI) and 12.8 ± 0.47%,
respectively
Elevated O3 significantly reduced total biomass, stem
diameter, stem biomass, and leaf biomass.
Interactions between O3 and water stress were
significant for leaf, stem, and total biomass of the
plants, with lower relative biomass reductions in
drought-stressed plants. Leaf senescence was also
reduced in reduced watered plants in comparison to
well-watered plants. For O3 dose-response, modeled
as biomass changes, model performance was
significantly better when using POD (flux) compared
with AOT40 (R2 = 0.829, p = 0.012 vs. R2 = 0.560,
p = 0.087). Using the flux model, the O3 critical level
(CL) for preventing a 4% biomass loss in this poplar
clone under different water regimes was between
5.27 mmol/m2 PLA and 4.09 mmol/m2 PLA.
Hoshika etal. (2018)
FACE; Sapporo
Experimental Forest,
Hokkaido University,
northern Japan
Betula platyphylla
var. japonica
(Japanese white
birch) and Quercus
mongolica var.
crispula (Mongolian
deciduous oak)
There were two plots, one with
elevated O3 (target of 60 ppb)
and one with ambient O3.
Fumigation occurred August to
November 2011, and May to
November 2012. Daytime hourly
mean O3 concentrations in
ambient and elevated O3 were
25.7 ±11.4 ppb and
56.7 ± 10.5 ppb during the
experimental period in 2011,
and 27.5 ±11.6 ppb and
61.5 ± 13.0 ppb in 2012
Elevated O3 significantly decreased white birch
stomatal conductance 28% in early summer, and
10% in late summer. Elevated O3 reduced stomatal
sensitivity of white birch to VPD and increased
stomatal conductance under low light conditions. In
contrast, no significant effects of O3 were observed in
deciduous oak.
8-171
-------
Table 8-21 (Continued): Ozone exposure and water cycling.
Study
Study Type and Location Study Species
Ozone Exposure
Effects on Water Cycling
Xu etal. (2017)
OTC; Shenyang
Arboretum, China
Greenhouse grown 30 days of O3 treatments in ppb All treatments significantly decreased stomatal size
1 yr old Lonicera
maackii (bush
honeysuckle)
(range, mean, AOT40): control
(1.2-58.3, 41.5, 165.3), drought
(0.9-62.1, 39.8, 170.9),
elevated O3 (75.4-125.0, 85.3,
12,073.5), drought * O3
(68.9-119.0, 84.9, 11,685.2);
soil water content: control
(65.9%), drought (35.6%),
elevated O3 (62.5%),
drought * O3 (38.4%)
as compared with control, and O3 significantly
decreased WUE in single and combined treatments
(about 30%).
Bohleretal. (2013)
Greenhouse study; 14-h
light period, location not
given
10 cm tall clones of
Populus tremula * P.
alba (Populus *
canescens
[poplar]—clone INRA
717-1-B4)
Factorial design of O3 by Differences in Gs were observed on Day 10, when
drought: O3 treatments: O3 x drought treatment had a lower Gs than control
charcoal-filtered air, (roughly 0.18 mmol/m2-s vs. 0.3 mmol/m2-s), with no
charcoal-filtered air + 120 ppb of significant effect of drought alone or O3 alone.
O3 for 13 h/day. Drought
treatment: maintained soil water
content at 35%
Wagg et al. (2013) Greenhouse; England
Ranunculus acris
(meadow buttercup),
Dactylis glomerata
(orchard grass)
L0WO3: 16-34 ppb seasonal
mean, high O3: 73-90 ppb
seasonal mean
For D. glomerata, the maximum stomatal
conductance increased 50% in high O3 compared to
low O3 D. glomerata grown in high O3 exhibited
reduced sensitivity of stomatal closing response to
the environmental parameters of light, vapor pressure
deficit, temperature, soil moisture. At high O3, R. acris
stomatal conductance was less responsive to light,
vapor pressure deficit, and temperature.
8-172
-------
Table 8-21 (Continued): Ozone exposure and water cycling.
Study
Study Type and Location Study Species
Ozone Exposure
Effects on Water Cycling
Haves et al. (2012a)
Greenhouse; Bangor,
North Wales, U.K.
(elevation 610 m, grid ref
SH613619)
Dactylis glomerata
(orchardgrass)
Treatments in ppb and ppb-h
(season mean, mean daily
maximum, season AOT40):
ambient-20: (16.2, 20.8, 0);
ambient: (33.9, 41.56, 2.00);
ambient+12 (44.1, 53.85,
13.95)
24.80)
44.33)
58.04)
108.4,
ambient+24 (50.7, 62.4,
ambient+36 (62.0, 80.6,
ambient+48 (72.6, 89.1,
ambient+60 (88.9,
84.65); ambient+72
(89.5, 110.7, 85.12)
At 3 weeks, there was drought-induced lowering of
stomatal conductance. After 9 weeks of elevated O3
exposure, stomatal conductance was only
significantly different between watering treatments at
low/moderate levels of increased O3. At 19 weeks,
stomatal conductance (i.e., drought response) in
newly developed leaves of reduced water plants was
only significantly lower at ambient O3. Modifying
models of cumulative O3 flux to incorporate this loss
of stomatal control results in significantly higher
values of cumulative O3 uptake in leaves. Dactylis
glomerata shows increasing sensitivity with
increasing O3, due to 50% reduction in root biomass
between highest and lowest O3 treatments. Shoot
biomass increased slightly with increasing O3.
Super et al. (2015)
Model using data
collected in Cabauw
pasture in the
Netherlands (51.91°N,
4.93°E)
Unspecified C3
grasses
Measured O3 exposures ranged
diurnally from 0 to 28 ppb
In a conceptual model of leaf atmospheric boundaries
to assess changes in evapotranspiration, O3 reduced
max evapotranspiration 7.5%.
Lombardozzi et al.
(2015)
Mode based on literature Vegetation
reviews; global
Global concentrations
2002-2009, generated by CAM
model. Global growing season
mean hourly O3 concentrations
from 2002 to 2009 ranged
approximately from 0 to 55 ppb
The model estimated that ambient O3 reduced GPP
by 8-12% and transpiration by 2-2.4% globally. GPP
and transpiration decreased as much as 20 and 15%,
respectively, in the eastern U.S., Europe, and
southeast Asia. Model did not include stomatal
sluggishness responses.
8-173
-------
Table 8-21 (Continued): Ozone exposure and water cycling.
Study
Study Type and Location Study Species
Ozone Exposure
Effects on Water Cycling
Mills etal. (2016)
Reanalysis of papers
reviewed from numerous
study locations
68 species (including
trees, crops, and
seminatural
grassland species)
Multiple studies with multiple
exposure values
Of the 68 species, 22% showed no change in
stomatal conductance, 10% showed a slowed
(sluggish) stomatal response to elevated O3, 23.5%
showed an increased stomatal opening under
elevated O3, and 44% displayed stomatal closure in
response to O3. Tree species were the most
adversely affected with 73% of species showing an
altered stomatal response, with 13 species showing
stomatal opening and 15 showing stomatal closure in
response to O3. Crops tended to respond to O3 stress
with stomatal closure (occurring in 75% of the
species), while increased or sluggish stomatal
response was only reported in 19% of the crops. For
the eight grassland species included, responses were
more or less evenly spread across the four categories
of stomatal response. There is a tendency for
stomatal opening to occur at lower concentrations.
Calatavud et al.
(2011)
OTC; Spain
Lamottea dianae, a 24-h avg (ppb): CF = 11,
Visible symptoms began in 3 days at an AOT40 of
perennial forb
endemic to the
Mediterranean
NF+30 = 40, ambient = 32; 12-h 2 ppm-h in the NF+30 treatment. Mature leaves in
avg (ppb): CF = 10,
NF+30 = 66, ambient = 46;
avg (ppb): CF = 13,
NF+30 = 74, ambient = 49;
AOT40 (ppm-h): CF = 0,
NF+30 = 36, ambient = 11
NF+30 had a 26% reduction in photosynthesis and a
25% decrease in WUE at saturating light conditions
compared to CF. NF+30 significantly reduced
belowground biomass compared to CF.
C3 = plants that use only the Calvin cycle for fixing the carbon dioxide from the air; C4 = plants that use the Hatch-Slack cycle for fixing the carbon dioxide from the air; CAM = plants
that use the crassulacean acid metabolism for fixing the carbon dioxide from the air; CF = charcoal-filtered air; C02 = carbon dioxide; FACE = free-air C02 enrichment; GPP = gross
primary production; Gs = stomatal conductance; kPa = kilopascal; NF+30 = nonfiltered air plus 30 ppb ozone; nmol/m2 = nanomole per meter squared; 03 = ozone; OTC = open-top
chamber; POD = phytotoxic ozone dose; ppb = parts per billion; ppm = parts per million; SUM06 = seasonal sum of all hourly average concentrations > 0.06 ppm;
|jmol/m2/s = micromoles per meter squared per second; VPD = vapor pressure deficit; W126 = cumulative integrated exposure index with a sigmoidal weighting function;
WUE = water use efficiency.
8-174
-------
8.12 Modifying Factors
It is important to acknowledge that ozone is just one of the environmental and anthropogenic
factors simultaneously influencing ecosystem function and that the human influence on ecosystems is
ubiquitous (Lewis and Maslin. 2015). To varying degrees, these other factors may exacerbate or negate
the effects of ozone. Research into how interactions with biotic and abiotic factors, both natural and
anthropogenic, is diverse and includes topics such as UV-B radiation (Bao et al.. 2014). pathogens (Mina
et al.. 2016; Chieppa et al.. 2015). shifts in community genetic and species composition (Mcncndcz et al..
2017; Moran and Kubiske. 2013). and land use/land cover change (Tian et al.. 2012). The degree to which
biotic and abiotic factors may modify the effects of ozone on ecosystems was comprehensively reviewed
in the 2006 Ozone AQCD and updated in the 2013 Ozone ISA. Consequently, this section focuses on
three factors that have received considerable research attention since the 2013 Ozone ISA: nitrogen
enrichment, increases in atmospheric CO2, and climate change. These topics were not systematically
reviewed for this ISA; however, some key citations are highlighted.
8.12.1 Nitrogen
Because oxidized nitrogen is a precursor to ozone formation, many ecosystems that are exposed
to chronic ozone pollution also experience elevated rates of N deposition (Fowler etal.. 1998). and the
combined effects of these two anthropogenic pollutants on plants and ecosystems have been a topic of
research for decades (Takcmoto et al.. 2001; Grulke et al.. 1998; Darrall. 1989). The 2013 Ozone ISA
reviewed several mechanisms wherein elevated N deposition might exacerbate or negate the effects of
ozone. First, because leaf-level photosynthesis is positively correlated with both foliar N concentrations
and stomatal conductance (Wright et al.. 2004) and N deposition has been linked to increased
photosynthetic capacity in some ecosystems (Fleischer et al.. 2013). greater N deposition may lead to
higher ozone flux into the leaf and further ozone damage. Conversely, because N limitation often limits
plant productivity (Yue et al.. 2016). N deposition can stimulate plant growth and NPP (Horn et al.. 2018;
Thomas et al.. 2010). overshadowing the effects of ozone on these processes. Additionally, the 2013
Ozone ISA also cited the possibility that increased photosynthesis as a result of N deposition could help
plants produce antioxidants that neutralized ozone damage.
Nitrogen deposition can also decrease plant biodiversity, eliminating rare species and favoring
species with rapid growth rates (Suding et al.. 2005). In a national analysis of the growth and mortality of
71 tree species, 53 of the species exhibited positive, negative, or unimodal (threshold) changes in growth
or mortality with increasing N deposition (Horn et al.. 2018). Among forests in the northeastern U.S., the
growth of some ozone-sensitive trees such as black cherry (Prunus serotina) and tulip poplar
{Liriodendron tulipifera) increased with greater N deposition, while the only three species showing
8-175
-------
decreased growth rates with greater N deposition were evergreen conifers, which tend to be less sensitive
to ozone (Thomas et al.. 2010). Each of these effects of N deposition on plant communities—decreased
biodiversity, greater abundance of rapidly growing plants, and the promotion of ozone-sensitive broadleaf
species at the expense of conifers—could make ecosystems more sensitive to the negative effects of
ozone. However, in a survey of grassland community composition in the U.K., Payne etal. (2011) found
that while N deposition and ozone each had effects on community composition, there were no interactions
between these effects. Similarly, Bassin et al. (2013) observed no significant interactive effects between
ozone and N for plant community composition or biomass in a 7-year factorial N addition and ozone
experiment in alpine grassland mesocosms in Switzerland. In a short-term (39 days) OTC experiment in a
Mediterranean annual grassland in Spain, Calvete-Sogo et al. (2016) observed several significant
ozone x N for the growth of individual species. In another ozone and N deposition experiment in a
Mediterranean ecosystem in Spain with the annual grass Briza maxima, the high N deposition treatment
negated the increase in leaf senescence caused by ozone in other treatments, but the increase in grass
lignin concentration caused by ozone persisted (Sanz etal.. 2011). In applying the DLEM to 20th century
C cycling in the southeastern U.S., Tian et al. (2012) found significant effects of ozone and N deposition
on NPP and C sequestration, but no interactive effects between the two.
Although gas-phase forms of N such as NOx and PAN can cause direct foliar injury and
phytotoxicity (Greaver et al.. 2012; Riddell et al.. 2012) that would be potentially additive to similar
damage caused by ozone, the most recent Integrated Science Assessment for Oxides of Nitrogen, Oxides
of Sulfur, and Particulate Matter-Ecological Criteria [Second External Review Draft; U.S. EPA (2018)1
concluded that concentrations of these gas-phase N forms in the U.S. rarely reach levels high enough to
be damaging.
8.12.2 Carbon Dioxide
The effects of elevated atmospheric CO2 on plants and terrestrial ecosystems have been well
studied over the past several decades (Norbv and Zak. 2011; Curtis and Wang. 1998). Elevated CO2
broadly stimulates photosynthesis and often increases plant growth and NPP (Norbv and Zak. 2011).
Because these effects contrast with those of ozone, it has long been hypothesized that elevated CO2 may
counteract the effects of ozone on plants (Dickson et al.. 2000). However, like ozone and N deposition,
the effects of CO2 extend beyond changes in photosynthesis and growth to include changes in other
properties and processes such as tissue chemistry (Norbv and Zak. 2011). ecosystem water use (Norbv
and Zak. 2011). and trophic interactions (Andrew et al.. 2014; Couture et al.. 2012). Further, like N
deposition, the effects of elevated CO2 are often species specific, and the shifts in plant community
composition observed under elevated CO2 can have consequences for biogeochemical cycling and other
processes (Bradley and Pregitzer. 2007). The diverse and complex effects of elevated CO2 make it
difficult to fully predict how ozone might interact with elevated CO2.
8-176
-------
Research on the combined effects of CO2 and ozone were reviewed in detail in the 2006 Ozone
AQCD and the 2013 Ozone ISA concluded that the bulk of the research to date showed that increased
CO2 could protect plants from ozone damage. In the time since the 2013 Ozone ISA, numerous new
papers have been published from the Aspen FACE and SoyFACE experiments, both of which include
elevated CO2 and ozone treatments. At SoyFACE, there was a significant interaction between CO2 and
ozone wherein elevated CO2 did not affect snap bean (Phaseolus vulgaris) pod yield, but did ameliorate
the negative effect of ozone (Burkev et al.. 2012). In addition, some elements of soil microbial
community composition also responded uniquely to the combined CO2 and ozone treatment (He et al..
2014). However, the effects of CO2 and ozone at SoyFACE were generally additive and often offsetting,
including in aspects of the agroecosystem that are less directly tied to the leaf-level physiological effects,
such as mycorrhizal community composition (Cotton et al.. 2015) and rates of soil N cycling (Decocket
al.. 2012).
As at SoyFACE, CO2 and ozone had largely offsetting effects on most ecosystem properties and
processes at Aspen FACE. In analyses conducted at the end of the Aspen FACE experiment, Talhelm et
al. (2014) and Zak et al. (2011) found few statistically significant interactions between ozone and CO2 in
measures of growth, productivity, ecosystem C pools, or ecosystem N pools. Instead, the combined
treatment (CO2 + ozone) created changes that were similar to the additive effects of the CO2 and ozone
treatments individually (Talhelm et al.. 2014; Zak etal.. 2011). Likewise, Hofmockel et al. (2011)
observed no significant CO2 x ozone interactions among particulate and mineral-associated fractions of
soil organic matter. The relative competitive ability of tree species or aspen genotypes to acquire soil N
was altered by CO2 and ozone as individual treatments, but there were no significant interactions between
the gases (Zak et al.. 2012). Similarly, although CO2 and ozone each individually affected the growth and
survival of the five different aspen genotypes in ways that altered community genetic composition,
community composition in the CO2 + ozone treatment was similar to ambient conditions (Moran and
Kubiske. 2013). There were some significant CO2 x ozone interactions at higher trophic levels, such as
for ectomycorrhizal root tip community composition (Andrew and Lilleskov. 2014). arthropod species
abundance (Hillstrom et al.. 2014). and gypsy moth larvae growth (Couture et al.. 2012). but these effects
tended to be small, vary by year, or to ameliorate the effect of ozone.
Overall, this body of research suggests that increases in atmospheric CO2 to levels predicted by
midcentury can help ameliorate many of the effects of ozone on terrestrial ecosystems. However, because
responses to CO2 and ozone are each species specific, the combined exposure to CO2 and ozone can cause
some shifts in community composition and concomitant shifts in ecosystem function. Moreover, results
from combined CO2 and ozone exposure experiments do not suggest that elevated CO2 will prevent ozone
damage, but instead create ecosystem outcomes that look similar to the current period (early 21st century)
rather than the potential increases in NPP, greater C sequestration, and other changes that could be
observed under low ozone conditions in a higher-C02 environment.
8-177
-------
8.12.3
Weather and Climate
Variation in climate and weather can potentially alter both conditions that lead to the formation,
transport, and persistence of ozone in the troposphere (Jacob and Winner. 2009) as well as the
vulnerability of plants and ecosystems (Anav et al.. 2018; Anav et al.. 2017); this section of text focuses
on how changes in climate and weather have already and may in the future modify the effects of ozone on
ecosystems. The degree to which climate and weather alter the effects of ozone is context specific
because damage to terrestrial ecosystems caused by ozone is largely a function of stomatal uptake (Anav
et al.. 2017). Changes in climate that cause shifts in plant cover from ozone-insensitive species such as
needle-leaf evergreens and C4 grasses to ozone-sensitive species, such as some broadleaf deciduous trees
and C3 grasses, may increase the portion of the landscape at risk for ozone damage. Conversely, changes
in climate that restrict stomatal conductance during periods of the day and growing season that experience
high ozone concentrations would limit damage to vegetation. As an example, central and southern
California has some of the highest ozone concentrations in the U.S. (Mahmud et al.. 2008). Seasonal
peaks in ozone in California occur during the summer (Mahmud et al.. 2008; Gevh et al.. 2000) and daily
peaks occur during late afternoon (Fares et al.. 2013). However, ozone damage to natural vegetation in
California is constrained by the predominance of conifers and other species with low stomatal
conductance, as well as the presence of a Mediterranean climate that concentrates precipitation and the
growth of vegetation to winter and spring months (Fares et al.. 2013). Climate change is expected to
increase the number of high ozone summer days in California (Mahmud et al.. 2008). but also accelerate
the timing of seasonal snowmelt, shift the growing season to earlier in the year, and increase summer
plant moisture deficits (Westerling et al.. 2011). Thus, climate change in California could simultaneously
increase exposure while limiting plant vulnerability by creating conditions that would decrease stomatal
conductance during high ozone periods. Conversely, when applying the DLEM model to 20th century C
cycling in the southeastern U.S., Tian et al. (2012) found a significant interaction between ozone and
climate that decreased NPP. These examples highlight both the range of potential ozone-climate
interactions, as well as the degree to which these interactions are context specific.
There have been relatively few field experiments that manipulated both ozone and temperature. A
warming (~+0.8°C) and ozone (1.2* ambient) experiment was conducted in Finland, first using potted
birch (Betulapendula) trees and then with potted Scots pine (Pinus sylvestris). For birch, warming
increased tree growth and ozone tended to decrease tree growth (particularly a decrease in foliar biomass),
but the only significant interactive effect between warming and ozone was that warming ameliorated the
acceleration of leaf senescence caused by ozone (Kasurinen et al.. 2012). As part of the birch portion of
the experiment, Kasurinen et al. (2017) observed that the two treatments each altered the leaf litter
chemistry and the soil abundance of bacteria and fungi, but there were no meaningful interactions
between the treatments, and treatment effects on litter decomposition were weak. Growth responses for
the pine were similar: warming increased growth rates, whereas ozone caused negative effects that were
weak aside from decreases in older needle biomass (Rasheed et al.. 2017). Only in the 1st year of the
experiment was there a significant warming x ozone effect on growth, wherein the negative effect of
8-178
-------
ozone on needle biomass was larger in the warming treatment. The pine portion of the experiment also
included fertilization and herbivory treatments, and belowground processes such as allocation to root
biomass, mycorrhizal colonization, and the rate of root ramification were subject to complex three- and
four-way interactions between experimental factors (Rasheed ct al.. 2017). There were no interactive
effects on pine sawfly (Acantholyda posticalis) foliar herbivory, but the combination of ozone and
warming increased herbivory-induced emissions of sesquiterpenes and oxidated monoterpenes (Ghimire
et al.. 2017).
There have been more experiments involving ozone and drought stress. Because drought stress
decreases stomatal conductance, it can limit ozone effects on plant growth and leaf gas exchange (Gao et
al.. 2017; Xu et al.. 2017; Bohler et al.. 2013; Hoshika et al.. 201IV However, outcome of these
experiments are not always straightforward; in a drought and ozone experiment on Shantung maple (Acer
truncatum) seedlings, drought stress reduced or alleviated ozone effects on leaf chlorophyll, tree height
growth, and stem diameter growth, but tended to exacerbate ozone effects on photosynthesis and stomatal
conductance (Li et al.. 2015). In the San Joaquin Valley of California, neither drought nor ozone affected
daytime stomatal conductance in the invasive weed Amaranthuspalmeri (Paudcl et al.. 2016).
Ozone may also exacerbate the effects of climate change on vegetation. Although ozone often
decreases stomatal conductance, there is also evidence from multiple experiments that ozone may lead to
decreased stomatal responsiveness to changing environmental conditions such as water stress (Wagg et
al.. 2013; Uddling et al.. 2009). At Aspen FACE, this loss of stomatal control was apparently linked to
increases in canopy conductance, particularly later in the growing season (Sun et al.. 2012; Uddling et al..
2009). The accuracy of model predictions of streamflow in six Appalachian watersheds improved when
both ozone and climate were included with in the model, with higher ozone linked to increases in water
use, decreases in soil moisture, and lower streamflow (Sun et al.. 2012). In addition to hydrologic effects,
the decrease in stomatal responsiveness to drought stress caused by ozone may increase stomatal fluxes of
ozone and ozone damage under low moisture conditions (Haves et al.. 2012a).
Overall, the body of research examining ozone interactions with climate has grown considerably
since the 2013 Ozone ISA. However, the context-specific nature of the outcomes and key mechanisms
makes it difficult to make broad generalizations about how climate and ozone interact to influence
ecosystems.
8.12.4 Summary
Other factors may exacerbate or negate the effects of ozone on plants, these include nitrogen
deposition, CO2, and climate variables. Nitrogen deposition often co-occurs with increased ozone
exposure. At the individual plant level, nitrogen deposition may either increase ozone flux and damage
through increased stomatal conductance and higher amounts of photosynthetic machinery or decrease
ozone damage through increased antioxidant production. Effects of increased nitrogen on plant growth
8-179
-------
may overshadow the detrimental effects of ozone on the same. At community and ecosystem scales, the
species-specific responses to both increased nitrogen and tropospheric ozone result in significant impacts
on species composition, although there is very little evidence of interactive effects between the two. For
CO2, research found the response to elevated CO2 was also species specific. The diverse and complex
effects of elevated CO2 on plant physiology make it difficult to fully predict how ozone might interact
with elevated CO2 in plants. At the individual plant level, increased CO2 exposure may either protect
plants from ozone exposure or overshadow the negative effects. In general, at larger scales, research finds
combined exposure to CO2 and ozone can cause some shifts in community composition and concomitant
shifts in ecosystem function. With respect to climate, modeling studies found a significant interaction
between ozone and climate that decreased NPP. Relatively few field experiments have manipulated both
ozone and temperature, but they find stronger effects of warming on plant growth and function than the
effects of ozone, with little evidence of interactive effects. Drought was once thought to have a protective
effect from ozone exposure, limiting ozone flux into the leaf as stomata close to prevent water loss.
However, more research into ozone-mediated impairment of stomatal function suggest that for some
species this assumption is false, and ozone exposures may be much higher than once thought.
8.13 Exposure Indices/Exposure Response
Exposure indices and exposure-response information for vegetation and related ecosystem effects
of ozone are critical for understanding the effects of current and future ozone exposures and evaluating
potential air quality standards. For over 60 years, controlled ozone exposure experiments have yielded a
wealth of information on exposure indices appropriate for vegetation and exposure response relationships.
This topic has been thoroughly described and supported by hundreds of studies in the 2013 Ozone ISA
(U.S. EPA. 2013) and previous AQCDs (U.S. EPA. 2006. 1996). In this section, new relevant information
was considered with what was previously known pertaining to species in the U.S. There is some brief
discussion of advances in European dose and exposure models for context of potential advances in this
area.
The main conclusions from the 1996 and 2006 Ozone AQCDs and 2013 Ozone ISA regarding
indices based on ambient exposure are still valid. These key conclusions can be restated as follows:
• Ozone effects in plants are cumulative;
• Higher ozone concentrations appear to be more important than lower concentrations in eliciting a
response;
• Plant sensitivity to ozone varies with time of day and plant development stage; and
• Quantifying exposure with indices that accumulate the ozone hourly 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.
8-180
-------
No recent information available since the 2013 Ozone ISA alters these basic conclusions. The
2013 Ozone ISA and previous AQCDs focused on the research used to develop various exposure indices
(e.g., SUM06, AOTx, W126, see Section 8.1.2.2) to help quantify effects on growth and yield in crops,
perennials, and trees (primarily seedlings). The performance of indices was compared through regression
analyses of earlier studies designed to support the estimation of predictive ozone exposure-response
models for growth and/or yield of crops and tree (seedling) species.
8.13.1 Exposure Indices
Exposure indices are metrics that quantify exposure as it relates to measured plant damage
(e.g., reduced growth). In the over 60 years of research, many forms of exposure metrics have been used,
including 7-, 12-, and 24- hour avg. The current secondary standard form of the 4th highest 8-hour max
avg over 3 years is rarely reported in the vegetation research. The most useful metrics in vegetation
research have been differentially weighted hourly concentrations that are cumulative during the growth of
plants. The 2013 Ozone ISA primarily discussed SUM06, AOTx, and W126 exposure metrics (see
Section 8.1.2.2 for definitions). These remain the common concentration-based indices discussed in the
literature since the 2013 Ozone ISA. These three types of metrics performed well in a recent study of
observations of maize and soybean yield and W126 was the preferred metric because it was potentially
the most sensitive index (Mcgrath et al.. 2015). Other studies also report various types of mean
concentration exposures, which are generally less robust than the metrics discussed above. The indices
described in Section 8.1.2.2 have a variety of relevant time windows that may be applied based on time of
day and season. In general, ozone concentrations have applied time windows during the daytime
(e.g., 8:00 a.m.-8:00 p.m.) when stomata are open and during the active growing season (e.g., 3 months
during the warm season; see Section 9.5.3 of 2013 Ozone ISA). In recent study, Mills et al. (2018)
described the distributions and trends of W126, AOT40 and 12-hour avg metrics at vegetated sites across
the globe and found the highest values were in the mid latitudes of the northern hemisphere where the
density of ozone monitors are the greatest.
Another approach for improving risk assessment of vegetation response to ambient ozone is
based on determining the ozone concentration from the atmosphere that enters the leaf (i.e., flux or
deposition). Much work has been published in recent years, particularly in Europe, in using
mathematically tractable flux models for ozone assessments at the regional, national, and European scale
(Feng et al.. 2017: Mills etal.. 2011: Matvssek et al.. 2008: Paoletti and Manning. 2007: Emberson et al..
2000b: Emberson et al.. 2000a). While some efforts have been made in the U.S. to calculate ozone flux
into leaves and canopies (Turnipsccd et al.. 2009: Uddling et al.. 2009: Bergweiler et al.. 2008: Hogg et
al.. 2007: Grulke et al.. 2004: Grantz et al.. 1997: Grantz et al.. 1995). little information has been
published relating these fluxes to effects on vegetation. Recently, Grantz et al. (2013) reported short-term
ozone flux and related it to leaf injury in cotton in California. The authors reported that cotton leaves were
most sensitive in the midafternoon, possibly due to changes in detoxification. They suggested with more
8-181
-------
research a sensitivity parameter may function well with the W126 metric. However, there remains much
unknown about ozone stomatal uptake in vegetation at larger scales and how much uptake results in an
injury or damage, which depends to some degree on the amount of internal detoxification occurring with
each particular species. Those species having high amounts of detoxification potential may, in fact, show
little relationship between ozone stomatal uptake and plant response (Musselman and Massman. 1999).
The lack of data in the U.S. and the lack of understanding of detoxification processes have made this
technique less viable for vulnerability and risk assessments in the U.S.
8.13.2 Exposure Response
The characterization of the effects of ozone on plants is contingent not only on the choice of the
index used (i.e., W126, AOT40) to summarize ozone exposure (see above), but also on quantifying the
response of the plant variables of interest at specific values of the selected index. The many factors that
determine the response include species, genotype and other genetic characteristics, biochemical and
physiological status, previous and current exposure to other stressors, and characteristics of the exposure
itself. This section reviews results that have related specific quantitative observations of ozone exposure
with quantitative observations of plant responses, and the predictions of responses that have been derived
from those observations through empirical models.
Extensive exposure-response information on a wide variety of plant species has been produced by
two long-term projects that were designed with the explicit aim of obtaining quantitative characterizations
of the response of such an assortment of crop plants and tree seedlings to ozone under North American
conditions: the NCLAN project for crops, and the U.S. EPA National Health and Environmental Effects
Research Laboratory, Western Ecology Division (NHEERL-WED) tree seedling project. Both projects
used OTCs to expose plants to three to five levels of ozone. These two programs are explained in detail in
Section 9.5 of the 2013 Ozone ISA.
The 1996 and 2006 Ozone AQCDs relied extensively on analyses of NCLAN and
NHEERL-WED projects by Lee et al. (1994); Lee et al. (1989. 1988. 1987) Hogsett et al. (1997). Lee and
Hogsett (1999). Heck et al. (1984). Rawlings and Cure (1985). Lesser et al. (1990). and Gumpertz and
Rawlings (1992). Those analyses concluded that a three-parameter Weibull model
fW 126\C
Yield or Biomass = Ae v b )
Equation 8-2
is the most appropriate model for the response of absolute yield and growth to ozone exposure, because of
the interpretability of its parameters, its flexibility (given the small number of parameters), and its
tractability for estimation. In addition, removing the intercept^ results in a model of relative
yield/biomass without any further reparameterization. Formulating the model in terms of relative yield or
biomass loss (yield or biomass loss = [1 - relative yield or biomass]) is essential in comparing
8-182
-------
exposure-response across species, genotypes, or experiments for which absolute values of the response
may vary greatly. In the 1996 and 2006 Ozone AQCDs, the two-parameter model of relative yield or
biomass loss was used in deriving common models for multiple species, multiple genotypes within
species, and multiple locations. The two-parameter Weibull model for relative yield or biomass is:
/W126\C
Relative Yield or Biomass Loss = 1 — e v b )
Equation 8-3
The NCLAN project was initiated by the U.S. EPA in 1980 primarily to improve estimates of
yield loss under field conditions and to estimate the magnitude of crop losses caused by ozone throughout
the U.S. (Heck et al.. 1991; Heck et al.. 1982V The cultural conditions used in the NCLAN studies
approximated typical agronomic practices, and the primary objectives were (1) to define relationships
between yields of major agricultural crops and ozone exposure as required to provide data necessary for
economic assessments and development of ozone NAAQS, (2) to assess the national economic
consequences resulting from ozone exposure of major agricultural crops, and (3) to advance
understanding of cause-and-effect relationships that determine crop responses to pollutant exposures.
NCLAN experiments yielded 54 exposure-response curves for 12 crop species, some of which were
represented by multiple cultivars at several of six locations throughout the U.S. Eight of the 54 crop data
sets were from plants grown under a combination of ozone exposure and experimental drought
conditions. Table 8-22 reports the minimum, median, and maximum nondroughted exposure response
parameters for 10 crop species available from Lee and Hogsett (1996).1 Outside of NCLAN, some
exposure response information is available for U.S. fruits and vegetables such as onions, broccoli,
Valencia oranges, and tomatoes (Olszvk et al.. 1990; Temple. 1990; Temple etal.. 1990).
The NHEERL-WED project was initiated by U.S. EPA in 1988 with the same objectives for tree
seedlings, and yielded 51 exposure-response functions for multiple genotypes of 11 tree species grown for
up to 3 years in Oregon, Michigan, Alabama, and the Great Smoky Mountains National Park (Lee and
Hogsett. 1996). The Weibull parameters and related information for all 51 seedling studies are given in
Table 8-23. Figure 8-14 shows graphical illustrations of some of the exposure-response information from
the NHEERL-WED tree seedling database. Table 8-24 shows the median composite functions of 11 tree
seedlings species plus aspen clones. These functions were calculated from the 51 seedling studies in
Table 8-23. adjusted to 92-day exposures to compare across studies and were reported originally in
Table 12 of Lee and Hogsett (1996). Figure 8-15 summarizes the information contained in Table 8-23 and
Table 8-24. The top graph shows a broad range of seedling relative biomass loss to 92-day, 12-hour
W126 exposures, while the bottom graph shows the same responses plotted in a narrower range ofW126
to focus on the median curves. The gray lines plot the exposure-response function for each of the 51 tree
seedling studies. The solid red line is the 50th percentile (i.e., median) composite exposure-response
1 Some individual studies within crop species were excluded from the min, max, and median determinations because
these studies had considerably shorter exposure durations than other replicate studies.
8-183
-------
function of 51 studies of tree seedling responses (Table 8-23). that includes 11 species and aspen clone
studies (Lee and Hogsett. 1996). At any given exposure, half of the studies (i.e.; 25 studies) were more
sensitive and half were less sensitive than the solid red line. The dotted green line is the 50th percentile
(i.e., median) of only the 11 species-specific median tree seedling functions in Table 8-24 with the
exception of the aspen clones. In this case, at any given exposure, half of the individual tree species
(i.e., 5 species) were predicted to more sensitive and half were less sensitive than dotted green line. The
median of the 11 species-specific functions (dotted green line) shows less sensitivity to ozone than the
median of the 51 studies (solid red line), as a result of differences in the number of studies associated with
each species. Tree species that are consistently more sensitive than the median functions include black
cherry and quaking aspen, found in many regions of the continental U.S.
8-184
-------
Table 8-22 Ozone exposure-response functions for selected National Crop Loss
Assessment Network (NCLAN) crops.
Ozone Index
Type
Crop
Relative Yield Loss Function
W126
Max
Cotton
1-exp[-(index/74.6)1 068]
W126
Min
Cotton
1 — exp[—(index/113.3)1 397]
W126
Median
Cotton
1 -exp[-(index/96.1 )1 482]
W126
Max
Field corn
1-exp[-(index/92.7)2585]
W126
Min
Field corn
1-exp[-(index/94.2)4167]
W126
Median
Field corn
1 -exp[-(index/97.9)2 96S]
W126
Median
Grain sorghum3
1-exp[-(index/205.9)1 963]
W126
Median
Peanut3
1-exp[-(index/96.8)1 890]
W126
Max
Soybean
1 — exp[—(index/130.1 )1]
W126
Min
Soybean
1-exp[-(index/476.7)1113]
W126
Median
Soybean
1 -exp[-(index/110.2)1 359]
W126
Max
Winter wheat
1-exp[-(index/24.7)1 °]
W126
Min
Winter wheat
1-exp[-(index/76.8)2031]
W126
Median
Winter wheat
1 -exp[-(index/53.4)2 367]
W126
Median
Lettuce3
1-exp[-(index/54.6)4917]
W126
Median
Barley
1-exp[-(index/6998.5)1 388]
W126
Median
Kidney bean3
1 -exp[-(index/43.1 )2 219]
W126
Min
Potato
1 —exp[—(index/113.8)1 299]
W126
Max
Potato
1-exp[-(index/96.3)1]
W126
Median
Potato
1-exp[-(index/99.5)1 242]
aPeanuts, grain sorghum, lettuce, barley, and kidney bean only have one exposure-response function and, therefore, do not have
a max and min.
Source: Lee and Hoasett (19961. Table 10.
8-185
-------
Table 8-23 Weibull exposure-response curves relating relative biomass loss as
a function of 12-hour W126 in ppm-hour for 51 seedling studies as
reported in Tables 12 and 13 of Lee and Hoqsett (1996). Relative
Biomass Loss = 1 - exp[-(W126/B)c]
Duration
Days3
Weibull Parameters
Species
Site
Year
Harvest No.b
B
C
Aspen-wild
Michigan
98
1991
1
124.7
1
Aspen-wild
Oregon
84
1989
1
74.6
1.328
Aspen-wild
Oregon
84
1989
2
127.9
1
Aspen-wild
Oregon
118
1991
1
103.1
3.162
Aspen-wild
Oregon
118
1991
2
102.4
5.174
Aspen-wild
Oregon
112
1990
1
85.7
6.744
Aspen-wild
Oregon
112
1990
2
109.2
1.29
Aspen-216
Michigan
82
1990
1
57.3
1.621
Aspen-253
Michigan
82
1990
1
66.9
1
Aspen-259
Michigan
82
1990
1
40.2
1
Aspen-271
Michigan
82
1990
1
33.6
2.35
Aspen-216
Michigan
98
1991
1
86.1
1
Aspen-259
Michigan
98
1991
1
47
1
Aspen-271
Michigan
98
1991
1
44.9
4.856
Black cherry
SMNP(c)
76
1989
1
42.2
1.04
Black cherry
SMNP
140
1992
1
44.5
1.017
Douglas fir
Oregon
113
1989-1990
1
357.7
1.844
Douglas fir
Oregon
113
1989-1990
2
1.30 x 1017
1
Douglas fir
Oregon
234
1989-1990
3
345.5
5.291
Douglas fir
Oregon
234
1989-1990
4
2,250.8
1
Douglas fir
Oregon
118
1991-1992
1
88.5
47.58
Douglas fir
Oregon
118
1991-1992
2
100.8
12.215
Douglas fir
Oregon
230
1991-1992
3
171.6
12.808
8-186
-------
Table 8-23 (Continued): Weibull exposure response curves relating total biomass
(g) as a function of 12-hour W126 in ppm-hour for
51 seedling cases as reported in Tables 12 and 13 of Lee
and Hogsett (1996). Relative Biomass
Loss = 1 -exp[-(W126/B)c].
Weibull Parameters
Species
Site
Duration
Days3
Year
Harvest No.b
B
C
E. white pine
Michigan
83
1990-1991
1
33.2
4.306
E. white pine
Michigan
180
1990-1991
3
524.2
1
Ponderosa pine
Oregon
111
1989
1
192.2
1
Ponderosa pine
Oregon
111
1989
2
283.8
1
Ponderosa pine
Oregon
113
1989-1990
1
181.8
1
Ponderosa pine
Oregon
113
1989-1990
2
279.2
1
Ponderosa pine
Oregon
234
1989-1990
3
255.9
1
Ponderosa pine
Oregon
234
1989-1990
4
494.5
1
Ponderosa pine
Oregon
118
1991-1992
1
213.7
1
Ponderosa pine
Oregon
118
1991-1992
2
165.9
1
Ponderosa pine
Oregon
230
1991-1992
3
349.5
1.297
Ponderosa pine
Oregon
84
1991
1
359.2
1
Ponderosa pine
Oregon
140
1992
1
173.3
2.774
Red alder
Oregon
113
1989
1
196.5
1
Red alder
Oregon
113
1989
2
142.2
5.068
Red alder
Oregon
118
1991
1
399.3
1
Red alder
Oregon
118
1991
2
1.20 x 1015
1
Red alder
Oregon
121
1990
1
172
1.427
Red alder
Oregon
112
1992
1
200.9
1.041
Red maple
SMNP
55
1988
1
190.2
1.376
Sugar maple
Michigan
83
1990-1991
1
34.7
7.189
Sugar maple
Michigan
180
1990-1991
3
65.2
5.806
Tulip poplar
SMNP
75
1990-1991
1
26.5
5.794
Tulip poplar
SMNP
184
1990-1991
3
352.6
1
8-187
-------
Table 8-23 (Continued): Weibull exposure response curves relating total biomass
(g) as a function of 12-hour W126 in ppm-hour for
51 seedling cases as reported in Tables 12 and 13 of Lee
and Hogsett (1996). Relative Biomass
Loss = 1 -exp[-(W126/B)c].
Duration
Days3
Weibull Parameters
Species
Site
Year
Harvest No.b
B
C
Tulip poplar
SMNP
81
1992
1
45.1
2.018
Virginia pine
SMNP
98
1992
1
1,807.9
1
Loblolly pine
Alabama
555
1988-1989
3
3,966.3
1
Loblolly pine
Alabama
555
1988-1989
3
12,626.4
1
SMNP = Smoky Mountain National Park
aDuration corresponds to the length in days of the 1 st year of exposure for Harvests 1 and 2 and to the total length of the 1 st and
2nd years of exposure for Harvests 3 and 4.
bHarvest 1 occurs immediately following the end of the 1st year of exposure. Harvest 2 occurs in the spring following the 1st year
of exposure. Harvest 3 occurs immediately following the end of the 2nd year of exposure. Harvest 4 occurs in the spring following
the 2nd year of exposure.
8-188
-------
75th Pctile
50th Pctile
25th Pctile
7 Douglas firdatasets
3 Tulip poplardatasets
50lh Pctile
90 day 12 hr W126 (ppm-hr)
90 day 12 hr W126 (ppm-hr)
Note: Curves were standardized to 90-day W126. The number of studies available for each species is indicated on each plot.
Source of Weibull parameters: Lee and Hoasett d 9961.Reprinted from U.S. EPA (20131
Figure 8-14 Quantiles of predicted relative biomass loss for four tree species
in NHEERL-WED experiments. Quantiles of the predicted relative
aboveground biomass loss at seven exposure values of 12-hour
W126 for Weibull curves estimated using nonlinear regression on
data for four tree species grown under well-watered conditions
for 1 or 2 years.
8-189
-------
Table 8-24 Median composite ozone exposure-response functions3 for tree
seedlings adjusted to 92-days exposure.
Ozone Index
Type
Species
Relative Biomass Loss Function
W126
Median
Ponderosa pine
1 -exp[-(index/159.63)1190]
W126
Median
Red alder
1-exp[-(index/179.06)1 2377]
W126
Median
Black cherry
1-exp[-(index/38.92)0"21]
W126
Median
Tulip poplar
1-exp[-(index/51.38)20889]
W126
Median
Sugar maple
1-exp[-(index/36.35)57785]
W126
Median
E. white pine
1-exp[-(index/63.23)1 6582]
W126
Median
Red maple
1-exp[-(index/3 1 8.12)1 3756]
W126
Median
Douglas fir
1 -exp[-(index/106.83)59631]
W126
Median
Aspen-wild
1-exp[-(index/109.81 )1 2198]
W126
Median
Aspen-clonesb
1-exp[-(index/49.56)1 4967]
W126
Median
Loblolly pinec
1 -exp[-(index/1,021.63)°9954]
W126
Median
Virginia pine
1 — exp[—(index/1,714.64)1 ]
individual exposure-response curves are reported using the 12-h W126 index adjusted to a 92-day exposure duration.
bAspen clones 216, 253, 259, and 271 with varying levels of ozone tolerance.
°Loblolly pine median parameters were derived from the two loblolly pine studies listed in Table 8-23 in the same manner as
described in Lee and Hoasett d9961 for the other tree species.
Source: Lee and Hoasett (19961. Tables 12 and 13.
8-190
-------
son
Lee and Hogsett (1996) median of 51 studies
Median of 11 species medians
45-
40-
35-
IS 30-
ro
£
o
£ 25-
0)
>
20-
15-
10-
26
0
2
6
10
12
20
22
24
30
4
8
14
16
18
28
92-day 12-hour W126 (ppm-hr)
15-.
Lee and Hogsett (1996) median of 51 studies
Median of 11 species medians
Of
q
_i
10-
in
ra
£
o
m
0)
>
22
10
12
20
6
8
14
16
18
92-day 12-hour W126 (ppm-hr)
The solid gray lines are Weibull functions of the 51 studies reported in Table 8-23 of Lee and Hogsett (1996) standardized to a
92-day exposure. The red line is the composite function for the 50th percentile (i.e., median) of the 51 studies (Lee and Hogsett,
1996). The dash-dotted green line is the 50th percentile (i.e., median) of the 11 individual species median functions in Table 8-24.
The top graph shows the range of responses from a W126 of 0 to 30 ppm-hour. The bottom graph is the same responses plotted in
a narrower range of W126 of 6 to 22 ppm-hour.
Figure 8-15 Relative biomass loss predictions from Weibull
exposure-response functions that relate percent aboveground
biomass loss to 12-hour W126 exposures adjusted to 92 days.
8-191
-------
In the 2013 Ozone ISA, yield and growth results for aspen trees and soybean that had provided
extensive exposure-response information in those projects have become available from studies that used
FACE technology, which is intended to provide conditions much closer to natural environments
(Prcgitzcr et al.. 2008; Morgan et al.. 2006; Morgan et al.. 2004; Dickson et al.. 2000). The NCLAN and
NHEERL-WED data with exposure measured as W126, was used to derive single-species median models
for soybean and aspen from studies involving different genotypes, years, and locations. The resulting
models were used to predict the change in yield of soybean and biomass of aspen between the two levels
of exposure reported in the later FACE experiments. Results from these new experiments were
exceptionally close to predictions from the models (Figure 8-16. Figure 8-17). The accuracy of model
predictions for two widely different plant species provided support for the validity of the corresponding
multiple-species models for crops and trees in the NCLAN and NHEERL-WED projects. However,
variability among species in those projects indicates that the range of sensitivity is likely quite wide. This
was confirmed by a study with cottonwood in a naturally occurring gradient of exposure (Gregg et al..
2006). which established the occurrence of species with responses substantially more severe than
predicted by the median model for multiple species.
100
(/)
(/)
O
_l
T3
Q)
>-
NCLAN 75th Pctile
FACE75th Pctile
NCLAN median FACE median
FACE25th Pctile
NCLAN 25th Pctile
H—'
c
0
o
-------
3000 n
2500 -
2000 -
1500 -
T 20(
¦««. -L
1000 -I 9 | 2000
2002
** $ 2001
1999
500 -
$ 1998
10 20 30 40 50
90 day 12 hr W126 (yearly average)
60
70
Note: Black dots are aspen biomass/m2 for three FACE rings filled with an assemblage of five clonal genotypes of aspen at Aspen
FACE; bars are SE for three rings; dashed line is median composite model for four clonal genotypes and wild-type seedlings in
11 NHEERL-WED 1-year OTC studies. Aspen FACE ozone data updated from Kubiske and Foss (20151. Single year 12-hour W126
is shown rather than the cumulative yearly average (average of each current and previous year) shown in Figure 9-20 of the 2013
Ozone ISA.
Source: Modified with permission from the publisher, adapted from King et al. (20051 and Lee and Hoasett (19961.
Figure 8-17 Comparison between aboveground biomass observed in Aspen
FACE experiment in 6 years and biomass predicted by the median
composite function based on NHEERL-WED.
Since the 2013 Ozone ISA, there have been a few new experimental studies that add more
exposure-response relationship information to the large historical database available on U.S. plants. In a
new experimental study, Betzelberger et al. (2012) studied seven soybean cultivars at the SoyFACE
experiment in Illinois. They found that the cultivars showed similar responses in a range of ozone
exposures expressed as AOT40. These results support conclusions of previous studies (Betzelberger et al..
2010) and the 2013 Ozone ISA that sensitivity of current soybean genotypes is not different than early
genotypes; therefore, soybean response functions developed in the NCLAN program remain valid. A
study by Neufeld et al. (2018) provided information on foliar injury response on two varieties of cutleaf
coneflower (Rudbeckict laciniata). For example, one variety had statistically detectable foliar injury when
the 24-hour W126 index reached 23 ppm-hour (12-hour AOT40 = 12 ppm-hour). Gao et al. (2017)
studied an ozone-sensitive hybrid cottonwood and found a strong relationship between biomass loss and
8-193
-------
ozone exposure measured as AOT40 and phytotoxic ozone dose. This study was performed in China, but
this species does occur in the U.S.
Despite the limited number of recent U.S. exposure-response studies, U.S. and international
syntheses have highlighted response function information for grassland and other plant species that occur
in the U.S. In a study by van Goethem et al. (2013). AOT40 response relationships were calculated for
87 grassland species that occur in Europe. Seventeen of these species are native to the U.S. and
65 additional species have been introduced to the U.S. and may have significant ecological, horticultural,
or agricultural value (USDA. 2015). This study has the most significant amount of new
exposure-response information for plants in the U.S. (see Table 8-25). A soybean synthesis study used
some U.S. studies along with studies from other countries to create composite exposure-response
functions based on a 7-hour mean metric. This study had limitations because the 7-hour means for many
studies had to be converted from other published metrics and some soybean cultivars included in the
study may not be used in the U.S. However, the same general patterns were seen with sensitivity of
soybean yield to ozone as reported in the 2013 ISA (Osborne et al.. 2016). Tai and Martin (2017)
developed an empirical model (partial derivative linear regression [PDLR] model) from multidecadal data
sets to estimate geographical variations across the U.S. in sensitivity to ozone of wheat, maize, and
soybean. This approach takes into consideration strong ozone-temperature covariation and does not rely
on pooled concentration-response functions. Several European studies have added to the
exposure-response information to the literature, but these studies mainly focused on European plant
species (Abeli et al.. 2017; Payne et al.. 2017; Sanz et al.. 2016; Haves etal.. 2011).
8-194
-------
Table 8-25 Grassland species that occur in the U.S. with biomass loss exposure-response functions as a
function of AOT40 calculated from previously published open-top chamber (OTC) experiments by
van Goethem et al. (2013).ab
Species
Duration
a
b
Exposure
(ppm-h) for
10% Biomass
Reduction
R2
Status in
U.S.
Reference
Trifolium striatum
Annual
-0.046
0.9
1.94
0.95
Introduced
Gimeno et al. (2004)
Medicago minima
Annual
-0.049
0.97
1.98
0.98
Introduced
Gimeno et al. (2004)
Trifolium angustifoiium
Annual
-0.046
0.96
2.06
0.85
Introduced
Gimeno et al. (2004)
Matricaria chamomiiia
Annual
-0.051
1.06
2.08
0.16
Introduced
Berqmann et al. (1995),
Beramann et al. (1996b)
Rumex acetosa
Perennial
-0.048
1.02
2.14
0.22
Introduced
Haves et al. (2006), Pleiiel
and Danielsson (1997),
Power and Ashmore (2002),
Ashmore et al. (1996)
Maiva syivestris
Perennial
-0.047
1.06
2.28
0.63
Introduced
Beramann et al. (1995).
Beramann et al. (1996b)
Papaver dubium
Annual
-0.041
0.98
2.4
0.93
Introduced
Beramann et al. (1996b)
Vaccinium vitis-idaea
Perennial
-0.042
1.03
2.46
0.98
Native
Mortensen and Nilsen
(1992)
Phieum aipinum
Perennial
-0.04
1.14
2.84
0.8
Native
Danielsson et al. (1999),
Pleiiel and Danielsson
(1997)
Leontodon hispidus
Perennial
-0.031
0.97
3.18
0.49
Introduced
Pleiiel and Danielsson
(1997), Ashmore et al.
(1996)
8-195
-------
Table 8-25 (Continued): Grassland species that occur in the U.S. with biomass loss exposure-response functions
as a function of AOT40 calculated from previously published open-top chamber (OTC)
experiments by van Goethem et al. (2013).ab
Species
Duration
a
b
Exposure
(ppm-h) for
10% Biomass
Reduction
R2
Status in
U.S.
Reference
Cirsium arvense
Perennial
-0.03
0.99
3.24
0.04
Introduced
Berqmann et al. (1995),
Haves et al. (2006), Power
and Ashmore (2002).
Beramann et al. (1996b)
Phleum pratense
Perennial
-0.027
1.09
3.96
0.46
Introduced
Danielsson et al. (1999),
Kohut et al. (1988).
Mortensen and Nilsen
(1992)
Dianthus deltoides
Perennial
-0.024
0.97
3.98
1
Introduced
Pleiiel and Danielsson
(1997)
Trifolium subterraneum
Perennial
-0.026
1.08
4.14
0.46
Introduced
Gimeno et al. (2004)
Lolium rigidum
Annual
-0.021
0.89
4.26
0.5
Introduced
Gimeno et al. (2004)
Trifolium glomeratum
Annual
-0.019
0.95
4.92
0.66
Introduced
Gimeno et al. (2004)
Campanula rotundifolia
Perennial
-0.021
1.02
4.98
0.07
Native
Haves et al. (2006).
Ashmore et al. (1996)
Matricaria matricarioides
Annual
-0.021
1.1
5.3
0.25
Introduced
Beramann et al. (1995),
Beramann et al. (1996b)
Avena sterilis
Annual
-0.019
1.01
5.4
0.09
Introduced
Gimeno et al. (2004)
Senecio vulgaris
Annual
-0.017
1.14
6.72
0.39
Introduced
Beramann et al. (1995),
Pleiiel and Danielsson
(1997)
Aegilops geniculata
Annual
-0.013
0.96
7.6
0.51
Introduced
Gimeno et al. (2004)
8-196
-------
Table 8-25 (Continued): Grassland species that occur in the U.S. with biomass loss exposure-response functions
as a function of AOT40 calculated from previously published open-top chamber (OTC)
experiments by van Goethem et al. (2013).ab
Species
Duration
a
b
Exposure
(ppm-h) for
10% Biomass
Reduction
R2
Status in
U.S.
Reference
Nardus stricta
Perennial
-0.012
0.99
8.3
0.73
Introduced
Haves et al. (2006).
Ashmore et al. (1996)
Trifolium repens
Perennial
-0.011
0.94
8.76
0.89
Introduced
Bunaener et al. (1999b).
Ashmore et al. (1996)
Hieracium pilosella
Perennial
-0.011
0.97
8.78
0.06
Introduced
Pleiiel and Danielsson
(1997). Ashmore et al.
(1996)
Silene acaulis
Perennial
-0.009
0.94
10.2
0.52
Native
Mortensen and Nilsen
(1992)
Bromus sterilis
Annual
-0.009
0.98
10.92
0.19
Introduced
Gimeno et al. (2004)
Chrysanthemum leucanthemum
Perennial
-0.009
1.01
11.26
0
Introduced
Bunaener et al. (1999b)
Lychnis flos-cuculi
Perennial
-0.013
1.5
11.38
0.52
Introduced
Bunaener et al. (1999b).
Power and Ashmore (2002).
Tonneiick et al. (2004).
Franzarina et al. (2000)
Holcus lanatus
Perennial
-0.009
1
11.52
0.6
Introduced
Haves et al. (2006).
Tonneiick et al. (2004).
Ashmore et al. (1996)
Chrysanthemum segetum
Annual
-0.008
0.96
11.96
0
Introduced
Pleiiel and Danielsson
(1997)
Festuca rubra
Perennial
-0.008
1
12.5
0.22
Native
Bunaener et al. (1999b).
Haves et al. (2006).
Ashmore et al. (1996)
8-197
-------
Table 8-25 (Continued): Grassland species that occur in the U.S. with biomass loss exposure-response functions
as a function of AOT40 calculated from previously published open-top chamber (OTC)
experiments by van Goethem et al. (2013).ab
Species
Duration
a
b
Exposure
(ppm-h) for
10% Biomass
Reduction
R2
Status in
U.S.
Reference
Chenopodium album
Annual
-0.007
0.94
13.28
0
Native
Berqmann et al. (1995),
Pleiiel and Danielsson
(1997). Beramann et al.
(1996b)
Briza maxima
Annual
-0.006
0.87
13.82
0.08
Introduced
Gimeno et al. (2004)
Tragopogon orientalis
Perennial
-0.007
1.01
14.26
0.85
Introduced
Bunaener et al. (1999b)
Hypochaeris radicata
Perennial
-0.006
0.95
15.02
0.54
Introduced
Pleiiel and Danielsson
(1997), Ashmore et al.
(1996)
Centaurea jacea
Perennial
-0.006
1.01
15.76
0.74
Introduced
Bunaener et al. (1999b)
Trifolium pratense
Perennial
-0.007
1.09
15.82
0.8
Introduced
Bunaener et al. (1999b),
Kohut et al. (1988).
Ashmore et al. (1996)
Bromus arvensis
Annual
-0.006
1.03
16.3
0.01
Introduced
Pleiiel and Danielsson
(1997)
Taraxacum officinale
Perennial
-0.006
1.08
17.96
0.37
Native
Bunaener et al. (1999b)
Poa annua
Annual
-0.005
0.98
18.14
0.88
Introduced
Pleiiel and Danielsson
(1997)
Poa pratensis
Perennial
-0.005
0.96
18.48
0.06
Native*
Bunaener et al. (1999b),
Ashmore et al. (1996)
Papaver rhoeas
Annual
-0.005
0.91
19.06
0.78
Introduced
Pleiiel and Danielsson
(1997)
Eupatorium cannabinum
Perennial
-0.005
1.07
19.78
0.72
Introduced
Franzarina et al. (2000)
8-198
-------
Table 8-25 (Continued): Grassland species that occur in the U.S. with biomass loss exposure-response functions
as a function of AOT40 calculated from previously published open-top chamber (OTC)
experiments by van Goethem et al. (2013).ab
Species
Duration
a
b
Exposure
(ppm-h) for
10% Biomass
Reduction
R2
Status in
U.S.
Reference
Briza media
Perennial
-0.005
0.99
21.5
0.75
Introduced
Pleiiel and Danielsson
(1997), Ashmore et al.
(1996)
Anthoxanthum odoratum
Perennial
-0.004
0.94
21.9
0.38
Introduced
Haves et al. (2006). Pleiiel
and Danielsson (1997),
Haves et al. (2010).
Ashmore et al. (1996)
Bromus hordeaceus
Annual
-0.004
0.98
23.34
0.7
Introduced
Gimeno et al. (2004)
Saxifraga cernua
Perennial
-0.004
1.04
24.26
0.17
Native
Mortensen and Nilsen
(1992)
Polygonum viviparum
Perennial
-0.003
1
29.32
0.99
Native
Mortensen and Nilsen
(1992)
Achillea millefolium
Perennial
-0.003
1.04
31.4
0.99
Native
Bunqener et al. (1999b)
Achillea ptarmica
Perennial
-0.003
1.02
35.08
0.53
Introduced
Franzarina et al. (2000)
Lotus corniculatus
Perennial
-0.003
0.96
36.82
0.04
Introduced
Bunqener et al. (1999b),
Ashmore et al. (1996)
Knautia arvensis
Perennial
-0.003
1.02
40.88
0.58
Introduced
Bunaener et al. (1999b)
Deschampsia flexuosa
Perennial
-0.002
1.08
59.96
0.64
Native
Ashmore et al. (1996)
Crepis biennis
Annual
-0.002
1.08
67.36
0.04
Introduced
Bunaener et al. (1999b)
Salvia pratensis
Perennial
-0.001
1.08
83.38
0.75
Introduced
Bunaener et al. (1999b)
8-199
-------
Table 8-25 (Continued): Grassland species that occur in the U.S. with biomass loss exposure-response functions
as a function of AOT40 calculated from previously published open-top chamber (OTC)
experiments by van Goethem et al. (2013).ab
Species
Duration
a
b
Exposure
(ppm-h) for
10% Biomass
Reduction
R2
Status in
U.S.
Reference
Dactylis glomerata
Perennial
-0.001
0.9
150.42
0.34
Introduced
Bunqener et al. (1999b),
Pleiiel and Danielsson
(1997). Ashmore et al.
(1996)
Plantago lanceolata
Perennial
-0.001
0.96
192.58
0.96
Introduced
Bunqener et al. (1999b),
Haves et al. (2006), Pleiiel
and Danielsson (1997),
Tonneiick et al. (2004),
Franzarinq et al. (2000).
Ashmore et al. (1996)
Phalaris arundinacea
Perennial
0.027
0.89
-
0.76
Native*
Pleiiel and Danielsson
(1997)
Festuca pratensis
Perennial
0.018
0.92
-
0.13
Introduced
Pleiiel and Danielsson
(1997)
Anthyllis vulneraria
Perennial
0.017
0.98
-
0.72
Introduced
Pleiiel and Danielsson
(1997), Ashmore et al.
(1996)
Silene dioica
Perennial
0.015
0.91
-
0.92
Introduced
Bunqener et al. (1999b)
Silene vulgaris
Perennial
0.014
0.97
-
0.83
Introduced
Pleiiel and Danielsson
(1997)
Galium saxatile
Perennial
0.011
0.91
-
0.05
Introduced
Haves et al. (2006), Haves
et al. (2010), Ashmore et al.
(1996)
Molinia caerulea
Perennial
0.01
0.88
-
0.39
Introduced
Tonneiick et al. (2004).
Franzarinq et al. (2000)
8-200
-------
Table 8-25 (Continued): Grassland species that occur in the U.S. with biomass loss exposure-response functions
as a function of AOT40 calculated from previously published open-top chamber (OTC)
experiments by van Goethem et al. (2013).ab
Species
Duration
a
b
Exposure
(ppm-h) for
10% Biomass
Reduction
R2
Status in
U.S.
Reference
Salix herbacea
Perennial
0.008
1.07
-
0.98
Native
Mortensen and Nilsen
(1992)
Deschampsia caespitosa
Perennial
0.008
1.01
-
0.83
Native
Ashmore et al. (1996)
Carex bigelowii
Perennial
0.007
1
-
0.1
Native
Haves et al. (2010)
Agrostemma githago
Annual
0.006
1.05
-
0.83
Introduced
Pleiiel and Danielsson
(1997)
Saxifraga cespitosa
Perennial
0.006
0.92
-
0.15
Native
Mortensen and Nilsen
(1992)
Calluna vulgaris
Perennial
0.006
0.92
-
0.04
Introduced
Foot et al. (1996)
Festuca ovina
Perennial
0.005
1.04
0.03
Introduced
Haves et al. (2006), Pleiiel
and Danielsson (1997),
Haves et al. (2010).
Ashmore et al. (1996)
Lolium perenne
Perennial
0.003
0.98
-
0.1
Introduced
Bunqener et al. (1999b),
Ashmore et al. (1996)
Agrostis capillaris
Perennial
0.003
1
-
0.84
Introduced
Haves et al. (2006). Haves
et al. (2010). Ashmore et al.
(1996)
Centaurea cyanus
Annual
0.002
1.03
-
0.02
Introduced
Pleiiel and Danielsson
(1997)
Aegilops triuncialis
Annual
0.002
1.04
-
0.81
Introduced
Gimeno et al. (2004)
Carum carvi
Perennial
0.002
0.96
-
0.03
Introduced
Bunqener et al. (1999b)
8-201
-------
Table 8-25 (Continued): Grassland species that occur in the U.S. with biomass loss exposure-response functions
as a function of AOT40 calculated from previously published open-top chamber (OTC)
experiments by van Goethem et al. (2013).ab
Species
Duration
a
b
Exposure
(ppm-h) for
10% Biomass
Reduction
R2
Status in
U.S.
Reference
Onobrychis viciifolia
Perennial
0.002
1.1
-
0.43
Introduced
Bunqener et al. (1999b)
Arrhenatherum elatius
Perennial
0.001
0.91
-
0.51
Introduced
Bunaener et al. (1999b).
Ashmore et al. (1996)
Trisetum flavescens
Perennial
0.001
1.1
-
0.61
Introduced
Bunaener et al. (1999b)
Bromus erectus
Perennial
0.001
1.05
-
0.99
Introduced
Bunaener et al. (1999b).
Ashmore et al. (1996)
Alopecurus pratensis
Perennial
0.001
0.96
-
0.24
Introduced
Pleiiel and Danielsson
(1997), Ashmore et al.
(1996)
aBoth native and introduced/naturalized plant species documented to occur in the U.S. are included.
bData are found in the Supplemental Information in this publication.
Note: "Duration describes the life cycle of the plant (annual or perennial). Columns "a" and "b" represent variables in the exposure-response relationship with ozone, y= ax + b
derived by linear regression for exposure (in AOT40) and proportion of biomass compared to charcoal-filtered air treatment. Column "Exposure..." represents the AOT40 ozone
exposure that reduces species biomass by 10%. Species that exhibited a biomass reduction in response to ozone have a negative value for a, and species appear in the table in
descending order of sensitivity to ozone (i.e., most sensitive species at the top, most tolerant species at the bottom of table). Column "R2" is the coefficient of determination from
linear regression for the exposure-response relationship. The column "Status in U.S." is based on the USDA (20151 determination of whether species are native to the U.S. (Native),
are introduced to the U.S. (Introduced), or have populations with native progenitors as well as populations with introduced progenitors (Native*).
ER functions for this table are from the OZOVEG database CHaves et al.. 20071. Six out of the sixteen studies above have been cited in previous ISAs or AQCDs."
8-202
-------
8.13.3
Summary
Exposure indices are metrics that quantify exposure as it relates to plant response (e.g., reduced
growth). These indices are summary measures of ozone concentrations overtime intended to provide a
consistent metric for reviewing and comparing exposure-response effects obtained from various studies.
Given the current state of knowledge and the best available data, exposure indices that cumulate and
differentially weight the higher hourly average concentrations and also include the midlevel values
(e.g., the W126 or AOT40 metrics) continue to offer the most defensible approach for use in developing
response functions and comparing studies, as well as for defining future indices for vegetation protection.
Since the 2013 Ozone ISA, there have been a limited number of new experimental studies that
add more exposure-response relationship information (see Table 8-26). However, U.S. and international
syntheses have highlighted response function information for grassland and other plant species that occur
in the U.S. (see Table 8-25). thus adding many new species with exposure-response information. Previous
reviews of the NAAQS have included exposure-response functions for the yield of many crop species
(see Table 8-22). and for the biomass accumulation of tree species (see Table 8-23. Table 8-24.
Figure 8-14. and Figure 8-15). They were based on large-scale experiments designed to obtain clear
exposure-response data and are updated by using the W126 metric to quantify exposure. In more recent
years, extensive exposure-response data obtained in more naturalistic settings have become available for
yield of soybean and growth of aspen. In the 2013 Ozone ISA, the exposure-response median functions
were validated based on previous data by comparing their predictions with the newer observations (see
Section 9.6 of the 2013 Ozone ISA). These functions continue to provide very accurate predictions of
effects in naturalistic settings (see Figure 8-16 and Figure 8-17). Although these median functions provide
reliable models for groups of species or group of genotypes within a species, the original data along with
recent results consistently show that some species, and some genotypes within species are much more
severely affected by exposure to ozone.
Finally, with what was known at the time of the 2013 Ozone ISA added to the new information
reported in this ISA, the knowledge base is stronger on exposure indices and exposure-response for
vegetation. The cumulative weighted indices (W126 and AOT40) and exposure-response relationships
presented in this section continue to be used in analyses in the scientific literature and are the best
available approach for studying the effects of ozone exposure on vegetation in the U.S.
8-203
-------
Table 8-26 Exposure indices and exposure response.
Study
Study Type and
Study Location
Study Species
Ozone Exposure
Relevant Results
Betzelberaer et al. (2012)
FACE; SoyFACE,
Champaign, IL
Seven cultivars of
soybean (Glycine
max)
Soybeans in eight 20-m diameter
soyFACE plots with different O3
concentrations were exposed for ~8 h
each day in two different growing
seasons (2009, 2010). Target
concentrations were ambient, 40, 55,
70, 85, 110, 130, 160, 200 in 2009, and
ambient, 55, 70, 85, 110, 130, 150,
170, 190 in 2010. 8-, 24-, and 1-h max
mean as well as AOT40 and SUM06
calculated for each plot shown in
Table 2 of this study.
All seven cultivars showed similar responses
to O3 with the range of responses between
18 to 30 kg ha per nL/L cumulative exposure
over 40 nL/L. At the highest target
concentration of 200 nL/L (AOT40 of
67.4 ppm-h) yields were reduced 64%. This
paper improves the estimate of soybean
response from an earlier paper where one
concentration was used over multiple years
to develop an exposure-response curve. For
the first time, a significant effect on duration
of canopy and size of canopy was observed
with O3 exposure. Interception efficiency was
estimated to be reduced by 20% at the
highest target concentration.
Grantzet al. (2013)
Greenhouse;
Kearney Research
and Extension
Center, Parlier, CA
Gossypium
barbadense (Pima
cotton)
Each plant was exposed to a single
15-m pulse of O3 (0.0, 0.5, 1.0, 1.5,
2.0 |jmol/mol). Pulses were done at 2-h
intervals throughout the daylight period.
After a single pulse, plants were
returned to greenhouse bench and left
undisturbed for 6 days
All three leaf injury measures declined with
increasing dose as indications of O3 induced
injury. For chlorophyll content, early in the
photoperiod (700 h), the slope was shallow
and nonsignificant, but in midafternoon
(1,500 h), the sensitivity increased
substantially and the slope (a) became
significant (a = -1.84 m2/mmol, r2 = 0.3).
Similar results for D-R of conductance (at
1,500 h a = -1.84 m2/mmol, r2 = 0.83). The
slope for D-R of noninjured leaf area was
shallow but significant at 700 h (a = -0.54,
r2 = 0.15). The slope and its significance
increased to maxima at 1,700 h
r= 0.52).
a = -2.57,
8-204
-------
Table 8-26 (Continued): Exposure indices and exposure response.
Study
Study Type and
Study Location
Study Species
Ozone Exposure
Relevant Results
Osborne et al. (2016)
Other;
28 experimental
studies (OTC or
FACE) between
1982-2014 from the
U.S., Asia, and China
48 soybean Ozone exposure data all converted to
cultivars seasonal 7-h mean from studies that
reported concentration as 8-h mean,
12-h mean, 24-h mean, or 3-mo
AOT40. Duration of O3 exposure was at
least 60% of growing season
This study updates the exposure-response
function for O3 in soybean using data after
1998 from the U.S., China, and India and
examines temporal and geographical trends
in sensitivity. The exposure-response
function was calculated by pooling relative
yield data and plotting against the seasonal
mean at 7 h (M7). All data was scaled to
theoretical yield at 0 ppb, 55 ppb was used to
represent present-day background levels.
Relative yield reduction at present
concentration was 17.3%. significant (5%)
loss of yield can occur is 32.3 ppb M7.
Previous exposure-response function for
soybean based on U.S. data may have
underestimated yield losses in Asia where
some cultivars appear to be more sensitive.
Cultivars varied in sensitivity to ozone with a
yield loss at a 7-h mean concentration of
55 ppb ranging from 13.3 to 37.9%.
Sensitivity to O3 increased by an avg of
32.5% between 1960 and 2000. Sensitivity
was higher in India and China compared with
the U.S. Also, sensitivity has appeared
increase overtime in soybeans.
8-205
-------
Table 8-26 (Continued): Exposure indices and exposure response.
Study
Study Type and
Study Location
Study Species
Ozone Exposure
Relevant Results
Tai and Martin (2017)
Other; modeling
study using
multidecadal U.S.
crop yield and
climate data
Soybean (Glycine
max), wheat
(Triticum), maize
(Zea mays)
Three cumulative ozone annual
exposure metrics AOT40, SUM06, and
W126 calculated from hourly ozone
observations from the AQS and
CASTNET networks averaged over
1993-2010
Instead of relying on pooled concentration
response functions which do not account for
cultivar sensitivity to ozone and temperature
differences, authors developed an empirical
model (partial derivative linear regression
[PDLR] model) from multidecadal data sets
to estimate geographical variations across
the U.S. in sensitivity of wheat, maize, and
soybean to ozone. This approach takes into
consideration the strong ozone-temperature
covariation. For all three crops, the revised
sensitivities (calculated in latitude-longitude
grid cells to account for regional differences
in temperature, water, and nutrient
availability) are, in general, higher than
previously indicated by
concentration-response functions derived
from experimental studies. Wheat yield
sensitivities to ozone were statistically
significant spatially along the northern U.S.
border, maize sensitivity was spatially
statistically significant at various locations
across the U.S., and soybean sensitivity was
spatially statistically significant in a band
from the Great Plains to the south-central
U.S. Crops in regions of elevated ozone and
high-water use, were more tolerant to ozone.
The PDLR model coupled with ozone and
temperature projections from 2000 to 2050
by the Community Earth System model
predict average declines of U.S. wheat,
maize, and soybean of 13, 43, and 28%
respectively.
Feng et al. (2017)
Other; global data set
of O3 experiments in
temperate,
Mediterranean, and
subtropical climates
57 tree species for
foliar injury,
9 European tree
species for biomass
Elevated O3 experiments; O3 exposures
are expressed as AOT40 values and
Phytotoxic O3 dose
Phytotoxic O3 dose (POD) based on leaf
mass is a stronger predictor of biomass
reduction (r2 = 0.56) than is POD based on
leaf area (r2 = 0.42).
8-206
-------
Table 8-26 (Continued): Exposure indices and exposure response.
Study
Study Type and
Study Location
Study Species
Ozone Exposure
Relevant Results
Gao etal. (2017)
OTC; Seed Station
Field of Changping,
northwest Beijing,
China
O3 sensitive clone
(546) of eastern
cottonwood
(Populus deltoides)
Three ozone treatments: charcoal
filtered, ambient, and elevated O3;
Plants fumigated 96 days
(June-September). Mean C>3was
33.5 ± 2.4 ppb in the CF treatment,
51.1 ± 4.1 ppb in the NF treatment, and
78.2 ± 5.5 ppb in the E-O3 treatment.
AOT40 were 4.3, 16.0, and 38.7 ppm-h,
respectively. Two irrigation treatments
were also applied
For O3 exposure-response, which was
modeled in response to biomass changes,
model performance was significantly better
when using POD (flux) compared with
AOT40 (R2 = 0.829, p = 0.012 vs.
R2 = 0.560, p = 0.087). Using this
accumulated flux model, The O3 critical level
(CL) for preventing a 4% biomass loss in this
poplar clone under different water regimes
was between 5.27 mmol/m2 PLA and
4.09 mmol/m2 PLA, depending on which
threshold (maximum biomass at zero O3
exposure) was used.
Neufeld etal. (2018)
OTC; experiments
conducted in Boone,
NC. Rhizomes
collected from Great
Smoky Mountains
National Park and
Rocky Mountains
National Park
Rudbeckia laciniata
var. ampla and var.
digitata (cutleaf
coneflower)
Three treatment groups:
charcoal-filtered air (CF), nonfiltered air
(NF), and nonfiltered air + 50 ppb O3
(2012) or +30 ppb/+ 50 ppb (2013)
(E-Os). In 2012, 24-h W126 was
0.1 ppm-h in the CF treatment,
2.0 ppm-h in the NF treatment, and
74.2 ppb in the E-O3 treatment. 12-h
AOT40 were 0.0, 2.0, and 24.1 ppm-h,
respectively. In 2013, 24-h W126 were
0.1, 1.8, and 80.5 ppm-h, respectively.
12-h AOT40 were 1.0, 2.0, and
53.8 ppm-h, respectively. Plants were
exposed for 47 days in 2012 and for
77 days in 2013.
In 2012 and 2013, injury levels in both
varieties were higher in the E-O3 treatment
than in either the CF or NF treatments, which
did not differ, but there were no statistically
significant differences between the varieties.
Stippling increased with time. Effects of O3
on biomass accumulation were
nonsignificant.
8-207
-------
Table 8-26 (Continued): Exposure indices and exposure response.
Study
Study Type and
Study Location
Study Species
Ozone Exposure
Relevant Results
Haves et al. (2011)
Greenhouse; near
Marchlyn Mawr, U.K.
Two communities:
four plants of forb
Leontodon hispidus
and three plants of
grass Dactylis
glomerata', four
plants of forb
Leontodon hispidus
and three plants of
Anthoxanthum
odoratum
Eight treatments: (1) seasonal 24-h
mean 21.4 ppb (12-h mean 21.1 ppb,
daylight [7:00 a.m.-6:00 p.m.]
AOT40 = 0.07 ppm-h, 24-h
AOT40 = 0.07 ppm-h); (2) seasonal
mean 39.9 ppb (12 h = 39.2 ppb,
daylight AOT40 = 4.93 ppm-h, 24-h
AOT40 = 10.91 ppm-h); (3) seasonal
mean 50.2 ppb (12 h = 49.6 ppb,
daylight AOT40 = 21.44 ppm-h, 24-h
AOT40 = 41.29 ppm-h); (4) seasonal
mean 59.4 ppb (12 h = 58.7 ppb,
daylight AOT40 = 38.04 ppm-h, 24-h
AOT40 = 72.19 ppm-h); (5) seasonal
mean 74.9 ppb (12 h = 73.3 ppb,
daylight AOT40 = 62.49 ppm-h, 24-h
AOT40 = 119.82 ppm-h); (6) seasonal
mean 83.3 ppb (12 h = 81.6 ppb,
daylight AOT40 = 77.13 ppm-h, 24-h
AOT40 = 147.42 ppm-h); (7) seasonal
mean 101.3 ppb (12 h = 99.0 ppb,
daylight AOT40 = 108.43 ppm-h, 24-h
AOT40 = 206.70 ppm-h); (8) seasonal
mean 102.5 ppb (12 h = 100.5, daylight
AOT40 = 112.47 ppm-h, 24-h
AOT40 = 214.34 ppm-h)
Exposure indices: there was a linear
relationship between 24-h mean ozone and
12-h mean ozone treatments (r2 = 0.9999).
There were linear relationships between
seasonal O3 concentration and root biomass,
leaf retention, reproductive phenology in the
following season, and grass cover.
Sanz et al. (2016)
OTC; OTC
experimental field
located in a rural
area in the northeast
of the Iberian
Peninsula,
Tarragona, Spain
Pasture species, Data analyzed from independent
Leguminosae (three experiments, 45-day avg O3 exposure
species) length
An O3 critical level for reproductive capacity
AOT40 = 2.0 (1.5, 2.8) ppm-h and phytotoxic
ozone dose (POD) 1 = 7.2 (1.1,
13.3) mmol/m2 was developed from linear
exposure-response functions based on seed
and flower production (see Figure 1 for
AOT40 and Figure 2 for POD1).
Reproductive capacity had the lowest critical
level of the endpoints evaluated.
8-208
-------
Table 8-26 (Continued): Exposure indices and exposure response.
Study
Study Type and
Study Location
Study Species
Ozone Exposure
Relevant Results
van Goethem et al. (2013)
Other; northwestern
Europe (mapping of
sensitivity is for a
square area of 50 to
61°N, and 11°Eto
11°W)
25 annual
grassland species,
62 perennial
grassland species,
9 tree species
OTC, FACE, or solardomes. All
experimental treatments were at
>40 ppb for at least 21 days, with mean
hourly O3 never exceeding 100 ppb.
Control treatments were
charcoal-filtered air or ambient air
Annual grassland species were significantly
more sensitive to O3 >40 ppb than were
perennial grassland species. Mean 10%
reduction in biomass occurred at 0.84 ppm-h
for annual species and 1.14 ppm-h for
perennial grassland species.
Exposure-response relationships for
96 European plants (biomass reduction vs.
AOT40) reported.
Abeli et al. (2017)
Lab; alpine seeds
collected on Mt.
Cimone, Mt.
Prado-Cusna, and in
the Dolomites in Italy;
O3 exposure inside
incubators
Achillea clavennae
Aster alpinus,
Festuca rubra
subs p. commutata
(Gaudin) Markgr.-
Dann, Festuca
violacea subs p.
puccinellii (Pari.)
Foggi, GrazRossi &
Signorini, Plantago
alpina L., Silene
acaulis (L.) J acq.,
Silene nutans L.,
Silene suecica
(Lodd) Greuter &
Burdet, Vaccinium
myrtillus L.
Control: ambient air (0-1 ppb),
125_5 treatment: 125 ppb O3 24 h/day
for 5 days, 125_10 treatment: 125 ppb
O3 24 h/day for 10 days,
185_5 treatment: 185 ppb O3 24 h/day
for 5 days
Combining all species, each treatment
(compared to control) significantly delayed
germination (125_5 = 0.71, 185_5 = 0.87,
125_10 = 1.17-day delay). Individual species
varied in their responses. Six of nine species
had reduction in germination percentage for
one or more of the O3 treatments at end of
the O3 exposure. Seven of nine species
showed significant effect of at least one O3
treatment at 28 days after sowing, and effect
ranged from increasing to decreasing
germination percentage. Combining all
species, 125_5 and 185_5 treatments did not
affect mean germination time either at the
end of the O3 exposure or at the end of the
experiment. The 125_10 treatment
significantly increased mean germination
time by 1.25 days after O3 exposure, but by
the end of the experiment, that difference did
not exist.
8-209
-------
Table 8-26 (Continued): Exposure indices and exposure response.
Study
Study Type and
Study Location
Study Species
Ozone Exposure
Relevant Results
Pavne et al. (2017)
Mesocosm; peat
sampled from wet,
heathy peatland,
U.K.
Microscopic algae
(desmids, diatoms),
protozoa (ciliates,
flagellates, testate
amoebae), and
microscopic animal
consumers
(nematodes,
rotifers) sampled
from Sphagnum
papillosum stems
Experimental O3 for 3.5 yr: ambient
(avg 25 ppb O3), low O3
(ambient + 10 ppb for 24 h/day),
moderate O3 (ambient + 25 ppb O3
24 h/day), elevated O3
(ambient + 35 ppb daytime 8 h/day in
summer, +10 ppb rest of year)
Authors indicated that O3 effects on
microscopic food web in peat generally start
at moderate O3 exposures.
AOT40 = seasonal sum of the difference between an hourly concentration at the threshold value of 40 ppb, minus the threshold value of 40 ppb; CF = charcoal-filtered air; D = dose;
E-O3 = elevated-ozone treatment; kg/ha = kilograms per hectare; NF = nonfiltered air; nL/L = nanoliters per liter; OTC = open-top chamber; PLA = projected leaf area; ppm = parts
per million; S = sensitivity; SUM06 = seasonal sum of all hourly average concentrations > 0.06 ppm; |jmol/mol = micromoies/mole; W126 = cumulative integrated exposure index with
a sigmoidal weighting function.
8-210
-------
8.14 References
Abeli. T; Guasconi. DB; Mondoni. A; Dondi. D; Bentivoglio. A; Buttafava. A; Cristofanelli. P;
Bonasoni. P; Rossi. G; Orsenigo. S. (2017). Acute and chronic ozone exposure temporarily affects
seed germination in alpine plants. Plant Biosyst 151: 304-315.
http://dx.doi.org/10.1080/11263504.2Q16.1174169
Adams. MB: Kelly. JM: Taylor. GE. Jr: Edwards. NT. (1990). Growth of five families of Pinus taeda
L during three years of ozone exposure. New Phytol 116: 689-694.
http://dx.doi.Org/10.llll/i.1469-8137.1990.tb00555.x
Adams. MB: O'Neill. EG. (1991). Effects of ozone and acidic deposition on carbon allocation and
mycorrhizal colonization of Pinus taeda L seedlings. Forest Sci 37: 5-16.
Agathokleous. E; Saitanis. CJ: Koike. T. (2015). Tropospheric 0-3, the nightmare of wild plants: a
review study. Journal of Agricultural Meteorology 71: 142-152.
http://dx.doi.org/10.2480/agrmet.D-14-000Q8
Agathokleous. E; Sakikawa. T; Abu ElEla. SA; Mochizuki. T; Nakamura. M; Watanabe. M;
Kawamura. K: Koike. T. (2017). Ozone alters the feeding behavior of the leaf beetle Agelastica
coerulea (Coleoptera: Chrysomelidae) into leaves of Japanese white birch (Betula platyphylla var.
japonica). Environ Sci Pollut Res Int 24: 17577-17583. http://dx.doi.org/10.1007/sl 1356-017-
9369-7
Ainsworth. EA. (2008). Rice production in a changing climate: a meta-analysis of responses to
elevated carbon dioxide and elevated ozone concentration. Global Change Biol 14: 1642-1650.
http://dx.doi.org/10.1111/i. 1365-2486.2008.01594 ,x
Ainsworth. EA: Serbin. SP: Skoneczka. JA: Townsend. PA. (2014). Using leaf optical properties to
detect ozone effects on foliar biochemistry. Photosynth Res 119: 65-76.
http://dx.doi.org/10.1007/slll20-Q13-9837-v
Allen. EB; Temple. PJ: Bytnerowicz. A: Arbaugh. MJ; Sirulnik. AG: Rao. LE. (2007). Patterns of
understory diversity in mixed coniferous forests of southern California impacted by air pollution.
ScientificWorldJournal 7: 247-263. http://dx.doi.org/10.1100/tsw.20Q7.72
Amundson. RG: Alscher. RG: Fellows. S: Rubin. G: Fincher. J: Van Leuken. P; Weinstein. LH.
(1991). Seasonal changes in the pigments, carbohydrates and growth of red spruce as affected by
exposure to ozone for two growing seasons. New Phytol 118: 127-137.
http://dx.doi.Org/10.llll/i.1469-8137.1991.tb00573.x
Anav. A: De Marco. A: Sicard. P: Vitale. M: Paoletti. E. (2017). Response on 'comparing
concentration-based (AOT40) and stomatal uptake (PODY) metrics for ozone risk assessment to
European forests' [Comment]. Global Change Biol 23: e3-e4. http://dx.doi.org/10.Ill 1/gcb. 13511
Anav. A: Liu. Q: De Marco. A: Proietti. C: Savi. F; Paoletti. E; Piao. S. (2018). The role of plant
phenology in stomatal ozone flux modeling. Global Change Biol 24: 235-248.
http://dx.doi.Org/10.l 111/gcb. 13 823
Andersen. CP. (2003). Source-sink balance and carbon allocation below ground in plants exposed to
ozone. New Phytol 157: 213-228. http://dx.doi.Org/10.1046/i.1469-8137.2003.00674.x
Andrew. C: Lilleskov. EA. (2014). Elevated C02 and 03 effects on ectomycorrhizal fungal root tip
communities in consideration of a post-agricultural soil nutrient gradient legacy. Mycorrhiza 24:
581-593. http://dx.doi.org/10.1007/s00572-014-Q577-4
8-211
-------
Andrew. CJ; van Diepen. LTA; Miller. RM: Lilleskov. EA. (2014). Aspen-associated mycorrhizal
fungal production and respiration as a function of changing C02, 03 and climatic variables. Fungal
Ecol 10: 70-80. http://dx.doi.Org/10.1016/i.funeco.2013.10.005
Arbaugh. M; Bvtnerowicz. A; Grulke. N; Fenn. M; Poth. M; Temple. P; Miller. P. (2003).
Photochemical smog effects in mixed conifer forests along a natural gradient of ozone and nitrogen
deposition in the San Bernardino Mountains. Environ Int 29: 401-406.
http://dx.doi.org/10.1016/S0160-4120(02)00176-9
Ashmore. MR; Bell. JNB; Mimmack. A. (1988). Crop growth along a gradient of ambient air
pollution. Environ Pollut 53: 99-121. http://dx.doi.org/10.1016/0269-7491(88)90028-0
Ashmore. MR: Power. SA: Cousins. DA: Ainsworth. N. (1996). Effects of ozone on native grass and
forb species: A comparison of responses of individual plants and artificial communities. In L
Karenlampi; L Skarby (Eds.), Critical levels for ozone: testing and finalizing the concepts UNECE
workshop report (pp. 193-197). Kuopio, Finland: United Nations Economic Commission for
Europe (UN-ECE).
Ashmore. MR: Thwaites. RH: Ainsworth. N: Cousins. DA: Power. SA: Morton. AJ. (1995). Effects of
ozone on calcareous grassland communities. Water Air Soil Pollut 85: 1527-1532.
http://dx.doi.org/10.1007/BF0Q477198
Avnery. S: Mauzerall. PL: Liu. J: Horowitz. LW. (2011). Global crop yield reductions due to surface
ozone exposure: 1. Year 2000 crop production losses and economic damage. Atmos Environ 45:
2284-2296. http://dx.doi.Org/10.1016/i.atmosenv.2010.l 1.045
Awmack. CS: Harrington. R; Lindroth. RL. (2004). Aphid individual performance may not predict
population responses to elevated C02 or 03. Global Change Biol Biol. 10: 1414-1423.
http://dx.doi.org/10.1111/i. 1365-2486.2004.00800.X
Baldantoni. D: Fagnano. M: Alfani. A. (2011). Tropospheric ozone effects on chemical composition
and decomposition rate of Quercus ilex L. leaves. Sci Total Environ 409: 979-984.
http ://dx.doi .org/10.1016/i. scitotenv.2010.11.022
Bao. X: Li. O: Hua. J: Zhao. T: Liang. W. (2014). Interactive effects of elevated ozone and UV-B
radiation on soil nematode diversity. Ecotoxicology 23: 11-20. http://dx.doi.org/10.1007/slQ646-
013-1146-x
Bassin. S: Yolk. M: Fuhrer. J. (2013). Species composition of subalpine grassland is sensitive to
nitrogen deposition, but not to ozone, after seven years of treatment. Ecosystems 16: 1105-1117.
http://dx.doi.org/10.1007/sl0021-013-967Q-3
Bassin. S: Yolk. M: Suter. M: Buchmann. N: Fuhrer. J. (2007). Nitrogen deposition but not ozone
affects productivity and community composition of subalpine grassland after 3 yr of treatment.
New Phytol 175: 523-534. http://dx.doi.Org/10.llll/i.1469-8137.2007.02140.x
Battv. K: Ashmore. M: Power. SA. (2001). Assessment of relative sensitivity of wetland plant species
to ozone. In The ozone umbrella project (CEH project C00970): Final report. Edinburgh, Scotland:
Centre for Ecology & Hydrology.
Bell. JNB: Ashmore. MR: Wilson. GB. (1991). Ecological genetics and chemical modifications of the
atmosphere. In GE Taylor, Jr.; LF Pitelka; MT Clegg (Eds.), Ecological genetics and air pollution
(pp. 33-59). New York, NY: Springer-Verlag.
Bender. J: Bergmann. E; Dohrmann. A: Tebbe. CC: Weigel. HJ. (2002). Impact of ozone on plant
competition and structural diversity of rhizosphere microbial communities in grassland mesocosms.
Phyton-Annales Rei Botanicae 42: 7-12.
8-212
-------
Bender. J; Bcrgmann. E; Wei gel. HJ. (2006). Responses of biomass production and reproductive
development to ozone exposure differ between european wild plant species. Water Air Soil Pollut
176: 253-267. http://dx.doi.org/10.1007/sll270-0Q6-9167-l
Bennett. JP; Jepsen. EA; Roth. JA. (2006). Field responses of Prunus serotina and Asclepias syriacato
ozone around southern Lake Michigan. Environ Pollut 142: 354-366.
http://dx.doi.Org/10.1016/i.envpol.2005.09.024
Benoit. LF: Skellv. JM: Moore. LP: Dochinger. LS. (1982). Radial growth reductions of Pinus strobus
L correlated with foliar ozone sensitivity as an indicator of ozone-induced losses in eastern forests.
Can J For Res 12: 673-678. http://dx.doi.org/10.1139/x82-101
Bergmann. E: Bender. J: Weigel. HJ. (1995). Growth responses and foliar sensitivities of native
herbaceous species to ozone exposures. Water Air Soil Pollut 85: 1437-1442.
http://dx.doi.org/10.1007/BF0Q477183
Bergmann. E: Bender. J: Weigel. HJ. (1999). Ozone threshold doses and exposure-response
relationships for the development of ozone injury symptoms in wild plant species. New Phytol 144:
423-435. http://dx.doi.Org/10.1046/i.1469-8137.1999.00534.x
Bergmann. E: Bender. J: Weigel. HJ. (2017). Impact of tropospheric ozone on terrestrial biodiversity:
A literature analysis to identify ozone sensitive taxa. J Appl Bot Food Qual 90: 83-105.
http://dx.doi.org/10.5073/JABFO.2017.090.Q12
Bergmann. E: Bender. J: Weigel. HJ. (1996a). Effects of chronic ozone stress on growth and
reproduction capacity of native herbaceous plants. In JS M. Knopflacher; G Soja (Eds.),
Exceedance of Critical Loads and Levels - Spatial and Temporal Interpretation of Elements in
Landscape Sensitivity to Atmospheric Pollutants (pp. 177-185). Vienna, Austria:
Umweltbundesamt Osterreich.
Bergmann. E: Bender. J: Weigel. HJ. (1996b). Ozone and natural vegetation: Native species sensitivity
to different ozone exposure regimes. In L Karenlampi; L Skarby (Eds.), Critical levels for ozone in
Europe: Testing and finalizing the concepts, UNECE workshop report (pp. 205-209). Kuopio,
Finland: University of Kuopio.
Bergweiler. CJ; Manning. WJ. (1999). Inhibition of flowering and reproductive success in spreading
dogbane (Apocynum androsaemifolium) by exposure to ambient ozone. Environ Pollut 105: 333-
339. http://dx.doi.org/10.1016/S0269-7491(99)00044-5
Bergweiler. CJ; Manning. WJ; Chevone. BI. (2008). Seasonal and diurnal gas exchange differences in
ozone-sensitive common milkweed (Asclepias syriaca L.) in relation to ozone uptake. Environ
Pollut 152: 403-415. http://dx.doi.Org/10.1016/i.envpol.2007.06.019
Bernacchi. CJ; Leakey. AD; Kimball. BA; Ort. DR. (2011). Growth of soybean at future tropospheric
ozone concentrations decreases canopy evapotranspiration and soil water depletion. Environ Pollut
159: 1464-1472. http://dx.doi.Org/10.1016/i.envpol.2011.03.011
Betzelberger. A; Yendrek. CR; Sun. J; Leisner. CP; Nelson. RL; Ort. PR: Ainsworth. EA. (2012).
Ozone exposure response for U.S. soybean cultivars: Linear reductions in photosynthetic potential,
biomass, and yield. Plant Physiol 160: 1827-1839. http://dx.doi.Org/10.1104/pp.l 12.205591
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.org/ 10.1111/i.l365-
3040.2010.02165.x
8-213
-------
Beyers. JL; Riechers. GH; Temple. PJ. (1992). Effects of long-term ozone exposure and drought on
the photosynthetic capacity of ponderosa pine (Pinus ponderosa Laws). New Phytol 122: 81-90.
http://dx.doi.Org/10.llll/i.1469-8137.1992.tb00055.x
Black. VJ; Black. CR; Roberts. JA; Stewart. CA. (2000). Tansley Review No. 115: Impact of ozone
on the reproductive development of plants. New Phytol 147: 421-447.
http://dx.doi.Org/10.1046/i.1469-8137.2000.00721.x
Black. VJ: Stewart. CA: Roberts. JA: Black. CR. (2012). Timing of exposure to ozone affects
reproductive sensitivity and compensatory ability in Brassica campestris. Environ Exp Bot 75: 225-
234. http://dx.doi.Org/10.1016/i.envexpbot.2011.07.006
Blande. JD: Holopainen. JK: Li. T. (2010). Air pollution impedes plant-to-plant communication by
volatiles. EcolLettl3: 1172-1181. http://dx.doi.Org/10.llll/i.1461-0248.2010.01510.x
Blande. JD: Holopainen. JK: Niinemets. U. (2014). Plant volatiles in polluted atmospheres: stress
responses and signal degradation [Review]. Plant Cell Environ 37: 1892-1904.
http://dx.doi.org/10.1111/pce. 12352
Blum. O: Bytnerowicz. A: Manning. W: Popovicheva. L. (1997). Ambient tropospheric ozone in the
Ukrainian Carpathian mountains and kiev region: detection with passive samplers and bioindicator
plants. Environ Pollut 98: 299-304. http://dx.doi.org/10.1016/S0269-7491(97)00158-9
Bohler. S: Sergeant. K; Jolivet. Y; Hoffmann. L; Hausman. JF; Dizengremel. P; Renaut. J. (2013). A
physiological and proteomic study of poplar leaves during ozone exposure combined with mild
drought. Proteomics 13: 1737-1754. http://dx.doi.org/10.1002/pmic.20120Q193
Bolsinger. M; Lier. ME: Hughes. PR. (1992). Influence of ozone air pollution on plant-herbivore
interactions Part 2: effects of ozone on feeding preference, growth and consumption rates of
monarch butterflies (Danaus plexippus). Environ Pollut 77: 31-37. http://dx.doi.org/10.1016/Q269-
7491(92)90155-4
Bolsinger. M; Lier. ME: Lansky. DM: Hughes. PR. (1991). Influence of ozone air pollution on plant-
herbivore interactions Part I: Biochemical changes in ornamental milkweed (Asclepias curassavica
L; Asclepiadaceae) induced by ozone. Environ Pollut 72: 69-83. http://dx.doi.org/10.1016/Q269-
7491(91)90156-0
Bradley. KL; Pregitzer. KS. (2007). Ecosystem assembly and terrestrial carbon balance under elevated
C02. Trends Ecol Evol 22: 538-547. http://dx.doi.Org/10.1016/i.tree.2007.08.005
Broberg. MC: Feng. Z; Xin. Y; Pleiiel. H. (2015). Ozone effects on wheat grain quality - a summary
[Review]. Environ Pollut 197: 203-213. http://dx.doi.Org/10.1016/i.envpol.2014.12.009
Brown. VC: McNeill. S: Ashmore. MR. (1992). The effects of ozone fumigation on the performance
of the black bean aphid, Aphis fabae Scop, feeding on broad beans, Vicia faba L. Agric Ecosyst
Environ 38: 71-78.
Bungener. P: Balls. GR: Nussbaum. S: Geissmann. M: Grub. A: Fuhrer. J. (1999a). Leaf injury
characteristics of grassland species exposed to ozone in relation to soil moisture condition and
vapour pressure deficit. New Phytol 142: 271-282. http://dx.doi.org/10.1046/i. 1469-
8137.1999.00390.x
Bungener. P: Nussbaum. S: Grub. A: Fuhrer. J. (1999b). Growth response of grassland species to
ozone in relation to soil moisture and plant strategy. New Phytol 142: 283-293.
http://dx.doi.Org/10.1046/i.1469-8137.1999.00389.x
8-214
-------
Burkev. KO; Booker. FL; Ainsworth. EA; Nelson. RL. (2012). Field assessment of a snap bean ozone
bioindicator system under elevated ozone and carbon dioxide in a free air system. Environ Pollut
166: 167-171. http://dx.doi.Org/10.1016/i.envpol.2012.03.020
Burton. A; Burkev. KO; Carter. TE. Jr; Orf. J; Crcgan. PB. (2016). Phenotypic variation and
identification of quantitative trait loci for ozone tolerance in a Fiskeby III x Mandarin (Ottawa)
soybean population. Theor Appl Genet 129: 1113-1125. http://dx.doi.org/10.1007/sQ0122-016-
2687-1
Bussotti. F; Cozzi. A; Bettini. D. (2003a). Ozone-like visible foliar symptoms at the permanent
monitoring plots of the CONECO-FOR programme in Italy. Ann 1st Speriment Selvicoltura 30: 99-
106.
Bussotti. F; Cozzi. A; Ferretti. M. (2006). Field surveys of ozone symptoms on spontaneous
vegetation. Limitations and potentialities of the European programme. Environ Monit Assess 115:
335-348. http://dx.doi.org/10.1007/sl0661-0Q6-6558-0
Bussotti. F; Gerosa. G. (2002). Are the Mediterranean forests in Southern Europe threatened from
ozone? J Mediterr Ecol 3: 23-34.
Bussotti. F: Mazzali. C; Cozzi. A: Ferretti. M: Gravano. E: Gerosa. G; Ballarin-Denti. A. (2003b).
Ozone injury symptoms on vegetation in an Alpine valley, North Italy. In Developments in
Environmental Science.
Byres. DP; Dean. TJ: Johnson. JD. (1992). Long-term effects of ozone and simulated acid rain on the
foliage dynamics of slash pine (Pinus elliottii var elliottii Engelm). New Phytol 120: 61-67.
http://dx.doi.Org/10.llll/i.1469-8137.1992.tb01058.x
Bytnerowicz. A; Olszvk. DM; Fox. CA; Dawson. PJ; Kats. G; Morrison. CL; Wolf. J. (1988).
Responses of desert annual plants to ozone and water stress in an in situ experiment. J Air Waste
Manag Assoc 38: 1145-1151. http://dx.doi.org/10.1080/0894063Q.1988.10466463
Calatavud. V; Garcia-Breiio. FJ; Cervero. J; Reig-Arminana. J; Sanz. MJ. (2011). Physiological,
anatomical and biomass partitioning responses to ozone in the Mediterranean endemic plant
Lamottea dianae. Ecotoxicol Environ Saf 74: 1131-1138.
http://dx.doi.Org/10.1016/i.ecoenv.2011.02.023
Calvete-Sogo. H; Gonzalez-Fernandez. I; Sanz. J; Elvira. S; Alonso. R; Garcia-Gomez. H; Ibanez-
Ruiz. MA; Bermeio-Bermeio. V. (2016). Heterogeneous responses to ozone and nitrogen alter the
species composition of Mediterranean annual pastures. Oecologia 181: 1055-1067.
http://dx.doi.org/10.1007/sQ0442-016-3628-z
Cannon Jr. W; Roberts. B; Barger. J. (1993). Growth and physiological response of water-stressed
yellow-poplar seedlings exposed to chronic ozone fumigation and ethylenediurea. For Ecol
Manage 61: 61-73. http://dx.doi.org/10.1016/0378-l 127(93)90190-X
Cannon. WN. (1990). Olfactory response of eastern spruce budworm larvae to red spruce needles
exposed to acid rain and elevated levels of ozone. J Chem Ecol 16: 3255-3261.
http://dx.doi.org/10.1007/BF00982Q96
Cano. I; Calatavud. V; Cervero. J; Sanz. MJ. (2007). Ozone effects on three Sambucus species.
Environ Monit Assess 128: 83-91. http://dx.doi.org/10.1007/slQ661-006-9417-0
Capps. SL; Driscoll. CT; Fakhraei. H; Templer. PH; Craig. KJ; Milford. JB; Lambert. KF. (2016).
Estimating potential productivity cobenefits for crops and trees from reduced ozone with US coal
power plant carbon standards. J Geophys Res Atmos 121: 14679-14690.
http://dx.doi.org/10.1002/2016JDQ25141
8-215
-------
Ceron-Breton. JG; Ceron-Breton. RM: Guerra-Santos. JJ; Aguilar-Ucan. C; Montalvo-Romero. C;
Acosta-Morales. D; Calderon-Canto. R. (2009). Ozone-exposure responses of three Mangrove
species by using open top chambers. In Energy and Environmental Engineering Series.
Chapin. FS. Ill; Matson. PA; Moonev. HA. (2002). Principles of terrestrial ecosystem ecology. New
York, NY: Springer-Verlag New York, Inc.
Chappelka. A; Renfro. J; Somers. G; Nash. B. (1997). Evaluation of ozone injury on foliage of black
cherry (Prunus serotina) and tall milkweed (Asclepias exaltata) in Great Smoky Mountains
National Park. Environ Pollut 95: 13-18. http://dx.doi.org/10.1016/S0269-7491(96)00120-0
Chappelka. A; Skellv. J; Somers. G; Renfro. J; Hildebrand. E. (1999a). Mature black cherry used as a
bioindicator of ozone injury. Water Air Soil Pollut 116: 261-266.
http://dx.doi.Org/10.1023/A:1005260422738
Chappelka. A; Somers. G; Renfro. J. (1999b). Visible ozone injury on forest trees in Great Smoky
Mountains National Park, USA. Water Air Soil Pollut 116: 255-260.
http://dx.doi.org/10.1023/A: 1005204305900
Chappelka. AH. (2002). Reproductive development of blackberry (Rubus cuneifolius) as influenced
by ozone. New Phytol 155: 249-255. http://dx.doi.org/10.1046/i. 1469-8137.2002.00464.x
Chappelka. AH; Kush. JS; Meldahl. RS; Lockabv. BG. (1990). An ozone-low temperature interaction
in loblolly pine (Pinus taeda L). New Phytol 114: 721-726. http://dx.doi.org/10.1111/i.1469-
8137.1990.tb00444.x
Chappelka. AH; Neufeld. HS; Davison. AW; Somers. GL; Renfro. JR. (2003). Ozone injury on cutleaf
coneflower (Rudbeckia laciniata) and crown-beard (Verbesina occidentalis) in Great Smoky
Mountains National Park. Environ Pollut 125: 53-60. http://dx.doi.org/10.1016/SQ269-
7491(03)00086-1
Chappelka. AH; Samuelson. LJ. (1998). Ambient ozone effects on forest trees of the eastern United
States: A review [Review]. New Phytol 139: 91-108. http://dx.doi.org/10.1046/i. 1469-
8137.1998.00166.x
Chappelka. AH; Somers. GL; Renfro. JR. (2007). Temporal patterns of foliar ozone symptoms on tall
milkweed (Asclepias exaltata L) in Great Smoky Mountains National Park. Environ Pollut 149:
358-365. http://dx.doi.Org/10.1016/i.envpol.2007.05.015
Chen. Z; Wang. X; Shang. H. (2015). Structure and function of rhizosphere and non-rhizosphere soil
microbial community respond differently to elevated ozone in field-planted wheat. J Environ Sci
32: 126-134. http://dx.doi.Org/10.1016/i.ies.2014.12.018
Chen. Z; Wang. XK; Yao. FF; Zheng. FX; Feng. ZZ. (2010). Elevated ozone changed soil microbial
community in a rice paddy. Soil Sci Soc Am J 74: 829-837.
http://dx.doi.org/10.2136/sssai2009.0258
Cheng. L; Booker. FL; Burkev. KO; Tu. C; Shew. HP; Ruftv. TW; Fiscus. EL; Deforest. JL; Hu. S.
(2011). Soil microbial responses to elevated C02 and 03 in a nitrogen-aggrading agroecosystem.
PLoS ONE 6: e21377. http://dx.doi.org/10.1371/iournal.pone.0021377
Chieppa. J; Chappelka. A; Eckhardt. L. (2015). Effects of tropospheric ozone on loblolly pine
seedlings inoculated with root infecting ophiostomatoid fungi. Environ Pollut 207: 130-137.
http://dx.doi.Org/10.1016/i.envpol.2015.08.053
Chung. HG; Zak. PR; Lilleskov. EA. (2006). Fungal community composition and metabolism under
elevated C02 and 03. Oecologia 147: 143-154. http://dx.doi.org/10.1007/sQ0442-005-0249-3
8-216
-------
Coleman. JS; Jones. CG. (1988). Plant stress and insect performance: cottonwood, ozone and a leaf
beetle. Oecologia76: 57-61.
Coleman. MP; Dickson. RE; Isebrands. JG; Karnoskv. DF. (1996). Root growth physiology of potted
and field-grown trembling aspen exposed to tropospheric ozone. Tree Physiol 16: 145-152.
http://dx.doi.org/10.1093/treephvs/16.l-2.145
Cornelissen. T. (2011). Climate change and its effects on terrestrial insects and herbivory patterns
[Review], Neotrop Entomol 40: 155-163. http://dx.doi.org/10.1590/S1519-566X2011000200Q01
Costa. SD; Kennedy. GG; Heagle. AS. (2001). Effect of host plant ozone stress on Colorado potato
beetles. Environ Entomol 30: 824-831. http://dx.doi.Org/10.1603/0046-225X-30.5.824
Costanza. R: De Groot. R: Braat. L; Kubiszewski. I; Fioramonti. L; Sutton. P; Farber. S: Grasso. M.
(2017). Twenty years of ecosystem services: How far have we come and how far do we still need
to go? Ecosyst Serv 28: 1-16. http://dx.doi.Org/10.1016/i.ecoser.2017.09.008
Cotton. TEA; Fitter. AH; Miller. RM: Dumbrell. AJ; Helgason. T. (2015). Fungi in the future:
interannual variation and effects of atmospheric change on arbuscular mycorrhizal fungal
communities. New Phytol 205: 1598-1607. http://dx.doi.Org/10.l 111/nph. 13224
Couture. JJ; Lindroth. RL. (2012). Atmospheric change alters performance of an invasive forest insect.
Global Change Biol 18: 3543-3557. http://dx.doi.org/10.ll 1 l/gcb.12014
Couture. JJ; Lindroth. RL. (2014). Atmospheric change alters frass quality of forest canopy
herbivores. Arthropod-Plant Inte 8: 33-47. http://dx.doi.org/10.1007/sl 1829-013-9286-8
Couture. JJ; Meehan. TP; Kruger. EL; Lindroth. RL. (2015). Insect herbivory alters impact of
atmospheric change on northern temperate forests. 1: 15016.
http://dx.doi.org/10.1038/NPLANTS.2015.16
Couture. JJ; Meehan. TP; Lindroth. RL. (2012). Atmospheric change alters foliar quality of host trees
and performance of two outbreak insect species. Oecologia 168: 863-876.
http://dx.doi.org/10.1007/sQ0442-011-2139-l
Cui. H; Su. J; Wei. J; Hu. Y; Ge. F. (2014). Elevated 0-3 enhances the attraction of whitefly-infested
tomato plants to Encarsia formosa. Sci Rep 4: 5350. http://dx.doi.org/10.1038/srep05350
Cui. H; Sun. Y; Chen. F; Zhang. Y; Ge. F. (2016a). Elevated 03 and TYLCV infection reduce the
suitability of tomato as a host for the whitefly Bemisiatabaci. International Journal of Molecular
Sciences 17: 1964. http://dx.doi.org/10.3390/iimsl7121964
Cui. H; Sun. Y; Su. J; Ren. Q; Li. C; Ge. F. (2012). Elevated 0-3 reduces the fitness of Bemisia tabaci
via enhancement of the SA-dependent defense of the tomato plant. Arthropod-Plant Inte 6: 425-
437. http://dx.doi.org/10.1007/sll829-Q12-9189-0
Cui. H; Wei. J; Su. J; Li. C; Ge. F. (2016b). Elevated 03 increases volatile organic compounds via
jasmonic acid pathway that promote the preference of parasitoid Encarsia formosa for tomato
plants. Plant Sci 253: 243-250. http://dx.doi.Org/10.1016/i.plantsci.2016.09.019
Curtis. PS; Wang. X. (1998). A meta-analysis of elevated CO 2 effects on woody plant mass, form,
and physiology. Oecologia 113: 299-313. http://dx.doi.org/10.1007/sQ04420050381
Panielsson. H; Gelang. J; Pleiiel. H. (1999). Ozone sensitivity, growth and flower development in
Phleum genotypes of different geographic origin in the Nordic countries. Environ Exp Bot 42: 41-
49. http://dx.doi.org/10.1016/S0098-8472(99')00016-7
8-217
-------
Darbah. INT; Kubiske. ME; Neilson. N; Oksanen. E; Vaapavuori. E; Karnoskv. DF. (2007). Impacts
of elevated atmospheric C02 and 03 on paper birch (Betula papyrifera): Reproductive fitness.
ScientificWorldJournal 7: 240-246. http://dx.doi.org/10.1100/tsw.20Q7.42
Darbah. INT; Kubiske. ME; Nelson. N; Oksanen. E; Vapaavuori. E; Karnoskv. DF. (2008). Effects of
decadal exposure to interacting elevated C02 and/or 0-3 on paper birch (Betula papyrifera)
reproduction. Environ Pollut 155: 446-452. http://dx.doi.Org/10.1016/i.envpol.2008.01.033
Darrall. NM. (1989). The effect of air pollutants on physiological processes in plants. Plant Cell
Environ 12: 1-30. http://dx.doi.Org/10.llll/i.1365-3040.1989.tb01913.x
Davis. DP. (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) 14r403:OSOVWT12.Q.CO;2
Davis. DP. (2007b). Ozone injury to plants within the Seney National Wildlife Refuge in northern
Michigan. Northeast Nat 14: 415-424. http://dx.doi.org/10.1656/lQ92-
6194(2007) 14T415 :OITPWT12.0.CO;2
Pavis. PP. (2009). Ozone-induced stipple on plants in the Cape Romain National Wildlife Refuge,
South Carolina. Southeastern Naturalist 8: 471-478. http://dx.doi.org/10.1656/058.008.03Q8
Pavis. PP. (2011). Ozone-induced leaf symptoms on vegetation in the Mingo National Wildlife
Refuge, Missouri. Northeast Nat 18: 115-122. http://dx.doi.org/10.1656/045.Q18.0111
Pavis. PP; 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.Org/10.1016/i.envpol.2005.10.051
Pavis. PP; Skellv. JM. (1992). Growth response of four species of eastern hardwood tree seedlings
exposed to ozone, acidic precipitation, and sulfur dioxide. J Air Waste Manag Assoc 42: 309-311.
Pavison. AW; Neufeld. HS; Chappelka. AH; Wolff. KW; Finkelstein. PL. (2003). Interpreting spatial
variation in ozone symptoms shown by cutleaf cone flower, Rudbeckia laciniata L. Environ Pollut
125: 61-70. http://dx.doi.org/10.1016/S0269-7491(03)00087-3
Pe Bauer. LI; Hernandez-Teieda. T; Skellv. JM. (2000). Air pollution problems in the forested areas
of Mexico and Central America. In JL Innes; AH Haron (Eds.), Air pollution and the forests of
developing and rapidly industrializing regions Report No 4 of the IUFRO Task Force on
Environmental Change (pp. 35-61). Vienna, Austria: CABI Pub.; Published in association with the
International Union of Forestry Research Organizations, c2000.
http://dx.doi.org/10.1079/9780851994819.0Q35
de Lourdes de Bauer. M; Hernandez-Teieda. T. (2007). A review of ozone-induced effects on the
forests of central Mexico [Review]. Environ Pollut 147: 446-453.
http://dx.doi.Org/10.1016/i.envpol.2006.12.020
de Vries. W; Posch. M; Simpson. P; Reinds. GJ. (2017). Modelling long-term impacts of changes in
climate, nitrogen deposition and ozone exposure on carbon sequestration of European forest
ecosystems. Sci Total Environ 605-606: 1097-1116.
http ://dx.doi .org/10.1016/i. scitotenv.2017.06.132
Pe Vries. W; Reinds. GJ; Vel. E. (2003). Intensive monitoring of forest ecosystems in Europe: 2:
atmospheric deposition and its impacts on soil solution chemistry. For Ecol Manage 174: 97-115.
Pean. TJ; Johnson. JP. (1992). Growth response of young slash pine trees to simulated acid rain and
ozone stress. Can J For Res 22: 839-848.
8-218
-------
Decock. C; Chung. H; Venterea. R; Gray. SB; Leakey. ADB; Six. J. (2012). Elevated C02 and 0-3
modify N turnover rates, but not N20 emissions in a soybean agroecosystem. Soil Biol Biochem
51: 104-114. http://dx.doi.Org/10.1016/i.soilbio.2012.04.015
Decock. C; Six. J. (2012). Effects of elevated C02 and 0-3 on N-cycling and N20 emissions: a short-
term laboratory assessment. Plant Soil 351: 277-292. http://dx.doi.org/10.1007/sl 1104-011-0961-1
Diaz-De-Ouiiano. M: Kefauver. S; Qgava. R; Vollenweider. P: Ribas. A: Penuelas. J. (2016). Visible
ozone-like injury, defoliation, and mortality in two Pinus uncinata stands in the Catalan Pyrenees
(NE Spain). European Journal of Forest Research 135: 687-696. http://dx.doi.org/10.1007/slQ342-
016-0964-9
Dickson. RE: Lewin. KF: Isebrands. JG; Coleman. MP: Heilman. WE: Riemenschneider. DE: Sober.
J: Host. GE; Zak. PR: Hendrev. GR; Pregitzer. KS: Karnoskv. DF. (2000). Forest Atmosphere
Carbon Transfer and Storage (FACTS-II) the Aspen Free-Air C02 and 03 Enrichment (FACE)
project: An overview. (General Technical Report NC-214). St. Paul, MN: U.S. Dept. of
Agriculture, Forest Service, http://nrs.fs.fed.us/pubs/278
Dietze. MC: Moorcroft. PR. (2011). Tree mortality in the eastern and central United States: Patterns
and drivers. Global Change Biol 17: 3312-3326. http://dx.doi.Org/10.l 11 l/i.1365-
2486.2011.02477.x
Dohmen. GP. (1988). Indirect effects of air pollutants: changes in plant/parasite interactions. Environ
Pollut 53: 197-207.
Dohrmann. AB: Tebbe. CC. (2005). Effect of elevated tropospheric ozone on the structure of bacterial
communities inhabiting the rhizosphere of herbaceous plants native to Germany. Appl Environ
Microbiol 71: 7750-7758. http://dx.doi.org/10.1128/AEM.71.12.7750-7758.2005
Dumont. J: Spicher. F: Montpied. P: Dizengremel. P: Jolivet. Y: Le Thiec. D. (2013). Effects of ozone
on stomatal responses to environmental parameters (blue light, red light, C02 and vapour pressure
deficit) in three Populus deltoides Populus nigra genotypes. Environ Pollut 173: 85-96.
http://dx.doi.Org/10.1016/i.envpol.2012.09.026
Dunbar. J: Gallegos-Graves. L: Steven. B: Mueller. R: Hesse. C: Zak. PR: Kuske. CR. (2014). Surface
soil fungal and bacterial communities in aspen stands are resilient to eleven years of elevated C02
and 0-3. Soil Biol Biochem 76: 227-234. http://dx.doi.org/10.1016/i .soilbio.2014.05.027
Ebanvenle. E: Burton. AJ: Storer. AJ: Richter. PL: Glaeser. JA. (2016). Elevated tropospheric C02
and 0-3 may not alter initial wood decomposition rate or wood-decaying fungal community
composition of Northern Hardwoods. Int Biodeterior Biodegradation 111: 74-77.
http://dx.doi.Org/10.1016/i.ibiod.2016.04.026
Edwards. GS: Wullschleger. SP: Kelly. JM. (1994). Growth and physiology of northern red oak:
preliminary comparisons of mature tree and seedling responses to ozone. Environ Pollut 83: 215-
221. http://dx.doi.org/10.1016/0269-7491(94)90036-1
Edwards. IP: Zak. PR. (2011). Fungal community composition and function after long-term exposure
of northern forests to elevated atmospheric C02 and tropospheric 03. Global Change Biol 17:
2184-2195. http://dx.doi.org/10.1111/i. 1365-2486.2010.02376.X
Edwards. NT: Edwards. GL: Kelly. JM: Taylor. GE. Jr. (1992). Three-year growth responses of Pinus
taeda L to simulated rain chemistry, soil magnesium status, and ozone. Water Air Soil Pollut 63:
105-118. http://dx.doi.org/10.1007/BF0Q475625
Emberson. L: Ashmore. MR: Cambridge. HM: Simpson. P: Tuovinen. JP. (2000a). Modelling
stomatal ozone flux across Europe. Environ Pollut 109: 403-413. http://dx.doi.org/10.1016/SQ269-
7491(00)00043-9
8-219
-------
Emberson. LP; Pleiiel. H; Ainsworth. EA; van Den Berg. M; Ren. W; Osborne. S; Mills. G; Pandev.
D; Dentener. F; Biiker. P; Ewert. F; Koebl. R; Van Dingcncnd. R. (2018). Ozone effects on crops
and consideration in crop models. Eur J Agron 100: 19-34.
http://dx.doi.Org/10.1016/i.eia.2018.06.002
Emberson. LP; Wieser. G; Ashmore. MR. (2000b). Modelling of stomatal conductance and ozone flux
of Norway spruce: Comparison with field data. Environ Pollut 109: 393-402.
http://dx.doi.org/10.1016/S0269-7491(00)00042-7
Evans. LS; Fitzgerald. GA. (1993). Histological effects of ozone on slash pine (Pinus elliotti var.
Pensa). Environ Exp Bot 33: 505-513. http://dx.doi.org/10.1016/0098-8472(93)90024-A
Fares. S; Vargas. R; Petto. M: Goldstein. AH: Karlik. J: Paoletti. E: Vitale. M. (2013). Tropospheric
ozone reduces carbon assimilation in trees: estimates from analysis of continuous flux
measurements. Global Change Biol 19: 2427-2443. http://dx.doi.org/10. Ill 1/gcb. 12222
Farre-Armengol. G: Penuelas. J: Li. T: Yli-Pirila. P: Filella. I: Llusia. J: Blande. JP. (2015). Ozone
degrades floral scent and reduces pollinator attraction to flowers. New Phytol 209: 152-160.
http://dx.doi.Org/10.l 111/nph. 13620
Felzer. B: Kicklighter. P: Melillo. J: Wang. C: Xhuang. O: Prinn. R (2004). Effects of ozone on net
primary production and carbon sequestration in the conterminous United States using a
biogeochemistry model. Tellus B Chem Phys Meteorol 56: 230-248.
http://dx.doi.org/10.1111/i. 1600-0889.2004.00097.X
Felzer. B: Reillv. J: Melillo. J: Kicklighter. P: Sarofim. M: Wang. C: Prinn. R; Zhuang. O. (2005).
Future effects of ozone on carbon sequestration and climate change policy using a global
biogeochemical model. Clim Change 73: 345-373. http://dx.doi.org/10.1007/slQ584-005-6776-4
Felzer. BS: Cronin. TW: Melillo. JM: Kicklighter. PW: Schlosser. CA. (2009). Importance of carbon-
nitrogen interactions and ozone on ecosystem hydrology during the 21st century. J Geophys Res
114: GO 1020. http://dx.doi.org/10.1029/2008JG00Q826
Feng. Y: Lin. X: Yu. Y: Zhang. H: Chu. H: Zhu. J. (2013). Elevated ground-level 03 negatively
influences paddy methanogenic archaeal community. Sci Rep 3: 3193.
http://dx.doi.org/10.1038/srep03193
Feng. Y; Lin. X: Yu. Y; Zhu. J. (2011). Elevated ground-level 03 changes the diversity of anoxygenic
purple phototrophic bacteria in paddy field. Microb Ecol 62: 789-799.
http://dx.doi.org/10.1007/sQ0248-011-9895-7
Feng. Y; Yu. Y; Tang. H; Zu. O: Zhu. J: Lin. X. (2015). The contrasting responses of soil
microorganisms in two rice cultivars to elevated ground-level ozone. Environ Pollut 197: 195-202.
http://dx.doi.Org/10.1016/i.envpol.2014.l 1.032
Feng. Z; Biiker. P; Pleiiel. H; Emberson. L; Karlsson. PE; Uddling. J. (2017). A unifying explanation
for variation in ozone sensitivity among woody plants. Global Change Biol 24: 78-84.
http://dx.doi.Org/10.l 111/gcb. 13 824
Feng. ZZ; Kobavashi. K. (2009). Assessing the impacts of current and future concentrations of surface
ozone on crop yield with meta-analysis. Atmos Environ 43: 1510-1519.
http://dx.doi.Org/10.1016/i.atmosenv.2008.l 1.033
Feng. ZZ: Kobavashi. K; Ainsworth. EA. (2008). Impact of elevated ozone concentration on growth,
physiology, and yield of wheat (Triticum aestivum L.): A meta-analysis. Global Change Biol 14:
2696-2708. http://dx.doi.org/10.1111/i. 1365-2486.2008.01673,x
8-220
-------
Fenn. ME; de Bauer. LI; Hernandez-Teieda. T. (2002). Summary of air pollution impacts on forests in
the Mexico City air basin. In Urban air pollution and forests. New York, NY: Springer-Verlag.
Ferreira. B; Ribeiro. H; Pereira. MS; Cruz. A; Abreu. I. (2016). Effects of ozone in Plantago
lanceolata and Salix atrocinerea pollen. Aerobiologia 32: 421-430.
http://dx.doi.org/10.1007/slQ453-015-9415-l
Findlev. DA; Keever. GJ; Chappelka. AH; Eakes. DJ; Gilliam. CH. (1996). Sensitivity of red maple
cultivars to acute and chronic exposures of ozone. J arboric 22: 264-269.
Finnveden. G. (2000). On the limitations of life cycle assessment and environmental systems analysis
tools in general. Int J Life Cycle Assess 5: 229-238.
Fiscus. EL; Booker. FL; Sadok. W; Burkev. KO. (2012). Influence of atmospheric vapour pressure
deficit on ozone responses of snap bean (Phaseolus vulgaris L.) genotypes. J Exp Bot 63: 2557-
2564. http://dx.doi.org/10.1093/ixb/err443
Fishman. J; Belina. KM; Encarnacion. CH. (2014). The St. Louis ozone gardens: Visualizing the
impact of a changing atmosphere. Bull Am Meteorol Soc 95: 1171-1177.
http://dx.doi.Org/10.1175/BAMS-D-13-00009.l
Fleischer. K; Rebel. KT; van der Molen. MK; Erisman. JW; Wassen. MJ; van Loon. EE; Montagnani.
L; Gough. CM; Herbst. M; Janssens. IA; Gianelle. D; Dolman. AJ. (2013). The contribution of
nitrogen deposition to the photosynthetic capacity of forests. Global Biogeochem Cycles 27: 187-
199. http://dx.doi.org/10.1002/gbc.20Q26
Flowers. MP; Fiscus. EL; Burkev. KO; Booker. FL; Dubois. JJB. (2007). Photosynthesis, chlorophyll
fluorescence, and yield of snap bean (Phaseolus vulgaris L.) genotypes differing in sensitivity to
ozone. Environ Exp Bot 61: 190-198. http://dx.doi.Org/10.1016/i.envexpbot.2007.05.009
Foot. JP; Caporn. SJM; Lee. JA; Ashenden. TW. (1996). The effect of long-term ozone fumigation on
the growth, physiology and frost sensitivity of Calluna vulgaris. New Phytol 133: 503-511.
http://dx.doi.Org/10.llll/i.1469-8137.1996.tb01918.x
Fowler. D; Flechard. C; Skiba. U; Covle. M; Cape. JN. (1998). The atmospheric budget of oxidized
nitrogen and its role in ozone formation and deposition. New Phytol 139: 11-23.
http://dx.doi.Org/10.1046/i.1469-8137.1998.00167.x
Franzaring. J; Tonneiick. AEG; Kooijman. AWN; Dueck. TA. (2000). Growth responses to ozone in
plant species from wetlands. Environ Exp Bot 44: 39-48. http://dx.doi.org/10.1016/S0Q98-
8472(00)00052-6
Fuentes. JD; Chamecki. M; Roulston. T; Chen. B; Pratt. KR. (2016). Air pollutants degrade floral
scents and increase insect foraging times. Atmos Environ 141: 361-374.
http://dx.doi.Org/10.1016/i.atmosenv.2016.07.002
Fuentes. JD; Roulston. TH; Zenker. J. (2013). Ozone impedes the ability of a herbivore to find its host.
Environ Res Lett 8. http://dx.doi.Org/10.1088/1748-9326/8/l/014048
Fuhrer. J. (1994). Effects of ozone on managed pasture: 1. Effects of open-top chambers on
microclimate, ozone flux, and plant growth. Environ Pollut 86: 297-305.
http://dx.doi.org/10.1016/0269-7491(94)90170-8
Gao. F; Catalavud. V; Paoletti. E; Hoshika. Y; Feng. Z. (2017). Water stress mitigates the negative
effects of ozone on photosynthesis and biomass in poplar plants. Environ Pollut 230: 268-279.
http://dx.doi.Org/10.1016/i.envpol.2017.06.044
Gate. IM; Mcneill. S; Ashmore. MR. (1995). Effects of air pollution on the searching behaviour of an
insect parasitoid. Water Air Soil Pollut 85: 1425-1430. http://dx.doi.org/10.1007/BF0Q477181
8-221
-------
Gaucher. C; Dizcngrcmcl. P; Mauffette. Y; Chevrier. N. (2005). Response of Acer saccharum
seedlings to elevated 0-3 and C02 concentrations. Phytoprotection 86: 7-17.
Gerosa. G; Ballarin-Denti. A. (2003). Regional scale risk assessment of ozone and forests. In DF
Karnosky; KE Percy; AH Chappelka; C Simpson; J Pikkarainen (Eds.), Air pollution, global
change and forests in the new millennium (pp. 119-139). Amsterdam, Netherlands: Elsevier.
Gevh. AS; Xue. J; Ozkavnak. H: Spengler. JD. (2000). The Harvard Southern California chronic
ozone exposure study: Assessing ozone exposure of grade-school-age children in two southern
California communities. Environ Health Perspect 108: 265-270.
http://dx.doi.org/10.1289/ehp.00108265
Ghimire. RP; Kivimaenpaa. M: Kasurinen. A: Haikio. E: Holopainen. T: Holopainen. JK. (2017).
Herbivore-induced BVOC emissions of Scots pine under warming, elevated ozone and increased
nitrogen availability in an open-field exposure. Agr Forest Meteorol 242: 21-32.
http://dx.doi.Org/10.1016/i.agrformet.2017.04.008
Gillespie. C; Stabler. D; Tallentire. E; Goumenaki. E; Barnes. J. (2015). Exposure to environmentally-
relevant levels of ozone negatively influence pollen and fruit development. Environ Pollut 206:
494-501. http://dx.doi.Org/10.1016/i.envpol.2015.08.003
Gilliland. NJ; Chappelka. AH: Muntifering. RB: Booker. FL: Ditchkoff. SS. (2012). Digestive
utilization of ozone-exposed forage by rabbits (Oryctolagus cuniculus). Environ Pollut 163: 281-
286. http://dx.doi.Org/10.1016/i.envpol.2012.01.003
Gilliland. NJ: Chappelka. AH: Muntifering. RB: Ditchkoff. SS. (2016). Changes in southern Piedmont
grassland community structure and nutritive quality with future climate scenarios of elevated
tropospheric ozone and altered rainfall patterns. Plant Biol (Stuttg) 18: 47-55.
http://dx.doi.org/10.1111/plb. 12324
Gimeno. BS: Bermeio. V: Sanz. J: De La Torre. D: Elvira. S. (2004). Growth response to ozone of
annual species from Mediterranean pastures. Environ Pollut 132: 297-306.
http://dx.doi.Org/10.1016/i.envpol.2004.04.022
Giron-Calva. PS: Li. T: Mande. JD. (2016). Plant-plant interactions affect the susceptibility of plants
to oviposition by pests but are disrupted by ozone pollution. Agric Ecosyst Environ 233: 352-360.
http://dx.doi.Org/10.1016/i.agee.2016.09.028
Godzik. B: Grodzinska. K. (2002). Ground level ozone concentration and its effect on plants of Polish
national parks. Kosmos 5: 427.
Grantz. DA: Gunn. S: Vu. HB. (2006). 03 impacts on plant development: A meta-analysis of
root/shoot allocation and growth. Plant Cell Environ 29: 1193-1209.
http://dx.doi.org/10.1111/i. 1365-3040.2006.01521.x
Grantz. DA: Jackson. A: Vu. HB: Burkev. KO; Mcgrath. MT; Harvey. G. (2014). High ozone
increases soil perchlorate but does not affect foliar perchlorate content. J Environ Qual 43: 1460-
1466. http://dx.doi.org/10.2134/iea2013.ll.Q464
Grantz. DA: Paudel. R; Vu. HB: Shrestha. A: Grulke. N. (2016). Diel trends in stomatal response to
ozone and water deficit: a unique relationship of midday values to growth and allometry in Pima
cotton? Plant Biol (Stuttg) 18: 37-46. http://dx.doi.org/10.1111/plb. 12355
Grantz. DA; Shrestha. A; Vu. HB. (2008). Early vigor and ozone response in horseweed (Conyza
canadensis) biotypes differing in glyphosate resistance. Weed Sci 56: 224-230.
http://dx.doi.Org/10.1614/ws-07-130.l
8-222
-------
Grantz. DA; Vu. HB. (2009). 03 sensitivity in a potential C4 bioenergy crop: Sugarcane in California.
Crop Sci 49: 643-650. http://dx.doi.org/10.2135/cropsci2008.04.021Q
Grantz. DA; Vu. HB; Heath. RL; Burkev. KO. (2013). Demonstration of a diel trend in sensitivity of
Gossypium to ozone: a step toward relating 0-3 injury to exposure or flux. J Exp Bot 64: 1703-
1713. http://dx.doi.org/10.1093/ixb/ert032
Grantz. DA; Vu. HB; Tew. TL; Veremis. JC. (2012). Sensitivity of gas exchange parameters to ozone
in diverse C-4 sugarcane hybrids. Crop Sci 52: 1270-1280.
http://dx.doi.org/10.2135/cropsci2011.08.0413
Grantz. DA; Zhang. XJ; Massman. W; Delanv. A; Pederson. R. (1997). Ozone deposition to a cotton
(Gossypium hirsutum L) field: Stomatal and surface wetness effects during the California Ozone
Deposition experiment. Agr Forest Meteorol 85: 19-31. http://dx.doi.org/10.1016/SQ168-
1923(96)02396-9
Grantz. DA; Zhang. XJ; Massman. WJ; Den Hartog. G; Neumann. HH; Pederson. J. R. (1995). Effects
of stomatal conductance and surface wetness on ozone deposition in field-grown grape. Atmos
Environ 29: 3189-3198. http://dx.doi.org/10.1016/1352-2310(95)00129-M
Greaver. TL; Sullivan. TJ; Herrick. JD; Barber. MC; Baron. JS; Cosby. BJ; Deerhake. ME; Dennis.
RL; Dubois. J. -JB; Goodale. CL; Herlihv. AT; Lawrence. GB; Liu. L; Lynch. JA; Novak. KJ.
(2012). Ecological effects of nitrogen and sulfur air pollution in the US: What do we know? Front
Ecol Environ 10: 365-372. http://dx.doi.Org/10.1890/l 10049
Grebenc. T; Kraigher. H. (2007). Changes in the community of ectomycorrhizal fungi and increased
fine root number under adult beech trees chronically fumigated with double ambient ozone
concentration. Plant Biol (Stuttg) 9: 279-287. http://dx.doi.org/10.1055/s-2006-924489
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
Gregg. JW; Jones. CG; Dawson. TE. (2006). Physiological and developmental effects of 03 on
cottonwood growth in urban and rural sites. Ecol Appl 16: 2368-2381.
http://dx.doi.org/10.1890/1051-0761(2006)016r2368:PADEQO12.0.CO;2
Grulke. NE. (1999). Physiological responses of ponderosa pine to gradients of environmental
stressors. In PR Miller; JR McBride (Eds.), Oxidant air pollution impacts in the montane forests of
southern California: A case study of the San Bernardino Mountains (pp. 126-163). New York, NY:
Springer. http://dx.doi.org/10.10Q7/978-l-4612-1436-6 7
Grulke. NE; Alonso. R; Nguyen. T; Cascio. C; Dobrowolski. W. (2004). Stomata open at night in
pole-sized and mature ponderosa pine: Implications for 03 exposure metrics. Tree Physiol 24:
1001-1010. http://dx.doi.org/10.1093/treephvs/24.9.1001
Grulke. NE; Andersen. CP; Fenn. ME; Miller. PR. (1998). Ozone exposure and nitrogen deposition
lowers root biomass of ponderosa pine in the San Bernardino Mountains, California. Environ Pollut
103: 63-73. http://dx.doi.org/10.1016/S0269-7491(98)00130-4
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: 03 uptake and injury response. Trees
Struct Funct 17: 292-298. http://dx.doi.org/10.1007/s00468-002-Q237-8
Gumpertz. ML; Rawlings. JO. (1992). Nonlinear regression with variance components: Modeling
effects of ozone on crop yield. Crop Sci 32: 219-224.
http://dx.doi.org/10.2135/cropscil992.0011183X003200010Q45x
8-223
-------
Gundel. PE; Sorzoli. N; Ueno. AC; Ghersa. CM; Seal. CE; Bastias. DA; Martinez-Ghersa. MA.
(2015). Impact of ozone on the viability and antioxidant content of grass seeds is affected by a
vertically transmitted symbiotic fungus. Environ Exp Bot 113: 40-46.
http://dx.doi.Org/10.1016/i.envexpbot.2015.01.001
Gunthardt-Goerg. MS; Maurer. S; Bolligcr. J; Clark. AJ; Landolt. W; Bucher. JB. (1999). Responses
of young trees (five species in a chamber exposure) to near-ambient ozone concentrations. Water
Air Soil Pollut 116: 323-332.
Gustafson. EJ; Kubiske. ME; Sturtevant. BR; Miranda. BR. (2013). Scaling Aspen-FACE
experimental results to century and landscape scales. Landsc Ecol 28: 1785-1800.
http://dx.doi.org/10.1007/slQ980-013-9921-x
Habeck. CW; Lindroth. RL. (2013). Influence of global atmospheric change on the feeding behavior
and growth performance of a mammalian herbivore, Microtus ochrogaster. PLoS ONE 8: e72717.
http://dx.doi.org/10.1371/iournal.pone.0Q72717
Haesler. F; Hagn. A; Engel. M; Schloter. M. (2014). Impact of elevated atmospheric 0-3 on the
actinobacterial community structure and function in the rhizosphere of European beech (Fagus
sylvatica L.). FMICB 5: 36. http://dx.doi.org/10.3389/fmicb.2014.00Q36
Handlev. T; Grulke. NE. (2008). Interactive effects of 03 exposure on California black oak (Quercus
kelloggii Newb.) seedlings with and without N amendment. Environ Pollut 156: 53-60.
http://dx.doi.Org/10.1016/i.envpol.2008.01.002
Hanson. PJ; Wullschleger. SD; Norbv. RJ; Tschaplinski. TJ; Gunderson. CA. (2005). Importance of
changing C02, temperature, precipitation, and ozone on carbon and water cycles of an upland-oak
forest: incorporating experimental results into model simulations. Global Change Biol 11: 1402-
1423. http://dx.doi.org/10.1111/i. 1365-2486.2005.00991 ,x
Haves. F; Jones. MLM; Mills. G; Ashmore. M. (2007). Meta-analysis of the relative sensitivity of
semi-natural vegetation species to ozone. Environ Pollut 146: 754-762.
http://dx.doi.Org/10.1016/i.envpol.2006.06.011
Haves. F; Mills. G; Ashmore. M. (2009). Effects of ozone on inter- and intra-species competition and
photosynthesis in mesocosms of Lolium perenne and Trifolium repens. Environ Pollut 157: 208-
214. http://dx.doi.Org/10.1016/i.envpol.2008.07.002
Haves. F; Mills. G; Harmens. H; Wvness. K. (2011). Within season and carry-over effects following
exposure of grassland species mixtures to increasing background ozone. Environ Pollut 159: 2420-
2426. http://dx.doi.Org/10.1016/i.envpol.2011.06.034
Haves. F; Mills. G; Jones. L; Abbott. J; Ashmore. M; Barnes. J; Cape. JN; Covle. M; Peacock. S;
Rintoul. N; Toet. S; Wedlich. K; Wvness. K. (2016). Consistent ozone-induced decreases in
pasture forage quality across several grassland types and consequences for UK lamb production.
Sci Total Environ 543: 336-346. http://dx.doi.org/10.1016/i.scitotenv.2015.10.128
Haves. F; Mills. G; Jones. L; Ashmore. M. (2010). Does a simulated upland grassland community
respond to increasing background, peak or accumulated exposure of ozone? Atmos Environ 44:
4155-4164. http://dx.doi.Org/10.1016/i.atmosenv.2010.07.037
Haves. F; Mills. G; Williams. P; Harmens. H; Buker. P. (2006). Impacts of summer ozone exposure on
the growth and overwintering of UK upland vegetation. Atmos Environ 40: 4088-4097.
http://dx.doi.Org/10.1016/i.atmosenv.2006.03.012
Haves. F; Wagg. S; Mills. G; Wilkinson. S; Davies. W. (2012a). Ozone effects in a drier climate:
implications for stomatal fluxes of reduced stomatal sensitivity to soil drying in a typical grassland
species. Global Change Biol 18: 948-959. http://dx.doi.Org/10.l 111/i. 1365-2486.2011.02613.x
8-224
-------
Haves. F; Williamson. J; Mills. G. (2012b). Ozone pollution affects flower numbers and timing in a
simulated BAP priority calcareous grassland community. Environ Pollut 163: 40-47.
http://dx.doi.Org/10.1016/i.envpol.2011.12.032
He. Z; Xiong. J; Kent. AD; Deng. Y; Xue. K; Wang. G; Wu. L; Van Nostrand. J; Zhou. J. (2014).
Distinct responses of soil microbial communities to elevated C02 and 03 in a soybean agro-
ecosystem. ISME J 8: 714-726. http://dx.doi.org/10.1038/ismei.2Q13.177
Heagle. AS; Body. DE: Heck. WW. (1973). An open-top field chamber to assess the impact of air
pollution on plants. J Environ Qual 2: 365-368.
http://dx.doi.org/10.2134/ieal973.00472425000200030Q14x
Heagle. AS; Brandenburg. RL; Burns. JC; Miller. JE. (1994). Ozone and carbon dioxide effects on
spider mites in white clover and peanut. J Environ Qual 23: 1168-1176.
http://dx.doi.org/10.2134/ieal994.004724250023000600Q6x
Heck. WW; Cure. WW; Rawlings. JO; Zaragoza. LJ; Heagle. AS; Heggestad. HE; Kohut. RJ; Kress.
LW; Temple. PJ. (1984). Assessing impacts of ozone on agricultural crops: II. Crop yield functions
and alternative exposure statistics. J Air Pollut Control Assoc 34: 810-817.
Heck. WW; Heagle. AS; Miller. JE; Rawlings. JO. (1991). A national program (NCLAN) to assess the
impact of ozone on agricultural resources. In RL Berglund; DR Lawson; DJ McKee (Eds.),
Tropospheric ozone and the environment: papers from an international conference; March 1990;
Los Angeles, CA (pp. 225-254). Pittsburgh, PA: Air & Waste Management Association.
Heck. WW; Philbeck. RB; Dunning. JA. (1978). A continuous stirred tank reactor (CSTR) system for
exposing plants to gaseous air contaminants: Principles, specifications, construction, and operation.
(ARS-S-181). New Orleans, LA: U.S. Department of Agriculture, Agricultural Research Service.
Heck. WW; Taylor. OC; Adams. R; Bingham. G; Miller. J; Preston. E; Weinstein. L. (1982).
Assessment of crop loss from ozone. J Air Pollut Control Assoc 32: 353-361.
http://dx.doi.org/10.1080/00022470.1982.104654Q8
Hendrev. GR; Ellsworth. DS; Lewin. KF; Nagy. J. (1999). A free-air enrichment system for exposing
tall forest vegetation to elevated atmospheric C02. Global Change Biol 5: 293-309.
http://dx.doi.Org/10.1046/i.1365-2486.1999.00228.x
Hendrev. GR; Kimball. BA. (1994). The FACE program. Agr Forest Meteorol 70: 3-14.
http://dx.doi.org/10.1016/0168-1923(94')90044-2
Hildebrand. E; Skellv. JM; Fredericksen. TS. (1996). Foliar response of ozone-sensitive hardwood tree
species from 1991 to 1993 in the Shenandoah National Park, Virginia. Can J For Res 26: 658-669.
http://dx.doi.Org/10.l 139/x26-076
Hillstrom. ML; Couture. JJ; Lindroth. RL. (2014). Elevated carbon dioxide and ozone have weak,
idiosyncratic effects on herbivorous forest insect abundance, species richness, and community
composition. Insect Conservation and Diversity 7: 553-562. http://dx.doi.org/10. Ill 1/icad. 12078
Hillstrom. ML; Lindroth. RL. (2008). Elevated atmospheric carbon dioxide and ozone alter forest
insect abundance and community composition. Insect Conservation and Diversity 1: 233-241.
http://dx.doi.Org/10.l 111/i. 1752-4598.2008.00031.x
Hofmockel. KS; Zak. DR; Moran. KK; Jastrow. JD. (2011). Changes in forest soil organic matter
pools after a decade of elevated C02 and 03. Soil Biol Biochem 43: 1518-1527.
http://dx.doi.Org/10.1016/i.soilbio.2011.03.030
8-225
-------
Hogg. A; Uddling. J; Ellsworth. D; Carroll. MA; Presslev. S; Lamb. B; Vogel. C. (2007). Stomatal
and non-stomatal fluxes of ozone to a northern mixed hardwood forest. Tellus B Chem Phys
Meteorol 59: 514-525. http://dx.doi.org/10.1111/i. 1600-0889.2007.00269.X
Hogsett. WE; Olszvk. D; Ormrod. DP; Taylor. GE. Jr; Tingev. DT. (1987a). Air pollution exposure
systems and experimental protocols: Volume 2: Description of facilities. (EPA/600/3-87/037b).
Corvallis, OR: U.S. Environmental Protection Agency, Environmental Research Laboratory.
http://nepis.epa.gov/Exe/ZvPURL.cgi?Dockev=30000KQH.txt
Hogsett. WE; Olszvk. D; Ormrod. DP; Taylor. GE. Jr; Tingev. DT. (1987b). Air pollution exposure
systems and experimental protocols: Volume I: A review and evaluation of performance.
(EPA/600/3-87/037a). Corvallis, OR: U.S. Environmental Protection Agency.
Hogsett. WE; Tingev. DT; Holman. SR. (1985). A programmable exposure control system for
determination of the effects of pollutant exposure regimes on plant growth. Atmos Environ 19:
1135-1145. http://dx.doi.org/10.1016/0004-6981(85)90198-2
Hogsett. WE; Weber. JE; Tingev. 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/s00267990001Q
Holmes. CD. (2014). Air pollution and forest water use [Comment]. Nature 507: E1-E2.
http ://dx.doi .org/10.103 8/nature 13113
Holmes. WE; Zak. PR; Pregitzer. KS; King. JS. (2006). Elevated C02 and 03 alter soil nitrogen
transformations beneath trembling aspen, paper birch, and sugar maple. Ecosystems 9: 1354-1363.
http://dx.doi.org/10.1007/sl0021-006-Q163-5
Hong. B; Weinstein. D; Swanev. D. (2006). Assessment of ozone effects on nitrate export from
Hubbard Brook Watershed 6. Environ Pollut 141: 8-21.
http://dx.doi.Org/10.1016/i.envpol.2005.08.030
Hong. Y; Yi. T; Tan. X; Zhao. Z; Ge. F. (2016). High ozone (03) affects the fitness associated with
the microbial composition and abundance of Q biotype Bemisia tabaci. FMICB 7: 1593.
http://dx.doi.org/10.3389/fmicb.2016.01593
Horn. KJ; Thomas. RQ; Clark. CM; Pardo. LH; Fenn. ME; Lawrence. GB; Perakis. SS; Smithwick.
EAH; Baldwin. D; Braun. S; Nordin. A; Perry. CH; Phelan. JN; Schaberg. PG; St Clair. SB;
Warbv. R; Watmough. S. (2018). Growth and survival relationships of 71 tree species with
nitrogen and sulfur deposition across the conterminous. PLoS ONE 13: Article #e0205296.
http://dx.doi.org/10.1371/iournal.pone.0205296
Hoshika. Y; Carriero. G; Feng. Z; Zhang. Y; Paoletti. E. (2014). Determinants of stomatal
sluggishness in ozone-exposed deciduous tree species. Sci Total Environ 481: 453-458.
http://dx.doi.Org/10.1016/i.scitotenv.2014.02.080
Hoshika. Y; De Marco. A; Materassi. A; Paoletti. E. (2016). Light intensity affects ozone-induced
stomatal sluggishness in snapbean. Water Air Soil Pollut 227. http://dx.doi.org/10.1007/sll27Q-
016-3127-1
Hoshika. Y; Katata. G; Deushi. M; Watanabe. M; Koike. T; Paoletti. E. (2015). Ozone-induced
stomatal sluggishness changes carbon and water balance of temperate deciduous forests. Sci Rep 5:
9871. http://dx.doi.org/10.1038/srepQ9871
Hoshika. Y; Omasa. K; Paoletti. E. (2011). Both ozone exposure and soil water stress are able to
induce stomatal sluggishness. Environ Exp Bot 88: 19-23.
http://dx.doi.Org/10.1016/i.envexpbot.2011.12.004
8-226
-------
Hoshika. Y; Watanabe. M; Carrari. E; Paoletti. E; Koike. T. (2018). Ozone-induced stomatal
sluggishness changes stomatal parameters of Jarvis-type model in white birch and deciduous oak.
Plant Biol (Stuttg) 20: 20-28. http://dx.doi.org/10. Ill 1/plb. 12632
Hoshika. Y; Watanabe. M; Inada. N; Koike. T. (2012a). Modeling of stomatal conductance for
estimating ozone uptake of Fagus crenata under experimentally enhanced free-air ozone exposure.
Water Air Soil Pollut 223: 3893-3901. http://dx.doi.org/10.1007/sll270-012-1158-9
Hoshika. Y: Watanabe. M: Inada. N; Koike. T. (2012b). Ozone-induced stomatal sluggishness
develops progressively in Siebold's beech (Fagus crenata). Environ Pollut 166: 152-156.
http://dx.doi.Org/10.1016/i.envpol.2012.03.013
Huang. YZ: Zhong. M. (2015). Influence of elevated ozone concentration on methanotrophic bacterial
communities in soil under field condition. Atmos Environ 108: 59-66.
http://dx.doi.Org/10.1016/i.atmosenv.2015.02.075
Hunova. I; Matouskova. L: Srnenskv. R: Kozelkova. K. (2011). Ozone influence on native vegetation
in the Jizerske hory Mts. of the Czech Republic: Results based on ozone exposure and ozone-
induced visible symptoms. Environ Monit Assess 183: 501-515. http://dx.doi.org/10.1007/slQ661-
011-1935-8
Iriti. M: Faoro. F. (2009). Chemical diversity and defence metabolism: How plants cope with
pathogens and ozone pollution [Review]. International Journal of Molecular Sciences 10: 3371-
3399. http://dx.doi.org/10.3390/iimsl0083371
Jackson. DM: Heagle. AS; Eckel. RVW. (1999). Ovipositional response of tobacco hornworm moths
(Lepidoptera: Sphyngidae) to tobacco plants grown under elevated levels of ozone. Environ
Entomol 28: 566-571. http://dx.doi.Org/10.1093/ee/28.4.566
Jackson. DM: Ruftv. TW: Heagle. AS: Severson. RF: Eckel. RWV. (2000). Survival and development
of tobacco hornworm larvae on tobacco plants grown under elevated levels of ozone. J Chem Ecol
26: 1-18.
Jacob. DJ: Winner. DA (2009). Effect of climate change on air quality. Atmos Environ 43: 51-63.
http://dx.doi.Org/10.1016/i.atmosenv.2008.09.051
Jamieson. MA: Burkle. LA: Manson. JS: Runvon. JB; Trowbridge. AM: Zientek. J. (2017). Global
change effects on plant-insect interactions: the role of phytochemistry [Review]. Curr Opin Insect
Sci 23: 70-80. http://dx.doi.Org/10.1016/i.cois.2017.07.009
Jones. ME: Paine. TP. (2006). Detecting changes in insect herbivore communities along a pollution
gradient. Environ Pollut 143: 377-387. http://dx.doi.Org/10.1016/i.envpol.2005.12.013
Jones. MLM: Hodges. G: Mills. G. (2010). Nitrogen mediates above-ground effects of ozone but not
below-ground effects in a rhizomatous sedge. Environ Pollut 158: 559-565.
http://dx.doi.Org/10.1016/i.envpol.2009.08.002
Jones. TG: Freeman. C: Llovd. A: Mills. G. (2009). Impacts of elevated atmospheric ozone on
peatland below-ground doc characteristics. Ecol Eng 35: 971-977.
http://dx.doi.Org/10.1016/i.ecoleng.2008.08.009
Juergens. A: Bischoff. M. (2017). Changing odour landscapes: the effect of anthropogenic volatile
pollutants on plant-pollinator olfactory communication. Funct Ecol 31: 56-64.
http://dx.doi.org/10.1111/1365-2435.12774
Kanerva. T; Paloiarvi. A: Ramo. K; Manninen. S. (2008). Changes in soil microbial community
structure under elevated tropospheric 03 and C02. Soil Biol Biochem 40: 2502-2510.
http://dx.doi.Org/10.1016/i.soilbio.2008.06.007
8-227
-------
Kanerva. T; Paloiarvi. A; Ramo. K; Oianpera. K; Esala. M; Manninen. S. (2006). A 3-year exposure to
C02 and 03 induced minor changes in soil N cycling in a meadow ecosystem. Plant Soil 286: 61-
73. http://dx.doi.org/10.1007/slll04-006-9Q26-2
Kanerva. T; Rcgina. K; Ramo. K; Oianpera. K; Manninen. S. (2007). Fluxes of N20, CH4 and C02 in
a meadow ecosystem exposed to elevated ozone and carbon dioxide for three years. Environ Pollut
145: 818-828. http://dx.doi.Org/10.1016/i.envpol.2006.03.055
Karnoskv. DF: Gagnon. ZE: Dickson. RE: Coleman. MP: Lee. EH: Isebrands. JG. (1996). Changes in
growth, leaf abscission, biomass associated with seasonal tropospheric ozone exposures of Populus
tremuloides clones and seedlings. Can J For Res 26: 23-37.
Karnoskv. DF: Mankovska. B: Percy. K: Dickson. RE: Podila. GK: Sober. J: Noormets. A: Hendrev.
G: Coleman. MP: Kubiske. M; Pregitzer. KS: Isebrands. JG. (1999). Effects of tropospheric ozone
on trembling aspen and interaction with C02: Results from an 03-gradient and a FACE
experiment. Water Air Soil Pollut 116: 311-322. http://dx.doi.org/10.1023/A: 1005276824459
Kasurinen. A: Biasi. C: Holopainen. T; Rousi. M; Maenpaa. M; Oksanen. E. (2012). Interactive effects
of elevated ozone and temperature on carbon allocation of silver birch (Betula pendula) genotypes
in an open-airfield exposure. Tree Physiol 32: 737-751. http://dx.doi.org/10.1093/treephvs/tps005
Kasurinen. A: Keinanen. MM: Kaipainen. S: Nilsson. LO: Vapaavuori. E: Kontro. MH: Holopainen.
T. (2005). Below-ground responses of silver birch trees exposed to elevated C02 and 03 levels
during three growing seasons. Global Change Biol 11: 1167-1179.
http://dx.doi.org/10.1111/i. 1365-2486.2005.00970.X
Kasurinen. A: Riikonen. J: Oksanen. E; Vapaavuori. E; Holopainen. T. (2006). Chemical composition
and decomposition of silver birch leaf litter produced under elevated C02 and 03. Plant Soil 282:
261-280. http://dx.doi.org/10.1007/sl 1104-005-6026-6
Kasurinen. A: Silfver. T: Rousi. M: Mikola. J. (2017). Warming and ozone exposure effects on silver
birch (Betula pendula Roth) leaf litter quality, microbial growth and decomposition. Plant Soil 414:
127-142. http://dx.doi.org/10.1007/slllQ4-016-3122-8
Kats. G: Olszvk. DM: Thompson. CR. (1985). Open top experimental chambers for trees. J Air Waste
Manag Assoc 35: 1298-1301. http://dx.doi.org/10.1080/00022470.1985.10466Q33
Kats. G: Thompson. CR: Kubv. WC. (1976). Improved ventilation of open top greenhouses. J Air
Pollut Control Assoc 26: 1089-1090. http://dx.doi.org/10.1080/00022470.1976.1047Q367
Kefauver. SC: Penuelas. J: Ustin. S. (2013). Using topographic and remotely sensed variables to
assess ozone injury to conifers in the Sierra Nevada (USA) and Catalonia (Spain). Rem Sens
Environ 139: 138-148. http://dx.doi.Org/10.1016/i.rse.2013.07.037
Kefauver. SC: Penuelas. J: Ustin. SL. (2012a). Improving assessments of tropospheric ozone injury to
Mediterranean montane conifer forests in California (USA) and Catalonia (Spain) with GIS models
related to plant water relations. Atmos Environ 62: 41-49.
http://dx.doi.Org/10.1016/i.atmosenv.2012.08.013
Kefauver. SC: Penuelas. J: Ustin. SL. (2012b). Applications of hyperspectral remote sensing and GIS
for assessing forest health and air pollution. In 2012 IEEE International Geoscience and Remote
Sensing Symposium (IGARSS). Piscataway, NJ: IEEE.
http://dx.doi.org/10.1109/IGARSS.2012.6350696
Keller. T; Hasler. R. (1984). The influence of a fall fumigation with ozone on the stomatal behavior of
spruce and fir. Oecologia 64: 284-286. http://dx.doi.org/10.1007/BF0Q376884
8-228
-------
Kelting. PL; Burger. JA; Edwards. GS. (1995). The effects of ozone on the root dynamics of seedlings
and mature red oak (Quercus rubra L). For Ecol Manage 79: 197-206.
http://dx.doi.org/10.1016/Q378-l 127(95)03603-2
Khaling. E; Li. T; Holopainen. JK; Blande. JD. (2016). Elevated ozone modulates herbivore-induced
volatile emissions of Brassica nigra and alters atritrophic interaction. J Chem Ecol 42: 368-381.
http://dx.doi.org/10.1007/slQ886-016-0697-8
Khaling. E: Papazian. S; Poelman. EH: Holopainen. JK: Albrectsen. BR: Blande. JD. (2015). Ozone
affects growth and development of Pieris brassicae on the wild host plant Brassica nigra. Environ
Pollut 199: 119-129. http://dx.doi.Org/10.1016/i.envpol.2015.01.019
King. JS: Kubiske. ME: Pregitzer. KS: Hendrev. GR: Mcdonald. EP: Giardina. CP: Ouinn. VS:
Karnosky. DF. (2005). Tropospheric 03 compromises net primary production in young stands of
trembling aspen, paper birch and sugar maple in response to elevated atmospheric C02. New
Phytol 168: 623-635. http://dx.doi.org/10.1111/i. 1469-8137.2005.01557.X
King. JS: Pregitzer. KS: Zak. PR: Sober. J: Isebrands. JG: Dickson. RE: Hendrev. GR: Karnosky. DF.
(2001). Fine-root biomass and fluxes of soil carbon in young stands of paper birch and trembling
aspen as affected by elevated atmospheric C02 and tropospheric 03. Oecologia 128: 237-250.
Kline. LJ: Davis. DP: Skellv. JM: Decoteau. DR. (2009). Variation in ozone sensitivity within Indian
hemp and common milkweed selections from the Midwest. Northeast Nat 16: 307-313.
http://dx.doi.org/10.1656/045.016.021Q
Kline. LJ: Pavis. PP: Skellv. 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 r57:OSOPSE12.Q.CO:2
Koerner. C. (2006). Plant C02 responses: an issue of definition, time and resource supply. New Phytol
172: 393-411. http://dx.doi.Org/10.llll/i.1469-8137.2006.01886.x
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.
http://dx.doi.Org/10.1016/i.envpol.2007.04.022
Kohut. R: Flanagan. C: Cheatham. J: Porter. E. (2012). Foliar ozone injury on cutleaf coneflower at
Rocky Mountain National Park, Colorado. West N Am Nat 72: 32-42.
http://dx.doi.org/10.3398/064.072.01Q4
Kohut. RJ: Laurence. JA: Amundson. RG. (1988). Effects of ozone and sulfur dioxide on yield of red
clover and timothy. J Environ Qual 17: 580-585.
http://dx.doi.org/10.2134/ieal988.00472425001700040Q10x
Kostiainen. K: Saranpaa. P: Lundqvist. SO: Kubiske. ME: Vapaavuori. E. (2014). Wood properties of
Populus and Betula in long-term exposure to elevated C02 and 0-3. Plant Cell Environ 37: 1452-
1463. http://dx.doi.org/10.Ill 1/pce. 12261
Kozovits. AR: Matvssek. R: Blaschke. H: Gottlein. A: Grams. TEE. (2005). Competition increasingly
dominates the responsiveness of juvenile beech and spruce to elevated C02 and/or 03
concentrations throughout two subsequent growing seasons. Global Change Biol 11: 1387-1401.
http://dx.doi.org/10.1111/i. 1365-2486.2005.00993 ,x
Kress. LW: Skellv. JM. (1982). Response of several eastern forest tree species to chronic doses of
ozone and nitrogen dioxide. Plant Pis 66: 1149-1152. http://dx.doi.org/10.1094/PP-66-l 149
8-229
-------
Krupa. SV; Grunhagc. L; Jager. HJ; Nosal. M; Manning. WJ; Lcggc. AH; Hanewald. K. (1995).
Ambient ozone (03) and adverse crop response: A unified view of cause and effect. Environ Pollut
87: 119-126. http://dx.doi.org/10.1016/S0269-7491(99)80014-1
Kubiske. ME; Foss. AR. (2015). Carbon dioxide and ozone data from the Aspen FACE Experiment,
1998-2009, and Phase II Experiment, 2010 [Database]. Fort Collins, CO: Forest Service Research
Data Archive. Retrieved from https://www.fs.usda.gov/rds/archive/Product/RDS-2015-0001
Kubiske. ME; Ouinn. VS; Marquardt. PE; Karnoskv. DF. (2007). Effects of elevated atmospheric C02
and/or 03 on intra- and interspecific competitive ability of aspen. Plant Biol (Stuttg) 9: 342-355.
http://dx.doi.org/10.1055/s-2006-92476Q
Kvalevag. MM; Mvhre. G. (2013). The effect of carbon-nitrogen coupling on the reduced land carbon
sink caused by tropospheric ozone. Geophys Res Lett 40: 3227-3231.
http://dx.doi.org/10.1002/grl.50572
Landesmann. JB; Gundel. PE; Martinez-Ghersa. MA; Ghersa. CM. (2013). Ozone exposure of a weed
community produces adaptive changes in seed populations of Spergula arvensis. PLoS ONE 8:
e75820. http://dx.doi.org/10.1371 hournal.pone.0075820
Lapina. K; Henze. DK; Milford. JB; Travis. K. (2016). Impacts of Foreign, Domestic, and State-Level
Emissions on Ozone-Induced Vegetation Loss in the United States. Environ Sci Technol 50: 806-
813. http://dx.doi.org/10.1021/acs.est.5b04887
Laurence. JA; Amundson. RG; Kohut. RJ; Weinstein. DA. (1997). Growth and water use of red spruce
(Picea rubens Sarg) exposed to ozone and simulated acidic precipitation for four growing seasons.
Forest Sci 43: 355-361.
Laurence. JA; Kohut RJ; Amundson. RG; Weinstein. DA; MacLean. DC. (1996). Response of sugar
maple to multiple year exposure to ozone and simulated acidic precipitation. Environ Pollut 92:
119-126. http://dx.doi.org/10.1016/0269-7491(95)00105-0
Lee. EH; Hogsett. WE. (1996). Methodology for calculating inputs for ozone secondary standard
benefits anaylsis: Part II. Research Triangle Park, NC: U.S. Environmental Protection Agency.
Lee. EH; Hogsett. WE. (1999). Role of concentrations and time of day in developing ozone exposure
indices for a secondary standard. J Air Waste Manag Assoc 49: 669-681.
http://dx.doi.org/10.108Q/10473289.1999.10463835
Lee. EH; Hogsett. WE; Tingev. DT. (1994). Attainment and effects issues regarding alternative
secondary ozone air quality standards. J Environ Qual 23: 1129-1140.
http://dx.doi.org/10.2134/ieal994.004724250023000600Q2x
Lee. EH; Tingev. DT; Hogsett. WE. (1987). Selection of the best exposure-response model using
various 7-hour ozone exposure statistics. Research Triangle Park, NC: U.S. Environmental
Protection Agency.
Lee. EH; Tingev. DT; Hogsett. WE. (1988). Evaluation of ozone exposure indices in exposure-
response modeling. Environ Pollut 53: 43-62. http://dx.doi.org/10.1016/0269-7491(88)90024-3
Lee. EH; Tingev. DT; Hogsett. WE. (1989). Interrelation of experimental exposure and ambient air
quality data for comparison of ozone exposure indices and estimating agricultural losses.
(EPA/600/3-89/047). Corvallis, OR: U.S. Environmental Protection Agency.
Lee. EH; Tingev. DT; Waschmann. RS; Phillips. PL; Olszvk. DM; Johnson. MG; Hogsett. WE.
(2009). Seasonal and long-term effects of C02 and 03 on water loss in ponderosa pine and their
interaction with climate and soil moisture. Tree Physiol 29: 1381-1393.
http://dx.doi.org/10.1093/treephvs/tppQ71
8-230
-------
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/10.1016/00Q4-
6981(87)90038-2
Legge. AH; Grunhage. L; Nosal. M; Jager. HJ; Krupa. SV. (1995). Ambient ozone and adverse crop
response: An evaluation of North American and European data as they relate to exposure indices
and critical levels. J Appl Bot Food Qual 69: 192-205.
Leisner. CP; Ainsworth. EA. (2012). Quantifying the effects of ozone on plant reproductive growth
and development. Global Change Biol 18: 606-616. http://dx.doi.org/10.1111/i. 1365-
2486.2011.02535.x
Leisner. CP; Yendrek. CR; Ainsworth. EA. (2017). Physiological and transcriptomic responses in the
seed coat of field-grown soybean (Glycine max L. Merr.) to abiotic stress. BMC Plant Biol 17:
242. http://dx.doi.org/10.1186/sl2870-017-1188-v
Lesser. VM; Rawlings. JO; Spruill. SE; Somerville. MC. (1990). Ozone effects on agricultural crops:
Statistical methodologies and estimated dose-response relationships. Crop Sci 30: 148-155.
http://dx.doi.org/10.2135/cropscil990.0011183X003000010Q33x
Lewis. JS; Ditchkoff. SS; Lin. JC; Muntifering. RB; Chappelka. AH. (2006). Nutritive quality of big
bluestem (Andropogon gerardii) and eastern gamagrass (Tripsacum dactyloides) exposed to
tropospheric ozone. Rangeland Ecol Manag 59: 267-274. http://dx.doi.org/10.2111/05-025R2.1
Lewis. SL; Maslin. MA. (2015). Defining the anthropocene. Nature 519: 171-180.
http://dx.doi.org/10.1038/naturel4258
Li. CH; Wang. TZ; Li. Y; Zheng. YH; Jiang. GM. (2013a). Flixweed is more competitive than winter
wheat under ozone pollution: evidences from membrane lipid peroxidation, antioxidant enzymes
and biomass. PLoS ONE 8: e60109. http://dx.doi.org/10.1371/iournal.pone.0060109
Li. L; Manning. WJ; Tong. L; Wang. X. (2015). Chronic drought stress reduced but not protected
Shantung maple (Acer truncatum Bunge) from adverse effects of ozone (03) on growth and
physiology in the suburb of Beijing, China. Environ Pollut 201: 34-41.
http://dx.doi.Org/10.1016/i.envpol.2015.02.023
Li. Q; Yang. Y; Bao. X; Zhu. J; Liang. W; Bezemer. TM. (2016a). Cultivar specific plant-soil
feedback overrules soil legacy effects of elevated ozone in a rice-wheat rotation system. Agric
Ecosyst Environ 232: 85-92. http://dx.doi.Org/10.1016/i.agee.2016.07.025
Li. T; Blande. JD. (2015). Associational susceptibility in broccoli: mediated by plant volatiles,
impeded by ozone. Global Change Biol 21: 1993-2004. http://dx.doi.org/10. Ill 1/gcb. 12835
Li. T; Blande. JD; Holopainen. JK. (2016b). Atmospheric transformation of plant volatiles disrupts
host plant finding. Sci Rep 6: 33851. http://dx.doi.org/10.1038/srep33851
Li. X; Deng. Y; Li. Q; Lu. C; Wang. J; Zhang. H; Zhu. J; Zhou. J; He. Z. (2013b). Shifts of functional
gene representation in wheat rhizosphere microbial communities under elevated ozone. ISME J 7:
660-671. http://dx.doi.org/10.1038/ismei.2012.120
Lindroth. RL; Reich. PB; Tjoelker. MG; Volin. JC; Oleksyn. J. (1993). Light environment alters
response to ozone stress in seedlings of Acer saccharum March, and hybrid Populus L III
Consequences for performance of gypsy moth. New Phytol 124: 647-654.
http://dx.doi.org/10.1111/i. 1469-8137.1993.tb03854.x
8-231
-------
Liu. L; King. JS; Giardina. CP. (2007). Effects of elevated atmospheric C02 and tropospheric 03 on
nutrient dynamics: Decomposition of leaf litter in trembling aspen and paper birch communities.
Plant Soil 299: 65-82. http://dx.doi.org/10.1007/slllQ4-007-9361-v
Lloyd. KL; Davis. DP; Marini. RP: Decoteau. DR. (2018). Effects of nighttime ozone treatment at
ambient concentrations on sensitive and resistant snap bean genotypes. J Am Soc Hortic Sci 143:
23-33. http://dx.doi.org/10.21273/JASHS04253-17
Loats. KV; Rebbeck. J. (1999). Interactive effects of ozone and elevated carbon dioxide on the growth
and physiology of black cherry, green ash, and yellow poplar seedlings. Environ Pollut 106: 237-
248. http://dx.doi.org/10.1016/S0269-7491(99)00069-X
Lombardozzi- D; Levis. S: Bonan. G: Hess. PG: Sparks. JP. (2015). The Influence of Chronic Ozone
Exposure on Global Carbon and Water Cycles. J Clim 28: 292-305.
http://dx.doi.Org/10.1175/JCLI-D-14-00223.l
Lorenz. M; Becher. G: Mues. V: Fischer. R; Becker. R; Catalavud. V: Pise. N: Krause. GHM; Sanz.
M; Ulrich. E. (2005). Forest Condition in Europe, 2005 Technical Report of IOCP Forests.
Hamburg, Germany: Federal Research Centre for Forestry and Forest Products (BFH) and
Pepartment of Wood Science University of Hamburg.
Lucas. PW: Cottam. PA; Sheppard. LJ; Francis. BJ. (1988). Growth responses and delayed winter
hardening in Sitka spruce following summer exposure to ozone. New Phytol 108: 495-504.
http://dx.doi.Org/10.llll/i.1469-8137.1988.tb04192.x
Lucas. PW: Rantanen. L; Mehlhorn. H. (1993). Needle chlorosis in Sitka spruce following a three-year
exposure to low concentrations of ozone: changes in mineral content, pigmentation and ascorbic
acid. New Phytol 124: 265-275. http://dx.doi.Org/10.llll/i.1469-8137.1993.tb03816.x
Mahmud. A: Tvree. M: Cavan. P: Motallebi. N: Kleeman. MJ. (2008). Statistical downscaling of
climate change impacts on ozone concentrations in California. J Geophys Res 113: P21103.
http://dx.doi.org/10.1029/2007JP0Q9534
Mandl. RH: Laurence. JA: Kohut. RJ. (1989). Pevelopment and testing of open-top chambers for
exposing large, perennial plants to air pollutants. J Environ Qual 18: 534-540.
http://dx.doi.org/10.2134/ieal989.00472425001800040Q26x
Mandl. RH: Weinstein. LH; McCune. PC: Kevenv. M. (1973). A cylindrical, open-top chamber for
the exposure of plants to air pollutants in the field. J Environ Qual 2: 371-376.
Mankovska. B; Percy. KE; Karnoskv. PF. (2005). Impacts of greenhouse gases on epicuticular waxes
of Populus tremuloides Michx.: Results from an open-air exposure and a natural 03 gradient.
Environ Pollut 137: 580-586. http://dx.doi.Org/10.1016/i.envpol.2005.01.043
Manninen. AM: Holopainen. T: Lvvtikainen-Saarenmaa. P: Holopainen. JK. (2000). The role of low-
level ozone exposure and mycorrhizas in chemical quality and insect herbivore performance on
Scots pine seedlings. Global Change Biol 6: 111-121. http://dx.doi.org/10.1046/i. 1365-
2486.2000.00290.x
Manning. WJ: Godzik. B. (2004). Bioindicator plants for ambient ozone in Central and Eastern
Europe. Environ Pollut 130: 33-39. http://dx.doi.Org/10.1016/i.envpol.2003.10.033
Manning. WJ: Godzik. B: Musselman. R. (2002). Potential bioindicator plant species for ambient
ozone in forested mountain areas of central Europe. Environ Pollut 119: 283-290.
http://dx.doi.org/10.1016/S0269-7491(02)00102-1
8-232
-------
Manning. WJ; Krupa. SV. (1992). Experimental methodology for studying the effects of ozone on
crops and trees. In AS Lefohn (Ed.), Surface level ozone exposures and their effects on vegetation
(pp. 93-156). Chelsea, MI: Lewis Publishers.
Martinez-Ghersa. MA; Menendez. AI; Gundel. PE; Folcia. AM; Romero. AM; Landesmann. JB;
Ventura. L; Ghersa. CM. (2017). Legacy of historic ozone exposure on plant community and food
web structure. PLoS ONE 12: e0182796. http://dx.doi.org/10.1371/iournal.pone.0182796
Matvssek. R; Le Thiec. D; Low. M; Dizengremel. P; Nunn. AT: Haberle. KH. (2006). Interactions
between drought and 03 stress in forest trees [Review]. Plant Biol (Stuttg) 8: 11-17.
http://dx.doi.org/10.1055/s-2005-873Q25
Matvssek. R; Sandermann. H; Wieser. G; Booker. F; Cieslik. S; Musselman. R; Ernst. D. (2008). The
challenge of making ozone risk assessment for forest trees more mechanistic [Review]. Environ
Pollut 156: 567-582. http://dx.doi.Org/10.1016/i.envpol.2008.04.017
Matvssek. R; Wieser. G; Ceulemans. R: Rennenberg. H; Pretzsch. H; Haberer. K; Low. M; Nunn. AT:
Werner. H; Wipfler. P; Obwald. W; Nikolova. P; Hanke. DE; Kraigher. H; Tausz. M; Bahnweg. G;
Kitao. M; Dieler. T; Sandermann. H; Herbinger. K; Grebenc. T; Blumenrother. M; Deckmvn. G;
Grams. TEE; Heerdt. C; Leuchner. M; Fabian. P; Haberle. KH. (2010). Enhanced ozone strongly
reduces carbon sink strength of adult beech (Fagus sylvatica): Resume from the free-air fumigation
study at Kranzberg Forest. Environ Pollut 158: 2527-2532.
http://dx.doi.Org/10.1016/i.envpol.2010.05.009
Mavitv. E; Berrang. P. (1994). Effects of ozone on Solidago albopilosa, and endangered species of
goldenrod from central Kentucky [Abstract]. Bull Ecol Soc Am 75: 146.
McBride. TR; Laven. RD. (1999). Impact of oxidant air pollutants on forest succession in the mixed
conifer forests of the San Bernardino Mountains. In PR Miller; TR McBride (Eds.), Oxidant air
pollution impacts in the montane forests of southern California: A case study of the San Bernardino
Mountains (pp. 338-352). New York, NY: Springer. http://dx.doi.org/10.10Q7/978-l-4612-1436-
6 16
Mcfrederick. OS; Kathilankal. TC; Fuentes. TP. (2008). Air pollution modifies floral scent trails.
Atmos Environ 42: 2336-2348. http://dx.doi.Org/10.1016/i.atmosenv.2007.12.033
Mcgrath. TM; Betzelberger. AM; Wang. S; Shook. E; Zhu. XG; Long. SP; Ainsworth. EA. (2015). An
analysis of ozone damage to historical maize and soybean yields in the United States. Proc Natl
Acad Sci USA 112: 14390-14395. http://dx.doi.org/10.1073/pnas.1509777112
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.Org/10.llll/i.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.org/10.111 l/i.l469-8137.2007.01970.x
Meehan. TP; Couture. TT; Bennett. AE; Lindroth. RL. (2014). Herbivore-mediated material fluxes in a
northern deciduous forest under elevated carbon dioxide and ozone concentrations. New Phytol
204: 397-407. http://dx.doi.org/10.1111/nph. 12947
Menendez. AI; Gundel. PE; Lores. LM; Aleiandra Martinez-Ghersa. M. (2017). Assessing the impacts
of intra- and interspecific competition between Triticum aestivum and Trifolium repens on the
species' responses to ozone. Botany 95: 923-932. http://dx.doi.org/10.1139/cib-2016-Q275
8-233
-------
Menendez. AI; Romero. AM; Folcia. AM; Martinez-Ghersa. MA. (2009). Getting the interactions
right: Will higher 03 levels interfere with induced defenses to aphid feeding? Basic Appl Ecol 10:
255-264. http://dx.doi.Org/10.1016/i.baae.2008.03.010
Menendez. AI; Romero. AM; Folcia. AM; Martinez-Ghersa. MA. (2010). Aphid and episodic 03
injury in arugula plants (Eruca sativa Mill) grown in open-top field chambers. Agric Ecosyst
Environ 135: 10-14. htto://dx.doi.org/10.1016/i.agee.2009.08.005
Miglietta. F; Peressotti. A; Vaccari. FP; Zaldei. A; Deangelis. P; Scarascia-Mugnozza. G. (2001).
Free-air C02 enrichment (FACE) of a poplar plantation: the POPFACE fumigation system. New
Phytol 150: 465-476. htto://dx.doi.org/10.1046/i. 1469-8137.2001.00115.x
Miller. P; Guthrev. R; Schilling. S; Carroll. J. (1998). Ozone injury responses of ponderosa and Jeffrey
pine in the Sierra Nevada and San Bernardino Mountains in California. General Technical Report -
Pacific Southwest Research Station. U.S. Department of Agriculture, Forest Service :: USDA.
Miller. PL. (1973). Oxidant-induced community change in a mixed conifer forest. In JA Naegele
(Ed.), Air pollution damage to vegetation (pp. 101-117). Washington, DC: American Chemical
Society.
Mills. G; Harmens. H; Wagg. S; Sharps. K; Haves. F; Fowler. D; Sutton. M; Davies. B. (2016). Ozone
impacts on vegetation in a nitrogen enriched and changing climate. Environ Pollut 208: 898-908.
http://dx.doi.Org/10.1016/i.envpol.2015.09.038
Mills. G; Haves. F; Jones. MLM; Cinderbv. S. (2007). Identifying ozone-sensitive communities of
(semi-)natural vegetation suitable for mapping exceedance of critical levels. Environ Pollut 146:
736-743. http://dx.doi.Org/10.1016/i.envpol.2006.04.005
Mills. G; Pleiiel. H; Braun. S; Biiker. P; Bermeio. V; Calvo. E; Danielsson. H; Emberson. L;
Fernandez. IG; Griinhage. L; Harmens. H; Haves. F; Karlsson. P. erE; Simpson. D. (2011). New
stomatal flux-based critical levels for ozone effects on vegetation. Atmos Environ 45: 5064-5068.
http://dx.doi.Org/10.1016/i.atmosenv.2011.06.009
Mills. G; Pleiiel. H; Mallev. CS; Sinha. B; Cooper. OR; Schultz. MG; Neufeld. HS; Simpson. D;
Sharps. K; Feng. Z; Gerosa. G; Harmens. H; Kobavashi. K; Saxena. P; Paoletti. E; Sinha. V; Xu.
X. (2018). Tropospheric Ozone Assessment Report: Present-day tropospheric ozone distribution
and trends relevant to vegetation. Elementa: Science of the Anthropocene 6.
http://dx.doi.org/10.1525/elementa.302
Mills. G; Wagg. S; Harmens. H. (2013). Ozone pollution: Impacts on ecosystem services and
biodiversity. Bangor, UK: NERC/Centre for Ecology & Hydrology.
http://nora.nerc.ac.uk/id/eprint/502675/
Mina. U; Fuloria. A; Aggarwal. R. (2016). Effect of ozone and antioxidants on wheat and its pathogen
- Bipolaris sorokininana. Cereal Research Communications 44: 594-604.
http://dx.doi.org/10.1556/0806.44.2016.Q39
Mofikova. AO; Kim. TH; Abd El-Raheem. AM; Blande. JD; Kivimaenpaa. M; Holopainen. JK.
(2017). Passive adsorption of volatile monoterpene in pest control: Aided by proximity and
disrupted by ozone. J Agric Food Chem 65: 9579-9586. http://dx.doi.org/10.1021/acs.iafc.7b03251
Mondor. EB; Tremblav. MN; Awmack. CS; Lindroth. RL. (2005). Altered genotypic and phenotypic
frequencies of aphid populations under enriched C02 and 03 atmospheres. Global Change Biol 11:
1990-1996. http://dx.doi.org/10.1111/i. 1365-2486.2005.01054.x
Moran. EV; Kubiske. ME. (2013). Can elevated C02 and ozone shift the genetic composition of aspen
(Populus tremuloides) stands? New Phytol 198: 466-475. http://dx.doi.org/10.Ill 1/nph. 12153
8-234
-------
Morgan. PB; Ainsworth. EA; Long. SP. (2003). How does elevated ozone impact soybean? A meta-
analysis of photosynthesis, growth and yield. Plant Cell Environ 26: 1317-1328.
http://dx.doi.Org/10.1046/i.0016-8025.2003.01056.x
Morgan. PB; Bernacchi. CJ; Ort. PR; Long. SP. (2004). An in vivo analysis of the effect of season-
long open-air elevation of ozone to anticipated 2050 levels on photosynthesis in soybean. J Plant
Physiol 135: 2348-2357. http://dx.doi.org/10.1104/pp.104.043968
Morgan. PB: Mies. TA: Bollero. GA: Nelson. RL; Long. SP. (2006). Season-long elevation of ozone
concentration to projected 2050 levels under fully open-air conditions substantially decreases the
growth and production of soybean. New Phytol 170: 333-343. http://dx.doi.Org/10.l 11 l/i.1469-
8137.2006.01679.x
Morskv. SK; Haapala. JK; Rinnan. R; Tiiva. P; Saarnio. S: Silvola. J: Holopainen. T; Martikainen. PJ.
(2008). Long-term ozone effects on vegetation, microbial community and methane dynamics of
boreal peatland microcosms in open-field conditions. Global Change Biol 14: 1891-1903.
http://dx.doi.Org/10.llll/i.1365-2486.2008.01615.x
Mortensen. L. (1994a). The influence of carbon dioxide or ozone concentration on growth and
assimilate partitioning in seedlings of nine conifers. Acta Agric Scand B Soil Plant Sci 44: 157-
163. http://dx.doi.org/10.1080/0906471940941Q239
Mortensen. LM. (1992). Effects of ozone on growth of seven grass and one clover species. Acta Agric
Scand B Soil Plant Sci 42: 235-239. http://dx.doi.org/10.1080/09064719209410218
Mortensen. LM. (1993). Effects of ozone on growth of several subalpine plant species. Norweg J Agr
Sci 7: 129-138.
Mortensen. LM. (1994b). Further studies on the effects of ozone concentration on growth of subalpine
plant species. Norweg J Agr Sci 8: 91-97.
Mortensen. LM: Nilsen. J. (1992). Effects of ozone and temperature on growth of several wild plant
species. Norweg J Agr Sci 6: 195-204.
Mortensen. LM: Skre. O. (1990). Effects of low ozone concentrations on growth of Betula pubescens
Ehrh, Betula verrucosa Ehrh and Alnus incana (L) Moench. New Phytol 115: 165-170.
http://dx.doi.Org/10.l 111/i. 1469-8137.1990.tb00934.x
Muntifering. RB: Crosby. DP: Powell. MC: Chappelka. AH. (2000). Yield and quality characteristics
of bahiagrass (Paspalum notatum) exposed to ground-level ozone. Anim Feed Sci Technol 84: 243-
256. http://dx.doi.org/10.1016/S0377-8401(00)00124-3
Musselman. RC: Massman. WJ. (1999). Ozone flux to vegetation and its relationship to plant response
and ambient air quality standards. Atmos Environ 33: 65-73.
NAPCA (National Air Pollution Control Administration). (1970). Air quality criteria for
photochemical oxidants [EPA Report]. (AP-63). Washington, DC: U.S. Department of Health,
Education, and Welfare, http://nepis.epa.gov/Exe/ZvPURL.cgi?Dockev=9100E2Z7.txt
Neufeld. HS: Johnson. J: Kohut. R. (2018). Comparative ozone responses of cutleaf coneflowers
(Rudbeckia laciniata var. digitata, var. ampla) from Rocky Mountain and Great Smoky Mountains
National Parks, USA. Sci Total Environ 610-611: 591-601.
http ://dx.doi .org/10.1016/i. scitotenv.2017.08.046
Neufeld. HS: Lee. EH: Renfro. JR; Hacker. WD. (2000). Seedling insensitivity to ozone for three
conifer species native to Great Smoky Mountains National Park. Environ Pollut 108: 141-151.
http://dx.doi.org/10.1016/S0269-749 K99)00247-X
8-235
-------
Neufeld. HS; Lee. EH; Renfro. JR; Hacker. WD; Yu. BH. (1995). Sensitivity of seedlings of black
cherry (Prunus serotina Ehrh) to ozone in Great Smoky Mountains National Park I Exposure-
response curves for biomass. New Phytol 130: 447-459. http://dx.doi.Org/10.l 11 l/j.1469-
8137.1995 .tbO 1839.x
Noble. R; Jensen. KF; Ruff. BS; Loats. K. (1992). Response of Acer saccharum seedlings to elevated
carbon dioxide and ozone. Ohio J Sci 92: 60-62.
Norbv. RJ; Zak. DR. (2011). Ecological Lessons from Free-Air C02 Enrichment (FACE)
Experiments. Annu Rev Ecol Evol Systemat 42: 181-203. http://dx.doi.org/10.1146/annurev-
ecolsvs-102209-144647
Novak. K; Skellv. JM; Schaub. M; Krauchi. N; Hug. C; Landolt. W; Bleuler. P. (2003). Ozone air
pollution and foliar injury development on native plants of Switzerland. Environ Pollut 125: 41-52.
http://dx.doi.org/10.1016/S0269-7491(03')00085-X
Nunn. AJ; Reiter. IM; Haberle. KH; Werner. H; Langebartels. C; Sandermann. H; Heerdt. C; Fabian.
P; Matvssek. R. (2002). "Free-air" ozone canopy fumigation in an old-growth mixed forest:
Concept and observations in beech. Phyton-Annales Rei Botanicae 42: 105-119.
Oikawa. S; Ainsworth. EA. (2016). Changes in leaf area, nitrogen content and canopy photosynthesis
in soybean exposed to an ozone concentration gradient. Environ Pollut 215: 347-355.
http://dx.doi.Org/10.1016/i.envpol.2016.05.005
Oksanen. E. (2003). Responses of selected birch (Betula pendula Roth) clones to ozone change over
time. Plant Cell Environ 26: 875-886. http://dx.doi.Org/10.1046/i.1365-3040.2003.01020.x
Olszvk. DM; Kats. G; Morrison. CL; Dawson. PJ; Gocka. I; Wolf. J; Thompson. CR. (1990).
Valencia' orange fruit yield with ambient oxidant or sulfur dioxide exposures. J Am Soc Hortic Sci
115: 878-883. http://dx.doi.org/10.21273/JASHS.115.6.878
Orendovici. T; Skellv. JM; Ferdinand. JA; Savage. JE; Sanz. MJ; Smith. GC. (2003). Response of
native plants of northeastern United States and southern Spain to ozone exposures; determining
exposure/response relationships. Environ Pollut 125: 31-40. http://dx.doi.org/10.1016/SQ269-
7491(03)00089-7
Ormrod. DP; Marie. BA; Allen. OB. (1988). Research approaches to pollutant crop loss functions. In
WW Heck; OC Taylor; DT Tingey (Eds.), Assessment of crop loss from air pollutants: proceedings
of an international conference; October 1987; Raleigh, NC (pp. 27-44). Essex, United Kingdom:
Elsevier Science Publishers, Ltd.
Osborne. SA; Mills. G; Haves. F; Ainsworth. EA; Biiker. P; Emberson. L. (2016). Has the sensitivity
of soybean cultivars to ozone pollution increased with time? An analysis of published dose-
response data. Global Change Biol 22: 3097-3111. http://dx.doi.org/10.llll/gcb.13318
Oshima. RJ; Braegelmann. PK; Baldwin. DW; Van Way. V; Taylor. OC. (1977). Reduction of tomato
fruit size and yield by ozone. J Am Soc Hortic Sci 102: 289-293.
Oshima. RJ; Poe. MP; Braegelmann. PK; Baldwin. DW; Van Way. V. (1976). Ozone dosage-crop loss
function for alfalfa: A standardized method for assessing crop losses from air pollutants. J Air
Pollut Control Assoc 26: 861-865. http://dx.doi.org/10.1080/00022470.1976.10470330
Ozolincius. R; Serafinaviciute. B. (2003). Ozone-induced visible foliar injuries in Lithuania. Baltic
Forestry 9: 51-57.
Panek. JA. (2004). Ozone uptake, water loss and carbon exchange dynamics in annually drought-
stressed Pinus ponderosa forests: Measured trends and parameters for uptake modeling. Tree
Physiol 24: 277-290. http://dx.doi.Org/10.1093/treephvs/24.3.277
8-236
-------
Panek. JA; Goldstein. AH. (2001). Responses of stomatal conductance to drought in ponderosa pine:
Implications for carbon and ozone uptake. Tree Physiol 21: 337-344.
http://dx.doi.Org/10.1093/treephvs/21.5.337
Paoletti. E; Grulke. NE. (2010). Ozone exposure and stomatal sluggishness in different plant
physiognomic classes. Environ Pollut 158: 2664-2671.
http://dx.doi.Org/10.1016/i.envpol.2010.04.024
Paoletti. E: Manning. WJ. (2007). Toward a biologically significant and usable standard for ozone that
will also protect plants [Review]. Environ Pollut 150: 85-95.
http://dx.doi.Org/10.1016/i.envpol.2007.06.037
Parsons. WFJ: Bockheim. JG: Lindroth. RL. (2008). Independent, interactive, and species-specific
responses of leaf litter decomposition to elevated C02 and 03 in a northern hardwood forest.
Ecosystems 11: 505-519. http://dx.doi.org/10.1007/slQ021-008-9148-x
Paudel. R: Grantz. DA: Vu. HB: Shrestha. A. (2016). Tolerance of elevated ozone and water stress in
a California population of Palmer amaranth (Amaranthus palmeri). Weed Sci 64: 276-284.
http://dx.doi.Org/10.1614AVS-D-15-00146.l
Payne. RJ: Stevens. CJ: Pise. NB: Gowing. D.T: Pilkington. MG: Phoenix. GK: Emmett. BA:
Ashmore. MR. (2011). Impacts of atmospheric pollution on the plant communities of British acid
grasslands. Environ Pollut 159: 2602-2608. http://dx.doi.Org/10.1016/i.envpol.2011.06.009
Payne. RJ: Toet. S: Ashmore. M: Jassev. VEJ: Gilbert. D. (2017). Impacts of tropospheric ozone
exposure on peatland microbial consumers. Soil Biol Biochem 115: 124-128.
http ://dx.doi .org/10.1016/i. soilbio .2017.08.012
Pell. EJ: Sinn. JP; Brendlev. BW: Samuelson. L; Vinten-Johansen. C: Tien. M; Skillman. J. (1999).
Differential response of four tree species to ozone-induced acceleration of foliar senescence. Plant
Cell Environ 22: 779-790. http://dx.doi.Org/10.1046/i.1365-3040.1999.00449.x
Peltonen. PA: Vapaavuori. E; Heinonen. J: Julkunen-Tiitto. R; Holopainen. JK. (2010). Do elevated
atmospheric C02 and 0-3 affect food quality and performance of folivorous insects on silver
birch? Global Change Biol 16: 918-935. http://dx.doi.org/10.1111/i.1365-2486.2009.02073,x
Percy. KE; Awmack. CS: Lindroth. RL: Kubiske. ME: Kopper. BJ; Isebrands. JG: Pregitzer. KS:
Hendry. GR; Dickson. RE: Zak. PR: Oksanen. E; Sober. J: Harrington. R; Karnosky. DF. (2002).
Altered performance of forest pests under atmospheres enriched with C02 and 03. Nature 420:
403-407. http://dx.doi.org/10.1038/natureO 1028
Peterson. PL: Arbaugh. MJ; Wakefield. VA; Miller. PR. (1987). Evidence of growth reduction in
ozone-injured Jeffrey pine (Pinus jeffreyi Grev and Balf) in Sequoia and Kings Canyon National
Parks. J Air Waste Manag Assoc 37: 906-912. http://dx.doi.org/10.1080/0894063Q.1987.10466283
Pfleeger. TG: Plocher. M; Bichel. P. (2010). Response of pioneer plant communities to elevated ozone
exposure. Agric Ecosyst Environ 138: 116-126. http://dx.doi.Org/10.1016/i.agee.2010.04.009
Piikki. K: Pe Temmerman. L: Hogy. P: Pleiiel. H. (2008). The open-top chamber impact on vapour
pressure deficit and its consequences for stomatal ozone uptake. Atmos Environ 42: 6513-6522.
http://dx.doi.Org/10.1016/i.atmosenv.2008.04.014
Pinto. PM: Blande. JP: Nvkanen. R: Pong. WX: Nerg. AM: Holopainen. JK. (2007). Ozone degrades
common herbivore-induced plant volatiles: Poes this affect herbivore prey location by predators
and parasitoids? J Chem Ecol 33: 683-694. http://dx.doi.org/10.1007/slQ886-007-9255-8
8-237
-------
Pinto. DM; Blande. JD; Souza. SR; Nerg. AM; Holopainen. JK. (2010). Plant volatile organic
compounds (VOCs) in ozone (03) polluted atmospheres: The ecological effects [Review]. J Chem
Ecol 36: 22-34. http://dx.doi.org/10.1007/slQ886-009-9732-3
Pinto. DM; Himanen. SJ; Nissinen. A; Nerg. AM; Holopainen. JK. (2008). Host location behavior of
Cotesia plutellae Kurdjumov (Hymenoptera: Braconidae) in ambient and moderately elevated
ozone in field conditions. Environ Pollut 156: 227-231.
http://dx.doi.Org/10.1016/i.envpol.2007.12.009
Pleiiel. H; Broberg. MC; Uddling. J; Mills. G. (2018). Current surface ozone concentrations
significantly decrease wheat growth, yield and quality. Sci Total Environ 613-614: 687-692.
http ://dx.doi .org/10.1016/i. scitotenv.2017.09.111
Pleiiel. H; Danielsson. H. (1997). Growth of 27 herbs and grasses in relation to ozone exposure and
plant strategy. New Phytol 135: 361-367. http://dx.doi.Org/10.1046/i.1469-8137.1997.00648.x
Pleiiel. H; Danielsson. H; Simpson. D; Mills. G. (2014). Have ozone effects on carbon sequestration
been overestimated? A new biomass response function for wheat. Biogeosciences 11: 4521-4528.
http://dx.doi.org/10.5194/bg-ll-4521-2014
Powell. MC; Muntifering. RB; Lin. JC; Chappelka. AH. (2003). Yield and nutritive quality of sericea
lespedeza (Lespedeza cuneata) and little bluestem (Schizachyrium scoparium) exposured to
ground-level ozone. Environ Pollut 122: 313-322. http://dx.doi.org/10.1016/S0269-7491(02)00331-
7
Power. SA; Ashmore. MR. (2002). Responses of fen and fen-meadow communities to ozone. New
Phytol 156: 399-408. http://dx.doi.Org/10.1046/i.1469-8137.2002.00540.x
Pregitzer. K; Lova. W; Kubiske. M; Zak. D. (2006). Soil respiration in northern forests exposed to
elevated atmospheric carbon dioxide and ozone. Oecologia 148: 503-516.
http://dx.doi.org/10.1007/s00442-006-Q381-8
Pregitzer. KS; Burton. AJ; King. JS; Zak. DR. (2008). Soil respiration, root biomass, and root turnover
following long-term exposure of northern forests to elevated atmospheric Co-2 and tropospheric O-
3. New Phytol 180: 153-161. http://dx.doi.Org/10.llll/i.1469-8137.2008.02564.x
Pregitzer. KS; Euskirchen. ES. (2004). Carbon cycling and storage in world forests: biome patterns
related to forest age. Global Change Biol 10: 2052-2077. http://dx.doi.Org/10.l 11 l/i.1365-
2486.2004.00866.x
Pregitzer. KS; Talhelm. AF. (2013). Belowground carbon cycling at Aspen FACE: Dynamic responses
to C02 and 03 in developing forests. In R Matyssek; N Clarke; P Cudlin; TN Mikkelsen; JP
Tuovinen; G Wieser; E Paoletti (Eds.), Climate change, air pollution and global challenges:
Understanding and perspectives from forest research (pp. 209-226). Amsterdam, Netherlands:
Elsevier. http://dx.doi.org/10.1016/B978-0-08-098349-3.0001Q-4
Pritsch. K; Esperschuetz. J; Haesler. F; Raidl. S; Winkler. B; Schloter. M. (2009). Structure and
activities of ectomycorrhizal and microbial communities in the rhizosphere of Fagus sylvatica
under ozone and pathogen stress in a lysimeter study. Plant Soil 323: 97-109.
http://dx.doi.org/10.1007/slllQ4-009-9972-6
Puiol Pereira. EI; Chung. H; Scow. K; Sadowskv. MJ; van Kessel. C; Six. J. (2011). Soil nitrogen
transformations under elevated atmospheric CO and O during the soybean growing season. Environ
Pollut 159: 401-407. http://dx.doi.Org/10.1016/i.envpol.2010.10.033
Oiu. Z; Chappelka. AH; Somers. GL; Lockabv. BG; Meldahl. RS. (1992). Effects of ozone and
simulated acidic precipitation on above- and below-ground growth of loblolly pine (Pinus taeda).
Can J For Res 22: 582-587.
8-238
-------
Rao. MY; Hale. BA; Ormrod. DP. (1995). Amelioration of ozone-induced oxidative damage in wheat
plants grown under high carbon dioxide: Role of antioxidant enzymes. J Plant Physiol 109: 421-
432.
Rasheed. MU; Kasurinen. A; Kivimaenpaa. M; Ghimire. R; Haikio. E; Mpamah. P; Holopainen. JK;
Holopainen. T. (2017). The responses of shoot-root-rhizosphere continuum to simultaneous
fertilizer addition, warming, ozone and herbivory in young Scots pine seedlings in a high latitude
field experiment. Soil Biol Biochem 114: 279-294. http://dx.doi.org/10.1016/i.soilbio.2017.07.024
Rawlings. JO; Cure. WW. (1985). The Weibull function as a dose-response model to describe ozone
effects on crop yields. Crop Sci 25: 807-814.
Rebbeck. J. (1996a). Chronic ozone effects on three northeastern hardwood species: growth and
biomass. Can J For Res 26: 1788-1798. http://dx.doi.Org/10.l 1397x26-203
Rebbeck. J. (1996b). The chronic response of yellow-poplar and eastern white pine to ozone and
elevated carbon dioxide: three-year summary (pp. 23-30). U.S. Department of Agriculture, Forest
Service :: USDA. https://www,fs.usda.gov/treesearch/pubs/13377
Reiling. K; Davison. AW. (1992). The response of native, herbaceous species to ozone: growth and
fluorescence screening. New Phytol 120: 29-37. http://dx.doi.org/10.1111/i .1469-
8137.1992.tb01055.x
Reinert. RA; Eason. G; Barton. J. (1997). Growth and fruiting of tomato as influenced by elevated
carbon dioxide and ozone. New Phytol 137: 411-420. http://dx.doi.org/10.1046/i. 1469-
8137.1997.00846.x
Reinert. RA; Ho. MC. (1995). Vegetative growth of soybean as affected by elevated carbon dioxide
and ozone. Environ Pollut 89: 89-96. http://dx.doi.org/10.1016/0269-7491(94)00039-G
Ren. W; Tian. H; Tao. B; Huang. Y; Pan. S. (2012). China's crop productivity and soil carbon storage
as influenced by multifactor global change. Global Change Biol 18: 2945-2957.
http://dx.doi.Org/10.l 111/i. 1365-2486.2012.02741.x
Resco de Dios. V; Mereed. TE; Ferrio. JP; Tissue. DT; Voltas. J. (2016). Intraspecific variation in
juvenile tree growth under elevated C02 alone and with 03: a meta-analysis. Tree Physiol 36: 682-
693. http://dx.doi.org/10.1093/treephvs/tpw026
Rhea. LK; King. JS. (2012). Depth-dependency of trembling aspen and paper birch small-root
responses to eCO(2) and eO(3). Plant Soil 355: 215-229. http://dx.doi.org/10.1007/sl 1104-011-
1094-2
Richards. BL; Middleton. JT; Hewitt. WB. (1958). Air pollution with relation to agronomic crops: V:
Oxidant stipple of grape. J Am Soc Agron 50: 559-561.
http://dx.doi.org/10.2134/agronil958.00021962005000090Q19x
Riddell. J; Padgett. PE; Nash. TH. III. (2012). Physiological responses of lichens to factorial
fumigations with nitric acid and ozone. Environ Pollut 170: 202-210.
http://dx.doi.Org/10.1016/i.envpol.2012.06.014
Ritter. W; Andersen. CP; Matvssek. R; Grams. TEE. (2011). Carbon flux to woody tissues in a
beech/spruce forest during summer and in response to chronic 0-3 exposure. Biogeosciences 8:
3127-3138. http://dx.doi.org/10.5194/bg-8-3127-2011
Rodenkirchen. H; Gottlein. A; Kozovits. AR; Matvssek. R; Grams. TEE. (2009). Nutrient contents and
efficiencies of beech and spruce saplings as influenced by competition and 03/C02 regime.
European Journal of Forest Research 128: 117-128. http://dx.doi.org/10.1007/slQ342-008-0221-v
8-239
-------
Rogers. A; Allen. DJ; Davev. PA; Morgan. PB; Ainsworth. EA; Bernacchi. CJ; Comic. G; Dermodv.
OC; Dohleman. FG; Heaton. EA; Mahonev. J; Zhu. XG; Delucia. EH; Ort. PR; Long. SP. (2004).
Leaf photosynthesis and carbohydrate dynamics of soybean grown throughout their life-cycle
under free-air carbon dioxide enrichment. Plant Cell Environ 27: 449-458.
Romaneckiene. R; Pilipavicius. V; Romaneckas. K. (2008). The influence of ozone and uv-b radiation
on fat-hen growth in different temperature conditions. Zemdirbyste-Agriculture 95: 122-132.
Runeckles. VC; Wright. EF. (1996). Delayed impact of chronic ozone stress on young Douglas-fir
grown under field conditions. Can J For Res 26: 629-638. http://dx.doi.org/10.11397x26-073
Samuelson. LJ; Kelly. JM; Mays. PA; Edwards. GS. (1996). Growth and nutrition of Quercus rubra L
seedlings and mature trees after three seasons of ozone exposure. Environ Pollut 91: 317-323.
http://dx.doi.org/10.1016/0269-7491(95)00067-4
Sanz. J; Bermeio. V; Muntifering. R; Gonzalez-Fernandez. I; Gimeno. BS; Elvira. S; Alonso. R.
(2011). Plant phenology, growth and nutritive quality of Briza maxima: responses induced by
enhanced ozone atmospheric levels and nitrogen enrichment. Environ Pollut 159: 423-430.
http://dx.doi.Org/10.1016/i.envpol.2010.10.026
Sanz. J: Gonzalez-Fernandez. I; Elvira. S; Muntifering. R; Alonso. R; Bermeio-Bermeio. V. (2016).
Setting ozone critical levels for annual Mediterranean pasture species: Combined analysis of open-
top chamber experiments. Sci Total Environ 571: 670-679.
http://dx.doi.Org/10.1016/i.scitotenv.2016.07.035
Saviranta. NMM; Julkunen-Tiitto. R; Oksanen. E; Karialainen. RO. (2010). Leaf phenolic compounds
in red clover (Trifolium pratense L.) induced by exposure to moderately elevated ozone. Environ
Pollut 158: 440-446. http://dx.doi.Org/10.1016/i.envpol.2009.08.029
Schmidt. MWI; Torn. MS; Abiven. S; Dittmar. T; Guggenberger. G; Janssens. IA; Kleber. M; Koegel-
Knabner. I; Lehmann. J: Manning. DAC; Nannipieri. P; Rasse. DP; Weiner. S; Trumbore. SE.
(2011). Persistence of soil organic matter as an ecosystem property. Nature 478: 49-56.
http://dx.doi.org/10.1038/naturel0386
Seiler. LK; Decoteau. PR; Davis. DP. (2014). Evaluation of Ailanthus altissima as a bioindicator to
detect phytotoxic levels of ozone. Northeast Nat 21: 541-553.
http://dx.doi.org/10.1656/045.021.04Q5
Shafer. SR; Reinert. RA; Eason. G; Spruill. SE. (1993). Analysis of ozone concentration - biomass
response relationships among open-pollinated families of loblolly pine. Can J For Res 23: 706-715.
Shelburne. VB; Reardon. JC; Pavnter. VA. (1993). The effects of acid rain and ozone on biomass and
leaf area parameters of shortleaf pine (Pinus echinata Mill). Tree Physiol 12: 163-172.
Simini. M; Skellv. JM; Pavis. PP; Savage. JE; Comrie. AC. (1992). Sensitivity of four hardwood
species to ambient ozone in north central Pennsylvania. Can J For Res 22: 1789-1799.
http://dx.doi.org/10.1139/x92-234
Sitch. S; Cox. PM; Collins. WJ; Huntingford. C. (2007). Indirect radiative forcing of climate change
through ozone effects on the land-carbon sink. Nature 448: 791-794.
http://dx.doi.org/10.1038/nature06059
Skellv. JM; Innes. JL; Savage. JS; Snyder. KR; Vanderhevden. PZ; Sanz. MJ. (1999V Observation
and confirmation of foliar injury ozone symptoms of native plant species of Switzerland and
southern Spain. Water Air Soil Pollut 116: 227-234. http://dx.doi.Org/10.1023/A: 1005275431399
8-240
-------
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/slQ661-011-
2243-z
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. http://dx.doi.org/10.1023/A: 1024879527764
Smith. GC; Coulston. JW; O'Connell. BM. (2008). Ozone bioindicators and forest health: a guide to
the evaluation, analysis, and interpretation of the ozone injury data in the Forest Inventory and
Analysis Program. (NRS-34). Newtown Square, PA: U.S. Department of Agriculture, Forest
Service. http://dx.doi.org/10.2737/NRS-GTR-34
Somers. GL; Chappelka. AH; Rosseau. P; Renfro. JR. (1998). Empirical evidence of growth decline
related to visible ozone injury. For Ecol Manage 104: 129-137. http://dx.doi.org/10.1016/SQ378-
1127(97)00252-1
Souza. L; Neufeld. HS; Chappelka. AH; Burkev. KO; Davison. AW. (2006). Seasonal development of
ozone-induced foliar injury on tall milkweed (Asclepias exaltata) in Great Smoky Mountains
National Park. Environ Pollut 141: 175-183. http://dx.doi.Org/10.1016/i.envpol.2005.07.022
Spence. RD; Rvkiel. EJ. Jr; Sharpe. PJH. (1990). Ozone alters carbon allocation in loblolly pine:
assessment with carbon-11 labeling. Environ Pollut 64: 93-106. http://dx.doi.org/10.1016/Q269-
7491(90)90107-N
Stampfli. A; Fuhrer. J. (2010). Spatial heterogeneity confounded ozone-exposure experiment in semi-
natural grassland. Oecologia 162: 515-522. http://dx.doi.org/10.1007/s00442-009-1462-2
Stoelken. G; Pritsch. K; Simon. J; Mueller. CW; Grams. TEE; Esperschuetz. J; Gavler. S; Buegger. F;
Brueggemann. N: Meier. R; Zeller. B: Winkler. JB; Rennenherg. H (2010). Enhanced ozone
exposure of European beech (Fagus sylvatica) stimulates nitrogen mobilization from leaf litter and
nitrogen accumulation in the soil. Plant Biosyst 144: 537-546.
http://dx.doi.org/10.1080/112635009Q3429346
Subramanian. N; Karlsson. PE; Bergh. J; Nilsson. U. (2015). Impact of ozone on sequestration of
carbon by Swedish forests under a changing climate: A modeling study. Forest Sci 61: 445-457.
http://dx.doi.org/10.5849/forsci.14-026
Suding. KN; Collins. SL; Gough. L; Clark. C; Cleland. EE; Gross. KL; Milchunas. DG; Pennings. S.
(2005). Functional- and abundance-based mechanisms explain diversity loss due to N fertilization.
Proc Natl Acad Sci USA 102: 4387-4392. http://dx.doi.org/10.1073/pnas.04086481Q2
Sun. G; Mclaughlin. SB; Porter. JH; Uddling. J; Mulholland. PJ; Adams. MB; Pederson. N. (2012).
Interactive influences of ozone and climate on streamflow of forested watersheds. Global Change
Biol 18: 3395-3409. http://dx.doi.org/10.1111/i. 1365-2486.2012.02787.X
Sun. J; Feng. Z; Ort. DR. (2014). Impacts of rising tropospheric ozone on photosynthesis and
metabolite levels on field grown soybean. Plant Sci 226: 147-161.
http://dx.doi.Org/10.1016/i.plantsci.2014.06.012
Super. I; de Arellano. JVG; Krol. MC. (2015). Cumulative ozone effect on canopy stomatal resistance
and the impact on boundary layer dynamics and C02 assimilation at the diurnal scale: A case study
for grassland in the Netherlands. Jour Geo Res: Biog 120: 1348-1365.
http://dx.doi.org/10.1002/2015JG0Q2996
Szantoi. Z; Chappelka. AH; Muntifering. RB; Somers. GL. (2007). Use of ethylenediurea (EDU) to
ameliorate ozone effects on purple coneflower (Echinacea purpurea). Environ Pollut 150: 200-208.
http://dx.doi.Org/10.1016/i.envpol.2007.01.020
8-241
-------
Szantoi. Z; Chappelka. AH; Muntifcring. RB: Somers. GL. (2009). Cutleaf coneflower (Rudbeckia
laciniata L.) response to ozone and ethylenediurea (EDU). Environ Pollut 157: 840-846.
http://dx.doi.Org/10.1016/i.envpol.2008.l 1.014
Tai. APK; Martin. MY. (2017). Impacts of ozone air pollution and temperature extremes on crop
yields: Spatial variability, adaptation and implications for future food security. Atmos Environ 169:
11-21. http://dx.doi.Org/10.1016/i.atmosenv.2017.09.002
Taia. W; Basahi. J: Hassan. I. (2013). Impact of ambient air on physiology, pollen tube growth, pollen
germination and yield in pepper (Capsicum annuum L.). Pakistan J Bot 45: 921-926.
Takemoto. BK: Bvtnerowicz. A: Dawson. PJ: Morrison. CL: Temple. PJ. (1997). Effects of ozone on
Pinus ponderosa seedlings: comparison of responses in the first and second growing seasons of
exposure. Can J For Res 27: 23-30. http://dx.doi.org/10.1139/cifr-27-l-23
Takemoto. BK; Bvtnerowicz. A; Fenn. ME. (2001). Current and future effects of ozone and
atmospheric nitrogen deposition on California's mixed conifer forests. For Ecol Manage 144: 159-
173. http://dx.doi.org/10.1016/S0378-l 127(00)00368-6
Talhelm. AF; Pregitzer. KS; Giardina. CP. (2012). Long-term leaf production response to elevated
atmospheric carbon dioxide and tropospheric ozone. Ecosystems 15: 71-82.
http://dx.doi.org/10.1007/slQ021-011-9493-z
Talhelm. AF; Pregitzer. KS; Kubiske. ME; Zak. PR; Campanv. CE; Burton. AJ; Dickson. RE;
Hendrev. GR; Isebrands. JG; Lewin. KF; Nagy. J; Karnoskv. DF. (2014). Elevated carbon dioxide
and ozone alter productivity and ecosystem carbon content in northern temperate forests. Global
Change Biol 20: 2492-2504. http://dx.doi.org/10. Ill 1/gcb. 12564
Taylor. M. (2002). The impact of ozone on a salt marsh cordgrass (Spartina alterniflora). Environ
Pollut 120: 701-705. http://dx.doi.org/10.1016/S0269-7491(02)00178-1
Telesnicki. MC; Aleiandra Martinez-Ghersa. M; Arneodo. JD; Ghersa. CM. (2015). Direct effect of
ozone pollution on aphids: Revisiting the evidence at individual and population scales. Entomol
Exp Appl 155: 71-79. http://dx.doi.Org/10.l 111/eea. 12288
Temple. P; Bvtnerowicz. A: Fenn. M; Poth. M. (2005). Air pollution impacts in the mixed conifer
forests of southern California. USDA Forest Service General Technical Report (pp. 145-164).
(PSW-GTR-195). U.S. Department of Agriculture, Forest Service :: USDA.
Temple. PJ. (1990). Growth and yield responses of processing tomato (Lycopersicon esculentum Mill)
cultivars to ozone. Environ Exp Bot 30: 283-291.
Temple. PJ. (1999). Effects of ozone on understory vegetation in the mixed conifer forest. In PR
Miller; JR McBride (Eds.), Oxidant air pollution impacts in the montane forests of southern
California: A case study of the San Bernardino Mountains (pp. 208-222). New York, NY: Springer.
http://dx.doi.org/10.1007/978-l-4612-1436-6 10
Temple. PJ; Jones. TE; Lennox. RW. (1990). YIELD LOSS ASSESSMENTS FOR CULTIVARS OF
BROCCOLI, LETTUCE, AND ONION EXPOSED TO OZONE. Environ Pollut 66: 289-299.
http://dx.doi.org/10.1016/0269-7491(90)90146-4
Temple. PJ; Kupper. RS; Lennox. RW; Rohr. K. (1988). Injury and yield responses of differentially
irrigated cotton to ozone. Agron J 80: 751-755.
http://dx.doi.org/10.2134/agronil988.0002196200800005Q011x
Temple. PJ; Miller. PR. (1994). Foliar ozone injury and radial growth of ponderosa pine. Can J For
Res 24: 1877-1882.
8-242
-------
Temple. PJ; Riechers. GH; Miller. PR. (1992). Foliar injury responses of ponderosa pine seedlings to
ozone, wet and dry acidic deposition, and drought. Environ Exp Bot 32: 101-113.
http://dx.doi.org/10.1016/0098-8472(92')90035-Z
Temple. PJ; Riechers. GH; Miller. PR; Lennox. RW. (1993). Growth responses of ponderosa pine to
long-term exposure to ozone, wet and dry acidic deposition, and drought. Can J For Res 23: 59-66.
http://dx.doi.Org/10.l 139/x93-010
Terrer. C; Vicca. S; Stocker. BP; Hungate. BA: Phillips. RP; Reich. PB; Finzi. AC; Prentice. IC.
(2018). Ecosystem responses to elevated C02 governed by plant-soil interactions and the cost of
nitrogen acquisition [Review]. New Phytol 217: 507-522. http://dx.doi.org/10.1111/nph. 14872
Tetteh. R; Yamaguchi. M; Wada. Y; Funada. R; Izuta. T. (2015). Effects of ozone on growth, net
photosynthesis and yield of two African varieties of Vigna unguiculata. Environ Pollut 196: 230-
238. http://dx.doi.Org/10.1016/i.envpol.2014.10.008
Thomas. RO; Canham. CD; Weathers. KC; Goodale. CL. (2010). Increased tree carbon storage in
response to nitrogen deposition in the US. Nat Geosci 3: 13-17.
http://dx.doi.org/10.1038/NGEQ721
Thompson. CR; Olszvk. DM; Kats. G; Bvtnerowicz. A; Dawson. PJ; Wolf. JW. (1984). Effects of
ozone or sulfur-dioxide on annual plants of the Mojave Desert. J Air Pollut Control Assoc 34:
1017-1022.
Tian. H; Chen. G; Zhang. C; Liu. M; Sun. G; Chappelka. A; Ren. W; Xu. X; Lu. C; Pan. S; Chen. H;
Hui. D; McNultv. S; Lockabv. G; Vance. E. (2012). Century-scale responses of ecosystem carbon
storage and flux to multiple environmental changes in the southern United States. Ecosystems 15:
674-694. http://dx.doi.org/10.1007/sl0021-Q12-9539-x
Tioelker. MG; Volin. JC; Oleksvn. J; Reich. PB. (1993). Light environment alters response to ozone
stress in seedlings of Acer saccharum Marsh and hybrid Populus L I In situ net photosynthesis,
dark respiration and growth. New Phytol 124: 627-636. http://dx.doi.org/10.1111/i.1469-
8137.1993.tb03852.x
Toet. S; Ineson. P; Peacock. S; Ashmore. M. (2011). Elevated ozone reduces methane emissions from
peatland mesocosms. Global Change Biol 17: 288-296. http://dx.doi.org/10.1111/i .1365-
2486.2010.02267.x
Tong. D; Mathur. R; Schere. K; Kang. D; Yu. S. (2007). The use of air quality forecasts to assess
impacts of air pollution on crops: Methodology and case study. Atmos Environ 41: 8772-8784.
http://dx.doi.Org/10.1016/i.atmosenv.2007.07.060
Tonneiick. AEG; Franzaring. J; Brouwer. G; Metselaar. K; Dueck. TA. (2004). Does interspecific
competition alter effects of early season ozone exposure on plants from wet grasslands? Results of
athree-year experiment in open-top chambers. Environ Pollut 131: 205-213.
http://dx.doi.Org/10.1016/i.envpol.2004.02.005
Turnipseed. AA; Burns. SP; Moore. DJP; Hu. J; Guenther. AB; Monson. RK. (2009). Controls over
ozone deposition to a high elevation subalpine forest. Agr Forest Meteorol 149: 1447-1459.
http://dx.doi.Org/10.1016/i.agrformet.2009.04.001
U.S. EPA (U.S. Environmental Protection Agency). (1977). Photochemical oxidant air pollutant
effects on a mixed conifer forest ecosystem: A progress report [EPA Report]. (EPA-600/3-77-104).
Corvallis, OR. http://nepis.epa.gov/Exe/ZvPURL.cgi?Dockev=20015IYP.txt
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, DC.
http://nepis.epa.g0v/exe/Z vPURL.cgi?Dockev=200089CW.txt
8-243
-------
U.S. EPA (U.S. Environmental Protection Agency). (1980). Photochemical oxidant air pollution
effects on a mixed conifer forest ecosystem: Final report, 1977 [EPA Report]. (EPA-600/3-80-
002). Corvallis, OR.
https://ntrl.ntis.gov/NTRL/dashboard/scarch Results.xhtmr.'scarchQucn^PBSO 176779
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.
https://ntrl.ntis.gov/NTRL/dashboard/searchResults.xhtml?searchQuery=PB87142949
U.S. EPA (U.S. Environmental Protection Agency). (1996). Air quality criteria for ozone and related
photochemical oxidants, Vol. II of III [EPA Report]. (EPA/600/P-93/004BF). Research Triangle
Park, NC.
U.S. EPA (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: U.S.
Environmental Protection Agency, Office of Research and Development, National Center for
Environmental Assessment-RTP Office.
http://cfbub.epa.gov/ncea/cfm/recordisplav.cfm?deid=149923
U.S. EPA (U.S. Environmental Protection Agency). (2008). HERO.
U.S. EPA (U.S. Environmental Protection Agency). (2009). HERO, the Database. Available online at
http://www.epa.gov/hero
U.S. EPA (U.S. Environmental Protection Agency). (2013). Integrated science assessment for ozone
and related photochemical oxidants [EPA Report]. (EPA/600/R-10/076F). Research Triangle Park,
NC: U.S. Environmental Protection Agency, Office of Research and Development, National Center
for Environmental Assessment-RTP Division.
http://cfpub.epa.gov/ncea/isa/recordisplav.cfm?deid=247492
U.S. EPA (U.S. Environmental Protection Agency). (2014). Welfare Risk and Exposure Assessment
for Ozone, Final. (EPA-452/P-14-005a). Research Triangle Park, NC: Office of Air Quality
Planning and Standards.
https://www3.epa.gov/ttn/naaqs/standards/ozone/data/20141021welfarerea.pdf
U.S. EPA (U.S. Environmental Protection Agency). (2015). Preamble to the integrated science
assessments [EPA Report]. (EPA/600/R-15/067). Research Triangle Park, NC: U.S. Environmental
Protection Agency, Office of Research and Development, National Center for Environmental
Assessment, RTP Division. https://cfpub.epa.gov/ncea/isa/recordisplav.cfm?deid=310244
U.S. EPA (U.S. Environmental Protection Agency). (2018). Integrated science assessment for oxides
of nitrogen, oxides of sulfur and particulate matterecological criteria (2nd external review draft).
(EPA/600/R-18/097). Research Triangle Park, NC: U.S. Environmental Protection Agency, Office
of Research and Development, National Center for Environmental Assessment.
https://cfpub. epa.gov/ncea/isa/recordisplav. cfm?deid=340671
Uddling. J; Teclaw. RM: Pregitzer. KS; Ellsworth. DS. (2009). Leaf and canopy conductance in aspen
and aspen-birch forests under free-air enrichment of carbon dioxide and ozone. Tree Physiol 29:
1367-1380. http://dx.doi.org/10.1093/treephvs/tpp070
Ueda. Y; Frindte. K; Knief. C; Ashrafuzzaman. M; Frei. M. (2016). Effects of elevated tropospheric
ozone concentration on the bacterial community in the phyllosphere and rhizoplane of rice. PLoS
ONE 11: e0163178. http://dx.doi.org/10.1371/iournal.pone.0163178
8-244
-------
Ueno. AC; Gundel. PE; Omacini. M; Ghersa. CM; Bush. LP; Aleiandra Martinez-Ghersa. M. (2016).
Mutualism effectiveness of a fungal endophyte in an annual grass is impaired by ozone. Funct Ecol
30: 226-234. http://dx.doi.org/10.1111/1365-2435.12519
Unsworth. MH; Heagle. AS; Heck. WW. (1984a). Gas exchange in open-top field chambers: I.
Measurement and analysis of atmospheric resistances to gas exchange. Atmos Environ 18: 373-
380. http://dx.doi.org/10.1016/0004-6981(84)90111-2
Unsworth. MH; Heagle. AS; Heck. WW. (1984b). Gas exchange in open-top field chambers: II.
Resistances to ozone uptake by soybeans. Atmos Environ 18: 381-385.
http ://dx. doi.org/10.1016/0004-6981(84)90112-4
USD A (U.S. Department of Agriculture). (2015). Natural Resources Conservation Service PLANTS
database. Available online at http://plants.usda.gov/iava/
Valkama. E; Koricheva. J; Oksanen. E. (2007). Effects of elevated 03, alone and in combination with
elevated C02, on tree leaf chemistry and insect herbivore performance: A meta-analysis. Global
Change Biol 13: 184-201. http://dx.doi.org/10.1111/i. 1365-2486.01284.X
Van Dingenen. R; Dentener. FJ; Raes. F; Krol. MC; Emberson. L; Cofala. J. (2009). The global
impact of ozone on agricultural crop yields under current and future air quality legislation. Atmos
Environ 43: 604-618. http://dx.doi.Org/10.1016/i.atmosenv.2008.10.033
van Goethem. TM; Azevedo. LB; van Zelm. R; Haves. F; Ashmore. MR; Huiibregts. MA. (2013).
Plant species sensitivity distributions for ozone exposure. Environ Pollut 178: 1-6.
http://dx.doi.Org/10.1016/i.envpol.2013.02.023
van Groenigen. KJ; Qi. X; Osenberg. CW; Luo. Y; Hungate. BA. (2014). Faster decomposition under
increased atmospheric C02 limits soil carbon storage. Science 344: 508-509.
http://dx.doi.org/10.1126/science. 1249534
Vanderhevden. D; Skellv. J; Innes. J; Hug. C; Zhang. J; Landolt. W; Bleuler. P. (2001). Ozone
exposure thresholds and foliar injury on forest plants in Switzerland. Environ Pollut 111: 321-331.
http://dx.doi.org/10.1016/50269-7491(00)00060-9
Vanloocke. A; Betzelberger. A: Ainsworth. EA; Bernacchi. CJ. (2012). Rising ozone concentrations
decrease soybean evapotranspiration and water use efficiency whilst increasing canopy
temperature. New Phytol 195: 164-171. http://dx.doi.Org/10.l 111/i. 1469-8137.2012.04152.x
Vigue. LM; Lindroth. RL. (2010). Effects of genotype, elevated C02 and elevated 03 on aspen
phytochemistry and aspen leaf beetle Chrysomela crotchi performance. Agr Forest Entomol 12:
267-276. http://dx.doi.org/10.1111/i. 1461-9563,2010.00475.x
Volin. JC; Reich. PB; Givnish. TJ. (1998). Elevated carbon dioxide ameliorates the effects of ozone on
photosynthesis and growth: species respond similarly regardless of photosynthetic pathway or plant
functional group. New Phytol 138: 315-325. http://dx.doi.Org/10.1046/i.1469-8137.1998.00100.x
Yolk. M; Bungener. P; Contat. F; Montani. M; Fuhrer. J. (2006). Grassland yield declined by a quarter
in 5 years of free-air ozone fumigation. Global Change Biol 12: 74-83.
http://dx.doi.Org/10.l 111/i. 1365-2486.2005.01083.x
Yolk. M; Geissmann. M; Blatter. A; Contat. F; Fuhrer. J. (2003). Design and performance of a free-air
exposure system to study long-term effects of ozone on grasslands. Atmos Environ 37: 1341-1350.
http://dx.doi.Org/10.1016/S 1352-2310(02)01012-9
Vuorinen. T; Nerg. AM; Holopainen. JK. (2004). Ozone exposure triggers the emission of herbivore-
induced plant volatiles, but does not disturb tritrophic signalling. Environ Pollut 131: 305-311.
http://dx.doi.Org/10.1016/i.envpol.2004.02.027
8-245
-------
Wagg. S; Mills. G; Haves. F; Wilkinson. S; Davies. WJ. (2013). Stomata are less responsive to
environmental stimuli in high background ozone in Dactylis glomerata and Ranunculus acris.
Environ Pollut 175: 82-91. http://dx.doi.Org/10.1016/i.envpol.2012.l 1.027
Wahid. A; Ahmad. SS; Butt. ZA; Ahmad. M. (2011). Exploring the hidden threat of gaseous
Pollutants using rice (Oryza sativa 1.) Plants in Pakistan. Pakistan J Bot 43: 365-382.
Wan. W; Manning. WJ; Wang. X: Zhang. H: Sun. X: Zhang. O. (2014). Ozone and ozone injury on
plants in and around Beijing, China. Environ Pollut 191: 215-222.
http://dx.doi.Org/10.1016/i.envpol.2014.02.035
Wan. W; Xia. Y: Zhang. H: Wang. J: Wang. X. (2013). [The ambient ozone pollution and foliar injury
of the sensitive woody plants in Beijing exurban region]. Acta Ecol Sin 33: 1098-1105.
Wang. B. in; Shugart. HH: Shuman. JK; Lerdau. MT. (2016). Forests and ozone: productivity, carbon
storage, and feedbacks. Sci Rep 6: 22133. http://dx.doi.org/10.1038/srep22133
Wang. D; Bormann. FH; Karnoskv. DF. (1986). Regional tree growth reductions due to ambient
ozone: Evidence from field experiments. Environ Sci Technol 20: 1122-1125.
http://dx.doi.org/10.1021/esQ0153a007
Wang. H; Zhang. L; Ma. X; Zou. J; Siemann. E. (2018). The effects of elevated ozone and C02 on
growth and defense of native, exotic and invader trees. Journal of Plant Ecology 11: 266-272.
http://dx.doi.org/10.1093/ipe/rtwl42
Wang. P; Baines. A; Lavine. M; Smith. G. (2012). Modelling ozone injury to US forests. Environ Ecol
Stat 19: 461-472. http://dx.doi.org/10.1007/slQ651-012-0195-2
Wang. P; Marsh. EL; Ainsworth. EA; Leakey. ADB; Sheflin. AM; Schachtman. DP. (2017). Shifts in
microbial communities in soil, rhizosphere and roots of two major crop systems under elevated
C02and 03. Sci Rep 7: 15019. http://dx.doi.org/10.1038/s41598-017-14936-2
Wang. S; Wang. F; Diao. X; He. L. (2014). Effects of elevated 03 on microbes in the rhizosphere of
mycorrhizal snap bean with different 03 sensitivity. Can J Microbiol 60: 93-103.
http://dx.doi.Org/10.l 139/cim-2013-0851
Wang. X; Taub. PR; Jablonski. LM. (2015). Reproductive allocation in plants as affected by elevated
carbon dioxide and other environmental changes: a synthesis using meta-analysis and graphical
vector analysis. Oecologia 177: 1075-1087. http://dx.doi.org/10.1007/sQ0442-014-3191-4
Warwick. KR; Taylor. G. (1995). Contrasting effects of tropospheric ozone on five native herbs which
coexist in calcareous grassland. Global Change Biol 1: 143-151. http://dx.doi.org/10.1111/i. 1365-
2486.1995 .tbOOO 14.x
Wedlich. KV; Rintoul. N; Peacock. S; Cape. JN; Covle. M; Toet. S; Barnes. J; Ashmore. M. (2012).
Effects of ozone on species composition in an upland grassland. Oecologia 168: 1137-1146.
http://dx.doi.org/10.1007/sQ0442-011-2154-2
Werner. H; Fabian. P. (2002). Free-air fumigation of mature trees: A novel system for controlled
ozone enrichment in grown-up beech and spruce canopies. Environ Sci Pollut Res Int9: 117-121.
http://dx.doi.org/10.1007/BF02987458
Westerling. AL; Bryant. BP; Preisler. HK; Holmes. TP; Hidalgo. HG; Das. T; Shrestha. SR. SR.
(2011). Climate change and growth scenarios for California wildfire. Clim Change 109: 445-463.
http://dx.doi.org/10.1007/slQ584-011-0329-9
Wieser. G; Manning. WJ; Tausz. M; Bvtnerowicz. A. (2006). Evidence for potential impacts of ozone
on Pinus cembra L. at mountain sites in Europe: An overview. Environ Pollut 139: 53-58.
http://dx.doi.Org/10.1016/i.envpol.2005.04.037
8-246
-------
Williams. W; Macgregor. N. (1975). Oxidant induced air pollution damage to forest trees in the
Southern Sierra-Sevada Moun-tains of California. Proc Am Phytopathol Soc 2: 120.
Williams. WT; Brady. M; Willison. SC. (1977). Air pollution damage to the forests of the Sierra
Nevada Mountains of California. J Air Pollut Control Assoc 27: 230-234.
http://dx.doi.org/10.1080/00022470.1977.1047Q415
Winner. WE: Lefohn. AS; Cotter. IS; Greitner. CS; Nellessen. J: Mcevov. LR. Jr: Olson. RL;
Atkinson. CJ; Moore. LP. (1989). Plant responses to elevational gradients of 03 exposures in
Virginia. Proc Natl Acad Sci USA 86: 8828-8832. http://dx.doi.org/10.1073/pnas.86.22.8828
Wiselogel. AE; Bailey. JK; Newton. RJ; Fong. F. (1991). Growth response of loblolly pine (Pinus
taeda L) seedlings to ozone fumigation. Environ Pollut 71: 43-56. http://dx.doi.org/10.1016/Q269-
7491(91)90043-V
Wittig. YE; 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: 115 0-1162.
http://dx.doi.org/10.1111/i. 1365-3040.2007.01717.X
Wittig. YE; Ainsworth. EA; Naidu. SL; Karnoskv. 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.org/10.1111/i. 1365-
2486.2008.01774.x
Wright. GA; Lutmerding. A; Dudareva. N; Smith. BH. (2005). Intensity and the ratios of compounds
in the scent of snapdragon flowers affect scent discrimination by honeybees (Apis mellifera). J
Comp Physiol A Neuroethol Sens Neural Behav Physiol 191: 105-114.
Wright. IJ; Reich. PB; Westobv. M; Ackerlv. DP; Baruch. Z; Bongers. F; Cavender-Bares. J; Chapin.
T; Cornelissen. JHC; Diemer. M; Flexas. J; Gamier. E; Groom. PK; Gulias. J; Hikosaka. K;
Lamont. BB; Lee. T; Lee. W; Lusk. C; Midglev. JJ; Navas. ML; Niinemets. U; Oleksvn. J; Osada.
N; Poorter. H; Poot. P; Prior. L; Pvankov. VI; Roumet. C; Thomas. SC; Tioelker. MG; Veneklaas.
EJ; Villar. R. (2004). The worldwide leaf economics spectrum. Nature 428: 821-827.
http://dx.doi.org/10.1038/nature02403
Wvness. K; Mills. G; Jones. L; Barnes. JD; Jones. PL. (2011). Enhanced nitrogen deposition
exacerbates the negative effect of increasing background ozone in Dactylis glomerata, but not
Ranunculus acris. Environ Pollut 159: 2493-2499. http://dx.doi.Org/10.1016/i.envpol.2011.06.022
Xu. S; Fu. W; He. X; Chen. W; Zhang. W; Li. B; Huang. Y. (2017). Drought alleviated the negative
effects of elevated 03 on Lonicera maackii in urban area. Bull Environ Contam Toxicol 99: 648-
653. http://dx.doi.org/10.1007/sQ0128-017-2179-2
Yang. N; Wang. X; Zheng. F; Chen. Y. (2017). The impact of elevated ozone on the ornamental
features of two flowering plants (Tagetes erecta linn. Petunia hybrida Vilm.). Int J Environ Pollut
61: 29-45. http://dx.doi.org/10.1504/IJEP.2017.100Q3698
Yendrek. CR; Koester. RP; Ainsworth. EA. (2015). A comparative analysis of transcriptomic,
biochemical, and physiological responses to elevated ozone identifies species-specific mechanisms
of resilience in legume crops. J Exp Bot 66: 7101-7112. http://dx.doi.org/10.1093/ixb/erv404
Yuan. JS; Himanen. SJ; Holopainen. JK; Chen. F; Stewart. CN. Jr. (2009). Smelling global climate
change: Mitigation of function for plant volatile organic compounds [Review]. Trends Ecol Evol
24: 323-331. http://dx.doi.Org/10.1016/i.tree.2009.01.012
8-247
-------
Yue. K; Peng. Y; Peng. C; Yang. W; Peng. X; Wu. F. (2016). Stimulation of terrestrial ecosystem
carbon storage by nitrogen addition: A meta-analysis. Sci Rep 6: 19895.
http://dx.doi.org/10.1038/srepl9895
Yue. X; Unger. N. (2014). Ozone vegetation damage effects on gross primary productivity in the
United States. Atmos Chem Phys 14: 9137-9153. http://dx.doi.org/10.5194/acp-14-9137-2014
Yun. SC; Laurence. JA. (1999). The response of clones of Populus tremuloides differing in sensitivity
to ozone in the field. New Phytol 141: 411-421. http://dx.doi.org/10.1046/i .1469-
8137.1999.00359.x
Yuska. DE: Skellv. JM: Ferdinand. JA: Stevenson. RE: Savage. JE: Mulki. J: Hines. A. (2003). Use of
bioindicators and passive sampling devices to evaluate ambient ozone concentrations in north
central Pennsylvania. Environ Pollut 125: 71-80. http://dx.doi.org/10.1016/S0269-7491(03)00096-
4
Zak. PR: Kubiske. ME: Pregitzer. KS: Burton. AJ. (2012). Atmospheric C02 and 03 alter
competition for soil nitrogen in developing forests. Global Change Biol 18: 1480-1488.
http://dx.doi.org/10.1111/i. 1365-2486.2011.02596.x
Zak. PR: Pregitzer. KS: Kubiske. ME: Burton. AJ. (2011). Forest productivity under elevated C02
and 03: positive feedbacks to soil N cycling sustain decade-long net primary productivity
enhancement by C02. Ecol Lett 14: 1220-1226. http://dx.doi.Org/10.l 111/i.1461-
0248.2011.01692.x
Zhang. J: Tang. H: Zhu. J: Lin. X: Feng. Y. (2016). Pivergent responses of methanogenic archaeal
communities in two rice cultivars to elevated ground-level 03. Environ Pollut 213: 127-134.
http://dx.doi.Org/10.1016/i.envpol.2016.01.062
Zheng. F: Wang. X: Lu. F: Hou. P: Zhang. W: Puan. X: Zhou. X: Ai. Y: Zheng. H: Ouvang. Z: Feng.
(2011). Effects of elevated ozone concentration on methane emission from a rice paddy in
Yangtze River Pelta, China. Global Change Biol 17: 898-910. http://dx.doi.0rg/lO.l 111/i. 1365-
2486.2010.02258.x
8-248
-------
APPENDIX 9 THE ROLE OF TROPOSPHERIC
OZONE ON CLIMATE EFFECTS
Summary of C ansality Determinations
keeeul e\ idenee eoulMiues lo support ;i e;ius;il relationship between tropospherie
o/onc ;iik.I r;idi;iti\e I'oreiim. ;ind ;i likelv to he e;ius;il relationship. \ i;i r;idi;iti\e I'oreiim.
hem con tropospherie oA>ue ;md temperature. preeipit;iliou. ;md related I ' (l/one IS \ None of llie new siudies
support ;i elimme to either e;iiis;iln> delerniiii;iiiou ineluded in the 2<> I i ()A»ne IS \ More
detmls on the e;ius;il I'rmnework used to re;ieh these eoiielusious ;ire ineluded in the I'remnhle to
the IS \ il S. \ .\> V 21»15)
1". ITiils
Ki-I;iliiinshi|>
I ropusplu rii ii/iiiu- ;iml iliiiiini-
Radialive forcing
Causal
leniperaliire. precipitation. and related
9.1 Introduction
9.1.1 Summary from the 2013 Ozone ISA
Changes in the abundance of tropospheric ozone perturb the radiative balance of the atmosphere
by interacting with incoming solar radiation and outgoing longwave radiation. This effect is quantified by
the radiative forcing (RF) metric. Through this effect on the Earth's radiation balance, tropospheric ozone
plays a major role in the climate system, and increases in ozone abundance contribute to climate change
(Forster et al.. 2007V
• Increases in tropospheric ozone are tied to the rise in emissions of ozone precursors from human
activity, mainly from fossil fuel consumption and agricultural processes. Models estimate that the
global average tropospheric ozone concentration has increased 30-70% since preindustrial times
(Gauss et al.. 2006). and observations indicate that tropospheric ozone concentrations may have
increased by factors of 4 or 5 in some regions (Marcnco et al.. 1994; Staehelin et al.. 1994V In the
21st century, as the Earth's population continues to grow and energy technology spreads to
developing countries, a further rise in the global concentration of tropospheric ozone is likely
(Forster et al.. 2007V
9-1
-------
• The 2007 IPCC report estimated RF from tropospheric ozone since preindustrial times (1750 to
2005) to be 0.35 W/m2 (avg) from multimodel studies, with 95th percentile error bars ranging
from 0.25 to 0.65 W/m2 (Forster et al.. 2007).
• The transport of ozone to the Arctic from the midlatitudes leads to RF estimates greater than
1.0 W/m2 in this region, especially in summer (Shindcll et al.. 2006; Liao et al.. 2004; Micklev et
al.. 1999V
• Based on RF, increasing tropospheric ozone concentration since preindustrial times is estimated
to contribute about 25-40% of the total warming effects of anthropogenic carbon dioxide (CO2)
and about 75% of the effects of anthropogenic methane [CH4; Forster et al. (2007)1. ranking
ozone third in importance for affecting global climate behind these two major greenhouse gases.
• The Earth's land-atmosphere-ocean system responds to the RF with a change in climate, typically
expressed as a change in near-surface air temperature (Forster et al.. 2007). The effect of
tropospheric ozone on global surface temperature is approximately proportional to the change in
tropospheric ozone concentration. The Earth's surface temperatures are most sensitive to ozone
abundance perturbations in the upper troposphere.
• The increase of tropospheric ozone abundance has contributed an estimated 0.1-0.3 °C warming
to the global climate since 1750 (Hansen et al.. 2005; Micklev et al.. 2004).
• Tropospheric ozone also has the potential to contribute to UV-B shielding, with further
implications for human health and ecosystem effects. Almost no studies were found in the 2013
Ozone ISA examining the incremental effects of changes in tropospheric ozone concentrations on
UV-B—and, of those, such effects of tropospheric ozone concentrations on UV-B were found to
be small (Madronich et al.. 201IV No studies were found that adequately examined the
incremental health or welfare effects attributable specifically to changes in UV-B exposure
resulting from perturbations in tropospheric ozone concentrations.
9.1.2 Scope for the Current Review
The scope of this section is defined by a scoping tool that generally describes the relevant
Population, Exposure, Comparison, Outcome, and Study Design (PECOS). The PECOS tool defines the
parameters and provides a framework to help identify the relevant literature to inform the current Ozone
ISA. This review builds on the findings from the 2013 Ozone ISA and draws on peer-reviewed research,
including the Intergovernmental Panel on Climate Change (IPCC) Fifth Assessment Report (AR5)
(Mvhre et al.. 2013). to integrate evidence on how changing tropospheric ozone concentrations might
affect climate. None of the new studies support a change to either climate-related causality determination
included in the 2013 Ozone ISA. The studies evaluated and subsequently discussed within this section
were included if they satisfied all of the components of the following PECOS tool summarized in
Table 9-1.
9-2
-------
Table 9-1
Population, exposure, comparison, outcome, and study design
(PECOS) tool for radiative forcing and climate change.
Radiative Forcing
Temperature, Precipitation, and
Related Climate Variables
Population/geographical
scope (P)
Regional, continental, and/or global
scale
Regional, continental, and/or global scale
Exposure/environmental
variable (E)
Tropospheric ozone concentration
distributions in 3D (observed/modeled)
Tropospheric ozone concentration
distributions in 3D (observed/modeled)
Comparison (C)
Relevant baseline or nonperturbed
scenarios/conditions
Relevant baseline or nonperturbed
scenarios/conditions
Outcome (O)
Changes in RF resulting from change in
tropospheric ozone
Subsequent climate effects (via RF)
(e.g., global surface temperature)
resulting from change in tropospheric
ozone
Study design (S)
Observations or modeling studies
Observations or modeling studies
The following sections of this Appendix provide a brief background on the Earth's climate
system and the pathways through which ozone influences it (Section 9.1.3) and on the new scientific
evidence contributing to the causality determinations for RF (Section 9.2) and temperature, precipitation,
and related climate variables (Section 9.3). In addition, this Appendix provides a brief background on the
issue of UV-B shielding and reports on the results of a literature screening carried out to determine if any
new evidence on incremental effects of tropospheric ozone concentrations on UV-B has emerged since
the 2013 Ozone ISA (Section 9.1.3.4).
9.1.3 Introduction to Climate, Ozone Chemistry, and Radiative Forcing
9.1.3.1 Climate Change
Comprehensively assessing the role of anthropogenic activity in past and future climate change,
including the influence of changing tropospheric concentrations of ozone, is the mandate of the IPCC, an
initiative begun in 1988 by the World Meteorological Organization (WMO) and the United Nations
Environment Programme (UNEP). The IPCC supports the work of the Conference of Parties to the
United Nations Framework Convention on Climate Change (UNFCCC). New IPCC reports are issued
every 5 to 7 years, and the climate discussion in the 2013 Ozone ISA relied heavily on the IPCC Fourth
Assessment Report (AR4), published in 2007. The next iteration, AR5 (Mvhre et al.. 2013). reports on the
9-3
-------
key scientific advances in understanding the climate effects of ozone since AR4. This Appendix draws
substantially upon AR5 in summarizing these effects, in particular Chapter 8 of AR5, on RF (Mvhrc ct
al.. 2013). as well as more recent literature published subsequent to AR5.
Human activity has led to observable increases of greenhouse gases (GHGs) in the atmosphere
since the start of the Industrial Revolution, mainly through fossil fuel combustion. Over the last century,
global-average surface air temperature has increased by approximately 1.0°C, and emissions of
greenhouse gases are the main cause (Wucbblcs et al.. 2017; IPC'C. 2013). There are many other aspects
of the global climate system that are changing in addition to this warming, including melting glaciers,
reductions in snow cover and sea ice, sea level rise, ocean acidification, and increases in the frequency or
intensity of many types of extreme weather events (Wucbblcs et al.. 2017). The magnitude of future
climate change, globally and regionally, and in terms of both temperature increases and these other types
of associated effects, will depend primarily on the amount of greenhouse gases emitted globally
fWuebbles et al.. 2017; IPCC. 2013).
9.1.3.2 Ozone Chemistry and Role in Climate
The atmosphere can be divided into several distinct vertical layers, based primarily on the major
mechanisms by which they are heated and cooled. The lowest major layer is the troposphere, which
extends from the earth's surface to about 8 km above polar regions and to about 16 km above tropical
regions. Lying above the troposphere is the stratosphere, which extends from the top of the troposphere to
an altitude of about 50 km. The boundary between the troposphere and stratosphere is known as the
tropopause.
As with the 2013 Ozone ISA, the emphasis in this Appendix is on the climate effects of
"tropospheric ozone," which refers broadly to ozone occurring throughout the total depth of the
troposphere, consistent with the usage in the primary scientific literature on ozone and climate. Within the
troposphere, the lowest sublayer is the planetary boundary layer, extending from the surface to about
1-2 km, which is most strongly affected by surface conditions, including local emissions of ozone
precursors. The portion of the troposphere lying above the planetary boundary layer, in which
atmospheric transport processes occur over much larger spatial scales, is often referred to as the "free
troposphere," with the "upper troposphere" referring to the high-altitude portion of the free troposphere
nearest the tropopause.
The 2013 Ozone ISA described how tropospheric ozone differs in important ways from other
GHGs, in that it is not emitted directly, but is produced through photochemical oxidation of carbon
monoxide (CO), CH4, and nonmethane volatile organic compounds (VOCs) in the presence of nitrogen
oxides (NOx = NO + NO2). It is also supplied by vertical transport from the stratosphere, where it is
formed naturally by other chemical processes. In addition, because the lifetime of ozone in the
troposphere is typically a few weeks, it is not distributed uniformly like the well-mixed GHGs (e.g., CO2
9-4
-------
and CH4), but instead has an inhomogeneous distribution that also varies seasonally. See Appendix 1 for
additional context.
Ozone influences the Earth's radiation budget by its longwave absorption, mainly in the
9.6 micron band, where absorption by the long-lived GHGs and water vapor is relatively weak. In
addition, unlike other major GHGs, ozone absorbs in the shortwave as well as in the longwave part of the
spectrum.
Figure 9-1 depicts the influence of tropospheric ozone on climate. Global emissions of ozone
precursors including carbon monoxide (CO), volatile organic compounds (VOCs), methane (CH4), and
oxides of nitrogen (NOx), of which U.S. emissions are a subset, lead to the production of tropospheric
ozone globally. A change in the abundance of tropospheric ozone perturbs the radiative balance of the
atmosphere, an effect quantified by RF (defined in more detail in the next subsection). As discussed in the
2013 Ozone ISA, the Earth's land-atmosphere-ocean system responds to this RF with a change in climate,
including a change in near-surface air temperature with associated effects on precipitation and
atmospheric circulation patterns. This climate response causes downstream climate-related health and
ecosystem effects. Feedbacks from both the direct climate response and such downstream effects can, in
turn, affect the abundance of tropospheric ozone and ozone precursors through multiple mechanisms. This
Appendix provides a brief discussion of some of the most direct feedbacks, but the downstream effects
and their longer term feedbacks can be extremely complex and outside the scope of this assessment.
9-5
-------
, t
I Climate Effects *
I on Human Health r
and Ecosystems )
Changes in Tropospheric
O, Abundance
L (I®) .
Radiative Forcing
Due to O, Change
(W/m2)
Climate Response
(°C)
Precursor Emissions of
CO, VOCs, CH4, NOx
(Tg/y)
Note: Climate effects and their feedbacks are de-emphasized in this figure since these downstream effects are extremely complex
and outside the scope of this assessment.
Figure 9-1 Schematic diagram of the effects of tropospheric ozone on
climate.
9.1.3.3 Radiative Forcing
RF is a perturbation in net radiative flux at the tropopause (or top of the atmosphere) caused by a
change in radiatively active forcing agent(s) after stratospheric temperatures have readjusted to radiative
equilibrium [stratospherically adjusted RF; Fiore et al. (2015); Mvhre et al. (2013)1. It is commonly
expressed in units of W/m2. All else being equal, a positive RF results in net warming of the Earth's
surface, while negative RF leads to a net cooling. Effective radiative forcing (ERF) accounts for both the
effects of the forcing agent and the rapid adjustments to that forcing agent [e.g., atmospheric temperature,
cloud cover, water vapor; Mvhre et al. (2013)1. While global mean RF and ERF are both important
measures of climate response to radiative effects, this assessment focuses on RF as an endpoint, because
ozone ERF estimates in the published literature are more limited, and because differences between RF and
ERF for tropospheric ozone tend to be small compared to existing uncertainties in RF and ERF (Mvhre et
al.. 2013). The nonuniform distribution of ozone (horizontally, vertically, and temporally) also makes
quantifying global and regional ozone RF challenging. Unlike RF estimates for well-mixed GHGs, which
can be and are determined from observations at a few surface sites (Mvhre et al.. 2013). ozone RF
9-6
-------
estimates are generally calculated using a combination of general circulation model (GCM) radiative
transfer parameterization schemes and more detailed line-by-line radiative transfer models. As a result, it
is more difficult to provide precise quantitative estimates of the RF associated with changes in surface
ozone at regional scales. As noted above, tropospheric ozone RF is estimated to be about 25-40% of the
total warming effects of anthropogenic carbon dioxide (CO2) and about 75% of the effects of
anthropogenic methane (CH4), globally, ranking ozone third in importance for global climate behind these
two major greenhouse gases (Forster et al.. 2007).
9.1.3.4 Tropospheric Ozone and Ultraviolet-B (UV-B) Shielding
UV radiation from the sun contains sufficient energy when it reaches the Earth to photolyze
chemical bonds in molecules, thereby leading to damaging effects on living organisms and materials. It is
well understood that stratospheric ozone plays a crucial role in reducing exposure to solar UV radiation at
the Earth's surface. The question of whether tropospheric ozone has a supplemental UV-B shielding
effect with significance for human health and ecosystems was considered in the 2013 Ozone ISA (as part
of Chapter 10, in addition to the discussion of climate effects). The 2013 Ozone ISA concluded:
"EPA has found no published studies that adequately examine the incremental health or
welfare effects (adverse or beneficial) attributable specifically to changes in UV-B
exposure resulting from perturbations in tropospheric O3 concentrations. While the
effects are expected to be small, they cannot yet be critically assessed within reasonable
uncertainty. Overall, the evidence is inadequate to determine if a causal relationship
exists between changes in tropospheric O3 concentrations and effects on health and
welfare related to UV-B shielding" [pg. 10-32],
For the current review, a literature screening on tropospheric ozone and UV-B was conducted.
This screening determined that there was no new evidence since the 2013 Ozone ISA relevant to the
question of UV-B shielding by tropospheric ozone, including the incremental effects of tropospheric
ozone concentration changes on UV-B. While the literature screening identified a small number of studies
that examined tropospheric ozone and UV-B together in some way, these studies addressed different
scientific and technical issues, such as the impact of UV-B on tropospheric ozone production
[e.g., Jasaitis et al. (2016)1. or the development of improved model parameterizations for better capturing
the influence of total column ozone, in combination with other input parameters, on surface UV radiation
[e.g., Lamv et al. (2018)1.
9-7
-------
9.2
Ozone Impacts on Radiative Forcing
lh\'hlights of Recent lividence for limitative I 'orving
( linimcs mi ilie ;ihiiiid;incc nf imposphcric i»a»iic ;il lecl kT I lie 2t> P ( )a>iic IS \
rcpi>ris ;i kl'iil'n '5 \V in I'mni imposphcric i>a>iic since prciiidiisiri;il limes (I ~5o in 2iki5i
h:ised on iniilnmodel studies i loisicr el ;il. 2uii~i The nmsi reeeni IP( ( ;isscssnicni. \k.\
reports iroposphcric oa»iic kl' ;is t> 4t> (<> ,.2<> lotUiU) W in iM\ lire el ;il . 2nI 'i. which is
w 11 In n I lie i" i line nl' pre\ ions ;issessnienis 11 e.. Ak4i ' There h;i\ e ;ilso been ;i lew studies since
\k5. iiicludiim i he studs of iroposphcric oa»iic kl' h;ised on I lie Coupled Model
hilercnnipiirisnii hojccl Mi;ise (> (( \ 111 > i d;il;i sel ;md I he \tniosphcric ( heniisir\ ;md ( hni;ile
Mudel hilercnnipiirisnii I'rnjecl ( \( (Mll'i niiiliiniKdel siud> nT irDpnspheric chemisir\. hnili
of w Inch reinforce I he \k5 csiininlcs ;md ilie c;ins;il relnliDiiship helween u\>pi/i>ne
;md kl'
9.2.1 Recent Evidence for Historical Period
This section summarizes the new scientific evidence contributing to the causality determination
for RF. The new evidence comes from the IPCC AR5 (Mvhrc et al.. 2013) and its supporting references,
as well as a few additional, more recent studies. It builds on evidence presented in the 2013 Ozone ISA.
• The AR5 best estimate of tropospheric ozone RF is 0.40 (0.20 to 0.60) W/m2 [from 1750 to 2011;
Table 9-2; Shindell et al. (2013); Sovde et al. (2012); Skeie et al. (2011)1. There are uncertainties
from inter-model spread across atmospheric models (-0.11 to 0.11 W/m2) and differences
between standalone radiative transfer models (-0.07 to 0.07 W/m2), where all ranges represent
the 95% confidence intervals (Mvhre et al.. 2013). One of the largest uncertainties in calculating
ozone RF is estimating ozone concentrations in preindustrial times. Trends in free tropospheric
ozone and upper tropospheric ozone (where RF is particularly sensitive to changes in ozone) are
not captured well by models (Hu et al.. 2017; Sherwen et al.. 2017; Parrish et al.. 2014).
Uncertainties also remain in preindustrial emissions and the representation of chemical and
physical processes beyond those already included in the current models. These additional model
uncertainties include those associated with specific rate constants for important reactions
[e.g., NO2 + OH^HNOs and O3 + NO^N02 + O2; Newsome and Evans (2017)1 and halogen
chemistry (Sherwen et al.. 2017). Despite this, the IPCC AR5 overall confidence in the
tropospheric ozone RF is high (Table 9-3. Figure 9-2). AR5 concluded there were no major
changes in understanding of this confidence level since AR4.
• Additionally, there have been a few studies since AR5, including the study of tropospheric ozone
RF based on the Coupled Model Intercomparison Project Phase 6 (CMIP6) data set (Checa-
Garcia et al.. 2018) and the Atmospheric Chemistry and Climate Model Intercomparison Project
(ACCMIP) multimodel study of tropospheric chemistry (Conlev et al.. 2013; Stevenson et al..
2013). both of which reinforce the AR5 estimates. The latest individual estimates of tropospheric
ozone RF, based on the CMIP6 data set for ozone, give a present-day (2000-2014 relative to
1850-1860) tropospheric ozone RF of 0.33 ± 0.17 W/m2 (Checa-Garcia et al.. 2018).Mvhre et al.
(2017) also recently estimated ozone RF for the 1990-2015 time period with a multimodel mean
of 0.06 W/m2, which is -50% greater than the AR5 estimate for this same time period. The
9-8
-------
difference is likely due to an increase in NOx [in Mvhre et al. (2017)1 that is more than twice that
in the AR5 emission data.
Table 9-2 Contributions of tropospheric ozone changes to radiative forcing
(W/m2) from 1750 to 2011.a
Troposphere
Total
AR5a
0.40
(0.20 to 0.60)
ACCMIP (multimodel results)15
0.41
(0.21 to 0.61)
Shindell et al. (20131
0.33
(0.31 to 0.35)
Sovde etal. (20121c
0.45
0.38
Skeieet al. (20111
0.41
(0.21 to 0.61)
AR4C
0.35
(0.25 to 0.65)
aTable 9-2 is adapted from IPCC AR5 Table 8.3.
bStevenson et al. (20131.
°0.45 based on REF chemistry, 0.38 based on R2 chemistry, see Sovde et al. (20121.
dForster et al. (20071.
Source: Mvhre et al. (20131.
Table 9-3 Confidence level for ozone forcing for the 1750-2011 period.3
Basis for Uncertainty
Confidence Estimates (More Change in Understanding
Evidence Agreement Level Certain/Less Certain) Since AR4
Tropospheric Robust Medium High Observed trends of No major change
ozone ozone in the troposphere
and model results for the
industrial era/differences
between model estimates
of RF
aTable 9-3 is an abbreviated version of IPCC AR5 Table 8.5.
Source: Mvhre et al. (20131.
9-9
-------
Radiative forcing of climate between 1750 and 2011
Forcing agent
Well Mixed
Greenhouse Gases
Ozone
Stratospheric water
vapour from CH4
Surface Albedo
Contrails
Aerosol-Radiation Interac.
Aerosol-Cloud Interac.
CO,
Other WMGHG
h
CH,
E9
Halocarbons
I
Stratospheric \ > -| ¦ ~ I Tropospheric
iW
Land Use | -| W Black carbon
on snow
Contrail induced cirrus
Total anthropogenic
Solar irradiance
•~I
1
Radiative Forcing (W rrr
Reproduced from IPCC AR5, Mvhre et al. (20131.
Note: Uncertainties (5 to 95% confidence range) are given for RF (dotted lines) and ERF (solid lines). For ozone, differences
between RF and ERF tend to be small compared to existing uncertainties in RF and ERF, so only RF is reported
Figure 9-2 Bar chart for radiative forcing (RF; hatched) and effective
radiative forcing {ERF; solid) for the period 1750-2011.
• Ozone-depleting substances (ODSs) also affect tropospheric ozone. Recent simulations with the
NCAR-CAM3.5 model show that ODSs contribute an ozone RF of-0.15 (-0.3 to 0.0) W/m2,.
some of which is in the troposphere, and tropospheric ozone precursors contribute an ozone RF of
0.50 (0.30 to 0.70) W/m2, some of which is in the stratosphere (Mvhre et al.. 2013; Shindell et al..
2013). See Figure 9-3.
• Instantaneous radiative kernels (IRKs), which represent the sensitivity of top-of-the-atmosphere
(TOA) radiative flux to each observed (satellite) ozone profile, have been used to estimate all-sky
global average TOA longwave radiative effect (LWRE) of tropospheric ozone as
0.33 ± 0.02 W/m2 (Worden et al.. 2011). The LWRE due to ozone is computed with respect to the
TOA radiative flux as observed from space by the Aura satellite's Tropospheric Emission
Spectrometer (TES) instrument; this is distinguished from the IPCC AR5 RF definition, which
represents the difference in total irradiance at the tropopause due to changes between
preindustrial and present tropospheric ozone concentrations. More recently. Bowman et al. (2013)
applied IRKs and the ACCMIP model results to estimate a multimodel mean tropospheric ozone
RF of 0.39 ± 0.042 W/m2 (one standard deviation).
9-10
-------
• Tropospheric ozone concentrations and RF are also sensitive to changes in ozone precursor
emissions, which can alter the radiative balance of the atmosphere—sometimes in competing
ways. Ozone and its precursors exert a strong control on the oxidizing capacity of the
troposphere, thereby affecting the lifetime of CH4 and other gases (Dcrwcnt et al.. 2001). with
further implications for RF and climate. CO and VOC emissions exert an overall positive RF
(warming) by increasing tropospheric ozone and CH4 concentrations. NOx emissions contribute a
positive RF by increasing tropospheric ozone, but exert a negative RF by lowering global CH4
(via hydroxyl radical [OH] increases) and increasing nitrate aerosols. VOC emissions also
contribute to organic particulate matter production, which influences RF. Methane is itself a
powerful greenhouse gas and a precursor to ozone, leading to further warming (Fiorc et al..
2015). Short-lived ozone precursors (CO, VOCs, and NOx) influence ozone RF globally and
regionally (Fry et al.. 2012) and additionally can affect its seasonality (Bellouin et al.. 2016).
Figure 9-3. from the IPCC AR5, summarizes the magnitude of direct RF over the industrial era
associated with ozone precursor species themselves, as well as their indirect RF due to their
influence on ozone concentrations (Mvhrc et al.. 2013V
9-11
-------
Components of Radiative Forcing
o
x
o
!J
a>
X
tft
CO
o
o
~o
>
o
J=
00
o
tf>
w
3
a>
¦_
CL
"O
C
o
o
CO
O
0>
<
at
_c
CH4
HaloCarbons
H20(Strot.)
HFCs-PFCs—SF
NMVOC
trote
| Nitrate
Sulphate
Sulphate
BC on
snow
Fossil and
Burning
ERFaci
Contrails
i.2«r
Surface
Albedo
-0,5 0,0 0,5 1.0
Radiative Forcing (W m"2)
^ Black Carbon
Organic Carbon
Mineral Dust
Aerosol-Cloud
Aircraft
Land Use
Solar Irradiance
1-5
Reproduced from IPCC AR5, Mvhre et al. (2013).
Figure 9-3 Radiative forcing (RF) over the industrial era associated with
emitted compounds, including ozone (green bars) and its
precursors.
9-12
-------
9.2.2
Recent Evidence of Radiative Forcing Temporal and Spatial Trends
This section continues the summary of new scientific evidence contributing to the causality
determination for RF, with a focus on temporal and spatial trends in RF. As with the previous section, the
new evidence comes primarily from the IPCC AR5 (Mvhrc et al.. 2013) and its supporting references,
supported by the findings from a few additional, more recent studies. This new evidence builds on the
evidence presented in the 2013 Ozone ISA.
• Recent evidence from AR5 also provides the temporal trends of ozone RF. Tropospheric ozone
RF increased slowly before 1950, grew rapidly from 1950 to 1990, and increased slowly again
after 1990, matching the trends in anthropogenic ozone precursor emissions (Figure 9-4).
Changes in NOx emissions related to traffic and industry, for example, contributed to increasing
ozone RF trends over recent decades (Dahlmann et al.. 2011). In general, changes in NOx, CO,
and VOC emissions affect tropospheric ozone on short time scales (days to months), while
CFLrinduced changes in tropospheric ozone (via CH4 oxidation) occur on decadal timescales
(Fiore et al.. 2015).
0,6
—1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 r
. Tropospheric
Total
Stratospheric
-0,2
—I I I I I I I I I I I I I L
1750
1800
1B50
1900
1950
2000
Reproduced from IPCC AR5, Mvhre et al. (20131.
Note: Stratospheric ozone radiative forcing is also shown but is not discussed as part of this Appendix.
Figure 9-4 Time evolution of the radiative forcing (RF) from tropospheric
ozone from 1750 to 2010.
9-13
-------
• 1PCC has also examined future RF trends using multimodel estimates, with the magnitude of
future ozone RF effects consistent with our general understanding of the relationship between RF
and ozone concentrations described above. As shown in AR5, future ozone RF projections are
contingent on the specific emissions scenarios chosen for the model simulations (Mvhrc et al..
2013).
• The spatial distribution of net ozone RF (1850 to 2000) is depicted in Figure 9-5 based on the
ACCMIP models, with the greatest ozone RF occurring in subtropical latitudes. The positive
tropospheric ozone forcing in the Northern Hemisphere is associated with increases in
tropospheric ozone, while the negative forcing in the Southern Hemisphere's polar region is
related to stratospheric ozone loss. The ACCMIP models also show the largest standard deviation
in the polar regions, where lower stratosphere/upper troposphere changes vary among models
(Young et al.. 2013). Shifts m global tropospheric ozone concentrations may be driven most
strongly by the spatial distribution of anthropogenic emissions (1980 to 2010), compared to
changes in the overall magnitude of emissions and global CFL concentrations (Zhang et al..
2016).
y+rtfttt
WSt -
Preindustriai to Present-Day Forcing
Multi-model mean Standard deviation
0.29W m3
-.88 -.62 -.38 -.12 .12 .38 .62 .88
Reproduced from IPCC AR5, Mvhre et al. (20131.
Note: Values are the average of the area-weighted global means (11 models), with the area-weighted mean of the standard
deviation of models at each point provided in parenthesis.
Figure 9-5 Radiative forcing (RF) spatial distribution of 1850 to 2000 ozone
RF among the atmospheric chemistry and Climate Model
Intercomparison Project models, mean values (left) and standard
deviation (right).
• AR5 rates the confidence in the spatial distribution of ozone RF as medium (lower than global
mean ozone RF) because of the large regional standard deviations between models.
9-14
-------
• To address present-day model biases and improve spatial distribution estimates, modeled ozone
RF distributions can be adjusted by vertical information retrieved from the Tropospheric
Emission Spectrometer [TES; Shindell et al. (2013)1. For example, Figure 9-6 shows the
difference in the annual average RF between modeled (GISS-E2-R) and observed TES
present-day (2005-2009) total natural plus anthropogenic ozone throughout the atmosphere. The
global mean tropospheric ozone RF difference is 0.016 W/m2. Shindell et al. (2013 found good
agreement between the modeled and observed global mean RFs, in part due to a cancellation of
positive biases in the Northern Hemisphere and negative biases in the Southern Hemisphere.
i i i i i i i i i i i i
-0.54 -0.42 -0.30 -0.1 B -0.06 0.06 0.1 B 0.30 0.42 0.54
Reprinted with permission of publisher, Shindell et al. (2013).
Figure 9-6 Difference in annual average radiative forcing (W/m2) between
modeled (GISS-E2-R) and observed Tropospheric Emission
Spectrometer present-day (2005-2009) total natural plus
anthropogenic ozone throughout the atmosphere.
9.2.3 Summary and Causality Determination
Recent evidence continues to support a causal relationship between tropospheric ozone and
RF as concluded in the 2013 Ozone ISA (Table 9-4). The new evidence comes from the IPCC AR5
(Mvhre et al.. 2013) and its supporting references—in addition to a few more recent studies—and builds
on evidence presented in the 2013 Ozone ISA. None of the new studies support a change to the causality
determination included in the 2013 Ozone ISA.
9-15
-------
The AR5 best estimate of tropospheric ozone RF is 0.40 (0.20 to 0.60) W/m2 (from 1750 to 2011)
ITable 9-2; Mvhre et al. (2013)1. The overall confidence in the tropospheric ozone RF is high
(Table 9-3. Figure 9-2). Additionally, there have been a few studies since AR5, including the
study of tropospheric ozone RF based on the CMIP6 data set (Checa-Garcia et al.. 2018) and the
ACCMIP multimodel study of tropospheric chemistry (Conlev et al.. 2013; Stevenson et al..
2013). both of which reinforce the AR5 estimates.
Tropospheric ozone concentrations and RF are sensitive to changes in ozone precursor emissions.
Ozone precursors themselves exert a strong control on the oxidizing capacity of the troposphere
and can alter the radiative balance of the atmosphere—sometimes in competing ways. Figure 9-3
from the IPCC AR5 summarizes the magnitude of RF over the industrial era associated with
ozone precursor species (Mvhre et al.. 2013).
Evidence from AR5 and more recent literature addresses temporal trends and spatial patterns of
ozone RF. Tropospheric ozone RF increased slowly before 1950, grew rapidly from 1950 to
1990, and increased slowly again after 1990, matching the trends in anthropogenic ozone
precursor emissions (Figure 9-4). Spatially, ozone RF is estimated to be highest in the Northern
Hemisphere, in association with spatial emissions patterns.
One of the largest contributors to uncertainty in ozone RF is estimating preindustrial ozone
concentrations. In addition, trends in free tropospheric ozone and upper tropospheric ozone
(where RF is particularly sensitive to changes in ozone) are not captured well by models.
Uncertainties also remain in preindustrial emissions and the representation of chemical and
physical processes beyond those included in the current models, such as specific rate constants
and halogen chemistry. These uncertainties make it difficult to provide precise quantitative
estimates of the RF associated with changes in surface ozone at regional scales (e.g., how does
U.S. surface ozone affect climate forcing). Despite these uncertainties, the overall confidence in
current estimates of tropospheric ozone RF at the global level remains high (Table 9-3).
Further research in a number of areas can help address remaining uncertainties in the relationship
between RF and tropospheric ozone. These areas include: improving the quantification of
observed trends in ozone concentrations in the free troposphere, upper troposphere, and remote
regions; increasing understanding of the CFU budget and of ozone coupling with temperature,
water vapor, and clouds (with implications for the height and latitude dependence of ozone RF);
and improving estimates of ozone spatiotemporal structure developed using global models
constrained by observations. Additional areas of potential future research related to tropospheric
ozone and climate effects are described at the end of Section 9.3.3.
9-16
-------
Table 9-4 Summary of evidence for a causal relationship between tropospheric
ozone and radiative forcing.
Rationale for Causality
Determination
Key Evidence
Key References
Consistent evidence from
multiple, high-quality studies
Multidecadal, global chemistry-climate
modeling ensemble studies constrained
by historical observations of ozone
concentrations (e.g., IPCCAR5;
ACCMIP; CMIP6)
Mvhre et al. (2013); Section 9.2.1
Robust physical
understanding
Robust, well-understood relationship
between tropospheric ozone
concentration and RF
Mvhre et al. (2013); Section 9.2.1;
Section 9.1.3.3
Spatial/temporal effect Temporal trends in ozone RF match the Mvhre et al. (2013): Section 9.2.2
correspondence historical trends in anthropogenic ozone
precursor emissions; spatially, largest
effects seen in the subtropics, again
consistent with observed emissions
patterns
9.3 Ozone Impacts on Temperature, Precipitation, and Related
Climate Variables
Highlights of Recent Evidence for Impacts on Temperature, Precipitation, and
Related Climate Variables
Consistent with previous estimates, the effect of tropospheric ozone on global surface
temperature, through its impact on RF, continues to be estimated at roughly 0.1-0.3°C since
preindustrial times (Xie et al„ 2016: Mvhre et al.. 2013). with larger effects in some regions. In
addition to temperature, tropospheric ozone changes have impacts on other climate metrics
such as precipitation and atmospheric circulation patterns (Macintosh et al„ 2016: Allen et al..
2012: Shindell et al.. 2012a). While the wanning effect of tropospheric ozone in the climate
system is well established in general, precisely quantifying changes in surface temperature due
to tropospheric ozone changes, along with related climate effects, requires complex climate
simulations, including important feedbacks and interactions. Current limitations in climate
modeling tools, variation across models, and the need for more comprehensive observational
data on these effects represent sources of uncertainty in quantifying the precise magnitude of
climate responses to ozone changes, particularly at regional scales (Mvhre et al„ 2013). All of
this evidence reinforces the likely to be causal relationship between tropospheric ozone and
temperature, precipitation, and related climate variables (referred to as "climate change" in the
2013 Ozone ISA).
9-17
-------
9.3.1
Recent Evidence for Effects on Temperature
As described above, estimates of RF provide a first-order metric of the effect of tropospheric
ozone on climate. The Earth's land-atmosphere-ocean system then responds to this RF with a change in
climate, beginning with changes in temperature, followed by "downstream" impacts on other climate
variables, including changes in precipitation and shifts in atmospheric circulation patterns. Even further
downstream effects resulting from these changes in climatic conditions occur as well, for example, on
ecosystems, and may in turn create complex climate system feedbacks, including those with potential
impacts on tropospheric ozone concentrations, as summarized in the 2013 Ozone ISA (Figure 9-IV A
comprehensive accounting of all possible Earth system impacts of ozone-induced climate change and
associated potential feedback loops, however, is beyond the scope of this review.
• Literature cited in IPCC AR4, and referenced in the 2013 Ozone ISA, estimated that the increase
in global tropospheric ozone abundance since 1750, and thus associated RF, has likely
contributed roughly 0.1-0.3°C warming to near-surface air temperatures globally, in the context
of the roughly 1.0°C total warming from preindustrial times to the present (Forstcr et al.. 2007).
These temperature effects are the result of ozone concentration changes throughout the total depth
of the troposphere: it is well understood that Earth's surface temperatures should be most
sensitive to ozone concentration changes in the upper troposphere, rather than those closer to the
surface, in the absence of additional feedbacks, as summarized in the 2013 Ozone ISA.
• Literature cited in IPCC AR5 (Mvhrc et al.. 2013) continues to be consistent with these earlier
estimates of the effects of tropospheric ozone on global surface temperatures. Xie et al. (2016). in
a more recent modeling study consisting of a series of 15-year simulations with a global coupled
chemistry-climate model, found a similar increase in global- and annual-mean surface
temperature of 0.36°C (averaged over the last 10 years of each simulation) from preindustrial
times to the present (i.e., calculated as the difference in tropospheric ozone concentration between
1850 and 2013).
• Regional temperature effects may be larger. For example, earlier modeling studies indicated that
increased tropospheric ozone over the second half of the 20th century has caused proportionally
more warming in the Northern Hemisphere than globally, particularly in the Arctic and in
continental interiors (Chang et al.. 2009; Shindell and Faluvegi. 2009; Shindell et al.. 2006;
Micklev et al.. 2004). More recent evidence from Xie et al. (2016) found stronger surface
temperature increases over the high latitudes in both hemispheres, with the maximum increase
exceeding 1.4°C in Siberia (Figure 9-7). This type of regional warming pattern (e.g., Arctic
amplification) is broadly similar to that associated with other radiative forcing agents, including
well-mixed GHGs.
9-18
-------
Reprinted with permission of publisher, Xie et al. (2016).
Figure 9-7 Mean annual change in surface temperature (°C) resulting from
tropospheric ozone concentration changes from 1850-2013.
90°N
60°N
30°N
EQ
30°S
60°S
90°S
0° 60°E 120°E 180° 120°W 60° W 0°
• Idealized modeling studies also support this basic magnitude of the impact of ozone RF on global
and regional temperatures (Tluszar et al.. 2012; Yang et al.. 2012).
• New evidence from recent modeling studies find that the uneven spatial distribution of RF from
historical changes in both aerosols and tropospheric ozone leads to stronger climate response per
unit of global-mean RF than for the well-mixed GHGs [globally and in the Northern Hemisphere
extratropics; Shindell et al. (2015); Shindell (2014)1. This enhanced sensitivity occurs because
most of this RF is located in Northern Hemisphere extratropical latitudes where it triggers more
rapid land responses and stronger feedbacks.
• As described in detail in the 2013 Ozone ISA and the IPCC assessments, however, the
heterogeneous distribution of ozone in the troposphere complicates the direct attribution of spatial
patterns of temperature change to ozone-induced RF (e.g., horizontal gradients in ozone RF and
resulting induced heat transport may weaken the correlation between local RF and local
temperature response). Such effects may also create ozone-climate feedbacks that further alter the
relationship between ozone RF and temperature (and other climate variables) in complex ways
(Fiore et al.. 2015). In addition, the vertical distribution of ozone in the troposphere is also
nonuniform, which may influence atmospheric stability and convection and cloud processes,
leading to further effects on climate and the potential for additional feedbacks.
• Quantifying the climate effects of tropospheric ozone requires complex climate simulations that
include the important feedbacks and interactions discussed above. Current limitations in climate
modeling tools, such as uncertainties associated with simulating trends in upper tropospheric
ozone concentrations, as well as variation across models and the need for more comprehensive
observational data on these effects, represent sources of uncertainty in quantifying the precise
magnitude of climate responses to ozone changes, particularly at regional scales (Mvhre et al..
9-19
-------
2013). These are in addition to other key sources of uncertainty in quantifying ozone RF changes
discussed above in Section 9.2. such as emissions over the time period of interest and baseline
ozone concentrations during preindustrial times.
• As discussed in the 2013 Ozone ISA, future trends in tropospheric ozone concentrations, and
therefore, effects on RF and climate, depend largely on what emissions pathway the world
follows in the coming decades, as discussed in the IPCC AR5 (Mvhre et al.. 2013). Such ozone
effects trends will also depend on changes in a suite of climate-sensitive factors, such as the water
vapor content of the atmosphere. From the 2013 Ozone ISA: "Several studies have included
tropospheric ozone in their investigations of the response in the future atmosphere to a suite of
short-lived species [e.g.. Lew et al. (2008); Shindell et al. (2008); Shindell et al. (2007)/. Few
studies, however, have calculated the climate response to changes in tropospheric ozone
concentrations alone in the future atmosphere. " This conclusion remains the case with the
current literature fMvhre et al.. 2013).
9.3.2 Recent Evidence for Other Climate Effects
Some new research has explored certain additional aspects of the climate response to ozone RF
beyond global and regional temperature change. Specifically, ozone changes are understood to have
impacts on other climate metrics such as precipitation and atmospheric circulation patterns, and new
evidence has continued to support and further quantify this understanding. This new evidence is limited to
a relatively small number of studies, however, leaving various uncertainties still to be resolved. While less
work has been done on ozone, recent work on understanding impacts on precipitation from other
heterogeneously distributed radiative forcers, such as aerosols (Liu et al.. 2018; Westervelt et al.. 2018).
could improve estimates of ozone RF effects on precipitation going forward.
9.3.2.1 Precipitation
• The Xie et al. (2016) study, cited above for temperature, also examined precipitation. The study
authors found in their model simulations that the difference in tropospheric ozone concentration
between 1850 and 2013 caused an increase of 0.02 mm/day in global-average precipitation, with
regional shifts in precipitation patterns near the equator (Figure 9-8).
• Additional modeling studies have examined the relative influence of ozone on precipitation
compared with well-mixed GHGs. Using simple model calculations of the time-dependent
precipitation change due to an RF perturbation, Macintosh et al. (2016) found that the
contribution of ozone change from 1765-2011 to precipitation change over that period could
exceed 50% of that due to CO2 change (including both stratospheric and tropospheric ozone,
though the bulk of the RF change is from tropospheric ozone). Shindell et al. (2012b) also found
that both ozone and aerosol RF typically induce larger precipitation responses than the equivalent
CO2 forcing, and that these spatially heterogeneous forcers are therefore potentially disruptive to
the hydrologic cycle at regional scales.
9-20
-------
0° 60°E 120°E 180° 120°W 60° W 0°
^znnj^mm
-0.5 -0.3 -0.1 0.1 0.3 0.5
Reprinted with permission of publisher, Xie et al. (2016).
Figure 9-8 Mean annual change in precipitation (mm/day) resulting from
tropospheric ozone concentration changes from 1850-2013.
9.3.2.2 Atmospheric Circulation
• The climate modeling study of Allen et al. (2012) concluded that RF due to increases in black
carbon and tropospheric ozone were the most likely causes of an observed increase in the width
of the tropical belt and associated poleward shift in the Northern Hemisphere extratropical storm
tracks over the last few decades. This result is uncertain, however, because other studies have
found that increases in well-mixed GHGs alone can account for this widening of the tropical belt
and stonn track shift (Lau and Kim. 2015).
9.3.3 Summary and Causality Determination
Recent evidence is sufficient to conclude that there is a likely to be causal relationship between
tropospheric ozone and temperature, precipitation, and related climate variables concluded in the
2013 Ozone ISA (Table 9-5; referred to as "climate change" in the 2013 Ozone ISA). The new evidence
conies from the IPCC AR5 (Mvhre et al.. 2013) and its supporting references—in addition to a few more
recent studies—and builds on evidence presented in the 2013 Ozone ISA. None of the new studies
support a change to the causality determination included in the 2013 Ozone ISA.
9-21
-------
Consistent with previous estimates, the effect of global, total tropospheric ozone increases on
global mean surface temperatures continues to be estimated at roughly 0.1-0.3°C since
preindustrial times (Xic et al.. 2016; Mvhre et al.. 2013). with larger effects in some regions.
In addition to temperature, tropospheric ozone changes have impacts on other climate metrics
such as precipitation and atmospheric circulation patterns (Macintosh et al.. 2016; Allen et al..
2012; Shindell etal.. 2012a).
Various uncertainties render the precise magnitude of the overall effect of tropospheric ozone on
climate more uncertain than that of the well-mixed GHGs (Mvhre et al.. 2013). These include the
remaining uncertainties in the magnitude of RF estimated to be attributed to tropospheric ozone.
In addition, precisely quantifying the change in surface temperature (and other climate variables)
due to tropospheric ozone changes requires complex climate simulations that include all relevant
feedbacks and interactions. Current limitations in climate modeling tools, variation across
models, and the need for more comprehensive observational data on these effects represent
sources of uncertainty in quantifying the precise magnitude of climate responses to ozone
changes, particularly at regional scales (Mvhre et al.. 2013).
Even with these uncertainties, however, evidence from climate models indicates that tropospheric
ozone changes have likely contributed to observed increases in global and regional mean surface
temperatures.
Further research in the following areas can help address these remaining uncertainties:
quantifying more precise relationships between regional ozone RF and regional climate change;
improving understanding of the impacts of heterogeneously distributed RF, including from ozone,
aerosols, and other short-lived climate forcers, on the hydrologic cycle, precipitation, and
atmospheric circulation patterns; improving understanding of, and ability to model, critical
ozone-climate feedbacks; and continuing exploration of links between precursor pollutant control
strategies, climate, and ozone concentrations.
9-22
-------
Table 9-5 Summary of evidence for a likely to be causal relationship between
ozone and temperature, precipitation, and related climate variables.
Rationale for Causality
Determination Key Evidence Key References
Consistent evidence from Temperature: Multidecadal, global Xie et al. (2016); Mvhre et al. (2013)
multiple, high-quality studies chemistry-climate modeling ensemble
studies constrained by historical
observations of ozone concentrations;
not as many such studies as for RF
Other climate effects (precipitation. Xie et al. (2016); Allen et al. (2012)
atmospheric circulation): Multidecadal,
global chemistry-climate modeling
ensemble studies constrained by
historical observations of ozone
concentrations; only a limited number of
such modeling studies for these other
climate effects
Robust physical Temperature: Robust, well-understood Huszar et al. (2012); Yang et al. (2012);
understanding relationship between RF and Xie et al. (2016); Mvhre et al. (2013);
atmospheric and surface temperatures Fiore et al. (2015)
Other climate effects (precipitation. Xie et al. (2016); Macintosh et al. (2016);
atmospheric circulation): Good Allen et al. (2012); Shindell et al. (2012a)
understanding of the potential ways
ozone RF and temperature effects
further influence atmospheric
thermodynamics and dynamics;
however, multiple complex interactions
and feedbacks confound precise
quantification of the magnitude of the
ozone effect
Spatial/temporal effect Temperature: Largest effects seen in the Xie et al. (2016): Mvhre et al. (2013):
correspondence Northern Hemisphere's middle and high Shindell and Faluvegi (2009): Shindell et
latitudes, consistent with observed al. (2015): Shindell (2014)
patterns of pollutant emissions combined
with climate dynamical processes that
lead to Arctic amplification of
temperature responses to RF
Other climate effects (precipitation. Xie et al. (2016): Allen et al. (2012)
atmospheric circulation): Spatial
correspondence not as strong with these
other climate effects, due to complex
interactions and feedbacks in the climate
system at multiple space and time
scales; more limited studies to date
9-23
-------
9.4 References
Allen. RJ: Sherwood. SC; Norris. JR; Zender. CS. (2012). Recent Northern Hemisphere tropical
expansion primarily driven by black carbon and tropospheric ozone. Nature 485: 350-354.
http://dx.doi.org/10.1038/naturellQ97
Bellouin. N; Baker. L; Hodnebrog. O; Olivie. D; Cherian. R; Macintosh. C; Samset. B; Esteve. A;
Aamaas. B; Quaas. J; Myhre. G. (2016). Regional and seasonal radiative forcing by perturbations
to aerosol and ozone precursor emissions. Atmos Chem Phys 16: 13885-13910.
http://dx.doi.org/10.5194/acp-16-13885-2016
Bowman. KW; Shindell. DT; Worden. HM; Lamarque. JF; Young. PJ; Stevenson. DS; Qu. Z; De La
torre. M: Bergmann. D: Cameron-Smith. P.T: Collins. WJ; Dohertv. R; Dalsoren. SB: Faluvegi. G;
Folberth. G: Horowitz. LW: Josse. BM: Lee. YH: Mackenzie. IA: Myhre. G: Nagashima. T: Naik.
V: Plummer. DA: Rumbold. ST: Skeie. RB: Strode. SA; Sudo. K; Szopa. S: Voulgarakis. A: Zeng.
G: Kulawik. SS: Aghedo. AM: Worden. J. R. (2013). Evaluation of ACCMIP outgoing longwave
radiation from tropospheric ozone using TES satellite observations. Atmos Chem Phys 13: 4057-
4072. http://dx.doi.org/10.5194/acp-13-4057-2013
Chang. W: Liao. H; Wang. H. (2009). Climate responses to direct radiative forcing of anthropogenic
aerosols, tropospheric ozone, and long-lived greenhouse gases in eastern China over 1951-2000.
Adv Atmos Sci 26: 748-762. http://dx.doi.org/10.1007/s00376-009-9Q32-4
Checa-Garcia. R; Hegglin. MI; Kinnison. D; Plummer. DA: Shine. KP. (2018). Historical tropospheric
and stratospheric ozone radiative forcing using the CMIP6 database. Geophys Res Lett 45: 3264-
3273. http://dx.doi.org/10.1002/2017GL07677Q
Conlev. AJ; Lamarque. JF; Vitt. F; Collins. WD; Kiehl. J. (2013). PORT, a CESM tool for the
diagnosis of radiative forcing. GMD 6: 469-476. http://dx.doi.org/10.5194/gmd-6-469-2013
Dahlmann. K; Grewe. V; Ponater. M; Matthes. S. (2011). Quantifying the contributions of individual
NO(x) sources to the trend in ozone radiative forcing. Atmos Environ 45: 2860-2868.
http://dx.doi.Org/10.1016/i.atmosenv.2011.02.071
Derwent. RG; Collins. WJ; Johnson. CE; Stevenson. DS. (2001). Transient behaviour of tropospheric
ozone precursors in a global 3-D CTM and their indirect greenhouse effects. Clim Change 49: 463-
487. http://dx.doi.Org/10.1023/A:1010648913655
Fiore. AM; Naik. V; Leibensperger. EM. (2015). Air quality and climate connections [Review]. J Air
Waste Manag Assoc 65: 645-685. http://dx.doi.org/10.1080/10962247.2015.104Q526
Forster. P; Ramaswamv. V; Artaxo. P; Berntsen. T; Betts. R; Fahev. DW; Havwood. J; Lean. J; Lowe.
DC; Myhre. G; Nganga. J; Prinn. R; Raga. G; Schultz. M; Van Dorland. R. (2007). Changes in
atmospheric constituents and in radiative forcing. In Climate change 2007: The physical science
basis Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental
Panel on Climate Change (pp. 129-234). Cambridge, United Kingdom: Cambridge University
Press, http://www.ipcc.ch/pdf/assessment-report/ar4/wgl/ar4-wgl-chapter2.pdf
Fry. MM; Naik. V; West. JJ; Schwarzkopf. MP; Fiore. AM; Collins. WJ; Dentener. F.I: Shindell. DT;
Atherton. C; Bergmann. D; Duncan. BN; Hess. P; Mackenzie. I; Marmer. E; Schultz. MG; Szopa.
S; Wild. O; Zeng. G. (2012). The influence of ozone precursor emissions from four world regions
on tropospheric composition and radiative climate forcing. J Geophys Res 117: n/a-n/a.
http://dx.doi.org/10.1029/2011JD017134
9-24
-------
Gauss. M; Myhre. G; Isaksen. ISA; Grewe. V; Pitari. G; Wild. O; Collins. WJ; Dentener. FJ;
Ellingscn. K; Gohar. LK; Hauglustaine. DA; Iachetti. D; Lamarque. JF; Mancini. E; Micklev. LJ;
Prather. MJ; Pvle. JA; Sanderson. MG; Shine. KP; Stevenson. DS; Sudo. K; Szopa. S; Zeng. G.
(2006). Radiative forcing since preindustrial times due to ozone change in the troposphere and the
lower stratosphere. Atmos Chem Phys 6: 575-599. http://dx.doi.org/10.5194/acp-6-575-20Q6
Hansen. J; Sato. M; Ruedv. R; Nazarenko. L; Lacis. A; Schmidt. GA; Russell. G; Aleinov. I; Bauer.
M; Bauer. S; Bell. N; Cairns. B; Canuto. V; Chandler. M; Cheng. Y; Del Genio. A; Faluvegi. G;
Fleming. E; Friend. A; Hall. T; Jackman. C; Kellev. M; Kiang. N; Koch. D; Lean. J; Lerner. J; Lo.
K; Menon. S; Miller. R; Minnis. P; Novakov. T; Oinas. V; Perlwitz. J; Perlwitz. J; Rind. D;
Romanou. A; Shindell. D; Stone. P; Sun. S; Tausnev. N; Thresher. D; Wielicki. B; Wong. T; Yao.
M; Zhang. S. (2005). Efficacy of climate forcings. J Geophys Res 110: D18104.
http://dx.doi.org/10.1029/2005JD0Q5776
Hu. L; Jacob. DJ; Liu. X; Zhang. Y; Zhang. L; Kim. PS; Sulprizio. MP; Yantosca. RM. (2017). Global
budget of tropospheric ozone: Evaluating recent model advances with satellite (OMI), aircraft
(IAGOS), and ozonesonde observations. Atmos Environ 167: 323-334.
http://dx.doi.Org/10.1016/i.atmosenv.2017.08.036
Huszar. P; Miksovskv. J; Pisoft. P; Belda. M; Halenka. T. (2012). Interactive coupling of a regional
climate model and a chemical transport model: evaluation and preliminary results on ozone and
aerosol feedback. Clim Res 51: 59-88. http://dx.doi.org/10.3354/cr01054
IPCC (Intergovernmental Panel on Climate Change). (2013). Climate change 2013: The physical
science basis. Contribution of working group I to the fifth assessment report of the
Intergovernmental Panel on Climate Change. In TF Stacker; D Qin; GK Plattner; MMB Tignor;
SK Allen; J Boschung; A Nauels; Y Xia; V Bex; PM Midgley (Eds.). Cambridge, UK: Cambridge
University Press, http://www.ipcc.ch/report/ar5/wg 1 /
Jasaitis. D; Vasiliauskiene. V; Chadvsiene. R; Peciuliene. M. (2016). Surface Ozone Concentration
and Its Relationship with UV Radiation, Meteorological Parameters and Radon on the Eastern
Coast of the Baltic Sea. Atmosphere (Basel) 7. http://dx.doi.org/10.3390/atmos7020Q27
Lamv. K; Portafaix. T; Brogniez. C; Godin-Beekmann. S; Bencherifi H; Morel. B; Pazmino. A;
Metzger. JM; Auriol. F; Deroo. C; Duflot. V; Goloub. P; Long. CN. (2018). Ultraviolet radiation
modelling from ground-based and satellite measurements on Reunion Island, southern tropics.
Atmos Chem Phys 18: 227-246. http://dx.doi.org/10.5194/acp-18-227-2018
Lau. WK; Kim. KM. (2015). Robust Hadley Circulation changes and increasing global dryness due to
C02 warming from CMIP5 model projections. Proc Natl Acad Sci USA 112: 3630-3635.
http://dx.doi.org/10.1073/pnas.1418682112
Lew. H; Schwaarzkopf. MP; Horowitz. L; Ramaswamv. V; Findell. KL. (2008). Strong sensitivity of
late 21st century climate to projected changes in short-lived air pollutants. J Geophys Res 113:
D06102. http://dx.doi.org/10.1029/2007JD0Q9176
Liao. H; Seinfeld. JH; Adams. PJ; Micklev. LJ. (2004). Global radiative forcing of coupled
tropospheric ozone and aerosols in a unified general circulation model. J Geophys Res 109:
D16207. http://dx.doi.org/10.1029/2003JD0Q4456
Liu. L; Shawki. D; Voulgarakis. A; Kasoar. M; Samset. BH; Myhre. G; Forster. PM; Hodnebrog. O;
Sillmann. J; Aalbergsio. SG; Boucher. O; Faluvegi. G; Iversen. T; Kirkevag. A; Lamarque. JF;
Olivie. D; Richardson. T; Shindell. D; Takemura. T. (2018). A PDRMIP multimodel study on the
impacts of regional aerosol forcings on global and regional precipitation. J Clim 31: 4429-4447.
http://dx.doi.Org/10.1175/JCLI-D-17-0439.l
9-25
-------
Macintosh. CR; Allan. RP: Baker. LH; Bellouin. N; Collins. W; Mousavi. Z; Shine. KP. (2016).
Contrasting fast precipitation responses to tropospheric and stratospheric ozone forcing. Geophys
Res Lett 43: 1263-1271. http://dx.doi.org/10.1002/2015GLQ67231
Madronich. S; Wagner. M; Groth. P. (2011). Influence of tropospheric ozone control on exposure to
ultraviolet radiation at the surface. Environ Sci Technol 45: 6919-6923.
http://dx.doi.org/10.1021/es200701q
Marenco. A: Gouget. H: Nedelec. P; Pages. JP; Karcher. F. (1994). Evidence of a long-term increase
in tropospheric ozone from Pic du Midi data series: Consequences: Positive radiative forcing. J
Geophys Res 99: 16617-16632. htto://dx.doi.org/10.1029/94JD00021
Micklev. LJ: Jacob. DJ; Field. BP: Rind. D. (2004). Climate response to the increase in tropospheric
ozone since preindustrial times: A comparison between ozone and equivalent C02 forcings. J
Geophys Res 109: D05106. http://dx.doi.org/10.1029/2003JD0Q3653
Micklev. LJ: Murti. PP; Jacob. DJ; Logan. JA: Koch. DM: Rind. D. (1999). Radiative forcing from
tropospheric ozone calculated with a unified chemistry-climate model. J Geophys Res 104: 30153-
30172. http://dx.doi.org/10.1029/1999JD900439
Mvhre. G: Aas. W: Cherian. R: Collins. W: Faluvegi. G: Flanner. M: Forster. P: Hodnebrog. O:
Klimont. Z: Lund. MT: Muelmenstaedt. J: Mvhre. CL: Olivie. D: Prather. M: Ouaas. J: Samset.
BH; Schnell. JL; Schulz. M; Shindell. D; Skeie. RB: Takemura. T; Tsvro. S. (2017). Multi-model
simulations of aerosol and ozone radiative forcing due to anthropogenic emission changes during
the period 1990-2015. Atmos Chem Phys 17: 2709-2720. http://dx.doi.org/10.5194/acp-17-27Q9-
2017
Mvhre. G: Shindell. D; Breon. FM; Collins. W: Fuglestvedt. J: Huang. J: Koch. D; Lamarque. JF; Lee.
D: Mendoza. B: Nakaiima. T: Robock. A: Stephens. G: Takemura. T: Zhang. H. (2013).
Anthropogenic and natural radiative forcing. In TF Stacker; D Qin; GK Plattner; MMB Tignor; SK
Allen; J Boschung; A Nauels; Y Xia; V Bex; PM Midgley (Eds.), Climate change 2013: The
physical science basis Contribution of working group I to the fifth assessment report of the
Intergovernmental Panel on Climate Change (pp. 659-740). Cambridge, UK: Cambridge University
Press, http://www.ipcc.ch/report/ar5/wg 1/
Newsome. B; Evans. M. (2017). Impact of uncertainties in inorganic chemical rate constants on
tropospheric composition and ozone radiative forcing. Atmos Chem Phys 17: 14333-14352.
http://dx.doi.org/10.5194/acp-17-14333-2017
Parrish. DP: Lamarque. JF: Naik. V: Horowitz. L; Shindell. DT; Staehelin. J: Derwent. R; Cooper.
OR: Tanimoto. H: Volz-Thomas. A: Gilge. S: Scheel. HE: Steinbacher. M: Froehlich. M. (2014).
Long-term changes in lower tropospheric baseline ozone concentrations: Comparing chemistry-
climate models and observations at northern midlatitudes. J Geophys Res Atmos 119: 5719-5736.
http://dx.doi.org/10.1002/2013JDQ21435
Sherwen. T: Evans. MJ: Carpenter. LJ: Schmidt. JA: Micklev. LJ. (2017). Halogen chemistry reduces
tropospheric 03 radiative forcing. Atmos Chem Phys 17: 1557-1569.
http://dx.doi.org/10.5194/acp-17-1557-2017
Shindell. D: Faluvegi. G. (2009). Climate response to regional radiative forcing during the twentieth
century. Nat Geosci 2: 294-300. http://dx.doi.org/10.1038/ngeo473
Shindell. D; Faluvegi. G: Lacis. A: Hansen. J: Ruedv. R; Aguilar. E. (2006). Role of tropospheric
ozone increases in 20th-century climate change. J Geophys Res 111: D08302.
http://dx.doi.org/10.1029/2005JD0Q6348
9-26
-------
Shindell. D; Faluvcgi. G; Nazarenko. L; Bowman. K; Lamarque. JF; Voulgarakis. A; Schmidt. GA;
Pechonv. O; Ruedv. R. (2013). Attribution of historical ozone forcing to anthropogenic emissions.
Nat Clim Chang 3: 567-570. htto://dx.doi.org/10.1038/nclimate 1835
Shindell. D; Kuvlenstierna. JC; Vignati. E; van Dingenen. R; Amann. M; Klimont. Z; Anenberg. SC;
Muller. N; Janssens-Maenhout. G; Raes. F; Schwartz. J; Faluvegi. G; Pozzoli. L; Kupiainen. K;
Hoglund-Isaksson. L; Emberson. L; Streets. D; Ramanathan. V; Hicks. K; Oanh. NT; Millv. G;
Williams. M; Demkine. V; Fowler. D. (2012a). Simultaneously mitigating near-term climate
change and improving human health and food security. Science 335: 183-189.
http ://dx.doi .org/10.1126/science. 1210026
Shindell. DT. (2014). Inhomogeneous forcing and transient climate sensitivity. Nat Clim Chang 4:
274-277. http://dx.doi.org/10.1038/NCLIMATE2136
Shindell. DT; Faluvegi. G; Bauer. SE; Koch. DM; Unger. N; Menon. S; Miller. RL: Schmidt. GA;
Streets. DG. (2007). Climate response to projected changes in short-lived species under an A1B
scenario from 2000-2050 in the GISS climate model. J Geophys Res 112: D20103.
http://dx.doi.org/10.1029/2007id0Q8753
Shindell. DT; Faluvegi. G; Rotstavn. L; Millv. G. (2015). Spatial patterns of radiative forcing and
surface temperature response. J Geophys Res Atmos 120: 5385-5403.
http://dx.doi.org/10.1002/2014JDQ22752
Shindell. DT; Lew H. II; Schwarzkopf. MP; Horowitz. LW; Lamarque. JF; Faluvegi. G. (2008).
Multimodel projections of climate change from short-lived emissions due to human activities. J
Geophys Res 113: D11109. http://dx.doi.org/10.1029/2007JD0Q9152
Shindell. DT; Voulgarakis. A; Faluvegi. G; Millv. G. (2012b). Precipitation response to regional
radiative forcing. Atmos Chem Phys 12: 6969-6982. http://dx.doi.org/10.5194/acp-12-6969-2012
Skeie. RB; Berntsen. TK; Mvhre. G; Tanaka. K; Kvalevag. MM; Hovle. CR. (2011). Anthropogenic
radiative forcing time series from pre-industrial times until 2010. Atmos Chem Phys 11: 11827-
11857. http://dx.doi.org/10.5194/acp-l 1-11827-2011
Sovde. OA; Hovle. CR; Mvhre. G; Isaksen. ISA. (2012). The HN03 forming branch of the H02 + NO
reaction: pre-industrial-to-present trends in atmospheric species and radiative forcings (vol 11, pg
8929, 2011). Atmos Chem Phys 12: 7725-7725. http://dx.doi.org/10.5194/acp-12-7725-2012
Staehelin. J; Thudium. J; Buehler. R; Volz-Thomas. A; Graber. W. (1994). Trends in surface ozone
concentrations at Arosa (Switzerland). Atmos Environ 28: 75-87. http://dx.doi.org/10.1016/1352-
2310(94)90024-8
Stevenson. DS; Young. PJ; Naik. V; Lamarque. JF; Shindell. DT; Voulgarakis. A; Skeie. RB;
Dalsoren. SB; Mvhre. G; Berntsen. TK; Folberth. GA; Rumbold. ST; Collins. WJ; Mackenzie. TA:
Dohertv. RM; Zeng. G; van Noiie. TPC; Strunk. A; Bergmann. D; Cameron-Smith. P; Plummer.
DA; Strode. SA; Horowitz. L; Lee. YH; Szopa. S; Sudo. K; Nagashima. T; Josse. B; Cionni. I;
Rjghi. M; Evring. V; Conlev. A; Bowman. KW; Wild. O; Archibald. A. (2013). Tropospheric
ozone changes, radiative forcing and attribution to emissions in the Atmospheric Chemistry and
Climate Model Intercomparison Project (ACCMIP). Atmos Chem Phys 13: 3063-3085.
http://dx.doi.org/10.5194/acp-13-3063-2013
U.S. EPA (U.S. Environmental Protection Agency). (2015). Preamble to the integrated science
assessments [EPA Report]. (EPA/600/R-15/067). Research Triangle Park, NC: U.S. Environmental
Protection Agency, Office of Research and Development, National Center for Environmental
Assessment, RTP Division. https://cfpub.epa.gov/ncea/isa/recordisplav.cfm?deid=310244
9-27
-------
Westervelt. DM; Conlev. AJ; Fiore. AM; Lamarque. JF; Shindell. DT; Previdi. M; Mascioli. NR;
Faluvcgi. G; Correa. G; Horowitz. LW. (2018). Connecting regional aerosol emissions reductions
to local and remote precipitation responses. Atmos Chem Phys 18: 12461-12475.
http://dx.doi.org/10.5194/acp-18-12461-2018
Worden. HM; Bowman. KW; Kulawik. SS; Aghedo. AM. (2011). Sensitivity of outgoing longwave
radiative flux to the global vertical distribution of ozone characterized by instantaneous radiative
kernels from Aura-TES. J Geophys Res 116: D14115. http://dx.doi.org/10.1029/2010JDQ15101
Wuebbles. DJ; Fahev. DW; Hibbard. KA; Dokken. DJ; Stewart. BC; Mavcock. TK. (2017). USGCRP,
2017: Climate science special report: Fourth national climate assessment, Volume I. Washington,
DC: U.S. Global Change Research Program, https://science2017.globalchange.gov/
Xie. B; Zhang. H; Wang. Z; Zhao. S; Fu. Q. (2016). A modeling study of effective radiative forcing
and climate response due to tropospheric ozone. Adv Atmos Sci 33: 819-828.
http://dx.doi.org/10.1007/s00376-Q16-5193-0
Yang. J; Hu. Y; Peltier. WR. (2012). Radiative effects of ozone on the climate of a Snowball Earth.
Climate of the Past 8: 2019-2029. http://dx.doi.org/10.5194/cp-8-2019-2012
Young. PJ; Archibald. AT; Bowman. KW; Lamarque. JF; Naik. V; Stevenson. DS; Tilmes. S;
Voulgarakis. A; Wild. O; Bergmann. D; Cameron-Smith. P; Cionni. I; Collins. WJ; Dalsoren. SB;
Dohertv. RM; Evring. V; Faluvegi. G; Horowitz. LW; Josse. B; Lee. YH; Mackenzie. IA;
Nagashima. T; Plummer. DA; Righi. M; Rumbold. ST; Skeie. RB; Shindell. DT; Strode. SA; Sudo.
K; Szopa. S; Zeng. G. (2013). Pre-industrial to end 21st century projections of tropospheric ozone
from the Atmospheric Chemistry and Climate Model Intercomparison Project (ACCMIP). Atmos
Chem Phys 13: 2063-2090. http://dx.doi.org/10.5194/acp-13-2063-2013
Zhang. Y; Cooper. OR; Gaudel. A; Thompson. AM; Nedelec. P; Qgino. SY; West. JJ. (2016).
Tropospheric ozone change from 1980 to 2010 dominated by equatorward redistribution of
emissions. Nat Geosci 9: 875-+. http://dx.doi.org/10.1038/NGEQ2827
9-28
-------
APPENDIX 10 DEVELOPMENT OF THE
INTEGRATED SCIENCE
ASSESSMENT —PROCESS
10.1 Introduction
The Integrated Science Assessment (ISA) provides the scientific foundation for the review of the
National Ambient Air Quality Standards (NAAQS). The ISA contains a synthesis and evaluation of the
most policy-relevant science using methods and approaches described in the Preamble to the Integrated
Science Assessments (U.S. EPA. 2015b). hereafter "Preamble," which provides an overview of the ISA
development process. Early in the review, U.S. EPA releases an Integrated Review Plan (IRP) to
summarize the entire process for the NAAQS review, including a plan for developing the ISA
(https://www.epa.gov/naaas/ozone-o3-standards-planning-documents-current-review'). The ISA is
developed by U.S. EPA scientists in the Office of Research and Development (ORD) with extensive
knowledge in their respective fields, other U.S. EPA scientists with relevant expertise, and extramural
scientists who are solicited by the U.S. EPA for their subject-matter expertise. The general development
steps are presented in Figure 10-1. but the specific details can vary from assessment to assessment. This
Appendix builds off the Preamble and the IRP to describe the process for this current ISA. Appendix 10
was developed to describe the methods for literature review, study quality evaluation, and quality
assurance and quality control.
10.2 Literature Search and Initial Screen
Development of the Ozone ISA began with an initial literature search and screening process. For
this step, the ISA authors applied systematic review methodologies to identify relevant scientific findings
that have emerged since the 2013 Ozone ISA (U.S. EPA. 2013) which included peer-reviewed literature
published through July 2011. Search techniques for the current ISA identified and evaluated studies and
reports that have undergone scientific peer review and were published or accepted for publication
between January 1, 2011 (providing some overlap with the cutoff date from the last review) and March
30, 2018. Studies published after the literature cutoff date for this review were also considered if they
were submitted in response to the Call for Information or identified in subsequent phases of ISA
development (e.g., peer-input consultation; see Figure 10-1). particularly to the extent that they provide
new information that affects key scientific conclusions.
Peer-reviewed literature was identified and evaluated to provide a better understanding of the
following issues: (1) the natural and anthropogenic sources of ozone precursors in the ambient air;
(2) formation, transport, and fate of ozone in the environment; (3) measurement methods and ambient
10-1
-------
concentrations of ozone; (4) how exposure assessment methods used in epidemiologic studies can
influence inferences drawn about ozone health effects; (5) the independent effect of ozone exposure on
health and welfare1 effects; (6) the potential influence of other factors (e.g., other pollutants in the
ambient air, ambient air temperature) shown to be correlated with ozone and health or welfare effects;
(7) the shape of the concentration-response relationship at ozone concentrations at the low end of the
distribution; and (8) populations and lifestages at increased risk of ozone-related health effects.
The literature search and screening processes are described in the sections below and are
summarized in the Literature Flow Diagram shown below (Figure 10-2).
1 Section 302(h) of the Clean Air Act [42 U.S.C. 7602(h)] provides that all language referring to effects on welfare
includes, but is 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..." fCAA. 2005).
10-2
-------
Literature Search and
Study Selection
*
Evaluation of Individual Study Quality
After study selection, the quality of individual studies is evaluated by EPA or outside experts in the fields of
atmospheric science, exposure assessment, dosimetry, animal toxicology, controlled human exposure studies,
epidemiology, ecology, and otherwelfare effects, considering the design, methods, conduct, and documentation of
each study. Strengths and limitations of individual studies that may affect the interpretation of the study are
considered.
*
Develop Initial Sections
Review and summarize new study results as well
as findings and conclusions from previous
assessments by category of outcome/effectand
by discipline, e.g., toxicological studiesoflung
function.
Peer Input Consultation
Review of initial draft materials by scientists
from both outside and within EPA in public
meeting orpublic teleconference.
*
Evaluation, Synthesis, and Integration of Evidence
Integrate evidence from scientific disciplines - for example, toxicological, controlled human exposure, and
epidemiologic study findings for a particular health outcome. Evaluate evidence for related groups of endpoints or
outcomes to draw conclusions regarding health orwelfare effect categories, integrating health orwelfare effects
evidence with information on mode of action and exposure assessment.
Development of Scientific Conclusions and Causal Determinations
Characterize weight of evidence and develop judgments regarding causality for health or welfare effect categories.
Develop conclusions regarding concentration- or dose-response relationships, potentially at-risk populations,
lifestages, or ecosystems.
It
Draft Integrated Science Assessment
Evaluation and integration of newly published studies
after each draft.
Clean Air Scientific Advisory Committee
Independent review of draft documents for scientific
quality and sound implementation of causal
framework; anticipated review of two drafts of ISA in
public meetings.
Public Comments
Comments on draft ISA solicited by EPA
Final Integrated Science Assessment
Source: Modified from Figure II of the Preamble to the Integrated Science Assessments U.S. EPA (2015b1.
Figure 10-1 General process for development of Integrated Science
Assessments.
10-3
-------
Keyword Search
(26,253)
Keyword Search
Included
(9,612)
Topic Classified:
• Epidemiology (1,756)
• Experimental (3,312}
• Exposure(5,141)
Keyword Search
Excluded
(16,641)
• Not Topic
Classified
Citation Mapping
(22,901)
Topic Specific:
• Climate (10,193)
¦ Ecology (12,180)
• Atmospheric - Background
Ozone (2,581)
f
¦\
Literature Search Included
(31,906)
v
J
\
r
References
from Other
Sources
I (825) J
Level 1 Title-Abstract
Screening Included
(4,548)
r i
Level 2 Full-Text Screening Included
(1,704)
V
\
r
Level 1 Title-Abstract Screening Excluded
(27,358)
Level 2 Full-Text Screening Excluded
3.669
Included in Ozone Integrated Science Assessment
(1,704)
' Appendix 1: Atmospheric (211) ' Appendix 6: Mortality (72)
• Appendix 2; Exposure (201) ¦ Appendix 7: Other Health (226)
• Appendix 3. Respiratory (316) * Appendix 8: Ecology (536)
• Appendix 4: Cardiovascular (191) • Appendix 9: Climate (52)
• Appendix 5: Metabolic (89) • Appendix 10: Process (21)
Preface (20)
Executive Summary (13)
Integrated Synthesis (83)
Notes: These references are tagged in the HERO database to provide greater transparency in the approach used to identify
references for inclusion in the ISA.
Figure 10-2 Literature flow diagram for the Ozone Integrated Science
Assessment.
10-4
-------
10.2.1
Literature Search
The U.S. EPA uses a structured approach to identify relevant studies for consideration and
inclusion in the ISAs. The search for relevant literature in this review began with publication of the
Federal Register notice announcing the initiation of this ozone NAAQS review and requesting
information from the public including relevant literature (83 FR 29785, June 26, 2018). In addition, the
U.S. EPA identified publications by conducting multitiered systematic literature searches that included
extensive mining of literature databases on specific topics in a variety of disciplines. The search strategies
were designed a priori to optimize identification of pertinent published papers.
For this ISA, discipline-specific approaches were used to identify literature. In each case, careful
consideration was given to literature search strategies used in the development of previous assessments
and to the methods that would result in the best precision1 and recall2 for each of the disciplines. The
literature identification approaches included broad keyword searches in routinely used databases with
automatic topic classification and topic-specific citation mapping (see Section 10.2.2 for specific
approaches used for each discipline).
As has been done for past ISAs, a broad keyword search was developed as a starting point to
capture literature pertinent to the pollutant of interest. In this case, the main keyword string used was
"ozone OR O3," which is sufficiently broad to capture ozone-relevant literature in each database
(i.e., PubMed, Web of Science, TOXLINE). Following the broad keyword search for ozone, automatic
topic classification was used to categorize references by discipline (e.g., epidemiology, toxicology, etc.).
This step employs machine learning where positive and negative seed references3 for a particular
discipline are used to train an algorithm to identify discipline-specific references based on word use and
frequency in titles and abstracts. This method, used in several previous ISAs, varies in effectiveness
across disciplines due to the broad range of topics and the variability in term usage in some evidence
bases.
Another approach used in past ISAs that was employed in this review is topic-specific citation
mapping, also known as relational reference searching. In this approach, a set of relevant published
references are identified as a seed set and then more recent literature that has cited any of the references in
the seed set are collected. Topic-specific references from the 2013 Ozone ISA comprised the seed set for
this ISA. Because the seed set is highly relevant to the topic of interest, this targeted approach to reference
1 Precision refers to the number of relevant references identified divided by the total number of references identified.
2 Recall is the number of relevant references identified divided by the total number of relevant references that exist.
3 Positive seed references are those that are examples of references that are relevant (i.e., the references would be
selected for full-text screening). Negative seed references are those that are examples of references that are not
relevant (i.e., they would not be selected for full-text screening). For ISAs, the positive seed set includes references
from the prior ISA for the discipline of interest. The negative seed set includes the references from all of the other
disciplines in the previous ISA.
10-5
-------
identification is more precise than keyword searches, and it further allows for relevance ranking based on
the number of references in a bibliography that match references in the seed set.
References were also identified for consideration in this ISA in other ways, including
identification of relevant literature by U.S. EPA expert scientists, recommendations received in response
to the Call for Information and the external review process for the ISA (see Section 10.4.3). and by
review of citations included in previous assessments. This multipronged search strategy is consistent with
the approach used in previous ISAs as described in the Preamble.
10.2.2 Study Selection: Initial Screening (Level 1) of Studies from the
Literature Search
Once studies were identified, ISA authors (U.S. EPA staff and extramural scientists) reviewed
and screened the studies for a further refined list of references to be considered for inclusion in the ISA.
References for each discipline (i.e., atmospheric science, exposure assessment, experimental health
studies, epidemiology, ecology, and climate-related science) first went through title and abstract screening
using SWIFT-ActiveScreener (SWIFT-AS), which is referred to as Level 1 screening. Level 1 screening
criteria for inclusion were intended to be broad and err on the side of inclusion. For each discipline, title
and abstracts were selected for consideration if there was indication of ozone and a quantifiable effect
relevant to an identified discipline. SWIFT-AS is a software application that employs machine learning in
real time to identify relevant literature. The machine learning feature builds a model to predict relevant
references based on inclusion/exclusion screening decisions in real time as scientists screen each
reference. As title/abstract screening is conducted, references are queued based on the predicted relevance
and SWIFT-AS estimates when a 95% recall threshold has been reached,1 a level often used to evaluate
the performance of machine learning applications and considered comparable to human error rates
(Howard et al.. 2016; Cohen et al.. 2006).
The application of SWIFT-AS was tailored for each discipline. This included using a specific
seed set of 50-100 relevant references from the 2013 Ozone ISA to train the SWIFT-AS algorithm and
developing specific screening questions for each discipline to allow for the categorization of references
based on the information available in the title and abstract. Specific details about inclusion/exclusion
criteria and the screening questions for each discipline are described in more detail below. The references
identified for inclusion after Level 1 screening were then reviewed in Level 2 full-text screening.
Figure 10-3 demonstrates the efficiency gained by using SWIFT-AS for each discipline. Figure 10-4 is
another example of the efficiency of SWIFT-AS demonstrating for toxicological references the rate at
which 95% recall was achieved with the machine learning algorithm relative to that of manual screening.
1 A 95% recall threshold represents the point at which 95% of the potentially relevant references have been
identified.
10-6
-------
10000
"V
«*
,^°x
i\e"
fc**
nr
,0«*
aO*
.rf*
3®
s5°v
1270
rfe
0.9
<*S»
¦ Consider 0 Exclude ~ Not Screened
Figure 10-3 Summary of title/abstract screening in SWIFT-ActiveScreener.
100
-o
o
o
0 -7C
CC 7s
"O
0)
TJ
o
c
o
d)
o>
-2
c
aj
o
L.
a)
CL
50
25
0k
Your Predicted Recall
Predicted Normal Screening
95% Inclusion
1k 2k
Number of Records Screened
3k
Notes: The green line represents the predicted 95% recall threshold, while the blue line represents the screening progression to
achieve 95% recall using the machine learning algorithm. The red line represents the approximate rate of manual screening all
references.
Figure 10-4 illustrative example of screening efficiency using
SWI FT-Acti veScreener,
10-7
-------
10.2.2.1 Atmospheric Science
Literature related to the atmospheric science topics discussed in Appendix 1 was primarily
identified by topic-specific citation mapping methods that relied upon references cited in the 2013 Ozone
ISA. More specifically, references were collected from subtopics on physical and chemical processes,
atmospheric modeling, monitoring, and background ozone concentrations. Consistent with the ozone IRP,
the primary focus for evaluation of the recent literature was on studies estimating U.S. background (USB)
ozone and its sources. Studies obtained from citation mapping were systematically title/abstract screened
using SWIFT-AS.
Additionally, a recent critical review of the science in which the U.S. EPA played an active role
(Jaffe et al.. 20 IS) was selected to serve as the primary reference for most of the Appendix 1 discussion
concerning USB ozone. In addition to Jaffe et al. (2018). papers published as part of the International
Global Atmospheric Chemistry (IGAC) Project Tropospheric Ozone Assessment Report [TOAR; Gaudel
et al. (2018)1 were reviewed and evaluated. Articles found with citation mapping and subsequent
screening that were not cited and discussed in either the critical review by Jaffe et al. (2018) or the IGAC
TOAR report (Gaudel et al.. 2018) were reviewed independently. Additional topics, including newly
described photochemical mechanisms and the role of variability and change in large-scale meteorological
patterns, were also included in Appendix 1. Literature concerning these topics was identified by separate
keyword searches conducted in Web of Science (included in Figure 10-2 as part of "References from
Other Sources").
10.2.2.2 Exposure Assessment
Exposure literature relevant to ozone was identified using the broad keyword search and
automatic topic classification, as described in Section 10.2.1. Automatic topic classification for exposure
references included a sufficiently large set of positive and negative seeds from previous ISAs. Positive
seeds included references from the exposure chapter from the 2016 ISA for Oxides of Nitrogen—Health
Criteria1 (U.S. EPA. 2016) and the 2013 Ozone ISA; the negative seeds included nonrelevant references
(i.e., those from other disciplines in these two ISAs). Following identification and sorting of the literature
by topic, SWIFT-AS was used for Level 1 screening. Positive seeds to train the SWIFT-AS algorithm
included a subset of the exposure references cited in the 2013 Ozone ISA. Additionally, references were
categorized in Level 1 screening in SWIFT-AS by study type, study location, and exposure duration.
1 The 2016 ISA for Oxides of Nitrogen—Health Criteria is the most recent ISA using automatic topic classification.
Since automatic topic classification is not pollutant specific, including references from this literature set increased
the number of seed references to improve precision of the algorithm for each discipline.
10-8
-------
10.2.2.3 Health—Experimental Studies
Experimental (i.e., controlled human exposure and animal toxicology) studies examining the
health effects of ozone exposure were identified using the broad keyword search and Automatic Topic
Classification, as described in Section 10.2.1. The Automatic Topic Classification for experimental
references encompassed a sufficiently large set of positive seeds, including controlled human exposure
and animal toxicology references cited in the 2016 ISA for Oxides ofNitrogen—Health Criteria and the
2013 Ozone ISA, and a sufficiently large set of negative seeds, including nonexperimental references
cited in these two ISAs. Following identification of the literature, SWIFT-AS was used for Level 1
screening. The SWIFT-AS algorithm was trained using a set of positive seed references from a selection
of controlled human exposure and animal toxicological studies cited in the 2013 Ozone ISA.
Additionally, references were categorized in Level 1 screening in SWIFT-AS by health outcome category
(e.g., respiratory, cardiovascular, metabolic, etc.), exposure duration (e.g., short-term, long-term), and
study type (e.g., controlled human exposure, animal toxicology, etc.).
10.2.2.4 Health—Epidemiologic Studies
Identification of recent epidemiologic studies examining a health effect and ambient exposure to
ozone was accomplished using the broad keyword search and automatic topic classification, as described
in Section 10.2.1. The approach for automatic topic classification to identify epidemiologic studies from
the broad literature search results paralleled the approach described in Section 10.2.2.3 for the
experimental studies on health effects. A sufficiently large set of seed references cited in the 2016 ISA for
Oxides ofNitrogen—Health Criteria and the 2013 Ozone ISAs was used, with positive seeds consisting
of epidemiologic references in those ISAs and negative seeds consisting of all references other than
epidemiologic references. Following identification of the literature, SWIFT-AS was used for Level 1
screening. Positive seeds were also used to train the SWIFT-AS algorithm and included select
epidemiologic references cited in the 2013 Ozone ISA. Additionally, references were categorized in
Level 1 SWIFT-AS screening by health outcome category (e.g., mortality, respiratory, cardiovascular,
etc.), exposure duration (e.g., short term, long term), and study location (e.g., U.S., Canada, Europe, etc.).
10.2.2.5 Welfare—Ecological Studies
Studies relevant to the ecological effects of ozone exposure were primarily identified by
topic-specific citation mapping in Web of Science, based on ecological studies cited in the 2013 Ozone
ISA. The broad keyword searches and automatic topic classification have traditionally resulted in a poorly
targeted set of references for Level 1 screening for ecological endpoints in past ISAs. Following citation
mapping, Level 1 screening of the identified references was conducted in SWIFT-AS, including the use
of a positive seed set of ecological references from the 2013 Ozone ISA. Screening questions to facilitate
10-9
-------
organization of the literature included effect category (e.g., foliar injury, plant growth, biodiversity, etc.),
exposure conditions, location, and ecosystem type (e.g., wetland, crop, etc.).
10.2.2.6 Welfare—Effects on Climate
Studies examining the effect of tropospheric ozone on climate were identified in two ways. First,
references were identified by topic-specific citation mapping in Web of Science using references cited in
the 2013 Ozone ISA. In addition, relevant references were identified from recent national and
international climate assessments, such as the National Climate Assessment (Wuebbles et al.. 2017). the
Intergovernmental Panel on Climate Change Fifth Assessment Report (IPCC. 2013). and other recent,
more focused reports relevant to ozone climate forcing. Level 1 screening of the identified references was
conducted in SWIFT-AS aided by a seed set of selected references from the climate section of the 2013
Ozone ISA. Screening questions to aid in organizing the literature included radiative forcing, effects on
climate, precursor and copollutant effects, and factors and feedbacks.
10.2.3 Documentation
To improve transparency, all studies identified in the literature search for ozone are documented
in the Health and Environmental Research Online (HERO) database. The HERO project page for this ISA
(https://hero.epa.gov/hero/index.cfrn/proiect/page/proiect id/2737) contains the references that were
considered for inclusion in the ISA and provides bibliographic information and abstracts. It is accessible
to the public.
All inclusion and exclusion decisions, as shown in Figure 10-2. are documented in the HERO
database. References that passed through Level 1 screening are tagged in HERO as "Title-Abstract
Screening Included." Finally, inclusion and exclusion decisions from full-text screening of references
passing through Level 1 screening are also thoroughly documented. References identified from sources
other than literature search were also screened using the same discipline-specific criteria and
inclusion/exclusion decisions for these references are also documented. All references included in the ISA
are tagged as such and are further tagged by Appendix in which the reference is cited.
10.3 Study Selection: Full-Text Evaluation of Studies (Level 2)
Following Level 1 screening, subject-matter experts from U.S. EPA's Office of Research and
Development reviewed the group of references during full-text Level 2 screening. This level of screening
evaluated studies by relevance (Section 10.3.1) and quality (Section 10.3.2).
10-10
-------
10.3.1
Relevance
Relevance was evaluated in two stages. The first stage looked at the extent to which a study was
potentially policy-relevant and informative. Such studies included those that describe or provide a basis
for the relationship between ozone exposure and health or welfare effects. Also pertinent are studies that
provide insights concerning the sources of background ozone and its concentration patterns, those that
describe innovation in measurement methods or study design, or those that present novel information on
previously unidentified effects or issues. Informative studies were not limited to specific study designs,
model systems, or outcomes.
The second stage of the relevancy review included a determination of the specific scope for the
health and welfare portions of the ISA. This stage relied on scoping tools that explicitly define the
relevant Population, Exposure, Comparison, Outcome, and Study design (PECOS) serving as criteria for
inclusion/exclusion decisions. The PECOS tool characterizes the parameters and provides a framework to
help identify literature relevant to the ISA. Discipline-specific PECOS tools were developed for
experimental studies, epidemiologic studies, ecological studies, and for studies on the effects of
tropospheric ozone on climate. PECOS tools differ by study area depending on the types of questions to
be answered and by a priori knowledge related to that question. The use of PECOS tools is widely
accepted and increasingly applied for systematic review in risk assessment. The use of these tools is
consistent with recommendations by the National Academy of Sciences for improving the design of risk
assessment through planning, scoping, and problem formulation to better meet the needs of decision
makers (NASEM. 2018V PECOS tools are identified in each of the relevant appendices in this ISA
(Appendix 3-Appendix 9). Specific details about scope and the screening questions for each discipline,
including atmospheric science and exposure assessment, are described in more detail below.
While useful for research studies that evaluate a specific health or welfare outcome, PECOS tools
were less informative for the atmospheric science and exposure assessment portions of the ISA. Full-text
(Level 2) screening of references is further described for these disciplines in the sections below.
10.3.1.1 Atmospheric Sciences
The role of the atmospheric sciences discussion in the ISA is to provide necessary context and
insight into the processes that produce ambient ozone and its variable concentration patterns for the
assessment of human and ecosystem exposure and subsequent effects. The available information
generally falls into four categories: (1) emissions measurements and inventories, (2) ambient air
measurements, (3) discussions of insights gained through laboratory chemistry studies, and
(4) discussions of ozone photochemical models and their process, and concentration estimates. The
relevance criteria applied to the body of title/abstract screened peer-reviewed literature included:
10-11
-------
1. The study addresses ozone and its precursors;
2. The study addresses the human activities, natural processes, atmospheric
chemistry and dynamics that are responsible for formation in or introduction of
ozone into the lower troposphere (planetary boundary layer) where exposure to
ozone may lead to damaging effects to human health, ecosystems, the climate
system, or other elements of the environment relevant to human welfare. The
study is also relevant if it describes spatial and temporal patterns in lower
tropospheric ozone concentrations;
3. The study addresses the origins and concentrations of ozone in the U.S.; and
4. The study provides insight into the sources, causes, and concentrations of USB
ozone, the methodologies used to estimate USB ozone, and any technical issues
that must be considered in order to correctly interpret the scientific findings
relevant to USB ozone.
The studies cited in Appendix 1 met at least one of these criteria. Any supplemental studies that
were included as necessary supporting material or studies identified outside of topic-specific citation
mapping also met at least one of these criteria.
10.3.1.2 Exposure Assessment
The ISA describes the exposure assessment methods employed in epidemiologic studies
discussed in the health appendices. For these studies, an additional screening step with added search terms
was applied to further identify potentially relevant references. The additional step was needed because the
Level 1 title/abstract screening was insufficient to gauge whether each paper covered the methods
discussed in the epidemiology literature. It was also needed because of the disparate nature of exposure
assessment references and the large number of references relative to other disciplines. The search terms
correspond to the sections of Appendix 2. Each of the references obtained through this search was then
evaluated in the Level 2 full-text screening for relevance.
Relevance was based on whether an exposure assessment study was representative of the
population and conditions addressed in the epidemiology literature. If there was sufficient evidence that a
method could provide an adequate representation of exposure from the U.S., there was no need to
consider studies conducted outside of the U.S. or Canada. If there was not sufficient evidence that a
method could provide an adequate representation of exposure from the U.S., then there was a need to
consider western European and Australian studies, which were the next most similar to studies conducted
in the U.S. If there was not sufficient evidence that a method could provide an adequate representation of
exposure from the U.S., Canada, western Europe, or Australia, then it was necessary to consider all
studies regardless of geographic location.
10-12
-------
10.3.1.3 Health—Experimental Studies
For experimental studies, specifically, controlled human or animal exposure studies, the relevance
evaluation focused on those studies with appropriate study designs and relevant exposure concentrations,
as well as those that address key uncertainties and limitations in the evidence identified in the previous
review (Table lO-l). The scope of the experimental evidence encompassed studies of short-term
(i.e., hours to weeks) and long-term (i.e., months to years) exposures conducted at concentrations of
ozone that are relevant to the range of human exposures to ambient air (up to 2 ppm, which is one to two
orders of magnitude above ambient concentrations). Review studies were identified and included in the
ISA, as appropriate, to provide context for biological plausibility figures and/or causality determinations;
however, they were not considered primary literature and thus not considered in making causality
determinations.
10-13
-------
Table 10-1 Population, Exposure, Comparison, Outcome, and Study Design
(PECOS) tool to define the parameters and provide a framework for
identifying relevant experimental studies.
Exposure Duration and Health
Effect
Population, Exposure, Comparison, Outcome, and Study Design
(PECOS)
Short-term exposure and
respiratory, cardiovascular,
metabolic, nervous system,
reproductive or developmental
effects
Population: Study populations of any controlled human exposure or animal
toxicological study of mammals at any lifestage
Exposure: Short-term (in the order of minutes to weeks) inhalation exposure
to relevant ozone concentrations (i.e., 0.4 ppm or below for humans, 2 ppm or
below for other mammals)
Comparison: Human subjects that serve as their own controls with an
appropriate washout period or when comparison to a reference population
exposed to lower levels is available, or, in toxicological studies of mammals,
an appropriate comparison group that is exposed to a negative control
(i.e., clean air or filtered air control)
Outcome: Respiratory, cardiovascular, metabolic, or nervous system;
reproductive or developmental effects
Study Design: Controlled human exposure (i.e., chamber) studies; in vivo
acute, subacute, or repeated-dose toxicity studies in mammals; reproductive
toxicity or immunotoxicity studies
Long-term exposure and
respiratory, cardiovascular,
metabolic, nervous system,
carcinogenic, reproductive or
developmental effects
Population: Study population of any animal toxicological study of mammals
at any lifestage
Exposure: Long-term (in the order of months to years) inhalation exposure to
relevant ozone concentrations (i.e., 2 ppm or below)
Comparison: Appropriate comparison group exposed to a negative control
(i.e., clean air or filtered air control)
Outcome: Respiratory, cardiovascular, metabolic or nervous system;
carcinogenic, reproductive, or developmental effects
Study Design: In vivo chronic, subchronic, or repeated-dose toxicity studies
in mammals; reproductive toxicity or immunotoxicity studies;
genotoxicity/mutagenicity studies
Population: In controlled human exposure studies, generally healthy adults approved for study participation by the
appropriate IRB or ethics committee; for toxicological studies, well-defined/well-characterized strains of mammals at
any lifestage.
Exposure: Ozone concentrations deliberately delivered to subjects for a predefined duration.
Comparison: In controlled human exposure studies, subjects serve as their own controls with an appropriate
washout period, or a reference population exposed to lower ozone concentrations, or, in toxicological studies, an
appropriate comparison group that is exposed to a negative control (i.e., clean air or filtered air control).
Outcome: Clearly measurable health endpoint.
Study Design: Controlled human exposure (i.e., chamber) studies; in vivo acute, subacute, subchronic, chronic or
repeated-dose toxicity studies in mammals; reproductive toxicity or immunotoxicity studies;
genotoxicity/mutagenicity studies.
10-14
-------
10.3.1.4 Health—Epidemiologic Studies
The evaluation of epidemiologic studies focused on the associations between short- and long-term
exposure to ozone and a range of health effects, including respiratory, cardiovascular, reproductive and
developmental, metabolic, and nervous system outcomes, as well as cancer and mortality (Table 10-2). In
instances when a "causal" or "likely to be a causal" relationship was concluded in the 2013 Ozone ISA
(i.e., short-term ozone exposure and respiratory and cardiovascular effects and total mortality, and
long-term ozone exposure and respiratory effects), the epidemiologic studies evaluated for those
outcomes were more limited in scope and targeted towards study locations that include U.S. airsheds or
airsheds that are similar to those found in the U.S., as reflected in the PECOS tool. For outcomes for
which the 2013 Ozone ISA concluded that evidence was "suggestive of' or "inadequate to infer" a causal
relationship (i.e., short-term ozone exposure and nervous system effects and long-term ozone exposure
and cardiovascular, nervous system, reproductive or developmental effects, cancer, or mortality), the
epidemiologic studies evaluated were not limited geographically or by airshed characteristics, as reflected
in the PECOS tool. Review studies, meta-analyses, and studies that used a previously published effect
estimate to predict changes in future health outcomes were identified and included in the ISA, as
appropriate, to provide context for causality determinations, but they were not considered primary
literature and thus not considered in making causality determinations.
Table 10-2 Population, Exposure, Comparison, Outcome, and Study Design
(PECOS) tool to define the parameters and provide a framework for
identifying relevant epidemiologic studies.
Exposure Duration and
Health Effect
Population, Exposure, Comparison, Outcome, and Study Design (PECOS)
Short-term exposure and
respiratory effects
Population: Any U.S. or Canadian population, including populations or lifestages that
might be at increased risk
Exposure: Short-term (on the order of 1 to several days) ambient concentration of
ozone
Comparison: Per unit increase (in ppb)
Outcome: Change in risk (incidence/prevalence) of respiratory effects
Study Design: Epidemiologic studies consisting of panel, case-crossover, time-series,
and case-control studies; cross-sectional studies with appropriate timing of exposure
for the health endpoint of interest
Short-term exposure and
mortality
Population: Any U.S. or Canadian population, including populations or lifestages that
might be at increased risk
Exposure: Short-term exposure (on the order of 1 to several days) to ambient
concentrations of ozone
Comparison: Per unit increase (in ppb)
Outcome: Change in risk (incidence) of mortality
Study Design: Epidemiologic studies consisting of case-crossover or time-series
studies with appropriate timing of exposure for the health endpoint of interest
10-15
-------
Table 10-2 (Continued): Population, Exposure, Comparison, Outcome, and Study
Design (PECOS) tool to define the parameters and
provide a framework for identifying relevant
epidemiologic studies.
Exposure Duration and
Health Effect Population, Exposure, Comparison, Outcome, and Study Design (PECOS)
Long-term exposure and Population: Any U.S. or Canadian population, including populations or lifestages that
respiratory effects might be at increased risk
Exposure: Long-term (on the order of months to years) ambient concentration of
ozone
Comparison: Per unit increase (in ppb)
Outcome: Change in risk (incidence/prevalence) of respiratory effects
Study Design: Epidemiologic studies consisting of cohort and case-control studies;
time-series, case-crossover, and cross-sectional studies with appropriate timing of
exposure for the health endpoint of interest
Population: Any U.S., Canadian, European, or Australian population, including
populations or lifestages that might be at increased risk
Exposure: Short-term (on the order of 1 to several days)_ambient concentration of
ozone
Comparison: Per unit increase (in ppb)
Outcome: Change in risk (incidence/prevalence) of cardiovascular effects
Study Design: Epidemiologic studies consisting of panel, case-crossover, time-series,
and case-control studies; cross-sectional studies with appropriate timing of exposure
for the health endpoint of interest
Population: Any population, including populations or lifestages that might be at
increased risk
Exposure: Short-term (on the order of 1 to several days) ambient concentration of
ozone
Comparison: Per unit increase (in ppb)
Outcome: Change in risk (incidence/prevalence) of a nervous system effect
Study Design: Epidemiologic studies consisting of panel, case-crossover, time-series,
and case-control studies; cross-sectional studies with appropriate timing of exposure
for the health endpoint of interest
Short-term exposure and
cardiovascular effects
Short-term exposure and
nervous system effects
10-16
-------
Table 10-2 (Continued): Population, Exposure, Comparison, Outcome, and Study
Design (PECOS) tool to define the parameters and
provide a framework for identifying relevant
epidemiologic studies.
Exposure Duration and
Health Effect Population, Exposure, Comparison, Outcome, and Study Design (PECOS)
Long-term exposure and
cardiovascular, nervous
system, reproductive, or
developmental effects;
cancer, or mortality
Population: Any population, including populations or lifestages that might be at
increased risk
Exposure: Long-term (on the order of months to years) ambient concentration of
ozone
Comparison: Per unit increase (in ppb)
Outcome: Change in risk (incidence/prevalence) of a cardiovascular, nervous system,
reproductive, or developmental effect; cancer or mortality
Study Design: Epidemiologic studies consisting of cohort and case-control studies;
time-series, case-crossover, and cross-sectional studies with appropriate timing of
exposure for the health endpoint of interest
Population: The general population, all age groups, living both in urban and in rural areas exposed on a daily basis
to ozone through outdoor (ambient) air, and not exclusively in occupational settings or as a result of indoor
exposure. Populations and lifestages at increased risk are included, such as those with specific pre-existing health
conditions (e.g., respiratory or cardiovascular diseases), children, or older adults.
Exposure: Ambient ozone from any source measured as short-term (minutes to weeks) or long-term (months to
years).
Comparison: The health effect observed by unit increase in concentration of ozone in the same or in a control
population.
Outcome: Clearly measurable health endpoint.
Study Design: Epidemiologic studies on health effects of ozone consisting of cross-sectional, case-control,
case-crossover, cohort, panel, and time-series studies.
10.3.1.5 Welfare—Ecological Studies
Similar to health effects, this ISA builds on information available during the last review
(i.e., effects of ozone exposure on vegetation and ecosystems). For research evaluating ecological effects,
emphasis was placed on recent studies that: (1) evaluated effects at ozone concentrations likely to occur in
North American airsheds and (2) investigated effects on any individual, population (in the sense of a
group of individuals of the same species), community, or ecosystem in North America (Table 10-3). In
instances when a "causal relationship" was concluded in the 2013 Ozone ISA (i.e., visible foliar injury,
vegetation growth, reduced yield/quality of agricultural crops, reduced productivity, alteration of
belowground biogeochemical cycles) the current review only evaluated studies conducted in North
America. These endpoints that were causal in the 2013 Ozone ISA have a robust evidence base, supported
by decades of research, therefore new studies were limited to North American species. For all other
ecological endpoints in Table 10-3 (terrestrial water cycling, carbon sequestration, terrestrial community
composition, plant reproduction, phenology, or mortality, insects, other wildlife, plant-animal signaling)
there are no geographic constraints, the evidence base from prior reviews either supported a "likely to be
causal" relationship or the endpoint was not evaluated for causality in the 2013 Ozone ISA, therefore, all
10-17
-------
available evidence was considered. Studies that passed through Level 2 screening included primary
literature on the endpoints and study designs identified in the PECOS tool (see Table 10-3) and
meta-analyses. Reviews were included if a comprehensive list of affected species was available;
nonagricultural North American species were separated out from the larger data sets and the evidence on
the North American species was evaluated (e.g., foliar injury, biomass).
10-18
-------
Table 10-3 Population, Exposure, Comparison, Outcome, and Study Design
(PECOS) tool to define the parameters and provide a framework for
identifying relevant ecological studies.
Population, Exposure, Comparison, Outcome, and Study Design
Ecological Endpoint (PECOS) Tool
Visible foliar injury, vegetation growth, Population: For any species, an individual, population (in the sense
yield/quality of agricultural crops, of a group of individuals of the same species), community, or
productivity, belowground biogeochemical ecosystem in North America
cyc''n9 Exposure: Concentrations occurring in the environment or
experimental ozone concentrations within an order of magnitude of
recent concentrations (as described in Appendix 1)
Comparison: Relevant control sites, treatments, or parameters
Outcome: Visible foliar injury, alteration of vegetative growth,
yield/quality of agricultural crops, productivity, belowground
biogeochemical cycles
Study Design: Laboratory, greenhouse, OTC, FACE, field, gradient,
or modeling studies
Terrestrial water cycling; carbon
sequestration; terrestrial community
composition; plant reproduction, phenology,
or mortality; insects, other wildlife,
plant-animal signaling
Comparison: Relevant control sites, treatments, or parameters
Outcome: Alteration of terrestrial water cycling; carbon
sequestration; terrestrial community composition; plant reproduction,
phenology, mortality; growth, reproduction, and survival of insects
and other wildlife; plant-animal signaling
Study Design: Laboratory, greenhouse, OTC, FACE, field, gradient,
or modeling studies
Population: For any species, an individual, population (in the sense
of a group of individuals of the same species), community, or
ecosystem in any continent3
Exposure: Concentrations occurring in the environment or
experimental ozone concentrations within an order of magnitude of
recent concentrations (as described in Appendix 1)
Population: Unit of study.
Exposure: Environmental variable to which population is exposed.
Comparison: Change in endpoint observed by unit increase in concentration of ozone in the same or in a control
population.
Outcome: Measurable endpoint resulting from exposure.
Study Design: Laboratory, field, gradient, open-top chamber (OTC), free-air carbon dioxide enrichment (FACE),
greenhouse, and modeling studies.
Notes: This definition of population is for the purpose of applying PECOS to ecology. Ecological populations are defined as a
group of individuals of the same species.
aln cases where a comprehensive list of affected species was available, nonagricultural North American species were separated
out from the larger data sets and the evidence on the North American species was evaluated (e.g., foliar injury, biomass).
10-19
-------
10.3.1.6 Welfare—Effects on Climate
For effects on climate, the ISA focused on effects of tropospheric ozone on climate, consistent
with the inclusion of "climate" in the list of effects on welfare in Section 302(h) of the Clean Air Act. The
ISA does not focus on downstream ecosystem effects from changes in climate, climate-related human
health effects, or future air quality projections resulting from changes in climate. In addition, the ISA
assessed the available evidence on the effects of tropospheric ozone as an absorber of UV-B radiation.
Studies that assess the independent role of ozone in climate forcing as well as its effects on U.S. national
and regional climate are within the scope of the literature considered in the review (Table 10-4).
Table 10-4 Population, Exposure, Comparison, Outcome, and Study Design
(PECOS) tool to define the parameters and provide a framework for
identifying relevant studies on the effects of tropospheric ozone on
climate.
Population, Exposure, Comparison, Outcome, and Study Design
Effect on Climate (PECOS)
Changes in radiative forcing (RF) Population/Geographical Scope: Evaluations of radiative forcing at the
regional, continental, and/or global scale
Exposure: Tropospheric ozone concentration distributions in 3D
(observed/modeled)
Comparison: Relevant baseline or unperturbed scenarios/conditions
Outcome: Changes in RF resulting from change in tropospheric ozone
Study Design: Observations or modeling studies
Changes in climate (e.g., surface Population/Geographical Scope: Evaluations of climate effects at the
temperature, hydrological cycle) regional, continental, and/or global scale
Exposure: Tropospheric ozone concentration distributions in 3D
(observed/modeled)
Comparison: Relevant baseline or unperturbed scenarios/conditions
Outcome: Subsequent climate effects (via radiative forcing) (e.g., global
surface temperature) resulting from change in tropospheric ozone
Study Design: Observational or modeling studies
Population/geographical scope: Spatial extent of study.
Exposure: Environmental variable (tropospheric O3 concentrations).
Comparison: Radiative forcing or climate effects observed from unit change in tropospheric O3 concentration.
Outcome: Relevant radiative forcing or climate outcomes resulting from change in tropospheric O3.
Study Design: Observations/satellite, modeling.
10-20
-------
10.3.2
Individual Study Quality
After selecting studies for inclusion based on relevance, individual study quality was evaluated by
considering the design, methods, conduct, and documentation of each study, but not the study results. For
ISAs, the overall individual study quality evaluation process is described in the Preamble to the ISA (U.S.
EPA. 2015b). which outlines a base set of questions for consideration when evaluating the scientific
quality of studies, intended for use in both human health and ecological studies:
• Were the study designs, study groups, methods, data, and results clearly presented in relation to
the study objectives to allow for study evaluation? Were limitations and any underlying
assumptions of the design and other aspects of the study stated?
• Were the ecosystems, study site(s), study populations, subjects, or organism models adequately
selected, and are they adequately defined to allow for meaningful comparisons between study or
exposure groups?
• Are the air quality, exposure, or dose metrics of adequate quality and are they sufficiently
representative of or pertinent to ambient air?
• Are the welfare effect measurements meaningful, valid, and reliable?
• Were likely covariates or modifying factors adequately controlled or taken into account in the
study design and statistical analysis?
• Do the analytical methods provide adequate sensitivity and precision to support conclusions?
• Were the statistical analyses appropriate, properly performed, and properly interpreted?
The process for individual study quality evaluation has been refined by discipline with each
successive ISA. Prior to evaluating the evidence for the health disciplines (i.e., experimental and
epidemiology), study quality criteria tables were developed to define important aspects to consider in
evaluating if a study addresses issues specific to ozone in a proper manner. Study quality criteria tables
have been developed for each of the most recent ISAs (U.S. EPA. 2018. 2017. 2016) and are pollutant
specific. Study quality criteria tables mirror the process described in the Preamble and serve as the
foundation for review of health studies. Ecological studies were evaluated for study quality using a broad
set of questions from the Preamble, while other disciplines (e.g., atmospheric sciences) used similar
approaches appropriate for and consistent with the fields of study. The processes used are described in the
sections below.
Study quality was a final step in Level 2 screening in deciding whether to include a study in the
ISA. Any references that did not pass the study quality review, and deemed critically deficient, were
excluded from the ISA. Any study that passed both the relevance screening and the study quality
evaluation was included in the ISA. The combination of approaches described above are intended to
produce a comprehensive collection of pertinent studies needed to address the key scientific issues that
form the basis of the ISA.
10-21
-------
10.3.2.1 Atmospheric Science
The topics relevant to the atmospheric science of ground-level ozone formation include estimated
emissions of ozone precursors by anthropogenic, natural (i.e., nonanthropogenic) and international
precursor sources, photochemical formation mechanisms, atmospheric dynamics, photochemical
modeling, ambient ozone measurement methods, and the temporal and spatial concentration patterns of
anthropogenic and U.S. background ozone. Study quality in this context is established using the following
criteria: (1) quality assurance and quality control standards were applied in the development of the data
sets used to create figures showing national precursor emissions and ground-level ozone concentrations
and (2) studies describing new understanding of the relevant precursor emissions, atmospheric processes,
and photochemical modeling of the formation of ozone underwent peer review. Peer review represents the
current standard for establishing study quality in the atmospheric sciences.
10.3.2.2 Exposure Assessment
Exposure assessment studies that passed the PECOS relevance screening were then evaluated for
study quality. For exposure assessment methodology studies, study quality was deemed adequate if the
exposure assessment methods were clearly described, the selected exposure assessment methods were
appropriate for the research question evaluated, the assumptions of the method(s) were clearly stated, the
uncertainties and limitations of the methods were clearly stated, and if quality assurance testing had been
performed. These requirements were based on studies in the literature that provided good examples of
study quality evaluation for exposure assessment methods (Zou et al.. 2009; Ryan and Lemasters. 2007;
Nieuwenhuiisen et al.. 2006; Gilliland et al.. 2005; Jerrett et al.. 2005). Validation techniques for quality
assurance testing varied from study to study but often included some form of comparison with Federal
Reference Method monitoring data.
10.3.2.3 Health Approach
Consistent with the Preamble, conclusions across the body of evidence are made after
independently evaluating the overall quality of each study. This uniform approach considers the strengths,
limitations, and possible roles of chance, confounding, and other biases that may influence the results of
the study.
The evaluation of individual study quality in ISAs has evolved over the past several ISAs. U.S.
EPA developed the Preamble in 2015 based on input and feedback from numerous reviews by the Clean
Air Scientific Advisory Committee (CASAC) over several years. In recent ISAs (U.S. EPA. 2018. 2017.
2016). study quality tables for human health were developed to provide greater clarity on important
aspects of study quality. These tables describe the characteristics of multiple study domains (e.g., study
10-22
-------
design, exposure assessment, potential confounding) that can increase or decrease confidence in the study
results.
For this ISA, the study quality tables were tailored to address factors specific to health studies of
ozone exposure (Table Annex 3-1). For example, the study quality table for epidemiologic studies of
ozone describes the potential confounding factors, copollutants, lags, averaging times, and seasonal
considerations that are relevant to the analysis of ozone and health effects.
To further facilitate the study quality evaluation of health studies, prompting questions were
developed for each of the study domains included in the study quality table. These prompting questions
were designed to assist in the narrative documentation of study quality for each of the individual domains,
ensuring consistent information is included across reviewers. The specific prompting questions are
different for epidemiology and experimental (i.e., animal toxicology and controlled human exposure)
studies, reflecting differences in relevant study quality factors between the two disciplines. The goal of
the narrative approach is to provide nuanced and transparent documentation of the strengths and
limitations that support expert judgement on the inclusion/exclusion decisions for individual studies.
Narrative reviews were completed for the most policy-relevant studies,1 and the completed reviews for
each of the selected studies were recorded in the Health Assessment Workspace Collaborative (HAWC)
database and can be accessed from the HERO project page (see Section 10.3.3V The HAWC project page
also contains that guidance text and prompting questions (EPA. 2019).
10.3.2.4 Ecological Approach
Generally, the field of study quality evaluation is much more robust for human health research
than for ecological research. Study quality is still very important for ecological research, and U.S. EPA
staff have relied on the criteria listed in the Preamble (U.S. EPA. 2015b) for reviewing the quality of
individual ecological studies within this ISA. U.S. EPA also relied on discussions in previous ISAs about
the strengths and weaknesses of various ecological study designs (see Section 8.1.2.1). A limited number
of studies were excluded based on consideration of the study quality questions in the Preamble and
application of the PECOS tool. The main reasons that studies were eliminated were due to insufficient or
low replication, ozone exposures were unable to be determined, lack of statistical testing for endpoints of
interest, methods were inadequately described, selective reporting of data, flaws in study design, or a
model was based on flawed data or incorrect assumptions.
1 Studies reviewed included those related to health effects for which there was a "causal" or "likely to be a causal"
relationship in the 2013 Ozone ISA, or for which the determination about the causal relationship changed from the
determination made in the 2013 Ozone ISA.
10-23
-------
10.3.3
Documentation
During the review of references for Level 2 screening, inclusion and exclusion decisions were
carefully documented. The inclusion/exclusion results are recorded in HERO, while specific data about
concentrations, experimental design, and results are reported within the appendices. Reference-specific
information about study quality are documented in HAWC for select health studies and can be accessed
via the HERO project page associated with this ISA (EPA. 2019). All decisions about Level 2 screening,
including both relevance and quality, are documented in the HERO database and on the publicly available
HERO project page for this ISA.
10.4 Peer Review and Public Participation
Peer review is an important component of any scientific assessment. U.S. EPA has formal
guidance about peer review in the Peer Review Handbook (U.S. EPA. 2015a'). and this ISA follows all the
policies and procedures identified therein. Additionally, this ISA follows all the guidelines of the
Information Quality Guidelines (U.S. EPA. 2002).
U.S. EPA has designated this ISA as a Highly Influential Scientific Assessment, which is defined
by The Office of Management and Budget's Final Information Quality Bulletin for Peer Review
(hereafter, "Peer Review Bulletin") as:
A subset of Influential Scientific Information that is a scientific assessment (i.e., an
evaluation of a body of scientific or technical knowledge, which typically synthesizes
multiple factual inputs, data, models, assumptions and/or applies best professional
judgment to bridge uncertainties in the available information) that "could have a potential
impact of more than $500 million in any year on either the public or private sector" or "is
novel, controversial, or precedent-setting, or has significant interagency interest."
(https://obamawhitehouse.archives.gov/omb/memoranda fV2005 m05-03/).
As such, there are additional review and transparency steps required in the release of this
information. These steps are described below. CASAC also serves an important role in reviewing this ISA
(see Section 10.4.5V
10.4.1 Call for Information
Consistent with the Preamble, a Call for Information was published in the Federal Register on
June 26, 2018 (83 FR 29785). The purpose of this Call for Information was announcing the beginning of
the review cycle of the air quality criteria and the ozone NAAQS. Specifically, the Call for Information
10-24
-------
stated that U.S. EPA would be preparing an Integrated Review Plan and Integrated Science Assessment.
The public was given 30 days "...to assist the U.S. EPA by submitting information regarding significant
new ozone research and policy-relevant issues for consideration in this review of the primary
(health-based) and secondary (welfare-based) ozone standards." U.S. EPA received 14 comments via the
Federal eRulemaking Portal (http://www.regulations.gov).
In previous assessments, U.S. EPA has held a science and policy issue workshop to "kick-off' the
review cycle. The workshop brought together subject area experts and the public to highlight significant
new and emerging research and make recommendations to the U.S. EPA regarding the design and scope
of the review. However, certain process efficiencies were discussed in the Administrator's May 9, 2018
memorandum, "Back-to-Basics Process for Reviewing National Ambient Air Quality Standards."
(https://www.epa.gov/sites/production/files/2018-05/documents/image2018-05-09-173219.pdf). In
particular, the Administrator called for ".. .efficiencies and opportunities to streamline the NAAQS
review process to ensure they finish within a 5-year interval. For the next review of the ozone NAAQS,
U.S. EPA shall seek efficiencies through replacing the kick-off workshop with a more robust request for
information...." As a result, no science and policy issue workshop was held and the Call for Information
served as the only initiating event for this NAAQS review.
10.4.2 Integrated Review Plan
Following the Call for Information, U.S. EPA prepared an Integrated Review Plan (IRP) that
summarized the current plan for this NAAQS review, a projected timeline, and the process for conducting
the review. The IRP also identifies key policy-relevant issues or questions intended to guide the review.
The draft IRP for this review cycle was announced in the Federal Register on November 2, 2018 (83 FR
55163) for consultation with the CASAC and for public comment. The public was given 30 days to
respond, and U.S. EPA received 13 comments via the Federal eRulemaking Portal
(http: //www. regulations. gov).
A CASAC consultation was held in a public meeting on November 29, 2018, and documentation
of that occurrence along with written comments from individual CASAC members were sent to the U.S.
EPA Administrator in a letter dated December 10, 2018
(https://vosemite.epa.gov/sab/sabproduct.nsf/A286A0F015 lDC8238525835F007D348A/$File/EPA-
CASAC-19-001 .pdf). The final IRP was prepared in consideration of CASAC and public comments and
released in August 2019.
10.4.3 Peer Input
The role of peer input is described in the Preamble, as well as the Peer Review Handbook (U.S.
EPA. 2015a. b). After a thorough literature search and screening process, U.S. EPA staff developed
10-25
-------
preliminary drafts of all appendices for initial peer input. Peer input is a process that allows U.S. EPA to
solicit early-in-the-process feedback from subject-matter experts from within and external to the U.S.
EPA to ensure that the ISA is up-to-date and focused on the most policy-relevant findings. This review
also assists U.S. EPA with integration of evidence within and across disciplines. Peer input serves as a
supplement to other peer-review mechanisms and does not take the place of a thorough external peer
review.
For this ISA, U.S. EPA worked with ICF International to run the peer input process. Following
the guidelines in the peer-review handbook, U.S. EPA prepared a list of expertise needed for a first
review of the ISA material. ICF was solely responsible for inviting 24 Reviewers across four discipline
areas to participate in the peer input workshops and for coordinating all communication for this
consultation. A Federal Register Notice was issued on October 23, 2018 announcing the workshops and
included information for public access to the discussions (83 FR 53472). Preliminary draft materials were
made available to the expert peer reviewers; they were not publicly disseminated, and no formal public
comment was held at this early stage.
U.S. EPA developed a charge for the reviewers that included questions specific to each discipline,
as well as three overarching prompts:
• Please comment on the extent to which the initial draft materials capture the key studies from the
peer-reviewed literature that have been published since the completion of the 2013 Ozone ISA.
As noted in the following Session areas, please identify additional studies published since the
2013 Ozone ISA that should be included.
• The review of evidence is intended to be concise and coherent, not encyclopedic. To that end,
please comment on the extent to which the scientific information is accurately characterized and
with the appropriate level of detail in draft materials. Are there any additional topics, endpoints,
factors, etc. that should be included?
• Are there specific issues that should be considered or highlighted that will be important for
integrating evidence across disciplines?
Reviewers provided written responses to U.S. EPA both prior to the workshop discussion and a
revised version afterwards. Public webinar workshops were held during October and November 2018.
Following these consultation workshops, U.S. EPA considered reviewer comments and literature
suggestions to revise the preliminary drafts of the appendices.
10.4.4 Internal Technical Review
The U.S. EPA Office of Research and Development guidelines require an internal technical
review process prior to any external dissemination of scientific information. Consistent with this policy,
the draft ISA was reviewed by U.S. EPA subject-matter experts, both those who had been involved in
developing the draft document and those who had not. Following the technical review, U.S. EPA used the
reviewers' comments to revise the document.
10-26
-------
10.4.5
Clean Air Scientific Advisory Committee (CASAC) Peer Review
Two sections of the Clean Air Act, Sections 108 and 109 [42 U.S.C. 7408 and 7409], govern the
periodic review and establishment of the NAAQS (CAA. 2005). With respect to CASAC,
Section 109(d)(2) addresses the appointment and advisory functions of an independent scientific review
committee. Section 109(d)(2)(A) requires the Administrator to appoint this committee, which is to be
composed of "seven members including at least one member of the National Academy of Sciences, one
physician, and one person representing State air pollution control agencies." Section 109(d)(2)(B)
provides that the independent scientific review committee periodically "shall complete a review of the
criteria... and the national primary and secondary ambient air quality standards... and shall recommend to
the Administrator any new... standards and revisions of existing criteria and standards as may be
appropriate...." Since the early 1980s, this independent review function has been performed by the
CASAC of the U.S. EPA's Science Advisory Board.
The CASAC serves as the official peer review mechanism for this ISA. As a Highly Influential
Scientific Assessment, the review process is also governed by the Peer Review Bulletin. All requirements
in the Peer Review Bulletin about selection of reviewers, information access, opportunity for public
participation, transparency, and management of peer-review process and reviewer selection have been
met for the CASAC review of this ISA.
A draft ISA document was released for public comment and review on September 26, 2019, and a
Federal Register Notice [84 FR 50836] was published on September 26, 2019 to announce availability of
the draft ISA and seek public comment. A separate Federal Register Notice [84 FR 58713] was published
on November 1, 2019 announcing the public meeting for CASAC review of the draft ISA. CASAC met
December 3-6, 2019 to discuss the draft ISA and provide an independent scientific peer review of the
draft ISA. A draft of the CASAC response to U.S. EPA was shared, and the Committee held
teleconferences February 11-12, 2020 to further discuss their consensus. The final CASAC response was
sent to the U.S. EPA Administrator on February 21, 2020 and contained both consensus comments and
individual reviewer comments
(https://vosemite.epa.gov/sab/sabproduct.nsf/264cbl227d55e02c852574020Q7446a4/F228E5D4D848BB
ED85258515006354D0/$File/EPA-CASAC-20-002.pdf). The letter from the CASAC, as well as public
comments received on the draft Ozone ISA, can be found in Docket ID No. EPA-HQ-ORD-2018-0274.
The Administrator responded to the CASAC's letter on the External Review Draft of the Ozone
ISA on April 1, 2020, and the letter is available at:
https ://vosemite .epa. gov/sab/sabproduct.nsf/LookupWebReportsARsLastMonthCAS AC/
F228E5D4D848BBED85258515006354D0/$File/EPA-CASAC-20-002 Response.pdf. Administrator
Wheeler's letter to the CASAC indicated the Agency will "incorporate the CASAC's comments and
recommendations, to the extent possible, and create a final Ozone ISA so that it may be available to
inform a proposed decision on any necessary revisions of the NAAQS by the spring of 2020." The
consensus CASAC comments on the draft Ozone ISA (February 19, 2020) recommended that the Draft
Ozone ISA would benefit from:
10-27
-------
(1) critical review, synthesis, and discussion of available scientific evidence; (2) reassessment of causality
determinations and rationale for some new and altered causality determinations; and (3) consultation with
outside experts on high-level, over-arching process aspects related to ISA development and consideration
of causality. To address these comments while preparing the Final Ozone ISA, the EPA added new text
and clarified existing text in the Preface and in Appendix 10 to more clearly articulate how scientific
evidence is identified, evaluated and summarized in the ISA, revised the causality determination for long-
term ozone exposure and metabolic effects, and will take steps to both identify methods for improving the
ISA process and to solicit outside expertise on best practices for making causality determinations from
multiple lines of evidence. Additionally, the U.S. EPA focused on addressing those comments that
contributed to improving clarity, could be addressed in the near-term, and identified errors in the draft
Ozone ISA. Lastly, Administrator Wheeler noted, "for those comments and recommendations that are
more significant or cross-cutting and which were not fully addressed, the Agency will develop a plan to
incorporate these changes into future Ozone ISAs as well as ISAs for other criteria pollutant reviews."
10.5 Quality Assurance
The use of quality assurance (QA) helps ensure that the U.S. EPA conducts high-quality science
that can be used to inform policymakers, industry, and the public. Agency-wide, the U.S. EPA Quality
System provides the framework for planning, implementing, documenting, and assessing work performed
by the Agency, and for carrying out required quality assurance and quality control (QA/QC) activities.
Additionally, the Quality System covers the implementation of the U.S. EPA Information Quality
Guidelines (U.S. EPA. 2002). This ISA follows all Agency guidelines to ensure a high-quality document.
Within the U.S. EPA, Quality Assurance Project Plans (QAPPs) are developed to ensure that all
Agency materials meet a high standard for quality. U.S. EPA has developed a Program-level QAPP
(PQAPP) for the ISA Program to describe the technical approach and associated QA/QC procedures
associated with the ISA Program. All QA objectives and measurement criteria detailed in the PQAPP
have been employed in developing this ISA.
Quality assurance checks were conducted on numerical entries used in the appendices, and at a
minimum, the numbers obtained from every tenth reference cited in the appendices were verified against
the original source by an independent scientist for accuracy. Furthermore, publicly available databases
(e.g., National Emissions Inventory, Air Quality System database) from which data was used in analyses
were verified to have their own QA processes in place.
U.S. EPA QA staff are responsible for the review and approval of all quality-related
documentation. Because this is a Highly Influential Scientific Assessment (see Section 10.4). U.S. EPA
QA staff performed a Technical System Audit on the ISA in September 2019. This audit verified that the
appropriate QA procedures, criteria, reviews, and data verification were adequately performed and
documented.
10-28
-------
10.6 Conclusion
This Appendix describes the criteria used to select, evaluate, and summarize studies cited in the
ISA and provides details of how the criteria were applied in developing this ISA. Overall, U.S. EPA has a
robust set of policies and procedures in place to ensure the highest quality products. In developing this
ISA, the U.S. EPA has followed all current processes and endeavored to add additional steps as
practicable and needed (e.g., use of text mining applications, PECOS tools, and documentation of
individual study quality).
10-29
-------
10.7 References
CAA (Clean Air Act). (2005). Clean Air Act, section 302: Definitions, 42 USC 7602.
http://www.gpo.gov/fdsvs/pkg/USCODE-2005-title42/pdf/USCQDE-2005-title42-chap85-
subchapIII-sec7602 .pdf
Cohen. AM; Hersh. WR; Peterson. K; Yen. PY. (2006). Reducing workload in systematic review
preparation using automated citation classification. J Am Med Inform Assoc 13: 206-219.
http://dx.doi.org/10.1197/iamia.M1929
EPA (Environmental Protection Agency). (2019). Ozone ISA Study Quality Evaluations - Health.
Retrieved from https://hawcprd.epa.gov/studv/assessment/100500Q31/
Gaudel. A: Cooper. OR: Ancellet. G; Barret. B: Bovnard. A: Burrows. JP; Clerbaux. C; Coheur. PF:
Cuesta. J: Cuevas. E: Doniki. S: Dufour. G: Eboiie. F: Foret. G: Garcia. O: Granados Munos. MJ:
Hannigan. J: Hase. F; Huang. G: Hassler. B; Hurtmans. D; Jaffe. D; Jones. N: Kalabokas. P;
Kerridge. B; Kulawik. SS: Latter. B; Leblanc. T; Le Flochmoen. E; Lin. W: Liu. J: Liu. X: Mahieu.
E: McClure-Beglev. A: Neu. JL: Osman. M: Palm. M: Petetin. H: Petropavlovskikh. I: Ouerel. R:
Rahpoe. N: Rozanov. A: Schultz. MG: Schwab. J: Siddans. R: Smale. D: Steinbacher. M;
Tanimoto. H: Tarasick. D; Thouret. V: Thompson. AM: Trickl. T; Weatherhead. E: Wespes. C:
Worden. H: Vigouroux. C: Xu. X: Zeng. G: Ziemke. J. (2018). Tropospheric ozone assessment
report: Present-day distribution and trends of tropospheric ozone relevant to climate and global
atmospheric chemistry model evaluation. Elementa: Science of the Anthropocene 6: 39.
http://dx.doi.org/10.1525/elementa.291
Gilliland. F: Avol. E: Kinney. P: Jerrett. M: Dvonch. T: Lurmann. F: Buckley. T: Brevsse. P: Keeler.
G: de Villiers. T; Mcconnell. R. (2005). Air pollution exposure assessment for epidemiologic
studies of pregnant women and children: Lessons learned from the Centers for Children's
Environmental Health and Disease Prevention Research. Environ Health Perspect 113: 1447-1454.
http://dx.doi.org/10.1289/ehp.7673
Howard. BE: Phillips. J: Miller. K; Tandon. A: Mav. D; Shah. MR: Holmgren. S: Pelch. KE; Walker.
V: Roonev. AA: Macleod. M: Shah. RR: Thaver. K. (2016V SWIFT-Review: a text-mining
workbench for systematic review. Syst Rev 5: 87. http://dx.doi.Org/10.l 186/sl3643-016-0263-z
IPCC (Intergovernmental Panel on Climate Change). (2013). Climate change 2013: The physical
science basis. Contribution of working group I to the fifth assessment report of the
Intergovernmental Panel on Climate Change. In TF Stacker; D Qin; GK Plattner; MMB Tignor;
SK Allen; J Boschung; A Nauels; Y Xia; V Bex; PM Midgley (Eds.). Cambridge, UK: Cambridge
University Press, http://www.ipcc.ch/report/ar5/wg 1 /
Jaffe. DA: Cooper. OR: Fiore. AM: Henderson. BH: Tonnesen. GS: Russell. AG: Henze. DK:
Langford. AO: Lin. M: Moore. T. (2018). Scientific assessment of background ozone over the US:
Implications for air quality management. Elementa: Science of the Anthropocene 6.
http://dx.doi.org/10.1525/elementa.309
Jerrett. M; Arain. A; Kanaroglou. P; Beckerman. B; Potoglou. D; Sahsuvaroglu. T; Morrison. J;
Giovis. C. (2005). A review and evaluation of intraurban air pollution exposure models [Review]. J
Expo Anal Environ Epidemiol 15: 185-204. http://dx.doi.org/10.1038/si.iea.7500388
NASEM (National Academies of Sciences, Engineering, and Medicine). (2018). Progress toward
transforming the Integrated Risk Information System (IRIS) program. A 2018 evaluation.
Washington, DC: The National Academies Press, http://dx.doi.org/10.17226/25086
10-30
-------
Nieuwenhuiisen. M; Paustenbach. D; Duarte-Davidson. R. (2006). New developments in exposure
assessment: The impact on the practice of health risk assessment and epidemiological studies
[Review]. Environ Int 32: 996-1009. http://dx.doi.Org/10.1016/i.envint.2006.06.015
Ryan. PH; Lemasters. GK. (2007). A review of land-use regression models for characterizing
intraurban air pollution exposure [Review]. Inhal Toxicol 19: 127.
http://dx.doi.org/10.1080/089583707Q1495998
U.S. EPA (U.S. Environmental Protection Agency). (2002). Guidelines for ensuring and maximizing
the quality, objectivity, utility, and integrity of information disseminated by the Environmental
Protection Agency. (EPA/260/R-02/008). Washington, DC: U.S. Environmental Protection
Agency, Office of Environmental Information, https ://www.epa.gov/sites/production/files/2017-
03/documents/epa-info-qualitv-guidelines.pdf
U.S. EPA (U.S. Environmental Protection Agency). (2013). Integrated science assessment for ozone
and related photochemical oxidants [EPA Report]. (EPA/600/R-10/076F). Research Triangle Park,
NC: U.S. Environmental Protection Agency, Office of Research and Development, National Center
for Environmental Assessment-RTP Division.
http://cfbub.epa.gov/ncea/isa/recordisplav.cfm?deid=247492
U.S. EPA (U.S. Environmental Protection Agency). (2015a). Peer review handbook [EPA Report] (4th
ed.). (EPA/100/B-15/001). Washington, DC: U.S. Environmental Protection Agency, Science
Policy Council, https://www.epa.gov/osa/peer-review-handbook-4th-edition-2015
U.S. EPA (U.S. Environmental Protection Agency). (2015b). Preamble to the integrated science
assessments [EPA Report]. (EPA/600/R-15/067). Research Triangle Park, NC: U.S. Environmental
Protection Agency, Office of Research and Development, National Center for Environmental
Assessment, RTP Division. https://cfbub.epa.gov/ncea/isa/recordisplav.cfm?deid=310244
U.S. EPA (U.S. Environmental Protection Agency). (2016). Integrated science assessment for oxides
of nitrogen-health criteria (final report) [EPA Report]. (EPA/600/R-15/068). Research Triangle
Park, NC: U.S. Environmental Protection Agency, Office of Research and Development, National
Center for Environmental Assessment.
http://ofinpub.epa.gov/eims/eimscomm.getfile7p download id=526855
U.S. EPA (U.S. Environmental Protection Agency). (2017). Integrated science assessment for sulfur
oxides: Health criteria [EPA Report]. (EPA/600/R-17/451). Research Triangle Park, NC: U.S.
Environmental Protection Agency, Office of Research and Development, National Center for
Environmental Assessment- RTP. https://www.epa.gov/isa/integrated-science-assessment-isa-
sulfur-oxides-health-criteria
U.S. EPA (U.S. Environmental Protection Agency). (2018). U.S. EPA. Integrated Science Assessment
(ISA) for Particulate Matter (External Review Draft). (EPA/600/R-18/179). Washington, DC.
https://cfpub.epa.gov/ncea/isa/recordisplav.cfm?deid=341593
Wuebbles. DJ: Fahev. DW: Hibbard. KA; Dokken. DJ; Stewart. BC: Mavcock. TK. (2017). USGCRP,
2017: Climate science special report: Fourth national climate assessment, Volume I. Washington,
DC: U.S. Global Change Research Program, https://science2017.globalchange.gov/
Zou. B; Wilson. JG: Zhan. FB; Zeng. Y. (2009). Air pollution exposure assessment methods utilized
in epidemiological studies [Review]. J Environ Monit 11: 475-490.
http://dx.doi.org/10.1039/b813889c
10-31
-------
&EPA
United States
Environmental Protection
Agency
Office of Research and Development
National Center for Environmental Assessment,
Research Triangle Park, NC
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA
PERMIT NO. G-35
Official Business
Penalty for Private Use
$300
------- |