United States
Environmental Protection
^^Inal B % Agency
EPA/600/R-19/093
September 2019
www.epa.gov/isa
Integrated Science
Assessment for Ozone and
Related Photochemical
Oxidants
(External Review Draft)
September 2019
National Center for Environmental Assessment—RTP Division
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC
-------
Disclaimer
This document is an external review draft, for review purposes only. This information is
distributed solely for the purpose of predissemination peer review under applicable information quality
guidelines. It has not been formally disseminated by EPA. It does not represent and should not be
construed to represent any Agency determination or policy. Mention of trade names or commercial
products does not constitute endorsement or recommendation for use.
-------
Contents
LIST OF TABLES x
LIST OF FIGURES xxv
INTEGRATED SCIENCE ASSESSMENT TEAM, FOR OXIDES OF NITROGEN xxxiii
AUTHORS, CONTRIBUTORS, AND REVIEWERS xxxv
ACRONYMS AND ABBREVIATIONS xxxix
PREFACE I xx
Legislative Requirements for the Review of the National Ambient Air Quality Standards Ixx
History of the Reviews of the Primary and Secondary National Ambient Air Quality Standard
for Ozone Ixxii
Purpose and Overview of the Integrated Science Assessment Ixxvii
Process for Developing Integrated Science Assessments Ixxviii
Scope of the ISA Ixxx
Evaluation of the Evidence Ixxxi
References for Preface Ixxxiv
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-3
ES.4 Health and Welfare Effects of Ozone Exposure ES-4
ES.4.1 Health Effects of Ozone Exposure ES-4
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 for Executive Summary ES-20
INTEGRATED SYNTHESIS IS-1
15.1 Introduction IS-2
15.1.1 Purpose and Overview IS-2
15.1.2 Process and Development IS-3
15.1.3 New Evidence Evaluation and Causality Determinations IS-6
15.2 Atmospheric Chemistry, Ambient Air Ozone Concentrations, and Background Ozone IS-11
15.2.1 Ambient Air Ozone Anthropogenic Sources, Measurement, and Concentrations IS-11
15.2.2 Background Ozone IS-13
15.3 Exposure to Ambient Ozone IS-16
15.3.1 Human Exposure Assessment in Epidemiologic Studies IS-16
15.3.2 Ecological Exposure IS-18
15.4 Evaluation of the Health Effects of Ozone IS-19
15.4.1 Connections among Health Effects IS-19
15.4.2 Biological Plausibility IS-20
15.4.3 Summary of Health Effects Evidence IS-22
15.4.4 At-Risk Populations IS-50
-------
CONTENTS (Continued)
IS. 5 Evaluation of Welfare Effects of Ozone IS-63
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-89
15.7 References for Integrative Synthesis IS-94
APPENDIX 1 ATMOSPHERIC SOURCE, CHEMISTRY, METEOROLOGY, TRENDS,
AND BACKGROUND OZONE 1-1
1.1 Overview 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 1-8
1.3.2 Stratosphere-Troposphere Exchange Processes 1-23
1.4 Ozone Photochemistry 1-25
1.4.1 Winter Ozone in Western Intermountain Basins 1-26
1.4.2 Halogen Chemistry 1-28
1.5 Inter-annual Variability and LongerTerm Trends in Meteorological Effects on Anthropogenic
and U.S. Background (USB) Ozone 1-29
1.5.1 Meteorological Effects on Ozone Concentrations at the Ground Level 1-30
1.5.2 Inter-annual and Multidecadal Climate Variability 1-31
1.6 Measurements and Modeling 1-33
1.6.1 Advances in Ozone Measurement Methods 1-33
1.6.2 Advances in Regional Chemical Transport Modeling 1-35
1.7 Ambient Air Concentrations and Trends 1-36
1.8 U.S. Background Ozone Concentrations 1-49
1.8.1 Modeling Strategies Applied to Estimate U.S. Background Ozone 1-49
1.8.2 Concentrations and Trends of U.S. Background (USB) and Baseline Ozone 1-54
1.9 Summary 1-61
1.10 References 1-63
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 2-13
2.4.1 Time-Activity Data 2-13
2.4.2 Infiltration 2-21
2.4.3 Relationships between Personal Exposure and Ambient Concentration 2-29
2.5 Copollutant Correlations and Potential for Confounding 2-32
2.6 Interpreting Exposure Measurement Error for Use in Epidemiology Studies 2-34
2.6.1 Short-Term Exposure 2-35
2.6.2 Long-Term Exposure 2-40
2.7 Conclusions 2-52
2.8 Evidence Inventories—Data Tables to Provide Supporting Information 2-54
2.9 References 2-156
-------
CONTENTS (Continued)
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 Respiratory Effects in Populations with Asthma 3-39
3.1.6 Respiratory Effects in Other Populations with Pre-existing Conditions 3-53
3.1.7 Respiratory Infection and other Associated Health Effects 3-61
3.1.8 Combinations of Respiratory Related Hospital Admissions and Emergency
Department (ED) Visits 3-64
3.1.9 Respiratory Mortality 3-68
3.1.10 Relevant Issues for Interpreting Epidemiologic Evidence—Short-Term Ozone
Exposure and Respiratory Effects 3-68
3.1.11 Summary and Causality Determination 3-77
3.2 Long-Term Ozone Exposure 3-85
3.2.1 Introduction, Summary from the 2013 Ozone ISA, and Scope for Current Review 3-85
3.2.2 Population, Exposure, Comparison, Outcome, and Study Design (PECOS) Tool 3-86
3.2.3 Biological Plausibility 3-87
3.2.4 Respiratory Health Effects 3-90
3.2.5 Relevant Issues for Interpreting Epidemiologic Evidence—Long-Term Ozone
Exposure and Respiratory Effects 3-108
3.2.6 Summary and Causality Determination 3-109
3.3 Evidence Inventories—Data Tables to Summarize Study Details 3-116
3.3.1 Short-Term Exposure 3-116
3.3.2 Long-Term Exposure 3-174
Annex for Appendix 3: Evaluation of Studies on Health Effects of Ozone 3-187
3.4 References 3-194
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 4-26
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
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 4-45
4.2.1 Introduction, Summary from the 2013 Ozone ISA, and Scope for Current Review 4-45
-------
CONTENTS (Continued)
4.2.2 Biological Plausibility 4-47
4.2.3 Ischemic Heart Disease (IHD) and Associated Cardiovascular Effects 4-49
4.2.4 Atherosclerosis 4-50
4.2.5 Heart Failure and Impaired Heart Function 4-50
4.2.6 Vascular Function 4-51
4.2.7 Cardiac Depolarization, Repolarization, Arrhythmia, and Arrest 4-51
4.2.8 Blood Pressure Changes and Hypertension 4-52
4.2.9 Heart Rate and Heart Rate Variability 4-53
4.2.10 Coagulation 4-54
4.2.11 Systemic Inflammation and Oxidative Stress 4-54
4.2.12 Stroke and Associated Cardiovascular Effects 4-55
4.2.13 Other Cardiovascular Endpoints 4-56
4.2.14 Aggregate Cardiovascular Disease 4-56
4.2.15 Cardiovascular Mortality 4-57
4.2.16 Potential Copollutant Confounding of the Ozone-Cardiovascular Disease (CVD)
Relationship 4-58
4.2.17 Effect Modification of the Ozone-Cardiovascular Relationship 4-60
4.2.18 Summary and Causality Determination 4-62
4.3 Evidence Inventories—Data Tables to Summarize Study Details 4-65
4.3.1 Short-Term Ozone Exposure 4-65
4.3.2 Long-Term Ozone Exposure 4-129
Annex for Appendix 4: Evaluation of Studies on Health Effects of Ozone 4-147
4.4 References 4-154
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 Glucose and Insulin Homeostasis 5-7
5.1.4 Overweight and Obesity 5-11
5.1.5 Other Indicators of Metabolic Function 5-13
5.1.6 Potential Copollutant Confounding of the Ozone-Metabolic Effects Relationship 5-21
5.1.7 Effect Modification ofthe Ozone-Metabolic Effects Relationship 5-21
5.1.8 Summary and Causality Determination 5-23
5.2 Long-Term Ozone Exposure—Introduction, Summary from the 2013 Ozone ISA, and Scope
for Current Review 5-26
5.2.1 Population, Exposure, Comparison, Outcome, and Study Design (PECOS) Tool 5-26
5.2.2 Biological Plausibility 5-27
5.2.3 Glucose and Insulin Homeostasis 5-29
5.2.4 Adiposity, Weight Gain, and Obesity 5-30
5.2.5 Metabolic Syndrome and Type 2 Diabetes 5-32
5.2.6 Type 1 Diabetes 5-34
5.2.7 Gestational Diabetes 5-34
5.2.8 Metabolic Disease Mortality 5-35
5.2.9 Potential Copollutant Confounding ofthe Ozone-Metabolic Effects Relationship 5-35
5.2.10 Effect Modification ofthe Ozone-Metabolic Effects Relationship 5-35
5.2.11 Summary and Causality Determination 5-37
5.3 Evidence Inventories—Data Tables to Summarize Study Details 5-40
Annex for Appendix 5: Evaluation of Studies on Health Effects of Ozone
5-55
-------
CONTENTS (Continued)
5.4 References 5-62
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 6-4
6.1.4 Cause-Specific Mortality 6-7
6.1.5 Effect Modification of the Ozone-Mortality Relationship 6-9
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 6-33
6.2.5 Potential Copollutant Confounding of the Ozone-Mortality Relationship 6-34
6.2.6 Shape of the Concentration-Response Function 6-36
6.2.7 Summary and Causality Determination 6-39
6.3 Evidence Inventories—Data Tables to Summarize Study Details 6-44
6.3.1 Short-Term Ozone Exposure and Mortality: Data Tables 6-44
6.3.2 Long-Term Ozone Exposure and Mortality: Data Tables 6-56
Annex for Appendix 6: Evaluation of Studies on Health Effects of Ozone 6-67
6.4 References 6-74
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 7-114
7.6 Evidence Inventories—Data Tables to Summarize Cancer Study Details 7-119
7.6.1 Epidemiologic Studies 7-119
7.6.2 Toxicological Studies 7-124
-------
CONTENTS (Continued)
Annex for Appendix 7: Evaluation of Studies on Health Effects of Ozone 7-125
7.7 References 7-132
APPENDIX 8 ECOLOGICAL EFFECTS 8-1
8.1 Introduction 8-1
8.1.1 Scope 8-2
8.1.2 Assessing Ecological Response to Ozone 8-5
8.1.3 Mechanisms Governing Vegetation Response to Ozone 8-10
8.2 Visible Foliar Injury and Biomonitoring 8-12
8.2.1 Summary 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 8-37
8.4 Plant Reproduction, Phenology, and Mortality 8-43
8.4.1 Plant Reproduction 8-44
8.4.2 Plant Phenology 8-47
8.4.3 Plant Mortality 8-48
8.4.4 Summary 8-48
8.5 Reduced Crop Yield and Quality 8-62
8.5.1 Field Studies and Meta-Analyses 8-62
8.5.2 Yield Loss at Regional and National Scales 8-64
8.5.3 Summary 8-66
8.6 Herbivores: Growth, Reproduction, and Survival 8-73
8.6.1 Individual-Level Responses 8-74
8.6.2 Population- and Community-Level Responses 8-75
8.6.3 Summary 8-76
8.7 Plant-Insect Signaling 8-89
8.7.1 Emission and Chemical Composition of Volatile Plant Signaling Compounds
(VPSCs) 8-90
8.7.2 Pollinator Attraction and Plant Host Detection 8-91
8.7.3 Plant Attraction of Natural Enemies of Herbivores 8-92
8.7.4 Summary 8-92
8.8 Carbon Cycling in Terrestrial Ecosystems: Primary Productivity and Carbon Sequestration 8-100
8.8.1 Terrestrial Primary Productivity 8-100
8.8.2 Soil Carbon 8-103
8.8.3 Terrestrial Carbon Sequestration 8-104
8.8.4 Summary 8-105
8.9 Soil Biogeochemistry 8-116
8.9.1 Decomposition 8-117
8.9.2 Soil Carbon 8-118
8.9.3 Soil Nitrogen 8-120
8.9.4 Summary 8-122
8.10 Terrestrial Community Composition 8-131
8.10.1 Plant Community 8-132
8.10.2 Microbes 8-137
8.10.3 Consumer Communities 8-140
8.10.4 Summary 8-141
8.11 Water-Cycling 8-163
8.11.1 Structural Changes in Plants 8-163
8.11.2 Impaired Stomatal Function 8-164
-------
CONTENTS (Continued)
8.11.3 Models of Plant Water Use 8-165
8.11.4 Ecosystem Water Dynamics 8-166
8.11.5 Drought and Ozone 8-167
8.11.6 Summary 8-167
8.12 Modifying Factors 8-177
8.12.1 Nitrogen 8-177
8.12.2 Carbon Dioxide 8-178
8.12.3 Weather and Climate 8-180
8.12.4 Summary 8-181
8.13 Exposure Indices/Exposure Response 8-182
8.13.1 Exposure Indices 8-183
8.13.2 Exposure Response 8-184
8.13.3 Summary 8-197
8.14 References 8-205
APPENDIX 9 THE ROLE OF TROPOSPHERIC OZONE IN 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 9-7
9.2.1 Recent Evidence for Historical Period 9-7
9.2.2 Recent Evidence of Radiative Forcing Temporal and Spatial Trends 9-12
9.2.3 Summary and Causality Determination 9-14
9.3 Ozone Impacts on Temperature, Precipitation, and Related Climate Variables 9-16
9.3.1 Recent Evidence for Effects on Temperature 9-16
9.3.2 Recent Evidence for Other Climate Effects 9-18
9.3.3 Summary and Causality Determination 9-20
9.4 References 9-22
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 10-19
10.3.3 Documentation 10-22
10.4 Peer Review and Public Participation 10-23
10.4.1 Call for Information 10-23
10.4.2 Integrated Review Plan 10-24
10.4.3 Peer Input 10-24
10.4.4 Internal Technical Review 10-25
10.4.5 Clean Air Scientific Advisory Committee (CASAC) Peer Review 10-25
10.5 Quality Assurance 10-26
10.6 Conclusion 10-27
10.7 References 10-28
-------
LIST OF TABLES
Table I History of the National Ambient Air Quality Standards for Ozone,
1971-2015. lxxiv
Table II Weight of evidence for causality determinations. lxxix
Table ES-1 Summary of causality determinations by exposure duration and health
outcome. ES-5
Table ES-2 Summary of causality determinations for ecological effects. ES-13
Table ES-3 Summary of causality determinations for tropospheric ozone effects on
climate. ES-15
Table IS-1 Summary of causality determinations by exposure duration and health
outcome. IS-7
Table IS-2 Summary of causality determinations for ecological effects. IS-10
Table IS-3 Summary of causality determinations for tropospheric ozone effects on
climate. IS-11
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. IS-24
Table IS-5 Summary of evidence from epidemiologic and animal toxicological
studies on the respiratory effects associated with long-term ozone
exposure. IS-32
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. IS-3 8
Table IS-7 Summary of evidence from epidemiologic and animal toxicological
studies on the metabolic effects associated on long-term ozone
exposure. IS-40
Table IS-8 Summary of evidence from epidemiologic, controlled human exposure,
and animal toxicological studies on the cardiovascular effects of
short-term ozone exposure. IS-41
Table IS-9
Summary of evidence from epidemiologic studies on the association of
short-term ozone exposure with mortality. IS-46
-------
LIST OF TABLES (Continued)
Table IS-10 Characterization of evidence for factors potentially increasing the risk
for ozone-related health effects. IS-52
Table IS-11 Summary of evidence for populations at increased risk to the health
effects of ozone. IS-53
Table IS-12 Prevalence of respiratory diseases, cardiovascular diseases, diabetes,
and obesity among adults by age and region in the U.S. in 2012. IS-57
Table IS-13 Summary of evidence for welfare effects of ozone. IS-63
Table 1-1 Nationwide distributions of ozone concentrations (ppb) from the
year-round data set 2015-2017. 1-38
Table 1-2 Nationwide distributions of ozone concentrations (ppb) from the
warm-season data set 2015-2017. 1-40
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.). 2-17
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.). 2-18
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.). 2-20
Table 2-4 Summary of U.S. studies of ozone infiltration published after 2011. 2-22
Table 2-5 Studies reporting relationships between personal ozone exposures and
ambient ozone concentrations. 2-30
Table 2-6. Summary of exposure estimation methods, their typical use in ozone
epidemiologic studies, and related errors and uncertainties. 2-42
Table 2-7 Studies informing assessment of exposure measurement error when
concentrations measured by fixed-site monitors are used for exposure
surrogates. 2-56
Table 2-8 Studies informing assessment of exposure measurement error when
concentrations measured by personal and microenvironmental monitors
are used for exposure surrogates. 2-61
-------
LIST OF TABLES (Continued)
Table 2-9 Studies informing assessment of exposure measurement error when
concentrations modeled by spatial interpolations methods are used for
exposure surrogates. 2-63
Table 2-10 Studies informing assessment of exposure measurement error when
concentrations modeled by land use regression or spatiotemporal
models are used for exposure surrogates. 2-71
Table 2-11 Studies informing assessment of exposure measurement error when
concentrations modeled by chemical transport modeling are used for
exposure surrogates. 2-79
Table 2-12 Studies informing assessment of exposure measurement error when
concentrations modeled by hybrid approaches are used for exposure
surrogates. 2-138
Table 2-13 Studies informing assessment of exposure measurement error when
concentrations modeled by microenvironmental modeling are used for
exposure surrogates. 2-155
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. 3-73
Table 3-2 Summary of evidence indicating a causal relationship between
short-term ozone exposure and respiratory effects. 3-81
Table 3-3 Summary of evidence for a likely to be causal relationship between
long-term ozone exposure and respiratory effects. 3-112
Table 3-4 Study-specific details from controlled human exposure studies of lung
function in healthy populations. 3-116
Table 3-5 Study-specific details from animal toxicological studies of short-term
ozone exposure and lung function—healthy. 3-118
Table 3-6 Epidemiologic studies of short-term exposure to ozone and lung
function in healthy populations. 3-122
Table 3-7 Epidemiologic studies of short-term exposure to ozone and lung
function, airway inflammation, and oxidative stress in general
populations. 3-124
Table 3-8 Study-specific details from controlled human exposure studies of
respiratory symptoms in healthy populations. 3-125
Table 3-9
Study-specific details from controlled human exposure studies of
inflammation, oxidative stress, and injury in healthy populations. 3-126
-------
LIST OF TABLES (Continued)
Table 3-10 Study-specific details from animal toxicological studies of short-term
ozone exposure and allergic sensitization—healthy. 3-128
Table 3-11 Study-specific details from animal toxicological studies of short-term
ozone exposure and inflammation, oxidative stress, and
injury—healthy. 3-129
Table 3-12 Study-specific details from animal toxicological studies of short-term
ozone exposure and morphology—healthy. 3-139
Table 3-13 Epidemiologic studies of short-term exposure to ozone and hospital
admission for asthma. 3-142
Table 3-14 Epidemiologic studies of short-term exposure to ozone and emergency
department (ED) visits for asthma. 3-144
Table 3-15 Epidemiologic studies of short-term exposure to ozone and respiratory
symptoms in children with asthma. 3-149
Table 3-16 Study-specific details from controlled human exposure studies of lung
function in adults with asthma. 3-149
Table 3-17 Study-specific details from controlled human exposure studies of lung
function in healthy adults and adults with asthma. 3-150
Table 3-18 Study-specific details from animal toxicological studies of short-term
ozone exposure and lung function—allergy. 3-150
Table 3-19 Study-specific details from animal toxicological studies of short-term
ozone exposure and lung function—asthma. 3-151
Table 3-20 Study-specific details from controlled human exposure studies of
inflammation, oxidative stress, and injury in healthy adults and adults
with asthma. 3-152
Table 3-21 Study-specific details from controlled human exposure studies of
allergic sensitization—atopy. 3-152
Table 3-22 Study-specific details from animal toxicological studies of short-term
ozone exposure and inflammation, oxidative stress, and
injury—allergy. 3-153
Table 3-23 Study-specific details from animal toxicological studies of short-term
ozone exposure and inflammation, oxidative stress, and
injury—asthma. 3-153
Table 3-24
Study-specific details from animal toxicological studies of short-term
ozone exposure and morphology—asthma. 3-154
-------
LIST OF TABLES (Continued)
Table 3-25 Epidemiologic studies of short-term exposure to ozone and
inflammation, oxidative stress, and injury in children with asthma. 3-154
Table 3-26 Epidemiologic studies of short-term exposure to ozone and emergency
department (ED) visits for chronic obstructive pulmonary disease
(COPD). 3-155
Table 3-27 Epidemiologic studies of short-term exposure to ozone and medication
use in adults with chronic obstructive pulmonary disease (COPD). 3-157
Table 3-28 Study-specific details from animal toxicological studies of short-term
ozone exposure and lung function—chronic obstructive pulmonary
disease (COPD). 3-157
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). 3-158
Table 3-30 Study-specific details from animal toxicological studies of short-term
ozone exposure and morphology—chronic obstructive pulmonary
disease (COPD). 3-158
Table 3-31 Study-specific details from controlled human exposure studies of
respiratory effects in obese adults. 3-159
Table 3-32 Study-specific details from animal toxicological studies of short-term
ozone exposure and lung function—obesity. 3-159
Table 3-33 Study-specific details from animal toxicological studies of short-term
ozone exposure and inflammation, oxidative stress, and
injury—obesity. 3-160
Table 3-34 Study-specific details from animal toxicological studies of short-term
ozone exposure and morphology—obesity. 3-161
Table 3-35 Epidemiologic studies of short-term exposure to ozone and pulmonary
inflammation in populations with metabolic syndrome. 3-161
Table 3-36 Study-specific details from animal toxicological studies of short-term
ozone exposure and lung function—cardiovascular disease. 3-162
Table 3-37 Study-specific details from animal toxicological studies of short-term
ozone exposure and inflammation, oxidative stress, and
injury—cardiovascular disease. 3-163
Table 3-38 Study-specific details from animal toxicological studies of short-term
ozone exposure and morphology—cardiovascular disease. 3-164
-------
LIST OF TABLES (Continued)
Table 3-39 Epidemiologic studies of short-term exposure to ozone and emergency
department (ED) visits for respiratory infection. 3-165
Table 3-40 Study-specific details from animal toxicological studies of short-term
ozone exposure and host defense/infection—healthy. 3-169
Table 3-41 Epidemiologic studies of short-term exposure to ozone and hospital
admissions for aggregate respiratory diseases. 3-170
Table 3-42 Epidemiologic studies of short-term exposure to ozone and emergency
department (ED) visits for aggregate respiratory diseases. 3-171
Table 3-43 Epidemiologic studies of long-term exposure to ozone and
development of asthma. 3-174
Table 3-44 Study-specific details from animal toxicological studies of long-term
ozone exposure and inflammation, oxidative stress, and
injury—allergy. 3-175
Table 3-45 Study-specific details from animal toxicological studies of long-term
ozone exposure and lung function—allergy. 3-176
Table 3-46 Study-specific details from animal toxicological studies of long-term
ozone exposure and inflammation, oxidative stress, and
injury—healthy. 3-177
Table 3-47 Study-specific details from animal toxicological studies of long-term
ozone exposure and morphology and other endpoints—healthy. 3-179
Table 3-48 Epidemiologic studies of long-term exposure to ozone and lung
function and development. 3-180
Table 3-49 Study-specific details from animal toxicological studies of long-term
ozone exposure and morphology—allergy. 3-181
Table 3-50 Study-specific details from animal toxicological studies of long-term
ozone exposure and lung function—healthy. 3-182
Table 3-51 Epidemiologic studies of long-term exposure to ozone and
development of chronic obstructive pulmonary disease (COPD). 3-182
Table 3-52 Epidemiologic studies of long-term exposure to ozone and respiratory
infection. 3-183
Table 3-53
Epidemiologic studies of long-term exposure to ozone and severity of
respiratory disease. 3-184
-------
LIST OF TABLES (Continued)
Table 3-54 Epidemiologic studies of long-term exposure to ozone and allergic
sensitization. 3-185
Table 3-55 Study-specific details from animal toxicological studies of long-term
ozone exposure and allergic sensitization—healthy. 3-186
Table Annex 3-1 Scientific considerations for evaluating the strength of inference from
studies on the health effects of ozone. 3-188
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. 4-42
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. 4-64
Table 4-3 Epidemiologic studies of short-term exposure to ozone and heart
failure. 4-65
Table 4-4 Study-specific details from controlled human exposure studies of
impaired heart function. 4-69
Table 4-5 Study-specific details from short-term animal toxicological studies of
impaired heart function. 4-70
Table 4-6 Epidemiologic studies of short-term exposure to ozone and ischemic
heart disease. 4-73
Table 4-7 Epidemiologic panel studies of short-term exposure to ozone and
ischemic heart disease. 4-80
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. 4-81
Table 4-10 Epidemiologic panel studies of short-term exposure to ozone and
endothelial function. 4-82
Table 4-11 Study-specific details from controlled human exposure studies of
vascular function. 4-83
Table 4-12 Study-specific details from short-term animal toxicological studies of
vascular function. 4-84
-------
LIST OF TABLES (Continued)
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. 4-85
Table 4-14 Epidemiologic panel studies of short-term exposure to ozone and
electrophysiology, arrhythmia, and cardiac arrest. 4-89
Table 4-15 Study-specific details from controlled human exposure studies of
electrophysiology, arrhythmia, cardiac arrest. 4-90
Table 4-16 Study-specific details from short-term animal toxicological studies of
electrophysiology, arrhythmia, cardiac arrest. 4-91
Table 4-17 Epidemiologic studies of short-term exposure to ozone and blood
pressure. 4-92
Table 4-18 Epidemiologic panel studies of short-term exposure to ozone and blood
pressure. 4-94
Table 4-19 Study-specific details from controlled human exposure studies of blood
pressure. 4-96
Table 4-20 Study-specific details from short-term animal toxicological studies of
blood pressure. 4-97
Table 4-21 Epidemiologic panel studies of short-term exposure to ozone and heart
rate variability (HRV), and heart rate (HR). 4-99
Table 4-22 Study-specific details from controlled human exposure studies of heart
rate variability (HRV), heart rate (HR). 4-101
Table 4-23 Study-specific details from short-term animal toxicological studies of
heart rate variability (HRV), heart rate (HR). 4-102
Table 4-24 Epidemiologic studies of short-term exposure to ozone and pulmonary
vascular disease (PVD), thrombosis. 4-104
Table 4-25 Epidemiologic panel studies of short-term exposure to ozone and
coagulation. 4-106
Table 4-26 Study-specific details from controlled human exposure studies of
coagulation. 4-107
Table 4-27 Study-specific details from short-term animal toxicological studies of
coagulation. 4-108
Table 4-28
Epidemiologic panel studies of short-term exposure to ozone and
inflammation. 4-109
-------
LIST OF TABLES (Continued)
Table 4-29 Study-specific details from controlled human exposure studies of
systemic inflammation and oxidative stress. 4-111
Table 4-30 Study-specific details from short-term animal toxicological studies of
systemic inflammation and oxidative stress. 4-113
Table 4-31 Epidemiologic studies of short-term exposure to ozone and
cerebrovascular disease. 4-117
Table 4-32 Epidemiologic studies of short-term exposure to ozone and aggregate
cardiovascular disease. 4-125
Table 4-33 Epidemiologic studies of short-term exposure to ozone and
cardiovascular mortality. 4-128
Table 4-34 Epidemiologic studies of long-term exposure to ozone and ischemic
heart disease (IHD). 4-129
Table 4-35 Epidemiologic studies of long-term exposure to ozone and
atherosclerosis. 4-130
Table 4-36 Study-specific details from animal toxicological studies of
atherosclerosis. 4-131
Table 4-37 Epidemiologic studies of long-term exposure to ozone and heart failure. 4-132
Table 4-38 Study-specific details from animal toxicological studies of impaired
heart function. 4-133
Table 4-39 Study-specific details from animal toxicological studies of vascular
function. 4-133
Table 4-40 Epidemiologic studies of long-term exposure to ozone and blood
pressure. 4-134
Table 4-41 Study-specific details from animal toxicological studies of blood
pressure. 4-141
Table 4-42 Study-specific details from animal toxicological studies of heart rate
variability (HRV), heart rate (HR). 4-141
Table 4-43 Study-specific details from animal toxicological studies of coagulation. 4-142
Table 4-44 Study-specific details from animal toxicological studies of
inflammation. 4-143
Table 4-45 Epidemiologic studies of long-term exposure to ozone and
cerebrovascular disease. 4-144
-------
LIST OF TABLES (Continued)
Table 4-46 Epidemiologic studies of long-term exposure to ozone and aggregate
cardiovascular disease. 4-146
Table Annex 4-1 Scientific considerations for evaluating the strength of inference from
studies on the health effects of ozone. 4-148
Table 5-1 Summary of evidence for a likely to be causal relationship between
short-term ozone exposure and metabolic effects. 5-24
Table 5-2 Criteria for clinical diagnosis of metabolic syndrome. 5-32
Table 5-3 Criteria for clinical diagnosis of diabetes. 5-33
Table 5-4 Summary of evidence to support a likely to be causal relationship
between long-term ozone exposure and metabolic effects. 5-39
Table 5-5 Epidemiologic studies of short-term exposure to ozone and
glucose/insulin homeostasis. 5-40
Table 5-6 Controlled human exposure study of short-term exposure to ozone and
glucose/insulin homeostasis and other metabolic indicators. 5-42
Table 5-7 Study-specific details from animal toxicological studies of short-term
exposure to ozone and glucose/insulin homeostasis. 5-43
Table 5-8 Study-specific details from animal toxicological studies of short-term
overweight and obesity. 5-45
Table 5-9 Epidemiologic studies of short-term exposure to ozone and other
indicators. 5-46
Table 5-10 Study-specific details from animal toxicological studies of short-term,
other metabolic indicators. 5-47
Table 5-11 Epidemiologic studies of long-term exposure to ozone and overweight
and obesity. 5-50
Table 5-12 Epidemiologic studies of long-term exposure to ozone and metabolic
syndrome and type 2 diabetes. 5-51
Table 5-13 Study-specific details from long-term animal toxicological studies of
glucose and insulin homeostasis. 5-53
Table 5-14 Epidemiologic studies of long-term exposure to ozone and type 1
diabetes. 5-54
Table Annex 5-1 Scientific considerations for evaluating the strength of inference from
studies on the health effects of ozone. 5-56
-------
LIST OF TABLES (Continued)
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. 6-23
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. 6-42
Table 6-3 Epidemiologic studies of short-term exposure to ozone and total
(nonaccidental) mortality. 6-44
Table 6-4 Epidemiologic studies of short-term exposure to ozone and
cardiovascular mortality. 6-52
Table 6-5 Epidemiologic studies of short-term exposure to ozone and respiratory
mortality. 6-54
Table 6-6 Epidemiologic studies of long-term exposure to ozone and total
(nonaccidental) mortality. 6-56
Table 6-7 Epidemiologic studies of long-term exposure to ozone and
cardiovascular mortality. 6-60
Table 6-8 Epidemiologic studies of long-term exposure to ozone and respiratory
mortality. 6-63
Table 6-9 Epidemiologic studies of long-term exposure to ozone and other
mortality. 6-65
Table Anne 6-1 Scientific considerations for evaluating the strength of inference from
studies on the health effects of ozone. 6-68
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. 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. 7-20
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. 7-32
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. 7-43
-------
LIST OF TABLES (Continued)
Table 7-5 Summary of evidence that is inadequate to determine if a causal
relationship exists between long-term ozone exposure and cancer. 7-49
Table 7-6. Epidemiologic studies of exposure to ozone and reproduction—male. 7-50
Table 7-7 Epidemiologic studies of exposure to ozone and reproduction—female. 7-51
Table 7-8 Epidemiologic studies of exposure to ozone and
pregnancy/birth—hypertension disorders. 7-52
Table 7-9 Epidemiologic studies of exposure to ozone and
pregnancy/birth—diabetes. 7-55
Table 7-10 Epidemiologic studies of exposure to ozone and pregnancy/birth—fetal
growth. 7-56
Table 7-11 Epidemiologic studies of exposure to ozone and
pregnancy/birth—preterm birth. 7-64
Table 7-12 Epidemiologic studies of exposure to ozone and pregnancy/birth—birth
defects. 7-74
Table 7-13 Epidemiologic studies of exposure to ozone and
pregnancy/birth—infant and fetal mortality. 7-80
Table 7-14 Epidemiologic studies of exposure to ozone and developmental effects. 7-83
Table 7-15 Epidemiologic studies of exposure to ozone and other. 7-92
Table 7-16 Study specific details from studies of ozone (O3) and pregnancy/birth
outcomes. 7-96
Table 7-17 Study specific details from studies of ozone (O3) and developmental
effects. 7-98
Table 7-18 Epidemiologic studies of short-term exposure to ozone and effects on
cognition, motor activity, and mood. 7-101
Table 7-19 Epidemiologic studies of short-term exposure to ozone and hospital
admissions, emergency department, and outpatient visits. 7-102
Table 7-20 Epidemiologic studies of long-term exposure to ozone and
cognitive/behavioral effects. 7-106
Table 7-21 Epidemiologic studies of long-term exposure to ozone and
neurodegenerative diseases. 7-108
Table 7-22 Epidemiologic studies of long-term exposure to ozone and
neurodevelopmental effects. 7-111
-------
LIST OF TABLES (Continued)
Table 7-23 Study-specific details from short-term studies of brain inflammation
and morphology. 7-114
Table 7-24 Study-specific details from short-term studies of cognitive and
behavioral effects. 7-116
Table 7-25 Study-specific details from short-term studies of neuroendocrine
effects. 7-116
Table 7-26 Study-specific details from long-term studies of brain inflammation and
morphology. 7-117
Table 7-27 Study-specific details from long-term studies of cognitive and
behavioral effects. 7-118
Table 7-28 Study effects. 7-119
Table 7-29 Epidemiologic studies of long-term exposure to ozone and cancer
incidence. 7-120
Table 7-30 Epidemiologic studies of ozone exposure and lung cancer mortality. 7-121
Table 7-31 Epidemiologic studies of long-term exposure to ozone and other cancer
endpoints. 7-123
Table 7-32 Study specific details of ozone exposure and DNA damage. 7-124
Table Annex 7-1 Scientific considerations for evaluating the strength of inference from
studies on the health effects of ozone. 7-126
Table 8-1 Summary of ozone causality determinations for effects on vegetation
and ecosystems in the 2013 Ozone ISA. 8-3
Table 8-2 Population, exposure, comparison, outcome, and study design
(PECOS) tool for ozone effects on vegetation and ecosystems. 8-4
Table 8-3 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). 8-14
Table 8-4 Ozone exposure and foliar injury. 8-25
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 Bergmann et al. (2017). 8-32
Table 8-6 Ozone exposure and plant growth and biomass. 8-38
-------
LIST OF TABLES (Continued)
Table 8-7 Summary of evidence for causal relationship between ozone exposure
and plant reproduction. 8-50
Table 8-8 Summary of evidence for likely to be causal relationship between
ozone exposure and tree mortality. 8-51
Table 8-9 Ozone exposure and plant reproduction, phenology, and mortality. 8-52
Table 8-10 Ozone and crop yield and quality. 8-67
Table 8-11 Summary of studies reporting altered growth in herbivores. 8-77
Table 8-12 Summary of studies reporting altered reproduction in herbivores. 8-79
Table 8-13 Summary of evidence for likely to be causal relationship between
ozone exposure and alteration of herbivore growth and reproduction. 8-80
Table 8-14 Ozone exposure and effects on herbivores. 8-81
Table 8-15 Summary of evidence for a likely to be causal relationship between
ozone exposure and alteration of plant-insect signaling. 8-94
Table 8-16 Ozone exposure and plant insect signaling. 8-95
Table 8-17 Ozone exposure effects on productivity and carbon sequestration. 8-107
Table 8-18 Response of belowground processes and biogeochemical cycles to
ozone exposure. 8-124
Table 8-19 Summary of evidence for a causal relationship between ozone exposure
and terrestrial community composition, based on Table 2 from the
Preamble. 8-141
Table 8-20 Terrestrial community composition response to ozone exposure. 8-145
Table 8-21 Ozone exposure and water cycling. 8-168
Table 8-22 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) 8-189
Table 8-23 Exposure indices and exposure response. 8-198
Table 9-1 Population, exposure, comparison, outcome, and study design
(PECOS) tool for radiative forcing and climate change. 9-3
Table 9-2 Contributions of tropospheric ozone changes to radiative forcing
(W/m2) from 1750 to 2011. 9-8
-------
LIST OF TABLES (Continued)
Table 9-3
Table 9-4
Table 9-5
Table 10-1
Table 10-2
Table 10-3
Table 10-4
Confidence level for ozone forcing for the 1750-2011 period. 9-8
Summary of evidence for a causal relationship between tropospheric
ozone and radiative forcing. 9-15
Summary of evidence for a likely to be causal relationship between
ozone and temperature, precipitation, and related climate variables. 9-21
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. 10-18
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. 10-19
-------
LIST OF
Figure ES-1
Figure ES-2
Figure ES-3
Figure ES-4
Figure ES-5
Figure IS-1
Figure IS-2
Figure IS-3
Figure IS-4
Figure IS-5
Figure IS-6
Figure IS-7
Figure IS-8
Figure 1-1
FIGURES
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-2
Causality determinations for health effects of short- and long-term
exposure to ozone. ES-6
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-8
Illustrative diagram of ozone effects cascading from the cellular level
to plants and ecosystems. ES-11
Causality determinations for ozone across biological scales of
organization and taxonomic groups. ES-12
Illustrative figure for potential biological pathways for health effects
following ozone exposure. IS-22
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-29
Illustrative diagram of ozone effects cascading up through scales of
biological organization from the cellular level to plants and ecosystems._ IS-69
Representative ozone foliar injury in two common tree species in the
U.S. IS-70
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
Causality determinations for health effects of short- and long-term
exposure to ozone. IS-84
Causality determinations for ozone across biological scales of
organization and taxonomic groups. IS-90
Causality determinations for tropospheric ozone and climate change. IS-93
Major atmospheric processes and precursor sources contributing to
ambient ozone. 1-8
-------
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, is included for context. D) methane (CH4). 1-10
Figure 1-3 U.S. anthrogenic ozone precursor emission trends. 1-11
Figure 1-4 Asian anthrogenic ozone precursor emission trends.. 1-17
Figure 1-5 Anthropogenic ozone precursor emission trends derived using the
MEIC emissions model. 1-18
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. 1-32
Figure 1-7 Monitor locations for the warm-season and year-round data sets. 1-43
Figure 1-8 Individual monitor ozone concentrations in terms of design values for
2015-2017. 1-44
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). 1-45
Figure 1-10 Trend in mean 4th-highest 8-hour daily max ozone by U.S. region
2000-2017. 1-46
Figure 1-11 Individual monitor 3-year avg of the changes in ozone design values
from 2008-2010 to 2015-2017. 1-47
Figure 1-12 Individual monitor W126 exposure metric values for 2015-2017. 1-48
Figure 1-13 CMAQ (a) and CAMx (b) estimates of daily distributions of
bias-adjusted USB MDA8 ozone concentration (ppb) for the period
April-October 2007, binned by base model MDA8 ozone
concentration ranges. 1-58
Figure 1-14 CMAQ (a) and CAMx (b) estimates of daily distributions of
bias-adjusted USB ozone fraction at monitoring locations across the
western U.S. for the period April-October 2007, binned by base model
MDA8 ozone concentration ranges. 1-59
Figure 2-1. Year-round Pearson correlations of 8-hour daily max ozone
concentrations with copollutant concentrations measured in AQS
2015-2017. 2-33
-------
LIST OF FIGURES (Continued)
Figure 2-2. Seasonal Pearson correlations of 8-hour daily max ozone
concentrations with copollutant concentrations measured in AQS,
2015-2017. 2-34
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 one 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 one 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. 3-41
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. 3-43
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. 3-62
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. 3-67
Figure 3-8 Loess (locally estimated scatterplot smoothing) 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. 3-74
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. 3-75
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. 3-76
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. 3-77
-------
LIST OF FIGURES (Continued)
Figure 3-12 Potential biological pathways for respiratory effects following
long-term ozone exposure. 3-89
Figure 3-13 Associations between long-term exposure to ozone and respiratory
mortality in recent cohort studies. 3-107
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. 4-47
Figure 4-7 Associations between long-term exposure to ozone and cardiovascular
mortality in recent cohort studies. 4-58
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. 4-60
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. 5-28
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 evaluated in previous ozone assessments. 6-6
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. 6-8
-------
LIST OF FIGURES (Continued)
Figure 6-3
Figure 6-4
Figure 6-5
Figure 6-6
Figure 6-7
Figure 6-8
Figure 6-9
Figure 6-10
Figure 6-11
Figure 6-12
Figure 6-13
Figure 6-14
Figure 7-1
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-
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-
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-
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-
Percent 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-
Associations between long-term exposure to ozone and total
(nonaccidental) mortality in recent cohort studies. 6-
Associations between long-term exposure to ozone and respiratory
mortality in recent cohort studies. 6-
Associations between long-term exposure to ozone and cardiovascular
mortality in recent cohort studies. 6-
Associations between long-term exposure to ozone and mortality with
and without adjustment for PM2.5 concentrations in recent cohort
studies. 6-
The concentration-response relationship estimated with log-linear
model with a thin-plate spline (A) and the concentration-response
relationship estimated with threshold model (B), indicating the
potential for a threshold at 40 ppb (8-hour daily max). 6-
Concentration-response relationship between ozone concentrations
(ppb) and total (nonaccidental) mortality in the CanCHEC cohort
(mean 39.6; knots: 30.0, 38.9, 50.7 ppb). 6-
Concentration-response curve for ozone associated with respiratory
mortality using a natural spline model with 3 degrees of freedom. 6-
Potential biological pathways for male reproduction and fertility effects
following ozone exposure. 7
-------
LIST OF FIGURES (Continued)
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. 7-29
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. 8-11
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. 8-46
Figure 8-5 Estimated percentage reduction of soybean and maize yield in the U.S.
from ozone for 1980-2011. 8-65
Figure 8-6 Conceptual model of ozone effects on herbivore growth, reproduction,
and survival. 8-73
Figure 8-7 Conceptual model of ozone effects on volatile plant signaling
compounds and plant-insect signaling. 8-90
Figure 8-8 Conceptual diagram of ozone effects on belowground processes and
biogeochemical cycles. 8-117
Figure 8-9 Elevated ozone effect of accelerated senescence and reduced seed
production soil N. 8-121
Figure 8-10 Mechanisms by which ozone alters plant communities. 8-132
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. 8-136
-------
LIST OF FIGURES (Continued)
Figure 8-12
Figure 8-13
Figure 8-14
Figure 8-15
Figure 8-16
Figure 9-1
Figure 9-2
Figure 9-3
Figure 9-4
Figure 9-5
Figure 9-6
Figure 9-7
Figure 9-8
Biological plausibility of ozone effects on soil microbial communities
and soil invertebrate communities. 8-143
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. 8-166
Quantiles of predicted relative biomass loss for four tree species in
NHEERL-WED experiments. 8-185
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. 8-186
Comparison between aboveground biomass observed in Aspen FACE
experiment in 6 years and biomass predicted by the median composite
function based on NHEERL-WED. 8-187
Schematic diagram of the effects of tropospheric ozone on climate. 9-5
Bar chart for radiative forcing (RF; hatched) and effective radiative
forcing (ERF; solid) for the period 1750-2011. 9-9
Radiative forcing (RF) over the industrial era associated with emitted
compounds, including ozone (green bars) and its precursors. 9-11
Time evolution of the radiative forcing (RF) from tropospheric ozone
from 1750 to 2010. 9-12
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). 9-13
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. 9-14
Mean annual change in surface temperature (°C) resulting from
tropospheric ozone concentration changes from 1850-2013. 9-17
Mean annual change in precipitation (mm/day) resulting from
tropospheric ozone concentration changes from 1850-2013. 9-19
-------
LIST OF FIGURES (Continued)
Figure 10-1
General process for development of Integrated Science Assessments.
10-3
Figure 10-2
Literature Flow Diagram for Ozone.
10-4
Figure 10-3
Summary of title/abstract screening in SWIFT-ActiveScreener.
10-7
Figure 10-4
Example of screening efficiency using SWIFT-ActiveScreener.
10-7
-------
INTEGRATED SCIENCE ASSESSMENT TEAM, FOR OZONE
AND RELATED PHOTOCHEMICAL OXIDANTS
Executive Direction
Dr. John Vandenberg (Director)—National Center for Environmental Assessment—RTP
Division, Office of Research and Development, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. Steven J. Dutton (Deputy Director)— National Center for Environmental
Assessment—RTP Division, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Jennifer Nichols (Acting Branch Chief)— National Center for Environmental
Assessment—RTP Division, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Tara Greaver (Acting Branch Chief)— National Center for Environmental
Assessment—RTP Division, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Andrew Hotchkiss (Acting Branch Chief)— National Center for Environmental
Assessment—RTP Division, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Technical Support Staff
Ms. Marieka Boyd—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Ms. Annamarie Cory—Oak Ridge Associated Universities, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Ms. Madison Feshuk—Oak Ridge Associated Universities, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Ms. Erin Gallagher—Oak Ridge Associated Universities, National Center for Environmental
Assessment, Office of Research and Development, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Ms. Beth Gatling—Oak Ridge Associated Universities, National Center for Environmental
Assessment, Office of Research and Development, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Ms. Hillary Hollinger—Oak Ridge Associated Universities, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Mr. Ryan Jones—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
-------
INTEGRATED SCIENCE ASSESSMENT TEAM, FOR OZONE (Continued)
Mr. Lukas Kerr—Oak Ridge Associated Universities, National Center for Environmental
Assessment, Office of Research and Development, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Ms. Mckayla Lein—Oak Ridge Associated Universities, National Center for Environmental
Assessment, Office of Research and Development, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Ms. Danielle Moore—Senior Environmental Employment Program, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Mr. R. Byron Rice—Oak Ridge Associated Universities, National Center for Environmental
Assessment, Office of Research and Development, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Mr. Brayndon Stafford—Oak Ridge Associated Universities, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Mr. S. Shane Thacker—National Center for Environmental Assessment, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Ms. Erin Vining—Oak Ridge Associated Universities, National Center for Environmental
Assessment, Office of Research and Development, U.S. Environmental Protection
Agency, Research Triangle Park, NC
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS
Authors
Dr. Thomas Luben*1' (Health Assessment Team Lead, Integrated Science Assessment for
Ozone)—National Center for 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)—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Ms. Rebecca Daniels1' (Project Manager, Integrated Science Assessment for
Ozone)—National Exposure Research Laboratory, 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*—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Barbara Buckley*—National Center for Environmental Assessment, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Mr. Evan Coffman* '—National Center for Environmental Assessment, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Laura Dishaw*—National Center for Environmental Assessment, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Brian Eder—National Exposure Research Laboratory, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Emmi Felker-Quinn*— National Center for Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC; Air Resources Division, Natural Resource Stewardship and Science, National
Park Service, Lakewood, CO
Dr. Meridith Fry*—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Washington, D.C.
Dr. Barbara Glenn*—National Center for Environmental Assessment, Office of Research
and Development, U.S. Environmental Protection Agency, Washington, D.C.
Dr. Tara Greaver*—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Brooke Hemming*'—National Center for Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Dr. Jeffrey D. Herrick*1'—National Center for Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS (Continued)
Dr. Erin Hincs* '—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. S. Douglas Kaylor—Oak Ridge Institute for Science and Education, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Ellen Kirranc*'—National Center for 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—National Health and Environmental Effects Research Laboratory,
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*#—National Center for Environmental Assessment, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Jennifer Nichols*—National Center for 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—National Health and Environmental Effects Research Laboratory,
Office of Research and Development, U.S. Environmental Protection Agency, Research
Triangle Park, NC
Dr. Jeanette Reyes—Oak Ridge Institute for Science and Education, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Jennifer Richmond-Bryant* '—National Center for Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Dr. Caroline Ridley*—National Center for Environmental Assessment, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Mr. Jason Sacks*—National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Michael Stewart*1'—National Center for Environmental Assessment, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. James Szykman—National Exposure Research Laboratory, 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—National Exposure Research Laboratory, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS (Continued)
Dr. Christopher Weaver*1'—National Center for Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Dr. Chelsea Weitekamp—Oak Ridge Institute for Science and Education, National Center
for Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Research Triangle Park, NC
*National Center for Environmental Assessment Scientific Staff,
f Appendix 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—National Health and Environmental Effects Research Laboratory,
Western Ecology Division, U.S. Environmental Protection Agency, Corvallis, OR
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
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. Jan Dye—National Health and Environmental Effects Research Laboratory, Western
Ecology Division, 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. Mehdi Hazari—National Health and Environmental Effects Research Laboratory, Office
of Research and Development, U.S. Environmental Protection Agency, Research
Triangle Park, NC
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS (Continued)
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. Terry Keating—National Center for Environmental Research, Office of Research and
Development, U.S. Environmental Protection Agency, Washington, D.C.
Dr. Travis Knuckles—Department of Occupational and Environmental Health Sciences,
School of Public Health, West Virginia University, Morgantown, WV
Dr. Urmila Kodavanti—National Health and Environmental Effects Research Laboratory,
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. 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. Leigh Moorhead—Oak Ridge Institute for Science and Education, National Center for
Environmental Assessment, 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, D.C.
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
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, National Center for
Environmental Assessment, Office of Research and Development, U.S. Environmental
Protection Agency, Research Triangle Park, NC
-------
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
ADMS-Urban
ADOS
ADREX, ADX
AER
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
Hemoglobin A1C
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
Advanced Dispersion Modeling System
Atmospheric Dispersion Modelling System-Urban
Autism Diagnostic Observation Schedule
adrenalectomy
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, arylhydrocarbon 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 Information Report for Public Access 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(s)
AM3 Atmospheric Model 3 (chemical transport model from Geophysical Fluid Dynamics Laboratory)
AMI acute myocardial infarction
AMO Atlantic Multidecadal Oscillation
AMPK Adenosine monophosphate-activated protein kinase
AMS American Meteorological Society
AMX adrenal medullarectomy
ANPP annual net primary productivity
AO arctic oscillation
AOD aerosol optical depth
seasonal sum of the difference between an hourly concentration at the threshold value of 0 ppb,
AOTO minus the threshold value of 0 ppb
seasonal sum of the difference between an hourly concentration at the threshold value of 40 ppb,
AOT40 minus the threshold value of 40 ppb
seasonal sum of the difference between an hourly concentration at the threshold value of 60 ppb
AOT60 minus the threshold value of 60 ppb
September 2019
xl
DRAFT: Do Not Cite or Quote
-------
sum of the differences between hourly concentrations greater than a specified threshold (x) during
AOTx 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 apo lipoprotein A-I
ApoB Apo lipoprotein B
ApoE- apo lipoprotein 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 (AR4) from the IPCC
AR5 Fifth Assessment Report (AR5) from the IPCC
ARDS adult respiratory distress syndrome
ARG arginase
ARI acute respiratory infection
As 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 Adenosinetriphosphotase
AUC area under the curve
September 2019
xli
DRAFT: Do Not Cite or Quote
-------
AUP unpaired predicted to-observed peak ozone ratio
AURAMS Unified Regional Air Quality Modeling System
AV atrioventricular
AVCD atrioventricular conduction disorders
avg average
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
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
September 2019
xlii DRAFT: Do Not Cite or Quote
-------
C4
C57BL
Ca
CAA
CAD
CALGRID
CalNEX
CAM
CAMP
CAMS
CAMx
CanCHEC
CAP
CAPPS
CARB
CASAC
CASTNet
CAT
CATHGEN
CB05
CBVD
CCL
CCR2
CCSP
CD
CD36
CD4
cd56
plants that use the Hatch-Slack cycle for fixing the carbon dioxide from the air
wild type c57 black mouse
ambient ozone concentration
Clean Air Act
coronary artery disease
California Grid Simulations
California Research at the Nexus of Air Quality and Climate Change
Community Atmosphere Model; plants that use crassulacean acid metabolism for fixing the
carbon dioxide from the air
Constant Air Quality Model Performance
Continuous Monitoring Station
Comprehensive Air Quality Model with extensions
Canadian Census Health and Environment Cohort
concentrated ambient particle(s)
Canadian Asthma Primary Prevention Study
California Air Resources Board
Clean Air Scientific Advisory Committee
Clean Air Status and Trends Network
catalase
catheterization genetics
carbon bond mechanism developed in 2005
cerebrovascular disease
chemokine ligand
chemokine receptor type 2
U.S. Climate Change Science Program, forerunner to the U.S. Global Change Research Program;
club cell secretory protein
cluster of differentiation; confidence distribution
cluster of differentiation 36
cluster of differentiation 4
neural cell adhesion molecule
September 2019
xliii
DRAFT: Do Not Cite or Quote
-------
CDC
CDR
CEM
CF
CFR
CH2O
CH4
CHAD
CHARGE
CHASER
CHD
CHE
CHF
Chil4
CHIMERE
CHRONOS
CHS
CI
Ci
cIMT
CL
Cb
Clcal
CLM
CINCh
CLPPs
cm
CM
Centers for Disease Control and Prevention
Clinical Dementia Rating Sum of Boxes
continuous emissions modeling
charcoal filter; carbon filtered
Code of Federal Regulations
carbohydrate
methane
Consolidated Human Activity Database
Childhood Autism Risks from Genetics and the Environment
CHemical Atmospheric general circulation model for Study of atmospheric Environment and
Radiative Forcing (chemical transport model)
chronic heart disease; coronary heart disease
controlled human exposure
congestive heart failure
chitinase-like-4 protein
regional chemistry transport model Multi-scale chemistry-transport model for atmospheric
composition analysis and forecast
Canadian Hemispheric and Regional Ozone and NOX System
Children's Health Study
Confidence interval
intra-cellular carbon dioxide; substrate concentrations
carotid intimal-medial thickness
critical level
chlorine (gas)
Chloride channel accessory 1
chemiluminescence method
nitryl chloride
community-level physiological profiles
centimeter
conditioned medium
September 2019
xliv
DRAFT: Do Not Cite or Quote
-------
CMAQ Community Multi-scale Air Quality modeling system
CMAQ-HBM Community Multi-scale 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 United States
COPD Chronic Obstructive Pulmonary Disease
CP coverage prediction interval
CPC condensation particle counter
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)
CSTRs 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
September 2019
xlv
DRAFT: Do Not Cite or Quote
-------
Cypb5 cytochrome p450 b5
D dose
DA24 daily 24-hour average concentration
db decibel
DBP diastolic blood pressure
DC3 Deep Convective Cloud and Chemistry (field study)
DDM Decoupled Direct Method
DDS Department of Developmental Services
DECSO Daily Emission estimates Constrained by Satellite Observations (algorithm)
DEHM Danish Eulerian Hemispheric Model
DEMED Adrenal demedullation
df degrees of freedom
dg decigram
Deriving Information on Surface conditions from COlumn and VERtically resolved observations
DISCOVER-AQ relevant to Air Quality
DL distributed lag
dL deciliters
DLEM Dynamic Land Ecosystem Model
DLNM distributed lag nonlinear model
DM Dry Moderate; dry matter
DM8H 8-hour daily maximum 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
September 2019
xlvi
DRAFT: Do Not Cite or Quote
-------
Ea ambient ozone exposure
EBC exhaled breath condensate
EC elemental carbon
ECG electrocardiographic
ED emergency department
EDGAR Emissions Database for Global Atmospheric Research
EDMUS European Database for Multiple Sclerosis
EGAS Economic Growth Analysis System
EGF epidermal growth factor
EGUs electricity generating units
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 strain of rats
FA filtered air; adjusted forcings; fatty acid
September 2019
xlvii
DRAFT: Do Not Cite or Quote
-------
FACE free-air carbon dioxide/ozone enrichment
FACS florescence activated cell sorting
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 one second
FFA free fatty acid
FF1H fawn-hooded hypertensive
FIA (United States Department of Agriculture Forest 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 grams
GAM generalized additive model
GB gross bias
GCM General Circulation Model
September 2019
xlviii
DRAFT: Do Not Cite or Quote
-------
GD
GDAS
GDM
GE
GEM-MACH
GEMS
GEOS
GEOS-Chem
GFAP
GFDL
GGT
GHG
GHGs
GINI
GINIplus
GIS
GISS
GLM
GLP
GLVs
GMRF
GOES
GOME2a
GPP
GPS
GPx
GRP
GRPR
gestational day
Global Data Assimilation System
gestational diabetes mellitus
gross error
Global Environmental Multi-scale coupled with Model of Air quality and Chemistry
Global and regional Earth-system Monitoring using Satellite and in-situ data
Goddard Earth Observing System
GEOS-Chemistry
glial fibrillatory acidic protein
Geophysical Fluid Dynamics Laboratory
gamma-glutamyltranspeptidase
greenhouse gas
Greenhouse Gases
German Infant Nutritional Intervention study
German Infant Nutritional Intervention plus environmental and genetic influences
Geographic Information System
Goddard Institute for Space Studies
generalized linear model
good laboratory practices
green leaf volatiles
Gaussian Markov random field
Geostationary Operational Environmental Satellite
Global Ozone Monitoring Experiment 2A (instrument system borne on the European Remote
Sensing Satellite)
gross primary productivity
Global Positioning System
glutathione peroxidase
gastrin-releasing peptide
gastrin-releasing peptide receptor
September 2019
xlix
DRAFT: Do Not Cite or Quote
-------
Gs stomatal conductance
GSH reduced glutathione
GSHGSSG 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
HbAlc Hemoglobin A1C
HBM Hierarchical Bayesian Model
HCHO Formaldehyde
HDDM higher order decoupled direct method; Hierarchical Bayesian Diffusion Drift Model
HDL high density lipoprotein
HDM house dust mite; house dust mite allergen
HDMA house dust mite allergen
HEI Health Effects Institute
HERO Health and Environmental Research Online
HF high frequency component of HRV; heart failure; high fat diet
HFD high fat diet
HFr right heart failure
Hg mercury
HGB Houston-Galveston-Brazoria
September 2019
1
DRAFT: Do Not Cite or Quote
-------
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
HO heme oxygenase
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
HOx hydrogen & oxygen containing radicals (sum of hydroxyl 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 defibrillators
ICD10 International Classification of Diseases - version 10
ICD9 International Classification of Diseases - version 9
ICU intensive care unit
IDW inverse distance weighting
IFN interferon
September 2019
li
DRAFT: Do Not Cite or Quote
-------
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
iNOS inducible nitric oxide synthase
IO iodine monoxide
IOA index of agreement
IOM institute of medicine
IP inhalable particle; 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 intra-tracheal 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 mice
JNK c-Jun N-terminal kinase
K2SO4 potassium sulfate-measurement of extractable soil carbon
KC local neutrophil chemoattractant protein
September 2019
lii
DRAFT: Do Not Cite or Quote
-------
KCLurban
KEEP
kg
KK
KKAY
km
kPa
KROFEX
L
LAEI
LAI
LANDIS
lbs
LC50
LDH
LE
LF
LHID2000-NHIRD
LIF
LIFE
LISA
LISAplus
LNOx
LOOCV
LOTOS-EUROS
LPS
LT50
LTB4
King's College London urban
The Kidney Early Evaluation Program
kilogram
KK mouse strain
KKAY strain of mouse
kilometer
kilopascal
Kranzberg Ozone Fumigation Experiment
location
large artery elasticity index
leaf area index
forest landscape model
pounds
median lethal concentration
lactate dehydrogenase
Lake Elsinore; Long Evans rat
low-frequency component of HRV
Longitudinal Health Insurance Database 2000 - National Health Insurance Research Database
leukemia inhibitory factor
Longitudinal Investigation of Fertility and the Environment
Lifestyle-Related factors on the Immune System and the Development of Allergies in Childhood
Lifestyle-Related factors on the Immune System and the Development of Allergies in Childhood
plus the influence of traffic emissions and genetics
nitrogen oxides generated by lightning
leave-one-out cross-validation
Long Term Ozone Simulation European Operational Smog
lipopolysaccharide
median lethal time
leukotriene B4
September 2019
liii
DRAFT: Do Not Cite or Quote
-------
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 seven 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
MCP monocyte chemoattractant protein; monocyte chemotactic protein
Mcp-1 Monocyte chemotactic protein 1
MDA malondialdehyde
MDA1 daily maximum 1-hour average
MDA8 daily maximum 8-hour average
MDL method detection limit
ME microenvironmental exposure; mean error
September 2019
liv
DRAFT: Do Not Cite or Quote
-------
MEGAN
MEIC
MERRA
MESA
MESA-Air
MFB
MFE
mg
MI
min
MINAP
MtP
MIROC
mL
ML
MLN
mm
MM5
MNB
MNE
MNGE
MO
MODIS
MOSES
MOVES
MOZART
MP
MP AN
Model of Emissions of Gases and Aerosols from Nature
Multiresolution Emissions Inventory for China
Modern-Era Retrospective analysis for Research and Applications (a NASA reanalysis of satellite
ozone data)
Multi-Ethnic Study of Atherosclerosis
Multi-Ethnic Study of Atherosclerosis and Air Pollution
mean fractional bias
mean fractional error
milligrams
myocardial infarction; myocardial ischemia
minute(s); minimum
Myocardial Ischaemia National Audit Project
macrophage inflammatory protein
Model for Interdisciplinary Research On Climate
milliliter
Mira Loma
mediastinal lymph node
millimeters
Mesoscale Model Version 5
mean normalized bias
mean normalized error
mean normalized gross error
month
MODerate resolution Imaging Spectroradiometer
Met Office Surface Exchange Scheme
U.S. EPA Motor Vehicle Emission Simulator
MOdel for Ozone and Related chemical Tracers
mid polar; myelopeptide; moist polar
peroxymethacrylic nitrate
September 2019
lv
DRAFT: Do Not Cite or Quote
-------
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
MYJ Mellore-Yamadae-Janjic
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 N-acetyl-P-D-glucosaminidase
September 2019
lvi
DRAFT: Do Not Cite or Quote
-------
NAI net annual increment
NAM Northern Annular Mode; North American mesoscale
NAMS National Air Monitoring Stations
NAPCA U.S. National Air Pollution Control Administration
NAPDH reduced form of nicotinamide adenine dinucleotide phosphate
NAPS National Air Pollution Surveillance
NAQFC National Air Quality Forecasting Capability
NASA U.S. National Aeronautics and Space Administration
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 non-detectable
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 non-filtered air
NFkB nuclear factor kappa light-chain-enhancer of activated B cells
ng nanogram
NGE normalized gross error
NGF nerve growth factor
NH3 ammonia
September 2019
lvii
DRAFT: Do Not Cite or Quote
-------
nh4+
NH4N03
NHANES
NHEERL
NHIRD
NHIS
NHIS-NCI
NHIS-NSC
NHLBI
NHS
NK
nL
NLDN
Nlrp
nm
NMB
NME
NMMAPS
NNs
NO
NO2
NO3-
NOAA
NOS
Notch3
Notch4
NOx
NOy
ammonium
ammonium nitrate
National Health and Nutrition Examination Survey
U.S. EPA National Health and Environmental Effects Research Laboratory
National Health Insurance Research Database
National Health Insurance Service
National Health Insurance Service - National Sample Cohort
National Health Insurance Service - National Sample Cohort
National Heart, Lung, and Blood Institute
Nurses' Health Study
neurokinin
nanoliter
National Lightning Detection Network
Nucleotide-binding oligomerization domain, Leucine rich Repeat and Pyrin domain containing
nanometer(s)
normalized mean bias
normalized mean error
U.S. National Morbidity, Mortality, and Air Pollution Study
the interval between normal beats
nitric oxide
Nitrogen dioxide
nitrate
U.S. National Oceanic and Atmospheric Administration
nitric oxide synthase
neurogenic locus notch homolog protein 3
neurogenic locus notch homolog protein 4
oxides of nitrogen (NO + N02)
the sum of NOx with its related reservoir forms (gas- phase HN03, PAN, HONO, N03, N205,
organic nitrates [RN03], and nitrate in particles [pN03])
September 2019
lviii
DRAFT: Do Not Cite or Quote
-------
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
0 outdoor ozone air concentration
The oxygen "singlet D" radical (a high energy, electronically excited form of the monatomic
OlD oxygen radical)
02 oxygen
03 ozone
OA objective analysis
obs observed
OC organic carbon
ODSs Ozone Depleting Substances
OE elevated ozone treatment
OGG1 8 oxo-guanine repair enzyme
OGTT oral glucose tolerance test
OH hydroxide; hydroxyl radical
OHCA out-of-hospital cardiac arrests
01 optimal interpolation
Oil ozone injury index
OK ordinary kriging
September 2019
lix
DRAFT: Do Not Cite or Quote
-------
OMB
OMI
ONPHEC
OPEC
OR
ORD
OSAT
OTC
OTUs
OZOVEG
P
PA
PAH
PAI
PAMS
PAN
PAR
PAT
PBL
PCA
PCI
PCR
PD
PDLR
PDO
PE
PECOS
PEF
Per
September 2019
U.S. EPA Office of Management and Budget
Ozone Monitoring Instrument
Ontario Population Health and Environment Cohort
Outdoor Plant Environment Chamber
odds ratio(s)
U.S. EPA Office of Research and Development
Ozone Source Apportionment Tool (in CAMx)
open-top chamber
operational taxonomic units
Ozone Vegetation Database
population; probability value
photoacoustic analyzer; physical activity; plasminogen activator; pascal(s); policy assessment
polycyclic aromatic hydrocarbons
plasminogen activator inhibitor, (e.g. PAI-1)
Photochemical Assessment Monitoring Stations
peroxyacetyl nitrate; peroxyacl nitrate
photo synthetically 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
lx DRAFT: Do Not Cite or Quote
-------
PFT
Pg
Pgam5
PGE2
PGF2a
PH
PI
PIAMA
PKS
PLANTS
PLFA
PLS
PM
PMio
PM2.5
PMAE
PMN
PMNs
PMSE
PN
PND
PNDIO
PND15
PND21
PND28
pNN50
POD
POD6
pulmonary function test
Petagram, equal to 1015 grams or one billion tonnes
phosphoglycerate mutase 7
prostaglandin E2
prostaglandin 2 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
particulate matter
particulate matter with an aerodynamic diameter less than or equal to 10 |xm
particulate matter with an aerodynamic diameter less than or equal to 2.5 |xm
Predictive Mean Absolute Error
polymorphonuclear leukocytes
polymorphic neutrophil
Predictive Mean Squared Error
particle number
postnatal day
postnatal day 10
postnatal day 15
postnatal day 21
postnatal day 28
proportion of pairs of successive normal simus intervals exceeds 50 milliseconds divided by the
total number of successive pairs of normal simus intervals
probability of detection; phytotoxic ozone dose
phytotoxic ozone dose above a threshold of 6 nmol/m2/s
September 2019
lxi
DRAFT: Do Not Cite or Quote
-------
PPb
ppbv
PPL
ppm
ppm-h
PPN
PPROM
ppt
PQAPP
PR
PR protein
PRB
PROM
Prxd
PTB
pts
PIT
PVD
PVN
PWA
Qi
Q2
Q3
Q4
QA
QAPP
QBME
QC
parts per billion
parts per billion by volume
potential productivity loss
parts per million
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
peroxypropionyl nitrate
preterm premature rupture of membranes
parts per trillion
Program-level Quality Assurance Project Plan
time interval between the beginning of the P wave to the peak of the R wave
pathogenesis related protein
Policy-relevant Background, typically used in the phrase, "PRB ozone."
premature rupture of membranes
peroxiredoxin
preterm birth
points
partial thomboplastin time
peripheral vascular disease
paraventricular nucleus
population weighted average
1 st quartile or quintile
2nd quartile or quintile
3rd quartile or quintile
4th quartile or quintile
Quality Assurance
Quality Assurance Project Plan
quantile-based Bayesian maximum entropy
quality control
September 2019
lxii
DRAFT: Do Not Cite or Quote
-------
QNSE Quasi Normal Scale Elimination
qPCR quantitative polymerase chain reaction
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 the Q wave to end of the T wave
QTc QT interval corrected for heart rate
r correlation coefficient
R6MA1 running 6-month average of the 1 hour daily max
Rag Recombination activating gene
RAMP Regionalized Air Quality Model Performance
RCGC Research and Development Center for Global Change
RCTs randomized clinical trials
REA risk and exposure assessment
REAS Regional Emissions inventory in Asia
redox reduction-oxidation
RF radiative forcing(s)
RFLP restriction fragment length polymorphism
RH relative humidity
RISCAT Cardiovascular Risk and Air Pollution in Tuscany
rMSSD root-mean-square of successive differences
RMSE root mean squared error
RNA ribonucleic acid
RNS reactive nitrogen species
ROCK rho associated kinase
ROS reactive oxygen species
RP-N reducing power of protein-binding compounds on nitrogen digestibility
RR risk ratio, relative risk
rRNA ribosomal ribonucleic acid
RuBisCO riibulose-l,5-bisphosphate carboxylase/oxygenase
RV right ventricular
September 2019
lxiii
DRAFT: Do Not Cite or Quote
-------
s 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; Sprague-Dawley rat
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 soluble fms-like tyrosine kinase 1
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
September 2019
lxiv
DRAFT: Do Not Cite or Quote
-------
SHIS
Si
SILAM
SIP
SJV
SLAMS
SMARTFIRE
SMBD
SMOKE
SO2
SOC
SOD
SOEP
SOLVEG
SOx
SoyFACE
sp
SP
SP+
spp
SPSH
sRaw
ST
STAT3
std
STE
STEMI
STN
STROBE
September 2019
Shanghai Health Insurance Study
silicon
System for Integrated modeLling of Atmospheric coMposition
State Implementation Plan
San Joaquin Valley
State and Local Air Monitoring Stations
Satellite Mapping Automated Reanalysis Tool for Fire Incident Reconciliation
Spanish Minimum Basic Data
Spare-Matrix Operator Kernel Emissions system
sulfur dioxide
semi-volatile organic compound
superoxide dismutase
Socioeconomic Panel
atmosphere-soil-vegetation land surface model
sulfur oxides
Soybean Free Air gas Concentration Enrichment (Facility)
species
surfactant protein (e.g., SPA, SPD)
substance-P-positive
several species
stroke-prone spontaneously hypertensive
specific airway resistance
spatiotemporal
signal transducer and activator of transcription 3
standard
strata sphere-troposphere exchange
ST-Elevation Myocardial Infarction
Speciation Trends Network
Strengthening the Reporting of Observational Studies in Epidemiology
lxv DRAFT: Do Not Cite or Quote
-------
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
TB tracheobronchial
TBARS thiobarbituric acid reactive substances
TC total hydrocarbon
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
HA transient ischemic attack
HD ter in die, three times per day
HMP tissue inhibitor of metalloproteinase
September 2019
lxvi
DRAFT: Do Not Cite or Quote
-------
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
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; United Kingdom
UNEP United Nations Environment Programme
UNFCCC United Nations Framework Convention on Climate Change
September 2019
lxvii
DRAFT: Do Not Cite or Quote
-------
UPA
URI
URTI
USB
USBAB
USDA
UV
UVAFME
V
VABS
VBS
VCAM
VEGF
VIIc
VIS
VOC
VPD
VPSC(s)
VSD
VT
VTI
VW
vWF
W126
WBC
WDCGG
WED
WHI-OS
unpaired nonnalized bias
upper respiratory infection
upper respiratory tract infection
United States Background
U.S. Background apportionment-based
United States Department of Agriculture
ultraviolet radiation
University of Virginia Forest Model Enhanced
meridional velocity of wind vector
Vineland Adaptive Behavior Scales
volatility basis set
vascular cell adhesion protein
vascular endothelial growth factor
Factor VII coagulant activity
visible (spectrum)
volatile organic compound
vapor pressure deficit
volatile plant signaling compound(s)
Very Simple Dynamic Model-soil biogeochemical process model
tidal volume
velocity time interval
Volkswagen
von Willebrand factor
cumulative integrated exposure index with a sigmoidal weighting function
white blood cell(s)
World Data Centre for Greenhouse Gases
U.S. EPA National Health and Environmental Effects Research Laboratory Western Ecology
Division
Women's Health Initiative Observational Study
September 2019
lxviii
DRAFT: Do Not Cite or Quote
-------
WHO World Health Organization
WISH Women's Isoflavone Soy Health
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 Forecasting
WRF-ARW WRF-Advanced Research WRF
WRF-Chem WRF with Chemistry
WRF-NMM WRF-Nonhydrostatic Mesoscale Model
WS wood smoke
WT wild type
WUE Water Use Efficiency
XO radical containing a halogen atom and an oxygen atom, X = I or Br
YIBs Yale Interactive Terrestrial Biosphere Model
Ym2 chitinase-like-4 protein
yr year(s)
YSU Yonsei University
ZCTAs zip-code tabulation areas
September 2019
lxix
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
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 that 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); CAA (1990)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[b]). 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 A secondary
standard, as defined in Section 109(b)(2), must "specify a level of air quality the attainment and
maintenance of which, in the judgment of the Administrator, based on such criteria, is requisite to protect
the public welfare from any known or anticipated adverse effects associated with the presence of [the] 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.
September 2019
lxx
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
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 so doing,
protection is provided for both the population as a whole and those groups and lifestages potentially at
increased risk for health effects from exposure to the air pollutant for which each NAAQS is set.
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 standards that are "requisite" to protect public health and welfare 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
1 See Lead Industries Association v. EPA, 647 F.2d 1130, 1154 [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)'Association of Battery Recyclers v. EPA, 604 F. 3d 613, 617-18 (D.C.
Cir. 2010).
2 See Lead Industries 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.
September 2019
lxxi
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
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 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 pollutant to be measured in the ambient air for the purpose of determining
compliance with the standard. The averaging time defines the time period over which air quality
measurements are to be obtained and averaged or cumulated. 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 annual Ozone NAAQS is 0.070 ppm as a 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 provided by the primary and secondary
NAAQS.
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
(VOC), 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 can vary 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 VOC concentrations increase and is described as
"VOC-limited." Once formed, ozone near the Earth's surface can be transported by the prevailing winds
1 The List of CASAC members is available at:
https://vosemite.epa.gov/sab/sabpeople.nsfAVebExternalCommitteeRosters?OpenView&committee=CASAC&seco
ndname=Clean%20Air%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.
September 2019
lxxii
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
before eventually being removed from the atmosphere over the course of hours to weeks via chemical
reactions or deposition to surfaces.
The U.S. EPA initially set primary and secondary NAAQS for photochemical oxidants in 1971,
with a 1-hour averaging 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 other 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.
September 2019
lxxiii
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Table 1
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
photochemical
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
greaterthan 0.12 ppm, is equal to or less than
1
58 FR 13008
March 9, 1993
U.S. EPA decided revisions to the standard 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 distinct 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).
September 2019
lxxiv
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
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 petitioned for a rehearing en banc
on several issues related to that decision. The court granted the request for rehearing in part and denied it
in part, but declined to review its ruling with regard to the potential beneficial effects of 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, hereafter referred to as "ATA IIP').
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 comment 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. The U.S. EPA completed the review in March 2008 by
September 2019
lxxv
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
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 delays in reaching a final decision, 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), 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 PA3 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/85f90b771 Iacb0c88525763300617d0d.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
(Frev. 2014a 2014. 5408574 2014. 5408574. c; U.S. EPA. 2014a. b, c).
September 2019
lxxvi
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
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 are too stringent, while the environmental
and health petitioners' brief argued that the revised standards are not stringent enough to protect public
health and welfare as the Act requires. On August 23, 2019, the court issued an opinion that denied all the
petitions for review with respect to the 2015 primary standard while also concluding that the EPA had not
provided a sufficient rationale for aspects of its decision on the 2015 secondary standard and remanding
that standard to the U.S. EPA {Murray Energy v. EPA, No. 15-1,385, Order, Doc. No. 1803352 [D.C. Cir.
August 23, 2019]).
Purpose and Overview of the Integrated Science Assessment
The Integrated Science Assessment (ISA) is a comprehensive evaluation and synthesis of the
policy-relevant science "useful in indicating the kind and extent of identifiable effects on public health or
welfare which may be expected from the presence of [a] pollutant in 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 serves as the scientific foundation for the review of the current
primary (health-based) and secondary (welfare-based) National Ambient Air Quality Standards (NAAQS)
for ozone.
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
1 The current NAAQS for ozone are specified at 40 CFR 50.19.
September 2019
lxxvii
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
assessments of ozone, specifically, the 2006 AQCD for Ozone and Related Photochemical Oxidants (U.S.
EPA. 2004). the 2007 Staff Paper (U.S. EPA. 2007). the 1996 AQCD and Staff Paper for Ozone and
Other 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 several
policy-relevant questions identified in the Integrated Review Plan for the National Ambient Air Quality
Standards for Ozone (https://www.epa.gov/naaqs/ozone-o3-standards-planning-documents-current-
review). 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.
September 2019
lxxviii
DRAFT: Do Not Cite or Quote
-------
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.
September 2019
lxxix
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
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 (2015a).
As part of this process, the ISA is reviewed by the CASAC, which is a formal independent panel
of scientific experts, 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 ISA for ozone 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
conclusions. Section 10.2. "Literature Search and Initial Screen," details the study selection process in
further detail.
September 2019
lxxx
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
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?
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
September 2019
lxxxi
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
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 to focus overall conclusions specifically on birth
outcomes and on fertility and pregnancy effects separately.
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 [OTC], free-air carbon dioxide enrichment
[FACE]), 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 all the
policies and procedures identified therein. Additionally, this ISA follows all of 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
effects by integrating information across scientific disciplines and related health outcomes 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 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 loosely based on the 50th-95th percentile of
concentrations observed for each averaging time and exposure 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
September 2019
lxxxii
DRAFT: Do Not Cite or Quote
-------
1 evidence across scientific disciplines, supports a thorough evaluation of the current state of the science for
2 ozone.
3 In evaluating the evidence, determinations are made about causation, not just association, and are
4 based on judgments of aspects such as the consistency of evidence within a discipline, coherence of
5 effects across disciplines, and biological plausibility of observed effects. Determinations account for
6 related uncertainties. The ISA uses a formal causal framework [Table II of the Preamble to the ISAs (U.S.
7 EPA. 2015a) I to classify the weight of evidence according to the five-level hierarchy.
September 2019
lxxxiii
DRAFT: Do Not Cite or Quote
-------
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/fdsYs/pkg/USCODE-2005-title42/pdf/USCODE-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/264cbl227d55e02c85257402007446a4/5EFA32QCCAD326E88
5257D030071531C/$File/EPA-CASAC-14-004+unsigned.pdf (accessed
Frev. HC. (2014b). Letter from Dr. H. Christopher Frey, Chair, Clean Air Scientific Advisory Committee, to
Administrator Gina McCarthy. CASAC Review of the EPAs Welfare Risk and Exposure Assessment for
Ozone (Second External Review Draft). EPACASAC-14-003. Available online at
https://vosemite.epa.gov/sab/sabproduct.nsf/264cbl227d55e02c852574020Q7446a4/3D561C9413E49E8E85
257CFB0069E0DF/$File/EPA-CASAC-14-003+unsigned.pdf (accessed
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/ZvPURL.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://cfpub.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?searchOuerv=PB96185608.
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?searchOuerv=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/2017-03/documents/epa-info-qualitv-guidelines.pdf.
U.S. EPA (U.S. Environmental Protection Agency). (2004). Air quality criteria for particulate matter [EPA
Report]. (EPA/600/P-99/002aF-bF). Research Triangle Park, NC: U.S. Environmental Protection Agency,
Office of Research and Development, National Center for Environemntal Assessment- RTP Office.
http://cfpub.epa. gov/ncea/cfm/recordisplav.cfm?deid=87903.
September 2019
lxxxiv
DRAFT: Do Not Cite or Quote
-------
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/naaas/standards/ozone/data/20Q7 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://cfpub.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=P 100KBUF.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/naaas/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/naaas/standards/ozone/data/2014102 lwelfarerea.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.
September 2019
lxxxv
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
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 the 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 (i.e., causality determinations; Section ES.4) are
presented and explained. They 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
information across scientific disciplines and building off the evidence base evaluated in previous reviews.
The ISA thus provides the policy-relevant scientific information that supports the review of the NAAQS.
This executive summary provides an overview of the important conclusions drawn in the ISA
across the scientific disciplines, beginning with information on sources, concentrations, estimated
background and exposure, followed by health and welfare effects. A more detailed summary of the
evidence is presented in the Integrated Synthesis, and individual appendices for each topic area include
study-level information and an in-depth characterization of the weight-of-evidence conclusions.
1 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. 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).
September 2019
ES-1
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
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 throughout the troposphere also by reactions of
CO and CH4 with oxides of nitrogen (Section 1.4). Recent developments in understanding ozone
chemistry include observations of high ozone concentrations during the winter in some western U.S.
mountain basins (Section 1.4.1) 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).
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) o 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.
September 2019
ES-2
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
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. USB is not measured directly but is estimated based on 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.
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,
is often used as a surrogate 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. IV 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. Robust exposure indices that quantify exposure as it relates to measured plant response
(e.g., growth) have been in use for decades and only require ambient air quality data. Exposure duration
influences the degree of plant response, and ozone effects on plants are cumulative. Cumulative indices
summarize ozone concentrations over time and provide a consistent metric for reviewing and comparing
exposure-response effects obtained from various studies. Cumulative indices of exposure that
September 2019
ES-3
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
differentially weight hourly concentrations are, therefore, best suited to characterize vegetation exposure
to ozone (Section 8.1.2.1V
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 causal 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". These causality 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 5.2.3V 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
ozone-induced 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.
September 2019
ES-4
DRAFT: Do Not Cite or Quote
-------
Table ES-1 Summary of causality determinations by exposure duration and
health outcome.
Health Outcome3
Conclusions from 2013 Ozone
ISA
Conclusions in the Current ISA
Short-term exposure to ozone
Respiratory effects
Causal relationship
Causal relationship
Metabolic effects
No determination made
Likely to be a causal relationship15
Cardiovascular effects
Likely to be a causal relationship
Suggestive of, but not sufficient to infer, a causal
relationship0
Total mortality
Likely to be a causal relationship
Suggestive of, but not sufficient to infer, a causal
relationship
Central nervous system
effects
Suggestive of a causal
relationship
Suggestive of, but not sufficient to infer, a causal
relationship
Long-term exposure to ozone
Respiratory effects
Likely to be a causal relationship
Likely to be a causal relationship
Metabolic effects
No determination made
Likely to be a causal relationship15
Cardiovascular effects
Suggestive of a causal
relationship
Suggestive of, but not sufficient to infer, a causal
relationship
Total mortality
Suggestive of a causal
relationship
Suggestive of, but not sufficient to infer, a causal
relationship
Reproductive effects
Suggestive of a causal
relationship
Effects on fertility and reproduction: suggestive of a
causal relationship15
Effects on pregnancy and birth outcomes:
suggestive of a causal relationship15
Central nervous system
effects
Suggestive of a causal
relationship
Suggestive of, but not sufficient to infer, a causal
relationship
Cancer
Inadequate to infer a causal
relationship
Inadequate to infer 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.
September 2019
ES-5
DRAFT: Do Not Cite or Quote
-------
Causality Determinations for Health Effects of Ozone
ISA
Current Ozone Draft ISA
Health Outcome
Respiratory
Short-term
exposure
Long-term
exposure
Metabolic
Short-term
exposure
+
Long-term
exposure
+
Cardiovascular
Short-term
exposure
*
Long-term
exposure
Nervous System
Short-term
exposure
Long-term
exposure
Reproductive
Male/Female
Reproduction
and Fertility
Long-term
*
Pregnancyand
Birth Outcomes
exposure
*
Cancer
Long-term
exposure
Mortality
Short-term
exposure
*
Long-term
exposure
Causal ¦ Likely causal 1 Suggestive Inadequate
+ new causality determination
* change in causality determination from 2013 Ozone ISA
Figure ES-2 Causality determinations for health effects of short- and
long-term exposure to ozone.
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 3.1.11). and there
September 2019
ES-6
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
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). respiratory
symptoms, and respiratory tract inflammation. Epidemiologic studies continue to provide evidence that
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 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.
September 2019
ES-7
DRAFT: Do Not Cite or Quote
-------
8
"§ +£ 6
O c
3 g
TS E
C O
¦ u 4
m U H
= £
O ^
N -
o >
LU 2
LL.
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)
V
0
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 (2006) found statistically
significant responses to square-wave chamber exposures at 60 ppb based on the analysis of Brown et al. (2008) and Kim et al.
(2011). During each hour of the exposures, subjects were engaged in moderate quasi-continuous exercise (20 L/min 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 for illustrative purposes. The McDonnell et al. (2013) illustrates the predicted FE\A| 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. (2011) study, but only published in Figure 5 of McDonnell
et al. (2012).
Adapted from Figure 6-1 of 2013 Ozone ISA (U.S. EPA. 2013). Studies appearing in the figure legend are: Adams (2006). Adams
(2003). Adams (2002). Folinsbee et al. (1988). Horstman et al. (1990). Kim et al. (2011). McDonnell et al. (2013). McDonnell et al.
(1991). and Schelegle et al. (2009).
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.
1 Metabolic effects related to ozone exposure are evaluated as a separate health endpoint category
2 for the first time in this ISA (Appendix 5). Recent evidence from animal toxicological, controlled human
3 exposure, and epidemiologic studies indicate that there is a "likely to be causal relationship" between
4 short-term ozone exposure and metabolic effects (Section 5.1.8). The strongest evidence for this
5 determination is provided by animal toxicological studies that demonstrate impaired glucose tolerance,
6 increased triglycerides, fasting hyperglycemia, and increased hepatic gluconeogenesis in various strains
7 of animals across multiple laboratories. Biological plausibility is provided by results from controlled
8 human exposure and animal toxicological studies that demonstrate activation of sensory nerve pathways
9 following ozone exposure triggers the central neuroendocrine stress response, which includes increased
September 2019
ES-8
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
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.
Similarly, there is a "likely to be causal relationship" between long-term ozone exposure and
metabolic effects (Section 5.2.11). Animal toxicological studies of long term-ozone exposure also
provide evidence for impaired insulin signaling, glucose intolerance, hyperglycemia, and insulin
resistance. In prospective cohort studies conducted in the U.S. and Europe, increased incidence of type 2
diabetes was observed with long-term ozone exposure. In a large, population-based study in China, the
risk of metabolic syndrome was also increased. These results are also consistent with two long-term
ozone exposure studies in China that observed increased risk of obesity (a risk factor for type 2 diabetes)
in both adults and children. In epidemiologic studies, positive associations between long-term exposure to
ozone and diabetes-related mortality were reported in well-established cohorts in the U.S. and Canada.
The results of the morbidity and 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).
Notably, there are changes in the causality determinations for short-term ozone exposure and
cardiovascular effects (Appendix 4) as well as total mortality (Appendix 6). 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 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
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.
September 2019
ES-9
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
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 8) and effects on climate (Appendix 9). For ecological effects, damage to
terrestrial ecosystems caused by ozone 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). Ozone can result in
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).
September 2019
ES-10
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
...... ....
*r '
03 exposure
* 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
1 -Altered crop quality
1
Belowground processes
•Altered litter production and decomposition
•Altered soil carbon and nutrient cycling
•Altered soil fauna and microbial communities
D
CD
—i
CD
U
O
< .i
CO
CD
=3
CO
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 grow th
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.
September 2019 ES-11 DRAFT: Do Not Cite or Quote
-------
Causality Determinations for Ecological Effects of Ozone
Scale of Ecological Response
Belowground
Biogeochemical Cycles
"
\
*>
>
Water Cycling
D
u
Ui
Carbon Sequestration
Productivity
Community
Biodiversity
Terrestrial Community Composition*
Species Interactions
Plant-Insect Signaling +
Population
rO
3
JG
C
Survival
Trees+
Growth
Plants
Herbivores +
Reproduction
Plants*
Herbivores +
Yield
Agricultural Crops
Individual
Visible Foliar Injury
Causal Likely Causal |
new determination (+) or change in causality determination (*) from 2013 Ozone ISA
Figure ES-5 Causality determinations for ozone across biological scales of
organization and taxonomic groups.
September 2019
ES-12
DRAFT: Do Not Cite or Quote
-------
Table ES-2 Summary of causality determinations for ecological effects.
Endpoint
Conclusions from 2013 Ozone
ISA
Conclusions in the current 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 a 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 a causal
relationship3
Alteration of plant-insect signaling
Causality not assessed
Likely to be a causal
relationship3
Reduced productivity in terrestrial
ecosystems
Causal relationship
Causal relationship
Reduced carbon sequestration in terrestrial
ecosystems
Likely to be a causal relationship
Likely to be a causal relationship
Alteration of belowground biogeochemical
cycles
Causal relationship
Causal relationship
Alteration of terrestrial community
composition
Likely to be a causal relationship
Causal relationship15
Alteration of ecosystem water cycling
Likely to be a causal relationship
Likely to be a causal relationship
aDenotes new causality determination.
bDenotes change in causality determination from 2013 Ozone ISA.
1 Visible foliar injury resulting from exposure to ozone has been well characterized and
2 documented in over six decades of research involving many tree, shrub, herbaceous, and crop species and
3 using both long-term field studies and laboratory approaches. Recent experimental evidence (Section 8.2)
4 continues to show a consistent association between visible injury and ozone exposure supporting the
5 conclusion of the 2013 Ozone ISA that, there is a "causal relations hip" between ozone and visible
6 foliar injury. Changes in photosynthesis and carbon allocation in ozone-exposed plants scale up to
7 reduced growth documented in natural and managed (e.g., agriculture, forestry, landscaping) species
8 (Section 8.3). as well as impaired reproduction in individual plants (Section 8.4.1). Consistent with the
September 2019
ES-13
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
conclusions in the 2013 Ozone ISA, there is a "causal relationship" 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., 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 support 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 studies covering a range of species at varying levels of ozone
exposure frequently show statistically significant effects; however, they do not provide any consistent
pattern of response across endpoints of growth or reproduction (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). 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 8.8V As
described in the 2013 Ozone ISA and new 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
September 2019
ES-14
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
functioning, loss of leaf area, and 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
(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 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 current ISA
Radiative forcing
Causal relationship
Causal relationship
Temperature, precipitation, and related Likely to be a causal relationship Likely to be a 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.. 201311.
September 2019
ES-15
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
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, and the extent to which recent studies
have addressed or reduced uncertainties from previous assessments, as well as 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 ozone exposure. 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 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, 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.
There is strong and consistent experimental evidence linking short- and long-term ozone exposure
with respiratory and 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 reasonable study 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
September 2019
ES-16
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
controlled human exposure studies. Additionally, a number of animal toxicological studies are performed
in rodent disease models, while controlled human exposure studies generally are conducted in healthy
individuals. Controlled human exposure studies do not 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). Additionally, 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 any lack of coherence between results of experimental animal and human
studies.
Epidemiologic studies contribute important evidence supporting the relationship between short-
and long-term ozone exposure with respiratory and metabolic health 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. 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
September 2019
ES-17
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
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 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 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 tropospheric ozone on global 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
September 2019
ES-18
DRAFT: Do Not Cite or Quote
-------
1 understanding. While the warming effect of tropospheric ozone in the climate system is well established
2 in general, precisely quantifying changes in surface temperature due to tropospheric ozone changes, along
3 with related climate effects, requires complex climate simulations, including important feedbacks and
4 interactions. Current limitations in climate modeling tools, variation across models, and the need for more
5 comprehensive observational data on these effects represent sources of uncertainty in quantifying the
6 precise magnitude of climate responses to ozone changes, particularly at regional scales.
September 2019
ES-19
DRAFT: Do Not Cite or Quote
-------
ES.6 References for Executive Summary
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.
https://doi.org/10.1080/089583703045Q5.
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.
Browa 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.
Folinsbee. LJ: Mcdonnell. WF: Horstmaa PH. (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.
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/10.1164/airccm/142.5.1158.
Kim. CS: Alexis. NE: Rappold. AG: Kehrl. H: Hazucha. MJ: Lav. JC: Schmitt. MT: Case. M: Devlin. RB:
Pedea 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: Devlia 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. (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: Berntsea TK: Bian. H: Bellouin. N: Chin. M: Diehl.
T: Easter. RC: Feichter. J: Ghaa SJ: Hauglustaine. D: Iversen. T: Kinne. S: Kirkevag. A: Lamarque. JF: Lin.
G: Liu. X: Lund. MT: Luo. G: Ma. X: vanNoiie. 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: Yooa JH: Zhang. K: Zhang.
H: Zhou. C. (2013). Radiative forcing of the direct aerosol effect from AeroCom Phase II simulations. Atmos
ChemPhys 13: 1853-1877. http://dx.doi.org/10.5194/acp-13-1853-2013.
Schelegle. ES: Morales. CA: Walbv. WF: Marion. S: Allea 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.
September 2019
ES-20
DRAFT: Do Not Cite or Quote
-------
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.
September 2019
ES-21
DRAFT: Do Not Cite or Quote
-------
INTEGRATED SYNTHESIS
Overall Conclusions of the (hone Integrated Science. \ssessment (IS. \)
Human Health l:JJect\
• Recall studies support ;ind c\p;iud upon the stroim bod\ of c\ ideiicc. which h;is been
; 1 11111111;1111iu o\ cr llie List lew decides. lh;il short-term oaiiic exposure c;iuscs rcspir;ilor\
effects I lie strongest c\ ideiicc comes from controlled I in 11 i;i 11 exposure studies dcnioiisir;iliim
oAine-induced decreases mi hum I'm n (> hours of exposure In ;iddiliou. epidemiologic studies
continue lii pro\ ide strong c\ ideiicc lh;il oaiiic is ;issoci;ilcd w illi rcspir;ilor\ effects. includum
;isihni;i ;md ('()l'l) c\;iccrb;ilious. ;is well ;is hospital ;idniissious ;md cnicrucucs dcpminicul
\isiis for rcspir;itor\ discuses The results from to\icolomc;il siudics rurihci' ch;ii;iclei'i/c
pciienli;iI iiicch;iiiisiic p;ilhw;i\ s ;ind pro\ ide continued support liir ilic biolouicnl pkiusihiIii\ of
n/iHic-iiidiiccd iespii;iun\ cITecls
• I !nicrmim c\ ideiicc nidic;ilcs ilini short- ;md loim-lcrni i»a»iic exposure contributes in nicl;ibohc
disease, iiichidiiiu di;ihelcs Spccil'iciillx. ;iiiiiii;iI lo\icolomc;il siudics dcnioiisir;itc impaired
•ilucnsc lolcmucc. me reused tnuKccridcs. Iiisiiiii; h\ perul\ccnii;i. ;md nicrc;ised hep;ilic
ulucoucoucucsis in l;ibor;itor\ ;iniiii;iIs \ limned number i»l" epidemiologic siudics obser\ cd
;issoci;ilious between oaiiic ;md uicrc;ised iiicidcuce of i\pe 2 di;ibcles ;md niori;ilit\ from
di;ibelcs
• I jcc;iiisc reccni e\ ideiicc from controlled hiini;iu exposure siudics pro\ ides inconsistent c\ ideiicc
of OAine-induced c;irdio\ ;iseul;ir effects. ilic o\cr;ill bod\ of c\ ideiicc for;iu ;issi.>ci;ilk>ii of
slk»ri-ierni i»a»iic e\pi>sure w 11li c;udio\ ;iscul;ir cl lecls ;uid lot;il (iiou;iccidciii;il i niort;ilii\ is less
ccri;iiu lh;ui reported mi ilic 2d I ' ( )aiiic IS \. resiilliuu mi ;i ckiuuc mi ilic c;ius;ilit\
dclermiu;iiKd \e;irs cle;irl\ demonstrates ih;ii oaiiic
;illccls \ cucl;iliou ;md ecos\ stems The stroimcst c\ ideiicc comes from \ eizcl;ilK>ii-rel;ilctl
cudpoiuls. foli;ir injurs. reduced urowtli. ;md decreased \ icld ;ire well ch;ir;iclcri/ed m ;i \ ;iricl>
of crop ;iud noiicrop species. I !colomc;il cl lecls of oaiiic lire obser\ ed ncross sc\ cr;il sc;ilcs of
biolomc;il oru;uiiAiliou (i e . Ironi ilie ccllukir lc\cl throimh iudi\ idiuil oru;iinsnis in ilie lc\ cl of
communities ;uid ecos\ sicuisi. uliini;iicl> ;ilfccliim ;ibo\curouud ;md below mound processes
iiicludum products n\. c;irbou sa|iicsir;iiiou. bioucochcniic;il cscliuu ;md hulrolous lu most
c;iscs. new research sireuulliens ilic pre\ lousk reached conclusions mi ilic 2d I i Oaiiic IS \ New
cud points included mi this rc\ lew result from cnieruiuu ;ire;is of studs such ;is chcniic;il eciilnus
ic u . pl;uil-iusecl simiiilium or new c\ ideiicc cu;ibliuu I'lii'thcr reriiicnieui ii|"pre\ kmisIs
iiudcrstiuid i>/i>ue cllecls ic u . pl;uit repriiducliDii. tree iikiic IS \ ;md aa>iic inip;icls on r;idi;ili\ c liirciuu ;uid clini;ilc \ ;iri;ibles.
uicliidiuu tcniper;iiiirc mid precipitation i referred to ;is "chni;itc ch;iimc"' mi the 2d I ' ( )a>iic
IS \i
September 2019
IS-1
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
IS.1
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 identifiable effects on
public health or welfare2 which may be expected from the presence of [a] pollutant in ambient air," as
described in Section 108 of the Clean Air Act (CAA. 1990V3 This ISA reviews and 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 synthesize 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 which is for the annual fourth-highest daily maximum 8-hour concentration
averaged over 3 years5. The ozone primary NAAQS is established to protect public health with an
adequate margin of safety, including at-risk populations such as children and people with asthma. The
ozone secondary standard is intended to protect the public welfare from known or anticipated adverse
effects associated with the presence of ozone 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, animal toxicology, atmospheric science, exposure science, vegetation studies,
agricultural science, ecology, and climate-related science. Key scientific conclusions (i.e., causality
determinations; Section IS. 1.2.4) are presented that provide the basis for developing risk and exposure
analyses, evaluating policy, and making environmental health decisions. In characterizing the evidence
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, 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).
September 2019
IS-2
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
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 effects by integrating information across
scientific disciplines and related health and welfare outcomes and synthesizing evidence from previous
and recent studies. As in previous reviews, the ISA for this review will focus mainly on the assessment of
health and welfare effects resulting from exposure to surface-level concentrations of tropospheric ozone.
Less emphasis will be accorded to other photochemical oxidants for which there is distinctly much less
information. 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 The 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 generally 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 identifies and evaluates studies published since 2011, synthesizing and
integrating the new evidence into 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 to communicate 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, this
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 ISA development 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
and welfare effects [see Preamble (U.S. EPA. 2015)1. Development of the ISA includes approaches for
1 Ozone is the only photochemical oxidant other than nitrogen dioxide (NO2) that is routinely monitored in ambient
air (i.e., EPA's AQS database; https://www.epa.gov/aqs). 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.
September 2019
IS-3
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
literature searches, criteria for selecting and evaluating relevant studies, and a 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. 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 shielding was evaluated in the 2013 Ozone ISA and determined
to be inadequate to draw a causal conclusion; this continues to be the case in the current ISA
(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
scoping that generally defines the relevant Population, Exposure, Comparison, Outcome, and Study
Design (PECOS) for each of the scientific disciplines that form the basis of the evaluation of evidence for
the broad health and welfare effects categories for which this ISA forms causality determinations. 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
around the evidence base at the time of the last review (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).
September 2019
IS-4
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
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 will 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 study design, 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 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 is 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 draft ISA. Furthermore, the Ozone ISA is classified as providing
Highly Influential Scientific Assessment (HISA), which is defined by the Office of Management and
September 2019
IS-5
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
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 will be considered in the development of a final 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 health and welfare effects for categories of related 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, as well as 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 a causal relationship"; "suggestive of, but not sufficient to infer, a
causal relationship"; "inadequate to infer a causal relationship"; "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 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.
September 2019
IS-6
DRAFT: Do Not Cite or Quote
-------
Table IS-1 Summary of causality determinations by exposure duration and
health outcome.
Health Outcome3
Conclusions from 2013 Ozone ISA
Conclusions in the Current ISA
Short-term exposure to ozone
Respiratory effects
Causal relationship
Causal relationship
Metabolic effects
No determination made
Likely to be a causal relationship15
Cardiovascular effects
Likely to be a causal relationship
Suggestive of a causal relationship0
Total mortality
Likely to be a causal relationship
Suggestive of a causal relationship0
Central nervous system
effects
Suggestive of a causal relationship
Suggestive of a causal relationship
Long-term exposure to ozone
Respiratory effects
Likely to be a causal relationship
Likely to be a causal relationship
Metabolic effects
No determination made
Likely to be a causal relationship15
Cardiovascular effects
Suggestive of a causal relationship
Suggestive of a causal relationship
Total mortality
Suggestive of a causal relationship
Suggestive of a causal relationship
Reproductive effects
Suggestive of a causal relationship
Effects on fertility and reproduction:
suggestive of a causal relationship15
Effects on pregnancy and birth outcomes:
suggestive of a causal relationship15
Central nervous system
effects
Suggestive of a causal relationship
Suggestive of a causal relationship
Cancer
Inadequate to infer a causal relationship
Inadequate to infer a causal relationship
aHealth effects (e.g., respiratory effects, cardiovascular effects) include the spectrum of outcomes, from measurable subclinical
effects (e.g., FE\A, 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.
1 The strongest evidence for health effects due to ozone exposure continues to come from studies
2 of short- and long-term ozone exposure and respiratory health. Consistent with conclusions from the 2013
3 Ozone ISA, there is a "causal relationship" between short-term ozone exposure and respiratory effects,
4 and there is a "likely to be causal relationship" between long-term ozone exposure and respiratory effects.
5 For short-term ozone exposure, controlled human exposure studies provide experimental evidence for
6 ozone-induced lung function decrements, respiratory symptoms, and respiratory tract inflammation.
September 2019
IS-7
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
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 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, 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 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 indicate that there is a "likely to be causal relationship" between short-term ozone
exposure and metabolic effects. The strongest evidence is provided by animal toxicological studies that
demonstrate impaired glucose tolerance, increased triglycerides, fasting hyperglycemia, and increased
hepatic gluconeogenesis in various strains of animals across multiple laboratories. Biological plausibility
is provided by controlled human exposure and animal studies that demonstrate activation of sensory
pathways following ozone exposure, which triggers the central neuroendocrine stress response, and
results in increased corticosterone, Cortisol or epinephrine, as noted in the description of the controlled
human exposure study. These findings are coherent with epidemiologic studies that report associations
between ozone exposure and perturbations to glucose and insulin homeostasis. Animal toxicological
studies also provide evidence for impaired insulin signaling, glucose intolerance, hyperglycemia, and
insulin resistance after long-term exposure. In addition, these pathophysiological changes are often
accompanied by increased inflammatory markers in peripheral tissues and activation of the
neuroendocrine system. In prospective cohort studies in the U.S. and Europe, increased incidence of
type 2 diabetes was observed with long-term ozone exposure. In China, the odds of metabolic syndrome
increased as well. These findings are consistent with two long-term ozone exposure studies in China, one
in adults and one in children, that presented increased odds of obesity (a risk factor for type 2 diabetes) in
both adults and children. In epidemiologic studies, 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). There is a "likely to be causal relationship"
between long-term ozone exposure and metabolic effects.
September 2019
IS-8
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Notably, 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 draft 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 from a lack of control for potential 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 greater 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
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.
September 2019
IS-9
DRAFT: Do Not Cite or Quote
-------
Table IS-2 Summary of causality determinations for ecological effects.
Endpoint
Conclusions from 2013 Ozone
ISA
Conclusions in the Current
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 a 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 a causal
relationship3
Alteration of plant-insect signaling
Causality not assessed
Likely to be a causal
relationship3
Reduced productivity in terrestrial
ecosystems
Causal relationship
Causal relationship
Reduced carbon sequestration in terrestrial
ecosystems
Likely to be a causal relationship
Likely to be a causal relationship
Alteration of belowground biogeochemical
cycles
Causal relationship
Causal relationship
Alteration of terrestrial community
composition
Likely to be a causal relationship
Causal relationship15
Alteration of ecosystem water cycling
Likely to be a causal relationship
Likely to be a causal relationship
aDenotes new causality determination.
bDenotes change in causality determination from 2013 Ozone ISA.
IS.1.3.3 Welfare: Effects on Climate
1 Recent evidence continues to support a causal relationship between tropospheric ozone and
2 radiative forcing and a likely to be causal relationship, via radiative forcing, between tropospheric ozone
3 and temperature, precipitation, and related climate variables (referred to as "climate change" in the 2013
4 Ozone ISA; the revised title for this causality determination provides a more accurate reflection of the
5 available evidence ITablc IS-31). The new evidence comes from the Intergovernmental Panel on Climate
September 2019
IS-10
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Change (IPCC) Fifth Assessment Report (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. The new
studies further support the causality determinations included in the 2013 Ozone ISA.
Table IS-3 Summary of causality determinations for tropospheric ozone effects
on climate.
Conclusions in 2013 Ozone ISA
Conclusions in the Current ISA
Radiative forcing
Causal relationship
Causal relationship
Temperature, precipitation, and related
climate variables
Likely to be a causal relationship
Likely to be a 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 due to
natural processes and anthropogenic activities to ambient ozone concentrations. The section summarizes
recent developments in measurement and modeling methods, atmospheric chemistry, and ambient air
concentration trends (Section IS.2.1.). 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.2.2.3).
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 primarily produced 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
September 2019
IS-11
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
organic compounds, which may be found in the upper troposphere (Section 1.4). The most abundant
national and global sources of VOCs are biogenic (U.S. EPA. 2013b). 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 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). For
example, wintertime ozone concentrations in the 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, 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. 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 "maximum
daily 8-hour average" (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).
September 2019
IS-12
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
The 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 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 to 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
September 2019
IS-13
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
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
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
contribute about 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 1.3.1V 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
(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 form 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).
September 2019
IS-14
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
IS.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
alternative to model sensitivity approaches, source apportionment techniques track source contributions to
ozone formation without perturbing emissions (Section 1.8.1.2). 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 determining 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 determining 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 (Jaffe 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. 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 large 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
September 2019
IS-15
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
West and Southwest. Higher USB concentrations were also estimated at elevations greater 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
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, 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 little 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), indoor ozone concentrations were roughly 10-30% of
ambient air concentrations, and 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
September 2019
IS-16
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
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 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 concentrations
measured at fixed-site ambient air monitors are still widely used as surrogates for ozone exposure in
epidemiologic studies (Section 2.3.1.1). 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 2.3.2.4). 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
September 2019
IS-17
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
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
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 in regard
to 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).
September 2019
IS-18
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
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
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.
• 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.
• 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 independently in the appendices of this ISA, though
the mechanisms and disease progression leading to these health effects are not restricted to a single organ
system. Here, a high-level overview of how different health effects may be connected, and how insults to
one organ system are likely to affect others, is provided. This section provides a more holistic perspective
of the relationship between ozone and health than what is found in the individual health appendices.
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 ITJ.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 create 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 other body systems, there is the
opportunity for insults to multiple organ systems to contribute to a single health effect. The 2006 Ozone
AQCD rU.S. EPA (2006a): Chapter 4] and the 2013 Ozone ISA ITJ.S. EPA (2013b): Section 5.31 provide
extensive background on dosimetry and potential pathways and potential pathways underlying health
effects for these responses.
September 2019
IS-19
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
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, blood glucose, glycogenolysis/gluconeogenesis, hormone
release, and other organ functions (McCorrv. 2007). Endocrine, immune, and inflammatory responses can
also send signals capable of altering multiple pathways and eliciting cardiovascular, respiratory, and
metabolic health effects.
One example of a multisystem disruption resulting from ozone exposure is the decrease in core
body temperature observed in rats. This decrease affects metabolic rate, leading to decreased oxygen
consumption, reduced minute ventilation, decreased HR, decreased thyroid hormone concentrations, and
lowered blood pressure, among other physiological changes (Watkinson et al.. 2003; Mautz and Bufalino.
1989). As discussed in Appendix 5 (Section 5.1). high blood pressure is a component of metabolic
syndrome, while obesity, metabolic syndrome, and type 2 diabetes are risk factors for cardiovascular
disease, creating a two-way relationship for disease progression between the systems.
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
September 2019
IS-20
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
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
and effects at the population level. The boxes are color-coded according to their position in the exposure
to outcome continuum.
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
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.
September 2019
IS-21
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
Key Clinical Effect
Ozone
Exposure
Effect at Population
Level
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 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.
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 penod 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
plausibi lity 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
September 2019
IS-22
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
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 a "suggestive of but not sufficient to infer" or "inadequate" causality determination
has been concluded are noted in Table IS-1. and more fully discussed in the respective health effects
appendices.
IS.4.3.1 Short-Term Exposure and Respiratory Health Effects
Section 2.8 of 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.
Evidence from recent controlled human exposure studies augment 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
1 Concentrations from controlled human exposure studies are target concentrations, unadjusted for study-specific
measurement information.
September 2019
IS-23
DRAFT: Do Not Cite or Quote
-------
1 department visits observed associations at concentrations as low as 31 ppb. Controlled human exposure
2 studies also provide consistent evidence of ozone-induced increases in airway responsiveness
3 (Section 3.1.4.3 and Section 3.1.5.5) and inflammation in the respiratory tract (Section 3.1.4.4 and
4 Section 3.1.5.6V Recent animal toxicological studies are consistent with evidence summarized in the 2013
5 Ozone ISA (U.S. EPA. 2013b); these studies support the evidence observed in healthy humans.
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.
Conclusions from 2013 Ozone ISA
Results and Conclusions from 2019 ISAa
Respiratory effects
Evidence integrated across controlled
human exposure, epidemiologic, and
animal toxicological studies and across
the spectrum of respiratory health
endpoints demonstrated that there was
causal relationship between
short-term ozone exposure and
respiratory health effects.
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.
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
(7:50 a.m.-5:50 p.m.).
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.
September 2019
IS-24
DRAFT: Do Not Cite or Quote
-------
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 2019 ISAa
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 decreases 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 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.
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.
September 2019
IS-25
DRAFT: Do Not Cite or Quote
-------
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 2019 ISAa
Allergic and asthma
related responses
Controlled human exposure studies in
atopic individuals with asthma
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
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 max
ozone concentrations between 31 and 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 daily 8-h 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).
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 2019 ISA include evidence from recent studies integrated with evidence included in previous Ozone ISAs
and AQCDs.
September 2019
IS-26
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
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.
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
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 3.1.6.1.2).
September 2019
IS-27
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
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 2.5V 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.
Epidemiologic studies have included analyses to inform our understanding of the lag structure
(Section 3.1.10.3) for associations between short-term exposure to ozone and respiratory effects. The
largest evidence base for evaluating the lag structure of associations comes from studies of ozone
exposure and hospital admissions or ED visits for asthma. The strongest single-day associations were
generally observed with ozone concentrations on the same day as the outcome, but positive associations
were present across a range of lags, extending as far as 6 days prior to the health outcome of interest. This
range indicates that ozone may elicit both immediate and prolonged respiratory effects.
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
model based on these [and other controlled human exposure data; McDonnell et al. (2013); Figure IS-1;
Section 3.1.4.1.11.
September 2019
IS-28
DRAFT: Do Not Cite or Quote
-------
~
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 (2006) found statistically
significant responses to square-wave chamber exposures at 60 ppb based on the analysis of Brown et al. (2008) and Kim et al.
(2011). 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 for illustrative purposes. McDonnell et al. (2013) illustrates the predicted FE\A 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. (2011) study, but only published in Figure 5 of McDonnell
et al. (2012).
Source: Adapted from Figure 6-1 of 2013 Ozone ISA (U.S. EPA. 2013b). Studies appearing in the figure legend are: Adams (2006).
Adams (2003), Adams (2002), Folinsbee et al. (1988), Horstman et al. (1990), Kim et al. (2011). McDonnell et al. (2013), McDonnell
et al. (1991). and Scheleale et al. (2009).
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
1 Epidemiologic studies examining the shape of the relationship between ambient air
2 concentrations and the studied health outcome and/or the presence of a threshold in this relationship have
3 been inconsistent (Section 3.1.10.4). While most studies assume a no-threshold, log-linear C-R shape, a
4 limited number of studies have used more flexible models to test this assumption. Results from some of
5 these studies indicate approximately linear associations between ozone concentrations and hospital
6 admissions for asthma, while others indicate the presence of a threshold ranging from 20 to 40 ppb 8-hour
7 max ozone concentrations.
8 Most epidemiologic studies that examine the relationship between short-term concentrations of
9 ozone in ambient air and health effects rely primarily on a 1-hour max, 8-hour max, or 24-hour avg
Adams (2006)
Adams (2003)
Adams (2002)
Horstman et al. (1990)
Kim et al. (2011)*
McDonnell et al. (1991)
Schelegle et al. (2009)
McDonnell et al. (2013)
September 2019
IS-29
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
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 IXewis 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. 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 first 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.
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 70 ppb and greater ozone
concentrations following 6.6-hour exposures 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
September 2019
IS-30
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
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 a 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 a causal relationship" between
long-term exposure to ambient ozone and respiratory effects. Results from a limited number of
epidemiologic studies examining potential copollutant confounding suggested that the reported
associations were robust to adjustment for other pollutants, including PM2 5. 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.IV 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
September 2019
IS-31
DRAFT: Do Not Cite or Quote
-------
1 bronchitic symptoms in children with asthma. Notably, some uncertainty remains about the validity of the
2 results from studies examining long-term ozone exposure and hospital admissions and ED visits for
3 asthma, because most of these studies do not adjust for short-term ozone concentrations, despite the
4 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 2019 ISA3
Respiratory effects
Epidemiologic evidence, combined with
toxicological studies in rodents and nonhuman
primates, provided biologically plausible
evidence that there is likely to be a 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.
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
admissions
Epidemiologic studies provided evidence that
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.
September 2019
IS-32
DRAFT: Do Not Cite or Quote
-------
Table IS-5 (Continued): Summary of evidence from epidemiologic and animal
toxicological studies on the respiratory effects
associated on long-term ozone exposure.
Conclusions from 2013 Ozone ISA
Conclusions from 2019 ISA3
Pulmonary
structure and
function
Evidence for pulmonary function effects was
inconsistent, with some epidemiologic studies
observing positive associations (mean annual
8-h daily max ozone concentrations less than
65 ppb). Results from toxicological studies
indicated 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
inflammation,
injury and
oxidative stress
Several epidemiologic studies (mean 8-h max
ozone concentrations less than 69 ppb) and
animal toxicological studies (as low as 500 ppb)
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.
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. Recent animal toxicological studies
continue to provide evidence for
ozone-induced airway eosinophilia in infant
monkeys (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 f"13* 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.
September 2019
IS-33
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Table IS-5 (Continued): Summary of evidence from epidemiologic and animal
toxicological studies on the respiratory effects
associated on long-term ozone exposure.
Conclusions from 2013 Ozone ISA
Conclusions from 2019 ISA3
Respiratory
mortality
A single study demonstrated that exposure to
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 2019 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 3.2.4.6.2). 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, 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 flbrotic- 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
September 2019
IS-34
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
(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.
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. 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
September 2019
IS-35
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
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 2.3.1.1).
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 is sufficient to conclude that there is "likely to be a causal relationship" between
long-term ozone exposure and respiratory effects.
IS.4.3.3 Short-Term Exposure and Metabolic Effects
Metabolic syndrome is a term 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). The presence of these risk factors may predispose someone to an
increased risk of type 2 diabetes and cardiovascular disease. Diagnosis of metabolic syndrome is based on
the presence of three of these five risk factors (Alberti et al.. 2009). The metabolic effects reviewed in this
ISA include metabolic syndrome, diabetes, metabolic disease mortality, and indicators of metabolic
syndrome. Indicators of metabolic syndrome include alterations in glucose and insulin homeostasis,
peripheral inflammation, liver function, neuroendocrine signaling, and serum lipids, 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 short-term ozone exposure and metabolic effects has expanded substantially (Table IS-6).
Results from animal toxicological studies of metabolic effects demonstrate that short-term ozone
exposure impairs glucose and insulin homeostasis (e.g., glucose intolerance, hyperglycemia, dyslipidemia
of triglycerides, glucagon concentration, altered blood pressure, impaired (3-cell function, increased
hepatic gluconeogenesis, and neuroendocrine activation contributing to altered metabolic function) after
inhalation exposure to 0.25 to 1 ppm ozone. Controlled human exposure to ozone in intermittently
exercising subjects, for 2 hours at an exposure of 0.3 ppm ozone or fresh air with 15 minute on/off
exercise in a controlled chamber, confirms activation of the neuroendocrine stress response, and shows
the formation of ketone bodies, abiomarker of diabetes (Section 5.1.5). Previous epidemiologic studies
provide inconsistent evidence for elevated HbAlc (a biomarker of diabetes and an indicator of the degree
of glycemic control in diabetics), increased triglycerides, altered serum cholesterol, increased HOMA-IR,
and fasting glucose level instability associated with short-term ozone concentrations ranging from
19.4-64.4 ppb (mean 24-hour avg across study locations; Section 5.1.3).
September 2019
IS-36
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Recent studies of short-term ozone exposure and metabolic effects compared 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 an average ozone concentration of 64.4 ppb. In addition, an animal
toxicological study demonstrated greater metabolic effects (i.e., increased triglycerides and serum insulin)
in aged animals.
The strongest evidence for metabolic effects of short-term ozone exposure is provided by animal
toxicological studies that show impaired glucose tolerance, increased triglycerides, fasting
hyperglycemia, and increased hepatic gluconeogenesis in various strains of rodents in multiple
laboratories (Section 5.1.3.3). Biological plausibility for an effect of short-term ozone exposure on
metabolic effects is indicated by results from controlled human exposure studies and animal toxicological
studies showing that ozone activates sensory nerves and triggers the central neuroendocrine stress
response, which includes increased corticosterone, Cortisol, or epinephrine production. Additionally, a
recent controlled human exposure study reported that short-term ozone exposure increases ketone body
formation, a biomarker of diabetes. Ketone body formation begins when ozone acts as a sensory and
pulmonary irritant and activates sensory nerves in the respiratory tract that induce downstream effects on
the autonomic nervous system. Evidence demonstrates that the hypothalamus, pituitary, and adrenals are
activated by ozone exposure, and removal of the adrenal pathway (i.e., adrenalectomy or
pharmacologically) can block the induction of metabolic syndrome in rodents exposed to ozone. In
combination with limited epidemiologic and controlled human exposure evidence, the expanding animal
toxicological studies show robust evidence of short-term ozone exposure contributing to activation of
neuroendocrine pathways that lead to impairment of glucose and insulin homeostasis, decreased
glucagon, impaired pancreatic p-cell function, and dyslipidemia. Overall, the collective evidence is
sufficient to conclude that there is "likely to be a causal relationship" between short-term ozone
exposure and metabolic effects.
September 2019
IS-37
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
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 2019 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.
Animal toxicological studies demonstrate impaired glucose intolerance, hyperglycemia,
dyslipidemia of triglycerides, altered glucagon concentrations, altered blood pressure,
impaired (3-cell function, increased hepatic gluconeogenesis, and neuroendocrine
activations, all of which contribute to altered metabolic function after inhalation exposure
to ozone at concentrations from 0.25 to 1 ppm (6 h a day for 2 days). A controlled human
exposure study also shows activation of the neuroendocrine system (2 h at 0.3 ppm
ozone or fresh air exposure with 15 min on/off exercise in a controlled chamber).
A controlled human exposure study shows increased ketone body formation (2 h at
0.3 ppm ozone or fresh air exposure with 15 min on/off exercise in a controlled chamber),
a biomarker of diabetes. A limited number of epidemiologic studies provide some
evidence for some biomarkers of diabetes and other precursors to diabetes (increased
triglycerides, altered serum cholesterol, increased HOMA-IR, fasting glucose level
instability) in an exposure range from 19.4-64.4 ppb mean 24-h avg across study
locations although evidence is inconsistent across studies. An animal toxicological study
shows increased HOMA-IR after ozone exposure (0.8 ppm ozone).
Altered metabolic
function
Diabetes biomarkers
and precursors
IS.4.3.4 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 syndrome has expanded substantially
(Table IS-7). 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. (mean daily ozone
concentration across study locations 38.2 ppb) and Canada (mean annual average ozone concentration
across study locations 39.6 ppb). 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 in 33 communities in China demonstrate increases in metabolic
syndrome associated with a mean increased ozone concentration of 25.1 ppb. Additionally, in prospective
cohort studies in the U.S. and Europe, increased incidence of type 2 diabetes is observed in association
with long-term ozone exposure (mean annual average ozone concentration across study locations
37.5-49.4 ppb).
September 2019
IS-38
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Recent studies examined the potential for copollutant confounding by evaluating copollutant
models that included PM2 5, PM10, and NO2 (Section 5.2.9V The limited number of recent studies provide
some evidence that the metabolic effects associated with long-term ozone exposure are independent of
coexposure to correlated copollutants.
Epidemiologic studies evaluating long-term ozone exposure and metabolic effects did not show
stronger associations in older adults compared with other age groups. For example, a longitudinal cohort
study of diabetes incidence reported increased risk estimates for those under 50 years old, but not for
subjects aged 50 to 60 years, or those over 60 years (mean exposure 49.4 ppb). Similarly, a cohort study
of black women reported increased hazard ratios for all age groups evaluated, with the greatest risk
observed for women under 40 years in age, intermediate risk for women aged 40-54, and the lowest risk
for women over 55 years old with a mean exposure of 37.5 ppb of ozone. Conversely, a recent study that
examined the effect of age on health outcomes in rodents showed that senescent or aged animals were
more sensitive to ozone-dependent serum insulin changes. In the same study, young adult rodents
exposed to ozone did not have significant changes in serum insulin with ozone exposure.
Animal toxicological studies address some of the uncertainty in the epidemiologic evidence
related to the independent effect of ozone exposure by providing evidence of direct effects on metabolic
function. The animal toxicological studies showed evidence that long-term ozone exposure resulted in
impaired insulin signaling, glucose intolerance, hyperglycemia, and insulin resistance (Section 5.2.3.1). In
addition, these pathophysiological changes were often accompanied by increased inflammatory markers
in peripheral tissues and the activation of the neuroendocrine stress response (Section 7.2.1.5V
Importantly, short-term ozone exposure has been shown to contribute to the development of metabolic
syndrome in animals, which is coherent with the evidence that long-term ozone exposure leads to
development or worsening of metabolic syndrome or its risk factors. Overall, the collective evidence is
sufficient to conclude that there is "likely to be a causal relationship" between long-term ozone
exposure and metabolic effects.
September 2019
IS-39
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Table IS-7 Summary of evidence from epidemiologic and animal toxicological
studies on the metabolic effects associated on long-term ozone
exposure.
Results and Conclusions from 2019 ISA
Metabolic effects Recent evidence from controlled human exposure, epidemiologic, and animal toxicological
studies support that there is likely to be a causal relationship between long-term
ozone exposure and metabolic effects.
Epidemiologic and experimental animal studies report effects on glucose homeostasis and
serum lipids, as well as other indicators of metabolic function (e.g., peripheral
inflammation and neuroendocrine activation). Animal toxicological studies provide
evidence that long-term ozone exposure results in impaired insulin signaling, and induced
glucose intolerance, hyperglycemia, and insulin resistance (0.25 to 1 ppm ozone; 6 h a
day, 2 days a week for 13 weeks).
Epidemiologic evidence for increased incidence of type 2 diabetes is associated with
long-term ozone concentrations of 37.5-49.4 ppb (mean annual average ozone
concentration across study locations) in prospective cohort studies in the U.S. and
Europe.
Diabetes mortality Epidemiologic studies report positive associations between long-term exposure to ozone
and diabetes-related mortality in well-established cohorts in the U.S.(38.2 ppb; mean
annual average ozone concentration across study locations) and Canada (39.6 ppb, mean
annual average ozone concentration across study locations).
Altered metabolic
function
Metabolic syndrome
and diabetes
IS.4.3.5 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), while the
controlled human exposure studies provided some evidence, though limited coherence with the animal
studies. The epidemiologic evidence, while reporting associations between short-term ozone exposure and
cardiovascular mortality, did not observe associations between short-term ozone exposure and
cardiovascular morbidity. This lack of coherence between the results investigating associations with
cardiovascular morbidity and 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
studies do not indicate an effect of ozone on cardiac function, ST segment, endothelial dysfunction, or
September 2019
IS-40
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
HR, while other recent studies provide little evidence that 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-8V
Table IS-8 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 2019 ISA
Cardiovascular effects Evidence from animal toxicological
Recent animal toxicological studies continue to
provide evidence for impaired heart function
and endothelial dysfunction, with limited
evidence 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
and inconsistent evidence 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.
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 a causal
relationship between long-term
exposure to ozone and respiratory
effects.
September 2019
IS-41
DRAFT: Do Not Cite or Quote
-------
Table IS-8 (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 2019 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
0.4-1.0 ppm 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
NO-induced vascular reactivity.
Epidemiologic studies reported generally
positive associations for hospital
admissions or ED visits due to arrythmia
or dysrhythmia.
Recent animal toxicological studies
demonstrate limited 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.
September 2019
IS-42
DRAFT: Do Not Cite or Quote
-------
Table IS-8 (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 2019 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 the 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 (0.25-1.0 ppm 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.
September 2019
IS-43
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Table IS-8 (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 2019 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
mortality
Multicity epidemiologic studies observed
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 2019 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.4.3.6 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., maximum daily
1-hour max, max daily 8-hour avg, 8-hour avg, and 24-hour avg) (Table IS-9). The limited assessment of
cause-specific mortality (e.g., respiratory mortality r Section 3.1.91. cardiovascular mortality
[Section 4.1.141) in 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
September 2019
IS-44
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
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 PMio or NO2. Importantly, the issues surrounding the assessment of
potential copollutant confounding that complicate interpretation (as detailed in the 2013 Ozone ISA)
persist, specifically within studies that relied on different PM sampling schedules, such as every 3rd- and
6th-day PM sampling (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 and animal toxicological evidence of cardiovascular effects, a large number of
recent controlled human exposure studies are not consistent with the evidence presented in the 2013
Ozone ISA from controlled human exposure studies showing cardiovascular effects. 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.
September 2019
IS-45
DRAFT: Do Not Cite or Quote
-------
Table IS-9 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 2019 ISAa
Mortality
Consistent, positive associations were
reported across multicity and
multicontinent studies in combination with
strong evidence from studies of
respiratory morbidity. There was limited
evidence from 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
absent. 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,
indicating a lack of coherence between
experimental and epidemiologic studies and
providing limited evidence of a biologically
plausible pathway to cardiovascular mortality
or to other causes of mortality.
Conclusions from the 2019 ISA include evidence from recent studies integrated with evidence included in previous Ozone ISAs
and AQCDs.
September 2019
IS-46
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
IS.4.3.7 Other Health Endpoints
The evidence for the other health endpoints not discussed in previous sections, including
long-term ozone exposure and cardiovascular 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 6. and Appendix 7.
IS.4.3.7.1 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.
IS.4.3.7.2 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" 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
September 2019
IS-47
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
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.
IS.4.3.7.3 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" 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 indicate 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).
September 2019
IS-48
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
IS.4.3.7.4 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 indicate 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
supporting the biological plausibility of the associations with autism or ASD in epidemiologic studies is
limited.
IS.4.3.7.5 Long-Term Ozone Exposure and Cancer
The evidence describing the relationship between exposure to ozone and cancer remains
inadequate to determine whether a causal relationship exists. 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.
September 2019
IS-49
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
IS.4.3.7.6 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.4.4 At-Risk Populations
Interindividual variation in 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, as well as 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
September 2019
IS-50
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
concentrations (e.g., more time spent in areas with higher ambient concentrations). Some factors may lead
to a reduction in risk and are recognized during the evaluation process, but for identifying those
populations or lifestages at greater risk to inform decisions on the NAAQS, the focus in this ISA is on
characterizing those factors that may increase 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, described in detail
in the Preamble to the ISAs (U.S. EPA. 2015) and illustrated in Table IS-10. Briefly, in contrast to the
overall evaluation of ozone exposures and health effects presented in Appendix 3-Appendix 7. this
section specifically aims to summarize the consideration of evidence for populations and lifestages
potentially at increased risk of an ozone-related health effect. 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. In addition, the populations and lifestages explicitly 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.
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
September 2019
IS-51
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
observed in studies that formally examine effect measure modification. Similar to the characterization of
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 IS-10.
Table IS-10 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.
Suggestive The collective evidence suggests that a factor results in a population or lifestage being at
evidence 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.
Inadequate The collective evidence is inadequate to determine whether a factor results in a population or
evidence 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.
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
evidence of coherence in 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
September 2019
IS-52
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
ozone. There is relatively little recent evidence for older adults, individuals with reduced intake of certain
nutrients, and outdoor workers, and the evidence presented in the 2013 Ozone ISA is adequate to classify
them as at-risk populations.
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-11; Section IS.4.4.3.1).
Generally, studies comparing age groups also reported higher magnitude associations for
respiratory hospital admissions and ED visits among children (Section IS.4.4.4.1) than for adults. In
addition, recent evidence from studies of nonhuman primates demonstrate ozone-induced respiratory
effects and support 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. Also, children have
higher exposure and dose due to increased time spent outdoors and ventilation rate, and 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).
Table IS-11 Summary of evidence for populations at increased risk to the health
effects of ozone.
Conclusions from 2013 Ozone ISA
Conclusions from 2019 ISA
Adequate evidence
Pre-existing
asthma
Collective evidence from controlled human
exposure studies is supported by
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.
September 2019
IS-53
DRAFT: Do Not Cite or Quote
-------
Table IS-11 (Continued): Summary of evidence for populations at increased risk
to the health effects of ozone.
Conclusions from 2013 Ozone ISA
Conclusions from 2019 ISA
Children
Controlled human exposure and
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 indicate 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,
which demonstrated increased exposure,
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.
No recent information has been evaluated that
would inform or change prior conclusions.
Genetic factors Multiple genetic variants have been
observed in epidemiologic and controlled
human exposure studies to affect the risk of
ozone-related respiratory outcomes and
support is provided by animal toxicological
studies of genetic factors.
No recent information has been evaluated that
would inform or change prior conclusions.
Diet
Individuals with reduced intake of vitamins E
and C are at risk for ozone-related health
effects based on substantial, consistent
evidence both within and among disciplines.
No recent information has been evaluated that
would inform or change prior conclusions.
Suggestive evidence
Sex
Evidence for increased risk for ozone-related No recent information has been evaluated that
health effects present for females in some would inform or change prior conclusions.
studies and males in other studies; some
indication that females are increased risk of
ozone-related respiratory hospital
admissions and ED visits.
Pre-existing
obesity
Multiple epidemiologic, controlled human
exposure, and 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.
September 2019
IS-54
DRAFT: Do Not Cite or Quote
-------
Table IS-11 (Continued): Summary of evidence for populations at increased risk
to the health effects of ozone.
Conclusions from 2013 Ozone ISA
Conclusions from 2019 ISA
SES
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 limited number of studies indicate 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 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 ortoxicological studies to
determine whether pre-existing diabetes
modifies ozone effects on health.
A limited number of recent studies provides some
evidence that individuals with pre-existing
metabolic disease may be at greater risk of
mortality associated with long-term ozone
exposure.
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.
September 2019
IS-55
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
IS.4.4.3 Pre-existing Disease
Individuals with some pre-existing diseases may be considered 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-12 presents the prevalence of these diseases according to the Centers for Disease Control and
Prevention's (CDC's) National Center for Health Statistics (Blackwell 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
determining whether identifying populations that may be at increased risk for ozone-related health effects.
September 2019
IS-56
DRAFT: Do Not Cite or Quote
-------
Table IS-12 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. (2014): National Center for Health Statistics: Data from Tables 1 -4, 7, 8, 28, and 29 of the Centers for
Disease Control and Prevention report.
September 2019 IS-57 DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
IS.4.4.3.1 Pre-existing Asthma
Asthma is the leading chronic illness affecting children. Approximately 8.0% of adults and 9.3%
of children (age <18 years) in the U.S. currently have asthma (Blackwell 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 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 15 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.
September 2019
IS-58
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
IS.4.4.3.2 Pre-existing Chronic Obstructive Pulmonary Disease
Chronic obstructive pulmonary disease (COPD) comprises chronic bronchitis and emphysema
and affects approximately 6.8 million adults in the U.S. (Table IS-12). In the U.S., over 4% of adults
report having chronic bronchitis and almost 2% report having emphysema (Pleis et al.. 2009). 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.. 2017). 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 to 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
September 2019
IS-59
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
limited evidence within epidemiologic studies and some coherence from controlled human exposure and
animal toxicological studies.
IS.4.4.3.4 Pre-existing 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 central obesity), and increased fasting blood glucose (Albcrti ct 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 to
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.
IS.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-12). 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
or without 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
September 2019
IS-60
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
cardiovascular disease exacerbates the respiratory effects of ozone exposure. Injury, inflammation,
oxidative stress, 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, inflammation, 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-12; 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, although the effects were not consistently greater
September 2019
IS-61
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
in young animals than adult animals (U.S. EPA. 2013b). Compared with other age groups, there is
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.
IS.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.
IS.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 lung function decrements declining with age, but still
September 2019
IS-62
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
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 small lung
function decrements occurring in adults of age group ranging up to 70 years. 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-13.
Table IS-13 Summary of evidence for welfare effects of ozone.
Endpoint
Conclusions from 2013 Ozone ISA
Conclusions from Current Draft ISA
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.
September 2019
IS-63
DRAFT: Do Not Cite or Quote
-------
Table IS-13 (Continued): Summary of evidence for welfare effects of ozone.
Endpoint
Conclusions from 2013 Ozone ISA
Conclusions from Current Draft ISA
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 average 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 a 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 impacts 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 nonsoy 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.
September 2019
IS-64
DRAFT: Do Not Cite or Quote
-------
Table IS-13 (Continued): Summary of evidence for welfare effects of ozone.
Endpoint
Conclusions from 2013 Ozone ISA
Conclusions from Current Draft ISA
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 a 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 Causality not assessed
plant-insect ^ few experimental and modeling studies
signaling reported altered chemical signaling in
Section 8.7 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 a 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 ofthe
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 a causal relationship
Studies add to the strong and consistent
evidence in the 2006 AQCD that ozone
decreases plant photosynthesis. Most
assessments ofthe effects of ozone on
terrestrial C are from model simulations.
Likely to be a 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.
September 2019
IS-65
DRAFT: Do Not Cite or Quote
-------
Table IS-13 (Continued): Summary of evidence for welfare effects of ozone.
Endpoint
Conclusions from 2013 Ozone ISA
Conclusions from Current Draft ISA
Alteration of
belowground
biogeochemical
cycles
Section 8.9
Causal relationship
It has been documented since the 2006 Ozone
AQCD that while belowground roots and soil
organisms are not exposed directly to ozone,
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 a 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 a 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 a 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.
September 2019
IS-66
DRAFT: Do Not Cite or Quote
-------
16
17
18
19
20
21
22
23
24
25
26
27
28
Table IS-13 (Continued): Summary of evidence for welfare effects of ozone.
Endpoint Conclusions from 2013 Ozone ISA Conclusions from Current Draft ISA
Likely to be a 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.
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 community3; 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
1 A functional unit consisting of living organisms (biota), their nonliving environment and the interactions within
and between them (TeametaL 2014).
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).
Temperature,
precipitation and
related climate
variables
Section 9.3
Likely to be a 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.
September 2019
1S-67
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
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.
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, community composition, 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
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 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).
September 2019
IS-68
DRAFT: Do Not Cite or Quote
-------
..........
*r ¦
03 exposure
* 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
1 -Altered crop quality
1
Belowground processes
•Altered litter production and decomposition
•Altered soil carbon and nutrient cycling
•Altered soil fauna and microbial communities
D
CD
—1
CD
U
O
< .1
CO
CD
=3
CO
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.
IS.5.1.1 Visible Foliar Injury
1 In the 2013 Ozone ISA the evidence was sufficient to conclude a causal relationship between
2 ozone exposure and visible foliar injury on sensitive vegetation across the U.S. Visible foliar injury
3 (Figure IS-4) resulting from exposure to ozone has been well characterized and documented in over six
4 decades of research on many tree, shrub, herbaceous, and crop species using both long-term field studies
5 and laboratory approaches (U.S. EPA. 2013b. 2006a. 1996b. 1986. 1978; NAPCA. 1970; Richards et al..
6 1958). Recent experimental evidence continues to show a consistent association between visible injury
7 and ozone exposure (Section 8.2). In a recent global-scale synthesis documenting foliar injury from ozone
September 2019 IS-69 DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
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 van 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.
Note: Tulip poplar (Liriodendron tulipifera) ori 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 phototoxicity 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-induced 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
September 2019
IS-70
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
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
indicates 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
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 documenting reductions in biomass due to ozone exposure. At least 69 plant species that
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 [Section 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
September 2019
IS-71
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
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
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. (Dietze 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 AQCDs 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.
September 2019
IS-72
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
There is a considerable amount of new research on major U.S. crops, especially soybean, wheat,
and other nonsoy legumes at concentrations of ozone occurring in the environment (Section 8.5V 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.. 2016V 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.. 2012V 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.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. 2013bV 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),
Hemioptera (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 8.6V 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
September 2019
IS-73
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
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 S-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
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 8.8.1). 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
September 2019
IS-74
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
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
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.
IS.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.
September 2019
IS-75
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
IS.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 Onograceae) 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
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.
IS.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
September 2019
IS-76
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
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; I Section S. 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 (Kostiaincn 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.
IS.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
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. 1996b). For
example, studies from the Aspen FACE experiment have shown that ozone caused reduction in total
biomass in quaking aspen (Populus tremuloides), paper birch (Betulapapyrifera), 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.
September 2019
IS-77
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
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, four 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
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 (Yolk etal.. 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,
September 2019
IS-78
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
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.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. Weighting of hourly
concentrations and the diurnal and seasonal time window of exposure 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
September 2019
IS-79
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
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 8.13.2V 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.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,
September 2019
IS-80
DRAFT: Do Not Cite or Quote
-------
1 transport, and persistence of ozone in the troposphere (Appendix 1). as well as increased vulnerability of
2 plants and ecosystems. The degree to which climate and weather alter the effects of ozone is context and
3 species specific because damage to terrestrial ecosystems caused by ozone is largely a function of plant
4 uptake. Factors that modify the effects of ozone in ecosystems, including carbon dioxide, weather, and
5 climate are discussed in Section 8.12.
^3
Climate Effects
on Human Health r
and Ecosystems J
Changes in Tropospheric
O, Abundance
« ™ <
Radiative Forcing
Due to 03 Change
(W/m2)
Climate Response
(°C)
Precursor Emissions of
CO, VOCs, ch4, nox
(Tg/y)
Source: U.S. EPA (2013b).
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.
6 Characterization of ozone impacts on radiative forcing (Section 9.2) builds on the findings in the
7 2013 Ozone ISA and draws heavily on the IPCC Assessment Reports. In the 2013 Ozone ISA, the
8 evidence was sufficient to conclude a causal relationship between tropospheric ozone and radiative
9 forcing (U.S. EPA. 2013b). The 2013 Ozone ISA reported a radiative forcing (RF) of 0.35 W/m2 from the
fO change in tropospheric ozone abundance from preindustrial times to the present (1750 to 2005) based on
September 20 f 9
IS-81
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
multimodel studies (Forstcr et al.. 2007). The most recent IPCC assessment, AR5, reports tropospheric
ozone RF as 0.40 (0.20 to 0.60) W/m2 (Mvhrc 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 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) between tropospheric ozone and climate change (U.S. EPA.
2013b). New studies reviewed in Section 9.3 are consistent with previous estimates and the effect of
tropospheric ozone on global surface temperature continues to be estimated at roughly 0.1-0.3°C since
preindustrial times (Xie et al.. 2016; Myhre 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.
IS.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
September 2019
IS-82
DRAFT: Do Not Cite or Quote
-------
1 evidence linking ozone with health and welfare effects with or despite minimal remaining uncertainties,
2 the U.S. EPA makes a determination of a causal or likely to be causal relationship.
IS.6.1 Health Effects Evidence: Key Findings
3 A large body of scientific evidence spanning many decades clearly demonstrates there are health
4 effects related to both short- and long-term ozone exposure (Figure IS-6). The strongest evidence supports
5 a relationship between ozone exposure and respiratory health effects. The collective body of evidence for
6 each health outcome category evaluated in this ISA is systematically considered and assessed, including
7 the inherent strengths, limitations, and uncertainties in the overall body of evidence, resulting in the
8 causality determinations detailed in Table IS-1. Through identification of the strengths and limitations in
9 the evidence, this ISA may help in the prioritization of research efforts to support future ozone NAAQS
10 reviews.
September 2019
IS-83
DRAFT: Do Not Cite or Quote
-------
Causality Determinations for Health Effects of Ozone
ISA
Current Ozone Draft ISA
Health Outcome
Respiratory
Short-term 1 1
exposure | |
Long-term
exposure
Metabolic
Short-term
exposure
+
Long-term
exposure
+
Cardiovascular
Short-term
exposure
*
Long-term
exposure
Nervous System
Short-term
exposure
Long-term
exposure
Reproductive
Male/Female
Reproduction
and Fertility
Long-term
exposure
*
Pregnancy and
Birth Outcomes
*
Cancer
Long-term
exposure
Mortality
Short-term
exposure
*
Long-term
exposure
Causal ¦ Likely causal 1 Suggestive Inadequate 1 1
+ new causality determination
* change in causality determination from 2013 Ozone ISA
Figure IS-6 Causality determinations for health effects of short- and
long-term exposure to ozone.
1 An inherent strength of the evidence integration in this ISA is the extensive amount (in both
2 breadth and depth) of available evidence resulting from decades of scientific research that describes the
3 relationship between both short- and long-term ozone exposure and health effects. The breadth of the
4 enormous database is illustrated by the different scientific disciplines that provide evidence
5 (e.g., controlled human exposure, epidemiologic, animal toxicological studies), the range of health
6 outcomes examined (e.g., respiratory, cardiovascular, metabolic, reproductive, and nervous system
7 effects, as well as cancer and mortality), and the large number of studies within several of these outcome
September 2019
IS-84
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
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, modern systematic review methodologies are applied to identify and characterize this
expansive evidence base (see Appendix 10 for details). The evidence is effectively integrated from
(1) different scientific disciplines, (2) a variety of study designs within the same scientific discipline, and
(3) a span of different health endpoints within a health effect category. Finally, a formal framework is
systematically applied 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.
Older and more recent studies also provide evidence for biologically plausible pathways that may
underlie respiratory effects related to long-term ozone exposure, and metabolic effects related to both
short- and long-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 admission (Figure 4-1). However, the
evidence mainly comes 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
September 2019
IS-85
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
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
difference could explain some of the inconsistencies across studies from 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 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 rarely studied). Therefore, individuals that may be at greater risk are not included in controlled
September 2019
IS-86
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
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
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.
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 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
September 2019
IS-87
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
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 2.6).
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 better
measured pollutant 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 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 discipline 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
September 2019
IS-88
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
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
(U.S. EPA. 2013b. 2006a. 1996b. 1986. 1978: NAPCA. 1970: Richards ctal.. 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 a dozen 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 takes into account the important role of soil moisture in foliar injury. As documented in the 2013
Ozone ISA (Table IS-13) 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
September 2019
IS-89
DRAFT: Do Not Cite or Quote
-------
1 category of growth but the current body of evidence for this endpoint warrants a separate causality
2 category.
Causality Determinations for Ecological Effects of Ozone
Belowground
Biogeochemical Cycles
1
>-
Water Cycling
b
5
jj
u
Carbon Sequestration
M
c
o
a
Productivity
Of
en
"5
o
Community
Biodiversity
Terrestrial Community Composition*
tuo
o
u
UJ
Species Interactions
Plant-Insect Signaling +
o
J8
Survival
Trees+
tv
tf\
c
5
re
-
;g
Growth
Plants Herbivores +
3
EL
S
£
>
'•V
c
Reproduction
Plants+ Herbivores +
Yield
Agricultural Crops
Individual
Visible Foliar Injury
Causal Likely Causal |
new determination (+} or change in causality determination (*) from 2013 Ozone ISA
Figure IS-7 Causality determinations for ozone across biological scales of
organization and taxonomic groups.
3 While the effect of ozone on vegetation is well established in general, there are some knowledge
4 gaps regarding precisely which species are sensitive and what exposures elicit adverse responses for many
5 species. Currently there are over 40,000 plants and lichens occurring in the U.S. as documented by the
6 USDA PLANTS database (TJSDA. 2015). It not feasible to know what the effects are on all U.S. species
7 and what the ecological consequences of the differential sensitivities are of these species. However, there
September 2019
IS-90
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
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 studies can show causal
links of ecological endpoints to 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.
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 scaling ozone effects to
the ecosystem level remains a challenge, there is a large body of knowledge of how ecosystems work
through ecological observations and models that simulate processes at multiple scales. The models scale
up and 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).
September 2019
IS-91
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
More research on ecosystem-level responses will strengthen understanding of scaling across different
levels of biological organization.
In general, the most promising approaches to scaling 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 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 tropospheric ozone RF as 0.40 (0.20
to 0.60) W/m2 (Mvhre etal.. 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 indicate a change to either causality determination
(Figure IS-S).
September 2019
IS-92
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
Causality Determinations for Tropospheric Ozone arid
Climate Change
Radiative Forcing
Temperature, precipitation and
related climate variables
Causal
~
Likely Causal
Figure IS-8 Causality determinations for tropospheric ozone and climate
change.
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, 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). 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 affect
other climate metrics, such as precipitation and atmospheric circulation patterns, and new evidence has
continued to support and further quantify this understanding. 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.
September 2019
IS-93
DRAFT: Do Not Cite or Quote
-------
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.
https://doi.org/10.1080/089583703045Q5.
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: Cleemaa 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.
Allea 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/nature 11097.
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.
Bergmana 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: Nelsoa 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. 112.205591.
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.llll/i.1365-3040.2010.02165.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.
Bovd. J: Banzhaf. S. (2007). What are ecosystem services? The need for standardized environmental accounting
units. Ecol Econ 63: 616-626.
Browa JS: Batesoa 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.
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.
September 2019
IS-94
PRAFT: Po Not Cite or Quote
-------
Costanza. R: De Groot. R: Braat. L: Kubiszewski. I; Fioramonti. L: Suttoa 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: Kleia 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.
Pi. 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/10.1111/i. 1365-2486.2011.02477.x.
Folinsbee. LJ: Mcdonnell. WF: Horstmaa PH. (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.
Forster. P: Ramaswamv. V: Artaxo. P: Berntsea T: Betts. R: Fahev. PW: Havwood. J: Leaa J: Lowe. PC: Mvhre.
G: Nganga. J: Prina R: Raga. G: Schultz. M: VanPorland. 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.
Gustafsoa 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. CP: Qgdea CL. (2017). Prevalence of Obesity Among Adults and Youth: United
States, 20152016. NCHS Pata Brief 288.
Hill. AB. (1965). The environment and disease: Association or causation? Proc R Soc Med 58: 295-300.
Horstman. PH: 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/10.1164/airccm/142.5.1158.
Hovert. PL: Xu. J. (2012). Peaths: 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/wgl/.
Jaffe. PA: Cooper. OR: Fiore. AM: Hendersoa BH: Tonnesea 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. 6. http://dx.doi.org/10.1525/elementa.309.
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: Pevlin. RB: Pedea
PB: Piaz-Sanchez. P. (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.l 164/rccm.201011-
1813QC.
Kostiainen. K: Saranpaa. P: Lundavist. 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.
September 2019
IS-95
PRAFT: Po Not Cite or Quote
-------
Lefohn. AS: Runeckles. VC. (1987). Establishing standards to protect vegetation - ozone exposure/dose
considerations. Atmos Environ 21: 561-568. http://dx.doi.org/10.1016/0004-6981(87190038-2.
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.llll/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.ll.070.
Macintosh. CR: Allaa RP: Baker. LH: Bellouia 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.
Mautz. WJ: Bufalino. C. (1989). Breathing pattern and metabolic rate responses of rats exposed to ozone. Respir
Physiol Neurobiol 76: 69-77.
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: Devlia RB: O'Neil. JJ: Horstman. PH.
(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. (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: Sua 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/10.1111/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.
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/10.1007/978-l-4612-1436-6 19.
Mills. G: Harmens. H: Wagg. S: Sharps. K: Haves. F: Fowler. D: Suttoa 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/.
Moraa 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: Breoa 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/wgl/.
September 2019
IS-96
DRAFT: Do Not Cite or Quote
-------
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: Johnsoa 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/10.1787/5i lv 1 m9d 1 g3 2-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: Btiker. 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/10.1111/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.
Richards. BL: Middletoa 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: Allea 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: vanDingenen. R: Amann. M: Klimont. Z: Anenberg. SC: Muller. N:
Janssens-Maenhout. G: Raes. F: Schwartz. J: Faluvegi. G: Pozzoli. L: Kupiainen. K: Hoglund-Isakssoa L:
Embersoa L: Streets. D: Ramanathaa 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.
Sua 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/ZvPURL.cgi?Dockev=200089CW.txt.
September 2019
IS-97
DRAFT: Do Not Cite or Quote
-------
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.
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.epa.gov/ncea/cfm/recordisplav.cfm?deid=149923.
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.epa.gov/ncea/cfm/recordisplav.cfm?deid=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.cfm?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/naaas/standards/nox/data/2014Q6finalirpprimarvno2.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). (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://ofmpub.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/ (accessed
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.
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.
September 2019
IS-98
DRAFT: Do Not Cite or Quote
-------
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.Ill 1/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.Ill 1/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.
Watkinson. WP: Campen. MJ: Wichers. LB: Nolaa JP: Costa. PL. (2003). Cardiac and thermoregulatory responses
to inhaled pollutants in healthy and compromised rodents: Modulation via interaction with environmental factors.
Environ Res 92: 35-47. http://dx.doi.org/10.1016/S0013-9351(02)00023-3.
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/sQ0376-016-5193-0.
Zak. PR: Pregitzer. KS: Kubiske. ME: Burtoa 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.llll/i.1461-0248.2011.01692.x.
September 2019
IS-99
DRAFT: Do Not Cite or Quote
-------
APPENDIX 1 ATMOSPHERIC SOURCE,
CHEMISTRY, METEOROLOGY,
TRENDS, AND BACKGROUND
OZONE
• The sources of ground-level ozone vary widely, including emissions due to human
activities within the U.S. and internationally, natural and biological processes, and
dynamics within Earth's atmosphere. Uncertainties in the rates of precursor emissions
for the sources of background ozone are typically high due to the difficulties
associated with characterizing and quantifying such variables as biological variability
in vegetation species across U.S. regions, and the effects of meteorological variability
on sources sensitive to temperature, moisture, and atmospheric circulation patterns.
• Understanding concerning the production of ozone during wintertime in the
Intermountain West and the depletion of ozone on coastlines continues to advance.
Local emissions related to oil and gas production combined with strong atmospheric
inversions (cold pools) explain the unusually high concentrations of ground-level
ozone in Western mountain basins during winter. Photochemical mechanisms that
include halogen radical-based heterogeneous reactions in sea salt particles have been
found to account for previously unexplained reduced concentrations in surface ozone
along urban areas on U.S. maritime coastlines.
• Large-, regional-, and local-scale atmospheric circulation patterns influence both
observed U.S. background ozone and the local production of ground-level ozone.
Interannual (e.g., the El Nino-Southern Oscillation [ENSO] cycle) and multidecadal
(e.g., the Pacific Decadal Oscillation [PDO] and Atlantic Multidecadal Oscillation
[AMO]) 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.
• U.S. anthropogenic emissions of ozone precursors have declined over the past two
decades. For example, U.S. NOx emissions decreased by 48% between 2002 and
2014. As precursor concentrations decrease, the U.S. ambient ozone concentration
distribution is compressing (i.e., 95th percentile concentrations are decreasing at the
same time 5th percentile concentrations are increasing), consistent with chemistry
expected after reductions in NOx.
• U.S. background ozone continues to account for a large fraction of ambient ozone
concentrations as a result of stratospheric exchange, international transport, wildfires,
lightning, global methane emissions, and natural biogenic and geogenic precursor
emissions. New results concerning U.S. background ozone are (1) a wider range of
concentration estimates, and poorer agreement between models have been observed
than were reported in the 2013 Ozone ISA, with a range of uncertainty of ~10 ppb for
seasonal average concentrations, (2) U.S. background concentrations are uncorrelated
with local ground-level concentrations above -60 ppb, and (3) an increasing trend of
U.S. background concentration at high elevation western U.S. sites before
approximately 2010 now shows signs of slowing or even reversing, probably due to
decreasing East Asian precursor emissions.
September 2019
1-1
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
1.1 Overview
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. The
primary focus is 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. 2016b). 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. 2006a).
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 up-to-date 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
September 2019
1-2
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
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-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 maximum 1-hour average (MDA1), daily maximum 8-hour average (MDA8), and daily
24-hour average 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 average of the annual 4th-highest MDA8 ozone concentrations.
September 2019
1-3
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
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 day and seasonal time window. The sigmoidal weighting of hourly
ozone concentration is given by Wc = 1/(1 + 4.403c l 2"r). 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, 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 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 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).
September 2019
1-4
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
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. 2006a) and the 2013 Ozone ISA (U.S. EPA. 2013). which also used modeled
estimates of USB. The 2006 Ozone AQCD (U.S. EPA. 2006a) concluded that background ozone
concentrations could not be determined exclusively from ozone measurements at relatively
remote monitoring sites because of long-range transport of ozone originating from U.S.
anthropogenic precursors even at the most remote monitoring locations. Reliance on atmospheric
modeling for USB concentrations estimates continued in the 2013 Ozone ISA (U.S. EPA. 2013).
In earlier assessments, ozone estimates were based on measurements at monitoring sites with low
concentrations that appeared to be isolated from anthropogenic sources (Altshuller and Lefohn.
1996; Trainer et al.. 1993).
1.2.2.2 Apportionment-Based U.S. Background (USB)
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. 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 a
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.
September 2019
1-5
DRAFT: Do Not Cite or Quote
-------
1 • Seasonal mean USB is more relevant for understanding source contributions to long-term
2 concentrations.
3 • As discussed by Jaffe et al. (2018) and in Section 1.8.1. USB MDA8 estimates on specific days
4 are more uncertain than USB seasonal mean estimates, because of considerable daily variation
5 influenced by season, meteorology, and elevation.
1.2.2.4 Other Background Ozone Definitions
6 Other definitions besides USB have been used in previous U.S. EPA science assessments.
7 Although USB is emphasized in this document, research results based on North American background
8 (NAB) and natural background are also included. These terms were also widely used in the 2013 ozone
9 ISA (U.S. EPA. 2013) and in earlier ozone assessments.
10 • NAB has been defined as the ozone concentration that would occur in the U.S. in the absence of
11 anthropogenic emissions in continental North America (U.S. EPA. 2013). NAB has also been
12 referred to as policy-relevant background (PRB) in earlier publications (U.S. EPA. 2007).
13 • Emissions-influenced background (EIB) has been defined as another measure of background
14 ozone estimated from source apportionment modeling approaches while including chemical
15 interactions with anthropogenic emissions (Lefohn et al.. 2012).
16 • Natural background ozone is defined as the ozone concentrations that would occur if all
17 anthropogenic emissions were removed worldwide. Processes that contribute to natural
18 background ozone include ozone transport from the stratosphere and ozone formed from
19 precursor emissions originating from wildfires, lightning, natural methane sources, plants, and
20 other natural VOC and NOx emissions (see Section 1.3).
1.2.2.5 Baseline Ozone
21 Baseline ozone is an alternative metric to USB or NAB that has been defined as the measured
22 ozone concentration at rural or remote sites that have not been influenced by recent, local emissions (Jaffe
23 et al.. 2018). In contrast to USB, baseline ozone is by definition directly measured.
24 • Baseline measurements are typically from monitors in locations that are minimally influenced by
25 local anthropogenic sources, and samples used as baseline measurements are limited to those
26 monitored during meteorological conditions consistent with the relative absence of local
27 contamination.
28 • Baseline ozone can include the ozone produced from U.S. emissions that circle the globe and may
29 also include effects of same-state emissions. An example of the latter would be ozone from U.S.
30 emissions near the West Coast or Gulf Coast that is transported over the Pacific Ocean or Gulf of
31 Mexico, respectively, and then transported back onshore.
32 • In some cases, sources that impact baseline ozone may not similarly impact ozone in populated
33 locations. For instance, baseline ozone measured on a mountaintop may include stratospheric
34 influences that are not representative of contributions in nearby lower elevation locations.
September 2019
1-6
DRAFT: Do Not Cite or Quote
-------
1 • There are several reasons why baseline ozone measurements cannot be used as a proxy to
2 estimate USB ozone levels in urban areas. As previously described, baseline ozone can include
3 contributions from U.S. emissions. Additionally, baseline ozone monitors can be very distant
4 from urban sites, and ozone measured at the baseline site can be destroyed through surface
5 deposition or chemical reactions during transport from the baseline site to a downwind monitor.
6 In addition, atmospheric conditions may not favor transport of baseline ozone from the monitor
7 location to populated areas at lower elevations.
8 • Another reason why baseline ozone measurements cannot be used as a proxy for USB ozone
9 levels in urban areas is that meteorological conditions that favor mixing from the free troposphere
10 to ground level have strong ventilation and are not conducive to photochemical ozone episodes
11 that produce the highest urban ozone concentrations (see Section 1.5. IV Stratospheric intrusion
12 events are an exception (see Section 1.3.2).
13 • While baseline ozone measurements cannot be used directly to estimate USB ozone, baseline
14 ozone data are useful for evaluating the CTMs that are used to provide model estimates of USB
15 ozone.
1.3 Sources of U.S. Ozone and Its Precursors
16 U.S. tropospheric ozone (i.e., ozone that may have harmful health and environmental impacts) is
17 classified in this assessment as either being derived from U.S. anthropogenic sources or background
18 (USB). Anthropogenic ozone within the U.S. is further defined as the product of photochemical reactions
19 of precursors derived from human activities. USB ozone, as defined in Section 1.2.2.1. has a broader,
20 more complex array of sources. These include natural precursor sources as well as precursors transported
21 from across U.S. borders from both nearby and distant locations within the Northern Hemisphere. Ozone
22 derived from the stratosphere and from the reaction of internationally-transported precursors in the upper
23 troposphere can be drawn down into the lower troposphere through atmospheric dynamics (i.e., vertical
24 movement of large air masses between the stratosphere and the troposphere). Figure 1-1 illustrates the
25 complexities associated with attributing measured ground-level ozone to particular sources.
26 The main focus of this section is recent scientific findings concerning the sources of USB ozone.
27 To provide context for this discussion, updated information on U.S. anthropogenic ozone precursor
28 emissions and trends in those emissions is included.
September 2019
1-7
DRAFT: Do Not Cite or Quote
-------
Tropopause Folding
Aircraft Emissions
NOx CO CH,
Photochemistry
STRATOSPHERE
TROPOSPHERE
PAN
Cities
A ~
NOx VOCs
CH4 CO
Lightning
NOx
Convection
Photochemistry
Industry ± . Sj
r.jjyj II
Deposition
i »i'j
FREE
TROPOSPHERE
BOUNDARY
LAYER
9L .
Trans porta bo n
Agriculture and
Animal Husbandry
Forests and Other
Ecosystems
-
Landfill Gas 1,
fJL
Wild and
Prescribed
Fire
Source: Adapted from CCSP (2003).
Power
Generation
IJli ©"©
Fossil Fuel Extraction
Figure 1-1 Major atmospheric processes and precursor sources contributing
to ambient ozone.
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.
September 2019
1-8
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
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. These estimates are taken from the publicly available
versions of the U.S. EPA National Emissions Inventory [2014 NEI, Version 2; U.S. EPA (2018a)! and of
the U.S. Inventory of Greenhouse Gases and Sinks (U.S. EPA. 2016c). Emissions of each precursor are
shown as a function of source type.
The U.S. EPA maintains a 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 1,"source types (U.S. EPA. 2019b). National emissions estimates for these categories are derived
from the NEI and are included in the trends dataset 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 dataset 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 dataset. 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.
September 2019
1-9
DRAFT: Do Not Cite or Quote
-------
A) NOx (14,366 kTon/yr)
;Commercial
Marine
Vessels
Non-Road
Biogenics -
Equipment-
Vegetation and Soil
6%
Diesel
8%
Electric
Generation -
Combustion
Other
17%
Locomotives
5%
Oil & Gas
Production
n 5%
On-Road
Diesel
Heavy Duty
Vehicles
15%
On-Road non-Diesel
Light Duty Vehicles
17%
Residential-Natural
Gas Combustion
2%
Industrial
Boilers, ICEs-
Natural Gas
4%
Non-Road
Equipment
- Gasoline
2%
B) CO (72,353 kTons/yr)
On-Road non-
Diesel Light
Duty Vehicles
31%
Non- Road
Equipment-
Gasoline
16%
Other
11%
9
Wildfires
15%
Prescribed
Fires
12%
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)
69%
Other
Other
8%
Non-Road
Equipment
- Gasoline
3%
Oi & Gas
Production
Wildfires
Prescribed
Petroleum Systems
6%
Coal Mining
Landfills
Agriculture-
Animal
Husbandry
36%
Consumer &
Commercial
Solvent Use
3%
On-Road non-
Diesel Light
Duty Vehicles
3%
Natural Gas
Systems
25%
Sources: A)-C) 2014 U.S. EPA National Emissions Inventory, Version 2 (U.S. EPA. 2018a) and; D) 2016 U.S. Inventory of
Greenhouse Gases (U.S. EPA. 2016c).
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, is included for context.
D) methane (CH4).
September 2019
1-10
DRAFT: Do Not Cite or Quote
-------
A)
I
NOx
12000
10000
8000
6000
4000
2000
0
CI VOCs
u
B
s
B)
8
Combustt
EG Us
Petroleum
Highway
Vehicles
& Related
industries
Solvent
witatires
Off-Highway
Equipment
2002 2005 ZOOS 2011 2014 2017
Inventory Year
CO
60000
50000
40000
30000
20000
Wildfires
10000
Inventory Year
2002 2005 2008 20112014 2017
inventory Year
D)
Tf
M'nfnb
mm :!>os 200s 2011201a 20 w
Year
Agrwuliurp • Awm! Husbandry
-Natural Gas Systems :
•Landfills
-Coal Mining
•Petroleum Systems
01 her
• Petroleum & Related Industries — Fuel Combustion - E6Us
— Other Industrial Processes
= Storage and Transport
aste Disposal & Recycling
ghway Vehicles
— Off Hiphway Equipment
— • es
• Miscellaneous (w/o Wildfires)
• Chemical & Allied Product MFG
- Fuel Combustion - Industrial
Solvent
• Metals Processing
• l-iiel 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. 2016c).
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.
September 2019
1-11
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
1.3.1.1.1 U.S. Anthropogenic 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. 2018b) and
to the health effects of NOx (U.S. EPA. 2016b).
• 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 along with anthropogenic VOC emissions. Biogenic
sources contribute substantially more to the U.S. emissions inventory than anthropogenic sources, can
play an important role urban ozone formation and are, therefore, included 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
an hour, whereas methane has an atmospheric lifetime of about a decade. In urban areas, compounds
representing all classes of VOCs and CO are important for ozone formation. In nonurban vegetated areas,
biogenic VOCs emitted from vegetation tend to dominate the VOC budget.
September 2019
1-12
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
• 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.
2016c). 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.
September 2019
1-13
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
1.3.1.2.1
Global Methane
The 2013 Ozone ISA (U.S. EPA. 2013) reported an estimate by Zhang et al. (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 NOx-saturated 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 CHVyear (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 CHVyear 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 average 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
September 2019
1-14
DRAFT: Do Not Cite or Quote
-------
1 increased at a rate of 3.5 + 0.2 ppb/year (Saunois et al.. 2016). Some recent studies suggest that
2 the methane increases were mainly due to increases in fossil fuel production (e.g., coal and oil
3 and gas industry) and agricultural emissions, while other studies point to large uncertainties in
4 natural emissions (Van Dingenen et al.. 2018). Modeling studies also suggest that natural sources
5 contribute to the inter-annual variability of methane, while anthropogenic emissions, mainly
6 emitted in the Northern Hemisphere, have played a major role in the increase of methane
7 observed since 2005 (Bader et al.. 2017).
8 • Global tropospheric ozone levels are enhanced when methane increases. Studies suggest that
9 increases in global methane since the 1800s have yielded higher levels of global tropospheric
10 ozone (and ground-level ozone) worldwide (NASEM. 2018). Studies indicate that there is an
11 approximately linear relationship between anthropogenic methane emissions and tropospheric
12 ozone, such that for every teragram per year decrease in methane emissions, ozone could decrease
13 by 11 ppt to 15 ppt (Tiore et al.. 2008).
14 • Global methane abundance contributes to rising U.S. surface ozone during all months. Based on
15 a set of transient chemistry-climate model simulations between 2005 and 2100, the global
16 methane abundance contributes to rising surface ozone during all months, with the largest
17 influence during cooler months when the ozone lifetime is longer (Rieder et al.. 2018; Clifton et
18 al.. 2014). These simulations indicate that the sensitivity of the ozone mixing ratio to potential
19 changes in global methane abundance is about 7-16 ppb over the northeastern U.S. and by about
20 12-19 ppb over the intermountain western U.S. at the end of the 21st century (Clifton et al..
21 2014).
1.3.1.2.2 International Emissions of Ozone Precursors
22 Ozone precursor emissions by countries that are "upwind" of the U.S. can contribute to U.S.
23 ozone. As described in earlier assessments (U.S. EPA. 2013. 2006a. b), under certain atmospheric
24 conditions, precursors emitted by large cities and other sources can be lofted above the boundary layer
25 into the high-altitude zone referred to as the "free troposphere." (see Figure 1-1). NOx and ozone have
26 significantly longer atmospheric residence times in this colder atmospheric zone due to slower rates of
27 reaction than they have near Earth's surface (Rastigeiev et al.. 2010). Furthermore, NOx can react to form
28 reservoir species (i.e., species that can remain stable over very long distances) at these altitudes. These
29 reservoir species include PAN and similar compounds that become unstable at the warmer temperatures
30 of the lower troposphere, regenerating reactive NOx.
31 Large-scale atmospheric flows in the free troposphere can transport these pollutants and their
32 reaction products (i.e., ozone precursors and ozone formed within the plume) across continents and
33 oceans. Plumes from these international sources experience shear processes and dilution during advection
34 downwind. However, distinctive, coherent plumes have been observed by aircraft, sondes, and satellites
35 for a week or more. Downward mixing from the upper troposphere by way of other meteorological
36 processes, such as convective mixing, can then bring ozone down into the boundary layer.
37 International sources of ozone precursors do vary in significance for USB ozone, depending on
38 their relationship with the continental U.S. with respect to atmospheric dynamics and long-range
39 circulation patterns. Asia, as described in previous ISAs (U.S. EPA. 2013. 2006a. b), has been an
September 2019
1-15
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
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.. 2018). 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, DECSO-GOME2a, EnKF-MIROC, EnKF-CHASER) 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 ofFigure 1-4.
September 2019
1-16
DRAFT: Do Not Cite or Quote
-------
A)
C)
10
I 4
s
I M'l uri
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Mg N km'2 yr-1
, 'm'J
0*
% 1
/'
• - DCCSO-OMI
• - DCCSO-GOME2*
- - CnKF-MPiOC
• - EnXT-CHASER
— WAS v2.l
REAS v2.2
— mbc
— EDGAR
D)
1.2
1.0
0.8
0.6
Mainland China
I • • D€CSOOM|
-- OCCSO-GOME2*
-• EnKF-MIROC
1 -- CmcF-CMASER
— REAS v2,l
— REAS v2.2
/>::•. \
/ ~ • %
— MEIC
»/ -»\ •
||— EDGAR
ft/
X/
2002 2004 2006 2008 2010 2012 2014
Yetr
2000 2002 2004 2006 2008 2010 2012 2014
Year
Source: Adapted from Ding et ah (2017). Permission pending.
Figure 1-4 Asian anthropogenic ozone precursor emission trends. A) The
study domain, indicating annual NOx flux rates by location,
B) Annual 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 • Stringent air quality standards implemented in 2013 within China have markedly reduced
2 national emissions. Zheng et al. (2018' applied the bottom-up inventory model underlying the
3 Multi-resolution Emission Inventory for China (MEIC) to estimate anthropogenic emissions for
4 31 Chinese provinces. Figure 1-5 shows these estimates, aggregated to provide annual national
5 emissions values. The results of this accounting indicate that China's emissions of NOx and CO
6 have declined by 17 and 27%, respectively, while nonmethane VOCs grew by approximately
7 5 Tg/year between 2010 and 2017. Zheng et al. (2018) analyzed this inventory using index
8 decomposition analysis to identify the drivers behind these changes. The results of this analysis
DCCSO-OMI
0€CSO-GOM€2*
CnKF-mnOC
CnKF-CMASER
RE AS v2.1
REA5 v2J
CAPSS
EOGAA
°'VboO 2002 2004 2006 2008 2010 2012 2014
rur
September 2019
1-17
DRAFT: Do Not Cite or Quote
-------
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.
o>
co
.2 20
c/>
CO
LU
250
125
NMVOC A
A A A
2010 2012 2014 2016
2010 2012 2014 2016
2010 2012 2014 2016
Note: Red = power sector emissions; yellow = industrial emissions; green = residential emissions; blue = transportation emissions;
purple = solvent use.
Source: Adapted from Zheng et al. (2018). Permission pending.
Figure 1-5 Anthropogenic ozone precursor emission trends derived using
the MEIC emissions model. 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.
1.3.1.3 Natural Ozone Precursor Emissions
4 Ozone attributed to natural sources is formed through photochemical reactions involving natural
5 emissions of ozone precursors from vegetation, microbes, animals, burning biomass (e.g., forest fires),
6 and lightning.
1.3.1.3.1 Biogenic Nitrogen Oxide Emissions: Fertilized Soils
7 Biogenic sources of NOx were not discussed in either the 2013 Ozone ISA or the ISA for the
8 Oxides of Nitrogen—Health Criteria (U.S. EPA. 2016b. 2013). The topic was briefly mentioned in the
9 ISA for Oxides of Nitrogen, Oxides of Sulfur, and Particulate Matter—Ecological Criteria (2nd external
10 review draft) (U.S. EPA. 2018b). Microbial nitrification (NIL -> NO;, ) and denitrification (NO: -$¦ Na)
11 processes in soils produce NO, contributing to local and regional atmospheric NOx concentrations. Soil
September 2019
1-18
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
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
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.. 2018).
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. 2018b. 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. 2016a). However the fraction of total soil NOx due to fertilizer
applications versus from natural soils are not reported separately. U.S. EPA (2018b) 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. 2016a).
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
September 2019
1-19
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
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 Landuse 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.
• 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). cited in Jaffe et al. (2018). 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
September 2019
1-20
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
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
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
average 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..
September 2019
1-21
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
2013) further supports the estimate of biomass burning's contribution to mean tropospheric ozone
concentrations given by Jaffe et al. (20IS), 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
(Fischer et al.. 2018; Busilacchio et al.. 2016).
• 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 a typical 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
September 2019
1-22
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
conducted (Schreier etal.. 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.
1.3.1.3.4 Lightning 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 peroxylpropionyl 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 derivedfrom 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 CH4. 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 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
September 2019
1-23
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
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. 2013V The descent of stratospheric ozone into the troposphere is
along isentropic surfaces, and 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).
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 etal. (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. Langford 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, Nevada 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
time scales (up to 2 weeks) and may mix downward and affect surface ozone concentrations
(Stohl et al.. 2000). For example, the influence of stratospheric intrusions have been seen at
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
September 2019
1-24
DRAFT: Do Not Cite or Quote
-------
1 and daily data (Langford et al.. 2009) or to smaller ozone increases over several days (Lin et al..
2 2012a).
3 • Quantifying the contribution of STE to surface ozone remains challenging and is a source of
4 uncertainty in estimating USB ozone. Stratosphere-troposphere exchange of ozone has been
5 observed using ground measurements, in situ aircraft or balloon measurements; through remote
6 sensing (lidar, satellite); identified with reanalysis data; and modeled via chemical transport and
7 global chemistry models. However, STE's contribution to USB ozone remains hard to quantify.
8 As previously mentioned, the AM3 global chemistry model (Lin et al.. 2012b) has been used to
9 estimate the stratospheric ozone contribution from deep intrusion events to be between 17 and
10 40 ppb at high surface ozone sites during springtime in the western U.S. Stratospheric intrusion
11 events reaching the surface have less influence on surface ozone during the summer months when
12 total ground-level ozone concentrations tend to be highest.
1.3.2.2 Deep Convective Mixing
13 Since the previous assessment, studies of the dynamics within thunderstorm anvil clouds has
14 revealed that deep convection can entrain stratospheric ozone and draw it down into the upper
15 troposphere. The Deep Convective Cloud and Chemistry (DC3) (Barth et al.. 2015) aircraft field
16 campaign over the central U.S. in May and June of 2012 identified this process using in situ
17 measurements (Huntrieser et al.. 2016; Pan et al.. 2014). Pan et al. (2014) observed in situ ozone mixing
18 ratios as high as 150 ppb in the upper troposphere adjacent to the storm cloud edge. They postulated that
19 these high concentrations could be the result of the dynamical response to tropospheric air overshooting
20 the tropopause, with stratospheric air being mixed down into the upper troposphere and wrapping around
21 the cloud edges of the thunderstorm outflow. The high ozone concentrations found at the storm edges
22 were anticorrelated with mixing ratios of measured CO, indicating the stratosphere as the source of the
23 ozone-enriched air. The study found ozone enhancement in the upper troposphere near storm cloud edges
24 on numerous flight sample cases that indicated the prevalence of the deep convection
25 stratospheric-tropospheric exchange (STE) mechanism during the 2012 field campaign. Although the
26 studies of Pan et al. (2014) and Huntrieser et al. (2016) provided observed data of deep convection
27 leading to the downward flux of stratospheric air into the troposphere, the authors did not estimate the
28 contribution deep convection made to USB or other measures of background ozone at the surface.
1.4 Ozone Photochemistry
29 The general photochemistry of tropospheric ozone is well understood and described in detail in
30 previous U.S. EPA integrated science assessments and criteria documents (U.S. EPA. 2013. 2006a) and
31 textbooks (Seinfeld and Pandis. 2006; Finlavson-Pitts and Pitts. 2000). Ozone is a product of the
32 oxidation of carbonaceous precursor gases in the presence of NOx. The involvement of NOx as an oxidant
33 ensures rapid ozone formation in the presence of solar radiation. This mechanism differs greatly from the
34 chemistry of stratospheric ozone formation or of ozone formed by lightning. The former requires the hard
September 2019
1-25
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
solar ultraviolet radiation present above the troposphere and the latter requires an electrical discharge at a
voltage sufficient to ionize molecular nitrogen.
Recent developments in ground-level ozone chemistry include observations and studies
concerning 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 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 over marine environments.
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; Rappenglueck 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/second (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
September 2019
1-26
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
snow surface) (Ahmadov et al.. 2015; Field et al.. 2015; Edwards et al.. 2014; Rappenglueck et
al.. 2014; Warneke et al.. 2014). 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| 'D|). 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.
(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. (Zhou et al.) conducted photochemical box model simulations using the Master
Chemical Mechanism v3.3 and found similar sensitivity results to that of Edwards et al. (2014)
for the UB wintertime ozone 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, Rappenglueck 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. Rappenglueck 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 (Rappenglueck 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.
September 2019
1-27
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
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 ct al.. 2015V Tuite et al. (20 IS) 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 which incorporated halogen chemistry and concluded that iodine is the most
influential 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 et al. (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 D.C. 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 et al.. 2011) and
therefore may overestimate background ozone transported over marine environments.
September 2019
1-28
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
• 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 et al.. 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.
• 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 Inter-annual 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 looked into the role
of meteorological effects on ozone and the year-to-year trends in ozone concentrations. Large-, regional-,
September 2019
1-29
DRAFT: Do Not Cite or Quote
-------
1 and local-scale atmospheric circulation patterns have been shown to influence both observed U.S.
2 background ozone and the local production of ground-level ozone. More recently, Nolte et al. (2018)
3 described emerging, robust evidence that the effects of climate warming on meteorology are negatively
4 affecting ground-level ozone concentrations.
5 Large-scale meteorology patterns influence USB ozone in several ways, including the likelihood
6 of the occurrence of deep stratospheric intrusions events in the western U.S., the transport of Asian
7 pollution to the U.S., and regional temperature and precipitation patterns which can influence the
8 frequency and distribution of wildfires emissions of VOC precursors from vegetation and
9 combustion-derived NOx. During localized stagnation events conducive to ozone production,
10 ground-level ozone concentrations can be further influenced by regional and large-scale meteorology
11 patterns or by regional-scale background ozone aloft being mixed down to the surface at urban sites.
12 Large-scale meteorology patterns help create the local-scale conditions that are conducive to
13 photochemical production of the ground-level. Example conditions include stagnation events associated
14 with high temperatures and high ozone concentrations versus cool, wet meteorological conditions
15 associated with lower ambient ozone concentrations.
16 Large-scale atmospheric circulation patterns are also subject to variability on annual and decadal
17 scales, which is reflected in the patterns of regional- and local-scale U.S. ground-level ozone
18 concentrations.
1.5.1 Meteorological Effects on Ozone Concentrations at the Ground Level
19 Meteorology at the regional and local scales establishes the chemical conditions that govern the
20 formation of ozone. Meteorological variables of importance at these scales include temperature, relative
21 humidity, wind speed, and precipitation. Synoptic-scale circulation (i.e., meteorological processes at
22 scales on the order of 1,000 km) are particularly important in determining ozone formation at regional and
23 local scales.
24 • Ozone was found to be strongly correlated to meteorology over the Intermountain West. Reddv
25 and Pflster (2016) found that surface ozone in the western U.S. is well correlated with the
26 500 millibar (mb) pressure level height. The study showed that the July mean max 8-hour ozone
27 increased when the mean July 500 mb height also increased. Over the western U.S., increases in
28 the 500 mb level are often associated with weather (clear skies, low wind) that is conducive to
29 ozone formation. By using the 500 mb height variable to detrend and correct for the influence of
30 meteorology, the study found that July max 8-hour ozone has steadily decreased from 1995 to
31 2013 in the Wasatch Front area surrounding Salt Lake City, UT. Over the same time period, a
32 general increase in July max 8-hour ozone was found along the Front Range (Denver area) of
33 Colorado. The study hypothesizes that ozone increases in Denver areas may be the result of
34 emissions associated with population growth and/or emissions from the increased activity of
35 nearby oil and gas development.
September 2019
1-30
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
• Summertime ozone in the eastern U.S. and Midwest are affected by synoptic-scale meteorology
patterns. Shen et al. (2015) quantified the sensitivity of max daily 8-hour 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 etal.. 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 max daily 8-hour ozone had decreased by
4% between 2000 and 2012 in Atlanta, GA.
1.5.2 Inter-annual 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.
Inter-annual (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 inter-annual 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 inter-annual 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 inter-annual (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) inter-annual 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
September 2019
1-31
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
deep stratospheric intrusions more frequently reach the surface (Lin et al.. 2015). 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 inter-annual oscillation.
Pinatubo
El Nino
La Nina
70
T r , , , 1 , r
- r2 (OBS. AM3 )=0.56
r2 (OBS, O3Strat)=0.43
r2 (AM3,03Strat)=0.74
r 1^l
35
1992
1996
2000
2004
2008
2012
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: Permission pending 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.
• The Atlantic Mnltidecadal 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.
• Inter-annual variability of wintertime ozone in the Intermountain U.S. has been tied to the phase
of the Arctic Oscillation. The Arctic Oscillation (AO) refers to an atmospheric circulation pattern
over the mid to high latitudes of the Northern Hemisphere and can have a strong influence on
weather and climate in the U.S., especially during winter. (Zhou et al.) found that year-to-year
variability of wintertime ozone concentrations in the intermountain West were correlated to the
AO, where a negative phase of AO is associated with higher wintertime ozone and vice versa.
Within oil and gas basins of the intermountain West, the colder surface temperatures associated
with the negative AO, along with consistent snow cover, can lead to elevated wintertime daily
maximum 8-hour ozone above 70 ppb. For example, wintertime 8-hour ozone concentrations
reached greater than 100 ppb in the Uinta Basin in Utah during the 2013 negative phase of the
AO. In contrast, the positive AO winters of 2012 and 2014, which lacked snow cover and had
September 2019
1-32
DRAFT: Do Not Cite or Quote
-------
1 warmer surface temperatures, saw much lower ozone concentrations, at daily maximum 8-hour
2 levels well below 70 ppb.
3 • Variability in climate can influence the activity of wildfires in the western U.S. The frequency and
4 distribution of fire activity in the western U.S. is influenced by temperature and precipitation
5 patterns associated with climate variability (Abatzoglou et al.. 2016).
1.6 Measurements and Modeling
1.6.1 Advances in Ozone Measurement Methods
6 This section provides a concise overview of methods used in monitoring networks and advances
7 in remote sensing using satellite-based technology for ozone and ozone precursor measurements. While
8 there is growing literature on low-cost sensors for ozone measurement, they have not been widely applied
9 in studies of atmospheric concentration distributions, human exposure, or health impacts, so studies about
10 them will not be reviewed here.
1.6.1.1 Network Monitoring Methods
11 A new Federal Reference Method (FRM) for ozone measurement was established in 2015
12 (40 CFR Part 50 Appendix D). The new ozone FRM is based on the detection of chemiluminescence
13 from the reaction of ozone with nitric oxide (NO). It was adopted because instruments based on
14 chemiluminescence from the reaction of ozone with ethylene were no longer commercially available.
15 Further discussion of chemiluminescence and UV measurements of ozone are presented in the 2013
16 Ozone ISA (U.S. EPA. 2013). Almost all State and Local Air Monitoring Stations (SLAMS) that report
17 data to the U.S. EPA's Air Quality System (AQS) database use the Federal Equivalence Method (FEM)
18 based on UV absorption.
1.6.1.2 Satellite-Based Remote Sensing Methods
19 Satellite instruments used to retrieve data on trace gases provide a routine and systematic data set,
20 with the measurements used to provide important column observations of ozone and ozone precursors at
21 scales that range from regional to global.
22 • Satellite-based remote sensing methods measure the total ozone column rather than ppm or ppb in
23 the atmosphere, and mathematical methods to derive tropospheric or surface ozone concentrations
24 are needed. Thus, there is more uncertainty in surface estimates derived from satellite-based
25 measurements than from monitoring network measurements.
September 2019
1-33
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
• While the vertical sensitivity of space-based measurements to ozone and its precursors is variable
and depends on the method (Martin. 2008). various satellite data sets have been shown in recent
modeling studies to provide useful observational constraints for tropospheric ozone (Emili et al..
2014) or both tropospheric ozone and NO2 (Huang et al.. 2014).
• Satellite observations also provide useful measurements to characterize processes that can
contribute to USB ozone, such as stratospheric transport/intrusions (Lin et al.. 2012a). They can
improve the characterization of USB ozone precursor emissions such as LNOx, VOCs from
biogenic sources (Mebust and Cohen. 2014; Mebustetal.. 2011). or the vertical lofting of
emissions into the upper troposphere [e.g., CO:ozone ratios Kim et al. (2013); Voulgarakis et al.
(2011)1.
• Because the measurements are consistent over time, the observed trends, seasonal or inter-annual,
provide quantitative information that can be used to test the representation of processes relevant
to USB ozone in models. For example, year-to-year climate variability (Ziemke et al.. 2010)
affects the distribution of upper tropospheric ozone or CO, and day-to-day temperature variations
affect the emission and chemistry of biogenic VOC and NOx emissions.
• Because ground-level concentration estimates from satellites can have substantially greater
uncertainty than total column ozone measurements, these technologies are most suitable for
investigating trends in total column ozone or in the upper troposphere (Gaudel et al.. 2018).
Currently, satellite-based estimates of ground-level ozone concentrations require considerable
supplemental information and/or assumptions about atmospheric characteristics and conditions.
• While the use of satellite-based remote sensing methods is becoming more widespread in each of
these applications, it is useful to understand the strengths, limitations, and appropriate use of
satellite measurements for estimating ozone and ozone precursors in the atmosphere.
Satellite-based spectrometers provide measurements of backscattered sunlight and thermal
radiation in ultraviolet (UV), visible (VIS), and infrared (IR) spectral ranges at various spectral
resolutions and spatial sampling rates. These satellite-based radiance measurements can detect
and quantify tropospheric aerosols and several trace gases, including ozone and ozone precursors.
Space-based retrieval of ozone and other trace gases from instruments aboard satellite platforms
must account for variability of the radiance measured (e.g., solar spectrum, albedo, IR emissivity,
and skin temperature), the path of light through the atmosphere (e.g., Rayleigh scattering, clouds,
temperature gradients for thermal IR), and the vertical profile of the absorbing species. Whenever
the above factors are not well characterized, a priori assumptions can affect the retrieval products
to varying degrees (Duncan et al.. 2014; Martin. 2008).
• The quantitative findings of a satellite study must be evaluated in the context of the uncertainty of
the underlying satellite data set and associated analysis methods. Factors to consider include, but
are not limited to, the maturity of the underlying satellite retrieval algorithm and data product for
a particular type of satellite observation, the robustness of validation efforts (short term vs. long
term) of algorithms and data products, the length of the study, and a clear description of the data
quality flag used to screen the quality of the satellite data. Of all these factors, validation of the
algorithm and data products is often the most difficult to accomplish because of the paucity of
critical geophysical measurements (e.g., tropospheric column ozone, NO2, CO, HCHO, partial
column amounts, or profiles) that are spatially and temporally consistent with the satellite
measurement concentrations. Operational networks, such as the Pandora Global Network and
Total Carbon Column Observing Network (TCCON) (Wunchetal.. 2011) are emerging to
support these efforts.
• Reprocessing of geophysical data products from calibrated radiance data continues to develop as
input assumptions (Russell et al.. 2011) and techniques improve (Zoogman et al.. 2014; Cuesta et
September 2019
1-34
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
al.. 2013; Natrai etal.. 2011). This will allow satellite products to be used in characterizing USB
ozone in the free troposphere versus the boundary layer on a routine basis.
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 (Rvdsaa 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.
• 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 (Yahva et al.. 2015). A
meta-analysis examining 6 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
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 inter-connected physical, chemical, and biological systems. Many studies
September 2019
1-35
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
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 average 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
(Kotaet 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 over the 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 over the years
in that area. With more coarse spatial resolution, global-scale models have also been used to
examine trends overthe U.S. (Lin etal.. 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.
1.7 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
September 2019
1-36
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
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: (1) the
average of the hourly observations over a 24-hour period (DA24), (2) the maximum hourly observation
occurring in a 24-hour period (MDA1), and (3) the max 8-hour running average of the hourly
observations occurring in a 24-hour period (MDA8).
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.
September 2019
1-37
DRAFT: Do Not Cite or Quote
-------
Table 1-1 Nationwide distributions of ozone concentrations (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
Time Period3
N Sites
N Obs
Mean
SD
Min
1
5
10
25
50
75
90
95
98
99
Max
Max Site ID
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
September 2019
1-38
DRAFT: Do Not Cite or Quote
-------
Table-1-1 (Continued): Nationwide distributions of ozone concentrations (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.
Winter = December-February, spring = March-May, summer = June-August, autumn = September-November.
September 2019
1-39
DRAFT: Do Not Cite or Quote
-------
Table 1-2 Nationwide distributions of ozone concentrations (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
U.S.
Region3
N Sites
N Obs
Mean
SD
Min
1
5
10
25
50
75
90
95
98
99
Max
Max Site ID
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
September 2019
1-40
DRAFT: Do Not Cite or Quote
-------
Table 1-2 (Continued): Nationwide distributions of ozone concentrations (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
U.S.
Region3
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)
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
September 2019
1-41
DRAFT: Do Not Cite or Quote
-------
Table 1-2 (Continued): Nationwide distributions of ozone concentrations (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).
September 2019
1-42
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
if » V
* Ycjr Round Only (11 lines) «i Wd f m Sewon Oify i-JS 1 Site;} •tot* Data Sets (79S 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 Tabic 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.
• 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.
September 2019
1-43
DRAFT: Do Not Cite or Quote
-------
1 Figure 1-8 through Figure 1-11 summarize ambient ozone concentration patterns and trends.
2 These figures contain data from both warm-season and year-round monitors for all monitors that met the
3 completeness criterion of 75% data capture. The data sets used in Table 1-1 and Table 1-2 are combined,
4 and the data in Figure 1-8 through Figure 1-11 reflect concentration metrics applied to the entire period of
5 monitor operation, rather than the same season across all monitors.
6 • Figure 1-8 shows the design values, or the 3-year avg of the annual 4th-highest 8-hour daily max
7 (MDA8) ozone concentrations for 2015-2017 (see Section 1.2.1.1). The highest design values
8 (>76 ppb) occur in central and southern California, Arizona, Colorado, Utah, Texas, along the
9 shore of Lake Michigan, and in the Northeast Corridor.
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) 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.
September 2019
1-44
DRAFT: Do Not Cite or Quote
-------
1 • Figure 1-9 shows the decreasing trend in the annual 4th-highest MDA8 ozone concentration from
2 882 U.S. monitors. The median annual 4th-highest MDA8 ozone concentration across those sites
3 decreased from more than 80 ppb in 2000 to less than 70 ppb in 2017. Other studies also reported
4 a decreasing trend over periods of 15 years or more for 4th-highest MDA8 ozone concentration or
5 other ozone concentration metrics associated with higher concentrations (Lefohn et al.. 2017;
6 Simon et al.. 2015; Strode et al.. 2015).
Trend in Annual 4th Highest Daily Maximum 8-hour Ozone (ppb)
100
3* 90
Q_
CL
C
0
1
I 80
u
c
o
u
O)
c
o
N
° 70
60 -
90th Percentile
75th Percentile
50th Percentile
25th Percentile
10th Percentile
NAAQS Level
National Trend Based on 882 Monitoring Sites
o
t—
CM
CO
lO
CO
r-
00
CD
O
T—
CM
CO
in
CD
Is-
o
o
O
O
o
o
o
o
o
o
t—
t—
•t—
'S—
r—
—
t—
t—
o
o
O
O
o
o
o
o
o
o
o
o
O
O
o
o
O
o
CM
C\J
CM
CM
CM
CM
CM
CM
CM
CM
CM
CM
CM
CM
CM
CM
CM
CM
Year
ppb = parts per billion.
Note: Although the trend lines are annual values, the 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.e83a.gov/air-trends/ozone-tr6nds.
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).
September 2019
1-45
DRAFT: Do Not Cite or Quote
-------
• Figure 1-10 shows a regional breakdown of the trend in 4th-highest MDA8 ozone concentrations.
Declines are observed in most regions, with the strongest declines in regions that had the greatest
4th-highest MDA8 ozone concentrations.
100
90
¦Q'
a
Q-
c
o
ts
QJ
O
c
o
U
0)
c
o
M
O
80
70
60
Regional Trends in Annual 4th Highest Daily Maximum 8-hour Ozone (ppb)
CD
Is-
CO
<7)
O
O
o
o
O
¦t—
O
o
o
O
o
CM
CM
CM
CM
CM
Year
m
CD
Is-
—
T—¦
o
O
o
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)
ppb = parts per billion.
Note: All monitors with 75% or greater data capture included, both warm-season and year-round monitors. 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), 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).
Source: U.S. EPA, National Air Quality: Status and Trends of Key Air Pollutants, https://vwwtf.epa.gov/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.
September 2019 1-46 DRAFT: Do Not Cite or Quote
-------
1 • 111 contrast to the decreasing trend in ozone metrics associated with higher concentrations, 5th
2 percentile ozone concentrations at the lower end of the ozone concentration distribution have
3 exhibited both increasing and decreasing trends in summer, depending on individual monitors,
4 and a generally increasing trend in winter from 1998-2013 (Simon et al.. 2015). These
5 observations demonstrate that a compression of the ozone concentration distribution has occurred
6 over this period.
7 • Figure 1-11 shows the geographic difference in the design values for all U.S. monitors between
8 the 2008-2010 period and the 2015-2017 period. Since the 2008-2010 period was used to
9 designate attainment and nonattainment areas for the 2008 ozone NAAQS, this comparison
10 indicates progress achieved since efforts to meet that standard began.
Change in Ozone Design Values from 2003 - 2010 to 2015 - 2017
• Decrease of 8 to 17 ppb (192 sites) © Decrease of 3 to 7 ppb (378 sites) 0 Change of less than 3 ppb (253 sites)
• Increase of 8 to 12 ppb (8 sites) © 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.
September 2019
1-47
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
• 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 inter-annual 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.
• 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).
••
• •
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) © 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.
September 2019
1-48
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
1.8 U.S. Background 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 chemical transport models (CTMs) to estimate USB ozone have found that
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
(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. 2006a. 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 inter-annual variability.
1.8.1 Modeling Strategies Applied to Estimate U.S. Background Ozone
As described in Section 1.2.2.1. USB ozone cannot be reliably estimated using ambient
monitoring data because monitors can be influenced by U.S. emissions, including both relatively nearby
emissions and interstate and hemispheric transport of ozone produced from U.S. emissions. Instead, air
quality model simulations are used to estimate USB ozone. The 2006 Air Quality Criteria Document
September 2019
1-49
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
(AQCD) for ozone (U.S. EPA. 2006a) 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. 2006a') was published, GEOS-Chem
(v4.3.3) (Tiore et al.. 2003) was the only model documented in the literature for calculating background
ozone concentrations, and it was used for the 2006 AQCD estimates of background ozone. Global-scale
simulations like those obtained from GEOS-Chem for the 2006 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 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 (Jaffe 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.
September 2019
1-50
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
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.
• 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; Dunkcr. 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.. 2018). 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
September 2019
1-51
DRAFT: Do Not Cite or Quote
-------
1 emissions from those source groups (Cohan and Napelenok. 2011; Grewe et al.. 2010). A challenge in the
2 use of source apportionment techniques is that both VOC and NOx precursors contribute to the
3 production of ozone, so rules must be developed to assign ozone production to either the VOC or NOx
4 source groups.
5 • Tagging approaches include CAMx Ozone Source Apportionment Technology (OSAT)
6 (http://www.camx.com/') and CMAQ Integrated Source Apportionment Method (ISAM) (Kwok
7 et al.. 2015). These approaches assign ozone production to either the tagged VOC or NOx
8 precursors depending on whether the ozone is produced in a VOC sensitive or NOx sensitive
9 chemical regime.
10 • Tagging can be applied to track contributions to ozone production based on source regions or
11 source types (Fiore et al.. 2002; Wang et al.. 1998) or to ozone transportation from the
12 stratosphere (Zhang et al.. 2014; Lin et al.. 2012a).
13 • Other tagging approaches that have been developed attribute source contributions to a single
14 precursor, either NOx or VOC. For rural or remote areas in which ozone is mostly produced in
15 NOx-sensitive chemical regimes, tracers can be used to track the source contributions from NOx
16 emissions (Pfister et al.. 2013; Emmons et al.. 2012). For urban areas where ozone is mostly
17 produced in VOC-sensitive conditions, tracers can be used to track the source contributions from
18 VOC emissions (Butler et al.. 2011; Ying and Krishnan. 2010).
19 • Tagging of ozone source contributions is more complex when natural and anthropogenic
20 precursors react to produce ozone. The CAMx model source apportionment technique includes an
21 option for preferentially attributing ozone production to anthropogenic precursors (Jaffe et al..
22 2018) when anthropogenic precursors react with natural precursors.
23 • Tracking techniques have been used to define an emissions-influenced background (EIB) ozone
24 concentration (see Section 1.2.2.4) that addresses the reduced lifetime of ozone that is transported
25 from the stratosphere or produced from natural and international precursors due to reaction with
26 and is chemically destroyed by anthropogenic emissions (Lefohn et al.. 2014).
1.8.1.3 Differences between Zero-Out and Source Apportionment Approaches
27 Due to the nonlinear character of ozone chemistry, removing emissions in model sensitivity or
28 model zero-out simulations will give a slightly different answer than tracking emissions contributions to
29 ozone production in a source apportionment approach. For this reason, USB estimated with a source
30 apportionment approach is identified in this document as apportionment-based USB (USBab) following
31 (Dolwick et al.. 2015). while USB without qualification (and without a subscript) generally refers to USB
32 based on zero-out or other source sensitivity-based modeling approaches (see Section 1.2.2.1). The
33 zero-out approach is more suited for answering the question "what ozone levels would exist in the
34 absence of all U.S. emissions?" while the source apportionment approach is more suited for answering the
35 question "what amount of current ozone comes from background sources?" The difference between USB
36 and USBab is small in remote areas most strongly affected by USB sources, but can be substantial in
37 urban areas strongly affected by anthropogenic sources that influence both production and destruction of
38 ozone (Dolwick et al.. 2015).
September 2019
1-52
DRAFT: Do Not Cite or Quote
-------
1 • Comparison of U.S. background estimates between the zero-out approach using CMAQ and a
2 tagged source apportionment method using CAMx gave similar April to October mean estimates
3 in rural areas, but the CAMx source apportionment approach produced lower estimates in urban
4 areas (Dolwick et al.. 2015).
5 • Differences in seasonal mean MDA8 U.S. background estimates from the zero-out and source
6 apportionment approaches were less than 2.5 ppb at 75% of locations after base case model bias
7 correction (Dolwick et al.. 2015).
8 • Differences between USB and USBab in urban areas indicate that ozone reductions resulting from
9 a reduction of U.S. anthropogenic ozone precursor emissions could be partially offset by the
10 absence of interactions with U.S. anthropogenic emissions that destroy USB ozone. However, this
11 offset may not apply to other photochemical oxidants that are produced along with U.S.
12 anthropogenic ozone (see Section 1.2.1).
1.8.1.4 Other Approaches for Estimating Background Ozone
13 One additional recently developed approach to estimating background ozone involves fitting a
14 running average of ozone concentrations over a long period to an exponential decay function (Parrish et
15 al.. 2017b).
16 • This approach is difficult to compare with modeling studies that rely on a more rigorous
17 definition of background. The regression approach also requires numerous assumptions, including
18 that U.S. emissions asymptotically approach zero and that background estimates remain constant
19 overtime.
20 • In addition, results reported in Parrish et al. (2017b) suggest that estimates of background ozone
21 are sensitive to assumptions of the exponential decay rate and the years of data included in the
22 analysis.
23 • It has been suggested that estimates using this approach are more representative of baseline ozone
24 concentrations plus some additional unquantified amount of ozone produced from local U.S.
25 anthropogenic emissions, rather than background concentration as defined by various modeling
26 approaches (Jaffe et al.. 2018).
1.8.1.5 Uncertainties and Model Disagreement
27 Jaffe et al. (2018) reviewed recent modeling results and reported that USB ozone estimates
28 contain uncertainties of about 10 ppb for seasonal average concentrations, with higher uncertainty for
29 MDA8 average concentrations. Because of uncertainty in model predictions, simple bias correction
30 approaches are useful to adjust model results for bias and error. However, these approaches might not be
31 reliable if the model has large errors in USB ozone and locally produced ozone. Accordingly, days with
32 poor model performance are typically excluded when using model results to estimate USB or other
33 measures of background ozone (Fiore et al.. 2014). There have been continued efforts to improve model
34 performance and better understand biases and uncertainties involved in the application of CTMs to
35 estimating USB or other measures of background ozone:
September 2019
1-53
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
• 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).
• 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 (Hogrefe et al.. 2018).
• 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 (Jaffe et al.. 2018).
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 (Jaffe etal.. 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, and has provided new evidence for both an inverse relationship between relative USB
contribution and total ozone concentration and a leveling off of baseline ozone concentrations that have
been increasing since monitoring was begun.
September 2019
1-54
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
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.
• Fiore et al. (2014) estimated summer NAB 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 concentrations of 40 to 45 ppb at
intermountain west monitors, compared with 25 to 35 ppb along the Pacific Coast.
• Guo et al. (2018b) estimated seasonal means for spring of 41 ppb for U.S. EPA Region 8, the
region most closely corresponding to the intermountain West, but seasonal means for all other
U.S. EPA regions were narrowly distributed from 34 to 37 ppb.
1.8.2.2 Seasonal Trends 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 (Fiore 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 have come up with conflicting conclusions about seasonal trends in USB.
Higher seasonal mean USB concentrations in spring than in winter were reported for
intermountain western sites (Fiore 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.. 2018b). 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.
September 2019
1-55
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
• 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 1-11 and Figure 1-12 based on CMAQ and CAMx model results for USB from 2007 do
not show an inverse relationship between USB and total ozone concentrations across the U.S., but
the results do show that 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. (2018b) 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.. 2018a; Guo et al.. 2018b; 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. (2018b). 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
September 2019
1-56
DRAFT: Do Not Cite or Quote
-------
1 outliers above 0.75 in Figure 1-13a and Figure l-14a between 70 and 90 ppb. These are most
2 often associated with model-predicted ozone events from wildfires.
3 • There is consistent evidence across several studies using different background measurement
4 approaches that USB or other background concentration estimates on most days with high ozone
5 concentrations have been generally predicted to be similar to or smaller than seasonal mean USB
6 ozone estimates in the eastern U.S. and in urban and low-elevation areas of the western U.S., and
7 an inverse relationship between relative USB contribution and total ozone concentration in these
8 areas has been consistently predicted. This contrasts with high-elevation locations in the western
9 U.S., where USB and NAB have been consistently predicted to increase with total ozone
10 concentration.
September 2019
1-57
DRAFT: Do Not Cite or Quote
-------
c
8
CD
<
Q
2
m
§«-
o
<
s
o
T3
j3
H
ro 20 -
CO
0 "
<25 25-30 30-35 35-40 40-45 45-50 50-55 55-80 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)
b
gl
IIIIIIIIIIII
Sill
Sit
<25 25-30 30-35 35-40 40-45 45-50 50-55 55-60 60-65 65-70 70-75 75-80 80-85 85-90 90-95 95-100 >100
Bins of CAMx Base Model MDA8 Ozone (ppb)
ppb = parts per billion.
Source: Dolwick et al. (2015).
Figure 1-13 CMAQ (a) and CAMx (b) estimates of daily distributions of
bias-adjusted USB MDA8 ozone concentration (ppb) for the
period April-October 2007, binned by base model MDA8 ozone
concentration ranges.
a
i
¦ j
II ffl EI E
II
II
September 2019
1-58
DRAFT: Do Not Cite or Quote
-------
° 0 75
•r o.fo
2 0 25
<25 25-30 30-35 35-40 40-45 45-50 50-55 55-60 eO-65 65-70 70-75 75-80 80-85 85-90 90-95 95-100 >100
Bins ofCMAQ Base Model MDA8 Ozone (ppb)
¦o 1.00
y 0.75
ra 0.50
<25 25-30 30-35 35-40 40-45 45-50 50-55 55-60 60-65 65-70 70-75 75-30 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. (2015V
Figure 1-14 CMAQ (a) and CAMx (b) estimates of daily distributions of
bias-adjusted USB ozone fraction at monitoring locations across
the western U.S. for the period April-October 2007, binned by
base model MDA8 ozone concentration ranges.
September 2019
1-59
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
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 inter-annual 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.. 2015). 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 etal.. 2015). and satellites are poorly
suited for detecting ground-level ozone. In spite of 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 etal.. 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 inter-annual 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
September 2019
1-60
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
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 inter-annual
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. The
primary focus is 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 (Section 1.4).
• The effects of local precursor emissions controls can be masked by meteorological variability.
Inter-annual 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,
September 2019
1-61
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
stratosphere-troposphere exchange, biogenic emissions, and integration with meteorological
models (Section 1.6V
• For the 2015-2017 time period, the 98th percentile MDA1, MDA8, and DA24 concentrations are
78, 69, and 53 ppb, respectively. Over the 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).
September 2019
1-62
DRAFT: Do Not Cite or Quote
-------
1.10 References
Abatzoglou. JT: Koldea 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-
9326/11/4/045005.
Ahmadov. R: Mckeen. S: Trainer. M: Banta. R: Brewer. A: Browa S: Edwards. PM: de Gouw. JA: Frost. GJ:
Gilman. J: Helmig. D: Johnson. B: Karioq A: Koss. A: Langford. A: Lerner. B: Olsoa J: Oltmans. S:
Peischl. J: Petroa G: Pichugina. Y: Roberts. JM: Rversoa T: Schnell. R: Senff. C: Sweeney. C: Thompson.
C: Veres. PR: Warneke. C: Wild. R: Williams. EJ: Yuaa B: Zamora. R. (2015). Understanding high
wintertime ozone pollution events in an oil- and natural gas-producing region of the western US. Atmos
ChemPhys 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.
Altshuller. AP: Lefohn. AS. (1996). Background ozone in the planetary boundary layer over the United States. J
Air Waste Manag Assoc 46: 134-141. http://dx.doi.org/10.108Q/10473289.1996.10467445.
Appel. KW: Napelenok. SL: Foley. KM: Pve. HOT: Hogrefe. C: Lueckea 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 of
WRF-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: Hausmana P: Jones. N: Krummel. P: Murata. I: Morino. I:
Nakaiima. H: O'Dohertv. S: Paton-Walsh. C: Robinson. J: Sandria 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 ChemPhys 17: 2255-2277. http://dx.doi.org/10.5194/acp-17-2255-
2017.
Barth. MC: Cantrell. CA: Brune. WH: Rutledge. SA: Crawford. JH: Huntrieser. H: Carey. LP: Macgorman. D:
Weisman. M: Pickering. KE: Bruning. E: Anderson. B: Apel. E: Biggerstaff. M: Campos. T: Campuzano-
Jost. P: Cohea 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:
Paa LL: Peischl. J: Pollack. I: Richter. D: Riemer. D: Rversoa 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 AmMeteorol Soc 96: 1281-1309.
http://dx.doi.org/10.1175/BAMS-D-13-00290.1.
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.
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: Pi Carlo. P: Aruffo. E: Biancofiore. F: Salisburgo. CD: Giammaria. F: Bauguitte. S: Lee. J:
Moller. S: Hopkins. J: Punjabi. S: Andrews. S: Lewis. AC: Partington. M: Palmer. PI: Hver. E: Wolfe. GM.
(2016). Production of peroxy nitrates in boreal biomass burning plumes over Canada during the BORTAS
campaign. Atmos ChemPhys 16: 3485-3497. http://dx.doi.org/10.5194/acp-16-3485-2016.
September 2019
1-63
PRAFT: Po Not Cite or Quote
-------
Butler. TJ: Vermevlea FM: Run . 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.ll.040.
Bvua 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
MechRev 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: Herndoa 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.
Cliftoa 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.
Cohaa PS: Napelenok. SL. (2011). Air quality response modeling for decision support. Atmosphere (Basel) 2:
407-425. http://dx.doi.org/10.3390/atmos20304Q7.
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.
Cuesta. J: Eremenko. M: Liu. X: Dufour. G: Cai. Z: Hoepfner. M: von Clarmann. T: Sellitto. P: Foret. G:
Gaubert. B: Beekmana M: Orphal. J: Chance. K: Spurr. R: Flaud. JM. (2013). Satellite observation of
lowermost tropospheric ozone by multispectral synergism of IASI thermal infrared and GOME-2 ultraviolet
measurements over Europe. Atmos ChemPhys 13: 9675-9693. http://dx.doi.org/10.5194/acp-13-9675-2013.
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: Simoa H: Tonnesea 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/10.1016/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.
Duncaa BN: Prados. AI: Lamsal. LN: Liu. Y: Streets. DG: Gupta. P: Hilsenrath. E: Kaha RA: Nielsea JE:
Beversdorf. AJ: Burtoa SP: Fiore. AM: Fishman. J: Henze. DK: Hostetler. CA: Krotkov. NA: Lee. P: Lia
M: Pawsoa 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.
September 2019
1-64
DRAFT: Do Not Cite or Quote
-------
Dunker. AM. (1981). Efficient calculation of sensitivity coefficients for complex atmospheric models. Atmos
Environ 15: 1155-1161. http://dx.doi.org/10.1016/0004-6981(8 D90305-X.
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: Degouw. JA: Dube. WP: Field. RA: Flvna
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: Thompsoa CR: Trainer.
MK: Tsai. C: Veres. PR: Washenfelder. RA: Warneke. C: Wild. RJ: Young. CJ: Yuaa 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/naturel3767.
Emery. C: Jung. J: Downey. N: Johnsoa J: Jimenez. M: Yarwood. G: Morris. R. (2012). Regional and global
modeling estimates of policy relevant background ozone over the United States. Atmos Environ 47: 206-217.
http://dx.doi.Org/10.1016/i.atmosenv.2011.ll.012.
Emili. E: Barret. B: Massart. S: Le Flochmoen. E: Piacentini. A: El Amraoui. L: Pannekoucke. O: Cariolle. D.
(2014). Combined assimilation of IASI and MLS observations to constrain tropospheric and stratospheric
ozone in a global chemical transport model. Atmos Chem Phys 14: 177-198. http://dx.doi.org/10.5194/acp-
14-177-2014.
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/10.1029/2010JD014079.
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/B978-0-12-257060-5.X500Q-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/10.1029/2001JP00Q982.
Fiore. AM: Oberman. JT: Lia MY: Zhang. L: Cliftoa OE: Jacob. PJ: Naik. V: Horowitz. LW: Pinto. JP: Milly.
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.
Fischer. EV: Zhu. L: Payne. VH: Wordea JR: Jiang. Z. he: Kulawik. SS: Brev. S: Hecobiaa A: Gombos. P:
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.
Foley. KM: Hogrefe. C: Pouliot. G: Possiel. N: Roselle. SJ: Simoa H: Timin. B. (2015). Pynamic 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.
September 2019
1-65
PRAFT: Po Not Cite or Quote
-------
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 ChemPhys 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.
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 Mufios. 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.
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: Geroa 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/S1352-2310r99-)00465-3.
Guo. JJ: Fiore. AM: Murray. LT: Jaffe. DA: Schnell. JL: Moore. CT: Millv. GP. (2018). Average versus high
surface ozone levels over the continental USA: Model bias, background influences, and interannual
variability. Atmos ChemPhys 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.
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: Thompsoa CR: Evans. J: Bovlan. 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/es405046r.
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.
Hennemaa 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. Ill 1/gcb. 13644.
September 2019
1-66
DRAFT: Do Not Cite or Quote
-------
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: Lia M: Park. RJ. (2018). Impacts of
different characterizations of large-scale background on simulated regional-scale ozone over the continental
United States. Atmos ChemPhys 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/10.1002/2014JD022732.
Huang. LL: Pan. C: Yu. TT: Guo. K. un: Wang. X. hui: Zhang. J. unYan: Wang. H. zhi: Gao. S. (2014).
BENEFICAL THERAPEUTIC EFFECT OF CHINESE HERBAL XINJI'ERKANG FORMULA ON
HYPERTENSION-INDUCED RENAL INJURY IN THE 2-KIDNEY-l-CLIP HYPERTENSIVE RATS.
African Journal of Traditional, Complementary and Alternative Medicines 11: 16-27.
http://dx.doi.org/10.4314/aitcam.vl li5.3.
Huang. M: Bowmaa KW: Carmichael. GR: Pierce. RB: Wordea HM: Luo. M: Cooper. OR: Pollack. IB:
Rversoa TB: Browa 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: Diskia GS. (2013b). Impacts of transported background pollutants
on summertime western US air quality: model evaluation, sensitivity analysis and data assimilation. Atmos
ChemPhys 13: 359-391. http://dx.doi.org/10.5194/acp-13-359-2013.
Huntrieser. H: Lichtenstera M: Scheibe. M: Aufmhoff. H: Schlager. H: Pucik. T: Minikia A: Weinzierl. B:
Heimerl. K: Fuetterer. D: Rappengluck. B: Ackermann. L: Pickering. KE: Cummings. KA: Biggerstaff. MI:
Bettea 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.
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: Hogrefe. 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:
Lia M: Moore. T. (2018). Scientific assessment of background ozone over the US: Implications for air
quality management. 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.ll.063.
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.
Kim. PS: Jacob. DJ: Liu. X: Warner. JX: Yang. K: Chance. K: Thouret. V: Nedelec. P. (2013). Global ozone-CO
correlations fromOMI and AIRS: constraints on tropospheric ozone sources. Atmos ChemPhys 13: 9321-
9335. http://dx.doi.org/10.5194/acp-13-9321-2013.
September 2019
1-67
DRAFT: Do Not Cite or Quote
-------
Kirschke. S: Bousauet. P: Ciais. P: Saunois. M: Canadell. JG: Dlugokenckv. EJ: Bergamaschi. P: Bergmana 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; Simpsoa 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. D: Hogrefe. C: Forkel. R: Werhaha J; Hirtl. M: Perez. JL: San Jose. R: Giordano. L: Brunner. D:
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.ll.066.
Knowland. KE; Ott. LE; Duncan. BN; Wargaa 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. D; 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/10.1016/i.atmosenv.2015.03.027.
Krotkov. NA; McLindea 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; Dickersoa RR: He. H; Lu. Z;
Streets. DG. (2016). Aura OMI observations of regional S02 and N02 pollution changes from 2005 to 2015.
Atmos ChemPhys 16: 4605-4629. http://dx.doi.org/10.5194/acp-16-4605-2016.
Kwok. RHF; Baker. KR; Napelenok. SL; Tonnesea 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.
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; Aikia 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; Gustia MS; Lin. MY; Marchbanks. RD; Pierce. RB; Sandberg.
SP: Senff. CJ: Weickmana 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: Lia 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.
Lee. L; Teng. AP; Wennberg. PO; Crounse. JD: Cohea 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.ll.033.
Lefohn. AS: Mallev. CS: Simoa 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.
September 2019
1-68
DRAFT: Do Not Cite or Quote
-------
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: Cohea RC: Crounse. JD: Teng. AP: Wennberg. PO: Lee. BH: Lopez-Hilfiker. FD:
Thornton. JA: Peischl. J: Pollack. IB: Rversoa TB: Veres. P: Roberts. JM: Neumaa 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.
Lia M: Fiore. AM: Cooper. OR: Horowitz. LW: Langford. AO: Lew. 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.
Lia M: Fiore. AM: Horowitz. LW: Cooper. OR: Naik. V: Hollowav. J: Johnsoa BJ: Middlebrook. AM:
Oltmans. SJ: Pollack. IB: Rversoa TB: Warner. JX: Wiedinmver. C: Wilsoa J: Wvmaa 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.
Lia 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.
Lia M: Horowitz. LW: Oltmans. SJ: Fiore. AM: Faa 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: Pavtoa 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: Herndoa 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. O: 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.
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.
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/10.5194/acp-17-12449-2017.
McDonald-Buller. EC: Allea DT: Brown. N: Jacob. DJ: Jaffe. D: Kolb. CE: Lefoha 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.
September 2019
1-69
DRAFT: Do Not Cite or Quote
-------
Mebust. AK: Russell. AR: Hudmaa RC: Valia 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. Q: Covle. M: Derwent. R: Fowler. D: Granier. C: Law. KS:
Mills. GE: Stevenson. PS: 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.
Mufiiz-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: Hudmaa RC: Koshak. WJ. (2012). Optimized regional and interannual
variability of lightning in a global chemical transport model constrained by LIS/OTD satellite data. J
GeophysRes 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.
Natrai. V: Liu. X: Kulawik. S: Chance. K: Chatfield. R: Edwards. DP: Eldering. A: Francis. G: Kurosu. T:
Pickering. K: Spurr. R: Wordea 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.
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/.
Oetien. H: Payne. VH: Neu. JL: Kulawik. SS: Edwards. DP: Eldering. A: Worden. HM: Wordea 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.
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.
Paa LL: Homever. CR: Honomichl. S: Ridley. BA: Weismaa M: Barth. MC: Hair. JW: Fenn. MA: Butler. C:
Diskia 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: Lamarque. JF: Naik. V: Horowitz. L: Shindell. DT: Staehelia 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.
Parrish. DP: Law. KS: Staehelin. J: Derwent. R: Cooper. OR: Tanimoto. H: Volz-Thomas. A: Gilge. S: Scheel.
HE: Steinbacher. M: Chaa 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.
September 2019
1-70
DRAFT: Do Not Cite or Quote
-------
Parrish. DP: 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.
Parrish. DP: Young. LM: Newmaa MH: Aikin. KC: Rverson. TB. (2017b). Ozone design values in southern
California's air basins: Temporal evolution and US background contribution. J Geophys Res Atmos 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.
Rappenglueck. B: Ackermann. L: Alvarez. S: Golovko. J: Buhr. M: Field. RA: Soltis. J: Montague. DC: Hauze.
B: Adamson. S: Risch. D: Wilkersoa 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. PS: 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/10.1002/2015JD02384Q.
Reidmiller. PR: Fiore. AM: Jaffe. DA: Bergmana D: Cuvelier. C: Dentener. FJ: Duncan: Brvaa N: Folberth. G:
Gauss. M: Gong. S: Hess. P: Jonsoa JE: Keating. T: Lupu. A: Marnier. E: Park. R: Schultz. MG: Shindell.
PT: 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: Perring. AE: Valia LC: Bucsela. EJ: Browne. EC: Wooldridge. PJ: Cohen. RC. (2011). A high
spatial resolution retrieval of N02 column densities from OMI: Method and evaluation. Atmos Chem Phys
11: 8543-8554. http://dx.doi.org/10.5194/acp-11-8543-2011.
Rvdsaa. 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.
Rvu. YH: Hodzic. A: Barre. J: Pescombes. 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-18-7509-2018.
Sandu. A: Paescu. PN: 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.
September 2019
1-71
PRAFT: Po Not Cite or Quote
-------
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: Simoa 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: Peregoa A: Ciais. P: Canadell. JG: Dlugokenckv. EJ: Etiope. G:
Bastvikea D: Houweling. S: Janssens-Maenhout. G: Tubiello. FN: Castaldi. S: Jacksoa 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: Kleinea T: Krummel. P: Lamarque. JF: Langenfelds. R. av: Locatelli. R: Machida. T:
Maksvutov. S: Mcdonald. KC: Marshall. J: Meltoa JR: Morino. I: Naik. V: O'Dohertv. S: Parmentier. FJW:
Patra. PK: Peng. C: Peng. S: Peters. GP: Pisoa I: Prigent. C: Prinn. R: Ramonet. M: Rilev. WJ: Saito. M:
Santini. M: Schroeder. R: Simpsoa IJ: Spahni. R: Steele. P: Takizawa. A: Thornton. BF: Tian. H: Tohiima.
Y: Viow. 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.
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
(2nd ed.). Hoboken, NJ: John Wiley & Sons. https://www.wilev.com/en-us/Atmospheric+Chemistry+and
+Physics%3A+From+Air+Pollution+to+Climate+Change%2C+3rd+Edition-p-9781118947401.
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-
15-10925-2015.
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.
Simoa 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.
Simoa H: Kimura. Y: McGaughev. G: Allen. DT: Brown. SS: Osthoff. HP: Roberts. JM: Bvua 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/10.1029/2008id01Q732.
Simoa 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/10.102 l/es504514z.
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.
September 2019
1-72
DRAFT: Do Not Cite or Quote
-------
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/10.1016/81352-2310(99100320-9.
Stone. D: Sherwen. T: Evans. MJ: Vaughaa 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: Logaa 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.
Tang. O: Prather. MJ: Hsu. J. (2011). Stratosphere-troposphere exchange ozone flux related to deep convection.
Geophys Res Lett 38: L03806. http://dx.doi.org/10.1029/2010GL046Q39.
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.
Trainer. M: Parrish. DP: Buhr. MP: Norton. RB: Fehsenfeld. FC: Anlauf. KG: Bottenheim. JW: Tang. YZ:
Wiebe. HA: Roberts. JM: Tanner. RL: Newmaa L: Bowersox. VC: Meagher. JF: Olszvna. KJ: Rodgers.
MO: Wang. T: Berresheim. H: Demeriian. KL: Rovchowdhurv. UK. (1993). Correlation of ozone with NOy
in photochemically aged air. J Geophys Res 98: 2917-2925.
Travis. KR: Jacob. DJ: Fisher. JA: Kim. PS: Marais. EA: Zhu. L: Yu. K: Miller. CC: Yantosca. RM: Sulprizio.
MP: Thompsoa AM: Wennberg. PO: Crounse. JD: St Clair. JM: Cohea RC: Laughner. JL: Dibb. JE: Hall.
SR: Ullmana 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: Flvna 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). (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.epa.gov/ncea/cfm/recordisplav.cfm?deid=149923.
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). (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/naaas/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://cfbub.epa. gov/ncea/cfm/recordisplav.cfm?deid=218686.
September 2019
1-73
DRAFT: Do Not Cite or Quote
-------
U.S. EPA (U.S. Environmental Protection Agency). (2011). MOVES (Motor Vehicle Emission Simulator).
Available online at http://www.epa.gov/otaa/models/moves/index.htm (accessed January 28, 2011).
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). (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, version 1
technical support document (Draft). Research Triangle Park, NC.
U.S. EPA (U.S. Environmental Protection Agency). (2016b). 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://ofmpub.epa.gov/eims/eimscomm.getfile7p download id=526855.
U.S. EPA (U.S. Environmental Protection Agency). (2016c). 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). (2017). Profile of version 1 of the 2014 National Emissions
Inventory. Washington, DC. https://www.epa.gov/sites/production/files/2017-
04/documents/2014neivl profile final aprill82017.pdf.
U.S. EPA (U.S. Environmental Protection Agency). (2018a). 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). (2018b). Integrated science assessment for oxides of
nitrogen, oxides of sulfur and particulate matter -Ecological 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.
U.S. EPA (U.S. Environmental Protection Agency). (2019a). Air markets program data. Available online at
https ://ampd.epa. gov/ampd/ (accessed
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
VanDingenen. 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/10.2760/82Q175.
Vinkea GCM: Boersma. KF: Maasakkers. JD: Adon. M: Martin. RV. (2014). Worldwide biogenic soil NOx
emissions inferred from OMIN02 observations. Atmos ChemPhys 14: 10363-10381.
http://dx.doi.org/10.5194/acp-14-10363-2Q14.
Voulgarakis. A: Telford. PJ: Aghedo. AM: Braesicke. P: Faluvegi. G: Abraham. NL: Bowmaa KW: Pvle. JA:
Shindell. DT. (2011). Global multi-year 03-C0 correlation patterns from models and TES satellite
observations. Atmos ChemPhys 11: 5819-5838. http://dx.doi.org/10.5194/acp-ll-5819-2011.
Wang. YH: Jacob. DJ: Logaa 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.
September 2019
1-74
DRAFT: Do Not Cite or Quote
-------
Warneke. C: Geiger. F: Edwards. PM: Dube. W: Petroa G: Kofler. J: Zahn. A: Browa SS: Graus. M: Gilman.
JB: Lerner. BM; Peischl. J: Rversoa 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 ChemPhys 14: 10977-10988. htto://dx.doi.org/10.5194/acp-14-
10977-2014.
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 ChemPhys 13: 837-850. htto://dx.doi.org/10.5194/acp-13-
837-2013.
Wu. SL: Duncaa BN: Jacob. DJ: Fiore. AM: Wild. O. (2009). Chemical nonlinearities in relating
intercontinental ozone pollution to anthropogenic emissions. Geophys Res Lett 36: L05806.
htto://dx.doi.org/10.1029/2008gl036607.
Wunch. D: Tooa GC: Blavier. JF: Washenfelder. RA: Notholt. J: Connor. BJ: Griffith. DW: Sherlock. V:
Wennberg. PO. (2011). The total carbon column observing network. Philos Transact A Math Phys Eng Sci
369: 2087-2112. htto://dx.doi.org/10.1098/rsta.2010.0240.
Xie. Y: Paulot. F: Carter. WPL: Nolte. CG: Lueckea DJ: Hutzell. WT: Wennberg. PO: Cohea RC: Pinder. RW.
(2013). Understanding the impact of recent advances in isoprene photooxidation on simulations of regional
air quality. Atmos ChemPhys 13: 8439-8455. htto://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 ChemPhys 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 ChemPhys 15: 2723-2747. http://dx.doi.org/10.5194/acp-15-2723-2015.
Yahva. K: Wang. K: Zhang. Y: Kleindienst. TE. (2015). 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.
Ying. O: Krishnan. A. (2010). Source contributions of volatile organic compounds to ozone formation in
southeast Texas. J Geophys Res Atmos 115: Article #D17306. htto://dx.doi.org/10.1029/2010JD013931.
Young. PJ: Naik. V: Fiore. AM: Gaudel. A: Guo. J: Lia 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: Christensea 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. htto://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
Environ45: 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.
September 2019
1-75
DRAFT: Do Not Cite or Quote
-------
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 ChemPhys 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: Cohaa A: Biazar. AP: Cohan. PS. (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: Thompsoa 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 ChemPhys 18: 14095-14111. http://dx.doi.org/10.5194/acp-18-
14095-2018.
Zhou. Y: Mao. H: Sive. BC. (In Press) Decadal trends and variability in intermountain west surface ozone near
oil and gas extraction fields. Atmos Chem Phys. http://dx.doi.org/10.5194/acp-2019-164.
Ziemke. J. R.: Chandra. S: Omaa LP: Bhartia. PK. (2010). A new ENSO index derived from satellite
measurements of column ozone. Atmos Chem Phys 10: 3711-3721.
Zoogman. P: Jacob. PJ: Chance. K: Liu. X: Lin. M: Fiore. A: Travis. K. (2014). Monitoring high-ozone events
in the US Intermountain West using TEMPO geostationary satellite observations. Atmos Chem Phys 14:
6261-6271. http://dx.doi.org/10.5194/acp-14-6261-2014.
September 2019
1-76
PRAFT: Po Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
APPENDIX 2
EXPOSURE TO AMBIENT OZONE
Overall ('(inclusions regarding Estimates of Exposure to. inihient ():.one for I se in
Epidemiologic Studies
• Since llie 2<>l ' O/oue ISA. ;id\;iuces li;i\e heeu ni;ide in se\er;il ;ippro;ielies lor
prediclum ;inihieul o/oue coiicciiiriilious ;is siiitou;iIcs lor exposure I !itois ;issoei;iled
w nh exposure iissessnieui mclIkkK ;ire olieu siniikir o\or ui"h;iii senles hcc;iuse
jinihieul o/oue coiicciiir;ilious lend lo 11;I\ c low sp;ili;il \ ;iri;ihilil\.
• I'OI' epidemiologic sllldlCs 1)1" sIlol'l-ICI'lll exposure lo JllllhlCIll O/OIIC. lIlC IlssoCKIlioil
helWCCIl exposure CslMIKIICs ;illd llCIlllll cl'lcels I1KI\ he UlldcrcsllllKllcd In llie
nic;isurcnieui or model used lo represeui exposure. ;md ilie cl'l'cel esiim;iie m;i\ li;i\e
reduced precision I !\ en w lien llie m;iumliide of llie ;issoei;iliou is iiiiceri;iiii. ilic line
ell'eel would 11kel> he hiruer iIkiii ilie esinmied ;issoei;iliou in these e;ises. I lie hi;is
;iud redueliou in precision ;ire l\pie;ill> sm;ill mi iikiuiiiiucle
• I-'or epidemiologic siudies i>l" louu-ierni exposure lo ;iuihieiil o/oue. depeudiuu on i lie
model ;iud seeu;ino heiuu modeled, llie ;issoei;iliou helweeu exposure esiim;iles ;iud
lie;illli effects in;i\ he uuderesiim;iied oro\crcsiini;iied ll is mueli more common lor
llie ;issoci;iiioii lo he uuderesiini;iied hec;iuse ue;ir-ro;id o/oue sc;i\eimiim c;iu resuli mi
ure;iler sp;iii;il x;in;ihililx due lo ;i redueliou m o/oue couceninilioii compared Willi
;iinhleiH o/oue measured ;il ;i lixed-siie mouiior The hi;is ;ind redueliou mi precision
;irc Ix pic;iIIx siikiII mi m;iuiiilude
• I !siini;iiiuu exposure wiilioul ;iccoiiuiiim lor iinie-;icli\ ii\ d;il;i ni;i\ resuli mi
uiideres|im;iliou of llie ;issoci;iliou ;iud reduced precision Allliouuli llie ni;imiiludc ol'
llie ;issoci;iliou helweeu exposure esiinuiles ;md lienllli elleels is uuceri;uii. l lie I rue
ell'eel lends io he hiruer ikiu ilie esiininled ;issoei;ilious mi iliese e;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 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).
September 2019
2-1
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
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 air purifiers. Some 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 and CO 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
(Sarnat et al.. 2000). A fixed-site monitor (i.e., a monitor with a fixed position) is a type of ambient air
September 2019
2-2
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
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 (Lipfcrt and Wvzga. 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).
Low spatial correlation with negative bias of the effect estimate in a long-term study indicates
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 et al.. 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 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 etal.. 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. Bergson 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 et al.. 2011;
Reeves et al.. 1998). Berkson exposure measurement error may occur when the time series of ambient
September 2019
2-3
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
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 etal. (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 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
September 2019
2-4
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
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 ambient 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 2-1). Strengths include 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 cited in Table 2-1 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 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 to estimate long-term
exposure tended to be within 6 ppb, 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 over time (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 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 area monitoring. Variations include drawing air past a nitrite coated glass tube
for an integrated ozone sample or using UV photometry for continuous ozone monitoring. Studies in the
2013 Ozone ISA reported the MDL for the nitrite coated glass tube to be 10 ppb-hour, while studies
September 2019
2-5
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
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. Other biases and uncertainties in
passive badges and portable active monitors were not reported.
Table 2-8 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 samplers. In the case of the low-cost continuous
monitors, reduction of bias was dependent on the way 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 et al. (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 precision, uncertainty is reduced because these concentrations are well above the MDL reported
in the 2013 Ozone ISA (U.S. EPA. 2013V
2.3.2 Modeling
The 2013 Ozone ISA (U.S. EPA. 2013) reviewed three types of models: models that estimate
ambient ozone concentration, microenvironmental models, and air exchange models. Concentration
modeling estimates ambient 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 demographic, time-activity pattern, building characteristic,
and air quality data as inputs for stochastic exposure simulations. Air exchange models estimate air
exchange rates for buildings based on building characteristics and meteorological variables. Air exchange
models describe airflow and are not specific to ozone. These are used as inputs to microenvironmental
models and so are not discussed in this 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
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.
September 2019
2-6
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Recent spatial interpolation studies examine the strengths and limitations of three approaches:
data averaging, IDW, and kriging (Table 2-9). 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 might not be needed. However, these methods could lead to incorrect
model fitting if 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 noninterpolation 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 function to estimate the ambient 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 location. Misspecification of the covariance 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 error 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 dearth of
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
September 2019
2-7
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
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 If = 0.34 for
a model of 8-hour daytime ozone (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 avg 8-hour daily ozone (9:00 a.m.-5:00 p.m.) (Adam-Poupart et al..
2014). 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 avg 8-hour daily 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 autocorrelation (Wang et al.. 2015). Spatiotemporal models
were applied in several recent studies to improve predictions of ozone concentrations for exposure
assessment studies (Table 2-10). Several studies used BME approaches, although the prior information
varied across studies and included kriging, LUR, and an autoregressive 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). An important advantage of BME is that it incorporates
multiple sources of data into the model, allowing for minimization of errors (Adam-Poupart et al.. 2014:
Warren etal.. 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, spatially clustered monitors in the study domain tend to produce more accurate model results.
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 in most locations (except near roads) (Wang et al.. 2016: Xu et al.. 2016a:
Wang et al.. 2015). Like BME models, PLS approaches require sufficient input data to produce accurate
models.
September 2019
2-8
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
2.3.2.3
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 output to a
discretized grid. However, CTMs are limited by their grid cell resolution, may be resource intensive to
run, and contain no time-activity information for possible exposure assignment. The Community
Multiscale Air Quality (CMAQ) model (EPA. 201S) 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 (U.S. EPA.
2013) (Table 2-11). A few studies directly compared different CTMs. Bond 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 horizontal and vertical advection mechanism and
deposition. Both models mostly overpredicted ozone concentrations in January, most likely because
neither model accounted for vertical mixing. In July, overprediction occurred for CAMx while
underprediction occurred for CMAQ when compared to monitor observations, 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).
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
changed with the size of a given grid cell (Yu et al.. 2016; Schaap et al.. 2015; Thompson and Selin.
September 2019
2-9
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
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 in (Yu et al.. 2016). Schaap et al. (2015) compared the
Multiscale Chemistry-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 average ozone 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 fire 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 AOD >0.5, the
model had 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 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 due to uncertainties in the cloud fraction along with other meteorological
variables. Yahva et al. (2016) found that for a 10-year avg of certain cloud variables, ozone
concentrations were generally underpredicted for most regions of the U.S.
September 2019
2-10
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
CTMs have been shown to underestimate high concentrations and overestimate low
concentrations, which could impact estimates of peak exposure conditions. In Tsimpidi et al. (2012) the
normalized mean bias was slightly negative at a 4-km grid resolution in CMAQ when concentrations less
than 40 ppb were excluded from the bias calculations (-7.9%) for the Pacific northwest in July, 2006.
However, when all concentrations were included, the statistic became large and positive in magnitude
(42.7%). Tsimpidi etal. (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 concentration 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 etal. (2015) observed a
positive mean bias in the early morning hours that decreased through the 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 develop ambient concentration
estimates. Several hybrid models combine observed data from fixed-site regulatory monitors with CTMs
that are defined over a spatiotemporal grid (Table 2-12). 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 predicts ambient ozone
concentration at the centroid of the grid. In the "downscaler" method, Berrocal et al. (2012) adjusted
CTM data at any point in the domain based on a weighted average of the CTM predictions for
surrounding grids such that exposure estimates were predicted at spatial scales finer than the input CTM.
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 (MSE: 124 ppb2, MAE: 8.7 ppb) or the
CTM data alone (MSE: 136 ppb2, MAE: 9.1 ppb) when predicting ozone over the eastern CONUS in
summer, 2001 with CMAQ data. Its predictive power was more pronounced in areas far from monitoring
locations.
Other studies found similarly improved performance with use of the BME. Xu et al. (2016b) used
a BME approach to merge ambient ozone concentration data from the Air Quality System (AQS)
database with CAMx simulations modeled at a 36- * 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) estimated validation error as the RMSE between the predicted and
observed data and 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- x 36-km grid
September 2019
2-11
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
was used in the CAMx model compared with a 12- * 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 Multi-scale coupled with Model of Air
quality and Chemistry (GEM-MACH, 2012) with surface observations from the 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 mean
squared error 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 ozone column 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 observations from Terra and Aqua satellites. Incorporating
observed and satellite data improved the correlation between surface observed data when compared to
CMAQ data 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.
2.3.2.5 Microenvironmental Modeling
The 2013 Ozone ISA (U.S. EPA. 2013) presented several studies that evaluated integrated
microenvironmental exposure (ME) and dose models. ME models apply stochastic sampling of
distributions of data for air quality, time-activity patterns, demographic, physiological, and building
ventilation variables to predict population exposures in different locations. ME models predict
microenvironmental concentrations, exposures, and doses. Advantages identified in the 2013 Ozone ISA
September 2019
2-12
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
include ability of the user to design analyses for specific populations (assuming that demographic and
time-activity data are available) and to include of indoor air sources (which are uncommon for ozone).
Limitations include resource intensiveness of the ME models and that indoor exposures cannot be easily
validated (Georgopoulos et al.. 2005).
Strengths and limitations identified in recent studies (Table 2-13) 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 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 by personal
monitors, so the extent of this error was unknown.
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 concentration. 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. and 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).
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.-12:59 p.m.)
September 2019
2-13
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
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. 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 for older adults. A
separate 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
(Isaacs. 2014). Children and teens ages 4-18 years had the largest participation rate among 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, Isaacs (2014) calculated that
79% of time spent by children ages 4-18 years and 63% of time spent by adults ages 19-95 years
involved at least moderate exertion. When comparing time-activity data by race from the CHAD database
(Table 2-2). Hispanic study participants spent slightly more time indoors at home than the total
population, while white study participants spent the most time outdoors compared with Asian, black, and
Hispanic participants (Isaacs. 2014). However, 11% of participants had missing race/ethnicity data or
refused to provide information regarding race/ethnicity, so these results should be interpreted cautiously.
Males spent more time outdoors than females (Table 2-3). These studies 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 is useful because it provides a detailed picture of time-activity across
population groups, and it has a large number of days of data. 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 may have differed across studies input to CHAD, 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 breaking down the data by age, race, 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
September 2019
2-14
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
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 median positional 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 misclassification of time-activity patterns.
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 zero.
September 2019
2-15
DRAFT: Do Not Cite or Quote
-------
1 Updated time-activity data and tools for assessing time-activity data have improved the general
2 understanding of time-activity data and related uncertainties in recent years. Analysis of CHAD diaries
3 indicated that young children ages 0-5 years were found to spend less time outdoors than older children,
4 teens, and adults, and white respondents spent more time outdoors than their Asian, black, and Hispanic
5 counterparts (Isaacs. 2014). New technologies to assess study participant location, errors related to study
6 participant recall, and residential mobility have been used to determine that location-based errors are
7 within 6% for short- and long-term exposure assessment, while omission of residential mobility can
8 produce bias in the exposure estimate, resulting in negatively biasing the effect estimate for a study of
9 long-term ozone exposure.
September 2019
2-16
DRAFT: Do Not Cite or Quote
-------
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
2,253
2,010
1,080
5,785
1,403
(%)
(100)
(18)
(16)
(8.5)
(46)
(11)
Warm Season, 12:00 a.m.-
-11:59 p.m. (person-minutes [%])
Indoor-residential
31,038,736
5,932,419
4,474,880
1,846,578
13,593,134
5,012,405
(75)
(81)
(74)
(71)
(71)
(83)
Transit
2,359,073
284,770
242,706
134,854
1,352,166
329,628
(5.7)
(3.9)
(4.0)
(5.2)
(7.1)
(5.4)
Indoor-work/school/other
5,956,425
725,350
906,573
429,633
3,342,988
504,961
(14)
(9.9)
(15)
(17)
(18)
(8.3)
Outdoor
1,562,018
214,340
255,882
130,973
793,278
162,906
(3.8)
(2.9)
(4.2)
(5.1)
(4.1)
(2.7)
Uncertain or missing
521,188
163,109
166,890
45,274
74,463
48,180
(1.3)
(2.2)
(2.8)
(1.7)
(0.39)
(0.80)
Warm Season, 12:00 p.m.-8:00 p.m. (person-minutes [%])
Indoor-residential
9,234,040
1,867,690
1,327,384
529,438
3,853,625
1,601,243
(61)
(70)
(60)
(56)
(55)
(73)
Transit
1,407,828
184,339
157,601
81,709
781,252
194,099
(9.3)
(6.9)
(7.1)
(8.6)
(11)
(8.8)
Indoor-work/school/other
3,240,916
412,785
470,148
225,926
1,819,323
288,478
(21)
(15)
(21)
(24)
(26)
(13)
September 2019
2-17
DRAFT: Do Not Cite or Quote
-------
Table 2-1 (Continued): 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
Outdoor
1,015,749
154,838
198,851
92,928
476,679
89,982
(6.7)
(5.8)
(9.0)
(10)
(6.8)
(4.1)
Uncertain or missing
190,037
53,793
56,847
14,958
37,363
26,851
(1.3)
(2.0)
(2.6)
(1.6)
(0.50)
(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. Data were downloaded from: https://www.epa.gov/healthresearch/consolidated-human-activitv-
database-chad-use-human-exposure-and-health-studies-and.
1
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
individuals)
12,673
248
1,829
729
8,083
310
(%)
(100)
(2.0)
(14)
(5.8)
(64)
(2.4)
Warm Season, 12:00 a.m.-
11:59 p.m. (person-minutes [%])
Indoor-residential
31,038,736
693,831
4,026,861
2,278,661
20,590,280
968,084
(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)
September 2019
2-18
DRAFT: Do Not Cite or Quote
-------
Table 2-2 (Continued): 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
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 missing
521,188
12,810
69,533
23,721
357,941
13,674
(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-residential
9,234,040
201,679
1,166,578
693,833
6,157,342
297,935
(61)
(59)
(61)
(65)
(61)
(65)
Transit
1,407,828
28,237
188,304
93,949
935,729
41,811
(9.3)
(8.3)
(10)
(8.8)
(9.3)
(9.1)
Indoor-
3,240,916
90,168
413,189
208,599
2,151,117
90,752
work/school/other
(21)
(27)
(22)
(20)
(21)
(20)
Outdoor
1,015,749
15,146
113,455
62,517
727,574
25,541
(6.7)
(4.5)
(5.9)
(5.9)
(7.2)
(5.5)
Uncertain or missing
190,037
4,427
38,950
9,393
114,541
5,295
(1.3)
(1.3)
(2.0)
(0.9)
(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. Data were downloaded from: https://www.epa.gov/healthresearch/consolidated-human-activitv-
database-chad-use-human-exposure-and-health-studies-and.
September 2019
2-19
DRAFT: Do Not Cite or Quote
-------
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
236,982
261,327
(1.3)
(1.1)
(1.4)
Warm Season, 12:00 p.m.-8:00 p.m. (person-minutes [%])
Indoor-residential
9,234,040
5,069,614
4,162,147
(61)
(63)
(59)
Transit
1,407,828
752,611
654,873
(9.3)
(9.4)
(9.3)
Indoor-work/school/other
3,240,916
1,670,920
1,569,613
(21)
(21)
(22)
Outdoor
1,015,749
441,741
573,859
(6.7)
(5.5)
(8.1)
Uncertain or missing
190,037
96,730
93,307
(1.3)
(1.2)
(1.3)
September 2019
2-20
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
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
household air purifiers, I/O ratios generally ranged from 0.1-0.4. Higher ratios were observed during the
warm season when ambient ozone concentrations are highest.
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 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 etal. (2014) also examined ozone infiltration in
vehicles, and the mean and range of I/O ratios were between the values for I/O ratio obtained for open
versus closed windows or doors.
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.
September 2019
2-21
DRAFT: Do Not Cite or Quote
-------
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 etal. (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
September 2019
2-22
DRAFT: Do Not Cite or Quote
-------
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
Chen etal. (2012)
(cont.)
Miami, FL
1987-2000
All
residents
(cont)
Home (estimated
by model) (cont.)
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
September 2019
2-23
DRAFT: Do Not Cite or Quote
-------
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
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
September 2019
2-24
DRAFT: Do Not Cite or Quote
-------
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 etal. (2014)
San Francisco
September
Store Grocery stores
Average, 1-h
Mean (range): 0.40
NR
Across stores:
Bay Area,
2011-March
occupants
daily max
(0.18-0.59)
0.65-1.47/h
Sacramento Area,
2013
across stores:
Fresno, Los
24.1-66.7 ppb,
Angeles Area, CA
30.0-79.4 ppb
Furniture/hardware
Average, 1-h
Mean (range): 0.42
NR
Across stores:
stores
daily max
(0.29-0.47)
0.39-2.38/h
across stores:
30.1-62.1 ppb,
30.1-62.1 ppb
Apparel stores
Average, 1-h
daily max
across stores:
12.1-51.5 ppb,
15.1-59.6 ppb
Mean (range): 0.33
(0.11-0.47)
NR
Across stores:
0.52-2.33/h
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)
Gall etal. (2011) Houston. TX
Simulation
Simulated
homes
No passive removal
materials
NR
Average: 0.16
NR
0.5/h
Simulated
homes
Gypsum, activated
carbon cloth, or
other removal
materials
NR
Average: 0.047-0.12
NR
0.5/h
September 2019
2-25
DRAFT: Do Not Cite or Quote
-------
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
Nq et al. (2015)
Atlanta, GA
NR
Simulated ASHRAE
box store prescribed
ventilation
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
Volume-weighted
concentrations
Average of
daily average
(daily peak):
76 ppb
(93 ppb)
Calculated as mean
indoor/daily average
outdoor: daily
average (daily peak):
0.079 (0.075)
NR
1.2 L/s-m2
0.4 L/s-m2
Lai et al. (2015)
West Lafayette,
IN
NR
Test Infiltration
chamber
24.02-
53.5 ppb
0.050-0.099
NR
Simple mechanical 38.96-39.69
ventilation ppb
0.57-0.63
NR
HVAC
Window open
Fagade natural
ventilation
20.11-34.49
ppb
0.15-0.43
NR
30.85-51.02
ppb
0.23-0.42
NR
22.09-27.68
ppb
0.18-0.33
NR
Median
(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)
Median
(10th—90th
percentile) 3.67
(0.74-7.70)
September 2019
2-26
DRAFT: Do Not Cite or Quote
-------
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
Mechanical mixed
with added outdoor
air when
thermodynamically
favorable
Range of
means: 17 ppb
(Seattle, WA)-
35 ppb
(Albuquerque,
NM)
NR
NR
Natural ventilation:
ASHRAE 62.1
Range of
means: 17 ppb
(Seattle, WA)-
35 ppb
(Albuquerque,
NM)
Mean (5th—95th
percentile): 0.107
(0.0926, 0.128)
NR
Natural ventilation
with added outdoor
air when
thermodynamically
favorable
Range of
means: 17 ppb
(Seattle, WA)-
35 ppb
(Albuquerque,
NM)
NR
0.39 (for all
locations)
0.40 (Miami,
FL)—1.4 (Los
Angeles, CA)
0.33 (Miami,
FL)-0.39 (Los
Angeles, CA
and Seattle,
WA)
0.49 (Seattle,
WA and Boston,
MA)-1.6 (Los
Angeles, CA)
September 2019
2-27
DRAFT: Do Not Cite or Quote
-------
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
1-h daily max:
average per
ventilation
condition:
44-72 ppb
8-h daily max
average 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
average:
17.5 ppb
(Miami, FL)-
34.0 ppb
(Albuquerque,
NM)
0.18-0.49
NR
Variable air volume Hourly
ventilation average:
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 = indoor ozone
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 rather than the I/O provided in the table per 10 ppb of ozone.
September 2019
2-28
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
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). 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 average P/A of 0.48 with a 95th percentile of 0.83 and correlation of 0.98.
During the Moderate and Severe Asthmatics and Their Environment Study (MASAES), Williams et al.
(2012) observed no relationship between ozone exposure and personal activities, 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.
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.
September 2019
2-29
DRAFT: Do Not Cite or Quote
-------
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 Mean (SD): 3.4 ppb Mean (SD): 0.0665 (slope) P/A: 0.27
2008—April, moderate and (3.6 ppb) 29.7 ppb
2009 severe asthmatics (15.0 ppb)
and their
environment
study panel
Chen etal. (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
September 2019
2-30
DRAFT: Do Not Cite or Quote
-------
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 etal. (2012) (cont.)
Denver, CO 1987-2000
(cont.)
Los Angeles, CA
Miami, FL
Nashville, TN
New York City,
NY
Phoenix, AZ
Seattle, WA
St. Louis, MO
Washington, DC
Worcester, MA
All residents
(cont.)
NR (cont.;
NR (cont.;
0.27
0.25
0.26
0.27
0.30
0.25
0.30
0.29
0.27
0.27
NR (cont.;
Jones et al. (2013)
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
12.78 ppb
(0.83)
diagnoses
(20.78 ppb)
NR = not reported, NMMAPS = National Morbidity, Mortality, and Air Pollution Study; P/A = fo +fipa/(a+k) where P = personal exposure to ambient ozone, A = ambient ozone
concentration, fo = fraction of time spent outdoors, fi = 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. (2012) by inputting the data provided in Table 1 into the equation for P/A presented in the 2013 Ozone ISA (U.S. EPA. 2013).
September 2019
2-31
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
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 (Billionnet et al.. 2012). Potential confounding is
limited to copollutants in this section, because noise is 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 et al.. 2000). Winquist et al. (2014) compared joint effects
calculated with single-pollutant models 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.
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. 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.
September 2019
2-32
DRAFT: Do Not Cite or Quote
-------
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 on the form of the standards were used for aii pollutants (ozone: 8-h daily max, CO: 8-h daily max, N02:
1-h daily max, PM2.5: 24-h avg, PMi0: 24-h avg, S02: 1-h 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 AQS
2015-2017.
September 2019
2-33
DRAFT: Do Not Cite or Quote
-------
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-h daily max, CO: 8-h daily max, N02:
1-h daily max, PM2.5: 24-h avg, PM10: 24-h avg, S02: 1-h 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 2-2. Seasonal Pearson correlations of 8-hour daily max ozone
concentrations with copollutant concentrations measured in AQS,
2015-2017.
2.6 Interpreting Exposure Measurement Error for Use in
Epidemiology Studies
1 As described in the 2013 Ozone ISA (U.S. EPA. 2013). exposure measurement error, which
2 refers to the biases and uncertainties associated with using concentration metrics as surrogates for the
3 actual exposure of an individual or population (Section 2.2. Exposure Concepts), can be an important
4 contributor to error in epidemiologic study results. Short-term exposure studies include time-series
5 studies, case-crossover studies, and panel studies. Time-series studies generally assess the association of
6 daily health status of a population of thousands or millions of people over the course of multiple years
7 (i.e.. thousands of days) across an urban area with estimates of human exposure using a short monitoring
September 2019
2-34
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
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
Zcgcr. 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)).
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)). 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). 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).
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..
2017). 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).
September 2019
2-35
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
2.6.1.1 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 is 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
(Zegeret al.. 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 ambient monitoring is used to represent exposure in the statistical model. Differences
between personal exposure to ambient ozone and Ca due to unaccounted 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. 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 (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
inaccuracy when temporal variability in the concentration at the location of 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 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 a distance of 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
September 2019
2-36
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
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 spatio-temporal 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 spatio-temporal 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.
Geostatistical-simulated concentrations 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
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
September 2019
2-37
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
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 etal.. 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
more pronounced for the fixed-site ambient monitoring data. 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) recently compared effects in four cities (Athens, Greece;
London, UK; 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 city, the percentage change in total mortality corresponding to a 10 (ig/m3 increase in ozone
September 2019
2-38
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
concentration decreased from positive to negative (of equal or lesser magnitude) with larger confidence
intervals when the one-in-six 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 2-6). 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 year group, which spends the most time indoors (Table 2-1). was the only group for
which effect was underestimated by the fixed-site monitor.
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 of the effects of short-term exposure to
ozone typically use active or passive microenvironmental monitors to represent exposure (Table 2-6). 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 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.
September 2019
2-39
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
2.6.2
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. 2013). 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 2-6). 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 with high spatial resolution ambient ozone concentration predictions to
estimate ambient ozone exposures among the population.
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
September 2019
2-40
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
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 2-6). 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(SD) = -0.12 ppb
(0.093 ppb)] of the total bias in the effect estimate, and biases were negative. The standard deviation 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.
September 2019
2-41
DRAFT: Do Not Cite or Quote
-------
Table 2-6. 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 A FRM or FEM
TSection 2.3.1.1
Table 2-7; U.S. EPA
(2013)1
monitor located at a
fixed location to
measure ambient
ozone concentration
by
chemiluminescence population within a
Short-term
exposure studies:
surrogate for
ambient ozone
exposure
concentration of a
Ambient ozone
concentration
measurements
undergo rigorous
quality assurance
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
September 2019
2-42
DRAFT: Do Not Cite or Quote
-------
Table 2-6. (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
[Section 2.3.1.1
Table 2-7; U.S. EPA
(2013)1 (cont.)
A FRM or FEM
monitor located at a
fixed location to
measure ambient
ozone concentration
by
chemiluminescence
or UV absorption
(cont.)
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
(cont.)
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 (cont.)
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 2-8)
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 Instrument errors more
absorption typically lead to positive
instruments subject artifacts from interferences
to interference from
humidity, mercury,
and VOCs
Instrument errors are
typically small but
positive and so have
the potential to add
negative bias to the
effect estimate
September 2019
2-43
DRAFT: Do Not Cite or Quote
-------
Table 2-6. (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
Passive personal
exposure monitor
(Section 2.3.1.2
Table 2-8)
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
Modeling Methods
Data averaging
(Section 2.3.2.1
Table 2-9)
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
Ambient ozone
concentration
measurements
undergo rigorous
quality assurance;
averaging scheme
designed for
population or trend
of interest; simple to
implement
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
data
Negative bias and
decreased precision in
the effect estimate
September 2019
2-44
DRAFT: Do Not Cite or Quote
-------
Table 2-6. (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
Data averaging
(Section 2.3.2.1
Table 2-9) (cont.;
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 (cont.)
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
Localized ozone loss
processes near roads are
not captured
Potential for negative
bias in the effect
estimate
Assumption of constant
ozone concentration within
some geographic area
Potential for decreased
precision in the effect
estimate
Omission of time-activity
data
Negative bias and
decreased precision in
the effect estimate
September 2019
2-45
DRAFT: Do Not Cite or Quote
-------
Table 2-6. (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
Inverse-distance
weighting
(Section 2.3.2.1
Table 2-9)
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
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
Kriging
(Section 2.3.2.1
Table 2-9)
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
September 2019
2-46
DRAFT: Do Not Cite or Quote
-------
Table 2-6. (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
Land use regression
(Section 2.3.2.2
Table 2-10)
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
misspecifi cation
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
September 2019
2-47
DRAFT: Do Not Cite or Quote
-------
Table 2-6. (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 2-10)
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
misspecifi cation
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
September 2019
2-48
DRAFT: Do Not Cite or Quote
-------
Table 2-6. (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
Chemical transport
model
(Section 2.3.2.3
Table 2-11)
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
September 2019
2-49
DRAFT: Do Not Cite or Quote
-------
Table 2-6. (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
Hybrid approaches
(Section 2.3.2.4
Table 2-12)
Grid-based ambient
ozone
concentrations are
estimated from
precursor emissions,
meteorology, and
atmospheric
chemistry and
physics and bias
corrected 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
data 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
September 2019
2-50
DRAFT: Do Not Cite or Quote
-------
Table 2-6. (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
Microenvironmental Estimates
modeling
(e.g., APEX,
SHEDS
TSection 2.3.2.5
Table 2-131)
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,
indooroutdoor 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 = Federal Equivalent Method; FRM = Federal Reference Method;
IDW = inverse-distance weighting; SHEDS = stochastic human exposure and dose simulation.
September 2019
2-51
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
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 IYU.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 near NOx 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. Use of CTMs has
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, 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 2-6). 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
September 2019
2-52
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
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 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 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 2-6). 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. 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.
September 2019
2-53
DRAFT: Do Not Cite or Quote
-------
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 2-7 through
Table 2-13:
Unpaired predicted-to-observed
peak ozone ratio (AUP)
Mean bias (MB)
Mean error (ME)
Mean-squared error (MSE)
Root-mean-squared error (RMSE)
Fractional bias (FB)
Fractional error (FE)
Gross error (GE)
Mean normalized bias (MNB)
(-100% to +oo)
Mean normalized error (MNE)
(0% to +co)
Normalized mean bias (NMB)
(-100% to +oo)
Normalized mean error (NME)
(0% to +oo)
Mean fractional bias (MFB)
(-200 to +200%)
p — n
ri,peak wi,peak
p
1 i,peak
N
- "¦>
i= 1
N
i=1
1
»Zjp' -o,)
1 VnJV
Pj-Oj
Pi + Oi
Pj-Oj
Pj + Oi
i= 1
N
!v |Pi ~ °i
JvZjI 0;
i=1
Z?=i(Pj ~ Oj)
ZUOi
Zf=il Pj-Qj\
yN o¦
Zji=l ui
2 y (pi ~ OA
NZ^KPi + oJ
1=1
September 2019
2-54
DRAFT: Do Not Cite or Quote
-------
Mean fractional error (MFE) 2 v"11P- — 0
(o to +200%) ^2, I^Toi
i=1
Coefficient of determination (R2) _Q^p. -p)}2
lUiOt-ofll^Pt-P)2
Mean absolute error (MAE) Eililf; ~ 0;
N
Index of agreement (IOA) Z"=i(^£ _ °;)2
IZ=i(\Pi-o\ + \ot-o\y
NB, FB, FE, U R, NB, NE, NGE,
AUP, CSI, FAR, UPA, MNGE,
Pi and Oi are prediction and observation at the ith monitoring site,
respectively; N is the number of monitoring sites.
September 2019
2-55
DRAFT: Do Not Cite or Quote
-------
Table 2-7 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 fixed-
Long-term
Mean summer
Observed data
Spatial
NR
(2011)
fixed-site
the Atlanta
site ambient
exposure
quartiles of peak
collected
variability is
ambient
metropolitan
monitors only,
8-h ozone in ppb,
directly;
limited to
monitors only
area;
protocols of
44.85, 55.73,
speciation
monitoring
from SEARCH,
T: between the
observed data
64.93, 80.23
collected;
locations
U.S. EPA
years 1999
were referenced
several years
PAMS, U.S.
and 2007;
in previous
of data
EPA STN,
P: entire
publications
collected
IMPROVE
population
monitoring
considered
networks
Hackbarth et al.
Daily 8-h max
L: California;
Observed data
Short-term
Ozone daily
Observed data
Sensitivity
Ozone daily
(2011)
ozone within
T: 2005-2007;
used in inverse-
exposure
average in CA
used directed
analysis of
average across
20 miles of the
P: those with
distance
between 2005
with ER data
exposure
California =
zip code
ER visits for
weighting
and 2007
method not
39.9 ppb,
centra id
cardio, resp,
validated from
because
explored
standard
weighted by
asthma
U.S. EPA
37.1-73.8 ppb
(e.g., buffer
error = 0.225
inverse
size of nearest
ppb
distance from neighbor)
U.S. EPA's
fixed-site
monitors
September 2019
2-56
DRAFT: Do Not Cite or Quote
-------
Table 2-7 (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. (2013) Nearest
L: Houston and
Comparison
Short-term
Houston =
Simple to
Overfitting may
N = 10
monitor
Los Angeles;
with 10-20
exposure
31.0-112.5 ppb;
implement
be caused by
RMSE =.
T: select days
monitors in each
Los Angeles =
just one
54-21.01 ppb,
in 2009-2011
metropolitan
10.8-117.1 ppb
incorrect
n = 20 RMSE =
area
parameter;
8.41-19.06 ppb
does not
capture the
underlying
phenomena
Dionisio et al. Ambient
L: Atlanta, GA;
Comparison
Long-term
NR
Less spatial
Uncertainty in
Mean (SD)
(2014) monitor
T: 1999-2002;
with dispersion
exposure
variability in
areas where
exposure
P: Entire
model
ozone, so
there are
measurement
population
fixed-site
known sinks
error for
monitors do a
(e.g., near
omission of
better job than
roads)
spatial
for spatially
variability:
variable
-0.055 (0.037);
pollutants
bias on effect
estimate for
omission of
spatial
variability:
-0.12 (0.093)
September 2019
2-57
DRAFT: Do Not Cite or Quote
-------
Table 2-7 (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 etal. (2013)
Comparison
L: Houston,
Comparison
Short-term
8-h daily max =
Average and
Positive bias of
Differences
between FRM
TX;
between
exposure
average 22 ppb,
maximum data
FRM due to
between
(Thermo-
T: August 26-
ambient monitor
maximum 94 ppb,
compare well;
water vapor,
monitors
scientific 49C)
November 19,
types using air
19 values and
frequent
gas-phase
presented
and FEM
2010;
spiked with
6 days with 8-h
calibrations
mercury, and
graphically but
ambient
P: entire
known quantity
daily max above
and zero/span
VOCs
not reported
monitors (two
population
of ozone
75 ppb
improved data
models:
quality
Teledyne
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 value
Average
Potentially high
In the vicinity of
(2014)
monitor
NC;
with an FRM
exposure
= 33.0-55.2 ppb
values are only
positive errors
VOC sources,
T: September,
and a
slightly higher
in vicinity of
ozone
2012;
microenviron-
than FRM
VOC sources
concentration
P: Entire
mental model
from the FEM
population
was several
hundred
percentage
higher
(depending on
the source)
than a
microenviron-
mental monitor
that compared
well with the
FRM
September 2019
2-58
DRAFT: Do Not Cite or Quote
-------
Table 2-7 (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. (2017)
Fixed-site
L: Montreal,
Comparison
Short-term
8-h daily max;
Most accurate
Lacks spatial
ICC mean
monitors
Quebec,
with BME, IDW,
exposure
mean (SD) =
measure of
resolution
(95% CI) vs.
reporting data
Canada;
and back-
27.9 ppb
ozone
IDW 0.89 (0.89,
for 8-h daily
T: January 1,
extrapolation
(15.2 ppb) median
0.89), vs. LUR
max
1991—
LUR
= 26.3 ppb
w/back-
December 31,
extrapolation
2002;
0.67 (0.47,
P: Entire
0.78), vs. BME
population
0.64 (0.41,
0.77)
September 2019
2-59
DRAFT: Do Not Cite or Quote
-------
Table 2-7 (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, GA;
Comparison
Short-term
NR
Accurate
No spatial
Urban site:
monitor
T: 2011;
with fixed-site
exposure
capture of
resolution,
MB = -2.54
reporting data
P: Entire
monitors
temporal
autocorrelation
ppb,
for 8-h daily
population
variation
introduces bias
ME = 4.15 ppb,
max
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; U.S. EPA = Environmental Protection Agency;
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.
September 2019
2-60
DRAFT: Do Not Cite or Quote
-------
Table 2-8 Studies informing assessment of exposure measurement error when concentrations measured by
personal and microenvironmental monitors are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, 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
Bartet 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 real-
Canada;
monitors
time
T: May-
measurements
September,
2012;
P: Entire
population
Zimmerman et al.
Low cost
L: Pittsburgh,
Deployed a
Short-term
15-min avg
Allows for
Sensitivity to
Multiple linear
(2018)
sensor (Real-
PA;
dense network of
exposure
time; NR
better spatial
model quality
regression MAE avg
Time Affordable
T: August 3,
low cost
coverage
and input data
(SD) = 5.1 ppb
Multipollutant
2016-
samplers then
quality
(0.6 ppb) R = 0.81
sensors) data
February 7,
applied one of
Random forest MAE
filtered through
2017;
two models
average (SD) =
model to
P: Carnegie
(random forest
0.7 ppb (0.1 ppb)
improve data
Mellon
or multiple linear
R = 0.99
quality based
University
regression) to
on data across
campus
smooth data for
geographical population ozone
area
September 2019
2-61
DRAFT: Do Not Cite or Quote
-------
Table 2-8 (Continued): Studies informing assessment of exposure measurement error when concentrations
measured by personal and microenvironmental 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
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,
FEM;
65 ppb; indoor
compared
intercomparison;
slope = 0.82
at 254 nm
2014;
intercomparison
test range
with FEM
correlations
outdoor:
wavelength (2B
P: Panel of
of personal
(1-min avg time)
between
R = 0.991,
Technologies)
volunteers
monitors
= 30-55 ppb;
personal
slope = 1.08
chamber test
monitors
indoor;
range (1-min
dropped when
comparison with
avg time) =
VOCs were
FEM: outdoor
85-125 ppb
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; P = population; Pearson R = correlation coefficient; SD = standard deviation;
SE = standard error; T = time; VOC = volatile organic compound.
September 2019
2-62
DRAFT: Do Not Cite or Quote
-------
Table 2-9 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
Data Averaging
Joseph et
Simple
L: Houston
Comparison
Short-term Houston =
Simple to implement
Overfitting may be
Houston: n = 10
al. (2013)
averaging
and Los
with 10-20
exposure 31.0-112.5 ppb;
caused by just one
RMSE =
Angeles;
monitors in
Los Angeles =
incorrect
11.30-15.35 ppb,
T: Select days
each
10.8-117.1 ppb
parameter; does
n =20
in 2009-2011
metropolitan
not capture the
RMSE =
area
underlying
10.77-15.07 ppb;
phenomena
Los Angeles:
n = 10
RMSE =
15.16-25.13 ppb,
n =20
RMSE =
12.96-24.35
September 2019
2-63
DRAFT: Do Not Cite or Quote
-------
Table 2-9 (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 city
P: Entire
RMSE = 5.73 ppb,
reporting data
population
MNB = -1%,
for 8-h daily
MNE = 10%,
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 IDWofdata L: Montreal.
Comparison
Short-term
8-h daily max;
Low spatial
Quality of model
ICC mean
al. (2017) from fixed-site Quebec.
with BME,
exposure
mean (SD) =
variability of ozone
depends on spatial
(95% CI) vs. fixed-
monitors Canada;
back-
28.1 ppb
may negate
density of monitors
site monitor = 0.89
T: January 1,
extrapolation
(13.0 ppb) median
limitation
(0.89, 0.89), vs.
1991-
LUR, and
= 26.5 ppb
LUR w/back-
December 31,
fixed-site
extrapolation = 0.6
2002;
monitors
2 (0.59, 0.64), vs.
P: Entire
BME = 0.76 (0.72,
population
0.78)
September 2019
2-64
DRAFT: Do Not Cite or Quote
-------
Table 2-9 (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.
IDW of data
L: Atlanta,
Comparison
Short-term
NR
Better spatial
Quality of model
Urban site:
(2018)
from 10 fixed-
GA;
with fixed-site
exposure
resolution than
depends on spatial
MB = -1.27 ppb,
site monitors
T: 2011;
monitors
monitor-based
density of the
ME = 2.94 ppb,
P: Entire
approaches
monitors
RMSE = 4.31 ppb,
population
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
September 2019
2-65
DRAFT: Do Not Cite or Quote
-------
Table 2-9 (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 2009-
metropolitan
not capture the
n = 20
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
Long-term
NR
As a traditional
Assumes linearity
Kriging Model 1
(2011)
kriging
and
of model
exposure
method, this is
or some simplified
Daily RMSE = 8.58-
midwestern
points with
better established
function between
22.67 ppb; Kriging
U.S.;
concentratio
sampling points
Model 2 Daily RMSE
T: May 15-
ns from 375
= 8.65-21.11 ppb
September
monitors
11, 1995
reporting to
10:00-17:00;
AQS
P: Entire
population
September 2019
2-66
DRAFT: Do Not Cite or Quote
-------
Table 2-9 (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
Ordinary and
L: Houston
Comparison
Short-term
Houston =
Yields superior
Overfitting may be
Houston:
al. (2013)
universal
and Los
with 10-20
exposure
31.0-112.5 ppb
validation compared
caused by just one
OK
kriging
Angeles;
monitors in
Los Angeles =
to other methods
incorrect
n = 10 valid pts
T: Select days
each
10.8-117.1 ppb
parameter
RMSE =
in 2009-2011
metropolitan
area
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
September 2019
2-67
DRAFT: Do Not Cite or Quote
-------
Table 2-9 (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
cities; with monitors exposure
T: 2012; reporting to the
P: Entire Texas Air
population Monitoring
Information
System
Mean (SD) for
select hours
3/25/2012
2:00 p.m. =
68.2 ppb
(6.15 ppb)
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 = 0.7
14 to 1.099 avg
std
error = 0.00527 to
0.0105
September 2019
2-68
DRAFT: Do Not Cite or Quote
-------
Table 2-9 (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
Gelfand et
Kriging
L: California;
Compare
Long-term
NR
Comparison of
Preferential
RMSE high
al. (2012)
(approach is
T: 2008;
kriged ozone
exposure
monitor selection
sampling causes
monitors = 22.7 pp
unspecified)
P: Entire
surface to
allows for evaluation
overestimation or
b,
population
monitors
of best practices
underestimation
low
reporting to
(i.e., use of
monitors = 23.9 pp
AQS; kriged
randomly selected
b, randomly
surface is fit to
monitors);
selected
high
otherwise, selection
monitors = 18.0
concentration
of high monitors
ppb,
monitors, low
causes
all monitors:
concentration
overestimation of
18.0 ppb
monitors,
concentrations and
randomly
vice versa
selected
monitors, or all
monitors
September 2019
2-69
DRAFT: Do Not Cite or Quote
-------
Table 2-9 (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.
September 2019
2-70
DRAFT: Do Not Cite or Quote
-------
Table 2-10 Studies informing assessment of exposure measurement error when concentrations modeled by
land use regression or spatiotemporal models are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Land Use Regression
Clark et al.
Land use
L: 100 U.S.
Observed data
Short-term
8-h daytime avg
Observed data
Observed data
Model LUR
(2011)
regression with
urban areas
used in model
exposure
during ozone
used
were sparse
R2 = 0.34
variables related
across the
was validated
season
to the built
U.S.;
arithmetic mean
environment,
T: May-
is 45 ppb
climate,
September
transportation,
1990;
and income
P: Entire
population
Adam-Poupart
Land use
L: Montreal,
Cross-validation
Short- and
8-h daily max;
More accurate
Output quality
R2 = 0.466,
et al. (2014)
regression mixed-
Quebec,
against NAPS
long-term
NR
than BME
depends on
RMSE =
effects model
Canada;
monitoring data
exposure
variation in some
quality of input
8.747 ppb,
incorporating
T: May-
from 2005
cases
data
percentage
temperature,
September
change in
precipitation, day
1990-2009;
MSE = -19.9%
of year, road
P: Entire
(compared with a
density, and
population
BME-LUR hybrid
latitude
model)
September 2019
2-71
DRAFT: Do Not Cite or Quote
-------
Table 2-10 (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
Measurement
Exposure
Period, 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
September 2019
2-72
DRAFT: Do Not Cite or Quote
-------
Table 2-10 (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 Measurement Exposure
Period, 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 region 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%
September 2019
2-73
DRAFT: Do Not Cite or Quote
-------
Table 2-10 (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 Measurement Exposure
Period, and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Wang et al.
(2016)
Spatiotemporal-
LUR incorporating
the UCD-CIT
chemical
transport model
with meteorology
modeled by WRF
v.3.1.1 on a 4-km
grid, ST-LUR
alone
L: Los
Angeles and
Riverside
Counties;
T: 2000-
2008;
P: Entire
population
10-fold cross-
validation
against
37 monitors for
each variation of
model
Short- and
long-term
exposure
8-h daily max;
range =
10-50 ppb (long-
term avg)
Improved
integration of
different models,
takes advantage
of spatial
residuals
Modeling
approach used
2-week data, did
not look at 8-h
daily max
ST-LUR:
RMSE = 5.64 ppb
R2 = 0.86;
ST-LUR + CTM:
RMSE = 4.65
ppb, R2 = 0.87
Wang et al.
(2015)
Spatiotemporal
L: Baltimore,
10-fold cross-
Short- and
2-week avg; NR Low spatial
Missing ozone
Overall cross-
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;
more applicable
New York
= 0.60,
T: 1993-
St. Paul =
0.91,
2013;
Winston-
P: MESA Air
Salem = 0
.66
study
Overall cross-
participants
validation
R2:
Baltimore = 0.89,
Chicago = 0.72,
Los
Angeles = 0.78,
New York = 0.61,
St. Paul = 0.90,
Winston-
Salem = 0.76
September 2019
2-74
DRAFT: Do Not Cite or Quote
-------
Table 2-10 (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
Measurement
Exposure
Period, 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,
MD;
T: February
11-22, 2012
and June
18-27,
2012;
P:
Participants
in the
MESA-Air
study
LOOCV, using
comparison with
measurements
obtained on a
mobile
monitoring
platform
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-year avg
RMSE = 4.21
ppb, MAE = 3.36
ppb
September 2019
2-75
DRAFT: Do Not Cite or Quote
-------
Table 2-10 (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 Measurement Exposure
Period, 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.
(2017)
Bayesian
L: Montreal
maximum entropy
Quebec,
model drawing
Canada;
output from a land
T: January
use regression as
1, 1991 —
its prior
December
31, 2002;
P: Entire
population
Comparison with
IDW, back-
extrapolation
LUR, and fixed-
site monitors
Short-term
exposure
8-h daily max;
mean (SD) =
30.0 ppb
(9.1 ppb)
median =
29.8 ppb
Captures spatial
variability more
completely
Higher
complexity
compared with
other models
ICC mean
(95% CI) vs.
fixed-site
monitor = 0.64
(0.41, 0.77), vs.
IDW = 0.76 (0.72,
0.78), vs. LUR
w/back-
extrapolation = 0.
37 (0.16, 0.52)
September 2019
2-76
DRAFT: Do Not Cite or Quote
-------
Table 2-10 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by land use regression or spatiotemporal models are used for exposure
surrogates.
Reference
Model
Location,
Time
Period, and
Population
Measurement
Evaluation
Technique
Epidemiology
Applications
Concentrations
Measured
Strengths
Limitations
Exposure
Measurement
Errors
Gonq et al.
HYSPLIT v4.9
L: Eight
Model compared
Short- and
8-h daily max
Multiple models
Limited cities
Model-obs
(2017)
with GDAS 1° x
western U.S.
(and built) with
long-term
ozone for
used together
explored in
comparison
1° data with a
cities (i.e.,
surface, fixed-
exposure
Houston
(e.g., HYSPLIT
western U.S.
R2 = 0.816 for
GAM, satellite
Houston,
site monitoring
between 0 and
and HMS and
Houston
data from HMS
Boise,
data
120 ppb;
obs data)
Denver, Fort
Table 3: obs 8-h
Collins,
daily max ozone
Provo, Salt
mean =
Lake City,
39.29 ppb, no
Spokane);
smoke n = 1,082,
T: May to
smoke n = 41,
September
no smoke
for 2008 to
residuals =
2015;
-0.33,
P: Entire
smoke residuals
population
= 8.10; 8-h daily
max ozone for
Provo site
between 20 and
80 ppb
September 2019
2-77
DRAFT: Do Not Cite or Quote
-------
Table 2-10 (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
Measurement
Exposure
Period, 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 = Multiethnic 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.
September 2019
2-78
DRAFT: Do Not Cite or Quote
-------
Table 2-11 Studies informing assessment of exposure measurement error when concentrations modeled by
chemical transport modeling are used for exposure surrogates.
Location,
Time Measurement Exposure
Period, and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Hutzell et al.
CMAQ
L: Eastern
CMAQ with
Short-term
NR (shown
Specific
Modeling time is
Mechanisms and
(2012)
modeled
CONUS;
SAPRC07T
exposure
graphically)
mechanism in
relatively short, so
observed data
output version
T: January
mechanism
CMAQ
an epi application
compared
4.7.1 with a
and July in
compared with
investigated
may be limited
January
36-km grid
2002;
SAPRC-99
during and not
NMB = -16 to
and nested
P: Entire
mechanism; both
during the ozone
16 ppb, July
12-km grid
population
mechanisms
compared with
observed fixed-site
monitoring data
season
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
September 2019
2-79
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Measurement Exposure
Period, and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Li etal. (2012)
CMAQ 4.6
L: Southeast
CMAQ with
with nested
Texas;
SAPRC07
grids (36 and
T: 3 weeks
mechanism
12 km) going
of hourly
compared with
down to 4-km
ozone
SAPRC99
grid size
between
mechanism; both
August 16
mechanisms
and
compared with
September
observed fixed-site
6, 2000;
monitoring data
P: Entire
population
Short-term
exposure
8-h avg during
ozone episode;
hourly ozone
NR (shown
graphically)
The scale of the
CMAQ data was
very fine, and
the specific
mechanisms
were directly
compared
Only 3 weeks of
data investigated
MFE by site for
S99 = 0.14-0.33
ppb and 0.25 ppb
overall, MFE by
site for
S07 = 0.17-32
ppb and 0.25 ppb
overall, MFB by
site for
S99 = -0.21
to-0.04 ppb and
-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
September 2019
2-80
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (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 GEOS-
long-term
average for
modeling
GEOS-Chem may
CAMx = 0.34-
2- x 2.5-
ozone data
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
September 2019
2-81
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
McDonald-Buller
GEOS-Chem
L: CONUS,
Comparison done
Short- and
NR (shown
Modeled output
Assumption that
No exposure
etal. (2011)
(it is not clear
with global
with fixed-site
long-term
graphically)
compared to
observed data is
measurement
the version of
comparisons
CASTNet monitors,
exposure
observed data
representative
error presented
GEOS-Chem
ofTES, OMI,
global scales
and other
in tables
used with
GEOS-
compared with
modeled data
resolution 0.5
Chem (TES
OMI, TES,
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);
T: 8-h daily
max from
March-May
2006
compared
(OMI AK)
with June-
August
2006, with
results
displayed
between
2006 and
2008; P:
Entire
population
September 2019
2-82
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, 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
September 2019
2-83
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Schere et al.
CTM of CMAQ
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 standard
CHIMERE at a
conditions over
also explored
deviation of
0.25-degree
CONUS, CMAQ
modeled
horizontal
with AQMEII
ozone = 0-
resolution
boundary
2.5 ppb
(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
September 2019
2-84
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, 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 brute-
horizontal
CONUS;
observed ozone
applied to the
associated with
force method for
resolution
T: Hourly
data from AQS
method, method
introducing a new
selected stations
ozone in
coming from fixed-
allows for
process to the
by ozone
July and
site monitors;
specification
CMAQ model,
concentration,
August
CMAQ 4.7.1 using
with chemical
results only shown
RMSE = 6 ppb at
2005;
brute force
processes with
for selected
Charlotte site,
P: Entire
emissions changes
HDDM, method
monitors; model
RMSE = 4-7 at
population
clearly
explained;
method
compared to
observed data
only run for 2 mo
Detroit site
September 2019
2-85
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Measurement Exposure
Period, and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Tsimpidi et al.
(2012)
CMAQ 4.7
with 12-km
horizontal
resolution with
nested 4-km
grid in Seattle
with a
DDM-3D
L: Seattle,
WA;
T: Hourly
data for July
12-24,
2006;
P: Entire
population
CMAQ 4-km
resolution
compared with
12 and 36 km,
compared with
observed data from
fixed-site monitors
from AQS
Short-term
exposure
NR (shown
graphically)
Method was
compared to
observed data;
the paper
explored the
effect of grid
resolutions
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
Comparing
modeled to
observed to
hourly, maximum
hourly, max 8-h
ozone to different
grid resolutions
MB, GE, RMSE,
NMB, NME,
mean
obs = 30.2-
57.0 ppb for
4 km, 30.3-57.0
ppb for 12 km,
38.3-65.1 ppb for
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,
September 2019
2-86
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Measurement Exposure
Period, and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Tsimpidi et al.
(2012) (cont.)
CMAQ 4.7
with 12-km
horizontal
resolution with
nested 4-km
grid in Seattle
with a
DDM-3D
(cont)
L: Seattle,
WA;
T: Hourly
data for July
12-24,
2006;
P: Entire
population
(cont.)
CMAQ 4-km
resolution
compared with
12 and 36 km,
compared with
observed data from
fixed-site monitors
from AQS (cont.)
Short-term
exposure
(cont.)
NR (shown
graphically)
(cont.)
Method was
compared to
observed data;
the paper
explored the
effect of grid
resolutions
(cont.)
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
(cont.)
-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%
September 2019
2-87
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Measurement Exposure
Period, 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
September 2019
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Hu et al. (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 ppb
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%
September 2019
2-89
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Hu et al. (2012)
UCD-CIT
L: San
Compare model
Short-term
Mapped
Study designed
4-km resolution
gross
(cont.)
model (cont.)
Joaquin
results from two
exposure
concentrations
to interpret
not fine enough to
error =11.5 ppb
Valley, CA
photochemistry
(cont.)
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
(cont.)
concentrations
sources (cont.)
gross
August 2,
(cont.)
and to evaluate
error = 5.7 ppb
2000;
different model
NGE = 6.6%,
P: Entire
variations (cont.)
domain
population
bias = -2.7 ppb
(cont.)
NB = -3.3%
gross
error = 9.6 ppb
NGE = 12.0%
1-h R = 0.7-0.8,
1-h
NMB = 4.9-17.0
%, 1-h
NME = 15.6-25.7
%, 8-h daily max
R = 0.8-0.8, 8-h
daily max
NMB = 8.5-
25.0%, 8-h daily
max
NME = 17.0-
30.1%
14,619
Liu and Zhang
(2011)
CMAQ 4.4
over CONUS
at 32-km
horizontal
resolution with
MM5 3.4 with
NEI 3
L: CONUS;
T: hourly
ozone from
June 12-28
1999;
P: Entire
population
Comparison of U.S.
EPA observed
fixed-site monitors
Short-term
exposure
1-h mean obs =
53.0-60.3 ppb,
1-h mean mod =
62.0-67.4 ppb,
1 h n = 84-
14,659; 8-h
daily max mean
obs = 46.6-55.0
ppb, 8-h daily
max mean mod
= 58.3-62.2
ppb, 8-h daily
max n = 82-
Many evaluation
methods: both
the horizontal
grids and
vertical grids
(through flight
data) were
evaluated with
observed data,
satellite data
used
The horizontal
resolution was
coarse; modeling
period was short
and specific so epi
application will not
be representative
of longer term
exposure
September 2019
2-90
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Measurement Exposure
Period, and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Wanaetal. (2012)
) CMAQ4.7
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 CMAQ-
compared with
max mean mod
be representative
2.8%, NME =
Dust with
observed data from
= 45.4-51.1
of exposures
16.6-18.5%,
incorporation
different U.S. EPA
ppb; AIRS max
outside of this
R = 0.46-0.53,
of
fixed-site
1 h n = 29,993,
time window;
NMB = -4.7 to
ISORROPIA
monitoring
mean obs =
coarse resolution
6.8%, NME =
II, WRF 3.2,
networks
52.7 ppb, mean
is very coarse,
17.9-18.8%;
1999 NEI 1
mod = 48.7-
54.1 ppb; in
Beijing mean 1
max h ozone is
86.8-112.4 ppb,
max 1 h, n = 30,
mean obs =
95.8 ppb, mean
mod = 86.8-
109.9 ppb
making exposure
assignment in an
epi study have
potential
misclassification
R= -0.03 to 0.06,
NMB = -9.36 to
17.3%, NME =
25.5-30.6%;
ozone difference
spatially = -1.5 to
1.5 ppb
September 2019
2-91
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Measurement Exposure
Period, and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Yu etal. (2012)
CMAQ
L: Eastern
(version NR)
U.S. with
12 km where
flight data
WRF-ARW
over east
(Advanced
Texas; T:
Research
Hourly data
WRF 3.0) and
from August
WRF-NMM
1-October
are compared
15, 2006;
P: Entire
population
Two different
modeling methods
were compared
with observed data
from U.S. EPA's
fixed-site monitors
and observed data
from air planes with
flight paths and
ship data in port
Short-term
exposure
1 h for model
ARW (model
NMM),
n = 51,532,
mean obs =
48.6 ppb, mean
mod = 56.2 ppb
(56.7 ppb), 8-h
daily max mean
obs = 42.7,
mean
mod = 50.4 ppb
(52.0 ppb);
mean ± SD for
obs = 36.38 ±
24.13
Multiple
comparison
methods with
two different
types of
modeled and
multiple types of
observed data
(e.g., fixed-site,
flight data, ship
data)
Version of CMAQ
never stated;
vertical validation
not applicable for
an epi setting;
because the study
was short term,
ambient
concentrations
may not be
representative of
a longer term
exposure
1 h for model
ARW (model
NMM) MB = 7.5
ppb (8.1 ppb),
RMSE = 13.4
ppb (13.9 ppb),
NMB = 15.5%
(16.7%),
NME = 22.3%
(22.8%), R =
0.76 (0.75), 8-h
daily max mean
MB = 7.7 ppb
(9.3 ppb), RMSE
= 12.6 ppb (13.8
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,
September 2019
2-92
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, 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%,
(cont.)
(version NR)
U.S. with
modeling methods
exposure
ARW (model
comparison
never stated;
NME = 0.1-
12 km where
flight data
were compared
(cont.)
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; T:
from U.S. EPA's
mean obs =
types of
an epi setting;
40.07 ±22.46,
Research
Hourly data
fixed-site monitors
48.6 ppb, mean
modeled and
because the study
mod ARW =
WRF 3.0) and
from August
and observed data
mod = 56.2 ppb
multiple types of
was short term,
41.33 ±20.36,
WRF-NMM
1-October
from air planes with
(56.7 ppb), 8-h
observed data
ambient
NMB NMM =
are compared
15, 2006;
flight paths and
daily max mean
(e.g., fixed-site,
concentrations
10.1%, NMB
(cont.)
P: Entire
ship data in port
obs = 42.7,
flight data, ship
may not be
ARW = 13.6%
population
(cont.)
mean
data) (cont.)
representative of
(cont.)
mod = 50.4 ppb
(52.0 ppb);
mean ± SD for
obs = 36.38 ±
24.13 (cont.)
a longer term
exposure (cont.)
Godowitch et al.
CMAQ 4.7,
L: Eastern
Compared with
Short-term
Daily max 8 h
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
September 2019
2-93
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, 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 county-
No new method
Monthly NMB
12-km
U.S.;
outputs to
exposure
seasonal
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
1-1 line with
observed data
and difference
modeled data but
not correlation
calculated; no
clear exposure
measures related
to ozone found in
text, figures, or
tables
Carlton and Baker
(2011)
CMAQ 4.7.1
L: Oark
Biogenic emission
12-km
region of the
from both BEIS
horizontal
U.S.
3.14 and MEGAN
resolution,
covering
2.04 compared with
AER05,
Missouri,
fixed-site monitors
CB05, WRF
Illinois,
(AIRS,
3.1, 2001 NEI
Indiana,
CASTNet/IMPROV
2 and BEIS
Kentucky,
E network), balloon
3.14
and northern
and aircraft
compared with
Arkansas;
measurements
MEGAN 2.04
T: Hourly
ozone data
from June
15-July 31,
1998;
P: Entire
population
Short-term
exposure
Hourly ozone;
NR (shown
graphically)
Comparison of
biogenic
emissions has
the specificity to
understand the
crux of ozone
differences in
areas with
higher isoprene
emissions;
comparison
methods were
thorough with
comparing two
types of
modeled data
with multiple
sources of
observed data
Given the short
time period,
concentrations
over a month and
half may not be
indicative of more
long-term
exposure; the
area is relatively
rural so precursor
emissions may
vary in other parts
of the U.S. where
more population
may be affected
September 2019
2-94
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Stevn et al. (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)
September 2019
2-95
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Measurement Exposure
Period, and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Zhou etal. (2013)
CMAQ 4.7
12-km
horizontal
resolution,
MM5 3.6.3,
SMOKE 2.2
L: Eastern
U.S.;
T: 2002 and
2006;
P: Entire
population
Observed data
from surface fixed-
site monitors were
compared with
CMAQ; 2002
compared with
2006; NOxSIP Call
region compared
with data outside of
NOxSIP Call
region
Long-term
exposure
Only ozone
differences are
explored and
direct ozone
concentrations
are not explored
There are
multiple
comparison of
this paper:
modeled to
observed, data
from 2002
compared with
2006; ozone by
percentage;
inside NOx SIP
Call area vs.
outside
The paper
recognizes the
issues of long-
range transport of
ozone
Relative
difference
between
quantities of obs
and mod, in SIP
call region (obs is
reference):
average change
(2002-2006) in
8-h daily max =
42.5%, average
percentage
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%
September 2019
2-96
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Appel et al. (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%
Cho et 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%
September 2019
2-97
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Herwehe et al.
CMAQ 4.7,
L: CONUS,
WRF CMAQ
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 data
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 WRF-
Short-term
NR (shown
A highly
The model run is
Comparison
12-km
California
CMAQ compared
exposure
graphically)
specialized
only for a few
between mod
horizontal
and
with offline WRF
modeled
days during a
and obs: all data
resolution with
surrounding
with CMAQ, both
component pin
wildfire; therefore,
(daytime) slope =
two-way
states;
methods are
points the
short-term
0.98, R = 0.62;
coupled WRF-
T: June 20-
compared with
differences in
exposures may be
when AOD > 0.5
CMAQ model
29, 2008;
fixed-site observed
modeled ozone;
higherthan a
slope = 1.2,
P: Entire
monitoring data
comparisons are
typical
R = 0.75
population
from AQS
made with
observed data
concentration
September 2019
2-98
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Measurement Exposure
Period, and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Bond etal. (2013)
CMAQ 4.5.1
with CAMx
4.42 both with
MM5 3.7 and
a 4-km
horizontal
resolution
L: South-
eastern U.S.
in North
Carolina,
northeastern
Georgia,
South
Carolina,
Tennessee,
Virginia, and
Kentucky;
T: Hourly
ozone in
January and
July 2002;
P: Entire
population
Both CMAQ and
CAMx are
compared to each
other; both
modeled compared
to observed fixed-
site monitors from
AQS, CASTNet,
SEARCH, and
NCDENR
Short-term
exposure
January 2002:
Mean 1-h max =
31.8-42.0 ppb,
8-h max = 27.5-
39.0 ppb across
CMAQ and
CAMx at
locations of
AQS, CASTNet,
and SEARCH
monitors;
January 2002:
Mean 1-h max =
59.8-74.7 ppb,
8-h max =
55.7-67.3 ppb
Fine-scale
horizontal
resolution; two
different models
compared with
both model
compared to
observed data;
errors with
observed data
presented by
monitoring
network
Short-term
exposure during
only 2 mo may not
be indicative of
typical, long term
exposures
1-h max, n
between 62 and
384, R = 0.5-0.7,
NMB = -7.6 to
10.1%, NME =
15.4-24.5%, and
8-h max, n = 61-
384, R = 0.6-0.7,
NMB = 0.1-
15.8%, NME =
19.2-25.4%
Kavnak et al.
(2013)
CMAQ 4.5
36-km
horizontal
resolution
L: CONUS;
T: Hourly
ozone July
1-August
31, 2004;
P: Entire
population
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 data are
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%
September 2019
2-99
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, 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 ofthe
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; VR = 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; VR = 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
data; there is an
lose spatial
CMAQ data
(unclear which
participants
examination) =
epi application
heterogeneity;
CMAQ version
(considered
37.5-60.3 ppb
inverse-distance
was used and
represent-
mean = 51.5
weighting is a
how CMAQ
ative 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
September 2019
2-100
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Measurement Exposure
Period, and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Jaffeetal. (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
L: Salt Lake
CMAQ and WRF- Short- and
Obs n =
Several different
Modeled data
City, Boise,
Chem have been long-term
1,449-1,586,
data sources
never directly
Reno;
validated in exposures
obs min =
used (e.g.,
compared with
T: June-
previous
17.0-24.8 ppb,
monitoring data,
obs data; for the
September
publications;
obs maximum =
CMAQ, and
short time
2000-2012;
multilinear modeled
82-101.5 ppb,
WRF-Chem),
window, exposure
L: Western
not validated, but
obs mean =
variety of
may not be
U.S.;
assessed
50.9-55.8 ppb,
timescales
indicative of
T: June 10-
and obs SD =
explored
longer termed
July 10,
8.4-11.0
exposure
2008;
L: CONUS;
T: summer
2012;
P: Entire
population
NR
September 2019
2-101
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Chen etal. (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
May15-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 CMAQ-
horizontal
P: Entire
each other; each
emissions
typically
AQS = -31.3 to
resolution over
population
method compared
inventory and
exposures
-13.1%; CMAQ
August 2009
with surface, fixed-
pin point ozone
(with satellite-
with baseline
sited U.S. EPA
differences;
based
emissions
AQS data
comparison to
emissions)-AQS
compared with
observed data
= 9.6-38.1%
NOx satellite-
adjusted
emissions
September 2019
2-102
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, 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 data
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
September 2019
2-103
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Measurement Exposure
Period, and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Karamchandani et
al. (2014)
CMAQ 5.01
with APT with
12-km
horizontal
resolution
L: Eastern
U.S.;
T: January
1-15, July
1-15, 2005;
P: Entire
CMAQ APT
compared with
CMAQ base and
both models are
compared with
fixed-site surface
Short-term
exposure
population monitors from U.S.
EPA's AQS sites
Mean in July
above 40 ppb
was 55.4 ppb,
above 60 ppb
was 70.2;
around point
sources in July
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
Specific CMAQ
component
update to see a
pointed
difference; APT
differences
explored around
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%,
September 2019
2-104
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (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
Exposure
Measurement
Limitations Errors
Karamchandani et CMAQ 5.01
L: Eastern
CMAQ APT Short-term
Mean in July
Specific CMAQ
R2 = 0.30; in July
al. (2014) (cont.) with APT with
U.S.;
compared with exposure
above 40 ppb
component
with 40 ppb cut
12-km
T: January
CMAQ base and (cont.)
was 55.4 ppb,
update to see a
off with 5 x 5 grid
horizontal
1-15, July
both models are
above 60 ppb
pointed
CMAQ, n =
resolution
1-15, 2005;
compared with
was 70.2;
difference; APT
2,791, mean obs
(cont.)
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,
(cont.)
EPA's AQS sites
above 40 ppb
point sources
ratio of means =
(cont.)
was 55.1 ppb,
(cont.)
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,
(cont.)
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%,
September 2019
2-105
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (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
Exposure
Measurement
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) (cont.) with APT with
U.S.;
compared with exposure
above 40 ppb
component
with 40 ppb cut
12-km
T: January
CMAQ base and (cont.)
was 55.4 ppb,
update to see a
off with 9 x 9 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
(cont.)
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
(cont.)
EPA's AQS sites
above 40 ppb
point sources
means = 0.99,
(cont.)
was 55.1 ppb,
(cont.)
GB = -0.4 ppb,
above 60 ppb
NB = 0.7%, FB =
was 70.7 ppb;
-2.5%, GE = 9.7
around point
ppb, NE =
sources in July
18.2%, FE =
above 40 ppb
18.9%, NMB =
was 56.1 ppb,
-0.8%, NME =
above 60 ppb
17.4%, R2 = 0.27,
was 70.3 ppb
CMAQ APT, n =
(cont.)
7,197, mean obs
= 56.1 ppb, mean
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
September 2019
2-106
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Measurement Exposure
Period, and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Zhang et al.
(2014)
CMAQ 4.7.1
L: CONUS,
CMAQ output
Long-term
NR (shown
This paper has Limited obs for
Monthly mean
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; P:
Detroit, St.
Entire
Paul, Winston-
population
Salem)
September 2019
2-107
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (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
breaks down 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
SI LAM
-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
September 2019
2-108
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Measurement Exposure
Period, 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%,
September 2019
2-109
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, 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) (cont.)
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, fixed-
(cont.)
ozone season =
more indicative
forward is not an
74.0%, bias =
3.0, CB05,
September
site monitors from
41.4—48.6 ppb;
of typical
ideal exposure
1.1-1.7, CSI =
12-km
2009-2011,
AQS, CASTNet,
avg 8-h daily
exposes;
methodology in an
27.4-30.6%,
horizontal
December to
IMPROVE, and
max during
sensitivity
epi setting (cont)
FAR = 58.8-
resolution over
February
SEARCH (cont.)
winter = 31.1—
analysis was
63.3%, POD =
the
2009-2012,
36.0 ppb (cont.)
extensive (cont.)
44.2-62.1%
southeastern
sensitivity
U.S.
analysis
comparing
performed
biogenic
July 2011;
emission from
P: Entire
MEGAN2,
population
satellite-
(cont.)
derived fire
emissions
(SD-Fire), and
MEGAN2+
SD-Fire (cont.)
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:
typical long-term
exposures
Entire
population
September 2019
2-110
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Measurement Exposure
Period, and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Bias U.S.
emissions urban
= 7.08 ppb,
suburban =
7.48 ppb, rural =
7.80 ppb; U.S. +
Canadian
emissions urban
= 6.16 ppb,
suburban =
6.22 ppb, rural =
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
Pan etal. (2014)
NAQFC-beta
(coupled
NAM-CMAQ
4.7.1 with
mobile
sources
defined from
2005
MOBILE6 + 05
to 12
projections,
point sources
L: CONUS;
T: July 2011;
P: Entire
population
Modeled output
compared with
surface, fixed-site
monitors from U.S.
EPA's AQS and
CTM with base-
case emissions
Short-term Average hourly Complete cover Ozone
exposure ozone by hour of CONUS;
NR (shown
graphically)
incremental
change of model
inputs pinpoints
differences
between two
different models
concentrations in
1 summer month
is not indicative of
more long-term
exposures
September 2019
2-111
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (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
Cuchiara et al.
ARW-WRF,
L: Houston,
All four PBL
Short-term
NR (shown Very fine scale
Highly localized
Statistics across
(2014)
WRF/Chem
TX;
schemes compared
exposure
graphically) resolution;
space/time
sites for four
3.5 with 4-km
T: October
to each other and
small,
modeling scenario
boundary layer
horizontal
5, 2006;
to observed, fixed-
incremental
is not indicative of
schemes: YSU R
resolution,
P: Entire
site monitors from
changes in
more long-term
= 0.79-0.92, bias
with four
population
U.S. EPA's CAMS
model pin points
exposures
= 0.59-0.99,
planetary
and aloft observed
model
RMSE = 13.20-
boundary layer
data from
differences and
21.02 ppbv; MYJ
schemes from
ozonesonde and
assumptions
R = 0.70-0.90,
YSU, MJY,
a ire rafts
bias 0.64-1.05,
ACM2, QNSE.
RMSE = 12.17-
YSU, and
20.76 ppbv;
ACM2 are
ACM2 R = 0.37-
computed
0.77, bias =
based on the
0.75-1.26,
bulk
RMSE = 18.53-
Richardson
25.77 ppbv;
number, which
QNSE R = 0.54-
is the ratio of
0.71, bias =
buoyancy to
0.72-1.09,
turbulence
RMSE = 15.59-
caused by
24.85 ppbv
shear
stresses. MJY
and QNSE are
computed
based on
eddy-
diffusivity, or
atmospheric
mixing
September 2019
2-112
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Measurement Exposure
Period, 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 36-
4-, and 2-km
Brazoria
other and to
maximum ozone
resolution of the
coverage of the
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 long-
resolution = 63%,
13-
quality monitors in
graphically)
term exposures
4-km resolution =
September
the region
26%, 2-km
15, 2006;
resolution = 25%
P: Entire
population
Lu etal. (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
comparison
necessarily
ozone NMB =
resolution with
and July
modeled data
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
September 2019
2-113
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Measurement Exposure
Period, 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%;
September 2019
2-114
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, 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, fixed-
exposure
ozone before
coverage with
windows may not
ozone before
MOBILE 6.2,
U.S.;
site observed data
(after) smoke
CTM data;
be indicative of
(after) smoke
GVRD,
T: July 3-
from U.S. EPA's
event mean obs
multiple data
typical exposures
event R = 0.7
BEIS-3,
August
AQS
= 45.8 ppb
sources used to
(0.8), MB = -4.7
SMOKE 2.4,
2007, June
(42.3 ppb)
model 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
September 2019
2-115
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Measurement Exposure
Period, 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%)
September 2019
2-116
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Reference
Model
Period, and
Population
Evaluation
Technique
Epidemiology
Applications
Concentrations
Measured
Strengths
Limitations
Measurement
Errors
Hoarefe et al.
CMAQ 5.0.1,
L: CONUS;
Model compared
Short- and
Mean during
CMAQ has full
Number of months
Avg 8-h daily
(2015)
WRF 3.4 with
T: June to
with surface, fixed-
long-term
2006 of 8-h
coverage of
and different
max ozone
12-km
August
site monitors from
exposure
daily max ozone
spatial domain
years explored
June-August
horizontal
2006, May to
U.S. EPA's AQS
= 32.3-51.1
lends itself to both
2006: MB = -0.9
resolution
-September
ppb; mean
short- and
to 6.6 ppb, ME =
2010;
during 2010 of
long-term
5.6-10.9 ppb,
P: Entire
8-h daily max
exposures
RMSE = 7.4-
population
ozone = 33.6-
14.4 ppb, NMB =
47.5 ppb
-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 month
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
September 2019
2-117
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Koo etal. (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.
WRF/Chem
L: CONUS;
Model compared
Short- and
Mean maximum
Comparison
CTMs have
R 1-h daily max
(2015b)
3.4.1 with
T: January,
with surface,
long-term
1 h ozone =
method to obs
inherent errors
CASTNet = 0.40,
36-km
February,
observed fixed-site
exposure
between 33.2
data; time
and horizontal
AQS = 0.34; R =
horizontal
December
monitors from U.S.
and 48.4 ppb,
window lends
resolution is
8-h daily max
resolution for
2006 and
EPA's CASTNet
mean 8-h daily
itself to both
coarse
CASTNet = 0.40,
2006 and
2010 with
and AQS
max ozone =
short- and long-
AQS = 0.20;
2010
June, July,
between 32.7
term exposure
NMB 1-h daily
August 2006
and 43.8 ppb
max CASTNet =
and 2010;
-30.0%, AQS =
P: Entire
-15.8%; NMB =
population
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%
September 2019
2-118
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Pan etal. (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-
observations from
exposure
addition of
model,
resolution
25.3%, SD of
not reported)
2015,
fixed-site monitors
study
2.6 ppbv across
1,200 monitoring
(50 km), model
NMB = 17.9%
with adjoint
P: Entire
reporting to the
(contribution
U.S. due to
sites used for
based on 2005
(NMB calculated
population
AQS
from VW
emissions)
excess NOx
emissions
validation
emissions
inventory
(emissions have
dropped over
time)
from 1-h daily
max)
September 2019
2-119
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, 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 (cross-
and errors are
increases errors
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
observations from
long-term
50 ppbv for
analysis of the
may lead to bias
NB = -2.8%
focused on
U.S.;
fixed-site monitors
exposure
3-mo avg;
influence of
because not all
(-28.2, 28.0%),
impact of
T: March
reporting to the
impact of
meteorology and
cloud chemistry
NGE = 18.8%
trans-Pacific
21-June 30,
AQS
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
September 2019
2-120
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, 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
September 2019
2-121
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Measurement Exposure
Period, and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Yahva et al.
(2015a)
WRF-Chem
3.4.1 at 36-km
resolution with
initial and
boundary
conditions
downscaled
from global
models
L:
Contiguous
U.S.;
T: 2001,
2006, 2010;
P: Entire
population
Comparison with
observations from
fixed-site monitors
reporting to the
AQS and CASTNet
Long-term
exposure
(Model values at
sites for
networks
mentioned)
2001 CASTNet
8-h max =
39.0 ppb,
CASTNet 1-h
max = 38.3 ppb,
AQS 8-h max =
44.6 ppb, AQS
1-h max =
49.4 ppb; 2006
CASTNet 8-h
max = 38.4 ppb,
CASTNet 1-h
max = 39.3 ppb,
AQS 8-h max =
43.2 ppb, AQS
1-h max = 48.1
ppb; 2010
CASTNet 8-h
max = 38.2 ppb,
CASTNet 1-h
max = 38.6 ppb,
AQS 8-h max =
41.8 ppb, AQS
1-h max =
47.3 ppb
Extensive
Lower resolution
2001 CASTNet
comparisons
(36 km)
8-h max: MB =
made at
-4.8 ppb, NMB =
different time
-11.0%, NME =
averages and
28.2%, CASTNet
validation data
1-h max = MB:
sets, validation
-7.9 ppb, NMB =
on
-17.2%, NME =
meteorological
30.1%, AQS 8-h
variables as well
max = MB: -0.3
ppb, NMB =
-0.7%, NME =
29.9%, AQS 1-h
max MB = -1.7
ppb, NMB =
-3.3%, NME =
28.5%; 2006
CASTNet 8-h
max: MB =
-5.2 ppb, NMB =
-11.8%, NME =
27.1%, CASTNet
1-h max: MB =
-8.3 ppb, NMB =
-17.4%, NME =
28.7%, AQS 8-h
max:
September 2019
2-122
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Measurement Exposure
Period, and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Yahva et al.
(2015a) (cont.'
WRF-Chem
L:
Comparison with Long-term
Model values at
Extensive
Lower resolution
MB = -1.2 ppb,
3.4.1 at 36-km
Contiguous
observations from exposure
sites for
comparisons
(36 km) (cont.)
NMB = -2.8%,
resolution with
U.S.;
fixed-site monitors (cont.)
networks
made at
N ME = 27.5%,
initial and
T: 2001,
reporting to the
mentioned)
different time
AQS 1-h max MB
boundary
2006, 2010;
AQS and CASTNet
2001 CASTNet
averages and
= -2.2 ppb, NMB:
conditions
P: Entire
(cont.)
8-h max =
validation data
= -4.5%, NME =
downscaled
population
39.0 ppb,
sets, validation
26.3%; 2010
from global
(cont.)
CASTNet 1-h
on
CASTNet 8-h
models (cont.)
max = 38.3 ppb,
meteorological
max: MB =
AQS 8-h max =
variables as well
-5.7 ppb, NMB =
44.6 ppb, AQS
(cont.)
-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 (cont.)
September 2019
2-123
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (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 etal. (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
Ran etal. (2016) CMAQ
L:
Comparison with
Long-term
CMAQ: April
Addition of leaf 12-km resolution
CMAQ: April
5.0.2/WRF 3.4
Contiguous
observations from
exposure
2006 = 52.70
area index
2006: RMSE =
with MODIS
U.S.,
fixed-site monitors
ppb, August
allows for
9.51 ppb, MAE =
leaf area index
southern
reporting to the
2006 = 55.00
consideration of
7.33 ppb, MB =
model
Canada,
AQS
ppb, October
role of
3.94 ppb; August
included in
northern
2006 = 42.2.0
vegetation
2006: RMSE =
some runs
Mexico;
ppb; CMAQ +
12.80 ppb, MAE
T: April,
MODIS: April
= 9.70 ppb, MB =
August,
2006 = 55.40
4.84 ppb;
October
ppb, August
October 2006:
2006;
2006 = 57.10
RMSE = 10.10
P: Entire
ppb, October
ppb, MAE = 8.20
population
2006 = 44.60
ppb, MB = 5.34
ppb
ppb
September 2019
2-124
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, 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.;
observations from
long-term
meteorology
continually
still inhibits urban
meteorology
resolution with
T: June 1-
fixed-site monitors
exposure
Model 1 = 50.1
updated data
studies
Model 1 MB =
WRF/FDDA
August 31,
reporting to the
ppbv, Model 2 =
improves
9.8 ppbv, MAE
meteorology
2002;
AQS, tower
48.2 ppbv;
accuracy of
13.5 = ppb, Fp
and boundary
P: Entire
sensors at one
northeastern
model
42%, Model 2
conditions
population
location (Raleigh,
U.S.: Model 1 =
MB = 7.9 ppbv,
from a global
NC), and
49.6 ppbv,
MAE = 12.6 ppb,
GEOS-Chem
DISCOVER-AQ
Model 2 = 47.5
Fp 58%;
simulation
flight sensors
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%
Without
assimilation: IOA
= 0.78, RMSE =
14.9 ppb, MAE =
12.3 ppb, MB =
9.3 ppb; with
assimilation: IOA
= 0.83, RMSE =
13.8 ppb, MAE =
11.0 ppb, MB =
6.1 ppb
Li etal. (2016b)
CMAQ 5.0.2
with
WRF/FDDA
meteorological
model and
assimilation of
meteorological
data
L: Southeast
TX,
southwest
LA;
T:
September
2013;
P: Entire
population
Comparison with
observations from
fixed-site monitors
reporting to the
AQS
Short-term
exposure
Without
assimilation
mean = 33.7
ppb, SD = 14.1
ppb; with
assimilation
mean = 30.6
ppb, SD = 17.4
ppb
Data
assimilation
improves
representation
of short-term
variability in
concentration
field, better
captures hot
spots
4-km resolution
misses spatial
variation
September 2019
2-125
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, 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;
observations from
exposure
provides more
the model
0.69-0.82,
Forecast
T: July 2011;
fixed-site monitors
mechanistic
overpredict at low
RMSE =
Capability
P: Entire
reporting to the
information,
concentrations
13.59-18.75 ppb,
combines
population
AQS
despite having
and underpredict
MB = -1.15 to
CMAQ 4.5
higher error
at high
8.96 ppb, NMB =
with WRF-
concentrations
-2.28 to 22.34%;
NMM, also
NAQFC-beta:
tested beta
Corr = 0.67-0.81,
version with
RMSE = 15.81 —
full gas and
20.92 ppb, MB =
aerosol
3.40-13.84 ppb,
mechanism
NMB = 6.75-
34.49%
Wana et al. (2016)
UCD-CIT
L: Los
10-fold cross-
Short- and
8-h daily max,
CTM accounts
Positive bias in
RMSE =
chemical
Angeles and
validation against
long-term
annual
for atmospheric
the concentration,
8.83 ppb, R2 =
transport
Riverside
37 monitors for
exposure
averages
chemistry, long-
more variability
0.56
model with
counties,
each variation of
(annual average
range transport
compared with
meteorology
CA;
model
used for
of ozone and its
spatiotemporal
modeled by
T: 2000-
summary stats)
precursors, and
models
WRF 3.1.1 on
2008;
biogenic VOCs
a 4-km grid
P: Entire
population
September 2019
2-126
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, 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 observations
exposure
graph) March-
captured
summer and fall
across years
36-km domain
December
from monitors
May = 45-50
and negatively
March-May = 0.9
with met fields
31, 2010;
reporting to the
ppb, June-
biased in the
ppb (0.7 ppb),
downscaled
P: Entire
population
AQS
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)
September 2019
2-127
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Measurement Exposure
Period, and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Yahva et al.
(2016)
WRF/Chem
3.6.1 with
36-km
horizontal
resolution
L: CONUS;
T: 2001-
2010;
P: Entire
population
Model compared
with surface, fixed-
site monitors from
U.S. EPA's AQS
and CASTNet
Long-term
exposure
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
maximum 1-h
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
maximum 1-h
ozone R = 0.4,
MB = -8.9 ppb,
NMB = -18.8
ppb, NME =
31.4%; vs.
CASTNet
maximum 8 h
ozone R = 0.5,
MB = -5.4 ppb,
NMB = -12.5%,
NME = 29.6%
September 2019
2-128
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time Measurement Exposure
Period, 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
September 2019
2-129
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, 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
Average hourly
This incremental
Localized
Biases and errors
WRF 3.3,
and northern
to each other and
exposure
obs ozone
improvement in
spatiotemporal
when using
biogenic
California;
compared to
greater than
modeled inputs
modeling domain
satellite
emission from
T: June 3-
observed, fixed-site
60 ppb =
allow for seeing
may not be typical
parameterization
BEIS 3.61 with
July 31,
monitors from U.S.
70.9 ppb, less
pointed
of average ozone
of weather
4-km
2009;
EPA's AQS
than 60 =
concentration
exposures
model: ozone
horizontal
P: Entire
32.0 ppb,
changes; fine
greater than
resolution;
population
average mod
spatial
60 ppb: median
sensitivity
hourly ozone
resolution
bias = -8 to -9
analysis
greater than
ppb, median
includes BEIS
60 ppb =
error = 13-14
3.14, BEIS
between 62.1
ppb, MB = -6.2
3.61 WRF par,
and 64.8, less
to -5.5 ppb, ME
MEGAN 2.1
than 60 ppb =
= 11-12 ppb, FB
WRF par
between 40.7
= -10.1 to
and 41.7 ppb
-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%
Accel et al. (2017)
WRF 3.7 and
L: CONUS;
Annual, monthly,
Long-term
NR
Benchmark
12-km resolution
NR
CMAQ 5.1
T: 2011
seasonal and
exposure
study of state-of-
annual
diurnal evaluations
the-art CTM
simulation;
provided against
science and
P: Entire
AQS data
evaluation
population
September 2019
2-130
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Pleim etal. (2016)
WRF 3.7 and
L: CONUS;
Daily evaluation
Short-term
NR
Improvements
12-km resolution,
NR
CMAQ5.1
T: 3-week
against 1,144 sites
exposure
made in Land-
3-week simulation
simulation
Surface Model
August 10-
and PBL model
30, 2006;
provides more
P: Entire
accurate ozone
population
simulations
Pan etal. (2017a)
WRF 3.4 and
L: Houston,
Hourly evaluation
Short-term
NR
4-km resolution
1-day study
With
CMAQ 5.0.1
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
Muniz-
WRF 3.71,
L: Greater
Hourly and month-
Short- and
NR
4-km resolution
1-mo analysis
Inclusion of
Unamunzaqa et
CMAQ 5.1
Los Angeles
long aggregated
long-term
marine halogen
al. (2018)
area, CA;
evaluation against
exposure
and sulfur
T:
eight AQS sites
concentrations
September
reduced model
2006;
overprediction as
P: Entire
mean bias is
population reduced from
13.5 to 4.9%
across the
domain and
month
September 2019
2-131
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (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,
CMAQ 4.7.1
L: CONUS;
T: 2005
annual
simulation;
P: Entire
population
Hourly and 8-h
daily max
Short-term
exposure
NR CONUS annual
simulation
12-km resolution
For 8-h daily
max: bias ranged
from -2 to -9
ppb, error ranged
from 9 to 9 ppb,
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-year
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.
September 2019
2-132
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, 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
Nopmonqcol 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
September 2019
2-133
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (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
Seltzer et al.
GEOS-Chem
L: CONUS;
R6MA1 (running
Short-term
NR Annual
Grid resolutions
CONUS values
(2017)
T: Two,
6-mo avg of the 1-h
exposure
simulations
were 2.0 * 2.5°
of the NMB of 8-h
2-year
daily max) and 8-h
and 0.5 x 0.666°
daily max ranged
annual
daily max of over
from 2.0 to 6.6
simulations
1,000 AQS sites
depending on
(2004-2006,
simulation year
2009-2011);
P: Entire
population
Solazzo et al. WRF, CMAQ L: CONUS;
(2017) T: 2010
annual
simulation;
P: Entire
population
Hall et al. (2012) WRF 3.1, L: CONUS;
CMAQ 4.7.1 T: 2008
annual
simulation;
P: Entire
population
-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%
Hourly ozone Short-term NR CONUS 12-km resolution Annual MSE
concentrations exposure ranged from 28.6
to 79.3 ppb2
Hourly and 8-h Short-term NR Annual 12-km resolution Evaluation was
daily max ozone at exposure simulation segregated into
1,176 AQS sites seasons and
eight CONUS
subregions. NMB
ranged from
September 2019
2-134
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured Strengths
Limitations
Errors
Zhana and Yina
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
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
September 2019
2-135
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, 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
September 2019
2-136
DRAFT: Do Not Cite or Quote
-------
Table 2-11 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by chemical transport modeling are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
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; 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 = inter-quartile 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 = Multiethnic 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; NB = normalized bias; NC DENR = North Carolina Department of Environment and Natural Resources;
NEI = National Emissions Inventory; NEu = Northern Europe; NGE = normalized gross error; NAQFC = National Air Quality Forecasting Capability; 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; UAM = Urban Airshed Model; UCD-CIT = UC Davis-California Institute of Technology model; UK = universal kriging; UPA = unpaired normalized bias; 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.
September 2019
2-137
DRAFT: Do Not Cite or Quote
-------
Table 2-12 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 et al. (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 fourth-
0.05-0.085
RMSE
temporal R2 =
concentration
population
highest 8-h
ppm, Southeast
0.75, RMSE =
predicted by
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
September 2019
2-138
DRAFT: Do Not Cite or Quote
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by hybrid approaches are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured Strengths
Limitations
Errors
Hao etal. (2012) IDWofCMAQ
L: CONUS;
CMAQ grid
Short-term
98th percentile
CMAQ data
Methods are unclear
Number of
data at the block
T: 8-h daily
resolutions of
exposure
of ozone for
have been well
for many of the
monitoring sites
group level
max ozone
36 and 12 km
2006 = between
validated
figures in the paper
between 790
(version of CMAQ
for 2006;
compared
39.8 and 100.7
and 897, n
unclear) for two
P: Entire
ppb (unclear
between
different grid
population
which data
195,035 and
resolutions: 36
source
232,081, mean
and 12 km
produced these
absolute
concentrations),
deviation
90th percentile
(12 km) between
of ozone for
3.51 and 4.53
2006 = between
ppb, mean
36.7 and 84.4
absolute
ppb (again,
deviation
unclear which
(36 km) between
data source
2.98 and 3.23
produced these
ppb, R (12 km)
concentrations)
between 0.94
and 0.96, R (36
km) between
0.96 and 0.97
September 2019
2-139
DRAFT: Do Not Cite or Quote
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by hybrid approaches are used for exposure surrogates.
Location,
Time Measurement Exposure
Period, 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, maximum
= 35.4 ppb,
linear model N =
6,919, mean =
26.4 ppb, SD =
3.4 ppb, min =
18.1 ppb, IQR =
4.7 ppb,
maximum =
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
September 2019
2-140
DRAFT: Do Not Cite or Quote
-------
Table 2-12 (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 etal. (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 data
developed for each
ozone, obs R2 =
output and
cities across
modeled met
were 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
September 2019
2-141
DRAFT: Do Not Cite or Quote
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by hybrid approaches are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, 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.; T: July
with monitors
exposure
graphically)
assimilation
oversmoothing to
135.9, PMAE =
not stated) at
4, July 20,
reporting to
improves
blur predictive maps
9.1, 95% PI NR
12-km resolution
August 9,
AQS
model fit for
CP NR;
downscaled with
2001;
different
regressor PMSE
either a Gaussian
P: Entire
variations of
= 124.2, PMAE =
Markov random
population
downscaling,
8.7, 95% PI NR
field or a
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%
September 2019
2-142
DRAFT: Do Not Cite or Quote
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by hybrid approaches are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, 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
10:00-17:00;
P: Entire
population
AQS
model has
flexibility
model outputs are
Gaussian processes
September 2019
2-143
DRAFT: Do Not Cite or Quote
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by hybrid approaches are used for exposure surrogates.
Location,
Time Measurement Exposure
Period, and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Pongprueksa
(2013)
CMAQ 4.7.1, WRF
3.4, CONUS with
36-km horizontal
resolution for 2009
merged with
Tropospheric
Emission
Spectrometer
(TES) L3 (using
TES data as is
and shifted by
10 ppb to better
account for
boundary
conditions)
L: CONUS;
T: 2009;
P: Entire
population
CMAQ using
satellite data
were
compared to
the typical
conditions for
CMAQ, both
CMAQ
methods were
compared with
observed data
from U.S.
EPA's AQS
and
ozonesonde
data collected
across CONUS
from WOUDC,
NOAA, and
TOPP
Short- and
long-term
exposure
Ozonesonde;
Addition of
Satellite data
annual 8-h daily satellite data overestimates
max ozone in
southern states
for 2009; 8-h
daily max from
Texas; annual
8-h daily max
ozone across troposphere
CONUS
reduces error tropospheric ozone
and uncertainty
both in the
upper
atmosphere
and in the
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
September 2019
2-144
DRAFT: Do Not Cite or Quote
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by hybrid approaches are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Reich et al. (2014)
Spectral
L: CONOUS;
Spectral
Short-term
1-day avg
Method is
Short-term exposure
Spatial
downscaling using
T: July 2005;
downscaling
exposure
ozone
explicitly
is not indicative of
prediction:
surface, fixed-site
P: Entire
compared with
concentration
stated; hybrid
longer ozone
monitors only
monitors from U.S.
population
no CMAQ,
models allows
exposures;
MSE = 62.8
EPA's AQS and
linear
for strength of
collocation needed
ppb2, bias =
CASTNet and
downscaler,
both CTMs and
for validation
-0.14 ppb,
CMAQ 5.0.1 with
and kernel
obs data
preferentially selects
variance = 66.3
a 12-km horizontal
smoothed
higher ozone areas
ppb2 CP = 0.91,
resolution
downscaler, all
comparison
methods are
compared
against
observed data
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
September 2019
2-145
DRAFT: Do Not Cite or Quote
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by hybrid approaches are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Reich et al. (2014
Spectral
L: CONOUS;
Spectral
Short-term
1-day avg
Method is
Short-term exposure
= 0.91, kernel
(cont.)
downscaling using
T: July 2005;
downscaling
exposure
ozone
explicitly
is not indicative of
downscaler
surface, fixed-site
P: Entire
compared with
(cont.)
concentration
stated; hybrid
longer ozone
120-km
monitors from U.S.
population
no CMAQ,
(cont.)
models allows
exposures;
resolution MSE =
EPA's AQS and
(cont.)
linear
for strength of
collocation needed
60.9 ppb2, bias =
CASTNet and
downscaler,
both CTMs and
for validation
-0.14 ppb,
CMAQ 5.0.1 with
and kernel
obs data
preferentially selects
variance = 62.9
a 12-km horizontal
smoothed
(cont.)
higher ozone areas
ppb2 CP = 0.91;
resolution (cont.)
downscaler, all
comparison
methods are
compared
against
observed data
(cont.)
nonspatial
prediction:
monitors only
MSE = 339.7
ppb2, bias =
-6.17 ppb,
(cont.) 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,
September 2019
2-146
DRAFT: Do Not Cite or Quote
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by hybrid approaches are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Reich et al. (2014
Spectral
L: CONOUS;
Spectral
Short-term
1-day avg
Method is
Short-term exposure
bias = -0.06
(cont.)
downscaling using
T: July 2005;
downscaling
exposure
ozone
explicitly
is not indicative of
ppb, variance =
surface, fixed-site
P: Entire
compared with
(cont.)
concentration
stated; hybrid
longer ozone
134.8 ppb2 CP =
monitors from U.S.
population
no CMAQ,
(cont.)
models allows
exposures;
0.89, kernel
EPA's AQS and
(cont.)
linear
for strength of
collocation needed
downscaler
CASTNet and
downscaler,
both CTMs and
for validation
60-km resolution
CMAQ 5.0.1 with
and kernel
obs data
preferentially selects
MSE = 157.6
a 12-km horizontal
smoothed
(cont.)
higher ozone areas
ppb2, bias = 1.06
resolution (cont.)
downscaler, all
comparison
methods are
compared
against
observed data
(cont.)
ppb, variance =
142.9 ppb2 CP =
0.89, kernel
downscaler
120 km
resolution MSE =
(cont.) 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
September 2019
2-147
DRAFT: Do Not Cite or Quote
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by hybrid approaches are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Paci 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 Discover-
0.52-0.58, MB =
CMAQ 5.0.2,
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
September 2019
2-148
DRAFT: Do Not Cite or Quote
-------
Table 2-12 (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)
September 2019
2-149
DRAFT: Do Not Cite or Quote
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by hybrid approaches are used for exposure surrogates.
Location,
Time Measurement Exposure
Period, and Evaluation Epidemiology Concentrations Measurement
Reference Model Population Technique Applications Measured Strengths Limitations Errors
Fribera et al.
(2016)
CMAQ 4.5 fused
with observational
data by three
methods: fusing
CMAQ with
interpolated
observations,
scaling CMAQ
fields to
observations 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
bias, low
RMSE, and
relatively high
R2 (compared
to application
of the model
for other
pollutants), and
errors are
minimized
through the
model-
weighting
approach
Errors in
measurements used
as input are
propagated into the
model, limited spatial
coverage of monitors
increases errors
(although this is less
of a limitation for
ozone and other
secondary
pollutants)
Interpolated
observations
MFE = 0.16,
MFB = 0.02,
NME = 14.7%,
NMB = -0.57%,
MB = -2.7e-4,
RMSE = 0.01, R2
(cross-
validation):
68.7%; for
optimized
method MFE =
0.05, MFB =
0.01, NME =
4.49%, NMB =
0.03%, MB =
1.5e-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 =
1.3e-5, RMSE =
0.006, R2 (cross-
validation) =
87.1%
September 2019
2-150
DRAFT: Do Not Cite or Quote
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by hybrid approaches are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, and
Evaluation
Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique
Applications
Measured
Strengths
Limitations
Errors
Xu etal. (2016b)
CAMx 5.30 at
L:
Comparison
Long-term
NR
BME enables
Uncertainty in
CAMP: RMSE
nested 36-km
Contiguous
with
exposure
estimation of
concentration
0 km = 5.675
domain over the
U.S.;
observations
concentration
estimates increases
ppb, 36 km =
CONUS and
T: 2005;
from fixed-site
below scale of
with distance from
6.442 ppb, 72
12-km domain
P: Entire
monitors
CTM
the monitors
km = 6.966 ppb,
over the eastern
population
reporting to the
simulation,
108 km = 7.250
U.S., with BME
AQS, tower
better spatial
ppb R2 0 km =
implemented by a
sensors at one
and temporal
0.884, 36 km =
model with
location
validation with
0.853, 72 km =
parameters held
(Raleigh, NC),
RAMP model,
0.831, 108 km =
constant across
and
computational!
0.819; RAMP:
the CONUS
DISCOVER-
y efficient and
RMSE 0 km =
[CAMP] and with
AQ flight
straightforward
5.445 ppb, 36
regional
sensors
approach
km = 6.109 ppb,
parameters
72 km = 6.531
[RAMP]
ppb, 108 km =
6.732 ppb R2 0
km = 0.893, 36
km = 0.866, 72
km = 0.849, 108
km = 0.841
September 2019
2-151
DRAFT: Do Not Cite or Quote
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by hybrid approaches are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, 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
Average hourly
This
Localized
Biases and
3.3, biogenic
and northern
compared to
exposure
obs ozone
incremental
spatiotemporal
errors when
emission from
CA, U.S.;
each other and
greater than
improvement in
modeling domain
using satellite
BEIS 3.61 with
T: June 3-
compared to
60 ppb =
modeled inputs
may not be typical of
parameterization
4-km horizontal
July 31,
observed,
70.9 ppb, less
allow for
average ozone
of weather
resolution;
2009;
fixed-site
than 60 =
seeing pointed
exposures
model: ozone
sensitivity analysis
P: Entire
monitors from
32.0 ppb,
concentration
greater than 60
include BEIS 3.14,
population
U.S. EPA's
average mod
changes; fine
ppb = median
BEIS 3.61 SAT
AQS
hourly ozone
spatial
bias -9 to -12
(from MODIS) par,
greater than
resolution
ppb, median
MEGAN 2.1 SAT
60 ppb =
error = 13-14
(from MODIS) par
between 62.1
and 64.8, less
than 60 ppb =
between 40.7
and 41.7 ppb
ppb, MB = -6.6
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 range 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
September 2019
2-152
DRAFT: Do Not Cite or Quote
-------
Table 2-12 (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
observations
September 2019
2-153
DRAFT: Do Not Cite or Quote
-------
Table 2-12 (Continued): Studies informing assessment of exposure measurement error when concentrations
modeled by hybrid approaches are used for exposure surrogates.
Location,
Time
Measurement
Exposure
Period, and
Evaluation Epidemiology
Concentrations
Measurement
Reference
Model
Population
Technique Applications
Measured
Strengths
Limitations
Errors
Yu etal. (2018)
CMAQ-kriging
L: Atlanta,
Comparison Short-term
NR
Improves
Complex model
Urban site: MB =
hybrid model
GA;
with fixed-site exposure
spatial
design is more
-5.8 ppb, ME =
where CMAQ
T: 2011;
monitors
resolution over
difficult to implement
6 ppb, RMSE = 7
5.0.2, 36- x 36-km
P: Entire
CMAQ alone
compared with
ppb, MNB =
resolution was
population
and monitor-
CMAQ alone or the
-12%, MNE =
run, ratios
based
CMAQ-kriging hybrid
14%, NMB =
calculated
approaches
model
-13%, NME =
between CMAQ
14%, MFB =
and observations
-13%, MFE =
across the
15%, R2 = 0.95,
surface, CMAQ
slope = 1.20;
output was
Rural site: MB =
adjusted by those
-1.79 ppb, ME =
ratios, and then
3.98 ppb, RMSE
the surface was
= 5.16 ppb, MNB
kriged to
= -4%, MNE =
interpolate
10%, NMB =
between grid
-4%, NME =
centroids (Friberq
9%, MFB = -5%,
etal.. 2016)
MFE = 10%, R2
= 0.88, slope =
0.88
September 2019
2-154
DRAFT: Do Not Cite or Quote
-------
Table 2-13 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
Dionisio et al.
Stochastic
L: Atlanta,
Comparison with
Long-term
8-h daily max
More precise
Computationally
Mean (SD)
(20141®
Human
GA;
dispersion model
exposure
ozone, NR
model
intensive, but
exposure
Exposure
T: 1999-
might not be
measurement
and Dose
2002;
needed
error:
Simulation
P: Entire
population =
(SHEDS)
population
-0.66 (0.029)
model
spatial =
-0.055 (0.037),
total = -0.72
(0.010)
Reich et al. (2012)
Air Pollutant
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,
reporting to AQS
providing a
accuracy is
linear
2001;
surrogate for
limited to quality
relationship
P: Entire
exposure
of input data
between
population
predictions and
observations,
but there are
many instances
where the
model is
positively
biased
APEX = Air Pollution Exposure model; AQS = Air Quality System; NR = not reported; SHEDS = Stochastic Human Exposure and Dose Simulation
aData were obtained from the study author.
September 2019
2-155
DRAFT: Do Not Cite or Quote
-------
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.ll.016.
Appel. KW: Napelenok. SL: Foley. KM: Pve. HOT: Hogrefe. C: Lueckea 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. (1992). Principles of exposure measurement in epidemiology. New York,
NY: Oxford University Press.
Baker. KR: Woody. MC: Tonnesea 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.
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.
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.
http://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.
http://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.l lll/i.1541-
0420.2011.01725.x.
September 2019
2-156
DRAFT: Do Not Cite or Quote
-------
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.ll.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 Environ45: 1291-1302. http://dx.doi.Org/10.1016/i.atmosenv.2010.12.004.
Bond. TC: Dohertv. SJ: Fahev. DW: Forster. PM: Berntsea T: Deangelo. BJ: Flanner. MG: Ghaa S: Karcher.
B: Koch. D: Kinne. S: Kondo. Y: Ouina PK: Sarofim. MC: Schultz. MG: Schulz. M: Venkataramaa C:
Zhang. H: Zhang. S: Bellouin. N: Guttikunda. SK: Hopke. PK: Jacobson. MZ: Kaiser. JW: Klimont. Z:
Lohmann. U: Schwarz. JP: Shindell. D: Storelvmo. T: Warren. SG: Zender. CS. (2013). Bounding the role of
black carbon in the climate system: A scientific assessment. J Geophys Res Atmos 118: 5380-5552.
http://dx.doi.org/10.1002/igrd.5Q171.
Brauer. M: Amann. M: Burnett. RT: Cohea A: Dentener. F: Ezzati. M: Hendersoa SB: Krzvzanowski. M:
Martin. RV: VanDingenea R: van Donkelaar. A: Thurstoa 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/10.1021/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.ll.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.
Castellanos. P: Marufu. LT: Poddridge. BG: Taubmaa 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: Mcaueea 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.
Chaa WR: Coha S: Sidheswaran. M: Sullivaa PP: Fisk. WJ. (2014). Contaminant levels, source strengths, and
ventilation rates in California retail stores. Indoor Air 25: 381-392. http://dx.doi.org/10. Ill 1/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.
Chea 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. 1103970.
Chea P: 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.
September 2019
2-157
PRAFT: Po Not Cite or Quote
-------
Cho. S: Mceachern. P: Morris. R: Shah. T: Johnsoa 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/BF00890661.
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.
Cohaa PS: 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: Kleia 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: Lubea 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.el41.
http://dx.doi.Org/10.1016/i.annepidem.2016.ll.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.
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.
Pi. O: 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.
Pionisio. 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.
Puttoa SM: Banks. P: 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: Pownev. N: Johnsoa J: Jimenez. M: Yarwood. G: Morris. R. (2012). Regional and global
modeling estimates of policy relevant background ozone over the United States. Atmos Environ 47: 206-217.
http://dx.doi.Org/10.1016/i.atmosenv.2011.ll.012.
EPA. US. (2018). CMAQ (Version 5.2.1) [Computer Program]: Zenodo. Retrieved from
https ://zenodo .org/record/121260 l#.XGFto VxKi IV
Fadevi. MO: Weschler. CJ: Tham. KW: Wu. WY: Sultaa 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.
September 2019
2-158
PRAFT: Po Not Cite or Quote
-------
Ferreira. J: Rodriguez. A: Monteiro. A: Miranda. Al: 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.
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. Ill l/i.0006-
341X.2005.030821.x.
Gall. ET: Corsi. RL: Siegel. JA. (2011). Barriers and opportunities for passive removal of indoor ozone. Atmos
Environ45: 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.
Garner. GG: Thompsoa AM: Lee. P: Martins. DK. (2015). Evaluation of NAQFC model performance in
forecasting surface ozone during the 2011 DISCOVER-AQ campaign. J Atmos Chem72: 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
PollutRes 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.
Goldmaa GT: Mulholland. JA: Russell. AG: Srivastava. A: Strickland. MJ: Klein. M: Waller. LA: Tolbert. PE:
Edgertoa 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/es 101386r.
Goldmaa GT: Mulholland. JA: Russell. AG: Strickland. MJ: Kleia 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/10.1186/1476-069X-10-61.
September 2019
2-159
DRAFT: Do Not Cite or Quote
-------
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.7b0313Q.
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.
Hao. Y: Flowers. H: Monti. MM: Oualters. 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. uoJen: 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: Vaughaa 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: Toriaa 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/10.1080/10962247.2Q13.811127.
Hu. J: Howard. CJ: Mitloehner. F: Greea PG: Kleemaa 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/es203369p.
Huang. L: Mcdonald-Buller. E: Mcgaughev. G: Kimura. Y: Allea 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: Martia 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-11-22.
September 2019
2-160
DRAFT: Do Not Cite or Quote
-------
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.
Jaffe. DA: Wigder. N: Downey. N: Pfister. G: Bovnard. 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/es402164f.
Jerrett. M: Burnett. RT: Pope. CA. Ill: Ito. K: Thurston. G: Krewski. D: Shi. Y: Calle. E: Thua M. (2009).
Long-term ozone exposure and mortality. N Engl J Med 360: 1085-1095.
http://dx.doi.org/10.1056/NEJMoa08Q3894.
Johnson. T: Capel. J: Ollisoa 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.
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.
Karamchandani. P: Johnsoa 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/sl 1270-
013-1674-2.
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/iierphllQ100983.
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: Odmaa 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.
Lai. D: Karava. P: Chen. O. (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. http://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/10.1088/1748-9326/9/11/114019.
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.
September 2019
2-161
DRAFT: Do Not Cite or Quote
-------
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
ChemPhys 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. 4-295-294-302). Sacramento, CA: California Air Resources Board.
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. i: 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/sll869-014-0256-9.
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: Netherv. 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.
Matichuk. R: Tonnesea 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 Res Atmos 122: 13545-13572.
http://dx.doi.org/10.1002/2017JD027Q57.
McDonald-Buller. EC: Allea DT: Brown. N: Jacob. DJ: Jaffe. D: Kolb. CE: Lefoha 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.
Mufiiz-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
Environ45: 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.
Ngaa F: Byua 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.
September 2019
2-162
DRAFT: Do Not Cite or Quote
-------
Paci. L: Gelfand. AE: Holland. DM. (2013). Spatio-temporal modeling for real-time ozone forecasting. Spat Stat
4: 79-93. htto://dx.doi.org/10.1016/i.spasta.2013.04.003.
Paa L. i: 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: Jeoa 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.
Paa 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.
Paa S: Choi. Y: Rov. 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. InDG 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 Chem70: 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/(SICI) 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 111/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. JR Stat Soc SerC Appl Stat 61: 47-
66. http://dx.doi.org/10.1111/i. 1467-9876.2011.00772.x.
Rim P: 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.2017.11.036.
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.
September 2019
2-163
PRAFT: Po Not Cite or Quote
-------
Samoli. E: Peng. RD: Ramsay. T. im: Touloumi. G: Dominici. F: Atkinsoa RW: Zanobetti. A: Le Tertre. A:
Andersoa 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/sll869-014-025Q-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.
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.
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: Simpsoa 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.
Simoa H: Baker. KR: Akhtar. F: Napelenok. SL: Possiel. N: Wells. B: Timia 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.
Simoa 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-17-10435-2017.
Spalt. EW: Curl. CL: Allea RW: Cohen. M: Adar. SD: Stukovskv. KH: Avol. E: Castro-Diehl. C: Nuna 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.2015.29.
Springs. M: Wells. J. R.: Morrisoa GC. (2011). Reaction rates of ozone and terpenes adsorbed to model indoor
surfaces. Indoor Air 21: 319-327. http://dx.doi.org/10.1111/i. 1600-0668.2010.00707.x.
September 2019
2-164
DRAFT: Do Not Cite or Quote
-------
Steva 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: Goldmaa 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: Cohaa PS: 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: Cohaa PS: Pour-Biazar. A: Lamsal. LN: White. AT: Xiao. X: Zhou. W: Hendersoa BH: Lash. BF.
(2015a). Influence of satellite-derived photolysis rates and NOx emissions on Texas ozone modeling. Atmos
ChemPhys 15: 1601-1619. http://dx.doi.org/10.5194/acp-15-1601-2015.
Tang. Y: Chai. T: Pan. L: Lee. P: Tong. P: Kim. HC: Chea W. (2015b). Using optimal interpolation to
assimilate surface measurements and satellite AOP 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: Chia 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. JP: Marshall. JP. (2015). Twelve-month, 12 km resolution North American WRF-Chem v3.4
air quality simulation: performance evaluation. GMP 8: 957-973. http://dx.doi.org/10.5194/gmd-8-957-2015.
Thompson. TM: Selia NE. (2012). Influence of air quality model resolution on uncertainty associated with
health impacts. Atmos ChemPhys 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). (2011). Air quality modeling technical support document:
Final EGU NESHAP. (EPA-454/R-11-009). U.S. Environmental Protection Agency, Office of Air Quality
Planning and Standards, Air Quality Assessment Pivision.
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 Pevelopment, National Center for Environmental
Assessment-RTP Pivision. http://cfpub.epa.gov/ncea/isa/recordisplav.cfm?deid=247492.
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-P agricultural air quality modeling: Impacts of NH3/H2S gas-phase reactions
and bi-directional exchange of NH3. 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 ChemPhys 12: 10209-10237. http://dx.doi.org/10.5194/acp-12-10209-2Q12.
September 2019
2-165
PRAFT: Po Not Cite or Quote
-------
Wang. M: Keller. JP: Adar. SD: Kim. S: Larson. TV: Olives. C: Sampsoa PD: Sheppard. L: Szpiro. AA: Vedal.
S: Kaufmaa 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: Kleemaa M: Keller. JP: Olives. C: Szpiro. AA: Vedal. S: Kaufmaa 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.5b06001.
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: Coha RD: Chu. H: Zeldia DC: Londoa
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.
Wheeler. AJ: Xu. X: Kulka. R: You. H: Wallace. L: Mallach. G: VanRvswvk. 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: Kleia 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 ChemPhys 15: 2723-2747. http://dx.doi.org/10.5194/acp-15-2723-2015.
Xu. W: Rilev. EA: Austia E: Sasakura. M: Schaal. L: Gould. TR: Hartia 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: Reyes. 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: Reyes. 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.
September 2019
2-166
DRAFT: Do Not Cite or Quote
-------
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/10.1002/2015JDQ23819.
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.
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. http://dx.doi.Org/10.1016/i.atmosenv.2014.04.024.
Ying. O: 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/10.1016/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: Rverson. 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 ChemPhys 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, https://cfpub.epa.gov/si/si public record report.cfm?
Lab=NERL&dirEntrvId=80980.
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.ll.121.
Zhang. HL: Ying. O. (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.
Zhou. W: Cohaa PS: 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.
Zimmermaa N: Presto. AA: Kumar. SPN: Gu. J: Haurvliuk. 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-11-291-
2018.
September 2019
2-167
DRAFT: Do Not Cite or Quote
-------
APPENDIX 3 HEALTH EFFECTS —RESPIRATORY
Stuii/iuiry of ( utisiility Determinations for Short- and l.oii^-Tcrm Ozone
Exposure and Respiratory llffccts
This \ppcudi\ cli;ir;iclcri/cs I lie scicuiilic c\ idenee I li:il supporis c;ius;ihi\
dclcriui ikiI ions lor shori-;uid loim-lcrni ozone exposure ;md respinilorx I ic;i 1111 clfccls. I lie
l\ pes ii|" studies c\ ;ilu;ilcd w illiin llns \ppcudi\ ;irc cousisieul w illi I lie o\ei~il I scope of ilie ISA
;is (,lel;nleel in ilie IJiel":iee In ;isscssiuu ilie enill e\ idenee. sirenullis ;md liniii;iiious of
iiidi\ idn;il siudies were c\ ;ilu;ilcd h;ised mi seienl i lie cousidenilious del;nled in I lie \mie\ I'm'
\nneiidi\ V More del;nls on lhe e;ins;il fr;uncwork used lo re;ieli iliese eoiielnsions ;ire included
in ilie Pre;imhle lo ilie IS \ il S IT \. 2<>151 I lie c\ idenee preseuled lliroimlioui llns
\ppeiidi\ snppori ilie follow iiiu c;ius;iln\ conclusions
ll\|iiisiiiv Diiniiiiui
( iiusiilii\ IK'U'rmin;iliiin
Short-term exposure
Cuusul
Long-term exposure
Likely I
o be aius;il
3.1 Short-Term Ozone Exposure
3.1.1 Introduction, Summary from the 2013 Ozone ISA, and Scope for Current
Review
1 The 2013 Ozone ISA concluded that "short-term ozone exposure is causally associated with
2 respiratory health effects" [see Chapter 6 of (U.S. EPA. 2013aYI. This conclusion was based largely on
3 controlled human exposure studies demonstrating ozone-related respiratory effects in healthy individuals.
4 Specifically, statistically significant decreases in group mean pulmonary function relative to ozone
5 exposures as low as 60 ppb were observed in young, healthy adults. Additionally, controlled human
6 exposure and toxicological studies demonstrated ozone-induced increases in respiratory symptoms, lung
7 inflammation, airway permeability, and airway responsiveness. The experimental evidence was supported
8 by strong evidence from epidemiologic studies. Specifically, these studies demonstrated associations
9 between ozone concentrations and respiratory hospital admissions and emergency department (ED) visits
10 across the U.S., Europe, and Canada. Most effect estimates ranged from a 1.6 to 5.4% increase in daily
September 2019
3-1
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
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.
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
intra-cellular 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).
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
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.
September 2019
3-2
DRAFT: Do Not Cite or Quote
-------
1 of relevant issues for interpreting the epidemiologic evidence discussed in the preceding sections.
2 Throughout the sections on respiratory health effects, results from recent studies are evaluated in the
3 context of evidence provided by studies that were previously evaluated in the 2013 Ozone ISA (U.S.
4 EPA. 2013a). Study-specific details, including exposure time periods and exposure concentrations in
5 experimental studies, and study design, averaging times, and select results in epidemiologic studies are
6 presented in evidence inventories in Section 3.3.
3.1.2 Population, Exposure, Comparison, Outcome, and Study Design
(PECOS) Tool
7 The scope of this section is defined by a scoping tool that generally describes the relevant
8 Population, Exposure, Comparison, Outcome, and Study Design (PECOS). The PECOS tool defines the
9 parameters and provides a framework to help identify the relevant literature to inform the draft 2019
10 Ozone ISA. Because the 2013 Ozone ISA concluded there is a causal relationship between short-term
11 ozone exposure and respiratory health effects, the recent epidemiologic studies evaluated in this ISA are
12 limited to study locations in the U.S. and Canada to provide a focus on study populations and air quality
13 characteristics that are most relevant to circumstances in the U.S. The studies evaluated and subsequently
14 discussed within this section were included if they satisfied all of the components of the following
15 PECOS tool:
16 Experimental studies:
17 • Population: Study populations of any controlled human exposure or animal toxicological study of
18 mammals at any lifestage
19 • Exposure: Short-term (on the order of minutes to weeks) inhalation exposure to relevant ozone
20 concentrations (i.e., <0.4 ppm for humans, <2 ppm for other mammals); while ozone
21 concentrations in animal toxicological studies appear high, it should be noted that deposition of
22 ozone resulting from exposure to 2 ppm ozone in a resting rat is roughly equivalent to deposition
23 of ozone resulting from exposure to 0.4 ppm ozone in an exercising human (Hatch et al.. 1994).
24 • Comparison: Human subjects serve as their own controls with an appropriate washout period or
25 groups may be compared at the same or varied exposure concentrations; or, in toxicological
26 studies of mammals, an appropriate comparison group is exposed to a negative control (i.e., clean
27 air or filtered-air control)
28 • Outcome: Respiratory effects
29 • Study Design: Controlled human exposure studies and animal studies meeting the above criteria
30 Epidemiologic studies:
31 • Population: Any U.S. or Canadian population, including populations or lifestages that might be at
32 increased risk
33 • Exposure: Short-term exposure (on the order of hours to several days) to ambient concentrations
34 of ozone
September 2019
3-3
DRAFT: Do Not Cite or Quote
-------
1 • Comparison: Per unit increase (in ppb), or humans exposed to lower levels of ozone compared
2 with humans exposed to higher levels
3 • Outcome: Change in risk (incidence/prevalence) of respiratory effects
4 • Study Design: Epidemiologic studies consisting of panel, case-crossover, time-series studies, and
5 case-control studies, as well as cross-sectional studies with appropriate timing of exposure for the
6 health endpoint of interest
3.1.3 Biological Plausibility
7 This section describes biological pathways that potentially underlie respiratory health effects
8 resulting from short-term exposure to ozone. Figure 3-1 graphically depicts the proposed pathways as a
9 continuum of upstream events, connected by arrows, that may lead to downstream events observed in
10 epidemiologic studies. This discussion of how short-term exposure to ozone may lead to respiratory
11 health effects contributes to an understanding of the biological plausibility of epidemiologic results
12 evaluated later in Section 3.1.4 to Section 3.1.9.
13 Evidence that short-term exposure to ozone may affect the respiratory tract generally informs two
14 proposed pathways (Figure 3-1). The first pathway begins with the activation of sensory nerves in the
15 respiratory tract that can trigger local reflex responses and transmit signals to regions of the central
16 nervous system that regulate autonomic outflow. The second pathway begins with injury, inflammation,
17 and oxidative stress responses, which are difficult to disentangle. Inflammation generally occurs as a
18 consequence of injury and oxidative stress, but it can also lead to further oxidative stress and injury due to
19 secondary production of reactive oxygen species (ROS) by inflammatory cells.
September 2019
3-4
DRAFT: Do Not Cite or Quote
-------
Altered Heart
Rhythm
Decreased
Inspiratory
Capacity/
Pain on
Inspiration
Short
Exposure
Decrements in Lung
Function (FVC,
FEV,) and
Increased
Respiratory
Symptoms
Emergency
Department Visits/
Hospital
Admissions
Asthma
Exacerbation
Emergency
Department Visits/
Hospital
Admissions
Respiratory
Infections
A
\
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.
Figure 3-1 Potential biological pathways for respiratory effects following
short-term ozone exposure.
Activation of Sensory Nerves in the Respiratory Tract
1 Airway sensory nerves in the lower respiratory tract are vagal afferents that carry signals to the
2 nucleus tractus solitaries in the brain. Signals are integrated in the brain and may result in altered
3 autonomic activity that affects the lung (e.g., airway obstruction) or other organs (e.g., altered heart
4 rhythm). In addition, activation of some types of sensory nerves (e.g., C-fibers) leads to local axon reflex
5 responses in the airways that result in altered ventilatory parameters (e.g., altered breathing frequency and
6 inspiratory capacity) and airway obstruction. The release of substance P or other tachykinins from
7 C-fibers and subsequent binding to neurokinin receptors in the airway has been identified as a mechanism
September 2019
3-5
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
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 3-1V In studies using pharmacological tools, nociceptive sensory nerves,
presumably bronchial and pulmonary C-flbers, 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
ct 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
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
September 2019
3-6
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
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 (Verhein 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; Schelegle 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.
September 2019
3-7
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
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 ones published more recently,
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).
September 2019
3-8
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
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
costimulation 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 etal..
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.. 2012). 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.. 2016). 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 was 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
September 2019
3-9
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
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. 1997V 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.
September 2019
3-10
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
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 at 40 L/minute) and 300 ppb O3 (2 hours, heavy intermittent exercise at 70 L/minute) (Adams.
2003a. b, 2002). 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 step-wise 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 versus
step-wise) 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.
September 2019
3-11
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
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 periods). 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 (square-wave) ozone for 6.6 hours with moderate 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.
September 2019
3-12
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
~ 0 ppb Q40 ppb
0% 0%
60 ppb 1180 ppb
10% 30% with >10% FEVi
>. 0.5
O" 0.3 -¦
-10 to -5 -5 to 0 0 to 5 5 to 10 10 to 15
FEN^ decrement (%)
>15
The illustrated data are for 30 subjects in the study conducted by Adams (2006). FE\A decrements following each exposure were
calculated as pre-exposure FE\A| minus post-exposure FE\A| then divided by the pre-exposure FE\A|. The FE\A decrements for
filtered air (0 ppb ozone) were subtracted from the FE\A| 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 one 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 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 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.
September 2019
3-13
DRAFT: Do Not Cite or Quote
-------
16 1
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 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 FE\A| 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 60 ppb, statistically significant FE\A| responses to square-wave chamber exposures were found by Kim et
al. (2011) and in the Adams (2006) study based on the analysis of Brown et al. (2008). With the exception of the Scheleqle et al.
(2009) data, the data at 60, 80, and 120 ppb have been offset for illustrative purposes. The McDonnell et al. (2013) line illustrates
the predicted FE\A| 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. (2011) study, but only published in Figure 5 of McDonnell et al. (2012).
Adapted from Figure 6-1 of 2013 Ozone ISA (U.S. EPA. 2013a). Studies appearing in the figure legend are: Adams (2006). Adams
(2003a). Adams (2002). Adams (2000). Adams and Ollison (1997). Folinsbee et al. (1994). Folinsbee et al. (1988). Horstman et al.
(1990). Kim et al. (2011). McDonnell et al. (2013). McDonnell et al. (1991). and Scheleqle et al. (2009). *80 ppb data for 30 health
subjects were collected as part of the Kim et al. (2011) study, but only published in Figure 5 of McDonnell et al. (2012).
Figure 3-3 Cross-study comparisons of mean ozone-induced forced
expiratory volume in one 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 ^
N -
o >
LU
1 Since the 2013 Ozone ISA, one new study has evaluated the effect of ozone on lung function at
2 concentrations below 80 ppb. The results of this study of older adults (55-70 years) exposed for 3 hours
3 to 0, 70, and 120 ppb ozone appear in both an HEI report (Frampton et al.. 2017) and in the scientific
4 literature (Ariomandi et al.. 2018) and are discussed in a subsection on lifestage (Section 3.1.4.1.1.1).
5 Several new studies have investigated the effects of 100-300 ppb ozone exposure on lung function
6 [e.g., Biller et al. (2011). Ghio et al. (2014). Hoffmever et al. (2013). Madden et al. (2014). Stiegel et al.
7 (2017). Tank et al. (2011)1. Given that lower ambient concentrations are more common currently, any
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)*
o
McDonnell et al. (1991)
~
Schelegle et al. (2009)
-McDonnell et al. (2013)
— m
— s,m
September 2019
3-14
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
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) an or 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). 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. 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 overtime 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 etal. (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, i.e., 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
FEVi responses. Two new models (Hsieh 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 dataset
consisting of the FEVi responses of 741 young healthy adults (104 F, 637 M; mean age 23.8 years) from
September 2019
3-15
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
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-subj ect 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 FEV i response distribution to have improvements in FEV i
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 et al. (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. 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 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
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:
September 2019
3-16
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
• Hoffmever et al. (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,p = 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 (19 F; 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 et al.. 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°C 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 compared to 22°C.
• 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.
September 2019
3-17
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
• 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. 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; Arjomandi 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 p. 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 Mest: 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,
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 Arjomandi et al. (2018) are small, 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 and brief 3 hour duration of exposure. 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
September 2019
3-18
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
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 and showed that lung function decrements decline with age, 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 the
ambient environment 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 exposure
to 2 ppm ozone in a resting rat is roughly equivalent to deposition of ozone resulting from exposure to
0.4 ppm ozone in an exercising human (Hatch et al.. 1994).
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
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
September 2019
3-19
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
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,
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; Schelegle 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.. 2016).
o Increased enhanced pause—1 ppm ozone for 4 hours/day for 1 and 2 days (Miller et al..
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).
o Increased peak expiratory flow and enhanced pause—0.8 ppm ozone for 4 hours/day for
1 and 2 days (Henriquez 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 lifestages.
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.
September 2019
3-20
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
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; A vol 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.
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 (Lepeule et
al.. 2014; Rice et al.. 2013). These studies included both healthy participants and those with
September 2019
3-21
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
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 relative to ozone exposures as low as 60 to 70 ppb in young, healthy adults performing moderate
exercise. One recent study observed a small, but not statistically significant, group decrease in
post-exercise lung function in older adults following 70 ppb exposure.
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
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.
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.
September 2019
3-22
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
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 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 concentration
where statistically significant ozone-induced lung function decrements and subjective symptoms have
been reported (Schelegle 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, it is difficult or impossible to assess respiratory symptoms such as
pain on deep inspiration, shortness of breath, and cough in rodents. Rodents are obligate nasal breathers
September 2019
3-23
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
and in general do not cough. However, changes in ventilation may be consistent with dyspnea. Further,
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).
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 (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 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.
September 2019
3-24
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
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.
3.1.4.3.2 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,
September 2019
3-25
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
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 exposure
to 2 ppm ozone in a resting rat is roughly equivalent to deposition of ozone resulting from exposure to
0.4 ppm ozone in an exercising human (Hatch et al.. 1994). 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
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 (Verhein 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;
September 2019
3-26
DRAFT: Do Not Cite or Quote
-------
1 Cho et al.. 2013; Sunil et al.. 2013); however, no increases were observed in Mathews et al.
2 (2017b) or Choetal. (2013V
3 o This response was persistent over time in Sunil et al. (2013).
4 o Several studies provide evidence for cell signaling and other pathways underlying
5 increases in airway responsiveness resulting from acute ozone exposure.
6 ¦ TNF-stimulated gene 6 and hyaluronan-heavy chain complexes (Stober et al..
7 2017)
8 ¦ Rho-associated coiled-coil-containing protein kinase [ROCK, (Kasahara et al..
9 2015)1
10 ¦ Dietary short chain fatty acids/gut microbiome (Cho et al.. 2018)
11 ¦ IL-17 (Mathews et al.. 2018)
12 ¦ Osteopontin (Barreno et al.. 2013)
13 o Evidence for ozone exposure-induced release of tachykinins is provided by Barker et al.
14 (2015). Acute ozone exposure (2 ppm for 3 hours) increased levels of the tachykinin
15 substance P levels in the BALF through upstream effects on IL-ip and nerve growth
16 factor.
17 • Exposure to 2 ppm for 4 hours increased airway responsiveness measured as increased
18 bronchoconstriction in response to electrical stimulation of the vagal nerve (Verhein et al.. 2013).
19 Both decreased function of M2 muscarinic receptors and involvement of the p38/JNK pathway
20 were implicated in this response.
21 • Acute exposure to 0.8-1 ppm ozone for 1-6 hours resulted in increased airway responsiveness
22 (Zvchowski et al.. 2016; Groves et al.. 2012). Rho kinase was implicated in this response
23 (Zvchowski et al.. 2016).
24 • Repeated exposure to 1 ppm ozone (3 hours/day for 7 days) resulted in increased airway
25 responsiveness (Zhu et al.. 2016). This response was blocked by treatment with Vitamin E,
26 implicating oxidative stress in mediating ozone-induced increased airway responsiveness.
27 • Acute and repeated exposures to 0.25 and 0.5 ppm ozone did not result in increases in airway
28 responsiveness.
3.1.4.3.3 Integrated Summary for Airway Responsiveness
29 Controlled human exposure studies and animal toxicological studies evaluated in the 2006 Ozone
30 AQCD (U.S. EPA. 2006) and the 2013 Ozone ISA (U.S. EPA. 2013a) provide consistent evidence of
31 ozone-induced increases in airway responsiveness. In experimental studies in humans, changes in airway
32 responsiveness were less transient than the observed ozone-related lung function changes discussed in
33 Section 3.1.4.1.1. One recent study of healthy adult women showed a tendency for increased airway
34 responsiveness following ozone exposure. In recent experimental animal studies, increases in airway
35 responsiveness resulted from ozone exposures in the range of 0.8 to 2 ppm, but not in response to acute
36 and repeated exposures of 0.25 and 0.5 ppm. Mechanistic studies provide evidence that local reflex
37 responses and activation of parasympathetic pathways mediate increases in airway responsiveness due to
September 2019
3-27
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
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
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 p. 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.
September 2019
3-28
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
• Alexis et al. (2013) conducted a post hoc analysis of sputum PMN collected by Kim etal. (2011)
from 24 healthy adults (20-33 years) 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 Ariomandi 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 hour post-exposure in either bronchial wash (BW) fluid or BALF. Significant
PMN increases were apparent at 6 hours in both the BW (4 times,/? < 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,
p < 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 (Tank etal.. 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
these studies reported that %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.
September 2019
3-29
DRAFT: Do Not Cite or Quote
-------
3.1.4.4.1.1 Factors Affecting Pulmonary Inflammation, Injury, and Oxidative Stress
Ambient Temperature
1 Gomes et al. (201 lb) exposed nine male endurance runners (24 ± 6 years) to 0 and 100 ppb ozone
2 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
3 completed four exposures). Nasal lavage was conducted approximately 15 minutes post-exposure. There
4 were no differences in inflammatory markers among the exposures. Although there were no differences
5 between the heat only or ozone only compared to control, levels of nasal Club cells following the
6 high-temperature ozone exposure were significantly increased (p = 0.03) relative to the lower temperature
7 air control. Glutathione concentrations were also significantly increased (p = 0.001) following the
8 high-temperature ozone exposure relative to the lower temperature air control. The increases in Club cells
9 and glutathione appeared to be additive, but no trend analysis was reported.
Lifestage
10 As reported in the 1996 and 2006 Ozone AQCDs (U.S. EPA. 2006. 1996a). decrements in lung
11 function and increases in respiratory symptoms in response to ozone exposure decreased with increasing
12 age. However, whether inflammatory responses persisted with increasing age remained unstudied at the
13 time of the 2013 Ozone ISA (U.S. EPA. 2013a). Two recent studies demonstrated inflammatory
14 responses in older adults.
15 • Arjomandi et al. (2018) investigated changes in sputum markers of inflammation and injury in
16 healthy older adults (52 F, 35 M; 59.9 ± 4.5 years) exposed to 0, 70, and 120 ppb ozone for
17 3 hours during light to moderate, intermittent exercise. Sputum samples were obtained 22.5 hours
18 post-exposure. A mixed effects model showed marginally significant (p = 0.012)
19 concentration-dependent increases in PMNs by 4.1% of total (n.s.; p = 0.134) and 8.2% of total
20 (0.003) following 70 and 120 ppb ozone exposures, respectively. Sputum PMN increases
21 following ozone exposure showed no interaction with sex (52 F, 35 M), age (55-70 years), or
22 GSTM1 genotype (57% null, 43% positive). Due to the activity level and duration of exposure,
23 the total delivered ozone dose (120 ppb exposure) was estimated by Arjomandi et al. (2018) to be
24 about 60% of the delivered dose in the Kim etal. (2011) study, which identified a significant
25 increase in sputum PMN in young healthy adults following exposure to 60 ppb ozone. Sputum
26 IL-6, IL-8, TNF-a, and total protein concentrations did not show any significant changes due to
27 ozone exposure.
28 • Kirsten et al. (2011) studied Bimosiamose effectiveness in mitigating PMN response in healthy
29 older subjects (3 F, 15 M; 43.9 ± 7.4 years) who were found to be responsive (>20% increase in
30 sputum PMN) following exposure to 250 ppb ozone (no air control) for 3 hours with intermittent
31 exercise (alternating 15 minutes intervals of rest and exercise at 14 L/minute per m2 BSA).
32 Sputum was collected 3 hour post-exposure. Another nine individuals (age and sex not specified)
33 were also exposed to ozone, but did not experience a sufficient increase in sputum PMN for
34 inclusion in the drug trial. Bimosiamose pretreatment of the 18 PMN responders reduced PMN
35 after ozone exposure to approximately the pre-exposure baseline. This study shows that 2/3 of the
September 2019
3-30
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
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). These the studies discussed here and prior studies of younger
adults prevent quantification of PMN responses as a function of age; these 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
monkeys. Mucous cell metaplasia of nasal epithelium has been observed in both rodents and monkeys
exposed to ozone over several days.
September 2019
3-31
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
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) (a collectin), 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
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
September 2019
3-32
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
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 ofNFKB, 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.. 2011V
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.
September 2019
3-33
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
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 (C'ho et al.. 2013)1
¦ Osteopontin (Barrcno 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.
¦ 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).
September 2019
3-34
DRAFT: Do Not Cite or Quote
-------
1 ¦ Resolution of inflammation and injury within 72 hours following ozone exposure
2 was demonstrated (Groves et al.. 2012). However, airway resistance remained
3 increased, indicating that the lung was functionally compromised.
4 o Several studies focused on the accumulation of macrophages in the lung in response to
5 ozone exposure (Francis et al.. 2017b; Sunil et al.. 2015; Groves et al.. 2013).
6 ¦ Resident alveolar macrophages were not affected, but infiltrating monocytic and
7 granulocytic cells were increased.
8 ¦ Increases in both classically activated macrophages (Ml, proinflammatory) and
9 alternatively activated macrophages (M2, anti-inflammatory) were found.
10 "A time course study showed that Ml macrophages increased in number rapidly
11 and persisted for 72 hours post-exposure, while M2 macrophages were increased
12 beginning at 72 hours.
13 ¦ The spleen was found to be a source for these Ml and M2 cells.
14 o Some studies provide evidence for cell signaling and other pathways underlying the
15 inflammatory, injury, or oxidative effects of acute ozone exposure.
16 ¦ CD36 (Robertson et al.. 2013)
17 ¦ IL-33 and ST2 (Michaudel et al.. 2018)
18 ¦ Glucocorticoids (Thomson et al.. 2016)
19 ¦ TLR4 (Connor et al.. 2012)
20 ¦ Galectin (Sunil et al.. 2015)
21 ¦ C-C chemokine receptor type 2 (CCR2) (Francis et al.. 2017a)
22 o Other endpoints examined include mucus secretion (Gabehart et al.. 2015) and
23 upregulation of glucocorticoid-sensitive genes (Thomson et al.. 2016; Thomson et al..
24 2013).
25 ¦ Mucus secretion was not seen in juvenile or adult mice in response to ozone.
26 ¦ Transient changes in glucocorticoid-sensitive genes occurred immediately after
27 exposure to ozone.
28 o One study provides evidence for respiratory effects of acute ozone exposure in rodents of
29 varying lifestages. In Gabehart et al. (2015). inflammation and injury responses in 2-, 3-,
30 and 6-week-old mice, representing weanling, juvenile, and adult stages, respectively,
31 were examined. Results in 1-week-old mice (neonates) similarly exposed are discussed in
32 the long-term exposure section of this Appendix. In general, responses were smallest in
33 1-week-old mice and greatest in 6-week-old mice, with responses in the 2- and
34 3-week-old mice sometimes in between and sometimes as high as responses in the
35 6-week-old mice. The exception was mucus secretion, which was highest in 1-week-old
36 mice and minimal in 2-week-old mice.
37 • Acute exposure to 0.25-0.5 ppm ozone resulted in minimal or no changes in inflammation,
38 injury, or oxidative stress markers in BALF (Michaudel et al.. 2018; Kodavanti et al.. 2015;
39 Kumarathasan et al.. 2015; Kurhanewicz et al.. 2014; Mclntosh-Kastrinskv et al.. 2013; Thomson
40 et al.. 2013). Histopathological lesions were seen in response to 0.25 and 0.5 ppm ozone (Ramot
41 et al.. 2015).
September 2019
3-35
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
• 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 et al.. 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 etal.. 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).
o Some studies of repeated ozone exposures provide evidence for cell signaling and other
pathways underlying inflammation, injury, or oxidative stress.
September 2019
3-36
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
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 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.
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.
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
September 2019
3-37
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
and airway obstruction may share similar underlying mechanisms, while inflammation and lung volume
(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.
Recent studies provide some further evidence that GSTMl-null individuals are more susceptible to
ozone-related inflammatory responses, although the evidence is not entirely consistent.
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. Recent studies are consistent with
previous findings and expand on observed interindividual variability in inflammatory responses,
providing additional evidence that GSTMl-null individuals are more susceptible to ozone-related
inflammatory responses.
September 2019
3-38
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
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 panel studies of children in summer camps
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 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,
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
September 2019
3-39
DRAFT: Do Not Cite or Quote
-------
1 visits for asthma, respiratory symptoms and asthma medication use, lung function, and subclinical effects,
2 such as pulmonary inflammation and oxidative stress, in people with asthma. ED visits for asthma are
3 more common and often less serious than hospital admissions. Generally, only a small fraction of
4 respiratory ED visits result in a hospital admission. Accordingly, the two outcomes may reflect different
5 severities of asthma and are evaluated separately.
3.1.5.1 Hospital Admissions
6 A single study evaluated in the 2013 Ozone ISA (U.S. EPA. 2013a) examined the association
7 between short-term exposure to ozone and hospital admissions for asthma. In New York City, 8-hour max
8 ozone concentrations were associated with severe acute asthma admissions in the warm season
9 (Silverman and Ito. 2010). The authors reported positive associations with non-ICU asthma admissions
10 that were strongest (i.e., of greatest magnitude) for children ages 6 to 18 years, compared to the other age
11 groups examined (ages <6 years, 19-49, 50+, and all ages). The observed effect remained robust to
12 adjustment for PM25. The authors also performed an analysis examining the shape of the
13 concentration-response (C-R) relationship, which is discussed in more detail in Section 3.1.10.4.
14 Recent studies expand the existing evidence base and provide consistent evidence of an
15 association between ozone and hospital admissions for asthma (Figure 3-4). Generally, the evaluated
16 studies use 8-hour daily max averaging times, although there are studies that use daily 8-hour avg (Zu et
17 al.. 2017) and 24-hour avg (Shmool et al.. 2016). The averaging time used in each study, along with other
18 study-specific details, including air quality characteristics and select effect estimates, are highlighted in
19 Table 3-13 in Section 3.3.1. An overview of the evidence is provided below.
20 • Multicity studies in Texas (Goodman et al.. 2017b; Zu et al.. 2017) and single-city studies in New
21 York City (Goodman et al.. 2017a; Shmool et al.. 2016; Sheffield et al.. 2015) and St. Louis, MO
22 (Winquist et al.. 2012) reported evidence of an association between short-term ozone
23 concentrations and hospital admissions for asthma.
24 • Shmool et al. (2016) compared monitor-based ozone concentrations to ozone estimated at a
25 300-m spatial scale using a fusion of monitoring data and land-use regression (LUR). In short, the
26 authors used LUR with local monitoring inputs to estimate seasonal average concentrations
27 within 300 m radial buffers around geocoded participant addresses. The ratio of these
28 spatially-resolved seasonal average concentrations and the citywide averages were multiplied by
29 daily monitor averages to estimate spatially-refined daily exposures. The effect estimates derived
30 from the spatiotemporal model were similar to those estimated using monitored ozone
31 concentrations. These results indicate that the observed association of ozone concentrations with
32 asthma hospital admissions is robust to exposure assignment technique.
33 • Like previous findings from Silverman and Ito (2010). recent studies that reported age-stratified
34 results (Goodman et al.. 2017b; Goodman et al.. 2017a; Zu et al.. 2017; Sheffield et al.. 2015)
35 generally observed ozone-asthma hospital admission associations that were strongest (i.e., greater
36 in magnitude) in younger populations (5 to 18 years of age). Many studies exclude data for
37 children less than 5 years of age due to less reliable asthma diagnosis in young children.
38 Additionally, most studies that examined hospital admissions in adults older than 50 reported null
September 2019
3-40
DRAFT: Do Not Cite or Quote
-------
1 associations. While most studies observed associations in analyses of all ages combined, stratified
2 analyses suggest that these associations are likely being driven by hospital admissions among
3 children.
4 • In recent studies, there was some limited evaluation of the shape of the C-R relationship (Zu et
5 al.. 2017). potential copollutant confounding (Shmool et al.. 2016). and seasonal differences in
6 effect estimates (Goodman et al.. 2017a) across the evaluated studies. These topics are discussed
7 in more detail in the Relevant Issues for Interpreting Epidemiologic Evidence section
8 (Section 3.1.10).
Study
Location
Ages
Silverman et al. (2010)
New York, NY
<6
fGoodman etal. (2017)
New York, NY
<6
fZu et al. (2017)
6 Texas Cities
5-14
fGoodman etal. (2017)
3 Texas Cities
5-14
Silverman et al. (2010)
New York, NY
6-18
fGoodman etal. (2017)
New York, NY
6-18
fZu et al. (2017)
6 Texas Cities
15-64
fGoodman etal. (2017)
3 Texas Cities
15-64
Silverman et al. (2010)
New York, NY
19-49
fGoodman etal. (2017)
New York, NY
19-49
Silverman et al. (2010)
New York, NY
50+
fGoodman etal. (2017)
New York, NY
50+
fZu et al. (2017)
6 Texas Cities
65+
fGoodman etal. (2017)
3 Texas Cities
65+
fWinquistet al. (2012)
St. Louis, MO
All
fZu etal. (2017)
6 Texas Cities
All (5+)
fGoodman etal. (2017)
3 Texas Cities
All (5+)
Silverman et al. (2010)
New York, NY
All
fGoodman etal. (2017)
New York, NY
All
Mean
Concentrations3
41
30.7
32.2; 8-hr avg
41.8
41
30.7
32.2; 8-hr avg
41.8
41
30.7
41
30.7
32.2; 8-hr avg
41.8
41
32.2; 8-hr avg
41.8
41
30.7
Season
Warm
Warm
Year-Round
Year-Round
Warm
Warm
Year-Round
Year-Round
Warm
Warm
Warm
Warm
Year-Round
Year-Round
Year-Round
Year-Round
Year-Round
Warm
Warm
Lag
0-1
0-1
0-3
0
0-1
0-1
0-3
0
0-1 ¦*-
0-1
0-1
0-1
0-3
0
0-4 DL
0-3
0
0-1
0-1
-+-
0.9 1 1.1 1.2 1.3
Effect Estimate (95% CI)
DL = distributed lag.
Note: f Studies published since the 2013 Ozone ISA. Black text = studies included in the 2013 Ozone ISA.
aMean concentrations reported in ppb and are for 8-hour daily max averaging times 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.
September 2019
3-41
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
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
ED visits. The observed associations were consistently stronger in magnitude in the warm season.
Additionally, Ito et al. (2007) reported an association that was robust in copollutant models that adjusted
for PM2 5, NO2, SO2, 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 8-hour daily max averaging times, although there are instances in which the 1-hour
daily max (Malig etal.. 2016) and 24-hour avg (Szvszkowicz et al.. 2018; Sarnatetal.. 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 et al.. 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).
September 2019
3-42
DRAFT: Do Not Cite or Quote
-------
Mean
Study
Location
Ages
Concentration8
Lag
Season
i
i
|
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
>-
| -©¦
tSacksetal. (2014)
North Carolina
All
43.6
0-2
!~
tSarnat et al. (2013)
Atlanta, GA
All
41.9; 24-hr avg
0-2
i
i«
i
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
!-©-
lo-
<-©-
fSarnat et al. (2015)
St. Louis, MO
All
36.2
0-2 DL
I
tAlhanti etal. (2016)
3 U.S. Cities
0-4
37.3-43.7
0-2
:»
tStricklandetal. (2014)
Atlanta, GA
2-16
42.2
0-2
i
1 -o-
1
tXiao etal. (2016)
Georgia
2-18
42.1
0-3
;o
tO'Lenicket al. (2017)
Atlanta, GA
5-17
NR
0-2
! -©•
tAlhanti etal. (2016)
3 U.S. Cities
5-18
37.3-43.7
0-2
1
i -©•
1
fSzyszkowicz et al. (2018)
Multicity, Canada
<20; Females
<20; Males
22.5-29.2; 24-hr avg
1
1
rO
tAlhanti etal. (2016)
3 U.S. Cities
19-39
40-64
65+
37.3-43.7
0-2
1
H©-
V
1
tMalig etal. (2016)
California
All
33-55; 1-hr max
0-1
Warm
1
1 -©¦
tSacks et al. (2014)
North Carolina
All
50.1
0-2
k>
tByers et al. (2015)
Indianapolis, IN
All (5+)
48.5
0-2
i
+-•-
i
tGleason et al. (2014)
New Jersey
3-17
NR
0-2
1 ©¦
tWinquistet al. (2014)
Atlanta, GA
5-17
53.9
0-2
!-•-
tSheffieldetal. (2015)
New York, NY
5-17
NR
0-3
i
tByers et al. (2015)
Indianapolis, IN
5-17
18-44
45+
48.5
0-2
~r~°—
i—®—
-®-j—
tWinquistet al. (2014)
Atlanta, GA
5-17
53.9
0-2
Cold
r 1 r
0.9 1 1.1 1.2
Effect Estimate (95% CI)
Note: fStudies published since the 2013 Ozone ISA.
aMean concentrations reported in ppb and are for 8-hour daily max averaging times 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.
September 2019
3-43
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
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 support for the
observed 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 ofbreath, 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 an increased incidence of wheeze
in subjects with asthma exposed for 7.6 hours with light quasi-continuous exercise to 160 ppb. 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 3.1.4.1 of the 2013 Ozone ISA). Methodological distinctions in the evaluated multicity studies,
including lack of power and extended averaging times (e.g., 19-day averages), 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
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.
September 2019
3-44
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
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 %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 (Ariomandi 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 Ariomandi et al. (2015) nor Fry et al. (2012) reported FEVi responses to ozone
differentiated by the presence of asthma.
• 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) and a lack of inhaled corticosteroid
treatment (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
September 2019
3-45
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
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 (Lerov 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.
• Schclcglc 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.
3.1.5.4.3 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..
September 2019
3-46
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
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
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
September 2019
3-47
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
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.
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 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)!.
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.
3.1.5.5.2 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
effects 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.
• Schclcglc and Walbv (2012) evaluated the role of vagal afferents in mediating ozone-induced
increased airway responsiveness to allergen. Direct measurements of airway resistance were
September 2019
3-48
DRAFT: Do Not Cite or Quote
-------
1 made in naive and allergic rats (sensitized and challenged with nDer f 1) following ozone
2 exposure (1 ppm, 8 hours). Ozone exposure enhanced allergen-induced airway resistance
3 (i.e., increase in specific airway responsiveness) in allergic rats to a greater degree than in naive
4 rats. This was an early airway response; no late airway response was observed. The mechanisms
5 underlying this response were explored using vagotomy and pharmacological agents and
6 demonstrated the involvement of vagal C-fibers, vagal myelinated fibers, and possibly
7 neuropeptides released in the airway. Results indicated that vagal lung C-fibers mediated the
8 enhanced specific airway reactivity (to the allergen). Neuropeptides (e.g., substance P) may also
9 be involved in the bronchoconstrictive response to allergen.
3.1.5.5.3 Integrated Summary for Airway Responsiveness
10 Controlled human exposure studies previously evaluated in the 2013 Ozone ISA indicate that
11 individuals with and without asthma exhibit similar relative increases in ozone-induced airway
12 responsiveness. However, in general individuals with asthma having greater baseline airway
13 responsiveness than individuals without asthma. Increased airway responsiveness can result in the
14 narrowing of airways upon inhalation of a variety of stimuli, providing biological plausibility for
15 epidemiologic associations observed between increases in ozone and asthma exacerbation (i.e., hospital
16 admissions and ED visits for asthma and prevalence of respiratory symptoms in children with asthma).
17 No recent controlled human exposure studies or epidemiologic studies were identified for review.
18 Consistent with previously reviewed experimental studies in humans, animal toxicological studies
19 reviewed in the 2013 Ozone ISA observed increased airway responsiveness to inhaled allergens in
20 allergen-sensitized animal models. A recent study also found that ozone exposure resulted in enhanced
21 airway responsiveness in allergic mice compared to naive mice.
3.1.5.6 Respiratory Tract Inflammation, Injury, and Oxidative Stress
3.1.5.6.1 Controlled Human Exposure Studies
22 Studies reviewed in Section AX6.9.3 of the 2006 Ozone AQCD (U.S. EPA. 2006) and carried
23 forward into Section 6.2.3.1 starting on p. 6-77 of the 2013 Ozone ISA (U.S. EPA. 2013a') showed greater
24 ozone-induced neutrophilic responses in lavage samples collected at 18 hours post-exposure from
25 individuals with asthma than without asthma. Specifically, two studies showed that individuals with
26 asthma exposed to 200 ppb ozone for 4-6 hours with exercise exhibited significantly more neutrophils in
27 BALF (18 hours post-exposure) than similarly exposed healthy individuals. In another study, when lavage
28 samples were collected at 6 hours following a 2-hour exposure with exercise to 200 ppb ozone, there were
29 no observed differences in inflammatory responses between those with and without asthma. However, the
30 subjects with asthma were on average 5 years older than the healthy subjects in this study, and it is still
September 2019
3-49
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
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.
• 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, the authors concluded that allergens
exaggerate the response to ozone. However, the 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.
• (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.
September 2019
3-50
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
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.
• Schelegle 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).
3.1.5.6.3 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-Villarreal 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.
September 2019
3-51
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
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). See Table 3-25 in Section 3.3.1 for complete study details.
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
available for review reported a null association between ozone and FeNO 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.
September 2019
3-52
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
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 Chronic Obstructive Pulmonary Disease
Chronic obstructive pulmonary disease (COPD) is a chronic lung disorder characterized by
destruction of alveolar tissue, airway remodeling, and minimally reversible airflow limitation. Reduced
airflow is associated with decreased lung function, and clinical symptoms demonstrating exacerbation of
COPD include cough, sputum production, and shortness of breath. 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 (Sticb et al.. 2009). In a single-city study in Sao Paulo, Brazil, Arbex etal.
(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
September 2019
3-53
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
evaluated studies use 8-hour daily max averaging times, 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).
• 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 (Magzamen 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-deflcient mice that do not resolve by
September 2019
3-54
DRAFT: Do Not Cite or Quote
-------
1 72 hours. In contrast, resolution of responses occurred by 72 hours in sfpd-sufficient mice. Ozone
2 exposure resulted in altered lung mechanics that is indicative of central airway and peripheral
3 tissue involvement in sfpd-deficient mice. In sfpd-sufficient mice, ozone exposure resulted in
4 altered lung mechanics that is indicative of only central airway involvement. These
5 determinations were made by analysis of resistance and elastance spectra obtained from
6 impedance data. In a second study, Groves et al. (2013) found age-related increases in enlarged
7 vacuolated macrophages, alveolar wall rupture, type 2 hyperplasia, BALF protein and cell
8 number, and changes in lung mechanics consistent with COPD are observed in sfpd-deficient
9 mice. Acute ozone exposure (0.8 ppm, 3 hours) resulted in greater alveolar hyperplasia and
10 classically activated macrophages in sfpd-deficient than in sfpd-sufficient mice. The effects of
11 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)
12 In summary, recent large multicity epidemiologic studies of ED visits support an association
13 between short-term ozone exposure and COPD exacerbation. Associations are reported across a variety of
14 study locations, exposure levels, and exposure assignment methods, including nearest monitor
15 concentrations and CMAQ-fused models. In limited copollutant results, the observed association is robust
16 to adjustment for other gaseous pollutants (NO2, SO2, and CO). While none of the experimental animal
17 studies evaluated in the 2013 Ozone ISA examined acute ozone exposure in animals with chronic
18 inflammation, results from recent studies suggest that chronic inflammation enhances sensitivity to ozone
19 exposure, providing coherence with ozone-related COPD exacerbation observed in epidemiologic studies.
3.1.6.2 Obese Populations or Populations with Metabolic Syndrome
20 Metabolic syndrome is comprised of a cluster of metabolic abnormalities, including obesity,
21 dyslipidemia, hypertension, and type 2 diabetes. There is growing evidence that components of metabolic
22 syndrome, including obesity, may increase susceptibility to air pollution-related health effects (Jiu-Chiuan
23 and Schwartz. 2008). The following section evaluates studies examining the relationship between
24 short-term exposure to ozone and respiratory health effects in obese populations or populations with
25 metabolic syndrome.
3.1.6.2.1 Lung Function
3.1.6.2.1.1 Controlled Human Exposure Studies
26 In the 2013 Ozone ISA (U.S. EPA. 2013a'). two retrospective analyses of controlled human
27 exposure studies showed ozone-induced FEVi decrements increased with increasing BMI. Since the 2013
28 Ozone ISA, there is a new controlled human exposure study and a larger retrospective analysis
September 2019
3-55
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
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
magnitude but in the opposite direction of the age effect. Thus, the model predicts FEV i
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. Study specific details, including effect estimates, are
summarized in Table 3-7 in Section 3.3.1.
September 2019
3-56
DRAFT: Do Not Cite or Quote
-------
3.1.6.2.2
Airway Responsiveness
3.1.6.2.2.1 Controlled Human Exposure Studies
1 No controlled human exposure studies were available for review in the 2013 Ozone ISA that
2 examined airway responsiveness in obese individuals or individuals with metabolic syndrome (U.S. EPA.
3 2013a).
4 • A recent study showed an increase in airway responsiveness after ozone exposure did not differ
5 between normal-weight and obese women (Bennett et al.. 2016). Study-specific details, including
6 exposure concentrations and durations, are summarized in Table 3-31 in Section 3.3.1.
3.1.6.2.2.2 Animal Toxicological Studies
7 The 2013 Ozone ISA summarized the evidence of respiratory effects in obese animals resulting
8 from exposure to ozone (U.S. EPA. 2013a). In mouse models of obesity, airways were innately more
9 responsive and responded more vigorously to acute ozone exposure (2 ppm for 3 hours) than lean
10 controls. Newly available information confirms and extends these findings.
11 Several recent studies evaluated the respiratory effects of acute ozone exposure (2 ppm, 3 hours)
12 in mouse models of obesity (Mathews et al.. 2018; Mathews et al.. 2017a; Mathews et al.. 2017b;
13 Williams et al.. 2015). These studies compared responses in obese mice with those of lean mice. Changes
14 in airway responsiveness described below were statistically significant. Study-specific details are
15 summarized in Table 3-32 in Section 3.3.1.
16 • Pulmonary mechanics were assessed by using the flexiVent system. Baseline and nonspecific
17 (i.e., methacholine challenge) airway responsiveness were greater in obese mice than lean mice in
18 the absence of ozone exposure. Acute ozone exposure increased baseline and nonspecific airway
19 responsiveness in obese mice, but not in lean mice.
20 • Williams et al. (2015) probed the role of TNF-a and TNF-a receptor in the augmented responses
21 to ozone exposure in obese mice and found that deficiency in these factors enhanced the increase
22 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
23 No controlled human exposure studies were available for review in the 2013 Ozone ISA that
24 examined pulmonary inflammation, injury, or oxidative stress in obese individuals or individuals with
25 metabolic syndrome (U.S. EPA. 2013a). Study-specific details from recent studies, including exposure
26 concentrations and durations, are summarized in Table 3-21 and Table 3-31 in Section 3.3.1.
September 2019
3-57
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
• 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 hour 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
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
September 2019
3-58
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
in increased eosinophils and albumin (a marker of injury) in BALF. 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 a.m. to 9 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
(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.
September 2019
3-59
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
3.1.6.3
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.
• 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, tracheobronchiolar 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 (-48.0 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.
September 2019
3-60
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
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 8-hour daily max averaging times, 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 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 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).
September 2019
3-61
DRAFT: Do Not Cite or Quote
-------
Study Location
Combined Respiratory Infection
Stieb et al. (2009)
Acute Respiratory Infection
tMaligetal. (2016)
Upper Respiratory Infection
tDarrowet al. (2014)
fBarry et al. (2018)
tMaligetal. (2016)
tXiaoetal. (2016)
fSzyszkowicz et al. (2018)
Lower Respiratory Infection
fKousha et al. (2014)
fSzyszkowicz et al. (2018)
Pneumonia
fWinquist et al. (2012)
fDarrow et al. (2014)
tMaligetal. (2016)
fXiaoet al. (2016)
Bronchitis
tDarrowet al. (2014)
Multicity, Canada
California
Atlanta, GA
Atlanta, GA
Birmingham, AL
Dallas, TX
Pittsburgh, PA
St. Louis, MO
California
Georgia
Multicity, Canada
Edmonton, Canada
Multicity, Canada
St. Louis, MO
Atlanta, GA
California
Georgia
Atlanta, GA
Ages
All Ages
All
All
2-18
<20; Females
<20; Males
All
<20; Females
<20; Males
All
0-4
All
2-18
0-4
Mean
Concentration3
10.3-22.1; 24-hr avg
33-55; 1-hrmax
0-4 45.9
All 37.5-42.2
33-55; 1-hr max
42.1
22.5-29.2; 24-hr avg
18.6
22.5-29.2; 24-hr avg
NR
45.9
33-55; 1-hr max
42.1
45.9
Season
Year-Round
Year-Round
Warm
Year-Round
Year-Round
Year-Round
Warm
Year-Round
Year-Round
Year-Round
Year-Round
Year-Round
Year-Round
Year-Round
Year-Round
Year-Round
Warm
Year-Round
Year-Round
Lag
0-1
Lag 0-2
0-2
0-1
0-3
1
0-4 DL
Lag 0-2
0-1
0-3
Lag 0-2
0.95
Effect Estimate (95% CI)
DL = distributed lag.
Note: fStudies published since the 2013 Ozone ISA.
aMean concentrations reported in ppb and are for 8-hour daily max averaging-times unless otherw/ise 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.
September 2019
3-62
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
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 costimulation by innate immune cells such as macrophages and dendritic cells, which
may lead to T-cell activation. Controlled human exposure studies described in 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. 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.
(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
September 2019
3-63
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
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.. 2016V 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.
3.1.8 Combinations of Respiratory Related Hospital Admissions and
Emergency Department (ED) Visits
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'). The available studies 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; 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
September 2019
3-64
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
(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 old. 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 8-hour daily max averaging times, 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
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 city-wide 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.
September 2019
3-65
DRAFT: Do Not Cite or Quote
-------
1 • A limited number of studies evaluated lag structures (Malig et al.. 2016; Darrow et al.. 2011).
2 seasonal differences in associations (Malig et al.. 2016). and copollutant confounding (Malig et
3 al.. 2016). These results are discussed in more detail in the Relevant Issues for Interpreting
4 Epidemiologic Evidence section (Section 3.1.10).
5 In summary, studies conducted in diverse locations with a variety of exposure assignment
6 techniques continue to provide evidence of an association between ozone and both hospital admissions
7 and ED visits for combined respiratory diseases. Additionally, there is some evidence, previously
8 characterized in the 2013 Ozone ISA, that daily 8-hour max, 1-hour max, and daytime average ozone
9 concentrations may be most strongly associated with respiratory ED visits (Darrow et al.. 2011).
10
September 2019
3-66
DRAFT: Do Not Cite or Quote
-------
Study
ED Visits
Darrow et al. (2011)
fWinquistet al. (2012)
tDarrowetal. (2011)
tMalig etal. (2016)
Location
Atlanta, GA
Ages Mean Concentration Season Lag
All
St. Louis, MO All
Atlanta, GA All
California All
tBarry etal. (2018) Atlanta, GA All
Birmingham, AL
Dallas, TX
Pittsburgh, PA
St. Louis, MO
fO'Lenicketal. (2017) Atlanta, GA 5-18
Dallas, TX
St. Louis, MO
Hospital Admissions
Katsouyanni et al. (2009) 90 U.S. Cities 65+
Katsouyanni et al. (2009) 12 Canadian Cities 65+
Cakmaketal. (2006) 10 Canadian Cities All
fWinquistet 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
Year-Round
Year-Round
Warm
Year-Round
1
1
1
0-4 DL
1
0-1
0-1
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) ~
-• ~
-h
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 8-hour daily max averaging times 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.
September 2019
3-67
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
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 et al. (2011) conducted a study in Atlanta, GA that included 7.5 additional years of data
compared to 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.
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
were robust to statistical adjustment for PM2 5. Additionally, Tolbert et al. (2007) reported
September 2019
3-68
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
ozone-related increases in ED visits for combined respiratory diseases that were attenuated, but
still positive, in 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 NO3 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 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 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.
September 2019
3-69
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
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; Bvers 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).
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. 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
September 2019
3-70
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
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 average 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.
September 2019
3-71
DRAFT: Do Not Cite or Quote
-------
3.1.10.3.1 Asthma
1 • Associations between short-term ozone exposure and hospital admissions and ED visits for
2 asthma are generally present across daily lags ranging from 0 to 6 days (Table 3-1). Although
3 precision (e.g., 95% CIs) is not specified in the table, within-study precision was generally
4 consistent across single-day lags.
5 • The strongest single-day associations were generally observed with ozone concentrations on the
6 same day as the outcome, or within the first 3 days prior to the outcome.
7 • Studies that examined multiday average lag associations generally reported stronger, but less
8 precise associations than single-day lags (Goodman et al.. 2017b; Zu et al.. 2017; Malig et al..
9 2016; Bvers et al.. 2015).
3.1.10.3.2 Other Respiratory Effects
10 • A limited number of studies examined the lag structure of associations between short-term
11 exposure to ozone and COPD ED visits. In a statewide study in California, Malig et al. (2016)
12 observed associations between ED visits for COPD and single-day lagged ozone on Days 0
13 through 3. The largest and most precise (i.e., smallest 95% CI) effect estimate was observed with
14 ozone concentrations on the day prior to ED visit. Similarly, in a multicity study in Canada,
15 Szvszkowicz et al. (2018) reported evidence of more immediate effects of ozone in males. The
16 authors observed associations of similar magnitude and precision on lag Days 0 through 2.
17 Results for females were more delayed, with associations between ozone and COPD ED visits
18 noted on lag Days 2 through 4.
19 • In a study of combined respiratory-related ED visits in Atlanta, warm season associations with
20 same-day ozone concentrations were strongest (i.e., of greatest magnitude), compared to 1-, 2-,
21 and 3-day lags (Darrow et al.. 2011). Malig et al. (2016) similarly observed consistent warm
22 season associations on single-day lags from 0 to 3, but reported the strongest associations with 1
23 and 2 day lagged ozone. The authors additionally reported that moving average ozone
24 concentrations were associated with larger increases in respiratory ED visits, but the estimates
25 were less precise than single day lag estimates.
3.1.10.3.3 Summary of Evidence on Lag Structures
26 In summary, the largest evidence base for lag structure comes from studies examining the
27 association between ozone exposure and hospital admissions or ED visits for asthma. Associations were
28 generally observed across a range of lags, extending as far as 6 days prior to the health outcome of
29 interest. This range indicates that ozone may elicit both immediate and prolonged effects, with additional
30 evidence of potentially delayed respiratory effects in one study (Sheffield et al.. 2015). Additionally, the
31 strongest associations were observed with multiday averages of ozone that were indicative of more
32 immediate effects. Notably, effect estimates derived from multiday average concentrations were less
33 precise than effect estimates from single-day lag estimates. Finally, it is important to note that different
34 lag responses may be observed across different population subgroups (e.g., age or sex groups), as seen in
35 Szvszkowicz et al. (2018).
September 2019
3-72
DRAFT: Do Not Cite or Quote
-------
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
Gleason 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 Function
1 The 2013 Ozone ISA evaluated a large body of epidemiologic evidence that provided evidence of
2 an association between short-term exposure to ambient ozone and respiratory health effects. Of the
3 evaluated studies, a limited number attempted to characterize the shape of the C-R relationship or
4 determine the presence of a concentration threshold below which a positive association with health effects
5 does not occur. Studies examining asthma-related hospital admissions (Silverman and Ito. 2010) and ED
6 visits (Strickland et al.. 2010) used natural splines and locally weighted smoothing functions,
7 respectively, to examine the shape of the C-R relationship between ozone concentrations and
September 2019
3-73
DRAFT: Do Not Cite or Quote
-------
1 asthma-related hospital admissions or ED visits. Visual inspections of the plots revealed approximately
2 linear associations and no evidence of a threshold with 8-hour daily max ozone concentrations as low as
3 30 ppb (Figure 3-8 and Figure 3-9). There is increased uncertainty in the shape of the C-R curve at the
4 lower end of the distribution of ozone concentrations, starting around 30 ppb, due to the low density of
5 data in this range.
Ozone Warm Season
LO
CM
o
LO
o
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: Permission pending, Strickland et al. (2010).
Figure 3-8 Loess (locally estimated scatterplot smoothing) 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.
September 2019
3-74
DRAFT: Do Not Cite or Quote
-------
Ozone: All Ages
cc
cc
p
NAAQS
20
40
60
80
100
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: Permission pending, Silverman and Ito (2010).
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.
1 In addition, a small number of recent studies show conflicting evidence of C-R nonlinearity and
2 the presence of a threshold. In contrast to evidence from the 2013 Ozone ISA, a multicity study in Texas
3 estimated C-R curves using penalized spline models and observed evidence of nonlinearity in the
4 relationship between 8-hour daily avg ozone and asthma hospital admissions (Zu et al.. 2017). The C-R
5 curves indicate the potential presence of a threshold between 30 and 40 ppb for children aged 5-14-years
6 and adults aged 15-64-years (see Figure 3-10). The presence of a threshold in this range is supported by a
7 recent statewide study in New Jersey that examined associations between pediatric asthma ED visits and
8 quintiles of 8-hour daily max ozone exposure (Gleason et al.. 2014). In comparison to the lowest quintile
9 of ozone exposure, only quintiles 3 through 5 were associated with increased odds of ED visits. The third
10 quintile exposure range started at 42.48 ppb.
September 2019
3-75
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
10
c
o
g b. children (aged 5-14 yrs)
m
c
a
c. youngef adults {aged 15-64 yrs}
.£
i
o
re
S
1
i
¦a
m
in
1
o
c.
5
I
a
£
o
E
£
ra
E
%
ra
c
4>
9
%
lr.
"\ //
ID
&
rri
e
ra
3
«v1 «. / /
O-j -—,
O
m
"X
o
I 1—r i r i . 1 1 t—
0 1 0 20 30 50 60 70 0 10 20 30 4D 50 60 70
O?one LagD-3 fppti) Oaorte LagQ-3 (ppb)
Note: The solid lines are smoothed-fit data, with long broken lines indicating 95% confidence bands.
Source: Permission pending, Zu et al. (2017).
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.
September 2019
3-76
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
50 40 50 60 70
Concentration (ppb)
30 40 50 60 70
Concentration (ppb)
Ozone - Pneumonia
Ozone - URI
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: Permission pending, Darrow et al. (2014).
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 ppb. 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, increased airway responsiveness, and lung injury and
September 2019
3-77
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
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 as low
as 60 ppb in young adults (Section 3.1.4.1.1). There were no recent experimental studies in humans that
examined respiratory symptoms in relation to short-term ozone exposures. However, ozone-induced
respiratory symptoms in combination with FEVi decrements in young healthy adults at concentrations as
low as 70 ppb 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. Recent studies expand on
observed interindividual variability in inflammatory responses, providing additional evidence that
GSTMl-null individuals are 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 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). A number of panel studies of children in
summer camps 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
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. While there is a lack of recent epidemiologic studies conducted in the
September 2019
3-78
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
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 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). 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 mortalitv(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 ozone-related increases in ED visits for COPD (Section 3.1.6.1.2).
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
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.
September 2019
3-79
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Copollutant analyses were limited in epidemiologic studies evauated 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.
Epidemiologic studies have also attempted to 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. The strongest single-day
associations were generally observed with ozone concentrations on the same day as the outcome, but
positive associations were present across a range of lags, extending as far as 6 days prior to the health
outcome of interest. This range indicates that ozone may elicit both immediate and prolonged respiratory
effects. 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
exposure. The combination of lung function decrements and respiratory symptoms has been observed
following 70 ppb and greater ozone concentrations following 6.6 hour exposures. 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
September 2019
3-80
DRAFT: Do Not Cite or Quote
-------
1 upregulation of mucus synthesis and storage in allergic mice compared to naive mice. These toxicological
2 results further inform the potential mechanistic pathways that underlie downstream respiratory effects,
3 providing continued support for the biological plausibility of the observed epidemiologic results. Thus,
4 the recent evidence integrated across disciplines, along with the total body of evidence evaluated in
5 previous integrated reviews, is sufficient to conclude that there is a causal relationship between
6 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.
Rationale for Causality
Determination3
Key Evidence13
Key References'3
Ozone
Concentrations
Associated with
Effects0
Respiratory Effects in Healthy Populations
Consistent evidence from controlled
human exposure studies at relevant
concentrations
Studies show:
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
stress
Section 3.1.4.4.1
60-600 ppb
Consistent evidence from
toxicological studies at relevant
concentrations
Studies show:
Altered ventilatory parameters
Section 3.1.4.1.2
0.1-2 ppm
Cough response
Clav et al. (2016)
2 ppm
Increased airway responsiveness
Section 3.1.4.3.2
0.3-2 ppm
Inflammation, injury, and oxidative
stress
Section 3.1.4.4.2
0.15-2 ppm
Type 2 immune responses—upper Harkema et al. 0.5-0.8 ppm
and lower airways (2017): Kumaaai et
al. (2017): Ona et
al. (2016): Kumaqai
et al. (2016).
September 2019
3-81
DRAFT: Do Not Cite or Quote
-------
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
Coherence in epidemiologic studies Panel studies provide support for
of respiratory effects in healthy experimental studies with consistent
children associations for lung function and
pulmonary inflammation in healthy
children
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
Evidence supporting biological Controlled human exposure studies Section 3.1.4.1.1 400-420 ppb
plausibility 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
evidence that changes in lung
function may be attributed to
activation of sensory nerves and
involvement of parasympathetic
pathways
Section 3.1.4.1.2
Section 3.1.4.3.2
Clay et al. (2016) 2 ppm
Verhein et al.
(2013)
Respiratory Effects in Populations with Asthma
Consistent epidemiologic evidence
from multiple, high-quality studies
at relevant concentrations
Increases in asthma-related hospital
admissions and ED visits in children,
and all ages combined in studies
conducted in the U.S. and Canada
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
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
Section 3.1.5.4.1
Section 3.1.5.5.1
Section 3.1.5.6.1
>125 ppb
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
September 2019
3-82
DRAFT: Do Not Cite or Quote
-------
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
Coherence in epidemiologic studies Panel studies in children with asthma Section 3.1.5.3.2 1- h max:
across the continuum of effects provide support for asthma Section 3.1.5.4.3 43.0-65.8 ppb
exacerbation in children, with Section 3.1.5.6.3 3. ^ max-
consistent associations for respiratory 31 5-52 9 ppb
symptoms, lung function decrements,
and pulmonary inflammation
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 Scheleqle 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
Respiratory Effects in Populations with COPD
Consistent epidemiologic evidence Increases in ED visits for COPD in
from a limited number of studies conducted in the U.S. and
high-quality multicity studies at Canada
relevant concentrations
Stieb et al. (2009) 24- h avg:
18.4 ppb
Malia et al. (2016) 1 - h max:
33-55 ppb
Szvszkowicz et al. 24- h avg:
(2018) 22.5-29.2 ppb
Consistent evidence from a limited Results show enhanced injury, Groves et al. (2012) 0.8 ppm
number of toxicological studies at inflammation, oxidative stress, and Groves et al (2013)
relevant concentrations altered morphology and lung
mechanics in animal model of COPD
But, lack of coherence in Panel studies in adults with COPD do Peacock et al.
epidemiologic studies across the not observe ozone associations with (2011): Maqzamen
continuum of effects lung function or respiratory symptoms et al. (2018)
in adults with COPD
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
Malia et al. (2016)
September 2019
3-83
DRAFT: Do Not Cite or Quote
-------
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
Respiratory Infection
Generally consistent epidemiologic
evidence from multiple, high-quality
studies at relevant concentrations
Increases in ED visits for:
Pneumonia
Maliq etal. (2016):
Xiao etal. (2016)
1 - h max:
33-55 ppb
8- h max:
42.1 ppb
Acute respiratory infections
Malia etal. (2016)
1 - h max:
33-55 ppb
Upper respiratory tract infections
Maliq etal. (2016):
Xiao etal. (2016):
Szvszkowicz et al.
(2018): Barrv etal.
(2018).
1 - h max:
33-55 ppb
8- h max:
37.5-42.2 ppb
24- h avg:
22.5-29.2 ppb
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
infectious disease
Mikerov et al.
(2011)
2 ppm
Evidence of biological plausibility
Animal toxicological studies show
increased susceptibility to infections
Section 3.1.7.3
0.08-2 ppm
Combinations of Respiratory-Related Hospital Admissions and ED Visits
Epidemiologic studies provide
consistent evidence of positive
associations when examining
combined respiratory-related
diseases
Increases in hospital admissions and
ED visits for combined respiratory-
related diseases in multicity studies
Section 3.1.8
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 studies,
with evidence that associations
generally remain robust in models
with gaseous pollutants
Section 3.1.10.1
Respiratory Mortality
Generally consistent epidemiologic Generally consistent evidence of
evidence from multiple, high-quality increases in mortality in response to
studies at relevant concentrations 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
September 2019
3-84
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
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. 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. 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 3.3.1).
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 PM2 5 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
support for the observed respiratory health associations with long-term exposure to ozone. Building upon
September 2019
3-85
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
that evidence, the 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, averaging times, 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 draft 2019
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
• Exposure: Long-term (on the order of months to years) or perinatal inhalation exposure to
relevant ozone concentrations (i.e., <2 ppm)
September 2019
3-86
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
• 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.
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
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
September 2019
3-87
DRAFT: Do Not Cite or Quote
-------
1 host defense response was diminished. These types of alterations in structure and function in the
2 developing lung may underlie the development of asthma.
3 Recent studies include those conducted in adult and neonatal rodents and those conducted in
4 infant monkeys. Studies in adult rodents found that long-term exposure to ozone results in respiratory
5 tract oxidative stress, inflammation, and injury (Gordon et al.. 2016b; Gordon et al.. 2016a; Miller et al..
6 2016a; Snow et al.. 2016). Similar findings were reported in the developing lungs of rodents exposed
7 postnatally to ozone (Dye et al.. 2017; Gabehart et al.. 2015; Gabehart et al.. 2014). In addition, secretion
8 and upregulation of mucus expression, which can offer protection against injury, were increased, while
9 cell proliferation was decreased in the neonatal rodents.
September 2019
3-88
DRAFT: Do Not Cite or Quote
-------
A
Long-
Term
Ozone
Exposure
Respiratory
Tract
Oxidative
Stress
Respiratory
Tract
Inflammation
Fibrotic- or
Emphysema-
Like
Disease/COPD
~
Respiratory
Mortality
Increased Seventy
of Influenza
Altered Lung
Development
Increased
Airway
Responsiveness
Development of
Asthma
Altered Morphology
in Nasal Airways
Allergic Responses
Serotonin
Upregulation/
Altered Neural
innervation
Altered Innate
Immune 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.
Figure 3-12 Potential biological pathways for respiratory effects following
long-term ozone exposure.
1 Postnatal ozone exposure had morphological effects. For example, decreased sensory neuron
2 development (Zellner et al.. 2011) and altered airway architecture (Lee et al.. 2011) were demonstrated in
3 rodents. Studies in infant monkeys found altered components of a cell death pathway, altered expression
4 of serotonin, which is a neurotransmitter involved in airway smooth muscle contraction, and altered
5 innate immune function (Clay et al.. 2014; Murphy et al.. 2014; Murphy et al.. 2013). Alterations in cell
6 growth and cell death pathways observed in these long-term studies may underlie changes in structure
7 (i.e., airway architecture) in the developing lung. Effects on serotonin could potentially underlie changes
8 in function in the developing lung (i.e., increased airway responsiveness), while effects on innate immune
9 function may lead to altered immune response. Studies in the infant model of allergic airway disease
10 model found impaired alveolar morphogenesis (Herring et al.. 2015; Avdalovic et al.. 2012). airway
September 2019
3-89
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
smooth muscle hyperreactivity (Moore et al.. 2012). an enhanced allergic phenotype (Crowley et al..
2017; Chou et al.. 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).
3.2.4 Respiratory Health Effects
3.2.4.1 Development of Asthma
Asthma is a chronic inflammatory disease of the airways that develops over time (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.
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.
September 2019
3-90
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
Studies evaluated in the 2013 Ozone ISA did not provide evidence of a main effect of long-term
ozone exposure on asthma incidence 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 gene and variants in genes for arginase, respectively, related to the risk of
new-onset asthma in children. 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.
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).
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).
September 2019
3-91
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
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 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
September 2019
3-92
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
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.. 2012). 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
September 2019
3-93
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
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 (five 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.. 2014). 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
September 2019
3-94
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
IL-6 and IL-8 gene expression in response to LPS in the 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 etal. (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 development 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
September 2019
3-95
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
compromising airway growth and development, promoting the development of an allergic phenotype, 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). As such, 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 et al. (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).
September 2019
3-96
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
In 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
September 2019
3-97
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
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. 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.
Two 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. Effects described below were statistically significant.
Study-specific details are summarized in Table 3-50 in Section 3.3.2.
• 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.
September 2019
3-98
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
• In 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.
3.2.4.3.1 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).
September 2019
3-99
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
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 3.2.5V 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 glycosoaminoglycans, 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. (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, N-a.ccty 1 -g 1 utarni 11 idase) 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
September 2019
3-100
DRAFT: Do Not Cite or Quote
-------
1 or injury were observed at 0.25 ppm. A marker of inflammation (BALF total cell number) was
2 increased in young adult rats exposed to 1 ppm ozone.
3 • In Gordon et al. (2016a). rats that were exercise-trained or sedentary were exposed for 6 weeks to
4 ozone (0.25, 0.5, and 1.0 ppm for 5 hours/day, once a week). No effects on inflammation or
5 injury were observed at 0.25 ppm. Exposure to 0.5 ppm ozone increased markers of injury (BALF
6 protein and albumin) in exercise-trained rats. Exposure to 1.0 ppm ozone increased inflammatory
7 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
8 The 2013 Ozone ISA did not evaluate any epidemiologic studies that examined the relationship
9 between long-term exposure to ozone and the development of COPD. One recent epidemiologic study
10 provides evidence of an association between long-term ozone concentrations and incident COPD
11 hospitalizations. Animal toxicological studies reviewed in the 2013 Ozone ISA found that chronic ozone
12 exposure can damage the distal airways and proximal alveoli, resulting in persistent inflammation and
13 lung tissue remodeling that leads to irreversible changes including flbrotic- and emphysematous-like
14 changes in the lung. Additionally, recent animal toxicological studies provide consistent evidence that
15 subchronic ozone exposure can lead to airway injury and inflammation. In adult animals these changes
16 may underlie the progression and development of chronic lung disease and provide biological plausibility
17 for ozone-induced development of COPD.
3.2.4.4 Respiratory Infection and other Associated Respiratory Effects
3.2.4.4.1 Epidemiologic Studies
18 There were no epidemiologic studies examining the association between long-term exposure to
19 ozone and respiratory infection available for inclusion in the 2013 Ozone ISA (U.S. EPA. 2013a). Two
20 recent studies observed inverse associations between ozone and respiratory infection. Smith et al. (2016)
21 reported an inverse association between 2-year avg ozone concentrations and pulmonary tuberculosis in a
22 nested case-control study of adults in northern California. The authors did observe a strong positive
23 association with NO2 and a negative correlation between ozone and NO2, which may explain the inverse
24 association. In a study of otitis media in the first 2 years of life, 2-month avg ozone concentrations were
25 associated with decreased risk of infection (Maclntyre et al.. 2011). Study-specific details, including air
26 quality characteristics and select effect estimates, are highlighted in Table 3-52 in Section 3.3.2.
September 2019
3-101
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
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, two cross-sectional studies of asthma hospital admissions in children (Moore et al..
2008) and 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.
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
September 2019
3-102
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
ozone can be found in the "Relevant Issues for Interpreting Epidemiologic Evidence" section
(Section 3.2.5).
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.
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.
3.2.4.5.3 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
September 2019
3-103
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
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 forNC>2, 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.
3.2.4.6.2 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 etal. (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.
September 2019
3-104
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
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.
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 utaminidase) 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.
September 2019
3-105
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
• 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.
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 studies observing positive associations
between long-term ozone exposure and respiratory mortality (Jerrett et al.. 2009) included in the 2013
Ozone ISA, 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 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 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.
September 2019
3-106
DRAFT: Do Not Cite or Quote
-------
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. 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 Associations between long-term exposure to ozone and
respiratory mortality in recent cohort studies.
September 2019
3-107
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
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 (Tetreault et al.. 2016a).
3.2.5.3 Shape of the Concentration-Response Function
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.
September 2019
3-108
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
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 new-onset asthma, 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. 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. This 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. While 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 provide compelling 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 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).
September 2019
3-109
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
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 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 3.2.4.6). 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
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
September 2019
3-110
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
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. 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 given that ozone
concentrations do not vary over space as much as other criteria pollutants, such as NOx or SO2
(Section 2.3.1.1)
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
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.
September 2019
3-111
DRAFT: Do Not Cite or Quote
-------
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. (2012) q g 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
Tetreault et al. 32.1 ppb mean
(2016a): Garcia et summer ozone
al. (2019) concentration,
based on 8-h
midday avg
Longitudinal studies provide
evidence of associations with
asthma development in populations
with specific genetic variants
Islam et al. (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.)
Uncertainty regarding confounding
by copollutants
No examination of copollutant
confounding in models of gene-
environment interaction. Available
studies report low to moderate
copollutant correlations
September 2019
3-112
DRAFT: Do Not Cite or Quote
-------
Table 3-3 (Continued): Summary of evidence for a likely to be causal relationship
between long-term ozone exposure and respiratory effects.
Rationale for Causality
Determination3
Key Evidence13
Key References'3
Ozone
Concentrations
Associated with
Effects0
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):
Mena 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
by copollutants
Limited evidence that observed
associations were attenuated but still
positive in copollutant models
adjusting for NO2 or PM2.5
Berhane et al.
(2016): Gilliland et
al. (2017)
Other uncertainties
Studies of hospital admissions and
ED visits for asthma do not account
for the potential effect of short-term
exposures leading to these acute
events
Section 3.2.4.5.1
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
Development of Chronic Obstructive Pulmonary Disease
Consistent evidence from Animal toxicological evidence of Section 3.2.4.3.2 0.12-1 ppm
toxicological studies at relevant morphologic changes in distal
concentrations airways and proximal alveoli leading
to lung tissue remodeling and
fibrotic/emphysematous-like
changes in rodents and monkeys
September 2019
3-113
DRAFT: Do Not Cite or Quote
-------
Table 3-3 (Continued): Summary of evidence for a likely to be causal relationship
between long-term ozone exposure and respiratory effects.
Rationale for Causality
Determination3
Key Evidence13
Key References'3
Ozone
Concentrations
Associated with
Effects0
Limited epidemiologic evidence
from study of COPD incidence
The only study evaluated indicates
an association between ozone
concentrations and COPD incidence
in adults with asthma
Toetal. (2016")
38.4 ppb mean
ozone
concentration,
based on average
of monthly 24- h
max from time of
enrollment
Uncertainty regarding confounding
by copollutants
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
Allergic Response
Limited, but consistent
epidemiologic evidence from
studies of allergic response
Epidemiologic studies provide
consistent evidence of ozone
associations with hay fever/rhinitis
and allergen-specific IgE levels
Section 3.2.4.6.1
51.5 ppb annual
average, based on
8- h max
Uncertainty regarding confounding
by copollutants
Limited evidence from a single study
reported an association between
ozone and rhinitis that was
persistent in a copollutant model
adjusting for NO2.
Penard-Morand et
al. (2005)
Other uncertainties
All available studies were
cross-sectional. Additionally,
potential confounding by pollen
concentrations also remains a
considerable uncertainty
Section 3.2.4.6.1
Coherent evidence from
toxicological studies at relevant
concentrations
Animal toxicological evidence for
enhanced allergic responses
Section 3.2.4.6.2
0.1-0.5 ppm
Animal toxicological evidence from
short-term studies show type 2
immune responses in nasal airways
of rodents repeatedly exposed
Section 3.1.4.4.2
0.5-0.8 ppm
Respiratory Mortality
Inconsistent epidemiologic
evidence from multiple, high-
quality studies
Recent epidemiologic studies
provide some evidence of an
association with respiratory mortality,
but the evidence is not consistent.
New evidence from one study
demonstrating an association with
COPD mortality
Section 3.2.4.9
September 2019
3-114
DRAFT: Do Not Cite or Quote
-------
Table 3-3 (Continued): Summary of evidence for a likely to be causal relationship
between long-term ozone exposure and respiratory effects.
Rationale for Causality
Determination3
Key Evidence13
Key References'3
Ozone
Concentrations
Associated with
Effects0
Some coherence with underlying
causes of mortality
Studies of COPD development
provide coherence with COPD
mortality
Section 3.2.4.3
Biological plausibility
Animal toxicological studies show
the development of
emphysematous-like disease and
increased severity of infection-
related alveolitis
Section 3.2.4.3.2
Section 3.2.4.5.2
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.
September 2019
3-115
DRAFT: Do Not Cite or Quote
-------
3.3 Evidence Inventories—Data Tables to Summarize Study
Details
i
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)
Ghioetal. (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)
Frampton et al. (2015)
Healthy adults
n = 12 males, 12 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)
Kahle et al. (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
FEV-i/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)
September 2019
3-116
DRAFT: Do Not Cite or Quote
-------
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 etal. (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, 3h
120 ppb, 3h
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)
Tank etal. (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)
September 2019
3-117
DRAFT: Do Not Cite or Quote
-------
Table 3-5 Study-specific details from animal toxicological studies of short-term
ozone exposure and lung function—healthy.
Study
Species (Strain), n, Sex,
Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Scheleqle 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)
Cho et 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)
Ghioetal. (2014)
Mice (CD-1)
n = 6 females
Age: 4 weeks
2 ppm, 3 h
Enhanced pause, MCh challenge
(24 h PE)
September 2019
3-118
DRAFT: Do Not Cite or Quote
-------
Table 3-5 (Continued): Study-specific details from animal toxicological studies of
short-term ozone exposure and lung function—healthy.
Species (Strain), n, Sex, Exposure Details
Study Age (Concentration, Duration) Endpoints Examined
Razvi etal. (2015)
Mice (C57BL/6J WT and
2 ppm, 3 h
Pulmonary mechanics, MCh
resistin deficient)
challenge (24 h PE)
n = 6-8 males and
females
Age: 4-8 weeks
Dve etal. (2015)
Rats (WKY, WS, SD)
0.25 ppm, 4 h
Whole body plethysmography (0
n = 4-8 males
0.5 ppm, 4 h
and 20 h PE)
Age:12-14 weeks
1.0 ppm, 4 h
Clav etal. (2016)
Guinea pigs
2 ppm, 1 h
Cough response, pulmonary
(Dunkin-Hartley)
2 ppm, 30 min
mechanics, challenge with Mch
n = 4-32 males
(4 h or 3 days PE)
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
Snow et al. (2016)
Rats (BN)
0.25 ppm, 6 h/day for 2 days
Ventilatory parameters (18 h PE)
n = 6-8 males
1 ppm, 6 h/day for 2 days
Age: 1,4, 12, 24 mo
Gordon et al.
Rats (BN)
0.8 ppm, 5 h
Ventilatory parameters (18 h PE)
(2016b)
n = 9-10 males and
females
Age: 20 weeks
Kasahara et al.
Mice (C57BL/6 WT,
2 ppm, 3 h
Pulmonary mechanics, challenge
(2015)
ROCK1 insufficient,
with MCh (24 h PE)
ROCK2 insufficient)
n = 4-12 males
Age:20-25 weeks
Verhein etal. (2013)
Guinea pigs
2 ppm, 4 h
Pulmonary inflation pressure,
(Dunkin-Hartley)
challenge with i.v. of acetylcholine
n = 3-7 females
and electrical stimulation of the
Age: NR but weight was
vagal nerve (24 h PE)
300-470 g
Williams et al.
Mice (C57BL/6, TNF-a
2 ppm, 3 h
Pulmonary mechanics, challenge
(2015)
sufficient and deficient)
with MCh (24 h PE)
n = 5-9 females
Age:10-12 weeks
Hansen et al. (2016)
Mice (BALB/cJ)
2 ppm, 1 h/day for 3 days
Breathing frequency, tidal volume,
n = 5 females
time of brake, time of pause
Age: 6 weeks
(during exposure)
September 2019
3-119
DRAFT: Do Not Cite or Quote
-------
Table 3-5 (Continued): Study-specific details from animal toxicological studies of
short-term ozone exposure and lung function—healthy.
Study
Species (Strain), n, Sex,
Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
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 etal. (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)
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 etal. (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)
September 2019
3-120
DRAFT: Do Not Cite or Quote
-------
Table 3-5 (Continued): Study-specific details from animal toxicological studies of
short-term ozone exposure and lung function—healthy.
Species (Strain), n, Sex, Exposure Details
Study Age (Concentration, Duration) Endpoints Examined
Cho et al. (2018) Mice (C57BL/6) Specific 2 ppm, 3 h Pulmonary mechanics, challenge
pathogen free and germ with MCh (24 h PE)
free
n = 6-14 males
Age: 10 weeks
BN = brown Norway; LE = Long-Evans; MCh = methacholine; PE = post-exposure; S-D = Sprague-Dawley; WKY = Wistar Kyoto;
WS = Wistar; WT = wild type.
September 2019
3-121
DRAFT: Do Not Cite or Quote
-------
Table 3-6 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
Berry et al. (1991)
Hamilton, NJ, U.S.
July 1988
Panel study
n = 14
Campers
without asthma
Age: <14 yr
Regional monitor for Mean: NR Correlation FEVi (mL):
part of study
(<8 miles from
camps)
Mobile trailer monitor
onsite at one camp
1 - h max
Maximum:
204
(r): NR
Copollutant
models with:
NR
20.5 (4.3, 36.7)
PEF (mL/sec):
-25.3 (-82.6, 32.1)
Spektorand Lippmann
(1991)
Fairview Lake, NJ, U.S.
July-August 1988
Panel study
n = 46
Campers
without asthma
Age: 8-14 yr
On-site monitor
1-h avg
Mean: 69
Maximum:
137
Correlation
(r): NR
Copollutant
models with:
NR
Percent increase
FEVi:
-1.4 (-1.9, -0.8)
Avol et al. (1990)
Pine Springs, CA, U.S.
June-August 1988
Panel study
n = 295
Campers
without asthma
Age: 8-17 yr
On-site monitoring
1 - h avg
Mean: 94
Maximum:
161
Correlation
(r): NR
Copollutant
models with:
NR
Percent increase
FEVi:
-0.4 (-0.6, -0.1)
PEF:
1.2 (0.4, 1.9)
Burnett et al. (1990)
Lake Couchiching,
Ontario, Canada
June-July 1983
Panel study
n = 29
Campers
without asthma
Age: 7-15 yr
On-site monitoring
1-h avg
Mean: 59 Correlation Percent increase
Maximum: (r)- NR FEVi:
95 Copollutant -0.2 (-1.1, 0.7)
PEF:
-1.19 (-2.38, -0.03)
models with:
NR
Hiaains et al. (1990)
n = 43
On-site monitoring
Mean: 103
Correlation
Percent increase
San Bernardino, CA,
Campers
1-h avg
Maximum:
(r): NR
FEVi:
U.S.
without asthma
245
Copollutant
-1.0 (-1.5, -0.5)
June-July 1987
Age: 7-13 yr
models with:
PEF:
Panel study
NR
-0.5 (-1.3, 0.2)
Raizenne et al. (1989)
n = 112
On-site monitoring
Mean: 71
Correlation
Percent increase
Lake Erie, Ontario,
Campers
1-h avg
Maximum:
(r): NR
FEVi:
Canada
without asthma
143
Copollutant
-0.3 (-0.5, -0.1)
June-August 1986
Age: mean 11.6
models with:
PEF:
Panel study
yr
NR
-0.04 (-0.35, 0.26)
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 with:
NR
Percent Increase
FEVi:
-0.6 (-0.9, -0.2)
PEF:
-1.1 (-2.1, -0.3)
September 2019
3-122
DRAFT: Do Not Cite or Quote
-------
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
tDales etal. (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 with:
NR
Percent increase
FEVi: -0.47 (-1.00,
0.06)
FEVi/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)
Note: fStudies published since the 2013 Ozone ISA.
aResults 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.
September 2019
3-123
DRAFT: Do Not Cite or Quote
-------
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) n = 776
Boston, MA, U.S. Adult men
1999-2009
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 a.m. to
8 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 with:
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
Max: 59.6
Correlation
(r):
NO2: 0.01;
PM2.5: 0.33
Copollutant
models with:
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
U.S.
ages 14-19
schools
Copollutant
Unit change in exhaled
2005
50% asthma
8- h max
models with:
breath condensate pH
Panel Study
prevalence
May-June
NR
-0.14 (-0.33, 0.05)
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 a.m.
to 6 p.m.)
Mean: 35.1
Max: 63.7
Correlation
(r):
PM2.5: 0.07;
PM10: 0.34;
NO2: -0.41
Copollutant
models with:
NR
7-day avg
Inducible nitric oxide
synthase (iNOS)
% methylation
-0.08 (-1.40, 1.28)
Note: fStudies published since the 2013 Ozone ISA.
aResults 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.
September 2019
3-124
DRAFT: Do Not Cite or Quote
-------
Table 3-8 Study-specific details from controlled human exposure studies of
respiratory symptoms in healthy populations.
Population Exposure Details
n, Sex, Age (Range or (Concentration,
Study 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
September 2019
3-125
DRAFT: Do Not Cite or Quote
-------
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
n = 9 males,
0 females
Age: 30 ± 2.6 yr
100 ppb with heat, 0.5 h
Nasal lavage (0 and 6 h PE)
Alexis et al. (2013)
Healthy adults
n =24
Age: 20-33yr
60 ppb, 6.6 h
Sputum PMN (18 h PE)
Hoffmever et al. (2015)
Healthy adults
n = 5 males,
5 females
Age: 25.6 ± 2.5yr
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)
September 2019
3-126
DRAFT: Do Not Cite or Quote
-------
Table 3-9 (Continued): Study-specific details from controlled human exposure
studies of inflammation, oxidative stress, and injury in
healthy populations.
Population
n, Sex, Age (Range Exposure Details
Study or Mean ± SD) (Concentration, Duration) Endpoints Examined
Bennett et al. (2016) Healthy adults 0 ppm, 2 h Sputum PMN (4 h PE)
n = 19 normal weight 400 ppb, 2 h
females
Age: 24 ± 4 yr
n = 19 obese
females
Age: 28 ± 5 yr
Chenq 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)
Tank et 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 etal. (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)
1
September 2019
3-127
DRAFT: Do Not Cite or Quote
-------
Table 3-10 Study-specific details from animal toxicological studies of short-term
ozone exposure and allergic sensitization—healthy.
Study
Species (Strain), n,
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; PE = post-exposure; Th2 = T
helper 2.
September 2019
3-128
DRAFT: Do Not Cite or Quote
-------
Table 3-11 Study-specific details from animal toxicological studies of short-term
ozone exposure and inflammation, oxidative stress, and
injury—healthy.
Species (Strain), n, Sex, Exposure Details
Study Age (Concentration, Duration) Endpoints Examined
Hulo etal. (2011) Mice (C57BL6/SV129 WT 2 ppm, 3 h Markers of oxidative stress,
and AMPK-a deficient) inflammation, injury; AMPK
n = 3-10 males activation; Na/K-ATPase
. _. . abundance(24 h PE)
Age: 20-24 weeks v '
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 peptide- differentials, total protein
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 etal. (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
September 2019
3-129
DRAFT: Do Not Cite or Quote
-------
Table 3-11 (Continued): Study-specific details from animal toxicological studies
of short-term ozone exposure and inflammation, oxidative
stress, and injury—healthy.
Study
Species (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)
Bhoopalan 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)
Cho et al. (2013)
Mice (ICR WT and NRF2
deficient)
n = 3-12
Sex and age: NR
0.3 ppm, 6-72 h
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
BALFtotal 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 Mice (C57BL/6)
et al. (2013) n = 8 fema|es
Age: 7 mo
0.245 ppm, 4 h
BALF cells (12 h PE)
September 2019
3-130
DRAFT: Do Not Cite or Quote
-------
Table 3-11 (Continued): Study-specific details from animal toxicological studies
of short-term ozone exposure and inflammation, oxidative
stress, and injury—healthy.
Study
Species (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)
Wang 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, mRNAeNOS, iNOS,
ICAM-1 (24 h after
6th exposure)
Gonzalez-Guevara et Rats (WS)
al. (2014) n = 3_g ma|es
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)
Ghioetal. (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 (2,472 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)
September 2019
3-131
DRAFT: Do Not Cite or Quote
-------
Table 3-11 (Continued): Study-specific details from animal toxicological studies
of short-term ozone exposure and inflammation, oxidative
stress, and injury—healthy.
Study
Species (Strain), n, Sex,
Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Razvi etal. (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 etal. (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 etal. (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)
Ramotetal. (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 etal. (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)
September 2019
3-132
DRAFT: Do Not Cite or Quote
-------
Table 3-11 (Continued): Study-specific details from animal toxicological studies
of short-term ozone exposure and inflammation, oxidative
stress, and injury—healthy.
Study
Species (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)
Che et 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) 2 ppm, 4 h
n = 3-7 females
Age: NR but weight was
300-470 g
BALF total cells and cell
differentials (24 h PE)
Mathews et al. (2015) Mice (C57BL/6 WT, gamma 0.3 ppm, 24-72 h
delta T cell deficient)
n = 4-14 males
Age:10-13 weeks
BALF total cells and cell
differentials, cytokines, protein;
lung tissue mRNA; lung tissue
macrophage subpopulations,
histopathology (0, 1, 3, 5 days
PE)
September 2019
3-133
DRAFT: Do Not Cite or Quote
-------
Table 3-11 (Continued): Study-specific details from animal toxicological studies
of short-term ozone exposure and inflammation, oxidative
stress, and injury—healthy.
Study
Species (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 etal. (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)
September 2019
3-134
DRAFT: Do Not Cite or Quote
-------
Table 3-11 (Continued): Study-specific details from animal toxicological studies
of short-term ozone exposure and inflammation, oxidative
stress, and injury—healthy.
Study
Species (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)
September 2019
3-135
DRAFT: Do Not Cite or Quote
-------
Table 3-11 (Continued): Study-specific details from animal toxicological studies
of short-term ozone exposure and inflammation, oxidative
stress, and injury—healthy.
Study
Species (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)
Tiahe et al. (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 et al. (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)
September 2019
3-136
DRAFT: Do Not Cite or Quote
-------
Table 3-11 (Continued): Study-specific details from animal toxicological studies
of short-term ozone exposure and inflammation, oxidative
stress, and injury—healthy.
Study
Species (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
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,
ADAM17, 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
5 days
BALF inflammatory cell
subpopulations, IFN-y, IL-4,
IL-17, TGFp, PGE2
(immediately PE)
September 2019
3-137
DRAFT: Do Not Cite or Quote
-------
Table 3-11 (Continued): Study-specific details from animal toxicological studies
of short-term ozone exposure and inflammation, oxidative
stress, and injury—healthy.
Species (Strain), n, Sex, Exposure Details
Study 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 M|P.2i
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 = C-C 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 = inter-cellular 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;
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; 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;
5-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 receptor;
TNF = tumor necrosis factor; TSLP = thymic stromal lymphopoietin; WKY = Wistar Kyoto; WS = Wistar; WT = wild type.
1
September 2019
3-138
DRAFT: Do Not Cite or Quote
-------
Table 3-12 Study-specific details from animal toxicological studies of short-term
ozone exposure and morphology—healthy.
Species (Strain), n, Sex, Exposure Details
Study Age (Concentration, Duration) Endpoints Examined
Connor et al. (2012) Mice (C3H/HeOuJ and 0.8 ppm, 3 h Type 2 cell proliferation (12-72 h
C3H/HeJ TLR4 mutant) PE)
n = 3-18 males
Age:11-12 weeks
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)
Cho et 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—Muc-5AC (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)
September 2019
3-139
DRAFT: Do Not Cite or Quote
-------
Table 3-12 (Continued): Study-specific details from animal toxicological studies
of short-term ozone exposure and morphology—healthy.
Species (Strain), n, Sex, Exposure Details
Study Age (Concentration, Duration)
Endpoints Examined
Ona etal. (2016)
Mice (C57BL/6J WT, 0.5 ppm, 4 h for 1-9 days Histopathology, immunochemistry,
Rag2 deficient, and Il2rg mRNA expression (2-24 h PE)
deficient)
n = 6 males
Age: 6-8 weeks
Zhu etal. (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)
Kumaaai et al. Mice (C57BL/6 WT, Rag2 0.5 ppm, 4 h/day for 1 or
(2016) deficient, Il2rg deficient) 9 days
n = 6 males
Age: 6-8 week
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/cNTac)
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)
1 ppm, 1 h
Michaudel et al. Mice (C57BL/6J WT, ST2 0.3 ppm, 1 h
(2018) deficient, IL-33 deficient,
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)
September 2019
3-140
DRAFT: Do Not Cite or Quote
-------
Table 3-12 (Continued): Study-specific details from animal toxicological studies
of short-term ozone exposure and morphology—healthy.
Species (Strain), n, Sex, Exposure Details
Study 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; MLN = mediastinal lymph node; PE = post-exposure; SD = Sprague-Dawley; WKY = Wistar
Kyoto; WS = Wistar; WT = wild type.
September 2019
3-141
DRAFT: Do Not Cite or Quote
-------
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
HR (95% Cl)a
RR
All ages:
1.05 (0.99,
1.10)
<6 yr:
1.03 (0.92,
1.15)
6-18 yr:
1.23 (1.09,
1.39)
19-49 yr:
0.97 (0.89,
1.07)
50+ yr:
1.04 (0.96,
1.12)
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 ofthe
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 with:
PM2.5
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 with:
NR
RR
0-4 DL: 1.05 (0.99,
1.11)
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
24-h avg
Warm season
(May-September)
Max: 60
Correlation (r):
NR
Copollutant
models with:
NR
No quantitative
results. Results
presented
graphically
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 =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
Max:
Temporal:
60.0
Spatio-
temporal: 60.3
Correlation (r): No quantitative
NR results. Results
Copollutant presented
models with: graphically
NR
September 2019
3-142
DRAFT: Do Not Cite or Quote
-------
Table 3-13 (Continued): Epidemiologic studies of short-term exposure to ozone
and hospital admission for asthma.
Study Exposure Copollutant Effect Estimates
Population Assessment Mean (ppb) Examination HR (95% Cl)a
Study
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
Max: 105.4
Correlation (r): Lag 0-1 RRs
PM2.5: 0.2
Copollutant
models with:
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
Max: 82.8
Correlation (r): Lag 0-3 RRs
NR
Copollutant
models with:
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
Max: 107
Correlation (r):
NR
Copollutant
models with:
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)
Note: fStudies published since the 2013 Ozone ISA.
aResults 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.
September 2019
3-143
DRAFT: Do Not Cite or Quote
-------
Table 3-14 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 HR (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
Mean: 18.4
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
Percent increase
Lag 1: 2.6 (0.2,
5.0)
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 with:
NR
Villeneuve et al.
(2007)
Alberta, Canada
Ozone: 1992-2002
Follow-up:
1992-2002
Time-series study
n = 57,912 Average of three monitors.
All ages 8-h max
Year-round and seasonal
(April-September,
October-March)
Summer:
Mean: 38.0
75th: 46.0
Winter:
Mean: 24.3
75th: 31.5
Correlation
(r): NR
Copollutant
models with:
NR
Lag 1 OR
All ages
Year-round: 1.04
(1.02, 1.07)
Winter: 1.02
(0.98, 1.06)
Summer: 1.07
(1.04, 1.10)
Ito et al. (2007) All aqes
Average of 16 monitors
All year:
Correlation
New York City, NY,
within 20 miles of the
Mean: 30.4
(r): NR
U.S.
geographic city center
95th: 68.0
Copollutant
Ozone: 1999-2002
8-h max
Warm:
models with:
Follow-up:
1999-2002
Year-round and seasonal
Mean: 42.7
PM2.5, NO2,
(April-September,
95th: 77.0
SO2, CO
Time-series study
October-March)
Cold:
Mean: 18.0
95th: 33.0
Percent increase
Lag 0-1
Warm season:
11.0 (7.1, 15.0)
September 2019
3-144
DRAFT: Do Not Cite or Quote
-------
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 HR (95% CI)
tSarnat et al.
(2013)
Atlanta, GA, U.S.
Ozone: 1999-2002
Follow-up:
1999-2002
Time-series study
n = 270,816 Zip-code centroid
All ages
estimates from a hybrid
model fusing spatially
interpolated background
O3 concentrations with
local-scale AERMOD
output
24-h avg
Year-round
Mean: 41.9 Correlation Lag 0-2 RRs
Median: 39.3
75th: 53.8
95th: 76.2
Max: 132.7
(r):
PM2.5: 0.51;
NOx: -0.03
Copollutant
models with:
NR
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.
Oznoe: 2001-2007
Follow-up:
2001-2007
Time-series study
All ages
One monitor
8-h max
Year-round
Correlation
(r): NR
Copollutant
models with:
NR
RR
0-4 DL: 1.05
(1.02, 1.08)
tSacks et al. (2014) n = 122,607
North Carolina
(statewide), U.S.
Ozone: 2006-2008
Follow-up:
2006-2008
Case-crossover
study
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:
All-year: 43.6;
warm
season: 50.1
75th: All-year:
54.3; warm
season: 59.2
Max: All-year:
108.1; warm
season:
108.1
Correlation
(r): PM2.5:
0.54
Copollutant
models with:
PM2.5
Lag 0-2 ORs
All-year: 1.02
(1.00, 1.04)
Warm season:
1.02 (1.00, 1.04)
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 Correlation Lag 0-2 RRs
Median: 53.3 (r):
75th: 67.7
PM2.5: 0.66;
NO2: 0.54;
SO2: 0.27
Copollutant
models with:
NR
Cold season
(November-April):
1.05 (1.01, 1.09)
Warm season
(May-October):
1.05 (1.02, 1.09)
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
Max:
80.2-106.3
Correlation
(r): NR
Copollutant
models with:
NR
RR
Lag 0-2
Atlanta:
1.03 (1.01, 1.05)
Birmingham:
1.03 (0.98, 1.08)
Dallas:
1.03 (1.00, 1.06)
Pittsburgh:
1.03 (1.00, 1.06)
St. Louis:
1.06 (1.02, 1.09)
September 2019
3-145
DRAFT: Do Not Cite or Quote
-------
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 HR (95% CI)
tGleason et al.
(2014)
New Jersey
(statewide), U.S.
Ozone:
April-September,
2004-2007
Follow-up:
April-September,
2004-2007
Case-crossover
study
n =21,854
Age: 3-17 yr
Fusion of monitor and
CMAQ modeling
8-h max
Warm season
(April-September)
Correlation
(r): PM2.5:
0.56
Copollutant
models with:
PM2.5
OR
Lag 0-2: 1.06
(1.05, 1.08)
tStrickland et al. n = 109,758
Population-weighted
Mean: 42.22 Correlation
RR
(2014) Age: 2-16 yr
monitor averages
(r): NR
Lag 0-2: 1.07
Atlanta, GA, U.S.
8-h max
(1.04, 1.09)
Ozone: 2002-2010
Year-round
Copollutant
Follow-up:
models with:
2002-2010
NR
Time-series study
tSarnat et al.
(2015)
St. Louis, MO, U.S.
Ozone: 2001-2004
Follow-up:
2001-2003
Time-series study
n = 34,086
All ages
One monitor
8-h max
Year-round
Mean: 36.2
Correlation
(r): PM2.5:
0.23;
NO2: 0.37;
SO2: -0.04;
SO42": 0.49;
NOs-: -0.57;
OC: 0.30;
EC: -0.09
RR
0-2 DL: 1.05
(1.00, 1.09)
Copollutant
models with:
PM2.5, S042",
EC, OC, NO2
tBvers etal. (2015)
Indianapolis, IN,
U.S.
Ozone: 2007-2011
Follow-up:
2007-2011
Time-series study
n = 165,056
Age: >5 yr
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 with:
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)
September 2019
3-146
DRAFT: Do Not Cite or Quote
-------
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 HR (95% CI)
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 with:
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
24-h avg
Warm season
(May-September)
Max: 60
Correlation
(r): NR
Copollutant
models with:
NR
Percent increase
Lag 0-3: 10.81
(6.84, 15.03)
tMaliq et al. (2016)
California
(statewide), U.S.
Ozone: 2005-2009
Follow-up:
2005-2008
Case-crossover
study
All ages Nearest monitor within Mean: 33-55
20 km of population across
weighted zip-code centroid climate zones
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
Percent increase
Lag 0-1
Year-round: 3.85
(2.17, 5.56)
Warm season:
4.18 (1.93, 6.48)
Copollutant
models with:
NO2, CO,
SO2
September 2019
3-147
DRAFT: Do Not Cite or Quote
-------
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 HR (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
Max:
Temporal:
60.0;
spatio-
temporal:
60.3
Correlation
(r): NR
Copollutant
models with:
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-km grid
cells
8-h max
Year-round
Correlation
(r): NR
Copollutant
models with:
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-km grid
cells
8-h max
Year-round
Mean: 42.1
75th: 50.9
Max: 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
OR
Lag 0-3: 1.03
(1.01, 1.05)
Copollutant
models with:
NR
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
Max: 80
Correlation
(r): NR
Copollutant
models with:
NR
OR
Lag 1:
Males: 1.03 (1.00,
1.06)
Females: 1.04
(1.00, 1.08)
Note: fStudies published since the 2013 Ozone ISA.
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.
September 2019
3-148
DRAFT: Do Not Cite or Quote
-------
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 etal. (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):
PM2.5: 0.55
Copollutant
models with:
NR
No quantitative
results. Results
presented
graphically.
Note: fStudies 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
September 2019
3-149
DRAFT: Do Not Cite or Quote
-------
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)
Table 3-18 Study-specific details from animal toxicological studies of short-term
ozone exposure and lung function—allergy.
Study
Species (Strain), n, Sex, Age
Exposure Details
(Concentration,
Duration)
Endpoints Examined
Scheleqle and Walbv
(2012)
Rats (BN) naive and
sensitized/challenged with
allergen
n = 5-11 males
Age: 8-10 weeks
1 ppm, 8 h
Lung function, breathing pattern
(immediately PE)
BN = brown Norway; PE =
post-exposure.
September 2019 3-150 DRAFT: Do Not Cite or Quote
-------
Table 3-19 Study-specific details from animal toxicological studies of short-term
ozone exposure and lung function—asthma.
Exposure Details
Study Species (Strain), n, 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 (durin9 exposure)
Age: 6 weeks
MCh = methacholine. PE = post-exposure.
September 2019
3-151
DRAFT: Do Not Cite or Quote
-------
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)
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 (2013)
Adults with
atopy
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)
September 2019
3-152
DRAFT: Do Not Cite or Quote
-------
Table 3-22 Study-specific details from animal toxicological studies of short-term
ozone exposure and inflammation, oxidative stress, and
injury—allergy.
Study
Species (Strain), n, Sex,
Age
Exposure Details
(Concentration,
Duration)
Endpoints Examined
Scheleqle and Walbv
(2012)
Rats (BN) naive and
sensitized/
challenged with allergen
n = 5-11 males
Age: 8-10 weeks
1 ppm,
BALF markers of injury,
inflammation (immediately PE)
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 (Strain), n, Sex,
Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Bao etal. (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.
September 2019
3-153
DRAFT: Do Not Cite or Quote
-------
Table 3-24 Study-specific details from animal toxicological studies of short-term
ozone exposure and morphology—asthma.
Exposure Details
Study Species (Strain), n, Sex, Age (Concentration, Duration) Endpoints Examined
Bao et al. (2013) Mice (BALB/c naive and 2 ppm, 3 h Mucosubstance secretion
ovalbumin and Muc5AC, epithelial cell
sensitized/challenged) density (24 h PE)
n = 6-7 females
Age: 6-8 weeks
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
Exposure
Mean
Copollutant
Effect Estimates
Study
Population
Assessment
PPb
Examination
HR (95% Cl)a
tDelfino etal. (2013)
n =45
One monitor
Mean:
Correlation (r):
Lag 0: 0.42
Riverside and Whittier,
Children with
(Riverside);
52.9
PM2.5: 0.39;
(-1.33, 2.19)
CA, U.S.
asthma
average of two
Median:
NO2: 0.07;
Lag 1: 0.63
Ozone:
Age: 9-18 yr
monitors (Whittier)
46.8
EC: 0.55;
(-1.05, 2.35)
August-December
8-h max
Max:
OC: 0.71
Lag 0-2: 1.24
2003 (Riverside);
Year-round
120.8
Copollutant
(-1.04, 3.57)
July-November 2004
models with:
(Whittier)
NR
Follow-up:
August-December
2003 (Riverside);
July-November 2004
(Whittier)
Panel study
Note: fStudies published since the 2013 Ozone ISA.
aResults 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.
September 2019
3-154
DRAFT: Do Not Cite or Quote
-------
Table 3-26 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 HR 95% Cla
Arbex et al. (2009)
n = 1,769
Average of four
Mean: 48.8
Correlation (r):
Percent
Sao Paulo, Brazil
All ages
monitors
75th: 61.0
NR
increase
Ozone: 2001-2003
1-h max
Max: 143.8
Copollutant
Women
Follow-up:
Year-round
models with:
NR
Lag 0: 1.0
2001-2003
(0.0, 2.0)
Time-series study
tRodopoulou et al.
n = 12,511
One monitor
Mean: 40
Correlation (r):
Percent
(2015)
Ages 15+
8-h max
Median: 39
PM2.5: 0.33
increase
Little Rock, AR, U.S.
Seasonal:
75th: 50
Copollutant
Lag 2: 2.29
Ozone: 2002-2012
cold season
models with:
(-2.07,
Follow-up:
(October-March)
PM2.5
6.85)
2002-2012
and warm season
Time-series study
(April-September)
estimates
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
Year-round
NO2: 0.37;
SO2: -0.04;
0.98 (0.92,
1.06)
Follow-up:
SO42": 0.49;
2001-2003
NOs": -0.57;
Time-series study
OC: 0.30;
EC: -0.09
Copollutant
models with:
PM2.5, S042",
EC, OC, NO2
tMalia et al. (2016)
All ages
Nearest monitor
Mean: 33-55
Correlation (r):
Percent
California
within 20 km of
across climate
NO2: -0.01
increase
(statewide), U.S.
population-weighted
zones
YR; 0.26
Lag 0-1
zip-code centroid
1-h max
warm; SO2:
Ozone: 2005-2009
Follow-up:
-0.06 YR;
0.02 warm;
Year-round:
0.89 (-0.26,
2005-2008
Seasonal:
CO:
2.06)
Case-crossover
study
warm season
(May-October) and
year-round
estimates
-0.28 YR;
0.02 warm
Copollutant
models with:
Warm
season:
2.02 (0.46,
3.61)
NO2, CO, SO2
September 2019
3-155
DRAFT: Do Not Cite or Quote
-------
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
HR 95% Cla
tXiao et al. (2016)
Georgia (statewide),
U.S.
Ozone: 2002-2008
Follow-up:
2002-2008
Case-crossover
study
n = 84,597
Ages: 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
Max: 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 with:
NR
OR
Lag 0-3:
1.03 (1.00,
1.06)
tBarrv et al. (2018)
Five U.S. cities
Ozone: 2002-2008
Follow-up:
2002-2008
Time-series study
All ages
Fusion of CMAQ Mean:
model estimates 37.5-42.2
and ground-based 75^-
measurements; 50 1-54 4
population-weighted
average of 12-km
grid cells for each
city Max:
o u 80 2-106 3
8-h max
Year-round
90th:
59.3-63.5
models with:
NR
Correlation (r): RR
NR Lag 0-2
Copollutant At,anta:
1.00 (0.97,
1.03)
Birmingham:
0.99 (0.93,
1.05)
Dallas:
1.04 (0.99,
1.09)
Pittsburgh:
1.02 (0.98,
1.07)
St. Louis:
1.03 (0.99,
1.08)
tSzvszkowicz et al. n = 183,544
Average of all
Mean:
Correlation (r):
OR
£2018) Age: 55+ yr
monitors within
22.5-29.2
NR
Lag 1;
Multicity, Canada
35 km
across cities
Copollutant
females:
Ozone: 2004-2011
24-h avg
Max: 80
models with:
1.01 (0.99,
Follow-up:
Year-round
NR
1.03)
2004-2011
Lag 0;
Case-crossover
Study
males: 1.01
(0.99, 1.03)
Note: fStudies published since the 2013 Ozone ISA.
aResults 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.
September 2019
3-156
DRAFT: Do Not Cite or Quote
-------
Table 3-27 Epidemiologic studies of short-term exposure to ozone and
medication use in adults with chronic obstructive pulmonary disease
(COPD).
Study
Exposure
Mean
Copollutant
Effect Estimates
Study
Population
Assessment
PPb
Examination
95% Cla
tMaqzamen et al.
n = 35
Monitors
Median:
Correlation (r):
RR
(2018)
Age: 40+ yr
8-h max
17.21
NR
Lag 0: 0.98 (0.93,
Seattle and Tacoma,
Former
Year-round
75th:
Copollutant
1.45)
WA, U.S.
smokers with
24.37
models with:
Ozone: December
COPD but not
Max:
NR
2011 to October 2012
asthma
40.86
Follow-up: December
2011 to October 2012
Panel study
Note: fStudies published since the 2013 Ozone ISA.
aResults 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.
Table 3-28 Study-specific details from animal toxicological studies of short-term
ozone exposure and lung function—chronic obstructive pulmonary
disease (COPD).
Study
Species (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
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.
September 2019
3-157
DRAFT: Do Not Cite or Quote
-------
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).
Species (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
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).
Study
Species (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
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.
September 2019
3-158
DRAFT: Do Not Cite or Quote
-------
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)
Table 3-32 Study-specific details from animal toxicological studies of short-term
ozone exposure and lung function—obesity.
Study
Species (Strain), n, Sex,
Age
Exposure Details
(Concentration,
Duration)
Endpoints Examined
Williams et al. (2015)
Mice (C57BL/6) WT and Cpe
fat, 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)
Mathews et al. Mice (C57BL/6J)WT and 2 ppm, 3 h Pulmonary mechanics,
(2017b) db/db, TCR gamma delta challenge with MCh (24 h PE)
deficient mice on high-fat diet
for 24 weeks
n = 4-10 females
Age: 10 weeks and greater
than 24 weeks
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.
September 2019
3-159
DRAFT: Do Not Cite or Quote
-------
Table 3-33 Study-specific details from animal toxicological studies of short-term
ozone exposure and inflammation, oxidative stress, and
injury—obesity.
Species (Strain), n,
Study Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Yina etal. (2016)
Mice (Kkay)
n = 8 males
Age: 6-7 weeks
0.5 ppm, 4 h/day for 13 days
Lung tissue mRNA for
proinflammatory genes; T cell
subpopulations in lymph nodes (about
2 h PE)
Zhonq 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 olderthan
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 = C-C 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; PE = post-exposure; TCR = T cell receptor.
September 2019
3-160
DRAFT: Do Not Cite or Quote
-------
Table 3-34 Study-specific details from animal toxicological studies of short-term
ozone exposure and morphology—obesity.
Study
Species
(Strain), n,
Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Zhonq 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)
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
HR (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):
PM2.5: 0.22;
Other:
NOx: -0.35;
BC: 0.28;
OC: 0.24;
sulfate: 0.32;
PN: -0.78
Copollutant
models with:
PM2.5
Estimated change in
FeNO (ppb); Lag 1:
-5.95 (-10.79,
-0.90)
Note: fStudies published since the 2013 Ozone ISA.
aResults 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.
September 2019
3-161
DRAFT: Do Not Cite or Quote
-------
Table 3-36 Study-specific details from animal toxicological studies of short-term
ozone exposure and lung function—cardiovascular disease.
Species (Strain), n, Exposure Details
Study Sex, Age (Concentration, Duration) Endpoints Examined
Dye etal. (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. Mice (C57BL/6) 1 ppm, 4 h Airway responsiveness to MCh
(2016) control and mice with (18-20 h PE)
induced pulmonary
hypertension
n = 4-8 males
Age: 6-8 weeks
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.
September 2019
3-162
DRAFT: Do Not Cite or Quote
-------
Table 3-37 Study-specific details from animal toxicological studies of short-term
ozone exposure and inflammation, oxidative stress, and
injury—cardiovascular disease.
Study
Species (Strain), n, Sex,
Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Farrai et al. (2012)
Rats (SH)
n = 6 males
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 mRNA for HO-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
(2015)
Rats (WKY, SH, SPSH,
obese SHHF, obese
atherosclerosis prone JCR)
n = 3-4 males
Age: 10-12 weeks
1 ppm, 4 h
Lung gene expression profiling
(immediately PE)
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; 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 a; WKY = Wistar Kyoto, WS = Wistar.
September 2019
3-163
DRAFT: Do Not Cite or Quote
-------
Table 3-38 Study-specific details from animal toxicological studies of short-term
ozone exposure and morphology—cardiovascular disease.
Species (Strain), n, Sex,
Study Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Ramot etal. (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 = Wistar.
September 2019
3-164
DRAFT: Do Not Cite or Quote
-------
Table 3-39 Epidemiologic studies of short-term exposure to ozone and
emergency department (ED) visits for respiratory infection.
Effect
Estimates
Study Exposure Copollutant HR (95%
Population Assessment Mean ppb Examination CI)
Study
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
Mean: 18.4
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 with:
NR
Percent
increase
Lag 1: 1.00
(0.98, 1.02)
tWinquist et al. (2012)
All Ages
One monitor
Correlation
RR
St. Louis, MO, U.S.
8-h max
(r): NR
Pneumonia
Ozone: 2001-2007
Follow-up: 2001-2007
Time-series study
Year-round
Copollutant
models with:
NR
0-4 DL:
1.01 (0.98,
1.04)
tKousha and Rowe (2014)
n = 48,252
Three monitors
Mean: 16
S.6 Correlation
OR
Edmonton, Canada
All ages
8-h max
Median:
(r): NR
Lower
Ozone: 1999-2002
Follow-up: 1999-2002
Case-crossover study
Seasonal:
cold season
(October-March) and
warm season
(April-September)
estimates
17.8
Copollutant
models with:
NR
respiratory
disease
Lag 0;
Year-round:
1.07 (1.03,
1.10)
September 2019
3-165
DRAFT: Do Not Cite or Quote
-------
Table 3-39 (Continued): Epidemiologic studies of short-term exposure to ozone
and emergency department (ED) visits for respiratory
infection.
Effect
Estimates
Study
Exposure
Copollutant
HR (95%
Study
Population
Assessment
Mean ppb
Examination
CI)
tDarrow et al. (2014)
n = 80,399
Population-weighted
Mean: 45.9
Correlation
Lag 0-2
Atlanta, GA, U.S.
Age: 0-4 yr
monitor averages
Median:
(r):
RRs
Ozone: 1993-2010
8-h max
43.8
PM2.5: 0.3;
NO2: 0.37;
CO: 0.21
Year-round
Follow-up: 1993-2010
Seasonal:
75th: 58.7
Bronchitis:
Time-series study
cold season
(November-February)
and warm season
95th: 80.6
Max: 127.1
Copollutant
models with:
NR
1.02 (0.99,
1.05)
URI: 1.03
(March-October)
(1.01, 1.05)
estimates
Pneumonia:
1.06 (1.03,
1.09)
tRodoDoulou et al. (2015)
n = 13,650
One monitor
Mean: 40
Correlation
Percent
Little Rock, AR U.S.
Age: 15+ yr
8-h max
Median: 39
(r):
increase
Ozone: 2002-2012
Seasonal:
75th: 50
PM2.5: 0.33
Acute Rl;
Follow-up: 2002-2012
Time-series study
cold season
(October-March) and
warm season
(April-September)
estimates
Copollutant
models with:
NR
Lag 2: -1.49
(-5.79,
3.00)
Pneumonia;
Lag 2: -8.19
(-16.64,
1.16)
tBarrv et al. (2018)
All ages
Fusion of CMAQ model
Mean:
Correlation
RR
Five U.S. cities
estimates and
37.5-42.2
(r): NR
URI—Lag
Ozone: 2002-2008
Follow-up: 2002-2008
Time-series study
ground-based
measurements;
population-weighted
average of 12-km grid
cells for each city
75th:
50.1-54.4
90th:
59.3-63.5
Copollutant
models with:
NR
0-2
Atlanta: 1.02
(1.01, 1.04)
Birmingham:
8-h max
Max:
1.02 (1.00,
80.2-106.3
1.05)
Year-round
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
Year-round
Copollutant
Lag 0;
Follow-up: 2004-2010
models with:
NR
Year-round:
1.04 (0.86,
1.21)
Case-crossover study
September 2019
3-166
DRAFT: Do Not Cite or Quote
-------
Table 3-39 (Continued): Epidemiologic studies of short-term exposure to ozone
and emergency department (ED) visits for respiratory
infection.
Study
Effect
Estimates
Study Exposure Copollutant HR (95%
Population Assessment Mean ppb Examination 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 with:
NO2, CO,
SO2
Percent
increase
Lag 0-1
Pneumonia
Year-round:
1.32 (0.20,
2.46)
Warm
season:
2.27 (0.32,
4.26)
ARI
Year-round:
2.15 (1.45,
2.86)
Warm
Season:
2.30 (1.24,
3.37)
URTI
Year-round:
3.77 (0.40,
7.26)
Warm
season:
3.14 (0.16,
6.21)
tXiao et al. (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
Max: 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 with:
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)
September 2019
3-167
DRAFT: Do Not Cite or Quote
-------
Table 3-39 (Continued): Epidemiologic studies of short-term exposure to ozone
and emergency department (ED) visits for respiratory
infection.
Effect
Estimates
Study
Exposure
Copollutant
HR (95%
Study
Population
Assessment
Mean ppb
Examination
CI)
tSzvszkowicz et al. (2018)
n = 717,676
Average of all monitors
Mean:
Correlation
OR
Multicity, Canada
All ages
within 35 km
22.5-29.2
(r): NR
URI—Lag 0
Ozone: 2004-2011
24-h avg
across
cities
Max: 80
Copollutant
Females:
Follow-up: 2004-2011
Year-round
models with:
NR
1.03 (1.02,
1.05)
Case-crossover study
Males: 1.02
(1.01, 1.04)
ALR—Lag 0
Females:
1.05 (1.02,
1.07)
Males: 1.02
(1.00, 1.05)
Note: fStudies published since the 2013 Ozone ISA.
aResults 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.
September 2019
3-168
DRAFT: Do Not Cite or Quote
-------
Table 3-40 Study-specific details from animal toxicological studies of short-term
ozone exposure and host defense/infection—healthy.
Study
Species (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.
September 2019
3-169
DRAFT: Do Not Cite or Quote
-------
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
HR (95% Cl)a
Katsouvanni et al.
(2009)
90 U.S. cities
32 European cities
12 Canadian cities
NMMAPS
APHEA
All ages
in each city.
1-h max
Year-round and
warm season
(April-September)
NMMAPS:
Correlation (r):
50th:
No
34.9-60.0
quantitative
75th:
results.
46.8-68.8
Results
APHEA:
presented
50th:
graphically.
11.0-38.1
Copollutant
75th:
models with:
15.3-49.4
NR
12 Canadian
cities:
50th: 6.7 8.3
75th: 8.4-12.4
Percent Increase
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)
Cakmak et al. (2006) All ages
10 Canadian cities
Average of monitors
Mean: 17.4
Correlation (r):
Percent increase
in each city.
Max (across
NR
Lag 1
24-h avg
cities):
Copollutant
3.3 (1.7, 4.9)
Year-round
38.0-79.0
models with:
NR
One monitor
Correlation (r):
RR
8-h max
NR
All ages
Year-round
Copollutant
0-4 DL: 1.00 (0.98,
models with:
1.03)
NR
2-18 yr
0-4 DL: 1.07 (1.00,
1.16)
tWinquist et al.
(2012)
St. Louis, MO, U.S.
Ozone: 2001-2007
Follow-up:
2001-2007
Time-series study
All ages
Note: fStudies published since the 2013 Ozone ISA.
aResults 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.
September 2019
3-170
DRAFT: Do Not Cite or Quote
-------
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
HR (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
Max: 147.5
PM2.5: 0.62;
NO2: 0.44;
SO2: 0.21;
CO: 0.27
SO42": 0.56;
TC: 0.52;
OC: 0.54;
EC: 0.40
Copollutant
models with:
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
Max: 180
24-h avg:
Mean: 30
75th: 37
Max: 81
8-h max:
Mean: 53
75th: 67
Max: 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
NO2:
CO:
0.25;
¦0.15;
¦0.17;
8-h max O3:
PM2.5: 0.46;
NO2: 0.24;
CO: 0.15
Copollutant
models with:
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 with:
NR
0-4 DL: 1.01
(1.00, 1.03)
September 2019
3-171
DRAFT: Do Not Cite or Quote
-------
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
HR (95% Cl)a
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
Max: 148
Correlation (r): Lag 1 RRs
PM2.5: 0.46;
NO2: 0.24
Copollutant
models with:
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 Mean: 33-55
within 20 km of across
population weighted climate zones
zip-code centroid
1-h max
Seasonal:
warm season
(May-October) and
year-round
estimates
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 with:
NO2, CO, SO2
Percent increase
Lag 0-1
Year-round: 2.07
(1.63, 2.52)
Warm season:
2.39 (1.56, 3.23)
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
Copollutant
models with:
NR
Atlanta:
1.02 (1.01, 1.04)
2002-2008
average of 12-km
90th:
Birmingham:
Time-series study
grid cells for each
59.3-63.5
1.02 (1.00, 1.05)
city
Max:
Dallas:
8-h max
80.2-106.3
1.04 (1.02, 1.06)
Year-round
Pittsburgh
1.02 (1.01, 1.04)
St. Louis
1.02 (1.00, 1.03)
September 2019
3-172
DRAFT: Do Not Cite or Quote
-------
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
HR (95% Cl)a
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
Mean:
model estimates and 40.0-42.2
ground-based
measurements;
12-km grid cells
area weighted to
ZCTAs
8-h max
Year-round
across cities
Max: 125
Correlation (r): Lag 0-2 ORs
NR
Copollutant
models with:
NR
St. Louis:
1.02 (0.99, 1.06)
Dallas:
1.03 (1.01, 1.06)
Atlanta:
1.06 (1.05, 1.09)
Note: fStudies published since the 2013 Ozone ISA.
aResults 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.
September 2019
3-173
DRAFT: Do Not Cite or Quote
-------
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
Assessment
Mean ppb
Copollutant
Examination
Effect
Estimates
(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 a.m. to 6 p.m.
avg) 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 with:
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
Chronic Disease
Average summer
(June-August)
Surveillance System concentrations of
n = 1,183,865
Children born in
Quebec
8-h midday O3
estimated using a
BME-LUR model.
Mean: 32.07
Median: 32.19
75th: 33.76
Max: 43.12
Correlation
(r): NR
Copollutant
models with:
NR
Asthma onset
HRs
Birth address:
1.20 (1.16, 1.23)
Time-varying
exposure:
1.23 (1.20, 1.27)
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 with:
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)
Note: fStudies published since the 2013 Ozone ISA.
aResults standardized to a 10-ppb increase in long-term ozone concentrations.
September 2019
3-174
DRAFT: Do Not Cite or Quote
-------
Table 3-44 Study-specific details from animal toxicological studies of long-term
ozone exposure and inflammation, oxidative stress, and
injury—allergy.
Species
(Strain), n, Exposure Details
Study Sex, Age (Concentration, Duration) Endpoints Examined
Chou etal. (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 = C-C motif chemokine receptor 3; FA = filtered air, PND = post-natal day;
PE = post-exposure.
September 2019
3-175
DRAFT: Do Not Cite or Quote
-------
Table 3-45 Study-specific details from animal toxicological studies of long-term
ozone exposure and lung function—allergy.
Species
(Strain), n, Exposure Details
Study 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. (2012) 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.
September 2019
3-176
DRAFT: Do Not Cite or Quote
-------
Table 3-46 Study-specific details from animal toxicological studies of long-term
ozone exposure and inflammation, oxidative stress, and
injury—healthy.
Species (Strain), n, Sex, Exposure Details
Study Age (Concentration, Duration) Endpoints Examined
Hunter et al. (2011) Rats 2 ppm, 3 h
n =4-6
Strain and sex: NR
Age: PND 6-28
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)
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 1 ppm, 3 h BALF total cell number and
and TLR4 deficient differential cell counts, albumin,
n = 3-14 females Muc-5AC; lung tissue mRNA for
. _ „ . chemokines, antioxidants, TLR4,
Age: 1, 2, 3 weeks 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)
0.8 ppm, 5 h/day for
BALF total cells, cell differentials,
n = 9-10 males,
1 day/week for 4 weeks
albumin (18 h PE)
9-10 females
Age: 20 weeks
Gordon et al. (2016a)
Rats (S-D)
n = 10 females
Age: 20 weeks
0.25 ppm, 5 h/day for
1 day/week for 6 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
BALF total cells, cell differentials,
albumin, NAG, GGT (24 h PE)
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)
Gabehart et al. (2014) Mice (BALB/c) 1 ppm, 3 h
n = NR
Sex: NR
Age: 3 days
Clay et al. (2014) Rhesus macaque (Macaca 0.5 ppm, 8 h/day for 5 days
mulatta) followed by 9 days of
n = 3-5 males FA—11 cYcles followed by
FA until 12 mo
Age: 1 mo
September 2019
3-177
DRAFT: Do Not Cite or Quote
-------
Table 3-46 (Continued): Study-specific details from animal toxicological studies
of long term ozone exposure and inflammation, oxidative
stress, and injury—healthy.
Species (Strain), n, Sex,
Study 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 etal. (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 etal. (2017)
Rats (LE)
n = 9-10 females
Age : NR but weight was
200 g, pregnant
0.4 ppm, 4 h/day for 2 days;
GDs 5-6
0.8 ppm, 4 h/day for 2 days;
GDs 5-6
BALF total protein, albumin,
LDH, NAG, GGT, total cell, and
differential cell count (GD 21)
BALF = bronchoalveolar lavage fluid; BN = brown Norway; FA = filtered air; GGT = gamma glutamyl transferase; IL = interleukin;
LE = Long-Evans; LPS = lipopolysaccharide; NAG = /V-acetyl-glucosaminidase; NGF = nerve growth factor; PE = post-exposure;
TLR4 = toll receptor 4; WKY = Wistar Kyoto; WS = Wistar.
September 2019
3-178
DRAFT: Do Not Cite or Quote
-------
Table 3-47 Study-specific details from animal toxicological studies of long-term
ozone exposure and morphology and other endpoints—healthy.
Study
Species
(Strain), n,
Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Lee et al. (2011)
Rats (S-D)
0.5 ppm, 6 h/day for 3 weekly cycles
Airway architectural
n = 7-9
either 5 days ozone and 2 days
parameters—diameter, length,
males
recovery or 2 days ozone and 5 days
branching angles of conducting
Age: 7 days
recovery
airways (56 days PE)
Murphy et al. (2013) Rhesus 0.5 ppm, 8 h Serotonin pathway components
macaque g.5 ppm, 8 h/day for 5 days followed by (PE, at 2 and 6 mo of age)
(,Macaca g days of FA_-| or 1 1 cydes
mulatta)
n = 3-4
males
Age: 1 mo
Murphy et al. (2014) Rhesus 0.5 ppm, 8 h/day for 5 days followed by Lung tissue mRNA and
macaque 9 days of FA—1 or 11 cycles immunostaining for NK-1R,
(Macaca TAC1/SP, Nur77 (PE, up to
mulatta) 25 weeks)
n = 3-4
males
Age: 1 mo
September 2019
3-179
DRAFT: Do Not Cite or Quote
-------
Table 3-48 Epidemiologic studies of long-term exposure to ozone and lung
function and development.
Study
Study
Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect
Estimates
(95% Cl)a
tEckeletal. (2012)
Multicity, U.S.
Ozone: 1990-1997
Follow-up: 1990-1997
Cohort study
Cardiovascular Cumulative sum of Mean: 39.7
Health Study
n = 3,382
Age: >65 yr
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
Median: 39.7
75th: 51.3
95th: 64.1
Max: 79.6
Correlation (r):
PM10: 0.96
Copollutant
models with: 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 a.m.
to 6 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 with:
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 a.m.
to 6 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 with:
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)
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 with: NR
FEV1 (percent
increase)
First year of
life exposure:
-1.12 (-2.60,
0.40)
Average
lifetime
exposure:
-1.30 (-3.88,
1.36)
Note: fStudies published since the 2013 Ozone ISA.
aResults standardized to a 10-ppb increase in long-term ozone concentrations.
September 2019
3-180
DRAFT: Do Not Cite or Quote
-------
Table 3-49 Study-specific details from animal toxicological studies of long-term
ozone exposure and morphology—allergy.
Species
(Strain), n, Exposure Details
Study 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
FA = filtered air; PE = post-exposure.
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
September 2019
3-181
DRAFT: Do Not Cite or Quote
-------
Table 3-50 Study-specific details from animal toxicological studies of long-term
ozone exposure and lung function—healthy.
Species (Strain), n,
Exposure Details
Study
Sex, Age
(Concentration, Duration)
Endpoints Examined
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
BALF = bronchoalveolar lavage fluid; BN = brown Norway; GD = gestational day; GGT = gamma glutamyl transferase;
LDH = lactate dehydrogenase; LE = Long-Evans; NAG = /V-acetyl-glucosaminidase; PE = post-exposure, S-D = Sprague-Dawley.
Table 3-51 Epidemiologic studies of long-term exposure to ozone and
development of chronic obstructive pulmonary disease (COPD).
Study
Study
Population
Exposure
Assessment
Copollutant
Mean ppb Examination
Effect Estimates
95% Cla
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 with:
PM2.5
COPD in adults with
asthma:
HR 2.05 (1.17, 3.60)
Note: fStudies published since the 2013 Ozone ISA.
aResults standardized to a 10-ppb increase in long-term ozone concentrations.
September 2019
3-182
DRAFT: Do Not Cite or Quote
-------
Table 3-52 Epidemiologic studies of long-term exposure to ozone and
respiratory infection.
Study
Exposure
Mean
Copollutant
Effect Estimates
Study
Population
Assessment
PPb
Examination
HR (95% Cl)a
tMaclntvre et al.
n =45,513
IDW average of three
Mean:
Correlation (r):
Otitis media
(2011)
All singleton
closest monitors within
28.2
NR
HR 0.96 (0.95, 0.97)
Georgia Basin Airshed
live births in the
50 km
Median:
Copollutant
(including Vancouver
Georgia Basin
26.1
models with:
and Victoria, British
Airshed,
Max'
NR
Columbia), Canada
followed for the
71.8
Ozone: 1999-2002
first 2 yr of life
Follow-up: 1999-2002
Cohort study
tSmith etal. (2016)
Multicity, northern
California, U.S.
Ozone: 1996-2010
Follow-up: 1996-2010
Case-control study
n =6,913
2-yr avg from the
Median:
Correlation (r): I
Cases are adult
nearest monitor
31.5
PM25: 0.25; 1
members of
8-h avg
Max: 67
NO2: -0.33;
Kaiser
SO2: -0.24;
Permanente
Other:
/
Northern
CO: -0.28 1
California with a
Copollutant
clinical
models with: '
diagnosis of TB
NR
and a
(
corresponding
J
anti-TB
(
prescription or
a positive TB
culture.
Controls were
matched 2-1 on
age, sex, and
race/ethnicity.
Age: >21 yr
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)
Note: fStudies published since the 2013 Ozone ISA.
aResults standardized to a 10-ppb increase in long-term ozone concentrations.
September 2019
3-183
DRAFT: Do Not Cite or Quote
-------
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
(June-August)
concentrations of
8-h midday O3
estimated using a
BME-LUR model.
Other
Mean: 30.57
Median: 30.8
75th: 32.42
Max: 38.92
Correlation (r): Hospital/ED
NR
Copollutant
models with:
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 with:
PM2.5
COPD in adults
with asthma:
2.05 (1.17, 3.60)
tBerhane et al. (2016)
Multicity, southern
California, U.S.
Os: 1992-2011
Follow-up: 1992-2000;
1995-2003;
2002-2011
Cohort study
Children's
Health Study
n = 4,602
Age: 10 and
15 yr-olds
9- or 10-yr avg of
10 a.m. to 6 p.m.
Ozone measured at
one monitor in each
of the eight
communities
Other
Mean: Range
across cohorts:
44.8-47.7
Correlation (r): Absolute
PM2.5: 0.54;
NO2: 0.38
Copollutant
models with:
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)
Note: fStudies published since the 2013 Ozone ISA.
aResults standardized to a 10-ppb increase in long-term ozone concentrations.
September 2019
3-184
DRAFT: Do Not Cite or Quote
-------
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 with:
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)
Note: fStudies published since the 2013 Ozone ISA.
aResults standardized to a 10-ppb increase in long-term ozone concentrations.
September 2019
3-185
DRAFT: Do Not Cite or Quote
-------
Table 3-55 Study-specific details from animal toxicological studies of long-term
ozone exposure and allergic sensitization—healthy.
Study
Species
(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.
September 2019
3-186
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
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. 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 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.
September 2019
DRAFT: Do Not Cite or Quote
3-187
-------
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 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.
September 2019
3-188
DRAFT: Do Not Cite or Quote
-------
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 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.
September 2019
3-189
DRAFT: Do Not Cite or Quote
-------
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., 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 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.
September 2019
3-190
DRAFT: Do Not Cite or Quote
-------
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 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.
September 2019
3-191
DRAFT: Do Not Cite or Quote
-------
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, 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.
September 2019
3-192
DRAFT: Do Not Cite or Quote
-------
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 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.
a(U.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).
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).
September 2019
3-193
DRAFT: Do Not Cite or Quote
-------
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.
https://doi.org/10.1080/089583703045Q5.
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: Ollisoa WM. (1997). Effects of prolonged simulated ambient ozone dosing patterns on human
pulmonary function and symptomatology. Pittsburgh, PA: Air & Waste Management Association.
Alexeeff. SE: Litoniua. AA: Suh. H: Sparrow. D: Vokonas. PS: Schwartz. J. (2007). Ozone exposure and lung
function: Effect modified by obesity and airways hyperresponsiveness in the VA Normative Aging Study.
Chest 132: 1890-1897. http://dx.doi.org/10.1378/chest.07-l 126.
Alexis. NE: Lav. JC: Zhou. H: Kim CS: Hernandez. ML: Kehrl. H: Hazucha. MJ: Devlin. RB: Diaz-Sanchez. D:
Pedea 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: Devlia 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.
Arbex. MA: de Souza Conceicao. GM: Cendoa SP: Arbex. FF: Lopes. AC: Movses. EP: Santiago. SL: Saldiva.
PHN: Pereira. LAA: Braga. 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 inFEVl. Chest 107: 621-628.
Ariomandi. M: Balmes. JR: Framptoa 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: Daltoa 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.
September 2019
3-194
DRAFT: Do Not Cite or Quote
-------
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.
Avol. EL: Trim. SC: Little. DE: Spier. CE: Smith. MN: Peng. RC: Linn. WS: Hackney. JD: Gross. KB: D'Arcv.
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/10.1164/airccm.153.3.8630571.
Bao. A: Liang. L. i: Li. F: Zhang. M. in: Zhou. X. in. (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-Nufiez. MC: Sienra-Monge. JJ: Ramirez-
Aguilar. M: Cortez-Lugo. M: Holguin. F: Diaz-Sanchez. D: Olia 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: L118-L129. http://dx.doi.org/10.1152/aiplung.00080.2Q13.
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/10.1183/09031936.00021210.
Bates. ML: Brenza. TM: Ben-Jebria. A: Bascom. R: Eldridge. MW: Ultmaa 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: Londoa SJ. (2016). Effect of
obesity on acute ozone-induced changes in airway function, reactivity, and inflammation in adult females.
PLoS ONE 11: eO 160030. http://dx.doi.org/10.1371/iournal.pone.0160030.
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: Gaudermaa WJ: Avol. E: Rapapport. E: Urmaa 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.
September 2019
3-195
DRAFT: Do Not Cite or Quote
-------
Berhane. K: Zhang. Y: Linn. WS: Rappaport. EB: Bastaia TM: Salam. MT: Islam. T: Lurmana 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
EnvironMed71: 507-513. http://dx.doi.org/10.1136/oemed-2013-101874.
Berry. M: Liov. PJ: Gelperiq 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.
Bhoopalaa V: Haa 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: Mtiller. 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 abiomarker of ozone-induced pulmonary inflammation. PLoS ONE 8: e81816.
http://dx.doi.org/10.1371/iournal.pone.0081816.
Brand. JD: Ballinger. CA: Tuggle. 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.00185.2012.
Brand. JD: Mathews. JA: Kasahara. PI: 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.
Bretoa CV: Salam. MT: Vora. H: Gaudermaa 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.
Browa JS: Batesoa 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: China S: Poissoa 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.00018.2015.
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: Atkinsoa RW: Kent. AJ: van Staa. T: Cook. DG: Andersoa 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/10.1164/rccm.201210-1758QC.
Che. L: Jia 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: Chea Z: Shea 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.
September 2019
3-196
DRAFT: Do Not Cite or Quote
-------
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 inNrf2. OxidMed Cell Longev 2013: Article #254069.
http://dx.doi.org/10.1155/2013/254069.
Cho. Y: Abu-Ali. G: Tashiro. H: Kasahara. PI: Browa 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/10.1165/rcmb.2017-0404QC.
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.
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.0090401.
Clay. E: Patacchini. R: Trevisani. M: Preti. P: Brana. M: Spina. P: Page. C. (2016). Ozone-induced hypertussive
responses in rabbits and guinea pigs. J Pharmacol Exp Ther 357: 73-83.
http://dx.doi.org/10.1124/ipet.115.230227.
Connor. AJ: Laskin. JP: 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.yexmp.2012.01.004.
Crouse. PL: Peters. PA: Hvstad. P: Brook. JR: van Ponkelaar. A: Martia 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: VanRvswvk. K: Kumarathasaa 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: Kleia 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: Kleia 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. JExpo Sci Environ Epidemiol 21: 10-19.
http://dx.doi.org/10.1038/ies.2009.49.
Pelfino. RJ: Staimer. N: Tioa. T: Gillea 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.2Q13.25.
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. 112.094359.
September 2019
3-197
PRAFT: Po Not Cite or Quote
-------
Dokic. D: Traikovska-Dokic. E. (2013). Ozone exaggerates nasal allergic inflammation. Prilozi 34: 131-141.
Durrani. F: Phelps. PS: 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/10.3109/01902148.2011.654045.
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: Duaa L: Yaa 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.222Q2.
Folinsbee. LJ: Horstmaa 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/10.1164/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: Stedmaa 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/10.1136/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.
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.
Framptoa 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.
Framptoa 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.
September 2019
3-198
DRAFT: Do Not Cite or Quote
-------
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: Sua R: Cervelli. JA: Choi. H: Mandal. M: Abramova. EV: Gow. AJ: Laskia JD: Laskia 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/kfwl92.
Freiier. JI: VanEiikeren. JCH: VaaBree. 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: Sebastiaa 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: L1129-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. RespirRes 13: 89. http://dx.doi.org/10.1186/1465-9921-13-89.
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: Urmaa 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.2019.5357.
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/NEJMoa040610.
Gauderman. WJ: Urman. R: Avol. E: Berhane. K: McConnell. R: Rappaport. E: Chang. R: Lurmana 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/NEJMoal414123.
Ghio. AJ: Soukup. JM: Dailev. LA: Richards. JH: Duncaa KE: Lehmana J. (2014). Iron decreases biological
effects of ozone exposure. Inhal Toxicol 26: 391-399. http://dx.doi.org/10.3109/08958378.2014.90833Q.
Gielea MH: Van Per Zee. SC: Van Wiinea 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: Gaudermaa 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. (2011a). 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/s00421 -011 -1839-x.
September 2019
3-199
PRAFT: Po Not Cite or Quote
-------
Gonzalez-Guevara. E: Carlos Martinez-Lazcano. J: Custodio. V: Hernandez-Ceroa 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: e0180522. http://dx.doi.org/10.1371/iournal.pone.0180522.
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: Phillips. PM: Beaslev. TE: Ledbetter. A: Avdia 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.
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.
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/10.1097/00043764-199309000-00Q14.
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.00027.2013.
Groves. AM: Gow. AJ: Massa. CB: Laskin. JD: Laskia 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: Koponea IK: Sorli. JB: Paidi. MP: Hansea SW: Clausen. PA: Nielsea GD:
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. DN: 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. 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.
Hatch. GE: Slade. R: Harris. LP: Mcdonnell. WF: Devlia RB: Korea 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/10.1164/airccm.150.3.8087337.
Hazucha. MJ: Folinsbee. LJ: Bromberg. PA. (2003). Distribution and reproducibility of spirometric response to
ozone by gender and age. J Appl Physiol (1985) 95: 1917-1925.
http://dx.doi.org/10.1152/iapplphvsiol.00490.20Q3.
September 2019
3-200
DRAFT: Do Not Cite or Quote
-------
Hebbera 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.
Heckbert. SR: Kooperberg. C: Safford. MM: Psatv. BM: Hsia. J: McTiernaa 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: Dye. 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: Pedea 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.ll.007.
Hernandez. ML: Lav. JC: Harris. B: Esther. CR: Brickev. WJ: Bromberg. PA: Diaz-Sanchez. D: Devlia RB:
Kleeberger. SR: Alexis. NE: Pedea 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: Btinger. J. (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.
Hoffmever. F: Sucker. K: Monse. C: Berresheim. H: Rosenkranz. N: Jettkant. B: Beine. A: Briining. T: Btinger.
J. (2013). Relationship of pulmonary function response to ozone exposure and capsaicin cough sensitivity.
Inhal Toxicol 25: 569-576. http://dx.doi.org/10.3109/08958378.2Q13.812699.
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: Hanrahaa 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/s41590-017-0Q13-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/10.1177/07482337950110Q401.
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/10.1164/airccm/142.5.1158.
September 2019
3-201
DRAFT: Do Not Cite or Quote
-------
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: Gaudermaa WJ: Avol. E: Peters. JM: Gilliland. FD. (2008). Ozone, oxidant defense
genes and risk of asthma during adolescence. Am J Respir Crit Care Med 177: 388-395.
http://dx.doi.org/10.1164/rccm.200706-863QC.
Ito. K: Thurston. GD: Silvermaa 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: Beckermaa BS: Turner. MC: Krewski. D: Thurston. G: Martin. RV: van Donkelaar. A:
Hughes. E: Shi. Y: Gapstur. SM: Thua 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/10.1164/rccm.2013Q3-
0609QC.
Jerrett. M: Burnett. RT: Pope. CA. Ill: Ito. K: Thurston. G: Krewski. D: Shi. Y: Calle. E: Thua M. (2009).
Long-term ozone exposure and mortality. N Engl J Med 360: 1085-1095.
http://dx.doi.org/10.1056/NEJMoa08Q3894.
Jiu-Chiuaa C: Schwartz. J. (2008). Metabolic syndrome and inflammatory responses to long-term particulate air
pollutants. Environ Health Perspect 116: 612-617.
Jones. SL: Kittelsoa J: Cowaa JO: Flannerv. EM: Hancox. RJ: McLachlaa 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/10.1164/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: Maddea 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. PI: 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.00235.2Q13.
Kasahara. PI: Kim. HY: Williams. AS: Verbout. NG: Tran. J: Si. H: Wurmbrand. AP: Jastrab. J: Hug. C:
Umetsu. PT: 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/10.4Q49/iimmunol.1102363.
Kasahara. PI: 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.
September 2019
3-202
PRAFT: Po Not Cite or Quote
-------
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.0065829.
Katsouvanni. K: Samet. JM: Andersoa 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:
Pedea 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: Masoa 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: Tillmana 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.
Kousha. T: Castner. J. (2016). The air quality health index and emergency department visits for otitis media. J
Nurs Scholarsh48: 163-171. http://dx.doi.org/10.Ill 1/inu. 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/sl3382-014-0229-0.
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: VanDvkea 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/10.1165/rcmb.2015-0118QC.
Kumagai. K: Lewandowski. RP: Jackson-Humbles. DN: Buglak. N: Li. N: White. K: VanDvkea 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: Biarnasoa 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.
September 2019
3-203
DRAFT: Do Not Cite or Quote
-------
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.M112.43844Q.
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: Chea 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.llll/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: Mooa 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.
Lerov. P: Tham. A: Wong. H: Tennev. R: Chea C: Stiner. R: Balmes. JR: Paauet. AC: Ariomandi. M. (2015).
Inflammatory and repair pathways induced in human bronchoalveolar lavage cells with ozone inhalation.
PLoS ONE 10: e0127283. http://dx.doi.org/10.1371/iournal.pone.0127283.
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.ll.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.0b013e3181fdb60f.
Madden. MC: Stevens. T: Case. M: Schmitt. M: Diaz-Sanchez. D: Bassett. M: Montilla. TS: Berntsea 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/10.1186/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/10.1136/oemed-2017-104808.
Malig. BJ: Pearson. PL: Chang. Y: Broadwia 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: Jacksoa WT: Siddiaui. SR: Spencer. CY: Mitchell. NC: Hague.
IU: Johnstoa 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. PI: 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.0181017.
September 2019
3-204
PRAFT: Po Not Cite or Quote
-------
Mathews. JA: Kasahara. DI: Ribeiro. L: Wurmbrand. AP: Ninia 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. DI: Hutchinson. J: Cho. Y: Brand. JD: 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: Londoa 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.
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/10.1164/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: Devlia RB: O'Neil. JJ: Horstman. PH.
(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: Paa WK: Paa J. (1999b). Ozone-induced respiratory symptoms:
Exposure-response models and association with lung function. Eur Respir J 14:
845-853. https://eri.ersiournals.eom/content/14/4/845.long.
Mclntosh-Kastrinskv. R: Diaz-Sanchez. D: Sextoa KG: Jania. CM: Zavala. J: Tillev. SL: Jaspers. I: Gilmour.
MI: Devlia 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-Ramoa 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: Lombard! 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: Taa 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.
September 2019
3-205
DRAFT: Do Not Cite or Quote
-------
Mikerov. AN: Cooper. TK: Wang. G: Hu. S: Umstead. TM: Phelps. PS: 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: Dye. JA: Ledbetter. AD: Schladweiler. MC: Richards. JH: Snow. SJ: Wood. C'E: 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. 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/10.1186/sl3293-016-0069-7.
Moore. BP: Hyde. D: Miller. L: Wong. E: Frelinger. J: Schelegle. ES. (2012). 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/10.1159/000336835.
Moore. K: Neugebauer. R: Lurmana F: Hall. J: Braier. V: Alcora S: Tager. I. (2008). Ambient ozone
concentrations cause increased hospitalizations for asthma in children: An 18-year study in Southern
California. Environ Health Perspect 116: 1063-1070. http://dx.doi.org/10.1289/ehp. 10497.
Mortimer. KM: Neas. LM: Dockerv. DW: 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: Dockerv. DW: 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. DE. (1997). Comparison of survey and physician
claims data for detecting hypertension. J Clin Epidemiol 50: 711-718. http://dx.doi.org/10.1016/SQ895-
4356(97)00019-X.
Murgia. N: Brisman. J: Claessoa A: Muzi. G: Olin. AC: Torea 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. DM: 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.00324.2Q13.
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.2012-0110QC.
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.
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.
September 2019
3-206
DRAFT: Do Not Cite or Quote
-------
Neophvtou. AM: White. MJ: Oh. SS: Thakur. N: Galanter. JM: Nishimura. KK: Pino-Yanes. M: Torgersoa DG:
Gignoux. CR: Eng. C: Nguvea 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-Cintroa W: Bibbins-
Domingo. K: Thvne. 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/10.1164/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: Nguvea 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. J. R.: Sea S: Lurmann. F: Balmes. J. R.: 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/10.1164/rccm.201302-0264QC.
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/10.1186/sl2940-017-0244-2.
O'Connor. GT: Neas. L: Vaughn. B: Kattan. M: Mitchell. H: Crain. EF: Evans. R. Ill: Gruchalla. R: Morgaa W:
Stout. J: Adams. GK: Lippmana M. (2008). Acute respiratory health effects of air pollution on children with
asthma inUS 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-2015-206530.
Ong. CB: Kumagai. K: Brooks. PT: Brandenberger. C: Lewandowski. RP: Jackson-Humbles. DN: Nault. R:
Zacharewski. TR: Wagner. JG: Harkema. JR. (2016). Ozone-induced type 2 immunity in nasal airways.
Development 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: Robertsoa 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/klv093.
Parker. JD: 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: Deepti. 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: Marstoa L: Seemungal. TA: Strachaa DP: Wedzicha. JA. (2011).
Outdoor air pollution and respiratory health in patients with COPD. Thorax 66: 591-596.
http://dx.doi.org/10.1136/thx.2010.155358.
Penard-Morand. C: Charpin. D: Raherisoa 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.
September 2019
3-207
DRAFT: Do Not Cite or Quote
-------
Oiaa 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: Jacauemia B: Nadif. R: Orvszczvn. MP: Siroux. V: Aguilera. I: Kauffmann. F: Ktinzli. N. (2009).
Total serum IgE levels are associated with ambient ozone concentration in asthmatic adults. Allergy 64: 40-
46. http://dx.doi.org/10.1111/i. 1398-9995.2008.01800.X.
Raizenne. ME: Burnett. RT: Stern. B: Franklia 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: Charpia 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. IU: 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.00270.2015.
Rice. MB: Liungmaa 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/10.1164/rccm.201308-1414QC.
Robertsoa 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: 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.
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.
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.1306940.
Salam. MT: Bvua 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.
September 2019
3-208
DRAFT: Do Not Cite or Quote
-------
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: Allea 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/s00441-016-2545-Q.
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. PI: 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.
Silvermaa 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.
Smith. GS: VanDenEeden. SK: Garcia. C: Shan. J: Baxter. R: Herring. AH: Richardson. DB: VanRie. 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: Gordoa 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: Mever. M: Gowdv. KM: Fessler. MB: Duncaa 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.Ml 16.732362.
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: Thurstoa GD: Citak. K: James. DJ: Bock. N: Speizer. FE: Haves. C.
(1988). Effects of ambient ozone on respiratory function in active, normal children. Am Rev Respir Dis 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.
September 2019
3-209
DRAFT: Do Not Cite or Quote
-------
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.Ml 16.756627.
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/10.1164/rccm.2009Q8-
1201QC.
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: Laskia 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: Shea J: Laskin. JD: Laskia 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. JP: Laskia PL. (2013). Ozone-induced injury and
oxidative stress inbronchiolar epithelium is associated with altered pulmonary mechanics. Toxicol Sci 133:
309-319. http://dx.doi.org/10.1093/toxsci/kft071.
Szvszkowicz. 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:
Jordaa 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.0018737.
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/iierphl308Q771.
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.
Thomsoa EM: Vladisavlievic. P: Mohottalage. S: Kumarathasaa 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.
September 2019
3-210
PRAFT: Po Not Cite or Quote
-------
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/10.1165/rcmb.2017-0357QC.
To. T: Zhu. J: Larsea K: Simatovic. J: Feldman. L: Rvckmaa K: Gershoa A: Lougheed. MP: Licskai. C: Chea
H: Villeneuve. PJ: Crighton. E: Su. Y: Sadatsafavi. M: Williams. D: Carlstea C. (2016). Progression from
asthma to chronic obstructive pulmonary disease (COPD): is air pollution a risk factor? Am J Respir Crit
Care Med 194: 429-438. http://dx.doi.org/10.1164/rccm.201510-1932QC.
Tolbert. PE: Kleia 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.7500625.
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/10.1164/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://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). (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-95/001F). 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] (pp. 1-166). (EPA/630/P-03/001F). 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://cft)ub.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://cfpub.epa.gov/ncea/isa/recordisplav.cfm?deid=247492.
September 2019
3-211
DRAFT: Do Not Cite or Quote
-------
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/eele280e77586de
985257b65005d37e7!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.
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://ofmpub.epa.gov/eims/eimscomm.getfile7p download id=526855.
Urmaa R: McConnell. R: Islam. T: Avol. EL: Lurmana FW: Vora. H: Linn. WS: Rappaport. EB: Gilliland. FD:
Gaudermaa 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/10.1136/thoraxinl-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.
Verheia 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.
Verheia 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.
Villeneuve. PJ: Chea 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/10.1186/1476-069X-6-40.
von Elm. E: Altman. DG: Egger. M: Pocock. SJ: Gotzsche. 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.
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: Cohea HW: Hall. CB: Kelly. K: Prezaat. 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.
September 2019
3-212
DRAFT: Do Not Cite or Quote
-------
Weir. C'H: Yeatts. KB: Sarnat. JA: Vizuete. W: Salo. PM: Jaramillo. R: Coha RD: Chu. H: Zeldia DC: Londoa
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.
Wendt. JK: Svmanski. E: Stock. TH: Chaa 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. PI: Wurmbrand. AP: Chea 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: Kleia 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-ll-70.
Wolkoff. P: Clausen. PA: Larsea ST: Hammer. M: Nielsea 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/kfV018.
Wood. AM: Harrison. RM: Semple. S: Avres. JG: Stocklev. RA. (2009). Outdoor air pollution is associated with
disease severity in al-antitrypsin deficiency. EurRespir J 34: 346-353.
http://dx.doi.org/10.1183/09031936.000879Q8.
Xiang. Y: Oin. XO: Liu. HJ: Taa 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.0031158.
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-v.
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-ll-52.
Ying. Z: Allea K: Zhong. J: Chea 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: Logaa G: Wixted. WE: Kou. JP: Chalupowicz. DG: Feldser. HG:
Sanchez. Y: Nie. H: Callahaa JF: Kerns. JK: Podolia 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/jpet. 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.
September 2019
3-213
DRAFT: Do Not Cite or Quote
-------
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/fbhYS.2017.0Q249.
Zhong. J: Allen. K: Rao. X: Ying. Z: Braunstein. Z: Kankanala. SR: Xia. C: Wang. X: Bramble. LA: Wagner.
JG: Lewandowski. R: Sun. O: 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/10.1080/08958378.2Q16.1179373.
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/c5txQ0259a.
Zu. K: Liu. X: Shi. L: Tao. G: Loftus. CT: Lange. S: Goodmaa 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: Campea 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.
September 2019
3-214
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
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). 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
September 2019
4-1
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
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 a 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
September 2019
4-2
DRAFT: Do Not Cite or Quote
-------
1 • Exposure: Short-term (on the order of minutes to weeks) inhalation exposure to relevant ozone
2 concentrations (i.e., <0.4 ppm for humans; <2 ppm for other mammals)
3 • Comparison: Human subjects that serve as their own controls with an appropriate washout period
4 or subjects compared to a reference population exposed to lower levels (when available), or, in
5 toxicological studies of mammals, an appropriate comparison group that is exposed to a negative
6 control (e.g., filtered air)
7 • Outcome: Cardiovascular effects
8 • Study Design: Controlled human exposure (i.e., chamber) studies; in vivo acute, subacute, or
9 repeated-dose toxicity studies in mammals, immunotoxicity studies
10 Epidemiologic Studies:
11 • Population: Any U.S., Canadian, European, or Australian population, including populations or
12 lifestages that might be at increased risk
13 • Exposure: Short-term ambient concentration of ozone
14 • Comparison: Per unit increase (in ppb)
15 • Outcome: Change in risk (incidence/prevalence) of cardiovascular effects
16 • Study Design: Epidemiologic studies consisting of panel, case-crossover, time-series studies,
17 case-control studies, and cross-sectional studies with appropriate timing of exposure for the
18 health endpoint of interest
4.1.3 Biological Plausibility
19 This subsection describes the biological pathways that potentially underlie cardiovascular health
20 effects resulting from short-term inhalation exposure to ozone. Figure 4-1 graphically depicts these
21 proposed pathways as a continuum of pathophysiological responses—connected by arrows—that may
22 ultimately lead to the apical cardiovascular events associated with short-term exposures to ozone at
23 concentrations observed in epidemiologic studies (e.g., ED visits and hospital admissions). This
24 discussion of how short-term exposure to ozone may lead to these cardiovascular events also provides
25 biological plausibility for the epidemiologic results reported later in this Appendix. In addition, most
26 studies cited in this subsection are discussed in greater detail throughout this Appendix.
September 2019
4-3
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
Short-term
Ozone
Exposure
Activation of
Sensory
Nerves in the
Respiratory
Tract
Exacerbation
of Conduction
Abnormalities
or Arrhythmia
Exacerbation
of Ischemic
Heart Disease/
Potential
Myocardial
infarction or
Stroke
Respiratory
Tract
Inflammation/
Oxidative
Stress
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.
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. The 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
response and/or cause distal pathophysiological events that can contribute to overt cardiovascular disease.
September 2019 4-4 DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
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 ct 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; Biller et al..
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). 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; Biller etal.. 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)
demonstrated 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 (Karoly 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; Lanzinger 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
ozone-induced impaired vascular function. We also note that clinical indicators of potential ischemia
September 2019
4-5
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
(e.g., 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.. 2016).
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 (Khderetal.. 1998). which as noted above, could lead to a worsening of IHD or HF (Figure 4-1).
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; Mcintosh-Kastrinskv et al.. 2013; Wang et al.. 2013; Farrai et al..
2012) studies of autonomic nervous system modulation—including limited evidence 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 or
HF.
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 4-1). 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).
September 2019
4-6
DRAFT: Do Not Cite or Quote
-------
4.1.4 Heart Failure, Impaired Heart Function, and Associated Cardiovascular
Effects
1 Heart failure refers to a set of conditions in which the heart's pumping action is compromised. In
2 congestive heart failure (CHF), the flow of blood from the heart slows and fails to meet the body's
3 oxygen demand. Edema from heart failure frequently occurs from increased sodium reabsorption
4 resulting in an increase in blood volume (hypervolemia) and fluid retention, which often causes swelling
5 in the lungs or other tissues (typically in the legs and ankles). The effect of short-term ozone exposure on
6 people with CHF, which is a chronic condition, is generally evaluated using ICD codes recorded when a
7 patient is admitted or discharged from the hospital or ED. The relevant diagnostic codes for heart failure
8 are ICD9 428 and ICD10 150. These codes encompass left, systolic, diastolic, and combined heart failure.
9 In experimental studies, indicators of heart failure include decreased contractility and/or relaxation in
10 response to pharmacological challenge, reduced ejection fraction (i.e., the percentage of blood pumped
11 from the ventricles during each contraction), reduced stroke volume (i.e., the volume of blood pumped
12 per contraction) and reduced cardiac output (stroke volume multiplied by heart rate), as well as decreases
13 in left ventricular developed pressure (LVDP). Of note, the most prevalent form of heart failure is
14 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
15 The 2013 Ozone ISA reported the results of several studies in the U.S., Canada, and the U.K., all
16 of which observed null results for the association between CHF-related emergency department or hospital
17 visits and ozone exposure averaged over either 8 or 24 hours. A few additional studies have been
18 conducted since the 2013 Ozone ISA with mixed results (Table 4-3). Specifically:
19 • While studies conducted in the U.K. and U.S. did not observe positive associations between CHF
20 (alone or combined with hypertensive heart disease) and 8-hour max ozone concentrations
21 (Rodopoulou et al.. 2015; Miloievic et al.. 2014). a study in St. Louis, MO reported a 5% increase
22 in ED visits (95% CI: 1, 9%): and hospital admissions (95% CI: 2, 9%) for CHF (Winquist et al..
23 2012) associated with 8-hour max ozone. Similarly, an additional study in St. Louis observed a
24 4% (95% CI: -1, 10%) increase in ED visits for CHF, which increased to 6% (95% CI: 0, 12%)
25 when CO was included in the model (Sarnat et al.. 2015). Copollutant models with either PM2 5 or
26 NO2 did not change the predicted risk for ozone.
27 • Studies evaluating the role of lifestage in ozone's effects on heart failure reported no notable
28 differences for older adults (>65 or 70 years) compared with other adult age groups (19-64 or
29 <70 years) (Miloievic et al.. 2014; Winquist et al.. 2012).
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.
September 2019
4-7
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
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 (Tankcrslcv et al.. 2010). Moreover, these authors
also reported a decrease in fractional shortening—an indicator of impaired cardiac contraction
characterized by the percent 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!dtminim„m) relative to FA exposure in isolated, perfused murine hearts (p < 0.05).
September 2019
4-8
DRAFT: Do Not Cite or Quote
-------
1 • Wang et al. (2013) reported at least some evidence of dissolved myofilaments (a potential
2 indicator of cardiac damage) in right ventricles by microscopy following short-term ozone
3 (0.8 ppm) exposure.
4 Although results from the studies mentioned above demonstrated an effect of short-term ozone
5 exposure on changes in heart function, other results from these studies showed no effect. That is:
6 • Mclntosh-Kastrinsky et al. (2013) reported that short-term ozone exposure (0.245 ppm) did not
7 result in changes in LVDP, dP/dtmaximum, or coronary flow rate relative to FA exposure in isolated,
8 perfused murine hearts prior to ischemia. Moreover, following ischemia/reperfusion there was no
9 difference between ozone and FA exposure with respect to time to ischemic contracture, recovery
10 of LVDP, or ischemia-induced infarct size. Similarly, Kurhanewicz et al. (2014) reported that in
11 mice, there were no differences in time to ischemic contracture, or coronary flow rate prior or
12 after ischemia with ozone exposure. They also reported no differences in the recovery of left
13 ventricular developed pressure or pressure development over time post-ischemia.
14 • Tankerslev et al. (2013) did not report changes in left ventricular pressure over time (dP/dtmimmam
15 or dP/^/^maximum) following short-term ozone (0.5,0.8 ppm) exposure relative to FA in wild-type
16 mice. Similarly, Zychowski et al. (2016) reported that short-term ozone (1.0 ppm) exposure did
17 not result in appreciable right ventricular hypertrophy in mice kept in normal oxygen conditions,
18 nor did ozone exacerbate right ventricular hypertrophy in hypoxia-induced mice.
19 • Ramot et al. (2015) reported no effect of ozone effect on heart pathology in rats (0.25, 0.5,
20 1.0 ppm).
21 Although not demonstrated by all studies, most of the studies presented above report some
22 indicator of impaired cardiac function following short-term ozone exposure (Table 4-5). In addition, there
23 is evidence suggesting that the atrial natriuretic peptide gene may mediate some of these ozone-induced
24 cardiovascular effects.
4.1.5 Ischemic Heart Disease and Associated Cardiovascular Effects
25 IHD is a chronic condition characterized by atherosclerosis and reduced blood flow to the heart.
26 Myocardial infarction (MI), more commonly known as a heart attack, occurs when heart tissue death
27 occurs secondary to prolonged ischemia due to occlusion of the coronary artery. The effect of short-term
28 ozone exposure on acute MI, complications from recent MI, and other acute or chronic IHD are generally
29 evaluated using ICD codes recorded when a patient is admitted or discharged from the hospital or
30 emergency department (ICD9: 410-414 or ICD10: 120-125). In experimental or epidemiologic panel
31 studies, indicators of MI include ST segment depression as measured by an electrocardiograph (ECG).
32 The ST segment of an electrocardiogram recorded by surface electrodes corresponds to the electrical
33 activity of the heart registered between ventricular depolarization and repolarization, and is normally
34 isoelectric.
September 2019
4-9
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
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). In contrast, a study in
London, reported null results for angina (OR: 0.98; 95% CI: 0.94, 1.03) (Poloniecki et al.. 1997).
• 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).
• 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).
September 2019
4-10
DRAFT: Do Not Cite or Quote
-------
Reference
Location
Notes
Lag
Type
Poloniecki etal. 1997
London, England
0
i
i
Angina pectoris
IVidale et al. 2017
Como, Italy
0
AMI
IBhaskaran etal. 2011
UK
MINAP
0-3
0 |"
MI
IFinnbjornsdottir etal. 2013
Reykjavik, Iceland
0
O 1—
1
Angina pectoris
IClaeys et al. 2015
Belgium
with PCI
0-5
—1—•—
STEMI
IRasche et al. 2018
Jena, Germany
2
0.29 (0.11, 0.86)
1
STEMI
IWang et al. 2015
Canada
Calgary
0
1
—•—
Incident MI
0
1
STEMI
0
r*
NSTEMI
Edmonton
0
O
Incident MI
0
STEMI
0
r*
NSTEMI
IMilojevic et al. 2014
England, Wales
MINAP
0-4
MI
1
STEMI
NSTEMI
HES
~1
IHD
-o-
1
MI
IButland etal. 2016
England, Wales
MINAP
0-2
MI
STEMI
t
NSTEMI
IBard et al. 2014
Strasbourg, France
0-1
1
-• r
i
MI
ISarnat etal. 2015
St. Louis, MO
0-2
—•!—
IHD
INuvolone et al. 2013
Italy, 6 cities
0-1
—• ~
i
Incident MI
Von Klot et al. 2005
Europe, 5 cities
1
| 1
i
• ~
—i 1 1
—~ Angina pectoris
—f
0.8 0.85
0.9 0.95 1 1.05 1.1 1.15
1.2
Odds Ratio (95% CI)
AMI = acute myocardial infarction; HES = Hospital Episode Statistics; IHD = ischemic heart disease; Ml = myocardial infarction;
MINAP = Myocardial Ischaemia National Audit Project; NSTEMI = non-ST-elevation myocardial infarction; PCI = percutaneous
coronary intervention; STEMI = ST-elevation myocardial infarction.
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-h max
averaging times, 20 ppb increase for 8-h avg times, and 15 ppb increase for 24-h 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 4-2 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
1 The 2013 Ozone ISA reported inconsistent results with respect to an association between
2 short-term ozone exposure and MI. One study reported elevated risks for recurrent Mis (Henrotin et al..
3 2010). while another observed no associations between short-term ozone exposure and ST-segment
4 depression in elderly men with a history of coronary artery disease (Delfino et al.. 2011).
September 2019
4-11
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
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 studies 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
September 2019
4-12
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
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 ofbiomarkers 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 limited evidence 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:
• Lanzinger et al. (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 (Zanobctti 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
In the last review, Brook etal. (2009) found no effect of ozone exposure alone on clinical
indicators of endothelial dysfunction, such as FMD. 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.
September 2019
4-13
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
• 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.
September 2019
4-14
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
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) (Haloncn et al.. 2009). but null results
were reported in Atlanta, London, and a multicity study in Canada (Stieb 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 average ozone
concentration were 1.18 (95% CI: 1.00, 1.41) in Helsinki, Finland, 1.13 (95% CI: 1.03, 1.24) in
September 2019
4-15
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
Gironde Department in France, and 1.16 (95% CI: 1.03, 1.29) in Stockholm County, Sweden
(Pradeau et al.. 2015; Raza et al.. 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 (Stranev
et al.. 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 (Pradeau 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).
• 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 average 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).
September 2019
4-16
DRAFT: Do Not Cite or Quote
-------
Reference
Location
Lag
Type
Dennekamp et al. 2010
Australia
0-1
•
OHCA
IStraney et al. 2014
Perth, Australia
0
3
OHCA
IRosenthal etal. 2013
Helsinki, Finland
0-3
lEnsor et al. 2013
Houston, TX
0-1
—O—
OHCA
IPradeau et al. 2015
Gironde, France
1
o
OHCA
IRaza et al. 2014
Stockholm, Sweden
0
•
CA
IMilojevic etal. 2014
UK
0-4
-Oi
o
—
»-
Arrythmias
AF
AVCD
ISarnat etal. 2015
St. Louis, MO
0-2
1
—•
1
Dysrhythmia
IWinquist et al. 2012
St. Louis, MO
0-4
-1
—<
k—
—
Dysrhythmia - ED
Dysrhythmia - HA
IRodopoulou etal. 2015
Little Rock, AR
1
0
Dysrhythmias
ISadeetal. 2015
Negev, Israel
0
o—
AF
1
0.8
1 1 1 1
0.9 1 1.1 1.2
Risk Ratio (95% CI)
I
1.3
\
1.4
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-h max avg
times, 20-ppb increase for 8-h avg times, and 15-ppb increase for 24-h 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
1 The 2013 Ozone ISA stated that many studies reported positive associations for
2 arrhythmia-associated endpoints, yet collectively, results were inconsistent. In a population of subjects
3 with an implantable cardioverter defibrillator (ICD), Metzger et al. (2007) observed no evidence of an
4 association for tachyarrhythmic events with an increase in ozone concentrations. In contrast, in a study of
5 nonsmoking adults, increased odds were observed for supraventricular ectopy (Sarnat et al.. 2006). In the
September 2019
4-17
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
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 etal. (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). CHE
studies published since the 2013 Ozone ISA provide limited evidence that short-term ozone exposure
(2-3 hours) can appreciably effect 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
September 2019
4-18
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
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 very limited evidence from 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-hours ozone exposure, both
levels of ozone reduced increased sensitivity to aconitine-induced arrhythmia (p < 0.05). Similar
results were also found in an another study by this group (Farrai et al.. 2016V
• 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 etal. (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. Hypertension, the clinically relevant consequence of
chronically high blood pressure, typically develops over years. 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
September 2019
4-19
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
known as hypertension and can lead to a thickening of the ventricular wall resulting in diminished filling
during diastole. This 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 the
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
(Vencloviene 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 (Rodopoulou et al.. 2015) (Table 4-17).
4.1.8.2 Epidemiologic Panel Studies
Limited evidence was 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 (Zanobctti et al.. 2004). The evidence from recent
studies remains inconsistent and is characterized in Table 4-18. Specifically:
• Cakmaketal. (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). Additionally, in a cohort of Canadian children ages
September 2019
4-20
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
6-17 years of age, Dales and Cakmak (2016) observed increases in SBP (4.41; 95% CI: 1.91,
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
In the 2013 Ozone ISA, CHE studies indicated that short-term ozone exposure alone did not have
an effect on diastolic blood pressure (Sivagangabalan et al.. 2011; Brook et al.. 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; M 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 vs 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 to 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
September 2019
4-21
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
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), a key prognostic indicator, 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 (Fauchier et al.. 2004; Bigger etal.. 1992). 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 of NN 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 6-175], 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).
However, Cole-Hunter et al. (2018). who used 227 subjects from the TAPAS and EXPOsOMICS
cohorts in Barcelona, Spain, did not observe changes in heart rate when assigning spatially
September 2019
4-22
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
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) (Mirowskv et al.. 2017; Bartell et al.. 2013).
(Bartell et al.. 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, Mirowskv 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 limited
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 associated with 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.
• 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; Ariomandi et al.. 2015; Frampton et al.. 2015; Barath et al..
2013; Kusha et al.. 2012).
September 2019
4-23
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
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). Similarly.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 either time point.
• 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.
September 2019
4-24
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
• 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
(de 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
(Miloievic et al.. 2014) (Table 4-24).
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.
• 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).
September 2019
4-25
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
4.1.10.3 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 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 IYU.S. EPA. 2013a) pg 6-166] Since
the publication of the 2013 Ozone ISA, there is limited evidence from CHE studies that short-term ozone
exposure (1-2 hours) can result in changes to markers of coagulation or fibrinolysis. 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 (Ariomandi 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).
4.1.11 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
September 2019
4-26
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
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, these 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) 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 at 24-hours post-exposure. 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 day 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).
• A single study in a cohort of over 3,000 subjects evaluated 1-7 day moving avg exposure to
ozone reported no change in the oxidative stress biomarkers myeloperoxidase and indexed
8-epi-prostaglandin F2alpha (Li et al.. 2016).
September 2019
4-27
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
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:
• Billeretal. (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 hour 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 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; Billeretal.. 2011).
• Ramanathan et al. (2016) also demonstrated that ozone exposure (0.12 ppm) did not alter HDL
antioxidant or anti-inflammatory capacity in healthy adults.
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.
September 2019
4-28
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
4.1.11.3 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-tyrosine 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.
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).
September 2019
4-29
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
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, a multicity French study, and an analysis of stroke subtypes in Edmonton,
Canada. 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
(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
September 2019
4-30
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
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). 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; Miloievic
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 day 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).
September 2019
4-31
DRAFT: Do Not Cite or Quote
-------
Reference
Location
Lag
Type
i
Villeneuve et al. 2006
Canada
0-2
Ischemic
•
1 *
0-2
Hemorrhagic
!"•
m—>
i
0-2
TIA
*-j
-« A—
i
1 ¦
IChenet al. 2014
Canada
0-2
All
•—1
¦* A
1
¦
—~
IMaheswara et al. 2016
London, UK
0-6
Ischemic
I
1 O
IButland et al. 2017
London, UK
0
First-ever
O 1
Ischemic
o—!
T Trt i ,.,.1
IBedada et al. 2012
Manchester, UK
0
TIA, stroke
1
O
Liverpool, UK
0
TIA, stroke -*-•
1—
IVidale et al. 2017
Como, Italy
0
Ischemic
cj-
IMeclitouff et al. 2012
Rhone, France
NR
Ischemic <
•—1
IMilojevic et al. 2014
UK
0-4
o
IWing etal. 2015
Texas
0
First-ever
—1"°
IWingetal. 2017
Texas
1
First-ever, severe
1
-o
1
Recurrent
o i
IRodopoulou et al. 2015
Arkansas
1
All
-O 1
1
Henrotin et al. 2007
Dijon, France
1
Ischemic
1
1
•—
—~
ISuissa et al. 2013
Nice, France
1
All ischemic
1
Q
1
Recurrent
i
O—
—~
1
Large artery
•—1
i
IMontresor-Lopez et al.
South Carolina
0
All
—•—!
2015
0
Ischemic
• i
0
Hemorrhagic
O —
IXuetal. 2013
Pennsylvania
0-3
First-ever
0
1
Halonen et al. 2009
Finland
0-5
All
1—~
Larrieu et al. 2007
8 French cities
0-1
All
1
1 1
0.7 0.8
1 1 1 1
0.9 1 1.1 1.2
I
1.3
I
1.4
I
1.5
Risk Ratio (95% CI)
TIA = transient ischemic attack.
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-h max avg
times, 20-ppb increase for 8-h avg times, and 15-ppb increase for 24-h 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.
September 2019
4-32
DRAFT: Do Not Cite or Quote
-------
4.1.13 Nonspecific Cardiovascular Effects
4.1.13.1 Epidemiologic Studies of Emergency Department Visits and Hospital
Admissions
1 Several studies of ozone concentrations and cardiovascular hospital admissions and ED visits for
2 all CVD diagnoses combined were discussed in the 2013 Ozone ISA. With few exceptions, these studies
3 did not report an association between ozone concentrations and an increased risk of aggregated CVD in
4 populations in the U.S., Canada, Europe, and Australia.
5 • Recent studies that reported a risk ratio for combined cardiovascular disease outcomes show a
6 similar pattern to those studies included in the 2013 Ozone ISA (Table 4-32; Figure 4-5).
7 Although changes were small (<1%), associations were positive during the cold season and
8 negative during the warm season.
9 • Studies that evaluated effect modification by sex or age did not find notable differences
10 (Miloievic et al.. 2014; Rodopoulou et al.. 2014). Winquist et al. (2012) observed a higher
11 relative risk per 8-hour max ozone concentration among individuals residing in a poverty area.
September 2019
4-33
DRAFT: Do Not Cite or Quote
-------
Reference
Location
Lag
Notes
i
i
IVidale et al. 2017
Como, Italy
0
1
• o
1
tMilojevic et al. 2014
UK
0-4
1
°1
1
IRodopoulou et al. 2014
Dona Ana County, NM
1
ED
1 • ~
1
HA
i—0
1
ISarnat et al. 2015
St. Louis, MO
0-2
1
—a—
1
1
IWinquist et al. 2012
St. Louis MSA, MO
0-4
ED
1
+¦
0-4
HA
-or-
1
IRodopoulou et al. 2015
Little Rock, AR
1
1
-o
1
tHunova etal. 2013
Prague, Czech Republic
1
1
1
IHunova etal. 2017
Prague, Czech Republic
1
1
—#4
i
IChoi etal. 2011
Maryland
0-4
i
1 —0
1
1
1 1—
0.8 0.85
—1 1 1 1 1 1 \
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-h max avg
times, 20-ppb increase for 8-h avg times, and 15-ppb increase for 24-h 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.
September 2019
4-34
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
4.1.14 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 days for a 20-ppb increase in 8-hour max ozone concentrations.
• Klemm et al. (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 days
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 limited evidence from 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% (95% CI: -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.
September 2019
4-35
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
(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 et al.. 2014) or CHF (Sarnat et al.. 2015) after evaluating a copollutant
model with NO2. A study in St. Louis observed a 4% (95% CI: -1, 10%) elevation in ED visits
for CHF, which was increased to a 6% increased risk (95% CI: 0-12%) 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 et al.. 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 (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).
September 2019
4-36
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
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. Specifically, the evidence from controlled human exposure studies provided
support for increased decrements in FEVi and greater inflammatory responses to ozone in individuals
with asthma than in healthy individuals without a history of asthma. Studies of short-term ozone exposure
and mortality provided limited 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)
• Lanzinger et al. (2014) reported FMD decreases in individuals with type 2 diabetes. However,
Mirowsky 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).
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
September 2019
4-37
DRAFT: Do Not Cite or Quote
-------
1 studies then examining whether the magnitude of the ozone-cardiovascular health association varies either
2 across seasons or in the summer/warm season compared with the entire year. Studies evaluated in the
3 2013 Ozone ISA reported evidence of positive ozone-cardiovascular health associations in all-year
4 analyses that tended to be larger in magnitude during the warm or summer months. A limited number of
5 recent studies that conducted seasonal analyses reported associations that were similar for both warm
6 season and cool season analyses, Specifically, recent studies indicate:
7 • Evans et al. (2016) observed increased odds of STEMI in the cooler months (November to April)
8 for ozone exposure at 12-hours (OR= 1.43; 95% CI: 1.03, 1.98), 24-hours (OR= 1.45; 95% CI:
9 1.04, 2.03), and 72-hours (OR = 1.60; 95% CI: 1.05, 2.46), and no increased associations during
10 the warmer months.
11 • Increases in TNFR2 were associated with short-term ozone exposure in a cohort of over
12 3,000 subjects. When stratified by warm and cool seasons, the associations remained positive in
13 both season (Li et al.. 2017).
14 • Seasonality altered cardiovascular electrophysiology in a population of 8,662 Ottawa and
15 Gatineau patients referred for 24-hour ambulatory cardiac monitoring with exposure linked to the
16 3-hour max exposure for the 24-hours prior to the visit, based on the monitor closest to their
17 home address. During the warm season (April-September), Cakmak et al. (2014) reported
18 increases in the presence of heart block (1.23; 95% CI: 1.07, 1.42). However, in the cold season,
19 the same study reported increases in the number of supraventricular ectopic runs (defined as more
20 than three consecutive beats) (8.15; 95% CI: 0.34, 16.57) and the length of the longest ventricular
21 ectopic runs (20.68; 95% CI: 5.3, 38.31).
4.1.17 Summary and Causality Determination
22 The 2013 Ozone ISA concluded that the strongest evidence for an effect of short-term ozone
23 exposure on cardiovascular health was from animal toxicological studies demonstrating ozone-induced
24 impaired vascular and cardiac function, as well as changes in HR and HRV (U.S. EPA. 2013a). This
25 evidence was supported by a limited number of controlled human exposure studies in healthy adults
26 demonstrating changes in HRV, as well as in blood markers associated with an increase in coagulation,
27 systemic inflammation, and oxidative stress. Evidence of these effects in animals and humans was cited
28 as providing biological plausibility for the evidence from epidemiologic studies reporting positive
29 associations between short-term ozone exposure and cardiovascular-related mortality. However, there was
30 limited or no evidence from controlled human exposure or epidemiologic studies for short-term ozone
31 exposure and cardiovascular morbidity, such as effects related to HF, IHD and MI, arrhythmia and
32 cardiac arrest, or thromboembolic disease. The lack of evidence connecting the effects observed on
33 impaired vascular and cardiac function in animal toxicological studies and the association between
34 short-term ozone exposure and cardiovascular mortality observed in epidemiologic studies was a major
35 source of uncertainty in the 2013 Ozone ISA.
36 Animal toxicological studies published since the 2013 Ozone ISA provide generally consistent
37 evidence for impaired cardiac function and endothelial dysfunction, but limited or inconsistent evidence
September 2019
4-38
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
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; no effect of ozone was reported
from studies of cardiac function, indicators of IHD (i.e., ST segment), endothelial dysfunction, or HR,
while other studies provide limited evidence 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 ISA
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 limited evidence from 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 4.1.6.2V In addition, results from
September 2019
4-39
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
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').
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 4.1.9.2V 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 published
since the last review report inconsistent results. 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
September 2019
4-40
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
ozone exposure and cardiovascular health effects contribute evidence to determine if children are at a
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 4.1.16.1). 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).
September 2019
4-41
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
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 4-1). 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.
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 Indicators of impaired heart function, Section 4.1.4.3. ~0.2to0.3ppm
evidence from animal endothelial dysfunction Section 4.1.6.3
toxicological studies at
relevant ozone
concentrations
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
Farraietal. (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
September 2019
4-42
DRAFT: Do Not Cite or Quote
-------
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
Limited and inconsistent
evidence from controlled
human exposure studies
at relevant ozone
concentrations
No changes in a number of electrophysiology Rich et al. (2018)
measures by ECG, but there was increased
probability of ventricular but not
supraventricular ectopy couplets or runs,
Change in T-wave alternans during the first 5 Kusha et al. (2012)
min of exposure, but no change relative to
FA later in the exposure. The effect observed
in first 5-minutes is likely not meaningful.
0.07 but not
0.12 ppm
0.12 ppm
No meaningful changes in SBP and/or DBP
Frampton et al.
(2015)
Barath et al. (2013)
Ariomandi et al.
(2015)
Rich et al. (2018)
Changes in HRV
Barath et al. (2013)
Rich et al. (2018)
Ariomandi et al.
(2015)
Markers of coagulation, systemic
inflammation and oxidative stress
Kahle et al. (2015)
Barath et al. (2013)
Ariomandi et al.
(2015)
Frampton et al.
(2015)
Section 4.1.11.2
No evidence from
controlled human
exposure studies at
relevant ozone
concentrations
Changes in stroke volume or left ventricular Frampton et al.
ejection time (2015)
Changes in ST segment
Rich et al. (2018)
Clinical indicators of endothelial dysfunction Section 4.1.6.2
Changes in HR
Frampton et al.
(2015)
Barath et al. (2013)
Ariomandi et al.
(2015)
Rich et al. (2018)
Kusha et al. (2012)
Consistent evidence from A number of studies evaluated in the 2013
high-quality,
epidemiologic studies of
cardiovascular mortality
Section 4.1.14
Ozone ISA reported positive associations for
cardiovascular mortality in all-year and
seasonal analyses. A more limited number of
recent studies continue to report positive
associations.
September 2019
4-43
DRAFT: Do Not Cite or Quote
-------
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
Limited epidemiologic
evidence from multiple
studies of CVD hospital
admissions or ED visits
Section 4.1.4.
Section 4.1
Section 4.1
Section 4.1.12
Generally null or inconsistent associations
(both negative and positive direction)
observed in studies of CVD hospital
admissions or ED visits limited by low ozone Section 4.1.8
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.
Mean: 20-40 ppb
75th : 27-50 ppb
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
Limited epidemiologic
evidence from
copollutant models
provides some support
for an independent ozone
association
The magnitude of ozone associations
remains relatively unchanged, but in some
cases with wider confidence intervals in a
limited number of studies evaluating
copollutant models, including PM2.5 and
other gaseous pollutants.
When reported, correlations with PM2.5 or
gaseous copollutants were primarily in the
low to moderate range (r< 0.7).
Section 4.1.15
Uncertainty due to limited
coherence between CVD
morbidity and CVD
mortality
Consistent positive associations observed in
studies of short-term ozone exposure and
mortality, although limited evidence 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
September 2019
4-44
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
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
Uncertainty regarding
geographic heterogeneity
in ozone associations
Multicity U.S. studies demonstrate city-to-city
and regional heterogeneity in ozone-CVD ED
visit and hospital admission associations.
Evidence supports that a combination of
factors, including composition and exposure
factors may contribute to the observed
heterogeneity.
Section 4.1.5.1.
Section 4.1.7.1.
Section 4.1.12.1.
Section 4.1.13.1
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, 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.
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 limited evidence 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
September 2019
4-45
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
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
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
September 2019
4-46
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
• 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.
Modulation of
the Autonomic
Nervous System
(e.g., HRV, HR)
Long-term
Ozone
Exposure
Respiratory
Tract
Inflammation/
Oxidative
Stress
Systemic
Inflammation/
Oxidative Stress
Athens
clerosis
A
1
Impaired
Vascular
Function
'
Alte
Hemo
red
stasis
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.
Figure 4-6: Potential biological pathways for cardiovascular effects following
long-term exposure to ozone.
September 2019
4-47
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
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 into 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
response. Thus, it is important to note that there is evidence from long-term experimental studies in
animals (Miller etal.. 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 anti-coagulation 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 chemo attractants
that recruit inflammatory cells. Macrophages may then internalize circulating lipids, leading to the
formation of foam cells: a hallmark of atherosclerotic lesions. Over time, 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,
September 2019
4-48
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
one study in the 2013 Ozone ISA reported enhanced aortic atherosclerotic lesions in mice following
long-term ozone exposure m.S. EPA (2013a). pg. 7-38], However, considerable uncertainty remains in
how long-term ozone exposure may lead to mortality given that there is little epidemiologic evidence of
an association between long-term exposure to ozone and other cardiovascular endpoints such as IHD,
stroke, or thromboembolic disease. 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 ct 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.. 2014V
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 et al.. 2013) and a cohort study conducted in South Korea observed an
inverse association (Kim et al.. 2017) (Table 4-34).
September 2019
4-49
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
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
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 a negative 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 (Kim et al.. 2017) (Table 4-37).
September 2019
4-50
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
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) (Pcrcpu et al.. 2012;
Sethi et al.. 2012) (Table 4-38). 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.
4.2.6 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)
I (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).
September 2019
4-51
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
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:
• 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 et al.. 2017; Liu et al.. 2016; Dong et al.. 2013b; Zhao et al.. 2013;
Chuang et al.. 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 etal.. 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 etal.. 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 when in women. Similarly,
September 2019
4-52
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
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.
4.2.9 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.
September 2019
4-53
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
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 7-38], 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 animal toxicological studies that long-term exposure to ozone can result in changes in
mRNA levels of coagulation factors (Table 4-43).
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).
September 2019
4-54
DRAFT: Do Not Cite or Quote
-------
1 pg 7-39], In addition, there was evidence that long-term exposure to ozone decreased SOD enzyme
2 activity and increased levels of malondialdehyde. Recent studies provide some evidence that long-term
3 exposure (4-17 weeks) to ozone can result in an increase in markers of inflammation and oxidative stress
4 (Table 4-44). Specifically:
5 • The same laboratory cited in the 2013 Ozone ISA reported that in rats, long-term exposure to
6 ozone resulted in an increase in myocardial production of TNF-a (p < 0.05) (Perepu et al.. 2012;
7 Sethi etal.. 2012). This laboratory (Perepu et al.. 2012) also reported a decrease in the
8 anti-inflammatory cytokine IL-10 following long-term exposure to ozone (0.8 ppm).
9 • In rats, Miller etal. (2016) also reported an increase in serum levels of IL-4, IL-10, and IFN-y,
10 but no change in IL-1 or TNF-a following long-term ozone (1 ppm) exposure.
11 • Notably, some studies also found that long-term exposure to ozone (0.8, 1.0 ppm) resulted in no
12 appreciable changes in other inflammatory markers, including TNF-a, IL-1 (Miller et al.. 2016).
13 and total lymphocytes (Gordon et al.. 2013).
14 With respect to markers of oxidative stress, there is limited evidence that long-term exposure to
15 ozone can result in markers of oxidative stress. That is:
16 • In rats, Sethi et al. (2012) and Perepu et al. (2012) reported a decrease in SOD activity (p < 0.05)
17 following long-term ozone (0.8 ppm) exposure. Perepu et al. (2012) also reported an increase in
18 lipid peroxidation.
19 • However, Gordon et al. (2013) reported no appreciable change in HO-1 levels following
20 long-term ozone (0.8 ppm) exposure in rats.
4.2.12 Stroke and Associated Cardiovascular Effects
4.2.12.1 Epidemiologic Studies
21 No studies examining long-term ozone exposure and stroke or other cerebrovascular outcomes
22 were included in the 2013 Ozone ISA. A recent national cohort study conducted in England observed null
23 associations between long-term ozone exposure and both stroke and cerebrovascular disease (Atkinson et
24 al.. 2013). In addition, several recent publications report results from a cross-sectional study conducted in
25 33 Chinese communities, noting positive associations between long-term ozone exposure and stroke
26 (Dong et al.. 2013a). When stratified by obesity status, positive associations were observed between
27 long-term ozone exposure and stroke for adults that were overweight or obese, and null associations for
28 adults with normal weight (Oin et al.. 2015). These studies are characterized in Table 4-45.
September 2019
4-55
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
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 et al.. 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.
4.2.14 Aggregate Cardiovascular Disease
4.2.14.1 Epidemiologic Studies
No studies examining long-term ozone exposure and aggregate endpoints related to
cardiovascular disease 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 (Oin 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.
September 2019
4-56
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
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 limited 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:
• 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 etal.. 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.. 2017; Cakmak et al.. 2016; Crouse et al.. 2015).
• Cohort studies conducted in France (Bcntavcb 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).
September 2019
4-57
DRAFT: Do Not Cite or Quote
-------
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.
Corresponding quantitative results are reported in Supplemental Table A5-C (HERO).
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
1 The evaluation of potential confounding effects of copollutants on the relationship between
2 long-term ozone exposure and cardiovascular effects allows for examination of whether ozone risks are
3 changed in copollutant models. In the 2013 Ozone ISA, Jerrett et al. (2009) reported associations with
4 cardiovascular mortality that were attenuated, changing from positive to negative, after adjustment for
5 PM2 5 concentrations. Recent studies examined the potential for copollutant confounding by evaluating
September 2019
4-58
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
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 PM2 5 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.. 2015V
• 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).
September 2019
4-59
DRAFT: Do Not Cite or Quote
-------
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
September 2019
4-60
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
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 (Oin et al.. 2015V 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
September 2019
4-61
DRAFT: Do Not Cite or Quote
-------
1 associations between different age groups, but do not report consistent evidence that older adults are at
2 increased risk.
3 • In an English cohort, Atkinson et al. (2013) observed no difference in the association between
4 long-term ozone exposure and heart failure for participants aged 40-64 years compared with
5 those aged 65-89 years.
6 • In a cross-sectional study of 33 Chinese communities (Dong et al.. 2013a'). the association
7 between long-term ozone exposure and prevalent prehypertension was stronger among older
8 women (>55 years) compared with younger women (<35 years), whereas the association for
9 increases in blood pressure were stronger among younger adults (<35 years) compared with older
10 adults (>55 years). In an additional cross-sectional analysis of a Chinese population, stronger
11 associations were observed between long-term ozone exposure and hypertensions in both younger
12 (<55 years) and older (>65 years) adults, compared with adults that were between 55 and 64 years
13 old.
4.2.18 Summary and Causality Determination
14 This section evaluates evidence for cardiovascular health effects, with respect to the causality
15 determination for long-term exposures to ozone using the framework described in the Preamble to the
16 ISA (U.S. EPA. 2015). The key evidence, as it relates to the causal framework, is summarized in
17 Table 4-2. A small number of toxicological studies reviewed in the 2013 Ozone ISA provided some
18 evidence for enhanced atherosclerosis and impaired cardiac contraction in isolated perfused rat hearts
19 following long-term ozone exposure ITJ.S. EPA (2013a). see pg 7-40], In addition, an animal
20 toxicological study demonstrated increases in markers associated with inflammation, oxidative stress,
21 thrombosis, and vasoconstriction following long-term exposure ITJ.S. EPA (2013a). see pg 7-40], The
22 limited body of epidemiologic evidence included in the 2013 Ozone ISA included studies of long-term
23 ozone exposure and circulating biomarkers, as well as a study evaluating cardiovascular mortality. Recent
24 epidemiologic evidence remains limited, although several recent studies provide some evidence for
25 changes in measures of blood pressure or increases in hypertension outcomes. Further, the number of
26 studies of cardiovascular mortality has increased, and these studies generally report positive associations.
27 Overall, the limited number of recent studies are consistent with, and in some cases extend, the
28 conclusions in the 2013 Ozone ISA. This evidence is discussed in greater detail below.
29 Overall, the evidence base describing the relationship between long-term ozone exposure and
30 cardiovascular effects remains limited. A couple of recent animal toxicological studies continue to
31 demonstrate impaired cardiac function following long-term ozone exposure. Note that these studies were
32 conducted by the same laboratory and show similar effects to those studies included in the 2013 Ozone
33 ISA (Section 4.2.5.2). In addition, a limited number of recent animal toxicological studies show
34 inconsistent evidence with respect to increases in markers of inflammation, oxidative stress, and a
35 proatherosclerotic environment.
September 2019
4-62
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
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, a number of
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 PMio and NO2
(Section 4.2.16V 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 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. Collectively, the body of evidence for long-term ozone exposure and
cardiovascular effects is suggestive of, but not sufficient to infer, a causal relationship.
September 2019
4-63
DRAFT: Do Not Cite or Quote
-------
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
Section 4.2.4.2
Section 4.2.11
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
Section 4.2.15
14.3-57.5 ppb
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
Section 4.2.3.1
Section 4.2.5.1
Section 4.2.12.1
19.9-24.7 ppb
No evidence from a
limited number of animal
toxicological studies
Changes in blood pressure
Section 4.2.8.2
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.
September 2019
4-64
DRAFT: Do Not Cite or Quote
-------
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 HR
95% CI
IWinauist etal. (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. Mean: 36.3 Correlation
EPA AQS at Tudor Street
stationary monitor; data
missing 1.9% of days
8-h max
Maximum:
111.8
(r):PM2.s: 0.25
Copollutant
models with: NR
ED visits, lag 0-4: 1.05
(1.01, 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
HES
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: -0.096
NO2: -0.3489
SO2: -0.0849;
Other:
PM10 0.0302,
CO -0.2973
Copollutant
models with: NA
Heart failure, lag 0-
0.99 (0.98, 1.01)
September 2019
4-65
DRAFT: Do Not Cite or Quote
-------
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 HR
95% CI
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 in St. Louis
from U.S. EPAAQS
8-h max
Mean: 36.2
Correlation (r):
PM2.5: 0.23;
NO2: 0.37;
SO2: -0.04;
Other: CO -0.01
Copollutant
models with: 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 etal. (2015) n = 84,269
Little Rock, AR, U.S.
Ozone: 2002-2012
Follow-up: 2002-2012
Time-series study
Daily emergency
room visits among
persons 15 yr and
older, 19% 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 with: NR
Hypertensive heart
disease and heart
failure, lag 1: 0.97 (0.91,
1.05)
IWinauist etal. (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 AQS at Tudor Street
stationary monitor; data
missing 1.9% of days
8-h max
Mean: 36.G
Maximum:
111.8
Correlation (r):
PM2.5: 0.25;
Copollutant
models with: NR
ED visits, lag 0-4: 1.05
(1.01, 1.09)
HA, lag 0-4: 1.05 (1.02,
1.09)
September 2019
4-66
DRAFT: Do Not Cite or Quote
-------
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 HR
95% CI
IMiloievic et al. (2014)
England and Wales, U.K.
Ozone: 2003-2009
Follow-up: 2003-2009
HES
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: -0.096;
NO2
SO2
Heart failure, lag 0-
0.99 (0.98, 1.01)
-0.3489;
-0.0849;
Other:
PM10 0.0302,
CO -0.2973
Copollutant
models with: 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 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 in St. Louis
from U.S. EPAAQS
8-h max
Mean: 36.2
Correlation (r):
PM2.5: 0.23;
NO2: 0.37;
SO2: -0.04;
Other: CO -0.01
Copollutant
models with: NR
0-2 day distributed lag:
1.04 (0.99, 1.10)
Copollutant model with
NO2, 0-2 day
distributed lag: 1.02
(0.96, 1.08)
Copollutant model with
PM2.5, lag 1: 0.97 (0.90,
1.05)
Copollutant model with
PM2.5, 0-2 day
distributed lag: 1.02
(0.97, 1.08)
Copollutant model with
CO, 0-2 day distributed
lag: 1.06 (1.00, 1.12)
September 2019
4-67
DRAFT: Do Not Cite or Quote
-------
Table 4-3 (Continued): Epidemiologic studies of short-term exposure to ozone and heart failure.
Copollutant Effect Estimates HR
Study Study Population Exposure Assessment Mean (ppb) Examination 95% CI
tRodoDoulou etal. (2015)
n = 84,269
U.S. AQS data from
Mean: 40
Correlation (r):
Hypertensive heart
Little Rock, AR, U.S.
Daily emergency
stationary monitor in Little
Median: 39
NR
disease and heart
Ozone: 2002-2012
room visits among
Rock
75th: 50
Copollutant
failure, lag 1: 0.97 (0.91,
1.05)
Follow-up: 2002-2012
persons 15 yr and
older, 19% 65 yr
8-h max
models with: NR
Time-series study
and older, 42.5%
male
September 2019
4-68
DRAFT: Do Not Cite or Quote
-------
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 WT, GSTM null)
n = GSTM WT 8, GSTM null seven males, GSTM WT 4,
GSTM null five females
Age: GSTM null: 27.3 ± 4.2 yr, GSTM WT: 25.4 ± 2.8 yr
0.1, 0.2 ppm, 3 h
(alternating 15 min periods
of rest and exercise)
LVDP and 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; WT = wild type.
September 2019
4-69
DRAFT: Do Not Cite or Quote
-------
Table 4-5 Study-specific details from short-term animal toxicological studies of impaired heart function.
Study
Species (Strain), n, Sex,
Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
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 Measures of cardiac function (8-10 h PE)
by 3 h of FA
Mclntosh-Kastrinskv et al. Mice (C57BL/6) 0.245 ppm, 4 h (aged, FA, or ozone) on 3
(2013) n = o males, 14-15/group 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) 0.3 ppm, 4 h LVDP, contractility (24 h PE)
n = 5-8/group males, 0
females
Age:10-12 weeks
September 2019
4-70
DRAFT: Do Not Cite or Quote
-------
Table 4-5 (Continued): Study-specific details from short-term animal toxicological studies of impaired heart
function.
Study
Species (Strain), n, Sex,
Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Ramot etal. (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
September 2019
4-71
DRAFT: Do Not Cite or Quote
-------
Table 4-5 (Continued): Study-specific details from short-term animal toxicological studies of impaired heart
function.
Study
Species (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.
September 2019
4-72
DRAFT: Do Not Cite or Quote
-------
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 HR 95% CI
IFinnbiornsdottir et al. (2013)
Reykjavik, Iceland
Ozone: January 1,
2005-December31, 2009
Follow-up: January 1,
2005-December31, 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:
144.5 |jg/m3
Correlation (r): NO2:
-0.62; Other: PM10
0.13
Copollutant models
with: 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; Other:
CO -0.15,
PMio0.21
Copollutant models
with: NR
0-1 distributed lag: 1.05 (0.96,
1.16)
September 2019
4-73
DRAFT: Do Not Cite or Quote
-------
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 HR 95% CI
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; Other:
PM10 0.0302, CO
-0.2973
Copollutant models
with: 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
Heart Disease
Register, a WHO
MONICA center
n = 2,134
Fatal and nonfatal Ml
cases, aged 35-74 yr,
76.9% male
Modeled hourly concentrations
at census block level using
ADMS-Urban air dispersion
model. Control days selected
using a monthly time-stratified
design
8-h avg
Mean: 32.13
Median: 30.16
75th: 43.15
Maximum:
228.3 |jg/m3
Correlation (r): NO2:
-0.34; Other:
PM-io-0.16,
CO -0.34, benzene
-0.51
Copollutant models
with: 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)
tSarnat et al. (2015)
n = 69,679
Averaged hourly concentrations Mean: 36.2
Correlation (r):
Ischemic heart disease,
St. Louis, MO, U.S.
ED visit records of
in St. Louis from U.S. EPA AQS
PM2.5: 0.23; NO2:
0-2 day distributed lag: 0.99
Ozone: June 1, 2001-May 30,
2003
patients residing in St.
Louis MSA (eight
8-h max
0.37; SO2: -0.04;
Other: CO -0.01
(0.95, 1.04)
Follow-up: June 1, 2001-May
30, 2003
counties each in
Missouri and Illinois)
from 36 out of 43
Copollutant models
with: NR
Time-series study
acute care hospitals
September 2019
4-74
DRAFT: Do Not Cite or Quote
-------
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 HR 95% CI
IWana et al. (2015a)
Calgary, Canada
Ozone: April 1, 1999-March 31,
2010
Follow-up: April 1,
31, 2010
Case-crossover study
1999-March
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
with: NR
Analytical results were not
reported for main effects for
ozone, only statistically
significant results reported
September 2019
4-75
DRAFT: Do Not Cite or Quote
-------
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 HR 95% CI
IWanq et al. (2015b)
Calgary, Edmonton, Canada
Ozone: April 1, 1999-March 31,
2010
Follow-up: April 1,
31, 2010
Case-crossover study
1999-March
n = 12,066
AMI cases aged 20 or
older living in urban
Calgary and
Edmonton
Averaged hourly concentrations
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
Mean: NR
Correlation (r): NR
Copollutant models
with: NR
Calgary-whole, lag 0: 1.00
(0.96, 1.05)
Calgary-whole, lag 1: 0.97
(0.93, 1.01)
Calgary-whole, 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-whole, lag 0: 1.00
(0.96, 1.05)
Edmonton-whole, lag 1: 1.00
(0.95, 1.04)
Edmonton-whole, 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)
September 2019
4-76
DRAFT: Do Not Cite or Quote
-------
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 HR 95% CI
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):
PM2.5: -0.35; Other:
PM-io-0.24
Copollutant models
with: 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
Ml NAP
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
Copollutant models
with: NR
All Ml, lag 0-2: 1.00 (0.99,
1.01)
STEMI, lag 0-2: 0.99 (0.98,
1.01)
nonSTEMI, lag 0-2: 1.00
(0.99, 1.01)
September 2019
4-77
DRAFT: Do Not Cite or Quote
-------
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 HR 95% CI
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;
Other: PM10 -0.287
Copollutant models
with: NR
Not statistically significant,
results in figure
ICollart 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
with: NR
Analytic results displayed in
Figure 4. No associations
using any lag
TVidale etal. (2017)
Como, Italy
Ozone: January
2005-December 2014
Follow-up: January
2005-December 2014
Time-series study
n = 4,110
All residents of Como
with hospital
admission for acute Ml
between January 2005
and December 2014,
mean age 71 yr, 65%
male
Average daily concentrations
from two stationary monitors
24-h avg
Correlation (r): NR
Copollutant models
with: NR
Lag 0: 1.00 (0.99, 1.00)
Lag 1: 0.98 (0.97, 1.01)
IRasche etal. (2018)
Jena, Germany
Ozone: January 1,
2003-December31, 2010
Follow-up: January 1,
2003-December31, 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
Maximum:
117.29 |jg/m3
Correlation (r): NR
Copollutant models
with: NR
Lag 2: 0.29 (0.11, 0.86)
September 2019
4-78
DRAFT: Do Not Cite or Quote
-------
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 HR 95% CI
IHannaet al. (2011)
North Carolina, five cities, U.S.
Ozone: January 1,
1996-December31, 2004
Follow-up: January 1,
1996-December31, 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
Correlation (r): NR
Copollutant models
with: 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 etal. (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 Median: 19.29 Correlation (r): NO2: 1-h avg, lag 1-6 h: 0.99 (0.96,
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
75th: 28.43
-0.58; SO2: -0.14;
Other: CO -0.24
Copollutant models
with: NA
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)
September 2019
4-79
DRAFT: Do Not Cite or Quote
-------
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 HR (95% CI)
lEvans etal. (2016)
Rochester, NY, U.S.
Ozone: 2007-2012
Panel study
n = 362
Treated forSTEMI,
NSTEMI, or unstable
angina
Mean concentrations from Mean: 27.4 Correlation (r): Increased odds of STEMI
NYDEC monitor
1-h max, 12, 24, 48, and
72-h avg
Median: 27
75th: 36.9
Maximum:
104
NR
Copollutant
models with: NR
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)
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, 0.07,120 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.
September 2019
4-80
DRAFT: Do Not Cite or Quote
-------
Table 4-9 Study-specific details from short-term animal toxicological studies of ST-segment depression.
Study
Species (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 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.
September 2019
4-81
DRAFT: Do Not Cite or Quote
-------
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 HR (95% CI)
tZanobetti etal. (2014)
n = 64 Averaged hourly
Mean: 10
Correlation (r):
No change in BAD at 5-day avg
Boston, MA, U.S.
J2D concentrations from local sites
Median: 28
NR
exposure to ozone—qualitative
Ozone: 2006-2009
24-h avg
75th: 33
Copollutant
result
models with: NR
Panel study
Maximum: 47
tLiunqman et al. (2014)
Framingham
Hourly concentrations from
Mean: 23
Correlation (r):
Percentage increase PAT ratio:
Boston, MA, U.S.
Offspring/Third
Boston area monitors were
Maximum: 64
NR
1-day moving avg: -3.43
Ozone: 2003-2008
Generation
averaged to create moving
Copollutant
(-6.33, -0.53)
Panel study
n = 2,369
averages
24-h avg
models with: NR
2-day moving avg: -4.42
(-7.90, -0.93)
Lags examined: 1-7 day
5-day moving avg: -0.32
moving avg
(-5.01, 4.37)
tLanzinqer et al. (2014)
n = 22
Monitor data
Mean: 41
Correlation (r):
Percentage increase FMD
Chapel Hill, NC, U.S.
Subjects with T2D aged
8-h max
Median: 39
NR
Lag 0: -29.2 (-52.6, -5.80)
Ozone: 2004-2005
48-78 yr
75th: 52
Copollutant
Lag 1: -27.0 (-54.0, —0.08)
models with: NR
Panel study
Maximum: 82
IMirowskv et al. (2017)
Chapel Hill, NC, U.S.
Ozone: 2012-2014
Panel study
Lags examined: 0-4, 5-day
avg
Additional endpoints
reported: LAEI, SAEI
CATHGEN
n = 13
Have undergone cardiac
catheterization
Age 53-68
AQS monitor
24-h avg
Mean: 26
Median: 25
75th: 33
Maximum: 63
Correlation (r):
NR
Copollutant
models with:
PM2.5
Percentage increase FMD
Lag 0: -17.14 (-40.82, 15.11)
Lag 1: 4.82 (-27.21, 49.82)
5-day avg: -19.93
(-53.46, 34.39)
Percentage increase BAD
Lag 0: -2.25 (-5.46, 1.07)
Lag 1: -2.04 (-5.25, 1.29)
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.
September 2019
4-82
DRAFT: Do Not Cite or Quote
-------
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, 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
FramDton et al. (2015)
Healthy adults (GSTM WT,
GSTM null)
n = GSTM WT 8, GSTM null
seven males, GSTM WT 4,
GSTM null five females
Age: GSTM null: 27.3 ± 4.2 yr,
GSTM WT: 25.4 ± 2.8 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; WT = wild type.
September 2019 4-83 DRAFT: Do Not Cite or Quote
-------
Table 4-12 Study-specific details from short-term animal toxicological studies of vascular function.
Study
Species (Strain), n, Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Wanaetal. (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 (Fischer)
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.
September 2019
4-84
DRAFT: Do Not Cite or Quote
-------
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 HR
95% CI
lEnsoretal. (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; Other: CO
-0.32
Copollutant models
with: 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
with: 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)
September 2019
4-85
DRAFT: Do Not Cite or Quote
-------
Table 413 (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 HR
95% CI
IWinquist etal. (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
with: 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
week day stratified on month and year
24-h avg
Mean: 31.57
Maximum:
143.4 |jg/m3
Correlation (r): PM2.5:
0.22; NO2: -0.32
Copollutant models
with: 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)
September 2019
4-86
DRAFT: Do Not Cite or Quote
-------
Table 413 (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 HR
95% CI
IStranev etal. (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;
Other: PM10 0.0067,
CO -0.405
Copollutant models
with: NR
Lag 0-1 h: 1.00 (0.99, 1.01)
Lag 0-2 h: 1.00 (0.99, 1.01)
Lag 0-3 h: 1.00 (0.99, 1.01)
Lag 0-4 h: 1.00 (0.99, 1.01)
Lag 0-8 h: 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
IMiloievic 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 Mean: NR
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
Correlation (r): PM2.5: Arrythmias, lag 0-4: 0.99
Median: 30.96
75th: 38.58
-0.096;
NO2: -0.3489;
SO2: -0.0849;
Other: PM10 0.0302,
CO -0.2973
Copollutant models
with: NA
(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;
Other: CO -0.01
Copollutant models
with: NR
0-2 day distributed lag:
1.00 (0.94, 1.06)
September 2019
4-87
DRAFT: Do Not Cite or Quote
-------
Table 413 (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 HR
95% CI
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
with: NR
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)
ISadeet al. (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
with: 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:
114 |jg/m3
Copollutant models
with: NR
September 2019
4-88
DRAFT: Do Not Cite or Quote
-------
Table 4-14 Epidemiologic panel studies of short-term exposure to ozone and electrophysiology, arrhythmia,
and cardiac arrest.
Copollutant
Effect Estimates
Study
Study Population Exposure Assessment
Mean (ppb) Examination
HR (95% CI)
TBartell etal. (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
with: 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)
ILiunqman 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
with: 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
with: 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.
September 2019
4-89
DRAFT: Do Not Cite or Quote
-------
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, 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.
September 2019
4-90
DRAFT: Do Not Cite or Quote
-------
Table 4-16 Study-specific details from short-term animal toxicological studies of electrophysiology, arrhythmia,
cardiac arrest.
Study
Species (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.
September 2019
4-91
DRAFT: Do Not Cite or Quote
-------
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 HR (95%
CI)
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
with: 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 etal. (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
with: NR
Follow-up: 2002-2012
among persons 15 yr
and older, 19% 65 yr
Time-series study
and older, 42.5% male
September 2019
4-92
DRAFT: Do Not Cite or Quote
-------
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 HR (95%
CI)
IVencloviene etal. (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):
Other: PM10 -0.028,
CO -0.298
Copollutant models
with: 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)
September 2019
4-93
DRAFT: Do Not Cite or Quote
-------
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 HR
95% CI
tCakmak et al. (2011)
Canadian Health Measures
Monitor
Mean: 34.1
Correlation (r): NR
Absolute change DBP
Canada
Survey
1-h max
95th: 59.6
Copollutant models
(mm Hg), lag 0: 0.65 (0.06,
Ozone: 2007-2009
Panel study
n = 5,604
with: none
1.23)
Absolute change SBP
(mm Hg), lag 0: 1.17 (0.29,
2.05)
tHoffmann et al. (2012)
n = 70
Monitor
Mean: 25
Correlation (r):
Percentage Increase CMP
Boston, MA, U.S.
T2D;40-85 yr
24-h avg
Median: 24
PM2.5: 0.09;
2-day mean: -0.33 (-2.30,
Ozone: 2006-2010
Panel study
75th: 32
Copollutant models
with: PM2.5
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)
September 2019
4-94
DRAFT: Do Not Cite or Quote
-------
Table 4-18 (Continued): Epidemiologic panel studies of short-term exposure to ozone and blood pressure.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates HR
95% CI
IDales and Cakmak (2016)
National, Canada
Ozone: 2007-2009
Follow-up: 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):
PM2.5: NR;
NO2: NR;
SO2: NR;
Other: NR
Copollutant models
with: NA
Percentage increase SBP
Absence of mood disorder:
-0.52 (-1.18, 0.14)
Presence of mood disorder
4.41(1.91, 6.93)
Percentage Increase DBP
Absence of mood disorder
-0.24(-0.85, 0.36)
Presence of mood disorder
3.55(1.01, 6.08)
tMirowskv et al. (2017)
CATHGEN
AQS Monitor
Mean: 26
Correlation (r): NR
Percentage increase SBP
Chapel Hill, NC, U.S.
n = 13
24-h avg
Median: 25
Copollutant models
Lag 0: 2.46 (-2.14, 7.18)
Ozone: 2012-2014
Have undergone cardiac
75th: 33
with: PM2.5
Lag 1: -0.11 (-3.54, 3.64)
Panel study
catheterization
Maximum: 63
5-day avg: 1.50 (-3.75,
6.96)
Lags reported 0-4, 5 day avg
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
with: NR
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)
BP = blood pressure, DBP = diastolic blood pressure, SBP = systolic blood pressure, T2D = type 2 diabetes.
September 2019
4-95
DRAFT: Do Not Cite or Quote
-------
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 WT eight males, GSTM null
seven males; GSTM WT four females, GSTM null
five females
Age: GSTM null: 27.3 ± 4.2 yr, GSTM WT:
25.4 ± 2.8 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)
DBP 15 min, 4 and 22 h PE
Stieqel etal. (2017)
Healthy adults
n = 11 males, four 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; WT = wild type.
September 2019
4-96
DRAFT: Do Not Cite or Quote
-------
Table 4-20 Study-specific details from short-term animal toxicological studies of blood pressure.
Study
Species (Strain), n, Sex,
Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Wanaetal. (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)
September 2019
4-97
DRAFT: Do Not Cite or Quote
-------
Table 4-20 (Continued): Study-specific details from short-term animal toxicological studies of blood pressure.
Study
Species (Strain), n, Sex,
Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Ramot etal. (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 = Sprague-Dawley; SH = spontaneously hypertensive; SHHF = spontaneously
hypertensive heart failure; WKY = Wistar Kyoto.
September 2019
4-98
DRAFT: Do Not Cite or Quote
-------
Table 4-21 Epidemiologic panel studies of short-term exposure to ozone and heart rate variability (HRV), and
heart rate (HR).
Copollutant
Effect Estimates HR
Study
Study Population
Exposure Assessment
Mean (ppb)
Examination
95% CI
tCakmak et al. (2011)
Canadian Health Measures
Monitor
Mean: 34.1
Correlation (r): NR
Absolute change resting
Canada
Survey
1-h max
95th: 59.6
Copollutant models
heart rate (bpm), lag 0: 0.90
Ozone: 2007-2009
n = 5,604
with: None
(0.18, 1.63)
Panel study
TBartell etal. (2013)
n = 55
Hourly monitor values
Mean: 27.1
Correlation (r): NR
Percentage increase
Los Angeles, CA, U.S.
Elderly nonsmokers
24-h avg
Maximum:
Copollutant models
rMSSD, 24-h: 0.54 (-3.04,
Ozone: 2005-2007
60.7
with: NR
4.13)
Panel study
3-day avg: -1.68 (-7.71,
4.34)
Lags reported: 4, 8, 24 h, or
3-, and 5-day avg
5-day avg: -9.03 (-19.23,
1.16)
Additional endpoints reported:
pNN50
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)
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)
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
with: NR
September 2019
4-99
DRAFT: Do Not Cite or Quote
-------
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 HR
95% CI
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):
PM2.5: NR;
NO2: NR;
SO2: NR;
Other: NR
Copollutant models
with: 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
with: 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)
tCole-Hunter et al. (2018)
TAPAS/EXPOsOM ICS
Monitored values used to
Mean: 22
Correlation (r): NR
Percentage increase HR,
Barcelona, Spain
n = 62
model daily time weighted
Maximum:
Copollutant models
3-days prior: 0.41 (-0.66,
Ozone: 2011-2014
Healthy nonsmokers
based on location
(home/work)
32.9
with: NR
1.49)
Panel study
24-h avg
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.
September 2019
4-100
DRAFT: Do Not Cite or Quote
-------
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 = eight males, nine 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 WT, GSTM null)
n = GSTM WT 8, GSTM null seven males,
GSTM WT 4, GSTM null five females
Age: GSTM null: 27.3 ± 4.2 yr, GSTM WT:
25.4 ± 2.8 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; WT = wild type.
September 2019
4-101
DRAFT: Do Not Cite or Quote
-------
Table 4-23 Study-specific details from short-term animal toxicological studies of heart rate variability (HRV),
heart rate (HR).
Study
Species (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
HRV (before, during, and after
exposure)
HRV: time and frequency
domains (before, during, and
after exposure)
Heart rate (before, during, and
after exposure)
Wanqetal. (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)
September 2019
4-102
DRAFT: Do Not Cite or Quote
-------
Table 4-23 (Continued): Study-specific details from short-term animal toxicological studies of heart rate
variability (HRV), heart rate (HR).
Exposure Details
Study Species (Strain), n, Sex, 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 ^A 'n 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.
September 2019 4-103 DRAFT: Do Not Cite or Quote
-------
Table 4-24 Epidemiologic studies of short-term exposure to ozone and pulmonary vascular disease (PVD),
thrombosis.
Copollutant Effect Estimates HR
Study Study Population Exposure Assessment Mean (ppb) Examination 95% CI
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
with: NR
No associations with monthly
average ozone >37 ppb
September 2019
4-104
DRAFT: Do Not Cite or Quote
-------
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 HR
95% CI
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;
Other: PM10 0.0302,
CO -0.2973
Copollutant models
with: NA
Lag 0-4: 0.99 (0.96, 1.02)
tde Miauel-Diez et al. (2016)
National, Spain
Ozone: January 1,
2000-December31, 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
with: 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.
September 2019
4-105
DRAFT: Do Not Cite or Quote
-------
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 HR (95% CI)
tGreen etal. (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
with: 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 etal. (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
with: 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
with: 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.
September 2019 4-106 DRAFT: Do Not Cite or Quote
-------
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 WT, GSTM null)
n = GSTM WT 8, GSTM null seven males,
GSTM WT 4, GSTM null five females
Age: GSTM null: 27.3 ± 4.2 yr, GSTM WT:
25.4 ± 2.8 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
Kahle et al. (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 0.1, 0.2 ppm, 4 h (alternating 30-min Markers of coagulation in blood
asthma (n = 16) periods of exercise and rest) (before, immediately after and 20
n = 13 males, 13 females h PE)
Age: Asthma: 33.5 ± 8.8 yr, healthy:
30.8 ± 6.9 yr
GSTM = glutathione S-transferase M1; PE = post-exposure; WT = wild type.
September 2019
4-107
DRAFT: Do Not Cite or Quote
-------
Table 4-27 Study-specific details from short-term animal toxicological studies of coagulation.
Species (Strain), n, Sex, Exposure Details
Study Age (Concentration, Duration) Endpoints Examined
Snow et al. (2018) Rats (WKY) 0.8 ppm, 4 h/day for 2 consecutive days (diets Circulating platelets
n = 6-8/group males enriched with coconut, olive, or fish oil for 8 weeks
0 females prior)
Age: -12 weeks
WKY = Wistar Kyoto
September 2019
4-108
DRAFT: Do Not Cite or Quote
-------
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 HR
95% CI
TBind etal. (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
with: NR
Panel study
tGandhietal. (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):
PM2.5: -0.05,
S04:-0.05,
NOx:-0.52
Copollutant models
with: NR
Percentage increase plasma
nitrite
Lag 0: -5.61 (-20.61, 9.47)
Lag 1: -4.91 (-18.33, 8.42)
IGreen etal. (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
with: NR
Percentage increase CRP
1-day : 0.80 (-2.00, 3.60)
TLi etal. (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
with: NR
Qualitative results for
myeloperoxidase and
indexed 8-epi-prostaglandin
F2alpha show no change
September 2019
4-109
DRAFT: Do Not Cite or Quote
-------
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 HR
95% CI
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
with: 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)
ILi et al. (2017) Framingham
Boston, MA, U.S. Offspring Cohort
Ozone: 2005-2008 n = 3.396
Panel study
Lags reported: 1-7 day moving
avg
Averaged ozone monitors in Mean: 23.7
the area and made moving
averages per lag
24-h avg
Correlation (r): NR Percentage increase TNFR2
Copollutant models 1-day moving avg: 1.69
with: NR (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.
September 2019
4-110
DRAFT: Do Not Cite or Quote
-------
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)
Kahle et al. (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)
Stieqel etal. (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 etal. (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)
September 2019
4-111
DRAFT: Do Not Cite or Quote
-------
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
Stieqel etal. (2017)
Healthy adults
0.3 ppm, 2 h (four 15-min periods of
Markers of systemic
n = 11 males, 4 females
exercise)
inflammation in blood (before,
Age: 23-31 yr
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)
HDL = high-density lipoproteins; PE = post-exposure.
September 2019
4-112
DRAFT: Do Not Cite or Quote
-------
Table 4-30 Study-specific details from short-term animal toxicological studies of systemic inflammation and
oxidative stress.
Study
Species (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)
Wanqetal. (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)
September 2019
4-113
DRAFT: Do Not Cite or Quote
-------
Table 4-30 (Continued): Study-specific details from short-term animal toxicological studies of systemic
inflammation and oxidative stress.
Study
Species (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)
September 2019
4-114
DRAFT: Do Not Cite or Quote
-------
Table 4-30 (Continued): Study-specific details from short-term animal toxicological studies of systemic
inflammation and oxidative stress.
Study
Species (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)
Martinez-CamDos et al. (2012)
Rats
6-10 per group
0.5 ppm 4 h/day for 2 weeks then
exercise 4 h/day for 2 weeks no
exercise
Markers of oxidative stress in
blood (at the end of 2-week
exposure)
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,
mifepristine, or propranolol followed by
mifepristine)
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)
September 2019
4-115
DRAFT: Do Not Cite or Quote
-------
Table 4-30 (Continued): Study-specific details from short-term animal toxicological studies of systemic
inflammation and oxidative stress.
Study
Species (Strain), n, Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Francis et al. (2017)
Mice (C57BL/6J WT and CCR2 null)
n = 0 males, 3-4/group females
0.8 ppm, 3 h
Markers of systemic
inflammation in blood (24-72 h
PE)
Age: 8-11 weeks
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.
September 2019
4-116
DRAFT: Do Not Cite or Quote
-------
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 HR
95% CI
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;
Lag NR: 0.95 (0.68, 1.31)
Copollutant models
with: NA
IBedada etal. (2012)
Manchester and Liverpool, U.K.
Ozone: 2003-2007
Follow-up: 2003-2007
Case-crossover study
NORTHSTAR
n = 335 Manchester
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;
Other: CO -0.54,
PM-io-0.23
Copollutant models
with: 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)
September 2019
4-117
DRAFT: Do Not Cite or Quote
-------
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 HR
95% CI
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
with: 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
Correlation (r): NR
Copollutant models
with: NR
Lag 0
Lag 1
Lag 2
Lag 3
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)
September 2019
4-118
DRAFT: Do Not Cite or Quote
-------
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 HR
95% CI
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):
N02: -0.54;
Other: minimal
temperature 0.67
Copollutant models
with: NR
Recurrent stroke, 8
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)
September 2019
4-119
DRAFT: Do Not Cite or Quote
-------
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 HR
95% CI
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
SO2
-0.59;
-0.02;
Other: CO -0.47
Copollutant models
with: 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)
September 2019
4-120
DRAFT: Do Not Cite or Quote
-------
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 HR
95% CI
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;
Other: PM10 0.0302,
CO -0.2973
Copollutant models
with: 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 etal. (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
with: NR
Follow-up: 2002-2012
15 yr and older,
19% 65 yr and older,
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 week day stratifying on
month and year
8-h max
Mean: NR
Median: 35.7
75th: 46.3
Correlation (r): NR
Copollutant models
with: 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)
September 2019
4-121
DRAFT: Do Not Cite or Quote
-------
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 HR
95% CI
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
with: 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
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
Averaged hourly concentrations
from monitors nearest to
residential postal code centroid.
Control exposure days selected
using time-stratified design
matching on week day stratified
by season
24-h avg
Mean: 15.3
Correlation (r): NR
Copollutant models
with: 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)
September 2019
4-122
DRAFT: Do Not Cite or Quote
-------
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 HR
95% CI
IWina etal. (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
with: NR
Lag 1: recurrent stroke: 0.94
(0.76, 1.14)
Lag 1: severe incident stroke:
1.27 (1.12, 1.41)
IButland etal. (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 week day 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;
Other: NOx -0.72,
PMio-0.33
Copollutant models
with: 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)
September 2019
4-123
DRAFT: Do Not Cite or Quote
-------
Table 4-31 (Continued): Epidemiologic studies of short-term exposure to ozone and cerebrovascular disease.
Copollutant
Effect Estimates HR
Study
Study Population
Exposure Assessment
Mean (ppb) Examination
95% CI
tVidale etal. (2017)
n =4,110
Average daily concentrations
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
with hospital
24-h avg
with: NR
Ischemic stroke, lag 1: 1.00
2005-December 2014
admission for acute Ml
(0.99, 1.01)
Follow-up: January
or ischemic stroke
between January 2005
and December 2014,
2005-December 2014
Time-series study
mean age 71 yr,
65% male
tWina etal. (2017b)
Brain Attack
Daily maximal 8-h concentration
Correlation (r): NR
Severe incident stroke risk,
Nueces County, TX, U.S.
Surveillance in Corpus
from one central monitor in
Copollutant models
lag 1: 1.27 (1.12, 1.41)
Ozone: 2000-2012
Christi register
TCEQ TAMIS
with: NR
Severe incident stroke risk,
Follow-up: 2000-2012
n = 3,035
Cases recorded in
24-h avg
lag 1, with neighborhood
disadvantage: 1.27 (1.12,
Time-series study
registry in Nueces
County, TX, mean age
70 yr, 48.7% male,
53% Mexican
American
1.41)
September 2019
4-124
DRAFT: Do Not Cite or Quote
-------
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 HR
95% CI
IKalantzi etal. (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 Lag 0: 1.02 (1.01, 1.03)
Copollutant models Lag ^ 1-02 (1.01, 1.03)
with: NR
IHunova etal. (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
with: 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 etal. (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 111 8
days
8-h max
Correlation (r):
PM2.5: 0.25;
Copollutant models
with: NR
HA, lag 0-4: 0.99 (0.95, 1.00)
ED, lag 0-4: 1.00 (0.98, 1.02)
September 2019
4-125
DRAFT: Do Not Cite or Quote
-------
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 HR
95% CI
IRodopoulou etal. (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
from three sites in the county,
U.S. EPAAQS data
8-h max
Mean: 43.2
Median: 43
75th: 51
Maximum: 70
Correlation (r):
PM2.5: -0.05;
Other: PM10 0.18
Copollutant models
with: 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: -0.3489;
SO2: -0.0849;
Other: PM10 0.0302,
CO -0.2973
Copollutant models
with: NA
All CVD, lag 0-4: 0.99 (0.99,
1.00)
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. EPAAQS
8-h max
Correlation (r):
PM2.5: 0.23;
NO2: 0.37;
SO2: -0.04;
Other: CO -0.01
Copollutant models
with: NR
Lag 0-2: 0.99 (0.97, 1.02)
IRodopoulou etal. (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
with: NR
15 yr and older,
19% 65 yr and older,
Follow-up: 2002-2012
Time-series study
42.5% male
September 2019
4-126
DRAFT: Do Not Cite or Quote
-------
Table 4-32 (Continued): Epidemiologic studies of short-term exposure to ozone and aggregate cardiovascular
disease.
Study
Study Population
Exposure Assessment
Copollutant
Mean (ppb) Examination
Effect Estimates HR
95% CI
tVidale etal. (2017)
n =4,110
Average daily concentrations
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
with: NR
Time-series study
mean age 71 yr, 65%
male
IHunova etal. (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
Correlation (r):
Other: PM10 lag-1
0.457
Copollutant models
with: NR
0.98 (0.95, 1.01)
IChoietal. (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
with: NR
Lag 0-4: 1.07 (1.03, 1.11)
September 2019
4-127
DRAFT: Do Not Cite or Quote
-------
Table 4-33 Epidemiologic studies of short-term exposure to ozone and cardiovascular mortality.
Study
Study Population
Exposure Assessment
Mean (|jg/m3
Copollutant
Examination
Effect Estimates
HR (95% CI)
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
with: 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):
PM2.5: 0.43;
NO2: 0.18;
SO2: -0.19;
Other: CO: -0.35
Copollutant models
with: 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) All ages
10 Canadian cities
1981-1999
Time-series study
Monitor located downtown or at Mean: 19.3
city airports within 27 km of
downtown in each city
24-h avg
Correlation (r): All-year (lag 0): 4.65 (1.86,
NR 7.43)
Copollutant models Spring (lag 0): 3.16 (0.25,
with: NR 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)
September 2019
4-128
DRAFT: Do Not Cite or Quote
-------
4.3.2 Long-Term Ozone Exposure
1
Table 4-34 Epidemiologic studies of long-term exposure to ozone and ischemic heart disease (IHD).
Copollutant
Effect Estimates HR
Study
Study Population
Exposure Assessment
Mean (|jg/m3)
Examination
95% CI
tAtkinson et al. (2013)
English cohort
Annual average from
Correlation (r):
Ml; 2003-2007 exposure
Nationwide, U.K.
n = 836,557
emission-based model with
PM2.5: -0.43
period; NO2 copollutant:
Ozone: 2002-2007
Age: 40-89 yr
1- x 1-km resolution
Copollutant models
0.71 (0.57, 0.87)
Follow-up: 2003-2007
with: NR
Ml; 2003-2007 exposure
period; PM10 copollutant:
Cohort study
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)
Kim et 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' zip
Median: 18.75
PM2.5: 0.67;
(0.75, 0.88)
codes
NO2: 0.68;
Ozone: NR
Healthy adults
75th: 27.08
SO2: 0.84;
Follow-up: 2007-2013
Maximum:
Other: CO: 0.55;
Cohort study
71.12
PMio-2.5: 0.37
Copollutant models
with: NR
September 2019
4-129
DRAFT: Do Not Cite or Quote
-------
Table 4-35 Epidemiologic studies of long-term exposure to ozone and atherosclerosis.
Study
Study
Population
Exposure Assessment Mean (ppb)
Copollutant
Examination
Effect Estimates HR
95% CI
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;
Other: PM10:
-0.05
Copollutant
models with:
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)
September 2019
4-130
DRAFT: Do Not Cite or Quote
-------
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 HR
95% CI
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;
Other: PM10:
0.21-0.66
Copollutant
models with: 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)
Table 4-36 Study-specific details from animal toxicological studies of atherosclerosis.
Study
Species (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, 6hr/day, 1 day/week for
17 weeks
Potential markers of
atherosclerosis at the end of the
given exposure (28 or 56 days)
Sethi etal. (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.
September 2019
4-131
DRAFT: Do Not Cite or Quote
-------
Table 4-37 Epidemiologic studies of long-term exposure to ozone and heart failure.
Copollutant Effect Estimates HR
Study Study Population Exposure Assessment Mean (|jg/m3) Examination 95% CI
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
Correlation (r):
PM2.5: -0.43
Copollutant models
with: 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)
tKimet al. (2017)
NHIS-NSC
Average from monitors linked to
Mean: 19.93
Correlation (r): HR for CHF: 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;
Ozone: NR
Healthy adults
75th: 27.08
Follow-up: 2007-2013
Maximum:
Other: CO: 0.55;
Cohort study
71.12
PM10-2.5: 0.37
Copollutant models
with: NR
September 2019
4-132
DRAFT: Do Not Cite or Quote
-------
Table 4-38 Study-specific details from animal toxicological studies of impaired heart function.
Study
Species (Strain), n, Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Perepu 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
LVDP (28 and 56 days PE)
Sethi etal. (2012)
Rats (S-D)
n = 6/treatment group males, 0 females
Age: adult
0.8 ppm, 8 h/day for 28 or 56 days
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
Species (Strain), n, Sex, Age
Exposure Details
(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, 17 weeks
0 females
Age: 4 and 20 mo
Markers of endothelial function
blood drawn day after final
exposure
BN = brown Norway.
September 2019
4-133
DRAFT: Do Not Cite or Quote
-------
Table 4-40 Epidemiologic studies of long-term exposure to ozone and blood pressure.
Study
Study Population
Exposure
Assessment
Mean (|jg/m3
Copollutant
Examination
Effect Estimates HR (95% CI)
IDona et al. (2013b) 33 Communities
Three northeastern
cities, China
Ozone: 2006-2008
Follow-up:
2009-2010
Cross-sectional
study
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
models with
NR
Correlation Absolute increase in
(r). NR Absolute increase in
Copollutant 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)
September 2019
4-134
DRAFT: Do Not Cite or Quote
-------
Table 4-40 (Continued): Epidemiologic studies of long-term exposure to ozone and blood pressure.
Study
Study Population
Exposure
Assessment
Mean (|jg/m3
Copollutant
Examination
Effect Estimates HR (95% CI)
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
Median: 25
Maximum:
35.5
Correlation
(r): NR
Copollutant
models with:
NR
Absolute difference for SBP; normal weight—female: -0.12 (-0.66,
0.44)
Absolute difference for DBP; normal weight—female: -0.13 (-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)
September 2019
4-135
DRAFT: Do Not Cite or Quote
-------
Table 4-40 (Continued): Epidemiologic studies of long-term exposure to ozone and blood pressure.
Study
Study Population
Exposure
Assessment
Mean (|jg/m3
Copollutant
Examination
Effect Estimates HR (95% CI)
IDonq 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 with:
NR
Absolute increase in
Absolute Increase in
Absolute increase in
Absolute increase in
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;
SBP; girls: 0.20 (0.16, 0.24)
SBP; breastfeeding only: 0.22 (0.18, 0.25)
SBP; all: 0.22 (0.19, 0.25)
SBP; no breastfeeding: 0.22 (0.16, 0.28)
DBP; breastfeeding only: 0.23 (0.21, 0.27)
SBP; boys: 0.23 (0.19, 0.28)
DBP; all: 0.25 (0.22, 0.27)
DBP; girls: 0.25 (0.22, 0.29)
DBP; boys: 0.25 (0.21, 0.28)
DBP; no breastfeeding: 0.27 (0.22, 0.32)
breastfeeding only: 1.04 (1.03, 1.05)
boys: 1.05 (1.04, 1.06)
girls: 1.05 (1.04, 1.06)
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
average of AQS
monitors (n = 4)
averaged over
trimesters
24-h avg
Median: 23.4
75th: 29.2
Correlation
(r): PM2.5:
-0.13; NO2:
-0.69; Other:
BC: -0.35;
NOx: -0.92
Copollutant
models with:
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)
September 2019
4-136
DRAFT: Do Not Cite or Quote
-------
Table 4-40 (Continued): Epidemiologic studies of long-term exposure to ozone and blood pressure.
Study
Study Population
Exposure
Assessment
Mean (|jg/m3
Copollutant
Examination
Effect Estimates HR (95% CI)
IDonq 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
Absolute
increase
in
SBP
normal weight children: 0.13 (0.10, 0.17)
(r): NO2:
0.33;
SO2: 0.6;
Absolute
increase
in
SBP
normal weight girls: 0.13 (0.09, 0.18)
Absolute
increase
in
SBP
normal weight boys: 0.14 (0.08, 0.19)
Other: PM10:
Absolute
increase
in
SBP
overweight boys: 0.14 (0.03, 0.25)
0.85
Absolute
increase
in
DBP
normal weight boys: 0.17 (0.13, 0.22)
Copollutant
Absolute
increase
in
SBP
overweight children: 0.17 (0.09, 0.25)
models with:
NR
Absolute
increase
in
DBP
normal weight children: 0.19 (0.16, 0.22)
Absolute
increase
in
DBP
normal weight girls: 0.19 (0.16, 0.23)
Absolute
increase
in
SBP
overweight girls: 0.20 (0.08, 0.32)
Absolute
increase
in
DBP
overweight girls: 0.24 (0.13, 0.35)
Absolute
increase
in
DBP
overweight children: 0.25 (0.18, 0.33)
Absolute
increase
in
SBP
obese boys: 0.25 (0.13, 0.36)
Absolute
increase
in
SBP
obese children: 0.25 (0.16, 0.34)
Absolute
increase
in
DBP
obese boys: 0.26 (0.17, 0.35)
Absolute
increase
in
SBP
obese girls: 0.26 (0.10, 0.42)
Absolute
increase
in
DBP
obese children: 0.27 (0.20, 0.35)
Absolute
increase
in
DBP
overweight boys: 0.29 (0.19, 0.38)
Absolute
increase
in
DBP
obese girls: 0.30 (0.17, 0.44)
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 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)
September 2019
4-137
DRAFT: Do Not Cite or Quote
-------
Table 4-40 (Continued): Epidemiologic studies of long-term exposure to ozone and blood pressure.
Study
Study Population
Exposure
Assessment
Mean (|jg/m3
Copollutant
Examination
Effect Estimates HR (95% CI)
TLiu etal. (2016)
Taipei, Taiwan
Ozone: 2005-2012
Follow-up:
2005-2012
Cohort study
n = 3,762
Age: 20-80 yr
Annual average Mean: 27
of nearest Maximum:
monitor 28 7
Correlation
(r): NR
Copollutant
models with:
NR
Absolute
difference
for
DBP
AHI 0-4: -0.10 (-0.85, 0.65)
Absolute
difference
for
SBP
AH I 5-29: -1.20 (-2.12, -0.27)
Absolute
difference
for
SBP
AHI 30+: -1.54 (-2.48, -0.61)
Absolute
difference
for
SBP
all: -1.54 (-2.11, -0.98)
Absolute
difference
for
SBP
AHI 0-4: -1.92 (-2.96, -0.87)
Absolute
difference
for
DBP
AHI 30+: 0.19 (-0.46, 0.84)
Absolute
difference
for
DBP
all: 0.27 (-0.12, 0.66)
Absolute
difference
for
DBP
AHI 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;
Other: PM10:
0.21-0.66
Copollutant
models with:
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 with:
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)
September 2019
4-138
DRAFT: Do Not Cite or Quote
-------
Table 4-40 (Continued): Epidemiologic studies of long-term exposure to ozone and blood pressure.
Study
Study Population
Exposure
Assessment
Mean (|jg/m3
Copollutant
Examination
Effect Estimates HR (95% CI)
lYanq 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 stations
Age: 18-74 yr 8-h avg
Mean: 24.7
Maximum:
35.5
Correlation
(r): NR
Copollutant
models with:
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)
September 2019
4-139
DRAFT: Do Not Cite or Quote
-------
Table 4-40 (Continued): Epidemiologic studies of long-term exposure to ozone and blood pressure.
Study
Study Population
Exposure
Assessment
Mean (|jg/m3
Copollutant
Examination
Effect Estimates HR (95% CI)
ICole-Hunter et al.
(2018)
Barcelona, Spain
Ozone: 2011-2014
Follow-up:
2011-2014
Cohort study
n = 57
Healthy adults
Age: 18-60 yr
Annual average
Mean: 22
Correlation
assigned to
Maximum:
(r): PM2.5:
participant
32.9
-0.4; NO2:
address from
-0.21; Other:
closest reference
PM10: -0.56;
station
NOx: -0.37
Copollutant
models with:
PM10
City- or
Mean: 22.95
Correlation
countywide
Maximum:
(r): NR
annual average
42.3
Copollutant
from monitoring
models with:
stations
NR
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
Change in DBP (mm Hg): 22.97 (20.27, 25.66)
Change in SBP (mm Hg): 24.03 (18.88, 29.20)
September 2019
4-140
DRAFT: Do Not Cite or Quote
-------
Table 4-41 Study-specific details from animal toxicological studies of blood pressure.
Study
Exposure Details
Species (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, 0 females 17 weeks
Age: 4 and 20 mo
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
Exposure Details
Species (Strain), n, Sex, Age (Concentration, Duration)
Endpoints Examined
Gordon et al. (2014)
Rats (BN) 1 ppm, 6 h/day, 2 day/week for
n = 12/treatment group males, 0 females 13 weeks
Age: 4 and 20 mo
Heart rate (rats implanted with
telemeter)
Gordon et al. (2013)
Rats (BN) 0.8 ppm, 6 h/day, 1 day/week for
n = 12/treatment group males, 0 females 17 weeks
Age: 4 and 20 mo
Heart rate (biweekly through
Week 15)
BN = brown Norway; HR = heart rate; HRV =
heart rate variability.
September 2019
4-141
DRAFT: Do Not Cite or Quote
-------
Table 4-43 Study-specific details from animal toxicological studies of coagulation.
Study
Species (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.
September 2019
4-142
DRAFT: Do Not Cite or Quote
-------
Table 4-44 Study-specific details from animal toxicological studies of inflammation.
Study
Species (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 etal. (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.
September 2019
4-143
DRAFT: Do Not Cite or Quote
-------
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 HR
95% CI
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
Correlation (r):
PM2.5: -0.43
Copollutant models
with: NR
Stroke; 2003-2007 exposure
period; PM10 copollutant:
0.94 (0.76, 1.22)
CBVD; 2003-2007 exposure
period: 0.96 (0.90, 1.02)
Stroke; 2002 exposure
period: 1.00 (0.82, 1.30)
Stroke; 2003-2007 exposure
period; NO2 copollutant: 1.00
(0.76, 1.22)
Stroke; 2003-2007 exposure
period: 1.02 (0.81, 1.28)
Stroke; 2003-2007 exposure
period; SO2 copollutant: 1.07
(0.87, 1.38)
IDonq et al. (2013a) 33 Communities
Three northeastern cities, China Chinese Health Study
Ozone: 2006-2008 n = 24,845
Follow-up: 2009-2010 A9e: 18-74 V
Cross-sectional study
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):
NO2: 0.45
;S02: 0.87;
Other: PM10: 0.80
Copollutant models
with: NR
OR for stroke; female: 1.13
(0.92, 1.39)
OR for stroke; all: 1.14 (0.99,
1.30)
OR for stroke; male: 1.14
(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
with: NR
OR compares exposures
>37 ppb to those lower than
37 ppb: 0.83 (0.26, 2.70)
September 2019
4-144
DRAFT: Do Not Cite or Quote
-------
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 HR
95% CI
TQin etal. (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
with: 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)
IKimet al. (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;
Other: CO: 0.55;
PM10-2.5: 0.37
Copollutant models
with: NR
0.73
HR for ischemic stroke
(0.68, 0.77)
HR for stroke: 0.73 (0.69,
0.76)
HR for hemorrhagic stroke:
0.74 (0.67, 0.81)
September 2019
4-145
DRAFT: Do Not Cite or Quote
-------
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 HR
95% CI
IDonq et al. (2013a)
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):
NO2: 0.45;
SO2: 0.87;
Other: PM10: 0.80
Copollutant models
with: 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
with: NR
OR for CVDs; BMI
<25 kg/m2—female: 0.71
(0.42, 1.22)
OR for CVDs; normal weight:
1.07 (0.88, 1.31)
OR for CVDs; overweight:
1.07 (0.87, 1.31)
OR for CVDs; BMI
<25 kg/m2—all: 1.09 (0.89,
1.33)
OR for CVDs; BMI
>25 kg/m2—male: 1.15 (0.94,
1.41)
OR for CVDs; BMI
>25 kg/m2—all: 1.16 (0.96,
1.39)
OR for CVDs; BMI
<25 kg/m2—male: 1.17 (0.94,
1.45)
OR for CVDs; BMI
>25 kg/m2—female: 1.23
(0.82, 1.86)
OR for CVDs; obese: 1.50
(1.02, 2.21)
September 2019
4-146
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
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.
September 2019
DRAFT: Do Not Cite or Quote
4-147
-------
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.
September 2019
4-148
DRAFT: Do Not Cite or Quote
-------
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 forage, 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.
September 2019
4-149
DRAFT: Do Not Cite or Quote
-------
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 oron-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.
September 2019
4-150
DRAFT: Do Not Cite or Quote
-------
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.
September 2019
4-151
DRAFT: Do Not Cite or Quote
-------
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.
September 2019
4-152
DRAFT: Do Not Cite or Quote
-------
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 (2008).
bMurqia et al. (2014); Weakley et al. (2013); Yang et al. (2011); Heckbert et al. (2004); Barret al. (2002); Muhaiarine
etal. (1997); Toren et al. (1993).
cBurnev et al. (1989).
September 2019
4-153
DRAFT: Do Not Cite or Quote
-------
4.4 References
Argacha. JF: Collart. P: Wauters. A: Kavaert. P: Lochv. S: Schoors. D: Sonck. J: de Vos. T: Forton. M: Brasseur.
Q: 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: Daltoa 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: Andersoa 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: el00307. http://dx.doi.org/10.1371/iournal.pone.0100307.
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-ll-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: 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.
Bhaskaraa 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: Steinmaa RC: Rolnitzkv. LM: Kleiger. RE: Rottmaa JN. (1992). Frequency domain
measures of heart period variability and mortality after myocardial infarction. Circulation 85: 164-171.
Biller. H: Holz. O: Windt. H: Koch. W: Mtiller. 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/fphvs.2013.00026.
Bind. MA: Baccarelli. A: Zanobetti. A: Tarantini. L: Suh. H: Vokonas. 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.lQ97/EDE.0b013e31824523f0.
September 2019
4-154
DRAFT: Do Not Cite or Quote
-------
Bossoa JA: Blomberg. A: Stenfors. N: Helledav. R: Kelly. FJ: Behndig. AF: Mudwav. I. (2013). Peripheral
blood neutrophilia as abiomarker of ozone-induced pulmonary inflammation. PLoS ONE 8: e81816.
http://dx.doi.org/10.1371/iournal.pone.0081816.
Bretoa C"V: Wang. X: Mack. WJ: Berhane. K: Lopez. M: Islam TS: Feng. M: Lurmann. F: McConnell. R:
Hodis. HN: Ktinzli. 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.l 161/CIRCULATIONAHA. 112.096164.
Bretoa CV: Yao. J: Millsteia 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: Silvermaa F: Gold. PR: Wellenius. G: Mittleman. MA: Raiagopalaa 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/hvpertensionaha.109.130237.
Burnev. PG: Laitinen. LA: Perdrizet. S: Huckauf. H: Tattersfield. AE: China S: Poissoa 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: Crichtoa 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-2016-208025.
Butland. BK: Atkinson. RW: Miloievic. A: Heal. MR: Dohertv. RM: Armstrong. BG: Mackenzie. IA: Vieno. M:
Lia C: Wilkinsoa 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/10.1136/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.
Cakmak. S: Hebbern. C: Pinault. L: Lavigne. E: Vanos. J: Crouse. PL: Tiepkema. M. (2017). Associations
between long-term PM2.5 and ozone exposure and mortality in the Canadian Census Health and
Environment Cohort (CANCHEC), by spatial synoptic classification zone. Environ Int 111: 200-211.
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: Atkinsoa RW: Kent. AJ: van Staa. T: Cook. PG: Andersoa 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/10.1164/rccm.201210-1758QC.
September 2019
4-155
PRAFT: Po Not Cite or Quote
-------
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.
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.
Chea L: Villeneuve. PJ: Rowe. BH: Liu. L: Stieb. DM. (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: Johantgea 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. DG: Pompilius. M: Ballinger. CA: White. RC: Krzvwanski. DM:
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/10.1164/rccm.200611-1627QC.
Chuang. KJ: Yaa 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/10.1136/oem.2009.0527Q4.
Claevs. MJ: Coenea S: Colpaert. C: Bilcke. J: Beutels. P: Wouters. K: Legrand. V: Van Damme. 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: Donaire-Gonzalez. D: Kubesch. N: Carrasco-Turigas. G: Matt. F: Foraster. M:
Martinez. T: Ambros. A: Cirach. M: Martinez. D: 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.ll.024.
Collart. P: Dramaix. M: Leveaue. 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: el 1.
http://dx.doi.org/10.4081/ni.2012.ell.
Crouse. PL: Peters. PA: Hvstad. P: Brook. JR: van Donkelaar. A: Martia 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.
September 2019
4-156
DRAFT: Do Not Cite or Quote
-------
de Miguel-Piez. 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: Abramsoa 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/10.1097/EDE.Qb013e3181e093db.
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. 112.094359.
Dong. G: Oian. Z: Wang. J: Chea W: Ma. W: Trevathaa E: Xaverius. PK: DeClue. R: Wiese. A: Langstoa M:
Liu. MM: Wang. D: Rea 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: Oiaa ZM: Trevathaa E: Zeng. XW: Vaugha MG: Wang. J: Zhao. Y: Liu. YO: Rea WH: Oia 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: Oiaa ZM: Xaverius. PK: Trevathaa E: Maalouf. S: Parker. J: Yang. L: Liu. MM: Wang. D: Rea
WH: Ma. W: Wang. J: Zelicoff. A: Fu. O: 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.l 161/HYPERTENSIONAHA.l 11.00003.
Dong. GH: Wang. J: Zeng. XW: Chea L: Oia XD: Zhou. Y: Li. M: Yang. M: Zhao. Y: Rea 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: Raua 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.l 161/CIRCULATIONAHA. 113.000027.
Evans. KA: Hopke. PK: Utell. MJ: Kane. C: Thurston. SW: Ling. FS: Chalupa. D: Rich. DO. (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: Silvermaa 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. P: Terrell. P: 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. P: 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: Rafnssoa 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/10.1186/1476-069X-12-38.
September 2019
4-157
PRAFT: Po Not Cite or Quote
-------
Framptoa 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/10.3109/08958378.2Q14.996272.
Francis. M: Groves. AM: Sua R: Cervelli. JA: Choi. H: Laskia JD: Laskia 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.lQ93/toxsci/kfw226.
Gandhi. SK: Rich. DO: 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/10.3109/08958378.2014.913216.
Gong. H. Jr: Wong. R: Sarma. RJ: Lina WS: Sullivan. ED: Shamoo. DA: Andersoa 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: Avdia 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: Broadwia 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: Greea 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.
Halonea JI: Lanki. T: Tiittanea 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.
Halvorsea 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/10.1186/1476-069X-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: McTiernaa A: Gaziano. JM: Frishmaa 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: Dye. 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.
Henrotia 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/10.1136/oem.2006.029306.
September 2019
4-158
DRAFT: Do Not Cite or Quote
-------
Henrotia 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.2010.20Q337.
Hoffmann. B: Luttmann-Gibson. H: Cohea A: Zanobetti. A: de Souza. C: Foley. C: Suh. HH: Coull. BA:
Schwartz. J: Mittlemaa M: Stone. P: Hortoa 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. 1103647.
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-Q751-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-
Jerrett. M: Burnett. RT: Beckermaa 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/10.1164/rccm.2013Q3-
0609QC.
Jerrett. M: Burnett. RT: Pope. CA. Ill: Ito. K: Thurston. G: Krewski. D: Shi. Y: Calle. E: Thua M. (2009).
Long-term ozone exposure and mortality. N Engl J Med 360: 1085-1095.
http://dx.doi.org/10.1056/NEJMoa08Q3894.
Kahle. JJ: Neas. LM: Devlin. RB: Case. MW: Schmitt. MT: Maddea 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: Duauenne. 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: Briancoa S: Petermana R: Ouilliot. 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. H.T: Kim. H: Heo. J: Yi. SM: Kim. K: Youa 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. 117.007170.
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: Masoa 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: Tillmana 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.
September 2019
4-159
DRAFT: Do Not Cite or Quote
-------
Kotecha. D: New. G: Flather. MP: Ecclestoa 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/10.1136/heartinl-
2011-300033.
Kumarathasaa P: Blais. E: Saravanamuthu. A: Bielecki. A: Mukheriee. B: Biarnasoa 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: Diaz-Sanchez. D: Hinderliter. A: Peters. A: Devlia 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.
Li. W: Dorans. KS: Wilker. EH: Rice. MB: Liungmaa PL: Schwartz. JD: 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: Dorans. KS: Rice. MB: Schwartz. J: Coull. BA: Koutrakis. P: Gold. PR: Keanev. JF: Lia H:
Vasan. RS: Benjamin. EJ: Mittlemaa 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. D: 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.
Liungman. PL: Wilker. EH: Rice. MB: Schwartz. J: Gold. PR: Koutrakis. P: Vita. JA: Mitchell. GF: Vasaa RS:
Beniamia EJ: Mittlemaa 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.
M. A: H. W: A. D: J. F: S. D: W. C: K. P: JR. B. (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.
Maheswaran. R: Pearsoa 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:
e0158556. 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. ScientificWorldJournal 2012: 1-5.
http://dx.doi.org/10.1100/2012/135921.
Mcintosh-Kastrinskv. R: Diaz-Sanchez. D: Sextoa KG: Jania. CM: Zavala. J: Tillev. SL: Jaspers. I: Gilmour.
MI: Devlia 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.
September 2019
4-160
DRAFT: Do Not Cite or Quote
-------
Mechtouff. L: Canoui-Poitrine. F: Schott. AM: Nighoghossiaa 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. Ill 1/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 inpatients with implantable defibrillators. Epidemiology 18: 585-592.
http://dx.doi.org/10.1097/EDE.0b013e318124fF0e.
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: Wilkinsoa P: Armstrong. B: Bhaskaraa 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/10.1136/heartinl-2Q13-
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/sl2940-017-0335-0.
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/SQ895-
4356(97)00019-X.
Murgia. N: Brisman. J: Claessoa A: Muzi. G: Olin. AC: Torea 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.
Notarius. CF: Butler. GC: Ando. S: Pollard. MJ: Senn. BL: Floras. JS. (1999). Pissociation 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. P: Balzi. P: Pepe. P: Chini. M: Scala. P: 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: Robertsoa 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: Kleia M: Flanders. WP: 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. P: Sethi. R. (2010). Enhanced death signaling in ozone-exposed ischemic -
reperfused hearts. Mol Cell Biochem 336: 55-64. http://dx.doi.org/10.1007/sll010-009-Q265-4.
Perepu. RSP: Postal. PE: 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/sllQ10-011-1073-l.
Pilz. V: Wolf. K: Breitner. S: Rtickerl. 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.
September 2019
4-161
PRAFT: Po Not Cite or Quote
-------
Poloniecki. JD: Atkinsoa RW: Ponce de Leoa 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: Guernioa 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.
Oia XD: Oian. Z: Vaughn. MG: Trevathan. E: Emo. B: Paul. G: Rea 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 line particulate matter and ozone onHDL 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.
Raza. A: Bellander. T: Bero-Bedada. G: Dahlquist. M: Hollenberg. J: Jonsson. M: Lind. T: Rosenqvist. M:
Svensson. L: Liungmaa 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. DO: Balmes. JR: Framptoa 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.
Robertsoa 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: 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.
Rosenthal. FS: Kuisma. M: Lanki. T: Husseia 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: Campea 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.0800437.
September 2019
4-162
DRAFT: Do Not Cite or Quote
-------
Sacks. JD: Ito. K: Wilsoa 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-Q295-2.
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.Q27292.
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.
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/sll010-012-1363-2.
Silvermaa RA: Ito. K: Freese. J: Kaufmaa 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/kwa217.
Sivagangabalaa G: Spears. D: Masse. S: Urch. B: Brook. RD: Silvermaa 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: Boa 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. L: Fina 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/10.1136/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/10.1136/bmiopen-2013-00406Q.
September 2019
4-163
DRAFT: Do Not Cite or Quote
-------
Tallon. LA: Maniourides. J: Pun. VC: Mittlemaa 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/10.1186/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: a016295. http://dx.doi.org/10.1101/cshperspect.a016295.
Tankerslev. CG: Georgakopoulos. 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/10.1093/toxsci/kfs273.
Tankerslev. CG: Peng. RD: Bedga. D: Gabrielsoa K: Champioa 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: Silvermaa 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: Kumarathasaa 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/10.1164/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-95/001F). 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] (pp. 1-166). (EPA/630/P-03/001F). 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.
September 2019
4-164
DRAFT: Do Not Cite or Quote
-------
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://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/b5d8alce9b07293485257375007Q12b7/eele280e77586de
985257b65005d37e7!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.
van Rossem. L: Rifas-Shiman. SL: Mellv. SJ: Kloog. I: Luttmann-Gibsoa H: Zanobetti. A: Coull. BA:
Schwartz. JD: Mittleman. MA: Okea E: Gillmaa 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: Chea 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: Gotzsche. 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.
vonKlot. S: Peters. A: Aalto. P: Bellander. T: Berglind. N: D'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: Naa B: Morishita. M: Mukheriee. B: Dvonch. JT: Spino. C: Fink. GD:
Raiagopalan. S: Sua 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.
September 2019
4-165
DRAFT: Do Not Cite or Quote
-------
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/10.1136/bmiopen-2015-0Q9169.
Weakley. J: Webber. MP: Ye. F: Zeig-Owens. R: Cohea 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.
Wing. JJ: Adar. SD: Sanchez. BN: Morgenstera 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.ll.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/STROKE AH A. 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-ll-7Q.
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/10.1159/000345138.
Yang. BY: Oiaa ZM: Vaughn. MG: Nelson. EJ: Pharmage. SC: Heinrich. J: Lin. S: Lawrence. WR: Ma. H:
Chea 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.
Yang. CL: To. T: Fotv. RG: Stieb. PM: Pell. SP. (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-ll-52.
Ying. Z: Allea K: Zhong. J: Chea M: Williams. KM: Wagner. JG: Lewandowski. R: Sun. O: Raiagopalaa 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. P: Eagan-Bengstoa E: Gates. KA: Hartley. LH: Suh.
H: Gold. PR. (2004). Ambient pollution and blood pressure in cardiac rehabilitation patients. Circulation
110: 2184-2189. http://dx.doi.org/10.1161/01.cir.000Q143831.33243.d8.
Zanobetti. A: Luttmann-Gibson. H: Hortoa ES: Cohea A: Coull. BA: Hoffmann. B: Schwartz. JP: Mittlemaa
MA: Li. Y: Stone. PH: de Souza. C: Lamparello. B: Koutrakis. P: Gold. PR. (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: Oiaa Z: Wang. J: Vaughn. MG: Liu. Y: Rea W: Pong. G. (2013). Poes 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.
September 2019
4-166
PRAFT: Po Not Cite or Quote
-------
Zhong. J: Allen. K: Rao. X: Ying. Z: Braunstein. Z: Kankanala. SR: Xia. C: Wang. X: Bramble. LA: Wagner.
JG: Lewandowski. R: Sun. O: 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/10.1080/08958378.2Q16.1179373.
Zvchowski. KE: Lucas. SN: Sanchez. B: Herbert. G: Campea 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.
September 2019
4-167
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
APPENDIX 5 HEALTH EFFECTS —METABOLIC
EFFECTS
Summary o f Causal Determinations for Short- and l.onx- Term Metabolic
Health Infects
This \ppendi\ cli;ir;iclcri/cs I lie sciciililic c\ idcucc I li:il supports c:ius;ilil\
dclcriiii ikiI ions lor short-;iud loim-lcrui o/ouc c\posurc ;iik.I uicl;iholic effects I lie l> pes of
si i id ics c\ ;ilu;itcd w n In ii ilns \ppcudi\ ;irc consistent \x illi I lie o\ ei~ill scope ol' ilic IS \ ;is
detailed mi llie hel;ice. In ;issessinu ilie o\er;ill c\ ideiicc. llie sirenullis ;ind liniiiiitious of
iiidi\ idii;il studies were e\:ilu;ilcd h;ised dm sciciilil'ic cuiisideiiilinns del:nled in I lie \mie\ I'm-
\nneiidi\ 5 Mure del;nls on I lie c;iiis;iI l'r;iinc\uirk used Id rc;icli lliesc cdiicIiiskiiis ;iiv nicliidcd
in llie Preiinihle in ilic IS \ il S IllJ \. 2
-------
1 Diabetes is characterized by a continuum of hyperglycemia (i.e., elevated glucose level) resulting
2 from defects in insulin signaling, secretion, or both. Several types of diabetes have been classified by the
3 American Diabetes Association (ADA. 2014). Type 1 diabetes (T1D) is caused by (3-cell dysfunction or
4 destruction that leads to insulin deficiency, while type 2 diabetes is characterized by defects in insulin
5 secretion in an insulin resistant environment. Gestational diabetes mellitus (GDM) is generally diagnosed
6 during the second or third trimester of pregnancy.
7 The subsections below provide an evaluation of the most policy-relevant scientific evidence
8 relating short-term ozone exposure to metabolic health effects. These sections focus on studies published
9 since the completion of the 2013 Ozone ISA. There are a limited number of recent epidemiologic studies
10 examining the effects of short-term ozone exposure on glucose tolerance, insulin sensitivity, and diabetes
11 control. In addition, multiple animal toxicological studies evaluate ozone-mediated effects, and these
12 studies indicate that short-term exposure to ozone affects glucose homeostasis and other factors that
13 contribute to metabolic syndrome.
5.1.1 Population, Exposure, Comparison, Outcome, and Study Design
(PECOS) Tool
14 The scope of this section is defined by a scoping tool that generally defines the relevant
15 Population, Exposure, Comparison, Outcome, and Study Design (PECOS). The PECOS tool defines the
16 parameters and provides a framework to help identify the relevant evidence in the literature to inform the
17 ISA. Because the 2013 Ozone ISA did not make a causality determination for short-term ozone exposure
18 and metabolic health effects, the epidemiologic studies evaluated are less limited in scope and not
19 targeted towards specific study locations, as reflected in the PECOS tool. The studies evaluated and
20 subsequently discussed within this section were identified using the following PECOS tool:
21 Experimental Studies:
22 • Population: Study populations of any controlled human exposure or animal toxicological study of
23 mammals at any lifestage
24 • Exposure: Short-term (in the order of minutes to weeks) inhalation exposure to relevant ozone
25 concentrations (i.e., 0.4 ppm or below for humans, 2 ppm or below for other mammals)
26 • Comparison: Human subjects that serve as their own controls with an appropriate washout period
27 or when comparison to a reference population exposed to lower levels is available, or, in
28 toxicological studies of mammals, an appropriate comparison group that is exposed to a negative
29 control (i.e., clean air or filtered air control)
30 • Outcome: Metabolic effects (e.g., diabetes, metabolic syndrome, dyslipidemia, glucose
31 intolerance, insulin resistance, overweight, obesity)
32 • Study Design: Controlled human exposure (e.g., chamber) studies; in vivo acute, subacute, or
33 repeated-dose toxicity studies in mammals, immunotoxicity studies
September 2019
5-2
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
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.
September 2019
5-3
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
Diabetes
Metabolic
Syndrome
Activation of
Neuro
Endocrine
Sympathetic
Adrenal
Medullary
Pathway
Short-
term
Ozone
Exposure
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.
Figure 5-1 Potential biological pathways for metabolic outcomes following
short-term ozone exposure.
Ozone inhalation can contribute to metabolic syndrome or diabetic health outcomes starting with
upstream events that impact the nervous system leading to changes in HPA axis that impact overall
endocrine and energy homeostasis. Ozone inhalation interacts with the central nervous system in the
respiratory tract activating a stress response. Specifically, the pulmonary irritant ozone stimulates
nasopharyngeal and pulmonary nerves and receptors including the trigeminal and vagal nerves, which
induces downstream effects to the autonomic nervous system. The sensory nerves innervate the lungs and
also communicate with brain regions and other areas of the body like the vagus jugular or nodose
ganglions and is sensitive to this ozone irritant response. Sensory nerve activation can affect metabolic
pathways as well as the pulmonary and cardiovascular systems. The developing rat nodose and jugular
ganglions are structurally altered with early-life ozone exposure, with fewer neurons present after ozone
exposure (Zellner et al.. 2011). The communication through the nodose and jugular ganglion is
transmitted to the hypothalamic paraventricular nucleus [PVN; Gackiere et al. (2011)1 and the brainstem's
September 2019
5-4
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
nucleus tractus solitarius (NTS) where ozone can have an effect as seen with increased c-fos staining, a
marker for neuronal activation Gackiere et al. (201IV Also, the paratrigeminal nucleus (Pa5) can
potentially signal to the hypothalamus via norepinephrine (NE) and second-order neurons or via
glucagon-like peptide-1, or it can communicate back to the heart or airways. These studies show that
ozone exposure induces neuronal communication from the lung to the brain affecting areas including the
brainstem and the hypothalamus. Hypothalamic effects of ozone have been demonstrated in multiple
studies (Mumaw et al.. 2016; Dorado-Martinez et al.. 2001). The hypothalamus is the control center for
stress activation in the brain and it activates downstream targets including the pituitary and the adrenals
contributing further to the stress axis activation through the sympathetic adrenomedullary (SAM)
pathway. For example, ozone induces neuro-inflammatory activation of brain microglia 4 hours after
exposure and out to 24 hours post-exposure. These changes are mediated by soluble factors in the blood
(TNF-a, H202) as seen with microglial activation during ex vivo testing of the bioactivity of serum from
ozone-exposed animals (Mumaw et al.. 2016). Thus, ozone exposure along the lung-brain axis involves
initial neuronal irritant activation in the lungs that then activates multiple downstream pathways in the
brain that contribute directly to activation of the neuroendocrine system starting with the hypothalamus,
which is detailed further below as it progresses to the pituitary and adrenals and other stress pathways that
contribute to perturbed energy homeostasis.
Ozone exposure also induces activation of the neuro-endocrine hypothalamic pituitary adrenal
(HPA) axis stimulating the sympathetic adrenal medullary (SAM) pathways. The hypothalamus, pituitary,
and adrenals respectively release CRH, ACTH and Cortisol forming the neuro-endocrine system that
controls and mediates reactions to stress and regulates body systems accordingly. There is direct evidence
of activation of the neuro-endocrine pathways in the brain, lung, and metabolic organs mediated through
the release of stress hormones [ACTH, norepinephrine, Cortisol, and corticosterone; Miller et al. (2016a);
Miller et al. (2015); Bass et al. (2013); Thomson et al. (2013)1 into the circulation of rats with ozone
exposure. The upstream mediators like adrenocorticotropin hormone (ACTH) from the pituitary
(Thomson et al.. 2013) are released into the circulation of rats (Miller et al.. 2016a) and contribute to a
multiorgan stress response upon ozone exposure. Diabetes and metabolic syndrome are disorders of the
autonomic nervous system. Ozone exposure increases stress hormones and multiple downstream
metabolic effects including glucose intolerance, fasting hyperglycemia, and hepatic gluconeogenesis
(Miller etal.. 2016b; Zhong et al.. 2016; Miller et al.. 2015; Bass et al.. 2013). With ozone exposure,
changes in the biomarkers of glucose intolerance and insulin resistance, including HOMA-IR (Li et al..
2017; Miller etal.. 2016a; Kim and Hong. 2012) or HbAlc (Chuang et al.. 2011). as well as increased
ketone-body formation (Miller et al.. 2016a). have been noted in humans. Impaired insulin signaling is a
pathophysiological effect leading to clinical outcomes such as insulin resistance, increased blood glucose,
and increased blood lipids. Specifically, insulin stimulates sensitive tissues to take up glucose, lipids, and
amino acids. In muscle, insulin stimulates glucose oxidation or storage as glycogen and protein synthesis;
in liver, insulin stimulates glycogen synthesis; and in adipose tissue, insulin stimulates lipid synthesis and
storage. During a fast (overnight) plasma glucose and insulin levels are low; glucagon levels rise, and
lipids are mobilized from adipose tissue into the circulation; glycogenolysis and gluconeogenesis increase
September 2019
5-5
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
in the liver; and striated muscle metabolizes lipids and degrades proteins into amino acids (Boron and
Boulpaep. 2017). When individuals do not respond properly to glucose and insulin levels (as in T2D),
body fuels (glucose, lipid, and amino acid) are mobilized into the blood, putting a burden on liver, kidney,
and vascular function.
Other biomarkers of ozone-dependent altered energy homeostasis are elevated triglycerides in
animals (Miller et al.. 2016c; Bass et al.. 2013) and humans (Chuang et al.. 2011). increased levels of
circulating free fatty acids in both humans and animal toxicological models (Miller etal.. 2016a; Miller et
al.. 2015). and altered cholesterol levels (Thomson et al.. 2013). Skeletal muscle insulin resistance
develops with ozone exposure (Vella etal.. 2014).
Further verification of the importance of the HPA axis in ozone-induced metabolic perturbations
(denoted by the solid line in the Figure 5-1) comes from the attenuation or amelioration of ozone-induced
metabolic effects (glucose intolerance, hyperglycemia, elevated stress hormones epinephrine and Cortisol)
after surgical removal of the adrenal glands (Miller et al.. 2016c; Miller et al.. 2015). As mentioned
earlier, under normal physiological conditions, the hypothalamus, pituitary and adrenals work together to
respond to a potential stressor with Cortisol or corticosterone produced by the adrenal cortex.
Administration of glucocorticoid receptor antagonists reduces ozone-dependent inflammation (Henriquez
etal.. 2017b).
In addition to effects mediated by the HPA axis, there are also immediate changes to baseline
metabolic rate in animals after ozone exposure as well as changes to the thyroid and the pituitary. Adult
male rats immediately become hypothermic with an associated bradycardia during exposure (Mautz and
Bufalino. 1989). Once exposure ceases, there is a delayed daytime hyperthermia that manifests a couple
of days after exposure stops (Gordon et al.. 2014). Baseline metabolism and thermoregulation can be
influenced by thyroid function, and people with thyroid disease are at increased risk of developing type 2
diabetes (Chaker et al.. 2016). Under normal physiological conditions, the thyroid regulates metabolism
and thermoregulation; thyroid hormone status is associated with body weight and energy expenditure
(Chaker et al.. 2016). Hyperthyroidism or excess thyroid hormone production causes a hypermetabolic
state with increased resting energy expenditure, decreased body weight, reduced cholesterol levels,
increased lipolysis, and increased gluconeogenesis. Alternatively, hypothyroidism or decreased thyroid
hormone levels is associated with decreased metabolism and reduced resting energy expenditure,
increased body weight, elevated serum cholesterol, decreased lipolysis, and decreased gluconeogenesis.
Changes in thyroid function are seen after acute ozone exposure (1 hour, 1 ppm, adult male rodents):
circulating serum TSH levels significantly decreased, thyroid hormone (T3 and T4) levels significantly
decreased, circulating protein-bound iodine concentrations significantly decreased, and thyroid weight
went down. Circulating prolactin was significantly increased. Pituitary TSH and prolactin content were
considerably increased, but only TSH was statistically significantly increased in the pituitary,
demonstrating perturbation of the pituitary-thyroid axis following ozone exposure (Clemons and Garcia.
1980) and responsive upregulation of pituitary TSH in response to the drop in serum thyroid hormone
September 2019
5-6
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
levels. The pituitary-thyroid axis may be depressed with acute ozone exposure by decreasing
hypothalamic stimulation by thyrotropin releasing hormone while at the same time removing the
hypothalamic catecholamine inhibition of prolactin release. In an adjacent organ, histological analysis of
the parathyroid gland is consistent with hyperactivity of the parathyroid gland after ozone exposure in
rabbits [4-8 hours exposure to 0.75 ppm ozone; Atwal and Wilson (1974)1.
In summary, short-term ozone exposure has extrapulmonary effects that contribute to metabolic
disturbances, including hypothermia, decreased metabolic rate, increased corticosterone, and
hyperglycemia. These pathways may be initially stimulated by irritant receptors in the pulmonary tract.
The thyroid system is also affected with decreased circulating TSH, T4, and T3. Cytokines, including
IL-33 and IL-17a, contribute to development of metabolic syndrome in animal models of obesity. The
entire cascade begins when pulmonary signaling at irritant receptors and pulmonary nerves (trigeminal
and vagus nerve) is activated by ozone exposure. The HPA axis is then stimulated and systemic
inflammation and oxidative stress ensues. After these initial events, downstream events are activated,
including microglial activation, hypothalamic changes at the level of the PVN and brainstem changes in
the NTS, decreased TSH and thyroid hormones, and changes to core body temperature. Serum
triglycerides increase with ozone exposure. Serum cholesterol can also be affected by ozone exposure, but
varies by model; free fatty acids increase in serum with ozone exposure. ACTH is elevated with ozone
exposure as is its downstream corticosterone or Cortisol. The sympathetic activation also includes
increased levels of norepinephrine. Hyperglycemia, glucose intolerance, and hepatic gluconeogenesis
follow. Recent epidemiologic studies provide evidence for impaired glucose regulation and altered
HbAlc, and human clinical studies show increased ketone body formation in participants exposed to
ozone. All of these upstream factors of autonomic activation and homeostatic imbalance can contribute to
an animal model or humans being at a greater risk for developing metabolic syndrome or diabetes with
ozone exposure. Together, these proposed pathways provide biological plausibility for epidemiologic
evidence of metabolic syndrome and/or diabetes with ozone exposure and will be used to inform a
causality determination, which is discussed later in this Appendix.
5.1.3 Glucose and Insulin Homeostasis
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. Thus, during feeding, blood insulin levels rise, stimulating glucose uptake and replenishing
body fuel reserves in the form of triglycerides and glycogen. When insulin levels decrease (e.g., during
fasting), fuels, such as lipids from adipose tissue and amino acids from muscle, are mobilized to the
bloodstream where they are used by the liver to synthesize glucose.
Clinical outcomes of impaired glucose regulation include diabetic ketoacidosis and diabetic
coma. Diabetic ketoacidosis, which is usually seen in type 1 diabetics, can result in unconsciousness from
September 2019
5-7
DRAFT: Do Not Cite or Quote
-------
1 a combination of a severely increased blood sugar level, dehydration, and accumulation of ketones or
2 acids that were formed as the diabetic body used fat for fuel instead of sugar. Diabetic coma is a
3 reversible form of coma found in people with diabetes which involves extremely low blood sugar.
4 The effects of short-term exposure to ozone on glucose and insulin homeostasis are characterized
5 below and utilize various techniques. The glucose tolerance test (GTT) involves the sampling of blood
6 glucose levels at multiple time points after glucose injection or ingestion to measure the body's response
7 to glucose and is used to diagnose or monitor diabetes or gestational diabetes. The insulin tolerance test
8 (ITT) involves insulin injection to fasting animals and glucose monitoring after injection. Ketone bodies
9 can be formed in diabetic ketoacidosis when energy production pathways are altered and higher levels of
10 ketones are generated in response to low insulin. The Homeostatic Model Assessment (HOMA) is a
11 method for assessing (3-cell function and insulin resistance.
5.1.3.1 Epidemiologic Studies
12 One epidemiologic study of short-term ozone exposure and glucose or insulin homeostasis was
13 reviewed in the 2013 Ozone ISA. Chuang et al. (2010) found increases in fasting glucose 5 days
14 following increased exposure to ozone. Recent epidemiologic studies provide some evidence of
15 associations between short-term ozone exposure and these endpoints (Table 5-5). Specifically:
16 • Kim and Hong (2012) found increases in fasting glucose (0.19%; 95% CI: 0.09, 0.28%), insulin
17 (0.71%; 95% CI: 0.02, 1.38%)1, and HOMA (0.30%; 95% CI: 0.06, 0.53%) in the Korean Elderly
18 Environmental Panel (KEEP). The KEEP cohort consisted of 560 Koreans over 60 years old. The
19 association of 5-day avg ozone concentration with glucose, insulin, and HOMA-IR was
20 approximately threefold larger in people with a previous diagnosis of type 2 diabetes (glucose
21 [0.68%; 95% CI: 0.28, 1.07%], insulin [2.76%; 95% CI: 0.78, 4.75%], HOMA [1.21%; 95% CI:
22 0.44, 1.99%]). In subjects without type 2 diabetes, an association with glucose was observed
23 (0.09%; 95% CI: 0.02, 0.16%), while associations with insulin and HOMA were not observed.
24 Copollutant models with NO2 and PM10 were also evaluated. The associations with glucose
25 remained after adjustment for NO2 (0.16%; 95% CI: 0.06, 0.25%) and PM10 (0.15%; 95% CI:
26 0.01,0.14%).
27 • Using 5,958 participants from the Framingham Offspring Cohort and Third Generation Cohort, Li
28 et al. (2017) completed a panel study evaluating the association of fasting glucose, insulin,
29 HOMA-IR, and other metabolic endpoints with 1- to 7-day moving avg ozone concentrations.
30 Decreases in fasting glucose were observed at 3-, 5-, and 7-day moving avg. No other endpoints
31 differed based on the qualitative results presented in the study.
32 • In a study of 1,023 Mexican Americans in southern California, Chen et al. (2016b) evaluated
33 changes to HOMA-IR, fasting glucose, and insulin resulting from short-term exposure to ozone.
34 The study used cumulative averages of daily ozone concentrations from 0-90 days prior to
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.
September 2019
5-8
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
testing. No associations were reported for any metabolic outcomes. Results were presented
qualitatively.
• One study evaluated hospital admissions for diabetic ketoacidosis and diabetic coma in the
Santiago region of Chile (Dales et al.. 2012V Insulin resistance may lead to these hospitalizations
in both type 1 and 2 diabetics. Using a 6-day distributed lag, a null association for the
relationships of hospital admissions for diabetic ketoacidosis or diabetic coma (RR: 1.02; 95%
CI: 0.996, 1.04) was observed. The effect remained null when divided into subregions of Santiago
and when CO, PMio, PM2.5, or SO2 were individually added to evaluate two-pollutant models.
5.1.3.2 Controlled Human Exposure Studies
Controlled human exposure studies of short-term ozone and glucose or insulin homeostasis were
not reviewed in the 2013 Ozone ISA. A recent study (Table 5-6) did not show evidence that short-term
ozone exposure affected these endpoints. In a study of 24 health volunteers aged 22-30 years, Miller et
al. (2016a) randomly exposed subjects to ozone (0.3 ppm) or filtered air over 2 hours while alternating
15 minutes of exercise and 15 minutes of rest during two clinic visits. After a 2-week wash-out period, the
subjects had the alternate exposure. Serum samples found no change in HOMA-IR or insulin levels
immediately following ozone exposure when compared with HOMA-IR and insulin levels immediately
following exposure to filtered air.
5.1.3.3 Animal Toxicological Studies
No animal toxicological studies of short-term ozone and glucose or insulin homeostasis were
reviewed in the 2013 Ozone ISA. A number of recent animal toxicological studies provide evidence for
changes in glucose and insulin homeostasis following short-term ozone exposure. As detailed below and
in the evidence inventory table that follows (Table 5-7). short-term ozone exposure is associated with
elevated fasting blood glucose, hyperglycemia, and glucose intolerance, endpoints which are controlled
by adrenal-cortex derived hormones, as demonstrated by the fact that removal of the adrenals attenuates
the ozone-dependent metabolic dysfunction (Miller et al.. 2016c). Exercise attenuates the
ozone-dependent glucose intolerance. Also, insulin homeostasis is altered by ozone exposure and varies
by animal model with mixed effects including some studies showing insulin resistance and others null
effects. Ozone exposure also induces impaired insulin secretion from (3-cells. Multiple metabolic
indicators demonstrate that ozone exposure impairs glucose and insulin homeostasis in animal
toxicological studies. Evidence inventory tables provide detailed information on experimental design and
the studies are characterized in greater detail below.
• In multiple studies of healthy animals, ozone induced hyperglycemia and impaired glucose
tolerance (glucose tolerance test) have been noted. Ozone induced hyperglycemia and glucose
intolerance after acute ozone exposure [Brown Norway rats, 1.0 ppm, 6 hours/day for 2 days;
Bass et al. (2013)1 or [Wistar Kyoto rats, 6 hours/day, 0, 0.5, or 1.0 ppm ozone; Miller et al.
(2015)1. These ozone effects were slightly reduced in animals that were intermittently exposed to
September 2019
5-9
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
ozone for 13 weeks. In a separate study, after 1 day of ozone exposure (5 hours/day, 1 or 3 days,
1.0 ppm exposure), adult male Wistar Kyoto rats had fasting hyperglycemia (Miller et al.. 2016b)
and glucose tolerance testing demonstrated that the rats had glucose intolerance across time
(statistically significantly increased area under the curve [AUC] over the time course of 2 hours,
measurements taken every 30 minutes). Adult male Fischer 344 rats that were acutely exposed to
ozone (4 hours/day, 1 day, 0.8 ppm ozone) had impaired glucose tolerance with the GTT
(Thomson et al.. 2018); specifically, peak glucose levels in ozone-exposed animals (30 minutes
post-glucose injection) were significantly higher than air-exposed controls.
• Ozone is known to increase circulating corticosterone in rats and humans (Miller et al.. 2016c;
Miller et al.. 2016a'). and removal of the adrenal corticosterone (bilateral total adrenalectomy
[ADX] or bilateral adrenal medullar ablation [AMX] in male Wistar Kyoto rats [12-13 weeks of
age, 1.0 ppm ozone, 4 hours/day for 1 or 2 days]) significantly attenuates or ameliorates the
perturbed metabolic response to ozone (Miller et al.. 2016c'). With ozone exposure,
hyperglycemia and glucose intolerance were significantly ameliorated with ADX and
significantly attenuated with AMX (Miller etal.. 2016c).
• Exercise's effect on glucose use after ozone challenge (0.25, 0.5, or 1.0 ppm, exercising or
sedentary female Long-Evans [LE] rats): The study showed that all ozone-exposed animals
(1.0 ppm) had elevated blood glucose after ozone exposure (Gordon et al.. 2017b). Also, glucose
tolerance was impaired in ozone-exposed animals; however, the exercising rats (0.25 and 0.5 ppm
ozone) recovered better from the glucose challenge 30 minutes post-exposure than did the
sedentary animals with a smaller glucose peak. Exercise confounded the effect of ozone on
glucose tolerance, and the highest dose of ozone increased the time required for serum glucose
levels to return to baseline, as measured over 2 hours after an initial glucose challenge.
• The effect of ozone on insulin homeostasis was measured in multiple toxicology studies of
healthy animals. In one study, ozone exposure decreased serum insulin (2 days of ozone exposure
1.0 ppm), but insulin levels returned to baseline after the period of recovery [18 hours later;
Miller et al. (2015)1. In a separate study using the insulin tolerance test (ITT), adult male Wistar
Kyoto rats exposed to ozone (5 hours/day, 1 day, 1.0 ppm exposure) had fasting hyperglycemia
(Miller et al.. 2016b). and glucose remained significantly elevated at the first measurements
(30 minutes after insulin injection), showing initial insulin resistance or residual hyperglycemia
that resolved and returned to baseline over the remainder of the testing period (every 30 minutes
out to 2 hours) indicating no prolonged insulin resistance. The ITT addresses pituitary or adrenal
function. Another endpoint studied (Miller et al.. 2016b) also demonstrated that ozone exposure
caused decreased glucose-stimulated insulin secretion in response to glucose injection with the
P-cell function test (serum insulin measurement, 30 minutes after glucose injection to fasting
animals) suggesting impairment of insulin secretion which has been linked to stress mediated
changes in metabolic response. In another study, ozone exposure (0.8 ppm for 16 hours) induced
systematic and peripheral insulin resistance (HOMA-IR, ITT, and the EH clamp technique) and
impaired insulin sensitivity in skeletal muscle (Vella et al.. 2014). Glucagon and insulin are
hormones secreted by the pancreas that counterbalance glucose changes. Glucagon keeps blood
glucose from dropping too low by stimulating the release of glucose into the bloodstream from
storage depots in the body. Whereas insulin controls glucose by signaling the body (liver and
adipose) to move glucose from the serum to storage depots. Nose-only ozone exposure (4 hours,
0.8 ppm) to male Fischer 344 rats resulted in significantly decreased plasma glucagon (Thomson
etal.. 2016V
• Multiple distinct diabetic and overweight rodent models have been used to explore the effects of
ozone in overweight or obese animals. KKAy mice are obese, diabetic, and have severe
hyperglycemia and insulin resistance at baseline in control filtered-air animals; with ozone
exposure (males, 0.5 ppm ozone for 13 consecutive weekdays), there were significant decreases
September 2019
5-10
DRAFT: Do Not Cite or Quote
-------
1 in fasting plasma insulin and HOMA-IR, but not altered insulin resistance (ITT, AUC); muscle
2 insulin signaling increased (Ying et al.. 2016). In a separate study by the same lab, another strain
3 of diabetes-prone mice (adult male KK mice) were exposed to ozone for 13 consecutive
4 weekdays [0.5 ppm, 4 hours/day; Zhong et al. (2016)1. While the fasting glucose levels were
5 unchanged between ozone and filtered-air controls, fasting insulin was significantly decreased,
6 insulin resistance was significantly elevated (AUC ITT), and P-cell insulin secretory function
7 (HOMA -%B) was significantly decreased. In summary, the KKAY obese diabetic mice have
8 decreased HOMA-IR, no change in insulin resistance and decreased fasting insulin with ozone
9 exposure. In a separate model, the KK diabetes-prone mice, weekday ozone exposure exacerbated
10 insulin resistance and impaired P-cell insulin secretion.
11 • The effects of age on metabolic response to ozone (1.0 ppm ozone, 6 hours/day for 2 days) were
12 studied by exploring effects in rats (male brown Norway rats) at ages 1, 4, 12, and 24-months.
13 Bass et al. (2013) reported reduced glucose intolerance in rats exposed to ozone across all age
14 groups. Also, ozone induced hyperglycemia in fasting rats at 1, 12, and 24 months of age;
15 4-month-old rats were refractory to hyperglycemia with ozone exposure.
5.1.3.4 Summary
16 Recent evidence from the epidemiologic literature shows associations between short-term ozone
17 exposure and fasting glucose levels, although some studies showed opposing associations. The one
18 controlled human exposure study showed no changes in glucose and insulin homeostasis with ozone
19 exposure to healthy volunteers but there were marked changes in lipid metabolism that were associated
20 with increased plasma Cortisol. It was concluded that intermittent exercise during ozone exposure lead to
21 the lack of hyperglycemia. However, the animal toxicological literature shows short-term ozone exposure
22 induces hyperglycemia, impaired glucose tolerance, increased gluconeogenesis, P-cell dysfunction, and
23 decreased glucagon levels. Further, animal toxicological literature shows that these ozone-dependent
24 effects can be ameliorated with removal of the adrenal glands, indicating the importance of the
25 neuro-endocrine system's sympathetic adrenal medullary axis and the hypothalamic-pituitary-adrenal axis
26 on the ozone-mediated effects. Despite limited epidemiologic and controlled human exposure literature,
27 the expanding animal toxicological literature shows robust evidence of short-term ozone exposure
28 contributing to impairment of glucose and insulin homeostasis.
5.1.4 Overweight and Obesity
5.1.4.1 Animal Toxicological Studies
29 The 2013 Ozone ISA included studies in overweight/obese animals demonstrated altered
30 respiratory responses and respiratory inflammation with ozone exposure compared to lean controls.
31 Genetically obese mice had airway hyper-responsiveness and responded more vigorously to acute ozone
September 2019
5-11
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
exposure than did lean controls (Shore. 2007). Studies done at the U.S. EPA examining effects of ozone
at various concentrations (0.25, 0.5, 1.0 ppm) in healthy and obese rat models with leptin receptor
mutation and associated cardiovascular disease demonstrated low sensitivity to ozone-induced lung injury
and neutrophilic inflammation (Kodavanti. 2015). Pulmonary inflammation and injury in response to
ozone were also enhanced [2 ppm ozone for 3 hours; Shore (2007)1. The 2013 Ozone ISA also included
studies in diet-induced obese animals, showing obesity-augmented inflammation and injury, as measured
by BALF markers, and enhanced AHR in mice exposed acutely to ozone [2 ppm ozone for 3 hours;
Johnston et al. (2008)1. Another study from the 2013 Ozone ISA at a lower exposure level (0.3 ppm
ozone, 72 hours) with the same genetically obese mice reported the inflammatory response following
exposure to ozone was dampened by obesity (Shore et al.. 2009). The ozone-dependent pulmonary injury
and inflammation (PMNs in the lung), and reduced pulmonary compliance seen in lean mice was
attenuated or absent in the obese animals (Shore et al.. 2009). Recent toxicological studies provided some
evidence that ozone may impair metabolism and affect body weight, BMI, and body composition, as well
as effect caloric intake. More detailed information on these studies is contained in the evidence inventory
(Table 5-8).
• Male Brown Norway rats exposed to ozone consumed more food and water than did air control
animals. High fat and high fructose were included in diet to determine whether animals with
ozone exposure on different diets had similar eating patterns (0.8 ppm ozone, 4 day/week for
3 weeks). Ozone exposure caused males on the control and high-fat diets to eat statistically
significantly more food and trended toward statistically significant increases on high fructose diet
(Gordon et al.. 2016). Ozone exposure caused males on normal diet and high-fructose diets to
drink statistically significantly more water. Ozone exposure caused animals on high-fat diets to
statistically significantly increase caloric intake compared with filtered-air controls on a high-fat
diet. Females were refractory to ozone-dependent changes. Other rodent strains developed
metabolic syndrome with high-fat or high-fructose diets, but Brown Norway rats less susceptible
to this.
• A diabetic mouse model (male KKAy mice, 0.5 ppm ozone for 13 consecutive weekdays)
provided evidence for reduced body-weight gain with ozone exposure (Ying et al.. 2016). KKAy
mice are diabetic, obese, and have severe hyperglycemia and insulin resistance at baseline; they
are a genetic model of obese type 2 diabetes, as described above. Reduction in body weight gain
has also been noted in healthy nondiabetic rats exposed to ozone (Henriquez et al.. 2018).
• Obesity is a risk factor for the development of type 2 diabetes and exercise can improve glucose
tolerance. To determine the role of maternal exercise and diet on ozone's effect on glucose
homeostasis and obesity in offspring, a study was conducted with multiple diet and exercise
options. A control (CD) or high-fat diet (HF) with or without exercise (run wheel, RW) was
provided to pregnant dams creating four exposure groups (CD, CD-RW, HF, and HF-RW). The
dams on the high-fat diet weighed more at the onset of pregnancy versus control dams and
produced offspring that weighed more at weaning (PND 27) but not in adulthood (PND 133),
independent of exercise status. When these offspring were challenged with ozone in adulthood
(0.8 ppm 4 hours/day, PND 161-162), baseline glucose levels in ozone-exposed males were
increased; females were refractory at baseline. Male and female offspring in all four exposure
groups were statistically significantly glucose intolerant at one or two time points over the 2-hour
glucose tolerance test when compared to filtered-air animals on the same diet and exercise
rcgimcn(Gordon et al.. 2017a). Glucose area under the curve during the glucose tolerance test
September 2019
5-12
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
was not measured, and comparisons were not made across groups. With a glucose challenge in
the glucose tolerance test, male and female animals on various diets, whether exercising or
sedentary had significantly elevated glucose versus air exposed animals; baseline glucose levels
were only elevated in male ozone-exposed animals.
• Genetically obese mice (dg/db) that were exposed short-term to ozone showed increased
pulmonary inflammation after ozone exposure (2 ppm ozone, 3 hours) compared with lean,
wild-type mice with mechanistic contribution from IL-17a and gastrin releasing peptide receptor
(Mathews ct al.. 20 IS). It was conclude that the type 2 inflammatory/cytokine reaction that
contributed to increased effects of ozone in obese mice may be driven by IL-33 (Mathews et al..
2017b). Further work examined how the metabolome differed between these lean and obese mice
and explored those differences with ozone exposure [2 ppm for 3 hours; Mathews et al. (2017aVI.
The lung metabolomes of the lean versus the obese mice differed at baseline, and pathways like
the glutathione pathway were differentially altered with ozone exposure. More information on
these studies is included in Appendix 3—Respiratory Health Effects.
In summary, ozone exposure changes eating patterns in control rodents on various diets, leading
males to eat more food and drink more water. Ozone induces glucose intolerance in multiple animal
models independent of diet (high-fat or control diet) or exercise status (exercising or sedentary). In one
genetic model of severe type 2 diabetes and overweight status, ozone exposure caused the animals to lose
weight as is seen with acute ozone exposure in healthy nonobese animals. Obese and diabetic animals
have a different pulmonary inflammatory response to ozone than lean animals.
5.1.5 Other Indicators of Metabolic Function
5.1.5.1 Inflammation
It is widely believed that inflammation plays a critical role in the development of type 2 diabetes
and atherosclerosis leading to CHD. As outlined in the Section 5.1.3 (Biological Plausibility),
inflammation may promote a peripheral inflammatory response in organs and tissues, such as liver and
adipose tissues. The role of systemic inflammation after acute ozone exposure may be seen in some but
not all strains of rodents used in animal toxicology studies. The role of systemic inflammation in ozone
exposure is covered in Section 4.1.11. of the Cardiovascular appendix. New evidence for peripheral
inflammation in adipose tissue following short-term exposure to ozone is presented below.
5.1.5.1.1 Animal Toxicological Studies
Inflammatory markers in adipose tissue are significantly elevated with ozone exposure in obese
and diabetic animals. Specific examples are detailed below and in the evidence inventory tables that
follow (Table 5-10). The inflammatory effects of ozone reach peripheral tissue like adipose.
September 2019
5-13
DRAFT: Do Not Cite or Quote
-------
1 • Obesity-prone mice (adult male KK mice) were exposed to ozone for 13 consecutive weekdays
2 [4 hours/day; Zhong et al. (2016)1. Epididymal adipose showed significantly increased
3 inflammation (increased monocytes/macrophages), increased expression of the chemokine
4 CXC-11, and significant increases in inflammatory gene expression (Ifn-g, IL-12, iNOS, cd56).
5 Ozone exposure to obesity-prone mice leads to increased visceral adipose inflammation as
6 measured by multiple aforementioned biomarkers.
7 • Inflammatory and oxidative stress biomarkers (Tnf-a, Mcp-1) were upregulated and
8 anti-inflammatory genes were downregulated (IL-10) in epicardial and perirenal adipose tissue in
9 rats (8-week-old Male Sprague-Dawley rats were fed a normal diet [ND] or high fructose diet
10 [HFr] for 8 weeks) exposed to ozone [0.5 ppm, 8 hours/day, 5 days/week, for 9 days over
11 2 weeks; Sun et al. (2013)1. There was significantly increased infiltration of macrophages that
12 was associated with increased expression of tumor necrosis factor a and iNOS.
13 • Inhalation exposure to ozone increased proinflammatory macrophages in adipose tissue of a
14 diabetic mouse model [male KKAy mice, 0.5 ppm ozone for 13 consecutive weekdays; Ying et
15 al. (2016)1.
16 A limited number of animal toxicological studies provide additional evidence that short-term
17 exposure to ozone may result in inflammation of the visceral or perirenal adipose tissue, which is
18 particularly relevant to metabolic function and a risk factor for metabolic syndrome.
5.1.5.2 Liver Outcomes
19 The liver, which is between the portal and systemic circulation, is the site for primary energy and
20 xenobiotic metabolism (Boron and Boulpaep. 2017V The liver is a crucial organ for the maintenance of
21 glucose homeostasis. It can be stimulated to increase blood glucose by inducing gluconeogenesis during
22 fasting or to store glucose after feeding. The liver can also synthesize and degrade protein, carbohydrates,
23 and lipids for distribution to extrahepatic tissues depending on energy needs. Finally, the liver regulates
24 whole-body cholesterol balance via biliary excretion of cholesterol, cholesterol conversion to bile acids,
25 and by regulating cholesterol synthesis (Boron and Boulpaep. 2017). The liver is also the site of
26 generation of ketone bodies, which are a biomarker for diabetes, because the diabetic body can switch to
27 using fats as its fuel source. Consequently, the liver is an essential regulator of whole-body metabolism
28 and energy homeostasis.
29 Acute-phase liver proteins, such as CRP, can act as sensors of liver function and are discussed in
30 more detail in Appendix 4. Section 4.1.11. An epidemiologic study found associations between CRP, a
31 protein that is produced in response to acute systemic inflammation, and ozone exposure. These proteins,
32 in combination with other liver enzymes can give information about overall health, including liver
33 function.
September 2019
5-14
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
5.1.5.2.1
Controlled Human Exposure Studies
No controlled human exposure studies of liver outcomes 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 5-6). Ketone body formation is a biomarker for diabetes and ketone bodies are
formed by the liver from fatty acids as a result of gluconeogenesis. 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 was associated with increased
carnitine conjugates of long-chain FFA and acetyl carnitine suggestive of accelerated (3-oxidation and
increased ketone body gcncration(Millcr et al.. 2016a).
5.1.5.2.2 Animal Toxicological Studies
Ozone exposure in animal models impacts various pathways that are mediated through the liver,
including increasing hepatic glucose production through gluconeogenesis (pyruvate tolerance test),
decreasing bile acid production, altering gut microbiome, impairing glycolytic pathways, altering
(3-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, but like other
pathways, the liver contributes to increased blood glucose with ozone exposure.
• Ozone induced hyperglycemia, impaired glucose tolerance, and altered cholesterol after 1 or
2 days of ozone exposure [Wistar Kyoto rats or Brown Norway rats, 6 hours/day, 0, 0.25, 0.5, or
1.0 ppm ozone; Miller etal. (2015); Bass et al. (2013)1. and pathways that may contribute to this
in the liver were delineated with metabolomic analysis of serum. Bile acids are made in the liver
from cholesterol, further processed by the gut microbiome, and released to the intestine to
facilitate absorption of dietary fat. Serum cholesterol and bile acid metabolites were significant
decreased by ozone exposure. Ozone increased circulating free fatty acids. Ozone also impaired
glucose homeostasis by perturbing glycolytic pathways (decreased anhydro glucitol [a biomarker
of glycemic control], increased fructose levels, increased pyruvate [Day 1], and decreased lactate
[Day 2]-glycolysis/glycolytic pathways). Mitochondrial [3-oxidation metabolites were reduced
with ozone exposure (Miller etal.. 2015). This metabolomic analysis demonstrates multiple
pathways that are affected by ozone exposure.
• Acute exposure of male Wistar Kyoto Rats to 1 ppm ozone (5 hours/day for 1 day) resulted in
hyperglycemia (Miller et al.. 2016b). To determine whether this ozone-dependent hyperglycemia
was controlled by liver gluconeogenesis, a pyruvate tolerance test (PTT) was performed where
pyruvate was injected and blood glucose was measured overtime. The PTT showed statistically
significant increased blood glucose with ozone exposure (1.0 ppm) compared with filtered-air
controls, confirming the stimulation of gluconeogenesis with ozone exposure. Further, at 1.0 ppm
ozone, glucose AUC was statistically significantly increased, confirming these findings.
• Short-term exposure to ozone (8 hours/day for 5 days to male Sprague-Dawley rats) is 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
September 2019
5-15
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
GRP-78 is a novel autoantigen in human type 1 diabetes), protein disulfide isomerase, and
glutathione S-transferase Ml (Thcis et al.. 2014).
5.1.5.2.3 Summary
Multiple metabolic indicators from the liver provide evidence that ozone exposure induces
changes 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.
5.1.5.3 Endocrine Hormones
Ozone exposure activates the autonomic sensory pathway, which triggers central neuroendocrine
stress response including responses like increased corticosterone, Cortisol, or epinephrine (Snow et al..
2018). Ozone acts as a pulmonary irritant and stimulates nasopharyngeal and pulmonary nerves and
receptors, including the trigeminal and vagal nerves, which induces downstream effects to the autonomic
nervous system and increases the levels of epinephrine (Snow et al.. 2018). The hypothalamus and
adrenals are activated with ozone exposure, and removal of the adrenal pathway with adrenalectomy or
pharmacologically can ameliorate the ability of ozone to induce metabolic homeostatic changes in
rodents.
5.1.5.3.1 Epidemiologic Studies
No epidemiologic studies in the 2013 Ozone ISA assessed the association between short-term
ozone exposure and endocrine hormones. As noted in Table 5-9. one recent study evaluated the
association between short-term ozone exposure and endocrine hormones. Using 5,958 participants from
the Framingham Offspring Cohort and Third Generation Cohort, Li et al. (2017) completed a panel study
evaluating adiponectin, leptin, and resistin over a 1- to 7-day moving avg. Based on the published
qualitative results, there were no changes due to short-term exposure to ozone, but adiponectin had a
positive trend and resistin had a negative trend.
5.1.5.3.2 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 5-6) used healthy adult human volunteers who
September 2019
5-16
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
were intermittently exercised and 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 minute on/off exercise in a controlled chamber).
Acute ozone exposure increased circulating stress hormones (Cortisol and corticosterone) in these
volunteers (Miller et al.. 2016a').
5.1.5.3.3 Animal Toxicological Studies
Ozone exposure activates the hypothalamic-pituitary adrenocortical stress pathway and its
associated release of stress hormones into the circulation (adrenaline, epinephrine, and
cortisol/corticosterone) in animal studies. Ozone also affects the hormones leptin and ghrelin, which are
related to energy balance and hunger/satiety control. Specific details of these studies are included in the
Evidence Inventory (Table 5-10).
• Circulating adrenaline, epinephrine, and cortisol/corticosterone are significantly increased in
laboratory animals after acute ozone exposure (Henriquez et al.. 2017a; Henriquez et al.. 2017b;
Miller et al.. 2016c; Miller et al.. 2016a; Miller et al.. 2015; Bass et al.. 2013; Thomson et al..
2013). Removal of input from the adrenals or the adrenal medullary system significantly
ameliorates or attenuates the metabolic effects of ozone exposure, respectively (Henriquez et al..
2017a; Henriquez et al.. 2017b; Miller et al.. 2016c).
• In healthy rodent models (Thomson et al.. 2016). the adrenocorticoid axis's contribution to
ozone-induced metabolic changes was monitored using metyrapone, a glucocorticoid synthesis
inhibitor. Ghrelin was statistically significantly decreased with ozone exposure, but pretreatment
with metyrapone did not alter the effect of ozone on ghrelin. Thus, ghrelin is significantly
decreased with ozone exposure, and this effect is independent of modification of the
adrenocortical pathway. Likewise; ozone-induced hypothermia which might be linked to global
metabolic changes was also not prevented by adrenalectomy (Henriquez et al.. 2017a) suggesting
that multiple neuroendocrine mechanisms might be altered after ozone exposure.
• In healthy rodent models, short-term ozone exposure was associated with either elevated serum
leptin (Miller et al.. 2015; Bass et al.. 2013; Sun et al.. 2013) or a trend toward increased leptin
(Gordon et al.. 2017b). In obese animals (Zhong et al.. 2016) and diabetic animals (Ying et al..
2016). there was significantly decreased serum leptin with ozone exposure. Thus, healthy and
diseased animal models have significantly different leptin responses to ozone exposure or the
temporality differences between studies might explain the directionality differences.
5.1.5.3.4 Summary
Recent evidence shows that neuroendocrine activation is essential to the perturbed metabolic
pathways that develop after ozone exposure. Elevated circulating stress hormones are consistently
observed in animal models and in controlled human exposure studies after short-term ozone exposure.
Removal of the neuroendocrine input by surgically removing the adrenal glands removes the
neuroendocrine stress activation, ameliorates the stress hormone response and attenuates glucose
intolerance and other factors that contribute to metabolic syndrome in rodents exposed to ozone. Thus,
September 2019
5-17
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
neuroendocrine stress activation is essential to the development of adverse metabolic outcomes after
short-term ozone exposure.
5.1.5.4 Serum Lipids
In the 2013 Ozone ISA, one epidemiologic study provided evidence of ozone exposure
association with 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. Apolipoprotein B (ApoB), as a lipid carrier
which transports triglycerides and cholesterol around the body, was associated with 3-day avg ozone
concentration. The 5-day mean ozone concentration was associated with increased fasting glucose levels
and triglycerides. In addition, 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.
5.1.5.4.1 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 5-9) evaluated the effects of
short-term ozone exposure on blood lipids. Chen et al. (2016a) used the (3-Gene cohort of 1,023 Mexican
Americans living in southern California. The study considered LDL levels and HDL-to-LDL ratios. The
study used cumulative averages of daily concentrations from 0-90 days prior to testing. No outcomes
were reported for any metabolic endpoints evaluated with short-term increases of ozone exposure. Results
were presented qualitatively.
5.1.5.4.2 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 5-6. there is one study of healthy adult human volunteers (n = 24) who
were exercised intermittently and 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 medium and long-chain FFA and plasma glycerol
consistent with enhanced lipolysis (Miller et al.. 2016a).
September 2019
5-18
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
5.1.5.4.3 Animal Toxicological Studies
Animal toxicology studies demonstrate that ozone exposure induces changes in serum lipids
including cholesterol, free fatty acids, and triglycerides. Animals exposed to ozone show increased serum
triglycerides, elevated free fatty acids, and altered serum cholesterol levels. The majority of the studies
mentioned below use male rodents and some study these outcomes in rodents that are obese, diabetic, or
have cardiovascular disease (CVD). More specific information follows below and detailed study design
can be found in the table that follows (Table 5-10).
• Ozone exposure to rodents can alter serum lipids, and this may differ by strain of rodent or by
rodent disease model (Ramot et al.. 2015). Some rodent models have significantly elevated serum
lipids before ozone challenge at baseline versus other rodents, especially the rodent models of
CVD, diabetes, or obesity.
• Ozone exposure has been associated with changes to serum triglycerides. In healthy brown
Norway rats, ozone induced increased serum triglycerides in a dose-dependent manner with
increasing age of the animal; the statistically significant triglyceride changes were highest in the
oldest animals exposed to ozone (1-, 4-, 12-, and 24-month-old males) which was measured
immediately after an exposure of 6 hours/day for 2 days to 1 ppm ozone (Bass et al.. 2013). In
healthy animals, ozone-induced statistically significantly increased triglycerides were ameliorated
with adrenalectomy or demedularization of the adrenal glands, which removes the section of the
adrenal gland that produces stress hormones [Wistar Kyoto rats, 1.0 ppm ozone, 4 hours/day for
2 days; Miller et al. (2016c)l. In animal models of CVD, ozone exposure increased serum
triglycerides [0.3 ppm ozone, 3 hours, 1 day, 12-week-old male spontaneously hypertensive (SH)
rats; Farrai et al. (2016)1.
• Ozone exposure can alter serum cholesterol. In healthy animal models, ozone exposure (1.0 ppm,
6 hours/day for 2 days, 10-week-old male Wistar Kyoto rats) resulted in statistically significantly
increased LDL cholesterol (Miller et al.. 2015). In animal models of CVD, ozone induced
statistically significant decreases in HDL cholesterol [0.8 ppm ozone, 4 hours; 12-week-old male
spontaneously hypertensive rats; Farrai et al. (2012); 1.0 ppm ozone, obese FHH rats, and obese
diabetic JCR rats Ramot et al. (2015)1. But some animals are refractory to cholesterol changes
with ozone exposure; ozone exposure did not significantly affect HDL or LDL cholesterol
[0.3 ppm ozone, 3 hours, 1 day exposure of 12-week-old male SH rats; Farrai et al. (2016)1. or
HDL cholesterol [0.2 ppm ozone, 4 hours, 12-week-old male SH rats; Farrai et al. (2012)1. In
CVD animal models, LDL cholesterol was statistically significantly decreased with a greater
ozone exposure [SH rats, 0.5 and 1.0 ppm ozone, 4 hours; Ramot et al. (2015)1. Ozone exposure
alters serum cholesterol in multiple animal models.
• Ozone exposure affects serum lipids immediately after exposure and can continue to have more
prolonged effects after a period of recovery. After a period of recovery from ozone exposure,
healthy animals had statistically significantly increased LDL [male WKY rats, 20 hours recovery
after 4 hours 1.0 ppm ozone; Ramot et al. (2015)1. Effects in healthy rodents of different ages
(male brown Norway rats; 1-, 4-, 12-, and 24-month-old males, 6 hours/day for 2 consecutive
days, 1 ppm ozone) included statistically significantly increased levels of serum HDL in
12-month-old animals with 1 ppm ozone exposure compared with filtered-air control measured
after 18 hours of recovery from ozone exposure; all other endpoints (HDL, LDL, and total
cholesterol), ages, and doses (0.25 ppm ozone) were refractory to change (Bass et al.. 2013).
Animals with CVD maintained altered cholesterol levels, including statistically significantly
decreased HDL [0.5 and 1.0 ppm ozone 4 hours, diabetic obese male JCR rats; 1.0 ppm 4 hours
September 2019
5-19
DRAFT: Do Not Cite or Quote
-------
1 ozone, obese male FHH rats, 20 hours recovery after ozone exposure; Ramot et al. (2015)1. and
2 statistically significantly increased total cholesterol [male SH rats, 4 hours exposure to 1.0 ppm
3 ozone, 20 hours recovery; Ramot et al. (2015)1. In healthy animals, there were significant
4 increases in all types of cholesterol measured [total, LDL and HDL; 1.0 ppm, 5 hours/day for 1 or
5 2 days, 10-week-old male Wistar Kyoto rats, measured 18 hours after exposure; Miller et al.
6 (2015)1. These studies show that cholesterol does not recover to baseline levels after recovery
7 from ozone exposure in these animals.
8 • The effect of exercise on ozone-dependent changes in serum lipids was examined in female LE
9 rats, specifically, the role of exercise training (active vs. sedentary lifestyle) in its contribution to
10 cholesterol changes after a 1-day ozone challenge. Rats exercised or remained sedentary from
11 weaning to age 10 weeks, whereupon they were exposed to ozone (0.8 ppm ozone, for 5 hours)
12 and their cholesterol levels measured (Gordon et al.. 2017b). All serum cholesterol measurements
13 showed no significant changes in cholesterol with ozone exposure (total cholesterol, HDL). Most
14 studies of the effects of ozone on serum lipids were conducted in male animals. Thus, female LE
15 rats were refractory to ozone-dependent changes in serum cholesterol. Interestingly, there was a
16 statistically significant decrease in running wheel activity the night after ozone exposure,
17 demonstrating changed behavior after ozone exposure.
18 • The effect of high-fat and high-fructose diets was tested in male brown Norway rats. With ozone
19 exposure (0.8 ppm ozone, 4 days/week for 3 weeks), there was significantly decreased serum
20 cholesterol in animals on control diet, an effect which was ameliorated with high-fat or
21 high-fructose diets (Gordon et al.. 2016). In fact, ozone induced statistically significantly
22 increased cholesterol in male animals on the high-fat diet versus high fat filter air controls. Serum
23 triglycerides were significantly increased in ozone-exposed male rodents on the control or
24 high-fat diets versus filter air controls. Females were refractory to change.
5.1.5.4.4 Summary
25 Multiple studies provide additional evidence that short-term exposure to ozone may result in
26 altered lipid homeostasis (cholesterol and triglycerides). Additionally, increases in free-fatty acid release
27 into the circulation, an indicator that the body has shifted toward using fats as its source of fuel in place of
28 glucose, as can be seen in diabetics, demonstrates neuroendocrine activation with ozone exposure. In
29 animal toxicology studies, removal of the neuroendocrine activation by adrenalectomy ameliorates the
30 ozone-dependent dyslipidemia. Ozone exposure induced metabolic changes in humans and animals,
31 including neuroendocrine activation and altered lipid metabolism, which is particularly relevant to
32 metabolic function and a risk factor for metabolic syndrome, especially with chronic exposure.
5.1.5.5 Blood Pressure
33 Short-term ozone exposure mediated effects on blood pressure are discussed in detail in the
34 Cardiovascular Appendix (see Appendix 4) (Akcilar et al.. 2015; Wagner et al.. 2014; Gordon et al..
35 2013; Uchivama and Yokovama. 1989; Uchivama et al.. 1986). Hypertension is a clinically relevant
36 consequence of chronically high blood pressure, which typically develops over years. High blood
37 pressure, dyslipidemia, increased fasting blood glucose, and obesity are criteria for metabolic syndrome,
September 2019
5-20
DRAFT: Do Not Cite or Quote
-------
1 which is a risk factor for heart disease, stroke, and diabetes. Recent epidemiologic evidence and human
2 exposure evidence is limited in number and generally inconsistent. Animal toxicological studies show
3 some evidence to suggest that short-term exposure to ozone can result in changes in blood pressure in
4 animals. However, some results also suggest that changes in diet may mediate these effects (see
5 Section 4.1.8).
5.1.6 Potential Copollutant Confounding of the Ozone-Metabolic Effects
Relationship
6 Few studies have examined potential short-term ozone exposure copollutant confounding with
7 PM2 5 or PM10, or gaseous copollutants. When associations were noted, the association with ozone
8 remained, and with null associations, the null effect also remained. This suggests that these findings may
9 not be substantially impacted by copollutant confounding.
10 • Kim and Hong (2012) analyzed the KEEP cohort consisting of 560 Koreans over 60 years old and
11 observed increases in fasting glucose (0.19%; 95% CI: 0.09, 0.28%). Copollutant models of NO2
12 and PM10 were also evaluated. The associations with glucose remained after adjustment for NO2
13 (0.16%; 95% CI: 0.06, 0.25%) and PM10 (0.15%; 95% CI: 0.01, 0.14%).
14 • One study evaluated hospital admissions for diabetic ketoacidosis and diabetic coma in Chile
15 (Dales et al.. 2012). Using a 6-day distributed lag for ozone, a weak, positive association was
16 observed for the risk of hospital admissions for diabetic ketoacidosis or diabetic coma in the
17 greater Santiago area (RR: 1.02; 95% CI: 1.00, 1.04). The effect remained relatively unchanged
18 when divided into subregions of Santiago and when the model added CO, PM10, PM2.5, and SO2.
5.1.7 Effect Modification of the Ozone-Metabolic Effects Relationship
5.1.7.1 Lifestage
19 The 1996 and 2006 Ozone AQCDs identified children, especially those with asthma, and older
20 adults as at-risk populations (U.S. EPA. 2006. 1996). In addition, the 2013 Ozone ISA confirmed that
21 there was adequate evidence to conclude that children and older adults are at increased risk of
22 ozone-related health effects (U.S. EPA. 2013). Collectively, the majority of evidence for older adults has
23 come from studies of short-term ozone exposure and mortality. A couple of recent studies of short-term
24 ozone exposure and metabolic effects compared associations between different age groups. One
25 epidemiologic study did not report consistent evidence that older adults are at increased risk for metabolic
26 effects; however, the animal toxicology study did see greater effects in aged animals.
27 • One study evaluated associations of short-term ozone exposure and hospital admissions for
28 diabetic ketoacidosis and diabetic coma in the Santiago region of Chile (Dales et al.. 2012). Using
September 2019
5-21
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
a 6-day distributed lag, a null association was observed for the relationships of hospital
admissions for diabetic ketoacidosis or diabetic coma (1.02; 95% CI: 1.00, 1.04). However, the
effect increased in populations aged 75-84 (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).
• 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.. 2013V Ozone-induced elevated blood glucose area under the curve is increased in an
age-dependent manner in rats with the greatest glucose elevation seen in the oldest animals (age
1,4, 12, and 24 months).
5.1.7.2 Pre-existing Disease
Individuals with certain pre-existing diseases may be considered at greater risk of an air
pollution-related health effect because their health is compromised depending on the type and severity of
their disease. The 2013 Ozone ISA concluded that there was adequate evidence for increased
ozone-related health effects among individuals with asthma (U.S. EPA. 2013). 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.
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:
• 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 5-day avg ozone concentration with glucose, insulin, and HOMA-IR 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%]). In subjects without type 2 diabetes, an association with glucose was observed
(0.09%; 95% CI: 0.02, 0.16%), while associations with insulin and HOMA were not found in
those without pre-existing type 2 diabetes.
Animal toxicological studies with animal models of cardiovascular disease with or without
obesity have shown differences in sensitivity to ozone in terms of circulating triglycerides and
cholesterol (Ramot et al.. 2015) and changes in transcriptional profile of the lung metabolic
pathways indicating animal model and disease specific expression signatures at baseline and after
ozone exposure in rat models of obesity with or without CVD (Ward and Kodavanti. 2015V
September 2019
5-22
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
5.1.8
Summary and Causality Determination
There were no causality conclusions for metabolic effects in the 2013 Ozone ISA (U.S. EPA.
2013). The literature pertaining to outcomes from short-term exposure to ozone and metabolic effects has
expanded substantially since the 2013 Ozone ISA (U.S. EPA. 2013). with multiple epidemiologic and
experimental studies and a few human clinical studies currently available for review. Findings from
animal toxicological studies of metabolic effects showed short-term ozone exposure impaired glucose and
insulin homeostasis (glucose intolerance, hyperglycemia, dyslipidemia of triglycerides, altered blood
pressure, impaired (3-cell function, increased hepatic gluconeogenesis, and neuroendocrine activation
contributing to altered metabolic function). Controlled human exposure to ozone in exercising
participants confirmed activation of the neuroendocrine system and showed formation of ketone bodies, a
biomarker of diabetes. Previous epidemiologic evidence demonstrates elevated HbAlc (a biomarker of
diabetes and an indicator of the degree of glycemic control in diabetics), increased triglycerides, altered
serum cholesterol, increased HOMA-IR, and fasting glucose level instability associated with short-term
ozone exposure. The information pertaining to the relationship between short-term exposure to ozone and
metabolic effects is summarized in Table 5-1. using the framework for causality determinations described
in the Preamble to the IS As (U.S. EPA. 2015).
The strongest evidence for an effect of short-term exposure to ozone on metabolic effects is
provided by animal toxicological studies that show impaired glucose tolerance, increased triglycerides,
fasting hyperglycemia, decreased insulin, and increased hepatic gluconeogenesis in various strains of
animals across multiple labs following short-term exposure to ozone. Biological plausibility is indicated
by results from controlled human exposure studies and animal studies that show ozone activates the
autonomic sensory pathway, which triggers central neuroendocrine stress response including responses
like increased corticosterone, Cortisol or epinephrine, as noted in the controlled human exposure study.
Ketone body formation, a biomarker of diabetes, is induced in controlled human exposure studies with
ozone exposure. This begins when ozone acts as a pulmonary irritant and stimulates nasopharageal and
pulmonary nerves and receptors including the trigeminal and vagal nerves, which induces downstream
effects on the autonomic nervous system. The sympathetic adrenal-medullary (SAM) and
hypothalamus-pituitary-adrenal (HP A) axes are activated with ozone exposure and removal of the adrenal
pathway with adrenalectomy or pharmacologically can ameliorate the ability of ozone to induce
metabolic syndrome in rodents. Despite limited epidemiologic and controlled human exposure literature,
the expanding animal toxicological literature shows robust evidence of short term ozone exposure
contributing to impairment of glucose and insulin homeostasis. These findings are coherent with
epidemiologic studies that report associations with perturbations to glucose and insulin homeostasis with
ozone exposure. Overall, the collective evidence is sufficient to conclude that a likely to be causal
relationship exists between short-term ozone exposure and metabolic effects.
September 2019
5-23
DRAFT: Do Not Cite or Quote
-------
Table 5-1 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
Animal toxicological studies of impaired
glucose tolerance, fasting
hyperglycemia, dyslipidemia, hepatic
gluconeogenesis, and activation of the
neuroendocrine pathway with ozone
exposure
Animal toxicological evidence of
increased inflammation
Section 5.1. 0.25-1 ppm
Miller et al. (2016c). Miller
etal. (2015) Miller et al.
(2016b), Thomson et al.
(2018)
Yinq et al. (2016); Zhonq 0.25-1 ppm
etal. (2016); Sun etal.
(2013)
Animal toxicological evidence of Bass et al. (2013). Farrai 0.25-1 ppm
dyslipidemia et al. (2012), Farrai et al.
(2016), Gordon et al.
(2016). Miller et al.
(2016c), Ramot et al.
(2015)
Animal toxicological evidence of
liver-mediated effects
Miller etal. (2016b); Miller 0.25-1 ppm
etal. (2015); Theis etal.
(2014)
Consistent
epidemiologic
evidence of
increased risk
diabetes or
metabolic syndrome
Epidemiologic evidence for positive
associations between short-term ozone
exposure and increased indicators of
impaired glucose and insulin
homeostasis, including HOMA-IR,
dyslipidemia, elevated HbA1c, and
increased fasting glucose
Chuana et al. (2011)
Section 5.1.4.1
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
Li etal. (2017); Chen etal.
(2016b); Dales et al.
(2012); Kim and Hong
(2012)
Section 5.1.3.1
Controlled human
exposure evidence
of increased
metabolic changes
with ozone
exposure at
relevant
concentrations
A limited number of studies observed
ketone body formation and
neuroendocrine system activation with
ozone exposure
Miller etal. (2016a)
0.3 ppm
September 2019
5-24
DRAFT: Do Not Cite or Quote
-------
Table 5-1 (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
Limited
The magnitude of ozone associations
Kim and Honq (2012)
Section 5.1.6
epidemiologic
remains relatively unchanged in a limited
Dales et al. (2012)
evidence from
number of studies evaluating copollutant
copollutant models
models, including PM2.5 and other
provides some
gaseous pollutants
support for an
independent ozone
association
Biological
Experimental studies provide evidence
Section 5.1.2
0.3-2.0 ppm
plausibility
of metabolic syndrome mediated by
pulmonary irritant receptor stimulation
and activation of the neuroendocrine
system with short-term ozone exposure
provides biological plausibility to the
effects of ozone on metabolic syndrome
and diabetes
HbA1c = hemoglobin A1c; HOMA-IR = Homeostatic Model Assessment of Insulin Resistance; PM25 = particulate matter with a
nominal aerodynamic diameter less than or equal to 2.5 |jm; ppm = parts per million.
September 2019
5-25
DRAFT: Do Not Cite or Quote
-------
5.2 Long-Term Ozone Exposure—Introduction, Summary from
the 2013 Ozone ISA, and Scope for Current Review
1 Metabolic effects were not included in the 2013 Ozone ISA as a distinct section because there
2 were limited studies evaluating the effects of long-term ozone exposure on metabolic effects. One study
3 presented in the cardiovascular disease appendix evaluated the association between long-term exposure of
4 ozone and effects in blood lipids and glucose homeostasis (Chuang etal.. 2011); it reported increases in
5 total cholesterol, fasting glucose, and hemoglobin Ale. Multiple experimental animal studies have
6 evaluated ozone-mediated effects, and these studies indicate that long-term exposure to ozone may affect
7 glucose homeostasis and factors that may contribute to metabolic syndrome.
8 The metabolic effects from long-term ozone exposure reviewed here include indicators of
9 metabolic function that underlie metabolic and cardiovascular diseases. These include glucose and insulin
10 homeostasis, adiposity, weight gain, metabolic syndrome, type 1 and 2 diabetes, and mortality from
11 diabetes or cardiometabolic diseases. The subsections below provide an evaluation of the most
12 policy-relevant scientific evidence relating long-term ozone exposure to metabolic effects. These sections
13 focus on studies published since the completion of the 2013 Ozone ISA.
5.2.1 Population, Exposure, Comparison, Outcome, and Study Design
(PECOS) Tool
14 The scope of this section is defined by a scoping tool that generally defines the relevant
15 Population, Exposure, Comparison, Outcome, and Study Design (PECOS). The PECOS tool defines the
16 parameters and provides a framework to help identify the relevant evidence in the literature to inform the
17 ISA. Because the 2013 Ozone ISA did not make a causality determination for long-term ozone exposure
18 and metabolic health effects, the epidemiologic studies evaluated are less limited in scope and not
19 targeted towards specific study locations, as reflected in the PECOS tool. The studies evaluated and
20 subsequently discussed within this section were identified using the following PECOS tool:
21 Experimental Studies:
22 • Population: Study populations of any controlled human exposure or animal toxicological study of
23 mammals at any lifestage
24 • Exposure: Long-term (over 30 days) inhalation exposure to relevant ozone concentrations
25 (i.e., <2 ppm for mammals)
26 • Comparison: In toxicological studies of mammals and in controlled human exposures, an
27 appropriate comparison group that is exposed to a negative control (i.e., clean air or filtered-air
28 control)
29 • Outcome: Metabolic effects
September 2019
5-26
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
• Study Design: In vivo acute, subacute, or repeated-dose toxicity studies in mammals,
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.
September 2019
5-27
DRAFT: Do Not Cite or Quote
-------
Long-te?rr<
Ozone
Ejpoaure
At&vattixv oT
5*ivt0ry
m
R«plr#tory
T<#Ct
GesiatC'na
Diabetes
mwxm
~ |,.»| MwttWy |
: IMul
hcrta$«j
S»rum
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.
Figure 5-2 Potential biological pathways for metabolic outcomes following
long-term ozone exposure.
1 Ozone exposure can induce irritant signaling, both nasopharyngeal and pulmonary, and activate
2 the trigeminal and vagus nerves. Ozone exposure can also directly lead to hypothalamic-pituitary-adrenal
3 axis activation or this pathway can be activated through irritant signaling as further described in the
4 pulmonary appendix. The ozone-dependent activation of the HPA axis has been shown to occur with
5 short-term ozone exposure (Miller et al.. 2016c). As in the short-term controlled human exposure studies
6 and epidemiologic studies, ozone exposure in animals activates the autonomic sensory pathway, which
7 triggers central neuroendocrine stress response including responses like increased corticosterone, Cortisol,
8 or epinephrine. Ozone acts as a pulmonary irritant and stimulates nasopharyngeal and pulmonary nerves
9 and receptors, including the trigeminal and vagal nerves, which induces downstream effects to the
10 autonomic nervous system. The hypothalamus and adrenals are activated with ozone exposure and
September 2019 5-28 DRAFT: Do Not Cite or Quote
-------
1 removal of the adrenal pathway either with adrenalectomy or pharmacologically can ameliorate the ability
2 of ozone to induce metabolic syndrome in rodents.
3 With long-term ozone exposure, animals develop hyperglycemia, glucose intolerance, peripheral
4 muscle insulin resistance, decreases circulating insulin, and inhibition of glucose-dependent insulin
5 release rSection 5.2.3; Miller et al. (2016b); Bass et al. (2013)1. Studies also show dyslipidemia (elevated
6 triglycerides and decreased HDL cholesterol) with ozone exposure (Bass et al.. 2013). Fewer animal
7 toxicological studies exist on long-term ozone exposure but outcomes studied in both short- and
8 long-term studies both show consistently impaired metabolic homeostasis with ozone exposure.
9 The animal toxicology demonstrates the pathways of disruption, provides a plausibility for the
10 long-term adverse human health effects, including diagnosis of metabolic syndrome, diabetes, and
11 mortality resulting from diabetes or cardiometabolic diseases.
5.2.3 Glucose and Insulin Homeostasis
12 Insulin is secreted by (3-cells within the pancreas in response to glucose levels. When glucose
13 levels rise, depolarization of the pancreatic (3-cells or modulation by other hormones stimulate insulin
14 secretion. Thus, during feeding, blood insulin levels rise stimulating glucose uptake and replenishing the
15 body's fuel reserves in the form of triglycerides and glycogen. When insulin levels decrease (e.g., during
16 fasting) fuels such as lipids from adipose tissue and amino acids from muscle are mobilized to the
17 bloodstream where they are used by the liver to synthesize glucose. Ozone has been shown to impair
18 glucose and insulin homeostasis in health animals. Details of these studies follow below.
5.2.3.1 Animal Toxicological Studies
19 The 2013 Ozone ISA did not contain information on long-term ozone exposure and metabolic
20 effects. Since that time, several new animal toxicology studies have been published showing the effects of
21 long-term ozone on glucose and insulin homeostasis (e.g., glucose tolerance test, insulin tolerance test
22 fasting glucose and insulin, blood glucose and insulin levels, (3-cell insulin secretion test). With long-term
23 exposure, healthy animals develop hyperglycemia, glucose intolerance, and insulin resistance, and
24 inhibition of glucose-dependent insulin release. Senescent animals were more sensitive to
25 ozone-dependent serum insulin changes in a study that examined an age by ozone effect. Specific
26 information is detailed below and in the evidence inventory tables that follow (Table 5-13).
27 • Subchronic ozone-induced glucose intolerance was evaluated in young and old male brown
28 Norway rats (4, 12, and 24 months of age) exposed 2 days/week for 13 weeks [0.25 or 1.0 ppm
29 ozone; Bass et al. (2013)1. Glucose tolerance testing was performed immediately after the last
30 ozone exposure. Glucose tolerance was statistically significantly impaired in all age groups
31 (1.0 ppm ozone). AUC was not measured in this study, but in the 12-month-old animals,
September 2019
5-29
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
0.25 ozone exposure trended toward glucose intolerance with higher blood glucose levels and a
longer decay to baseline compared with control animals. This study compared acute exposure
(Section 5.1.2) with this subchronic exposure, and the acutely exposed animals had greater
glucose intolerance than did animals with subchronic ozone exposure. Nonetheless, all of the
animals exposed to ozone had impaired glucose tolerance.
• Ozone-induced glucose intolerance was followed in adult male Wistar Kyoto rats (250-300 g)
after 12 weeks of ozone exposure [1.0 ppm, 5 hours/day for 3 consecutive days/week; Miller et
al. (2016b) I. With ozone exposure, these animals had fasting hyperglycemia and statistically
significant glucose intolerance. With the glucose tolerance test, glucose AUC was statistically
significantly increased with ozone exposure versus filtered-air exposure. With the insulin
tolerance test, ozone-exposed fasting animals were hyperglycemic at baseline versus air control
and remained significantly elevated at the first two measurements (30 and 60 minutes after insulin
injection during the insulin tolerance test but not at 1 and 2 hours), but ozone AUC was not
significantly increased over control air animals, demonstrating that insulin resistance did not
remain over the 2-hour time course. Ozone caused an impaired insulin response with significantly
decreased serum insulin as measured with the [3-cell insulin secretion test (serum insulin
measurement, 30 minutes after glucose injection to fasting animals, 12 weeks after ozone
exposure). Effects were not seen in animals exposed to 0.25 ppm ozone. Long-term ozone
(1.0 ppm) significantly lowered circulating insulin and significantly impaired glucose-stimulated
(3-cell insulin secretion. Ozone-exposed animals had fasting hyperglycemia and were less able to
respond to a glucose challenge.
• Permanence of effects in these same animals was tested by 1 week of recovery with filtered air
after 13 weeks of ozone exposure (Miller et al.. 2016b). Glucose intolerance was ameliorated
1 week after the end of ozone exposure; ozone-exposed animals were no longer hyperglycemic at
baseline, and glucose tolerance testing was no different from air controls. Thus, ozone-dependent
systemic metabolic change was reversible after a period of recovery or exposure to clean air.
• The effect of ozone on metabolism was assessed in aged animals versus young adult animals.
Ozone-exposed aged (senescent) males had significantly increased serum insulin versus aged
filtered-air controls and aged ozone-exposed animals had significantly increased insulin versus
adult ozone-exposed animals [male brown Norway rats chronically exposed to 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 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 nonsenescent younger adult rodents.
5.2.4 Adiposity, Weight Gain, and Obesity
Adiposity (particularly visceral adiposity) and weight gain are risk factors for metabolic
syndrome, type 2 diabetes, and cardiovascular disease. Although most epidemiologic studies consider
BMI as a potential confounder or modifier of the association between ozone and cardiovascular disease,
there were no studies of the association of long-term exposure to ozone with adiposity or weight gain
reviewed in the 2013 Ozone ISA.
September 2019
5-30
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
5.2.4.1
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
(Table 5-11). 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 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) in a large cohort of African American adult women.
• 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 Center for Disease Control and Prevention (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. Additionally, the study did not
report correlations for the 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 (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). Similar to the other study,
copollutant correlations were not reported, and both PMio and SO2 observations were high in the
33 communities, so it is difficult to estimate the level of confounding from other ambient
pollutant exposures.
5.2.4.2 Animal Toxicological Studies
Elevated body weight, BMI, and adiposity are risk factors for metabolic syndrome as is
dyslipidemia (altered serum cholesterol or triglycerides). The 2013 Ozone ISA contained no animal
toxicological studies on these endpoints with ozone exposure. The effect of long-term exposure to ozone
on body weight was studied in one recent animal toxicological study and the rodents displayed no
ozone-dependent changes to body weight or body composition. Serum lipids (triglycerides and HDL
cholesterol) were significantly changed with ozone exposure in aged animals [24-month-old males,
September 2019
5-31
DRAFT: Do Not Cite or Quote
-------
1 0.25 ppm ozone, 6 hours/day, 2 days/week for 13 weeks; Bass et al. (2013)1. an effect not seen in younger
2 animals with the same exposure.
3 • Ozone had no effect on body composition and body weight of brown Norway rats (young adult
4 4 months old and aged 20 months old) with long-term ozone exposure [6 hours/day, 1 day/week
5 for 17 weeks; Gordon et al. (2013)1; also these animals displayed no changes to body composition
6 (fat or lean mass) with ozone exposure (Gordon et al.. 2013). Brown Norway rats tend to be less
7 susceptible to metabolic changes than do some other strains of rodents.
5.2.5 Metabolic Syndrome and Type 2 Diabetes
8 Criteria for metabolic syndrome include high blood pressure, dyslipidemia (elevated triglycerides
9 and low levels of high density lipoprotein [HDL] cholesterol), obesity (particularly central obesity), and
10 increased fasting blood glucose (FBG). Table 5-2 provides the criteria for a clinical diagnosis for
11 metabolic syndrome.
Table 5-2 Criteria for clinical diagnosis of metabolic syndrome.
Risk Factor
Threshold
Waist circumference
>89 cm in women and >102 cm in males
Triglycerides3
>150 mg/dL (1.7 mmol/L)
HDL-Ca
<40 mg/dL (1.0 mmol/L in males); <50 mg/dL (1.3 mmol/L) in females
Blood pressure15
Systolic >130 and/or diastolic >85 mm Hg
Fasting glucose0
>100 mg/dL (5.6 mmol/L)
HDL-C = HDL cholesterol; mg/dL = milligrams per deciliter; mm Hg = millimeters of mercury; mmol/L = millimoles per liter.
aA person taking drugs used to lower triglycerides or raise HDL-C is considered to exceed the threshold.
bA person taking blood pressure medication is considered to exceed the threshold.
°A person taking glucose-regulating medication is considered to exceed the threshold.
Source: Permission pending. Adapted from Alberti et al. (2009).
12 The diagnostic testing criteria for diabetes are listed in Table 5-3. The Ale, which is also known
13 as the hemoglobin Ale, HbAlc, or glycohemoglobin, is a blood test that provides information about a
14 person's average blood glucose over the past 3 months by measuring the percentage of hemoglobin (i.e., a
15 blood protein with a 3-month lifespan) modified by glucose. In controlled human exposure, animal
16 toxicological, and epidemiologic studies, the homeostasis model assessment (HOMA) has been widely
17 used to quantify insulin resistance (HOMA-IR) and pancreatic p-cell (HOMA-(3) function and used to
September 2019
5-32
DRAFT: Do Not Cite or Quote
-------
1 infer diabetes risk. The HOMA-IR index is given by the product of basal insulin and glucose levels
2 divided by 22.5, whereas the HOMA-J3 index is derived from the product of 20 and basal insulin levels
3 divided by glucose concentration minus 3.5 (Wallace et al.. 2004; Matthews et al.. 1985).
Table 5-3 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 2-hour plasma glucose >200 mg/dL (11.1 mmol/L) during OGTT. The test should be
test (OGTT) 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 (2014).
5.2.5.1 Epidemiologic Studies
4 No long-term epidemiologic studies of metabolic syndrome or type 2 diabetes were evaluated in
5 the 2013 Ozone ISA. Recent studies, which are listed in Table 5-12. include large cohort studies around
6 the world; they provide evidence for increased incidence for type 2 diabetes and metabolic syndrome.
7 Specifically:
8 • Jerrett et al. (2017) analyzed data from the Black Women's Health Study Cohort in a prospective
9 study of type 2 diabetes. The 43,003 women were greater than 30 years old, resided in
10 56 metropolitan areas, and had BMI information at baseline. Ozone was estimated using the
11 CMAQ model 8-hour max concentration for the centroid of the census tract of residence between
12 2007-2008 to approximate long-term averages. The study observed increased hazard ratios for
13 incident diabetes (1.28; 95% CI: 1.06, 1.55); however, when adjusted for NO2, this relationship
14 was slightly weaker and had wider confidence intervals (1.20; 95% CI: 0.96, 1.50).
September 2019
5-33
DRAFT: Do Not Cite or Quote
-------
1 • Using the Rome Longitudinal Study Cohort, Renzi et al. (2017) evaluated the effects of ozone
2 exposure in over one million subjects over 35 years old without diabetes at baseline. The study
3 used the Flexible Air Quality Regional Model (FARM) with a 1-km grid dispersion and 2005
4 seasonal ozone (May-September 8-hour avg) to predict the spatial distribution of ozone in Rome
5 between 2008-2013. The study showed increased hazard ratios for incidence of diabetes for those
6 living in Rome (1.01; 95% CI: 1.00, 1.02). Additionally, when the ozone model was adjusted for
7 NOx, the increased incidence remained significant (1.02; 95% CI: 1.00, 1.03).
8 • Yang et al. (2018) looked at the odds of developing metabolic syndrome due to exposure to ozone
9 in adults from the 33 Communities Chinese Health Study Cohort in participants that were
10 18-74 years of age and had lived in the same location for more than 5 years. Ozone exposure was
11 measured at municipal air monitoring stations, and the 8-hour daily mean concentrations were
12 aggregated into a 3-year avg. In a study population of 15,477, odds of metabolic syndrome
13 increased (1.16; 95% CI: 1.12, 1.23) according to the American Heart Association definition. The
14 study reported high correlations of ozone with PMio (0.81 ± 0.002 SD) and SO2
15 (0.84 ± 0.001 SD); these high correlations provide potential for copollutant confounding, and are
16 a source of uncertainty in estimating the direct effect of ozone on metabolic syndrome.
5.2.6 Type 1 Diabetes
17 Type 1 diabetes mellitus (T1D), which typically affects children and young adults, is a chronic
18 condition that results when the pancreas fails to produce the insulin needed for glucose homeostasis.
19 There were no epidemiologic studies of TID in the 2013 Ozone ISA. The evidence relating to the effect of
20 long-term exposure to ozone on TID is limited to a prospective study in Scania, Sweden rMalmqvist et al.
21 (2015); Table 5-141. The study evaluated prenatal exposure during first, second, and third trimesters of
22 pregnancies for children born between 1999-2005. Ozone exposure was measured by the nearest
23 monitoring station, averaging the 24-hour ozone concentrations and aggregating them into trimester
24 averages. The levels were categorized in quartiles with the reference exposure being set at a level less
25 than 22 ppb and the highest quartile exposure over 30.6 ppb. There were elevated ORs for type 1 diabetes
26 in the highest quartile of ozone concentrations in the first trimester (1.52; 95% CI: 0.88, 2.61) and second
27 trimester (1.62; 95% CI: 0.99, 2.65), although confidence intervals were wide. There was no evidence of
28 association with third-trimester exposure levels.
5.2.7 Gestational Diabetes
29 Several studies of gestational diabetes were conducted. Generally, the results of the studies were
30 inconsistent, although several reported positive associations with gestational diabetes or impaired glucose
31 tolerance with ozone exposures during the second trimester. While the evidence base for gestational
32 diabetes is growing, it is still limited to a relatively small number of studies which report generally
33 inconsistent results (see the "Pregnancy and Birth Outcomes" section for more details [Section 7.1.31).
September 2019
5-34
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
5.2.8
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, specifically whether there is evidence of 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 rTurner et al. (2016); Crouse et al. (2015); see Section 6.2.3.2. Figure 6-10 for more details].
5.2.9 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 increased 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 (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), however when adjusted for PM2 5 it 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).
5.2.10 Effect Modification of the Ozone-Metabolic Effects Relationship
5.2.10.1 Lifestage
The 1996 and 2006 Ozone AQCDs identified children, especially those with asthma, and older
adults as at-risk populations (U.S. EPA. 2006. 1996). In addition, the 2013 Ozone ISA confirmed that
September 2019
5-35
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
there was adequate evidence to conclude that children and older adults are at increased risk of
ozone-related health effects (U.S. EPA. 2013V Collectively, the majority of evidence for older adults has
come from studies of short-term ozone exposure and mortality. One recent study of short-term ozone
exposure and metabolic health effects compared associations between different age groups
(Section 5.1.6). but it does not report consistent evidence that older adults are at increased risk. Long-term
exposure to ozone associations with lifestage are described below.
• Using the Rome Longitudinal Study Cohort, Renzi et al. (2017) evaluated the effects of ozone
exposure in over one million subjects over the 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 (1.05; 95% CI: 1.02, 1.08) but not those from 50-60 (1.02; 95% CI: 0.99, 1.04), or over
60 (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). There was no difference between the groups.
• 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 Center for Disease Control and Prevention (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 (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), the population over 50 had
increased odds of being overweight (1.12; 95% CI: 1.05, 1.19), and females over 50 also had
increased odds of obesity (1.23; 95% CI: 1.04, 1.44). There were no differences found in the
under 50 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 versus 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.
5.2.10.2 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 varying with 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. 2013). The results of controlled
September 2019
5-36
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
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 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 the 35 years old without diabetes at baseline. When
stratified by subjects that had comorbidities (myocardial infarction, COPD, hypertension, or
hyperlipidemia) had an increased incidence of diabetes (1.02; 95% CI: 1.00, 1.03), but did not
differ from those without comorbidities had an increased HR for diabetes (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).
5.2.11 Summary and Causality Determination
There were no causality determinations for metabolic effects in the 2013 Ozone ISA (U.S. EPA.
2013). The recent literature pertaining to long-term exposure to ozone and metabolic effects has expanded
substantially since the 2013 Ozone ISA, with multiple epidemiologic and experimental studies currently
available for review. In prospective cohort studies in the U.S. and Europe increased incidence of type 2
diabetes was observed with long-term exposure to ozone. In China, the odds of metabolic syndrome
increased as well. These findings are consistent with two long-term ozone exposure studies in China, one
in adults and one in children, presented increased odds of obesity in both adults and children as obesity is
a risk factor for type 2 diabetes (T2D). 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
mortality findings 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). Findings from the one epidemiologic study of metabolic
disease showed increases in metabolic syndrome for both the Joint International and American Heart
Association criteria in 33 communities in China. Additionally, in prospective cohort studies in the U.S.
and Europe, increased incidence of type 2 diabetes was observed with ozone exposure. The information
pertaining to the relationship between long-term exposure to ozone and metabolic effects is summarized
in Table 5-4. using the framework for causality determination described in the Preamble to the ISAs (U.S.
EPA. 2015V
Experimental animal studies address some of the uncertainty in the epidemiologic evidence
related to the independent effect of ozone exposure by providing evidence of direct effects on metabolic
function. The animal toxicological studies provided evidence that long-term ozone exposure resulted in
September 2019
5-37
DRAFT: Do Not Cite or Quote
-------
1 impaired insulin signaling, glucose intolerance, hyperglycemia, and insulin resistance (Section 5.2.3.1). In
2 addition, these pathophysiological changes were often accompanied by increased inflammatory markers
3 in peripheral tissues, and activation of the neuroendocrine system (Section 7.2.1.5). A limited number of
4 epidemiologic studies have evaluated potential copollutant cofounding for PM or NOx rJerrett et al.
5 (2017); Renzi et al. (2017); Section 5.2.31. Importantly, short-term ozone exposure studies also provided
6 evidence that ozone exposure could contribute to the development of metabolic syndrome and show
7 consistency with the evidence that long-term ozone exposure could lead to development or worsening of
8 metabolic syndrome or its risk factors. Overall, the collective evidence is sufficient to conclude that a
9 likely to be causal relationship exists between long-term ozone exposure and metabolic effects.
September 2019
5-38
DRAFT: Do Not Cite or Quote
-------
Table 5-4 Summary of evidence to support a likely to be 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
animal
toxicology
evidence from
multiple,
high-quality
studies at
relevant ozone
concentrations
Animal toxicological studies of impaired glucose
tolerance, fasting hyperglycemia, dyslipidemia, insulin
resistance, and activation of the neuroendocrine
pathway with ozone exposure
Section 5.2.3
Miller et al. (2016b);
Bass et al. (2013)
0.25, 1.0 ppm
Consistent
epidemiologic
evidence of
increased risk
diabetes or
metabolic
syndrome
Increased odds of metabolic syndrome, increased
hazard ratio for incidence of diabetes, increased
hazard ratio for incident diabetes in U.S. cohort
Yang et al. (2018)
Jerrett et al. (2017);
Renzi et al. (2017)
See Section 5.2.5.1
for exposure
information
Increased odds of developing of gestational diabetes
with ozone exposure in the second trimester.
Elevated ORs for type 1 diabetes with higher ozone
concentrations in first and second trimester.
Malmavist 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)
Section 5.2.8
Limited Limited number of epidemiologic studies evaluate Jerrett et al. (2017); Section 5.2.3
epidemiologic potential copollutant cofounding for PM or NOx Renzi et al. (2017)
evidence from
copollutant
models
provides some
support for an
independent
ozone
association
Biological Experimental studies provide evidence of metabolic Section 5.2.2
plausibility syndrome mediated by neuroendocrine activation with
long-term ozone exposure provides biological
plausibility to the effects of ozone on metabolic
syndrome and diabetes
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.
September 2019
5-39
DRAFT: Do Not Cite or Quote
-------
5.3 Evidence Inventories—Data Tables to Summarize Study
Details
Table 5-5 Epidemiologic studies of short-term exposure to ozone and
glucose/insulin homeostasis.
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 years old in the
Seongbuk-Gu area
of Seoul, South
Korea
Daily mean
concentration
of monitor
nearest
residence
24-h avg
Mean: 19.38 Correlation
Median:
19.34
75th
90th
95th
26.67
29.56
31.33
(r):
N02: -0.35;
SO2: -0.3;
Other: PM10:
-0.12
Copollutant
models with:
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)
September 2019
5-40
DRAFT: Do Not Cite or Quote
-------
Table 5-5 (Continued): Epidemiologic studies of short-term exposure to ozone
and glucose/insulin homeostasis.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
HR (95% CI)
tChen et al.
(2016b)
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
of monitored air
quality data
spatially
mapped to
residence
locations using
inverse
distance
squared
interpolation
with a
maximum
radius of 50 km
24-h avg
Mean:
Correlation
Qualitative results only, no
30-day
(r): PM2.5:
change in fractional
cumulative:
30 day:
disappearance rate
43.4 ppb;
-0.02;
Glucose
Annual
annual: 0.04;
HOMA-IR
average:
NO2: 30 day:
40.8 ppb
-0.37;
Insulin
annual: -0.31
Metabolic clearance
Copollutant
Insulin sensitivity
models with:
NR
tLi et al. (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)
Daily averages
of two ozone
monitors in the
greater Boston,
MA area
24-h avg
Mean: 23.7
Correlation
(r):
PM2.5: 0.01;
NO2: -0.54;
SO2: 0.13;
Other:
BC: -0.26
Copollutant
models with:
NR
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
tDales et al.
n = general
Daily averaged
Mean: 64.41 Correlation
Increased risk for hospital
(2012)
population of
monitor(s) in
(r):
admission for diabetic
Santiago
Province, Chile
five sectors was
the sector of
PM2.5: -0.31;
coma or diabetic
5 million
residence
NO2: -0.31;
ketoacidosis: 1.02 (1.00,
Ozone:
Daily hospital
24-h avg
SO2: -0.08
1.04)
2001-2008
admissions where
Copollutant
Cross-sectional
study
diabetes was the
models with:
principal diagnosis
NR
(insulin dependent
and noninsulin-
dependent) with
coma or
ketoacidosis
BC = black carbon; HOMA-IR = Homeostatic Model Assessment of Insulin Resistance; mg/dL = milligrams per deciliter;
mmol/L = millimoles per liter; N02 = nitrogen dioxide; 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.
fStudies published since the 2009 PM ISA.
September 2019
5-41
DRAFT: Do Not Cite or Quote
-------
Table 5-6 Controlled human exposure study of short-term exposure to ozone
and glucose/insulin homeostasis and other metabolic indicators.
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
0.3 ppm, 2 h (15 min of exercise
alternating with 15 min of rest)
HOMA-IR, insulin, Cortisol,
corticosterone, cortisone, leptin,
ketone bodies, free fatty acids
September 2019
5-42
DRAFT: Do Not Cite or Quote
-------
Table 5-7 Study-specific details from animal toxicological studies of short-term
exposure to ozone and glucose/insulin homeostasis.
Species (Strain), N, Sex,
Study Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Wagner et al.
(2014)
Rats (S-D)
n = 4-8/group males,
0 females
Age: 8 weeks
0.5 ppm, 8 h/day for 2 weeks (O3 Blood pressure
insulin
and O3 + CAPs; high fructose or
normal diet for 8 weeks prior) for
9 consecutive weekdays
resistance, fasting levels of
blood glucose and
triglycerides, HOMA-IR,
body weight, heart rate
(24 h PE)
Bass et al. (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
GTT, AUC, epinephrine,
cholesterol (total, HDL,
LDL), triglycerides, serum
leptin, IL-6, Insulin, mRNA
biomarkers in liver and
adipose (NR)
Vellaetal. (2014)
Rats (Wistar) 0.8 ppm, 16 h (with and without
n = 4-10 males, 0 females pretreatment of N-acetylcysteine)
Age: adult (400-450 g)
HOMA-IR, glucose (fasting
glucose, insulin, ITT,
triacylglycerols, total
cholesterol, 1ST [l-arginine]),
serum oxidative stress
biomarkers (GSH/GSSG,
MDA), glucose-dep JNK,
AKT, or ER pathways
PE
Zhonq et al. (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,
days
4 h/day for 3 consecutive
Glucose metabolic
hormones (insulin, leptin,
adiponectin), visceral
adipose characterization
(oil-red-o stain),
inflammatory genes in
adipose (CXCL-11, IFN-g,
TNF-a, IL-12, and iNOS)
22 h PE
Insulin tolerance test (IP)
2 h PE
Thomson et al.
(2016)
Rats (F344)
n = NR males, 0 females
Age: adult (200-250 g)
0.8 ppm, 4 h
Glucose met hormones
(glucagon, insulin, ghrelin,
PAI-1; NR)
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; ADREX),
or sham surgery (SHAM)
GTT (blood glucose, AUC)
Immediately PE, lipids, free
fatty acids, branched chain
amino acids, leptin
September 2019
5-43
DRAFT: Do Not Cite or Quote
-------
Table 5-7 (Continued): Study-specific details from animal toxicological studies of
short-term exposure to ozone and glucose/insulin
homeostasis.
Study
Species (Strain), N, Sex,
Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
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
Glucose tolerance test,
body composition (lean, fat,
fluid percentage), BALF,
EMKA plethysmography,
Beta cell insulin secretion
test-inhibition, insulin
resistance test in liver,
insulin resistance test
muscle.
Immediately PE Day 1
Miller et al. (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
GTT (blood glucose, AUC),
insulin tolerance test,
pyruvate tolerance test
(hepatic gluconeogenesis)
cholesterol, catecholamines,
adrenaline and
noradrenaline, AKT (NR)
Miller et al. (2015)
Rats (WKY)
Male
Age: 10 weeks
0.25, 0.50, or 1.0 ppm ozone,
6 h/day for 2 days
Cholesterol, LDL
Farrai etal. (2012)
Rats (SH)
Male
Age: 12 weeks
0.8 ppm ozone, 4 h, whole body
exposure
Cholesterol, HDL
Farrai et al. (2016)
Rats (SH)
Male
Age: 12 weeks
0.3 ppm ozone, 3 h, 1 day whole
body exposure
Cholesterol
Ramotetal. (2015)
Rats (WKY)
Male
1.0 ppm ozone, 4 h, 1 day whole
body exposure
Cholesterol, LDL
Thomson et al. Rats (Fisher-344) 0.8 (with or without Glucose tolerance test,
(2018) n = 6-8/group males metyrapone) ppm, 4 h, whole body HOMA IR, plasma
0 females ' exposure triglycerides, HPA axis (cort
synth inhibitor metyrapone,
9e' a u or exogenous cort), glucose
met hormones (glucagon,
insulin, leptin, GLP-1,
ghrelin, cort), inf cytokines
(TNF, IL-6, VEGF, PAI-1;
NR)
AKT = protein kinase B; AUC = area under the curve; BALF = bronchoalveolar lavage fluid; BN = brown Norway; CAP = criteria air
pollutants; ER = estrogen receptor; F344 = Fischer 344; GSH/GSSG = ratio of reduced to oxidized glutathione; GTT = glucose
tolerance test; HDL = high density lipoproteins; HOMA-IR = Homeostatic Model Assessment of Insulin Resistance;
HPA = hypothalamic-pituitary-adrenal; ITT = insulin tolerance test; JNK = c-Jun N-terminal kinase; LDL = low density lipoproteins;
LE = Long-Evans; MDA = malondialdehyde; NR = not reported; 03 = ozone; PE = post-exposure; ppm = parts per million;
S-D = Sprague-Dawley; SH = spontaneously hypertensive.
September 2019
5-44
DRAFT: Do Not Cite or Quote
-------
Table 5-8 Study-specific details from animal toxicological studies of short-term
overweight and obesity.
Study
Species (Strain), N, Sex, Age
Exposure Details
(Concentration,
Duration)
Endpoints Examined
Yina etal. (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
HOMA-IR, ITT, inflammatory
cytokines, body weight, leptin,
hyperglycemia, in vitro insulin
treatment, GTT, area under the
curve, plasma insulin, plasma
glucose, white adipose cell
inflammation (NR)
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 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)
Serum cholesterol,
triglycerides, body weight,
effect of diet and exercise on
endpoints, changes in body
composition (fat, lean, liquid
mass)
18 h PE
Mathews et al.
(2017b)
Mice (wild type lean or obese;
dg/db)
2 ppm ozone, 3 h, whole
body exposure
Gastrin releasing peptide
receptor, IL-17a and IL-33
signaling
Mathews et al.
(2017a)
Mice (wild type lean or obese;
dg/db)
2 ppm ozone, 3 h, whole
body exposure
Lung metabolome, antioxidant
signaling in lung (glutathione
pathway)
Gordon etal. (2017a) Rats (LE)
n = 8 offspring total, four males,
four females when possible.
10 dams/treatment group.
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
0.8 ppm, 4 h/day for
2 consecutive days, Age:
adult (30 days) offspring
ozone challenge, PND
161-162 ozone exposure
Glucose tolerance test,
ventilation, BALF counts
(PND 162, offspring
measurements)
Mathews et al. (2018) Mice (wild type lean or obese; 2 ppm ozone, 3 h, whole Gastrin releasing peptide
dg/db) body exposure receptor, IL-17a and IL-33
signaling
BALF = bronchoalveolar lavage fluid; BN = brown Norway; GTT = glucose tolerance test; HOMA-IR = Homeostatic Model
Assessment of Insulin Resistance; ITT = insulin tolerance test; NR = not reported; PND = postnatal day; ppm = parts per million.
September 2019
5-45
DRAFT: Do Not Cite or Quote
-------
Table 5-9 Epidemiologic studies of short-term exposure to ozone and other
indicators.
Study
Study
Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
HR (95% CI)
tChen etal. (2016b)
California, U.S.
Ozone: 2002-2008
Follow-up: 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: 30-day
Correlation (r):
of monitored
cumulative:
PM2.5:
air quality data
43.4 ppb;
30 day: -0.02;
spatially
annual average:
annual: 0.04;
mapped to
40.8 ppb
NO2: 30 day:
residence
-0.37;
locations using
annual: -0.31
inverse
Copollutant
distance
models with:
squared
NR
interpolation
with a
maximum
radius of
50 km
24-h avg
Daily averages
Mean: 23.7
Correlation (r):
of two ozone
PM2.5: 0.01;
monitors in the
NO2: -0.54;
greater
SO2: 0.13;
Boston, MA
Other:
area
BC: -0.26
24-h avg
Copollutant
models with:
NR
Qualitative results
No change in:
HDL-to-LDL ratio,
LDL
tLi etal. (2017)
Northeastern U.S.
Ozone: 2002-2005 and
2008-2011
Follow-up: panel study
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)
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; LDL = low-density lipoproteins;
mg/dL = milligrams per deciliter; mmol/L = millimoles per liter; N02 = nitrogen dioxide; NR = not reported; PM2.5 = particulate matter
with a nominal aerodynamic diameter less than or equal to 2.5 |jm; ppb = parts per billion; S02 = sulfur dioxide.
fStudies published since the 2009 PM ISA.
September 2019
5-46
DRAFT: Do Not Cite or Quote
-------
Table 5-10 Study-specific details from animal toxicological studies of
short-term, other metabolic indicators.
Species (Strain), N,
Study Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Thomson et al. (2013)
Rats (F344)
n = 4-6/group
males, 0 females
Age: adult
(200-250 g)
0.4 ppm, 4 h (nose only)
0.8 ppm, 4 h (nose only)
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)
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)
Perirenal and epicardial
adipose tissue (PAT and
EAT), inflammation in PAT
and EAT, 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
24 h PE
Bass et al. (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
GTT, AUC, epinephrine,
cholesterol (total, HDL, LDL),
triglycerides, serum leptin,
IL-6, insulin, 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
GTT, insulin, leptin, IL-6,
cholesterol (total, LDL, HDL),
metabolomics, liver
transcriptomics
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)
September 2019
5-47
DRAFT: Do Not Cite or Quote
-------
Table 5-10 (Continued): Study-specific details from animal toxicological studies
of short-term, other metabolic indicators.
Species (Strain), N, Exposure Details
Study Sex, Age (Concentration, Duration) Endpoints Examined
0.5 ppm, 4 h/day for 13 consecutive HOMA-IR, ITT, inflammatory
days cytokines, body weight, leptin,
hyperglycemia, in vitro insulin
treatment, GTT, area under
the curve, plasma insulin,
plasma glucose, white adipose
cell inflammation (NR)
Gordon et al. (2016) Rats (BN)
n = 10/group males,
10/group females
Age: 30 days
18 h PE
Yina et al. (2016) Mice (KKAy, diabetic
prone)
n = 8/group males,
0 females
Age: 4-5 weeks
0.8 ppm, 5 h (high fructose or high
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)
Serum cholesterol,
triglycerides, body weight,
effect of diet and exercise on
endpoints, changes in body
composition (fat, lean, liquid
mass)
Zhona et al. (2016) Mice (KK; 0.5 ppm,
obesity-prone days
develops moderate
degrees of obesity,
insulin resistance,
and diabetes)
n = 8/group males,
no females
Age: adult
h/day for 3 consecutive Glucose metabolic hormones
(insulin, leptin, adiponectin),
visceral adipose
characterization (oil-red-o
stain), inflammatory genes in
adipose (CXCL-11, IFN-g,
TNF-a, IL-12, and iNOS)
22 h PE
Insulin tolerance test (IP)
2 h PE
Miller et al. (2016c) Rats (WKY) 1 ppm, 4 h/day for 1 or 2 days (rats GTT (blood glucose, AUC)
n = 5/group males, underwent bilateral adrenal Immediately PE
0 females demedullation [DEMED], total
bilateral adrenalectomy; ADX), or
Age: adult shgm surgery (SHAM)
Gordon et al. (2017b) Rats (LE) 0.25 ppm, 5 h/day for 2 days Glucose tolerance test, body
n = 0 males, 0.5 ppm, 5 h/day for 2 days composition (lean, fat, fluid
10/group females , ppm, 5 may for 2 „ays ^ThTsmogiap^'' ^
Age: 22 days at start v 3 . v 3
of exercise regimen Immediately PE Day 1
Miller etal. (2016b)
Male Rats (WKY)
n = 8-10/group
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
GTT (blood glucose, AUC),
Insulin tolerance test, beta cell
insulin secretion test, pyruvate
tolerance test (hepatic
gluconeogenesis) cholesterol,
catecholamines, adrenaline
and noradrenaline,
corticosterone, insulin
resistance-AKT (NR)
September 2019
5-48
DRAFT: Do Not Cite or Quote
-------
Table 5-10 (Continued): Study-specific details from animal toxicological studies
of short-term, other metabolic indicators.
Species (Strain), N,
Study Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Henriauez et al.
(2017b)
Male Rats (WKY)
Age: 12 weeks old,
n = 6-8/group
Rodents were pretreated daily for
7 days with propranolol (PROP; a
nonselective (3 adrenergic receptor
antagonist), mifepristone (a
glucocorticoid receptor antagonist),
both drugs, or respective vehicles,
and then exposed to air or ozone
(0.8 ppm), 4 h/day for 1 or
2 consecutive days while continuing
drug treatment
Inflammation, epinephrine,
Cortisol, lung transcriptomic
assessment
AKT = protein kinase B; AUC = area under the curve; BALF = bronchoalveolar lavage fluid; BN = brown Norway; FA = filtered air;
HDL = high-density lipoproteins; LDL = low-density lipoproteins; NR = not reported; 03 = ozone; PE = post-exposure; ppm = parts
per million; S-D = Sprague-Dawley; WKY = Wistar Kyoto.
September 2019
5-49
DRAFT: Do Not Cite or Quote
-------
Table 5-11 Epidemiologic studies of long-term exposure to ozone and
overweight and obesity.
Study
Study
Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
HR (95% CI)
tDonq 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
Mean: 27.4
within 1 km of Maximum:
school
8-h max
44.5
Correlation (r): Increased odds of obesity
NR
Copollutant
models with:
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 with:
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 males
obesity: 0.99 (0.88, 1.12)
Increased odds of
females obesity: 1.12
(1.01, 1.26)
tWhite etal. (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
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 with:
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.
fStudies published since the 2009 PM ISA.
September 2019
5-50
DRAFT: Do Not Cite or Quote
-------
Table 5-12 Epidemiologic studies of long-term exposure to ozone and metabolic
syndrome and type 2 diabetes.
Exposure
Study Study Population Assessment Mean (ppb)
Copollutant Effect Estimates
Examination HR (95% CI)
tYanq et al. (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 Mean: 25.1
1 km of the
household,
using the daily
8-h avg to
create a 3-yr
avg
concentration
8-h avg
Correlation (r): Increased odds
Maximum: 36.0
PM2.5: 0.45;
NO2: 0.45;
SO2: 0.84;
Other:
PM10O.8I
Copollutant
models with:
NR
of metabolic
syndrome
diagnosis
American Heart
Association
criteria: 1.164
(1.12, 1.23)
Joint international
criteria: 1.21
(1.02, 1.39)
tJerrett et al. (2017)
56 metropolitan areas
U.S.
Ozone: 2007-2008
Follow-up: 1995-201'
Cohort study
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: 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)
Black Women's
Health Study
n = 43,003
Black women
living in
56 metropolitan
areas in the U.S.,
aged 30 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 with:
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)
September 2019
5-51
DRAFT: Do Not Cite or Quote
-------
Table 5-12 (Continued): Epidemiologic studies of long term exposure to ozone
and metabolic syndrome and type 2 diabetes.
Exposure Copollutant Effect Estimates
Study Study Population Assessment Mean (ppb) Examination HR (95% CI)
tRenzi etal. (2017)
Rome, Italy
Ozone: 2005
Follow-up: 2008-2013
Cohort study
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)
Rome Longitudinal Flexible Air Mean: 49.4
Correlation (r): Increased
Study
n = 1,319,193
Individuals over
35 yr of age
without prevalent
diabetes at
baseline
Quality
Regional
Model (FARM)
using a 1-km
grid dispersion
model
8-h avg
Maximum: 57.3
PM2.5: -0.01;
NO2: -0.16
Copollutant
models with:
NR
prevalence of
diabetes at
baseline: 1.001
(0.991, 1.012)
Increased
incidence of
diabetes: 1.012
33CCHS = 33 Communities Chinese Health Study; CMAQ = Community Multiscale Air; HR = hazard ratio; N02 = nitrogen dioxide;
NR = not reported; 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.
fStudies published since the 2009 PM ISA.
September 2019
5-52
DRAFT: Do Not Cite or Quote
-------
Table 5-13 Study-specific details from long-term animal toxicological studies of
glucose and insulin homeostasis.
Study
Species (Strain),
Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Bass et al. (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
GTT, AUC, epinephrine,
cholesterol (total, HDL, LDL),
triglycerides, serum leptin, IL-6,
insulin, mRNA biomarkers in liver
and adipose (NR)
Gordon et al. (2013)
Rats (BN)
n = 7-10/group males,
0 females
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
0, 0.8 ppm, NR
1 ppm, 2 h
HOMA-IR, glucose (fasting
glucose, insulin, ITT,
triacylglycerols, total cholesterol,
1ST [l-arginine]), serum oxidative
stress biomarkers (GSH/GSSG,
MDA), glucose-dep JNK, AKT, or
ER pathways (1 day PE)
Pre- and post-ozone levels of
lung antioxidants (e.g., total
glutathione, ascorbic acid, uric
acid, alpha-tocopherol),
superoxide dismutase (SOD)
and enzyme content/activity
related to glutathione recycling
and differences across strains
(NR)
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
GTT (blood glucose, AUC),
insulin tolerance test, pyruvate
tolerance test (hepatic
gluconeogenesis), cholesterol,
catecholamines, adrenaline and
noradrenaline, AKT (NR)
AKT = protein kinase B; AUC = area under the curve; BN = brown Norway; ER = estrogen receptor; GSH/GSSG = ratio of reduced
to oxidized glutathione; GTT = glucose tolerance test; HDL = high-density lipoproteins; ITT = insulin tolerance test; JNK = c-Jun
N-terminal kinase; LDL = low-density lipoproteins; LE = Long-Evans; MDA = malondialdehyde; NR = not reported;
PE = post-exposure; ppm = parts per million; S-D = Sprague-Dawley; WKY = Wistar Kyoto.
September 2019
5-53
DRAFT: Do Not Cite or Quote
-------
Table 5-14 Epidemiologic studies of long-term exposure to ozone and type 1
diabetes.
Study
Study
Population
Exposure
Assessment
Mean
(PPb)
Copollutant
Examination
Effect Estimates
HR (95% CI)
tMalmavist 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 with:
NR
Quartile 4 vs.
reference exposure
HR = hazard ratio; NR = not reported; T1 D= type 1 diabetes.
fStudies published since the 2009 PM ISA.
September 2019
5-54
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
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 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
(Klimischetal.. 19971. STROBE guidelines (von Elm et al.. 20071. and ARRIVE guidelines (Kilkenny et al.. 20101.
September 2019
DRAFT: Do Not Cite or Quote
5-55
-------
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 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.
September 2019
5-56
DRAFT: Do Not Cite or Quote
-------
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 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.
September 2019
5-57
DRAFT: Do Not Cite or Quote
-------
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., 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 oron-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.
September 2019
5-58
DRAFT: Do Not Cite or Quote
-------
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 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.
September 2019
5-59
DRAFT: Do Not Cite or Quote
-------
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, 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.
September 2019
5-60
DRAFT: Do Not Cite or Quote
-------
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 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 (2008).
bMurqia et al. (2014); Weakley et al. (2013); Yang et al. (2011); Heckbert et al. (2004); Barret al. (2002); Muhaiarine
etal. (1997); Toren et al. (1993).
cBurnev et al. (1989).
September 2019
5-61
DRAFT: Do Not Cite or Quote
-------
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-S081.
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.
Alberti. KG: Eckel. RH: Grundy. SM: Zimmet. PZ: Cleemaa 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.
Atwal. OS: Wilson. T. (1974). Parathyroid gland changes following ozone inhalation: a morphologic study. Arch
Environ Occup Health 28: 91-100.
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: China S: Poissoa 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.
Chea BY: Chea CH: Chuang. YC: Kim. H: Honda. Y: Chiang. HC: Guo. YL. (2016a). Schoolchildren's
antioxidation genotypes are susceptible factors for reduced lung function and airway inflammation caused by
air pollution. Environ Res 149: 145-150. http://dx.doi.Org/10.1016/i.envres.2016.05.007.
Chea Z: Salam. MT: Toledo-Corral. C: Watanabe. RM: Xiang. AH: Buchanaa TA: Habre. R: Bastain. TM:
Lurmana F: Wilsoa JP: Trigo. E: Gilliland. FD. (2016b). 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: Yaa 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.
Chuang. KJ: Yaa 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/10.1136/oem.2009.0527Q4.
Clemons. GK: Garcia. JF. (1980). Changes in thyroid function after short-term ozone exposure in rats. J Environ
Pathol Toxicol 4: 359-369.
September 2019
5-62
DRAFT: Do Not Cite or Quote
-------
Crouse. PL: Peters. PA: Hvstad. P: Brook. JR: van Donkelaar. A: Martig 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: Rea W: Flick. LH: Fu. J: Wang. J: Chen. W: Simckes. M: Trevathaa 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.
Dorado-Martinez. C: Parades-Carbaial. C: Mascher. D: Borgonio-Perez. G: Rivas-Arancibia. S. (2001). Effects
of different ozone doses on memory, motor activity and lipid peroxidation levels, in rats. Int J Neurosci 108:
149-161. http://dx.doi.org/10.3109/00207450108986511.
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.
Gackiere. F: Saliba. L: Baude. A: Bosler. O: Strube. C. (2011). Ozone inhalation activates stress-responsive
regions of the CNS. J Neurochem 117: 961-972. http://dx.doi.Org/10.llll/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: 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.
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: McTiernaa A: Gaziano. JM: Frishmaa 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. A: House. J: Miller. DB: Snow. SJ: Fisher. A: Ren. H: Schladweiler. MC: Ledbetter. AD: Wright. F:
Kodavanti. UP. (2017a). Adrenal-derived stress hormones modulate ozone-induced lung injury and
inflammation. Toxicol Appl Pharmacol 329: 249-258. http://dx.doi.Org/10.1016/i.taap.2017.06.009.
September 2019
5-63
DRAFT: Do Not Cite or Quote
-------
Henriguez. AR: Snow. SJ: Schladweiler. MC: Miller. CN: Dye. 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.
Henriguez. AR: Snow. SJ: Schladweiler. MC: Miller. CN: Dye. JA: Ledbetter. AD: Richards. JE: Mauge-Lewis.
K: Mcgee. MA: Kodavanti. UP. (2017b). 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.
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.
Johnston. RA: Theman. TA: Lu. FL: Terry. RD: Williams. ES: Shore. SA. (2008). Diet-induced obesity causes
innate airway hyperresponsiveness to methacholine and enhances ozone-induced pulmonary inflammation. J
Appl Physiol (1985) 104: 1727-1735. http://dx.doi.org/10.1152/iapplphvsiol.00075.20Q8.
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. 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: Tillmana 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. (2015). Air pollution and insulin resistance: do all roads lead to Rome? [Comment]. Diabetes 64:
712-714. http://dx.doi.org/10.2337/dbl4-1682.
Li. M: Oian. Z: Vaughn. M: Boutwell. B: Ward. P: Lu. T. ao: Lia S: Zhao. Y: Zeng. XW: Liu. RO: Oin. XD:
Zhu. Y: Chea 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.
Li. W: Dorans. KS: Wilker. EH: Rice. MB: Kloog. I: Schwartz. JD: Koutrakis. P: Coull. BA: Gold. PR: Meigs.
JB: Fox. CS: Mittlemaa MA. (2017). Ambient air pollution, adipokines, and glucose homeostasis: The
FraminghamHeart Study. Environ Int 111: 14-22. http://dx.doi.Org/10.1016/i.envint.2017.ll.010.
Malmgvist. 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.
Mathews. JA: Kasahara. PI: 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.0181017.
Mathews. JA: Krishnamoorthv. N: Kasahara. PI: Cho. Y: Wurmbrand. AP: Ribeiro. L: Smith. P: Umetsu. P:
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.
Matthews. PR: Hosker. JP: Rudenski. AS: Navlor. BA: Treacher. PF: Turner. RC. (1985). Homeostasis model
assessment: Insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations
in man. Piabetologia 28: 412-419. http://dx.doi.org/10.1007/BF00280883.
September 2019
5-64
PRAFT: Po Not Cite or Quote
-------
Mautz. WJ: Bufalino. C. (1989). Breathing pattern and metabolic rate responses of rats exposed to ozone. Respir
Physiol Neurobiol 76: 69-77.
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/10.1164/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/SQ895-
4356(97)00019-X.
Mumaw. CL: Levesaue. S: Mcgraw. C: Robertson. S: Lucas. S: Stafflinger. JE: Campea MJ: Hall. P:
Norenberg. JP: Anderson. T: Lund. AK: Mcdonald. JP: 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.lQ96/fi.201500047.
Murgia. N: Brisman. J: Claessoa A: Muzi. G: Olin. AC: Torea 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.
Ramot. Y: Kodavanti. UP: Kissling. GE: Ledbetter. AP: 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.
Renzi. M: Cerza. F: Gariazzo. C: Agabiti. N: Cascini. S: Pi Pomenicantonio. R: Pavoli. 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: 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.
Shore. SA. (2007). Obesity and asthma: Lessons from animal models [Review]. J Appl Physiol (1985) 102: 516-
528. http://dx.doi.org/10.1152/iapplphvsiol.00847.20Q6.
Shore. SA: Lang. JE: Kasahara. PI: Lu. FL: Verbout. NG: Si. H: Williams. ES: Terry. RP: Lee. A: Johnston.
RA. (2009). Pulmonary responses to subacute ozone exposure in obese vs. lean mice. J Appl Physiol (1985)
107: 1445-1452. http://dx.doi.org/10.1152/iapplphvsiol.00456.2009.
Snow. SJ: Henriquez. AR: Costa. PL: Kodavanti. UP. (2018). Neuroendocrine regulation of air pollution health
effects: emerging insights. Toxicol Sci 164: 9-20. http://dx.doi.org/10.1093/toxsci/kfV129.
Sua L: Liu. C: Xu. X: Ying. Z: Maiseveu. A: Wang. A: Allea K: Lewandowski. RP: Bramble. LA: Morishita.
M: Wagner. JG: Pvonch. JT: Sua Z: Yan. X: Brook. RP: Raiagopalaa S: Harkema. JR: Sua O: Faa 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.
September 2019
5-65
PRAFT: Po Not Cite or Quote
-------
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.ll.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: Piloa 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: Kumarathasaa 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: Beckermaa 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/10.1164/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-95/001F). 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] (pp. 1-166). (EPA/630/P-03/001F). 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://cfpub.epa.gov/ncea/isa/recordisplav.cfm?deid=247492.
September 2019
5-66
DRAFT: Do Not Cite or Quote
-------
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/eele280e77586de
985257b65005d37e7!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.
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.
Uchivama. 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/S0013-9351(89)80087-8.
Vella. RE: Pilloa NJ: Zarrouki. B: Croze. ML: Koppe. L: Guichardant. M: Pesenti. S: Chauvin. MA: Rieusset. J:
Geloea 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: Gotzsche. 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: Naa B: Morishita. M: Mukheriee. B: Dvonch. JT: Spino. C: Fink. GD:
Raiagopalan. S: Sua 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.
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.
Weakley. J: Webber. MP: Ye. F: Zeig-Owens. R: Cohea 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.
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: Oiaa ZM: Li. S: Faa S: Chea G: Svberg. KM: Xiaa H: Wang. SO: Ma. H: Chea 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-ll-52.
Ying. Z: Allea K: Zhong. J: Chea M: Williams. KM: Wagner. JG: Lewandowski. R: Sun. O: Raiagopalaa 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.
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.
September 2019
5-67
DRAFT: Do Not Cite or Quote
-------
Zhong. J: Allen. K: Rao. X: Ying. Z: Braunstein. Z: Kankanala. SR: Xia. C: Wang. X: Bramble. LA: Wagner.
JG: Lewandowski. R: Sun. O: 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/10.1080/08958378.2Q16.1179373.
September 2019
5-68
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
APPENDIX 6
HEALTH EFFECTS-
MORTALITY
Stimniary of Causality Determinations for Short- and l.on^-lcrni Ozone
Exposure and Total (\onaccidcntal) Mortality
This Appendix eh;n;ielei"i/es ihe seienlifie e\ idenee Ih;il supports e;ius;ilii\
delei'iiiiii;ilKiiis lV»r short-;ind loim-iei'in o/one exposure ;md loi;il nioi"i;ihi\ The i\pes of
siudies e\ ;ilu;iled u ilhin lliis \ppench n ;ne eousisieul \x illi I he o\ ercill scope of llie IS \ ;is
decided in llie hcllicc In ;isscssiuu ihe o\cr;ill c\ idenee. sirenullis ;md liniiuiiioiis of mdi\ idu;il
siudies were e\;ilu;iled h;ised mi seienlil'ie a>iisideniliniis del;iiled in llie \mie\ I'm- \nnendi\ (¦
More del;nls on ihe e;ius;il Innnewiiik used li< ie;ieh iliese ei)iiehiskiiis ;ne inehided in llie
I'lejimhle In llie IS \ 11 S I llJ \. 2<> 15. 2u| ^;p
I'A|>hmiiv Dumliiiii
C ;ius;ilil\ Di'U'rmiiiiiliiin
Short-term exposure
Suggestive of.
but not sufficient to infer, a causal relationship
I .ong-lerm exposure
Suggestive of.
but not sufficient to 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
September 2019
6-1
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
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 morbidity evidence supporting cardiovascular mortality was more limited.
Although 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, there was inconsistency in results between experimental and epidemiologic
studies. Specifically, epidemiologic studies did not consistently demonstrate positive associations with
other apical 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 PM size fractions, but also 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.
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
September 2019
6-2
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
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 2013 Ozone ISA informed a series of
uncertainties and limitations, specifically: 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
concentration-response (C-R) relationship. Recent studies, however, support and in some cases further
address the uncertainties and limitations in the evidence base examined in those earlier studies. While the
evidence in this section will focus on epidemiologic studies, the overall conclusions will draw on the
morbidity evidence presented for different health endpoints across the 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 draft 2019
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 of the components of the following
PECOS tool:
• 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: Mortality
• Study Design: Epidemiologic studies consisting of case-crossover or time-series studies
September 2019
6-3
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
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
characterized 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.
Appendix 4 outlined 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 due to cardiovascular disease, particularly ischemic
heart disease and congestive heart failure. Collectively, the progression demonstrated in the available
evidence for respiratory morbidity supports potential biological pathways by which short-term ozone
exposures could result in mortality; the evidence, however, for cardiovascular morbidity is more limited
due to the inconsistency in results between experimental and epidemiologic studies.
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 conducted 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, many of the recent studies continued 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 of 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.. 2016). 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:
September 2019
6-4
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
• The 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 PM25 (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
due to larger uncertainty from interpolating ozone concentrations from the fewer monitors located
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).
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.
September 2019
6-5
DRAFT: Do Not Cite or Quote
-------
Study
Bell etal. (2004)a
Levyetal. (2005) a
Bell etal. (2005)a
Ito et al. (2005)a
Schwartz (2005)b
Bell etal. (2007)b
Bell and Dommici (2008)b
Katsouyanni et al. (2009)b
Katsouyanni et al. (2009)b
Katsouyanni etal. (2009)b,c
jMoolgavkar etal. (2013)
jVanos et al. (2013)
jPengetal. (2013)
jPengetal. (2013)
jPenget al. (2013)c
Bell etal. (2004)a
Levyetal. (2005)a
Bell etal. (2005)a
Ito et al. (2005)a
Schwartz (2005)b
Franklin and Schwartz (2008)b
Zanobetti and Schwartz (2008)b
Zanobetti and Schwartz (2008)b
Medina-Ramon and Schwarlz (2008)b
Katsouyanni et al. (2009)b
Katsouyanni et al. (2009)b
Katsouyanni etal. (2009)b,c
|Liu et al. (2016)
|Liu et al. (2016)
jVanos et al. (2014)
"Penget al. (2013)
"Penget al. (2013)
Penget al. (2013)c
Jhun etal. (2014)
Madrigano et al. (2015)d
jDi etal. (2017)e
Location
95 U.S. communities
U.S. and Non-U. S.
U.S. and Non-U. S.
U.S. and Non-U. S.
14 U.S. cities
98 U.S. communities
98 U.S. communities
APHENA-U.S.
APHENA-Canada
APHENA-Canada
NMMAPS (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.
14 U.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(10 S. U.S. cities)
NMMAPS (10 N. U.S. cities)
10 Canadian cities
APHENA-U.S.
APHENA-Canada
APHENA-Canada
NMMAPS (97U.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: f and red text = recent multicity 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. (2005), Ito et al. (2005) and Levy et al. (2005) used a range of lag days in the meta-analysis: Lag 0, 1, 2,
or average 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 evaluated in previous ozone
assessments.
September 2019
6-6
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
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
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 et al. (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 etal. (2011) included 7.5 more years of data than in Klemm and Mason (2000) and
Klemm et al. (2004). and reported evidence of a positive association with cardiovascular morality
(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]), contradicting the consistent primarily
positive associations reported in multicity studies (Figure 6-2).
September 2019
6-7
DRAFT: Do Not Cite or Quote
-------
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: f and red text = recent multicity 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 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. Bell et al. (2005) used a range of lag days in the meta-analysis: Lag 0, 1, 2, or average 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.
September 2019
6-8
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
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 percent 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
September 2019
6-9
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
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 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.
• 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, case-crossover and nested case-control, each of which
addresses different questions. 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 methods 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 LUR methods (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:
September 2019
6-10
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
• 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) in
Montreal, Canada across each of the pre-existing cardiovascular disease outcomes examined.
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. 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 evidence of a synergistic effect between ozone and temperature at higher ozone
concentrations, 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
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.
September 2019
6-11
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
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
average daily temperature modified the ozone-mortality relationship depending on where the
temperature fell along the distribution of mean temperatures across the study duration. In three
separate analyses, where low and high temperatures were defined as the 25th and 75th percentile,
10th and 90th percentile, and 5th and 95th percentile, the authors reported evidence of a U-shaped
curve. Across these analyses, the ozone-mortality risk was highest at low temperatures when
using the 25th/75th percentile cutoff, but highest when high temperatures were defined as above
the 90th and 95th percentiles. However, the higher ozone-mortality risk estimates at high
temperatures were found to be attenuated when examining risks across the distribution of AC
prevalence, specifically above the 75th percentile.
• While previous studies focused on whether ozone-mortality risk estimates were modified by
long-term average temperature or 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 temperature, the authors reported evidence of a
synergistic effect of temperature on mortality risk at higher ozone concentrations (Figure 6-4).
When comparing results across the three different models examined, additive linear, additive
nonlinear, and the monotone spatial risk surface model, the percent increase (for an increase from
the median of ozone concentrations and temperature to the 95th percentile) in the national
estimate was found to vary by 3.06, 3.54, and 3.98%, 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 interaction in northern U.S. cities where there is a larger difference between
temperatures on high temperature and moderate temperature days.
September 2019
6-12
DRAFT: Do Not Cite or Quote
-------
Spring
1 02 -
51.01 H
ca
11.00 -\*-
a
x 0 99 -
0 98 H
"T"
10
15
~~T~
20
Average temperature ("C)
Summer
1,02
1.01 -
-£
B
o;
>
I 1-00 i
or
0.99 H
—i 1 1 1
22 24 25 28
Average temperature (XT)
Note: Dot size represents the central estimate and 95% confidence intervals for each individual city, while the solid line and dotted
lines represent the central estimate and 95% confidence intervals from the metaregression.
Source: Permission pending, adapted from Liu et al. (2016).
Figure 6-3 Results of a metaregression analysis in Liu et al. (20161 indicating
larger ozone-mortality risk estimates in cities with lower average
temperatures in the spring and summer seasons.
September 2019
6-13
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
99th pctl, 90F
0.4
O
o 0.3
N
o
0.2-
o
ir
cc
o 0.1 -
0.0
95th pctl, 85F
75th pctl. 78F
50th pctl, 70F
100 200
Max 1-hour ozone, ppb
300
Source: Permission pending Wilson et al. (2014).
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).
Similar to Wilson et al. (2014). Chen et al. (2018) examined a bivariate response surface of ozone
and temperature to capture the joint effect on daily mortality in 86 U.S. cities from NMMAPS.
The authors observed that 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 temperatures within
each city. Overall, the authors reported evidence of modification of the ozone-mortality
association at high 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.
September 2019
6-14
DRAFT: Do Not Cite or Quote
-------
0
ts
3 g
O JZ
Is 1
™ .£
"c 0
'a ^
if) 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: Permission pending from Chen et al. (2018).
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.
6.1.5.5 Weather Patterns
1 While the majority of studies to date focus on whether season or temperature modify the
2 ozone-mortality relationship, a series of recent studies (Vanos et al.. 2015; Vanos et al.. 2014; Vanos et
3 al.. 2013) conducted in multiple cities in Canada examined the role of specific weather patterns
4 (i.e., synoptic weather categories). The weather categories examined included dry moderate (DM), dry
5 polar (DP), dry tropical (DT), moist moderate (MM), moist polar (MP), moist tropical (MT), and a
6 transitional category (TR), representing the shift from one weather category to another. Although each of
7 the aforementioned studies conducted analyses that differed by lag days, mortality outcome, and years
September 2019
6-15
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
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:
• Using more recent air quality data, Di et al. (2017a) reported that ozone-mortality risk estimates
were relatively unchanged 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 PM10. 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
September 2019
6-16
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
concentrations; Peng et al. (2013). lag 1: ozone = 0.89% (95% CI: 0.00, 1.73),
ozone + PMio = 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 percentile, 25th-75th percentile,
>75th percentile). In copollutant models with PMio 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 PMio and ozone data during the study duration (lag 0:
ozone = 5.14% [95% CI: 1.57, 8.85], ozone + PMio = 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 PMio (ozone = 2.78% [95%
CI: 1.38, 4.14], ozone + PMio = 2.38% [95% CI: -0.88, 6.02]), 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:
• 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.
September 2019
6-17
DRAFT: Do Not Cite or Quote
-------
1 • The summer season results of Liu et al. (2016) are consistent with the warm season analysis
2 conducted by Madrigano et al. (2015) in the 12-county analysis using observed ozone
3 concentrations. Associations of similar magnitude and precision were observed when using either
4 4 or 7 df/year.
5 • While the aforementioned studies focused on assessing the control for temporal/seasonal trends,
6 Di et al. (2017a) examined whether the appropriate df were instituted to control for
7 meteorological factors in the statistical model. The authors did not report any evidence that the
8 magnitude of ozone-mortality risk estimates changed when increasing the df for meteorological
9 variables (i.e., temperature and dew point temperature) from 6 to 9.
10 While the studies detailed above focused primarily on examining how changing the df for
11 temporal/seasonal trends or temperature influenced ozone-mortality associations, additional studies
12 conducted systematic evaluations of the relationship between alternative model specifications and
13 ozone-mortality risk estimates and provided new information on the potential for residual confounding:
14 • Sacks et al. (2012) examined whether similar results were observed across the different statistical
15 models used in multicity studies using a common data set. The authors observed variability in the
16 ozone-cardiovascular mortality relationship corresponding to differing levels of adjustment for
17 temperature. Specifically, those statistical models that more thoroughly controlled for
18 temperature, such as by including multiple temperature terms or a term for apparent temperature,
19 were found to have larger risk estimates (1.3 to 2.2% for a 20-ppb increase in 8-hour max ozone
20 concentrations) compared with those models that included only one temperature term for a single
21 lag day (-1.6 to 0.5%).
22 • In examining the ozone-mortality relationship at different temperatures, Chen et al. (2018)
23 employed multiple methods to explore whether hot or cold temperatures confounded
24 temperature-stratified ozone-mortality associations. There was evidence of residual confounding
25 and the overestimation of ozone-mortality risk estimates, specifically at high temperatures, which
26 was attributed to not adequately controlling for heat effects.
6.1.7 Shape of the Concentration-Response (C-R) Relationship
27 Studies included in the 2013 Ozone ISA conducted a variety of statistical analyses to characterize
28 the shape of the concentration-response (C-R) relationship between short-term ozone exposure and
29 mortality and did not observe any evidence of a threshold or deviations from linearity within the range of
30 ozone concentrations observed within the U.S. However, it is important to note that the examination of
31 the ozone-mortality C-R relationship is complicated by previously identified city-to-city and regional
32 heterogeneity in ozone-mortality risk estimates (U.S. EPA. 2013a). Recent studies continue to provide
33 evidence of a linear C-R relationship with no evidence of a threshold below which mortality effects do
34 not occur along the distribution of ozone concentrations observed within the U.S. as described below:
35 • Moolgavkar et al. (2013) reported evidence of a linear relationship down to concentrations of
36 60 ppb, with less certainty in the shape of the curve below 60 ppb when examining lag 1, 24-hour
37 avg ozone concentrations in a flexible model using 6 df (Figure 6-6).
September 2019
6-18
DRAFT: Do Not Cite or Quote
-------
106
III
0.01
0.02
/
-0.02
-r~
20
-1-
10
-r~
EC
Lag-1 0,
Source: Permission pending, Moolaavkar et al. (2013).
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 Moolaavkar 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).
September 2019
6-19
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
-1-1 1 1 1 1 r
0 20 40 60 80 100
Ozone, ppb
Source: Permission pending, Pi et al. (2017a).
Figure 6-7 Percent 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). The limited assessment of cause-specific mortality by Vanos et al.
(2014) is consistent with the pattern of positive associations reported for studies evaluated in the 2013
Ozone ISA (Figure 6-2). However, the experimental evidence, specifically from controlled human
September 2019
6-20
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
exposure studies, is not consistent with the studies evaluated in the 2013 Ozone ISA. This contributes
additional uncertainty for a biologically plausible mechanism by which short-term ozone exposure could
lead to cardiovascular mortality. Lastly, most of the recent studies examined associations between
short-term ozone exposure and mortality using ozone data prior to the year 2000, with only Di et al.
(2017a) focusing on more recent ozone concentrations.
Recent studies continue to assess the influence of 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 used to
control for each confounder. 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 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.
The 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
increased risk in individuals with CHF, and more limited evidence for other cardiovascular-related
diseases, including acute coronary artery disease, hypertension, and cerebrovascular disease (Buteau 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
September 2019
6-21
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
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
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 higher 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)). 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 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 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 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 limited
experimental evidence in combination with the lack of coherence between experimental and
epidemiologic studies of cardiovascular morbidity that do not demonstrate consistent evidence of
ozone-induced cardiovascular effects complicates the evidence for a biological pathway of events leading
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,
September 2019
6-22
DRAFT: Do Not Cite or Quote
-------
there remains relatively limited biological plausibility and coherence within and across disciplines to
support the relatively strong evidence for cardiovascular 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 Effects0
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 etal. (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
Epidemiologic
evidence from
copollutant models
provides some support
for an independent
ozone association
The few recent multicity
studies that examined
potential copollutant
confounding provide
evidence supporting that
ozone-mortality risk
estimates are relatively
unchanged or slightly
attenuated, but remain
positive, in copollutant
models with PM2.5, PM-io, and
NO2.
Studies that reported
correlations between ozone
and PM2.5 or PM10 were
generally low (<0.40).
Moolqavkar et al. (2013)
Peng etal. (2013)
Chen etal. (2018)
Pi etal. (2017a)
Section 6.1.5.1
Epidemiologic
evidence continues to
support a linear C-R
relationship with no
evidence of a threshold
Studies continue to provide
evidence of a linear C-R
relationship with no evidence
of a threshold. There is less
certainty in the shape of the
C-R relationship at the lower
end of concentrations
observed in the U.S.
Moolqavkar et al. (2013)
Pi etal. (2017a)
Peng etal. (2013)
Section 6.1.5.3
24-h avg
>40 ppb
8-h max
>30 ppb
September 2019
6-23
DRAFT: Do Not Cite or Quote
-------
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
Rationale for
Concentrations
Causality
Associated
Determination3
Key Evidence13
Key References'3 with Effects0
Limited biological
Animal toxicological and
Appendix 4
plausibility from studies
controlled human exposure
of cardiovascular
studies do not provide
morbidity
consistent evidence of
potential biological
pathways. Additionally, there
is a lack of coherence with
epidemiologic studies of
cardiovascular morbidity to
support more overt effects,
such as 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.
Uncertainty regarding
Recent studies indicate
Section 6.1.3
geographic
latitude and temperature
Section 6.1.5.4
heterogeneity in
may account for some of the
ozone-mortality
observed heterogeneity, but
associations
more extensive evaluations
have not been conducted.
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.
September 2019
6-24
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
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
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 will be 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 will be emphasized. Another important consideration will be
characterizing the shape of the C-R relationship across the full concentration range observed in
epidemiologic studies. The evidence in this section will focus on epidemiologic studies because
experimental studies of long-term exposure and mortality are generally not conducted. However, this
section will draw 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.
September 2019
6-25
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
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. 2017V This evidence is
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.
September 2019
6-26
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
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 the 8-hour daily max ozone
metric, though there are instances in some that use the 24-hour avg [e.g., Sese et al. (2017)1. or the 1-hour
daily max Ic.g.. Jerrett et al. (2009)1. 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 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 (Zanobetti 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 (Di 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.. 2017;
Weichenthal et al.. 2017; Cakmak et al.. 2016; Crouse et al.. 2015).
• 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).
September 2019
6-27
DRAFT: Do Not Cite or Quote
-------
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—
0.6
!•
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: fStudies 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.
September 2019
6-28
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
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 (Jcrrctt 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 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.
September 2019
6-29
DRAFT: Do Not Cite or Quote
-------
Reference
Cohort
Jerrett et al. 2009 ACS
tTurner etal. 2016 ACS
tJerrett et al. 2013 ACS
tCrouse etal. 2015 CanCHEC
tWeicherrthal 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 etal. 2016 ACS
tTurner etal. 2016 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: fStudies 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
1 Recent cohort studies extend the body of evidence for the relationship between long-term ozone
2 exposure and cardiovascular mortality. The 2013 Ozone ISA noted inconsistent evidence for
3 cardiopulmonary mortality, and there was limited evidence for the association between long-term ozone
4 exposure and cardiovascular mortality based on an analysis of the ACS cohort (Jerrett et al.. 2009).
5 Recent analyses from the ACS cohort in the U.S. and the CanCHEC cohort in Canada provide consistent
6 evidence for positive associations between long-term ozone exposure and cardiovascular and IHD
7 mortality, as well as mortality due to diabetes or cardiometabolic diseases. Associations with mortality
8 due to cerebrovascular disease (e.g., stroke) were less consistent, and generally closer to the null value.
9 Other recent studies conducted in the Europe and Asia report null or negative associations. Similar to total
September 2019
6-30
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
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 etal.. 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.. 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 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..
2015).
September 2019
6-31
DRAFT: Do Not Cite or Quote
-------
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: fStudies 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
1 A recent study adds to the body of evidence on the relationship between long-term ozone
2 exposure and mortality by examining temporal trends in ozone concentrations and changes in life
3 expectancy, testing the hypothesis that populations living in areas with higher ozone concentrations have
4 lower life expectancies. Li et al. (2016) reported the mean life expectancy for males and females in the
5 U.S. from 2002 to 2008 at the county level, separating counties into three classes based on average ozone
6 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
September 2019
6-32
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
(45.7-54.5). Nationwide, ozone concentrations reduced an average 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
Class 2 and Class 3 counties (respectively, compared to counties in Class 1) and by 0.3 and 0.6 years for
females in Class 2 and Class 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. Specifically, the evidence from controlled human exposure studies provided
support for larger decrements in FEVi and greater inflammatory responses to ozone in individuals with
asthma than in healthy individuals without a history of asthma. 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
September 2019
6-33
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
statistically significant associations for each of the disease cohorts (Zanobctti and Schwartz.
2011) (Table 6-6). 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.
• 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 (Vicira 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
September 2019
6-34
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
effects associated with long-term ozone exposure are independent of coexposure to correlated
copollutants in long-term analyses:
• 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 (Figure 6-11).
September 2019
6-35
DRAFT: Do Not Cite or Quote
-------
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: fStudies 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
1 An important consideration in characterizing the ozone-mortality association is whether the
2 concentration-response (C-R) relationship is linear across the full concentration range that is encountered
3 or there are concentration ranges that depart from linearity. The epidemiologic studies included in the
4 2013 Ozone ISA indicated a "generally linear C-R function with no indication of a threshold" (U.S. EPA.
5 2013a). With regard to studies of long-term ozone exposure and mortality, a threshold analysis indicated
6 that the linear model was not a better fit to the data (p > 0.05) than a threshold representation of the
7 overall ozone-mortality association (Jerrett et al.. 2009); however, the authors reported "limited evidence"
8 for an effect threshold at an ozone concentration (seasonal avg of 1-hour max) of 56 ppb (p = 0.06).
September 2019
6-36
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
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 etal. (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.
September 2019
6-37
DRAFT: Do Not Cite or Quote
-------
40
45
Ozone (ppb)
40 45 50
Ozone (ppb)
Note: The solid line represents the estimate and the dotted lines represent the 95% confidence interval for the estimate.
Source: Pi et al. (2017b)—Permission Pending.
Figure 6-12 The concentration-response relationship estimated with log-linear
model with a thin-plate spline (A) and the concentration-response
relationship estimated with threshold model (B), indicating the
potential for a threshold at 40 ppb (8-hour daily max).
03
en
TO
N
CC
~3 - ppb
Note: The solid blue line represents the estimate and the gray shaded areas represent the 95% confidence interval for the estimate.
Source: Crouse et al. (2015)—Permission Pending.
Figure 6-13 Concentration-response relationship between ozone
concentrations (ppb) and total (nonaccidental) mortality in the
CanCHEC cohort (mean 39.6; knots: 30.0, 38.9, 50.7 ppb).
September 2019
6-38
DRAFT: Do Not Cite or Quote
-------
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. Grey 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.
Source: Turner et al. (2016)—Permission Pending.
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
1 This section describes the evaluation of evidence for total (nonaccidental) mortality based on the
2 scientific considerations detailed in the Annex for Appendix 6. with respect to the causality determination
3 for long-term exposures to ozone using the framework described in the Preamble to the ISAs (U.S. EPA.
4 2015). The key evidence, as it relates to the causal framework, is summarized in Table 6-2. Recent cohort
5 studies provide limited support for the association between long-term ozone exposure and total mortality,
6 with some U.S. and Canadian cohorts reporting modest positive associations between long-term ozone
7 exposure and total mortality, while other recent studies conducted in the U.S, Europe, and Asia report null
8 or negative associations. The strongest evidence for an association between long-term ozone exposure and
9 total mortality continues to come from analyses of the Medicare cohort data included in the 2013 Ozone
September 2019
6-39
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
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
PM25 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, controlled human
exposure studies, and epidemiology) and biological plausibility for ozone-related cardiovascular
(Appendix 4) and 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
September 2019
6-40
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
these studies do not help to explain the inconsistent associations observed across studies, although they
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.
September 2019
6-41
DRAFT: Do Not Cite or Quote
-------
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 (2011)
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 etal. (2016):
Jerrett et al. (2013)
Crouse et al. (2015): Cakmak et al.
(2016): Cakmak et al. (2017)
Section 6.2.3.2
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
September 2019
6-42
DRAFT: Do Not Cite or Quote
-------
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.
Ozone
Rationale for
Concentrations
Causality
Associated
Determination3
Key Evidence13
Key References'3 with Effects0
Epidemiologic
Positive associations
Section 6.2.4.1
evidence from
observed between long-term
Fiqure 6-9
copollutant models
ozone exposure and total
provides some support
mortality remain relatively
for an independent
unchanged after adjustment
ozone association
for PM2.5, and NO2.
When reported, correlations
with copollutants were highly
variable (low to high).
Limited epidemiologic
Some studies provide
Section 6.2.6
evidence supports a
evidence of a generally
linear C-R relationship;
linear C-R relationship;
some evidence for a
others observed a sublinear
sublinear C-R
relationship, indicating larger
relationship
changes in risk for higher
compared with lower ozone
concentrations.
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; PM25 = 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.
September 2019
6-43
DRAFT: Do Not Cite or Quote
-------
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.
Study
Study
Population
Exposure Assessment
Mean and
Upper
Percentiles
(Hg/m3)
Copollutant
Examination
Effect Estimates
Percent Increase (95% CI)
Bell et al. (2004)
95 U.S. cities
1987-2000
Times-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: 26.0
Correlation (r): PM10:
-0.38 to 0.63
Copollutant models
with: PM10
All-year (lag 0-6 DL)
0.78 (0.40, 1.16)
Warm/summer (lag 0-6 DL)
0.58 (0.19, 0.97)
Lew 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
with: NR
All-year
1.23 (0.94, 1.52)
Warm/summer
2.52 (1.70, 3.30)
Bell 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
with: NR
All-year
1.31 (0.82, 1.77)
Warm/summer
2.26 (1.09, 3.45)
September 2019
6-44
DRAFT: Do Not Cite or Quote
-------
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
(Hg/m3)
Copollutant
Examination
Effect Estimates
Percent Increase (95% CI)
Ito et al. (2005)
U.S. and
24-h avg
NR
Correlation (r): NR
All-year
U.S. and non-U.S.
non-U.S.
Copollutant models
1.65 (0.60, 2.69)
Meta-analysis
with: NR
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 each county
1-h max
Mean:
35.1-60.0
75th:
46.3-69.0
Correlation (r): NR
Copollutant models
with: 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
with: 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
with: PM2.5, PM10
All-year (lag 0-6)
0.78 (0.42, 1.16)
September 2019
6-45
DRAFT: Do Not Cite or Quote
-------
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
(Hg/m3)
Copollutant
Examination
Effect Estimates
Percent Increase (95% CI)
Katsouvanni et al. (2009)
APHENA
1987-1996
Time-series study
NMMAPS
12 Canadian
cities
All ages
Exposure assignment
approach detailed in
original studies and
based on all available
monitoring data
1-h max
Mean:
U.S.:
13.0-38.0
Canada:
6.7-8.4
75th:
U.S.:
21.0-52.0
Canada:
8.7-12.5
Correlation (r): NR
Copollutant models
with: 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
(2008)
18 U.S. communities
Time-series study
All ages
Average of all monitors
in each county based on
the method detailed in
Schwartz (2000)
24-h avg
Mean:
21.4-48.7
Correlation (r): PM2.5:
0.43; SO42": 0.34;
OC: 0.50; NOs": 0.24
Copollutant models
with: PM2.5, SO42",
oc, no3-
Warm/summer (lag 0)
1.34 (0.68, 2.00)
Zanobetti and Schwartz
All ages
Average of all monitors
Mean
Correlation (r): NR
Warm/summer (lag 0)
(2008a)
in each city
(across
Copollutant models
1.00 (0.76, 1.24)
48 U.S. cities
8-h avg
seasons):
with: NR
Case-crossover study
16.5-47.8
Maximum
(across
seasons):
40.6-103.0
Zanobetti and Schwartz
All ages
Average of all monitors
Mean:
Correlation (r): NR
Warm/summer (lag 0-3)
(2008b)
in each city
15.1-62.8
Copollutant models
1.06 (0.56, 1.55)
48 U.S. Cities
8-h avg
75th:
with: NR
Time-series study
19.8-75.4
Maximum:
34.3-146.2
Medina-Ramon and
All ages
Average of all monitors
Median:
Correlation (r): NR
Warm/summer (lag 0-2)
Schwartz (2008)
in each county based on
16.1-58.8
Copollutant models
1.30 (0.76, 1.87)
48 U.S. cities
the method detailed in
with: NR
Case-only study
Schwartz (2000)
8-h avg
September 2019
6-46
DRAFT: Do Not Cite or Quote
-------
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
(Hg/m3)
Copollutant
Examination
Effect Estimates
Percent Increase (95% CI)
tKlemm et al. (20111
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
75th: 47.82 Copollutant models
Maximum:
109.07
Lag 0-1: 1.36 (-0.50, 3.25)
with: NR
tMoolaavkar et al.
(2013)
98 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
with: PM10
100 dftemporal trends (lag 1): 0.60 (0.44,
0.80)
100 dftemporal trends with PM10 (lag 1):
0.33 (-0.07, 0.72)
tGoldberg etal. (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
with: 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)
September 2019
6-47
DRAFT: Do Not Cite or Quote
-------
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
(Hg/m3)
Copollutant
Examination
Effect Estimates
Percent Increase (95% CI)
tPenaetal. (2013)
50 U.S. cities
12 Canadian cities
1987-1996
Time-series study
APHENA
All ages
Average of all monitors
in each city
1-h max
Mean: NR
Median:
6.6-19.4
Correlation (r): NR
Copollutant models
with: PM10
50 U.S. cities
All-year (lag 0-2 DL): 2.13 (0.54, 3.73)
All-year with PM10 (lag 1): 0.64 (-0.88,
2.18)
Summer (lag 0-2 DL): 3.23 (1.63, 4.85)
12 Canadian cities
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)
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
with: 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
with: 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)
September 2019
6-48
DRAFT: Do Not Cite or Quote
-------
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
(Hg/m3)
Copollutant
Examination
Effect Estimates
Percent Increase (95% CI)
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
with: 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)
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
with: 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 et al. (2016)
20 U.S. cities
(10 northern;
10 southern);
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
8-h max
Mean: 39.7
75th: 41.2
Maximum:
44.7
Correlation (r): NR
Copollutant models
with: 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)
September 2019
6-49
DRAFT: Do Not Cite or Quote
-------
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
(Hg/m3)
Copollutant
Examination
Effect Estimates
Percent Increase (95% CI)
tDi etal. (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
with: 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)
tButeau etal. (2018)
Montreal, Canada
1991-2002
Case-crossover study
Case-control study
n =63,534
65+ with CHF
Nearest monitoring
station
8-h avg
Mean: 28.9
Median:
27.3
75th: 37.5
95th: 57.5
Maximum:
108.8
Correlation (r): NR
Copollutant models
with: NR
Case-crossover
Nearest monitoring station (0-3 DLNM):
-2.24 (-9.38, 5.31)
BME (0-3 DLNM): -5.12 (-16.61, 7.88)
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)
tWilson etal. (2014)
NMMAPS
Average across all NR
Correlation (r): NR
April-October (lag 0)
95 U.S. cities
All ages
monitors in each city;
Copollutant models
Additive linear
1987-2000
10% trimmed mean to
correct for yearly
with: NR
National: 3.06 (SE = 0.30)
Time-series study
averages of each
Additive nonlinear
monitor
National: 3.54 (SE = 0.75)
1-h max
Surface
National: 3.98 (SE = 0.24)
September 2019
6-50
DRAFT: Do Not Cite or Quote
-------
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
(Hg/m3)
Copollutant
Examination
Effect Estimates
Percent Increase (95% CI)
tMadriaano et al. (2015)
91 northeast U.S.
counties
1988-1999
Time-series study
New York:
62 counties
New Jersey:
21 counties
Connecticut:
8 counties
All ages
Analysis of the average
across all monitors in
each county for
12 counties; kriging for
all 91 counties analysis
8-h max
Mean:
12 counties
(observed
data): 45.6
91 counties
(kriging
data): 45.7
23 urban
counties:
45.6
68
nonurban
counties:
45.7
Maximum:
133.5-149
Correlation (r): NR
Copollutant models
with: 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. counties
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
with: 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.
Note: f = 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.
September 2019
6-51
DRAFT: Do Not Cite or Quote
-------
Table 6-4 Epidemiologic studies of short-term exposure to ozone and cardiovascular mortality.
Mean and
Upper
Percentiles Copollutant Effect Estimates
Study Study Population Exposure Assessment (|jg/m3) Examination Percent 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 with: 1-67 (1.02,2.30)
NR
Katsouvanni et al.
APHENA
1987-1996
Time-series study
(2009) 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 with:
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 Mean:
each city 15.1-62.8
8-h avg 75th:
19.8-75.4
Maximum:
34.3-146.2
Correlation (r): NR Summer (lag 0-3)
Copollutant models with: (0.96, 2.27)
NR
September 2019
6-52
DRAFT: Do Not Cite or Quote
-------
Table 6 4 (Continued): Epidemiologic studies of short-term exposure to ozone and cardiovascular mortality.
Study
Study Population Exposure Assessment
Mean and
Upper
Percentiles
(Hg/m3)
Copollutant
Examination
Effect Estimates
Percent Increase (95% CI)
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 with:
Maximum:
109.07
NR
tSacks et al. (2012)
Philadelphia, PA, U.S.
5/12/1992-9/30/1995
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): PM2.5:
0.43; NO2: 0.18; SO2:
-0.19; Other: CO: -0.35
Copollutant models with:
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 et al. (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 with:
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.
Note: f = U.S. and Canadian studies published since the 2013 Ozone ISA.
September 2019
6-53
DRAFT: Do Not Cite or Quote
-------
Table 6-5 Epidemiologic studies of short-term exposure to ozone and respiratory mortality.
Study
Study Population Exposure Assessment
Mean and
Upper
Percentiles
(Hg/m3)
Copollutant
Examination
Effect Estimates
Percent Increase (95% CI)
Bell et al. (2005)
Meta-analysis 24-h avg
NR
Correlation (r): NR
All-year:0.70 (-0.77, 2.21)
Meta-analysis
Copollutant models with:
NR
Katsouvanni et al.
APHENA
1987-1996
Time-series study
(2009)
NMMAPS
12 Canadian cities
All ages
Exposure assignment
approach detailed in
original studies and based
on all available monitoring
data
1-h max
Mean:
U.S.:
13.-38.4
Canada:
6.7-8.4
75th:
U.S.:
21.0-52.0
Canada:
8.7-12.5
Correlation (r): NR
Copollutant models with:
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,
Canada: -2.91
Summer (lag 0
All
U.S.: 2.73 (-1.32,
Canada: 15.59 (8.
>75 yr
U.S.: 2.52 (-2.67,
Canada: 11.79 (1.
5.66)
12.56, 8.09)
df/yr-NS):
2; 8
6.90)
09, 24.08)
7.94)
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 with:
NR
Summer (lag 0-3):
1.67 (0.76, 2.58)
September 2019
6-54
DRAFT: Do Not Cite or Quote
-------
Table 6 5 (Continued): Epidemiologic studies of short-term exposure to ozone and respiratory mortality.
Study
Study Population Exposure Assessment
Mean and
Upper
Percentiles
(Hg/m3)
Copollutant
Examination
Effect Estimates
Percent Increase (95% CI)
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 with:
Maximum:
109.07
NR
tVanos et al. (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 with:
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.
Note: f = U.S. and Canadian studies published since the 2013 Ozone ISA.
September 2019
6-55
DRAFT: Do Not Cite or Quote
-------
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
Mean
Exposure Assessment (pg/m3)
Copollutant
Examination
Effect Estimates
HR (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
with: PM2.5
Total mortality (96 MSAs): 1.00 (1.00, 1.01)
Total mortality (86 MSAs): 1.00 (1.00, 1.01)
Total mortality (86 MSAs + PM2.5): 0.99 (0.98, 1.00)
Zanobetti and Schwartz
Medicare
Citywide 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
with: 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.
(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 Total mortality: 1.1 (0.99, 1.21)
Copollutant models Total mortality + PM10: 1.09 (0.99, 1.21)
with: PM10
September 2019
6-56
DRAFT: Do Not Cite or Quote
-------
Table 6-6 (Continued): Epidemiologic studies of long-term exposure to ozone and total (nonaccidental) mortality.
Study
Study Population Exposure Assessment
Mean
(Hg/m3)
Copollutant
Examination
Effect Estimates
HR (95% CI)
tCarev et al. (2013)
English Medical
Annual mean estimates
Mean: 25.85
Correlation (r): Total mortality: 0.76 (0.62, 0.87)
Nationwide, U.K.
Practice
from dispersion model
Maximum:
PM2.5: -0.39; NO2:
Ozone: 2002
n = 835,607
for 1-km grid cells linked
to nearest residential
31.5
-0.46; SO2: -0.41;
PM10: -0.40
Follow-up: 2003-2007
Adults, 40-89 yr,
from English
medical practices
postal code centroid
Copollutant models
Cohort study
with: NR
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 4 monitors
within 50 km of
residence
Mean: 50.35 Correlation (r):
Median: 50.8 PM2.5: 0.56; NO2:
-0.0071;
75th: 61
90th: 68.56
95th: 74.18
Maximum:
89.33
Copollutant models
with: 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
with: 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 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
with: NR
Total mortality: 1.03 (1.03, 1.04)
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
with: PM2.5
Total mortality (year-round): 1.02 (1.01, 1.04)
Total mortality (year-round; + PM2.5): 1.02(1.01,
1.04)
September 2019
6-57
DRAFT: Do Not Cite or Quote
-------
Table 6-6 (Continued): Epidemiologic studies of long-term exposure to ozone and total (nonaccidental) mortality.
Study
Study Population Exposure Assessment
Mean
(Hg/m3)
Copollutant
Examination
Effect Estimates
HR (95% CI)
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; NO2:
Ozone: 1988-2011
with newly
from up to 4 monitors
within 50 km of
residence
-0.01; PM10: 0.36
Follow-up: 1988-2011
diagnosed cancer
Copollutant models
with: NR
Cohort study
8-h max
tDi etal. (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
with: PM2.5
Total mortality (main analysis): 1.01 (1.01, 1.01)
Total mortality (single pollutant): 1.02(1.02, 1.02)
Total mortality (low 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
with: NR
Total mortality: 1.06 (1.05, 1.07)
tCakmak et al. (2017)
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
with: PM2.5
Total mortality: 1.08 (1.02, 1.14)
Total mortality (+ PM=): 1.05 (0.99, 1.11)
September 2019
6-58
DRAFT: Do Not Cite or Quote
-------
Table 6-6 (Continued): Epidemiologic studies of long-term exposure to ozone and total (nonaccidental) mortality.
Study
Study Population Exposure Assessment
Mean
(Hg/m3)
Copollutant
Examination
Effect Estimates
HR (95% CI)
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
with: NR
Total mortality: 0.78 (0.75, 0.82)
tSese etal. (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
with: NR
September 2019
6-59
DRAFT: Do Not Cite or Quote
-------
Table 6-7 Epidemiologic studies of long-term exposure to ozone and cardiovascular mortality.
Study
Study Population Exposure Assessment Mean (pg/m3)
Copollutant
Examination
Effect Estimates
HR (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
with: 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-Hwana 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)
with: PM10
tCarev et al. (2013)
English medical
Annual mean estimates
Mean: 25.85
Correlation (r): Circulatory mortality: 0.76 (0.66, 0.87)
Nationwide, U.K.
practice
from dispersion model
Maximum:
PM2.5: -0.39; NO2:
Ozone: 2002
n = 835,607
for 1-km grid cells linked
to nearest residential
31.5
-0.46; SO2: -0.41;
PM10: -0.40
Follow-up: 2003-2007
Adults, 40-89 yr,
from English
medical practices
postal code centroid
Copollutant models
Cohort study
with: NR
September 2019
6-60
DRAFT: Do Not Cite or Quote
-------
Table 6-7 (Continued): Epidemiologic studies of long-term exposure to ozone and cardiovascular mortality.
Study
Study Population Exposure Assessment Mean (pg/m3
Copollutant
Examination
Effect Estimates
HR (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 4 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
with: 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
with: NR
CVD mortality: 0.83 (0.39, 1.75)
tCrouse et al. (2015)
CanCHEC
Model of warm season
Mean: 39.6
Correlation (r):
CVD: 1.04 (1.03, 1.05)
Nationwide, Canada
n =2,521,525
concentration at 21-km
Median: 39
PM2.5: 0.73; NO2:
Cardiometabolic: 1.05 (1.04, 1.06)
Ozone: 2002-2009
25+ yr
horizontal resolution
0.19;
assigned at postal code
75th: 44.2
Copollutant models
with: NR
IHD: 1.09 (1.08, 1.11)
Follow-up: 1991-2006
8-h max
Maximum: 60
CBVD: 0.98 (0.96, 1.00)
Cohort study
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
with: 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)
September 2019
6-61
DRAFT: Do Not Cite or Quote
-------
Table 6-7 (Continued): Epidemiologic studies of long-term exposure to ozone and cardiovascular mortality.
Study
Study Population Exposure Assessment Mean (pg/m3
Copollutant
Examination
Effect Estimates
HR (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
with: 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.
CanCHEC
Model of warm season
Mean: 38.29
Correlation (r): NR CVD mortality: 1.16 (1.14, 1.18)
(2017)
n =2,448,500
concentration at 21-km
Median: 38.11
Copollutant models
Nationwide, Canada
25+ yr
horizontal resolution
75th: 42.63
95th: 50.51
with: NR
Ozone: 2002-2009
assigned at postal code
8-h max
Follow-up: 1991-2011
Maximum:
Cohort study
60.46
tCakmak etal. (2017)
Nationwide, Canada
Ozone: 2002-2009
Follow-up: 1991-2011
Cohort study
CanCHEC
n =2,291,250
25+ yr
Model of warm season
concentration at21-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
with: 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
with: NR
Cardiovascular mortality: 0.72 (0.64, 0.81)
September 2019
6-62
DRAFT: Do Not Cite or Quote
-------
Table 6-8 Epidemiologic studies of long-term exposure to ozone and respiratory mortality.
Study
Study Population
Exposure
Assessment
Mean (pg/m3
Copollutant
Examination
Effect Estimates
HR (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
Resp mortality (96 MSAs): 1.03 (1.01, 1.05)
Resp mortality (86 MSAs): 1.03 (1.01, 1.05)
Copollutant models ReSp mortality (86 MSAs + PM2 5): 1.04(1.01,
with: PM2.5 -i 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
with: 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 4 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
with: 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)
September 2019
6-63
DRAFT: Do Not Cite or Quote
-------
Table 6-8 (Continued): Epidemiologic studies of long-term exposure to ozone and respiratory mortality.
Study
Study Population
Exposure
Assessment
Mean (pg/m3
Copollutant
Examination
Effect Estimates
HR (95% CI)
tBentaveb et al. (2015) Gazel
Nationwide, France
Ozone: 1989-2008
Follow-up: 1989-2013
Cohort study
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
with: 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; NO2:
COPD: 0.97 (0.95, 1.00)
horizontal resolution
0.19;
Ozone: 2002-2009
25+ yr
assigned at postal
75th: 44.2
Follow-up: 1991-2006
Copollutant models
with: NR
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; 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
with: 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
with: NR
Respiratory mortality: 1.04 (1.01, 1.07)
September 2019
6-64
DRAFT: Do Not Cite or Quote
-------
Table 6-9 Epidemiologic studies of long-term exposure to ozone and other mortality.
Study
Study Population
Exposure
Assessment
Averaging Time Mean (pg/m3)
Copollutant
Examination
Effect Estimates
HR (95% CI)
tXu etal. (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
with: 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
with: 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
tLi etal. (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
with: 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 etal. (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
with: NR
In-hospital mortality (15 high ozone cities): 1.11
Follow-up: 2011
for ARDS
(1.08, 1.15)
Cohort study
September 2019
6-65
DRAFT: Do Not Cite or Quote
-------
Table 6 9 (Continued: Epidemiologic studies of long-term exposure to ozone and other mortality.
Study
Study Population
Exposure
Assessment
Averaging Time Mean (pg/m3)
Copollutant
Examination
Effect Estimates
HR (95% CI)
tVieira etal. (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
95th:
265-727
Correlation (r): NR NA
Copollutant models
with: NR
September 2019
6-66
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
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. 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. 2013b'). ToxRTool
(Klimisch et al.. 19971. STROBE guidelines (von Elm et al.. 20071. and ARRIVE guidelines (Kilkenny et al.. 20101.
September 2019
DRAFT: Do Not Cite or Quote
6-67
-------
Table Anne 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 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.
September 2019
6-68
DRAFT: Do Not Cite or Quote
-------
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 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.b 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.b 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.
September 2019
6-69
DRAFT: Do Not Cite or Quote
-------
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 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.
September 2019
6-70
DRAFT: Do Not Cite or Quote
-------
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 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.b 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,c 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.
September 2019
6-71
DRAFT: Do Not Cite or Quote
-------
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, 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.
September 2019
6-72
DRAFT: Do Not Cite or Quote
-------
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 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.
a(U.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).
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).
September 2019
6-73
DRAFT: Do Not Cite or Quote
-------
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 inmultisite 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: 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.
Burnev. PG: Laitinen. LA: Perdrizet. S: Huckauf. H: Tattersfield. AE: China S: Poissoa 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. (2017). Associations
between long-term PM2.5 and ozone exposure and mortality in the Canadian Census Health and
Environment Cohort (CANCHEC), by spatial synoptic classification zone. Environ Int 111: 200-211.
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.
Carey. IM: Atkinsoa RW: Kent. AJ: van Staa. T: Cook. PG: Andersoa 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/10.1164/rccm.201210-1758QC.
Chea K: Wolf. K: Hampel. R: Stafoggia. M: Breitner. S: Cvrvs. J: Samoli. E: Andersen. ZJ: Bero-Bedada. G:
Bellander. T: Hennig. F: Jacauemia B: Pekkanea J: Peters. A: Schneider. A. (2018). Poes 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 Ponkelaar. A: Martia 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.
September 2019
6-74
PRAFT: Po Not Cite or Quote
-------
Pi. O: 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.2Q17.17923.
Pi. O: 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: Peng. H: Lurmana FW: Liu. L: Gilliland. FP. (2016). Air pollution affects
lung cancer survival. Thorax 71: 891-898. http://dx.doi.org/10.1136/thoraxinl-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. PM: Brophv. JM: Paskalopoulou. 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: McTiernaa A: Gaziano. JM: Frishmaa WH:
Curb. JP. (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: Pe Leoa 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.0000165821.90114.7f.
Jerrett. M: Burnett. RT: Beckermaa BS: Turner. MC: Krewski. P: Thurston. G: Martin. RV: van Ponkelaar. A:
Hughes. E: Shi. Y: Gapstur. SM: Thua 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/10.1164/rccm.2013Q3-
0609QC.
Jerrett. M: Burnett. RT: Pope. CA. Ill: Ito. K: Thurstoa G: Krewski. P: Shi. Y: Calle. E: Thua M. (2009).
Long-term ozone exposure and mortality. N Engl J Med 360: 1085-1095.
http://dx.doi.org/10.1056/NEJMoa0803894.
Jhun. I: Fana 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.
Katsouvanni. K: Samet. JM: Andersoa HR: Atkinson. R: Le Tertre. A: Medina. S: Samoli. E: Touloumi. G:
Burnett. RT: Krewski. P: Ramsay. T: Pominici. F: Peng. RP: 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. IC: Emersoa M: Altman. PG. (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. H: Kim. J: Kim. S: Kang. SH: Kim. H.T: Kim. H: Heo. J: Yi. SM: Kim. K: Youa 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. 117.007170.
Klemm. RJ: Lipfert. FW: Wvzga. RE: Gust. C. (2004). Paily 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: Masoa 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: Tillmana 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.
September 2019
6-75
PRAFT: Po Not Cite or Quote
-------
Lew. Jl: Chcmcrvnski. 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: Wea X: Hao. Y: Oualters. 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: Lia H: Rutherford. S: Xiao. J: Li. X: Li. Z: Oiaa 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-Ramoa 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. 1104507.
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/SQ895-
4356(97)00019-X.
Murgia. N: Brisman. J: Claessoa A: Muzi. G: Olin. AC: Torea 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:
Cohea A: Atkinson. RW: Andersoa 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/sll869-012-018Q-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: Wilsoa 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/10.1164/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: Juvia 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/10.1136/thoraxinl-2017-209967.
September 2019
6-76
DRAFT: Do Not Cite or Quote
-------
Spencer-Hwang. R: Knutsea SF: Soret. S: Ghamsarv. M: Beesoa 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: Beckermaa 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/10.1164/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-95/001F). 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] (pp. 1-166). (EPA/630/P-03/001F). 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://cft)ub.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://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/b5d8alce9b07293485257375007Q12b7/eele280e77586de
985257b65005d37e7!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.
September 2019
6-77
DRAFT: Do Not Cite or Quote
-------
Vanos. JK: Cakmak. S: Bristow. C: Brion. V: Tremblav. N: Martin. SL: Sheridaa 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.
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: Gotzsche. 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.
Weakley. J: Webber. MP: Ye. F: Zeig-Owens. R: Cohea 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.
Wilsoa 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-ll-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/10.1164/rccm.200706-823QC.
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.
September 2019
6-78
DRAFT: Do Not Cite or Quote
-------
APPENDIX 7 HEALTH EFFECTS —OTHER
HEALTH ENDPOINTS
Stimmiiry of Ciitisnlity Determinations/or Other Health Effects
This \ppendi\ eh;ir;ieleri/es ihe seieiiiilie e\ idenee lh;ii supports c;ius;ilii\
delerini n;il mils I'nr slum- ;md loim-lemi n/nne exposure ;iud he;illh el'leels. iiicludinu
keprodueli\e ;md l)e\elopnieul;il I Tl'eels (see Seelion ~ 11. \erums S\ siem I Tl'eels (see
Seelion ~2 i. ;md ( ;ineer (see Seel ion ~ ' i The l> pes of siudies e\ ;ilu;iled willmi llns \ppendix
;ire eousisieul Willi l lie n\ cnill scope ol'lhc IS \ ;is delniled in I lie hchicc In ;isscssiuu ihe
o\ ercill c\ idcucc. ihe siiviiullis ;ind IiiiiiI;iIK)iis of indi\ idii;il studies were c\ ;ilu;ilcd h;ised mi
scicul i lie coiisidenilious dcl;iilcd in I lie \nne\ lor \nncndi.x ~ Mure del;iils mi 1 he c;ius;ilil\
I'nnicuork used In re;ieh these conclusions ;ire inehided mi 1 he hc;inihlc In lhe IS \ 11 S. I T \.
2i 1151
lk';illh 1'. I'l'i'ii
( ;ms;ilil\ IK'k'riniiiiiliiin
Sliiirl-krm i-\|)iisuiv
Nervous system effects
Suggestive of but not sufficient to infer, a causal relationship
l.iui^-k'nii i'\|)nsuiv
Reproductive and developmental
effects
Male and female reproduction
Suggestive of. but not sufficient to infer a causal relationship
and fertility
Pregnancy and birth outcomes
Suggestive of, but not sufficient to infer, a causal relationship
]¦ fleets of exposure durina
Evidence is summarized in Section 7.1.4 but contributes to
developmental periods
the causalitv determinations for relevant oman sv stems
(i.e.. Respiratorv-Appendix 3, Cardiovascular-Appendix 4.
Metabolic-Appendix 5. and Nervous Svstem—Section 7.1.4)
Nervous system effects
Suggestive of. but not sufficient to infer, a causal relationship
Cancer
Inadequate evidence to determine if a causal relationship
exists
September 2019
7-1
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
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 effects of exposure during developmental periods
(referred to as "developmental effects") are summarized in Section 7.1.4. but are fully evaluated and
contribute to causality determinations in the ISA section for the relevant organ system (i.e., respiratory
[see Appendix 31. cardiovascular [see Appendix 41. metabolic [see Appendix 51. and nervous system
effects [see Section 7.21). Many studies have been added to the body of literature since the 2013 Ozone
ISA (U.S. EPA. 2013a) including epidemiologic studies and short- and long-term animal toxicological
and developmental studies. Because the average length of gestation in rodents is 18-24 days, animal
toxicological studies investigating the effects of ozone generally are considered short-term exposure
periods. For comparison, an epidemiologic study that uses the entire pregnancy as the exposure period is
considered to have a long-term exposure period (about 40 weeks, on average). Results from both
short- and long-term exposure periods are included in a single section (Section 7.1) and are identified
accordingly in the text and tables throughout this section. Well-designed studies that consider sources of
bias, including potential confounding by copollutant exposures, are emphasized.
A major 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 not completely understood. Thus, multiple exposure
periods are evaluated in many epidemiologic studies, including long-term (months to years) 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 ozone exposure at these critical windows of development may have
permanent, lifelong effects.
The 2013 Ozone ISA (U.S. EPA. 2013a) 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, 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
September 2019
7-2
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
exposure on central nervous system effects, including laterality, brain morphology, neurobehavioral
abnormalities, and sleep aberration. Epidemiologic studies examining the effects of ozone on sperm
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 separate causality
determinations are made for the male and female fertility and reproduction (see Section 7.1.2). and
pregnancy and birth outcomes (see Section 7.1.3). as they are likely to have different etiologies and
critical exposure windows over different lifestages. For effects of exposure during developmental periods
see Section 7.1.4; however, summaries are included in this section of the ISA, while 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 draft 2019
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::
• 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
• 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
September 2019
7-3
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
determinations in Appendix 3. Appendix 4. and Appendix 5. and Section 7.2.2.5 and are
summarized in Section 7.1.4.')
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, the studies evaluated are
less limited in scope and not targeted towards specific study locations, as reflected in the PECOS tool.
The epidemiologic 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 (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 et al.. 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 to 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-IV 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.
September 2019
7-4
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
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. 2013a) and by new evidence. Once these pathways are initiated, there is
evidence from experimental and epidemiologic studies that ozone inhalation may result 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.
Respiratory tract
Inflammation/
Altered Testicular
Pathology with
Ozone Exposure
Spermatogenesis
and Sperm Quality
Decreased
Reduced Sperm
Quality in Humans
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
(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.
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
(Jedlinska-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
September 2019 7-5 DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
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 most of pregnancy (Gestation Days 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 limited epidemiologic evidence from few studies for
an association between ozone and sperm quality, with associations between reductions in sperm
concentration and both short- and long-term ozone exposures. Since then, additional evidence is limited
to: 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.
September 2019
7-6
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
7.1.2.2.3 Summary
Overall, there is evidence of impaired spermatogenesis and decreased sperm count and
concentration from epidemiologic studies, and decremental effects on testicular morphology and impaired
spermatogenesis from toxicological studies with ozone exposures.
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 (Legro 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 30 day periods
before ovulation. Another study found no evidence of association with exposure up to 2 months before
conception, but did show an improvement in fecundity with ozone exposure post-conception, likely
indicating unmeasured confounding (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) 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 food 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) showed that ozone acted
September 2019
7-7
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
synergistically with sodium salicylate to 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 appeared 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 with ozone exposures, while toxicological studies show limited 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 observed in epidemiologic studies. This
discussion of "how" exposure to ozone may lead to reproductive and developmental health effects
contributes to an understanding of the biological plausibility of epidemiologic results evaluated in
Section 7.1.3.2 through Section 7.1.3.5.
September 2019
7-8
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Ozone
Exposure
J-
Activation of Sensory
Nerves in
Respiratory Tract
Respiratory
Inflammation and
Oxidative Stress
¦JMIiili
Elevated sFit-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 Fetal Growth
and/or Birth Weight
Preterm birth
Altered Litter Size
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.
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. 2013a) and recent evidence indicates multiple initial
events after ozone inhalation contribute to effects on pregnancy and birth outcomes, including systemic
inflammation or oxidative stress. 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 that provide biological plausibility for the associations reported in epidemiologic
studies and animal toxicological studies that contribute to the apical endpoint 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). The initial event of altered systemic oxidative
stress is demonstrated in the epidemiologic literature with ozone-dependent increased odds of elevated
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 reproductive organ-specific
September 2019 7-9 DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
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
significantly altered in ozone-exposed animals. Evidence from the 2013 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 etal.. 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.
2013a). 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
lst-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 (Wu et al.. 2011).
• Studies of "hypertensive disorders of pregnancy" generally reported positive associations (Hu et
al.. 2016; Michikawa et al.. 2015; Mobasher et al.. 2013).
September 2019
7-10
DRAFT: Do Not Cite or Quote
-------
1 • Two studies examining hypertension reported mixed associations with 1st trimester exposure,
2 with Lee et al. (2013) showing positive associations and Xu et al. (2014) showing no evidence of
3 association.
4 • Increased odds of gestational diabetes were observed for higher ozone exposures during the 1st
5 and 2nd trimesters in a Florida population compared to lower ozone exposures (Hu et al.. 2015)
6 and for weekly exposures during the 2nd trimester in a national study (Robledo et al.. 2015).
7 • No evidence of association with gestational diabetes was observed in the national study for ozone
8 exposures 90 days before conception and in the 1st trimester (Robledo et al.. 2015).
9 • The single study of hypertensive disorders of pregnancy examined the potential for copollutant
10 confounding and showed an odds ratio increase (from 1.05 to 1.11) with adjustment for NO2
11 (Olsson et al.. 2013). In both studies of gestational diabetes, adjustment for copollutants did not
12 change effect estimates (Hu et al.. 2015; Robledo et al.. 2015). reducing uncertainties that the
13 associations observed with ozone are due to copollutant confounding.
14 Data from current studies of maternal health during pregnancy are extracted and summarized in
15 the evidence inventories (see Table 7-8 and Table 7-9).
7.1.3.2.2 Toxicological Evidence of Effects on Pregnancy
16 Studies from the 2013 Ozone ISA demonstrated a transient anorexiogenic effect of ozone on
17 pregnant dam weight gain during pregnancy. Initially, dam body weight (0.8 and 1.2 ppm ozone), water
18 consumption (0.4, 0.8, and 1.2 ppm ozone), and food consumption (0.4, 0.8, and 1.2 ppm ozone) during
19 pregnancy were decreased with ozone exposure but these deficits dissipated a week or two after the initial
20 exposure (Bignami et al.. 1994). The anorexigenic effect of ozone exposure on the pregnant dam appears
21 to dissipate with time; the dams seem to adapt to the ozone exposure. Studies from the 2013 Ozone ISA
22 also demonstrated enhanced pulmonary inflammatory response in BALF of pregnant and lactating rodents
23 to ozone exposure (1.0 ppm, 6 hours); there was significantly enhanced sensitivity to ozone-induced
24 pulmonary inflammation during pregnancy, which was maintained during lactation, and disappeared after
25 lactation ceased at weaning (Gunnison et al.. 1992). Research since the 2013 Ozone ISA also shows that
26 ozone affects weight gain during pregnancy. Pregnant rats exposed to ozone (0.8 ppm ozone) during the
27 period of implantation (GDs 5-6) showed significantly lower body-weight gain during this period (Miller
28 et al.. 2017). demonstrating a similar anorexigenic effect as documented in the 2013 Ozone ISA.
29 Exposure to 0.4 ppm ozone during implantation did not affect dam body weight gain during pregnancy.
30 Miller et al. (2017) also assessed dam blood pressure (GD 15, GD 19, GD 21) and kidney histopathology
31 in near-term ozone exposed dams to evaluate whether ozone exposure might contribute to gestational
32 hypertension/preeclampsia, with data showing null findings. Peri-implantation ozone exposure (1.2 ppm,
33 GD 5) caused increased homeostatic model assessment for insulin resistance (HOMA-IR) and increased
34 area under the curve with the glucose tolerance test in dams immediately after ozone exposure; exposure
35 to 0.4 or 0.8 ppm ozone did not induce these metabolic changes in the dam (Miller et al.. 2019). Data
36 from current studies of maternal health during pregnancy are extracted and summarized in the evidence
37 inventories (see Section 7.6.1. Table 7-7.)
September 2019
7-11
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
7.1.3.2.3 Summary
Evidence for effects on maternal health during pregnancy is mixed, with epidemiologic studies
showing limited 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 intra-uterine growth and
gestational age, among others, so studies of these outcomes will often restrict to term births. Vulnerability
to 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.
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 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
concentrations were associated with decrements in birth weight, although the magnitude of the
decrement varied, ranging from -4.61 to -27.27 (per 10 ppb increase in ozone).1
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.
September 2019
7-12
DRAFT: Do Not Cite or Quote
-------
1 • One study using geographically weighted regression indicated variation by spatial characteristics,
2 with lower birth weight associated with higher ozone concentration in less urbanized
3 communities (Tu et al.. 2016V
4 • Some studies of odds of low birth weight (<2,500 g), including well designed studies [e.g., Chen
5 et al. (2017b); Laurent et al. (2016a); Vinikoor-Imler et al. (2014); Laurent et al. (2013)1. reported
6 increased odds of low birth weight with increased ozone concentrations; however, those
7 associations are inconsistent across exposure windows.
8 • In studies with copollutant adjusted models, effect estimates were largely similar to those
9 reported for single-pollutant models for ozone (Smith et al.. 2017; Ha et al.. 2014; Olsson et al..
10 2013).
7.1.3.3.2 Toxicological Evidence for Fetal Growth, Birth Weight, and Body Length at Birth
11 Evidence from the 2013 Ozone ISA showed decreased birth weight in pups whose pregnant dams
12 were exposed to ozone during pregnancy (Sharkhuu et al.. 2011; Haro and Paz. 1993). but no effects on
13 the number of pups born. A few studies reported that mice or rats exposed developmentally
14 (gestationally ± lactationally) to ozone had deficits in postnatal body-weight gain (Bignami et al.. 1994;
15 Haro and Paz. 1993; Kavlock et al.. 1980). Recent animal toxicological evidence also shows that ozone
16 exposure during pregnancy causes decreased fetal weight near term. In summary, animal toxicological
17 models show ozone exposure caused decreased birth weight and decreased postnatal body-weight gain
18 but did not affect litter number.
19 • There is recent evidence that fetuses whose dams were exposed to ozone (0.8 ppm for both sexes,
20 0.4 ppm ozone for male fetuses) during the period of implantation (GDs 5-6) weighed
21 significantly less than the air-exposed control pups at GD 21, near the end of pregnancy. There
22 exists a sexual dimorphism with male pups more sensitive to ozone exposure than females.
23 Further examination showed that dams exposed to 0.8 ppm ozone had male and female fetuses
24 with significantly lower lean mass and fat mass compared with control-air dams at GD 21 (Miller
25 et al.. 2017).
7.1.3.3.3 Summary
26 Overall, there is some epidemiologic evidence for the effects of ozone on fetal growth, especially
27 for continuous-term birth weight, a conclusion supported by toxicological evidence in rodents.
7.1.3.4 Preterm Birth
28 Preterm birth (PTB), delivery that occurs before 37 weeks of completed gestation, is a marker for
29 fetal underdevelopment and is related to subsequent adverse health outcomes (Saigal and Doyle. 2008;
30 IOM. 2007; MacDorman et al.. 2007; Gilbert et al.. 2003). PTB is characterized by multiple etiologies
September 2019
7-13
DRAFT: Do Not Cite or Quote
-------
1
2
(spontaneous, premature rupture of membranes [PROM], or medically induced), which may have either
individual or shared mechanistic pathways.
7.1.3.4.1 Epidemiologic Evidence of Preterm Birth
3 In the 2013 Ozone ISA, short-term exposure to ozone during late pregnancy was consistently not
4 associated with preterm birth. However, associations with long-term exposures were inconsistent across
5 studies, particularly across study locations. Since then, the number of studies examining ozone exposure
6 and preterm birth has doubled. All studies that examined ozone exposures during the 1st or 2nd trimesters
7 reported associations elevated from the null. Effects are more mixed with 3rd trimester and
8 entire-pregnancy exposure, with both positive and null associations present. As in the 2013 Ozone ISA,
9 studies of short-term, near-birth exposures generally reported no evidence of association. Data from
10 current studies of preterm birth are extracted and summarized in the evidence inventories (see
11 Table 7-1IV
12 • One study divided PTB into three categories and looked at 4-week intervals. The study authors
13 observed elevated odds ratios (ORs) for late and moderate PTB (but not severe/very PTB;
14 20-28 weeks) with exposures during Gestation Weeks 9-12; for 2nd trimester exposures, they
15 observed elevated ORs across preterm birth groups for ozone exposures during Gestation Weeks
16 17-21. 21-24. and 25-28 (Svmanski et al.. 2016).
17 • A single study was conducted on PROM (including both preterm and term births) examining
18 exposures at 0 to 4 hours before delivery and across the entire pregnancy. The association with
19 entire pregnancy exposure was null, however, the associations with short-term, near-birth hourly
20 exposures were all elevated form the null [OR range 1.05 to 1.07; Wallace et al. (2016)1.
21 • Adjustment for copollutants generally moved effect estimates slightly away from the null (Haet
22 al.. 2014; Olsson et al.. 2013; Olsson et al.. 2012V
23 • There were no apparent differences in effect estimates based on study location or exposure
24 assessment method used for recent studies.
7.1.3.4.2 Summary
25 Overall, there is evidence of an association between ozone exposures during early to
26 midpregnancy with preterm birth in epidemiologic studies. However, there are no toxicologic studies
27 specific to preterm birth.
7.1.3.5 Birth Defects
28 Birth defects are structural and functional abnormalities that can cause physical and intellectual
29 disability and other health problems; they are a leading cause of infant mortality and developmental
30 disability in the U.S. Critical periods for birth defect development are generally known, reducing
September 2019
7-14
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
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 themselves a grouping of separate defects, 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
odds ratios with higher levels of ozone exposure for neurological defects [neural tube defects,
anencephaly, and spina bifida; Padula et 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
Earlier research found eyelid malformation following gestational and postnatal exposure to
0.2 ppm ozone (Veninga. 1967). No recent 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 in epidemiologic studies,
and there are few animal studies on birth defects.
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 first year of life. In the 2013 Ozone ISA studies of infant mortality provided no evidence
September 2019
7-15
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
for an association between ozone exposure and infant mortality. In the current review, studies are
primarily of stillbirth, with a U.S.-based study (Haetal.. 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. In the studies that examined copollutant models, effect
estimates were robust to copollutant inclusion. Data from current studies of fetal and infant mortality are
extracted and summarized in the evidence inventories (see Table 7-13).
7.1.3.6.1 Toxicological Evidence of Birth Defects
There are no toxicological studies of fetal and infant mortality.
7.1.3.6.2 Summary
Findings for ozone associated fetal and infant mortality are generally inconsistent across exposure
windows in epidemiologic studies, and there are no animal studies.
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 (Hcindcl et al.). 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 began nearly 30 years ago with
Barker's hypothesis (Barker and Osmond. 1986) that 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 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.
September 2019
7-16
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
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 (see Appendix 3. Section 3.2.4.1) 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. 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 fully 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
September 2019
7-17
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
artery 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 et al.. 2012). A 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
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. All available evidence
examining the relationship between exposure to ozone and reproductive effects was evaluated using the
framework described in the Preamble to the IS As (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.
The 2013 Ozone ISA (U.S. EPA. 2013a) 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 (U.S. EPA. 2013a) came from
studies of sperm quality and on continuous birth weight. Current evidence continues to support
conclusions for studies of sperm quality and continuous birth weight. There is also new supporting
evidence for effects on preterm birth with exposures to ozone, particularly in the first and
second trimesters.
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 determination is supported by evidence
across epidemiologic studies, including those from the 2013 Ozone ISA (U.S. EPA. 2013a) of 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.
September 2019
7-18
DRAFT: Do Not Cite or Quote
-------
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
Limited but consistent
evidence for male
Limited evidence for decrements on
sperm count and concentration
Farhat et al. (2016)
-42 ppb
reproduction, effects on
sperm
Sokol et al. (2006)
-21.68 ppb
Hansen et al. (2010)
-30.8 ppb
Limited evidence for changes to
testicular morphology and
spermatogenesis
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
Farhat et al. (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 smaller 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. 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 the full body of evidence
is described.
°Describes the ozone concentrations with which the evidence is substantiated.
1 Overall the evidence is suggestive of, but not sufficient to infer, a causal relationship
2 between ozone exposure and pregnancy and birth outcomes. The key evidence as it relates to the
3 causal framework is summarized in Table 7-2. There are several well-designed, well-conducted studies
4 that indicate an association between ozone and poorer birth outcomes, particularly for outcomes of
5 continuous birth weight and preterm birth. In particular, studies of preterm birth that examine exposures
6 in the first and second trimesters show fairly consistent positive associations (increased ozone exposures
7 associated with increased odds of preterm birth). In addition, some animal toxicological studies
8 demonstrate decreased birth weight and changes in uterine blood flow. Studies of continuous birth weight
9 and preterm birth did not generally adjust for potential copollutant confounding, although studies that did
September 2019
7-19
DRAFT: Do Not Cite or Quote
-------
appeared to show limited impacts. There is also inconsistency across exposure windows for associations
with continuous birth weight, and the magnitude of effect estimates varies.
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
Evidence from multiple
epidemiologic studies of
continuous birth weight
but uncertainties remain
Positive associations from many
studies, but variability in timing of
exposures and magnitude of effects,
and limited assessment of copollutant
confounding contribute to uncertainty
Section 7.4.1
Mean
concentrations
across studies:
4-43 ppb
Evidence from multiple
epidemiologic studies
and preterm birth but
uncertainties remain
Positive associations from many
studies that examine exposure
windows in the first and second
trimesters, but magnitude of effects
differ across studies. Copollutant
adjustment generally not changing
observed effect estimates
Section 7.4.1
Mean
concentrations
across studies:
16-51 ppb
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. (2011)
Miller et al. (2017)
Limited assessment of
copollutant confounding
Few studies adjust for potential
confounding by NO2 and PM2.5
Section 7.4.1
Lack of information on
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.
September 2019
7-20
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
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 ozone
exposure (i.e., on the order of minutes to weeks) to ozone on the nervous system. The 2013 Ozone ISA
(U.S. EPA. 2013a) 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. 2013aV
September 2019
7-21
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
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 draft 2019
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:
• 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
Because the 2013 Ozone ISA concluded that evidence existed to suggest a causal relationship
between short-term ozone exposure and nervous system effects, the studies evaluated are less limited in
scope and not targeted towards specific study locations, as reflected in the PECOS tool. The
epidemiologic studies were evaluated and subsequently discussed using the PECOS tool below:
• 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 observed in
epidemiologic studies (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 epidemiologic results
September 2019
7-22
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
evaluated later. The biological plausibility for ozone-induced effects on the nervous system is supported
by evidence from the 2013 Ozone ISA and by recent evidence.
Modulation of the
Autonom ic Nervous
System
Short-T erm
Exposure
Neuroinflammation:
Whole Brain
Olfactory Bulb
Cerebral Cortex
Cerebellum
Hippocampus
Altered
Neurotransm itter Levels
(serotonin, dopamine,
acetylcholine)
Sleep Disturbances
Symptoms
Mental Health
Hospitalizations
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
(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.
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
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.
September 2019
7-23
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
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 beta-amyloid proteins, a strong predictor of Alzheimer's disease in
humans, in aged mice after a short-term exposure (Tyler etal.. 2018); 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 (Vcrhein 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 (Figure 7-3; solid lines) (Snow et al.. 2018; Kodavanti. 2016). 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.
September 2019
7-24
DRAFT: Do Not Cite or Quote
-------
7.2.1.3
Brain Inflammation and Morphology
7.2.1.3.1 Toxicological Studies
1 In the 2013 Ozone ISA, short-term ozone exposure resulted in increases in markers of oxidative
2 stress and inflammatory responses. These effects were observed in many regions of the brain, including
3 the olfactory bulbs, striatum, cortex, substantia nigra, and cerebellum and were associated with changes in
4 neuronal morphology, increased apoptosis, and decreased numbers of dopaminergic neurons in the
5 substantia nigra.
6 Recent studies (see Table 7-23) support the results summarized in the 2013 Ozone ISA, showing
7 increases in inflammatory responses and markers of oxidative stress in various regions of the brain. Most
8 studies evaluated a single concentration of ozone with exposure durations ranging from hours (single
9 exposure) to >15 days depending on the study. In studies with multiple time points, the magnitude or
10 severity of effects generally increased with exposure duration. Several studies evaluated both short- and
11 long-term exposures. Cellular markers of oxidative stress were generally seen at the earlier
12 (i.e., short-term) time points; effects on apoptosis/cell counts were primarily observed at the later time
13 points (i.e., long term).
14 • Increased brain inflammation and oxidative stress was commonly reported following short-term
15 ozone exposure in rodents (Tyler et al.. 2018; Mumaw et al.. 2016; Mokoena et al.. 2015; Rivas-
16 Arancibia et al.. 2015; Gomez-Crisostomo et al.. 2014; Gonzalez-Guevara et al.. 2014; Pinto-
17 Almazan et al.. 2014; Rodriguez-Martinez et al.. 2013; Mokoena et al.. 2011).
18 • Inflammation and oxidative stress were associated with increased mitochondrial damage (Rivas-
19 Arancibia et al.. 2015; Gomez-Crisostomo et al.. 2014; Rodriguez-Martinez et al.. 2013).
20 • Most of these data were generated in adult male rats (Mokoena et al.. 2015; Rivas-Arancibia et
21 al.. 2015; Gomez-Crisostomo et al.. 2014; Gonzalez-Guevara et al.. 2014; Pinto-Almazan et al..
22 2014; Rodriguez-Martinez et al.. 2013; Mokoena et al.. 2011). although Mumaw et al. (2016) did
23 report similar effects in both male and female CD-/- mice (pulmonary immune function
24 impaired).
25 • Some evidence suggests that aged populations may be more susceptible to ozone-induced
26 inflammation in the brain. One study evaluated the effects of ozone in both adult and aged mice,
27 and although there was a clear main effect of ozone exposure, inflammatory outcomes were more
28 pronounced in the aged animals (Tyler et al.. 2018).
29 • Brain inflammation and oxidative stress were largely observed in the hippocampus (Tyler etal..
30 2018; Mokoena et al.. 2015; Gomez-Crisostomo et al.. 2014; Pinto-Almazan et al.. 2014;
31 Rodriguez-Martinez et al.. 2013) and cerebral cortex (Tyler etal.. 2018; Mumaw et al.. 2016;
32 Mokoena et al.. 2015; Gonzalez-Guevara et al.. 2014; Mokoena et al.. 2011). with more limited
33 data for other regions of the brain (Tyler et al.. 2018; Rivas-Arancibia et al.. 2015).
34 • There is some evidence in rats to suggest that ozone exposure may affect glial morphology and
35 blood-brain barrier permeability. Changes in glial morphology in the nucleus tractus solitarius
36 were reported following a 24-hour continuous ozone exposure, with treated animals showing
37 increased gial wrapping of synapses. The overall increase in glial coverage was driven by a
September 2019
7-25
DRAFT: Do Not Cite or Quote
-------
1 decrease in the proportion of synapses with no glial coverage. There were no changes in
2 expression of proteins associated with astrocyte activation (Chounlamountrv et al.. 2015). In
3 contrast, adult and aged mice exposed to ozone showed effects on blood brain barrier
4 permeability, resulting in increased infiltration of circulatory inflammatory cells and structural
5 changes in the microglia. Notably, this effect was only statistically significant in aged animals.
6 These effects were observed in the cortex, dentate gyrus, hippocampus, and hypothalamus: brain
7 regions that are known to have increased permeability of the blood-brain barrier or high
8 sensitivity of the cells to toxic insult (Tyler etal.. 2018).
9 • The effect of ozone exposure on (3-amyloid accumulation and structure was investigated in
10 several studies. Tyler et al. (2018) found that short-term ozone exposure increased (3-amyloid
11 formation in aged mice, but several other studies reported no effect in adult rats at the 7 or 15 day
12 time points (Rivas-Arancibia et al.. 2017; Fernando Hernandez-Zimbron and Rivas-Arancibia.
13 2016; Hernandez-Zimbron and Rivas-Arancibia. 2015). (3-Amyloid accumulation is strongly
14 associated with Alzheimer's disease in humans; therefore, these data are discussed further in the
15 long-term exposure section (see Section 7.2.2).
7.2.1.4 Cognitive and Behavioral Effects
7.2.1.4.1 Epidemiologic Studies
16 No epidemiologic studies of short-term ozone exposure and its effects on cognitive and
17 behavioral effects were reviewed in the 2013 Ozone ISA. In a recent study, Lim et al. (2012) examined
18 older adults in South Korea during a 3-year, follow-up study using the Korean Geriatric Depression
19 Scale-Short Form (SGDS-K). An increase in SGDS-K score, indicating increased depressive symptoms,
20 largely driven by emotional symptoms, was associated with 3-day moving avg ozone concentration (see
21 Table 7-18).
7.2.1.4.2 Toxicological Studies
22 In the 2013 Ozone ISA, short-term exposure to ozone was associated with changes in behavior in
23 rodents, including decreased motor activity, impaired performance on learning and memory tasks, and
24 altered sleep-wake cycles. In general, these effects were more pronounced with increasing exposure
25 durations. Effects on sleep-wake cycles were associated with decreases in acetylcholine levels in the
26 medial preoptic area, a region of the brain that regulates sleep.
27 Several recent studies (see Table 7-24) have reported cognitive and behavioral changes in rodent
28 models following short-term exposure to ozone. Observed effects on cognition and behavior included
29 increases in depressive-like behaviors in a rodent model of depression and anxiety, decreases in
30 performance on learning and memory tasks, and declines in motor activity. Effects on neurotransmitter
31 levels were also reported.
September 2019
7-26
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
• One study reported increases in depression and anxiety behaviors in Flinders sensitive line (FSL)
rats, a rodent model of depression (Mokocna ct al.. 2015). 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 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, with exposed
animals spending 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
(Mokoena etal.. 2015).
• Three studies evaluated motor activity following short-term ozone exposure. Of these, two
reported statistically significant decreases in motor activity associated with ozone treatment
(Gordon et al.. 2016; Pinto-Almazan et al.. 2014). 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
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
Section 7.2.1 describes and characterizes the epidemiologic and toxicological evidence relating to
the effect of short-term ozone exposure on cognition and behavior. There are no epidemiologic studies of
cognition or motor-function-related effects. 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). In addition, 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). Some of the behavioral effects in
animals are supported by data showing effects on neurotransmitter levels that are associated with these
outcomes.
September 2019
7-27
DRAFT: Do Not Cite or Quote
-------
1 Biological plausibility for the short-term effect of ozone on the nervous system is derived from
2 multiple studies demonstrating that short-term exposure to ozone can to lead to inflammation and
3 oxidative-stress responses in the brain, as well as modulation of the neuroendocrine system.
4 • Overall, the available evidence pertaining to cognitive and behavioral effects is limited. Increased
5 depressive symptoms were observed in humans and in animals, providing some coherence across
6 scientific disciplines.
7.2.1.5 Neuroendocrine Effects
7.2.1.5.1 Toxicological Evidence
7 In the 2013 Ozone ISA, two studies provided evidence that ozone alters neuroendocrine function,
8 affecting levels of thyroid hormones and corticosterone following short-term exposure. Since then,
9 several studies have been published investigating the potential effects of ozone on the HPA axis;
10 however, most of the data examine outcomes related to metabolic function and are therefore discussed in
11 detail in Appendix 5.
12 A recent study (see Table 7-25) evaluated potential neuroendocrine effects of ozone in the
13 nervous system following a short-term ozone exposure in rats (Thomson et al.. 2013). A 4-hour exposure
14 induced a transient effect on a wide array of genes involved in antioxidant response, xenobiotic
15 metabolism, inflammation, and endothelial dysfunction. The pattern of gene responses was largely
16 consistent across several organs, including the brain and pituitary, supporting systemic effects of
17 neuroendocrine changes. Notably, the effects observed in the present study were transient, largely
18 disappearing by 24 hour post-exposure; however, chronic exposure could result in prolonged
19 neuroendocrine modulation. As described previously (see Section 7.2.1.2). ozone likely modulates HPA
20 axis function by activating the sensory nerves in the lung and thereby altering autonomic nervous system
21 activity.
7.2.1.6 Hospital Admissions and Emergency Department Visits
22 There were no studies of hospital admissions, emergency department (ED), or outpatient visits for
23 diseases of the nervous system in the 2012 Ozone ISA. Recent studies (see Table 7-19) examining the
24 association of short-term ozone exposure with hospital admissions, ED visits, or outpatient visits for
25 diseases of the nervous system or mental health are presented in Figure 7-4. Outcomes that are presented
26 on the plot and included in this section generally include hospitalizations for International Classification
27 of Diseases version 9 (ICD-9) codes 290-319 or 320-359 and version 10 (ICD-10) codes F1-F99 or
28 G00-G99. Several of the studies shown in Figure 7-4 are stratified by season, reporting separate
29 associations for the warm and cold seasons.
September 2019
7-28
DRAFT: Do Not Cite or Quote
-------
1 • Some positive associations with hospitalizations for migraine, dementia, and multiple sclerosis
2 were observed in single studies. Several studies also reported associations of short-term ozone
3 exposure with mental health hospital admissions or ED visits for conditions such as depression
4 and panic attack, but the results were not entirely consistent (Figure 7-4).
5 • Because hospitalizations or ED visits among those with chronic diseases may be related to
6 comorbid conditions, the extent to which these studies are informative regarding the effect of
7 short-term ozone exposure on nervous system health is uncertain.
Study
Cohort
Outcome
Lag
Mean
f Jeanjean et al. 2018
Strasbourg, France
HA: multiple sclerosis relapse
0-3
44.3
f Guo et al. 2018 *
Guangzhou, China
ED visit: neurologic disease
0-2
49.8
t Chiu al. 2015
Taipei, Taiwan
Outpatient visit: migrane
0-2
24.6
f Szyszkowicz et al. 2016
9 cities, Canada
ED visit: depression
0, Male
22.5-29.2
ED visit: depression
0, Female
22.5-29.2
f Oudin et al. 2018
Gothenberg, Sweden
ED visit: pychiatric emergencies
0
29.5
t Jeanjean et al. 2018
Strasbourg, France
HA: multiple sclerosis relapse
0-3
19.0
f Guo et al. 2018 *
Guangzhou, China
ED visit: neurologic disease
0-2
49.8
t Chiu al. 2015
Taipei, Taiwan
Outpatient visit: migrane
0-2
24.6
t Oudin etal. 2018
Gothenberg, Sweden
ED visit: pychiatric emergencies
0
21.1
f Linares et al. 2017
Madrid, Spain
HA: dementia related
5 d
18.2
f Xu etal. 2016
Xi'an, China
Outpatient visit: epilepsy
0
51.0
t Chen etal. 2018*
Shanghai, China
HA: mental disorder
0-1
51.0
f Cho et al. 2015
Seoul, South Korea
ED visit: panic attack
0-3
18.0
Warm Season
i
'•<>•
i
i
4
i
i
1 _o
i
i
i
i
>~
i
i
%¦
i
i
i
i
i
i
Cold Season
All Year
0.6 0.8 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.
fStudies 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.
September 2019
7-29
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
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
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 (Jeaniean 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. 2013a) 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. Current evidence continues to support conclusions for related endpoints, including brain
inflammation and changes in brain morphology, oxidative stress, and neurotransmitter levels. 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.
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 etal.. 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
September 2019
7-30
DRAFT: Do Not Cite or Quote
-------
1 showing depression-like behaviors in rodents (Mokoena et al.. 2015). Biological plausibility of these
2 effects is supported by multiple toxicological studies showing inflammation and morphological changes
3 in the brain following short-term ozone exposure (see Section 7.2.1.2). As noted in Section 7.2.1.6. these
4 hospital admission and ED visit studies provide limited information regarding the effect of short-term
5 ozone exposures on the nervous system because the extent to which people are treated for comorbid
6 conditions may not be discernable.
7 Overall, the evidence is suggestive of, but not sufficient to infer, a causal relationship
8 between short-term exposure to ozone and nervous system effects. This conclusion remains based
9 largely on multiple toxicological studies demonstrating the effect of short-term exposure to ozone on the
10 brain.
September 2019
7-31
DRAFT: Do Not Cite or Quote
-------
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.
Rationale for Causality
Determination3
Key Evidence13
Key References'3
Ozone
Concentrations
Associated with
Effects0
Limited epidemiologic
evidence
An increase in depressive
symptoms reported in single study
Relevance of studies of hospital
admissions, ED visits, and
outpatient visits to nervous system
effects is uncertain
~ met al. (2012)
Section 7.2.1.6
Mean: 48.1 ppb
Coherence with experimental
animal study
Study of FSL rats demonstrates
enhanced depressive-like
symptoms
Mokoena et al. (2015)
Section 7.2.1.4
0.3 ppm
Single toxicological studies
demonstrate effects on motor
activity and cognition
0.25-0.8 ppm
Multiple toxicological studies
generally support effects on
the brain and provide
biological plausibility
Multiple studies show brain
inflammation and morphological
changes following short-term
ozone exposure
Section 7.2.1.3
0.25-2 ppm
Epidemiologic evidence from
copollutant models lacking
Evaluation of copollutant
confounding limited to studies of
hospital admissions, ED visits, and
outpatient visits, which are subject
to limitations
Section 7.2.1.7
C-R = concentration-response; N02 = nitrogen dioxide; ppb = parts per billion; 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.
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.
September 2019
7-32
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
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. 2013a) 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 1.2.23): effects on cognition, motor activity, and mood (Section 7.2.2.4); and
neurodevelopmental effects (Section 7.2.2.5). 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. The experimental studies evaluated and subsequently
discussed within this section were identified using the PECOS statements below:
• Population: Study population from 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)
September 2019
7-33
DRAFT: Do Not Cite or Quote
-------
1 • Comparison: Appropriate comparison group exposed to a negative control (i.e., clean air or
2 filtered-air control)
3 • Outcome: Nervous system effects
4 • Study Design: In vivo chronic, subchronic, or repeated-dose toxicity studies in mammals
5 • Because the 2013 Ozone ISA concluded that there was evidence to suggest a causal relationship
6 between long-term ozone exposure and nervous system effects, the studies evaluated are less
7 limited in scope and not targeted towards specific study locations, as reflected in the PECOS
8 statement. The epidemiologic studies evaluated and discussed within this section were identified
9 using the following PECOS statement:
10 • Population: Any population, including populations or lifestages that might be at increased risk
11 • Exposure: Long-term ambient concentration of ozone
12 • Comparison: Per unit increase (in ppb)
13 • Outcome: Change in risk (incidence/prevalence) of a nervous system effect
14 • Study Design: Epidemiologic studies consisting of cohort and case-control studies, time-series,
15 case-crossover, and cross-sectional studies with appropriate timing of exposure for the health
16 endpoint of interest
7.2.2.2 Biological Plausibility
17 This section describes biological pathways that potentially underlie nervous system effects
18 resulting from long-term and developmental exposure to ozone. Studies that include exposure during the
19 perinatal period are discussed in the long-term exposure section, regardless of the duration of the
20 exposure because of the sensitivity of this lifestage to nervous system effects and potential for long-term
21 health impacts. Biological plausibility is depicted via the proposed pathways as a continuum of upstream
22 events, connected by arrows, that may lead to downstream events observed in epidemiologic studies
23 (Figure 7-5). This discussion of "how" exposure to ozone may lead to effects on the nervous system
24 contributes to an understanding of the biological plausibility of epidemiologic results evaluated later. The
25 biological plausibility for ozone-induced effects on the nervous system is supported by evidence from the
26 2013 Ozone ISA and by new evidence.
27 As discussed in the short-term exposure section (see Section 7.2.1.2). inflammation is also
28 expected to be an important mechanism driving nervous system effects following long-term ozone
29 exposure. The first proposed pathway (Figure 7-5) is largely conserved across the short- and long-term
30 exposure durations, however, there is a stronger link to neurodegenerative outcomes in humans following
31 long-term exposures. Briefly, inhaled ozone elicits inflammation releasing inflammatory cytokines and
32 ROS into the bloodstream that trigger systemic inflammation. Proinflammatory markers interact with, and
33 in some cases infiltrate, the blood-brain barrier initiating neuroinflammation, as indicated by altered gene
34 expression, increased apoptosis, lipid/protein oxidation, and microglial activation (see Section 7.2.2.3:
35 Table 7-26). These effects are associated with changes to nervous system function
September 2019
7-34
DRAFT: Do Not Cite or Quote
-------
1 (e.g., behavior/cognition, sleep disturbances, neurotransmitter levels) and structure (e.g., blood brain
2 barrier, fS-am\ loid accumulation, morphology) that are associated with neurodegenerative diseases such
3 as Alzheimer's and Parkinson's diseases, and mood disorders.
Long-Term
r ^
Neuroinflammation in
Developing Animals
L.
i
i
,
,
Altered
Neurotransmitter
Levels
Altered
Neurodevelopmental
Processes
Neurodevelopmental
Disorders
Neuroinflammation
in Adult Animals:
Whole Brain
Olfactory Bulb
Cerebral Cortex
Cerebellum
Hippocampus
Altered
Neurotransmitter Levels
(Seratonin, Dopamine,
Acetylcholine)
Cognitive and Behavioral
Changes (Learning,
Memory, Motor Activity)
Neurodegenerative
Diseases
\
Mood Disorders
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
(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.
Figure 7-5 Potential biological pathways for nervous system effects
following long-term exposure to ozone.
4 In the second pathway (Figure 7-5). adverse nervous system effects have also been reported when
5 exposure occurs during development. Inflammation is expected to be a critical pathway for
6 neurodevelopmental effects of ozone on developing offspring, just as it is in adults. In animal models,
7 respiratory and systemic inflammation, either from direct (i.e., inhalation) or indirect (i.e., via the dam)
8 exposure, are expected to elicit neuroinflammation that is associated with altered neurotransmitter levels,
9 cognitive and behavioral changes, and altered development of the peripheral nervous system. Together,
10 these effects may contribute to neurodevelopmental disorders. Neuroinflammation in developing animals
September 2019
7-35
DRAFT: Do Not Cite or Quote
-------
1 may also be triggered by activation of sensory nerves in the lung, leading to altered neurodevelopment of
2 the nodose and jugular ganglia that transmit sensory information from the lung to the brain (see
3 Section 7.2.2.5; Table 7-28).
4 The pathway(s) described here provide biological plausibility for evidence of neurodegenerative
5 diseases, mood disorders, and sleep disturbances in adults (see Section 7.2.2.4') and neurodevelopmental
6 disorders in children (see Section 7.2.2.5) in association with long-term exposure to ozone. These
7 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
8 In the 2013 Ozone ISA, long-term ozone exposure elicited similar effects on the brain versus
9 short-term exposure (see Section 7.2.1.3.1). with many studies showing increases of inflammatory and
10 oxidative-stress responses, elevated cell death, and changes in neuronal morphpology in various regions
11 of the brain. In general, the magnitude and severity of the effects was generally increased with longer
12 exposure durations; however, some studies found these effects could be mitigated by coexposure with
13 antioxidants.
14 As discussed below, the effects of long-term exposure on brain inflammation and morphology
15 were similar to those described in the short-term exposure section (see Section 1.2.1.1): however, the
16 magnitude and severity of the effects were generally increased with longer exposure durations. These
17 effects were observed in multiple brain regions. There is also some evidence to suggest that males may be
18 more susceptible than females to inflammation and oxidative damage. Study details are provided in
19 Table 7-26.
20 • Multiple studies measured elevated levels of oxidative stress and inflammation in the brains of
21 rats and mice following long-term exposure to ozone (Rodriguez-Martinez et al.. 2016; Akhter et
22 al.. 2015; Rivas-Arancibia et al.. 2015; Gomez-Crisostomo et al.. 2014; Pinto-Almazan et al..
23 2014; Rodriguez-Martinez et al.. 2013; Mokoena et al.. 2011). Histological analyses revealed
24 reduced cell counts and increased apoptosis and oxidative damage in several regions of the brain,
25 including the hippocampus (Rodriguez-Martinez et al.. 2016; Gomez-Crisostomo et al.. 2014;
26 Pinto-Almazan et al.. 2014; Rodriguez-Martinez et al.. 2013). frontal cortex (Mokoena et al..
27 2011). and substantia nigra (Rivas-Arancibia et al.. 2015). These regions were also found to have
28 damage to the mitochondria and endoplasmic reticulum (Rivas-Arancibia et al.. 2015; Gomez-
29 Crisostomo et al.. 2014; Rodriguez-Martinez et al.. 2013).
30 • Although the majority of the data were generated in male Wistar rats, the study by Akhter et al.
31 (2015) evaluated the oxidative effects of ozone exposure in both males and females using a
32 mouse model of Alzheimer's disease. Ozone-exposed Alzheimer's disease model males exhibited
33 significantly greater apoptosis in the hippocampus relative to the other experimental groups
34 (i.e., wild-type males and females + ozone, Alzheimer's disease model males and
September 2019
7-36
DRAFT: Do Not Cite or Quote
-------
1 females + filtered air, and Alzheimer's disease females + ozone). Male Alzheimer's disease
2 model mice were found to have significantly lower baseline antioxidant levels than wild-type
3 animals or Alzheimer's disease model females which may make them more susceptible to
4 oxidative stressors (Akhter et al.. 2015).
5 • A series of studies from the same research group reported that long term exposure to ozone
6 affected (3-amyloid accumulation and regulation in the brain. These effects are strongly associated
7 with development of Alzheimer's disease. Accumulation of (3-amyloid proteins was increased,
8 and alterations in several proteins and genes that regulate (3-amyloid formation and degradation
9 were observed in the hippocampus and cortex of male Wistar rats following exposure to 0.25 ppm
10 ozone (Rivas-Arancibia et al.. 2017; Fernando Hernandez-Zimbron and Rivas-Arancibia. 2016;
11 Hernandez-Zimbron and Rivas-Arancibia. 2015V In general, these results showed an
12 exposure-dependent trend, with the magnitude of effect increasing with exposure duration. Rivas-
13 Arancibia et al. (2017) also found that ozone exposure induced exposure-dependent changes in
14 the folding of (3-amyloid proteins in a manner consistent with those observed in (3-amyloid
15 plaques associated with Alzheimer's disease. In some cases, (3-amyloid was found to be
16 colocalized with mitochondria (Hernandez-Zimbron and Rivas-Arancibia. 2015) and the
17 endoplasmic reticulum (Fernando Hernandez-Zimbron and Rivas-Arancibia. 2016). In contrast, a
18 single study from a different laboratory did not find an effect of ozone exposure on (3-amyloid
19 accumulation in a transgenic Alzheimer's disease mouse model. These animals were
20 intermittently exposed to ozone for 4 months and while all Alzheimer's disease model animals
21 showed (3-amyloid accumulation in the hippocampus and cortex, there was no effect of ozone
22 exposure (Akhter et al.. 2015).
7.2.2.4 Effects on Cognition, Motor Activity, and Mood
23 The cognitive and behavioral effects measured in the epidemiologic studies reviewed in this
24 section include scores on the Mini-Mental State Examination (MMSE), which is a questionnaire used to
25 screen for dementia, and performance on neurobehavioral tests of cognitive function. Depression was
26 evaluated using self-reported information on depression diagnosis and use of antidepressant medication.
27 Clinically diagnosed dementia, including Alzheimer's disease and vascular dementia, and Parkinson's
28 disease were also examined in a small number of studies. In a few animal toxicological studies, effects on
29 learning and memory, motor activity, and anxiety were evaluated.
7.2.2.4.1 Epidemiologic Studies
Cognition and Dementia-Related Effects
30 The 2013 Ozone ISA reported declines on tests of cognitive function measured using
31 Neurobehavioral Evaluation System-2 (NES2), in a cross-sectional analysis of NHANES III (1988-1991)
32 data (Chen and Schwartz. 2009). A small number of recent studies examine the effect of long-term
33 exposure to ozone with performance on neurobehavioral tests (see Table 7-20). Alzheimer's disease, and
34 other forms of dementia (see Table 7-21). Overall, the limited number of epidemiologic studies support
September 2019
7-37
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
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. Clearv 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 reported associations
of ozone exposure with Parkinson's disease. Shin et al. (2018) and Cerza et al. (2018) reported a
positive associations (HR: 1.06; 95% CI: 1.02, 1.11 and HR: 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 an 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,
other researchers using the same database but a quantile-based Bayesian maximum entropy
spatio-temporal model to characterize long-term exposure, Lee et al. (2016) reported a null
September 2019
7-38
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
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 7-201V
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
varied. Most of the studies reported decreased activity in rats (Gordon et al.. 2016; Pinto-
Almazan et al.. 2014; Gordon et al.. 2013) and is in agreement with the data included 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.
September 2019
7-39
DRAFT: Do Not Cite or Quote
-------
7.2.2.4.3
Summary
1 The current section describes and characterizes the epidemiologic and toxicological evidence
2 relating to the effect of long-term ozone exposure on cognition, motor activity, and mood.
3 • Biological plausibility for the long-term effect of ozone on the nervous system is derived from
4 multiple studies demonstrating that long-term exposure to ozone can to lead to inflammation and
5 oxidative stress responses in the brain.
6 • Limited epidemiologic evidence reports associations with decrements on tests of cognitive
7 function that may be associated with neurodegenerative diseases, such as Alzheimer's and
8 Parkinson's disease. Toxicological studies provide coherence for these findings, but
9 epidemiologic studies of Alzheimer's and Parkinson's disease are not consistent. The animal data
10 do not support an association between long-term ozone exposure and mood disorders. The
11 epidemiologic evidence is limited to a study reporting an association with self-reported
12 depression.
7.2.2.5 Neurodevelopmental Effects
13 In the 2013 Ozone ISA, discussion of the data on neurodevelopmental effects was split across the
14 short- and long-term exposure sections, however, in the current ISA these data are only reviewed in the
15 long-term exposure section due to the sensitivity of the developing nervous system to toxicants and the
16 potential for long-term impacts. The 2013 Ozone ISA reviewed toxicological evidence for
17 neurodevelopmental effects of prenatal and early life ozone exposure. Exposure that was limited the
18 prenatal period altered gene expression of nerve growth factors, affected regulation of neurotransmitter
19 levels and altered neuroadaptive responses to stress. Social interaction, defensive/submissive behavior,
20 and turning preferences were also affected in animals exposed either prenatally or during both gestation
21 and lactation. Notably, some of these outcomes persisted into adulthood, suggesting early life exposure
22 can have long lasting impacts on neurological function. There were no epidemiologic studies of long-term
23 exposure to ozone and neurodevelopmental outcomes. A recent study by Lin et al. (2014a) examined the
24 effect of prenatal ozone exposure and neurobehavioral outcomes but reported no evidence of an
25 association. The current evidence base also includes several epidemiologic studies of autism spectrum
26 disorder (ASD) and toxicological studies that focus on the peripheral nervous system.
7.2.2.5.1 Epidemiologic Studies
27 There were no studies of long-term exposure to ozone and autism reviewed in the 2013 Ozone
28 ISA. Several recent studies conducted in the U.S. and Taiwan are currently available (see Table 7-28).
29 Overall, these studies report positive associations, but associations are imprecise (i.e., wide confidence
30 intervals) and are not consistently observed across pregnancy periods. In addition, outcome definitions for
31 autism, which is a heterogenous condition with potentially different etiologies, varied across studies. For
September 2019
7-40
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
example, Becerra et al. (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).
• Becerra et al. (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
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, Zellner etal. (2011) reported no effect of ozone exposure on
pulmonary innervation.
• Hunter etal. (2011) studied the effects of ozone on lung 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
September 2019
7-41
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
window for neuronal development in the lung, eliciting statistically significant increases in NGF
in the short term (24-hour PE) and potentiating the effects of subsequent ozone exposures.
7.2.2.5.3 Summary
Section 7.2.2.5 describes and characterizes the epidemiologic and toxicological evidence relating
to the effect of long-term ozone exposure on neurodevelopment. 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.. 2013). 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. 2013a) 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 reproductive 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 aged and young populations may have
increased sensitivity to ozone exposure. Functional deficits in tasks of learning and memory and
September 2019
7-42
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
decreased motor activity were correlated with biochemical and 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 aged animals showed differences in the patterns of oxidative stress, with young animals
showing higher levels in the striatum and aged 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. The potential for these effects have been
supported by data derived from toxicological studies.
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. 2013a). 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)
September 2019
7-43
DRAFT: Do Not Cite or Quote
-------
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 etal. (2017c)
NR
disease
studies of Alzheimer's
Junq et al. (2014)
92.6 ppb
disease or dementia are
Wu etal. (2015)
NR
imprecise and vary in
magnitude
Associations with
Kirrane et al. (2015)
40.6
Parkinson's disease are not
Chen etal. (2017a)
NR
consistently observed and
Lee etal. (2016)
26.1
generally lack precision
Shin etal. (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.
September 2019
7-44
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
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. 2013a'). 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 draft 2019
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:
• 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
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, the studies evaluated are less
limited in scope and not targeted towards specific study locations, as reflected in the PECOS tool. The
epidemiologic studies evaluated and discussed in this section were identified using the following PECOS
tool:
September 2019
7-45
DRAFT: Do Not Cite or Quote
-------
1 • Population: Any population, including populations or lifestages that might be at increased risk
2 • Exposure: Long-term ambient concentration of ozone
3 • Comparison: Per unit increase (in ppb)
4 • Outcome: Change in risk (incidence/prevalence) of a cancer effect
5 • Study Design: Epidemiologic studies consisting of cohort, case-control and cross-sectional
6 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
7 As noted in the 2013 Ozone ISA, the potential for genotoxic effects relating to ozone exposure
8 was predicted from the radiomimetic properties of ozone. The decomposition of ozone in water produces
9 OH and HO2 radicals, the same species that are generally considered to be the biologically active products
10 of ionizing radiation. Ozone has been observed to cause degradation of DNA in a number of different
11 models and bacterial strains. The toxic effects of ozone have been generally assumed to be confined to the
12 tissues directly in contact with the gas, such as the respiratory epithelium.
13 Several epidemiologic studies evaluated in the 2013 Ozone ISA observed positive associations
14 between long-term ozone exposure and DNA damage (i.e., DNA adduct levels, oxidative DNA damage,
15 DNA strand breaks). In addition, there was some evidence of cytogenetic damage (i.e., micronuclei
16 frequency among lymphocytes and buccal cells) after long-term, but not short-term ozone exposure. Such
17 DNA and cytogenic damage may be relevant to mechanisms leading to cancer development and serve as
18 early indicators of an elevated risk of mutagenicity.
19 Since the 2013 Ozone ISA, a few additional studies have looked at the relationship between
20 ozone exposure and the potential for DNA damage and found inconsistent results:
21 • Holland et al. (2014) exposed healthy volunteers to filtered air (FA), 100, and 200 ppb ozone and
22 collected blood lymphocytes 24-hours post-exposure. A statistically significant increase in the
23 frequency of micronuclei in binucleated cells was reported with increasing ozone concentrations
24 (p < 0.05). However, these authors also reported no appreciable changes in neoplasmic bridges
25 (an indicator of radiation and other types of genotoxic exposure) and no difference in cell
26 proliferation following ozone exposure.
27 • Finkenwirth et al. (2014) exposed healthy volunteers to FA and ozone, collected lymphocytes and
28 analyzed them for single stranded breaks. No appreciable difference in single stranded breaks
29 were observed at either 30 minutes or 4.5 hours post-exposure.
30 • In rats, Cestonaro et al. (2017) evaluated exposure to 0.05 ppm ozone from an air purifier for
31 3 hours or 24 hours per day for 14 or 28 days. The results indicated no significant effects on
September 2019
7-46
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
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 rat lung tissue, 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 ACS CPSII
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. 2013a') 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. 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).
• 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.
September 2019
7-47
DRAFT: Do Not Cite or Quote
-------
1 • A number of recent studies conducted in the U.S., Canada, and Europe provided limited and
2 inconsistent evidence for an association between long-term ozone exposure and lung cancer
3 mortality (Cakmak et al.. 2017; Turner et al.. 2016; Crouse et al.. 2015; Carey et al.. 2013; Jerrett
4 et al.. 2013) (Table 7-30).
5 • A case-crossover study conducted in Shenyang, China observed null associations between
6 short-term ozone exposure and lung cancer mortality (Xuc et al.. 2018).
7 • Studies of childhood leukemia (Badaloni et al.. 2013) and breast tissue density, an indicator of
8 breast cancer (Yaghjyan et al.. 2017). observed null associations with long-term ozone exposure.
7.3.3 Summary and Causality Determination
9 In the 2013 Ozone ISA, very few studies were available to assess the relationship between
10 long-term exposure to ozone and carcinogenesis. The few available toxicological and epidemiologic
11 studies suggested that ozone exposure may contribute to DNA damage. However, given the overall lack
12 of studies, the 2013 Ozone ISA concluded that the evidence was inadequate to determine whether a causal
13 relationship existed between ambient ozone exposures and cancer.
14 The number of studies examining the relationship between ozone exposure and the potential for
15 carcinogenesis reman few. Studies published since the 2013 Ozone ISA provide some additional animal
16 toxicological evidence that ozone exposure can lead to DNA damage. In addition, several but not all
17 recent cohort and case-control studies have observed positive associations between long-term ozone
18 exposure and lung cancer incidence or mortality. Several of the studies evaluating lung cancer mortality
19 were conducted among populations that had already been diagnosed with cancer in a different organ
20 system. Associations between ozone exposure and other types of cancer were generally null. Given the
21 limited evidence base, the lack of an evaluation of copollutant confounding in epidemiologic studies
22 reporting associations, and the evaluation of study populations that had already been diagnosed with
23 cancer in several of the epidemiologic studies, the evidence is not sufficient to draw a conclusion
24 regarding causality (Table 7-5). Thus, the evidence describing the relationship between exposure to
25 ozone and carcinogenesis remains inadequate to determine if a causal relationship exists.
September 2019
7-48
DRAFT: Do Not Cite or Quote
-------
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
Guo et al. (2016)
56.9
A limited number of recent studies Eckel et al. (2016) 28.5
observed positive associations
between long-term ozone exposure
and lung cancer or respiratory mortality xu et al (2013) 40 2
in study populations already diagnosed
with cancer
No epidemiologic A limited number of recent studies Badaloni et al. (2013) 24.2
evidence for other observed null associations between
cancers long-term ozone exposure and
childhood leukemia and breast cancer Yagh|yan et al. (2017) 36.0
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.
1
September 2019
7-49
DRAFT: Do Not Cite or Quote
-------
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
Other, 90 days before
sample collection
Mean: 42
75th: 48
Maximum: 54
Correlation (r):
NO2: 0.6;
Other: 0.28
Copollutant models
with: 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)
tLiuetal. (2017)
Wuhan, China
Ozone: NR
Follow-up: 2013-2015
Cohort study
n = 1,759 men
Monitor
Other, 0-90 days before
sample collection
Mean: 25-64
Correlation (r): NR
Copollutant models
with: NR
Sperm concentration (A million/mL): 0.082
(-0.077, 0.240)
Sperm count (A million): 0.018 (-0.145, 0.181)
Total motility (A percentage): 0.082 (-0.068,
0.236)
Progressive motility (A percentage): 0.068
(-0.086, 0.217)
Total motile sperm count (A million): 0.041
(-0.113, 0.199)
fRecent studies evaluated since the 2013 Ozone ISA.
September 2019
7-50
DRAFT: Do Not Cite or Quote
-------
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
Other
Mean: NR Correlation (r) (1st mo
unprotected
intercourse):
PM2.5: -0.41;
NO2
SO2
-0.49;
-0.69
Copollutant models
with: NO2, PM2.5
Fecundity (FR)
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 etal. (2016)
Region NR, France
Ozone: 2012-2015
Follow-up: April
2012-December 2015
Cohort study
n = 292 couples Monitors
Couples undergoing Other
IVF attempts
NR
Correlation (r):NR
Copollutant models
with: NR
Ovarian response to stimulation and
number of top embryos were increased
with short- or long-term exposures to
high levels of ozone
tNobles etal. (2018)
Michigan and Texas, U.S.
Ozone: 2005-2010
Follow-up: 2005-2009
Cohort study
Longitudinal
Investigation of
Fertility and the
Environment (LIFE)
Study
n = 500 couples
Couples had
presumed exposure to
persistent organic
pollutants
Model, modified
CMAQ
Other
75th (cycle prior Correlation (r):NR
to observed
cycle): 27.85
90th: 34.2
Maximum:
40.54
Copollutant models
with: PM2.5, NOx,
SO2, CO
Fecundability (FOR)
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.
fRecent studies evaluated since the 2013 Ozone ISA.
September 2019
7-51
DRAFT: Do Not Cite or Quote
-------
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
+Lee et al. (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: 46.8
Correlation (r):
PM2.5: 0.5;
Other: 0.7
Copollutant models with:
NR
Gestational hypertension (OR): 1.07
(0.98, 1.16)
Preeclampsia (OR): 1.07 (0.93, 1.23)
tMobasher et al. (2013)
Los Angeles, CA, U.S.
Ozone: NR
Follow-up: 1996-2008
Case-control study
n = 298 Monitor
Attended the Other
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;
Other: -0.64
Copollutant models with:
NR
2nd trimester
PM2.5: -0.53;
NO2: -0.74;
Other: -0.57
Copollutant models with:
NR
3rd trimester
PM2.5: -0.55;
NO2: -0.78;
Other: -0.66
Copollutant models with:
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 et al. (2013)
Stockholm, Sweden
Ozone: 1997-2006
Follow-up: 1998-2006
Cohort study
n = 120,755 Monitor
Swedish 1st trimester
medical birth
registry
Mean: 35
Correlation (r):
NO2: -0.48
Copollutant models with:
NO2
Preeclampsia (OR)
1.10 (1.04, 1.17)
Adjusted for NO2: 1.23(1.06, 1.44)
September 2019
7-52
DRAFT: Do Not Cite or Quote
-------
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 etal. (2014)
Florida, U.S.
Ozone: 2003-2005
Follow-up: 2004-2005
Cohort study
n = 22,041
birth records
Monitor
24-h avg
Other
Mean:
1st trimester: 40
2nd trimester: 41
3rd trimester: 40
Correlation (r):NR
Copollutant models with:
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)
tNahidi etal. (2014)
Tehran, Iran
Ozone: 2010-2011
Follow-up: September
2010-March 2011
Case-control study
n = 65 cases;
130 controls
admitted to
hospitals in
Tehran
Monitor
Entire pregnancy
Mean: NR
Correlation (r):NR
Copollutant models with:
NR
Preeclampsia (OR)
High vs. low exposure: 1.00 (0.49,
2.03)
tMichikawa et al. (2015)
Kyushu-Okinawa District,
Japan
Ozone: NR
Follow-up: 2005-2010
Cohort study
Japan
Perinatal
Registry
Network
n = 36,620
Monitor
1st trimester
Mean: 41.3
Median: 40.1
75th: 48
Correlation (r):
PM2.5: 0.12
NO2: -0.18
SO2: -0.17
Copollutant models with:
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) Consortium on Model, modified Median:
U.S.
Ozone: NR
Follow-up: 2002-2008
Cohort study
Safe Labor
n = 192,687
women
recruited from
12 centers
(19 hospitals)
across U.S.
CMAQ
Other
Preconception: 29.7
1st trimester: 29.2
2nd trimester: 29.4
Entire pregnancy: 28.5
Correlation (r):NR
Copollutant models with:
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)
September 2019
7-53
DRAFT: Do Not Cite or Quote
-------
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
tHu et al. (2016)
Florida, U.S.
Ozone: NR
Follow-up: 2005-2007
Cohort study
n = 655,529
birth records
Model, CMAQ
hierarchical
Bayesian
Other
Mean:
1st trimester: 38.65
2nd trimester: 38.59
Other: 38.63
Median:
1st trimester: 37.91
2nd trimester: 36.95
Other: 5.33
Correlation (r):NR
Copollutant models with:
NR
Hypertensive disorders of pregnancy
(each week of 1 st two trimesters)
1st trimester: 1.08 (1.06, 1.12)
2nd trimester: 1.06 (1.04, 1.08)
1st and 2nd trimesters: 1.14 (1.10,
1.17)
ORs elevated with ozone exposure at
each week of pregnancy (1-26)
tWu et al. (2011)
Los Angeles and Orange
counties, CA, U.S.
Ozone: 1997-2006
Follow-up: NR
Cohort study
fRecent studies evaluated since the 2013 Ozone ISA.
September 2019 7-54 DRAFT: Do Not Cite or Quote
n = 81,186
hospital-based
birth database
Monitor
Entire pregnancy
Mean: 36.5
Correlation (r):
PM2.5: -0.61;
NO2: -0.81;
Other: -0.74
Copollutant models with:
NR
Preeclampsia (OR)
Los Angeles: 1.00 (0.74, 1.35)
Orange: 1.46 (1.12, 1.90)
-------
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
tRobledo et al. (2015) Consortium on Safe
U.S.
Ozone: 2001-2008
Labor
n = 220,264
Model, modified
CMAQ
Other
Follow-up: 2002-2008 12 clinical centers/19
~ x , hospitals
Cohort study
Median:
Other: 29.71
1st trimester: 29.21
75th:
Other: 35.82
1st trimester: 35.17
Correlation (r):
PM2.5: -0.38;
NO2
SO2
-0.39;
-0.42
Copollutant models with:
NR
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 etal. (2015) n = 410.267 birth
Model, CMAQ
Mean:
Correlation (r):
Gestational diabetes (OR)
Florida, U.S. records
hierarchical Bayesian
1st trimester: 37.22
PM2.5: 0.39
1st trimester: 1.19 (1.14, 1.23)
Ozone: NR
Other
2nd trimester: 37.54
Copollutant models with:
2nd trimester: 1.25 (1.21, 1.30)
Entire pregnancy: 37.84
NR
Entire pregnancy: 1.39 (1.32,
Follow-up: 2004-2005
Median:
1.46)
Cohort study
1st trimester: 36.48
2nd trimester: 36.95
Entire pregnancy: 7.09
fRecent studies evaluated since the 2013 Ozone ISA.
September 2019
7-55
DRAFT: Do Not Cite or Quote
-------
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
German Monitor
Socioeconomic nthor
Panel (SOEP)
NR
Correlation (r): NR
Copollutant models
with: 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
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;
Other: 0.7
Copollutant models
with: 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 Entire pregnancy
births
Mean: 23
Correlation (r):
PM2.5: -0.12;
NO2: -0.77;
Other: -0.28
Copollutant models
with: NR
Low birth weight (percentage change
per 7 ppb):
-6.3 (-11, -1.3)
tLaurent et al. (2013)
Los Angeles and Orange
counties CA, U.S.
Ozone: NR
Follow-up: 1996-2006
Cohort study
n = 70,000 births
Hospital-based
obstetric database
Monitor
Entire pregnancy
Mean: 35.66
Correlation (r):
PM2.5: -0.61;
NO2: -0.81;
Other: -0.74
Copollutant models
with: NR
Term birth weight (A g):
-27.27 (-32.02, -22.51)
Low birth weight (OR):
1.11 (1.02, 1.21)
September 2019
7-56
DRAFT: Do Not Cite or Quote
-------
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
tOlsson etal. (2013)
Stockholm, Sweden
Ozone: 1997-2006
Follow-up: 1998-2006
Cohort study
n = 120,755
Swedish medical
birth registry
Monitor
1st trimester
Mean: 35
Correlation (r):
N02: -0.48
Copollutant models
with: NO2
Small for gestational age (OR):
0.98 (0.96, 1.02)
Adjusted for NO2: 0.98 (0.90, 1.06)
tGeeretal. (2012)
Texas, U.S.
Ozone: NR
Follow-up: 1998-2004
Cohort study
n = 1,548,904
Birth records,
40 Texas counties
Monitor
Entire pregnancy
Mean: 25.4
Correlation (r): NR
Copollutant models
with: NR
Term birth weight (A g):
-4.61 (-7.34, -1.88)
tRitz etal. (2014)
New York City, NY, U.S.
Ozone: 1993-1996
Follow-up: 1993-1996
Cohort study
Behavior in
pregnancy study
n = 688
Monitor
Other
Mean: 40.2
Maximum: 96.1
Correlation (r): NR
Copollutant models
with: NR
Biparietal diameter (A mm)
Estimated date of conception to 1st
ultrasound date -0-19 weeks
gestation: 0.026 (-0.153, 0.199)
1 st to 2nd ultrasound date
(-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)
tLaurent et al. (2014)
Los Angeles County CA,
U.S.
Ozone: 2000-2008
Follow-up: 2001-2008
n = 960,945
Birth records, term
births
Model, empirical
Bayesian kriging based
on monitor
Entire pregnancy
Mean: 38.95
Correlation (r): NR
Copollutant models
with: NR
Low birth weight (OR): 0.99 (0.98,
1.00)
Cohort study
September 2019
7-57
DRAFT: Do Not Cite or Quote
-------
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
tGrav etal. (2014)
n = 457,642
CMAQ downscaler
Mean: 43.2
Correlation (r): NR
Birth weight (A g): -12.54 (-16.10,
North Carolina, U.S.
Ozone: 2001-2006
Follow-up: 2002-2006
Cohort study
Birth records
Entire pregnancy
Copollutant models
with: NR
-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
Other
Mean: 40.85
Median: 41.86
75th: 48.86
Maximum: 70.35
Correlation (r):NR
Copollutant models
with: 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)
+Ha et 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
Other
Mean: 37.2
Median: 36.5
75th: 41
Maximum: 56.2
Correlation (r):NR
Copollutant models
with: 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)
tSmith etal. (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
with: 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"
September 2019
7-58
DRAFT: Do Not Cite or Quote
-------
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
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
Other
Median: 38.77
75th: 42.03
Maximum: 60.35
Correlation (r):NR
Copollutant models
with: 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.
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.
90, 1.02)
84, 0.96)
90, 1.02)
83, 0.94)
81, 0.92)
92, 1.04)
tCapobussi et al. (2016)
Como, Italy
Ozone: NR
Follow-up: 2005-2012
Cohort study
n = 27,128
Birth records
Monitor, within 5 km
Entire pregnancy
Correlation (r):NR
Copollutant models
with: NR
Low birth weight (OR): 0.96 (0.85,
1.08)
Small for gestational age (OR): 1.00
(0.95, 1.06)
tLaviqne et al. (2016)
Ontario, Canada
Ozone: 2002-2009
Follow-up: January 1,
2005-March 31, 2012
Cohort study
Better Outcomes
Registry &
Network Ontario
n = 362,800
Singleton live
births
Model
Entire pregnancy
Mean: 27.8
Median: 28
95th: 33.05
Correlation (r):
PM2.5: -0.14;
NO2: -0.53
Copollutant models
with: NR
Term low birth weight (OR): 1.17
(1.09, 1.24)
Small for gestational age (OR): 1.24
(1.21, 1.28)
September 2019
7-59
DRAFT: Do Not Cite or Quote
-------
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
tYitshak-Sade et al.
(2016)
Negev, Israel
Ozone: NR
Follow-up: December
2011-April 2013
Cohort study
n = 959
Bedouin-Arab
population in
southern Israel,
seminomadic
Monitor, IDW
Other
Mean: 11
Correlation (r): NR
Copollutant models
with: 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)
tLaurent et al. (2016a)
California, U.S.
Ozone: 2000-2008
Follow-up: 2001-2008
Case-control study
n = 72,632 cases;
x five controls
Birth records
Model, empirical
Bayesian kriging based
on monitor
Entire pregnancy
Mean: 39.55
Correlation (r): NR
Copollutant models
with: NR
Low birth weight (OR): 1.03 (1.02,
1.05)
+Tu et al. (2016)
Atlanta, GA, U.S.
Ozone: 2001
Follow-up: 2000
Cohort study
n = 105,818
Term births, birth
records
Model, CMAQ downscaler
2001 annual average
Mean: 42.76
Maximum: 48.99
Correlation (r):NR
Copollutant models
with: 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.
September 2019
7-60
DRAFT: Do Not Cite or Quote
-------
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
Procriar
n = 453
Pregnant women
from three
prenatal care units
Personal sampler
24-h avg
Other—1 day during
specified trimester
Mean: 4
Median: 4
Maximum: 7
Correlation (r):NR
Copollutant models
with: 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
All live singleton
births in Madrid
Monitors
Other
Mean: 18
Median: NA
75th: NA
Maximum: 38
Correlation (r): NR
Copollutant models
with: 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
Other
Mean: 17
Maximum: 38
Correlation (r): NR
Copollutant models
with: NR
Low birth weight (weekly)
Reported only statistically significant
effects, no associations reported for
ozone
September 2019
7-61
DRAFT: Do Not Cite or Quote
-------
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
Japan Perinatal
Registry Network
n = 29,177
Monitor
Other
Mean: 41.2
Median: 40
75th: 47.8
Correlation (r): NR
Copollutant models
with: 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
Entire pregnancy
Mean: 16
Correlation (r): NR
Copollutant models
with: 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 Other
births singletons
no birth defects
Mean: 28
Median: 28
75th: 30
Maximum: 36
Correlation (r): NR
Copollutant models
with: 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
Other
Mean: 16.82
Median: 16.78
75th: 17.58
Maximum: 22.34
Correlation (r):
PM2.5: 0.27;
NO2
SO2
-0.04;
-0.04;
Other: -0.16
Copollutant models
with: 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
September 2019
7-62
DRAFT: Do Not Cite or Quote
-------
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
tReiset al. (2017) n = 13,660 birth
Volta Redonda and Rio de records
Janeiro, Brazil
Ozone: NR
Follow-up: 2003-2006
Cohort study
Monitor
Other
Mean: 30
Correlation (r):NR
Copollutant models
with: NR
Low birth weight
Exposure increment not reported,
ORs elevated from null reported for
2nd and 3rd trimester exposures but
not 1st
fRecent studies evaluated since the 2013 Ozone ISA.
September 2019
7-63
DRAFT: Do Not Cite or Quote
-------
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
Entire
pregnancy
Mean: 36.5 Correlation (r):
PM2.5: -0.61;
NO2: -0.81;
Other: -0.74
Copollutant models with: NR
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 et al.
(2012)
Stockholm, Sweden
Ozone: 1986-1995
Follow-up:
1987-1995
Cohort study
n = 115,588 Monitor
Swedish medical Other
birth registry
Mean:
1st trimester: 29
2nd trimester: 29
Other: 30
Median: 1st
trimester: 28
2nd trimester: 28
Other: 30
Correlation (r):
NO2:
1st trimester: -0.43;
2nd trimester: -0.39;
Other: -0.26
Copollutant models with: 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)
September 2019
7-64
DRAFT: Do Not Cite or Quote
-------
Table 7-11 (Continued): Epidemiologic studies of exposure to ozone and pregnancy/birth—preterm birth.
Study Exposure Effect Estimates
Study Population Assessment Mean (ppb) Copollutant Examination 95% CI
Correlation (r): PTB (OR)
PM2.5: 0.5; 1.13(1.01,1.27)
Other: 0.7;
Copollutant models with: NR
Cohort study
tOlsson et al.
n = 120,755
Monitor
Mean: 35
Correlation (r):
PTB (OR), 1st trimester:
(2013)
Swedish medical
1st trimester
NO2: -0.48
1.08 (1.02, 1.17)
Stockholm, Sweden
birth registry
Copollutant models with:
NO2
Asthmatic mother: 1.17 (1.02, 1.32)
Ozone: 1997-2006
Nonasthmatic mother: 1.08 (1.00, 1.17)
Follow-up:
Adjusted for NO2:
1998-2006
1.10 (0.98, 1.23)
Cohort study
Asthmatic mother: 1.17 (1.00, 1.37)
Nonasthmatic mother: 1.10 (0.98, 1.23)
tWarren et al.
n = 32,170
Monitor and
Correlation (r):NR
PTB, results presented as figures
(2012)
observations
Model (fused
Copollutant models with:
NR
Effect estimates elevated from null with
Texas, U.S.
Birth records,
CMAQ)
exposures in early weeks of pregnancy,
Ozone: 2002-2004
singleton live birth
Other
and for 1st and 2nd trimester exposures.
Follow-up:
2002-2004
Cohort study
tSchifano et al.
n = 132,691
Monitor
Median: 19
Correlation (r): NR
PTB (percentage increase), lag 1-2:
(2013)
births
8-h max
75th: 29
Copollutant models with:
NR
1.01 (0.94, 1.09)
Rome, Italy
Birth records
Maximum: 66
Ozone: 2001-2007
Follow-up:
2001-2010
Cohort study
tLee etal. (2013)
Pittsburgh, PA, U.S.
Ozone: 1996-2002
Follow-up:
1997-2002
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.i
September 2019
7-65
DRAFT: Do Not Cite or Quote
-------
Table 7-11 (Continued): Epidemiologic studies of exposure to ozone and pregnancy/birth—preterm birth.
Study Exposure Effect Estimates
Study Population Assessment Mean (ppb) Copollutant Examination 95% CI
tGrav etal. (2014) n = 457,642
North Carolina, U.S. Birth records
Ozone: 2001-2006
Follow-up:
2002-2006
Cohort study
CMAQ Mean: 43.2 Correlation (r):NR PTB (OR):
downscaler Copollutant models with: NR 1.03 (0.98,1.07)
Entire
pregnancy
September 2019
7-66
DRAFT: Do Not Cite or Quote
-------
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
tHa et al. (2014)
Florida, U.S.
Ozone: 2003-2005
Follow-up:
2004-2005
Case-control study
n =423,719
Birth records,
singleton live
births
Model,
Mean:
CMAQ
1st trimester: 37.2
hierarchical
2nd trimester:
Bayesian
37.6
Other
3rd trimester:
37.4
Entire pregnancy:
37.4
Median:
1st trimester: 36.5
2nd trimester: 37
3rd trimester: 37
Entire pregnancy:
37.9
75th:
PTB:
1st trimester: 41
2nd trimester:
41.3
3rd trimester:
41.1
Entire pregnancy:
41
Maximum:
PTB:
1st trimester: 56.2
2nd trimester:
57.3
3rd trimester:
69.2
Entire pregnancy:
51.3
Correlation (r):NR
Copollutant models with: 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)
September 2019
7-67
DRAFT: Do Not Cite or Quote
-------
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 etal. (2016)
Georgia, U.S.
Ozone: 2002-2006
Follow-up: January
1, 2002-February
28, 2006
Cohort study
n = 511,658
births
Birth records
Model, Median: 40.88
CMAQ fused 75th: 51.14
with monitors Maximum: 75.06
8-h max
Other
Correlation (r):
PM2.5: 0.47;
NO2: -0.19;
SO2: 0.63;
Other: 0.7
Copollutant models with: 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)
tLinetal. (2015)
Region NR, Taiwan
Ozone: 2000-2007
Follow-up:
2001-2007
Case-control study
n = 86,224 cases; Monitor
344,896 controls 3.^ max
Birth registry other
Mean: 1st
trimester: 42.74
2nd trimester:
48.43
3rd trimester:
48.98
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;
Other: 0.53
Copollutant models with: 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 etal. (2015)
Wuhan, China
Ozone: NR
Follow-up:
August 19, 2010-
September 9, 2013
Cohort study
n = 95,911
Monitor
Entire
pregnancy
Mean: 38
Maximum: 74
Correlation (r):
PM2.5: -0.16;
NO2
SO2
-0.12;
-0.13;
Other: -0.12
Copollutant models with: 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)
September 2019
7-68
DRAFT: Do Not Cite or Quote
-------
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: ~51
Correlation (r):
Length of gestation (HR)
(2016)
Rome, Italy and
Barcelona, Spain
n = Barcelona:
27,255
8-h avg
Other
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:
Other:
Rome: 0.17;
Barcelona: -0.19
Copollutant models with: NR
2007-2012
tCapobussi et al.
n =27,128
Monitor,
Correlation (r):NR
PTB (OR) 0.99 (0.92, 1.06)
(2016)
Birth records
within 5 km
Copollutant models with: NR
Como, Italy
Entire
Ozone: NR
pregnancy
Follow-up:
2005-2012
Cohort study
tLaviqne et al.
(2016)
Ontario, Canada
Ozone: 2002-2009
Follow-up: January
1, 2005-March 31,
2012
Cohort study
Better Outcomes
Registry &
Network Ontario
n = 362,800
Singleton live
births
Model
Entire
pregnancy
Mean: 27.8 Correlation (r):
Median: 28 PM2.5: -0.14;
95th: 33.05 NO2: -0.53
Copollutant models with: 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
Consortium on
Safe Labor, Air
Quality, and
Reproductive
Health Study
n = 223,375
Singleton
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 with: 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)
September 2019
7-69
DRAFT: Do Not Cite or Quote
-------
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.
(2016a)
U.S.
Ozone: NR
Follow-up:
2002-2008
Cohort study
Consortium on
Model,
Safe Labor, Air
modified
Quality, and
CMAQ
Reproductive
Other
Health Study
n = 223,502
singleton
pregnancies
Recruited from
12 centers
(19 hospitals)
across the U.S.
Mean: 29.65
Median: 37.3
Correlation (r): NR
Copollutant models with: NR
PTB
Elevated ORs with ozone exposure for
early PTB (<34 weeks) in mothers
without asthma for exposures during
Weeks 8-14 and 15-21
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 other
births
Mean: 37.6
Median: 33.7
75th: 47.5
90th: 62.4
Correlation (r): NR
Copollutant models with: 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)
Week 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)
September 2019
7-70
DRAFT: Do Not Cite or Quote
-------
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
n =442,314
cases; * two
controls
Birth records
Model,
empirical
Bayesian
kriging
based on
Mean: 39.71
Correlation (r):
PM2.5: -0.24;
NO2: -0.07
Copollutant models with: 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)
Follow-up:
2001-2008
Case-control study
monitor
Entire
pregnancy
tArrovo et al.
n = 470 weeks
Monitors
Mean: 18
Correlation (r): NR
PTB (OR)
(2016)
Madrid, Spain
Ozone: 2001-2009
All live singleton
births in Madrid
Other
Median: NA
75th: NA
Maximum: 38
Copollutant models with: NR
Week 12: 1.02 (1.01, 1.03)
Only significant results reported
Follow-up:
2001-2009
Time-series study
September 2019
7-71
DRAFT: Do Not Cite or Quote
-------
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
tChen et al.
(2017b)
Brisbane, Australia
Ozone: 2002-2013
Follow-up: July 1
2003-Dec 31 2013
Cohort study
173,720 birth
records
Monitors
24-h avg
Mean:
Entire pregnancy:
16.82
1st trimester:
16.82
2nd trimester:
16.76
3rd trimester:
16.91
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
Correlation (r): PM2.5: 0.27
NO2: -0.04;
SO2: -0.04;
Other: -0.16
Copollutant models with: NR
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
September 2019
7-72
DRAFT: Do Not Cite or Quote
-------
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
tDastoorpoor et al.
n =49,173 births
Monitor
Mean: 32
Correlation (r):NR
PTB (RR)
(2017)
Khuzestan
Province, Iran
Ozone: March
Ahvaz Imam
Khomeini
Hospital
24-h avg
Median: 24
Maximum: 3,316
Copollutant models with: NR
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)
2008-March 2015
Follow-up: March
2008-March 2015
Time-series study
tSmith etal. (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 with: 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
fRecent studies evaluated since the 2013 Ozone ISA.
September 2019
7-73
DRAFT: Do Not Cite or Quote
-------
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. (2011) Northern
Northeast England
Ozone: 1993-2003
Follow-up:
Case-control study
Congenital
Abnormality
Survey
n pooled
cases = 2,140;
controls =
14,256
Monitor
Weeks 3-
Mean: NR
Correlation
(r):NR
Copollutant
models with:
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
National Birth
Defects
Prevention
Study, CA
n = 1,651
subjects
Monitor
8-h max
1 st 2 mo of
pregnancy
Median: 46.95
75th: 65.65
Maximum: 91.92
Correlation (r):
PM2.5: -0.61;
NO2: -0.35;
Other: -0.71
Copollutant
models with:
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
10 km, inverse
distance weighing
Weeks 3-8
Mean: 25.1
Median: 26.5
75th: 39.1
Maximum: 128
Correlation (r):
NR
Copollutant
models with:
NR
Multiple congenital heart defects
Atrial and atrial septal defects
Isolated ventricular septal defects
Patent ductus arteriosus (BW > 2,500 g)
September 2019
7-74
DRAFT: Do Not Cite or Quote
-------
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)
n = 322,969 Model, hierarchical Mean: 40.74
Birth defect Bayesian model Median: 42.15
surveillance CMAQ and monitor Maximum: 74.99
birth registry Weeks 3-8
Correlation
(r):NR
Copollutant
models with:
NR
Spina bifida
Hydrocephalus
Anophthalmia/microphthalmia
Congenital cataract
Microtia/anotia
Transposition of great vessels
North Carolina, U.S.
Ozone: 2002-2005
Follow-up: 2003-2005
Cohort study
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
September 2019
7-75
DRAFT: Do Not Cite or Quote
-------
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-December31,
2006
Case-control study
National Birth
Defects
Prevention Study
n = 3,328 cases;
4,632 controls
Monitor
Weeks 2-8
Mean: 42.9
90th: 51.8
Correlation
(r):NR
Copollutant
models with:
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
September 2019
7-76
DRAFT: Do Not Cite or Quote
-------
Table 7-12 (Continued): Epidemiologic studies of exposure to ozone and pregnancy/birth—birth defects.
Study
Exposure
Copollutant
Study
Population
Assessment
Mean (ppb)
Examination
Outcomes Examined
tLinetal. (2014b)
n = 1,687 cases,
Monitor
Mean: 41
Correlation (r):
Limb defects
Taiwan
10* controls
Weeks 1-4
PM2.5: 0.52;
Syndactyly
Ozone: 2000-2007
Birth registry,
Weeks 5-8
NO2: -0.07;
SO2: 0.18;
Polydactyly
Follow-up: 2001-2007
isolated cases
Weeks 9-12
Other: -0.17
Reduction deformities of limb
Case-control study
Copollutant
models with:
NR
tJurewicz et al. (2014)
Environmental
Monitor
Mean: 23
Correlation (r):
Sperm chromosomal disomy
Poland
factors and male
90 days before
Median: 23
PM2.5: -0.41;
Ozone: NR
infertility
n = 212 men
sample collection
Maximum: 41
NO2: -0.44;
SO2: -0.26;
Follow-up: NR
Other: -0.43
Cohort study
Men attending
an infertility clinic
Copollutant
for diagnostic
models with:
NR
purposes
tFarhietal. (2014)
n =216,730
Monitor, kriging
Mean:
Correlation
Total birth defects
Israel
infants; 207,825
1st trimester
1st trimester: 32.4
(r):NR
Ozone: NR
spontaneous
conceptions,
2nd trimester: 32.7
Entire pregnancy: 32.1
Copollutant
models with:
Follow-up: 1997-2004
8,905 assistive
Median:
NR
Cohort study
reproductive
technology
conceptions
Birth records
1st trimester: 32.3
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
September 2019
7-77
DRAFT: Do Not Cite or Quote
-------
Table 7-12 (Continued): Epidemiologic studies of exposure to ozone and pregnancy/birth—birth defects.
Study Exposure Copollutant
Study Population Assessment Mean (ppb) Examination Outcomes Examined
tVinikoor-lmler et al.
21,060 cases;
Model, hierarchical
Mean:
Correlation (r):
Total birth defects
(2015)
1,401,611
Bayesian model
Texas: 40.3
NR
Texas, U.S.
controls
CMAQ and monitor
NBDPS: 37.2
Copollutant
Ozone: NR
Birth defect
1st trimester
Median:
models with:
registry
Texas: 40.5
NR
Follow-up: 2002-2006
NBDPS: 34.9
Case-control study
75th:
Texas: 46.8
NBDPS: 43.4
Maximum:
Texas: 65.1
NBDPS: 62.3
tHwanq et al. (2015)
n = 1,087 cases;
Monitor
Mean: 44.53
Correlation (r):
Ventricular septal defects
Taiwan
10,870 controls
1st trimester
Median: 44.14
NR
Atrial septal defects
Ozone: NR
Birth records
Copollutant
models with:
Patent ductus arteriosus
Follow-up: 2001-2007
NR
Pulmonary artery and valve
Case-control study
Tetralogy of Fallot
Transposition of the great arteries
Conotuncal defects
tZhana et al. (2016)
n = 105,988
Central site
Mean: 37
Correlation (r):
Congenital heart defects
Wuhan, China
births
monitor, nearest
75th: 54
NO2: -0.12;
Ventricular septal defect
Ozone: 2010-2012
8-h avg
SO2: -0.16;
Other: -0.2
Tetralogy of Fallot
Follow-up: June 10,
2011-June 9, 2013
1st trimester
Copollutant
models with:
Cohort study
NO2, SO2, CO
September 2019
7-78
DRAFT: Do Not Cite or Quote
-------
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
NBDPS
n = 7,035 cases;
4,697,523 total
live births
Model, CMAQ
downscaler
8-h max
5-10 weeks
Mean:
Correlation
All oral clefts: 40.9
(r):NR
Cleft lip with/without palate:
40.9
Copollutant
models with:
Cleft palate: 40.7
NR
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
All oral clefts
Cleft palate
Cleft lip with/without palate
fRecent studies evaluated since the 2013 Ozone ISA.
September 2019
7-79
DRAFT: Do Not Cite or Quote
-------
Table 7-13 Epidemiologic studies of exposure to ozone and pregnancy/birth—infant and fetal mortality.
Exposure
Copollutant
Effect Estimates
Study
Study Population
Assessment
Mean (ppb)
Examination
95% CI
tHwanq et al. (2011)
n = 9,325 stillbirths;
Monitor
Mean: 35.93
Correlation (r):
Stillbirth (OR)
Taiwan
93,250 controls
1st trimester
Maximum: 61.27
NO2: -0.31;
1st trimester: 1.01 (0.96, 1.06)
Birth records
SO2: 0.13;
2nd trimester: 0.96 (0.91, 1.01)
Other: 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
with: NR
tMoridi et al. (2014)
n = 148 cases;
Monitor
Mean:
Correlation (r): NR
Spontaneous abortion before 14 weeks
Tehran, Iran
148 controls
1st trimester
22.29-28.88
Copollutant models
of pregnancy (OR): 2.43 (1.72, 3.42)
Ozone: NR
with: NR
Follow-up: June
2010-February 2011
Case-control study
tGreen et al. (2015)
n = 13,999 stillbirths;
Monitor
Mean: 48.48
Correlation (r):NR
Stillbirth (OR)
California, U.S.
3,012,270 livebirths
1st trimester
Median: 47.27
Copollutant models
1st trimester: 1.00 (0.98, 1.02)
Birth records
75th: 55.52
with: NR
2nd trimester: 1.01 (0.99, 1.03)
third pregnancy: 1.03 (1.01, 1.05)
Follow-up: 1999-2009
Entire pregnancy: 1.01 (0.99, 1.04)
Cohort study
tArrovo et al. (2016)
n = 470 weeks
Monitors
Mean: 18
Correlation (r):
Late fetal death (<24 h)
Madrid, Spain
All live singleton
Other
Median: NA
75th: NA
Maximum: 38
PM2.5: NR;
NO2: NR;
SO2: NR
Copollutant models
Only reports statistically significant
Ozone: 2001-2009
Follow-up: 2001-2009
births in Madrid
results, examined exposure during
each week of pregnancy
Week 24: 1.33 (1.32, 1.35)
Time-series study
with: NR
September 2019
7-80
DRAFT: Do Not Cite or Quote
-------
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
tMendola etal. (2017)
U.S.
Ozone: NR
Follow-up: 2002-2008
Cohort study
Consortium on Safe
Labor
n = 223,375 singleton
pregnancies
Recruited from
12 centers
(19 hospitals) across
the U.S.
Model, modified
CMAQ
Other
Lag-1:
Lag-
Lag-
Lag-
Lag-
Lag-
Lag-
Mean:
entire pregnancy:
29.3
1st trimester: 29
Lag-0: 29.9
30
30.1
30.1
30.1
30.1
30
29.9
Median:
entire pregnancy:
28.5
1st trimester: 29.2
75th:
entire pregnancy:
32.7
1st trimester: 35.2
95th:
entire
pregnancy:7.8
1st trimester: 12.3
Correlation (r): NR
Copollutant models
with: NR
17.9
17.8
17.7
17.7
17.7
17.7
17.8
17.8
Maximum:
entire pregnancy:
46.4
1st trimester: 48.7
Lag-0
Lag-1
Lag
Lag
Lag
Lag
Lag
Lag
Stillbirth (OR)
Entire pregnancy: 1.53 (1.06,
Asthma: 1.29 (0.76, 2.20)
No asthma: 1.33 (0.89, 1.97)
1st trimester: 1.14 (1.00, 1.31
2.19)
Lag
0
1
07
(0.97,
1.20)
Lag
1
1
07
(0.96,
1.19)
Lag
2
1
13
(1.01,
1.26)
Lag
3
1
11
(1.00,
1.24)
Lag
4
1
10
(0.99,
1.22)
Lag
5
1
18
(1.06,
1.31)
Lag
6
1
15
(1.03,
1.27)
Lag
7
1
12
(1.01,
1.25)
September 2019
7-81
DRAFT: Do Not Cite or Quote
-------
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
Other
Mean: 32
Median: 24
Maximum: 3,316
Correlation (r): NR
Copollutant models
with: 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)
tHa et al. (2017b)
Michigan and Texas, U.S.
Ozone: 2005-2009
Follow-up: 2005-2009
Cohort study
Longitudinal
Investigation of
Fertility and the
Environment
n = 343
Couples trying to get
pregnant followed
through pregnancy
Model, modified
CMAQ
24-h avg
Other
Median: 25
75th: 29.5
Maximum: 42.6
Correlation (r):
PM2.5: -0.25;
NO2
SO2
-0.42;
-0.04
Copollutant models
with: 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
SO2
-0.698;
-0.468;
Other: -0.499
Copollutant models
with: NR
Stillbirth (OR):
Entire pregnancy: 0.85 (0.71, 1.02)
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)
fRecent studies evaluated since the 2013 Ozone ISA.
September 2019
7-82
DRAFT: Do Not Cite or Quote
-------
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 et al. (2011)
n = 4,277 infants
Monitor
Mean: 43.9
Correlation (r):
Apnea (OR)
Atlanta, GA, U.S.
Apnea Center of
8-h max
Median: 39.6
PM2.5: 42;
Lag 0-1: 1.03 (0.99,
1.07)
Ozone: August 1,
1998-December 31,
Children's
90th: 78
Maximum:
130.8
NO2: 0.45;
SO2: -0.11;
Other: 0.48
Copollutant models with:
NR
Bradycardia (OR)
2002
Healthcare of
Lag 0-1: 1.04 (1.02,
1.06)
Follow-up: August 1,
1998-December 31,
2002
Atlanta at Egleston,
children at high risk
for cardiorespiratory
Cohort study
events
tConeus and Spiess
(2012)
Nationally representative
sample, Germany
Ozone: 2002-2007
Follow-up: 2002-2007
Cohort study
German
Socioeconomic
Panel (SOEP)
Monitor
Other
NR
Correlation (r): NR
Copollutant models with:
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)
Southern California, U.S.
Ozone: 1980-2009
Follow-up: 2007-2009
Cohort study
Testing Responses Monitors, inverse
on Youth (TROY)
n = 768
College students
distance squared
weighing, 4 within 50 km
24-h avg
0-5 yr:
Mean: 23.1
Maximum: 41.
Correlation (r):
PM2.5: 0.09;
NO2: 0.09;
SO2: NR;
Other: 0.18
Copollutant models with:
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)
CHARGE
Monitor
Correlation (r): NR
Autism (OR)
California, U.S.
n = 524 mother-
8-h max
Copollutant models with:
1st trimester: 1.05 (0.97, 1.20)
Ozone: 1997-2008
child pairs
Other
NR
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)
September 2019
7-83
DRAFT: Do Not Cite or Quote
-------
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: -0.33;
SO2: NR;
Other: -0.55
Copollutant models with:
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
tFuertes et al. (2013a)
Canada, Sweden,
Germany
Ozone: NR
Follow-up: recruitment
1994-1999
Cohort study
Traffic, Asthma, and
Genetics (TAG),
includes Canadian
Asthma Primary
Prevention Study
(CAPPS), the Study
of Asthma, Genes,
and the
Environment
(SAGE), the
Children, Allergy,
Milieu, Stockholm,
Epidemiological
(BAMSE) survey,
the Prevention and
Incidence of Asthma
and Mite Allergy
(PIAMA) study,
German Infant
Nutritional
Intervention
(GIN I plus) study,
and the Lifestyle
related factors,
Immune System
and the
development of
Allergies in
Childhood
(LISAplus) study
n = 15,299 children
Six birth cohorts
from Canada and
Europe
Model, APMoSPHERE
(1 x 1 km)
For 2001
Correlation (r):
PM2.5: -0.18;
NO2: -0.25;
Other: -0.15
Copollutant models with:
NR
Allergic rhinitis (OR)
0.83 (0.59, 1.17)
Aeroallergen sensitization (OR)
0.90 (0.66, 1.25)
September 2019
7-85
DRAFT: Do Not Cite or Quote
-------
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
The traffic, asthma
and genetics (TAG)
study
n = 5,115
Metacohort:
combination of
several cohorts:
Canadian Asthma
Primary Prevention
Study (CAPPS),
Study of Asthma,
Genes, and the
Environment
(SAGE), Children,
Allergy, Milieu,
Stockholm,
Epidemiological
Survey (BAMSE),
German Infant
Nutrition
Intervention (GINI)
study plus
environmental and
genetic influences
on allergy, Influence
of Life-Style Factors
on the Development
of the Immune
System and
Allergies in East and
West Germany plus
the influence of
traffic emissions and
genetics (LISA),
Prevention and
Incidence of Asthma
and Mite Allergy
(PIAMA) study
Model for European
populations,
APMoSPHERE (Air
Pollution Modelling for
Support to Policy on
Health and
Environmental Risks in
Europe); monitor for
Canadian populations
using inverse distance
weighting of the closest
three ambient monitors
(within 50 km)
1st yr of life
Mean: 19
Maximum: 28
Correlation (r):
N02: -0.19
Copollutant models with:
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)
September 2019
7-86
DRAFT: Do Not Cite or Quote
-------
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
tFuertes et al. (2013b)
Germany
Ozone: NR
Follow-up: recruitment
1995-1998
Cohort study
GINIplus and
LISAplus
n = 6,604 children
Birth cohort full term
normal weight
APMoSPHERE models,
only 2001
Other
Mean: 22
Median: 22
75th: 23
Maximum: 30
Correlation (r): NR
Copollutant models with:
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)
Childhood Autism
Monitor
Mean:
NR
Correlation (r): NR
Autism Spectrum Disorder (OR)
California, U.S.
Risk from Genetics
Entire pregnancy
Copollutant models with:
<41.8 ppb ozone, CG/GG SNP:
Ozone: 1997-2009
and the
Environment Study
NR
reference
<41.8 ppb ozone, CC SNP: 1.0 (0.59,
Follow-up:
n = 252 cases;
1.9)
Case-control study
156 controls
>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)
tLinetal. (2014a)
Taiwan Birth Cohort
Monitor
Mean:
31-39
Correlation (r): NR
No associations between ozone
Taiwan
Pilot Study
Other
Copollutant models with:
exposure at any time period
Ozone: NR
Follow-up: October
n = 511 mother-
NR
(1st trimester, 2nd and 3rd trimesters,
child pairs
birth—12 mo, or 13-18 mo) and
neurodevelopmental scores (gross
2003-January 2004,
motor, fine motor, language, and social-
recruitment
personal)
Cohort study
tOrione et al. (2014)
n = 20 cases;
Monitor
Mean:
NR
Correlation (r): NR
Juvenile dermatomyositis (OR)
Sao Paulo, Brazil
56 controls
Entire pregnancy
Copollutant models with:
No association with ozone for entire
Ozone: NR
Cases from
NR
pregnancy exposures, some
Follow-up: August
2011-August 2012
Pediatric
inconsistent associations with trimester
Rheumatology Unit
of the Children's
specific exposures (e.g., elevated OR
for middle fertile exposure in
Case-control study
Institute
1st trimester)
September 2019
7-87
DRAFT: Do Not Cite or Quote
-------
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
tvan Rossem et al.
(2015)
Boston, MA, U.S.
Ozone: NR
Follow-up: recruited April
1999—July 2002
Cohort study
Project Viva
n = 1,131 mother-
infant pairs recruited
from eight urban
and suburban
observation offices
Monitor
Other
Median: 15-24
Correlation (r): NR
Copollutant models with:
NR
Increases in newborn systolic blood
pressure with 1st and 2nd trimester
ozone increases (13.0 and 12.8 ppb,
respectively), and decreases with 3rd
trimester exposure (13.6 ppb).
Decreases in blood pressure with
exposure lagged from birth up to
90 days.
tMalmqvist et al. (2015)
Scania, Sweden
Ozone: NR
Follow-up: 1999-2013
Cohort study
Skane study Monitor
(1999-2005), Better 1st trimester
Diabetes Diagnosis
n = 262 cases;
682 controls
Median: 26.5
75th: 30.6
Correlation (r): NR
Copollutant models with:
NR
Type I diabetes
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)
tHuana et al. (2015)
Taiwan
Ozone: 2004-2006
Follow-up: 2005
Cohort study
Taiwan Birth Cohort
Study
n = 16,686
Monitor, kriging
1st trimester
Mean: 27.9
Median: 27.5
75th: 32.5
Correlation (r): NR
Copollutant models with:
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:
(0.86, 1.47)
1.13
September 2019
DRAFT: Do Not Cite or Quote
-------
Table 7-14 (Continued): Epidemiologic studies of exposure to ozone and developmental effects.
Copollutant Effect Estimates
Study Study Population Exposure Assessment Mean (ppb) Examination 95% CI
tBreton et al. (2016)
California, U.S.
Ozone: NR
Follow-up: Kindergarten
(2002-2003) through
age 11
Cohort study
Percentage AluYb8 methylation (OR):
0.94 (0.82, 1.08)
2nd trimester
Left CIMT: 0.00 (-0.00, 0.00)
Right CIMT: 0.00 (-0.00, 0.00)
Systolic BP: 0.05 (-0.33, 0.43)
Diastolic BP: -0.04 (-0.32, 0.24)
LINE 1: 0.05 (-0.08, 0.18)
Percentage AluYb8 methylation: 0.95
(0.83, 1.10)
3rd Trimester
Left CIMT: -0.00 (-0.00, 0.00)
Right CIMT: -0.00 (-0.00, 0.00)
Systolic BP: 0.05 (-0.39, 0.48)
Diastolic BP: 0.07 (-0.25, 0.39)
LINE 1: 0.15 (0.00, 0.31)
Percentage AluYb8 methylation: 1.02
(0.87, 1.19)
Children's Health
Study
n =459
Children enrolled in
kindergarten or
first grade from
public schools
Monitors, IDW2, four
within 50 km
Other
Mean: NR
Median: NA
75th: NA
Correlation (r):
PM2.5: 0.41;
NO2: 0.01;
Other: 0.7
Copollutant models with:
NR
1st trimester
Left CIMT (A mm): -0.00 (-0.00, 0.00)
Right CIMT: -0.00 (-0.00, 0.00)
Systolic BP (A mm Hg): -0.14 (-0.53,
0.25)
Diastolic BP: -0.15 (-0.43, 0.13)
LINE 1 (A methylation): -0.20 (-0.32,
-0.07)
tNishimura et al. (2016) GALA II Monitor
Chicago, IL; Bronx, NY; n = 1,032 asthma 8-h max
Houston, TX; San cases
Francisco Bay Area, CA;
and Puerto Rico, U.S.
Ozone: NR
Follow-up: 2006-2011
Cohort study
Correlation (r): NR Atopic status (OR)
Copollutant models with: 1st year of life: 1.74 (1.23, 2.46)
NR
September 2019
7-89
DRAFT: Do Not Cite or Quote
-------
Table 7-14 (Continued): Epidemiologic studies of exposure to ozone and developmental effects.
Study
Study Population
Exposure Assessment
Copollutant
Mean (ppb) Examination
Effect Estimates
95% CI
tKimetal. (2017)
California, U.S.
Ozone: 1999-2010
CHARGE
n = 158 cases;
147 controls
Monitor
8-h max
Correlation (r): NR
Copollutant models with:
NR
Autism (OR)
Entire pregnancy + 1 st two yr of life:
1.34 (0.89, 2.01)
Follow-up: 2002-2011
Case-control study
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 with:
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
CHARGE Monitor
n = 346 ASD cases; 8-h max
260 typical
development
controls
Median: 17 Correlation (r):
PM2.5: -0.463;
NO2: -0.425;
Other: -0.022
Copollutant models with:
NR
Autism spectrum disorder (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)
tFranca et al. (2018)
Sao Paulo, Brazil
Ozone: NR
Follow-up: 2013-2014
Case-control study
n = 66 cases;
124 controls
Hospital recruits
Monitor
Other
Mean: 44 Correlation (r): NR
Copollutant models with:
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
September 2019
7-90
DRAFT: Do Not Cite or Quote
-------
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
tKerin et al. (2017)
California, U.S.
Ozone: 1998-2008
Follow-up:
Cohort study
CHARGE Monitor
n = 325 ASD cases 8-h max
Mean: 37.3
Correlation (r): NR
Copollutant models with:
NR
Neurodevelopmental (A score)
No evidence of association between
prenatal or Yr 1 ozone exposure and
any neurodevelopmental score (VABS,
MSEL, ADOS CSS)
fRecent studies evaluated since the 2013 Ozone ISA.
September 2019
7-91
DRAFT: Do Not Cite or Quote
-------
Table 7-15 Epidemiologic studies of exposure to ozone and other.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
HR (95% CI)
tLee etal. (2011)
Pittsburgh, PA, U.S.
Ozone: 1996-2001
Follow-up: 1997-2001
Cohort study
Prenatal Exposures and
Preeclampsia Prevention
Study
n = 1,696 women
Enrolled in clinics and
private practices early in
pregnancy
Model, space-time
ordinary kriging
Other
Mean: 29.9
Median: 30.3
75th: 34.3
95th: 41.6
Maximum: 51.4
Correlation (r):
PM2.5: 0.5;
NO2: 0;
SO2: 0.1;
Other: 0.5
Copollutant
models with: 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
Prenatal Exposures and Model, space-time Mean: 22.7
Preeclampsia Prevention
Study
n = 1,684 women
Enrolled in clinics and
private practices early in
pregnancy
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
Other: 0.7
Copollutant
models with: 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)
September 2019
7-92
DRAFT: Do Not Cite or Quote
-------
Table 7-15 (Continued): Epidemiologic studies of exposure to ozone and other.
tMannisto et al. (2015b)
U.S.
Ozone: 2006
Follow-up: 2006
Cohort study
Consortium on Safe
Labor
n = 500
Random sample from
cohort
Model, modified
CMAQ
Other
Mean: 41.4
Median: 42.2
Maximum: 68.3
Correlation (r): NR
Copollutant
models with: NR
Per 10% increase in Ozone
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)
September 2019
7-93
DRAFT: Do Not Cite or Quote
-------
Table 7-15 (Continued): Epidemiologic studies of exposure to ozone and other.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
HR (95% CI)
tMannisto et al. (2015a)
Consortium on Safe
Model, modified
Mean: 29.9
Correlation (r):
Cardiovascular events during labor and
U.S.
Labor
CMAQ
Maximum: 79.8
PM2.5: -0.1;
delivery.
Ozone: NR
n = 223,502 singleton
24-h avg
NO2: -0.35;
SO2: -0.18;
Other: -0.3
Odds ratios below the null for
Follow-up: 2002-2008
pregnancies
Recruited from
Other
exposures at lag Days 5, 6, and 7. No
association with exposures for lag
Cohort study
12 centers (19 hospitals)
across U.S.
Copollutant
models with: NR
Days 0 to 4
tMichikawa et al. (2016)
Kyushu-Okinawa District,
Japan
Ozone: NR
Follow-up: 2005-2010
Cohort study
Japan Perinatal Registry Monitor
Network other
n = 40,573
Mean: 41.3
Median: 39.9
Correlation (r):
PM2.5: 0.17;
NO2
SO2
-0.16;
-0.09
Copollutant
models with: 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
Other
Mean: 4
Correlation (r):
NO2: 0.088
Copollutant
models with: NR
No association between placental flow
index, placental vascularization index,
or placental vascularization flow index
with 1 week of ozone exposures
tMichikawa et al. Japan Perinatal Registry Monitor
(2017b) Network Other
Kyushu-Okinawa District, n = 821 cases
Japan
Ozone: NR
Follow-up: 2005-2010
Other study
Mean: 41.1
Median: 40.2
75th: 51.2
90th: 62.6
Correlation (r): NR
Copollutant
models with: NR
No association with placental abruption
with daily ozone exposure, lags of 1 to
5 before event
September 2019
7-94
DRAFT: Do Not Cite or Quote
-------
Table 7-15 (Continued): Epidemiologic studies of exposure to ozone and other.
Study
Study Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
HR (95% CI)
tHa et al. (2017a)
U.S.
Ozone: NR
Follow-up: 2002-2008
Other study
Consortium on Safe
Labor
n =680
12 U.S. clinical sites
Model, modified
CMAQ
24-h avg
Other
Mean: 30.32
75th: 37.85
Maximum: 54.47
Correlation (r): NR
Copollutant
models with: 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
Japan Perinatal Registry Monitor
Network other
n = 23,782
Mean:
1st trimester: 41.3
2nd trimester: 42
3rd trimester: 41.6
Median:
1st trimester: 40.2
2nd trimester: 41
3rd trimester: 40.2
75th:
1st trimester: 47.9
2nd trimester: 48.3
3rd trimester: 50.4
Correlation (r): NR
Copollutant
models with: NR
Fetal heart rate false positives (OR):
0.98 (0.92, 1.05)
Fetal heart rate false positives (OR):
1.01 (0.95, 1.08)
Fetal heart rate false positives (OR):
1.04 (0.99, 1.10)
fRecent studies evaluated since the 2013 Ozone ISA.
September 2019
7-95
DRAFT: Do Not Cite or Quote
-------
7.4.2
Toxicological Studies
Table 7-16 Study specific details from studies of ozone (O3) 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 Long-evans
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
September 2019
7-96
DRAFT: Do Not Cite or Quote
-------
Table 7-16 (Continued): Study specific details from studies of ozone (O3) 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 b|ooc| g|UCose and serum
n = 3 females, n = 3 males per treatment group ^ree acids (GD 21)
Age' adult Fetal 9rowth parameters (body
weight, length, percent lean
mass, percent 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)
fRecent studies evaluated since the 2013 Ozone ISA.
September 2019
7-97
DRAFT: Do Not Cite or Quote
-------
Table 7-17 Study specific details from studies of ozone (O3) and developmental effects.
Study
Species (Strain), n, Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
tCarev et al. (2011)
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 one 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. (2011)
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
sm 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)
September 2019
7-98
DRAFT: Do Not Cite or Quote
-------
Table 7-17 (Continued): Study specific details from studies of ozone (O3) and developmental effects.
Study
Species (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)
Long-Evans rats
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.
0.8 ppm, 4 h/day, 2 days
Glucose tolerance testing, BALF
immune cells, metabolic function
indicators
tDve et al. (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
September 2019
7-99
DRAFT: Do Not Cite or Quote
-------
Table 7-17 (Continued): Study specific details from studies of ozone (O3) and developmental effects.
Exposure Details
(Concentration, Duration)
Endpoints Examined
0, 0.4, or 0.8 ppm, 4 h/day for 2 days
(GD 5, GD 6)
BALF (GD 21)
Dam blood glucose and serum
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, percent lean
mass, percent 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)
Study
Species (Strain), n, Sex, Age
tMiller et al. (2017)
Rats (LE)
n = 0 males, 10/group females
Age: adult
fRecent studies evaluated since the 2013 Ozone ISA.
September 2019
7-100
DRAFT: Do Not Cite or Quote
-------
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.
Study
Study Population
Exposure Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
HR (95% CI)
tLimetal. (2012)
Seongbuk-Gu, South Korea
Ozone: 2008-2010
Follow-up: August 2008-August
2010
Other study
n = 560
Older adults
Nearest monitor
8-h max
Mean: 48.1
Median: 44
Maximum: 140
Correlation (r):
PM2.5: NR;
NO2
SO2
-0.15;
-0.18;
Other: CO: -0.30
Copollutant models
with: NR
Factor 3—affective
symptoms: 1.07 (0.83, 1.38)
Factor 2—somatic
symptoms: 1.25 (0.90, 1.74)
Factor 1—emotional
symptoms: 1.58 (1.16, 2.14)
fRecent studies evaluated since the 2013 Ozone ISA.
September 2019
7-101
DRAFT: Do Not Cite or Quote
-------
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
HR (95% CI)
tChiu and Yana (2015)
Taipei, Taiwan
Ozone: 2006-2011
Follow-up: 2006-2011
Case-crossover study
TNHIP
n = 13,676
Random sample of
enrollees
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;
Other: CO: -0.22
Copollutant models
with: yes
Outpatient visit (migraine),
>23°C: 1.08 (1.02, 1.15)
Outpatient visit (migraine),
<23°C: 1.28 (1.19, 1.38)
tXu etal. (2016)
Xi'an, China
Ozone: 2013-2014
Follow-up: 2013-2014
n = 20,368
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;
Other: -0.454 ozone
Epilepsy outpatient visit:
0.98 (0.95, 1.00)
Time-series study
Copollutant models
with: yes—NO2, SO2
tLinares et al. (2017)
Madrid, Spain
Ozone: 2001-2009
HMS
n = 1,175
27-monitor avg
24-h avg
Mean: 18.21
Maximum:
45.59
Correlation (r):NR
Copollutant models
with: NR
Dementia-related hospital
admission: 1.29 (1.12,
1.51)
Follow-up: 2001-2009
Time-series study
tCulqui etal. (2017)
Madrid, Spain
Ozone: 2001-2009
HMS
n = 1,183
27-monitor avg
24-h avg
Mean: 18.21
Median: 18.21
Correlation (r): NR
Copollutant models
with: NR
Alzheimer's-related hospital
admission: (NR, not
statistically significant)
Follow-up: 2001-2009
Time-series study
September 2019
7-102
DRAFT: Do Not Cite or Quote
-------
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
HR (95% CI)
tGuoetal. (2018)
Guangzhou, China
Ozone: 2013-2015
Follow-up: 2013-2015
Time-series study
Daily avg of 36 monitors
8-h max
Mean: 49.81
Maximum:
125.89
Correlation (r): NR
Copollutant models
with: 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
with: 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
EDMUS
n = 1,783
Relapse occurrence in
registry
ADMS-Urban Air Dispersion
model
24-h avg
Mean: 44.29
Median: 42.55
Maximum:
112.71
Correlation (r):
N02: -0.06;
Other: -0.21
Copollutant models
with: 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)
tLee et al. (2017)
Seoul, South Korea
Ozone: 2002-2013
Follow-up: 2002-2013
NHIS-NSC
n = 314
27-monitor avg
8-h avg
Mean: 24.2
Correlation (r): NR
Copollutant models
with: NR
Parkinson's disease
hospital admission: (NR,
figure only: no statistically
significant associations)
Case-crossover study
tLimetal. (2012)
Seongbuk-Gu, South Korea
Ozone: 2008-2010
Follow-up: August 2008-August
2010
Other study
n = 560
Older adults
Nearest monitor
8-h max
Mean: 48.1
Median: 44
Maximum: 140
Correlation (r):
PM2.5: NR;
NO2: -0.15;
SO2: -0.18;
Other: CO: -0.30
Copollutant models
with: NR
Factor 3—affective
symptoms: 1.07 (0.83,
1.38)
Factor 2—somatic
symptoms: 1.25 (0.90,
1.74)
Factor 1—emotional
symptoms: 1.58 (1.16,
2.14)
September 2019
7-103
DRAFT: Do Not Cite or Quote
-------
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
HR (95% CI)
tCho etal. (2015)
Seoul, South Korea
Ozone: 2005-2009
Follow-up: 2005-2009
Time-series study
HIRA
n = 2,320
27-monitor avg
24-h avg
Mean: 18
Median: 16
90th: 31
Correlation (r): PM2.5:
NR
Copollutant models
with: NR
ED visit for panic attack:
1.11 (1.05, 1.17)
tChen etal. (2018)
Shanghai, China
Ozone: 2013-2015
Follow-up: 2013-2015
Time-series study
SHIS
n = 39,143
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;
Other: -0.22
Copollutant models
with: NR
Mental disorder hospital
admission: 1.01 (0.96,
1.07)
Mental disorder hospital
admission: 1.03 (0.95,
1.12)
tOudin et al. (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
with: 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)
September 2019
7-104
DRAFT: Do Not Cite or Quote
-------
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
HR (95% CI)
tSzvszkowicz et al. (2016)
Nine urban areas, Canada
Ozone: April 2004-December
2011
Follow-up: April 2004-December
2011
Case-crossover study
NACRS
n = 118,602
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
with: 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)
fRecent studies evaluated since the 2013 Ozone ISA.
September 2019
7-105
DRAFT: Do Not Cite or Quote
-------
Table 7-20 Epidemiologic studies of long-term exposure to ozone and cognitive/behavioral effects.
Study
Study
Population
Exposure
Assessment
Mean (ppb)
Copollutant
Examination
Effect Estimates
HR (95% CI)
tGatto etal. (2014)
Los Angeles, U.S.
Ozone: 2000-2006
Follow-up: 2000-2006
Cross-sectional study
BVAIT, WISH,
ELITE
n = 1,496
3 RCTs
Monitor within 5 km of
residence, average of
monitors within 100 km
(IDW)
8-h max
Mean: NR—geographic
variability in
concentration shown in
Figure 2
Maximum: >25
Correlation (r):
PM2.5: -0.62;
NO2: -0.77;
Copollutant models
with: NR
Global cognition (>49 ppb vs. <34
[reference]): -0.08 (-0.45, 0.28)
Semantic memory (>49 ppb vs. <34
[reference]): -0.12 (-0.5, 0.26)
Verbal learning (34-49 ppb vs. <34
[reference]): -0.13 (-0.41, 0.16)
Visual processing (34-49 ppb vs. <34
[reference]): -0.18 (-0.43, 0.07)
Verbal learning (>49 ppb vs. <34
[reference]): -0.2 (-0.63, 0.23)
Visual processing (>49 ppb vs. <34
[reference]): -0.2 (-0.59, 0.18)
Executive function (34-49 ppb vs. <34
[reference]): -0.23 (-0.68, 0.22)
Executive function (>49 ppb vs. <34
[reference]): -0.66 (-1.35, 0.03)
Visual memory (>49 ppb vs. <34
[reference]): 0.01 (-0.42, 0.44)
Global cognition (34-49 ppb vs. <34
[reference]): 0.05 (-0.19, 0.29)
Semantic memory (34-49 ppb vs. <34
[reference]): 0.08 (-0.17, 0.33)
Visual memory (34-49 ppb vs. <34
[reference]): 0.12 (-0.16, 0.4)
Logical memory (>49 ppb vs. <34
[reference]): 0.24 (-0.21, 0.68)
Logical memory (34-49 ppb vs. <34
[reference]): 0.31 (0.01, 0.6)
September 2019
7-106
DRAFT: Do Not Cite or Quote
-------
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
HR (95% CI)
tClearv et al. (2018)
Nation-wide, U.S.
Ozone: 2001-2008
Follow-up: 2004-2008
NACC
Alzheimer's
Disease Center
participants
HBM to combine
monitored and
predicted (CMAQ)
concentrations
8-h max
Mean: NR (figure only)
Correlation (r): NR
Copollutant models
with: NR
Cognitive decline on MMSE and
CDR-SB with ozone exposure among
those with no baseline impairment
tKioumourtzoalou et al. NHS
(2017) n= 41,844
48 continental states, U.S. women
Ozone: 1996-2008
Follow-up: 1996-2008
Cohort study
Summer average
(May-Sept) of up to
five monitors (at least
one monitor within
50 km (IDW), at
residential address
Other
Mean: 31.9
Correlation (r): NR
Copollutant models
with: NR
Depression onset (depression
diagnosis): 1.00 (0.92, 1.08)
Depression onset (antidepressant or
depression): 1.06 (1.00, 1.12)
Depression onset (use of
antidepressant medication): 1.08 (1.02,
1.14)
fRecent studies evaluated since the 2013 Ozone ISA.
September 2019
7-107
DRAFT: Do Not Cite or Quote
-------
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
HR (95% CI)
tJuna et al. (2014)
National, Taiwan
Ozone: 2000-2010
Follow-up: 2000-2010
Case-control study
LHID2000-NHIRD
n = 97,627
Annual avg of three
Mean:
nearest monitors within 92.64
25 km of grid cell
(IDW), assigned to
postal code of
residence
8-h max
Maximum:
137.65
Correlation (r):
N02: -0.05;
SO2: 0.27;
Other: CO: 0.10;
PM10: -0.26
Copollutant models
with: yes
Alzheimer's disease
Baseline ozone
concentration: 1.06 (1.00,
1.12)
Change in ozone
concentration at follow-up
minus concentration at
baseline: 2.84 (2.67, 3.01)
tKirrane et al. (2015)
North Carolina and Iowa, U.S.
Ozone: 2002-2006
Enrollment: 1993-2005
Follow-up: 1997-2010
Case-control study
AHS
North Carolina:
n = 104 cases;
29,612 controls
Iowa: n = 195 cases;
53,024 controls
Farmer pesticide
applicators and their
spouses
Annual, seasonal
(April-October) 4-yr
avg of daily predictions
using measured
concentrations and
CMAQ
8-h max
Mean: 40.6 Correlation (r):
PM2.5: -0.15 to 0.06,
depending on metric
Copollutant models
with: NR
Parkinson's disease (Iowa
4-yr avg): 0.46 (0.13, 1.69)
Parkinson's disease (Iowa
warm season average):
0.46 (0.11, 1.84)
Parkinson's disease
(North Carolina 4-yr avg):
1.49 (0.43, 5.16)
Parkinson's disease
(North Carolina warm
season average): 2.60
(0.94, 7.24)
tChen et al. (2017a)
National, Taiwan
Ozone: 2000-2013
Follow-up: 2000-2013
Case-control study
TNHIP-NHIRD
n = 249 cases;
497 controls
<40 yr
Monthly average during
follow-up in areas
where participants
reside
Mean: NR
Correlation (r): NR
Copollutant models
with: NR
Parkinson's disease: 1.10
(0.74, 1.48)
September 2019
7-108
DRAFT: Do Not Cite or Quote
-------
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
HR (95% CI)
tChen etal. (2017c)
Ontario, Canada
Ozone: 1994-2013
Follow-up: April 2001-March 2013
Cohort study
ONPHEC
n = 2,066,639
55-85 yr at
enrollment
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
with: NR
(multipollutant only)
Dementia: 0.97 (0.94,
1.00)
tWu etal. (2015)
Multicity, Taiwan
Ozone: 2007-2010
Follow-up: 2007-2010
Case-control study
Hospitals and clinics
n = 1,060 cases;
4,240 controls
<60 yr
Spatiotemporal model,
cumulative annual
average
Mean: NR
Correlation (r): NR
Copollutant models
with: NR
Alzheimer's disease
(20.20-21.56 vs. <20.20
[reference]): 0.60 (0.33,
1.09)
Vascular dementia
(>20.20-21.56 vs. <20.20
[reference]): 0.62 (0.28,
1.38)
Alzheimer's disease
(>21.56 vs. <20.20 ppb
[reference]): 2.00 (1.14,
3.50)
Vascular dementia
(>21.56 vs. <20.20 ppb
[reference]): 2.09 (1.01,
4.33)
tLee etal. (2016)
NHIRD
QBME spatio-temporal Mean 26.1
Correlation (r):
Parkinson's disease
National, Taiwan
First clinic visit for PD
model
SO2: 0.01;
1.00 (0.93, 1.07)
Ozone: 1998-2009
(patients >35 yr)
CO: -0.60
Follow-up: 2007-2009
n = 11,117 cases;
4-to-1 match
Copollutant models
with: NR
Case-control study
September 2019
7-109
DRAFT: Do Not Cite or Quote
-------
Table 7-21 (Continued): Epidemiologic studies of long-term exposure to ozone and neurodegenerative diseases.
Exposure Copollutant Effect Estimates
Study Study Population Assessment Mean (ppb) Examination HR (95% CI)
tShin et al. (2018)
ONPHEC
Summer average,
Mean: 49.8 Correlation (r): NR
Parkinson's disease
Ontario, Canada
Registry record for
fused-based optimal
Copollutant models
1.06 (1.02, 1.11)
Follow-up: 2001-2013
Parkinson's disease
healthcare or
interpolation of
measured and
with: NR
1994-2013
medication
predicted ozone,
55+ yr old
21 x 21 km grid
8 h max
n = 38,745 cases
(-2.2 million followed)
tCerza et al. (2018)
Regional Health
Summer average,
Mean: 45.5 Correlation (r): NR
Parkinson's disease
Rome Italy
Information System
chemical dispersion
Copollutant models
1.04 (1.00, 1.11)
Ozone: 2008
Insurance registry
claim for Parkinson's
model with grid
resolution of 1 * 1 km
with: NO2
Follow-up: 2008-2013
disease
8-h avg
Cohort study
50+ yr
n = 1,008,253
fRecent studies evaluated since the 2013 Ozone ISA.
1
September 2019
7-110
DRAFT: Do Not Cite or Quote
-------
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
HR (95% CI)
tBecerra et al. (2013)
DDS registry
Nearest monitor,
Mean: 36.8
Correlation (r):
Autism disorder: 1.05
Los Angeles County, U.S
n = 7,594 cases;
trimester and whole
PM2.5: -0.47;
(1.01, 1.10)
Ozone: 1995-2006
75,635 controls
pregnancy averages
8-h avg
NO2: -0.34;
Other: CO:
Follow-up: 1998-2009
Diagnosis between 36
-0.55, PM10:
Case-control study
and 71 mo
-0.17, NO : -0.73
Copollutant models
with: yes
tJunq et al. (2013)
LHID2000-NHIRD
Annual average of three
Mean: 90-120
Correlation (r):
ASD: 1.59 (1.42,
National, Taiwan
n =49,073
nearest monitors within
depending on
NO2: 0;
1.78)
Ozone: 2000-2010
ASD
25 km of grid cell (IDW),
assigned to postal code
season
SO2: 0.22; Other:
PM10: 0.66
Follow-up: 2000-2010
of residence
Copollutant models
Cohort study
8-h max
with: yes
tKerin et al. (2017)
California, U.S.
Ozone: NR
Follow-up:
Case-control study
CHARGE
n = 325 cases born
1999-2007
Diagnosed with ASD
between 24 and 60 mo
Pregnancy, Yr 1 average Mean: 37.3
of up to four monitors
within 50 km (IDW) or
one monitor within 5 km
8-h max
Correlation (r):
PM2.5: -0.21;
NO2: -0.45;
Other: PM10: 0.04
Copollutant models
with: NR
ADDS-CSS: 0.99
(0.94, 1.05)
MSEL: 0.99 (0.88,
1.10)
VABS: 1.00 (0.96,
1.04)
September 2019
7-111
DRAFT: Do Not Cite or Quote
-------
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
HR (95% CI)
tVolket al. (2013)
California, U.S.
Ozone: 1997-2009
Follow-up: 1997-2008
Case-control study
CHARGE
Full-syndrome autism
(ADOS and autism
diagnostic
interview—revised.)
n = 534 cases and
controls
Diagnosed with autism
between 24 and 60 mo
1st yr, entire pregnancy,
1st trimester, 2nd
trimester, 3rd trimester
Average of four closest
monitors within 50 km
(IDW) or one monitor
within 5 km
8-h max
Mean:
NR
Correlation (r): NR
Copollutant model:
NR
Autism: 1.05 (0.84,
1.31), entire
pregnancy
tKimetal. (2017)
California, U.S.
Follow-up: 1999-2008
Ozone: 1997-2009
Case-control study
CHARGE
Confirmed autism
n = 158 cases;
147 controls
Diagnosed with autism
between 24 and 60 mo
Pregnancy, 1 st yr, 2nd
trimester
Average of four closest
monitors within 50 km
(IDW) or one monitor
within 5 km
8-h max
Mean:
NR
Correlation (r): NR
Copollutant model:
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
CHARGE
n = 297 confirmed
autism; 143 ASD;
326 controls
Diagnosed with autism
between 24 and 60 mo
1st trimester
Average of four closest
monitors within 50 km
(IDW) or one monitor
within 5 km
8-h max
Mean:
NR
Correlation (r): NR
Copollutant model:
NR
No interaction
between ozone
exposure and folic
acid intake
September 2019
7-112
DRAFT: Do Not Cite or Quote
-------
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
HR (95% CI)
tLinetal. (2014a)
11 towns, Taiwan
Follow-up: October 2003-January 2004
Cohort study
n = 511 mother-infant 1st, 2nd, 3rd trimester Mean: NR
pairs, neurodevelopment monitor average
assessed by parent
report
Correlation (r): NR No associations with
Copollutant model: ozone reported
NR
ADOS-CSS = Autism Diagnostic Observation Schedule derived-Calculated Severity Score; AHS = Agricultural Health Study; ASD = autism spectrum disorder; 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; HIRA = Health Insurance Review and Assessment Service; HMS = Hospital Morbidity Survey; LHID2000 = Longitudinal
Health Insurance Database 2000; 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; ONPHEC = Ontario Population Health and Environment Cohort; SHIS = Shanghai Health Insurance
System; TNHIP = Taiwan National Health Insurance Program; WISH = Women's Isoflavone Soy Health.
fRecent studies evaluated since the 2013 Ozone ISA.
1
September 2019
7-113
DRAFT: Do Not Cite or Quote
-------
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 etal. (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
September 2019
7-114
DRAFT: Do Not Cite or Quote
-------
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)
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
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)
Tvleret al. (2018)
Mice (C57BL/6)
n = 3-11 males, 0 females
Age: adult (8-10 week); aged
(12-18 mo)
1 ppm, 4 h
Blood-brain barrier permeability/infiltration;
microglial activation; B-amyloid accumulation
(20 h PE)
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
September 2019
7-115
DRAFT: Do Not Cite or Quote
-------
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) 0.8 ppm, 5 h/day, 1 day/week, 4 week Motor activity (measured after 1 week
n = 9-10 males, 9-10 females exposure)
Age: adult (-20 week)
Table 7-25 Study-specific details from short-term studies of neuroendocrine effects.
Study
Population
Species (Strain), N, Sex, Age
Exposure Details
(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
September 2019
7-116
DRAFT: Do Not Cite or Quote
-------
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
Akhteret al. (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)
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,
90 days
or
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,
90 days
or
p-Amyloid structure (2 h PE)
September 2019
7-117
DRAFT: Do Not Cite or Quote
-------
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
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
Rodrfquez-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
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
Akhter et al. (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 week
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 week
Motor activity
September 2019
7-118
DRAFT: Do Not Cite or Quote
-------
Table 7-28 Study effects.
Study
Population
Species (Strain), N, Sex, Age
Exposure Details
(Concentration, Duration)
Endpoints Examined
Hunter et al. (2011)
Rats (NR)
2 ppm, 3 h
Lung innervation, NGF
n =4-5 NR
production
Age: PNDs 6, 10, 15, 21, 28
Zellner et al. (2011)
Rats (F344)
n = 3, ganglia weight; n = 6-13, nerve cell count,
males and females combined
Age: PNDs 10, 15, 21, 28
2 ppm, 3 h
Neurodevelopment (nodose and
jugular sensory ganglion)
(5-23 days PE); lung innervation
7.6 Evidence Inventories—Data Tables to Summarize Cancer Study Details
7.6.1 Epidemiologic Studies
September 2019
7-119
DRAFT: Do Not Cite or Quote
-------
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
HR (95% CI)
tHvstad etal. (2013)
Nationwide, Canada
Ozone: 1975-1994
Follow-up: 1994-1997
Case-control study
NECSS
n = 2,390 cases
Spatio-temporal model;
25 x 25 km;
May-September; includes
residential history
Mean: 20.3
Maximum: 33.:
Correlation (r):
PM2.5: 0.25;
NO2: 0.11
Copollutant models
with: 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)
tGuoetal. (2016)
Nationwide, China
Ozone: 19,902,005
Follow-up: 1990-2009
Cohort study
n = 368,762 lung
cancer cases
30+ yrs old
Hybrid model from Global
Burden of Disease
Mean: 56.9
Median: 56.8
75th: 60.5
Maximum: 76.:
Correlation (r): NR
Copollutant models
with: 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)
fRecent studies evaluated since the 2013 Ozone ISA.
September 2019
7-120
DRAFT: Do Not Cite or Quote
-------
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
English Medical
Practice
n = 835,607
Age: adults, 40-89 yr,
from English medical
practices
Annual mean estimates from Mean: 25.85
dispersion model for 1-km grid Maximum: 31.5
cells linked to nearest
residential postal code
centra id
Correlation (r):
PM2.5: -0.39;
NO2: -0.46;
SO2: -0.41;
PM10: -0.40
Copollutant models
with: NR
Lung cancer: 0.66 (0.50,
0.87)
tJerrett et al. (2013)
ACS
Monthly averages calculated
Mean: 50.35
Correlation (r):
Lung cancer: 0.94 (0.89,
California, U.S.
n = 73,711
from IDW from up to four
Median: 50.8
PM2.5: 0.56;
1.00)
Ozone: 1988-2002
Follow-up: 1982-2000
Cohort study
monitors within 50 km of
residence
75th: 61
90th: 68.56
95th: 74.18
Maximum: 89.33
NO2: -0.0071
Copollutant models
with: PM2.5; NO2
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
CanCHEC
n =2,521,525
Age: 25+ yr
Model of warm season Mean: 39.6
concentration at 21-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
with: 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
ACS
n = 669,046
Age: 35+
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
with: PM2.5
Lung cancer: 0.96 (0.91,
1.00)
tCakmak et al. (2017)
Nationwide, Canada
Ozone: 2002-2009
Follow-up: 1991-2011
Cohort study
CanCHEC
n =2,291,250
Age: 25+ yr
Model of warm season
concentration at21-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
with: PM2.5
Lung cancer: 1.05 (0.97,
1.13)
Lung cancer (+ PM2.5):
1.01 (0.93, 1.09)
September 2019
7-121
DRAFT: Do Not Cite or Quote
-------
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;
Other: PM10: 0.21;
CO: 0.25
Copollutant models
with: 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)
tEckeletal. (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;
Other: PM10: 0.36
Copollutant models
with: NR
Lung cancer mortality:
1.03 (1.02, 1.03)
fRecent studies evaluated since the 2013 Ozone ISA.
September 2019
7-122
DRAFT: Do Not Cite or Quote
-------
Table 7-31 Epidemiologic studies of long-term exposure to ozone and other cancer endpoints.
Copollutant Effect Estimates
Study Study Population Exposure Assessment Mean (ppb) Examination HR (95% CI)
tBadaloni et al. (2013)
SETIL
LUR—6-yr mean
Mean: 24.2
Correlation (r): NR
Q2 vs. Q1 ozone
Nationwide, Italy
n = 620 cases
Maximum: 50.1
Copollutant models
exposure—incident
Ozone: NR
Age: children <10 yr
with: NR
leukemia: 0.88 (0.65, 1.19)
Follow-up: children born between
1998-2001
Q3 vs. Q1 ozone
exposure—incident
leukemia: 1.2 (0.87, 1.65)
Case-control study
Q4 vs. Q1 ozone
exposure—incident
leukemia: 1.1 (0.76, 1.59)
tTurner et al. (2017)
ACS
HBM with inputs from
Mean: 38.2
Correlation (r): NO2:
(Selected results; highest
Nationwide, U.S.
n = 623,048
NAMS/SLAMS and CMAQ
Median: 38.1
-0.09
and lowest magnitude
Ozone: 2002-2004
Age: 30+ yr old
75th: 40.1
95th: 44.9
Maximum: 59.3
Copollutant models
results presented)
Follow-up: 1982-2004
8-h max
with: NR
Salivary gland cancer
(n = 58): 1.70 (0.87, 3.34)
Cohort study
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)
tYaqhivan et al. (2017)
BCSC
CMAQ-HBM
Median: 36
Correlation (r): NR
Q4 vs. Q1 ozone exposure;
Nationwide, U.S.
n = 279,967
8-h max
75th: 37.9
Copollutant models
breast tissue density: 0.8
Ozone: 2001-2008
Age: 40+ yr old
with: NR
(0.73, 0.87)
Follow-up: 2001-2009
women with no history
of breast cancer
Cohort study
fRecent studies evaluated since the 2013 Ozone ISA.
September 2019
7-123
DRAFT: Do Not Cite or Quote
-------
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 etal. (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)
0.05 ppm, 24 h/day for 14 or 28 days
DNA in tail/olive tail moment
n = 12/group males, 0 females
Age: 9-10 weeks
0.05 ppm, 3 h/day for 14 and 28 days
(PE)
Micronuclei induction
(post-exposure)
Zhanq etal. (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-argenine and L-NAME.
Production of 8-oxoG/OGG1
during lung injury (baseline 4, 8,
and 12 days PE)
PE = post-exposure
September 2019
7-124
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
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. 2013b'). ToxRTool
(Klimisch et al.. 19971. STROBE guidelines (von Elm et al.. 20071. and ARRIVE guidelines (Kilkenny et al.. 20101.
September 2019
DRAFT: Do Not Cite or Quote
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.
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.b The loss or withdrawal of recruited subjects during the course
September 2019
7-126
DRAFT: Do Not Cite or Quote
-------
Table Annex 7-1 (Continued): Scientific considerations for evaluating the strength
of inference from studies on the health effects of
ozone.
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.b 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.
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 for size fractions that cannot penetrate the airway of a study animal
September 2019
7-127
DRAFT: Do Not Cite or Quote
-------
Table Annex 7-1 (Continued): Scientific considerations for evaluating the strength
of inference from studies on the health effects of
ozone.
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.
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.
September 2019
7-128
DRAFT: Do Not Cite or Quote
-------
Table Annex 7-1 (Continued): Scientific considerations for evaluating the strength
of inference from studies on the health effects of
ozone.
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.b 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,c 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.
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).
September 2019
7-129
DRAFT: Do Not Cite or Quote
-------
Table Annex 7-1 (Continued): Scientific considerations for evaluating the strength
of inference from studies on the health effects of
ozone.
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.
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.
September 2019
7-130
DRAFT: Do Not Cite or Quote
-------
Table Annex 7-1 (Continued): Scientific considerations for evaluating the strength
of inference from studies on the health effects of
ozone.
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.
a(U.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).
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).
September 2019
7-131
DRAFT: Do Not Cite or Quote
-------
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/10.1186/sl2958-015-0032-l.
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 Groea 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/klV124.
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/10.1002/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.
Bhoopalaa V: Haa 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: Cohea A: Dentener. F: Ezzati. M: Hendersoa SB: Krzvzanowski. M:
Martin. RV: VanDingenea 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/10.1021/es2025752.
Bretoa CV: Wang. X: Mack. WJ: Berhane. K: Lopez. M: Islam TS: Feng. M: Lurmann. F: McConnell. R:
Hodis. HN: Ktinzli. 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.l 161/CIRCULATIONAHA. 112.096164.
Bretoa CV: Yao. J: Millsteia 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.
September 2019
7-132
DRAFT: Do Not Cite or Quote
-------
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. https://iournals.lww.com/iphmp/Abstract/2015/03Q01/
Linking Air Pollution Data and Adverse Birth. 10.aspx.
Burnev. PG: Laitinen. LA: Perdrizet. S: Huckauf. H: Tattersfield. AE: China S: Poissoa 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: Tiepkema. M. (2017). Associations
between long-term PM2.5 and ozone exposure and mortality in the Canadian Census Health and
Environment Cohort (CANCHEC), by spatial synoptic classification zone. Environ Int 111: 200-211.
http://dx.doi.Org/10.1016/i.envint.2017.ll.030.
Capobussi. M: Tettamanti. R: Marcolin. L: Piovesaa 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: Atkinsoa RW: Kent. AJ: van Staa. T: Cook. DG: Andersoa 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/10.1164/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: Copea CE: Stephea 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. ede. gov/nchs/data/nhsr/nhsr067.pdf.
Chea 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.
Chea 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.
Chea G: Guo. Y: Abramsoa 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.
September 2019
7-133
DRAFT: Do Not Cite or Quote
-------
Chea 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.
Chea 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/10.1080/15287394.2Q15.983218.
Cho. J: Choi. YJ: Sohn. J: Suh. M: Cho. SK: Ha. KH: Kim. C: Shia 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-Bellaa AM: Bosler. O: Kessler. JP: Strube. C. (2015).
Remodeling of glial coverage of glutamatergic synapses in the rat nucleus tractus solitarii after ozone
inhalation. J Neurochem 134: 857-864. http://dx.doi.org/10.llll/inc.13193.
Clearv. EG: Cifuentes. M: Grinsteia 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: Martia 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: Diaz. 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.
Padvand. P: Rankia J: Rushtoa 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.
Pastoorpoor. M: Idani. E: Goudarzi. G: Khaniani. 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-Q692-9.
Piaz. 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.
Pve. 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.
September 2019
7-134
PRAFT: Po Not Cite or Quote
-------
Eckel. SP: Cockburn. M: Shu. YH: Deng. H: Lurmana 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: Stera 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/10.1590/1516-3180.2016.026210Q317.
Fernando Hernandez-Zimbron. L: Rivas-Arancibia. S. (2016). Syntaxin5 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/10.1155/2016/2125643.
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.16150Q.
Fuertes. E: Brauer. M: Maclntvre. E: Bauer. M: Bellander. T: von Berg. A: Berdel. D: Brunekreef. B: Chan-
Yeung. M: Gehring. U: Herbarth. O: Hoffmann. B: Kerkhof. M: Kltimper. C: Koletzko. S: Kozvrskvi. A:
Kull. I: Heinrich. J: Melen. E: Pershagea 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.
Fuertes. E: Standi. M: Cvrvs. 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/S0029-7844(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/10.1155/2014/805764.
September 2019
7-135
DRAFT: Do Not Cite or Quote
-------
Gonzalez-Guevara. E: Carlos Martinez-Lazcano. J: Custodio. V: Hernandez-Ceroa 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: Lurmana 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/10.1002/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/10.3109/08958378.2013.764946.
Gordon. CJ: Johnstone. AF: Avdia 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.
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.
Gunnisoa 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: Chea 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.ll.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: Shermaa 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.
September 2019
7-136
DRAFT: Do Not Cite or Quote
-------
Hansea C: Lubea TJ: Sacks. JD: Olshaa 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.
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. (In Press) Review of developmental origins of
health and disease publications in environmental epidemiology. Reprod Toxicol.
http://dx.doi.Org/10.1016/i.reprotox.2016.ll.011.
Hernandez-Zimbroa 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.Oil.
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: Hendersoa BH: Warner. TP: Roth. J: Kaa 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: Wea 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. Ill 1/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.Qb013e3182949ae7.
September 2019
7-137
DRAFT: Do Not Cite or Quote
-------
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.
Jeanieaa 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.
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: Beckermaa BS: Turner. MC: Krewski. D: Thurstoa G: Martia RV: van Donkelaar. A:
Hughes. E: Shi. Y: Gapstur. SM: Thua 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/10.1164/rccm.2013Q3-
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.0075510.
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: Polariska. 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: Dastoa 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/S0041-008X(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-l.
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/sl0803-017-3304-Q.
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: Giriraiaa 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. 01: 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: Bowmaa C: Davis. JA: Hoppia JA: Blair. A: Chen. H: Patel. MM: Sandler. DP: Tanner. CM:
Vinikoor-Imler. L: Ward. MH: Lubea 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.
September 2019
7-138
DRAFT: Do Not Cite or Quote
-------
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/10.10Q6/rtph.1996.1076.
Kodavanti. UP. (2016). Stretching the stress boundary: Linking air pollution health effects to a neurohormonal
stress response [Review]. BiochimBiophys 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: Thurstoa G: Calle.
EE: Thun. MJ: Beckermaa B: Deluca. P: Finkelsteia 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: Cockbura 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: Kleemaa 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: Kleemaa 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: Martia RV: Brook. JR: Crouse. PL:
Burnett. RT: Chea H: Weichenthal. S: Johnsoa 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: Dodsoa WC: Liao. D. (2010). Effect of air quality on
assisted human reproduction. Hum Reprod 25: 1317-1324. http://dx.doi.org/10.1093/humrep/dea021.
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.
Lia CC: Yang. SK: Lin. KC: Ho. WC: Hsieh. WS: Shu. BC: Chea 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/iierphllQ706827.
September 2019
7-139
DRAFT: Do Not Cite or Quote
-------
Lia 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.
Lia 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: Chea 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.
MacDormaa MF: Callaghaa 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: Melea E: Bauer. CP: Bauer. M: Berdel. D: Bergstrom. A: Brunekreef. B: Chan-
Yeung. M: Kltimper. C: Fuertes. E: Gehring. U: Gref. A: Heinrich. J: Herbarth. O: Kerkhof. M: Koppelmaa
GH: Kozvrskvi. AL: Pershagen. G: Postma. PS: Thiering. E: Tiesler. CM: Carlstea 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: Shermaa S: Ying. O: Liu. P. (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. P: 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/10.1007/sll869-014-Q268-5.
Mendola. P: Ha. S: Pollack. AZ: Zhu. Y: Seeni. I: Kim. SS: Shermaa S: Liu. P. (2017). Chronic and Acute
Ozone Exposure in the Week Prior to Pelivery Is Associated with the Risk of Stillbirth. Int J Environ Res
Public Health 14. http://dx.doi.org/10.3390/iierphl4070731.
Mendola. P: Wallace. M: Hwang. BS: Liu. P: Robledo. C: Mannisto. T: Sundaram. R: Shermaa 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.
Mendola. P: Wallace. M: Liu. P: 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.
September 2019
7-140
PRAFT: Po Not Cite or Quote
-------
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.00000000000006Q5.
Miller. CN: Dye. 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: Thompsoa LC: Schladweiler. MC: Farrai. AK:
Kodavanti. UP: Dye. 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/kfz003.
Mobasher. Z: Salam. MT: Goodwia TM: Lurmann. F: Ingles. SA: Wilsoa 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 ChemRes 20: 1687-1695.
http://dx.doi.org/10.1007/s00044-010-9493-Q.
Mokoena. ML: Harvey. BH: Vilioea 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/10.1186/1476-069X-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/SQ895-
4356(97)00019-X.
Mumaw. CL: Levesaue. S: Mcgraw. C: Robertson. S: Lucas. S: Stafflinger. JE: Campea 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.lQ96/fi.201500047.
Murgia. N: Brisman. J: Claessoa A: Muzi. G: Olin. AC: Torea 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.2012-0110QC.
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.
September 2019
7-141
DRAFT: Do Not Cite or Quote
-------
Nishimura. KK: Galanter. JM: Roth. LA: Oh. SS: Thakur. N: Nguvea 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. J. R.: Sea S: Lurmann. F: Balmes. J. R.: 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/10.1164/rccm.201302-0264QC.
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-Cintroq 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: Schistermaa EF: Ha. S: Buck Louis. GM: Shermaa 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: Mogrea 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/10.1136/bmiopen-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/10.1002/acr.22358.
Oudin. A: Astrom DO: Asplund. P: Steingrimsson. S: Szabo. Z: Carlsea 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/10.1186/sl2940-017-0348-8.
Padula. AM: Tager. IB: Carmichael. SL: Hammond. SK: Lurmana 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: Kleia 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.
Oiaa Z: Liang. S: Yang. S: Trevathan. E: Huang. Z: Yang. R: Wang. J: Hu. K: Zhang. Y: Vaugha M: Shea L:
Liu. W: Li. P: Ward. P: Yang. L: Zhang. W: Chea 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.ll.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: Oiu. 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 Zimbroa 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/fnagi.2015.00Q65.
September 2019
7-142
DRAFT: Do Not Cite or Quote
-------
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/fnmol.2017.0Q137.
Robledo. CA: Mendola. P: Yeung. E: Mannisto. T: Sundaram. R: Liu. D: Ying. O: 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/fnagi.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.
Salam MT: Millsteia 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: Hvstad. 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/dwl72.
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. Ill 1/biom. 12294.
Smith. RB: Fecht. D: Gulliver. J: Beevers. SD: Dainak. D: Blangiardo. M: Ghosh. RE: Hansell. AL: Kelly. FJ:
Andersoa 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. Toxicol Sci 164: 9-20. http://dx.doi.org/10.1093/toxsci/kfvl29.
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.
September 2019
7-143
DRAFT: Do Not Cite or Quote
-------
Stingone. JA: Lubea 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: Olshaa 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.
Szvszkowicz. M: Kousha. T: Kingsbury. M: Colman. I. (2016). Air Pollution and Emergency Department Visits
for Depression: AMulticity 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: Mclaia 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: 1324-133 l.el321. http://dx.doi.Org/10.1016/i.fertnstert.2012.ll.037.
Thomson. EM: Vladisavlievic. D: Mohottalage. S: Kumarathasaa 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. 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/10.1164/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: Milligaa ED: Campen. MJ.
(2018). Aging Exacerbates Neuroinflammatory Outcomes Induced by Acute Ozone Exposure. Toxicol Sci.
http://dx.doi.org/10.1093/toxsci/kfvQ14.
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-95/001F). 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] (pp. 1-166). (EPA/630/P-03/001F). 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.
September 2019
7-144
DRAFT: Do Not Cite or Quote
-------
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://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/b5d8alce9b07293485257375007Q12b7/eele280e77586de
985257b65005d37e7!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.
van Rossem. L: Rifas-Shiman. SL: Mellv. SJ: Kloog. I: Luttmann-Gibsoa H: Zanobetti. A: Coull. BA:
Schwartz. JD: Mittleman. MA: Oken. E: Gillmaa 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.
Veninga. TS. (1967). Toxicity of ozone in comparison with ionizing radiation. Strahlentherapie 134: 469-477.
Verheia 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: Mever. RE: Lubea 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: Lubea 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: Gotzsche. 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.
September 2019
7-145
DRAFT: Do Not Cite or Quote
-------
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: Pre/ant. D.I (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: Sua 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: Faa 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: eO 161992.
http://dx.doi.org/10.1371/iournal.pone.0161992.
Xu. X: Ha. S: Kaa 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: Sua 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.
Yaghivaa 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/sl3058-017-0828-3.
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-ll-52.
Yang. S: Tan. Y: Mei. H: Wang. F: Li. N: Zhao. J: Zhang. Y: Oian. 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: Oian. Z: Liang. S: Bassig. BA: Zhang. Y: Hu. K: Xu. S: Pong. 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.
Zhang. J: Merialdi. M: Piatt. LP: Kramer. MS. (2010). Pefining 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/fphvs.2017.00249.
Zhou. Y: Gilboa. SM: Herdt. ML: Lupo. PJ: Flanders. WP: 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.ll.007..
September 2019
7-146
PRAFT: Po Not Cite or Quote
-------
APPENDIX 8
ECOLOGICAL EFFECTS
Summary of Causality Determinations for Ecological Effects
This Appendix ch;ii;iclei'i/es Hie seieiiiilie e\ idenee lluii supports e;ins;ilii\
delernu ikiI imis I'mn/mie e\pnsme ;ind eenlm:ic;il effects \1rne delnils mi ilie dclci'iiiiii;ilimis ih;il ;nc new m' iv\ isal since llie l;is| ie\ ie\\ ;ne indie;iled w ith
;iii ;isiensk
Yi'!!ii;ilinn ;iihI sk in 1.ITiiis
( ;ius;ilil\ l)i-UTiiiin;ilinn
Visible foliar injury
Causal
Reduced vegetation growth
Causal
Reduced plant reproduction
Causal*
Increased tree mortality
Likely 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 belowground biogeochemical cycles
Causal
Alteration ofterrestrial community composition
Causal*
Alteration of ecosystem water cycling
Likely to be causal
8.1 Introduction
1 This Appendix evaluates the relevant scientific information on ecological effects as part of the
2 review of the air quality criteria for ozone and other photochemical oxidants and to help form the
3 scientific foundation for the review of the secondary National Ambient Air Quality Standard (NAAQS)
4 for ozone. It serves as a concise update to Chapter 9 of the 2013 Ozone ISA (U.S. EPA. 2013) and
5 Appendix 9 of the 2006 Ozone Air Quality Criteria Document [AQCD; U.S. EPA (2006)1. Numerous
6 studies on the effects of ozone on vegetation and ecosystems were reviewed in the 2013 Ozone ISA. The
September 2019
8-1
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
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 et al.. 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
•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
QJ_
O
CO
(/>
CD
<
•<
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.
8.1.1 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
September 2019 8-2 DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
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 (cite IRP; 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 magnitude of recent concentrations observed in the U.S. (as described in Appendix IV 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
1 The Clean Air Act definition of welfare includes, but is not limited to, effects on soils, water, wildlife, vegetation,
visibility, weather, and climate, as well as effects on man-made materials, economic values, and personal comfort
and well-being.
September 2019
8-3
DRAFT: Do Not Cite or Quote
-------
1 no geographic constraints for all the other endpoints evaluated (terrestrial water cycling; carbon
2 sequestration; terrestrial community composition; plant reproduction, phenology, and survival; insects
3 and other wildlife; and plant-animal signaling). In the PECOS for ecological effects, relevant study
4 designs include laboratory, greenhouse, field, gradient, open top chamber (OTC), Free-Air Carbon
5 Dioxide Enrichment (FACE), and modeling studies.
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
cycling 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
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)
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
Comparator = change in endpoint observed by unit increase in concentration of ozone in the same or in a control population;
exposure = environmental variable to which population is exposed; outcome = measurable endpoint resulting from exposure;
population = unit of study; study design = laboratory, field, gradient, open top chamber (OTC), Free-Air Carbon Dioxide
Enrichment (FACE), greenhouse, and modeling studies.
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 was evaluated (e.g., foliar injury, biomass).
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.
Terrestrial water cycling; carbon
sequestration; terrestrial community
composition; plant reproduction, phenology,
or mortality; insects, other wildlife,
plant-animal signaling
September 2019
8-4
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
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 8.5). 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 8.7). 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.
8.1.2 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.
8.1.2.1 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.
2013). Exposure methodologies such as greenhouse studies, continuous stirred tank reactors (CSTRs),
open top chambers (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.
September 2019
8-5
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
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. (1978). 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 et al.. 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.
8.1.2.1.2 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. Hogsett 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
et al.. 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 etal.. 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
September 2019
8-6
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
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 National Crop Loss Assessment Network (NCLAN) and similar
programs, the database 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
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).
September 2019
8-7
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
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 (Karnoskv et al..
1999 Morgan. 2004. 72764 Morgan. 2004. 72764). This method has been employed in Illinois
(SoyFACE) to study soybeans [Glycine /wax;((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
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
September 2019
8-8
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
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 (Oshima et al.. 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 et aL 1988). Studies of forest tree species in
national parks in the eastern U.S. (Winner etal.. 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; Karnosky 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).
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
September 2019
8-9
DRAFT: Do Not Cite or Quote
-------
1 other pollutants, soil composition, moisture, or temperature were either controlled or accounted for in the
2 analysis.
8.1.2.2 Definitions of Exposure Metrics and Indices
3 Exposure indices are metrics that quantify exposure as it relates to measured plant damage
4 (e.g., reduced growth). The details of these metrics are discussed in Section 8.13.1. In the over 60 years of
5 research, many forms of exposure metrics have been used, including 7-, 12-, and 24-hour averages. The
6 current secondary standard form of the 4th highest 8-hour max avg over 3 years is rarely reported in the
7 vegetation research.
8 The most useful metrics in vegetation research have been differentially weighted hourly
9 concentrations that are cumulative during the growth of plants. The 2013 Ozone ISA primarily discussed
10 SUM06, AOTx, and W126 exposure metrics. Below are the definitions of the three cumulative index
11 forms:
12 •
13
14 •
15
16
17 •
18
19
20
1
Wc ~ 1 + 4403e~126C
Equation 8-1
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.
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:
8.1.3 Mechanisms Governing Vegetation Response to Ozone
21 The ecological effects of ozone are observed across multiple levels of biological organization,
22 starting at the subcellular and cellular level, then to individual organisms, and finally to ecosystem-level
23 processes. The 2013 Ozone ISA summarized in detail the mechanisms for ozone's effects at the leaf level
24 (Section 9.3 of 2013 Ozone ISA). Figure 8-2 summarizes current scientific understanding of effects of
25 ozone on plant physiology at the biochemical and leaf level. These effects lead to changes in
26 photosynthesis and carbon allocation to different plant carbon pools. Carbon allocation links ozone effects
27 at the subleaf and leaf level to changes at larger scales.
September 2019
8-10
DRAFT: Do Not Cite or Quote
-------
Ethylene
H20
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 = intra-cellular 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: Permission pending, adapted from Emberson et al. (2018).
Figure 8-2 Schematic representation of the cellular and metabolic effects of
ozone on vegetation.
1 As seen in Figure 8-2. ozone ( 0,": represented in gray) enters the plant through leaf stomatal
2 openings ("gsto") during gas exchange, although some reproductive tissues are also directly affected by
3 ozone exposure (see Section 8.4). Ozone and its derivatives, referred to as reactive oxygen species
4 ("ROS"), are phytotoxic sources of oxidative stress in plants [Section 9.3.2 in U.S. EPA (2013)1. They
5 may be partially detoxified by "antioxidants" [Section 9.3.4 in U.S. EPA (2013)1; however, any
6 remaining effective ozone flux causes damage to photosynthetic machinery and results in declines in
7 "gross photosynthesis." Ozone flux into the leaf may also cause "downregulation" of RuBisCO (the
8 enzyme responsible for carbon fixation), which also results in declines in gross photosynthesis
9 [Section 9.3.5 in U.S. EPA (2013)1. Ozone exposure may cause elevated "ethylene" production, a
10 multifunctional plant hormone [Section 9.3.3 in U.S. EPA (2013)1. This ozone-induced elevated ethylene
September 2019
8-11
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
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 ("H2O"), 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.
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
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
September 2019
8-12
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
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 (Lloyd et al..
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 (Chieppa et al.. 2015).
The use of bioindicator 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
September 2019
8-13
DRAFT: Do Not Cite or Quote
-------
1 Rocky Mountain National Park, displayed similar visible injury and may also serve as a
2 bioindicator (Ncufcld et al.. 2018).
3 "Tree of heaven (Ailanthus altissima), an established invasive species found widely across the
4 U.S., has been identified as an effective ozone bioindicator species by the National Park Service
5 and Forest Service (Smith et al.. 2008; Kohut. 2007). In greenhouse exposures, foliar injury
6 occurred at 8-hour avg ozone exposure levels of 60 to 120 ppb, with greater injury corresponding
7 to higher exposures (Seiler et al.. 2014). In the field, an ambient ozone 3-month, 12-hour W126
8 value of 11.6 ppm-hour induced foliar injury (Seiler et al.. 2014).
9 In addition to these studies, a recent global-scale synthesis of published ozone exposure studies
10 documents foliar injury from ozone exposure in the field, across gradients, or in controlled ozone
11 experiments in hundreds of species (Bcrgmann et al.. 2017). In field and gradient studies involving ozone
12 concentrations in ambient air, 245 plant species from 28 plant genera experienced ozone foliar injury
13 (Bcrgmann et al.. 2017). Many of the species that experience ozone foliar injury have populations native
14 to the U.S. (see Table 8-3).
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 Macqreqor (1975)
Acer macrophyllum
Y
TemDle 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): Bunqener et al. (1999a): Bunqener
et al. (1999b)
Agrostis vinealis
Y
Haves et al. (2006)
Alchemilla sp.
Y
Manninq et al. (2002)
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)
Alnus viridis or Alnus alnobetula
Y
Vanderhevden et al. (2001): Skellv et al. (1999): Lorenz
et al. (2005): De Vries et al. (2003)
September 2019
8-14
DRAFT: Do Not Cite or Quote
-------
Table 8 3 (Continued): 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
Bergmann et al. (2017).
Species
Ozone Causes
Foliar Injury
References
Amorpha californica
Y
U.S. EPA (1980): Temple (1999)
Apocynum androsaemifolium
Y
Berqweiler and Manninq (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 dougiasiana
Y
Temple (1999): U.S. EPA (1980)
Artemisia dracuncuius
Y
Temple (1999)
Aruncus dioicus
Y
Bussotti et al. (2003a)
Asciepias californica
Y
Temple (1999)
Asciepias exaitata
Y
Chappelka et al. (2007): Souza et al. (2006)
Asciepias fascicularis
Y
Temple (1999)
Asciepias incarnata
Y
Orendovici et al. (2003)
Asciepias 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)
Bignonia 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 etal. (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 aren aria
Y
Jones et al. (2010)
Carex atrofusca
Y
Mortensen (1994b)
Carex echinata
Y
Haves et al. (2006)
September 2019
8-15
DRAFT: Do Not Cite or Quote
-------
Table 8 3 (Continued): 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
Bergmann et al. (2017).
Species
Ozone Causes
Foliar Injury
References
Carex nigra
Y
Franzarinq et al. (2000)
Centaurea spp.
Y
Bussotti et al. (2006)
Cephalanthus occidentalis
Y
Kline et al. (2008)
Cercis canadensis
Y
Kline et al. (2008)
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)
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)
Conocarpus erectus
Y
Ceron-Breton et al. (2009)
Conyza canadensis
Y
Grantz et al. (2008)
Cordyianthus rigidus
Y
Temple (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)
September 2019
8-16
DRAFT: Do Not Cite or Quote
-------
Table 8 3 (Continued): 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
Bergmann et al. (2017).
Species
Ozone Causes
Foliar Injury
References
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):
Bunqener 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)
Galium aparine
Y
U.S. EPA (1980)
Gayophytum diffusum
Y
Wahid et al. (2011)
Geum rivale
N
Battvetal. (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)
September 2019
8-17
DRAFT: Do Not Cite or Quote
-------
Table 8 3 (Continued): 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
Bergmann et al. (2017).
Species
Ozone Causes
Foliar Injury
References
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 elata
Y
Wahid et al. (2011)
Oenothera rosea
Y
Skellv et al. (1999)
Oenothera sp.
Y
Skellv etal. (1999)
Oxalis acetosella
Y
Haves et al. (2006)
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): Manninq etal. (2002)
Pectocarya heterocarpa
Y
Thompson etal. (1984)
Pectocarya piatycarpa
Y
Thompson etal. (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 etal. (1992)
September 2019
8-18
DRAFT: Do Not Cite or Quote
-------
Table 8 3 (Continued): 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
Bergmann et al. (2017).
Species
Ozone Causes
Foliar Injury
References
Pinus jeffrey
Y
Temple et al. (2005): Miller et al. (1998): Williams and
Macqreqor (1975)
Pinus lambertiana
Y
Williams and Macqreqor (1975): Williams et al. (1977)
Pinus leiophylla
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 Macqreqor (1975): Williams et al. (1977)
Pinus pungens
Y
Neufeld etal. (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):
Wiseloqel 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)
Piatanus racemosa
Y
Temple et al. (2005)
Poa pratensis
Y
Bender et al. (2002): Bender etal. (2006): Bunqeneret
al. (1999b): Bunqener et al. (1999a): Mortensen (1992):
Ashmore et al. (1996)
Poiygonatum sp.
Y
Bussotti et al. (2006)
Popuius spp.
Y
Davis (2007a): Bussotti and Gerosa (2002): De Vries et
al. (2003)
Popuius tremuioides
Y
Volin et al. (1998): Karnoskv et al. (1999): Karnoskv et
al. (1996): Coleman et al. (1996): Yun and Laurence
(1999)
Potentiiia gianduiosa
Y
Wahid etal. (2011): U.S. EPA(1980)
Prunus emerginata
Y
Temple (1999)
Prunus pensyivanica
Y
Davis (2007a)
September 2019
8-19
DRAFT: Do Not Cite or Quote
-------
Table 8 3 (Continued): 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
Bergmann et al. (2017).
Species
Ozone Causes
Foliar Injury
References
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): 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)
Rhamnus spp.
Y
Bussotti and Gerosa (2002)
Rhizophora mangle
Y
Ceron-Breton et al. (2009)
Rhus aromatica
Y
Kline et al. (2008)
Rhus copallina
Y
Davis and Orendovici (2006): Davis (2009)
Rhus typhina
Y
Wan et al. (2013): Wan et al. (2014)
Ribes spp.
N
Temple (1999)
Robinia pseudoacacia
Y
Skellv et al. (1999): Wana 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)
Rubusidaeus
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)
September 2019
8-20
DRAFT: Do Not Cite or Quote
-------
Table 8 3 (Continued): 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
Bergmann et al. (2017).
Species
Ozone Causes
Foliar Injury
References
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)
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): Manninq et al. (2002): Blum et al. (1997)
Sambucus spp.
Y
Bussotti and Gerosa (2002)
Sassafras albidum
Y
Chappelka et al. (1999b): Davis and Orendovici (2006):
Davis (2011)
Scirpus cespitosus
Y
Haves et al. (2006)
Scrophularia nodosa
Y
Bussotti et al. (2003a)
September 2019
8-21
DRAFT: Do Not Cite or Quote
-------
Table 8 3 (Continued): 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
Bergmann et al. (2017).
Species
Ozone Causes
Foliar Injury
References
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 (1993)
Solidago 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
Battvetal. (2001)
Stachys spp.
Y
Bussotti et al. (2006)
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): Beramann et al. (1999): Reilina
and Davison (1992): Beramann 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): Mannina et
al. (2002)
Vaccinium uliginosum
Y
De Vries et al. (2003): Gerosa and Ballarin-Denti (2003)
September 2019
8-22
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Table 8 3 (Continued): 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
Bergmann et al. (2017).
Species
Ozone Causes
Foliar Injury
References
Verbesina occidentalis
Y
Chappelka et al. (2003)
Vernonia noveboracensis
Y
Orendovici et al. (2003)
Viburnum nudum
Y
Beramann et al. (2017): Davis (2007a): Davis (2007b)
Viburnum spp.
Y
Bussotti and Gerosa (2002): Manninq et al. (2002)
Vicia californica
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 (Matvssek et al.. 2006; Panek. 2004; Grulke et al.. 2003; Panek and Goldstein. 2001;
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
September 2019
8-23
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
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
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.
September 2019
8-24
DRAFT: Do Not Cite or Quote
-------
Table 8-4 Ozone exposure and foliar injury.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Foliar Injury
Grantz et al. (2013)
Greenhouse; Kearney Gossypium
Research and Extension barbadense. (Pima
Center, Parlier, CA cotton)
(36.598°N, 119.503°W)
Each plant was exposed to a single
15-min pulse of O3 (0.0, 0.5, 1.0, 1.5,
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 average 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.
September 2019
8-25
DRAFT: Do Not Cite or Quote
-------
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.
September 2019
8-26
DRAFT: Do Not Cite or Quote
-------
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 1st 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.
September 2019
8-27
DRAFT: Do Not Cite or Quote
-------
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 +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.
September 2019
8-28
DRAFT: Do Not Cite or Quote
-------
Table 8-4 (Continued): Ozone exposure and foliar injury.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Foliar Injury
Seileret al. (2014)
Greenhouse, field; Penn Ailanthus altissima
State Russell E. Larson (tree of heaven)
Agricultural Research
Center, Rock Springs,
PA
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; 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;
|jmol/mol = micromoles/mole; W126 = cumulative integrated exposure index with a sigmoidal weighting function; W126-3 mo = the running maximum 3-month, cumulative 12-hour
W126 weighted value.
September 2019
8-29
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
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. Biomass declines were reported to be greater
(11 to 17%) with elevated ozone exposures [average 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 [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 ozone levels 1.5 times
ambient overthe period from 1998-2008 [Ambient W126 2.1-8.8 ppm-hour, Elevated
12.7-35.1 ppm-hour; Moran and Kubiske (2013)1.
September 2019
8-30
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
• 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
intra-specific 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 (Pinus ponderosa) 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 (Bergmann
et al.. 2017). Many of these species have populations native to the U.S., and a comprehensive list of U.S.
species identified by Bergmann et al. (2017) as sensitive to ozone are presented below in Table 8-5.
September 2019
8-31
DRAFT: Do Not Cite or Quote
-------
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).a b
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 etal. (1999): Rebbeck (1996a): Laurence
et al. (1996): Kress and Skellv (1982): Noble etal. (1992): Tioelkeret
al. (1993)
Achillea millefolium
N
Bender et al. (2002): Bunaener et al. (1999a): Bunaeneret al.
(1999b)
Agropyron smithii
Y
Volin etal. (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 etal. (1996)
Carex aren aria
Y
Jones et al. (2010)
Carex atrofusca
Y
Mortensen (1994b)
Carex echinata
N
Haves et al. (2006)
Carex nigra
N
Franzarinq etal. (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)
September 2019
8-32
DRAFT: Do Not Cite or Quote
-------
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 (2017).ab
Species
Ozone
Reduces
Growth
References
Comarum palustre
Y
Battvetal. (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
Battvetal. (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):
Chappelka 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)
September 2019
8-33
DRAFT: Do Not Cite or Quote
-------
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 (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 Macqreqor (1975): Williams et
al. (1977)
Pinus pungens
N
Neufeld etal. (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): Shaferetal. (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 etal. (2006): Bunqener et al. (1999b):
Bunqener et al. (1999a): Mortensen (1992): Ashmore et al. (1996)
Polygonum viviparum
Y
Mortensen and Nilsen (1992)
Popuius deitoides
Y
Wanq etal. (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-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 etal. (2000):
Bussotti and Gerosa (2002): Yuska et al. (2003)
Pseudotsuga menziesii
Y
Runeckles and Wriqht (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): 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)
September 2019
8-34
DRAFT: Do Not Cite or Quote
-------
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 (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): 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)
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): Berqmann et al. (1995): Berqmann et al.
(1999): Berqmann et al. (1996a)
Solidago albopilosa
Y
Mavitv and Berranq (1993)
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): Berqmann et al. (1999): Reilinq and Davison
(1992): Berqmann 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.
September 2019
8-35
DRAFT: Do Not Cite or Quote
-------
8.3.2
Changes in Biomass Allocation
1 In addition to declines in plant growth rates, ozone alters patterns of carbon allocation, both
2 belowground and aboveground (the portion of energy expended by the plant toward roots, stems, or
3 leaves; see Figure 8-2). Changes in biomass allocation alter plant nutrient uptake, plant water use, and
4 carbon fixation.
5 • Over the course of the Aspen FACE experiment (1998-2008), the effects of ozone on plant
6 carbon allocation were dynamic through time and varied among the forest communities (Talhclm
7 et al.. 2014; Pregitzer and Talhelm. 2013; Rhea and King. 2012). Elevated ozone consistently
8 suppressed leaf production in each of the three communities. There were effects on root biomass
9 in 2006 consistent with Aspen FACE studies of previous years, with elevated ozone increasing
10 small root (0-2 mm diameter) biomass in the aspen-only rings and decreasing small root biomass
11 in the aspen-birch rings (Rhea and King. 2012). There were also effects of ozone on the
12 distribution of roots across the soil profile, which are discussed in more detail in Section 8.9.2.
13 • Shifts in wood anatomy (change in growth, cell size, vessel density, and proportion) also occurred
14 with elevated ozone at the Aspen FACE site (Kostiainen et al.. 2014). Elevated ozone
15 significantly decreased radial growth and diameters of wood fibers and vessels in quaking aspen.
16 Most treatment responses were observed in the early phase of the experiment, indicating
17 ontogenetic changes during wood maturation that are consistent with shifts in the trees' metabolic
18 priority from growth to hydraulic transport in response to ozone.
19 • A study of the effects of short-term ozone exposure on loblolly pine seedlings found positive
20 effects on aboveground growth, but the study authors attribute this finding to reduction in
21 photosynthate transport to roots, which contributed to declines in seedling vigor (Chieppa et al..
22 2015). Even with the increased aboveground growth observed, ozone alterations to carbon
23 transport and subsequent declines in seedling vigor and longevity may have negative impacts on
24 forest establishment and regeneration.
8.3.3 Connections with Community Composition and Water Cycling
25 Studies published since the 2013 Ozone ISA have provided insight on ozone-mediated alterations
26 to biomass allocations within an individual plant that are relevant to whole-plant growth and function.
27 Additionally, the studies provide context for scaling up the long-known detrimental effects of ozone on
28 photosynthesis and growth in numerous plant species to changes at the community and ecosystem level.
29 While outside the scope of this assessment, decreases in photosynthesis due to ozone are well studied and
30 quantified and are directly related to declines in plant biomass discussed here. Ozone-caused declines in
31 canopy density and leaf area index, an important component of plant biomass, have similarly been well
32 studied but are outside the scope of the current assessment (see Section 8.1.1.). These effects were,
33 however, thoroughly reviewed in the 2013 Ozone ISA (U.S. EPA. 2013). and studies continue to be
34 published in this area (U.S. EPA. 2008).
September 2019
8-36
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
• 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
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
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.
September 2019
8-37
DRAFT: Do Not Cite or Quote
-------
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 Littler 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.
September 2019
8-38
DRAFT: Do Not Cite or Quote
-------
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 yrofthe 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.
September 2019
-39
DRAFT: Do Not Cite or Quote
-------
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. 1st
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 (2*) 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 2* O3 increased significantly compared with CF
and NF seedlings and had greater aboveground and
total dry matter yield than CF seedlings.
September 2019
-40
DRAFT: Do Not Cite or Quote
-------
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. 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) (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.
September 2019
8-41
DRAFT: Do Not Cite or Quote
-------
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 ofW126
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.
September 2019
8-42
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
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 evidence was not sufficient to determine causality (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).
September 2019
8-43
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
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 8-4). Furthermore, compared with
ambient air (an average of 33 ppb across all studies), all metrics of plant reproduction were
reduced under elevated ozone (Figure 8-6). For instance, in experiments that used
charcoal-filtered air as the control, seed number decreased 16% (at an average exposure of
70 ppb), and fruit number decreased 9% (at an average exposure of 90 ppb). In experiments that
used ambient air as the control, average fruit weight decreased 51% (at an average 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 average 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.
September 2019
8-44
DRAFT: Do Not Cite or Quote
-------
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)
(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)
-100 -50 0 50 100
Percentage change from charcoal-filtered air (%)
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: Permission pending Leisner and Ainsworth (2012).
Figure 8-3 Meta-analysis of the effects of ozone exposure (relative to
charcoal-filtered air) on plant reproduction.
September 2019
8-45
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
FLN
1 ~ 1
¦
(28, 56 ppb)
Y
-
(164, 79 ppb)
HI
(—•—I
¦
(39, 58 ppb)
FN
1 ~ 1
-
(51. 85 ppb)
FW
1 • 1
¦
(38, 98 ppb)
SF
1—#-H
¦
(29, 74 ppb)
SN
-
(21, 68 ppb)
WP
1—•—1
-
(68, 77 ppb)
sw
h*H
-
(158, 73 ppb)
i ¦ i
-80 -60 -40 -20 0 20
Percentage change from ambient air (%}
HI = harvest index; FLN = flower number; FN = fruit number; FW = fruit weight; 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: Permission pending, Leisner and Ainsworth (2012).
Figure 8-4 Meta-analysis of the effects of ozone exposure (relative to
ambient air) on plant reproduction.
• 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).
• 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
September 2019
8-46
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
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% maximum 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).
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).
September 2019
8-47
DRAFT: Do Not Cite or Quote
-------
8.4.3
Plant Mortality
1 Several new studies examined the impact of ozone exposure on plant mortality (i.e., the fraction
2 of individuals in population that die over a given interval). All were focused on tree species, and the study
3 results are consistent with previous evidence showing that ozone can affect tree mortality (Table S-S).
4 • In the Aspen FACE experiment in Rhinelander, WI, the survival of sensitive aspen (Populus
5 tremuloides) genotypes 271 and 259 declined significantly between 1997 and 2008 under
6 elevated ozone concentrations relative to ambient conditions (Moran and Kubiske. 2013). In
7 contrast, the survival of tolerant genotype 8L increased 14.8% under elevated ozone. Genetically
8 based differences in ozone sensitivity could have implications for intra-specific diversity and
9 evolution of wild populations (see Section 8.4.1).
10 • Dietze and Moorcroft (2011) conducted a large-scale analysis of factors contributing to annual
11 mortality of trees in functional types (each characterized by different species) in the forests of the
12 eastern and central U.S. The U.S. Forest Service's FIA database (http://www.fia.fs.fed.us/.
13 version 2.1) was queried for data on tree mortality, and the analysis only included trees that were
14 measured in consecutive censuses. Overall, ozone was ranked 9th on the list of 13 factors that
15 forests were sensitive to, and ozone's effect was similar in magnitude to that of precipitation.
16 Mortality in eight out of ten plant functional types was significantly correlated with ozone 8-hour
17 max: seven experienced increasing mortality with increasing ozone exposure, while late
18 successional conifers showed a slight decrease in mortality with increasing ozone exposure.
19 Evergreen hardwoods were the functional type most sensitive to increasing ozone; they showed
20 annual mortality ranging from 1% in areas of the country with relatively low ozone to 3% in areas
21 of relatively high ozone. Assuming no replacement, a change in mortality rate from 1 to 3%
22 would shift the time to 50% loss of a species from 69 to 24 years. Such changes in mortality are
23 consistent with documented changes in community composition (Section 8.10).
8.4.4 Summary
24 Ozone exposure can affect plant reproduction. Over 100 studies of cultivated and noncultivated
25 species have now been synthesized qualitatively and quantitatively. They show that diverse metrics of
26 plant reproduction decline under ozone concentrations occurring either in the environment or under
27 experimental conditions within an order of magnitude of recent concentrations. The biological
28 mechanisms underlying ozone's effect on plant reproduction are twofold. They include both direct
29 negative effects on reproductive tissues and indirect negative effects that result from decreased
30 photosynthesis and other whole-plant physiological changes. Two metrics of plant reproduction, fruit
31 number and fruit weight, show greater reductions under increased ozone when combined across species
32 for ozone concentrations that span 40 to >100 ppb; other metrics do not show such reductions or do so
33 across a narrower range of ozone concentrations. An exposure-response and a dose-response curve
34 developed for legume species in Europe both show significant negative effects of ozone on plant
35 reproductive biomass. Finally, experimental ozone exposure at multiple experimental settings
36 (e.g., in vitro, whole plants in the laboratory, whole plants and/or reproductive structures in the
37 greenhouse, and whole plant communities in field settings) convincingly show ozone independently
September 2019
8-48
DRAFT: Do Not Cite or Quote
-------
1 reduces plant reproduction. Therefore, previous evidence and new evidence reviewed here is sufficient to
2 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).
September 2019
8-49
DRAFT: Do Not Cite or Quote
-------
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. AX9.5.4.4. AX9.5.5.1,
AX9.6.2.4, AX9.6.2.5, AX9.6.4.2;
U.S. EPA (2013). Section 9.4.3.3
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
for some metrics of reproduction and
not well resolved for others; exposure-
response and dose-response curve
exists for a set of legume species
Leisner and Ainsworth (2012).
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 a 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.
AX9.6.2.5, AX9.6.4.2;
U.S. EPA (2013). Section 9.4.3.3
AX9.5.4.4, AX9.5.5.1,
September 2019
8-50
DRAFT: Do Not Cite or Quote
-------
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. AX9.6.2.2.
U.S. EPA (2013). Section 9.4.7.1
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.
U.S. EPA (2013). Section 9.4.7.1
AX9.6.2.3;
Limited evidence from experimental
studies that isolate the effect of ozone
on tree mortality
The Aspen FACE study shows sensitive
genotypes have increased mortality with ozone
exposure compared with a tolerant genotype.
Moran and Kubiske (2013),
U.S. EPA (2013). Section 9.4.7.1
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.
September 2019
8-51
DRAFT: Do Not Cite or Quote
-------
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 etal. (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 etal. (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.
September 2019
8-52
DRAFT: Do Not Cite or Quote
-------
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). For the 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%.
September 2019
8-53
DRAFT: Do Not Cite or Quote
-------
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.
(2015)
Greenhouse; Close
House,
Northumberland, U.K.
Lycopersicon Pollen and ovule experiments:
esculentum control = charcoal-filtered air
(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.
September 2019
8-54
DRAFT: Do Not Cite or Quote
-------
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 etal. (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
four 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.
September 2019
8-55
DRAFT: Do Not Cite or Quote
-------
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 average of 35 ppb at
stands 1,040 m 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% maximum 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).
September 2019
8-56
DRAFT: Do Not Cite or Quote
-------
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. (20111
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.
September 2019
8-57
DRAFT: Do Not Cite or Quote
-------
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 et al. (2016) OTC; experimental field Dehasa-type
located in a rural area
in the northeastern
Iberian Peninsula,
Tarragona (40.41°N,
0.47°E)
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/m2 was
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
coast, Spain, (40.68°N,
0.78°E)
Briza maxima O3 as AOT40 index—ozone:
charcoal-filtered (mean
O3 <10 ppb, AOT40 = 0); Ambient
(mean 63 <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.
September 2019
8-58
DRAFT: Do Not Cite or Quote
-------
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
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).
September 2019
8-59
DRAFT: Do Not Cite or Quote
-------
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
PI ant ago
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.
September 2019
-60
DRAFT: Do Not Cite or Quote
-------
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
average 37.1 ppb O3
(AOT40 = 1.6 ppm-h); elevated
average 99.2 ppb
(AOT40 = 20.4 ppm-h); highly
elevated average 145.2 ppb
(AOT40 = 36.4 ppm-h); Ozone
exposure during growth period of
petunia: Ambient average 40.0 ppb
O3 (AOT40 = 4.0 ppm-h); elevated
average 96.0 ppb
(AOT40 = 25.0 ppm-h); highly
elevated average 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/meters squared; nmol/mol = nanomoles/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.
September 2019
8-61
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
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 (U.S. EPA. 2013). 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 (U.S. EPA. 2013).
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
September 2019
8-62
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
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 (Lcisncr
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 a 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.
September 2019
8-63
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
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 etal.. 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 (Mcgrath 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.
September 2019
8-64
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
(D
">>
O -10
-t—'
0
S
L_
0)
O)
g
1 -20
L_
CD
CL
-30
„ 1 , 1 1 1
Soybean
Maize
i
i
l>
on
i *
n "<
LIL-L-
i i
li UHI
1 >
LA
^ir—H
T.J i" '
V ¦
1
1 O
41
nu
i y
11
¦ >
o
1)
* 1
11
1
i i
I 1
i
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 ofW126 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: Permission pending, Mcqrath et al. (2015).
Figure 8-5 Estimated percentage reduction of soybean and maize yield in the
U.S. from ozone for 1980-2011.
• 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
September 2019
8-65
DRAFT: Do Not Cite or Quote
-------
1 regional differences in temperature, water, and nutrient availability) are generally higher than
2 previously indicated by concentration-response functions derived from experimental studies.
3 Wheat yield sensitivities to ozone were statistically significant spatially along the northern U.S.
4 border, maize sensitivity was spatially statistically significant at various locations across the U.S.,
5 and soybean sensitivity was spatially statistically significant in a band from the Great Plains to
6 the south-central U.S. Crops in regions of elevated ozone and high water use, were more tolerant
7 to ozone. The PDLR model coupled with ozone and temperature projections by the Community
8 Earth System model from 2000-2050 have predicted average declines of U.S. wheat, maize, and
9 soybean of 13, 43, and 28%, respectively.
10 • A modeling study considering the cobenefits associated with decreases of NOx under the U.S.
11 EPA Clean Power Plan (to regulate emissions of CO2) estimated the effects on total production
12 and biomass loss of four U.S. crops (potatoes, soybean, cotton, maize) under three policy
13 scenarios and a reference (ambient air) scenario for the year 2020 (Capps et al.. 2016). In this
14 analysis, the CMAQ model was used to model exposure values of W126 and then apply these to
15 crop distribution maps using published data to estimate biomass loss and potential productivity
16 loss (PPL). At ambient ozone concentrations, modeled production loss is greatest for potatoes,
17 soybean, and cotton, with these losses ranging from 1.5 to 1.9%. Scenario 1 (which is closest to
18 current levels) results in an ozone impact reduction of <2% for each crop. Reductions in PPL of
19 8.4% (soybean) and 6. 7% (cotton) in Scenario 2 (which is most similar to the final Clean Power
20 Plan) and 6.6 and 3.8% in Scenario 3 (most stringent policy option) suggest that reduction in NOx
21 with CO2 regulation will decrease agricultural yield losses associated with ozone.
8.5.3 Summary
22 The relationship between ozone exposure and reduced crop yield is well established in the
23 scientific literature and continues to be an area of active research with hundreds of papers on this topic
24 published since the 2013 Ozone ISA in the U.S. and other countries (U.S. EPA. 2009). There is a
25 considerable amount of new research on major U.S. crops, especially soybean, wheat, and non-soybean
26 legumes, including updated soybean exposure-response curves. Meta-analyses published since the 2013
27 Ozone ISA provide further supporting evidence that ozone decreases growth and yield of wheat and
28 affects reproductive and developmental plant traits important to crop yield. Recent advances in
29 characterizing ozone's effects on U.S. crop yield include further geographic and temporal refinement of
30 ozone sensitivity and national-scale estimates of maize and soybean losses from ozone based on actual
31 yield data. A few studies on grassland species add to the existing body of evidence in the 2013 Ozone ISA
32 for ozone effects on nutritive quality. New information is consistent with the conclusions of the 2013
33 Ozone ISA that the body of evidence is sufficient to infer a causal relationship between ozone
34 exposure and reduced yield and quality of agricultural crops.
September 2019
8-66
DRAFT: Do Not Cite or Quote
-------
Table 8-10 Ozone and crop yield and quality.
Study
Study Type
and Location
Study Species
Ozone Exposure
Effects on Crop Yield
Burkev etal. (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.
Betzelberqer 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 etal. (2012)
Other;
California
(36.6°N,
119.5°W)
Saccharum sp.
(sugarcane)
Four hybrids
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.
September 2019
8-67
DRAFT: Do Not Cite or Quote
-------
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; Glycine max
SoyFACE, (soybean)
Champaign, IL u 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 average
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-, or 24-
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
September 2019
8-68
DRAFT: Do Not Cite or Quote
-------
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 Os 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.
September 2019
8-69
DRAFT: Do Not Cite or Quote
-------
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; Glycine max
SoyFACE, (soybean)
Champaign, IL
(40.04°N,
88.24°W)
Elevated O3 fumigation system increased
O3 to 100 ppb from 10:00 a.m. to 5:00 p.m.
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. Average 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 63 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 (average peak 49 ppb).
Mean monthly 2x ambient was 37 to
56 ppb (average 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.
September 2019
8-70
DRAFT: Do Not Cite or Quote
-------
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 average 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.
September 2019
8-71
DRAFT: Do Not Cite or Quote
-------
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.
AQS = (U.S. EPA) Air Quality System (database); AOT40 = seasonal sum of the difference between an hourly concentration at the threshold value of 40 ppb, minus the threshold
value of 40 ppb; 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/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.
September 2019
8-72
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
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.
September 2019 8-73 DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
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) and L. 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..
2016). 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).
September 2019
8-74
DRAFT: Do Not Cite or Quote
-------
1 Reproduction:
2 • In B. tabaci, exposure to elevated ozone (72 vs. 37 ppb) decreased fecundity (Cui et al.. 2016b;
3 Cui etal.. 2012). However, at higher levels (238 vs. 50 ppb), fecundity increased (Hong et al..
4 2016). In L. dispar, exposure to 1.5x ambient levels decreased fecundity (Couture and Lindroth.
5 2012).
6 • Egg laying in the diamondback moth (Plutella zylostella) was significantly higher in the absence
7 of ozone (when given a choice between artificial leaves fumigated with plant volatiles mixed with
8 clean air or elevated ozone [80 ppb; Li and Blande (2015); see Section 8.71. In the same lab
9 study, plants exposed to herbivore-damaged neighbor plants had more eggs deposited on them at
10 ambient ozone (10 ppb) than plants exposed to undamaged control plants. In presence of 80 ppb
11 ozone, the preference for egg-laying on damaged plants was lost. Under field conditions,
12 P. zylostella laid more eggs on plants exposed to control levels (10 ppb) compared with elevated
13 ozone [30-80 ppb, Giron-Calva et al. (2016)1.
14 • In B. tabaci, adults preferred control plants for oviposition [37 vs. 72 ppb, Cui et al. (2014)1.
15 Survival:
16 • In I', brassicae, there was a nonsignificant trend whereby larval mortality tended to increase with
17 increasing ozone levels [15-20 ppb, 70, 120 ppb; Khaling et al. (2015)1. Elevated ozone (50 and
18 150 ppb vs. 0.5 ppb) increased mortality in Metopolophium dirhodum aphids (Telesnicki et al..
19 2015).
20 • In L. dispar, survival of early instars decreased in response to feeding on leaves exposed to
21 elevated ozone [ 1.5 x ambient; Couture and Lindroth (2012)1.
22 • 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
23 Changes in host plant quality resulting from elevated ozone can alter the population density and
24 structure of associated insect herbivore communities, ultimately affecting ecosystem processes
25 (Cornelissen. 2011). In the 2013 Ozone ISA, these population- and community-level responses included
26 altered population growth rates in aphids (Menendez et al.. 2010; Awmack et al.. 2004). reduced total
27 arthropod abundance at the Aspen FACE site (Hillstrom and Lindroth. 2008). and changes in genotypic
28 frequencies of aphids over multiple generations (Mondor et al.. 2005). Recent studies report metrics of
29 altered population and community structure (e.g., population size, relative species abundance) adding to
30 the evidence base for herbivore responses to ozone at higher levels of biological organization. New
31 studies include:
32 • In a study from Aspen FACE, elevated ozone did not consistently influence arthropod community
33 composition (Hillstrom et al.. 2014).
34 • In a mesocosm study, past ozone exposure had no effect on the richness, diversity, or evenness of
35 the arthropod community associated with the descendant plant community but did increase the
36 relative abundance of carnivore arthropods while decreasing the relative abundance of herbivore
37 arthropods (Martinez-Ghersa et al.. 2017).
September 2019
8-75
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
• 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
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, they did not provide any consistent pattern of
response across growth (Table 8-11). reproduction (Table 8-12). and mortality endpoints. 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.
September 2019
8-76
DRAFT: Do Not Cite or Quote
-------
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)
Soli dago
canadensis;
Taraxacum
officinale
50-100
Habeck and
Lindroth (2013)
Pieris brassicae
Brassica nigra
120
Khalinq 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
Honq et al. (2016)
Aphid (Rhopaio-
siphum padi)
Lolium
multiflorum
120
Ueno et al. (2016)
Aspen leaf
beetle
(Chrysomeia
crotchi)
Trembling
aspen
(Popuius
tremuioides)
50-100
T
Vique and Lindroth
(2010)°
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)°
Aphid
Broad bean
85
Dohmen (1988)a
Aphid
Broad bean
100
(>24 h)
1
Brown et al. (1992)a
September 2019
8-77
DRAFT: Do Not Cite or Quote
-------
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.
(2000V
Sawfly (Gilpinia
pallida)
Scots pine
T
Manninen et al.
(2000V5
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.
(2000V5
a1996 Ozone AQCD.
"2006 AQCD.
°2013 ISA
September 2019
8-78
DRAFT: Do Not Cite or Quote
-------
Table 8-12 Summary of studies reporting altered reproduction in herbivores.
Herbivore
Plant
Exposure (ppb)
Fecundity
Oviposition
Preference
Reference
Whitefly (Bemisia
tabaci)
Tomato
(Lycopersicon
escuientum)
72
Cui etal. (2016a)
Whitefly (Bemisia
tabaci)
Tomato
(Lycopersicon
escuientum)
72
Cui etal. (2012)
Whitefly (Bemisia
tabaci)
Tomato
(Lycopersicon
escuientum)
238
T
Hona etal. (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 etal. (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.
September 2019
8-79
DRAFT: Do Not Cite or Quote
-------
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
Hong etal. (2016). Ueno etal. (2016). Khalina et al. (2015).
Habeck and 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
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)
September 2019
8-80
DRAFT: Do Not Cite or Quote
-------
Table 8-14 Ozone exposure and effects on herbivores.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Insects and Other Wildlife
Cui etal. (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
1st 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.
September 2019
8-81
DRAFT: Do Not Cite or Quote
-------
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
Cui etal. (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)
defence-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 6. 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.
Cui etal. (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.
September 2019
8-82
DRAFT: Do Not Cite or Quote
-------
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 O3 W126 = 2.1 -8.8
ppm-h and elevated 03= 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 03W126 = 2.9-8.8
ppm-h and elevated O3 =
13.1-35.1 ppm-h. Ambient air
C02and 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.
September 2019
-83
DRAFT: Do Not Cite or Quote
-------
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 O3 W126 = 5.6, 4.9, 2.1
ppm-h and elevated O3 = 14.6,
13.1, 12.7 ppm-h. Ambient air
C02and 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 O3 = 14.6,
13.1, 12.7 ppm-h. Ambient air
CChand 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 Os W126 = 7.3, 5.6, 4.9
ppm-h and elevated O3 = 29.6,
14.6, 13.1 ppm-h. Ambient air
CChand 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.
September 2019
8-84
DRAFT: Do Not Cite or Quote
-------
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. 2nd 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.
September 2019
8-85
DRAFT: Do Not Cite or Quote
-------
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
arundina),
dallisgrass and
(Paspalum
dilatatum),
common Bermuda
grass (Cynodon
dactylon), and
white clover
(Trifolium repens)
fed to New
Zealand white
rabbits
(Oryctolagus
cuniculus)
Six OTC chambers (three
chambers per treatment).
Grasses were exposed for
8 weeks. Mean monthly 12-h
ambient was 21 -32 ppb
(average peak conc. 49 ppb).
Mean monthly 2x ambient was
37 to 56 ppb (average 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 Os
conc. increase from 20 to 30 ppb, liveweight gain predicted
to decrease by 12%.
Ueno et 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-C>3 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.
September 2019
8-86
DRAFT: Do Not Cite or Quote
-------
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 O3 W126 = 2.9-8.8
ppm-h and elevated O3 =
13.1-35.1 ppm-h. Ambient air
C02and 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.
September 2019
8-87
DRAFT: Do Not Cite or Quote
-------
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 O3, 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 O3 than 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; CN = 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.
September 2019
8-88
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
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).
September 2019
8-89
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Plant
Herbivory
Pollination
Ozone
Insect pollinators
Insect herbivores
Insect parasitoids
Pa rasitized
insect herbivores
Volatile plant signaling compounds in air
Production of volatile plant signaling compounds
Emission of volatile plant signaling compounds
Insect detection of volatile plant signaling 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 (Juergens 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
September 2019
8-90
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
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 (Fuentes 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.. 2016). 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.. 2016). 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 meters 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; Farrc-Armcngol 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. 1990). 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
September 2019
8-91
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
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. (2016bVI. In
OTCs, both ozone and herbivory by whiteflies (Bemisia tabaci) increased emissions of tomato
plant (Solanum 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 etal.. 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 \2/ 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
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,
September 2019
8-92
DRAFT: Do Not Cite or Quote
-------
1 dispersion, and lifespan of these compounds. Numerous preference studies in insects show altered plant
2 host detection, reduced pollinator attraction, and shifts in plant host preference in the presence of
3 elevated, yet environmentally relevant, ozone concentrations. Plant defense mechanisms (i.e., attraction of
4 predators and parasitoids that target phytophagous insects) were either reduced, enhanced, or unaffected
5 by elevated ozone. Considering the available evidence (i.e., the 2013 Ozone ISA and more recent research
6 efforts) and recognizing uncertainties around how chemical signaling responses observed in the
7 laboratory translate to natural environments (Table 8-13). this ISA makes a new causality determination
8 that the body of evidence is sufficient to infer a likely to be causal relationship between ozone
9 exposure and alteration of plant insect signaling.
September 2019
8-93
DRAFT: Do Not Cite or Quote
-------
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, Vuorinen et al. (2004). Pinto et al. (2007). Mcfrederick et al. (2008)
by multiple research groups show direct emission, dispersion, and lifespan of (model), Blande et al. (2010), Cui et al. (2014), Li and Blande (2015),
effects of ozone on VPSCs VPSCs Farre-Armenqol et al. (2015), Cui et al. (2016b), Fuentes et al. (2016)
(model), Khalina etal. (2016). Li et al. (2016b). Mofikova et al. (2017).
Ghimire et al. (2017)
Multiple experimental studies by multiple Altered plant host detection and insect Cannon (1990). Gate et al. (1995). Fuentes et al. (2013). Cui et al. (2014).
research groups show altered insect herbivory; reduced pollinator attraction Farre-Armenqol et al. (2015), Li et al. (2016b), Khalinq et al. (2016), Cui et
response to VPSCs in presence of ozone and altered parasitoid attraction by al. (2016b)
plants
VPSC(s) = volatile plant signaling compound(s).
September 2019
8-94
DRAFT: Do Not Cite or Quote
-------
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 30 PPb at ni9ht)for 5 daYs
oleracea) (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.
Khalinq 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; average
value from 9:00 a.m.-5:00
p.m.
2x ambient (72.2 ppb;
average 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.
September 2019
8-95
DRAFT: Do Not Cite or Quote
-------
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+ 03 (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: (average 29 ppb),
elevated: (average 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.
September 2019
8-96
DRAFT: Do Not Cite or Quote
-------
Table 8-16 (Continued): Ozone exposure and plant insect signaling.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Plant Insect Signaling
Cui et al. (2016b) OTC; Observation Insect: Bemisia tabaci Ambient (37.3 ppb; avg value Elevated O3 increased the total amount of plant VOC
Station for Global (whitefly)
Change Biology, P|ant: solanum
Beijing China lycopersicum (wild-type
(40.183°N, (om^o)
116.4°E)
from 9:00 a.m.-5:00 p.m.) for emissions 4.78-fold in the wild-type tomato plants. Whitefly
3 weeks, in both 2010 and herbivory increased the total amount of plant VOC emissions
2011 5.66-fold in the wild-type tomato plants. Production of VPSCs
2* ambient (72 2 ppb' avg was highest in the treatment of O3 + herbivory. In a
value from dual-choice Y-tube assay, adult parasitoids preferred
9 00 a m -5 00 p m ) <->3 + herbivory plants over all other treatments.
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.
September 2019
8-97
DRAFT: Do Not Cite or Quote
-------
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 + O3 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.
September 2019
8-98
DRAFT: Do Not Cite or Quote
-------
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/minute; mL/min = milliliters/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).
September 2019
8-99
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
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
September 2019
8-100
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
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 et al. (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 (ambientW126 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 et al.. 2011). Zak et al. (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 average of
13.5 ppm-hour) compared to the first 8 years of the experiment (W126 average of 27.4 ppm-hour)
(Kubiske and Foss. 2015). 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
September 2019
8-101
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
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 averages 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 Kranzberger 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.
September 2019
8-102
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
• 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.. 2015V
• 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 8.12: 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
September 2019
8-103
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
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 et al.. 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 Mvhre (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 (Kvalevag and Mvhre. 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%.
September 2019
8-104
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
• 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 or biome 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
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 further
September 2019
8-105
DRAFT: Do Not Cite or Quote
-------
1 evidence that is consistent with the conclusions of the 2013 Ozone ISA that the body of evidence is
2 sufficient to conclude that there is a likely to be causal relationship between ozone exposure and
3 reduced carbon sequestration in ecosystems.
September 2019
8-106
DRAFT: Do Not Cite or Quote
-------
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, Wl
(45.675°N,
89.625°W)
Betula papyrifera
(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
average 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%.
September 2019
3-107
DRAFT: Do Not Cite or Quote
-------
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;
Rhinelander, Wl
Betula papyrifera
(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 average
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 average
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.
September 2019
8-108
DRAFT: Do Not Cite or Quote
-------
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, Wl
(45.70°N,
89.50°W)
(1) Foliar herbivore
insects: individual
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.
Zak et 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 O3 = 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.
September 2019
3-109
DRAFT: Do Not Cite or Quote
-------
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.
September 2019
8-110
DRAFT: Do Not Cite or Quote
-------
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.
September 2019
8-111
DRAFT: Do Not Cite or Quote
-------
Table 8-17 (Continued): Ozone exposure effects on productivity and carbon sequestration.
Study
Study Type
and Location
Study Species
Ozone Exposure
Effects on Productivity
Betzelberqer 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.
September 2019
8-112
DRAFT: Do Not Cite or Quote
-------
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.
Ritteretal. (2011)
FACE;
Kranzberger
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.
September 2019
8-113
DRAFT: Do Not Cite or Quote
-------
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.
September 2019
8-114
DRAFT: Do Not Cite or Quote
-------
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
CChand 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/meters squared; NPP = net primary productivity; 03 = ozone; PLFA = phospholipid fatty acid; Pg C = petagrams (gigaton)
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.
September 2019
8-115
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
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 8.10.1). soil carbon (Section 8.10.2). and soil nitrogen
(Section 8.10.3) 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.
September 2019
8-116
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
CO., HjO
CO», HsO
Altered stomatai function
Run off
Allocation of C
retention
Litter production
and chemistry
Altered species competition
Rootsisymbionts
Root, leaf litter
exudation
Nutrients
Soil foodweb
Eacterta
Fin#
Mlsro i marco hverstrates
Organic matter
CO., release
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
exposure decreases leaf litter nutrients (Liu et al.. 2007; Kasurinen et al.. 2006). increases leaf litter
nutrients (Rodenkirchen et al.. 2009; Parsons ef 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
September 2019
8-117
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
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 Be tula papyrifera 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
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 (Decock and Six. 2012; Decock et al.. 2012).
September 2019
8-118
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
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
et al.. 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.
The 2013 Ozone ISA (U.S. EPA. 2013) found O3 alters CH4 emissions (Toet et al.. 2011; Zheng
et al.. 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.
September 2019
8-119
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
The 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 (Kancrva 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 et al.. 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).
• 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 (Decock and Six. 2012). but decreased the abundance of
September 2019
8-120
DRAFT: Do Not Cite or Quote
-------
1 multiple microbial N cycling genes in surface soils (He et al.. 2014). A model constructed using
2 SoyFACE natural abundance 15N values suggests that elevated O;, accelerated N cycling by
3 increasing both soybean belowground N allocation and N2 emissions from soil I Decock et al
4 (2012); see Figure 8-9 below].
5 • Nemissions from meadow: The 2013 ISA (U.S. EPA. 2013) found that elevated O3 emissions
6 decreased daily N2O emissions in a Finnish meadow (Kanerva et al.. 2007). At SoyFACE in
7 Illinois, elevated O3 did not alter N2O emissions, but a model suggested that O3 may affect
8 emissions (Decock et al.. 2012).
Total
gaseous N
loss
• Grain N •
Biological
N-fixation
P ant N
Soil N
Note: Dashed lines indicate decreases, thin solid lines indicate no major change, and thick solid lines represent increases.
Source: Permission pending Decock et al. (2012).
Figure 8-9 Elevated ozone effect of accelerated senescence and reduced
seed production soil N.
September 2019
8-121
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
8.9.4
Summary
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 8.10.1). soil carbon (Section 8.10.2). and soil nitrogen (Section 8.10.3).
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. EPA.
September 2019
8-122
DRAFT: Do Not Cite or Quote
-------
1 2013) showing that in forests ozone may decrease soil N content in some studies, but have no effect in
2 others. Also ozone did not affect forest root uptake of N. In crops, ozone did not affect soil N in field
3 studies and showed mixed results, depending on the N chemical species, in laboratory studies.
4 Overall, the evidence does not change the conclusions from the 2013 Ozone ISA (U.S. EPA.
5 2013). and therefore, suggests that ozone can alter soil biogeochemical cycling of carbon and nitrogen,
6 although the direction and magnitude of these changes often depends on the species, site, and time of
7 exposure. Currently, there does not appear to be a consistent exposure-response relationship. The body of
8 evidence is sufficient to conclude that there is a causal relationship between ozone exposure and the
9 alteration of belowground biogeochemical cycles.
September 2019
8-123
DRAFT: Do Not Cite or Quote
-------
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, Study in soybean (Glycine Soils for the study collected from Elevated O3 significantly increased soil
Illinois (40.056°N, max) agroecosystem soybean plots at the SoyFACE NO3", and briefly increased soil NhV early
88.201 °W) agroecosystem with ambient and in the incubation. Elevated O3 did not affect
elevated O3 (target concentration mineral N transformation rates as
of 1.23* ambient) determined by 15N tracers, and did not
affect potential CO2 or N2O emission rates.
September 2019 8-124 DRAFT: Do Not Cite or Quote
-------
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
He et al. (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 exochitanase decreased
significantly. Soil N was quantified and NH
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.
September 2019
8-125
DRAFT: Do Not Cite or Quote
-------
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 over4-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.
September 2019
8-126
DRAFT: Do Not Cite or Quote
-------
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 average
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.
September 2019
8-127
DRAFT: Do Not Cite or Quote
-------
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 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)
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 average
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.
Zak et 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 average
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).
September 2019
8-128
DRAFT: Do Not Cite or Quote
-------
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
Zaketal. (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 1st 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
(2x) ppm. Seasonal W126 were
0.033 (CF), 0.423 (NF), and
21.913 (2x)
Seedlings under 2x 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,
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 quaking aspen (Populus
tremuloides) and paper birch
(Betula papyrifera). Logs were
grown and placed in each of
the 4 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 average
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.
September 2019
8-129
DRAFT: Do Not Cite or Quote
-------
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.
Grantz et 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/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/liter; W126 = cumulative integrated exposure index with a sigmoidal weighting function.
September 2019
8-130
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
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 S-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.
September 2019
8-131
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
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.
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).
8.10.1 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
September 2019
8-132
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Bauer and Hernandez-Teieda. 2007; Wieseretal.. 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 (Kubiske 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).
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. 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).
• 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 (Stain pfli 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
September 2019
8-133
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
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).
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
herbaceaus species from experiments around the world (Bergmann et al.. 2017; van Goethem et 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 (Bergmann 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, intra-species 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:
September 2019
8-134
DRAFT: Do Not Cite or Quote
-------
1 • In a seeded pasture of leguminous clover and three grass species in Alabama, experimentally
2 elevated ozone (56 ppb ozone) increased the biomass of grass species but had no effect on clover
3 biomass (Gilliland et al.. 2016). effectively increasing the relative biomass of grass to legumes in
4 the community.
5 • In a greenhouse study of grass and forbs in competition, grass cover increased linearly with
6 elevated ozone in a 12-hour mean range of 21 to 103 ppb rHaves etal. (2011); Figure 8-111.
7 • In an experimental fumigation in a Swiss high-elevation pasture reported in Bassin et al. (2007).
8 there was no effect of 7 years of elevated ozone on relative abundance of plants grouped as forbs,
9 grasses, or sedges, but elevated ozone did increase the abundance of a dominant grass species in
10 the community (Bassin et al.. 2013).
September 2019
8-135
DRAFT: Do Not Cite or Quote
-------
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
50 100 150
Seasonal mean 03 conc. (24 h, ppb)
Source: Permission pending 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.
September 2019
8-136
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
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).
8.10.2 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
(Kanerva et al.. 2008; Dohrmann and Tebbe. 2005). peatland mesocosms (Morskv et al.. 2008). and forest
mesocosms (Pritsch etal.. 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,
September 2019
8-137
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
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
analysis (Cheng et al.. 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 (Kasurinen 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 (Ueda et al.. 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 et al.. 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,
September 2019 8-138 DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
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 (Kancrva et al.. 2008). marginally increased
fungal abundance (quantified by PLFA profiling) in peatland mesoscosms (Morsky 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:
• 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 (Kasurinen et al.. 2005). 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 (Kasurinen 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).
September 2019
8-139
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
• 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 etal.. 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).
8.10.2.3 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 methanogenMethanosaeta (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 arthopod 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
September 2019
8-140
DRAFT: Do Not Cite or Quote
-------
1
2
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
3 The 2013 Ozone ISA found the evidence sufficient to conclude that there is a likely to be causal
4 relationship between ozone exposure and the alteration of community composition of some ecosystems.
5 Evidence of this relationship was presented for forest communities of trees; grassland communities of
6 grasses, herbs, and legumes; and soil microbial communities of bacteria and fungi. Recently published
7 papers extend the evidence for each of these topics (Table 8-19 and Table 8-20).
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
reasonable confidence
Models of forest tree community composition in the
U.S.
Gustafson et al. (2013): Wanq et al.
(2016)
Grassland studies
Calvete-Soao et al. (2016): Wedlich
et al. (2012): Pavne et al. (2011)
Controlled exposure
studies (lab or small- to
medium-scale field
study)
Forest: Aspen FACE
2013 Ozone ISA (U.S. EPA. 2013):
Zak et al. (2012).
Grassland plants
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
of inference
Models of regional forest composition in the U.S.
Gustafson et al. (2013): Wanq et al.
(2016)
Global synthesis of woody and herbaceous plant
responses to controlled exposure of O3, grouped by
plant family (relevant to natural plant communities)
Berqmann et al. (2017)
Grassland plant studies at national or European
scale
Pavne et al. (2011): van Goethem
et al. (2013): U.S. EPA (2013)
September 2019
8-141
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
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
Previous U.S. EPA syntheses
2006 Ozone AQCD (U.S. EPA
2006): 2013 Ozone ISA (U.S. EPA.
2013)
Multiple studies by
multiple research
groups
Forest (studies in the U.S., Europe)
Section 8.10.1.1
Grassland (studies in the U.S., Argentina, China,
U.K., Spain, Switzerland, Europe)
Section 8.10.1.2
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
et al. (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
September 2019
8-142
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
ozone and community composition (Haves et al.. 2011; Payne et al.. 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.
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
September 2019
8-143
DRAFT: Do Not Cite or Quote
-------
1 plant communities, while confirming that the effects of ozone on soil microbial communities are diverse.
2 The body of evidence is sufficient to conclude that there is a causal relationship between ozone
3 exposure and the alteration of community composition of some ecosystems.
September 2019
8-144
DRAFT: Do Not Cite or Quote
-------
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°E)
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 detected genes. Elevated O3 reduced
May 31, 2010 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 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.
September 2019
8-145
DRAFT: Do Not Cite or Quote
-------
Table 8-20 (Continued): Terrestrial community composition response to ozone exposure.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects
He et al. (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 endoglucanse
increased significantly and xylanase, cellobiase
and exochitanase decreased significantly. Soil
N was quantified; NhU-N was significantly lower
in the surface soil and NO3-N significantly
higher in subsoil of elevated O3 plots compared
to ambient.
Li et al. (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.
September 2019
8-146
DRAFT: Do Not Cite or Quote
-------
Table 8-20 (Continued): Terrestrial community composition response to ozone exposure.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects
Chen etal. (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
Os)
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 Zhonq
(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%.
September 2019
8-147
DRAFT: Do Not Cite or Quote
-------
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 ofarbuscular
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.
September 2019
8-148
DRAFT: Do Not Cite or Quote
-------
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.
September 2019
8-149
DRAFT: Do Not Cite or Quote
-------
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.
OTC; University of
6-week-old Trifolium repens
Charcoal-filtered ambient O3 in
O3 exposure reduced the number of Rhizobium
(2017)
Buenos Aires,
(white clover) and
2010-2011 (concentrations not
nodules on clover roots by 35%.
Argentina (34.59°S,
Rhizobium spp. (N fixing
reported), elevated O3 for 4 h/day
58.58°W)
bacteria in root nodules on
over 5 days at maximum
clover)
90-120 ppb O3
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.
September 2019
8-150
DRAFT: Do Not Cite or Quote
-------
Table 8-20 (Continued): Terrestrial community composition response to ozone exposure.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects
Mina etal. (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 etal. (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.
September 2019
8-151
DRAFT: Do Not Cite or Quote
-------
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
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 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 average 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.
September 2019
8-152
DRAFT: Do Not Cite or Quote
-------
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
year 180, elevated O3 decreased productivity
by half. Elevated O3 reduced landscape
biomass.
Zak et 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 average 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).
September 2019
8-153
DRAFT: Do Not Cite or Quote
-------
Table 8-20 (Continued): Terrestrial community composition response to ozone exposure.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects
Wanqetal. (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
(2x) 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.
September 2019
8-154
DRAFT: Do Not Cite or Quote
-------
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) Os: ~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 by t-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).
September 2019
8-155
DRAFT: Do Not Cite or Quote
-------
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 63 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.
September 2019
8-156
DRAFT: Do Not Cite or Quote
-------
Table 8-20 (Continued): Terrestrial community composition response to ozone exposure.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects
Ueda etal. (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 Os)
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 average 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.
September 2019
8-157
DRAFT: Do Not Cite or Quote
-------
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 giomerata\
(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)
September 2019
8-158
DRAFT: Do Not Cite or Quote
-------
Table 8-20 (Continued): Terrestrial community composition response to ozone exposure.
Study Type and
Study Location Study Species Ozone Exposure Effects
Payne et al. (2011) Gradient; 64 acidic Entire plant community, Site-specific O3 exposures from the In terms of single factor models, O3 is the
grassland sites acidic grasslands U.K. Air Pollution Information strongest predictor of species cover. Within the
stratified by N System multiple-factor model, only current total
deposition and inorganic N deposition and mean annual total
climate, U.K. 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).
September 2019
8-159
DRAFT: Do Not Cite or Quote
-------
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) ofthe 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
dactyion)
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 average 1-h 63 = 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. Meta-analysis;
(2013)
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.
September 2019
8-160
DRAFT: Do Not Cite or Quote
-------
Table 8-20 (Continued): Terrestrial community composition response to ozone exposure.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects
Wanqetal. (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 O3 changed
soil microbial community structure of both bean
genotypes based on PCA of PLFA data.
Berqmann et al. Meta-analysis;
(2017) global;
peer-reviewed
papers, book
chapters, reports,
and conference
proceedings
published 1980 to
unspecified mid
2010s
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.
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
September 2019
8-161
DRAFT: Do Not Cite or Quote
-------
Table 8-20 (Continued): Terrestrial community composition response to ozone exposure.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects
Payne et al. (2017) Mesocosm; peat Microscopic algae
sampled from wet, (desmids, diatoms),
heath peatland, U.K. protozoa (ciliates,
flagellates, testate
amoebae), and microscopic
animal consumers
(nematodes, rotifers)
sampled from Sphagnum
papillosum stems
Experimental O3 for 3.5 yr: ambient
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)
Testate amoeba community structure was
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/hectare; Lsu rRNA = fungi-specific primer; N = nitrogen; N/h/yr = kg nitrogen/hectare/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.
September 2019 8-162 DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
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 stomata to 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.
September 2019
8-163
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
• 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 8-21).
• 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).
September 2019
8-164
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
• 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; Hoshika et al. (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 take into account 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 et al.. 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 IHoshika et al. (2015); Figure 8-131.
September 2019
8-165
DRAFT: Do Not Cite or Quote
-------
I
i
V
O
T?
rs
-
ra
r*
c
Q
&
c
J
10 20 JO 40 50 0 30 M) 90 120
Dttylinx O, (raitejKrau™ (umoJ mot'' > Canopy cumulative 0; uptake (nrniol m-' t
Note: Effects of ozone-induced stornatal 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: Permission pending 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 2008-2009). (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
1 New work examines the influence of environmental measures, inclusive of ozone exposure and
2 climate, on late-season stream flow in forests in the eastern U.S. and shows that ozone effects scale up
3 from leaf level through to ecosystem level. The 2013 Ozone ISA reviewed the work of Mclaughlin et al.
4 (2007a); Mclaughlin et al. (2007b). which used field measurements to link ozone to changes in tree sap
5 flow and scale up to the ecosystem level. Building on this. Sun et al. (2012 built empirical statistical
6 models from data collected in six watersheds in Tennessee, North Carolina, Virginia, and West Virginia
a CO. nssirniknti^
- fMMto+LOl JPHX4J 0
m >—91JB- *145 O
— J-JR1flmuw £>-0.92 TJ
0 0
0 0 0
lijff jfc O
~
a* /
a &Q a
b Transpiration
m ff-O 17
tife-o.05
m jn) Ci>2k--0 15n • 1 -24 HMl-23
X e
0 0
C Water use efficiency <¦
jPMfclla- I44JT-&.53 0
°o Q
0
_ yHlOfct-WH 0
m * 264
September 2019
8-166
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
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
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.
September 2019
8-167
DRAFT: Do Not Cite or Quote
-------
Table 8-21 Ozone exposure and water cycling.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Water Cycling
Hoshika et al. (2011)
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%.
Grantz et al. (2016)
Greenhouse; Parlier, CA
(36.60°N, 119.50°W)
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.
September 2019
8-168
DRAFT: Do Not Cite or Quote
-------
Table 8-21 (Continued): Ozone exposure and water cycling.
Study Type and
Study Location Study Species Ozone Exposure Effects on Water Cycling
Paudel et al. (2016) Greenhouse; Parlier, CA Amaranthus palmeri Two runs of exposure 30 and Elevated O3 exposure and water stress had no effect
(36.60°N, 119.50°W) (Palmer amaranth) 35 days. 12-h means of 4, 59, on the daytime stomatal conductance, shoot growth,
and 114 ppb and root growth. This agricultural weed species may
be more tolerant to elevated O3 and moisture stress
than crop species.
Vanloocke et al. FACE; SoyFACE, Glycine max 12-h means of 40, 46, 54, 58, With increasing ozone treatment, yield (-64%),
(2012) Champaign, IL (soybean) 71, 88, 94, 116 ppb 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 *
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 Os. 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.
September 2019
8-169
DRAFT: Do Not Cite or Quote
-------
Table 8-21 (Continued): Ozone exposure and water cycling.
Study
Study Type and
Location
Study Species
Ozone Exposure
Effects on Water Cycling
Sun etal. (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); maximum 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 average
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).
September 2019
8-170
DRAFT: Do Not Cite or Quote
-------
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.
September 2019
8-171
DRAFT: Do Not Cite or Quote
-------
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 C>3: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
September 2019
8-172
DRAFT: Do Not Cite or Quote
-------
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.
September 2019
8-173
DRAFT: Do Not Cite or Quote
-------
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 * Os
(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
drought: O3 treatments:
charcoal-filtered air,
charcoal-filtered air + 120 ppb of
O3 for 13 h/day. Drought
treatment: maintained soil water
content at 35%
Differences in Gs were observed on Day 10, when
O3 x drought treatment had a lower Gs than control
(roughly 0.18 mmol/m2-s vs. 0.3 mmol/m2-s), with no
significant effect of drought alone or O3 alone.
Wagg et al. (2013) Greenhouse; England
Ranunculus acris
(meadow buttercup),
Dactylis glomerata
(orchard grass)
Low O3: 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.
September 2019
8-174
DRAFT: Do Not Cite or Quote
-------
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); ambient+24 (50.7, 62.4,
24.80); ambient+36 (62.0, 80.6,
44.33); ambient+48 (72.6, 89.1,
58.04); ambient+60 (88.9,
108.4, 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
maximum 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.
September 2019
8-175
DRAFT: Do Not Cite or Quote
-------
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,
perennial forb
endemic to the
Mediterranean
24-h avg (ppb): CF = 11,
NF+30 = 40, ambient = 32;
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
Visible symptoms began in 3 days at an AOT40 of
12-h 2 ppm-h in the NF+30 treatment. Mature leaves in
NF+30 had a 26% reduction in photosynthesis and a
8-h 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/meters 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/meters squared/second; VPD = vapor pressure deficit; W126 = cumulative integrated exposure index with a sigmoidal weighting function; WUE = water use
efficiency.
September 2019
8-176
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
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 (Menendez 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 (Takemoto 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 (Primus serotina) and tulip poplar
(Liriodendron tulipifera) increased with greater N deposition, while the only three species showing
September 2019
8-177
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
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 ofN 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 et al.. 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.
September 2019
8-178
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
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.
September 2019
8-179
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
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 (Betula pendula) 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
September 2019
8-180
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
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 et 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 (Ghimirc
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.. 2011). 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
September 2019
8-181
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
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.
September 2019
8-182
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
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 ofW126, 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 (Turnipseed et al.. 2009: Uddling et al.. 2009: Bergweiler et al.. 2008: Hogg et
al.. 2007: Grulke et al.. 2004: Grantz et al.. 1997: Grantzetal.. 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
September 2019
8-183
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
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. 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; Hecketal.. 1982). 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. The NHEERL-WED
project was initiated by U.S. EPA in 1988 with the same objectives for tree species, and yielded
49 exposure-responses curves for multiple genotypes of 10 tree species grown for up to 3 years in
Oregon, Michigan, and the Great Smoky Mountains National Park. Both projects used OTCs to expose
plants to three to five levels of ozone. Eight of the 54 crop data sets were from plants grown under a
September 2019
8-184
DRAFT: Do Not Cite or Quote
-------
1
2
3
combination of ozone exposure and experimental drought conditions. These two programs are explained
in detail in Section 9.5 of the 2013 Ozone ISA. Figure 8-14 shows an example of some of the
exposure-response information from the NHEERL-WED on tree seedlings.
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: permission pending Lee and Hoasett (1996).
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.
4 In the 2013 Ozone ISA, yield and growth results for aspen trees and soybean that had provided
5 extensive exposure-response information in those projects have become available from studies that used
6 FACE technology, which is intended to provide conditions much closer to natural environments
7 (Pregitzer et al.. 2008; Morgan et al.. 2006; Morgan et al.. 2004; Dickson et al.. 2000). The NCLAN and
September 2019 8-185 DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
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 8-15. Figure 8-16). 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 ct 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
-)—'
c
0
o
-------
1
2
3
4
5
6
7
8
9
10
11
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 3 FACE rings filled with an assemblage of 5 clonal genotypes of aspen at Aspen FACE;
bars are SE for 3 rings; dashed line is median composite model for 4 clonal genotypes and wild-type seedlings in 11 NHEERL-WED
1-year OTC studies. Aspen FACE ozone data updated from Kubiske and Foss (2015). 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: Permission pending King et al. (2005). and Lee and Hoasett (1996).
Figure 8-16 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
September 2019
8-187
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
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 8-22). 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 et al.. 2011).
September 2019
8-188
DRAFT: Do Not Cite or Quote
-------
Table 8-22 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
Beramann et al. (1995).
Berqmann 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)
September 2019
8-189
DRAFT: Do Not Cite or Quote
-------
Table 8 22 (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).
Berqmann 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
Berqmann et al. (1995),
Berqmann 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
Berqmann et al. (1995).
Pleiiel and Danielsson
(1997)
Aegilops geniculata
Annual
-0.013
0.96
7.6
0.51
Introduced
Gimeno et al. (2004)
September 2019
8-190
DRAFT: Do Not Cite or Quote
-------
Table 8 22 (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)
September 2019
8-191
DRAFT: Do Not Cite or Quote
-------
Table 8 22 (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)
September 2019
8-192
DRAFT: Do Not Cite or Quote
-------
Table 8 22 (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
Bunaener 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
Bunaener 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)
September 2019
8-193
DRAFT: Do Not Cite or Quote
-------
Table 8 22 (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)
September 2019
8-194
DRAFT: Do Not Cite or Quote
-------
Table 8 22 (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
Bunaener et al. (1999b)
September 2019
8-195
DRAFT: Do Not Cite or Quote
-------
Table 8 22 (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
Bunqener et al. (1999b),
Ashmore et al. (1996)
Trisetum flavescens
Perennial
0.001
1.1
-
0.61
Introduced
Bunqener et al. (1999b)
Bromus erectus
Perennial
0.001
1.05
-
0.99
Introduced
Bunqener 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 (2015) 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 (Haves et al.. 2007). Six out of the sixteen studies above have been cited in previous ISAs or AQCDs."
September 2019
8-196
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
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 8-23). 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 8-22). 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,
and for the biomass accumulation of tree species. 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. 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.
September 2019
8-197
DRAFT: Do Not Cite or Quote
-------
Table 8-23 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.
Grantz et 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 at2-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 (a = -2.57,
r= 0.52).
September 2019
8-198
DRAFT: Do Not Cite or Quote
-------
Table 8-23 (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 average 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.
September 2019
8-199
DRAFT: Do Not Cite or Quote
-------
Table 8-23 (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).
September 2019
3-200
DRAFT: Do Not Cite or Quote
-------
Table 8-23 (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 Oswas
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)
(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-hour 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
nonsignificant.
September 2019
8-201
DRAFT: Do Not Cite or Quote
-------
Table 8-23 (Continued): Exposure indices and exposure response.
Study
Study Type and
Study Location
Study Species
Ozone Exposure
Relevant Results
Haves et al. (20111
Greenhouse; near Two communities:
Marchlyn Mawr, U.K. 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.
September 2019
8-202
DRAFT: Do Not Cite or Quote
-------
Table 8-23 (Continued): Exposure indices and exposure response.
Study
Study Type and
Study Location
Study Species
Ozone Exposure
Relevant Results
Sanz etal. (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.
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 etal. (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
(Gaudin) Markgr.-
Dann, Festuca
violacea subsp.
puccinellii (Pari.)
Foggi, GrazRossi &
Signorini, Plantago
alpina L., Silene
acaulis (L.) Jacq.,
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.
September 2019
8-203
DRAFT: Do Not Cite or Quote
-------
Table 8-23 (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
(average 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;
EO = elevated-ozone treatment; E-03 = elevated-ozone treatment; kg/ha = kilograms/hectare; NF = nonfiltered air; nL/L = nanoliters/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 = micromoles/mole; W126 = cumulative integrated
exposure index with a sigmoidal weighting function.
September 2019
8-204
DRAFT: Do Not Cite or Quote
-------
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/10.1080/11263504.2016.1174169.
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. Ill 1/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/sll356-017-9369-7.
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: Serbia 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-y.
Allea EB: Temple. PJ: Bvtnerowicz. A: Arbaugh. MJ: Sirulnik. AG: Rao. LE. (2007). Patterns of understory
diversity in mixed coniferous forests of southern California impacted by air pollution. Scientific WorldJournal
7: 247-263. http://dx.doi.org/10.1100/tsw.20Q7.72.
Amundsoa RG: Alscher. RG: Fellows. S: Rubia G: Fincher. J: VanLeukea P: Weinsteia 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. O: 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.Ill 1/gcb. 13823.
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.
Andrew. CJ: vanDiepea 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/SQ 160-4120(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.
September 2019
8-205
DRAFT: Do Not Cite or Quote
-------
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: Mortoa AJ. (1995). Effects of ozone on
calcareous grassland communities. Water Air Soil Pollut 85: 1527-1532.
http://dx.doi.org/10.1007/BF0Q477198.
Avnerv. 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.ll.045.
Awmack. CS: Harringtoa 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.ll.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/sl0646-013-1146-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.
Bender. J: Bergmann. E: Weigel. 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: Jepsea EA: Roth. JA. (2006). Field responses of Prunus serotina and Asclepias syriaca to 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.
Bergmana 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.
September 2019
8-206
DRAFT: Do Not Cite or Quote
-------
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.
Bergmana 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: Nelsoa 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. 112.205591.
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.Ill 1/i. 1365-3040.2010.02165.x.
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/10.1111/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 inBrassica campestris. Environ Exp Bot 75: 225-234.
http://dx.doi.Org/10.1016/i.envexpbot.2011.07.006.
Blande. JD: Holopainea JK: Li. T. (2010). Air pollution impedes plant-to-plant communication by volatiles.
Ecol Lett 13: 1172-1181. http://dx.doi.Org/10.llll/i.1461-0248.2010.01510.x.
Blande. JD: Holopainea 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.Ill 1/pce. 12352.
Blum. O: Bvtnerowicz. 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/S0269-749ir97')00158-9.
Bohler. S: Sergeant. K: Jolivet. Y: Hoffmana L: Hausmaa 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/10.1016/0269-7491(92)90155-4.
September 2019
8-207
DRAFT: Do Not Cite or Quote
-------
Bolsinger. M: Lier. ME: Lanskv. 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/10.1016/0269-7491(91190156-0.
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: Xia 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.
Browa 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.
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.
Burtoa A: Burkev. KO: Carter. TE. Jr: Orf. J: Cregan. 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-l.
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/sl0661-006-6558-Q.
Bussotti. F: Gerosa. G. (2002). Are the Mediterranean forests in Southern Europe threatened from ozone? J
MediterrEcol 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: Deaa 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/10.1111/i. 1469-8137.1992.tb01058.x.
Bvtnerowicz. A: Olszvk. DM: Fox. CA: Dawsoa PJ: Kats. G: Morrisoa 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/10.1080/08940630.1988.10466463.
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 Saf74: 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: Ibafiez-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/10.1016/0378-1127(93)90190-X.
Cannoa 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.
September 2019
8-208
DRAFT: Do Not Cite or Quote
-------
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 GeophysRes Atmos 121: 14679-14690. http://dx.doi.org/10.1002/2016JDQ25141.
Ceron-Bretoa JG: Ceron-Breton. RM: Guerra-Santos. JJ: Aguilar-Ucaa 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.
Chapia 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.
NewPhytol 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.llll/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/S0269-7491(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.
Chea Z: Wang. X: Shang. H. (2015). Structure and function of rhizosphere and non-rhizosphere soil microbial
community respond differently to elevated ozone infield-planted wheat. J Environ Sci 32: 126-134.
http://dx.doi.Org/10.1016/i.ies.2014.12.018.
Chea 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/s00442-005-0249-3.
Coleman. JS: Jones. CG. (1988). Plant stress and insect performance: cottonwood, ozone and a leaf beetle.
Oecologia 76: 57-61.
September 2019
8-209
PRAFT: Po Not Cite or Quote
-------
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/16.1-2.145.
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: Helgasoa 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/10.1111/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. Ill 1/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/sll829-013-9286-8.
Couture. JJ: Meehaa 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: Meehaa 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/s00442-011-
2139-1.
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: Chea F: Zhang. Y: Ge. F. (2016a). Elevated 03 and TYLCV infection reduce the suitability of
tomato as a host for the whitefly Bemisia tabaci. International Journal of Molecular Sciences 17: 1964.
http://dx.doi.org/10.3390/iimsl7121964.
Cui. H: Sun. Y: Su. J: Rea O. in: 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/sll829-012-9189-Q.
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/s00442005Q381.
Panielssoa 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.
Parbah. JNT: Kubiske. ME: Neilsoa N: Oksanea E: Vaapavuori. E: Karnoskv. PF. (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.2007.42.
Parbah. JNT: Kubiske. ME: Nelson. N: Oksanea E: Vapaavuori. E: Karnoskv. PF. (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.
Parrall. NM. (1989). The effect of air pollutants on physiological processes in plants. Plant Cell Environ 12: 1-
30. http://dx.doi.org/10.1111/i. 1365-3040.1989.tb01913.x.
September 2019
8-210
PRAFT: Po Not Cite or Quote
-------
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:OSOVWT12.0.CQ: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/10.1656/1092-6194(2007)14r415:OITPWT12.0.CQ:2.
Davis. DP. (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/10.1656/045.018.0111.
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.
Pavisoa 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: Simpsoa 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.
Peaa 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.
Pecock. C: Chung. H: Venterea. R: Gray. SB: Leakey. APB: 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.
Pecock. 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.
Piaz-Pe-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.
Picksoa RE: Lewin. KF: Isebrands. JG: Colemaa MP: Heilmaa WE: Riemenschneider. PE: Sober. J: Host.
GE: Zak. PR: Hendrev. GR: Pregitzer. KS: Karnoskv. PF. (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. Pept. of Agriculture, Forest Service.
http://nrs.fs.fed.us/pubs/278.
Pietze. 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/10.1111/i. 1365-2486.2011.02477.x.
Pohmea GP. (1988). Indirect effects of air pollutants: changes in plant/parasite interactions. Environ Pollut 53:
197-207.
September 2019
8-211
PRAFT: Po Not Cite or Quote
-------
Pohrmana 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.71.12.7750-7758.20Q5.
Dumont. J: Spicher. F: Montpied. P: Pizengremel. 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: Stevea 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 Biochem76: 227-234. http://dx.doi.Org/10.1016/i.soilbio.2014.05.027.
Ebanvenle. E: Burtoa 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. IntBiodeteriorBiodegradation 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.
Embersoa L: Ashmore. MR: Cambridge. HM: Simpson. P: Tuovinea JP. (2000a). Modelling stomatal ozone
flux across Europe. Environ Pollut 109: 403-413. http://dx.doi.org/10.1016/S0269-7491(00)00043-9.
Embersoa LP: Pleiiel. H: Ainsworth. EA: van Pen Berg. M: Rea W: Osborne. S: Mills. G: Pandev. P:
Pentener. F: Buker. P: Ewert. F: Koebl. R: Van Pingenend. 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.
Embersoa 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/10.1111/gcb. 12222.
Farre-Armengol. G: Pefiuelas. 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. Ill 1/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 ChemPhys Meteorol 56: 230-248. http://dx.doi.org/10.Ill 1/i. 1600-0889.2004.00097.x.
Felzer. B: Reillv. J: Melillo. J: Kicklighter. P: Sarofim. M: Wang. C: Prina 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: G01020.
http://dx.doi.org/10.1029/2008JG00Q826.
September 2019
8-212
PRAFT: Po Not Cite or Quote
-------
Feng. Y: Lia 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: Lia X: Yu. Y: Zhu. J. (2011). Elevated ground-level 03 changes the diversity of anoxygenie purple
phototrophic bacteria in paddy field. Microb Ecol 62: 789-799. http://dx.doi.org/10.1007/s00248-011-9895-7.
Feng. Y: Yu. Y: Tang. H: Zu. O: Zhu. J: Lia 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.ll.032.
Feng. Z: Btiker. P: Pleiiel. H: Embersoa 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. Ill 1/gcb. 13824.
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.ll.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.
Fena 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.
Finnvedea 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.
Fishmaa J: Belina. KM: Encarnacion. CH. (2014). The St. Louis ozone gardens: Visualizing the impact of a
changing atmosphere. Bull AmMeteorol Soc 95: 1171-1177. http://dx.doi.org/10.1175/BAMS-D-13-
00009.1.
Fleischer. K: Rebel. KT: van der Molen. MK: Erismaa JW: Wassea MJ: van Loon. EE: Montagnani. L: Gough.
CM: Herbst. M: Janssens. IA: Gianelle. D: Dolmaa 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.
EnvironExpBot 61: 190-198. http://dx.doi.Org/10.1016/i.envexpbot.2007.05.009.
Foot. JP: Capora SJM: Lee. JA: Ashendea 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/10.1111/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: Kooiiman. 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/S0098-8472(00)00052-6.
September 2019
8-213
DRAFT: Do Not Cite or Quote
-------
Fuentes. JD: Chamecki. M: Roulstoa 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.
Gaucher. C: Dizengremel. 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: Kasurinea A: Haikio. E: Holopainea T: Holopainea 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. Ill 1/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: Guna 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.
September 2019
8-214
DRAFT: Do Not Cite or Quote
-------
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.Ill 1/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-07-130.1.
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: Pedersoa 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/S0168-1923(96)02396-9.
Grantz. DA: Zhang. XJ: Massman. WJ: DenHartog. 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: Sullivaa 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/10.1890/110049.
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-20Q6-924489.
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/natureO 1728.
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/10.189Q/1051-
0761(2006)016r2368:PADE0012.0.CQ: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: Nguvea 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: Andersea 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/SQ269-7491(98)00130-4.
Grulke. NE: Johnson. R: Esperanza. A: Jones. D: Nguvea 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.
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.
September 2019
8-215
DRAFT: Do Not Cite or Quote
-------
Gunthardt-Goerg. MS: Maurer. S: Bolliger. 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.
Gustafsoa 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.0072717.
Haesler. F: Hagq 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: Gundersoa 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: Wilkinsoa 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.Ill 1/i. 1365-2486.2011.02613.x.
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: VanNostrand. 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.2013.177.
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.
September 2019
8-216
DRAFT: Do Not Cite or Quote
-------
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: 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 AirPollut 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: Fredericksea 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/10.1139/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/10.1111/i. 1752-4598.2008.00031.x.
Hofmockel. KS: Zak. DR: Moraa 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.
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: Holmaa 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.
Holmes. CD. (2014). Air pollution and forest water use [Comment]. Nature 507: E1-E2.
http://dx.doi.org/10.1038/naturel3113.
Holmes. WE: Zak. DR: 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.
September 2019
8-217
DRAFT: Do Not Cite or Quote
-------
Hong. B: Weinsteia D: Swanev. D. (2006). Assessment of ozone effects on nitrate export from Hubbard Brook
Watershed 6. EnvironPollut 141: 8-21. http://dx.doi.Org/10.1016/i.envpol.2005.08.030.
Hong. Y: Yi. T: Taa 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.Q1593.
Horn. KJ: Thomas. RO: Clark. CM: Pardo. LH: Fena ME: Lawrence. GB: Perakis. SS: Smithwick. EAH:
Baldwin. D: Braua S: Nordia 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/sl 1270-016-3127-1.
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.
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/10.1111/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-Q12-1158-9.
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.
Jacksoa 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.
Jacksoa DM: Ruftv. TW: Heagle. AS: Seversoa 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.
Jamiesoa 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.
September 2019
8-218
DRAFT: Do Not Cite or Quote
-------
Jones. ME: Paine. TP. (2006). Detecting changes in insect herbivore communities along a pollution gradient.
EnvironPollut 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: Freemaa 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. FunctEcol 31: 56-64. http://dx.doi.org/10. Ill 1/1365-2435.12774.
Kanerva. T: Paloiarvi. A: Ramo. K: Manninea 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.
Kanerva. T: Paloiarvi. A: Ramo. K: Oianpera. K: Esala. M: Manninea 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: Regina. K: Ramo. K: Oianpera. K: Manninea 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: Gagnoa ZE: Dicksoa 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: Dicksoa 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.
Kasurinea A: Biasi. C: Holopainea T: Rousi. M: Maenpaa. M: Oksanea E. (2012). Interactive effects of
elevated ozone and temperature on carbon allocation of silver birch (Betula pendula) genotypes in an open-
air field exposure. Tree Physiol 32: 737-751. http://dx.doi.org/10.1093/treephvs/tps005.
Kasurinea A: Keinanea MM: Kaipainen. S: Nilsson. LO: Vapaavuori. E: Kontro. MH: Holopainea 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.llll/i.1365-2486.2005.00970.x.
Kasurinea A: Riikonea J: Oksanea E: Vapaavuori. E: Holopainea 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/slll04-005-6Q26-6.
Kasurinea A: SillVer. 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: Thompsoa CR. (1985). Open top experimental chambers for trees. J Air Waste Manag
Assoc 35: 1298-1301. http://dx.doi.org/10.1080/00022470.1985.10466033.
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/10.1080/00022470.1976.10470367.
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.
September 2019
8-219
DRAFT: Do Not Cite or Quote
-------
Kefauver. SC: Pefiuelas. 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/BF00376884.
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-
1127(95)03603-2.
Khaling. E: Li. T. ao: Holopainen. JK: Blande. JD. (2016). Elevated Ozone Modulates Herbivore-Induced
Volatile Emissions of Brassica nigra and Alters a Tritrophic Interaction. J Chem Ecol 42: 368-381.
http://dx.doi.org/10.1007/sl0886-016-Q697-8.
Khaling. E: Papaziaa S: Poelmaa 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: Karnoskv. 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: Karnoskv. 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: Lundavist. 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-66-1149.
Krupa. SV: Grunhage. L: Jager. HJ: Nosal. M: Manning. WJ: Legge. 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/SQ269-7491(99)80014-1.
September 2019
8-220
PRAFT: Po Not Cite or Quote
-------
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-2015-0Q01
Kubiske. ME: Ouinn. VS: Marguardt. 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.
Landesmana 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/iournal.pone.007582Q.
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: Amundsoa 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: Amundsoa RG: Weinsteia 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: Tingev. DT: Waschmana 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/tpp071.
Lefohn. AS: Laurence. JA: Kohut. RJ. (1988). A comparison of indices that describe the relationship between
exposure to ozone and reduction in the yield of agricultural crops. Atmos Environ 22: 1229-1240.
http://dx.doi.org/10.1016/0004-6981(88)90353-8.
Lefohn. AS: Runeckles. VC. (1987). Establishing standards to protect vegetation - ozone exposure/dose
considerations. Atmos Environ 21: 561-568. http://dx.doi.org/10.1016/0004-6981(87)90038-2.
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
ApplBotFoodQual69: 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.llll/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.
Lewis. JS: Ditchkoff. SS: Lia 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.l.
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.
September 2019
8-221
DRAFT: Do Not Cite or Quote
-------
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. O: 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.llll/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. O: 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.2012.12Q.
Lindroth. RL: Reich. PB: Tioelker. MG: Volia JC: Oleksvn. 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.llll/i.1469-8137.1993.tb03854.x.
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.
Llovd. 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/10.1111/i. 1469-8137.1988.tb04192.x.
Lucas. PW: Rantanea L: Mehlhora 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: Kleemaa 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: Weinsteia 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.
September 2019
8-222
PRAFT: Po Not Cite or Quote
-------
Manninen. AM: Holopainen. T: Lwtikainen-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.
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: Landesmana 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. AJ: 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-20Q5-
873025.
Matvssek. R: Sandermana 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. AJ: Werner.
H: Wipfler. P: Obwald. W: Nikolova. P: Hanke. DE: Kraigher. H: Tausz. M: Bahnweg. G: Kitao. M: Dieler.
J: Sandermann. H: Herbinger. K: Grebenc. T: Blumenrother. M: Deckmvn. G: Grams. TEE: Heerdt. C:
Leuchner. M: Fabiaa 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. (1993). Effects of ozone on Solidago albopilosa, and endangered species of goldenrod
from central Kentucky. Bull Ecol Soc Am 75: 146.
McBride. JR: Laven. RD. (1999). Impact of oxidant air pollutants on forest succession in the mixed conifer
forests of 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. 338-352).
New York, NY: Springer, http://dx.doi.org/10.1007/978-l-4612-1436-6 16.
Mcfrederick. OS: Kathilankal. JC: Fuentes. JD. (2008). Air pollution modifies floral scent trails. Atmos Environ
42: 2336-2348. http://dx.doi.Org/10.1016/i.atmosenv.2007.12.033.
Mcgrath. JM: 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: Sua 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/10.1111/i. 1469-8137.2007.02018.X.
Mclaughlia SB: Wullschleger. SD: Sua 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.llll/i.1469-8137.2007.01970.x.
Meehan. TP: Couture. JJ: 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. Ill 1/nph. 12947.
September 2019
8-223
DRAFT: Do Not Cite or Quote
-------
Menendez. Al: 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.
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.
http://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.
http://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 :: USD A.
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: Suttoa 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: Braua S: Btiker. P: Bermeio. V: Calvo. E: Danielssoa H: Embersoa L: Fernandez. IG:
Griinhage. L: Harmens. H: Haves. F: Karlssoa P. erE: Simpsoa 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: Simpsoa 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
Chem65: 9579-9586. http://dx.doi.org/10.102l/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.l 111/i. 1365-2486.2005.1054.x.
Moraa 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.
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.
September 2019
8-224
DRAFT: Do Not Cite or Quote
-------
Morgaa 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: Nelsoa 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. Ill 1/i. 1469-8137.2006.01679.x.
Morskv. SK: Haapala. JK: Rinnaa R: Tiiva. P: Saarnio. S: Silvola. J: Holopainea 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.Ill 1/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/10.1080/0906471920941Q218.
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: Nilsea 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/10.1111/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.
Musselmaa RC: Massmaa 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: Johnsoa 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/S0269-7491(99)00247-X.
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.Ill 1/i. 1469-8137.1995.tb01839.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.
September 2019
8-225
DRAFT: Do Not Cite or Quote
-------
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'l365-3040.2003.01020.x.
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/S0269-7491(03)00089-7.
Ormrod. DP: Marie. BA: Allea 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: Btiker. 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. Ill 1/gcb. 13318.
Oshima. RJ: Braegelmann. PK: Baldwia 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/10.1080/00022470.1976.1047033Q.
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/treephys/24.3.277.
Panek. JA: Goldsteia 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/treephys/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/sl0021-0Q8-9148-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/10.1614/WS-D-15-00146.1.
Payne. RJ: Stevens. CJ: Pise. NB: Gowing. DJ: 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.
September 2019
8-226
DRAFT: Do Not Cite or Quote
-------
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-Johansea C: Tiea M: Skillmaa 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: Heinonea 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:
Dicksoa RE: Zak. PR: Oksanen. E: Sober. J: Harrington. R: Karnoskv. DF. (2002). Altered performance of
forest pests under atmospheres enriched with C02 and 03. Nature 420: 403-407.
http://dx.doi.org/10.1038/nature01028.
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: Nykanea 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
ChemEcol 33: 683-694. http://dx.doi.org/10.1007/slQ886-007-9255-8.
Pinto. PM: Blande. JP: Souza. SR: Nerg. AM: Holopainea JK. (2010). Plant volatile organic compounds
(VOCs) in ozone (03) polluted atmospheres: The ecological effects [Review]. J ChemEcol 36: 22-34.
http://dx.doi.org/10.1007/slQ886-009-9732-3.
Pinto. PM: 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. EnvironPollut 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.lll.
Pleiiel. H: Panielsson. 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: Panielsson. H: Simpson. P: 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. P. (2006). Soil respiration in northern forests exposed to elevated
atmospheric carbon dioxide and ozone. Oecologia 148: 503-516. http://dx.doi.org/10.1007/s00442-006-0381 -
8.
September 2019
8-227
PRAFT: Po Not Cite or Quote
-------
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 0-3. 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/10.1111/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-Q8-
098349-3.00010-4.
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/slll04-009-9972-6.
Puiol Pereira. EI: Chung. H: Scow. K: Sadowskv. MJ: vanKessel. 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.
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: Holopainea 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.
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/10.1139/x26-203.
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 :: USD A.
https ://www.fs.usda. gov/treesearch/pubs/13 377.
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.llll/i.1469-8137.1992.tb01055.x.
Reinert. RA: Easoa 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.llll/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: Middletoa 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.
September 2019
8-228
DRAFT: Do Not Cite or Quote
-------
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.
Rodenkirchea H: Gottleia 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.
Rogers. A: Allen. DJ: Davev. PA: Morgan. PB: Ainsworth. EA: Bernacchi. CJ: Cornic. G: Dermodv. OC:
Dohlemaa FG: Heatoa 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/10.1139/x26-073.
Samuelsoa 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: Oksanea E: Karialainea 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: Abivea S: Dittmar. T: Guggenberger. G: Janssens. IA: Kleber. M: Koegel-Knabner.
I: Lehmana 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: Easoa 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. (1999). 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.
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/s 10661 -011 -2243-z.
September 2019
8-229
PRAFT: Po Not Cite or Quote
-------
Smith. G: Coulstoa 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: Coulstoa 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. ForEcol Manage 104: 129-137. http://dx.doi.org/10.1016/SQ378-l 127(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.
EnvironPollut 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
withcarbon-11 labeling. Environ Pollut 64: 93-106. http://dx.doi.org/10.1016/0269-7491(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/sQ0442-009-1462-2.
Stoelkea G: Pritsch. K: Simon. J: Mueller. CW: Grams. TEE: Esperschuetz. J: Gavler. S: Buegger. F:
Brueggemann. N: Meier. R: Zeller. B: Winkler. JB: Rennenberg. 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/10.1080/11263500903429346.
Subramaniaa N: Karlssoa 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.
Sua 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.
Sua 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.
Szantoi. Z: Chappelka. AH: Muntifering. 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.ll.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: Dawsoa PJ: Morrison. CL: Temple. PJ. (1997). Effects of ozone onPinus
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.
September 2019
8-230
DRAFT: Do Not Cite or Quote
-------
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/SQ378-l 127(00100368-6.
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/sl0021-011-9493-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.Ill 1/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. (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/10.10Q7/978-l-
4612-1436-6 10.
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.
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/10.1016/0Q98-
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/10.1139/x93-01Q.
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.Ill 1/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 CP: 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. PM: Kats. G: Bvtnerowicz. A: Pawsoa PJ: Wolf. JW. (1984). Effects of ozone or
sulfur-dioxide on annual plants of the Mojave Pesert. J Air Pollut Control Assoc 34: 1017-1022.
Tiaa H: Chea G: Zhang. C: Liu. M: Sun. G: Chappelka. A: Ren. W: Xu. X: Lu. C: Paa S: Chea H: Hui. P:
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/slQ021-012-9539-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 LI 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.
September 2019
8-231
PRAFT: Po Not Cite or Quote
-------
Toet. S: Ineson. P: Peacock. S: Ashmore. M. (2011). Elevated ozone reduces methane emissions frompeatland
mesocosms. Global Change Biol 17: 288-296. http://dx.doi.Org/10.llll/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 a three-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.gov/exe/ZvPURL.cgi?Dockev=200089CW.txt.
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/searchResults.xhtml?searchOuerv=PB80176779.
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). (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://cfpub.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 (accessed
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/naaas/standards/ozone/data/2014102 lwelfarerea.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.
September 2019
8-232
DRAFT: Do Not Cite or Quote
-------
U.S. EPA (U.S. Environmental Protection Agency). (2018). Integrated science assessment for oxides of nitrogen,
oxides of sulfur and particulate matter -Ecological criteria (2nd external review draft). (EPA/600/R-18/097).
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. PS. (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: Ashrafuzzamaa 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.
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/ (accessed
Valkama. E: Koricheva. J: Oksanea 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. https://doi.org/10.ll 1 l/i.l365-2486.2006.01284.x.
VanDingenea R: Dentener. FJ: Raes. F: Krol. MC: Embersoa 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: Oi. 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.
Vanderhevdea 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/S0269-7491(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.llll/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.llll/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. Ill 1/i. 1365-
2486.2005.01083.x.
September 2019
8-233
DRAFT: Do Not Cite or Quote
-------
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(02101012-9.
Vuorinen. T: Nerg. AM: Holopainea 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.
Wagg. S: Mills. G: Haves. F: Wilkinsoa 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.ll.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.
Waa W: Manning. WJ: Wang. X: Zhang. H: Sua 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.
Waa 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], ActaEcol 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/sl0651-012-Q195-2.
Wang. P: Marsh. EL: Ainsworth. EA: Leakey. ADB: Sheflin. AM: Schachtmaa 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-Q17-14936-2.
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/10.1139/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.Ill 1/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 Int 9: 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.
September 2019
8-234
DRAFT: Do Not Cite or Quote
-------
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.
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 Waste Manag Assoc 27: 230-234.
Winner. WE: Lefoha 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/10.1016/0269-7491(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: 1150-1162. http://dx.doi.org/10. Ill 1/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.Ill 1/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: Chapia 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/10.15Q4/IJEP.2017.10003698.
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.
Yuaa JS: Himanea SJ: Holopainen. JK: Chea 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.
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 ChemPhys 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.
September 2019
8-235
DRAFT: Do Not Cite or Quote
-------
Yuska. DE: Skellv. JM: Ferdinand. JA: Stevensoa 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(03100096-4.
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. Ill 1/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.llll/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. Z. (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.org/10.1111/i. 1365-2486.2010.02258.X.
September 2019
8-236
PRAFT: Po Not Cite or Quote
-------
APPENDIX 9 THE ROLE OF TROPOSPHERIC
OZONE IN CLIMATE EFFECTS
Summary of Causality Determinations
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 causal
statement provides a more accurate reflection of the available evidence). The new evidence
comes from the Intergovernmental Panel on Climate Change (IPCC) Fifth Assessment Report
(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
indicate a change to either causality determination included in the 2013 Ozone ISA.
Effects Relationship
Tropospheric ozone and climate change
Radiative forcing Causal
Temperature, precipitation, and related T ¦, i A i ,
,. A . ' , Likely-to-be-causal
climate variables
9.1 Introduction
9.1.1 Summary from the 2013 Ozone ISA
1 Changes in the abundance of tropospheric ozone perturb the radiative balance of the atmosphere
2 by interacting with incoming solar radiation and outgoing longwave radiation. This effect is quantified by
3 the radiative forcing (RF) metric. Through this effect on the Earth's radiation balance, tropospheric ozone
4 plays a major role in the climate system, and increases in ozone abundance contribute to climate change
5 (Forster et al.. 2007).
6 • Increases in tropospheric ozone are tied to the rise in emissions of ozone precursors from human
7 activity, mainly from fossil fuel consumption and agricultural processes. Models estimate that the
8 global average tropospheric ozone concentration has increased 30-70% since preindustrial times
9 (Gauss et al.. 2006). and observations indicate that tropospheric ozone concentrations may have
10 increased by factors of 4 or 5 in some regions (Marenco et al.. 1994; Staehelin et al.. 1994). In the
11 21st century, as the Earth's population continues to grow and energy technology spreads to
September 2019
9-1
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
developing countries, a further rise in the global concentration of tropospheric ozone is likely
(Forster et al.. 2007).
• The 2007 IPCC report estimated RF from tropospheric ozone since preindustrial times (1750 to
2005) to be 0.35 W/m2 (average) 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 (Shindell et al.. 2006; Liao et al.. 2004; Micklev et
al.. 1999).
• 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). 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 mid to upper troposphere (Forster et al.. 2007).
• 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.. 2011). 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 draft 2019
Ozone ISA. The current 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 indicate 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.
September 2019
9-2
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
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
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 (Wuebbles et al.. 2017; IPCC. 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
September 2019
9-3
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
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
(Wuebbles et al.. 2017; IPCC. 2013).
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 (Wuebbles et al.. 2017; IPCC.
2013). reports on the 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 et al.. 2013). as well as more recent literature published subsequent to AR5.
9.1.3.2 Ozone Chemistry and Role in Climate
The 2013 Ozone ISA described how tropospheric ozone differs in important ways from other
greenhouse gases, 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
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 greenhouse gases and water vapor is relatively weak.
In addition, unlike other major greenhouse gases, ozone absorbs in the shortwave as well as in the
longwave part of the spectrum. This absorption leads to RF and the consequences described below.
Figure 9-1 depicts the influence of tropospheric ozone on climate. Emissions of ozone precursors
including CO, VOCs, CH4, and NOx lead to the production of tropospheric ozone. 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
September 2019
9-4
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
discussion of some of the most direct feedbacks, but the downstream effects and their longer term
feedbacks are extremely complex and outside the scope of this assessment.
, t
I Climate Effects *
I on Human Health r
^ and Ecosystems )
Changes in Tropospheric
O Abundance
l (18) .
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). 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 [i.e., 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 ozone RF as an endpoint
because ozone ERF estimates in the published literature are more limited, and differences between RF and
September 2019
9-5
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
ERF for ozone tend to be small compared to existing uncertainties in RF and ERF (Mvhre et al.. 2013).
The nonuniform distribution of ozone (spatially 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 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 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), ranking ozone third
in importance for global climate behind these two major greenhouse gases (Forstcr 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."
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. Based on this review of the literature, evidence is inadequate to determine if a
causal relationship exists between changes in tropospheric ozone concentrations and UV-B effects.
September 2019
9-6
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
9.2
Ozone Impacts on Radiative Forcing
Highlights ofRcccnt ICriilcnccfor Radiative / 'lining
CIkiiiucs iii llic ;ilmiid;iiicc ol" iroposphcric oaiiic ;ilfccl kl' I lie 2d I i Oaiiic IS \
reporls ;i kl' nf i) '5 W in from iroposphcric oahic since prciiidiisirul nines < I ~5u in 2<>u5i
h;ised on niullinuidcl sindics (I 'orslcr el al . 2<)()7i I lie mosi recenl ll'( ( ;isscssinciil. \k5.
reports iroposphcric oahic kl ' ;is () 4o d) 2o In (> (>ui \\ in i \l\ lire cl al . 2<) I3 i. u Inch is
w 11 hi ii r;iimc of pre\ ions ;isscssinciils (i c . \K4i. There li;i\ c ;ilso Kvn ;i lew siudics since
AR5. iiicliidinu I lie sind\ ill" li'iipiisplieric n/nne kl ' h;ised iin 1 lie Cnnpled Model
liiiei"eiinip;iiistiii kiiijeel Mi;ise<> (("MI P(> i d;il;i sel ;md I lie Xiniiispherie Clieniisii\ ;nid (lini;iie
Model Iiilei"eonip;ii'ison I'lojeel ( \('(A11lJ) ninlliniodel sind\ ol' liopospliene elieniis|i\. holli
ol' w Inch reinforce I lie \R5 esiini;iies ;md ilie c;ins;il rehilioiiship helween iropospheric o/one
;md kl'
9.2.1 Recent Evidence for Historical Period
• 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 (Shindcll 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 (Mvhrc 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^HNCh and O3 + NO^NCh + O2) (Newsome and Evans, 2017) 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 (Conlcv 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 (Chcca-Garcia et al.. 2018).Mvhrc 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
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.
September 2019
9-7
DRAFT: Do Not Cite or Quote
-------
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. (2013)
0.33
(0.31 to 0.35)
(Sovde etal.. 2012)c
0.45
0.38
Skeieet al. (2011)
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. (2013).
°0.45 based on REF chemistry, 0.38 based on R2 chemistry, see Sovde et al. (2012).
dForster et al. (2007).
Source: Mvhre et al. (2013).
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. (2013).
September 2019
9-8
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Radiative forcing of climate between 1750 and 2011
Forcing agent
a
c
Q
O)
o
Q.
O
c
<
Well Mixed
Greenhouse Gases
Ozone
Stratospheric water
vapour from CH4
Surface Albedo
Contrails
Aerosol-Radiation Interac.
Aerosol-Cloud Interac.
CO,
Other WMGHG
¦
¦
Halocarbons
-I
CH- ! ! ffn
Stratospheric 1-fftxKxf H Tropospheric
ftl
Lar>d Use
uj Black carbon
^ on snow
h
Contrail induced cirrus
1
03
*_
Z3
-t—•
03
Solar irradiance
¦*l
-1
0 1 2
Radiative Forcing (W m 2)
Note: Figure 9-2 is taken from IPCC AR5 Figure 8.15.
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.
Source: Mvhre et al. (2013).
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).
• 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-Tropospheric Emission Spectrometer (TES);
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 (1 standard
deviation).
• 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
September 2019
9-9
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
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. CH4 is itself a powerful
greenhouse gas and a precursor to ozone, leading to further warming (Fiore 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 (Mvhre et al.. 2013).
September 2019
9-10
DRAFT: Do Not Cite or Quote
-------
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
Note: Figure 9-3 is taken from IPCC AR5 Figure 8.17.
Source: Myhre et al. (2013).
Figure 9-3 Radiative forcing (RF) over the industrial era associated with
emitted compounds, including ozone (green bars) and its
precursors.
September 2019
9-11
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
9.2.2 Recent Evidence of Radiative Forcing Temporal and Spatial Trends
• 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).
. Tropospheric
Total
Stratospheric
0.2
1750
1800
1B50
1900
1950
2000
Note: Figure 9-4 is taken from IPCC AR5 Figure 8.7.
Note: Stratospheric ozone radiative forcing is also shown but is not discussed as part of this Appendix.
Source: Mvhre et al. (2013).
Figure 9-4 Time evolution of the radiative forcing (RF) from tropospheric
ozone from 1750 to 2010.
• IPCC 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 (Mvhre et al..
2013).
September 2019
9-12
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
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 van among models
(Young et al.. 2013). Shifts in 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 CH4 concentrations (Zhang et al..
2016).
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.
Preindustriai to Present-Day Forcing
Multi-model mean Standard deviation
0.29 W m3
0.16(0.19) Wm3
cc
o
c
o
N
O
I I I
-.88 -.62 -.38 -.12 .12 .38 .62 .88
.06 .12
Note: Figure 9-5 is taken from IPCC AR5 Figure 8.23, fourth row.
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.
Source: Mvhre et al. (2013).
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).
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). 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/nr. Shindell et al. (2013) found good
September 2019
9-13
DRAFT: Do Not Cite or Quote
-------
1
2
agreement between the modeled and observed global mean RPs. 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 '
-0.54 -0.42 -0.30 -0.1 B -0.06 0.06 0.1 B 0.30 0.42 0.54
Note: Figure 9-6 is taken from Shindell et al (2013), Figure 3.
Source: 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
3 Recent evidence continues to indicate a causal relationship between tropospheric ozone and
4 RF as concluded in the 2013 Ozone ISA (Table 9-4). The new evidence comes from the IPCC AR5
5 (Mvhre et al.. 2013) and its supporting references—in addition to a few more recent studies—and builds
6 on evidence presented in the 2013 Ozone ISA. None of the new studies indicate a change to the causality
7 determination included in the 2013 Ozone ISA.
8 • The AR5 best estimate of tropospheric ozone RF is 0.40 (0.20 to 0.60) W/m3 (from 1750 to 2011)
9 rTable 9-2; (Mvhre et al.. 2013)1. The overall confidence in the tropospheric ozone RF is high
10 (Table 9-3. Figure 9-2). Additionally, there have been a few studies since AR5, including the
11 study of tropospheric ozone RF based on the CMIP6 data set (Checa-Garcia et al.. 2018) and the
September 2019
9-14
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
ACCMIP multimodel study of tropospheric chemistry (Conlcv 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 (above atmospheric boundary layer)
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. Despite these uncertainties, the overall
confidence in current estimates of tropospheric ozone RF is high (Table 9-3).
• Further research in the following areas can help in address remaining uncertainties. These areas
include improving (1) the quantification of observed trends in ozone concentrations in the free
troposphere, upper troposphere, and remote regions; (2) the understanding of the CH4 budget and
of ozone coupling with temperature, water vapor, and clouds (with implications for the
height- and latitude-dependence of ozone RF); and (3) the estimates of ozone spatio-temporal
structure developed using global models constrained by observations.
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 Multidecadal, global chemistry-climate Mvhre et al. (2013): Section 9.2.1
multiple, high-quality studies modeling ensemble studies constrained
by historical observations of ozone
concentrations (e.g., IPCC AR5;
ACCMIP; CMIP6)
Robust physical Robust, well-understood relationship Mvhre et al. (2013); Section 9.2.1;
understanding between tropospheric ozone Section 9.1.3.3
concentration and RF
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
September 2019
9-15
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
9.3 Ozone Impacts on Temperature, Precipitation, and Related
Climate Variables
Highlights of Recent Evidence for Impacts on t emperature. Precipitation, ami
Related ( limate I ariahles
( iiiisisieiil Willi pre\ inns cMim;ilcv 1 lie cITccl ill" I I'lipiispllCNC ii/OMC Oil ulollMl MII'lMCC
ICMipci'Mllirc. llll'Oimll lis IllipMCl DM kl'. COIll I lilies III he CslllMMlCtl Ml roildlK II. | II i (' slIICC
pre 11 id i isi ri;i 11 lines (\ic el mI . 2. Allen el
mI.. 2
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
• Literature cited in IPCC AR5 (Mvhre et al.. 2013) continues to be consistent with these earlier
estimates of the effect of tropospheric ozone on global surface temperature. Xie et al. (2016). in a
more recent modeling study consisting of a series of 15-year sim ulations 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; Sluadell 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.
90°N
60°N
30°N
EQ
30°S
60°S
90°S
Note: Haure 9-7 is taken from Xie et al. (2016). Figure 4a.
Source: Xie et al. (2016).
Figure 9-7 Mean annual change in surface temperature (°C) resulting from
tropospheric ozone concentration changes from 1850-2013.
• Idealized modeling studies also support this basic magnitude of the impact of ozone RF on global
and regional temperatures (Huszar 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
September 2019
9-17
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
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, 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 (Mvhrc et al.. 2013). These are in
addition to the key sources of uncertainty in quantifying ozone RF changes, 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 (Mvhrc 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., fLew 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 (Mvhre 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.
September 2019
9-18
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
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 m global-average precipitation, with
regional shifts in precipitation patterns near the equator (Figure 9-8).
90°N —
60°S - --
90°S ' | ¦ ¦ | ¦ ' yvty f ^¥f:rr"
0° 60°E 120°E 180° 120°W 60° W 0°
i i i i
-0.5 -0.3 -0.1 0.1 0.3 0.5
Note: Figure 9-8 is taken from Xie et al. (2016). Figure 4c.
Source: Xie et al. (2016).
Figure 9-8 Mean annual change in precipitation (mm/day) resulting from
tropospheric ozone concentration changes from 1850-2013.
• 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.
September 2019
9-19
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
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 storm track shift (Lau and Kim. 2015V
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
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
indicate a change to the causality determination included in the 2013 Ozone ISA.
• 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 (Xie et al..
2016; Mvhre et al.. 2013). with larger effects regionally. In addition to temperature, ozone
changes have impacts on other climate metrics such as precipitation and atmospheric circulation
patterns (Macintosh et al.. 2016; Allen etal.. 2012; Shindell et al.. 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 requires complex climate simulations that include 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 mean and regional surface
temperatures.
Further research in the following areas can help address these remaining uncertainties, which
include quantifying a more precise relationship 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.
September 2019
9-20
DRAFT: Do Not Cite or Quote
-------
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
Robust physical
understanding
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
Temperature: Robust, well-understood
relationship between RF and
atmospheric and surface temperatures
Huszaret al. (2012); Yang etal. (2012);
Xie et al. (2016): Mvhre etal. (2013):
Fioreet 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 Faluveai (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
September 2019
9-21
DRAFT: Do Not Cite or Quote
-------
9.4 References
Allea 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/naturel 1097.
Bellouia N: Baker. L: Hodnebrog. O: Olivie. D: Cheriaa R: Macintosh. C: Samset. B: Esteve. A: Aamaas. B:
Ouaas. J: Mvhre. G. (2016). Regional and seasonal radiative forcing by perturbations to aerosol and ozone
precursor emissions. Atmos ChemPhys 16: 13885-13910. http://dx.doi.org/10.5194/acp-16-13885-2016.
Bowmaa KW: Shindell. DT: Wordea HM: Lamarque. JF: Young. PJ: Stevenson. PS: Ou. Z: De La torre. M:
Bergmana D: Cameron-Smith. PJ: Collins. WJ: Dohertv. R: Dalsoren. SB: Faluvegi. G: Folberth. G:
Horowitz. LW: Josse. BM: Lee. YH: Mackenzie. IA: Mvhre. 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: Wordea J. R. (2013). Evaluation of ACCMIP outgoing longwave radiation from tropospheric ozone
using TES satellite observations. Atmos ChemPhys 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: Hegglia 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.
Dahlmana 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: Stevensoa PS. (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:
Mvhre. G: Nganga. J: Prina R: Raga. G: Schultz. M: VanDorland. 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. FJ: Shindell. DT: Atherton.
C: Bergmana D: Duncaa 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.
Gauss. M: Mvhre. G: Isaksen. ISA: Grewe. V: Pitari. G: Wild. O: Collins. WJ: Dentener. FJ: Ellingsen. K:
Gohar. LK: Hauglustaine. DA: Iachetti. D: Lamarque. JF: Mancini. E: Micklev. LJ: Prather. MJ: Pvle. JA:
Sanderson. MG: Shine. KP: Stevenson. PS: 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-6-575-2006.
September 2019
9-22
PRAFT: Po Not Cite or Quote
-------
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: Jackmaa 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: Sua 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.
ClimRes 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/wgl/.
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/atmos7020027.
Lamv. K: Portafaix. T: Brogniez. C: Godin-Beekmana S: Bencherif. 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 ChemPhys 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: Mvhre. 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-17-0439.1.
Macintosh. CR: Allaa RP: Baker. LH: Bellouia 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. http://dx.doi.org/10.1029/94JD00Q21.
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/10.1029/2003JD003653.
September 2019
9-23
DRAFT: Do Not Cite or Quote
-------
Micklev. LJ: Murti. PP: Jacob. DJ: Logaa 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/10.1029/1999JD90Q439.
Mvhre. G: Aas. W: Cheriaa 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-17-2709-2017.
Mvhre. G: Shindell. D: Breoa 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; VBex; PMMidgley (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/wgl/.
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. htto://dx.doi.org/10.5194/acp-
17-14333-2017.
Parrish. DP: Lamarque. JF: Naik. V: Horowitz. L: Shindell. DT: Staehelia 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.
Sherwea 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.
Shindell. D: Faluvegi. G: Nazarenko. L: Bowmaa 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. http://dx.doi.org/10.1038/nclimatel835.
Shindell. D: Kuvlenstierna. JC: Vignati. E: vanDingenen. R: Amann. M: Klimont. Z: Anenberg. SC: Muller. N:
Janssens-Maenhout. G: Raes. F: Schwartz. J: Faluvegi. G: Pozzoli. L: Kupiainen. K: Hoglund-Isakssoa L:
Embersoa L: Streets. D: Ramanathaa 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/10.1029/2007id008753.
Shindell. DT: Faluvegi. G: Rotstava 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.
September 2019
9-24
DRAFT: Do Not Cite or Quote
-------
Shindell. DT: Voulgarakis. A: Faluvegi. G: Millv. G. (2012b). Precipitation response to regional radiative
forcing. Atmos ChemPhys 12: 6969-6982. http://dx.doi.org/10.5194/acp-12-6969-2012.
Skeie. RB: Berntsea TK: Mvhre. G: Tanaka. K: Kvalevag. MM: Hovle. CR. (2011). Anthropogenic radiative
forcing time series frompre-industrial times until 2010. Atmos ChemPhys 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
ChemPhys 12: 7725-7725. http://dx.doi.org/10.5194/acp-12-7725-2012.
Staehelia 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.
Stevensoa PS: Young. PJ: Naik. V: Lamarque. JF: Shindell. DT: Voulgarakis. A: Skeie. RB: Dalsorea SB:
Mvhre. G: Berntsea TK: Folberth. GA: Rumbold. ST: Collins. WJ: Mackenzie. I A: Dohertv. RM: Zeng. G:
vanNoiie. 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: Righi. 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 IntercomparisonProject (ACCMIP). Atmos ChemPhys 13:
3063-3085. http://dx.doi.org/10.5194/acp-13-3063-2013.
Westervelt. DM: Conlev. AJ: Fiore. AM: Lamarque. JF: Shindell. DT: Previdi. M: Mascioli. NR: Faluvegi. G:
Correa. G: Horowitz. LW. (2018). Connecting regional aerosol emissions reductions to local and remote
precipitation responses. Atmos ChemPhys 18: 12461-12475. http://dx.doi.org/10.5194/acp-18-12461-2018.
Wordea HM: Bowmaa 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. 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/sQ0376-016-
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: Stevensoa PS: Tilmes. S: Voulgarakis. A:
Wild. O: Bergmana D: Cameron-Smith. P: Cionni. I: Collins. WJ: Dalsorea SB: Dohertv. RM: Evring. V:
Faluvegi. G: Horowitz. LW: Josse. B: Lee. YH: Mackenzie. IA: Nagashima. T: Plummer. PA: Righi. M:
Rumbold. ST: Skeie. RB: Shindell. PT: 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 ChemPhys 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.
September 2019
9-25
PRAFT: Po Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
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. 2015a). 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/naaqs/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. Details related to specific quantitative analyses are not described in this
Appendix, but have been included in those specific appendices where the analyses are presented.
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 previous ISA for ozone, which 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 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;
September 2019
10-1
DRAFT: Do Not Cite or Quote
-------
1 (2) formation, transport, and fate of ozone in the environment; (3) measurement methods and ambient
2 concentrations of ozone; (4) how exposure assessment methods used in epidemiologic studies can
3 influence inferences drawn about ozone health effects; (5) the independent effect of ozone exposure on
4 health and welfare1 effects; (6) the potential influence of other factors (e.g., other pollutants in the
5 ambient air, ambient air temperature) shown to be correlated with ozone and health or welfare effects;
6 (7) the shape of the concentration-response relationship at ozone concentrations at the low end of the
7 distribution; and (8) populations and lifestages at increased risk of ozone-related health effects.
8 The literature search and screening process are described in the sections below and are
9 summarized in the Literature Flow Diagram shown below (Figure 10-2).
1 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.
September 2019 10-2 DRAFT: Do Not Cite or Quote
-------
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 (2015a).
Figure 10-1 General process for development of Integrated Science
Assessments.
September 2019
10-3
DRAFT: Do Not Cite or Quote
-------
Keyword Search
29,516
Keyword Search
Included
TopicClassified:
• Epidemiology
• Experimental
• Exposure
Keyword Search
Excluded
NotTopic
Classified
Citation Mapping
(22,801)
TopicSpecific:
• Climate
• Ecology
• Atmospheric-Background
Ozone
Literature Search Included
r
References
from Other
Sources
r A
Level 1 Title-Abstract
Screening Included
Level 2 Full-Text Screening Included
Level 1 Title-Abstract Screening Excluded
Level 2 Full-Text Screening Excluded
Included in Draft Ozone Integrated Science Assessment
(1,678)
Preface (20)
Executive Summary (13)
I ntegrated Sy nth esis (85)
Appendix 1: Atmospheric (187)
Appendix2: Exposure (193)
Appendix3: Respiratory (310)
Appendix4: Cardiovascular (188)
Appendix5: Metabolic(71)
Appendix6: Mortality (82)
Appendix7: Other Health (226)
Appendix8: Ecology (532)
Appendix9: Climate (53)
AppendixlO: Process(22)
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.
September 2019
10-4
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
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 draft ISA. Because the seed set is highly relevant to the topic of interest, this targeted approach to
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.
September 2019
10-5
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
reference 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.
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.
September 2019
10-6
DRAFT: Do Not Cite or Quote
-------
10000
"V
.rf*
3®
s5°v
1270
oSe
0.9
/S*
0^
~ Consider ~ Exclude ~ Not Screened
Figure 10-3 Summary of title/abstract screening in SWIFT-ActiveScreener.
Your Predicted Recall
-+~ Predicted Normal Screening
> 95% Inclusion
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 Example of screening efficiency using SWIFT-ActiveScreener,
September 2019
10-7
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
10.2.2.1 Atmospheric Science
Literature related to the atmospheric science topics discussed in Appendix 1 was identified by
topic-specific citation mapping methods that relied upon references cited in the 2013 Ozone ISA. More
specifically, references were collected from the atmospheric science sections of the 2013 Ozone ISA,
including 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 selected
from citation mapping were systematically title/abstract screened using SWIFT-AS. A recent critical
review of the science in which the U.S. EPA played an active role (Jaffa et al.. 2018) 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 individually reviewed. 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.
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 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.
September 2019
10-8
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
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 NOx ISA and the 2013 Ozone ISA, and a sufficiently
large set of negative seeds, including nonexperimental references cited in these two IS As. 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 identified 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 of
Nitrogen - 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 in past ISAs for ecological endpoints. 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
September 2019
10-9
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
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.cfm/proi ect/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.
References that were identified by topic classification from the keyword literature search are
tagged in the HERO database, as described in Figure 10-2. References from topic-specific citation
mapping are also tagged for each discipline. Finally, the references that passed through Level 1 screening
are tagged in HERO as Title-Abstract Screening Included. All inclusion and exclusion decisions are
documented in the HERO database, as well as discipline tags from automatic topic classification.
10.3 Study Selection: Full-Text Evaluation of Studies (Level 2)
Following Level 1 screening, NCEA subject matter experts 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 10.3.2V
September 2019
10-10
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
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. 2018). 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:
1. The study addresses ozone and its precursors;
September 2019
10-11
DRAFT: Do Not Cite or Quote
-------
1 2. The study addresses the human activities, natural processes, atmospheric
2 chemistry and dynamics that are responsible for formation in or introduction of
3 ozone into the lower troposphere (planetary boundary layer) where exposure to
4 ozone may lead to damaging effects to human health, ecosystems, the climate
5 system, and other elements of the environment relevant to human welfare. The
6 study is also relevant if it describes spatial and temporal patterns in lower
7 tropospheric ozone concentrations;
8 3. The study addresses the origins and concentrations of ozone in the U.S.; and
9 4. The study provides insight into the sources, causes, and concentrations of USB
10 ozone, the methodologies used to estimate USB ozone, and any technical issues
11 that must be considered in order to correctly interpret the scientific findings
12 relevant to USB ozone.
13 The studies cited in Appendix 1 met all four criteria. Also meeting these criteria were
14 supplemental studies that were included as necessary supporting material or studies identified outside of
15 topic-specific citation mapping.
10.3.1.2 Exposure Assessment
16 The ISA describes the exposure assessment methods employed in epidemiologic studies
17 discussed in the health appendices. For these studies, an additional screening step with added search terms
18 was applied to further identify potentially relevant references. The additional step was needed because the
19 Level 1 title/abstract screening was insufficient to gauge whether each paper covered the methods
20 discussed in the epidemiology literature. It was also needed because of the disparate nature of exposure
21 assessment references and the large number of references relative to other disciplines. The search terms
22 correspond to the sections of Appendix 2. Each of the references obtained through this search was then
23 evaluated in the Level 2 full-text screening for relevance.
24 Relevance was based on whether an exposure assessment study was representative of the
25 population and conditions addressed in the epidemiology literature. If there was sufficient evidence that a
26 method could provide an adequate representation of exposure from the U.S., there was no need to
27 consider studies conducted outside of the U.S. or Canada. If there was not sufficient evidence that a
28 method could provide an adequate representation of exposure from the U.S., then there was a need to
29 consider western European and Australian studies, which were the next most similar to studies conducted
30 in the U.S. If there was not sufficient evidence that a method could provide an adequate representation of
31 exposure from the U.S., Canada, western Europe, or Australia, then it was necessary to consider all
32 studies regardless of geographic location.
September 2019
10-12
DRAFT: Do Not Cite or Quote
-------
10.3.1.3 Health—Experimental Studies
1 For experimental studies, specifically, controlled human or animal exposure studies, the relevance
2 evaluation focused on those studies with appropriate study designs and relevant exposure concentrations,
3 as well as those that address key uncertainties and limitations in the evidence identified in the previous
4 review (Table 10-1). The scope of the experimental evidence encompassed studies of short-term
5 (i.e., hours to weeks) and long-term (i.e., months to years) exposures conducted at concentrations of
6 ozone that are relevant to the range of human exposures to ambient air (up to 2 ppm, which is one to two
7 orders of magnitude above ambient concentrations).
September 2019
10-13
DRAFT: Do Not Cite or Quote
-------
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.
Comparator: 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.
September 2019
10-14
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
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 (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.
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 Population: Any U.S. or Canadian population, including populations or lifestages that
respiratory effects might be at increased risk
Exposure: Short-term (on the order of one 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
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 one 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
Short-term exposure and
mortality
September 2019
10-15
DRAFT: Do Not Cite or Quote
-------
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 one 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 one 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
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
Short-term exposure and
cardiovascular effects
Short-term exposure and
nervous system effects
Long-term exposure and
cardiovascular, nervous
system, reproductive, or
developmental effects;
cancer, or mortality
September 2019
10-16
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
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)
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).
Comparator: 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. 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 and all available evidence was considered.
September 2019
10-17
DRAFT: Do Not Cite or Quote
-------
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
cycling 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
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)
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
Comparator = change in endpoint observed by unit increase in concentration of ozone in the same or in a control population;
exposure = environmental variable to which population is exposed; outcome = measurable endpoint resulting from exposure;
population = unit of study; 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, non-agricultural North American species were separated
out from the larger data sets and the evidence was evaluated (e.g., foliar injury, biomass).
10.3.1.6 Welfare—Effects on Climate
1 For effects on climate, the ISA focused on effects of tropospheric ozone on climate, consistent
2 with the inclusion of "climate" in the list of effects on welfare in Section 302(h) of the Clean Air Act. The
3 ISA does not focus on downstream ecosystem effects from changes in climate, climate-related human
4 health effects, or future air quality projections resulting from changes in climate. In addition, the ISA
Terrestrial water cycling; carbon
sequestration; terrestrial community
composition; plant reproduction, phenology,
or mortality; insects, other wildlife,
plant-animal signaling
September 2019
10-18
DRAFT: Do Not Cite or Quote
-------
1
2
3
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)
Comparator: 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
10.3.2 Individual Study Quality
4 After selecting studies for inclusion based on relevance, individual study quality was evaluated by
5 considering the design, methods, conduct, and documentation of each study, but not the study results. For
6 ISAs, the overall individual study quality process is described in the Preamble to the ISA (U.S. EPA.
7 2015a).
September 2019
10-19
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
The process for individual study quality criteria has been refined by discipline with each 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 used a broad set of questions from the Preamble to review study quality,
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.
10.3.2.1 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 USB ozone. Study quality in this context is established using the following criteria:
(1) quality assurance and 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 were subject to 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
September 2019
10-20
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
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
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 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 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, the study quality tables were
used to develop prompting questions 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).
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.
September 2019
10-21
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
10.3.2.4 Ecological Approach
Worldwide, 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 Preamble as criteria for reviewing the quality of individual ecological studies
within this ISA. The Preamble provides 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?
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 these study quality questions 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.
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.
September 2019
10-22
DRAFT: Do Not Cite or Quote
-------
10.4 Peer Review and Public Participation
1 Peer review is an important component of any scientific assessment. U.S. EPA has formal
2 guidance about peer review in the Peer Review Handbook (U.S. EPA. 2015b). and this ISA follows all
3 the policies and procedures identified therein. Additionally, this ISA follows all the guidelines of the
4 Information Quality Guidelines (U.S. EPA. 2002).
5 U.S. EPA has designated this ISA as a Highly Influential Scientific Assessment, which is defined
6 by The Office of Management and Budget's Final Information Quality Bulletin for Peer Review
1 (hereafter, "Peer Review Bulletin") as:
8
9
10 A subset of Influential Scientific Information that is a scientific assessment (i.e., an
11 evaluation of a body of scientific or technical knowledge, which typically synthesizes
12 multiple factual inputs, data, models, assumptions and/or applies best professional
13 judgment to bridge uncertainties in the available information) that "could have a potential
14 impact of more than $500 million in any year on either the public or private sector" or "is
15 novel, controversial, or precedent-setting, or has significant interagency interest."
16 (https://obamawhitehouse.archives.gov/omb/memoranda fv2005 m05-03/).
17
18 As such, there are additional review and transparency steps required in the release of this
19 information. These steps are described below. CASAC serves an important role in reviewing this
20 ISA (see Section 10.4.5).
10.4.1 Call for Information
21 Consistent with the Preamble, a Call for Information was published in the Federal Register on
22 June 26, 2018 (83 FR 29785). The purpose of this Call for Information was announcing the beginning of
23 the review cycle of the air quality criteria and the ozone NAAQS. Specifically, the Call for Information
24 stated that U.S. EPA would be preparing an Integrated Review Plan and Integrated Science Assessment.
25 The public was given 30 days ".. .to assist the U.S. EPA by submitting information regarding significant
26 new ozone research and policy-relevant issues for consideration in this review of the primary
27 (health-based) and secondary (welfare-based) ozone standards." U.S. EPA received 14 comments via the
28 Federal eRulemaking Portal (http ://www.regulations.gov).
29 In previous assessments, U.S. EPA has held a kick-off workshop to begin a review cycle. The
30 workshop brought together subject area experts and the public to highlight significant new and emerging
31 research and make recommendations to the U.S. EPA regarding the design and scope of the review.
32 However, certain process efficiencies were discussed in the Administrator's May 9, 2018 memorandum,
33 "Back-to-Basics Process for Reviewing National Ambient Air Quality Standards."
September 2019
10-23
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
(https://www.epa.gov/sites/production/files/2018-05/documents/image2018-05-Q9-173219.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 kick-off workshop was held for this review cycle and the Call for
Information served as the formal initiation of the NAAQS review process.
10.4.2 Integrated Review Plan
Following the Call for Information, U.S. EPA prepared an Integrated Review Plan (IRP) that
summarizes 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 .re gulations. 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
preliminary drafts of all appendices for initial peer input. Peer input is a process that allows U.S. EPA to
solicit feedback from subject-matter experts 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 experts across four discipline areas
(link to Front Matter page with list of reviewers') to participate in the peer input workshops and for
coordinating all communication for this consultation. A Federal Register Notice was issued on October
September 2019
10-24
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
23, 2018 announcing the workshops and included information for public access to the discussions (83 FR
53472). 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. 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 document.
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.
10.4.5 Clean Air Scientific Advisory Committee (CASAC) Peer Review
The Clean Air Act governs the NAAQS review process, and also includes instruction about
review of science and policy documents developed by U.S. EPA (CAA. 1990). Section 109(d)(2)
addresses the appointment and advisory functions of an independent scientific review committee.
Section 109(d)(2)(A) requires the Administrator to appoint this committee, which is to be composed of
"seven members including at least one member of the National Academy of Sciences, one physician, and
one person representing State air pollution control agencies." Section 109(d)(2)(B) provides that the
independent scientific review committee "shall complete a review of the criteria.. .and the national
primary and secondary ambient air quality standards...and shall recommend to the Administrator any
September 2019
10-25
DRAFT: Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
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.
Release of this draft ISA has been announced in the Federal Register. This also begins a period of
time for the public to provide comment on this draft. Additionally, the CASAC will hold a public meeting
to discuss the draft ISA and provide an independent scientific peer review of the document.
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 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 draft ISA. More specifically, QA was conducted on all
appendices, and the numbers from every tenth reference, or in some instances more, were checked 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 QA processes in place.
Additionally, 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 will perform a Technical System Audit on this draft before final release. This audit verifies that
the appropriate QA procedures, criteria, reviews, and data verification are adequately performed and
documented.
September 2019
10-26
DRAFT: Do Not Cite or Quote
-------
10.6 Conclusion
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
endeavors to add additional steps as needed. This Appendix will be updated before final release to include
information about the CAS AC review, QA audit, and any other process developments.
September 2019
10-27
DRAFT: Do Not Cite or Quote
-------
10.7 References
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.
Cohea 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.
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 Mufios. 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: Thayer. K. (2016). SWIFT-Review: a text-mining workbench for
systematic review. SystRev 5: 87. http://dx.doi.org/10.1186/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/wgl/.
Jaffe. DA: Cooper. OR: Fiore. AM: Henderson. BH: Tonnesen. GS: Russell. AG: Henze. DK: Langford. AO:
Lia M: Moore. T. (2018). Scientific assessment of background ozone over the US: Implications for air
quality management. 6. http://dx.doi.org/10.1525/elementa.309.
Jerrett. M: Arain. A: Kanaroglou. P: Beckermaa 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.750Q388.
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.
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.
Rvaa 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.
September 2019
10-28
DRAFT: Do Not Cite or Quote
-------
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). (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.
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://ofmpub.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: Wilsoa JG: Zhaa 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.
September 2019
10-29
DRAFT: Do Not Cite or Quote
------- |