Review of the National Ambient Air Quality
Standards for Ozone:
Policy Assessment of Scientific
and Technical Information
OAQPS Staff Paper - First Draft
U S EPA Headquarters Library
Mail code 3404T
1200 Pennsylvania Avenue NW
Washington, DC 20460
202-566-0556
D
-------
-------
EPA-452/D-05-002
November 2005
Review of the National Ambient Air Quality
Standards for Ozone:
Policy Assessment of Scientific
and Technical Information
OAQPS Staff Paper - First Draft
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina
-------
DISCLAIMER
This document has been reviewed by the Office of Air Quality Planning and Standards
(OAQPS), U.S. Environmental Protection Agency (EPA), and approved for publication. This
first draft OAQPS Staff Paper contains the preliminary draft conclusions of the staff of the
OAQPS and does not necessarily represent those of the EPA. Mention of trade names or
commercial products is not intended to constitute endorsement or recommendation for use.
-------
TABLE OF CONTENTS
List of Tables viii
.List of Figures x
1. INTRODUCTION 1-1
1.1 PURPOSE 1-1
1,2 BACKGROUND '. 1-2
1.2.1 Legislative Requirements 1-2
1.2.2 History of Ozone NAAQS Reviews 1-3
1.2.3 Litigation Related to the 1997 Ozone Standards 1-5
1.2.4 Current Ozone NAAQS Review '. 1-6
1.3 GENERAL APPROACH AND ORGANIZATION OF THE DOCUMENT 1-7
REFERENCES ..: 1-9
2. AIR QUALITY CHARACTERIZATION 2-1
2.1 INTRODUCTION ': 2-1
2.2 CHEMICAL AND PHYSICAL PROPERTIES, FORMATION, AND
TRANSPORT 2-1
2.2.1 Chemical and Physical Properties 2-1
2.2.2 Formation 2-2
2.2.3 Transport 2-2
2.2.4 Precursors, Sources and Emissions 2-3
2.2.5 Tropospheric vs. Stratospheric Ozone 2-3
2.3 DATA SOURCES 2-8
2.3.1 Air Quality System (AQS) 2-8
2.3.2 CASTNET 2-10
2.4 OZONE MONITORING METHODS AND ISSUES 2-10
2.5 CHARACTERIZATION OF GROUND-LEVEL OZONE
CONCENTRATIONS 2-11
2.5.1 Metrics 2-11
2.5:2 Spatial Variability 2-12
2.5.2.1 8-hour and 1-hour Statistics 2-12
2,5.2.2 Cumulative Seasonal Statistics 2-12
2.5.3 Temporal Variability 2-15
2.5.3.1 Long Term Variability - Trends 2-15
2.5.3.2 Seasonal Variability 2-24
2.5.3.3 Short Term Variability-Diurnal 2-24
2.6 CHARACTERIZATION OF OZONE EPISODES 2-29
'2.7 BACKGROUND LEVELS 2-36
REFERENCES 2-43
November 2005 i Draft - Do Not Quote or Cite
-------
3. POLICY-RELEVANT ASSESSMENT OF HEALTH EFFECTS EVIDENCE 3-1
3.1 INTRODUCTION 3,1
3.2 MECHANISMS 3-3
3.2.1 Direct Pulmonaiy Effects 3-3
3.2.1.1 Breathing Pattern Changes 3-4
3.2.1.2 Symptoms and Lung Function Changes 3-4
3.2.1.3 Airway Hyperresponsiveness 3-4
3.2.2 Extrapulmonary Effects 3-5
3.3 NATURE OF EFFECTS 3-6
3.3.1 Morbidity 3-6
3.3.1.1 Emergency Department Visits/Hospital Admissions for Respiratory
Causes 3-7
3.3.1.2 Effects on the Respiratory System from Short-term Exposures 3-8
3.3.1.2.1 Pulmonary Function Decrements, Respiratory Symptoms,
and Asthma Medication Use 3-9
3.3.1.2.2 Airway Responsiveness 3-12
3.3.1.2.3 Respiratory Inflammation and Permeability 3-13
3.3.1.2.4 Changes in Host Defense Capability 3-15
3.3.1.2.5 Morphological Effects 3-16
3.3.1.2.6 Effects on Exercise Performance 3-17
3.3.1.2.7 Increased School Absences 3-17
3.3.1.3 Effects on the Respiratory System from Long-term Exposures 3-18
3.3.1.3.1 Seasonal Ozone Effects on Lung Function 3-18
3.3.1.3.2 Reduced Baseline Lung Function and Respiratory Symptoms 3-19
3.3.1.3.3 Long-term Oa Exposure and Respiratory Inflammation 3-21
3.3.1.3.4 Risk of Asthma Development 3-22
3.3.1.3.5 Morphological Effects : 3-23
3.3.1.3.6 Summary 3-23
3.3.1.4 Effects on the Cardiovascular System 3-23
3.3.2 Premature Mortality 3-25
3.3.2.1 Mortality and Short-term O3 Exposure 3-25
3.3.2.2 Mortality and Long-term O3 Exposure 3-28
3.3.3 Summary 3-30
3.4 INTEGRATIVE ASSESSMENT OF EVIDENCE FROM EPIDEMIOLOGICAL
STUDIES 3-31
3.4.1 Strength of Associations 3-32
3.4.2 Robustness of Associations 3-33
3.4.2.1 Exposure Error 3-33
3.4.2.2 Confounding by Copollutants 3-34
3.4.2.3 Model Specification -. 3-35
3.4.3 Consistency .- 3-36
3.4.4 Temporality 3-37
3.4.5 Lag Structure in Short-term Exposure Studies 3-37
November 2005 ii Draft - Do Not Quote or Cite
-------
3.4.6 Concentration-Response Relationships and Potential Thresholds 3-39
3.5 BIOLOGICAL PLAUSIBILITY AND COHERENCE OF EVIDENCE 3-40
3.5.1 Coherence and Plausibility of Short-term Effects on the Respiratory System
3-41
3.5.2 Coherence and Plausibility of Effects on the Cardiovascular System 3-48
3.5.3 Coherence and Plausibility of Effects Related to Long-Term O3
Exposure 3-49
3.5.4 Coherence and Plausibility of Mortality-Related Health Endpoints 3-50
3.6 OZONE-RELATED IMPACTS ON PUBLIC HEALTH 3-51
3.6.1 Factors which Modify Responsiveness to Ozone 3-51
3.6.2 Susceptible Population Groups : 3-53
3.6.2.1 Active People 3-53
3.6.2.2 People with Lung Disease 3-53
3.6.2.3 Children and Older Adults 3-54
3.6.2.4 People with Increased Responsiveness to Ozone 3-56
3.6.2.5 Other Population Groups 3-57
3.6.3 What Constitutes an Adverse Health Impact from Ozone Exposure? 3-57
3.7 SUMMARY AND CONCLUSIONS FOR OZONE HEALTH EFFECTS 3-63
1 3.7.1 Morbidity Health Effects of Acute (Short-term) Exposures to Ozone 3-63
3.7.2 Mortality-Related Health Effects of Short-term Exposures to Ozone 3-67
3.7.3 Heallh Effects of Repeated Short-term Exposures to Ozone 3-67
3.7.4 Health Effects of Long-term Exposures to Ozone 3-68
3.7.5 Health Effects of Binary Pollutant Mixtures Containing Ozone 3-68
3.7.6 Populations At Risk and Susceptibility Factors Associated With Ozone
Exposure 3-69
REFERENCES 3-71
4. CHARACTERIZATION OF HUMAN EXPOSURE TO OZONE 4-1
4.1 INTRODUCTION 4-1
4.2 OZONE EXPOSURE STUDIES 4-2
4.2.1 Exposure Concepts and Definitions 4-2
4.2.2 Monitoring Equipment Considerations 4-3
4.2.3 Personal Ozone Exposure Assessment Studies 4-4
4.2.4 Microenvironmental Studies 4-5
4.3 EXPOSURE MODELING 4-5
4.3.1 The APEX Model 4-6
4.3.2 Key Algorithms 4-7
4.3.3 Model Output 4-11
4.3.4 Limitations of the Model 4-13
4.3.4.1 Estimation of Ambient Air Quality 4-13
4.3.4.2 Estimation of Concentrations in Indoor Microenvironments 4-14
4.3.4.2.1 Air Exchange Processes 4-14
4.3.4.2.2 Deposition Processes... 4-15
4.3.4.2.3 Chemical Reaction Processes 4-15
November 2005 iii Draft - Do Not Quote or Cite
-------
4.3.4.3 Characterization of Population Demographics and Activity Patterns 4-15
4.3.4.4 Modeling Physiological Processes 4-16
4.4 SCOPE OF EXPOSURE ASSESSMENTY 4-16
4.4.1 Selection of Urban Areas to be Modeled 4-16
4.4.2 Time Periods Modeled 4-16
4.4.3 Populations Modeled '. 4-17
4.5 INPUTS TO THE EXPOSURE MODEL 4-18
4.5.1 Population Demographics 4-18
4.5.2 Population Commuting Patterns 4-19
4.5.3 Human Activity Data 4-19
4.5.4 Physiological Data 4-20
4.5.5 Microenvironments Modeled 4-23
4.5.6 Ambient Ozone Concentrations 4-23
4.5.7 Meteorological Data 4-25
4.6 EXPOSURE ASSESSMENT RESULTS 4-25
REFERENCES 4-33
5. CHARACTERIZATION OF HEALTH RISKS 5-1
5.1 INTRODUCTION 5-1
5.1.1 Overview of Risk Assessment From Last Review 5-1
5.1.2 Development of Approach for Current Risk Assessment 5-2
5.2 SCOPE OF OZONE HEALTH RISK ASSESSMENT 5-4
5.2.1 Selection of Health Endpoint Categories 5-5
5.2.2 Selection of Study Areas.. 5-7
5.2.3 Air Quality Considerations 5-8
5.3 COMPONENTS OF THE RISK MODEL 5-9
5.3.1 Assessment of Risk Based on Controlled Human Exposure Studies 5-9
5.3.1.1 General Approach 5-9
5.3.1.2 Exposure Estimates 5-13
5.3.1.3 Exposure Response Functions 5-13
5.3.1.4 Characterizing Uncertainty and Variability 5-14
5.3.2 Assessment of Risk Based on Epidemiological Studies.: 5-16
5.3.2.1 General Approach 5-16
5.3.2.2 Air Quality Considerations 5-19
5.3.2.3 Concentration-Response Functions 5-20
5.3.2.4 Baseline Health Effects Incidence and Population Estimates 5-23
5.3.2.5 Characterizing Uncertainty and Variability 5-27
5.4 OZONE RISK ESTIMATES 5-30
5.4.1 Recent Air Quality 5-30
5.4.2 Just Meeting Current Ozone Standards 5-48
5.4.3 Just Meeting Alternative Ozone Standards [to be included in next draft
of Staff Paper] : 5-61
5.4.4 Key Observations [to be included in next draft of Staff Paper] 5-61
REFERENCES 5-62
November 2005 iv Draft - Do Not Quote or Cite
-------
6. STAFF CONCLUSIONS AND RECOMMENDATIONS ON PRIMARY O3
NAAQS : 6-1
6.1 INTRODUCTION 6-1
6.2 APPROACH 6-2
6.3 PRIMARY O3 STANDARD...: 6-4
6.3.1 Adequacy of Current O3 Standard 6-4
6.3.1.1 Evidence-based Considerations 6-5
6.3.1.1.1 Respiratory Morbidity : 6-5
6.3.1,1.2 Cardiovascular Morbidity 6-7
6.3.1.1.3 Mortality 6-8
6.3.1.2 Risk-based Considerations 6-8
6.3.1.3 Summary 6-14
6.3.2 Indicator 6-14
6.3.3 Averaging Time 6-15
6.3.3.1 Short-Term and Prolonged (1 to 8 hours) : 6-15
6.3.3.2 Long-Term..... .' 6-16
6.3.4 Form ! 6-17
6.3.5 Level 6-18
6.3.6 Summary of Alternative Standards to Be Considered in Additional Exposure
and Risk Analyses '. 6-22
6.4 SUMMARY OF KEY UNCERTAINTIES AND RESEARCH
RECOMMENDATIONS RELATED TO SETTING A PRIMARY 03 STANDARD
[TO BE INCLUDED IN THE NEXT DRAFT STAFF PAPER]. 6-22
REFERENCES 6-23
7. POLICY-RELEVANT ASSESSMENT OF WELFARE EFFECTS EVIDENCE.. 7-1
7.1 INTRODUCTION 7-1
7.2 EFFECTS ON VEGETATION 7-2
7.2.1 Exposure Methodologies Used in Vegetation Research 7-3
7.2.2 Species (Intra-Plant) Response/Mode of Action 7-6
7.2.2.1 Entry of Ozone into the Leaf 7-6
7.2.2.2 Reactions of Oa and possible reaction product(s) at cell surfaces 7-8
7.2.2.3 Movement of an O3 reaction product(s) into the cell with enzymatic or
chemical transformation of those products in the cell 7-8
7.2.2.4 Ozone Initiated Wounding and Pathogen Attack Response 7-9
7.2.2.5 Defense and Compensation Mechanisms 7-10
7.2.2.6 Changes to Plant Metabolism 7-10
7.2.2.7 Relationships Between Age and Size and Ozone Response 7-12
7.2.3 Factors That Modify Functional and Growth Response 7-12
7.2.3.1 Genetics 7-13
7.2.3.2 Biological Factors 7-13
7.2.3.3 Physical Factors •. 7-14
7.2.3.4 Chemical Factors 7-15
November 2005 v Draft - Do Not Quote or Cite
-------
7.2.4 Effects-Based Air Quality Exposure Indices 7-17
7.2.4.1 Concentration-based Forms 7-18
7.2.4.2 Flux-based forms 7-20
7.2.4.3 The Critical Level Approach 7-21
7.2.4.4 Summary 7-22
7.2.5 Ozone Exposure-Plant Response Relationships 7-23
7.3 VEGETATION IMPACT ASSESSMENT 7-27
7.3.1 Overview 7-27
7.3.2 Air Quality Analysis 7-29
7.3.2.1 Monitor coverage 7-29
7.3.2.2 Modeling tools 7-30
7.3.2.3 Generation of Potential Ozone Exposure Surfaces 7-33
7.3.2.4 Alternative Air quality scenarios 7-34
7.3.3 Crop Risk/Benefits Assessments 7-34
7.3.3.1 Exposure Assessment 7-36
7.3.3.2 Crop yield loss Assessment 7-36
7.3.3.3 Economic Benefits Assessment - AGSIM 7-38
7.3.4 Tree Seedling/Mature Tree/Forest Species Quantitative/Qualitative Risk
Assessments 7-38
7.3.4.1 Selected Species Exposure Assessment 7-39
7.3.4.2 Selected Tree Seedling Biomass Loss 7-39
7.3.4.3 Foliar Injury Incidence/Epidemiology - FIA Data 7-39
7.3.4.4 Ponderosa Pine case study for mature tree 7-44
7.4 ECOSYSTEM CONDITION, FUNCTION AND SERVICES 7-45
7.4.1 Evidence Demonstrating the Potential for Ozone to Alter Ecosystem Structure
and Function 7-48
7.4.2 Effects on Ecosystem Products and Services 7-49
7.4.2.1 Carbon Sequestration 7-49
7.4.2.2 Water Resources 7-50
7.4.3 Research needs 7-50
REFERENCES 7-53
APPENDIX 3A:
Ozone Epidemiological Study Results 3A-1
APPENDIX 4A:
Microenvironment Modeling Parameters 4A-1
APPENDIX 4B:
Frequency Distributions of Daily Maximum 8-Hour Average Ozone Concentrations
4B-1
APPENDIX 4C:
Exposures for Exercising Children 4C-1
November 2005
VI
Draft - Do Not Quote or Cite
-------
APPENDIX 4D:
Grid Points used for Policy Relevant Background in 12 Cities.
.4D-1
APPENDIX 5A:
Monitor-Specific Oj Air Quality Information 5A-1
APPENDIX SB:
Summary of Locations, Concentration-Response Functions, Months Included and
Counties Included 5B-1
APPENDIX 5C:
Tables of Study-Specific Information 5C-1
November 2005
VII
Draft - Do Not Quote or Cite
-------
List of Tables
Number
Page
2-1, VOC emission sources, 1970-2004 2-4
2-2. NOX emission sources, 1970-2004 2-6
3-1. Acute Oa-induced Physiological and Biochemical Changes in Humans and Animals 3-42
3-2. Gradation of Individual Responses to Short-Term Ozone Exposure in Healthy Persons. 3-60
3-3. Gradation of Individual Responses to Short-Term Ozone Exposure in Persons with
Impaired Respiratory Systems 3-61
3-4. Summary of Ozone-Induced Respiratory Health Effects from Clinical Studies 3-65
4-1. Exertion Levels in Terms of Equivalent Ventilation Rates 4-12
4-2. Urban Areas and Time Periods Modeled 4-17
4-3. Population Coverage of Modeled Areas 4-18
4-4. Studies in CHAD 4-21
4-5. Microenvironments Modeled ^ 4-23
4-6. 2002-2004 8-Hour Ozone Design Values for the Modeled Areas 4-24
4-7. Numbers of Person-Days Over the Modeled Periods With Daily Maximum 8-Hour Average
Exposures Above 0.08 ppm Under Moderate Exertion 4-27
4-8. Numbers of Persons With At Least One Daily Maximum 8-Hour Average Exposure Above
0.08 ppm Under Moderate Exertion Over the Modeled Periods 4-28
4-9. Numbers of Persons With At Least One Daily Maximum 8-Hour Average Exposure Above
0.08 ppm Under Moderate Exertion Over the Modeled Periods (percent of population
group) 4-29
4-10. Numbers of Person-Days Over the Modeled Periods with Daily Maximum 8-Hour
Average Exposures Above 0.07 ppm Under Moderate Exertion 4-30
4-11. Numbers of Persons With at Least One Daily Maximum 8-hour Average Exposure Above
0.07 ppm Under Moderate Exertion Over the Modeled Period 4-31
4-12. Numbers of Persons With at Least One Daily Maximum 8-hour Average Exposure Above
0.07 ppm Under Moderate Exertion Over the Modeled Period (percent of population
group) 4-32
5-1. Locations and Health Endpoints Included in the Ozone Risk Assessment Based on
Epidemiological Studies 5-22
5-2. Relevant Population Sizes for Ozpme Risk Assessment Locations* 5-24
5-3. Baseline Mortality Rates (per 100,000 Population) for 2002 for Ozone Risk Assessment
Locations* 5-25
5-4. Baseline Rates for Hospital Admissions .• 5-26
5-5. Comparison of Number and Percent of School Age Children and Active School Age
Children Estimated to Experience Lung Function Responses Associated with 8-Hour
Ozone Exposure While Engaged in Moderate Exertion 5-32
November 2005
vm
Draft - Do Not Quote or Cite
-------
5-6. Comparison of Number and Percent of Occurrences of Lung Function Responses
Associated with 8-Hour Ozone Exposure While Engaged in Moderate Exertion 5-34
5-7. Estimated Non-Accidental Mortality Associated with Recent Ozone Concentrations (April -
September, 2004) 5-41
5-8. Estimated Cardiovascular and Respiratory Mortality Associated with Recent Ozone
Concentrations (April - September, 2004) 5-44
5-9. Estimated Hospital Admissions Associated with Recent Ozone Concentrations in New
York, NY* (April - September, 2004) 5-46
5-10. Estimated Non-Accidental Mortality Associated with Ozone Above Background When the
Current 8-Hour Standard is Just Met (April - September) 5-54
5-11. Estimated Cardiovascular and Respiratory Mortality Associated with Ozone
Concentrations: that Just Meet the Current 8-Hour Daily Maximum Standard (April -
September): 5-57
5-12. Estimated Hospital Admissions Associated with Ozone Above Background In New York,
NY* When the Current 8-Hour Standard is Just Met (April - September) 5-59
6-1. Summary of Estimates of Number of people Exposed and Number of Occurrences
Associated with 8-Hour Daily Maximum Ozone Concentrations Above 0.08 ppm for 12
Urban Areas in the U.S. (from Tables 4-7 and 4-8) 6-11
6-2. Summary of Comparison of Median Number of Children and Median Number of
Occurrences Associated with 8-Hour Ozone Exposure Among All Children and Among
Active Children While Engaged in Moderate Exertion in 12 Urban Areas in the U.S. (from
Tables 5-5 and 5-6) , 6-12
November 2005
IX
Draft - Do Not Quote or Cite
-------
List of Figures
Number
Page
2-1. Locations of Monitors from AQS and CASTNET 2-11
2-2. 4th Highest Daily Maximum 8-Hour Values in U.S. Counties, 2002-2004 2-13
2-3 . 2nd Highest Daily Maximum 1-Hour Values in U.S. Counties, 2002-2004 2-14
2-4. Highest 3-month SUM06 Exposure Index in US. Counties, 2001 AQS Data 2-16
2-5. Highest 3-month W126 Exposure Index in U.S. Counties, 2001 AQS Data 2-17
2-6. Highest 3-month SUM06 Exposure Index in U.S. Counties, 2001
CASTNET Data. 2-18
2-7. Highest 3-month W126 Exposure Index in U.S. Counties, 2001
CASTNET Data 2-19
2-8. 4th highest daily maximum 8-hour ozone values 1990-2004 (Urban) 2-20
2-9. 4th highest daily maximum 8-hour ozone values 1990-2004 (Rural) 2-21
2-10. 2nd highest daily maximum 1-hour ozone values 1990-2004 (Urban) 2-22
2-11. 2nd highest daily maximum 1-hour ozone values 1990-2004 (Rural) 2-23
2-12. 2nd highest daily maximum 1 -hour ozone values from 2004 by month '. 2-25
2-13. 4th highest daily maximum 8-hour ozone values from 2004 by month 2-26
2-14. 1-Hour Diurnal Week Day Pattern for Urban Sites, May through September 2004 2-27
2-15. 8-Hour Diurnal Week Day Pattern for Urban Sites, May through
September 2004 2-28
2-16. 1-Hour Week Day Diurnal Pattern for Rural Sites, May through
September 2004 '. 2-30
2-17. 1-Hour Week End Diurnal Pattern for Rural Sites, May through
September 2004 2-31
2-18. 8-Hour Week Day Diurnal Pattern for Rural Sites, May through
September 2004 2-32
2-19. 8-Hour Week End Diurnal Pattern for Rural Sites, May through
September 2004 2-33
2-20. Length of Episodes over 0.12 ppm by Year for 1-hour O3 Data 2-34
2-21. Length of Episodes over 0.08 ppm by Year for 8-hour 63 Data. 2-35
2-22. Length of Episodes over Levels for 1-hour O3 Data (2000-2004) 2-37
2-23. Length of Episodes over Levels for 8-hour O3 Data (2000-2004) 2-38
2-24. Length of Gaps in Days Between Episodes over O.OSppm for 8-hour O3
Data (2000-2004) 2-39
2-25. Length of Gaps in Days Between Episodes over 0.12ppm for 8-hour O3
Data (2000-2004) 2-40
3-1. Resolution Time-Line for Acute Ozone-Induced Physiological and Biochemical
Responses in Humans 3-43
3-2. Postulated Cellular and Molecular Changes in Human Airway Epithelial Cells on Acute
Exposure to Ozone..... 3-44
3-3. Effect Estimates from U.S. and Canadian Studies 3-46
November 2005
Draft-Do Not Quote or Cite
-------
4-1 APEX Diagram 4-8
5-1. Major Components of Ozone Health Risk Assessment Based on Controlled Human
Exposure Studies 5-11
5-2. Major Components of Ozone Health Risk Assessment Based on Epidemiological
Studies ; 5-17
5-3. Estimated (Non-Accidental) Mortality Associated with Short-Term Exposure to Ozone
Above Background: Single-Pollutant, Single-City Models ,
(April - September, 2004) 5-37
5-4. Estimated (Non-Accidental) Mortality Associated with Short-Term Exposure to Ozone
Above Background (April — September, 2004): Single-City Model (left bar) vs.
Multi-City Model (right bar) ^ 5-38
5-5. Estimated Cardiovascular and Respiratory Mortality Associated with Short-Term
Exposure to Ozone Above Background (April - September, 2004): Single-City Model
(left bar) vs. Multi-City Model (right bar) - Based on Huang et al. (2004) .' 5-39
5-6. Estimated Cardiovascular and Respiratory Mortality Associated with Short-Term Exposure
to Ozone Above Background (April - September, 2004): Single-Pollutant vs. Multi-
Pollutant Models [Huang et al. (2004), additional pollutants, from left to right: none,
PM10,NO2,SO2,CO].: ". 5-40
5-7. Estimated (Unscheduled) Hospital Admissions in Detroit Associated with Short-Term
Exposure to Ozone Above Background (April - September, 2004): Different Lag Models
- Based on Ito (2003) [bars from left to right are 0-day, 1-day, 2-day, and 3-day lag
models] * 5-50
5-8. Estimated (Non-Accidental) Mortality Associated with Short-Term Exposure to Ozone
Above Background When the Current 8-Hour Standard is Just Met (April - September):
Single-City Model (left bar) vs. Multi-City Model (right bar)..:. 5-51
5-9. Estimated Cardiovascular and Respiratory Mortality Associated with Short-Term Exposure
to Ozone Above Background When the Current 8-Hour Standard is Just Met (April -
September): Single-City Model (left bar) vs. Multi-City Model (right bar) - Based on
Huang et al. (2004) ; 5-52
5-10. Estimated Cardiovascular and Respiratory Mortality Associated with Short-Term Exposure
to Ozone Above Background When the Current 8-Hour Standard is Just Met (April -
September): Single-Pollutant vs. Multi-Pollutant Models [Huang et al. (2004), additional
pollutants, from left to'right: none,PM10,NO2, SO2, CO] 5-53
6-1. Odds ratios for associations between Os and respiratory symptoms; and effect estimates for
associations between'Os and hospitalization for respiratory diseases 6-21
7-l(a-c). Major Components of Planned Environmental Assessment 7-28
7-2. Locations of AQS monitors (top) .and CASTNET monitoring stations 7-32
7-3. Range of planted soybean for the year 2001 7-37
7-4. County-level 4ht highest maximum 8-hour averages for 2001 7-42
7-5. County-level max 3month 12-hour SUM06 for 2001 7-43
7-6. Common anthropogenic stressors and the essential ecological attributes they affect.
Modified from Young and Sanzone (2002) 7-47
November 2005
XI
Draft - Do Not Quote or Cite
-------
1 1. INTRODUCTION
2 1.1 PURPOSE
3 This draft Staff Paper, prepared by staff in the U.S. Environmental Protection Agency's
4 (EPA) Office of Air Quality Planning and Standards (OAQPS), evaluates the policy implications
5 of the key studies and scientific information contained in the draft document, Air Qualify
6 Criteria for Ozone and Related Photochemical Oxidants: T* External Review Draft (USEPA,
7 2005b; henceforth referred to as the draft CD), prepared by EPA's National Center for
8 Environmental Assessment (NCEA). This draft Staff Paper also presents and interprets initial
9 results from several staff analyses (e.g., air quality analyses, human exposure analyses, and
10 human health risk assessments) and discusses plans for a staff environmental assessment of
11 vegetation-related impacts. Staff believes that these analyses should be considered in EPA's
12 current review of the national ambient air quality standards (NAAQS) for ozone (63). This draft
13 Staff Paper identifies alternative standard options for purposes of conducting additional exposure
14 and risk analyses but does not present staff conclusions and recommendations as to potential
15 revisions of the primary (health-based) and secondary (welfare-based) Oa NAAQS.
16 The policy assessment to be presented in the final version of this Staff Paper is intended
17 to help "bridge the gap" between the scientific review contained in the draft CD and the
18 judgments required of the EPA Administrator in determining whether it is appropriate to revise
19 the NAAQS for Os. Emphasis will be placed on identifying those conclusions and uncertainties
20 in the available scientific literature that the staff believes should be considered in selecting an
21 indicator, averaging times, forms1, and levels for the primary (health-based) and secondary
22 (welfare-based) standards, which must be considered collectively in evaluating the health and
23 welfare protection afforded by O3 standards. The final Staff Paper will evaluate the policy
24 implications of the key studies and scientific information contained in the final CD (targeted for
25 completion by February 2006), identify the critical elements that EPA staff believes should be
26 considered in the current review of the NAAQS for Os, and present factors relevant to the
27 evaluation of current primary and secondary Os NAAQS, as well as staff conclusions and
28 recommendations of options for the Administrator to consider.
29 This draft Staff Paper is being provided to CASAC and the public for review at a meeting
30 planned for December 2005. Following that meeting, staff will complete the human exposure
" - The "form" of a standard defines the air quality statistic that is to be compared to the level of the standard
in determining whether an area attains the standard.
November 2005
1-1
Draft - Do Not Quote or Cite
-------
1 analyses and health risk assessment and conduct an environmental assessment of vegetation-
2 related impacts. Taking CAS AC and public comments on this first draft Staff Paper into
3 account, staff will prepare a second draft Staff Paper, to be based on the final CD, and will make
4 that draft document available for further review and comment by CASAC and the public.
5 While this document should be of use to all parties interested in the 63 NAAQS review, it
6 is written for those decision makers; scientists, and staff who have some familiarity with the
7 technical discussions contained in the draft CD.
8 1.2 BACKGROUND
9 1.2.1 Legislative Requirements
10 Two sections of the Clean Air Act (Act) govern the establishment and revision of the
11 NAAQS. Section 108 (42 U.S.C: 7408) directs the Administrator to identify "air pollutants" that
12 "in his judgment, may reasonably be anticipated to endanger public health and welfare" and
13 whose "presence... in the ambient air results from numerous or diverse mobile or stationary
14 sources" and, if listed, to issue air quality criteria for them. These air quality criteria are intended
15 to "accurately reflect the latest scientific knowledge useful in indicating the kind and extent of
16 identifiable effects on public health or welfare which may be expected from the presence of [a]
17 pollutant in ambient air...."
18 Section 109 (42 U.S.C. 7409) directs the Administrator to propose and promulgate
19 "primary" and "secondary" NAAQS for pollutants identified under section 108. Section
20 109(b)(l) defines a primary standard as one "the attainment and maintenance of which in the
21 judgment of the Administrator, based on such criteria and allowing an adequate margin of safety,
22 are requisite to protect the public health."2 A secondary standard, as defined in Section
23 109(b)(2), must "specify a level of air quality the attainment and maintenance of which, in the
24 judgment of the Administrator, based on such criteria, is requisite to protect the public welfare
25 from any known or anticipated adverse effects associated with the presence of [the] pollutant in
26 the ambient air."3
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, 9l" Cong., 2d Sess. 10 (1970)].
3 Welfare effects as defined in section 302(h) [42 U.S.C. 7602(h)] include, but are not limited to, "effects
on soils, water, crops, vegetation, man-made materials, animals, wildlife, weather, visibility and climate, damage to
and deterioration of property, and hazards to transportation, as well as effects on economic values and on personal
comfort and well-being."
November 2005
1-2
Draft - Do Not Quote or Cite
-------
1 . In setting standards that are "requisite" to protect public health and welfare, as provided
2 in section 109(b), EPA's task is to establish standards that are neither more nor less stringent
3 than necessary for these purposes. In so doing, EPA may not consider the costs of implementing
4 the standards. See generally Whitman v. American Trucking Associations, 531 U.S. 457,464,
5 475-76(2001). . • .
6 The requirement that primary standards include an adequate margin of safety was
7 intended to address uncertainties associated with inconclusive scientific and technical
8 information available at the time of standard setting. It was also intended to provide a reasonable
9 degree of protection against hazards that research has not yet identified. Lead Industries
10 Association v. EPA, 647 F.2d 1130,1154 (D.C. Cir 1980), cert, denied. 101 S. Ct. 621 (1980);
11 American Petroleum Institute v. Costle, 665 F.2d 1176,1186 (D.C. Cir. 1981), cert, denied. 102
12 S.Ct. 1737(1982). Bom kinds of uncertainties are components of the risk associated with
13 pollution at levels below those at which human health effects can be said to occur with
14 reasonable scientific certainty. Thus, in selecting primary standards that include an adequate
15 margin of safety, the Administrator is seeking not only to prevent pollution levels that have been
16 demonstrated to be harmful but also to prevent lower pollutant levels that may pose an
17 unacceptable risk of harm, even if the risk is not precisely identified as to nature or degree.
18 In selecting a margin of safety, the EPA considers such factors as the nature and severity
19 of the health effects, the size of the sensitive population(s) at risk, and the kind and degree of the
20 uncertainties that must be addressed. The selection of any particular approach to providing an
21 adequate margin of safety is a policy choice left specifically to the Administrator's judgment.
22 Lead Industries Association v. EPA, supra. 647 F.2d at 1161 -62.
23 Section 109(d)(l) of the Act requires that "not later than December 31,1980, and at 5-
24 year intervals thereafter, the Administrator shall complete a thorough review of the criteria
25 published under section 108 and the national ambient air quality standards ... and shall make
26 such revisions in such criteria and standards and promulgate such new standards as may be
27 appropriate . ..." Section 109(d)(2) requires that an independent scientific review committee
28 "shall complete a review of the criteria... and the national primary and secondary ambient air
29 quality standards .. . and shall recommend to the Administrator any new'. .. standards and
30 revisions of existing criteria and standards as may be appropriate...." Since the early 1980's,
31 this independent review function has been performed by the Clean Air Scientific Advisory
32 Committee (CAS AC), a standing committee of EPA's Science Advisory Board.
33 1.2.2 History of Ozone NAAQS Reviews
34 Tropospheric (ground-level) Os is formed from biogenic precursor emissions and as a
35 result of anthropogenic precursor emissions. Naturally occurring Os in the troposphere can result
November 2005 1 -3 Draft -Do Not Quote or Cite
-------
1 from biogenic organic precursors reacting with naturally occurring nitrogen oxides (NOX) and by
2 stratospheric Os intrusion into the troposphere. Anthropogenic precursors of O3, specifically
3 NOX and volatile organic compounds (VOC), originate from a wide variety of stationary and
4 mobile sources. Ambient 63 concentrations produced by these emissions are directly affected by
5 temperature, solar radiation, wind speed and other meteorological factors. •
6 The EPA initially established primary and secondary NAAQS for photochemical
7 oxidants on April 30,1971 (36 FR 8186). Both primary and secondary standards were set at an
8 hourly average of 0.08 parts per million (ppm), total photochemical oxidants, not to be exceeded
9 more than one hour per year.
10 On February 8,1979, EPA completed its first periodic review of the criteria and
11 standards for O3 and other photochemical oxidants (44 FR 8202). In that action, EPA made
12 significant revisions to the original standard: the level of the primary and secondary NAAQS
13 was changed to 0.12 ppm; the indicator was changed to O3; and the form of the standards was
14 changed to be based on the expected number of days per calendar year with a maximum hourly
15 average concentration above 0.12 ppm (i.e., attainment of the standard occurs when that number
16 is equal to or less than one).
17 On March 9,1993, EPA concluded its second periodic review of the criteria and
18 standards for O3 by deciding that revisions to the O3 NAAQS were not warranted at that time (58
19 FR 13008). The timing of this decision was required by a court order issued to resolve a lawsuit
20 filed to compel EPA to complete its review of the criteria and standards for O3 in accordance
21 with the Act. This decision reflected EPA's review of relevant scientific information assembled
22 since the last review, as contained in the 1986 03 CD (USEPA, 1986), its Supplement (USEPA,
23 1992) and the 1989 O3 Staff Paper (USEPA, 1989), although it did not take into consideration a
24 large number of studies on the health and welfare effects of O3 published since the literature was
25 last assessed in the O3 Supplement. The final decision emphasized the Administrator's intention
26 to proceed as rapidly as possible with the next periodic review of the air quality criteria and
27 standards to consider the more recent information.
28 Under a court-ordered schedule and a highly accelerated review process, EPA completed
29 its third review of the O3 NAAQS on July 18,1997, based on the 1996 O3 CD (USEPA, 1996a)
30 and 1996 O3 Staff Paper (USEPA, 1996b). EPA revised the primary and secondary O3 standards
31 on the basis of the then latest scientific evidence linking exposures to ambient O3 to adverse
32 health and welfare effects at levels allowed by the 1-hr average standards (62 FR 38856). The
33 O3 standards were revised by replacing the existing primary 1-hr average standard with an 8-hr
34 average O3 standard set at a level of 0.08 ppm. The form of the primary standard was changed to
35 the annual fourth-highest daily maximum 8-hr average concentration, averaged over three years.
November 2005 1 -4 Draft - Do Not Quote or Cite
-------
1 The secondary O3 standard was changed by making it identical in all respects to the revised
2 primary standard.
3 1.2.3 Litigation Related to the 1997 Ozone Standards
4 Following promulgation of the revised O3 NAAQS, petitions for review were filed
5 addressing a broad range of issues. On May 14,1999, in response to those challenges to EPA's
6 1997 decision by industry and others, the U. S. Court of Appeals for the District of Columbia
7 Circuit (D.C. Circuit) remanded the O3 NAAQS to EPA, finding that section 109 of the Act, as
8 interpreted by EPA, effected an unconstitutional delegation of legislative authority.5 In addition,
9 the D.C. Circuit Court directed that, in responding to the remand, EPA should consider the
10 potential beneficial health effects of O3 pollution in shielding the public from the effects of solar
11 ultraviolet (UV) radiation.
12 On January 27, 2000, EPA petitioned the U.S. Supreme Court for certiorari on the
13 constitutional issue (and two other issues) but did not request review of the D.C. Circuit ruling
14 regarding the potential beneficial health effects of 03. On February 27,2001, the U. S. Supreme
15 Court unanimously reversed the judgment of the D.C. Circuit on the constitutional issue, holding
16 that section 109 of the CAA does not delegate legislative power to the EPA in contravention of
17 the Constitution, and remanded the case to the D.C. Circuit Court to consider challenges to the
18 O3 NAAQS that had not been addressed by that Court's earlier decisions.6 On March 26,2002,
19 the D.C. Circuit Court issued its final decision, finding the 1997 O3 NAAQS to be "neither
20 arbitrary nor capricious," and denying the remaining petitions for review.7
21 On November 14,2001, EPA proposed to respond to the D.C. Circuit's remand to
22 consider the potential beneficial health effects of Os pollution in shielding the public from the
23 effects of solar (UV) radiation by leaving the 1997 8-hr NAAQS unchanged (66 FR 52768).
24 Taking into account public comment on the proposed decision, EPA published its final response
25 to this remand on January 6, 2003, reaffirming the 8-hr O3 NAAQS set in 1997 (68 FR 614).
26 Finally, on April 30,2004, EPA announced (69 FR 23966) the decision to make the 1 -hr O3
27 NAAQS no longer applicable to areas one year after the effective date of the designation of those
28 areas for the 8-hr NAAQS. For most areas the date that the 1 -hr NAAQS no longer applied was
29 June 15,2005. (See 40 CFR 50.9 for details.)
5 American Trucking Associations v. EPA,.l15 F.3d 1027 (D.C. Cir., 1999)
6 Whitman v. American Trucking Associations, 531 U.S. 457 (2001)
7 American Trucking Associations v. EPA, 283 F.3d 355, (D.C. Cir. 2002)
November 2005
1-5
Draft - Do Not Quote or Cite
-------
1 1.2.4 Current Ozone NAAQS Review
2 EPA initiated the current NAAQS review in September 2000 with a call for information
3 (65 FR 57810). A project work plan (USEPA, 2002) for the preparation of the CD was released
4 in November 2002 for CASAC and public review. EPA held a series of workshops in mid-2003
5 on several draft chapters of the CD to obtain broad input from the relevant scientific
6 communities. These workshops helped to inform the preparation of the first draft CD (EPA,
7 2005a), which was released for CASAC and public review on January 31,2005.
8 During the process of preparing the first draft CD, NCEA revised the planned format of
9 the CD described in the 2002 work plan. These decisions were made as part of a collaborative
10 effort with OAQPS staff to modify the review process so as to enhance the Agency's ability to
11 meet this and future NAAQS review schedules. As described in Chapter 1 of the first draft CD,
12 emphasis is placed on interpretative evaluation and integration of evidence in the main body of
13 the document, with more detailed descriptions of individual studies being provided in a series of
14 accompanying annexes. This change is intended to streamline the document so as to facilitate
15 timely CASAC and public review and to focus more clearly on issues most relevant to the policy
16 assessment to be developed in the Staff Paper. The modified review process envisions that key
17 policy-relevant issues will be identified earlier in the review process through enhanced
18 collaboration between NCEA and OAQPS staff, leading to a more efficient linkage between the
19 CD and the Staff Paper. At the CASAC meeting held on May 4-5,2005, to review the first draft
20 CD, this new format for the CD was met with general approval of CASAC and the public. A
21 second draft CD (EPA, 2005b) was released for CASAC and public review on August 31, 2005.
22 The schedule for completion of this review is governed by a consent decree resolving a
23 lawsuit filed in March 2003 by a group of plaintiffs representing national environmental
24 organizations. The lawsuit alleged that EPA had failed to perform its mandatory duty, under
25 section 109(d)(l), of completing the current review within the period provided by statute.
26 American Lung Association v. Whitman (No. 1:03CV00778, D.D.C. 2003). An initial consent
27 decree, entered by the court in July 2003, after an opportunity for public comment, was
28 subsequently modified in December 2003 and in April, July, and December 2004. The modified
29 consent decree that now governs this review, entered by the court on December 16,2004,
30 provides that EPA will sign for publication notices of proposed and final rulemaking concerning
31 its review of the O3 NAAQS no later than March 28, 2007 and December 19,2007, respectively.
32 These dates are premised on the expectation that a series of interim milestones will be met,
33 including the release of a final CD by February 28, 2006, and release of a second draft Staff
34 Paper by April 2006, followed by CASAC and public review by July 2006, with completion of a
35 final Staff Paper by September 2006.
November 2005 1 -6 Draft - Do Not'Quote or Cite
-------
1 13 GENERAL APPROACH AND ORGANIZATION OF THE DOCUMENT
2 The policy assessment in this draft Staff Paper is based on staff's evaluation of the policy
3 implications of the scientific evidence contained in the draft CD and initial results of quantitative
4 analyses based on that evidence. Taken together, this information informs preliminary staff
5 conclusions and recommendations on certain elements of the Os standards under review. While
6 the draft CD focuses on new scientific information available since the last review, it
7 appropriately integrates that information with scientific criteria from previous reviews. The
8 quantitative analyses presented in this draft Staff Paper (and described in more detail in technical
9 support documents) are based on the most recently available air quality information, so as to
10 provide current characterizations of Os air quality patterns and estimated health and
11 environmental risks related to exposure to ambient Os concentrations.
12 Following this introductory chapter, this draft Staff Paper is organized into three main
13 parts: the characterization of ambient Oj air quality data; Os-related health effects and primary
14 Oa NAAQS; and Os-related welfare effects and secondary Os NAAQS. The content of these
15 parts is discussed more fully below.
16 The characterization of ambient 63 air quality data is presented in Chapter 2 and includes
17 information on Oi properties, current OB air quality patterns, historic trends, and background
18 levels. This chapter provides a frame of reference for subsequent discussion of current and
19 alternative Oj NAAQS and alternative forms of 63 standards.
20 Chapters 3 through 6 comprise the second main part of this draft Staff Paper dealing with
21 human health and primary standards. Chapter 3 presents an overview of key policy-relevant
22 health effects evidence, major health-related conclusions from the draft CD, and an examination
23 of issues related to the quantitative assessment of evidence from controlled human exposure and
24 epidemiological studies. Chapters 4 and 5 describe the scope and methods used in conducting
25 human exposure and health risk assessments and present initial results from those assessments.
26 Chapter 6 includes a preliminary discussion of the adequacy of the current primary standard and
27 identifies alternative primary standards that staff believes are appropriate to consider in
28 completing the human exposure and health risk assessments.
29 Chapters 7 and 8 will comprise the third main part of this draft Staff Paper. Chapter 7
30 presents a policy-relevant assessment of Os welfare effects evidence and discusses the scope and
31 methods that staff is planning to use in conducting vegetation-related exposure and risk
32 assessments. Chapter 8 will be added in the second draft Staff Paper, and it will include
33 discussions of the adequacy of the current secondary standard and staff recommendations related
34 to the secondary Os standard.
35 Staff conclusions and recommendations, to be presented in the second draft Staff Paper,
36 will be informed by comments received from CASAC and the public in their reviews of this draft
November 2005 1 -7 Draft - Do Not Quote or Cite
-------
1 Staff Paper. The final Staff Paper will be informed by further comments received from CAS AC
2 and the public in their review of the second draft Staff Paper. The final Staff Paper will take into
3 account the scientific evidence reviewed in the final CD and will include: 1) the results of
4 comparative air quality analyses, human exposure and health risk assessments, and vegetation-
5 related environmental assessments; 2) the staffs overall evaluation of the adequacy of the
6 current primary and secondary NAAQS; and 3) staff conclusions and recommendations as to
7 whether any revisions are appropriate to address public health and welfare effects associated
8 with exposure to Os. For these purposes, the staff will assess and integrate new scientific and
9 technical findings with information gained in previous reviews in the context of those critical
10 elements that the staff believes should be considered.
November 2005
1-8
Draft - Do Not Quote or Cite
-------
1 REFERENCES
2
3 Federal Register (1971) National Primary and Secondary Ambient Air Quality Standards for
4 Photochemical Oxidants; Final rule. 40 CFR 50; Federal Register,36: 8186.
5
6 Federal Register (1979) National Primary and Secondary Ambient Air Quality Standards: Revisions to the National
7 Ambient Air Quality Standards for Photochemical Oxidants, Final Rule, 40 CFR 50; Federal Register
8 44:8202.
9 • .
10 Federal Register (1993) National Ambient Air Quality Standards for Ozone, Final rule. 40 CFR 50; Federal Register
11 58:13008.
12 . .
13 Federal Register (1997) National Ambient Air Quality Standards for Ozone; Final Rule. 40 CFR 50; Federal
14 Register 62:38856.
15
16 Federal Register (2001) National Ambient Air Quality Standards for Ozone; Proposed Response to Remand;
17 Proposed Rule. Federal Register 66: 57368.
18
19 Federal Register (2003) National Ambient Air Quality Standards for Ozone; Proposed Response to Remand. Final
20 Rule. Federal Register 68: 614.
21
22 U.S. Environmental Protection Agency. (1986) Air Quality Criteria for Ozone and Other Photochemical Oxidants.
23 Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and
24 Assessment Office; EPA report nos. EPA-600/8-84-020aF-eF. Available from NTIS, Springfield, VA;
25 PB87-142949.
26
27 U.S. Environmental Protection Agency (1992) Summary of selected new information on effects of ozone on health
28 and vegetation: supplement to 1986 air quality criteria for ozone and other photochemical oxidants.
29 Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and
30 Assessment Office; EPA report no. EPA/600/8-88/105F. Available from NTIS, Springfield, VA; PB92-
31 235670.
32
33 U.S. Environmental Protection Agency (1996) Air Quality Criteria for Ozone and Related Photochemical
3 4 Oxidants. Research Triangle Park, NC: Office of Research and Development; report nos. EPA/6001AP-
35 93/004aF-cF. 3v. Available from: NTIS, Springfield, VA; PB96-185582, PB96-185590, and PB96-
36 185608. Available online at: http://cfpub.epa.gQv/ncea/cfnvrecordisplav.cfm7deid~2831.
37
38 U.S. Environmental Protection Agency (2002) Project Work Plan for Revised Air Quality
39 Criteria for Ozone and Related Photochemical Oxidants. Research Triangle Park, NC: National Center for
40 Environmental Assessment-RTP Report no. NCEA-R-1068.
41
42 U.S. Environmental Protection Agency (2005a) Air Quality Criteria for Ozone and Related
43 Photochemical Oxidants (First External Review Draft). Washington, DC, EPA/600/R-05/004aA-cA.
44 Available online at: www.ega.gpy/hceu/
45
46 U.S. Environmental Protection Agency (2005b) Air Quality Criteria for Ozone and Related
47 Photochemical Oxidants (Second External Review Draft). Washington, DC, EPA/600/R-05/004aB-cB.
48 Available online at: hijfe//cfpi!b..e^^
. November 2005 1 -9 Draft - Do Not Quote or Cite
-------
1 2. AIR QUALITY CHARACTERIZATION
2 2.1 INTRODUCTION
3 This Chapter summarizes the origin and status of ozone (Oa) concentrations in the
4 ambient air. Section 2.2 summarizes the physical and chemical properties of 63 including its
5 formation, transport, and fate. The section also summarizes the differences between tropospheric
6 and stratospheric 63 and how O3 and ultraviolet radiation interact. Section 2.3 describes the
7 various sources of ground level Os data and the monitoring methodology that gives rise to these
8 data and issues regarding the monitoring methodology. Section 2.4 describes 63 monitoring
9 methods and issues. Section 2.5 describes the spatial and temporal variation found in ground
10 level Os data Section 2'.6 characterizes Os episodes and Section 2.7 describes background levels
11 of ground level Os-
12 2.2 CHEMICAL AND PHYSICAL PROPERTIES, FORMATION, AND
13 TRANSPORT
14 2.2.1 Chemical and Physical Properties
15 Ozone and other oxidants form in polluted areas mainly by chemical reactions in the
16 atmosphere involving two classes of precursor pollutants, volatile organic compounds or VOCs
17 and nitrogen oxides (NOX). Ozone is, therefore, a secondary pollutant. Carbon monoxide (CO) is
18 also important for ozone formation in urban areas. The formation of Os, other oxidants and
19 oxidation products from these precursors is a complex process involving many factors: the
20 intensity and spectral distribution of sunlight; atmospheric mixing and processing on cloud and
21 aerosol particles; the concentrations of the precursors in ambient air; and the rates of chemical
22 reactions of the precursors. A more detailed discussion of these processes can be found in
23 Chapter 2 of Volume 1 of Air Quality Criteria for Ozone and Related Photochemical Oxidants
24 (draft CD, pp.2-1 - 2-27).
25 The effects of sunlight on Os formation, aside from the role of solar radiation in
26 meteorological processes, depend on its intensity and its spectral distribution. Intensity varies
27 diumally, seasonally, and with latitude, but the effect of latitude is strongest in the winter.
28 Ultraviolet radiation from the sun plays a key role in initiating the photochemical processes
29 leading to O? formation and affects individual photolytic reaction steps. However, there is little
30 empirical evidence in the literature, directly linking day-to-day variations in observed surface
31 UV radiation levels with variations in tropospheric O3levels (draft CD, p. AX2-88).
November 2005
2-1
Draft - Do Not Quote or Cite
-------
1 2.2.2 Formation
2 The chemical formation of Os in the troposphere results from the oxidation of nitric oxide
3 (NO) to nitrogen dioxide (NOa) by organic (RO2) or hydro-peroxy (HO2) radicals. Photolysis
4 (the chemical process of breaking down molecules into smaller units through the absorption of
5 light) of NO2yields nitric oxide (NO) and a ground-state oxygen atom, O(3P), which then reacts
6 with molecular oxygen to form Oa (draft CD, p.2-2).
7 In urban areas, compounds representing all forms of VOCs are important for Os
8 formation. In non-urban, vegetated areas, biogenic VOCs emitted from vegetation tend to be the
9 most important. In the remote troposphere, CH4and CO are the main carbon-containing
10 precursors to Os formation. In coastal environments and other selected environments, atomic Cl
11 and Br radicals can also initiate the oxidation of VOCs (draft CD, p.2-2 and 2-3).
12 Oxidized nitrogen containing compounds are essential to the formation of Os in the air.
13 There are a large number of oxidized nitrogen containing compounds in the atmosphere
14 including NO, NO2, NO3, HNO2, HN03, N2O5, HNO4, PAN and its homologues, other organic
15 nitrates and particulate nitrate. Collectively these species are referred to as NOy. Oxidized
16 nitrogen compounds are emitted to the atmosphere mainly as NO which rapidly interconverts
17 with NO2. Consequently, NO and NO2 are often grouped together into their own family called
18 NOX (draft CD, p. 2-3). NOx is considered a good surrogate for N0y and is monitored more often
19 and its emissions are more widely reported (see Table 2-2).
20 2.2.3 Transport
21 The transport of OB and other secondary pollutants is determined by meteorological and
22 chemical processes extending typically over spatial scales of several hundred kilometers (e.g.,
23 Civerolo et al., 2003; Rao et al., 2003). An analysis of the output of regional model studies
24 conducted by Kasibhatla and Chameides (2000) suggests that Oa can be transported over a few
25 thousand kilometers in the upper boundary layer of the eastern half of the United States during
26 specific 03 episodes. Convection is capable of transporting Oj and its precursors vertically
27 through the troposphere as shown in Annex AX2.3.2 of the CD. Nocturnal low level jets (LLJs)
28 can also transport pollutants hundreds of kilometers. They are observed over the mid-Atlantic
29 region, the central U.S. and California Turbulence associated with LLJs can bring these
30 pollutants to the surface and result in secondary Os maxima in the early morning in many
31 locations. However, the presence of mountain barriers limits mixing as in Los Angeles and
32 Mexico City and will result in a higher frequency and duration of days with high Os
33 concentrations (draft CD, p.2-9).
November 2005 2-2 Draft - Do Not Quote or Cite
-------
1 2.2.4 Precursors, Sources and Emissions
2 Although there are direct sources of ozone (electrical discharges, lightning), ambient CH
3 pollution problems are generally acknowledged to result from the secondary formation of 63 via
4 the processes described in section 2.2.1.
5 Table 2-1 (see http://w7wvv,epaagov/airtrends/econ-einissions.html) lists the main sources
6 of VOC emissions from 1970-2004, The categories in the table are self explanatory with the
7 exception of the fires and miscellaneous categories. The fires category includes both wild fires
8 and prescribed burns. The miscellaneous category includes mainly structural fires and sources
9 from agricultural activities. One category not in either table is biogenic emissions. Biogenic
10 emissions are an important factor on warm to hot days in heavily vegetated areas. As can be
11 seen in the table, highway vehicles have been the single largest source of VOC emissions over
12 the years ranging from about 49% of total emissions in 1970 to about 27% of total emissions in
13 2004. Starting in 2001, solvent use and highway vehicles were the two main sources of VOCs
14 with roughly equal contributions to the total emissions.
15 Table 2-2 contains the same emission information but for NOx emissions. Again,
16 highway vehicles are the single largest source of NOx emissions over the years ranging from
17 about 47% of total emissions in 1970 to about 37% of total emissions in 2004.
18 2.2.5 Tropospheric vs. Stratospheric Ozone
19 The atmosphere can be divided into several distinct vertical layers, based primarily on the
20 major mechanisms by which they are heated and cooled. The lowest major layer is the
21 troposphere, which extends from the earth's surface to about 8 km above the surface in polar
22 regions and to about 16 km above the surface in tropical regions. The planetary boundary layer
23 (PEL) is the lower sub-layer of the troposphere, extending from the surface to about 1 or 2 km,
24 and is most strongly affected by surface conditions. The stratosphere extends from the top of the
25 troposphere, to about 50 km in altitude. The emphasis in this chapter is placed on concentrations
26 of Oa occurring in the troposphere, in particular in the PBL (draft CD, p.2-1).
27 In urban environments the rate of O3 formation is sensitive to the rate of photolysis of
28 several species including H2CO, HjOj, 03, and especially NO2. Monte Carlo calculations suggest
29 that model simulations of photochemical O3 production are most sensitive to uncertainty in the
30 photolysis rate coefficient for NO2 (draft CD, p.AX2-88).
November 2005
2-3
Draft - Do Not Quote or Cite
-------
comcocNT-^cncNCOOu .
O>'J-OOCDOCN-*r'<3-'-O1^-Oen __
o> T-r-r--T-co'3-ocooooi--in
^8
CD T- T- r- CN
o oo CN
co r-- in
i-- ^- CN
CN" CN"
CN CN
i^-^-m'^-CN^'T-coocNfco
oo oo T— CN co T— o oo *~ co in oo
T— OOI*^T—co^oinocof1^^1
in" T-" t- ao CN"
S8
o
co"
CN
o>
m
CO
i
o«
gv~ t"- ^* ^"
•*• -*t Oi <3)
*"" if) CO CO ^
^r *]• tf o CD *-
r-- in oo T- T- o
T- O> O> O) CD T~
rC ,-" *-" co" T-~ W
O)
CN
co o> r-
•f" co"
co co
CO
o
SlOCDOOT-CON*COCNO>COh-lOO
^TO>^'O>CN^rcOCDCN^'fv-^t'CN
o T^-r-co'-cO'^-'«-CDOCNoor-.
'•" co" T-" *-" N." CN"
O) CO' ^~
CO CO CO
w CD o)
CN
CN
CN
§
O)
I
**•*
o
(0
|
o
Q.
E
o
o
ja
c
T- CO
•
Sh- CN CO
^3" OO f^-
<3> T- f-
T™
« CN v ro
CO CO CO O
e» T- T
^ t— '
o in r- oo
^
U) O O O
1*^1 ^f 10 r*^
-
«"
T- O
o in
in"
o o>
en o>
CO CO_
in"
CO 00
CD"
in t-
co in
CN CO
iri"
CN Ol
co a>
in en
O CO
O) OO
^ co
,_7
h- O)
•- ir>
o> r-
^r
T- ^~
00 CO
T- 0)
CN
O
CD
00
00"
00
8
en
s
CO
CN"
O)
CO
oo
CO"
CN
CD
CO
m"
en
o
r-
CN"
CN
CO
CO
CN"
en
CO
^
CN"
?
CN"
i-~
^~
O)
T-"
CD
in
h-
o>
in
o
^mj
CO
IO
CO
^_ n
CO
Is-.
r^
10
co"
CN
OO
o
"fr"
CN
O
^
|S^J"
CN
g
^
CO
in
CO
N.
Q"
00 O)
r- o>
CO 00
CN
co in
OO CN
e» T-
co"
CM
in o>
CD CO
^ O)
co"
CM
•sr CN
ss
T-" O"
CO
N. 00
00 !•-
in T-
o"
CN
<
£
« UJ yj
O LU — O
f. . .
m CD CQ
o o
3 LU
W LL.
O O
o o
LU LU
LL LL
^co
^
Q W
Z LU
— O ._
00 ^
¥£ = 0
< . O CL
o
o »o
O 5
CK ^
a- Lu
UJ
at N]
CO d
Q ^
? 1-
CL Lu
o
y D
<
ui cr w i~ -S
a. O (/) U) >
T.
o
co
LU
"3
O
it o
§
-o
J
-------
•*t CM O O "fr O) LO
^y yj fx.' CO ^~ {£) f^»
O T- r>- CN «
lO h- CO 00 CO
o in ^ co oo
CN •* m co oo
in en
oo in
r- r-
m
Q
k
CO CO
CM
CNOO^-CDinCDCNinh-COOOCO
lOh-O5T-CDN-OO)OlO1*OOCO
•«-h-cN w^i-cD-CN^rincDco
^r LO co
Sco in
N. h-
co in
in co
oo in
m
a
I
l-T-COOCNi-T-OCNCNOCNCNCN »-
ocor>-incoN-^rcNT-o>CNincNco T-
o T-CDCM ^•^•ot-^tcncDin T-
cn in i- ^ CN r>-
o CN co en
co^-T-coinin^co
8CDi^-^-incocNincor>-T-cN'«rco
T-C»CM •^rtrcpT-^-
(M
in CM
CM ^"CMO>'*~COf*"COCO^Is^'~CN^~
^o cninh-T— mcDincococooocoooo>
C.** CD i-CBCM ^•^•OCN^tCDCDI1"-
w •§ *• m ^ « CM
O I*- CO
r-, co o
CM CD CD
Is
00 CO
51 ?
•^-a>^-flomcoo)r-.Lf)O>cOGO
Is* OO O> l*1^ CO ^" ^ -C^ CO lO t** ^"
T'" oo co ^~ ^i ^~ co if) oo co r*™
in T- in CM
co
co
r»- co T-
CO CO
n is
o> "8 h-CNinrococoN-ooT-cocoincNCO
Is H o>Lnr^c»coh-oococNCNT-ooinT-
O^ o> T— coco '^'^•cocoinc3ir».oo
,. •« T- in T- m CM
CO
00
CN
11
SCOcooDcof^ini^-^coT— LOO
r-c»eo!--i*-coi>-cnocNcO''i-
T-OOCO .^•^^•cNincMOO)
T- O 1-
r- r»- o
co co o
o *- en
CN T-
§
^
3
S
£
o\
i
CA
§
w
U
O
fS
a>
1
W'TCDcooincNococMr^-coo'-
o>isrocNcocMTj-inooincD'^-o>in
o>
-------
S
CD •«»• o CD en
ie>_ T-_ n_ i-
co" co T-
oo
co Z
_
of
O) CM h-
CO_ CO O)_
If) •&
CN CM
U
§
en
m m
CD T-
co" co"
CO CM h-
T- CO
O_ CO_ T-
co" t-"
T"L °l N
ro" co"
CO O5
o>" co"
r- ^ co
in co CM
p> CN T-_
LO~ in"
CN CN
CN CN O
10" w"
CN CN
a
I
a
10
•«-o>o>T-incoT-cM
o> in o CM T- T-CO
*~ co" CM" T-"
o>
•^r oo CM Z
oT co"
f^- CM
^r co
CM CM r^
CM CD CO
in co T-
in
CN
in
CM
in
(ft
-
§CNOT-(OOO(NCN
•«-rMr--cN T-CO
^~ to co"
o" co"
r-
m
m"
CM
in
N-
«
CNIO^TT-UJN-O
o> o^ ir^ r- ot • CN
^ h-T ro"
CN Z
^<
°. z
CM
O)
O
r-"
CN
O)CDOr-
o S> ~ z
CN" CN"
e
r- r-
co Z co
CD
CN
co
CM
o> en co oo CN
" "
CM
co co co
CM" CN"
CO .f
00 <
oo z
co"
CN
co
oo
oo
8
o
OT
C
5
<0
<0
1
UJ
1-
d
>.UJ
0 111
f.
S
Source Ca
FUELCOIV
D:
IB. INDUST
IB. OTHER
MOO land
MOO land
RODUCT MFG
Q_
O
UJ
_J
_j
00
I
CHEMICAl
6
ROCESSIN
METALS P
ED INDUSTRIES
ROCESSES
i— n
JM & RELA
DUSTRIAL 1
PETROLEt
OTHER IN
Z
UTILIZATIC
SOLVENT
RT
ECYCLING
O a:
& TRANSP
SPOSAL & 1
STORAGE
WASTE Dl
VEHICLES
WAY
HIGHWAY
OFF-HIGH
CO
O
UJ
MISCELLA
CO
O
UJ
Z
MISCELLA
TOTAL
ut FIRES
FIRES
Total witho
••o
I
I
-------
ooinLn^rcnh-cococMincocDO CM *-
TT r*. CM o oo •<*• oo i- m co oo in -*^rjr
CM o oo •<*• oo
f-- 00 f-
CO CM
CD oo in
t- CO O CO
Is- •
Qomin^-o>N.aocDCMincDcoo
r«- r- CM o oo ^r co r- in CD co in
CM oo i**- •*•" T™^* T- ro o co
^r CN • r-. •«•
CM O)
CM
U
o
&
o
a
-
OCDOD^,-
*"- CO O CO
h- ^1"
CM
CM
«n h- o en CM
CM o T- co r- in
T- in , t- t- «-
CM
T- 00
N. h-
v . •
co CM CD o co
to I**-
in CM
i a ie.
^19 f in CM
-. . _ Is-
T- CO T- CM
oo ^~
OCOOcocMin T-
-
COOCM
oo •*
in
•sr
co
CM
CM
co in
CD m
CM co
CM4"
Oo OOCMO*~O3O)fOr*~COCDCOO>CDO)
v x o>comocMoo- OJCNOT-*- T-^- r-CDT-T-
OI«I>T-CDCOT- 00^-
111
15
o
^^ (O ^~ ^~ t^~ O) co
O) t**» ^3 t*" ^^ QO ^"
O> CM •«-•<- T- T-
«- CD CO T-
00 CN <0
s
m en
gTj-T-r-incocucocMincococflCMO r-mcM
coinrocMoococo T-mpr>-T:- - - -
^- CD fo ^ ^~ ^~ "*
OJCO'-CM'-
*" CD CO" T-~
co o 00
r-- i- oo
s s"
r- »- CD 2 in in o>
oo_ »-_ CM Z CD_ CM co_
f
s
o
(O
o
'»
.2
UJ
x
O
c
CN
n
to
CD t±J
a £
iu
CO
LU
^
Li.
-i-»
O
x:
^cof
2: ui -5
0^1
I- U- H
I
I
-------
1 23 DATA SOURCES
2 Two main sources of data were used for this assessment, the state-supplied data from
3 various types of monitors housed in the Air Quality System (AQS) data base and the Clean Air
4 Status and Trends Network (CASTNET). The vegetation exposure/risk analyses (see Chapter 7)
5 will depend heavily on using the Community Multi-scale Air Quality (CMAQ) modeled data for
6 2001 in conjunction with 2001 monitor data. Both monitor and model values will be combined
7 using the spatial interpolation tool in BenMAP to create an interpolated surface to fill in the gaps
8 left by a sparse rural monitoring network. Air quality models are often used to simulate the
9 formation, transport, and decay of air pollution/ The CMAQ modeling system is a
10 comprehensive three-dimensional grid-based Eulerian air quality model designed to estimate
11 ozone and particulate concentrations and deposition over large spatial scales (Dennis et al, 1996;
12 Byun and Ching, 1999). The CMAQ model is a publicly available, widely-used, peer-reviewed,
13 state-of-the-science model consisting of a number of science attributes that are critical for
14 simulating the oxidant precursors and nonlinear organic and inorganic chemical relationships
15 associated with the formation of ozone, as well as sulfate, nitrate, and organic aerosols.
16 2.3.1 Air Quality System (AQS)
17 The Code of Federal Regulations, Title 40, Part 5 8 (40 CFR Part 58) contains the EPA's
18 ambient air quality surveillance regulations. Section 58.20 requires States to provide for the
19 establishment of air quality surveillance systems in their State Implementation Plans (SIP). The
20 air quality surveillance system consists of a network of monitoring stations designated as State
21 and Local Air Monitoring Stations (SLAMS), which measure ambient concentrations of those
22 pollutants for which standards have been established in 40 CFR Part 50. SLAMS, National Air
23 Monitoring Stations (NAMS), which are a subset of SLAMS, and Photochemical Assessment
24 Monitoring Stations (PAMS) must meet the requirements of 40 CFR Part 5 8, Appendices A
25 (Quality Assurance Requirements), C (Ambient Air Quality Monitoring Methodology), D
26 (Network Design Criteria), and E (Probe and Path Siting Criteria).
27 The Air Quality System (AQS) is EPA's repository of ambient air quality data. AQS
28 stores data from over 10,000 monitors; 5000 of which are currently active. Of these, over 3000
29 measure and report Oa concentration data (See Figure 2-1). These monitors make up the
30 SLAMS, PAMS, NAMS and other special purpose monitors used and operated by the States.
31 AQS also contains meteorological data, descriptive information about each monitoring station
32 (including its geographic location and its operator), and data quality assurance/quality control
November 2005 2-8 Draft - Do Not Quote or Cite
-------
0
o>
I
t
$
'S
o
s
^
2
&
i
O)
I
a
i
-------
1 information. The Office of Air Quality Planning and Standards (OAQPS) and other AQS users
2 rely upon the data system to assess air quality, assist in Attainment/Non-Attainment
3 designations, evaluate State Implementation Plans for Non-Attainment Areas, perform modeling
4 for permit review analysis, and other air quality management functions. AQS information is also
5 used to prepare reports for Congress as mandated by the Clean Air Act (see
6 http;//ww\y,ej^^
7 The NAMS/PAMS/SLAMS ozone monitor network achieved an overall average bias
8 (upper bound) of 0.2% and an overall mean precision of 3% for 2002. If special purpose and
9 other ozone monitors are also included the average upper bounds of bias and precision were
10 0.4% and 2.9% respectively (U.S. EPA 2004a).
11 ' 23.2 CASTNET
12 CASTNET is the nation's primary source for data on dry acidic deposition and rural,
13 ground-level ozone. Operating since 1987, GASTNET is used in conjunction with other national
14 monitoring networks to provide information for evaluating the effectiveness of national emission
15 control strategies. CASTNET consists of over 80 sites across the eastern and western United
16 States (see Figure 2-1) and is cooperatively operated and funded with the National Park Service.
17 In 1986, EPA established the National Dry Deposition Network (NDDN) to obtain field data on
18 rural deposition patterns and trends at different locations throughout the United States. The
19 network consisted of 50 monitoring sites that derived dry deposition based on measured air
20 pollutant concentrations and modeled dry deposition velocities estimated from meteorology, land
21 use, and site characteristic data. In 1990, amendments to the Clean Air Act necessitated a long-
22. term, national program to monitor the status and trends of air pollutant emissions, ambient air
23 quality, and pollutant deposition. In response, EPA, in cooperation with the National Oceanic
24 Atmospheric Administration (NOAA), created CASTNET from NDDN. In terms of data quality,
25 CASTNET achieved 98% to 99% of all precision and accuracy audits being within the ±10%
26 criteria for both precision and accuracy. Overall, CASTNET ozone monitors are stable and show
27 only very small variation (U.S. U.S. EPA 2003, p.22).
28 2.4 OZONE MONITORING METHODS AND ISSUES
29 Ozone monitoring is conducted almost exclusively with UV absorption spectrometry with
30 commercial short path instruments, a method that has been thoroughly evaluated in clean air. The
31 ultimate reference method is a relatively long-path UV absorption instrument maintained under
32 carefully controlled conditions at the National Institute of Standards and Technology (NIST)
33 (draft CD, p.2-21).
November 2005 2-10 Draft - Do Not Quote or Cite
-------
1 Several reports in the reviewed scientific literature have investigated interferences in 03
2 detection via UV radiation absorption. These include the effects of water vapor, VOC's,
3 aromatic compounds and their oxidation products, and other organic and inorganic compounds
4 on instruments based on both UV absorption and chemiluminescence. Water vapor had no
5 significant impact on UV absorption-based instruments, but could cause a positive interference
6 of up to 9% in chemilurninescence-based detectors at high humidities (dew point of 24 C).
7 Aromatic compounds and their oxidation products were found to generate a positive but small
8 interference in the UV absorption instruments. However, when the results are applied to ambient
9 concentrations of toluene and NOx, the effect appears to be very minor (about 3 percent under
10 the study conditions). Other organic and inorganic compounds displayed interferences, but not at
11 levels likely to interfere with accurate determination of 03 in an urban environment The
12 possibility for substantive interferences in O3 detection exists, but such interferences have not
13 been observed even in urban plumes (draft CD, p. 2-24).
14 Ozone is also measured by differential optical absorption spectroscopy (DOAS) at a
15 variety of wavelengths in the UV and visible parts of the spectrum. Comparisons of DOAS
16 results to those from a UV absorption instrument showed good agreement, on the order of 10%.
17 Researchers have reported a positive interference due to an unidentified absorber in the 279 to
18 289 nm spectral region used by many commercial short-path DOAS systems for the
19 measurement of O3. Results of that study suggest that compounds from wood burning, used for
20 domestic heating, may be responsible (draft CD, p.AX2-146).
21 New techniques are being developed, but UV absorption remains the method of choice
22 for ambient 0,monitoring near the Earth's surface. These commercial UV absorption detectors
23 are available at a moderate price. They show good absolute accuracy with only minor cross
24 sensitivity in clean to moderately polluted environments; they are stable, reliable, and sensitive
25 (draftCD,p.AX2-147).
26 2.5 CHARACTERIZATION OF GROUND-LEVEL OZONE CONCENTRATIONS
27 2.5.1 Metrics
28 This section characterizes ground level Os concentrations based on several metrics. Two
29 daily maximum statistics, 1-hr and 8-hr, and two seasonal cumulative statistics, SUM06 and
30 W126 are summarized to show how Os varies over space and time. The daily maximum 8-hr '
31 values are found by first calculating running or moving 8-hr values for all 24 hours in a day (for
32 example averaging the 1-hr concentrations from 1:00am to 8:am, then average the 1-hr values
33 from 2:00am to 9:am, etc.). Then the maximum value for each day is found (note that any 8-hr
34 time period that starts in a day is assigned to that day). On an annual basis, the fourth highest of
November 2005 2-11 Draft - Do Not Quote or Cite
-------
1 these values is summarized. The daily maximum 1-hr statistic is the maximum value of all 1-hr
2 values in a day. On an annual basis, the second highest of these values in a year is summarized.
3 Both die SUM06 and the W126 statistics were calculated using all 1-hr values from 8:00am to
4 8:00pm and finding the largest 3-month sum of these values in an Os monitoring season
5 according to the secondary standard proposed in 1996 (FR Dec 13, 61(241) 1996 p. 65750). The
6 SUM06 cumulative statistic is calculated by summing all 1-hr values that are greater than or
7 equal to 0.06ppm for every day in a month. The W126 seasonal cumulative statistic is calculated
8 similarly to the SUM06 statistic. The only difference is the weighting function. SUM06 has a
9 weighting function that is 0 when the concentration is less than 0.06 and is 1.0 when the
10 concentration is greater than or equal to 0.06. The W126 statistic is a continuous, sigmoidal
11 weighting function with an inflection point between 0.06ppm and 0.07ppm (Lefohn and
12 Runeckles, 1987).
13 2.5.2 Spatial Variability
14 This section characterizes the spatial variability of Os based on all the metrics discussed
15 above. Spatial variability is based on maps displaying county levels of the various metrics. In
16 this way different levels of Oa for different areas of the country are displayed.
17 2.5.2.1 8-Hour and 1-Hour Statistics
18 High 8-hr average 63 concentrations tend to occur near larger urban areas in the
19 same patterns as the 8-hr concentrations. Elevated levels occurring in smaller urban and non-
20 urban areas are most likely caused by transport (see Figure 2-2). These smaller urban and non-
21 urban areas are most obvious at the end of the northeast corridor (the highly urbanized area
22 running from Washington, DC to Boston, MA), North-central New York, and the Northern coast
23 of Lake Michigan. Some of the highest levels occur not in California but in Texas, some
24 counties in the Northeast Corridor, and isolated counties in the East (see Figure 2-2) (Fitz-
25 Simons, et al, 2005). High 1-hr Os concentrations levels occur in the same patterns as the 8-hr
26 concentrations (see Figure 2-3). The highest levels occur in California. (Fitz-Simons, et al,
27 2005).
28 2.5.2.2 Cumulative Seasonal Statistics
29 High SUM06 and Wl 26 levels in 2001 (most of the analyses in Chapter 7 center on 2001
30 data) occurred in most of the agricultural areas of California. When the data were from
31 CASTNET sites more purely rural counties show higher values. The spatial patterns for SUM06
32 and W126 are very similar (See Figure 2-4,2-5,2-6, 2-7). (Fitz-Simons, et al, 2005).
November 2005 2-12 Draft - Do Not Quote or Cite
-------
0
en
D
I
u
o
I
i
i
g,
M
(0
o
o
o
Q.
CM O
— o
_«» o_
O CO
s. i
o £
g "
Q.
O
in i«-
o> o>
at
£
3
O
O
at
CM
o
.- O
o
n
w
•2
w "£
s i
O m
a
I
00
O
V
X
n
: J
! i
Q.
Q.
C
o
1
o
O
CO
u- X X
§O)
. O.
O O
I
-------
S
s_
X
i—<
|
i
I
3
10
i ^
0) O
Q. °.
o •«-
« o
Q. °. Q
o r- o
o
o_
«" uf O>
55 £ S
«M CQ
10 3 S
I
a
a
0)
Q.
O
O
CM
C*»
s
O
V
X
£ 3
T o>
O W
V
X X
11 it
V V
N- O>
O> *-
o *-m
o o
I
a.
a.
at
o
c
o
o
1
-------
1 2.5.3 Temporal Variability
2 Temporal variability consists of several time frames when considering characterization of
3 ground level air quality data. Multi-year trends characterize long term variability or year to year
4 variability. Trends can usually give evidence as to whether or not a situation is getting better or
5 worse over time. For the purposes of displaying long term trends, the data from both AQS and
6 CASTNET are screened for temporally consistent data (only data from sites that meet a data
7 completeness criteria of 12 complete years out of 15 and no gaps of more than 3 consecutive
8 years are included). Seasonal variability characterizes month to month variability to demonstrate
9 when, in the year the highest concentrations occur. Diurnal variability characterizes hour-to-hr
10 changes demonstrating when, in the day, the highest concentrations occur (Fitz-Simons, et al.,
11 2005).
12 2.5.3.1 Long Term Variability - Trends
13 Long term, nationwide trends for 8-hr ozone values are presented in Figures 2-8 and 2-9.
14 Figure 2-8 presents data from sites in the AQS that meet trends criteria and have locations
15 described as Urban and Center City. Figure 2-9 presents data from CASTNET which are rural
16 locations.
17 The trends are similar. The urban trends have more data and more variation. The rural
18 means are slightly lower that the urban means; however the largest urban concentrations are
19 much higher than the largest rural concentrations (Fitz-Simons, etal, 2005).
20 Long term trends for 1 -hr ozone values are presented in Figures 2-10 and 2-11. Figure 2-
21 10 presents data from sites in the AQS that meet trends criteria and have locations described as
22 Urban and Center City. Figure 2-11 presents data from CASTNET which are rural locations.
23 The trends are similar. As in the 8-hr data, urban trends have more data and more variation with
24 the means for the urban trends being higher than the means for the rural trends. This difference
25 is more pronounced than in the 8-hr trends (Fitz-Simons, et al., 2005).
26 The long term trends for both 1-brand 8-hr ozone data are similar. The 8-hr
27 concentrations are lower, but the trends are basically parallel. The highest means occur in
28 1990,1991,1995,1998 and 2002. The highest extreme values are clearly in the 1990s. In many
29 cases, short term variation (3 years or less) is associated with meteorological conditions that are
30 generally more or less conducive to Oa formation in a particular year. One high year between
31 two low years or one low year between two higher years are examples of this short term
32 variation (see Evaluating Ozone Control Programs in the Eastern United States: NOx P-17, U.S.
33 EPA 2005V).
34
November 2005 2-15 Draft - Do Not Quote or Cite
-------
ci
«
Q
•s
o
U
=
i
i
§
CO
-
3
t
.. \
W
vv
UNO
I
3
V
o
Q.
Q.
C
_o
'•&
2
I
0
fc
a
I
I
I
(N
tN
-------
I
0
t
o
U
CO
p
e
'i
•c
M
w
I
I
OJ
I
a s
v \
8S t
iX *
H <°
HCVI
?§
u>
V
to
0>OI
vi ^
SM A-
T- V
go>
-------
i
p
H
a
en
3
8
f
e
o
u
en
I
2
o
fr
s
Q.
O.
O
1
•s
0)
g
o
O
G
fc
I
a
I
00
1
-------
0>
CM
V
to o
CM IO
T- II
§A
n <°
II CM
?§
i
0)
s
£
a
i
$
$
'%
I
o
u>
V
— *O
0> CM
0)
CM
to
CM
Z
|
E
'x
-------
•e
3
§
o
00
Q
Ir
oo
.1
b
%
a
t
*•>
-------
O
A
1
€
T
a
i*;;;::«:!::*:::;«i:;;;e::::^:::::g::::^:::::g::::^:
_i_- - • • • ^k _j_ j.t_ •.'• • ." •'!'• • .-• • • • • • .•• •
19 - •' • fCp • • • • Gj v Cj O O O • * ' 'O "
&MJ)
BUDZQ
§
I
a
I
I
I
I
-------
>—'
S
o
TH
I
i
t
Q
a
IS
2
I
I
a
I
1
-------
1
i
o»
c
s.
o
I
i
a
•M
i.
on
£
%
:O:O::O::O::O::O': !P
h:: HI «S : CC tt •*: CM
i::g::§::q::5:.q::q:
'::P::P. ::0::0. : O::0:
Olldd) UO|IB4Ja3UOO BUDZQ:
O
^
a
§
Oi
s
a
I
•n
I
-------
1 2.5.3.2 Seasonal Variability
2 Monthly statistics are the best method to characterize seasonal variation in Oa
3 concentrations. However in many areas, monitors are not active during cooler months. As a
4 result, data from May through September are the only universally available data for all monitors.
5 Although this is a limited characterization of seasonal variability, it is consistent across the entire
6 national network.
7 Figure 2-12 shows box-plots of all 2004 data from May through September for the
8 second highest daily 1-hr maximums. The center of the distribution shows a slight, steady
9 increase from May to September while the extreme values show a more pronounced but more
10 variable increase for the same period (Fitz-Simons, et al, 2005).
11 Figure 2-13 shows box-plots of all 2004 data from May through September for the fourth
12 highest daily 8-hr maximums. The center of the distribution and the extremes show a slight,
13 steady increase from May to July followed by a slight decrease from July through September
14 (Fitz-Simons, etal., 2005).
15 2.5.3.3 Short Term Variability - Diurnal
16 The daily cycles of human activity and the solar phase drive the hour-to-hour daily cycle
17 seen in ground level Os concentrations. The daily 1 -hr peak levels generally occur in the
18 afternoon with the lowest concentration occurring in the early morning. However, on any given
19 day when conditions are right, this phase can be reversed with the highest values occurring at
20 night or early morning.
21 In order to examine diurnal patterns, box-plots summarize 1-hr values and 8-hr for each
22 hour in the day. The most recently available data, 2004, was used to generate all the box-plots.
23 Figure 2-14 summarizes 1-hr data from AQS that was classified as urban and center city. The
24 pattern is similar for both weekend and week day data. The pattern of the center of the
25 distribution of values shows a smooth sinusoidal portion of the curve from 6:00AM until 8:OOPM
26 and reaches a peak at 1:00 PM to 3:00 PM. Then the pattern alters to a gradual decrease from
27 9:00 PM to 6:00AM (Fitz-Simons, et al., 2005).
28 Figure 2-15 shows the same set of summaries for 8-hr data. 8-hr values run from 0 to 23
29 hours. Hourl is the average of 1-hr values from 1 to 8 while hour 2 is the average of hours 2 to 9
30 and so on. The main difference between the 1-hr data and the 8-hr data is that the 8-hr data
31 exhibit a smoother sinusoidal pattern throughout the day with a peak for the center of the
2-24
-------
fc
o
S
£
8
3
|
a
Ml
£
»s
(uiddj UO|PBMBOUOQ SUOZQ
i
o
I
rs
-c>
I
-------
£
o
S
i
j&
•a
Q
+*
2
•a
a
I
a
-------
u
QZH
I
Q.
y.
•a
H
I
•e
1
o.
I
•a
Q
u
I
CM
CM
CM
CM
0>
T-
00
-------
C/3
f
I
5?
V)
I
£
5?
a
^
I
.
o
S
V)
3.
o
8-
o
*
^3
auozojnoy
CM
o
*•
00
N.
«0
10
M-
N.
O
(O
S
I
i
a
oo
0
3
0
|
^
-------
1 distribution occurring at 10:00 AM or 11:00 AM and a minimum at about 12:00 midnight. The
2 week end pattern is similar to the week day pattern (Fitz-Simons, et al., 2005).
3 Figures 2-16 through 2-19 summarize 1 -hr and 8-hr data from CASTNET sites which are
4 considered rural. Several differences are noted here. The patterns for the center of the
5 distribution is similar to the patterns for the urban sites. The largest values of the 1-hr data
6 exhibit no pattern but the largest values for the 8-hr data have a discemable pattern that differs
7 from the patterns for the values in the center of the distribution. The weekday pattern for the
8 highest values has a smooth sinusoidal pattern but reaches 2 peaks in the day (12:00 midnight
9 and 12:00 noon). The week end pattern also shows a pronounced peak in the afternoon at about
10 1:00 PM which occurs about 2 hours after the peak for the values in the center of the distribution
11 (Fitz-Simons, et al., 2005).
12 2.6 CHARACTERIZATION OF OZONE EPISODES
13 Major episodes of high Oa concentrations in the eastern United States are associated with
14 slow moving, high pressure systems. High pressure systems during the warmer seasons are
15 associated with the sinking of air, resulting in warm, generally cloudless skies, with light winds.
16 These conditions result in the development of stable air masses near the surface which inhibit the
17 vertical mixing of 63 precursors. The combination of inhibited vertical mixing and light winds
18 minimizes the dispersal of pollutants emitted in urban areas, allowing their concentrations to
19 build up. Photochemical activity involving these precursors is also enhanced because of higher
20 temperatures and the availability of sunlight. In the eastern United States, high Os concentrations
21 during an episode can extend over hundreds of thousands of square kilometers for several days.
22 Episodes have two main characteristics, the concentration level reached and the length of
23 time that this level is reached in consecutive days. The following discussion addresses how these
24 characteristics of episodes have varied through both space and time.
25 Numbers of episodes defined by daily maximum 1-hr 63 concentrations reaching a level
26 of 0.12ppm for 1 day generally follow the long term trend of central values (means or medians)
27 of the 1-hr O3 data (See Figures 2-10 and 2-20). As the length of these episodes increase, the
28 frequency of these episodes decreases. In the most recent years (1997-2004) episodes lasting 5
29 days or more often have not occurred at all (Fitz-Simons, et al., 2005).
30 Numbers of episodes defined by daily maximum 8-hr 63 concentrations reaching a level
31 of O.OSppm for 1 day generally follow the long term trend of central values of the 8-hr Os data
32 (See Figures 2-8 and 2-21). As the length of these episodes increase, the frequency of these
33 episodes decreases. However, some of the longer episodes (6 days of more) continue to occur at
34 this level even in the most recent years. In fact the episode must be defined by a level of 0.10
35 ppm before these longer episodes disappear in the most recent years (Fitz-Simons, et al., 2005).
2-29
-------
I
Q.
V
CO
f.
cc
1
I
a.
8?
a
1
^
•a
S
§
£
CM
O
CM
Ot
to
V)
*»
-5
SIS
^ Q
C
cS
a
o
CO
«o
2
1-5
0C/OZO jnoif-
•o
^
1
-------
I
I
i
en
•a
i
i
»•*
o
•O
,
I
I
Si
OS
*-
CO
«•>
»•
£! »•
*• ,3
*-s
-8
-8
6
-§ .
Cil
CN
a
§
CM
o
«o
fe
•o
I
auozo jnoq-f.
-------
31
9
«S
3.
±.
±
A.
M)
I
^
I
in
I
I
_£
-Li
3Z>
i_F
•Q
-r^rr
•Qi
I
L.
o
S
-C^>
HHI>
UJ
CM
N. Uj
Quozojnoif-s
CM
-5 a
s
- O)
. CQ
. K
a
0
5
W
§hl
X
CM
SJ
t
CM v,
** 3 tN
O ^
S C «N
. O
§
§
hS
8
o
§
-------
v
A
I
a.
f
I
-on
^n
-OE
_^__^
-L3I
-On
-n~T
T^~>
-nn-
$»
00
T"
»v
»•
(O
^
10
1
I
I
•a
S
e
=
oe
CZIF
CN
o
(O
10
CM
to
1
3-hour ozone
-------
fi
o
I
T-4
i.
I
I
a
I
o
.2
ex
m
**»
2
£
R
d
d
d I
D -
\
C3
°
d *
\
d
d
-\
d
\
d
\
d
d §
~^,
d
•\
d
d
\
Cl
•^.
C]
d
-^,
d
•"-^
d
\
d
•^s,
d
•^,
d
\
d
•^^
d
•\
d
•-^
d
Cl -
d
~~^
d
d -
^\
d ?
^\
d ?
\.
d °
•^-s
d
-\
d
--^
d ?
*\
d -
d
d
d -
"^-v
d
-^
d
"^
d
--^
d
•^^.
d s
d -
fc
a
I
I
I
-------
$
rf
t,
e
oo
.0
o.
00
q
d
h
•e
o
2
I
fc
I
I
I
-------
1 As episode length and level increase for both 1-hr and 8-hr Os data the frequency
2 decreases (Figure 2-22 and 2-23). The longer periods and higher levels disappear altogether in
3 the period from 2000-2004 (Fitz-Simons, et al., 2005).
4 One final aspect of episodes to examine is the return time or the number of days between
'5 episodes. Looking at the intervals between episodes of O.OSppm for 8-hr data, the most
6 prevalent gap length in days is 1 day. There is a slight peak again at 4 days followed by a
7 gradual decrease in frequency as the gap-length increases (see Figure 2-24). Looking at the same
8 data for episodes of 0.12ppm, it appears that some periodicities appear at 1 day, 5-6 days, 21
9 days, and 33-34 days (see Figure 2-25). The frequencies for these episodes are so small
10 compared to the frequencies of the lower level episodes that these indications should not be
11 considered real or significant indications of periodicities. The 1-hr Oa data exhibit much the
12 same lack of periodicity as the 8-hr data (Fitz-Simons, et al., 2005).
13 2.7 BACKGROUND LEVELS
14 Policy relevant background (PRB) concentrations are those concentrations that would
15 result in the United States in the absence of anthropogenic emissions in North America
16 (including Canada and Mexico) (U.S. EPA, 2005a, AX3-130). Background concentrations
17 include contributions from global natural sources and from anthropogenic sources outside North
18 America. As discussed in Chapter 5 of this Staff Paper, PRB concentrations enter into the
19 assessment of risk to human health.
20 Contributions to background levels of O3 include: photochemical interactions involving
21 natural emissions of VOCs, NOX, and CO; the long-range transport of O3 and its precursors
22 from outside North America; and stratospheric-tropospheric exchange (STE). Processes involved
23 in STE are described in detail in Annex AX2.3 of the CD. Natural sources of O3 precursors
24 include biogenic emissions, wildfires, and lightning. Biogenic emissions from agricultural
25 activities are not considered in the formation of PRB (draft CD, p. AX2-145).
26 As a result of long-range transport from anthropogenic source regions within North
27 America, estimates of PRB O3 concentrations cannot be derived solely from measurements of
28 O3, and must be based on modeling. The global photochemical transport model GEOS-CHEM
29 (Fiore et al., 2003) has been applied to estimate PRB O3 concentrations across the U.S. (U.S.
30 EPA, 2005a, AX3-131). This model shows that PRB O3 concentrations are a function of season,
31 altitude and total surface O3 concentration. PRB 03 concentrations at the surface are generally
32 predicted to be in the range of 0.015 to 0.035 ppm in the afternoon, and they decline under
33 conditions conducive to O3 episodes. They are highest during spring and decline into summer.
34
2-36
-------
a
o
o
e
&
«
o
r>
o
I
•H
I
GO
•3
I
o
e
vt
'5.
w
e
€
B*
J
t
I
-------
1
o
00
.fe
1
00
O>
T9
O
en
'5.
w
Ml
(D
g
ci
r-.
q
o
g
o
I
I
a
I
oo
i
Is
§
-------
AON3n03Hd
-------
-------
1 Higher values tend to occur at higher elevations during spring due to contributions from
2 hemispheric pollution and stratospheric intrusions. The stratospheric contribution to surface O3 is
3 typically well below 0.020 ppm and only rarely elevate O3 concentrations at low-altitude sites
4 and only slightly more often elevate them at high-altitude sites (U.S. EPA, 2005a, AX3-148).
5 Ozone concentrations measured at remote monitoring sites in'the Northern Hemisphere
6 show a broad rarige in mean and median concentrations, as do annual maximum 1-hr
7 concentrations. Generally, concentrations increase with elevation and the highest concentrations
8 are found during spring. The overall average of the annual median O3 concentrations at all remote
9 sites in the continental United States is about 30 ppb and excluding higher elevation sites it is
10 about 24 ppb. Maximum concentrations may be related to stratospheric intrusions, wildfires, and
11 intercontinental or regional transport 'of pollution. However, it should be noted that all of these
12 sites are affected by North American anthropogenic emissions to some extent making an
13 interpretation based on these data alone problematic (draft CD, p.AX2-147).
14 Background estimates of hour Oj concentrations from observations tend to be at the
15 higher end of the range of PRB estimates (25 to 45 ppbv), while the observational results, as well
16 as those from prior modeling studies, indicate that background O3 concentrations in surface air
17 are usually below 40 ppbv. The background O3 concentrations derived from observations may be
18 overestimated if observations at remote and rural sites contain some influence from regional
19 pollution. Natural O3 concentrations are generally in the 10 to 25 ppbv range and never exceed
20 40 ppbv. The range of the hemispheric pollution enhancement (the difference between the
21 background and natural 03 concentrations) is typically 4 to 12 ppbv and only rarely exceeds 20
22 ppbv (< 1% total incidences). The stratospheric contribution is usually below 10 ppbv (draft CD,
23 p.AX2-149).
24 In addition to policy relevant background concentrations, a second component of more
25 rare episodic high-concentration events over shorter periods of time (e.g., days or weeks) both
26 within and outside the U.S., Canada, and Mexico (e.g., volcanic eruptions, large forest fires).
27 Specific natural events such as stratospheric intrusions (STE), wildfires, and volcanic eruptions,
28 both of U.S. and international origin, can lead to very high levels of OB comparable to, or greater
29 man, those driven by man-made emissions in polluted urban atmospheres. Because such
30 excursions can be essentially uncontrollable, EPA has in place policies that can remove
31 consideration of them, where appropriate, from decisions when implementing the NAAQS (U.S.
32 EPA, 1986).
33 Currently, estimates of PRB Os concentrations used in this document are based on
34 predictions obtained by the global scale, three dimensional, chemical transport model GEOS-
35 CHEM (Fiore et al., 2003). Estimates of PRB O3 concentrations cannot be derived solely from
November 2005
2-41
Draft - Do Not Quote or Cite
-------
1 measurements of Oa at relatively unpolluted sites because of long-range transport from
2 anthropogenic source regions within North America. PRB Oa concentrations are a function of
3 season, altitude and total surface 63 concentration. PRB 03 concentrations at the surface are
4 generally estimated to be in the range of 0.015 to 0.035 ppm from 1300 to 1700 local time, and
5 they decline under conditions conducive to Os episodes. They are highest during spring and
6 decline into summer. Higher values tend to occur at higher elevations during spring due to
7 contributions from hemispheric pollution and stratospheric intrusions. The stratospheric
8 contribution to surface Oj is typically well below 0.020 ppm. The maximum probability of
9 stratospheric intrusions reaching about 1800 m altitude was less than 1 %. Stratospheric
10 intrusions occur with higher probabilities of 1 to 2 % at about 4100 m, and 10 % at about 5400
11 m. Thus, stratospheric intrusions only rarely elevate 03 concentrations at low-altitude sites and
12 only slightly more often elevate them at high-altitude sites.
13 The analysis used in Chapter 4 and 5 uses the estimates of PRB from the GEOS-CHEM
14 model. The data consist of gridded values with latitude running from 12° to 80 ° in 2 ° steps and
15 longitude running from -177.5 ° to -47.5° in 2.5 ° steps. These data are in hourly values for the
16 2001 warm season. The model estimated the PRB and total ozone concentrations at each grid
17 point.
November 2005 2-42 Draft - Do Not Quote or Cite
-------
1 REFERENCES
2
3 Byun, D. W., and Ching, J.K.S., Eds, 1999. Science algorithms of EPA Models-3 Community Multiscale Air Quality
4 (CMAQ modeling system, EPA/600/R-99/030, Office of Research and Development, U.S. Environmental
5 Protection Agency.
6
7 Civerolo, K. L.; Mao, H. T.; Rao, S. T. (2003) The airshed for ozone and fine particulate pollution in the eastern
8 United States; Pure Appl. Geophys. 160: 81-105.
9
10 Dennis, R.L., Byun, D.W., Novak, J.H., Galluppi, K.J., Coats, C.J., and Vouk, M.A., 1996. The next generation of
11 integrated air quality modeling: EPA's Models-3, Atmospheric Environment, 30,1925-1938.
12
13 Fiore, A.; Jacob, D.J.;Liu, H.; Yantosca, R.M.; Fairlie, T.D.; Li, Q. (2003). Variability in Surface Ozone
14 Background over the United States: Implications for Air Quality Policy. J. of Geophysical Research,
15 108(024)19-1-19-12.
16
17 Fits-Simons, T.; McCluney, L.; Rizzo, M.(2005) U.S. EPA Memorandum to File. Subject: Analysis of 2004 Ozone
18 Data for the Ozone NAAQS Review, November 7.
19
20 Kasibhatla, P.; Chameides, W. L. (2000) Seasonal modeling of regional ozone pollution in the eastern United States.
21 Geophys. Res. Lett. 27:1415-1418.
22
23 Lefohn A.S. and Runeckles'V.C.-, Establishing Standards to Protect Vegetation - Ozone Exposure/Dose
24 Considerations. Atmospheric Environment 21:561-568,1987.
25
26 Rao, S. T.; Ku, J.-Y.; Berman, S.; Zhang, K.; Mao, H. (2003) Summertime characteristics of the atmospheric
27 boundary layer and relationships to ozone levels over the eastern United States. Pure Appl. Geophys. 160:21-
28 55.
29 U.S. Environmental Protection Agency (1986). Guideline on the Identification and Use of Air Quality Data Affected
30 by Exceptional Events. EPA-450/4-86-007.
31
32 U.S. Environmental Protection Agency (2003).Clean Air Status and Trends Network (CASTNet) 2001 Quality
33 Assurance Report; Research Triangle Park, NC: Office of Air Quality Planning and Standards. Report from
34 EPA Contract No. 68-D-98-112.
35
36 U.S. Environmental Protection Agency (2004a). 2003 Criteria Pollutant Quality Indicator Summary Report for July
37 14,2004, AQS Data; Research Triangle Park, NC: Office of Air Quality Planning and Standards. Report from
3 8 EPA Contract No. 68-D-02-061.
39
40 U.S. Environmental Protection Agency (2004b).The Ozone Report: Measuring Progress through Research Triangle
41 Park, NC: Office of Air Quality Planning and Standards; Report no.EPA-454-K-04-001.
42
43 U.S. Environmental Protection Agency (2005a).Air Quality Criteria for Ozone and Related Photochemical Oxidants.
44 Research Triangle Park, NC: Office of Research and Development; Report no. EPA/600/R-05/0054aB.
45
46 U.S. Environmental Protection Agency (2005b).Evaluating Ozone Control Programs in the Eastern United States:
47 NOX Budget Trading Program Progress and Compliance. Research Triangle Park, NC: Office of Air Quality
48 Planning and Standards; Report no. EPA-454- K-05-001.
49
50
November 2005 2-43 Draft - Do Not Quote or Cite
-------
1 3. POLICY-RELEVANT ASSESSMENT OF HEALTH
2 EFFECTS EVIDENCE
3 3.1 INTRODUCTION
4 This chapter assesses key policy-relevant information on the known and potential health
5 effects associated with exposure to ambient Os, alone and in combination with other pollutants
6 lhat are routinely present in ambient air. This assessment focuses specifically on the health
7 effects evidence contained in Chapters 4 through 7 of the draft CD, with particular emphasis on
8 the integrative synthesis presented in Chapter 8. That integrative synthesis focuses on
9 integrating newly available scientific information with that available from the last review, as well
10 as integrating information from various disciplines, to address a set of issues central to the
11 assessment of scientific information upon which this review of Os NAAQS is to be based. It is
12 intended to provide a coherent framework for assessment of human health effects posed by
13 ambient Os in the U.S. and to facilitate consideration of the key policy-related issues to be
14 addressed in this draft Staff Paper.
15 Thisrchapter summarizes that evidence as a basis for future development of staff
16 conclusions and recommendations related to primary standards for O3. These staff conclusions
17 and recommendations will be discussed in Chapter 6 of the second draft Staff Paper.
18 Furthermore, this assessment addresses key issues relevant to quantitative assessment of
19 controlled human exposure and epidemiological evidence so as to provide a foundation for '
20 quantitative health risk assessment, as discussed in Chapter 5 of this draft Staff Paper.
21 The decision in the last Os primary NAAQS review was focused substantially on short-
22 term and prolonged controlled-exposure studies reporting lung function decrements, respiratory
23 symptoms, and respiratory inflammation in humans. The current review of the draft CD has
24 emphasized a large number of epidemiological studies with these and additional health
25 endpoints, including acute and long-term effects of Os on premature mortality, enhanced
26 respiratory symptoms and lung function decrements in asthmatic individuals, school absences,
27 emergency room visits and hospital admissions for respiratory causes, published since the last
28 review, with some important new information also coming from toxicology, dosimetry, and
29 controlled human exposure studies. As discussed in more detail in section 3.3 of this draft Staff
30 Paper, highlights of the new evidence include:
November 2005 3-1 Draft - Do Not Cite or Quote
-------
1 • Newly available, large multicity studies, designed specifically to examine the acute
2 effects of PM and O3 on mortality, provide much more robust and credible information
3 than was previously available in the last NAAQS review. The results from two key
4 .studies carried out in 95 U.S. communities (U.S. National Morbidity, Mortality Air
5 Pollution Study [NMMAPS]) and in 23 European cities (Air Pollution and Health:
6 European Approach [APHEA]) showed positive and significant O3 effect estimates for
7 all cause (nonaccidental) mortality.
8 • Numerous acute exposure epidemiological studies published during the past decade
9 offer added evidence of ambient Correlated lung function decrements and respiratory
10 symptoms in exercising healthy subjects and asthmatic subjects, as well as evidence on
11 new health endpoints, such as the relationships between ambient O3 concentrations and
12 school absenteeism and between ambient O3 and cardiac physiologic endpoints.
13 • Several new studies have been published over the last decade examining the temporal
14 associations between O3 exposures and emergency department visits for respiratory
15 diseases and a significant O3 effect on respiratory hospital admissions.
16 • New controlled human-exposure studies offer evidence of increased airway
17 responsiveness to allergens in subjects with allergic asthma exposed to O3 and of
18 increased airway allergen responsiveness in subjects with allergic rhinitis following
19 exposure to 03.
20 • Numerous controlled human-exposure studies have reported indicators of O3-induced
21 inflammatory response in both the upper respiratory tract (URT) and lower respiratory
22 tract (LRT), while other studies have shown significant changes in host defense
23 capability following O3 exposure of healthy young adults.
24 • Animal toxicology studies provide new information regarding mechanisms of action,
25 dosimetry, increased susceptibility to respiratory infection, and the biological
26 plausibility of acute effects and chronic, irreversible respiratory damage.
27 The scientific evidence and conclusions presented in this chapter are based upon
28 information contained in the draft CD's evaluation of health evidence from various disciplines.
29 Section 3.2 is an overview of potential mechanisms by which exposure to O3 may result in health
30 effects, discussed in Chapters 5, 6, and 7 of the draft CD. Section 3,3 summarizes the nature of
31 effects induced by O3 exposure, and draws on information contained in Chapters 5, 6 and 7 of the
32 draft CD. Section 3.4 presents several issues important to staffs interpretation and quantitative
33 assessment of epidemiological studies. Section 3.5 discusses biological plausibility and
34 coherence of evidence for 03-induced adverse health effects, including short-term respiratory
35 effects, short-term cardiovascular effects, long-term health effects, and mortality-related health
36 endpoints. Also drawing from the draft CD's integrative synthesis, section 3.6 discusses factors
37 that modify responsiveness to O3; potentially susceptible and vulnerable populations groups;
38 public health impacts of exposure to ambient O3; and what constitutes an adverse health impact
39 from ambient O3 exposure. Finally, in section 3.7, staff builds upon the draft CD's detailed
November 2005
3-2
Draft - Do Not Cite or Quote
-------
1 evaluation and integration of the scientific evidence available on these issues to draw
2 conclusions regarding the use of health study results in quantitative evaluation and risk
3 assessments that will be relied upon in developing future staff recommendations on potential
4 revisions to the primary O3 NAAQS to be presented in Chapter 6 of the second draft of the Staff
5 Paper.
6 3.2 MECHANISMS
7 This section provides an overview of evidence covered in the draft CD (chapters 5 and 6)
8 on possible mechanisms by which exposure to 03 may result in acute and chronic health effects.
9 Evidence from dosimetry, toxicology, and human exposure studies has contributed to an
10 understanding of the mechanisms which help to explain the biological plausibility and coherence
11 of evidence for O3-induced respiratory health effects reported in epidemiological studies. In the
12 past, however, little information was available to help explain potential biological mechanisms
13 which linked O3 exposure to premature mortality or cardiovascular effects.
14 Scientific evidence discussed in the draft CD indicates that reactions with lipids and
15 antioxidants are the initial step in mediating deleterious health effects of O3. Subsequent
16 activation of a cascade of events starting with inflammation, altered permeability of the epithelial
17 barrier, impaired clearance mechanisms (including host defense), and pulmonary structural
18 alterations that potentially exacerbate a preexisting disease status (CD, sec. 8.4.1). According to
19 the draft CD, the scientific evidence is still lacking for clearly establishing a role for one or a
20 group of mechanistic pathways underlying O3 health effects observed in epidemiological studies.
21 Most of these mechanisms of action were based on animal toxicology studies with some support
22 from human exposure studies.
23 3.2.1 Direct Pulmonary Effects
24 Potential direct pulmonary effects of O3 include changes in breathing pattern, symptoms
25 of breathing discomfort, lung function changes, and airway hyperreactivity. Subjects who
26 engage in physical activity for multiple hrs while exposed to O3 may experience subjective
27 respiratory tract symptoms and acute physiological changes. Airway irritation is consistently the
28 most typical symptomatic response reported in studies and is accompanied by several
29 physiological changes, depending on individual responsiveness to O3. These physiological
30 changes include bronchoconstriction, airway hyperresponsiveness, airway inflammation,
31 immune system activation, and epithelial injury. Severity of symptoms and magnitude of
32 response depend on dose of inhaled 03, individual sensitivity to O3, and extent of tolerance
33 resulting from previous 03 exposures. Development of effects is time-dependent with a
34 substantial degree of overlap of increasing and receding effects. Time sequences, magnitudes,
35 and types of responses of this series of events, in terms of development and recovery, indicate
November 2005 3-3 Draft - Do Not Cite or Quote
-------
1 that several mechanisms, activated at different times, contribute to the overall lung function
2 response. (CD, pp. 6-8 to 6-9) For the full discussion of the mechanisms of pulmonary function
3 responses, see section 6.2.5 of the draft CD.
4 32.1.1 Breathing Pattern Changes
5 Human controlled-exposure studies have consistently found that inhalation of Oj alters
6 the breathing pattern without significantly affecting minute ventilation (CD, pp. 6-10 to 6-11). A
7 progressive decrease in tidal volume and an increase in frequency of breathing to maintain steady
8 ventilation during exposure of human subjects suggest a direct impact on ventilation. These
9 changes are similar to responses in many animal species exposed to O3 and other respiratory
10 irritants. Bronchial C-fibers and rapidly adapting receptors appear to be the primary modulators
11 of Oj-induced changes in ventilation rate and O3 penetration in both humans and animals (CD,
12 section 6.2.5.1).
13 3.2.1.2 Symptoms and Lung Function Changes
14 In addition to changes in ventilatory control, 03 inhalation by humans induces a variety
15 of symptoms, reduces vital capacity (VC) and related functional measures, and increases airway
16 resistance (CD, pp. 6-11 to 6-12). The reduction in VC caused by exposure to O3 is a reflex
17 action and not a voluntary cessation of breathing resulting from discomfort. While O3-induced
18 symptom responses (mediated in part by bronchial C-fibers) are substantially reduced by inhaled
19 topical anesthetic, the anesthetic had a minor and irregular effect on pulmonary function
20 decrements and rapid, shallow breathing. Since respiratory symptom responses were largely
21 abolished, these findings support reflex inhibition of VC due to stimulation of both bronchial and
t
22 pulmonary C-fibers. Intersubject variability in FEVi responses is not explained by differences in
23 O3doses between similarly exposed individuals (CD, section 6.2.5.1).
24 3.2.1.3 Airway Hyperresponsiveness
25 Bronchial or airway hyperresponsiveness (AHR) refers to a condition in which the
26 propensity for the airways to constrict due to specific (e.g., allergens and antigens) or nonspecific
27 (e.g., histamine and cold air) stimuli becomes increased (CD, section 6.8). Despite a common
28 mechanism (CD, p. 5-49, pp.6-12 to 6-13), post-O3 exposure pulmonary function changes and
29 AHR (either early or late phase) are poorly correlated. Neither does post-O3 exposure AHR
30 seem to be related to baseline airway reactivity. These findings imply that the mechanisms are
31 either not related or are activated independent in time. Indeed, O3-induced increases in AHR
32 appear to persist longer than pulmonary function and symptom responses to O3. Animal studies
33 (with limited support from human studies) have suggested that that stimulation of C-fibers can
34 lead to increased responsiveness of bronchial smooth muscle independently of systemic and
November 2005 3-4 Draft - Do Not Cite or Quote
-------
1 inflammatory changes which may be absent. Characteristic O3-induced inflammatory, airway
2 neutrophilia, which at one time was considered a leading AHR mechanism, appears to be only
3 temporally correlated with AHR. This observation does not rule out involvement of other cells in
4 AHR modulation. However, there is some evidence that release of inflammatory mediators
5 might sustain AHR and bronchoconstriction. Late AHR observed in some studies is plausibly
6 due to sustained damage of the airway epithelium and continual release of inflammatory
7 mediators. Thus, O3-induced AHR appears to be a product of many mechanisms acting at
8 different time periods.
9 3.2.2 Extrapulmonary Effects
10 Ozone reacts rapidly on contact with lipids and antioxidants in the epithelial lining fluid
11 (ELF) and is not absorbed or transported to extrapulmonary sites (CD, p. 6-40, p. AX6-126).
12 Laboratory animal studies suggest that reaction products formed by the interaction of O3 with
13 respiratory system fluids or tissues may produce effects measured outside the respiratory tract—
14 either in the blood, as changes in circulating blood or as changes in the structure or function of
15 other organs, such as the parathyroid gland, the heart, the liver and the central nervous system.
16 However, very little is known about the mechanisms by which O3 could cause extrapulmonary
17 effects (CD, sections AX6.10 and 5.4). Human exposure studies discussed in previous criteria
18 documents (U.S. EPA, 1986,1996) failed to demonstrate any consistent extrapulmonary effects.
19 More recent human exposure studies have identified specific markers of exposure to O3 in the
20 blood, such as a reduction in the serum levels of a free radical scavenger after O3 exposure (CD,
21 sections 6.10 and AX6.10).
22 Alterations in heart rate and/or rhythm are thought to reflect pathophysiologic changes
23 which may represent mechanisms by which pollutants like O3 may induce acute adverse
24 cardiovascular health effects. Decreased heart rate variability (HRV) has been suggested as a
25 predictor of increased cardiovascular morbidity and mortality (CD, p. 7-55). Heart rate
26 variability, resting heart rate, and blood pressure are modulated by a balance between the
27 sympathetic arid parasympathetic nervous systems, and decreased HRV predicts an increased
28 risk of cardiovascular morbidity and mortality in the elderly and those with significant heart
29 disease. Decreased parasympathetic input to the heart may provide an important mechanistic
30 link between air pollution and cardiovascular mortality by promoting fatal arrhythmias (CD,
31 section 7.2.7.1). The impact of O3 exposure on the cardiovascular system is discussed in section
32 3.3.1.4 of this draft Staff Paper.
November 2005 3 -5 Draft - Do Not Cite or Quote
-------
1 3.3 NATURE OF EFFECTS
2 When the primary O3 NAAQS was last reviewed, the decision was based largely upon
3 evidence of pulmonary function decrements, respiratory symptoms, and indicators of
4 inflammation collected in experimental controlled exposure studies of human subjects. Less
5 consistent, but supportive evidence was provided by epidemiological studies that reported
6 associations between ambient O3 levels and hospital admissions or emergency room (ER) visits
7 for respiratory causes, as well as school absences, work loss days, and restricted activity days. In
8 addition, lexicological studies demonstrated morphological changes and altered host defense
9 mechanisms with O3 exposure (U.S. EPA, 1996)
10 The current draft CD provides some new evidence of lung function, symptom, and
11 inflammatory effects in controlled exposure studies that supports the findings of the previous
12 CD. However, the most significant body of new studies in this review is the recent
i.
13 epidemiological evidence of associations between short-term exposure to O3 and effects such as
14 premature mortality, hospital admissions and ER visits for respiratory (e.g., asthma) and
15 cardiovascular causes.
16 The following discussions of O3-induced health effects are based on scientific evidence
17 critically reviewed in chapters 5, 6, and 7 of the draft CD, as well as the draft CD's integration of
18 scientific evidence contained in Chapter 8. hi addition, these health effects discussions rely on
19 the more detailed information and tables presented in the draft CD's annexes AX5, AX6, and
20 AX7. Conclusions drawn about O3-induced health effects depend on the full body of evidence
21 from controlled-exposure human, epidemiological and toxicological data contained in the draft
22 CD. Staff first focuses on the broad array of morbidity effects that have been associated with O3
23 exposure, including both acute and chronic exposures, in section 3.3.1. Section 3.3.2 then
24 discusses the expanded body of evidence on associations between acute O3 exposure and
25 mortality, as well as the more limited evidence on chronic O3 exposures and mortality.
26 3.3.1 Morbidity
27 This section briefly summarizes evidence from toxicological, controlled human exposure
28 and epidemiological studies on respiratory and cardiovascular effects associated with exposure to
29 O3, based on the draft CD's assessment in Chapters 4, 6 and 7. Evidence of O3-related hospital
30 admissions and emergency room visits is discussed in section 3.3.2.1, followed by discussion of
31 the effects of short-term and long-term exposure to O3 on the respiratory system in sections
32 3.3.2.2 and 3.3.2.3, and Cvrelated cardiovascular effects in section 3.3.2.4.
November 2005 3-6 Draft - Do Not Cite or Quote
-------
1 33.1.1 Emergency Department Visits/Hospital Admissions for Respiratory
2 Causes
3 In the last review of the O3 NAAQS, the 1996 CD evaluated emergency room visits and
4 hospital admissions as possible outcomes following exposure to O3 (CD, section 7.3.1). The
5 evidence was limited for emergency department visits, but results of several studies generally
6 indicated that short-term exposures to O3were associated with respiratory emergency department
7 visits (Bates et al., 1990; Cody et al., 1992; Weisel et al., 1995; White et al., 1994). The
8 strongest and most consistent evidence, both below and above 0.12 ppm 1-hr max O3, was found
9 in the group of studies which investigated summertime daily hospital admissions for respiratory
10 causes in different eastern North American cities (Bates and Sizto, 1983,1987,1989; Burnett et
11 al., 1994; Lipfert and Hammerstrom, 1992; Thurston et al., 1992,1994). These studies were
12 consistent in demonstrating that ambient O3 levels were associated with increased hospital
13 admissions and accounted for about one to three excess respiratory hospital admissions per
14 million persons with each 100 ppb increase in 1 -hr max O3, with adjustment for possible
15 confounding effects of temperature and copollutants. Overall, the 1996 CD concluded that there
16 was strong evidence that ambient O3 exposures were associated with exacerbation of respiratory
17 disease (CD, p. 7-57,7-58).
18 In the past decade, a number of studies have examined the temporal associations between
19 O3 exposures and emergency department visits for respiratory causes (CD, section 7.3.2). These
20 studies are summarized in the draft CD (Table AX7-3, Chapter 7 Annex). Respiratory causes for
21 emergency room visits include asthma, bronchitis, emphysema, pneumonia, and other upper and
22 lower respiratory infections, such as influenza, but asthma visits typically dominate the daily
23 incidence counts. Among studies with adequate controls for seasonal patterns, many reported at
24 least one significant positive association involving O3. These studies examined emergency
25 department visits for total respiratory complaints (Delfmo et al., 1997b, 1998b; Hemandez-
26 Garduno et al., 1997; Ilabaca et al., 1999; Jones et al., 1995; Lin et al., 1999), asthma (Friedman
27 et al., 2001; Jaffe et al., 2003; Stieb et al., 1996; Tenias et al., 1998; Tobias et al., 1999; Tolbert
28 et al., 2000; Weisel et al., 2002), and COPD (Tenias et al., 2002).
29 Figure 7-8 (CD, p. 7-59) provides effect estimates for associations between emergency
30 department visits for asthma and short-term O3 exposures". In general, O3 effect estimates from
31 summer only analyses tended to be positive and larger compared to results from cool season or
32 all year analyses (CD, p. 7-60). Several of the studies reported significant associations between
33 O3 concentrations and emergency department visits for respiratory causes. However,
34 inconsistencies were observed which were at least partially attributable to differences in model
35 specifications and analysis approach among various studies. For example, ambient O3
36 concentrations, length of the study period, and statistical methods used to control confounding by
November 2005 3-7 Draft - Do Not Cite or Quote
-------
1 seasonal patterns and copollutants appear to affect the observed 03 effect on emergency
2 department visits. This has led the draft CD (p. 7-62) to conclude that Hie body of evidence
3 remains inconclusive regarding effects of O3 on the risk of emergency department visits.
4 Unscheduled hospital admissions occur in response to unanticipated disease
5 exacerbations and are more likely to be affected by environmental factors, such as high 03 levels.
6 Thus, hospital admissions studies focus specifically on unscheduled admissions. Results of a
7 fairly large number of these studies published during the past decade are summarized in Table -
8 AX7-4 (CD, Chapter 7 Annex). As a-group, these hospital admissions studies tend to be larger
9 geographically and temporally than the emergency department visit studies and provide results
10 that are generally more consistent. The largest and most significant associations of respiratory
11 hospital admissions with O3 concentrations were observed using short lag periods, in particular
12 for a 0-day lag (same day exposure) and a 1-day lag (previous day exposure). In a 16-city
13 Canadian study, Burnett et al. (1997) found the strongest association at a 1 -day lag and a decline
14 in magnitude and significance of effect estimates with longer lag periods for O3 exposure. In
15 five European cities, Anderson et al., (1997) investigated the association between O3 and
16 unscheduled daily hospital admissions for COPD and found that the largest risk estimates were
17 for 0- and 1-day lags: Also, among all pollutants examined, the most consistent and significant
18 findings were for O3. According to the draft CD (p. 7-67), other studies conducted in one or two
19 cities over a five year period or longer provided substantial additional evidence regarding O3
20 effects on respiratory hospital admissions (Anderson et al., 1998; Burnett et al., 1999, 2001;
21 Moolgavkar et al., 1997;.Petroeschevsky et al., 2001; Ponce de Leon et al., 1996; Sheppard et al.,
22 1999 [reanalysis Sheppard, 2003]; Yang et al., 2003). The draft CD observes that in some areas
23 with low ambient O3 concentrations, authors reported that significant associations between O3
24 exposure and respiratory hospitalization were not found but did not provide quantitative results
25 (e.g., Lin et al., 2004).
26 Overall, the draft CD concludes that positive and robust associations were found between
27 ambient 63 concentrations and hospital admissions, when focusing particularly on results of
28 warm-season analyses. Recent studies have also reported associations with ER visits for
29 respiratory diseases, though the draft CD finds this evidence to be less consistent (CD, p. 7-177).
30 33.1.2 Effects on the Respiratory System from Short-term Exposures
31 Short-term exposures to O3 have been reported to induce a wide variety of respiratory
32 health effects. These effects include a range of more subtle effects such as including
33 morphological changes in the respiratory tract, pulmonary function decrements, respiratory
34 inflammation, increased airway responsiveness, morphological effects, changes in host defense
November 2005 3-8 Draft - Do Not Cite or Quote
-------
1 capability, and effects on exercise performance. Short-term O3 exposure has also been
2 associated with increases in respiratory symptoms, restricted activity days, and school absences.
3 3.3.1.2.1 Pulmonary Function Decrements, Respiratory Symptoms, and Asthma
4 Medication Use
5 A very large literature base of studies published prior to 1996, which investigated the
6 health effects on the respiratory system from short-term 03 exposures, was reviewed in the 1986
7 and 1996 CDs (U.S. Environmental Protection Agency, 1986,1996). In the last review, the
8 lowest 03 concentration at which statistically significant reductions in forced vital capacity
9 (FVC) and forced expiratory volume in 1 second (FEV() had been reported in sedentary subjects
10 was 0.5 ppm (CD, p 6-3). During exercise, spirometric and symptomatic responses were
11 observed at much lower O3 exposures. When minute ventilation was considerably increased by
12 continuous exercise (CE) during O3 exposures lasting 2 hr or less at > 0.12 ppm, healthy subjects
13 generally experienced decreases in FEVl3 FVC, total lung capacity (TLC), inspiratory capacity
14 (1C), mean forced expiratory flow from 25% to 75% of FVC (FEF25_75), and tidal volume (VT);
15 increases in specific airway resistance (sRaw), breathing frequency (fB), and airway
16 responsiveness; and symptoms such as cough, pain on deep inspiration, shortness of breath,
17 throat irritation, and wheezing. When exposures were increased to 4- to 8-hr in duration,
18 statistically significant spirometric and symptom responses were reported at lower O3,
19 concentrations, as low as 0.08 ppm, and at lower minute ventilation (i.e., moderate exercise) than
20 the shorter duration studies (CD. p. 6-5).
21 The most important observations drawn from studies reviewed in the 1996 CD were that:
22 (1) young healthy adults exposed to O3 concentrations > 0.08 ppm develop significant,
23 reversible, transient decrements in pulmonary function if VE or duration of exposure is increased
24 sufficiently, (2) children experience similar spirometric responses but lesser symptoms from O3
25 exposure relative to young adults, (3) 03-induced spirometric responses are decreased in the
26 elderly relative to young adults, (4) there is a large degree of intersubjecf variability in
27 physiologic and symptomatic responses to O3 but responses tend to be reproducible within a
28 given individual over a period of several months, and (5) subjects exposed repeatedly to O3 for
29 several days develop a tolerance to successive exposures, as demonstrated by an attenuation of
30 responses, which is lost after about a week without exposure (CD, p. 6-1).
31 Since 1996, there have been a number of studies published investigating spirometric and
32 symptomatic responses, and they generally support the observations previously drawn. Recent
33 studies for acute exposures of 1 to 2 hrs in duration are summarized in Table AX6-1 of the draft
34 CD (p. AX6-5 to AX 6-7). Among the more important of the recent studies was McDonnell et
35 al. (1997), which examined reported changes in FEVt in 485 white males (ages 18-36) exposed
36 for 2 hrs to O3 concentrations from as low as 0.08 ppm up to 0.40 ppm, at rest or with IE.
November 2005 3-9 Draft - Do Not Cite or Quote
-------
1 Decrements in FEVi were modeled by sigmoid-shaped curve as a function of subject age, O3
2 concentration, minute ventilation, and duration of exposure. In another study, Ultman et al.
3 (2004) found that exposing 60 young, healthy subjects to 0.25 ppm O3 for 1 hr with CE produced
4 considerable intersubject variability in FEVi decrements ranging from 4% improvement to a
5 56% decrement, which was consistent with findings in the 1996 CD.' One third of subjects had
6 FEV, decrements of > 15% and 7% had decrements of > 40%. Foster et al. (1993,1997)
7 examined the effects of O3 on ventilation distribution and reported results suggesting a prolonged
8 O3 effect on the small airways and ventilation distribution (CD, p. 6-5).
9 For prolonged exposures (4 to 8 hr) in the range of 0.08 to 0.16 ppm O3 using moderate
10 quasi continuous exercise (QCE; 50 min exercise [minute ventilation of 35 to 40 L/min] and 10
11 min rest per hr), several pre- and post-1996 studies (Folinsbee et al.; 1988; Folinsbee et al., 1994;
12 Horstman et al., 1990; Adams 2002; Adams 2003) have reported statistically significant
13 responses and increased symptoms and spirometric responses with increasing duration of
14 exposure, O3 concentration, and total minute ventilation. Based on review of several prolonged
15 exposure studies, the draft CD (p. 6-6) concluded that FEVt decrements are a function of minute
16 ventilation in 6.6 hr exposure studies and that data from recent studies do not support the
17 contention that minute ventilation should be normalized to BSA. Triangular exposure studies
18 (Hazucha et al., 1992; Adams,- 2003) suggest that depending upon the profile of the exposure, the
19 triangular exposure, which is more representative of actual ambient exposures, can potentially
20 lead to greater FEV, decrements than square wave exposures at overall equivalent O3 doses (CD,
21 p. 6-8). McDonnell (1996) used data from a series of studies to investigate the frequency
22 distributions of FEVi decrements following 6.6 hr exposures and found that average FEVi
23 responses were relatively small (between 5 and 10 %) at 0,08 ppm O3. However, about 18% of
24 the exposed subjects had moderate functional decrements (10 to 20%), and about 8%
25 experienced large decrements (>20%). This demonstrates that while average responses may
26 appear small and insignificant, some individuals can experience much more significant and
27 severe effects.
28 A relatively large number of field studies investigating the effects of ambient O3
29 concentrations on lung function decrements and respiratory symptoms have been published since
30 1996 and are reviewed and summarized in the draft CD (sections 7.2.3, 7.2.4, and 8.4.2.1.1).
31 These newer studies support the major findings of the 1996 CD that lung function changes, as
32 measured by decrements in FEVi or peak expiratory flow (PEF), and respiratory symptoms are
33 closely correlated to ambient O3 concentrations. Pre-1996 field studies focused primarily on
34 children attending summer camps (Spektor et al., 1988a, 1991; Avol et al., 1990; Raizenne et al.,
35 1987,1989; Higgins et al., 1990). The newer studies have expanded into looking at O3-induced
36 effects on outdoor workers (Brauer et al., 1996; Romieu et al., 1998), athletes (Hoppe et al.,
November 2005 3-10 Draft - Do Not Cite or Quote
-------
1 1998, 2003), the elderly (Schindler et al, 2001; H6ppe et al., 2003), hikers (Korrick et al, 1998),
2 school children (Uhner et al., 1997; Linn et al., 1996; Scarlett et al., 1996), and asthmatics
3 (Hoppe et al., 2003; Romieu et al, 2002). Collectively, these studies confirm and extend clinical
4 observations that prolonged exposure periods, combined with elevated levels of exertion or
5 exercise, may magnify the effect of O3 on lung function. The most representative data come
6 from the Korrick et al. (1998) hiker study, which provided-outcome measures stratified by.
7 several factors (e.g., gender, age, smoking status, presence of asthma) within a population
8 capable of more than normal exertion. In this study, lung function was measured before and.
9 after hiking, and both ambient and personal 03 exposure measurements were made. Decreased
10 lung function was associated with O3 exposure, with the greatest effect estimates reported for
11 subgroup who reported having asthma or wheezing, and those who hiked for longer periods of
12 time, thus increasing the exposure period (CD, p. 7-27, 7-31).
13 Asthma panel studies, conducted both in the U.S. and in other countries, have reported
14 that decrements in PEF are associated with O3 exposures among asthmatic and healthy persons
15 (CD, sections 7.2.3.2 and 8.4.2.1.1). One large U.S. multicity study (Mortimer et al., 2002)
16 examined 03-related changes in PEF in 846 asthmatic children from 8 urban areas and reported
17 that the incidence of > 10% decrements in morning PEF are associated with 30 ppb increase in 8-
18 hr average O3 for a 5-day cumulative lag, suggesting that 03 exposure may be associated with
19 clinically significant .changes in PEF in asthmatic children; however, no associations were
20 reported with evening PEF (CD, p. 7-40). The authors also reported that the associations
21 reported with morning PEF remained statistically significant when days with 8-hr O3
22 concentrations above 80 ppb were excluded (CD, p. 7-40). Two studies (Romieu et al, 1996,
23 1997) carried out simultaneously in northern and southwestern Mexico City with mildly
24 asthmatic school children reported statistically significant Q3-related reductions in PEF, with
25 variations in effect depending on lag time and time of day. While several other studies report
26 statistically significant associations between O3 exposure and reduced PEF in asthmatics (Gielen
27 et al., 1997; Jalaludin et al., 2000; Just et al., 2002; Ross et al., 2002; Thurston et al., 1997), other
28 studies did not (Hilterman et al., 1988; Delfino et al., 1997a), possibly due to very low levels of
29 O3. Collectively, however, these studies indicate that O3 may, be associated with declines in lung
30 function in asthmatic individuals (CD, p. 7-40).
31 Mortimer et al. (2002) discussed biological mechanisms for delayed effects on pulmonary
32 function, which included increased bronchial reactivity secondary to airway inflammation
33 associated with irritant exposure. Animal toxicological and human chamber studies (CD,
34 Chapters 5 and 6) provide supporting evidence that exposure to 03 may augment cellular
35 infiltration and cellular activation, enhance release of cytotoxic inflammatory mediators, and
36 alter membrane permeability (CD, p.7-38). In most laboratory animals studied, biochemical
November 2005 3-11 Draft - Do Not Cite or Quote
-------
1 markers of lung injury and associated morphological changes were not found to be attenuated.
2 Although there are no data available on pulmonary function changes in animals chronically
3 exposed to O3, earlier work of repeated exposures of rats to an episodic profile of O3
4 demonstrated small but significant decrements in lung function consistent with early indicators
5 of focal fibrogenesis in the proximal alveolar region (CD, p. 8-26).
6 Most of the panel studies which have investigated associations between O3 exposure and
7 respiratory symptoms or increased use of asthma medication are focused on asthmatic children
8 (CD, sections 7.2.4 and 8.4.2.1.1). Two large U.S. studies (Mortimer et al., 2002; Gent et al.,
9 2003), as well as several small er U. S. (Delfino et al., 2003; Just et al., 2002; Newhouse et al.,
10 2004; Romieu et al., 1996,1997; Ross et al., 2002; Thurston et al., 1997) and international
11 studies (Hilterman et al., 1998; Desqueyroux et al., 2002a,b), have reported fairly robust
12 associations between ambient O3 concentrations and daily symptoms/asthma medication use,
13 even after adjustment for copollutants. The CD observes mat there are a number of well-
14 conducted, albeit smaller, studies (Avol et al., 1998; Chen et al., 1998; Delfino et al., 1996,
15 1997a, 1998a; Gielen et al., 1997; Jalaludin et al., 2004; Ostro et al., 2001; Taggart et al., 1996)
16 which showed only limited or a lack of evidence for symptom increases associated with O3
17 exposure.
18 The draft CD (p. 7-48) concludes that the asthma panel studies as a group indicate a
19 positive association between ambient concentrations and respiratory symptoms and increased
20 medication use in asthmatics. The evidence has continued to expand since 1996 and now is
21 considered to be much stronger than in the previous review of the 03 primary standard.
22 3.3,1.2,2 Airway Responsiveness
23 Airway hyperresponsiveness (AHR), also know as bronchial hyperreactivity, refers to a
24 condition in which the propensity for the airways to bronchoconstrict due to a variety of stimuli
25 (e.g., exposure to cold air, allergens, or exercise) becomes augmented. This condition is
26 typically quantified by measuring the decrement in pulmonary function (e.g., spirometry or
27 plethysmography) after inhalation exposure to specific (e.g., antigen, allergen) or nonspecific
28 (e.g., methacholine, histamine) bronchoconstrictor stimuli. Exposure to Os causes an increase in
29 nonspecific airway responsiveness as indicated by a reduction in the concentration of
30 methacholine or histamine required to produce a given reduction in' FEVj or increase in SRaw.
31 Increased airway responsiveness is an important consequence of exposure to O3 because its
32 presence means that the airways are predisposed to narrowing on inhalation of various stimuli,
33 such as specific allergens, cold air or SO2 (CD, p. 8-29). Significant, clinically relevant
34 decreases in pulmonary function have been observed in early phase allergen response in subjects
35 with rhinitis after consecutive (4-day) exposure to 0.125 ppm O3 (Holz et al., 2002). Similar
36 increased airway responsiveness in asthmatics to house dust mite antigen 16 to 18 hrs after
November 2005 3-12 Draft - Do Not Cite or Quote
-------
1 exposure to a single dose of O3 (0.16 ppm for 7.6 hrs) was observed. These observations suggest
2 that O3 exposure may be a clinically important factor that can exacerbate the response to ambient
3 bronchoconstrictor substances in individual with preexisting allergic asthma and that its
4 influence may be both immediate and persist for long periods (CD, p. 8-29).
5 An important aspect of increased airway responsiveness after O3 exposure is that it
6 represents a plausible link between 03 exposure and increased hospital admissions. Kreit et al.
7 (1989) found that O3 can induce increased airway responsiveness in asthmatic subjects to O3,
8 who typically have increased airway responsiveness at baseline. A subsequent study (Jdrres et
9 al., 1996) suggested an increase in specific (i.e., allergen-induced) airway reactivity in subjects
10 with allergic asthma, and to a lesser extent in subjects with allergic rhinitis after exposure to 0.25
11 ppm O3 for 3 hrs; other studies (Molfino et al., 1991; Kehrl et al., 1999) reported similar results.
12 According to one study (Folinsbee and Hazucha, 2000), changes in airway responsiveness after
13 O3 exposure resolve more slowly than changes in FEVt or respiratory symptoms. Other studies
14 of repeated exposure to O3 suggest that changes in airway responsiveness tend to be somewhat
15 less affected by attenuation with consecutive exposures than changes in FEVi (Dimeo et al.,
16 1981; Folinsbee et al., 1994; Gong et al., 1997a; Kulle et al., 1982).
17 An extensive laboratory animal data base exploring the effects of acute, long-term, and
18 repeated exposure to O3 indicates that induction of AHR occurs at relatively high O3
19 concentrations. These studies provide clues to the roles of physiological and biochemical
20 components involved in this process, but caution should be exercised in interpreting these
21 results, as different mechanisms may be involved in mediating high- and low-dose responses. As
22 observed in humans, the acute changes in AHR do not persist after long-term exposure of
23 animals exposed to near-ambient concentrations of O3, and attenuation has been reported.
24 The draft CD concludes that exposure is linked with increased AHR (CD, p. 6-43,6-44).
25 Both human and animal studies indicate that airway responses are not associated with
26 inflammation, but they do suggest a likely role for neuronal involvement (CD, 8.4.2.4.2).
27 Increases in AHR do not appear to be strongly associated with decrements in lung function or
28 increases in symptoms (CD, p. 6-30). These findings suggest that different mechanisms underlie
29 reported changes in AHR and changes in lung function or respiratory symptoms.
30 3,3.1.2.3 Respiratory Inflammation and Permeability
31 Short-term exposures to O3 can cause acute respiratory inflammation and increased
32 permeability in the lungs of humans and experimental animals (CD, sections 5.2.3, 6.9, 7.2.5 and
33 8.4.1.4.3). Lung inflammation and increased permeability, which are distinct events controlled
34 by different mechanisms, are two well characterized effects of O3 exposure observed in all
35 species studied. Disruption of the lung barrier leads to leakage of serum proteins, influx of
November 2005 3-13 Draft - Do Not Cite or Quote
-------
1 polymorphonuclear leukocytes (PMNs), release of bioactive mediators, and movement of
2 compounds from the airspaces into the blood.
3 In the animal lexicological studies discussed in the draft CD (Chapter 5), the lowest O3
4 concentration that induced inflammation in the mouse lung was 0.11 ppm for 24 hr exposures.
5 Shorter exposures of 8 hrs required concentrations of 0,26 ppm to induce epithelial permeability
6 thought there was no effect on inflammation. The lowest O3 concentration that affected-
7 epithelial permeability of inflammation in the rat was 0.5 ppm for a 3 hr exposure (CD, p. 8-32).
8 After acute exposures, the influence of the time of exposure increases as the concentration of 03
9 increases. Hie exact role of inflammation in causation of lung disease is not known, nor is the
10 relationship between inflammation and lung function (CD, section 5.2.3).
11 A number of human Cvexposure studies have analyzed bronchoalveolar lavage (BAL)
12 and nasal lavage (ML) fluids and cells for markers of inflammation and lung damage. These
13 studies are summarized in the draft CD (Annex AX6, Tables AX6-12 and AX6-13). Increased
14 lung inflammation is demonstrated by the presence of neutrophils (PMNs) found in BAL fluid in
15 the lungs, which has long been accepted as a hallmark of inflammation. It is apparent, however,
16 that inflammation within airway tissues may persist beyond the point that inflammatory cells are
17 found in the BAL fluid. Soluble mediators of inflammation, such as cytokines and arachidonic
18 acid metabolites have been measured in the BAL fluid of humans exposed to O3. In addition to
19 their role in inflammation, many of these compounds have bronchoconstrictive properties and
20 may be involved in increased airway responsiveness (CD, p. 6-31).
21 In the 1996 CD, assessment of human exposure studies indicated that a single, acute (1 to
22 4 hrs) O3 exposure (0.2 to 0.6 ppm) of subjects engaged in moderate to heavy exercise could
23 induce a number of cellular and biochemical changes suggestive of pulmonary inflammation and
24 lung permeability (CD, p. 8-33). These changes persisted for at least 18 hrs. Graham and Koren
25 (1990) compared inflammatory mediators present in NL and BAL fluids of humans exposed to
26 0.4 ppm O3 for 2 hrs and found similar increases in PMNs in both fluids, suggesting a qualitative
27 correlation between inflammatory changes in the lower airways (BAL) and upper respiratory
28 tract (NL). Acute airway inflammation was shown to occur among adults exposed to 0.08 ppm
29 O3 for 6.6 hrs with exercise (Devlin et al., 1990). Another study (McBride et al., 1994) reported
30 that asthmatic subjects were more sensitive than non-asthmatics to upper airway inflammation
31 for O3 exposures (0.24 ppm, 1.5 hrs, with light IE) that did not affect pulmonary function.
32 Since 1996, a substantial number of human exposure studies have been published which
33 have provided important new information on lung inflammation and epithelial permeability.
34 One publication (Mudway and Kelly, 2004) examined 03-induced inflammatory responses and
35 epithelial permeability with a meta-analysis of 21 controlled human exposure studies. Results of
36 the analysis suggest that for a 1 hr exposure to 0.12 ppm.O3, the threshold dose for early phase
November 2005
3-14
Draft - Do Not Cite or Quote
-------
1 PMN response would not be exceeded unless an individual was engaged in very heavy exercise
2 (minute ventilation of 90 L/min), For 8 hr exposures to 0.08 ppm O3, early phase PMN dose
3 threshold could be reached during relatively light sustained activity (minute ventilation of 17
4 L/min). For these 1- and 8-hr exposure scenarios, BAL fluid protein levels would be predicted to
5 increase by about 1.1 fold.. For late phase PMN responses, threshold dose was 26% greater than
6 for early phase responses. Mudway and Kelly (2004) indicated.that their threshold doses were
7 for average PMN responses in healthy adults and that some individuals would respond at lower
8 doses.
9 A number of more recent studies (Peden et al., 1997; Scannell et al., 1996; Hilterman et
10 al., 1999; Bosson et al., 2003) have provided evidence suggesting that asthmatics show greater
11 inflammatory response than healthy subjects when exposed to similar O3 levels. Markers from
12 BAL fluid following both 2-hr (Devlin et al., 1997) and 4-hr (Christian et al., 1998; Jorres et al.,
13 2000) O3 exposures repeated up to 5 days indicate that there is ongoing cellular damage
14 irrespective of attenuation of some cellular inflammatory responses of the airways, pulmonary
15 function, and symptom responses.
16 The draft CD (p. 8-35) concludes that interaction of O3 with constituents of ELF-and
17 induction of oxidative stress is implicated in injury and inflammation. Alterations in the
18 expression of cytokines, chemokines, and adhesion molecules, indicative of an ongoing active
19 stress response, as well as injury repair and regeneration processes, have been reported in animal
20 toxicology and human in vitro studies evaluating biochemical mediators implicated in injury and
21 inflammation. .
22 3.3.1.2.4 Changes in Host Defense Capability
23 As discussed in section 8.4.1.4.2 of the draft CD, acute exposures to O3 have been shown
24 to impair host defense capabilities in both humans and experimental animals by depressing
25 alveolar macrophage functions and by altering the mucociliary clearance of inhaled particles and
26 microbes. Acute O3 exposures also interfere with the clearance process by accelerating clearance
27 for low doses and slowing clearance for high doses. Animal toxicological studies have reported
28 that acute O3 exposures suppress alveolar phagocytes and immune functions. Dysfunction of
29 host defenses and subsequent increased susceptibility to bacterial lung infection in laboratory
30 animals has been induced by acute exposures to O3 levels as low as 0.08 ppm (CD, p. 8-36).
31 Changes in antibacterial defenses are dependent on exposure regimens, species and strain of lab
32 animals, species of bacteria, and age of the animals used. Acute O3-induced suppression of
33 alveolar phagocytosis and immune function in experimental animals appeared to be transient and
34 attenuated with continuous or repeated exposures (CD, p. 8-36).
35 Ozone exposure has also been show to interfere with AM-mediated clearance in the
36 respiratory region of the lung and with mucociliary clearance of the tracheobronchial airways.
November 2005
3-15
Draft - Do Not Cite or Quote
-------
1 These interferences with clearance are dose dependent, with low doses accelerating clearance
2 and high doses slowing the process. An in vitro study of epithelial cells from nonatopic and
3 atopic asthmatics exposed to 0.01 to 0.10 ppm O3 showed significantly increased permeability
4 compared to cells from normal persons. This indicates a potentially inherent susceptibility of •
5 cells from asthmatic individuals for O3-induced permeability. Although the available
6 information suggests that acute O3 exposures can impair host defense capability by interfering
7 with AM functions, there is no compelling evidence from animal or human studies that O3
8 increases the incidence of respiratory viral infection in humans (CD, p, 8-36).
9 A single human exposure study (Devlin et al., 1991) reviewed in the 1996 CD reported
10 that exposure to 0.08 to 0.10 ppm 03 for 6.6 hrs induced decrements in the ability of alveolar
11 macrophages (AMs) to phagocytose microorganisms; several other studies reported similar
12 effects but with higher exposure concentrations (CD, p. 6-39). Few new controlled human
13 exposure studies are available in this draft CD (section 6.9.6). Integrating the recent study -
14 results with evidence available in the 1996 CD, the draft CD concludes that evidence supports
15 the conclusion that short-term 03 exposures have an adverse impact on host defense capability,
16 even for exposures as low as 0.08 ppm to 0.10 ppm (CD, p. 6-39).
17 3.3.1.2.5 Morphological Effects
18 .In the 1996 CD, it was found that short-term 03 exposures cause similar alterations in
19 lung morphology in all laboratory animal species"studied, including primates. Cells in the
20 centriacinar region (CAR) of the lung (the segment between the last conducting airway and the
21 gas exchange region) have been recognized as a primary target of Cvinduced damage (epithelial
22 cell necrosis and remodeling of respiratory bronchioles), possibly because this region receives
23 the greatest dose of O3 delivered to the lower respiratory tract (CD, p. 8-30). Ciliated cells in the
24 nasal cavity and airways, as wellvas Type I cells in the gas-exchange region, are also identified as
25 targets. Differences in distribution of antioxidants in the CAR of the lungs are responsible for
26 differences in injury and morphological effects observed in nonhuman primates and rodents.
27 While short-term O3 exposures can cause structural changes such as fibrosis in the CAR, these
28 changes appear to be transient with recovery time after exposure dependent on species and O3
29 dose.
30 • Recent studies continue to show that short-term exposures to O3 cause similar alterations
31 in lung structure in a variety of experimental animal species, at concentrations of 0.15 ppm in
32 rats and even lower concentrations in primates (CD, section 5.2.4.1 and 5.2.4.2). New work
33 (Hotchkiss et al., 1998) has shown that a topical anti-inflammatory corticosteroid can prevent
34 these effects in nasal epithelia, while exposure to bacterial endotoxin can potentiate effects
35 (Fanucchi et al., 1998). Ozone-induced fibrotic changes in the CAR are maximal at 3 days of
November 2005
3-16
Draft - Do Not Cite or Quote
-------
1 exposure and recover 3 days post-exposure with exposures of 0,2 ppm 03 in rodents (Dormans et
2 al., 1999). • • .
3 New studies of susceptibility factors demonstrated that ferrets and monkeys have similar
4 inflammatory and necrotic responses to 1 ppm O3, which differs from lesser injury seen in rats
5 (Stemer-Kock et al., 2000). Rats with induced allergic rhinitis are more susceptible to 0.5 ppm>
6 O3 than are controls (Cho et al., 1999b). Important new work has demonstrated variability of
7 local O3 dose and subsequent injury in the respiratory tract due to depletion of glutathione (GSH)
8 (Plopper et al., 1998). The proximal respiratory bronchiole receives the most acute epithelial
9 injury from exposures < 1 ppm, while metabolic effects were greatest in the distal bronchioles
10 and minor daughter airways (CD, p. 5-32).
11 Based on evidence from animal toxicological studies, the draft CD concludes that short-
12 term exposures to O3 can cause morphological changes in the respiratory systems of a number of
13 laboratory animal species (CD, p. 5-38). Little evidence is available from human studies.
14 3.3.1.2.6 Effects on Exercise Performance
15 The effects of O3 exposure on exercise performance of healthy individuals have been •
16 investigated in a number of controlled exposure studies (CD, section 6.7). Several studies
17 reported that endurance exercise performance and VOama* may be limited by acute exposure to O3
18 (Adams and Schelegle, 1983; Schelegle and Adams, 1986; Gong et al., 1986; Foxcroft and
19 Adams, 1986; Folinsbee et al., 1977; Linder et al., 1988). Gong et al. (1986) and Schelegle and
20 Adams (1986) found that significant reductions in maximal endurance exercise performance may
21 occur in well-conditioned athletes while they perform CE (VB >-80 L/min) for 1 hr at O3
22 concentrations > 0.18 ppm. There are no new studies available in the draft CD.
23 Thus, as in the 1996 CD, the draft CD concludes that reports from studies of O3 exposure
24 during high-intensity exercise indicate that breathing discomfort associated with maximal
25 ventilation may be an important factor in limiting exercise performance in some, but not all,
26 subjects (CD; p. 6-29).
27 3.3.1.2.7 Increased School Absences
28 The association between school absenteeism and ambient O3 concentrations was assessed
29 in three relatively large field studies (CD, section 7.2.6). One study (Chen et al., 2000) examined
30 daily school absenteeism in 27,793 elementary school students in Nevada over a 2-year period
31 (after adjusting for PMio and CO concentrations) found that ambient O3 concentrations were
32 associated with 10.41 % excess rate of school absences per 40 ppb increase in 1 -hr max O3.
33 Another study (Gilliland et al., 2001) studied O3-related absences among 1,933 4th grade students
34 in 12 southern California communities and found significant associations between 30-day
35 distributed lag of 8-hr average O3 concentrations and all absence categories, particularly for
November 2005 3-17 Draft - Do Not Cite or Quote
-------
1 respiratory causes. Neither PMio nor N02 were associated with any respiratory or nonrespiratory
2 illness-related absences in single pollutant models. A third study (Park et al., 2002) investigated
3 associations between air pollution and school absences in 1,264 students (1st to 6th grades) in
4 Seoul, Korea and reported that same day O3 concentrations were positively associated with
5 illness-related absences. Models of PMio alone and of both PMio and 03 showed positive
6 associations with illness-related absences, with a slightly greater effect seen for O3.
7 • The draft CD concludes that these studies of school absences suggest that ambient O3
8 concentrations on the same day, as well as accumulated over two to four weeks, may be
9 associated with school absenteeism, but further replication is needed before firm conclusions can
10 be reached regarding this effect (CD, p. 7-52).
11 3.3.1.3 Effects on the Respiratory System from Long-term Exposures
12 The 1996 CD concluded that there was insufficient evidence from the limited number of
13 studies to determine whether long-term O3 exposures resulted in chronic health effects. •
14 However, the aggregate evidence suggested that 03 exposure, along with other environmental
15 factors, could be responsible for health effects in exposed populations (CD, section 7.5). Animal
16 toxicological studies carried out in the 1980' s and 1990' s demonstrated that long-term exposures
17 can result in a variety of morphological effects, including permanent changes in the small
18 airways of the lungs, including remodeling of the distal airways and CAR and deposition of
19 collagen. These changes result from the damage and repair processes that occur with repeated
20 exposure. Fibrotic changes were also found to persist after months of exposure providing a
21 potential pathophysiologic basis for changes in airway function observed in children in some
22 recent epidemiological studies. It appears that variable seasonal ambient patterns of exposure
23 may be of greater concern than continuous daily exposures. One important toxicological study
24 (Tyler et al., 1988) reported that young monkeys exposed to seasonal ambient O3 patterns, but not
25 daily exposures, experienced increases in total lung collagen content, chest wall compliance, and
26 inspiratory capacity, suggesting a delay in lung maturation in seasonally-exposed animals. This
27 section reviews studies published since 1996 in which health effects were assessed for O3
28 exposures lasting from weeks to several years. Summaries of new morphological effects studies
29 of long-term exposures are listed in Table AX5-6 (CD, Annex AX5) and of new morbidity
30 effects epidemiological studies long-term exposure are listed in Table AX7-7 (CD, Annex AX7).
31 3.3.1.3.1 Seasonal Ozone Effects on Lung Function
32 It is well documented in controlled human exposure and field studies that daily multi-
33 hour exposures to O3 produce transient declines in lung function; however, lung function effects
34 of repeated exposures to O3 over extended periods are far less studied. Several studies published
35 since 1996 have investigated lung function changes over seasonal time periods (CD, section
November 2005 3-18 Draft - Do Not Cite or Quote
-------
1 7.5.3). One large, three-year study (Frischer et al., 1999) collected repeated lung function
2 measurements in 1,150 young, Austrian school children and found that growth-related increases
3 in lung function over the summer season were reduced in relation to seasonal mean O3 levels. A
4 one-year extension of this study by Horak et al. (2002a,b) confirmed the results that seasonal
5 mean O3 levels are associated with a negative effect on increases in lung function in children.
6 Another study (Kopp et al., 2000) of 797 children in Austria and southwestern Germany reported
7 smaller increases in lung function in children exposed to higher levels of ambient O3 compared
8 to children living in areas with lower ambient 03 levels. Another Austrian study (Ihorst et al.,
9 2004) of 2,153 young children found significantly lower FVC and FEVi increases associated
10 with higher O3 exposures in the summer but not in the winter. A pilot study (Kinney and
11 Lippmann, 2000) of 72 young adult, military academy students provided results that are
12 consistent with a seasonal decline in lung function that may be due, in part, to O3 exposures.
13 According to the draft CD (p. 7-102), these studies collectively indicate that seasonal O3
14 exposure is associated with smaller increases in lung function in children and that there is limited
15 evidence that seasonal O3 also may affect young adults.
16 . 3,3.1.3.2 Reduced Baseline Lung Function and Respiratory Symptoms
17 Lung capacity grows during childhood and adolescence as body size increases, reaches a
18 maximum during the twenties, and then begins to decline steadily and progressively with age.
19 Long-term exposure to air pollution has long been thought to contribute to slower growth in lung
20 capacity, diminished maximally attained capacity, and/or more rapid decline in lung capacity
21 with age (CD, section 7.5.4). Toxicological findings evaluated in the 1996 CD demonstrated that
22 repeated daily exposure of rats to an episodic profile of Os caused small, but significant,
23 decrements in lung function that were consistent with early indicators of focal fibrogenesis in the
24 proximal alveolar region, without overt fibrosis (CD, section 5.2.5.2). Because.O3 is a strong
25 respiratory irritant and has been shown to cause inflammation and restructuring of the respiratory
26 airways, it is plausible that long-term O3 exposures might have a negative impact on baseline
27 lung function, particularly during childhood when these exposures might have long-term risks.
28 In the current draft CD, however, no recent toxicological studies have been published on effects
29 of chronic Os exposure.
30 Several epidemiological studies published since 1996 have examined the relationship
31 between lung function and long-term 03 exposure. The most extensive and robust study of
32 respiratory effects in relation to long-term air pollution exposures among children in the U.S. is
33 the Children's Health Study carried out in 12 communities of southern California starting in
34 1993 (Avol et al., 2001; Gauderman et al., 2000, 2002, 2004a,b; Peters et al., 1999a,b). One
35 study (Peters et al.,1999a) examined the relationship between long-term O3 exposures and self
36 reports of respiratory symptoms and asthma in a cross sectional analysis and found a limited
November 2005
3-19
Draft - Do Not Cite or Quote
-------
1 relationship between outcomes of current asthma, bronchitis, cough and wheeze and a 40 ppb
2 increase in 1-hr max O3. Another analysis (Peters et al., 1999b) examined the relationship
3 between lung function at baseline and levels of air pollution in the community and reported
4 evidence that annual mean 03 levels were associated with decreases in FVC, FEVi, PEF and
5 FEF25-?5 (the latter two being statistically significant) among females but not males. In a separate
6 study (Gauderman et al., 2000) of 4th, 7th, and 10th grade students, a longitudinal analysis of lung
7 function over four years found no association with 03 exposure. Subsequent studies by the same
8 group (Gauderman et al., 2002,2004a,b) led the authors to conclude that results provide little
9 evidence that ambient O3 at current levels is associated with chronic deficits in the rate of
10 increase in lung function in children. Avol et al. (2001) examined children who had moved from
11 participating communities in southern California to other states with improved air quality and
12 found, with the exception of FEVi, the 03 effect estimates for all other spirometric parameters
13 were negative, but the associations were not as strong as those observed for PMi0. Collectively,
14 the results of these reports from the children's health cohorts provide little evidence for impact of
15 long-term O3 exposures on smaller increases in lung function, but further study is needed to
16 address mis question (CD, p. 7-104). Three other studies (Frisher et al., 1999; Ihorst et al, 2004;
17 Horak et al., 2002a) conducted in Austria and Germany also found no associations between
18 increases in lung function parameters and mean summer O3levels over a three- to five-year
19 period of study.
20 Evidence for a significant relationship between long-term O3 exposures and decrements
21 in maximally attained lung function was reported in a nationwide study (Galizia and Kinney,
22 1999; Kinney et al., 1998) of first year Yale students. Males had much larger effect estimates
23 than females, which might reflect higher outdoor activity levels and correspondingly higher Oj
24 exposures during childhood. A similar study (Kunzli et al., 1997; Tager et al., 1998) of college
25 freshmen at University of California at Berkeley also reported significant effects of long-term O3
26 exposures on lung-function. In a comparison of students whose city of origin was either Los
27 Angeles or San Francisco, long-term O3 exposures were associated with significant changes in .
28 mid- and end-expiratory flow measures, which could be considered early indicators for
29 pathologic changes that might progress to COPD. An autopsy pathologic study (Sherwin et al.,
30 2000) examined subjects for CAR alterations in LA and in Miami, FL. A trend toward greater -
31 extent and severity of CAR alterations, in LA residents was observed,' and it was suggested that
32 this effect might be related to higher O3 exposures in LA.
33 Recent publications from the southern California children's cohort study provide no
34 evidence for an association between long-term Os exposure and lung function in children (CD, p.
35 7-106). Limited evidence is available from studies of adults and college studies to suggest that
36 long-term Os exposure may affect lung function or respiratory symptoms (CD, p. 7-105).
November 2005 3-20 Draft - Do Not Cite or Quote
-------
1 Overall, the draft CD concluded that this body of evidence was generally inconclusive for effects
2 of long-term Os exposure on respiratory symptoms or lung function (CD, p. 7-178).
3 3.3. L 3.3 Long-term O3 Exposure and Respiratory Inflammation
4 As noted earlier in section 3.3.1.2 of this Staff Paper and in the draft CD (Chapter 6),
5 chamber studies of exercising humans exposed to 03 for 2 to 6.6 hrs have demonstrated
6 inflammation in the lungs, including the alveolar region where gas exchange takes place. The
7 potential long-term significance of short-term exposures to O3 is that they can result in the -
8 release of reactive substances from inflammatory cells that can damage the sensitive cells lining
9 the lungs. Over time repeated inflammation can lead to permanent lung damage and
10 • restructuring of the small airways and alveoli. Also since inflammation is a hallmark
11 characteristic of asthma, there is the possibility that O3-induced inflammation may exacerbate
12 existing asthma or contribute to the development of asthma in genetically predisposed
13 individuals (CD, section 7.5.5).
14 For subchronic exposures of animals, permeability changes are transient (and species-
15 dependent) and return to control levels even with continuing exposure. For long-term O3
16 exposures, persistent O3-induced inflammation plays an important role in alterations of lung
17 structure and function. Significant remodeling of the epithelium and underlying connective
18 tissues in distal airways have been reported in rats exposed to 0.25 ppm O3 (12 hr/day for 6
19 weeks) and in monkeys exposed to 0.2 ppm 03 (8hr/day for 90 days) (CD, p. 8-32).
20 In one'epidemiological field study (Kinney et al, 1996), BAL fluids were taken in the
21 summer and winter from a group of joggers in New York and were compared for evidence of
22 acute inflammation and of enhanced cell damage. There was little evidence of acute
23 inflammation in the summer BAL fluids compared to winter, but there was evidence of enhanced
24 cell damage. This suggests that even though inflammation may diminish over the summer, cell
25 damage may be continuing. A series of studies (Calderon-Garciduenas et al., 1995,1997,1999)
26 conducted in Mexico City provides evidence of inflammation and genetic damage to cells in the
27 nasal passages of children chronically exposed to O3 and other air pollutants. Significantly
28 higher DNA damage was reported in children living in Mexico City compared to nonurban
29 children and in older compared to younger children. A more recent study (Calderon-
30 Garciduenas et al., 2003) of Mexico City children reported that they exhibited nasal
31 abnormalities (22%), hyperinflation (67%), interstitial markings (49%), and a mild restrictive
32 pattern by spirometry (10%), while a control group of nonurban children showed no significant
33 abnormalities. Another marker of inflammation, urinary eosinophil, was analyzed in an Austrian
34 school children study (Frischer et al., 2001), and it was reported that O3 exposure was
35 significantly associated with eosinophil inflammation.
November 2005 3 -21 Draft - Do Not Cite or Quote
-------
1 In assessing these studies, the draft CD (p. 7-110) concluded that specific attribution of
2 these adverse respiratory and genotoxic effects to 03 is difficult given the complex mixture in
3 ambient air, although inflammatory changes like eosinophil levels observed in the Austrian study
4 would be consistent with known effects of O3.
5 3.3.1.3.4 Risk of Asthma Development
6 There have been a few studies investigating associations between long-term O3 exposures
7 and the onset of new cases of asthma (CD, section 7.5.6). The Adventist Health and Smog
8 (AHSMOG) study cohort of 3,914 was drawn from nonsmoking, non-Hispanic white adult
9 Seventh Day Adventists living in California (Greer et al., 1993; McDonnell et al., '1999).
10 Subjects were surveyed in 1977,1987, and 1992, and new cases of asthma were defined as self-
11 reported doctor-diagnosed asthma at either the 1987 or 1992 follow-up questionnaire among
12 those who had not reported having asthma when the enrolled in 1977. During the ten-year
13 follow-up in 1987, it was reported that the incidence of new asthma was 2.1 % for males and
14 2.2% for females (Greer etal., 1993). A statistically significant relative risk of 3.12 per 10 ppb
15 increase in annual mean O3 (no exposure metric given) as observed in males, compared to a
16 nonsignificant relative risk of 0.94 in females. In the 15-year follow-up in 1992, it was reported
17 that 3.2% of eligible males and 4.3% of eligible females had developed adult asthma (McDonnell
18 etal., 1999). For males, the relative risk of developing asthma was 2.27 per 30 ppb increase in
19 8-hr average O3, but there was no evidence of an association in females. The lack of an
20 association in females does not necessarily mean there is no effect but may be due to differences.
21 in time-activity patterns in males and females, which could lead to greater misclassification of
22 exposure in females. Consistency of results in the two follow-up studies provides evidence of an
23 association between long-term O3 exposure and asthma incidence in adult males; however,
24 representativeness of this cohort to the general U.S. population may be limited (CD, p. 7-111).
25 • In a similarstudy (McConnell et al., 2002) of incident asthma among children (ages 9 to
26 16 at enrollment), annual surveys of 3,535 children initially without asthma were used to identify
27 new-onset asthma cases as part of the Children's Health Study. Six high-O3(mean 1-hr max O3
28 of 75.4 ppb over four years) and six low-03(50.1 ppb) communities were identified where the
29 children resided. There were 265 children who reported new-onset asthma during the follow-up
30 period. Although asthma risk was no higher for all residents of the six high-O3 versus six low-O3
31 communities, asthma risk was 3.3 times greater for children who played three or more sports as
32 compared with children who played no sports within the high-O3 communities. This association
33 was absent in the communities with lower O3 concentrations. No other pollutants were found to
34 be associated with new-onset asthma.
35 The draft CD (p. 7-111) suggests that playing sports may indicate outdoor activity when
36 O3 levels would be higher, as well as an increased ventilation rate, both of which could increase
November 2005 3-22 Draft - Do Not Cite or Quote
-------
1 O3 exposure and dose. These results, however, were based on a small number (29) of new-onset
2 asthma cases among children who played three or more sports. Future replication of these
3 findings in other cohorts would help support a causal interpretation.
4 3.3.1.3. S Morphological Effects
5 The progression of morphological effects during and after a chronic exposure in the range
6 of 0.5 to 1.0 ppm O3is complex, with inflammation peaking over the first few days of exposure,
7 then dropping, then plateauing, and finally, largely disappearing (CD, section 5.2.4.3). Epithelial
8 hyperplasia follows a somewhat similar pattern. In contrast, fibrotic changes in the tissue -
9 increase very slowly over months of exposure, and, after exposure ceases, the changes
10 sometimes persist or increase. Patterns of exposure in this same concentration range determine -.
11 effects, with 18 months of daily exposure causing less morphologic damage than exposures on
12 alternating months. This is important as environmental O3 exposure is typically seasonal. The
13 long-term study of Plopper et al. (1998) investigated infant rhesus monkeys exposed to
14 simulated, seasonal O3 (0.5 ppm, 8 hrs/day for 5 days, every 14 days for 11 episodes) and
15 demonstrated: 1) remodeling in the distal airways, 2) abnormalities in trachea! basement
16 membrane; 3) eosinophil accumulation in conducting airways; and 4). decrements in airway
17 innervation. The only epidemiological study (Sherwin et al., 2000) that investigated severity of
18 CAR alterations in human lungs compared autopsy results of lungs for Miami and LA residents.
19 The results indicate a significantly greater extent and severity of CAR alterations in LA
20 residents, suggesting that the CAR alterations may be related to higher O3 levels in LA. These
21 findings advance earlier information regarding possible injury-repair processes occurring with
22 seasonal O3 exposures (CD, p. 5-36).
23 3.3.L3.6 Summary
24 In the past decade, important new longitudinal studies have examined the effect of
. 25 chronic O3 exposure on respiratory health outcomes. Evidence from recent long-term morbidity
26 studies have suggested in some cases that chronic exposure to 03 may be associated with
27 seasonal declines in lung function, increases in inflammation, and development of asthma in
28 children and adults. Seasonal decrements or smaller increases in lung function measures have
29 been reported in several studies; however, it remains uncertain to what extent these changes are
30 transient. In contrast to the supportive evidence from animal studies with chronic exposures, •
31 epidemiological studies of new asthma development and longer-term lung function declines
32 remain inconclusive at present (CD, p. 7-118).
33 3.3.1.4 Effects on the Cardiovascular System
34 Epidemiological studies of cardiovascular effects have investigated associations with
35 several air pollutants, including O3, PM, CO, NO2 and SO2, most often with a focus on PM health
November 2005
3-23
Draft ~ Do Not Cite or Quote
-------
1 endpoints (CD, p. 7-54). Several studies also have examined associations of O3 and other
2 gaseous pollutants with heart rate variability (HRV) in the elderly and the increased risk of
3 myocardial infarction associated with exposure to air pollution. One study (Liao et al. 2004)
4 reported larger effect estimates for PM,0 than those for the gaseous pollutants such as 03;
5 however, the findings are suggestive of short-term effects of 03 and other pollutants on HRV.
6 Another study (Park et al., 2005) reported that in analyses by ischemic heart disease,
7 hypertension, and diabetes status, stronger associations of HRV and both O3 and PM^s were
8 observed for individuals with ischemic heart disease and hypertension than those without these
9 preexisting conditions. These results were consistent with another study (Holguin et al., 2003)
10 conducted in Mexico City, which reported an association between decreased HRV with O3 in
11 subjects with hypertension. Other studies did not provide evidence for an O3 effect on cardiac
12 arrhythmias; however, the draft CD notes that the 03 concentrations in these studies were low
13 (CD, p. 7-57).
14 Some hew epidemiological studies have reported associations with more severe effects,
15 such as myocardial infarction. Peters et al. (2001) reported positive, but not statistically
16 significant associations between ozone and the incidence of myocardial infarction in Boston
17 (CD, p. 7-55). The effect estimate for the association with Oa averaged over 2 hrs prior to the
18 myocardial infarction was substantially larger than that reported for an association with 24-hr
19 average Oj (Peters et al., 2001). In France, Ruidavets et al. (2005) reported a statistically
20 significant association between ambient Oj concentrations and incidence of myocardial
21 infarction (CD, p. 7-55).
22 A number of epidemiological studies have also reported associations between short-term
23 exposures and hospitalization for cardiovascular diseases. As shown in Figure 7-13 of the draft
24 CD, the results of these studies are inconsistent (CD, p. 7-72). In addition, in Denver, (Koken et
25 al., 2003) reported associations between hospitalization for specific cardiovascular diseases and
26 ambient O3 concentrations; the associations were inconsistent, however, and generally not
27 statistically significant.
28 There is limited controlled-exposure human experimental information available which
29 suggests that O3 exposure induces cardiovascular effects.(CD, sections 6.10 and AX6.10). In one
30 study, Gong et al. (1998) monitored numerous cardiac variables in both healthy and hypertensive
31 subjects. Results suggested that by impairing the alveolar-to-arterial oxygen transfer, the O3
32 exposure could potentially lead to adverse cardiac events by decreasing oxygen supply to the
33 myocardium. The subjects in the Gong et al. (1998) study had sufficient functional reserve so as
34 not to experience significant ECG changes or myocardial ischemia and/or injury; however, it was
35 concluded that O3 exposure could pose a cardiopulmonary risk to persons with preexisting
36 cardiovascular disease, with or without concomitant respiratory disease. In other research,
November 2005 3-24 Draft - Do Not Cite or Quote
-------
1 Foster et al. (1993) demonstrated that even in relatively healthy young adults, O3 exposure can
2 cause ventilation to shift away from the well perfused basal lung, an effect on small airways
3 which Foster et al. (1997) showed may persist for over 24 hr after exposure. Hypoxic pulmonary
4 artery vasoconstriction acts to shift perfusion away from areas of low ventilation and moderate
5 ventilation-perfusion mismatches (Santak et al., 1998). This arterial vasoconstriction is thought
6 to be mediated by protein kinase C (Barman, 2001; Tsai et al., 2004). A more generalized (i.e.,
7 not localized to poor ventilated ^areas) increase in pulmonary vascular resistance in response to
8 03 exposure would presumably act against the ability of the hypoxic vasoconstriction in
9 mediating ventilation-perfusion mismatches (CD, sections 6-10 and AX6-10).
10 Based on epidemiological study results, the draft CD concludes that the current evidence
11 from field studies is rather limited, but supportive of a potential effect of short-term O3 exposure
12 and heart rate variability, cardiac arrhythmia and incidence of myocardial infarction (CD, p. 7-
13 57). In the draft CD's evaluation of studies of hospital admissions for cardiovascular disease
14 (CD, section 7.3.4), it is concluded that evidence from this growing group of studies is generally
15 inconsistent, but is suggestive of an association with 03 in studies conducted during the warm
16 season (CD, p. 7-73,7-74).
17 3.3.2 Premature Mortality
18 There were only a limited number of studies which examined the relationship between 03
19 and mortality available for review in the 1996 CD. Some studies suggested that mortality was
20 associated with short-term exposure to ozone, but conclusions could not be drawn regarding such
21 associations (U.S EPA, 1996, p. 42). Numerous recent studies have provided new and more
22 substantial evidence supporting such an association, as discussed below in section 3.3.2.1.
23 At the time of the last review, little epidemiological evidence was available on potential
24 associations between long-term exposure to O3 and mortality. Among the recent studies are
25 some that have evaluated this relationship, and still provided limited, if any, evidence for an
26 association between chronic ozone exposure and mortality, as described in section 3.3.2.2.
27 3.3.2.1 Mortality and Short-term 63 Exposure
28 In the 1996 CD, some, but not all, epidemiological studies had reported associations
29 between short-term ozone exposure and mortality. Given the information available from the
30 studies modeling for O3 and weather, it was difficult to evaluate whether associations were
31 possibly not achieving statistical significance due to overspecification of the weather model. •• .
32 Based on Ihe limited evidence available from the epidemiological studies and the uncertainties
33 regarding weather model specifications, the 1996 CD was unable to quantitatively assess the O3-
34 mortality relationship, or even provide qualitative assessment of the likelihood of Os-mortality
35 associations (CD, p. 7-74).
November 2005
3-25
Draft - Do Not Cite or Quote
-------
1 The draft CD includes results from numerous epidemiological analyses of the
2 relationship between O3 and mortality. Key findings are available from multicity time-series
3 studies that report associations between O3 and mortality. These studies include analyses using
4 data from 90 U.S. cities in the National Mortality, Morbidity and Air Pollution (NMMAPS)
5 study (Samet et al., 2000, reanalyzed in Dominici, 2003) and from 95 U.S. cities in an extension
6 to the NMMAPS analyses (Bell et al., 2004), and further analyses using a subset of 19 U.S. cities
7 and focusing on cause-specific mortality associations (Huang et al.5 2005). An additional study
8 used case-crossover design and data from 14 U.S. cities (Schwartz, 2005), to further investigate
9 the influence of adjustment for weather variables in the O3-mortality relationship. Finally,
10 results are available from a European study, Air Pollution and Health: a European Approach
11 (APHEA), an analysis using data from 23 cities (Gryparis et al., 2004) and 4 cities (Touloumi et
12 al., 1997).
13 In the original 90-city NMMAPS analysis, which was primarily focused on investigating
14 effects of PMio on mortality, a significant association was reported between mortality and 24-hr
15 average O3 concentrations during the warm season, but the association was not significant in
16 analyses for the full year (Samet et al., 2000; CD, Figure 7-19; p. 7-86). The extended
17 NMMAPS analysis included data from 95 U.S. cities and included data from 1987-2000 (Bell, et
18 al., 2004), and significant associations were reported between ozone and mortality. The effect
19 estimate for increased mortality was 0.5% per 24-hr average O3 measured on the same day (20
20 ppb change; 95% PI: 0.24, 0.78), and 1.04% per 24-hr average O3 in a 7-day distributed lag
21 model (20 ppb change, 95% PI: 0.54,135) (CD, p. 7-78). In analyses using only data from the
22 warm season, the results were not significantly different from the full-year results; the effect
23 estimate for increased mortality was 0.44% per 24-hr average O3 measured on the same day (20
24 ppb change; 95% PI: 0.14,0.74), and 0.78% per 24-hr average O3 in a 7-day distributed lag
25 model (20 ppb change, 95% PI: 0.26,1.30) (CD, p. 7-85). The authors also report that O3-
26 mortality associations were robust to adjustment for PM (CD, p. 7-87).
27 Using a subset of the NMMAPS data set, Huang et al. (2005) focused.on associations
28 between cardiopulmonary mortality and ozone exposure (24-hr average) during the summer
29 season only. The authors report a 1.47% increase per 20 ppb change in O3 concentration
30 measured on the same day (95% PI: 0.54, 2.39), and a 2.52% increase per 20 ppb change in O3
31 concentration using a 7-day distributed lag model (95% PI: 0.94,4.10) (CD, p. 7-78, 7-80).
32 As discussed in section 3.4, assessment of confounding by weather, especially
33 temperature, is complicated by the fact that higher temperatures are important for O3 formation.
34 Using case-crossover study design, Schwartz (2005) assessed associations between daily
35 maximum concentrations and mortality, matching case and control periods by temperature, and
36 using data only from the warm season. The reported effect estimate of 0.92% change in
November 2005 3-26 Draft - Do Not Cite or Quote
-------
1 mortality per 40 ppb O3 (1-hr maximum, 95% PI: 0.06,1.80) was similar to time-series analysis
2 results with adjustment for temperature (0.76% per 40 ppb O3; 95% Probability Interval: 0.13,
3 1.40), suggesting that associations between O3 and mortality are not sensitive to these adjustment
4 methods for temperature (CD, p. 7-80, 7-81).
5 An initial publication from APHEA, a European multi-city study, reported statistically
6 significant associations between daily maximum O3 concentrations and mortality, with an effect
7 estimate of 4.5% change in mortality per 40 ppb O3 (95% CI: 1.6, 7.7) in four cities (Touloumi et
8 al., 1997). An extended analysis was done using data from 23 cities throughout Europe
9 (Gryparis et al., 2004). In this report, a positive but not statistically significant association was
10 found between mortality and 1-hr daily maximum O3 in a full year analysis (CD, p. 7-81).
11 Focusing on analyses using summer measurements, the authors report statistically significant
12 associations with total mortality [1.8% increase per 30 ppb 8-hr ozone (95% CI: 0.8,2.9)],
13 cardiovascular mortality [2.7% increase per 30 ppb 8-hr ozone (95% CI: 1.2, 4.3)] and with
14 respiratory mortality [6.8% increase per 30 ppb 8-hr ozone (95% CI: 4.5, 9.2)] (Gryparis et al.,
15 2004; CD, p. 7-85).
16 Two of the recent multi-city mortality studies have also reported associations for multiple
17 averaging times (Bell et al., 2004; Gryparis et al., 2004). Bell and colleagues (2004) reported
18 associations between mortality and 1-hr daily maximum, 8-hr daily maximum and 24-hr average
19 O3 concentrations. Effect estimates for associations with 1-hr O3 was slightly larger than that
20 reported for 8-hr O3 concentrations, and both were distinctly larger than the association with 24-
21 hr average O3, but the effect estimates did not differ statistically. Gryparis et al. (2004) also
22 reported effect estimates that were slightly larger with 1 -hr than with 8-hr O3 concentrations, but
23 not significantly so.
24 Numerous single-city analyses have also reported associations between mortality and
25 short-term O3 exposure, especially for those analyses using warm season data. As shown in
26 Figure 7-19 of the draft CD, the results of recent publications show a pattern of positive, often
27 statistically significant associations between short-term O3 exposure and mortality during the
28 warm season (CD, p. 7-86). For example, statistically significant associations were reported in
29 southern California (Ostro et al., 1995), Philadelphia (Moolgavkar et al., 1995), Dallas (Gamble
30 et al., 1998), and Vancouver (Vedal et al., 2003) as well as numerous studies conducted in other
31 countries. However, no evidence of an association was seen in a study conducted in Pittsburgh
32 (Chock et al., 2000). In considering results from year-round analyses, there remains a pattern of
33 positive results but the findings are less consistent. For example, statistically significant
34 associations were reported in Philadelphia (Moolgavkar et al., 1995) and Dallas (Gamble et al.,
35 1998) while positive but not statistically significant associations were reported in Detroit
36 (Lippmann et al., 2000; reanalyzed in Ito, 2003), San Jose (Fairley, 1999; reanalyzed Fairley,
November 2005 3-27 Draft - Do Not Cite or Quote
-------
1 2003), and Atlanta (Klemm et al., 2004). No evidence for associations was reported in Los
2 Angeles (Kinney et al., 1995), Coachella Valley (Ostro et al., 2003) and St. Louis and Eastern
3 Tennessee (Dockery et al., 1992). In most single-city analyses, effect estimates were not
4 substantially changed with adjustment for PM (CD Figure 7-20, p. 7-88).
5 In addition, several meta-analyses have been conducted on the relationship between 03
6 and mortality. As described in section 7.4.4 of the draft CD, these analyses reported.fairly
7 consistent and positive combined effect estimates ranging from 1.5 to 2.5% increase in mortality
8 for a standardized change in O3 (CD Figure 7-18, page 7-84). Three recent meta-analyses
9 evaluated potential sources of heterogeneity in Cvmortality associations (Bell et al., 2005; Ito et
10 al., 2005; Levy et al., 2005). The draft CD observes common findings across all three analyses,
11 in that all reported that effect estimates were larger in warm season analyses, reanalysis of results
12 using default GAM criteria did not change the effect estimates, and there was no strong evidence
13 of confounding by PM (CD, p. 7-84). Bell et al. (2005) and Ito et al. (2005) both provided
14 suggestive evidence of publication bias, but Cvmortality associations remained after accounting
15 for that potential bias. The draft CD concludes that Ihese studies "provide strong evidence that
16 O3 is associated with mortality." (CD, p. 7-84)
17 Taken together, the draft CD concludes that the epidemiological evidence shows robust
18 associations between daily O3 concentrations and mortality (CD, p. 7-96). For standardized
19 ozone increments, effect estimates range from 0.5 to 2.5% increases in mortality in the multi-city
20 studies, and from 0.5 to 5% in single-city studies. For most studies that conducted season-
21 specific analyses, effects were larger and more precise in warm-season analyses (CD, p. 7-96).
22 The results of multi-city studies, single-city studies and the recent meta-analyses all indicate that
23 associations reported between O3 and mortality are also robust to adjustment for PM in the
24 models (CD, p. 7-89). Finally, from those studies that included assessment of associations with
25 specific causes of death, it appears that effect estimates for associations with cardiovascular
26 mortality are larger than those for total mortality; effect estimates for respiratory mortality are
27 less consistent in size, possibly due to reduced statistical power (CD, p. 7-94).
28 33.2.2 Mortality and Long-term Os Exposure
29 Little evidence was available in the previous O3 NAAQS review on the potential for
30 associations between mortality and long-term exposure to O3. In the Harvard Six City
31 prospective cohort analysis, the authors report that mortality was not associated with long-term
32 exposure to O3; the mortality and 03 concentration data are presented in a figure but quantitative
33 concentration-response functions.are not available in this report (Dockery et al., 1993). The
34 authors note that the range of O3 concentrations across the six cities was small (19;7 to 28.0 ppb
35 in average 24-hr concentrations over the 7-year study period) which may have limited the power
November 2005
3-28
Draft - Do Not Cite or Quote
-------
1 of the study to detect associations between mortality and O3 levels (Dockery et al., 1993, p.
2 1755).
3 As discussed in section 7.5.8 of the draft CD, in this review there are results available
4 from three prospective cohort studies: the American Cancer Society (ACS) study, the Adventist
5 Health and Smog (AHSMOG) study, and the U.S. Veterans Cohort study. In addition, a major
6 reanalysis report includes evaluation of data from the Harvard Six City cohort study (Krewski et
7 al., 2000). This reanalysis also includes additional evaluation of data from the initial ACS cohort
8 study report that had only reported results of associations between mortality and long-term
9 exposure to fine particles and sulfates (Pope et al., 1995).'
10 In this reanalysis of data from previous prospective cohort study reports, the investigators
11 replicated and validated the findings of the original studies, and the report included additional
12 quantitative results beyond those available in the original report (Krewski et al., 2000). In the
13 reanalysis of data from the Harvard Six Cities study, the effect estimate for the association
14 between long-term O3 concentrations (8.3 ppb between the highest and lowest concentrations in
15 the cities) and mortality was negative and nearly statistically significant (relative risk ~ 0.87,
16 95% CI 0.76,1.00) (Krewski et al., 2000, p. 150).
17 The ACS study is based on health data from a large prospective cohort of approximately
18 500,000 adults and air quality data from about 150 U.S. cities. The initial report (Pope et al.,
19 1995) focused on associations with fine particles and sulfates, for which significant associations
20 had been reported in the earlier Harvard Six Cities report (Dockery et al,, 1993). As part of the
21 major reanalysis of these data, results for associations with other air pollutants were also
22 reported, and the authors report that no significant associations were found with O3. However,
23 results of seasonal analyses show a small positive association between long-term O3
24 concentrations in the warm months (April-September) with a relative risk of 1.02 for all-cause
25 mortality (95% CI 0.96-1.07) and a stronger association was reported for cardiopulmonary
26 mortality (relative risk 1.08, 95% CI 1.01 -1.16) (Krewski et al., 2000, p. 174).
27 A recent report from this study reported results of associations with an extended data
28 base; the mortality records for the cohort had been updated to include 16 years of follow-up
29 (compared with 8 years in the first report) and more recent air quality data were included in the
30 analyses (Pope et al., 2002). Results are presented for full-year analyses, and show no evidence
31 for a significant association between long-term exposure to O3 and mortality. As shown in
32 Figure 7-24 of the draft CD, the effect estimates are near zero and sometimes negative (though
1 This reanalysis report and the original prospective cohort study findings are discussed in more detail in section
8.2.3 mAir Quality Criteria for Particulate Matter (EPA, 2004).
November 2005
3-29
Draft - Do Not Cite or Quote
-------
1 not statistically significant) for associations between long-term O3 exposure and all-cause,
2 cardiopulmonary, and lung cancer mortality (CD, p. 7-112).
3 The Adventist Health and Smog (AHSMOG) cohort includes about 6,000 adults living in
4 California. In two studies from this cohort, a significant association has been reported between
5 long-term 03 exposure and increased lung cancer risk among males only (Beeson et al, 1998;
6 Abbey et al., 1999). No significant associations were reported between long-term 03 exposure
7 and mortality from all causes or cardiopulmonary causes, however (Abbey et al., 1999). Due to
8 the small numbers of lung cancer deaths (12 for males, 18 for females) and the precision of the
9 effect estimate (i.e., the wide confidence intervals), the draft CD raised concerns about the
10 plausibility of the reported association with lung cancer (CD, p. 7-114.)
11 The U.S. Veterans Cohort study (Lipfert et al., 2000b) of approximately 50,000 middle-
12 aged males diagnosed with hypertension, reported some positive'associations between mortality
13 and peak O3 exposures (95th percentile level for several years of data). The analysis included
14 numerous analyses using subsets of exposure and mortality follow-up periods which spanned the
15 years 1960 to 1996. In the results of analyses using deaths and O3 exposure estimates
16 concurrently across the study period, there were positive, statistically significant associations
17 between peak O3 and mortality, with a 9.4% excess risk (95% CI: 0.4,18.4) per mean 95%
18 percentile O3 (CD, p. 7-112).
19 Thus, the results from all-year analyses in the Harvard Six Cities and ACS cohorts
20 provide no evidence for associations between long-term O3 exposure and mortality, though the
21 warm-season results in the reanalysis of the ACS cohort study suggest a potential association.
22 Imprecise and inconclusive associations were reported in analyses for the AHSMOG cohort
23 study. Significant associations between long-term O3 exposure and mortality were only reported
24 for the Veterans cohort study; while this study used an indicator of peak ozone concentrations.,
25 the cohort is also a rather specific subgroup of the U. S. population. Overall, the draft CD
26 concludes that consistent associations have not been reported between long-term O3 exposure
27 and all-cause, cardiopulmonary or lung cancer mortality (CD, p. 7-114).
28 3.3.3 Summary
29 The draft CD (Chapters 4-8) summarizes and assesses substantial new evidence which
30 builds upon what was previously known about the health effects of O3. The new information
31 supports previous findings that short-term O3 is associated with lung function decrements and
32 respiratory symptoms, as well as numerous more subtle effects on the respiratory system such as
33 morphological changes and altered host defense mechanisms (CD, p. 7-177). Short-term O3
34 exposure has also been associated with hospital admissions for respiratory causes in numerous
35 new studies that support findings in the 1996 CD. The draft CD reports that warm-season
November 2005 3-30 Draft - Do Not Cite or Quote
-------
1 studies show evidence for positive and robust associations between ambient O3 concentrations
2 and respiratory admissions, and positive but less conclusive evidence for associations with
3 respiratory ER visits (CD, p. 7-177).
4 Some new studies have suggested associations between increased incidence of asthma or
5 reduced lung function and long-term exposure to elevated ambient O3 levels. The findings of
6 this small group of studies are inconsistent, however, and the draft CD concludes lhat the
7 evidence for this group of associations is inconclusive (CD, p. 7-178).
8 A new body of studies has suggested associations between short-term 03 exposure and
9 effects on the cardiovascular system, including changes in heart rate variability, cardiac
10 arrhythmia, incidence of myocardial infarction and hospitalization for cardiovascular diseases.
11 The draft CD finds this body of evidence to be limited but supportive of potential effects of O3
12 on the cardiovascular system (CD, p. 7-177).
13 The major new information presented in the draft CD is evidence of an elevated risk of
14 mortality associated with acute exposure to O3, especially in the summer or warm season when
15 O3 levels are typically high. Recent meta-analyses also showed risk estimates that are consistent
16 across studies and robust to control for potential confounders, providing strong evidence for an
17 association between short-term O3 exposure and mortality (CD, p. 7-78). The limited evidence
18 from long-term O3 exposure studies, however, does not provide conclusive evidence of an
19 association with mortality.
20 3.4 INTEGRATIVE ASSESSMENT OF EVIDENCE FROM EPIDEMIOLOGICAL
21 STUDIES
22 In Chapter 8, the draft CD assesses the new health evidence, integrating findings from
23 experimental (e.g., toxicological, dosimetric and controlled human exposure) and
24 epidemiological studies, to make judgments about the extent to which causal inferences can be
25 made about observed associations between health endpoints and exposure to Oj. In evaluating
26 the evidence from epidemiological studies in section 8.4, the draft CD focuses on well-
27 recognized criteria, including: (1) the quality of the exposure metrics; (2) the quality and size of
28 the study population; (3) the robustness of reported associations to the use of alternative model
29 specifications and potential confounding by co-pollutants; (4) the strength of reported
30 associations; (5) evidence for temporality between exposure and observed effects (CD, p. 8-10).
31 Integrating more broadly across epidemiological and experimental evidence, the draft CD also
32 focuses on the coherence and.plausibility of observed Os-related health effects to reach
33 judgments about causality (Section 8.4.8); this will be the focus of section 3.5.
34 The following discussion summarizes the conclusions and judgments from the draft CD's
35 integrative assessment, focusing in particular on discussions of strength, consistency, robustness,
November 2005
3-31
Draft - Do Not Cite or Quote
-------
1 temporality and in the epidemiolbgical evidence. This section also addresses several issues
2 relevant to the interpretation of epidemiological evidence. In particular, staff has focused on
3 issues related to exposure error in 03 epidemiological studies, potential confounding by co-
4 pollutants, the affect of alternative model specifications on O3-health associations, the lag periods
5 between O3 ambient exposure levels and health outcomes, and the nature of Orhealth
6 concentration-response relationships.
7 3.4.1 Strength of Associations
8 The strength of associations most directly refers to the magnitude of the reported relative
9 risk estimates. Taking a broader view, the CD draws upon the criteria summarized in a recent
10 report from the U.S. Surgeon General, which define strength of an association as "the magnitude
11 of the association and its statistical strength" which includes assessment of both effect estimate
12 size and precision, which is related to the statistical power of the study (CD, p. 8-10; CDC,
13 2004). In general, when associations are strong in terms of yielding large relative risk estimates,
14 it is less likely that the association could be completely accounted for by a potential confounder
15 or some other source of bias (CDC, 2004). With associations that yield small relative risk
16 estimates it is especially important to consider potential confounding and other factors in
17 assessing causality.
18 Effect estimates between O3 and various health outcomes are generally small in size. For
19 example, the draft CD reports effect estimates in the range of 0.5 to 5% increase in mortality per
20 40 ppb increase in Ch or equivalent (CD, p. 7-80). The magnitude of these associations, while
21 small, is found to be relatively consistent between studies (CD, p. 8-12). As shown in Figure 8-5
22 in the draft CD, effect estimates for associations with emergency department visits and hospital
23 admissions range up to 40% increases per incremental change in Os (CD, p. 8-47).
24 In considering both the magnitude and statistical strength of the associations, a pattern of
25 positive and often statistically significant associations can be seen between mortality and
26 respiratory morbidity and short-term exposures to Os (CD, pp. 7-45, 7-46,7-74). Such
27 associations are strong in terms of the precision of the studies; that is, the associations were
28 strong enough to have been reliably measured by the studies such that many of the associations
29 can be distinguished from the null hypothesis with statistical confidence. Thus, while the
30 associations reported in the more recent body of epidemiological studies are approximately
31 characterized as being modest in terms of the magnitude of the relative risk estimates, such
32 modest associations are generally coherent with outcomes that may reasonably be expected.
November 2005 3-32 Draft - Do Not Cite or Quote
-------
1 3.4.2 Robustness of Associations
2 In section 8.4.7, the draft CD evaluates the robustness of epidemiological associations,
3 reporting that the health associations reported with short-term exposure to Os are generally
4 robust (CD, p. 8-58). In this evaluation, the draft CD focuses on the impact of exposure error,
5 statistical model specification, and potential confounding by co-pollutants on Os-health
6 associations; these issues are discussed below.
7 3.4.2.1 Exposure Error
8 In time-series epidemiological studies, concentrations measured at ambient monitoring
9 stations are generally used to represent a community's exposure to ambient Os. For time-series
10 studies, the emphasis is on the temporal (e.g., daily or hourly) changes in ambient Os. In cohort
11 or cross-sectional studies, air quality data averaged over a period of months to years are used as
12 indicators of a community's long-term exposure to ambient 03 and other pollutants. In both
13 types of analyses, exposure error is an important consideration as actual exposures to individuals
14 in the population will vary across the community. As described in the draft CD, there are very
15 few sources of Os exposure for most people other than ambient air; one potential source is 03
16 emitted from office equipment (CD, p. 7-6). Exposure to ambient Os for individuals is
17 influenced by factors related to the infiltration of Os into buildings, as well as the time spent out
18 of doors by the individuals.
19 The draft CD discusses the potential influence of exposure error on epidemiological study
20 results in section 7.1.3.1. Three components to exposure measurement error are outlined: (1) the
21 use of average population rather than individual exposure data; (2) the difference between
22 average personal ambient exposure and ambient concentrations at central monitoring sites; and
23 (3) the difference between true and measured ambient concentrations (CD, p. 7-7). These
24 components are expected to have different effects, with the first and third likely not causing bias
25 in a particular direction ("nondifferential error") but increasing the standard error, while the
26 second component may result in downward bias, or attenuation of the risk estimate (CD, p. 7-7).
27 Some recent studies have evaluated the impact of exposure measurements error on Os
28 effect estimates. Navidi et al. (1999) used data from a children's cohort study to compare effect
29 estimates from a simulated "true" exposure level to results of analyses from Os exposures
30 determined by several methods. The results indicated that the use of Os exposures from personal
31 sampling or microenvironmental approaches is associated with nondifferential error in Os effect
32 estimates, compared with effect estimates from "true" exposures. However, Os exposures based
33 on the use of ambient monitoring data overestimates the individual's O3 exposure and thus
34 generally results in Os effect estimates that are biased downward (CD, p. 7-7). Similarly, Zidek
35 (1997) used data from a previous epidemiological study and observed that accounting for
November 2005
3-33
Draft - Do Not Cite or Quote
-------
1 exposure measurement error produced results that were similar to the conclusions of the original
2 report, but the effect estimates were considerably larger in magnitude (CD, p. 7-7).
3 The draft CD concludes that ambient Os concentrations "may serve as valid surrogate
4 measures for aggregate personal Oa exposures in time-series studies. However, using ambient
5 concentrations to determine exposure generally overestimates true personal Os exposures,
6 resulting in biased descriptions of underlying concentration-response relationships." (CD, p. 7-
7 8) The draft CD recommends caution in the quantitative use of these effect estimates as they
8 may lead to underestimation of the overall health impact of air pollution. In using
9 epidemiological study results for quantification of health risks for certain health outcomes, staff
10 recognizes that the risk estimates may be underestimating true public health risk. However, staff
11 observes that the use of risk estimates for comparing relative risk reductions between alternative
12 Cb standards considered in the risk assessment is less likely to suffer-from this concern. In
13 addition, as discussed in Chapter 5, staff has conducted an exposure assessment in conjunction
14 with a portion of the health risk assessment that incorporates estimated population exposures in
15 developing risk estimates for health outcomes based on controlled human exposure studies,
16 where concern about use of ambient concentrations as a surrogate for exposure is not an issue.
17 3.4.2.2 Confounding by Copollutants
18 Confounding occurs when a health effect that is caused by one risk factor is attributed to
19 another variable that is correlated with the causal risk factor; epidemiological analyses attempt to
20 adjust or control for potential confounders. Copollutants (e.g., PM, CO, SC*2 and NCh) can meet
21 the criteria for potential confounding in O3-health associations if they are potential risk factors
22 for the health effect under study and are correlated with O3. Effect modifiers include variables
23 that may influence the health response to the pollutant exposure (e.g., co-pollutants, individual
24 susceptibility, smoking or age). Both are important considerations for evaluating effects in a
25 mixture of pollutants, but for confounding, the emphasis is on controlling or adjusting for
26 potential confounders in estimating the effects of one pollutant, while the emphasis for effect
27 modification is on identifying and assessing the level of effect modification.
28 The draft CD observes that 03 is generally not highly correlated with other criteria
29 pollutants (e.g., PMio, CO, SO2 and NO2), but may be more highly correlated with fine particles,
30 especially during the summer months (CD, p. 7-130). In addition, the correlation between O3
31 and other pollutants may vary across seasons, since O3 concentrations are generally,higher in the
32 summer months. This may lead to negative correlations between 03 and other pollutants during
33 the cooler months, but positive associations between O3 and pollutants such as fine particles
34 during the wanner months (CD, p. 7-14). Thus, the draft CD pays particular attention to the
November 2005 3-34 Draft - Do Not Cite or Quote
-------
1 results of season-specific analyses and studies that assess effects of PM in potential confounding
2 of Da-health relationships in its discussions in section 7.6.4.
3 Multipollutant models are commonly used to assess potential confounding in
4 epidemiological studies. As discussed in the draft CD, the limitations to the use of
5 multipollutant models include the difficulty in interpreting results where the copollutants are
6 highly colinear, or where correlations between pollutants change by season (CD, p. 7-131). This
7 is particularly the situations where 03 and a copollutant, such as sulfates, are formed under the
8 same atmospheric condition; in such cases multipollutant models would produce unstable and
9 possibly misleading results (CD, p. 7-132).
10 For mortality, the results from numerous multi-city and single-city studies are shown in
11 Figure 7-20 of the draft CD. These results indicate that 03-mortality associations do not appear
12 to be substantially changed in multipollutant models including PMio.or PMi5 (CD, p. 7-88).
13 Focusing on results of warm season analyses, Figure 7-21 of the draft CD shows risk estimates
14 for Cvmortality associations that are fairly robust to adjustment for PM in multipollutant models
15 (CD, p. 7-90).
16 Similarly, multipollutant models are presented for associations between short-term O3
17 exposures and respiratory hospitalization in Figure 7-12 of the draft CD; the draft CD concludes
18 that copollutants generally do not confound the relationship between O3 and respiratory
19 hospitalization (CD, p. 7-70,7-71). Multipollutant models were not used as commonly in studies
20 of relationships between respiratory symptoms or lung function with O3, but the CD reports that
21 results of available analyses indicate that such associations were robust to adjustment for PMzs
22 (CD, p. 7-134). In reports from U.S. multi-city studies of respiratory symptoms, associations
23 with O3 were found to be remain statistically significant and little changed in magnitude in two-
24 pollutant models including PMio or PM2.3 (Mortimer et al., 2002; Gent et al., 2003; CD pp. 7-45,
25 7-46)."
26 Considering this body of studies, the draft CD concludes: "These findings indicate that
27 the effects of O3 on various health outcomes are robust and independent of the effects of other
28 copollutants." (CD, p. 7-151) Staff will use the results of single-pollutant model results in
29 presentation of results in this chapter and in quantitative assessments conducted as part of mis
30 review for purposes of comparing results from different studies. However, staff will also include
31 the use of multi-pollutant model results in presenting risk estimates, Where available to more
32 completely assess the quantitative health risks associated with alternative Os standards.
33 3.4.2.3 Model Specification
34 The draft CD observes that one challenge of time-series epidemiological analysis is
35 assessing the relationship between 03 and health outcomes while avoiding bias due to
November 2005
3-35
Draft - Do Not Cite or Quote
-------
1 confounding by other time-varying factors, particularly seasonal trends and weather variables.
2 (CD;'p. 7-12) These variables are of particular interest because O3 concentrations have a well-
3 characterized seasonal pattern (see Chapter 2) and are also highly correlated with changes in
4 temperature. Thus it can be difficult distinguish effects associated with O3 or the seasonal or
5 weather variables in statistical analyses.
6 Section 7.1.3.4 of the draft CD discusses statistical modeling approaches that have been
7 used to adjust for time-varying factors, highlighting a series of analyses that were done in a
8 Health Effects Institute-funded reanalysis of numerous time-series studies. While the focus of
9 these reanalyses was on associations with PM, a number of investigators also'examined the
10 sensitivity of 03 coefficients to the extent of adjustment for temporal trends and weather factors.
11 In addition, several recent studies, including U.S. multi-city studies (Bell et al., 2005; Huang et
12 al., 2005; Schwartz et al., 2005) and a meta-analysis study (Ito et al., 2005) evaluated the effect
13 of model specification on Cvmortality associations (CD, p. 7-14). These studies generally report
14 that associations reported with O3 are not substantially changed with alternative modeling
15 strategies (CD, p. 7-122). Overall, the draft CD concludes that O3 effects are generally robust to
16 various model specifications for temporal trend adjustment though additional research would
17 help better understand the influence of different modeling approaches. (CD, p. 7-150)
18 A number of epidemiological studies have conducted season-specific analyses, as
19 discussed in section 7.6.3.2 of the draft CD. As observed above in section 3.3, such'studies have
20 generally reported stronger and more precise effect estimates for O3 associations in the warm
21 season than in analyses conducted in the cool seasons or over the full year. For assessing
22 relationships between O3 and health outcomes, the draft CD highlights several reasons to focus
23 on warm season analyses: (1) the seasonal nature of O3 concentrations; (2) the relationship
24 between O3 formation and temperature; (3) correlations between other pollutants, particularly
25 fine particles, and O3 varies across seasons in some areas; and (4) factors affecting exposure to
26 ambient O3, such as air conditioning use, varies seasonally in most areas of the U.S. The draft
27 CD concludes: "Given the potentially significant effect of season, O3 effect estimates computed
28 for year-round data need to be interpreted with caution." (CD, p. 7-130) Staff has therefore
29 focused on epidemiological findings from warm season analyses, where available, for
30 quantitative and qualitative assessments.
31 3.4.3 Consistency
32 Consistency refers to the persistent finding of an association between exposure and
33 outcome in multiple studies of adequate power in different persons, places, circumstances and
34 times (CDC, 2004). The draft CD observes that the magnitude of effect estimates is relatively
35 consistent across the results of recently published studies of associations between short-term
November 2005 3-36 Draft - Do Not Cite or Quote
-------
1 ozone exposure and mortality (CD, p. 8-41). Including results from studies conducted on several
2 continents, the draft CD finds that effect estimates range from 0.5 to 5% increases in mortality
3 per incremental change in Os (CD, p. 8-56). For associations with morbidity, the draft CD
4 reports that there is greater variability in effect estimates between studies, but that fairly
5 consistent positive associations between short-term ozone exposure and various respiratory
6 health outcomes have been observed (CD, p. ^51). ;
1 While concluding that Os-health associations are found to be generally consistent, the
8 recent Da-mortality meta^analyses indicate that some heterogeneity exists across studies (CD, p.
9 7-84). The draft CD discusses a number of factors that could result in heterogeneity in
10 associations between different geographic areas, focusing particularly on variables that can affect
11 exposure to ambient Os. For example, the use of air conditioning can reduce ambient exposures
12 during the warm season, while increased outdoor activity can increase exposure. In addition, as
13 discussed previously, stronger and more consistent associations have been reported in analyses
14 for the warm season. The meta-analysis results indicated that these two factors were important
15 sources of variability in Os-mortality relationships (Levy et al., 2005; CD, p. 8-57), but also
16 reported that the associations were fairly consistent between studies; addressing these sources of
17 heterogeneity resulted in stronger and more consistent findings.. Overall, the epidemiological
18 findings are fairly consistent for associations between short-term exposure to Os and both
19 respiratory morbidity and mortality.
20 3.4.4 Temporality
21 Temporality refers to the occurrence of a cause before its purported effect, and is most
22 relevant to studies of diseases that develop over time. This factor is difficult to investigate in
23 situations where the pollutant concentrations are correlated over time as is the case to some
24 degree in time series studies and to a greater degree in cohort studies. The short-term exposure
25 studies evaluate associations between acute health outcomes and ozone measured on an hourly
26 basis. In many studies, associations have been reported between health events and ozone
27 measured contemporaneously. For example, in studies of total and cardiovascular mortality, the
28 draft CD observes that effects have been most clearly linked with ozone measured on the same
29 day or the preceding day (CD, p. 8-59): This would be expected for acute health outcomes;
30 however, it is difficult to characterize these associations in terms of temporality. Issues related
31 to the evaluation and selection of lag periods among studies are discussed in the next section.
32 3.4.5 Lag Structure in Short-term Exposure Studies
33 In the short-term exposure epidemiological studies, many investigators have tested
34 associations for a range of lag periods between the health outcome and ozone concentration (see
November 2005
3-37
Draft - Do Not Cite or Quote
-------
1 CD, sections 7,1,3.3 and 8.4.2.3). The draft CD observes that the selection of an appropriate lag
2 period can depend on the health outcome under study. For example, if cough is resulting from
3 irritant action of ozone, that would be expected to occur with a short lag time; however,
4 exacerbation of asthma through an inflammatory response might occur up to several days after
5 initial exposure CD, p. 7-9). For both mortality and respiratory hospital admissions, the draft CD
6 reports that most significant associations between ozone and mortality were observed with ozone
7 measured on the same day or a 1-day lag period in studies using individual lag periods (CD, p, 8-
8 59). In U.S. multi-city studies, larger effect estimate sizes were reported for the ozone-mortality
9 ' relationship with the distributed lag structure (CD, p. 7-80). Field studies of lung function or *
10 respiratory symptoms reported associations with ozone across a range of lag periods from the
11 exposure on the same day to exposures averaged over several days (CD, Sections 7.2.3 and
12 7.2.4). Cardiovascular effects appeared to be associated with ozone at shorter lag periods;
13 cardiovascular health outcomes such as changes in cardiac autonomic control were associated
14 with ozone measured on the same day (CD, section 7.2.7.1). In addition, Peters et al. (2001)
15 reported a positive but not statistically significant association between myocardial infarction
16 onset and ozone witii very short lag times of 1- to 4 hrs (CD, p. 7-55).
17 In focusing on an effect estimate reported for any individual lag period, the draft CD
18 observes that it is important to consider the pattern of results across the series of lag periods. If
19 there is an apparent pattern of results across the different lags, then selecting the single-day lag
20 with the largest effect from a series of positive associations is likely to underestimate the overall
21 effect size, since single-day lag effect estimates do not fully capture the risk that may be
22 distributed over adjacent or other days (CD, p. 7-10). However, if the reported effect estimates
23 vary substantially across lag periods, any result for a single day may well be biased (CD, p. 7- •
24 11). If the effect of ozone on health outcomes persists over several days, distributed lag model
25 results can provide more accurate effect estimates for quantitative assessment than an effect
26 estimate for a single lag period (CD, p. 7-10). Conversely, if the underlying ozone-health
27 relationship is truly an acute effect, then a distributed lag model would likely result in reduced
28 effect estimate size that may underestimate the effect (CD, p. 7-10).
29 On this basis, the draft CD focuses on effect estimates from models using 0- or 1-day lag
30 periods, with some consideration of multi-day lag effects (CD, p. 7-11). For quantitative
31 assessments, staff concludes that it is appropriate to use results from lag period analyses
32 consistent with those reported in the draft CD, focusing on single day lag periods of 0-1 days for
33 associations with mortality or respiratory hospitalization, depending on availability of results
34 (CD, p. 8-59). When available, distributed lag model results also have been used in the
35 quantitative risk assessment. However, for those few studies that show inconsistent patterns, the
36 use of single-day lag results is not appropriate for inclusion in the quantitative assessment.
November 2005 3-38 Draft - Do Not Cite or Quote
-------
1 3.4.6 Concentration-Response Relationships and Potential Thresholds
2 It has been recognized that it is reasonable to expect that there likely are biological
3 thresholds for different health effects in individuals or groups of individuals with similar innate
4 characteristics and health status. For ozone exposure, individual thresholds would presumably
5 vary substantially from person to person due to individual differences in genetic-level •
6 susceptibility, pre-existing disease conditions and possibly individual risk factors such as diet or
7 exercise levels (and could even vary from one time to another for a given person). Thus, it
8 would be difficult to detect a distinct threshold at the population level, below which no
9 individual would experience a given effect, especially if some members of a population are
10 unusually sensitive even down to very low concentrations. (U.S. EPA, 2004, p. 9-43, 9-44)
11 Some time-series epidemiological studies have used statistical modeling approaches to
12 evaluate whether thresholds exist in associations between short-term ozone exposure and
13 mortality. As discussed in section 7.6.5 of the draft CD, one European multi-city study included
14 evaluation of the shape of the concentration-response curve, and observed no deviation from a
15 linear function across the range of ozone measurements from the study (Gryparis et al., 2004;
16 CD p. 7-135). Several single-city studies also observed amonotonic increase in associations
17 between ozone and morbidity that suggest that no threshold exists (CD, p. 7-137).
18 On the other hand, a study in Korea used several different modeling approaches and
19 reported that a threshold model provided the best fit for the data. The results suggested a
20 potential threshold level of about 45 ppb (1-hr maximum concentration) for an association
21 between mortality and short-term ozone exposure during the summer months (Kim et al., 2004;
22 CD, p. 7-136). The authors reported larger effect estimates for the association for data above the
23 potential threshold level, suggesting that an ozone-mortality association might be underestimated
24 in the non-threshold model. In addition, Burnett and colleagues (1997) plotted the relationships
25 between air pollutant concentrations and both respiratory and cardiovascular hospitalization, and
26 it appears in these results that the associations with ozone are found in the concentration range
27 above about 30 ppb (1-hr maximum).
28 Other studies have tested associations between ozone and health outcomes after removal
29 of days with higher ozone levels from the data set; such analyses do not necessarily indicate the
30 presence or absence of a threshold, but provide some information on whether the relationship is
31 found using only lower-concentration data,. For example, using data from 95 U.S. cities, Bell et
32 al. (2004) found that the effect estimate for an association between short-term ozone exposure
33 and mortality was little changed when days exceeding 60 ppb (24-hr average) were excluded in
34 the analysis (CD, p. 7-135). Using data from 8 U.S. cities, Mortimer and colleagues (2002) also
35 reported that associations between ozone and both lung function and respiratory symptoms
36 remained statistically significant and of the same or greater magnitude in effect size when
November 2005
3-39
Draft - Do Not Cite or Quote
-------
1 concentrations greater than 80 ppb (8-hr average) were excluded (CD, p. 7-137). Several single-
2 city studies are also summarized in section 7.6.5 that report similar findings of associations that
3 remain or are increased in magnitude and statistical significance when data at the upper end of
4 the concentration range are removed.
5 Finally, Vedal and colleagues (2003) reported a significant association between ozone
6 and mortality in British Columbia where ozone concentrations were quite low (mean
7 concentration of 27.3 ppb). The authors did not specifically test for threshold levels, but the fact
8 that the association was found in an area with such low ozone concentrations suggests that any
9 potential threshold level would be quite low in this data set.
10 In summary, some epidemiological analyses have suggested that no threshold levels can
11 be found in associations between ozone and mortality or morbidity. In those studies that provide
12 suggestive evidence of thresholds, the potential thresholds are at low concentrations (CD, p. 7-
13 138). The draft CD finds that no definitive conclusion can be reached with regard to the
14 existence of thresholds in epidemiological studies (CD, p. 8-24). Staff recognizes, however, the
15 possibility that thresholds for individuals may exist in reported associations at fairly low levels
16 within the range of air quality observed in the studies, but not be detectable as population
17 thresholds in epidemiological analyses. Based on the draft CD's conclusions, staff judges that
18 there is insufficient evidence to support use of potential threshold levels in quantitative
19 assessments, but that data should be used within the range of air quality concentrations down to a
20 policy-relevant background level.
21 3.5 BIOLOGICAL PLAUSIBILITY AND COHERENCE OF EVIDENCE
22 This section summarizes material contained in section 8.5 of the CD, which integrates
23 epidemiological studies with mechanistic information from controlled human exposure studies
24 and animal toxicological studies to draw conclusions regarding the coherence of evidence and
25 biological plausibility of Oa-related health effects. For its assessment, the draft CD's discussion
26 draws from epidemiological evidence on a range of relevant health endpoints (from
27 cardiopulmonary and physiological changes to morbidity and mortality) and assessment of
28 available toxicological and biochemical evidence on potential plausible causal relationships for
29 the observed epidemiological associations (CD, p. 8-59).
30 Table 3-1 (Table 8-1, CD, p. 8-60) summarizes physiological and biochemical
31 observations which represent the knowledge base available from toxicological studies in humans
32 and animals that underlie biological alterations that cause acute Os-induced health effects. Table
33 3-1 was based upon experimental data (contained in CD Chapters 5 and 6, as well as the chapter
34 annexes),-which used environmentally relevant exposure regimens. Although most of the acute
35 O3-induced health effects are transient and attenuate over time, the time-line for resolution of
November 2005 3-40 Draft - Do Not Cite or Quote
-------
1 many of the physiological and biological parameters presented in Figure 3-1 (Figure 8-9, CD, p.
2 8-61) differ for healthy human subjects and those with underlying cardiopulmonary diseases.
3 The CD further notes that alterations in acute Oa-induced cellular and molecular changes
4 observed in human airway epithelium evolve over time, as depicted in Figure 3-2 (Figure 8-10,
5 CD, p. 8-62), and that the knowledge of this profile is important in assessing biological
6 plausibility to integrate across evidence of various health endpoints.
7 The similarities in physiological, biochemical and pathological processes between
8 humans and many animal species are due to the high level of genome sequence homology that
9 exists across species (CD, p. 8-62). It is this homology that supports the use of knowledge
10 gained on initiation, progression, and treatment regimes for disease processes across species,
11 especially on the acute Os-induced effects in the respiratory tracts of humans and various animal
12 species, as depicted in CD Table 3-1 and Figures 3-1 and 3-2. The similarities observed in
13 human and rat respiratory system effects (e.g., in spirometry, ventilatory response, host defense),
14 attenuation, and at higher levels of cellular of organization (e.g., neutrophilic inflammation,
15 macrophage phagocytosis processes) lend support to animal-to-hurnan extrapolation. This is
16 particularly important in collecting information that would not be possible to gather in human
17 exposure or epidemiological studies but may corroborate data from both types of studies. Since
18 quantitative extrapolation requires a combination of dosimetry, end point homology, and species
19 sensitivity and because uncertainties continue to exist, extrapolation models have not been
20 completely validated. However, one study (Hatch etal., 1994) of inflammatory markers
21 suggests mat a 2 ppm Oa exposure in sedentary rats approximates a 0.4 ppm exposure in
22 exercising humans, which supports the use of some animal data collected at higher Os exposures
23 to help understand molecular changes in acutely exposed humans. Also of great importance are
24 the chronic exposure studies (12 to 24 months) reporting lesions caused by long-term Oa
25 exposures, which may facilitate a more direct assessment of chronic health effects in humans
26 (CD, p. 8-63).
27 3.5.1 Coherence and Plausibility of Short-term Effects on the Respiratory
28 System
29 Acute respiratory morbidity effects that have been associated with short-term exposure to
30 Os include such health endpoints as decrements in lung function, bronchoconstriction, increased
31 airway responsiveness, airway inflammation, epithelial injury, and immune system effects,
32 emergency department visits for respiratory diseases, and hospitalization due to respiratory
33 illness.
34 Recent epidemiological studies have supported evidence available in the previous Oa
35 NAAQS review on associations between ambient Oa exposure and decline in lung function for
November 2005
3-41
Draft - Do Not Cite or Quote
-------
Table 3-1. Acute Oj-induced Physiological and Biochemical
Changes in Human and Animals
Physiological/biochemical
Alterations
Human Exposure Studies u
Animal Toxicology Studies3-4
Pulmonary Function:
Airway Responsiveness:
Inflammation:
ROS
Host Defense:
Lung injury:
Morphology
Susceptibility:
Cardiovascular Changes:
1FEV,
I Frequency of breathing
(rapid, shallow)
1 inspiratory capacity
(cough, breathing discomfort,
throat irritation, wheezing)
Mild bronchoconstriction
I (neuronal involvement)
change in lung resistance
Yes
T inflammatory mediators
I
t Particle clearance
I permeability
1 AM phagocytosis
Yes
Age,
Inter individual variability
Disease status
Polymorphism in certain genes
being recognized
Impairment in arterial O2 transfer
Ventilation-perfusion mismatch
(suggesting potential arterial
vasoconstriction)
I
1 Frequency of breathing
(Rapid, shallow )
J inspiratory capacity
t (vagal mediation)
Change in lung resistance
Yes
I inflammatory mediators
I
t Particle clearance
t permeability
1 clearance of bacteria
t severity of infection
I Mortality & morbidity
Yes
Species specific differences
Genetic basis for susceptibility
indicated
Heart rate variability (HRV)
i core body temperature
T ANF
Role for PAF indicated
increased pulmonary vascular
resistance
1 Controlled chamber exposure studies in human volunteers were carried out for a duration of 1-6.6 h with O}
concentration in the range of 0.08-0.4 ppm with intermittent exercise.
2 Data on some of the biochemical parameters were obtained from in vitro studies using cells recovered
from BALF.
3 Responses were observed in animal toxicology studies with exposure for a duration of 2-72h with O3
concentration in the range of 0.1- 2.0 ppm.
* Various species (mice, rat, guinea pigs and rabbit) and strains.
(Reproduced from Table 8-1, CD, p. 8-60).
November 2005
3-42
Draft - Do Not Cite or Quote
-------
Ol
fe
8
•o
-B
tw
E^
o ^
£ ^>
•s°°
•c a
S cT
-a o
i ^
^ 00
ii
S "o
§ £
81
«•*
1"S
kj
a
i
i
^
^
(J
It
II
11
p
Is
•*•• ^2
It
?6
«|
£ «
0> •-
= 1
Is
II
|r
i -
(2 2.
r^
£
I
E
en
1
-------
Figure 3-2. O3-induced cellular and molecular changes and their evolution depicted here
is derived from the data reported in Leikauf et al. (1995) and Mudway and
Kelly (2000). (Reproduced from Figure 8-10, CD, p. 8-62)
November 2005
3-44
Draft - Do Not Cite or Quote
-------
1 children. Earlier observations that children and asthmatic individuals are particularly susceptible
2 to ambient Os are supported by a meta-analysis (Kinney et al, 1996) of summer camp studies
3 and a multicity study (Mortimer et al., 2002). The draft CD concludes that exposure to ambient
4 Os has a significant effect on lung function, is associated with increased respiratory symptoms
5 and medication use, particularly in asthmatics.
6 Short-term exposure to Oa has also been associated with more severe morbidity
7 endpoints, such as emergency department visits and hospital admissions for respiratory cases,
8 including specific respiratory illness (e.g., asthma) (CD, sections 7.3.2 and 7.3.3). In addition, a
9 few epidemiological studies have reported positive associations between short-term Oj exposure
10 and respiratory mortality, though the associations are not generally statistically significant,
11 possibly due to a lack of statistical power for this mortality subcategory. (CD, p. 7-94).
12 Considering Hie evidence from epidemiological studies, the results described above
13 provide evidence for coherence in Os-related effects on the respiratory system. Effect estimates
14 from U.S. and Canadian studies are shown in Figure 3-3, where it can be seen that mostly
15 positive associations have been reported with respiratory effects ranging from respiratory
16 symptoms, such as cough or wheeze, to hospitalization for various respiratory diseases, and there
17 is suggestive evidence for associations with respiratory mortality. Many of the reported
18 associations are statistically significant.
19 Considering also evidence from toxicological studies, the CD (section 8.5.1) discusses
20 biological plausibility and coherence of evidence for acute Os-induced respiratory health effects.
21 Inhalation of O3 for several hrs while subjects are physically active can elicit both acute
22 pathophysiological changes and subjective respiratory tract symptoms (CD, section 8.4.2.4.1).
23 Acute pulmonary responses observed in healthy humans exposed to O3 at ambient concentrations
24 include: decreased inspiratory capacity; mild bronchoconstriction; rapid, shallow breathing
25 during exercise; subjective symptoms of tracheobronchial airway irritation, including cough and
26 pain on deep inspiration; decreases in FVC and FEVi; and increased SRaw The severity of
27 symptoms and magnitude of response depends on inhaled dose, individual O3 sensitivity, and the
28 degree of attenuation resulting from previous 03 exposures. Pulmonary function studies of
29 several animal species acutely exposed to relatively low O3 levels (0.25 to 0.4 ppm) show
30 responses similar to those observed in humans, including increased breathing frequency,
31 decreased tidal volume, increased resistance, and decreased FVC. Alterations in breathing
32 pattern return to normal within hrs of exposure, and attenuation in functional responses following
33 repeated O3 exposures is similar to those observed in humans.
November 2005
3-45
Draft - Do Not Cite or Quote
-------
U.S EPA Headquarters Library
Mail code 3404T
1200 Pennsylvania Avenue NW
Washington, DC 20460 .
202-566-0556
fc
i
en
1
-------
1 Physiological and biochemical alterations investigated in controlled human exposure and
2 animal toxicology studies tend to support certain hypotheses of underlying pathological
3 mechanisms which lead to the development of respiratory-related effects reported in
4 epidemiology studies (e.g., increased hospitalization and medication use). Some of these are:
5 (a) decrements in lung function, (b) bronchoconstriction, (c) increased airway responsiveness, (d)
6 airway inflammation, (e) epithelial injury, (f) immune system activation, (g) host defense
7 impairment, and sensitivity of individuals, such as age, genetic susceptibility, and the degree of
8 attenuation present due to prior exposures. The time sequence, magnitude, and overlap of these
9 complex events, both in terms of development and recovery (as depicted in Figures 3-1 and 3-2),
10 illustrate the inherent difficulty of interpreting the biological plausibility of Oa-induced
11 cardiopulmonary health effects (CD, p. 8-64).
12 The interaction of Os with airway epithelial cell membranes and ELF to form lipid
13 ozonation products and ROS is supported by numerous human, animal and in vitro studies.
14 Ozonation products and ROS initiate a cascade of events that lead to oxidative stress, injury,
15 inflammation, airway epithelial damage and increased epithelial damage and increased alveolar
16 permeability to vascular fluids. Repeated respiratory inflammation can lead to a chronic
17 inflammatory state with altered lung structure and lung function and may lead to chronic
18 respiratory diseases such as fibrosis and emphysema (CD, p. 8-65). Continued respiratory
19 inflammation also can alter the ability to respond to infectious agents, allergens and toxins.
20 Acute inflammatory responses to Os are well documented, and lung injury can become apparent
21 within 3 hr after exposure in humans. Initial inflammatory response phase is characterized by
22 increased PMNs in the BAL fluid and increased levels of inflammatory mediators (e.g.,
23 interleukins, prostaglandins, complement component C3a). Late inflammatory phase in the
24 lungs is characterized by increased levels of monocytes and eosinophils and mediators, such as
25 cytokines, leukotrienes, proteinases, and ROS. Ozone-induced lung injury and subsequent
26 disruption of the airway epithelial barrier has been implicated in increased mucociliary clearance
27 of particles in human subjects. Similarly, animal toxicology studies have shown that Os
28 exposure increased clearance of particles and increased mortality resulting from bacterial and
29 viral infections (CD, p. 8-65).
30 The draft CD concludes that the findings of Os-mduced lung function changes,
31 respiratory symptoms, inflammation, and permeability changes in animal toxicological studies
32 are consistent with the associations with respiratory health outcomes reported in epidemiological
33 studies. Taken together, there appears to be substantial evidence for coherence and biological
34 plausibility for effects of Os on the respiratory system.
November 2005
'
3-47
Draft - Do Not Cite or Quote
-------
1 3.5.2 Coherence and Plausibility of Effects on the Cardiovascular System
2 As described in sections 7.2.7 and 7.3.4, some new epidetniological studies have reported
3 associations between short-term Os exposure and effects on the cardiovascular system. In the
4 recent studies on incidence of myocardial infarction and some more subtle cardiovascular health
5 endpoints, such as changes in heart rate variability or cardiac arrhythmia, some but not all studies
6 reported associations with short-term exposure to Os (CD, section 7.2.7.1). From these studies,
7 the draft CD concludes that the "current evidence is rather limited but supportive of a potential
8 effect on [heart rate variability], ventricular arrhythmias, and the incidence of MI." (CD, p. 7-57)
9 Some studies have reported associations with hospitalization or emergency department
10 visits for cardiovascular diseases. As shown in Figure 7-13 of the draft CD, the results are
11 somewhat inconsistent, but a number of the associations are positive and statistically significant,
12 especially those conducted during the warm season. The draft CD finds that the evidence from
13 these studies is inconclusive, but the results from studies using warm season analyses suggest
14 potential effects of Os on the cardiovascular system (CD, p. 7-73).
15 Finally, several recent studies report associations between short-term 63 exposure and
16 mortality from cardiovascular or cardiopulmonary causes, and numerous studies have reported
17 statistically significant associations with total nonaccidental mortality, and cardiopulmonary
18 causes constitute the largest category. Based on results of multi-city analyses, meta-analyses and
19 numerous single-city analyses, the draft CD concludes that there is strong evidence for
20 associations between short-term 63 exposure and mortality (CD, p. 7-74). As shown in Figure 7-
21 22 of the draft CD, all associations reported in these studies are positive, and many are
22 statistically significant, particularly the associations reported from warm season analyses (CD, p.
23 7-92).
24 There are very few human or animal experimental studies that have investigated potential
25 cardiovascular effects of acute Oj exposure, in contrast to the large data base of studies
26 indicating that 03 exposure induces lung injury, inflammation, impaired mucociliary clearance,
27 and increased epithelial permeability. The draft CD observes that generation of lipid ozonation
28 products and ROS in lung tissue can influence pulmonary hemodynamics and the cardiovascular
29 system (CD, section 5.4.2). Recent reports of interaction between Os and cholesterol in the lung
30 surfactant and the generation of highly reactive products, such as oxysterols and B-epoxide,
31 indicate a role for cardiovascular effects and atherosclerosis (CD, p. 8-66). Changes in heart rate
32 variability, heart tissue edema, and increased tissue and serum levels of ANF have been reported
33 in experimental animals and support the potential for cardiovascular effects induced by acute Os
34 exposures. Such effects are the result of stimulation of airway irritant receptors and C-fiber
35 activation and may result from either local or CNS involvement. Ozone-induced changes in
36 alveolar-arterial oxygen transfer observed in controlled exposure human studies of subjects with
November 2005 3-48 Draft - Do Not Cite or Quote
-------
1 hypertension suggest potential complex ANF effects that require further investigation (CD, p. 8-
2 66).
3 While still limited, the evidence from recent epidemiological studies suggests some
4 coherence in effects of short-term Os exposure on the cardiovascular system, from subtle
v t
5 changes in physiological cardiac responses to hospitalization for cardiovascular disease to
6 cardiopulmonary mortality. The draft CD also describes toxicological evidence that provides
7 some potential plausible mechanisms or pathways for effects on the cardiovascular system.
8 3.5.3 Coherence and Plausibility of Effects Related to Long-Term O3 Exposure
9 As discussed in section 8.5.2 of the draft CD, previous epidemiological studies have
10 provided only inconclusive evidence for either mortality or morbidity effects of long-term Oa
11 exposure. The draft CD observes mat the inconsistency in findings may be due to a lack of
12 precise exposure information, the possibility of selection bias, and the difficulty of controlling
13 for confounders (CD, p. 8-69). Several new longitudinal epidemiology studies hav.e evaluated
14 associations between long-term OB exposures and morbidity and mortality and suggest that these
15 long-term exposures may be related to changes in lung function, increased incidence in asthma,.
16 mortality, and possibly lung cancer (CD, section 7.5). Considering these new findings, the draft
17 CD concludes: "The strongest evidence is for negative seasonal effects of Os on lung function in
18 adults and children. Less conclusive are longer-term studies investigating the association of
19 chronic Os exposure on yearly lung function, asthma incidence, and respiratory symptoms.
20 Chronic Os-mortality studies observed inconsistencies across exposure perio'ds, cause-specific
21 mortality outcomes, and gender." (CD, p. 7-150) Thus, the epidemiological evidence suggests
22 some potential effects on the respiratory system, but provides no evidence for associations
23 between long-term Os exposure and mortality.
24 Human chamber studies have not evaluated effects with long-term exposures to Os, but
25 some evidence is available from toxicological studies. While early animal toxicology studies of
26 long-term Os exposures were conducted using continuous exposures, more recent studies have
27 focused on exposures which mimic diurnal and seasonal patterns and more realistic Os exposure
28 levels (CD, p. 8-69). Studies of monkeys that compared these two exposure scenarios found
29 increased airway pathology only'with the latter. A long-term study of rhesus monkeys exposed
30 to simulated seasonal Os (0.5 ppm, 8 hr/day for 5 days every 14 days for 11 episodes) reported
31 remodeling of the distal airways, abnormalities in trachea! basement membrane, accumulation of
32 eosinophils in conducting airways, and decrements in airway innervation. Another seasonal
33 exposure study of monkeys exposed (0.61 ppm Os) for a year reported increased deposition of
34 collagen and thickening of the CAR of the deep lung. Long-term seasonal Os exposure studies of
35 rats also have provided evidence of biochemical and morphological changes suggestive of
November 2005
3-49
Draft - Do Not Cite or Quote
-------
1 irreversible chronic damage to the lungs. One of these studies exposed rats for 20 months (0.5-
2 1.0 ppm Os for 6 hr/day) and found increased deposition of collagen and thickening of the CAR
3 of the deep lung. Although some earlier seasonal exposure studies of rats reported small, but
4 significant, decrements in lung function consistent with focal fibrogenesis in the proximal
5 alveolar region, other chronic exposure studies with exposures of 0.5 to 1.0 ppm (^report
6 epithelial hyperplasia that disappears in a few days.
7 The draft CD concludes that the totality of evidence from animal toxicology studies
8 strongly suggests that Ps exposure can damage the distal airways and proximal alveoli, resulting
9 in lung tissue remodeling thus leading to apparent irreversible damage. Such structural changes
10 and compromised pulmonary function caused by persistent inflammation may exacerbate the
11 progression and development of chronic lung disease (CD, p. 8-70). Together with the limited
12 evidence available from epidemiological studies, these studies suggest coherence and plausibility
13 for associations between long-term exposure to Oa and adverse effects on the respiratory system;
14 however, as noted previously, there is not clear evidence linking long-term exposure to Os and
15 mortality.
16 3.5.4 Coherence and Plausibility of Mortality-Related Health Endpoints
17 An extensive epidemiological literature on air pollution related mortality risk estimates
18 from the U.S., Canada, and Europe is discussed in the draft CD (section 7-4). These mortality
19 studies suggest a pattern of effects for causality that has a biologically plausible explanation, but
20 our knowledge regarding potential underlying mechanisms is very limited at this time and
21 requires further research. Most of the physiological and biochemical parameters investigated in
22 human and animal studies suggest that (^-induced biochemical effects are relatively transient
23 and attenuate over time. A generic pathway of Os-induced lung damage, potentially involving
24 oxidative lung damage with subsequent inflammation and/or decline in lung function leading to
25 respiratory distress is hypothesized in the CD (p. 8-70).
26 The third National Health and Nutrition Examination Followup data analysis indicates
27 mat about 20% of the adult population has reduced FEVi values, suggesting impaired lung
28 function. Most of these individuals have COPD, asthma or fibrotic lung disease (Manino et al.,
29 2003), which are associated with persistent low-grade inflammation. Furthermore, patients with
30 COPD are at increased risk for cardiovascular disease, and lung disease with underlying
31 inflammation may be linked to low-grade systemic inflammation associated with atherosclerosis,
32 independent of cigarette smoking (CD, p. 8-71). Lung function decrements in persons with
33 cardiopuhnonary disease have been associated with inflammatory markers, such as C-reactive
34 protein (CRP) in the blood. At a population level it has been found that individuals with the
35 lowest FEVi values have the highest levels of CRP, and those with the highest FEVi values have
November 2005 • 3-50 Draft - Do Not Cite or Quote
-------
1 the lowest CRP levels (Manino et al., 2003; Sin and Man, 2003). Although this complex series
2 of physiological and biochemical reactions following Os exposure could lead to adverse health
3 effects in those with cardiopulmonary disease, there is no experimental data available to support
4 such a hypothesis at this time to explain the cardiovascular mortality, evaluated in the
5 epidemiological studies. Future reevaluation may help in understanding the biological
6 plausibility of changes in CRP in the context of Os and other air pollution related mortality (CD,
7 section 8.5.3).
8 3.6 OZONE-RELATED IMPACTS ON PUBLIC HEALTH
9 The following discussion draws from section 8.6 of the draft CD to characterize factors
10 which modify responsiveness to O^t subpopulations potentially at risk for Os-related health
11 effects, and potential public health impacts associated with exposure to ambient Og. Providing
12 appropriate protection of public health requires that a distinction be made between those health
13 effects that are considered adverse and those that are not adverse. What constitutes an adverse
14 health effect depends not only on the type and magnitude of effect but also on the population
15 group being affected. While some changes in healthy individuals would not be considered
16 adverse, similar changes in susceptible individuals would be seen as adverse. In order to
17 estimate the potential public health impact, it is important to consider both the susceptible
18 subpopulations for 03 exposure and the definition of adversity for Os health effects.
19 3.6.1 Factors which Modify Responsiveness to Ozone
20 There are numerous factors which can modify individual responsiveness to 63. These
21 include: influence of physical activity; age; gender and hormonal influences; racial, ethnic and
22 socioeconomic status (SES) factors; environmental factors; and oxidant-antioxidant balance.
23 These factors are discussed in more detail in section 6.5 of the CD.
24 It is well established that physical activity increases an individual's minute ventilation
25 and will thus increase the dose of Os inhaled (CD, section 6.5.4). Increased physical activity
26 results in deeper penetration of Os into more peripheral regions of the lungs, which are more
27 sensitive to acute Os response and injury. This will result in greater lung function decrements for
28 acute exposures of individuals during increased physical activity. Research has shown that
29 respiratory effects are observed at lower O$ concentrations if the level of exertion is increased
30 and/or duration of exposure and exertion are extended. Predicted O3-induced decrements in lung
31 function have been shown to be a function of exposure duration and exercise level for healthy,
32 young adults (McDonnell etal.51997)
33 Most of the studies investigating the influence of age have used lung function decrements
34 and symptoms as measures of response. After 18 to 20 years of age, lung function and symptom
November 2005
3-51
Draft - Do Not Cite or Quote
-------
1 responses to Os decline as age increases. The rate of decline in Os responsiveness appears
2 greater in those 18 to 35 years old compared to those 35 to 55 years old, while there is very little
3 change after age 55. In one study (Seal et al., 1996) analyzing a large data set, a 5.4% decrement
4 in FEVi was predicted for 20 year old individuals exposed to 0.12 ppm Os, whereas similar
5 exposure of 35 year old individuals were predicted to have a 2.6% decrement. While healthy
6 children tend not to report respiratory symptoms when exposed to low levels of Os, for subjects
7 18 to 36 years old symptom responses induced by Os tend to decrease with increasing age
8 (McDonnell etal., 1999).
9 Limited evidence of gender differences in response to Os exposure has suggested that
10 females may be predisposed to a greater susceptibility to Os. Lower plasma and NL fluid levels
11 of the most prevalent antioxidant, uric acid, in females relative to males may be a contributing
12 factor (Housley et al., 1996). Consequently, reduced removal of Os in the upper airways may
13 promote deeper penetration. However, most of the evidence on gender differences appears to be
14 equivocal, with one study (Hazucha et al., 2003) suggesting that physiological responses of
15 young healthy males and females may be comparable (CD, section 6.5.2).
16 A few studies have suggested that ethnic minorities might be more responsive to Os than
17 Caucasian population groups (CD, section 6.5.3). This may be more the result of a lack of
18 adequate health care and socioeconomic status than any differences in sensitivity to Os. The
19 ' limited data available, which have investigated the influence of race, ethnic or other related
20 factors on responsiveness to Os, prevent drawing any clear conclusions at this time.
21 Few human studies have examined the potential influence of environmental factors such
22 as the sensitivity of individuals who voluntarily smoke tobacco (i.e., smokers) and the effect of
23 high temperatures. New controlled human exposure studies have confirmed that smokers are
24 less responsive to Os than nonsmokers; however, time course of development and recovery of
25 these effects, as well as reproducibility, was not different from nonsmokers (CD, section 6.5.5).
26 Influence of ambient temperature on pulmonary effects induced by Os has been studied very
27 little, but additive effects of heat and Os exposure have been reported.
28 Antioxidants, which scavenge free radicals and limit lipid peroxidation present in the
29 ELF, are the first line of defense against oxidative stress. Ozone exposure depletes the
30 antioxidant level in the nasal ELF by scrubbing of Os, but concentration and antioxidant enzyme
31 activity in ELF or plasma don't appear related to Os responsiveness (CD, section 6.5.6).
32 Controlled studies of the protective effects of dietary antioxidant supplements have shown some
33 protective effects of lung function but not of subjective symptoms or inflammatory response.
34 Dietary antioxidant supplements have provided some protection to asthmatics by attenuating
35 post-exposure airway hyperresponsiveness. Animal studies have also supported the protective
36 effects of ELF antioxidants.
November 2005 3-52 Draft - Do Not Cite or Quote
-------
1 3.6.2 , Susceptible Population Groups
2 Several characteristics that may increase the extent to which a population group shows
3 sensitivity to Os have been discussed in the draft CD, in the sections on clinical studies in
4 Chapter 6, epidemiological studies in Chapter 7, and in the integrated assessment in Chapter 8;
5 this section will draw on all of these. The characteristics that likely increase susceptibility to 63
6 are based on: (1) activity patterns; (2) lung disease; (3) age; and (4) biological responsiveness to
7 Cb. Other groups that might have enhanced sensitivity to Oj, but for which there is currently
8 very little evidence, include: people with heart disease; groups based on race, gender and
9 socioeconomic status; and those with nutritional deficiencies.
10 , 3.6.2.1 Active People
11 A large group of individuals at risk from Os exposure consists of children, adolescents,
12 and adults who engage in outdoor activities involving exertion or exercise during summer
13 daylight hours when ambient Os concentrations tend to be higher. This conclusion is based on a
14 large number of controlled-exposure human studies which have been conducted with healthy
15 children and adults and those with preexisting respiratory diseases (CD, section 6.2 and 6.3).
16 These studies show a clear 63 exposure-response relationship with increasing spirometric and
17 symptomatic response as exercise level increases. Furthermore, Os-induced response increases
18 as time of exposure increases. Although these studies show a wide variability of response and
19 sensitivity among subjects and the factors contributing to this variability continue to be
20 incompletely understood, as discussed below, the effect of increased exertion is consistent.
21 3.6.2.2 People with Lung Disease
22 People with preexisting pulmonary disease may be at increased risk from Os exposure.
23 Altered physiological, morphological and biochemical states typical of respiratory diseases like
24 aslhma, COPD and chronic bronchitis may render people sensitive to additional oxidative burden
25 induced by Os exposure. Evidence contained in the 1996 CD indicated that people with asthma
26 appear to be at least as sensitive, or more, to the acute effects of Os as healthy nonasthmatic
27 individuals. The new results reviewed in Chapters 6 and 7 of the draft CD from controlled .
28 human exposure, field, and epidemiological studies continue to suggest that people with aslhma
29 are a potentially sensitive subpopulation for Os health effects. This information expands the
30 understanding of the physiological basis for increased sensitivity in people with asthma. It
31 indicates that there are slightly increased spirometric responses in mild asthmatics compared to
32 healthy subjects. Small airways tend to be more affected in people with asthma than in healthy
33 subjects (Alexis et al., 2000), Oa-induced FEVi decrements tended to be greater in people with
34 asthma and allergic rhinitis compared to healthy controls (J6rres et al., 1996), and Oa-induced
35 increases in sRaW were reported to be greater in people with asthma compared to healthy subjects
36 (Aris et al., 1995; Balmes et al., 1996). In a study (Horstman et al., 1995) of people with mild-
37 to-moderate asthma exposed for longer durations (6.6 hr), FEVi and FEF25-75 decrements were
November 2005
3-53
Draft - Do Not Cite or Quote
-------
1 found to be significantly greater than for healthy subjects. Moreover, a significant positive
2 correlation reported in people with asthma between Os-induced spirometric responses and
3 baseline lung function suggested that responses increased with severity of disease. Ozone-
4 induced inflammatory responses also have been reported to be greater in people with asthma as
5 compared to healthy individuals (Peden et al., 1995; Scannell et al., 1996; Holz et al. 1999).
6 New evidence also indicates that people with asthma may have increased occurrence and
7 duration of nonspecific airway responsiveness, and that people with pre-existing allergic asthma
8 may have increased airway responsiveness to allergens following Os exposure (CD, p. 8-29).
9 A number of epidemiological studies have been conducted using asthmatic study
10 populations. The majority of epidemiological panel studies that evaluated respiratory symptoms
11 and medication use related to Os exposures focused on children (CD, p. 8-44). These studies
12 suggest that Os exposure may be associated with increased respiratory symptoms and medication
13 use in children with asthma. Other reported effects include respiratory symptoms, lung function
14 decrements, and emergency department visits, as discussed in the draft CD (section 7.6.7.1).
15 Strong evidence from a large multi-city study (Mortimer et al., 2002), along with support from
16 several single-city studies suggest that Os exposure may be associated with increased respiratory
17 symptoms and medication use in children with asthma (CD, p. 8-44). With regard to ambient Os
18 levels and increased hospital admissions and emergency department visits for asthma and other
19 respiratory causes, strong evidence establishes a correlation between Os exposure and increased
20 exacerbations of preexisting respiratory disease for 1-hr maximum Os concentrations <0.12 ppm
21 (CD, p. 8-46). Several hospital admission and emergency department visit studies in the U.S.
22 (Peel et al., 2005), Canada (Burnett et al., 1997a; Anderson et al., 1997), and Europe (Anderson
23 et al., 1997) have reported positive associations between increase in Os and increased risk of
24 emergency department visits and hospital admissions.
25 3.6.2.3 Children and Older Adults
26 Supporting evidence exists for heterogeneity in the effects of 03 by age. As discussed in
27 section 6.5.1 of the draft CD, children, adolescents, and young adults (<18 yrs of age) appear, on
28 average, to have nearly equivalent spirometric responses to Os, but have greater responses than
29 middle-aged and older adults when exposed to comparable Os doses. Symptomatic responses to
30 Os exposure, however, appear to increase with age until early adulthood and then gradually
31 decrease with increasing age. In contrast to young adults, the diminished symptomatic responses
32 in children and symptomatic and spirometric responses in the elderly may put them at an
33 increased risk for continued exposure.
34 As described in the section 7.6.7.2 of the draft CD, many epidemiological field studies
35 focused on the effect of Os on the respiratory health of school children. In general, children
36 experienced decrements in pulmonary function parameters, including PEF, FEVj, and FVC.
37 Increases in respiratory symptoms and asthma medication use were also observed in asthmatic
38 children. In one German study, children with and without asthma were found to be particularly
November 2005 3-54 Draft - Do Not Cite or Quote
-------
1 susceptible to 63 effects on lung function. Approximately 20% of the children, both with and
2 without asthma, experienced a greater than 10% change in FEVi, compared to only 5% of the
3 elderly population and athletes (Hoppe et al., 2003).
4 The American Academy of Pediatrics (2004) notes that children and infants are among
.5 the population groups most susceptible to many air pollutants, including Os. This is in part
6 because their lungs are still developing. For example, eighty percent of alveoli are formed after
7 birth, and changes in lung development continue through adolescence (Dietert et al., 2000).
8 Children are also likely to spend more time outdoors than adults do, which results in increased
9 exposure to air pollutants (Wiley et al., 1991a,b). Moreover, children have high minute
10 ventilation rates and high levels of physical activity which also increases their dose (Plunkett et
11 al., 1992).
12 Several mortality studies have investigated age-related differences in Os effects. Among
13 the studies that observed positive associations between Os and mortality, a comparison of all age
14 or younger age (^65 years of age) Os-mortality risk estimates to that of the elderly .population
15 (>65 years) indicates that, in general, the elderly population is more susceptible to p3 effects
16 (Borja-Aburto et al. 1997; Bremner et al., 1999; Gouveia and Fletcher 2000b; O'Neill et al.,
17 2004; Simpson et al., 1997; Sartor et al., 1995; Sunyer et al., 2002). For example, a study by
18 Gouveia and Fletcher (2000b) examined the Os-mortality effect by age in Sao Paulo, Brazil.
19 Among all ages, Os was associated with a 0.6% excess risk in all cause mortality per 40 ppb
20 increase in 1-hr max 03. In comparison, in the elderly population, the Os-mortality risk estimate
21 was nearly threefold greater, at 1.7%. Similarly, a Mexico City study found mat Oa-mortalily
22 risk estimates were 1.3% and 2.8% per 20 ppb increase in 24-hr average Os concentration in all
23 ages and the elderly, respectively (O'Neill et al., 2004).
24 The meta-analysis by Bell et al. (2005) found a larger effect estimate for the elderly
25 (2.92% per 20 ppb increase in 24-hr average O3) man for all ages (1.75%). In the large U.S. 95
26 communities study (Bell et al., 2004), effect estimates were slightly higher for those aged 65 to
27 74 years, 1.40% excess risk per 20 ppb increase in 24-hr average Oa, compared to individuals
28 less than 65 years and 75 years or greater, 1.00% and 1.04%, respectively, using a constrained
29 distributed 7-day lag model. Bell et al. (2004) note that despite similar effects estimates, the
30 absolute effect of Oa is substantially greater in the elderly population due to the higher
31 underlying mortality rates, which lead to a larger number of extra deaths for the elderly
32 compared to the general population. The draft CD concludes that the elderly population (>65
33 years of age) appear to be at greater risk of Os-related mortality and hospitalizations compared to
34 all age or younger populations (CD, p. 7-151).
35 The draft CD notes that, collectively, there is supporting evidence of age-related
36 differences in susceptibility to Os health effects. The elderly population (>65 years of age)
37 appear to be at increased risk of Oj-related mortality and hospitalizations, and children (<18
38 years of age) experience other potentially adverse respiratory health outcomes with increased Os
39 exposure (CD, section 7.6.7.2). •
November 2005
3-55
Draft - Do Not Cite or Quote
-------
1 3.6.2.4 People with Increased Responsiveness to Ozone
2 New animal toxicology studies using various strains of mice and rats have identified Os-
3 sensitive and resistant strains and illustrated the importance of genetic background in
4 determining O3 susceptibility. Using subacute low exposure regimen (0.3 ppm Os, 48h) studies
5 on inbred strains that have been designated as inflammation prone or resistant, Kleeberger et al.,
6 (1997) identified the pro-inflammatory cytokine gene, Tnf-most of these
28 studies were carried out using relatively high doses of Os, making the relevance of these studies
29 questionable in human health effects assessment. No doubt, the molecular parameters identified
30 in these studies may serve as useful biomarkers with the availability of suitable technologies and,
31 ultimately, can likely be integrated with epidemiological studies. Interindividual differences in
32 Os responsiveness have been observed across a spectrum of symptoms and lung function
33 responses do not yet allow identification of important underlying factors, except a significant
34 role for age.
November 2005
3-56
Draft - Do Not Cite or Quote
-------
1 3.6.2.5 Other Population Groups
2 There is limited, new evidence supporting associations between short-term Os exposures
3 and a range of effects on the cardiovascular system. Some but not all, epidemiological studies
4 have reported associations between short-term 63 exposures and the incidence of myocardial
5 infarction and more subtle cardiovascular health endpoints, such as changes in heart rate
6 variability and cardiac arrhythmia. Others have reported associations with hospitalization or
7 emergency department visits for cardiovascular diseases, although the results across the studies
8 are not consistent. Studies also report associations between short-term 63 exposure and
9 mortality from cardiovascular or cardiopulmonary causes. Based on epidemiological study
10 results, the draft CD concludes that the current evidence from field studies is rather limited but
11 supportive of a potential effect of short-term 03 exposure and heart rate variability, cardiac
12 arrhythmia and incidence of myocardial infarction (CD, p. 7-57). In the draft CD's evaluation of
13 studies of hospital admissions for cardiovascular disease (CD, section 7.3.4), it is concluded that
14 evidence from this growing group of studies is generally inconsistent but is suggestive of an
15 association with O3 in studies conducted during the warm season (CD, p. 7-73,7-74). This body
16 of evidence suggests that people with heart disease may be at increased risk from short-term
17 exposures to 63; however, more evidence is needed to conclude that people with heart disease
18 are a susceptible population.
19 Other groups that might have enhanced sensitivity to Os, but for which there is currently
20 very little evidence, include, groups based on race, gender and socioeconomic status, and those
21 with nutritional deficiencies, as discussed in section about factors which modify responsiveness
22 to Os, above.
23 3.6.3 What Constitutes an Adverse Health Impact from Ozone Exposure?
24 . In making judgments as to when various Os-related effects become significant enough
25 that they should be regarded as adverse to the health of individuals, in previous NAAQS reviews
26 staff has relied upon the guidelines published by the American Thoracic Society (ATS) and the
27 advice of CAS AC. While recognizing that perceptions of "medical significance" and "normal
28 activity" may differ among physicians, lung physiologists and experimental subjects, the ATS
29 (1985) defined adverse respiratory health effects "medically significant physiologic changes
30 generally evidenced by one or more of the following: (1) interference with the normal activity of
31 the affected person or persons, (2) episodic respiratory illness, (3) incapacitating illness, (4) ,
32 permanent respiratory injury, and/or (5) progressive respiratory dysfunction."
33 During the 1997 review, it was concluded that there was evidence of causal associations
34 from controlled human exposure studies for effects in the first category, evidence of statistically
35 significant associations from epidemiological studies for effects in the second and third
36 categories, and evidence from animal toxicology studies, which could be extrapolated to humans
37 only with a significant degree of uncertainty, for the last two categories. For the current review
November 2005
3-57
Draft - Do Not Cite or Quote
-------
1 the evidence is much stronger across the categories. For ethical reasons, clear causal evidence
2 from controlled human exposure studies still covers only effects in the first category. However,
3 for this review there are results from epidemiolpgical studies, upon which to base judgments
4 about adversity, for effects in all of the categories. Statistically significant and robust
5 associations have been reported in epidemiology studies fall into the second and third categories.
6 These more serious effects include respiratory illness that may require medication (e.g., asthma),
7 but not necessarily hospitalization,-as well as respiratory hospital admissions. Less conclusive,
8 but still positive associations have been reported for school absences, emergency room visits for
9 respiratory causes, and cardiovascular hospital admissions. Human health effects for which
10 associations have been suggested through evidence from epidemiological and animal toxicology
11 studies, but have not been conclusively demonstrated still fall primarily into the last two
12 categories, but the evidence is much stronger in this review than in the 1997 review. In the last
13 review of the 63 standard, evidence for these more serious effects came from studies of effects in
14 laboratory animals, and could be extrapolated to humans only with a significant degree of
15 uncertainty. Evidence from animal studies evaluated in this draft CD strongly suggests that Os is
16 capable of damaging the distal airways and proximal alveoli, resulting in lung tissue remodeling
17 leading to apparently irreversible changes. Recent advancements of dosimetry modeling also
18 provide a better basis for extrapolation from animals to humans. Information from
19 epidemiological studies provides supporting, but limited evidence of irreversible respiratory
20 effects in humans (as described in section 6.3.3.2 below). Moreover, the draft CD concludes that
21 the findings from multi-city times series, single city, and meta-analyses epidemiology studies
22 support a causal association between short-term Oa exposure and mortality particularly in the
23 warm season.
24 While Os has been associated with effects that are clearly adverse, application of these
25 guidelines, in particular to the least serious category of effects related to ambient Os exposures,
26 involves judgments about which medical experts on the CASAC panel and public commenters
27 have in the past expressed diverse views. -To help frame such judgments, staff defined
28 gradations of individual functional responses (e.g., decrements FEVi and airway responsiveness)
29 and symptomatic responses (e.g., cough, chest pain, wheeze), together with judgments as to the
30 potential impact on individuals experiencing varying degrees of severity of these responses, that
31 have been used in previous NAAQS reviews. These gradations and impacts are summarized
32 below in Tables 3-2 and 3-3 below.
3 3 For active healthy people, it has been j udged that moderate levels of functional responses
34 (e.g., FEVi decrements of >10% but < 20%, lasting up to 24 hrs) and/or moderate symptomatic
35 responses (e.g., frequent spontaneous cough, marked discomfort on exercise or deep breath, .
36 lasting up to 24 hrs) would likely interfere with normal activity for relatively few sensitive
November 2005 3-58 Draft - Do Not Cite or Quote
-------
1 individuals; whereas large functional responses (e.g., FEVi decrements > 20%, lasting longer
2 than 24 hrs) and/or severe symptomatic responses (e.g., persistent uncontrollable cough, severe
3 discomfort on exercise or deep breath, lasting longer than 24 hrs) would likely interfere with
4 normal activities for many sensitive individuals and therefore would be considered adverse under
5 ATS guidelines. However, for people with lung disease, even moderate functional (e.g., FEVi
6 decrements > 10% but < 20%, lasting up to 24 hrs) or symptomatic responses (e.g., frequent
7 spontaneous cough, marked discomfort on exercise or with.deep breath, wheeze accompanied by
8 shortness of breath, lasting up to 24 hrs) would likely interfere with normal activity for many
9 individuals, and would likely result in additional and more frequent use of medication. For
10 people wilh lung disease, large functional responses (e.g., FEVi decrements > 20%, lasting
11 longer than 24 hrs) and/or severe symptomatic responses (e.g., persistent uncontrollable cough,
12 severe discomfort on exercise or deep breath, persistent wheeze accompanied by shortness of
13 breath, lasting longer than 24 hrs) would likely interfere with normal activity for most
14 individuals and would increase the likelihood that these individuals would seek medical
15 .treatment or go to an emergency room for relief.
16 In judging the extent to which these impacts represent effects that should be regarded as
17 adverse to the health status of individuals, an additional factor that has been considered in
18 previous NAAQS reviews is whether such effects are experienced repeatedly during the course
19 of ayear or only on a single occasion. While some experts would judge single occurrences of
20 moderate responses to be a "nuisance," especially for healthy individuals, a more general
21 consensus view of the adversity of such moderate responses emerges as the frequency of
22 occurrence increases. Thus it has been judged that repeated occurrences of moderate responses,
23 even in otherwise healthy individuals, may be considered to be adverse since they could well set
24 the stage for more serious illness (61 FR 65723). The CAS AC panel.in the last review expressed
25 a consensus view that these "criteria for the determination of an adverse physiological response
26 was reasonable" (Wolff, 1995b).
27 In 2000, the American Thoracic Society (ATS) published an official statement on "What
28 Constitutes an Adverse Health Effect of Air Pollution?" (ATS, 2000), which updated the
29 guidance for defining adverse respiratory health effects (ATS, 1985) that was published fifteen
30 years earlier and has been used in past NAAQS reviews. The revised guidance was intended to
31 address new investigative approaches used to identify the effects of air pollution, and to reflect
32 the concern for the impacts of air pollution on specific groups that had been expressed through
33 the environmental justice movement.
34 The new guidance builds upon and expands the 1985 definition of adversity in several
35 ways. There is an increased focus on quality of life measures as indicators of adversity. There is
November 2005
3-59
Draft - Do Not Cite or Quote
-------
Table 3-2. Gradation of Individual Responses to Short-Term Ozone Exposure in Healthy
Persons2
Functional
Response
FEVi
Nonspecific
bronchial
responsivenessS
Duration of response
Symptom Response
Cough
Chest pain
Duration of response
Impact of Responses
Interference with normal
activity
None
Within
normal
range (±3%)
Within
normal range
None
Normal
Infrequent
cough
None
None
Normal
None
Small
Decrements of
3 to slO%
Increases of
<100%
<4hrs
Mild
Cough with deep
breath
Discomfort just
noticeable on
exercise or
deep breath
<4hrs
Normal
None
Moderate
Decrements of
>10but<20%
Increases of
*300%
>4 hrs but
*24hrs
Moderate
Frequent
spontaneous cough
Marked discomfort
on exercise or deep
breath
>4 hrs but £24 hrs
Mild
A few sensitive
individuals choose
to limit activity
Large
Decrements of
a20%
Increases of
>300%
>24 hrs
Severe
Persistent
uncontrollable
cough
Severe discomfort
on exercise or
deep breath
>24 hrs
Moderate
Many sensitive
individuals
choose to limit
activity
2 This table is reproduced from the 1996 O3 AQCD (Table 9-1, page 9-24) (U.S. Environmental Protection
Agency, 1996). ,
3 An increase in nonspecific bronchial responsiveness of 100% is equivalent to a 50% decrease in PDjo or PDjoo-
November 2005
3-60
Draft - Do Not Cite or Quote
-------
Table 3-3. Gradation of Individual Responses to Short-Term Ozone Exposure in Persons
with Impaired Respiratory Systems
Functional
Response
FEVt change
Nonspecific
bronchial
responsiveness 4
Airway resistance
(SRaw)
Duration of
response
Symptom
Response
Wheeze
Cough
Chest pain
Duration of
response
Impact of
Responses
Interference with
normal activity
Medical treatment
None
Decrements of
<3%
Within normal
range
Within normal
range (±20%)
None
Normal
None
Infrequent
cough '
None
None
Normal
None
No change
Small
Decrements of
3 to slO%
Increases of <1 00%
•* ^
SRaw increased
<100%
<4hr
Mild
With otherwise
normal breathing
Cough with deep
breath
Discomfort just
noticeable on exercise
or deep breath
<4hr
Mild
Few individuals
choose to limit
activity
Normal medication as
needed
Moderate
Decrements of >10
but <20%
Increases of s300%
SRaw increased up to
200% or up to 15cm
H2O/S
>4hrbut<:24hr
Moderate
With shortness of
breath
Frequent spontaneous
cough
Marked discomfort on
exercise or deep
breath
>4hrbut<24hr
Moderate
Many individuals
choose to limit
activity
Increased frequency
of medication use or
additional medication
Large
Decrements of
i20%
Increases of >300%
SRaw increased
>200% or more than
15cmH2O/s
>24hr
Severe
Persistent with
shortness of breath
Persistent
uncontrollable
cough
Severe discomfort
on exercise or deep
breath
>24 hr
Severe
Most individuals
choose to limit
activity
Physician or
emergency room
visit
An increase in nonspecific bronchial responsiveness of 100% is equivalent to a 50% decrease in PD2o or PDn».
November 2005
3-61
Draft - Do Not Cite or Quote
-------
1 also a more specific consideration of population risk. Exposure to air pollution that increases the
2 risk of an adverse effect to the entire population is adverse, even though it may not increase the
3 risk of any individual to an unacceptable level. For example, a population of asthmatics could
4 have a distribution of lung function such that no individual has a level associated with significant
5 impairment. Exposure to air pollution could shift the distribution to lower levels that still do not
6 bring any individual to a level that is associated with clinically relevant effects. However, this
7 would be considered to be adverse because individuals within the population would have
8 diminished reserve function, and therefore would be at increased risk if affected by another
9 agent
10 Of the various effects of 63 exposure that have been studied, many would meet the ATS
11 definition of adversity. Such effects include, for example, any detectible level of permanent lung
12 function loss attributable to air pollution, including both reductions in lung growth or
13 acceleration of the age-related decline of lung function; exacerbations of disease in individuals
14 with chronic cardiopulmonary diseases; reversible loss of lung function in combination with the
15 presence of symptoms; as well as more serious effects such as those requiring medical care
16 including hospitalization and, obviously, mortality.
17 As discussed above, relatively small, reversible declines in lung function parameters may
18 be of questionable significance in healthy people. However, a 5 to 15 % change in FEVi is
19 considered to have clinical importance to asthma morbidity (ATS 1991; Lebowitz et al. 1987;
20 Lippmann, 1988). The National Institutes of Health (1997) has stated that a PEF below 80% of a
21 person's personal best indicates a need for continued medication use in asthmatics. In Mortimer
22 et al. (2002), Os was associated with increased incidence of > 10%,declines in morning PEF as
23 well as morning symptoms, suggesting that Oa exposure may have clinically significant effects
24 on asthmatic children.
25 Reflecting new investigative approaches, the ATS statement describes the potential
26 usefulness of research into the genetic basis for disease, including responses to environmental
27 agents that will provide insights into the mechanistic basis for susceptibility, and provide
28 markers of risk status. Likewise biomarkers, that are indicators of exposure, effect or
29 susceptibility, may someday be useful in defining the point at which a response should be
30 equated with an adverse effect. Based on concern for segments of the population that may be
31 disproportionately exposed to environmental contaminants, or have other factors that may
32 increase susceptibility (e.g., genetic or nutritional factors), there was a call for increased research
33 in these areas.
34 Overall, the new guidance does not fundamentally change the approach previously taken
35 to define adversity, nor does it suggest a need at this time to change the structure or content of
36 the tables describing gradation of severity and adversity of effects in Tables 3-2 or 3-2 above.
November 2005 3 -62 Draft - Do Not Cite or Quote
-------
1 3.7 SUMMARY AND CONCLUSIONS FOR OZONE HEALTH EFFECTS
2 Based on dosimetric, experimental, and epidemiological evidence assessed in the 1996
3 CD, a set of findings and conclusions were drawn regarding potential health effects of Os
4 exposure as of 1996, and is reproduced in the draft CD (section 8.4.1). Similarly, section 8.7 of
5 the draft CD has summarized the main conclusions derived from the integrated analysis of
6 animal toxicology (CD, Chapter 5), human experimental (CD, Chapter 6) and epidemiological
7 (CD, Chapter 7) studies that evaluated evidence of health effects associated with short-term,
8 repeated, and long-term exposures to 63 alone or in combination with other pollutants commonly
9 found in the ambient air. This section of the draft Staff Paper integrates conclusions drawn from
10 both sections of the draft CD to provide an overview of health effects of Os alone and in
11 mixtures and to identify susceptibility factors associated with exposure to O3.
12 3.7.1 Morbidity Health Effects of Acute (Short-term) Exposures to Ozone
13 In the 1996 CD, it was concluded mat short-term 63 exposures cause: changes in
14 pulmonary function, including tachypnea (rapid, shallow breathing), decreased lung volumes and
15 flows, and increased airway responsiveness to nonspecific stimuli; increased airway resistance;
16 and airway irritation such as cough or chest pain (CD, p. 8-12). Changes in pulmonary function
17 and respiratory symptoms occur as a function of exposure concentration, duration and level of
18 exercise. According to the draft CD, results from the maj ority of acute exposure studies continue
19 to support the conclusions reported in the 1996 CD.
20 The 1996 CD concluded that group mean data from numerous controlled human exposure
21 and field studies of healthy subjects (8 to 45 years of age) exposed for 1 to 3 hr indicate that, in
22 general, statistically significant pulmonary function decrements beyond the range of normal
23 measurement variability (e.g., 3 to 5% for FEV]) occur
24 • at >0.12 ppm Os with very heavy exercise (competitive running).
25 • atXXISppmOs with heavy exercise (easy jogging),
26 * at >0.30 ppm 63 with moderate exercise (brisk walking),
27 • at >0.37 ppm Os with light exercise (slow walking), arid
28 • at XX50 ppm Os when at rest.
29 Small group mean changes (e.g., <5%) in FEVi have been observed in healthy young .
30 adults at levels as low as 0.12 ppm Oa. Also, pulmonary function decrements have been
31 observed in children and adolescents at concentrations of 0.12 and 0.14 ppm Os with heavy
32 exercise. Some individuals within a study may experience FEVi decrements in excess of 15%
33 under these conditions, even when group mean decrements are less than 5%.
November 2005 3-63 Draft - Do Not Cite or Quote
-------
1 For exposures of healthy subjects performing moderate exercise during longer duration
2 exposures (6 to 8 hr), 5% group mean decrements in FEVi were observed at
3 • 0.08 ppm after O3 5.6 hr,
4 • 0.10 ppm after Os 4.6 hr, and
5 * 0.12 ppm after O3 3 hr.
6 For these same subjects, 10% group mean FEVi decrements were observed at 0.12 ppm O3 after
7 5.6 and 6.6 hr. As in the shorter duration studies, some individuals experience changes larger
8 than those represented by group mean changes.
9 The draft CD (section 8.7) concludes that newer meta-analyses confirmed interindividual '
10 differences in lung function decrements reported in the 1996 CD. Age-specific differences in
11 lung function responses were also observed. Spirometric responses (due to decrements in lung
12 function) in healthy adults exposed to near ambient Os levels typically resolve to near baseline
13 within 4-6 hr. Meta-analyses of four controlled human exposure studies (two new and two
14 assessed in the 1996 CD) reporting the effects of prolonged (6.6 hr) exposures to 0.08 ppm Os
15 during moderate exercise on pulmonary function in young healthy adults (M=90, F=30; mean
16 age 23 years) indicate an absolute FEVi decrease of 6%, whereas FEVt increased by 1%
17 following free air exposures.
18 The 1996 CD concluded that an increase in the incidence of cough has been reported at
19 Os concentrations as low as 0.12 ppm in healthy adults during 1 to 3 hr of exposure with very
20 heavy exercise. Other respiratory symptoms, such as pain on deep inspiration, shortness of
21 breath, and lower respiratory scores (i.e., a combination of several symptoms), have been
22 observed at 0.16 ppm to 0.18 ppm Os with heavy and very heavy exercise. Respiratory
23 symptoms also have been observed following exposure to 0.08,0.10 and 0.12 ppm Os for 6.6 hr
24 with moderate exercise levels. Also, increases in nonspecific airway responsiveness in healthy
25 adults at rest have been observed after 1 to 3 hr of exposures to 0.40 ppm but not to 0.20 ppm Os;
26 during very heavy exercise, these increases were observed at concentrations as low as 0.18 ppm
27 but not at 0.12 ppm Os. Increases in nonspecific airway responsiveness during the 6.6 hr
28 exposures with moderate levels of exercise have been observed at 0.08,0.10 and 0.12 ppm Os.
29 See Table 3.4 for a summary of short-term health effects of Os based on clinical studies.
30 The 1996 CD concluded that increased Os levels are associated with increased hospital
31 admissions and emergency department visits for respiratory causes. Analyses from data in the
32 northeastern U.S. suggest that Os air pollution is associated with a substantial portion (on the
33 order of 10 to 20%) of all summertime respiratory hospital visits and admissions. The draft CD
34 concludes that a large multi-city and several single-city studies have indicated a positive
35
November 2005 3 -64 Draft - Do Not Cite or Quote
-------
2
3
Table 3-4. Summary of Ozone-Induced Respiratory Health Effects from Clinical Studies3
Health Effect
Pulmonary
Function
Decrements
Increased
Respiratory
Symptoms
Airway
Responsiveness
Respiratory
Inflammation
Changes in Host
Defenses
Decreased
Exercise
Performance
Exercise Level
Moderate
Moderate
Moderate
Competitive
Very Heavy
Heavy
Moderate
Light
At rest
Moderate
Very Heavy
Moderate
Very Heavy
At rest
Moderate
Very Heavy
Moderate
Competitive
Prolonged
Exposure
6.6 hr
4.6 hr
3.0 hr
6.6 hr
6.6 hr
6.6 hr
6.6 hr
Short-term
Exposure
Ihr
1-3 hr
1-3 hr
1-3 hr
1-3 hr
1-3 hr
1-3 hr
1-3 hr
1-3 hr
1-3 hr
Ihr
Lowest Ozone Effect
Level
0.08 ppm
0.10 ppm
0.1 2 ppm
0.1 2-0. 14 ppm
0.1 6 ppm
0.18 ppm
0.30 ppm
0.37 ppm
0.50 ppm
0.08 ppm
0.1 2 ppm
0.08 ppm
0.1 8 ppm
0.40 ppm
0.08 ppm
0.20 ppm
0.08 ppm
0.1 8 ppm
Information contained in this table is based on scientific data assessed in Chapters 6 and 8 of the CD.
November 2005
3-65
Draft - Do Not Cite or Quote
-------
1 association between increased 63 levels (especially during the warm season) and increased risk
2 for hospital admissions.
3 Pulmonary function in children at summer camps in southern Ontario, Canada, in the
4 northeastern U.S., and in southern California is associated with ambient 03 levels (CD, p. 8-13).
5 Meta-analyses indicate that a 0.50-mL decrease in FEVi is associated with a 1 ppb increase in Os
6 concentration. For preadolescent children exposed to 120 ppb (0.12 ppm) ambient 0% this
7 amounts to an average decrement of 2.4 to 3.0% in FEVi. Similar responses are reported for
8 exercising children and adolescents exposed to Os in ambient air or 63 in purified air for 1 -2 hr.
9 Short-term Os exposure of lab animals and humans disrupts the barrier function of the •
10 lung epithelium, permitting materials in the airspaces to enter lung tissue, allowing cells and
11 serum proteins to enter the airspaces, and setting off a cascade of responses (CD, p. 8-14).
12 Increased levels of PMNs and protein in lung lavage fluid have been observed following
13 exposure of healthy adults to 0.20,0.30, and 0.40 ppm Os with very heavy exercise and have not
14 been studied at lower concentrations for 1-to 3-hr exposures. Increases in lung lavage protein
15 and PMNs also have been observed at 0.08 and 0.10 ppm 0$ during 6.6 hr exposures with
16 moderate exercise; lower levels have not been tested. Short-term Os exposure of lab animals and
17 humans impairs alveolar macrophage (AM) clearance of viable and nonviable particles from the
18 lungs and decreases the effectiveness of host defenses against bacterial lung infections in animals
19 and perhaps humans. The ability of AMs to engulf microorganisms is decreased in humans
20 exposed to 0.08 and 0.10 ppm Os for 6.6 hr with moderate exercise.
21 The draft CD (p. 8-83) concludes that inflammatory responses (PMNs, inflammatory
22 mediators such as cytokines and chemokines) and permeability change (proteins, albumin),
23 typically measured in BAL fluid, also exhibit intersubject variability. Recent meta-analyses on
24 numerous clinical studies indicate interindividual differences in response to short-term 63
25 exposures. Also, inflammatory and permeability responses resolve (in some instances complete
26 recovery) and exhibit differential attenuation profiles between normal healthy subjects and
27 people with preexisting respiratory diseases. Some inflammation markers may not resolve
28 readily, and mild persistent inflammation has been reported. It was also concluded that short-
29 term O3-induced lung function decrements, respiratory symptoms, inflammation, and
30 permeability changes observed in animal toxicology studies are consistent with human studies.
31 Two health endpoints that were not addressed in the 1996 CD conclusions, but are
32 addressed in the draft CD, are school absenteeism and cardiovascular effects. The draft CD (p.
33 8-83) concludes that an association between short-term Os exposures and school absenteeism
34 (due to respiratory illness) has been suggested. Also, with regard to cardiac outcomes, a limited
35 number of field studies that examined the relationship between short-term Os exposures and
36 cardiovascular effects (heart rate variability, myocardial infarction) suggest an association.
November 2005 3-66 Draft - Do Not Cite or Quote
-------
1 3.7.2 Mortality-Related Health Effects of Short-tenn Exposures to Ozone
2 The 1996 CD concluded that an association between daily mortality and O3 concentration
3 for areas with high 03 levels (e.g., Los Angeles) was suggested. However, due to a very limited
4 number of studies available at that time, the magnitude of the effect was unclear. Since 1996,
5 new data are available from large multicity studies conducted in the U.S. and several single-city
6 studies conducted all over the world, as well as from several meta-analyses that have combined
7 information from multiple studies. The majority of these studies suggest an elevated risk of total
8 nonaccidental mortality associated with acute exposure to Os, especially in the summer or warm
9 season when Os levels are typically high, with somewhat larger effect estimate sizes for
10 associations with cardiovascular mortality (CD, p. 7-177,7-178). The draft CD finds that the
11 results from U.S. multicity time-series studies provide the strongest evidence to-date for
12 associations between short-term Os exposure and mortality. These studies, along with recent
13 meta-analyses, showed consistent risk estimates that are unlikely to be confounded by PM,
14 though the CD observes that future work is needed to better understand the influence of model
15 specifications on the risk coefficient (CD, p. 7-177 to 7-178). For cardiovascular mortality, the
16 draft CD reports that effect estimates are consistently positive, falling in the range of 1 to 8%
17 increases per 40 ppb in 1 -hr O3 (CD, p. 7-108). Overall, the draft CD concludes that these
18 findings appear to be consistent with a causal association between short-term Os exposure and
19 mortality particularly in the warm season (CD, p. 8-52).
20 3.7.3 Health Effects of Repeated Short-term Exposures to Ozone
21 The 1996 CD drew several conclusions regarding repeated short-term Os exposures (CD,
22 p. 8-15). Partial or complete attenuation is observed for some of the Os-induced responses.
23 After 5 days of exposure, pulmonary function changes return to control levels with the greatest
24 changes usually occurring on the second day, but the attenuation was reversed after 7 to 10 days
25 without Os exposure. Most inflammatory markers (e.g., PMN influx) attenuate after 5 days of
26 exposure, but markers of cell damage (e.g., LDH enzyme activity) do not attenuate and continue
27 to increase. Recovery of some inflammatory markers occurred a week to 10 days after exposure
28 ceased, but some responses were not normal after 20 days. Animal studies suggest underlying
29 cell damage continues throughout the attenuation process. Also, attenuation may alter normal
30 distribution of Os within the lungs, allowing more Os to reach sensitive regions, possibly
31 affecting lung defenses. Newer studies assessed in the draft CD (p. 8-84) supported all of these
32 conclusions in addition to which it was concluded that repeated daily, multi-hour exposure to
33 lower concentrations of Os (0.125 ppm for 4 days) causes an increased response to bronchial
34 allergen challenge in subjects with preexisting allergic airway disease, with or without asthma.
November 2005 3-67 Draft - Do Not Cite or Quote
-------
1 In these subjects, changes in airway responsiveness after Os exposure appear to be resolved more
2 slowly than changes in FEVi or respiratory symptoms.
3 3.7.4 Health Effects of Long-term Exposures to Ozone
4 In the 1996 CD, available data, primarily from animal toxicology studies, indicated that
5 exposure to Os for months to years causes structural changes in several regions of the respiratory
6 tract (CD, p. 8-15). Effects may be of greatest importance in the CAR, where the alveoli and
7 conducting airways meet. This region of the lungs is typically affected in most human airway
8 diseases. However, data from epidemiological and clinical studies is lacking, and most
9 information on chronic Os effects in the distal lungs continue to be extrapolated from animal
10 toxicology studies.
11 What had been previously been viewed as an apparent lack of reversibility of effects
12 during clean air exposures has been investigated since 1996 with animal toxicology studies using
13 exposure regimens simulating a seasonal exposure pattern. One long-term study exposed rhesus
14 monkeys to a simulated seasonal Os pattern (0.5 ppm Os 8hr/day for 5 days, every 14 days for 11
15 episodes) and reported: (1) remodeling in the distal airways; (2) abnormalities in trachea!
16 basement membrane; (3) eosinopHil accumulation in conducting airways; and (4) decrements in
17 airway innervatiori. These findings support and advance the earlier information suggestive of
18 injury and repair processes which are caused by seasonal Os exposures (CD, p.8-85). Although
19 pathophysiological changes associated with long-term Os exposures reported in animal studies
20 suggest similar changes in humans, interspecies differences in sensitivity to chronic effects of 03
21 continue to be a limiting factor in extrapolation of effect responses in animals to human health
22 effects. • •
23 • Epidemiological studies investigating chronic effects in humans following long-term
24 exposures to Oa previously provided only limited suggestive evidence. However, recent studies
25 of pulmonary function changes observed in children living in cities with high Os levels as well as
26 alterations in lung structure alterations reported in an autopsy study in LA support the conclusion
27 that long-term Os exposure may play a role in causing irreversible lung damage. Further
28 investigation will be necessary to draw firmer conclusions about chronic health effects of Os.
29 3.7.5 Health Effects of Binary Pollutant Mixtures Containing Ozone
30 In the 1996 CD, it was recognized that coexposure of humans and animals to Os and
31 other pollutants, such as NO2, SO2, H2SO4, HNO3, or CO, showed additive response for lung
32 spirometry or respiratory symptoms (CD, p..8-16). Since 1996, most animal toxicology studies
33 investigating Oj in a mixture with NO2 and H2SO4 have shown that effects can be additive,
34 synergistic, or even antagonistic, depending on the exposure regimen and the endpoint studied.
November 2005 3-68 Draft - Do Not Cite or Quote
-------
1 Although the issue of exposure to copollutants was previously described as poorly understood,
2 especially with regard to chronic effects, newer information from human and animal studies of
3 binary mixtures containing O3 suggest potential interactions depending on the exposure regimen
4 and pollutant mix (CD, p. 8-87). Examples of this newer information include:' (1) continuous
5 exposure to SO2 and NO2 increased inhaled Oj bolus absorption,'while continuous exposure to
6 O3 decreased Oa bolus absorption; (2) asthmatics exhibited enhanced airway reactivity to house
7 dust mite allergen following exposures to 03, N02 and the combination of the two gases;
8 however, spirometric response was impaired only by O3 and O3+ NO2 at higher concentrations;
9 and (3) animal toxicology studies with Os in mixture with NO2, formaldehyde, and PM
10 demonstrated additive, synergistic, or antagonistic effects depending on the exposure regimen
11 and the endpoints evaluated. One controlled-exposure study of children, designed to
12 approximate conditions of an epidemiological study by matching population and exposure
13 atmosphere (0.1 ppm 63,0.1 ppm SCh, and 101 ug/m2 H2SO4), failed to support the findings of
14 the epidemiological study. This points out the difficulty of trying to link outcomes of
15 epidemiological studies and controlled-exposure studies with binary pollutant mixtures.
16 3.7.6 Populations at Risk/Susceptibility Factors Associated with Ozone Exposure
17 The 1996 CD (p. 8-16) identified several factors that may increase sensitivity to 03 of
18 population groups, including: (1) biological variation in responsiveness to O3; (2) preexisting
19 lung disease (e.g., asthma); (3) activity patterns (e.g., exercise level); (4) personal exposure
20 history (e.g., indoor v. outdoor); and (5) personal factors (e.g., age, nutritional status, gender,
21 smoking history, ethnicity). Based on the information assessed in the 1996 CD (p. 8-18),
22 population groups that demonstrated increased responsiveness to ambient concentrations of 03
23 consist of exercising, healthy and asthmatic individuals, including children, adolescents, and
24 adults (CD, p. 8-18). Since 1996, evidence from conttolled-exposure human and animal studies,
25 as well as from epidemiological studies, has provided further support for these and other
26 susceptibility factors and populations at risk. For example, controlled-exposure human studies
27 continue to show differential biological response to O3 based on physical activity (exertion) and
28 age. These studies demonstrate a large variation in sensitivity and responsiveness to Os,
29 although specific factors that contribute to mis intersubject variability are yet to be identified.
30 Associations of increased summertime hospital admissions for asthma and COPD with ambient
31 O3 levels suggest that individuals with these respiratory diseases are populations at risk to O3
32 exposure effects. Also, based on O3-induced differential response in lung inflammation and
33 airway responsiveness, asthmatic adults and children appear to have potentially increased
34 susceptibility to O3. There is no evidence from controlled-exposure human studies which
35 suggests that individuals with COPD are more sensitive to health effects of O3. There is some
November 2005
3-69
Draft - Do Not Cite or Quote
-------
1 animal toxicology evidence which has demonstrated the importance of genetic background in Os
2 susceptibility. Genetic and molecular characterization studies of experimental animals have
3 identified genetic loci responsible for both sensitivity and resistance. Taking all of this.
4 information into account, the draft CD (p. 8-86) concludes that even though the role of ethnic,
5 racial, and nutrition status factors for O3 susceptibility remains inconclusive, all exercising
6 (moderate to high physical exertion) healthy and asthmatic adults, adolescents, and children
7 appear to exhibit increased responsiveness to ambient O3 levels and continue to be considered at
8 increased risk of O3-induced health effects. Also, any individual with respiratory or
9 cardiovascular disease or any healthy individual who is engaged in vigorous physical activity
10 outdoors during periods when Os levels are high (e.g., active outdoor children) is potentially at
11 increased risk to (Vinduced health effects, hi addition, healthy individuals and those with
« • " .*
12 cardiorespiratory impairment (e.g., those with COPD or cardiovascular disease) who are
13 "hyperresponsive" to Oj exposure (i.e., exhibit much higher than normal lung function
14 decrements) would be considered at greater risk to Os exposure. . .
November 2005 3-70 Draft - Do Not Cite or Quote
-------
1 REFERENCES
2 -
3 Abbey, D. E.; Nishino, R; McDonnell, W. F.; Burchette, R. J.; Knutsen, S. F.; Beeson, W. L.; Yang, J. X. (1999)
4 Long-term inhalable particles and other air pollutants related to mortality in nonsmokers. Am. J. Respir. Crit.
5 Care Med. 159:373-382.
6
7 Adams, W. C. (2002) Comparison of chamber arid face-mask 6.6-hr exposures to ozone on pulmonary function and
8 symptoms responses. Inhalation Toxicol. 14: 745-764.
9
10 Adams, W. C. (2003) Comparison of chamber and face mask 6.6-hr exposure to 0.08 ppm ozone via square-wave
11 and triangular profiles on pulmonary responses. Inhalation Toxicol. 15: 265-281.
12
13 Adams, W. C.; Schelegle, E. S. (1983) Ozone and high ventilation effects on pulmonary function and endurance
14 performance. J. Appl. PhysioL: Respir. Environ. Exercise Physiol. 55: 805-812.
15
16 Alexis, N.; Urch, B.; Tarlo, S.; Corey, P.; Pengelly, D.; O'Byme, P.; Silvemian, F. (2000) Cyclooxygenase
17 metabolites play a different role in ozone-induced pulmonary function decline in asthmatics compared to
18 normals. Inhalation Toxicol. 12: 1205-1224.
19
20 American Thoracic Society (1985) Guidelines as to what constitutes an adverse respiratory health effect, with
21 special reference to epidemic logical studies of air pollution. Am. Rev. Respir. Dis. 131: 666-668.
22
23 American Thoracic Society (1991) Lung function testing: selection of reference values and interpreatative
24 strategies. Am. Rev. Respir. Dis. 144: 1202-1218.
25
26 American Thoracic Society (2000) What constitutes an adverse health effect of air pollution? Am. J. Respir. Crit.
27 Care Med. 161:665-673.
28
29 Anderson, H. R.; Ponce de Leon, A.; Bland, J. M; Bower, J. S.; Emberlin, J.; Strachen, D. P. (1998) Air pollution,
30 pollens, and daily admissions for asthma in London 1987-92. Thorax 53: 842-848.
31
32 American Academy of Pediatrics, Committee on Environmental Health. (2004) Ambient air pollution: health
33 hazards to children. Pediatrics 114: 1699-1707.
34
35 Anderson, H. R.; Spix, C.; Medina, S.; Schouten, J. P.; Castellsague, J.; Rossi, G.; Zmirou, D.; Touloumi, G.;
36 Wojtyniak, B.; Ponka, A.; Bacharova, L.; Schwartz, J.; Katsouyanni, K. (1997) Air pollution and daily
37 admissions for chronic obstructive pulmonary disease in 6 European cities: results from the APHEA project.
38 Eur. Respir. J. 10: 1064-1071.
39
40 Aris, R. M.; Tager, I.; Christian, D.; Kelly, T.; Balmes, J. R. (1995) Methacholine responsiveness is not associated
41 withO3-induced decreases inFEVl. Chest 107:621-628.
42
43 Avol, E. L.; Trim, S. C.; Little, D. E.; Spier, C. E.; Smith, M. N.; Peng, R.-C.; Linn, W. S.; Hackney, J. D.; Gross,
44 K. B.; D'Arcy, J. B.; Gibbons, D.; Higgins, I. T. T. (1990) Ozone exposure and lung function in children
45 attending a southern California summer camp. Presented at: 83rd annual meeting and exhibition of the Air &
46 Waste Management Association; June; Pittsburgh, PA. Pittsburgh, PA: Air & Waste Management
47 Association; paper no. 90-150.3.
48
49 Avol, E. L.; Navidi, W. C.; Rappaport, E. B.; Peters, J. M. (1998) Acute effects of ambient ozone on asthmatic,
50 wheezy, and healthy children. Cambridge, MA: Health Effects Institute; research report no. 82.
51
52 Avol, E. L.; Gauderman, W. J.; Tan, S. M.; London, S. J.; Peters, J, M. (2001) Respiratory effects of relocating to
53 areas of differing air pollution levels. Am. J. Respir. Crit. Care Med. 164: 2067-2072.
54
November. 2005 3-71 Draft - Do Not Cite or Quote
-------
1 Balmes, J. R.; Chen, L. L.; Scannell, C.; Tager, I.; Christian, D.; Heame, P. Q.; Kelly, T.; Aris, R. M (1996) Ozone-
2 induced decrements in FEV1 and FVC do not correlate with measures of inflammation. Am. J. Respir. Crit.
3 Care Med. 153:904-909.
4 j
5 Barman, S. A. (2001) Effect of protein kinaseC inhibition on hypoxic pulmonary vasoconstriction. Am. J. Physiol.
6 Lung Cell Mol. Physiol. 280: L888
7
8 Bates, D. V.; Baker-Anderson, M.; Sizto, R. (1990) Asthma attack periodicity: a study of hospital emergency visits
9 in Vancouver. Environ. Res. 51: 51-70.
10
11 Bates, D. V.; Sizto, R. (1983) Relationship between air pollutant levels and hospital admissions in Southern Ontario.
12 Can. J. Public Health 74: 117-122.
13
14 Bates, D. V.; Sizto, R. (1987) Air pollution and hospital admissions in southern Ontario: the acid summer haze
15 effect. Environ. Res. 43: 317-331.
16
17 Bates, D. V.; Sizto, R. (1989) The Ontario Air Pollution study: identification of the causative agent. Environ. Health
18 Perspect. 79:69-72.
19
20 Beeson, W. L.; Abbey, D. E.; Knutsen, S. F. (1998) Long-term concentrations of ambient air pollutants and incident
21 lung cancer in California adults: results from the AHSMOG study. Environ. Health Perspect 106: 813-823.
22
23 Bell, M. L.; McDermott, A.; Zeger, S. L.; Samet, J. M.; Dominici, F. (2004) Ozone and short-teim mortality in
24 95 US urban communities, 1987-2000. JAMA J. Am. Med. Assoc. 292: 2372-2378.
25
26 Bell, M. L.; Dominici, F.; Samet, J. M. (2005) A meta-analysis of time-series studies of ozone and mortality with
27 comparison to the national morbidity, mortality, and air pollution study. Epidemiology 16: 436-445.
28
29 Boria-Aburto, V. H.; Loomis, D. P.; Bangdiwala, S. I.; Shy, C. M.; Rascon-Pacheco, R. A. (1997) Ozone, suspended
30 particulates, and daily mortality in Mexico City. Am. J. Epidemiol. 145: 258-268.
31
32 Bosson, J.; Stenfors,N.; Bucht, A.; Helleday, R.; Pourazar, J.; Holgate, S. T.; Kelly, F. J.; SandstrCm, T.; Wilson,
33 S.; Frew, A. J.; Blomberg, A. (2003) Ozone-induced bronchial epithelial cytokine expression differs between
34 healthy and asthmatic subjects. Clin. Exp. Allergy 33: 777-782.
35
36 Brauer, M.; Blair, J.; Vedal, S. (1996) Effect of ambient ozone exposure on lung function in farm workers. Am. J.
37 Respir. Crit. Care Med. 154: 981-987.
38
39 Bremner, S. A.; Anderson, H. R.; Atkinson, R. W,; McMichael, A. J.; Strachan, D. P.; Bland, J. M.; Bower, J. S.
40 (1999) Short term associations between outdoor air pollution and mortality in London 1992-4. Occup.
41 Environ. Med. 56: 237-244.
42 . .
43 Broeckaert, F.; Clippe, A.; Wattiez, R.; Fahnagne, P.; Bernard, A. (2003) Lung hyperpermeability, Clara-cell
44 secretory potein(CC 16), and susceptibility to ozone of five inbred strains of mice. Inhalation Toxicol.
45 15:1209-1230.
46
47 Brook, R. D.; Franklin, B.; Cascio, W.; Hong, Y.; Howard, G.; Lipsett, M.; Luepke'r, R.; Mittleman, M.; Samet, J.;
48 Smith, S. C., Jr.; Tager, I. (2004) Air pollution and cardiovascular disease. A statement for healthcare
49 professionals from the Expert Panel on Population and Prevention Science of the American Heart
50 Association. Circulation 109: 2655-2671.
51
52 Burnett, R. T.; Dales, R. E.; Raizenne, M. E.; Krewski, D.; Summers, P. W.; Roberts, G. R.; Raad-Young, M.;
53 Dann, T.; Brook, J. (1994) Effects of low ambient levels of ozone and sulfates on the frequency of respiratory
54 ' admissions to Ontario hospitals. Environ. Res. 65: 172-194.
55
November 2005 3-72 Draft - Do Not Cite or Quote
-------
1 Burnett, R. T.; Brook, J. R,; Yung, W. T.; Dales, R. E.; Krewski, D. (1997a) Association between ozone and
2 hospitalization for respiratory diseases in 16 Canadian cities. Environ. Res. 72: 24-31.
3
4 Burnett, R. T.; Brook, J. R.; Yung, W. T.; Dales, R. E.; Krewski, D. (1997b) Association between ozone and
5 hospitalization for respiratory diseases in 16 Canadian cities. Environ. Res. 72: 24-31.
6
7 Burnett, R. T.; Smith-Doiron, M.; Stieb, D.; Cakmak, S.; Brook, J. R. (1999) Effects of particulate and gaseous air
8 pollution on cardiorespiratory hospitalizations. Arch. Environ. Health 54:130-139.
9
10 Burnett, R. T.; Smith-Doiron, M.; Stieb, D.; Raizenne, M. E.; Brook, J. R.; Dales, R. E.; Leech, J. A.; Cakmak, S.;
11 Krewski, D. (2001) Association between ozone and hospitalization for acute respiratory diseases in children
12 less than 2 years of age. Am. J. Epidemiol. 153: 444-452.
13
14 Calder6n-Garcidue as, L.; RodriguezrAlcaraz, A.; Garcia, R.; Ramirez, L.; Barragan, G. (1995) Nasal
15 inflammatory responses in children exposed to a polluted urban atmosphere. J. Toxicol. Environ. Health 45:
16 427-437.
17
18 Calder6n-Garciduenas, L.; Osnaya, N.; Rodriguez-Alcaraz, A.; Villarreal-Calderon, A. (1997) DNA damage in
19 nasal respiratory epithelium from children exposed to urban pollution. Environ. Mol. Mutagen. 30:11-20.
20
21 Calderon-Garciduenas, L.; Wen-Wang, L.; Zhang, Y.-J.; Rodriguez-Alcaraz, A.; Osnaya, N.; Villarreal-Calder6n,
22 A.; Santella, R. M. (1999) 8-hydroxy-2'-deoxyguanosine, a major mutagenic oxidative DNA lesion, and DNA
23 strand breaks in nasal respiratory epithelium of children exposed to urban pollution. Environ. Health
24 Perspect 107: 469-474.
25
26 Calderon-Garciduenas, L.; Valencia-Salazar, G.; Rodriguez-Alcaraz, A.; Gambling, T. M.; Garcia, R.; Osnaya, N.;
27 Villarreal-Calderon, A.; Devlin, R. B.; Carson, J. L. (2001) Ultrastructural nasal pathology.in children
28 chronically and sequentially exposed to air pollutants. Am. J. Respir. Cell Mol. Biol. 24: 132-138.
29
30 Calderon-Garciduenas, L.; Mora-Tiscareno, A.; Fordham, L. A.; Valencia-Salazar, G.; Chung, C. J.; Rodriguez-
31 Alcaraz, A.; Paredes, R.; Variakojis, D.; Villarreal-Calderon, A.; Flores-Camacho, L.; Antunez-Solis, A.;
32 Henriquez-Roldan, C.; Hazucha, M. J. (2003) Respiratory damage in children exposed to urban pollution.
33 Pediatr. Pulmonol. 36:148-161.
34 '
35 Centers for Disease Control and Prevention. (2004) The health consequences of smoking: a report of the Surgeon
36 General. Atlanta, GA: U.S. Department of Health and Human Services, National Center for Chronic Disease
37 Prevention and Health Promotion, Office on Smoking and Health. Available:
38 http://www.cdc.gov/tobacco/sgr/sgr_2004/chapters.htm (18 August, 2004).
39
40 Chen, P.-C.; Lai, Y.-M.; Wang, J.-D.; Yang, C.-Y.; Hwang, J.-S.; Kuo, H'.-W.; Huang, S.-L.; Chan, C.-C. (1998)
41 Adverse effect of air pollution on respiratory health of primary school children in Taiwan. Environ. Health
42 Perspect 106: 331-335.
43
44 Chen, L.; Jennison, B. L.; Yang, W.; Omaye, S. T. (2000) Elementary school absenteeism and air pollution.
45 Inhalation Toxicol. 12: 997-1016.
46
47 Cho, H. Y.; Hotchkiss, J. A.; Bennett, C. B.; Harkema, J. R. (1999) Effects of pre-existing rhinitis on ozone-induced
48 mucous cell metaplasia in rat nasal epithelium. Toxicol. Appl. Pharmacol. 158: 92-102. ,
49
50 Cho, H.-Y.; Zhang, L.-Y.; Kleeberger, S. R. (2001) Ozone-induced lung inflammation and hyperreactivity are
51 mediated via tumor necrosis factor-a receptors. Am. J. Physiol. 280: L537-L546.
52
53 Chock, D. P.; Winkler, S. L.; Chen, C. (2000) A study of the association between daily mortality and ambient air
54 pollutant concentrations in Pittsburgh, Pennsylvania. J. Air Waste Manage. Assoc. 50: 1481-1500.
55
November 2005 3-73 Draft - Do Not Cite or Quote
-------
1 Christian, D. L.; Chen, L. L.; Scannell, C. H.; Ferrando, R. E.; Welch, B. S.; Bahnes, J. R. (1998) Ozone-induced
2 inflammation is attenuated with multiday exposure. Am. J. Respir. Crit. Care Med. 158; 532-537.
3
4 Cody, R. P.; Weisel, C. P.; Bimbaum, G.; Lioy, P. J. (1992) The effect of ozone associated with summertime
5 photochemical smog on the frequency of asthma visits to hospital emergency departments. Environ. Res.
6 58: 184-194.
7
8 David, G. L.; Romieu, I.; Sienra-Monge, J. J.; Collins, W. J.; Ramirez-Aguilar, M.; Del Rio-Navarro, B. E.; Reyes-
9 Ruiz, N. I.; Morris, R. W.; Marzec, J. M.; London, S. J. (2003)Nicotinamide adenine dinucleotide
10 (Phosphate) reduced:quinone oxidoreductase and glutathione s-transferase ml polymorphism and childhood
11 asthma. Am. J. Respir. Crit. Care Med. 168: 1199-1204.
12
13 De Leon, S. F.; Thurston, G. D.; Ito, K. (2003) Contribution of respiratory disease to nonrespiratory mortality
14 associations with air pollution. Am. J. Respir. Crit. Care Med. 167: 1117-1123.
15
16 Delfino, R. J.; Coate, B. D.; Zeiger, R. S.; Seltzer, J. M.; Street, D. H.; Koutrakis, P. (1996) Daily asthma severity in
17 relation to personal ozone exposure and outdoor fungal spores. Am. J. Respir. Crit. Care Med. 154: 633-641.
18
19 Delfino, R. J.; Zeiger, R. S.; Seltzer, J. M.; Street, D. H.; Matteucci, R. M.; Anderson, P. R.; Koutrakis, P. (1997a)
20 The effect of outdoor fungal spore concentrations on daily asthma severity. Environ. Health Perspect
21 105:622-635.
22
23 Delfino, R. J.; Murphy-Moulton, A. M.; Burnett, R. T.; Brook, J. R.; Becklake, M. R. (1997b) Effects of air
24 pollution on emergency room visits for respiratory illnesses in Montreal, Quebec. Am. J. Respir. Crit. Care
25 Med. 155: 568-576.
26
27 Delfino, R. J.; Gone, H.; Linn, W. S.; Pellizzari, E. D.; Hu, Y. (2003) Asthma symptoms in Hispanic children and
28 daily ambient exposures to toxic and criteria air pollutants. Environ. Health Perspect. Ill: 647-656.
29
30 Desqueyroux, H.; Pujet, J.-C.; Prosper, M.; Squinazi, F.; Moraas, I. (2002a) Short-term effects of low-level air
31 pollution on respiratory health of adults suffering from moderate to severe asthma. Environ. Res. A 89: 29^
32 37.
33
34 Desqueyroux, H.; Pujet, J.-C.; Prosper, M.; Le Moullec, Y.; Momas, I. (2002b) Effects of air pollution on adults
35 with chronic obstructive pulmonary disease. Arch. Environ. Health 57: 554-560.
36
37 Devlin, R. B.; Folinsbee, L. J.; Biscardi, F.; Hatch, G.; Becker, S.; Madden, M. C.; Robbins, M.; Keren, H. S. (1997)
3 8 Inflammation and cell damage induced by repeated exposure of humans to ozone. Inhalation Toxicol. 9: 211-
39 235.
40
41 Devlin, R. B.; McDonnell, W. F.; Mann, R.; Becker, S.; House, D. E.; Schreinemachers, D.; Keren, H. S. (1991)
42 Exposure of humans to ambient levels of ozone for 6.6 hrs causes cellular and biochemical changes in the
43 lung. Am. J. Respir. Cell Mol. Biol. 4: 72-81.
44
45 Devlin, R. B.; McDonnell, W. F.; Koren, H. S.; Becker, S. (1990) Prolonged exposure of humans to 0.10 and
46 0.08 ppm ozone results in inflammation in the lung. Presented at: 83rd annual meeting of the Air & Waste
47 Management Association; June; Pittsburgh, PA. Pittsburgh, PA: Air & Waste Management Association;
48 paper no. 90-150.2.
49
50 Dietert, R. R.; Etzel, R. A.; Chen, D.; Halonen, M.; Holladay, S. D.; Jarabek, A. M.; Landreth, K.; Peden, D. B.;
51 Pinkerton, K.; Smialowicz, R. J.; Zoetis, T. (2000) Workshop to identify critical window of exposure for
52 children's health: immune and respiratory systems work group summary. Environ. Health Perspect. Suppl.
53 108(3): 483-490.
54
55
November 2005 3-74 Draft - Do Not Cite or Quote
-------
1 Dimeo, M. J.; Glenn, M. G.; Holtzman, M. J.; Shelter, J. R.; Nadel, J. A.; Boushey, H. A. (1981) Threshold
2 concentration of ozone causing an increase in bronchial reactivity in humans and adaptation with repeated
3 exposures. Am. Rev. Respir. Dis. 124:245-248.
4 ..
5 Dockery.D. W.; Schwartz, J.; Spengler, J. D. (1992) Air pollution and daily mortality: associations with
6 participates and acid aerosols. Environ. Res. 59: 362-373.
7
8 Dockery, D. W.; Pope, C. A., Ill, Xu, X.; Spengler, J. D.; Ware, J. H.; Fay, M. E.; Ferris, B. G., Jr.; Speizer, F. E.
9 (1993) An association between air pollution and mortality in six U.S. cities. N. Engl. J. Med. 329:1753-1759.
10 '
11 Dominici, P.; McDennott, A.; Daniels, M.; Zeger, S. L.; Samet, J. M. (2003) Mortality among residents of 90 cities.
12 In: Revised analyses of time-series studies of air pollution and health. Special report. Boston, MA: Health
13 Effects Institute; pp. 9-24. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdf [12 May, 2004].
14 '
15 Dormans, J. A. M. A.; Van Bree, L.; Boere, A. J. F.; Marra, M.; Rombout, P. J. A. (1999) Interspecies differences in
16 time course of pulmonary toxicity following repeated exposure to ozone. Inhalation Toxicol. 11: 309-329.
17
18 Fairley, D. (1999) Daily mortality and air pollution in Santa Clara County, California: 1989-1996. Environ. Health
19 Perspect. 107: 637-641.
.20 ,
21 Fairley, D. (2003) Mortality and ak pollution for Santa Clara County, California, 1989-1996. In: Revised analyses
22 of time-series studies of air pollution and health. Special report. Boston, MA: Health Effects Institute;
23 pp. 97-106. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdftl8 October, 2004].
24
25 Fanucchi, M. V.; Hotchkiss, J. A.; Harkema, J. R. (1998) Endotoxin potentiates ozone-induced mucous cell
26 metaplasia in rat nasal epithelium. Toxicol. Appl. Pharmacol. 152:1 -9.
27
28 Folinsbee, L. J.; Silverman, F.; Shephard, R. J. (1977) Decrease of maximum work performance following ozone
29 exposure. J. Appl. Physiol,: Respir. Environ. Exercise Physiol. 42: 531-536.
30 .
31 Folinsbee, L. J.; McDonnell, W. F.; Horstaan, D. H. (1988) Pulmonary function and symptom responses after
32 6.6-hr exposure to 0.12 ppm ozone with moderate exercise. JAPCA 38:28-35.
33
34 Folinsbee, L. J.; Horstman, D. H.; Kehrl, H. R.; Harder, S.; Abdul-Salaam, S.; Ives, P. J. (1994)Respiratory
35 responses to repeated prolonged exposure to 0.12 ppm ozone. Am. J. Respir. Grit Care Med. 149: 98-105.
36
37 Folinsbee, L. J.; Hazucha, M. J. (2000) Time course of response to ozone exposure in healthy adult females.
38 Inhalation Toxicol. 12:151-167.
39
40 Foster, W. M.; Silver,.!. A.; Groth, M. L. (1993) Exposure to ozone alters regional function and particle dosimetry
41 in the human lung. J. Appl. Physiol. 75: 1938-1945.
42 . " '
43 Foster, W. M.; Weinmann, G. G.; Menkes, E.; Macri, K. (1997) Acute exposure of humans to ozone impairs small
44 airway function. Ann. Occup. Hyg. 41(suppl. 1): 659-666.
45
46 Foxcroft, W. J.; Adams, W. C. (1986) Effects of ozone exposure on four consecutive days on work performance and
47 VO2max. J. Appl. Physiol. 61: 960-966.
48
49 Friedman, M. S.; Powell, K. E.; Hutwagner, L.; Graham, L. M.; Teague, W. G. (2001) Impact of changes in
50 transportation and commuting behaviors during the 1996 summer Olympic games in Atlanta on air quality
51 and childhood asthma. JAMA J. Am. Med. Assoc. 285: 897-905.
52
53 Frischer, T.; Studnicka, M.; Gartner, C.; Tauber, E.; Horak, F.; Veiter, A.; Spengler, J.; Kuhr, J.; Urbanek, R. (1999)
54 Lung function growth and ambient ozone: a three-year population study in school children. Am. J. Respir.
55 Crit. Care Med. 160: 390-396.
56
November 2005 3-75 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
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
Frischer, T.; Studnicka, M; Halmerbauer, G.; Horak, F.; Gartner, C.; Tauber, E.; Koller, D. Y. (2001) Ambient
ozone exposure is associated with eosinophil activation in healthy children. Clin. Exp. Allergy 31:1213-
1219.
Galizia, A.; Kinney, P. L. (1999) Long-term residence in areas of high ozone: associations with respiratory health in
a nationwide sample of nonsmoking young adults. Environ. Health Perspect. 107: 675-679.
Gamble, J. L. (1998) Effects of ambient air pollution on daily mortality: a time series analysis of Dallas, Texas,
1990-1994. Presented at: 91 st annual meeting and exhibition of the Air & Waste Management Association;
June; San Diego, CA. Pittsburgh, PA: Air & Waste Management Association; paper no. 98-MP26.03.
Gauderman, W. J.; McConnell, R.; Gilliland, F.; London, S.; Thomas, D.; Avol, E.; Vora, H.; Berhane, K.;
Rappaport, E. B.; Lurmann, F.; Margolis, H. G.; Peters, J. (2000) Association between air pollution and lung
function growth in southern California children. Am. J. Respir. Crit. Care Med. 162: 1383-1390.
Gauderman, W. J.; Gilliland, G. F.; Vora, H.; Avol, E.; Stram, D.; McConnell, R.; Thomas, D.; Lurmann, F.;
Margolis, H. G.; Rappaport, E. B.; Berhane, K.; Peters, J. M. (2002) Association between air pollution and
lung function growth in southern California children: results from a second cohort. Am. J. Respir. Crit. Care
Med. 166:76-84.
Gauderman, W. J.; Avol, E.; Gilliland, F.; Vora, H.; Thomas, D.; Berhane, K.; McConnell, R.; Kuenzli, N.;
Lurmann, F.; Rappaport, E.; Margolis, H.; Bates, D.; Peters, J. (2004a) The effect of air pollution on lung
development from 10 to 18'years of age. N. Engl. J. Med. 351:1057-1067.
Gauderman, W. J.; Avol, E.; Gilliland, F. (2004b) Air pollution and lung function [reply letter]. N. Engl. J. Med.
351:2653.
Gent, J. F.; Triche, E. W.; Holford, T. R; Beknger, K.; Bracken, M. B.; Beckett, W. S.; Leaderer, B. P. (2003)
Association of low-level ozone and fine particles with respiratory symptoms in children with asthma. JAMA
J. Am. Med. Assoc. 290: 1859-1867.
Gielen, M. H.; Van Der Zee, S. C.; Van Wijnen, J. H.; Van Steen, C. J.; Brunekreef, B. (1997) Acute effects of
siimmer air pollution on respiratory health of asthmatic children. Am. J. Respir. Crit. Care Med. 155: 2105-
2108.
Gilliland, F. D.; Berhane, K.; Rappaport, E. B.; Thomas, D. C.; Avol, E.; Gauderman, W. J.; London, S. J.;
. Margolis, H. G.; McConnell, R; Islam, K. f.; Peters, J. M. (2001) The effects of ambient air pollution on
school absenteeism due to respiratory illnesses. Epidemiology 12: 43-54.
Gong, H., Jr.; Bradley, P. W.; Simmons, M. S.; Tashkin, D. P. (1986) Impaired exercise performance and pulmonary
function in elite cyclists during low-level ozone exposure in a hot environment. Am. Rev. Respir. Dis.
134:726-733.
Gong, H., Jr.; McManus, M. S.; Linn, W. S. (1997a) Attenuated response to repeated daily ozone exposures in
asthmatic subjects. Arch. Environ. Health 52: 34-41.
Gong, H., Jr.; Shamoo, D. A; Anderson, K. R.; Linn, W. S. (1997b) Responses of older men with and without
chronic obstructive pulmonary disease to prolonged ozone exposure. Arch. Environ. Health 52: 18-25.
Gong, H., Jr.; Wong, R; Sarma, R. J.; Linn, W. S.; Sullivan, E. D.; Shamoo, D. A.; Anderson, K. R.; Prasad, S, B.
(1998a) Cardiovascular effects of ozone exposure in human volunteers. Am. J. Respir. Crit. Care Med.
158:538-546.
Gong, H., Jr.; Simmons, M. S.; Linn, W. S.; McDonnell, W. F.; Westerdahl, D. (1998b) Relationship between acute
ozone responsiveness and chronic loss of lung function in residents of a high-ozone community. Arch.
Environ. Health 53: 313-319.
November 2005
3-76
Draft - Do Not Cite or Quote
-------
1
2
3
4
5
6
7
g
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
50
51
52
53
54
55
56
Gouveia, N.; Fletcher, T. (2000b) Time series analysis of air pollution and mortality: effects by cause, age and
socioeconomic status. J. Epidemiol. Community Health 54: 750-755.
Graham, D. E.; Koren, H. S. (1990) Biomarkers of inflammation in ozone-exposed humans: comparison of the nasal
and bronchoalveolar lavage. Am. Rev. Respir. Dis. 142:152-156.
Greer, J. R.; Abbey, D. E.; Burchette, R. J. (1993) Asthma related to occupational and ambient air pollutants in
nonsmokers. J. Occup. Med. 35: 909-915..
Gryparis, A.; Forsberg,B.; Katsouyanni, K.; Analitis, A.; Touloumi, G.; Schwartz, I.; Samoli, E.; Medina, S.;
Anderson, H. R.; Niciu, E. M.; Wichmann, H.-E.; Kriz, B.; Kosnik, M.; Skorkovsky, J.; Vonk, J. M.;
Dortbudak, Z. (2004) Acute effects of ozone on mortality from the "air pollution and health: a European
approach" project. Am. J. Respir. Crit. Care Med. 170: 1080-1087.
Hatch, G. E.; Slade, R.; Harris, L.P.; McDonnell, W.F.; Devlin, R.B.; Koren, H.S.; Costa, D. L., McKee, J. (1994)
Ozone dose and effect in humans and rats: a comparison using oxygen-18 labeling and bornchoalveolar
. lavage. Am.J. Respir. Crit Care Med. 150:676-683.
Hazucha, M. J.; Folinsbee, L. J.; Bromberg, P. A. (2003) Distribution and reproducibility of spirometric response to
ozone by gender and age. J. Appl. Physiol. 95: 1917-1925.
Hazucha, M. J.; Folinsbee, L. J.; Seal, E., Jr. (1992) Effects of steady-state and variable ozone concentration profiles
on pulmonary function. Am. Rev. Respir. Dis. 146:1487-1493.
Hemandez-Garduno,E.; Perez-Neria, J.; Paccagnella, A. M.; Pi a-Garcia, M.;Munguia-Castro, M.; Catalan-
Vazquez, M.; Rojas-Ramos, M. (1997) Air pollution and respiratory health in Mexico City. J. Occup.
Environ. Med. 39:299-307.
Higgins, I. T. T.; D'Arcy, J. B.; Gibbons, D. I.; Avol, E. L.; Gross, K. B. (1990) Effect of exposures to ambient
ozone on ventilatory lung function in children. Am. Rev. Respir. Dis. 141:1136-1146.
Hutemann, T. J. N.; Stolk, J.; Van der Zee, S. C.; Brunekreef, B.; De Bruijne, C. R.; Fischer, P. H.; Ameling, C. B.;
Sterk, P. J.; Hiemstra, P. S.; Van Bree, L. (1998) Asthma severity and susceptibility to air pollution. Eur.
Respir. 1.11:686-693.
Hiltermann, J. T. N.; Lapperre,T. S.; Van Bree, L.; Steerenberg, P. A.; Brahim, J. J.; Sont, J. K.; Sterk, P. J.;
Hiemstra, P. S.; Stolk, J. (1999) Ozone-induced inflammation assessed in sputum and bronchial lavage fluid
from asthmatics: a new noninvasive tool in epidemiological studies on air pollution and asthma. Free Radical
Biol. Med. 27: 1448-1454.
Holguin, F.; Tellez-Rojo, M. M.; Hernandez, M.; Cortez, M.; Chow, J. C.; Watson, J. G.; Mannino, D.; Romieu, I.
(2003) Air pollution and heart rate variability among the elderly in Mexico City. Epidemiology 14: 521-527.
Holz, O.; Milcke, M.; Paasch, K.; Bohme, S.; Timm, P.; Richter, K.; Magnussen, H.; Jorres, R. A. (2002) Repeated
ozone exposures enhance bronchial allergen responses in subjects with rhinitis or asthma. Clin. Exp. Allergy.
32: 681-689.
Holz, O.; Jftrres, R. A.; Timm, P.; Mucke, M.; Richter, K.; Koschyk, S.; Magnussen, H. (1999) Ozone-induced
airway inflammatory changes differ between individuals and are reproducible. Am. J. Respir. Crit. Care Med.
159:776-784.
Hoppe, P.; Peters, A.; Rabe, G.; Praml, G.; Lindner, J.; Jakobi, G.; Fruhmann, G.; Nowak, D. (2003) Environmental
ozone effects in different population subgroups. Int. J. Hyg. Environ. Health 206: 505-516.
November 2005
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
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
Horak, F,, Jr.; Studnicka, M.; Gartner, C.; Spengler, J. D.; Tauber, E.; Urbanek, R.; Veiter, A.; Frischer, T. (2002a)
Particulate matter and lung function growth in children: a 3-yr follow-up study in Austrian schoolchildren.
Eur. Respir. J. 19:838-845.
Horak, F., Jr.; Studnicka, M.; Gartner, C.; Spengler, J. D.; Tauber, E.; Urbanek, R.; Veiter, A.; Frischer, T. (2002b)
Particulate matter and lung function growth in children: a 3-yr follow-up study in Austrian schoolchildren
[author response], Eur. Respir. J. 20: 1355.
Horstman, D. H.; Folinsbee, L. J.; Ives, P. J.; Abdul-Salaam, S.; McDonnell, W. F. (1990) Ozone concentration and
pulmonary response relationships for 6.6-hr exposures with five hrs of moderate exercise to 0.08,0.10, and
0.12 ppm. Am. Rev. Respir. Dis. 142: 1158-1163.
Horstman, D. H.; Ball, B. A.; Brown, J.; Gerrity, T.; Folinsbee, L. J. (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.
Hotchkiss, J. A.; Hilaski, R.; Cho, H.; Regan, K.; Spencer, P.; Slack, K.; Harkema, J. R. (1998) Fluticasone
propionate attenuates ozone-induced rhinitis and mucous cell metaplasia in rat nasal airway epithelium.
Am. J. Respir. Cell Mol. Biol. 18: 91-99.
Housley, D. G.; Eccles, R.; Richards, R. J. (1996) Gender difference in the concentration of the antioxidant uric acid
in human nasal lavage. Acta Oto-Laryngol. 116: 751-754.
Huang, Y.; Dominici, F.; Bell, M. L. (2005) Bayesian hierarchical distributed lag models for summer ozone
exposure and cardio-respiratory mortality. Environmetrics 16: 547-562.
Ihorst, G.; Frischer, T.; Horak, F.; Schumacher, M.; Kopp, M.; Forster, J.; Mattes, J.; Kuehr, J. (2004) Long- and
medium-term ozone effects on lung growth including a broad spectrum of exposure. Eur. Respir. J. 23: 292-
299.
Ilabaca, M.; Olaeta, L; Campos, E.; Villaire, J.; Tellez-Rojo, M. M.; Romieu, I. (1999) Association between levels •
of fine particulate and emergency visits for pneumonia and other respiratory illnesses among children in
Santiago, Chile. J. Air Waste Manage. Assoc. 49: 154-163.
Ito, K.; De Leon, S, F.; Lippmann, M. (2005) Associations between ozone and daily mortality, analysis and meta-
analysis. Epidemiology 16:446-457.
Ito, K. (2003) Associations of particulate matter components with daily mortality and morbidity in Detroit,
Michigan. In: Revised analyses of time-series studies of air pollution and health. Special report. Boston, MA:
Health Effects Institute; pp. 143-156. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdf
[12 May, 2004],
Jaffe, D. H.; Singer, M. E.; Rimm, A. A. (2003) Air pollution and emergency department visits for asthma among
Ohio Medicaid recipients, 1991-1996. Environ. Res. 91: 21-28.
Jalaludin, B. B.; Chey,-T.; OToole, B. L; Smith, W. T.; Capon, A. G.; Leeder, S. R. (2000) Acute effects of low
levels of ambient ozone on peak expiratory flow rate in a cohort of Australian children. Int. J. Epidemiol.
29: 549-557.
Jalaludin, B. B.; OToole, B. I.; Leeder, S. R. (2004) Acute effects of urban ambient air pollution on respiratory
symptoms, asthma medication use, and doctor visits for asthma in a cohort of Australian children. Environ
Res. 95: 32-42.
Jones, G. N.; Sietten, C.; Mandry, C.; Brantley, P. J. (1995) Ozone level effect on respiratory illness: an
investigation of emergency department visits. South. Med. J. 88:1049-1056.
November 2005
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
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
Jdrres, R. A.; Holz, O.; Zachgo, W.; Timm, P.; Koschyk, S.; Mttller, B.; Grimminger, F.; Seeger, W.; Kelly, F. J.;
Dunster, C.; Frischer, T.; Lubec, G.; Waschewski, M.; Niendorf, A.; Magnussen, H. (2000) The effect of
repeated ozone exposures on inflammatory markers in bronchoalveolar lavage fluid and mucosal biopsies.
Am. J. Respir. Crit. Care Med. 161: 1855-1861.
Jorres, R.; Nowak, D.; Magnussen, H.; Speckin, P.; Koschyk, S. (1996) The effect of ozone exposure on allergen
responsiveness in subjects with asthma or rhinitis. Am. J. Respir. Crit. Care Med. 153: 56-64.
Just, J.; Segala, C.; Sahraoui, F.; Priol, G.; Grimfeld, A; Neukirch, F. (2002) Short-term health effects of particulate
and photochemical air pollution in asthmatic children. Eur. Respir. J. 20: 899-906.
Kehrl, H. R.; Peden, D. B.; Ball, B. A; Folinsbee, L. J.; Horstman, D. H. (1999) Increased specific airway reactivity
of persons with mild allergic asthma after 7.6 hrs of exposure to 0.16 ppm ozone. J. Allergy. Clin. Immunol.
104: 1198-1204.
Kim, S.-Y.; Lee, J.-T.; Hong, Y.-C.; Ahn, K.-J.; Kim, H. (2004) Determining the threshold effect of ozone on.daily
mortality: an analysis of ozone and mortality in Seoul, Korea, 1995-1999. Environ. Res. 94:113-119.
Kinney, P. L.; Lippmann, M. (2000) Respiratory effects of seasonal exposures to ozone and particles. Arch.
Environ. Health 55: 210-2161
Kinney, P. L.; Aggarwal, M.; Nikiforov, S. V.; Nadas, A. (1998) Methods development for epidemiological
investigations of the health effects of prolonged ozone exposure. Part III: an approach to retrospective
estimation of lifetime ozone exposure using a questionnaire and ambient monitoring data (U.S. sites).
Cambridge, MA: Health Effects Institute; research report no. 81; pp. 79-108.
Kinney, P. L.; Nilsen, D. M.; Lippmann, M.; Brescia, M.; Gordon, T.; McGovern, T.; ElFawal, H.; Devlin, R. B.;
Rom, W. N. (1996) Biomarkers of lung inflammation in recreational joggers exposed to ozone. Am. J.
Respir. Crit. Care Med. 154:1430-1435.
Kinney, P. L.; Ito, K.; Thurston, G. D. (1995) A sensitivity analysis of mortality/PMlO associations in Los Angeles.
In: Phalen, R. F.; Bates, D. V., eds. Proceedings of the colloquium on particulate air pollution and human
mortality and morbidity; January 1994; Irvine, CA. Inhalation Toxicol. 7: 59-69.
Kinney, P. L.; Ozkaynak, H. (1991) Associations of daily mortality and air pollution in Los Angeles County.
Environ. Res. 54: 99-120.
Kleeberger, S. R.; Levitt, R. C.; Zhang, L.-Y.; Longphre, M.; Harkema, J.; Jedlicka, A.; Eleff, S. M.; DiSilvestre,
D.; Holroyd, K. J. (1997) Linkage analysis of susceptibility to ozone-induced lung inflammation in inbred
mice. Nat. Genet. 17: 475-478.
Klemm, R. J.; Lipfert, F. W.; Wyzga, R. E.; Gust, C. (2004) Daily mortality and air pollution in Atlanta: two years
of data from ARIES. Inhalation Toxicol. 16(suppl. 1): 131-141.
Klemm, R. J.; Mason, R. M., Jr. (2000) Aerosol Research and Inhalation Epidemiological Study (ARIES): air
quality and daily mortality statistical modeling-interim results. J. Air. Waste Manage. Assoc. 50: 1433-1439.
Koken, P. J.; Piver, W. T.; Ye, F.; Elixhauser, A.; Olsen, L. M.; Portier, C. J. (2003) Temperature, air pollution,
and hospitalization for cardiovascular diseases among elderly people in Denver. Environ. Health Perspect.
111:1312-1317.
Kopp, M. V.; Ulmer, C.; Ihorst, G.; Seydewitz, H. H.; Frischer, T.; Forster, J.; Kuehr, J. (1999) Upper airway
inflammation in children exposed to ambient ozone and potential signs of adaptation. Eur. Respir. J. 14: 854-
861.
November 2005
3-79
Draft - Do Not Cite or Quote
-------
1 Kopp, M. V.; Bonnet, W.; Frischer, T.; Ulmer, C.; Studnicka, M.; Ihorst, G.; Gardner, C.; Forster, J.; Urbanek, R.;
2 Kuehr, J. (2000) Effects of ambient ozone on lung function in children over a two-summer period. Eur.
3 Respir. J. 16:893-900.
4
5 Korrick, S. A.; Neas, L. M.; Dockery, D. W.; Gold, D. R.; Allen, G. A.; Hill, L. B.; Kimball, K, D.; Rosner, B. A.;
6 Speizer, F. E. (1998) Effects of ozone and other pollutants on the pulmonary function of adult hikers.
7 Environ. Health Perspect 106: 93-99.
8
9 Kreit, J. W.; Gross, K. B.; Moore, T. B.; Lorenzen, T. J.; D'Arcy, J.; Eschenbacher, W. L. (1989) Ozone-induced
10 changes in pulmonary function and bronchial responsiveness in asthmatics. J. Appl. Physiol. 66: 217-222.
11
12 Krewski, D.; Burnett, R. T.; Goldberg, M. S.; Hoover, K.; Siemiatycki, J,; Jerrett,'M; Abrahamowicz, M.; White,
13 W. H. (2000) Reanalysis of the Harvard Six Cities Study and the American Cancer Society Study of
14 particulate air pollution and mortality. A special report of the Institute's particle epidemiology reanalysis
15 project. Cambridge, MA: Health Effects Institute.
16
17 Krishna, M. T.; Springall, D.; Meng, Q.-H.; Withers, N.; Macleod, D.; Biscione, G.; Frew, A.; Polak, J.; Holgate, S.
18 (1997) Effects of ozone on epithelium and sensory nerves in the bronchial mucosa of healthy humans. Am. J.
19 Respir. Crit. Care Med. 156: 943-950.
20
21 Kulle, T. J.; Sauder, L. R.; Kerr, H. D.; Farrell, B. P.; Bermel, M. S.; Smith, D. M. (1982) Duration of pulmonary
22 function adaptation to ozone in humans. Am. Ind. Hyg. Assoc. J. 43: 832-837.
23
24 Kttnzli, N.; Lurmann, F.; Segal, M.; Ngo, L.; Balmes, J.; Tager, I. B. (1997) Association between lifetime ambient
25 ozone exposure and pulmonary function in college freshmen—results of a pilot study. Environ. Res. 72: 8-23.
26
27 Lebowitz, M. D.; Camilli, A. E.; Bronnimam, D.; Quackenboss, J. (1987) The significance and meaningfulness of
28 intraindividual changes in objective test results as responses to air contaminants. Presented at: 80th annual
29 meeting of the Air Pollution Control Association; June; New York, NY. Pittsburgh, PA: Air Pollution Control
30 Association; paper no. 87-32.1.
31
32 Levy, J. I-; Chemerynski, S. M.; Samat, J. A. (2005) Ozone exposure and mortality, an empiric Bayes
33 metaregression analysis. Epidemiology 16: 458-468.
34
35 Liao, D.; Duan, Y.; Whitsel, E. A.; Zheng, Z.-J.; Heiss, G.; Chinchilli, V. M.; Lin, H.-M. (2004) Association of
36 higher levels of ambient criteria pollutants with impaired cardiac autonomic control: a population-based
37 study. Am. J. Epidemiol. 159: 768-777.
38
39 Lin, M.; Chen, Y.; Villeneuve, P. J.; Burnett, R. T.; Lemyre, L.; Hertzman, C.; McGrail, K. M.; Krewski, D. (2004)
40 Gaseous air pollutants and asthma hospitalization of children with low household income in Vancouver,
41 British Columbia, Canada. Am. J. Epidemiol. 159: 294-303.
42
43 Lin, C. A.; Martins, M. A.; Farhat, S. C. L.; Pope, C. A., Ill; Concei? o, G. M. S.; Anastacio, V. M.; Hatanaka, M.;
44 Andrade, W. C.; Hamaue, W. R.; BOhm, G. M.; Saldiva, P. H. N. (1999) Air pollution and respiratory illness
45 of children in S o Paulo, Brazil: Paediatr. Perinat. Epidemiol. 13: 475-488.
46
47 Linder, J.; Herren, D.; Monn, C.; Wanner, H.-U. (1988) Die Wirkung von Ozon auf die korperliche
48 Leistungsfahigkeit [The effect of ozone on physical activity]. Schweiz Z. Sportmed. 36: 5-10.
49 .
50 Linn, W. S.; Shamoo, D. A.; Anderson, K. R.; Peng, R.-C.; Avol, E. L.; Hackney, J. D.; Gong, H., Jr. (1996) Short-
51 term air pollution exposures and responses in Los Angeles area schoolchildren. J. Exposure Anal. Environ.
52 Epidemiol. 6: 449-472.
53
54 Lipfert, F. W.; Perry, H. M., Jr.; Miller, J. P.; Baty, J. D.; Wyzga, R. E.; Carmody, S. E. (2000) The Washington
55 University-EPRI veterans' cohort mortality study: preliminary results. In: Grant, L. D., ed. PM2000:
56 particulate matter and health. Inhalation Toxicol. 12(suppl. 4): 41-73.
November 2005 3-80 Draft - Do Not Cite or Quote
-------
1
2
3
4
5
6
7
8
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
50
51
52
53
54
55
Lipfert, F. W.; Hammerstrom, T, (1992) Temporal patterns in air pollution and hospital admissions. Environ. Res.
59: 374-399.
Lippmann, M.; Ito, K.; Nadas, A.; Burnett, R. T. (2000) Association of particulate matter components with daily
mortality and morbidity in urban populations. Cambridge, MA: Health Effects Institute; research report
no. 95.
Lippmann, M. (1988) Health significance of pulmonary function responses to airborne irritants. JAPCA 38: 881-
887.
McBride, D. E.; Koenig, J. Q.; Luchtel, D. L.; Williams, P. V.; Henderson, W. R., Jr. (1994) Inflammatory effects of
ozone in the upper airways of subjects with asthma. Am. J. Respir. Crit. CareMed. 149:1192-1197.
McConnell, R.; Berhane, K.; Gilliland, F.; London, S. J.; Vora, H.; Avol; E.; Gauderman, W. J.; Margolis, H. G.;
Lurmann, F.; Thomas, D. C.; Peters, J. M. (1999) Air pollution and bronchitic symptoms in southern
California children with asthma. Environ. Health Perspect. 107: 757-760.
McConnell, R.; Berhane, K.; Gilliland, F., London, S. J.; Islam, T.; Gauderman, W. J.; Avol, E.; Margolis, H. G.;
Peters, J, M. (2002) Asthma in exercising children exposed to ozone: a cohort study. Lancet 359: 386-391.
McDonnell, W. F. (1996) Individual variability in human lung function responses to ozone exposure. Environ.
Toxicol. Pharmacol. 2: 171-175.
McDonnell, W. F.; Abbey, D. E.; Nishino, N.; Lebowitz, M. D. (1999) Long-term ambient ozone concentration and
the incidence of asthma in nonsmoking adults: the ahsmog study. Environ. Res. 80:110-121.
McDonnell, W. F.; Stewart, P. W.; Andreoni, S.; Seal, E., Jr.; Kehrl, H. R.; Horstman, D. H.; Folinsbee, L. J.; Smith,
M. V. (1997) Prediction of ozone-induced FEV1 changes: effects of concentration, duration, and ventilation.
Am. J. Respir. Crit. Care Med. 156: 715-722.
Molfino,N. A.; Wright, S. C.; Katz, L; Tarlo, S.; Silverman, F.; McClean, P. A.; Szalai, J. P.; Raizenne, M.;
Slutsky, A. S.; Zamel, N. (1991) Effect of low concentrations of ozone on inhaled allergen responses in
asthmatic subjects. Lancet 338(8761): 199-203.
Moolgavkar, S. H.; Luebeck, E. G.; Hall, T. A; Anderson, E. L. (1995) Air pollution and daily mortality in
Philadelphia. Epidemiology 6: 476-484.
Moolgavkar, S, H.; Luebeck, E. G.; Anderson, E. L. (1997) Air pollution and hospital admissions for respiratory
causes in Minneapolis-St. Paul and Birmingham. Epidemiology 8: 364-370.
Mortimer, K. M.; Neas, L. M.; Dockery, D. W.; Redline, S.; Tager, I. B. (2002) The effect of air pollution on inner-
city children with asthma. Eur. Respir. J. 19: 699-705.
Mudway, I. S.; Kelly, F. J. (2004) An investigation of inhaled ozone dose and the magnitude of airway inflammation
in healthy adults. Am. J. Respir. Crit. Care Med. 169:1089-1095.
National Institutes of Health. (1997) Guidelines for the diagnosis and management of asthma: expert panel report 2.
Bethesda, MD: U.S. Department of Health and Human Services, National Heart, Lung, and Blood Institute;
publication no. 97-4051. Available: http://www.nhlbi.nih.gov/guidelines/asthma/asthgdm.pdf (11 April
2003).
Navidi, W.; Thomas, D.; Langholz, B.; Stram, D. (1999) Statistical methods for epidemiologic studies of the health
effects of air pollution. Cambridge, MA: Health Effects Institute; research report no. 86.
November 2005
3-81
Draft - Do Not Cite or Quote
-------
1 Newhouse, C. P.; Levetin, B. S.; Levetin, E. (2004) Correlation of environmental factors with asthma and rhinitis
2 symptoms in Tulsa, OK. Ann. Allergy Asthma Immunol. 92: 356-366.
3
4 O'Neill, M. S.; Loomis, D.; Borja-Aburto, V. H. (2004) Ozone, area social conditions, and mortality in Mexico City.
5 Environ. Res. 94: 234-242.
6
7 Ostro, B. (1995) Fine participate air pollution and mortality in two Southern California counties. Environ. Res.
8 70:98-104.
9
10 Ostro, B.; Lipsett, M.; Mann, J.; Braxton-Owens, H.; White, M. (2001) Air pollution and exacerbation of asthma in
11 African-American children in Los Angeles. Epidemiology 12: 200-208.
12
13 Ostro, B. D.; Broadwin, R.; Lipsett, M. J. (2003) Coarse particles and daily mortality in Coachella Valley,
14 California. In: Revised analyses of time-series studies of air pollution and health. Special report. Boston, MA:
15 Health Effects Institute; pp. 199-204. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdfI18
16 October, 2004].
17
18 Park, S. K.; O'Neill, M. S.; Vokonas, P. S.; Sparrow, D.; Schwartz, J. (2005) Effects of air pollution on heart rate
19 variability: the VA normative aging study. Environ. HealthPerspect. 113: 304-309.
20
21 Park, H.; Lee, B.; Ha, E.-H.; Lee, J.-T.; Kim, H.; Hong, Y.-C. (2002) Association of air pollution with school
22 absenteeism due to illness. Arch. Pediatr. Adolesc. Med. 156:1235-1239.
23
24 Peden, D. B.; Boehlecke, B.; Horstman, D.; Devlin, R. (1997) Prolonged acute exposure to 0.16 ppm ozone induces
25 eosinophilic airway inflammation in asthmatic subjects with allergies. J. Allergy Clin. Immunol. 100:
26 802-808.
27
28 Peden, D. B.; Setzer, R. W., Jr.; Devlin, R. B. (1995) Ozone exposure has both a priming effect on allergen-induced
29 responses and an intrinsic inflammatory action in the nasal airways of perennially allergic asthmatics. Am. J.
30 Respir. Grit. Care Med.' 151:1336-1345.
31
32 Peel, J. L.; Tolbert, P. E.; Klein, M.; Metzger, K. B.; Flanders, W. D.; Knox, T.; Mulholland, J. A.; Ryan, P. B.;
33 Frumkin, H. (2005) Ambient air pollution and respiratory emergency department visits. Epidemiology
34 16: 164-174.
35
36 Peters, J. M.; Avol, E.; Navidi, W.; London, S. J.; Gauderman, W. J.; Lurmann, F.; Linn, W. S.; Margolis, H.;
37 Rappaport, E.; Gong, H., Jr.; Thomas, D. C. (1999a) A study of twelve southern California communities with
38 differing levels and types of air pollution. I. Prevalence of respiratory morbidity. Am. J. Respir. Crit Care
39 Med. 159: 760-767.
40
41 Peters, J. M.; Avol, E.; Gauderman, W. J.; Linn, W. S.; Navidi, W,; London, S. J.; Margolis, H.; Rappaport, E.;
42 V°ra, H.; Gong, H., Jr.; Thomas, D. C. (1999b) A study of twelve southern California communities with
43 differing levels and types of air pollution. II. Effects on pulmonary function. Am. J. Respir. Crit Care Med. •
44 159:768-775.
45
46 Peters, A.; Dockery, D. W.; Muller,, J. E.; Mittleman, M. A. (2001) Increased particulate air pollution and the
47 triggering of myocardial"infarction. Circulation 103: 2810-2815.
48
49 Petroeschevsky, A.; Simpson, R. W.; Thalib, L.; Rutherford, S. (2001) Associations between outdoor air pollution
50 and hospital admissions in Brisbane, Australia. Arch. Environ. Health 56: 37*52.
51 . '
52 Plopper, C. G.; Hatch, G. E.; Wong, V.; Duan, X.; Weir, A. J.; Tarkington, B. K.; Devlin, R. B.; Becker, S.;
53 Buckpitt, A. R. (1998) Relationship of inhaled ozone concentration to acute tracheobronchial epithelial
54 injury, site-specific ozone dose and glutathione depletion in rhesus monkeys. Am. J. Respir. Cell Mol. Biol.
55 19:387-399.
56
November 2005 3-82 Draft - Do Not Cite or Quote
-------
1 Plunkett, L. M.; Tumbull, D.; Rodricks, J. V. (1992) Differences between adults and children affecting exposure
2 assessment. In: Guzelian, P. S.; Henry, D. J.; Olin, S. S., eds. Similarities and differences between children
3 and adults: implications for risk assessment. Washington, DC: ILSI Press, pp. 79-96.
4
5 Ponce de Leon, A.; Anderson, H. R.; Bland, J. M.; Strachan, D. P.; Bower, J. (1996) Effects of air pollution on daily
6 hospital admissions for respiratory disease in London between 1987-88 and 1991-92. In: St Leger, S., ed.
7 The APHEA project. Short term effects of air pollution on health: a European approach using
8 epidemiological time series data. J. Epidemiol, Common. Health 50(suppl. 1): S63-S70.
9
10 Pope, C. A., Ill; Thun, M. J.; Namboodiri, M. M.; Dockery, D. W.; Evans, J. S.; Speizer, F. E.; Heath, C. W., Jr.
11 (1995) Particulate air pollution as a predictor of mortality in a prospective study of U.S. adults. Am. J. Respir.
12 Crit. Care Med. 151: 669-674.
13 . .
14 Pope, C. A, III; Burnett, R. T.; Thun, M. J.; Calle, E. E.; Krewski, D.; Ito, K.; Thurston, G. D. (2002) Lung cancer,
15 cardiopulmonary mortality, and long-term exposure to fine particulate air pollution. JAMA J. Am. Med.
16 Assoc. 287: 1132-1141.
I?
18 Raizenne, M.; Stem, B.; Burnett, R.; Spengler, J. (1987) Acute respiratory function and transported air pollutants:
19 observational studies. Presented at: 80th annual meeting of the Air Pollution Control Association; June;
20 New York, NY. Pittsburgh, PA: Air Pollution Control Association; paper no. 87-32.6.
21
22 Raizenne, M. E.; Burnett, R. T.; Stem, B.; Franklin, C. A.; Spengler, J. D. (1989) Acute lung function responses to
23 ambient acid aerosol exposures in children. Environ. Health Perspect. 79:179-185.
24
25 Rich, D. Q.; Schwartz, J.; Mittleman, M. A.; Link, M.; Luttmann-Gibson, H.; Catalano, P. J.; Speizer, F. E.;
26 Dockery, D. W. (2005) Association of short-term ambient air pollution concentrations and ventricular
27 arrhythmias. Am. J. Epidemiol. 161: 1123-1132.
28
29 Romieu, L; Meneses, F.; Ruiz, S.; Sienra, J. J.; Huerta, J.; White, M. C.; Etzel, R. A. (1996) Effects of air pollution
30 on the respiratory health of asthmatic children living in Mexico City. Am. J. Respir. Crit. Care Med.
31 154:300-307.
32 ,
33 Romieu, I.; Meneses, F.; Ruiz, S.; Huerta, J.; Sienra, J. J.; White, M.; Etzel, R; Hernandez, M. (1997) Effects of
34 intermittent ozone exposure on peak expiratory flow and respiratory symptoms among asthmatic children in
35 Mexico City. Arch. Environ. Health 52: 368-376.
36 .
37 Romieu, I.; Meneses, F.; Ramirez, M.; Ruiz, S.; Padilla, R. P.; Sienra, J. J.; Gerber, M.; Grievink, L.; Dekker, R.;
38 . Walda, I.; Brunekreef, B. (1998) Antioxidant supplementation and respiratory functions among workers
39 exposed to high levels of ozone. Am. J. Respir. Crit. Care Med. 158: 226-232.
40
41 Romieu, I.; Sienra-Monge, J. J.; Ramirez-Aguilar, M.; Tellez-Rojo, M M.; Moreno-Macias, H.; Reyes-Ruiz,N. I.;
42 Del Rio-Navarro, B. E.; Ruiz-Navarro, M. X,; Hatch, G.; Slade, R.; Hemandez-Avila, M. (2002) Antioxidant
43 supplementation and lung functions among children with asthma exposed to high levels of air pollutants.
44 Am. J. Respir. Crit. Care Med. 166: 703-709.
45
46 Romieu, I.; Sienra-Monge, J. J.; Ramirez-Aguilar, M.; Moreno-Macias, H.; Reyes-Ruiz, N. I.; Estela del Rio-
47 Navarro, B.; Hemandez-Avila, M.; London, S. J. (2004) Genetic polymorphism of GSTM1 and antioxidant
48 supplementation influence lung function in relation to ozone exposure in asthmatic children in Mexico City.
49 Thorax 59: 8-10.
50
51 Ross, M. A.; Persky, V. W.; Scheff, P. A.; Chung, J.; Curtis, L.; Ramakrishnan, V.; Wadden, R. A; Hryhorczuk,
52 D. O. (2002) Effect of ozone and aeroallergens on the respiratory health of asthmatics. Arch. Environ. Health
53 57:568-578.
54
55 Ruidavets, J.-B.; Coumot, M.; Cassadou, S.; Giroux, M.; Meybeck, M.; Ferri res, J. (2005) Ozone air pollution is
56 associated with acute myocardial infarction. Circulation 111: 563-569.
November 2005 3-83 Draft - Do Not Cite or Quote
-------
1
2 Samet, J. M.; Zeger, S. L.; Dominici, F.; Curriero, F.; Coursac, I.; Dockery, D. W,; Schwartz, J.; Zanobetti, A.
3 (2000) The national morbidity, mortality, and air pollution study. Part II:' morbidity, mortality, and air
4 pollution in the United States. Cambridge, MA: Health Effects Institute; research report no. 94, part II.
6 Santak, B.; Calzia, E.; Radermacher, P. (1998) Ventilation-perfusion relationships: clinical implications. Acta
7 Anesthio}. Scand. 42[Suppl 112]: 18-20.
8
9 Sartor, F.; Snacken, R.; Demuth, C.; Walckiers, D. (1995) Temperature, ambient ozone levels, and mortality during
10 summer, 1994, inBelgium. Environ. Res. 70:105-113.
11
12 Scannell, C.; Chen; L.; Aris, R. M.; Tager, I.; Christian, D.; Ferrando, R.; Welch, B.; Kelly, T.; Balmes, J. R. (1996)
13 Greater ozone-induced inflammatory responses in subjects with asthma. Am. J. Respir. Crit. Care Med.
14 154:24-29.
15
16 Scarlett, J. F.; Abbott, K. J.; Peacock, J. L.; Strachan, D. P.; Anderson, H. R. (1996) Acute effects of summer air
17 pollution on respiratory function in primary school children in southern England. Thorax 51:1109-1114.
18
19 Schelegle, E. S.; Miller, L. A.; Gershwin, L. J.; Fanucchi, M. V.; Van Winkle, L. S.; Gerriets, J. E.; Walby, W. F.;
20 Mitchell, V.; Tarkington, B. K.; Wong, V. J.; Baker, G. L.; Pantle, L. M.; load, J. P.; Pinkerton, K. E.;
21 Wu, R.; Evans, M. J.; Hyde, D. M.; Plopper, C. G. (2003) Repeated episodes of ozone inhalation amplifies
22 the effects of allergen sensitization and inhalation on airway immune and structural development in Rhesus
23 monkeys. Toxicol. Appl. Pharmacol. 191: 74-85.
24
25 Schelegle, E. S.; Adams, W. C. (1986) Reduced exercise time in competitive simulations consequent to low level'
26 ozone exposure. Med. Sci. Sports Exercise 18: 408-414.
27
28 Schindler, C.; Kunzli, N.; Bongard, J.-P.; Leuenberger, P.; Karrer, W.; Rapp, R.; Monn, C.; Ackermann-Liebrich,
29 U.; Swiss Study on Air Pollution and Lung Diseases in Adults Investigators. (2001) Short-term variation in
30 air pollution and in average lung function among never-smokers. Am. J. Respir. Crit. Care Med. 163: 356-
31 361.
32 Schwartz, J. (2005) How sensitive is the association between ozone and daily deaths to control for temperature?
33 Am. J. Respir. Crit. Care Med. 171: 627-631.
34
35 Schwartz, J.; Spix, C.; Touloumi, G.; Bacharova, L.; Barumamdzadeh, T.; le Tertre, A.; Piekarksi, T.;
36 Ponce de Leon, A.; POnkS, A.; Rossi, G.; Saez, M.; Schouten, J. P. (1996) Methodological issues in studies of
37 air pollution and daily counts of deaths or hospital admissions. In: St Leger, S., ed. The APHEA project.
38 Short term effects of air pollution on health: a European approach using epidemiological time series data.
39 J. Epidemiol. Commun. Health 50(suppl. 1): S3-S11.
40
41 Seal, E., Jr.; McDonnell, W. F.; House, D. E. (1996) Effects of age, socioeconomic status, and menstrual cycle on
42 pulmonary response to ozone. Arch. Environ. Health 51: 132-137.
43
44 Sheppard, L.; Levy, D.; Morris, G.; Larson, T. V.; Koenig, J. Q. (1999) Effects of ambient air pollution on
45 nonelderly asthma hospital admissions in Seattle, Washington, 1987-1994. Epidemiology 10: 23-30.
46
47 Sheppard, L. (2003) Ambient air pollution and nonelderly asthma hospital admissions in Seattle, Washington, 1987-
48 1994. In: Revised analyses of time-series studies of air pollution and health. Special report. Boston, MA:
49 Health Effects Institute; pp. 227-230. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdf
50 [18 October, 2004].
51
52 Sherwin, R. P.; Richters, V.; Kraft, P.; Richters, A. (2000) Centriacinar region inflammatory disease in young
53 individuals: a comparative study of Miami and Los Angeles residents. Virchows Arch. 437: 422-428.
54
55 Simpson, R. W.; Williams, G.; Petroeschevsky, A.; Morgan, G.; Rutherford, S. (1997) Associations between
56 outdoor air pollution and daily mortality in Brisbane, Australia. Arch. Environ. Health 52: 442-454.
November 2005 3-84 Draft - Do Not Cite or Quote
-------
1
2 Spektor, D. M.; Lippmann, M.; Lioy, P. J.; Thurston, G. D.; Citak, K.; James, D. J.; Bock, N.; Speizer,.F. E.;
3 Hayes, C. (1988a) Effects of ambient ozone on respiratory function in active, normal children. Am. Rev.
4 Respir. Dis. 137: 313-320.
5
6 Spektor, D. M.; Lippmann, M.; Thurston, G. D.; Lioy, P. J.; Stecko, J.; O'Connor, G.; Garshick, E.; Speizer, F. E.;
7 Hayes, C. (1988b) Effects of ambient ozone on respiratory function in healthy adults exercising outdoors.
8 Am. Rev. Respir. Dis. 138: 821-828.
9
10 Spektor, D. M.; Lippmann, M. (1991) Health effects of ambient ozone on healthy children at a summer camp. In:
11 Berglund, R. L.; Lawson, D. R.; McKee, D. J., eds. Tropospheric ozone and the environment: papers from an
12 international conference; March 1990; Los Angeles, CA. Pittsburgh, PA: Air & Waste Management
13 Association; pp. 83-89. (A&WMA transactions series no. TR-19).
14
15 Stenfors, N.; Pourazar, J.; Blomberg, A.; Krishna, M. T.; Mudway, I.; Helleday, R.; Kelly, F. J.; Frew, A. J.;
16 Sandstrflm, T. (2002) Effect of ozone on bronchial mucosal inflammation in asthmatic and healthy subjects.
17 Respir. Med. 96: 352-358.
18
19 Stemer-Kock, A.; Kock, M.; Braun, R.; Hyde, D. M. (2000) Ozone-induced epithelial injury in the ferret is similar
20 tononhumanprimates. Am. J.Respir. Crit. Care Med. 162:1152-1156.
21 .
22 Stieb, D. M.; Burnett, R. T.; Beveridge, R. C.; Brook, J. R. (1996) Association between ozone and asthma
23 emergency department visits in Saint John, New Brunswick, Canada. Environ. Health Perspect 104:13 54-
24 1360.
25
26 Sunyer, J.; Basaga a, X.; Belmonte, J.; Anto, J. M. (2002) Effect of nitrogen dioxide and ozone on the risk of dying
27 in patients with severe asthma. Thorax 57: 687-693.
28
29 Tager, I. B.; Kunzli, N.; Lurmann, F.; Ngo, L.; Segal, M.; Balmes, J. (1998) Methods development for
30 epidemiological investigations of the health effects of prolonged ozone exposure. Part II: an approach to
31 retrospective estimation of lifetime ozone exposure using a questionnaire and ambient monitoring data
32 (California sites). Cambridge, MA: Health Effects Institute; research report no. 81; pp. 27-78.
33
34 Taggart, S. C. O.; Custovic, A.; Francis, H. C.; Faragher, E. B.; Yates, C. J.; Higgins, B. G.; Woodcock, A. (1996)
35 Asthmatic bronchial hyperresponsiveness varies with ambient levels of summertime air pollution. Eur.
36 Respir. 19:1146-1154.
37
38 Tenias, J. M.; Ballester, F.; Rivera, M. L. (1998) Association between hospital emergency visits for asthma and air
39 pollution in Valencia, Spain. Occup. Environ. Med. 55: 541-547.
40
41 Tenias, J. M.; Ballester., F.; Perez-Hoyos, S.; Rivera, M. L. (2002) Air pollution and hospital emergency room
42 admissions for chronic obstructive pulmonary disease in Valencia, Spain. Arch. Environ. Health 57: 41-47.
43
44 Thurston, G. D.; Ito, K.; Kinney, P. L.; Lippmann, M. (1992) A multi-year study of air pollution and respiratory
45 hospital admissions in three New York State metropolitan areas: results for 1988 and 1989 summers.
46 J. Exposure Anal. Environ. Epidemiol. 2: 429-450.
47
48 Thurston, G. D.; Ito, K.; Hayes,-C. G.; Bates, D. V.; Lippmann, M. (1994) Respiratory hospital admissions and
49 summertime haze air pollution in Toronto, Ontario: consideration of the role of acid aerosols. Environ. Res.
50 65:271-290.
51
52 Thurston, G. D.; Lippmann, M.; Scott, M. B.; Fine, J. M. (1997) Summertime haze air pollution and children with
53 asthma. Am. J. Respir. Crit Care Med. 155: 654-660.
54
55 Tobias, A.; Campbell, M. J.; Saez, M. (1999) Modelling asthma epidemics on the relationship between air pollution
56 and asthma emergency visits in Barcelona, Spain. Eur. J. Epidemiol. 15: 799-803.
November 2005 3-85 Draft - Do Not Cite or Quote
-------
1
2 Tolbert, P. E.; Mulholland, J. A.; Macintosh, D. L.; Xu, F.; Daniels, D.; Devine, O. I; Carlin, B. P.; Klein, M;
3 Dorley, J.; Butler, A. J.; Nordenberg, D. F.; Frumkin, H.; Ryan, P. B,; White, M. C. (2000) Air quality and
4 pediatric emergency room visits for asthma in Atlanta, Georgia, Am. J. Epidemiol. 151: 798-810.
5
6 Touloumi, G,; Katsouyanni, K.; Zmirou, D.; Schwartz, I; Spix, C.; Ponce de Leon, A.; Tobias, A.; Quennel, P.;
7 Rabczenko, D.; Bacharova, L.; Bisanti, L.; Vonk, J. M.; Ponka, A. (1997) Short-term effects of ambient
8 oxidant exposure onmortality: a combined analysis within the APHEA project. Am. J. Epidemiol. 146:177-
9 185.
10
11 Trenga, C. A.; Koenig, J. Q.; Williams, P. V. (2001) Dietary antioxidants and ozone-induced bronchial
12 hyperresponsiveness in adults with asthma. Arch. Environ. Health 56: 242-249.
13
14 Tsai, J.-J.; Lin, Y.-C.; Kwan, Z.-H.; Kao, H.-L. (1998) Effects of ozone on ovalbumin sensitization in guinea pigs
15 J. Microbiol. Imniunol. Infect. 31: 225-232.
16
17 Tsai, B. M.; Wang, M.; Pitcher, J. M.; Meldrum, K. K.; Meldrum, D. R. (2004) Hypoxic pulmonary
18 vasoconstriction and pulmonary artery tissue' cytokine expression are mediated by protein kinase C. Am. J.
19 Physiol. Lung Cell Mol. Physiol. 287: L1215-L1219.
20
21 Tyler, W. S.; Tyler, N. K.; Last, J. A; Gillespie, M. J.; Barstow, T. J. (1988) Comparison of daily and seasonal
22 exposures of young monkeys to ozone. Toxicology 50:131-144.
23
24 U.S. Environmental Protection Agency. (1996) Air quality criteria for ozone and related photochemical oxidants.
25 Research Triangle Park, NC: Office of Research and Development; report nos. EPA/600/AP-93/004aF-cF.
26 3v. Available from: NTIS, Springfield, VA; PB96-185582, PB96-185590, and PB96-185608. Available:
27 http://cfpub2.epa.gov/ncea/.
28
29 U.S. Environmental Protection Agency. (2004) Air quality criteria for particulate matter. Research Triangle Park,
30 NC: National Center for Environmental Assessment; report no. EPA/600/P-99/002aF-bF. 2v. Available:
31 http ://www.epa.gov/ttn/naaqs/standards/pm/s_pm_cr_cd.html [9 November, 2004].
32
33 U.S. Environmental Protection Agency (2005) Air quality criteria for ozone and related photochemical oxidants
34 (second external review draft) Research Triangle Park, NC: National Center for Environmental Assessment;
35 report no. EPA/600R-05/004aB-cB, 3v. Available:
36 http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=l37307 [August 2005]
37
38 Ulmer, C.; Kopp, M.; Ihorst, G.; Frischer, T.; Forster, J.; Kuehr, J. (1997) Effects of ambient ozone exposures
39 during the spring and summer of 1994 on pulmonary function of schoolchildren. Pediatr. Pulmonol. 23: 344-
40 353.
41
42 Ultman, J. S.; Ben-Jebria, A.; Arnold, S. F. (2004) Uptake distribution of ozone in human lungs: intersubject
43 variability in physiologic response. Boston, MA: Health Effects Institute.
44
45 Vedal, S.; Brauer, M.; White, R.; Petkau, J. (2003) Air pollution and daily mortality in a city with low levels of
46 pollution. Environ. Health Perspect 111: 45-51.
47
48 Wattiez, R.; Noel-Georis, I.; Cruyt, C.; Broeckaert, F,; Bernard, A.; Falmagne, P. (2003) Susceptibility to oxidative
49 stress: proteomic analysis of bronchoalveolar lavage from ozone-sensitive and ozone-resistant strains of mice.
50 Proteomics 3: 658-665.
51
52 Weisel, C. P.; Cody, R. P.; Lioy, P. J. (1995) Relationship between summertime ambient ozone levels and
53 emergency department visits for asthma in central New Jersey. Environ. Health Perspect. 103(suppl. 2): 97-
54 102.
55
November 2005 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
Weisel, C. P.; Cody, R. P.; Georgopoulos, P. G.; Purushothaman, V.; Weiss, S. H.; Bielory, L.; Gregory, P.; Stem,
A. H. (2002) Concepts in developing health-based indicators for ozone. Int. Arch. Occup. Environ. Health
75: 415-422.
White, M. C.; Etzel, R. A.; Wilcox, W. D.; Lloyd, C. (1994) Exacerbations of childhood asthma and ozone pollution
in Atlanta. Environ. Res. 65:56-68. - .
Wiley, J. A.; Robinson, J. P.; Piazza, T.; Garrett, K.; Cirksena, K.; Cheng, Y.-T.; Martin, G. (199la) Activity
patterns of California residents. Final report. Sacramento, CA: California Air Resources Board; report no.
ARB/R93/487. Available from: NTIS, Springfield, VA.; PB94-108719.
Wiley, J. A.; Robinson, J. P.; Cheng, Y.-T.; Piazza, T.; Stork, L.; Pladsen, K. (1991b) Study of children's activity
patterns: final report. Sacramento, CA: California Air Resources Board; report no. ARB-R
Wolff, G.T. (1995) Letter to EPA Administrator Carol Browner: "CASAC Closure on the Primary Standard
Portion of the Staff Paper for Ozone" EPA-SAB-CASAC-LTR-96-002, November 30,1995.
Yang, Q.; Chen, Y.; Shi, Y.; Burnett, R. T.; McGrail, K. M.; Krewski, D. (2003) Association between ozone and
respiratory admissions among children and the elderly in Vancouver, Canada. Inhalation Toxicol. 15: 1297-
1308.
Zidek, J. V. (1997) Interpolating air pollution for health impact assessment. In: Barnett, E. V.; Turkman, K. F., eds.
Pollution Assessment and Control. New York, NY: John Wiley & Sons. (Statistics for the Environment,
no. 3).
November 2005
3-87
Draft - Do Not Cite or Quote
-------
1
2 4. CHARACTERIZATION OF HUMAN EXPOSURE TO OZONE
3 4.1 INTRODUCTION
4 As part of the last 63 NAAQS review, EPA conducted exposure analyses for the general
5 population, children who spent more time outdoors, and outdoor workers. Exposure estimates
6 were generated for nine urban areas for "as is" (i.e., a recent year) air quality and for just meeting
7 the existing 1-hr standard and several alternative 8-hr standards. EPA also conducted a health
8 risk assessment that produced risk estimates for the number of children and percent of children
9 experiencing lung function and respiratory symptoms associated with the exposures estimated
10 for these same nine urban areas.
11 The exposure analysis conducted for the current review builds upon the methodology and
12 lessons learned from the exposure analyses conducted for the last review (US EPA, 1996a). The
13 methodology used to conduct the exposure analysis as well as summary results from the
14 exposure analysis are described in this chapter. The exposure analysis technical support
15 document, Ozone Population Exposure Analysis for Selected Urban Areas (US EPA, 2005a)
16 (hereafter cited as "draft Exposure Analysis TSD") presents a more detailed description of the
17 exposure analysis methodology,
18 Population exposures to ambient Os levels are modeled for 12 urban areas located across
19 the U. S. using the Air Pollutants Exposure (APEX) model, also referred to as the Total Risk
20 Integrated Methodology Inhalation Exposure (TRIM. Expo) model. Exposure estimates are
21 developed for current Ojlevels, based on 2004 ambient air quality measurements, and for Oj
22 levels associated with just meeting the current 8-hr Oa NAAQS, based on adjusting the 2004 air
23 quality data. Exposures to background levels of O3 are also estimated, based on Oa
24 concentrations predicted by the GEOS-CHEM atmospheric photochemical model. The model
25 estimated the PRB and total ozone concentrations at each grid point and the grid points selected
26 to represent the 12 urban areas used in the exposure analysis are included in Table 4D-1 in
27 Appendix 4D. ,
28 Exposures are modeled for 1) the general population, 2) all school-age children, and 3)
29 active school-age children (defined below). The strong emphasis on children reflects the rinding
30 of the last O3 NAAQS-review that children, especially those who are active outdoors, are an
31 important at-risk group. An assessment of exposures of asthmatic school-age children will be
32 presented in the next draft Os Staff Paper.
November 2005 4-1 Draft - Do Not Quote or Cite
-------
1 This chapter provides a brief overview of the types of studies that provide data which this
2 analysis is based on, followed by a description of the exposure model used for this analysis, the
3 model input data, and the results of the analysis.
4 4.2 OZONE EXPOSURE STUDIES
5 Many studies have produced information and data supporting the development of
6 methods for estimating human exposure to ambient Os over the past several decades. These
7 studies have been reviewed in the current and previous EPA Ozone Air Quality Criteria
8 Documents (US EPA, 1986,1996b, 2005b). :
9 The types of measurements that have proven to be useful for understanding and
10 estimating exposure obtained from personal exposure assessment studies include fixed-site
11 ambient concentrations, concentrations in specific indoor and outdoor microenvironments,
12 personal exposure measurements, personal activity patterns, air exchange rates, infiltration rates,
13 deposition and decay rates, and meteorology.
14 4.2.1 Exposure Concepts and Definitions
15 Human exposure to a contaminant is defined as "contact at a boundary between a human
16 and the environment at a specific contaminant concentration for a specific interval of time"
17 (National Research Council, 1991). For airborne pollutants the contact boundary is nasal and
18 oral openings in the body, and personal exposure of an individual to a chemical in the air for a
19 discrete time period is quantified as (Lioy, 1990; National Research Council, 1991):
20 £[*,,'3]= . ^,(t)ftt (4-1)
"i
21 where Eft,^] is the personal exposure during the time period from 11 to ?2, and C(0 is the
22 concentration at time t in the breathing zone. The breathing rate (ventilation rate) at the time of
23 exposure is also an important determinant of the dose received by the individual.
24 Personal exposure to Oa can be estimated directly, by monitoring the concentration of Os
25 in the person's breathing zone (close to the nose/mouth), using a personal exposure monitor
26 (PEM). Exposure can also be estimated indirectly, by estimating or monitoring the
27 concentrations over time in locations in which the individual spends time and estimating the time
28 and duration the individual spends in each location. In both of these methods, Equation 4-1 is
29 used to calculate an estimate of personal exposure.
30 A key concept in modeling exposure is the microenviroranent, a term that refers to the
31 immediate surroundings of an individual. A microenvironment is a location in which pollutant
32 concentrations are relatively homogeneous for short periods of time. Microenvironments can be
November 2005 4-2 Draft - Do Not Quote or Cite
-------
1 outdoors or indoors; some examples are outdoors near the home, outdoors near the place of
2 work, bedrooms, kitchens, vehicles, stores, restaurants, street-comer bus stops, schools, and
3 places of work. A bedroom may be treated as a different microenvironment than a kitchen if the
4 concentrations are significantly different in the two rooms. The concentrations in a
5 microenvironment typically change over time; for example, Os concentrations in a kitchen while
6 cooking with a gas stove may be lower than when these activities are not being performed, due to
7 scavenging of Os by NOX emissions from the gas burned.
8 An important factor affecting the concentrations of Oa indoors is the degree to which the
9 ambient outdoor air is transported indoors. This can be estimated using physical factors such as
10 air exchange rates, deposition and decay rates, and penetration factors. The volumetric exchange
11 rate (m3/hour) is the rate of air exchange between the indoor and outdoor air. The air exchange
12 rate (AER) between indoors and outdoors is the number of complete air exchanges per hour and
13 is equal to the volumetric exchange rate divided by the volume of the well-mixed indoor air.
14 Indoor concentrations of Os can be decreased by uptake of 63 by surfaces and by chemical
15 reactions. The deposition and decay rates are the rates (per hour) at which Os is removed from'
16 the air by surface uptake and chemical reactions. Some exposure models employ an infiltration
17 factor, which is conceptually useful if distinguishing between the air exchange processes of air
18 blowing through open doors and windows and the infiltration of air through smaller openings.
19 Since measurements of air exchange rates account for all of these processes (including
20 "infiltration"), this distinction is not useful in applied modeling of Os exposures and will not be
21 discussed further here. Simplistic exposure models use a "factor model" approach to estimate
22 indoor Os concentrations by multiplying the ambient outdoor concentrations by an
23 indoor/outdoor concentration ratio, referred to as a penetration factor.
24 4.2.2 Monitoring Equipment Considerations
25 Exposure assessment studies involve monitoring airborne Os and/or other pollutants, and
26 monitor design and placement play a critical role in interpreting the results of these studies. For
27 exposure assessment purposes there are two general classes of monitors, personal exposure
28 monitors (PEMs) and fixed site monitors.
29 PEMs are designed to be worn or carried easily by individuals and to measure the
30 concentrations experienced by individuals over a period of hours, days, or weeks. The
31 placement of PEMs is important; the desired placement is usually in the breathing zone near the
32 mouth and nose, but where the monitor will not be excessively impacted by exhaled air. This
33 placement is intended to represent the concentrations the individual breathes in. PEMs typically
34 report continuously measured Os concentrations with averaging times ranging from 1 to 24
35 hours.
November 2005 4-3 Draft - Do Not Quote or Cite
-------
1 The draft CD reviews 03 PEMS (draft CD Appendix AX3, p. 163-4) and notes that
2 humidity, wind velocity, badge placement, and interference with other copollutants may result in
3 measurement error. The draft CD reports PEM detection limits ranging from 5 to 23 ppb for
4 averaging times from 24 hours to 1 hour.
5 Fixed-site monitors measure concentrations over time at a given location. There are
6 numerous fixed-site O3 monitors which are part of national, state, and local air monitoring
7 networks. In addition to their role of being used to determine which areas are in compliance with
8 existing Os NAAQS, these are also useful for alerting the public to high O^ days, providing air
9 quality data in support of exposure assessments for a study area, for tracking Os levels and
10 trends, and for studying the representativeness of measurements at these monitors for the study.
11 area. Existing fixed-site monitors usually report hourly averaged concentrations, and are in
12 operation over a period of years. A discussion of monitoring equipment and networks can be
13 found in Chapter 2 of this Staff Paper and in section 2.6 in the draft CD.
14 There are also stationary monitors expressly set up for particular exposure field studies.
15 These are used to measure concentrations over time in microenvironments, such as rooms in a
16 home, just outside a home, roadsides, and so forth. The stationary monitors which are outdoors
17 can provide information about community-scale representativeness of fixed-site monitors in or
18 near the community.
19 4.2.3 Personal Ozone Exposure Assessment Studies
20 The most useful PEM studies have data collected repeatedly from each individual in the
21 study over a period of time, yielding a longitudinal time series of concentrations each individual
22 is exposed to. These studies permit analysis of both the temporal and spatial variability of each
23 person's personal exposure to Os.
24 Some studies are designed so that the data are sampled randomly from the population,
25 which reduces bias and allows one to make inferences about exposure in the broader population.
26 Most studies addressing Os exposure have not been random. They might have specific goals for
27 which randomness is not required, or be subject to constraints which do not allow for random
28 sampling. Some studies draw upon data from existing measurement systems or historical data
29 collection efforts. These non-random studies can be very helpful in the development of models
30 of exposure; however, their use must recognize that they may not be.representative of the
31 broader population.
32 The draft CD summarizes results from several personal exposure studies that measured
33 Os conducted in the U.S., Canada, and France (draft CD, p. 3-60 to 3-61, Appendix AX3, p. 185-
34 189).
November 2005
4-4
Draft - Do Not Quote or Cite
-------
1 4.2.4 Microenvironmental Studies
2 The focus of microenvironmental studies is on measuring concentrations in different
3 locations that people spend time in, as well as on measuring the movement of pollutants from
4 one microenvironment to another and on measuring other parameters that contribute to
5 variability in exposure. Typically, microenvironmental measurements include indoor and
6 outdoor concentrations of Os and other pollutants, air exchange rates, infiltration factors,
7 deposition rates, decay rates, emissions of Oa, NOX, VOCs, and other pollutants, operating
8 characteristics of air conditioning systems, and meteorological data such as wind velocity,
9 temperature, and humidity. The draft CD discusses several studies of microenvironments that
10 contribute to our understanding of the factors and processes that affect exposure to Os (draft CD
11 Appendix AX3, p. 190-216).
12 There is a great deal of variability among individuals in the amount of time spent indoors,
13 but the majority of people spend most of their time indoors (Graham & McCurdy, 2004), and
14 therefore the concentrations of Os indoors can be an important determinant of people's exposure
15 to Os. There are several factors affecting Os concentrations indoors. The ambient outdoor
16 concentration of Os and the air exchange rate are the primary determinants of the indoor
17 concentrations. Removal processes are also significant, the most important of which is
18 deposition onto indoor surfaces such as carpets, furnishings, and ventilation ductwork. Chemical
19 reactions of 63 with other compounds, such as solvents from consumer products or NOX
20 emissions from gas stoves, also deplete Os indoors.
21 There are very few sources of O3 indoors. The draft CD reports on 63 emissions from
22 ' photocopiers and from home/office Os generators. Some older photocopiers, if run continuously
23 in an enclosed area, can increase O3 concentrations by as much as 20 ppb. Ozone generators can
24 increase indoor concentrations by more than 200 ppb.
25 4.3 EXPOSURE MODELING
26 Models of human exposure to airborne pollutants are typically driven by estimates of
27 ambient outdoor concentrations of the pollutants, which vary by time of day as well as by
28 location. These concentration estimates may be provided by measurements, by air quality
j
29 models, or by a combination of these. It is only possible to address hypothetical future scenarios
30 using modeling. The main purpose of this exposure analysis is to allow comparisons of
31 population exposures to Os within each urban area, associated with current air quality levels and
32 with alternative air quality standards or scenarios. Human exposure, regardless of the pollutant,.
33 depends on where an individual is located and what they are doing. Exposure models are useful
34 in realistically estimating personal air concentrations and intake dose based on activity-specific
November 2005 4-5 Draft- Do Not Quote or Cite
-------
1 ventilation rates, particularly when recognizing that these measurements cannot be performed for
2 a given population. This section provides a brief overview of the model used by EPA staff to
3 estimate Os population exposure. Details about the application of the model to estimate Os
4 population exposure are provided in sections 4.4 and 4.5 and in the draft Exposure Analysis
5 TSD.
6 4.3.1 The APEX Model
7 The EPA has developed the APEX model for estimating human population exposure to
8 criteria and air toxic pollutants. APEX also serves as the human inhalation exposure model'
9 within the Total Risk Integrated Methodology (TRIM) framework (Richmond et al., 2002; EPA
10 2005c). APEX is conceptually based on the probabilistic NAAQS Exposure Model (pNEM) that
11 was used in the last O3 NAAQS review (Johnson et al., 1996a, 1996b, 1996c). Since that time
12 the model has been restructured, improved, and expanded to reflect conceptual advances in the
13 science of exposure modeling and newer input data needed for the model. Key improvements to
14 algorithms include replacement of the cohort approach with a probabilistic sampling approach
15 focused on individuals, accounting for fatigue and oxygen debt after exercise in the calculation
16 of ventilation rates, and anew approach for construction of longitudinal activity patterns for
17 simulated persons. Major improvements to data input to the model include air exchange rates
18 and the people's daily activities database. These improvements are described later in this
19 chapter.
20 APEX simulates the movement of individuals through time and space and their exposure
21 to a given pollutant in indoor, outdoor, and in-vehicle microenvironments. Figure 4-1 provides a
22 schematic overview of the APEX model. The model stochastically generates simulated
23 individuals using census-derived probability distributions for demographic characteristics
24 (Figure 4-1, steps 1-3). The population demographics are drawn from the year 2000 Census at
25 the tract level, and a national commuting database based on 2000 census data provides home-to-
26 work commuting flows between tracts.J Any number of simulated individuals can be modeled,
27 and collectively they approximate a random sampling of people residing in a particular study
28 area
29 Daily activity patterns for individuals in a study area, an input to APEX, are obtained
30 from detailed diaries that are compiled in the Consolidated Human Activity Database (CHAD)
31 (McCurdy et al., 2000; EPA, 2002). The diaries are used to construct a sequence of activity
32 events for simulated individuals consistent with their demographic characteristics, day type, and
33 season of the year, as defined by ambient temperature regimes (Graham & McCurdy, 2004)
1 There are approximately 65,400 census tracts in the ~3,200 counties in the U.S. „
November 2005 4-6 Draft - Do Not Quote or Cite
-------
1 (Figure 4-1, step 4). APEX calculates the concentration in the microenvironment associated with
2 each event in an individual's activity pattern and sums the event-specific exposures within each
3 hour to obtain a continuous series of hourly exposures spanning the time period of interest
4 (Figure 4-1, steps 5,6).
5 APEX has a flexible approach for modeling microenvironmental concentrations, where
6 the user can define the microenvironments to be modeled and their characteristics. Typical
7 indoor microenvironments include residences, schools and offices. Outdoor microenvironments
8 might include near roadways, at bus stops, and playgrounds. Inside cars, trucks, and mass transit
9 vehicles are microenvironments which are classified separately from indoors and outdoors.
10 Activity-specific simulated breathing rates of individuals are used in APEX to
11 characterize intake dose received from an exposure. These breathing, or ventilation, rates are
12 derived from energy expenditure estimates for each activity included in CHAD and are adjusted
13 for age- and gender-specific physiological parameters associated with each simulated individual.
14 Energy expenditure estimates themselves are derived from METS (metabolic equivalents of
15 work) distributions associated with every activity in CHAD (McCurdy et al., 2000), largely
16 based upon the Ainsworth et al. (1993) "Compendium of Physical Activities." METS are a
17 dimensionless ratio of the activity-specific energy expenditure rate to the basal or resting energy
18 expenditure rate, and the metric is used by exercise physiologists and clinical nutritionists to
19 estimate work undertaken by individuals as they go through their daily life (Montoye et al.,
20 1996). This approach is discussed more thoroughly in McCurdy (2000).
21 4.3.2 Key Algorithms
22 Ozone concentrations in each microenvironment are estimated using either a mass-
23 balance or transfer factors approach, and the user specifies probability distributions for the
24 parameters that are used in the microenvironment model (e.g., indoor-outdoor air exchange
25 rates). These distributions can depend on the values of other variables calculated in the model or
26 input to APEX. For example, the distribution of air exchange rates in a home, office, or car can
27 depend on the type of heating and air conditioning present, which are stochastic inputs to the
28 model, as well as the ambient temperature. The user can choose to keep the value of a stochastic
29 parameter constant for the entire simulation (which would be appropriate for the volume of a
30 house), or can specify that a new value shall be drawn hourly, daily, or seasonally from specified
31 distributions. APEX also allows the user to specify, diurnal, weekly, or seasonal patterns for
32 various microenvironmental parameters.
November 2005
4-7
Draft - Do Not Quote or Cite
-------
z
I
I
o
2
a
js
£
rfl
i
s
o
a
•e
5
B5
I
6
IS
O
cx
I
o
Q
i
Q
oo
4
o
o
1
S
z
-------
t
1
o
e
s
x
8,
§
55
B
|
| -ol
•i * J*
w ^"S
a S -3
•S " a, s
Scj w S
S ™ fe
&. £ a -2
I'S.lTf
115'
II
1-^il
tM -f I 5
13 § 5
li.a
1^-ls
I 111
to &.O- a
£
o
•*
S
§
u
o
I
o
S
!
ex
-^-t
o
o
Q
15
•8J
•
O
o
o
2
U.S EPA Headquarters Library
Mail code 3404T
1200 Pennsylvania Avenue NW
Washington, DC 20460
202-566-0556
-------
J
3
|
i
o
•c
"s
Q
O
a
i
§
u
o
s
Si
o
I
e
£
u
•a
1
«
!
a
•«
•e
|
d)
«H»
o
u*
o
o
o
Q
o
r-H
•4-
s
o
(U
x>
-------
1 The mass balance method assumes that the air in an enclosed microenvironment is well-
2 mixed and that the air concentration is fairly spatially uniform at a given time within the
3 microenvironment. The following four processes are modeled to predict the concentration of an
4 air pollutant in such a microenvironment:
5 • Inflow of air into the microenvironment;
6 • Outflow of air from the microenvironment;
7 • Removal of a pollutant from the microenvironment due to deposition, filtration, and
8 chemical degradation; and
9 • Emissions from sources of a pollutant inside the microenvironment.
10 The transfer factors model is simpler than the mass balance model, however, still most
11 parameters are derived from distributions rather than single values, to account for observed
12 variability. It does not calculate concentration in a microenvironment from the concentration in
13 the previous hour and it has only two parameters, a proximity factor, used to account for
14 proximity of the microenvironment to sources or sinks of pollution, or other systematic
15 differences between concentrations just outside the microenvironment and the ambient
16 concentrations (at the measurements site), and a penetration factor, which quantifies the degree
17 to which Ihe outdoor air penetrates into the microenvironment and is essentially the ratio of the
18 concentration in the microenvironment to the outdoor concentration.
19 Regardless of the method used to estimate the microenvironmental concentrations, APEX
20 calculates a time series of exposure concentrations that a simulated individual experiences during
21 the modeled time period. APEX estimates the exposure using the concentrations calculated for
22 each microenvironment and the time spent in each of a sequence of microenvironments visited
23 according to the "activity diary" of each individual. The hourly average exposures of each
24 simulated individual are time-weighted averages of the within-hour exposures. From hourly
25 exposures, APEX calculates the time series of 8-hr and daily average exposure concentrations
26 that simulated individuals experience during the simulation period. APEX then statistically
27 summarizes and tabulates the hourly, 8-hr, and daily exposures.
28 4.3.3 Model Output
29 There are several useful indicators of exposure and intake dose rate of people to O3 air
30 pollution. Factors that are important include the magnitude and duration of exposure, frequency
31 of repeated high exposures, and the breathing rate of individuals at the time of exposure. In this
32 analysis, exposure indicators include daily maximum 1-hr and 8-hr average Oa exposures,
33 stratified by a measure of the level of exertion at the times of exposure. The level of exertion of
34 individuals engaged in particular activities is measured by an equivalent ventilation rate (EVR),
November 2005 4-11 Draft - Do Not Quote or Cite
-------
1 ventilation normalized by body surface area, which is calculated as Ve/BSA, where Ve is the
2 ventilation rate "and BSA is the body surface area of the individual. Table 4-1 lists the ranges of
3 EVR corresponding to "moderate" and "heavy" levels of exertion. .
4
5 Table 4-1. Exertion Levels in Terms of Equivalent Ventilation Rates (liters/min-mZ)
Averaging time Moderate exertion Heavy exertion
Ihour 16-30 EVR > 30 EVR
8-hr 13-27 EVR > 27 EVR
6 from Whitfield et al, 1996, page 15.
7
8 APEX calculates two general types of exposure estimates: counts of the estimated
9 number of people exposed to a specified Oa concentration level and the number of times per Os
10 season2 mat they are so exposed; the latter metric is in terms of "person-occurrences." The
11. former highlights the number of individuals exposed one or more times per 63 season to the
12 exposure indicator of interest. In the case where the exposure indicator is a benchmark
13 concentration level, the model estimates the number of people who are expected to experience
14 that level of air pollution, or higher, at least once during the modeled period. The person-
15 occurrences measure estimates the number of times per season that individuals are exposed to the
16 exposure indicator of interest and then accumulates these estimates for the entire population
17 residing in an area. The latter metric conflates people and occurrences: one occurrence for each
18 of 10 people is counted the same as 10 occurrences for one person.
19 APEX tabulates and displays the two measures for exposures above levels ranging from 0
20 to 0.16 ppm by 0,01 ppm increments, where the exposures are:
21 • Daily maximum 1-hr average exposures
22 • Daily maximum 8-hr average exposures
23 • Daily average exposures.
24 These results are tabulated for the following population groups:
25 • All ages and activity levels
26 • Children at all activity levels
27 • Active people of all ages
28 • Active children.
29 Separate output tables are produced for different levels of exertion concomitant with the
30 exposures:
2 For purposes of the current exposure analysis, the Oj season was defined as a fixed period from April 1 through
September 30 for all urban areas included in the analysis.
November 2005 4-12 Draft - Do Not Quote or Cite
-------
1 «A11 exertion levels
2 • Moderate exertion levels
3 • Heavy exertion levels.
4
5 APEX also produces tables of the time spent in different microenvironments, stratified by
6 exposure levels.
7 4.3.4 Limitations of the Model
8 APEX has a strong scientific foundation and incorporates several significant algorithmic
9 improvements and updates to input data since its predecessor pNEM was used in the last review.
10 However, significant uncertainties in the predictions of APEX remain. In this section we discuss
11 qualitatively some of the limitations of this application of APEX to model population exposures
12 to Os pollution. A quantitative uncertainty analysis will be presented in the next draft of this
13 Staff Paper, which will address the impacts of these limitations.
14 We divide our discussion of the limitations of APEX into four areas: estimation of
15 ambient air quality, estimati on of concentrations in microenvironments, characterization of
16 population demographics and activity patterns, and modeling physiological processes. In
17 general, limitations and uncertainties result from variability not modeled or modeled incorrectly,
18 erroneous or uncertain inputs, errors in coding, simplifications of physical, chemical, and
19 biological processes to form the conceptual model, and flaws in the conceptual model. We
20 restrict the discussion here to limitations of the modeling of variability and the quality of input
21 data
22 4.3.4.1 Estimation of Ambient Air Quality
23 For estimating ambient Os concentrations to use in the exposure model, the urban areas
24 modeled have several monitors measuring hourly Os concentrations. The primary uncertainties
25 in the air quality data input to the model result from errors in estimating concentrations at ,
26 locations which are not close to monitoring sites (spatial interpolation) and from the estimation
27 of missing data. Concentrations of Oj near roadways are particularly difficult to estimate due to
28 tfie rapid reaction of 03 with N02 emitted from motor vehicles.
29 If a single Os season is modeled, another source of uncertainty results from the year-to-
30 year variability of CH concentrations. We have modeled the year 2004, the most recent year with
31 air quality and meteorological data. For most of the 12 areas modeled, Oa concentrations were
32 lower than previous years, due to a combination of reduced emissions of precursors and weather
33 patterns less conducive to the formation of 03. We plan to also model the year 2002 as part of an
November 2005 4-13 Draft - Do Not Quote or Cite
-------
1 analysis of the sensitivity of the exposure modeling results to year-to-year variability of air
2 quality and meteorology, and present these results in the next draft Staff Paper.
3 Modeling exposures for an unspecified future year simulated to just meet alternative air
4 quality standards has, in addition to the uncertainties involved with modeling historical
5 scenarios, Ihe uncertainties of the complex process of projecting to future years air quality,
6 population demographics, activity patterns, and other changing parameters. For the purpose of
7 estimating population exposure as an input to decisions about the appropriate level of a NAAQS,
8 EPA has historically not incorporated any projections in population demographics, activity
9 patterns, or other factors (e.g., air conditioning use, changes in housing types, etc). This allows
10 policy makers to focus on the impact of changing the allowed air quality distribution on
11 population exposure and public health while avoiding the additional uncertainties that inclusion
12 of these other factors would introduce.
13 4.3.4.2 Estimation of Concentrations in Indoor Microenvironments
14 The importance of estimation of concentrations in indoor microenvironments (homes,
15 offices, schools, restaurants, vehicles, etc.) is underscored by the finding that personal exposure
16 measurements of O$ are often not well-correlated with ambient measurements (draft CD, pages
17 3-59 to 3-61). However, in some cases, particularly where air exchange rates are high, indoor Oa.
18 concentrations generally closely track outdoor Os concentrations (draft CD, Appendix AX3,
19 page 175).
20 The niicroenvironmental characteristics used to model the concentrations in
21 microenvironments tend to be highly variable, both between microenvironments (e.g., different
22 houses have varying characteristics) and within microenvironments (e.g., the characteristics of a
23 given house can vary over time). Since APEX is a probabilistic model, if data accurately
24 characterizing this variability could be provided to the model, such variability would not result in
25 uncertainties. However, input data are always a limiting factor. Even if we can accurately
26 characterize the distributions of each individual microenvironmental parameter, we still need to
27 account for the relationships between the different parameters, as well as the relationships
28 between the microenvironmental parameters, human activities, physiology, and other
29 components of the exposure model.
/-
30 43.4.2.1 Air Exchange Processes
31 The AER is the single most important factor in determining the ratio of outdoor to indoor
32 concentrations of 63. AERs are highly variable, both within a microenvironment over time and
33 between microenvironments of the same type. AERs depend on the physical characteristics of a
34 microenvironment and also on the behavior of the occupants of the microenvironment. There is
35 also some dependence on the atmospheric conditions. APEX uses probabilistic distributions of
November 2005 4-14 Draft - Do Not Quote or Cite
-------
1 AERs which were derived from seven measurement studies in a number of locations, thought to
2 be sufficient to adequately characterize AERs for this analysis (see Appendix A of the draft
3 Exposure Analysis TSD),
4 4.3.4.2.2 Deposition Processes
5 The rate of deposition of Os to a surface depends on the material the surface is made of,
6 the humidity, and the concentration of 63. The rate of removal of Os from a microenvironment
7 depends on the dimensions, the ratio of surface area to volume, surface coverings, and
8 furnishings in the microenvironment. This is modeled in APEX by a distribution of decay rates
9 based on a study which measured decay rates in 26 homes in Southern California (Lee et al.5
10 1999). Although we do not expect inter-city differences in decay rates to be more important than
11 differences between homes within cities, there is some uncertainty associated with the small
12 sample size of this study. We do not expect this to be a major contributor to the uncertainty of
13 the modeling results.
14 4.3.4.2.3 Chemical Reaction Processes
15 Ozone reacts with a number of indoor pollutants, such as NOX from gas stoves and VOCs
16 from consumer products. However, Os reacts slowly with most indoor pollutants, and this is a
17 minor removal process compared to air exchange and surface removal (Weschler, 2000). Thus
18 the lack of a better treatment of indoor air chemistry is not considered to be a significant
19 limitation of APEX for modeling Os.
20 4.3.4.3 Characterization of Population Demographics and Activity Patterns
21 In addition to the uncertainty inherent in the human activity data input to APEX, there are
22 a number of population characteristics or attributes that contribute to the variability of exposures
23 which are modeled in APEX, but for which the assignment to simulated individuals is not
24 entirely reflective of real people:
25 • Occupational category
26 • Longitudinal stability in occupation, exercise levels, and leisure activities
27 • Geographical locations of activities away from the home
28 • The specific microenvironments visited away from home
29 • Representativeness of CHAD diaries (numbers of diaries used (20,000 used to represent
30 several million people over long periods of time), age of diaries (some are more than 20
31 years old), diary structure differences, etc.)
32
November 2005 4-15 Draft - Do Not Quote or Cite
-------
1 In addition, the extent to which the human activity database provides a balanced
2 representation of the population being modeled is likely to vary across areas. Although the
3 algorithm (hat constructs activity sequences accounts to some extent for the effects of population
4 demographics and local climate on activity, this adjustment procedure is unlikely to fully account
5 for all intercity differences in people's activities. Activity patterns are likely to be affected by
6 many local factors, including topography, land use, traffic patterns, mass transit systems, and
7 recreational opportunities.
8 43.4.4 Modeling Physiological Processes
9 The modeling of physiological processes that are relevant to the exposure and dose of 63
10 is a complicated endeavor. APEX currently has a physiological model for ventilation rates,
11 which is the primary driver of dose of Os. The limitations of this model have not yet been
12 characterized. •
13
14 4.4 SCOPE OF EXPOSURE ASSESSMENT
15 4.4.1 Selection of Urban Areas to be Modeled
16 The selection of urban areas to include in the exposure analysis takes into consideration
17 the location of Os epidemiologic studies, the availability of ambient 03 data, and the desire to
18 represent a range of geographic areas, population demographics, and Os climatology. These
19 selection criteria are discussed further in Chapter 5. Based on these criteria, staff chose the 12
20 urban areas listed in Table 4-2 to develop population exposure estimates. The geographic extent
21 of each modeled area consists of the census tracts in the combined statistical area (CS A) as
22 defined by OMB (OMB, 2005).
23 4.4.2 Time Periods Modeled
24 The exposure period modeled is from April 1 through September 30 for the most recent
25 year. This period encompasses all or most of the O3 season in 9 of the 12 urban study areas
26 when high ambient Os levels are most likely to occur, and for which routine hourly Os
27 monitoring data are available. For three of the study areas (Houston, Los Angeles, and
28 Sacramento) the Os season is the entire year, but the current exposure analysis only includes the
29 six month period from April through September. The geographic scope and time period modeled
30 are summarized in Table 4-2.
31
32
33
November 2005 4-16 Draft - Do Not Quote or Cite
-------
1 Table 4-2. Urban Areas and Time Periods Modeled
Urban Area (CSA) Period modeled
Atlanta-Sandy Springs-Gainesville, GA-AL April 1 to Sept. 30
Boston-Worcester-Manchester, MA-NH April 1 to Sept 30
Chicago-Naperville-Michigan City, IL-IN-WI April 1 to Sept 30
Cleveland-Akron-Elyria, OH April 1 to Sept. 30
Detroit-Warren-Flint, MI April 1 to Sept. 30
Houston-Baytown-Huntsville, TX April 1 to Sept 30
Los Angeles-Long Beach-Riverside, CA April 1 to Sept 30
New York-Newark-Bridgeport, NY-NJ-CT-PA April 1 to Sept. 30
Philadelphia-Camden-Vineland, PA-NJ-DE-MD April 1 to Sept. 30
Sacramento-Arden-Arcade-Truckee, CA-NV April 1 to Sept. 30
St. Louis-St Charles-Farmington, MO-IL April 1 to Sept 30
Washington-Baltimore-N. Virginia, DC-MD-VA-WV April 1 to Sept. 30
3 4.4.3 Populations Modeled
4 Exposure modeling is conducted for the general population residing in each area
5 modeled, as well as for school-age children (ages 5 to 18) and active school-age children. Due to
6 the increased amount of time spent outdoors engaged in relatively high-levels of physical activity
7 (which increases dose rates), school-age children as a group are particularly at risk for •
8 experiencing Os-related health effects. The next draft of this Staff Paper will include exposure
9 modeling for asthmatic school-age children.
10 Levels of physical activity are categorized by a daily physical activity index (PAI), a time
11 integrated measure of METS (discussed in section 4.3.1 above). Children are characterized as
12 active if their median daily PAI over the period modeled is greater than 1.75, a level
13 characterized by exercise physiologists as being "moderately active" or "active" (McCurdy,
14 2000).
15 Table 4-3 lists the year 2000 populations of the modeled areas. The 12 modeled areas
16 combined represent 40 percent of the total U.S. urban population (approximately 222 million) in
17 2000.
November 2005 4-17 Draft - Do Not Quote or Cite
-------
1 Table 4-3. Population Coverage of Modeled Areas
Urban Area (CSA) Modeled Modeled Active
population children children
(thousands) (thousands) (thousands)
Atlanta
Boston
Chicago
Cleveland
Detroit
Houston
Los Angeles
New York
Philadelphia
Sacramento
St. Louis
Washington, DC
Population in all 12 areas
4,548
5,714
9,311
2,945
5,357
4,815
16,349
21,357
5,832
1,930
2,754
7,572
88,484
942
1,098
1,946
582
1,110
1,076
3,594
4,084
1,179
418
572
1,473
18,074
519
529
933
295
553
598
1,951
2,009
609
226
309
759
9,290
3 4.5 INPUTS TO THE EXPOSURE MODEL
4 The data inputs to the APEX model are briefly described in this section. A more detailed
5 description of the development of these data and the derivation of input distributions can be
6 found in the draft Exposure Analysis TSD.
7 4.5.1 Population Demographics
8 APEX takes population characteristics into account to develop accurate representations of
9 study area demographics. Population counts and employment probabilities by age and gender
10 are used to develop representative profiles of hypothetical individuals for the simulation. Tract-
11 level population counts by age in one-year increments, from birth to 99 years, come from the
12 2000 Census of Population and Housing Summary File 1. Summary File 1 contains the 100-
13 percent data, which is the information compiled from the questions asked of all people and about
14 every housing unit.
15 Employment data from the 2000 Census provide employment probabilities for each
16 gender and specific age groups for every Census tract. The employment age groupings are: 16-
November 2005
4-18
Draft - Do Not Quote or Cite
-------
1 19, 20-21, 22-24, 25-29, 30-34, 35-44, 45-54, 55-59, 60-613 62-64, 65-69, 70-74, and >75 years
2 of age. Children under the age of 16 are assigned employment probabilities of zero.
3 4.5.2 Population Commuting Patterns
4 To ensure that the people's daily activities are accurately represented within APEX, it is
5 important to integrate working patterns into the assessment. The APEX commuting data were
6 originally derived from the 2000 Census and were collected as part of the Census Transportation
7 Planning Package (CTPP). CTPP contains tabulations by place of residence, place of work, and
8 the flows between the residence and work. These data are available from the U.S. Department of
9 Transportation, Bureau of Transportation Statistics (U.S. Department of Transportation and U.S.
10 Census Bureau, 2000).
11 It was assumed that all persons with home-to-work distances up to 120 km are daily
12 commuters, and that persons who travel further than 120 km do not commute daily. Therefore
13 the list of commuting destinations for each home tract is restricted to only those work tracts that
14 are within 120 km of the home tract.
15 APEX allows the user to specify how to handle individuals who commute to destinations
16 outside the study area. One option is to drop them from the simulation. If they are included, the
17 user specifies values for two additional parameters, called LU and LA (Multiplicative and
18 Additive factors for commuters who Leave the area). While a commuter is at work, if the
19 workplace is outside the study area, then the ambient concentration cannot be determined from
" ;
20 any air district (since districts are inside the study area). Instead, it is assumed to be related to
21 the average concentration CAVB(() over all air districts at the time in question. The ambient
22 concentration outside the study area at time t, Cour(f), is estimated as:
23 COUT(0 = LM * CAVE (0 + LA (4-2)
24 The microenvironmental concentration (for example, in an office outside the study, area)
25 is determined from this ambient concentration by the same model (mass balance or factor) as
26 applies inside the study area. The parameters LM and LA were both' set to zero for this modeling
27 analysis; thus, exposures to individuals are set to zero when they are outside of the study area,
28 This was done since we have not estimated ambient concentrations of Oa in counties outside of
29 the modeled areas.
30 4.5.3 Human Activity Data
31 The human activity data are drawn from the Consolidated Human Activity Database
32 (CHAD) (McCurdy et al., 2000; EPA, 2002), developed and maintained by the Office of
33 Research and Development's (ORD) National Exposure Research Laboratory (NERL). The
34 CHAD includes data from several surveys covering specific time periods at city, state, and
November 2005 4-19 Draft - Do Not Quote or Cite
-------
1 national levels, with varying degrees of representativeness. Table 4-4 summarizes the studies in
2 the version of CHAD used in this modeling analysis.
3 A key issue in this assessment is Hie development of an approach for creating (Vseason
4 or year-long activity sequences for individuals based on a cross-sectional activity data base that
5 includes 24-hr records. The average subject in the time/activity studies in CHAD provided less
6 than two days of diary data. For this reason, the construction of a season-long activity sequence
7 for each individual requires some combination of repeating data from one subject and using data
8 from multiple subjects. An appropriate approach should adequately account for the day-to-day
9 and week-to-week repetition of activities common to individuals while maintaining realistic
10 variability between individuals. The method in APEX for creating longitudinal diaries which
11 reflect the tendency of individuals to repeat activities is based on reproducing realistic variation
12 in a key diary variable, which is a user-selected function of diary variables. For this analysis the
13 key variable is set to the amount of time an individual spends outdoors each day.
14 The actual diary construction method targets two statistics, a diversity statistic (£>) and an
15 autocorrelation statistic (A). The D statistic reflects the relative importance of within-person
16 variance and between-person variance in the key variable. The A statistic quantifies the lag-one
17 (day-to-day) key variable autocorrelation. Desired D and A values for the key variable are
18 selected by the user and set in the APEX parameters file, and the method algorithm construct
19 longitudinal diaries that preserve these parameters. Longitudinal diary data from a field study of
20 school-age children (Geyh et al., 2000) and subsequent analyses (Xue et al., 2004) suggest thatD
21 and ./Tare stable over time (and perhaps over cohorts as well). Based on these studies,
22 appropriate target values for the two statistics for outdoor time for children are determined to be
23 0.22 for D and 0.19 for A. In the absence of data for estimating these statistics for younger
24 children and for adults, these values are also used for adults. This method for constructing
25 longitudinal diaries from the CHAD data is described in detail in Appendix C of the Exposure
26 Analysis TSD.
27 4.5.4 Physiological Data
28 APEX requires values for various physiological parameters for subjects in order to
29 accurately model their pollutant intake via metabolic processes. This is because physiological
30 differences may cause people with the same exposure and activity scenarios to have different
31 pollutant intake levels. The physiological parameters file distributed with APEX contains
32 physiological data or distributions by age and gender for maximum ventilatory capacity (in terms
33 of age- and gender-specific maximum oxygen consumption potential), body mass, resting
34 metabolic rate, and oxygen consumption-to-ventilation rate relationships.
November 2005
4-20
Draft -Do Not Quote or Cite
-------
u
e
i
1
3'
«
u
Si
0*
•c*
S§
S »-
% °
il
i •§
ss -a
ll
.£"-'
t*
85 M
s-s
1 A
e » „
P s» w
I*, s i?
Z e Q.TS
b a
« i
g i
! d
d *
o t>
JU
s^
g
0
1
1
(2
ON
pL
S8 5.
1— 1 p-4
R
g 5
r~ •*•
>— i OS
es oo
1^ «— 1
t
r~~ oo
ss
0 OS
<-H O
A
1
u
California
Adolescents and
Adults (CARB)
S
^H
S
N^*'
"3
u
1
s
5
1
1
^
O
/Hk
S
pml
O
o
cs
*.
S
(
o\ o
00 ON
ON ON
5S
0 0
£
c3
California
Children
(CARB)
/«\
N^
c
JS
g
§
1
s
'n
rr
V£)
*s
ts
1
SO
00
o
1
w%
-------
t!
II
a *
S-8
S g
t §
£*&
3 S
CA a
ft ^
s|
o 8
•dS
.
"3
t
Q
\o
VO
-------
1 4,5.5 Microenvironments Modeled
2 In APEX, microenvironments provide the exposure locations for modeled individuals.
3 For exposures to be measured accurately, it is important to have realistic microenvironments that
4 are matched closely to what actual people experience on a daily basis. As discussed in section
5 4.3.2 above, the two methods available in APEX for calculating pollutant concentrations within
6 microenvironments are a mass balance model and a transfer factor approach. Table 4-5 lists the
7 12 microenvironments selected for this analysis and the exposure calculation method for each. .
8 The parameters used in mis analysis for modeling these microenvironments are described in
9 Appendix 4A.
10 Table 4-5. Microenvironments Modeled
11
12
13
14
15
16
17
18
19
20
Microenvironment
Indoors - Residence
Indoors - Bars and restaurants
Indoors - Schools
Indoors - Day -care centers
Indoors - Office
Indoors - Shopping
Indoors - Other
Outdoors - Near road
Outdoors - Public garage/parking lot
Outdoors - Other
In-vehicle - Cars and Trucks
Calculation Method
Mass balance
Mass balance
Mass balance
Mass balance
Mass balance
Mass balance
Mass balance
Factors
Factors
Factors
Factors
Parameters1
AER and DE
AER and DE
AER and DE
AER and DE
AER and DE
AER and DE
AER and DE
PR
PR
None
PE and PR
1 A£R=air exchange rate, DE=decay-deposition rate, PR=proximity factor, PE=penetration factor
4.5.6 Ambient Ozone Concentrations
APEX requires hourly ambient Os concentrations at a set of locations in the study area.
Data from the EPA AIRS Air Quality Subsystem were used to prepare the ambient air quality
input files, using the most recent year (2004) of Os measurements. The hourly Os concentrations
at the AIRS sites in each CS A were interpolated to a 20 by 20 km rectangular grid covering the •
CS A using a simple inverse squared-distance weighted average for each hour. Grid locations
further than 75 km from the closest Os monitor were dropped. The hourly gridded and site
concentrations were used as input to APEX to represent the ambient concentrations within each
urban area. For near road and parking garage microenvironments the ambient concentrations are
November 2005
4-23
Draft - Do Not Quote or Cite
-------
1 adjusted by a proximity factor. An analysis of the interpolation errors and uncertainty of the 63
2 concentrations input to APEX will be presented in the next draft Staff Paper.
3 In addition to modeling exposures based on 2004 air quality, an analysis was conducted
4 using air quality representative of just meeting the current 8-hr Os NAAQS of 0.08 ppm. This
5 was done using a quadratic rollback approach to adjust the hourly 63 concentrations observed in •
6 2002-2004 to yield a design value of 0.084 ppm (based on rounding conventions, concentrations
7 at or below 0.084 ppm are considered to meet the standard). Design values for the 8-hr Os
8 NAAQS are calculated as the 3-year averages of the annual 4th daily maximum 8-hr average
9 concentration based on the maximum monitor within an urban area and are given in Table 4-6
10 for the 2002-2004 period. The quadratic rollback technique combines both linear and quadratic
11 elements to reduce higher concentrations more than lower concentrations near ambient
12 background levels. The quadratic rollback adjustment procedure was considered in a sensitivity
13 analysis during the last review of the Oj NAAQS and has been shown to be more realistic than
14 the linear proportional rollback method, where all of the ambient measurements are reduced by a
15 constant multiplicative factor regardless of their individual magnitudes. The rollback approach
16 and evaluation of this approach are described by Johnson (1997), Duff, Horst, and Johnson
17 (1998), and Rizzo (2005).
18 Table 4-6. 2002-2004 8-Hour Ozone Design Values for die Modeled Areas
Urban Area (CSA)
2002-2004 design Ratio of 0.084 to
value (ppm) the design value
19
Atlanta
Boston
Chicago
Cleveland
Detroit
Houston
Los Angeles
New York
Philadelphia
Sacramento
St. Louis ,
Washington, DC
0.093
0.091
0.094
0.095 1 .
0.092
0.101
0.127
0.094
0.094
0.102
0.089
0.089
0.90
0.92
0.89
0.88
0.91
0.83
0.66
0.89
0.89
0.82
0.94
0.94
November 2005
4-24
Draft - Do Not Quote or Cite
-------
1 The observed concentrations and the concentrations representative of just meeting the
2 current standard are summarized (for 2004 only) in figures in Appendix 4B in terms of the
3 frequencies of daily maximum 8-hr average Os concentrations above different concentration
4 levels. These figures also illustrate the percent reduction in exceedances of levels from the 2004
5 base year to when the current standard is just met.
6 4.5.7 Meteorological Data
7 Daily average and maximum 1 -hr temperatures are computed from hourly surface
8 temperature measurements obtained from the National Weather Service. These data are not
9 spatially interpolated; APEX uses the data from the closest weather station to each Census tract.
10 Temperatures are used in APEX both in selecting human activity data and in estimating air
11 exchange rates for indoor microenvironments, as discussed above.
12 4.6 EXPOSURE ASSESSMENT RESULTS
13
14 The results of the exposure analysis are presented in a series of tables in this section and
15 in graphs in Appendix 4C. The tables summarize exposures to Os concentrations above two
16 specific levels, 0.08 and 0.07 ppm, while the graphs in Appendix 4C depict these results over the
17 entire range of observed concentrations. The 0.08 ppm level corresponds to the lowest exposure
18 level used in the controlled human studies (see section 3.6.1). The exposure-response
19 relationships based on these studies are extrapolated down to background levels (section 5.3.1),
20 and the 0.07 ppm level was chosen to illustrate the sensitivity of the exposure modeling results to
21 the selected level. The tables summarize two measures of the extent of population exposures
22 over the modeled period (April 1 through September 30):
23 » numbers of person-days with daily maximum 8-hr average exposures under moderate
24 exertion above 0.08 and 0.07 ppm (Table 4-77 and Table 4-10),
25 • numbers of persons who experience one or more daily maximum 8-hr average
26 exposures under moderate exertion above 0.08 and 0.07 ppm (Table 4-8 and Table
27 4-111), also expressed as percentages of the population (Table 4-9 and Table 4-22) and
28 Exertion is characterized by equivalent ventilation rates (section 4.3.3). Results are
29 presented for the 2004 base case, "just meeting" the current standard, and the difference between
30 these scenarios (as a percent reduction), for the general population, children (ages 5-18), and for
31 active children. For the tables reporting counts of persons, there are companion tables with the
32 same results expressed as percentages of the population group totals given in Table 4-3.
33 For example, for Atlanta, Table 4-8 reports an estimated 13.8 thousand active children
34 who experience an 8-hr average exposure above 0.08 ppm under moderate exertion at least once
November 2005 4-25 Draft - Do Not Quote or Cite
-------
1 during the 2004 Oa season. Under the scenario of just meeting the current 8-hr standard, this
2 number is reduced by 87 percent to 1.8 thousand active children. These correspond to 2.7 and
3 0.4 percent respectively (Table 4-9) of the population of active children in Atlanta (519
4 thousand).
5 The tables in this section have counts of exposures above a level of 0.08 ppm. Appendix
6 4C shows graphs of counts of person-days of 8-hr exposures above different levels ranging from
7 0 to 0.16 ppm, concomitant with moderate exertion, for children ages 5 to 18.
8 These results indicate a significant reduction in 8-hr average exposures above 0.08 ppm
9 under attainment of the current 8-hr Os NAAQS. In addition, under the current standard it is rare
10 for an individual to experience more than one 8-hr exposure above 0.08 ppm, as evidenced by
11 the average number of 8-hr exposures above 0.08 for individuals, which is less than 1.01 for each
12 of the three population groups in these tables (compared to 1.2 for historical 2004 air quality).
November 2005
4-26
Draft - Do Not Quote or Cite
-------
00
•«.
o
9
09
O
Ed
I
I
00
I
i
s
t
a
5
09
•o
o
I
If
•O R
II
I
«
5
«
I
t^
4
••o
a
1
81
II
3 9
II
U I
13
City (CSA
OO
OOCS
O,— ;OS
Ox-
O
OOO
00
Ch
r-i o O O O
O O O O O
G
I
O O O OO CS
>/~i O
m
xP
tfx
00
m o o o o o o vo o o o «>
CS O O O O l> O O O O O VO
O OS SC
oo —
O OO
CS
•4
so
O O O O i—' O •—« i—' O O OO
n "-< cs o o o f-<
-------
V
•e
B
^** «
|
a.
yP
o
^
o
s '
8
o.
tt
1
o
B
s i
i 1
•- 9
a|
•5" -3
'S o
Q *C
™ U
II
*fc* ^
52 «^
u o
T M
tj 4J
-S
.ts 5;
*J
il
£ |
*|
_• KM
si
II
00
4
H
es en
/_4
2
*£
jU
- O O O O O *>D
n<— !•— ioo»noooor-ieSi— "
ON so i— i i— i c i i— i o iin ii— i TJ- o f-
<— ' oo ON ^ i— i *n
en
s^ s^ s^ NP Sp 's^ xP *sP vP xP xP vP
P*1* 0s- px tfs ^N QX P^ px QX 0s- O** O^
t^ ON ^5 ^^ ^^ ^5 ^^ oo NO ^5 ^^ ^^
OOONOOOONOONONOOOO
ONtsoooenOooenooso
\O O O O O <— • O i— i i— i O O OO'
n f— « es<— lOeni— i •<*
es OK r— c ^
Q
» frt
litltffltllf
<«OWQKNJ25o8ffi^
vP
ON
en
,_!
(-H
en
NO
ft
r-
ON
SO
s'-
Tf
oo
F-H
ON
^_
O
fS)
OO
ON
|
O
fe
o
Q
I
oo
M
4
1
-------
0
s
£
a
W&
o
n.
A --.
Ed Q.
v S
SB o
Sla
W>
!§
•< -c
>• «
s 3
si
1 ft.
««?
g e
MM <4M»
1 1
« w
S3
•*"> -S3
1.1
o £
i *•
n "*
2-«
to v
« •=>
S e
^s
^^
ft
SJ
5» ^^
S s
s s
II
•2 S
«s«
tn ji
fli S
•2 «
S "o
5 e
Is
ON
4
9)
&
&
H
y^^
00
^M
V)
1
xS
e
2
IS
u
«
w
^^
0?
1
V)
cu
v5
i
•e
1
^o
3
£
"3
j^
c
5
11
sl
i 1
03 w
S§
31
U %
O9 UO
» u
*• -o
S "«3
U 'O
o S
«1
en g
ea «
^
£
*
a
? 0 0 0
o o o o
P- §0 O i—
cs o o o
M O O O
o o o o
^ ^ ^ s?
(S| O O O
P P XP NO
?S 0 0 0
o o o o
o e o P
^ ^ O ^<
^ O O O
TJ
0 g
8 t>5 ." gj
< m 6 D
oSooooo^o
o o o o o o" o o
OPCOOPOO
0«^MC4000^
OOO""1!— I!— ' ^ O •xf
of-OOOOO«A
o o o o o o o o
xP xP xP xP x® xP xP xP
O i> ' ' •—' i— ' 1—l o r*i
xPxOxPxPxOxPvO-x?
o?r§50ooor>i
o o o o o o o o
xOxOxOxPxOxOxOxP
o •rf' vd o" o o o i— i
r\
^?
S? «* n C
T5 ^ S I §
c! w» S £" S S3 «>
.t- § J^ j2 !i | 3 .S
•bgcn^^atS.S
QKj^Sww^
I
a
I
rji
•xf
en
-------
o
o
§
2
§
00
s
•a
•a
_o
S-3
o i
e
II
01 d)
^ M
If
** s-
* 5
e
^H
4
£
£
0
c
Z
1
i/>
e
I
a
o
i
S'-g
v s
I-B
*3
11
a!
M
= 1
U S
4» V
QQ w
City
od
co o r>j o N
m
m
CO
-— I«NTJ-
r<
xp
0^
»n
cn
oc
(X)
i— i a\
00
cs
c«i i
fS
rn I ri oi i rn'
fS Os
$
«n
o
a
al
en
X?
5s-
Tt
OS
OS
-------
i
s «
3 "8
O
« o
ft -C
21
= 0
o e
E
2 *
v
4 -
S
H
00
If)
!!
w s
31
en
M «
II
I 3
0* -0
BQ
41 3
31
aa «
City (CSA
cs o O o o (S
00 -H O rt O
o o i— i
«o
0< (S
o ^^ o
v *
o o
os
-q-
ts
ts
o\
r-H en
06 od o
o o ri os' ts ri rs vo
(T) IS
-------
fe
90
i-H
If)
Wl
«5
SM
•e
§
i
w
0?
1
V3
£
M
>s>
e
^
25
U
JO
vt
"3
e
PH
1
55
u
"g 9
?* 3
-fc -g
3 S
U S
«J
U U
1 3
**
s* ^
*• IQ
31
in
a* a>
SB
^
(J 42
SB
« «
0$ C8
j,_
1
•
O^O^O^O^oS.pS.O^O^oXo4?
fS >O O <— ' O i— i i— i OO T) . fS
rt p O O O vS O O O O
^^XOSOSOXP^^^XO
1-1 vi o cft s 0 •& S
llslIisBly
<»UOQEJ?;S«
0?
f — (
0
xP
0s-
vo
0
g
o
0
Sn
O
S
0
xO •
^f
0
V)
'3
3
on
^
VD
g
I-N
*
m
>n
0
0s-
0?
cs
XO
5s
VO
Q
*•*
Washingtoi
I
-------
1 REFERENCES
2 Ainsworth, B. E.; Haskell, W. L.; Leon, A. S.; Jacobs, D. R., Jr.; Montoye, H. J.; Sallis, J. F.; Paffenbarger, R. S.,
3 Jr. (1993) Compendium of physical activities: classification of energy costs of human physical activities.
4 Med. Sci. Sports Exer. 25: 71-80.
5 Akland, G. G., Hartwell, T. D., Johnson, T. R., and Whitmore, R. W. (1985). Measuring.human exposure to carbon
6 monoxide in Washington, D. C. and Denver, Colorado during the winter of 1982-83. Environ. Sci. Technol.
7 19:911-918.
8 CASTNet (2004). Clean Air Status and Trends Network (CASTNet) 2003 Annual Report, Prepared by: MACTEC
9 Engineering and Consulting, Inc. Prepared for: U. S. Environmental Protection Agency, Office of Air and
10 Radiation, Clean Air Markets Division, Washington, DC.
11 Duff, M., Horst, R. L., Johnson, T. R. (1998). "Quadratic Rollback: A Technique to Model Ambient Concentrations
12 Due to Undefined Emission Controls." Presented at the Air and Waste Management Annual Meeting. San
13 Diego, California. June 14-18,1998.
14 Geyh, A. S.; Xue, J.; Ozkaynak, H.; Spengler, J. D. (2000). The Harvard Southern California chronic ozone
15 exposure study: assessing ozone exposure of grade-school-age children in two southern California
16 communities. Environ. Health Perspect. 108: 265-270.
17 Goldstein, B. R.; Tardiff, G.; Hoffnagle, G.; Kester, R. (1992). Valdez health study: summary report. Anchorage,
18 AK: Alyeska Pipeline Service Company.
19 Graham, S.E. and T. McCurdy (2004). Developing meaningful cohorts for human exposure models. J. Exp. Anal.
20 and Env. Epi., 14,23-43.
21 Hartwell, T. D.; Clayton, C. A.; Ritchie, R. M.; Whitmore, R. W.; Zelon, H. S.; Jones, S. M.; Whitehurst, D. A.
22 (1984). Study of carbon monoxide exposure of residents of Washington, DC and Denver, Colorado.
23 Research Triangle Park, NC:_U.S. Environmental Protection Agency, Office of Research and Development,
24 Environmental Monitoring Systems Laboratory; report no. EPA-600/4-84-031.
25 Hofferth, S., Davis-Kean. P. E., Davis, J., Finkelstein, J. (1997). The Child Development Supplement To The Panel
26 Study Of Income Dynamics. 1997 User Guide. Survey Research Center, Institute for Social Research, The
27 University of Michigan, Ann Arbor, MI. Available at: ht{p://ps'idori!ine.isr.tanich.edu/CDS/usergd.h1ml
28 Johnson, T. (1984). A study of personal exposure to carbon monoxide in Denver, Colorado. Research Triangle
29 Park, NC: U.S. Environmental Protection Agency, Environmental Monitoring Systems Laboratory; report no.
30 EPA-600/4-84-014.
31 Johnson, T. (1989). Human activity patterns in Cincinnati, Ohio. Palo Alto, CA: Electric Power Research Institute;
32 report no. EPRI EN-6204.
33 Johnson, T. (1997). "Sensitivity of Exposure Estimates to Air Quality Adjustment Procedure," Letter to Harvey
34 Richmond, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency, Research
3 5 Triangle Park, North Carolina.
36 Johnson, T., Capel, J., and McCoy, M. (1996a). Estimation of Ozone Exposures Experienced by Urban Residents
37 Using a Probabilistic Version of NEM and 1990 Population Data. Prepared by International Technology Air
38 Quality Services for Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
39 Research Triangle Park, NC; contract no. 63-D-30094; April 1996. Available at:
40 http:/Avwvv.cpa.gQv/ttn/naaqs/standards/o20ne/s o3 pr td.html.
41 Johnson, T., Capel, J., McCoy, M., and Wamasch, J. (1996b). Estimation of Ozone Exposures Experienced by
42 Outdoor Children in Nine Urban Areas Using a Probabilistic Version of NEM. Prepared by International
43 Technology Air Quality Services for Office of Air Quality Planning and Standards, EPA, Research Triangle
44 Park, NC; contract no. 63-D-30094; April 1996. Available at:
45 htto://www.epa.gov/ttn/t3aaas/standards/ozone/s o3__pr td.html.
November 2005 4-33 Draft - Do Not Quote or Cite
-------
Johnson, T., Capel, M. McCoy, and J. Wamasch. (1 996c). Estimation of Ozone Exposures Experienced by
Outdoor Workers in Nine Urban Areas Using a Probabilistic Version of NEM. 1 996. Prepared by
International Technology Air Quality Services for U, S. Environmental Protection Agency, OAQPS; contract
no. 63-D-30094; Research Triangle Park, NC, April, 1996. Available at: .
.go^^
1
2
3
4
6 Klepeis, N. E.; Tsang, A. M.; Behar, J. V. (1996). Analysis of the national human activity pattern survey (NHAPS)
7 respondents from a standpoint of exposure assessment. Washington, DC: U.S. Environmental Protection
8 Agency, Office of Research and Development, report no. EPA/600/R-96/074.
9 Lee, Vallarino, Dumyahn, Ozkaynak, and Spengler. 1999. Ozone decay rates in residences. JAWMA. 49: 1238-
10 1244. .
1 1 Lioy, PJ. (1 990). The analysis of total human exposure for exposure assessment: A multi-discipline science for
1 2 examining human contact with contaminants. Environmental Science and Technology 24:938-945.
13 • McCurdy, T. (2000). "Conceptual basis for multi-route intake dose modeling using an energy expenditure
14 approach." J. Exposure Anal. Environ. Epidemiol. 10:1-12.
1 5 McCurdy, T., Glen, G., Smith, L., Lakkadi, Y. (2000). "The National Exposure Research Laboratory's Consolidated
16 Human Activity Database." J. Exposure Anal. Environ. Epidemiol. 10:566-578.
17 Montoye, H. J., Kemper, H. C. G., etal. (1996). Measuring Physical Activity and Energy Expenditure. Champaign
1 8 IL: Human Kinetics.
19 National Research .Council (1991). Human exposure assessment for airborne pollutants: advances and opportunities.
20 Washington, DC: National Academy of Sciences.
21 OMB (2005). U.S. OMB Bulletin No. 05-02. Subject: Update of Statistical Area Definitions and Guidance on
22 Their Uses. February 22, 2005.
23 Richmond H., T. Palma, J. Langstaff, T. McCurdy, G. Glenn, and L. Smith (2002). "Further refinements and testing
24 of APEX (3.0): EPA's population exposure model for criteria and air toxic inhalation exposures." Poster
25 presentation. Joint meeting of the International Society of Exposure Analysis and International Society of
26 Environmental Epidemiology, August 11-15, 2002, Vancouver, Canada,
27 Rizzo, M. (2005). Evaluation of a quadratic approach for adjusting distributions of hourly ozone concentrations to
28 meet air quality standards. November 7, 2005.
29 Robinson, J. P., Wiley, J. A., Piazza, T., Garrett, K., and Cirksena, K. (1 989). Activity patterns of California
30 residents and their implications for potential exposure to pollution. California Air Resources Board (CARB-
31 . A6-177-33), Sacramento, CA.
32 Spier, C. E.; Little, D. E.; Trim, S. C.; Johnson, T. R.; Linn, W. S.; Hackney, J. D. (1992). Activity patterns in
33 elementary and high school students exposed to oxidant pollution. J. Exposure Anal. Environ. Epidemiol. 2:
34 277-293. .
35 Tsang A. M. and N. E. Klepeis (1 996). Descriptive Statistics Tables from a Detailed Analysis of the National
36 Human Activity Pattern Survey rNHAPS~) Data. U.S. Environmental Protection Agency (EPA/600/R-
37 96/148).
38 U.S. Department of Transportation and U.S. Census Bureau (2000). Census of Population and Housing, 2000 long-
39 form (sample) data, Census Transportation Planning Package (CTPP) 2000. Available at:
40 http://transtats.bts.gov/.
41 U.S..Environmental Protection Agency (1986). Air Quality Criteria for Ozone and Other Photochemical Oxidants.
42 EPA-600/8-84-020aF-eF. Office of Health and Environmental Assessment, Environmental Criteria and
43 Assessment Office, U.S. Environmental Protection Agency, Research Triangle Park, NC. Available from:
44 NTIS, Springfield, VA; PB87-142949.
November 2005
4-34
Draft - Do Not Quote or Cite
-------
1 U.S. Environmental Protection Agency (1996a). Review of National Ambient Air Quality Standards for Ozone:
2 Assessment of Scientific and Technical Information - OAQPS Staff Paper. EPA/452/R-96-007. Office of
3 Air Quality Planning and Standards, U.S. Environmental Protection Agency, Research Triangle Park, NC.
4 Available at: http://www.epa.gov/itrt/naaqs/siandards/ozone/s o3 pr sp.html
5 U.S. Environmental Protection Agency (1996b). Air Quality Criteria for Ozone and Related Photochemical
6 Oxidanis! EPA/600/P-93/004aF-cF. Office of Research and Development, National Center for
7 Environmental Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC. Available
8 at: http://cfpub.epa.gov/ncea/ciin/recordispluv.crm7deid~2831.
9 U.S. Environmental Protection Agency (2002). Consolidated Human Activities Database (CHAD) Users Guide.
10 The database and documentation are available electronically on the internet at: h^//ww.epa.gov/chadnetl/.
II U.S. Environmental Protection Agency (2005a). Ozone Population Exposure'Analysis for Selected Urban Areas
12 (draft). Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency, Research
13 Triangle Park, NC. November. Available electronically on the internet at:
14 http://\vww.epa.gov/(tn/naiiqs/'s(3ndards/ozone/s_o3 cr_td.html.
15 U.S. Environmental Protection Agency (2005b). Air Quality Criteria for Ozone and Other Related Photochemical
16 Oxidants. Second External Review Draft. National Center for Environmental Assessment, U.S.
17 Environmental Protection Agency, Research Triangle Park, NC. Available electronically on the internet at:
18 http://cfpxib.epa.gov/ncea/cfm/recordisplav.crni7deid~137307
19 U. S. Environmental Protection Agency (2005c). Total Risk Integrated Methodology (TRIM) - Air Pollutants
20 Exposure Model Documentation (TRIM-Expo / APEX, Version 4) Volume I: User's Guide. Office of Air
21 Quality Planning and Standards, U.S. Environmental Protection Agency, Research Triangle Park, NC.
22 . November 2005. Available at: htta://www.e»a.aov.''tJn/fCTa/human apex .him!.
23 Weschler, C. J. (2000) Ozone in indoor environments: concentration and chemistry. Indoor Air 10: 269-288.
24 Whitfield, R., Biller, W,, Jusko, M., and Keisler, J. (1996). A Probabilistic Assessment of Health Risks Associated
25 with Short- and Long-Term Exposure to Tropospheric Ozone. Argonne National Laboratory, Argonne, IL.
26 Wiley, J. A.; Robinson, J. P.; Piazza, T., Garrett, K.; Cirksena, K.; Cheng, Y.-T.; Martin, G. (1991a). Activity
27 patterns of California residents. Final report. Sacramento, CA: California Air Resources Board; report no.
28 ARB/R93/487. Available from: NTIS, Springfield, VA.; PB94-108719.
29 Wiley, J. A.; Robinson, J. P.; Cheng, Y.-T.; Piazza, T.; Stork, L.; Pladsen, K. (1991b). Study of children's activity
30 patterns: final report. Sacramento, CA: California Air Resources Board; report no. ARB-R-93/489.
31 Williams, R., Suggs, J., Creason, J., Rodes, C., Lawless, P., Kwok, R., Zweidinger, R., and Sheldon, L. (2000). The
32 1998 Baltimore particulate matter epidemiology-exposure study: Part 2. Personal exposure associated with an
33 elderly population. J. Expos. Anal. Environ. Epidemiol.
34 Xue, J., McCurdy, T., Spengler, J., Ozkaynak, H. (2004). Understanding variability in time spent in selected
35 locations for 7-12-year old children. J Expo Anal Environ Epidemiol 14(3):222-33.
November 2005 4-35 Draft - Do Not Quote or Cite
-------
5. CHARACTERIZATION OF HEALTH RISKS
2 5.1 INTRODUCTION
3 This chapter presents information regarding the results from an updated ozone (Os) health
4 risk assessment that builds supon the methodology used in the assessment conducted as part of the
5 last Os NAAQS review. This updated assessment includes estimates of (1) risks of lung function
6 decrements, hospital admissions, and mortality associated with recent ambient Oa levels; and (2)
7 risk reductions associated with just meeting the current 8-hr Oa NAAQS. The next draft of the
8 Staff Paper will also include risk reductions associated with just meeting various alternative Os
9 standards. The current risk assessment is more fully described and presented in a draft technical
10 support document, Ozone Health Risk Assessment for Selected Urban Areas (Abt Associates,
11 2005a; henceforth referred to as the Risk Assessment Technical Support Document and cited as
12 draft Risk Assessment TSD).
13 The goals of this Os risk assessment are: (1) to provide estimates of the potential
14 magnitude of mortality and morbidity effects associated with current 63 levels, and with meeting
15 the current Os NAAQS and alternative Os standards, in specific urban areas; (2) to develop a
16 better understanding of the influence of various inputs and assumptions on the risk estimates; and
17 (3) to gain insights into the distribution of risks and patterns of risk reductions associated with
18 meeting alternative Oj standards. Staff recognizes that while there are many sources of
19 uncertainty and variability inherent in the inputs to this assessment which make the specific
20 estimates uncertain, there is sufficient confidence in the direction and general indications
21 provided by the assessment for the assessment to serve as a useful input to decisions on the
22 adequacy of the 63 standard. While some of these uncertainties have been addressed
23 quantitatively in the form of estimated confidence ranges around central risk estimates, other
24 uncertainties and the variability in key inputs are not reflected in these confidence ranges, but
25 rather are addressed through separate sensitivity analyses or characterized qualitatively.
26 Following this introductory section, this chapter discusses the scope of the risk
27 assessment, including selection of urban areas and health endpoints; components of the risk
28 model; characterization of uncertainty and variability associated with the risk estimates; and key
29 results from the assessment. The draft Risk Assessment TSD provides a more detailed
30 discussion of the risk assessment methodology and includes additional risk estimates beyond
31 those summarized herein.
32 5.1.1 Overview of Risk Assessment From Last Review
33 EPA conducted a health risk assessment that produced risk estimates for the number and
34 percent of children and outdoor workers experiencing lung function and respiratory symptoms
November 2005 5-1 Draft - Do Not Quote or Cite
-------
1 associated with the exposures estimated for 9 urban areas. This portion of the risk assessment
2 was based on exposure-response relationships developed from analysis of data from several
3 controlled human exposure studies which was combined with exposure estimates developed for
4 children who spent more time outdoors and for outdoor workers. The risk assessment for the last
5 review also included risk estimates for excess respiratory-related hospital admissions related to
6 Os concentrations for New York City based on a concentration-response relationship reported in
7 an epidemiological study (Thurston et al., 1992). Risk estimates for lung function decrements,
8 respiratory symptoms, and hospital admissions were developed associated with recent air quality
9 levels (referred to as "as is" air quality) and for just meeting the existing 1-hr standard and
10 several alternative 8-hr standards. The methodological approach followed in Conducting the last
11 risk assessment and risk estimates resulting from that assessment are described in Chapter 6 of
12 the 1996 Staff Paper (EPA, 1996b) and in several technical reports and publications (Whitfield et
13 al, 1996; Whitfield, 1997; Whitfield et al., 1998).
14 In me 1997 review of the Os NAAQS,' the risk estimates played a significant role in both
15 the staff recommendations and in the proposed and final decisions to revise the 63 standards.
16 CAS AC stated (Wolff, 1995) in its advice and recommendations to the Administrator on the Os
17 Staff Paper that "EPA's risk assessments must play a central role in identifying an appropriate
18 level," while also noting that "because of the myriad of assumptions that are made to estimate
19 population exposure and risk, large uncertainties exist in these estimates." In the 1997
20 promulgation notice (62 FR 38856) announcing the decision to revise the Os standards EPA
21 indicated that the Administrator considered the results of the exposure and risk analyses and key
22 observations and conclusions from these analyses in putting effects considered to be adverse to
23 individuals into a broader public health perspective and in making judgments about the level of a
24 standard that would be requisite to protect public health with an adequate margin of safety.
25 5.1.2 Development of Approach for Current Risk Assessment
26 The health risk assessment described in this Chapter and in the draft Risk Assessment
27 TSD builds upon the methodology and lessons learned from the risk assessment work conducted
28 for the last review. The current risk assessment also is based on the information evaluated in the
29 second external review draft of the Os CD (EPA, 2005 c). Some aspects of the current risk
30 assessment may change based on changes that may be incorporated in the final Oj CD. The
31 general approach used in the current risk assessment was described in the draft Health
32 Assessment Plan (EPA, 2005a), that was released to the CASAC and general public in April
33 2005 for review and comment and which was the subject of a consultation with the CASAC Os
34 Panel on May 5,2005. The approach used in the current risk assessment reflects consideration
November 2005 5-2 Draft - Do Not Quote or Cite
-------
1 of the comments offered by CAS AC members and the public on the draft Health Assessment
2 Plan.
3 The basic structure of the current risk assessment reflects the two different types of
4 human studies on which the Os health risk assessment is based: controlled human exposure
5 studies and epidemiological studies. Controlled human exposure studies involve volunteer
6 subjects who are exposed while engaged in different exercise regimens to specified levels of
7 under controlled conditions for specified amounts of time. For the current health risk
8 assessment, staff is using probabilistic exposure-response relationships developed during the last
9 review which were based on analysis of individual data that describe the relationship between a
10 measure of personal exposure to 03 and measures of lung function recorded in the studies. The
11 measure of personal exposure to ambient Os is typically some function of hourly exposures -
12 e.g., 1-hr maximum or 8-hr maximum. Therefore, a risk assessment based on exposure-response
13 relationships derived from controlled human exposure study data requires estimates of personal
14 exposure to Os, typically on a 1 -hr or multi-hour basis. Because data on personal hourly Os
15 exposures are not available, estimates of personal exposures to varying ambient concentrations
16 are derived through exposure modeling, as described in Chapter 4.
17 In contrast to the exposure-response relationships derived from controlled human
18 exposure studies, epidemiological studies provide estimated concentration-response
19 relationships based on data collected in real world settings. Ambient O3 concentration is
20 typically measured as the average of monitor-specific measurements, using population-oriented
21 monitors. Population health responses for Os have included population counts of.hospital
22 admissions for respiratory and cardiac illness and premature mortality. As described more fully
23 below, a risk assessment based on epidemiological studies typically requires baseline incidence
24 rates and population data for the risk assessment locations.
25 The characteristics that are relevant to carrying out a risk assessment based on controlled
26 human exposure studies versus one based on epidemiology studies can be summarized as
27 follows:
28 • A risk assessment based on controlled human exposure studies uses exposure-
29 response functions, and therefore requires as input (modeled) personal exposures to
30 ambient Os. A risk assessment based on epidemiological studies uses concentration-
31 response functions, and therefore requires as input (monitored) ambient Os
32 concentrations.
33 • Epidemiological studies are carried out in specific real world locations (e.g., specific
34 urban areas). To minimize uncertainty, a risk assessment based on epidemiological
35 studies has been performed for the locations in which the studies were carried out
36 Controlled human exposure studies, carried out in laboratory settings, are generally not
37 specific to any particular real world location. A controlled human exposure studies-
November 2005
5-3
Draft - Do Not Quote or Cite
-------
1 based risk assessment can therefore appropriately be carried out for any location for
2 which there are adequate air quality and other data on which to base the modeling of
3 personal exposures. There are, therefore, some locations for which a controlled human
4 exposure studies-based risk assessment could appropriately be carried out but an
5 epidemiological studies-based risk assessment could not.
6 • The adequate modeling of hourly personal exposures associated with ambient
7 concentrations for use with exposure-response relationships requires more complete
8 ambient monitoring data than are necessary to estimate average ambient concentrations
9 used to calculate risks based on concentration-response relationships. Therefore, there
10 may be some locations in which an epidemiological studies-based risk assessment
11 could appropriately be carried out but a controlled human exposure studies-based risk
12 assessment could not.
13 • To derive estimates of risk from concentration-response relationships estimated in
14 epidemiological studies, it is usually necessary to, have estimates of the baseline
15 incidences of the health effects involved. Such baseline incidence estimates are not
16 needed in a controlled human exposure studies-based risk assessment
17 The scope of the current Os risk assessment is described in the next section along with air
18 quality considerations that are relevant to both parts of the risk assessment. Then, the methods
19 for the two parts of the risk assessment - the part based on controlled human exposure studies
20 and the part based on epidemiological and field studies - are discussed in sections 5.3.1 and 5.3.2
21 below, followed by presentation and discussion of the 03 risk estimates in section 5.4.
22 S.2 SCOPE OF OZONE HEALTH RISK ASSESSMENT
23 The current Oj health risk assessment estimates risks of various health effects associated
24 with exposure to ambient 63 in a number of urban areas selected to illustrate the public health
25 impacts of this pollutant. The short-term exposure related health endpoints selected for the Os
26 risk assessment, discussed in section 5.2.1, include those for which the draft CD concludes that
27 the weight of the evidence supports the general conclusion that Os, acting alone and/or in
28 combination with other co-pollutants is likely causal. The current risk assessment includes risk
29 estimates for 12 urban areas. The basis for selection of these areas is discussed below (section
30 5.2.2).
31 Another important aspect of the current risk assessment is that the risks estimated are
32 only those associated with ambient 63 concentrations exceeding estimated policy-relevant
33 background levels (hereafter, referred to as "background" in this Chapter).' Risks associated
34" with concentrations above this background are judged to be more relevant to policy decisions
1 Policy relevant background is defined in section 2.7 of this Staff Paper and development of estimates for policy
relevant background for use in the risk assessment are discussed in section 5.2.3.
November 2005 5-4 Draft - Do Not Quote or Cite
-------
1 about the NAAQS than estimates that include risks potentially attributable to uncontrollable
2 background concentrations.
3 5.2.1 Selection of Health Endpoint Categories
4 As noted above, in the last review a significant portion of the health risk assessment
5 involved developing risk estimates for both lung function decrements (> 10, > 15, and > 20%
6 changes in FEV0 and respiratory symptoms in children (age 6 to 18 years old) who spend more
7 time-outdoors and outdoor workers with 1 -hr exposures at moderate and heavy exertion and 8-hr
8 exposures at moderate exertion. As discussed in section 3.3.2.2 and Chapter 6 of the draft CD,
9 there is a significant body of controlled human exposure studies reporting lung function
10 decrements and respiratory symptoms in adults associated with 1- and 6-8-hr exposures to ozone.
11 Consistent with the approach used in the last review, staff judges that it is reasonable to
12 estimate exposure-response relationships for lung function decrements associated with ozone
13 exposures in children 5-18 years old based on data from adult subjects (18-35 years old). As
14 discussed in the 1996 Staff Paper and 1996 CD, findings from other chamber studies
15 (McDonnell et al., 1985) for children 8-11 years old and summer camp field studies in at least
16 six different locations in the U.S. and Canada found lung function decrements in healthy children
17 similar to those observed in heallhy adults exposed to Os under controlled chamber conditions.
18 The same approach is being used in the current assessment.
19 In the prior risk assessment, staff focused on the risk estimates for lung function
20 decrements associated with 1-hr heavy exertion, 1-hr moderate exertion, and 8-hr moderate
21 exertion exposures in children age 5-18 years of age. Since the 8-hr moderate exertion exposure
22 scenario in children who spend more time outdoors clearly resulted in the greatest health risks in
23 terms of lung function decrements, and since no new information published since the last review
24 suggests any changes that would impact this conclusion, staff has included only the lung function
25 decrements (>10,15, and 20% FEVO associated with 8-hr moderate exertion exposures in
26 children (age 5 to 18 years old) in the current risk assessment.
27 Although respiratory symptoms in healthy children were estimated in the last review,
28 staff has not included this endpoint in the current quantitative risk assessment This is because
29 several field studies conducted since the last review failed to find respiratory symptoms in field
30 studies examining responses in healthy children. The draft CD concludes (p. 7-48) that "these
31 studies indicate mat there is no consistent evidence of an association between Os and respiratory
32 symptoms among children."
33 As discussed in section 3.3.2.2, the draft CD also concludes that collectively, the results
34 of field studies suggest that respiratory symptoms and increased medication use in asthmatic
35 children are associated with acute exposure to Os. While these recent studies provide strong
November 2005
5-5
Draft - Do Not Quote or Cite
-------
1 evidence that some asthmatic children are likely to experience O3-related effects, the selective
2 recruitment of inner city asthmatic subjects makes it very difficult to develop quantitative risk
3 estimates for asthmatic children based on these studies.
4 While a number of controlled human exposure studies have reported additional health
5 endpoints associated with short-term exposures to Os, including airway hyperresponsiveness,
6 inflammation, and immune system effects, there is insufficient exposure-response data at
7 different concentrations to develop quantitative risk estimates for these effects. These additional
8 effects are discussed in Chapter 3 and it is important-to recognize that the current quantitative
9 risk assessment only presents a partial picture of the risks to public health associated with short-
10 term Os exposures.
11 As discussed in the draft CD and Chapter 3, a significant number of epidemiological
12 studies examining a variety of health effects associated with ambient Os concentrations in
13 various locations-throughout the U.S., Canada, Europe, and other regions of the world have been
14 published since the last Oj NAAQS review. Chapter 3 reviews the epidemiological evidence
15 evaluated in Chapter 7 of the draft CD. In selecting health endpoints to be included in the
16 current quantitative risk assessment, staff has focused on health endpoints that are better
17 understood in terms of health consequences (i.e., adversity) and endpoint categories for which
18 the weight of the evidence supports the inference of a likely causal relationship between Os and
19 the effect category. Based on these considerations, the following endpoints associated with
20 short-term exposures to Os have been included:
21 • Hospital admissions for respiratory and cardiovascular illness;
22 • Premature total, respiratory, and cardiovascular mortality.
23 As noted in the draft CD (p.7-149), "large multi-city studies, as well as many studies from
24 individual cities have reported an association of Os concentrations with respiratory and
25 cardiovascular hospital admissions. Studies with data restricted to the summer or warm season,
26 in general indicated positive and robust associations between ambient Os concentrations and
27 cardiopulmonary hospital admissions." With respect to acute Os effects on mortality, the draft
28 CD concludes (p.7-149) that "The majority of the studies suggest an elevated risk of all cause
29 mortality associated with acute exposure to Os, especially in the summer or warm season when
30 Os levels are typically high."
31 As discussed in Chapter 7 of the draft CD and section 3.3.2, several additional health
32 endpoints including emergency department visits for respiratory illness, increased respiratory
33 symptoms in asthmatic children, and increased school absences have been reported to be
34 associated with short-term Os exposures. The current quantitative risk assessment does not
35 include these additional health endpoints. Emergency department visits were excluded from the
November 2005 5-6 Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
1 0
11
12
13
14
15
16
17
1 8
1 9
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
quantitative risk assessment because of the limited and less consistent database as well as the
lack of baseline incidence data for emergency department visits, With respect to increased
respiratory symptoms in asthmatic children, the evidence is strong that ambient 63 exposures can
lead to these effects, but the design of these epidemiological studies, which focus on subjects
recruited in inner-city neighborhoods, pose significant problems since there is a lack of baseline
incidence data for this endpoint for asthmatic children residing in urban areas. Further, the staff
judges that the data reporting an association between short-term 03 exposures and school
absences is too limited to include in the current risk assessment.
5.2.2 Selection of Study Areas
The criteria and considerations that went into selection of urban areas for the
assessment included the following:
risk
• The overall set of urban locations should represent a range of geographic areas, urban
populations demographics, and climatology and be focused on areas that do not meet
the current 8-hr O3 NAAQS.
• The largest areas with major 63 nonattainment problems should be included.
• There must be sufficient air quality data for a recent three year period.
• An area should be the same or close to the location where at least one concentration-
response function for the health endpoints included in the assessment has been
estimated by a study that satisfies the study selection criteria (see below). If the study
was a hospital admissions study, then relatively recent location-specific baseline
incidence data had to be available.
• Locations in which more health endpoints have been assessed were preferred to those
with fewer.
Since the exposure-response functions for lung function decrements based on the controlled
human exposure studies were based on controlled laboratory conditions, the location of these
studies played no role in selecting urban locations for the risk assessment.
Based on the selection criteria and considerations listed above, the following urban areas
were included in the risk assessment:
• Atlanta
• Boston
• Chicago
• Cleveland
• Detroit
• Houston
• Los Angeles
• New York City
* Philadelphia
November 2005
5-7
Draft - Do Not Quote or Cite
-------
1 • Sacramento
2 • St. Louis
3 • Washington, D.C.
4
5 As discussed in Chapter 4, for the purposes of estimating population exposure and the
6 risk of lung function decrements associated with these population exposure estimates, the 12
7 urban areas have been defined based on consolidated statistical areas (CSAs). In contrast, for the
8 risk estimates for premature mortality and excess hospital admissions, the urban areas have been
9 defined to be generally consistent with the geographic boundaries used in the epidemiological
10 studies which were the source of the concentration-response functions used in this risk
11 assessment. In most cases the epidemiological studies only included the core urban county or a
12 limited number of counties in each of the 12 urban areas.
13 5.2.3 Air Quality Considerations
14 Both the controlled human exposure and epidemiolgic-based portions of the risk
i
15 assessment include risk estimates for a recent year of air quality (labeled "as is" air quality in the
16 draft Risk Assessment TSD) and for air quality adjusted so that it simulates just meeting the
17 current 8-hr 63 NAAQS. The year selected to represent recent air quality data is 2004, since this
18 is the most recent year for which complete data are currently available.
19 In order to estimate health risks associated with just meeting the current 8-hr 03 NAAQS,
20 it is necessary to estimate the distribution of hourly Oa concentrations that would occur under
21 any given standard. Since compliance with the current 63 standard is based on a 3-year average,
22 air quality data from 2002 to 2004 have been used to determine the amount of reduction in Os
23 concentrations required to meet the current standard. Estimated design values2 are used to
24 determine the adjustment necessary to just meet the current 8-hr daily maximum standard. The
25 amount of control has then been applied to a single year of data (2004) to estimate risks for a
26 single season (April through September) in a single year. As described in section 4.5.6 and in
27 more detail in Rizzo (2005), after considering several approaches, including proportional
28 rollback and Weibull adjustment procedures, staff concluded that the Quadratic air quality
29 adjustment procedure generally best represented the pattern of reductions across the Os air
30 quality distribution observed over the last decade. The Quadratic air quality adjustment
31 procedure was applied to the filled in 2004 Os monitoring data, based on the 3-year period
32 (2002-2004) Os design value, to generate a new time series of hourly Os concentrations that
2 A design value is a statistic that describes the air quality status of a given area relative to the level of the NAAQS..
Design values are often based on multiple years of data, consistent with the specification of the NAAQS in Part 50
of the CFR. For example, for the current OjNAAQS, the 3-year averages of the annual 4th daily maximum 8-hr
average concentration based on the maximum monitor within an urban area are the design values.
November 2005 5-8 Draft - Do Not Quote or Cite
-------
1 reflects air quality levels just meeting the current 8-hr Os standard for this single year. It should
2 " be noted that since compliance with the current standard is based on the 3-year average of the 4th
3 daily maximum 8-hr values, the air quality distribution in each of the 3 years can and generally
4 does vary. As a consequence, the risk estimates associated with air quality just meeting the
5 current standard also will vary depending on the year chosen for the analysis. Staff plans to
6 explore the magnitude of this year-to-year variability in the next draft of the assessment by
7 conducting assessments involving adjusting 2002 air quality (a year with generally higher 03
8 levels in many of the 12 urban study areas) to just meet the current 8-hr standard.
9 As noted earlier, the risk estimates developed for both the recent air quality scenario and
10 just meeting the current 8-hr standard represent risks in excess of estimated background
11 concentrations. The results of the global tropospheric Os model GEOS-CHEM have been used
12 to estimate monthly average background Oa levels for different geographic regions across the
13 U.S. These GEOS-CHEM simulations include a background simulation in which North
14 American anthropogenic emissions of nitrogen oxides, non-methane volatile organic compounds,
15 and carbon monoxide are set to zero, as described in' Fiore et al. (2003). Staff estimated monthly
16 background concentrations for each of the 12 urban areas based on the GEOS-CHEM
17 simulations (Langstaff, 2005a).
18
19 5.3 COMPONENTS OF THE RISK MODEL
20 As noted above in section 5.1.2, there are two parts to the health risk assessment: one
21 based on combining information from controlled human exposure studies with modeled
22 population exposure and the other based on combining information from community
23 epidemiological studies with either monitored or adjusted ambient concentrations levels. Section
24 5.3.1 below discusses the portion of the current risk assessment related to effects reported in
25 controlled human exposure studies and section 5.3.2 below discusses the portion of the current
26 risk assessment related to health effects reported in community epidemiological studies.
27 5.3.1 Assessment of Risk Based on Controlled Human Exposure Studies
28 5.3.1.1 General Approach
29 The maj or components of the portion of the health risk assessment based on data from
30 controlled human exposure studies are illustrated in Figure 5-1. As shown in Figure 5-1, under
31 this portion of the risk assessment, exposure estimates for a number of different air quality
32 scenarios (i.e, recent year of air quality, just meeting the current 8-hr standard, just meeting
33 alternative standards, and background) are combined with probabilistic exposure-response
34 relationships derived from the controlled human exposure studies to develop risk estimates
November 2005
5-9
Draft - Dp Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
for recent air quality and just meeting the current and alternative standards in excess of
background. As discussed above, the health effect included in this portion of the risk assessment
is lung function decrement, as measured by FEVi. The air quality and exposure analysis
components that are integral to this portion of the risk assessment are discussed in greater detail
in Chapter 4 and in the draft Exposure Assessment TSD.
Several risk measures were generated for this portion of the risk assessment. In addition
to the estimates of the number of school age children and active children experiencing 1
or more occurrences of a lung function decrement >10, >15, and >20% in an Os season, risk
estimates have been developed for the total number of occurrences of these lung function
decrements in school age children and active school age children.
A population risk estimate for a given lung function decrement (e.g., ^20% change in
FEVi) is an estimate of the expected number of people who will experience that lung function
decrement Since staff is interested in risk estimates associated with Oa concentrations in excess
of background concentrations, the following steps were taken: (1) expected risk given the
personal exposures associated with recent ambient Oa concentrations was estimated, (2) expected
risk given the personal exposures associated with estimated background ambient Os
concentrations was estimated, and (3) the latter was subtracted from the former. As shown in
Equation 5-1 below, the population risk is then calculated by multiplying the resulting expected
risk by the number of people in the relevant population. Because response rates are calculated
for 21 fractiles, estimated population risks are similarly tractile-specific.
November 2005
5-10
Draft - Do Not Quote or Cite
-------
W
15 -3 § °>
2 < ,_ -s 1 -s
.E -p >. n eo to -o c TZ
UJ g « § S J £
3 KOO w< w
ic . . .
t
£ .* *«
ro .2 ^
4*S
j L
W «^
£ £ T3
Cu ^^ ™
E£ » 3-g
S ? T. g 0 §
»Ic!l*«J3
§s l« 1 1 1 S
88iS|Sg51
Sl*3*y?*
t
a
1 S, '
2 5? "S
B, o o
^ 8-5
Ed UJ
ik j i j k i
^ •« «»
?ffll
8^t^
*6|0«
"
Jf St^s Sll III Si
i ii|| |l! oil |i
§ <§S Isl 5fe S|
Z, =5 m3 <
^
fe
1
t
%
•-i (N
-------
1 The risk (i.e., expected fractional response rate) for the k* fractile, Rt is:
2
N «t
3 Rk = £PJ x W I e>•) - ZP-* x W Ie-*) (Equati<>n 5-1)
>i w
4 where:
5
6 ey = (the midpoint of) the jth category of personal exposure to ozone, given recent
7 ambient Os concentrations;
8
9 ef = (the midpoint of) the ith category of personal exposure to ozone, given background
10 ambient 03 concentrations;
11
12 Pj = the fraction of the population having personal exposures to Os concentration of BJ
*•
13 ppm, given recent ambient Os concentrations;
14
15 Pf = the fraction of the population having personal exposures to Oj concentration of
16 ef ppm, given background ambient Os concentrations;
17
18 RRk | Cj = k-fractile response rate at Os concentration q;
19
20 RRk | ef = k-fractile response rate at Os concentration e\; and
21
22 N=number of intervals (categories) of Os personal exposure concentration, given recent
23 ambient Os concentrations; and
24
25 Nb - number of intervals of Os personal exposure concentration, given background
26 ambient Os concentrations.
27
28 For example, if the median expected response rate for recent ambient concentrations is
*
29 0.065 (i.e., the median expected fraction of the population responding is 6.5%) and the. median
30 expected response rate for background ambient concentrations is 0.001 (i.e., the median expected
31 fraction of the population responding is 0.1 %), then the median expected response rate
32 associated with recent ambient concentrations above background concentrations is 0.065 - 0.001
t
November 2005 5-12 Draft - Do Not Quote or Cite
-------
1 = 0.064. If there are 300,000 people in the relevant population, then the population risk is 0.064
2 x 300,000 = 19,200.
3 5.3.1.2 Exposure Estimates
4 Exposure estimates used in this portion of the risk assessment were obtained from
5 running TRIM.Expo for each of the 12 urban areas for the various air quality scenarios (i.e., for
6 2004 air quality representing a recent year, for 2004 air quality adjusted to just meet the current
7 8-hr standard, and for air quality levels representing background based on estimates from the
8 GEOS-CHEM model). Chapter 4 and the draft Exposure Assessment TSD (EPA, 2005d)
9 provide additional details about the inputs and methodology used to estimate population
10 exposure in the 12 urban areas. Exposure estimates for all and "active" school age'children (ages
11 5 to 18) were separately combined with probabilistic exposure-response relationships for lung
12 function decrements associated with 8-hr exposure while engaged in moderate exertion.
13 Children were characterized as active if their median daily physical activity index (see section
14 4.4.3) over the period modeled was 1.75 or higher, a level characterized by exercise
15 physiologists as being "moderately active" or "active." Individuals engaged in activities that
16 resulted in an average equivalent ventilation rate (EVR) for the 8-hr period in the range of 13 to
17 271/min-m2 were included in the exposure estimates for 8-hr moderate exertion. This range was
18 selected to match the range of EVR for the group of subj ects in the controlled human exposure
19 studies that were the basis for the exposure-response relationships used in this portion of the risk
20 assessment.
21 53.1.3 Exposure Response Functions
22 A similar methodology to that developed in the prior risk assessment has been used to
23 estimate probabilistic exposure-response relationships for lung function decrements in school age
24 children and active school age children associated with 8-hr moderate exertion exposures. As in
25 the prior assessment, the combined data set from the Folinsbee et al. (1988), Horstman et al.
26 (1990), and McDonnell et al. (1991) studies have been used to estimate exposure-response
27 relationships for 8-hr exposures. Data from these controlled human exposure studies were
28 corrected for the effect of exercise in clean air to remove any systematic bias that might be
29 present in the data attributable to an exercise effect. Generally, this correction for exercise in
30 clean air was small relative to the total effects measures in the Os-exposed cases. Regression
31 techniques were then used .to fit a function to the data for each of the three measures1 of lung
32 function decrement. In each case, a linear function provided a'good fit.3
3 As noted in Whitfield et al., 1996, the response data point associated with 0.12 ppm for the response measure
FEV1 > 15% appeared to be inconsistent with the other data points (see Whitfield et al., 1996, Table 10, footnote c).
November 2005 • 5-13 Draft - Do Not Quote or Cite
-------
1 53.1.4 Characterizing Uncertainty and Variability
2 An important issue associated with any population health risk assessment is the
3 characterization of uncertainty and variability. Uncertainty refers to the lack of knowledge
4 regarding both the actual values of model input variables (parameter uncertainty) and the
5 physical systems or relationships (model uncertainty - e.g., the shapes of concentration-response
6 functions). In any risk assessment uncertainly is, ideally, reduced to the maximum extent
7 possible, .but significant uncertainty often remains. It can be reduced by improved measurement
8 and improved model formulation. In addition, the degree of uncertainty can be characterized,
9 sometimes quantitatively. For example, the statistical uncertainty surrounding the estimated Oa
10 coefficients in the exposure-response functions is reflected in the credible intervals provided for
11 the risk estimates in this chapter and in the draft Risk Assessment TSD.
12 A Bayesian approach was used to characterize uncertainty attributable to sampling error
13 based on sample size considerations at each of the three Oa concentrations for which there were
14 data from the underlying studies (0.08, 0.10, and 0.12 ppm). Using diffuse Beta distributions as
15 priors distributions, the resulting posterior distributions are also Beta distributions (see Appendix
16 A in Whitfield et al., 1996). Response rates were calculated for 21 fractiles (for cumulative
17 probabilities from 0.05 to 0.95 in steps of 0.05) plus probabilities of 0.01 and 0.990 using these
18 posterior Beta distributions. For other concentrations, regression techniques were used to fit a
19 linear function through the three points at Os concentrations of 0.08, 0.10, and 0.12 ppm and then
20 this function was used to generate response rates for all 21 fractiles at specified Og
21 concentrations. If the estimated response rate was less than 0.0, it was set equal to 0.0 and if the
22 response rate was greater than 1.0, it was set to 1.0.
23 In addition to uncertainties arising from sample size considerations, other uncertainties
24 associated with the use of the exposure-response relationships for lung function responses are
25 briefly summarized below. Additional uncertainties with respect to the exposure inputs to the
26 risk assessment are described in Chapter 4 and the draft Exposure Assessment TSD. These
27 additional uncertainties include:
28 • Length of exposure. The 8-hr moderate exertion risk estimates are based on a
29 combined data set from three controlled human exposure studies conducted using 6.6-
30 hr exposures. The use of these data to estimate responses associated with an 8-hr
31 exposure are reasonable, in our judgment, because lung function response appears to
32 level off after exposure for 4 to 6 hours. It is unlikely that the exposure-response
33 relationships would have been appreciably different had the studies been conducted
34 over an 8-hr period.
Because of this, we estimated the probability of a response of FEV1 Z 15% at an O3 concentration of 0.12 ppm by
interpolating between the FEV1 £ 10% and FEV1 £ 20% response rates at that O3 concentration.
November 2005 5-14 Draft - Do Not Quote or Cite
-------
1 * Extrapolation of exposure-response relationships. It was necessary to estimate
2 responses at Oa levels below the lowest exposure levels used in the controlled human
3 studies (i.e., 0.08 ppm). In the prior review the CASAC Os Panel supported the staffs
4 decision to use a linear extrapolation approach down to background levels. The same •'
5 extrapolation has been applied in the current risk assessment.
6 • Reproducibilitv of Ounduced responses. The risk assessment assumed that the (V
7 induced responses for individuals are reproducible. This assumption is supported by
8 the evaluation in the draft CD (see section AX6.4) which cites studies by McDonnell et
9 al. (1985b) and Hazucha et al. (2003) as showing significant reproducibility of
10 response.
11 • Age and lung function response. As in the prior review, exposure-response
12 relationships based on controlled human exposure studies involving 18-35 year old
13 subjects were used in the risk assessment to estimate responses for school age children
14 (ages 5-18). This approach is supported by evaluation in the draft CD (see section
15 AX6.4) which cites the findings of McDonnell et al. (1985a) who reported that children
16 8-11 years old experienced FEVi responses similar to those observed in adults 18-35
17 years old when both groups were exposed to concentrations of 0.12 ppm at an EVR of
18 35 L/min/m2. The draft CD also notes that Hazucha et al. (2003) similarly observed
19 generally reproducible Ch-induced lung function responses in a controlled human
20 exposure study. In addition, a number of summer camp studies of school age children
21 exposed in outdoor environments in the Northeast also showed Os-induced lung
22 function changes similar to those observed in controlled human exposure studies.
23 • Exposure history. The risk assessment assumed that the Os-induced response on any
24 given day is independent of previous Os exposures. As discussed in Chapter 3 and in
25 the draft CD, Cvinduced responses can be enhanced or attenuated as a result of recent
26 prior exposures. The possible impact of exposure history on the risk estimates is an
27 additional source of uncertainty that is not quantified in Ihis assessment.
28 • Interaction between 0^ and other pollutants. Because the controlled human exposure
29 studies used in the risk assessment involved only 03 exposures, it was assumed that
30 estimates of Os-induced health responses would not be affected by the presence of
31 olher pollutants (e.g., SOi, PM2.5, etc). Some evidence exists that other pollutants may
32 enhance the respiratory effects associated with exposure to C% but the evidence is not
33 consistent across studies.
34 Variability refers to the heterogeneity in a population or variable of interest that is
35 inherent and cannot be reduced through further research. The current controlled human exposure
36 studies portion of the risk assessment incorporates some of the variability in key inputs to the
37 analysis by using location-specific inputs for the exposure analysis (e.g., location-specific
38 population data, air exchange rates, air quality and temperature data). Although spatial
39 variability in these key inputs across all U.S. locations has not been fully characterized,
40 variability across the selected locations is embedded in the analysis by using, to the extent
41 possible, inputs specific to each urban area. Temporal variability is more difficult to address,
42 because the risk assessment focuses on some unspecified time in the future. To minimize the
November 2005 5-15 Draft - Do Not Quote or Cite
-------
1 degree to which values of inputs to the analysis may be different from the values of those inputs
2 at that unspecified time, we have used the most current inputs available - for example, year 2004
3 air quality data for all of the urban locations, and the most recent available population data (from
4 the 2000 Census). However, future changes in inputs have not been predicted (e.g., future
5 population levels).
6 5.3.2 Assessment of Risk Based on Epidemiological Studies
7 As discussed above, the current quantitative risk assessment based on epidemiological
8 studies includes risk estimates for cardiopulmonary-related hospital admissions and total and
9 cardiopulmonary mortality associated with short-term 63 exposures in selected urban locations
10 in the U.S. The methods used in this portion of the risk assessment are described below.
11 5.3.2.1 General Approach
12 In order to estimate the incidence of a particular health effect associated with recent
13 conditions in a specific county or set of counties attributable to ambient Os exposures in excess
14 of background, as well as the change in incidence of the health effect in that county or set of
15 counties corresponding to a given change in Os levels resulting from just meeting the current or
16 alternative 8-hr Oa standards, the following three elements are required:
17 • Air quality information including: (1) recent air quality data for Oa from population-
18 oriented monitors in the assessment location, (2) estimates of background Os
19 concentrations appropriate to this location, and (3) recent concentrations adjusted to
20 reflect patterns of air quality estimated to occur when the area just meets the specified
21 standards. (These air quality inputs are discussed in more detail in section 4.5.6)
22 * Concentration-response function(s) which provide an estimate of the relationship
23 between the health endpoint of interest and ambient Os concentrations, preferably derived
24 in the assessment location, although functions estimated in other locations can be used at
25 the cost of increased uncertainty.
26 • Seasonal baseline health effects incidence rate and population. The baseline
27 incidence rate provides an estimate of the incidence rate in the assessment location
28 corresponding to recent Oa levels in that location
29
30 Figure 5-2 provides a broad schematic depicting the role of these components in this part of the
31 risk assessment. Each of the key components (i.e., air quality information, estimated
32 concentration-response functions, and baseline incidence and population data) is discussed
33 below, highlighting those points at which judgments have been made.
34 These inputs are combined to estimate health effect incidence changes associated with
35 specified changes in Os levels. Although some epidemiological studies have estimated linear or
36 logistic concentration-response functions, by far the most common form is the exponential (or
37 log-linear) form:
November 2005 5-16 Draft - Do Not Quote or Cite
-------
O
(0
I *
I < _ "S J "S
*s -g £. "c
J
^t
D
3
k
It
0
'S
<
s
i
i
j
c5
->
0
a
)
i
a
8
j
)
D
i
Modeled Background
i
<
<
<
.
c
k
Concentrations
_
10
1)
00
i
ta_
J
c
.O
Q
i
•s
f
t
Current and Alternative
L
Proposed Standards
en
S
.2
1
u
• o
(
,
c
o
ro
-------
1 y = Beftf, (Equation 5-2)
2
3 where x is the ambient O$ level, y is the incidence of the health endpoint of interest at 03 level x,
4 pis the coefficient of ambient 63 concentration, and B is the incidence at x=0, i.e., when there is
5 no ambient 63. The relationship between a specified ambient Os level, XQ, for example, and the
6 incidence of a given health endpoint associated with that level (denoted as yo) is then
7
8 y0=Beftc°. (Equation 5-3)
9
10 Because the log-linear form of concentration-response function (equation (5-2)) is by far the
11 most common form, we use this form to illustrate the derivation of the "health impact function"
12 used in this portion of the risk assessment.
13
14 The difference in health effects incidence, Ay = y0 - y, from yfl to the baseline incidence
15 rate, y, corresponding to a given difference in ambient 03 levels, Ax = XQ - x, can be derived by
16 dividing equation (5-3) by equation (5-2), which yields:
17
18 Ay = y[ep** -1] . (Equation 5-4)
19
20 Alternatively, the difference in health effects incidence can be calculated indirectly using relative
21 risk. Relative risk (RR) is a measure commonly used by epidemiologists to characterize the
22 comparative health effects associated with a particular air quality comparison. The risk of
23 mortality at ambient Os level xo relative to the risk of mortality at ambient Os level x, for
24 example, may be characterized by the ratio of the two mortality rates: the mortality rate among
25 individuals when the ambient Os level is XQ and the mortality rate among (otherwise identical)
26 individuals when the ambient Os level is x. This is the RR for mortality associated with the
27 difference between the two ambient Os levels, x0 and x. Given a concentration-response function
28 of the form shown in equation (5-1) and a particular difference in ambient 63 levels, Ax, the RR
29 associated with that difference in ambient O3, denoted as RR^, is equal to epAx . The difference
30 in health effects incidence, Ay, corresponding to a given difference in ambient Os levels, Ax, can
31 men be calculated based on this RRax:
32
33 &y = y[RRfr -1]. (Equation 5-5)
34
November 2005 5-18 Draft - Do Not Quote or Cite
-------
1 Equations (5-4) and (5-5) are simply alternative ways of expressing the relationship between a
2 given difference in ambient Os levels, Ax, and the corresponding difference in health effects
3 incidence, Ay. These health impact equations are the key equations that combine air quality
4 information, concentration-response function information, and baseline health effects incidence
5 information to estimate ambient Os health risk.
6 5.3.2.2 Air Quality Considerations
7 As illustrated in Figure 5-2, and noted earlier, air quality information required to conduct
8 the Oa risk assessment includes: (1) recent air quality data for Os from suitable monitors for each
9 selected location, (2) estimates of background concentrations for each selected location, and (3)
10 air quality adjustment procedures to modify the recent data to reflect changes in the distribution
11 of hourly 63 air quality estimated to occur when an area just meets a given Os standard. Staff
12 retrieved Os ambient air quality data for the years 2002 through 2004 from EPA's Air Quality
13 System(AQS).
14 To estimate the change in incidence of a health effect associated with a change in Os
15 concentrations from recent levels to background levels in an assessment location, two time series
16 of Os concentrations are needed for that location: (1) hourly Os concentrations from a recent
17 year, and (2) hourly background Os concentrations. In order to be consistent with the approach
18 generally used in the epidemiological studies that estimated Os concentration-response functions,
19 the (spatial) average ambient Os concentration on each hour for which measured data are
20 available is deemed most appropriate for the risk assessment. A composite monitor data set was
21 created for each assessment location based on averaging each hourly value from all monitors
22 eligible for comparison with the current standard for each hour of the day. Table 4-6 provides a
23 summary of the design values for the 12 urban study areas. Table 5B-1 (Appendix 5B) provides
24 more detailed information on ambient Os concentrations for these locations, including 1 -hr and
25 24-hr average statistics across monitors in each location and the composite monitor values used
26 in this part of the risk assessment.
27 Different exposure metrics have been used in epidemiological Os studies, including the
28 24-hr average and the daily 1-hr maximum. Therefore, daily changes at the composite monitor
29 in the Os exposure metric appropriate to a given concentration-response function were calculated
30 (see Table A-13, Appendix A for summary statistics for the composite monitor Os concentrations
31 in the 12 urban locations). For example, if a concentration-response function related daily
32 mortality to daily 1-hr maximum Os concentrations, the daily changes in 1-hr maximum Os
33 concentrations at the composite monitor were calculated. In the first part of the epidemiology -
34 based risk assessment, in which risks associated with the recent levels of Os above background
35 levels were estimated, this required the following steps:
November 2005 ' 5-19 Draft - Do Not Quote or Cite
-------
1 • Using the monitor-specific input streams of hourly Os concentrations from a recent
2 year, calculate a stream of hourly Os concentrations at the composite monitor. The
3 recent Os concentration at the composite monitor for a given hour on a given day is the
4 average of the monitor-specific Os concentrations for that hour on that day.
5 • Using the stream of hourly Os concentrations from a recent year at the composite
6 monitor, just created, calculate the 1-hr maximum Os concentration for each day at the
7 composite monitor.
8 • Using the monitor-specific input streams of hourly background 03 concentrations,
9 calculate a stream of hourly background Os concentrations at the composite monitor.
10 • Using the stream of background hourly Os concentrations at the composite monitor,
11 just created, calculate the 1 -hr maximum background Os concentration for each day at
12 the composite monitor.
13 • For each day, calculate Ax = (the 1 -hr maximum Os concentration for that day at the
14 composite monitor) (the 1 -hr maximum background Os concentration for that day at the
15 composite monitor).
16
17 The calculations for the second part of the epidemiology-based risk assessment, in which
18 risks associated with estimated Os levels that just meet the current standard above background
19 levels were estimated, were done analogously. For this case the monitor-specific series of
20 adjusted hourly concentrations were used rather than the monitor-specific series of recent
21 monitored hourly concentrations. Similarly, calculations for concentration-response functions
22 that used a different exposure metric (e.g., the 24-hr average) were done analogously, using the
23 exposure metric appropriate to the concentration-response function.
24 5.3.2.3 Concentration-Response Functions
25 As indicated in Figure 5-2, another key component in the epidemiological-based risk
26 model is the set of concentration-response functions which provide estimates of the relationships
27 between each health endpoint of interest and ambient concentrations. As discussed above, the
28 health endpoints that have been included in the Os risk assessment include mortality and hospital
29 admissions associated with short-term exposures. Once it has been determined that a health
30 endpoint is to be included in the assessment, the assessment includes all estimates of response
31 magnitude from studies judged suitable for inclusion in this assessment, including those which
32 are not statistically significant. Effect estimates that are not statistically significant are used from
33 studies judged suitable for inclusion in this assessment to avoid introducing bias into the estimate
34 of the magnitude of the effect. Table 5-1 summarizes the studies included in this part of the risk
35 assessment for each of the urban locations.
36 Studies often report more than one estimated concentration-response function for the
37 same location and health endpoint. Sometimes models including different sets of co-pollutants
November 2005 5-20 Draft - Do Not Quote or Cite
-------
1 are estimated in a study; sometimes different lags are estimated. In some cases, two or more
2 studies estimated a concentration-response function for Os and the same health endpoint in the
3 same location (this is the case, for example, with Os and mortality associated with short-term
4 exposures). For some health endpoints, there are studies that estimated multi-city Os
5 concentration-response functions, while other studies estimated single-city functions.
6 All else being equal, a concentration-response function estimated in the assessment
7 location is preferable to a function estimated elsewhere, since it avoids uncertainties related to
8 potential differences due to geographic location. That is why the urban areas selected for the
9 epidemiological studies-based part of the Os risk assessment are those locations in which
10 concentration-response functions have been estimated. There are several advantages, however,
11 to using estimates from multi-city studies versus studies carried out in single cities. These
12 advantages include, but are not limited to: (1) more precise effect estimates due to larger data
13 sets, (2) greater consistency in data handling and model specification that can eliminate city-to-
14 city variation due to study design, and (3) less likelihood of publication bias or exclusion of
15 reporting of negative or nonsignificant findings. Multi-city studies are applicable to a variety of
16 settings, since they estimate a central tendency across multiple locations. When they are
17 estimating a single concentration-response function based on several cities, multi-city studies
18 also tend to have more statistical power and provide effect estimates with relatively greater
19 precision than single city studies due to larger sample sizes, reducing the uncertainty around the
20 estimated coefficient. Because single-city and multi-city studies have different advantages,
21 where both are available for a given location, risk estimates are presented for both functions.
22 In summary:
23 • if a single-city concentration-response function was estimated in a risk assessment
24 location and a multi-city function which includes that location was also available for
25 the same health endpoint, both functions were included for that location in the risk
26 assessment;
27 • risk estimates based on both single- and multi-pollutant models were used when both
28 were available;
November 2005
5-21
Draft - Do Not Quote or Cite
-------
1 Table 5-1. Locations and Health Endpoints Included in the 03 Risk Assessment Based on
2 Epidemiological Studies*
Urban Area
Atlanta
Boston
Chicago
Cleveland
Detroit
Houston
Los Angeles
New York
Philadelphia
Sacramento
St. Louis
Washington, D.C.
Premature Total Mortality or
Cardiorespiratory Mortality
Bell etal, (2004)
Bell et al. (2004) - 95 cities
Huang et al. (2004)
Huang et al. (2004) - 19 cities
Bell et al. (2004) - 95 cities
Bell et al. (2004) - 95 cities
Huang et al. (2004)
Huang et al. (2004) - 1 9 cities
Schwartz (2004)
Schwartz (2004) - 14 cities
Bell et al. (2004)
Bell et al. (2004) - 95 cities
Huang et al. (2004)
Huang et al. (2004) - 19 cities
Bell etal. (2004)
Bell et al. (2004) - 95 cities
Huang et al. (2004)
Huang et al. (2004) - 1 9 cities
Schwartz (2004)
Schwartz (2004) - 14 cities
Ito (2003)
Bell et al. (2004)
Bell et al. (2004) - 95 cities
Huang etal. (2004)
Huang et al. (2004) - 1 9 cities
Schwartz (2004)
Schwartz (2004) - 14 cities
Bell et al. (2004)
Bell etal. (2004) -95 cities
Huang etal. (2004)
Huang et al. (2004) - 1 9 cities
Bell et al. (2004) - 95 cities
Huang etal. (2004)
Huang et al. (2004) - 19 cities
Bell et al. (2004) - 95 cities
Huang et al. (2004)
Huang et al. (2004) - 19 cities
Moolgavkar et al. (1995)
Bell et al. (2004)
Bell et al. (2004) - 95 cities
Bell et al. (2004)
Bell et al. (2004) - 95 cities
Bell et al. (2004) - 95 cities
Hospital Admissions for Respiratory and
Cardiovascular Dines ses
Schwartz etal. (1996)
ItO (2003)
Linn et al. (2000)
Thurston etal. (1992)
November 2005
5-22
Draft - Do Not Quote or Cite
-------
1 • distributed lag models were used, when available; when a study reported several single
2 lag models for a health effect, the initial selection of the appropriate lag structure for
3 the health effect was based on the overall assessment in the draft CD, considering all
4 studies reporting concentration-response functions for that health effect.
5 The locations, health endpoints, studies, and concentration-response functions included in
6 that portion of the risk assessment based on epidemiological studies are summarized in Table
7 5A-1 (Appendix 5A).
8 5.3.2.4 Baseline Health Effects Incidence and Population Estimates
9 As illustrated in Equation 5-4, the most common epidemiological-based health risk
10 model expresses the reduction in health risk (Ay) associated with a given reduction in Os
11 concentrations (Ax) as a percentage of the baseline incidence (y). To accurately assess the
12 impact of changes in Os air quality on health risk in the selected urban areas, information on the
13 baseline incidence of health effects (i.e., the incidence under recent air quality conditions) in
14 each location is therefore needed. Population sizes, for both total population and various age
15 ranges used in the risk assessment were obtained for the year 2002 (need citation) and are
16 summarized in Table 5-2. Where possible, county-specific incidence or incidence rates have
17 been used in the assessment. County specific mortality incidences were available for the year
18 2002 from CDC Wonder (CDC, 2005), an interface for public health data dissemination
19 provided by the Centers for Disease Control (CDC). The baseline mortality rates for each risk
20 assessment location are provided in Table 5-3 and are expressed as a rate per 100,000
21 population.4
22 County-specific rates for cardiovascular and respiratory hospital discharges, and various
23 subcategories (e.g., asthma, pneumonia, ischemic heart disease), have been obtained, where
24 possible, from state, local, and regional health departments and hospital planning commissions
25 for each of the risk assessment locations.5 Baseline hospitalization rates used in each risk
26 assessment location are summarized in Table 5-4 and are expressed as a rate per 100,000
27 relevant population.
4 Since the baseline incidence rates are expressed in terms of cases per 100,000 general population, the general
population estimates have been used in combination with these rates to generate the baseline incidence in each
location for non-accidental and various types of mortality in calculating the risk estimates.
5 The data were annual hospital discharge data, which were used as a proxy for hospital admissions. Hospital
discharges are issued to all people who are admitted to the hospital, including those who die in the hospital. Use of
the annual or seasonal discharge rate is based on the assumption that the admissions at the end of the year (or
season) that carry over to the beginning of the next year (or season), and are therefore not included in the discharge
data, are offset by the admissions in the previous year (or season) that cany over to the beginning of the current year
(or season).
November 2005 5-23 Draft - Do Not Quote or Cite
-------
I
•-S
en
e
o
i
a.
e
0.
I
i
z
i
millions
H
&
a
•I
Q,
£
At
M
Al
•t
2
o
H
V
B
ti
0
O
0
,
j
i
in
»— i
Fulton, DeKalb
89
1
<
,
i
'•
r-
d
|
Boston
,
1
i
in
•S
O
u
1
'S
u
ts
Q
i
l-H
Cuyahoga
Cleveland
1
i
i
fS
I
Detroit
,
!
•
3
33
Houston
,
I
j
o"
Los Angeles
Los Angeles
,
i
00
!
u
°
,y
•3
ra
E
oa
1
Los Angeles, Riversidi
Los Angeles
j
i
Os
00
1
O
§
1
o
(5
•*
"B
:2
1
a
w"
DC
.S
I
b
e
Z
1
!
i
o
00°
•ti
1
o
OS
«
•S
(2
^
z
I
w
I
n
|
z
1
1
in
Philadelphia
Philadelphia
,
!
i
2
j
Sacramento
,
!
i
o
St Louis City
J
,
i
i
VO
0
Washington, D.C.
U
ft
Washington,
1
'So
1
i
o
1
r— t
-------
i
6
.&
¥
I
a
o -
CM
V
a.
^M^
Oi
I
^
e
S
I
1
es
oa
s
Ji
5
^
& 4)
t A
Type of Mo
(ICD-9 Ct
Is
-is
oT NO
o ^t
^ 6?
1 |
U *— '
vascular and
yr (390-448; 490-
, 480-486; 507)
O ti'tk
'•3 $ cc
uS*
a!
1
8c?
99
i 0.x
R
O
S5
I
o
•3
a
^
0
i
i
,
(S
VO
Fulton, DeKalb
1 Atlanta
i
i
vo
r-
Suffolk
1 Boston
j
i
ON
00
00
o
o
1
1
i
i
00
00
m
n
^H
Cuyahoga
1 Cleveland
JS
in
1-H
^
2
ON
|
[Detroit
i
i
CO
(S
T-H
co
E
1 Houston
i
i
m
ON
V)
Los Angeles
W9
1 Los Angele
i
i
ON
ON
r— 1
s
r~
*!.
J2I ^
«" «
§1
A-o
Bronx, Kings, C
York, Richmon
1
i
i
i
cs
s
£?;
,1
r--
Philadelphia
.a
| Philadelph
I
I
00
Sacramento
o
| Sacrament
i
i
i
£
I"""!
St. Louis City
•3
•
i
i
cs
ON
U
Q
O
Q
*.
o
1 Washingto
o
oo
§
f-H
^—t
S
[***
i
| National
S
(2 w
^ d
s2.
O o
1 -
R 1)
||
5;°°"
8 S
§2
••a &
So-
il
-< -
-------
1 Table 5-4. Baseline Rates for Hospital Admissions
Relevant Population
Rate per 100,000 Relevant Population*
Los
Angeles'
Ages 30+
New
York2
All Ages
Detroit3
Ages
65+
Cleveland4
Ages 65+
Admissions for:
Cardiovascular illness (DRG Codes 103 - 144) - spring
Cardiovascular illness (DRG Codes 103'- 144) - summer
Pulmonary illness (DRG Codes 75 - 1 01) - spring
Pulmonary illness (DRG Codes 75 - 101) - summer
Respiratory illness (ICD codes 466, 480-486, 490, 491,
492, 493)
Asthma (ICD code 493)
Pneumonia (ICD codes 480-486)
COPD (ICD codes 490-496)
Ischemic heart disease (ICD codes 410-414)
Heart failure (ICD code, 428)
Dysrhythmias (ICD code 427)
Respiratory illness ((ICD codes 460-519)
-
431
421
208
174
—
—
—
...
--.-
—
—
—
.„
—
—
„.
800
327
—
-.
—
—
—
...
...
—
„.
.
...
2,068
1,593
4,030
2,822
1,330
...
—
—
...
-"
—
—
...
...
—.
—
—
3,632
2
3
4
5
6
7
8
Rates of unscheduled hospital admissions were calculated from patient discharge data for 1999, obtained from
California's Office of Statewide Health Planning and Development, which also provided records of hospital
admissions for the study by Linn et al. (2000).
2 Rates of unscheduled hospital admissions were calculated from patient discharge data for 2001, obtained from the
New York Statewide Planning and Research Cooperative.
3 Rates were calculated from hospitalization data for Wayne Comity for the year 2000, obtained from the Michigan
Health and Hospital Association in April 2002.
November 2005
5-26
Draft - Do Not Quote or Cite
-------
1 5.3.2.5 Characterizing Uncertainty and Variability
2 Section 5.3.1.4 previously defined what is meant by uncertainty and variability in the
3 context of this risk assessment. For the epidemiological-based portion of the risk assessment, the
4 statistical uncertainty surrounding the estimated 0.3 coefficients in the reported concentration-
5 response functions is reflected in the confidence or credible intervals provided for the risk
6 estimates in this chapter and in the draft Risk Assessment TSD. Additional uncertainties have
7 been addressed quantitatively through sensitivity analyses and/or have been discussed throughout
8 section 5.3.
9 With respect to variability within the epidemioiogical-based portion of the risk
10 assessment, there may be variability among concentration-response functions describing the
11 relation between Q3 and mortality across urban areas. This variability may be due to differences
12 in population (e.g., age distribution), population activities that affect exposure to Oa (e.g., use of
13 air conditioning), levels and composition of co-pollutants, and/or other factors that vary across
14 urban areas.
15 ' The current risk assessment incorporates some of the variability in key inputs to the
16 analysis by using location-specific inputs (e.g., location-specific concentration-response
17 functions, baseline incidence rates, and air quality data). Although spatial variability in these
18 key inputs across all U.S. locations has not been fully characterized, variability across the
19 selected locations is imbedded in the analysis by using, to the extent possible, inputs specific to
20 each urban area Temporal variability is more difficult to address, because the risk assessment
21 focuses on some unspecified time in the future. To minimize the degree to which values of
22 inputs to the analysis may be different from the values of those inputs at that unspecified time,
23 we have used the most current inputs available - for example, year 2004 air quality data for all of
24 the urban locations, and the most recent available mortality baseline incidence rates (from 2002).
25 However, future changes in inputs have not been predicted (e.g., future population levels or
26 possible changes in baseline incidence rates).
27 A number of important sources of uncertainty were addressed where possible. Section
28 4.19 in the draft Risk Assessment TSD discusses in greater detail the uncertainties and variability
29 present in the health risk assessment. The following is a brief discussion of the major sources of
30 uncertainty and variability in the epidemiological portion of the risk assessment and how they are
31 dealt with or considered in the risk assessment:
32 • Causality. There is uncertainty about whether each of the estimated associations
33 between Oj indicators and the various health endpoints included in this risk assessment
34 actually reflect a causal relationship. The staffs judgment, as discussed in more detail
35 in Chapter 3, is that for the health effects included in the risk assessment (i.e, total non-
36 accidental mortality, cardiorespiratory mortality, cardiovascular and respiratory
November 2005
5-27
Draft - Do Not Quote or Cite
-------
1 hospital admissions) there is a likely causal relationship with short-term 0$ exposures,
2 especially during the warm Os season.
3 • Empirically estimated concentration-response relationships. In estimating the
4 concentration-response relationships, there are uncertainties: (1) surrounding estimates
5 of 63 coefficients in concentration-response functions used in the assessment, (2)
6 concerning the specification of the concentration-response model (including the shape
7 of the relationship) and whether or not a population threshold or non-linear relationship
8 exists within the range of concentrations examined in the studies, (3) related to the
9 extent to which concentration-response relationships derived from studies in a given
10 location and time when Os levels were higher or behavior and/or housing conditions
11 were different provide accurate representations of the relationships for the same
12 locations with lower air quality distributions and different behavior and/or housing
13 conditions, and (4) concerning the possible role of co-pollutants which also may have
14 varied between the time of the studies and the current assessment period. The
15 approach taken to characterize uncertainties in the concentration-response functions
16 arising from sample size considerations is discussed below. With respect to the shape
17 of the function and whether or not a population threshold may exist, as discussed in
18 Chapter 3, the draft CD concludes (section 8.4.3.4, p.8-56) that "the limited evidence
19 suggests that if there is a threshold level in Oj health effects, it is likely near the lower
20 limit of ambient Oj concentrations in the United States." The draft CD also concludes
21 (section 8.4.3.2.3, p.8-54) that the U.S. time series studies conducted in multiple cities
22 "provide substantial epidemiological evidence indicating that associations for Os with
23 mortality and morbidity are robust to confounding by copollutants."
24 • Adequacy of ambient (^monitors as surrogate for population exposure. The extent to
25 which there are differences in the relationship between spatial variation in ambient Os
26 concentrations and ambient exposures in the original epidemiology studies compared
27 to more recent ambient Os data introduces additional uncertainty in the risk estimates.
28 The draft CD (section 8.4.3.2.1, p.8-53) states that "ambient concentrations generally
29 overestimate true personal Oa exposures" and that the "use of ambient concentrations
30 in risk calculations will likely result in effect estimates that are biased towards the null,
31 resulting in biased descriptions of underlying concentration-response relationships."
32 Thus, the risk estimates presented here may underestimate the overall impact of Os
33 exposures on mortality and hospital admissions.
34 • Adjustment of air quality distributions to simulate just meeting the current standard.
35 The shape of the distribution of hourly Os concentrations that would result upon just
36 meeting the current or alternative 8-hr standards is unknown. Based on an analysis of
37 historical data, staff believes mat the Quadratic air quality adjustment procedure
38 . provides reasonable estimates of the shape of the distribution; however, there is greater
39 uncertainty for those urban areas that have air quality well above the current standard
40 (e.g., Los Angeles, Houston). Staff plans to include additional sensitivity analyses
41 exploring the potential impact of using alternative air quality adjustment procedures on
42 the risk estimates associated with just meeting the current.and alternative Oa standards
43 in the next draft of the Risk Assessment TSD. Staff also plans to develop risk
44 estimates associated with just meeting the current standard based on a different year
November 2005 5-28 Draft - Do Not Quote or Cite
-------
1 within the three-year period on which the design value is based (e.g., 2002, in which Os
2 levels were generally higher in most of the locations).
3 • Estimated background concentrations for each location. The calculation of risk
4 associated with recent air quality in excess of background requires as an input
5 estimates of background concentrations for each location throughout the period of the
6 assessment. The estimated background concentrations have been obtained from runs
7 of the GEOS-CHEM global model (see section 2.7) and introduce some uncertainty
8 into the risk estimates for both the recent air quality scenario and the just meeting the
9 current 8-hr standard, both of which are calculated as risk in excess of background.
10 • Baseline incidence rates and population data. There are uncertainties related to: (1) the
11 extent to which baseline incidence rates, age distributions, and other relevant
12 demographic variables that impact the risk estimates vary for the year(s) when the
13 actual epidemiological studies were conducted, the recent year of air quality used in
14 mis assessment, and some unspecified future year when air quality is adjusted to
15 simulate just meeting the current or alternative standards and (2) the use of annual or
16 seasonal incidence rate data to develop daily health effects incidence data. Spatial
17 variability in baseline incidence and population data is taken into account by use of
18 city-specific data in most cases.
19 One of the most critical elements in the risk assessment is the concentration-response
20 relationships used in the assessment. The uncertainty resulting from the statistical uncertainty
21 associated with the estimate of the 63 coefficient in the concentration-response function was
22 characterized either by confidence intervals or by Bayesian credible intervals around the
23 corresponding point estimates of risk. Confidence and credible intervals express the range
24 within which the true risk is likely to fall if the only uncertainty surrounding the Os coefficient
25 involved sample size considerations. Other uncertainties, such as differences in study location,
26 time period, and model uncertainties are not represented by the confidence or credible intervals
27 presented.
28 The two multi-city mortality studies, Bell et al. (2004) and Huang et al. (2004), reported
29 both multi-location and single-location concentration-response functions, using a Bayesian two-
30 stage hierarchical model. In these cases, the single-location estimates can be adjusted to make
31 more efficient use of the data from all locations. The resulting "shrinkage" estimates are so
32 called because they "shrink" the location-specific estimates towards the overall mean estimate
33 (the mean of the posteri or distribution of the multi-location concentration-response function
34 coefficient). The greater the uncertainty about the estimate of the location-specific coefficient
35 relative to the estimate of between-study heterogeneity, the more the location-specific estimate is
36 "pulled in" towards the overall mean estimate. Bell et al. (2004) calculated these shrinkage
37 estimates, which were presented in Figure 2 of that paper. These location-specific shrinkage
38 estimates, and their adjusted standard errors were provided by the study authors and were used in
39 the risk assessment.
November 2005
5-29
Draft - Do Not Quote or Cite
-------
1 The location-specific estimates reported in Table 1 of Huang et al. (2004) are not
2 "shrinkage" estimates. However, the study authors provided the posterior distribution for the
3 heterogeneity parameter, T, for their distributed lag model, shown in Figure 4(b) of their paper.
4 Given this posterior distribution, and the original location-specific estimates presented in Table 1
5 of their paper, we calculated location-specific "shrinkage" estimates using a Bayesian method
6 described in DuMoucheL (1994) (see Appendix B of the draft Risk Assessment TSD for details
7 about the calculation). As with the shrinkage estimates presented in Bell et al. (2004), the
8 resulting Bayesian shrinkage estimates use the data from all of the locations considered in the
9 study more efficiently than do the original location-specific estimates. The calculation of these
10 shrinkage estimates is thus one way to address the relatively large uncertainty surrounding
11 estimates of coefficients in location-specific concentration-response functions.
12 With respect to model form, most of the epidemiological studies estimated Oa coefficients
13 using log-linear models. However, there still is substantial uncertainty about the correct
14 functional form of the relationship between 63 and various health endpoints, especially at the low
15 end of the range of observed' concentrations. While there are likely biological thresholds in
16 individuals for specific health responses, the available epidemiological studies do not support or
17 refute the existence of thresholds as Os levels approach background concentrations.
18 Several meta-analy ses addressing the impact of various factors on estimates of mortality
19 associated with short-term exposures to Os were recently published and are discussed in the draft
20 CD. Staff plans to review these analyses and explore whether they provide additional
21 information that can be used to assist in characterizing the uncertainties associated with risk
22 estimates for this health outcome.
23 5.4 OZONE RISK ESTIMATES
24 The risk estimates associated with two air quality scenarios, a recent year of air quality as
25 represented by 2004 monitoring data and air quality adjusted to simulate just meeting the current
26 8-hr standard are presented in this draft Staff Paper in the sections below. The next draft Staff •
27 Paper will include risk estimates for alternative 8-hr standards.
28 5.4.1 Recent Air Quality
29 In the prior risk assessment, risks for lung function decrements associated with 1-hr
30 heavy exertion, 1-hr moderate exertion, and 8-hr moderate exertion exposures were estimated.
31 Since the 8-hr moderate exertion exposure scenario for children clearly resulted in the greatest
32 health risks in terms of lung function decrements, staff has chosen to include only the 8-hr
33 moderate exertion exposures in the current risk assessment for this health endpoint. Thus, the
34 risk estimates presented here are most useful for making relative comparisons across alternative
November 2005 5-30 Draft - Do Not Quote or Cite
-------
1 air quality scenarios and do not represent the total risks for lung function decrements in children
2 or other groups within the general population associated with any of the air quality scenarios.
3 Tables 5-5 and 5-6 display the risk estimates for all and "active" school age children
4 (ages 5-18) associated with'2004 63 concentrations for two different levels of lung function
5 decrement responses for the 12 urban areas. These two tables also include risk estimates
6 associated.with air quality adjusted to simulate just meeting the current 0.08 ppm, 8-hr standard,
7 which will be discussed further in section 5.4.2. All estimates in both tables reflect responses
8 associated with exposure to 63 in excess of exposures associated with background Os
9 concentrations. Table 5^5 shows the number and percent of children estimated to have at least 1
10 lung function response during the Os season. Table 5-6 displays the total number of occurrences
11 for the specified lung function responses during the Oj season. As illustrated by the estimates
12 shown in these two tables, a child may experience multiple occurrences of a lung function
13 response during the Os season. For example, in Atlanta the median estimate is that 96,000 school
14 age children experienced an FEVi decrement > 15% during the Os season with a median estimate
15 of 738,000 occurrences of this same response in this population. Thus, for this example on
16 average each child is estimated to have over 7 occurrences of this lung function response during
17 the Os season.
18 As shown in Table 5-5, across the 12 urban areas, the ranges in median estimates of the
19 percent of all and "active".school age children estimated to experience at least one FEVi,
20 decrement > 15% during the OB season are 6.8-12.7 and 7.2-13.5%, respectively. The ranges in
21 median estimates of the percent of all and "active" school age children estimated to experience at
22 least one FEVi decrement > 20% during the Oa season across these same 12 urban areas are 0.8-
23 4.6% and 0.9-5.1 %, respectively.
24 In terms of total occurrences of FEVi decrement > 15% during the Os season, Table 5-6
25 shows a range of median estimates from 339,000 to over 6.1 million responses during the 63
26 season for all school age children and from 202,000 to 3.9 million responses for "active" school
27 age children across the 12 urban areas associated with 2004 Os concentrations. For FEVi
28 decrement > 20% during the Os season, Table 5-6 shows a range of median estimates from
29 13,000 to 670,000 for all school age children and from 9,000 to 440,000 for "active" school age
30 children.
November 2005
5-31
Draft - Do Not Quote or Cite
-------
I
t
5
U
*.
3 e
t
11
si
•I
IS
•* i
S &•
[2 O£
%
u
|U
• ti
\S
!i
8*
ctive Chlldr
Associate
pon
*
Ex
I
EX
5-1 8) Ha
8-Hour 03
« >
II
^ *T*
6 «
= <
ent
s.'
3!
ol
a.
cf
S
-a
|
a
§
y; IN
K i"
0) ,
G
"
to
o *-
8*
87
IO
J
M
S
"8
CO
R
SBT
ci
S"
a>
K »-
a> •
•8
S3.
s?
g
"
S
*
U
|
U
sa
R
S
5
u
ge
CM
S
S3.
'S
-8
o
8
±!
1
o
"
(0,
ge
ci
~
co m
CM 1
(O
I
#•»'
CN
in >n
o
CO
a 7
ci
sa
®
8
§
^
S?
01
?
.*
in
S?7
-------
« s
« —
OL C
Lea
ire
II
61
ponse
km**
t 1 Lung
Under M
«s.
•* •"
ll
5"
h
3!
c i
«
I
to
Rece
S3
8
ST
s
ff
I
ic
Q.
si
tN
in '
2n
0} i
UJ
s
01
si
a
a.
CM
IT) >
m
M
1
58T
«
G
e
i
e
2?
O
Q
I
a>
c
i
.1
£
a
I
re
in.
to
3 t!
«
G
I
a
*
%
I
-------
Q
|
8
^
g?
CM
0
n*
1?
8? °
to r~
T— 1
8?
i
s
^
S
CM ,
;£
CM
d
in
ST
67
o •*
a
w
c
^
S '
»
"
**:
s
o—
§
1
CM Is-
CO CO
if
10 «
CM
CO CO
52
£
«
C
CM
dg?
CO
d
s
3.
co •
CM"
a.
CM
CM
IT
s
S
o
I
dgs
CM
8T
C!
52
CO?
" co
o
O
i5
I
I
a
I
-------
j
Per
•5 «
3s
to
t
SB
31
1
a
<
a>
i
«
M
d
o •
,c
a
?- o
.S
4
£
•?
05?
*t
d
C?'
8®
o >
^
G
(O i
"§
sRci
o •
fff
W
5
55
. c>
03
g? rvi
CM '
CJ
O 0>
N
a
as d
O i
£
?- °
°S
p.
ss
S
?
t;
c
o
*
I
I
o .c
O •£
# S
K
-------
1 The risk estimates associated with 2004 Os concentrations for the health endpoints based
2 on epidemiological studies are shown in Figures 5-3 and 5-4 and Table 5-7 for non-accidental
3 mortality and Figures 5-5 and 5-6 and Table 5-8 for cardiovascular and respiratory mortality in
4 all of Ihe urban locations included in the assessment. Table 5-9 presents risk estimates for excess
5 hospital admissions for total respiratory illness and asthma (which is a subset of total respiratory
6 illness admissions) for the New York City urban area. Additional hospital admission estimates
7 for three other locations are provided in the draft Risk Assessment TSD. All results are for
8 health risks associated with short-term exposures to Os concentrations in excess of background
9 levels from April through September 2004. The percent of total incidence that is Os-related is
10 shown in the top portion of Figures 5-3 through 5-7; the incidence per 100,000 relevant
11 population is shown in the bottom portion of Figures 5-3 through 5-7.
12 The central tendency estimates in all of the figures and in Tables 5-7 through 5-9 are
13 based on the Os coefficients estimated in the studies, or, in the case of the location-specific
14 estimates from Huang et al. (2004) and Bell et al. (2004), on "shrinkage" estimates based on the
15 Os coefficients estimated in the study (see section 4.1.9.1.2 of the draft Risk Assessment TSD).
16 The ranges are based either on the 95 percent confidence intervals (CIs) around those estimates
17 (if the coefficients were estimated using classical statistical techniques) or on the 95 percent
18 credible intervals (if the coefficients were estimated using Bayesian statistical techniques).
19 As discussed previously, assessment locations were chosen in part on the basis of
20 whether an acceptable concentration-response function had been reported for that location. As a
21 result, risks were estimated in a given assessment location only for those health endpoints for
22 which there is at least one acceptable concentration-response function reported for that location.
23 For non-accidental mortality associated with short-term exposure to Os, Figure 5-3 displays
24 estimates for only nine of the twelve risk assessment locations because acceptable (single-city)
25 concentration-response functions for this health outcome were not available for the other three
26 locations. Figure 5-3 shows estimated percent of non-accidental mortality and cases per 100,000
27 relevant population related to recent O3 concentrations over background levels, based on single-
28 pollutant, single-city models across all locations for which such models were available.
29 Table 5-7 shows estimates of incidence, incidence per 100,000 relevant population, and
30 percent of total incidence of non-accidental mortality related to recent Os concentrations over
31 background levels in all locations, based on both single-city and multi-city models. Estimates of
32 Os-related (non-accidental) mortality ranged from 0.4 per 100,000 relevant population in Atlanta
33 (Bell et al., 2004) to 7.3 per 100,000 relevant population in Chicago (Schwartz, 2004).
34 Estimated Os-related (non-accidental) mortality reported by Schwartz (2004) for Chicago,
35 Detroit, and Houston, based on both the single-city and the multi-city concentration-response
36 functions, tend to be higher than other estimates in those locations in large part because Schwartz
November 2005 5-36 Draft - Do Not Quote or Cite
-------
1
2
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20,
21
22
23
24
26
27
28
29
Figure 5-3. Estimated (Non-Accidental) Mortality Associated with Short-Term Exposure to
Ozone Above Background: Single-Pollutant, Single-City Models (April - September, 2004)
5-3a. Estimated Percent of Total Incidence that is O3-Related
3.0%
2.5%
1.5%
1.0%
0.0%
-0.5%
-1.0% J-
H 1-
_f__ ,_
5-3b. Estimated O3-Related Cases per 100,000 Relevant Population
g 14.0
•g 12-°
3, 10.0
o 8.0
| 6.0
£ 4'°
S 2'°
1 0-0
H -2.0
g -4.0
November 2005
5-37
- Do Not Quote or Cite
-------
1 Figure 5-4. Estimated (Non-Accidental) Mortality Associated with Short-Term Exposure to
2 Ozone Above Background (April - September, 2004): Single-City Model (left
3 bar) vs. Multi-City Model (right bar)
5-4a. Estimated Percent of Total Incidence that is O3-Related
5
6
. 7
8
9
10
11
12
13
14
15
5" 35%
» 30%
"c 25% •
W
8"~~ 20% .
C 1 0% .
o 05% -
"5 00% -
5 -05% -
J* 1 ran. .
1
, J,*,.,
f
i
i
Atlanta: Bel et al.. 2004
Chicago: Schwartz, 2004
I «r e
: 1 1
j Ji
Houston: Bell etal., 2004
hHouston: Schwartz. 2004
^
'in
, r
Sacramento: Bell et al., 2004
(0
-
•§
3
fO
Figure 5-4b. Estimated O3-Related Cases per 100,000 Relevant Population
D 16.0 T
14.0
12.0
10.0
8.0
6.0
4.0
2.0
0.0
-2.0
I ' I 1 1 1
-l-1-! 1-
TB -a
•« v
~S
•s
November 2005
5-38
Dra/f - £)o Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Figure 5-5. Estimated Cardiovascular and Respiratory Mortality Associated with Short*
Term Exposure to Ozone Above Background (April - September, 2004):
Single-City Model (left bar) vs. Multi-City Model (right bar) - Based on Huang
etal. (2004)
5-Sa. Estimated Percent of Total Incidence that is Oa-Related
3 0%
5
§> 2 0% •
1,«
8
fl Q% -
i.u-w
fo5% .
£
£ 0 0% -
fL n «?9A
-1.0% •
A
.
j
i
I
tlanta
Chicago Cleveland Detroit
Houston Los New York Philadelphia
Angeles
Figure 5-5b. Estimated O3-Related Cases per 100.000 Relevant Population
3 Q
o
* 25-
10 f-3
01
•Q
W n n
C
5
15 15-
3 1.3
Q.
£
~ 10.
— I.U -
"" O "? -
K ••'
§" 0 0 -
u.u
si
a -05 .
I
n
0 -1.0 J
A
tlanta
C hicago C 1e ve land De t r oit
.
f
:
|
Houston Los New York Philadelphia
Angeles
November 2005
5-39
Draft - Do Not Quote or Cite
-------
1 Figure 5-6. Estimated Cardiovascular and Respiratory Mortality Associated with Short-
2 Term Exposure to Ozone Above Background (April - September, 2004):
3 Single-Pollutant vs. Multi-Pollutant Models [Huang et al. (2004), additional
4 pollutants, from left to right: none, PM10, NO2, SO2, CO]
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
5-6a. Estimated Percent of Total Incidence that is Oj-Related
3.0%
=• 2.5%
O
£ 2.0%
1 1.5%
° 10%
D.
0.5%
0.0%
-0.5%
„.„...„..,..„,..,. ™w™™.™.™. ~™™™.m..,..
— ~
„ «. ^- ^-^...vv,,-...,™™..
A
tlanta
Cl
llcago C
Figure 5-6b.
4 2°
S 05 - •
s
1 OQ • -
-05 • -
leveland Ot
troit
Hi
>uston
L
""""
1
s
New York Phflad
!
elphla
Estimated O3-ReIated Cases per 100,000 Relevant Population
•
Atlanta
C
hicago
Cl<
tveland De
roit
Ho
,
UStOQ
L.
Ang
»s No
AT York
Philad
elphia
November 2005
5-40
Draft - Do Not Quote or Cite
U.S EPA Headquarters Library
Mail code 3404T
1200 Pennsylvania Avenue NW
Washington, DC 20460
202-566-0556
-------
s
«
«\
I
s
&
a.
w
CO
Q.
•<
O
1
1
e
o
U
v
§
S
•^
I
£
!
•e
I
*u
e
M
3
£>
9
I
g
S
g
u
I
C
O
Z;
£
t^
i/>
jj
A
^2
S,
'o
^
01
1
m «
Q *
551
*
« ^
1 1
1 Mortality Asso
[elevant Backgr
S *
e
0)
Non-Accid
£
3
iff
1
£
5
I
^
£
!
•c
«
5
£ S
"5 S
*• T3
« ~
°"
0
° 5
II
Incidence per 11
Relevant Popul
Incidence
o
i
g
j
!
^
i
1
ob
"S O
C3 er*
CO
9
d co
CO
CO '
?
TO
£
3
0)
J2
T5
"I
.Q
*^
'°-
s
0
CN,
15
•5
"53
CD
5
e
<
ir
CO i
d £
d
o.
CN CN
T— F
S-
d,
ra
.c
oi
CD
jo
^
|
1
TJ
O
W,
8
'*;
O
CO
^
en
i
i
15
0)
"33
CD
'
En
CO '
d ^
d
^~ CO
s.
oT
f1^ i
0,
CO
£
cli
U)
(Q
^~
T3
0)
1
T3
O
0
CN,
I
-4-*
O
CO
in
en
4
1
15
•5
"55
m
e
s
<0
o
ffl
^
CN '
d ^
d
\n
en ,
d,
(0
^
S
0)
(0
~"~
?
1
T3
O
CN,
I
•4-*
O
CO
•D
OS
]
'
"(5
0)
"35
CQ
T~
O> i
^ cl
d
_
CN
CN
K '
CO
CN
S
^- CO
co m
CN
x
E
^
O)
JO
CO
-S
g
o
CM
"~^
•S
i
•§
CO
o
D)
CO
U
IE
u
CN
^ "7
d ^
CN
d
"^
(D
1
T3
O
CN
"5
"53
^
m
X
E
f
£
d
i
C
Sr
OJ '
d ^
d
CN"
CN i
h- CN
d,
co
^
cli
0)
JO
"S
1
TD
O
CN.
S
•t;
O
CO
m
en
i
i
15
"5
"S
m
i
t
i
>
i
i
CO
^ ?
° #
9
9
£"
co i
CO ,_
1
?
CO
.c
S
CO
(Q
^~
JB
T)
O
CN
^~^
"S
•s
"55 •
CO
s
2
a!
a
CO
^ d
CN '
d s£
T—
d
2-
t ?
.d.
co
£
S
en
CO
"8
2
1
•o
3
o
¥
±^
o
CO
CF)
i
i
15
aj
"S
03
cn
?T '
CN
9
CN $2
CO i
f
CO CN
CN i
CN
ci
E
•<-
DJ
jo
CO
7
o
1
CN
* — '
•^
§
•g
CO
1
I
-------
!?
£
1
o
«
O (-
£^j
s
53
_>
w
$ "°
"5 p
§ °>
< *
£2
s c
^ ^
o g
E "3
i -Accident
0
S
3
in
i
X
LLI
«
_
"
CN
^
E
-
D)
J2
Jc
S
D)
JS
istributed
"°
o
Q
Ct
S
O
CO
8
1
"S
^_
"55
CO
c
t
I
o>
T~ ^
T-
CM"
r- "^
i
CM m
d
8?
Si
X
to
-
0)
JO
o
|
Kl
•^
r—
o
CO
I
1
1
1
O) '
d ^
CO
00
00 ™
™ t*~
d
8
oo ••"
CM.
X
co
L_
-
O)
JO
TO
S
o
Q
Ci.
•S
0
to
I
1
*—
o
CO
CJ ^
O IQ
9
oo"
to ^
d CD
^7
CN
i
sj
?
S
o>
istributed
T3
|
• — '
"ro
^_
"53
CD
V
i
1
i
<
i
CO
^ d
in i
d ^
(N
"""'
co"
"^ ^
*~ If)
d
*^ 1
^~ m
ci
CO
s
o
JS
istributed
T3
o
Q
JN
(0
,0)
0
CO
1
"S
75
"53
m
|
h
IJ
i
i
CO
S? d
CM i
d ^?
o_
^™^
T^
K *•
°" N
d
8
o T~
{O l
d ^
- ^~^
in"
tn ^
^ in
d
CN '
• — •
ti>
CO
S
O)
JO
istributed
•a
o
Q
JN.
15
0
CO
13
S
1
"co
^
"33
m
.2
JC
"!
]
X
a
1
sO •«-
3* i
T~ ^
CO
' 2.
*~*
^
TT !^
10 rr
CO
CM *~
00 i
CM
in
d,
CO
S
D)
JO
CO
if
o
*— ^
"ro
1
D>
O
O
5
!
.
•
;
* T
co 'o
o ^**
CD
9
ocT
o -^
^j i
CO
5T
in
CM i
~
d,
CO
s
CD
JO
istributed
13
|
• — •
"TO
^_
"53
CD
3
t>
E
2
1
^
d
•*r
^~
oo
d>
CO
s
CO
JO
istributed
73
S
Q
^j.
1
a
CO
i
1
^
"33
CD
fc
§
I
-------
.1"
5
rith O3 Above 1
svels**
* 3
1 1
tal Mortality Associa
Relevant Backgrou
w
2
8
0
z
f/— *.
^sg
Percent of
Incident
(0.1%-0.
e
o O
Incidence per 100,
Relevant Populati
O4
10
«
1 a
•o i
o to
C ~"^
s«
11
s
OI
_j
v
S
CO
c
o
i
tO
*^ C)
JN ^
O 5s
CO
CO '
d ^
T~
^~ "?
» T
CO
!
s
D)
2
^
•=
«
T3
f
0
I
**
b
CO
13.
vn
O)
I
"ro
"53
13
m
§-
HI
c O
1°
a
t
* £ 2 fe
It! i
I a s -g
™ -D TJ |
<« CO u) O)
g> £ fe £
111 i
t I c |
o °" .a ~
* S n 2
| gtl
w J c z
jj 13 ts ..
? f I z
-------
m
C
^O
g
e
U
|
N
O
OS
6
|
•S
o
VI
<
&
i
b
o
!
»PM
B.
I
•8
13
s
w
i
S
« s
u^
-o u
SI
11
O9 >u
W en
j
u
I
c
n
3*
0)
i
>»
o
i
S!
o
<
n
O
£
I
11
11
o ^j
< "°
II
^j S
£
IB
ll
a.
a.
TJ
S
L.
n
3
O
HI
>
_O
IB
O
i
i
*
*
1
c
i
i
.. j
*
i
i
i
i
t
<
^
i
i
01
o
§
2
"5
c
1
i-
S
+*
c
41
I
^
g c
ll
§.£
§ 1
gs
01
o
41
O
S
I
*
3
3
9
I
3
9
*
i
!
1
C
t
>
E
OJ
O
1
!
1
i
C
CN
5 ^
CO
o
CO
q T]'
o
.H
in
2
I
>
!
5
«-^
W
I ^
d ^
d
i
**""
in
i^ *7
ci T-
9
in ^
CN
1
Q
14
1
OJ
5
•
#
^o W5
"i
• —
q CN
T— 1
0
gg
'
I
(0
o
1
1
«
1
8
^ CN
CO i
CO
d
Si"
CN ^
"*" m
c^
in
in ^
T— '
^,
S
rt:
0
CO
D
O)
8
>
n
>
f
00 I
CO
0
5-
°°. T
o
-------
^
5
0)
£
X
"o
Q. '
1
O
€
isociated v
Levels**
3?
*§
tory Mortal
Backgr
2
Q.
£
oe
T3
IB
"5
<3
in
IB
>
_0
f5
O
01
o
c
01
1 3?
5 ^ T
-£ S ^
* ( 0
^
r 100,000
pulaflon
a£
g C
Inciden
Releva
CD i
Ci CN
O
o> —
u r-
2 •" '
u ^
|
1
O
_J
•a
j3
CO
W
0>
•e:
O
CO
n'
O)
e
«
|
e
e
C
• in
w
V
8
<
Je
u>
£
«?
Jj! CN
T— 1
o
S"
2 '
o 8
OJ i
^
.2
Q.
i
o ^
CO
o
eq
1
N- ^
T— 1
CD
.
O
CO
Z)
CT)
.S3
Q.
"5
13
a
"S
Q.
fc
a
t
= X
E »
i c &
I
I
-------
1 Table 5-9. Estimated Hospital Admissions Associated with Recent Ozone Concentrations
2 in New York, NY* (April - September, 2004)
Health
Effects**
Respiratory
Illness
(unscheduled)
Asthma
(unscheduled)
Ages
All
All
Lag
3-day
1-day
Exposure
Metric
1-hr max
1-hrmax
Health Effects Associated with O3 Above Policy
Relevant Background Levels***
Incidence
(95% CI)
447
(108-786)
382
(81-683)
Incidence per
100,000 '
Relevant
Population
(95% CI)
5.6
(1.4-9.8)
4.8
(1.0-8.5)
Percent of Total
Incidence (95%
CI)
1.3
(0.3-2.2)
2.9
(0.6-5.2)
3
4
5
6
7
8
*New York in this study is defined as the five boroughs of New York City
"""Concentration-response relationships are from Thurston et el. (1992) and are associated with short-term
exposures.
"""""Incidence was quantified down to estimated policy relevant background levels. Incidences per 100,000 relevant
population and percent of total incidence are rounded to nearest tenth.
November 2005
5-46
Draft - Do Not Quote or Cite
-------
1 used the 1 -hr maximum Os concentration, rather than the 24-hr average, as the exposure metric.
2 The changes from recent (2004) 1-hr maximum to background 1-hr maximum Os concentrations
3 were generally larger in the assessment locations than the corresponding changes from recent 24-
4 hr average to background 24-hr average Oj concentrations. As a percent of total incidence,
5 estimated (Vrelated (non-accidental) mortality ranged from 0.1 percent in Atlanta (Bell et al.,
6 2004) to 1.9 percent in Chicago (Schwartz, 2004).
7 Figure 5-4 shows estimated percent of non-accidental mortality that is Os-related and Oa-
8 related cases per 100,000 relevant population, based on single-city versus multi-city models
9 across all locations for which both types of model were available. Estimates of Os-related non-
10 accidental mortality based on single-city models tended to have wider confidence or credible
11 intervals than those based on multi-city models, with both multi-city models (from Bell et al.,
12 2004 and Schwartz, 2004) producing statistically significant results. However, the choice of
13 single-city versus multi-city model did not have a uniform effect on the magnitude of the point
14 estimate. In some cases (Atlanta, Los Angeles, and Sacramento), the multi-city models produced
15 larger estimates than the single-city models, while in other cases (Chicago, Cleveland, Detroit,
16 Houston, and St. Louis) the reverse was true.
17 As shown in Figure 5-5, Bayesian credible intervals around the "shrinkage" estimates
18 (see section 5.3.2.5) of Os-related cardiovascular and respiratory mortality based on single-city
19 models in Huang et al. (2004) were uniformly larger than the corresponding credible intervals
20 around estimates based on the multi-city model from that study. As noted above, all of the
21 estimates were positive and, with the exception of the single-city estimate for Chicago, all were
22 statistically significant.
23 Figure 5-6 shows estimated percent of cardiovascular and respiratory mortality and cases
24 per 100,000 relevant population related to recent Os concentrations over background levels,
25 based on multi-city models for a single-pollutant versus multi-pollutant models from Huang et al.
26 (2004) across all locations for which such models were available. Table 5-8 shows estimates of
27 incidence, incidence per 100,000 relevant population, and percent of total incidence of
28 cardiovascular and respiratory mortality related to recent 63 concentrations over background
29 levels in all risk assessment locations covered in Huang et al. (2004), based on both single-city
30 and multi-city single-pollutant models from that study. Estimates of Oa-related cardiovascular
31 and respiratory mortality ranged from 0.4 per 100,000 relevant population in Chicago (using the
32 single-city concentration-response function) and Houston (using both the single-city and the
33 multi-city concentration-response functions) to 1.3 per 100,000 relevant population in
34 Philadelphia (using the single-city concentration-response function). As a percent of total
35 incidence, estimated Os-related cardiovascular and respiratory mortality ranged from 0.4 percent
November 2005 5-47 Draft - Do Not Quote or Cite
-------
1 in Chicago (using the single-city concentration-response function) to 1.6 percent in Los Angeles
2 (using the multi-city concentration-response function). All of the estimates of Oa-related
3 cardiovascular and respiratory mortality based on Huang et:al. (2004), from both single-city and
4 multi-city models, and from both single-pollutant and multi-pollutant models, were positive.
5 The shrinkage-based single-city single-pollutant models for Atlanta, Chicago, Cleveland,
6 Detroit, and Houston for Os-related cardiovascular and respiratory mortality based on Huang et
7 al. (2004) were not statistically significant. The single city," single pollutant model for the other
8 three locations (Los Angeles, New York, and Philadelphia) and all of the single pollutant, multi-
9 city models were statistically significant for this same health endpoint based on Huang et al.
10 (2004). With respect to the multi-pollutant models for this health endpoint and study, all of the
11 multi-pollutant models were statistically significant, with the exception of the models which
12 included PMio.
13 Table 5-9 shows estimates of unscheduled hospital admissions for respiratory illness in
14 the New York City area associated with Os levels above background of about 450 cases or 5.6
15 cases per 100,000 relevant population, which represents 1.3% of total incidence for recent (2004)
16 Os levels. For asthma-related hospital admissions, the estimates are about 380 cases or 4.8 cases
17 per 100,000 relevant population, which represents about 2.9% of total incidence.
18 5.4.2 Just Meeting Current Ozone Standards
19 As described in Chapter 4 and briefly in section 5.3.2.2, the risk estimates described in
20 this section represent the risks for a single example year based on adjusting the Os levels
21 observed in 2004 to Os levels predicted when just meeting the current 0.08 ppm standard, using
22 the 3-year design value from the 2002-2004 time period This section first discusses the risk
23 estimates for lung function responses, which are based on exposure-response relationships
24 derived from controlled human exposure studies, and then risk estimates are explored for
25 mortality and hospital admissions, which are based on concentration-response relationships
26 obtained from epidemiological studies.
27 The risk estimates for lung function responses are for the.Oa season, which is all year in 3
28 of the study areas (Houston, Los Angeles, and Sacramento) and which is generally 6-7 months
29 long in the other 9 urban study areas (e.g., April to September). The risk estimates for lung
30 function responses in all and "active" school age children (ages 5 to 18) for just meeting the
31 current 8-hr standard for 12 urban areas are summarized in Tables 5-5 and 5-6. As shown in
32 Table 5-5, across the 12 urban areas, the ranges in median estimates of the percent of all and
33 "active" school age children estimated to experience at least one FEVi decrement > 15% during
34 the Os season are 5.0-8.3% and 5.3-9.0%, respectively. The ranges in median estimates of the
35 percent of all and "active" school age children estimated to experience at least one FEVi
November 2005 5-48 Draft - Do Not Quote or Cite
-------
1 decrement > 20% during the 63 season across these same 12 urban areas are 0.2-1.4% and 0.3-
2 1.7%, respectively.
3 In terms of total-occurrences of FEVi decrement > 15% during the 63 season, Table 5-6
4 shows a range of median estimates from 222,000 to over 1.9 million responses during the Os
5 season for all school age children and from 129,000 to nearly 1.3 million responses for "active"
6 school age children across .the 12 urban areas associated with 2004 Os concentrations. For FEVi
7 decrement > 20% during the Os season, Table 5-6 shows a range of median estimates from 5,000
8 to 33,000 for all school age children and from 3,000 to 22-000 for "active" school age children.
9 The results of the assessment of the reduced mortality risks associated with Os
10 concentrations above background that j ust meet the current 8-hr daily maximum standard are
11 summarized across urban areas in Figures 5-7 through 5-10, and in Tables 5-10 and 5-11. In
12 addition, Table 5-12 summarizes unscheduled hospital admission risk estimates for respiratory
13 illness and asthma in New York City associated with Os concentrations above background and
14 just meeting the current 8-hr standard. Additional hospital admission estimates for three other
15 locations are provided in the draft Risk Assessment TSD. All results for the epidemiological
16 based part of the risk assessment are for health risks associated with short-term exposures to Os
17 concentrations in excess of background levels from April through September. The percent of
18 total incidence that is Os-related is shown in top portion of Figures 5-7 through 5-10; the
19 incidence per 100,000 relevant population is shown in the bottom portion of these same figures.
20 As described in the previous section, the central tendency estimates in all of the figures
21 and tables are based on the Os coefficients estimated in the studies, or, in the case of the location-
22 specific estimates from Huang et al. (2004), on "shrinkage" estimates based on the Oj
23 coefficients estimated in the study (see section 5.3.2.5). The ranges are based either on the 95
24 percent confidence intervals around those estimates (if the coefficients were estimated using
25 classical statistical techniques) or on the 95 percent credible intervals (if the coefficients were
26 estimated using Bayesian statistical techniques).
27 The results in this portion of the risk assessment follow the same patterns as the results
28 discussed in section 5.4.2 for risks associated with recent Oj concentrations, because they are
29 largely driven by the same concentration-response function coefficient estimates and confidence
30 or credible intervals.
31 Figure 5-7 shows estimated percent of non-accidental mortality and cases per 100,000
32 relevant population related to Os concentrations that just meet the current 8-hr Os standard, based
33 on single-pollutant, single-city models across all locations for which such models were available.
34 Table 5-10 shows estimates of incidence, incidence per 100,000 relevant population, and percent
35 of total incidence of non-accidental mortality related to 63 concentrations that just meet the
36 current 8-hr Os standard, based on both single-city and multi-city models. Estimates of Os-
November 2005 5-49 Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Figure 5-7. Estimated (Non-Accidental) Mortality Associated with Short-Term Exposure to
O3 Above Background When the Current 8-Hour Standard is Just Met: Single-
Pollutant, Single-City Models (April - September)
5-7a. Estimated Percent of Total Incidence that is Oa-Related
g 2.5%
S 2.0%
7 1.5%
u
g 1.0%
1 0.5%
3 0.0%
O
£ -0.5%
-1.0% -L
, 1 ,
1
!
*
I
I
4
i
Figure 5-7b. Estimated O3-Related Cases per 100,000 Relevant Population
12-0
10.0
4.0
2.0
0.0
-2.0
-4.0
1 > 1
1 \ 1
1 I '
-G
V
1
* tS
November 2005
5-50
Dra/f - Do JVof guote or Cite
-------
1 Figure 5-8. Estimated (Non-Accidental) Mortality Associated with Short-Term Exposure to
2 Ozone Above Background When the Current 8-Hour Standard is Just Met
3 (April - September): Single-City Model (left bar) vs. Multi-City Model (right
4 bar)
5
6
7
9
10
11
12
13
14
15
16
17
18
1Q
17
20
21
22
23
24
25
jg
1
11 5% *
1 0% <
3 05%
J~ ft n% .
8100 -
ty
i 80 •
^ 60-
3
a 40 -
£ 4'°
^ 9 n .
OL
8 -20-
? ./in .
Cases per
,
<
5-Sa. Estimated Percent of Total Incidence that is O3-Related
«|'B's|-S|vV«
Figure S-8b. Estimated O3-Related Cases per 100,000 Relevant Population
[ 1
•i £ «r "5r e «r g or «r «r
•5 | 4J 'S |« | -5 'E -S
*l**l?|S4fi
s a -g. -| « | .H jj | |
3S| oSo||B
•November 2005
5-51
Dra/C - Do //or Quote or Cife
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Figure 5-9. Estimated Cardiovascular and Respiratory Mortality Associated with Short-
Term Exposure to Ozone Above Background When the Current 8-Hour
Standard is Just Met (April - September): Single-City Model (left bar) vs.
Multi-City Model (right bar) - Based on Huang et al. (2004)
5-9a. Estimated Percent of Total Incidence that is O3-Related
2.0%-,
3
•o
e
(V
SI 0% .
e
TJ
'5
~s nsvn .
,2
•5
o 00%
£.
-0.5%
A
tla
nta
' Chicago
Cleveland Detroit
Houston
Los Nsw York Philadelphia
Angeles
Figure 5-9b. Estimated O3-Related Cases per 100,000 Relevant Population
1.4-1
o
vg 1 •> .
£
•o
c in.
s.
c
•2 08-
«
£ ne .
0.0
1 04
J 0.4
o 02-
o 00 •
h.
o
5 -0.4
A
tit
nta Chicago
[
Cleveland Detroit Houston
Los New York
Angeles
Philadelphia
November 2005
5-52
Draft - Do Not Quote or Cite
-------
1 1
2
3
4
5
6
7
8
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
figure 5-10. Estimated Cardiovascular and Respiratory Mortality Associated with Short-
Term Exposure to Ozone Above Background When the Current 8-Hour
Standard is Just Met (April - September): Single-Pollutant vs. Multi-Pollutant
Models [Huang et al. (2004), additional pollutants, from left to right: none,
PM10,NO2,SO2,CO]
5-10a. Estimated Percent of Total Incidence that is O3-Related
1.4% -
1.2%-
O
i
I 1D%-
| Q.8% -
| 0.6% -
•s
1 0.4% -
0.2% -
0.0% •
-0.2% -
f
••
tlanta Chicago Cleveland Detroit Houston L
is New York Philad
etohia :
Angeles
Figure 5-1 Ob. Estimated O3-Related Cases per 100,000 Relevant Population
1.4 -T-
12--
5
a 10-
1
c
O ft 0 _
15
I
•g 06 - •
tQ 4 . .
s
S 02-.
o "••* J
00 -
•0.2 - -
Atlanta (
:hicaao
C
1 veland
Detroit Houston Lc
>s New York Philae
Blphia
Angeles
November- 2005
5-53
Draft - Do Not Quote or Cite
-------
t
jg
f» "°
o i
fffl
v»
o
is
-2 fc
0 3
in a>
< £
_g> —
I H
O S
S S
2 *-"
C *
TI
1 a
< s
§S
2 S
u
g
o
u
i
I
IU
j
«
^
T
0
i
c
2
<
_
mfm
0 „
£ s
0 *
*• w
c TJ
§_E
°"
§ g
o o
§ 1
*- 3
fc §"
a) *«
o e
B 5
01 >
•o «
75 «
C Q.
Ol
u
e
•U
u
_c
o
s.
(U
S
j
»
c£
CO
-§ d
^^ i
f^i ^^
•sr
9
_____
o>
« V
d oo
i
in ?
S
1
§p
CO
|
CN
O)
_____
T3
0)
|
^
-— •
§
§.
75
"S3
i
i
0
^
<
^D"
CO
^ 0
O _N
T-
o_
^_
T-
CO i
d CM
2.
,
u?
O> i
co__
s?
CO
^
CM
S
"g
.n
^
tu
^
b
to
D
in
°\
75
_S 0
CM i
d ..S
o^
T
CO ^
O Jj
d
CO i
G.
i
J
CM
O5
_(0
«
i
10 .
10
£
0
CO
in
en
1
75
1 1
c
P
00
"?
c^
sg CNI
£_ O
d o
&_•
— -
O)
CD °
d CM
d
R?
o
^.
CO
^
CM
D)
JS
"g
.Q
_C
^
U)
X
^ti
O
CO
in
T
75
(U ^"
1 1
c£
U7
8l "
T~ ^
in
d
N.
oo ^
>o oo
r^
5T
o in
GO
21
E
£
""
O)
CO
c?'
0
^_-
g
§.
¥
u
W
1
o
^"
O)
*§ d
CO '
d gS
CM
d
N.
CM "
CM J
d
FT
T— f^_
£2.
X
E
k-.
^
o>
_«
!
CO
_u
^i-3 •
b
CO
Tj-
1
1 1
^
1^
s^ d
cb '
d ^
CM
d
N.
CO
^_ |
*" O)
2,
in
8 CM
i
(0
Tf
CM
O)
™
"g
"3
=6
^
o
§.
"to
QQ
t
I
c
^
c«)
^ CD
CM i
d ^
d
in
en ^
d -L
d
CM CM
^>
^— '
(0
•«f
CM
i
_0)
fi
15
09
§
§-
"ra
"S
m
|
o
—V
^
>? CM
T- °
d o
&»
S
^— .
CO '
d CM
£•
S"
CM CM
^
•*^
CO
^
CM
f
"g
"3
'•&.
d)
S
O
w
D
m
1
75
IB ?
1 1
^
•*
^ csi
?- '
T- e^
CM
9
^^
0)
o
in '
o
i
in CM
O i
T~ r*^
T™
***• '
X
CO
-C
""
D)
JS
CO
s
^
o
§.
s
•g
CO
a
C
fe
I
a
I
i
-------
$
1
ted with O3
ant Os Stand;
•2 E
O a
8 3
< |
£* *•*
i 1
II
e «
41 £
1 g
Non-Ac
Concentratio
S.
31
(A
Q,
^
UJ
c
«
_j
^
^
j
0
c
1
—
5
3 g
Fpercent of
Inciden
§ •=
® .g
§ 1
T- a
g. £
8 c
« i
Ii
~"
Incidence
s»
i
1
.
^
^ CM]
h^
00 "*
W
2-
10 oo
£
E
i_
JC
T™
f
CO
*o
CJ
_
ie
O
CO
^
!
re ~
1: °
CO d
^
0)
0 ^
C^ o^
9
^
M: ^
i
«"
00
*
re
l_
CN
CO
o
S
o
CN,
Q
—
^
n
. — .
CN
to T-:
O i
o^
CN "*
CN '
0,
CO
k~
*+•
"
f
S
*-J
1
"w
T3
i
CN,
15
•5
H>
m
^
!|
10
0 T^-
o
"^^^
. S"
d>
OJ
x:
S
O)
CO
"g
S
.Q
1
T3
cn
OJ
O
CO
£
!
IB
5 c?
"c5 o
CD Ci
1
3
^^
V)
ob i
o. ^
o.
*
03
0,
o ?
g
E
t*m
Jd
T™
S1
(0
S
S
o
d
£
i
0
CO
!
p
i
^^
^
CO '
d g?
CM
0^
0)
o
^*
oo"
S ™
X.
E
k»
x:
T~
S5
5-
S
CO
0)
O
CO
D
!
£
5 §
o 2
CO d
~
in
o ^
fO
1
10
co V
0 00
0
•—•
o sJS
o.
CM
O 0j
o
1^--'
v>
O) ^~
CO i
CO
CM,
d>
ra
i —
CN
O)
T3
(U
•4-f
1
^
T3
CO
V
O
CO
ol
!
15
— 0
"55 2
CO d
i
?
i
i
i
]
e?
S|
00
10 °
° CM
O
*~^
CO "7
d,
re
I'M,
CN
C?
?
+-»
E
B
_M
T3
CO
(U
Jtt
O
CO
S
1
15
o ^
— O
® S
CD Et
1
z
^^
^
CM "
0 ^
O)
v- ^
0,
0?
»- w
T- 1
CO,
d>
S
1^
x:
CM
O>
(0
"S
3
_Q
't_
i!>
-o
CO
CD
O
CO
ol
1
15
"S 5"
"53 2
CO Si
^
oo ^
°' V)
iF
10
10
CM
*™^ *
5?
CO
oo
CO,
g,
re
l_
CM
f
f
5?
O3
o>
^-*
"re
<5
1
OJ
o
0
S
ladelphia
!E
Q.
^
CM ^
O §3
9
Tf
^ "?
O t-
CN
—
00 '
10
?
re
k*.
_^
>T
CM
W
i
-*-j
1
T3
S
O
£i
'ra
5
"53
CD
;ramento
S
CO
1
I
a
I
S
1
-------
t
s
withOa
03 Standa
1 *
•J V
- E
O 3
g 3
10 41
< €
|1
•g S
1 1
3 2
i*
*l
e S
O &
ZS
c
u
I -5-
i2 * !n
§•30 sj?
s£ 5
(^ "•""•"'
S g
Is
II
at **
u e
i >
!S •"•
V
T- 1
CO
s g.
S CM IN
TJ ^~ i
o ^r
JS ~"
1 _a
10 t
O 41
UJ
D)
J
•o
2
co
D)
(0
u.
CM
TO
1
£
§
T3
Q)
•t±
O
CO
13
m
T
"S
Q) ^
"5 S
m £i
I
5
3
^
Si
* t
O ^*
CN
•n**-
J5
O (\|
i
,_,
rt '
T
ra
(0
JE
S
0>
IS
1
•c
]S
^
^
Q
§.
"S
•8
•jj
CD
^
vP N
&• 0
^
CD ^
O
__^
CM '
t™
o,
(0
IM.
S
S
1
^
•4^
^
Q)
'*i
b
CO
D
m
T
To
"S g"
T5 S
m CL
.2
—i
w
^
w o
^
°
csi
(
o_
^-.
^ 1
CM
g,
ra
-c
s
f
1
•c
•s
ID
(0
0)
iS
0
CO
D
in
*?
"S
•53 g"
1C S
m CL
c
0)
Z D
s
s
D
at
(A
8
I
•5
.3
_£
O
I
*•
"S =
3 1
s f
a, §
<0 D
^
1
T3
C
0 IB
S S
_0) ^
£ "S
III
"g £ "E
« S
m
II
TJ
C
c »
= = » i 1
I I -s i
-
a
IT)
^2 -
01
^1
0)
'o
r e 2 8
I
-------
o
1
V
u
g
U
S
•B
«^
11
ce C
03 S
|l
1=2
So.
<
b
s-g
•«•> 2
B-S
f*
^^
if
II
3£>
»1
'•S ^
II
S BO
« -JS
S S
« s-
H =
W U
_
tfl
a
ts
5
(ft
i
1
§
g
o
u
ft
0 ,
ft
5
il
lity Associa
urrentOaS
s S
3 E
S
'5.
!0
at
•o
2
3
CA
IB
_O
O
e
\
i
3
•c
ff
1
*
*
c
<
1
V
I
u
«
<
A
u
a
u
u
§
2
O
c
1
t—
•5
*ri
C
£.
fl»
»
o c
nee per 100
rant Popula
11
«
o
e
0)
•D
'o
C
1
:
i
i
*
i
!
i
i
i
1
i
i
i
i
i
i
C
e.
i
;
•^
^
g *f
9
a>
d r
9
?
(O i
CM
•"•n-""
Atlanta
1
«
<
^
£ 4
° &
s|
o"
CM,
O
CO
Z)
0)
"o*
ob
•sP O
G) 5^
CO
9
00
0 CN
C3
^^^
•a- i
T~ CO
T™
"
O
I
6
c
(
if
c.
<-»
^
£ °
d ^
FT
o.
co
;
1
?
i
ir?
CM 0
O i
o.
(D
T_
GO i
T"
Houston
— .
to
<=> 2?
0,
.
d
,
=0 r
CO,
«
O
CO
D
O)
I
CO
1
S
a
t>
W"l
•Q
1
-------
IS
g
4*
IS
£
entraflons
U
§
u
n
o ,,
£ i
Mortality Associated wl
the Current O3 Standar
|1
2
£
•a
C
IB
"5
u
^
a
U
8
B
0
2
o
c
it of Total
^
Ol
o
o!
icldence per 100,000
Relevant Population
u
B
W
-a
"u
c
o
i
X
1
^
0 i
O
§
m
CO t
CO
0^
H— N
to >
0 CN
g?
£1
s
o
£0
13
01
<0
1
I
^
1 g
C3 tN
« S
U5 ^
C.
^
^
1
^
^? d
CM
0^
° 2
co"
O> |
C
(U
±i
0
Cfl
D
o>
^
o
5T
IT) «N
T- >
H.
(0
!c
Q.
0)
•o
_ra
!S
Q.
&^
r*~ '
0 ^
co
o_
to"
o"
CM ™
T— I
S-
t/>
0>
b
(0
D
o>
re
1 I
i i
jf ™
^— I»
1
oo
£ »
1 1
Jr W 0- 0 .£
•1 s I E
-
Si
-------
1 Table 5-12. Estimated Hospital Admissions Associated with Ozone Above Background In
2 New York, NY* When the Current 8-Hour Standard is Just Met (April -
3 September)
Health
Effects**
Respiratory
Illness
(unscheduled)
Asthma
(unscheduled)
Ages
All
All
Lag
3-day
1-day
Exposure
Metric
1-hr max
1-hrmax
Health Effects Associated with 63 Above Policy
Relevant Background Levels***
Incidence
(95% CI)
364
(88-639)
310
(66-555)
Incidence per
100,000
Relevant
Population
(95% CI)
4.5
(1.1-8.0)
3.9
(0.8-6.9)
Percent of Total
Incidence (95%
CI)
1.0
(0.2-1.8)
2.4
(0.5-4.2)
4
5
6
7
8
9
10
11
*New York in this study is defined as the five boroughs of New York City
*'Concentration-response relationships are from Thurston et al. (1992) and are associated with short-term
exposures.
***Incidence was quantified down to estimated policy relevant background levels. Incidences per 100,000 relevant
population and percent of total incidence are rounded to nearest tenth.
November 2005
5-59
Draft - Do Not Quote or Cite
-------
1 related (non-accidental) mortality ranged from 0.3 per 100,000 relevant population in Atlanta
2 (Bell et at, 2004), Houston (Bell et al., 2004 - 95 U.S. Cities), and Los Angeles (Bell et al.,
3 2004) to 5.8 per 100,000 relevant population in Chicago (Schwartz, 2004).
4 As was the case for the analysis of effects associated with recent Os concentrations,
5 estimated Os-related (non-accidental) mortality reported by Schwartz (2004) for Chicago,
6 Detroit, and Houston, based on both the single-city and the multi-city concentration-response
7 functions, tend to be higher than other estimates in those locations in large part because Schwartz
8 used the 1 -hr maximum Os concentration, rather than the 24-hr average, as the exposure metric.
9 The changes from 1 -hr maximum Os concentrations that just meet the current 8-hr Os standard to
10 background 1-hr maximum Os concentrations were generally larger in the assessment locations
11 than tiie corresponding changes using the 24-hr average metric.
12 As a percent of total incidence, estimated non-accidental mortality related to Os
13 concentrations that just meet the current 8-hr Os standard ranged from 0.1 percent in several
14 locations (Atlanta, Chicago, Detroit, Houston, Los Angeles, New York, and St. Louis) to 1.5
15 percent in Chicago (Schwartz, 2004). Although 7 of the 12 estimates from single-city single-
16 pollutant models shown in Figure 5-7 were not statistically significant, all 12 were positive. In
17 addition, it should be noted that the multi-city model estimates for non-accidental mortality were
18 statistically significant based on Bell et al. (2004) and S chwartz (2004).
19 Figure 5-8 shows estimated percent of non-accidental mortality and cases per 100,000
20 relevant population related to Os concentrations that just meet the current 8-hr Os standard, based
21 on single-city versus multi-city models across all locations for which both types of model were
22 available. The results followed the same patterns as were observed in the analysis of effects
23 associated with recent Os concentrations above background levels, discussed in section 4.2.1
24 above (see also Figures 4-4a and b). Similarly, the results seen in Figure 5-9, for cardiovascular
25 and respiratory mortality, followed the same patterns as are evident in the corresponding analysis
26 of recent Os concentrations (see Figures 5-4 and 5-5).
27 Figure 5-10 shows estimated percent of cardiovascular and respiratory mortality and
28 cases per 100,000 relevant population related to Os concentrations that just meet the current 8-hr
29 Os standard, based on multi-city models which include a single-pollutant versus multi-city
30 models with multiple pollutants from Huang et al. (2004) across all locations for which such
31 models were available. Table 5-11 shows estimates of incidence, incidence per 100,000 relevant
32 population, and percent of total incidence of cardiovascular and respiratory mortality related to
33 Os concentrations above background that just meet the current 8-hr standard in all risk
34 assessment locations covered in Huang et al. (2004), based on both single-city and multi-city
35 models from that study. Estimates of Os-related cardiovascular and respiratory mortality ranged
36 from 0.2 per 100,000 relevant population in Houston (using both the single-city and the multi-
November 2005 5-60 Draft - Do Not Quote or Cite
-------
1 city concentration-response functions) to 1.0 per 100,000 relevant population in Philadelphia
2 (using the single-city concentration-response function). As a percent of total incidence,
3 estimated Os-related cardiovascular and respiratory mortality ranged from 0.3 percent in Chicago
4 (using the single-city function) to 0.8 percent in Los Angeles (using the multi-city function) and
5 Philadelphia (using the single-city function).
6 The staff notes that all of the estimates of Os-related cardiovascular and respiratory
7 mortality based on Huang et al. (2004), from both single-pollutant and multi-pollutant models
8 (see Figure 5-10) and from both single-city and multi-city models (see Figure 5-9and Table 5-11)
9 were positive.
10 Table 5-12 shows estimates of unscheduled hospital admissions for both respiratory
11 illness and asthma in New York City associated with Os levels above background for the period
12 from April to September with air quality adjusted to just meet the current 8-hr standard. For
13 total respiratory illness, Table 5-12 shows about 450 cases or 5.6 cases per 100,000 relevant
14 population, which represents 1.3% of total incidence for recent (2004) Os levels. For asthma-
15 related hospital admissions, which are a subset of total respiratory illness admissions, the
16 estimates are about 380 cases or 4.8 cases per 100,000 relevant population, which represents
17 about 2.9% of total incidence. The reduction in incidence for both total respiratory illness and.
18 asthma admissions from the recent Os levels to air quality just meeting the current 8-hr standard
19 is about 19%. Staff also notes that the estimates for this health endpoint for New York City are
20 higher than the estimates in the risk assessment conducted during the prior Os NAAQS review
21 which used the same concentration-response function. The main reason for this is the use of, a
22 single value of 0.04 ppm for background which is higher than the current modeled values for
23 background in the current assessment and, thus Os levels below 0.04 ppm are contributing
24 additional estimated cases in the current assessment
25
26 5.4.3 Just Meeting Alternative Ozone Standards [to be included in next draft of
27 Staff Paper]
28
29 5.4.4 Key Observations [to be included in next draft of Staff Paper]
30 • .
31
November 2005 5-61 Draft - Do Not Quote or Cite
-------
1 REFERENCES
2
3 Abt Associates Inc. (2005a). Ozone Health Risk Assessment for Selected Urban Areas: Draft Report. Prepared for
4 Office of Air Quality Planning and Standards, U. S. Environmental Protection Agency, Research Triangle Park, NC.
5 October 2005. Available electronically on the internet at:
6 hUp;//www.epa.gov/ttn/naaqs/standards.''ozone/so3_cr_td.himl.
7
8 Abt Associates Inc. (2005b). Particulate Matter Health Risk Assessment for Selected Urban Areas. Prepared for
9 Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency, Research Triangle Park, NC.
10 June 2005. Available electronically on.the internet at:
11 httB://\vwvv.epa.eov/tt.n/naaqs/staadards/pm/s nm cr td.htrol.
12
13 Bell, M,A. McDermott, S.L. Zeger, J.M. Samel, and F. Dominici (2004). "Ozone and short-term mortality in 95 US
14 urban communities, 1987-2000." JAMA 292(19): 2372-2378.
15
16 CDC, 2005. United States Department of Health and Human Services (US DHHS), Centers for Disease Control and
17 Prevention (CDC), National Center for Health Statistics (NCH S), Compressed Mortality File (CMF) compiled from
18 CMF 1968-1988, Series 20, No. 2A 2000, CMF 1989-1998, Series 20, No. 2E 2003 and CMF 1999-2002, Series 20,
19 No. 2H 2004 on CDC WONDER On-line Database. See hltt)://wonder.cdc.gov/.
20
21 DuMouchel, W. (1994). "Hierarchical Bayes Linear Models for Meta-Analysis." Technical Report #27.
22 September, 1994. National Institute of Statistical Sciences, P. O. Box 14162, Research Triangle Park, N.C. 27709.
23
24 EPA (1996a). Review of National Ambient Air Quality Standards for Ozone: Assessment of Scientific and
25 Technical Information - OAQPS Staff Paper. EPA/452/R-96-007. Office of Air Quality Planning and Standards,
26 Research Triangle Park, NC. Available electronically on the internet at:
27 bth):/Avww.epii.gr sp.httnl.
28
29 EPA(1996b). Air Quality. Criteria for Ozone and Related Photochemical Oxidants. EPA/600/P-93/004aF-cF.
30 Office of Research and Development, National Center for Environmental Assessment, Research Triangle Park, NC.
31 Available electronically on the internet at: http://cfpub.epa.gov/ncea/ofm/recordisclay.cfm7deid~2831.
32 . -
33 EPA (2002). Consolidated Human Activities Database Users Guide. The database and documentation are available
34 electronically on the internet at: bttp://www.epa.gov/chadnet 1 /.
35
36 EPA (2003). Total Risk Integrated Methodology TRIM.Expo/Inhalation User's Document
37 Volume I: Air Pollutants Exposure Model (APEX, version 3) User's Guide. Office of Air Quality Planning and
3 8 Standards, Research Triangle Park, NC. Available electronically on the internet at:
39 http://\vw'w.epa.gov/ttn/'fera/human_apex.html.
40
41 EPA (2004). Air Quality Criteria for Particulate Matter. EPA 600/P-99/002bF, 2v. National Center for
42 Environmental Assessment, Research Triangle Park, NC. Available electronically on the internet at:
43 htlrj://vyww.epa.eoy/ttn/nafigs/standards/Dm/s_pm cr_cd.htal
44
45 EPA (2005a). Health Assessment Plan for Review of the National Ambient Air Quality Standards for Ozone.
46 Office of Air Quality Planning and Standards, Research Triangle Park, NC. April. Available electronically on the
47 internet at l^.:/h^w.es&.s^'i^f^ftq^st:^M^^£B^s^3L^j^.^ai.
48
49 EPA (2005b). Review of National Ambient Air Quality Standards for Ozone: Assessment of Scientific and
50 Technical Information - OAQPS Staff Paper (first draft). EPA-452/D-05-001. Office of Air Quality Planning and
51 Standards, Research Triangle Park, NC. Available electronically on the internet at:
52 http://www.epa.eov/ttn/naaqs/standards/dzoiie/'s o3 cr sp.html.
53
November 2005 5-62 Draft - Do Not Quote or Cite
-------
1 EPA (2005c). Air Quality Criteria for Ozone and Other Related Photochemical Oxidants. Second External Review
2 Draft. National Center for Environmental Assessment, Research Triangle Park, NC. Available electronically on the
3 internet at: http://cfpub.et)a.govAn:ea/cim/recordispiay.cfin?deid-1373()7
4
5 EPA(2005d). Ozone Population Exposure Analysis for Selected Urban Areas (draft). Office of Air Quality
6 Planning and Standards, Research Triangle Park, NC. October. Available electronically on the internet at:
7 bttp7Mww.epa.gov/ittiyhaaqs/siandards/oaon.e/s__o3_CTjd.hfanl.
8
9 EPA (2005e).' Review of National Ambient Air Quality Standards for Particulate Matter: Assessment of Scientific
10 and Technical Information-OAQPS Staff Paper. EPA-452/D-05-001. Office of Air Quality Planning and
11 Standards, Research Triangle Park, NC. Available electronically on the internet at:
12 htte://\vmv.eBa.£oyAt^
13
14 Fiore, AM, D.J. Jacob, I. Bey, R.M. Yantosca, B.D. Field, AC. Fusco, and J.G. Wilkinson (2002a). "Background
15 ozone over the United States in summer: Origin, trend, and contribution to pollution episodes." J. Geophys. Res.,
16 107(D15),4275.
17
18 Fiore, A.M., D.J. Jacob, B.D. Field, D.G. Streets, S.D. Fernandes, and C. Jang (2002b). "Linking ozone pollution
19 with climate change: The case for controlling methane." Geophys. Res. Lett., 29(19), 1919.
20
21 Fiore, AM., D.J. Jacob, H. Liu, R.M. Yantosca, T.D. Fairlie, and Q. Li (2003). "Variability in surface ozone
22 background over the United States: Implications.for air quality policy." Journal Of Geophysical Research Vol.
23 108(D24), 4787.
24
25 Friedman, M.S., K.E. Powell, L. Hutwagner, L.M. Graham, and W.G. Teague (2001). "Impact of changes in
26 transportation and commuting behaviors during the 1996 summer Olympic games in Atlanta on air quality and
27 childhood asthma." JAMA 285:897-905.
28
29 Fusco, A. C., and J. A. Logan (2003). "Analysis of 1970- 1995 trends in tropospheric ozone at Northern
30 Hemisphere midlatitudes with the GEOSCHEM model." J. Geophys. Res., 108(D15), 4449.
31 '
32 Gent, J.F., E.W. Triche, T.R. Holford, K. Belanger, M.B. Bracken, W. S. Beckett, B.P. Leaderer (2003).
33 "Association of low-level ozone and fine particles with respiratory symptoms in children with asthma." JAMA
34 290(14):1859-1867.
35
36 Graham, S. and T. McCurdy (2004). "Developing meaningful cohorts for human exposure models." Journal of
37 Exposure Analysis and Environmental Epidemiology 14:23-43.
38
39 Gilliland, F.D., K. Berhane, E.B. Rappaport, D.C. Thomas, E. Avol, W.J. Gauderman, S.'J. London, H.G. Margolis,
40 R. McConnell, K.T. Islam and J.M. Peters (2001). "The effects of ambient air pollution on school absenteeism due
41 to respiratory illnesses." Epidemiology 12(l):43-54.
42
43 Hazucha, M. J.; Folinsbee, L. J.; Bromberg, P. A. (2003) "Distribution and reproducibility of spirometric response to
44 ozone by gender and age." J. Appl. Physiol. 95: 1917-1925.
45
46 Horstman, D.H. et al. (1990). "Ozone concentration and pulmonary response relationships for 6.6-hr exposures with
47 fivehoursof moderate exercise to 0.08,0.10, and 0.12 ppm." American Review of Respiratory Disease 142:1158-
48 1163.
49
50 Huang, Y., F. Dominici, M.L. Bell (2004). "Bayesian hierarchical distributed lag models for summer ozone
51 exposure and cardio-respiratory mortality." John Hopkins University, Department of Biostatistics Working Paper.
52 46.
53
54 Ito, K. (2003). Associations of particulate matter components with daily mortality and morbidity in Detroit,
55 Michigan. In: "Revised Analyses of Time-Series Studies of Air Pollution and Health," Health Effects Institute
56 Special Report, May.
November 2005 5-63 Draft - Do Not Quote or Cite
-------
1
2 Jaffe, D.H., Singer, M.E., and Rimm, A.A. (2003). "Air pollution and emergency department visits for asthma
3 among Ohio Medicaid recipients, 1991-1996." Environmental Research 91:21-28,
4
5 Johnson, T., Capel, J., and McCoy, M (1996a). Estimation of Ozone Exposures Experienced by Urban Residents
6 Using a Probabilistic Version of MEM and 1990 Population Data. Prepared by International Technology Air Quality
7 Services for Office of Air Quallity Planning and Standards, EPA, Research Triangle Park, NC. Available
8 electronically on the internet at: hltp://;w\yw.epa.Epy/^
9
10 Johnson, T., Capel, J., McCoy, M., and Warnasch, J. (1996b). Estimation of Ozone Exposures Experienced by
11 Outdoor Children in Nine Urban Areas Using a Probabilistic Version of NEM. Prepared by International
12 Technology Air Quality Services for Office of Air Quallity Planning and Standards, EPA, Research Triangle Park,
13 NC. Available electronically on the internet at:
14 http://wvvw.epa.eov/ttn/iiaaqs/sta!idard5/ozone/s o3 pr td.html.
15
16 Johnson, T. (1997). "Sensitivity of Exposure Estimates to Air Quality Adjustment Procedure," Letter to Harvey
17 Richmond, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency, Research
18 Triangle Park, North Carolina.
19
20 Kann, L. et al. (2000). "Youth risk behavior surveillance-United States, 1999." Mortality and Morbidity Weekly
21 Report 49(SS05):l-96.
22
23 Linn, W., Y. Szlachcis, H.J. Gong, P. Kinney, K. Berhane (2000). "Air pollution and daily hospital admissions in
24 metropolitan Los Angeles." Environmental Health Perspective 108:427-434.
25
26 McCurdy, T. (2000). "Conceptual basis for multi-route intake dose modeling using an energy expenditure
27 approach." J. Exposure Anal. Environ. Epidemiol. 10:1-12.
28
29 McCurdy, T., Glen, G., Smith, L., Lakkadi, Y. (2000). "The National Exposure Research Laboratory's Consolidated
30 Human Activity Database." J. Exposure Anal. Environ. Epidemiol. 10:566-578.
31
32 McDonnell, W.F. et al. (1985a). "Reproducibility of individual responses to ozone exposure." American Review of
33 Respiratory Disease 131:36-40.
34
35 McDonnell, W. F., Ill; Chapman, R. S.; Leigh, M W.; Strope, G. L.; Collier, A. M. (1985b) "Respiratory responses
36 of vigorously exercising children to 0.12 ppm ozone exposure." Am. Rev. Respir. Dis. 132: 875-879.
37
38 McDonnell, W.F. et al. (1991). "Respiratory response of humans exposed to low levels of ozone for 6.6 hours."
39 American Review of Respiratory Disease 147:804-810.
40
41 Moolgavkar, S. H.; Luebeck, E. G.; Hall, T. A.; Anderson, E. L. (1995). "Air pollution and daily mortality in
42 Philadelphia." Epidemiology 6: 476-484.
43
44 Morgan and Henrion (1990). Uncertainty: A Guide To Dealing with Uncertainty in Qualitative Risk and Policy
45 Analysis. Cambridge University Press.
46
47 Mortimer, K.M., L.M. Neas, D.W. Dockery, S. Redline, and I .B. Tager (2002). "The effects on air pollution on
48 inner-city children with asthma." European Respiratory Journal. 19:699-705.
49
50 Peel, J.L., P.E. Tolbert, M. Klein, K.B. Metzger, W.D. Flanders, K. Todd, J.M. Mulholland, P..B. Ryan, and H.
51 Frumkin, (2005). "Ambient air pollution and respiratory emergency department visits." Epidemiology 16(2): 164-
52 174.
53
54 Pope, C. A., R. T. Burnett, M. J. Thun, E. E. Calle, D. Krewski, K. Ito, and G. D. Thurston. 2002. Lung Cancer,
55 Cardiopulmonary Mortality, and Long-term Exposure to Fine Particulate Air Pollution. Journal of the American
56 Medical Association, vol 287,no 9: 287:1132-1141.
November 2005 5-64 Draft - Do Not Quote or Cite
-------
1
2 Post, E,, D. Hoaglin, L. Deck, and K. Lamtz (2001). "An Empirical Bayes approach to estimating the relation of .
3 mortality to exposure to particulate matter," Risk Analysis 21 (5): 837-842
4
5 Richmond H., T. Palma, J. Langstaff, T. McCurdy, G, Glenn, and L. Smith (2002). "Further refinements and testing
6 of APEX (3.0): EPA's population exposure model for criteria and air toxic inhalation exposures:" Poster
7 presentation. Joint meeting of the International Society of Exposure Analysis and International Society of
8 Environmental Epidemiology, August 11-15, 2002, Vancouver, Canada.
9
10 Schwartz, J. (2000). "The distributed lag between air pollution and daily deaths." Epidemiology 11(3):320-326.
11
12 Schwartz, J. (2004). "How sensitive is the association between ozone and daily deaths to control for temperature?"
13 Am. J. Resp. Crit. Care Med.
14
15 Schwartz, J., C. Spix, G. Touloumi, L. Bacharova, T. Bammamdzadeh, A. le Tertre, T. Piekarksi, A. Ponce de Leon,
16 A. Ponka, G. Rossi, M. Saez, J.P. Schouten (1996). "Methodological issues in studies of air pollution and daily
17 counts of deaths or hospital admissions."!. Epid. andComm. Health 50(Suppl 1):S3-S11.
18
19 Thurston, G.D., K. Ito, P.L. Kinney, M. Lippmann (1992). "A multi-year study of air pollution and respiratory
20 hospital admission in three New York State metropolitan areas: Results for 1988 and 1989 summers." J. Exposure
21 Anal. Environ. Epidemiol. 2(4):429-450.
22
23 Thurston, G.D. and Ito, K. (2001). "Epidemiological studies of acute ozone exposures and mortality." J. Exposure
24 Anal. Environ. Epidemiol. 11:286.
25
26 Tolbert, P.E., J.A. Mulholland, D.L. Macintosh, F. Xu, et al. (2000). "Air quality and Pediatric Emergency Room
27 Visits for Asthma in Atlanta, GA, USA." American Journal of Epidemiology 151(8):798-810.
28
29 Whitfield, R., Biller, W., Jusko, M., andKeisler, J. (1996). A Probabilistic Assessment of Health Risks Associated
30 with Short- and Long-Term Exposure to Tropospheric Ozone. Argonne National Laboratory, Argonne, IL.
31
32 Whitfield, R. (1997). A Probabilistic Assessment of Health Risks Associated with Short-term Exposure to
33 Tropospheric Ozone: A Supplement. Argonne National Laboratory, Argonne, IL.
34
35 Whitfield, R.G., Richmond, H.M. and Johnson, T.R. (1998). "Overview of Ozone Human Exposure and Health
36 Risk Analyses Used in the U.S. EPA's Review of the Ozone Air Quality Standard," pp. 483-516 in: T. Schneider, ed.
37 Air Pollution in the 215t Century: Priority Issues and Policy Elsevier; Amsterdam.
38 ,
39
40
41
November 2005 5-65 Draft - Do Not Quote or Cite
-------
1 6. STAFF CONCLUSIONS AND RECOMMENDATIONS ON
2 PRIMARY O3 NAAQS
3 6.1 INTRODUCTION
4 This chapter presents preliminary staff conclusions as to whether consideration should be
5 given to revising the existing primary 63 standard and, if so, what alternative standards should be
6 considered for additional exposure and risk assessments beyond those presented in Chapters 4
7 and 5. The results of these additional assessments will then be used to inform the staff
8 recommendations on the primary Oa NAAQS to be included in the next draft Staff Paper.
9 The existing 63 standard is an 8-hour standard set to protect public health from short-
10 term and prolonged exposures to Os. The standard is defined in terms of four basic elements:
11 indicator, averaging time, level and form. Preliminary staff conclusions on this standard and on
12 alternatives' for additional analyses are based on the assessment and integrative synthesis of
13 information presented in the draft CD and on initial staff analyses and evaluations presented in
14 Chapters 2 through 5 herein. As noted in Chapter 1, staff conclusions and recommendations
15 presented in the next draft Staff Paper will be further informed by consideration of the
16 information and analyses in the final CD, additional staff analyses and the results of the
17 completed population exposure and human health risk assessments, and CAS AC and public
18 comments received on this draft.
19 Staff notes that the final decision on retaining or revising the current standard is largely a
20 public health policy judgment. A final decision must draw upon scientific information and
21 analyses about health effects, population exposure, and risks, as well as judgments about how to
22 deal with the range of uncertainties that are inherent in the scientific evidence and analyses. The
23 staffs approach to these judgments, discussed more fully below, is based on a recognition that
24 the available health effects evidence generally reflects a continuum consisting of ambient levels
25 at which scientists generally agree that health effects are likely to occur through lower levels at
26 which the likelihood and magnitude of the response become increasingly uncertain. This
27 approach is consistent with the requirements of the NAAQS provisions of the Act and with how
28 EPA and.the courts have historically interpreted the Act. These provisions require the
29 Administrator to establish primary standards that, in the Administrator's judgment, are requisite
30 to protect public health with an adequate margin of safety. In so doing, the Administrator seeks
31 to establish standards that are neither more nor less stringent than necessary for this purpose.
32 The Act does not require that primary standards be set at a zero-risk level but rather at a level
33 that avoids unacceptable risks to public health.
November 2005 6-1 Draft - Do Not Quote or Cite
-------
1 6.2 APPROACH
2 In evaluating whether the current primary standard is adequate or whether consideration
3 of revisions is appropriate, and in developing recommendations on the elements of possible
4 alternative standards for further analyses, staffs approach in this review builds upon the general
5 approach used in the last review by expanding the exposure and risk assessments to reflect the
6 larger body of evidencenow available. The 1997 final decision notice (62 FR 38861) outlined
7 the key factors considered in selecting the elements of a standard for 63: the averaging time; 03
8 concentration (i.e., level); and the form (i. e., the air quality statistic to be used as a basis for
9 determining compliance with the standard). These factors represent an integration of
10 information on acute and chronic health effects associated with exposure to ambient Oj, expert
11 judgment on the adversity of such effects on individuals; and policy judgments, informed by air
12 quality, exposure assessment, and quantitative risk assessment when possible, as to the point at
13 which the standards are requisite to protect public health with an adequate margin of safely. This
14 approach to selecting a primary standard was endorsed by CASAC in the last review (Wolff,
15 1995b), particularly through its advice that "EPA's risk assessments must play a central role in
16 identifying an appropriate level" and recognition that "the selection of a specific level and [form]
17 is a policy judgment."
18 In developing these preliminary conclusions on the Os standard, staff has taken into
19 account evidence-based considerations primarily by assessing the evidence from
20 epidemiological, controlled human exposure, animal toxicological and field studies for a variety
21 of health endpoints. For those endpoints based on epidemiological studies, staff has placed
22 greater weight on associations with health endpoints that the draft CD has judged to be likely
23 causal based on an integrative synthesis of the entire body of evidence, including not only all
24 available epidemiological evidence but also evidence from animal toxicological and controlled
25 human exposure studies. Less weight is given to evidence of associations that are judged to be
26 only suggestive of possible causal relationships, although we have taken this information into
27 account as part of margin of safety considerations. For the purpose of evaluating the level of the
28 Oa standard in this review, staff has placed greater weight on U.S. and Canadian studies
29 reporting statistically significant associations. This is because findings of U.S. and Canadian
30 studies are more directly applicable for quantitative considerations in this review, since studies
31 conducted in other countries may well reflect quite different populations and air pollution
32 characteristics.
33 Staff has also taken into account quantitative exposure- and risk-based considerations,
34 drawn from the results of the exposure and risk assessments conducted in as many as twelve
35 urban areas (discussed in Chapters 4 and 5). More specifically, staff has considered estimates of
36 the magnitude of Os-related exposures and risks associated with current air quality levels, as well
November 2005 . 6-2 Draft - Do Not Quote or Cite
-------
1 as the exposure and risk reductions likely to be associated with meeting the current 8-hour
2 primary Os NAAQS. Staff recognizes the considerable uncertainties inherent in such estimates,
3 and, as described in Chapters 4 and 5, will take such uncertainties into account by considering
4 the sensitivity of the exposure and risk estimates to alternative assumptions likely to have
5 substantial impact on the estimates.
6 In mis review, as in the previous review, a series of general questions frames staffs
7 approach to reaching conclusions and recommendations, based on the available evidence and
8 information, as to whether consideration should be given to retaining or revising the current
9 primary Cb standard. Staffs preliminary review of the adequacy of the current standards begins
10 by considering whether the currently available body of evidence assessed in the draft CD
11 suggests that revision of any of the basic elements of the standards would be appropriate. More
12 specifically, mis evaluation of the adequacy of the current standard involves addressing
13 questions such as the following:
14 • To what extent does newly available information reinforce or call into question
15 evidence of associations with effects identified in the last review?
16 • To what extent does newly available information reinforce or call into question any of
17 the basic elements of the current standards?
18 • To what extent have important uncertainties identified in the last review been reduced
19 and have new uncertainties emerged?
20 To the extent that the evidence suggests that revision of the current standards would be
21 appropriate, staff then considers whether the currently available body of evidence supports
22 consideration of standards that are either more or less protective by addressing the following
23 questions:
24 • Is mere evidence that associations, especially likely causal associations, extend to air
25 quality levels mat are as low as or lower than had previously been observed, and what
26 are the important uncertainties associated with that evidence?
27 • Are exposures of concern and health risks estimated to occur in areas that meet the
28 current standard; are they important from a public health perspective; and what are the
29 important uncertainties associated with the estimated risks?
30 To the extent that mere is support for consideration of revised standards, staff then identifies
31 ranges of standards (in terms of indicators, averaging times, levels and forms) that would reflect
32 a range of alternative public health policy judgments, based on the currently available evidence,
33 as to the degree of protection that is requisite to protect public health with an adequate margin of
34 safety. In so doing, staff addresses the following questions:
35 • Does the evidence provide support for considering a different Os indicator?
November 2005 6-3 Draft - Do Not Quote or Cite
-------
1 • Does the evidence provide support for considering different averaging times?
2 * What ranges of levels and forms of alternative standards are supported by the evidence,
3 and what are the uncertainties and limitations in that evidence?
4 • To what extent do specific levels and forms of alternative standards reduce the
5 estimated exposures of concern and risks attributable to Oa, and what are the
6 uncertainties in the estimated exposure and risk reductions?
7 In this draft Staff Paper, staff develops preliminary recommendations on alternative standards to
8 be analyzed in additional exposure and risk assessments that will in turn be considered in the
9 next draft Staff Paper to help inform staff conclusions and recommendations as to whether
10 consideration should be given to retaining or revising the primary Oa NAAQS. The primary
11 standard for Oa is addressed in section 6.3 below, beginning with staffs consideration of the
12 adequacy of (he current primary Os standard in subsection 6.3.1. Subsequent subsections address
13 each of the major elements that define the Oa standard: pollutant indicator, averaging time, form
14 and level. This chapter concludes with a summary of alternative standards to be considered in
15 additional exposure and risk assessments in section 6.3.6. An additional section summarizing
16 key uncertainties and recommendations for additional research related to setting a primary 0^
17 standard will be included in the next draft Staff Paper.
18 63 PRIMARY O3 STANDARD
19 6.3.1 Adequacy of Current O3 Standard
20 In me last review, an important input to the primary NAAQS decision was the evidence
21 from human controlled exposure studies of healthy young subjects exposed for 1 to 8 hours. The
22 best documented health endpoints in these studies were decrements in forced expiratory volume
23 (FEV), also known as lung function decrements, and respiratory symptoms, such as cough and
24 chest pain on deep inspiration. For short-term exposures of one to three hours, group mean FEV
25 decrements were statistically significant for Oa concentrations only at and above 0.12 ppm, but
26 only when subjects engaged in very heavy exercise. By contrast, prolonged exposures of six to ,
27 eight hours can produce statistically significant FEV decrements at the lowest Oa concentrations
28 evaluated in studies, 0.08 ppm, even when experimental subjects are engaged in more realistic
29 intermittent moderate exercise levels. The health significance of this newer evidence led to the
30 conclusion in the 1997 final decision (62 FR 38856) that "the 8-hour averaging time is more
31 directly associated with health effects of concern at lower Oa concentrations than is .the 1 -hour
32 averaging time."
33 Also of particular importance in the last review were the following observations: (1) the
34 then-existing 1-hour standard provided little, if any, margin of safety for public health protection;
November 2005
6-4
Draft - Do Not Quote or Cite
-------
1 (2) there was clinical evidence of statistically significant responses at 6- to 8-hour exposures to
2 the lowest concentration evaluated, 0.08 ppm Oa, at moderate exertion, including: lung function
3 decrements, respiratory symptoms (e.g., cough, pain on deep inspiration), nonspecific bronchial
4 responsiveness, and biochemical indicators of pulmonary inflammation; (3) there was
5 epidemiological evidence of associations between ambient Os and increased respiratory hospital
6 admissions and emergency room visits; (4) toxicological evidence suggested that repeated long-
7 term exposures to Oa-induced lung tissue damage in experimental animals; however, uncertainty
8 regarding dosimetry and species sensitivity differences limited the quantitative use of these data;
9 and (5) concentration-based forms, within the range considered up to the fifth-highest
10 concentration form, were considered appropriate for a health-based primary Oj standard. This
11 form of the standard reflected recognition "... that Oa exhibits a continuum of effects, such that
12 there is no discernible threshold above which public health protection requires that no exposures
13 be allowed or below which all risks to public health can be avoided." (62 FR 38856) In making
14 the final decision in the last review, the Administrator recognized that important uncertainties
15 remained with regard to interpreting the role of other pollutants co-occurring with Oa, biological
16 mechanisms of health effects, human exposure, and quantitative risk assessment of the health
17 endpoints analyzed.
18 63.1.1 Evidence-based Considerations
19 Since the last review, important new information on O3-related health effects has
20 emerged, including new findings from:
21 ' • Dosimetry studies that clarify factors potentially affecting the regional distribution of
22 Oa in the respiratory tract and that provide improved bases for animal-to-human
23 extrapolation of experimental results.
24 • Experimental toxicological studies using controlled exposures of humans and
25 laboratory animals to delineate exposure-response relationships and the biochemical
26 mechanisms underlying the toxic effects, pathology and susceptibility.
27 • Epidemiological studies that provide important information about real-world exposures
28 and the effects of Os, including premature mortality, hospital admissions, emergency
29 room and doctors visits, on the general population, as well as in susceptible
30 populations, and that address many research needs identified during the last review.
31 The new evidence about (Vrelated effects on respiratory morbidity, cardiovascular morbidity
32 and total, as well as cardiovascular and respiratory, mortality is discussed below.
33 6.3.1.1.1 Respiratory Morbidity
34 As described in the draft CD (CD, section 8.4) and in Chapter 3 above, the integrative
35 assessment of these studies continues to support a causal association between short-term Oa
November 2005 6-5 Draft - Do Not Quote or Cite
-------
1 exposures and lung function decrements, respiratory symptoms and pulmonary inflammation that
2 were the most important basis for revising the 63 NAAQS in 1997. Statistically significant
3 associations between of lung function decrements and respiratory symptoms were found in
4 epidemiological studies even where the 98ttl/99lh percentile1 air quality values are well below the
5 level of the current standard. Not only are the newer findings consistent with the previous
6 review, but also there is better evidence about the physiological mechanisms by which Os causes
7 these effects. For all of these health endpoints, there exist considerable inter-individual
8 differences in the magnitude of responses to 0$. Inter-individual differences in lung function
9 and, to a lesser extent, respiratory symptoms are reproducible over a period of time, indicating
10 that some individuals are.consistently more responsive than others to Os. Identification of
11 population groups that are at increased risk to 0% due to either increased susceptibility or
12 increased potential for exposure, is based on their (1) biological responses to Oa, (2) existing
13 lung disease, (3) activity patterns, and/or (4) personal factors (e.g., age, nutritional status).
14 Newer information expands understanding of the physiological basis for increased
15 sensitivity in people with asthma. Newer studies continue to indicate that, relative to healthy
16 controls, people with asthma have somewhat larger decreases in pulmonary function in response
17 to Os. (CD, p. 8-27) New evidence also indicates that people with asthma may have increased
18 occurrence and duration of nonspecific airway responsiveness, and that people with pre-existing
19 allergic asthma may have increased airway responsiveness to allergens following Os exposure.
20 (CD, p. 8-29) Newly available human exposure studies suggest that people with asthma may
21 also have increased inflammatory responses relative to non-asthmatic subjects. (CD, p. 8-73)
22 The majority of epidemiological panel studies that evaluated respiratory symptoms and
23 medication use related to Os exposures focused on children. (CD, p. 8-44) These studies
24 suggest that Ch exposure may be associated with increased respiratory symptoms and medication
25 use in children with asthma. Taken together, these findings suggest that 63 exposure may be a
26 clinically important factor for people with asthma that can exacerbate the response to ambient
27 bronchoconstrictor substances and increase respiratory morbidity and possibly mortality (as
28 discussed further below).
29 At the time of the last review there was some epidemiological evidence of associations
30 between ambient Os and increased respiratory hospital admissions and emergency room visits
31 and only very limited evidence of associations for school absences and premature mortality.
32 Since the last review of the O3 standard, additional epidemiological studies have evaluated the
33 association between short-term exposures to Os and hospital admissions for respiratory causes.
1 The reason for using percentile values to describe maximum concentrations in the data from ozone studies
is discussed more fully in the footnote to Appendix 3A and in the section on the form of the standard below.
November 2005 6-6 Draft - Do Not Quote or Cite
-------
1 Large multi-city studies as well as several individual city studies (CD, Figure 8-5) have reported
2 positive, often statistically significant associations with total respiratory, asthma and COPD
3 hospitalizations, with statistical significance more often found in those studies that analyzed the
4 effect of 63 during the warm season. The draft CD indicates 1hat despite some inconsistencies
5 noted across the studies, the collective evidence supports the finding of significant and robust
6 effects of Os on respiratory hospitalization outcomes.
7 Although many new studies have evaluated the association between ambient 63 levels
8 and emergency department visits for respiratory causes, the evidence is still unclear. In general,
9 Os effect estimates from summer only analyses tended to be positive and larger, and the
10 estimates were more likely to be statistically significant, compared to results from cool season or
11 all year analyses. (CD, section 7.3.2, Figure 8-5) While several studies observed significant
12 associations between Os concentrations and emergency department visits for respiratory causes,
13 inconsistencies were observed which were at least partially attributable to differences in model
14 specifications and differences in the analyses. Because of this, the draft CD concludes that the
15 evidence remains inconclusive regarding effects of Os on the risk of emergency department
16 visits.
17 With regard to school absenteeism, two large U.S. studies and one study from Seoul,
18 Korea have investigated the relationship between ambient Os and this effect. All of the studies
19 found statistically significant positive associations between Os levels and absences from school.
20 Because of differences in the analyses, the draft CD concludes that results from these three
21 studies suggest that ambient Os concentrations may be associated with school absences,
22 especially illness-related absences. Additional studies and analyses using similar lag periods are
23 needed to more clearly delineate quantitative relationships between ambient Os and school
24 absences. (CD, p. 8r45)
25 6.3.1.1.2 Cardiovascular Morbidity
26 There is limited, new evidence supporting associations between short-term Os exposures
27 and a range of effects on the cardiovascular system. An increasing body of animal toxicology
28 evidence suggests that hematological and thermoregulatory alterations (in heart rate variability
29 and/or core body temperature) may mediate acute cardiovascular effects. A few controlled
30 human exposure studies have examined the potential effects,of O3 exposure on cardiovascular
31 functions. These studies have reported impairment of alveolar-arterial oxygen transfer and Os-
32 induced ventilation-perfusion mismatch, suggesting gas exchange abnormalities that could affect
33 cardiac function. Some but not all, epidemiological studies have reported associations between
34 short-term Os exposures and the incidence of myocardial infarction and more subtle
35 cardiovascular health endpoints, such as changes in heart rate variability and cardiac arrhythmia.
November 2005 6-7 Draft - Do Not Quote or Cite
-------
1 Based on epidemiological study results, the draft CD concludes that the current evidence from
2 field studies is rather limited but supportive of a potential effect of short-term 03 exposure and
3 * heart rate variability, cardiac arrhythmia and incidence of myocardial infarction (CD, p. 7-57).
4 A subset of hospital admission studies examined the effect of Oa exposure on
5 cardiovascular outcomes. The evidence is inconclusive on the association between Os exposure
6 and cardiovascular hospitalizations with regard to year-round data. However, the draft CD
7 concludes that in studies that adjusted for seasonal or meteorological factors, there is suggestive
8 evidence that 63 is associated with increased risk for cardiovascular hospital admissions in the
9 warm season. (CD, p. 8-48) Studies also report associations between short-term Oa exposure and
10 mortality from cardiovascular or cardiopulmonary causes (as discussed further below).
11 6.3.1.1.3 Mortality
12 The 1996 CD concluded that an association between daily mortality and Oa concentration
13 for areas with high Oa levels (e.g., Los Angeles) was suggested. However, due to a very limited
14 number of studies available at that time, the magnitude of rne effect was unclear. Since 1996,
15 newly available large multi-city studies designed specifically to examine the effect of Oa on
16 mortality have provided much more robust and credible information. New data are also
17 available from several single-city studies conducted all over the world, as well as from several
18 meta-analyses that have combined information from multiple studies. The majority of these
19 studies suggest an elevated risk of total non-accidental mortality associated with acute exposure
20 to Oa, especially in the summer or warm season when Oa levels are typically high, with
21 somewhat larger effect estimate sizes for associations with cardiovascular mortality. (CD, p. 7-
22 149) Some of the single city studies with positive and statistically significant results have 98th or
23 99th percentile air quality values well below the level of the current 8-hour Oa standard.
24 (Appendix 3A) The draft CD finds that the results from U.S. multi-city time-series studies,
25 along with the meta-analyses, provide strong evidence for associations between short-term Oa
26 exposure and mortality. (CD, p. 7-84) The results of these analyses show that the effects of
27 ozone on mortality are generally robust to confounding by copollutants. (CD, p. 7-149, 8-54)
28 For cardiovascular mortality, the draft CD reports that effect estimates are consistently positive,
29 and are more likely to be larger and statistically significant in the warm season analyses. (CD, p.
30 7-108, Figure 7-22) The findings regarding the effects size for respiratory mortality have been
31 less consistent, possibly due to lower statistical power in this group. (CD, p. 7-94) Overall, the
32 draft CD concludes that these findings suggest a causal association between short-term Oa
33 exposure and mortality particularly in the warm season. (CD, p. 8-84)
34 6.3.1.2 Risk-based Considerations
35 In discussing risk-based considerations, this section will focus first on the results of the
36 exposure assessment and then on the results of the risk assessments that were based on clinical .
37 and epidemiological evidence. As described in Chapter 4, for this review estimates of exposures
November 2005
6-8
Draft - Do Not Quote or Cite
-------
1 were calculated for active people of all ages, school age children (ages 5-18), and "active" school
2 age children.2 For the initial exposure analyses in this review, an "exposure of concern" was
3 defined the same way as in the previous review. -An "exposure of concern," as defined in the
4 1997 review of the O3 standard, is an 8-hour average exposure to 0.08 ppm O3 while
5 intermittently at moderate or greater exertion levels. (62 FR 38860) Exposure results are
6 displayed in Tables 4-7 and 4-8 for daily maximum 8-hour average exposures above 0.08 ppm
7 O3, in 12 urban areas across the U.S., for two cases (i.e., recent (2004) air quality, and just
8 meeting Hie 8-hour primary standard), at moderate or greater exertion levels for three groups:
9 active people of all ages; all school age children; and active school age children. Estimates,
10 aggregated across these 12 urban areas of the number of people exposed, in each of the three
11 groups, and the number of person-days (occurrences) of exposures, with daily maximum 8-hour
12 average exposures above 0.08 ppm while at moderate or greater exertion, are shown in Table 6-1
13 below.3
14 Under the recent (2004) air quality scenario, 1.6 million people, or 2 % of the total
15 population of the 12 urban areas, are estimated to experience one or more exposures of concern.
16 More than' 600,000 children, or'3% of the total number of children ages 5-18, and almost
17 400,000 active children, or 4 % of active children ages 5-18, are estimated to experience one or
18 more exposures of concern. When air quality is adjusted to simulate just meeting the 8-hour
19 standard, the number of people exposed drops substantially. Approximately 50,000 people, or
20 less than 0.1% of the total population of the 12 urban areas, are estimated to experience one or
21 more exposures of concern. Approximately 17,000 children (< 0.1 % of all children) and 11,000
22 active children (0.1 % of active children) are estimated to experience exposures of concern when
23 air quality just meets the 8-hour standard. These results suggest a substantial reduction in
24 estimated 8-hour average exposures above 0.08 ppm when the current 8-hour O3 standard is just
25 met. The estimated reduction in the number of total occurrences of exposures of concern in
26 these 12 urban areas was more than 97% in each of the three population groups. Moreover, a'
27 comparison of the number of people exposed with the number of occurrences indicates that very
28 few people are likely to be exposed more than one time during the O3 season. Under the current
29 standard it is estimated to be rare for individuals to experience more than one 8-hour exposure
30 above 0.08 ppm O3 while at moderate or greater exertion levels in these 12 urban areas during an
31 O3 season with air quality similar to 2004.
2As indicated in section 4.4.3 above, "active" school age children were defined as those with a physical
activity index>\.75.
3For greater discussion of the analyses and a breakdown of exposures by city, see Chapter 4.
November 2005 6-9 Draft - Do Not Quote or Cite
-------
W3
i
%
iS
¥ g
H *S
t3
3 §
«> ^3
ii
^-^
^
oo K
I-H no
^1S
••^ O
§ §
2 'J3
2 1|
U «
9 •""!
^,
M _^
If
S 3
1 1
fl
1 §
^v « vH
^ £
jl»
X'o J
»2 8
£s&
w
11
O ^
2 o
PH S
g
•S 'Q
ra O
s'i
o 3
1^1
en *S 2
C vo -S
2 T3 8
(^ C3 C5
PU 51 O.
00
rj_, }-| S3
o o S
"5> o i-
o § o
^i
Al
Recent Air
Quality (2004)
„
o
•§
IJ
1,
S ^
§
<=> ^
o
= 1
s^
I-H
Al
Just Meeting
Current 8-hour
Standard
-
vi S^-
N_/
00
rn sp
m f^
t
-•
g
cs.
AI;
Recent Air
Quality (2004)
r-,
0
U
"I
o
oo
So?
_0
o
£"2
xP
O
oo i
2|
-
o
(N
Al
Just Meeting
Current 8-hour
Standard
oo
a*
I
•3
§
O)
a
I
"8
I
vo
-------
1 Several aspects of the risk information presented are important to consider. The first is
2 that there is some degree of consistency in the estimated population risk across the 12 urban
3 areas, as indicated by the percent of the population estimated to be affected, which describes the
4 risk normalized across the populations. In Table 6-2, the percent of the all children likely to
5 experience one or more moderate or greater lung function responses under recent (2004) air
6 quality and when air quality just meets the current 8-hour standard are 10% and 6%,
7 respectively. The range across the 12 urban areas, from Table 5-5, is about 7% to 13% under the
8 recent (2004) air quality, and about 5% to 8 % when air quality just meets the current 8-hour
9 standard. More than one million children are estimated to experience one or more moderate or
10 greater lung function responses in these 12 urban areas when the current 8-hour primary standard
11 is met. Many of these children will experience repeated occurrences of moderate or greater lung
12 function responses. These results indicate that each of these children is likely to .experience, on
13 average, 7 occurrences of moderate or greater lung function responses during an Os season.4
14 However, based on the distribution of exposures estimated from the 1997 review, the more likely
15 distribution will be that many children will experience one or a just few moderate or greater lung
16 function responses, while a smaller number of children will experience large numbers of such
17 responses. This range of estimated number of occurrences (i.e., from one to many, with a mean
18 of approximately 7) of moderate or greater lung function decrements in an Os season is
19 important in considering the implications for the health status of individuals likely to experience
20 these effects. Moderate or greater lung function decrements are transient and reversible, so the
21 extent to which such effects are considered to be adverse to the health status of the individual
22 depends not only on the severity and duration of the effect, but also on the frequency with which
23 an individual experiences such effects throughout an Os season. (62 FR 38864)
24 For non-accidental mortality associated with "as is" O3 concentrations (Table 5-7), the
25 estimates of the percent of total mortality attributable to Os exposure ranges from 0.1 to 1.9%;
26 with an incidence per 100,000 relevant population ranging from 0.4 to 7.3. Estimated non-
27 accidental mortality associated with Os concentrations that just meet the current 8-hour standard
28 (Table 5-10) ranges from 0.1 to 1.5% of total incidence; with an incidence per 100,000 relevant
29 population ranging from 0.3 to 5.7. Estimated cardiovascular and respiratory mortality shows a
30 similar pattern (Tables 5-8, 5-11). For a recent year (2004) of Os concentrations, the estimates of
31 the percent of cardiovascular and respiratory mortality attributable to 63 ranges from 0.4 to
32 1.6%; the incidence per 100,000 relevant population ranges from 0.4 to 1.3. Estimated
33 cardiovascular and respiratory mortality associated with 63 concentrations that just meet the
4 This number is estimated for example for all children, by dividing the estimated number of children
(1,156,000) into the estimated number of occurrences (7,640,000) resulting in an average of 7 occurrences per child.
November 2005 6-13 Draft - Do Not Quote or Cite
-------
1 current 8-hour Os standard ranges from 0.2 to 0.8% of total incidence; with the incidence per
2 100,000 relevant population ranging from 0.2 to 0.9.
3 For unscheduled hospital admissions, risk estimates for the New York City area are
4 shown in Table 5-9 .for recent (2004) air quality, and Table 5-12 for just meeting the current 8-
5 hour standard. Table 5-9 shows estimates of unscheduled hospital admissions for respiratory
6 illness in the New York City area associated with Oj levels above background of about 450 cases
7 or 5.6 cases per 100,000 relevant population, which represents 1.3% of total incidence for recent
8 (2004) Oa levels. For asthma-related hospital admissions, the estimates are about 380 cases or
9 4.8 cases per 100,000 relevant population,- which represents about 2.9% of total incidence. Table
10 5-12 shows estimates of unscheduled hospital admissions for both respiratory illness and asthma
11 in New York City associated with Oj levels above background for the period from April to
12 September with air quality adjusted to just meet the current 8-hr standard. For total respiratory
13 illness, Table 5-12 shows about 450 cases or 5.6 cases per 100,000 relevant population, which
14 represents 1.3% of total incidence for recent (2004) Os levels. For asthma-related hospital
15 admissions, which are a subset of total respiratory illness admissions, the estimates are about 380
16 cases or 4.8 cases per 100,000 relevant population, which represents about 2.9% of total
17 incidence. The reduction in incidence for both total respiratory illness and asthma admissions
18 from the recent O? levels to air quality just meeting the current 8-hr standard is about 19%.
19 6.3.1.3 Summary
20 These initial analyses suggest that meeting the current 8-hour C*3 standard would likely
21 result in substantial reductions in exposures of concern and associated risks of serious health
22 effects above a level of 0.08 pprh 03. On the other hand, these analyses also suggest that there is
23 risk of moderate or greater lung function decrements in children, hospital admissions, and
24 mortality from Os resulting from exposures across the range of levels allowed by the current
25 standard. Staff concludes that the estimates discussed above are indicative of risk that some
26 might reasonably judge to be important from a public health perspective. Thus, staff believes
27 that it is appropriate to perform additional analyses so as to be able to consider the potential
28 reduction in exposures and risks from alternative standards that may provide more health
29 protection beyond that afforded by the current Os primary standard.
30 6.3.2 Indicator
31 The staff believes that the conclusions on the appropriate indicator for the primary Oa
32 NAAQS that were reflected in the 1996 Staff Paper remain valid today. It is generally
33 recognized that control of ambient Os levels provides the best means of controlling
34 photochemical oxidants of potential health concern. Further, among the photochemical oxidants,
35 the acute exposure chamber, panel and field epidemiological human health database raises .
November 2005
6-14
Draft - Do Not Quote or Cite
-------
1 concern only for Os at levels of photochemical oxidants commonly reported in the ambient air.
2 Thus the staff does not believe that it is appropriate to consider any other indicator for additional
3 analyses.
4 6.3.3 Averaging Time
5 63.3.1 Short-Term and Prolonged (1 to 8 hours)
6 The current 8-hour averaging time for the primary Oj NAAQS was set in 1997. The
7 decision to revise the averaging time of the primary standard from 1 to 8 hours was supported by
8 the following key observations and conclusions (62 FR 38861):
9 (1) The 1-hour averaging time of the previous NAAQS was originally selected on the
10 basis of health effects associated with short-term (i.e., 1 - to 3-hour) exposures.
11 (2) Substantial health effects information was available for the 1997 review that
12 demonstrated associations between a wide range of health effects (e.g., moderate to large lung
13 function decrements, moderate to severe symptoms and pulmonary inflammation) and prolonged
14 (i.e., 6- to 8-hour) exposures below the level of the NAAQS.
15 (3) Results of the quantitative risk analyses showed that the reductions in risks from both
16 short-term and prolonged exposures could be achieved through a primary standard with an
17 averaging period of either 1 or 8 hours.
18 (4) The 8-hour averaging time is more directly associated with health effects of concern
19 at lower Os concentrations than the 1 -hour averaging time. It was thus the consensus of CAS AC
20 "that an 8-hour standard was more appropriate for a human health-based standard than a 1-hour
21 standard." (Wolff, 1995b)
22 In looking at the new information that is discussed in section 7,6.2 of the draft CD,
23 epidemiological studies have used various averaging periods for Os concentrations, most
24 commonly 1-hour, 8-hour and 24-hour averages. As described more specifically below, in
25 general the results presented from U.S. and Canadian studies (Appendix 3A) show no consistent
26 difference for various averaging times in different studies.
27 Only a few studies presented results for different Os averaging periods using the same
28 data set. Two of the recent multi-city mortality studies reported associations for multiple
29 averaging times (Bell et al., 2004; Gryparis et al., 2004). Both reported that the effect estimates
30 for different averaging times were not statistically different, though the effect estimates for
31 associations with 1-hour daily maximum OB concentrations were somewhat larger than those for
32 longer averaging times, especially 24-hour average Os. In addition, Gent et al., (2003) reported
33 that associations for 1-hour and 8-hour average OB with respiratory symptoms were not
34 significantly different.
November 2005 6-15 Draft - Do Not Quote or Cite
-------
1 Among the single-city epidemiological studies, Peters et al. (2001) reported positive, but
2 not statistically significant associations between Os and the incidence of myocardial infarction
3 (CD, p. 7-55); this study differs from most since the short-term Os concentration used was the
4 time period preceding the health event, not the highest daily short-term average concentration.
5 The effect estimate for the association with Os averaged over 2 hours prior to the myocardial
6 .infarction was substantially larger than that reported for an association with 24-hour average Os
7 (Peters et al., 2001). The draft CD reports results for a number of single-city results that
8 generally reported effect estimate sizes that were larger when comparing 1-hour or 8-hour daily
9 maximum Os concentrations with 24-hour concentration, but the results did not differ
10 statistically (CD, p. 7-120).
11 The CD observes that the various Os average concentrations were generally very highly
12 'correlated with one another, so it is not surprising that effect estimates would be similar. The •
13 draft CD concludes that the epidemiological study results were generally comparable for the
14 three Os averaging times (CD, p. 7-120). Given that the 8-hour averaging time continues to be
15 more directly associated with health effects of concern from controlled human exposure studies
16 at lower concentrations than do shorter averaging periods, staff concludes that it is appropriate to
17 continue to base any additional exposure and risk analyses for short-term or prolonged effects on
18 8-hour average Os concentrations.
19 63.3.2 Long-Term
20 • For consideration during the last review, there was a large animal toxicological database
21 providing clear evidence of associations between long-term (e.g., from several months to years)
22 exposures and lung tissue damage, with additional evidence of reduced lung elasticity and
23 accelerated loss of lung function, but there was no corresponding evidence for humans.
24 Furthermore, the state of the science had not progressed sufficiently to allow quantitative
25 extrapolation of the animal study findings to humans. For these reasons, consideration of a
26 separate long-term primary Os standard was not judged to be appropriate at that time.
27 In the current review, long-term animal toxicological studies continue to support the
28 relationship between Os exposure and structural alterations in several regions of the respiratory
29 tract and identify the CAR as the most affected region. In addition, animal'toxicological studies
30 that utilized exposure regimens to simulate seasonal exposure patterns also report increased lung
31 injury compared to conventional long-term, stable exposures. (CD, p. 8-85) Collectively, the
32 evidence from animal studies strongly suggest that Os is capable of damaging the distal airways
33 and proximal alveoli, resulting in lung tissue remodeling leading to apparently irreversible
34 changes. Compromised pulmonary function and structural changes due to persistent
35 inflammation may exacerbate the progression and development of chronic lung disease. (CD, p.
November 2005 6-16 Draft - Do Not Quote or Cite
-------
1 8-70) There is also some new information about the effects of long-term exposures in humans.
2 Epidemiological studies investigating chronic effects in humans following long-term exposures
3 to Os previously provided only limited suggestive evidence; however, recent studies of
4 pulmonary function changes observed in children living in cities with high Os levels as well as
5 alterations in lung structure reported in an autopsy study in Los Angeles and Miami provide
6 additional evidence that long-term Os exposure may play a role in causing irreversible lung
7 damage. However, the strength of the evidence overall does not allow establishment of a likely
8 causal relationship between long-term Os exposures and increased respiratory morbidity and
9 mortality. In addition, although there have been recent advancements of do'simetry modeling,
10 providing a better basis for extrapolation, especially of long-term exposures, from animals to
11 humans (CD, p. 8-11), further research must be conducted before quantitative linkages to
12 specific health effects in humans can be established with sufficient certainty to include in a
13 quantitative manner in the review. For these reasons, staff concludes that it is not appropriate at
14 this time to base any exposure or risk assessments on long-term exposures to Os.
15 63.4 Form
16 In evaluating alternative forms for the primary standard, the adequacy of the public health
17 protection provided is the foremost consideration. Staff recognizes that it is important to have a
18 form of the standard that is stable and insulated from the impacts of extreme meteorological
19 events that are conducive to Os formation. Instability can have the effect of reducing public
20 health protection, because when areas are subject to shifting in and out of attainment simply
21 because of meteorological conditions it can disrupt ongoing implementation plans and associated
22 control programs. Providing more stability is one of the reasons that in 1997 the primary Os
23 NAAQS was changed from a "1-expected-exceedance" form5 to a less extreme, concentration-
24 based statistic, specifically the 3-year average of the annual fourth-highest daily maximum 8-
25 hour concentrations. The principal advantage of the concentration-based form is that it is more
26 directly related to the ambient Os concentrations that are associated with the health effects. With
27 a concentration-based form, days on which higher Os concentrations occur would weigh
28 proportionally more than days with lower concentrations, since the actual concentrations are
29 used in determining whether the standard is attained. That is, given that there is a continuum of
30 effects associated with exposures to varying levels of Os, the extent to which public health is
31 affected by exposure to ambient Os is related to the actual magnitude of the Os concentration, not
32 just whether the concentration is above a specified level.
The 1-expected-exceedance form essentially requires that the fourth-highest air quality value in 3 years,
based on adjustments for missing data, to be less than or equal to the level of the standard for the standard to be met
at an air quality monitoring site.
November 2005 6-17 Draft - Do Not Quote or Cite
-------
1 During the 1997 review, consideration was given to a range of alternative forms,
2 including the second-, third-:, fourth- and fifth-highest daily maximum 8-hour concentrations,
3 recognizing that the public health risks associated with exposure to a pollutant without a clear,
4 discernable threshold are best addressed through a standard that allows for multiple exceedances
5 to provide increased stability, but that also significantly limits the number of days on which the
6 level may be exceeded and the magnitude of such exceedances. Consideration was given to
7 setting a standard with a form at the lower end of the range to provide a margin of safety against
8 possible, but uncertain chronic effects, or at the upper end of the range to provide greater
9 stability to ongoing control programs. The fourth-highest daily maximum was selected because
10 it was decided that the difference between the protection against potential chronic effects
11 afforded by the alternatives within the range was not well enough understood to. use as a basis for
12 choosing the most restrictive forms. On the other hand, the relatively large percentage of sites
13 that would experience Os peaks well above 0.08 ppm and the number of days on which the level
14 of the standard may be exceeded even when attaining a fifth-highest 0.08 ppm concentration
15 standard, argued against choosing that form.
16 For the purposes of making recommendations of alternative standards for additional
17 exposure and risk analyses, staff considered two concentration-based forms, the nth highest
18 maximum concentration and the percentile-based form of the standard. A percentile-based
19 statistic, as is used in Figure 6-la and b, is useful for comparing datasets of varying length
20 because it samples approximately the same place in the distribution of air quality values, whether
21 the dataset is several months or several years long. However, a percentile-based form would
22 allow more days with higher air quality values in locations with longer Os seasons relative to
23 places with shorter Os seasons. An nth highest maximum concentration form would do a better
24 job of ensuring that people who live in areas with different length Os seasons receive the same
25 degree of public health protection. For this reason, the staff recommends that further analyses be
26 based on a form specified in terms of an nth-highest concentration over a three-year period, with
27 n ranging from 3 to 5. Staff believes that his range is sufficiently broad to allow for
28 consideration of the degree of protection requisite to protect public health and also to ensure
29 stability of the standard. ;
30 6.3.5 Level
31 Since the last review, the body of evidence from animal toxicology and dosimetry
32 studies, controlled human exposure and epidemiological studies has confirmed associations
33 between exposure to Os and the effects that were the basis for the 1997 Os standard: lung
34 function decrements; symptoms; airway hyper-responsiveness; pulmonary inflammation; and
35 respiratory-related hospital admissions. New evidence has provided a better understanding of
November 2005 . 6-18 Draft - Do Not Quote or Cite
-------
1 the pathophysiological mechanisms of these effects. This information has also provided new
2 evidence of likely causal associations with more serious effects including associations between
3 increasing Os concentrations and hospital admissions for respiratory causes and all cause (non-
4 accidental) mortality. Positive, but inconclusive associations have been found, between ambient
5 Os concentrations and respiratory symptoms, school absences, emergency department visits,
6 cardiovascular system effects, and cardiovascular and respiratory mortality.
7 Although it is reasonable to expect that there are biological thresholds for different health
8 effects in individuals or groups of individuals with similar health status, most studies designed to
9 evaluate the existence of a threshold have observed no deviation from a linear function across the
10 range of Os measurements from the study, as discussed above in Chapter 3^ section 3.4.6. One
11 study found a potential threshold level of about 45 ppb (1 -hour maximum concentration) for an
12 association between mortality and short-term Oa exposure during the summer months. -Another
13 found some evidence of a threshold at about 30 ppb (1-hour maximum concentration) for the
14 association between Os concentrations and both respiratory and cardiovascular hospital
15 admissions. Other studies, such as ones that have removed days with higher concentrations of
16 Os from the data set to test the association between Os at lower levels and health outcomes have
17 found that associations remain or are increased in magnitude. In summary, many
18 epidemiological studies have suggested that no threshold levels can be found. In those studies
19 that provide suggestive evidence of threshold, the potential thresholds are at low concentrations,
20 much lower than the current 8-hour standard.
21 Staff concludes that the results from the initial quantitative exposure and risk assessments
22 done to date suggest that sufficient population exposures and risk of various health endpoints
23 (including pulmonary function decrements, all cause (non-accidental) and respiratory and
24 cardiovascular mortality, and respiratory hospital admissions) remain after meeting the current 8-
25 hour standard that additional analyses for standards below the level of the current 8-hour
26 standard are appropriate. In addition to consideration of exposures and effects that are
27 quantifiable in population exposure and risk assessments, staff believes that it is also appropriate
28 to consider effects that it is not currently possible to quantify, for example the effects of chronic
29 Os exposure or the effects on people with asthma, to the extent these considerations would
30 ultimately cause closer examination of the lower end of the range, in developing
31 recommendations about alternative standards for further exposure and risk assessment.
32. The question then becomes, what level or levels should be evaluated? To answer this
33 question, staff turns to epidemiological studies of associations between O3 and respiratory
34 symptoms and hospital admissions for respiratory causes from warm season analyses. (Figures
35 6-1 a and b) Not included in Figure 6-1, but also considered are the reported associations
36 between 8-hour average Oa levels and daily mortality (Appendix 3 A). All of these effects
November 2005 6-19 Draff -Do Not Quote or Cite
-------
1 estimates are consistently positive and many of the reported associations are statistically
2 significant. In this body of evidence, small but statistically significant effects estimates have
3 been reported in studies with 98th/99th percentile values as.low as approximately 0.06 ppm Oa, 8-
4 hour average. The information is presented in terms of the 98th and.99th percentile values
5 because in the studies being compared, the data spanned widely varying periods of time
6 After consideration of the entire body of experimental and epidemiological evidence, the
7 results of exposure and risk assessments and the consideration of non-quantifiable effects, such
8 as the effects of repeated exposures-and potentially greater effects on people with asthma, it is
9 staffs view that it is appropriate to conduct additional exposure and risk assessments down to an
10 alternative standard level as low as 0.06 ppm. The level of 0.06 ppm represents the lowest air
11 quality statistic credibly and significantly associated with increased respiratory morbidity effects
12 such as symptoms and hospital admissions, and also with daily mortality.
November 2005
6-20
Draft - Do Not Quote or Cite
-------
1 Figui
2 (a)
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19 (b)
re 6-1 a and b
Odds Ratios (with 95% confidence intervals) for associations between O3
and respiratory symptoms, from warm season analyses, in order of
increasing 98th and 99th percentile 8-hr O3 concentrations (in boxes)
?5IJ
2.00
1.50 .
1.00
0.60
LEGEND:
S = asthma symptoms
MS= morning symptoms
ES = evening symptoms
MD = med use
CT = chest tightness
SB = shortness of breath
C = cough
W = wheeze
Effect estimate
64.3
66
.T
85.3
I 98
68.8 90''
75
. , \
S MS ES MD CT
C
107
I 109 p—. ,
121
^
SB s MD s w c'
VI
-y- Mortimer et al., 2002. 6 US cities
W
3-
i £ i ! 1
s t S> c iA C O
i s ill 1
a: z o ? a o
95% confidence intervals) for associations betwee
O3 and hospitalization for respiratory diseases, in warm season analyses, in
order of increasing 98th and 99th percentile 8-hr O3 concentrations (in boxes)
40 -
30 .
10 -
0 .
-10 .
f
f
F
1
A
?
ff\
73
A
\
A
68.4
74.8
1
C
\
Rl
69
73.8
\
R A
77.7
83.7
•x
79
81.5
r
R
P C
91
99
\
R
i
5
F
U
4
?
125.5
133
I A
\ +
i r T \
'
R A R A R
•a «i
•s *
m m
ra (0
•s -s
e §
I
S.
ra
•a
•s
+ 98" percentile values not available; are ikely high-end concentrations based on data provided In study
November 2005
6-21
Draft - Do Not Quote or Cite
-------
2 6.3.6 Summary of Alternative Standards to Be Considered in Additional Exposure
3 and Risk Analyses
4 Staff has considered the evidence from animal toxicology, epidemiological and
5 controlled human exposure studies, estimates of exposures and risk for a recent (2004) year and a
6 year when air quality is estimate to just meet the current 8-hour standard, and the related
7 limitations and uncertainties. For the reasons described above, staff provisionally concludes that
8 it is appropriate to conduct additional assessments of exposure and risk associated with
9 alternative standards that may provide increased protection beyond that afforded by the current
10 8-hour primary Os standard. Staff recommends that additional exposure and risk assessments be
11 conducted for alternative, 8-hour average standards at Os levels of 0.08 ppm, third-highest
12 concentration, 0.07 ppm third- through fifth-highest concentration, and if appropriate based on
13 these results, 0.06 ppm, third- through fifth-highest concentration, over a three-year period. This
14 combination of alternative levels and forms will provide a more complete picture of the risk from
15 lower concentrations where the public health risks are relatively small and more uncertain to
16 higher concentrations where the public health risks are greater, but the effects are more certain.
17 The results of these additional assessments will then be used to inform staff recommendations to
18 be included in the next draft Staff Paper on the primary Os NAAQS.
19 6.4 SUMMARY OF KEY UNCERTAINTIES AND RESEARCH
20 RECOMMENDATIONS RELATED TO SETTING A PRIMARY O3 STANDARD
21 [To be included in the next draft Staff Paper]
November 2005
6-22
Draft - Do Not Quote or Cite
-------
1 REFERENCES
2
3 Bell, M. L.; McDerraott, A.; Zeger, S. L.; Samet, J. M.; Dominici, F. (2004) Ozone and short-term mortality in 95
4 US urban communities, 1987-2000. JAMA J. Am. Med. Assoc. 292: 2372-2378.
5
6 Gent, J. F.; Triche, E. W.; Holford, T. R.; Belanger, K.; Bracken, M. B.; Beckett, W. S.; Leaderer, B. P. (2003)
7 Association of low-level ozone and fine particles with respiratory symptoms in children with asthma. JAMA
8
9 Gryparis, A.; Forsberg, B.; Katsouyanni, K.; Analitis, A.; Touloumi, G.; Schwartz, J.; Samoli, E.; Medina, S.;
10 Anderson, H. R.; Niciu, E. M.; Wichmann, H.-E.; Kriz, B.; Kosnik, M.; Skorkovsky, J.; Vonk, J. M.;
11 Ddrtbudak, Z. (2004) Acute effects of ozone on mortality from the "air pollution and health: a European
12 approach" project. Am. J. Respir. Grit. Care Med. 170:1080-1087.
13
14 Peters, A.; Dockery, D. W.; Muller, J. E.; Mittleman, M. A. (2001) Increased particulate air pollution and the
15 triggering of myocardial infarction. Circulation 103: 2810-2815.
16
17 U.S. Environmental Protection Agency (I996b) Review of the national ambient air quality standards for ozone:
18 assessment of scientific and technical information. OAQPS staff paper. Research Triangle Park, NC: Office
19 of Air Quality Planning and Standards; EPA report no. EPA-^52/R-96-007, Available from: NTIS,
20 Springfield, VA; PB96-203435.
21
22 Wolff, G. T., (1995b) Letter from Chairman of Clean Air Scientific Advisory Committee to the EPA Administrator,
23 dated November 30,1995. EPA-SAB-CASAC-LTR-96-002.
November 2005 6-23 Draft - Do Not Quote or Cite
-------
1 7. POLICY-RELEVANT ASSESSMENT OF WELFARE EFFECTS
2 EVIDENCE
3 7.1 INTRODUCTION
4 This chapter presents critical information for the review of the secondary NAAQS for Os.
5 Welfare effects addressed by a secondary NAAQS include, but are not limited to, effects on
6 soils, water, crops, vegetation, man-made materials, animals, wildlife, weather, visibility and
7 climate, damage to and deterioration of property, and hazards to transportation, as well as effects
8 on economic values and on personal comfort and well-being. Of these welfare effects categories,
9 the effects of Os on vegetation, including agricultural crops, trees in managed and unmanaged
10 forests, herbaceous and woody species growing in natural settings are of most concern at
11 concentrations typically occurring in the U.S. As stated in earlier reviews, "of the phytotoxic
12 compounds commonly found in the ambient air, Os is the most prevalent, impairing crop
13 production and injuring native vegetation and ecosystems more than any other air pollutant"
14 (U.S. EPA, 1989,1996).
15 Ozone can also affect other ecosystem components such as soils, water, animals, and
16 wildlife, either directly, or indirectly through its effects on vegetation. These individual
17 ecosystem components are associated with one or more of six essential ecological attributes
18 (EEAs) recently described in A Framework for Assessing and Reporting on Ecological
19 Condition: an SAB report (Young and Sanzone, 2002) as part of a conceptual framework useful
20 for assessing and reporting on ecological condition (see 7.4). This framework can be used to link
21 ozone effects at the species level to potential impacts at higher levels in the hierarchy (e.g.,
22 EEAs). Some of these species level impacts have direct, quantifiable economic value, while
23 others are currently not quantifiable, but still have societal value. In the absence of sufficient
24 research to allow quantification of Os impacts at the ecosystem level, including impacts on
25 ecosystem goods and services, only a qualitative discussion is included. However, the staff
26 infers, based on the linkages described in the SAB framework, that increasing protection for
27 vegetation from ozone related effects would also improve the protection afforded to ecosystems
28 and their related public welfare categories.
29 Other welfare effects categories affected by Os include damage to certain manmade
30 materials (e.g., elastomers, textile fibers, dyes, paints, and pigments) and effects on and by
31 climate. The amount of damage to actual in-use materials and the economic consequences of
32 that damage are poorly characterized, however, and the scientific literature contains very little
33 new information to adequately quantify estimates of materials damage from photochemical
34 oxidants (EPA, 1996a,b, 2005b). Therefore, staff judges that there is insufficient information in
November 2005
7-1
Draft - Do Not Quote or Cite
-------
1 the materials damage literature to inform secondary standard setting and so will not be further
2 discussed. Interested readers are referred to chapter 11 in the draft CD (EPA, 2005b). In
3 contrast, the welfare impact of Os on local, regional and global climates has received more
4 attention in recent years. Ozone enhances the heat capacity of the atmosphere. The overall body
5 of scientific evidence suggests that high concentrations of Oa on a regional scale could have a
6 discernable influence on climate, leading to surface temperature and hydrological cycle changes.
7 However, the CD states that confirming this effect will require further advances in monitoring
8 and improvement in chemical transport and regional-scale modeling. Thus, staff concludes that .
9 insufficient information is available at this time to quantitatively inform the secondary NAAQS
10 process. Though this topic will not be addressed further, its corollary, e.g., climate change
11 impacts on plant response to Os, will be considered under the discussion of factors that can
12 modify the vegetation responses to 0)3 and the implications of these interactions for future field
13 exposure conditions. Thus, this chapter will focus primarily on the well established body of
14 science regarding Os-related effects on vegetation, as discussed both in the last review and
15 summarized along with relevant new research in the current draft CD (EPA, 2005b).
16 Included in this discussion are plans for a number of analyses that update the exposure,
17 risk and benefits assessments conducted in the last review (EPA, 1996b). The EPA held a
18 consultation with the CAS AC 63 Panel on October 3, 2005 on the scope and methods being
19 considered in the planned assessments. The planned assessments described in this chapter take
20 into account the range of comments received from individual Panel members and the discussion
21 that occurred during the consultation. Staff recognizes that while these updated assessments
22 incorporate newer data, models, and approaches, and take into account current and alternative
23 ozone air quality scenarios under consideration, they are still limited by important data gaps and
24 uncertainties in currently available models and approaches. Due to the limitations of time and
25 staff resources, this first draft Staff Paper includes only descriptions of the planned assessments.
26 The second draft Staff Paper will contain results from the assessments, a more complete
27 discussion of the associated uncertainties and limitations, and staff conclusions and
28 recommendations with regard to the secondary O^ NAAQS.
29 7.2 EFFECTS ON VEGETATION
30 Science published since the conclusion of the 1996 review has not fundamentally
31 altered the understanding and conclusions regarding ozone effects on vegetation and ecosystems
32 found in the previous CD, though in some cases, recent work has expanded or clarified our
33 understanding, especially at the level of plant cell and tissue response (EPA, 2005b). In addition,
34 only a few new studies focus on addressing the data gaps or uncertainties identified in the last
35 review. Some notable exceptions within the U.S. include a shift and slight increase in research
November 2005 7-2 Draft - Do Not Quote or Cite
-------
1 designed to examine the joint interaction'of elevated COi and Os on the productivity of U.S.
2 vegetation, and the expansion of data collection and analysis of O3-induced visible foliar injuiy
3 occurrences at United States Department of Agriculture Forest Service Forest Inventory and
4 Analysis (USDA FS FIA) biomonitoring sites. In Europe, Os vegetation research has continued,
5 though with a shift in focus from the use of ambient exposure measures to modeled Os uptake by
6 vegetation in the context of developing an exposure index that can be used as a planning tool
7 (i.e., critical levels).
8 Significantly, however, a number of recent advances in the development and use of
9 mechanistic process models, leaf and canopy flux models, and improved understanding of the
10 relationships between impacts to vegetation and impacts to ecosystem structure and function, can
11 be linked to this expanded and further elucidated scientific base underpinning these
12 developments. Specifically, the expanded understanding regarding Os impacts at the genetic,
13 physiological, and mechanistic levels informs the interpretation of risk associated with
14 vegetation response at current Os levels. For example, staff is increasingly aware that Os
15 impacts at the genetic and cellular level may hold the key to understanding the more subtle but
16 equally important implications of rising C02 levels, temperature, and Os levels on plant
17 adaptability under conditions of climate change. Therefore, this section reviews the key scientific
18 conclusions identified in 1996 Os CD (EPA, 1996a), and incorporates new information from the
19 current draft CD where it expands or changes our understanding of the Os-plant interactions.
20 7.2.1 Exposure Methodologies Used in Vegetation Research
21 In the 1996 review, Os exposure studies were dominated by the use of various versions of
22 the open-top chamber (OTC), first described by Heagle et al. (1973) and Mandl et al. (1973).
23 Most OTC's consist of a cylindrical aluminum frame covered with transparent film and are
24 approximately 3m in diameter with 2.5-m-high walls. Charcoal filtered air, non-filtered air or
25 Os-supplemented air is blown through a perforated panel at the bottom of the chamber, into the
26 plant canopy and then escapes through the top of the chamber. Hogsett et al. (1987a) described
27 in detail many of the various modifications to the original OTC designs that appeared
28 subsequently, e.g., the use of larger chambers to permit exposing small trees (Kats et al., 1985)
29 and grapevines (Mandl et al., 1989). Several other modifications were made over the years to
30 improve ventilation reduce ambient air incursions, and a plastic rain-cap to exclude precipitation
31 (Hogsett et al., 1985).
32 Chambered systems, including open-top chambers, have several advantages. For
33 instance, they can provide a range of treatment levels including charcoal-filtered (CF), clean-air
34 control, and above ambient for Os experiments. Depending on experimental intent, a replicated,
35 clean-air control treatment is an essential component in many experimental designs. OTCs can
November 2005
7-3
Draft - Do Not Quote or Cite
-------
1 provide a consistent, definable exposure because of the constant wind speed and delivery
2 systems. From a policy prospective, the statistically robust C-R functions developed using such
3 systems are necessary for evaluating the implications of various alternative air quality scenarios
4 on crop response.
5 Nonetheless, there are several characteristics of the OTC design and operation that can
6 lead to unrealistic exposures. First, the plants are subjected to constant turbulence, which, by
7 lowering the boundary layer resistance to diffusion, results in increased uptake. This may lead to
8 an overestimation of the cause-effect relationships (Krupa et al., 1995; Legge et al., 1995).
9 However, in at least one case where canopy resistances were quantified in OTCs and in the field,
10 it was determined that gaseous pollutant exposure to crops in OTCs was similar to that which
11 would have occurred at the same concentration in the field (Unsworth et al., 1984a, 1984b). A
12 second concern is that the introduction of the (Venriched air into the lower part of chambers as
13 described by Heagle et al. (1973) and Mahdl et al. (1973) results in a Os concentration gradient
14 that decreases with increasing height, the converse of the situation observed in ambient air in
15 which the Os concentration decreases from above a plant canopy to ground level (Grunhage and
16 Jager, 1994; Pleijel et al., 1995,1996). Finally, as with all methods that expose vegetation to
17 modified 63 concentrations in the field, OTCs create internal environments that differ from
18 ambient air. For OTC's the so-called "chamber effect" refers to the modification of
19 microclimatic variables, including reduced and uneven light intensity, uneven rainfall, constant
20 wind speed, reduced dew formation, and increased air temperatures (Fuhrer, 1994; Manning and
21 Krupa, 1992). However, staff notes that the uncertainties associated with the influence of other
22 modifying factors occurring in the field such as water and nutrient availability are likely to be
23 greater than the uncertainties in the data due to the influence of OTCs. Because of the
24 standardized methodology and protocols used in NCLAN, the database can be assumed to be
25 internally consistent.
26 While it is clear that OTCs can alter some aspects of the microenvironment and plant
27 growth, the question to be answered is whether or not these differences affect the plant's
28 response to OB. As noted in the 1996 Os CD (EPA, 1996a), evidence from a number of
29 comparative studies of OTCs and other exposure systems suggested that, responses were, in
30 general, essentially the same regardless of exposure system used and chamber effects did not
31 significantly affect response. A recent example is the study of chamber effects examined the -
32 responses of tolerant and sensitive white clover clones (Trifoliwn repens) to ambient Oa in
33 greenhouse, open-top, and ambient plots (Heagle et al., 1996). The response found in OTCs was
34 the same as in ambient plots.
35 Though the OTC method has remained a widely used technique in the U.S. and Europe
36 for exposing plants to varying levels of Os (EPA, 2005b), in recent years, a few studies have
November 2005 7-4 Draft - Do Not Quote or Cite
-------
1 employed a modified Free Air COa Enrichment (FACE) method to expose vegetation to elevated
2 Os. This is an exposure methodology originally developed to expose vegetation without
3 chambers to elevated levels of COa. FACE has been modified in Illinois and Wisconsin to
4 include exposure of soybean and deciduous trees to Os, respectively (Dickson et al., 2000;
5 Morgan et al., 2004). The FACE method releases gas (e.g., CO2, Os) from a series of orifices
6 placed along the length of the vertical pipes surrounding a circular field plot and uses the
7 prevailing wind to distribute it. This exposure method may more closely replicate conditions in
8 the field and, more importantly for tree/forest research, has the benefit of being able to expand
9 vertically with the growth of the trees, allowing for exposure experiments to span numerous
10 years. The FACE methodology has a different set of limitations. Specifically, it is not possible
11 to produce a number of replicated treatment levels, including Os concentrations below ambient,
12 or a control where ambient Os levels are already at phytotoxic levels. Thus, FACE sites cannot
13 be used to build the statistically robust C-R functions like those produced with OTCs and is of
14 limited value in developing empirical models for predicting Os effects. In addition, FACE
15 systems are relatively expensive to operate, likely limiting the number of new sites that will
16 employ these systems and the variety of species studied.
17 Despite the differences in these two exposure methods, recent evidence obtained using
18 FACE and OTC systems appear to support the results observed in OTC studies used in the 1996
19 review. Specifically, a series of studies undertaken using free-air Os enrichment in Rhinelander,
20 WI (Isebrands et al., 2000,2001) showed that Os-symptom expression was generally similar in
21 OTCs, FACE, and ambient-Os gradient sites, supporting the previously observed variation
22 among trembling aspen clones (Populus tremuloides L.) using OTCs (Karnosky et al., 1999).
23 The FACE study also evaluated the effects of 3 years of exposure to combinations of elevated
24 COa and Os on growth responses in mixtures of five trembling aspen clones (Isebrands et al.,
25 2000,2001). Height, diameter, and stem volume (diameter2 * height) were decreased by
26 elevated Os. On average for all clones, stem volume was decreased by 20% over the 3 years in
27 the elevated Os treatment as compared with the Ix-ambient treatment. These results also appear
28 to show similar response patterns reported previously with the same clones grown in soil in pots
29 or the ground in OTCs without the alterations of microclimate induced by chambers (EPA,
30 2005b). As more FACE data become available, a more quantitative comparison of findings from
31 these two systems would be useful.
32 Staff recognizes there are other exposure methods described both in the 1996 CD and in
33 the current draft CD that have and can provide useful information on plant responses to Os
34 exposure including, chemical protectants, exclusion, passive monitors and naturally occurring
35 gradients. However, based on the considerations described above, especially the policy need for
November 2005
7-5
Draft - Do Not Quote or Cite
-------
1 robust C-R functions, staff concludes that the OTC methodology is currently the most useful in a
2 policy context.
3 7.2.2 Species (Intra-Plant) Response/Mode of Action
4 This section emphasizes reactions of Oa with the plant cell and tissue, rather than the
5 whole plant, to describe the fundamental mechanisms known to govern the response of the plant
6 to Os exposure. This section discusses the movement of Oj into plant leaves and their
7 biochemical and physiological responses to Os. In most cases, the mechanisms of response are
8 similar regardless of the degree of sensitivity of the species. The information assessed in the
9 1996 CD regarding the fundamental hypotheses concerning Os-induced changes in physiology
10 continues to be valid. However, during the last decade, this understanding of the cellular
11 processes wimin plants has been further clarified and enhanced. This section describes: (l)the
12 regulation of Oa entry into the leaf, (2) reactions of Os or its reaction products at the cell surface,
13 (3) movement and/or transformation of reaction products into the cell, (4) Os -triggered wound or
14 pathogen attack response, (5) plant defense and compensation mechanisms, (6) Os-induced
15 changes to plant metabolism, and (7) delayed expression of plant response.
16 7.2.2.1 Entry of Ozone into the Leaf
17 To cause injury, Os must first enter the plant. Ozone-induced changes to a leafs cuticle
18 (the outer layer of the leaf surface) are minimal, and Os does not penetrate the cuticle (Kerstiens
19 and Lendzian, 1989). Thus, only the 63 that diffuses into a plant through the stomata (which
20 exert some control on Os uptake) to the active sites within a leaf has the potential to impair plant
21 processes or performance. Once inside the leaf, a phytotoxic effect .will occur only if sufficient
22 amounts of Os reach sensitive cellular sites that are subject to the various physiological and
23 biochemical controls within the leaf cells. Ozone injury will not occur if (1) the rate and amount
24- of Os uptake is small enough for the plant to detoxify or metabolize Os or its metabolites or (2)
25 the plant is able to repair or compensate for the Os impacts (Tingey and Taylor, 1982; EPA,
26 1996).
27 For Os to be absorbed into a leaf, it must first reach the stomatal openings in the leaf.
28 Foliar absorption is controlled by the leaf boundary layer and stomatal conductances, which
29 together determine leaf conductance. Although the movement of pollutants through a boundary
30 layer into the stomata region is known to be important, and even rate limiting in many cases of
31 low wind velocity, its description has been defined from aeronautical concepts and usually
32 relates to smooth surfaces that are not typical of leaf-surface morphology; however, it is nearly
33 the only treatment available (Gates, 1968). Once through the boundary layer, the gas must enter
34 the leaf through the stomata.
November 2005 7 -6 Draft - Do Not Quote or Cite
-------
1 The entry or flux of gases through the stomata into a leaf is dependent upon the physical
2 and chemical processes of gas phase and surfaces and has been well-defined. In the past, the
3 internal concentration of Os has been assumed to be zero (Laisk et al., 1989), due to early studies
4 that found that virtually no Os could pass through a leaf. That was expected because Os is
5 extremely reactive with cellular biochemicals. However, a recent study by Moldau and Bichele
6 (2002) indicated that the internal Os concentration may not be zero as previous assumed.
7 However, because Os has no easily measured isotope, virtually no measurements have been done
8 on an actual dose of Os, i.e., the amount of Os which reacts with individual biochemicals inside
9 the leaf. In addition, only a limited number of studies have measured Os concentration or its •
10 reaction products within the leaf (e.g., Moldau and Bichele (2002)), and only a few instances of
11 direct measures of Os flux to foliage in the field are reported.
12 Several factors complicate estimates of flux into leaves and at the whole plant and canopy
13 scales. First, in some species, Os.modifies the opening of the stomata, usually closing it
14 partially, so that the flux rate will change (see next section for more discussion). Secondly,
15 leaves exist as part of a three dimensional canopy. Thus, the relationship of any given leaf.to the
16 ambient air is unique so that the amount of Os absorbed can vary from leaf to leaf, making it
17 difficult to extrapolate uptake from the level of an individual leaf to that of a whole plant or
18 canopy. Thirdly, Os uptake in a plant canopy is a complex process involving Os adsorption to
19 surfaces e.g., leaf cuticles, stems, and soil (termed non-stomatal deposition) and scavenging
20 reactions of Os with intra-canopy biogenic VOCs and naturally occurring NOx emissions from
21 soils, so that less Os is ultimately available for absorption into leaves.
22 Not surprisingly, as understanding of these complicating factors has grown, the issue of
23 how to characterize and define uptake or flux has received more attention. Specifically, interest
24 has been increasing in recent years, particularly in Europe, in using mathematically tractable flux
25 models for Os assessments at the regional and national scale (Emberson et al., 2000a,b).
26 Uptake or flux models have to distinguish between stomatal and non-stomatal components of Os
27 deposition to adequately estimate actual concentration reaching the interior of a leaf.
28 Determining this Os uptake via canopy and stomatal conductance by necessity relies on models
29 to predict total flux and ultimately the "effective" flux (Grunhage et al., 2004; Massman et al.,
30 2000; Massman, 2004). "Effective flux" has been defined as the balance between the Os flux and
31 the detoxification process (DSmmgen et al., 1993; Grunhage and Haenel, 1997; Musselman and
32 Massman, 1999). The time-integrated "effective flux" is termed "effective dose".
33 As described more fully below, and in the CD, scientific understanding of the detoxification
34 mechanisms is not yet complete so that the ability to model this component of flux would require
35 an intensive research effort.
November 2005
7-7
Draft - Do Not Quote or Cite
-------
1 The current state of the science with respect to flux models and their relevance to the
2 standard setting process will be revisited in greater detail in section 7,2.4, on Indices.
3 72.2.2 Reactions of 03 and possible reaction product(s) at cell surfaces
4 Ozone diffuses into the leaf air spaces and reacts either with varied biochemical
5 compounds that are exposed to the air (path 1) or is solubilized into the water lining the cell wall
6 of the air spaces'(path 2). Each reaction has the possibility of transforming Os into another
7 chemical species (a toxicant) which, in turn, may react with other chemical species and lead to a
8 cascade of reactions.
9 Within the stomata, gases react with the water at the cell's surface and generate new
10 species with the components within the cell wall region. Although these chemical reactions are
11 poorly understood, some of the fundamentals are known (Heath, 1987,1988; Wellburri, 1990).
12 Ozone reacts with organic molecules at the double bonds to form carbonyl groups and, under •
13 certain circumstances, generates peroxides, including hydrogen peroxides (^02). The role of
14 hydrogen peroxide as a signaling molecule in plants, is now better understood. One example is
15 its link to the hormone ABA-induced closure of the stomata (Pel et al., 2000). Pel etal. (2000)
16 also found that ABA induced the production of H2O2 through Reactive Oxidative Species (ROS)
17 accumulation. This complex interaction between H2O2 and ABA could help explain why Os
18 would often decrease conductance in some cases, but not always (Heath, 1994b). Other
19 chemicals present in the water phase can lead to many other oxygenated moieties. Each of the
20 steps is generally pH dependent (Jans and Hoigne, 2000; Walcek et al., 1997. Effective
21 detoxification reactions can occur here via antioxidant metabolites and enzymes if they are •
22 present at high enough concentrations (Castillo et al., 1987; Matters and Scandalios, 1987).
23 7.2.2.3 Movement of an O3 reaction produces) into the cell with enzymatic or
24 chemical transformation of those products in the cell
25 It is believed that the initial site of 63 injury is near or within the plasma membrane.
26 Ozone is soluble in water and once having entered the aqueous phase, it can be rapidly altered to
27 form oxidative products that can diffuse more readily into and through the cell and react with
28 many biochemicals. A toxic product of 03 may migrate through the cytoplast to react with
29 photosynthetic processes, or a spurious signal generated at the membrane may affect some
30 control process or signal transduction pathway (Schraudner et al., 1998; Overmyer et al., 2000,
31 2003; DeCaria et al., 2000; Rao et al., 2002; Booker et al., 2004; Leitao et al., 2003; Rao and
32 Davis, 2001; Sandermann, 2000; Vahala et al., 2003). Certainly, membrane functions, such as
33 membrane fluidity (Pauls and Thompson, 1980), permeability (Elkiey and Ormrod, 1979), K+-
34 exchange via ATPase reactions (Dominy and Heath, 1985), and Ca2+ exclusion (Castillo and
35 Heath, 1990), are changed. The initial sites of membrane reactions seem to involve transport
November 2005 7-8 Draft - Do Not Quote or Cite
-------
1 properties and, possibly, the external signal transducer molecules. It would seem that one of the
2 primary triggers of Os-induced cell responses is a change in internal Ca2+ levels (CD, 2005). The
3 ' presence of an internal antioxidant would be critical to reduce the concentration of most
4 oxidants.
5 7.2.2.4 Ozone Initiated Wounding and Pathogen Attack Response
6 The primary set of metabolic reactions that Os triggers now clearly includes those typical
7 of "wounding" responses generated by cutting of the leaf or by pathogen/insect attack. The
8 similarity of wounding responses (Langebartels et al., 1991) and Os-induced membrane
9 disruption suggests the induction of normal wound-regulated genes (Mehlhorn et al., 1991;
10 Sandermann, 1998). The sequence of the plant response to a pathogen is (1) recognition of the
11 gene products of the pathogen by the plant (elicitor), (2) generation of an immediate
12 phytoresponse to attempt to localize the attack and its products, and (3) generation of a systemic
13 acquired resistance (SAR) to subsequent attack by the pathogen. One aspect of this total response
14 is the production of O? and HbC^ by the cell (Lamb and Dixon, 1997).
15 Ozone per se does not generate the I-tOi, but rather triggers stress-related H2O2
16 formation ^similar to what occurs in a pathogen attack (ROS reaction). The presence of higher
17 than normal levels of HaO2 within the apoplastic space is a potential trigger for the normal, well-
18 studied pathogen defense pathway. SAR has been heavily investigated, and DNA probes have
19 existed for some time for a series of expressed genes. Several enzyme classes are associated with
20 Os injury, including glucanases and peroxidases and others. Thus, strong evidence exists from
21 enzyme function and genetic material that Os induces an activation of a SAR-like response.
22 Ethylene (ET) is another compound produced when plants are subj ected to biotic
23 stressors, e.g., attacks by insects, fungi, and bacteria or abiotic stressors such as wounding or
24 environmental stressors such as heat, cold, or oxidative stress and Os. Increased ET production
25 by plants exposed to Os stress was identified as a consistent marker for Os exposure decades ago
26 (Tingey et al., 1976). These studies suggested that increased production of stress-ET correlated
27 well with the degree of foliar injury that developed within hours or days after Os exposure. The
28 amount of ET released was exponentially related to the Os exposure, with peaks of high Os
29 (rather than accumulated dose) generating a higher rate of ET release, at least for a single Os
30 exposure under an acute dose. Furthermore, the amount of Os-induced ET declined with
31 repeated exposure, indicating an acclimatization to Os. This acclimatization effect associated
32 with repeated wounding has not yet been well described. Thus, one could postulate that Os
33 generates a wounding response with a production of ET, which would, in turn, generate a change
34 in stomatal conductance and photosynthesis. Clearly, these multiple events may have
35 confounded some earlier studies.
November 2005
7-9
Draft - Do Not Quote or Cite
-------
1 7.2.2.5 Defense and Compensation Mechanisms
2 The first line of defense against Os is a closure of the stomata to exclude its uptake. This
3 is counterproductive for efficient photosynthesis, but some amount of closure limits the rate of
4 Os deposition into the leaf tissue to allow for a secondary line of defense to detoxify the 03. The
5 secondary line of defense involves a range of antioxidants, including ascorbate, glutathione
6 peroxidase (GSH-Px), and sulfuroxide dimutase (SOD) which are highly reactive to the types of
7 chemicals that can be generated by Os. The timescales for changes in their levels vary: some rise
8 rapidly, while others rise more slowly. The pattern of changes in these particular proteins varies
9 greatly among different species and conditions.
10 Since 1996, the role of detoxification in providing a level of resistance to Os has been
11 further investigated. For example, most recent reports indicate that ascorbate within the cell wall
12 is the real first line of all defense. Ascorbate is water soluble, present in the solution where Os
13 can dissolve, and is highly reactive. Ascorbate concentration declines when the tissue is exposed
14 to Os (Luwe et al.,1993; Moldau, 1998), and appears to be closely linked to the amount of Os
15 penetrating the leaf tissue. Ascorbate is present within the cell wall, cytoplasm, and chloroplasts
16 (Burkey, 1999; Moldau, 1998); and can move between the cytoplasm and the cell wall with
17 relative ease (Bichele et al., 2000. It is likely that ascorbate is in higher concentration than ET,
18 and that the rate reaction of ascorbate with Os would therefore greatly dominate any -possible
19 reaction of Os with ET.
20 In spite of the new research, however, it is still not clear as to what extent detoxification
21 can protect against Os injury. Data are needed especially on the potential rates of antioxidant
22 production and on the subcellular location of the antioxidants. Potential rates of antioxidant
23 production are needed to assess whether they are sufficient to detoxify the Os as it enters the cell.
24 The subcellular location(s) is needed to assess whether the antioxidants are in cell wall or
25 plasmalemma locations that permit contact with the Os before it has a chance to damage
26 subcellular systems. Although these detoxification and compensation processes divert resources
27 away from other sinks and expend energy, they may counteract the reduction in canopy carbon
28 fixation caused by Os. The quantitative importance of these processes requires further
29 investigation.
30 7.2.2.6 Changes to Plant Metabolism
31 Ozone-related physiological effects within the leaves can 1) inhibit photosynthesis; 2)
32 alter the assimilation of photosynthate and shift its allocation patterns; and 3) can lead to reduced
33 biomass production, growth, and yield (EPA, 1986,1996a). The working hypothesis is that Os
34 which is not eliminated by antioxidants in the cell wall alters the properties of the plasma
35 membrane. Once this membrane disruption occurs, the cell must mobilize repair systems to
November 2005 7-10 Draft - Do Not Quote or Cite
-------
1 overcome the injury. Thus, carbon and energy sources once destined for productivity, must be
2 used in repair processes. Some of these repairs are thought to result from the induction of
3 specific genes.
4 Photosynthesis is inhibited by exposure to 63. A large body of literature published since
5 1996 has further elucidated the mechanism of mis effect. Pell et al., 1997 show that Oa exposure
6 results in a loss of Rubisco, the central carboxylating enzyme that plays an important role in the
7 production of carbohydrates. Due to its central importance, any loss of Rubisco may have severe
8 consequences for the plant's productivity. Several studies have found that Os had a greater effect
9 as leaves aged, with greatest impact of Os on the oldest leaves (Fiscus et al., 1997; Reid and
10 Fiscus, 1998; Noormets et al., 2001; Morgan et al., 2004). The results of these studies and others
11 suggest that alterations to the dark reactions are much more common than to light reactions
12 (Farage et al., 1991; Farage and Long, 1999).
13 The rate of leaf senescence has been shown to increase as a function of increasing Os
14 exposure. The loss of Rubisco and its messenger RNA is linked to an early senescence or a
15 speeding up of normal development leading to senescence. However, the mechanism of the
16 increased senescence is not known, and, hence, deserves further study. Most studies have shown
17 that 63 decreases allocation of photosynthate to roots. In some cases, allocation to leaf
18 production has increased. Whether these changes are driven entirely by changes in carbohydrate
19 availability or are controlled by other factors (e.g., hormones) is not known. The loss of
20 productivity is not yet clearly explained. However, several studies provide some insight into
21 possible mechanisms. A study by Grantz and Yang (2000) suggests that Os can trigger a plant-
22 wide response that may be linked to alterations in signal transduction and the generation of
23 whole plant signals. Stitt (1996) suggested that"... allocation is regulated by long-distance
24 signals that act to influence growth of selected sinks and to modify the delivery of resources to
25 these sinks in parallel." In addition, Cooley and Manning (1987), citing McLaughlin and
26 McConathy (1983), suggested three possible ways that Oa fumigation might alter translocation:
27 (1) malfunction of the phloem loading process, (2) increased translocation to leaf injury repair,
28 and (3) an altered balance between the leaf and sinks caused by reduced carbon fixation and a
29 greater demand for assimilate in the leaf. Alternatively, ethylene appears to be able to repress
30 the expression of extracellular invertase, which is critical for control and downloading of sucrose
31 derived from the translocational stream (Roitsch, 1999). More work is needed on the
32 interactions between assimilation, translocation, and source/sink relations with Os exposure. In
33 these interactions, one must be aware of the developmental age of the plants and their
34 phytohormonal status.
November 2005
7-11
Draft - Do Not Quote or Cite
-------
1 7.2.2.7 Relationships Between Age and Size and Ozone Response
2 Many changes that occur with Os exposure can be observed within hours, or perhaps
3 days, of the exposure, including those connected with wounding and elicitor-induced changes in
4 gene expression. Other effects due to 63, however, take longer to occur and tend to become
5 most obvious under long periods of low-Cb concentrations. These have been linked to '
6 senescence or some other physiological response very closely linked to senescence. The
7 understanding of how CH affects long-term growth and resistance to other biotic and abiotic
8 insults in long-lived trees is unclear. Often, the conditions to which a tree is subjected in one
9 year will affect the response of that tree in the next year. This has been called "memory effect",
10 although the term "carry-over" is preferred. In other words, a condition in an earlier year sets the
11 stage for a reaction in the next year; thereby giving a "cause-effect" scenario. In perennial plant
12 species, growth affected by a reduction in carbohydrate storage may result in the limitation of
13 growth the following year (Andersen et al., 1997). Carry-over effects have been documented in
14 the growth of tree seedlings (Hogsett et al., 1989; Sasek et al., 1991; Temple et al., 1993; U.S.
15 Environmental Protection Agency, 1996) and in roots (Andersen et al., 1991; EPA, 1996a).
16 Accumulation of carry-over effects over time will affect survival and reproduction.
17 It is important to note that while understanding of how Os interacts with the plant at a
18 cellular level has dramatically improved in recent y ears, the translation of those mechanisms into
19 how Os is involved with altered cell metabolism, with whole plant productivity, and with other
20 physiological facts remain to be more fully elucidated.
21 7.2.3 Factors That Modify Functional and Growth Response
22 The caveat that must be placed on results from any experimental study on the response of
23 living organisms to a stressor in a specific setting is that uncertainty is introduced when
24 attempting to extrapolate or apply those results outside that specific setting (e.g., to a different set
25 of organisms, scales, or exposure/growing conditions). The description of plant response to Oa
26 is no different. Because staff must necessarily rely on experimental data produced under very
27 specific sets of conditions in conducting this assessment, it is important to understand the range
28 of factors that can influence plant response to Os and the magnitude and direction of that
29 response, in order to better assess the likelihood of observing the experimentally predicted
30 response in the ambient environment.
31 The 1996 O3 CD (EPA, 1996) concluded with a statement that our understanding
32 regarding modifying factors was too fragmented to permit drawing many general conclusions.
33 Unfortunately, in the interval since the 1996 criteria document, rigorous, systematic
34 investigations of interactions have been rare, and most of the new information is as fragmented
35 as before. This is inevitable, partly in view of the vast scope of the possible interactions between
November 2005 7-12 Draft - Do Not Quote or Cite
-------
1 Os and other environmental variables and partly due to the overall lack of funding for research in
2 these areas (EPA, 2005b). Therefore, only a brief overview of the current understanding from
3 this research is provided. The reader is referred to the 1996 03 CD and the current draft CD for
4 further information.
5 Plant response to Os exposure is a function of the plant's ongoing integration of genetic,
6 biological, physical and chemical factors both within and external to the plant. The corollary is
7 also true that Os exposure can modify the plant's subsequent integrated response to other
8 environmental factors, both by influencing the plant response directly, and by contributing to
9 altered climatic factors that influence plant response through its greenhouse gas forcing
10 properties..
11 7.23.1 Genetics
12 Plant response to Os is determined by genes that are directly related to oxidant stress and
13 to an unknown number of genes that are not specifically related to oxidants but instead mat
14 control leaf and cell wall thickness, stomatal conductance, and the internal architecture of the air
15 spaces. Because the genetic code is species specific, species vary greatly in their responsiveness
16 to Os. Even within a given species, individual genotypes or populations can also vary
17 significantly with respect to Os sensitivity. Thus, caution should be taken when ranking species
18 categorically as having an absolute degree of sensitivity to Os.
19 Recent studies using molecular biological tools and with transgenic plants have begun to
20 positively verify the role of various genes and gene products in Os tolerance and are beginning to
21 increase the understanding of Os toxicity and differences in Os sensitivity. Specifically, Os has
22 been shown to trigger the production of a number of compounds (e.g. salicylic acid, ethylene and
23 jasmonic acid) and the signaling of these molecules determines in some cases the Os
24 susceptibility of plants (CD, 2005). It is unlikely that single genes are responsible to Os
25 tolerance responses, except in rare cases (Engle and Gabelman, 1966).
26 7.2.3.2 Biological Factors
27 The biological factors within the plant's environment that may directly or indirectly
28 influence its response to Os in a positive or negative manner encompass insects, other animal
29 pests, diseases, weeds, and other competing plant species. Ozone and other photochemical
30 oxidants may influence the severity of a disease or infestation by either direct effects on the
31 causal species, or indirectly by affecting the host, or both. • Likewise, mutually beneficial
32 relationships or symbioses involving higher plants and bacteria or fungi may also be affected by
33 Os. Ozone can also have indirect effects on higher herbivorous animals due to Os-induced1
34 changes in feed quality.
November 2005 7-13 Draft - Do Not Quote or Cite
-------
1 New evidence with regard to insect pests and diseases has done little to remove the
2 uncertainties noted in the 1996 CD. "Most of the large number of such interactions that may
3 affect crops, forest trees, and other natural vegetation have yet to be studied. The trend suggested
4 previously that Oa increases the likelihood and success of insect attack has received some
5 support from recent studies, but only with respect to chewing insects. With the economically
6 important group of sucking insects such as the aphids, no clear trends have been revealed by the
7 latest studies. We are still far from being able to predict ihe nature of any particular Oa plant
8 insect interaction, its likelihood, or its severity.
9 . The situation is a little clearer with respect to interactions involving facultative
10 necrotrophic plant pathogens with Oa, generally leading to increased disease. With obligate
11 biotrophic fungal, bacterial, and nematode diseases, there are twice as many reports indicating
12 Oa-induced inhibitions than enhancements. The frequent reports that infection by obligate
13 biotrophs reduces the severity of Oa-induced foliar injury should not be interpreted as
14 "protection", because of the negative effects on the host plant of the disease per se. With
15 obligate biotrophs, the nature of any interaction with Oa is probably dictated by the unique,
16 highly specific biochemical relationships between pathogen and host plant. At this time,
17 therefore, although some diseases may become more widespread or severe as a result of
18 exposure to Oa, it is still not possible to predict which diseases are likely to present the greatest
19 risks to crops and forests.
20 Several studies have indicated that the functioning of tree root symbioses with
21 mycorrhizae may be adversely affected-by Oa, but there is also evidence that the presence of
22 mycorrhizae may overcome root diseases stimulated by Oa and that Oa may encourage the spread
23 of mycorrhizae to the roots of uninfected trees. The latest studies, therefore, present no clearer
24 picture of the likely nature of simple interactions of Oa and root symbionts.
25 The few recent studies of the impact of Oa on intraspecific plant competition have again
26 confirmed that grasses frequently show greater resilience than other types of plants. In grass-
27 legume pastures, the leguminous species suffer greater growth inhibition. And the suppression of
28 Ponderosa pine seedling growth by blue wild-rye grass was markedly increased by Oa. However,
29 we are far from being able to predict the outcome of the impact of Oa on specific competitive
30 situations, such as successional plant communities or crop-weed interactions.
31 7.2.3.3 Physical Factors
32 Light, a component of the plant's physical environment, is an essential "resource" whose
33 energy content drives photosynthesis and COi assimilation. It has been suggested that increased
34 light intensity may increase the sensitivity to Oa of light-tolerant species while decreasing that of
35 shade-tolerant species, but this appears to be an oversimplification with many exceptions.
November 2005
7-14
Draft - Do Not Quote or Cite
-------
1 Temperature affects the rates of all physiological processes based on enzyme-catalysis
2 and diffusion, and each process and overall growth (the integral of all processes) has a distinct
3 optimal temperature range. Although some recent field studies have indicated that Os impact
4 significantly increases with increased ambient temperature, olher studies have revealed little
5 effect of temperature. But temperature is unquestionably an important variable affecting plant
6 response to Os in the presence of the elevated CO2 levels contributing to global climate change
7 (see below). In contrast, evidence continues to accumulate to indicate that exposure to Os
8 sensitizes plants to low temperature stress by reducing below-ground carbohydrate reserves,
9 possibly leading to responses in perennial species ranging from rapid demise to impaired growth
10 in subsequent seasons,
11 Although the relative humidity of the ambient air has generally been found to increase the
12 adverse effects of Oa by increasing stomatal conductance and thereby increasing Oa flux,
13 abundant evidence indicates that the ready availability of soil moisture results in greater
14 sensitivity to Oa. The partial "protection" against the adverse effects of Oa afforded by drought
15 has been observed in field experiments and modeled in computer simulations. There is also
16 compelling evidence that Oa can predispose plants to drought stress. Hence, the response will
17 depend to some extent upon the sequence in which the stresses occur, but, even though the nature
18 of the response is largely species-specific, successful applications of model simulations will lead
19 to larger-scale predictions of the consequences of Oa x drought interactions. However, it must be
20 recognized that regardless of the interaction., the net result on growth in the short-term is
21 negative, although in the case of tree species, other responses such as increased water use
22 efficiency could be a benefit to long-term survival.
23 Wind speed and air turbulence, affects the thickness of the boundary layers over leaves
24 and canopies and, hence, affects gas exchange rates. These factors can have a significant impact
25 on the relationship between ambient air exposures and actual exposure concentrations at the leaf
26 or canopy surface.
27 7.2.3.4 Chemical Factors
28 Mineral nutrients in the soil, other gaseous air pollutants, and agricultural chemicals
29 constitute chemical factors in the environment. The evidence regarding interactions with specific
30 nutrients is still contradictory. Some experimental evidence indicates that low general fertility
31 increases sensitivity to Oj, while simulation modeling of trees suggests that nutrient deficiency
32 and Oa act less than additively; however there are too many examples of contrary trends to
33 permit any sweeping conclusions. Somewhat analogously with temperature, it appears that any
34 shift away from the nutritional optimum may lead to greater sensitivity, but the shift would have
35 to be substantial before a significant effect on response to Oa was observed.
November 2005 7-15 Draft - Do Not Quote or Cite
-------
1 Interactions of Oa with other air pollutants have received relatively little recent attention.
2 The situation with SC>2 remains inconsistent, but seems unlikely to pose any additional risk to
3 those related to the individual pollutants. With NO and NO2, the situation is complicated by their
4 nutritional value as N sources. In leguminous species, it appears that NO2 may reduce the impact
5 of Oa on growth, with the reverse in other species, but the nature of the exposure pattern, i.e.,
6 sequential or concurrent, also determines the outcome. Much more investigation is needed before
7 we will be able to predict the outcomes of different Os-NO-NCh scenarios. The latest research
8 into Oa * acid rain interactions has confirmed that, at realistic acidities, significant interactions
9 are unlikely. A continuing lack of information precludes offering any generalizations about
10 interactive effects of Oa with NHa, HF, of heavy metals. More evidence has been reported that
11 the application of fungicides affords some protective effects against Oa.
12 Over the last decade, considerable emphasis has been placed on research into Oa
13 interactions with two components of global climate change: increased atmospheric COa and
14 increased mean global temperature. Most of these studies, however, have tended to regard
15 increased CO2 levels and increased mean temperatures as unrelated phenomena, in spite of the
16 crucial role of temperature as a climatic determinant (Monteith and Elston, 1993). Thus,
17 experiments that examime the effects of doubled COa levels at today's mean ambient
18 temperatures are not particularly helpful in trying to assess the impact of climate change on
19 responses to Oa, since most of the biotic and chemical interactions with oxidants (as discussed in
20 7.2.3 above) may be modified by these climatic changes. Though it is now known from limited
21 experimental evidence and evidence obtained by computer simulation that an atmosphere
22 sufficiently enriched with CO2 (e.g., 600 + ppm) would more than offset the impact of Oa on
23 responses as varied as wheat yield or the growth of young Ponderosa pine trees, the concurrent
24 increase in temperature would reduce, but probably not eliminate, the net gain.
25 Little if any experimental evidence exists related to three-way interactions, such as Oa *
26 CO2 x disease or Oa * CO2 x nutrient availability, although such interactions cannot be predicted
27 from the component two-way interactions. Increased use of computer simulations may be
28 important in suggesting outcomes of the many complex interactions of Oa and various
29 combinations of environmental factors. However, the results obtained will only be as reliable as
30 the input data used for their parameterization. Thus, additional data from organized, systematic
31 study is needed.
32 It is important to recognize that wide variations in net impacts of climate change in
33 different geographic areas are expected. Although many regions are predicted to experience
34 severe, possibly irreversible, adverse effects due to climate change, beneficial changes may also
35 take place. Findings from the U.S. Global Change Research Program (USGCRP) (NAST, 2000)
36 and related reports illustrate the considerable uncertainties and difficulties in projecting likely
November 2005 7-16 Draft - Do Not Quote or Cite
-------
1 climate change impacts at the regional or local scale. The USGCRP findings also reflect the
2 mixed nature of projected potential climate change impacts, i.e., combinations of deleterious and
3 beneficial effects, for U.S. regions and the variation of projected impacts across different
4 regions. The EPA is currently leading a research effort that uses regional-scale climate models
5 with the goal of identifying changes to OB and PM concentrations that may occur in a warming
6 climate. An assessment of the results of this effort is expected to be available for consideration
7 in the next review of the OsNAAQS.
8 7.2.4 Effects-Based Air Quality Exposure Indices
9 The language in sections 108 and 109 of the Clean Air Act indicates that the.secondary
10 NAAQS is to specify a level of ambient air quality that when met, will protect the public welfare
11 from any known or anticipated adverse effects associated with the presence of such air pollutant
12 in the ambient air. Since those words were written, Ihe vegetation and ecosystem science has
13 evolved to demonstrate that the presence of a pollutant (Oa) in ambient air is only one of a
14 multitude of factors influencing the likelihood of an adverse vegetation and/or ecosystem effect
15 occurring as a result of exposure to a pollutant. However, since most of the other factors that
16 play a role in regulating the potential impact of an air pollutant on vegetation are outside of
17 human control, except in controlled experiments or heavily managed agricultural settings, it
18 seems reasonable to continue to focus on the potential contribution of anthropogenically derived
19 ambient air concentrations of Os in producing adverse effects to vegetation, recognizing that it is
20 not possible to predict for all plants occurring in the U.S. at any given time and ambient pollutant
21 concentration, which and to what degree modifying factors are influencing either the rate of that
22 pollutant uptake from the ambient air or the plant's response to that uptake. Thus, any ambient
23 air quality exposure index will by necessity be a simplification of the actual relationship between
24 pollutant concentrations in the ambient air and plant response. That said, there may be ways to
25 improve upon or more carefully focus the application of existing air quality exposure indices to
26 improve their predictive power.
27 Most of what is known about vegetation response to Os is a result of controlled
28 experiments that sought to minimize the influence of other confounding variables so that a clear
29 Os signal could be measured. Experimental exposure profiles were typified by the episodic
30 occurrence of a large number of higher Os concentrations. Though not atypical, growth or yield
31 effects may be over- or underestimated in regions of the country where a different type of
32 temporal pattern is prevalent. Therefore, it should be recognized that the conclusions drawn
33 about the importance of different exposure features are heavily influenced by the nature of the
34 experiments conducted.
November 2005
7-17
Draft - Do Not Quote or Cite
-------
1 In the last review, the aspects of exposure known to affect plant response included a)
2 exposure duration (i.e., Oa effects are cumulative); b) higher concentrations appear to be more
3 important than lower; c) plant sensitivity to Os varies with time of day and crop development
4 stage. Exposure indices that accumulate the hourly Os concentrations and preferentially weight
5 the higher concentrations had better statistical fits to growth/yield response than did the mean
6 and peak indices. No experiments were conducted to test the performance of these indices in the
7 field. Instead, the testing of adequacy of available indices was accomplished through regression
8 analyses of earlier studies. Therefore, indices selected for further consideration based on the
9 regression analysis were those mat could best quantify growth and yield effects in crops,
10 perennials and trees (primarily seedlings) and were cumulative and peak weighted (e.g., SUM06,
11 Wl 26, and AOT40). These indices are also known as "concentration based" as they only
12 consider ambient concentrations in deriving the value of the index.
13 Other issues raised during the last review regarding the most relevant aspects of exposure
14 for inclusion in an air quality exposure index included the question of the relative importance of
15 cumulative peak (>0.10 ppm) versus mid-range (0.05-0.099 ppm) concentrations, given the
16 concern mat higher concentrations do not always occur at the time of maximum plant uptake in
17 the field. This coincidence was considered to be the critical factor in determining peak
18 concentration impacts on plants. Based on evidence at that time, it was not possible to resolve
19 this issue and no experimental studies had addressed this question. A multicomponent index was
20 suggested that combined a concentration-weighted, cumulative index with the number of
21 occurrences of hourly averaged concentrations >=0.10 ppm but no direct experimental studies
22 have been conducted to address the usefulness of this approach in reducing uncertainty. Another
23 element considered was the appropriate diurnal window (e.g., 7,12 or 24 hours) over which to
24 cumulate exposures. At that time, staff concluded that the 12 hour, daylight period Was the most
25 appropriate, and widely applicable based on the information available at that time that the
26 majority of plants, although not all, have significantly reduced stomatal conductance at night, so
27 that the potential for significant impacts from nighttime Os exposures was considered low.
28 7.2.4.1 Concentration-based Forms
29 Concentration-based air quality exposure indices focus on a particular feature or features
30 (e.g., concentration, duration, and exposure patterns) of the ambient pollutant profile that has
31 been experimentally shown to be predictive of plant response. A few recent studies have focused
32 on the role of these different components of Oa exposure and have substantiated the earlier
33 conclusions in the 1996 CD and Staff Paper of the importance of higher concentration, shape of
34 the peak, and the episodicity of peak occurrence in eliciting the plant response to Os exposure. A
35 few studies have also further clarified the role of nocturnal conductance. Because it has
November 2005
7-18
Draft - Do Not Quote or Cite
-------
1 implications for selecting the appropriate diurnal timeframe over which to cumulate exposures,
2 this latter topic is reviewed in more detail below.
3 Musselman and Minnick (2000) performed an extensive review of the literature and
4 reported that a large number of species had varying degrees of nocturnal stomatal conductance
5 (Musselman and Minnick, 2000). Although stomatal conductance was lower at night than
6 during the day for most plants, nocturnal conductance could result in some measurable 63 flux
7 into the plants. In addition, plants might be more susceptible to 63 exposure at night than during
8 the daytime, because of possibly lower plarit defenses at night (Musselman and Minnick, 2000).
9 Nocturnal O? flux also depends on the level of turbulence that intermittently occurs at night.
10 Based on their review, Musselman and Minnick (2000) recommended that any O$ exposure
11 index used to relate air quality to plant response should use the 24-h cumulative exposure period
12 for both exposure-response and effective flux models.
13 The role of nighttime stomatal conductance and Oa exposure was demonstrated
14 experimentally as well. Grulke et al. (2003) showed that the stomatal conductance at night for
15 Ponderosa pine in the San Bernardino NF (CA) ranged from 10% to 25% that of maximum
16 daytime gas exchange. In June, at tile high-elevation site, 11% of the total daily Os uptake of
17 pole-sized trees occurred at night. In late summer, however, Os uptake at night was negligible.
18 Birch seedlings exposed to OB at night show greater reductions in growth than those exposed to
19 Os in daylight (Matyssek et al., 1995). Massman (2004) suggested that nocturnal stomatal Oa
20 uptake accounted for about 15% of the cumulative daily effective 63 dose that was related to
21 predicted injury.
22 A number of findings, however, confound the generalization of the importance of
23 nocturnal exposures. For example, field mustard (Brassica rapa L.) plants exposed to Os during
24 the day or night showed little significant difference in the amounts of injury or reduced growth
25 response to Os treatment, and the stomatal conductance was 70 to 80% lower at night (Winner .et
26 al., 1989). Tissue biomass of Ponderosa pine seedlings was significantly reduced when seedlings
27 were exposed to either daytime or nighttime episodic profiles (Lee and Hogsett, 1999) and the
28 biomass reductions were much greater with daytime peak concentrations than with nighttime
29 peak concentrations. In an evaluation of a very large number of indices that described the Os
30 impact on spring wheat, Finnan et al. (1997) did not find any improvement in performance of the
31 cumulative concentration-weighted indices by weighting those concentrations occurring during
32 sunlight hours. Thus, it would appear that the importance of nocturnal conductance is species
33 specific.
November 2005
7-19
Draft - Do Not Quote or Cite
-------
1 7.2.4.2 Flux-based forms
2 A large number of recent studies have focused on the development of a flux-based index
3 to better relate ambient Os to observed vegetation effects. Though a few such studies were
4 published prior to 1996 and reviewed in the 1996 O3 CD (U.S. Environmental Protection
5 Agency, 1996b), a large body of recent literature has further highlighted and elucidated the
6 multiple controlling factors and complexities (outlined below) that are associated with linking
7 ambient ozone concentrations to observed plant response. Specific factors that influence
8 stomatal uptake and subsequent plant response to OB include:
9
10 (1) The potential for maximum flux of Os into a leaf depends on synchronicity between
11 the timing of peak exposure events and maximal stomatal conductance. In cases where there is
12 disconnect between these two diunal patterns, the predicted effect of the exposure for that
13 species/individual is an overestimation. This concern is especially apparent when assessing the
14 impact of Os across all the varied climatic regions and species occurring within the U.S..
15
16 (2) Multiple meteorological, species-and site-specific factors influence Os uptake. In
17 order to integrate those factors that drive the patterns of stomatal conductance and exposure,
18 some studies use stomatal (Ashmore et al., 2004a) or physiological process-based models
19 (Laurence et al., 2001). However, the species- and site-specific scope of these models limits
20 their usefulness in national or regional scale risk assessments.
21 ;•
22 (3) Not all 63 stomatal uptake results in a reduction in yield. This nonlinear relationship
23 between Os uptake and plant injury (not growth alteration) response depends to some degree on
24 the amount of internal detoxification occurring with each particular species; species having high
25 amounts of detoxification potential may show less of a relationship between 63 stomatal uptake
26 and plant response. Because detoxification potential is genetically determined, it cannot be
27 generalized across species. Much more needs to be learned about the detoxification processes
28 available to plants and to what extent they modify the potentially phytotoxic dose in the leaf
29 interior before this factor can be meaningfully considered in a biologically-relevant index.
30
31 It is anticipated that, as the overlapping mathematical relationships of conductance,.
32 concentration, and defense mechanisms are better defined, Os-flux-based models may be able to
33 predict vegetation injury and/or damage at least for some categories of canopy-types with more
34 accuracy than the currently available exposure-response models. The results of these studies and
35 reviews indicate the need to continue to develop indices that are more physiologically and
36 meteorologically connected to the actual dose of Oa the plant receives. The flux approach should
November 2005 7-20 Draft - Do Not Quote or Cite
-------
1 provide an opportunity to improve upon the concentration-based exposure index in the future,
'2 recognizing that a concerted research effort is needed to develop the necessary experimental data
3 and modeling tools that will provide the scientific basis for such critical levels for Og (Dammgen
4 et al., 1994; Fuhrer et al, 1997; Grunhage et al., 2004).
5 7:2.4.3 The Critical Level Approach
6 Both the concentration-based and flux-based exposure index forms can be used to
7 establish "a "critical level" for plant exposure to 63. One definition of a critical level is "the
8 concentration of pollutant in the atmosphere above which direct adverse effects on receptors,
9 such as plants, ecosystems, or materials may occur according to present knowledge" (UNECE,
10 1988). As used by the United Nations Economic Commission for Europe International
11 Cooperative Programme (UNECE ICP), the critical levels are not air quality regulatory standards
12 in the U.S. 'sense, but rather planning targets for reductions in pollutant emissions to protect
13 ecological resources. Critical levels for Os are intended to prevent long-term deleterious effects
14 on the most sensitive plant species under the most sensitive environmental conditions, but not to
15 quantify Os effects. The nature of the "adverse effects" was not specified in the original
16 definition, which provided for different levels for different types of harmful effect (e.g., visible
17 injury or loss of crop yield). There are also different levels for crops, forests, and seminatural
18 vegetation. The caveat, "according to present knowledge," is important because critical levels
19 are not rigid; they are revised periodically as new scientific information becomes available. To
20 date, critical levels (Level I) have been set for agricultural crops, for foliar injury symptoms in
21 the field and for forest trees (see section 7.2.5 and EPA, 2005b). Level I critical levels are
22 currently used to map and identify areas in Europe in which the levels are exceeded, and that
23 information is then used to plan optimized and effects-based abatement strategies.
24 In the 1990s, however, many exposure studies demonstrated that the simple, exposure-
25 based approach led to the overestimation of effects in some regions and underestimation in
26 others (Fuhrer et al., 1997; Karenlampi and Skarby, 1996) because it did not differentiate
27 between plant species, and it did not include modifying site and micrometeorological factors of
28 Os uptake such vapor pressure deficit (VPD), water stress, temperature, and light and variation in
29 canopy height. At that time, a decision was made by the UNECE ICP to work towards a flux- •
30 based approach for the critical levels ("Level II"), with the goal of modeling Os flux-effect
31 relationships for three vegetation types: crops, forests, and seminatural vegetation (Grunhage and
32 Jager, 2003). Progress has been made in modeling flux (Ashmore et al., 2004a,b) and the
33 Mapping Manual is being revised (Ashmore et al., 2004a,b; Grennfelt, 2004; Karlsson et al.,
34 2003). The revisions may include a flux-based approach for three crops: wheat, potatoes, and
35 cotton. However, because of a lack of flux-response data, a cumulative, cutoff concentration-
November 2005
7-21
Draft - Do Not Quote or Cite
-------
1 based (AOTx) exposure index will remain in use for the near future for most crops and for
2 forests and seminatural herbaceous vegetation (Ashmore et al., 2004a).
3 7.2.4.4 Summary
4 From the above discussion, several cautionary statements emerge that must be kept in
5 mind when considering the most appropriate and useful concentration- or flux-based forms for
6 characterizing the air quality that is associated with adverse vegetation effects (e.g., when
• 7 defining a critical level). First, current understanding of the important components of exposure
8 in eliciting plant response are based on exposure regimes that favored Os uptake, contained large
9 numbers of peak concentrations, and closely controlled other environmental factors. In the
10 absence of further study, it is unclear how well indices selected on this basis perform under
11 different exposure and growth scenarios. Second, flux-based models are currently limited by the
12 species-specific information required and by the observed nonlinearity between total flux and
13 plant response. Better understanding of the detoxification and compensation processes would be
14 required to account for this nonlinearity in future models. In some cases, Os exposure has been
15 shown to explain 63 effects as well or better than calculated internal Os dose (Grulke, et al.
16 2002; Hanson et al., 1994).
17 Other relevant information that should be evaluated include the extent to which: (1)
18 nighttime exposures represent a significant percentage of total diurnal exposures, and whether
19 their impact on growth or foliar injury effects are proportional; (2) elevation and nocturnal
20 turbulence effects may alter actual nocturnal uptake; and (3) plant defense mechanisms and other
21 processes may differ at night, leading to either more or less susceptibility than that associated
22 with daytime exposures. Staff will take into account the expanded evidence on the importance of
23 nocturnal conductance/exposures that has become available since the last review, along with the
24 associated caveats, in its consideration of appropriate averaging time windows.
25 Until such research can be done, the current draft CD concludes that, at this time, based
26 on the current state of knowledge, exposure indices that differentially weight the higher hourly
27 average Os concentrations but include the mid-level values still represent the best approach for
28 relating vegetation effects to Os exposure in the U. S.. This is due in part to the existence of a
29 large database that has been used for establishing exposure-response relationships. Such a
30 database does not yet exist for relating Oj flux to growth response. The draft CD further
31 concludes that the disconnects between period of maximum uptake and peak 63 occurrence,, as
32 well as the potential for nocturnal uptake, should be considered by adding some weighting
33 functions into the currently used exposure indices. Of particular consideration would be their
34 inclusion in regional-to-national estimations of 03 impacts on vegetation. In evaluating the
35 information now available and described in the current CD, staff will consider whether and how .
November 2005
7-22
Draft - Do Not Quote or Cite
-------
1 additional flux/uptake-related factors could be combined with, existing cumulative, peak
2 weighted indices in order to develop an air quality index that is a better surrogate predictor of
3 vegetation risk.
4
5
6
7
8
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
7.2.5 * Ozone Exposure-Plant Response Relationships
Much of what is known about Os exposure-plant response relationships, as summarized
below, is based.on research that was available in the last review. Thus, the present discussion is
largely based on the conclusions of the 1978,1986, and 1996 CDs EPA, 1978,1986,1996a).
These earlier conclusions were derived from basically two types of studies: (1) studies that
developed predictive equations relating Osexposure to plant response, and (2) studies that
compared the effects of discrete treatment level(s) to a control. The advantage of the regression
approach is that exposure-response models can be used to interpolate results between treatment
levels. During the 1980s, the most commonly used indices for expressing Oa exposure were 7-,
8-, or 12-h daytime average values over the duration of Os exposure, which was often 3 months
or somewhat less for experimental studies with crops. Studies into the 1990's also began to use
cumulative, peak weighted forms.
(1) Ozone can cause a range of effects, beginning with individual cells, leaves, and
plants, and proceeding to plant populations and communities. These effects may be classified as
either "injury" or "damage". Injury encompasses all plant reactions, such as reversible changes
in plant metabolism (e.g., altered photosynthetic rate), altered plant quality, or reduced growth
that does not impair yield or the intended use or value of the plant (Guderian, 1977). In contrast,
damage includes all effects that reduce or impair the intended use or value of the plant. Damage
includes reductions in aesthetic values (e.g., foliar injury in ornamental species) as well as losses
in terms of weight, number, or size of the plant part that is harvested (yield loss). Yield loss also
may include changes in crop quality, i.e., physical appearance, chemical composition, or the
ability to withstand storage.
(2) Research results since 1978 did not invalidate EPA conclusions (EPA, 1978,1986)
that foliar symptoms due to Os exposures reduce the market value of certain crops and
ornamentals where leaves are the product (such as spinach, petunia, geranium, and poinsettia)
and that such damage occurs at ambient Os concentrations observed in the U.S.. In addition, the
results of OTC studies that compared yields at ambient Os exposures with those in filtered air
and retrospective analyses of crop data also established that ambient Os concentrations were
sufficient to reduce the yield of maj or crops in the U. S:.' .
November 2005
7-23
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
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) A 3-month SUM06 exposure of 24.4 ppm h, corresponding to a 7-h mean of 49 ppb
and a 2HDM of 94 ppb O3 may prevent a 10% loss in 50% of the 49 experimental cases analyzed
by Tingey et al. (1991). A 12-h growing season mean of 0.045 ppb should restrict yield losses to
10% in major crop species (Lesser et al., 1990).
(4) Depending on duration, concentrations of Oa and SUM06 exposures currently in the
United States are sufficient to affect the growth of a number of tree species. Given the fact that
multiple-year exposures may cause a cumulative effect on the growth of some trees (Simini et
al., 1992; Temple et al., 1992), it is likely that a number of species currently are being impacted,
even at ambient Os exposures.
(5) Exposure-response functions for 51 cases of seedling response to O3 (Hogsett et al.,
1995), including 11 species representing deciduous and evergreen growth habits, suggest that a
SUM06 exposure for 5 months of 31.5 ppm h would protect hardwoods from a 10% growth loss
in 50% of the cases studied. A SUM06 exposure of 42.6 ppm h should provide the same level of
protection for evergreen seedlings. These conclusions do not take into the account the possibility
of effects on growth in subsequent years, an important consideration in the case of long-lived
species.
(6) Studies of the response of trees to Os have established that, in some cases (for
instance, poplars and black cherry), trees are as sensitive to Oa as are annual plants, in spite of
the fact that trees are longer-lived and generally have lower gas exchange rates, and, therefore,
lower Os uptake.
(7) Use of the chemical protectant, EDU, is of value in estimating Oa-related losses in
crop yield and tree growth, provided that care is exercised in establishing appropriate EDU
dosages to protect the plants without affecting growth.
Studies conducted since the conclusion of the last review have not fundamentally altered
the conclusions stated above. Unfortunately, no single exposure index has been used
consistently in the recent literature. Of the cumulative indices that preferentially weight higher
concentrations, the SUM06 index has been used most commonly in the U.S. literature, while the
use of the AOT40 index has become quite common in Europe. This lack of consistency makes it
difficult to compare experimentally derived exposure-response data expressed as AOT40 to
ambient U.S. Os exposures. The paragraphs below touch on areas where new research has
confirmed or expanded the knowledge base that existed in the last review.
November 2005
7-24
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
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) Much of the research on the effects of Os on growth, biomass, and yield effects in
annual and biennial species during the last decade has been conducted in Europe. Substantial
effort has gone into developing a Level-I critical level for crops. Based on regression analysis of
15 OTC studies of spring wheat, an AOT40 value of 2.8 ppm-h corresponded to a 5% yield loss
(Fuhrer et al., 1997). Because a 4 to 5% decrease could be detected with a 99% confidence level,
a critical level of an AOT40 value of 3 ppm-h was selected in 1996 (Karenlampi and SkSrby,
1996).
(2) Several studies during recent decades have further demonstrated Os effects on
different stages of reproduction. Effects of Os have been observed on pollen germination, pollen
tube growth, fertilization, and abortion of reproductive structures, as reviewed by Black et al.
(2000). Reproductive effects will culminate for seed-bearing plants in seed production. The
recent scientific literature supports the conclusions of the 1996 CD that ambient Cb
concentrations are reducing the yield of major crops in the U. S. and that there may be
economically important effects of ambient Ch on the quality of crop and forage species. For
example, the yield reductions for soybean are generally similar to those reported previously
(EPA, 2005b). Ozone may also reduce the quality or nutritive value of annual species. Foliar
symptoms are important if they reduce the marketability of the crop. In Europe, Level I critical
levels have been determined for such effects.
(3) During the past 10 years, much of the research in the U.S. has focused on perennial
species, including forage crops. Recent results confirm that yields and quality of multiple-year
forage crops are reduced at sufficient magnitude to have nutritional and possibly economic
implications to their use for ruminant animal feed at Oa exposures that occur in some years over
large areas of the U.S.. When species are grown in mixtures, Qa exposure can increase the
growth of O3-tolerant species while exacerbating the growth decrease of Ch-sensitive species. In
Europe, a provisional critical level for perennials of an AOT40 value of 7 ppm-h over 6 months
has been proposed to protect sensitive plant species from the adverse effects of Os.
(4) A few investigations reported since the last review have examined saplings or mature
trees, notably of oak species in the southern Appalachian Mountains and pine species in
California. Most of these studies have been of natural (uncontrolled) Os exposures. Additional
studies have examined foliar symptoms on mature trees, and in recent years such surveys have
become more common and with greater attention to the standardization of methods and the use
of reliable indicator species (Campbell et al., 2000; Smith et al., 2003). Previous CDs have
November 2005
7-25
Draft - Do Not Quote or Cite
-------
1 noted the difficulty in relating foliar symptoms to effects on individual tree growth, stand
2 growth, or ecosystem characteristics (EPA, 1996a). This difficulty remains to the present day.
3
4 (5) Some investigators have suggested that a comprehensive risk assessment of the
5 effects of Os on mature tree species might best be accomplished by extrapolating measured
6 effects of Oaon seedlings to effects on forests using models based on tree physiology and forest
7 stand dynamics (Chappelka and Samuelson, 1998; Laurence et al.3 2000, 2001).
8
9 (6) An ongoing study was undertaken using a FACE carbon dioxide and Oa enrichment
10 facility in Rhinelander, WI (Isebrands et al., 2000, 2001). These studies showed that Ch-
11 symptom expression was generally similar in OTCs, FACE, and gradient sites, supporting the
12 previously observed variation among aspen clones (Karnosky et al., 1999).
13
14 (7) Many studies have demonstrated that root growth is more sensitive to Cbexposure
15 than is stem growth. In Finland, reduced root growth was found for a number of clones of silver.
16 birch (Oksanen and Saleem, 1999). After 5 years, root growth was decreased by 33%, but shoot
17 growth was not affected by Cb exposures of a 7-h mean of 15 ppm-h over 5 years in a FACE
18 system (Oksanen, 2001). Data from a long-studied pollution gradient in the San Bernardino
19 Mountains of southern California suggests that Cb substantially reduces root growth in natural
20 stands of ponderosa pine. Root growth in mature trees was decreased at least 87% in a high
21 pollution site as compared to a low pollution site (Grulke et al., 1998), and a similar pattern was
22 found in a separate study with whole tree harvest along this gradient (Grulke and Balduman,
23 1999). In contrast, a study in Great Smoky Mountains National Park in Tennessee (Neufeld et
24 al., 2000) found no statistically significant effects of Cb exposure on stem or root biomass for
25 several tree species.
26
27 (8) European beech was selected for development of a Level I critical level, because data
28 from several studies were available for this species and because deciduous tree species were
29 judged to be more sensitive to Cb compared to evergreen tree species (Fuhrer et al., 1997;
30 Karenlampi and Skarby, 1996). A critical level was defined as an AOT40 value of 10 ppmh for
31 daylight hours for a 6-month growing season (Karenlampi and Skarby, 1996). However, other
32 studies have shown that other species such as silver birch may be more sensitive to Cb than beech
33 (Paakkonen et al., 1996).
34
x
35 (9) Recent results support the conclusions of the 1996 CD (EPA, 1996) that individual
36 deciduous trees are generally less sensitive to Cb than are most annual plants, with the exception
November 2005 7-26 Draft - Do Not Quote or Cite
-------
1 of a few. genera such as Populus, which are highly sensitive. However, the data suggest that
2 ambient exposures that occur in different regions of the U.S. can sometimes reduce the growth of
3 seedlings of deciduous species. The (^sensitivity of seedlings and mature trees within species
4 and between species varies widely. In general, mature deciduous trees are likely to be more
5 sensitive to Os compared to seedlings, while mature evergreen trees are likely to be less sensitive
6 than seedlings. Based on these results, stomatal conductance, Chuptake, and Os effects cannot be
7 assumed to be equivalent in seedlings and mature trees.
8 73
VEGETATION IMPACT ASSESSMENT
9 7.3.1 Overview
10 The planned assessments for the current review are outlined in Figure 7.1(a-c). These
11 assessments build upon the 1996 review and ozone response relationships contained in that
12 review. This section is organized based on each of the main components of the assessment.
13 First, the air quality data, modeling and interpolation analyses that will be input into the rest of
14 the vegetation assessment are discussed. A description then follows of the four areas in which
15 Os exposures and impacts will be assessed. The vegetation exposure and risk discussion is
16 divided between the crop and tree analyses. The crop part of the analysis will focus on the
17 estimated effect of current Os on crop yield and how this, in turn, affects modeled economic
18 parameters for the agriculture sector. The tree analyses have three parts: (1) seedling growth
19 will be updated to estimate growth loss under estimated 2001 Os conditions; 2.) foliar injury will
20 be linked to current monitored and estimated air quality levels; and (3) the TREGRO model will
21 be used to model ponderosa pine growth under recent air quality data and under scenarios where
22 alternative standards are just met. The plans discussed in the following sections take into
23 account the range of views that were expressed during the recent consultation with the CASAC
24 Os Panel. Differing views were expressed in some areas by Panel members and the description
25 of the assessments has addressed the comments to the extent possible. Generally the comments
26 were concerned with better characterization of uncertainties in the 63 exposure interpolation for
27 2001, Os concentration measurements at the height of monitor inlets relative to the actual Os
28 concentration at the vegetation height, the continued relevance of C-R functions from NCLAN
29 studies and their incorporation into the agricultural model used in the crop analysis. This first
30 draft Staff Paper acknowledges and seeks to identify all the potentially relevant sources of
31 uncertainty that will be more fully quantitatively and qualitatively characterized in the second
32 draft Staff Paper.
November 2005
7-27
Draft - Do Not Quote or Cite
-------
1 Figure 7. l(a-c). Major Components of Planned Environmental Assessment
2
November 2005
7-28
- Do Not Quote or Cite
-------
1 73.2 Air Quality Analysis
2 To accomplish an assessment of the effects of ambient Os exposures on vegetation and
3 ecosystems, it is important to characterize recent Os air quality over broad geographical areas of
4 concern. This presents a great challenge since vast rural areas of the U.S., where important crops
5 and natural vegetation occur, do not have Os monitors. Thus, staff is currently evaluating a
6 variety of approaches for estimating potential Os exposures in these non-monitored areas. Three
7 main approaches available for characterizing ozone exposures in non-monitored areas are: (1)
8 use of sophisticated models of air quality to model the entire U. S.; (2) use of data collected from
9 monitors to interpolate non-monitored areas; and (3) combining monitored and modeled
10 information into an interpolated surface. The section below describes the main approach staff
11 plans to use to combine monitored observations and modeled 63 predictions from CMAQ to.
12 estimate Oa exposures in many areas as possible. Staff also intends to evaluate a whether
13 interpolation without modeled data would be adequate in areas with relatively, dense monitoring
14 coverage. As suggested during the CASAC consultation, staff will also interpolate across
15 smaller, more homogeneous regions rather than across the entire contiguous U.S. The staff also
16 plans to rely on monitoring data where possible (see sections 7.2.4)
17 73.2.1 Monitor coverage
18 EPA monitored data is currently available through 2004. The monitored hourly Os data
19 is available from two networks: (1) Air Quality System (AQS;
20 http://www.epa.gov7ttn/airs/airsaqs) and (2) Clean Air Status and Trends Network (CASTNET;
21 http://www.epa.gov/castnetA. The locations of these monitors are presented in Figure 7.2a-b and
22 are described in section 2.3.1 and 2.3.2 of Chapter 2. The AQS monitoring network currently
23 has over 3000 monitors measuring Os concentrations and monitors are generally sited near
24 population centers. However, there are approximately 36 monitored located in National Parks.
25 CASTNET is the nation's primary source for data on dry acidic deposition and rural, ground-
26 level ozone. It consists of over 80 sites across the eastern and western U.S. (see Figure 7-2b).
27 During the recent CASAC consultation, the question was raised about how 63
28 concentrations measured by the monitoring network compare to concentrations occurring at the
29 surface of vegetation. Inlets to ambient monitors are typically at heights of 3 to 5 meters, and
30 thus are located in the inner part of the planetary boundary layer (EPA, 2005b). It is well known
31 that within this layer 03 reacts with vegetation and other surfaces on the ground and can create
32 vertical gradient of decreasing Os concentration from the inlet height of the monitors to the
33 surface of vegetation and crops (Regener, 1957). The magnitude of the gradient is determined
34 by the intensity of turbulent mixing in the surface layer. During the daytime hours the vertical
35 Os gradient is relatively small because turbulent mixing maintains the downward flux of Oj. For
November 2005
7-29
Draft - Do Not Quote or Cite
-------
1 example, Hovath et al. (1995) calculated a 7% decrease in Os going from a height of 4 meters
2 down to 0.5 meters above the surface during unstable (or turbulent) conditions in a study over
3 low vegetation in Germany. This is compared to a 20% decrease during stable conditions which
4 usually occur during the night. While staff acknowledges there is likely to be bias in using Oa
5 data from inlets of ambient monitors to characterize Os exposures close to the surface, this bias is
6 likely to be relatively small under typical turbulent mixing during the day. The day-time versus
7 the night-time bias is an important distinction considering the assessments outlined below will
8 rely heavily on daytime metrics such as the!2hr SUM06.
9 7.3.2.2 Modeling tools
10 Staff plans to use 03 outputs from the EPA/NOAA CMAQ model system
11 (http://www.epa.gov/asmdnerl/CMAO. Byun and Ching, 1999; Arnold et al. 2003; Hogrefe et al.
12 2004;,Eder and Yu, 2005) to spatially scale an interpolation of 63 monitoring data for different
13 regions of the U.S.. The CMAQ modeling system has undergone two external peer reviews
14 through the Community Modeling and Analysis System (CMAS) based at the University of
15 North Carolina at Chapel Hill (UNO Carolina Environmental Program (Amar et al. 2005,2004).
16 The CMAQ model is a multi-pollutant, multiscale air quality model that contains state-of-science
17 techniques for simulating all atmospheric and land processes that affect the transport,
18 transformation, and deposition of atmospheric pollutants and/or their precursors on both regional
19 and urban scales. It is designed as a science-based modeling tool for handling many major
20 pollutants (including photochemical oxidants/Os, PM, and nutrient deposition) holistically. The
21 CMAQ model incorporates output fields from emissions and meteorological modeling systems
22 and several other data sources through special interface processors into the CMAQ Chemical
23 Transport Model (CCTM). Currently, the Sparse Matrix Operator Kernel Emissions (SMOKE)
24 System produces the emissions factors and the Fifth Generation Perm State University/ National
25 Center for Atmospheric Research Mesoscale Model (MM5) provides the meteorological fields.
26 CCTM then performs chemical transport modeling for multiple pollutants on multiple scales.
27 The CMAQ model can generate estimates of hourly Os concentrations for the contiguous
28 U.S., making it possible to express model outputs in terms of a variety of exposure indices (e.g.,
29 SUM06, 8-hr average). Due to the significant resources required to run CMAQ, however, model
30 outputs are currently only available for limited years. Currently, 2001 outputs from CMAQ
31 version 4.5 are the most recent data available. This version of CMAQ utilizes the more refined
32 12 km x 12 km grid for the eastern U.S., while using the 36 km x 36 km grid for the western U.S.
33 The 12 km x 12 km domain covers an area from roughly central Texas, north to North Dakota,
34 east to Maine, and south to central Florida Emission inventories of SO2, CO, NOx, and VOCs
35 are based on EPA's 2001 National Emission Inventory (NEI) and are consistent with inventories
November'2005
7-30
Draft - Do Not Quote or Cite
-------
1 used for the analysis of the Clean Air Interstate Rule (CAIR) rule (EPA, 2005c). Biogenic
2 emissions were processed using the Biogenic Emissions Inventory .System (BEIS) version 3.13.
3 The staff recognizes that 03 exposures vary between years depending on meteorology and other
4 factors. Therefore, staff will add additional years for comparison if more CMAQ outputs
5 become available.
6 Recently EPA/NOAA conducted an initial evaluation of the eastern U.S. domain of
7 CMAQ version 4.5 (Appeletal., 2005;
8 http://www.cmascenter.org./docs/CMAO/v4.5/CMAOv4.5 EvaluationDocument-Final2005.pdf).
9 This evaluation used the same metrics published by Eder and Yu (2005) for the CMAQ version
10 4.4 model release. For the modeled summer months of June, July and August of 2001, CMAQ
11 version 4.5 predictions were compared to AQS monitor sites. The prediction of daily 8hr-max
12 Os was relatively good, showing a slight positive normalized mean bias of 1.62% and a
13 normalized mean error of 17.4%. Hourly ozone patterns are predicted well during the daytime.
14 However, the CMAQ consistently over-predicted hourly Os at night. Nighttime over-predictions
15 in Os have been improved over CMAQ version 4.4 by modifications to the minimum Kz
16 approximation in CMAQ version 4.5, but additional investigations are needed. Again, since
17 many of the assessments outlined below rely daytime Os accumulated in the 12hr SUM06 (8AM-
18 8PM), the night-time over-prediction is less of an issue. Overall, CMAQ predictions of daily 8hr
19 Os averages were improved in the 12km x 12km grid size when compared to the 36km x 36krn
20 grid size.
21 Since CMAQ output is averaged over large square blocks and monitor observations are
22 effectively averages over much smaller regions, CMAQ output and monitor observations have a
23 mismatch in spatial resolution. (Fuentes and Rafterty 2005). The problem is well known, but is
24 often ignored since there are not standard operational methods that can be applied to the CMAQ
25 model output to deal with this problem. The CMAQ version 4.5 evaluation described above
26 ignores the mismatch of spatial resolution and treats CMAQ output as a point-value. The staff
27 believes this simplification is reasonable because Os is a secondary pollutant and its
28 concentration generally varies fairly smoothly in flat rural areas where important crops and
29 vegetation grow. However, Q$ is notably variable in complex terrain, across in urban/rural
30 gradients and along coastal areas. In these cases significant differences in Os concentration
31 could occur with a 12xl2km call and the uncertainties associated with areas are unknown. The
32 current assessment is most concerned with rural areas and it is recognized that any estimates of
33 Os exposure in complex terrain are very uncertain.
November 2005
7-31
Draft - Do Not Quote or Cite
-------
Figure 7.2 Locations of AQS monitors (top) and CASTNET monitoring stations
a.
3 illtlllJi III 11:1 111 plBISS llfll
November 2005
7-32
Draft - Do Not Quote or Cite
-------
1 7.3.2.3 Generation of Potential Ozone Exposure Surfaces
2 Staff will interpolate Oa exposures on a regional basis rather than a continuous surface for
3 the US. The result of these interpolations will be the creation of regional Potential Ozone
4 Exposure Surfaces (rPOES). First staff will separate the eastern and western regions of the U.S.
5 based on the 12km and 36km CMAQ model domains. The eastern domain will be partitioned
6 into homogeneous sub-regions as defined by Eder et al. (1993). For the western domain, staff
7 will follow Hie subregions as defined by Lee et al (2001).
8 To estimate a rPOES for different regions of the contiguous U.S., staff plans to primarily
9 use the interpolation module in BenMAP, which uses the enhanced Voronoi Neighbor Averaging
10 (eVNA) interpolation method to combine monitored data with spatial scaling from CMAQ
11 model outputs (see appendix C.3.2 ofhttp://w^y\^epa.gpv/1te/ecas/models/mQd8idoc,pdft. This
12 method employs inverse-distance weighting of monitoring data scaled by the CMAQ model.
13 This method makes the implicit assumption that CMAQ adequately characterizes the general
14 spatial pattern of 63 exposures. It also retains the true monitored values at monitored sites.
15 Exceptions to this approach might be warranted in certain cases. For example, in areas with
16 fairly dense monitor coverage (e.g., northeastern U.S.) staff will investigate whether
17 interpolation without spatial scaling would be a satisfactory. Likewise, in areas of the country
18 with little or no monitor coverage, where an interpolation would depend on data from distant
19 monitors that may have very little correlation with the true Os exposure at the unmonitored cell,
20 staff is evaluating the benefit of identifying criteria that could be used to define the appropriate
21 spatial "window" within which monitored sites can be used to interpolate values for the non-
22 monitored area and/or whether it is more suitable in these cases to use CMAQ modeled 63
23 exposures only, instead of relying on interpolated Oj values. At a minimum, staff plans to
24 generate the rPOES in terms of both the 12-hr (8 am to 8 pm), maximum 3-month SUM06 index
25 and the 8-hr average index that reflects the form of the current secondary standard.
26 Staff recognizes there are inherent uncertainties in the interpolation that must rely on
27 sparse data representative of urban and near-urban areas with little representation of rural areas.
28 This network could bias the picture of the ozone exposure estimate. This limitation will be noted
29 when results are presented in the second draft of the staff paper. It is expected that this eVNA
30 with spatial scaling from CMAQ approach will be an improvement over a simple interpolation
31 between monitors that does not use spatial scaling from an air quality model. The interpolation
32 technique will be run without spatial scaling to test this expectation and to determine if
33 interpolation without CMAQ scaling could be used to characterize Os exposures in regions with
34 relatively dense monitoring coverage. This will allow for characterization of ozone exposures in
35 other years besides 2001. Finally, the uncertainties associated with estimating exposures at non-
November 2005
7-33
Draft - Do Not Quote or Cite
-------
1 monitored areas will be quantified by removing a subset of monitors and interpolating the
2 missing monitors with the remaining data.
3 During the recent CASAC consultation, the question was raised as to whether other
4 approaches, such as kriging, would be appropriate for the interpolation of Os exposures in non-
5 monitored areas. The eVNA approach outlined above is simple, but does not allow for
6 calculation of the model error in the same way kriging does. While staff agrees that kriging may
7 be an appropriate interpolation method in general, it is not available as an option in the BenMAP
8 model that will be used for this assessment,
9 7.3.2.4 Alternative Air quality scenarios
10 The following air quality scenarios will be generated:
11 • "As is" air quality (using base year 2001)
12 • Meeting the current standard: 4th highest daily maximum 8-hr average of 0.084 ppm
13 • Meeting alternative Os standards
14 In conjunction with work being done as part of the health risk assessment, the quadratic
15 air quality adjustment that was used in the last review will be used to simulate just meeting the
16 current and alternative standards (Johnson, 1997). This technique combines both linear and
17 quadratic elements to reduce larger Oa concentrations more than smaller ones. In this regard, the
18 quadratic method attempts to account for reductions in emissions without greatly affecting lower
19 concentrations near ambient background levels. EPA has recently evaluated the implementation
20 of this method on a subset of monitors and found that when this method is used to roll-back Os
21 values from "high" years, it yields a similar distribution of hourly Os values to that of "low" Oj
22 years (Rizzo, 2005). Further, the quadratic method performs better than proportional and peak-
23 shaving methods when compared to monitor data. In addition, EPA NHEERL-WED lab is
24 conducting an evaluation of the quadratic method for the rural Crestline monitor in the San
25 Bernardino Mountains of Southern California. Results of that evaluation will be included in the
26 next draft Staff Paper.
27 7.3.3 Crop Risk/Benefits Assessments
28 In light of a number of developments since the last review, including the potential to
29 better characterize Os exposures in crop growing regions using the CMAQ model, changes in air
30 quality, updated crop planting information, and an alternative agricultural economic model that
31 reflects the most up-to-date market forces, staff plans to update the previous review/s crop risk
32 and economic benefits assessments in order to better evaluate the adequacy of the level of
33 protection afforded by the current standard.
November 2005 7-34 Draft - Do Not Quote or Cite
-------
1 One element of the analysis that has not changed since the last review is the source of the
2 crop yield loss concentration response (C-R) functions. The 1996 crop assessment was built
3 upon the NCLAN (National Crop Loss Assessment Network) Os C-R functions. Since very few
4 new studies have published C-R functions that would be useful in an updated assessment, C-R
5 functions from NCLAN remain the best data available for a national assessment of crop loss
6 under various Os air quality scenarios. The NCLAN protocol was designed to produce crop C-R
7 data representative of the areas in which the crops were typically grown. The U.S. was divided
8 into 5 regions over which a network of field sites was established. In total, 15 crop species
9 , (com, soybean, winter wheat, tobacco, sorghum, cotton, barley, peanuts, dry beans, potato,
10 lettuce, turnip, and hay [alfalfa, clover, and fescue]), were studied. The first 12 of these 15 listed
11 species were analyzed for the .1996 review and included 38 different cultivars and were studied
12 under a variety of unique combinations of sites, water regimes, and exposure conditions,
13 producing a total of 54 separate cases. According to the most recent USDA National
14 Agricultural Statistical Survey (NASS) data, these 12 species account for greater than 70% of
15 principal crops acreage planted in the U.S. in 2004.1 Com, soybean, and winter wheat alone
16 accounted for 61% of principal crops acreage planted.
17 Since the NCLAN studies were performed in 1980-1988 there is some uncertainty
18 whether the crop cultivars tested in NCLAN are representative of crops grown today. In general,
19 new crop varieties are not specifically bred for Os tolerance. The fact that Os levels are not
20 consistent from year to year does not allow crop breeders to select for ozone tolerance under
21 natural conditions. Since the last review there has been little evidence that crops are becoming
22 more tolerant or more sensitive to ozone (EPA, 2005b). Crops are bred for higher yield and this
23 may even make them more susceptible to Os through higher stomatal conductance. In cotton,
24 some newer varieties have been found to have higher yield loss to ozone compared to older
25 varieties (Olszyk et al. 1993, Grantz and McCool 1992). In a meta-analysis of 53 studies
26 Morgan et al. (2003), found consistent deleterious effects of Os exposures on soybean from
, 27 studies published between 1973 and 2001. Further, early results from the Soy FACE experiment
28 in Illinois indicate a lack of any apparent difference in the Os tolerance of old and recent
29 cultivars of soybean in a study of 22 soybean varieties (Long et al. 2003). Given the limited
30 amount of information available on the Os sensitivity of current cultivars of different crops, staff
31 plans to focus this analysis most heavily on crops (such as soybean, cotton, and wheat) with
32 information on Os sensitivity of different cultivars.
1 Principal crops as defined by the USDA include com, sorghum, oats, barley, winter wheat, rye, Durum
wheat, other spring wheat, rice, soybeans, peanuts, sunflower, cotton, dry edible beans, potatoes, sugar beets, canola,
proso millet, hay, tobacco, and sugarcane. Acreage data for the principal crops was taken form the USDA NASS
2005 Acreage Report ftii
November 2005
7-35
Draft - Do Not Quote or Cite
-------
1 73.3.1 Exposure Assessment
2 In the last review, C-R functions were developed in terms of the SUM06 and Wl 26
3 indices for most NCLAN crops (Lee and Hogsett, 1996). Currently, work is underway to re-
4 analyze the NCLAN database to recalculate the C-R functions in terms of an 8-hr average index.
5 Specifically, staff plans to plot relative crop yield loss against an 8-hr average index calculated
6 from the 1-hr averages contained in the NCLAN database. The benefits of this re-analysis are
7 two-fold: (1) it permits evaluation of the appropriateness of the 8-hr average index for predicting
8 growth effects of the NCLAN studies as compared to a SUM06 index, and (2) it permits direct
9 evaluation of estimated yield effects expected to occur under air quality scenarios expressed in
10 terms of the current 8-hr, 0.08 ppm standard level.
11 73,3.2 Crop yield loss Assessment
12 County-level crop planting data will be obtained from USDA-NASS (National
13 Agricultural Statistics Service; http://www.asda.gov/nass) for 2001 for each NCLAN crop as
14 available (Figure 7-3). This information will be used to create GIS maps containing the planting
15 data for each species/cultivar of commodity crop. Staff plans to overlay the rPOES (as discussed
16 in section 7.3.2) with GIS maps of the crop growing regions and then calculate yield loss using
17 the relevant C-R functions. This combination of data will result in an estimate of county-level
18 percent yield loss for each selected NCLAN crop. Staff plans to create GIS maps of percent
19 yield loss of each crop for the counties in which they were planted in 2001. This analysis will
20 also be performed for just meeting the current standard and other alternative standards. The
21 change in crop county-level percent yield loss estimates between 'as is' 2001 air quality and
22 meeting various standards will serve as inputs to the AGSIM agricultural economic benefits
23 model.
November 2005
7-36
Draft - Do Not Quote or Cite
-------
>
I
I
a,
•8
O
S!
O)
I
a
*?
(•!•
I
-------
I
2 73.3.3 Economic Benefits Assessment -AGSIM
3 The Agriculture Simulation Model (AGSIM) model (Taylor 1994, Taylor et al., 1993)
4 has been utilized recently in many major policy evaluations.2 AGSIM is an econometric-
5 simulation model used to calculate agricultural benefits of changes in Os exposure and is based
6 on a large set of statistically estimated demand and supply equations for agricultural
7 commodities produced in the U.S.. A number of updates to AGSIM will be performed before
8 running the analysis: (1) an update of the commodity data for 2001, (2) incorporation of the
9 most recent version of the official USD A baseline model, (3) an econometric component added
10 to AGSIM to compute total farm program payments for different levels of farm program
11 parameters, and (4) farm payment program component added to the economic surplus module.
12 Initially, AGSIM will be used to calculate the economic benefits of yield changes between the
13 'as is' and 'just meet' scenarios for base year 2001. This approach will also be used to calculate
14 benefits from any alternative standards under consideration. If data are available, the same
15 analysis will be performed using air quality data from other years. Information will be used from
16 a range of crop cultivars as they are available for each crop. This will allow for some bounding
17 of possible effects of tolerant and sensitive cultivars.
18 7.3.4 Tree Seedling/Mature Tree/Forest Species Quantitative/Qualitative Risk
19 Assessments
20 In the last review, analyses of the effects of Os on trees were limited to 11 tree species for
21 which C-R functions for the seedling growth stage had been developed from OTC studies
22 conducted by NHEERL-WED. Since the last review, only a few studies have developed C-R
23 functions for additional tree seedling species (EPA, 2005b). Section 7.2.4.1 outlines plans to re-
24 analyze the OTC C-R functions in terms of an 8-hr average index. Section 7.2.4.2 describes how
25 staff plans to update the tree seedling risk analysis performed in the last review. Section 7.2.4.4
26 discusses the planned approach for modeling Oa impacts on mature trees. Section 7.2.4.3
27 discusses the planned approach for assessing Os effects on vegetation in natural settings using
28 visible foliar injury data. The tree and/or forest analyses outlined below will enable staff to
29 begin to assess important long-term effects of various secondary standard levels on'forest
30 ecosystem health and services.
2 For example, AGSIM© has been used in EPA's prospective study of the benefits deriving from the Clean
Air Act Amendments of 1990 required by section 812-B of the Clean Air Act, non-road land-based diesel engine
rule, and proposed Clear Skies legislation.
November 2005
7-38
Draft-
-------
1 7.3.4.1 Selected Species Exposure Assessment
2 Similar to crops (section 7.3.3), C-R functions for tree seedling biomass loss due to Os
3 exposures have not been reported in terms of an 8-hr average index. Staff plans to re-analyze the
4 11 OTC tree seedling C-R functions described in the 1996 0$ Staff Paper in terms of the current
5 8-hr exposure metric. This re-analysis will enable staff to evaluate the appropriateness of using
6 an 8-hr average index as a predictor of tree seedling growth/biomass losses and to directly
7 evaluate estimated seedling biomass loss values expected to occur under air quality, exposure
8 scenarios expressed in terms of the current 8-hr, 0.08 ppm secondary standard.
9 7.3.4.2 Selected Tree Seedling Biomass Loss
10 In the 1996 03 Staff Paper, information on tree species growing regions was derived from
11 the USDA Atlas of United States Trees (Little, 1971). Staff plans to use more recent information
12 from the USDA Forest Service FIA database in order to update tree growing ranges for the 11
13 tree species studied by NHEERL-WED. In a process similar to that used for crops staff plans to
14 combine the POES with the C-R function for each of the tree seedling species and information
15 on each tree species growing region to produce estimates of biomass loss for each of the 11 tree
16 seedling species. From this information, staff plans to generate GIS maps depicting these results
17 for each POES scenario. . ,
18 7.3.4.3 Foliar Injury Incidence/Epidemiology-FIA Data
19 Since the last review, there have been large amounts of data generated regarding visible
20 foliar injury to native plant species resulting from ozone exposure. The current draft CD
21 discusses the breadth of this new information.
22 • Foliar injury is a valuable indicator of phytotoxic concentrations of ozone in ambient air
23 which, at high enough concentrations, can have wide ranging effects from altering plant fitness
24 and aesthetics, possibly altering species composition of natural systems, and reducing the
25 marketability and yield of commercially valuable species. The degree of visible foliar injury is
26 dependent on a range of environmental and genetic factors and may vary widely even among co-
27 members of a population at similar exposure levels. While sensitivity to ozone may vary
28 considerably between and within taxonomic groups, common patterns of injury have been
29 discovered. There are generally four types of lesions that form on leaves and needles as a result
30 of ozone exposure in sensitive species: pigmented lesions (stippling), surface bleaching, bifacial
31 necrosis, and chlorosis.
32 The Unites States Forest Service, first as the Forest Health Monitoring Program (FHM)
33 and then as the Forest Inventory and Analysis (FIA) Program, has been collecting data since
34 1990 regarding the incidence and severity of visible foliar injury to several plants throughout ihe
35 U.S. shown to be sensitive to ozone (Coulston et al. 2003, 2004; Smith et al. 2003). This injury
November 2005
7-39
Draft-
-------
1 has been well documented as an effect of ozone on these species and is as such a useful indicator
2 of elevated ozone concentrations. Biomonitoring sites are located throughout the country and
3 analysis of foliar injury within these sites follows an established and rigorous protocol which
4 records the incidence of characteristic stippling on the leaves of species known to exhibit
5 sensitivity to ozone. The relative severity of injury is calculated based on the percentage of
6 leaves showing characteristic injury for multiple individual plants of several species per site.
7 These values are then used to give Biotic Index values for each site and each species for each
8 year (Smith et al., 2002).
9 Staff intends to compare the incidence and severity of visible foliar injury attributable to
10 ozone with available EPA maps of counties considered to fall below the level of Hie current 8hr.,
11 0.08 ppm secondary ozone standard (See Figure 7.4) for the same period to determine the extent
12 to which foliar injury continues to occur at levels below the current standard. Because the EPA
13 definition of attainment for these counties incorporates a three-year average of ambient air
14 concentration that does not capture year-to-year changes in ozone concentrations, this analysis
15 will compare year-to-year data to account for anomalous years in the ambient average.
16 Similarly, staff will also compare yearly county-level 12-hour SUM06 to incidence of foliar
17 injury (See Figure 7.5). Particular attention will be paid to visible foliar injury in national parks
18 (Class I areas) and other areas of aesthetic interest since the perception of visible injury in these
19 areas may be more valued by the public.
20 Interpolation of ozone monitors for 2001 will be done using the CMAQ model for spatial
21 scaling. This potential ozone exposure surface (rPOES), as described in section 7.2.2, will then
22 be used to determine the amount of correlation between modeled Os exposures, interpolated FIA
23 Os exposures, and visible foliar injury as measured in the FIA data for that year. If a close
24 correlation exists, further analysis of foliar injury for additional years using FIA interpolated
25 ozone values may be possible using site or species specific data to determine the extent of
26 impacts under the current ambient air quality. An analysis of this exposure surface will also
27 compare the degree to which the POES generated by EPA predicts relative changes in the
28 severity of foliar inj ury.
29 Important uncertainties associated with this approach of comparing 63 exposures to foliar
30 injury data include the following:
31 • The monitoring protocol used by the FIA changed significantly between the 2001 and
32 2002 data collection seasons. These changes added to the robustness of the data set
33 and increased the number of sites and species for which visible foliar injury is
34 collected. Due to the availability of CMAQ data, EPA must use 2001 data to generate
35 the POES. Since the FIA data for 2001 is concentrated in the northeast and north-
36 central regions of the US, those are the regions for which a comparison can be made.
November 2005
7-40
Draft-
-------
1
2
3
4
5.
6
7
8
The geographical locations of the FIA biomonitoring sites are considered confidential
by the Forest Service. The data provided to EPA may have the coordinates "fuzzed"
up to 8.5 miles to maintain this confidentiality. This introduces some error in
determining the actual location of the foliar injury with relation to county boundaries
and relative elevations.
The SUM06 values calculated by the FIA program include all air quality monitors in
the AQS but do not include CASTNET monitors. In addition, the elevation of the
monitors has not been considered.
November 2005
7-41
Draft-
-------
o
o
es
43
oo
•a
IS
-*.
f-
l
'
c nj
.s
sg
in
CO
o
O'T
I
I
V»"
XII
cov
S1"
«=OD
£9
a
a.
I
oi
I
I
«N
1
-------
B!
ujo9
85
CO
O
4=
-a
I
•s
o
O
D>
wco
£§
II
tn
Sw
Q.
Q.
a
I
I
-------
1 73.4.4 Ponderosa Pine case study for mature tree
2 In the 1996 Os Staff Paper, analyses on trees were limited to the seedling growth stage.
3 Recent experiments using the FACE methodology have been able to expose 3 tree species to Os
4 beyond the seedling growth stage. However, this methodology has not yielded C-R functions at
5 this time. Therefore, in order to go beyond the seedling stage, staff is planning to use a tree
6 growth model as a tool to evaluate the effect of changing Os air quality scenarios from just
7 meeting alternative Os standards on the growth of ponderosa pine. This method offers a means
8 to evaluate effects on mature trees capturing the interaction of Os uptake climate and
9 meteorology. .
10 A tree growth simulation model, TREGRO (Weinstein et al, 1990) has been used to
11 evaluate the effects of a variety of Os scenarios and linked with concurrent climate data to
12 account for ozone and climate/metorology interactions on several species of trees in different
13 regions of the US (Tingey et al., 2001; Weinstein et al., 1991; Retzlaff et al., 2000; Laurence et
14 al., 1993; Laurence et al., 2001; Weinstein et al., 2005). Staff is collaborating with the EPA
15 NHEERL-WED lab to use the TREGRO to assess long-term ponderosa pine growth at the
16 Crestline site in the San Bernardino Mountains of California associated with 'as is' air quality,
17 and air quality adjusted to j ust meet alternative Os standards. An earlier assessment of the
18 effectiveness of national air quality standards in place since the early 1970s took advantage of 40
19 years of air quality and climate data for this site to simulate ponderosa pine growth over time
20 with the improving air quality using TREGRO (Tingey etal., 2004).
21 Staff and NHEERL-WED scientists plan to use Crestline air quality and climate data
22 from the years 1995 to 2000 and 1980 tol 985. The years 1980 to 1985 will be used to represent
23 "bad" air quality years and 1995 to 2000 will represent recent improved air quality. Air quality
24 from each year will be rolled back using the quadratic method to 'just meet' the current 8-hr
25 secondary standard (4* highest maximum average = O.OSppm). TREGRO will be run for "as is"
26 and "just meet" in four 3 year increments to increase the accountability of climate variability and
27 the annual average biomass determined from these 4 simulations to yield an annual average
28 biomass change over the 6 years of ozone exposure. The differences between growth under 'just
29 meet' air quality and "as is" air quality will be compared and allow for the evaluation of the
30 potential effectiveness of the current secondary standard in protecting ponderosa pine as an
31 ecological resource. The comparisons between the annual average biomass change under the
32 two different 6 year periods (1980-1985 and 1995-2000) will allow staff to evaluate the
33 additional potential benefits from the current 8-hour standard compared to the indicated
34 improved growth achieved with the previous secondary standard put into place in the late 1970s.
November 2005
7-44
Draft - Do Not Quote or Cite
-------
1 7.4 ECOSYSTEM CONDITION, FUNCTION AND SERVICES
2 Ecosystems are comprised of complex assemblages of organisms that provide distinct
3 ecological attributes, many of which may be adversely affected by ozone (EPA, 2005b). A new
4 effort has been initiated within the Agency to identify indicators of ecological condition whose
5 responses can be clearly linked to changes in air quality and be used to track improvements in
6 environmental protection attributable to environmental program actions/implementation.
7 Moreover, a recent critique of the secondary N AAQS review process published in the report by
8 the National Academy of Sciences on Air Quality Management in the United States (NRC, 2004)
9 stated that "EPA's current practice for setting secondary standards for most criteria pollutants
10 does not appear to be sufficiently protective of sensitive crops and ecosystems...." This report
11 made several specific recommendations for improving the secondary NAAQS process and
12 concluded that "There is growing evidence that tighter standards to protect sensitive ecosystems
13 in the United States are needed " However, the vast majority of information regarding the
14 effects of ozone involves the sensitivity of individual species. Therefore, this section lays out
15 some examples of our current understanding of how the O? may be affecting ecosystems and
16 identifies areas of research needed to address this issue,
17 An ecosystem is defined as comprising all of the organisms in a given area interacting
18 with the physical environment, so that a flow of energy leads to a clearly defined trophic
19 structure, biotic diversity, and cycling of materials between living and nonliving parts (Odum,
20 1963). Individuals within a species and populations of species are the building blocks from
21 which communities and ecosystems are constructed. Classes of natural ecosystems, e.g., tundra,
22 wetland, deciduous forest, and conifer forest, are distinguished by their dominant vegetation
23 forms. Ecosystems boundaries are delineated when an integral unit is formed by their physical
24 and biological parts. Defined pathways for material transport and cycling and for the flow of
25 energy are contained within a given integrated unit."
26 Each level of organization within an ecosystem has functional and structural
27 characteristics. At the ecosystem level, functional characteristics include, but are not limited to,
28 energy flow; nutrient, hydrologic, and biogeochemical cycling; and maintenance of.food chains.
29 The sum of the functions carried out by ecosystem components provides many benefits to
30 mankind, as in the case of forest ecosystems (Smith, 1992). Some of these benefits include food,
31 fiber production, aesthetics, genetic diversity, and energy exchange.
32 A conceptual framework for discussing the effects of Os on ecosystems was developed
33 by the EPA Science Advisory Board (Young and Sanzone, 2002). Their six essential ecological
34 attributes (EEAs) include landscape condition, biotic condition, organism condition, ecological
35 processes, hydrologjcal and geomorphological processes, and natural disturbance regimes.
November 2005
7-45
Draft - Do Not Quote or Cite
-------
1 Figure 7.6 outlines the how common anthropogenic stressors, including tropospheric 63, might
2 affect the essential ecological attributes outlined by the SAB.
3 There is evidence that tropospheric Os is an important stressor of ecosystems, with
4 documented impacts on the biotic condition, ecological processes, and chemical/physical nature
5 of natural ecosystems (EPA, 2005b). Most of the effects on ecosystems must be inferred from
6 Os exposure to individual plants and processes that are scaled up through the ecosystem affecting
7 processes such as energy and material flow, inter- and intraspecies competition, and net primary
8 productivity (NPP). Thus, effects on individual keystone species and their associated microflora
9 and fauna, which have been shown experimentally, may cascade through the ecosystem to the
10 landscape level. By affecting water balance, cold hardiness, tolerance to wind and by
11 predisposing plants to insect and disease pests, 03 may even impact the occurrence and impact of
12 natural disturbance (e.g., fire, erosion).
13 Another approach to assessing Oa effects on ecosystems is the identification and use of
14 indicators. For example, the main indicators of phytotoxic Oj exposures used for forest
15 ecosystems are visible foliar injury (as described in section 7.3.4.3 above) and radial growth of
16 trees. Systematic injury surveys demonstrate that foliar injury occurs on Os-sensitive species in
17 many regions of the United States. However, there is not always a direct relationship between
18 the severity of the visible foliar symptoms and growth. This essentially means it is difficult to
19 quantify or characterize the degree which EEAs may be impacted when foliar injury is found in
20 the field. Investigations of the relationship between changes in radial growth of mature trees and
21 ambient Os, in combination with data from many controlled studies with seedlings, suggest that
22 ambient Os is reducing the growth of mature trees in some locations. However, definitively
23 attributing growth losses in the field to 63 in a wide array of ecosystems is often difficult
24 because of confounding factors with other pollutants, climate, insect damage and disease.
25 The draft CD (EPA, 2005b) outlines seven case studies where O3 effects on ecosystems
26 have either been documented or are suspected. However, in most cases, only a few components
27 in each of these ecosystems have been examined and characterized for ozone effects, and
28 therefore the full extent of ecosystem changes in these example ecosystems is not fully
29 understood. Clearly, there is a need for highly integrated ecosystem studies that specifically
30 investigate the effect of Os on ecosystem structure and function in order to fully determine the
31 extent to which ozone is altering ecosystem services.
November 2005
7-46
Draft - Do Not Quote or Cite
-------
1
2
3
Figure 7.6 Common anthropogenic stressors and the essential .ecological attributes they affect.
Modified-from Young and Sanzone (2002)
;Hydro
iKabUat Conversion
Climate Change
ibver-Hafvestins
invssive
Spfrcies IrttrcducJion
Large-Scale Disease/Pest
Outbreaks
Habitat Conversion
Climate Change
Over-Harvesting Vegttation
Di**s$e/Pe$t Outbreaks
Altered Fife Regime
Altered Flood Regime
Hy
-------
1 7.4.1 Evidence Demonstrating the Potential for Ozone to Alter Ecosystem
2 Structure and Function
3 The seven case studies listed in the draft CD demonstrate the potential for Os to alter
4 ecosystem structure and function. The oldest and best example involves the San Bernardino
5 Mountain forest ecosystem. In this example, Os appeared to be a,predisposing factor leading to
6 increased drought stress, windthrow, root diseases, and insect infestation (Takemoto et al., 2001).
7 Increased mortality of susceptible tree species including ponderosa and Jeffrey pine resulting
8 from these combined stresses has shifted community composition towards white fir and incense
9 cedar and has altered forest stand structure (Miller et al., 1989). Although the role of Os was
10 extremely difficult to separate from other confounding factors, such as high N deposition, there
11 is evidence that this shift in species composition has altered trophic structure and food web
12 dynamics (Pronos et al., 1999) and C and N cycling (Arbaugh et al., 2003). Ongoing research in
13 this important ecosystem will reveal the extent to which ecosystem services have been affected.
14 One of the most well-documented studies of population and community response to Os
15 effects are the long-term studies of sunflower (Plantago major) in native plant communities in
16 the United Kingdom (Davison and Reiling, 1995; Lyons et al., 1997; Reiling and Davison, .
17 1992c). Sensitive populations of sunflower had significant growth decreases in elevated Os
18 (Pearson et al., 1996; Reiling and Davison, 1992a,b; Whitfield et al., 1997) and reduced fitness
19 as determined by decreased reproductive success (Pearson et al., 1996; Reiling and Davison,
20 1992a). While spatial comparisons of population responses to Os are complicated by other
21 environmental factors, rapid changes in Os resistance were imposed by ambient levels and
22 variations in Os exposure (Davison and Reiling, 1995). At the site of sunflower seed collection
23 the highest correlations occurred between Os resistance and ambient Os concentrations (Lyons et
24 al., 1997). In this case study, it appears that Os- sensitive individuals are being removed by Os
25 stress and the genetic variation represented in the population could be declining. If genetic
26 diversity and variation is lost in ecosystems, there may be increased vulnerability of the system
27 to other biotic and abiotic stressors, and ultimately a change in the services provided by those
28 ecosystems.
29 Reconstructed ecosystems in artificial exposure experiments have also provided new
30 insight into how ozone may be altering ecosystem structure and function (Karnosky et al, 2005).
31 For example, the Aspen Free-Air CO2 Enrichment facility was designed to examine the effects of
32 both elevated CO2and Oson aspen (Populus tremuloides), birch (Betulapapyrifera), and sugar
33 maple in a simple reconstructed plantation characteristic of Great Lakes Aspen-dominated
34 forests (Karnosky et al., 2003b; Karnosky et al., 1999). They found evidence that the effects on
35 above- and below-ground growth and physiological processes have cascaded through the
November 2005 7-48 Draft - Do Not Quote or Cite
-------
1 ecosystem, even affecting microbial communities (Larson et al., 2002; Phillips et al., 2002),
2 This study also confirmed earlier observations of Os-induced changes in trophic interactions
3 involving keystone tree species, as well as important insect pests and their natural enemies
4 (Awmack et at., 2003; Holtonetal., 2003; Percy etal., 2002).
5 Collectively these examples suggest that Os is an important stressor in natural
6 ecosystems, but it is difficult to quantify the contribution of Os due to the combination of stresses
7 present in ecosystems. Continued research, employing new approaches, will be necessary to
8 fully understand the extent to which 63 is affecting ecosystem services.
9 7.4.2 Effects on Ecosystem Products and Services
10 Since it has been established that 63 affects photosynthesis and growth of plants, Os is
11 most likely affecting the productivity of crop and forest ecosystems. Therefore, it is desirable to
12 link effects on growth and productivity to essential ecosystem services. However, it is very
13 difficult to quantify ecosystem-level productivity losses because of the amount of complexity in
14 scaling from the leaf-level or individual plant to the ecosystem level, and because not all
15 organisms in an ecosystem are equally affected by ozone. Below is a discussion of potential
16 effects of Oa on two important ecological services.
17 7.4.2.1 Carbon Sequestration
18 Terrestrial ecosystems are important in the Earth's carbon (C) balance and potentially
19 have a role in offsetting emissions of CCh by humans. Temperate forests of the northern
20 hemisphere have been estimated to be a net sink of 0.6 to 0.7 Pg of C per year (Goodale et al.'
21 2002). COj enters vegetation through stomates and is used in photosynthesis to produce
22 carbohydrates that form plant tissues. Ozone interferes with photosynthesis, causes some plants
23 to senesce leaves prematurely and in some cases, reduces allocation to stem and root tissue.
24 Thus, 03 decreases the amount of net productivity and C sequestration of the individual plants
25 and entire ecosystems. One issue in the O3 effects on C sequestration is the interaction of rising
26 Os pollution and rising CO2 concentrations in the coming decades. Models generally predict that
27 in the future C sequestration will increase with increasing CO2, but often there is not an
28 accounting of the decrease in productivity due to the local effects of tropospheric 0$. In some
29 cases, the stimulatory effect of rising COi concentrations on forest productivity has been
30 estimated to be reduced by more that 20% (Tingey et al 2001; Ollinger et al. 2002; Karnosky et
31 al. 2003). Another issue, separate from the stimulation of productivity by COi, is how Os is
32 currently affecting C sequestration. In a study including all ecosystem types, Felzer et al. (2004),
33 estimated that US Net Primary Production (NPP) was decreased by 2.6-6.8% due to Os pollution
34 in Ihe late 1980's to early 1990's. Ozone not only reduces C sequestration in existing forests, it
35 can also affect reforestation projects. This effect, in turn, has been found to ultimately inhibit C
November 2005
7-49
Draft - Do Not Quote or Cite
-------
1 sequestration in forest soils which act as long-term C storage (Loya et al., 2003; Beedlow et al.
2 2005).
3 7.4.2.2 Water Resources
4 At present, there are no publications on the effects of Oa exposure that are carried through
5 at the ecosystem level to changes in mass water flow, channel morphology, riparian habitat
6 complexity, or sediment movement. It is possible that processes occurring at smaller scales are
7 affecting geomorphological processes in ecosystems; however, difficulties in scaling these
8 responses spatially and temporally have made this challenging to show experimentally. It is
9 possible that Oj exposure affects water quality through changes in energy and material flows.
10 One mechanism for how the amount of water quantity may be affected by Os exposure at the
11 landscape level is through loss of tight stomatal control. Moderately high Os exposure may
12 affect the mechanism of stomatal opening (McAinsh et al., 2002), resulting in sluggish stomatal
13 opening and closing (Reich and Lassoie, 1984). During moderately high Os exposure in a
14 drought year, canopy transpiration was greater for yellow poplar than on adjacent days with
15 lower Os exposure, which could alter water use at the landscape level. Oxidant exposure (Oa and
16 NOx) may decrease the ability of exposed plants to close stomata at night (Grulke et al., 2004),
17 thus increasing water loss from the landscape. Ecosystem models should aid in interpreting Os-
18 exposure effects at the landscape level.
19 7.4.3 Research needs
20 The knowledge base for examining the range of ecological effects of Oa on natural
21 ecosystems has changed little from the last review, however there is currently greater recognition
22 that ecosystem response to ozone needs to be examined in a more holistic way mat include
23 recognition of important ecosystem services and how to quantify and value changes to these
24 services. Below are listed areas of research that would improve our understanding of the role of
25 ozone, and how important ecosystem attributes may be protected. For example, there is a need
26 for information on the following ecosystem-level responses:
27 • Ecosystem Processes. Little is known about the effects of Os on water, C, and nutrient
28 cycling, particularly at the stand and community levels. Effects on below-ground
29 ecosystem processes in response to Os exposure alone and in combination with other
30 stressors are critical to projections at the watershed and landscape scales. Little is yet
31 known about the effects of Os on structural or functional components of soil food
32 webs, or how these impacts could affect plant species diversity (Andersen, 2003).
33 • Biodiversity and Genetic Diversity. The study of genetic aspects of Os impacts on
34 natural ecosystems has been largely correlational in nature and it remains to be shown
35 conclusively whether Os affects biodiversity or genetic diversity (Davison and Barnes,
36 1998; Pitelka; 1988; Winner et al., 1991). Studies of competitive interactions under
November 2005
7-50
Draft - Do Not Quote or Cite
-------
1 elevated Os levels are needed (Laurence and Andersen, 2003), and reexamination via
2 new sampling of population studies to bring in a time component to previous studies
3 showing spatial variability in population responses to 03 are needed. These studies
4 could be strengthened by modern molecular methods to quantify impacts on diversity.
5 • Natural Ecosystem Interactions with the Atmosphere, Little is known about feedbacks
6 between Oa and climate change on VOC production, which in turn., could affect Os
7 production (Fuentes et al., 2001). At moderate-to-high Os exposure sites, aberrations
8 in stomatal behavior could significantly affect individual tree water balance of sensitive
9 trees, and if the sensitive tree species is dominant, hydrologic balance at the watershed
10 , and landscape level could be affected. This has not been addressed in any model
11 because Os exposure effects, if included in the modeling effort have assumed a linear
12 ' relationship between assimilation and stomatal conductance.
13 * Below-Ground Interactions. While the negative effects of Os on below-ground growth
14 are well characterized, interactions of roots with the soil or microorganisms are not
15 • Other interactive Effects. Interaction studies with other components of global change
16 (e.g., warming, increasing atmospheric CC>2, N deposition) or with various biotic
17 stressors are needed to belter predict complex interactions likely in the future
18 (Laurence and Andersen, 2003). Whether 63 will negate the positive effects of an
19 elevated C(>2 environment on plant carbon and water balances is not yet known; nor is
20 it known if these effects will scale up through the ecosystem. How Os might affect the
21 progress of pest epidemics and insect outbreaks as concentrations increase is unclear
22 (Ball et al., 1998).
23 • Reproduction Effects. Information concerning the impact of Os on reproductive
24 processes and reproductive development under realistic field or forest conditions are
25 needed, as well as examination of reproductive effects under interacting pollutants
26 (Black et al., 2000).
27 • Comparative Extrapolation. The vast majority of O3 studies of trees have been
28 conducted with young, immature trees and in trees that have not yet formed a closed
29 canopy. Questions remain as to the comparability of O3 effects on juvenile and mature
30 trees and on trees grown in the open versus those in a closed forest canopy in a
31 competitive environment (Chappelka and Samuelson, 1998; Kolb and Matyssek, 2001;
32 Samuelson and Kelly, 2001).
33 • Scaling-Up Issues. Scaling the effects of Os from the responses of single or a few
34 plants to effects on communities and ecosystems is a complicated matter that will
35 . require a combination of manipulative experiments with model ecosystems,
36 community and ecosystem studies along natural Os gradients, and extensive modeling
37 efforts to project landscape-level, regional, national and international impacts of Oa.
38 Linking these various studies via impacts on common research quantification across
39 various scales using measures of such factors as leaf area index or spectral reflective
40 data, which could eventually be remotely sensed (Kraft et al., 1996; Panek et al., 2003),
41 would provide powerful new tools for ecologists.
42 • Comparative Risk Assessment Methodologies. Methodologies to determine the
43 important values of services and benefits derived from natural ecosystems such that
November 2005
7-51
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
these could be used in comprehensive risk assessment for 63 effects on natural
ecosystems (Heck et al.} 1998).
Economic Valuation. There is a critical need for research to support the development
of methods to value ecosystem services affected by Os in order to estimate the potential
economic and non-monetized benefits accruing from Os reductions.
November 2005
7-52
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
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
50
51
52
53
54
55
REFERENCES
Abt Associates, Inc. (1995). Ozone NAAQS benefits analysis: California crops. Report to U.S. EPA, July.
Adams, DM and Haynes R.W. 1996. USDA Forest Service Pacific Northwest Research Station. General Technical
Report GTR-PNW-368. The 1993 timber assessment market model: structure, projections, and policy
simulations. November 1996.
Amar, P., D. Chock, A. Hansen, M Moran, A. Russell, D. Steyn, and W. Stockwell, 2005: Final Report: Second
Peer Review of the CMAQ Model, Report submitted to Community Modeling and Analysis System Center,
University of North Carolina at Chapel Hill, May, 28 pp.
(http://WAW.craascenter.org/PDI;VC?,lAOScd_Peer Rev_July_5.pdf)
Amar, P., R. Bomstein, H. Feldman, H. Jeffries, D. Steyn, R. Yamartino, and Y. Zhang, 2004: Final Report:
December 2003 Peer Review of the CMAQ Model. Report submitted to Community Modeling and
Analysis System Center, University of North Carolina at Chapel Hill, July, 24 pp.
Andersen, C. P.; Hogsett, W. E.; Wessling, R.; Plocher, M. (1991) Ozone decreases spring root growth and root
carbohydrate content in ponderosa pine the year following exposure. Can. J. For. Res. 21:1288-1291.
Andersen, C. P.; Wilson, R.; Plocher, M.; Hogsett, W. E. (1997) Carry-over effects of ozone on root growth and
carbohydrate concentrations of ponderosa pine seedlings. Tree Physiol. 17: 805-811.
Andersen, C. P. (2003) Source-sink balance and carbon allocation below ground in plants exposed to ozone. New
Phytol. 157:213-228.
Appel,K.W.; A. Gilliland;B. Eder(2005) An Operational Evaluation of Hie 2005 Release of Models-3 CMAQ
Version 4.5. National Oceanic and Atmospheric Administration - Air Resources Laboratory, Atmospheric
Sciences Modeling Division; In partnership with the National Exposure Research Laboratory, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Arbaugh, M.; Bytnerowicz, 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.
Arnold J.R.; R. L. Dennis; G. S. Tonnesen, (2004) Diagnostic evaluation of numerical air quality models with
specialized ambient observations: testing the Community Multiscale Air Quality modeling system (CMAQ)
at selected SOS 95 ground sites, Atmospheric Environment 37:1185-1198.
Awmack, C. S.; Harrington, R.; Lindroth, R. L. (2004) Aphid individual performance may not predict population
responses to elevated CO2 or O3. Global Change Biol. 10:1414-1423.
Ball, G. R.; Benton, J.; Palmer-Brown, D.; Fuhrer, J.; Skarby, L.; Gimeno, B. S.; Mills, G. (1998)Identifying factors
which modify the effects of ambient ozone on white clover (Trifolium repens) in Europe. Environ. Pollut
103: 7-16.
Bichele, L; Moldau, H.; Padu, E. (2000) Estimation of plasmalemma conductivity to ascorbic acid in intact leaves
exposed to ozone. Physiol. Plant 108:405-412.
Booker, F. L.; Reid, C. D.; Brunschon-Harti, S.; Fiscus, E. L.; Miller, J. E. (1997) Photosynthesis and
photorespiration in soybean [Glycine max (L) Merr] chronically exposed to elevated carbon dioxide and
ozone. J. Exp. Bot 48:1843-1852.
Booker, F. L.; Fiscus, E. L.; Miller, J. E. (2004) Combined effects of elevated atmospheric carbon dioxide and ozone
on soybean whole-plant water use. Environ. Manage. 33: S355-S362.
November 2005
7-53
Draft - Do Not Quote or Cite
-------
I
2
3
4
5
6
7
8
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
50
51
52
53
54
55
Booker, F. L.; Miller, J. E.; Fiscus, E. L; Pursley, W. A.; Stefanski, L. A. (2005) Comparative responses of
container- versus ground-grown soybean to elevated CO2 and O3. Crop Sci. 45: 883-895.
Burkey, K. O. (1999) Effects of ozone on apoplast/cytoplasm partitioning of ascorbic acid in snap bean. Physiol.
Plant. 107:188-193.
Byun, D.W., Ching, J.K.S. (Eds.), 1999. Science Algorithms of the EPA Models-3 Community Multiscale Air
Quality Model (CMAQ) Modeling System. EPA/600/R-99/030, US Environmental Protection Agency,
Office of Research and Development, Washington, DC 20460.
Castillo, F. J.; Heath, R. L. (1990) CA2+ transport in membrane vesicles from pinto bean leaves and its alteration
after ozone exposure. Plant Physiol. 94: 788-795.
Castillo, F. J.; Miller, P. R.; Greppin, H. (1987) Extracellular biochemical markers of photochemical oxidant air
pollution damage to Norway spruce. Experientia (Basel) 43:111-115.
Chappelka, A. H.; Samuelson, L. J. (1998) Ambient ozone effects on forest trees of the eastern United States: a
review. New Phytol. 139: 91-108.
Cooley, D. R.; Manning, W. J. (1987) The impact of ozone on assimilate partitioning in plants: a review. Environ.
Pollut 47:95-113.
Coulston, J. W., Smith, G. C. and Smith, W. D. (2003). "Regional assessment of ozone sensitive tree species using
bioindicator plants." Environmental Monitoring and Assessment 83: 113-127.
Coulston, J.W., K. H. Riitters and G. C. Smith (2004). "A Preliminary Assessment of the Montreal Process
Indicators of Air Pollution for the United States." Environmental Monitoring and Assessment 95: 57-74.
DeCaria, A. J.; Pickering, K. E.; Stenchikov, G. L.; Scala, J. R.; Stith, J. L.; Dye, J. E.; Ridley, B. A.; Laroche, P.
(2000) A cloud-scale model study of lightning-generate NOx in an individual thunderstorm during
STERAO-A.J. Geophys. Res. [Atmos. ] 105:11,601-11,616.
Davison, A. W.; Barnes, J. D. (1998) Effects of ozone on wild plants. New Phytol. 139: 135-151.
Davison, A. W.;Reiling, K. (1995) A rapid change in ozone resistance of Plantago major after summers with high
ozone concentrations. New Phytol. 131: 337-344.
Dickson, R.E., Lewin K.F., Isebrands J.G., Coleman M.D., Heilman W.E., Riemenschneider D.E., Sober J., Host
G.E., Zak D.F., Hendrey G.R., Pregitzer K.S. and Kamosky D.F. (2000) Forest atmosphere carbon transfer
storage-II (FACTS II) - The aspen free-air CO2 and Oj enrichment (FACE) project in an overview. USDA
Forest Service North Central Research Station. General Tech. Rep. NC-214. 68pp.
Dominy, P. J.; Heath, R. L. (1985) Inhibition of the K+-stimulated ATpase of the plasmalemma of pinto bean leaves
by ozone. Plant Physiol. 77: 43-45.
Eder, B. and S. Yu, 2005: A performance evaluation of the 2004 release of Models-3 CMAQ, Atmos. Environ., in
press
Eder, B.K., Davis, J.M. and Bloomfield, P., 1993. A characterization of the spatiotemporal variability of non-urban
ozone concentrations over the eastern United States. Atmospheric Environment 27A: 2645-2668
Eder BK, Davis JM, Bloomfield P (1994) An automated classification scheme designed to better elucidate the
dependence of ozone on meteorology, J Appl Meteorol 33:1182-1199
November 2005
7-54
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
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
50
51
52
53
54
EPA (1989). Review of National Ambient Air Quality Standards for Ozone: Assessment of Scientific and Technical
. Information - OAQPS Staff Paper. Office of Air Quality Planning and Standards, Research Triangle Park,
NC. .
EPA (1996a). Air Quality Criteria for Ozone and Related Photochemical Oxidants. EPA/600/P-93/004aF-cF.
Office of Research and Development, National Center for Environmental Assessment, Research Triangle
Park,NC.
EPA (1996b). Review of National Ambient Air Quality Standards for Ozone: Assessment of Scientific and
Technical Information-OAQPS Staff Paper. EPA/452/R-96-007. Office of Air Quality Planning and
Standards, Research Triangle Park, NC.
EPA (2005a). Plan for Review of the National Ambient Air Quality Standards for Ozone. Office of Air Quality .
Planning and Standards, Research Triangle Park, NC. March. Available electronically on the internet at
htQ>://www.epa.gov/ttn/naaqs/standards/ozone/s_O3_cr_pd.htinl.
EPA (2005b). Air Quality Criteria for Ozone and Other Related Photochemical Oxidants. Second External Review
Draft National Center for Environmental Assessment, Research Triangle Park, NC. Available
electronically on the internet at: http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=l 14523.
EPA (2005c). Technical Support Document for the Final Clean Air Interstate Rule: Air Quality Modeling. Office of
Air Quality Planning and Standards, Research Triangle Park, NC.
Farage, P. K.; Long, S. P. (1999) The effects of O3 fumigation during leaf development on photosynthesis of wheat
and pea: an in vivo analysis. Photosynth. Res. 59:1-7.
Farage, P. K.; Long, S. P.; Lechner, E. G.; Baker, N. R. (1991) The sequence of changes within the photosynthetic
apparatus of wheat following short-term exposure to ozone. Plant Physiol. 95: 529-535.
Felzer, B.; Kickligher, D.; Melillo, J.; Wang, C.; Zhuang, Q.; Prinn, R. (2004) Effects of ozone on net primary
production and carbon sequestration in the conterminous United States using a biogeochemistry model.
Tellus B 56 (3), 230-248.
Fiscus, E. L.; Reid, C. D.; .Miller, I.E.; Heagle, A. S. (1997) Elevated CO2 reduces O3 flux and O3-induced yield
losses in soybeans: possible implications for elevated CO2 studies. J. Exp. Bot. 48: 307-313.
Fiscus, E. L.; Miller, J. E.; Booker, F. L.; Heagle, A. S.; Reid, C. D. (2002) The impact of ozone and other
limitations on the crop productivity response to CO2. Technology 8:181-192.
Fiscus, E. L.; Booker, F. L.; Burkey, K. O. (2005) Crop responses to ozone: uptake, modes of action, carbon
assimilation and partitioning. Plant Cell Environ.: in press.
Fuentes, M and Raftery, AE (2005). Model evaluation and spatial interpolation by Bayesian combination of
observations with outputs from numerical models. Biometrics, 61,36-45.
Fuentes, J. D.; Hayden, B. P.; Garstang, M.; Lerdau, M.; Fitzjarrald, D.; Baldocchi, D. D.; Monson, R.; Lamb, B.;
Geron, C. (2001) New directions: VOCs and biosphere-atmosphere feedbacks. Atmos. Environ. 35:189-
191.
Fuhrer, J. H. (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.
Gates, D. M. (1968) Transpiration and leaf temperature. Annu. Rev. Plant Physiol. 19: 211-238.
November 2005
7-55
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
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
50
51
52
53
54
55
Granlz, D.A., McCool, P.H., 1992. Effect of ozone onPima and Acala cottons in the San Joaquin Valley. In: Berber,
D.J., Richter, D.A. (Eds.), Proceedings 1992 Beltwide Cotton Conferences, vol 3 .National Cotton Council
of America, Memphis, TN, pp. 1082-1084.
Grantz, D. A.; Farrar, J. F. (2000) Ozone inhibits phloem loading from a transport pool: compartmental efflux
analysis in Pima cotton. Aust. J. Plant Physiol. 27: 859-868.
Gregg, J. W., C.G. Jones and I.E. Dawson (2003). "Urbanization effects on tree growth in the vicinity of New York
City." Nature 424:183-187.
Grilnhage, L.; Jager, H, J. (1994) Influence of the atmospheric conductivity on the ozone exposure of plants under
ambient conditions: considerations for establishing ozone standards to protect vegetation. Environ. Pollut
85: 125-129.
Grulke, N. E.; Alonso, R.; Nguyen, T.; Cascio, C.; Dobrowolski, W. (2004) Stomata open at night in pole-sized and
mature ponderosa pine: implications for O3 exposure metrics. Tree Physiol. 24: 1001-1010.
Hanson, P., Samuelson, L., Wullschleger, S., Tabberer, T. and Edwards, G. (1994) "Seasonal patterns of light-
saturated photosynthesis and leaf conductance for mature and seedling Quercus rubra L. foliage:
differential sensitivity to ozone exposure." Tree Physiology 14:1351-1366.
Heagle, A. S.; Body, D. E.; Heck, W. W. (1973) An open-top field chamber to assess the impact of air pollution on
plants. J. Environ. Qua!. 2: 365-368.
Heagle, A. S.; Letchworth, M. B.; Mitchell, C. A. (1983) Effects of growth medium and fertilizer rate on the yield
response of soybeans exposed to chronic doses of ozone: Phytopathology 73: 134-139.
Heagle, A.S., L.W. Kress, P.J. Temple, R.J. Kohut, J.E. Miller, H.E. Heggestad (1988) "Factors influencing ozone
dose-yield response relationships in open-top field chamber studies." In: W.W. Heck, O.C. Taylor, D.T.
Tingey eds. Assessment of crop loss from air pollutants: proceedings of an international conference;
October, 1987; Raleigh, NC. New York, NY: Elsevier Applied Science; pp. 141-179.
Heagle, A. S.; Reinert, R. A.; Miller, J. E. (1996) Response of white clover to ozone in different environments. J.
Environ. Qual. 25: 273-278.
Heagle, A. S.; Miller, J. E.; Pursley, W. A. (1998) Influence of ozone stress on soybean response to carbon dioxide
enrichment: III Yield and seed quality. Crop Sci. 38:128-134.
Heath, R. L. (1987) The biochemistry of ozone attack on the plasma membrane of plant cells. Recent Adv.
Phytochem. 21:29-54.
Heath, R. L. (1988) Biochemical mechanisms of pollutant stress. In: Heck, W. W.; Taylor, O. C.; Tingey, D. T.,
eds. Assessment of crop loss from air pollutants: proceedings of an international conference; October,
1987; Raleigh, NC. London, United Kingdom: Elsevier Applied Science; pp. 259-286.
Heath, R. L. (1994a) Alterations of plant metabolism by ozone exposure. In: Alscher, R. G.; Wellbum, A. R., eds.
Plant Responses to the Gaseous Environment: molecular, metabolic, and physiological aspects, molecular,
metabolic, and physiological aspects. London, United Kingdom: Chapman & Hall; pp. 121-146.
Heath, R. L. (1994b) Possible mechanisms for the inhibition of photosynthesis by ozone. Photosynt. Res. 39:
439451.
Heck, W. W.; Furiness, C. S.; Cowling, E. B.; Sims, C. K. (1998) Effects of ozone on crop, forest, and natural
ecosystems: assessment of research needs. EM (October): 11-22.
November 2005
7-56
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
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
50
51
52
53
54
55
56
Herstrom, A., W. Hogsett, D. Tingey, E. Lee, and D. Phillips (1995). Using a geographical information system to
estimate ozone exposure over forests across the United States
Holton, M. K.; Lindroth, R. L.; Nordheim, E. V. (2003) Foliar quality influences tree-herbivore-parisitoid
interactions: effects of elevated CO2,O3, and plant genotype. Oecologia 137: 233-244.
Horst, R.; Duff, M. (1995). Concentration data transformation and the quadratic rollback methodology (Round 2,
Revised). Unpublished memorandum to R. Rodriguez, U.S. EPA, June 8.
C. Hogrefe, J. Biswas, B. Lynn, K, Civerolo, J. -Y. Ku, J. Rosenthal, C. Rosenzweig, R. Goldberg and P. L.
Kinney, Simulating regional-scale ozone climatology over the eastern United States: model evaluation
results, Atmospheric Environment, Volume 38, Issue 17, June 2004, Pages 2627-2638.
Hogsett, W. E.; Tingey, D. T.; Holman, S. R. (1985) A programmable exposure control system for determination of
the effects of pollutant exposure regimes on plant growth. Atmos. Environ. 19:1135-1145.
Hogsett, W. E.; Olszyk, D.; Oimrod, D; P.; Taylor, G. E., Jr.; Tingey, D. T. (1987a) Air pollution exposure systems
and experimental protocols: volume 1: a review and evaluation of performance. Corvallis, OR: U.S.
Environmental Protection Agency, Environmental Research Laboratory; EPA report no. EPA/600/3-
87/037a. Available from: NTIS, Springfield, VA; PB88-181680.
Hogsett, W. E., Weber, J. E., Tingey, D., Herstrom, A., Lee, E. H., Laurence, J. A. (1997) "Environmental auditing:
an approach for characterizing tropospheric ozone risk to forests." Environ. Manage, 21:105-120
Hogsett, W. E.; Tingey, D. T.; Hendricks, C.; Rossi, D. (1989) Sensitivity of western conifers to SO2 and seasonal
interaction of acid fog and ozone. In: Olson, R. K.; Lefohn, A. S., eds. Effects of air pollution on western
forests [an A&WMA symposium; June; Anaheim, CA]. Air Pollution Control Association; pp. 469-491
(APCA transactions series: no. 16).
Horvath, L.;Nagy, Z.; Weidinger, T.; Artz, R.; Luke, W. T.; Valigura, R.; Pinto, J. P.; Womack, J. (1995)
Measurement effluxes of trace gases (Oj, NOx, SOi, CO2, HNO3), particulate sulfate and nitrate, water
vapour over short vegetation by gradient and eddy correlation techniques in Hungary. EGS XX. General
Assembly; April; Hamburg, Germany. Ann. Geophys. 13(suppl. 2): C490.
Howitt,R. E. (1995b), "Positive mathematical programming." American Journal of Agricultural Economics 77:
329.
Isebrands, J. G.; Dickson, R. E.; Rebbeck, J.; Kamosky, D. F. (2000) Interacting effects of multiple stresses on
growth and physiological processes in northern forest trees. In: Mickler, R. A.; Birsdey, R. A.; Horn, J.,
eds. Responses of northern U.S. forests to environmental change. New York, NY: Springer-Verlag; pp.
149-180. (Ecological studies: v. 139).
Isebrands, J. G.; McDonald, E. P.; Kruger, E.~; Hendrey, G.; Percy, K..; Pregitzer, K.; Sober, J.; Kamosky, D. F.
(2001) Growth responses of Populus tremuloides clones to interacting carbon dioxide and tropospheric
ozone. Environ. Pollut 115: 359-371.
Jans, U.; Hoigne, J. (2000) Atmospheric water: transformation of ozone into OH-radicals by sensitized
photoreactions or black carbon. Atmos. Environ. 34: 1069-1085.
Johnson, T. (1997). "Sensitivity; of Exposure Estimates to Air Quality Adjustment Procedure," Letter to Harvey
Richmond, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency, Research
Triangle Park, North Carolina.
Kolb, T. E.; Matyssek, R. (2001) Limitations and perspectives about scaling ozone impacts in trees. Environ. Pollut
115:373-393.
November 2005
7-57
Draft ~ Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
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
50
51
52
53
54
55
Karaosky, D.F., B. Mankovska, K. Percy, R.E. Dickson, O.K. Podila, J. Sober, A. Nooimets, G. Hendrey, M.D.
Coleman, M. Kubiske, K.S. Pregitzer, and J.G. Isebrands (1999). "Effects of tropospheric O3 on trembling
aspen and interaction with CO2: Results from an Oygradient and a FACE experiment." J. Water, Air and
SoilPollut. 116:311-322.
Karnosky, D. F., Z.E.Gagnon, R.E. Dickson, M.D. Coleman, E.H. Lee, J.G. Isebrands, (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.
Karnosky, D. F.; Zak, D. R.; Pregitzer, K. S.; Awmack, C. S.; Bockheim, J. G.; Dickson, R. E.; Hendrey, G. R.;
Host, G. E.; King, J. S.; Kopper, B. J.; Kruger, E. L.; Kubiske, M. E.; Lindroth, R. L.; Mattson, W. J.;
McDonald, E. P. (2003b) Tropospheric O3 moderates responses of temperate hardwood forests to elevated
CO2: A synthesis of molecular to ecosystem results from the Aspen FACE project. Fund Ecol. 17: 289-
304.
Kats, G.; Thompson, C. R.; Kuby, W. C. (1976) Improved ventilation of open top greenhouses. J. Air Pollut.
Control Assoc. 26:1089-1090.
i
Kats, G.; Olszyk, D. M.; Thompson, C. R. (1985) Open top experimental chambers for trees. J. Air Pollut. Control
Assoc. 35:1298-1301.
Kraft, M.; Weigel, H. J.; Mejer, G. J.; Brandes, F. (1996) Reflectance measurements of leaves for detecting visible
and non-visible ozone damage to crops. J. Plant Physiol. 148:148-154.
Kerstiens, G.; Lendzian, K. J. (1989) Interaction between ozone and plant cuticles: I ozone deposition and
permeability. NewPhytol. 112: 13-19.
Krupa, S. V.; Grunhage, L.; Jager, H. J.; Nosal, M.; Manning, W. J.; Legge, A. H.; Hanewald, K. (1995) Ambient
ozone (O3) and adverse crop response: a unified view of cause and effect. Environ. Pollut. 87: 119-126.
Laisk, A.; Kull, O.; Moldau, H. (1989) Ozone concentration in leaf intercellular air spaces is close to zero. Plant
Physiol. 90:1163-1167.
Lamb, C.; Dixon, R. A. (1997) The oxidative burst in plant disease resistance. Ann. Rev. Plant Physiol. Mol. Biol.
48: 251-275.
Langebartels, C.; Kemer, K.; Leonardi, S.; Schraudner, M.; Trost, M.; Heller, W.; Sandermann, H., Jr. (1991)
Biochemical plant responses to ozone: I differential induction of polyamine and ethylene biosynthesis in
tobacco. Plant Physiol. 95: 882-889.
Larson, J, L.; Zak, D. R.; Sinsabaugh, R. L. (2002) Extracellular enzyme activity beneath temperate trees growing
under elevated carbon dioxide and ozone: Soil Sci. Soc. Am. J. 66:1848-1856.
Laurence, J. A.; Andersen, C. P. (2003) Ozone and natural systems: understanding exposure, response, and risk.
Environ. Int 29:155-160.
Lee, E. H.; Hogsett, W. E (2001) Interpolation of temperature and non-urban ozone exposure at high spatial
resolution over the western United States. Climate Research. 18:163-179
Lee, E; H.; Hogsett, W. E. (1996) Methodology for calculating inputs for ozone secondary standard benefits
analysis: part II. Report prepared for Office of Air Quality Planning and Standards, Air Quality Strategies
and Standards Division, U.S. Environmental Protection Agency, Research Triangle Park, N.C., March.
Lee, E. H.; Tingey, D. T.; Hogsett, W. E. (1989) Interrelation of experimental exposure and ambient air quality data
for comparison of ozone exposure indices and estimating agricultural losses. Corvallis, OR: U.S.
November 2005
7-58
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
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
50
51
52
53
54
55
56
Environmental Protection Agency, Environmental Research Laboratory; EPA report no.
EPA-600/3-89-047. Available from: NTIS, Springfield, VA; PB89-195036.
Legge, A. H.; Grunhage, L.; Nosal, M.; Jager, H. J.; Krupa, S. V. (1995) Ambient ozone and adverse crop response:
an evaluation of North American and European data as they relate to exposure indices and critical levels.
Angew. Bot 69:192-205.
Leitao, L.; Goulas, P.; Biolley, J. P. (2003) Time-course of Rubisco oxidation in beans (Phaseolus vulgaris L)
. subjected to a long-term ozone stress. Plant Sci. 165:613-620.
Little, E.L., Jr. (1971) Atlas of United States Trees. U.S. Department of Agriculture, Forest Service;
Miscellaneous Publication No. 1146.
Laurence, J.A., Retzlaff.W.A, Kem, J.S., Lee, E.H., Hogsett,W.E., Weinstein, D.A., (2001). "Predicting the
regional impact of ozone and precipitation on the growth of loblolly pine and yellow poplar using linked
TREGRO and ZELIG models." For. Ecol. Mgmt. 146: 251-267.
Long, S.,Nelson, R.L., Ainsworth, L., Hollis, K., Mies, T., Morgan, P., Naidu, S., Ort, D.R., Webster, R., Zhu, X.
Adapting Soybean To Current And Future Change In Atmospheric Composition. Do We Need More Than
Field Selection Under Current Conditions?. Cellular And Molecular Biology Of Soybean Biennial
Conference. 2002. P. 401.
Lyons, T. M.; Barnes, J. D.; Davison, A. W. (1997) Relationships between ozone resistance and climate in
European populations of Plantago major. New Phytol. 136: 503-510.
Luwe, M. W. F.; Takahama, U.;Heber, U. (1993) Role of ascorbate in detoxifying ozone in the apoplast of spinach
(Spinacia oleracea L) leaves. Plant Physi ol. 101: 969-976.
Mandl, R. H.; Weinstein, L. H.; McCune, D. C.; Keveny, M (1973) A cylindrical, open-top chamber for the
exposure of plants to air pollutants in the field. J. Environ. Qual. 2: 371-376.
Manning, W. J.; Krupa, S. V. (1992) Experimental methodology for studying the effects of ozone on crops and
• trees. In: Lefohn, A. S., ed. Surface level ozone exposures and their effects on vegetation. Chelsea, MI:
,- Lewis Publishers, Inc.; pp. 93-156.
Matters, G. L.; Scandalios, J. G. (1987) Synthesis of isozymes of superoxide dismutase in maize leaves in response
to O3, SO2, and elevated O2. J. Exp. Bot 38: 842-852.
McAinsh, M. R.; Evans, N. G.; Montgomery, L. T. (2002) Calcium signalling in stomatal responses to pollutants.
New Phytol. 153:441-447.
McLaughlin, S. B.; McConathy. (1983) Effects of SO2 and O3 on alocation of 14C-labeled photosynthate in
Phaseolus vulgaris. Plant Physiol. 73: 630-635.
Mehlhom, H.; O'Shea, J. M.; Wellburn, A. R. (1991) Atmospheric ozone interacts with stress ethylene formation by
plants to cause visible plant injury. J. Exp. Bot. 42:17-24.
Miller, P. R.; McBride, J. R.; Schilling, S. L.; Gomez, A. P. (1989) Trend of ozone damage to conifer forests
between 1974 and 1988 in the San Bernardino Mountains of southern California. In: Olson, R. K.; Lefohn,
S., eds. Effects of air pollution on western forests [an A&WMA symposium; June; Anaheim, CA]. Air and
Waste Management Association; pp. 309-323. (APCA transactions series, no. 16).
Moldau, H. (1998) Hierarchy of ozone scavenging reactions in the plant cell wall. Physiol. Plant 104:617-622.
Moldau, H.; Bichele, I. (2002) Plasmalemma protection by the apoplast as assessed from above zero ozone
concentration in leaf intercellular air spaces. Planta 214: 484-487.
November 2005
7-59
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
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
50
51
52
53
54
Morgan, P. B.; Ainsworth, E, A.; Long, S. P. (2003) How does elevated ozone impact soybean? A meta-analysis of
photosynthesis, growth and yield. Plant Cell Environ. 26:1317-1328.
Morgan, P.B., Bernacchi, C.J., Ort, D.R., Long, S.P. (2004). "An in vivo analysis of the effect of season-long open-
air elevation of ozone to anticipated 2050 levels on photosynthesis in soybean." Plant Physiology 135:
2348-2357.
Noormets, A.; Sober, A.; Pell, E. J.; Dickson, R. E.; Posila, G. K.; Sober, J.; Isebrands, J. G.; Karnosky, D. F.
(2001) Stomatal and non-stomatal limitation to photosynthesis in two trembling aspen (Populus
tremuloides Michx) clones exposed to elevated CO2 and/or O3. Plant Cell Environ. 24: 327-336.
Olszyk, D., Bytnerowlez, A., Kats, G., Reagan, C., Hake, S., Kerby, T., Millhouse, D., Roberts, B., Anderson,
C., Lee, H. (1993) Cotton yield losses and ambient ozone concentrations in California's San Joaquin
Valley. Journal of Environmental Quality 22,602-611.
Overmyer, K.; Tuominen, H.; Kettunen, R.; Betz, C.; Langebartels, C.; Sandermann, H., Jr.; Kangasjarvi, J.
(2000) Ozone-sensitive Arabidopsis rcdl mutant reveals opposite roles for ethylene and jasmonate
signaling pathways in regulating superoxide-dependent cell death. Plant Cell 12:1849-1862.
Pauls, K. P.; Thompson, J. E. (1980) In vitro simulation of senescence-related membrane damage by ozone-
induced lipid peroxidation. Nature (London) 283: 504-506.
Panek, J. A.; Baldocchi, D. D.; Goldstein, A. H. (2003) The need for spatially and functionally integrated
models of ozone deposition to Sierra Nevada forests. In: Bytnerowicz, A.; Arbaugh, M. J.; Alonso, R.,
eds. Ozone Air Pollution in the Sierra Nevada: Distribution and Effects on Forests. Distribution and
Effects on Forests. New York, NY: Elsevier; pp. 325-357.
Percy, K. E.; Awmack, C. S.; Lindrotii, R. L.; Kubiske, M. E.; Kopper, B. J.; Isebrands, J. G.; Pregitzer, K. S.;
Hendry, G. R.; Dickson, R. E.; Zak, D. R.; Oksanen, E.; Sober, J.; Harrington, R; Kamosky, D. F. (2002)
Altered performance of forest pests under atmospheres enriched with CO2 and O3. Nature (London) 420:
403-407.
Pearson, S.; Davison, A. W.; Reiling, K.; Ashenden, T.; Ollerenshaw, J. H. (1996) The effects of different ozone
exposures on three contrasting populations ofPlantago major. NewPhytol. 132:493-502.
Pel, Z. M.; Murata, Y.; Benning, G.; Thomine, S.; Klusener, B.; Allen, G. J.; Grill, E.; Schroeder, J..I. (2000)
Calcium channels activated by hydrogen peroxide mediate abscisic acid signalling in guard cells. Nature
(London) 406:731-734.
Pell, E. J.; Schlagnhaufer, C. D.; Arteca, R. N. (1997) Ozone-induced oxidative stress: mechanisms of action and
reaction. PhysioL Plant. 100: 264-273.
Phillips, R. L.; Zak, D. R.; Holmes, W. E.; White, D. C. (2002) Microbial community composition and function
beneath temperate trees exposed to elevated atmospheric carbon dioxide and ozone. Oecologia 131: 236-
244.
Pleijel, H.; Wallin, G.; Karlsson, P.; Skarby, L.; Sellden, G. (1995) Gradients of ozone at a forest site and over a
field crop - consequences for the AOT4O concept of critical level. Water Air Soil Pollut. 85: 2033-2038.
Pleijel, H.; Karlssori, G. P.; Sild, E.; Danielsson, H.; Skarby, L.; Sellden, G. (1996) Exposure of a grass-clover
mixture to ozone in open-top chambers—effects on yield, quality and botanical composition. Agric.
Ecosyst. Environ. 59: 55-62.
November 2005
7-60
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
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
50
51
52
53
54
55
Pronos, J.; Merrill, L.; Dahlsten, D. (1999) Insects and pathogens in a pollution-stressed forest. In: Miller, P. R.;
McBride, J. R., eds. Oxidant air pollution impacts in the montane forests of southern California. Springer,
pp. 317-337.
Pitelka, L. F. (1988) Evolutionary responses of plants to anthropogenic pollutants. Trends in Ecology and
Evolution 3: 23 3-236.
Rao, M. V.; Davis, K. R. (2001) The physiology of ozone induced cell death. Planta 213: 682-690.
Rao, M. V.; Lee, H. I.; Davis, K. R. (2002) Ozone-induced ethylene production is dependent on salicylic acid, and
both salicylic acid and ethylene act in concert to regulate ozone-induced cell death. Plant J. 32: 447-456.
Regener, V.H. (1957). Vertical flux of atmospheric ozone. J. Geophys. Res. 62: 221-228.
Reid, C. D.; Fiscus, E. L. (1998) Effects of elevated [CO2] and/or ozone on limitations to CO2 assimilation in
soybean (Glycine max). J. Exp. Bot. 49: 885-895.
Reich, P. B.; Lassoie, J. P. (1984) Effects of low level O3 exposure on leaf diffusive conductance and water-use
efficiency in hybrid poplar. Plant Cell Environ. 7: 661-668.
Reiling, K.; Davison, A. W. (1992a) Effects of a short ozone exposure given at different stages in the development
otPlantago major L. New Phytol. 121: 643:647.
Reiling, K.; Davison, A. W. (1992b) The response of native, herbaceous species to ozone: growth and
fluorescence screening. New Phytol. 120: 29-37.
Reiling, K.; Davison, A. W. (1992c) Spatial variation in ozone resistance of British populations of Plantago major
L. New Phytol. 122: 699-708.
Retzlaff, W. A.; Arthur, M. A.; Grulke, N. E.; Weinstein, D. A.; Gollands, B. (2000) Use of a single-tree simulation
model to predict effects of ozone and drought on growth of a white fir tree. Tree Phy sioL 20: 195-202.
Rizzo, M (2005) Evaluation of a quadratic approach for adjusting distributions of hourly ozone concentrations to
meet air quality standards. Novembber 7,2005.
Roitsch, T. (1999) Source-sink regulation by sugar and stress. Curr. Opin. Plant Biol. 2:198-206.
Samuelson, L.; Kelly, J. M. (2001) Scaling ozone effects from seedlings to forest trees Tansley review no 21. New
Phytol. 149:21-41.
Sandermann, H., Jr. (1998) Ozone: an air pollutant acting as a plant-signaling molecule. Naturwissenschaften 85:
369-375.
Sandermann, H., Jr. (2000) Ozone/biotic disease interactions: molecular biomarkers as a new experimental tool.
Environ. Pollut. 108: 327-332.
Sasek, T. W.; Richardson, C. J.; Fendick, E. A; Bevington, S. R.; Kress, L. W. (1991) Carryover effects of acid rain
and ozone on the physiology of multiple'fiushes of loblolly pine seedlings. For. Sci. 37:1078-1098.
Schraudner, M.; Moeder, W.; Wiese, C.; Van Camp, W.; Inze, D.; Langebartels, C.; Sandermann, H., Jr. (1998)
Ozone-induced oxidative burst in the ozone biomonitor plant, tobacco Bel W3. Plant J. 16:235-245.
Smith, W. H. (1992) Air pollution effects on ecosystem processes. In: Barker, J. R.; Tingey, D. T., eds. Air pollution
effects on biodiversity. VanNostrand Reinhold; pp. 234-260.
November 2005
7-61
Draft ~ Do Not Quote or Cite
-------
1 Smith, G., Coulston J., Jepsen, J. and Prichard, T. (2003), "A national ozone biomonitoring program: Results from
2 field surveys of ozone sensitive plants in northeastern forest (1994-2000)" Environmental Monitoring and
3 Assessment 87(3): 271-291.
4
5 Stitt, M. (1996) Plasmodesmata play an essential role in sucrose export from leaves: a step towards an integration of
6 metabolic biochemistry and cell biology. Plant Cell 8: 565-571.
8 Takemoto, B. K.; Bytnerowicz, A.; Fenn, M. E. (2001) Current and future effects of ozone and atmospheric nitrogen
9 deposition on California's mixed conifer forests. Far. Ecot. Manage. 144:159-173.
10
11 Taylor R. (1994) "Deterministic versus stochastic evaluation of the aggregate economic effects of price support
12 programs" Agricultural Systems 44: 461-473.
13
14 Taylor, C.R. "AGSIM: Model Description and Documentation." Agricultural Sector Models for the United States.
15 C.R. Taylor, K.H. Reichelderfer, and S.R. Johnson, eds. Ames IA: Iowa State University Press, 1993.
16
17 Temple, P. J.; Riechers, G. H., Miller, P. R.; Lennox, R. W. (1993) Growth responses of ponderosa pine to
18 longterm exposure to ozone, wet and dry acidic deposition, and drought. Can. J. For. Res. 23: 59-66.
19
20 Tingey, D. T.; Fites, R. C.; Wickliff, C. (1975) Activity changes in selected enzymes from soybean leaves
21 following ozone exposure. Physiol. Plant 33: 316-320.
22
23 Tingey, D. T.; Standley, C.; Field, R. W. (1976) Stress ethylene evolution: a measure of ozone effects on plants.
24 Atmos. Environ. 10: 969-974.
25
26 Tingey, D. T.; Taylor, G. E., Jr. (1982) Variation in plant response to ozone: a conceptual model of physiological
27 events. In: Unsworth, M. H.; Ormrod, D. P., eds. Effects of gaseous air pollution in agriculture and
28 horticulture. London, United Kingdom: Butterworth Scientific; pp. 113-138.
29
30 Tingey, D. T.; Laurence, J. A.; Weber, J. A.; Greene, J.; Hogsett, W. E.; Brown, S.; Lee, E. H. (2001) Elevated CO2
31 and temperature alter the response of Pinus ponderosa to ozone: A simulation analysis. Ecol. Appl.
32 11: 1412-1424.
33
34 Tingey, D. T.; Hogsett, W. E.; Lee, E. H.; Laurence, J. A. (2004) Stricter ozone ambient air quality standard has
35 beneficial effect on Ponderosa pine in California. Environ. Manage. 34: 397-405.
36
37 Unsworth, M. H.; Heagle, A. S.; Heck, W. W. (1984a) Gas exchange in open-top field chambers -1 measurement
38 and analysis of atmospheric resistances to gas exchange. Atmos. Environ. 18: 373-380.
39
40 Unsworth, M. H.; Heagle, A. S.; Heck, W. W. (1984b) Gas exchange in open-top field chambers - II resistances to
41 ozone uptake by soybeans. Atmos. Environ. 18: 381-385.
42
43 Urban, D.L., G.B. Bonan, T.M. Smith, H.H. Shugart (1991). "Spatial applications of gap models." For. Ecol Mgmt.
44 42:95-110.
45
46 USDA Forest Service: 1999, 'Forest Health Monitoring 1999 Field Methods Guide. USDA Forest Service',
47 National Forest Health Monitoring Program, Research Triangle Park, NC.
48
49 Vahala, J.; Ruonala, R.; Keinanen, M.; Tuominen, H.; Kangasjarvi, J. (2003) Etiiylene insensitivity modulates
50 ozone-induced cell death in birch. Plant Physiol. 132: 185-195.
51
52 Walcek, C. J.; Yuan, H. H.; Stockwell, W. R. (1997) The influence of aqueous-phase chemical reactions on ozone
53 formation in polluted and nonpolluted clouds. Atmos. Environ. 31:1221-1237.
54
55 Weinstein, D.A., Beloin, R.M., R.D. Yanai(1991). "Modeling changes in red spruce carbon balance and allocation
56 in response to interacting ozone and nutrient stress." Tree Physiology 9:127-146.
November 2005 7-62 Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Wellbum, A. R. (1990) Why are atmospheric oxides of nitrogen usually phytotoxic and not alternative fertilizers?
New Phytol. 115:395-429.
Whitfield, C. P.; Davison, A. W.; Ashenden, T. W. (1997) Artificial selection and heritability of ozone resistance in
two populations of Plantago major. New Phytol. 137: 645-655.
Winner, W. R; Lefohn, A. S.; Cotter, I. S.; Greitner, C. S.; Nellessen, J.; McEvoy, L. R., Jr.; Olson, R. L.; Atkinson,
C. J.; Moore, L. D. (1989) Plant responses to elevational gradients of O3 exposures in Virginia. Proc. Natl.
Acad. Sci. U. S. A. 86: 8828-8832.
Winner, W. E.; Coleman, J. S.; Gillespie, C.; Mooney, H. A.; Pell, E. J. (1991) Consequences of evolving reistance
to air pollutants. In: Taylor, G. E.; Pitelka, L. F.; Clegg, M. T., eds. Ecological genetics and air pollution.
Springer-Verlag; pp. 177-202.
Young, T. F.; Sanzone, S., eds. (2002) A framework for assessing and reporting on ecological condition: an SAB
report. Washington, DC: U.S. Environmental Protection Agency, Science Advisory Board; report no. EPA-
. SAB-EPEC-02:009. Available: http://www.epa.gov/sab/pdf/epec02009.pdf [9 December, 2003].
November 2005
7-63
Draft - Do Not Quote or Cite
-------
1 APPENDIX 4A. MICROENVIRONMENT MODELING PARAMETERS
2 Air Exchange Rates for Indoor Residential Environments
3 Distributions of AERs for the indoor microenvironments were developed using data from
4 several studies. The analysis of these data and the development of the distributions used in the
5 modeling is described in detail in the draft Exposure Analysis TSD. This analysis showed that
6 the AER distributions for the residential microenvironments depend on the type of air
7 conditioning (A/C) and on the outdoor temperature, as well as other variables for which we do
8 not have sufficient data to estimate. This analysis clearly demonstrates that the AER
9 distributions vary greatly across cities and A/C types and temperatures, so that the selected AER
10 distributions for the modeled cities should also depend upon the city, A/C type and temperature.
11 For example, the mean AER for residences with A/C ranges from 0.39 for Los Angeles between
12 30 and 40 °C to 1.73 for New York between 20 and 25 °C. The mean AER for residences without
13 A/C ranges from 0.46 for San Francisco between 10 and 20 °C to 2.29 for New York between 20
14 and 25 °C. The need to account for the city as well as the A/C type and temperature is illustrated
15 by the result that for residences with A/C and between 20 and 25 °C, the mean AER ranges from,
16 0,52 for Research Triangle Park to 1.73 for New York. For each combination of A/C type, city,
17 and temperature with a minimum of 11 AER values, exponential, lognormal, normal, and
18 Weibull distributions were fit to the AER values and compared. Generally, the lognormal
19 distribution was the best-fitting of the four distributions, and so, for consistency, the fitted
20 lognormal distributions are used for all the cases.
21 One limitation of this analysis was that distributions were available only for selected
22 cities, and yet the summary statistics and comparisons demonstrate that the AER distributions
23 depend upon the city as well as the temperature range and A/C type. Another important
24 limitation of the analysis was that distributions were not able to be fitted to all of the temperature
25 ranges due to inadequate data. A description of how these limitations were addressed can be
26 found in the draft Exposure Analysis TSD.
27 City-specific, AER distributions were used where possible; otherwise data for a similar
28 city were used. We obtained estimates of A/C prevalence from the American Housing Survey
29(AHS, 2003) for each metropolitan area (Table A-l). The final AER distributions used
for the exposure modeling are given in Table A-2.
November 2005 4A-1 Draft - Do Not Quote or Cite
-------
Table .A-l. City-specific air conditioning prevalence rates (percentage of
residences with central or room A/C) .
City
Atlanta
Boston
Chicago
Cleveland
Detroit
Houston
Los Angeles
New York
Philadelphia
Sacramento
St. Louis
Washington DC
4
5
Survey Area and Year
Atlanta, 2003
Boston, 2003
Chicago, 2003
. Cleveland, 2003
Detroit, 2003
Houston, 2003
Los Angeles, 2003
New York, 2003 .
Philadelphia, 2003
Sacramento, 2003 '
St. Louis, 2003
Washington DC, 2003
'
Percentage
97
75
87
75
81
99
55
82
91
95
96
96
Table A-2. Lognormal distributions used for
residential air exchange rates
City
Houston
Air
Conditioner
in residence
(Yes/No)
Yes
No
Yes
Temperature
Range
(Celsius)
520
20-25
25-30
>30
<10
10-20
>20
520
Scale
-0.898
-0.760
-0.862
-0.695
-0.422
-0.469
-0.088
-0.529
Shape
0.748
0.662
0.814
.0.541
0.518
1.070
0.897
0,639
Geometric
Mean
0.407
0.467 '
0.422
0.499
0.656
0.625
0.916
0.589
Geometric
Standard
Deviation
2.113
1.938
2.258
1.717
1.679
2.916
2.451
1.894
November 2005
4A-2
Draft - Do Not Quote or Cite
-------
Los'Angeles
Boston, Chicago,
Cleveland, Detroit,
New York City,
Philadelphia'
Atlanta,
Washington D,C.
Sacramento
St. Louis
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
520
20-25
25-30
>30
510
10-20
20-25
>25
510
10-25
>25
510
10-20
>20
510
10-20
20-25
>25
510
10-20
20-25
>25
5 25
>25
510 '
10-20
20-25
>25
510
10-20
20-25
25-30
>30
510
-0.529
0.096
-0.207
-1.323
-0.611
-0.292
0.316
-0.012
-0.341
0.130
0.218
0.016
-0.235
0.474
-0.045
-0.660
-0.646
-0.937
-0.283
-0.359
0.313
0.065 .
-0.687
-0.186
-0.643
-0.408
0.052
-0.190 '
-0.153
-0.642
-0.405
-0.770
-0.466
-0.077
0.639
0.861
0.882
1.026
1.127
0.735
0.825
0.676
0.702
0.985
0.778 -
0.760
0.714
0.751
0.674
0.702
0.783
0.730
, 0.840
0.779
0.829
0.687
0.653
0.856
1.161
0.777
0.537
0.817
0.703
0.739
0.923
0.850
0.729
0.734
0.589
1.100
0.813
0.266
0.543
0.747
1.372
0.988
0.711
1.139
1.244
1.016
0.791
1.606
0.956
0.517
0.524
0.392
0.754
0.698
1.367
1.067
0.503
. . 0.830
0.526
0.665
1.054
- 0.827
0.858
0.526
0.667
0.463
0.627
0.926
1.894
2.365
2.415
2.790
3.087
2.085
2.283
1.967
2.018
2.677
2.177
2.138
2.042
2.119
1.962
2.017
2.189
2.076
2.317
2.180
2.292
1.989
1.921
2.353
3.192
2.174
1.711
2.265
2.020
2.094
2.517
2.339
2.073
2.084
November 2005
4A-3
Draft - Do Not Quote or Cite
-------
10-20
>20
-0.310
0.321
0.846
0.822
0.733
1.378 .
2.330
2.276
27
2 AER Distributions for Other Indoor Environments
3 To estimate AER distributions for non-residential, indoor environments (e.g., offices and
4 schools), we obtained and analyzed two AER data sets: "Turk" (Turk et al, 1989); and "Persiiy"
5 (Persiiy and Gorfain, 2004; Persiiy et al., 2005). The earlier 'Turk" data set (Turk et al., 1989)
6 includes 40 AER measurements from offices (25 values), schools (7 values), libraries (3 values),
7 and multi-purpose (5 values), each measured using an SFe tracer over two or four hours in
8 different seasons of the year. The more recent "Persiiy" data (Persiiy and Gorfain, 2004; Persiiy
9 et al., 2005) were derived from the U.S. EPA Building Assessment Survey and Evaluation
10 (BASE) study, which was conducted to assess indoor air quality, including ventilation, in a large
11 number of randomly selected office buildings throughout the U. S. This data base consists of a
12 total of 390 AER measurements in 96 large, mechanically ventilated offices; each office was
13 measured up to four times over two days, Wednesday and Thursday, AM and PM. The office
14 spaces were relatively large, with at least 25 occupants, and preferably 50 to 60 occupants. AERs
15 were measured both by a volumetric method and by a CC>2 ratio method, and included their
16 uncertainty estimates. For these analyses, we used the recommended "Best Estimates" defined by
17 the values with the lower estimated uncertainty; in the vast majority of cases the best estimate
18 was from the volumetric method.
19 Due to the small sample size of the Turk data, the data were analyzed without
20 stratification by building type and/or season. For the Persiiy data, the AER values for each office
21 space were averaged, rather using the'individual measurements, to account for the strong
22 dependence of the AER measurements for the same office space over a relatively short period.
23 The mean values are similar for the two studies, but the standard deviations are about twice as
24 high for the Persiiy data. The proposed AER distributions were derived from the more recent
25 Persiiy data only.
26 ' We fitted exponential, lognorrnal, normal, and Weibull distributions to the 96 office
space average AER values, and the best fitting of these was the lognorrnal. Table A-3
gives the fitted parameters for this distribution, which is
29 used for AER distributions for the indoor, non-residential microenvironments.
November 2005
4A-4
Draft - Do Not Quote or Cite
-------
Table A-3. Lognormal distributions used for all
non-residential indoor air exchange rates for all cities
Scale
0.1038
Shape
0.1036
Geometric
Mean
1.1094
Geometric
Standard
Deviation
3.0150
Lower
Bound
0.07
Upper
Bound
13.8
5 Proximity and Penetration Factors For Outdoors, In-vehicle, and Mass Transit
6 For the outdoors near-road, public garage/parking lot, and in-vehicle proximity factors,
7 and for the in-vehicle penetration factors, we use distributions developed from the Cincinnati
8 Ozone Study (American Petroleum Institute, 1997, Appendix B; Johnson et al., 1995). This field
9 study was conducted in the greater Cincinnati metropolitan area in August and September, 1994.
10 Vehicle tests were conducted according to an experimental design specifying the vehicle typej
11 road type, vehicle speed, and ventilation mode. Vehicle types were defined by the three study
12 vehicles: a minivan, a full-size car, and a compact car. Road types were interstate highways
13 (interstate), principal urban arterial roads (urban), and local roads (local). Nominal vehicle
14 speeds (typically met over one minute intervals within 5 mph) were at 35 mph, 45 mph, or 55
15 mph. Ventilation modes were as follows:
16 • Vent Open: Air conditioner off. Ventilation fan at medium. Driver's window half open.
17 Other windows closed.
18 • Normal A/C: Air conditioner at normal. All windows closed.
19 * Max A/C: Air conditioner at maximum. All windows closed.
20 Ozone concentrations were measured inside the vehicle, outside the vehicle, and at six fixed-site
21 monitors in the Cincinnati area.
22 The draft Exposure Analysis TSD documents the rationale for the selection of
23 distributions of penetration and proximity factors for outdoors and in-vehicle microenvironments
used in this modeling analysis, which are listed in Table A4.
25
November 2005
4A-5
Draft - Do Not Quote or Cite
-------
Table A-4. Distributions of penetration and
proximity factors for outdoors and in-vehicle microenvironments
Microenvironment
Outdoors near road;
outdoors, public
garage / parking lot
Other outdoors
microenvironments
In-vehicle
Parameter
Penetration
factor
Proximity
factor
Penetration
factor
Proximity
factor
Penetration
factor
Proximity
factor
Road
Type
All
'All
All
All
All
Local
Roads
Urban
Roads
Interstates
Distribution
Constant
Normal
Constant
Constant
Normal
Normal
Normal '
Normal
Mean
1.0
0.755
1.0
•T
1.0
0.300
0.755
0.754
0.364
Standard Lower
• Deviation Bound
-
0.203 0.422
0.232 0
0.203 0.422
0.243 ' 0.355
0.165 0.093
Upper
Bound
1.0
1
1
1
1
5
6
7
8
9
10
11
12
Ozone Decay and Deposition Rates
A distribution for combined Os decay and deposition rates was obtained from the analysis
of measurements from a study by Lee et al. (1999). This study measured decay rates in the
living rooms of 43 residences in Southern California. Measurements of decay rates in a second
room were made in 24 of these residences. The 67 decay rates range from 0.95 to 8.05 hour . A
lognormal distribution was fit to the measurements from this study, yielding a geometric mean of
2.5 and a geometric standard deviation of 1.5. This distribution is used for all indoor
microenvironments.
November 2005
4A-6
Draft - Do Not Quote or Cite
-------
1 REFERENCES
2 AHS, 2003. U.S. Bureau of the Census and U.S. Department of Housing and Urban Development. 2003 American
3 Housing Survey (AHS): National Survey Data. Available at:
4 http :/Avww.census. gov/fahes/www/hous inc/ahs/ahs .html. and http://www .huduser.org/tiatasets/ahs.html.
5 American Petroleum Institute. (1997). Sensitivity testing ofpNEM/O3 exposure to changes in the model algorithms.
6 Health and Environmental Sciences Department.
7 Johnson, T., Pakrasi, A., Wisfaeth, A., Meiners, G., Ollison, W. (1995). Ozone exposures within motor vehicles -
8 results of a field study in Cincinnati, Ohio. Proceedings 88* annual meeting and exposition of the Air &
9 Waste Management Association, June 18-23,1995. San Antonio, TX. Preprint paper 95-WA84A.02.
10 Persily, A. and J. Gorfain.(2004). Analysis of ventilation data from the U.S. Environmental Protection Agency
11 Building Assessment Survey and Evaluation (BASE) Study. National Institute of Standards and Technology,
12 NISTIR 7145, December 2004.
13 Persily, A., J. Gorfain, G. Brunner (2005). Ventilation design and performance in U.S. office buildings. ASHRAE
14 Journal. April 2005, 30-3 5.
15 Turk, B. H., D. T. Grimsrud, J. T. Brown, K. L. Geisling-Sobotka, J. Harrison, R. J. Prill (1989). Commercial
16 building ventilation rates and particle concentrations. ASHRAE, No. 3248.
November 2005
4A-7
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
APPENDIX 4B. FREQUENCY DISTRIBUTIONS OF DAILY MAXIMUM 8-HOUR
AVERAGE OZONE CONCENTRATIONS
The measured concentrations and the concentrations representing just meeting the current
8-hour standard, for 2004 only, are summarized in the graphs in this Appendix in terms of the
frequencies of daily maximum 8-hour average Os concentrations above various concentration
levels. Since these are summations of "exceedances" for all monitors in an urban area, having
more monitors will lead to higher counts in these figures. These graphs also show the percent
reduction in exceedances of levels from the 2004 base year to when the current 8-hour standard
is just met.
The lines in these graphs with hollow triangles and squares indicate frequencies for the
2004 base and current standard respectively, and their values can be read off the left axis. The
lines with solid circles give the percent reduction from the base to current standard, and
correspond to the scale of the axis on the right side of the figures.
November 2005
4B-1
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
Figure 4B-1
Atlanta CSA: Frequencies of doily maximum 8-hour ozone above different levels
10000
1000
1005!
0*
0.00 0,02 0.04 0.06 0.08 0.10 0.12 0.14 0.16
03 level (ppm)
Right axis: •*• Percent reduction from base
Left axis: ***Bosecase •»• Current standards
Figure 4B- 2
Boston CSA: Frequencies of daily maximum 8—hour ozone above different levels
100001 :» 1 ioox
10001
H
0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14
03 level (ppm)
Right axis; •*• Percent reduction from base
Left axis: *** Base case ""Current standards
0.16
November 2005
4B-2
Draft ~ Do Not Quote or Cite
-------
1
2
3
4
Figure 4B-3
Chicago CSA: Frequencies of doily maximum 8-hour ozone above different levels
100001 : 7*=* \ 100*
1000
05!
0.00 0.02 0.04 . 0.06 0.08 0.10 0.12 0.14 0.16
03 level (ppm)
Right axis: **• Percent reduction from base
Left axis: *** Base case " . ***Current standards
Figure 4B-4
Cleveland CSA: Frequencies of doily maximum 8—hour ozone above different levels
10000
1000
100
100*
1
0:00 0.02 0.04 0.06 0.08 0.10
03 level (ppm)
Right axis: *** Percent reduction from base
Left axis: •** Base case """Current standards
November 2005
4B-3
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
Figure 4B-5
Detroit CSA: Frequencies of daily maximum 8-hour ozone above different levels
10000
1000
100-
10Qx
0.00 0.02 0.04 0.06 0,08 0.10 0.12 0.14 0.16
03 level (ppm)
Right axis: *** Percent reduction from base
Left axis: *~*Basecase ""Current standards
Figure 4B-6
Houston CSA: Frequencies of daily maximum 8-hour ozone above different levels
10000
1000
100
100%
0.00 0.02 0.04 0.06 O.OB 0.10 0.12 0.14 0.16
03 level (ppm)
Right axis: *** Percent reduction from base
Left axis: *** Base case "*"Current standards
November 2005
4B-4
Draft - Do Not Quote or Cite
-------
2
3
4
Figure 4B-7
Los Angeles CSA: Frequencies of doily maximum 8—hour ozone above different levels
IQOOODi ^ •••••*- 1 1.QO.
10000
1000
100
0.00 0.02 0.04 0.06 O.OB 0.10 0.12 0.14 0.16
03 level (ppm)
Right axis: *** Percent reduction from base
Left axis: *** Base case ""Current standards
Figure 4B-8
New York CSA: Frequencies of daily maximum 8—hour ozone above different levels
100001
1000
100
100*
O.'OO 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16
03 level (ppm)
Right axis: *** Percent reduction from base
Left axis: •"•"• Base case ""Current standards
November 2005
4B-5
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
Philadelphia
10000
1000
100
Figure 4B-9
CSA: Frequencies of daily maximum 8-hour ozone above different levels
100*
0.00 0.02 0,04 0.06 O.OB 0.10 0.12 0.14 0.16
03 level (ppm)
Right axis: *** Percent reduction from base
Left axis: **• Base case "-"Current standards
07.
Sacramento
10000
1000
100
Figure 4B-10
CSA: Frequencies of daily maximum 8-hour ozone above different levels
100*
0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16
03 level (ppm)
Right axis: •** Percent reduction from base
Left axis: *** Base case •**Current standards
Figure 4B-10 1
November 2005
4B-6
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
Figure 4B-11
St. Louis CSA: Frequencies of daily maximum 8-hour ozone above different levels
10000
tooo
1
0,00 0.02 0,04
0.12 0.14
.16
0.06 0.08 0.10
03 level (ppm)
Right axis: •** Percent reduction from base
Left oxis: ***Basecase """"Current standards
Figure 4B-12
Washington CSA: Frequencies of daily maximum B-hour ozone above different
10000f
1000
100
1
0.00 0.02 0.04 0.06 0.08 0,10 0.12 0.14 0.11
03 level (ppm)
Right axis: •*• Percent reduction from base -
Left axis: •*• Base case ^Current standards
November 2005
4B-7
Draft - Do Not Quote or Cite
-------
APPENDIX 4C. EXPOSURES FOR EXERCISING CHILDREN
This appendix presents graphs for each modeled city of counts of person-days of
daily maxima 8-hour average exposures above different levels, concomitant with
moderate exertion, for children ages 5 to 18. The lines with hollow triangles and squares
indicate counts for the 2004 base and current standard respectively, and their values can
be read off the left axis. The lines with solid circles give the percent reduction from the
base to current standard counts, and correspond to the axis on the right.
November 2005
4C-1
Draft - Do Not Quote or Cite
-------
Figure 4C-1
Atlanta CSA: Children ages 5—18, Person—days with moderate
exertion above daily maximum 8—hour exposure levels
100%
90?.
80%
70%
60*
' 50%
40%
30%
20%
10%
07.
0.06 0.07 0.08 0.09 0.10 0.11 0.12
Exposure level (ppm)
Right axis: *** Percent reduction from base
Left axis: *•*"* Base case ***Current standard
Figure 4C-2 1
Boston CSA: Children ages 5-18. Person-days with moderate
exertion above daily maximum 8—hour exposure levels
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
0.06 0.07 0.08 0.09 0.10 0.11 0.12
Exposure level (ppm)
Right axis: *** Percent reduction from base
Left axis: *"•-* Base case ***Current standard
November 2005
4C-2
Draft - Do Not Quote or Cite
-------
Figure 4C-3
Chicago CSA: Children ages 5—18. Person—days with moderate
exertion above daily maximum 8—hour exposure levels
100%
90%
80%
70%
60%
50*
407.
30%
20%
10*
0.06
0.11 0'.12
0.07 0.08 0.09 0.10
Exposure level (p%pm)
Right axis: *** Percent reduction from base
Left axis: ***Base case ""Current standard
Figure 4C-4
Cleveland CSA: Children ages 5—18. Person—days with moderate
exertion above daily maximum 8—hour exposure levels
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
0.06 0.07 0.08 0.09 0.10 0.11 0.12
Exposure level (ppm)
Right axis: *** Percent reduction from base
Left axis: *** Base case *** Current standard
November 2005
4C-3
Draft - Do Not Quote or Cite
-------
Figure 4C-5
Detroit CSA: Children ages 5—18. Person—days with moderate
exertion above daily maximum 8—hour exposure levels
1007.
90%
80%
70%
60%
50%
40%
30%
20%
10%
0.06, 0.07 0.08 0.09 0.10- 0.11 0.12
Exposure level (ppm)
Right axis: *•* Percent reduction from base
Left axis: *** Base case ***Current standard
Figure 4C-6
Houston CSA: Children ages 5—18. Person—days with moderate
exertion above daily maximum 8—hour exposure levels
100%
90%
80%
70%
60%
50%
10,000,000
1,000,000
100.000
' 10,000
1.000
100
10
1
40%
2.0%
10%
0%
0.11 0.12
0.06 0.07 0:08 0.09 0.10
Exposure level (ppm)
Right axis: **• Percent reduction from base
Left axis: **•• Base case **•Current standard
November 2005
4C-4
Draft - Do Not Quote or Cite
-------
10,000,000
1,000,000
Figure 4C-7
Los Angeles CSA: Children ages 5-.18. Person-days with moderate
exertion above daily maximum 8—hour exposure levels
1007.
907.
807.
707.
60*
507.
40%
307.
207.
10%
V n n n n Q%
0.06 0.07 0.08 0.09 0.10 0.11 0.12
Exposure level (ppm)
Right axis: •-"Percent reduction from base
Left axis: *-*-*Base case •**Current standard
Figure 4C-8
New York CSA: Children ages 5—18. Person—days with moderate
exertion above daily maximum 8—hour exposure levels
1007.
90%
807.
707.
60%
507.
40%
307.
207.
10%
07.
0.06
0.11 0.12
0.07 0.08 0.09 0.10
Exposure level (ppm)
Right axis: *** Percent reduction from base
Left axis: *** Base case *** Current standard
November 2005
4C-5
Draft - Do Not Quote or Cite
-------
r
Figure 4C-9
Philadelphia CSA: Children ages 5-1S. Person-days with moderate
exertion above daily maximum 8—hour exposure levels
100%
90%
80*
70%
60%
50%
403
30%
2 Of.
0.06 0.07 0.08 0.09 0.10 0.11 0.12
Exposure level (ppm)
Right axis: •"*•• Percent reduction from base
Left axis: **-* Base case ""**Current standard
Figure 4C-10
Sacramento CSA: Children ages 5—18. Person—days with moderate
exertion above daily maximum 8—hour exposure levels
100%
90%
80%
70%
60%
50%
40%
30%
20%
0%
0.06 0.07
0.08 0.09 0.10
Exposure level (ppm)
Right axis: ~* Percent reduction from base
Left axis: *"*"*Base case ***Current standard
0.12
November 2005
4C-6
Draft - Do Not Quote or Cite
-------
Figure 4C-11
St. Louis CSA: Children ages 5—18. Person—days with moderate
exertion above daily maximum 8—hour exposure levels
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
0.06 0.07
0.11 0.12
0.08 0.09 0.10
Exposure level (ppm)
Right axis: ™~* Percent reduction from base
Left axis: *~"Base case ""^ Current standard
Figure 4C-12
Washington CSA: Children ages 5—18. Person—days with moderate
exertion above daily maximum 8—hour exposure levels
100%
90%
0.06 0.07 0.08 0.09 0.10 0.11 0.12
Exposure level (ppm)
Right axis: *** Percent reduction from base
Left axis: **~*Base case ""Current standard
November 2005
4C-7
Draft - Do Not Quote or Cite
-------
Appendix 4D. Table 4D-1. GEOS-CHEM Grids Used for 12 Urban Locations
City
Atlanta
Boston
Chicago
Cleveland
Detroit
Houston
Los Angeles
New York
Philadelphia
Sacramento
St Louis
Washington, DC
1 J
38 12
43 16
37 16
39 16
39 16
34 10
25 12
42 15
42 15
23 14
36 14
41 14
latitude longitude
34
42
42
42
42
30
34
40
40
38
38
38
-85
-72.5
-87.5
-82.5
-82.5
-95
-117.5
-75
-75
-122.5
-90
-77.5
November 2005
4D-1
Draft - Do Not Quote or Cite
-------
1 APPENDICES FOR CHAPTER 5
-------
Table 5A-1. Monitor-Specific O3 Air Quality Information: Atlanta, GA
AIRS Monitor ID
1305700011
1306700031
1307700021
1308500012
1308900021
1308930011
1309700041
1311300011
1312100551
1313500021
1315100021
1322300031
1324700011
Average:
Fourth
2002
0.089
0.100
0.099
0.088
0.095
0.090
0.098
0.088
0.100
0.089
0.099
0.099
0.099
0.095
Daily Maximum 8-Hour
Average (ppm)
2003 2004
0.084 0.073
0.077 0.083
0.077 0.068
0.080 0.084
0.091 0.088
0.085 0.080
0.077 0.084
0.091 0.089
0.088 0.092
0.082 0.085
0.083 0.073
0.078 0.087
0.083 0.082
Design Value*:
Average of the 3
Year-Specific
Values (ppm)
0.085
0.086
0.077
0.086
0.089
0.087
0.083
0.093
0.089
0.088
0.085
0.088
0.093
"The design value is the maximum of the monitor-specific averages of the annual
fourth daily maximum 8-hour average over the 3 year period.
Table 5A-2. Monitor-Specific O3 Air Quality Information: Boston, MA
AIRS Monitor ID
2500900051
2500920061
2500940041
2501711021
2502130031
2502500411
2502500421
2502700151
Average:
Fourth
2002
0.088
0.100
0.094
0.096
0.107
0.102
0.074
0.091
0.094
Daily Maximum 8-Hour
Average (ppm)
2003 2004
0.079 0.081
0.080 0.077
0.073 0.070
0.088 0.078
0.078 0.079
0.074 0.064
0.080 0.074
0.079 0.075
Design Value*:
Average of the 3
Year -Specific
Values (ppm)
0.086
0.083
0.079
0.091
0.086
0.07
0.081
0.091
*The design value is the
fourth daily maximum 8
maximum of the monitor-specific averages of the annual
i-hour average over the 3 year period.
November 2005
5A-1
Draft - Do Not Quote or Cite
-------
Table 5A-3. Monitor-Specific 03 Air Quality Information: Chicago, IL
AIRS Monitor ID
1703100011
1703100321
1703100422
1703100501
1703100641
1703100721
1703100761
1703110032
1703116011
1703140021
1703140071
1703142011
1703142012
1703170021
1703180031
1704360011
1708900051
1709710021
1709710071
1709730011
1711100011
1719710081
1719710111
1808900221
1808900241
1808900301
1808920081
1809100051
1809100101
1812700202
1812700241
1812700261
5505900021
5505900191
5505900221
Average:
Fourth
2002
0.094
0.096
0.103
0.084
0.085
0.085
0.092
0.081
0.084
0.093
0.087
0.067
0.091
0.074
0.084
0.082
0.090
0.100
0.087
0.090
0.086
0.087
0.094
0.086
. 0.101 .
0.107
0.100
0.097
0.101
0.100
0.110
0.116
0.096
0.092
Daily Maximum 8-Hour
Average (ppm)
2003
0.077
0.080
0.069
0.067
0.075
0.071
0.075
0.070
0.073
0.080
0.082
0.066
0.076
0.074
0.078
0.079
0.077
0.073
0.076
0.081
0.081
0.082
0.084
0.079
0.077
0.082
0.085
0.088
0.088
0.077
2004
\ 0.065
0.067
0.054
0.060
0.068
0.067
0.067
0.059
0.064
0.067
0.051
0.071
0.065
0.069
0.068
0.071
0.068
0.063
0.068
0.064
0.064
0.067
0.070
0.069
0.072
0.078
0.066
Design Value*:
Average of the 3
Year-Specific
Values (ppm)
0.078
0.081
0.068
0.073
0.076
0.074
0.071
0.076
0.078
0.081
0.071
0.075
0.077
0.083
0.079
0.075
0.076
0.078
0.083
0.086
0.082
0.084
0.094
0.094
*The design value is the maximum of the monitor-specific averages of the annual
fourth daily maximum 8-hour average over the 3 year period.
November 2005
5A-2
Draft - Do Not Quote or Cite
-------
Table 5A-4. Monitor-Specific Os Air Quality Information: Cleveland, OH
AIRS Monitor ID
3900710011
3903500341
3903500641
3903550021
3905500041
3908500031
3908530021
3909300171
3910300031
3913310011
3915300201
Average:
Fourth
2002
0.103
0.090
0.090
0.098
0.115
- 0.104
0.088
0.099
0.091
0.097
0.103
0.098
Daily Maximum 8-Hour
Average (ppm)
2003 2004
0.099 0.081
0.076 0.057
0.079 0.063
0.089 0.077
0.097 0.075
0.092 0.079
0.080 0.076
0.085 0.074
0.086 0.077
0.091 0.081
0.089 0.077
0.088 0.074
Design Value*:
Average of the 3
Year-Specific
Values (ppm)
0.094
0.074
0.077
0.088
0.095
0.091
0.081
0.086
0.084
. 0.089
0.089
0.095
The design value is the
fourth daily maximum 8
maximum of the monitor-specific averages of the annual
i-hour average over the 3 year period.
Table 5A-5. Monitor-Specific O3 Air Quality Information: Detroit, Ml
AIRS Monitor ID
2604900211
2604920011
2609900091
2609910031
2612500012
2614700051
2616100081
2616300012
2616300161
2616300192
Average:
Fourth
2002
0.088
0.089
0.095
0.092
0.093
0.100
0.091
0.088
0.092
0.083
0.091
Daily Maximum 8-Hour
Average (ppm)
2003 2004
0.087 0.075
0.091 0.077
0.102 0.081
0.101 0.071
0.090 0.075
0.086 0.074
0.091 • 0.071
0.085 0.065
' 0.084 0.066
0.098 0.066
0.092 0.072
Design Value*:
Average of the 3
Year-Specific
Values (ppm)
0.083
0.085
0.092
0.088
0.086
0.086
0.084
' 0.079
0.08
" 0.082
0.092
*The design value is the
fourth daily maximum 8
maximum of the monitor-specific averages of the annual
-hour average over the 3 year period.
November 2005
5A-3
Draft - Do Not Quote or Cite
-------
Table 5A-6. Monitor-Specific O3 Air Quality Information: Houston, TX
AIRS Monitor ID
4803910032
4803910041
4803910161
4816700141
4816710022
4820100242
4820100263
4820100292
4820100461
4820100472
4820100512
4820100551
4820100621
4820100661
4820100701
4820100751
4820110151
4820110342
4820110353
4820110391
4820110411
4820110501
4833900781
. Average:
Fourth
2002
0,095
0.092
0.093
0.083
0.096
0.088
0.098
0.078
0.072
0.101
0.094
0.095
0.084
0.088
0.078
0.093
0.092
0.095
0.090
0.094
0.082
0.090
Daily Maximum
Average (ppm)
2003
0.097
0.092
0.082
0.095
0.098
0.096
0.093
0.082
0.103
0.107
0.094
0.081
0.100
0.096
0.108
0.102
0.105
0.113
0.092
0.094
0.097
8-Hour
2004
0.103
0.081
0.088
0.096
0.085
0.090
0.084
0.083
0.095
0.104
0.097
0.097
0.078
0.093
0.093
0.091
0.092
0.097
0.097
0.080
0.091
Design Value*:
Average of the 3
Year -Specific
Values (ppm)
0.097
0.091
0.095
0:09
0.094
0.085
0.079
0.099
0.101
0.095
0.087
0.088
0.089
0.095
0.096
0.101
0.094
0.085
0.101
The design value is the
fourth daily maximum 8
maximum of the monitor-specific averages of the annual
.-hour average over the 3 year period.
November 2005
5A-4
Draft - Do Not Quote or Cite
-------
Table 5A-7. Monitor-Specific 03 Air Quality Information: Los Angeles, CA
AIRS Monitor ID
0603700021
0603700161
0603701131
0603710021
0603711031
0603712011
0603713011
0603716011
0603717011
0603720051
0603740021
0603750011
0603750051
0603760121
0603790331
0605900071
0605910031
0605920221
060595001 1
0606500121
0606520021
0606550011
0606560011
060658001 1
0606590011
0606590031
0607100011
0607100051
0607100121
0607100171
0607103061
'0607110042
0607112341
0607120021
0607140011
0607140031
0607190021
0607190041
0611100051
0611100071
0611100091
0611110041
0611120021
0611120031
0611130011
Average:
Fourth
2002
0.097
0.111
0.073
0.091
0.077
0.111
0.049
0.074
0.099
0.095
0.059
0.064
0.131
0.102
0.069
0.066
0.081
0.071
0.113
0.097
0.109
0.107
0.109
0.104
0.092
0.131
0.115
0.087
0.106
0.105
0.089
0.114
0.113
0.117
0.101
0.105
0.076
0.080
0.087
0.097
0.092
0.064
0.064
0.093
Dally Maximum
Averaqe (ppfn)
2003
0.104
0.123
0.083
0.096
0.082
0.119
0.057
0.082
0.109
0.101 '
0.063
0.070
0.137
0.103
0.080
0.079
0.095
0.080
0.127
0.100
0.105
0.116
0.120
0.112
0.088
0.130
0.103
0.084 .
0.104
0.114
0.087
0.132
0.110
0.137
0.111
0.123
0.087
0.093
0.093
0.093
0.074
0.069
0.099
8-Hour
2004
0.092
0.095
0.076
0.089
0.078
0.101
0.065
0.079
0.095
0.093
0.070
0.085
0.107 .
0.095
0.088
0.076.
0.085
0.075
0.112
0.094
0.099
0.095
0.111
0.100
0.060
0.082
0.122
0.097
0.087
0.085
0.102
0.082
0.111
0.099
0.119
0.102
0.112
0.086
0.086
0.092
0.092
0.069
0.065
0.091
Design Value*:
Average of the 3
Year-Specific
Values (ppm)
0.097
0.109
0.077
0.092
0.079
0.11
0.057
0.078
0.101
0.096
0.064
0.125
0.1
0.079
0.073
0.087
0.075
0.117
0.097
0.104
0.106
0.113
0.105
0.087
0.127
0.105
0.086
0.098
0.107
0.086
0.119
0.107
0.124
0.104
0.113
0.084
0.088
0.094
0.092
0.069
0.066
0.127
*The design value is the maximum of the monitor-specific averages of the annual
fourth daily maximum 8-hour average over the 3 year period.
November 2005
5A-5
Draft - Do Not Quote or Cite
-------
Table 5A-8. Monitor-Specific O3 Air Quality Information: New York, NY
AIRS Monitor ID
3600500831
3600501101
3602700071
3607150011
3607900051
3608100981
3608101241
3608500671
3610300021
3610300041
3610300092
3611110051
3611920041
Average:
Fourth
2002 .
0.096
0.089
0.111
0.082
0.102
0.082
0.089
0.099
0.108
0.090
0.103
0.084
0.102
0.095
Daily Maximum 8-Hour
Average (ppm)
2003 2004
0.079 0.074
0.082 0.069
0.081 0.076
0.087 0.078
0.082 0.082
0.072 0.064
0.086 0.075
0.086 0.083
0.094 0.081
0.082
0.102 0.079
0.082 0.076
0.091 0.078
0.085 0.076
Design Value*:
Average of the 3
Year-Specific
Values (ppm)
0.083
0.08
0.089
0.082
0.088
0.072
0.083
0.089
0.094
0.094
0.08 ' '
0.09
0.094
*The design value is the maximum of the monitor-specific averages of the annual
fourth daily maximum 8-hour average over the 3 year period.
Table 5A-9. Monitor-Specific O3 Air Quality Information: Philadelphia, PA
AIRS Monitor ID
4201700121
4202900501
4202901001
4204500021
4209100131
4210100041
4210100141
4210100241
4210101361
Average:
Fourth
2002
0.111
0.104
0.112
0.106
0.101
0.082
0.098
0.110
0.094
0.102
Daily Maximum 8-Hour
Average (ppm)
2003 2004
0.087 0.082
0.085
0.085 0.085
0.080 0.081
0.085 0.083
0.069 0.054
0.083 0.077
0.082 0.091
0.070 0.073
0.081 . 0.078
Design Value*:
Average of the 3
Year -Specific
Values (ppm)
0.093
0.094
0.089
0.089
0.068
0.086
0.094
0.079
0.094
*The design value is the maximum of the monitor-specific averages of the annual
fourth daily maximum 8-hour average over the 3 year period.
November 2005
5A-6
Draft - Do Not Quote or Cite
-------
Table 5A-10. Monitor-Specific O3 Air Quality Information: Sacramento, CA
AIRS Monitor ID
0601700101
0601700111
0601700121
0601700201
0605700051
0605700071
0605710011
0606100021
0606100041
0606100061
0606100071
0606130011
0606700021
0606700061
0606700101
0606700111
0606700121
0606700131
0606750031
0611300041
0611310031
Average:
Fourth
2002
0.098
0.067
0.077
0.111
0.099
0.093
0.065
0.101
0.101
0.095-
0.097 -
0.095
0.105
0.083
0.069
0.104
0.079
0.097
0.076
0.088
0.090
Daily Maximum
Average (ppm)
2003
0.096
0.065
0.075
0.106.
0.098
0.090
0.094
0.089
0.085
0.068
*
0.086
0.097
0.076
0.087
0.098
0.075
0.097
0.077
0.082
0.086 '
8-Hour
2004
0.089
0.073
0.089
0.093
0.085
0.092
0.087
0.082
0.076
0.083
0.067
0.077
0.087
0.067
0.089
0.071
0.069
0.081
Design Value*:
Average of the 3
Year -Specific
Values (ppm)
0.094
0.075
0.102
0.096
0.089
0.095
0.092
0.087
0.085
0.095
0.075
0.077
0.096
0.073
0.094
0.074
0.079
0.102
*The design value is the maximum of the monitor-specific averages of the annual
fourth daily maximum 8-hour average over the 3 year period.
November 2005
5A-7
Draft - Do Not Quote or Cite
-------
Table 5A-11. Monitor-Specific O3 Air Quality Information: St. Louis, MO
AIRS Monitor ID
1708310011
1711700021
1711900081
1711910091
1711920072
1711930071
1716300102
2909900121
2918310021
2918310041
2918900041 '
2918900061
2918930011
2918950011
2918970031
2951000071
2951000721
2951000861
Average:
Fourth
2002
0.100
0.085.
0.094
0.090
0.090
0.084
0.093
0.093
0.099
0.098
0.098
0.094
0.094
0.095
0.093
0.090 '
0.081
0.098
0.093
Daily Maximum
Average (ppm)
2003
0.083
0.077
0.089
0.088
0.082
0.083
0.079
0.082
0.091
0.090
0.088
0.086
0.082
0.088
0.088
0.084
0.071
0.090
0.085
8-Hour
2004
0.073
0.068
0.074
0.078
0.068
0.073
0.073
0.070
0.077
0.076
0.070
0.067
0.067
0.068
0.069
0.058
0.072
0.071
Design Value*:
Average of the 3
Year -Specific
Values (ppm)
0.085
0.076
0.085
0.085
0.08
0.08
0.081
0.081
0.089
0.088
0.085
0.082
0.081
0.083
0.083
0.07
0.086
0.089
The design value is the maximum of the monitor-specific averages of the annual
fourth daily maximum 8-hour average over the 3 year period.
Table 5A-12. Monitor-Specific O3 Air Quality Information: Washington, D.C.
AIRS Monitor ID
1100100251
1100100411
1100100431
Average:
Fourth Daily Maximum 8-Hour
Average (ppm)
2002 2003 2004
0.097 0.079 0.080
0.102 0.082 0.070
0.106 0.081 0.081
0.102 0.081 0.077
Design Value*:
Average of the 3
Year-Specific
Values (ppm)
0.085
0.084
0.089
0.089
*The design value is the maximum of the monitor-specific averages of the annual
fourth daily maximum 8-hour average over the 3 year period.
November 2005
5A-8
Draft - Do Not Quote or Cite
-------
•2
•a
1
o
U
•o
e
41
tJ
a
I
a
.S
8
a
a
i
S
,3
a!
r \
<* L- m
« - B
151
•« -S g
"u
>s
*|-i
§ 1 «
s 1 g
1 *
o 48
*" i
iS
_fl
tft _^
|||
1
•**
g
e
O.
•a
a
o
•-8
u
B
(3
a
I
w
i
£
q •£
i S S
MM
» * e
B 2 M
o •<*
S
1 §
•* a -B
*> R U
a | -
•
1 - October
1
t
J3
4
cs
1
f
S
1
i
i-
•
§
G/
3 «
"S a
« J
u 73
m <
i
- September
1
r
M
4
cs
§
•° f
1 1
« &
C3 ^g
§
1 8
jg_S.
"^f
S
&
^
^3 w
o
i ^
x 2
i
:
- September
1
s?
^
^J.
§
ci
•a
« s .
g1 '3
ra O
ffl -
- September
1
i
^
4
O
O
if
•1 i
u >>
-,
OT Kji
5 -2
o 2
1 *
8 a
i
^
8
•a
"1
w
-------
S «N a
S fi -S
§ -o -5
2 « a
u | g
*§
*
i S j-
. H "O U
S « a
s •§. g
"3 ^
&
1 |
M *
e
111
0 * S
I
.s
f
w
a
K
a
o
•-g
£
K
i
i
.s
1 s §
J3 §«
*" & £ u
2 S «
fl "i kJ
o • y
m Q
K ^J
1
1
1
1
I
•"5
'*
r
_g
^_
CS
i
1
accidental moi
r
g
t
^>
g
ci
e8 yi
^3 :!
!=J 0
» in
ffl ON
1
ft
< w
a
.8
u
(
j
^
1
U
t— >
r
_g
^
cs
!
^
iovascular and
ratory mortalii
*& PI
§ S
i
^.
3
CJ-
^
" S
g) '^
3 o
1 2
i
i
i
1
03
-,
r
^
TJ-
ts
IX,
>p
iovascular and
ratory mortali
"O p.
§ §
|
^.
g
cs
•3
» 1
53 o
E 2
i
1
03
t
Q>
t--,
r
^
^
rs
M
i
iovascular and
ratory mortali
•^3 f\.
1 8
t
Q
•53
1u «
K 2
(
!
x>
!
i
(U
H^
i
J3
4
(N
0
CO
£>
iovascular and
ratory mortali
*O Pi.
§ 1
Ci-
"3
IO tA
U
S ' 'o
(
!
XI
1
i
*
i
^
. 4
tM
8
Lovascular and
iatory mortali
*O Pi
§ a
(
^
8
_ i
cd
« s
r-,5
o
T1 ^
JH •— i
..
0
O
O
1
CO
4)
1
r
1
i
j
g
/— s
§
G'
1 -f
1 ±
•«•
i
o
'1
1
£*•>
1
S
g
.£<
•—
i
f
accidental mo
L
a
i
§
£J,
3 M
1 S
(
i
1
i
"C
•£
r
M
4
(S
u
f
accidental moi
e!
g
-------
PC
s *, s
•a ^ .0
ill
« s g
1
«s"s
'a "d u
1 -a B
.« s £
«rt
1 -
n
is
d
fe 1 "»
5 5 "o
•o 5 S ^
£
•«•*
d
'S
o.
•o
"eS
«
d '
'•S
fa
tf
u
1
.a
g 1 «
5 - E -J
? S 1 1
i ^
o
III
-------
S i- 8
a iS 9
•*•* t^4 •!•*
eg •**
a tJ a
3|I
J
5*1
I I I '
S 1 S
I*
il
S
.3
111
5 1 s
o
b
"S
1
o
a
1
d
e
33
o
1
OH
6
1
I
1 1 «
1 -3 a -J
i a i 8
a S j
o •<;
* 1 i
••B O -Si
•M ** W
«3 (A U
QJi (U **
w Q
3 ^
i
i
b
x>
i
1
1
J
1
1 |
1 §
I!
S 2
•B &
1 a
,
^-*-
g
G-
cd
"S co
3 "o
a -
i
1
I
i
1
g?
•?
^
(X,
it
S3 °
IS
§ S
(
^^
8
Cd
^> 05
x 2
i
1
1
1
?
•?
ts
i"
it
1 i
I!
3 4)
1
^
8
cs
Od
« s
00 -^
a 3
a 2
!
1
w
1
r
CM
0
1}
0 >v
1 1
1 !
i
,^.
8
99
« s
a '3
a 2
!
S
Xi
i
1
i
1
8
11
13 °
1 6
1 |
1 g
1
^3.
8
ci
td
3 '3
a 2
•S
1
1
i
1
oh
1 •
1
1 1
1 1
1. 1
1
^-
1
cd
15
00 -tS
a Q
• i
1
i
1
*
1
I
I
1
'3
i
•
/-.
§
cs
ft
1 2
1
*
a
1 '
1
1
1
!
1
1
§
^.
§
^,^
43 ^
l
-------
I
•a S 1
1 S ?
311
S
111
1*
>• U
1 1
S" S
b — —
III
£
ealth Endpoint
S3
a
"o
0
£
y
•§•
^
&
G **
I •* 1 1
w 2 s S
a S J
0 •<
S
Jill
VI ^
I
- October
1
00
^
cs
s
i
.tl
&
K
S
O
i
1 1
*H M
1
£
- October
'I
00
J3
Si
1
eduled hospital
For pnuemonia
"5 *
6Q _§
3 "1
s
O
I
8 |
1
1
- October
:i
gp
£
i
eduled hospital
For COPD
•S a
6
3 a
|
0
1
I 1
o, E?
53 01
a
•1
1
1
gp
^
i
eduled hospital
for ischemic heart
,6
•§ « S
B*3 ft 3J
3 ea -3
S
O
1 .
i 1
>"~/ S
HH O5
i.
1
- October
i
gp
^
^ ,
s
eduled hospital
For heart failure
•g en
wa S
§ "1
2
o
i
fi *j
S s
1
- October
1
g?
^
s
1
eduled hospital
For dysrhythmias
•S s
§ «
2
o
i
""" "S
8 S
1
1
- October
i
^>
^
s;
i
eduled hospital
For pnuemonia
•S |
9 "i
^
c^
1
8
s
1
1
- October
i
DO
1
S
,
eduled hospital
For COPD
•S |
i "1
2
o
i
0
Cx
3
§
1
1
o
i
I
00
^
s
1
eduled hospital
for ischemic heart
e
3 H vi
3 "1 53
S
$
1
Ci-
^
Wl
-------
°?
S *<• e
•s
A
g
G-
3
CQ '3
1
a
g
>>
5
£
^
fS
1
1
."2
§
/— N
I
£i
I!
0 0
CQ 53
r
1
CO
u
1-1
i
,.}.
ts
!
cardiovascular and
respiratory mortality
p
^^
0
(S
•a
|,|
1 2
j
1
£
CO
l
1
£
^
cs
o
cardiovascular and
respiratory mortality
|
g
Ci
•3
^ u
S o
a 2
1
18
A
I
„
£
4
CS
1
cardiovascular and
respiratory mortality
t
^_^
g
CJ.
"«3
|l
1 2
i
S3
XI
!
•
i
£
4
O
cardiovascular and
respiratory mortality
(
/^s
i
Ci
"3
4-1
|l
i
1
1
00
t
1— >
!
4
O
U
cardiovascular and
respiratory mortality
§
/_^_
0
CJ-
"3
f» |
y
a 2
•a
a
1
1
i
1
i
^r
ts
§
cardiovascular and
respkatory mortality
p
g
Ci
•a
ll
a a
£
^
5
§
1
o
a
VO
PQ
-------
I ! §
§; i
u
0
s 5 ••§
a ^3 u
I -8 §
u
B
! s
III
' £
"1
•e
I
B!
%
I
a ^
« 3
a .3
B -S
»
-g
a
w ^
•a I |
s
I ^
I v
& 2
1
•
1
1
U
1
g
!
T3
'S
3
^ 01
1 I
u
=3
i
I
s
U
ffl
*
a o
ffi 2
t
f
Ul* 11_J
II
s 2
t
f
'•
ro u
8
•a
w
•i
1!
K
S3
.a
I
i
i
1
g
JJ
I
•a |
"S 00
9 •§
-------
OS
S^ 3
IRM O O
•a * s
9 T3 t!
5|l
•> u"
•
a!, m o
-S 53 S-
K PQ O
M £». G
>3 (2 m
1 -s 1
S 1 £
tn
x>
i
u
I/I
|
I
«!
5 .a
II
Sd
1
r
f
avg
24
11
0 s S
.
n.
W
"3 5
11
i
&
« 5
I i o
i ,B I
S
8
•a
u
•a
t>
•s
2 s
u '^
ffl 'o.
a
c3
15
\
ffi ^
1 °
>s ^^
«
0> wi £
* <2 1
2 I
I
-------
as
i
H cj Cj *S
1
I
u
00
1
>-»
f
^
4
ts
1
T» ^
§ 1
1-J ^
*3 ^
2 ^i
cS R
1 1
1
/—v
i
"c3 j-j
-*-* "fcH
u o
M) >H
a z
l|1-»
1 1 i£ d
1
1
X
^
•^
I— I
.1 .
ed g
I" «
•a o
(U -+g
p fa
1 ^
§ «s
•g
•5
1 ^
«M ^
E^ N^-'
of
H 1 1 .
111!
i
1
a
. ^
•3
--:
i
ai
t'
-^
« si
3 6
1-J
•3
•s
1 1
•
.
!
b
i
I
f
jg
,1
CS
1
1
o
"S
•3
u
^
non-ace
>r\
o\
^>
g
^
BeU et al.
cities
1
1
CO
1
i
^
3
i ••
&
•i i
I-J Q
'S fi
s >1
2 &
1 1
tj w
w 1>
1
g
si
"a
•S g
§ 'o
a 2
tH,
B
1
B
<*
"R
1
.-2
cr>
ffl
-------
PS
1
<8 fu ss
1 SI
o * S
"u *
"-1
y _
ajjJ
q TJ U
^ 0 [j^
J *
£ u
5 S
2 1
w
1 ||
"e
B.
^g
d
W
«
u
a.
e
o
••s
d
(2
0$
U
?,
T3
i
.&
« ^ 8 -E
U V9 « 1±
* S 8 |
§ % 2
$
IN
M J
•<
j
i
September
i
1
»— i
«
DO
•?
(S
1
S 3
tovascular
j^toiy mo:
*o PI
«7
O C
§
^j-
8
•3
« S3
5P -r^
3 "o
E •— •
j
!
September
I
i
£
r
1
cs
cs
§
it
1 1
1 1
1 s
t
^
§
ci
3
5P *^
1 2
!
!
September
i
1
i-»
r
I
ts
d
00
Pr*
iovascular
ratory moi
"S °-
b «>
C3 (D
f
^-
8
Si-
's
tJ «
00 *ft
1 2
-
,
\
September
i
i
i->
r
j
CS
O
U
if
iovascular
iratory moj
i
•^j.
g
CN,
13
tS g
r*^
o
1
T3
3
a.
September
i
1
>-.
f
j
r<
1
ll
iovascular
iratory mo:
**3 Cu
8 s
i
^-
8
"3
U
r.
X 1
0]
|
H
PH
1
1
|
1
5
•'f
1
fS
1
0
j
Jj
§
•a
-s
S3
Is
OJD ^T*
ll
i
•i
1
1
i
i
i
.••5
f
j
01
s
H
i
accidental
i
1
•3
tJ
S
"1
bO i^1
§^\
Ov
.
'
!
IH
U
&
3
fS_
"5
^j
^ K
=a .*>
._j:
^
1
i
53
g
M
1
i
"1
%
u
—
<
r
5
CJ
1
i
accidental
j^
g
i
^
8
¥ 1
^i 5
ffl OT
•
,
t
fe
1:
1
1
<.
r
I
«s
I •
i
accidental
jj
§
^
i
/-N
1
5
^n
a -1
v -3
ffl '3
S3
"o
°
i
.2
Q
55
O
^H
PQ
-------
OS
o „
.1 1 1
§ £ '•§
511
T!
,3
I
(A
5 II
"S- -B "w
| | §
£
1 5 •
&• M
N S
U •
.5
u -S —
2 c -S
O 3 S
£
5
•&
A
a
u
1
PS
u
!••
^
.s
I 1 §
III
I ja
s
"S fl
- g .2
•as e :s
j*2 M £J
PH w
^
•& &
' o w
l»-l f!Li
S?
4
ts
g
§
t •
g
1
• i
j
CJ/
•a -s
•s s-
1 ^
-
i
to
1
O
£
3
«s
OJ
g
|
g
1
i
i
i
(S
^—S
•a
? s
T> '-3
m 'S
'
1
O
1
i
i
«t CJ
^ Q
.3
a —
G- ~&
ill- 111
« 3 1 | ^ 8 ^
^1|I?-S
"*« I
o
g
IS
u
I
d
« d
•s-1
•K S
g> S,
a *
13
o
••
* -s .s
-|!1
? 11
S B S
•I ^ "S
6P 05 "J
-------
o
5"
to
tf
o
«
•o
o
U)
d
10
1
«o
V
I
H
1
!!
8
5*
«
o»
I all
S a
S ro
i
S
S
*B
8
l
•O
«
<$
&
u
8.
co
0)
2
ii
o
I
-------
J
o"
o
tc
ID
|
V>
6
&
o
I
}
i
o
rf
it.
s a
f
§o
ii
i
So.
S 15
4U
f
IN
1
fl
]l
ill
•o
o
.£
I
rf
fi
I
*
«»
i^
oS
•D
I
o
3
u
§
o'
8
o
s
8
ci
a.
(j
B
.1
It
E =
w
3
-------
S
i
ii
1
•s
Q
£
rf
£
o
i
g.
M
3
10
01
2
1
99
s
GO
§
0
p
9'
8 5
•2 &
8 .5
8
3
9
is]
?g^
1
o
!*l
8-
d
00
i
d
8
d
s
P
9
It" I
M
Ml
3
6
5
™ -5 S
OT S .2
en
u
O)
I
a
1
-------
!
3
I
i
II
i*l
=!s|
ill
i
S
o
-
(QAM
s
t
A
1
i tr
I:
I:
H
J!
'.i
ac.
0)
-------
O
d
£
S.
t
I
55
u>
«
D
o"
t
f
i
60 to
IE?
sf
n
ii
i.t
sg
«S
3
i.»
••
s
8
o
s-
"O
6
8
o
s
9
3
-
"D
6
§ 1
If I
^11
5| 5
ffl?
8
d
i
1
I
ti
. >•
il
« 8
. J o
F t
I
a
t
I
-------
8
CD
8
Q
8
o
S
I
1
i
8
d
8
C)
8
S
I
a
d
•8
Ci
<
Z
J=
E
II
8
II
r
o
S
8
I
I
_L
•f
^
•»
i
i
cn
<
I
1
fi
1
T5
1
2
,a
1
8
§
g
z
a.
s
y
t
ti
•K (9 •
n (3
111
$
S^
£»
5 s
1 =
S jj
S
01
71
I.
w
3
3
5s
LQ
S
0>
«!
o'
§;
•s
c
rf
£
u
1
S.
I
3
10
»
1
o
i 2 n
CO
D
0)
is|
i
I
-------
u
IB
(O
M
.£
•o
I
o
£
!
S5
£
A
o
o»
U
f
S
S
8 .
II-
8 I a
s S S
i !
1 8«
»fj.
I
1
o"
I
i
•a
I
in
£
If
II
o"
o
S3
I
-------
'53
fr,
-------
INL/EXT-05-Q00767
Atmospheric Mercury
Concentrations near
Salmon Falls Creek
Reservoir - Phase 1
Michael L. Abbott
October 2005
The INL is a U.S. Department of Energy National Laboratory
operated by Battelle Energy Alliance
-------
------- |