&EPA
Uirtsd Stales
Environmental Protection
Agency
Review of the National Ambient Air Quality
Standards for Ozone:


Policy Assessment of Scientific
and Technical Information

OAQPS Staff Paper

-------
                                              EPA-452/R-07-003
                                                 January 2007
Review of the National Ambient Air Quality
Standards for Ozone:

Policy Assessment of Scientific
and Technical Information

OAQPS Staff Paper
                 U.S. Environmental Protection Agency
                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
final OAQPS Staff Paper contains the conclusions and recommendations of staff of 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.

-------
                                 ACKNOWLEDGMENTS

       This Staff Paper is the product of the Office of Air Quality Planning and Standards
(OAQPS).  For the chapters on ozone-related health effects, exposure, risk, and primary
standards, the principal authors include Karen Martin, David McKee, Harvey Richmond, Susan
Lyon Stone, and John Langstaff For the chapters on ozone-related welfare effects and
secondary standards, the principal authors include Vicki Sandiford and Jeffrey Herrick.  The
principal authors of the chapter on air quality characterization include Lance McCluney and
Michael Rizzo. Staff from other EPA offices, including the Office of Research and
Development, the Office of General Counsel, and the Office of Transportation and Air Quality,
also provided valuable comments.
       Earlier drafts of this document were formally reviewed by the Clean Air Scientific
Advisory Committee (CASAC) and made available for public comment.  This document has
been informed by the expert advice and comments received from CASAC, as well as by public
comments submitted by a  number of independent scientists, officials from State and local air
pollution organizations, environmental groups, and industrial groups and companies.

-------
                            Table of Contents



List of Tables	ix

List of Figures	xii

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	1-4
      1.2.4     Current Ozone NAAQS Review	1-5
   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     Properties and Formation	2-1
      2.2.2     Relationship of Ozone to Photochemical Oxidants	2-7
      2.2.3     Transport	2-7
   2.3   DATA SOURCES	2-9
      2.3.1     Air Quality System (AQS)	2-9
      2.3.2     CASTNET	2-10
   2.4   OZONE MONITORING METHODS AND DATA QUALITY	2-12
      2.4.1     Effect of Measurement Precision on 8 hour Ozone Averages	2-12
   2.5   CHARACTERIZATION OF GROUND-LEVEL OZONE
         CONCENTRATIONS	2-13
      2.5.1     Metrics	2-13
      2.5.2     Spatial Variability	2-15
         2.5.2.1   Distributions of 1-hr, 8-hr, and 24-hr Ozone Metrics	2-15
         2.5.2.2   8-Hour and 1-Hour Statistics	2-19
         2.5.2.3   Cumulative Concentration-Weighted Statistics	2-19
      2.5.3     Temporal Variability	2-30
         2.5.3.1   Long Term Variability - Trends	2-30
         2.5.3.2   Short Term Variability - Annual	2-35
         2.5.3.3   Seasonal Variability	2-35
         2.5.3.4   Short Term Variability - Diurnal	2-35

-------
   2.6    CHARACTERIZATION OF OZONE EPISODES	2-39
   2.7    POLICY RELEVANT BACKGROUND LEVELS	2-48
   REFERENCES	2-56

3.     POLICY-RELEVANT ASSESSMENT OF HEALTH EFFECTS
      EVIDENCE	3-1

   3.1    INTRODUCTION	3-1
   3.2    MECHANISMS OF TOXICITY	3-3
   3.3    NATURE OF EFFECTS	3-3
      3.3.1     Morbidity	3-4
         3.3.1.1   Effects on the Respiratory System from Short-term Exposures	3-4
         3.3.1.2   Effects on the Respiratory System from Long-term Exposures	3-20
         3.3.1.3   Effects on the Cardiovascular System	3-26
      3.3.2     Premature Mortality	3-27
         3.3.2.1   Mortality and Short-term Os Exposure	3-27
         3.3.2.2   Mortality and Long-term Os Exposure	3-34
      3.3.3     Ozone Effects on UV-B Flux	3-36
      3.3.4     Summary	3-36
   3.4    ASSESSMENT OF EVIDENCE FROM EPIDEMIOLOGICAL STUDIES	3-37
      3.4.1     Strength of Associations	3-38
      3.4.2     Robustness of Associations	3-39
         3.4.2.1   Exposure Error	3-39
         3.4.2.2   Confounding by Copollutants	3-42
         3.4.2.3   Model Specification	3-44
      3.4.3     Consistency	3-45
      3.4.4     Lag Structure in Short-term Exposure Studies	3-45
      3.4.5     Concentration-Response Relationships and Potential Thresholds	3-46
      3.4.6     Health Effects of Pollutant Mixtures Containing O3	3-48
   3.5    BIOLOGICAL PLAUSIBILITY AND COHERENCE OF EVIDENCE	3-50
      3.5.1     Animal-to-Human Extrapolation Issues	3-51
      3.5.2     Coherence and Plausibility of Short-term Effects on the Respiratory
                System	3-55
      3.5.3     Coherence and Plausibility of Effects on the Cardiovascular System	3-58
      3.5.4     Coherence and Plausibility of Effects Related to Long-Term Os
                Exposure	3-60
      3.5.5     Coherence and Plausibility of Short-Term  Mortality-Related Health
                Endpoints	3-61
   3.6    OZONE-RELATED IMPACTS ON PUBLIC HEALTH	3-62
      3.6.1     Factors that Modify Responsiveness to Ozone	3-63
      3.6.2     Susceptible Population Groups	3-64
         3.6.2.1   Active People	3-64
         3.6.2.2   People with Lung Disease	3-66
         3.6.2.3   Children and Older Adults	3-69
                                      11

-------
         3.6.2.4   People with Increased Responsiveness to Ozone	3-70
         3.6.2.5   Other Population Groups	3-71
       3.6.3    What Constitutes an Adverse Health Impact from Ozone Exposure?	3-72
       3.6.4    Estimation of Potential Numbers of People in At-Risk Susceptible
               Population Groups in the United States	3-78
   3.7    SUMMARY AND CONCLUSIONS FOR OZONE HEALTH EFFECTS	3-79
       3.7.1    Respiratory Morbidity Effects of Short-term Exposures to Ozone	3-80
       3.7.2    Cardiovascular Morbidity Effects of Short-term Exposures to Ozone	3-83
       3.7.3    Mortality-Related Effects of Short-term Exposures to Ozone	3-84
       3.7.4    Health Effects of Repeated Short-term Exposures to Ozone	3-85
       3.7.5    Confidence in Various Health Outcomes Associated with Short-term
               Exposures to Ozone	3-86
       3.7.6    Health Effects of Long-term Exposures to Ozone	3-87
       3.7.7    Health Effects of Pollutant Mixtures Containing Ozone	3-89
       3.7.8    Populations at Risk/Susceptibility Factors Associated with Ozone
               Exposure	3-90
   REFERENCES	3-91

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	4-4
       4.2.3    Personal Ozone Exposure Assessment Studies	4-5
       4.2.4    Microenvironmental Studies	4-5
   4.3    EXPO SURE MODELING	4-6
       4.3.1    The APEX Model	4-6
       4.3.2    Key Algorithms	4-11
       4.3.3    Model Output	4-12
       4.3.4    Strengths and Limitations of the Model	4-13
         4.3.4.1   Estimation of Ambient Air Quality	4-14
         4.3.4.2   Estimation of Concentrations in Indoor Microenvironments	4-15
         4.3.4.3   Characterization of Population Demographics and Activity Patterns	4-16
         4.3.4.4   Modeling Physiological Processes	4-17
   4.4    SCOPE OF EXPOSURE ASSESSMENT	4-18
       4.4.1    Selection of Urban Areas to be Modeled	4-18
       4.4.2    Time Periods Modeled	4-18
       4.4.3    Populations Modeled	4-18
   4.5    INPUTS TO THE EXPOSURE MODEL	4-20
       4.5.1    Population Demographics	4-20
       4.5.2    Population Commuting Patterns	4-21
       4.5.3    Human Activity Data	4-22
       4.5.4    Physiological Data	4-24
                                      in

-------
      4.5.5    Microenvironments Modeled	4-24
         4.5.5.1   Air Exchange Rates for Indoor Residential Environments	4-24
         4.5.5.2   AER Distributions for Other Indoor Environments	4-26
         4.5.5.3   Proximity and Penetration Factors for Outdoors and In-vehicle	4-27
         4.5.5.4   Ozone Decay and Deposition Rates	4-27
      4.5.6    Meteorological Data	4-27
      4.5.7    Ambient Ozone Concentrations	4-28
      4.5.8    Modeling Alternative Standards	4-28
   4.6    MODEL EVALUATION, SENSITIVITY, & UNCERTAINTY ANALYSES	4-30
      4.6.1    Comparison with Exposure Estimates from the Prior Review	4-30
      4.6.2    Comparison of Model Estimates with Measured Personal Exposures	4-31
      4.6.3    Sensitivity Analyses	4-33
         4.6.3.1   Near-Road Residential Exposures	4-33
         4.6.3.2   Air Exchange Rates and Prevalence of Residential Air Conditioning	4-33
         4.6.3.3   Activity Patterns: Representativeness of CHAD	4-34
         4.6.3.4   Activity Patterns: Underestimation of Repeated Exposures	4-34
      4.6.4    Uncertainty Analysis	4-37
      4.6.5    Key Findings	4-41
   4.7    EXPOSURE ASSESSMENT RESULTS	4-43
      4.7.1    APEX Modeling Results	4-43
      4.7.2    Estimated Exposures above Selected Benchmark Levels	4-43
      4.7.3    Estimates of Repeated Exposures	4-64
   REFERENCES	4-68


5.     CHARACTERIZATION OF HEALTH RISKS	5-1

   5.1    INTRODUCTION	5-1
      5.1.1      Overview of Risk Assessment From Last Review	5-2
      5.1.2      Development of Approach for Current Risk Assessment	5-2
   5.2    SCOPE OF  OZONE HEALTH RISK ASSESSMENT	5-5
      5.2.1      Selection of Health Endpoint Categories	5-6
      5.2.2      Selection of Study Areas	5-9
      5.2.3      Air Quality Considerations	5-11
   5.3    COMPONENTS OF THE RISK MODEL	5-14
      5.3.1      Assessment of Risk Based on Controlled Human Exposure Studies	5-14
         5.3.1.1   General  Approach	5-14
         5.3.1.2   Exposure Estimates	5-18
         5.3.1.3   Exposure-Response Functions	5-18
         5.3.1.4   Characterizing Uncertainty and Variability	5-22
      5.3.2      Assessment of Risk Based on Epidemiological  Studies	5-29
         5.3.2.1   General  Approach	5-29
         5.3.2.2   Air Quality Considerations	5-32
         5.3.2.3   Concentration-Response Functions	5-34
         5.3.2.4   Baseline Health Effects Incidence and Population Estimates	5-37
                                      IV

-------
         5.3.2.5   Characterizing Uncertainty and Variability	5-40
   5.4    OZONE RISK ESTIMATES	5-46
      5.4.1       Recent Air Quality	5-47
      5.4.2       Just Meeting Current and Alternative Ozone Standards	5-63
      5.4.3       Sensitivity Analyses	5-81
         5.4.3.1   Impact of Alternative Assumptions About Background	5-80
         5.4.3.2   Impact of Alternative Assumptions About the Shape of Exposure-
                 Response Relationships for Lung Function Decrements	5-82
      5.4.4       Comparison with Risk Estimates from Prior Review	5-87
      5.4.5       Key Observations	5-92
   REFERENCES	5-100

6.     STAFF CONCLUSIONS AND RECOMMENDATIONS ON THE
      PRIMARY O3 NAAQS	6-1
   6.1    INTRODUCTION	6-1
   6.2    APPROACH	6-1
   6.3    PRIMARY O3 STANDARD	6-4
      6.3.1    Adequacy of Current O3 Standard	6-5
         6.3.1.1   Evidence-based Considerations	6-8
         6.3.1.2   Exposure- and Risk-based Considerations	6-17
         6.3.1.3   CASAC and Public Commenters' Views on the Adequacy of the
                 Current Standard	6-42
         6.3.1.4   Staff Conclusions on the Adequacy of the Current Standard	6-46
      6.3.2    Indicator	6-53
      6.3.3    Averaging Time	6-53
         6.3.3.1   Short-Term and Prolonged (1 to 8 Hours)	6-53
         6.3.3.2   Long-Term	6-56
      6.3.4    Level	6-57
         6.3.4.1   Evidence-based Considerations	6-58
         6.3.4.2   Exposure/Risk-based Considerations	6-61
         6.3.4.3   CASAC and Public Commenters' Views on the Level of the
                 Standard	6-76
         6.3.4.4   Staff Conclusions on the Level of the Standard	6-77
      6.3.5    Form	6-82
      6.3.6     Summary of Staff Conclusions and Recommendations on the Primary
               O3 NAAQS	6-85
   6.4    SUMMARY OF KEY UNCERTAINTIES AND RESEARCH
          RECOMMENDATIONS RELATED TO SETTING A PRIMARY O3
          STANDARD	6-87
   REFERENCES	6-92

-------
7.     POLICY-RELEVANT ASSESSMENT OF WELFARE EFFECTS
      EVIDENCE	7-1
   7.1    INTRODUCTION	7-1
   7.2    MECHANISMS GOVERNING PLANT RESPONSE TO OZONE	7-2
      7.2.1     Ozone Uptake: Canopy, Plant and Leaf	7-3
      7.2.2     Cellular to Systemic Response	7-4
      7.2.3     Compensation and Detoxification	7-4
      7.2.4     Changes to Plant Metabolism	7-5
      7.2.5     Plant Response to Chronic/Long-term Os Exposures	7-6
   7.3    NATURE OF  EFFECTS ON VEGETATION	7-6
      7.3.1     Ozone Sensitive Plants and Their Relationship to Public Welfare	7-7
      7.3.2     Vegetation Effects Endpoints	7-7
         7.3.2.1   Visible Foliar Injury and Premature Senescence	7-7
         7.3.2.2   Carbohydrate Production and Allocation	7-8
         7.3.2.3   Growth and Reproduction	7-9
         7.3.2.4   Reduced Plant Vigor	7-10
   7.4    IMPACTS ON PUBLIC WELFARE	7-10
      7.4.1     What Constitutes an Adverse Vegetation Impact from Ozone Exposure?.... 7-10
      7.4.2     Factors That Modify Functional and Growth Response	7-11
         7.4.2.1   Genetics	7-12
         7.4.2.2   Biological Factors	7-12
         7.4.2.3   Physical Factors	7-13
         7.4.2.4   Chemical Factors	7-14
   7.5    CHARACTERIZATION OF VEGETATION EXPOSURES TO OZONE	7-15
      7.5.1     Key  Considerations in Vegetation Exposure Characterization	7-15
      7.5.2     Monitor Networks: National Coverage	7-24
      7.5.3     Community Multi-scale Air Quality Model (CMAQ)	7-26
      7.5.4     Generation of Potential Ozone Exposure Surfaces (POES)	7-26
      7.5.5     Uncertainties in the Os Exposure Analysis	7-34
   7.6    CHARACTERIZATION OF VEGETATION RISKS	7-38
      7.6.1     Scope of Vegetation Risk Assessment	7-38
      7.6.2     Characterization of Crop Risks and Associated Economic Benefits	7-40
         7.6.2.1   Exposure Methodologies Used in Vegetation Research	7-40
         7.6.2.2   Basis for C-R Functions	7-42
         7.6.2.3   Considerations for Exposures at Crop Canopy Height	7-46
         7.6.2.4   Quantifiable Risk of Yield Loss In Select Commodity, Fruit and
                  Vegetable Crops	7-47
         7.6.2.5   Economic Benefits Assessment- AGSEVI	7-49
         7.6.2.6   Uncertainties In the Crop Risk and Benefit Analyses	7-52
      7.6.3     Tree Risk Assessments	7-55
         7.6.3.1   Quantifiable Risk of Biomass Loss In Select Tree  Seedling Species	7-58
         7.6.3.2   Visible Foliar Injury Incidence	7-60
         7.6.3.3   Modeled Mature Tree Growth Response: Eastern and Western
                  Species Case Studies	7-65
                                      VI

-------
         7.6.3.4   Uncertainties In the Tree Risk Analyses	7-69
   7.7   QUALITATIVE RISK: ECOSYSTEM CONDITION, FUNCTION AND
         SERVICES	7-70
      7.7.1    Evidence of Potential Ozone Alteration of Ecosystem Structure and
               Function	7-73
      7.7.2    Effects on Ecosystem Services and Carbon Sequestration	7-74
    REFERENCES	7-76

8.     STAFF CONCLUSIONS AND RECOMMENDATIONS ON THE
      SECONDARY O3 NAAQS	8-1
   8.1   INTRODUCTION	8-1
   8.2   APPROACH	8-1
   8.3   SECONDARY O3 STANDARD	8-4
      8.3.1    Adequacy of Current O3 Standard	8-5
         8.3.1.1   Vegetation Evidence-, Exposure- and Risk-Based Considerations	8-6
         8.3.1.2   CASAC and Public Commenter Views on the Adequacy of the
                 Current Standard	8-13
         8.3.1.3   Staff Conclusions on the Adequacy of the Current Standard	8-14
      8.3.2    Pollutant Indicator	8-15
      8.3.3    Averaging Times	8-16
         8.3.3.1   Seasonal Window	8-16
         8.3.3.2   Diurnal Window	8-17
         8.3.3.3   Alternative Views and Staff Conclusions	8-17
      8.3.4    Form of the Standard	8-18
         8.3.4.1   Comparison of 8-Hour Average and Cumulative Seasonal Forms	8-19
         8.3.4.2   Comparison of SUM06 and W126  Cumulative, Concentration-
                 Weighted Forms	8-21
      8.3.5    Level of the Standard	8-22
      8.3.6    Summary of Staff Conclusions and Recommendations on the Secondary
               O3 Standard	8-24
   8.4   SUMMARY OF KEY UNCERTAINTIES AND RESEARCH
         RECOMMENDATIONS RELATED TO SETTING A SECONDARY O3
         STANDARD	8-27
   REFERENCES	8-30
ATTACHMENT: Clean Air Scientific Advisory Committee Letter (October 24, 2006)
                                    vn

-------
APPENDICES

APPENDIX 2A. PLOTS OF DIURNAL POLICY RELEVANT BACKGROUND
OZONE PATTERNS FOR 12 URBAN AREAS BASED ON RUNS OF THE GEOS-
CHEM MODEL FOR APRIL-OCTOBER 2001	2A-1

APPENDIX 3A. MECHANISMS OF TOXICITY	3A-1

APPENDIX 3B. TABLE OF KEY EPIDEMIOLOGICAL STUDIES	3B-1

APPENDIX 3C. TABLE OF KEY CONTROLLED HUMAN EXPOSURE STUDIES	3C-1

APPENDIX 4A: EXPOSURE TABLES	4A-1

APPENDIX 5A.I: OZONE AIR QUALITY INFORMATION FOR 12 URBAN AREAS... 5A-1

APPENDIX 5A.2: SCATTER PLOTS	5A-10

APPENDIX 5B1: TABLES OF STUDY-SPECIFIC INFORMATION	5B-1

APPENDIX 5B2: CONCENTRATION-RESPONSE FUNCTIONS AND HEALTH
IMPACT FUNCTIONS	5B-8

APPENDIX 5B3: THE CALCULATION OF "SHRINKAGE" ESTIMATES FROM
THE LOCATION-SPECIFIC ESTIMATES REPORTED IN HUANG ET AL. (2004)	5B-11

APPENDIX 5C: ADDITIONAL HEALTH RISK ASSESSMENT ESTIMATES	5C-1

APPENDIX 6A: PREDICTED PERCENT OF COUNTIES WITH MONITORS (AND
PERCENT OF POPULATION IN COUNTIES) NOT LIKELY TO MEET
ALTERNATIVE OZONE STANDARDS	6A-1
                             Vlll

-------
                                List of Tables
Number                                                                          Page

2-1.  NOx Emission Sources, 1970-2004	2-3
2-2.  VOC Emission Sources, 1970-2004	2-5
2-3.  Relationship between Precision of 1-hour Ozone Data and Corresponding Standard
          Deviation of 8-hour Design Values	2-14

3-1.  Acute 03-induced Physiological and Biochemical Changes in Humans and Animals .... 3-52
3-2.  Gradation of Individual Responses to Short-Term Ozone Exposure in Healthy
          Persons	3-74
3-3.  Gradation of Individual Responses to Short-Term Ozone Exposure in Persons with
          Impaired Respiratory Systems	3-75

4-1.  Exertion levels in terms of equivalent ventilation rates (liters/min-m2)	4-2
4-2.  Urban areas and time periods modeled	4-19
4-3.  Population coverage of modeled areas (2002 analysis)	4-20
4-4.  Studies in CHAD used in this analysis	4-23
4-5.  Microenvironments modeled by APEX	4-25
4-6.  Alternative 8-hr ozone standard scenarios	4-29
4-7.  2002-2004 8-hr ozone design values for the modeled areas	4-29
4-8.  Comparison of APEX 2002 base case simulations: All CHAD vs. the NHAPS part
          of CHAD.  Percent of children at moderate exertion with 8-hour exposures
          above levels of 0.06, 0.07, 0.08 ppm-8hr	4-35
4-9.  Comparison of APEX simulations:  All CHAD vs. the NHAPS part of CHAD.
          Percent reduction1 from the 2002 base case to the current standard of the
          number of children at moderate exertion with 8-hour exposures above levels
          of 0.06, 0.07, 0.08 ppm-8hr	4-35
4-10. Comparison of estimated outdoor workers' repeated exposures with APEX results
          for all workers, in Atlanta and Sacramento, 2002. Numbers of people with at
          least six repeated 8-hour exposures above 0.06, 0.07, and 0.08 ppm-8hr	4-36
4-11. Uncertainty of the estimated percent of children exposed at moderate exertion,
          Boston, 2002	4-40
4-12. Uncertainty of the estimated percent of asthmatic children exposed at moderate
          exertion, Boston, 2002	4-40
4-13. Uncertainty of the estimated percent reduction, from the base case to the current
          standard, of all children and asthmatic children exposed at moderate exertion,
          Boston, 2002	4-41

5-1.  Health Effects and Associated Population Groups Addressed in Quantitative Risk
          assessment	5-10
5-2.  Health Endpoints and Associated Population Groups Not Included in the
          Quantitative Risk Assessment	5-10
5-3.  Study-Specific Exposure-Response Data for Lung Function Decrements	5-19
5-4.  Locations, Health Endpoints, and Epidemiological Studies Included in the Os Risk
          Assessment	5-36
                                       IX

-------
5-5. Relevant Population Sizes for Os Risk Assessment Locations	5-38
5-6. Baseline Mortality Rates (per 100,000 Population) Used in the Os Risk Assessment
          	5-39
5-7. Baseline Rates for Hospital Admissions Used in the Oj Risk Assessment	5-41
5-8. Number and Percent of All School Age Children Estimated to Experience Lung
          Function Responses (FEVi > 15%) Associated with 8-Hour Os Exposure
          While Engaged in Moderate Exertion for Location-Specific Os Seasons	5-49
5-9. Number of Occurrences of Lung Function Responses (FEVI > 15%) Among All
          School Age Children Associated with 8-Hour O3 Exposure While Engaged in
          Moderate Exertion for Location-Specific O3 Seasons	5-50
5-10. Number and Percent of Asthmatic School Age Children Estimated to Experience
          Lung Function Responses (FEVi > 10%) Associated with 8-Hour O3
          Exposure While Engaged in Moderate Exertion for Location Specific 63
          Seasons	5-52
5-11. Number of Occurrences of Lung Function Responses (FEVI > 10%) Among
          Asthmatic School Age Children Associated with 8-Hour O3 Exposure While
          Engaged in Moderate Exertion for Location Specific O3 Seasons	5-53
5-12. Estimated Respiratory Symptoms Associated with Recent (April - September,
          2004) 63 Concentrations Above Background in Boston, MA	5-54
5-13. Estimated Respiratory Symptoms Associated with Recent (April - September,
          2002) Os Concentrations Above Background in Boston, MA	5-55
5-14. Estimated Hospital Admissions Associated with Recent (April - September, 2004)
          O3 Concentrations in NY, NY	5-56
5-15. Estimated Hospital Admissions Associated with Recent (April - September, 2002)
          O3 Concentrations in NY, NY	5-56
5-16. Estimated Non-Accidental Mortality Associated with Recent (April - September,
          2004) O3 Concentrations	5-57
5-17. Estimated Non-Accidental Mortality Associated with Recent (April - September,
          2002) O3 Concentrations	5-59

6-la. Summary of Estimates of Number of People Exposed and Number of Occurrences
          at Moderate Exertion Associated with 8-Hour Daily Maximum Ozone
          Concentrations > 0.080 ppm for 12 Urban Areas in the U.S	6-26
6-lb. Summary of Estimates of Number of People Exposed and Number of Occurrences
          at Moderate Exertion Associated with 8-Hour Daily Maximum Ozone
          Concentrations > 0.070 ppm for 12 Urban Areas in the U.S	6-27
6-lc. Summary of Estimates of Number of People Exposed and Number of Occurrences
          at Moderate Exertion Associated with 8-Hour Daily Maximum Ozone
          Concentrations > 0.060 ppm for 12 Urban Areas in the U.S	6-28
6-2. Summary of Number and Percent of All School Age Children (5-18) in  12 Urban
          Areas Estimated to Experience Lung Function Responses and the Number of
          Occurrences Associated with 8-Hour Ozone Exposures While Engaged in
          Moderate Exertion for 2002, 2003, and 2004 Air Quality and Just Meeting the
          Current 8-Hour Standard	6-31
6-3. Summary of Number and Percent of Asthmatic School Age Children (5-18) in 5
          Urban Areas Estimated to Experience Lung Function Responses and the

-------
          Number of Occurrences Associated with 8-Hour Ozone Exposures While
          Engaged in Moderate Exertion for 2002, 2003, and 2004 Air Quality and Just
          Meeting the Current 8-Hour Standard	6-35
6-4.  Incidence of Respiratory Symptom Days for Chest Tightness Associated with
          Recent (2004, 2002) Air Quality and Just Meets the Current Standard Based
          on Adjusting 2004 and 2002 Air Quality in Moderate to Severe Asthmatic
          Children in Boston, MA	6-38
6-5.  Risks of Respiratory- and Asthma-related Hospital Admissions Associated with
          Recent (2004, 2002) Air Quality and Air Quality Adjusted to Just Meets
          Current Standard Based on Adjusting 2004 and 2002 Air Quality in New
          York City, NY	6-40
6-6.  Risks of Non-accidental Mortality Associated with Recent (2004, 2002) Air Quality
          and Air Quality Adjusted to Just Meets Current Standard Based on Adjusting
          2004 and 2002 Air Quality	6-41
6-7.  Estimates of Percent of Children Exposed While at Moderate Exertion to 8-Hour
          Daily Maximum Ozone Concentrations > 0.070 ppm and > 0.060 ppm
          Combined for 12 Urban Areas in the U.S., and the Range of Estimates for
          Each of the 12 Cities - Just Meeting Current Standard	6-51
6-8.  Daily Maximum Ozone Concentrations > 0.07 ppm and > 0.06 ppm Combined for
          12 Urban Areas in the U.S., and the Range of Estimates for Each of the 12
          Cities - Just Meeting Alternative Standards	6-66
6-9.  Risks of Respiratory Symptom Days for Chest Tightness Associated with Just
          Meeting the Current and Alternative Ozone Standards Based on Adjusting
          2002 and 2004 Air Quality in Moderate to Severe Asthmatic Children in
          Boston, MA	6-75
6-10. Risks of Hospital Admissions for Respiratory Illness Associated with Just Meeting
          the Current and Alternative Ozone Standards Based on Adjusting 2002 and
          2004 Air Quality in New York, NY	6-76

7-la. Evaluation statistics for the 3-month 12-hr SUM06 interpolations of the Eastern
          and Western U.S. domains	7-36
7-lb. Evaluation statistics for the 3-month 12-hr W126 interpolations of the Eastern and
          Western U.S.  domains	7-36
7-3 A-B. Agricultural model results with (A) and without (B) a 10% adjustment of hourly
          Os exposures	7-54
7-4.  Percentage and number of counties with visible foliar injury (injured) and without
          injury (not injured) below various standard levels for the years 2001-2004	7-64
7-5.  Relative increase in total annual tree biomass  growth, simulated with the TREGRO
          model, if the level of the current (0.08 ppm) and alternative standards are met	7-66
                                       XI

-------
                               List of Figures

Number                                                                       Page

2-1. Atmospheric Processes Affecting the Formation of Photochemical Oxidants and
          Particulate Matter	2-8
2-2. Locations of Ozone Monitors from AQS and CASTNET	2-11
2-3. 1-hr Ozone Distributions across 12 Risk Areas, 2002-2004	2-16
2-4. 8-hr Ozone Distributions across 12 Risk Areas, 2002-2004	2-17
2-5. 24-hr Ozone Distributions across 12 Risk Areas, 2002-2004	2-18
2-6. Average 2nd Highest Daily Maximum 1-hour Values in U.S. Counties, 2002-2004
          AQS Data	2-20
2-7. Average 4th Highest Daily Maximum 8-hour Values in U.S. Counties, 2002-2004
          AQS Data	2-21
2-8. Highest 3-month 12-hour SUM06 Exposure Index in U.S. Counties, 2001 AQS
          Data	2-22
2-9. Highest 3-month 12-hour W126 Exposure Index in U.S. Counties, 2001 AQS Data	2-23
2-10.  Highest 3-month 12-hour SUM06 Exposure Index in U.S. Counties, 2001
          CASTNET Data	2-24
2-11.  Highest 3-month 12-hour W126 Exposure Index in U.S. Counties, 2001
          CASTNET Data	2-25
2-12. Highest 3-month 12-hour SUM06 Exposure Index in U.S. Counties, 2002 AQS and
          CASTNET Data	2-26
2-13.  Highest 3-month 12-hour Wl 126 Exposure Index in U.S. Counties, 2002 AQS and
          CASTNET Data	2-27
2-14.  Highest 3-month 12-hour SUM06 Exposure Index in U.S. Counties, 2004 AQS
          and CASTNET Data	2-28
2-15. Highest 3-month 12-hour W126 Exposure Index in U.S. Counties, 2004 AQS and
          CASTNET Data	2-29
2-16.  4th Highest Daily Maximum 8-hour Ozone Values 1990-2004 (Urban)	2-31
2-17.  4th Highest Daily Maximum 8-hour Ozone Values 1990-2004 (Rural)	2-32
2-18.  2nd Highest Daily Maximum 1-hour Ozone Values 1990-2004 (Urban)	2-33
2-19. 2nd Highest Daily Maximum 1-hour Ozone Values 1990-2004 (Rural)	2-34
2-20. Comparison of 1-hr, 8-hr, and 24-hr Metrics for 2002 and 2004, 12 Risk Areas	2-36
2-21.  2nd Highest Daily Maximum 1-hour Ozone Values from 2004 by Month	2-37
2-22.  4th Highest Daily Maximum 8-hour Ozone Values from 2004 by Month	2-38
2-23.  1 -Hour Diurnal Week Day Pattern for Urban  Sites, May through  September 2004	2-40
2-24.  1 -Hour Diurnal Week End Pattern for Urban  Sites, May through September 2004	2-41
2-25.  8-Hour Diurnal Week Day Pattern for Urban  Sites, May through  September 2004	2-42
2-26.  8-Hour Diurnal Week End Pattern for Urban  Sites, May through September 2004	2-43
2-27 1 -Hour Week Day Diurnal Pattern for Rural Sites, May through September 2004	2-44
2-28.  1-Hour Week End Diurnal Pattern for Rural Sites, May through September 2004	2-45
2-29.  8-Hour Week Day Diurnal Pattern for Rural Sites, May through September 2004	2-46
2-30. 8-Hour Week End Diurnal Pattern for Rural Sites, May through September 2004	2-47
2-31.  Length of Consecutive Day Episodes over 0.12 ppm by Year for  1-hour Ozone
          Data across  all Monitors	2-49
                                     xn

-------
2-32.  Length of Consecutive Day Episodes over 0.08 ppm by Year for 8-hour Ozone
          Data across all Monitors	2-50
2-33.  Length of Consecutive Day Episodes over Displayed Levels for 1-hour Ozone
          Data (2000-2004) across all Monitors	2-51
2-34.  Length of Consecutive Day Episodes over Displayed Levels for 8-hour Ozone
          Data (2000-2004) across all Monitors	2-52
2-35.  Length of Gaps in Days Between Episodes over 0.08 ppm for 8-hour Ozone Data
          (2000-2004)	2-53

3-lA,B- Frequency distributions of FEVi changes following 6.6-hr exposures to a
          constant concentration of Os or filtered air	3-7
3-1 C.  Frequency distributions of FEVi changes following 6.6-hr exposures to a constant
          concentration of Os or filtered air	3-7
3-2. Resolution Time-Line for Acute Ozone-Induced Physiological and Biochemical
          Responses in Humans	3-53
3-3. Postulated Cellular and Molecular Changes in Human Airway Epithelial Cells on
          Acute Exposure to Ozone	3-54

4-1. APEX Diagram	4-8
4-2. Modeled and measured 6-day average personal exposures, Upland	4-32
4-3. Modeled and measured 6-day average personal exposures, Lake Arrowhead,
     Crestline, and Running Springs	4-32
4-4. Uncertainty of percent of children with exposures above 0.06 ppm-8hr 	4-38
4-5. Uncertainty of percent of children with exposures above 0.07 ppm-8hr 	4-39
4-6. Uncertainty of percent of children with exposures above 0.08 ppm-8hr 	4-39
4-7. Comparison of exposures for different population groups	4-44
4-8. Comparison of exposures for different years	4-44
4-9. Percent of children with repeated exposures > 0.08 ppm-8hr	4-65
4-10.  Percent of children with repeated exposures > 0.07 ppm-8hr	4-66
4-11.  Percent of children with repeated exposures > 0.06 ppm-8hr	4-67

5-1. Major Components of Ozone Health Risk Assessment Based on Controlled Human
          Exposure Studies	5-17
5-2. a, b, c. Probabilistic Exposure-Response Relationships for FEVI Decrement >
          10%,  > 15%, and > 20% for 8-Hour Exposures Under Moderate Exertion	5-24
5-3a, b, c.  Probabilistic Exposure-Response Relationships for FEVI Decrement > 10%
          and > 15% for 8-Hour Exposures Under Moderate Exertion:  Comparison of
          90% Logistic/10% Linear (Hockeystick) Split and 80% Logistic/20% Linear
          (Hockeystick) and 50%  Logistic/50% Linear (Hockeystick) Splits in
          Assumed Relationship Between Exposure and Response	5-27
5-4. Median Exposure-Response Functions Using Three Different Combinations of
          Logistic and Linear (Hockeystick) Models	5-26
5-5. Major Components of Ozone Health Risk Assessment Based on Epidemiological
          Studies	5-31
5-6. Estimated Annual Percent of Non-Accidental Mortality Associated with Short-Term
          Exposure to Recent Os Concentrations Above Background for the Period
                                      Xlll

-------
          April - September (Based on Bell et al., 2004) - Total and Contribution of
          24-Hour Average O3 Ranges	5-62
5-7. Percent of All Children (Ages 5-18) Engaged in Moderate Exertion Estimated to
          Experience At Least One Lung Function Response (Decrement in FEV1 >
          15%) Associated with Exposure to O3 Concentrations That Just Meet the
          Current and Alternative Average 4th-highest Daily Maximum 8-Hour
          Standards, for Location-Specific O3 Seasons (Based on Adjusting 2002 Air
          Quality	5-66
5-8. Percent of All Children (Ages 5-18) Engaged in Moderate Exertion Estimated to
          Experience At Least One Lung Function Response (Decrement in FEVi > 15
          %) Associated with Recent Air Quality (2002) and Just Meeting the Current
          and Alternative Average nth Daily Maximum 8-Hour Standards, for
          Location-Specific O3 Seasons (Based on Adjusting 2002 Air Quality)*	5-67
5-9. Percent of Asthmatic Children (Ages  5-18) Engaged in Moderate Exertion
          Estimated to Experience At Least One Lung Function Response (Decrement
          in FEVi > 10%) Associated with Exposure to O3 Concentrations That Just
          Meet the Current and Alternative Average 4th-highest Daily Maximum 8-
          Hour Standards, for Location-Specific O3 Seasons (Based on Adjusting 2002
          Air Quality)	5-69
5-10. Percent of Asthmatic Children (Ages 5-18) Engaged in Moderate Exertion
          Estimated to Experience At Least One Lung Function Response (Decrement
          in FEVI > 10 %) Associated with Recent Air Quality (2002) and Exposure to
          O3 Concentrations That Just Meet the Current and Alternative 8-Hour
          Standards, for Location-Specific O3 Seasons: Based on Adjusting 2002 O3
          Concentrations	5-70
5-11. Estimated Symptom-Days for Chest Tightness Among Moderate/Severe Asthmatic
          Children (Ages 0 - 12) in Boston Associated with Recent (April-September
          2002) O3 Levels and with Levels Just Meeting Alternative Average 4th-
          Highest Daily Maximum  8-Hour Ozone Standards	5-72
5-12. Estimated Incidence of (Unscheduled) Respiratory Hospital Admissions per
          100,000 Relevant Population in New York Associated with Recent (April -
          September, 2002) O3 Levels and with O3 Levels Just Meeting Alternative
          Average 4th-Highest Daily Maximum 8-Hour Standards	5-73
5-13. Estimated Incidence of Non-Accidental Mortality per 100,000 Relevant Population
          Associated with Recent Air Quality (2002) and with Just Meeting Alternative
          Average 4th-Highest Daily Maximum 8-Hour Ozone Standards (Using Bell et
          al., 2004 - 95 U.S. Cities Function), Based on 2002 Ozone Concentrations	5-74
5-14. Annual Warm Season (April to September) Estimated O3-Related Non-Accidental
          Mortality Associated with Recent (2002) O3 Levels and Levels Just Meeting
          Alternative 8-hr O3 Standards  (Using Bell  et al., 2004 - 95 U.S. Cities
          Function)	5-76
5-15. Estimated Annual Percent of Non-Accidental Mortality Associated with Short-
          Term Exposure to O3 Above Policy Relevant Background for the Period April
          - September When the Current 8-Hour Standard is Just Met (Based on Bell et
          al., 2004) - Total and Contribution of 24-Hour Average O3 Ranges	5-79
                                      xiv

-------
5-16.  Sensitivity Analysis of Estimated Percent Change in (Vrelated Non-Accidental
          Mortality (Using Bell et al., 2004 - 95 Cities) From the Current Standard to
          Alternative 8-hr Standards and a Recent Year of Air Quality, Using Base
          Case, Higher, and Lower PRB Estimates	5-83
5-17  Sensitivity Analysis: Impact of Alternative Estimates of Exposure-Response
          Function on Estimated Percent Changes From the Current Standard in
          Numbers of All Children (Ages 5-18) Engaged in Moderate Exertion
          Experiencing at Least One Decrement in FEVi  >15%	5-88
5-18.  Sensitivity Analysis: Impact of Alternative Estimates of Exposure-Response
          Function on Estimated Percent Changes From the Current Standard in
          Numbers of Asthmatic Children (Ages 5-18) Engaged in Moderate Exertion
          Experiencing at Least One Decrement in FEVi  >10%	5-90

6-1. Percent Changes in Numbers of School Age Children Experiencing at Least One
          Decrement in FEVI >15% when O3 Concentrations are Reduced from Those
          Just Meeting the Current Standard to Those that Would Just Meet Each
          Alternative Standard, Based on Adjusting 2002 Data	6-68
6-2. Percent Changes in Numbers of School Age Children Experiencing at Least One
          Decrement in FEVI >15% when O3 Concentrations are Reduced from Those
          Just Meeting the Current Standard to Those that Would Just Meet Each
          Alternative Standard, Based on Adjusting 2004 Data	6-69
6-3. Percent Changes in Numbers of Asthmatic School Age Children Experiencing at
          Least One Decrement in FEVI >10% when O3 Concentrations are Reduced
          from Those Just Meeting the Current Standard to Those that Would Just Meet
          Each Alternative Standard, Based on Adjusting 2002 Data	6-70
6-4. Percent Changes in Numbers of Asthmatic School Age Children Experiencing at
          Least One Decrement in FEVI > 10% when O3 Concentrations are Reduced
          from Those Just Meeting the Current Standard to Those that Would Just Meet
          Each Alternative Standard, Based on Adjusting 2004 Data	6-71
6-5. Percent Changes in O3-Related Non-Accidental Mortality Incidence  when O3
          Concentrations are Reduced from Those Just Meeting the Current Standard to
          Those that Would Just Meet Each Alternative Standard, Based on Adjusting
          2002 Data (Using Bell et al., 2004 - 95 U.S. Cities)	6-72
6-6. Percent Changes in O3-Related Non-Accidental Mortality Incidence  When O3
          Concentrations are Reduced from Those Just Meeting the Current Standard to
          Those that Would Just Meet Each Alternative Standard, Based on Adjusting
          2004 Data (Using Bell et al., 2004 - 95 U.S. Cities)	6-73

7-1. The 3-year average (2002-2004) of the 4th-highest maximum 8-hr average (current
          standard form) versus the 3-year average of the highest 3-month  12-hr W126,
          by county	7-18
7-2. Highest 3-month 12-hr W126 values from monitors in National Parks and other
          natural areas in the Southeast (A) and Northeast (B). Monitors designated as
          GSMNP are found in different areas of the Great Smoky Mountain National
          Park	7-20
                                      xv

-------
7-3. Highest 3 month 12-hr W126 values from monitors in National Parks in the
          Mountain West (A) and California (B)	7-21
7-4. Maximum 3-month 12-hr SUM06 plotted against maximum 3-month 12-hr W126.
          Data points are from the AQS and CASTNET Oj monitors for the year 2001	7-23
7-5. Locations of AQS monitors (top) and CASTNET monitoring stations (bottom)	7-25
7-6. Estimated 12-hr W126 Ozone Exposure - Max 3-months for 2001:  "As Is" scenario	7-28
7-7. Estimated 12-hr W126 Ozone Exposure - Max 3-months for 2001: Quadratic
          Rollback to just meet 4th-Highest 8-hr Maximum of >0.084	7-30
7-8. Estimated 12-hr W126 Ozone Exposure - Max 3-months for 2001:  Quadratic
          Rollback to just meet 4th Highest 8-hr Maximum of >0.070	7-31
7-9. Estimated 12-hr W126 Ozone Exposure - Max 3-months for 2001:  Quadratic
          Rollback to just meet 12-hr SUM06 of 25 ppm-hr, secondary standard
          proposed in 1996	7-32
7-10. Estimated 12-hr W126 Ozone Exposure - Max 3-months for 2001: Quadratic
          Rollback to just meet 12-hr SUM06 of 15 ppm-hr	7-33
7-11. Comparison of predicted versus observed 12-hr W126 at CASTNET and "rural"
          AQS monitors. Monitor data was predicted by dropping out each monitor
          sequentially and interpolated with the all remaining monitors	7-37
7-12 (a-c).  Major Components of Vegetation Risk Assessment	7-39
7-13. Median crop yield loss from NCLAN crops characterized with the 12-hr W126	7-44
7-14 (A-D). Median soybean (A), wheat (B), cotton (C) and corn (D) yield loss from
          NCLAN crops characterized with the 12-hr W126	7-45
7-15. Estimated soybean yield loss based on interpolated 2001 3-month 12-hr W126 with
          a 10% downward adjustment of hourly O3 concentrations	7-49
7-16. Median tree seedling biomass loss for all 49 cases (A), quaking aspen (B), and
          ponderosa pine (C) characterized with the 12-hr W126	7-56
7-17. Cottonwood (Populus deltoides) shoot biomass (mean ± s.e.) at urban (filled) and
          rural (open) sites in the vicinity of New York City versus ambient Os
          exposure (growing period 12-hr W126,  July 7- Sept. 20)	7-57
                                     xvi

-------
                               1.   INTRODUCTION

1.1  PURPOSE
       This Staff Paper, prepared by staff in the U.S. Environmental Protection Agency's (EPA)
Office of Air Quality Planning and Standards (OAQPS), evaluates the policy implications of the
key studies and scientific information contained in the document, Air Quality Criteria for Ozone
and Related Photochemical Oxidants (USEPA, 2006; henceforth referred to as the CD), prepared
by EPA's National Center for Environmental Assessment (NCEA). This document also presents
and interprets results from several quantitative analyses (e.g., air quality analyses, human
exposure analyses, human health risk assessments, and an environmental assessment of
vegetation-related impacts) that we believe should also be considered in EPA's current review of
the national ambient air quality standards (NAAQS) for ozone (O3), and presents factors relevant
to the evaluation of current primary and secondary 03 standards. Finally, this document presents
staff conclusions and recommendations on a range of policy options that we believe are
appropriate for the Administrator to consider concerning whether, and if so how, to revise the
primary (health-based) and secondary (welfare-based) O3 NAAQS.
       The policy assessment presented in this Staff Paper is intended to help "bridge the gap"
between the scientific assessment contained in the CD and the judgments required of the EPA
Administrator  in determining whether it is appropriate to retain or revise the NAAQS for O3.
This policy assessment considers the available scientific evidence and quantitative risk-based
analyses, together with related limitations and uncertainties, and focuses on the basic elements of
air quality standards: indicator, averaging times, forms,1 and levels.  These elements, which
serve to define each standard, must be considered collectively in evaluating the health and
welfare protection afforded by the O3 standards. Our conclusions and policy recommendations
on whether, and if so how, to revise these standard elements are based on the assessment and
integrative synthesis of information presented in the CD and on staff analyses and evaluations
presented in this document,  and are further informed by comments and advice received from an
independent scientific review committee, the Clean Air Scientific Advisory Committee
(CAS AC),2 in  their review of earlier drafts of this document, as well as comments on earlier
drafts submitted by public commenters.
       1 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.

       2 The statutory requirements for CASAC are discussed below in the next section.
                                           1-1

-------
       While this Staff Paper should be of use to all parties interested in the 63 NAAQS review,
it is written with an expectation that the reader has some familiarity with the technical
discussions contained in the CD.

1.2   BACKGROUND
       1.2.1  Legislative Requirements
       Two sections of the Clean Air Act (Act) govern the establishment and revision of the
NAAQS. Section 108 (42 U.S.C. 7408) directs the Administrator to identify and list "air
pollutants" that "in his judgment, may reasonably be anticipated to endanger public health and
welfare" and whose "presence ... in the ambient air results from numerous or diverse mobile or
stationary sources" and, if listed, to issue air quality criteria for them. These air quality criteria
are intended to "accurately reflect the latest scientific knowledge useful in indicating the kind
and extent of identifiable effects on public health or welfare which may be expected from the
presence of [a] pollutant in ambient air ... ."
       Section 109 (42 U.S.C. 7409) directs the Administrator to propose and promulgate
"primary" and "secondary" NAAQS  for pollutants identified under section 108.  Section
109(b)(l) defines a primary standard as one "the attainment and maintenance of which in the
judgment of the Administrator, based on such criteria and allowing an adequate margin of safety,
are requisite to protect the public health."3 A secondary standard, as defined in Section
109(b)(2), must "specify a level of air quality the attainment and maintenance of which, in the
judgment of the Administrator, based on such criteria, is requisite to protect the public welfare
from any known or anticipated adverse effects associated with the presence of [the] pollutant in
the ambient air."4
       The requirement that primary standards include an adequate margin of safety was
intended to  address uncertainties associated with inconclusive scientific and technical
information available at the time of standard setting. It was also intended to provide a reasonable
degree of protection against hazards that research has not yet identified.  Lead Industries
Association v. EPA, 647 F.2d 1130, 1154 (D.C. Cir 1980), cert, denied.  101 S. Ct. 621 (1980);
       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, 91st Cong., 2d Sess. 10 (1970)].

       4 Welfare effects as defined in section 302(h) [42 U.S.C. 7602(h)] include, but are not limited to, "effects
on soils, water, crops, vegetation, man-made materials, animals, wildlife, weather, visibility and climate, damage to
and deterioration of property, and hazards to transportation, as well as effects on economic values and on personal
comfort and well-being."
                                             1-2

-------
American Petroleum Institute v. Costle, 665 F.2d 1176, 1186 (D.C. Cir. 1981), cert, denied. 102
S.Ct.  1737 (1982). Both kinds of uncertainties are components of the risk associated with
pollution at levels below those at which human health effects can be said to occur with
reasonable scientific certainty. Thus, in selecting primary standards that include an adequate
margin of safety, the Administrator is seeking not only to prevent pollution levels that have been
demonstrated to be harmful but also to prevent lower pollutant levels that may pose an
unacceptable risk of harm, even if the risk is not precisely identified as to  nature or degree.
       In selecting a margin of safety, the EPA considers  such factors as the nature and severity
of the health effects, the size of the sensitive population(s) at risk, and the kind and degree of the
uncertainties that must be addressed. The selection of any particular approach to providing an
adequate margin of safety is a policy choice left  specifically to the Administrator's judgment.
Lead Industries Association v. EPA, supra, 647 F.2d at 1161-62.
       In setting standards that are "requisite" to protect public health and welfare, as provided
in section 109(b), EPA's task is to establish standards that are neither more nor less stringent
than necessary for these purposes. In so doing, EPA may not consider the costs of implementing
the standards.  See generally Whitman v. American Trucking Associations, 531 U.S. 457, 464,
475-76 (2001).
       Section  109(d)(l) of the Act requires that "not later than December 31, 1980, and at 5-
year intervals thereafter, the Administrator shall  complete a thorough review of the criteria
published under section 108 and the national ambient air quality  standards . . . and shall make
such revisions in such criteria and standards and promulgate such new standards as may be
appropriate . . .  ." Section 109(d)(2) requires that an independent scientific review committee
"shall complete a review of the criteria . . .  and the national primary and secondary ambient air
quality standards . . . and shall recommend to the Administrator any new . . . standards and
revisions of existing criteria and standards as may be appropriate . . . ."  Since the early 1980's,
this independent review function has been performed by the Clean Air Scientific Advisory
Committee (CASAC), a standing committee of EPA's Science Advisory Board.

       1.2.2  History of Ozone NAAQS  Reviews
       Tropospheric (ground-level) 63 is formed from biogenic precursor emissions and as a
result of anthropogenic precursor emissions. Naturally occurring Os in  the troposphere can result
from biogenic organic precursors reacting with naturally occurring nitrogen oxides (NOX) and by
stratospheric 63 intrusion into the troposphere. Anthropogenic precursors of 63,  specifically
NOX and volatile organic compounds (VOC), originate from a wide variety of stationary and
mobile sources. Ambient 63 concentrations produced by these emissions are directly affected by
temperature, solar radiation, wind speed and other meteorological factors.
                                           1-3

-------
       The EPA initially established primary and secondary NAAQS for photochemical
oxidants on April 30, 1971 (36 FR 8186). Both primary and secondary standards were set at an
hourly average of 0.08 parts per million (ppm), total photochemical oxidants, not to be exceeded
more than one hour per year.
       On February 8, 1979, EPA completed its first periodic review of the criteria and
standards for O3 and other photochemical oxidants (44 FR 8202). In that action, EPA made
significant revisions to the original standard:  the level of the primary and secondary NAAQS
was changed to 0.12 ppm; the indicator was changed to O3; and the form  of the standards was
changed to be based on the expected number of days per calendar year with a maximum hourly
average concentration above 0.12 ppm (i.e., attainment of the standard occurs when that number
is equal to or less than one).
       On March 9, 1993, EPA concluded its second periodic review of the criteria and
standards for O3 by deciding that revisions to the O3 NAAQS were not warranted at that time (58
FR 13008).  The timing of this decision was required by a court order issued to resolve a lawsuit
filed to compel EPA to complete its review of the criteria and standards for O3 in accordance
with the Act. This decision reflected EPA's review of relevant scientific and other information
assembled since the last review, as contained in the 1986 O3 CD (U.S. EPA, 1986), its
Supplement (U.S. EPA, 1992), and the 1989 O3 Staff Paper (U.S. EPA, 1989), although it did not
take  into consideration a large number of studies on the health and welfare effects of O3
published  since the literature was last assessed in the O3  Supplement. The final decision
emphasized the Administrator's intention to proceed as rapidly as possible with the next periodic
review of the air quality criteria and  standards to consider the more recent information.
       Under a court-ordered schedule and a highly accelerated review process, EPA completed
its third review of the O3 NAAQS on July 18, 1997, based on the 1996 O3 CD (U.S. EPA, 1996a)
and 1996 O3 Staff Paper (U.S. EPA,  1996b). EPA revised the primary and secondary O3
standards on the basis of the then latest scientific evidence linking exposures to ambient O3 to
adverse health and welfare effects at levels allowed by the 1-hr average standards (62 FR 38856).
The O3 standards were revised by replacing the existing primary 1-hr average standard with an 8-
hr average O3 standard set at a level of 0.08 ppm.  The form of the primary standard was changed
to the annual fourth-highest daily maximum 8-hr average concentration, averaged over three
years.  The secondary O3 standard was changed by making it identical in  all respects to the
revised primary standard.

       1.2.3  Litigation  Related to the 1997 Ozone Standards
       Following promulgation of the revised O3 NAAQS, petitions for review were filed
addressing a broad range of issues. On May 14, 1999, in response to those challenges, the U.S.
                                          1-4

-------
Court of Appeals for the District of Columbia Circuit (D.C. Circuit Court) remanded the 63
NAAQS to EPA, finding that section 109 of the Act, as interpreted by EPA, effected an
unconstitutional  delegation of legislative authority.5 In addition, the D.C. Circuit Court directed
that EPA should consider the potential beneficial health effects of 63 pollution in shielding the
public from the effects of solar ultraviolet (UV) radiation, as well as the adverse health effects.
        EPA petitioned the U.S. Supreme Court for certiorari on the constitutional issue but did
not request review of the D.C. Circuit Court ruling regarding its obligation to consider the
potential beneficial health effects of Os. On February 27, 2001, the U.S. Supreme Court
unanimously reversed the judgment of the D.C. Circuit Court on the constitutional issue, holding
that section 109 of the CAA does not delegate legislative power to the EPA in contravention of
the Constitution, and remanded the case to the D.C. Circuit Court to consider those challenges to
the 63 NAAQS that had not been addressed by that Court's earlier decisions.6  On March 26,
2002, the D.C. Circuit Court issued its final decision, finding the 1997 O3 NAAQS to be "neither
arbitrary nor capricious," and denying the remaining petitions for review.7
        In response to the D.C. Circuit Court's remand to consider the potential beneficial health
effects of Os pollution in shielding the public from the effects of solar (UV) radiation, on
November 14, 2001, EPA proposed to leave the 1997 8-hr NAAQS unchanged (66 FR 52768).
After considering public comment on the proposed decision, EPA reaffirmed the 8-hr Os
NAAQS set in 1997 (68 FR 614).  Finally,  on April 30, 2004, EPA announced the decision to
make the 1-hr 63 NAAQS no longer applicable to areas one year after the effective date of the
designation of those areas for the 8-hr NAAQS (69 FR 23966). For most areas, the date that the
1-hr NAAQS no longer applied was June 15, 2005. (See 40 CFR 50.9 for details.)

       1.2.4   Current Ozone NAAQS Review
       EPA initiated the current NAAQS review in September 2000 with a call for information
(65 FR 57810). A project work plan (U.S. EPA, 2002) for the preparation of the CD was
released in November 2002 for CAS AC and public review. EPA held a series of workshops in
mid-2003  on several draft chapters of the CD to obtain broad input from the relevant scientific
communities.  These workshops helped to inform the preparation of the first draft CD (U.S.
EPA, 2005a), which was released for CASAC and public review on January  31, 2005.
       5 American Trucking Associations v. EPA, 175 F.3d 1027 (D.C. Cir, 1999).

       6 Whitman v. American Trucking Associations, 531 U.S. 457 (2001).

       7 Whitman v. American Trucking Associations, 283 F.3d 355 (D.C. Cir. 2002).
                                           1-5

-------
       During the process of preparing the first draft CD, NCEA revised the planned format of
the CD described in the 2002 work plan.  These revisions were made as part of a collaborative
effort with OAQPS staff to modify the review process so as to enhance the Agency's ability to
meet this and future NAAQS review schedules. As described in Chapter 1 of the first draft CD,
emphasis is placed on interpretative evaluation and integration of evidence in the main body of
the document, with more detailed descriptions of individual studies being provided in a series of
accompanying annexes. This change is intended to streamline the document so as to facilitate
timely CASAC and public review and to focus more clearly on issues most relevant to the policy
decisions to be made by the Administrator. The modified review process envisions that key
policy-relevant issues will be identified earlier in the review process through enhanced
collaboration between NCEA and OAQPS, leading to a more efficient linkage between the CD
and the Staff Paper. At the CASAC meeting held on May 4-5, 2005, to review the first draft CD,
this new format for the CD was met with  general approval of CASAC and the public.  A second
draft CD  (EPA, 2005b) was released for CASAC and public review on August 31, 2005.  In a
February 16, 2006 letter to the Administrator, CASAC offered final comments on all chapters of
the CD (Henderson, 2006a),  and the final CD was released on March 21, 2006. In a June 8, 2006
letter to the Administrator, CASAC offered additional advice to the Agency concerning Chapter
8 of the final CD (Integrative Synthesis) to help inform the second draft Staff Paper (Henderson,
2006b). The second draft Staff Paper was released on July 17, 2006 and reviewed by the
CASAC Ozone Panel on August 24 and 25, 2006.  In an October 24, 2006 letter to the
Administrator, CASAC provided advice and recommendations to the Agency concerning the
second draft Staff Paper (Henderson, 2006c).  Advice and recommendations from CASAC as
well as public comments have been taken into account in preparing this final Staff Paper.
       The schedule for completion of this review is governed by a consent decree resolving a
lawsuit filed in March 2003 by a group of plaintiffs representing national environmental
organizations, alleging that EPA had failed to complete the current review within the period
provided by statute. American Lung Association v.  Whitman (No. 1:03CV00778, D.D.C. 2003).
The modified consent decree that now governs this review, entered by the court on December 16,
2004, provides that EPA sign for publication notices of proposed and final rulemaking
concerning its review of the O3 NAAQS no later than March 28, 2007 and December 19, 2007,
respectively. This consent decree was further modified in October 2006 to change these
proposed and final rulemaking dates to no later than May 30, 2007 and February 20, 2008,
respectively. The EPA expects that these dates for signing the publication notices of proposed
and final  rulemaking will now be extended to no later than June 20, 2007 and March 12, 2008,
respectively.
                                         1-6

-------
1.3   GENERAL APPROACH AND ORGANIZATION OF THE DOCUMENT
       The policy assessment in this Staff Paper is based on staffs evaluation of the policy
implications of the scientific evidence contained in the CD and results of quantitative analyses
based on that evidence, as well as the views presented by CASAC and various stakeholders.
Taken together, this information informs various conclusions and the identification of a range of
options on certain elements of the O?, standards under review. While the CD focuses on new
scientific information available  since the last review, it appropriately integrates that information
with scientific criteria from previous reviews.  The quantitative analyses presented in this Staff
Paper (and described in more detail in technical support documents) are based on the most
recently available air quality information, so as to provide current characterizations of O3 air
quality patterns and estimated health and environmental risks related to exposure to ambient Os
concentrations.
       Following this introductory chapter, the Staff Paper is organized into three main parts: the
characterization of ambient Os air quality data; Os-related health effects and primary Os
NAAQS; and (Vrelated welfare effects and secondary 63 NAAQS.  The content of these parts is
discussed more fully below.
       The characterization of ambient 63 and related photochemical oxidants is presented in
Chapter 2 and includes information on O3 properties, current O3 air quality patterns, historic
trends, and background levels. This chapter provides a frame of reference for subsequent
discussion of current and alternative 63 NAAQS and alternative forms of 63 standards.
       Chapters 3 through 6 comprise the second main part of this Staff Paper dealing with
human health and primary standards. Chapter 3 presents an overview of key policy-relevant
health effects evidence, major health-related conclusions from the CD, and an examination of
issues related to the quantitative assessment of evidence from controlled human exposure and
epidemiological studies.  Chapters 4 and 5 describe the scope and methods used in conducting
human exposure and health risk assessments and present results from those assessments.
Chapter 6 includes staff conclusions and policy recommendations on the adequacy of the current
primary standard and on an appropriate range of alternative primary standards for the
Administrator's consideration, together with a discussion of the science and public health policy
judgments underlying such standards.
       Chapters 7 and 8 comprise the third main part of this Staff Paper. Chapter 7 presents a
policy-relevant assessment of Os welfare effects evidence and discusses the scope and methods
used in conducting vegetation-related exposure and risk assessments. Chapter 8 includes staff
conclusions and policy recommendations on the adequacy of the current secondary standard and
on an appropriate range of alternative secondary standards that for the Administrator's
                                           1-7

-------
consideration, together with a discussion of the science and public welfare policy judgments
underlying such standards.

-------
REFERENCES


Federal Register (1971) National Primary and Secondary Ambient Air Quality Standards for
        Photochemical Oxidants; Final rule. 40 CFR 50; Federal Register 36: 8186.

Federal Register (1979) National Primary and Secondary Ambient Air Quality Standards: Revisions to the National
        Ambient Air Quality Standards for Photochemical Oxidants, Final Rule. 40 CFR 50; Federal Register
        44: 8202.

Federal Register (1993) National Ambient Air Quality Standards for Ozone, Final rule. 40 CFR 50; Federal Register
        58: 13008.

Federal Register (1997) National Ambient Air Quality Standards for Ozone; Final Rule. 40 CFR 50; Federal
        Register 62: 38856.

Federal Register (2001) National Ambient Air Quality Standards for Ozone; Proposed Response to Remand;
        Proposed Rule.  Federal Register 66: 57368.

Federal Register (2003) National Ambient Air Quality Standards for Ozone; Proposed Response to Remand. Final
        Rule. Federal Register 68: 614.

Henderson, R. (2006a) Letter from CASAC Chairman Rogene Henderson to EPA Administrator Stephen Johnson,
        February 16, 2006, EPA-CASAC-06-003.

Henderson, R. (2006b) Letter from CASAC Chairman Rogene Henderson to EPA Administrator Stephen Johnson,
        June 5, 2006, EPA-CASAC-06-007.

Henderson, R. (2006c) Letter from CASAC Chairman Rogene Henderson to EPA Administrator Stephen Johnson,
      October 24, 2006, EPA-CASAC-07-001.

U.S. Environmental Protection Agency. (1986) Air Quality Criteria for Ozone and Other Photochemical Oxidants.
        Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and
        Assessment Office; EPA report nos. EPA-600/8-84-020aF-eF. Available from NTIS, Springfield, VA;
        PB87-142949.

U.S. Environmental Protection Agency (1992) Summary of selected new information on effects of ozone on health
        and vegetation: supplement to 1986 air quality criteria for ozone and other photochemical oxidants.
        Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and
        Assessment Office; EPA report no. EPA/600/8-88/105F.  Available from NTIS, Springfield, VA; PB92-
        235670.

U.S. Environmental Protection Agency (1996) Air Quality Criteria for Ozone and Related Photochemical Oxidants.
        Research Triangle Park, NC: Office of Research and Development; report nos. EPA/600/AP-93/004aF-cF.
        3v. Available from: NTIS, Springfield, VA;  PB96-185582, PB96-185590, and PB96-185608. Available
        online at: http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=2831.

U.S. Environmental Protection Agency (2002) Project Work Plan for Revised Air Quality
        Criteria for Ozone and Related Photochemical Oxidants. Research Triangle Park, NC: National Center for
        Environmental Assessment-RTF Report no. NCEA-R-1068.

U.S. Environmental Protection Agency (2005a) Air Quality Criteria for Ozone and Related
        Photochemical Oxidants (First External Review Draft). Washington, DC, EPA/600/R-05/004aA-cA.
        Available online at: www.epa.gov/ncea/
                                                 1-9

-------
U.S. Environmental Protection Agency (2005b) Air Quality Criteria for Ozone and Related Photochemical Oxidants
        (Second External Review Draft). Washington, DC, EPA/600/R-05/004aB-cB. Available online at:
        www. epa. gov/ncea/

U.S. Environmental Protection Agency (2006) Air Quality Criteria for Ozone and Related
        Photochemical Oxidants (Final). Washington, DC, EPA/600/R-05/004aB-cB. Available online at:
        http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=137307
                                                 1-10

-------
                     2.   AIR QUALITY CHARACTERIZATION

2.1  INTRODUCTION
       This chapter generally characterizes ambient ozone (O^) and related photochemical
oxidants in terms of measurement methods, recent concentrations and trends, relationships
between different air quality indicators, and estimates of policy-relevant background. This
information is useful for interpreting the available exposure, health, and welfare effects
information, and for evaluating the adequacy of the current primary and secondary standards for
Os and developing options for alternative standards. The information presented in this chapter
was drawn from the 2006 Criteria Document (CD) and additional analyses of data from various
Os monitoring networks.
       This chapter particularly focuses on 1-hr, 8-hr, and 24-hr average concentrations metrics
in characterizing urban 63 air quality because these are the metrics most frequently used in the
health effect studies discussed in the CD and Chapter 3 of this Staff Paper.  For the vegetation
exposure and risk assessment discussed in Chapter 7 of this Staff Paper, both the current
secondary standard 8-hr, metric and the cumulative, concentration-weighted metrics, SUM06
and W126 are used.
       Although this chapter focuses on 2002-2004 air quality data in order to be consistent with
the CD, recent observations demonstrates the 2005 Os data show that national ambient
concentrations have decreased 20% since 1980 and 8% since 1990.  2005 concentrations were
the second lowest on record with only 2004 levels being lower by 5%. Meteorological
conditions for 2005 were similar to those observed in 2002 which was much hotter than 2004
and, therefore, more conducive to 63 formation.  Furthermore, the 63 levels in 2005 were
approximately 9% lower than those seen in 2002 for sites east of 100 degrees west longitude.
One explanation for this difference can possibly be attributed to the implementation of the NOX
SIP Call1 which occurred in 2002 for many states east of the Mississippi River.

2.2  CHEMICAL AND PHYSICAL PROPERTIES, FORMATION, AND TRANSPORT
       2.2.1  Properties  and Formation
       The atmosphere can be divided into several distinct vertical layers, based primarily on the
major mechanisms by which they are heated and cooled. The lowest major layer is the
troposphere, which extends from the earth's surface to about 8 km above the surface in polar
       1 EPA's rule, known as the NOX SIP Call, was designed to reduce regional transport of O3 and O3-forming
pollutants in the eastern half of the United States by requiring 21 states to reduce O3- season NOX emissions that
contribute to nonattainment in other states.
                                           2-1

-------
regions and to about 16 km above the surface in tropical regions. The planetary boundary layer
(PEL) is the lower sub-layer of the troposphere, extending from the surface to about 1 or 2 km,
and is most strongly affected by surface conditions. The stratosphere extends from the top of the
troposphere, to about 50 km in altitude. The emphasis in this chapter is placed on concentrations
of Os occurring in the lower troposphere, in particular in the PEL (CD, p.2-1).
       Ozone chemistry in the presence of sunlight, nitrogen oxides (NOX) and volatile organic
carbon (VOC) is well understood and a central component of modern air quality models. The
chemical formation of Os in the troposphere results from the oxidation of nitric oxide (NO) to
nitrogen dioxide (NO2) by organic (R£>2) or hydro-peroxy (HO2) radicals. Photolysis (the
chemical process of breaking down molecules into smaller units through the absorption of light)
of NO2 yields NO and a ground-state oxygen atom, O(3P), which then reacts with molecular
oxygen to form O3 (CD, p.2-2).
       Oxidized nitrogen compounds are emitted to the atmosphere mainly as NO, which is
oxidized to NO2 which subsequently can be reduced back to NO.  Consequently, NO and NO2 are
often grouped together into their own family called NOX (CD, p.2-3).  Oxidized nitrogen
containing compounds are essential to the formation of Os in the air. There are a large number of
oxidized nitrogen containing compounds in the atmosphere including NO, NO2, nitrate (NOs),
nitrous acid (HNO2), nitric acid (HNOs), nitrogen pentoxide (TS^Os), pernitric acid (HNO4),
peroxy acetyl nitrate (PAN) and its homologues, other organic nitrates and particulate nitrate.
Collectively these species are referred to as NOy. NOX is considered a good surrogate for NOy
and, thus,  is commonly monitored and reported (see Table 2-1).
       In  urban areas, both biogenic and anthropogenic VOCs  are important for Os formation.
Table 2-2  lists a variety of VOC sources (see http://www.epa.gov/airtrends/econ-
emissions.html).  The categories in the table are self explanatory with the exception of the fires
and miscellaneous categories.  The fires category includes both wild fires and prescribed burns.
The miscellaneous category includes mainly structural fires  and sources from agricultural
activities.  One category not in either table due to insufficient estimates is biogenic emissions. As
can be seen in the table, highway vehicles have been the single largest source of anthropogenic
VOC emissions over the years ranging from about 49% of total emissions in 1970 to about 27%
of total emissions in 2004. Starting in 2001, solvent use and highway vehicles were the two
main sources of VOCs with roughly equal contributions to the total  emissions.
       In  non-urban, vegetated areas, biogenic VOCs emitted from  the vegetation tend to be the
most important. In the remote troposphere,  CH4 and  CO are the main carbon-containing
precursors to O3 formation. In coastal environments  and other selected environments, atomic Cl
and Br radicals can also initiate the oxidation of VOCs (CD, p.2-2 and 2-3).
                                           2-2

-------
Table 2-1. NOx Emission Sources, 1970-2004

                                                             Nitrogen Oxides (NOx)
                                                    National Emissions Totals (thousands of tons)

Source Category                       1970     1975     1980     1985     1990      1991     1992     1993     1994
FUEL COMB. ELEC. UTIL.                4,900     5,694     7,024    6,127    6,663     6,519     6,504     6,651    6,565
FUEL COMB. INDUSTRIAL                4,325     4,007     3,555    3,209    3,035     2,979     3,071     3,151    3,147
FUEL COMB. OTHER                     836      785      741      712    1,196     1,281     1,353     1,308    1,303
CHEMICAL & ALLIED PRODUCT MFC       271      221      213      262      168       165      163      155      160
METALS PROCESSING                    77       73       65       87       97       76       81       83       91
PETROLEUM & RELATED INDUSTRIES      240       63       72      124      153       121      148      123      117
OTHER INDUSTRIAL PROCESSES          187      182      205      327      378       352      361      370      389
SOLVENT UTILIZATION                     000212333
STORAGE & TRANSPORT                   000236555
WASTE DISPOSAL & RECYCLING          440      159      111       87       91       95       96      123      114
HIGHWAY VEHICLES                   12,624    12,061    11,493   10,932    9,592     9,449     9,306     9,162    9,019
OFF-HIGHWAY                         2,652     2,968     3,353    3,576    3,781     3,849     3,915     3,981    4,047
MISCELLANEOUS                       330      165      248      310      369       286      255      241      390
MISCELLANEOUS                   NA      NA      NA      NA       NA      NA      NA      NA      NA

TOTAL                              26,883    26,377    27,079   25,757   25,529    25,179    25,260    25,357   25,349
Total without FIRES                     26,883    26,377    27,079   25,757   25,167    24,932    25,026    25,123   24,967
FIRES                                  NA       NA       NA      NA      362       247      234      234      382
                                                       2-3

-------
Table 2-1. NOx Emission Sources, 1970-2004 (cont'd)
Source Category
FUEL COMB. ELEC. UTIL.
FUEL COMB. INDUSTRIAL
FUEL COMB. OTHER
CHEMICAL & ALLIED PRODUCT MFC
METALS PROCESSING
PETROLEUM & RELATED INDUSTRIES
OTHER INDUSTRIAL PROCESSES
SOLVENT UTILIZATION
STORAGE & TRANSPORT
WASTE DISPOSAL & RECYCLING
HIGHWAY VEHICLES
OFF-HIGHWAY
MISCELLANEOUS
MISCELLANEOUS

TOTAL
FIRES
Total without FIRES
                                                         Nitrogen Oxides (NOX)
                                                 National Emissions Totals (thousands of tons)
1995
6,384
3,144
1,298
158
98
110
399
3
6
99
8,876
4,113
267
NA
24,956
258
24,698
1996
6164
3151
1197
125
83
139
433
2
15
153
8733
4179
412
0
24787
405
24,382
1997
6276
3101
1177
127
89
143
460
3
16
157
8792
4178
187
0
24705
179
24,526
1998
6232
3050
1101
129
89
143
467
3
16
163
8619
4156
179
0
24348
172
24,176
1999
5721
2709
768
102
86
120
451
4
14
162
8371
4084
251
0
22845
236
22,609
2000
5330
2723
766
105
89
122
479
4
15
129
8394
4167
276
0
22598
263
22,335
2001
4917
2757
779
107
94
124
504
4
16
130
7774
4156
184
0
21549
171
21,378
2002
4699
2870
725
105
84
149
487
8
16
152
7365
4086
356
0
21102
341
20,761
2003
4270
2870
725
105
84
149
487
8
16
152
7365
4086
356
0
20672
341

2004
3740
2870
725
105
84
149
487
8
16
152
7365
4086
356
0
20142
341

                                                                                                  20,331    19,801
                                                    2-4

-------
Table 2-2. VOC Emission Sources, 1970-2004
                                         Volatile Organic Compounds (VOC)
                                         National Totals (thousands of tons)
 Source Category
 FUEL COMB. ELEC. UTIL.
 FUEL COMB. INDUSTRIAL
 FUEL COMB. OTHER
 CHEMICAL & ALLIED PRODUCT MFC
 METALS PROCESSING
 PETROLEUM & RELATED INDUSTRIES
 OTHER INDUSTRIAL PROCESSES
 SOLVENT UTILIZATION
 STORAGE & TRANSPORT
 WASTE DISPOSAL & RECYCLING
 HIGHWAY VEHICLES
 OFF-HIGHWAY
 MISCELLANEOUS

 TOTAL
 FIRES
 Total without FIRES
1970
30
150
541
1,341
394
1,194
270
7,174
1,954
1,984
16,910
1,616
1,101
34,659
917
33,742
1975
40
150
470
1,351
336
1,342
235
5,651
2,181
984
15,392
1,917
716
30,765
587
30,178
1980
45
157
848
1,595
273
1,440
237
6,584
1,975
758
13,869
2,192
1,134
31,106
1,024
30,082
1985
32
134
1,403
881
76
703
390
5,699
1,747
979
12,354
2,439
566
27,404
465
26,939
1990
47
182
776
634
122
611
401
5,750
1,490
986
9,388
2,662
1,059
24,108
983
23,125
1991
44
196
835
710
123
640
391
5,782
1,532
999
8,860
2,709
756
23,577
678
22,899
1992
44
187
884
715
124
632
414
5,901
1,583
1,010
8,332
2,754
486
23,066
407
22,659
1993
45
186
762
701
124
649
442
6,016
1,600
1,046
7,804
2,799
556
22,730
478
22,252
1994
45
196
748
691
126
647
438
6,162
1,629
1,046
7,277
2,845
720
22,569
638
21,931
                                                   2-5

-------
Table 2-2. VOC Emission Sources, 1970-2004 (cont'd)
                                         Volatile Organic Compounds (VOC)
                                         National Totals (thousands of tons)
 Source Category
 FUEL COMB. ELEC. UTIL.
 FUEL COMB. INDUSTRIAL
 FUEL COMB. OTHER
 CHEMICAL & ALLIED PRODUCT MFC
 METALS PROCESSING
 PETROLEUM & RELATED INDUSTRIES
 OTHER INDUSTRIAL PROCESSES
 SOLVENT UTILIZATION
 STORAGE & TRANSPORT
 WASTE DISPOSAL & RECYCLING
 HIGHWAY VEHICLES
 OFF-HIGHWAY
 MISCELLANEOUS

 TOTAL
 FIRES
 Total without FIRES
1995
44
206
823
660
125
642
450
6,183
1,652
1,067
6,749
2,890
551
22,041
464
21,577
1996
50
179
893
388
73
477
435
5477
1294
509
6221
2935
1940
20871
1870
19001
1997
52
175
893
388
78
487
438
5621
1328
518
5985
2752
816
19530
744
18786
1998
56
174
889
394
78
485
443
5149
1327
535
5859
2673
718
18782
645
18136
1999
54
172
919
251
66
457
438
5036
1237
487
5681
2682
791
18270
667
17603
2000
62
173
949
254
67
428
454
4831
1176
415
5325
2644
733
17512
615
16898
2001
61
176
950
262
71
441
420
5012
1192
420
4952
2622
532
17111
412
16699
2002
52
170
790
214
69
375
406
4692
1205
457
4543
2688
883
16544
785
15759
2003
52
170
790
214
69
375
406
4692
1205
457
4543
2688
883
16544
785
15759
2004
52
170
790
214
69
375
406
4692
1205
457
4543
2688
883
16544
785
15759
                                                    2-6

-------
       The effects of sunlight on 63 formation, aside from the role of solar radiation in
meteorological processes, depend on its intensity and its spectral distribution. Intensity varies
diurnally, seasonally, and with latitude, but the effect of latitude is strongest in the winter.
Ultraviolet radiation from the sun plays a key role in initiating the photochemical processes
leading to Os formation and affects individual photolytic reaction steps. However, there is little
empirical evidence in the literature directly linking day-to-day variations in observed surface UV
radiation levels with variations in tropospheric O3 levels (CD, p.AX2-90).

       2.2.2   Relationship of Ozone to Photochemical Oxidants
       The relationship between 63, other oxidants, and oxidation products is complex and
involves  many factors.  Most notably, Os acts as a generator of hydroxyl radicals (OH)
propagating a variety of integrated and cascading reactions yielding additional oxidizing species
(e.g., hydroperoxy radical, HO2; organic peroxy radicals, RO2; hydrogen peroxide, H2O2;
nitrogen  dioxide, NO2;  many of which are short lived in the atmosphere. These "oxidizing"
species would be expected, like Os, to interfere with cellular processes leading to adverse health
effects.  Oxidants also drive the formation of particle bound sulfate, nitrate and organic carbon;
which  are principal  components of secondarily formed PM^.s; and influence positively the
formation or reaction of toxic gases such as peroxy acetyl nitrate (PAN), aldehydes (e.g.,
formaldehyde, acetaldehyde, acrolein), organic nitrates  and organic amines.  Furthermore,
deposition and subsequent adverse aquatic and terrestrial effects generally are enhanced through
availability of reactive  oxidant species.   These "enhanced" deposition impacts include
acidification, near field deposition of reactive mercury,  and other nitrogen-based eutrophication
effects (See Figure 2-1).  Benefits of reducing Os extend beyond that associated with Os
autonomously, given the  role of O3 as an oxidizing agent elevating levels of atmospheric
components responsible for a broad range of adverse human health and environmental impacts.

       2.2.3   Transport
       The transport of Os  and other secondary pollutants is determined by meteorological and
chemical processes  extending typically over spatial scales of several hundred kilometers (e.g.,
Civerolo et al., 2003; Rao et al., 2003).  An analysis  of the output of regional model studies
conducted by Kasibhatla  and Chameides (2000) suggests that Os can be transported over a few
thousand kilometers in the upper boundary layer of the eastern half of the United States during
specific Os episodes. Convection is capable of transporting Os and its precursors vertically
through the troposphere as shown in Annex AX2.3.2 of the CD.  Nocturnal low-level jets (LLJs)
                                           2-7

-------
                                                      Primary Sources
Figure 2-1. Atmospheric Processes Affecting the Formation of Photochemical Oxidants and Particulate Matter. Chemical
links illustrating relationships across, criteria pollutants and HAPs including mercury, as well as connections across sources,
secondarily formed species, gases, particulate matter and deposition.   Primary emissions are distinguished from secondarily formed
species.  Note that this diagram is a highly condensed model that does not capture numerous various heterogeneous processes and
complex chemical pathways.  Key atmospheric species that are involved in many reactions across pollutant categories include O3 and
the hydroxyl radical, OH.  Primary PM emissions are not included as they interact marginally with other other atmospheric species
(adopted from NARSTO, 2002).
                                                          2-8

-------
can also transport pollutants hundreds of kilometers over the mid-Atlantic region, the central
U.S. and California (Zhang et al., 2001).  Turbulence associated with LLJs can bring these
pollutants to the surface and result in secondary O?, maxima in the early morning in many
locations. However, the presence of mountain barriers can limit both horizontal and vertical
dispersion such as observed in Los Angeles and Mexico City resulting in a greater frequency and
duration of days with high 63 concentrations (CD, p.2-10).

2.3  DATA SOURCES
       Two main sources of monitoring data were used for this assessment, the state-supplied
data from various types of monitors housed in the Air Quality System (AQS) data base (which
includes National Park Service monitors) and the Clean Air Status and Trends Network
(CASTNET). The vegetation exposure analysis also uses an enhanced Veroni Neighborhood
Average (eVNA) spatial interpolation technique to combine 2001 monitor data from both AQS
and CASTNET with 2001 modeled data from the Community Multi-scale Air Quality (CMAQ)
model. This interpolated surface is used to fill in the gaps left by a sparse rural monitoring
network in the western United States.
       Air quality models are often used to simulate the formation, transport, and decay of air
pollution.  The CMAQ modeling system is a comprehensive three-dimensional grid-based
Eulerian air quality model designed to estimate Os and particulate concentrations and deposition
over large spatial scales (Dennis et al., 1996; Byun and Ching, 1999; Byun and Schere, 2006).
The CMAQ  model is a publicly available, widely-used, peer-reviewed, state-of-the-science
model consisting of a number of science attributes that are critical for simulating the oxidant
precursors and nonlinear organic and inorganic chemical relationships associated with the
formation of 63,  as well as sulfate, nitrate, and organic aerosols.

       2.3.1   Air Quality System (AQS)
       EPA's ambient air quality surveillance  regulations are found at 40 CFR Part 58.  Section
58.20 requires States to provide for the establishment of air quality surveillance systems in their
State Implementation Plans (SIP). The air quality surveillance system consists of a network of
monitoring stations designated as State and Local Air Monitoring Stations (SLAMS), which
measure ambient concentrations of those pollutants for which standards have been established in
40 CFR Part 50.  SLAMS, National Air Monitoring Stations (NAMS), which are a subset of
SLAMS, and Photochemical Assessment Monitoring Stations (PAMS) must meet the
requirements of 40 CFR Part 58, Appendices A (Quality Assurance Requirements), C (Ambient
Air Quality Monitoring Methodology), D (Network Design Criteria), and E (Probe and Path
Siting Criteria).
                                          2-9

-------
       The Air Quality System (AQS) is EPA's repository of ambient air quality data. AQS
stores data from over 10,000 monitors; 5000 of which are currently active.  Of these, over 3000
measure and report Os concentration data (See Figure 2-2). These monitors make up the
SLAMS, PAMS, NAMS, and other special purpose monitors used and operated by the States.
AQS also contains meteorological data, descriptive information about each monitoring station
(including its geographic location and its operator), and data quality assurance/quality control
information. The Office of Air Quality Planning and Standards (OAQPS) and other AQS users
rely upon the data system to assess air quality, assist in Attainment/Non-Attainment
designations, evaluate State Implementation Plans for Non-Attainment Areas, perform modeling
for permit review analysis, and other air quality management functions.  AQS information is also
used to prepare reports for Congress as mandated by the Clean Air Act (see
http://www.epa.gov/ttn/airs/airsaqs/sysoverview.htm).
       The NAMS/PAMS/SLAMS Os monitor network achieved an overall average bias (upper
bound) of 0.2% and an overall mean precision of 3% for 2002.  If special purpose and other Os
monitors are also included the average upper bounds of bias and precision were 0.4% and 2.9%
respectively (U.S. EPA 2004a).

       2.3.2   CASTNET
       CASTNET is the nation's primary source for data on dry acidic deposition and rural,
ground-level Os.  Operating since 1987, CASTNET is used in conjunction with other national
monitoring networks to provide information for evaluating the effectiveness of national emission
control strategies. CASTNET consists of over 80 sites across the eastern and western United
States (see Figure 2-2) and is cooperatively operated and funded with the National Park Service.
In 1986, EPA established the National Dry Deposition Network (NDDN) to obtain field data on
rural deposition patterns and trends  at different locations throughout the United States. The
network consisted of 50 monitoring sites that derived dry deposition data based on measured air
pollutant concentrations and modeled dry deposition velocities estimated from meteorology, land
use, and site characteristic data. In 1990, amendments to the Clean Air Act necessitated a long-
term, national program to monitor the status and trends of air pollutant emissions, ambient air
quality, and pollutant deposition. In response, EPA, in cooperation with the National Oceanic
Atmospheric Administration (NOAA), created CASTNET from NDDN. In terms of data  quality,
CASTNET achieved 98% to 99% of all precision and accuracy audits being within the ±10%
criteria for both precision and accuracy.  Overall, CASTNET Os monitors are stable showing
only very small variation (U.S. EPA 2003, p.22).
                                          2-10

-------
SOURCE
  (     AQS
       CASTNET
Figure 2-2. Locations of Ozone Monitors from AQS and CASTNET
                                 2-11

-------
2.4  OZONE MONITORING METHODS AND DATA QUALITY
       2.4.1  Ozone Monitoring Methods
       Ozone monitoring is conducted almost exclusively with UV absorption spectrometry with
commercial short path instruments, a method that has been thoroughly evaluated in clean air. The
ultimate reference method is a relatively long-path UV absorption instrument maintained under
carefully controlled conditions at the National Institute of Standards and Technology (NIST)
(CD, p.2-22). Most Os UV instruments reference the NIST method through a network of
Standard Reference Photometers (SRPs) that are maintained and operated by EPA.
       Several reports in the reviewed scientific literature have investigated interferences in O3
detection via UV radiation absorption and chemiluminescence which is the reference method but
almost never employed. These include the effects of water vapor, VOC's, aromatic compounds
and their oxidation products, and other organic and inorganic compounds. Water vapor had no
significant impact on UV absorption-based instruments, but could cause a positive interference
of up to 9% in chemiluminescence-based detectors at high humidities (dew point of 24° C).
Aromatic compounds and their oxidation products were found to generate a positive but small
interference in the UV absorption instruments. However, when the results are  applied to ambient
concentrations of toluene and NOx, the effect appears to be very minor (about 3 percent under
the study conditions). Other organic and inorganic compounds displayed interferences, but not at
levels likely to interfere with accurate determination of O3 in an urban environment (CD, p.2-25).
       Although not widely used, Os measurements by differential optical absorption
spectroscopy (DO AS) at a variety of wavelengths in the UV and visible parts of the spectrum.
Comparisons of DOAS results to those from a UV absorption instrument showed good
agreement on the order of 10%. Researchers have reported a positive interference due to an
unidentified absorber in the 279 to 289 nm spectral region used by many commercial short-path
DOAS systems for the measurement of O3. Results of that study suggest that compounds from
wood burning, used for domestic heating, may be responsible (CD, p.AX2-149).

       2.4.1  Effect of Measurement Precision on 8 hour Ozone Averages
       Staff conducted an analysis to determine the precision of an 8-hr averaged Os
concentration (Cox and Camalier, 2006). Daily maximum 8-hr Os values were simulated using a
Weibull distribution to yield a "true" three-year averaged Os design value without the influence
of measurement error.
       From 2002 to 2004, the average precision in the collected Os measurements is
approximately 3%. This means, for example, that a 1-hr measured concentration of 0.100 ppm
could be between 0.097 ppm and 0.103 ppm. Utilizing precision data from 900 Os monitors for
                                         2-12

-------
the 2002 through 2004 63 seasons, a second set of 8-hr 63 concentrations was generated to
incorporate the precision data from the Os monitoring network to account for instrument
measurement error. The result was a value which reflected the "true" Os design value plus
measurement error. The difference between the value with measurement error and the "true
value" reflects the impact of the instrument measurement error on the calculated 8-hr design
value.
       The exercise was repeated 1000 times and the differences between the two previously
described design values were summarized.  Table 2-3 shows the results of the analysis.  The
percentiles presented in the table reflect the percentage  of sites at or below the corresponding 1-
hour precision value.  The table shows that even at a precision of approximately 4.5%, of which
95% of the Os sites are at or below, the standard deviation of the difference between the 8-hr
design values is less than 0.001 ppm.
       A second exercise was performed to incorporate systematic bias error which includes the
instrument drift, noise, precision and calibration error associated with the  UV absorption method.
It was assumed that each 8-hr measurement was subjected to this randomly occurring bias which
had an average of zero and a standard deviation of approximately 0.004 ppm.  The mean and
standard deviation utilized for the simulation were believed to be reasonable estimates for
monitors operating under normal conditions.  The results  of this exercise show that assuming a
random bias of 0.004 ppm produced an uncertainty in the 8-hr design value of approximately
0.001 ppm.  This analysis supports expressing the level of the standard to  the nearest thousandth
(three decimal places) part per million (ppm), which is equivalent to the nearest part per billion
(ppb).2  The State of California also reached a similar conclusion regarding the precision of the
existing ozone monitoring methodology (California Environmental Protection Agency, 2005).

2.5  CHARACTERIZATION OF GROUND-LEVEL OZONE CONCENTRATIONS
       2.5.1   Metrics
       This section characterizes ground level Os concentrations based on several metrics. Two
daily maximum statistics, 1-hr and 8-hr averages,  and one daily average statistic in the form of a
24-hr concentration, and two cumulative concentration weighted statistics, SUM06 and W126,
are summarized to show how Os varies over space and time.  The 1-hr and 8-hr daily maximum
averaging times reflect the former and current 63  standards, and much of the health effects
literature for Os has focused on effects associated with these averaging times.  The 24-hr daily
       2Under the current standard, a rounding convention is used to determine attainment where the design value is rounded
to the nearest 0.01 ppm.  A National Ambient Air Quality Standard expressed to the nearest 0.001 ppm would mean that the
current rounding conventions become trivial. However, it is envisioned that the data handling guideline within the current
regulation where digits past the third decimal place are truncated would be retained.

                                           2-13

-------
Table 2-3. Relationship between Precision of 1-hour Ozone Data and Corresponding Standard Deviation of 8-hour Design
         Values
    Precision of 1- hour
      ozone value (%)
Nationwide Percentile
 Standard Deviation of
Difference in DV's (ppb)
1.63
2.22
2.97
3.89
4.52
25
50
75
90
95
0.27
0.34
0.45
0.57
0.63
                                                2-14

-------
average has been used for several personal exposure studies (CD, pp.3-72 - 74).  The SUM06
and W126 have been used frequently in the scientific literature and CD in studying and assessing
the relationship between Os exposures and adverse effects on vegetation. The daily maximum 8-
hr values are found by first calculating running or moving 8-hr values for all 24 hours in a day
(for example, averaging the 1-hr concentrations from l:00am to 8:00 am, then average the 1-hr
values from 2:00am to 9:00 am, etc.).  Then the maximum value for each day is found (note that
any 8-hr time period that starts in a day is assigned to that day). On an annual basis, the fourth
highest of these values is summarized. The daily maximum  1-hr statistic is the maximum value
of all 1-hr values in a day.  On an annual basis, the second highest of these values in a year is
summarized. The 24-hr average is a mean of the 24 individual hourly concentrations measured
from midnight to midnight.
       The maximum, 3 month, 12 hour SUM06 statistic is calculated by cumulating all 1-hr
values greater than or equal to 0.06ppm that occur during the 12 hour daytime window (8:00am
to 8:00pm Local Standard Time).  For each month of the Os monitoring season, the largest
consecutive 3-month sum of the daily values is calculated according to the secondary standard
proposed in 1996 (61 FR 65638), but not adopted in 1997 (62 FR 38856).  The SUM06 has a
weighting function that is 0 when the concentration is less than 0.06 and is 1.0 when the
concentration is greater than or equal to 0.06. The W126 seasonal cumulative statistics is
calculated similarly to the SUM06 statistic. The only difference is the weighting function where
the W126 statistic is a continuous, sigmoidal weighting function with an inflections point
between 0.06ppm and 0.07ppm (Lefohn and Runeckles, 1987).

       2.5.2   Spatial Variability
       This section characterizes the spatial variability of Os based on all the metrics discussed
above. Spatial variability is based on maps displaying county levels of the various metrics. In
this way different levels of Os for different areas of the country are displayed. It  should be noted
that county areas can be much larger in the West than in the East, but monitors are not spread
evenly within a county.  As a result, the assigned concentration range might not represent
conditions throughout a particular county.

        2.5.2.1     Distributions of 1-hr, 8-hr, and 24-hr Ozone Metrics
       Figures 2-3 to 2-5 show the distributions for measured 1-hr, 8-hr, and 24-hour daily
average Os concentrations for 12 major urban areas in the United States. The Los Angeles area
clearly has a distribution which is different from the other 11 cities, in that the hourly
concentration interquartile range is within 0.057 to 0.089 ppm as opposed to the next highest
interquartile range of Sacramento where 50% of the hourly concentrations lie between 0.056 and
0.079 ppm. In comparison, Houston which also has several 1-hr concentrations greater than
                                          2-15

-------
             X
                                                            X
                                                            X
                                                                   X
                            X
                            X
                                                             X
      Kv
CX
                                                                                  1°
                                                \  os
                                                       City
Figure 2-3. 1-hr Ozone Distributions across 12 Risk Areas, 2002-2004. Box Depicts interquartile range and median; the dot
depicts the mean; whiskers depict 10th and 90th percentile; and 'x' depicts values outside the 10th and 90th percentile.
Data Source: AQS
                                                          2-16

-------
                                                                                            X
                                                             City
Figure 2-4. 8-hr Ozone Distributions across 12 Risk Areas, 2002-2004. Box Depicts interquartile range and median; the dot
depicts the mean; whiskers depict 10th and 90th percentile; and 'x' depicts values outside the 10th and 90th percentile.
Data Source: AQS
                                                          2-17

-------
     0.05

     0.04

     0.03

     0.02

     0.01

     0.00
                       X
                                        X
                                        X
                                                         X
                                                            City
Figure 2-5. 24-hr Ozone Distributions across 12 Risk Areas, 2002-2004. Box Depicts interquartile range and median; the dot
depicts the mean; whiskers depict 10th and 90th percentile; and 'x' depicts values outside the 10th and 90th percentile.
Data Source: AQS
                                                          2-18

-------
0.125 ppm has a lower interquartile range of 0.034 to 0.07 ppm with 10% of its hourly values
greater than 0.089 ppm as opposed to approximately 0.106 ppm for Los Angeles.  Houston also
has a larger interquartile range of 0.036 ppm when compared to the average of the remaining 11
cities of 0.025 ppm. This trend is also observed in the 8-hr averaged concentrations. The
remaining 9 cities exhibit similar distributions to one another for all three metrics (1-hr, 8-hr, and
24-hr).
        For the 24-hour daily averaged concentration distributions, Houston shows a lower 75th
percentile than the other cities with  areas like Cleveland, Philadelphia and New York having
higher distributions. The lower 24 hour concentrations in Houston indicate a wider range
between the daily Os minima and maxima unlike an area like Cleveland, which has a higher
interquartile range.

        2.5.2.2     8-Hour and 1-Hour Statistics
       High 8-hr average 63 concentrations tend to occur near larger urban areas exhibiting
similar patterns as corresponding 1-hr concentrations (see Figure 2-6). Elevated 8-hr levels
occurring in smaller urban and non-urban areas are most likely caused by transport.  Higher 8-hr
63 levels observed in smaller urban and non-urban areas are most obvious at the end of the
northeast corridor (the highly urbanized area running from Washington, DC to Boston, MA),
North-central New York, and the Northern coast of Lake Michigan. Some of the highest levels
occur not only in California, but also in Texas as well as some counties in the Northeast Corridor
and isolated counties in the East (see Figure 2-7) (Fitz-Simons, et al., 2005).  The highest 1-hr
levels occur in California. (Fitz-Simons, et al., 2005).

        2.5.2.3     Cumulative Concentration-Weighted Statistics
       The highest SUM06 and W126 levels in 2001 (most of the analyses in Chapter 7 center
on 2001 data) occurred in most of the agricultural areas of California. When the data were from
CASTNET sites, more purely rural  counties showed higher values (See Figures 2-8 through 2-
11) (Fitz-Simons, et al., 2005).  The SUM06 and W126 values experienced a sharp decline in
2004 when compared to 2002 primarily in the eastern part of the United States (See Figure 2-12
through 2-15). Although there were reductions in the West, the decreases in the East were more
substantial. As discussed in section 2.1, the overall reductions across the country could possibly
be due to lower temperatures experienced during the Os season. However, the eastern half of the
country was also subject to the emission control requirements implemented under the NOX SIP
Call which occurred after 2002.  The improvements seen in 2004 for the East are most likely due
to a combination of cooler weather, the emission reductions from the NOX SIP Call, and emission
reductions from mobile source and other stationary source rules.
                                          2-19

-------
Concentration PPM
                                       3 X< 0.100;  364 Counties; 63,252,165 People (2000 census)
                                 I	| 0.100 <= X < 0.125; 264 Counties;  100,568,984 People
                                 ^^m 0.125 <=X; 16 Counties; 26,021,093 People
Figure 2-6. Average 2nd Highest Daily Maximum 1-hour Values in U.S. Counties, 2002-2004 AQS Data.
                                                2-20

-------
           Concentration PPM
X< 0.074; 140 Counties; 27,502,811 People (2000 census)
0.074 <= X < 0.085; 294 Counties; 66,167,168 People
0.085 <=X;  210 Counties;  96,172,263 People
Figure 2-7. Average 4th Highest Daily Maximum 8-hour Values in U.S. Counties, 2002-2004 AQS Data.
                                                           2-21

-------
                Concentration PPM-Hour        SUM06<15
                                           25<=SUM06<38
15<=SUM06<25
38<=SUM06
Figure 2-8. Highest 3-month 12-hour SUM06 Exposure Index in U.S. Counties, 2001 AQS Data.
                                                      2-22

-------
          Concentration PPM-Hour
W126
-------
               Concentration PPM-Hour        SUM06<15
                                          25<=SUM06<38
15<=SUM06<25
38<=SUM06
Figure 2-10. Highest 3-month 12-hour SUM06 Exposure Index in U.S. Counties, 2001 CASTNET Data.
                                                     2-24

-------
          Concentration PPM-Hour
W126
-------
             Concentration PPM-Hour
5UM06<15
?5<=	
                                              =SUM06<38
15<=SUM06<25

38<=SUM06
Figure 2-12. Highest 3-month 12-hour SUM06 Exposure Index in U.S. Counties, 2002 AQS and CASTNET Data.
                                                  2-26

-------
          Concentration PPM-Hour
W126
-------
             Concentration PPM-Hour
5UM06<15
?5<=	
                                              =SUM06<38
15<=SUM06<25

38<=SUM06
Figure 2-14. Highest 3-month 12-hour SUM06 Exposure Index in U.S. Counties, 2004 AQS and CASTNET Data.
                                                  2-28

-------
           Concentration PPM-Hour
W126
-------
       2.5.3   Temporal Variability
       Temporal variability consists of several time frames when considering characterization of
ground level air quality data. Multi-year variability characterizes long term variability or year to
year variability. Trends usually provide evidence on whether or not air quality is improving over
time.  For the purposes of displaying long term trends, the data from both AQS and CASTNET
are screened for temporally consistent data (only data from sites that meet a data completeness
criteria of 12 complete years out of 15 and no gaps of more than 3 consecutive years are
included). Seasonal variability characterizes month to month variability to demonstrate when in
the year the highest concentrations occur. Diurnal variability characterizes hour-to-hour changes
demonstrating when, in the day, the highest concentrations occur (Fitz-Simons, et al., 2005).

        2.5.3.1     Long Term Variability - Trends
       Long term, nationwide trends for 8-hr Os values are presented in Figures 2-16 and 2-17.
Figure 2-16 presents data from sites in the AQS that meet trends criteria and have locations
described as Urban and Center City.  Figure 2-17 presents data from CASTNET which are rural
locations.
       The rural and urban trends are similar, but the urban trends have more data and more
variation. The rural means are slightly lower than the urban means; however the largest urban
concentrations are much higher than the largest rural concentrations (Fitz-Simons, et al., 2005).
       Long term trends for 1-hr 63 values are presented in Figures 2-18 and 2-19. Figure 2-18
presents data from sites in the AQS that meet trends criteria and have locations described as
Urban and Center City. Figure 2-19 presents data from CASTNET which are rural locations. As
with the 8-hr data, the 1-hr urban trends and rural trends are similar, but urban have more data
and more variation. The 1-hr means for the urban trends are higher than the means for the rural
trends. This difference is more pronounced than in the 8-hr trends (Fitz-Simons, et al., 2005).
       The long term trends for both 1-hr and 8-hr Oj data are similar.  The 8-hr concentrations
are lower, but the  trends are basically parallel.  The urban area peak values in both the 1-hr and
8-hr concentrations have shown a gradual decline during the 15-year period, while the mean and
median concentrations have not varied much.  The highest means occur in 1990, 1991, 1995,
1998 and 2002. The highest extreme values are clearly in the 1990s.  In many cases, short term
variation (3 years  or less) is associated with meteorological conditions that are generally more or
less conducive to Os formation in a particular year. One high year between two low years or one
low year between  two higher years are examples of this 3 years or less variation (see Evaluating
Ozone Control Programs in the Eastern United States:  NOX p. 17,  U.S. EPA, 2005b)
                                          2-30

-------
                                        Urban
    0.20
    0.18
    0.16
 ~ 0.14
 Q.
 Q.
        487   505   505   543   545   569   560  592   602  618   633  650   670   670   670
 o
 '•S
 5
 0)
 o
 c
 o
 O
 0)
 c
 o
 N
 O
    0.12
    0.10
0.08
    0.06
    0.04
    0.02
    0.00
           X
                X
                               X    X
                                          X
                                          X
                                                X
                                                                        x
                                                                  X
                                                                                   X
                                                                            X    X
        1990  1991  1992  1993  1994  1995 1996  1997 1998  1999 2000  2001  2002  2003  2004
Figure 2-16. 4th Highest Daily Maximum 8-hour Ozone Values 1990-2004 (Urban). Box Depicts interquartile range and median;
the dot depicts the mean; whiskers depict 10th and 90th percentile; 'x' depicts values outside the 10th and 90th percentile; and
numbers above the boxes depicts the number of sites.
Data Source: AQS
                                                        2-31

-------
    0.20
    0.18
    0.16
 — 0.14
 Q.
 Q.
 O
 '•S
 5
 0)
 O
 c
 O
 O
 0)
 8
 O
0.12

0.10

0.08

0.06

0.04

0.02

0.00
X
         X
                                         Rural
         52    55    55    51    52    55   55    66    72   75    75   75    76    77    81
      x    x
                          X
                            X
                                 X
                                                X
X
I
•
I

X
X
                                                                       X
                                                                       X
                                                                              X
                                                                              X
        1990  1991  1992  1993 1994  1995  1996  1997  1998  1999  2000 2001  2002  2003  2004
Figure 2-17. 4th Highest Daily Maximum 8-hour Ozone Values 1990-2004 (Rural). Box Depicts interquartile range and median;
the dot depicts the mean; whiskers depict 10th and 90th percentile; 'x' depicts values outside the 10th and 90th percentile; and
numbers above the boxes depicts the number of sites.
Data Source: CASTNET
                                                        2-32

-------
 E
 Q.
 Q.
 0)
 O
 c
 O
 O
 0)
 c
 O
 N
 O
0.32

0.30

0.28

0.26

0.24

0.22

0.20

0.18

0.16

0.14

0.12

0.10

0.08

0.06

0.04

0.02

0.00
                                        Urban

        494   5$7   513  547   546  572   566   595   604   620  631   650   668   667   671
         X
X


X

X


X

X
X
                                     X
                                                X
                                                            X
           X
                                                                            X
                                           X
                                                X
                                       X
                                                                 X
X    X
X     X
        1990  1991  1992  1993 1994  1995  1996  1997  1998 1999  2000  2001  2002  2003 2004

Figure 2-18. 2nd Highest Daily Maximum 1-hour Ozone Values 1990-2004 (Urban). Box Depicts interquartile range and median;

the dot depicts the mean; whiskers depict 10th and 90th percentile; 'x' depicts values outside the 10th and 90th percentile; and

numbers above the boxes depicts the number of sites.

Data Source: AQS
                                                        2-33

-------
                                         Rural
    0.32
    0.30
    0.28
    0.26
    0.24
 •§ 0.22
 1 0.20
 c
 5  0.18
 s
 I  0.16
 o
 o  0.14
 O
 £  0.12
 o
 O  0.10
    0.08
    0.06
    0.04
    0.02
    0.00
52    55   55    51   52    55    55   66    72    75    75    75   76    77
X
                                             X
                                                                         80
        1990  1991  1992  1993  1994  1995  1996 1997  1998  1999  2000  2001  2002  2003  2004
Figure 2-19. 2nd Highest Daily Maximum 1-hour Ozone Values 1990-2004 (Rural).  Box Depicts interquartile range and median;
the dot depicts the mean; whiskers depict 10th and 90th percentile; 'x' depicts values outside the 10th and 90th percentile; and
numbers above the boxes depicts the number of sites.
Data Source: CASTNET
                                                         2-34

-------
        2.5.3.2    Short Term Variability - Annual
       Figure 2-20 shows a map of the number of exceedance days for 2002 and 2004 at 12
urban locations in the United States. Each grouping of two bars represents the number of
exceedance days for 1-hr, 8-hr averaged and 24-hr averaged O3 concentrations. The 1-hr
measured concentrations were compared to the previous 1-hr Os standard of 0.12 ppm, 8 hour
averaged concentrations were compared to the current 63 standard of 0.08 ppm and the average
24 hour concentrations were compared to 0.055 ppm which is the 95th percentile for 24 hour Oj
concentrations across the United States for 2002 through 2004. The data show that in all sites in
the Midwest and the East, Os concentrations were down dramatically in 2004 when compared to
2002. This is due in part to the fact that 2004 was much cooler than 2002. The reduction in peak
63 concentrations also reflects the improvement in air quality due to emission reductions from
the NOX SIP Call, mobile source and other stationary source rules. The NOx  SIP Call provided
large NOX reductions in the eastern part of the country in 2003 and 2004, thereby reducing peak
63 concentrations (U.S. EPA, 2005b). The number of 8-hr exceedance days actually increased
for Houston while showing a decrease in Los Angeles.  The difference between 2002 and 2004
for days greater than 0.055 ppm for the 24 hour averaged concentrations is smaller for the three
cities west of the Mississippi  River than it is for cities in the eastern United States.

        2.5.3.3    Seasonal Variability
       Monthly statistics are the best method to characterize seasonal variation in Os
concentrations. However in many areas, monitors are not active during  cooler months.  As a
result, data from May through September are the only universally available data for all monitors.
Although this is a limited characterization of seasonal variability, it is consistent across the entire
national network.
       Figure 2-21 shows box-plots of all 2004 data from May through  September for the
second highest daily 1-hr maximums.  The center of the distribution shows a slight, steady
increase from May to September while the extreme values show a more pronounced but more
variable increase for the same period (Fitz-Simons, et al., 2005).
       Figure 2-22 shows box-plots of all 2004 data from May through  September for the fourth
highest daily 8-hr maximums. The center of the distribution and the extremes show a slight,
steady increase from May to July followed by a slight decrease from July through September
(Fitz-Simons, et al., 2005).

        2.5.3.4    Short Term Variability - Diurnal
       The daily cycles of human activity and the solar phase drive the hour-to-hour daily cycle
seen in ground level Os  concentrations. The daily 1-hr peak levels generally occur in the
afternoon with the lowest concentration occurring in the early morning.  However, on any given
                                         2-35

-------
                                                          1-hr   8-hr  24-hr
                                                        2002 2004 2002 2004 2002 2004
Figure 2-20. Comparison of 1-hr, 8-hr, and 24-hr Metrics for 2002 and 2004,12 Risk Areas




Data Source: AQS
                                                          2-36

-------
0.32
0.30
0.28
0.26
0.24
 — 0.22
 S 0.20
 I  0.18
 §  0.16

 5  °-14
 o>
 §  0-12
 N
 0  0.10
    0.08
    0.06
    0.04
    0.02
    0.00
          301
1002
976

378
 X
          May               June               July              August           September
Figure 2-21. 2nd Highest Daily Maximum 1-hour Ozone Values from 2004 by Month. Box Depicts interquartile range and
median; the dot depicts the mean; whiskers depict 10th and 90th percentile; 'x' depicts values outside the 10th and 90th percentile;
and numbers above the boxes depicts the number of sites.
Data Source: AQS
                                                          2-37

-------
    0.20
    0.18
    0.16
          360
1053
800
392
                                                                                       X
    0.14
 E
 Q.
 o
 i
 o
 c
 o
 o
 N
 O
    0.12
    0.10
    0.08
    0.06
    0.04
    0.02
    0.00
            I                   I                   I                  I                   I

          May               June               July              August           September


Figure 2-22. 4th Highest Daily Maximum 8-hour Ozone Values from 2004 by Month.  Box Depicts interquartile range and


median; the dot depicts the mean; whiskers depict 10th and 90th percentile; 'x' depicts values outside the 10th and 90th percentile;


and numbers above the boxes depicts the number of sites.


Data Source: AQS
                                                          2-38

-------
day when conditions are right, this phase can be reversed with the highest values occurring at
night or early morning. Ozone transport can also effect at what time peaks can occur. For
example, some sites in Maine peak late in the evening due to transport.
       In order to examine diurnal patterns, box-plots summarize 1-hr values and 8-hr for each
hour in the day. Figures 2-23 and 2-24 summarize 1-hr data from AQS that was classified as
urban and center city.  The pattern is similar for both weekend and week day data.  The pattern
of the center of the distribution of values shows a smooth sinusoidal portion of the curve from
6:00AM until 8:00PM and reaches a peak at 1:00 PM to 3:00 PM.  Then the pattern alters to a
gradual decrease from 9:00 PM to 6:OOAM (Fitz-Simons, et al., 2005).
       Figures 2-25 and 2-26 show the same set of summaries for 8-hr data. 8-hr values run
from 0 to 23 hours. Hourl is the average of 1-hr values from 1 to 8 while hour 2 is the average of
hours 2 to 9 and so on. The main difference between the 1-hr data and the 8-hr data is that the  8-
hr data exhibit a smoother sinusoidal pattern throughout the day with a peak for the center of the
distribution occurring  at 10:00 AM or 11:00 AM and a minimum at about 12:00 midnight.  The
week end pattern is similar to the week day pattern (Fitz-Simons, et al., 2005).
       Figures 2-27 through 2-30 summarize 1-hr and 8-hr data from CASTNET sites which are
considered rural. Several differences are noted here. The patterns for the center of the
distribution are similar to the patterns for the urban sites. The largest values of the 1-hr data
exhibit no pattern but the largest values for the 8-hr data have a discernable pattern that differs
from the patterns for the values in the center of the distribution. The weekday pattern for the
highest values, shown in figure 2-29, has a  smooth sinusoidal pattern but reaches 2 peaks in the
day (12:00 midnight and 12:00 noon). The weekend pattern, shown in figure  2-30, also shows a
pronounced peak in the afternoon at about 1:00 PM which occurs about 2 hours after the peak for
the values in the center of the distribution (Fitz-Simons,  et al., 2005).

2.6   CHARACTERIZATION OF OZONE EPISODES
       Major episodes of high O3 concentrations in the United States are associated with slow
moving, high pressure systems. High pressure systems during the warmer seasons are associated
with the sinking of air, resulting in warm, generally cloudless skies, with light winds. These
conditions result in the development of stable air masses near the surface which inhibit the
vertical mixing of Os precursors. The combination of inhibited limited vertical mixing and light
winds minimizes the dispersal of pollutants emitted in urban areas, allowing their concentrations
to build up. Photochemical activity involving these precursors is also enhanced because of higher
temperatures and the availability of sunlight. Downward entrainment of overnight transported 63
and precursors trapped aloft begins on the following day as the PEL starts growing.  In the
                                          2-39

-------
   0.200
   0.750
 8 0.100

 s
 o
   0.050
         22 23  00  01  02 03 04 05  06  07 08 09 10 11  12  13  14 15 16 17  18  19  20 21 22 23 00  01


                                                   hour


Figure 2-23. 1-Hour Diurnal Week Day Pattern for Urban Sites, May through September 2004. Box Depicts interquartile range


and median; whiskers depict maximum and minimum values; and '+' depicts the mean.


Data Source: AQS
                                                        2-40

-------
   0.200
   0.750
 N 0.700
 o
 o
   0.050
          22 23 00 01  02  03 04 05  06  07 08 09  10  11  12 13  14  15 16 17 18  19  20 21 22 23  00 01
                                                    hour
Figure 2-24. 1-Hour Diurnal Week End Pattern for Urban Sites, May through September 2004. Box Depicts interquartile range
and median; whiskers depict maximum and minimum values; and '+' depicts the mean.
Data Source: AQS
                                                        2-41

-------
 s
 §
 g
 o
oo
      0.125
      0.100
      0.075
      0.050
      0.025
      0.000
             22 23  00  01  02 03 04 05 06 07  08  09 10 11 12 13 14  15  16  17 18 19 20 21 22 23  00  01
                                                      hour
Figure 2-25. 8-Hour Diurnal Week Day Pattern for Urban Sites, May through September 2004. Box Depicts interquartile range
and median; whiskers depict maximum and minimum values; and '+' depicts the mean.
Data Source: AQS
                                                        2-42

-------
    0.725
    0.700
    0.075
 c
 I
 3
 O
•c
    0.050
    0.025
    0.000
            22 23 00 01 02 03 04  05  06 07 08 09 10 11 12  13  14  15 16 17 18 19 20 21 22  23  00 01
                                                    hour
Figure 2-26. 8-Hour Diurnal Week End Pattern for Urban Sites, May through September 2004. Box Depicts interquartile range
and median; whiskers depict maximum and minimum values; and '+' depicts the mean.
Data Source: AQS
                                                        2-43

-------
  0.750
  0.700
I
3
O
•c
   0.050
  -0.050
          \   \    \   \   \   \    \   \   \   \    \   \   \   \    \   \   \   \   \    \   \   \   \    \   \   \   \    \
         22 23  00  01 02 03  04  05 06 07  08  09 10 11  12  13 14 15  16  17 18 19 20  21  22 23 00  01
                                                   hour
Figure 2-27 1-Hour Week Day Diurnal Pattern for Rural Sites, May through September 2004. Box Depicts interquartile range
and median; whiskers depict maximum and minimum values; and '+' depicts the mean.
Data Source: CASTNET
                                                        2-44

-------
   0.200
   0.750
<   0.700
 i
 o
 o
    0.050
   -0.050
          22 23 00  01  02 03 04  05 06 07  08  09 10 11  12  13 14  15  16 17 18  19  20 21 22  23 00 01
                                                     hour
Figure 2-28. 1-Hour Week End Diurnal Pattern for Rural Sites, May through September 2004. Box Depicts interquartile range
and median; whiskers depict maximum and minimum values; and '+' depicts the mean.
Data Source: CASTNET
                                                        2-45

-------
   0.750
   0.725
   0.700
   0.075
 §
 g
   0.050H
 o
00
   0.025
  -0.025
          \   \   \    \   \   \   \    \   \   \    \   \   \   \    \   \   \    \   \   \    \   \   \   \    \   \   \    \
         22 23 00  01  02 03 04  05 06 07  08  09 10  11  12 13 14  15  16 17  18  19 20 21  22  23 00 01
                                                    hour
Figure 2-29. 8-Hour Week Day Diurnal Pattern for Rural Sites, May through September 2004.  Box Depicts interquartile range
and median; whiskers depict maximum and minimum values; and '+' depicts the mean.
Data Source: CASTNET
                                                        2-46

-------
   0.750
   0.725
   0.700
   0.075
 §
   0.050
   0.025
 -0.025
          i   i   i    i   i   i   i    i   i   i    i   i   i   i    i   i   i    i   i   i   i    i   i   i    i   i   i   i
         22 23 00  01  02 03  04  05 06 07  08  09 10 11  12 13 14  15  16 17 18  19 20 21 22  23 00 01
                                                    hour
Figure 2-30. 8-Hour Week End Diurnal Pattern for Rural Sites, May through September 2004. Box Depicts interquartile range
and median; whiskers depict maximum and minimum values; and '+' depicts the mean.
Data Source: CASTNET
                                                        2-47

-------
eastern United States, high 63 concentrations during an episode can extend over hundreds of
thousands of square kilometers for several days.
       Episodes have two main characteristics: the concentration level reached and the length of
time that this level is reached in consecutive days. The following discussion addresses how these
characteristics of episodes have varied through both space and time.
       Numbers of episodes defined by daily maximum 1-hr 63 concentrations reaching a level
of 0.12 ppm for 1  day generally follow the long term trend of central values (means or medians)
of the 1-hr Os data (See Figures 2-18 and 2-31).  As the length of these episodes increases, the
frequency of these episodes decreases.  In the most recent years (1997-2004) episodes lasting 5
days or more often have not occurred at all (Fitz-Simons, et al.,  2005). From this we conclude
that control strategies have been reducing peak 1-hr Os across major urban areas.
       Numbers of episodes defined by daily maximum 8-hr 63 concentrations reaching a level
of 0.08 ppm for 1  day generally follow the long term trend of central values of the 8-hr Os data
(See Figures 2-16 and 2-32). As the length of these episodes increase, the frequency of these
episodes decreases. However, some of the longer episodes (6 days of more) continue to occur at
this level even in the most recent years. In fact, the episode must be defined by a level of 0.10
ppm before these longer episodes disappear in the most recent years (Fitz-Simons, et al., 2005).
       As episode length and level increase for both 1-hr and 8-hr Os data the frequency
decreases (Figure  2-33 and 2-34).  The longer periods and higher levels disappear altogether in
the period from 2000-2004 (Fitz-Simons, et al., 2005).
       One final aspect of episodes to examine is the return time or the number of days between
episodes.  Looking at the intervals between episodes of O.OSppm for 8-hr data, the most prevalent
gap length in days is 1 day.  There is a slight peak again at 4 days followed by a gradual decrease
in frequency as the gap-length increases (see Figure 2-35). Looking at the same data for
episodes of 0.12ppm, it appears that some periodicities appear at 1 day, 5-6  days, 21 days, and
33-34 days.  The frequencies for these episodes are so small  compared to frequencies lower level
episodes that these indications should not be considered real or significant indications of
periodicities. The 1-hr 63 data exhibit much the same lack of periodicity as the 8-hr data (Fitz-
Simons, et al., 2005).

2.7   POLICY RELEVANT BACKGROUND LEVELS
       For purposes of this document, background or policy relevant background (PRB) 63 is
defined as the distribution of Os concentrations that would be observed in the U.S. in the absence
of anthropogenic (man-made)  emissions of precursor emissions (e.g., VOC, NOx, and CO) in the
U.S., Canada, and Mexico. This is referred to as policy-relevant background, since  this
definition of background facilitates separating pollution levels that can be controlled by U.S.
                                          2-48

-------
  /
            yl/1/l/l/l/l/l/l/l/l/l/l/l/l/i
             /  1123 / 1205 / 891 / 894 / 819 /  1133 /  667 /  649 / 866 / 822 / 553 /  514 /  711 /  528 / 214

          2 /I /I /i /i /§ /I /i /i /i /i /a /e /i /a ju /
           /  317 /  336 /  208 / 188 /  189 /  243 /  147 /  154 /  164 / 148 /  99 /  92 /  137 /  110 /  31 /
           L   -L   -L   -J.    -J.


         /  134 /  137 /  105 / 84 / 75 /
                  / U   LJ  I i

                              100   71
                                   LJ  n
                                 /  Ui / LJ  /
                                 /    /
                               43 /  57 / 25 /
                                           LJ   LJ   0  U   LJ
                                           19 /  22 /  48 / 38 /   7 /
4 / r~i  Hi  n>  n
 /  LJ / U> / U  / U
/  49 / 71 / 60 / 44
/	/	/	/	
              36
                     u   u   a
                       /
                     34 /  28 /  10
                             Quau/aa
                             40 /  10 /  3 /  16 / 20 /  6
       36 /  46 /  41
           u  a
          ' 23 / 18
                          a
                          19
                   LJ
                   9
                           a
                           10

              I I—I  LJ
               23 /  3
                                                 r~?
                                                  3
>6
   a
   16
U

24
a
15
               13
            LJ
            7
LJ
6
LJ
4 /
LJ   LJ
  1990   1991   1992  1993  1994  1995  1996   1997   1998  1999  2000  2001  2002  2003   2004
                             year


Figure 2-31. Length of Consecutive Day Episodes over 0.12 ppm by Year for 1-hour Ozone Data across all Monitors.

Data Source: AQS
                                   2-49

-------
              / 3369 / 3285 /  2561 / 3260 / 3221 / 3828 /  3068 / 3299 / 4460 /  4864 /  3132 / 3257 / 3906 /  2634 / 1589 /
                                    §
                                                                      i   »
        '/•/•/w/i/1/i/i/I/i/i/l/l/i/w/fl/
         / 1041 /  1277 / 788 / 958 /  1085 /  1270 / 1063 / 1153 /  1644 /  1640 / 991  / 1107 /  1641 / 789 / 440 /
            y  / •  / y / y
               /    I    I
            396 / 558  /  323 /  358
                             387   579   468   390   654
                                                                    416   121
              LJ
          172
                               371
                                a  / a  / u
                                180   147   313
                                                372   163
                                                         27 A
                                                             375
                                                              a   a
                                                              166 / 60
 5 / °  / ° I
  /  78 / 154 /
             uauu
                      91 /  155
             85   93
 /    /    /    /    /
 / / 1  /  /—i  / I 1  / i n  / i—i
  u    u   u   u   u
/ 104 /  71 /  161 /  202 / 67 /
                                                     / a   Q   a   a
                                                       101 /  269 / 66 /  36 /
  a   a   a
I  41 / 61 /  62
               a   a   a   a
               47 / 43 /  65 /  61
                                    aauaauaa
                                  '  47 /  74 / 70    46  /  45 /  110 / 44 /  18 /
>6
a   ^ / ^
33 / 43 /  37 /
                a   a   a   a  a   a   a   a   a   a
                27 /  25 /  31 /  34 / 39 /  33 /  34 /  24 /  17 /  35
                                                   a
                                                        25 /
  1990  1991   1992   1993  1994  1995   1996   1997  1998   1999   2000  2001  2002   2003   2004
                                year


Figure 2-32. Length of Consecutive Day Episodes over 0.08 ppm by Year for 8-hour Ozone Data across all Monitors.

Data Source: AQS
                                      2-50

-------
    W/a
                                                           _
                                     5433  /     2655  /     1168  /     470   /     209   /      74
                                                                         33
                               2285
          LJ
           952
                                                326  /      134   /     29   /     12
                         _

                         1029
     359   /     131   /      46    /     11
                    597   /     191   /     63    /     13
              303   /     64
      10
                                            1    /      1
>6
         355   /      95
17
0.09         0.1         0.11
                                       0.12
                                       level
                   0.13         0.14         0.15
Figure 2-33. Length of Consecutive Day Episodes over Displayed Levels for 1-hour Ozone Data (2000-2004) across all
          Monitors.

Data Source: AQS
                                                      2-51

-------
             5    /  ^

                     4536  /     1835  /      539   /      112   /      18
               _
               2957
      968   /     263
41
>6
         .
         5861
1423   /     315   /      57    /      13
      0.06         0.07         0.08
                     0.09
                     level
  0.1          0.11         0.12
Figure 2-34. Length of Consecutive Day Episodes over Displayed Levels for 8-hour Ozone Data (2000-2004) across all
           Monitors.

Data Source: AQS
                                                          2-52

-------
   LU


   o
   LU
   o:
                              lllhhllhihli.iiiiii.......
       1 2 34 5 6  7 8 9  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

                                   gap length (days)





Figure 2-35. Length of Gaps in Days Between Episodes over 0.08 ppm for 8-hour Ozone Data (2000-2004).
                                           2-53

-------
regulations (or through international agreements with neighboring countries) from levels that are
not generally controllable in this manner. As defined here, PRB includes (1) Os in the U.S. from
natural sources of emissions in the U.S., Canada, and Mexico and (2) Os in the U.S. from the
transport of 63 or the transport of emissions from both natural and man-made sources, from
outside of the U.S. and its neighboring countries. As discussed in Chapter 5 of this Staff Paper,
PRB concentrations enter into the assessments of risk to human health.
       Contributions to PRB levels of O3 include: photochemical interactions involving natural
emissions of VOCs, NOx, and CO; the long-range transport of O3 and its precursors from outside
North America; and stratospheric-tropospheric exchange (STE). Processes involved in STE are
described in detail in Annex AX2.3 of the CD. Natural sources of O3 precursors include biogenic
emissions, wildfires, and lightning. Biogenic emissions from agricultural activities are not
considered in the formation of PRB (CD, p.AX2-145).
       As a result of long-range transport of O3  and its precursors from anthropogenic sources
within North America, estimates of PRB O3 concentrations cannot be derived solely from
measurements of O3, and must be based on modeling. The global photochemical transport model
GEOS-CHEM (Fiore et al., 2003) has been applied to estimate PRB O3 concentrations across the
U.S. (U.S. EPA, 2005a, AX3-131). The CD refers to a number of GEOS-CHEM publications
(Bey et al., 2001; Liu et al., 2002; Martin et al., 2002; Fusco and Logan, 2003; Li et al., 2002,
2005), summarizing their conclusions as "results indicate no significant bias, and agreement to
generally within 5 ppbv (parts per billion volume) for monthly mean concentrations at different
altitudes." The CD goes on to review detailed evaluations of GEOS-CHEM with Os
observations at U.S. surface sites (Fiore et al., 2002, 2003) and comparisons of GEOS-CHEM
predictions with observations at Trinidad Head, CA (Goldstein et al., 2004). The comparisons at
Trinidad Head are especially relevant because sources of the Os found there are often limited to
those in the PRB definition. The observations, filtered to remove local influence, averaged 41 ±
5 ppbv,  as compared to GEOS-CHEM predictions of 39 ± 5 ppbv, indicating no significant
differences between the model predictions and observations for conditions suggestive of PRB.
The CD further notes that "several other papers have evaluated the GEOS-CHEM simulation for
surface Os and its precursors over the United States." Summarizing their assessment of the
validity of the GEOS-CHEM model,  the CD states "in conclusion, we estimate that the PRB O3
values reported by Fiore et al. (2003) for afternoon surface air over the United States are likely
10 ppbv too high in the southeast in summer, and accurate within 5 ppbv in other regions and
seasons." These error estimates are based on comparison of model output with observations for
conditions which most nearly reflect those given in the PRB definition, i.e., at the lower end of
the probability distribution. For Os (cf Figures 8 and 9 of Fiore et al.,  (2003) for the SE and
                                         2-54

-------
Figure 3 of Fiore et al. (2002) for the ME.) it can be seen that GEOS-CHEM overestimates O3 for
the SE and underestimates it for the NE.
      The GEOS-CHEM model shows that PRB O3 concentrations are related to season,
altitude and total surface O3 concentration. PRB O3 concentrations at the surface are generally
predicted to be in the range of 0.015 to 0.035 ppm in the afternoon, and they tend to decline
under conditions conducive to high O3 episodes. They are highest during spring and decline into
summer.  Higher values tend to occur at higher elevations during spring due to contributions
from hemispheric pollution and stratospheric intrusions. The stratospheric contribution to
surface O3 is typically well below 0.020 ppm and only rarely elevates O3 concentrations at low-
altitude sites and only slightly more often elevates them at high-altitude sites (U.S. EPA, 2005a,
AX3-148).
      In the previous review of the 63 NAAQS, the criteria document and staff paper adopted a
value of 40 ppb for PRB 03. However, Figure 3-17 in the CD shows that mean daily maximum
8-h Os concentrations were less than 40 ppb at over 10 % of U.S. sites. In 2004, the mean daily
maximum 8-h Os concentrations were less than 40 ppb at 25 % of U.S. sites. It is highly unlikely
that this fraction of Os monitoring sites would be unaffected by Os generated by sources from
within continental North America.  Figure 3-19, in the CD, shows that over the past 15 years,
mean daily maximum 8-h 63 concentrations at Voyageurs National Park were typically less than
40 ppb.  Simulations of Os at Voyageurs, the site with the lowest Os  show that there is still a
substantial regional contribution to 63 (Figure AX2-86). Thus, 40 ppb is likely to be too high for
the mean PRB Oj concentration.
      The exposure and health risk analyses described in Chapter 4 and 5 use estimates of PRB
based on runs of the GEOS-CHEM model applied for the 2001 warm season (i.e., April to
September).  The GEOS-CHEM data consist of hourly gridded values with  latitude running from
12° to 80 ° in 2 ° steps and longitude running from -177.5 ° to -47.5° in 2.5 ° steps.  These data are
used to create monthly average  diurnal profiles which are fixed for each month during the Os
season.  The PRB estimates from the grid nearest each of the 12 urban areas included  in the
exposure and risk analyses have been used to estimate PRB  in each of these areas. Appendix 2A
provides plots of the PRB estimates by month for each of the 12 urban areas.
                                         2-55

-------
REFERENCES


Bey, I.; Jacob, D. I; Logan, J. A.; Yantosca, R. M. (200la) Asian chemical outflow to the Pacific in spring: origins,
      pathways, and budgets. J. Geophys. Res. [Atmos.] 106: 23,097-23,113.

Byun, D.W., and Ching, J.K.S., Eds, 1999. Science algorithms of EPA Models-3 Community Multiscale Air Quality
      (CMAQ) modeling system, EPA/600/R-99/030, Office of Research and Development, U.S. Environmental
      Protection Agency.

Byun, D,W, and Schere, K.L. 2006, Review of the governing equations, computational algorithms, and other
      components of the Models-3 Community Multiscale Air Quality (CMAQ) modeling system, Applied
      Mechanics Reviews, Vol 59, pp 51-77.

California Environmental Protection Agency: Air Resources Board (2005) Review of the California Ambient Air
      Quality Standard for Ozone.

Civerolo, K. L.; Mao, H. T.; Rao, S. T. (2003) The airshed for ozone and fine paniculate  pollution in the eastern
      United States. Pure Appl. Geophys. 160: 81-105.

Cox, W. M.;  Camalier, L. (2006) The effect of measurement error on 8-hour ozone design concentrations. Memo to
      Ozone NAAQS Review Docket.

Dennis, R.L., Byun, D.W., Novak, J.H., Galluppi, K.J., Coats, C.J., and Vouk, M.A., 1996. The next generation of
      integrated air quality modeling: EPA's Models-3, Atmospheric Environment, 30, 1925-1938.

Fiore, A. M.; Jacob, D. J.; Bey, L; Yantosca, R. M.; Field, B. D.; Fusco, A. C.; Wilkinson, J.  G. (2002) Background
      ozone over the United States in summer: origin, trend, and contribution to pollution episodes. J. Geophys.
      Res. (Atmos.) 107(D15): 10.1029/2001JD000982.

Fiore, A.; Jacob, D.J.;Liu, H.; Yantosca, R.M.; Fairlie, T.D.; Li, Q.  (2003). Variability in Surface Ozone
      Background over the United States: Implications for Air Quality Policy. J. of Geophysical Research,
      108(D24)19-1 - 19-12.

Fitz-Simons, T.; McCluney, L.; Rizzo, M.(2005) U.S. EPA Memorandum to File.  Subject: Analysis of 2004 Ozone
      Data for the Ozone NAAQS Review, November 7.

Kasibhatla, P.; Chameides, W. L. (2000) Seasonal modeling of regional ozone pollution in the eastern United States.
      Geophys. Res. Lett. 27: 1415-1418.


Lefohn A.S. and Runeckles V.C., Establishing Standards to Protect Vegetation - Ozone Exposure/Dose
      Considerations. Atmospheric Environment 21:561-568, 1987.

Rao, S. T.; Ku, J.-Y.; Berman, S.; Zhang, K.; Mao, H. (2003) Summertime characteristics of the atmospheric
      boundary layer and relationships to ozone levels over the eastern United States. Pure Appl. Geophys. 160:21-
      55.
U.S. Environmental Protection Agency (1986). Guideline on the Identification and Use of Air Quality Data Affected
      by Exceptional Events. EPA-450/4-86-007.

U.S. Environmental Protection Agency (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants.
      Research Triangle Park, NC: Office of Research and Development; Report no. EPA/600/P-93/004aF.

U.S. Environmental Protection Agency (2003).Clean Air Status and Trends Network (CASTNet) 2001 Quality
      Assurance Report; Research Triangle Park, NC: Office of Air Quality Planning and Standards. Report from
      EPA Contract No. 68-D-98-112.
                                                2-56

-------
U.S. Environmental Protection Agency (2004a). 2003 Criteria Pollutant Quality Indicator Summary Report for July
      14, 2004, AQS Data; Research Triangle Park, NC: Office of Air Quality Planning and Standards. Report from
      EPA Contract No. 68-D-02-061.

U.S. Environmental Protection Agency (2004b).The Ozone Report: Measuring Progress through 2003. Research
      Triangle Park, NC: Office of Air Quality Planning and Standards; Report no.EPA-454-K-04-001.

U.S. Environmental Protection Agency (2005a).Air Quality Criteria for Ozone and Related Photochemical Oxidants.
      Research Triangle Park, NC: Office of Research and Development; Report no. EPA/600/R-05/0054aB.

U.S. Environmental Protection Agency (2005b).Evaluating Ozone Control Programs in the Eastern United States:
      NOX Budget Trading Program Progress and Compliance. Research Triangle Park, NC: Office of Air Quality
      Planning and Standards; Report no. EPA-454- K-05-001.

Zhang, H. Mao, K. Civerolo, S. Herman, J. Ku, S.T. Rao, B. Doddridge,  C.R. Philbrick, and R. Clark (2001).
      Numerical investigation of boundary layer evolution and nocturnal low-level jets: Local versus non-local
      PEL schemes. Environmental Fluid Mechanics.  1: 171-208.
                                                2-57

-------
                  3.    POLICY-RELEVANT ASSESSMENT OF HEALTH
                                   EFFECTS EVIDENCE

3.1  INTRODUCTION
       This chapter assesses key policy-relevant information on the known and potential health
effects associated with exposure to ambient 63, alone and in combination with other pollutants
that are routinely present in ambient air. This assessment focuses specifically on the health
effects evidence evaluated in Chapters 4 through 7 of the CD with particular emphasis on the
integrative synthesis presented in Chapter 8. That integrative synthesis focuses on integrating
newly available scientific information with that available from the last review, as well as
integrating information from various disciplines, to address a set of issues central to the
assessment of scientific information upon which this review of the Os NAAQS is based. This
chapter also addresses key issues relevant to quantitative assessment of controlled-human
exposure and epidemiological evidence, to provide  a foundation for the quantitative human
exposure and health risk assessments presented in Chapters 4 and 5. Those quantitative
assessments, together with this evidence-based assessment, provide the foundation for the
development of staff conclusions and identification of options for consideration related to
primary standards for Os presented in Chapter 6.
       The decision in the last review focused primarily on evidence from short-term and
prolonged controlled-exposure studies reporting lung function decrements, respiratory
symptoms, and respiratory inflammation in  humans, as well as epidemiology studies reporting
excess hospital admissions and emergency department (ED) visits for respiratory causes. The
CD prepared for this review emphasizes a large number of epidemiological studies published
since the last review with these and additional health endpoints, including the effects of acute
and chronic exposures to 63 on premature mortality, enhanced respiratory symptoms and lung
function decrements in asthmatic individuals, and school absences. It also emphasizes important
new information from toxicology, dosimetry, and controlled human exposure studies.
       As discussed in more detail below (section 3.3),  highlights of the new evidence include:
     •   New controlled human-exposure studies have  observed that very small (<5%) lung
         function decrements occur in  some healthy adults under moderate exertion for 6.6 hr
         exposures to levels as low as 0.04 and  0.06 ppm; however, in a few subjects,
         decrements  of >10% were observed.
     •   New controlled human-exposure studies offer evidence of increased airway
         responsiveness to allergens in subjects with allergic asthma and allergic rhinitis
         exposed to Os.
     •   Numerous controlled human-exposure studies have reported indicators of O3-induced
         inflammatory response in both the upper respiratory tract (URT) and lower respiratory
                                         3-1

-------
         tract (LRT), while other studies have shown significant changes in host defense
         capability following Os exposure of healthy young adults.

      •   Animal toxicology studies provide new information regarding mechanisms of action,
         increased susceptibility to respiratory infection, and the biological plausibility of acute
         effects and chronic, irreversible respiratory damage.

      •   Numerous acute exposure epidemiological studies published during the past decade
         offer added evidence of ambient O3-related lung function decrements and respiratory
         symptoms in physically active healthy subjects and asthmatic subjects, as well as
         evidence on new health endpoints, such as the relationships between ambient O3
         concentrations and school absenteeism and between ambient O3and cardiac
         physiologic endpoints.

      •   Several new studies have been published over the last decade examining the temporal
         associations between O3 exposures and ED visits for respiratory diseases and on
         respiratory-related hospital admissions.
      •   Newly available, large multicity studies, designed specifically to examine the effects of
         acute exposure to PM and Os on mortality, provide much more robust and credible
         information than was available in the last review. The results from two key studies
         carried out in 95 U.S. communities (U.S. National Morbidity, Mortality Air Pollution
         Study [NMMAPS]) and in 23 European cities (Air Pollution and Health: European
         Approach [APHEA]) reported positive and significant O3 effect estimates for all cause
         (nonaccidental) mortality.

      •   In a recent study, Bell et al. (2006) applied several statistical models to data on air
         pollution, weather, and mortality for the 98 NMMAPS communities to evaluate
         whether a threshold level exists for premature mortality.  The results indicate that even
         low levels of tropospheric Os are associated with premature mortality.

      •   Three recent meta-analyses evaluated potential sources of heterogeneity in O3-mortality
         associations, and these studies provide evidence of a robust association between
         ambient O3 and mortality, especially for the warm O3 season.
      Section 3.2 provides an overview of mechanisms of toxicity, with more detailed discussion

in Appendix 3 A. Section 3.3  summarizes the nature of effects induced by O3 exposure or

associated with exposure to Os, alone and in combination with other pollutants, drawing on

information in Chapters 5-8 of the CD.  Section 3.4 summarizes conclusions and judgments from

the CD's integrative assessment of the epidemiological evidence regarding the extent to which

causal inferences can be made about observed associations between health endpoints and

exposure to 63, and discusses key issues related to quantitative risk assessment based on such

evidence.  Section 3.5 discusses biological plausibility and coherence of evidence for O3-related

adverse health effects, including short-term respiratory effects, short-term cardiovascular effects,
                                          3-2

-------
long-term health effects, and mortality-related health endpoint.  Drawing from the CD's
integrative synthesis, section 3.6 discusses factors that modify responsiveness to O3; potentially
susceptible and vulnerable populations groups; and public health impacts of exposure to ambient
O3. Finally, section 3.7, summarizes key policy-relevant conclusions from the CD about (V
related health effects, in the context of a discussion of issues related to our confidence in and the
utility of the underlying evidence.

3.2  MECHANISMS OF TOXICITY
       Evidence is covered in Chapters 5 and 6 of the CD on possible mechanisms by which
exposure to O3 may result in acute and chronic health effects. While most of the available
evidence addresses mechanisms for O3,  we recognize that O3 serves as an indicator for the total
photochemical oxidant mixture found in the ambient air.  Some effects may be caused by one or
more components in the overall  pollutant mix, either separately or in combination with O3.
Evidence from dosimetry, toxicology, and human exposure studies has contributed to an
understanding of the mechanisms that help to explain the biological plausibility and coherence of
evidence for O3-induced respiratory health effects reported in epidemiological studies. In the
past, however, little information was available to help explain potential biological mechanisms
which linked O3 exposure to premature mortality or cardiovascular effects.  More recently,
however, an emerging body of animal toxicology evidence is beginning to  suggest mechanisms
that may mediate acute 63 cardiovascular effects.
       Scientific evidence discussed in the CD (section 5.2) indicates that reactions with lipids
and antioxidants are the initial step in mediating deleterious health effects of O3.  There is
subsequent activation of a cascade of events starting with inflammation, altered permeability of
the epithelial barrier, impaired host defense (including clearance mechanisms), and pulmonary
structural alterations that can potentially exacerbate a preexisting disease status.  According to
the CD, the scientific evidence is still lacking for clearly establishing a role for one or a group of
mechanistic pathways underlying O3 health effects observed in epidemiological studies.
Appendix 3 A provides a further discussion of mechanisms of toxicity.

3.3  NATURE OF EFFECTS
        The CD provides new evidence that notably enhances our understanding of short-term
exposure effects, including effects on lung function, symptoms, and inflammatory effects
reported in controlled exposure studies. These studies support and extend the findings of the
previous CD. There is also a significant body of new epidemiological evidence of associations
between short-term exposure to  O3 and effects such as premature mortality, hospital admissions
                                          3-3

-------
and ED visits for respiratory (e.g., asthma) causes. Key epidemiological and human controlled
exposure studies are summarized in Appendices 3B and 3C, respectively.
       The following discussions of O3-related health effects are based on scientific evidence
critically reviewed in chapters 5, 6, and 7 of the CD, as well as the CD's integration of scientific
evidence contained in Chapter 8. In addition, these health effects discussions rely on the more
detailed information and tables presented in the CD's annexes AX5, AX6, and AX7.
Conclusions drawn about O3-related health effects depend on the full body of evidence from
controlled-exposure human, epidemiological and toxicological data contained in the CD.
Section 3.3.1 focuses on a broad array of morbidity effects, including both acute and chronic
exposures.  Section 3.3.2 focuses on the expanded body of evidence on associations between
acute O3 exposure and mortality, as well as the more limited evidence on chronic O3 exposures
and mortality.

       3.3.1   Morbidity
       This section summarizes scientific information contained in the CD on respiratory and
cardiovascular effects associated with exposure to O3.  Evidence of the effects of short-term and
long-term exposure to O3 on the respiratory system is discussed in sections 3.3.1.1 and 3.3.1.2,
and evidence of O3-related cardiovascular effects in section 3.3.1.3.

       3.3.1.1 Effects on the Respiratory System  from Short-term Exposures
       Short-term  exposures to O?, have been reported to induce a wide variety of respiratory
health effects. These effects include a range of effects, such as morphological changes in the
respiratory tract, pulmonary function  decrements, respiratory symptoms, respiratory
inflammation, increased airway responsiveness, changes in host defense capability,  acute
morphological effects, increased ED visits and hospital admissions, and effects on exercise
performance. Short-term Os exposure has also been associated with increases in restricted
activity days and school absences but evidence  is limited for these effects.

        3.3.1.1.1  Pulmonary Function Decrements, Respiratory Symptoms, and Asthma
                  Medication Use
       A very large literature base of studies published prior to 1996, which investigated the
health effects on the respiratory system from short-term Os exposures, was reviewed in the 1986
and 1996 CDs (U.S. Environmental Protection Agency, 1986, 1996). In the last review, the
lowest 63 concentration at which statistically significant reductions in forced vital capacity
(FVC) and forced expiratory volume in 1 second (FEVi) had been reported in sedentary subjects
was 0.5 ppm (CD,  p 6-3). During exercise, spirometric and symptomatic responses were
observed at much lower O3 exposures. When minute ventilation was considerably increased by
                                          3-4

-------
continuous exercise (CE) during O?, exposures lasting 2 hr or less at > 0.12 ppm, healthy subjects
generally experienced decreases in FEVi, FVC, total lung capacity (TLC), inspiratory capacity
(1C), mean forced expiratory flow from 25% to 75% of FVC (FEF 25.75), and tidal volume (VT);
increases in specific airway resistance (sRaw), breathing frequency (fB), and airway
responsiveness; and symptoms such as cough, pain on deep inspiration, shortness of breath,
throat irritation, and wheezing.  When exposures were increased to 4- to 8-hr in duration,
statistically significant spirometric and symptom responses were reported at 63 concentrations as
low as 0.08 ppm and at lower minute ventilation (i.e., moderate rather than high level exercise)
than the shorter duration studies (CD. p. 6-6).
       The most important observations drawn from studies reviewed in the 1996 CD were that:
(1) young healthy adults exposed to Os concentrations > 0.08 ppm develop significant,
reversible, transient decrements in pulmonary function if minute ventilation or duration of
exposure is increased sufficiently, (2) children experience similar spirometric responses but
lesser symptoms from Os exposure relative to young adults, (3) Os-induced spirometric
responses are decreased in  the elderly relative to young adults, (4) there is a large degree of
intersubject variability in physiologic and symptomatic responses to Os but responses tend to be
reproducible within a given individual over a period of several months, (5) subjects exposed
repeatedly to 63 for several days show an attenuation of response upon successive exposures;
this attenuation is lost after about a week without exposure; and (6) acute Os exposure initiates an
inflammatory response which may persist for at least 18 to 24 hr post exposure (CD, p. 6-2).
       Since 1996, there have been a number of studies published investigating spirometric and
symptomatic responses, and they generally support the observations previously drawn.  Recent
studies for acute exposures of 1 to 2 hr and 6 to 8 hr in duration are summarized in Tables AX6-1
and AX6-2 of the CD (p. AX6-5 to AX 6-7 and p. AX6-11 to AX6-12) and reproduced as Tables
3C-1 and 3C-2 in Appendix 3C. Among the more important of the recent studies was
McDonnell et al. (1997) which examined reported changes in FEVi in 485 white males (ages 18-
36) exposed for 2 hr to Os concentrations from as low as 0.08 ppm up to 0.40 ppm, at rest or with
intermittent exercise (IE).  Decrements in FEVi were modeled by sigmoid-shaped curve as a
function of subject age, 63  concentration, minute ventilation, and duration of exposure. In
another study, Ultman et al. (2004) found that exposing  60 young, healthy subjects to 0.25 ppm
63 for 1 hr with continuous exercise produced considerable intersubject variability in FEVi
decrements ranging from 4% improvement to a 56% decrement, which was consistent with
findings in the 1996 CD. One third of subjects had FEVi decrements > 15% and 7% had
decrements > 40%. Foster et al. (1993, 1997) examined the effects of Os on ventilation
distribution and reported results suggesting a prolonged Os effect on the small airways and
ventilation distribution (CD, p. 6-5).
                                         3-5

-------
       For prolonged exposures (4 to 8 hr) in the range of 0.08 to 0.16 ppm Causing moderate
quasi-continuous exercise (QCE; 50 min exercise [minute ventilation of 35 to 40 L/min] and 10
min rest per hr), several pre- and post-1996 studies (Folinsbee et al., 1988,1994; Horstman et al.,
1990; Adams, 2002, 2003a, 2006) have reported statistically significant spirometric responses
and increased symptoms in healthy adults with increasing duration of exposure, O3 concentration,
and minute ventilation. Based on review of several prolonged exposure studies, the CD (p. 6-6)
concluded that FEVi decrements are a function of minute ventilation in 6.6 hr exposure studies
and that data  from recent studies do not support the contention that minute ventilation should be
normalized to body surface area (BSA) for adults. Triangular exposure (i.e., integrated
exposures that begin at a low level, rise to a peak, and return to a low level during time of
exposure) studies (Hazucha et al.,  1992; Adams 2003a, 2006) suggest that, depending upon the
profile of the  exposure, the triangular exposure, which may reflect the pattern of ambient
exposures in  some locations, can potentially lead to greater FEVi decrements than square wave
exposures (i.e., a constant exposure level during time of exposure) when the overall Os doses are
equal (CD, p. 6-10), suggesting that peak exposures are important in terms of O3 toxicology.
       McDonnell (1996) used data from a series of studies to investigate the frequency
distributions of FEVi decrements following 6.6 hr exposures and found that average FEVi
responses were relatively small (between 5 and 10 %) at 0.08 ppm 63 (CD, p. 8-17)1.  However,
about 18% of the exposed subjects had moderate functional decrements (10 to 20%), and about
8% experienced large decrements (>20%). Figure 3-lA,B,C (CD, Figures 8-lA,B and 8-2, pp.
8-17 and 8-19) is based on study data that are in McDonnell (1996) together with data from
Adams (2002, 2006) that were not published  but were obtained from the author. This figure
demonstrates that while average responses may appear small and insignificant, some individuals
can experience much more significant and severe effects that may be clinically significant. The
FEVi responses illustrated in this figure were not corrected for the effect of exercise in clear air.
When that is  done for the Adams (2002, 2006) data, the percentage of subjects experiencing
>10% FEVi decrements changes to 7% at Os  exposures of 0.04 ppm, to 7% at Os exposures of
       1 The studies conducted in EPA's clinical research facility in Chapel Hill, NC that are considered in the
lung function risk assessment measured ozone concentrations to within +/- 5% or +/- 0.004 ppm at the 0.08 ppm
exposure level. The accuracy of these measurements was confirmed in an email sent by Steve Jackson, Program
Manager for the Human Studies Division in the National Health & Environmental Effects Research Laboratory,
Office of Research and Development, USEPA. He has overall responsibility for the monitoring program at the
USEPA clinical research facility in Chapel Hill, NC. This email has been placed in the Ozone NAAQS review
docket.
                                          3-6

-------
A.
Ja su
O
§40-
W 30-
2 20-
n n.



nn
_







0 ppm
n = 87
0%


n
•
•
•
-
0,08 ppm
n " 60
26%


nil
r-i


i-i




OH n-n
•
•
•
-
0.10 ppm
n = 32
31%

n
f
n
I
~




n
n
In n
•
-
•
-
0.12 ppm
n = 49
46%


fin





fl HP n r.
         10  0 -10 -20 -30 -40
                             10  0  -10 -20 -30 -40
                                                  10  0  -10 -20 -30 -40
                                                                       10  0  -10 -20 -30 -40
B.
   o
   0)
50-
40-
30-
20-
10-
n-




n





0 ppm
n = 30
0%




n
-









-




0.04 ppm
n = 30
0%


In
.

•

•














0.06 ppm
n = 30
3%


IL
.
-
-
-
-
0.08 ppm
n = 30
17%


n




nH
I h n
       10   0   -10 -20  -30 -40  10
                                   -10  -20  -30 -40  10   0  -10  -20  -30  -40  10
                                                                            -10  -20  -30  -40
                                               (%change)
Figure 3-1A and B. Frequency distributions of FEVi changes following 6.6-h exposures to
a constant concentration of Os or filtered air.  Note that the percentage in each panel
indicates the distributions of %  decrement.
Source:Panel A, McDonnell (1996); Panel B, Adams (2002, 2006), pre- and post-FEV: data for each subject
provided by author.

c.
jg ou-
o
_g>< 50-
3 40-
co
•5 30-
C 20-
S 10-
$ n-






n







0 ppm
n = 30
0%





r~i

-
.

'
-
-
0.04 ppm
n = 30




^





7%







Hn













.n







0.06 ppm
n = 30
7%



n

•
,

-
.
-
0.08 ppm
n-30
23%






]n

1


n, „
     10
            -10 -20  -30  -40  10
                                 -10  -20  -30 -40  10
                                                       -10  -20  -30  -40 10
                                                                            -10  -20  -30  -40
                                             (%change)
Figure 3-1C. Frequency distributions of FEVi changes following 6.6-h exposures to a
constant concentration of Os or filtered air. The FEVi changes following Os exposures
have been corrected for filtered air responses, i.e., they are Os-induced FEVi changes. Note
that the percentage in each panel indicates the distributions of % decrement.
Source:  Adams (2002, 2006), pre- and post- FEVi data for each subject provided by author.
                                          5-7

-------
0.06 ppm, and to 23% at Os exposures of 0.08 ppm, in studies conducted in California (CD, p. 8-
18).  The development of these effects is time-dependent during both exposure and recovery
periods, with great overlap for development and disappearance of the effects. In healthy human
subjects exposed to typical ambient 63levels near 0.12 ppm, spirometric responses largely
resolve within 4 to 6 hr postexposure, but cellular effects persist for about 24 hr. In these healthy
subjects, small residual lung function effects are almost completely gone within 24 hr, while in
hyperresponsive subjects, recovery can take as much as 48 hr to return to baseline.  The majority
of these responses are attenuated after repeated exposure, but such attenuation to Os is lost one
week postexposure (CD, p. 8-19).
       In the Adams (2006) investigation of the effects of square-wave (0.00, 0.06, and 0.08
ppm Os) and triangular (averaging 0.04, 0.06, and 0.08 ppm Os) exposures for 6.6 hr during
quasi continuous exercise on pulmonary function in 30 healthy adults, the study was designed to
compare pulmonary function responses between the six exposure protocols at each of six time
points (1, 2, 3, 4.6, 5.6, and 6.6 hr)2.  Accordingly, the author utilized a multiple comparison
technique to avoid Type I error (falsely rejecting the null hypothesis of no difference). At 6.6 hr,
FEV1 responses from the 0.08 ppm Os exposures were found to be significantly different from
the responses observed for the 0.0, 0.04, and 0.06 ppm Os exposures.  The FEVi responses did
not differ significantly at 6.6 hr between the two 0.08 ppm 63 exposures (i.e., the square-wave
vs. the triangular).  Another statistically insignificant comparison was between the FEVi
responses at 0.06 ppm 63 and filtered air (0.0 ppm 63) at 6.6 hr.
       On examination of the group mean FEVi responses in Figure 1 of Adams (2006),
however, responses during the 0.06 ppm Os exposures appear to diverge from responses for
filtered-air and 0.04 ppm 63 (CD, 8-42). In addition to reducing Type I error, the correction for
the multiple comparisons by Adams  (2006) may have also increased Type II error (falsely
accepting the null) for the simple evaluation of pre- to postexposure effects of Os versus filtered
air on FEVi, as has been commonly assessed by others (e.g., Horstman et al., 1990; McDonnell
et al., 1991). A cursory evaluation of pre- to postexposure effects can be  completed utilizing the
summary data in Table 3 of the Adams (2006) publication. For the filtered air, 0.06 ppm Os
(square-wave), and 0.06 ppm 63 (triangular) exposures, the FEVi responses were 1.35±0.54
[mean±standard error (SE)], -1.51±0.77, and -1.43±1.09%, respectively. Under the null
hypothesis of no pre- to postexposure difference in FEVi responses between filtered air and Os
exposure, the lack of an overlap in the range of responses  (i.e., the means±SEs) at 0.06 ppm Os
versus filtered air is suggestive of a significant effect on FEVi.  Furthermore, in a prior
publication (Adams, 2002), the author stated that, "some sensitive  subjects experience notable
       2 These studies reported O3 concentrations to be accurate within +/- 0.003 ppm over the range of
concentrations included in these studies.
                                         3-8

-------
effects at 0.06 ppm," based on the observation that 20% of subjects exposed to 0.06 ppm Os (in a
face mask exposure study) had greater than a 10% decrement in FEVi even though the group
mean response was not statistically different from the filtered air response.  The effects described
by Adams (2002), along with the cursory evaluation of the Adams (2006) data as described
above, strongly suggest that exposure to 0.06 ppm O3 causes small group mean FEV! decrements
in healthy adults with some individuals having notable  effects.
       Although not mentioned in the CD, Adams (2006) reported that total subjective symptom
scores (TSS) during the triangular 0.06 ppm exposure reached statistical significance (relative to
preexposure) at 5.6 and 6.6 hr, whereas they did not reach significance during the square-wave
0.06 ppm exposure. Data in Table 4 of the Adams (2006) publication allow further evaluation of
pre- to postexposure effects on respiratory symptoms, both TSS and pain on deep inspiration
(PDI).  For the filtered air, 0.06 ppm (square-wave), and 0.06 ppm (triangular) exposures, the
TSS responses were 0.6±0.40 (mean±SE), 2.5±0.89, and 3.9±1.35, respectively, and the PDI
responses were 0.2±0.16 (mean±SE), 1.4±0.53, and 2.0±0.80, respectively. As noted above for
FEVi changes, the lack of an overlap in the ranges of responses (i.e., the means±SEs) at 0.06
ppm Os versus filtered air for those two sympom scores is suggestive of a significant effect on
respiratory symptoms.
       A relatively large number of field studies investigating the effects of ambient 63
concentrations, in combination with other air pollutants, on lung function decrements and
respiratory symptoms have been published since 1996 (see CD, sections 7.2.3, 7.2.4, and
8.4.4.1).  These newer studies  support the major findings of the 1996 CD that lung function
changes, as measured by decrements in FEVi or peak expiratory flow (PEF), and respiratory
symptoms in healthy adults and asthmatic children are closely correlated to ambient 63
concentrations. Pre-1996 field studies focused primarily on children attending summer camps
and found Os-related impacts on measures of lung function, but not respiratory symptoms, in
healthy children.  The newer studies have expanded to evaluate (Vrelated effects on outdoor
workers, athletes, the elderly, hikers, school children, and asthmatics. Collectively, these studies
confirm and extend clinical observations that prolonged exposure periods, combined with
elevated levels of exertion or exercise, may magnify the effect  of Os on lung function.  The most
representative data come from the hiker study (Korrick et al., 1998), which provided outcome
measures stratified by several factors (e.g., gender, age, smoking status, presence of asthma)
within a population capable of more than normal exertion.  In this study, lung function was
measured before  and after hiking, and both ambient and personal  Os exposure measurements
were made. The  mean 8-hr average 63 concentration was 40 ppb (SD 12).  Decreased lung
function was associated with Os exposure, with the greatest effect estimates reported for the
                                         3-9

-------
subgroup that reported having asthma or wheezing, and for those who hiked for longer periods of
time, thus increasing the exposure period (CD, p. 7-36).
       Asthma panel studies, conducted both in the U.S. and in other countries, have reported
that decrements in PEF are associated with routine 63 exposures among asthmatic and healthy
persons (CD, sections 7.2.3.2 and 8.4.4.1). One large U.S. multicity study (Mortimer  et al.,
2002) examined Os-related changes in PEF in 846 asthmatic children from 8 urban areas and
reported that the incidence of > 10% decrements in morning PEF are associated with a 30 ppb
increase in 8-hr average Os for a 5-day cumulative lag, suggesting that Os exposure may be
associated with clinically significant changes in PEF in asthmatic children; however, no
associations were reported with evening PEF (CD, p. 7-43).  The authors also reported that the
associations reported with morning PEF remained statistically significant when days with 8-hr
63 concentrations above 80 ppb were excluded (CD, p. 7-46). Two studies (Romieu et al., 1996,
1997) carried out simultaneously in northern and southwestern Mexico City with mildly
asthmatic school children reported statistically significant Os-related reductions in PEF, with
variations in effect depending on lag time and time of day.  In the northern study, the mean 1-hr
max Os concentrations were 190 ppb (SD 80), and in the southwestern study, mean 1-hr max
levels were 196 ppb (SD 78). While several  studies (Gielen et al., 1997; Jalaludin et al., 2000;
Ross et al., 2002; Thurston et al., 1997) report statistically significant associations between 63
exposure and reduced PEF in asthmatics, other studies (Hiltermann et al., 1998; Delfmo et al.,
1997a) did not, possibly due to very low levels of 63. Collectively, however, these studies
indicate that 63 may be associated with short-term declines in lung function in asthmatic
individuals and that they occurred at concentrations below those used in chamber studies using
exercise (CD, p. 7-40 to 7-46).
       Mortimer et  al. (2002) discussed biological mechanisms for delayed effects on pulmonary
function in asthma, which included increased nonspecific airway responsiveness secondary to
airway inflammation due to 63 exposure (CD, p. 7-43). Animal toxicological and human
chamber studies (CD, Chapters 5 and 6) provide supporting evidence that exposure to  Os may
augment cellular infiltration and cellular activation, enhance release of cytotoxic inflammatory
mediators, and alter membrane permeability  (CD, p.7-44). In most laboratory animals studied,
biochemical markers of lung injury and associated morphological changes were not found to be
attenuated, even though at similar exposures  pulmonary function changes might be attenuated.
       Most of the panel studies which have investigated associations between O3 exposure and
respiratory symptoms or increased use of asthma medication are focused on asthmatic children
(CD, sections 7.2.4 and 8.4.4.1).  Two large U.S. studies (Mortimer et al., 2002; Gent  et al.,
2003), as well as several smaller U.S.  (Delfmo et al., 2003; Just et al., 2002; Newhouse et al.,
2004; Romieu  et al., 1996, 1997; Ross et al.,  2002; Thurston et al., 1997) and international
                                         3-10

-------
studies (Hilterman et al., 1998; Desqueyroux et al., 2002a,b), have reported fairly robust
associations between ambient 63 concentrations and daily symptoms/asthma medication use,
even after adjustment for copollutants.
       The National Cooperative Inner-City Asthma Study (NCICAS) reported morning
symptoms in 846 asthmatic children from 8 U.S. urban areas to be most strongly associated with
a cumulative 1- to 4-day lag of Os concentrations (Mortimer et al., 2002). The NCICAS used
standard protocols that included instructing caretakers of the subjects to record symptoms in the
daily diary by observing or asking the child (Mitchell et al., 1997).  Symptoms reported included
cough, chest tightness, and wheeze. In the analysis pooling individual subject data from all eight
cities, the odds ratio for the incidence of symptoms was 1.35 (95% CI:  1.04, 1.69) per 30 ppb
increase in 8-hr avg Os (10 a.m.-6 p.m.). The mean 8-hr avg Os was 48 ppb across the 8 cities.
Excluding days when 8-hr avg 63 was greater than 80 ppb (less than 5% of days), the odds  ratio
was 1.37 (95% CI:  1.02, 1.82) for incidence of morning symptoms.
       Gent and colleagues (2003) followed 271 asthmatic children under age 12 and living in
southern New England for 6 months (April through September) in a diary study of daily
symptoms in relation to Os and PM2.5. Mean 1-hr max Os and 8-hr max Os concentrations were
58.6 ppb (SD 19.0) and 51.3 ppb (SD 15.5), respectively. The data were analyzed for two
separate groups of subjects, 130 who used maintenance asthma medications during the follow-up
period and 141 who did not.  The need for regular medication was considered to be a proxy for
more severe asthma. Not taking any  medication on a regular basis and not needing to use a
bronchodilator would suggest the presence of very mild asthma. Effects of 1-day lag 63 were
observed on a variety of respiratory symptoms only in the medication user group. Both daily 1-
hr max and 8-hr max 63 concentrations were similarly related to symptoms such as chest
tightness and shortness of breath. Effects of O3, but not PM2.s, remained significant and even
increased in magnitude in two-pollutant models. Some of the associations were noted at 1-hr
max 63 levels below 60 ppb.  In contrast, no effects were observed among asthmatics not using
maintenance medication.  In terms of person days of follow-up, this is one of the larger studies
currently available that address symptom outcomes in relation to Os, and provides supportive
evidence for effects of 63 independent of PM2.5. Study limitations include limited control for
meteorological factors and the post-hoc nature of the population stratification by medication use
(CD, p. 7-53).
       The multicities study by Mortimer et al. (2002), which provides an asthmatic population
most representative of the United States, and several  single-city studies indicate a robust
association of 63 concentrations with respiratory symptoms and increased medication use in
asthmatics. While there are a number of well-conducted, albeit relatively smaller, studies which
showed only limited or a lack of evidence for symptom increases associated with Os exposure,
                                         3-11

-------
these studies had less statistical power and/or were conducted in areas with relatively low Os
levels (CD, p. 7-54).  The CD (p. 7-55) concludes that the asthma panel studies, as a group, and
the NCICAS in particular, indicate a positive association between ambient concentrations and
respiratory symptoms and increased medication use in asthmatics.  The evidence has continued
to expand since 1996 and now is considered to be much stronger than in the previous review of
the Os primary standard.
       The association between school absenteeism and ambient 63 concentrations was assessed
in two relatively large field studies (CD, section 7.2.6). Chen et al. (2000) examined daily
school absenteeism in 27,793 elementary school students in Nevada over a 2-year period (after
adjusting for PM10 and CO concentrations) and found that ambient 63 concentrations were
associated with 10.41% excess rate of school absences per 40 ppb increase in 1-hr max Os with a
distributed lag of 1 to 14 days. Gilliland et al. (2001) studied (Vrelated absences among 1,933
4th grade students in 12 southern California communities and found significant associations
between 30-day distributed lag of 8-hr average Os concentrations and all absence categories, and
particularly for respiratory causes. Neither PM10 nor NO2 were associated with any respiratory
or nonrespiratory illness-related absences in single pollutant models. The CD concludes that
these studies of school absences suggest that ambient Os concentrations, accumulated over two
to four weeks, may be associated with school absenteeism, and particularly illness-related
absences, but further replication is needed before firm conclusions can be reached regarding the
effect of 63 on school absences (CD, p. 7-60).

          3.3.1.1.2  Airway  Responsiveness
       Airway hyperresponsiveness (AHR), also know as bronchial hyperreactivity, refers to a
condition in which the propensity for the airways to bronchoconstrict due to a variety of stimuli
(e.g., exposure to cold air, allergens, or exercise) becomes augmented (CD, section 6.8). This
condition is typically quantified by measuring the decrement in pulmonary function (e.g.,
spirometry or plethysmography) after inhalation exposure to specific (e.g., antigen, allergen) or
nonspecific (e.g., methacholine, histamine) bronchoconstrictor stimuli. Exposure to 63 causes an
increase in nonspecific  airway responsiveness as indicated by a reduction in the concentration of
methacholine or histamine required to produce a given reduction in FEVi or increase in SRaw.
Increased airway responsiveness is an important consequence of exposure to 63 because its
presence means that the airways are predisposed to narrowing on inhalation of various stimuli,
such as  specific allergens, cold air or SO2 (CD, p.  8-21). Significant, clinically relevant
decreases in pulmonary function have been observed in early phase allergen response in subjects
with rhinitis after consecutive (4-day) exposure to 0.125 ppm Os (Holz et al.,  2002).  Similar
increased airway responsiveness in asthmatics to house dust mite antigen 16 to 18 hrs after
                                          3-12

-------
exposure to a single dose of Os (0.16 ppm for 7.6 hrs) was observed. These observations suggest
that 63 exposure may be a clinically important factor that can exacerbate the response to ambient
bronchoconstrictor substances in individuals with preexisting allergic asthma. Further, Os may
have an immediate impact on asthmatics as well as contribute to effects that persist for longer
periods  (CD, p. 8-21).
       An important aspect of increased airway responsiveness after Os exposure is that it
represents a plausible link between 63 exposure and increased hospital admissions. Kreit et al.
(1989) found that Os can induce increased airway responsiveness in asthmatic subjects to Os,
who typically have increased airway responsiveness at baseline.  A subsequent study (Torres et
al., 1996) suggested an increase in specific (i.e., allergen-induced) airway reactivity in subjects
with allergic asthma, and to a lesser extent in subjects with allergic rhinitis after exposure to 0.25
ppm 63 for 3 hrs; other studies (Molfino et al., 1991; Kehrl et al., 1999) reported similar results.
According to one study (Folinsbee and Hazucha, 2000), changes in airway responsiveness after
Os exposure resolve more slowly than changes in FEVi or respiratory symptoms. Other studies
of repeated exposure to 63 suggest that changes in airway responsiveness tend to be somewhat
less affected by attenuation with consecutive exposures than changes in FEVi (Dimeo et al.,
1981; Folinsbee et al., 1994; Gong et al., 1997a; Kulle et al., 1982) (CD, p. 6-31).
       An extensive laboratory animal data base exploring the effects of acute, long-term, and
repeated exposure,  at rest, to O3 indicates that induction of AHR occurs at relatively high
(>lppm) 63 concentrations (p.  8-21).  These studies provide clues to the roles of physiological
and biochemical components involved in this process, but caution should be exercised in
interpreting these results, as different mechanisms may be involved in mediating high- and low-
dose responses. As observed in humans, the acute changes in AHR do not persist after long-term
exposure of animals exposed to near-ambient concentrations of O3, and attenuation has been
reported. In addition, dosimetric adjustments potentially could be made to allow better
estimation of levels that would be relevant to human exposure effect levels.
       The CD concludes that Os exposure is linked with increased AHR (CD, section 6.8).
Both human and animal studies indicate that AHR is not mechanistically associated with
inflammation, but they do suggest a likely role for neuronal involvement (CD, p. 8-21). Increases
in AHR do not appear to be strongly associated with decrements in lung function or increases in
symptoms (CD, p. 6-31).

         3.3.1.1.3  Respiratory Inflammation and Permeability
       Based on evidence from the previous review, acute inflammatory responses in the lung
have been observed subsequent to 6.6 hr Os exposures to the lowest tested level of 0.08 ppm in
healthy  adults engaged in moderately high exercise. Some studies suggest that inflammatory
                                         3-13

-------
responses may be detected in some individuals following Os exposures in the absence of O^-
induced pulmonary decrements in those subjects. Short-term exposures to 63 also can cause
increased permeability in the lungs of humans and experimental animals (CD, sections 5.2.3, 6.9,
7.2.5 and 8.4.3).  Not only are the newer findings consistent with the previous review, but also
there is better characterization of the physiological mechanisms by which O3 causes these
effects.
      Lung inflammation and  increased permeability, which are distinct events controlled by
different mechanisms, are two well characterized effects of Os exposure observed in all species
studied.  Disruption of the lung  barrier leads to leakage of serum proteins, influx of
polymorphonuclear leukocytes  (PMNs), release of bioactive mediators, and movement of
compounds from the airspaces into the blood.
      In the animal toxicological studies discussed in the CD (Chapter 5), the lowest 63
concentration that induced inflammation in the mouse lung was 0.11 ppm for 24 hr exposures.
Shorter exposures of 8 hours required concentrations of 0.26 ppm to induce epithelial
permeability, although there was no effect on inflammation. The lowest 63 concentration that
affected epithelial permeability  or inflammation in the rat was 0.5 ppm for a 3 hr exposure or
0.12 ppm for 6 hr (CD, p. 8-23). After acute exposures, the influence of the duration of exposure
increases as the concentration of 63 increases; however, dosimetric adjustments would need to be
done before one can compare levels.  The exact role of inflammation in causation of lung disease
is not known; nor is the relationship between inflammation and lung function (CD, p. 5-23).
      A number of human  (Vexposure studies have analyzed bronchoalveolar lavage (BAL)
and nasal lavage  (NL) fluids and cells for markers of inflammation and lung damage. These
studies are summarized in the CD (Annex AX6, Tables AX6-12 and AX6-13).  Increased  lung
inflammation is demonstrated by the presence of neutrophils (PMNs) found in BAL fluid in the
lungs, which has  long been accepted as a hallmark of inflammation. It is apparent, however, that
inflammation within airway  tissues may persist beyond the point that inflammatory cells are
found in the BAL fluid. Soluble mediators of inflammation, such as cytokines and arachidonic
acid metabolites have been measured in the BAL fluid of humans exposed to 03. In addition to
their role in inflammation, many of these compounds have bronchoconstrictive properties  and
may be involved  in increased airway responsiveness following Os exposure (CD, p. 6-31,  p. 8-
22). An in vitro study of epithelial cells from nonatopic and atopic asthmatics exposed to  0.01 to
0.10 ppm O3 showed significantly increased permeability compared to cells from normal
persons. This indicates a potentially inherent susceptibility of cells from asthmatic individuals
for (Vinduced permeability.
      In the 1996  CD, assessment of human exposure studies indicated that a single, acute (1 to
4 hr) Os exposure (> 0.08 to 0.1 ppm) of subjects engaged in moderate to heavy exercise could
                                         3-14

-------
induce a number of cellular and biochemical changes suggestive of pulmonary inflammation and
lung permeability (CD, p. 8-22). These changes persisted for at least 18 hrs.  Graham and Koren
(1990) compared inflammatory mediators present in NL and BAL fluids of humans exposed to
0.4 ppm 63 for 2 hrs and found similar increases in PMNs in both fluids, suggesting a qualitative
correlation between inflammatory changes in the lower airways (BAL) and upper respiratory
tract (NL). Acute airway inflammation was shown in Devlin et al. (1990) to occur among adults
exposed to 0.08 ppm 63 for 6.6 hr with exercise, and McBride et al. (1994) reported that
asthmatic  subjects were more sensitive than non-asthmatics to upper airway inflammation for Os
exposures (0.24 ppm, 1.5 hr, with light IE) that did not affect pulmonary function (CD, p. 6-33).
       The studies reporting inflammatory responses and markers of lung injury have clearly
acknowledged that there is significant variation in response of subjects exposed, especially to 6.6
hour 63 exposures at 0.08 and 0.10 ppm.  To provide some perspective on the public health
impact for these effects, we note that one study (Devlin et al., 1991, Figure 5) showed that
roughly 10 to 50% of the 18 young healthy adult subjects experienced notable increases (i.e., > 2
fold increase) in most of the inflammatory and cellular injury indicators analyzed, associated
with 6.6-hour exposures  at 0.08 ppm.  Similar, although in some cases higher, fractions of the
population of 10 healthy adults tested saw > 2 fold increases associated with 6.6-hour exposures
to 0.10 ppm.  The authors of this study suggest that "susceptible subpopulations such as the very
young, elderly, and people with pulmonary impairment or disease may be even more affected"
(Devlin et al., 1991).
       Since 1996, a substantial number of human exposure studies have been published which
have provided important new information on lung inflammation and epithelial permeability.
Mudway and Kelly (2004) examined (Vinduced inflammatory responses and epithelial
permeability with a meta-analysis of 21 controlled human exposure studies and showed that
PMN influx in healthy subjects is associated with total  Os dose ( product of Os concentration,
exposure duration, and minute ventilation) (CD, p. 6-34). Results of the analysis suggest that the
time course for inflammatory responses (including recruitment of neutrophils and other soluble
mediators) is not clearly  established, but differential attenuation profiles for many of these
parameters are evident (CD, p. 8-22).
       A number of studies (Peden  et al., 1997; Scannell et al., 1996; Hiltermann et al., 1999;
Bosson et  al., 2003) have provided evidence suggesting that asthmatics show greater
inflammatory response than healthy subjects when exposed to similar O3 levels (CD, section
6.9). Markers from BAL fluid following both 2-hr (Devlin et al., 1997) and 4-hr (Christian et al.,
1998; Torres et al., 2000) 63 exposures repeated up to 5 days indicate that there is ongoing
cellular damage irrespective of attenuation of some cellular inflammatory responses of the
airways, pulmonary function, and symptom responses (CD, p. 8-22).
                                         3-15

-------
       The CD (p. 8-24) concludes that interaction of Os with lipid constituents of epithelial
lining fluid (ELF) and cell membranes and the induction of oxidative stress is implicated in
injury and inflammation.  Alterations in the expression of cytokines, chemokines, and adhesion
molecules, indicative of an ongoing oxidative stress response, as well as injury repair and
regeneration processes, have been reported in animal toxicology and human in vitro studies
evaluating biochemical mediators implicated in injury and inflammation. While antioxidants in
ELF confer some protection, 63 reactivity is not eliminated at environmentally relevant
exposures.  Further, antioxidant reactivity with Os is both species-specific and dose-dependent
(CD, p. 8-24).

         3.3.1.1.4 Changes in Host Defense Capability
       As discussed in the CD (sections 5.2.2, 6.9.6, and 8.4.2), short-term exposures to 63 have
been shown to impair host defense capabilities in both humans and experimental animals by
depressing alveolar macrophage (AM) functions and by altering the mucociliary clearance of
inhaled particles and microbes.  Short-term O3 exposures also interfere with the clearance
process by accelerating clearance for low doses and slowing clearance for high doses.  Animal
toxicological studies have reported that acute 63 exposures suppress alveolar phagocytosis and
immune functions. Dysfunction of host defenses and subsequent increased susceptibility to
bacterial lung infection in laboratory animals has been induced by short-term exposures to Os
levels as low as 0.08 ppm (CD, p. 8-26).
       Changes in antibacterial defenses are dependent on exposure regimens, species and strain
of lab animals, species of bacteria, and age of the animals used. Acute Os-induced suppression
of alveolar phagocytosis and immune function in experimental animals appeared to be transient
and attenuated with continuous or repeated exposures. Ozone exposure has also been shown to
interfere with AM-mediated clearance in the respiratory region of the lung and with mucociliary
clearance of the tracheobronchial airways. These interferences with clearance are dose
dependent, with low doses accelerating clearance and high doses  slowing the process (CD, p. 8-
26).
       A single controlled human exposure study (Devlin et al., 1991) reviewed in the 1996 CD
reported that exposure to 0.08 to 0.10 ppm Os for 6.6 hrs (with moderate exercise) induced
decrements in the ability of AMs to phagocytose microorganisms (CD, p. 8-26).  Integrating the
recent study results with evidence available in the 1996 CD, the CD concludes that available
evidence indicates that short-term 63 exposures have the potential to impair host defenses,
primarily by interfering with AM function. Any impairment in AM function may lead to
decreased clearance of microorganisms or nonviable particles. Compromised AM functions in
                                         3-16

-------
asthmatics may increase their susceptibility to other O?, effects, the effects of particles, and
respiratory infections (CD, p. 8-26).

          3.3.1.1.5 Morphological Effects
       The 1996 CD found that short-term Os exposures cause similar alterations in lung
morphology in all laboratory animal species studied, including primates. Cells in the
centriacinar region (CAR) of the lung (the segment between the last conducting airway and the
gas exchange region) have been recognized as a primary target of Os-induced damage (epithelial
cell necrosis and remodeling of respiratory bronchioles), possibly because epithelium in this
region receives the greatest dose of Os delivered to the lower respiratory tract. Following
chronic Os exposure, structural changes have been observed in the CAR, the region typically
affected in most chronic airway diseases of the human lung (CD, p. 8-24).
       Ciliated cells in the nasal cavity and airways, as well as Type I cells in the gas-exchange
region, are also identified as targets. While short-term 63 exposures can cause epithelial cell
profileration and fibrolitic changes in the CAR, these changes appear to be transient with
recovery time after exposure, depending on species and Os dose. The potential impacts of
repeated  short-term and chronic morphological effects of 63 exposure are discussed later in
section 3.3.1.2.5.
       Recent studies continue to show that short-term and sub-chronic exposures to Os cause
similar alterations in lung structure in a variety of experimental animal species, at concentrations
of 0.15 ppm in rats (12 hr/day for 6 weeks) and even lower concentrations in primates (8 hr/day
for 90 days) (CD, section 5.2.4.). Recent work has shown that a topical anti-inflammatory
corticosteroid can prevent these effects in nasal epithelia, while exposure to bacterial endotoxin
can potentiate effects. Ozone-induced fibrotic changes in the CAR are maximal at 3 days of
exposure and recover 3 days post-exposure with exposures of 0.2 ppm 63 in rodents. One study
has demonstrated variability of local O3 dose and subsequent injury in the respiratory tract due to
depletion of glutathione (GSH). The proximal respiratory bronchiole receives the most acute
epithelial injury from exposures < 1 ppm, while metabolic effects were greatest in the distal
bronchioles and minor daughter airways (CD, p. 5-38).
       Based on evidence from animal toxicological studies, short-term and sub-chronic
exposures to 63 can cause morphological changes in the respiratory systems, particularly in the
CAR, of a number of laboratory animal species (CD, section 5.2.4).

          3.3.1.1.6 Emergency Department Visits/Hospital A dmissions for Respiratory
                   Causes
       The 1996 CD evaluated ED visits and hospital admissions as possible outcomes
following exposure to O3 (CD, section 7.3).  The evidence was limited for ED visits, but results
                                          3-17

-------
of several studies generally indicated that short-term exposures to O3were associated with
respiratory ED visits.  The strongest and most consistent evidence, both below and above 0.12
ppm 1-hr max O3, was found in the group of studies which investigated summertime daily
hospital admissions for respiratory causes in different eastern North American cities. These
studies were consistent in demonstrating that ambient O3 levels were associated with increased
hospital admissions and accounted for about one to three excess respiratory hospital admissions
per million persons with each 100 ppb increase in 1-hr max O3, with adjustment for possible
confounding effects of temperature and copollutants. Overall, the 1996 CD concluded that there
was strong evidence that ambient O3 exposures can cause significant exacerbations of preexisting
respiratory disease in the general public (CD, p. 7-66).  Excess respiratory-related hospital
admissions associated with O3 exposures for the New York City area (based on Thurston et al.,
1992) were included in the quantitative risk assessment in the prior review and are included in
the current assessment along with estimates for respiratory-related hospital admissions in
Cleveland, Detroit, and Los Angeles based on more recent studies (see Chapter 5). Significant
uncertainties and the difficulty of obtaining reliable baseline incidence numbers resulted in ED
visits not being used in the quantitative risk assessment conducted in the last O3 NAAQS review.
       In the past decade,  a number of studies have examined the temporal pattern associations
between O3 exposures and  ED visits for respiratory causes (CD, section 7.3.2).  These studies are
summarized in the CD (Table AX7-3, Chapter 7 Annex).  Respiratory causes for ED visits
include asthma, bronchitis, emphysema, pneumonia, and other upper and lower respiratory
infections, such as influenza, but asthma visits typically dominate the daily incidence counts.
Among studies with adequate controls for seasonal patterns, many reported at least one
significant positive association involving O3. These studies examined ED visits for total
respiratory complaints (Delfmo et al., 1997b, 1998b; Hernandez-Garduno et al., 1997; Ilabaca et
al., 1999; Lin et al., 1999), asthma (Friedman et al., 2001; Jaffe et al., 2003; Stieb et al.,  1996;
Tenias et al.,  1998; Tobias et al., 1999 ; Tolbert et al., 2000 ; Weisel et al., 2002), and COPD
(Tenias et al., 2002).
       Figure 7-8 (CD, p.  7-68) provides effect estimates for associations between ED visits for
asthma and short-term O3 exposures.  In general, O3 effect estimates from summer only analyses
tended to be positive and larger compared to results from cool season or all year analyses (CD, p.
7-67).  Several of the studies reported significant associations between O3 concentrations and ED
visits for respiratory causes.  However, inconsistencies were observed which were at least
partially attributable to differences in model specifications and analysis approach among various
studies. For example, ambient O3 concentrations, length of the study period, and statistical
methods used to control confounding by seasonal patterns and copollutants appear to affect the
observed O3 effect on ED visits. Thus, the CD (p. 7-71) has concluded that stratified analyses by
                                          3-18

-------
season generally supported a positive association between O3 concentrations and ED visits for
asthma in the warm season.
       Unscheduled hospital admissions occur in response to unanticipated disease
exacerbations and are more likely to be affected by environmental factors, such as high O3 levels.
Thus, hospital admissions studies focus specifically on unscheduled admissions. Results of a
fairly large number of these studies published during the past decade are summarized in Table
AX7-4 (CD, Chapter 7 Annex). As a group, these hospital admissions studies tend to be larger
geographically and temporally than the ED visit studies and provide results that are generally
more consistent. The largest and most significant associations of respiratory hospital admissions
with O3 concentrations were observed using short lag periods, in particular for a 0-day lag (same
day exposure) and a 1-day lag (previous day exposure).  Most studies in the United States and
Canada indicated positive, statistically significant associations between ambient 63
concentrations and respiratory hospital admissions in the warm season, including studies with
98th percentile 8-hr maximum Os levels as low as about 50 ppb. However, not all studies found
a statistically significant relationship with 63, possibly because of very low ambient 63 levels.
Analyses for confounding using multipollutant regression models suggest that copollutants
generally do not confound the association between Os and respiratory hospitalizations. Ozone
effect estimates were robust to PM adjustment in all-year and warm-season only data.
       Overall, the CD concludes that positive and robust associations were found between
ambient Os concentrations and various respiratory disease hospitalization outcomes, when
focusing particularly on results of warm-season analyses. Recent studies also generally
supported a positive association between O3 concentrations and ED visits for asthma during the
warm season (CD,  p. 7-175).  These observations are strongly supported by the human clinical,
animal toxicologic, and epidemiologic evidence for lung function decrements, increased
respiratory  symptoms, airway inflammation, and increased airway responsiveness.  Taken
together, the overall evidence supports a causal relationship between acute ambient Os exposures
and increased respiratory morbidity outcomes resulting in increased ED visits and
hospitalizations during the warm season (CD, p. 8-77).

         3.3.1.1.7 Effects on Exercise Performance
       The effects of Os exposure on exercise performance of healthy individuals have been
investigated in a number of controlled exposure studies (CD, section 6.7). Several  studies
discussed in the  1996 CD reported that endurance exercise performance and VO2max may be
limited by acute exposure to O3. Other studies found that significant reductions in maximal
endurance exercise performance may occur in well-conditioned athletes while they perform CE
(VE > 80 L/min) for 1 hr at Os concentrations > 0.18 ppm. There are no new studies available in
                                          3-19

-------
the CD. Thus, as in the 1996 CD, the CD concludes that reports from studies of Os exposure
during high-intensity exercise indicate that breathing discomfort associated with maximal
ventilation may be an important factor in limiting exercise performance in some, but not all,
subjects (CD, p. 6-30).

        3.3.1.2      Effects on the Respiratory System from Long-term Exposures
       The 1996 CD concluded that there was insufficient evidence from the limited number of
studies to determine whether long-term Os exposures resulted in chronic health effects at
ambient levels observed in the U.S. However, the aggregate evidence suggested that 63
exposure, along with other environmental factors, could be responsible for health effects in
exposed populations (CD, section 7.5).  Animal toxicological studies carried out in the 1980's
and 1990's demonstrated that long-term exposures can result in a variety of morphological
effects, including permanent changes in the small airways of the lungs, including remodeling of
the distal airways and CAR and deposition of collagen, possibly representing fibrotic changes.
These changes result from the damage and repair processes that occur with repeated exposure.
Fibrotic changes were also found to persist after months of exposure providing a potential
pathophysiologic basis for changes in airway function observed in children in some recent
epidemiological studies. It appears that variable seasonal ambient patterns  of exposure may be
of greater concern than continuous daily exposures.
       This section reviews studies published since 1996 in which health effects were assessed
for Os exposures lasting from weeks to several years.  Summaries of recent morphological
effects studies of subchronic and chronic exposures are listed in Table AX5-10 (CD, Annex
AX5).  Summaries of recent morbidity effects epidemiological studies of long-term exposure are
listed in Table AX7-6 (CD, Annex AX7).

         3.3.1.2.1 Seasonal Ozone Effects on Lung Function
       It is well documented in controlled human exposure and field studies that daily multi-
hour exposures to O3 produce transient declines in lung function; however, lung function effects
of repeated exposures to Os over extended periods are not as well characterized. Several studies
published since 1996 have investigated lung function changes over seasonal time periods (CD,
section 7.5.3). One large, three-year study (Frischer et al., 1999) collected repeated lung
function measurements in 1,150 young, Austrian school children in 9 communities and reported
that there may be an association between developmental changes in lung function over the
summer season and seasonal mean Os levels.  Mean summertime 24-hr avg Os concentrations
during the three summers was 34.8 ppb (SD 8.7). The number of days with half-hour maximum
O3 concentrations greater than 60 ppb ranged from 44 to 99 days across the 9 communities.
Seasonal mean Os was associated with reduced lung function development.  It was cautioned that
                                         3-20

-------
it was difficult to attribute the reported effects to Os alone independently of copollutants (CD, p.
7-113).  A one-year extension of this study by Horak et al. (2002a,b) confirmed the results that
seasonal mean Os levels may be related to a negative effect on increases in lung function in
children. A study (Kopp et al., 2000) of 797 children in Austria and southwestern Germany
reported smaller increases in lung function in children exposed to higher levels of ambient Oj
(mean Os concentration of 44 to 52 ppb) compared to children living in areas with lower ambient
63 levels (25 to 33 ppb).  Another Austrian study (Ihorst et al., 2000) of 2,153 young children
found significantly lower FVC  and FEVi increases associated with higher Os exposures in the
summer but not in the winter.  A pilot study (Kinney and Lippmann, 2000) of 72  young adult,
military academy  students provided results that are consistent with a seasonal decline in lung
function that may be due, in part, to Os exposures.  According to the CD (p. 7-114), these studies
collectively indicate that seasonal 63 exposure is associated with smaller growth-related
increases in lung function in children than they would have experienced living in clean air and
that there is some  limited evidence that seasonal Os also may affect lung function in young
adults, although uncertainty about the role of copollutants makes it difficult to attribute the
effects to Os alone.

         3.3.1.2.2  Reduced Baseline Lung Function and Respiratory Symptoms
      Lung capacity grows during childhood and adolescence as body size increases, reaches a
maximum during the twenties, and then begins to decline steadily and progressively with age.
Long-term  exposure to air pollution has long been thought to contribute to slower growth in lung
capacity, diminished maximally attained capacity, and/or more rapid decline in lung capacity
with age (CD, section 7.5.4). Toxicological findings evaluated in the 1996 CD demonstrated that
repeated daily exposure of rats to an episodic profile of Os caused small, but significant,
decrements in growth-related lung function that were consistent with early indicators of focal
fibrogenesis in the proximal alveolar region, without overt fibrosis (CD, section 5.2.5.2).
Because Os is a strong respiratory irritant and has been  shown to cause inflammation and
restructuring of the respiratory airways, it is plausible that long-term 63 exposures might have a
negative impact on baseline lung function, particularly during childhood when these exposures
might have long-term risks. As noted in the current CD, however, no recent toxicological studies
have been published on effects of chronic 63 exposure.
       Several epidemiological studies published since 1996 have examined the relationship
between lung function development and long-term 63 exposure. The most extensive and robust
study of respiratory effects in relation to long-term air pollution exposures among children in the
U.S. is the Children's Health Study carried out in 12 communities of southern California starting
in 1993  (Avol et al., 2001; Gauderman et al., 2000, 2002, 2004a,b; Peters et al., 1999a,b). One
                                         3-21

-------
study (Peters et al., 1999a) examined the relationship between long-term Os exposures and self
reports of respiratory symptoms and asthma in a cross sectional analysis and found a limited
relationship between outcomes of current asthma, bronchitis, cough and wheeze and a 40 ppb
increase in 1-hr max 63 (CD, p. 7-115). Another analysis (Peters et al., 1999b) examined the
relationship between lung function at baseline and levels of air pollution in the community and
reported evidence that annual mean Os levels were associated with decreases in FVC, FEVi, PEF
and FEF25-75  (the latter two being statistically significant) among females but not males (CD, p. 7-
116). In a separate study (Gauderman et al.,  2000) of 4th, 7th, and  10th grade students, a
longitudinal  analysis of lung function development over four years found no association with Os
exposure.  Subsequent studies by the same group (Gauderman et al., 2002, 2004a,b) led the
authors to conclude that results provide little evidence that ambient Os at current levels is
associated with chronic deficits in the rate of increase in growth-related lung function in children
(CD, p. 7-116 to 7-118). Avol et al. (2001) examined children who had moved from
participating communities in southern California to other states with improved air quality and
found, with the exception of FEVi, the 63 effect estimates for all other spirometric parameters
were negative, but the associations were not as strong as  those observed for PMi0 (CD, p. 7-116).
Collectively, the results of these reports from the children's health cohorts provide little evidence
for impact of long-term 63 exposures on lung function development (CD, p. 7-122).
      Evidence for a significant relationship between long-term Os exposures and decrements
in maximally attained lung function was reported in a nationwide  study of first year Yale
students (CD, p. 7-120). Males had much larger effect estimates than females, which might
reflect higher outdoor activity levels and correspondingly higher Os exposures during childhood.
A similar study (Kunzli et al.,  1997; Tager et al., 1998) of college freshmen at University of
California at Berkeley also reported significant effects of long-term O3 exposures on lung
function (CD, p. 7-121). In a comparison of students whose city of origin was either Los
Angeles or San Francisco, long-term 63 exposures were associated with significant changes in
mid- and end-expiratory flow measures, which could be considered early indicators for
pathologic changes that might progress to COPD.
      In summary, recent publications from the southern California children's cohort study
provide no evidence for an association between long-term Os exposure  and lung function
development in children (CD,  p. 7-118), while limited evidence available from studies of adults
and college students suggest that long-term O3 exposure may affect lung function or respiratory
symptoms (CD, pp. 7-120, 7-121). Overall, the CD concluded that this body of evidence was
inconclusive for effects of long-term 63 exposure on respiratory symptoms or lung function (CD,
p. 7-175).
                                         3-22

-------
         3.3.1.2.3  Long-term O3 Exposure and Respiratory Inflammation
       As noted above in section 3.3.1.1.3 and in the CD (Chapter 6), chamber studies of
exercising humans exposed to Os for 2 to 6.6 hrs have demonstrated inflammation in the lungs,
including the alveolar region where gas exchange takes place. The potential long-term
significance of short-term exposures to 63 is that they can result in the release of reactive
substances from inflammatory cells that can damage the sensitive cells lining the lungs. Over
time repeated inflammation can lead to permanent lung damage and restructuring of the small
airways and alveoli. Also, since inflammation is a hallmark characteristic of asthma, there is the
possibility that Os-induced inflammation may exacerbate existing asthma or contribute to the
development of asthma in genetically predisposed individuals (CD, section 7.5.5).
       For subchronic exposures of animals, permeability changes are transient (and species-
dependent) and return to control levels even with continuing exposure. For long-term Os
exposures, persistent (Vinduced inflammation plays an important role in alterations of lung
structure and function. Significant remodeling of the epithelium and underlying connective
tissues in distal  airways have been reported in rats exposed to 0.25 ppm 63 (12 hr/day for 6
weeks) and in monkeys exposed to 0.15 ppm O3 (8 hr/day for 90 days)(CD, p. 8-23).
       In one epidemiological field study (Kinney et al.,  1996), BAL fluids were taken in the
summer and winter from a group of joggers in New  York and were compared for evidence of
acute inflammation and of enhanced cell damage (CD, p. 7-122). The mean 1-hr max
concentrations for a 3-month period were 58 ppb (max 110 ppb) in the summer and 32 ppb (max
64 ppb) in the winter. There was little evidence of an association between 63 and acute
inflammation in the summer BAL fluids compared to winter, but there was evidence of enhanced
cell damage. This suggests that even though inflammation may diminish over the summer, cell
damage may be continuing.  A cross-sectional cohort study  (Calderon-Garciduenas et al., 1995)
conducted in Mexico City provides evidence of inflammation and genetic damage to cells in the
nasal passages of children chronically exposed to 63 and other air pollutants (CD, p. 7-123).  In
Mexico City, the 1-hr avg O3 concentrations exceeded 120 ppb for 4.4  hr/day, on average.
Significantly higher DNA damage was reported in children  living in Mexico City compared to
nonurban children and in older compared to younger children. Another marker of inflammation,
urinary eosinophils, was analyzed in an Austrian school children study (Frischer et al., 2001),
and it was reported that Os exposure (mean 30 day avg Os concentration before sample collection
was 31.6 ppb) was significantly associated with eosinophil inflammation (CD, p. 7-122).
       In assessing these studies, the CD (p. 7-123) concluded that specific attribution of these
adverse respiratory and genotoxic effects to 63 is difficult given the complex mixture in ambient
air, although inflammatory changes like eosinophil levels observed in the Austrian study would
be consistent with known effects of Os.
                                         3-23

-------
         3.3.1.2.4  Risk of Asthma Development
       There have been a few studies investigating associations between long-term 63 exposures
and the onset of new cases of asthma (CD, section 7.5.6). The Adventist Health and Smog
(AHSMOG) study cohort of 3,914 was drawn from nonsmoking, non-Hispanic white adult
Seventh Day Adventists living in California (Greer et al., 1993; McDonnell et al., 1999).
Subjects were surveyed in 1977, 1987, and 1992. During the ten-year follow-up in 1987, it was
reported that the incidence of new asthma was 2.1% for males and 2.2% for females (Greer et al.,
1993). A statistically significant relative risk of 3.12 (95% CI: 1.16, 5.85) per 10 ppb increase in
annual mean Os was observed in males, compared to a nonsignificant relative risk of 0.94 (95%
CI: 0.65, 1.34) in females. In the 15-year follow-up in  1992, it was reported that 3.2% of eligible
males and 4.3% of eligible females had developed adult asthma (McDonnell et al., 1999).  For
males, the relative risk of developing asthma was 2.27 (95% CI: 1.03, 4.87) per 30 ppb increase
in 8-hr average 63, but there was no evidence of an association in females. The lack of an
association in females does not necessarily mean there is no effect but may be due to differences
in time-activity patterns in males and females, which could lead to greater misclassification of
exposure in females.  Consistency of results in the two  studies with different follow-up times
provides supportive evidence of an association between long-term Os exposure and asthma
incidence in adult males; however, representativeness of this cohort to the general U.S.
population may be limited (CD, p. 7-125).
       In a similar study (McConnell et al., 2002) of incident  asthma among children (ages 9 to
16 at enrollment), annual surveys of 3,535 children initially without asthma were used to identify
new-onset asthma cases as part of the Children's Health Study. Six high-Os (75.4 ppb mean 1-hr
max over four years) and six low-Os (50.1 ppb, mean 1-hr max) communities were identified
where the children resided.  There were 265 children who reported new-onset asthma during the
follow-up period.  Although asthma risk was no  higher for all  residents of the six high-Os
communities versus the six low-Os communities, asthma risk was 3.3 times greater for children
who played three or more sports as compared with children who played no sports within the
high-Os communities.  This association was absent in the communities with lower Os
concentrations. No other pollutants were  found to be associated with new-onset asthma (CD, p.
7-125).
       Playing sports may result in extended outdoor activity  and  exposure occurring during
periods when 63 levels are higher.  The sports activities would cause an increased ventilation
rate, thus resulting in increased Os dose. It should be noted, however, that the results of the
Children's Health Study (McConnell et al., 2002) were based on a small number (20 in high-Os
areas and 9 in low- O3 areas) of new-onset asthma cases among children who played three or
                                         3-24

-------
more sports (CD, p. 7-125). Future replication of these findings in other cohorts would help
determine whether a causal interpretation is appropriate.

          3.3.1.2.5 Morphological Effects
       In animal toxicology studies, the progression of morphological effects reported during
and after a chronic exposure in the range of 0.5 to 1.0 ppm 63 is complex, with inflammation
peaking over the first few days of exposure, then dropping, then plateauing, and finally, largely
disappearing (CD,  section 5.2.4.4).  By contrast, fibrotic changes in the tissue increase very
slowly over months of exposure, and, after exposure ceases, the changes sometimes persist or
increase. Epithelial hyperplasia peaks soon after the inflammatory response but is usually
maintained in both the nose and lungs with continuous exposure. Epithelial
hyperplasia/metaplasia also does not return to pre-exposure levels after the end of exposure.
Patterns of exposure in this same  concentration range determine effects, with 18 months of daily
exposure, causing less morphologic damage than exposures on alternating months.  This is
important as environmental O3 exposure is typically seasonal. Long-term studies of Plopper and
colleagues (Evans et al., 2003; Schelegle et al., 2003; Chen et al., 2003; Plopper and Fanucchi,
2000) investigated infant rhesus monkeys exposed to simulated, seasonal 63 (0.5 ppm, 8 hrs/day
for 5 days, every 14 days for 11 episodes) and demonstrated: 1) remodeling in the distal airways,
2) abnormalities in tracheal basement membrane; 3) eosinophil accumulation in conducting
airways;  and 4) decrements in airway  innervation (CD,  p. 5-45). As with other effects, these
findings advance earlier information regarding possible injury-repair processes occurring with
long-term Os exposures suggesting that these processes are only partially reversible and may
progress  following cessation of 63 exposure. Further, these processes may lead to nonreversible
structural damage to lung tissue; however, there is still too much uncertainty to quantitatively
extrapolate these levels to human effect levels at this time (CD, p. 8-25).

          3.3.1.2.6 Summary
       In the past decade, important new longitudinal studies have examined the effect of
chronic Os exposure on respiratory health outcomes.  Evidence from recent long-term morbidity
studies have suggested in some cases that chronic exposure to 63 may be associated with
seasonal  declines in lung function or reduced lung function development, increases in
inflammation, and  development of asthma in children and adults. Seasonal decrements or
smaller increases in lung function measures have been reported in several studies; however, it
remains uncertain to what extent these changes are transient.  While there is supportive evidence
from animal studies involving chronic exposures, large uncertainties still remain as to whether
current ambient levels and exposure patterns might cause these same effects in human
                                         3-25

-------
populations. The CD also concludes that epidemiological studies of new asthma development
and longer-term lung function declines remain inconclusive at present (CD, p. 7-134).

        3.3.1.3     Effects on the Cardiovascular System
       At the time of the 1997 review, the possibility of (Vinduced cardiovascular effects was a
largely unrecognized issue.  Since then, a very limited body of evidence from animal, controlled
human exposure and epidemiologic studies has emerged that provides some potential plausible
mechanisms for how Os exposures might exert cardiovascular system effects, however much
needs to be done to substantiate these effects.  Possible mechanisms may involve (Vinduced
secretions of vasoconstrictive substances and/or effects on neuronal reflexes that may result in
increased arterial blood pressure and/or altered electrophysiologic control of heart rate or
rhythm. Some animal toxicology studies have shown (Vinduced decreases in heart rate, mean
arterial pressure, and core temperature.  One controlled human exposure study that evaluated
effects of 63 exposure on cardiovascular health outcomes found no significant (Vinduced
differences in ECG or blood pressure in healthy or hypertensive subjects but did observe a
significant (Vinduced increase the alveolar-to-arterial PC>2 gradient and heart rate in both groups
resulting in an overall increase in myocardial work and impairment in pulmonary gas exchange
(Gong et al., 1998). In another controlled human exposure study, inhalation of a mixture of
PM2 5 and  63by healthy subjects increased brachial artery tone and reactivity (Brook et al.,
2002).
       The evidence from a few animal studies also includes potential direct effects such as 
-------
hospital admissions in Toronto, Canada in a summer-only analysis (mean 1-hr max O3 of 41.2
ppb) (Burnett et al., 1997b).  The results were robust to adjustment for various PM indices,
whereas the PM effects diminished when adjusting for gaseous pollutants.  Other studies
stratified their analysis by temperature, i.e., by warms days (^ 20°C) versus cool days (< 20°C).
Several analyses using warms days consistently produced positive associations.
       The epidemiologic evidence for cardiovascular morbidity is much more mixed than for
respiratory morbidity, with only one of several U.S./Canadian studies showing statistically
significant positive associations of cardiovascular hospitalizations with warm-season O3
concentrations.  Most of the available European and Australian studies, all of which conducted
all-year O3 analyses,  did not find an association between short-term O3 concentrations and
cardiovascular hospitalizations.  Overall, the currently available evidence is inconclusive
regarding an association between cardiovascular hospital admissions and ambient O3 exposure
(CD, p. 7-83).
       Based on the  evidence from animal toxicology, human controlled exposure, and
epidemiologic studies, the CD concludes that this generally limited body of evidence is highly
suggestive that O3 can directly and/or indirectly contribute to cardiovascular-related morbidity,
but that much needs to be done to more fully substantiate links between ambient O3 exposures
and adverse cardiovascular outcomes (CD, p. 8-77).

       3.3.2  Premature Mortality
       There were only a limited number of studies which examined the relationship between O3
and mortality available for review in the 1996 CD. Some studies suggested that mortality was
associated with short-term exposure to O3, but conclusions could not be drawn regarding such
associations (CD, p. 7-84). Numerous recent studies have provided new and more substantial
evidence supporting such an association, as discussed below in section 3.3.2.1.
       At the time of the last review, little epidemiological evidence was available on potential
associations between long-term exposure to  O3 and mortality. Some recent studies have
evaluated this relationship and provide limited, if any, evidence for an association between
chronic O3 exposure  and mortality, as described in section 3.3.2.2.

        3.3.2.1    Mortality and Short-term Os Exposure
       The 1996 CD concluded that an association between daily mortality and O3 concentration
for areas with high O3 levels  (e.g., Los Angeles) was suggested. However, due to a very limited
number of studies available at that time, there was insufficient evidence to conclude that the
observed association  was likely causal, and thus the possibility that O3 exposure may be
associated with mortality was not relied upon in the 1997 decision on the O3 primary standard.
                                          3-27

-------
       The 2006 CD includes results from numerous epidemiological analyses of the
relationship between 63 and mortality. Key findings are available from multi-city time-series
studies that report associations between Os and mortality. These studies include analyses using
data from 90 U.S. cities in the National Mortality, Morbidity and Air Pollution (NMMAPS)
study (Dominici et al., 2003) and from 95 U.S. communities in an extension to the NMMAPS
analyses (Bell et al., 2004).  The analyses conducted by Huang et al. (2005) used a subset of 19
U.S. cities and focused primarily on cause-specific mortality associations during the warm
season.  An additional study (Schwartz, 2005) used case-crossover design and data from 14 U.S.
cities to further investigate the influence of adjustment for weather variables in the Os-mortality
relationship (CD, p. 8-38). Finally, results are available from a European study, Air Pollution
and Health: a European Approach (APHEA),  using data from 23 cities (Gryparis et al., 2004)
and 4 cities (Toulomi et al., 1997) (CD, p. 7-93).
        The original 90-city NMMAPS analysis, with data from 1987 to 1994, was primarily
focused on investigating effects of PMio on mortality. A significant association was reported
between mortality and 24-hr average 63 concentrations during the warm season, but the
association was not significant in analyses for the full year (Samet et al., 2000) (CD, Figure 7-21;
p. 7-98). This is because the estimate using all available data was about half that for the
summer-only data at a lag of 1-day. The extended NMMAPS  analysis included data from 95
U.S. cities and included an additional 6 years  of data, from 1987-2000 (Bell et al., 2004), and
significant associations were reported between 63 and mortality. The effect estimate for
increased mortality was 0.5% per 24-hr average 63 measured on the same day (20 ppb change;
95% PI: 0.24, 0.78), and 1.04% per 24-hr average O3 in a 7-day distributed lag model (20 ppb
change; 95% PI: 0.54, 1.55) (CD, p. 7-88).  In analyses using only data from the warm season,
the results were not significantly different from the full-year results; the effect estimate for
increased mortality was 0.44% per 24-hr average Os measured on the same day (20 ppb change;
95% PI: 0.14, 0.74), and 0.78% per 24-hr average O3 in a 7-day distributed lag model (20 ppb
change; 95% PI: 0.26, 1.30). The authors also report that Os-mortality associations were robust
to adjustment for PM (CD, p. 7-100).
       Using a subset of the NMMAPS data set, Huang et al. (2005) focused on associations
between cardiopulmonary mortality and Os exposure (24-hr avg) during the summer season only.
The authors report a 1.47% increase per 20 ppb change in 63 concentration measured on the
same day (95% PI: 0.54, 2.39) and a 2.52% increase per 20 ppb change in O3 concentration using
a 7-day distributed lag model (95% PI: 0.94, 4.10)(CD, p. 7-92). These findings suggest that the
effect of 63 on mortality is immediate but also persists for several days.
       As discussed below in section 3.4, confounding by weather, especially temperature, is
complicated by the fact that higher temperatures are associated with the increased photochemical
                                         3-28

-------
activities that are important for O3 formation.  Using a case-crossover study design, Schwartz
(2005) assessed associations between daily maximum concentrations and mortality, matching
case and control periods by temperature, and using data only from the warm season.  The
reported effect estimate of 0.92% change in mortality per 40 ppb O3 (1-hr max, 95% PI: 0.06,
1.80) was similar to time-series analysis results with adjustment for temperature (0.76% per 40
ppb O3, 95% PI, 0.13, 1.40), suggesting that associations between O3 and mortality were robust
to the different adjustment methods for temperature (CD, p. 7-93).
       An initial publication from APHEA, a European multi-city study, reported statistically
significant associations between daily maximum O3 concentrations and mortality, with an effect
estimate of a 4.5% increase in mortality per 40 ppb O3 (95% CI:  1.6, 7.7) in four cities (Toulomi
et al., 1997).  An extended analysis was done using data from 23 cities throughout Europe
(Gryparis et al., 2004). In this report, a positive but not statistically significant association was
found between mortality and 1-hr daily maximum O3 in a full year analysis (CD, p. 7-93).
Gryparis et al. (2004) noted that there was a considerable seasonal difference in the O3 effect on
mortality; thus, the small effect for the all-year data might be attributable to inadequate
adjustment for confounding by seasonality.  Focusing on analyses using summer measurements,
the authors report statistically significant associations with total mortality [1.8% increase per 30
ppb 8-hr O3 (95% CI: 0.8, 2.9)], cardiovascular mortality [2.7% increase per 30 ppb 8-hr O3
(95% CI:  1.2, 4.3)] and with respiratory mortality (6.8% increase per 30 ppb 8-hr O3, 95% CI:
4.5, 9.2) (CD, p. 7-93, 7-99).
       Two of the recent multi-city mortality studies (Bell et al., 2004; Gryparis et al., 2004)
have also  reported associations for multiple averaging times (CD, p. 8-38).  Bell and colleagues
(2004) reported associations between mortality and  1-hr daily max, 8-hr daily max and 24-hr avg
O3 concentrations. Effect estimates for associations with 1-hr O3 was slightly larger than that
reported for 8-hr O3 concentrations, and both were slightly larger than the association with 24-hr
avg O3, but the effect estimates did not differ statistically.  The APHEA study (Gryparis et al.,
2004) also reported effect estimates that were slightly larger with 1-hr than with 8-hr O3
concentrations, but not significantly so.
       Numerous single-city analyses have also reported associations between mortality and
short-term O3 exposure, especially for those analyses using warm season data. As shown in
Figure 7-21 of the CD, the results of recent publications show a pattern of positive, often
statistically significant associations between short-term O3 exposure and mortality during the
warm season (CD, p. 7-97).  For example, statistically significant associations were reported in
southern California (Ostro,  1995), Philadelphia (Moolgavkar et al., 1995), Dallas (Gamble et al.,
1998), and Vancouver (Vedal et al., 2003), as well as numerous studies conducted in other
countries. However, no evidence of an association was seen in a study  conducted in Pittsburgh
                                          3-29

-------
(Chock et al., 2000). In considering results from year-round analyses, there remains a pattern of
positive results but the findings are less consistent. For example, statistically significant
associations were reported in Philadelphia (Moolgavkar et al., 1995) and Dallas (Gamble et al.,
1998), while positive but not statistically significant associations were reported in Detroit
(Lippmann et al., 2000, reanalyzed in Ito, 2003), San Jose (Fairley, 1999, reanalyzed Fairley,
2003), and Atlanta (Klemm et al., 2004).  No evidence for associations was reported in Los
Angeles (Kinney et al., 1995), Coachella Valley (Ostro et al., 2003), and St. Louis and Eastern
Tennessee (Dockery et al., 1992). In most single-city analyses, effect estimates were not
substantially changed with adjustment for PM (CD Figure 7-22, p. 7-101).
       In addition, several meta-analyses have been conducted  on the relationship between 63
and mortality.  As described in section 7.4.4 of the CD, these analyses reported fairly consistent
and positive combined effect estimates ranging from 1.5 to 2.5% increase in mortality for a
standardized change in O3 (CD, Figure 7-20, p. 7-95).  Three recent meta-analyses evaluated
potential sources of heterogeneity in Os-mortality associations (Bell et al., 2005; Ito et al., 2005;
Levy et al., 2005). The CD (p. 7-96) observes common findings across all three analyses, in that
all reported that effect estimates were larger in warm season analyses, reanalysis of results  using
default GAM criteria did not change the effect estimates, and there was no strong evidence of
confounding by PM (CD, p.  7-97). Bell et al. (2005) and Ito et  al. (2005) both provided
suggestive evidence of publication bias, but O3-mortality associations remained after accounting
for that potential bias. The CD (7-97)  concludes that the "positive 63 effects estimates, along
with the sensitivity analyses  in these three meta-analyses, provide evidence of a robust
association between ambient Os and mortality."
       Most of the single-pollutant model estimates from single-city studies range from 0.5 to
5% excess deaths per  standardized increments. Corresponding  summary estimates  in large U.S.
multi-city studies ranged between 0.5 to 1% with some studies noting heterogeneity across cities
and studies (CD, p.  7-110).
       In the CD (p. 7-101), Figure 7-22 shows the Os risk estimates with and without
adjustment for PM indices using  all-year data in studies that conducted two-pollutant analyses.
Approximately half of the 63 risk estimates increased slightly, whereas the other half decreased
slightly with the inclusion of PM in the models.  In general, the Os-mortality risk estimates were
robust to adjustment for PM  in the models, with the exception of Los Angeles, CA  data with
PMio (Kinney et al., 1995) and Mexico City data with TSP (Borja-Aburto et al., 1997).  The U.S.
95 communities  study (Bell et al., 2004) examined the sensitivity of acute Os-mortality effects to
potential confounding by PMio (CD, 7-100). Restricting analysis to days when both 63 and PMio
data were available, the community-specific Os-mortality effect estimates as well as the national
average results indicated that Os was robust to adjustment for PMio (Bell et al., 2004).
                                          3-30

-------
       Several Os-mortality studies examined the effect of confounding by PM indices in
different seasons (CD, p. 7-102, Figure 7-23). In analyses using all-year data and warm-season
only data, Os effect estimates were once again fairly robust to adjustment for PM indices, with
values showing both slight increases and decreases with the inclusion of PM in the model. In the
analyses using cool season data only, the O3 effect estimates all increased slightly with the
adjustment of PM indices, although none reached statistical significance.
       The three recent meta-analyses (Bell et al., 2005; Ito et al., 2005; Levy et al., 2005) all
examined the influence of PM on Os risk estimates.  No substantial influence was observed in
any of these studies. In the analysis by Bell et al. (2005), the combined estimate without PM
adjustment was 1.7% (95% PI: 1.10, 2.37) from 41 estimates, and the combined estimate with
PM adjustment was 1.95% (95% PI: 1.06, 4.00) from  11 estimates per 20 ppb increase in 24-hr
avg 63. In the meta-analysis of 15 cities (Ito et al., 2005), the combined estimate was 1.6%
(95% PI: 1.1, 2.2) and 1.5% (95% PI: 0.8, 2.2) per 20  ppb in 24-hr avg O3 without and with PM
adjustment, respectively (CD, p. 7-103).  The additional time-series analysis of six cities by Ito et
al. (2005) found that the influence of PM by season varied across alternative weather models but
was never substantial. Levy et al.  (2005) examined the regression relationships between Os and
PM indices (PMi0 and PM2 5) with Os-mortality effect estimates for all year and by season.
Positive slopes, which might indicate potential confounding, were observed for PM2.5 on 63
effect estimates in the summer and all-year periods, but the relationships were weak. The effect
of one causal variable (i.e.,  63) is expected to be overestimated when a second  causal variable
(e.g., PM) is excluded from the analysis, if the two variables are positively correlated and act in
the same direction. However, the  results from these meta-analyses, as well as several single- and
multiple-city studies, indicate that copollutants generally do not appear to substantially confound
the association between O3  and mortality (CD, p. 7-103).
       Finally, from those studies that included assessment of associations with specific causes
of death, it appears that effect estimates for associations with cardiovascular mortality are larger
than those for total mortality; effect estimates for respiratory mortality are less  consistent in size,
possibly due to reduced statistical  power in this subcategory of mortality (CD, p. 7-108). The
U.S. 95 communities study  (1987-2000) analyzed 63effect estimates from cardiovascular and
respiratory mortality. The analysis by Bell et al. (2005) used all available data,  which included
all-year data from 55 communities and warm-season only data from 40 communities. The
national average estimate from the constrained distributed lag model was slightly greater for
cardiopulmonary deaths than deaths from all causes, with an excess risk of 1.28% (95% PI: 0.62,
1.97) compared to  1.04% (95% PI: 0.54, 1.55) per 20  ppb increase in 24-hr avg O3 in the
preceding week.
                                          3-31

-------
       One recent multi-city study (Bell et al., 2006) examined the shape of the concentration-
response function for the Os-mortality relationship in 98 U.S. urban communities for the period
1987 to 2000 specifically to evaluate whether a "safe" threshold level exists.  Results from
various analytic methods all indicated that any threshold would exist at very low concentrations,
far below the level of the current O3 NAAQS and other, lower international O3 standards,3 and
nearing background levels. Notably, in a subset analysis using only days that were below the
level of the current 63 NAAQS, the (Vmortality association remained statistically significant
with only a small change in the size of the effect estimate.  Further, in a subset analysis based on
24-hr average Os concentrations, the effect estimates declined and lost statistical significance
only when the maximum daily average concentration included was < 10 ppb (Bell et al., 2006, p.
14 and Figure 2), which corresponds to daily maximum 8-hr average concentrations in U.S.
cities that are within the range of background concentrations. The authors conclude that
"interventions to further reduce ozone pollution would benefit public health, even in regions that
meet current regulatory standards and guidelines" (Bell et al., 2006, p. 3).
       A related study (Huang et al., 2005) examined 63 effects on cardiopulmonary mortality
during the summers (June to September) of 1987 to 1994 in 19 large U.S. cities from the
NMMAPS database. Figure 7-24 in the CD (p. 7-104), presents the Bayesian city-specific and
overall average 63 effect estimates for cardiopulmonary mortality per 20 ppb increase in 24-hr
avg Os from a constrained 7-day distributed lag model.  The Os effect estimate was 2.52% (95%
PI: 0.94, 4.10) excess risk in cardiopulmonary mortality per 20 ppb increase in 24-hr avg Os in
the preceding week for the combined analysis of all cities.  For analyses of summer data,
confounding of the Os effect by PM is of concern as daily variations in Os may be positively
correlated to PM during the  summer months. Huang et al. (2005) observed that when PMio was
included in the model, the O3 effect estimate, on average, remained positive and significant.  As
PMio measurements were available only every 1 to 6 days, only  single-day lags were examined.
At a 0-day lag, O3 was associated with a 1.47% (95% PI: 0.54, 2.39) excess risk versus a 1.49%
(95% PI: 0.66, 3.47) excess risk in cardiopulmonary mortality in the Os-only model and after
adjustment for PMio, respectively. The slight sensitivity of the O3 health effects to the inclusion
of PMio in the model may indicate a true confounding effect. However, as only the days with
PMio data available were included in the analysis, the lack of significance is likely attributable to
higher statistical uncertainty due to the reduced data availability (CD, p. 7-105).
       Figure 7-25 in the CD (p., 7-106), presents effect estimates for associations between O3
and cardiovascular mortality for all-year and warm-season analyses. All studies, with the
       3 Other international 8-hr O3 standards considered by Bell et al. (2006, Table 1) include the California
standard of 70 ppb, the Canadian standard of 65 ppb, and the World Health Organization guideline and European
Commission target value of approximately 61 ppb.

                                          3-32

-------
exception of Ponka et al. (1998), showed positive associations between Os and cardiovascular
mortality (CD, p. 7-105). As with all-cause mortality, there appears to be heterogeneity in the
effect estimates across studies.  The cardiovascular mortality estimate from one meta-analysis
(Bell et al., 2005) appears to be close to the mode of the effect estimates from the various
studies, as shown in Figure 7-25, in the CD (p. 7-106).  This is expected, given that many of
these studies were also included in the meta-analysis.  This study observed that the posterior
mean estimate for cardiovascular causes (2.23% excess risk per 20 ppb increase in 24-hr avg 63
from 25 estimates) was slightly larger than that for total mortality (1.75% excess risk from 41
estimates). However, since cardiovascular deaths account for the largest fraction (over 40%) of
total deaths, it is not surprising that the risk estimates for cardiovascular mortality are somewhat
similar to those from all-cause mortality. Overall, the cardiovascular mortality risk estimates in
the current literature show consistently positive associations with some heterogeneity (most
estimates fall within the range of 1 to 8% per 40 ppb increase in 1-hr avg O3 (CD, p. 7-107).
       Several studies observed that the risk estimates for the respiratory category were larger
than the cardiovascular and total nonaccidental categories (Anderson et al., 1996; Gouveia and
Fletcher, 2000; Gryparis et al., 2004; Zmirou et al., 1998). The apparent inconsistencies across
studies may be due in part to the differences in model specifications, but they may also reflect
the lower statistical power associated with the smaller daily counts of the respiratory category
(usually accounting for less than 10% of total deaths) compared to the  larger daily counts for the
cardiovascular category (approximately 40 to 50% of total deaths).  Thus, an examination of the
differences in risk estimates across specific causes requires a large population and/or a long
period of data collection.
       In summary, several  single-city studies observed positive associations of ambient 63
concentrations with total nonaccidental and cardiopulmonary mortality. The CD finds that the
results from U.S. multi-city time-series studies provide the strongest evidence to date for 63
effects on acute mortality.  Recent meta-analyses also indicate positive risk estimates that are
unlikely to be confounded by PM; however, future work is needed to better understand the
influence of model specifications on the risk coefficient (CD, p. 7-175). A meta-analysis that
examined specific causes of mortality found that the cardiovascular mortality risk estimates were
higher than those for total mortality. For cardiovascular mortality, the  CD (Figure 7-25, p. 7-
106) suggests that effect estimates are consistently positive and more likely to be larger and
statistically significant in warm season analyses.  The findings regarding the effect size for
respiratory mortality have been less consistent, possibly because of lower statistical power in this
subcategory of mortality. The CD (p. 8-78) concludes that these  findings are highly suggestive
that short-term 63 exposure directly or indirectly contribute to non-accidental and
cardiopulmonary-related mortality, but additional research is needed to more fully establish
underlying mechanisms by which such effects occur.
                                          3-33

-------
        3.3.2.2    Mortality and Long-term O3 Exposure
       Little evidence was available in the last review on the potential for associations between
mortality and long-term exposure to O3. In the Harvard Six City prospective cohort analysis, the
authors report that mortality was not associated with long-term exposure to O?, (Dockery et al.,
1993). The authors note that the range of 63 concentrations across the six cities was small (19.7
to 28.0 ppb in average 24-hr concentrations over the 7-year study period), which may have
limited the power of the study to detect associations between mortality and 63 levels (CD, p. 7-
127).
       As discussed in section 7.5.8 of the CD, in this review there are results available from
three prospective cohort studies: the American Cancer Society (ACS) study, the Adventist
Health and Smog (AHSMOG) study, and the U.S. Veterans Cohort study. In addition, a major
reanalysis report includes evaluation of data from the Harvard Six City cohort study (Krewski et
al., 2000).  This reanalysis also includes additional evaluation of data from the initial ACS cohort
study report that had only reported results of associations between mortality and long-term
exposure to fine particles and sulfates (Pope et al.,  1995).4
       In this reanalysis of data from the previous Harvard Six City prospective cohort study,
the investigators replicated and validated the findings of the original studies, and the report
included additional quantitative results beyond those available in the original report (Krewski et
al., 2000).  In the reanalysis of data from the Harvard Six Cities study, the effect estimate for the
association between long-term Os concentrations (8.3 ppb between the highest and lowest
concentrations in the cities) and mortality was  negative and nearly statistically significant
(relative risk  = 0.87, 95% CI: 0.76, 1.00).
       The ACS study is based on health data  from a large prospective cohort of approximately
500,000 adults and air quality data from about  150 U.S. cities.  The initial report (Pope et al.,
1995) focused on associations with fine particles and sulfates, for which  significant associations
had been reported in the earlier Harvard Six Cities  study (Dockery et al., 1993).  As part of the
major reanalysis of these data, results for associations with other air pollutants were also
reported, and the authors report that no significant associations were found between Os and all
cause mortality (95%  CI: 0.96-1.07). A significant association was reported for
cardiopulmonary mortality (relative risk=1.08, 95% CI: 1.01, 1.16) (Krewski et al., 2000, p.
174). For some specifications of Os exposure in the ACS study, there was an effect in the warm
quarter, as there was in the reanalysis of the Harvard Six Cities study.
       4 This reanalysis report and the original prospective cohort study findings are discussed in more detail in
section 8.2.3 in Air Quality Criteria for P'articulate Matter (EPA, 2004).

                                          3-34

-------
       The ACS II study (Pope et al., 2002) reported results of associations with an extended
data base; the mortality records for the cohort had been updated to include 16 years of follow-up
(compared with 8 years in the first report) and more recent air quality data were included in the
analyses.  Results are presented for full-year and summer season analyses, and show no evidence
for a significant association between long-term exposure to O3 and mortality. As shown in
Figure 7-27 of the CD, the effect estimates are not statistically significant for associations
between long-term 63 exposure and all-cause, cardiopulmonary, and lung cancer mortality (CD,
p.  7-128) in all year analyses. However, in the summer season, marginally significant
associations were observed for cardiopulmonary mortality.
       The Adventist Health and Smog (AHSMOG)  cohort includes about 6,000 adults living in
California. In two studies from this cohort, a  significant association has been reported between
long-term 63  exposure and increased risk of lung cancer mortality among males only (Beeson et
al., 1998; Abbey et al., 1999). No significant associations were reported between long-term O3
exposure and mortality from all causes or cardiopulmonary causes. Due to the small numbers of
lung cancer deaths (12 for males, 18  for females) and the precision of the effect estimate (i.e., the
wide confidence intervals), the CD raised concerns about the plausibility of the reported
association with lung cancer (CD, p.  7-130).
       The U.S. Veterans Cohort study (Lipfert et al., 2000b, 2003) of approximately 50,000
middle-aged males diagnosed with hypertension, reported some positive associations between
mortality and peak 63 exposures (95th percentile level for several years of data).  The study
included numerous analyses using subsets of exposure and mortality  follow-up periods which
spanned the years 1960 to 1996. In the results of analyses using deaths and Os exposure
estimates concurrently across the study period, there were positive, statistically significant
associations between peak O3 and mortality, with a 9.4% excess risk  (95% CI:  0.4,  18.4) per
mean 95% percentile O3 (CD, p. 7-129).
       Thus,  the results from all-year analyses in the Harvard Six Cities and ACS cohorts
provide no evidence for associations between long-term Os exposure and mortality, though the
warm-season results in the reanalysis of the ACS cohort study suggest a potential association.
Imprecise and inconclusive associations were reported in analyses for the AHSMOG cohort
study.  Significant associations between long-term Os exposure and mortality were  only reported
for the Veterans cohort study; however, this study used an indicator of peak 63 concentrations
and the cohort is also a rather specific subgroup of the U.S. population.  Overall, the CD
concludes that consistent associations have not been reported between long-term Os exposure
and all-cause, cardiopulmonary or lung cancer mortality (CD, p. 7-130).
                                         3-35

-------
       3.3.3   Ozone Effects on UV-B Flux
       The CD (Chapter 10) provides a thorough analysis of the current understanding of the
relationship between reducing tropospheric Os concentrations  and the potential impact these
reductions might have on increasing UV-B surface fluxes and  indirectly contributing to increased
UV-B related health effects.  It is clear that there are many factors that influence UV-B radiation
penetration to the earth's surface, including cloud cover, surface albedo, PM concentration and
composition, and gas phase pollution. A risk assessment of UV-B related health effects would
need to take into account human habits, such as outdoor activities, dress and skin care. However,
little is known about the impact of these factors on  individual exposure to UV-B, and detailed
information does not exist regarding type (e.g., peak or cumulative) and time period (e.g.,
childhood, lifetime, current) of exposure, wavelength dependency of biological responses, and
interindividual variability in UV-B resistance. In fact there have been recent reports indicating
the necessity of UV-B in producing vitamin D, suggesting that increased risks of human disease
due to slight excess UV-B exposure may be offset by the benefits of enhanced vitamin D
production. However, as with other impacts of UV-B on human health, this beneficial effect of
UV-B radiation has not been studied in sufficient detail to allow for a credible health benefits or
risk assessment.  The CD (p.  10-38) concluded that the effects of changes in surface-level Os
concentrations on UV-induced health effects cannot be critically assessed given the significant
uncertainties summarized above.

       3.3.4   Summary
       The CD (Chapters 4-8) summarizes and assesses substantial new evidence which builds
upon what was previously known about the health effects of 63. The new information supports
previous findings that short-term  Os is associated with lung function decrements and respiratory
symptoms, as well as numerous more subtle effects on the respiratory system such as
morphological changes and altered host defense mechanisms.  Short-term 63 exposure has also
been associated with hospital admissions for respiratory causes in numerous new studies that
further confirm the findings evaluated in the 1996 CD.  The CD reports that warm-season studies
show evidence for positive and robust associations  between ambient O3 concentrations and
respiratory hospital admissions, asthma ED visits, and respiratory symptoms and lung function
effects in asthmatic children (CD, p. 7-175).
       Some new studies have suggested associations between increased incidence of asthma or
reduced lung function and long-term exposure to elevated ambient Os levels. The findings of
this small group of studies are inconsistent, however, and the CD concludes that the evidence for
this group of associations is inconclusive (CD, p. 7-175).
                                         3-36

-------
       A new body of studies has suggested associations between short-term Os exposure and
effects on the cardiovascular system, including changes in heart rate variability, cardiac
arrhythmia, incidence of MI and hospitalization for cardiovascular diseases. The CD finds this
body of evidence to be limited but supportive of potential effects of 63 on the cardiovascular
system (CD, p. 8-77).
       A major area where new information presented in the CD has significantly expanded our
knowledge on health effects is evidence of an elevated risk of mortality associated with acute
exposure to Os, especially in the summer or warm season when O3 levels are typically high.
Results from recent large U.S. multicity time-series studies and meta-analyses provide the
strongest evidence for associations between short-term 63 exposure and mortality (CD, p. 7-
175). The risk estimates shown are consistent across studies and robust to control for potential
confounders. This overall body of evidence is highly suggestive that 63 directly or indirectly
contributes to nonaccidental and cardiopulmonary-related mortality, but additional research is
needed to more fully establish underlying mechanisms by which such effects occur (CD, p. 8-
78).

3.4  ASSESSMENT OF EVIDENCE FROM EPIDEMIOLOGICAL STUDIES
       In Chapter 8, the CD assesses the  new health evidence, integrating findings from
experimental (e.g., lexicological, dosimetric and controlled human exposure) and
epidemiological studies, to make judgments about the extent to which causal inferences can be
made about observed associations between health endpoints and exposure to 03. Section 8.4.4.3
of the CD indicates that strength of the 63 effects (including the magnitude and precision of
reported Os effect estimates and their statistical  significance), robustness of epidemiological
associations (i.e., stability in the effect estimates after considering a number of factors), and
consistency of effects associations (looking across results of multiple- and single-city studies
conducted by different investigators in different places and times) are all important in forming
judgments as to the likely  causal significance of observed associations (CD, p. 8-40).
       In evaluating the evidence from epidemiological  studies in sections 7.1.3 and 8.4.4.3, the
CD focuses on well-recognized criteria, including:  (1) the strength of reported associations,
including the magnitude and precision of reported effect estimates and their statistical
significance; (2) the robustness of reported associations, or stability in the effect estimates after
considering factors such as alternative models and model specification, potential confounding by
co-pollutants, and issues related to the consequences of exposure measurement error; and (3) the
consistency of the effects associations as observed by looking across results of multiple- and
single-city studies conducted by different investigators in different places and times (CD, p. 8-
40). Integrating more broadly across epidemiological and experimental evidence, the CD also
                                          3-37

-------
focuses on the coherence and plausibility of observed Os-related health effects to reach
judgments about causality (CD, section 8.6).
       Subsequent to the final CD being published, CASAC sent a letter to the Administrator
(Henderson, 2006b) providing additional advice on some key issues in order to inform
specifically the preparation of this draft Staff Paper specifically and the review of the O3
NAAQS in general. The issues related to assessment of epidemiological studies are addressed in
this section and include the general issue of the utility of time-series epidemiological studies in
assessing the risks from exposure to Os and other criteria pollutants, as well as related issues
about exposure measurement error in Os mortality time-series studies and Os as a surrogate for
the broader mix of photochemical oxidant pollution in time-series studies.  Implications of these
issues for staff conclusions about the adequacy of the current Os NAAQS and the identification
of options for consideration will be considered in Chapter 6.
       The following discussion summarizes the conclusions and judgments from the CD's
summary of epidemiologic evidence and integrative assessment, focusing in particular on
discussions of strength, robustness, and consistency in the epidemiological evidence; judgments
in the CD about coherence and plausibility are summarized below in section 3.5.  This section
also addresses issues related to lag periods between O3 ambient exposure levels and health
outcomes, the nature of O3-health effect concentration-response relationships, and the assessment
of air pollutant mixtures containing Os in time-series epidemiological studies.

       3.4.1   Strength of Associations
       The strength of associations most directly refers to the magnitude of the reported relative
risk estimates. Taking  a broader view, the CD draws upon the criteria summarized in a recent
report from the U.S. Surgeon General, which define strength of an association as "the magnitude
of the association and its statistical strength" which includes assessment of both effect estimate
size and precision, which is related to the statistical power of the study (CDC, 2004). In general,
when associations are strong in terms of yielding large relative risk estimates, it is less likely that
the association could be completely  accounted for by a potential confounder or some other
source of bias (CDC, 2004). With associations that yield small relative risk estimates it is
especially important to consider potential confounding and other factors in assessing causality.
       Effect estimates between 63  and many health outcomes are generally small in size and
could thus be characterized as weak. For example, effect estimates for associations with
mortality generally range from 0.5 to 5% increases per 40 ppb increase in 1-hr max 63 or
equivalent, whereas associations for hospitalization range up to 50% increases per standardized
Os increment. The CD particularly notes that there are several multicity studies for associations
between short-term 63  exposure and mortality or morbidity that, although small in size, have
                                          3-38

-------
great precision due to the statistical power of the studies (CD, p. 8-40).  That is, the associations
were strong enough to have been reliably measured by the studies such that many of the
associations can be distinguished from the null hypothesis with statistical confidence.

       3.4.2  Robustness of Associations
       Factors considered in assessing robustness include impact of exposure error, potential
confounding by copollutants, and alternative models and model specifications, as evaluated in
the CD (sections 7.1.3 and 8.4.4.3) and discussed below.

        3.4.2.1    Exposure Error
       In time-series epidemiologic studies, concentrations measured at centrally-located
ambient monitoring stations are generally used to represent a community's exposure to ambient
03. For time-series studies and panel studies that use these ambient concentrations, the emphasis
is on the temporal (e.g., daily or hourly) changes in ambient 63.  In other cohort or cross-
sectional studies, air quality data averaged over a period of months to years are used as indicators
of a community's long-term exposure to ambient 63 and other pollutants.  In both types of
analyses, exposure error is an important consideration, as actual exposures to individuals in the
population will vary across the community.
       In considering exposure error, it should be noted that total personal exposure can be
partitioned into two types of sources, ambient and nonambient. As described in the CD, there
are few sources of Os exposure for most people other than ambient air; potential indoor sources
of 63 include office equipment, air cleaners, and small electric motors (CD, p. 7-6).  Sheppard
(2005) notes that nonambient source exposures typically vary across individuals, but the
community averages do not vary across communities. In addition, nonambient exposures are not
likely to have strong temporal correlations. In contrast, ambient concentrations across
individuals should be highly correlated, as they tend to vary over time similarly for everyone
because of changes in source generation, weather, and season. The independence of ambient and
non-ambient exposure sources has important implications. Sheppard et al. (2005) observes that
when ambient and nonambient sources are independent, exposure variation due to nonambient
source  exposures behaves like Berkson measurement error and does not bias the effect estimates.
       Ozone concentrations measured at central ambient monitoring sites may explain, at least
partially, the variance in individual exposures to ambient Os; however, this relationship is
influenced by various factors related to building ventilation practices and personal behaviors.
Further, the pattern of exposure misclassification error and the influence of confounders may
differ across the outcomes of interest as well as in susceptible populations. As discussed in the
CD Section 3.9, only a limited number of studies have examined the relationship between
ambient Os concentrations and personal exposures to ambient Os.  One of the strongest
                                         3-39

-------
predictors of the relationship between ambient concentrations and personal exposures appears to
be time spent outdoors. The strongest relationships were observed in outdoor workers (Brauer
and Brook, 1995, 1997; O'Neill et al., 2003).  For example, Brauer and Brook (1995, 1997)
observed that in farmers who worked 6-14 hours outdoors each day, the personal to ambient 63
concentration ratio was 0.96, with a Spearman correlation coefficient of 0.64. Statistically
significant correlations between ambient concentrations and personal exposures were also
observed for children, who likely spend more time outdoors in the warm season (Linn et al.,
1996; Xueetal., 2005).
       There is some concern about the extent to which ambient concentrations are
representative of personal 63 exposures of another particularly susceptible group of individuals,
the debilitated elderly, and what impact that may have on mortality and hospitalization time-
series studies. Those who suffer from chronic cardiovascular or respiratory conditions may tend
to protect themselves more from environmental threats by reducing their exposure to both O3 and
its confounders, such as high temperature and PM, than those who are healthy. The  correlation
between ambient concentrations and personal exposure measurements in older adults (mean age
75 years) has been  examined by Sarnat et al. (2001, 2005). These studies by Sarnat et al. also
included children and COPD patients, and only results for the combined populations are
reported. The first study conducted in Baltimore, MD observed no relationship between ambient
concentrations and personal exposures in both the summer and the winter. However, the second
study conducted in Boston, MA, found statistically significant associations between  ambient 63
concentrations and personal exposures to 63.  The regression coefficient was larger in the
summer (P = 0.27), compared to the winter (P = 0.04), suggesting once again that time spent
outdoors had a large influence on the relationship between ambient concentrations and personal
exposures.  Eight of 29 subjects had personal-ambient O3 correlations greater than 0.8 during the
summer.
       Collectively, these studies observed that the daily averaged personal 63 exposures from
the population were well correlated with ambient Os concentrations despite the substantial
variability that existed among the personal measurements.  Averaging likely removes the noise
associated with  other sources of variation.  These studies provide supportive evidence that
ambient Os concentrations from central  monitors may serve as valid surrogate measures for
mean personal exposures experienced by the population, which is of most relevance for time-
series studies. A better understanding of the relationship between ambient concentrations and
personal exposures, as well as of the other factors that affect relationship will improve the
interpretation of concentration-population health response associations observed.
       The CD  discusses the potential influence of exposure error on epidemiologic study results
in Section 7.1.3.1.  Zeger et al.  (2000) outlined the three components to exposure measurement
                                         340

-------
error: (1) the use of average population rather than individual exposure data; (2) the difference
between average personal ambient exposure and ambient concentrations at central monitoring
sites; and (3) the difference between true and measured ambient concentrations (CD, p. 7-7).
These components are expected to have different effects, with the first and third likely not
causing bias in a particular direction ("nondifferential error") but increasing the standard error,
while the second component may result in downward bias, or attenuation of the risk estimate
(CD, pp. 7-7 to 7-8).  Ambient exposure can be assumed to be the product of the ambient
concentration and an attenuation factor (i.e., building filter).  Panel studies and time-series
studies that use ambient concentrations instead of personal exposure measurements will estimate
a health risk that is attenuated by that factor.  Navidi et al. (1999) used data from a children's
cohort study to compare effect estimates from a simulated "true" exposure level to results of
analyses from 63 exposures determined by several methods.  The results indicated that the use of
O3 exposures from personal sampling or microenvironmental approaches is associated with
nondifferential error in Os effect estimates, compared with effect estimates from "true"
exposures.  However, 63 exposures based on the use of ambient monitoring data overestimates
the individual's Os exposure and thus generally results in Os effect estimates that are biased
downward (CD, p. 7-8). Similarly, Zidek (1997) observed that a statistical analysis must balance
bias  and imprecision (error variance).  For example, in a reanalysis of a study by Burnett et al.
(1994) on the acute respiratory effects of ambient air pollution, Zidek et al. (1998) reported that
accounting for measurement, as well as making a few additional changes to the analysis, resulted
in qualitatively similar conclusions, but the effects estimates were considerably larger in
magnitude (CD, p. 7-8).
       A simulation  study by Sheppard et al. (2005) also considered attenuation of the risk based
on personal behavior, their microenvironment, and the qualities of the pollutant in time-series
studies.  Of particular interest is their finding that significant variation in nonambient exposure or
in ambient source exposure that is independent of ambient concentration does not further bias the
effect estimate.  In other words, risk estimates were not further attenuated in time-series studies
even when the correlations between personal exposures and ambient concentrations were weak.
       In addition to overestimation of exposure and the resulting underestimation of effects, the
use of ambient Os concentrations may obscure the presence of thresholds in epidemiologic
studies (CD p. 7-9). Brauer et al. (2002) concluded that surrogate measures of exposure, such as
those from centrally located ambient monitors, that were not highly correlated with personal
exposures obscured the presence of thresholds in epidemiologic studies at the population level,
even if a common threshold exists for individuals within the  population.
       Existing epidemiologic models may not fully take into consideration all of the
biologically relevant exposure history or reflect the complexities of all of the underlying
                                          341

-------
biological processes. As discussed in the CD, Section 3.9, using ambient concentrations to
determine exposure generally overestimates true personal Os exposures by approximately 2- to
4-fold in available studies, resulting in attenuated risk estimates. The implication is that the
effects being estimated occur at fairly low exposures and the potency of 63 is greater than these
effects estimates indicate.  As very few studies evaluating O3 health effects with personal O3
exposure measurements exist in the literature, effect estimates determined from ambient Os
concentrations must be evaluated and used with caution to assess the health risks of 63. Until
more data on personal Os exposure becomes available, the use of routinely monitored ambient
Os concentrations as a surrogate for personal exposures is not generally expected to change the
principal conclusions from 63 epidemiologic studies.  Therefore, population health risk estimates
derived using ambient Os levels from currently available observational studies, with appropriate
caveats about personal exposure considerations, remain useful.
       The CD recommends caution in the quantitative use  of effect estimates calculated using
ambient Os concentrations as they may lead to underestimation of the potency of Os. However,
staff observes that the use of these risk estimates for comparing relative risk reductions between
alternative ambient Os standards considered in the risk assessment is less likely to suffer from
this concern. In addition, as discussed in Chapter 5, staff has conducted an exposure assessment
in conjunction with a portion of the health risk assessment that incorporates estimated population
exposures in developing risk estimates for health outcomes based on controlled human exposure
studies.

        3.4.2.2    Confounding by Copollutants
       Confounding occurs when a health effect that  is caused by one risk factor is  attributed to
another variable that is correlated with the causal risk factor; epidemiological analyses attempt to
adjust or control for potential confounders.  Copollutants (e.g., PM, CO, SO2 and NO2) can meet
the criteria for potential confounding in O3-health associations if they are potential risk factors
for the health effect under study and are correlated with Os.  Effect modifiers include variables
that may influence the health response to the pollutant exposure (e.g., co-pollutants, individual
susceptibility,  smoking or age). Both are important considerations for evaluating effects in a
mixture of pollutants, but for confounding, the emphasis is on controlling or adjusting for
potential confounders in estimating the effects of one pollutant, while the emphasis for effect
modification is on identifying and assessing the effects for different modifiers.
       The CD observes that Os is generally not highly correlated with other criteria pollutants
(e.g., PMio,  CO, SO2 and NO2), but may be more highly correlated with secondary fine particles,
especially during the summer months (CD, p. 7-148). In addition, the correlation between O3
and other pollutants may vary across seasons, since Os concentrations are generally  higher in the
                                          342

-------
summer months. For example, positive associations are observed between Os and pollutants
such as fine particles during the warmer months, but negative correlations may be observed
between Os and these pollutants during the cooler months (CD, p. 7-17).  Thus, the CD pays
particular attention to the results of season-specific analyses and studies that assess effects of PM
in potential confounding of O3-health relationships in its discussions in section 7.6.4.
       Multipollutant models are commonly used to assess potential confounding in
epidemiological studies.  As discussed in the CD, the limitations to the use of multipollutant
models include the difficulty in interpreting results where the copollutants are highly colinear, or
where correlations between pollutants change by  season (CD, p. 7-150). This is particularly the
situation where 63 and a copollutant, such as sulfates, are formed under the same atmospheric
condition; in such cases multipollutant models would produce unstable and possibly misleading
results (CD, p. 7-152).
       For mortality, the results from numerous multi-city and  single-city studies are shown in
Figure 7-22 of the CD.  These results indicate that Os-mortality associations do not appear to be
substantially changed in multipollutant models including PMio or PM2.5 (CD, p. 7-101).
Focusing on results of warm season analyses, Figure 7-23 of the CD shows effect estimates for
Os-mortality associations that are fairly robust to  adjustment for PM in multipollutant models
(CD, p. 7-102).  In general, based on results from several single- and multiple-city studies, and
on recent meta-analyses, the CD (p. 7-103) concludes that "copollutants generally do not appear
to substantially confound the association between 63 and mortality."
       Similarly, multipollutant models are presented for associations between short-term 63
exposures and respiratory hospitalization in Figure 7-12 of the CD; the CD concludes that
copollutants generally do not confound the relationship between 63  and respiratory
hospitalization (CD, p.  7-79 to 7-80). Multipollutant models were not used as commonly in
studies of relationships between respiratory symptoms or lung function with Os, but the CD
reports that results of available analyses indicate that such associations generally were robust to
adjustment for PM2.5 (CD, p. 7-154). For various co-pollutant models, in a large multicity study
of asthmatic children (Mortimer et al., 2002), the  O?, effect was attenuated, but there was still a
positive association.  In Gent et al. (2003), effects of 63, but not PM2.5, remained statistically
significant and even increased in magnitude in two-pollutant models (CD, p. 7-53).
       Considering this body of studies, the CD concludes: "Multipollultant regression analyses
indicated that O3 risk estimates, in general, were not sensitive to the inclusion of copollutants,
including PM2.5 and sulfate. These results suggest that the effects of Os on respiratory health
outcomes appear to be robust and independent of the effects of other copollutants (CD, p. 7-
154)." We use the results of single-pollutant model results in presentation of results in this
chapter and in quantitative risk assessments conducted as  part of this review (see Chapter 5) for
                                          343

-------
purposes of comparing results from different studies. However, we also include the use of multi-
pollutant model results in presenting risk estimates, when available, to more completely
characterize the quantitative health risks associated with ambient Os levels

        3.4.2.3     Model Specification
       The CD observes that one challenge of time-series epidemiological analysis is assessing
the relationship between O3 and health outcomes while avoiding bias due to confounding by
other time-varying factors, particularly seasonal trends and weather variables (CD, p. 7-14).
These variables are of particular interest because 63 concentrations have a well-characterized
seasonal pattern (see Chapter 2) and are  also highly correlated with changes in temperature.
Thus it can be difficult to distinguish whether effects are associated with Os or with seasonal or
weather variables in statistical analyses.
       Section 7.1.3.4  of the CD discusses statistical modeling approaches that have been used
to adjust for time-varying factors, highlighting a series of analyses that were done in a Health
Effects Institute-funded reanalysis of numerous  time-series studies. While the focus of these
reanalyses was on associations with PM, a number of investigators also examined the sensitivity
of 63 coefficients to the extent of adjustment for temporal trends and weather factors. In
addition, several recent studies, including U.S. multi-city  studies (Bell et al., 2005; Huang et al.,
2005; Schwartz et al., 2005) and a meta-analysis study (Ito et al., 2005), evaluated the effect of
model specification on (Vmortality associations.  As discussed in the CD (section 7.6.3.1), these
studies generally report that associations reported with Os are not substantially changed with
alternative modeling strategies for adjusting for  temporal trends and meteorologic effects. In the
meta-analysis by Ito et al. (2005), a separate multicity analysis was presented that found that
alternative adjustments for weather resulted in up to 2-fold difference in the Os effect estimate.
However, significant confounding can occur when strong seasonal cycles are present, suggesting
that season-specific results are more generally robust than year-round results in such cases.  The
CD concludes that "seasonal dependence of Os-mortality  effects complicates interpretation of Os
risk estimates calculated from year-round data without adequate adjustment of temporal trends"
(CD, p. 7-99), and that more work is needed in this area to reduce the uncertainty involved in the
epidemiologic interpretation of Os effect estimates (CD, p. 7-141).
       A number of epidemiological studies have conducted season-specific analyses, as
discussed in section 7.6.3.2 of the CD. As observed above in section 3.3, such studies have
generally reported stronger and more precise effect estimates for O3 associations in the warm
season than in analyses conducted in the cool  seasons or over the full year. For assessing
relationships between O3 and health outcomes, the CD highlights several reasons to focus on
warm season analyses:  (1) the seasonal nature of O3 concentrations; (2) the relationship between
                                          344

-------
O3 formation and temperature; (3) correlations between other pollutants, particularly fine
particles, and O3 variations across seasons in some areas; and (4) factors affecting exposure to
ambient O3, such as air conditioning use, varies seasonally in most areas of the U.S.. We have
therefore focused on epidemiological findings from warm season analyses, where available, for
qualitative assessments and for the quantitative risk assessment discussed in Chapter 5.

       3.4.3  Consistency
       Consistency refers to the persistent finding of an association between exposure and
outcome in multiple studies of adequate power in different persons, places, circumstances and
times (CDC, 2004). In considering results from multeity studies and single-city studies in
different areas, the CD observes general consistency in effects of short-term Os exposure on
mortality, respiratory hospitalization and other respiratory health outcomes (CD, p. 8-41).  The
variations in effects that are observed may be attributable to differences in relative personal
exposure to 63, as well as varying concentrations and composition of copollutants present in
different regions. Thus, the CD concludes that "consideration of consistency or heterogeneity of
effects is appropriately understood as an evaluation of the similarity or general concordance of
results, rather than an expectation of finding quantitative results with a very narrow range" (CD,
p.8-41).

       3.4.4  Lag Structure in Short-term Exposure Studies
       In the short-term exposure epidemiological studies, many investigators have tested
associations for a range of lag periods between the health outcome and 63 concentration (see
CD, sections 7.1.3.3). The CD observes that the selection of an appropriate lag period can
depend on the health outcome under study.  For example, if cough is resulting from the irritant
action of 63, that would be expected to occur with a short lag time; however, exacerbation  of
asthma through an inflammatory response might occur up to several days after initial exposure
(CD, p. 7-12). For both mortality and respiratory hospital admissions, the CD reports that most
significant associations between O3 and mortality were observed with O3 measured on the same
day or a 1-day lag period in studies using  individual lag periods (CD, p. 7-14). In U.S. multi-city
studies, larger effect estimate sizes were reported for the (Vmortality relationship with the
distributed lag structure (CD, p. 7-88).  Field studies of lung function or respiratory symptoms
reported associations with Os across a range of lag periods from exposure on the  same day to
exposures averaged over several days (CD,  sections 7.2.3 and 7.2.4). Cardiovascular effects
appeared to be associated with Os at shorter lag periods; cardiovascular health outcomes such as
changes in cardiac autonomic control were associated with 63 measured on the same day (CD,
section 7.2.7.1).  In addition, Peters et al. (2001) reported a positive but not statistically
                                          345

-------
significant association between myocardial infarction onset and O?, with very short lag times of
l-to4hr(CD, p. 7-64).
       In focusing on an effect estimate reported for any individual lag period, the CD observes
that it is important to consider the pattern of results across the series of lag periods.  If there is an
apparent pattern of results across the different lags, then selecting the single-day lag with the
largest effect from a series of positive associations is still likely to underestimate the overall
effect size, since single-day lag effect estimates do not fully capture the risk that may be
distributed over adjacent or other days (CD, p. 7-13). However, if the reported effect estimates
vary substantially across lag periods, any result for a single day may well be biased (CD, p. 7-
14).  If the effect of 63 on health outcomes persists over several days, distributed lag model
results may provide more accurate effect estimates for quantitative assessment than an effect
estimate for a single lag period (CD, p. 7-12).  Conversely, if the underlying (Vhealth
relationship  is truly  an acute effect, then a distributed lag model would likely result in a reduced
effect estimate size that may underestimate the effect (CD, p. 7-12).
       On this basis, the CD focuses on effect estimates from models using 0- or 1-day lag
periods, with some consideration of multi-day lag effects (CD, p. 7-14). For quantitative
assessments, we conclude that it is appropriate to use results from  lag period analyses consistent
with those reported in the CD, focusing on single day lag periods of 0-1 days for associations
with mortality or respiratory hospitalization, depending on availability of results (CD, p. 7-14).
When available, distributed lag model results also have been used in the quantitative risk
assessment.  However, for those few studies that show inconsistent patterns, the use of single-
day lag results is not appropriate for inclusion in the quantitative assessment.

       3.4.5  Concentration-Response Relationships and Potential Thresholds
       It has been recognized that it is reasonable to expect that there likely are biological
thresholds for different health effects in individuals or groups of individuals with similar innate
characteristics and health status.  For Os exposure, individual thresholds would presumably vary
substantially from person to person due to individual differences in genetic susceptibility, pre-
existing disease conditions and possibly individual risk factors such as diet or exercise levels
(and could even vary from one time to another for a given person). Thus,  it would be difficult to
detect a distinct threshold at the population level, below which no  individual would experience a
given effect, especially if some members of a population are unusually sensitive even down to
very low concentrations (U.S. EPA, 2004, p. 9-43, 9-44).
       Some studies have tested associations between O3 and health outcomes after removal of
days with higher Os levels from the data set;  such analyses do not necessarily indicate the
presence or absence of a threshold, but provide some information on whether the relationship is
                                          346

-------
found using only lower-concentration data. For example, using data from 95 U.S. cities, Bell et
al. (2004) found that the effect estimate for an association between short-term O3 exposure and
mortality was little changed when days exceeding 60 ppb (24-hr average) were excluded in the
analysis. Bell et al. (2006) found no difference in estimated effect even when all days with 24-hr
O3 concentrations <20 ppb were excluded (CD, p. 8-43). Using data from 8 U.S. cities,
Mortimer and colleagues (2002) also reported that associations between O3 and both lung
function and respiratory symptoms remained statistically significant and of the same or greater
magnitude in effect size when concentrations greater than 80 ppb (8-hr avg) were excluded (CD,
p. 7-46). Several single-city studies  are also summarized in section 7.6.5 of the CD that report
similar findings of associations that remain or are increased in magnitude and statistical
significance when data at the upper end of the concentration range are removed.
       Other time-series epidemiological studies have used statistical modeling  approaches to
evaluate whether thresholds exist in associations between short-term O3 exposure and mortality.
As discussed in section 7.6.5 of the CD, one European multi-city study included evaluation of
the shape of the concentration-response curve, and observed no deviation from a linear function
across the range of O3 measurements from the study (Gryparis  et al., 2004; CD p. 7-154).
Several single-city studies also observed a monotonic increase  in associations between O3 and
morbidity that suggest that no population threshold exists (CD, p. 7-159).
       On the other hand, a study in Korea used several different modeling approaches and
reported that a threshold model provided the best fit for the data. The results suggested a
potential threshold level of about 45  ppb (1-hr maximum concentration; < 35 ppb, 8-hr avg) for
an association between mortality and short-term O3 exposure during the summer months (Kim et
al., 2004; CD, p. 8-43).  The authors reported larger effect estimates for the association for data
above the potential threshold level, suggesting that an O3-mortality association might be
underestimated in the non-threshold  model.  A threshold analysis recently reported by Bell et al.
(2006) for 98 U.S.  communities, including the same 95 communities in Bell et al. (2004),
indicated that if a population threshold existed for mortality, it would likely fall below a 24-h
average O3 concentration of 15 ppb (< 25 ppb, 8-hr avg).  In addition, Burnett and colleagues
(1997) plotted the relationships between  air pollutant concentrations and both respiratory and
cardiovascular hospitalization, and it appears in these results that the associations with O3 are
found in the concentration range above about 30 ppb (1-hr maximum; < 25 ppb,  8-hr avg).
       Vedal and colleagues (2003)  reported a significant association between O3 and mortality
in British Columbia where O3 concentrations were quite low (mean concentration of 27.3 ppb).
The authors did not specifically test for threshold levels, but the fact that the association was
found in an area with such low O3 concentrations suggests that any potential threshold level
would be quite low in this data set.
                                          347

-------
       In summary, the CD finds that, taken together, the available evidence from clinical and
epidemiological studies suggests that no clear conclusion can now be reached with regard to
possible threshold levels for Os-related effects (CD, p. 8-44). We also recognize that the
available epidemiological evidence neither supports nor refutes the existence of thresholds at the
population level for effects such as increased hospital admissions and premature mortality. There
are limitations in epidemiological studies that make discerning thresholds in populations
difficult, including low data density in the lower concentration ranges, the possible influence of
exposure measurement error, and interindividual differences in susceptibility to Ch-related
effects in populations. We recognize, however, the possibility that thresholds for individuals
may exist in reported associations at fairly low levels within the range of air quality observed in
the studies but not be detectable as population thresholds in epidemiological analyses. Based on
the CD's conclusions, we judge that there is insufficient evidence to support use of potential
threshold levels in quantitative risk assessments and that it is appropriate to estimate risks within
the range of air quality concentrations down to estimated policy-relevant background level.

       3.4.6  Health Effects of Pollutant Mixtures Containing O3
       The potential for (Vrelated enhancements of PM formation,  particle uptake, and
exacerbation of PM-induced cardiovascular effects underscores the importance of considering
contributions of Os interactions with other often co-occurring air pollutants to health effects due
to Os-containing pollutant mixes. Chapters 4, 5, and 6 of the CD provide a discussion of
experimental studies that evaluate interactions of Os with other co-occurring pollutants. Some
examples of important pollutant mixture effects noted there are highlighted below.
       In Chapter 4, the CD noted some important interactive effects of coexposures to 63, and
NC>2 and SC>2, two other common gaseous copollutants found in ambient air mixes.  A study by
Rigas et al. (1997) showed that continuous exposure of healthy human adults to 862 or to NC>2
increased inhaled bolus O3 absorption,  while continuous exposure to O3  alone decreased bolus
absorption of Os. This suggests enhancement of Os uptake by NC>2 or SC>2 coexposure in ambient
air mixes. Another study by Jenkins et al. (1999) showed that asthmatics exhibited enhanced
airway responsiveness to house dust mite following exposures to O3, NO2, and the combination
of the two gases (CD,  Chapter 6).  Spirometric responses, however, were impaired only by Os
and O3+NO2 at higher concentrations.  On the other hand, animal toxicology studies (CD,
Chapter 5) that evaluated exposures to Os in mixture with NC>2, formaldehyde, and PM
demonstrated additive, synergistic or antagonistic effects, depending on the exposure regimen
and the specific health endpoints evaluated.
       Several studies have demonstrated the enhancement by Os exposure of various respiratory
responses of sensitive individuals to allergens. For example, Peden et al. (1995) showed (V
                                          348

-------
induced increased response to nasal allergen challenge among allergic asthmatic subjects, and
Michelson et al. (1999) showed promotion by 0.4 ppm Os exposure of inflammatory cell influx
in response to nasal allergen challenge in asymptomatic dust-mite sensitive asthmatics.  In
addition, Torres et al. (1996) demonstrated enhancement by 0.25 ppm 63 exposure of airway
responsiveness in mildly allergic asthmatics that was increased in response to an individual's
historical allergen (grass and birch pollen, house dust mite, animal dander). These results were
further extended by Holz et al. (2002) who showed that repeated daily exposure to 0.125 ppm 63
for 4 days exacerbated lung function decrements (e.g., decreased FEVi) in response to bronchial
allergen challenges among subjects with preexisting allergic airway disease, with or without
asthma (see Chapter 6 of the CD).  This suggests that 63 exposure can place allergic people who
do not have asthma, as well as people who do have asthma, at increased risk for allergic
respiratory  effects. Consistent with and supporting the above findings are animal toxicology
studies reviewed in detail by Harkema and Wagner (2005), which indicate that (a) O3-induced
epithelial and inflammatory responses in laboratory rodents are markedly enhanced by
coexposure to inhaled biogenic substances (e.g., bacterial  endotoxin or ovalbumin, an
experimental aeroallergen) and (b) adverse airway effects of biogenic substances can be
exacerbated by coexposure to  03.
       Also of much note is a newly emerging literature which indicates that  Os can modify the
biological potency of certain types of ambient PM, as shown by experimental tests. For
example, as described in the CD, Section 5.4.2, the reaction of diesel PM with 0.1 ppm 63 for 48
hr increased the potency (compared to non-exposed or air-exposed diesel PM) to induce
neutrophil influx, total protein, and LDH in lung lavage fluid in response to intratracheal PM
instillation in rats (Madden et  al., 2000). However, the potency of carbon black particles was not
enhanced by exposure to O3, suggesting that O3 reaction with organic components of the diesel
PM were responsible for the observed increased diesel PM effects.
       Potential interaction of Os with fine PM in aged rats was examined by Kleinman et al.
(2000).  In this study the effects of fine PM containing two common toxic constituents,
                                        o                                       o
ammonium bisulfate (ABS, 0.3 jim 70 |ig/m ) and elemental carbon (C, 0.3 jim 50 |ig/m ) and  a
mixture (ABS + C) with 0.2 ppm O3 was evaluated on aged rat lung structure  and macrophage
function. Exposures of Os, elemental carbon or ABS alone did not cause significant lung injury,
lung tissue  collagen content or respiratory burst activity. On the other hand, mixtures (ABS + C
+ Os) caused significant lung injury as assessed by increased cell proliferation response in lung
epithelial and interstitial cells, loss of lung tissue collagen and increase in respiratory burst and
phagocytic activity.
       The majority of toxicological studies discussed in the CD evaluated effects of individual
pollutants or simple mixtures of the constituents of urban  smog mixtures, and these toxicology
                                         349

-------
studies may not fully explain epidemiologic findings that have increasingly shown ambient Os,
other gaseous pollutants, and/or PM to be associated with various health effects at relatively low
concentrations. In a recent report, Sexton et al. (2004) utilized "smog chambers", i.e.,
environmental irradiation chambers to generate synthetic photochemical oxidants mixtures
similar to urban smog, and studied the toxicity of such mixtures on the inflammatory response of
A549 cells in an in vitro exposure system.  In this preliminary study, the authors found the
simulated urban photochemical oxidant mixture generated with the addition of 63 to have
enhanced toxicity (as assessed by the expression of IL-8 mRNA). Additional toxicology studies
using similar realistic air pollution smog mixtures in the future may provide more relevant
biological understanding for the potential interactions that occur in the ambient air among
various pollutants.
       The body of epidemiological studies discussed in this Staff Paper emphasizes the role of
O3 acting autonomously, from a statistical sense, in association with a variety of adverse
respiratory and cardiovascular effects.  Despite a variety of plausible mechanisms, there exists a
general consensus suggesting that 63, either directly  or through initiation, interferes with basic
cellular oxidation processes responsible for inflammation, reduced antioxidant capacity,
atherosclerosis and other effects. Reasoning that Os influences cellular chemistry through basic
oxidative properties (as opposed to a unique chemical interaction), other reactive oxidizing
species (ROS) in the atmosphere acting either independently or in combination with Os may also
contribute to a number of adverse respiratory and cardiovascular health effects.  Consequently,
the role of 63 should be considered more broadly as 63 behaves as a generator of numerous
oxidizing species  in the atmosphere.
       All of the above types of interactive effects of 63 with other co-occurring gaseous and
nongaseous viable and nonviable PM components of ambient air mixes argue for not only being
concerned about direct effects of Os acting alone, but also the need for viewing Os as a surrogate
indicator for air pollution mixes which may enhance risk of adverse effects due to 63 acting in
combination with other pollutants.  Viewed from this perspective, those epidemiologic findings
of morbidity and mortality associations, with ambient Os concentrations extending to
concentrations below 0.08 ppm, become more understandable and plausible.

3.5   BIOLOGICAL PLAUSIBILITY AND COHERENCE OF EVIDENCE
       This section summarizes material contained in section 8.4.3 and section 8.6 of the CD,
which integrates epidemiological studies with mechanistic information from controlled human
exposure studies and animal toxicological studies to draw conclusions regarding the coherence of
evidence and biological plausibility of (Vrelated health effects.  For its assessment, the CD's
discussion draws from epidemiological evidence on a range of relevant health endpoints (from
                                          3-50

-------
cardiopulmonary and physiological changes to morbidity and mortality) and assessment of
available toxicological and biochemical evidence on potential plausible causal relationships for
the observed epidemiological associations (CD, p. 8-45).

       3.5.1   Animal-to-Human Extrapolation Issues
       Table 3-1 (Table 8-1, CD, p. 8-29) summarizes physiological and biochemical
observations which represent the knowledge base available from studies in humans and animals
that support conclusions drawn about biological alterations that cause acute Os-induced health
effects.  Table 3-1 was based upon experimental data (contained in CD Chapters 5 and 6, as well
as the chapter annexes), which used environmentally  relevant exposure regimens.  Although
most of the acute Os-induced biological alterations are transient and attenuate over time, this
does not mean that injury at the cellular and tissue level does not continue. Most inflammatory
markers (e.g., PMN influx) attenuate after 5 days of exposure, but markers of cell damage (e.g.,
LDH enzyme activity) do not attenuate and continue to increase. The time-line for resolution of
many of the physiological and biological parameters presented in Figure 3-2 (Figure 8-3, CD, p.
8-30) differ for healthy human subjects and those with underlying cardiopulmonary diseases.
The CD also notes that alterations in acute (Vinduced cellular and molecular changes observed
in human airway epithelium evolve overtime, depicted in Figure 3-3 (Figure 8-4, CD, p. 8-31),
and that the knowledge of this profile is important in  assessing biological plausibility to integrate
across evidence of various health endpoints.
       The similarities in physiological, biochemical and pathological processes between
humans and many animal species are due to the high  level of genome sequence homology that
exists across species (CD, p. 8-28).  It is this homology that supports the use of knowledge
gained on initiation, progression, and treatment regimes for disease processes across species,
especially on the acute (Vinduced effects in the respiratory tracts of humans and various animal
species, as depicted in CD Table 3-1 and Figures 3-2  and 3-3.  The similarities observed in
human and rat respiratory system effects (e.g., in measures of lung function,  ventilatory
response, host defense), attenuation, and at higher levels of cellular organization (e.g.,
neutrophilic inflammation, macrophage phagocytosis processes) lend support to animal-to-
human extrapolation. This is particularly important in collecting information that would not be
possible to gather in human exposure or epidemiological studies but may corroborate data from
both types of studies.
                                         3-51

-------
                            Table 3-1. Acute O3-induced Physiological and Biochemical
                                         Changes in Human and Animals
Physiological/Biochemical
Alterations
Human Exposure Studies1>2
Animal Toxicology Studies 3>4
Pulmonary Function:
Airway Responsiveness:


Inflammation:


Reactive Oxygen Species:

Host Defense:
Lung Injury:
Morphology:

Susceptibility:
Cardiovascular Changes:
iFEV;
T Frequency of breathing
(rapid, shallow)
JFVC
(cough, breathing discomfort,
throat irritation, wheezing)
Mild bronchoconstriction

T (neuronal involvement)
Change in lung resistance

Yes
T inflammatory mediators

T

T particle clearance
T permeability
J AM phagocytosis
Yes

Age,
Interindividual variability
Disease status
Polymorphism in certain genes
being recognized

Impairment in arterial O2 transfer
Ventilation-perfusion mismatch
(suggesting potential arterial
vasoconstriction)
T rate pressure product5
T myocardial work5
T Frequency of breathing
   (rapid, shallow)
iFVC
T (vagal mediation)
Change in lung resistance

Yes
T inflammatory mediators

T

T particle clearance
T permeability
i clearance of bacteria
T severity of infection
T mortality & morbidity


Yes

Species-specific differences
Genetic basis for susceptibility indicated
Heart rate
i core body temperature
T atrial natriuretic factor
Role for platelet activity factor (PAF)
indicated
Increased pulmonary vascular resistance
1 Controlled chamber exposure studies in human volunteers were carried out for a duration of 1 to 6.6 h with O3
 concentration in the range of 0.08-0.40 ppm with intermittent exercise.
2 Data on some of the biochemical parameters were obtained from in vitro studies using cells recovered
 fromBALF.
3 Responses were observed in animal toxicology studies with exposure for a duration of 2 to 72 h with O3
 concentration in the range of 0.1 to 2.0 ppm.
4 Various species (mice, rat, guinea pigs and rabbit) and strains.
5 In hypertensive subjects.
                                                    3-52

-------
     Resolution Time-Line for Acute Ozone-Induced Physiological and Biochemical Responses in Humans
c
Pulmonary Function:

ma 1 airways unctions ( V2vra)





Neutrophil influx
Mediators


irway esponsiveness.

Hoss Defense;
Injyry/permeabiliSy.


) 2 4 8 12 1 S 20 24 36 48 72 Hours
1 I 1 Sf \ Sf \ /S \ \ ft \ tt \ tt \

















* Hyperresponsive Subjects
** Asthmatics
4 	 > Partial
•« 	 * Complete
Figure 3-2.  Resolution time-line for the respiratory, physiological, and biochemical parameters
            are derived from studies reported in the CD, Chapter 6 and Chapter 6 Annex.
                                      5-53

-------
        Postulated Cellular and Molecular Changes in Human Airway
                Cells In Response to Acute  Exposure to Ozone
  Response Time

  Immediate 0-2 h
Chemical reaction with ELF and epithelial cell membrane,
Generation of ozonation products, lipid peroxides
  Early 2-24 h
  (Neutrophil infiltration)
 Lipid ozonation products
 Pro-inflammatory mediators (neutrophil chemotaxins,
 Anti-inflammatory mediators (prostanoids)
 Cytokines, proteases
  Late (12-24 h)
  (Eosinophil/monocyte
  infiltration)
Lipid ozonation products
Increase in pro-inflammatory mediators (monocyte chemotaxins,
Decrease in anti-inflammatory mediators (prostanoids)
Release of cytokines
Increased expression of intracellular adhesion molecules
Increased synthesis of collagen, fibronectin
Release of leukocyte proteinase inhibitors
Increased synthesis of antioxidants (SOD, GSH, catalase)
Figure 3-3. Acute (1-8 h) Os exposure-induced cellular and molecular changes and
timelines for their resolution depicted here are derived from the data reported in Leikauf
et al. (1995) and Mudway and Kelly (2000)3-4.Acute (1-8 h) Os exposure-induced cellular
and molecular changes and timelines for their resolution depicted here are derived from
the data reported in Leikauf et al. (1995) and Mudway and Kelly (2000).
                                          5-54

-------
       Quantitative extrapolation requires a combination of dosimetry, end point homology, and
species sensitivity. Although uncertainties continue to exist, animal-to-human extrapolation can
be done for a number of health endpoints with sufficient accuracy to be useful in evaluating the
potential for human health effects. For example, the amount of protein in lavage fluid shows a
striking relationship when interspecies dosimetric adjustments are applied to the individual
species and exposure studies. One study (Hatch et al., 1994) of inflammatory markers suggests
that a 2 ppm 63 exposure in sedentary rats approximates a 0.4 ppm exposure in exercising
humans (i.e., if one considers the dosimetry, the sensitivities of rats and humans are consistent).
This supports the use of some animal data collected at higher Os exposures to help understand
molecular changes in acutely exposed humans (CD, 8-31).  Also of importance are the chronic
exposure studies (12 to 24 months) reporting lesions in animals caused by long-term Os
exposures that may analogously occur in humans with long-term (months, years) exposure to
relatively high levels of O3. However, specific exposure patterns of O3 concentrations that could
produce comparable alterations in human lungs remain to be substantiated (CD, p.  8-32).

       3.5.2   Coherence and Plausibility of Short-term Effects on the Respiratory  System
       Acute respiratory morbidity effects that have been associated with short-term exposure to
O3 include such health endpoints as decrements in lung function, increased airway
responsiveness, airway inflammation, epithelial injury, immune system effects, ED visits for
respiratory diseases, and hospitalization due to respiratory illness
       Recent epidemiological studies have supported evidence available in the previous Os
NAAQS review on associations between ambient Os exposure and decline in lung function for
children. Earlier observations that children and asthmatic individuals are particularly susceptible
to ambient Os are supported by a meta-analysis (Kinney et al., 1996) of summer camp studies.
The CD (p. 8-34) concludes that exposure to ambient 63 has a significant effect on lung function
and is associated with increased respiratory symptoms and medication use, particularly in
asthmatics.
       Short-term exposure to 63 has also been associated with more severe morbidity
endpoints, such as ED visits and hospital admissions for respiratory cases, including specific
respiratory illness (e.g., asthma) (CD, sections 7.3.2 and 7.3.3). In addition, a few
epidemiological studies have reported positive associations between short-term 63 exposure and
respiratory mortality, though the associations are not generally statistically significant (CD, p. 7-
108).
       Considering the evidence from epidemiological studies, the results described above
provide evidence for coherence in Os-related effects on the respiratory system.  Effect estimates
from U.S. and Canadian studies are shown in Figure 3-4, where it can be seen that
                                          3-55

-------

.uu
4.00 -
3.00
2.00
1 00

.uu
Respiratory Emergency Respiratory
Department Visits Respiratory Hospital Admissions Mortality
oU
60-
Respiratory Symptoms 40
(Odds Ratios)
t & 20-
(40) c

-------
mostly positive associations have been reported with respiratory effects ranging from respiratory
symptoms, such as cough or wheeze, to hospitalization for various respiratory diseases, and there
is suggestive evidence for associations with respiratory mortality.  Many of the reported
associations are statistically significant.
       Considering also evidence from lexicological, chamber, and field studies, the CD (section
8.6) discusses biological plausibility and coherence of evidence for acute Os-induced respiratory
health effects.  Inhalation of O3 for several hours while subjects are physically active can elicit
both acute adverse pathophysiological changes and subjective respiratory tract symptoms (CD,
section 8.4.2).  Acute pulmonary responses observed in healthy humans exposed to O3  at ambient
concentrations include: decreased inspiratory capacity; mild bronchoconstriction; rapid,  shallow
breathing during exercise; subjective symptoms of tracheobronchial airway irritation, including
cough and pain on deep inspiration; decreases in measures of lung function (e.g., FVC  and
FEVi); and increased airway resistance (SRaw). The severity of symptoms  and magnitude of
response depends on inhaled dose, individual O3 sensitivity, and the degree of attenuation or
enhancement of response resulting from previous O3 exposures. Lung function studies of several
animal species acutely exposed to relatively low O3 levels (0.25 to 0.4 ppm) show responses
similar to those observed in humans, including increased breathing frequency, decreased tidal
volume, increased resistance, and  decreased FVC. Alterations in breathing pattern return to
normal within hours of exposure, and attenuation in functional responses following repeated O3
exposures is similar to those observed in humans.
       Physiological and biochemical alterations investigated in controlled human exposure and
animal toxicology studies tend to support certain hypotheses of underlying pathological
mechanisms which lead to the development of respiratory-related  effects reported  in
epidemiology studies (e.g.,  increased hospitalization and medication use).  Some of these are:
(a) decrements in lung function, (b) bronchoconstriction, (c) increased airway responsiveness, (d)
airway inflammation, (e) epithelial injury, (f) immune system activation, (g) host defense
impairment, and sensitivity  of individuals, such as age, genetic susceptibility, and  the degree of
attenuation present due to prior exposures. The time sequence, magnitude, and overlap of these
complex events, both in terms of development and recovery (as depicted in Figures 3-2 and 3-3),
illustrate the inherent difficulty of interpreting the biological plausibility of Os-induced
cardiopulmonary health effects (CD, p.  8-48).
       The interaction  of Oj with  airway epithelial cell membranes and epithelial  lining fluid
(ELF) to form lipid ozonation products  and ROS is supported by numerous human, animal and in
vitro studies. Ozonation products  and ROS initiate a cascade of events that lead to oxidative
stress, injury, inflammation, airway epithelial damage and increased epithelial damage and
increased alveolar permeability to vascular fluids.  Repeated respiratory inflammation can lead to
                                          3-57

-------
a chronic inflammatory state with altered lung structure and lung function and may lead to
chronic respiratory diseases such as fibrosis and emphysema (CD, section 8.6.2).  Continued
respiratory inflammation also can alter the ability to respond to infectious agents, allergens and
toxins. Acute inflammatory responses to 63 are well documented, and lung injury can become
apparent within 3 hr after exposure in humans. Ozone-induced lung injury and subsequent
disruption of the airway epithelial barrier has been implicated in increased mucociliary clearance
of particles in human subjects.
       Taken together, the CD concludes that the evidence from experimental human and animal
toxicology studies indicates that acute Os exposure is causally associated with respiratory system
effects, including (Vinduced pulmonary function decrements, respiratory symptoms, lung
inflammation, and increased lung permeability, airway hyperresponsiveness, increased uptake of
nonviable and viable particles, and consequent increased susceptibility to PM-related toxic
effects and respiratory infections (CD, p. 8-48).

       3.5.3  Coherence and Plausibility of Effects on the Cardiovascular System
       There is very limited experimental evidence  of animals and humans that has evaluated
possible mechanisms or physiological pathways by which acute 63 exposures may induce
cardiovascular system effects. Ozone induces lung injury, inflammation, and impaired
mucociliary clearance, with a host of associated biochemical changes all leading to increased
lung epithelial permeability.  As discussed in section 3.2.1.3, the generation of lipid ozonation
products and reactive oxygen species in lung tissues can influence pulmonary hemodynamics,
and ultimately the cardiovascular system.
       Other potential mechanisms by which Os exposure may be associated with cardiovascular
disease outcomes have been described.  Laboratory animals exposed to relatively high Os
concentrations (^ 0.5 ppm) demonstrate tissue edema in the heart and lungs.  Ozone-induced
changes in heart rate, edema of heart tissue, and increased tissue and serum levels of ANF found
with 8-h 0.5 ppm O?, exposure in animal toxicology studies (Vesely et al., 1994a,b,c) also raise
the possibility of potential cardiovascular effects of acute ambient Os exposures.
       Animal toxicology studies have found both transient and persistent ventilatory responses
with and without progressive decreases in heart rate (Arito et al., 1997).  Observations of Os-
induced vasoconstriction in a controlled human exposure study by Brook et al. (2002) suggests
another possible mechanism  for Os-related exacerbations of preexisting cardiovascular disease.
One controlled human study  (Gong et al., 1998) evaluated potential cardiovascular health effects
of Os exposure. The overall  results did not indicate  acute cardiovascular effects of Os in either
the hypertensive or control subjects. The authors observed an increase in rate-pressure product
and heart rate, a decrement for FEVi, and a >10 mm Hg increase in the alveolar/arterial pressure
                                         3-58

-------
difference for O2 following Os exposure. The mechanism for the decrease in arterial oxygen (02)
tension study could be due to an (Vinduced ventilation-perfusion mismatch. Foster et al. (1993)
demonstrated that even in relatively young healthy adults, Os exposure can cause ventilation to
shift away from the well-perfused basal lung. This effect of 63 on ventilation distribution may
persist beyond 24-hr post-exposure (Foster et al., 1997). These findings suggest that O3 may
exert cardiovascular effects indirectly by impairing alveolar-arterial O2 transfer and potentially
reducing O2 supply to the myocardium. Ozone exposure may increase myocardial work and
impair pulmonary gas exchange to a degree that could perhaps be clinically important in persons
with significant preexisting cardiovascular impairment.
       As noted in section 3.3.1.3, a limited number of new epidemiological studies have
reported associations between short-term Os exposure and effects on the cardiovascular system.
Among these studies, three were population-based and involved relatively large cohorts. Two
studies, the ARIC (Liao at al., 2004) and the NAS (Parks et al., 2005) evaluated associations
between O3 and HRV.  The other study, MONICA (Ruidavets et al., 2005) evaluated the
association between Os levels and the relative risk of MI.  Such studies may offer more
informative results based on their large subject-pool and design. Results from these three studies
were suggestive of an association between Os exposure and the cardiovascular endpoints studies.
In other recent studies on incidence of myocardial infarction  and some more subtle
cardiovascular health endpoints, such as changes in heart rate variability or cardiac arrhythmia,
some but not all studies reported associations with short-term exposure to Os (CD, section
7.2.7.1).  From these studies, the CD concludes that the "current evidence is rather limited but
suggestive of a potential effect on HRV, ventricular arrhythmias, and MI incidence" (CD, p. 7-
65).
       An increasing number of studies have evaluated the association between O3 exposure and
cardiovascular hospital  admissions. As shown in Figure 7-13 and discussed in section 7.3.4 of
the CD, many reported  negative or inconsistent associations,  whereas other studies, especially
those that examined the relationship when Os exposures were higher, have found positive and
robust associations between Os and cardiovascular hospital admissions. The CD finds that the
overall evidence from these studies remains inconclusive regarding the effect of Os on
cardiovascular hospitalizations (CD, p. 7-83).
       The CD notes that the suggestive positive epidemiologic findings of Os exposure on
cardiac autonomic control, including effects on HRV, ventricular arrhythmias and MI, and
reported associations between Os exposure and cardiovascular hospitalizations in the warm
season gain credibility and scientific support from the results of experimental animal toxicology
and human clinical studies, which are indicative of plausible  pathways by which Os may exert
cardiovascular effects (CD, Section 8.6.1).
                                         3-59

-------
       3.5.4   Coherence and Plausibility of Effects Related to Long-Term O3 Exposure
       Human chamber studies can not evaluate effects of long-term exposures to 63; there is
some evidence available from toxicological studies.  While early animal toxicology studies of
long-term Os exposures were conducted using continuous exposures, more recent studies have
focused on exposures which mimic diurnal and seasonal patterns and more realistic 63 exposure
levels (CD, p.  8-50).  Studies of monkeys that compared these two exposure scenarios found
increased airway pathology only with the latter design.  Persistent and irreversible effects
reported in chronic animal toxicology studies suggest that additional complementary human data
are needed from epidemiologic studies  (CD, p. 8-50).
       A long-term study of infant rhesus monkeys exposed to simulated seasonal 63 (0.5 ppm ,
8 hr/day for 5 days every 14 days for 11 episodes) reported remodeling of the distal airways,
abnormalities in tracheal basement membrane, accumulation of eosinophils in conducting
airways, and decrements in airway innervation.  Another long-term exposure study of monkeys
exposed to 0.61 ppm Os for a year and studies of rats exposed for 20 months (0.5-1.0 ppm Os for
6 hr/day) reported increased deposition of collagen and thickening of the CAR, suggestive of
irreversible long-term O3 impacts on the lungs.  Although some earlier seasonal exposure studies
of rats reported small, but significant, decrements in lung function consistent with focal
fibrogenesis in the proximal alveolar region, other chronic exposure studies with exposures of
0.5 to 1.0 ppm Os report epithelial hyperplasia that disappears in a few days.
       At this time, there is limited evidence from human studies for long-term Os-induced
effects on lung function. As discussed in section 8.6.2 of the CD, previous epidemiological
studies have provided only inconclusive evidence for either mortality or morbidity effects of
long-term Os exposure. The CD observes that the inconsistency in findings may be due to a lack
of precise exposure information, the possibility of selection bias, and the difficulty of controlling
for confounders (CD, p. 8-50). Several new longitudinal epidemiology studies have evaluated
associations between long-term 63 exposures and morbidity and mortality and suggest that these
long-term exposures may be related to changes in lung function in children; however, little
evidence is available to support a relationship between chronic Os exposure and mortality or lung
cancer incidence (CD, p. 8-50).
       The CD (p. 8-51) concludes that evidence from animal toxicology studies strongly
suggests that chronic Os exposure is capable of damaging the distal airways and proximal alveoli,
resulting in lung tissue remodeling leading to apparent irreversible changes.  Such structural
changes and compromised pulmonary function caused by persistent inflammation may
exacerbate the progression and development of chronic lung disease. Together with the limited
evidence available from epidemiological studies, these findings offer some insight into potential
biological mechanisms for suggested associations between long-term or seasonal  exposures to Os
                                         3-60

-------
and reduced lung function development in children which have been observed in epidemiologic
studies (CD, p. 8-51).

       3.5.5  Coherence and Plausibility of Short-Term Mortality-Related Health
             Endpoints
       An extensive epidemiological literature on air pollution related mortality risk estimates
from the U.S., Canada, and Europe is discussed in the CD (sections 7.4 and 8.6.3).  These single-
and multi-city mortality studies coupled with meta-analyses generally indicate associations
between acute Os exposure and elevated risk for all-cause mortality, even after adjustment for the
influence of season and PM.  Several single-city studies that specifically evaluated the
relationship between 63 exposure and cardiopulmonary mortality also reported results suggestive
of a positive association (CD, p. 8-51). These mortality studies suggest a pattern of effects for
causality that have biologically plausible explanations, but our knowledge regarding potential
underlying mechanisms is very limited at this time and requires further research. Most of the
physiological and biochemical parameters investigated in human and animal studies suggest that
(Vinduced biochemical effects are relatively transient and attenuate over time.  The CD (p. 8-
52) hypothesizes a generic pathway of Os-induced lung damage, potentially involving oxidative
lung damage with subsequent inflammation and/or decline in lung function leading to respiratory
distress in some sensitive population groups (e.g., asthmatics), or other plausible pathways noted
below that may lead to Os-related contributions to cardiovascular effects that ultimately increase
risk of mortality.
       The third National Health and Nutrition Examination Follow-up data analysis indicates
that about 20% of the adult population has reduced FEVi values, suggesting impaired lung
function in some portion of the population. Most of these individuals have COPD,  asthma or
fibrotic lung disease (Manino et al., 2003), which are associated with persistent low-grade
inflammation. Furthermore, patients with COPD are at increased risk for cardiovascular disease.
Also, lung disease with underlying inflammation may be linked to low-grade systemic
inflammation associated with atherosclerosis, independent of cigarette smoking (CD, p. 8-52).
Lung function decrements in persons with cardiopulmonary disease have been associated with
inflammatory markers, such as C-reactive protein (CRP) in the blood. At a population level it
has been found that individuals with the lowest FEVi values have the highest levels of CRP, and
those with the highest FEVi values have the lowest CRP levels (Manino et al., 2003; Sin and
Man, 2003). This complex series of physiological and biochemical reactions following Oj
exposure may tilt the biological homeostasis mechanisms which could lead to adverse health
effects in people with compromised cardiopulmonary systems.
                                         3-61

-------
       Of much interest are several other types of newly available data that support reasonable
hypotheses that may help to explain the findings of (Vrelated increases in cardiovascular
mortality observed in some epidemiological studies.  These include the direct effect of Os on
increasing PAF in lung tissue that can then enter the general circulation and possibly contribute
to increased risk of blood clot formation and the consequent increased risk of MI,
cerebrovascular events (stroke), or associated cardiovascular-related mortality.  Ozone reactions
with cholesterol in lung surfactant to form epoxides and oxysterols that are cytotoxic to lung and
heart muscles and that contribute to atherosclerotic plaque formation in arterial walls represent
another potential pathway. Stimulation of airway irritant receptors may lead to increases in
tissue and serum levels of ANF, changes in heart rate, and edema of heart tissue.  A few new
field and panel studies of human adults have reported associations between ambient Os
concentrations and changes in cardiac autonomic control (e.g., HRV, ventricular arrhythmias,
and MI). These represent plausible pathways that may lead to O3-related contributions to
cardiovascular effects that ultimately  increase the risk of mortality.
       In addition, Os-induced increases in lung permeability allow more ready entry for inhaled
PM into the blood stream, and Os exposure may increase the risk of PM-related cardiovascular
effects.  Furthermore, increased ambient Os levels contribute to ultrafine PM formation in the
ambient air and indoor environments.  Thus, the contributions of elevated ambient Os
concentrations to ultrafine PM formation and human exposure, along with the enhanced uptake
of inhaled fine particles, consequently may contribute to exacerbation  of PM-induced
cardiovascular effects in addition to those more directly induced by Os (CD, p.  8-53).

3.6   OZONE-RELATED IMPACTS ON PUBLIC HEALTH
       The following discussion draws from section 8.7 of the CD to characterize factors which
modify responsiveness to Os, subpopulations potentially at risk for Os-related health effects, and
potential public health impacts associated with exposure to ambient Os. Providing appropriate
protection of public health requires that a distinction be made between those effects that are
considered adverse health effects and those that are not adverse.  What constitutes an adverse
health effect depends not only on the type and magnitude of effect but also on the population
group being affected. While some changes in healthy individuals would not be considered
adverse, similar changes in susceptible individuals would be seen as adverse. In order to
estimate the potential public health impact, it is important to consider both the susceptible
subpopulations for Os exposure and the definition of adversity for Os health effects.
                                          3-62

-------
       3.6.1   Factors that Modify Responsiveness to Ozone
       There are numerous factors that can modify individual responsiveness to 63. These
include: influence of physical activity; age; gender and hormonal influences; racial, ethnic and
socioeconomic status (SES) factors; environmental factors; and oxidant-antioxidant balance.
These factors are discussed in more detail in section 6.5 of the CD.
       It is well established that physical activity increases an individual's minute ventilation
and will thus increase the dose of 63 inhaled (CD, section 6.5.4). Increased physical activity
results in deeper penetration of O3 into more distal regions of the lungs, which are more sensitive
to acute Os response and injury.  This will result in greater lung function decrements for acute
exposures of individuals during increased physical activity.  Research has shown that respiratory
effects are observed at lower Os concentrations if the level of exertion is increased and/or
duration of exposure and exertion are extended. Predicted Os-induced decrements in lung
function have been shown to be a function of exposure duration and exercise level for healthy,
young adults (McDonnell et al., 1997).
       Most of the studies investigating the influence of age have used lung function decrements
and symptoms as measures of response. For healthy adults, lung function and symptom
responses to Os decline as age increases.  The rate of decline in Os responsiveness appears
greater in those 18 to 35 years old compared to those 35 to 55 years old, while there is very little
change after age 55. In one study (Seal et al., 1996) analyzing a large data set, a 5.4% decrement
in FEVi was estimated for 20 year old individuals exposed to 0.12 ppm Os, whereas similar
exposure of 35 year old individuals were estimated to have a 2.6% decrement. While healthy
children tend not to report respiratory symptoms when exposed to low levels of Os, for subjects
18 to 36 years old symptom responses induced by Os tend to decrease with increasing age
(McDonnell et al., 1999).
       Limited evidence of gender differences in response to Os exposure has suggested that
females may be predisposed to a greater susceptibility to 63. Lower plasma and NL fluid levels
of the most prevalent antioxidant, uric acid, in females relative to males may be a contributing
factor (Housley et al., 1996). Consequently, reduced removal of Os in the upper airways may
promote deeper penetration.  However, most of the evidence on gender differences appears to be
equivocal, with one study (Hazucha et al., 2003) suggesting that physiological responses of
young healthy males and females may be comparable (CD, section 6.5.2).
       A few studies have suggested that ethnic minorities might be more responsive to 63 than
Caucasian population groups (CD, section 6.5.3).  This may be more the result of a lack of
adequate health care and socioeconomic status than any differences in sensitivity to 63. The
limited data available, which have investigated the influence of race, ethnic or other related
factors on responsiveness to Os, prevent drawing any clear conclusions at this time.
                                         3-63

-------
       Few human studies have examined the potential influence of environmental factors such
as the sensitivity of individuals who voluntarily smoke tobacco (i.e., smokers) and the effect of
high temperatures. New controlled human exposure studies have confirmed that smokers are
less responsive to 63 than nonsmokers; however, time course of development and recovery of
these effects, as well as reproducibility, was not different from nonsmokers (CD, section 6.5.5).
Influence of ambient temperature on pulmonary effects induced by Os has been studied very
little, but additive effects of heat and 63 exposure have been reported.
       Antioxidants, which scavenge free radicals and limit lipid peroxidation in the ELF, are
the first line of defense against oxidative stress. Ozone exposure leads to absorption of O3 in the
ELF with subsequent depletion of ELF antioxidant level in the nasal ELF, but concentration and
antioxidant enzyme activity in ELF or plasma do not appear related to O3 responsiveness (CD,
section 6.5.6).  Controlled studies of dietary antioxidant supplements have shown some
protective effects on lung function decrements but not on symptoms and airway inflammatory
responses.  Dietary antioxidant supplements have provided some protection to asthmatics by
attenuating post-exposure airway hyperresponsiveness.  Animal studies have also supported the
protective effects of ELF antioxidants.

       3.6.2   Susceptible Population Groups
       Several characteristics that may increase the extent to which a population group shows
sensitivity to O3 have been discussed in the CD, in the sections on clinical studies in Chapter 6,
epidemiological studies in Chapter 7, and in the integrated assessment in Chapter 8; this section
will draw on all of these. The characteristics that likely increase susceptibility to O3 are based
on:  (1) activity patterns; (2) lung disease; (3) age; and (4) biological responsiveness to O3.
Other groups that might have enhanced sensitivity to O3, but for which there is currently very
little evidence, include: people  with heart disease; groups based on race, gender and
socioeconomic status; and those with nutritional deficiencies.
        3.6.2.1     Active People
       A large group of individuals at risk from O3 exposure consists of outdoor workers and
children, adolescents, and adults who engage  in outdoor activities involving exertion or exercise
during summer daylight hours when ambient O3 concentrations tend to be higher. This
conclusion is based on a large number of controlled-exposure human studies and several
epidemiologic field/panel studies which have been conducted with healthy children and adults
and those with preexisting respiratory diseases (CD, sections 6.2 and 6.3).  The controlled
human exposure studies show a clear O3 exposure-response relationship with increasing
spirometric and symptomatic response as exercise level increases. Furthermore, O3-induced
response increases as time of exposure increases. Studies of outdoor workers and others who
                                         3-64

-------
participate in outdoor activities indicate that extended exposures to Os at elevated exertion levels
can produce marked effects on lung function.
       The effects of Os on the respiratory health of outdoor workers and others who participate
in outdoor activities have been investigated in several recent epidemiologic studies. These
individuals may experience increased vulnerability for O3 health effects, because they are
typically exposed to high doses of Os as they spend long hours outdoors often at elevated
exertion levels. In a group of berry pickers in Fraser Valley, Canada, decrements in lung
function (-5% decrease in FEVi per 40 ppb increase in 1-hr max Os) were associated with acute
exposure to relatively low concentrations of Os (Brauer et al., 1996). The mean ambient 1-hr
max 63 was 40.3 ppb (SD 15.2) over the study period of June to August 1993.  The berry pickers
worked outdoors for an average of 11 hr at elevated heart rates (on average, 36% higher than
resting levels). These results indicate that extended exposures to 63 at elevated exertion levels
can produce marked effects on lung function among outdoor workers.
       Hoppe  et al. (1995) examined forestry workers for Os-related changes in pulmonary
function in Munich, Germany. Ventilation rates, estimated from their average activity levels,
were elevated.  When comparisons were made between high Os days (mean l/2-hr max Os of 64
ppb) and low Os days (mean %-hr max  Os of 32 ppb),  59% of the forestry workers experienced a
notable decrement in lung function (i.e., at least a 20% increase in specific airway resistance or
at least a 10% decrease in FEVi, FVC,  or PEF) on high Os days. None experienced improved
lung function.  This study also examined athletes following a 2-hr outdoor training period in the
afternoon yielding a ventilation rate double the estimate for the forestry workers. Though a
significant association between ambient Os levels and decrements in FEVi was  observed overall,
a smaller percentage of the athletes (14%) experienced a notable decrement in lung function on
high O3 days compared to the forestry workers; and 19% of the athletes actually showed an
improvement.
       A large field study by Korrick et al. (1998) examined the effects of multi-hour 63
exposures (on average, 8 hr) on adults hiking outdoors on Mount Washington, in NH. The mean
of the hourly Os concentrations during the hike was 40 ppb  (range 21-74).  After the hike, all
subjects combined experienced a relatively small mean decline in FEVi (1.5% decrease per 30
ppb increase in mean hourly Os concentrations) during the hike. Ozone-related changes in lung
function parameters were estimated. Stratifying the data by hiking duration indicated that
individuals who hiked 8 to 12 hr experienced a 2-fold greater decline in FEVi versus those only
hiking 2 to 8 hr.
       Results from the above field studies are consistent with those from earlier summer camp
studies (Avol et al., 1990; Higgins et al., 1990; Raizenne et al., 1987, 1989; Spektor et al., 1988,
                                         3-65

-------
1991), which also observed strong associations between acute O3 exposure and decrements in
lung function among children who spent long hours outdoors. In a recent analysis by the
Southern California Children's Health Study, a total of 3,535 initially nonasthmatic children
(ages 9 to 16 years at enrollment) were followed for up to 5 years to identify new-onset asthma
cases associated with higher long-term ambient O3 concentrations (McConnell et al., 2002).
Communities were stratified by pollution levels, with six high-O3 communities (mean 1-hr max
O3 of 75.4 ppb [SD 6.8] over four years) and six low-O3 communities (mean 50.1 ppb
[SD 11.0]).  In the combined analysis using all children, asthma risk was not found to be higher
for residents of the six high-O3 communities versus those from the six low-O3 communities.
However, within the high-O3 communities, asthma risk was more than 3 times greater for
children who played three or more sports versus those who played no sports, an association not
observed in the low-O3 communities.  Therefore, among children repeatedly exposed to higher
O3 levels, increased exertion outdoors (and resulting increased O3 dose) was associated with
excess asthma risk.
       These field studies with subjects at elevated exertion levels support the extensive
evidence derived from controlled human exposure studies.  The majority of human chamber
studies have examined the effects of O3 exposure in subjects performing continuous or
intermittent exercise for variable periods of time. Significant O3-induced respiratory responses
have been observed in clinical studies of exercising individuals. The epidemiologic studies
discussed above also indicate that prolonged exposure periods, combined with elevated levels of
exertion or exercise, may magnify O3 effects on lung function. Thus,  outdoor workers and others
who participate in higher exertion activities outdoors during the time of day when high peak O3
concentrations occur appear to be particularly vulnerable to O3 effects on respiratory health.
Although these studies show a wide variability of response and sensitivity among subjects and
the factors contributing to this variability continue to be incompletely understood, the effect of
increased exertion is consistent. It should be noted that this wide variability of response and
sensitivity among subjects may be in part due to the wide range of other highly reactive
photochemical oxidants coexisting with O3 in the ambient air.

        3.6.2.2     People with Lung Disease
       People with preexisting pulmonary disease are likely to be among those at increased risk
from O3 exposure. Altered physiological, morphological and biochemical states typical of
respiratory diseases like asthma, COPD and chronic bronchitis may render people sensitive to
additional oxidative burden induced by O3 exposure. The new results from  controlled exposure
and epidemiologic studies continue to indicate that asthmatics are a sensitive subpopulation for
O3 health effects.
                                         3-66

-------
       A number of epidemiological studies have been conducted using asthmatic study
populations. The majority of epidemiological panel studies that evaluated respiratory symptoms
and medication use related to Os exposures focused on children. These studies suggest that Os
exposure may be associated with increased respiratory symptoms and medication use in children
with asthma. Other reported effects include respiratory symptoms, lung function decrements,
and ED visits, as discussed in the CD (section 7.6.7.1).  Strong evidence from a large multi-city
study (Mortimer et al., 2002), along with support from several single-city studies suggest that 63
exposure may be associated with increased respiratory symptoms and medication use in children
with asthma. With regard to ambient Os levels and increased hospital admissions and ED visits
for asthma and other respiratory causes, strong and consistent evidence establishes a correlation
between Os exposure and increased exacerbations of preexisting respiratory disease for 1-hr
maximum 63 concentrations <0.12 ppm. Several hospital admission and ED visit studies in the
U.S. (Peel et al., 2005), Canada (Burnett et al., 1997a; Anderson et al., 1997), and Europe
(Anderson et al., 1997) have reported positive associations between increase in Os and increased
risk of ED visits and hospital admissions, especially during the warm season.
       Several clinical studies reviewed in the 1996 CD on atopic and asthmatic subjects had
suggested but not clearly demonstrated enhanced responsiveness to acute Os exposure compared
to healthy subjects.  The majority of the newer studies reviewed in Chapter 6 of the CD indicate
that asthmatics are as sensitive  as, if not more sensitive than, normal subjects in manifesting
induced pulmonary function decrements.
       Ozone-induced increases in neutrophils, protein, and IL-8 were found to be significantly
higher in the BAL fluid from asthmatics compared to healthy subjects, suggesting mechanisms
for the increased sensitivity of asthmatics.  Similarly, subjects with allergic asthma exhibited
increased  airway responsiveness to inhaled allergens upon acute O3 exposure. Asthmatics
present a differential response profile for cellular, molecular, and biochemical parameters (CD,
Figure 8-1) that are altered in response to acute 63 exposure.  Increases in (Vinduced
nonspecific airway responsiveness incidence and duration could have important clinical
implications for asthmatics.
       Bronchial constriction following provocation with allergens presents a two-phase
response.  The early response is mediated by release of histamine and leukotrienes that leads to
contraction of smooth muscle cells in the bronchi, narrowing the lumen and decreasing the
airflow. In asthmatics, these mediators also cause accumulation of eosinophils, followed by
production of mucus and a late-phase bronchial constriction and reduced airflow.  Holz et al.
(2002) reported an early phase response in subjects with rhinitis after a consecutive 4-day
exposure to 0.125 ppm Osthat resulted in a clinically relevant (>20%) decrease in FEVi.
Allergen challenge in mild asthmatics 24 hr before exposure to 0.27 ppm O?, for 2 hr resulted in
                                         3-67

-------
significantly increased eosinophil counts in samples of respiratory tract lining fluid, obtained by
sputum induction, compared to results in healthy subjects (Vagaggini et al., 2002). Epithelial
cells from mucosal biopsies of allergic asthmatics indicated significant increases in the
expression of IL-5, IL-8 and GM-CSF, suggesting increased neutrophilic inflammation
compared to healthy subjects (Bosson et al., 2003).
       Several human exposure studies have shown differences between asthmatics and healthy
human subjects with regard to PMN influx in BAL fluid.  In vitro studies (Schierhorn et al.,
1999) of nasal mucosal biopsies from atopic and nonatopic subjects exposed to 0.1 ppm Os found
significant differences in release of IL-4, IL-6, IL-8, and TNF-a. Another study by Schierhorn et
al. (2002) found significant differences in the Os-induced release of the neuropeptides neurokinin
A and substance P for allergic patients in comparison to nonallergic controls, suggesting
increased activation of sensory nerves by O3 in the allergic tissues. Another study by Bayram et
al. (2002) using in vitro culture of bronchial epithelial cells recovered from atopic and nonatopic
asthmatics also found significant increases in epithelial permeability in response to 63 exposure.
In addition, some controlled human Os exposure studies in asthmatics (Hiltermann et al., 1999;
Scannell et al., 1996) reported increased secretion of IL-8, suggesting increased neutrophilic
inflammation. Two studies (Torres et al., 1996; Holz et al., 2002) observed increased airway
responsiveness to repeated daily Os exposure to bronchial allergen challenge in subjects with
preexisting allergic airway disease.
       Newly available reports from controlled human exposure studies (see Chapter 6 in the
CD) utilized subjects with preexisting cardiopulmonary diseases such as COPD, asthma, allergic
rhinitis, and hypertension. The data generated from these studies that evaluated pulmonary
function changes in spirometry did not find clear differences between filtered air and O?, exposure
in COPD subjects. However, the new data on airway responsiveness, inflammation, and various
molecular markers of inflammation and bronchoconstriction indicate that people with atopic
asthma and allergic rhinitis comprise  susceptible groups for O3-induced adverse health effects.
       Although controlled human exposure studies have not found  evidence of larger
spirometric changes in people with COPD relative to healthy subjects, this may be due to the fact
that most people with COPD are older adults who would not be expected to have  such changes
based on their age. However, in Section 8.7.1, the CD notes that new epidemiological evidence
indicates that people with COPD may be more likely to experience other effects, including
emergency room visits, hospital admissions, or premature mortality.  For example, results from
an analysis of five European cities indicated strong and consistent Os effects on unscheduled
respiratory hospital admissions, including COPD (Anderson et al., 1997).  Also, an analysis of a
9-year data set for the whole population of the Netherlands provided risk estimates for more
specific causes of mortality, including COPD (Hoek et al., 2000, 2001; reanalysis Hoek, 2003); a
                                          3-68

-------
positive, but nonsignificant, excess risk of COPD-related mortality was found to be associated
with short-term 63 concentrations. Moreover, as indicated by Gong et al. (1998), the effects of
Os exposure on alveolar-arterial oxygen gradients may be more pronounced in patients with
preexisting obstructive lung diseases.  Relative to healthy elderly subjects, COPD patients have
reduced gas exchange and low SaO2.  Any inflammatory or edematous responses due to O3
delivered to the well-ventilated regions of the COPD lung could further inhibit gas exchange and
reduce oxygen saturation. In addition, (Vinduced vasoconstriction could also acutely induce
pulmonary hypertension.  Inducing pulmonary vasoconstriction and hypertension in these
patients would perhaps worsen their condition, especially if their right ventricular function was
already compromised (CD, Section 6.10).

        3.6.2.3    Children and Older Adults
       Supporting evidence exists for heterogeneity in the effects of Os by age.  As discussed in
section 6.5.1 of the CD, children, adolescents, and young adults (<18 yrs of age) appear, on
average, to have nearly equivalent spirometric responses to O3, but have greater responses than
middle-aged and older adults when exposed to comparable Os doses. Symptomatic responses to
63 exposure, however, do not appear to occur in healthy children, but are observed in asthmatic
children, particularly those who use maintenance medications. For adults (>17 yrs of age)
symptoms gradually decrease with increasing age. In contrast to young adults, the diminished
symptomatic responses in children and symptomatic and spirometric responses in the elderly
may put them at an increased risk for continued exposure.
       As described in the section 7.6.7.2 of the CD, many epidemiological field studies focused
on the effect of Os on the respiratory health of school  children. In general, children experienced
decrements in pulmonary function parameters, including PEF, FEVi, and FVC. Increases in
respiratory symptoms and asthma medication use were also observed in asthmatic children. In
one German study, children with and without asthma were found to be particularly susceptible to
Os effects on lung function. Approximately 20% of the children, both with and without asthma,
experienced a greater than 10% change in FEVi, compared to only 5% of the elderly population
and athletes (Hoppe et al., 2003).
       The American Academy  of Pediatrics (2004) notes that children and infants are among
the population groups most susceptible to many air pollutants, including 63.  This is in part
because their lungs are still developing.  For example, eighty percent of alveoli are formed after
birth, and changes in lung development continue through adolescence (Dietert et al., 2000).
Children are also likely to spend more time outdoors than adults do, which results in increased
exposure to air pollutants (Wiley et al., 1991a,b). Moreover, children have high minute
                                         3-69

-------
ventilation rates and high levels of physical activity which also increases their dose (Plunkett et
al., 1992).
       Several mortality studies have investigated age-related differences in Os effects. Among
the studies that observed positive associations between 63 and mortality, a comparison of all age
or younger age (<65 years of age) Os-mortality effect estimates to that of the elderly population
(>65 years) indicates that, in general, the elderly population is more susceptible to Os mortality
effects (Borja-Aburto et al., 1997; Bremner et al., 1999; Gouveia and Fletcher, 2000; O'Neill
et al., 2004;  Simpson et al., 1997;  Sartor et al., 1995; Sunyer et al., 2002).  For example, a study
by Gouveia and Fletcher (2000) examined the (Vmortality effect by age in Sao Paulo, Brazil.
Among all ages, Os was associated with a 0.6% excess risk in all cause mortality per 40 ppb
increase in 1-hr max (V  In comparison, in the elderly population,  the (Vmortality risk estimate
was nearly threefold greater, at 1.7%.  Similarly, a Mexico City study found that CVmortality
effect estimates were 1.3% and 2.8% per 20 ppb increase in 24-hr average 63 concentration in all
ages and the elderly, respectively (O'Neill et al., 2004).
       The meta-analysis by Bell  et al. (2005) found a larger effect estimate for the elderly
(2.92% per 20 ppb increase in 24-hr average O3) than for all ages (1.75%).  In the large U.S. 95
communities study (Bell et al., 2004), effect estimates were slightly higher for those aged 65 to
74 years, 1.40% excess risk per  20 ppb increase in 24-hr average Os, compared to individuals
less than 65 years and 75 years or greater, 1.00% and 1.04%, respectively, using a constrained
distributed 7-day lag model. Bell  et al. (2004) note that despite similar effects  estimates, the
absolute effect of 63 is substantially greater in the elderly population due to the higher
underlying mortality rates, which lead to a larger number of extra deaths for the elderly
compared to the general population. The CD concludes that the elderly population (>65 years of
age) appear to be  at greater risk  of (Vrelated mortality and hospitalizations compared to all ages
or younger populations (CD, p. 7-177).
       The CD notes that, collectively, there is supporting evidence of age-related differences in
susceptibility to 63 lung function effects.  The elderly population (>65 years of age) appear to be
at increased risk of (Vrelated mortality and hospitalizations, and children (<18 years of age)
experience other potentially adverse respiratory health outcomes with increased 63 exposure
(CD, section 7.6.7.2).

        3.6.2.4      People with Increased Responsiveness to Ozone
       New animal toxicology studies using various strains of mice and rats have identified O3-
sensitive and resistant strains and illustrated the importance of genetic background in
determining 63 susceptibility (CD, section 8.7.4). Using subacute low exposure regimen (0.3
ppm Os, 48h) studies on inbred strains that have been designated as inflammation prone or
                                          3-70

-------
resistant, Kleeberger et al., (1997) identified the pro-inflammatory cytokine gene, Tnf-a, as a
susceptibility gene.  Further characterization of this model indicated a role for TNF receptors
(TNFR1, TNFR2) in CVinduced pulmonary epithelial injury and inflammation (Cho et al.,
2001). Studies on five inbred strains of mouse with differing response to O3 exposure (acute
high dose or low dose continuous exposure for 3 days), reported a protective role for clara cell
secretory protein (CCSP) against (Vinduced oxidative damage (Broeckaert et al., 2003; Wattiez
et al., 2003).  The role for these genes and/or their orthologs in human susceptibility to Os
exposure is yet to be examined.
       Apart from age at the time of exposure, controlled human exposure studies have also
indicated a high degree of interindividual variability in some of the pulmonary physiological
parameters.  Recent studies by David et al. (2003) and Romieu et al. (2004) reported a role for
genetic polymorphism in antioxidant enzymes and genes involved in inflammation to modulate
pulmonary function and inflammatory responses to Os exposure. Similar to mouse studies
referred above, polymorphism in Tnf-a has been implicated in (Vinduced lung function changes
in healthy, mild asthmatics and individuals with rhinitis. These observations suggest a potential
role for these markers in the innate susceptibility to O3, however, the validity of these markers
and their relevance in the context of prediction to population studies needs additional
experimentation.
       Biochemical and molecular parameters extensively evaluated in these experiments were
used to identify specific loci on the chromosomes and, in some cases, to relate the differential
expression of specific genes to biochemical and physiological differences observed among these
species. Utilizing CVsensitive and CVresistant species, it has been possible to identify the
involvement of AHR and inflammation processes in Os susceptibility. However, most of these
studies were carried out using relatively high doses of O3, making the relevance of these studies
questionable in human health effects assessment.  The molecular parameters identified in these
studies may serve as useful biomarkers with the availability of suitable technologies and,
ultimately, can likely be integrated with epidemiological studies.  Interindividual differences in
Os responsiveness have been observed across a spectrum of symptoms and lung function
responses do not yet allow identification of important underlying factors, except a significant
role for age.

        3.6.2.5     Other Population Groups
       There is limited, new evidence supporting associations between short-term Os exposures
and a range of effects on the cardiovascular system.  Some but not all, epidemiological studies
have reported associations  between short-term Os exposures and the incidence of myocardial
infarction and more subtle  cardiovascular health endpoints, such as changes in heart rate
                                         3-71

-------
variability and cardiac arrhythmia.  Others have reported associations with hospitalization or ED
visits for cardiovascular diseases, although the results across the studies are not consistent.
Studies also report associations between short-term Os exposure and mortality from
cardiovascular or cardiopulmonary causes. The CD concludes that current cardiac physiologic
effects evidence from some field studies is rather limited but supportive of a potential effect of
short-term O3 exposure and HRV, cardiac arrhythmia, and MI incidence (CD, p. 7-65).  In the
CD's evaluation of studies of hospital admissions for cardiovascular disease (CD, section 7.3.4),
it is concluded that evidence from this growing group of studies is generally inconclusive
regarding an association with O3 in studies conducted during the warm  season (CD, p. 7-83).
This body of evidence suggests that people with heart disease may be at increased risk from
short-term exposures to Os; however, more evidence is needed to conclude that people with heart
disease are a susceptible population.
       Other groups that might have enhanced sensitivity to O3, but for which there is currently
very little evidence, include groups based on race, gender and socioeconomic status, and those
with nutritional deficiencies, as discussed in section 3.6.1 which presents factors which  modify
responsiveness to Os.

       3.6.3  What Constitutes an Adverse Health Impact from Ozone Exposure?
       In making judgments as to when various Os-related effects become regarded as adverse
to the health of individuals, in previous NAAQS reviews staff has relied upon the guidelines
published by the American Thoracic Society (ATS) and the advice of CASAC. While
recognizing that perceptions of "medical significance" and "normal  activity" may differ among
physicians, lung physiologists and experimental subjects, the ATS (1985) defined adverse
respiratory health effects as "medically significant physiologic changes generally evidenced by
one or more of the  following: (1) interference with the normal activity of the affected person or
persons, (2) episodic respiratory illness, (3) incapacitating illness, (4) permanent respiratory
injury, and/or (5) progressive respiratory dysfunction."
       During the  1997 review, it was concluded that there was evidence of causal associations
from controlled human exposure studies for effects in the first of these five ATS-defined
categories, evidence of statistically significant associations from epidemiological studies for
effects in the second and third categories, and evidence from animal toxicology studies, which
could be extrapolated to humans only with a significant  degree of uncertainty, for the last two
categories. For the current review, the evidence of Os-related effects is stronger across all the
categories. For ethical  reasons, clear causal evidence from controlled human exposure studies
still covers only effects in the first category. However, for this review there are results from
epidemiological studies, upon which to base judgments about adversity, for effects in all of the
                                          3-72

-------
categories. Statistically significant and robust associations have been reported in epidemiology
studies falling into the second and third categories.  These more serious effects include
respiratory illness that may require medication (e.g., asthma), but not necessarily hospitalization,
as well as respiratory hospital admissions and ED visits for respiratory causes. Less conclusive,
but still positive associations have been reported for school absences and cardiovascular hospital
admissions. Human health effects for which associations have been suggested through evidence
from epidemiological and animal toxicology studies, but have not been conclusively
demonstrated still fall primarily into the last two categories.  In the last review of the Os
standard, evidence for these more serious effects came from  studies of effects in laboratory
animals. Evidence from animal studies evaluated in this CD strongly suggests that 63 is capable
of damaging the distal airways and proximal alveoli, resulting in lung tissue remodeling leading
to apparently irreversible changes.  Recent advancements of dosimetry modeling also provide a
better basis for extrapolation from animals to humans.  Information from epidemiological studies
provides supporting, but limited evidence of irreversible respiratory effects in humans (as
described in section 6.3.3.2 below). Moreover, the CD concludes that the findings from single-
city and multi-city time-series epidemiology studies and meta-analyses of these epidemiology
studies support a likely causal association between short-term Os exposure and mortality
particularly in the warm season.
       While Os has been associated with effects that are clearly adverse, application of these
guidelines, in particular to the least serious category of effects related to ambient 63 exposures,
involves judgments about which medical experts on the CASAC panel and public commenters
have in the past expressed diverse views. To help frame such judgments, we have defined
gradations of individual functional  responses (e.g., decrements in FEVi and airway
responsiveness) and symptomatic responses (e.g., cough, chest pain, wheeze), together with
judgments as to the potential impact on individuals experiencing varying degrees of severity of
these responses, that have been used in previous NAAQS reviews. These gradations and impacts
are summarized in Tables 3-2 and 3-3.
       For active healthy people, moderate levels of functional responses (e.g., FEVi
decrements of >10% but < 20%, lasting up to 24 hrs) and/or  moderate symptomatic responses
(e.g., frequent spontaneous cough, marked discomfort on exercise or deep breath, lasting up to
24 hrs) would likely interfere with normal activity for relatively few sensitive individuals;
whereas large functional responses (e.g.,  FEVi decrements > 20%, lasting longer than 24 hrs)
and/or severe symptomatic responses (e.g., persistent uncontrollable cough,  severe discomfort on
exercise or deep breath, lasting longer than 24 hrs) would likely interfere with normal activities
                                          3-73

-------
Table 3-2. Gradation of Individual Responses to Short-Term Ozone Exposure in Healthy
           Persons1
Functional Response
FEVi


Nonspecific
airway responsiveness2
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 < 10%

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
<24 hrs
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
>20%

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
                                            5-74
 1 This table is reproduced from the 1996 O3 AQCD (Table 9-1, page 9-24) (U.S. Environmental Protection
 Agency, 1996).
 2 An increase in nonspecific airway responsiveness of 100% is equivalent to a 50% decrease in PD20 or PD100.

-------
 Table 3-3. Gradation of Individual Responses to Short-Term Ozone Exposure in Persons

            with Impaired Respiratory Systems3
Functional
Response
FEVi change
Nonspecific
airway
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 < 10%
Increases of < 100%

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 <300%

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
>20%
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
>24hr

Severe
Most individuals
choose to limit
activity
Physician or
emergency room
visit
3-75
 This table is reproduced from the 1996 O3 AQCD (Table 9-1, page 9-25) (U.S. Environmental Protection

Agency, 1996).
4

 An increase in nonspecific airway responsiveness of 100% is equivalent to a 50% decrease in PD20 or PD100.

-------
for many sensitive individuals and therefore would be considered adverse under ATS guidelines.
For the purpose of estimating potentially adverse lung function decrements in active healthy
people, the CAS AC indicated that a focus on the mid to upper end of the range of moderate
levels of functional responses is most appropriate (e.g., FEVi decrements > 15% but < 20%)5.
However, for people with lung disease, even moderate functional (e.g., FEVi decrements > 10%
but < 20%, lasting up to 24 hr) or symptomatic responses (e.g., frequent spontaneous cough,
marked discomfort on exercise or with deep breath, wheeze accompanied by  shortness of breath,
lasting up to 24 hr) would likely interfere with normal activity for many individuals, and would
likely result in additional and more frequent use of medication. For people with lung disease,
large functional responses (e.g., FEVi decrements > 20%, lasting longer than 24 hrs) and/or
severe symptomatic responses (e.g., persistent uncontrollable cough, severe discomfort on
exercise or deep breath, persistent wheeze accompanied by shortness of breath, lasting longer
than 24 hrs) would  likely interfere with normal activity for most individuals and would increase
the likelihood that these individuals would seek medical treatment.  For the purpose of
estimating potentially adverse lung function decrements in people with lung disease, the CASAC
indicated that a focus on the lower end of the range of moderate levels of functional responses is
most appropriate (e.g., FEVi decrements > 10%).
      In judging the extent to which these impacts represent effects that should be regarded as
adverse to the health status of individuals, an additional  factor that has been considered in
previous NAAQS reviews is whether such effects are experienced repeatedly during the course
of a year or only on a single occasion. While  some experts would judge single occurrences of
moderate responses to be a "nuisance," especially for healthy individuals, a more general
consensus view of the adversity of such moderate responses emerges as the frequency of
occurrence increases. Thus it has been judged that repeated occurrences  of moderate responses,
even in otherwise healthy individuals, may be considered to be adverse since they could well set
the stage for more serious illness (61 FR 65723). The CASAC panel in the last review expressed
a consensus view that these "criteria for the determination of an adverse physiological response
were reasonable" (Wolff, 1995).
      In 2000, the American Thoracic Society (ATS) published an official statement on "What
Constitutes an Adverse Health Effect of Air Pollution?" (ATS, 2000), which updated its earlier
guidance (ATS, 1985).  The revised guidance was intended to address  new investigative
approaches used to identify the effects of air pollution, and to reflect the concern for the impacts
of air pollution on specific groups that had been expressed through the environmental justice
movement.
        Transcript of CASAC meeting, day 8/24/06, page 149.
                                         3-76

-------
       The new guidance builds upon and expands the 1985 definition of adversity in several
ways. There is an increased focus on quality of life measures as indicators of adversity. There is
also a more specific consideration of population risk. Exposure to air pollution that increases the
risk of an adverse effect to the entire population is adverse, even though it may not increase the
risk of any individual to an unacceptable level.  For example, a population of asthmatics could
have a distribution  of lung function such that no individual has a level associated with significant
impairment. Exposure to air pollution could shift the distribution to lower levels that still do not
bring any individual to a level that is associated with clinically relevant effects. However, this
would be considered to be adverse because individuals within the population would have
diminished reserve function, and therefore would be at increased risk if affected by another
agent.
       Of the various  effects of 63 exposure that have been studied, many would meet the ATS
definition of adversity. Such effects include, for example, any detectible level of permanent lung
function loss attributable to air pollution, including both reductions in lung growth or
acceleration of the  age-related decline of lung function; exacerbations of disease in individuals
with chronic cardiopulmonary diseases; reversible loss of lung function in combination with the
presence of symptoms; as well as more serious  effects such as those requiring medical  care
including hospitalization and, obviously, mortality.
       As discussed above, relatively small, reversible declines in lung function parameters may
be of questionable  significance in healthy people.  However, a 5 to 15 % change in FEVi is
considered to have clinical importance to asthma morbidity (ATS 1991; Lebowitz et al. 1987;
Lippmann, 1988).  This is in line with the view expressed by the CASAC that a focus on the
lower end of the range of moderate levels of functional responses is most appropriate (e.g., FEVi
decrements > 10%) to estimate the risk of potentially adverse lung function responses in people
with lung disease.  The National Institutes of Health (1997) has stated that a PEF below 80% of a
person's personal best indicates a need for continued medication use in asthmatics. In Mortimer
et al. (2002), 63 was associated with increased incidence of > 10% declines in morning PEF as
well as morning symptoms, indicating that O3 exposure may have clinically significant effects on
asthmatic children.
       Reflecting new investigative approaches, the ATS statement describes the potential
usefulness of research into the genetic basis for disease, including responses to  environmental
agents that will provide insights into the mechanistic basis for susceptibility, and provide
markers of risk status.  Likewise biomarkers, that are indicators of exposure, effect or
susceptibility, may someday be useful in defining the point at which a response should be
equated with an adverse effect. Based on concern for segments of the population that may be
disproportionately  exposed to environmental contaminants, or have other factors that may
                                          3-77

-------
increase susceptibility (e.g., genetic or nutritional factors), there was a call for increased research
in these areas.
       Overall, the new guidance does not fundamentally change the approach previously taken
to define adversity, nor does it suggest a need at this time to change the structure or content of
the tables describing gradation of severity and adversity of effects in Tables 3-2 or 3-3 above.

       3.6.4   Estimation of Potential Numbers of People in At-Risk Susceptible Population
              Groups in the United States
       Although Ch-related health risk estimates may appear to be numerically small, their
significance from an overall  public health perspective is affected by the large numbers of
individuals in potential risk groups.  Several subpopulations may be identified as having
increased susceptibility or vulnerability to adverse health effects from Os, including: older adults,
children, individuals with preexisting pulmonary disease, and those with higher exposure levels,
such as outdoor workers.
       One consideration in the assessment of potential public health impacts is the  size  of
various population groups that may be at increased risk for health effects associated with Os-
related air pollution exposure.  Table 8-4 in the CD  summarizes information on the prevalence of
chronic respiratory conditions in the U.S. population in 2002 and 2003 (Dey and Bloom, 2005;
Lethbridge-Cejku et al., 2004).  Individuals with preexisting cardiopulmonary disease constitute
a fairly large  proportion of the population, with tens of millions of people included in each
disease category.  Of most concern here are those individuals with preexisting respiratory
conditions, with approximately 11% of U.S. adults and 13% of children having been diagnosed
with asthma and 6% of adults having COPD (chronic bronchitis and/or emphysema). Table 8-5
in the CD provides further information on the number of various specific respiratory conditions
per 100 persons by age among the U.S. population during the mid-1990s. Asthma prevalence
tends to be higher in children than adults.
       In addition, subpopulations based on age group also comprise substantial segments  of the
population that may be potentially at risk for Os-related health impacts. Based on U.S. census
data from 2003, about 26% of the U.S. population are under 18 years of age and 12% are 65
years of age or older. Hence, large proportions of the U.S. population are included in age groups
that are considered likely to have increased susceptibility and vulnerability for health effects
from ambient Os exposure.
       The health statistics data illustrate what is known as the "pyramid" of effects. At the top
of the pyramid, there are approximately 2.5 millions deaths from all causes per year in the U.S.
population, with about 100,000 deaths from chronic lower respiratory diseases (Kochanek et al.,
2004). For respiratory health diseases, there are nearly 4 million hospital discharges per year
                                          3-78

-------
(DeFrances et al., 2005), 14 million ED visits (McCaig and Burt, 2005), 112 million ambulatory
care visits (Woodwell and Cherry, 2004), and an estimated 700 million restricted activity days
per year due to respiratory conditions (Adams et al., 1999). Combining small risk estimates with
relatively large baseline levels of health outcomes can result in quite large public health impacts.
Thus, even a small percentage reduction in O3 health impacts on cardiopulmonary diseases would
reflect a large number of avoided cases.
       Another key input for public health impact assessment is the range of concentration
response functions for various health outcomes. Epidemiologic studies have reported
associations between short-term exposure to O3 with mortality, hospitalizations for pulmonary
diseases, ED visits for asthma, reduced lung function, and incidence of respiratory symptoms.
Effect estimates for morbidity responses to short-term changes in O3 tend to be larger and more
variable in magnitude than those for mortality.
       In addition to attribution of risks for various health outcomes related to O3 and other
copollutants, important considerations in assessing the impact of O3 on public health include the
size of population groups at risk, as well as the concentration-response relationship  and potential
identification of threshold levels. Taken together, based on the above information, it can be
concluded that exposure to ambient O3 likely has a significant impact on public health in the U.S.

3.7   SUMMARY AND CONCLUSIONS FOR OZONE HEALTH EFFECTS
       Based on dosimetric, experimental, and epidemiological evidence assessed in the 1996
CD, a set of findings and conclusions were drawn  regarding potential health effects of O3
exposure as  of 1996. These conclusions are  integrated into the Summary and  Conclusions for
Ozone Health Effects in the 2006 CD (section 8.8). The revised CD will be referred to as the
"2006 CD" in this section in order to be more easily distinguished from the "1996 CD." Section
8.8 of the 2006 CD also has summarized the main conclusions derived from the integrated
analysis of animal toxicology (2006 CD, Chapter 5), human experimental (2006 CD, Chapter 6)
and epidemiological (2006 CD, Chapter 7) studies that evaluated evidence of health effects
associated with short-term, prolonged, and long-term exposures to O3 alone or in  combination
with other pollutants commonly found in the ambient air.  This section summarizes conclusions
drawn from  section 8.8 of the 2006 CD with respect to the health effects associated with
exposure to O3 that are most relevant to our assessment of the adequacy of the current primary
O3 standard  and the identification of options to consider concerning potential alternative
standards to protect public health with an adequate margin of safety.
                                         3-79

-------
       3.7.1  Respiratory Morbidity Effects of Short-term Exposures to Ozone
       In the 1996 CD, it was concluded from assessment of controlled human exposure studies
that short-term Os exposures to Os concentrations of > 0.08 ppm for 6.6 to 8 hr under moderate
exertion and > 0.12 ppm for 1 hr under heavy exertion cause decrements in lung function in
children and increased lung function and respiratory symptoms in healthy adults and asthmatic
individuals exposed (2006 CD, p.  8-73). Lung inflammatory responses have been observed in
healthy human adults following 6.6 hr Os exposures as low as 0.08 ppm (2006 CD, p. 8-75).
Changes in lung function, respiratory symptoms, and lung inflammatory responses occur as a
function of exposure concentration, duration, and level of exertion.  Such experimentally
demonstrated effects were consistent with and helped support the plausibility of epidemiological
findings assessed in the 1996 CD regarding daily hospital admissions and ED visits for
respiratory causes.
       The 1996 CD  concluded that group mean data from numerous controlled human exposure
and field studies of healthy subjects (18 to 45 years of age) exposed for 1 to 3 hr indicate that, in
general, statistically significant pulmonary function decrements beyond the range of normal
measurement variability (e.g., 3 to 5% for FEVi) occur in subjects exposed:
     •  at >0.12 ppm Os after very heavy exercise (competitive running).
     •  at >0.18 ppm Os after heavy exercise (easy jogging),
     •  at >0.30 ppm Os after moderate exercise (brisk walking),
     •  at >0.37 ppm Os after light exercise (slow walking), and
     •  at >0.50 ppm Os when at rest.
       Small group mean changes (e.g., <5%) in FEVi have been observed in healthy young
adults at levels as low as 0.12 ppm Os for 1 to 3 hr exposure periods. Also, lung function
decrements have been observed in children and adolescents at concentrations of 0.12 and 0.14
ppm Os with heavy exercise.  Some individuals within a study may experience FEVi decrements
in excess of 15% under these conditions, even when group mean decrements are less than 5%.
       For exposures of healthy, young adult subjects performing moderate exercise during
longer duration exposures (6 to 8 hr), 5% group mean decrements in FEVi were observed at
     •  0.08 ppm after O3 5.6 hr,
     •  0.10 ppm after Os 4.6 hr, and
     •  0.12ppmafterO33hr.
                                         3-80

-------
For these same subjects, 10% group mean FEVi decrements were observed at 0.12 ppm Os after
5.6 and 6.6 hr. As in the shorter duration studies, some individuals experience changes larger
than those represented by group mean changes.
      The 2006 CD (section 8.8) concludes that newer meta-analyses confirmed interindividual
differences in lung function decrements reported in the 1996 CD. Age-specific differences in
lung function responses were also observed. Spirometric responses (i.e., decrements in lung
function) in healthy adults exposed to near ambient 63 levels typically resolve to near baseline
within 4-6 hr. Meta-analyses of four controlled human exposure studies (two new and two
assessed in the 1996 CD) reporting the effects of prolonged (6.6 hr) exposures to 0.08 ppm Os
during moderate exertion on lung function in young healthy adults (M=90, F=30; mean age 23
years) indicate an absolute FEVi decrease of 6%,  whereas FEVi increased by 1% following fresh
air exposures. Newer studies from Adams (2002, 2006), as illustrated earlier in Figure 3-1B,
demonstrate  notable interindividual variability for O3 exposure concentrations of 0.04, 0.06 and
0.08 ppm. Following a continuous exposure to 0.08 ppm Os during intermittent, moderate
exertion, the group mean FEVi decrement (corrected for filtered air) was 6%, but 23 % of
subjects had  FEVi decrements of 10% or more. Following exposure to 0.06 ppm Os, the group
mean FEVi decrement was less than 3%, but 7% of subjects had greater than 10% FEVi
decrements (2006 CD, p. 8-18). However, as discussed in  Section 3.3.1.1.1,  we note that the
pre- and post-exposure data presented in the Adams (2006) study show a small (< 3%) group
mean FEVi decrement following the 6.6-hr exposure at 0.06 ppm, which may be statistically
significantly different from filtered air responses.
      A few controlled human exposure studies  (Adams, 2003; 2006; Hazucha et al., 1992)
investigated  a triangular exposure pattern at 63 concentrations that had 6.6 to 8-hr averages
between 0.08 and 0.12 ppm in order to more closely mimic typical ambient O3 exposure patterns.
Greater overall FEVi decrements were observed with triangular exposures compared to the
constant or square-wave exposures. Furthermore, peak FEVi decrements observed during
triangular exposures were greater than those observed during square-wave patterns. At a lower
average O3 concentration of 0.06 ppm, no temporal (i.e., hour by hour responses) differences
were observed in FEVi decrements between square-wave and triangular exposure patterns.
There was, however, a statistically significant effect of the  0.06 ppm triangular exposure on total
respiratory symptoms following 5.6 and 6.6 h of exposure that was not observed for the 0.06
ppm square-wave exposure protocol. Results of these studies suggest the potential for somewhat
greater effects on lung function in ambient O3  exposure scenarios that typically involve gradually
increasing daily exposure up to a peak in the late afternoon and a subsequent gradual decline
(2006 CD, p. 8-19).  The quantitative risk assessment, discussed below in Chapter 5, provides
                                         3-81

-------
estimates of the percentages of school age children likely to experience FEVi decrements greater
than or equal to 10, 15, and 20% after 8-hr exposures to 63 while engaged in moderate exertion.
       Decrements in lung function associated with ambient Os levels have also been found in
children attending summer camps in southern Ontario, Canada, in the northeastern U.S., and in
southern California (2006 CD, p. 8-74).  The U.S. multeities study by Mortimer et al. (2002)
observed an association between acute Os exposure and the incidence of a >10% decrement in
morning PEF in asthmatic children. Meta-analyses indicate that a 0.50-mL decrease in FEVi is
associated with a 1 ppb increase in Os concentration.  For preadolescent children exposed to 120
ppb (0.12 ppm) ambient Os, this amounts to an average decrement of 2.4 to 3.0% in FEVi.
Similar responses are reported for exercising children and adolescents  exposed to 63 in ambient
air or Os in purified air for 1-2 hours.
      The 1996 CD concluded that an increase in the incidence of cough has been reported at 63
concentrations as low as 0.12 ppm in healthy adults during  1 to 3 hr of exposure with very heavy
exercise. Other respiratory symptoms, such as pain on deep inspiration, shortness of breath, and
lower respiratory scores (i.e., a combination of several symptoms), have been observed at 0.16
ppm to 0.18 ppm Os,  1-hr average, with heavy and very heavy exertion. Respiratory symptoms
also have been observed following exposure to 0.08, 0.10 and 0.12 ppm Os for 6.6 hr with
moderate exertion levels.  Also, increases in nonspecific airway responsiveness in healthy adults
at rest have been observed after 1 to 3 hr of exposures to 0.40 ppm but not to 0.20 ppm Os;
during very heavy exertion, these increases were observed at concentrations as low as 0.18 ppm
but not at 0.12 ppm 63. Increases in nonspecific airway responsiveness during the 6.6 hr
exposures with moderate levels of exertion have been observed at 0.08, 0.10 and 0.12  ppm Os.
       The majority  of asthma panel studies evaluated the associations of ambient 63 with lung
function and respiratory symptoms in asthmatic children. Results obtained from these studies
show some inconsistencies, with some indicating significant positive associations and other
smaller studies not finding such effects.  Overall, however,  the multicity study by Mortimer et al.
(2002)  and several credible single-city studies (e.g., Gent et al., 2003) indicate a fairly robust
association between ambient Os concentrations and increased respiratory symptoms in moderate
to severe asthmatic children (2006 CD, p. 8-35).
       The 2006 CD  (p. 8-75) concludes that lung inflammatory responses have been observed
in healthy human adults following 6.6 hr  63 exposures as low as 0.08 ppm.  These responses
have been found even in the absence  of O3-induced lung function decrements for some
individuals. Attenuation of most inflammatory markers occurs with repeated exposures over
several days, but none of the several markers of lung injury and permeability show attenuation,
which is indicative of continued lung tissue damage during repeated exposure. Laboratory
animal studies have reported that 1 to 3 hr Os exposures as low as 0.1 to 0.5 ppm can cause  (1)
                                         3-82

-------
lung inflammatory responses (e.g., increased ROS and inflammatory cytokines, influx of PMNs,
and activation of AMs); (2) damage to epithelial airway tissues, (3) increases in permeability of
both lung endothelium and epithelium, and (4) increases in susceptibility to infectious diseases
due to modulation of lung host defenses. Consistent with the above results of human and animal
experimental studies, there is limited epidemiologic evidence of an association between acute
ambient O3 exposure (1-hr max of about 0.1 ppm) and airway inflammation in children, all of
which taken together is indicative of a causal role for O3 in inflammatory responses in the
airways (2006 CD, p. 8-76).  See Table 3.4 for a summary of short-term health effects of O3
based on clinical studies assessed in both the 1996 CD and 2006 CD.
       The 1996 CD concluded that increased O3 levels are associated with increased hospital
admissions and ED visits for respiratory causes. Analyses from data in the northeastern U.S.
suggested that O3 air pollution is associated with a substantial portion of all summertime
respiratory hospital visits and admissions. The 2006 CD concludes (CD, p. 8-36) that a large
multi-city and several single-city studies have indicated a positive association between increased
O3 levels (especially during the warm season) and increased  risk for respiratory hospital
admissions and asthma ED visits.

       3.7.2  Cardiovascular Morbidity Effects of Short-term Exposures to Ozone
       One health endpoint that was unrecognized in the 1996 CD, but is addressed in the 2006
CD, is O3-induced cardiovascular effects.  Newly available evidence has emerged since 1996
which provides considerable plausibility for how O3 could exert cardiovascular effects (2006
CD, p. 8-77). Examples of such O3-induced cardiovascular effects include: (1) O3-induced
release from lung epithelial cells of PAF that may contribute to blood clot formation that would
increase the risk of serious cardiovascular outcomes (e.g., heart attack, stroke, mortality); (2)
interactions of O3 with surfactant components in ELF  of the lung resulting  in production of
oxysterols and ROS that may exhibit PAF-like activity contributing to clotting and/or exerting
cytotoxic effects on lung and heart cells; (3) possible mechanisms that may involve O3-induced
secretions of vasoconstrictive substances and/or effects on neuronal reflexes that may result in
increased arterial blood pressure  and/or altered electrophysiologic of heart rate or rhythm; (4)
associations between O3 and various cardiac physiologic endpoints suggesting a potential
relationship between O3 exposure and altered HRV, ventricular arrhythmias, and incidence of
MI; and (5) positive associations during the warm season only between ambient O3
concentrations and cardiovascular hospitalizations. While the only controlled human exposure
study that evaluated effects of O3 exposure on the cardiovascular system found no O3-induced
differences in ECG or blood pressure in healthy or hypertensive subjects, the study did report
overall increases in myocardial work and heart rate, and impairment in pulmonary gas exchange.
                                          3-83

-------
Table 3-4. Summary of Ozone-Induced Respiratory Health Effects from Clinical Studies

Health Effect
Pulmonary
Function
Decrements
Increased
Respiratory
Symptoms
Airway
Responsiveness
Respiratory
Inflammation
Changes in Host
Defenses
Decreased Exercise
Performance
Exercise Level
Moderate
Moderate
Moderate
Moderate
Competitive
Very Heavy
Heavy
Moderate
Light
At rest
Moderate
Moderate
Very Heavy
Moderate
Very Heavy
At rest
Moderate
Very Heavy
Moderate
Competitive
Prolonged
Exposure
6.6 hr
6.6 hr
4.6 hr
3.0hr
6.6 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
0.06 ppm
0.08 ppm
0.10 ppm
0.12 ppm
0.12-0. 14 ppm
0.16 ppm
0.1 8 ppm
0.30 ppm
0.37 ppm
0.50 ppm
0.06 ppm
0.08 ppm
0.12 ppm
0.08 ppm
0.1 8 ppm
0.40 ppm
0.08 ppm
0.20 ppm
0.08 ppm
0.1 8 ppm
       Also, animal toxicological studies have reported O3-induced decreases in heart rate, mean
arterial pressure and core temperature. Overall, the 2006 CD (p. 8-77) concludes that this
generally limited body of evidence is highly suggestive that Os directly and/or indirectly
contributes to cardiovascular-related morbidity, but much remains to be done to more fully
substantiate links between short-term ambient Os exposures and adverse cardiovascular effects.

       3.7.3  Mortality-Related Effects of Short-term Exposures to Ozone
       The  1996 CD concluded that an association between daily mortality and Os concentration
for areas with high Os levels (e.g., Los Angeles) was suggested.  However, due to a very limited
number of studies available at that time, there was insufficient evidence to conclude that the
observed association was likely causal.  Since 1996, new data are available from large multicity
       ' Information contained in this table is based on scientific data assessed in Chapters 6 and 8 of the 2006
CD.
                                           3-84

-------
studies conducted in the U.S. and Europe, and several single-city studies conducted all over the
world, as well as from several meta-analyses that have combined information from multiple
studies.  The majority of these studies suggest an elevated risk of total nonaccidental mortality
associated with acute exposure to 63, especially in the summer or warm season when 63 levels
are typically high, with somewhat larger effect estimate sizes for associations with
cardiovascular mortality (2006 CD, p. 7-175). The 2006 CD finds that the results from U.S.
multicity time-series studies provide the strongest evidence to-date for associations between
short-term Os exposure and mortality.  These studies, along with recent meta-analyses, showed
consistent effect estimates that are unlikely to be confounded by PM, though the 2006 CD
observes that future work is needed to better understand the influence of model specifications on
the effect estimates (2006 CD, p. 7-175). For cardiovascular mortality, the 2006 CD reports that
effect estimates are consistently  positive, falling in the range of 1 to 8% increases  per 40 ppb in
1-hr max O3 (2006 CD, p. 7-107).  Overall, the 2006 CD concludes that the majority of these
findings suggest an elevated risk of all-cause  mortality associated with short-term Os exposure,
especially in the summer or warm season when 63 levels are typically high.  Slightly greater
effects were observed for cardiovascular mortality (2006 CD, p. 7-175).

       3.7.4   Health Effects of Repeated Short-term Exposures to Ozone
       The 1996 CD drew several conclusions regarding repeated short-term Os exposures (2006
CD, p. 8-15). Partial or complete attenuation is observed for some of the (Vinduced responses
after more than 2 days of exposure. After 5 days of exposure, lung function changes return to
control levels with the greatest changes usually occurring on the second day, but the attenuation
was reversed  after 7 to 10 days without 63 exposure.  Most inflammatory markers (e.g., PMN
influx) attenuate after 5 days of exposure, but markers of cell damage (e.g., LDH enzyme
activity) do not attenuate and continue to increase. Recovery of some inflammatory markers
occurred a week to 10 days after exposure ceased, but some responses were not normal after 20
days.  Animal studies suggest underlying cell damage continues throughout the attenuation
process. Also, attenuation may alter normal  distribution of 63 within the lungs, allowing more
Os to reach sensitive regions, possibly affecting lung defenses.  Newer studies assessed in the
2006 CD (p. 8-74 and 8-75) supported all of these conclusions in addition to which it was
concluded that repeated daily, multi-hour exposure to lower concentrations of 63 (0.125 ppm for
4 days) causes an increased response to bronchial  allergen  challenge in subjects with preexisting
allergic airway disease, with or without asthma. In these subjects,  changes in airway
responsiveness after O3 exposure appear to be resolved more slowly than changes  in FEV! or
respiratory symptoms.
                                          3-85

-------
       3.7.5   Confidence in Various Health Outcomes Associated with Short-term
              Exposures to Ozone
       In characterizing the extent to which relationships between the various health outcomes
discussed above and short-term exposures to ambient 63 are likely causal, we note that several
different factors have informed the judgments made in the CD and here.  These factors include
the nature of the evidence (i.e., controlled human exposure, epidemiological, and/or toxicological
studies) and the weight of evidence, including such considerations as biological plausibility,
coherence of evidence, strength of association, and consistency of evidence.
       In assessing the health effects data base for 63, it is clear that human studies provide the
most directly applicable information because they are not limited by the uncertainties of
dosimetry differences and species sensitivity differences, which would need to be addressed in
extrapolating animal toxicology data to human health effects.  Controlled human exposure
studies provide dat
       a with the highest level of confidence since they provide human effects data under closely
monitored conditions and can provide clear exposure-response relationships. Epidemiological
data provide evidence of associations between ambient Os levels and more serious acute and
chronic health effects (e.g., hospital admissions and mortality) that cannot be assessed in
controlled human exposure studies. For these studies the degree of uncertainty regarding
potential confounding variables (e.g., other pollutants, temperature) and other factors affects the
level of confidence that the health effects being investigated are attributable to 63 exposures,
alone and in combination with other copollutants.
       In using a weight of evidence approach to inform judgments  about the degree of
confidence that various health outcomes are likely to be caused by exposure to 63, confidence
increases as the number of studies and other factors, such as strength, consistency, and coherence
of evidence, consistently reporting a particular health endpoint grows. For example, there is a
very high level of confidence that O3 induces lung function decrements in healthy adults and
children due in part to the dozens of studies consistently showing that these effects were
observed. As noted above, the 2006 CD (p.  8-74) states that studies  provide clear evidence of
causality for associations between short-term Os exposures and statistically significant declines
in lung function in children, asthmatics and adults who exercise outdoors. An increase in
respiratory  symptoms (e.g., cough, shortness of breath) has been observed in controlled human
exposure studies of short-term Os exposures, and significant associations between ambient Os
exposures and a wide variety of symptoms have been reported in epidemiology studies (2006
CD, p. 8-75). Aggregate population time-series studies showing robust associations with
respiratory  hospital admissions and ED visits are strongly supported by human clinical, animal
toxicologic, and epidemiologic evidence for lung function decrements, respiratory symptoms,
                                          3-86

-------
airway inflammation, and airway hyperreactivity. Taken together, the 2006 CD (p. 8-77)
concludes that the overall evidence supports the inference of a causal relationship between acute
ambient Os exposures and increased respiratory morbidity outcomes resulting in increased
asthma ED visits and respiratory hospitalizations during the warm season.  Recent epidemiologic
evidence has been characterized in the CD (p. 8-78) as highly suggestive that O3 directly or
indirectly contributes to non-accidental and cardiopulmonary-related mortality.
       As discussed above in section 3.5 and in section 8.6 of the 2006 CD, conclusions
regarding biological plausibility, consistency, and coherence of evidence of Os-related health
effects are drawn from the integration of epidemiological studies with mechanistic information
from controlled human exposure studies and  animal toxicological studies.  This  type of
mechanistic linkage has been firmly established for several respiratory endpoints (e.g., lung
function decrements, lung inflammation) but remains far more equivocal for cardiovascular
endpoints (e.g., cardiovascular-related hospital  admissions). Finally, for epidemiological studies,
strength of association refers to the magnitude of the association and its statistical strength,
which includes assessment of both effects estimate size and precision (section 3.4.1).  In general,
when associations yield large relative risk estimates,  it is less likely that the association could be
completely accounted for by a potential confounder or some other bias. Consistency refers to the
persistent finding of an association between exposure and outcome in multiple studies of
adequate power in different persons, places, circumstances and times (section 3.4.3).  For
example, the magnitude of effect estimates is relatively consistent across recent  studies showing
association between short-term, but not long-term, 63 exposure and mortality.
       Figure 3-5 summarizes our judgments for the various health outcomes discussed above
concerning the extent to which relationships between various health outcomes and ambient 63
exposures are likely causal. These judgments are informed by the conclusions and discussion in
the CD and in earlier sections of this chapter, reflecting the nature of the evidence and overall
weight of the evidence, and are taken into consideration in our quantitative risk assessment,
presented below in Chapter 5.

       3.7.6  Health Effects of Long-term Exposures to Ozone
       In the 1996 CD, available data, primarily from animal toxicology studies, indicated that
exposure to 63 for periods of months to years causes structural changes in several regions of the
respiratory tract (2006 CD, p. 8-79). Effects  may be  of greatest importance in the CAR, where
the gas exchange region and conducting airways meet. This region of the lungs  is typically
affected in most human airway diseases.  However, data from epidemiological are limited or
inconclusive, and data from clinical  studies are lacking.  Most information on chronic Os effects
in the distal lungs continues to come from animal toxicology studies.
                                          3-87

-------
Figure 3-5.  Qualitative Characterization of Ozone-Related Health Effect Outcomes
         Characterization
 Overall Confidence in Causal Relationship With Ambient
 Ozone
            Causal
           Suggestive
-Lung function decrements in healthy children
-Lung function decrements in asthmatic children
-Lung function decrements in healthy adults
-Respiratory symptoms in asthmatic children
-Respiratory symptoms in healthy adults
-Increased lung inflammation
-Aggravation of asthma (i.e., increased medication usage,
increased asthma attacks)
                             -Respiratory-related hospital admissions
                             -Respiratory related emergency department visits
                             -Respiratory-related doctors visits
                             -Increased school absences
                             -Respiratory-related mortality during the O3 season
                             -Cardiorespiratory-related mortality during the O3 season
                             -Total nonaccidental mortality during the O3 season
-Cardiovascular-related hospital admissions
                                        3-S

-------
What had been viewed previously as an apparent lack of reversibility of Os-induced effects with
a clean air recovery period has been investigated since 1996 with animal toxicology studies using
exposure regimens simulating a seasonal exposure pattern.  One long-term study exposed rhesus
monkeys to a simulated seasonal 63 pattern (0.5 ppm 63 8hr/day for 5 days, every 14 days for 11
episodes) and reported:  (1) remodeling in the distal airways; (2) abnormalities in tracheal
basement membrane; (3) eosinophil accumulation in conducting airways;  and (4) decrements in
airway innervation.  These findings support and advance the earlier information suggestive of
injury and repair processes which are caused by seasonal Os exposures (2006 CD, p.8-79).
Epidemiological studies investigating chronic effects in humans following long-term exposures
to 63 have provided only limited suggestive evidence.  Further investigation will be necessary
before we are able to draw firmer conclusions about chronic health effects of Os in human
populations.

       3.7.7  Health Effects of Pollutant Mixtures Containing Ozone
       In the 1996 CD, it was recognized that coexposure to O3 and other pollutants, such as
NO2, SO2, H2SO4, HNO3, or CO, showed additive response  for lung spirometry or respiratory
symptoms (2006 CD, p. 8-82). Since 1996, most animal toxicology  studies investigating 63 in a
mixture with NO2 and H2SO4 have shown that effects can be additive, synergistic, or even
antagonistic, depending on the exposure regimen and the endpoint studied. Ozone has served for
a long time as a surrogate or indicator for the overall photochemical  oxidant mix. It is well
recognized that the observed effects may be due to components of that mix alone or in
combination with Os and other gases and PM in the ambient air. Although the issue of exposure
to copollutants was previously described as poorly understood, especially  with regard to chronic
effects, newer information from human and animal studies of binary mixtures containing Os
suggest potential interactions depending on the exposure regimen and pollutant mix (CD, p. 8-
82).  Examples of this newer information include:  (1) continuous exposure to SO2 and NO2
increased inhaled Os bolus absorption, while continuous exposure to Os decreased Os bolus
absorption; (2) asthmatics exhibited enhanced airway reactivity to house dust mite allergen
following exposures to O3, NO2 and the combination of the  two gases; however, spirometric
response was impaired only by O3 and O3+ NO2 at higher concentrations; and (3) animal
toxicology studies with O3 in mixture with NO2, formaldehyde, and PM demonstrated additive,
synergistic, or antagonistic effects depending on the exposure regimen and the endpoints
evaluated.
       One controlled-exposure study of children, designed to approximate conditions of an
epidemiological study by matching population and exposure atmosphere (0.1 ppm Os,  0.1 ppm
SO2, and 101  ug/m2 H2SO4), failed to support the findings of the epidemiological study.  This
                                          3-89

-------
demonstrates the difficulty of trying to link outcomes of epidemiological studies and controlled-
exposure studies with pollutant mixtures.

       3.7.8   Populations at Risk/Susceptibility Factors Associated with Ozone Exposure
       The 1996 CD (2006 CD, p. 8-80) identified several factors that may increase sensitivity
to O3 of population groups, including: (1) biological variation in responsiveness to O3; (2)
preexisting lung disease (e.g., asthma); (3) activity patterns (e.g., exertion levels); (4) personal
exposure history (e.g., time spent indoors v. outdoors); and (5) personal factors (e.g., age,
nutritional status, gender, smoking history, ethnicity). Based on the information assessed in the
1996 CD (2006 CD, p. 8-80), population groups that demonstrated increased responsiveness to
ambient concentrations of O3 consisted of exercising, healthy and asthmatic individuals,
including children, adolescents, and adults.  Since 1996, evidence from controlled-exposure
human and animal studies, as well as from epidemiological studies, has provided further support
for these and other susceptibility factors  and populations at risk. For example, controlled-
exposure human studies continue to show differential biological response to O3 based on
physical activity (exertion) and age. These studies demonstrate a large variation in sensitivity
and responsiveness to O3, although specific factors that  contribute to this intersubject variability
are yet to be identified.  Associations of  increased summertime hospital admissions for asthma
with ambient O3 levels suggest that individuals with these respiratory diseases are populations  at
risk to O3 exposure effects. Also, based  on O3-induced  differential response in lung
inflammation and airway responsiveness, asthmatic adults and children appear to have
potentially increased susceptibility to O3. There is limited evidence from epidemiologic studies
and no evidence from controlled-exposure human studies which suggest that individuals with
COPD are more sensitive to  health effects of O3.
       There is some animal toxicology and limited epidemiologic evidence which has
demonstrated the importance of genetic background in O3 susceptibility.  Genetic and molecular
characterization studies of experimental  animals have identified genetic loci responsible for both
sensitivity and resistance.
       Taking all of this information into account, the CD (p. 8-80 to 8-81) concludes that
exercising (moderate to high physical exertion) healthy  and asthmatic adults, adolescents, and
children appear to exhibit increased responsiveness to ambient O3 levels and continue to be
considered at increased risk of O3-induced health effects.  Also, any individual with respiratory
or cardiovascular disease or  any healthy  individual who is engaged in vigorous physical activity
outdoors during periods when 63 levels  are high (e.g., active outdoor children) is potentially at
increased risk to (Vinduced health effects.  In addition, healthy individuals and those with
                                           3-90

-------
cardiorespiratory impairment (e.g., those with asthma or cardiovascular disease) who are
"hyperresponsive" to Os exposure (i.e., exhibit much higher than normal lung function
decrements and/or respiratory symptoms) would be considered at greater risk to 63 exposure.
Finally, individuals who are more likely to be exposed to air pollution while engaged in physical
activity (e.g., outdoor workers) and those with genetic polymorphisms for antioxidant enzymes
and inflammatory genes may be at heightened risk of effects of Os (2006 CD, p. 8-81).
                                          3-91

-------
References

Abbey, D. E.; Nishino, N.; McDonnell, W. F.; Burchette, R. I; Knutsen, S. F.; Beeson, W. L.; Yang, J. X. (1999)
      Long-term inhalable particles and other air pollutants related to mortality in nonsmokers. Am. J. Respir. Crit.
      Care Med. 159: 373-382.
Adams, W. C. (2000a) Feasibility study of prolonged ozone inhalation exposure via face mask. Inhalation Toxicol.
      12:299-313.
Adams, W. C. (2000b) Ozone dose-response effects of varied equivalent minute ventilation rates. J. Exposure Anal.
      Environ. Epidemiol. 10: 217-226.
Adams, W. C. (2002) Comparison of chamber and face-mask 6.6-hr exposures to ozone on pulmonary function and
      symptoms responses. Inhalation Toxicol. 14: 745-764.
Adams, W. C. (2003a) Comparison of chamber and face mask 6.6-hr exposure to 0.08 ppm ozone via square-wave
      and triangular profiles on pulmonary responses. Inhalation Toxicol. 15: 265-281.
Adams, W. C. (2003b) Relation of pulmonary responses induced by 6.6 h exposures to 0.08 ppm ozone and 2-h
      exposures to 0.30 ppm via chamber and face-mask inhalation. Inhalation Toxicol. 15: 745-759.
Adams, W. C. (2006) Comparison of chamber 6.6 h exposures to 0.04-0.08 ppm ozone via square-wave and
      triangular profiles on pulmonary responses. Inhalation Toxicol. 18: 127-136.
Adams, W. C.; Schelegle,  E. S. (1983) Ozone and high ventilation effects on pulmonary function and endurance
      performance. J. Appl. Physiol.: Respir. Environ. Exercise Physiol. 55: 805-812.
Adams, W. C.; Ollison, W. M. (1997) Effects of prolonged simulated ambient ozone dosing patterns on human
      pulmonary function and symptomatology. Presented at: 90th annual meeting of the Air & Waste Management
      Association; June; Toronto, Ontario, Canada. Pittsburgh, PA: Air & Waste Management Association; paper
      no. 97-MP9.02.
Adams, P. F.; Hendershot, G. E.; Marano, M. A. (1999) Current estimates from the National Health Interview
      Survey, 1996. Hyattsville, MD: U.S. Department of Health and Human Services, Public Health Service,
      National Center for Health Statistics; DHHS publication no. (PHS) 99-1528. (Vital and health statistics:
      v.  10, data from the National Health Survey, no. 200). Available:
      http://www.cdc.gov/nchs/products/pubs/pubd/series/srlO/pre-200/pre-200.htm [12 March, 2001].
Alexis, N.; Urch, B.; Tarlo, S.; Corey, P.; Pengelly, D.; O'Byrne, P.; Silverman, F. (2000)  Cyclooxygenase
      metabolites play a different role in ozone-induced pulmonary function decline in asthmatics compared to
      normals. Inhalation Toxicol. 12: 1205-1224.
American Academy of Pediatrics, Committee on Environmental Health. (2004) Ambient air pollution: health
      hazards to children. Pediatrics 114: 1699-1707.
American Thoracic Society (1985) Guidelines as to what constitutes an adverse respiratory health effect, with
      special reference to epidemiological studies of air pollution.  Am. Rev. Respir. Dis. 131:  666-668.
American Thoracic Society (1991) Lung function testing: selection of reference values and interpreatative
      strategies. Am. Rev. Respir. Dis. 144: 1202-1218.
American Thoracic Society (2000) What constitutes an adverse health effect of air pollution? Am. J. Respir. Crit.
      Care Med. 161:665-673.
Anderson, H. R.; Ponce de Leon, A.; Bland, J. M.; Bower, J. S.; Strachan, D. P.  (1996) Air pollution and daily
      mortality in London:  1987-92. Br. Med. J. 312: 665-669.
Anderson, H. R.; Spix, C.; Medina, S.; Schouten, J. P.; Castellsague, J.; Rossi, G.; Zmirou, D.; Touloumi,  G.;
      Wojtyniak, B.; Ponka, A.; Bacharova, L.; Schwartz, J.; Katsouyanni, K. (1997) Air pollution and daily
      admissions for chronic obstructive pulmonary disease in 6 European cities: results from the APHEA project.
      Eur. Respir. J.  10: 1064-1071.
Anderson, H. R.; Ponce de Leon, A.; Bland, J. M.; Bower, J. S.; Emberlin, J.; Strachen, D. P. (1998) Air pollution,
      pollens, and daily admissions for asthma in London 1987-92. Thorax 53:  842-848.
Arito, H.; Takahashi, M.; Iwasaki, T.; Uchiyama, I. (1997) Age-related changes in ventilatory and heart rate
      responses to acute ozone exposure in the conscious rat. Ind. Health 35: 78-86.
Aris, R. M.; Christian, D.; Hearne, P. Q.; Kerr, K.; Finkbeiner, W. E.; Balmes, J. R. (1993) Ozone-induced airway
      inflammation in human subjects as determined by airway lavage and biopsy. Am. Rev. Respir. Dis.
      148: 1363-1372.
Aris, R. M.; Tager, I.; Christian, D.; Kelly, T.; Balmes, J. R. (1995) Methacholine responsiveness is not associated
      withOs-induced decreases  inFEY:. Chest 107: 621-628.
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,
      K. B.; D'Arcy, J. B.; Gibbons, D.; Higgins, I. T. T. (1990) Ozone exposure and lung function in children
                                                  3-92

-------
      attending a southern California summer camp. Presented at: 83rd annual meeting and exhibition of the Air &
      Waste Management Association; June; Pittsburgh, PA. Pittsburgh, PA: Air & Waste Management
      Association; paper no. 90-150.3.
Avol, E. L.; Navidi, W. C.; Rappaport, E. B.; Peters, J. M. (1998) Acute effects of ambient ozone on asthmatic,
      wheezy, and healthy children. Cambridge, MA:  Health Effects Institute; research report no. 82.
Avol, E. L.; Gauderman, W. J.; Tan, S. M.; London, S. J.; Peters, J. M. (2001) Respiratory effects of relocating to
      areas of differing air pollution levels. Am. J. Respir. Crit. Care Med. 164: 2067-2072.
Bates, D. V.; Baker-Anderson, M.; Sizto, R. (1990) Asthma attack periodicity: a study of hospital emergency visits
      in Vancouver. Environ. Res.  51: 51-70.
Bates, D. V.; Sizto, R. (1983) Relationship between air pollutant levels and hospital admissions in Southern Ontario.
      Can. J. Public Health 74: 117-122.
Bates, D. V.; Sizto, R. (1987) Air pollution and hospital admissions in southern Ontario: the acid summer haze
      effect. Environ. Res. 43: 317-331.
Bates, D. V.; Sizto, R. (1989) The Ontario Air Pollution study: identification of the causative agent. Environ. Health
      Perspect. 79: 69-72.
Bayram, H.; Rusznak, C.; Khair, O. A.; Sapsford, R. J.; Abdelaziz, M. M. (2002) Effect of ozone and nitrogen
      dioxide on the permeability of bronchial epithelial cell cultures of non-asthmatic and asthmatic subjects.
      Clin. Exp. Allergy 32: 1285-1292.
Beeson, W. L.; Abbey, D. E.; Knutsen, S. F. (1998) Long-term concentrations of ambient air pollutants and incident
      lung cancer in California adults: results from the AHSMOG study. Environ. Health Perspect. 106: 813-823.
Bell, M. L.; McDermott, A.; Zeger,  S. L.; Samet, J. M.; Dominici, F. (2004) Ozone and short-term mortality in
      95 US urban communities, 1987-2000. JAMA J. Am. Med. Assoc. 292:  2372-2378.
Bell, M. L.; Dominici, F.; Samet, J.  M. (2005) A meta-analysis of time-series studies of ozone and mortality with
      comparison to the national morbidity, mortality, and air pollution study. Epidemiology 16: 436-445.
Bell, M. L.; Peng, R. D.; Dominici,  F. (2006) The exposure-response curve for ozone and risk of mortality and the
      adequacy of current ozone regulations. Environ. Health Perspect.: in press, doi:10.1289/ehp.8816.
      Available:  http://dx.doi.org/  [23 January, 2006].
Bell, M. L. (2006) Community-specific maximum likelihood estimates of O3-related excess risk in mortality for the
      NMMAPS U.S. 95 communities study  [personal communication with attachments to Jee Young Kim].
      New Haven, CT: Yale University School of Forestry and Environmental Studies; January 6.
Blomberg, A.; Mudway, I. S.; Nordenhall, C.; Hedenstrom, H.; Kelly, F. J.; Frew, A. J.; Holgate, S. T.;
      Sandstrom, T. (1999) Ozone-induced lung function decrements do not correlate with early airway
      inflammatory or antioxidant responses. Eur. Respir. J. 13: 1418-1428.
Blomberg, A.; Mudway, I.; Svensson, M.; Hagenbjork-Gustafsson, A.; Thomasson, L.; Helleday, R.; Dumont, X.;
      Forsberg, B.; Nordberg, G.; Bernard, A. (2003) Clara cell protein as a biomarker for ozone-induced lung
      injury in humans. Eur. Respir. J. 22: 883-888.
Borja-Aburto, V. H.; Loomis, D. P.; Bangdiwala, S. I.; Shy, C. M.; Rascon-Pacheco, R. A. (1997) Ozone, suspended
      particulates, and daily mortality in Mexico City. Am. J. Epidemiol. 145: 258-268.
Bosson, J.; Stenfors, N.; Bucht, A.; Helleday, R.; Pourazar, J.; Holgate, S. T.; Kelly, F. J.; Sandstrom, T.; Wilson,
      S.; Frew, A. J.; Blomberg, A. (2003) Ozone-induced bronchial epithelial cytokine expression differs between
      healthy and asthmatic subjects. Clin. Exp. Allergy 33: 777-782.
Brauer, M.; Blair, J.; Vedal, S. (1996) Effect of ambient ozone exposure on lung function in farm workers. Am. J.
      Respir. Crit. Care Med. 154:  981-987.
Brauer, M; Brumm, J.; Vedal, S.; Petkau, A.J. (2002) Exposure  misclassification and threshold concentrations in
      time-series analyses of air pollution health effects. Risk Anal.  22: 1183-1193.
Bremner, S. A.; Anderson, H. R.; Atkinson, R. W.; McMichael, A. J.; Strachan, D. P.; Bland, J. M.; Bower, J. S.
      (1999) Short term associations between outdoor air pollution and mortality in London 1992-4. Occup.
      Environ. Med. 56: 237-244.
Brook, R.  D.; Brook, J. R.; Urch, B.; Vincent, R.; Rajagopalan,  S.; Silverman,  F. (2002) Inhalation of fine
      paniculate air pollution and ozone causes acute arterial vasoconstriction in healthy adults. Circulation 105:
      1534-1536.
Brook, R.  D.; Franklin, B.; Cascio, W.; Hong, Y.; Howard, G.; Lipsett, M.; Luepker, R.; Mittleman, M.; Samet, J.;
      Smith, S. C., Jr.; Tager, I. (2004) Air pollution and cardiovascular disease. A statement for healthcare
      professionals from the Expert Panel on Population and Prevention Science of the American Heart
      Association. Circulation 109: 2655-2671.
                                                  3-93

-------
Burnett, R. T.; Dales, R. E.; Raizenne, M. E.; Krewski, D.; Summers, P. W.; Roberts, G. R.; Raad-Young, M.;
      Darin, T.; Brook, J. (1994) Effects of low ambient levels of ozone and sulfates on the frequency of respiratory
      admissions to Ontario hospitals. Environ. Res. 65: 172-194.
Burnett, R. T.; Brook, J. R.; Yung, W. T.; Dales, R. E.; Krewski, D. (1997a) Association between ozone and
      hospitalization for respiratory diseases in 16 Canadian cities. Environ. Res. 72: 24-31.
Burnett, R. T.; Brook, J. R.; Yung, W. T.; Dales, R. E.; Krewski, D. (1997b) Association between ozone and
      hospitalization for respiratory diseases in 16 Canadian cities. Environ. Res. 72:  24-31.
Burnett, R. T.; Smith-Doiron, M.; Stieb, D.; Cakmak, S.; Brook, J. R. (1999) Effects of paniculate and gaseous air
      pollution on cardiorespiratory hospitalizations. Arch. Environ. Health 54: 130-139.
Burnett, R. T.; Smith-Doiron, M.; Stieb, D.; Raizenne, M. E.; Brook, J. R.; Dales, R. E.; Leech, J. A.; Cakmak, S.;
      Krewski, D. (2001) Association between ozone and hospitalization for acute respiratory diseases in children
      less than 2 years of age. Am. J. Epidemiol.  153: 444-452.
Calderon-Garciduenas, L.; Rodriguez-Alcaraz, A.; Garcia, R.; Ramirez, L.; Barragan, G. (1995) Nasal inflammatory
      responses in children exposed to a polluted urban atmosphere. J. Toxicol. Environ. Health 45: 427-437.
Calderon-Garciduenas, L.; Osnaya, N.; Rodriguez-Alcaraz, A.; Villarreal-Calderon, A. (1997) DNA damage in
      nasal respiratory epithelium from children exposed to urban pollution. Environ. Mol.  Mutagen. 30: 11-20.
Calderon-Garciduenas, L.; Wen-Wang, L.; Zhang, Y.-J.; Rodriguez-Alcaraz, A.; Osnaya, N.; Villarreal-Calderon,
      A.; Santella, R. M. (1999) 8-hydroxy-2'-deoxyguanosine, a major mutagenic oxidative DNA lesion, and DNA
      strand breaks in nasal respiratory epithelium of children exposed to urban pollution. Environ. Health
      Perspect. 107: 469-474.
Calderon-Garciduenas, L.; Valencia-Salazar, G.; Rodriguez-Alcaraz, A.; Gambling, T. M.; Garcia, R.; Osnaya, N.;
      Villarreal-Calderon, A.; Devlin, R. B.; Carson, J. L. (2001) Ultrastructural nasal pathology in children
      chronically and sequentially exposed to  air pollutants. Am. J. Respir. Cell Mol. Biol.  24: 132-138.
Calderon-Garciduenas, L.; Mora-Tiscareno, A.; Fordham, L. A.; Valencia-Salazar, G.; Chung, C.  J.; Rodriguez-
      Alcaraz, A.; Paredes, R.; Variakojis, D.; Villarreal-Calderon, A.; Flores-Camacho, L.; Antunez-Solis, A.;
      Henriquez-Roldan, C.; Hazucha, M.  J. (2003) Respiratory damage in children exposed to urban pollution.
      Pediatr. Pulmonol. 36: 148-161.
Cassino, C.; Ito, K.; Bader, L; Ciotoli, C.; Thurston, G.; Reibman, J. (1999) Cigarette smoking and ozone-associated
      emergency department use for asthma by adults in New York City. Am. J. Respir. Crit. Care Med. 159: 1773-
      1779.
Centers for Disease Control and Prevention. (2004) The health consequences of smoking: a report of the Surgeon
      General. Atlanta,  GA: U.S. Department  of Health and Human Services, National Center for Chronic Disease
      Prevention and Health Promotion, Office on Smoking and Health. Available:
      http://www.cdc.gov/tobacco/sgr/sgr_2004/chapters.htm (18 August, 2004).
Chen, L.; Jennison, B. L.; Yang, W.; Omaye, S. T. (2000) Elementary school absenteeism and air pollution.
      Inhalation Toxicol. 12: 997-1016.
Chen, C.-Y.; Bonham, A. C.; Plopper, C. G.; Joad, J. P. (2003) Plasticity in respiratory motor control: selected
      contribution: neuroplasticity in nucleus tractus solitarius neurons following episodic ozone exposure in infant
      primates. J. Appl. Physiol. 94: 819-827.
Chock, D. P.; Winkler, S. L.; Chen, C. (2000) A study of the association between daily mortality and ambient air
      pollutant concentrations in Pittsburgh, Pennsylvania. J.  Air Waste Manage. Assoc. 50: 1481-1500.
Christian, D. L.; Chen, L. L.; Scannell, C. H.; Ferrando, R. E.; Welch, B. S.; Balmes, J. R. (1998) Ozone-induced
      inflammation is attenuated with multiday exposure. Am. J. Respir. Crit.  Care Med.  158:  532-537.
Cody, R. P.; Weisel, C. P.; Birnbaum, G.; Lioy, P. J. (1992) The effect of ozone associated with summertime
      photochemical smog on the frequency of asthma visits to hospital emergency departments. Environ. Res.
      58: 184-194.
Daly, C.; Fox, K.; Henein, M. (2002) Natriuretic peptides in the diagnosis of heart disease-first amongst equals?
      Int. J. Cardiol. 84: 107-113.
DeFrances, C. J.; Hall, M. J.; Podgornik, M. N. (2005) 2003 National Hospital Discharge Survey. Hyattsville, MD:
      National Center for Health Statistics; DHHS publication no. (PHS) 2004-1250. (Advance data from vital  and
      health statistics; no. 359). Available: http://www.cdc.gov/nchs/data/ad/ad359.pdf [3 August, 2005].
Delfmo, R. J.; Becklake, M. R.; Hartley, J.  A.  (1994) The relationship of urgent hospital admissions for respiratory
      illnesses to photochemical air pollution levels in Montreal. Environ. Res. 67:  1-19.
Delfino, R. J.; Coate, B.  D.; Zeiger, R. S.; Seltzer, J. M.; Street, D. H.; Koutrakis, P. (1996) Daily asthma severity in
      relation to personal ozone exposure and outdoor fungal spores. Am. J. Respir. Crit.  Care  Med. 154: 633-641.
                                                  3-94

-------
Derfino, R. I; Zeiger, R. S.; Seltzer, J. M; Street, D. H.; Matteucci, R. M.; Anderson, P. R.; Koutrakis, P. (1997a)
      The effect of outdoor fungal spore concentrations on daily asthma severity. Environ. Health Perspect.
      105: 622-635.
Delfino, R. J.; Murphy-Moulton, A. M.; Burnett, R. T.; Brook, J. R.; Becklake, M. R. (1997b) Effects of air
      pollution on emergency room visits for respiratory illnesses in Montreal, Quebec. Am. J. Respir. Crit. Care
      Med. 155: 568-576.
Delfino, R. J.; Zeiger, R. S.; Seltzer, J. M.; Street, D. H. (1998) Symptoms inpediatric asthmatics and air pollution:
      differences in effects by symptom severity, anti-inflammatory medication use and paniculate averaging time.
      Environ. Health Perspect. 106: 751-761.
Delfino, R. J.; Gone, H.; Linn, W.  S.; Pellizzari, E. D.; Hu, Y. (2003) Asthma symptoms in Hispanic children and
      daily ambient exposures to toxic and criteria air pollutants. Environ. Health Perspect. Ill: 647-656.
Desqueyroux, H.; Pujet, J.-C.; Prosper, M.; Squinazi, F.; Momas, I. (2002a) Short-term effects of low-level air
      pollution on respiratory health of adults suffering from moderate to severe asthma. Environ. Res. A 89: 29-
      37.
Desqueyroux, H.; Pujet, J.-C.; Prosper, M.; LeMoullec, Y.; Momas, I. (2002b) Effects of air pollution on adults
      with chronic obstructive pulmonary disease.  Arch. Environ. Health 57: 554-560.
Devlin, R. B.; Folinsbee, L. J.; Biscardi, F.; Hatch, G.; Becker, S.; Madden, M. C.; Robbins, M.;  Koren, H. S. (1997)
      Inflammation and cell damage induced by repeated exposure of humans to ozone. Inhalation Toxicol. 9: 211-
      235.
Devlin, R. B.; McDonnell, W. F.; Mann, R.; Becker, S.; House, D. E.; Schreinemachers, D.; Koren, H. S.  (1991)
      Exposure of humans to ambient levels of ozone for 6.6 hrs causes cellular and biochemical changes in the
      lung. Am. J. Respir. Cell Mol. Biol. 4: 72-81.
Devlin, R. B.; McDonnell, W. F.; Koren, H. S.; Becker, S. (1990) Prolonged exposure of humans to 0.10 and
      0.08 ppm ozone results in inflammation in the lung. Presented at: 83rd annual meeting of the Air & Waste
      Management Association; June; Pittsburgh, PA. Pittsburgh, PA: Air & Waste Management Association;
      paper no. 90-150.2.
Dey, A. N.; Bloom, B. (2005) Summary health statistics for U.S. children: National Health Interview Survey, 2003.
      Hyattsville, MD: U.S. Department of Health & Human Services, National Center for Health Statistics. (Vital
      and health statistics, series 10, no. 223). Available:  http://www.cdc.gov/nchs/data/series/sr_10/srlO_223.pdf
      [3 August, 2005].
Dietert, R. R.; Etzel, R. A.; Chen, D.; Halonen, M.; Holladay, S. D.; Jarabek, A. M.; Landreth, K.; Peden, D. B.;
      Pinkerton, K.; Smialowicz, R. J.; Zoetis, T. (2000) Workshop to identify critical window of exposure for
      children's health: immune and respiratory systems work group summary. Environ. Health Perspect. Suppl.
      108(3): 483-490.
Dimeo, M. J.; Glenn, M. G.; Holtzman, M. J.; Sheller, J. R.; Nadel, J. A.;  Boushey, H. A. (1981) Threshold
      concentration of ozone causing an increase in bronchial reactivity in humans and adaptation with repeated
      exposures. Am. Rev. Respir. Dis. 124: 245-248.
Dockery, D. W.; Schwartz, J.; Spengler, J. D. (1992) Air pollution and daily mortality: associations with
      particulates and acid aerosols. Environ. Res.  59:  362-373.
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.
      (1993) An association between air pollution and mortality in six U.S. cities. N. Engl. J. Med. 329: 1753-1759.
Dockery, D. W.; Luttmann-Gibson, H.; Rich, D. Q.; Link, M. S.; Mittleman, M. A.; Gold,  D. R.; Koutrakis, P.;
      Schwartz, J. D.; Verrier, R. L. (2005) Association of air pollution with increased incidence of ventricular
      tachyarrhythmias recorded by implanted cardioverter defibrillators. Environ. Health Perspect. 113:  670-674.
Dominici, F.; McDermott, A.; Daniels, M.; Zeger, S. L.; Samet, J. M. (2003) Mortality among residents of 90 cities.
      In: Revised analyses of time-series studies of air pollution and health. Special report. Boston, MA: Health
      Effects Institute; pp. 9-24. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdf [12 May,  2004].
Dormans, J. A. M. A.; Van Bree, L.; Boere, A. J. F.; Marra, M.; Rombout, P. J. A. (1999) Interspecies  differences in
      time course of pulmonary toxicity following repeated exposure to ozone. Inhalation Toxicol. 11: 309-329.
Evans, M. J.; Fanucchi, M. V; Baker, G. L.; Van Winkle, L. S.; Pantle, L. M.; Nishio, S. J.; Schelegle, E. S.;
      Gershwhin, L. J.; Miller, L. A.; Hyde, D. M.; Sannes, P. L.; Plopper, C. G. (2003) Atypical development of
      the tracheal basement  membrane zone of infant rhesus monkeys exposed to ozone and allergen.  Am. J.
      Physiol. 285:L931-L939.
Fairley, D. (1999) Daily mortality and air pollution in Santa Clara County, California: 1989-1996. Environ. Health
      Perspect. 107: 637-641.
                                                  3-95

-------
Fairley, D. (2003) Mortality and air pollution for Santa Clara County, California, 1989-1996. In: Revised analyses
      of time-series studies of air pollution and health. Special report. Boston, MA: Health Effects Institute;
      pp. 97-106. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdf [18 October, 2004].
Fanucchi, M. V.; Hotchkiss, J. A.; Harkema, J. R. (1998) Endotoxin potentiates ozone-induced mucous cell
      metaplasia in rat nasal epithelium. Toxicol. Appl. Pharmacol. 152: 1-9.
Folinsbee, L. J.; Silverman, F.; Shephard, R. J. (1977) Decrease of maximum work performance following ozone
      exposure. J. Appl. Physiol.: Respir. Environ. Exercise Physiol. 42: 531-536.
Folinsbee, L. J.; McDonnell, W. F.; Horstman, D. H. (1988) Pulmonary function and symptom responses after
      6.6-hr exposure to 0.12 ppm ozone with moderate exercise. JAPCA 38: 28-35.
Folinsbee, L. J.; Horstman, D. H.; Kehrl, H. R.; Harder, S.; Abdul-Salaam, S.; Ives, P. J. (1994) Respiratory
      responses to repeated prolonged exposure to 0.12 ppm ozone. Am. J. Respir. Crit. Care Med. 149: 98-105.
Folinsbee, L. J.; Hazucha, M. J. (2000) Time course of response to ozone exposure in healthy adult females.
      Inhalation Toxicol. 12: 151-167.
Foster, W. M.; Silver, J. A.; Groth, M. L. (1993) Exposure to ozone alters regional function and particle dosimetry
      in the human lung. J. Appl. Physiol. 75: 1938-1945.
Foster, W. M.; Weinmann, G. G.; Menkes, E.; Macri, K. (1997) Acute exposure of humans to ozone impairs small
      airway function. Ann. Occup. Hyg. 41(suppl. 1): 659-666.
Foxcroft, W. J.; Adams, W.  C. (1986) Effects of ozone exposure on four consecutive days on work performance and
      yO2max. J. Appl. Physiol. 61: 960-966.
Friedman, M. S.; Powell, K. E.; Hutwagner, L.; Graham, L. M.; Teague, W. G. (2001) Impact of changes in
      transportation and commuting behaviors during the 1996 summer Olympic games in Atlanta on air quality
      and childhood asthma. JAMA J. Am. Med. Assoc. 285:  897-905.
Frischer, T.; Studnicka, M.;  Gartner, C.; Tauber, E.; Horak, F.; Veiter, A.; Spengler, J.; Ktihr, J.; Urbanek, R. (1999)
      Lung function growth and ambient ozone: a three-year population study in school children. Am. J. Respir.
      Crit. Care Med. 160:  390-396.
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.
Fung, K. Y.; Luginaah, L; Gorey, K. M.; Webster, G. (2005) Air pollution and daily hospital admissions for
      cardiovascular diseases in Windsor, Ontario. Can. J. Public Health 96: 29-33.
Galizia, J.L..; 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: 91st 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.; Belanger, 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.; VanDer Zee, S.  C.; Van Wijnen, J. H.; Van Steen, C. J.; Brunekreef, B. (1997) Acute  effects of
      summer 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. T.; Peters, J. M. (2001) The effects of ambient air pollution on
      school absenteeism due to respiratory illnesses. Epidemiology 12:  43-54.
                                                  3-96

-------
Gold, D. R.; Litonjua, A.; Schwartz, I; Lovett, E.; Larson, A.; Nearing, B.; Allen, G.; Verrier, M.; Cherry, R.;
      Verrier, R. (2000) Ambient pollution and heart rate variability. Circulation 101: 1267-1273.
Goldberg, M. S.; Burnett, R. T.; Brook, I; Bailar, J. C., III;Valois, M.-F.; Vincent, R. (2001) Associations between
      daily cause-specific mortality and concentrations of ground-level ozone in Montreal, Quebec. Am. J.
      Epidemiol.  154: 817-826.
Goldberg, M. S.; Burnett, R. T.; Valois, M.-F.; Flegel, K.; Bailar, J. C., Ill; Brook, J.; Vincent, R.; Radon, K. (2003)
      Associations between ambient air pollution and daily mortality among persons with congestive heart failure.
      Environ. Res. 91:8-20.
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.
Gouveia, N.; Fletcher, T. (2000) 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
      andbronchoalveolarlavage. 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, J.; 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.
Gwynn, R. C.; Burnett, R. T.; Thurston, G.  D. (2000) A time-series analysis of acidic paniculate matter and daily
      mortality and morbidity in the Buffalo, New York, region. Environ. Health Perspect. 108: 125-133.
Gwynn, R. C.; Thurston, G. D. (2001) The  burden of air pollution: impacts among racial minorities. Environ. Health
      Perspect. Suppl. 109(4): 501-506.
Harkema, J. R.; Plopper, C. G.; Hyde, D. M.; St. George, J. A.; Wilson, D. W.; Dungworth, D. L. (1993) Response
      of macaque bronchiolar epithelium to ambient concentrations of ozone. Am J. Pathol. 143: 857-66.
Harkema, J. R.; Wagner, J. G. (2005) Epithelial and inflammatory responses in the airways of laboratory rats
      coexposed to ozone and biogenic substances: enhancement of toxicant-induced airway injury. Exp. Toxicol.
      Pathol. 57(suppl. 1): 129-141.
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.
Hernandez-Garduno, E.; Perez-Neria, J.;  Paccagnella, A. M.; Pina-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.
Henderson, R. (2006a) Letter from CASAC Chairman Rogene Henderson to EPA Administrator Stephen Johnson,
      February 16, 2006, EPA-CASAC-06-003.
Henderson, R. (2006b) Letter from CASAC Chairman Rogene Henderson to EPA Administrator Stephen Johnson,
      June 5, 2006, EPA-CASAC-06-007.
                                                 3-97

-------
Henderson, R. (2006c) Letter from CASAC Chairman Rogene Henderson to EPA Administrator Stephen Johnson,
      October 24, 2006, EPA-CASAC-07-001.
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.
Hiltermann, T. J. N.; Stolk, J.; Van derZee, S. C.; Brunekreef, B.; De Bruijne, C. R.; Fischer, P. H.; Ameling, C. B.;
      Sterk, P. J.; Hiemstra, P. S.; VanBree, L. (1998) Asthma severity and susceptibility to air pollution. Eur.
      Respir. J. 11:686-693.
Hiltermann, J. T. N.; Lapperre, T. S.; VanBree, 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.
Hoek, G.; Brunekreef, B.; Verhoeff, A.;  Van Wijnen, J.; Fischer, P. (2000) Daily mortality and air pollution in the
      Netherlands. J. Air Waste Manage. Assoc. 50:  1380-1389.
Hoek, G.; Brunekreef, B.; Fischer, P.; Van Wijnen, J. (2001) The association between air pollution and heart failure,
      arrhythmia, embolism, thrombosis, and other cardiovascular causes of death in a time series study.
      Epidemiology  12: 355-357.
Hoek, G. (2003) Daily mortality and air pollution in The Netherlands. In: Revised analyses of time-series studies of
      air pollution and health. Special report. Boston, MA: Health Effects Institute; pp.  133-142. Available:
      http://www.healtheffects.org/Pubs/TimeSeries.pdf [12 May, 2004].
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.; Torres, R. A.;  Timm, P.; Mticke, M.; Richter, K.; Koschyk, S.; Magnussen, H. (1999) Ozone-induced
      airway inflammatory changes differ between individuals and are reproducible. Am. J. Respir. Crit. Care Med.
      159: 776-784.
Holz, O.; Mticke, M.; Paasch, K.; Bohme, S.; Timm, P.; Richter, K.; Magnussen, H.; Torres, R. A. (2002) Repeated
      ozone exposures enhance bronchial allergen responses in subjects with rhinitis or asthma. Clin. Exp. Allergy.
      32: 681-689.
Hoppe, P.; Praml, G.; Rabe, G.; Lindner, J.; Fruhmann, G.; Kessel, R. (1995) Environmental ozone field study on
      pulmonary and subjective responses of assumed risk groups. Environ. Res. 71: 109-121.
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.
Horak, F., Jr.; Studnicka, M.; Gartner, C.; Spengler, J. D.; Tauber, E.; Urbanek, R.; Veiter, A.; Frischer, T. (2002a)
      Paniculate 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)
      Paniculate 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.  CellMol. 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, I.; Campos, E.; Villaire, J.; Tellez-Rojo, M. M.; Romieu, I. (1999) Association between levels
      of fine paniculate and emergency visits for pneumonia and other respiratory illnesses among children in
      Santiago, Chile. J. Air Waste Manage.  Assoc. 49: 154-163.
                                                  3-98

-------
Ito, K.; Thurston, G. D. (1996) Daily PMio/mortality associations: an investigation of at-risk subpopulations.
      J. Exposure Anal. Environ. Epidemiol. 6: 79-95.
Ito, K. (2003) Associations of paniculate 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].
Ito, K.; De Leon, S. F.; Lippmann, M. (2005) Associations between ozone and daily mortality, analysis and meta-
      analysis. Epidemiology 16: 446-457.
Iwasaki, T.; Takahashi, M.; Saito, H.; Arito, H. (1998) Adaptation of extrapulmonary responses to ozone exposure
      in conscious rats. Ind. Health 36: 57-60.
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.; O'Toole, B. I.; 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.; O'Toole, 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.
Jenkins, H. S.; Devalia, J. L.; Mister, R. L.; Bevan,  A. M.; Rusznak, C.; Davies, R. J. (1999) The effect of exposure
      to ozone and nitrogen dioxide on the airway  response of atopic asthmatics to inhaled allergen: dose- and
      time-dependent effects. Am. J. Respir. Crit. Care Med. 160: 33-39.
Jones, G. N.; Sletten, 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.
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.
Jorres, R. A.; Holz, O.; Zachgo, W.; Timm, P.; Koschyk, S.; Mtiller, 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.
Just, J.; Segala, C.; Sahraoui, F.; Priol, G.; Grimfeld, A.; Neukirch, F. (2002) Short-term health effects of paniculate
      and photochemical air pollution in asthmatic children. Eur. Respir. J. 20: 899-906.
Kafoury, R. M.; Pryor, W. A.; Squadrito, G. L.;  Salgo, M. G.; Zou, X.; Friedman, M. (1999) Induction of
      inflammatory mediators in human airway epithelial cells by lipid ozonation  products. Am. J. Respir. Crit.
      Care Med. 160: 1934-1942.
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.
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.; Ozkaynak, H. (1991) Associations of daily mortality and air pollution in Los Angeles County.
      Environ. Res. 54: 99-120.
Kinney, P. L.; Ito, K.; Thurston, G. D. (1995) A sensitivity analysis of mortality/PM10 associations in Los Angeles.
      In: Phalen, R. F.; Bates, D. V., eds. Proceedings of the colloquium on paniculate air pollution and human
      mortality and morbidity; January 1994; Irvine, CA. Inhalation Toxicol. 7: 59-69.
Kinney, P. L.; Nilsen, D. M.; Lippmann, M.; Brescia, M.; Gordon, T.; McGovern,  T.; El Fawal, 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.; 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.; Lippmann, M. (2000) Respiratory effects of seasonal exposures to ozone and particles.  Arch.
      Environ. Health 55: 210-2161
Kleinman, M. T.; Bufalino, C.; Rasmussen, R.; Hyde, D.; Bhalla, D. K.; Mautz, W. J. (2000) Toxicity of chemical
      components of ambient fine paniculate matter (PM2.5) inhaled by aged rats. J. Appl. Toxicol. 20: 357-364.
                                                  3-99

-------
Klemm, R. I; 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.
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.
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.
      Ill: 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.
Kopp, M. V.; Bohnet, W.; Frischer, T.; Ulmer, C.; Studnicka, M.; Ihorst, G.; Gardner, C.; Forster, J.; Urbanek, R.;
      Kuehr, J. (2000) Effects of ambient ozone on lung function in children over a two-summer period. Eur.
      Respir. J. 16: 893-900.
Korrick, S. A.; Neas, L. M.; Dockery, D. W.; Gold, D. R.; Allen, G. A.; Hill, L. B.; Kimball, K. D.; Rosner, B. A.;
      Speizer, F. E. (1998) Effects of ozone and other pollutants on the pulmonary function of adult hikers.
      Environ. Health Perspect. 106: 93-99.
Kreit, J. W.; Gross, K. B.; Moore, T. B.; Lorenzen, T. J.; D'Arcy, J.; Eschenbacher, W. L. (1989) Ozone-induced
      changes in pulmonary function and bronchial responsiveness in asthmatics. J. Appl. Physiol. 66: 217-222.
Krewski, D.; Burnett, R. T.; Goldberg, M. S.; Hoover, K.; Siemiatycki, J,; Jerrett, M.; Abrahamowicz, M.; White,
      W. H.  (2000) Reanalysis of the Harvard Six Cities Study and the American Cancer Society Study of
      paniculate air pollution and mortality. A special report of the Institute's particle epidemiology reanalysis
      project. Cambridge, MA: Health Effects Institute.
Kulle, T. J.; Sauder, L. R.; Kerr, H. D.; Farrell, B. P.;  Bermel, M. S.; Smith, D. M. (1982) Duration of pulmonary
      function adaptation to ozone in humans. Am. Ind. Hyg. Assoc. J. 43: 832-837.
Ktinzli, N.; Lurmann, F.; Segal, M.; Ngo, L.; Balmes, J.; Tager, I. B. (1997) Association between lifetime ambient
      ozone exposure and pulmonary function in college freshmen—results of a pilot study. Environ. Res. 72: 8-23.
Lebowitz, M. D.; Camilli, A. E.; Bronnimann, D.; Quackenboss, J. (1987) The significance and meaningfulness of
      intraindividual changes in objective test results as responses to air contaminants. Presented at: 80th annual
      meeting of the Air Pollution Control Association; June; New York, NY. Pittsburgh, PA: Air Pollution Control
      Association; paper no. 87-32.1.
Lethbridge-Cejku, M.; Schiller, J. S.; Bernadel, L. (2004) Summary health statistics for U.S. adults: National Health
      Interview Survey, 2002. Hyattsville, MD: Centers for Disease Control and Prevention; DHHS publication no.
      (PHS) 2004-1550. (Vital and health statistics, series 10, number 222). Available:
      http://www.cdc.gov/nchs/data/series/sr_10/srlO_222.pdf [3, August, 2005].
Levy, J. L; Chemerynski, S. M.; Sarnat, J. A. (2005) Ozone exposure and mortality, an empiric Bayes
      metaregression analysis. Epidemiology 16: 458-468.
Liao, D.; Duan, Y.; Whitsel, E. A.; Zheng, Z.-J.; Heiss,  G.; Chinchilli, V. M.; Lin, H.-M. (2004) Association of
      higher levels of ambient criteria pollutants with impaired cardiac autonomic control: a population-based
      study. Am. J. Epidemiol. 159: 768-777.
Lin, C. A.; Martins, M. A.; Farhat, S. C. L.; Pope, C. A., Ill; Conceicno, G. M. S.; Anastacio, V. M.; Hatanaka, M.;
      Andrade, W. C.; Hamaue, W. R.; Bohm, G. M.; Saldiva, P. H. N. (1999) Air pollution and respiratory illness
      of children in Sao Paulo, Brazil. Paediatr. Perinat. Epidemiol. 13: 475-488.
Lin, M.; Chen, Y.; Burnett, R. T.; Villeneuve, P. J.; Krewski, D. (2003) Effect of short-term exposure to gaseous
      pollution on asthma hospitalisation in children: a bi-directional case-crossover analysis. J.  Epidemiol.
      Community Health 57: 50-55.
Lin, M.; Chen, Y.; Villeneuve,  P. J.; Burnett, R. T.; Lemyre, L.; Hertzman, C.; McGrail, K. M.; Krewski, D. (2004)
      Gaseous air pollutants and asthma hospitalization of children with low household income in Vancouver,
      British Columbia, Canada. Am. J. Epidemiol. 159: 294-303.
Linder, J.; Herren, D.; Monn, C.; Wanner, H.-U. (1988) Die Wirkung von Ozon auf die korperliche
      Leistungsfahigkeit [The  effect of ozone on physical activity]. SchweizZ.  Sportmed.  36: 5-10.
Linn, W. S.; Shamoo, D. A.; Anderson, K. R.; Peng, R.-C.; Avol, E. L.; Hackney, J. D.;  Gong, H., Jr. (1996) Short-
      term air pollution exposures and responses in Los Angeles area schoolchildren. J. Exposure Anal. Environ.
      Epidemiol. 6: 449-472.
Linn, W. S.; Szlachcic, Y.; Gong, H., Jr.; Kinney, P. L.; Berhane, K. T. (2000) Air pollution and daily hospital
      admissions in metropolitan Los Angeles.  Environ. Health Perspect. 108: 427-434.
                                                  3-100

-------
Lipfert, F. W.; Hammerstrom, T. (1992) Temporal patterns in air pollution and hospital admissions. Environ. Res.
      59: 374-399.
Lipfert, F. W.; Perry, H. M, Jr.; Miller, J. P.; Baty, J. D.; Wyzga, R. E.; Carmody, S. E. (2000) The Washington
      University-EPRI veterans' cohort mortality study: preliminary results. In: Grant, L. D., ed. PM2000:
      paniculate matter and health. Inhalation Toxicol.  12(suppl. 4): 41-73.
Lippmann, M. (1988) Health significance of pulmonary function responses to airborne irritants. JAPCA 38: 881-
      887.
Lippmann, M.; Ito, K.; Nadas, A.; Burnett, R. T. (2000) Association of paniculate matter components with daily
      mortality and morbidity in urban populations. Cambridge, MA:  Health Effects Institute; research report
      no. 95.
Longphre, M.; Zhang, L.-Y.; Harkema, J. R.; Kleeberger, S. R. (1999) Ozone-induced pulmonary inflammation and
      epithelial proliferation are partially mediated by PAF. J. Appl. Physiol. 86: 341-349.
Luginaah, I. N.; Fung, K. Y.; Gorey, K. M.; Webster, G.; Wills, C. (2005) Association of ambient air pollution with
      respiratory hospitalization in a government designated "area of concern": the case of Windsor, Ontario.
      Environ. Health Perspect. 113: 290-296.
Mannino, D. M.; Ford, E. S.; Redd, S. C. (2003) Obstructive and restrictive lung disease and markers of
      inflammation: data from the Third National Health and Nutrition Examination. Am. J. Med.  114: 758-762.
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. Care Med. 149: 1192-1197.
McCaig,  L. F.; Burt, C. W. (2005) National Hospital Ambulatory Medical Care Survey: 2003 Emergency
      Department Summary. Hyattsville, MD: National Center for Health Statistics; DHHS publication no. (PHS)
      2005-1250. (Advance data from vital and health statistics; no. 358). Available:
      http://www.cdc.gov/nchs/data/ad/ad358.pdf [3 August, 2005].
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.; 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.
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.
Metzger, K. B.; Tolbert, P. E.; Klein, M.; Peel, J. L.; Flanders, W. D.; Todd, K. H.; Mulholland, J. A.; Ryan, P. B.;
      Frumkin, H.  (2004) Ambient air pollution and cardiovascular emergency department visits. Epidemiology
      15: 46-56.
Michelson, P. H.; Dailey, L.; Devlin, R. B.; Peden, D. B. (1999) Ozone effects on the immediate-phase response to
      allergen in the nasal airways of allergic asthmatic subjects. Otolaryngol. Head Neck Surg. 120: 225-232.
Molfmo,  N. A.; Wright, S. C.; Katz, I.; 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.
Montuschi, P.; Nightingale, J. A.; Kharitonov, S. A.; Barnes, P. J. (2002) Ozone-induced increase in exhaled
      8-isoprostane in healthy subjects is resistant to inhaled budesonide. Free Radical Biol. Med.  33: 1403-1408.
Moolgavkar, S. H. (2003) Air pollution and daily mortality in two U.S. counties: season-specific analyses and
      exposure-response relationships. Inhalation Toxicol. 15: 877-907.
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.; Tager, I. B.; Dockery, D. W.; Neas, L. M.; Redline, S. (2000) The effect of ozone on inner-city
      children with asthma: identification of susceptible groups. Am. J. Respir. Crit. Care Med. 162: 1838-1845.
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.
                                                 3-101

-------
Mudway, I. S.; Stenfors, N.; Blomberg, A.; Helleday, R.; Dunster, C.; Marklund, S. L.; Frew, A. I; Sandstrom, T.;
      Kelly, F. J. (2001) Differences in basal airway antioxidant concentrations are not predictive of individual
      responsiveness to ozone: a comparison of healthy and mild asthmatic subjects. Free Radical Biol. Med.
      31:962-974.
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.
Naeher, L. P.; Holford, T. R.; Beckett, W.  S.; Belanger, K.; Triche, E. W.; Bracken, M. B.; Leaderer, B. P. (1999)
      Healthy women's PEF variations with ambient summer concentrations of PMio, PN2.5, SCte-, H+, and Cb. Am.
      J. Respir. Crit. Care Med. 160: 117-125.
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/asthgdln.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.
Neas, L. M.; Dockery, D. W.; Koutrakis, P.; Tollerud, D. J.; Speizer, F. E. (1995) The association of ambient air
      pollution with twice daily peak expiratory flow rate measurements in children. Am. J. Epidemiol.
      141: 111-122.
Neas, L. M.; Dockery, D. W.; Koutrakis, P.; Speizer, F. E. (1999) Fine particles and peak flow  in children: acidity
      versus mass. Epidemiology 10: 550-553.
Newhouse, C. P.; Levetin, B. S.; Levetin, E. (2004) Correlation of environmental factors with asthma and rhinitis
      symptoms in Tulsa, OK. Ann. Allergy Asthma Immunol. 92: 356-366.
Nightingale, J. A.; Rogers, D. F.; Chung, K. F.; Barnes, P. J. (2000) No effect of inhaled budesonide on the response
      to inhaled ozone in normal subjects. Am. J. Respir. Crit. Care Med. 161: 479-486.
O'Neill, M. S.; Loomis, D.; Borja-Aburto,  V. H. (2004) Ozone, area social conditions, and mortality in Mexico City.
      Environ. Res. 94: 234-242.
Ostro, B. (1995) Fine paniculate air pollution and mortality in two Southern California counties. Environ. Res.
      70: 98-104.
Ostro, B.; Lipsett, M.; Mann, J.; Braxton-Owens, H.; White, M. (2001)  Air pollution and exacerbation of asthma in
      African-American children in Los Angeles. Epidemiology 12: 200-208.
Ostro, B. D.; Broadwin, R.; Lipsett, M. J. (2003) Coarse particles and daily mortality in Coachella Valley,
      California. In: Revised analyses of time-series studies of air pollution and health. Special report. Boston, MA:
      Health Effects Institute; pp. 199-204. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdftl8
      October, 2004].
Park, H.; Lee, B.; Ha, E.-H.; Lee, J.-T.; Kim, H.; Hong, Y.-C. (2002) Association of air pollution with school
      absenteeism due to illness. Arch. Pediatr. Adolesc. Med.  156: 1235-1239.
Park, S. K.; O'Neill, M. S.; Vokonas, P. S.; Sparrow, D.;  Schwartz, J. (2005) Effects of air pollution on heart rate
      variability: the VA normative aging study. Environ. Health Perspect. 113: 304-309.
Passannante, A. N.; Hazucha, M. J.; Bromberg, P. A.; Seal, E.; Folinsbee, L.; Koch, G. (1998) Nociceptive
      mechanisms modulate ozone-induced human lung function decrements. J. Appl. Physiol. 85:  1863-1870.
Peden, D. B.;  Setzer, R. W., Jr.; Devlin, R. B.  (1995) Ozone exposure has both a priming effect on allergen-induced
      responses and an intrinsic inflammatory action in the nasal airways of perennially allergic asthmatics. Am. J.
      Respir. Crit. Care Med. 151: 1336-1345.
Peden, D. B.;  Boehlecke, B.; Horstman, D.; Devlin, R. (1997) Prolonged acute exposure to 0.16 ppm ozone induces
      eosinophilic airway inflammation in asthmatic subjects with allergies. J. Allergy Clin. Immunol. 100:
      802-808.
Peel, J. L.; Tolbert, P. E.; Klein, M.; Metzger, K. B.; Flanders, W. D.; Knox, T.; Mulholland, J. A.; Ryan, P. B.;
      Frumkin, H. (2005) Ambient air pollution and respiratory emergency department visits. Epidemiology
      16: 164-174.
Peters, J. M.; Avol, E.; Navidi, W.; London, S. J.; Gauderman, W. J.; Lurmann, F.; Linn, W. S.; Margolis, H.;
      Rappaport, E.; Gong, H., Jr.; Thomas, D. C. (1999a) A study of twelve southern California communities with
      differing levels and types of air pollution. I. Prevalence of respiratory morbidity. Am. J.  Respir. Crit. Care
      Med. 159: 760-767.
Peters, J. M.; Avol, E.; Gauderman, W. J.;  Linn, W. S.; Navidi, W.; London, S. J.; Margolis, H.; Rappaport, E.;
      Vora, H.; Gong, H., Jr.; Thomas, D. C.  (1999b) A study of twelve southern California communities with
                                                 3-102

-------
      differing levels and types of air pollution. II. Effects on pulmonary function. Am. J. Respir. Crit. Care Med.
      159: 768-775.
Peters, A.; Liu, E.; Verrier, R. L.; Schwartz, J.; Gold, D. R.; Mittleman, M; Baliff, J.; Oh, J. A.; Allen, G.;
      Monahan, K.; Dockery, D. W. (2000a) Air pollution and incidence of cardiac arrhythmia. Epidemiology
      11: 11-17.
Peters, A.; Skorkovsky, J.; Kotesovec, F.; Brynda, J.; Spix, C.; Wichmann, H. E.; Heinrich, J. (2000b) Associations
      between mortality and air pollution in central Europe. Environ. Health Perspect. 108: 283-287.
Peters, A.; Dockery, D. W.; Muller, J. E.; Mittleman, M. A. (2001) Increased paniculate air pollution and the
      triggering of myocardial infarction. Circulation 103: 2810-2815.
Petroeschevsky, A.;  Simpson, R. W.; Thalib, L.; Rutherford, S.  (2001) Associations between outdoor air pollution
      and hospital admissions in Brisbane, Australia. Arch. Environ. Health 56: 37-52.
Plopper, C. G.; Hatch, G. E.; Wong, V.; Duan, X.; Weir, A. J.; Tarkington, B. K.; Devlin, R. B.; Becker,  S.;
      Buckpitt, A. R. (1998) Relationship of inhaled ozone concentration to acute tracheobronchial epithelial
      injury, site-specific ozone dose and glutathione depletion in rhesus monkeys. Am. J. Respir. Cell Mol. Biol.
      19: 387-399.
Plopper, C. G.; Fanucchi, M. V. (2000) Do urban environmental pollutants exacerbate childhood lung diseases?
      Environ. Health Perspect. 108: A252-A253.
Plunkett, L. M.; Turnbull, D.; Rodricks, J. V.  (1992) Differences between adults and children affecting exposure
      assessment. In: Guzelian, P. S.; Henry, D. J.; Olin, S. S., eds. Similarities and differences between children
      and adults: implications for risk assessment. Washington, DC: ILSI Press, pp. 79-96.
Ponce de Leon, A.; Anderson, H. R.; Bland, J. M.; Strachan, D. P.; Bower, J. (1996) Effects of air pollution on daily
      hospital admissions for respiratory disease in London between 1987-88 and 1991-92. In: StLeger, S., ed.
      The APHEA project.  Short term effects of air pollution on health: a European approach using
      epidemiological time series data. J. Epidemiol. Commun. Health 50(suppl. 1): S63-S70.
Ponka, A.; Savela, M.; Virtanen, M. (1998) Mortality and air pollution in Helsinki. Arch. Environ. Health53: 281-
      286.
Pope, C. A., Ill; Thun, M. J.; Namboodiri, M. M.; Dockery, D. W.; Evans, J. S.; Speizer, F. E.; Heath, C. W., Jr.
      (1995) Paniculate air pollution as a predictor of mortality in a prospective study of U.S. adults. Am. J. Respir.
      Crit. Care Med. 151: 669-674.
Pope, C. A., Ill; Burnett, R. T.; Thun, M. J.; Calle, E. E.; Krewski, D.; Ito, K.; Thurston, G. D. (2002) Lung cancer,
      cardiopulmonary mortality, and long-term exposure to fine paniculate air pollution. JAMA J. Am. Med.
      Assoc. 287: 1132-1141.
Pulfer, M. K.; Murphy, R. C. (2004) Formation of biologically active oxysterols during ozonolysis of cholesterol
      present in lung surfactant. J. Biol. Chem. 279: 26331-26338.
Pulfer, M. K.; Taube, C.; Gelfand, E.; Murphy, R. C. (2005) Ozone exposure in vivo and formation of biologically
      active oxysterols in the lung. J. Pharmacol. Exp. Ther. 312: 256-264.
Raizenne, M.; Stern, B.; Burnett, R.; Spengler, J. (1987) Acute respiratory function and transported air pollutants:
      observational studies. Presented at: 80th annual meeting of the Air Pollution Control Association;  June;
      New York, NY. Pittsburgh, PA: Air Pollution Control Association; paper no. 87-32.6.
Raizenne, M. E.; Burnett, R. T.; Stern, B.; Franklin, C. A.; Spengler, J. D. (1989) Acute lung function responses to
      ambient acid aerosol exposures in children. Environ. Health Perspect. 79: 179-185.
Rich, D. Q.; Schwartz, J.; Mittleman, M. A.; Link, M.; Luttmann-Gibson, H.; Catalano, P. J.; Speizer, F.  E.;
      Dockery, D. W. (2005) Association of short-term ambient air pollution concentrations and ventricular
      arrhythmias. Am. J. Epidemiol. 161:  1123-1132.
Rigas, M. L.; Ben-Jebria, A.; Ultman, J. S. (1997) Longitudinal distribution of ozone absorption in the lung: effects
      of nitrogen dioxide, sulfur dioxide, and ozone exposures. Arch. Environ. Health 52: 173-178.
Romieu, L; Meneses, F.; Ruiz, S.; Sienra, J. J.; Huerta, J.; White, M. C.; Etzel, R. A. (1996) Effects of air pollution
      on the respiratory health of asthmatic children living in Mexico City. Am. J. Respir. Crit. Care Med.
      154: 300-307.
Romieu, I.; Meneses, F.; Ruiz, S.; Huerta, J.; Sienra, J. J.; White, M.; Etzel, R.; Hernandez, M. (1997) Effects of
      intermittent ozone exposure on peak expiratory flow and respiratory symptoms among asthmatic children in
      Mexico City.  Arch. Environ. Health 52: 368-376.
Romieu, I.; Meneses, F.; Ramirez, M.; Ruiz, S.; Padilla, R.  P.; Sienra, J. J.; Gerber, M.; Grievink, L.;  Dekker, R.;
      Walda, I.; Brunekreef, B. (1998) Antioxidant supplementation and respiratory functions among workers
      exposed to high levels of ozone. Am. J. Respir. Crit. Care Med. 158:  226-232.
                                                  3-103

-------
Romieu, I.; Sienra-Monge, J. J.; Ramirez-Aguilar, M.; Tellez-Rojo, M. M.; Moreno-Macias, H.; Reyes-Ruiz, N. I.;
      Del Rio-Navarro, B. E.; Ruiz-Navarro, M. X.; Hatch, G.; Slade, R.; Hernandez-Avila, M. (2002) Antioxidant
      supplementation and lung functions among children with asthma exposed to high levels of air pollutants.
      Am. J. Respir. Crit. Care Med. 166: 703-709.
Romieu, I.; Sienra-Monge, J. J.; Ramirez-Aguilar, M.; Moreno-Macias, H.; Reyes-Ruiz, N. I.; Estela del Rio-
      Navarro, B.; Hernandez-Avila, M.; London, S. J. (2004) Genetic polymorphism of GSTM1 and antioxidant
      supplementation influence lung function in relation to ozone exposure in asthmatic children in Mexico City.
      Thorax 59: 8-10.
Ross, M. A.; Persky, V. W.; Scheff, P. A.; Chung, J.; Curtis, L.; Ramakrishnan, V.; Wadden, R. A.; Hryhorczuk,
      D. O. (2002) Effect of ozone and aeroallergens on the respiratory health of asthmatics. Arch. Environ. Health
      57: 568-578.
Ruidavets, J.-B.;  Cournot, M.;  Cassadou, S.; Giroux, M.; Meybeck, M.; FerriDres, J. (2005) Ozone air pollution is
      associated with acute myocardial infarction. Circulation 111: 563-569.
Samet, J. M.; Zeger, S. L.; Dominici, F.; Curriero, F.; Coursac, L; Dockery, D. W.; Schwartz, J.; Zanobetti, A.
      (2000) The national morbidity, mortality, and air pollution study. Part II: morbidity, mortality, and air
      pollution in the United States. Cambridge, MA: Health Effects Institute; research report no. 94, part II.
Samet, J. M.; Hatch, G. E.; Horstman,  D.; Steck-Scott, S.; Arab, L.; Bromberg, P. A.; Levine, M.;  McDonnell, W.
      F.; Devlin, R. B. (2001) Effect of antioxidant supplementation on ozone-induced lung injury in human
      subjects. Am. J. Respir. Crit. Care Med.  164: 819-825.
Sartor, F.; Snacken, R.; Demuth, C.; Walckiers, D. (1995) Temperature, ambient ozone levels,  and mortality during
      summer, 1994, in Belgium. Environ. Res. 70:  105-113.
Sathishkumar, K.; Haque, M.; Perumal, T. E.; Francis, J.; Uppu, R. M. (2005) A major ozonation product of
      cholesterol, 3"p"-hydroxy-5-oxo-5,6-secocholestan-6-al, induces apoptosis inH9c2 cardiomyoblasts.
      FEES Lett. 579: 6444-6450.
Scannell, C.; Chen, L.; Aris, R. M.; Tager, I.; Christian, D.; Ferrando, R.; Welch, B.; Kelly, T.; Balmes, J. R. (1996)
      Greater ozone-induced inflammatory responses in subjects with asthma. Am. J. Respir. Crit. Care Med.
      154: 24-29.
Scarlett, J. F.; Abbott, K. J.; Peacock, J. L.; Strachan, D. P.; Anderson, H. R.  (1996) Acute effects  of summer air
      pollution on respiratory function in primary school children in southern England. Thorax 51: 1109-1114.
Schelegle, E. S.; Adams, W. C. (1986) Reduced exercise time in competitive simulations consequent to low level
      ozone exposure. Med. Sci. Sports Exercise 18: 408-414.
Schelegle, E. S.; Miller, L. A.; Gershwin,  L. J.; Fanucchi, M. V.; Van Winkle, L. S.; Gerriets, J. E.; Walby, W. F.;
      Mitchell, V.; Tarkington, B. K.; Wong, V. J.; Baker, G. L.; Pantle, L. M.; Joad, J. P.; Pinkerton, K. E.;
      Wu, R.;  Evans, M. J.; Hyde, D.  M.; Plopper, C. G. (2003) Repeated episodes of ozone inhalation amplifies
      the effects of allergen sensitization and inhalation on airway immune and structural development in Rhesus
      monkeys. Toxicol. Appl. Pharmacol. 191: 74-85.
Schierhorn, K.; Zhang, M.; Matthias, C.; Kunkel, G. (1999) Influence of ozone and nitrogen dioxide on histamine
      and interleukin formation in a human nasal mucosa culture system. Am. J. Respir. Cell Mol. Biol. 20: 1013-
      1019.
Schierhorn, K.; Hanf, G.; Fischer, A.; Umland, B.; Olze, H.; Kunkel, G. (2002) Ozone-induced release of
      neuropeptides from human nasal mucosa cells. Int. Arch. Allergy Immunol. 129: 145-151.
Schindler, C.; Ktinzli, N.; Bongard, J.-P.; Leuenberger, P.; Karrer, W.; Rapp, R.;  Monn, C.; Ackermann-Liebrich,
      U.;  Swiss Study on Air Pollution and Lung Diseases in Adults Investigators. (2001) Short-term variation in
      air pollution and in average lung function among never-smokers. Am. J. Respir. Crit. Care Med.  163: 356-
      361.
Schwartz, J. (1994) Air pollution and hospital admissions for the elderly in Detroit, Michigan.  Am. J. Respir. Crit.
      Care Med. 150: 648-655.
Schwartz, J. (1996) Air pollution and hospital admissions for respiratory disease. Epidemiology 7: 20-28.
Schwartz, J.; Spix, C.; Touloumi, G.; Bacharova, L.; Barumamdzadeh, T.; le Tertre, A.; Piekarksi, T.;
      Ponce de Leon, A.; Ponka, A.; Rossi, G.;  Saez, M.; Schouten, J. P. (1996)  Methodological issues in studies of
      air pollution and daily counts of deaths or hospital admissions. In:  St Leger, S., ed. The APHEA project.
      Short term effects of air pollution on health: a European approach using epidemiological time series data.
      J. Epidemiol. Commun. Health  50(suppl. 1):  S3-S11.
Schwartz, J. (2005) How sensitive is the association between ozone and daily deaths to control for temperature?
      Am. J. Respir. Crit. Care Med. 171: 627-631.
                                                 3-104

-------
Seal, E., Jr.; McDonnell, W. F.; House, D. E. (1996) Effects of age, socioeconomic status, and menstrual cycle on
      pulmonary response to ozone. Arch. Environ. Health 51:  132-137.
Sexton, K. G.; Jeffries, H. E.; Jang, M; Kamens, R. M; Doyle, M; Voicu, I.; Jaspers, I. (2004) Photochemical
      products in urban mixtures enhance inflammatory responses in lung cells. Inhalation Toxicol.
      16(suppl. 1): 107-114.
Sheppard, L.; Levy, D.; Norris, G.; Larson, T. V.; Koenig, J. Q.  (1999) Effects of ambient air pollution on
      nonelderly asthma hospital admissions in Seattle, Washington, 1987-1994. Epidemiology 10: 23-30.
Sheppard, L. (2003) Ambient air pollution and nonelderly asthma hospital admissions in Seattle, Washington, 1987-
      1994. In: Revised analyses of time-series studies of air pollution and health. Special report. Boston, MA:
      Health Effects Institute; pp. 227-230. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdf
      [18 October, 2004].
Simpson, R. W.; Williams, G.; Petroeschevsky, A.; Morgan, G.; Rutherford, S. (1997) Associations between
      outdoor air pollution and daily mortality in Brisbane, Australia. Arch. Environ. Health 52: 442-454.
Sin, D. D.; Man,  S. F. P. (2003) Why are patients with chronic obstructive pulmonary disease at increased risk of
      cardiovascular diseases? Circulation 107: 1514-1519.
Spektor, D. M.; Lippmann, M.; Lioy, P. J.; Thurston, G. D.; Citak, K.; James, D. J.; Bock, N.; Speizer, F. E.;
      Hayes, C.  (1988a) Effects of ambient ozone on respiratory function in active, normal children. Am. Rev.
      Respir. Dis. 137:313-320.
Spektor, D. M.; Lippmann, M.; Thurston, G. D.; Lioy, P. J.; Stecko, J.; O'Connor, G.; Garshick, E.; Speizer, F. E.;
      Hayes, C.  (1988b) Effects of ambient ozone on respiratory function in healthy adults exercising outdoors.
      Am.  Rev. Respir. Dis. 138: 821-828.
Spektor, D. M.; Lippmann, M. (1991) Health effects of ambient ozone on healthy children at a summer camp. In:
      Berglund,  R. L.; Lawson, D. R.;  McKee, D. J.,  eds. Tropospheric ozone and the environment: papers from an
      international conference; March  1990; Los Angeles, CA. Pittsburgh, PA: Air & Waste Management
      Association; pp. 83-89. (A&WMA transactions series no. TR-19).
Steck-Scott, S.; Arab, L.; Craft, N. E.; Samet, J. M. (2004) Plasma and lung macrophage responsiveness to
      carotenoid supplementation and ozone exposure in humans. Eur. J. Clin. Nutr. 1-9.
Stenfors, N.; Pourazar, J.; Blomberg, A.; Krishna, M. T.; Mudway, I.; Helleday, R.; Kelly, F. J.; Frew, A. J.;
      Sandstrom, T.  (2002) Effect of ozone on bronchial mucosal inflammation in asthmatic and healthy subjects.
      Respir. Med. 96: 352-358.
Stieb, D. M.; Burnett, R. T.; Beveridge, R. C.; Brook,  J. R. (1996) Association between ozone and asthma
      emergency department visits in Saint John, New Brunswick, Canada. Environ. Health Perspect. 104: 1354-
      1360.
Sunyer, J.; BasagaDa, X.; Belmonte, J.; Anto, J. M. (2002) Effect of nitrogen dioxide and ozone on the risk of dying
      in patients with severe asthma. Thorax 57: 687-693.
Tager, I. B.; Ktinzli, N.; Lurmann, F.; Ngo, L.; Segal,  M.; Balmes, J. (1998) Methods development for
      epidemiological investigations of the health effects of prolonged ozone exposure. Part II: an approach to
      retrospective estimation of lifetime ozone exposure using a questionnaire and ambient monitoring data
      (California sites). Cambridge, MA: Health Effects Institute; research report no. 81; pp. 27-78.
Taggart, S.  C. O.; Custovic, A.; Francis, H. C.; Faragher, E. B.; Yates, C. J.; Higgins, B. G.; Woodcock, A. (1996)
      Asthmatic bronchial hyperresponsiveness varies with ambient levels of summertime air pollution. Eur.
      Respir. J. 9: 1146-1154.
Tenias, J. M.; Ballester, F.; Rivera, M. L. (1998) Association between hospital emergency visits for asthma and air
      pollution in Valencia, Spain. Occup. Environ. Med. 55: 541-547.
Tenias, J. M.; Ballester, F.; Perez-Hoyos, S.; Rivera, M. L. (2002) Air pollution and hospital emergency room
      admissions for chronic obstructive pulmonary disease in Valencia, Spain.  Arch. Environ. Health 57: 41-47.
Thurston, G. D.; Ito, K.; Kinney, P. L.; Lippmann, M. (1992) A multi-year study of air pollution and respiratory
      hospital admissions in three New York State metropolitan areas: results for 1988 and 1989 summers.
      J. Exposure Anal. Environ. Epidemiol. 2: 429-450.
Thurston, G. D.; Ito, K.; Hayes, C. G.; Bates, D. V.; Lippmann, M. (1994) Respiratory hospital admissions and
      summertime haze air pollution in Toronto, Ontario: consideration of the role of acid aerosols. Environ. Res.
      65: 271-290.
Thurston, G. D.; Lippmann, M.; Scott, M. B.; Fine, J.  M. (1997) Summertime haze air pollution and children with
      asthma. Am. J. Respir. Crit. Care Med. 155: 654-660.
Tobias, A.;  Campbell, M. J.; Saez, M. (1999) Modelling asthma epidemics on the relationship between air pollution
      and asthma emergency visits in Barcelona, Spain. Eur. J. Epidemiol. 15: 799-803.
                                                 3-105

-------
Tolbert, P. E.; Mulholland, J. A.; Macintosh, D. L.; Xu, F.; Daniels, D.; Devine, O. I; Carlin, B. P.; Klein, M;
      Dorley, J.; Butler, A. J.; Nordenberg, D. F.; Frumkin, H.; Ryan, P. B.; White, M. C. (2000) Air quality and
      pediatric emergency room visits for asthma in Atlanta, Georgia. Am. J. Epidemiol. 151:  798-810.
Touloumi, G.; Katsouyanni, K.; Zmirou, D.; Schwartz, J.; Spix, C.; Ponce de Leon, A.; Tobias, A.; Quennel, P.;
      Rabczenko, D.; Bacharova, L.; Bisanti, L.; Vonk, J. M.; Ponka, A. (1997) Short-term effects of ambient
      oxidant exposure on mortality: a combined analysis within the APHEA project. Am. J. Epidemiol.  146: 177-
      185.
Tsai, J.-J.; Lin, Y.-C.; Kwan, Z.-H.; Kao, H.-L. (1998) Effects of ozone on ovalbumin sensitization in guinea pigs.
      J. Microbiol. Immunol. Infect. 31: 225-232.
Tsai, B. M.; Wang, M.; Pitcher, J. M.; Meldrum, K. K.; Meldrum, D. R. (2004) Hypoxic pulmonary
      vasoconstriction and pulmonary artery tissue cytokine expression are mediated by protein kinase C. Am. J.
      Physiol. Lung Cell Mol. Physiol. 287: L1215-L1219.
U.S. Environmental Protection Agency. (1986) Air quality criteria for ozone and other photochemical oxidants.
      Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and
      Assessment Office; report nos. EPA-600/8-84-020aF-eF. 5v. Available from: NTIS, Springfield, VA; PB87-
      142949.
U.S. Environmental Protection Agency. (1996) Air quality criteria for ozone and related photochemical oxidants.
      Research Triangle Park, NC: Office of Research and Development; report nos. EPA/600/AP-93/004aF-cF.
      3v. Available from:  NTIS, Springfield, VA; PB96-185582, PB96-185590, and PB96-185608. Available:
      http://cfpub2.epa.gov/ncea/.
U.S. Environmental Protection Agency. (2004) Air quality criteria for paniculate matter. Research Triangle Park,
      NC: National Center for Environmental Assessment; report no. EPA/600/P-99/002aF-bF. 2v. Available:
      http://www.epa.gov/ttn/naaqs/standards/pm/s_pm_cr_cd.html [9 November, 2004].
U.S. Environmental Protection Agency (2006) Air quality criteria for ozone and related photochemical oxidants
      (second external review draft) Research Triangle Park, NC: National Center for Environmental Assessment;
      report no. EPA/600R-05/004aB-cB, 3v. Available:
      http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid= 137307 [March 2006]
Ulmer, C.; Kopp, M.; Ihorst, G.; Frischer, T.; Forster, J.; Kuehr, J. (1997) Effects of ambient ozone exposures
      during the spring and summer of 1994 on pulmonary function of schoolchildren.  Pediatr. Pulmonol. 23: 344-
      353.
Ultman, J. S.; Ben-Jebria, A.; Arnold,  S. F. (2004) Uptake distribution of ozone in human lungs: intersubject
      variability in physiologic response. Boston, MA:  Health Effects Institute.
Vagaggini, B.; Taccola, M.; Clanchetti, S.; Carnevali, S.; Bartoli, M. L.; Bacci, E.; Dente, F. L.; Di Franco, A.;
      Giannini, D.; Paggiaro, P. L. (2002) Ozone exposure increases eosinophilic airway response induced by
      previous  allergen challenge. Am. J. Respir. Crit. Care Med. 166: 1073-1077.
Vedal, S.; Brauer, M.; White, R.; Petkau, J. (2003) Air pollution and daily mortality in a city with low levels of
      pollution. Environ. HealthPerspect. Ill:  45-51.
Vesely, D. L.; Giordano, A. T.; Raska-Emery, P.; Montgomery, M. R. (1994a) Ozone increases amino- and carboxy-
      terminal atrial natriuretic factor prohormone peptides in lung, heart, and circulation. J. Biochem. Toxicol. 9:
      107-112.
Vesely, D. L.; Giordano, A. T.; Raska-Emery, P.; Montgomery, M. R. (1994b) Increase in atrial natriuretic factor in
      the lungs, heart, and circulatory system owing to ozone. Chest 105: 1551-1554.
Vesely, D. L.; Giordano, A. T.; Raska-Emery, P.; Montgomery, M. R. (1994c) Ozone increases atrial natriuretic
      peptides in heart, lung and circulation of aged vs. adult animals. Gerontology (Basel) 40: 227-236.
Villeneuve, P. J.; Burnett, R. T.; Shi, Y.; Krewski, D.; Goldberg, M. S.; Hertzman, C.; Chen, Y.; Brook, J. (2003)
      A time-series study of air pollution, socioeconomic status, and mortality in Vancouver, Canada. J. Exposure
      Anal. Environ. Epidemiol. 13: 427-435.
Watkinson, W. P.; Wiester, M. J.; Highfill, J. W. (1995) Ozone toxicity in the rat. I. Effect of changes in ambient
      temperature on extrapulmonary physiological parameters. J. Appl. Physiol. 78: 1108-1120.
Watkinson, W. P.; Campen, M.  J.; Nolan, J. P.; Costa, D. L. (2001) Cardiovascular and systemic responses to
      inhaled pollutants in rodents: effects of ozone and paniculate matter. Environ. Health Perspect.
      109(suppl. 4): 539-546.
Watkinson, W. P.; Campen, M.  J.; Wichers, L. B.; Nolan, J. P.; Costa, D. L. (2003) Cardiac and thermoregulatory
      responses to inhaled pollutants in healthy and compromised rodents: modulation via interaction with
      environmental factors. Environ. Res. 92: 35-47.
                                                 3-106

-------
Weisel, C. P.; Cody, R. P.; Lioy, P. J. (1995) Relationship between summertime ambient ozone levels and
      emergency department visits for asthma in central New Jersey. Environ. Health Perspect. 103(suppl. 2): 97-
      102.
Weisel, C. P.; Cody, R. P.; Georgopoulos, P. G.; Purushothaman, V.; Weiss, S. H.; Bielory, L.; Gregory, P.; Stern,
      A. H. (2002) Concepts in developing health-based indicators for ozone. Int. Arch. Occup. Environ. Health
      75: 415-422.
Wentworth, P., Jr.; Nieva, J.; Takeuchi, C.; Galve, R.; Wentworth, A. D.; Dilley, R. B.; DeLaria, G. A.; Saven, A.;
      Babior, B. M; Janda, K. D.; Eschenmoser, A.; Lerner, R. A. (2003) Evidence for ozone formation in human
      atherosclerotic arteries. Science (Washington, DC, U.S.) 302: 1053-1056.
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. (1991a) 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.
Wilson, A. M.; Wake,  C. P.; Kelly, T.; Salloway, J. C. (2005) Air pollution, weather, and respiratory emergency
      room visits in two northern New England cities: an ecological time-series study. Environ. Res. 97: 312-321.
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.
Woodwell, D. A.; Cherry, D. K. (2004) National Ambulatory Medical Care Survey: 2002 summary. Hyattsville,
      MD: National Center for Health Statistics; DHHS publication no. (PHS) 2004-1250. (Advance data from vital
      and health statistics; no. 346). Available: http://www.cdc.gov/nchs/data/ad/ad346.pdf [3 August, 2005].
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).
Zhu, L.; Carlin, B. P.; Gelfand, A. E. (2003) Hierarchical regression with misaligned spatial data: relating ambient
      ozone andpediatric asthma visits in Atlanta. Environmetrics 14: 537-557.
Zidek, J.V.; Shaddick,  G.; White, R.; Meloche, J.; Chatfield, C. (2005) Using a probabilistic model (pCNEM) to
      estimate personal exposure air pollution. Environmetrics 16: 481-493.
Zmirou, D.; Schwartz,  J.; Saez, M.; Zanobetti, A.; Wojtyniak, B.; Touloumi, G.; Spix, C.; Ponce de Leon, A.;
      Le Moullec, Y.; Bacharova, L.; Schouten, J.; Ponka, A.; Katsouyanni, K.  (1998) Time-series analysis of air
      pollution and cause-specific mortality. Epidemiology 9: 495-503.
                                                 3-107

-------
        4.   CHARACTERIZATION OF HUMAN EXPOSURE TO OZONE
 4.1   INTRODUCTION
       As part of the last 63 NAAQS review, EPA conducted exposure analyses for the general
population, children who spent more time outdoors, and outdoor workers.  Exposure estimates
were generated for nine urban areas for "as is" (i.e., a recent year) air quality and for just meeting
the existing 1-hr standard and several alternative 8-hr standards. EPA also conducted a health
risk assessment that produced risk estimates for the number of children and percent of children
experiencing impaired lung function and other respiratory symptoms associated with the
exposures estimated for these same nine urban areas.
       The exposure analysis conducted for the current review builds upon the methodology and
lessons learned from the exposure analyses conducted for the last review (US EPA, 1996a).  The
methodology used to conduct the exposure analysis as well as summary results from the
exposure analysis are described in this chapter.  The exposure analysis technical support
document, Ozone Population Exposure Analysis for Selected Urban Areas (US EPA, 2007)
(hereafter cited as "Exposure Analysis TSD") presents a detailed description of the exposure
analysis methodology.
       Population exposures to ambient Os levels are modeled for 12 urban areas located across
the U.S. using the Air Pollutants Exposure (APEX) model, also referred to as the Total Risk
Integrated Methodology Inhalation Exposure (TRIM.Expo) model (US EPA, 2006a,b).
Exposure estimates are developed for Os levels in recent years, based on 2002, 2003, and 2004
ambient air quality measurements. Exposures are also estimated for 63 levels associated with
just meeting the current 8-hr O3 NAAQS and several potential alternative standards, based on
adjusting data derived from the ambient monitoring network as described in section 4.5.8.
Exposures to background levels of 63 are also estimated, based on 63 concentrations predicted
by the GEOS-CHEM atmospheric photochemical model.
       Exposures are modeled for 1) the general population, 2) all school-age children (ages
5-18), 3) active school-age children, and 4) asthmatic school-age children.1  The strong
emphasis on children reflects the finding of the last Os NAAQS review that children are an
important at-risk group.  Two groups at increased risk that are not analyzed by exposure
modeling are older adults and outdoor workers. Two groups at increased risk for which exposure
       1  Subsequent to completion of this modeling, EPA analysis of uncertainty of the exposure modeling results
uncovered an error in how children are characterized as active. This error resulted in an overestimate of the number
of active children in the population. Thus, exposure estimates for active children are not included in this chapter.

                                          4-1

-------
modeling results are not presented are older adults and outdoor workers.  Although older adults
are in the general population exposure estimates, we did not separately tabulate exposures for
older adults since we do not have specific exposure-response data for this group. Exposure to
outdoor workers has not been modeled due to insufficient data to properly characterize this
population.
       This chapter provides a brief overview of the types of studies that provide data on which
this analysis is based, followed by a description of the exposure model used for this analysis, the
model input data, and the results of the analysis. The final sections of this chapter discuss the
exposure estimates in comparison to those from the prior review and summarize the sensitivity
analyses and model evaluation that have been conducted for the Os exposure model described in
this chapter.  The uncertainty assessment and a technical description of the modeling effort are
provided in separate documents (Langstaff, 2007; US EPA, 2007).

4.2   OZONE EXPOSURE STUDIES
       Many studies have produced information and data supporting the development of
methods for estimating human exposure to ambient 63 over the past several decades.  These
studies have been reviewed in the current and previous EPA Ozone Air Quality Criteria
Documents (US EPA,  1986, 1996b, 2006c).
       The types of studies which provide the basis for modeling human exposure  to O3 include
studies of people's activities, work and exercise patterns, physiology, physics and Os-related
chemistry in microenvironments, atmospheric modeling of Os, chamber studies of atmospheric
chemistry, and modeling of meteorology. Measurements that have proven to be useful for
understanding and estimating exposure obtained from personal exposure assessment studies
include fixed-site ambient concentrations, concentrations in specific indoor and outdoor
microenvironments, personal exposure levels, personal activity patterns, air exchange rates,
infiltration rates, deposition and decay rates, and meteorology.

       4.2.1  Exposure Concepts  and Definitions
       Human exposure to a contaminant is defined as "contact at a boundary between a human
and the environment at a specific contaminant concentration for a specific interval of time," and
has units of concentration times time (National Research Council, 1991). For airborne pollutants
the contact boundary is nasal and oral openings in the body, and personal exposure of an
individual to a chemical in the air for a discrete time period is quantified as (Lioy, 1990; National
Research Council, 1991):
                   ?2
                     C(t)dt                                                    (4.1)
                                          4-2

-------
where E[rl5r2] is the personal exposure during the time period from t\ to fc, and C(f) is the
concentration at time t in the breathing zone. We refer to the exposure concentration to mean the
concentration to which one is exposed.  The breathing rate (ventilation rate) at the time of
exposure is an important determinant of the dose received by the individual.  Although we do not
estimate dose, we refer to intake as the total amount of O3 inhaled (product of exposure
concentration, duration, and minute ventilation rate).
       Personal exposure to O3 can be estimated directly by monitoring the concentration of O3
in the person's breathing zone (close to the nose/mouth) using a personal exposure monitor.
Exposure can also be estimated indirectly, by estimating or monitoring the concentrations over
time in locations in which the individual spends time and estimating the time and duration the
individual spends in each location. In both of these methods, Equation 4-1 is used to calculate an
estimate of personal exposure. A key concept in modeling exposure is the microenvironment, a
term that refers to the immediate surroundings of an individual. A microenvironment is a
location in which pollutant concentrations are relatively homogeneous for short periods of time.
Microenvironments can be outdoors or indoors; some examples are outdoors near the home,
outdoors near the place of work, bedrooms, kitchens, vehicles, stores, restaurants, street-corner
bus stops, schools, and places of work.  A bedroom may be treated as a different
microenvironment than a kitchen if the concentrations are significantly different in the two
rooms. The concentrations in a microenvironment typically change over time; for example, O3
concentrations in a kitchen while cooking with a gas stove may be lower than when these
activities are not being performed, due to scavenging of O3  by nitric oxide (NO) emissions from
the gas burned.
       An important factor affecting the concentrations of O3 indoors is the degree to which the
ambient outdoor air is transported indoors. This can be modeled using physical factors such as
air exchange rates (AERs), deposition and decay rates, and penetration factors.  The volumetric
exchange rate (m3/hour) is the rate of air exchange between the indoor and outdoor air. The AER
between indoors and outdoors is the number of complete air exchanges per hour and is equal to
the volumetric exchange rate divided by the volume of the well-mixed indoor air. Indoor
concentrations of O3 can be decreased by uptake of O3 by surfaces and by chemical  reactions.
The deposition and chemical decay rates  are the rates (per hour) at which O3 is removed from
the air by surface uptake and chemical reactions.  Some exposure models employ an infiltration
factor, which is conceptually useful if distinguishing between the air exchange processes  of air
blowing through open doors and windows and the infiltration of air through smaller openings.
Since measurements of AERs account for both of these processes (including infiltration), this
distinction is not useful in applied modeling of O3 exposures and will not be discussed further
here. Simpler exposure models use a "factor model" approach to estimate indoor O3
                                          4-3

-------
concentrations by multiplying the ambient outdoor concentrations by an indoor/outdoor
concentration ratio, referred to as a, penetration factor.

       4.2.2   Monitoring Equipment Considerations
       Exposure assessment studies involve monitoring airborne Os and/or other pollutants, and
monitor design and placement play a critical role in interpreting the results of these studies.  For
exposure assessment purposes there are two general classes of monitors, personal exposure
monitors (PEMs) and fixed site monitors.
       PEMs are designed to be worn or carried easily by individuals and to measure the
concentrations experienced by individuals over a period of hours, days, or weeks.  The
placement of PEMs is important; the desired placement is usually in the breathing zone near the
mouth and nose, but where the monitor will not be excessively impacted by exhaled air. This
placement is intended to represent  the concentrations the individual breathes in. PEMs typically
report continuously measured 63 concentrations with averaging times ranging from 1  to 24
hours.
       The CD reviews 63 PEMS  (CD, Appendix AX3, p. 163-5)  and notes that humidity, wind
velocity, badge placement, and interference with other pollutants may result in measurement
error. The CD reports PEM detection limits ranging from 5 to 23 ppb for averaging times from
24-hr to 1-hr, respectively.
       Fixed-site monitors measure concentrations over time at a given location. There are
numerous fixed-site Os monitors which are part of national, state, and local air monitoring
networks. In addition to their role  of being used to determine which areas are in compliance with
existing Os NAAQS, these are also useful for alerting the public to high Os days, providing air
quality data in support of photochemical modeling and exposure assessments for a study area, for
tracking 63 levels and trends, and for studying the representativeness of measurements at these
monitors for the study area.  Existing fixed-site monitors usually report hourly averaged
concentrations, and are in operation over a period of years. Federal reference and equivalent 63
monitoring methods are required to have a lower detectable limit of 0.01 ppm and precision of
0.01 ppm for 1-hr average concentrations (40 CFR Ch.  1, §53.21).  A discussion of monitoring
equipment and networks can be found in Chapter 2 of this draft Staff Paper and in section 2.6 in
the CD.
       There are also stationary monitors expressly set up for particular exposure field studies.
These are used to measure concentrations over time in microenvironments,  such as rooms in a
home, just outside a home, roadsides, and so forth. The stationary  monitors which are outdoors
can provide information about community-scale representativeness of routinely operated fixed-
site monitors in or near the community.
                                           4-4

-------
       4.2.3  Personal Ozone Exposure Assessment Studies
       Useful PEM studies would have data collected repeatedly from each individual in the
study over a period of time, yielding a longitudinal time series of hourly (or shorter) average
concentrations each individual is exposed to. These studies would permit analysis of both the
temporal and spatial variability of each person's personal exposure to O3.
       Some studies could be designed so that the data are sampled randomly from the
population, which reduces bias and allows one to make inferences about exposure in the broader
population. Most studies addressing O3 exposure have not been of random design and the
measurement averaging times are longer than hourly.  They might have specific goals for which
randomness is not required, or be subject to constraints which do not allow for random sampling.
These non-random studies have been helpful in the development of models of exposure;
however, we recognize that they may not be representative of the broader population.

       4.2.4  Microenvironmental Studies
       The focus of microenvironmental studies is on measuring concentrations in different
locations that people spend time in, as well as on measuring the movement of pollutants from
one microenvironment to another and on measuring other parameters that contribute to
variability in exposure. Typically, microenvironmental measurements include indoor and
outdoor concentrations of O3 and other pollutants, AERs, infiltration factors, deposition rates,
decay rates, emissions of O3, NOX, VOCs,  and other pollutants, operating characteristics of air
conditioning systems, and meteorological data such as wind velocity, temperature, and humidity.
The CD discusses several studies of microenvironments that contribute to our understanding of
the factors and processes that affect exposure to O3 (CD Appendix AX3, p. 191-216).
       There is a great deal of variability among individuals in the amount of time spent indoors,
but the majority of people spend most of their time indoors (Graham & McCurdy, 2004), and
therefore the concentrations of O3 indoors  can be an important determinant of people's exposure
to O3.  There are several factors affecting O3 concentrations indoors. The ambient outdoor
concentration of O3 and the AER are the primary determinants of the indoor concentrations.
Removal processes are also significant, the most important of which is deposition onto indoor
surfaces such as carpets, furnishings, and ventilation ductwork.  Chemical reactions of O3 with
other compounds, such as solvents from consumer products or nitric oxide emissions from gas
stoves, also deplete O3 indoors. (Weschler, 2000; Monn, 2001.)
       The primary sources of O3 indoors are O3-generating air cleaners and  some photocopiers
and laser printers. Ozone generators can increase indoor concentrations by more than 0.05 ppm.
Some older photocopiers, if run continuously in an enclosed area, can increase O3 concentrations
                                          4-5

-------
by as much as 0.15 ppm. Older laser printers can produce concentrations of up to 0.18 ppm
indoors. (US EPA, 1995; CARS, 2005.)

4.3  EXPOSURE MODELING
       Models of human exposure to airborne pollutants are typically driven by estimates of
ambient outdoor concentrations of the pollutants, which vary by time of day as well as by
location. These outdoor concentration estimates may be provided by measurements, by air
quality models, or by a combination of these. It is only possible to address hypothetical future
scenarios using some form of modeling.  The main purpose of this exposure analysis is to allow
comparisons of population exposures to 63 within each urban area, associated with current air
quality levels and with several potential alternative air quality standards or scenarios.  Human
exposure, regardless of the pollutant, depends on where an individual is located and what they
are doing. Inhalation exposure models are useful in realistically estimating personal exposures
and intake based on activity-specific ventilation rates, particularly when recognizing that these
measurements cannot be performed  for a given population.  This section provides a brief
overview of the model used by EPA to estimate 63 population exposure.  Details about the
application of the model to estimate Os population exposure are provided in the following
sections and in the Exposure Analysis TSD (EPA, 2006a).
       4.3.1  The APEX Model
       The EPA has developed the APEX model for estimating human population exposure to
criteria and air toxic pollutants. APEX also serves as the human inhalation exposure model
within the Total Risk Integrated Methodology (TRIM) framework (Richmond et al., 2002; EPA
2006c). APEX is conceptually based on the probabilistic NAAQS Exposure Model (pNEM) that
was used in the last 63 NAAQS review (Johnson et al., 1996a;  1996b: 1996c).  Since that time
the model has been restructured, improved, and expanded to reflect conceptual advances in the
science of exposure modeling and newer input data available for the model. Key improvements
to algorithms include replacement of the cohort approach with a probabilistic sampling approach
focused on individuals, accounting for fatigue and oxygen debt after exercise in the calculation
of ventilation rates, and a new approach for construction of longitudinal activity patterns for
simulated persons.  Major improvements to data input to the model include updated AERs,
census and commuting data, and the daily time-activities database. These improvements are
described later in this chapter.
       APEX is a probabilistic model designed to account for the numerous sources of
variability that affect people's  exposures.  APEX simulates the movement of individuals through
time and space and estimates their exposure to a given pollutant in indoor, outdoor, and in-
                                          4-6

-------
vehicle microenvironments. Figure 4-1 provides a schematic overview of the APEX model.  The
model stochastically generates simulated individuals using census-derived probability
distributions for demographic characteristics (Figure 4-1, steps 1-3).  The population
demographics are drawn from the year 2000 Census at the tract level, and a national commuting
database based on 2000 census data provides home-to-work commuting flows between tracts.2
Any number of simulated individuals can be modeled, and collectively they approximate a
random sampling of people residing in a particular study area.
       Daily activity patterns for individuals in a study area, an input to APEX, are obtained
from detailed diaries that are compiled in the Consolidated Human Activity Database (CHAD)
(McCurdy et al., 2000; EPA, 2002).  The diaries are used to construct a sequence of activity
events for simulated individuals consistent with their demographic characteristics, day type, and
season of the year, as defined by ambient temperature regimes (Graham & McCurdy, 2004)
(Figure 4-1, step 4). The time-location-activity diaries input to APEX contain information
regarding an individuals' age, gender, race, employment status, occupation, day-of-week, daily
maximum hourly average temperature, the location, start time, duration, and type of each activity
performed. Much of this information is used to best match the activity diary with the generated
personal profile, using age, gender, employment status, day of week, and temperature as first-
order characteristics. The approach is designed to capture the important attributes contributing to
an individuals' behavior, and of particular relevance here, time spent outdoors (Graham and
McCurdy, 2004).  Furthermore,  these diary selection criteria give credence to the use of the
variable data that comprise CHAD (e.g., data collected were from different seasons, different
states of origin, etc.). APEX calculates the concentration in the microenvironment associated
with each event in an individual's activity pattern and sums the event-specific exposures within
each hour to obtain a continuous series of hourly exposures spanning the time period of interest
(Figure 4-1, steps 5, 6).
       APEX has a flexible approach for modeling microenvironmental concentrations, where
the user can define the microenvironments to be modeled and their characteristics.  Typical
indoor microenvironments include residences,  schools, and offices.  Outdoor microenvironments
include near roadways, at bus stops, and playgrounds.  Inside cars, trucks,  and mass transit
vehicles are microenvironments which are classified separately from indoors and outdoors.
       ! There are approximately 65,400 census tracts in the -3,200 counties in the U.S.
                                           4-7

-------
                                              Figure 4-1.  Overview of the APEX Model
      1.  Characterize study area
                2. Characterize study population
                                    3. Generate N number of
                                 simulated individuals (profiles)
              2000 Census tract-level data for the entire U.S. (sectors=tracts for the NAAQS ozone exposure application)
     Sector location data f
     (latitude, longitude) 1
Defined study area (sectors
within a city radius and with air
quality and meteorological data
within their radii of influence)
  Locations of air quality and
 meteorological measurements;
       radii of influence
Sector population data
  (age/gender/race)
  Commuting flow data
(origin/destination sectors)


-^


-N
i
>
r
1^
Population within
the study area




Age/gender/tract-specific
employment probabilities
                                                physiological
                                                distribution data (body
                                                weight, height, etc)
                                                                   Stochastic
                                                                profile generator
                Distribution functions for
                profile variables
                (e.g, probability of air
                conditioning)
                      Distribution functions
                      for seasonal and daily
                      varying profile variables
                      (e.g., window status, car
                      speed)

(
)
- National
database
(^ J) - Simulation <^
step ^"~~~--.

- Area-specific
input data
j> - Data processor




- Intermediate step
or data
±\ - Output data
                 A simulated individual with the
                 following profile:
                 • Home sector
                 • Work sector (if employed)
                 •Age
                 • Gender
                 • Race
                 • Employment status
                 • Home gas stove
                 • Home gas pilot
                 • Home air conditioner
                 • Car air conditioner
                 • Physiological parameters
                  (height, weight, etc.)
                                                                  4-8

-------
                                      Figure 4-1.  Overview of the APEX Model, continued
                                              4. Construct sequence of activity events
                                                   for each simulated individual
                             Diary events/activities
                            and personal information
                              (e.g., from CHAD)
                            Activity diary pools by day
                            type/temperature category
Profile for an
 individual
Stochastic diary
 selector using
age, gender, and
  employment
                                    Selected diary records for each day in the simulation
                                    period, resulting in a sequence of events
                                    (microenvironments visited, minutes spent, and
                                    activity) in the simulation period, for an individual
                          Each day in the simulation
                          period is assigned to an
                          activity pool based on day type
                          and temperature category
                              Maximum/mean daily
                                temperature data
          Stochastic
     calculation of energy
expended per event (adjusted for
  hysiological limits and EPOC)
        and ventilation
            rates
                                         Physiological
                                        parameters from
                                            profile
                                                                        Sequence of events for an
                                                                              individual
                                                                4-9

-------
                                         Figure 4-1. Overview of the APEX Model, concluded
                    5. Calculate concentrations in
                microenvironments for all events for
                      each simulated individual
                                              6. Calculate hourly
                                              exposures for each
                                              simulated individual
                                            7. Calculate
                                       population exposure
                                             statistics
 Microenvironments defined by
grouping of CHAD location codes
Hourly air quality
data for all sectors
                          Select calculation method for
                          each microenvironment:
                          • Factors
                          • Mass balance
          Calculate
      concentrations in all
      microenvironments
                                                                    I
                                            Average exposures
                                            for simulated person,
                                            stratified by ventilation
                                            rate:
                                            • Hourly
                                            • Daily 1-hour max
                                            • Daily 8-hour max
                                              Daily
                                       Population exposure
                                       indicators for:
                                       • Total population
                                       • Children
                                       • Asthmatic children
Hourly concentrations and
minutes spent in each
 microenvironment visited by
the simulated individual
                           Concentrations for all events
                           for each simulated individual
   Sequence of events for
 each simulated individual
                                                 Calculate hourly
                                                 concentrations in
                                                microenvironments
                                                     visited
                                                                 4-10

-------
       Activity-specific simulated breathing rates of individuals are used in APEX to
characterize intake received from an exposure. These breathing, or ventilation, rates are derived
from energy expenditure estimates for each activity included in CHAD and are adjusted for age-
and gender-specific physiological parameters associated with each simulated individual. Energy
expenditure estimates themselves are derived from METS (metabolic equivalents of work)
distributions associated with every activity in CHAD (McCurdy et al., 2000), largely based upon
the Ainsworth et al. (1993) "Compendium of Physical Activities." METS are a dimensionless
ratio of the activity-specific energy expenditure rate to the basal or resting energy expenditure
rate, and the metric is used by exercise physiologists and clinical nutritionists to estimate work
undertaken by individuals as they go through their daily life (Montoye et al., 1996). This
approach is discussed more thoroughly in McCurdy (2000).

       4.3.2 Key Algorithms
       Ozone concentrations in each microenvironment are estimated using either a mass-
balance or transfer factors approach, and the user specifies probability distributions for the
parameters that are used in the microenvironment model that reflect the observed variabilities in
the parameters.  These distributions can depend on the values of other variables calculated in the
model or input to APEX. For example, the distribution of AERs in a home, office, or car  can
depend on the type of heating and air conditioning present, which are also stochastic inputs to the
model, as well as the ambient temperature. The user can choose to keep the value of a stochastic
parameter constant for the entire simulation (which would be appropriate for the volume of a
house), or can specify that a new value shall be drawn hourly, daily, or seasonally from specified
distributions. APEX also allows the user to specify diurnal, weekly, or seasonal patterns for
various microenvironmental parameters. The distributions of parameters input to APEX
characterize the variability of parameter values, and are not intended to reflect uncertainties in
the parameter estimates.
       The mass balance method used within APEX assumes that the air in an enclosed
microenvironment is well-mixed and that the air concentration is fairly spatially uniform at a
given time within the microenvironment. The following four processes are modeled to  predict
the concentration of an air pollutant in such a microenvironment:
        .   Inflow of air into the microenvironment;
        .   Outflow of air from the microenvironment;
        .   Removal of a pollutant from the microenvironment due to deposition, filtration, and
           chemical degradation; and
        .   Emissions from sources of a pollutant inside the microenvironment.
                                          4-11

-------
       The transfer factors model is simpler than the mass balance model, however, still most
parameters are derived from distributions rather than single values, to account for observed
variability. It does not calculate concentration in a microenvironment from the concentration in
the previous hour and it has only two parameters, a proximity factor, used to account for
proximity of the microenvironment to sources or sinks of pollution, or other systematic
differences between concentrations just outside the microenvironment and the ambient
concentrations (at the measurements site), and a penetration factor, which quantifies the degree
to which the outdoor air penetrates into the microenvironment.  When there are no indoor
sources, the penetration factor is essentially the ratio of the concentration in the
microenvironment to the outdoor concentration.
       Regardless of the method used to estimate the microenvironmental concentrations, APEX
calculates a time series of exposure concentrations that a simulated individual experiences during
the modeled time period.  APEX estimates the exposure using the concentrations calculated for
each microenvironment and the time spent in each of a sequence of microenvironments visited
according to the "activity diary" of each individual.  The hourly average exposures of each
simulated individual are time-weighted averages of the within-hour exposures. From hourly
exposures, APEX calculates the time series of 8-hr and daily average exposures that simulated
individuals experience during the simulation period.  APEX then statistically summarizes and
tabulates the hourly, 8-hr, and daily exposures.

       4.3.3   Model Output
       There are several useful indicators of exposure and intake of people to 63 air pollution.
Factors that are important include the magnitude and duration of exposure, frequency of repeated
high exposures, and the breathing rate of individuals at the time of exposure. In this analysis,
exposure indicators include daily maximum  1-hr and 8-hr average 63 exposures, stratified by a
measure of the level of exertion at the time of exposure.  The level of exertion of individuals
engaged in particular activities is measured by an equivalent ventilation rate (EVR), ventilation
normalized by body surface area (BSA, in m2), which is calculated as VE/BSA, where VE is the
ventilation rate (liters/minute).  Table 4-1 lists the ranges of EVR corresponding to moderate and
heavy levels of exertion.

     Table 4-1. Exertion levels in terms of equivalent ventilation rates (liters/min-m2)
              Averaging time       Moderate exertion   Heavy exertion
                    1 hour             16-30 EVR           > 30 EVR
                    8 hour             13-27 EVR           > 27 EVR
              fromWhitfieldetal., 1996, page 15.
                                          4-12

-------
       APEX calculates two general types of exposure estimates: counts of the estimated
number of people exposed to a specified Os concentration level and the number of times per Os
season that they are so exposed; the latter metric is in terms of person-occurrences or person-
days.  The former highlights the number of individuals exposed one or more times per Os season
to the exposure indicator of interest.  In the case where the exposure indicator is a benchmark
concentration level, the model estimates the number of people who are expected to experience
exposures to that level of air pollution, or higher,  at least once during the modeled period.  APEX
also reports counts  of individuals with multiple exposures.  The person-occurrences measure
estimates the number of times  per season that individuals are exposed to the exposure indicator
of interest and then accumulates these estimates for the entire population residing in an area.
This metric conflates people and occurrences: one occurrence for each of 10 people is counted
the  same as 10 occurrences for one person.
       APEX tabulates and displays the two measures for exposures  above levels ranging from
0.0  to 0.16 ppm by  0.01 ppm increments, where the exposures are:
    •   Daily maximum 1-hour average exposures
    •   Daily maximum 8-hour average exposures
    •   Daily average exposures.
These results are tabulated for the following population groups:
    •   All ages and activity levels
    •   Children at  all activity  levels
    •   Asthmatic children.
Separate output tables are produced for different levels of exertion concomitant with the
exposures:
    •   All exertion levels
    •   Moderate exertion levels
    •   Heavy exertion levels.
APEX also produces tables of the time spent in different microenvironments, stratified by
exposure levels.

       4.3.4   Strengths and  Limitations of the Model
       APEX has a strong scientific foundation and incorporates several significant algorithmic
improvements and updates to input data since it's predecessor, pNEM, was used in the last
review. In this section we discuss qualitatively some of the general strengths and limitations of
the  application of APEX to model population exposures to 63 pollution.  The discussion is
divided into four fundamental  areas: estimation of ambient air quality, estimation of
concentrations in microenvironments, characterization of population demographics  and activity
patterns, and modeling physiological processes as applicable to this exposure assessment.
                                          4-13

-------
Additional details for advancements made to the model of specific relevance for this ozone
exposure assessment are given in section 4.5.
       In general, limitations and uncertainties result from variability not modeled or modeled
incorrectly, erroneous or uncertain inputs, errors in coding, simplifications of physical, chemical,
and biological processes to form the conceptual model, and flaws in the conceptual model. The
implications of these limitations for the uncertainty of the APEX results is discussed in Langstaff
(2007).

       4.3.4.1 Estimation of Ambient Air Quality
       For estimating ambient Os concentrations to use in the exposure model, the urban areas
modeled here have several monitors measuring hourly 63 concentrations. Having multiple
monitors in the simulated areas collecting time-resolved data allows for the utilization of APEX
spatial and temporal capabilities in estimating exposure.  Since APEX uses actual records  of
where individuals are located at specific times of the day, more realistic exposure estimates are
obtained in simulating the contact of individuals with these spatially and temporally diverse
concentrations. Primary uncertainties in the air quality data input to the model result from
estimating concentrations at locations which may not be in close proximity to monitoring  sites
(as estimated by  spatial interpolation of actual data points) and from the method used to estimate
missing data. In addition, concentrations of 63 near roadways are particularly difficult to
estimate due to the rapid reaction of Os with nitric oxide  emitted from motor vehicles.
       We have  modeled the Os seasons for 2002, 2003, and 2004, to better account for year-to-
year variability of air quality and meteorology. For most of the 12 areas modeled, 63
concentrations were lower in 2004 than previous years, due to a combination of reduced
emissions of precursors and weather patterns less conducive to the formation of Os. Having this
wide range of air quality data across multiple years available for use in the exposure simulation
has a direct impact on more realistically estimating the range of exposures, rather than using a
single year of air quality data.
       Modeling exposures for an unspecified future year simulated to just meet alternative air
quality standards has, in addition to the uncertainties involved with modeling historical
scenarios, the uncertainties of the complex process of projecting to future years air quality,
population demographics, activity patterns, and other changing parameters. For the purpose of
estimating population exposure as an input to decisions about the appropriate level of a NAAQS,
EPA has historically not incorporated any projections in population demographics, activity
patterns, or other factors (e.g., air conditioning use, changes  in housing types, etc). This allows
policy makers to focus on the impact of changing the allowed air quality distribution on
                                          4-14

-------
population exposure and public health while avoiding the additional uncertainties that inclusion
of these other factors would introduce.

       4.3.4.2 Estimation of Concentrations in Indoor Microenvironments
       The importance of estimation of concentrations in indoor microenvironments (e.g.,
homes, offices, schools, restaurants, vehicles) is underscored by the finding that personal
exposure measurements of Os are often not well-correlated with ambient measurements (CD,
pages 3-59 to 3-61).
       APEX has been designed to better estimate human exposure through use of algorithms
that attempt to capture the full range of Os concentrations expected within several important
microenvironments. Parameters used to estimate the concentrations in microenvironments can
be highly variable, both between microenvironments (e.g., different houses have varying
characteristics) and within microenvironments (e.g., the characteristics of a given house can vary
over time). Since APEX is a probabilistic model, if data accurately characterizing this variability
are provided to the model, then such variabilities would not result in uncertainties in the
estimation of the microenvironmental concentrations. Thus, it is the input data used in
development of the parameters that are the limiting factor, and to date, APEX uses the most
current available data to develop required input parameters for estimation of microenvironmental
concentrations.
Air Exchange Processes
       The AER is the single most important factor in determining the relationship between
outdoor and indoor concentrations of Os. AERs are highly variable, both within a
microenvironment over time and between microenvironments of the same type. AERs depend
on the physical characteristics of a microenvironment and also on the behavior of the occupants
of the microenvironment. There is a strong dependence on temperature, and some dependence
on other atmospheric conditions.  APEX uses probabilistic distributions of AERs which were
derived from several measurement studies in a number of locations, and are stratified by both
temperature and the presence or absence of air conditioning. These two variables are the most
influential variables influencing AER distributions (see Appendix A of the Exposure Analysis
TSD).
Removal Processes
       Concentrations within indoor microenvironments can be reduced due to removal
processes such as deposition to  surfaces and by reaction with other chemicals in the air. The rate
of deposition of Os to a surface depends on the surface composition, the humidity, and the
concentration of Os. The rate of removal of Os from a microenvironment depends on its
dimensions, the ratio of surface area to volume, and of course the presence, amount, and type of
                                          4-15

-------
surface coverings and furnishings in the microenvironment. Deposition is modeled
probabilistically in APEX by a using a distribution of decay rates derived from a study that
measured decay rates in 26 homes in Southern California (Lee et al., 1999). Although it is not
expected that inter-city differences in decay rates would be more important than differences
between homes within cities, there is uncertainty given the small sample size the distribution was
derived from.  The lack of capturing inter-city variability in decay rates is not anticipated to be a
large contributor to the uncertainty in the modeling results. There can be additional O3 loss,
which is not currently modeled, due to the use of HVAC systems, which significantly increase
the effective surface area as air recirculates through ductwork and filters.
       Ozone reacts with a number of indoor pollutants, such as nitric oxide from gas stoves and
VOCs from consumer products. With the exception of nitric oxide, O3 reacts slowly with most
indoor pollutants, and this is typically a less influential removal process than air exchange and
surface removal (Weschler, 2000).  The lack of a better treatment of indoor air chemistry is not
considered to be a significant limitation of APEX for modeling Os.

       4.3.4.3 Characterization of Population Demographics and Activity Patterns
       The approach to reasonably estimating exposure considering a variety of alternative
scenarios is best done using models that better represent the contact of a human with the
contaminant of concern. By using actual time-location-activity  diaries that capture the duration
and frequency of occurrence of visitations/activities performed, APEX can simulate expected
variability in human behavior, both within and between individuals.  Fundamentals of energy
expenditure are then used to estimate relative intensity of activities performed.  This, combined
with microenvironmental concentrations, allows for the reasonable estimation of the magnitude,
frequency,  pattern, and duration of exposures an individual experiences.
       CHAD is the best source of human activity data for use in exposure modeling. The
database contains time-location-activity patterns for individuals of both genders across a wide
range of ages (0-99).  The database is geographically diverse, containing diaries from  individuals
residing in  major cities, suburban and rural areas across the U.S. Time spent performing
activities within particular locations can be on a minute-by minute basis, thus avoiding the
smoothing  of potential peak exposures longer time periods would give.
       There are some  limitations to the database, however, many of which are founded in the
individual studies from which activity patterns were derived (Graham and McCurdy, 2004).  A
few questions remain regarding the representativeness of CHAD diaries to  the simulated
population, such as the  numbers of diaries available for use in a simulation (i.e., 20,000 used to
represent several million people over long periods of time), the age of diary data (i.e.,  some data
were generated in the 1980s), and diary structure differences (i.e., real-time versus recall method
                                          4-16

-------
of data collection). Many of the assumptions about use of these activity patterns in exposure
modeling are strengthened by the manner in which they are used by APEX, through focusing on
the most important individual attributes that contribute to variability in human behavior (e.g.,
age, gender, time spent outdoors, day of week, ambient temperature, occupation).
       The extent to which the human activity database provides a balanced representation of
the population being modeled is likely to vary across areas. Although the algorithm that
constructs activity sequences accounts to some extent for the effects of population demographics
and local climate on activity, this adjustment procedure may not account for all inter-city
differences in people's activities. A new methodology has been  developed to more appropriately
assign individual diaries to reflect time-location-activity patterns in simulated individuals
(discussed further in section 4.5.3).  Input distributions used in the new procedure for
constructing multi-day activity patterns are based on longitudinal activity data from children of a
specific age range (appropriate for this application where similar aged children are simulated),
however the data used were limited to one study and may not be appropriate for other simulated
individuals.  Thus, there are limitations in approximating within-person variance and between-
person variance for certain variables (e.g., time spent outdoors).  Personal activity patterns are
also likely to be affected by many local factors, including topography, land use, traffic patterns,
mass transit systems, and recreational opportunities, which are not incorporated in the current
exposure analysis approach due to the complexity of scale and lack of data to support the
development of a reasonable approach.

       4.3.4.4 Modeling Physiological Processes
       The modeling of physiological processes that are relevant to the exposure and intake of
Os is a complicated endeavor, particularly when attempting to capture inter- and intra-personal
variability in these rates.  APEX has a physiological module capable of estimating ventilation
rates (VE) for every activity performed by an individual, which primarily drives Os intake dose
rate estimates. See section 2.5 of the draft Exposure Assessment TSD for a discussion of this
module.  Briefly, the module is based on the relationship between energy expenditure and
oxygen consumption rate, thus both within- and between-person variability  in ventilation can be
addressed through utilization of the unique sequence of events individuals go through each
simulated day. These activity-specific VE estimates, when normalized by BSA, are then used to
characterize an individual's exertion level in compiling the summary exposure tables (Table
4-1).  One of the key determinants of estimated VE is the exertion level of an individual's
activity, where exertion levels have units of metabolic equivalents of work (MET), which is the
ratio of energy expenditure for an activity to the person's basal, or resting, metabolic rate.
                                           4-17

-------
       There are some limitations in using MET values for this purpose, due mostly to the
manner in which the time-location-activity diaries were generated and subsequent estimates of
exertion level. An individual (or their caregiver if younger than eight years old) would record
the activity performed with a start and end time, with no information on the associated exertion
level of the activity. Exertion level (MET) was then inferred by developers of the CHAD
database (McCurdy et al., 2000) using standard values and distributions of those values reported
by a expert panel of exercise physiologists (Ainsworth et al., 1993).  Although this approach
allows for an appropriate range of exertion levels to be assigned to the individuals' activities
(and to the simulated population), children's activity levels fluctuate widely within a single
activity category; their pattern is often characterized as having bursts of high  energy expenditure
within a longer time frame of less energy expenditure (Freedson, 1989).  These fluctuations in
energy expenditure that occur within an activity (and thus a simulated event)  are not well
captured by the MET assignment procedure.

4.4   SCOPE OF  EXPOSURE ASSESSMENT
       4.4.1   Selection of Urban Areas to be Modeled
       The selection of urban areas to include in the exposure analysis takes  into consideration
the location of Os epidemiological studies, the availability of ambient Os data, and the desire to
represent a range of geographic areas, population demographics, and 63 climatology.  These
selection criteria are discussed further in Chapter 5. Based on these criteria, we chose the 12
urban areas listed in Table 4-2 to develop population exposure estimates.3  The geographic  extent
of each modeled area consists of the census tracts in the combined statistical  area (CSA) as
defined by OMB (OMB, 2005).  Maps of the modeled areas are presented in  Appendix 4-A.

       4.4.2   Time Periods Modeled
       The exposure periods modeled are the Os seasons for which routine hourly Os monitoring
data are available.  These periods include most of the high-ozone events in each area.  The
seasons modeled for each area are listed in Table 4-2.

       4.4.3   Populations Modeled
       Exposure modeling was conducted for the general population residing in each area
modeled, as well as for school-age children (ages 5 to 18) and asthmatic school-age children.
Due to the increased amount of time spent outdoors engaged in relatively high levels of physical
activity (which increases intake), school-age children as a group are particularly at risk for
experiencing Os-related health effects. We report results for school-age children down to age
       1 In the remainder of this chapter the city name in bold in Table 4-2 is used to represent the entire CSA.
                                          4-18

-------
Table 4-2. Urban areas and time periods modeled
Urban Area (CSA)                                   Period modeled
Atlanta-Sandy Springs-Gainesville, GA-AL             March 1 to Oct. 31
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 Oct. 31
Detroit-Warren-Flint, MI                             April  1 to Sept. 30
Houston-Baytown-Huntsville, TX                     Jan. 1 to Dec. 30
Los Angeles-Long Beach-Riverside, CA                Jan. 1 to Dec. 30
New York-Newark-Bridgeport, NY-NJ-CT-PA          April  1 to Sept. 30
Philadelphia-Camden-Vineland, PA-NJ-DE-MD         April  1 to Oct. 31
Sacramento-Arden-Arcade-Truckee, CA-NV          Jan. 1 to Dec. 30
St. Louis-St. Charles-Farmington, MO-JL               April  1 to Oct. 31
Washington-Baltimore-N. Virginia, DC-MD-VA-WV    April  1 to Oct. 31
five, however, there is a trend for younger children to attend school.  Some states allow 4-year-
olds to attend kindergarten, and more than 40 states have preschool programs for children
younger than five (Blank and Mitchell, 2001). In 2000, six percent of U.S. children ages 3 to 19
who attend school were younger than five years old (2000 Census Summary File 3, Table QT-
P19: School Enrollment). We are not taking these younger children into account in our analysis
due to a lack of information which would let us characterize this group of children.
       The population of asthmatic children is estimated for each city using asthma prevalence
data from the National Health Interview Survey (NHIS) for 2003 (Dey and Bloom, 2005).
Asthma prevalence rates for children aged 0 to 17 years were calculated for each age, gender,
and geographic region. The regions defined by NHIS are "Midwest," "Northeast," "South," and
"West." For this analysis, asthma prevalence was defined as the probability of a "Yes" response
to the question: "Ever been told that... had asthma?" among those that responded "Yes" or "No"
to this question. The responses were weighted to take into account the complex survey design of
the NHIS survey. Standard errors and confidence intervals for the prevalence were calculated
using a logistic model, taking into account the survey design.  A scatter plot smoothing technique
using the LOESS smoother was applied to smooth the prevalence curves and compute the
standard errors and confidence intervals for the smoothed prevalence estimates.  Logistic
analysis of the prevalence curves shows statistically significant differences in prevalence by
                                         4-19

-------
gender and by region. Therefore we did not combine the prevalence rates for different genders
or regions. A detailed description of this analysis is presented in the Exposure Analysis TSD.
       The "modeled population" column in Table 4-3 lists the year 2000 populations of the
modeled CSAs. The 12 modeled areas combined represent 40 percent of the total U.S. urban
population (approximately 222 million in 2000). Table 4-3 also gives the modeled populations
of children ages 5-18 and children ages  5-18 characterized as asthmatic.

Table 4-3. Population coverage of modeled areas (2002 analysis)
Urban Area (CSA)
Atlanta
Boston
Chicago
Cleveland
Detroit
Houston
Los Angeles
New York
Philadelphia
Sacramento
St. Louis
Washington, DC
Total of all 12 areas
Modeled Modeled Asthmatic
population children1 children
(thousands) (thousands) (thousands)
4,548
5,714
9,311
2,945
5,357
4,815
16,371
21,357
5,832
1,930
2,754
7,572
88,506
943
1,096
1,951
594
1,110
1,089
3,667
4,147
1,186
412
582
1,485
18,262
117
182
279
89
162
136
457
643
193
51
83
187
2579
 ages 5-18.

4.5  INPUTS TO THE EXPOSURE MODEL
       The data inputs to the APEX model are briefly described in this section.  A more detailed
description of the development of these data and the derivation of input distributions can be
found in the Exposure Analysis TSD.

       4.5.1  Population Demographics
       APEX takes population characteristics into account to develop accurate representations of
study area demographics. Population counts and employment probabilities by age and gender
are used to develop representative profiles of hypothetical individuals for the simulation.  Tract-
level population counts by age in one-year increments, from birth to 99 years, come from the
                                         4-20

-------
2000 Census of Population and Housing Summary File 1.  The Summary File 1 contains the 100-
percent data, which is the information compiled from the questions asked of all people and about
every housing unit.
       Employment data from the 2000 Census provide employment probabilities for each
gender and specific age groups for every Census tract. The employment age groupings are: 16-
19, 20-21, 22-24, 25-29, 30-34, 35-44, 45-54, 55-59, 60-61, 62-64, 65-69, 70-74, and >75 years
of age. Children under the age of 16 are assigned employment probabilities of zero.

       4.5.2  Population Commuting Patterns
       To ensure that individual's daily activities are accurately represented within APEX, it is
important to integrate working patterns into the assessment. The APEX commuting data are
derived from the 2000 Census and collected as part of the Census Transportation Planning
Package (CTPP). CTPP contains tabulations by place of residence, place of work, and the flows
between the residence and work. These data are available  from the U.S. Department of
Transportation, Bureau of Transportation Statistics (U.S. Department of Transportation and U.S.
Census Bureau, 2000).
       For school age children we have not included commuting to and from school. This
results in the implicit assumption that children attend a school with ambient Os concentrations
similar to concentrations near their residence. To the extent that the highest ozone levels are
generally in the period June through August when most students are not in school, the absence of
school commuting is less likely to have a significant impact on the exposure estimates. As more
communities go to year-round schools,  school commuting  patterns may become important to
model.
       It is assumed that all persons with home-to-work distances up to 120 km are  daily
commuters, and that persons who travel further than 120 km do not commute daily.  Therefore
the list of commuting destinations for each home tract is restricted to only those work tracts that
are within 120 km of the home tract.
       APEX allows the user to specify how to handle individuals who commute to destinations
outside the study area. One option is to drop them from the simulation. If they are included, the
user specifies values for two additional  parameters, called LM and LA (Multiplicative and
Additive factors for commuters who Leave the area). While a commuter is at work, if the
workplace is outside the study area, then the ambient concentration cannot be determined from
any air district (since districts are inside the study area).  Instead, it is assumed to be related to
the average concentration CAvs(t) over  all air districts at the time in question.  The ambient
concentration outside the study area at time t, Cour(i), is estimated as:

       COUT (t) = LM * CAVE (0  + LA                                               (4-2)
                                         4-21

-------
       The microenvironmental concentration (for example, in an office outside the study area)
is determined from this ambient concentration by the same model (mass balance or factor) as
applies inside the study area. The parameters LM and LA were both set to zero for this modeling
analysis; thus, exposures to individuals are set to zero when they are outside of the study area.
Although this tends to underestimate exposures, it is a small effect and this was done since we
have not estimated ambient concentrations of 63 in counties outside of the modeled areas.

       4.5.3   Human Activity Data
       The human activity data are drawn from the most recent version (December 2000) of the
Consolidated Human Activity Database (CHAD) (McCurdy et al., 2000; EPA, 2002), developed
and maintained by the Office of Research and Development's (ORD) National Exposure
Research Laboratory (NERL).  The CHAD includes data from several surveys covering specific
time periods at city, state, and national levels, with varying degrees of representativeness. Table
4-4 summarizes the studies in CHAD used in this modeling analysis, providing nearly 16,000
diary-days of activity data (3,075 diary-days for ages 5-18) collected between 1982 and 1998.
       A key issue in this assessment is the development of an approach for creating (Vseason
or year-long activity sequences for individuals based on a cross-sectional activity data base of
24-hour records. The typical subject in the time/activity studies in CHAD provided less than two
days of diary data.  For this reason, the construction of a season-long activity sequence  for each
individual requires some combination of repeating the same data from one subject and using data
from multiple subjects. An appropriate approach should adequately account for the day-to-day
and week-to-week repetition of activities common to individuals while maintaining realistic
variability between individuals. The method in APEX for creating longitudinal diaries  was
designed to capture the tendency of individuals to repeat activities, based on reproducing  realistic
variation in a key diary variable, which is a user-selected function of diary variables.  For this
analysis the key variable is set to the amount of time an individual spends outdoors each day,
which is one of the most important determinants of exposure to high levels of Os.
       The actual diary construction method targets two statistics, a population diversity  statistic
(D) and a within-person autocorrelation statistic (A).  The D statistic reflects the relative
importance of within-person variance and between-person variance in the key variable. The A
statistic quantifies the lag-one (day-to-day) key variable autocorrelation.  Desired D and A values
for the key variable are selected by the user and set in the APEX parameters file, and the method
algorithm constructs longitudinal diaries that preserve these parameters. Longitudinal diary data
from a field study of school-age children (Geyh et al., 2000) and subsequent analyses (Xue et al.,
2004) suggest that D and A are stable over time (and perhaps over cohorts as well). Based on
these studies of children ages 7-12, appropriate target values for the two statistics for outdoor
                                          4-22

-------
Table 4-4.  Studies in CHAD used in this analysis
Study name
Baltimore
California
Adolescents (CARB)
California Adults
(CARB)
California Children
(CARB)
Cincinnati (EPRI)
Denver (EPA)
Los Angeles:
Elementary School
Los Angeles: High
School
National: NHAPS-Air
National: NHAPS-
Water
Washington, D.C.
(EPA)
Total diary days
Geographic
coverage
One building
in Baltimore
California
California
California
Cincinnati
metro, area
Denver
metro, area
Los Angeles
Los Angeles
National
National
Wash., D.C.
metro, area

Study time
period
01/1997-02/1997,
07/1998-08/1998
10/1987-09/1988
10/1987-09/1988
04/1989- 02/1990
03/1985-04/1985,
08/1985
11/1982-02/1983
10/1989
09/1990-10/1990
09/1992-10/1994
09/1992-10/1994
11/1982-02/1983

Subject
ages
72-93
12-17
18-94
<1- 11

-------
time are determined to be 0.2 for D and 0.2 for A In the absence of data for estimating these
statistics for younger children and others outside the study age range, these values are used for all
ages. This new method for constructing longitudinal diaries from the CHAD data is described in
detail in the Exposure Analysis TSD.

       4.5.4   Physiological Data
       APEX requires values for various physiological parameters for subjects in order to
accurately model their metabolic processes that affect pollutant intake.  This is because
physiological differences may cause people with the same exposure and activity scenarios to have
different pollutant intake levels. The physiological parameters file distributed with  APEX
contains physiological data or distributions by age and gender for maximum ventilatory capacity
(in terms of age- and gender-specific maximum oxygen consumption potential), body mass,
resting metabolic rate, and oxygen consumption-to-ventilation rate relationships.

       4.5.5   Microenvironments Modeled
       In APEX, microenvironments provide the exposure locations for modeled individuals.  For
exposures to be accurately estimated, it is important to have realistic microenvironments that are
matched closely to where people are physically located on a daily and hourly basis.  As discussed
in section 4.3.2 above, the two methods available in APEX for calculating pollutant concentrations
within microenvironments are a mass balance model and a transfer factor approach. Table 4-5
lists the 12 microenvironments selected for this analysis and the exposure calculation method for
each. The parameters used in this analysis for modeling these microenvironments are described in
this section.

       4.5.5.1 Air Exchange Rates for Indoor Residential Environments
       Distributions of AERs for the indoor microenvironments were developed using data from
several studies. The analysis of these data and the development of the distributions  used in the
modeling are described in detail in the Exposure Analysis TSD.  This analysis showed that the
AER distributions for the residential microenvironments depend on the type of air conditioning
(A/C) and on the outdoor temperature, as well as other variables for which we do not have
sufficient data to estimate. This analysis clearly demonstrates that the AER distributions vary
greatly across cities  and A/C types and temperatures, so that the selected AER distributions for the
modeled cities should also depend upon the city,  A/C type, and temperature. For example, the
mean AER for residences with A/C ranges from 0.39 for Los Angeles between 30 and 40 °C to
1.73 for New York between 20 and 25 °C. The mean AER for residences without A/C ranges
from 0.46 for San Francisco on days with temperature between 10 and 20 °C to 2.29 for
                                          4-24

-------
Table 4-5.  Microenvironments modeled by APEX
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
In-vehicle - Mass Transit
Calculation Method
Mass balance
Mass balance
Mass balance
Mass balance
Mass balance
Mass balance
Mass balance
Factors
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
PE and PR
    1 AER=air exchange rate, DE=decay-deposition rate, PR=proximity factor, PE=penetration factor

New York on days with temperature between 20 and 25 °C. The need to account for the city as
well as the A/C type and temperature is illustrated by the result that for residences with A/C on
days with temperature between 20 and 25 °C, the mean AER ranges from 0.52 for Research
Triangle Park to 1.73 for New York.  For each combination of A/C type, city, and temperature
with a minimum of 11 AER values, exponential, lognormal, normal, and Weibull distributions
were fit to the AER values and compared.  Generally, the lognormal distribution was the best-
fitting of the four distributions, and so, for consistency, the fitted lognormal distributions are
used for all the cases.
       One limitation of this analysis was that distributions were available only for selected
cities, and yet the summary statistics  and comparisons demonstrate that the AER distributions
depend upon the city as well as the temperature range and A/C type. Another important
limitation of the analysis was that distributions were not able to be fitted to all  of the temperature
ranges due to limited data in these ranges.  A description of how these limitations were addressed
can be found in the Exposure Analysis TSD.
       City-specific AER distributions were used where possible; otherwise data for a similar
city were used. We obtained estimates of A/C prevalence from the American Housing Survey
(AHS, 2003) for each metropolitan area. The final AER distributions used for the exposure
modeling are given the Exposure Analysis TSD.
                                          4-25

-------
       Some residences, particularly in the Southwest, use evaporative coolers, also known as
"swamp coolers," for cooling. We performed an analysis of AER distributions of residences
without A/C, with and without evaporative coolers, using data from three AER measurement
studies. This comparison is described in Appendix F in the Exposure Analysis TSD.  This
analysis showed no improvement in the statistical air exchange model when the data were also
stratified by evaporative cooler presence or absence, given that they are already stratified by
CSA, air conditioner presence or absence,  and outdoor temperature range.

       4.5.5.2 AER Distributions for Other Indoor Environments
       To estimate AER distributions for non-residential,  indoor environments (e.g.,  offices and
schools), we obtained and analyzed two AER data sets: "Turk" (Turk et al.,  1989); and "Persily"
(Persily and Gorfain, 2004; Persily et al., 2005). The earlier "Turk" data set (Turk et al., 1989)
includes 40 AER measurements from offices (25 values), schools (7 values), libraries (3  values),
and multi-purpose buildings (5 values), each measured using an SF6 tracer over two or four hours
in different seasons of the year. The more recent "Persily" data (Persily and Gorfain, 2004;
Persily et al., 2005) were derived from the U.S. EPA Building Assessment Survey and
Evaluation (BASE) study, which was conducted to assess indoor air quality, including
ventilation, in a large number of randomly selected office buildings throughout the U.S.  This
data base consists of a total of 390 AER measurements in 96 large, mechanically ventilated
offices. AERs were measured both by a volumetric method and by a  CC>2 ratio method, and
included their uncertainty estimates.  For these analyses, we used the recommended "Best
Estimates" defined by the values with the lower estimated uncertainty; in the vast majority of
cases the best estimate was from the volumetric method.
       Due to the small sample size of the Turk data, the data were analyzed without
stratification by building type and/or season. For the Persily data, the AER values for each office
space were averaged, rather using the individual measurements, to account for the strong
dependence of the AER measurements for the same office space over a relatively short period.
The mean values are similar for the two studies, but the standard  deviations are about twice as
high for the Persily data. The proposed AER distributions were derived from the more recent
Persily data only.
       We fitted exponential, lognormal, normal, and Weibull distributions to the 96 office
space average AER values, and the best fitting of these was the lognormal.  The fitted parameters
for this distribution,  used for AER distributions for the indoor non-residential
microenvironments, can be found in the Exposure Analysis TSD.
                                      4-26

-------
       4.5.5.3 Proximity and Penetration Factors for Outdoors and In-vehicle
             Microenvironments
       For the outdoors near-road, public garage/parking lot, and in-vehicle proximity factors,
and for the in-vehicle penetration factors, we use distributions developed from the Cincinnati
Ozone Study (American Petroleum Institute,  1997, Appendix B; Johnson et al., 1995). This field
study was conducted in the greater Cincinnati metropolitan area in August and September, 1994.
Vehicle tests were conducted according to an experimental design specifying the vehicle type,
road type, vehicle speed, and ventilation mode. Vehicle types were defined by the three study
vehicles:  a minivan, a full-size car, and a compact car.  Road types were interstate highways
(interstate), principal urban arterial roads (urban), and local roads (local). Nominal vehicle
speeds (typically met over one minute intervals within 5 mph) were at 35 mph, 45 mph, or 55
mph. Ventilation modes were as follows:
   •  Vent Open:  Air conditioner off. Ventilation fan at medium. Driver's window half open.
       Other windows closed.
   •  Normal A/C: Air conditioner at normal. All windows closed.
   •  Max A/C: Air conditioner at maximum. All windows closed.
Ozone concentrations were measured inside the vehicle, outside the vehicle, and at six fixed-site
monitors  in the Cincinnati area.
       The Exposure Analysis TSD documents the distributions and the rationale for the
selection  of distributions of penetration and proximity factors for outdoors and in-vehicle
microenvironments used in this modeling analysis.

       4.5.5.4 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"1. 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.

       4.5.6  Meteorological Data
       Daily average and maximum 1-hour temperatures were obtained from hourly surface
temperature  measurements obtained from the National Weather Service.  APEX uses the data
from the  closest weather  station to each Census tract.  Temperatures are used in APEX both in
selecting  human activity  data and in estimating AERs for indoor microenvironments.
                                      4-27

-------
       4.5.7  Ambient Ozone Concentrations
       APEX requires hourly ambient O3 concentrations at a set of locations in the study area.
Data from EPA's AIRS Air Quality Subsystem were used to prepare the ambient air quality
input files for 2002, 2003, and 2004. The hourly Os concentrations at the AIRS sites in each
CSA 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
adjusted by proximity factors, as described in the Exposure Analysis TSD.

       4.5.8  Modeling Alternative Standards
       In addition to modeling exposures based on historical air quality, an analysis was
conducted using air quality representative of just meeting the current 8-hr 63 standard of 0.08
ppm, and considering the previous rounding convention that allowed for concentrations up to
0.084 ppm. Several alternative standards, reflecting different combinations of standard levels
and form were also considered. Alternatives examined were intended to reflect improved
precision in the measurement of the ambient concentrations, where the precision would extend to
three instead of two decimal places (in ppm) (e.g., 0.080 ppm rather than 0.08 ppm) for several
different levels.  Differing forms of the standard were also explored outside of the average 4th
daily maximum 8-hr average scenario currently used. For example, a 3rd-highest form (the
average of the annual 3rd-highest daily maximum 8-hour concentrations averaged over the three
year period) was considered for 0.084 and 0.074 ppm levels, and a 5th-highest form for the 0.074
ppm level (the average of the annual 5th-highest daily maximum 8-hour concentrations averaged
over the three year period). These alternative scenarios are modeled using a quadratic rollback
approach to adjust the hourly O3 concentrations observed in 2002-2004 to yield a design value
corresponding to the standard being modeled. Table 4-6 shows the attainment thresholds (to
which the design values are rolled back), the form of the standard used for each scenario, and the
notation used in the remainder of this chapter. Design values for the current 8-hr Os standard are
calculated as the 3-year averages of the annual 4th-highest daily maximum 8-hr average
concentration based on the maximum monitor within an urban area. These are given in Table
4-7 for the 2002-2004 period.
       The quadratic rollback technique combines both linear and quadratic elements to reduce
higher concentrations more than lower concentrations near ambient background levels. The
quadratic rollback adjustment procedure was considered in a sensitivity analysis during the last
review of the Os NAAQS and has been shown to be more realistic than the linear proportional
rollback method, where all of the  ambient measurements are reduced by a constant multiplicative
factor regardless of their individual magnitudes. The quadratic rollback approach and evaluation
                                      4-28

-------
of this approach are described by Johnson (1997), Duff, Horst, and Johnson (1998), and Rizzo
(2005, 2006).
Table 4-6. Alternative 8-hr ozone standard scenarios
Attain. Threshold
0.084 ppm
0.080 ppm
0.074 ppm
0.070 ppm
0.064 ppm
Form of Standard
3rd-highest form
4th-highest form
4th-highest form
3rd-highest form
4th-highest form
5th-highest form
4th-highest form
4th-highest form
Notation
84/3
84/4
80/4
74/3
74/4
74/5
70/4
64/4
Table 4-7. 2002-2004 8-hr ozone design values for the modeled areas
Urban Area (CSA)
Atlanta
Boston
Chicago
Cleveland
Detroit
Houston
Los Angeles
New York
Philadelphia
Sacramento
St. Louis
Washington, DC
2002-2004 design
value1 (ppm)
0.093
0.091
0.094
0.095
0.092
0.101
0.127
0.094
0.094
0.102
0.089
0.089
Ratio of 0.084 to
the design value
0.90
0.92
0.89
0.88
0.91
0.83
0.66
0.89
0.89
0.82
0.94
0.94
 These design values are calculated based on the entire CSA, which in some
cases differ slightly from current non-attainment area definitions.
                                       4-29

-------
4.6  MODEL EVALUATION, SENSITIVITY, AND UNCERTAINTY ANALYSES
       The methods used to evaluate the APEX model and characterize the uncertainty of the
model predictions are described in this section. First, we discuss the results of the exposure
modeling in comparison to the modeling performed as part of the previous review of the 63
NAAQS, completed in 1997.
       Second, we report the results of a limited APEX model evaluation, which involves the
comparison of model exposure estimates to personal exposure measurement data. Although the
available exposure measurements are 6-day averages, this comparison can serve to gain insight
on whether the model is reasonably estimating exposures.
       Another approach for evaluating the model involves model sensitivity analyses,
evaluating how the model responds to variation in the input data and parameters used in several
of the key algorithms.  An analysis such as this is important for several reasons, such as
indicating those data that have the greatest impact on estimated exposures and the relative
confidence that one has in the model estimates as measured by the degree of certainty in the
model inputs and their appropriate use.
       A comprehensive analysis of the uncertainties of the exposure modeling was performed,
and the results of that analysis are described here. At the end of this section, the conclusions
drawn from the evaluation, sensitivity, and uncertainty analyses are summarized.

       4.6.1   Comparison with Exposure Estimates from the Prior Review
       There have been significant improvements to the exposure model  and the model inputs
since the review in 1997, as discussed in Section 4.3.1. In the previous review, six urban areas
were modeled using the pNEM model, Houston, Los Angeles, New York, Philadelphia, St.
Louis, and Washington (US EPA, 1996a,c).  These six cities (as well as six others) are also
modeled in the current review, although the geographic areas modeled are larger than in the
previous review, with over twice the population coverage.  When modeling a larger area,
extending well beyond the urban core, there will be more people exposed, but a smaller
percentage of the modeled population will be exposed at high levels, if 63 concentrations are
lower in the extended areas. Typical years, in terms of Os air quality, were modeled in the 1997
review (1990 for some cities and 1991 for others).  The only alternative standard for which we
have results for both reviews is the "84/3" standard. Exposures to children who tend to spend
more time outdoors were estimated in the previous review but not in the current review, and
there is no population group for which we can make a  direct comparison of the exposure
estimates for the two reviews. Therefore, a quantitative comparison of the exposure results is not
appropriate.
                                      4-30

-------
       4.6.2  Comparison of Model Estimates with Measured Personal Exposures
       APEX simulation results were compared to personal O3 concentration measurements
obtained from the Harvard Southern California Chronic Ozone Exposure Study (Xue et al. 2005,
Geyh et al. 2000). Although this study was limited in scope, and the measurements of ozone are
averaged over 6 days, it is the only study found that measured enough personal exposures to Os
to be useful for this evaluation and for which the data are available. In this study, children 7 to
12 years old were monitored from June 1995 to May 1996. There were 160 subjects on which
longitudinal Os concentrations were collected in at least 6 of the 12 months of the study period.
Passive 63 samplers were used to measure 6-day average personal 63 concentrations, as well as
indoor and outdoor concentrations at participants'  homes, for six days each month. The subjects
resided in two separate areas of San Bernardino County: urban Upland CA, and the small
mountain towns of Lake Arrowhead, Crestline, and Running Springs, CA.  There was a total  of
91 6-day periods with measurements used in this evaluation.
       For the APEX simulations we used the same model inputs as for the Los Angeles
simulations, described in Section 4.5 above, except for the air exchange rates.  The AERs used
were those developed for Sacramento from measurements taken in Sacramento and the inland
portions of the Los Angeles area: Riverside and San Bernardino Counties.  The hourly outdoor
O3 concentrations were from fixed site monitors located in Upland and Crestline.
       For each 6-day period for which personal measurements were available we simulated
10,000 subjects in the 7-12 age range in each of the two study areas. For each case the
distribution of simulated 6-day average exposure concentrations was compared to the
corresponding distribution of measured values, which ranged from 8 to 31 subjects.
Comparisons were also made between the continuous measurements made inside the  subjects'
homes and the APEX indoor residential concentration estimates during the times of exposure,
and between the Os concentrations measured outside the homes of the study subjects and those
measured at the nearby fixed site monitors.
       Figure 4-2 and Figure 4-3 illustrate the means of the population distributions of modeled
exposures from APEX (squares) and the ranges and means of the 8 to 31 measured personal
exposures (bars and diamonds) for each 6-day period, for Upland and for Lake Arrowhead,
Crestline, and Running Springs, arranged in order of increasing measured means for each week.
                                      4-31

-------
                      Means of Weekly Average
                Personal Ozone Exposure Concentrations
                             — Upland —
s- u.uo
Q.
"re
** n n-i
c u.u i
s
£ !
o u

i measurement
range
T
-p. i ^— T r i i

» Measured • APE>




f


"'




, i
I
_•




J



_•





;




';




'




. t
I
_•




i i
i














1


• -
....
M.-


Figure 4-2. Modeled and measured 6-day average personal exposures, Upland
                       Means of Weekly Average
                 Personal Ozone Exposure Concentrations
                          — Lake Arrowhead —
                               • Measured  • APEX
 Figure 4-3. Modeled and measured 6-day average personal exposures, Lake
 Arrowhead, Crestline, and Running Springs
                                4-32

-------
       Variability in exposure is reasonably estimated by APEX across the population, however
across most of the higher percentiles of the distributions, APEX estimates are lower than the
measured exposures. In general, APEX tends to underestimate the exposures by varying
amounts, ranging from 0 to less than 0.02 ppm. These exposures are low because they include
night time exposures. It is not clear how these results might translate to model performance of 8-
hour average exposures. Additional results and analyses of the evaluation are discussed in the
Exposure Assessment TSD. Since this evaluation is based on 6-day average exposures, it is only
of limited relevance for evaluating daily maximum 8-hr average exposure simulation results.

       4.6.3  Sensitivity Analyses
       The sensitivity analyses most relevant to evaluating APEX are summarized here.
Additional sensitivity analyses are described in the Exposure Assessment TSD.

       4.6.3.1 Near-Road Residential Exposures
       APEX does not take into account the effects of mobile source emissions of NO on the
concentrations of 63 in residences near roadways; therefore, we conducted a sensitivity analysis
to assess the potential effect on exposures in residences  of the titration of Os by mobile source
NO. We performed APEX simulations for the Boston and Houston 2002 base cases for each of
these three subsets of the population, defined by the distance of their residence from a major
roadway, based on the fractions of the population in each Census tract that live in three bands:
    a)  0-75 m from a major roadway,
    b)  75-200 m from a major roadway, and
    c)  >200 m from a major roadway.
       We used proximity factors to decrease the ambient concentrations outside their
residences in accordance with the distance from roadways.  The combined results of these three
simulations account for the decreased  exposures in residences near roadways. A comparison of
these model results with the standard simulations show decreases of one to three percent in the
estimated counts of exposures above 0.07 ppm-8hr. This demonstrates that the uncertainty
engendered by not accounting for the titration of 63 by NO in residences near roadways is  quite
small for high exposures.  Since high exposures are generally associated with outdoor and not
indoor activities, this is  not surprising.

       4.6.3.2 Air Exchange Rates and the Prevalence of Residential Air Conditioning
       Air exchange rates and the prevalence of residential air conditioning are two of the
influential determinants of exposures while indoors, and while the uncertainty of these two
model input parameters are accounted for in the uncertainty analysis described below, it is
possible that these inputs might be biased, and this sensitivity analysis assessed the effect of
                                      4-33

-------
potential biases in the values used in the APEX exposure modeling.  APEX simulations for the
Boston 2002 base case and current standard scenarios were performed, varying the air
conditioning prevalence rate from 85% to 70% and to 50%, and also varying the geometric
means of all of the AER distributions simultaneously by +0.1 and by -0.1. Biases of these
magnitudes in these parameters turn out to have little effect on the high 8-hour average
exposures predicted by APEX.  For example, changes in the percents of children experiencing
exposures over 0.06, 0.07, and 0.08 ppm-8hr change by less than one percentage point.

       4.6.3.3 Activity Patterns: Representativeness of CHAD
       Many of the studies included in the CHAD data base are not national in scope, nor do
they necessarily correspond to the modeled urban areas, and we conducted an analysis to assess
how similar the exposure results are when using individual component studies.  Strong similarity
would suggest that extrapolation of activity data gathered from one sample population to another
population is appropriate.  The largest and most comprehensive individual study in CHAD is the
National Human Activity Pattern Study (NHAPS), and we compared the APEX exposure results
using all of CHAD with corresponding results using only the NHAPS data.  NHAPS is national
in scope, with a random design, and comprises more than half of the CHAD diaries for all ages,
and 43 percent of the diary days in CHAD for children  ages 5 to 18.  Sensitivity analyses
conducted using other subsets of CHAD are described in the Exposure Assessment TSD.
       APEX simulations were performed using only NHAPS diaries for all 12 urban areas, for
the 2002 base case and the scenario of meeting the current NAAQS.  The results of this
comparison for the 2002 base case simulations are presented in Table 4-8 for the percent of
children at moderate exertion with 8-hour exposures above exposure levels of 0.06, 0.07, and
0.08 ppm-8hr. The comparison of estimated reductions in exposures to children at moderate
exertion in going from the base case to the current standard is presented in Table 4-9.
       There is very good agreement between the APEX results, whether all of CHAD or only
the NHAPS component of CHAD is used, indicating that the model results are not being unduly
influenced by any single study in CHAD.  This also indicates that the method for stratifying
diaries when assigning diaries to simulated individuals  (discussed in Section 4.3.2) is
appropriate.

       4.6.3.4 Activity Patterns: Underestimation of Repeated Exposures
       Not only is the actual exposure level important for the development of adverse health
outcomes but also the frequency of exposures at given levels of concern. In the absence of
specific data to directly evaluate the repeated exposure  results generated by APEX,
                                      4-34

-------
Table 4-8. Comparison of APEX 2002 base case simulations:  All CHAD vs.
Percent of children at moderate exertion with 8-hour exposures above levels
the NHAPS part of CHAD.
of 0.06, 0.07, 0.08 ppm-8hr.
Above 0.06 ppm-8hr
CSA
Atlanta
Boston
Chicago
Cleveland
Detroit
Houston
Los Angeles
New York
Philadelphia
Sacramento
St. Louis
Washington
All
CHAD
64%
62%
67%
74%
70%
55%
61%
71%
74%
64%
70%
72%
NHAPS
only
65%
60%
66%
73%
69%
55%
60%
70%
73%
65%
69%
72%
Absolute
difference
1%
( 2%)
( 0%)
( 1%)
( 0%)
( 0%)
( 1%)
( 1%)
( 1%)
1%
( 1%)
( 0%)
Above 0.07 ppm-8hr
All
CHAD
34%
41%
40%
57%
46%
26%
35%
49%
57%
36%
50%
50%
NHAPS
only
38%
41%
42%
58%
49%
28%
35%
49%
55%
39%
50%
51%
Absolute
difference
3%
0%
2%
0%
3%
2%
( 0%)
0%
( 1%)
2%
( 0%)
1%
Table 4-9. Comparison of APEX simulations: All CHAD vs. the NHAPS
reduction1 from the 2002 base case to the current standard of the number
exertion with 8-hour exposures above levels of 0.06, 0.07, 0.08 ppm-8hr.
Above 0.06 ppm-8hr
CSA
Atlanta
Boston
Chicago
Cleveland
Detroit
Houston
Los Angeles
New York
Philadelphia
Sacramento
St. Louis
Washington
All
CHAD
26%
21%
27%
17%
17%
58%
88%
37%
18%
51%
9%
22%
NHAPS
only
23%
19%
24%
16%
15%
54%
85%
34%
18%
48%
9%
20%
Absolute
difference
( 3%)
( 2%)
( 3%)
( 2%)
( 2%)
( 4%)
( 3%)
( 3%)
( 0%)
( 3%)
( 0%)
( 2%)
Above 0.07 ppm-8hr
All
CHAD
57%
41%
54%
46%
44%
77%
98%
71%
42%
81%
27%
49%
NHAPS
only
50%
39%
55%
41%
42%
73%
96%
67%
39%
75%
26%
47%
Absolute
difference
( 7%)
( 2%)
1%
( 5%)
( 3%)
( 3%)
( 1%)
( 4%)
( 2%)
( 6%)
( 1%)
( 2%)
Above 0.08 ppm-8hr
All
CHAD
11%
20%
15%
31%
18%
11%
16%
25%
34%
13%
21%
25%
part of CHAD
of children at
NHAPS Absolute
only difference
14%
21%
16%
34%
20%
12%
17%
25%
34%
16%
22%
27%
. Percent
moderate
4%
1%
1%
3%
2%
2%
1%
1%
0%
3%
1%
2%

Above 0.08 ppm-8hr
All
CHAD
76%
58%
84%
79%
84%
91%
100%
91%
74%
93%
53%
73%
NHAPS Absolute
only difference
71% (
55% (
82% (
79% (
79% (
89% (
99% (
88% (
69% (
91% (
50% (
73% (
4%)
3%)
2%)
1%)
5%)
2%)
0%)
3%)
5%)
2%)
3%)
0%)
1 The percent reductions are calculated as 100(base case results - current standard results)/(base case results).
                                         4-35

-------
the population of outdoor workers was targeted as a group that can be used indirectly for this
type of evaluation.  To this end, a comparison of estimated exposures of outdoor workers in two
urban areas (Atlanta and Sacramento) with exposures estimated by APEX for those urban areas
was performed.
      Due to data limitations, APEX has not been set up to specifically model outdoor workers,
but can estimate exposures for all employed adults in an urban area.  Comparison of estimates of
repeated exposures to outdoor workers with the corresponding APEX estimates for all workers
reveals that APEX significantly underestimates the number of multiple exposures for working
adults. For example, in Atlanta, APEX estimates 490,000 workers to be exposed at least once to
levels above 0.07 ppm-8hr in 2002; however, only 220 were estimated to experience repeated
exposures above 0.07 ppm-8hr six or more times.  Because there is no information on the
exertion levels of outdoor workers, these estimates include all exertion levels and are not
restricted to moderate or greater exertion. In contrast to the APEX estimates, a separate estimate
of exposures to outdoor workers, based on outdoor worker frequency estimates for occupation
categories, gives a range of from 62,000 to 140,000 outdoor workers experiencing six or more
repeated exposures to levels  above 0.07 ppm-8hr in Atlanta for 2002 air quality. Table 4-10
summarizes the comparisons performed for Atlanta and Sacramento for repeated exposures to
levels above 0.06, 0.07, and  0.08 ppm-8hr.  A description of the method used for estimating
these exposures of outdoor workers is given in Langstaff (2007). Since  outdoor workers are a
subset of all employed adults, it is clear from Table 4-10 that APEX is underestimating repeated
exposures to adult workers.

Table 4-10.  Comparison of estimated outdoor workers' repeated exposures with APEX
results for all workers, in Atlanta and Sacramento, 2002. Numbers  of people with at least
six repeated 8-hour exposures above 0.06, 0.07, and 0.08 ppm-Shr.1	
              # above 0.06 ppm-Shr    #  above 0.07 ppm-Shr    # above 0.08 ppm-Shr
             Est. outdoor  APEX all  Est. outdoor  APEX all  Est.  outdoor  APEX all
                  workers   workers     workers    workers      workers   workers
Atlanta

Sacramento

63,000 -
150,000
30,000 -
61,000
74,000

30,000

62,000 -
140,000
27,000 -
55,000
220

95

41,000-
94,000
21,000-
42,000
0

0

1 The numbers in this table have been rounded to two significant digits.

        This underestimation results primarily from the way that people's activities are modeled
using CHAD, which does not properly account for repeated behavior of individuals. The new
longitudinal methodology does increase the similarity of daily activities for a given simulated
individual in terms of the time spent outdoors, and some simulated individuals tend to spend
                                      4-36

-------
more time outdoors than others, compared to a more random assignment of diaries from CHAD
to modeled individuals. However, repeated routine behavior from one weekday to the next is not
simulated.  For example, there are no simulated individuals representing children in summer
camps who spend a large portion of their time outdoors, or adults with well-correlated weekday
schedules.  These limitations apply  to both children and adults, and therefore multiple exposures
to children are also expected to be underestimated by APEX.

       4.6.4   Uncertainty Analysis
       An understanding of the uncertainty of the APEX model predictions has been developed
through sets of complementary analyses addressing different aspects of the overall uncertainly.
A Monte Carlo analysis was performed which accounts for most of the uncertainties of the
APEX model inputs. Sensitivity analyses have been conducted to address the potential influence
of other sources of uncertainty, described in the previous  section.  This section provides a
summary of the results of the exposure modeling uncertainty analysis; the details of the
uncertainty analysis are described in Langstaff (2007).
       A Monte Carlo approach was selected for a detailed uncertainty analyses. Monte Carlo
methods for analysis of model uncertainty use statistical sampling techniques to estimate
statistics which characterize uncertainty.  Essentially, a Monte Carlo approach entails performing
many model runs with model inputs randomly sampled from distributions reflecting the
uncertainty of the inputs.  This propagates the uncertainty of the model inputs through to the
model results, taking into account input parameter dependencies and the interaction of
uncertainties within the model.  These simulations provide uncertainties  of model results in terms
of uncertainty distributions of the model outputs. From these we calculate 95 percent uncertainty
intervals (UI) for a particular model result as the interval from the 2.5th to the 97.5th percentile of
the uncertainty distribution for that  result.
       The Monte Carlo uncertainty analysis performed accounts for the following sources of
uncertainty:
   •   Ambient air concentrations measurement error
   •   Spatial interpolation of ambient concentrations
   •   Air exchange rates
   •   Air conditioning prevalence rates
   •   Ozone deposition and decay rates
   •   Vehicle penetration factors
   •   Longitudinal diary  assembly parameters
   •   Metabolic equivalents (MET)
   •   Model convergence
                                       4-37

-------
       The Monte Carlo uncertainty analysis was performed for Boston 2002, for the recent year
base case and the current standard scenarios, as well as for the estimated reductions in exposures
in going from the base case to the current standard. Uncertainties of model results for other areas
and years are expected to be similar.
       Figure 4-4 illustrates the uncertainty distributions for one model result, the percent of
children with exposures above 0.06 ppm-8hr while at moderate exertion.  This distribution
results from approximately 2000 Monte Carlo APEX simulations of the Boston 2002 base case
with model inputs varied randomly according to their uncertainty.  The "point estimate" of 62
percent is the result from the APEX simulation using our best estimates of the model inputs, as
described in Section 4.5. The corresponding result from the Monte Carlo simulations ranges
from 56 to 67 percent, with a 95 percent UI of 58 to 65 percent.  Figure 4-5 and Figure 4-6
illustrate the uncertainty distributions for two other model results, the percents of children with
exposures above 0.07 and 0.08 ppm-8hr while at moderate exertion.
          Uncertainty distribution for the estimated percent of children with any 8-hour exposures
                   above 0.06 ppm-8hr at moderate exertion (point estimate is 62%)
      600
         56%    57%    58%    59%    60%   61%    62%    63%
                                         Percent of children
64%
65%    66%
67%
Figure 4-4. Uncertainty of percent of children with exposures above 0.06 ppm-8hr (Boston
2002 base case)
                                       4-38

-------
          Uncertainty distribution for the estimated percent of children with any 8-hour exposures
                    above 0.07 ppm-8hr at moderate exertion (point estimate is 41%)
      600
          35%   36%    37%   38%    39%   40%    41%   42%   43%   44%   45%   46%   47%
                                           Percent of children


Figure 4-5. Uncertainty of percent of children with exposures above 0.07 ppm-8hr (Boston
2002 base case)
          Uncertainty distribution for the estimated percent of children with any 8-hour exposures
                    above 0.08 ppm-8hr at moderate exertion (point estimate is 20%)
      450
         17%      18%      19%     20%     21%      22%
                                           Percent of children
23%
24%
25%
26%
Figure 4-6. Uncertainty of percent of children with exposures above 0.08 ppm-8hr (Boston
2002 base case)

-------
       Uncertainty intervals are presented in Table 4-11 and Table 4-12 for the estimated
percentages of all children and asthmatic children with exposures above different 8-hour
exposure levels under moderate exertion. The UIs for the estimated reductions in exposures,
going from the 2002 base case to the current standard, for these two groups are given in Table
4-13.  Across these three tables, the spans of the 95 percent UIs range from 2 to 10 percentage
points, and the point estimates are generally within 5 percentage points of the UI endpoints.  The
uncertainties of the exposures to asthmatic children are slightly higher than for all children.
These results are very positive, and the modeling uncertainty is small enough to lend confidence
to the use of the model results.
 Table 4-11. Uncertainty of the estimated percent of children
 exposed at moderate exertion, Boston, 2002
Exposure level Air quality Point
(ppm-8hr) scenario estimate
0.06
0.07
0.08
0.06
0.07
0.08
base case
base case
base case
current standard
current standard
current standard
62%
41%
20%
49%
24%
8.5%
95% UI
58-65%
38-44%
19-24%
46-52%
23-27%
8-10%
 Table 4-12. Uncertainty of the estimated percent of asthmatic
 children exposed at moderate exertion, Boston, 2002
Exposure level Air quality Point
(ppm-8hr) scenario estimate
0.06
0.07
0.08
0.06
0.07
0.08
base case
base case
base case
current standard
current standard
current standard
65%
43%
21%
52%
24%
9%
95% UI
60-67%
39-46%
19-25%
48-56%
23-30%
8-11%
                                      4-40

-------
 Table 4-13. Uncertainty of the estimated percent reduction,
 from the base case to the current standard, of all children and
 asthmatic children exposed at moderate exertion, Boston, 2002
                        All children     Asthmatic children
    Exposure level      Point               Point
      (ppm-8hr)     estimate  95% UI  estimate   95% UI
0.06
0.07
0.08
21% 18-22%
41% 38-42%
58% 55-59%
19% 16-22%
43% 37-45%
58% 53-63%
       4.6.5   Key Findings
       Uncertainty of the APEX model predictions results from uncertainties in the spatial
interpolation of measured concentrations, the microenvironment models and parameters,
people's activity patterns, and, to a lesser extent, model structure.  The predominant sources of
uncertainty appear to be the activity pattern information and the spatial interpolation of ambient
concentrations from monitoring sites to other locations. The primary findings of these analysis
are the following:

   •   The Monte Carlo analysis of the uncertainties of the APEX model estimates of exposure
       distributions indicates that the uncertainty is relatively small.  The APEX estimates of the
       percent of children or asthmatic children with exposures above 0.06, 0.07, or 0.08 ppm-
       8hr under moderate exertion have 95% uncertainty intervals of at most ±6 percentage
       points.

   •   An investigation into the representativeness of the CHAD activity diaries with respect to
       the specific urban areas and time periods modeled indicates uncertainties of only a few
       percent in the APEX estimates of the numbers of children with exposures above 0.06,
       0.07, or 0.08 ppm-8hr under moderate exertion.

   •   Although the effect on exposures in residences of the titration of Os by mobile source NO
       is not explicitly  modeled by APEX, the resulting uncertainty is small, on the order of 1 to
       3 percent.

   •   APEX significantly underestimates the frequency of occurrence of individuals
       experiencing repeated 8-hour average exposures greater than 0.06,  0.07, and 0.08 ppm-
       8hr. The reasons for this are understood, and further research will be required to address
       this.
                                      4-41

-------
4-42

-------
4.7  EXPOSURE ASSESSMENT RESULTS

       4.7.1  APEX Modeling Results
       The results of the exposure analysis are presented as a series of exhibits and graphs
focusing on a range of benchmark levels, described in Chapters 3 and 6, as being of particular
health concern.  In addition, a wide range of concentrations in the air quality data collected over
the three year period  (2002-2004) were used in the exposure model, providing a broad range of
estimated exposures output by the model.  Exposure results are presented for the range of
alternative standard scenarios given in Table 4-6.  Estimates of exposures for the year 2003 were
developed since the second draft of this Staff Paper for only two alternative standard levels (74/4
and 64/4) due to time constraints. This section is organized into two main subsections, the first
addressing the exposures estimated for each of the particular benchmarks and the second
reporting on the estimates of repeated exposures.

       4.7.2  Estimated Exposures above Selected Benchmark Levels
       A series of exhibits are presented for each  of the benchmark levels of persons who
experience daily maximum 8-hour average exposures above 0.080, 0.070, and 0.060 ppm-8hr. A
few notes regarding the exhibits are necessary to mention. Exposure estimates are presented for
those individuals experiencing moderate levels of exertion during the same 8-hr period that the
exposure occurred. The exertion level is characterized by breathing rates, as described in section
4.3.3. Results for children exposed to Os while engaged in moderate exertion are presented in
each of the subsequent exhibits, however results for any other population group  could have been
presented with similar exposure outcomes and patterns across the 12 cities modeled. For
example, the comparison of three population groups, children, asthmatic children, and all
persons, indicates a very similar pattern of exposure estimates, regardless of the subgroup
considered (Figure 4-7). In addition,  use of the multiple years of ambient air quality data
generated a range of exposure concentrations, bracketed by the year 2002 (highest exposure
estimates) and year 2004 (lowest exposure estimates). Exposure estimates for year 2003
generally fell in between the  exposures estimated for the other two years (Figure 4-8).
                                      4-43

-------
   Population
   Subgroup
    children
asthmatic children
     all people
                                                         PHIL
                                                             SACR
                                          City
Figure 4-7. Comparison of population groups.  Percent of population subgroups with one or
more 8-hr ozone exposures above 0.070 ppm-8hr, concomitant with moderate or greater
exertion, with just reaching the current standard, based on 2002 ambient air quality.
   Year
                                                                _T
                                               City
Figure 4-8. Comparison of years. Percent of children with one or more 8-hr ozone exposures
above 0.07 ppm-8hr, concomitant with moderate or greater exertion, with just reaching the
current standard.
                                       4-44

-------
       Nine exhibits follow, representing the exposures estimated for years 2002, 2003, and
2004, each at the three benchmark levels (0.080, 0.070, and 0.060 ppm-8hr). Each exhibit
contains three exposure metrics: the estimated percent of children exposed, the estimated
number of children exposed, and the number of person-days exposure occurred above the
particular benchmark level. The notation used for the alternative standards is defined in Table
4-6.  Exposures were also estimated using the respective year air quality data without application
of the rollback procedure, and are presented as the "base" scenario in each exhibit. The numbers
for the base scenarios are omitted from some of the figures within the exhibits (i.e., the numbers
of children and person days) since they increased the vertical scale,  obscuring the details of
interest.
       The patterns of estimated exposures are variable from city to city, primarily due to
differences in air quality (local emissions and meteorology affect these), the  rollback procedure
as applied to each separate area, and people's time-location-activity patterns. The year-to-year
variability in  exposures is also evident.  All of the urban areas modeled except Houston and Los
Angeles have fewer exposures in 2003 and 2004 than in 2002, due in varying degrees to changes
in weather and emissions of precursors to Os.
       Inspection of these exhibits shows marked differences between urban areas in the levels
of exposures under alternative standards. For example, under the same 0.074 ppm, 4th daily
maximum 8-hr average alternative standard, it is estimated that 9 percent of the Boston  children
but very few of the Los Angeles children experience 8-hr Os exposures above 0.070 ppm-8hr
while engaged in moderate exertion using 2002 air quality.  This is primarily due to the larger
range of 2002-2004 4th-highest concentrations for Boston compared to the rest of its air quality
distribution, which in general, allowed for retention  of higher concentrations within the air
quality distribution (and therefore exposures) following the rollback of the air quality. In Los
Angeles,  much more of the upper range  of the air quality distribution needed to be rolled back to
allow for the  meeting of the alternative standards, thus significantly reducing the frequency of
occurrence of high ambient concentrations (and therefore exposures).
       The form of the standard had an  impact on the number of estimated exposures, the
general magnitude of which is dependent on the benchmark level selected and year of air quality
data used. For example in considering the series of 74/x scenarios (74/5, 74/4, 74/3), the impact
could be a low as a few percent (percent of children  with estimated  exposures above 0.060 ppm
using 2004 air quality go from, on average, 3% to 1%) or much greater (percent of children with
estimated exposures above 0.060 ppm using 2002 air quality go from, on average, 31% to 22%)
                                       4-45

-------
Exhibit 1.  Summary of exposure metrics regarding estimated exceedances of 0.080 ppm 8-hr
ozone exposures, concomitant with moderate or greater exertion, based on 2002 air quality. A)
Table of percent of children with at least one exposure above 0.080 ppm 8-hr, B) Illustration of
percent of children with at least one exposure above 0.080 ppm 8-hr, C) Illustration of the
number of children with at least one exposure above 0.080 ppm 8-hr, D) Illustration of the
number of children person days with exposure above 0.080 ppm 8-hr.
       1-A) Percent of children with exposures > 0.080 ppm, moderate exertion, 2002 data
City
Atlanta
Boston
Chicago
Cleveland
Detroit
Houston
Los Angeles
New York
Philadelphia
Sacramento
St. Louis
Washington
base
11%
20%
15%
31%
18%
11%
16%
25%
34%
13%
21%
25%
84/4
3%
8%
2%
6%
3%
1%
0%
2%
9%
1%
10%
7%
84/3
2%
6%
1%
2%
1%
1%
0%
1%
5%
0%
6%
3%
80/4
1%
5%
1%
1%
1%
0%
0%
1%
4%
0%
4%
3%
74/5
0%
4%
0%
0%
0%
0%
0%
0%
1%
0%
1%
1%
74/4
0%
1%
0%
0%
0%
0%
0%
0%
1%
0%
1%
0%
74/3
0%
1%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
70/4
0%
1%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
64/4
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
          base
           84I4
            so/4
    Alternative  74/4
    Standard    7Q/4
               64/4
    NY
        PHIL
            SACR
                  ,. ,
                     WASH
City
                                       4-46

-------
                                                                                                          1.2E+5
                                                                                                       O.OE+0
        84/4 \
          80/4
Alternative  74/4  \
 Standard    70/4
             64/4
                                                                                           „.
                                                                                                WASH
                     ATLA
    1-D)
         84/4
          8014  \
Alternative 74/4
 Standard   70/4
             64/4
                                                                                                          1.2E+5
                                                                                                          1 .OE+5
                                                                                                                 (D  (1)
                                                                                                         8.0E+4  "§  °
                                                                                                         6.0E+4
                                                                                                         4.0E+4
                                                                                                        2.0E+4
                                                                                                        O.OE+0
                                                                                           ST.L
                                                                                                WASH
                     ATLA
                             BOST
                                     CHIC
                                                                                                                 CO It

                                                                                                                 11
                                                                                                                 00 °-
                                                                                                                 ?%
                                                                                                                    tn
                                               4-47

-------
Exhibit 2.  Summary of exposure metrics regarding estimated exceedances of 0.070 ppm 8-hr
ozone exposures, concomitant with moderate or greater exertion, based on 2002 air quality. A)
Table of percent of children with at least one exposure above 0.070 ppm 8-hr, B) Illustration of
percent of children with at least one exposure above 0.070 ppm 8-hr, C) Illustration of the
number of children with at least one exposure above 0.070 ppm 8-hr, D) Illustration of the
number of children person days with exposure above 0.070 ppm 8-hr.

       2-A) Percent of children with exposures > 0.070 ppm, moderate exertion, 2002 data
City
Atlanta
Boston
Chicago
Cleveland
Detroit
Houston
Los Angeles
New York
Philadelphia
Sacramento
St. Louis
Washington
base
34%
41%
40%
57%
46%
26%
35%
49%
57%
36%
50%
50%
84/4
15%
24%
18%
31%
26%
6%
1%
14%
33%
7%
37%
26%
84/3
13%
18%
14%
23%
18%
3%
1%
11%
27%
5%
30%
18%
80/4
8%
17%
10%
22%
16%
3%
0%
8%
24%
3%
27%
17%
74/5
4%
15%
5%
10%
14%
1%
0%
2%
13%
1%
15%
11%
74/4
3%
9%
3%
8%
4%
1%
0%
2%
11%
1%
12%
8%
74/3
3%
6%
2%
3%
1%
0%
0%
1%
7%
0%
8%
4%
70/4
1%
5%
0%
2%
0%
0%
0%
1%
5%
0%
4%
3%
64/4
0%
1%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
       2-B)
      base \
       84/4
        80/4 \
Alternative  74/4 \
 Standard    j0/4
           64/4
                                                                              WASH
                     ATLA
                                       4-48

-------
    2-C)
        84/4
         80/4 A
Alternative 74/4 A,
 Standard   70/4
             64/4
                                                                                                    6.0E+5
                                                                                                   5.0E+5
                                                                                                   4.0E+5  "§
                                                                                                   3.0E+5
                                                                                                  2.0E+5
                                                                                                  1.0E+5
                                                                                                 O.OE+0
          PHIL
               SACR
                     STL
                          WASH
City
                                                                                                          »!•
                    ATLA
                                                                                                    8.0E+5
   2-D)
                                                                                          WASH
                    ATLA
                           BOST
                                   CHIC
                                            4-49

-------
Exhibit 3.  Summary of exposure metrics regarding estimated exceedances of 0.060 ppm 8-hr
ozone exposures, concomitant with moderate or greater exertion, based on 2002 air quality. A)
Table of percent of children with at least one exposure above 0.060 ppm 8-hr, B) Illustration of
percent of children with at least one exposure above 0.060 ppm 8-hr, C) Illustration of the
number of children with at least one exposure above 0.060 ppm 8-hr, D) Illustration of the
number of children person days with exposure above 0.060 ppm 8-hr.

       3-A) Percent of children with exposures > 0.060 ppm, moderate exertion, 2002 data
City
Atlanta
Boston
Chicago
Cleveland
Detroit
Houston
Los Angeles
New York
Philadelphia
Sacramento
St. Louis
Washington
base
64%
62%
67%
74%
70%
55%
61%
71%
74%
64%
70%
72%
84/4
47%
49%
49%
61%
58%
23%
7%
45%
61%
32%
64%
56%
84/3
47%
43%
44%
56%
51%
18%
7%
41%
56%
27%
61%
49%
80/4
38%
42%
40%
54%
50%
16%
5%
37%
54%
23%
59%
49%
74/5
29%
39%
32%
44%
48%
9%
1%
20%
44%
13%
51%
40%
74/4
24%
30%
25%
40%
34%
8%
1%
22%
42%
12%
46%
35%
74/3
23%
24%
20%
32%
26%
6%
1%
18%
36%
9%
41%
28%
70/4
13%
22%
15%
29%
22%
4%
1%
13%
32%
6%
35%
25%
64/4
3%
9%
3%
12%
6%
1%
0%
3%
15%
1%
15%
10%
      3-B)
          base
           84/4
            8014
    Alternative  74/4 A
    Standard    70/4 \
               64/4
        PHIL
            SACR
                  _T ,
                     WASH
City
                     ATLA
                           BOST
                                       4-50

-------
   3-C)
      84/4
       80/4 A
Alternative 74/4 A
 Standard  70/4 A
         64/4
              u   "MR  irt   vy^^^^^      LA   NY
              jU^^^ous    -
               ATLA
                                                       j	I  :^—r^
                                                       -^—1     _TI  WASH
                                                       PHIL SACR
   3-D)
                                                                           3.5E+6
      84/4
       80/4 \
Alternative ?4/4
 Standard  70/^
         64/4
                                                                         O.OE+0
                               CLEV
                                        ^   ^-^^=^^^STLWASH

                                        r^-^T^T^ SACR
                                         HOUS
                                                City
                                 4-51

-------
Exhibit 4.  Summary of exposure metrics regarding estimated exceedances of 0.080 ppm 8-hr
ozone exposures, concomitant with moderate or greater exertion, based on 2003 air quality.  A)
Table of percent of children with at least one exposure above 0.080 ppm 8-hr, B) Illustration of
percent of children with at least one exposure above 0.080 ppm 8-hr, C) Illustration of the
number of children with at least one exposure above 0.080 ppm 8-hr, D) Illustration of the
number of children person days with exposure  above 0.080 ppm 8-hr.

       4-A) Percent of children with exposures > 0.080 ppm, moderate exertion, 2003 data
City
Atlanta
Boston
Chicago
Cleveland
Detroit
Houston
Los Angeles
New York
Philadelphia
Sacramento
St. Louis
Washington
base
3%
2%
1%
7%
10%
15%
25%
11%
13%
7%
7%
7%
84/4
1%
1%
0%
1%
1%
0%
0%
1%
3%
0%
2%
1%
84/3












80/4












74/5












74/4
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
74/3












70/4












64/4
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
      4-B)
      base
       84/4  \
        80/4 A
Alternative  74/4 \
 Standard    70/4
           64/4
                                                                              WASH
                                       4-52

-------
                                                                                                      3.5E+4
 4-C)
                                                                                                   O.OE+0
                                                                  City
                     ATLA
                                                                                                      3.5E+4
 4-D)
        84/4
         80/4
Alternative 74/4 \
 Standard   7014 \
             64/4 A
                     ATLA
                                             4-53

-------
Exhibit 5.  Summary of exposure metrics regarding estimated exceedances of 0.070 ppm 8-hr
ozone exposures, concomitant with moderate or greater exertion, based on 2003 air quality. A)
Table of percent of children with at least one exposure above 0.070 ppm 8-hr, B) Illustration of
percent of children with at least one exposure above 0.070 ppm 8-hr, C) Illustration of the
number of children with at least one exposure above 0.070 ppm 8-hr, D) Illustration of the
number of children person days with exposure above 0.070 ppm 8-hr.

       5-A) Percent of children with exposures > 0.070 ppm, moderate exertion, 2003 data
City
Atlanta
Boston
Chicago
Cleveland
Detroit
Houston
Los Angeles
New York
Philadelphia
Sacramento
St. Louis
Washington
base
13%
9%
7%
18%
25%
38%
48%
27%
30%
26%
23%
16%
84/4
4%
5%
3%
4%
5%
2%
1%
5%
10%
2%
10%
5%
84/3












80/4












74/5












74/4
1%
1%
0%
0%
0%
0%
0%
0%
2%
0%
2%
1%
74/3












70/4












64/4
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
      5-B)
          base  \
           84/4
            80/4 \
    Alternative 74/4 \
    Standard    70/4 \
              64/4"
                                                                                     50
                                                                                     45
-j-40


^ -35


1
^-30


r25
4-20


4-15


-10
-5
-0

-O
?
o
ffi (D
X 3
0 ?
(/) -i
= 0
S i
V $
O (D
b =
•-1 s
•o ™
•o =•
3 -^
00 0
^ ^
o
s



               "  STL
      HOUS
                     ATLA
                                      4-54

-------
                                                                                                          2.5E+5
   5-C)
        84/4
          80/4
Alternative  74/4
 Standard    70/4
             64/4
                                                                                                       O.OE+0

                                                                                                WASH
                     ATLA
  5-D)
        84/4
          80/4

Alternative  74/4
 Standard    70/4  \
             64/4
                                                                                                          2.5E+5
                                                                                                         2.0E+5
                                                                                                         1.5E+5   = Z
                                                                                                                 "> 7?
                                                                                                                 w o-
                                                                                                        1.0E+5
                                                                                                       5.0E+4
                                                                                                       O.OE+0
                                                                                                WASH
                                                                                                                I
                     ATLA
                                               4-55

-------
Exhibit 6.  Summary of exposure metrics regarding estimated exceedances of 0.060 ppm 8-hr
ozone exposures, concomitant with moderate or greater exertion, based on 2003 air quality.  A)
Table of percent of children with at least one exposure above 0.060 ppm 8-hr, B) Illustration of
percent of children with at least one exposure above 0.060 ppm 8-hr, C) Illustration of the
number of children with at least one exposure above 0.060 ppm 8-hr, D) Illustration of the
number of children person days with exposure  above 0.060 ppm 8-hr.

       6-A) Percent of children with exposures > 0.060 ppm, moderate exertion, 2003 data
City
Atlanta
Boston
Chicago
Cleveland
Detroit
Houston
Los Angeles
New York
Philadelphia
Sacramento
St. Louis
Washington
base
47%
32%
29%
42%
50%
64%
71%
53%
55%
57%
54%
41%
84/4
28%
22%
19%
16%
20%
17%
8%
21%
31%
14%
38%
15%
84/3












80/4












74/5












74/4
7%
7%
5%
5%
6%
3%
2%
6%
11%
3%
14%
5%
74/3












70/4












64/4
0%
1%
0%
0%
0%
0%
0%
0%
1%
0%
1%
0%
      6-B)
          base  \
           84/4
            80/4 \
    Alternative 74/4 \
    Standard    70/4
              64/4
WASH
                                       4-56

-------
                                                                                                            9.0E+5
   6-C)
         84/4
          80/4
Alternative 74/4
 Standard   70/4  \^
              64/4
                                                                                                           4.0E+5  g E
                                                                                                                   |»
                                                                                                           3.0E+5  ?  |
                                                                                                                   39
                      ATLA
  6-D)
        84/4 \  I—"
         80/4
Alternative  74/4
 Standard    70/4
             64/4
                                                                                                        1.2E+6
                                                                                                       1.0E+6
                                                                                                       8.0E+5
                                                                                                      6.0E+5
                                                                                                      4.0E+5
                                                                                                      2.0E+5
                                                                                                     O.OE+0
                                                                                                                S-c
                                                                                                                 (D

                                                                                                              11
                     ATLA
                                                4-57

-------
Exhibit 7.  Summary of exposure metrics regarding estimated exceedances of 0.080 ppm 8-hr
ozone exposures, concomitant with moderate or greater exertion, based on 2004 air quality. A)
Table of percent of children with at least one exposure above 0.080 ppm 8-hr, B) Illustration of
percent of children with at least one exposure above 0.080 ppm 8-hr, C) Illustration of the
number of children with at least one exposure above 0.080 ppm 8-hr, D) Illustration of the
number of children person days with exposure above 0.080 ppm 8-hr.

       7-A) Percent of children with exposures > 0.080 ppm, moderate exertion, 2004 data
City
Atlanta
Boston
Chicago
Cleveland
Detroit
Houston
Los Angeles
New York
Philadelphia
Sacramento
St. Louis
Washington
base
3%
1%
0%
0%
0%
9%
13%
1%
2%
1%
0%
4%
84/4
0%
0%
0%
0%
0%
1%
0%
0%
0%
0%
0%
0%
84/3
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
80/4
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
74/5
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
74/4
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
74/3
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
70/4
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
64/4
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
       7-B)
      base
       84/4
        80/4  \
Alternative  74/4  \
 Standard    70/4 \
           64/4
                     ATLA
                           BOST
                                 CHIC
 ^^     LA
^^^1     HOUS   LA
 CLEV  DETR
                                                            NY
                                                                55^  I     WASH
                                                                 PHIL SACR
                                                         city
                                       4-58

-------
                                                                                                          1.0E+4
                                                                                                          9.0E+3
                                                                                                         8.0E+3
                                                                                                         7.0E+3  x
         84/4
          80/4 A
Alternative  74/4 A
 Standard   70/4
             64/4
                                                                                                                | I

                                                                                                         6.0E+3  | »
                                                                                                                 ft\ ~
                                                                                                         5.0E+3
                                                                                                        4.0E+3
                                                                                                                   n>
                                                                                                                •o
                                                                                                                •o

                                                                                                                 U
                     ATLA
    7-D
         84/4
          80/4
Alternative  74/4
 Standard   70/4
             64/4
                                                                                                          1.2E+4
                                                                                                          1 .OE+4
                                                                                                        4.0E+3
                                                                                                        2.0E+3
                                                                                                       O.OE+0
                                                                         NY
                                                                               PHIL
                                                                                    SACR
                                                                                           SJL
                                                                                                WASH
                     ATLA
                             BOST
                                     CHIC
                                            CLEV
                                                                    City
                                                                                                                   3
                                                                                                                   57
                                                                                                         8.0E+3  "§ °
                                                                                                         6.0E+3   o
                                                                                                                 li
                                               4-59

-------
Exhibit 8.  Summary of exposure metrics regarding estimated exceedances of 0.070 ppm 8-hr
ozone exposures, concomitant with moderate or greater exertion, based on 2004 air quality.  A)
Table of percent of children with at least one exposure above 0.070 ppm 8-hr, B) Illustration of
percent of children with at least one exposure above 0.070 ppm 8-hr, C) Illustration of the
number of children with at least one exposure above 0.070 ppm 8-hr, D) Illustration of the
number of children person days with exposure  above 0.070 ppm 8-hr.
       8-A) Percent of children with exposures > 0.070 ppm, moderate exertion, 2004 data
City
Atlanta
Boston
Chicago
Cleveland
Detroit
Houston
Los Angeles
New York
Philadelphia
Sacramento
St. Louis
Washington
base
13%
6%
1%
4%
3%
26%
37%
8%
11%
9%
1%
14%
84/4
4%
1%
0%
0%
0%
5%
0%
0%
1%
0%
0%
3%
84/3
4%
1%
0%
0%
0%
3%
0%
0%
0%
0%
0%
2%
80/4
2%
0%
0%
0%
0%
3%
0%
0%
0%
0%
0%
1%
74/5
1%
0%
0%
0%
0%
1%
0%
0%
0%
0%
0%
0%
74/4
0%
0%
0%
0%
0%
1%
0%
0%
0%
0%
0%
0%
74/3
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
70/4
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
64/4
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
       8-B)
          base
           84/4  \
            80/4
    Alternative 74/4 \
    Standard
WASH
              70/4
                     ATLA
                                       4-60

-------
    8-C)
         84/4
          80/4 A
Alternative 74/4 A,
 Standard   70/4
              64/4
                                                                                                          6.0E+4
                                                                                                          5.0E+4
                                                                                                          4.0E+4  "§
                                                                                                         3.0E+4
                                                                                                         2.0E+4
                                                                                                        1.0E+4
                                                                                                        O.OE+0
                                                                                                                    Q.
                                                                                                                    <5
                                                                                                                 »!•
                     ATLA
                                                                                                          6.0E+4
    8-D)
         84/4
          80/4 A
Alternative 74/4
 Standard   70/4
              64/4
                                                                                                        O.OE+0
WASH
                     ATLA
                                               4-61

-------
Exhibit 9.  Summary of exposure metrics regarding estimated exceedances of 0.060 ppm 8-hr
ozone exposures, concomitant with moderate or greater exertion, based on 2004 air quality. A)
Table of percent of children with at least one exposure above 0.060 ppm 8-hr, B) Illustration of
percent of children with at least one exposure above 0.060 ppm 8-hr, C) Illustration of the
number of children with at least one exposure above 0.060 ppm 8-hr, D) Illustration of the
number of children person days with exposure above 0.060 ppm 8-hr.
       9-A) Percent of children with exposures > 0.060 ppm, moderate exertion, 2004 data
City
Atlanta
Boston
Chicago
Cleveland
Detroit
Houston
Los Angeles
New York
Philadelphia
Sacramento
St. Louis
Washington
base
43%
23%
9%
23%
20%
54%
67%
29%
41%
40%
18%
41%
84/4
21%
9%
1%
4%
7%
22%
5%
4%
12%
5%
7%
16%
84/3
20%
6%
1%
2%
4%
17%
4%
3%
8%
4%
4%
11%
80/4
14%
5%
0%
2%
3%
15%
3%
2%
7%
2%
3%
10%
74/5
8%
4%
0%
0%
2%
9%
1%
0%
2%
1%
1%
5%
74/4
6%
2%
0%
0%
0%
7%
1%
0%
1%
0%
0%
4%
74/3
5%
1%
0%
0%
0%
5%
1%
0%
1%
0%
0%
2%
70/4
2%
0%
0%
0%
0%
4%
0%
0%
0%
0%
0%
1%
64/4
0%
0%
0%
0%
0%
1%
0%
0%
0%
0%
0%
0%
       9-B)
                                                                            WASH
                                      4-62

-------
    9-C)
        84/4  ,
         80/4 A

Alternative  74/4 \

 Standard    70/4
             64/4
                                                                                                       2.5E+5
                                                                                                       2.0E+5
                                                                                                      1.5E+5  5 »
                                                                                                              m ~
                                                                                                      1.0E+5
                                                                                                     5.0E+4
                                                                                                     O.OE+0
     WASH
                                                                                                              »!•
                     ATLA
    9-D)
        84/4
         8014 \

Alternative  74/4

 Standard    70/4
             64/4
                                                                                                       3.5E+5
                                                                                                       3.0E+5
                                                                                                      2.5E+5
                        3
                     s  ?
                     ?  a
                     g  &
             2.0E+5  S  =
                     V  (D
                                                                                                      1.5E+5
                                                                                                     1.0E+5
                                                                                                     5.0E+4
                                                                                                     O.OE+0
ST.L
     WASH
                     ATLA
                            BOST
                                    CHIC
                     CT) (D

                     11
                     §&
                     "^  to
                                              4-63

-------
       4.7.3   Estimates of Repeated Exposures
       As discussed in section 3.6.3, multiple exposures pose a greater health concern than
single exposures.  However, multiple repeated exposures are underestimated by APEX
(discussed previously in the model evaluation section), which should be kept in mind for
interpretation of these results. Figures 4-9 through 4-11 illustrate the effect of the current and
alternative standards has on the estimated percent of children experiencing 1, 2, and 3 or more
repeated exposures above 0.080, 0.070, and 0.060 ppm-8hr, respectively, concomitant with
moderate or greater exertion, for each of the urban areas modeled, based on rollback of 2002 O
concentrations. When considering the alternative standard scenarios, clear trends are evident.
The reduction in both the number of those experiencing exposures above a given benchmark
level and the frequency of occurrence of those exceedances is directly correlated with the
standard level.
                                       4-64

-------
                                      HASH
exposures
1=11
1=12
1=134
                                              ana
                                              Dz      Dz       Dz
                                          nan
                                      n  i  u      a
                                              Dz /   Dz
                                  a     a      a
                              n     n     n
                                        z  /     i  /     i /     z  /     i  /     i  /    r
                                   Dz  /    Dz  /    Dz  /    Dz  /    Dz  /    Dz  /    Dz
                      n     a     a     a      a
                       Dz  /    Dz  /    Dz  /    Dz  /   Dz /    Dz  /    |z
                  anna
                   Dz  /   Dz  /    Dz  /    Dz  /    Dz  /    1z
              a     n     a     a      B
               Dz  /    Dz  /    Dz  /    Dz  /   Dz  /    1z  /    2z
          a     a     n     a     a
           Dz  /    Dz  /    Dz  /    Dz
Dz       Iz  /    |z  /     |z
                                              5z  /    Gz  /    Bz
   a     a     a     u
                                   Dz  /     Iz  /    2z /    3z
           DETR
       CLEV
   CHIC
B4/4     ID/4     74/3      74/4     74/5     BO/4     B4/3     B4/4
Figure 4-9.  Percent of children (ages 5-18) with repeated 8-hr ozone exposures > 0.08 ppm-8hr,
for exposures concomitant with moderate or greater exertion, based on 2002 air quality data.
                                               4-65

-------
                                      HASH
                                              a
                                              Dz  /    3z
               Bz
        exposures
                                  ST.L
                                          a
          9
11:
                                                                                     /~1
17:
                                  n,
                                           D:  /    4z /    B:  /    12: /    15: /   27z  /   3Dz  /   37:
                               SUCH
                                      n     n      n      n     n
                                           5:  /    7:  /    ||i /    13: /    Ui //-, 111
                                           Dz  /    Dz  /    Dz  /    Dz  /    |z
                      n     a     a     a
                               Dz  /    Dz  /     I:
       I:  /    5:  /    Bz  /    9:  /    IS:  /    17:
   Dz       Iz  /    ]:  /    3:  /    4z  /    R:  /    13:
Bt/t     7D/<     J4/3
Bt/t
 Figure 4-10.  Percent of children (ages 5-18) with repeated 8-hr ozone exposures > 0.07 ppm-8hr,
 for exposures concomitant with moderate or greater exertion, based on 2002 air quality.
                                               4-66

-------
                                       WASH
        eiposures
                                   ST.L
                               SftCR
                                           15! /    35?  /   41!  /   (6!  /    51!  /    58!
                                   15! /    32! /    36!  /_ 43!  /   44!
                               3!  /    13! /    IB!  /   22!  /   20!  /    37!  /    4l! /    45?
                           a   I a     a     a     a
                                    I  I    1Z  /    l!  /    l!  /    5!  /    7!
               I2z  /    29z /    32!
        Sz  /    22z  /_  24z /_  30!
    3!  /    13!  /    23!  /    24! /    29! /    38!  /   47z  /   ( i
B4/4     ID/4     74/3     74/4     74/5
B4/3     B4/4
Figure 4-11.  Percent of children (ages 5-18) with repeated 8-hr ozone exposures > 0.06 ppm-8hr,
for exposures concomitant with moderate or greater exertion, based on 2002 air quality data.
                                               4-67

-------
REFERENCES

Adams, W. C. (2006). Comparison of chamber 6.6 h exposures to 0.04-0.08 ppm ozone via
     square-wave and triangular profiles on pulmonary responses.  Inhalation Toxicol.  18: 127-
     136.

AHS. (2003). U.S. Bureau of the Census and U.S. Department of Housing and Urban
     Development. 2003 American Housing Survey (AHS): National Survey Data. Available at:
     http://www.census.gov/hhes/www/housing/ahs/ahs.html, and
     http://www.huduser.org/datasets/ahs.html.

Ainsworth, B. E.; Haskell, W. L.; Leon, A. S.; Jacobs, D. R., Jr.; Montoye, H. J.; Sallis, J. F.;
     Paffenbarger, R. S., Jr. (1993).  Compendium of physical activities: classification  of energy
     costs of human physical activities. Med. Sci. Sports Exer. 25:  71-80.

Akland, G. G., Hartwell, T. D., Johnson, T. R., and Whitmore, R. W. (1985).  Measuring human
     exposure to carbon monoxide in Washington, D. C. and Denver, Colorado during  the
     winter of 1982-83. Environ.  Sci. Technol.  19: 911-918.

American Petroleum Institute. (1997). Sensitivity testing of pNEM/Os exposure to changes in
     the model algorithms. Health and Environmental Sciences Department.

Blank,  H. and Mitchell, A. (2001).  The Status of Preschool Policy in the States. Children's
     Defense Fund and Early Childhood Policy  Research. Available at:
     www.earlvchildhoodfinance.org/handouts/StatusOfPreschoolPolicyInTheStates.doc

CARB (2005). Review of the California Ambient Air Quality Standard for Ozone. California
     Environmental Protection Agency Air Resources Board. Available at:
     http://www.arb.ca.gov/research/aaqs/ozone-rs/rev-staff/rev-staff.htm

CASTNet. (2004).  Clean Air Status and Trends  Network (CASTNet) 2003 Annual Report,
     Prepared by: MACTEC Engineering and Consulting, Inc. Prepared for: U.S.
     Environmental Protection Agency, Office of Air and Radiation, Clean Air Markets
     Division, Washington, DC.

Dey A., Bloom B. (2005). Summary Health Statistics for U.S. Children: National Health
     Interview Survey, 2003. National Center for Health Statistics. Vital Health Stat 10(223).
     U.S. Department of Health and  Human Services, Centers for Disease Control and
     Prevention, National Center for Health Statistics, Hyattsville,  Maryland.

Duff, M., Horst, R. L., Johnson, T.  R. (1998).  "Quadratic Rollback: A Technique to Model
     Ambient Concentrations Due to Undefined Emission Controls." Presented at the Air and
     Waste Management Annual Meeting.  San Diego, California.  June 14-18, 1998.

Freedson, P. S. (1989).  "Field monitoring of physical activity in children." Pediatric Exercise
     Science 1: 8-18.

Geyh, A. S.; Xue, J.; Ozkaynak, H.; Spengler, J.  D. (2000). The Harvard Southern California
     chronic ozone exposure study: assessing ozone exposure of grade-school-age children in
     two southern  California communities. Environ. Health Perspect. 108: 265-270.
                                      4-68

-------
Goldstein, B. R.; Tardiff, G.; Hoffnagle, G.; Kester, R. (1992). Valdez health study: summary
     report. Anchorage, AK: Alyeska Pipeline Service Company.

Graham, S.E. and T. McCurdy (2004).  Developing meaningful cohorts for human exposure
     models. J. Exp. Anal.  andEnv. Epidemiol., 14, 23-43.

Hartwell, T. D.; Clayton, C.  A.; Ritchie, R. M.; Whitmore, R. W.; Zelon, H. S.; Jones, S. M.;
     Whitehurst, D. A. (1984). Study of carbon monoxide exposure of residents of Washington,
     DC and Denver, Colorado. Research Triangle Park, NC: U.S. Environmental Protection
     Agency, Office of Research and Development, Environmental Monitoring Systems
     Laboratory; report no.  EPA-600/4-84-031.

Hofferth, S., Davis-Kean. P. E., Davis, J., Finkelstein, J. (1997).  The Child Development
     Supplement To The Panel Study Of Income Dynamics. 1997 User Guide. Survey Research
     Center, Institute for Social Research, The University of Michigan, Ann Arbor, MI.
     Available at:  http://psidonline.isr.umich.edu/CDS/usergd.html

Johnson, T. (1984).  A Study of Personal Exposure to Carbon Monoxide in Denver, Colorado.
     Research Triangle Park, NC: U.S. Environmental Protection Agency, Environmental
     Monitoring Systems Laboratory; report no. EPA-600/4-84-014.

Johnson, T. (1989).  Human Activity Patterns in Cincinnati, Ohio. Palo Alto, CA: Electric Power
     Research Institute; report no. EPRIEN-6204.

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.

Johnson, T., Capel, J., and McCoy, M. (1996a). Estimation of Ozone Exposures Experienced by
     Urban Residents Using a Probabilistic Version of NEM and 1990 Population Data.
     Prepared by International Technology Air Quality Services for Office of Air Quality
     Planning and Standards, U.S. Environmental Protection Agency, Research Triangle Park,
     NC; contract no. 63-D-30094; April 1996. Available at:
     http://www.epa.gOv/ttn/naaqs/standards/ozone/s _o3_pr_td. html.

Johnson, T., Capel, J., McCoy, M., and Warnasch, J. (1996b). Estimation  of Ozone Exposures
     Experienced by Outdoor Children in Nine Urban Areas Using a Probabilistic Version of
     NEM. Prepared by International Technology Air Quality Services for Office of Air
     Quality Planning and Standards, EPA, Research Triangle Park, NC; contract no. 63-D-
     30094; April  1996. Available at:
     http://www.epa.gov/ttn/naaqs/standards/ozone/s_o3_pr_td.html.

Johnson, T., Capel, M. McCoy, and J. Warnasch.  (1996c).  Estimation of Ozone Exposures
     Experienced by Outdoor Workers in Nine Urban Areas Using a Probabilistic Version of
     NEM. 1996. 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:
     http://www.epa.gOv/ttn/naaqs/standards/ozone/s o3 pr td.html.
                                      4-69

-------
Johnson, T., Pakrasi, A., Wisbeth, A., Meiners, G., Ollison, W. (1995).  Ozone exposures Within
     Motor Vehicles - Results of a Field Study in Cincinnati, Ohio. Proceedings 88th annual
     meeting and exposition of the Air & Waste Management Association, June 18-23,  1995.
     San Antonio, TX. Preprint paper 95-WA84A.02.

Klepeis, N. E.; Tsang, A. M.; Behar, J. V. (1996). Analysis of the National Human Activity
     Pattern Survey (NHAPS) Respondents from a Standpoint of Exposure Assessment.
     Washington, DC: U.S. Environmental Protection Agency, Office of Research and
     Development; report no. EPA/600/R-96/074.

Langstaff, J. E. (2007). OAQPS Staff Memorandum to Ozone NAAQS Review Docket (OAR-
     2005-0172). Subject: Analysis of Uncertainty in Ozone Population Exposure Modeling.
     [January 31,2007]. Available at:
     http://www.epa.gOv/ttn/naaqs/standards/ozone/s _o3_cr_td. html.

Lee, Vallarino, Dumyahn, Ozkaynak, and Spengler. (1999).  Ozone decay rates in residences.
     JAWMA. 49: 1238-1244.

Lioy, P.J. (1990). The analysis of total human exposure for exposure assessment: A multi-
     discipline science for examining human contact with contaminants. Environmental Science
     and Technology 24:938-945.

McCurdy, T. (2000). "Conceptual basis for multi-route intake dose modeling using an energy
     expenditure approach." J. Exposure Anal. Environ. Epidemiol. 10:1-12.

McCurdy, T., Glen, G., Smith, L., Lakkadi, Y. (2000). "The National Exposure Research
     Laboratory's Consolidated Human Activity Database." J. Exposure Anal. Environ.
     Epidemiol.  10:566-578.

Monn C. (2001).  Exposure assessment of air pollutants: a review on spatial heterogeneity and
     indoor/outdoor/personal exposure to suspended particulate matter, nitrogen dioxide and
     ozone. Atmos. Env. 35(1).

Montoye, H. J., Kemper, H. C. G., et al. (1996). Measuring Physical Activity and Energy
     Expenditure. Champaign IL: Human Kinetics.

National Research Council (1991). Human Exposure Assessment for Airborne Pollutants:
     Advances and Opportunities. Washington, DC: National Academy of Sciences.

OMB (2005). U.S. OMB Bulletin No. 05-02. Subject: Update of Statistical Area Definitions
     and Guidance on Their Uses. February 22, 2005.

Persily, A. and J.  Gorfain.(2004).  Analysis of Ventilation Data from the U.S. Environmental
     Protection Agency Building Assessment Survey and Evaluation (BASE) Study. National
     Institute of Standards and Technology, NISTIR 7145, December 2004.

Persily, A., J. Gorfain, G. Brunner (2005). Ventilation Design and Performance in U.S. Office
     Buildings. ASHRAE Journal. April 2005, 30-35.
                                      4-70

-------
Richmond H., T. Palma, J. Langstaff, T. McCurdy, G. Glenn, and L. Smith (2002). "Further
     refinements and testing of APEX (3.0): EPA's population exposure model for criteria and
     air toxic inhalation exposures." Poster presentation. Joint meeting of the International
     Society of Exposure Analysis and International Society of Environmental Epidemiology,
     August 11-15, 2002, Vancouver, Canada.

Rizzo, M. (2005). Evaluation of a quadratic approach for adjusting distributions of hourly ozone
     concentrations to meet air quality standards. November 7, 2005.

Rizzo, M. (2006). A distributional comparison between different rollback methodologies applied
     to ambient ozone concentrations.  May 31, 2006.

Robinson, J. P., Wiley, J. A., Piazza, T., Garrett, K., and Cirksena, K. (1989). Activity patterns
     of California residents and their implications for potential exposure to pollution.  California
     Air Resources Board (CARB-A6-177-33), Sacramento, CA.

Spier, C.  E.; Little,  D. E.; Trim, S. C.; Johnson, T. R.; Linn, W. S.; Hackney, J. D. (1992).
     Activity patterns in elementary and high school students exposed to oxidant pollution. J.
     Exposure Anal. Environ. Epidemiol.2: 277-293.

Tsang A. M. and N. E. Klepeis (1996).  Descriptive Statistics Tables from a Detailed Analysis of
     the  National Human Activity Pattern Survey (NHAPS) Data.  U.S. Environmental
     Protection Agency (EPA/600/R-96/148).

Turk, B. H., D. T. Grimsrud, J. T. Brown, K. L. Geisling-Sobotka, J. Harrison, R. J. Prill (1989).
     Commercial building ventilation rates and particle concentrations. ASHRAE, No. 3248.

U.S. Department of Transportation and U.S. Census Bureau (2000).  Census of Population and
     Housing, 2000 long-form (sample) data, Census Transportation Planning Package (CTPP)
     2000.  Available at: http://transtats.bts.gov/.

U.S. Environmental Protection Agency (1986). Air Quality Criteria for Ozone and Other
     Photochemical Oxidants. EPA-600/8-84-020aF-eF. Office of Health and Environmental
     Assessment, Environmental Criteria and Assessment Office, U.S. Environmental
     Protection Agency, Research Triangle Park, NC. Available from: NTIS, Springfield, VA;
     PB87-142949.

U.S. Environmental Protection Agency (1995). Office Equipment: Design, Indoor Air
     Emissions,  and Pollution Prevention Opportunities. Office of Research and Development,
     U.S. Environmental Protection Agency, Washington, DC. Report no. EPA/600/R-95/045.

U.S. Environmental Protection Agency (1996a). 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, U.S.
     Environmental Protection Agency, Research Triangle Park, NC. Available at:
     http://www.epa.gOv/ttn/naaqs/standards/ozone/s o3 pr sp.html
                                      4-71

-------
U.S. Environmental Protection Agency (1996b). 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, U.S. Environmental Protection Agency,
     Research Triangle Park, NC.  Available at:
     http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=2831.

U.S. Environmental Protection Agency (2002). Consolidated Human Activities Database
     (CHAD) Users Guide. Database and documentation available at:
     http://www.epa.gov/chadnetl/.

U.S. Environmental Protection Agency (2007). Ozone Population Exposure Analysis for
     Selected Urban Areas. Office of Air Quality Planning and Standards, U.S. Environmental
     Protection Agency, Research Triangle Park, NC. Available at:
     http://www.epa.gOv/ttn/naaqs/standards/ozone/s _o3_cr_td. html.

U.S. Environmental Protection Agency (2006a). Total Risk Integrated Methodology (TRIM) -
     Air Pollutants Exposure Model Documentation (TRIM.Expo / APEX, Version 4) Volume
     I: User's Guide. Office of Air Quality Planning and Standards, U.S. Environmental
     Protection Agency, Research Triangle Park, NC. June 2006. Available at:
     http://www.epa.gov/ttn/f era/human  apex.html.

U.S. Environmental Protection Agency (2006b). Total Risk Integrated Methodology (TRIM) -
     Air Pollutants Exposure Model Documentation (TRIM.Expo / APEX, Version 4) Volume
     II: Technical  Support Document. Office of Air Quality Planning and Standards, U.S.
     Environmental Protection Agency, Research Triangle Park, NC. June 2006. Available at:
     http ://www. epa. gov/ttn/f era/human  apex.html.

U.S. Environmental Protection Agency (2006c). Air Quality Criteria for Ozone and Related
     Photochemical Oxidants (Final). National Center for Environmental Assessment, U.S.
     Environmental Protection Agency, Research Triangle Park, NC, EPA/600/R-05/004aF-cF.
     Available at: http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=149923

Weschler, C. J. (2000) Ozone in indoor environments: concentration and chemistry. Indoor Air
     10: 269-288.

Whitfield, R., Biller, W., Jusko, M., and Keisler, J. (1996). A Probabilistic Assessment of
     Health Risks Associated with Short- and Long-Term Exposure to Tropospheric Ozone.
     Argonne National Laboratory, Argonne, IL.

Wiley, J. A.; Robinson, J. P.; Piazza, T.; Garrett, K.; Cirksena, K.; Cheng, Y.-T.; Martin, G.
     (1991a). 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-93/489.

Williams, R., Suggs, J., Creason, J., Rodes, C., Lawless, P., Kwok, R., Zweidinger, R., and
     Sheldon, L. (2000). The 1998 Baltimore particulate matter epidemiology-exposure study:
                                      4-72

-------
     Part 2. Personal exposure associated with an elderly population. J. Expos. Anal. Environ.
     Epidemiol. 10(6):533-543.

Xue, J., McCurdy, T., Spengler, J., Ozkaynak, H. (2004). Understanding variability in time
     spent in selected locations for 7-12-year old children. J Expo Anal Environ Epidemiol
     14(3):222-33.
                                       4-73

-------
              5.     CHARACTERIZATION OF HEALTH RISKS

5.1    INTRODUCTION
       This chapter presents information regarding the results from an updated ozone (O3) health
risk assessment that builds upon the methodology used in the assessment conducted as part of the
last O3 NAAQS review. This updated assessment includes estimates of (1) risks of lung function
decrements in school age children, respiratory symptoms in asthmatic children, respiratory-
related hospital admissions, and respiratory, non-accidental, and cardiorespiratory mortality
associated with recent ambient O3 levels; and (2) risk reductions associated with just meeting the
current and several alternative 8-hr O3 NAAQS. The current risk assessment is more fully
described and presented in a technical support document, Ozone Health Risk Assessment for
Selected Urban Areas (Abt Associates, 2006a; henceforth referred to as the Risk Assessment
Technical Support Document and cited as Risk Assessment TSD).
       The goals of this O3 risk assessment are: (1) to provide estimates of the potential
magnitude of mortality and several morbidity effects associated with current O3 levels, and with
meeting the current 8-hr O3 NAAQS and alternative O3 8-hr standards, in specific urban areas;
(2) to develop a better understanding of the influence of various inputs and assumptions on the
risk estimates; and (3) to gain insights into the distribution of risks and patterns of risk reductions
associated with meeting alternative O3 standards. The risk assessment covers a variety of health
effects for which there is adequate information to develop quantitative risk estimates. However,
there are several health endpoints (e.g., increased lung inflammation, increased airway
responsiveness, impaired host defenses, increased medications use, increased asthma-elated
emergency department visits,  increased school  absences) for which there currently is insufficient
information to develop quantitative risk estimates. These additional health endpoints are
discussed qualitatively in Chapter 3 of this Staff Paper. We recognize that while there are many
sources of uncertainty and variability inherent in the inputs to this assessment, which make the
specific estimates uncertain, there is sufficient confidence in the direction and general magnitude
of the estimates provided by the assessment, particularly with respect to relative differences
between  alternative potential standards, for the assessment to serve as a useful input to decisions
on the adequacy of the O3 standard. While some of these uncertainties have been addressed
quantitatively in the form of estimated confidence ranges around central risk estimates, other
uncertainties and the variability in key inputs are not reflected in these confidence ranges, but
rather are addressed through separate sensitivity analyses or are characterized qualitatively.
       Following this introductory section, this chapter discusses the scope of the risk
assessment, including selection of urban areas and health endpoints and the degree of confidence
associated with the various health outcomes that have been associated with ambient O3

                                           5-1

-------
exposures; components of the risk model; characterization of uncertainty and variability
associated with the risk estimates; and key results from the assessment. The Risk Assessment
TSD provides a more detailed discussion of the risk assessment methodology and includes
additional risk estimates beyond those summarized herein.

       5.1.1   Overview of Risk Assessment From Last Review
       EPA conducted a health risk assessment that produced risk estimates for the number and
percent of children and outdoor workers experiencing lung function and respiratory symptoms
associated with O3 exposures for 9 urban areas. This portion of the risk assessment was based on
exposure-response relationships developed from analysis of data from several controlled human
exposure studies which was combined with exposure estimates developed for children who spent
more time outdoors and for outdoor workers.  The risk assessment for the last review also
included risk estimates for excess respiratory-related hospital admissions related to Os
concentrations for New York City, based on a concentration-response relationship reported in an
epidemiological study (Thurston et al.,  1992).  Risk estimates for lung function decrements,
respiratory symptoms, and hospital admissions were developed associated with recent air quality
levels (referred to as "as is" air quality) and for just meeting the existing 1-hr standard and
several alternative 8-hr standards. The methodological approach followed in conducting the last
risk assessment and risk estimates resulting from that assessment are described in Chapter 6 of
the 1996 Staff Paper (EPA, 1996b) and in several technical reports and publications (Whitfield et
al., 1996; Whitfield, 1997; Whitfield et al., 1998).
       In the 1997 review of the Os NAAQS, the risk estimates played a significant role in both
the staff recommendations and in the proposed and final decisions to revise the Os standards.
CAS AC stated (Wolff, 1995) in  its advice and recommendations to the Administrator on the Os
Staff Paper that "EPA's risk assessments must play a central role in identifying an appropriate
level," while also noting that "because of the myriad of assumptions that are made to estimate
population exposure and risk, large uncertainties exist in these estimates." In the 1997 notice (62
FR 38856) announcing the decision to revise the O3 standards, EPA indicated that the
Administrator considered the results of the exposure and risk analyses and key observations and
conclusions from these analyses in putting effects considered to be adverse to individuals into a
broader public health perspective and in making judgments about the level of a standard that
would be requisite to protect public health with an adequate  margin of safety.

       5.1.2   Development of Approach for Current Risk Assessment
       The health risk assessment described in this chapter and in the Risk Assessment TSD
builds upon the methodology and lessons learned from the risk assessment work conducted for
the last review. The current risk assessment also is based on the information evaluated in the

                                           5-2

-------
final CD. The general approach used in the current risk assessment was described in the draft
Health Assessment Plan (EPA, 2005a), that was released to the CASAC and general public in
April 2005 for review and comment and which was the subject of a consultation with the
CASAC Os Panel on May 5, 2005. The approach used in the current risk assessment reflects
consideration of the comments offered by CASAC members and the public on the draft Health
Assessment Plan, comments offered by CASAC members and the public on the first and second
drafts of the Staff Paper and first and second drafts of the Risk Assessment TSD at and
subsequent to a  consultation with CASAC on December 8, 2005, and at and subsequent to a
review by CASAC on August 24-25, 2006. CASAC comments reflecting both the Ozone
Panel's views and additional comments by individual members were provided to the Agency in
letters dated February 16, 2006 (Henderson, 2006a), June 5, 2006 (Henderson, 2006b) and
October 24, 2006 (Henderson, 2006c). This risk assessment chapter indicates where significant
new information has been added since the second  draft Staff Paper.
       The basic structure of the current risk  assessment reflects the two different types of
human studies on which the O3 health risk assessment is based: controlled human exposure
studies and epidemiological studies. Controlled human exposure studies involve volunteer
subjects who are exposed while engaged in different exercise regimens to specified levels of Os
under controlled conditions for specified amounts of time. For the current health risk
assessment, we  are using probabilistic exposure-response relationships based on analysis of
individual data that describe the relationship between a measure of personal exposure to Os and
measures of lung function recorded in the studies.  The measure of personal exposure to ambient
Os is typically some function of hourly exposures - e.g., 1-hr maximum or 8-hr maximum.
Therefore, a risk assessment based on exposure-response relationships derived from controlled
human exposure study data requires estimates of personal exposure to ambient  Os, typically on a
1-hr or multi-hour basis. Because data on personal hourly exposures to Os of ambient origin are
not available, estimates of personal exposures to varying ambient concentrations are derived
through exposure modeling, as described in Chapter 4. While the quantitative risk assessment
based on  controlled human exposure studies addresses only lung function responses, it is
important to note that other respiratory responses have been found to be related to  Os exposures
in these types of studies, including increased lung inflammation, increased respiratory symptoms,
increased airway responsiveness, and impaired host defenses.  Chapter 3 of this Staff Paper
provides  a more complete discussion of these additional health endpoints which are an important
part of the overall characterization of risks associated with ambient Os exposures.
       In contrast to the exposure-response relationships derived from controlled human
exposure studies, epidemiological studies provide estimated concentration-response
relationships based on data collected in real world community settings. Ambient Os
                                          5-3

-------
concentrations, measured as the average of monitor-specific measurements, using population-
oriented monitors, are used as a surrogate measure of population exposure. It is important to
consider that Os in ambient air is present in a complex mixture of air pollutants, and that some
other components of the mixture may play an important role in some of the health-related effects
observed.  It is also important to recognize that population health responses included in the
quantitative risk assessment for 63 (i.e., respiratory symptoms in asthmatic children, hospital
admissions for respiratory illness, and premature mortality) represent only a portion of the health
effects that are associated with Os exposures. As discussed more fully in Chapter 3, a wide
variety of respiratory and cardiorespiratory effects have been  shown to be related to 63
exposures including increased medication usage in asthmatics, increased doctor's visits and
emergency department visits, and increased school absences.  As described more fully below, a
risk assessment based on epidemiological studies typically requires baseline incidence rates and
population  data for the risk assessment locations.
       The characteristics that are relevant to carrying out a risk assessment based on controlled
human exposure studies versus one based on  epidemiology studies evaluated in the CD can be
summarized as follows:

   •  The relevant controlled human exposure studies in the CD provide data that can be used
       to estimate exposure-response functions, and therefore a risk assessment based on these
       studies requires as input (modeled) personal exposures to ambient 03. The relevant
       epidemiological studies in the CD provide concentration-response functions, and,
       therefore,  a risk assessment based on  these studies requires as input (actual monitored or
       adjusted based on monitored) ambient 63 concentrations, and personal exposures are not
       required as inputs to the assessment.

   •  Epidemiological studies are carried out in specific real world locations (e.g., specific
       urban areas).  To minimize uncertainty, a  risk assessment based on epidemiological
       studies has been performed for the locations in which  the studies were carried out.
       Controlled human exposure studies, carried out in laboratory settings, are generally not
       specific to any particular real world location. A risk assessment based on controlled
       human exposure studies can therefore appropriately be carried out for any location for
       which there are adequate air quality and other data on which to base the modeling of
       personal exposures. There are, therefore,  some locations for which a risk assessment
       based on controlled human exposure studies could appropriately be carried out but a risk
       assessment based on epidemiological  studies would involve greater uncertainty.

   •  The adequate modeling of hourly personal exposures associated with ambient
       concentrations for use with exposure-response relationships requires more complete
       ambient monitoring data than are necessary to estimate average ambient concentrations
       used to calculate risks based on concentration-response relationships. Therefore, there
       may be some locations in which an epidemiological studies-based risk assessment could
       appropriately be carried out, but a controlled human exposure studies-based risk
       assessment would involve greater uncertainty.
                                           5-4

-------
    •   To derive estimates of risk from concentration-response relationships estimated in
       epidemiological studies, it is usually necessary to have estimates of the baseline
       incidences of the health effects involved.  Such baseline incidence estimates are not
       needed in a controlled human exposure studies-based risk assessment.
       The scope of the current O3 risk assessment is described in the next section along with air
quality considerations that are relevant to both parts of the risk assessment.  Then, the methods
for the two parts of the risk assessment - the part based on controlled human exposure studies
and the part based on epidemiological and field studies - are discussed in sections 5.3.1 and 5.3.2
below, followed by presentation and discussion of the Os risk estimates in section 5.4.  Both
parts of the risk assessment were implemented within a new probabilistic version of TRIM.Risk,
the component of EPA's Total Risk Integrated Methodology (TRIM) model that estimates
human health risks.

5.2    SCOPE OF OZONE HEALTH RISK ASSESSMENT
       The current Os health risk assessment estimates risks of various health effects associated
with exposure to ambient 63 in a number of urban areas selected to illustrate the public health
impacts of this pollutant.  The short-term exposure related health endpoints selected for the O3
risk assessment, discussed in section 5.2.1, include those for which the CD concludes that the
evidence as a whole supports the general conclusion that 63, acting alone and/or in combination
with other components in the ambient air pollution mix is likely causal.1
       As discussed in section 3.7, we recognize that there are varying levels of confidence that
various health effect endpoints are associated with 63 at ambient levels.  As discussed in section
3.7.5 there is clear evidence of a causal relationship between lung function decrements and Os
exposures for school age children engaged in moderate exertion for 6 to  8-hours based on the
numerous controlled human exposure studies and summer camp field studies conducted by
various investigators over the last 30 years.  We also judge that there is clear evidence of a causal
relationship between increased respiratory symptoms in moderate to severe asthmatic children
and Os exposures.  There  also is strong evidence of a causal relationship between increased
respiratory-related hospital admissions and Os exposure during the warm O?, season, based on
extensive and fairly consistent epidemiological studies as well as evidence from controlled
human exposure studies reporting increased lung inflammation and airway responsiveness.
       The CD concludes that there is strong evidence which is highly suggestive of a causal
relationship between respiratory-related, non-accidental, and cardiorespiratory-related mortality
       1 As discussed in 5.2.1, certain endpoints met this criteria of likely causality, but were not included in the
risk assessment for other reasons, such as insufficient exposure-response data or lack of baseline incidence data.

                                           5-5

-------
and 63 exposures during the warm 63 season. Our judgment with respect to these health
outcomes is based on the fairly consistent positive associations found between elevated warm Os
season levels and these mortality outcomes even when the effect of PM is controlled for, and
supporting evidence about potential mechanisms of effects on the cardiovascular system from
animal toxicology, human clinical and epidemiological studies. There is certainly greater
uncertainty about these outcomes than the other effects discussed above. We also recognize, as
discussed in section 3.7.5,  that for some of the effects observed in epidemiological studies, such
as increased respiratory-related hospital admissions and non-accidental and cardiorespiratory
mortality, 63 may be serving as an indicator for reactive oxidant species in the overall
photochemical oxidant mix and that these other constituents may be responsible in whole or part
for the observed effects.
       The current risk assessment includes risk estimates for 12 urban areas. The basis for
selection of these areas is discussed below (section 5.2.2).
       Another important aspect of the current risk assessment is that the risks estimated are
only those associated with ambient O3 concentrations exceeding estimated policy-relevant
background (PRB) levels (hereafter, referred to as either "background" or "PRB" in this
chapter).2 Risks associated with concentrations above this background are judged to be more
relevant to policy decisions about the NAAQS than estimates that include risks potentially
attributable to uncontrollable background concentrations.

 5.2.1  Selection of Health Endpoint Categories
       As noted above, in the last review a significant portion of the health risk assessment
involved developing risk estimates for both lung function decrements (> 10, > 15, and > 20%
changes in FEVi) and respiratory symptoms in children (age 6 to 18 years old) who spend more
time outdoors and outdoor workers with 1-hr exposures at moderate and heavy exertion and 8-hr
exposures at moderate exertion.  As discussed in section 3.3.1.1 and Chapter 6 of the  CD, there is
a significant body of controlled  human exposure studies reporting lung function decrements and
respiratory symptoms in adults associated with 1- and 6 to 8-hr exposures to O3, as well as
similar responses in outdoor workers and others engaged in recreational outdoor activities.
        Consistent with the approach used in the last review, we judge that it is reasonable to
estimate exposure-response relationships for lung function decrements associated with Os
exposures in children  5-18 years old based on data from young adult subjects (18-35 years old).
As discussed in the 1996 Staff Paper and 1996  CD, findings from other chamber studies
       2Policy 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 is discussed in section 5.2.3.

                                           5-6

-------
(McDonnell et al., 1985) for children 8-11 years old at a single exposure level and summer camp
field studies in at least six different locations in the U.S. and Canada found lung function
decrements in healthy children similar to those observed in healthy adults exposed to Os under
controlled chamber conditions. The same approach is being used in the current assessment.
In the prior risk assessment, staff focused on the risk estimates for lung function decrements
associated with 1-hr heavy exertion, 1-hr moderate exertion, and 8-hr moderate exertion
exposures in children age 5-18 years of age. Since the 8-hr moderate exertion exposure scenario
in children who spend more time outdoors clearly resulted in the greatest health risks in terms of
both the magnitude of the lung function decrements and the percent of the population estimated
to experience these effects, and since no new information  published since the last review
suggests any changes that would impact this conclusion, we have included only the lung function
decrements (> 10, 15, and 20% FEVi) associated with 8-hr moderate exertion exposures in
children and asthmatic children (age 5 to 18 years old) in the current risk assessment.3  Risk
estimates for asthmatic school age children have been added to the risk assessment since the
second draft Staff Paper based on comments offered by the CAS AC emphasizing the importance
of health effects for this population.
       While outdoor workers and other adults who engage in moderate exertion for prolonged
periods during the day also are clearly at risk for experiencing similar lung function responses
when exposed to elevated ambient Os concentrations, the  exposure and risk assessment
conducted during the prior review suggested that school age children are at greatest risk in terms
of the number of individuals likely affected. Given the  lack of information about the number of
individuals who regularly work or exercise outdoors, we chose to focus the current quantitative
risk assessment for lung function decrements on all and asthmatic school age children.
Therefore, it is important to recognize that the current risk assessment does not account for all of
the lung function effects in the general population that would be associated with exposure to 63
under the various air quality scenarios examined.
       Although respiratory symptoms in healthy children were estimated in the last review, we
have not included this endpoint in  the current quantitative risk assessment. This is because
3Subsequent to completion of the Risk Assessment TSD, EPA analysis of uncertainty of the
exposure modeling results uncovered an error in how children are characterized as active. This
error resulted in an overestimate of the number of active children included in the exposure
estimates which are an input to the lung function risk estimates for active children.  Thus, the
lung function risk estimates provided for active  children in the Risk Assessment TSD are not
accurate and we did not include risk estimates for active children in this chapter.
                                           5-7

-------
several field studies conducted since the last review did not observe respiratory symptoms in
field studies examining responses in healthy children. The CD concludes that "collectively,
these studies indicate that there is no consistent evidence of an association between Os and
respiratory symptoms among healthy children" (CD, p. 7-55).  Thus, we decided to limit this
portion of the risk assessment to lung function decrements in children and to again base the
exposure-response relationships on data obtained for 18-35 year old subjects.
       While a number of controlled human exposure studies have reported additional health
endpoints associated with short-term exposures to Os, including airway hyperresponsiveness,
inflammation, and immune system effects, there is insufficient exposure-response data at
different concentrations to develop quantitative risk estimates for these effects. These important
additional effects are discussed in Chapter 3, and we want to emphasize that the current
quantitative risk assessment presents only a partial picture of the risks to public health associated
with short-term Os exposures.
       As discussed in the CD and Chapter 3, a significant number of epidemiological studies
examining a variety of health effects associated with ambient O3 concentrations in various
locations throughout the U.S., Canada, Europe, and other regions of the world have been
published since the last 63 NAAQS review. Chapter 3 reviews the epidemiological evidence
evaluated in Chapter 7 of the CD. In selecting health endpoints to be included in the current
quantitative risk assessment, we have focused on health endpoints that are better understood in
terms of health consequences (i.e., where there is greater consensus about the degree of response
that should be considered as  representing an adverse health effect in the population) and endpoint
categories for which the weight of the evidence supports  the inference of a likely causal
relationship between 63 and  the effect category.  Certain health endpoints met the criteria of
likely causality, but were not included in the risk assessment for other reasons, such as
insufficient exposure-response data or lack of baseline incidence data.  Based on these
considerations, the following endpoints associated with short-term exposures to O3 during the
"warm Os season" (April 1 to September 30) have been included:
   •  Respiratory  symptoms in moderate/severe asthmatic children (ages 0 to 12);
   •  Hospital admissions for respiratory illness and asthma;
   •  Premature total non-accidental and cardiorespiratory mortality.
       As discussed above in section 3.3.1.1.1, the CD also concludes that collectively, the
results of epidemiological studies suggest that respiratory symptoms and increased medication
use in asthmatic children are associated with acute exposure to Os.  These recent studies provide
very strong evidence that asthmatic children experience Os-related effects.

-------
       Large multi-city studies, as well as many studies from individual cities, have reported an
association of Os concentrations with respiratory-related hospital admissions. Studies with data
restricted to the summer or warm season, in general, indicated positive and robust associations
between ambient 63 concentrations and respiratory-related hospital admissions. With respect to
acute Os effects on mortality, the CD concludes (p.7-175) that, "The majority of the studies
suggest an elevated risk of all cause mortality associated with acute exposure to 63, especially in
the summer or warm season when Oj levels are typically high."
       As discussed in Chapter 7 of the CD and in sections 3.3.1.1.1 and 3.3.1.1.6 above, several
additional health endpoints including ED visits for respiratory illness and increased school
absences have been reported to be associated with short-term Os exposures.  The current
quantitative risk assessment does not include these additional health endpoints.  Emergency
department visits were excluded from the quantitative risk assessment because of the lack of
baseline incidence data for ED visits, as well as the more limited and less consistent database.
We also judge that the data reporting an association between short-term 63 exposures and school
absences is too limited to include in the current risk assessment.
       Table 5-1 provides a summary of the health effects and the corresponding populations for
each health effect which were included in the  quantitative risk assessment. Table 5-2 lists  some
of the health effects that have been associated with elevated Os exposures which were not
included in the quantitative risk assessment. Chapter 3 provides additional discussion of the
health effects not included in this risk assessment.

 5.2.2         Selection of Study Areas
       The criteria and considerations that went into selection of urban areas for the Os risk
assessment included the following:
    •   The overall set of urban locations should represent a range of geographic areas, urban
       population 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.
                                           5-9

-------
Table 5-1. Health Effects and Associated Population Groups Addressed in Quantitative
Risk assessment
Health Effect
Lung function decrements (FEVi)
Respiratory symptoms (chest
tightness, shortness of breath,
wheeze)
Hospital admissions:
-For respiratory illness
-Asthma-related
-Pneumonia
Mortality:
-Total (not including accidental)
-Cardiorespiratory
Population
All school age children (age 5-18)
School age asthmatic children (age 5-18)
Asthmatic children (age 0 - 12) in Boston
-Age 30+ in Los Angeles, Age 65+ in Cleveland, All ages
in New York
-All ages in New York
-Age 65+ in Detroit
All ages
Table 5-2. Health Endpoints and Associated Population Groups Not Included in the
           Quantitative Risk Assessment*
Health Effect
Lung function decrements
Respiratory symptoms (cough, chest
discomfort)
School absences for respiratory illness
Asthma-related emergency department visits
Doctors visits
Lung inflammation
Increased medication usage
Decreased resistance to infection, impaired
host defense
Population
Adults (outdoor workers, recreational exercisers,
athletes)
Adults (outdoor workers, recreational exercisers,
athletes)
Children
Asthmatics
Adults and children
Adults and children
Asthmatic children and adults
Adults and children
The list of health endpoints and populations not included in the risk assessment is not a comprehensive list, but
rather provides a general indication of the types of health endpoints that are associated with exposures to ozone but
not included in the quantitative risk assessment.
                                            5-10

-------
    •   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
       •     Sacramento
       •     St. Louis
       •     Washington, D.C.
As discussed in Chapter 4, for the purposes of estimating population exposure and the risk of
lung function decrements associated with these population exposure estimates, the 12 urban
areas have been defined based on consolidated statistical areas (CSAs).  The population
estimates for these 12 urban area CSAs are given in Table 4-3. About 40% of the total U.S.
urban population (88.5 million persons) resides in these 12 urban areas including 18.3 million
school age children (ages 5 to 18). In contrast to the risk assessment for lung function
decrements, for the risk estimates for premature mortality and excess hospital admissions, the
urban areas have been defined to be generally consistent with the geographic boundaries used in
the  epidemiological studies which were the source of the concentration-response functions used
in this risk assessment.  In  most cases the epidemiological studies only included the core urban
county or a limited number of counties in one or more of the 12 urban areas. In addition,
estimates of respiratory symptoms in asthmatic children were developed for one urban area (the
Boston CSA).

     5.2.3    Air Quality Considerations
      Both the portion of the risk assessment based on controlled human exposure and the
portion based on epidemiological studies include risk estimates for a recent year of air quality
(labeled "as is" air quality in the Risk Assessment TSD) and for  air quality adjusted so that it
simulates just meeting the current and several alternative 8-hr 63 NAAQS based on a recent
                                          5-11

-------
three-year period (2002-2004). This period was selected to represent the most recent air quality
data for which complete data were available when the risk assessment was conducted.
      In order to estimate health risks associated with just meeting the current and alternative 8-
hr 63 NAAQS, it is necessary to estimate the distribution of hourly 63 concentrations that would
occur under any given standard. Since compliance with the current Os standard is based on a 3-
year average,  air quality data from 2002 to 2004 have been used to determine the amount of
reduction in O3 concentrations required to meet the current standard.  Estimated design values4
are used to determine the adjustment necessary to just meet the current 8-hr daily maximum
standard. The amount of control has then been applied to each year of data (2002, 2003, and
2004) to estimate risks for a single Os season or  single warm Os season,  depending on the health
effect, in each of these individual years.  As described in section 4.5.6 and in more detail in
Rizzo (2006), after considering several approaches, we concluded that the Quadratic air quality
adjustment procedure generally best represented the pattern of reductions across the Os air
quality distribution observed over the last decade. The Quadratic air quality adjustment
procedure was applied in each of the 12 urban areas to the 2002 and 2004 O3 air quality data and
in a subset of 5 urban areas (Atlanta, Chicago, New York, Houston, and Los Angeles) to the
2003 63 air quality data, based on the 3-year period (2002-2004) 63 design values, to generate
new time series of hourly  Os concentrations for 2002, 2003, and 2004 that reflect air quality
levels that just meet the current 8-hr Os standard over this three year period. Risk estimates
associated with 2003 63 monitoring data and 2003 air quality adjusted to just meet the current
and alternative 8-hr standards have been  added to the assessment since the second draft Staff
Paper.
       We note that since compliance with the current standard is based on the 3-year average
of the 4th-highest daily maximum 8-hr values, the air quality distribution in each of the 3 years
can, and generally does, vary. As a consequence, the risk estimates associated with air quality
just meeting the current standard also will vary depending on the year chosen for the analysis.
We include assessments involving adjustment of both 2002 and 2004 air quality data to illustrate
the magnitude of this year-to-year variability in the risk estimates.  The year 2002 generally had
meteorology that was very conducive to producing Os over the eastern half of the U.S. and this
resulted in the highest Os levels over the  2002-2004 time period in the vast majority of the 12
urban study areas. In contrast, 2004 was a year associated with an unusually cool and rainy
       4A 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 O3 NAAQS, the 3-year average of the annual 4th-highest daily
maximum 8-hr average concentrations, based on the monitor within an urban area yielding the highest 3-yr average,
is the design value.

                                           5-12

-------
summer in the eastern half of the U.S. and this contributed to the fact that the lowest 63 levels
over this same three-year period were observed in this year in most of the urban areas included in
the assessment. The lower Os levels observed in 2004 also were lower, in part, as a result of
reductions in nitrogen oxides (NOX) emissions associated with implementation of additional
regional controls on large power plants in the eastern half of the U.S. Differences in
meteorology were less evident in Texas and California and these latter areas also were not
impacted by the recent additional regional controls imposed on large power plants. Thus, its not
surprising that the daily maximum 8-hr levels observed in Houston in 2004 were somewhat
higher than those observed in 2002 and that 8-hr levels were similar in Los Angeles between
these two years.  The risk results for 2002 and 2004, thus, provide generally lower-end and
upper-end estimates of the annual risks that can  occur over a three-year period when alternative
standards are just met. Daily maximum 1-hr and 8-hr 63 levels in 2003 generally fell
somewhere between 2002 and 2004 levels in most of the 12 urban areas.
       As noted earlier, the risk estimates developed for both the recent air quality scenario and
just meeting the current 8-hr standard represent risks associated with 63 levels in excess of
estimated background concentrations. The results of the global tropospheric 63 model GEOS-
CHEM have been used to estimate average background O3 levels for different geographic
regions across the U.S.  These GEOS-CHEM simulations include a background simulation in
which North American anthropogenic emissions of NOX, non-methane volatile organic
compounds, and carbon monoxide are set to zero, as described in Fiore et al. (2003). We
estimated monthly background concentrations for each of the 12 urban areas based on the
GEOS-CHEM simulations, including daily diurnal profiles which were fixed for each day of
each month during the 63 season (See Appendix 2-A of this  Staff Paper for plots of these
estimated background values).  The CD and section 2.7 above indicate that background 63
concentrations at the surface are generally predicted to be in the range of 0.015 to 0.035 ppm in
the afternoon, and they decline under conditions conducive to Cb episodes. They are highest
during spring and decline into summer. This range is lower than the estimated range of 0.03 to
0.05 ppm for typical summertime background levels included in the 1996 CD and the single
value of 0.04 ppm used for background in the prior risk assessment.
       As discussed in section 2.7, the CD summarizes its evaluation of the validity of the
GEOS-CHEM model, and states "in conclusion, we estimate that the PRB ozone values reported
by Fiore et al. (2003) for afternoon surface air over the United States are likely 10 ppbv too high
in the southeast in summer, and accurate within  5 ppbv in other regions and seasons" (CD, p. 3-
                                          5-13

-------
53 ). These error estimates are based on comparison of model output with observations for
conditions which most nearly reflect those given in the PRB definition (i.e., at the lower end of
the probability distribution). For O3 (cf Figures 8 and 9 of Fiore et al. (2003) for the southeast
and Figure  3 of Fiore et al. (2002) for the northeast) it can be seen that GEOS-CFIEM
overestimates Os for the southeast and underestimates it for the northeast. Sensitivity analyses
examining the impact of alternative estimates for background on lung function and mortality risk
estimates have been developed since the second draft Staff Paper and are presented in section
5.4.3. As discussed in section 5.4.3, estimated risk reductions associated with alternative
standards relative to just meeting the current standard are generally unaffected by the estimated
PRB levels within the range of alternative estimates examined.

5.3    COMPONENTS OF THE RISK MODEL
      As noted above in section 5.1.2, there are two parts to the health risk assessment: one
based on combining information from controlled human exposure studies with modeled
population  exposure and the other based on combining information from community
epidemiological  studies with either monitored or adjusted ambient concentrations levels.  Section
5.3.1 below discusses the portion of the current risk assessment related to effects reported in
controlled human exposure studies and section 5.3.2 below discusses the portion of the current
risk assessment related to health effects reported in community epidemiological studies.

       5.3.1   Assessment of Risk Based on Controlled Human Exposure Studies
        5.3.1.1 General Approach
       The major components of the portion of the health risk assessment based on data from
controlled human exposure studies are illustrated in Figure 5-1. As shown in Figure 5-1, under
this portion of the risk assessment, exposure estimates for a number of different air quality
scenarios (i.e, recent year of air quality, just meeting the current 8-hr standard, just meeting
alternative  standards, and background) are combined with probabilistic exposure-response
relationships derived from  the controlled human exposure studies to  develop risk estimates
associated with 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 in school aged children engaged in
moderate exertion for 8 hours.  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 Exposure Assessment TSD.
                                          5-14

-------
       Several risk measures were generated for this portion of the risk assessment. In addition
to the estimates of the number of all and asthmatic school age children experiencing one 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
these same population groups.  The mean number of occurrences per child has been calculated to
provide an indicator of the average number of times that a responder would experience the
specified effect during an O3 season. The population sizes for all and asthmatic school age
children for each of the 12 urban areas used in this part of the risk assessment are given in Table
4-3 of this Staff Paper.  We note that the asthmatic school age children subpopulation is a subset
of all school age children, and thus the risk estimates presented for these two groups should not
be combined.
       A population risk estimate for a given lung function decrement (e.g., > 15% change in
FEVi) is an estimate of the expected number of people who will experience that lung function
decrement. Since we are interested in risk estimates associated with Os concentrations in excess
of background concentrations, the following steps were taken to estimate the risk associated with
recent conditions in excess of background:  (1) expected risk given the personal exposures
associated with recent ambient 63 concentrations was estimated, (2) expected risk given the
personal exposures associated with estimated background ambient O3 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. See Appendix 5B.2 for additional information concerning
notation and the derivation of the risk estimate algorithms used in this risk assessment. Because
response rates are calculated for 21 fractiles (i.e., 0.01, 0.05, 0.10, 0.15 ... 0.50, 0.55, ... 0.90,
0.95, 0.99), estimated population risks are similarly fractile-specific.
       The risk (i.e., expected fractional response rate) for the kth fractile, Rk is:


              Rk=^PjX(RRk e^ -  ^Pfx(RRk   ef)   (Equation 5-1)
                   i=\                     t=\
where:
       Cj = (the midpoint of) they'th category of personal exposure to 63, given recent ambient
       Os concentrations;

       ef= (the midpoint of) the ith category of personal exposure to 63, given background
       ambient O3 concentrations;
                                          5-15

-------
       Pj = the fraction of the population having personal exposures to Os concentration
       ppm, given recent ambient Os concentrations;

       P* = the fraction of the population having personal exposures to Os concentration of
       ef ppm, given background ambient Os concentrations;

       RRk | ej = k-fractile response rate at 63 concentration BJ;

       RRk | e\ = k-fractile response rate at 63 concentration ef; and

       N= number of intervals (categories) of Os personal  exposure concentration, given recent
       ambient Os concentrations; and

       Nb = number of intervals of Os personal exposure concentration, given background
       ambient Os concentrations.

       For example, if the median expected response rate for recent ambient concentrations is
0.065 (i.e., the median expected fraction of the population responding is 6.5%) and the median
expected response rate for background ambient concentrations is 0.001 (i.e., the median expected
fraction of the population responding is 0.1%), then the median expected response rate
associated with recent ambient concentrations above background concentrations is 0.065 - 0.001
= 0.064.  If there are 300,000 people in the relevant population, then the population risk is 0.064
x 300,000= 19,200.5
       5A normalization procedure had to be applied to the number of responders (or the number of occurrences of
response) given personal exposures associated with recent ambient O3 concentrations (or concentrations rolled back
to simulate just meeting a standard) because the population size used in the exposure runs for background
concentrations were not identical to those used in the exposure runs conducted for the  recent air quality and
alternative standard scenarios. This normalization procedure is described in section 3.3.1 of the Risk Assessment
TSD.

                                            5-16

-------
Figure 5-1. Major Components of Ozone Health Risk Assessment Based on Controlled Human Exposure Studies
    Air Quality
         Ambient
       Monitoring for
      Selected Urban
          Areas
         Modeled
        Background
      Concentrations
        Air Quality
        Adjustment
        Procedures
        Current and
        Alternative
         Proposed
         Standards
 Recent
 ("As Is")
 Ambient
O3 Levels
                    Exposure
                    Exposure
                     Model
Exposure Estimates
Associated with:
•Background
Concentrations
•Recent Air Quality
•Current Standard
•Alternative
Standards
   Exposure-Response
       Controlled Human
       Exposure Studies
                         Probabilistic
                          Exposure -
                          Response
                         Relationships
                                          5-17
                                                            Health
                                                             Risk
                                                            Model
Risk Estimates:

• Recent Air
  Quality
• Current
  Standard
• Alternative
  Standards

-------
        5.3.1.2 Exposure Estimates
       Exposure estimates used in this portion of the risk assessment were obtained from
running TRIM.Expo for each of the 12 urban areas for the various air quality scenarios (i.e., for
2004 and 2002 air quality representing recent years, for 2004 and 2002 air quality adjusted to
just meet the current and several potential alternative 8-hr standards, for 2003 air quality and
2003 air quality adjusted to just meet several potential alternative standards in 5 urban areas, and
for air quality levels representing background based on estimates from the GEOS-CHEM
model). Chapter 4 and the Exposure Assessment TSD (EPA, 2006d) provide additional details
about the inputs and methodology used to estimate population exposure in the 12 urban areas.
Exposure estimates for all and asthmatic school age children (ages 5 to 18) were separately
combined with probabilistic exposure-response relationships for lung function decrements
associated with 8-hr exposure while engaged in moderate exertion. Individuals engaged in
activities that resulted in an average equivalent ventilation rate (EVR) for the 8-hr period at or
above 13 1/min-m2 were included in the exposure estimates for 8-hr moderate or greater exertion.
This range was selected to match the EVR for the group of subjects in the controlled human
exposure studies that were the basis for the exposure-response relationships used in this portion
of the risk assessment.

        5.3.1.3 Exposure-Response Functions
       As described in section 3.1.2 of the Risk Assessment TSD, a Bayesian Markov Chain
Monte Carlo approach was used to estimate probabilistic exposure-response relationships for
lung function decrements associated with 8-hr moderate exertion exposures using the WinBUGS
software (Spiegelhalter et al., 1996).6 The combined data set including the data from the
Folinsbee et al. (1988), Horstman et al. (1990), and McDonnell et al. (1991) studies used
previously and the more recent data from Adams (2002,  2003, 2006) have been used to estimate
exposure-response relationships for 8-hr exposures under moderate exertion for each of the three
measures of lung function decrement listed above.  Table 5-3 presents a summary of the study-
specific results based on correcting all individual responses for the effect on lung function
decrements of exercise in clean air.  The previously used studies were all conducted in EPA's
facility in Chapel Hill, while the Adams studies were conducted at the University of California at
       6See Gelman et al. (1995) or Gilks et al. (1996) for an explanation of these methods.
                                       5-18

-------
Table 5-3.  Study-Specific Exposure-Response Data for Lung Function Decrements
Study
Protocol
Change in FEV!>10%
Number
Exposed
Number
Responding
Change in FEV!>15%
Number
Exposed
Number
Responding
Change in FEV!>20%
Number
Exposed
Number
Responding
0.04ppmO3
Adams (2006)
Adams (2002)
Triangular
Square-wave, face mask
30
30
0
2
30
30
0
0
30
30
0
0
0.06ppmO3
Adams (2006)
Square-wave
Triangular
30
30
2
2
30
30
0
2
30
30
0
0
0.08ppmO3
Adams (2006)
Adams (2003)
Adams (2002)
F-H-M*
Square-wave
Triangular
Square-wave, chamber
Square-wave, face mask
Variable levels (0.08 ppm
avg), chamber
Variable levels (0.08 ppm
avg), face mask
Square-wave, face mask
Square-wave
30
30
30
30
30
30
30
60
7
9
6
9
6
5
6
18
30
30
30
30
30
30
30
60
2
3
2
3
1
3
5
11
30
30
30
30
30
30
30
60
1
1
1
1
1
0
2
5
0.1 ppm O3
F-H-M
Square-wave
32
13
32
9
32
5
0.12 ppm O3
Adams (2002)
F-H-M
Square-wave, chamber
Square-wave, face mask
Square-wave
30
30
30
17
21
15
30
30
30**
12
13
15**
30
30
30
10
7
6
               combined data from Folinsbee et al. (1988), Horstman et al. (1990), and McDonnell et al. (1991).
                fas sufficiently inconsistent with the other data from the F-H-M combined data set that it was considered an outlier and was not included in the
*F-H-M includes
**This data point
analysis.
                                                         5-19

-------
Davis. Data from these controlled human exposure studies were corrected for the effect of
exercise in clean air to remove any systematic bias that might be present in the data attributable
to an exercise effect.  Generally, this correction for exercise in clean air was small relative to the
total effects measures in the (Vexposed cases.
       For the risk assessment conducted during the last Os NAAQS review, there were data for
only 3 exposure levels (0.08, 0.10, and 0.12 ppm)7 and a linear exposure-response relationship
was estimated for use in the risk assessment.  With the addition of data from three more recent
Adams studies8 that included 0.04, 0.06, and/or 0.08 ppm, 6.6 hour exposures, the combined data
set appears to be more S-shaped, although there is still considerable uncertainty about the overall
functional form, given the limited data at exposure levels below 0.08 ppm.  Consistent with
advice from the CASAC O3 Panel in its October 24, 2006 letter (Henderson, 2006c), EPA
considered both linear and logistic functional forms in estimating the exposure-response
relationship and revised this aspect of the assessment by adopting a Bayesian Markov Chain
Monte Carlo approach. This Bayesian estimation approach incorporated both model uncertainty
and uncertainty due to sampling variability.
       We chose a 90 percent logistic/10 percent linear split as the base case for the current risk
assessment based on the following considerations: 1) the prior 1997 risk assessment had used a
linear form consistent with the advice from the CASAC Os Panel at the time that a linear model
reasonably fit the available lung function response data at 0.08, 0.10, and 0.12 ppm from three
6.6  hour exposure studies, 2) with the addition of data at 0.06 and 0.04 ppm, a logistic model
provided a very good fit to the data, and 3) as members of the current CASAC Os Panel have
noted there is only very limited data at the two lowest exposure levels and, therefore, a linear
model cannot entirely be ruled out. We have included a sensitivity analysis that examines the
impact on the lung function risk estimates of two alternative choices ,  an 80 percent logistic/20%
linear split and 50% logistic/50% linear split (see section 5.4.3.2).
       For each of the three measures of lung function decrement, we assumed for the base case
a 90% probability that the exposure-response function has the following 3-parameter logistic
form:9'10
       7The studies conducted in EPA's facility in Chapel Hill that are considered in the lung function risk
assessment measured O3 concentrations to within +/- 5% or +/- 0.004 ppm at the 0.08 ppm exposure level.
       8These studies reported O3 concentrations to be accuate within +/- 0.003 ppm over the range of
concentrations included in these studies.
       9As noted in Whitfield et al., 1996, the response data point associated with 0.12 ppm for the response
measure FEVi > 15% appeared to be inconsistent with the other data points (see Whitfield et al., 1996, Table 10,
footnote c). Because of this, we estimated the probability of a response of FEVi > 15% at an O3 concentration of
0.12 ppm by interpolating between the FEVi ^ 10% and FEVi ^ 20% response rates at that O3 concentration.

                                        5-20

-------
                      a*er(l-eftc}
         (x; a, 0, y) =        ^    J  ,                        (Equation 5-2)
         \    ,f,n                +'                        V  4         '
where x denotes the 63 concentration (in ppm) to which the individual is exposed, y denotes the
corresponding response (decrement in FEVi > 10%, > 15% or > 20%), and a, /?, and y are the
three parameters whose values are estimated.
       We assumed for the base case a 10 percent probability that the exposure-response
function has the following linear (hockeystick) form:

                                 [a + fie,  for a + ftc> 0
                     y(x;a,P) = \                            (Equation 5-3)
                                 [0, fora + flx<0

We assumed that the number of responses, S, out of TV subjects exposed to a given concentration,
x, has a binomial distribution with response probability given by model  (5-1) with 90 percent
probability and response probability given by model (5-2) with 10 percent probability.  In some
of the controlled human exposure studies, subjects were exposed to a given Os concentration
more than once - for example, using a square-wave exposure pattern in one protocol and a
triangular exposure pattern in another protocol.  However, because there were insufficient data to
estimate subject-specific response probabilities, we assumed a single response probability (for a
given definition of response) for all individuals and treated the repeated exposures for a single
subject as independent exposures in the binomial distribution.
       For each of the two functional forms (logistic and linear), we derived a Bayesian
posterior distribution using this binomial likelihood function in combination with prior
distributions for each of the unknown parameters. We assumed lognormal priors with maximum
likelihood estimates of the means and variances for the parameters of the logistic function, and
normal priors, similarly with maximum likelihood estimates for the means and variances, for the
parameters of the linear function. For each of the two functional forms  considered, we used
1000 iterations as the "burn-in" period followed by 9,000 iterations for the estimation.  Each
iteration corresponds to a set of values for the parameters of the (logistic or linear) exposure-
response function.  We then combined the 9,000 sets of values from the logistic model runs with
the last 1,000 sets of values from the linear model runs to get a single combined distribution of
       10The 3-parameter logistic function is a special case of the 4-parameter logistic, in which the function is
forced to go through the origin, so that the probability of response to 0.00 ppm is 0.

                                       5-21

-------
10,000 sets of values reflecting the 90 percent/10 percent assumption stated above.  As noted
above, sensitivity analyses examining the impact on lung function risk estimates of two
alternative choices are presented in section 5.4.3.2.
       For any 63 concentration, x, we could then derive the nth percentile response value, for
any n, by evaluating the exposure-response function at x using each of thelO,000 sets of
parameter values (9,000 of which were for a logistic model and 1,000 of which were for a linear
model). The resulting 2.5th percentile, median (50th percentile), and 97.5th percentile exposure-
response functions for changes in FEVi > 10%, > 15% and > 20% are shown separately in
Figure 5-2a, b, and c along with the response data to which they were fit.
       We note that the fraction of the population experiencing FEVi > 10, 15, and > 20%
associated with 0.08 ppm Os exposures was generally lower in the three Adams studies
compared to the combined data set based on the studies by Folinsbee et al. (1991), Horstmann et
al. (1990), and McDonell et al. (1991). For example, the fraction of the population experiencing
FEVi decrements >15% associated with 0.08 ppm 63 exposures ranged from 3.3 to!6.7% in the
three Adams studies compared to 18.3% in the combined data set from the Chapel Hill studies.
The 0.08 ppm level is the only common level tested in both sets of studies.  This observed
difference may be due to differences in sensitivity of the subjects tested, random variability due
to the relatively small number of subjects tested, and/or possibly greater attenuation of response
for subjects living in or near Davis, California (where the Adams studies were conducted)
compared to subjects living in or near Chapel Hill, NC (where the other studies were conducted).
Adams notes in his studies that they were conducted over a 6-month period when the 0.09 ppm,
1-hr California standard was not exceeded in the area where his subjects resided.  The difference
in observed responses between these two sets of studies is an additional uncertainty that should
be considered.
       As noted above, the Risk Assessment TSD includes risk estimates for all three measures
of lung function response (i.e., > 10, 15, and 20% decrements in FEVi). However, we  are
focusing on FEVi decrements >15% for all school age children and >10% FEVi decrements for
asthmatic school age children  in this Staff Paper, consistent with the advice from CAS AC
expressed in its October 24, 2006 letter (Henderson, 2006c) that these levels of response
represent indicators of adverse health effects for these populations.

        5.3.1.4 Characterizing Uncertainty and Variability
       An important issue associated with any population health risk assessment is the
characterization of uncertainty and variability.  Uncertainty refers to the lack of knowledge
regarding both the actual values of model input variables (parameter uncertainty) and the
physical systems or relationships (model uncertainty - e.g., the  shapes of concentration-response
                                      5-22

-------
functions). In any risk assessment, uncertainty is, ideally, reduced to the maximum extent
possible, but significant uncertainty often remains. It can be reduced by improved measurement
and improved model formulation. In addition, the degree of uncertainty can be characterized,
sometimes quantitatively.  For example, the statistical uncertainty surrounding the estimated 63
coefficients in the exposure-response functions is reflected in the credible intervals provided for
the risk estimates in this chapter and in the Risk Assessment TSD.
       As described in section 3.1.3 of the Risk Assessment TSD and section 5.3.1.3 above, we
have revised the approach used in the second draft Staff Paper and have now used a Bayesian
Markov Chain Monte  Carlo approach to characterize uncertainty attributable to sampling error.
Using this approach, we could derive the nih percentile response value, for any n, for any Oj
concentration, x, as described above.  Because the exposure estimates were generated at the
midpoints of 0.01 ppm intervals (i.e., for 0.005 ppm, 0.015 ppm, etc), we derived 2.5th percentile,
50th percentile (median), and 97.5th percentile response estimates for Oj concentrations at these
midpoint values. As illustrated in Figure 5-2a, b, and c, for each of the lung function response
definitions, the 2.5th percentile and  97.5th percentile response estimates comprise the lower and
upper bounds of the credible interval around each point estimate (median estimate) of response.
       As noted above, the exposure-response relationships shown in Figures 5-2a, b, and c are
based on the assumption that the relationship between exposure and response has a logistic form
with 90 percent probability and a linear (hockeystick) form with 10 percent probability. The
resulting 2.5th percentile, median (50th percentile), and 97.5th percentile exposure-response
functions for decrements in FEVi > 10% and > 15% for the base case and two alternative
exposure-response functions, based on an 80 percent logistic/20 percent linear split and a 50
percent logistic/50 percent linear split are shown in Figures 5-3a and b  To aid comparison
between the base case and the two alternative exposure-response functions,  Figures 5-4a and b
show the median exposure-response relationships in the same graph for these same two FEVi
endpoint definitions for the base  case and two alternative exposure-response functions.  Section
5.4.4 presents results from a sensitivity analysis that examines the impact of using these two
alternative exposure-response relationships on the lung function risk estimates for all and
asthmatic school age children.
       In addition to uncertainties arising from sampling variability considerations, there are
other uncertainties associated with the use of the exposure-response relationships for lung
function responses.  For example, while we have used the combined data set for the current risk
assessment, as it represents the best available data, we believe that the observed differences in
response between the Adams studies and the Chapel Hill  studies contribute to additional
uncertainty about the exact shape of the exposure-response relationship, especially for levels at
                                       5-23

-------
Figure 5-2.  a, b, c. Probabilistic Exposure-Response Relationships for FEVi Decrement > 10%,  >
                    15%, and  > 20% for 8-Hour Exposures Under Moderate Exertion*
                                        a)  FEVi Decrement > 10%
                                                                              •   Original Response Data
                                                                           	2.5th percentile curve
                                                                           	median curve
                                                                           	97.5th percentile curve
                       0.02     0.04    0.06     0.08     0.1     0.12     0.14
                                  Ozone Concentration (ppm)
                                        b)  FEVt Decrement > 15%
           100% -i
   *   Original Response Data
	2.5th percentile curve
	median curve
— •	97.5th percentile curve
                       0.02     0.04    0.06     0.08     0.1
                                  Ozone Concentration (ppm)
                                                             0.12
                                                                     0.14
                                        c)  FEVi Decrement > 20 %


HI
c
o
Q.
obability of a Re
o_




70%
60%
50%
40%
30%
20%
10%
0%
C




*— *•" """* """ *""
r^**^^
) 0.02 0.04 0.06 0.08 0.1 0.12 0.14
Ozone Concentration (ppm)


perceniecurve
— , , .- - 97.5th percentile curve



*Derived from Folinsbee et al, 1988; Horstman et al. 1990; McDonnell et al, 1991; Adams 2002, 2003, 2006).
Each curve assumes a 90% probability that the form of the exposure-response relationship is logistic and 10%
probability that the form is linear (see text above).
                                             5-24

-------
         Figure 5-3a, b, c.  Probabilistic Exposure-Response Relationships for FEVi Decrement > 10% and > 15% for 8-Hour Exposures
                           Under Moderate Exertion: Comparison of 90% Logistic/10% Linear (Hockeystick) Split and 80% Logistic/20%
                           Linear (Hockeystick) and 50% Logistic/50% Linear (Hockeystick) Splits in Assumed Relationship Between
                           Exposure and Response*
FEVj Decrement > 10%: 90% Logistic/10% Linear      FEVj Decrement > 10%: 80% Logistic/20% Linear
        o%
             0.02  0.04  0.06  0.08  0.1  0.12 0.14
               Ozone Concentration (ppm)
0  0.02 0.04 0.06  0.08  0.1  0.12 0.14
      Ozone Concentration (ppm)
FEVj Decrement > 15%: 90% Logistic/10% Linear      FEVj Decrement > 15%: 80% Logistic/20% Linear
         0   0.02  0.04 0.06 0.08 0.1  0.12  0.14
               Ozone Concentration (ppm)
0  0.02 0.04  0.06  0.08  0.1  0.12 0.14
      Ozone Concentration (ppm)
                                          FEV: Decrement > 10%: 50% Logistic/50% Linear
                                                      0.02  0.04 0.06  0.08  0.1  0.12  0.14
                                                         Ozone Concentration (ppm)
                                          FEY: Decrement> 15%: 50% Logistic/50% Linear
                                                      0.02  0.04 0.06  0.08  0.1  0.12  0.14
                                                         Ozone Concentration (ppm)
         *Derived from Folinsbee et al., 1988; Horstman et al. 1990; McDonnell et al., 1991; Adams 2002, 2003, 2006.
                                                              5-25

-------
Figure  5-4.  Median Exposure-Response Functions Using Three Different Combinations of
              Logistic and Linear (Hockeystick) Models
    70%
    60%
    50%
    40%
    30%
    20%
    10%
                               Figure 5-4a. FEVi Decrements > 10%
                                                                      90% logistic-10% linear
                                                                  - ••- 80% logistic - 20% linear
                                                                  - - * • -50% logistic - 50% linear
       0.00
50%

45%

40%

35%

30%

25%

20%

15%

10%

 5%
                    0.02
                                 0.04
                                               0.06           0.08
                                               Ozone Exposure (ppm)
                                                                         0.10
                                                                                      0.12
                               Figure 5-4b. FEVi Decrements > 15%
                                                                                                    0.14
                                                                                                   ..j
                                                                       —•—90% logistic -10% linear
                                                                       - -•- 80% logistic - 20% linear
                                                                       - - * - - 50% logistic - 50% linear
     0%
       0.00
              -m—•—«—
                    0.02
                                 0.04
                                               0.06           0.08
                                               Ozone Exposure (ppm)
                                                                         0.10
                                                                                      0.12
                                                                                                    0.14
                                              5-26

-------
   or below 0.08 ppm. Additional uncertainties with respect to the estimated exposure-response
   relationships are briefly summarized below.11  These additional uncertainties include:

   •   Length of exposure. The 8-hr moderate exertion risk estimates are based on a combined
       data set from six controlled human exposure studies conducted using 6.6-hr exposures.
       The use of these data to estimate responses associated with an 8-hr exposure are
       reasonable, in our judgment, because lung function response appears to level off after
       exposure for 4 to 6 hours.  It is unlikely that the exposure-response relationships would
       have been appreciably different had the studies been conducted over an 8-hr period.

   •   Extrapolation of exposure-response relationships.  It was necessary to estimate responses
       at 63 levels below the lowest exposure levels used in the controlled human studies (i.e.,
       0.04 ppm) down to background levels.

   •   Reproducibility of Conduced responses.  The risk assessment assumed that the (V
       induced responses for individuals are reproducible. This assumption is supported by the
       evaluation in the CD (see section AX6.4) which cites studies by McDonnell et al.
       (1985b) and Hazucha et al. (2003) as showing significant reproducibility of response.

   •   Age and lung function response. As in the prior review, exposure-response relationships
       based on  controlled human exposure studies involving 18-35 year old subjects were used
       in the risk assessment to estimate responses for school age children (ages 5-18). This
       approach is supported by evaluation in the CD (see section AX6.4) which cites the
       findings of McDonnell et al. (1985a) who reported that children 8-11 years old
       experienced FEVi responses similar to those observed in adults 18-35 years old when
       both groups were exposed to concentrations of 0.12 ppm at an EVR of 35 L/min/m2. In
       addition,  a number of summer camp studies of school age children exposed in outdoor
       environments in the northeast also showed (Vinduced lung function changes similar to
       and in some cases somewhat larger than those observed in controlled human exposure
       studies.

   •   Exposure history.  The risk assessment assumed that the (Vinduced response on any
       given day is independent of previous Os exposures. As discussed in Chapter 3 and in the
       CD, (Vinduced responses can be enhanced on the second day of exposure or attenuated
       after more than 2 consecutive days of exposure.  The possible impact of recent exposure
       history on the risk estimates is an additional source of uncertainty that is not quantified in
       this assessment. We note that the three Adams studies which were conducted in Davis,
       California reported a smaller fraction of the subjects experiencing FEVi decrements >15
       and 20%  associated with 63 exposures to 0.08 ppm for 6.6 hours than the
       Folinsbee/Horstman/McDonnell studies conducted in Chapel Hill, NC at this same level
       and exposure period.  While Adams indicates in each of these studies that Os levels did
       not exceed the 0.09 ppm, 1-hr California standard, we do not know whether the
       exposures outside the chamber played any role in the differences observed between these
       "Additional uncertainties with respect to the exposure inputs to the risk assessment are described in
Chapter 4 of this Staff Paper, in the Exposure Assessment TSD, and in Langstaff (2007).

                                         5-27

-------
       two sets of studies or whether the differences might reflect differential sensitivity among
       the pools of subjects tested.

    •   Exposure-response relationships for all and asthmatic school age children. The risk
       assessment used the same Os exposure-response relationships developed from data on
       "healthy" subjects, for all and asthmatic school age children.  Based on evidence from
       epidemiological studies, it is likely that moderate to severe asthmatic children would
       experience greater lung function decrements than other children without these conditions.
       This would tend to lead to underestimating the lung function decrements for asthmatic
       children in the current risk assessment. One consideration working in the opposite
       direction is that the activity patterns used in the exposure analysis to estimate exposures
       for asthmatic children were not specific to asthmatic individuals.  To the extent that
       asthmatic children, especially those with moderate to severe asthma, are less active or
       spend less time outdoors than other children of the same age, the estimates of their 8-hr
       exposures to 63 under moderate exertion may be overstated. This factor would tend to
       lead to overestimates of risks for lung function decrements in the asthmatic school age
       population.

    •   Interaction between O^ and other pollutants. Because the  controlled human exposure
       studies used in the risk assessment  involved only Os  exposures, it was assumed that
       estimates of (Vinduced health responses would not be affected by the presence of other
       pollutants (e.g., SC>2, PM2 5). Some evidence exists that other pollutants may enhance the
       respiratory effects associated with exposure to Os, but the evidence is not consistent
       across studies.
       Variability refers to the heterogeneity in a population or variable of interest that is
inherent and cannot be reduced through further research. The current controlled human exposure
studies portion of the risk assessment incorporates some of the variability in key inputs to the
analysis by using location-specific inputs for the exposure analysis (e.g., location-specific
population data, air exchange rates, air quality and temperature data). Although spatial
variability in these key inputs across all U.S. locations has not been fully  characterized,
variability across the selected locations is embedded in the analysis by using, to the extent
possible, inputs specific to each urban area.  The extent to which there is variability in the
exposure-response relationships for the populations included in the risk assessment across
different geographic areas is currently unknown. Temporal variability also is more difficult to
address, because the risk assessment focuses on some unspecified time in the future. To
minimize the degree to which values of inputs to the analysis may be different from the values of
those inputs at that unspecified time, we have used the most current inputs available - for
example, year 2002, 2003, and 2004 air quality data for the urban locations, and the most recent
available population data (from the 2000 Census).  However, future changes in inputs have not
been predicted (e.g., future population levels).
                                         5-28

-------
       5.3.2   Assessment of Risk Based on Epidemiological Studies
       As discussed above, the current quantitative risk assessment based on epidemiological
studies includes risk estimates for respiratory symptoms in moderate to severe asthmatic
children, respiratory-related hospital admissions, and total non-accidental and cardiorespiratory
mortality associated with short-term Os exposures in selected urban locations in the U.S.  We
want to emphasize that there is considerable evidence that Os exposures also results in additional
respiratory-related effects beyond those included in the quantitative risk assessment.  These
effects include, e.g., increased school absences, increased asthma-related emergency department
visits, and increased medication usage in asthmatics.  These additional effects are discussed in
Chapter 3 and considered in the overall risk characterization presented in Chapter 6. The
methods used in the epidemiological portion of the quantitative risk assessment are described
below.

        5.3.2.1 General Approach
       The general approach used in this part of the risk assessment relies upon concentration-
response functions which have been estimated in epidemiological studies evaluated in the CD.
Since these studies estimate concentration-response functions using ambient air quality data from
fixed-site, population-oriented monitors, the appropriate application of these functions in a risk
assessment similarly requires the use of ambient air quality data at fixed-site, population-oriented
monitors. In order to estimate the incidence of a particular health effect associated with recent
conditions in a specific county or set of counties attributable to ambient Os exposures in excess
of background, as well as the change in incidence of the health effect in that county or set of
counties corresponding to a given change in O3 levels resulting from just meeting the current or
alternative 8-hr Os standards, the following three elements are required:
    •   Air quality information including: (1) recent air quality data for Os from population-
       oriented monitors in the assessment location, (2) estimates of background 63
       concentrations appropriate to this location, and (3) recent concentrations adjusted to
       reflect patterns of air quality estimated to occur when the area just meets the specified
       standards.  (These air quality inputs are discussed in more detail in section 4.5.6)
    •   Concentration-response function(s) which provide an estimate of the relationship
       between the health endpoint of interest and ambient 63 concentrations, preferably derived
       in the assessment location, as use of functions estimated in other locations increases
       uncertainty.
    •   Seasonal baseline health effects incidence rate and population.  The baseline
       incidence rate provides an estimate of the incidence rate in the assessment location
       corresponding to recent 63 levels in that location.
                                         5-29

-------
       Figure 5-5 provides a broad schematic depicting the role of these components in this part
of the risk assessment. Each of the key components (i.e., air quality information, estimated
concentration-response functions, and baseline incidence and population data) is discussed
below, highlighting those points at which judgments have been made.
       These inputs are combined to estimate health effect incidence changes associated with
specified changes in 63 levels. Although some epidemiological studies have estimated linear or
logistic concentration-response functions, by far the most common form is the exponential (or
log-linear) form:
                                   y = Beftc,            (Equation 5-4)
where x is the ambient 63 level, y is the incidence of the  health endpoint of interest at 63 level x,
P is the coefficient of ambient Os concentration (describing the extent of change my with a unit
change in x\ and B is the incidence at x=0, i.e., when there is no ambient Os. The relationship
between a specified ambient O3 level, x0, for example, and the incidence of a given health
endpoint associated with that level (denoted as yo) is then

                                   y0=Beftc°.           (Equation 5-5)
Because the log-linear form of concentration-response function (equation (5-4)) is by far the
most common form, we use this form to illustrate the derivation of the "health impact function"
used in this portion of the risk assessment.12
       If we let XQ denote the baseline (upper) Os level, and xi denote the lower Os level, and yo
and yi denote the corresponding incidences of the health  effect, we can derive the following
relationship between the change in x, Ax= (XQ- xi), and the corresponding change my, Ay, from
equation (5-4)13:
                      4v = 0>0 - J>i) = Jot1 - e'1**}.        (Equation 5-6)

       Alternatively, the difference in health effects incidence can be calculated indirectly using
relative risk. Relative risk (RR) is a measure commonly  used by epidemiologists to characterize
       12The derivations of the health impact functions from concentration-response functions for all three
functional forms found in the epidemiological literature and used in this risk assessment - the log-linear, linear, and
logistic - are given in section B.2 of Appendix B in the Risk Assessment TSD.

        If Ax < 0 - i.e., if Ax = (KI- x0) - then the relationship between Ax and Ay can be shown to be
Ay = (yl - y0)= y0[e'3'" - 1].  If Ax < 0, Ay will similarly be negative. However, the magnitude of Ay will be the
same whether Ax > 0 or Ax < 0 - i.e., the absolute value of Ay does not depend on which equation is used.
                                          5-30

-------
  Figure 5-5. Major Components of Ozone Health Risk Assessment Based on Epidemiological Studies

 Air Quality
    Ambient Monitoring for
    Selected Urban Areas
    Modeled Background
    Concentrations
    Air Quality Adjustment
    Procedures
    Current and Alternative
    Proposed Standards
Concentration-Response
  Human Epidemiological
  Studies
  Estimates of City-specific
  Baseline Health Effects
  Incidence Rates and
  Population Data
Recent ("As Is")
Ambient O3 Levels
      Changes in
      Distribution
        of O3Air
        Quality
    Concentration
    Response
    Relationships
                              Health
                               Risk
                              Model
Risk Estimates:

• Recent Air
  Quality
• Current
  Standard
• Alternative
  Standards
                                            5-31

-------
the comparative health effects associated with a particular air quality comparison.  The risk of
mortality at ambient Os level XQ relative to the risk of mortality at ambient Os level xi, for
example, may be characterized by the ratio of the two mortality rates: the mortality rate among
individuals when the ambient 63 level is x0 and the mortality rate among (otherwise identical)
individuals when the ambient Os level is xi. This is the RR for mortality associated with the
difference between the two ambient 63 levels, x0 and XL  Given a concentration-response
function of the form shown in equation (5-4) and a particular difference in ambient O3 levels, Ax,
the RR associated with that difference in ambient Os, denoted as RRAx, is equal to  e13^  . The
difference in health effects incidence, Ay, corresponding to a given difference in ambient 63
levels, Ax, can then be calculated based on this
                      AX = (y0 ~ *) = JVotl - 0/^A,)]   (Equation 5-7)

Equations (5-6) and (5-7) are simply alternative ways of expressing the relationship between a
given difference in ambient Os levels, Ax, and the corresponding difference in health effects
incidence, Ay. These health impact equations are the key equations that combine air quality
information, concentration-response function information, and baseline health effects incidence
information to estimate ambient 63 health risk.

        5.3.2.2 Air Quality Considerations
       As illustrated in Figure 5-5, and noted earlier, air quality information required to conduct
the 63 risk assessment includes: (1) recent air quality data for 63 from suitable monitors for each
selected location, (2) estimates of background concentrations for each selected location, and (3)
air quality adjustment procedures to modify the recent data to reflect changes in the distribution
of hourly 63 air quality estimated to occur when an area just meets a given 63 standard. The
approach used to adjust air quality data to simulate just meeting alternative 8-hr standards is
discussed in more detail in Chapter 4 and in Rizzo (2005, 2006).
       We retrieved Os ambient air quality data for the years 2002 through 2004 from EPA's Air
Quality System (AQS). Although the Os season varies somewhat for different regions of the
country, for much of the country the season coincides with spring and summer. To allow
comparison across locations, and because 63 effects observed in epidemiological  studies have
been more clearly and consistently shown for warm season analyses, all analyses for this portion
of the risk assessment were carried out for the same time period, April through September.
Because O3 concentrations varied substantially over the 3-year period from 2002 through 2004,
separate analyses were carried out using air quality data from 2002, in which Os concentrations
were relatively higher in most locations over this 3-year period,  and air quality data from 2004,

                                        5-32

-------
in which Os concentrations were relatively lower in most locations for this 3-year period. These
two years provide generally upper- and lower-end cases within this 3-year period.  However, two
of the 12 urban areas, Houston and Los Angeles, had similar or higher Os concentrations in 2004
compared to 2002. In addition, a more limited set of analyses, focusing only on mortality in a
subset of five urban areas (Atlanta, Chicago, Houston, Los Angeles, and New York), was carried
out since the second draft Staff Paper, using air quality data from 2003.
       To estimate the change in incidence of a health effect associated with a change in O3
concentrations from recent levels to background levels in an assessment location, two time series
of 63 concentrations are needed for that location:  (1) hourly 63 concentrations from a recent
year for the period April 1 through September 30, and (2) hourly background Os concentrations
for the same time period.  In order to be consistent with the approach generally used in the
epidemiological studies that estimated 63 concentration-response functions, the (spatial) average
ambient Os concentration on each hour for which measured data are available is deemed most
appropriate for the risk assessment.  A composite monitor data set was created for each
assessment location based on averaging each hourly value from all monitors eligible for
comparison with the current standard for each hour of the day. Table 4-7 provides a summary of
the design values for the 12 urban study areas. Appendix 5A.I to this chapter provides more
detailed information on ambient Os concentrations for these locations.
       Different exposure metrics have been used in epidemiological Os studies, including the
24-hr average and the daily  1-hr and 8-hr maximum. Therefore, daily changes at the composite
monitor in the Os exposure metric appropriate to a given concentration-response function were
calculated for use in the risk assessment (see Tables 5A-13 and 5A-14, Appendix 5A.1 for
summary statistics for the composite monitor 63  concentrations in the 12 urban locations for
2002, 2003, and 2004).  For example, if a concentration-response function related daily mortality
to daily 1-hr maximum Os concentrations, the daily changes in 1-hr maximum Os concentrations
at the composite  monitor were calculated.  In the first part of the epidemiology-based risk
assessment, in which risks associated with the recent levels of Os above background levels were
estimated, this required the following steps:
   •  Using the monitor-specific input streams  of hourly Os concentrations from a recent year,
       calculate  a stream of hourly Os concentrations at the composite monitor. The recent Os
       concentration at the composite monitor for a given hour on a given day is the average of
       the monitor-specific 63 concentrations for that hour on that day.
   •  Using this stream of hourly 63 concentrations from a recent year at the composite
       monitor, calculate the  1-hr maximum Os concentration for each day at the composite
       monitor.
   •  Using the monitor-specific input streams  of hourly background Os concentrations,
       calculate  a stream of hourly background 63 concentrations at the composite monitor.
                                       5-33

-------
    •   Using this stream of background hourly 63 concentrations at the composite monitor,
       calculate the 1-hr maximum background Os concentration for each day at the composite
       monitor.
    •   For each day, calculate Ax = (the 1-hr maximum Os concentration for that day at the
       composite monitor) - (the 1-hr maximum background 63 concentration for that day at the
       composite monitor).
       The calculations for the second part of the epidemiology-based risk assessment, in which
risks associated with estimated Os levels that just meet the current and potential alternative  8-hr
standards above background levels, were done analogously.  For this case the series of monitor-
specific adjusted hourly concentrations were used rather than the series  of monitor-specific
recent monitored hourly concentrations. Similarly, calculations for concentration-response
functions that used a different exposure metric (e.g., the 8-hr daily maximum or 24-hr average)
were done analogously, using the exposure metric appropriate to the concentration-response
function.

        5.3.2.3 Concentration-Response Functions
       As indicated in Figure 5-5, another key component in the risk model based on
epidemiological  studies is the set of concentration-response functions which provide estimates of
the relationships between each health endpoint of interest and ambient concentrations. As
discussed above, the health endpoints that have been included in the 63  risk assessment include
respiratory symptoms  in moderate-to-severe asthmatic children, respiratory-related hospital
admissions, and premature mortality associated with short-term exposures. For those health
endpoints, the assessment includes all estimates of response magnitude  from studies judged
suitable for inclusion in this assessment, including those that did not yield statistically significant
results.  Effect estimates that are not statistically significant are used from studies judged suitable
for inclusion in this assessment to avoid introducing bias into the estimate of the magnitude of
the effect. Table 5-4 summarizes the studies included in this part of the risk assessment for each
of the urban locations.
       Studies often report more than one estimated concentration-response function for the
same location and health endpoint. Sometimes models including different sets of co-pollutants
are estimated in a study; sometimes different lags are estimated. In some cases, two  or more
studies estimated a concentration-response function for Os and the same health endpoint in the
same location (this is the case, for example, with 63 and mortality associated with short-term
exposures). For some health endpoints, there are studies that estimated  multi-city and single-city
Os concentration-response functions, while other studies estimated only single-city functions.
                                        5-34

-------
       All else being equal, a concentration-response function estimated in the assessment
location is preferable to a function estimated elsewhere, since it avoids uncertainties related to
potential differences due to geographic location. That is why the urban areas selected this part of
the 63 risk assessment are, generally, those locations in which concentration-response functions
have been estimated.  There are several advantages, however, to using estimates from multi-city
studies versus studies carried out in single cities. These advantages include, but are not limited
to: (1) more precise effect estimates due to larger data sets, reducing the uncertainty around the
estimated coefficient, (2) greater consistency in data handling and model specification that can
eliminate city- to-city variation due to study design, and (3) less likelihood of publication bias or
exclusion of reporting of negative or nonsignificant findings.  Multi-city studies are applicable to
a variety of settings, since they estimate a central tendency across multiple locations. Because
single-city and multi-city studies have different advantages, where both are available for a given
location, risk estimates have been developed for both functions.
       As discussed in the CD and section 3.3.2.1 of this Staff Paper, 63 epidemiological studies
have reported relationships based on single pollutant models and/or multi-pollutant models (i.e.,
where PM, NC>2, SC>2, or CO were  entered into the health effects model along with Os. To the
extent that any of the co-pollutants present in the ambient air may have contributed to the health
effects attributed to Os in single pollutant models, risks attributed to Os might be overestimated
where concentration-response functions are based on single pollutant models. However, if co-
pollutants are highly correlated with 63, their inclusion in an 63 health effects model can lead to
misleading conclusions in identifying a specific causal  pollutant. When collinearity exists,
inclusion of multiple pollutants in models often produces unstable and statistically insignificant
effect estimates for both 63 and the co-pollutants. Given that single and multi-pollutant models
each have both potential advantages and disadvantages, with neither type clearly preferable over
the other in all cases, risk estimates based on both single- and multi-pollutant models have been
included in the risk assessment where both are available.
       Epidemiological studies have reported effect estimates associated with varying lag
periods, but for the reasons discussed in the CD and summarized in section 3.4.5 above the CD
focuses on effect estimates from models using 0- or 1-day lag periods, with some consideration
of multi-day lag effects (CD, p. 7-11). For quantitative assessments, we conclude that it is
appropriate to use results from lag  period analyses consistent with those reported in the CD,
focusing on single day lag periods  of 0-1 days for associations with mortality or respiratory
hospitalization, depending on availability of results (CD, p. 8-59).  If the effect of 63 on health
outcomes persists over several days, distributed lag model results can provide more accurate
effect estimates for quantitative assessment than an effect estimate for a single lag period (CD, p.
7-10).  Therefore, we have used distributed lag models when they are available. Where only
                                         5-35

-------
Table 5-4. Locations, Health Endpoints, and Epidemiological Studies Included in the O3
            Risk Assessment*
Urban Area
Atlanta
Boston
Chicago
Cleveland
Detroit
Houston
Los Angeles
New York
Philadelphia
Sacramento
St. Louis
Washington, D.C.
Premature Mortality
Bell et al. (2004)
Bell et al. (2004) - 95 cities
Huang etal. (2004)**
Huang et al. (2004) - 19 cities**
Bell et al. (2004) - 95 cities
Bell et al. (2004) - 95 cities
Huang etal. (2004)**
Huang et al. (2004) - 19 cities**
Schwartz (2004)
Schwartz (2004) - 14 cities
Bell et al. (2004)
Bell et al. (2004) - 95 cities
Huang etal. (2004)**
Huang et al. (2004) - 19 cities**
Bell et al. (2004)
Bell et al. (2004) - 95 cities
Huang etal. (2004)**
Huang et al. (2004) - 19 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) - 19 cities**
Schwartz (2004)
Schwartz (2004) - 14 cities
Bell et al. (2004)
Bell et al. (2004) - 95 cities
Huang etal. (2004)**
Huang et al. (2004) - 19 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
Illnesses



Schwartz etal. (1996)
Ito (2003)

Linn et al. (2000)
Thurston etal. (1992)




Respiratory Symptoms in
Asthmatic Children

Gent et al. (2003)










*Where a study indicates "14 cities," "19 cities," or "95 cities," a multi-city concentration-response function was
used in the risk assessment and the assessment location was one of the cities included in the original epidemiological
study.
**This study estimated concentration-response functions for cardiorespiratory mortality.
                                            5-36

-------
single day lags are available we have focused on single day lag periods of 0-1 days for
associations with mortality or respiratory hospitalization, depending on availability of effect
estimates (CD, p. 8-59).
       In summary:
    •   if a single-city concentration-response function was estimated in a risk assessment
       location and a multi-city function which includes that location was also available for the
       same health endpoint, both functions were included for that location in the risk
       assessment;
    •   risk estimates based on both single- and multi-pollutant models were used when both
       were available;
    •   distributed  lag models were used, when available; when a study reported several single
       lag models for a health effect, the initial selection of the appropriate lag structure for the
       health  effect was based on the overall assessment in the CD, considering all studies
       reporting concentration-response functions for that health effect.
       The locations, health endpoints, studies, and concentration-response functions included in
that portion of the risk assessment based on epidemiological studies are summarized in Tables
5B-1 through 5B-12 in Appendix 5B.1.

        5.3.2.4 Baseline Health Effects Incidence and Population Estimates
       As  illustrated in Equation 5-4, the most common health risk model based on
epidemiological studies expresses the reduction in health risk (Ay) associated with a given
reduction in O?, concentrations (Ax) as a percentage of the baseline incidence (y). To accurately
assess the impact of changes in 63 air quality on health risk in the selected urban areas,
information on the baseline incidence of health effects in each location is therefore needed. For
this assessment, baseline incidence is the incidence under recent air quality conditions.
Population sizes, for both total population and various age ranges used in the risk assessment
were obtained for the year 2000 (U.S. Census) and are summarized in Table 5-5.  Where
possible, county-specific incidence or incidence rates have been used in the assessment. County
specific mortality incidences were available for the year 2002 from CDC Wonder (CDC, 2005),
an interface for public health data dissemination provided by the Centers for Disease Control
(CDC). The baseline mortality rates for each risk assessment location are provided in Table 5-6
and are expressed as a rate per 100,000 population.
       County-specific rates for respiratory hospital discharges, and various subcategories (e.g.,
asthma, pneumonia) have been obtained, where possible, from state, local, and regional health
                                        5-37

-------
Table 5-5.  Relevant Population Sizes for O3 Risk Assessment Locations*
City
Atlanta
Boston
Boston
Chicago
Cleveland
Detroit
Houston
Los Angeles
Los Angeles
New York
New York
Philadelphia
Sacramento
St. Louis
Washington, B.C.
Counties
Fulton, DeKalb
Suffolk
Essex, Middlesex, Norfolk, Suffolk, Worcester
Cook
Cuyahoga
Wayne
Harris
Los Angeles
Los Angeles, Riverside, San Bernardino, Orange
Bronx, Kings, Queens, New York, Richmond,
Westchester
Bronx, Kings, Queens, New York, Richmond
Philadelphia
Sacramento
St. Louis City
Washington, D.C.
Population (in millions)*
Total
1.5
0.7
—
5.4
1.4
2.1
3.4
9.5
—
8.9
8.0
1.5
1.2
0.3
0.6
Ages ^30
—
—
—
—
—
—
—
—
8.4
—
—
—
—
—
—
Ages S: 65
—
—
—
—
0.2
—
—
—
—
—
—
—
—
—
—
Children, Ages < 12, with
asthma**
—

0.025
—

—
—
—
—
—
—
—
—
—
—
  Total population and age-specific population estimates taken from the 2000 U.S. Census. Populations are rounded to the nearest 0.1 million. The urban areas
given in this table are those considered in the studies used in the risk assessment, with the exception of the larger Boston area, which is the CSA for Boston (since
the study that estimated a concentration-response function in moderate and severe asthmatic children (ages 0 - 12) was conducted in Springfield, MA and CT).
** Population derived as follows:  The populations of children <5 and 5 -12 in the counties listed were multiplied by corresponding percents of children [in each
age group] in New England with "current asthma"  ~ 5.1% and 10.7% for the two age groups, respectively (see "The Burden of Asthma in New England."
Asthma Regional Council. March 2006.  Table S-2. www.asthmaregionalcouncil.org). These estimated numbers of asthmatic children were then multiplied by
the estimated percent of asthmatic children using maintenance medications (40%) (obtained via email 4-05-06 from Jeanne Moorman, CDC) and the results were
summed.
                                                             5-38

-------
    Table 5-6.  Baseline Mortality Rates (per 100,000 Population) Used in the O3 Risk Assessment*
City
Atlanta
Boston
Chicago
Cleveland
Detroit
Houston
Los Angeles
New York
Philadelphia
Sacramento
St. Louis
Washington, D.C.
National
Counties
Fulton, DeKalb
Suffolk
Cook
Cuyahoga
Wayne
Harris
Los Angeles
Bronx, Kings, Queens, New
York, Richmond, Westchester
Philadelphia
Sacramento
St. Louis City
Washington, D.C.
—
Type of Mortality
(ICD-9 Codes)
Non-accidental
(<800)
623
736
781
1,058
913
533
569
704
1,057
686
1147
942
790
Cardiorespiratory
(390-448; 490-496;
487; 480-486; 507)
131
—
189
268
234
123
155
199
242
—
—
—
196
Respiratory
(460-519)
—
—
—
—
76
—
—
—
—
—
—
—
80
*Data for the year 2002 from United States Department of Health and Human Services (US DHHS), Centers for Disease Control and Prevention (CDC),
National Center for Health Statistics (NCHS), Compressed Mortality File (CMF) compiled from CMF 1968-1988, Series 20, No. 2A 2000, CMF 1989-
1998, Series 20, No. 2E 2003 and CMF 1999-2002, Series 20, No. 2H 2004 on CDC WONDER On-line Database.  See http://wonder.cdc.gov/.
                                                 5-39

-------
departments and hospital planning commissions for each of the risk assessment locations.14
Baseline hospitalization rates used in each risk assessment location are summarized in Table 5-7
and are expressed as a rate per 100,000 relevant population.
       Baseline rates of symptoms among asthmatic children who used maintenance
medications in the Boston area were estimated by using the median rates of the respiratory
symptoms reported in Table 3 of Gent et al. (2003). Each symptom rate, the percentage of days
on which the symptom occurred, was calculated for each subject by dividing the number of days
of the symptom by the number of days of participation in the study and then multiplying by 100.
Median symptom rates among maintenance medication users for wheeze, chest tightness, and
shortness of breath were 2.8%, 1.2%, and 1.5% of days, respectively.

         5.3.2.5 Characterizing Uncertainty and Variability
       Section 5.3.1.4 previously defined what is meant by uncertainty and variability in the
context of this risk assessment. For the portion of the risk assessment based on epidemiological
studies, the statistical uncertainty surrounding the estimated 0.3 coefficients in the reported
concentration-response functions is reflected in the confidence or credible intervals provided for
the risk estimates in this chapter and in the Risk Assessment TSD.  Additional uncertainties have
been addressed quantitatively  through sensitivity analyses and/or have been discussed throughout
section 5.3.
       With respect to variability within this portion of the risk assessment, there may be
variability among concentration-response functions describing the relation between O?, and
mortality across urban areas.  This variability may be due to differences in population (e.g., age
distribution), population activities that affect exposure to Os (e.g., use of air conditioning), levels
and composition of co-pollutants, and/or other factors that vary across urban areas.
       The current risk assessment incorporates some of the variability in key inputs to the
analysis by using location-specific inputs (e.g., location-specific concentration-response
functions, baseline incidence rates,  and air quality data).  Although spatial variability in these
key inputs across all U.S. locations has not been fully characterized, variability across the
selected locations is imbedded in the analysis by using, to the extent possible,  inputs specific to
each urban area.  Temporal variability is more  difficult to address, because the risk assessment
       14The 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 carry over to the beginning of the current year
(or season).

                                        5-40

-------
Table 5-7.  Baseline Rates for Hospital Admissions Used in the Os Risk Assessment
Relevant Population
Rate per 100,000 Relevant Population*
Los
Angeles1
Ages 30+
New
York2
All Ages
Detroit3
Ages
65+
Cleveland4
Ages 65+
Admissions for:
Pulmonary illness (DRG Codes 75 - 101) -
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)
Respiratory illness ((ICD codes 460-519)
208
174
—
—
—
—
—
—
800
327
—
—
—
—
—
—
2,068
—
—
—
—
—
—
3,632
1 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).
2Rates of unscheduled hospital admissions were calculated from patient discharge data for 2001, obtained from the
New York Statewide Planning and Research Cooperative.
3Rates were calculated from hospitalization data for Wayne County for the year 2000, obtained from the Michigan
Health and Hospital Association in April 2002. EPA expressly understands that the Michigan Health and Hospital
Association has not performed an analysis of the hospitalization data obtained or warranted the accuracy of this
information and, therefore, it cannot be held responsible in any manner for the outcome.
4Based on mean daily hospital admissions for ages 65+ for ICD-9 codes 460-519 ~ Table 1 in Schwartz et al.
(1996).
                                            5-41

-------
focuses on some unspecified time in the future.  To minimize the degree to which values of
inputs to the analysis may be different from the values of those inputs at that unspecified time,
 we have used recent input data - for example, years 2002 through 2004 air quality data for all
of the urban locations, and recent mortality baseline incidence rates (from 2002).  However,
future changes in inputs have not been predicted (e.g.,  future population levels or possible
changes in baseline incidence rates).
       A number of important sources of uncertainty were addressed where possible.  Section
4.1.9 in the Risk Assessment TSD discusses in greater detail the uncertainties and variability
present in the health risk assessment. The following is a brief discussion of the major sources of
uncertainty and variability in the epidemiological portion of the risk assessment and how they are
dealt with or considered in the risk assessment:

    •   Causality.  There is uncertainty about whether  each of the estimated associations between
       63 indicators and the various health endpoints included in this risk assessment actually
       reflect a causal relationship. Our judgment, drawing on the conclusions in the CD and as
       discussed in more detail in Chapter 3 (section 3.7.5), is that for the health effects included
       in the risk assessment (i.e, increased respiratory symptoms in moderate to severe
       asthmatic children, increased respiratory-related hospital admissions, total non-accidental
       mortality, and cardiorespiratory mortality) there is, at a minimum, a likely causal
       relationship with either short-term  Os exposure itself or with Os serving as an indicator
       for itself and other components of the photochemical oxidant mix, especially during the
       warm 63 season.

    •   Empirically estimated concentration-response relationships. In estimating the
       concentration-response relationships, there are  uncertainties: (1) surrounding estimates of
       Os coefficients in concentration-response functions used in the assessment, (2)
       concerning the specification of the concentration-response model (including the shape of
       the relationship) and whether or not a population threshold or non-linear relationship
       exists within the range  of concentrations examined in the studies , (3) related to the extent
       to which concentration-response relationships derived from studies in a given location
       and time when Os levels were higher or behavior and/or housing conditions were
       different provide accurate representations of the relationships for the same locations with
       lower air quality distributions and different behavior and/or housing conditions, and (4)
       concerning the possible role of co-pollutants which also may have varied between the
       time of the studies and the current assessment period.  The approach taken to characterize
       uncertainties in the concentration-response functions arising from sample size
       considerations is discussed below.  With respect to the shape of the function and whether
       or not a population threshold may exist, as discussed in Chapter 3, the CD states that in
       those studies that provide suggestive evidence  of thresholds, the potential thresholds are
       at low  concentrations at or approaching background levels (CD, p. 7-159).  As discussed
       in Chapter 3 and in the CD (CD, p.7-175), results from recent large U.S. multi-city time-
       series studies and meta-analyses also show effect estimates that are consistent across
       studies and robust to control for potential confounders.
                                       5-42

-------
•  Adequacy of ambient (^monitors as surrogate for population exposure. The extent to
   which there are differences in the relationship between spatial variation in ambient Os
   concentrations and ambient exposures in the original epidemiology studies compared to
   more recent ambient 63 data introduces additional uncertainty in the risk estimates. We
   recognize that ambient concentrations at central monitors may not provide a good
   representation of personal exposures.  The CD identifies the following three components
   to exposure measurement error: (1) the use of average population rather than individual
   exposure data; (2) the difference between average personal ambient exposure and
   ambient concentrations at central monitoring sites; and (3) the difference between true
   and measured ambient concentrations (CD, p. 7-7). The CD notes that "these
   components are expected to have different effects, with the first and third likely not
   causing bias in a particular direction ("nondifferential error") but increasing the standard
   error, while the second component may result in downward bias, or attenuation of the
   risk estimate" (CD, pp. 7-7 to 7-8). While a concentration-response function may
   understate the effect of personal  exposures to 63 on the incidence of a health effect, it
   will give an unbiased estimate of the effect of ambient concentrations on the incidence of
   the health effect if the ambient concentrations at monitoring stations provide an unbiased
   estimate of the ambient concentrations to which the population is exposed.  A more
   comprehensive discussion of exposure measurement is given in section 3.4.2.1 of this
   Staff Paper.

•  Adjustment of air quality distributions to  simulate just meeting the current standard. The
   shape of the distribution of hourly Os concentrations that would result upon just meeting
   the current or alternative 8-hr standards is unknown. Based on an analysis of historical
   data, we believe that the Quadratic air quality adjustment procedure provides reasonable
   estimates of the shape of the distribution; however, there is greater uncertainty for those
   urban areas that have air quality well above the current standard (e.g., Los Angeles,
   Houston).  As noted previously, there is considerable year-to-year variability in Os
   concentrations over a three-year  period in many of the urban areas examined.  This leads
   to substantial year-to-year variability in risk estimates associated with 63 concentrations
   when air quality is simulated to just meet the current and potential alternative standards.
•  Estimated background concentrations for each location. The calculation of risk
   associated with recent air quality in excess of background requires as an input estimates
   of background concentrations for each location throughout the period of the assessment.
   The estimated background concentrations for each location have been estimated based on
   runs of the GEOS-CHEM global model (see section 2.7) for all hours of an "average
   day" in a given month, for each of the months from April through September.  As
   discussed in section 2.7, evaluation of the GEOS-CHEM suggests that the model is
   generally within 5 ppb in most regions of the country and that it may be 10 ppb too high
   in the southeast.  Section 5.4.3 presents results from a sensitivity analysis that
   characterizes the impact of the uncertainty about background concentrations on the non-
   accidental mortality risk estimates associated with recent air quality and just meeting the
   current and several alternative 8-hr standards.

•  Baseline incidence rates and population data.  There are uncertainties related to: (1) the
   extent to which baseline incidence rates, age distributions, and other relevant
   demographic variables that impact the risk estimates vary for the year(s) when the actual
                                   5-43

-------
       epidemiological studies were conducted, the recent year of air quality used in this
       assessment, and some unspecified future year when air quality is adjusted to simulate just
       meeting the current or alternative standards and (2) the use of annual or seasonal
       incidence rate data to develop daily health effects incidence data. Spatial variability in
       baseline incidence and population data is taken into account by use of city-specific data
       in most cases.
       One of the most critical elements in the risk assessment is the concentration-response
relationships used in the assessment. The uncertainty resulting from the statistical uncertainty
associated with the estimate of the Os coefficient in the concentration-response function was
characterized either by confidence intervals or by Bayesian credible intervals around the
corresponding point estimates of risk. Confidence and credible intervals express the range
within which the true risk is likely to fall if the only uncertainty surrounding the Os coefficient
involved sampling error. Other uncertainties, such as differences in study location, time period,
and model uncertainties are not represented by the confidence or credible intervals presented.
       Two large scale multi-city mortality studies, Bell et al. (2004) and Huang et al. (2004),
reported both multi-location and single-location concentration-response functions, using a
Bayesian two-stage hierarchical model.  In these cases, the single-location estimates can be
adjusted to make more efficient use of the  data from all locations.  The resulting "shrinkage"
estimates are so called because they "shrink" the location-specific estimates towards the overall
mean estimate (the mean of the posterior distribution of the multi-location concentration-
response function coefficient). The greater the uncertainty about the estimate of the location-
specific coefficient relative to the estimate of between-study heterogeneity, the more the
location-specific estimate is "pulled in" towards the overall mean estimate. Bell et al. (2004)
calculated these shrinkage estimates, which were presented in Figure 2 of that paper. These
location-specific shrinkage estimates, and their adjusted standard errors were provided by the
study authors and were used in the risk assessment.
       The location-specific estimates reported in Table 1  of Huang et al.  (2004) are not
"shrinkage" estimates. However, the study authors provided the posterior distribution for the
heterogeneity parameter, T, for their distributed lag model,  shown in Figure 4(b) of their paper.
Given this posterior distribution, and the original location-specific estimates presented in Table 1
of their paper, we calculated location-specific "shrinkage"  estimates using a Bayesian method
described in DuMouchel (1994) (see section 5B.3  in Appendix 5B of this Staff Paper).  As with
the shrinkage estimates presented in Bell et al. (2004), the resulting Bayesian shrinkage estimates
use the data from all of the locations considered in the study more efficiently than do the original
location-specific estimates.  The calculation of these shrinkage estimates is thus one way to
address the relatively large uncertainty surrounding estimates of coefficients in location-specific
concentration-response functions.
                                       5-44

-------
       With respect to model form, most of the epidemiological studies estimated 63 coefficients
using log-linear models.  However, there still is substantial uncertainty about the correct
functional form of the relationship between O?, and various health endpoints, especially at the low
end of the range of observed concentrations. While there are likely biological thresholds in
individuals for specific health responses, as discussed in section 3.4.6 available studies have
found little evidence for population thresholds.  For example, in a recent study, Bell et al. (2006),
applied several statistical models to data on air pollution, weather, and mortality for the 98
NMMAPS communities to evaluate whether a threshold level exists for premature mortality.
The results suggested that even low levels of tropospheric 63, well below 0.08 ppm, are
associated with premature mortality. However, as discussed in section 3.4.6 and in the CD, the
use of ambient Os concentrations may obscure the presence of thresholds in epidemiological
studies (CD p. 7-158).  In those studies that provide suggestive evidence of thresholds,  the
potential thresholds are at low concentrations at or approaching background levels (CD, p. 7-
159).
       The CD finds that no definitive conclusion can be reached with regard to the existence of
thresholds in epidemiological studies (CD, p. 8-44).  We recognize, however, the possibility that
thresholds for individuals may exist for reported associations at fairly low levels within the range
of air quality observed in the studies, but not be detectable as population thresholds in
epidemiological analyses. Based on the CD's conclusions, we judge that there is insufficient
evidence to support use of potential threshold levels in the quantitative risk assessment, but we
do recognize there is increasing uncertainty about the concentration-response relationship at
lower concentrations that is not captured by the characterization of the statistical uncertainty due
to sampling error. Therefore, as discussed later in this chapter, the risk estimates for premature
mortality, respiratory symptoms in moderate to severe asthmatic children, and respiratory-related
hospital admissions associated with exposure to 63 must be considered in the light of
uncertainties about whether or not these O3-related effects occur in the population at very low
concentrations.
       Several recent meta-analyses (Bell et al. 2005; Levy et al., 2005; and Ito et al., 2005)
have addressed the impact of various factors on estimates of mortality associated with short-term
exposures to 03. We reviewed these meta-analyses for additional information that might be used
to assist in characterizing the uncertainties associated with concentration-response functions for
this health outcome. As discussed in Chapter 3, the CD observes common  findings across all
three analyses, in that all reported that effect estimates were larger in warm season analyses,
reanalysis of results using default GAM criteria did not change the effect estimates,  and there
was no strong evidence of confounding by PM (CD, p. 7-97).  Bell et al. (2005) and Ito et al.
(2005) both provided suggestive evidence of publication bias, but O3-mortality associations
                                       5-45

-------
remained after accounting for that potential bias. The results from these meta-analyses, as well
as several single- and multiple-city studies, also indicate that copollutants generally do not
appear to substantially confound the association between Os and mortality.
       As discussed in Chapter 3, while concluding that (Vhealth associations are found to be
generally consistent, the recent Os-mortality meta-analyses indicate that some heterogeneity
exists across studies (CD, pp. 7-96 - 7-97). The CD discusses a number of factors that could
result in heterogeneity in associations between different geographic areas, focusing particularly
on variables that can affect exposure to ambient Os.  For example, the use of air conditioning can
reduce ambient exposures during the warm season, while increased outdoor activity can increase
exposure.

5.4    OZONE RISK ESTIMATES
       We present risk estimates associated with several air quality scenarios, including three
recent years of air quality as represented by 2002, 2003, and 2004 monitoring data in section
5.4.1. In section 5.4.2 we summarize risk estimates associated with air quality adjusted to
simulate just meeting the current and several potential alternative 8-hr standards.  In section 5.4.3
we present and discuss the results of sensitivity analyses examining the influence of alternative
estimates of background 63 concentrations and alternative assumptions about the shape of the
exposure-response relationship for lung function decrements in all and asthmatic school age
children. In section 5.4.4 we discuss and compare the risk estimates developed for the current
review with the risk estimates developed for the prior Os NAAQS review completed in July
1997. Finally, in section 5.4.5 we present key observations from the health risk assessment.
       Throughout this section the uncertainty surrounding risk estimates resulting from the
statistical uncertainty of the 63 coefficients in the concentration- and exposure-response
functions used is characterized by ninety-five percent confidence or credible intervals around
estimates of incidence, incidence per 100,000 population, and the percent of total incidence that
is (Vrelated. In some cases, the lower bound of a confidence or credible interval falls below
zero.  This does not imply that additional exposure to Os has a beneficial effect, but only that the
estimated 63 coefficient in the concentration- or exposure-response function was not  statistically
significantly different from zero. Lack of statistical significance could reflect insufficient
statistical power to detect a relationship that exists or that a relationship does not exist.
Conversely, the fact that a study reports statistical significant associations does  not prove
causation.  The judgments about whether a causal relationship likely exists between Os and
various health endpoints rests on a variety  of types of supporting evidence and involves a weight
of the evidence judgment, as discussed in Chapter 3.
                                       5-46

-------
         5.4.1   Recent Air Quality
       In the prior 1997 risk assessment, risks for lung function decrements associated with 1-hr
heavy exertion, 1-hr moderate exertion, and 8-hr moderate exertion exposures were estimated.
Since the 8-hr moderate exertion exposure scenario for children clearly resulted in the greatest
health risks in terms of lung function decrements, we have chosen to include only the 8-hr
moderate exertion exposures in the current risk assessment for this health endpoint. Thus, the
risk estimates presented here are most useful for making relative comparisons across alternative
air quality scenarios and do not represent the total risks for lung function decrements in children
or other groups within the general population associated  with any of the air quality scenarios.
Thus, some outdoor workers and adults engaged in moderate exertion over multi-hour periods
(e.g., 6-8 hr exposures) also would be expected to experience similar lung function decrements.
However, the percentage of each of these other subpopulations expected to experience these
effects is expected to be smaller than all school age children who tend to spend more hours
outdoors while active based on the exposure analyses conducted during the prior review.
       We have included risk estimates for all and asthmatic school age children in this section.
As noted previously,  risk estimates for asthmatic school  age children have been added to the
assessment since the  second draft Staff Paper.  Risk estimates associated with recent air quality
(2002 and 2004) for up to 12 urban locations are presented in this section.  Additional risk
estimates developed since the second draft Staff Paper associated with 2003 air quality for a
subset of five locations (Atlanta, Chicago, Houston, Los  Angeles, and New York) are presented
in the Risk Assessment TSD and for all 12 urban areas in a recent memo (Post, 2007).
       Tables 5-8 and 5-9 display the risk estimates for all school age children (ages 5-18)
associated with 2004 and 2002 Os concentrations for > 15% lung function decrement responses
for the 12 urban areas.  Tables 5-8  and 5-9 also include risk estimates associated with air quality
adjusted to simulate just meeting the current 0.08 ppm, 8-hr standard, which will be discussed in
the next section.  Consistent with CASAC's advice contained in its October 24, 2006 letter
(Hendeson, 2006), we have focused on FEVi decrements > 15% for all school age children since
this level of response is judged to be an indicator of adverse health effects for healthy children.
Similar estimates for > 10 and X20%  decrement in lung  function for all school age children can
be found in Chapter 3 of the Risk Assessment TSD.  All  estimates in both tables reflect
responses associated  with exposure to 63 in excess of exposures associated with background 63
concentrations.  Table 5-8 shows the number and percent of all school age children estimated to
have at least 1 lung function response (defined as FEVi > 15%) during the Os season.  Table 5-9
displays the total number of occurrences for this same lung function response during the 63
season.  As illustrated by the estimates shown in these two tables, a child may experience
multiple occurrences of a lung function response during the 63 season. For example,  in Atlanta

                                      5-47

-------
the median estimate is that 30,000 school age children experienced an FEVi decrement > 15%
during the Os season with a median estimate of about 170,000 occurrences of this same response
in this population for 2004 air quality data.  Thus, for this example on average each child is
estimated to have over 5 occurrences of this lung function response during the 63 season.
       As shown in Table 5-8, across the 12 urban areas, the ranges in median estimates of the
percent of all  school age children estimated to experience at least one FEVi decrement > 15%
during the O3 season are 1.3 to 5.9% for 2004 and 4.8 to 9.1% for 2002. In terms of total
occurrences of FEVi decrement > 15% during the Os season, Table 5-9 shows a range of median
estimates from 64,000 to nearly 1.5 million responses in 2004 and from about 130,000 to about
1.4 million responses in 2002 for all school age children across the 12 urban areas associated
with Os concentrations. Both Tables 5-8 and 5-9 also include 95% credible intervals for the lung
function decrement risk estimates based on sample size considerations.  These credible  intervals
only represent part of the uncertainty associated with these risk estimates. Additional
uncertainties are summarized in section 5.3.2.5 and should be kept in mind as one considers the
risk estimates in these tables.
       Comparable tables to those discussed above for lung function responses in all school age
children are presented in Tables 5-10 and 5-11 for asthmatic school age children for 5 urban
areas that are  a subset of the 12 urban areas included for all children.  Again, the risk estimates
associated with just meeting the current 8-hr standard presented in these tables will be discussed
in the next section. For asthmatic children a lung function response defined in terms of FEVi
decrement > 10% is shown, consistent with CASAC's advice (Henderson, 2006c) that this level
of response serves as an indicator of adverse health effects for this population. As shown in
Table 5-10, across the 5 urban areas, the ranges in median estimates of the number of asthmatic
school age children estimated to experience at least one FEVi decrement > 10% during  the Os
season are 10,000 to 61,000 for 2004 air quality and  16,000 to about 110,000 for 2002 air
quality.  These median ranges represent 4.6 to 13.4% of asthmatic school age children for 2004
air quality and 11.5 to 17% of asthmatic school age children for 2002 air quality.  In terms of
total occurrences of FEVi decrement > 10% associated with 63 concentrations during the 63
season, Table 5-11 shows a range of median estimates from 98,000 to about 670,000 responses
in 2004 and from about 89,000 to 760,000 responses in 2002 for asthmatic school age children
across the 5 urban areas. Dividing the estimated total number of occurrences by the number of
asthmatic children estimated to experience this lung function response, results in each child
being estimated to have on average between 5 and 10 occurrences of this lung function  response
during the O3 season depending on the urban area and year of air quality analyzed.
                                      5-48

-------
 Table 5-8. Number and Percent of All School Age Children Estimated to Experience Lung Function Responses (FEVi > 15%)
              Associated with 8-Hour Os Exposure While Engaged in Moderate Exertion for Location-Specific Os Seasons*
Location
(Os Season)
Atlanta
(March-October)
Boston
(April-September)
Chicago
(April-September)
Cleveland
(April-October)
Detroit
(April-September)
Houston
(All year)
Los Angeles
(All year)
New York
(April-September)
Philadelphia
(April-October)
Sacramento
(All year)
St. Louis
(April-October)
Washington, DC
(April-October)
Number of Children (Ages 5-18) Having at Least 1 Lung Function Response (FEVi > 15%) Associated with 8-Hour Os Exposure**
Recent Air Quality (2004)
Number
(1000s)
30
(16-46)
22
(9 - 37)
25
(5 - 46)
11
(4-19)
19
(7 - 33)
51
(33 - 73)
215
(145-292)
97
(43-156)
34
(17-52)
11
(6-16)
10
(3-17)
48
(26 - 73)
Percent
3.2%
(1 .7% - 4.8%)
2%
(0.8% - 3.4%)
1 .3%
(0.2% - 2.4%)
1 .9%
(0.7% - 3.2%)
1 .7%
(0.6% - 3%)
4.7%
(3% - 6.7%)
5.9%
(4% - 8%)
2.3%
(1%-3.8%)
2.8%
(1 .40/0 . 4.40/0)
2.6%
(1 .3% - 4%)
1 .7%
(0.5% - 2.9%)
3.2%
(1 .7% - 4.9%)
Just Meeting Current 8-Hour
Standard (based on adjusting
2004 air quality)
Number
(1000s)
18
(6-31)
13
(3 - 25)
14
(0 - 30)
6
(1 - 12)
12
(2 - 23)
21
(8 - 35)
33
(5-61)
39
(4 - 77)
17
(4 - 30)
4
(1-7)
7
(1-13)
24
(8 - 42)
Percent
1 .9%
(0.7% - 3.3%)
1 .2%
(0.3% - 2.3%)
0.7%
(0% - 1 .5%)
1%
(0.1% -1.9%)
1.1%
(0.2% -2.1%)
1 .9%
(0.8% - 3.2%)
0.9%
(0.1% -1.7%)
0.9%
(0.1%- 1.9%)
1 .4%
(0.3% - 2.5%)
1%
(0.1% -1.8%)
1 .2%
(0.2% - 2.3%)
1 .6%
(0.5% - 2.8%)
Recent Air Quality (2002)
Number
(1000s)
50
(33-71)
76
(52-103)
115
(76-159)
50
(35 - 67)
71
(48 - 97)
52
(33 - 73)
220
(149-297)
316
(220 - 426)
108
(77- 143)
23
(15-31)
41
(28 - 55)
117
(82- 157)
Percent
5.3%
(3.5% - 7.5%)
6.9%
(4.7% - 9.4%)
5.9%
(3.9% - 8.2%)
8.5%
(6% - 1 1 .3%)
6.4%
(4.3% - 8.8%)
4.8%
(3% - 6.7%)
6%
(4.1% -8.1%)
7.6%
(5. 3% -10.3%)
9.1%
(6.5%- 12.1%)
5.6%
(3.7% - 7.6%)
7%
(4.8% - 9.5%)
7.9%
(5.5%- 10.6%)
Just Meeting Current 8-Hour
Standard (based on adjusting
2002 air quality)
Number
(1000s)
31
(17-47)
47
(29 - 68)
67
(37- 100)
27
(17-39)
45
(27 - 66)
22
(9 - 36)
35
(8 - 63)
131
(70 - 200)
58
(37 - 83)
9
(4-14)
30
(20 - 43)
64
(39 - 93)
Percent
3.3%
(1 .8% - 5%)
4.3%
(2.6% - 6.2%)
3.4%
(1.9% -5. 2%)
4.6%
(2.9% - 6.6%)
4%
(2.4% - 5.9%)
2%
(0.8% - 3.3%)
1%
(0.2% - 1 .7%)
3.1%
(1 .7% - 4.8%)
4.9%
(3.1% -7%)
2.2%
(1%-3.5%)
5.2%
(3.4% - 7.4%)
4.3%
(2.6% - 6.3%)
*Risks are estimated for exposures in excess of policy relevant background.
**Numbers are median (0.5 fractile) numbers of children.  Numbers in parentheses below the median are 95% credible intervals based on statistical uncertainty surrounding
the Os coefficient. Numbers are rounded to the nearest 1000. Percents are rounded to the nearest tenth.
                                                          5-49

-------
Table 5-9.  Number of Occurrences of Lung Function Responses (FEVi > 15%) Among All School Age Children Associated with
           8-Hour Os Exposure While Engaged in Moderate Exertion for Location-Specific Os Seasons*
Location
Atlanta
(March-October)
Boston
(April-September)
Chicago
(April-September)
Cleveland
(April-October)
Detroit
(April-September)
Houston
(All year)
Los Angeles
(All year)
New York
(April-September)
Philadelphia
(April-October)
Sacramento
(All year)
St. Louis
(April-October)
Washington, DC
(April-October)
Occurrences of Lung Function Response (FENAi > 15%) Associated with 8-Hour O3 Exposure Among
Children (Ages 5-18), in Thousands**
Recent Air Quality
(2004)
166
(22-410)
112
(1 1 - 295)
162
(5 - 455)
68
(6-179)
108
(8 - 293)
208
(53 - 436)
1465
(375-3109)
507
(56-1299)
192
(26 - 468)
80
(9-198)
64
(3-169)
225
(36 - 549)
Just Meeting Current 8-
Hour Standard (based on
adjusting 2004 air quality)
115
(8-312)
78
(3 - 225)
108
(0 - 327)
43
(1 - 126)
77
(2 - 225)
102
(10-242)
379
(6-1080)
274
(5-813)
116
(4-321)
40
(1 -113)
50
(1 -139)
140
(9 - 385)
Recent Air Quality
(2002)
233
(62-501)
272
(93 - 555)
462
(144-946)
221
(85 - 422)
300
(100-603)
187
(51 - 385)
1334
(368 - 2805)
1352
(484 - 2647)
502
(199-940)
132
(33 - 285)
172
(59 - 340)
500
(176-987)
Just Meeting Current 8-
Hour Standard (based on
adjusting 2002 air quality)
159
(27 - 376)
186
(45-419)
298
(58 - 684)
132
(33 - 290)
205
(47 - 457)
90
(1 1 - 208)
342
(10-952)
679
(112-1607)
296
(75 - 636)
67
(6-171)
133
(37 - 280)
307
(68 - 690)
    *Risks are estimated for exposures in excess of policy relevant background.
    **Numbers are median (0.5 fractile) numbers of occurrences in thousands. Numbers
    based on statistical uncertainty surrounding the O3 coefficient. Numbers are rounded
in parentheses below the median are 95% credible intervals
to the nearest 1000.
                                                     5-50

-------
       The risk estimates associated with 2004 and 2002 Os concentrations for morbidity health
endpoints based on epidemiological studies are shown in Tables 5-12 and 5-13 for respiratory
symptoms in moderate to severe asthmatic children for the Boston urban area and in Tables 5-14
and 5-15 for excess hospital admissions for total respiratory illness and asthma (which is a subset
of total respiratory illness admissions) for the New York City urban area.  Additional hospital
admission estimates for three other locations are provided in the Risk Assessment TSD. All
results for morbidity health endpoints based on epidemiological studies are for health risks
associated with short-term exposures to Os concentrations in excess of background levels from
April through September for 2004 and 2002, respectively.
       As discussed previously, risk estimates were developed for several respiratory symptoms
in asthmatic children ages 0 to 12 who use maintenance medications based on the concentration-
response functions provided in Gent et al. (2003). These estimates were developed only for the
Boston urban area which was near the location of the original epidemiological study.  Tables 5-
12 and 5-13  show risk estimates for three different respiratory symptoms (i.e., chest tightness,
shortness of breath, and wheeze) for the Boston area associated with O3 levels above background
for April through September of 2004 and 2002, respectively.  The risk estimates are expressed in
terms of cases, cases per 100,000 relevant population, and percent of total incidence
       Tables 5-14 and 5-15 show risk estimates of unscheduled hospital admissions for
respiratory illness in the New York City area associated with Os levels above background for
April through September of 2004 and 2002, respectively. The risk estimates are expressed in
terms of cases, cases per 100,000 relevant population, and percent of total incidence.
       Tables 5-16 and 5-17 show risk estimates for non-accidental mortality associated with Os
levels above background for April through September of 2004 and 2002, respectively.  Similar
tables for cardiorespiratory mortality are included in the Risk Assessment TSD.  The risk
estimates are presented in terms of estimated incidence, incidence per 100,000 relevant
population, and percent of total incidence.
       As discussed in section 5.3.2.5,, Bell et al. (2004) reported both multi-location and
single-location concentration-response functions in a variety of locations, using a Bayesian two-
stage hierarchical model. Thus, where available, risk estimates are included in Tables 5-16 and
5-17 based on both single-city  and multi-city functions. The ranges shown in these tables are
based either on the 95 percent confidence intervals around those estimates (if the coefficients
were estimated using classical  statistical techniques) or on the 95 percent credible intervals (if
the coefficients were estimated using Bayesian statistical techniques).
                                       5-51

-------
   Table 5-10. Number and Percent of Asthmatic School Age Children Estimated to Experience Lung Function Responses (FEVi
              > 10%) Associated with 8-Hour Os Exposure While Engaged in Moderate Exertion for Location Specific Os
              Seasons*
Location (O3
Season)
Atlanta
(March-
October
Chicago
(April-
September)
Houston
(All year)
Los Angeles
(All year)
New York
(April-
September)
Asthmatic Children (Ages 5-18) Having at Least 1 Lung Function Response (FEVi >10%) Associated with 8-Hour Os Exposure Under
Moderate Exertion**
Recent Air Quality (2004)
Number
(1000s)
10
(8-15)
13
(8-21)
16
(12-21)
61
(51 - 80)
47
(33 - 70)
Percent
8.9%
(6.5% -13.1%)
4.6%
(2.9% - 7.3%)
1 1 .5%
(9%- 15.7%)
13.4%
(11. 2% -17. 4%)
7.3%
(5.1% -10.9%)
Just Meeting Current 8-Hour
Standard (based on adjusting
2004 air quality)
Number
(1000s)
7
(5-11)
8
(5-13)
8
(6-13)
16
(1 1 - 25)
24
(14-39)
Percent
6.2%
(4.2% - 9.8%)
3%
(1 .7% - 4.8%)
6.1%
(4.3% - 9.4%)
3.4%
(2.4% - 5.4%)
3.7%
(2.2% - 6%)
Recent Air Quality (2002)
Number
(1000s)
16
(12-21)
38
(31 -51)
16
(12-21)
60
(50 - 78)
109
(89-138)
Percent
13.4%
(10.7%- 17.9%)
13.8%
(11%- 18.1%)
1 1 .5%
(9% -15. 5%)
13.2%
(11% -17.1%)
17%
(13.8% -21 .4%)
Just Meeting Current 8-Hour
Standard (based on adjusting
2002 air quality)
Number
(1000s)
11
(8-16)
26
(20 - 38)
8
(6-13)
16
(1 1 - 24)
58
(43 - 85)
Percent
9.6%
(7.2% -13.9%)
9.4%
(7%- 13.5%)
6.2%
(4.4% - 9.5%)
3.4%
(2.5% - 5.3%)
9.1%
(6.7% -13.3%)
*Risks are estimated for exposures in excess of policy relevant background.
**Numbers are median (0.5 fractile) numbers of children.  Numbers in parentheses below the median are 95% credible intervals based on statistical uncertainty
surrounding the Os coefficient. Numbers are rounded to the nearest 1000. Percents are rounded to the nearest tenth.
                                                          5-52

-------
Table 5-11. Number of Occurrences of Lung Function Responses (FEVi > 10%) Among Asthmatic School Age Children
           Associated with 8-Hour Os Exposure While Engaged in Moderate Exertion for Location Specific Os Seasons*
Location
Atlanta
(March-October)
Chicago
(April-October)
Houston
(All year)
Los Angeles
(All year)
New York
(April-September)
Occurrences of Lung Function Response (FEVi >10%) Associated with 8-Hour Os Exposure Among Asthmatic Children
(Ages 5-18) While Engaged in Moderate Exertion, in Thousands**
Recent Air Quality
(2004)
98
(31 -178)
113
(25-214)
103
(45-173)
671
(305-1133)
374
(110-685)
Just Meeting Current 8-Hour
Standard (based on adjusting
2004 air quality)
73
(18-138)
79
(11 -155)
58
(20-100)
226
(48 - 423)
228
(38 - 442)
Recent Air Quality
(2002)
121
(52 - 209)
241
(112-406)
89
(40- 148)
602
(271 -1015)
762
(377- 1260)
Just Meeting Current 8-Hour
Standard (based on adjusting
2002 air quality)
91
(33-163)
175
(67 - 308)
49
(17-84)
201
(45 - 372)
469
(170-837)
 *Risks are estimated for exposures in excess of policy relevant background.
 **Numbers are median (0.5 fractile) numbers of occurrences.  Numbers in parentheses below the median are 95% credible intervals based on statistical
 uncertainty surrounding the O3 coefficient. Numbers are rounded to the nearest 1000.
                                                       5-53

-------
     Table 5-12.  Estimated Respiratory Symptoms Associated with Recent  (April - September, 2004) O3 Concentrations Above
                Background in Boston, MA
Health Effects*
Respiratory symptoms among asthmatic
medication-users — chest tightness
Respiratory symptoms among asthmatic
medication-users — chest tightness
Respiratory symptoms among asthmatic
medication-users - chest tightness
Respiratory symptoms among asthmatic
medication-users - chest tightness
Respiratory symptoms among asthmatic
medication-users - shortness of breath
Respiratory symptoms among asthmatic
medication-users - shortness of breath
Respiratory symptoms among asthmatic
medication-users — wheeze
Study
Gent et al. (2003)
Gent et al. (2003)
Gent et al. (2003)
Gent et al. (2003)
Gent et al. (2003)
Gent et al. (2003)
Gent et al. (2003)
Ages
0-12
0-12
0-12
0-12
0-12
0-12
0-12
Lag
1-day lag
0-day lag
1-day lag
1-day lag
1-day lag
1-day lag
0-day lag
Exposure
Metric
1 hr max.
1 hr max.
1 hr max.
8 hr max.
1 hr max.
8 hr max.
1 hr max.
Other
Pollutants
in Model
none
PM2.5
PM2.5
none
none
none
PM2.5
Health Effects Associated with O3 Above Policy
Relevant Background Levels**
Incidence
5300
(800 - 9200)
8400
(3800-12400)
7700
(3000-11800)
5400
(1700-8700)
5700
(700-10200)
6300
(1200-10800)
15400
(5500 - 24200)
Incidence per
100,000 Relevant
Population
20700
(3300 - 36300)
33100
(14900-49100)
30400
(11800-46800)
21400
(6900 - 34500)
22500
(2700 - 40200)
24700
(4800 - 42500)
60800
(21800-95600)
Percent of Total
Incidence
9.4%
(1.5% -16.5%)
15.1%
(6.8% - 22.3%)
13.8%
(5.4% -21. 3%)
9.7%
(3.1% -15.7%)
8.2%
(1%-14.7%)
9%
(1.8% -15.5%)
11.9%
(4.3% -18.7%)
*Health effects are associated with short-term exposures to O3.
**lncidence was quantified down to estimated policy relevant background levels. Incidences of respiratory symptom-days and respiratory symptom-days per 100,000 relevant
population are rounded to the nearest 100.  Percents are rounded to the nearest tenth.
Note:  Numbers in parentheses are 95% confidence or credible intervals based on statistical uncertainty surrounding the O3 coefficient.
                                                                       5-54

-------
     Table 5-13.  Estimated Respiratory Symptoms Associated with Recent (April - September, 2002) O3 Concentrations Above
                Background in Boston, MA
Health Effects*
Respiratory symptoms among asthmatic
medication-users — chest tightness
Respiratory symptoms among asthmatic
medication-users — chest tightness
Respiratory symptoms among asthmatic
medication-users - chest tightness
Respiratory symptoms among asthmatic
medication-users - chest tightness
Respiratory symptoms among asthmatic
medication-users - shortness of breath
Respiratory symptoms among asthmatic
medication-users - shortness of breath
Respiratory symptoms among asthmatic
medication-users — wheeze
Study
Gent et al. (2003)
Gent et al. (2003)
Gent et al. (2003)
Gent et al. (2003)
Gent et al. (2003)
Gent et al. (2003)
Gent et al. (2003)
Ages
0-12
0-12
0-12
0-12
0-12
0-12
0-12
Lag
1-day lag
0-day lag
1-day lag
1-day lag
1-day lag
1-day lag
0-day lag
Exposure
Metric
1 hr max.
1 hr max.
1 hr max.
8 hr max.
1 hr max.
8 hr max.
1 hr max.
Other
Pollutants
in Model
none
PM2.5
PM2.5
none
none
none
PM2.5
Health Effects Associated with O3 Above Policy
Relevant Background Levels**
Incidence
6900
(1100-11800)
10800
(5000-15700)
10000
(4000-15000)
7200
(2400-11400)
7500
(900-13200)
8300
(1700-14000)
20100
(7400-31000)
Incidence per
100,000 Relevant
Population
27200
(4500 - 46600)
42700
(19700-62100)
39400
(15700-59400)
28400
(9300 - 44900)
29500
(3700 - 52000)
32800
(6600 - 55300)
79200
(29000-122300)
Percent of Total
Incidence
12.4%
(2% - 21 .2%)
19.5%
(9% -28. 3%)
17.9%
(7.1% -27%)
12.9%
(4.2% - 20.5%)
10.8%
(1.3% -19%)
11.9%
(2.4% - 20.2%)
15.5%
(5.7% - 23.9%)
*Health effects are associated with short-term exposures to O3.
**lncidence was quantified down to estimated policy relevant background levels. Incidences of respiratory symptom-days and respiratory symptom-days per 100,000 relevant
population are rounded to the nearest 100. Percents are rounded to the nearest tenth.
Note:  Numbers in parentheses are 95% credible intervals based on statistical uncertainty surrounding the O^ coefficient.
                                                                       5-55

-------
 Table 5-14. Estimated Hospital Admissions Associated with Recent (April - September, 2004) Os Concentrations in NY, NY"
Health Effects*
Hospital admissions
(unscheduled),
respiratory illness
Hospital admissions
(unscheduled), asthma
Study
Thurston et al. (1992)***
Thurston et al. (1992)***
Ages
all
all
Lag
3-day lag
1-day lag
Exposure
Metric
1 hr max.
1 hr max.
Other
Pollutants in
Model
none
none
Health Effects Associated with O3 Above Policy
Relevant Background Levels*
Incidence
447
(108-786)
382
(81 -683)
Incidence per
100,000
Relevant
Population
5.6
(1.4-9.8)
4.8
(1-8.5)
Percent of Total
Incidence
1.3%
(0.3% - 2.2%)
2.9%
(0.6% - 5.2%)
incidences are rounded to the nearest whole number; incidences per 100,000 relevant population and percents are rounded to the nearest tenth.
**New York in this study is defined as the five boroughs of New York City.
Note:  Numbers in parentheses are 95% confidence or credible intervals based on statistical uncertainty surrounding the O^ coefficient.
 Table 5-15. Estimated Hospital Admissions Associated with Recent (April - September, 2002) Os Concentrations in NY, NY"
Health Effects
Hospital admissions
[unscheduled),
respiratory illness
Hospital admissions
[unscheduled),
asthma
Study
Thurston et al. (1992)***
Thurston et al. (1992)***
Ages
all
all
Lag
3-day lag
1-day lag
Exposure
Metric
1 hr max.
1 hr max.
Other
Pollutants in
Model
none
none
Health Effects Associated with Os Above Policy
Relevant Background Levels*
Incidence
608
(147-1068)
519
(110-928)
Incidence per
100,000
Relevant
Population
7.6
(1.8-13.3)
6.5
(1.4-11.6)
Percent of Total
Incidence
1.7%
(0.4% - 3%)
4%
(0.8% -7.1%)
incidences are rounded to the nearest whole number; incidences per 100,000 relevant population and percents are rounded to the nearest tenth.
**New York in this study is defined as the five boroughs of New York City.
Note:  Numbers in parentheses are 95% confidence or credible intervals based on statistical uncertainty surrounding the Os coefficient.
                                                            5-56

-------
Table 5-16. Estimated Non-Accidental Mortality Associated with Recent (April - September, 2004) O3 Concentrations*
Location
Atlanta
Boston
Chicago
Cleveland
Detroit
Houston

Study
Bell et al. (2004)
Bell et al. - 95 US Cities (2004)
Bell et al. - 95 US Cities (2004)
Bell et al. - 95 US Cities (2004)
Schwartz (2004)
Schwartz - 14 US Cities (2004)
Bell et al. (2004)
Bell et al. - 95 US Cities (2004)
Bell et al. (2004)
Bell et al. - 95 US Cities (2004)
Schwartz (2004)
Schwartz - 14 US Cities (2004)
Ito (2003)
Bell et al. (2004)
Lag
distributed lag
distributed lag
distributed lag
distributed lag
0-day lag
0-day lag
distributed lag
distributed lag
distributed lag
distributed lag
0-day lag
0-day lag
0-day lag
distributed lag
5-57
Exposure Metric
24 hr avg.
24 hr avg.
24 hr avg.
24 hr avg.
1 hr max.
1 hr max.
24 hr avg.
24 hr avg.
24 hr avg.
24 hr avg.
1 hr max.
1 hr max.
24 hr avg.
24 hr avg.
Non-Accidental Mortality Associated with Os Above Policy Relevant
Background Levels**
Incidence
6
(-26 - 38)
12
(4-20)
7
(2-12)
49
(16-81)
394
(125-658)
148
(46 - 250)
27
(-17-69)
17
(6 - 28)
33
(-11 -76)
17
(6 - 28)
128
(-21 - 274)
70
(22-117)
40
(-37-116)
35

Incidence per 100,000
Relevant Population
0.4
(-1.8-2.6)
0.8
(0.3-1.4)
1.0
(0.3-1.7)
0.9
(0.3-1.5)
7.3
(2.3-12.2)
2.8
(0.9-4.6)
1.9
(-1.2-5)
1.2
(0.4-2)
1.6
(-0.5-3.7)
0.8
(0.3-1.4)
6.2
(-1 -13.3)
3.4
(1.1-5.7)
2.0
(-1.8-5.6)
1.0
Percent of Total
Incidence
0.1%
(-0.6% - 0.8%)
0.3%
(0.1% -0.4%)
0.3%
(0.1% -0.5%)
0.2%
(0.1% -0.4%)
1.9%
(0.6% -3.1%)
0.7%
(0.2% - 1 .2%)
0.4%
(-0.2% - 0.9%)
0.2%
(0.1% -0.4%)
0.4%
(-0.1% -0.8%)
0.2%
(0.1% -0.3%)
1.4%
(-0.2% -2.9%)
0.7%
(0.2% - 1 .2%)
0.4%
(-0.4% -1.2%)
0.4%


-------
Location

Los Angeles
New York
Philadelphia
Sacramento
St Louis
Washington, D.C.
Study

Bell et al. - 95 US Cities (2004)
Schwartz (2004)
Schwartz - 14 US Cities (2004)
Bell et al. (2004)
Bell et al. - 95 US Cities (2004)
Bell et al. - 95 US Cities (2004)
Bell et al. - 95 US Cities (2004)
Moolgavkar et al. (1995)
Bell et al. (2004)
Bell et al. - 95 US Cities (2004)
Bell et al. (2004)
Bell et al. - 95 US Cities (2004)
Bell et al. - 95 US Cities (2004)
Lag

distributed lag
0-day lag
0-day lag
distributed lag
distributed lag
distributed lag
distributed lag
1-day lag
distributed lag
distributed lag
distributed lag
distributed lag
distributed lag
Exposure Metric

24 hr avg.
1 hr max.
1 hr max.
24 hr avg.
24 hr avg.
24 hr avg.
24 hr avg.
24 hr avg.
24 hr avg.
24 hr avg.
24 hr avg.
24 hr avg.
24 hr avg.
Non-Accidental Mortality Associated with O3 Above Policy Relevant
Background Levels**
Incidence
(2-67)
17
(6 - 28)
93
(9-176)
78
(24-130)
62
(-149-271)
133
(45-221)
60
(20-100)
23
(8 - 38)
82
(52-112)
12
(-36 - 59)
18
(6 - 29)
3
(-6-13)
3
(1-5)
8
(3-14)
Incidence per 100,000
Relevant Population
(0.1-2)
0.5
(0.2-0.8)
2.7
(0.3-5.2)
2.3
(0.7-3.8)
0.6
(-1.6-2.8)
1.4
(0.5-2.3)
0.7
(0.2-1.1)
1.5
(0.5-2.5)
5.4
(3.4-7.4)
1.0
(-3-4.8)
1.4
(0.5-2.4)
1.0
(-1.7-3.6)
0.9
(0.3-1.5)
1.5
(0.5-2.4)
Percent of Total
Incidence
(0% - 0.7%)
0.2%
(0.1% -0.3%)
1%
(0.1% -1.9%)
0.9%
(0.3% -1.4%)
0.2%
(-0.5% -1%)
0.5%
(0.2% - 0.8%)
0.2%
(0.1% -0.3%)
0.3%
(0.1% -0.5%)
1%
(0.6% - 1 .4%)
0.3%
(-0.9% -1.4%)
0.4%
(0.1% -0.7%)
0.2%
(-0.3% -0.6%)
0.2%
(0.1% -0.3%)
0.3%
(0.1% -0.5%)
*AII results are for mortality (among all ages) associated with short-term exposures to O^. All results are based on single-pollutant models.
**lncidence was quantified down to estimated policy relevant background levels. Incidences are rounded to the nearest whole number; incidences per 100,000 relevant
population and percents are rounded to the nearest tenth.
Note: Numbers in parentheses are 95% confidence or credible intervals based on statistical uncertainty surrounding the O^ coefficient.
                                                                   5-58

-------
Table 5-17. Estimated Non-Accidental Mortality Associated with Recent (April - September, 2002) Os Concentrations*
Location
Atlanta
Boston
Chicago
Cleveland
Detroit
Study
Bell et al. (2004)
Bell et al. - 95 US Cities (2004)
Bell et al. - 95 US Cities (2004)
Bell et al. - 95 US Cities (2004)
Schwartz (2004)
Schwartz - 14 US Cities (2004)
Bell et al. (2004)
Bell et al. - 95 US Cities (2004)
Bell et al. (2004)
Bell et al. - 95 US Cities (2004)
Schwartz (2004)
Schwartz - 14 US Cities (2004)
Ito (2003)
Lag
distributed lag
distributed lag
distributed lag
distributed lag
0-day lag
0-day lag
distributed lag
distributed lag
distributed lag
distributed lag
0-day lag
0-day lag
0-day lag
Exposure Metric
24 hr avg.
24 hr avg.
24 hr avg.
24 hr avg.
1 hr max.
1 hr max.
24 hr avg.
24 hr avg.
24 hr avg.
24 hr avg.
1 hr max.
1 hr max.
24 hr avg.
Non-Accidental Mortality Associated with O3 Above Policy Relevant
Background Levels**
Incidence
9
(-37 - 54)
17
(6 - 29)
10
(3-17)
69
(23-115)
505
(161 -840)
191
(60-321)
61
(-38-157)
38
(13-64)
57
(-18-131)
29
(10-48)
181
(-30 - 385)
99
(31 -165)
69
(-64-198)
Incidence per 100,000
Relevant Population
0.6
(-2.5-3.6)
1.2
(0.4-1.9)
1.5
(0.5-2.5)
1.3
(0.4-2.1)
9.4
(3-15.6)
3.6
(1.1-6)
4.3
(-2.7-11.3)
2.8
(0.9-4.6)
2.8
(-0.9-6.3)
1.4
(0.5-2.3)
8.8
(-1.4-18.7)
4.8
(1.5-8)
3.4
(-3.1 -9.6)
Percent of Total
Incidence
0.2%
(-0.8% -1.2%)
0.4%
(0.1% -0.6%)
0.4%
(0.1% -0.7%)
0.3%
(0.1% -0.5%)
2.4%
(0.8% - 4%)
0.9%
(0.3% -1.5%)
0.8%
(-0.5% -2.1%)
0.5%
(0.2% - 0.9%)
0.6%
(-0.2% -1.4%)
0.3%
(0.1% -0.5%)
1.9%
(-0.3% -4.1%)
1%
(0.3% -1.8%)
0.7%
(-0.7% -2.1%)
                                                  5-59

-------
Location
Houston
Los Angeles
New York
Philadelphia
Sacramento
St Louis
Washington, D.C.
Study
Bell et al. (2004)
Bell et al. - 95 US Cities (2004)
Schwartz (2004)
Schwartz - 14 US Cities (2004)
Bell et al. (2004)
Bell et al. - 95 US Cities (2004)
Bell et al. - 95 US Cities (2004)
Bell et al. - 95 US Cities (2004)
Moolgavkar et al. (1995)
Bell et al. (2004)
Bell et al. - 95 US Cities (2004)
Bell et al. (2004)
Bell et al. - 95 US Cities (2004)
Bell et al. - 95 US Cities (2004)
Lag
distributed lag
distributed lag
0-day lag
0-day lag
distributed lag
distributed lag
distributed lag
distributed lag
1-day lag
distributed lag
distributed lag
distributed lag
distributed lag
distributed lag
Exposure Metric
24 hr avg.
24 hr avg.
1 hr max.
1 hr max.
24 hr avg.
24 hr avg.
24 hr avg.
24 hr avg.
24 hr avg.
24 hr avg.
24 hr avg.
24 hr avg.
24 hr avg.
24 hr avg.
Non-Accidental Mortality Associated with O3 Above Policy Relevant
Background Levels**
Incidence
29
(2 - 57)
14
(5 - 24)
85
(8-161)
71
(22-119)
51
(-124-224)
110
(37-184)
105
(35-174)
37
(12-62)
132
(83-180)
16
(-48 - 78)
23
(8 - 39)
6
(-11 -23)
6
(2-10)
15
(5 - 25)
Incidence per 100,000
Relevant Population
0.9
(0.1-1.7)
0.4
(0.1 -0.7)
2.5
(0.2-4.7)
2.1
(0.7-3.5)
0.5
(-1.3-2.4)
1.2
(0.4-1.9)
1.2
(0.4-2)
2.4
(0.8-4.1)
8.7
(5.5-11.9)
1.3
(-3.9-6.4)
1.9
(0.6-3.2)
1.9
(-3.1 -6.7)
1.7
(0.6-2.8)
2.6
(0.9-4.4)
Percent of Total
Incidence
0.3%
(0% - 0.6%)
0.2%
(0.1% -0.3%)
0.9%
(0.1% -1.8%)
0.8%
(0.2% -1.3%)
0.2%
(-0.5% -0.8%)
0.4%
(0.1% -0.7%)
0.3%
(0.1% -0.6%)
0.5%
(0.2% - 0.8%)
1.6%
(1%-2.2%)
0.4%
(-1.1% -1.9%)
0.6%
(0.2% - 0.9%)
0.3%
(-0.5% -1.2%)
0.3%
(0.1% -0.5%)
0.6%
(0.2% - 0.9%)
*AII results are for mortality (among all ages) associated with short-term exposures to 03.  All results are based on single-pollutant models.
**lncidence was quantified down to estimated policy relevant background levels.  Incidences are rounded to the nearest whole number; incidences per 100,000
relevant population and percents are rounded to the nearest tenth.
Note: Numbers in parentheses are 95% confidence or credible intervals based on statistical uncertainty surrounding the O3 coefficient.
                                                                    5-60

-------
       We observe from Tables 5-16 and 5-17 that estimates of (Vrelated non-accidental
mortality reported by Schwartz (2004) for Chicago, Detroit, and Houston, based on both single
city and multi-city concentration-response functions, tend to be higher than other estimates for
these locations. This is mainly due to the use of the 1-hr maximum Os concentration in Schwartz
(2004), rather than the 24-hr average, as the exposure metric. The changes from recent (2004 or
2002)) 1-hr maximum to background 1-hr maximum Os concentrations were generally larger in
the assessment locations than the corresponding changes from recent 24-hr average to
background 24-hr average Os concentrations. For example, for 2004 air quality the estimated
(Vrelated (non-accidental) mortality in Detroit based on Bell et al. (2004), which used a 24-hr
average indicator, ranged from 0.2% (based on 95 city model) to 0.4% of total incidence (based
on single-city model). In contrast, the estimated (Vrelated (non-accidental) mortality  in Detroit
based on Schwartz (2004), which used a 1-hr maximum 63 concentration as the indicator, ranged
from 0.7% (based on 14 city model)  to 1.4% (based on single-city model).
       Figures 5-6a and b show the estimated annual percent of non-accidental mortality
associated with short-term exposure to O3 concentrations within specified ranges for the warm
Os season (April 1 to September 30) in two recent years. While the current Os standard is
expressed in terms of an 8-hr daily maximum inidicator, the large multicity non-accidental (Bell
et al. (2004) and cardiorespiratory (Huang et al. (2004) mortality studies reported concentration-
response relationships for 24-hr average Os levels.  Thus, the intervals shown in this figure are
for 24-hr average concentrations. To provide some perspective on the 24-hr intervals shown,
scatter plots comparing 8-hr daily maximum concentrations at the highest monitor with the
average of the 24-hr average over all monitors within an urban area were developed and are
included in Appendix 5A.2. These scatter plots show that 8-hr daily maximum concentrations on
average are roughly twice the observed 24-hr average levels, although there is considerable
variability in this relationship from day-to- day within an urban area.
       As shown in Figure 5-6a, in 2004, all O3-related non-accidental mortality was associated
with Os concentrations less than 0.06 ppm, 24 hr average, and most of that was associated with
63 concentrations less than 0.04 ppm, 24-hr average.  As shown in Figure 5-6b, in 2002, all (V
related non-accidental mortality was associated with Os concentrations less than 0.08 ppm, 24-hr
average and the great majority was associated with Os concentrations less than 0.05 ppm, 24-hr
average.  The results for cardiorespiratory mortality follow a similar pattern and are shown in
Figure 4-15 in the Risk Assessment TSD.
                                      5-61

-------
Figure 5-6.  Estimated Annual Percent of Non-Accidental Mortality Associated with Short-
             Term Exposure to Recent Os Concentrations Above Background for the
             Period April - September (Based on Bell et al., 2004) - Total and Contribution
             of 24-Hour Average Os Ranges*

                              Figure 5-6a.  Based on 2004 Air Quality
                                                           d Attributable to 0.05 ppm<=ozone<0.06 ppm

                                                           D Attributable to 0.04 ppm<=ozone<0.05 ppm

                                                           • Attributable to 0.03 ppm<=ozone<0.04 ppm

                                                           D Attributable to ozone < 0.03 ppm
                              Figure 5-6b.  Based on 2002 Air Quality




0 1.00% -,
i
•O 0.80%
.§.
| 0.60%
- 0.40%
I
•5 0.20%
1 0.00%
























-
—
1



i

1
Boston



Chicago
[




























1

•







! !
-
B
d Attributable to 0.07 ppm<=ozone<0.08 ppm
• Attributable to 0.06 ppm<=ozone<0.07 ppm
n Attributable to 0.05 ppm<=ozone<0.06 ppm
n Attributable to 0.04 ppm<=ozone<0.05 ppm
• Attributable to 0.03 ppm<=ozone<0.04 ppm
n Attributable to ozone < 0.03 ppm




-
1 • !
3 3!



»

f

NewYoik




•

Philadelphia
i



"

•



!



!
i

•


1"

*Note that as shown in scatter plots in Appendix 5A.2, 8-hr daily maximum concentrations at the highest monitor are
roughly twice the level of the average of the 24-hr average O3 concentrations over all monitors within an urban area
which are used in this figure, this ratio varies across areas.
                                         5-62

-------
         5.4.2   Just Meeting Current and Alternative Ozone Standards
       As described in Chapter 4 and briefly in section 5.3.2.2, the risk estimates described in
this section represent the risks for two separate Os seasons based on adjusting the Os levels
observed in 2004 or 2002 to simulate Os levels associated with just meeting the current 0.08 ppm
standard and several potential alternative 8-hr standards, using the 3-year design value from the
2002-2004 time period.  To facilitate comparison of risk estimates across the urban areas, figures
used in this section present summaries of the risk estimates for the current and potential
alternative 8-hr daily maximum standards.  Most of the figures and tables in this section examine
the risks associated with alternative standards using the average 4th-highest daily maximum 8-hr
average form of the current standard. We present only limited results for several additional
alternative standards in this section.  Risk estimates for three additional alternative 8-hr standards
(0.084 and 0.074 ppm, using an average of the annual 3rd-highest daily maximum 8-hr
concentrations averaged over the three year period, and 0.074 ppm using an average of the
annual 5th-highest daily maximum 8-hr  averages over the three year period are more fully
presented in tables in the Risk Assessment TSD. Because we had to simulate the profiles of 63
concentrations that just meet the current and alternative 8-hour daily maximum O3 standards in
each location, there  is additional uncertainty surrounding estimates of the reduced incidence
associated with 63 concentrations that just meet these 63 standards.
       This section first discusses the risk estimates  for lung function responses in all and
asthmatic school age children, which are based on exposure-response relationships derived from
controlled human exposure studies, and then risk estimates are explored for respiratory
symptoms in asthmatic children, respiratory-related hospital admissions, and premature mortality
which are based on concentration-response relationships obtained from epidemiological studies
       The risk estimates for lung function responses are for the 63 season, which is all year in 3
of the study areas (Houston, Los Angeles, and Sacramento) and which is generally 6-7 months
long in the other 9 urban study areas (e.g., March or April to September or October).  The risk
estimates for lung function responses in all school age children (ages 5 to 18) for just meeting the
current 8-hr standard for 12 urban areas are summarized in Tables 5-6 and  5-7 presented in the
previous section.  Similarly, risk estimates for lung function responses in asthmatic school age
children (ages 5 to 18) for just meeting the current 8-hr standard for 5 urban areas are
summarized in Tables 5-8 and 5-9 in the previous section. Additional risk  estimates for all and
asthmatic school age children are presented in the Risk Assessment TSD and Post (2007),
including estimates based on adjusting 2003 air quality to just meet the current and several
alternative standards.
                                       5-63

-------
       Figure 5-7 shows the median estimates of the percent of all school age children estimated
to experience at least one FEVi decrement > 15% during the Os season across the 12 urban areas
for recent air quality (2002) and upon just meeting the current and several alternative 8-hr
standards. Figure 5C-1 in Appendix 5C of this Staff Paper shows a similar figure based on 2004
air quality data. For just meeting the current 8-hr standard the ranges of median estimates across
the 12 urban areas are  1.0 - 5.2% based on adjusting 2002 air quality data and 0.7 - 1.9% based
on adjusting 2004 air quality data.  In terms of total occurrences of FEVi decrement > 15%
during the Os season, Table 5-7 shows a range of median estimates from about 70,000 to nearly
680,000 responses during the 63 season for all school age children based on adjusting 2002 air
quality data to just meeting the current 8-hour standard and from 40,000 to nearly 380,000
responses across the 12 urban areas associated with adjusting 2004 Os concentrations to just
meeting the current 8-hour standard.
       As an illustration of the changes in the number of school age children estimated to
experience FEVi decrements > 15% across the range of alternative standards, under the current
standard the median estimates range from 9,000 to about 130,000 children per urban area across
the 12 urban areas and this would be reduced to a range of 3,000  to 41,000 children under the
most stringent alternative standard examined (i.e., 0.064 ppm, 4th-highest daily 8-hr maximum).
Somewhat lower estimates are observed based on adjusting 2004 air quality to just meet the
current and alternative 8-hr standards, with a range from 4,000 to about 67,000 children for just
meeting the current standard which is reduced to a range from 1,000 to 20,000 children under the
0.064 ppm, 4th-highest daily 8-hr maximum standard.  By comparing the estimated number of
occurrences  shown in Tables 5C-5  with the number of children estimated to experience 1 or
more responses shown in Tables 5C-1, one can get an estimate of the average number of
occurrences  of a given response in an Os season. For example, for Atlanta it is estimated that
31,000 children would have an FEVi decrement > 15% and that there would be 159,000
occurrences  of this response in this same population when 2002 air quality is  adjusted to just
meet the current 8-hr standard.  Thus, on average it is estimated that  there would be about 5
occurrences  per 63 season per responding child for air quality just meeting the current 8-hr
standard in this urban area. We recognize that some children in the population might have only 1
or 2 occurrences while others likely have 6 or more occurrences per  O?, season.
       Figure 5-8 shows the 95% confidence intervals  for the lung function risk estimates for
each of the 12 urban areas using the FEVi decrement >  15%  health response for recent Os levels
(2002) and for 2002 air quality adjusted to just meet the current and  alternative 8-hr average nth
daily maximum standards. A comparable figure (Figure 5C-2) using 2004  air quality and
adjusting 2004 air quality to just meet the current and alternative 8-hr standards is included in
Appendix 5C.
                                      5-64

-------
       Figures 5-9 shows the median estimates of the percent of asthmatic school age children
estimated to experience at least one FEVi decrement > 10% during the Os season across the five
urban areas for recent air quality (2002) and upon just meeting the current and several alternative
8-hr average 4th-highest daily maximum standards. Figure 5C-3 in Appendix 5C of this Staff
Paper shows a similar figure based on 2004 air quality data.  For just meeting the current 8-hr
standard the ranges of median estimates across the 5 urban areas are 3.0 - 6.2% based on
adjusting 2004 air quality data and 3.4 - 9.6% based on adjusting 2002 air quality data. In terms
of total occurrences of FEVi decrement > 15% during the Os season, Table 5-9 shows a range of
median estimates from about 60,000  to nearly 230,000 responses during the 63 season for
asthmatic school age  children based on adjusting 2004 air quality data to just meeting the current
8-hour standard and from about 40,000 to nearly 470,000 responses across the 5 urban areas
associated with adjusting 2002 63 concentrations to just meeting the current 8-hour standard.
       Figure 5-10 shows the 95% confidence intervals for the lung function risk estimates for
asthmatic school age  children in each of the 5 urban areas using the FEVi decrement > 10%
health response for recent O3 levels (2002) and for 2002 air quality adjusted to just meet the
current and alternative 8-hr average 4th-highest daily maximum standards. A comparable figure
(Figure 5C-4) using 2004 air quality  and adjusting 2004 air quality to just meet the current and
alternative 8-hr standards is included in Appendix 5C.
       Figure 5-11 summarizes respiratory symptom response risk estimates associated with O?,
exposures during the  April to September period for moderate/severe asthmatic children ages 0 to
12 in the Boston urban area based on the concentration-response relationships reported in Gent et
al. (2003) for 2002 air quality and the current and alternative 8-hr standards based on adjusting
2004 air quality data. Figure 5C-5 (Appendix 5C) presents comparable estimates associated with
2004 air quality and just meeting the current and alternative 8-hr standards based on adjusting
2004 air quality data. These figures include risk  estimates for chest tightness based on single
pollutant models and models that included PM2.5. Two additional symptom endpoints, shortness
of breath and wheeze are included in the tables in the Risk Assessment TSD and show similar
patterns as the risk estimates for chest tightness.
       The median estimated number of days involving chest tightness (using the concentration-
response relationship with only O?, in the model)  ranges from 4,500 (based on adjusting 2004 air
quality) to 6,100 (based on adjusting 2002 air quality) upon meeting the current 8-hr standard
and these are reduced to 3,100 (based on adjusting 2004 air quality) to 4,600 days upon meeting
the most stringent alternative examined (0.064 ppm, 4th-highest daily  maximum 8-hr average).
These same ranges correspond to 8 -  11% of total incidence  of chest tightness upon meeting the
                                      5-65

-------
Figure 5-7. Percent of All Children (Ages 5-18) Engaged in Moderate Exertion Estimated
           to Experience At Least One Lung Function Response (Decrement in FEVi >
           15%) Associated with Exposure to Os Concentrations That Just Meet the
           Current and Alternative Average 4th-highest Daily Maximum 8-Hour
           Standards, for Location-Specific Os Seasons (Based on Adjusting 2002 Air
           Quality)*
        Recent (2002)
               0.084/4
                 0.080/4
                  0.074/4
           Ozone  0.070/4 -
         Stds. (ppm)  0064/4
                                                 Urban Areas
   *An 8-hr average standard, denoted m/n is characterized by a concentration m ppm and an nth daily maximum. So
   for example, the current standard is 0.084/4 - 0.084 ppm, 4th-highest daily maximum 8-hr average. The 4th-highest
   daily maximum standards, denoted m/4, require that the average of the 3 annual nth-highest daily maxima over a
   three year period be at or below the specified level (m ppm).  95% credible intervals based on statistical uncertainty
   surrounding the O3 coefficient are presented in Table 5C-2 in Appendix 5C.
                                        5-66

-------
Figure 5-8.   Percent of All Children (Ages 5-18) Engaged in Moderate Exertion Estimated to Experience At
            Least One Lung Function Response (Decrement in FEVi > 15 %) Associated with Recent Air
            Quality (2002) and Just Meeting the Current and Alternative Average nth Daily Maximum 8-Hour
            Standards, for Location-Specific O3 Seasons (Based on Adjusting 2002 Air Quality)*
                  Atlanta
 Boston
12% T —
O
o -S 10%
i- 5j
= 1
< O) °
c •-
at ~a 4% -
O ^
£: o
o S- 2%
4)
0%
















nN
\fV

• .









i i
T t •
Ozone Concentrations/Standards
	



19%
£• xp <\ n%
~ °* fi°/
0 g
= -.i CO/
< !>
Si ^ 4%
" 1
0) Q- no/
Q. in £.K> -
n%








"""Mi
	 M.t.t
n^ \N ^(b OX ^D \N *C5 OX Ox
Cv^ 
£ c 4%
o 5.
< 2%
c
^ 0%
*
<*S


«










i

•





• i

i


• •







.
1
S"^ \^* ^ft \^* vD \^* ^ft \^* \^*
' O^ O^ 0? A^ A^ A^ A^ 0^
QP QP QP Ov Ov Ov Ov Q?
Ozone Concentrations/Standards
Detroit
o) 12%
"c 10%
o
Q.

£ "c 4% -
O n.
< 2% -
'c
dj
o 0% -





















• .


.


i



i I
t |

Ozone Concentrations/Standards

1007
"2 m%
o
a.
ao 8/°
S ^ R%
2 °>
In c 4%
cj ™
< 2%
'c
ai
o n%

Houston














1 i ,
1 T t f { •

• { i
\^ ^Tb \^ vD \^ ^Tb \^ \^
Ozone Concentrations/Standards
                                          5-67

-------
Figure 5-8.  (Continued)
               Los Angeles
                                                                              New York
TO
C
•o
2.
(/)
c
o
IE
<
"c
o
o
Q.

12% -,

10%

8%
v5 6%
? 4%
~ 2%
0%











T , T , f , + , + , + , + , *

rp\ \fc ft \fc A i> ifb \fc \^
^ ^r <§r ^ ^r ^r ^ ^P ^r
^,00000000
^ Ozone Concentrations/Standards
               Philadelphia
o) 12/°
"i 10%
8 8%
a o"
S s? 6%
ll 4%
^ & 2%
g 0%
^ ^



••

••
••


' " "
T T t
Ozone Concentrations/Standards
St. Louis
^oo/
O) l^/o
_C
"? m%
R
!fi ft%
o ~9 6%
!c -o 4%
< ^" 2%
"c



••
..
••


" "
1 T t
o 0^
JS

-------
       Figure 5-9.  Percent of Asthmatic Children (Ages 5-18) Engaged in Moderate Exertion
                  Estimated to Experience At Least One Lung Function Response (Decrement in
                  FEVi > 10%) Associated with Exposure to O3 Concentrations That Just Meet
                  the Current and Alternative Average 4th-highest Daily Maximum 8-Hour
                  Standards, for Location-Specific Os Seasons (Based on Adjusting 2002 Air
                  Quality)
                                                                                                1-20%
       Recent (2002)
              0.084/4
                 0.074/4
                    0.064/4
Ozone Stds. (ppm)
                                     0%
                                  Atlanta
                                               Chicago
Houston      Los Angeles


Urban Areas
                                                                                      New York
       * An 8-hr average standard, denoted m/n is characterized by a concentration of m ppm and an nth-highest daily
       maximum. So, for example, the current standard is 0.084/4 - 0.084 ppm, 4th-highest daily maximum 8-hr average.
       The 4th-highest daily maximum standards, denoted m/4, require that the average of the 3 annual nth-highest daily
       maxima over a 3-year period be at or below the specified level.  95% credible intervals based on statistical
       uncertainty surrounding the O3 coefficient are presented in Table 5C-5 in Appendix 5C.
                                               5-69

-------
Figure 5-10. Percent of Asthmatic Children (Ages 5-18) Engaged in Moderate Exertion Estimated
            to Experience At Least One Lung Function Response (Decrement in FEVi > 10 %)
            Associated with Recent Air Quality (2002) and Exposure to O3 Concentrations That
            Just Meet the Current and Alternative 8-Hour Standards, for Location-Specific O3
            Seasons: Based on Adjusting 2002 O3 Concentrations*
                  Atlanta
75%
J- ^o 9D%
is
= 
^ °* 1 n%
s ^

-------
current 8-hr standard and to about 5.5 - 8% of total incidence of chest tightness upon meeting a
0.064 ppm, 4th-highest daily maximum 8-hr average standard.  As shown in Tables 5C-7 and
5C-9 (Appendix 5C), the symptom with the greatest incidence is wheeze and is based on an 63
concentration-response relationship that included PM2.5 in the model.  These median estimates
range from about 13,000 days with wheeze (based on adjusting 2004 air quality) to nearly 18,000
days (based on adjusting 2002 air quality) upon meeting the current 8-hr standard and these
estimates are reduced to 9,000 (based on adjusting 2004 air quality) to about 13,000 (based on
adjusting 2002 air quality) upon meeting a 0.064 ppm, 4th-highest daily maximum 8-hr average
standard. Confidence intervals, based on statistical uncertainty reflecting sample size
considerations for incidence and percent of total incidence  are shown in Tables 5C-7 through 5C-
10 (Appendix 5C) based on adjusting 2004 and 2002 air quality.
       Figure 5-12 summarizes unscheduled hospital admission risk estimates for respiratory
illness and asthma in New York City associated with short-term exposures to 63 concentrations
in excess of background levels from April through September under recent air quality and when
the current and alternative 8-hr standards are just met based on adjusting 2004 and 2002 air
quality data, respectively. For total respiratory illness, Figure 5-12 shows about 6.4 cases per
100,000 relevant population, which represents 1.5% of total incidence or 513 cases when 2002
Os levels are adjusted to just meet the current 8-hr standard. For asthma-related hospital
admissions, which are a subset of total respiratory illness admissions, the estimates are about 5.5
cases per 100,000 relevant population, which represents about 3.3% of total incidence or 438
cases for this same air quality scenario. For increasingly more stringent alternative 8-hr
standards, Figure 5-12 shows a gradual reduction in respiratory illness cases per 100,000 relevant
population from 6.4 cases per 100,000 upon just meeting the current 8-hr standard to 4.6 cases
per 100,000 under the most  stringent 8-hr standard (i.e., 0.064 ppm, average 4th-highest daily
maximum) analyzed. The comparable estimates based on adjusting 2004 air quality are shown
in Figure 5C-6 (Appendix 5C) and are somewhat higher, but show a similar pattern of gradual
reduction. Confidence intervals, based on statistical uncertainty reflecting sample size
considerations for incidence, incidence per 100,000 relevant population, and percent of total
incidence are shown in Tables 5C-11 and 5C-12 (Appendix 5C) based on adjusting 2004 and
2002 air quality data to just  meet the current and potential alternative standards.
       Additional respiratory-related hospital admission estimates for three other locations are
provided in the Risk Assessment TSD. We note that the concentration-response functions for
each of these locations examined different outcomes in different age groups (e.g., > age 30 in
Los Angeles, >age 64 in Cleveland  and Detroit, vs. all ages in New York City),  making
comparison of the risk estimates across the areas very difficult. For hospital admissions in
Detroit, none of the  estimates were statistically  significant and the median estimates were
                                       5-71

-------
Figure 5-11. Estimated Symptom-Days for Chest Tightness Among Moderate/Severe
             Asthmatic Children (Ages 0 - 12) in Boston Associated with Recent (April-
             September 2002) Os Levels and with Levels Just Meeting Alternative Average
             4th-Highest Daily Maximum  8-Hour Ozone Standards*
               (Based on Gent  et al., 2003)
                                                                                       r11,000
Recent (2002)  ^
       0.084/4 A
         0.080/4
           0.074/4 ~^
             0.070/4^
   Ozone Stds.    0.064/4
      (ppm)
                           1-day lag/ 1 hr
                           max./ no other
                             pollutants
                                                        h     1-daylag/1hr
                                                        hr      J *        max./ no other
                                                       .5                    p0||utants
                                                    Concentration-Response Model
* An 8-hr average standard, denoted m/n is characterized by a concentration of m ppm and an nth daily maximum.
So, for example, the current standard is 0.084/4 - 0.084 ppm, 4th-highest daily maximum 8-hr average. The 4th-
highest daily maximum standards, denoted m/4, require that the average of the 3 annual nth daily maxima over a 3-
year period be at or below the specified level. 95% confidence intervals associated with these risk estimates are
presented in Table 5C-9 in Appendix 5C.
                                          5-72

-------
Figure 5-12.  Estimated Incidence of (Unscheduled) Respiratory Hospital Admissions per
            100,000 Relevant Population in New York Associated with Recent (April -
            September, 2002) Os Levels and with Os Levels Just Meeting Alternative
            Average 4th-Highest Daily Maximum 8-Hour Standards
            (based on Thurston et al.,  1992)
      Recent
      (2002)  0.084/4
                    0.080/4
 Ozone Stds. (ppm)
-—
1



—
"T-
^
c
— •



	 '
.074/4
-—_.

1

	 ,

-— _
1
-

•
~~~\
—

s


~-
-— _
^~~~~*m
1


--x.
>-
— -~
— ..



• —
________




.— •



1
L
^^


-— _

_JL





^
— —
^~^^~-



_- ^

-— _
-~-


==r
- — .
- — i —
                                                                              -8.0
                                                                              -7.0
                                                                             -6.0
                                                                             -5.0
                                                                             -4.0
                                                                             -3.0
                                                                             -2.0
-1.0
     Incidence per 100,000
      Relevant Population
                                   0.070/4
                                                                   Asthma
                                           0.064/4
                                                     Respiratory
                                                      Illness
                                                                         Diagnosis Category
* An 8-hr average standard, denoted m/n is characterized by a concentration of m ppm and an nth daily maximum.
So, for example, the current standard is 0.084/4 - 0.084 ppm, 4th-highest daily maximum 8-hr average. The 4th-
highest daily maximum standards, denoted m/4, require that the average of the 3 annual nth daily maxima over a 3-
year period be at or below the specified level.  95% confidence intervals associated with these risk estimates are
provided in Table 5 C-12 in Appendix 5C.
                                         5-73

-------
Figure 5-13. Estimated Incidence of Non-Accidental Mortality per 100,000 Relevant
           Population Associated with Recent Air Quality (2002) and with Just Meeting
           Alternative Average 4th-Highest Daily Maximum 8-Hour Ozone Standards
           (Using Bell et al., 2004 - 95 U.S. Cities Function), Based on 2002 Ozone
           Concentrations
   Ozone
    Stds.
   (ppm)
Recent (2002)
      0.084/4
        0.080/4 A
         0.074/4
          0.070/4
            0.064/4
                                        Urban Areas
* An 8-hr average standard, denoted m/n is characterized by a concentration of m ppm and an nth daily maximum.
So, for example, the current standard is 0.084/4 - 0.084 ppm, 4th-highest daily maximum 8-hr average. The 4th-
highest daily maximum standards, denoted m/4, require that the average of the 3 annual nth daily maxima over a 3-
year period be at or below the specified level. 95% confidence intervals associated with these risk estimates are
provided in Figure 5-14.
                                        5-74

-------
negative for 0- and 1-day lags and small but positive for 2- and 3-day lags for COPD-related and
pneumonia hospital admissions.
       Figure 5-13 summarizes the results of the assessment of the reduced non-accidental
mortality risks associated with 63 concentrations above background that just meet the current
and several potential alternative 8-hr daily maximum standards across the 12 urban areas for air
quality adjusted based on 2002 air quality data.  This figure shows the annual median risk
estimates for recent air quality and  for just meeting alternative 8-hr standards based on the O3
coefficients estimated in the studies based on adjusting 2002 air quality  data.  Ranges reflecting
the statistical uncertainty, taking into account sample size considerations, based either on the 95
percent confidence intervals around those estimates (if the coefficients were estimated using
classical statistical techniques) or on the 95 percent credible intervals (if the coefficients were
estimated using Bayesian statistical techniques) are presented in Tables  5C-13 through 5C-16
(Appendix 5C) and in the Risk Assessment TSD. The risk estimates in this figure are based on
the 95-city function reported in Bell et al. (2004) for non-accidental mortality. Additional  risk
estimates for cardiorespiratory mortality are included in the Risk Assessment TSD for 8  of the 12
urban areas. Also, Figure 5C-7 (Appendix 5C) shows comparable risk estimates based on
adjusting 2004 air quality data.
       Figure 5-14 shows the median estimates and 95% credible intervals for each of the 12
urban areas for non-accidental mortality based on the 95-cities concentration-response function
in Bell et al. (2004) for 2002 air quality data and just meeting alternative standards based on
adjusting 2002 air quality data.  Figure 5C-8 (Appendix 5C) presents the comparable figure for
2004 air quality and just meeting alternative standards based on adjusting 2004 air quality  data.
For example, Figure 5-14 shows a median risk estimate associated with just meeting the current
8-hr standard for non-accidental  mortality in Atlanta is around 0.9%  of total incidence and the
95% credible interval is about 0.3% to about 1.5% of total incidence. While the 95% credible
intervals get progressively smaller as one considers more stringent standards, as discussed
previously these credible intervals do not consider overall model uncertainty (e.g., whether or not
the shape of the concentration-response relationship is best represented by a log linear
relationship versus a more sigmoidal shape, particularly at lower Os concentration levels).
       The results in this portion of the risk assessment across the 12 urban areas follow the
same patterns as the results discussed in section 5.4.1 for risks associated with recent year  63
concentrations, because they are largely driven by the same concentration-response function
coefficient estimates and confidence or credible intervals. While there is a noticeable reduction
in the median risk estimates in some of the urban areas between that  associated with a recent
year of air quality and just meeting the current 8-hr standard, the reductions associated with
progressively more stringent alternative 8-hr standards are more modest. Based on adjusting
                                       5-75

-------
Figure 5-14. Annual Warm Season (April to September) Estimated O3-Related Non-Accidental
           Mortality Associated with Recent (2002) Os Levels and Levels Just Meeting
           Alternative 8-hr O3 Standards  (Using Bell et al., 2004 - 95 U.S. Cities Function)
               Atlanta
 Boston
-i- 48
C 4.0
ra =• . ~
g 0 4.^
oj S5 36
o -a 30
0 C
o" " 94
o — Z-H
^ O 18
ai is
Q. ro x o
>, 3 kz
•5 r» D 6
•e Q-
o 00












nN
\fV
-cS ^
0^







\^ ^ft
Q^ Q^
Ozone

• .


• 1 1 1 I i

Q? A^ A^ A^ A^ <&
y Q ^y Sy ^y Qr
Concentrations/Standards
„ 48
In 42
~Q> S^ 36
a: So •3-°
0 £ 30
P. c
o ' —
^ o 18
% «
12
>» 3
= g- 06
ra o u.o
o 00













•


•
i



• ,


JS 0- 0- 0
^P Ozone













• i


• i







Concentrations/Standards
               Chicago
*- 48
1 —
g 0 4-2
ai S? 36
Qi 10 °-D
§ S 30
0 C
o " 94
?r
i- o 18
?• — 12
>• = '^
==0 06
•E Q.
o n n
















n*
<^ ^
(j





• •

Ozone

• i


• i


• .


• •


.


• l
T
Q? A^ A^ A^ A^ . 3 !'Z
±; Q.
"c5 o 06
•e Q-
o n n
















Ox^
/ °






• i

Ozone

• i








• i


• i


• l
T
Q? A^ A^ A^ A^ 
o T:
cT "
T- C
i_ o
5 ~
Q. ro
>. 3
£ Q.
ra p
t Q-
o
S

^e
4 8
4 2

3.6

3.0
2 4
1 R

1.2

0.6
0 0

















~^"
Ozone Concentrations/Standards
                                     5-76

-------
Figure 5-14 (continued)
                  Los Angeles
New York
A ft
c
5s" 49
go 4-^
at s? -3 R
° ^ n
0 C
§" 94
1 — ^-^
^ o 18
ai ~ L0
°- ra
£a
15 o 06
° 00







111
t T + + f + +

t







^ 0- 0- 0- 0- 0- 0- 0- 0-
^ Ozone Concentrations/Standards
Philadelphia
Mortality per 100,000 Relevant
. Population (and 95% Cl)
%
^t OO->-->-K>COCO.N.N
/^> bcnk>boji.ba5k>bo







..
"




•







Ozone Concentrations/Standards
St. Louis
Mortality per 100,000 Relevant
^ Population (and 95% Cl)
%
*^t OO->-->-K>COCO.N.N
^ bcnKoboi-bcnkibo








* *
•


'




Ozone Concentrations/Standards

4.8
c
ra =• 42
^j O
§ ? 3.0
0 C
§ £ 2.4
>. 3 1-2
£ Q.
ra o 06
•e Q-
| 0.0











11


'III
T T T
O^N \^ ^ft \^ vD \^ ^ft \^ \^
Qi Qi 9i^ A.'1 A.'1 A.'1 ^t Oi
Ozone Concentrations/Standards
Sacramento
4 8
c
ra =• 4 9 -
go 4-2
o "^ 30-
o -c °-u
P. c
o ' —
T- C
5 ~ L°
?" — 19
>> 3 I -^
~S o n R
^ Q. U-D
o 00








, ,


















^ 0- 0- 0- 0- 0- 0- 0- 0-
^® Ozone Concentrations/Standards
Washington, D.C.
4 a
c
ra • — 42

o °* Q n
o -c -3-1-1
0 C
cT " 9 4 -
lr
*- o 18
Q. (Q
>^ •= 19
~ Q.
ra O OR
0 nn







••
••






















o"^ \^ ^ft \^ vD \^ ^ft \^ \^
^° Ozone Concentrations/Standards
* An 8-hr average standard, denoted m/n is characterized by a concentration of m ppm and an nth daily maximum.
So, for example, the current standard is 0.084/4 - 0.084 ppm, 4th-highest daily maximum 8-hr average. Percents are
median (0.5 fractile) percents of children.  The bars are 95% credible intervals based on statistical uncertainty
surrounding the O3 coefficient.

                                             5-77

-------
2002 air quality data, the range of median estimates associated with 63 upon just meeting the
current standard is about 0.3 to 2.4 cases per hundred thousand relevant population across the 12
urban areas and this range is reduced to about 0.1 to 1.7 cases per 100,000 relevant population
upon just meeting the most  stringent alternative standard analyzed (0.064 ppm, average 4th-
highest daily maximum 8-hr average) We also note that the risk estimates expressed in terms of
incidence per 100,000 population are noticeably smaller for Houston based on both 2002 and
2004 air quality data and for Los Angeles based on 2002 air quality, especially upon just meeting
the current or alternative 8-hr standards than the other urban areas. The risk estimates are
notably  higher in most of the urban areas for 2002 air quality data and air quality data simulated
to just meet the current and  alternative standards based on adjusting 2002 data.
      As shown in Table 5C-13 through 5C-16 in Appendix 5C of this chapter, estimated O^-
related (non-accidental) mortality reported by Schwartz (2004) for Chicago, Detroit, and
Houston, based on both the  single-city and the multi-city concentration-response functions,  tend
to be higher than the Bell et al. (2004) estimates in those locations in large part because Schwartz
used the 1-hr maximum O3  concentration, rather than the 24-hr average, as the exposure metric.
The changes from 1-hr maximum Os concentrations that just meet the current 8-hr Os standard to
background  1-hr maximum  63 concentrations were generally larger in these assessment locations
than the corresponding changes using the 24-hr average metric.
      Figure 5-15a and b shows the estimated annual percent of non-accidental mortality mortality
associated with short-term exposure to 63 concentrations that just meet the current 8-hour daily
maximum standard that fall  within specified ranges. The pattern of results is similar to the pattern
seen for recent year Os concentrations discussed in section 5.4.1.  Using simulated O?, concentrations
that just meet the current 8-hour standard based on 2004 air quality data, all  (Vrelated non-accidental
mortality was associated with Os concentrations less than 0.06 ppm, 24-hr average and most of that
was associated with 63 concentrations less than 0.04 ppm, 24-hr average. Using simulated 63
concentrations that just meet the current 8-hour standard based on 2002 air quality data, all O3-related
non-accidental mortality was associated with Os concentrations less than 0.08 ppm, 24-hr average and
the great majority was associated with 63 concentrations less than 0.05 ppm, 24-hr average.  The
results for cardiorespiratory mortality follow  a similar pattern.  As discussed in section 5.4.1, scatter
plots comparing 8-hr daily maximum concentrations at the highest monitor with the average of the 24-
hr average over all monitors within an urban area were developed and are included in Appendix 5A.2
to provide some perspective on the 24-hr intervals shown.  These scatter plots  show that 8-hr daily
maximum concentrations on average are roughly twice the observed 24-hr average levels,
although there is considerable variability in this relationship from day-to-day within an urban
area.  There  also is some variability in this relationship between 8-hr daily maximum and 24-hr
average levels across the 12 urban areas.
                                       5-78

-------
Figure 5-15. Estimated Annual Percent of Non-Accidental Mortality Associated with
              Short-Term Exposure to Os Above Policy Relevant Background for the Period
              April - September When the Current 8-Hour Standard is Just Met (Based on
              Bell et al., 2004) - Total and Contribution of 24-Hour Average Os Ranges

                        Figure 5-15a. Based on Adjusting 2004 Air Quality Data
             1.00%
D Attributable to 0.04 ppm<=ozone<0.05 ppm
• Attributable to 0.03 ppm<=ozone<0.04 ppm
D Attributable to ozone<0.03 ppm
             0.80%
             0.60%
             0.40%
             0.20%
             0.00%
                       Figure 5-15b. Based on Adjusting 2002 Air Quality Data
             1.00%
• Attributable to 0.06 ppm<=ozone<0.07 ppm
D Attributable to 0.05 ppm<=ozone<0.06 ppm
D Attributable to 0.04 ppm<=ozone<0.05 ppm
• Attributable to 0.03 ppm<=ozone<0.04 ppm
D Attributable to ozone<0.03 ppm
             0.80%
             0.60% -
          •=  0.40% -
             0.20% -
             0.00%
*Note that as shown in scatter plots in Appendix 5A.2, 8-hr daily maximum concentrations at the highest monitor are
roughly twice the level of the average of the 24-hr average O3 concentrations over all monitors within an urban area which
are used in this figure, although this ratio varies across areas.
                                           5-79

-------
         5.4.3   Sensitivity Analyses
       We have conducted sensitivity analyses examining the influence of alternative
assumptions about background (i.e., PRB) levels on the lung function and mortality risk
estimates and the impact of alternative assumptions about the shape of the lung function
exposure-response relationship on the lung function health risk estimates.  These sensitivity
analyses were motivated by the discussion at the August 24-25, 2006 CASAC meeting and are
presented below and in sections 3.3 and 4.3 of the Risk Assessment TSD.
       Reflecting the discussion at  the August CASAC meeting, the CASAC panel suggested
(Henderson, 2006c) that one approach to deal with the uncertainties surrounding estimation of
PRB levels would be to assess the change in total risk (i.e., all O3-related risks above 0 ppm)
associated with alternative standards relative to the total risks associated with just meeting the
current standard, without subtracting estimated risks associated with PRB levels from either
estimate. As shown in  Tables 4-48  and 4-49 of the Risk Assessment TSD, where non-accidental
mortality risks associated with O3 were estimated for all  days above 0 ppm for two recent years
of air quality (2002 and 2004), the largest part of the total risk was related to levels between 0
and our baseline estimated PRB levels.  Adopting the approach suggested by CASAC would
place emphasis on the region of the concentration-response relationship (i.e., at levels below
0.035 ppm down to 0 ppm) where there  is the greatest uncertainty about whether effects occur.
In addition, the approach suggested by CASAC only addresses risks relative to the current
standard and does not address the risk remaining upon meeting the current and alternative
standards, which is an important consideration in setting a NAAQS. For assessing risks
remaining upon just meeting a standard, EPA has decided as a matter of policy that only risks in
excess of PRB are relevant to the decision, and thus staff judges it is still appropriate to estimate
risks in excess of estimated PRB levels.
       As discussed below, we have examined the change in lung function risk estimates
associated with alternative standards relative to the current standard and have found that these
estimates are generally  insensitive to alternative assumptions about PRB. Similarly, changes in
non-accidental mortality risk estimates for just meeting alternative standards relative to the
current standard also are less sensitive to assumptions about the levels used to represent PRB.
We recognize that the lung function and non-accidental mortality risk estimates remaining upon
just meeting the current and alternative standards are impacted to varying degrees by the
assumptions about PRB levels depending on area, year of air quality, and health endpoint.

        5.4.3.1 Impact of Alternative Assumptions About Background
       Risk estimates associated with O3 concentrations discussed in this chapter and in the Risk
Assessment TSD have been developed- either based on O3 concentrations from a recent year of
                                      5-80

-------
air quality or 63 concentrations "rolled back" to just meet a standard - above PRB.  We selected
three locations - Atlanta, Los Angeles, and New York - for a sensitivity analysis for lung
function responses, and calculated lung function responses using (1) the original PRB estimates,
(2) lower PRB estimates for each location, and (3) higher PRB estimates for each location. We
also conducted a sensitivity analysis for non-accidental mortality associated with Os exposure for
all 12 urban areas. For all of the urban areas, except Atlanta, the lower PRB estimates were
calculated by subtracting 5 ppb from the original PRB estimates; for Atlanta, the lower PRB
estimates were calculated by subtracting 10 ppb from  the original PRB estimates. In all
locations, the higher PRB estimates were calculated by adding 5 ppb to the original PRB
estimates.15
       The lung function sensitivity analyses for alternative estimates of PRB were run for all
school age children, with response defined as a decrement in FEVi >15%, and for asthmatic
school age children, with lung function response defined as a decrement in FEVi >10%.  Table
5C-17 shows the results of this sensitivity analysis for all school age children in terms of the
number of children estimated to experience a lung function response of concern (i.e., FEVi
>15%) based on adjusting 2002 and 2004 air quality to just meet the current and two alternative
8-hr standards.  Additional tables showing the sensitivity analysis results for total occurrences of
this lung function response for all children are included in the Risk Assessment TSD (see section
3.3.1). The sensitivity analysis results for lung function responses in asthmatic children are
similar in pattern and also are presented in the Risk Assessment TSD (see section 3.3.1).
       The impact of alternative lower and higher assumed PRB levels on lung function
responses in all school age children (with responses defined as a decrement in FEVi > 15%,) was
relatively small, generally much less than +/- 3%.  Assuming lower PRB levels increased the
estimated number of children with a response, while assuming higher PRB levels decreased the
estimated number of children with a response.  In terms of total occurrences of moderate lung
function responses, different assumptions about PRB had a somewhat larger impact, but the
impact was still generally less than about +/- 10% relative to our base case assumption for PRB.
       Figure 5-16 shows the impact of lower and higher PRB assumed levels in terms of the relative
percent changes in non-accidental mortality risk from the current 8-hr standard based on adjusting
2002 air quality data in each of the 12 urban areas.  One observes that for most, but not all of the
locations, the general pattern is not significantly impacted by the choice of PRB assumptions. The
       1 Summarizing its assessment of the validity of the GEOS-CHEM model, the O3 CD states, "in conclusion,
we estimate that the PRB ozone values reported by Fiore et al. (2003) for afternoon surface air over the United
States are likely 10 ppbv too high in the southeast in summer, and accurate within 5 ppbv in other regions and
seasons." These error estimates are based on comparison of model output with observations for conditions that most
nearly reflect those given in the PRB definition, i.e., at the lower end of the probability distribution.
                                       5-81

-------
choice of PRB has a somewhat greater impact on the non-accidental mortality based on adjusting 2004
air quality data (see Figure 4-19 in the Risk Assessment TSD), which is likely due to the significantly
lower Os levels associated with just meeting the alternative standards based on this year of air quality
data.
       Results of the PRB sensitivity analysis for non-accidental mortality associated with Os
exposures expressed in terms of absolute estimates are presented in Table 5C-18 for 2002 air
quality adjusted to just meet the current standard.  Additional sensitivity analysis results for
recent air quality (both 2002 and 2004 air quality) and for 2004 and 2002 air quality adjusted to
just meet the current and two alternative standards are included in the Risk Assessment TSD (see
section 4.3 and Appendix I). Table 5C-18 illustrates the impact of alternative assumed PRB
levels on incidence of Os-related non-accidental mortality per 100,000 population. Lower and
higher assumed PRB levels generally resulted in increased and decreased, respectively, estimates
in the incidence of mortality per 100,000 population. As shown in this table, estimates assuming
lower PRB levels results in increased estimates of non-accidental mortality incidence per
100,000 that are often 50 to 100% greater than the base case estimates. Similarly, estimates
assuming higher PRB levels results in decreased estimates of non-accidental mortality  incidence
per 100,000 that are 50% or greater less than the base case estimates.
       As discussed in section 4.3 of the Risk Assessment TSD, because Os concentrations just
meeting the current standard are substantially lower than Os concentrations observed for the
recent years of air quality (for most of the urban areas), the change in the assumed PRB levels
had a greater impact on the estimates of mortality  associated with levels just meeting the current
standard, in terms of percent change in the estimate. Similarly, changing the estimates of PRB
tended to have progressively greater impacts on the estimates of mortality risk as progressively
more stringent standards were considered. Not surprisingly, assumptions about PRB have a
greater impact on risk estimates associated with the most stringent standard examined,  since a
greater percentage of days is impacted, in terms of being classified as above or below PRB, by
the assumptions concerning PRB levels.

        5.4.3.2 Impact of Alternative Assumptions About the Shape of Exposure-Response
              Relationships for Lung Function Decrements
       As described in section 5.3.1.3, the exposure-response functions used in the primary
analyses are based on the assumption that the relationship between exposure and response has a
                                       5-82

-------
       Figure 5-16.  Sensitivity Analysis of Estimated Percent Change in Os-related Non-
                       Accidental Mortality (Using Bell et al., 2004 - 95 Cities) From the Current
                       Standard  to Alternative 8-hr Standards and a Recent Year of Air Quality,
                       Using Base Case, Higher, and Lower PRB Estimates*

                                Figure 5-16a. Based on Adjusting 2004 O3 Concentrations
                         Atlanta
                                                                                    Boston
•0 lUU.UTo -
55



£ -100 0% -
0)
O)
c -150 0% -
O
+- -200 0% -
o> -250.0% -


••- r^z^^f?*1*
* 	 ?••*

A'




___---*^J:::t:t^

— • — Original PRB Estimates
- ••- Lower PRB Estimates
- - * - - Higher PRB Estimates












ra
•o
= 50 0% -
V)
£


£ 1 00 0% -
S)
c 1 50 0% -

0)
ir -950 0% -
.---*
—^***^=^1~~~~~'
p. j • - ' " '
*••"

— • — Original PRB Estimates
- ••- Lower PRB Estimates
--*- -Higher PRR Estimates


           2004 air quality    0.084/4      0.074/4      0.064/4
                       Alternative 8-Hr Standards
                                                                      2004 air quality    0.084/4      0.074/4      0.064/4
                                                                                   Alternative 8-Hr Standards
•E
2
3
O
e.
o
•E
    100.0%
     50.0%
0.0% - -
 -50.0%

-100.0%

-150.0%

-200.0%

-250.0%
                        Chicago
                                                                                   Cleveland
                          —•—Original PRB Estimates
                          - ••- Lower PRB Estimates
                          --*--Higher PRB Estimates
          2004 air quality   0.084/4       0.074/4      0.064/4
                       Alternative 8-Hr Standards
                                                                 in
                                                                 •E
                                                                     100.0%
                                                                      50.0%
                                                                      0.0%
                                                                -50.0%
o
E
2  -100.0%
                                                                    -150.0%
                                                               -200.0%
                                                                    -250.0%
—»—Original PRB Estimates
- *- Lower PRB Estimates
•-*•-Higher PRB Estimates
                                                                      2004 air quality    0.084/4      0.074/4      0.064/4
                                                                                   Alternative 8-Hr Standards
                         Detroit
                                                                                    Houston
in
•E
o
2
o
0)
ii
50 0% -




-1500% -

-9^0 0% -
.-••*
, , A* 	 . 	 ^r^— •
^^'••'''•'^^
A'
— » — Original PRB Estimates
-m- Lower PRB Estimates
• - * - - Higher PRR Estimates


           2004 air quality    0.084/4      0.074/4      0.064/4
                       Alternative 8-Hr Standards
                                                            in
                                                            •E
                                                            o
                                                            2
                                                            O

                                                            0)
                                                            ii
50 0% -




-1500% -

-9^0 n% -
.-••*
^^^^^"^
i ^^«B i i
wZ^*\-''

— » — Original PRB Estimates
- *- Lower PRB Estimates
• -*• -Higher PRR Estimates


                                                                      2004 air quality    0.084/4      0.074/4      0.064/4
                                                                                   Alternative 8-Hr Standards
                                                      5-83

-------
Sta
Figure 5-16a continued
                   Los Angeles
    100.0%
     50.0%
      0.0%
    -50.0%
o
E
2   -100.0%
o>
=   -150.0%

~   -200.0%
    -250.0%
                             —»—Original PRB Estimates
                             - m- Lower PRB Estimates
                             • - * - • Higher PRB Estimates
      2004 air quality    0.084/4       0.074/4       0.064/4
                    Alternative 8-Hr Standards
                                                                            100.0%
                                                                             50.0%
                                                                              0.0%
O    -50.0%
E
2   -100.0%
                                                                     -150.0%
                                                                           -200.0%
                                                                           -250.0%
                                                                                                 New York
—»—Original PRB Estimates
- m- Lower PRB Estimates
• - * • • Higher PRB Estimates
                                                                                    2004 air quality    0.084/4       0.074/4       0.064/4
                                                                                                 Alternative 8-Hr Standards
                   Philadelphia
                                                                                                 Sacramento
s
in
O

2
£
o
•£
50 0% -




-1500% -

-9^0 n% -

^M— ^^^^^^^
t^^^

— » — Original PRB Estimates
- m- Lower PRB Estimates
• - * - - Higher PRR Estimates


      2004 air quality    0.084/4       0.074/4       0.064/4
                    Alternative 8-Hr Standards
                                                                        o
50 0% -




-150 0% -

-9^0 n% -


r^^^
, • '
*' - -»- Original PRB Estimates
— • — Lower PRB Estimates
• - * - - Higher PRR Estimates


                                                                                    2004 air quality    0.084/4       0.074/4       0.064/4
                                                                                                 Alternative 8-Hr Standards
                           St. Louis
=   -150.0%

~   -200.0%
                                   —•—Original PRB Estimates
                                   - m- Lower PRB Estimates
                                   • - * - • Higher PRB Estimates
      2004 air quality
                           0.084/4       0.074/4
                          Alternative 8-Hr Standards
                                                    0.064/4
                                                                        S
                                                                        in
                                                                  O
                                                                  2
                                                                                        Washington, D.C.
50 0% -




-150 0% -

-9^n n% -
	 A
^^^^^^^^
Z^=^f

— • — Original PRB Estimates
- m- Lower PRB Estimates
• - * • • Higher PRR Estimates


            2004 air quality    0.084/4       0.074/4       0.064/4
                          Alternative 8-Hr Standards
      *The 8-hr average standards shown in these figures, denoted m/n, are characterized by a concentration of m ppm
      and an nth-highest daily maximum. So, for example, the current standard is 0.084/4 ~ 0.084 ppm, 4th-highest daily
      maximum 8-hr average. The figure also compares the current standard to a recent year of air quality.
                                                           5-84

-------
Figure 5-16b.  Based on Adjusting 2002 O3 Concentrations
                       Atlanta
                Boston
•5
3
§i
•£
S.
3


0)
O)
c
£
O

a!
"S.
ra
1
S
in
•£
1
£
P
0)
O)
ra
O
c
0)
s
DL
"E
ra
•O
c
J3
(0
-£
1
g
P
0)
O)
ra
O
c
0)
s














50 0% -




200 0% -




50 0% -




200 0% -
-9^0 n% -



^=^rr?**rr^
A"' • Original PRB Estimates
— ••— Lower PRB Estimates




2002 air quality 0.084/4 0.074/4 0.064/4
Alternative 8-Hr Standards
Chicago

^-^i^1-"""""""--"

f.=^**~-*

— « — Original PRB Estimates
-••- Lower PRB Estimates



2002 air guality 0.084/4 0.074/4 0.064/4
Alternative 8-Hr Standards
Detroit

m^^^=^^^=:=^
t^**^m

— » — Original PRB Estimates
-••- Lower PRB Estimates



                                                                    £  -100.0%
                                                                     0)
                                                                     |  -150.0%
                                                                     ii
                                                                     0)
                                                                     0.
                                                                    o
                                                                    I
                                                                    O
                                                                    •£
                         «   Original PRB Estimates
                       - *- Lower PRB Estimates
                       --^--Higher PRB Estimates
                                                                                 2002 air guality    0.084/4       0.074/4       0.064/4
                                                                                               Alternative 8-Hr Standards
                                                                                               Cleveland
srcent Change from Current Standarc


50 0% -


200 0% -
-9^0 n% -

_^— 4^^=^
1****^ ^"

— • — Original PRB Estimates
- *- Lower PRB Estimates



                                                                                 2002 air guality    0.084/4       0.074/4       0.064/4
                                                                                               Alternative 8-Hr Standards
                                                                                                Houston


50 0% -




200 0% -
-9^0 n% -


"•^^*X'
•^ , •'
A- ' ' — » — Original PRB Estimates
- *- Lower PRB Estimates



      2002 air guality     0.084/4        0.074/4        0.064/4
                     Alternative 8-Hr Standards
2002 air guality    0.084/4       0.074/4       0.064/4
              Alternative 8-Hr Standards
                                                        5-85

-------
      Figure 5-16b continued
                        Los Angeles
                                                                                            New York
in
I
o
I
o
•E


50 0% -


200 0% -
-9^0 n% -
-A- •

.---*

^'S7
m' jS •'
,S •' » Oriainal
PRB Estimates
/ -••- Lower PRB Estimates



A'
            2002 air quality    0.084/4       0.074/4       0.064/4
                         Alternative 8-Hr Standards
                                                                   OT

                                                                   I
                                                                   o
                                                                   I
                                                                   o
                                                                   •E


50 0% -




200 0% -
-9^0 n% -

^^^x^^*
t^^**

— • — Oriainal PRB Estimates
- -•- Lower PRB Estimates



                                                                               2002 air quality    0.084/4       0.074/4      0.064/4
                                                                                            Alternative 8-Hr Standards
I
O
•E
    100.0%
                        Philadelphia
                                                                                            Sacramento
     -50.0%
-100.0%
-150.0%
-200.0%
-250.0%
—•—Original PRB Estimates
- ••- Lower PRB Estimates
- - * - - Higher PRB Estimates
            2002 air quality    0.084/4       0.074/4       0.064/4
                         Alternative 8-Hr Standards
                                                                       
-------
logistic form with 90 percent probability and a linear (hockeystick) form with 10 percent
probability.  We have conducted sensitivity analyses which examine the influence of two
alternative assumptions about the exposure-response functions for lung function decrements in
all and asthmatic school age children. These alternative exposure-response functions are based
on an 80 percent logistic/20 percent linear split and a 50 percent logistic/50 percent linear split,
in five locations - Atlanta, Chicago, Houston, Los Angeles, and New York.  Appendix C of the
Risk Assessment TSD presents tables showing the results of the sensitivity analysis for a change
in FEVi > 15% for all school age children for a recent year of air quality as well as when Os
concentrations just meet each of three 4th-highest daily maximum standards - 0.084/4, 0.074/4
and 0.064/4, based on adjusting 2004 and 2002 data, respectively. Appendix C also shows the
corresponding impacts on the estimated number of asthmatic school age children experiencing at
least one lung function response, defined as a change in FEVi ^ 10%.
       Figures 5-17a and b and 5-18a and b show the impacts of alternative estimates of
exposure-response functions on estimated  percent changes in response among all school age
children and asthmatic school age children, respectively, when 63 concentrations are changed
from those just meeting the current standard to a recent year of air quality  (results are shown
based on both 2004 and 2002 air quality) and to those just meeting each of the two alternative
standards given above.  A positive percent change reflects a reduction in risk relative to just
meeting the current 8-hr standard. Since the comparisons are with respect to just meeting the
current standard, the  percent changes for the recent year of air quality are negative (i.e., going
from just meeting the current standard to 2002 or 2004 air quality represents an increase in risk).
The impacts of changing the functional form of the exposure-response relationship varied
substantially. Changing from the 90 percent logistic/10 percent  linear base case to the 80%/20%
split generally had only a small impact for all school age children, with most risk estimates being
within 5% of the base case estimates. The impacts of changing from the base case to the 50%
logistic/50% linear case were generally (although not always) larger. We observed greater
changes for all school age children between the 50/50 split and the base case in terms of percent
change in risk for the two more stringent alternative standards relative to the current standard.
With respect to the lung function risk estimates for asthmatic children, there were relatively
small changes observed between the 50/50 split and the base case in the percent changes
associated with the two more stringent alternative standards relative to the current standard.

         5.4.4  Comparison with Risk Estimates from Prior Review
       As noted in section 5.1.1, EPA conducted a health risk assessment during the prior Os
NAAQS review. We recognize that two of the health endpoints, lung function (FEVi)
decrements for children and respiratory-related and asthma hospital admissions were included in

                                       5-87

-------
Figure 5-17.  Sensitivity Analysis:  Impact of Alternative Estimates of Exposure-Response Function on
              Estimated Percent Changes From the Current Standard in Numbers of All Children (Ages 5-
              18) Engaged in Moderate Exertion Experiencing at Least One Decrement in FEVi >15%
                          Figure 5-17a.  Based on Adjusting 2004 O3 Concentrations
                    Atlanta
Percent Change from Current Standard
-|OQ% 	
50% -
-50% -
-100% -
-200% -
-250% -
-300% -
_-^^n°A -
	 A
A"'
— « — 90% logistic-10% linear
-m- 80% logistic-20% linear
- - * - -50% logistic-50% linear

   2004 air quality
                         0.084/4         0.074/4
                         Alternative 8-Hr Standards
                                              0.064/4
                                                               100%
                                                              -350% -1
                                                                                            Chicago
                                                                                                    90% logistic-10% linear
                                                                                                  - 80% logistic-20% linear
                                                                                                  - - 50% logistic-50% linear
                                                                     2004 air quality
                                                      0.084/4         0.074/4
                                                      Alternative 8-Hr Standards
                                                                                                                0.064/4
                    Houston
   100%
•E
•§   50% -
c
S5    o%

|  -50% -

  -100% -

  -150% -
0)
= -200% -
.c
H -250%
c
^ -300% -
£
  -350%

>—90% logistic-10% linear
I — 80% logistic-20% linear
t - -50% logistic-50% linear
   2004 air quality
                         0.084/4         0.074/4
                         Alternative 8-Hr Standards
                                              0.064/4
                                                                                         Los Angeles*
                                                                                                    90% logistic-10% linear
                                                                                                    80% logistic-20% linear
                                                                                                   - 50% logistic-50% linear
                                      2004 air quality
 0.084/4        0.074/4
Alternative 8-Hr Standards
                                                                                                                0.064/4
                                                    New York
•H
•S 50% -
x m
| -50% -
" -100% -
o
0)
| -200% -
£
" -250% -
^ -300% -
S.

	 A
yf
If / — • — 90% logistic-10% linear
— * - 80% logistic-20% linear
- - * - - 50% logistic-50% linear
A
2004 air quality 0.084/4 0.074/4 0.064/4
                                                                                          * The percent changes from
                                                                                          the current standard (0.084/4)
                                                                                          to 2004 air quality were
                                                                                          omitted for Los Angeles
                                                                                          because they were so large in
                                                                                          magnitude (-553%, -587%,
                                                                                          and -1027% for the 90/10,
                                                                                          80/20 and 50/50 splits,
                                                                                          respectively).
                                                   Alternative 8-Hr Standards
                                                  5-88

-------
                          Figure 5-17b.  Based on Adjusting 2002 O3 Concentrations
                  Atlanta
                             —*—90% logistic-10% linear
                             — -«	 80% logistic-20% linear
                             - - * - - 50% logistic-50% linear
2002 air quality
 0.084/4         0.074/4
Alternative 8-Hr Standards
                                                0.064/4
                                                                            Chicago
                                                                                        —•—90% logistic-10% linear
                                                                                        	« — 80% logistic-20% linear
                                                                                        - - * - -50% logistic-50% linear
2002 air quality
 0.084/4         0.074/4
Alternative 8-Hr Standards
                                                                                                                           0.064/4
                  Houston
                                   90% logistic-10% linear
                               -«	 80% logistic-20% linear
                                 - - 50% logistic-50% linear
2002 air quality
 0.084/4         0.074/4
Alternative 8-Hr Standards
                                                0.064/4
                                                                    100%
                                                                         Los Angeles*
                                                                                        — » — 90% logistic-10% linear
                                                                                        ........ -11- 80% logistic-20% linear
                                                                                        - - * - -50% logistic-50% linear
2002 air quality      0.084/4         0.074/4
                 Alternative 8-Hr Standards
                                                                                                                           0.064/4
                                                      New York
                                                                   —•—90% logistic-10% linear
                                                                   - -« - 80% logistic-20% linear
                                                                   - - * - - 50% logistic-50% linear
                                      2002 air quality
                                      0.084/4         0.074/4
                                      Alternative 8-Hr Standards
                                                                                     0.064/4
                                                                                 * The percent changes from
                                                                                 the current standard (0.084/4)
                                                                                 to 2002 air quality were
                                                                                 omitted for Los Angeles
                                                                                 because they were so large in
                                                                                 magnitude (-526%,
                                                                                 -549%, and -842% for the
                                                                                 90/10,  80/20 and 50/50 splits,
                                                                                 respectively).
                                                    5-89

-------
Figure 5-18. Sensitivity Analysis: Impact of Alternative Estimates of Exposure-Response Function on
            Estimated Percent Changes From the Current Standard in Numbers of Asthmatic Children (Ages
            5-18) Engaged in Moderate Exertion Experiencing at Least One Decrement in FEVi >10%
                       Figure 5-18a. Based on Adjusting 2004 O3 Concentrations
                  Atlanta
                              Chicago
1 80% -
1 60% -
8 40% -
= 20%-
0 -20% -
| -40% -
I -60%-
|> -80% -
| -100% -
~ -120% -
g -140% -
|j -160% -


^— ~~~*
^-^^

— » — 90% logistic-10% linear
— ••— 80% logistic-20% linear
- - * - -50% logistic-50% linear



Percent Change from Current Standard
2004 air quality 0.084/4 0.074/4 0.064/4
Alternative 8-Hr Standards
Houston
I 80% -
1 60% -
| 40% -
•£ 20% -
0)
0 -20% -
| -40% -
a -60% -
|> -80% -
5 -100% -
~ -120% -
$ -140% -
£ -160% -

— -^


— » — 90% logistic-10% linear
	 II- 80% logistic-20% linear
- - * - -50% logistic-50% linear

2004 air quality 0.084/4 0.074/4 0.064/4
Alternative 8-Hr Standards
Ne
Percent Change from Current Standard
80% -
60% -
40% -
20% -
0% -
-20% -
-40% -
-60% -
-80% -
-100% -
-120% -
-140% -
-160% -
-isn% -
3
*<
O
•?- Percent Change from Current Standard
80% -
60% -
40% -
20% -
-20% -
-40% -
-60% -
-80% -
-100% -
-120% -
-140% -
-160% -
100% -i
80% -
60% -
40% -
20% -
-20% -
-40% -
-60% -
-80% -
-100% -
-120% -
-140% -
-160% -
^^
^^
— » — 90% logistic-10% linear
- «!— 80% logistic-20% linear
- - * - -50% logistic-50% linear
2004 air quality 0.084/4 0.074/4 0.064/4
Alternative 8-Hr Standards
Los Angeles*
^^^
I"
// — » — 90% logistic-10% linear
.' / ••- -111- 80% logistic-20% linear
/ / • • * • • 50% logistic-50% linear
/I
2004 air quality 0.084/4 0.074/4 0.064/4
Alternative 8-Hr Standards
* The percent changes
• •^^ from tne current
. . iJK^^"^ standard (0.084/4) to
^^**°^""^ 2004 air quality, based
'
on the 90/10, 80/20, and
50/50 split exposure-
—•—90% logistic-10% linear leSpOUSe functions fol
- m- 80% logistic-20% linear LOS AngeleS W6r6 -
--*--50% logistic-50% linear 9Q4% -980% and
201%, respectively.
                                2004 air quality
 0.084/4      0.074/4
Alternative 8-Hr Standards
                                                                    0.064/4
                                           5-90

-------
                         Figure 5-18b. Based on Adjusting 2002 O3 Concentrations
                   Atlanta
                                               Chicago
   2002 air quality
   2002 air quality
_^—


— » — 90% logistic-10% linear
— •« - 80% logistic-20% linear
- - * - - 50% logistic-50% linear


ity 0.084/4 0.074/4 0.064/4
Alternative 8-Hr Standards
Houston
^^••^^


— * — 90% logistic-10% linear
— •• — 80% logistic-20% linear
- - * - - 50% logistic-50% linear


1 80%
1 60%
5J 40%
•£ 20%
t 0%
0 -20%
| -40%
* -60%
|> -80%
5 -100%
~ -120%
g -140%
£ -160%
I 80% -
1 60% -
£ 40% -
= 20%-
0 -20% -
| -40% -
I -60%-
|> -80% -
5 -100% -
^ -120% -
§ -140% -
* -160% -
^x*^
^^^
— • — 90% logistic-10% linear
-m- 80% logistic-20% linear
• • * - -50% logistic-50% linear

2002 air quality 0.084/4 0.074/4 0.064/4
Alternative 8-Hr Standards
Los Angeles*
^^^
/ '
:j — « — 90% logistic-10% linear
/ / - m- 80% logistic-20% linear
/ / - - * - - 50% logistic-50% linear
/I 	
ity 0.084/4 0.074/4 0.064/4 2002 air quality 0.084/4 0.074/4 0.064/4
Alternative 8-Hr Standards Alternative 8-Hr Standards
* The percent changes
I 80% -
1 60% -
% 40% -
•£ 20% -
at
5 -20% -
| -40% -
•J -60%-
|> -80% -
5 -100% -
~ -120% -
g -140% -
£ -160% -
-180% -
from the current standard
^^^ (0.084/4) to 2002 air
^^f^^^^^ quality, based on the
_^^ffff^^' ' 90/10, 80/20, and 50/50
A'
r




— • — 90% logistic-10% linear
_ m— 80% logistic-20% linear
- * - -50% logistic-50% linear

split exposure-response
functions for Los Angeles
were -287%, -274%, and
-198%, respectively.
                      2002 air quality
New York
 0.084/4        0.074/4
Alternative 8-Hr Standards
                                                               0.064/4
                                                5-91

-------
the current and prior reviews. The other health endpoints included in the current risk assessment,
lung function decrements in asthmatic children, respiratory symptoms in moderate/severe
asthmatic children, and non-accidental and cardiorespiratory mortality are based on more recent
scientific studies and, were not included in the prior review.
       The lung function risk estimates developed for the current and prior review are based on
exposure distributions generated by running 63 exposure models and exposure-response
relationships developed using the available controlled human exposure studies data. There have
been significant changes in the exposure model between the prior and current review.  As
discussed in Chapter 4, no direct comparison is being made of the differences in exposure
estimates for children engaged in moderate exertion associated with just meeting the current 8-hr
standard between the past and current reviews. This is due to differences in the exposure model,
differences in the population coverage within the urban areas included in the analyses, and
differences in the population definitions included in the past  and current exposure analyses.  Due
to the differences in the exposure analyses, as well as changes in the exposure-response
relationships (e.g., change in the shape of the exposure-response relationship from a linear
relationship to one that is more sigmoidal or s-shaped), and changes in estimates for PRB used in
the assessments, no  direct comparison is being made of the differences in lung function risk
estimates for children between the current and prior review.
       We note that the current estimates for Os-related hospital admissions for respiratory
illness and asthma for New York City are higher than the estimates in the risk assessment
conducted during the prior Os NAAQS review. Both the prior  and current assessments used the
same concentration-response functions for these health outcomes.  The main reason for higher
estimates in the current assessment is the use of a single value of 0.04 ppm for background in the
prior review which is higher than the current modeled values for background in the current
assessment which are in the range of about 0.015 to 0.035 ppm. Thus, under the current risk
assessment O3 levels above background but below 0.04 ppm are contributing additional
estimated cases that were not included in the assessment for the prior Os NAAQS review.

         5.4.5   Key Observations
       In considering the quantitative estimates from the risk assessment the limitations and
uncertainties associated with the risk estimates discussed in section 5.3.1.4 for lung function
decrements and section 5.3.2.5 for respiratory symptoms, hospital admissions, and pre-mature
mortality should be  kept in mind. It is also important to consider the degree of confidence about
the extent to which Os is causally related to each of the effects for which risk estimates were
produced (see section 3.7.5). For example, there is clear and convincing evidence of causality
for lung function decrements in healthy children under moderate exertion for 8-hr average O3

                                       5-92

-------
exposures.  We also judge that there is strong evidence for a causal relationship between
respiratory  symptoms in asthmatic children and Os exposures and between hospital admissions
for respiratory causes and ambient O?, exposures. There is greater uncertainty and somewhat less
confidence  about the relationship between 63 and non-accidental and cardiorespiratory mortality,
although the CD's overall evaluation is that it is highly suggestive that this relationship exists.


Recent Os Air Quality Levels
       Section 5.4.1 has presented risk estimates associated with two recent years of air quality
as represented by 2002 and 2004 monitoring data. Presented below are key observations
resulting from this part of the risk assessment.
   •   The ranges in median estimates of the number of school age children (ages 5-18)
       estimated to experience at least one FEVi decrement > 15% due to 8-hr O3 exposures
       during the 63 season across the 12 urban areas are 10,000 to nearly 220,000 (based on
       2004 air quality) and 23,000 to nearly 320,000 (based on 2002 air quality). In terms of
       percent of this population the ranges in  median estimates are 1.3 to 5.9% (based on 2004
       air quality) and 4.8 to 9.1% (based on 2002 air quality). In terms of estimated
       occurrences of this same response the ranges in median estimates are 64,000 to nearly 1.5
       million (based on 2004 air quality) and  about 130,000 to nearly 1.4 million (based on
       2002 air quality). The average number  of occurrences per child in an Os season ranged
       from about 4 to 7 depending on urban area and year.

   •   The ranges in median estimates of the number of asthmatic school age children (ages 5-
       18)  estimated to experience at least one FEVI decrement >10% due to 8-hr 63 exposures
       during the Os season across the five urban areas are 10,000 to 61,000 (based on 2004 air
       quality) and 16,000 to about 110,000 (based  on 2002 air quality).  In terms of the percent
       of this population the ranges in median  estimates are 4.6 to 13.4% (based on 2004 air
       quality) and 11.5 to 17% (based on 2002). In terms of total occurrences of this same
       response the ranges in median estimates are 98,000  to about 670,000 responses (based on
       2004 air quality) and from about 89,000 to 760,000 responses (based on 2002 air
       quality). Dividing the estimated total number of occurrences by the number of asthmatic
       children estimated to  experience this lung function response, results in each child being
       estimated to have on average from about 6 to 11 occurrences depending on urban area
       and year.

   •   Estimates for increased respiratory symptoms (i.e., chest tightness, shortness of breath,
       and wheeze) in asthmatic children (ages 0-12) who used maintenance medications were
       only developed for the Boston urban area. The ranges in median estimates of symptom
       days for these three health outcomes are about 5,000 to 15,000 (based on 2004 air
       quality) and about 7,000 to 20,000 (based on 2002 air quality). In terms of percent of
       total incidence  for these three health outcomes the ranges in median estimates are about 8
       to 14% (based on 2004 air quality) and  about 11 to 20% (based on 2002 air quality).

   •   Estimates for respiratory-related hospital admissions (e.g,, asthma-related) were
       developed for three urban areas (New York, Los Angeles, and Detroit).  The median

                                      5-93

-------
       estimates for New York are about 380 (based on 2004 air quality) and about 520 (based
       on 2002 air quality) Os-related excess hospital admissions for asthma. For 2004 and
       2002 air quality, these estimates represent about 3 and 4%, respectively, of total
       incidence.

   •   The risk assessment included a variety of estimates based on single- and multi-city
       studies for non-accidental and cardiorespiratory mortality.  Since the median estimates
       from single-city and multi-city studies and models were generally of similar magnitude,
       with a few notable exceptions, we have focused on the estimates based on the multi-city
       studies to compare risk estimates across the 12 urban areas. The median estimates for
       incidence for non-accidental mortality (based on Bell et al., 2004 - 95 cities
       concentration-response function) range from about 3 to 130 (based on 2004 air quality)
       which is about 0.2 to 0.4% of total incidence. These same  estimates based on 2002 air
       quality range from about 10 to 110 which is about 0.2 to 0.6% of total incidence.
       Estimates of Os-related non-accidental mortality reported by Schwartz (2004) for
       Chicago, Detroit, and Houston, based on both single city and multi-city concentration-
       response functions, are somewhat higher than other estimates for these locations.  This is
       mainly due to the use of the 1-hr maximum Os concentration in Schwartz (2004), rather
       than the 24-hr average, as the exposure metric.

   •   Examining the contribution of various Os ranges to these non-accidental mortality
       estimates, we found all of the mortality was associated with 24-hr average concentrations
       less than 0.06 ppm and most of it was associated with concentrations less than 0.04 ppm
       for 2004 air quality. For 2002, all of the Os-related non-accidental mortality was
       associated with 24-hr average concentrations less than 0.08 ppm and the great majority
       was associated with concentrations less than 0.05 ppm.  Based on an examination  of Os
       air quality relationships between 24-hr average concentrations averaged over the urban
       monitors in an urban area on a given day and the highest daily maximum 8-hr average at
       any of the monitors in the same area on the corresponding day, we note that the 8-hr daily
       maximum concentrations are on average about twice the 24-hr average level.  So,  for
       example, a range of 0.04 to 0.06 ppm, 24-hr average corresponds with roughly daily
       maximum 8-hr levels in the range 0.08 to 0.12 ppm measured at the highest fixed-site
       monitor within an urban area.

       Meeting the Current and Alternative 8-hr Standards

       Section 5.4.2 has presented risk estimates associated with just meeting the current and
several potential alternative 8-hr standards based on adjusting 2004 and 2002 monitoring  data
using design values for the 2002-2004 time period. Presented below are key observations
resulting from this part of the  risk assessment.


   •   In comparing risk estimates for alternative standards, uncertainties in quantifying  the
       health risks associated ambient 63 concentrations would be expected to remain relatively
       constant in different models.  Thus, we have greater confidence in relative comparisons
       in risk estimates between alternative standards than in the absolute magnitude of risk

                                       5-94

-------
    estimates associated with any particular standard.  As discussed in section 5.4.3.1,
    differences in risk between alternative standards (e.g., reductions in risk for alternative
    standards relative to the current standard or recent air quality levels) is in most situations
    not impacted by assumptions about estimated PRB levels. Reductions in risks for
    alternative standards relative to the risks associated with just meeting the current standard
    are presented in section 6.3.4.

•   Significant year-to-year variability in 63 concentrations combined with the use of a 3-
    year design value to determine the amount of air quality adjustment to be applied to each
    year analyzed, results in significant year-to-year variability in the annual health risk
    estimates associated with just meeting the current and potential alternative 8-hr standards.

•   The range in median estimates of the number of school age children (ages 5-18)
    estimated to experience at least one FEVi decrement > 15%  due to 8-hr O3 exposures
    during the 63 season across the 12 urban areas when the current 8-hr standard is just met
    (based on adjusting 2002 air quality) are about 9,000 to 131,000. These estimated risks
    would be reduced to a range of 3,000 to 41,000 children under the most stringent
    alternative standard examined (i.e., 0.064 ppm, 4th-highest daily 8-hr maximum -i.e., the
    0.064/4 alternative) based on adjusting 2002 air quality. In terms of percent of this
    population the ranges in median estimates are about 1.0 to 4.6% (based on 2002 air
    quality) for just meeting the current standard and these estimates are reduced to about 0.2
    to 1.4% upon just meeting the 0.064/4 alternative standard. In terms of estimated
    occurrences of this same response the range in median estimates  are 40,000 to about
    340,000 when the current 8-hr standard is just met (based on 2002 air quality)  These
    estimated risks would be reduced to a range of 16,000 to about 130,000 occurrences upon
   just meeting the most stringent alternative standard (0.064/4).  The average number of
    occurrences per child in an 63 season ranged from about 4 to 10 for air quality just
    meeting the current standard across the  12 urban areas (based on 2002 air quality). The
    average number of occurrences per child ranged from 4 to 10 for the most stringent
    alternative standard analyzed (0.064/4). The risk estimates associated with just meeting
    the current and alternative 8-hr standards based on adjusting 2004 air quality are
    generally of similar magnitude, although somewhat lower in 10 of the 12 urban areas.

•   The ranges in median estimates of the number of asthmatic school age children (ages 5-
    18) estimated to experience at least one FEVi decrement > 10% due to 8-hr Os exposures
    during the Os season across the five urban areas are about 8,000 to 58,000 or in terms of
    percentage of this population range from 3.4 to 9.6% associated with just meeting the
    current standard (based on adjusting 2002 air quality).  These estimated risks would be
    reduced to a range of about 3,000 to 26,000 asthmatic children under the most stringent
    alternative standard examined (i.e., 0.064 ppm, 4th-highest daily 8-hr maximum -i.e., the
    0.064/4 alternative) based on adjusting 2002 air quality. In terms of estimated
    occurrences of this same response the range of median estimates are 49,000 to nearly
    470,000 responses associated with just meeting the current standard (based on adjusting
    2002 air quality). These estimated risks would be reduced to a range of 13,000 to nearly
    260,000 occurrences upon just meeting the most stringent alternative standard (0.064/4).
    The average number of occurrences per asthmatic child in an 63  season ranged from

                                   5-95

-------
   about 6 to 13 associated with just meeting the current standard (based on 2002 air
   quality). The average number of occurrences per asthmatic child ranged from 4 to 10
   upon meeting the most stringent alternative examined (0.064 ppm, 4th-highest daily
   maximum 8-hr average). The risk estimates associated with just meeting the current and
   alternative 8-hr standards based on adjusting 2004 air quality are generally of similar
   magnitude, although somewhat lower in 3 of the 5 areas analyzed.

•  Estimates for increased respiratory symptoms (i.e., chest tightness, shortness of breath,
   and wheeze) in moderate/severe  asthmatic children (ages 0-12) were only developed for
   the Boston urban area. The median estimated number of days involving chest tightness
   (using the concentration-response relationship with only Os in the model) ranges from
   4,500 (based on  adjusting 2004 air quality) to 6,100 (based on adjusting 2002 air quality)
   upon meeting the current 8-hr standard and these are reduced to 3,100 (based on
   adjusting 2004 air quality) to 4,600 days upon meeting the most stringent alternative
   examined (0.064 ppm, 4th-highest daily maximum 8-hr average).  These same ranges
   correspond to 8 to 11% of total incidence of chest tightness upon meeting the current 8-hr
   standard and to about 5.5 to 8% of total incidence of chest tightness upon meeting a 0.064
   ppm, 4th-highest daily maximum 8-hr average standard.  Similar patterns of reduction
   were observed for each of the reported respiratory symptoms.

•  Estimates for respiratory-related hospital admissions (e.g,, respiratory illness, asthma-
   related) were developed for three urban locations (New York City, Los  Angeles, and
   Detroit). For asthma-related admissions in New York City the estimates are about 3.9
   cases per 100,000 relevant population, which represents about 2.4% of total incidence or
   313 cases upon just meeting the current standard based on adjusting 2004 air quality data.
   For increasingly more stringent alternative 8-hr standards, a gradual reduction in the
   cases per 100,000 relevant population is observed from 3.9 cases per 100,000 upon just
   meeting the current 8-hr standard to about 2.6 cases per  100,000 under the most stringent
   8-hr standard (i.e., 0.064 ppm, average 4th-highest daily maximum) analyzed. Based on
   adjusting 2002 air quality data, asthma-related admissions in New York City are about
   5.5 cases per 100,000 relevant population, which represents about 3.3% of total incidence
   or about 440 cases upon just meeting the current standard. For increasingly more
   stringent alternative 8-hr standards, a gradual reduction is observed from 5.5 cases per
   100,000 (3.3% of total incidence) upon just meeting the current 8-hr standard to about 3.9
   cases per 100,000 (2.4% of total  incidence).

•  Based  on the median estimates for incidence for non-accidental mortality (based on Bell
   et al., 2004 - 95  cities concentration-response function), meeting the most stringent
   standard shown (0.064 ppm, 4th-highest daily maximum) is estimated to reduce mortality
   by 55 percent of what it would be associated with just meeting the current standard
   (based on adjusting 2004 air quality data).  Adjusting 2002 air quality data to just meet
   the 0.064 ppm, standard results in a 40 percent reduction in non-accidental mortality
   relative to just meeting the current 8-hr standard .  The patterns for cardiorespiratory
   mortality are similar.  The aggregate Os-related cardiorespiratory mortality at the most
   stringent standard shown is estimated to be about 57  percent of what it would be at the
   current standard, using simulated 63 concentrations that just meet the current and

                                   5-96

-------
   alternative 8-hour standards based on 2004 air quality data. Using 2002 air quality data,
   the corresponding result is about 42 percent.

•  Much of the contribution to the risk estimates for non-accidental and cardiorespiratory
   mortality upon just meeting the current 8-hr standard is associated with 24-hr 63
   concentrations between background and 0.04 ppm.  Based on examining relationships
   between 24-hr concentrations averaged across the monitors within an urban area and 8-hr
   daily maximum concentrations, 8-hr daily maximum levels at the highest monitor in an
   urban area associated with these averaged 24-hr levels are generally about twice as high.

   Uncertainty and Variability

•  There is noticeable variability in estimated Os-related incidence of morbidity and
   mortality across the 12 urban areas analyzed for both recent years of air quality and for
   air quality  adjusted to simulate just meeting the current and several potential alternative
   8-hr standards. This variability is likely due to differences in air quality distributions,
   differences in exposure related to many factors including varying activity patterns and air
   exchange rates, differences in baseline incidence rates, and differences in susceptible
   populations and the age distribution across the 12 urban areas. For the lung function part
   of the risk  assessment, spatial variability in air quality and population exposure inputs has
   been included in the assessment by use of a location specific exposure analysis and
   location specific input data to that analysis. For the epidemiology-based health
   endpoints,  spatial variability in key inputs has been embedded in the analysis by use of
   location specific inputs  (e.g., air quality, population data, baseline incidence data,
   concentration-response  relationships).

•  An important uncertainty to consider is the extent to which the associations between 63
   and the health endpoints included in the assessment actually reflect causal relationships.
   For lung function decrements, respiratory symptoms in moderate to severe asthmatic
   children, and respiratory-related hospital admissions there is clear and very strong
   evidence supporting the judgment that the relationships are causal. With respect to non-
   accidental  and cardiorespiratory mortality, there is greater uncertainty, with the CD
   concluding that the overall body of evidence is highly suggestive that Os directly or
   indirectly contributes to nonaccidental and cardiopulmonary-related mortality (CD,  p. 8-
   78).  Given the overall evidence, including the strong evidence from the time-series
   studies reporting associations with non-accidental and cardiorespiratory mortality, along
   with information about  potential mechanisms and a range of health effects observed in
   controlled  human exposure studies (i.e., increased lung inflammation, impacts on host
   defense, and increased airway responsiveness), and strong evidence showing respiratory-
   related hospital admissions and emergency department visits for asthmatics, the staff
   judges that there is a likely causal relationship between Os exposures and non-accidental
   and cardiorespiratory mortality.

•  Statistical uncertainty in the exposure-response relationships associated with sampling
   error has been characterized in the lung function part of the risk assessment. Other
                                    5-97

-------
important uncertainties in the exposure-response relationship for the lung function health
outcomes include:

       uncertainty associated with the shape of the exposure-response relationship,
       which also has been considered using a Bayesian Markov Chain Monte Carlo
       approach recommended by members of the CAS AC panel and extrapolation of
       the relationship to levels below 0.04 ppm, the lowest tested  level in controlled
       human exposure studies;
       uncertainty due to use of 6.6-hr data for subjects engaged in moderate exertion to
       estimate response associated with 8-hr exposures under moderate or greater
       exertion;
       uncertainty about whether O3-induced responses are reproducible, although this
       is generally supported by other controlled human exposure studies showing
       significant reproducibility of response;
       uncertainty introduced by use of exposure-response relationships based on  18 to
       35 year old subjects to represent the relationships for all and asthmatic school age
       children age 5 to  18, although the use of adult data is supported by a study testing
       8 to 11 year olds  and observations from  a number of summer camp field studies
       of school age children which found comparable responses to those observed in
       adults;
       uncertainty in the estimated exposure-response relationship  due to assumption
       that response on any given day is independent of previous 63 exposure; and
       uncertainty in the estimated exposure-response due to assumption that the
       response would not be affected by the presence of other co-pollutants.

Uncertainties related to estimating the  concentration-response relationships for the
epidemiological-based part of the risk assessment include:

     -  statistical uncertainty due to sampling error which is characterized in the
       assessment;
     -  model uncertainty (i.e., uncertainty  about the shape and magnitude of the
       concentration-response relationship taking into account lags, other
       pollutants, etc.); and
       uncertainty about whether a concentration-response function provides an
       accurate representation of the relationship in the location of interest
       because of a) the possible role of associated co-pollutants, b) variations in
       the relationship of total ambient exposure to ambient monitoring in
       different location, and c) differences in population  characteristics and
       population behavior patterns across locations or over time in the same
       location.

  Uncertainties related to the air quality data affect both the controlled human exposure
  studies-based and epidemiological studies-based parts of the risk  assessment and
  include:
    -  uncertainties associated with the air  quality adjustment procedure that was used to
       simulate just meeting the current and alternative 8-hr standards; and

                                5-98

-------
              uncertainties about estimated background concentrations for each location, for
              which sensitivity analyses were conducted examining the impact of alternative
              lower and higher assumed background concentrations.

    •   Based on the sensitivity analyses presented in the prior review, alternative air quality
adjustment procedure had only a modest impact on the risk estimates.  With respect to the
uncertainties about estimated background concentrations, as discussed in section  5.4.3,
alternative assumptions about PRB levels had a variable impact depending on the health effect
considered and the location and standard analyzed in terms of the absolute magnitude of the risk
estimates. There was relatively little impact on either absolute or relative changes in lung
function risk estimates due to alternative assumptions about PRB levels. With respect to Os-
related non-accidental mortality, alternative assumptions about background levels had a greater
impact. Estimates of risk remaining upon just meeting the current or alternative standards were
most affected, with differences of+/-50% or larger observed in many of the areas. Alternative
assumptions about PRB levels had a greater impact on the non-accidental mortality risk
estimates associated with more stringent 8-hr standards. While noteable differences were
observed for non-accidental mortality in some areas, particularly for more stringent standards,
the overall pattern of reductions, expressed in terms of percentage reduction relative to the
current standard, was significantly less impacted by alternative assumptions for PRB than the
absolute magnitude of the risks.
                                       5-99

-------
References

Abt Associates Inc. (2005a).  Ozone Health Risk Assessment for Selected Urban Areas: Draft Report.  Prepared for
      Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency, Research Triangle
      Park, NC. October 2005. Available electronically on the internet at:
      http://www.epa.gOv/ttn/naaqs/standards/ozone/s o3cr td.html.

Abt Associates Inc. (2005b).  Paniculate Matter Health Risk Assessment for Selected Urban Areas.  Prepared for
      Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency, Research Triangle
      Park, NC. June 2005.  Available electronically on the internet at:
      http://www.epa.gov/ttn/naaqs/standards/pm/s_pm  cr  td.html.

Abt Associates Inc. (2006). Ozone Health Risk Assessment for Selected Urban Areas. Prepared for Office of Air
      Quality Planning and Standards, U.S. Environmental Protection Agency, Research Triangle Park, NC.
      December 2006.  Available electronically on the internet at:
      http://www.epa.gOv/ttn/naaqs/standards/ozone/s o3cr td.html.

Adams, W.C. (2002). "Comparison of Chamber and Face-Mask 6.6-Hour Exposures to Ozone on Pulmonary
      Function and Symptoms Responses." Inhalation Toxicology  14:745-764.

Adams, W.C. (2003). "Comparison of Chamber and Face Mask 6.6-Hour Exposure to 0.08 ppm Ozone via Square-
      Wave and Triangular Profiles on Pulmonary Responses." Inhalation Toxicology 15: 265-281.

Adams, W.C. (2006). "Comparison of Chamber 6.6-h Exposures to 0.04-0.08 ppm Ozone via Square-Wave and
      Triangular Profiles on Pulmonary Responses." Inhalation Toxicology 18: 127-136.

Bell, M,A. McDermott,  S.L. Zeger, J.M. Samet, and F. Dominici (2004). "Ozone and short-term mortality in 95 US
      urban communities, 1987-2000." JAMA292(19):2372-2378.

Bell, M,A. R.D. Peng, and F. Dominici (2006).  "The Exposure-Response Curve for Ozone and Risk of Mortality
      and the Adequacy of Current Ozone Regulations." Environmental Health Perspectives.  Available online at:
      http://dx.doi.org/

CDC, 2005. United States Department of Health and Human Services (US DHHS), Centers for Disease Control and
      Prevention (CDC), National Center for Health Statistics (NCHS), Compressed Mortality File (CMF)
      compiled from CMF 1968-1988, Series 20, No. 2A 2000, CMF 1989-1998, Series 20, No. 2E 2003 and CMF
      1999-2002, Series 20, No. 2H 2004 on CDC WONDER On-line Database.  See http://wonder.cdc.gov/.

DuMouchel, W. (1994).  "Hierarchical Bayes Linear Models for Meta-Analysis." Technical Report #27.
      September, 1994.  National Institute of Statistical Sciences, P. O. Box 14162, Research Triangle Park, N.C.
      27709.

Fiore, A.M., DJ. Jacob, I. Bey, R.M. Yantosca, B.D. Field, A.C. Fusco, and J.G. Wilkinson (2002a). "Background
      ozone over the United States in summer:  Origin, trend, and contribution to pollution episodes." J. Geophys.
      Res., 107(015), 4275.

Fiore, A.M., D.J. Jacob, B.D. Field, D.G. Streets, S.D. Fernandes, and C. Jang (2002b). "Linking ozone pollution
      with climate change: The case for controlling methane." Geophys. Res. Lett., 29(19), 1919.

Fiore, A.M., D.J. Jacob, H. Liu, R.M. Yantosca, T.D. Fairlie, and Q. Li (2003). "Variability in surface ozone
      background over the United States: Implications for air quality policy."  Journal Of Geophysical Research
      Vol. 108(D24), 4787.
                                            5-100

-------
Fitz-Simons, T., L. McCluney, and M. Rizzo (2005). OAQPS Staff Memorandum to Ozone NAAQS Review
      Docket (OAR-2005-1072). Subject:  Analysis of 2004 Ozone Data for the Ozone NAAQS Review.
      November 7, 2005. Available at: http://www.epa.gOv/ttn/naaqs/standards/ozone/s o3crtd.html

Friedman, M.S., K.E. Powell, L. Hutwagner, L.M. Graham, and W.G. Teague (2001).  "Impact of changes in
      transportation and commuting behaviors during the 1996 summer Olympic games in Atlanta on air quality
      and childhood asthma." JAMA 285:897-905.

Fusco, A. C., and J. A. Logan (2003). "Analysis of 1970- 1995 trends in tropospheric  ozone at Northern
      Hemisphere midlatitudes with the GEOSCHEM model." J. Geophys. Res.,  108(D15), 4449.

Gent, J.F., E.W. Triche, T.R. Holford, K. Belanger, M.B. Bracken, W.S. Beckett, B.P.  Leaderer (2003).
      "Association of low-level ozone and fine particles with respiratory symptoms in children with asthma."
      JAMA 290(14): 1859-1867.

Graham, S. and T. McCurdy (2004). "Developing meaningful cohorts for human exposure models." Journal of
      Exposure Analysis and Environmental Epidemiology 14:23-43.

Gilliland, F.D., K. Berhane, E.B. Rappaport, D.C. Thomas, E. Avol, W.J. Gauderman,  S.J. London, H.G. Margolis,
      R. McConnell, K.T. Islam and J.M. Peters (2001). "The effects of ambient air pollution on school
      absenteeism due to respiratory illnesses." Epidemiology 12(l):43-54.

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.

Henderson, R. (2006a).  Letter from CAS AC Chairman Rogene Henderson to EPA Administrator Stephen Johnson,
      June 5, 2006, EPA-CASAC-CON-06-003

Henderson, R. (2006b)  Letter from CASAC Chairman Rogene Henderson to EPA Administrator Stephen Johnson,
      June 5, 2006, EPA-CASAC-06-007.

Henderson, R. (2006c)  Letter from CASAC Chairman Rogene Henderson to EPA Administrator Stephen Johnson,
      October 24, 2006, EPA-CASAC-07-001.

Horstman, D.H. et al. (1990). "Ozone concentration and pulmonary response relationships for 6.6-hr exposures
      with five hours of moderate exercise to 0.08,  0.10, and 0.12 ppm."  American Review of Respiratory Disease
      142:1158-1163.

Huang, Y., F. Dominici, M.L. Bell (2004). "Bayesian hierarchical distributed lag  models for summer ozone
      exposure and cardio-respiratory mortality." John Hopkins University, Department of Biostatistics Working
      Paper. 46.

Ito, K. (2003). Associations of paniculate matter components with daily mortality and  morbidity  in Detroit,
      Michigan. In: "Revised Analyses of Time-Series Studies of Air Pollution and Health," Health Effects
      Institute Special Report, May.

Jaffe, D.H., Singer, M.E., and Rimm, A.A. (2003). "Air pollution and emergency department visits for asthma
      among Ohio Medicaid recipients, 1991-1996."  Environmental Research 91:21-28.

Johnson, T., Capel, J., and McCoy, M. (1996a). Estimation of Ozone Exposures Experienced by Urban Residents
      Using a Probabilistic Version of NEM and 1990 Population Data. Prepared by International Technology Air
      Quality Services for Office of Air Quallity Planning and Standards, EPA, Research Triangle Park, NC.
      Available electronically on the internet at: http://www.epa.gOv/ttn/naaqs/standards/ozone/s o3_prtd.html.

Johnson, T., Capel, J.,  McCoy, M., and Warnasch, J. (1996b). Estimation of Ozone Exposures Experienced by
      Outdoor Children in Nine Urban Areas Using a Probabilistic Version of NEM.  Prepared by International

                                            5-101

-------
      Technology Air Quality Services for Office of Air Quallity Planning and Standards, EPA, Research Triangle
      Park, NC. Available electronically on the internet at:
http://www.epa.gOv/ttn/naaqs/standards/ozone/s  o3_prtd.html.

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.

Kann, L. et al. (2000).  "Youth risk behavior surveillance-United States, 1999." Mortality and Morbidity Weekly
      Report 49(SS05): 1-96.

Langstaff, J. E. (2006). OAQPS Staff Memorandum to Ozone NAAQS Review Docket (OAR-2005-0172).
      Subject: Analysis of uncertainty in ozone population exposure modeling. July 17. Available at:
      http://www.epa.gOv/ttn/naaqs/standards/ozone/s o3crtd.html.

Linn, W., Y. Szlachcis, H. J. Gong, P. Kinney, K. Berhane (2000).  "Air pollution and daily hospital admissions in
      metropolitan Los Angeles." Environmental Health Perspective 108:427-434.

McCurdy, T. (2000). "Conceptual basis for multi-route intake dose modeling using an energy expenditure
      approach." J. Exposure Anal. Environ. Epidemiol. 10:1-12.

McCurdy, T., Glen, G., Smith, L., Lakkadi, Y. (2000). "The National Exposure Research Laboratory's Consolidated
      Human Activity Database." J. Exposure Anal. Environ. Epidemiol. 10:566-578.

McDonnell, W.F. et al. (1985a).  "Reproducibility of individual responses to ozone exposure." American Review of
      Respiratory Disease 131:36-40.

McDonnell, W. F., Ill; Chapman, R. S.; Leigh, M. W.; Strope, G. L.; Collier, A. M. (1985b) "Respiratory responses
      of vigorously exercising children to  0.12 ppm ozone exposure." Am. Rev. Respir. Dis.  132: 875-879.

McDonnell, W.F. et al. (1991). "Respiratory response of humans exposed to low levels of ozone for 6.6 hours."
      American Review of Respiratory Disease  147:804-810.

Moolgavkar, S. H.; Luebeck, E. G.; Hall, T. A.; Anderson, E. L. (1995).  "Air pollution and daily mortality in
      Philadelphia." Epidemiology 6: 476-484.

Morgan and Henrion (1990).  Uncertainty:  A Guide To Dealing with Uncertainty in Qualitative Risk and Policy
      Analysis.  Cambridge University Press.

Mortimer, K.M., L.M. Neas, D.W. Dockery, S. Redline, and I.E. Tager (2002). "The effects on air pollution on
      inner-city children with asthma." European Respiratory Journal. 19:699-705.

Peel, J.L.,  P.E. Tolbert, M. Klein, K.B. Metzger,  W.D. Flanders, K. Todd, J.M. Mulholland, P.B. Ryan, and H.
      Frumkin,  (2005).  "Ambient air pollution and respiratory emergency department visits."  Epidemiology
      16(2):164-174.

Pope, C. A., R. T. Burnett, M. J.  Thun, E. E. Calle, D. Krewski, K. Ito, and G. D.  Thurston.  2002.  Lung Cancer,
      Cardiopulmonary Mortality, and Long-term Exposure to Fine Paniculate Air Pollution. Journal of the
      American Medical Association, vol  287, no 9: 287:1132-1141.

Post, E., D. Hoaglin, L. Deck, and K. Larntz (2001). "An Empirical Bayes approach to estimating the relation of
      mortality to exposure to paniculate matter," Risk Analysis 21(5):  837-842.

Post, E.  (2007).  Memorandum - Subject: Additional Tables of Lung Function Risk Estimates Associated with
      Ozone Exposures Based on 2003 Air Quality Data. January 30. Available at:
      http: //www. epa. gov/ttn/naaqs/standards/ozone/s_o 3 _cr_td. html.

                                            5-102

-------
Richmond H., T. Palma, J. Langstaff, T. McCurdy, G. Glenn, and L. Smith (2002). "Further refinements and testing
      of APEX (3.0): EPA's population exposure model for criteria and air toxic inhalation exposures." Poster
      presentation.  Joint meeting of the International Society of Exposure Analysis and International Society of
      Environmental Epidemiology, August 11-15, 2002, Vancouver, Canada.

Rizzo, M. (2005).  OAQPS Staff Memorandum to Ozone NAAQS Review Docket (OAR-2005-1072).  Subject:
      Evaluation of a quadratic approach for adjusting distributions of hourly ozone concentrations to meet air
      quality standards.  November 7. Available at:
      http: //www. epa. gov/ttn/naaqs/standards/ozone/s_o 3 _cr_td. html.


Rizzo, M. (2006).  OAQPS Staff Memorandum to Ozone NAAQS Review Docket (OAR-2005-1072).  Subject: A
      Comparison between Different Rollback Methodologies Applied to Ambient Ozone Concentrations.  May 31.
      Available at:  http://www.epa.gOv/ttn/naaqs/standards/ozone/s o3crtd.html.

Schwartz,!. (2000).  "The distributed lag between air pollution and daily deaths."  Epidemiology ll(3):320-326.

Schwartz, J. (2004).  "How sensitive is the association between ozone and daily deaths to control for temperature?"
      Am. J. Resp. Crit.  Care Med.

Schwartz, J., C. Spix, G. Touloumi, L. Bacharova, T. Barumamdzadeh, A. le Tertre, T. Piekarksi, A. Ponce de Leon,
      A. Ponka, G. Rossi, M. Saez, J.P. Schouten (1996). "Methodological issues in studies of air pollution and
      daily counts of deaths or hospital admissions." J. Epid. andComm. Health 50(Suppl 1):S3-S11.

Thurston, G.D., K. Ito, P.L. Kinney, M. Lippmann (1992).  "A multi-year study of air pollution and respiratory
      hospital admission in three New York State metropolitan areas: Results for 1988 and 1989 summers." J.
      Exposure Anal. Environ. Epidemiol.  2(4):429-450.

Thurston, G.D. and Ito, K. (2001).  "Epidemiological studies of acute ozone exposures and mortality." J. Exposure
      Anal. Environ. Epidemiol. 11:286.

Tolbert, P.E., J.A. Mulholland, D.L. Macintosh, F. Xu, et al. (2000).  "Air quality and Pediatric Emergency Room
      Visits for Asthma in Atlanta, GA, USA."  American Journal of Epidemiology 151(8):798-810.

U. S. EPA (1996a). 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. Available electronically on the internet at:
      http://www.epa.gov/ttn/naaqs/standards/ozone/s_o3_pr_sp.html.

U.S. EPA (1996b).  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. Available electronically on the internet at:
      http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=2831.

U.S. EPA (2002).  Consolidated Human Activities Database Users Guide. The database and documentation are
      available electronically on the  internet at: http://www.epa.gov/chadnetl/.

U.S. EPA (2003).  Total Risk Integrated Methodology TRIM.Expo/Inhalation User's Document
Volume I:  Air Pollutants Exposure  Model (APEX, version 3) User's Guide. Office of Air Quality Planning and
      Standards, Research Triangle Park, NC. Available electronically on the internet at:
http://www.epa.gov/ttn/fera/human apex.html.

U.S. EPA (2004).  Air Quality Criteria for Paniculate Matter.  EPA 600/P-99/002bF, 2v. National Center for
      Environmental Assessment, Research Triangle Park, NC.  Available electronically on the internet at:
      http://www.epa.gOv/ttn/naaqs/standards/pm/s pm cr cd.html
                                            5-103

-------
U.S. EPA (2005a). Health Assessment Plan for Review of the National Ambient Air Quality Standards for Ozone.
Office of Air Quality Planning and Standards, Research Triangle Park, NC. April. Available electronically on the
      internet at  http://www.epa.gOv/ttn/naaqs/standards/ozone/s  o3 cr_pd.html.

U.S. EPA (2005b). Review of National Ambient Air Quality Standards for Paniculate Matter: Assessment of
      Scientific and Technical Information - OAQPS Staff Paper.  EPA-452/D-05-001. Office of Air Quality
      Planning and Standards, Research Triangle Park, NC.  Available electronically on the internet at:
      http://www.epa.gov/ttn/naaqs/standards/pm/s_pm cr  sp.html.

U. S. EPA (2006a).  Air Quality Criteria for Ozone and Other Related Photochemical Oxidants. National Center for
      Environmental Assessment, Research Triangle Park, NC. Available electronically on the internet at:
      http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=137307

U. S. EPA (2006b).  Ozone Population Exposure Analysis for Selected Urban Areas. Office of Air Quality
      Planning and Standards, Research Triangle Park, NC.  December.  Available electronically on the internet at:
      http://www.epa.gOv/ttn/naaqs/standards/ozone/s o3crtd.html.

Whitfield, R., Biller, W., Jusko, M, and Keisler, J. (1996). A Probabilistic Assessment of Health Risks Associated
      with Short- and Long-Term Exposure to Tropospheric Ozone.  Argonne National Laboratory, Argonne, IL.

Whitfield, R. (1997). A Probabilistic Assessment of Health Risks Associated with Short-term Exposure to
      Tropospheric Ozone: A Supplement.  Argonne National Laboratory, Argonne, IL.

Whitfield, R.G., Richmond, H.M. and Johnson, T.R. (1998).  "Overview of Ozone Human Exposure and Health
      Risk Analyses Used in the U.S. EPA's Review of the Ozone Air Quality Standard," pp.483-516 in: T.
      Schneider, ed. Air Pollution in the 21st Century: Priority Issues and Policy Elsevier; Amsterdam.
                                             5-104

-------
    6.     STAFF CONCLUSIONS AND RECOMMENDATIONS ON THE
                               PRIMARY O3 NAAQS
6.1    INTRODUCTION
       This chapter presents staff conclusions and recommendations for the Administrator to
consider in deciding whether the existing primary O?, standard should be revised and, if so, what
revised standard is appropriate. Our conclusions and recommendations are based on the
assessment and integrative synthesis of information presented in the CD, staff analyses and
evaluations presented in Chapters 2 through 5 herein, and the comments and advice of CAS AC
and interested parties who commented on earlier drafts of this document and related technical
support documents.
       In recommending  policy options for the Administrator's consideration, we note that the
final decision on retaining or revising the current O3 standard is largely a public health policy
judgment. A final decision should draw upon scientific information and analyses about health
effects, population exposure and risks, as well as judgments about the appropriate response to the
range of uncertainties that are inherent in the scientific evidence and analyses. Our approach to
informing these judgments, discussed more fully below, is based on a recognition that the
available health effects evidence generally reflects a continuum consisting of ambient levels at
which scientists generally agree that health effects are likely to occur through lower levels at
which the likelihood and magnitude of the response become increasingly uncertain.
       This approach is consistent with the requirements of the NAAQS provisions of the Act
and with how EPA and the courts have historically interpreted the Act.  These provisions require
the Administrator to establish primary standards that, in the Administrator's judgment, are
requisite to protect public health with an adequate margin of safety.  In so doing, the
Administrator seeks to establish standards  that are neither more nor less stringent than necessary
for this purpose. The Act does not require that primary standards be set at a zero-risk level but
rather at a level that avoids unacceptable risks to public health, including the health of sensitive
groups.

6.2    APPROACH
       To evaluate whether it is appropriate to consider retaining the current primary Os
standard, or whether consideration of revisions is appropriate, we adopted an approach in this
review that builds upon the general approach used in the last review and reflects the broader
body of evidence now available. The 1997 final decision notice (62 FR 38856) outlined the key
factors considered in selecting the elements of a standard for Os (judged to  be the most
                                           6-1

-------
appropriate indicator for photochemical oxidants):  the averaging time; the level (i.e., an ambient
Os concentration); and the form (i.e., the air quality statistic used as a basis for determining
compliance with the standard). Decisions on these elements were based on an integration of
information on acute and chronic health effects associated with exposure to ambient 63; expert
judgments on the adversity of such effects on individuals; and policy judgments as to when the
standard is requisite to protect public health with an adequate margin of safety, which were
informed by air quality analysis and quantitative exposure and risk assessments when possible,
as well as qualitative assessment of public health impacts that could not be quantified.
       As in the last review, in developing conclusions and recommendations on the primary Os
standard in this review, staff has  taken into account both evidence-based and quantitative
exposure- and risk-based considerations.  Evidence-based considerations include the assessment
of evidence from controlled human exposure, animal lexicological, field, and epidemiological
studies for a variety of health endpoints.  For those endpoints based on epidemiological studies,
we have placed greater weight on associations with health endpoints that the CD has judged to be
likely causal based on an integrative synthesis of the entire body of evidence, including not only
all available epidemiological evidence but also evidence from animal toxicological and
controlled human exposure studies.  Less weight has been placed on evidence of associations
that were judged to be only suggestive of possible causal relationships. For the purpose of
evaluating the level of the Os standard in this review, we have placed greater weight on U.S.  and
Canadian studies, taking into account the extent to which such studies have reported statistically
significant associations. This is because findings of U.S. and Canadian studies are more directly
applicable for quantitative considerations in this review as studies conducted in other countries
may well reflect quite different populations, exposure characteristics, and air pollution mixtures.
       Staffs consideration of quantitative exposure- and risk-based information draws from the
results of the exposure and risk assessments conducted for as many as twelve urban areas in the
U.S. (discussed in Chapters 4 and 5). More specifically, we have considered estimates of the
magnitude of Os-related exposures and risks associated with recent air quality levels, as well as
the exposure  and risk reductions  likely to be associated with just meeting the current 8-hr
primary Os NAAQS and potential alternative 8-hr standards. We recognize the uncertainties
inherent in such estimates, which are discussed in Chapters 4 and 5, in part by providing where
possible some sense of the direction and/or magnitude of the uncertainties which should be taken
into account as one considers these estimates.
       In this review, a series of general questions frames our approach to reaching conclusions
and recommendations in section  6.3 below, based on the available evidence and information, as
to whether consideration should be given to retaining or revising the current primary Os
                                            6-2

-------
standard.  Our review of the adequacy of the current standard in section 6.3.1 begins by
addressing questions such as the following:

      •   To what extent does newly available information reinforce or call into question
         evidence of associations of Os exposures with effects identified in the last review?

      •   To what extent has evidence of new effects and/or sensitive populations become
         available since the last review?

      •   To what extent have important uncertainties identified in the last review been reduced
         and have new uncertainties emerged?

      •   To what extent does newly available information reinforce or call into question any of
         the basic elements of the current standards?
To the extent that the available information  suggests that revision of the current standard may be
appropriate to consider, we also explore whether the currently available information supports
consideration of a standard that is either more or less protective by addressing the following
questions:

      •   Is there evidence that associations, especially likely causal associations, extend to
         ambient Os concentration levels that are as low as or lower than had previously been
         observed, and what are the important uncertainties associated with that evidence?

      •   Are exposures of concern and health risks estimated to occur in areas upon meeting the
         current standard; are they important from a public health perspective; and what are the
         important uncertainties associated with the estimated risks?
       To the extent that there is support for consideration of a revised standard, we then
consider the specific elements of the standard (in terms of an indicator for photochemical
oxidants, averaging time, level, and form in sections 6.3.2 through 6.3.5, respectively) and
identify policy options that we conclude would be appropriate for the Administrator to consider
in making public health policy judgments, based on the currently available information, as to the
degree of protection that is  requisite to protect public health with an adequate margin of safety.
In so doing, we address the following questions:

      •   Does the evidence provide support for considering a different Os indicator?

      •   Does the evidence provide support for considering different averaging times?

      •   What ranges of levels and forms of alternative standards are supported by the evidence,
         and what are the uncertainties and limitations in that evidence?

      •   To what extent do specific levels and forms  of alternative standards reduce the
         estimated exposures of concern and risks attributable to Os and other photochemical
                                            6-3

-------
         oxidants, and what are the uncertainties associated with the estimated exposure and risk
         reductions?
       A summary of staff conclusions and recommendations on policy options for the
Administrator's consideration concerning whether and, if so, how to revise the current primary
Os standard is presented in section 6.3.6.  This chapter concludes with a discussion of key
uncertainties and recommendations for additional research related to setting a primary Os
standard in section 6.4.

6.3    PRIMARY O3 STANDARD
       The current primary Os standard is an 8-hr standard set at a level of 0.08 ppm, with a
form of the annual fourth-highest daily maximum 8-hr average concentration, averaged over
three years, to provide protection to the public, especially children and other at-risk populations,
against a wide range of O3-induced effects.  As an introduction to the discussion in this section of
the adequacy of the current Os standard and potential options for alternative standards, it is
useful  to summarize the key factors that formed the basis of the decision in the last review to
revise  the averaging time, level, and form of the then current 1-hr standard.
       In the last review, the key factor in deciding to revise the averaging time of the primary
standard was evidence from controlled human exposure studies of healthy young adult subjects
exposed for 1 to 8 hr to Os. The best documented health endpoints in these studies were
decrements in indices of lung function, such as forced expiratory volume in 1  second (FEVi),
and respiratory symptoms, such as cough and chest pain on deep inspiration. For short-term
exposures of 1 to 3 hr, group mean FEVi decrements were statistically significant for Os
concentrations only at and above 0.12  ppm, and only when subjects engaged in very heavy
exertion. By contrast, prolonged exposures of 6 to 8 hr produced statistically significant group
mean FEVi decrements at the lowest Os concentrations evaluated in those studies, 0.08 ppm,
even when experimental subjects were engaged in more realistic intermittent moderate exertion
levels. The health significance of this  newer evidence led to the conclusion in the 1997 final
decision that the 8-hr averaging time is more directly associated with health effects of concern at
lower Os concentrations than is the 1-hr averaging time.
        Based on the available evidence of Os-related health effects, the following factors were
of particular importance in the last review in informing the selection of the level and form of a
new 8-hr standard:  (1) quantitative estimates of O3-related risks to active children, who were
judged to be a sensitive subgroup of concern, in terms of transient and reversible respiratory
effects judged to be adverse, including moderate to large decreases in lung function and
moderate to severe pain on deep inspiration, and the uncertainty and variability in such
estimates; (2) consideration of both the estimated percentages, total numbers of children, and
                                            6-4

-------
number of times they were likely to experience such effects; (3) epidemiological evidence of
associations between ambient Os and increased respiratory hospital admissions, and quantitative
estimates of percentages and total numbers of asthma-related admissions in one example urban
area that were judged to be indicative of a pyramid of much larger effects, including respiratory-
related hospital admissions, emergency department (ED) visits, doctor visits, and asthma attacks
and related increased medication use; (4) quantitative estimates of the number of "exposures of
concern" (defined as exposures > 0.08 ppm for 6 to 8 hr) that active children are likely to
experience, and the uncertainty and variability in such estimates; (5) the judgment that such
exposures are an important indicator of public health impacts of (Vrelated effects for which
information is too limited to develop quantitative risk estimates, including increased nonspecific
bronchial responsiveness (e.g., related to aggravation of asthma), decreased pulmonary defense
mechanisms  (suggestive of increased susceptibility to respiratory infection), and indicators of
pulmonary inflammation (related to potential aggravation of chronic bronchitis or long-term
damage to the lungs); (6) the broader public health perspective of the number of people living in
areas that would breathe cleaner air as a result of the revised standard; (7) consideration of the
relative seriousness of various health effects and the relative degree of certainty in both the
likelihood that people will experience various health effects and their medical significance; (8)
the relationship of a standard level to estimated "background" levels associated with
nonanthropogenic sources of O^; and (9) CASAC advice and recommendations. Additional
factors that were considered in selecting the form of the standard included the public health
implications of the expected number of times in an Os season that the standard level might be
exceeded in an area that is in attainment with the standard and the year-to-year stability of the air
quality statistic, so as to avoid disruptions to ongoing control programs which could interrupt
public health protections.
       In reaching a final decision  in the last review, the Administrator was mindful that Os
exhibits a continuum of effects, such that there is no discernible threshold above which public
health protection requires that no exposures be allowed or below which all risks to public health
can be avoided.  The final decision reflected a recognition that important uncertainties remained,
for example with regard to interpreting the role of other pollutants co-occurring with Os in
observed associations, understanding biological mechanisms of Os-related health effects, and
estimating human exposures and quantitative risks to at-risk populations for these health effects.

       6.3.1  Adequacy of Current Os Standard
       Overall, the new evidence available in this review generally supports and builds further
upon key health-related conclusions drawn in the previous review. New human clinical studies
provide information about lung function and respiratory symptom responses to prolonged
                                            6-5

-------
exposures at 63 levels at and below 0.08 ppm. There is an expanded body of evidence about the
mechanisms of respiratory effects, including important new evidence about increased
susceptibility of people with asthma and limited new evidence about plausible mechanisms by
which 63 exposure could induce effects on the cardiovascular system. In this review, there is
additional epidemiological evidence supporting associations between Os exposure and
respiratory symptoms in asthmatic children, ED visits and hospital admissions for respiratory
causes, and new evidence that links O3 exposure to premature mortality.
       As discussed in Chapter 3, the CD concludes that, based on the extensive body of human
clinical, lexicological, and epidemiological evidence, there is a causal relationship between
short-term Os exposure and a range of respiratory morbidity effects, including lung function
decrements, increased respiratory symptoms, airway inflammation, and increased airway
responsiveness. Aggregate population time-series studies provide evidence that ambient 63
concentrations are positively and robustly associated with respiratory-related hospitalizations and
ED visits during the warm season. The CD concludes that the overall body of evidence supports
a causal relationship between acute ambient O3 exposures and these respiratory morbidity
outcomes (CD, p. 8-77). Based on the evidence from animal toxicology, human clinical, and
epidemiological studies, the CD concludes that a generally limited body of evidence provides
considerable plausibility for cardiovascular mechanisms and is highly suggestive that Os can
directly or indirectly contribute to cardiovascular-related morbidity,  but that much needs to be
done to more fully substantiate links between ambient 63 exposures and adverse cardiovascular
outcomes (CD, p. 8-77-78).  The CD also finds that, consistent with observed Os-related
increases in respiratory- and cardiovascular-related morbidity, several newer U.S. multi-city
time-series studies, single-city  studies, and several meta-analyses of these studies, provide
relatively strong epidemiological evidence for associations between short-term Os exposure and
all-cause (non-accidental) mortality, even after adjustment for the influence of season and PM
(CD, p. 8-78).
       In considering this evidence as a basis for evaluating the adequacy of the current Os
standard, we recognize that important uncertainties remain.  For example, as discussed above in
section 3.4, we note that inherent limitations in time-series epidemiological studies raise
questions about the utility of such evidence to inform judgments about a NAAQS for an
individual pollutant such as 63 within a mix of highly correlated pollutants, such as the mix of
photochemical oxidants, especially at ambient Os concentrations below levels at which Os-
related effects have been observed in controlled human exposure studies.  We also recognize that
the available epidemiological evidence neither supports nor refutes the existence of thresholds at
the population level for effects such  as increased hospital admissions and premature mortality.
There are limitations in epidemiological studies that make discerning thresholds in populations
                                            6-6

-------
difficult, including low data density in the lower concentration ranges, the possible influence of
exposure measurement error, and interindividual differences in susceptibility to Ch-related
effects in populations.
       While there clearly are limitations in interpreting the findings from the epidemiological
studies, as noted above, we conclude for the following reasons that if a population threshold
level does exist, it would likely be well below the level of the current 63 standard and possibly
within the range of background levels. This conclusion is supported by the discussions in
Chapter 3 above (section 3.4.5)  and more fully in the CD (Chapter 7, section 7.6.5) of the
several epidemiological studies that have explored the question of potential thresholds and of the
studies that included analyses excluding days over 0.08 ppm or even lower Os levels. We note
that seasonal epidemiological studies show no consistent Os-related effects during the cold
season when 63 concentrations are generally low, in contrast to a pattern of generally positive
and statistically significant Os-related effects in the warm season when Os concentrations are
generally appreciably higher.  In addition to direct consideration of the epidemiological studies,
the findings from controlled human exposure studies strongly suggest that both lung function
decrements and respiratory symptoms occur in healthy adult subjects at levels down to at least
0.060 ppm (+/- 0.003 ppm) with some subjects experiencing notable effects (e.g., >10% FEVi
decrement, pain on deep inspiration). Controlled human exposure studies also found significant
responses in indicators of lung inflammation and cell injury at 0.080 ppm (+/- 0.004 ppm) in
healthy adult subjects.  These effects were observed in healthy young adult subjects and it is
likely that greater responses and responses at lower levels would occur in people with asthma
and other respiratory diseases. As discussed in Chapter 3, the physiological effects observed in
controlled human exposure studies have been linked to aggravation  of asthma and increased
susceptibility to respiratory infection, potentially leading to increased medication use, increased
school and work absences, increased visits to doctors'  offices and emergency departments (EDs),
and increased hospital admissions.  These observations provide additional support for the
conclusion that the associations observed in the epidemiological  studies, particularly for
respiratory-related effects and potentially for cardiovascular effects, extend down to ozone levels
well below 0.084 ppm.
       Based on the above considerations and findings from the CD, while being mindful of
important remaining uncertainties, staff concludes that the newly available information generally
reinforces our judgments about causal relationships between Os exposure and respiratory effects
observed in the last review and broadens the evidence of (Vrelated  associations to include
additional respiratory-related endpoints,  newly identified cardiovascular-related health
endpoints, and mortality.  Newly available evidence also has identified increased susceptibility in
people with asthma.  While recognizing that important uncertainties and research questions
                                            6-7

-------
remain, we also conclude that progress has been made since the last review in advancing the
understanding of potential mechanisms by which ambient Os, alone and in combination with
other pollutants, is causally linked to a range of respiratory- and cardiovascular-related health
endpoints.  Thus, we generally find support in the available evidence, including the direction of
the evidence developed since the last review, for consideration of an Os standard that is at least
as protective as the current standard and do not find support for consideration of an 63 standard
that is less protective than the current standard.  This general conclusion is consistent with the
advice and recommendations of CAS AC and with the views expressed by all interested parties
who provided comments on the previous draft of this document.  While CASAC and some
commenters supported revising the current standard to provide increased public health
protection, and other commenters supported retaining the current standard, no one who provided
comments supported a standard that would be less protective than the current standard.
       Having reached this general conclusion, we discuss in greater detail below the available
evidence (section 6.3.1.1) and exposure- and risk-based considerations (section 6.3.1.2) to more
fully inform consideration of the adequacy of the current standard.  We also take into account the
views expressed by CASAC and public commenters (section 6.3.1.3) in reaching  staff
conclusions on the adequacy of the current standard (section 6.3.1.4).

       6.3.1.1  Evidence-based Considerations
       In looking more specifically at the controlled human exposure and epidemiological
evidence summarized in Chapter 3 and Appendix 3B, staff first notes that controlled human
exposure studies provide the clearest and most compelling evidence for an array of human health
effects that are directly attributable to acute exposures to Os per se (CD, p. 8-73).  We also note
that evidence from such human studies, together with animal toxicological studies, help to
provide biological  plausibility for health effects observed in epidemiological studies.  In
considering the available evidence, we have focused on studies that examined health effects that
have been demonstrated to be caused by exposure to 63  or for which the CD judges associations
with O3 to be causal or likely causal.  In considering the epidemiological evidence as a basis for
reaching  conclusions about the adequacy of the current standard, we have focused on studies
reporting effects in the warm season, for which the effect estimates are more consistently
positive and statistically significant than those from all-year studies. We have considered the
extent to  which such studies provide evidence of associations that extend down to ambient Os
concentrations below the level of the current standard, which would call into question the
adequacy of the current standard. In so doing, we note, as discussed above,  that if a population
threshold level does exist for an effect observed in such studies, it would likely be at a level well
below the level of the current standard.  We have also sought to characterize whether the area in
                                            6-8

-------
which a study was conducted likely would or would not have met the current standard during the
time of the study, although we recognize that the confidence that would appropriately be placed
on the associations observed in any given study, or on the extent to which the association would
likely extend down to relatively low 63 concentrations, is not dependent on this distinction.
Further, we have considered studies that examined subsets of data that include only days with
ambient 63 concentrations below the level of the current 63 standard, or below even lower 63
concentrations, and continue to report statistically significant associations. We judge that such
studies are directly relevant to considering the adequacy of the current standard, particularly in
light of reported responses to 63 at levels below the current standard found in controlled human
exposure studies.
       In examining air quality information from the epidemiological studies for the purpose of
determining whether they were conducted in areas that likely would or would not have met the
current standard, we note that it is difficult to consistently characterize relevant air quality
statistics.1  These difficulties arise in  particular in panel studies of lung function or respiratory
symptoms in which the study periods were often shorter than a complete O3  season; Appendix
3B includes 98th and 99th percentile values as a way to approximate the fourth-highest value for
studies with differing study periods.  Difficulties also arise in all studies in which the air quality
data were averaged across multiple monitors in a study area (as are reported in Appendix 3B),
since an area's attainment status is determined by the monitor measuring the highest Os
concentrations in  an area, not averaged across monitors.  For studies with relatively low air
quality values that are based on averaging across multiple monitors, we  have further explored the
available air quality data so as to help inform a comparison with the level of the current standard,
as discussed below.
       Lung Function, Respiratory Symptoms,  and Other Respiratory Effects
       Health effects for which the CD continues to find clear evidence of causal associations
with short-term Os exposures include lung function decrements, respiratory symptoms,
pulmonary inflammation, and increased airway responsiveness.  In the last review, these (V
induced effects were demonstrated with statistical significance down to the lowest level tested in
controlled human exposure studies at that time (i.e., 0.08 ppm).  As discussed in Chapter 3
(section 3.3.1.1.1), in new controlled human  exposure studies, healthy adult volunteers were
exposed to 6.6-hr average Os levels down to  lower levels (i.e., 0.04 and  0.06 ppm) while engaged
in moderate exertion.  These studies did not report statistically significant changes in the group
mean FEVi decrements between lung function decrements associated with the 0.06 ppm or 0.04
       Determining attainment with the current standard is based on the 3-year average of the annual (over an O3
season) fourth-highest daily maximum 8-hr average O3 concentration at each monitor in an area.
                                            6-9

-------
ppm levels versus the filtered air (i.e., essentially 0 ppm 63) exposure when lung function
changes were analyzed for each hourly interval (i.e., after 1, 2, 3, 4.6, 5.6, and 6.6 hours of
exposure). However, as discussed in Section 3.3.1.1.1, we note that the pre- and post-exposure
data presented in the Adams (2006) study show a small (< 3%) group mean FEVi decrement
following the 6.6-hr exposure at 0.06 ppm, which may be statistically significantly different from
filtered air responses.  Notably, total respiratory symptoms (which includes pain on deep
inspiration, shortness of breath, and cough) following 5.6 and 6.6 h exposures at 0.06 ppm
(during a triangular exposure pattern, that is more representative of those encountered in summer
air pollution episodes than a square-wave exposure pattern) reached statistical significance. In
addition to information about group mean decrements, this study also reports that a small
percentage (7%) of healthy adult subjects experienced moderate lung function decrements (>
10% FEVi) with exposure to 0.06 and 0.04 ppm Os when corrected for the effects of exercise in
clean air. The distribution of individual responses related to lung function decrements (> 10,  15,
and 20% FEVi) found in these new studies are considered as part of the quantitative risk
assessment for lung  function responses in children discussed below.
      Newer information indicates that people with moderate-to-severe asthma have somewhat
larger decreases in lung function in response to 63 relative to healthy individuals and that lung
function responses in people with asthma appear to be affected by baseline lung function (i.e.,
responses increase with increasing disease severity, CD, p. 8-80). As discussed in the CD
(Chapter 6, sections 6.8 and 6.9; Chapter 8, sections 8.7 and 8.8) this newer information expands
our understanding of the physiological basis for increased sensitivity in people with asthma and
other airway diseases, recognizing that people with asthma present a different response profile
for cellular, molecular, and biochemical responses than people who do not have asthma. New
evidence indicates that some people with asthma have increased occurrence and duration of
nonspecific airway responsiveness, which is an increased bronchoconstrictive response to airway
irritants.  Controlled-human exposure studies also indicate that some people with allergic asthma
and rhinitis have increased airway responsiveness to allergens  following Os exposure.
Exposures to 63 exacerbated lung function decrements in people with pre-existing allergic
airway disease, with and without asthma. Ozone-induced exacerbation of airway responsiveness
persists longer and attenuates more slowly than Os-induced lung function decrements and
respiratory symptom responses and can have important clinical implications for asthmatics.
      Newly available human exposure studies suggest that some people with asthma also have
increased inflammatory responses, relative to non-asthmatic subjects, and that this inflammation
may take longer to resolve. The new data on airway responsiveness, inflammation, and various
molecular markers of inflammation and bronchoconstriction indicate that people with asthma
and allergic rhinitis (with  or without asthma) comprise susceptible groups for (Vinduced
                                           6-10

-------
adverse effects. This body of evidence qualitatively informs our evaluation of the adequacy of
the current Os standard in that it indicates that human clinical and epidemiological panel studies
of lung function decrements and respiratory symptoms that evaluate only healthy, non-asthmatic
subjects likely underestimate the effects of 63 exposure on asthmatics and other susceptible
populations.
       In addition to the experimental evidence of lung function decrements, respiratory
symptoms, and  other respiratory effects in healthy and asthmatic populations discussed above,
epidemiological studies have reported associations of lung function decrements and respiratory
symptoms in several locations (Appendix 3B; also Figure 3-4 for respiratory symptoms).  As
discussed in Chapter 3, two large U.S. studies (Mortimer et al., 2002 (the National Cooperative
Inner-City Asthma Study (NCICAS)), Gent et al., 2003), as well as several smaller U.S. and
international studies, have reported fairly robust associations between ambient O3 concentrations
and measures of lung function and daily symptoms (e.g., chest tightness, wheeze, shortness of
breath) in children with moderate to severe asthma and between 63 and increased asthma
medication use. The NCICAS reported statistically significant increases in incidence of > 10%
declines in morning lung function and respiratory symptoms  in asthmatic children for multi-day
lags in 8-hr average 63 concentrations in single pollutant models. For various co-pollutant
models, the Os effect was attenuated, but there was still a positive association.  Gent et al. (2003)
included asthmatic children in the area of southern New England and reported associations
between various respiratory symptoms and both daily 1-hr maximum and 8-hr maximum 63
levels for asthmatics who used maintenance medications and would be considered moderate to
severe asthmatics, while not finding an association for mild asthmatics, defined as not using
maintenance medication.  In this study, effects of 63, but not PM2.5, remained statistically
significant and even increased in magnitude in two-pollutant models (CD, p.7-53).  The CD
concludes that overall the multi-city NCICAS, Gent et al. (2003) and several other single-city
studies indicate a fairly robust positive association between ambient O3 concentrations and
increased respiratory symptoms in asthmatics. We recognize, however, that uncertainties remain
with regard to the relative contribution of 63 and other co-pollutants, some of which show
moderate correlations during the summer time, for the effects observed in asthmatic individuals.
       In considering the large number of single-city epidemiological studies reporting lung
function or respiratory symptoms in healthy or asthmatic populations (Appendix 3B), we note
that most such studies that reported positive and often statistically significant associations in the
warm season were conducted in areas with relevant air quality statistics that are indicative of
areas that likely would not have met the current standard (e.g., Gent et  al., 2003;  Ostro et al.,
2001; Neas et al.,  1995; Delfmo et al, 1998; Linn et al., 1996; Korrick et al.,  1998). In
considering the large multi-city NCICAS (Mortimer et al., 2002), we note that the 98th percentile
                                           6-11

-------
8-hr daily maximum Os concentrations at the monitor reporting the highest Os concentrations in
each of the study areas ranged from 0.084 ppm to > 0.10 ppm. The authors indicate that less
than 5 percent of the days in the eight areas had 8-hr daily Os concentrations exceeding 0.080
ppm.  The authors observed that when days with 8-hr average 63 levels greater than 0.080 ppm
were excluded, similar effect estimates were seen.
       There also are other studies in which the relevant air quality statistics provide some
indication that lung function and respiratory symptom effects may be occurring in areas that
likely would have met the current standard (e.g., Naeher et al., 1999; Ross et al., 2002; Brauer et
al., 1996).  We note that this last group of studies reported associations that were often but not
always statistically significant, and that Brauer et al. (1996) was an outdoor worker study of
berry pickers with exposure patterns that would not be typical of the general population.
       Respiratory Hospital Admissions and Emergency Department Visits
       As discussed in Chapter 3  (section 3.3.1.1.6), at the time of the last review, many time-
series studies indicated positive associations between ambient Os and increased respiratory
hospital admissions and emergency room visits, providing strong evidence for a relationship
between 63 exposure and increased exacerbations of preexisting lung disease at 63 levels below
the level of the then current 1-hr standard. Analyses of data from studies conducted in the
northeastern U.S. indicated that 63 air pollution was consistently and strongly associated with
summertime respiratory hospital admissions (CD, section 8.4.4).
       Since the last review, new epidemiological studies have evaluated the association
between short-term exposures to 63 and unscheduled hospital admissions for  respiratory causes
(CD, section 7.3.3). Large multi-city studies, as well as many studies from individual cities have
reported positive and often statistically significant Os associations with total respiratory
hospitalizations as well as asthma- and COPD-related hospitalizations, especially in studies
analyzing the Os effect during the summer or warm season.  Analyses using multipollutant
regression models suggest that copollutants generally do not confound the association between
63 and respiratory hospitalizations, and that the 63 effect estimates were generally robust to PM
adjustment in all-year and warm-season only data (CD, p. 7-79; Figure 7-12). The CD concludes
that the evidence supports a causal relationship between acute 63 exposures and increased
respiratory-related hospitalizations during the warm season (CD, p. 8-77).
       In looking specifically at U.S. and Canadian respiratory hospitalization studies that
reported positive and often statistically significant associations (and that either did not use GAM
or were reanalyzed to address GAM-related problems), we note that many such studies  were
conducted in areas that likely would not have met the current Os standard, with many providing
only all-year effect estimates, and with some reporting a statistically significant association in the
warm season  (e.g., Schwartz (1996) - Cleveland).  Of the studies that provide some indication

                                            6-12

-------
that (Vrelated respiratory hospitalizations may be occurring in areas that likely would have met
the current standard, we note that some are all-year studies (e.g., Sheppard et al., 2003; Yang et
al., 2003), whereas others reported statistically significant warm-season associations (e.g.,
Burnett et al., 1997a, in 16 Canadian cities; and Burnett et al., 1997b, 2001, in Toronto).  In
further examining the relevant air quality statistics in the 16 Canadian cities study (Burnett et al.,
1997a), we observe that while the aggregated 98th percentile 63 concentration was calculated as
47 ppb (Appendix 3B), the fourth-highest values at the monitors reporting the highest O3
concentrations in each of the cities ranged from approximately 35 to 110 ppb, making it difficult
to determine the extent to which the reported association can be attributed to effects occurring in
areas that likely would have met the current U.S. standard. We also further examined the
relevant air quality statistics in the Burnett et al. (1997b, 2001) studies in Toronto. We observed
that in one of those  studies (Burnett et al., 2001) the fourth-highest values at the highest monitor
ranged from approximately 80 to 150 ppb across the years of the study (from 1980 to 1994).  In
the other study (Burnett et al., 1997b) the calculated 98th percentile values averaged across the
several monitors used in the study ranged from 62 to 64 ppb in each of the three years of the
study (Appendix 3B), but individual monitor data were not available for further examination.
Based on these observations, we find it difficult to judge the extent to which these studies
provide evidence of an association  with respiratory-related hospitalizations in areas that likely
would have met the current standard. Nonetheless, as discussed above, we recognize that these
studies do provide evidence of associations that likely extend down to relatively low ambient 63
concentrations, well below the level of the current standard.
       Emergency department visits for respiratory causes have been the focus of a number of
new studies that have examined visits related to asthma, COPD, bronchitis, pneumonia,  and
other upper and lower respiratory infections, such as influenza, with asthma visits typically
dominating the daily incidence counts (CD, section 7.3.2).  Among  studies with adequate
controls for seasonal patterns, many reported at least one significant positive association
involving O3. However, inconsistencies were observed which were at least partially attributable
to differences in model specifications and analysis approach among various studies. In general,
O3 effect estimates from summer-only analyses tended to be positive and larger compared to
results from cool season or all-year analyses.  Almost all of the studies that reported statistically
significant effect estimates had calculated 98th percentile 63 concentrations (Appendix 3B),
averaged across monitors, that are indicative of areas that likely would not have met the current
standard.  The notable exception were two studies in Montreal (Delfmo et al., 1997, 1998) that
reported statistically significant warm-season associations with O3 and ED visits in a population
of older adults with a  calculated 98th percentile value, averaged across several monitors, of
approximately 60 ppb (and for which individual monitor data were not available for further
                                            6-13

-------
examination), although the CD raises questions about the plausibility of this result due to the low
Os concentrations and inconsistent results across years and age groups. The CD concluded that
analyses stratified by season generally supported a positive association between O3
concentrations and ED visits for asthma in the warm season.  These studies provide evidence of
effects in areas that likely would not have met the current standard, and evidence of associations
that likely extend down to relatively low ambient 63 concentrations.
       Mortality
       The 1996 CD concluded that an association between daily mortality and 63
concentrations for areas with high Os levels (e.g., Los Angeles) was suggested.  However, due to
a very limited number of studies available at that time, there was insufficient evidence to
conclude that the observed association was likely causal, and thus the possibility that O3
exposure may be associated with mortality was not relied upon in the 1997 decision on the Os
primary standard.
       Since the last review, as described above, the body of evidence with regard to Os-related
health effects has been expanded by animal, human clinical, and epidemiological studies, and
now includes biologically plausible mechanisms by which 63 may affect the cardiovascular
system. In addition, there is stronger information linking ozone to serious morbidity outcomes,
such as hospitalization, that are associated with increased mortality. Thus, there is now a
coherent body  of evidence that describes a range of health outcomes from  pulmonary function
decrements to hospitalization and premature mortality.
       Newly  available large multi-city studies designed specifically to examine the effect of 63
and other pollutants on mortality have provided much more robust and credible information.
The extended NMMAPS analysis included data from 95 U.S. cities and included an additional 6
years of data, from 1987-2000 (Bell et al., 2004), and significant associations were reported
between Os and mortality that were robust to adjustment for PM (CD, p. 7-100).  Using a subset
of the NMMAPS data set, Huang et al. (2005) focused on associations between cardiopulmonary
mortality and 63 exposure (24-hr average) during the summer season only. The authors report
the increase in mortality per 20 ppb change in Os concentration using a 7-day distributed lag
model was greater than the increase in mortality measured on the same day (CD, p. 7-92),
suggesting that the effect of O3 on mortality is immediate but also persists for several days.
Using a case-crossover study design, Schwartz (2005) assessed associations between daily
maximum concentrations and mortality in 14 cities, matching case and control  periods by
temperature, and using data only from the warm season. The reported effect estimate was
similar to time-series analysis results with adjustment for temperature, suggesting that
associations between 63 and mortality were robust to the different adjustment methods for
temperature (CD, p. 7-93).  Two of the recent multi-city mortality studies (Bell et al., 2004;

                                           6-14

-------
Gryparis et al., 2004) have also reported associations for multiple averaging times (CD, p. 8-38).
Bell and colleagues (2004) reported an effect estimate for the association with 1-hr Os
concentrations that was slightly larger than that reported for 8-hr Os concentrations, and both
were slightly larger than the association with 24-hr average 63, but the effect estimates did not
differ statistically.
       One recent multi-city study (Bell et al., 2006) examined the shape of the concentration-
response function for the O3-mortality relationship in 98 U.S. urban communities for the period
1987 to 2000 specifically to evaluate whether a "safe" threshold level exists. Results from
various analytic methods all indicated that any threshold, if it exists, would likely occur at very
low concentrations, far below the level of the current Os NAAQS and other, lower international
Os standards,2 and nearing background levels. Notably, in a subset analysis using only days that
were below the level of the current 63 NAAQS, the (Vmortality association remained
statistically significant with  only a small change in the size of the effect estimate. Further, in a
subset analysis based on 24-hr average 63 concentrations, the effect estimates declined and lost
statistical significance only when the maximum daily average concentration included was < 10
ppb (Bell et al., 2006, p. 14 and Figure 2), which corresponds to daily maximum 8-hr average
concentrations in U.S. cities that are within the  range of background concentrations. The authors
conclude that "interventions to further reduce ozone pollution would benefit public health, even
in regions that meet current regulatory standards and guidelines" (Bell et al., 2006,  p. 3).
       New data are also available from several single-city studies conducted world-wide, as
well as from  several meta-analyses that have combined information from multiple studies.  The
majority of these studies suggest that there is an elevated risk of total non-accidental mortality
associated with acute exposure to 63, especially in the summer or warm  season when 63 levels
are typically high, with  somewhat larger effect estimate sizes for associations with
cardiovascular mortality (CD, p. 7-175).  As shown in Figure 7-21 of the CD, the results of
recent publications show a pattern of positive, often statistically significant associations between
short-term Os exposure  and mortality during the warm season (CD, p.  7-97). For example,
statistically significant associations were reported in southern California (Ostro, 1995),
Philadelphia (Moolgavkar et al., 1995), Dallas (Gamble et al.,  1998), and Vancouver (Vedal et
al., 2003), as well as numerous studies conducted in other countries. In contrast, no evidence of
an association was seen in a study conducted in Pittsburgh (Chock et al., 2000).  In considering
results from year-round analyses, there remains a pattern of positive results but the findings are
       2 Other international 8-hr O3 standards considered by Bell et al. (2006, Table 1) include the California
standard of 0.070 ppm, the Canadian standard of 0.065 ppm, and the World Health Organization guideline and
European Commission target value of approximately 0.061 ppm.
                                            6-15

-------
less consistent. For example, statistically significant associations were reported in Philadelphia
(Moolgavkar et al., 1995) and Dallas (Gamble et al., 1998), while positive but not statistically
significant associations were reported in Detroit (Lippmann et al., 2000, reanalyzed in Ito, 2003),
San Jose (Fairley, 1999, reanalyzed Fairley, 2003), and Atlanta (Klemm et al., 2004).  No
evidence for associations was reported in Los Angeles (Kinney et al., 1995), Coachella Valley
(Ostro et al., 2003), and St. Louis and Eastern Tennessee (Dockery et al., 1992).  In most single-
city analyses, effect estimates were not substantially changed with adjustment for PM (CD
Figure 7-22, p. 7-101).
       Almost all single-city studies that show statistically significant associations with
mortality had calculated 98th percentile Os concentrations (Appendix 3B) that are indicative of
Os levels that likely would not have met the current standard.  The notable exception was a study
in Vancouver (Vedal et al., 2003) that reported a statistically significant warm-season association
with Os and total non-accidental mortality that was robust in two-pollutant models, with  a
calculated 98th percentile value, averaged across many monitors, of approximately 53 ppb. Upon
further examination, the relevant air quality statistics for each individual monitor in this study
ranged from 57 to 59 ppb.  This study provides evidence of an Os-related mortality association in
the warm season in an area with Os levels that were well below those that would have met the
current standard. However, the authors questioned whether Os, other gaseous pollutants, and PM
may be acting as surrogate markers of pollutant mixes that contain more toxic compounds, since
the low measured concentrations were unlikely, in their opinion, to cause the observed effects
(CD, p.7-155). Another study done in Vancouver over a much longer time period (Villeneuve et
al., 2003) did not provide evidence of Os-related mortality associations, but only all-year results
were presented which may be more likely confounded by other pollutants than the warm-season
results in Vedal et al. (2003).
       Three recent meta-analyses evaluated potential sources of heterogeneity in Os-mortality
associations (Bell et al., 2005;  Ito et al., 2005; Levy et al., 2005).  The CD (p. 7-96) observes
common findings across all three analyses, in that all reported that effect estimates were  larger in
warm season  analyses, reanalysis of results using default GAM criteria did not change the effect
estimates, and there was no strong evidence of confounding by PM (CD, p. 7-97). Bell et al.
(2005) and Ito et al. (2005) both provided suggestive evidence of publication bias, but O3-
mortality associations remained after accounting for that potential bias.  The CD (7-97)
concludes that the "positive Os effects estimates, along with the sensitivity analyses in these three
meta-analyses, provide evidence of a robust association between ambient Os and mortality."
       The CD finds that the results from U.S. multi-city time-series studies, along with the
meta-analyses, provide relatively strong evidence for associations between short-term Os
exposure and all-cause mortality even after adjustment for the influence of season and PM (CD,
                                            6-16

-------
p. 8-78). The results of these analyses indicate that copollutants generally do not appear to
substantially confound the association between Os and mortality (CD, p. 7-103; Figures 7-22 and
7-23).  In addition, several single-city studies observed positive associations of ambient Os
concentrations with total nonaccidental and cardiopulmonary mortality.  Finally, from those
studies that included assessment of associations with specific causes of death, it appears that
effect estimates for associations with cardiovascular mortality are larger than those for total
mortality; effect estimates for respiratory mortality are less consistent in size, possibly due to
reduced statistical power in this subcategory of mortality (CD, p. 7-108). For cardiovascular
mortality, the CD (Figure 7-25, p. 7-106) suggests that effect estimates are consistently positive
and more likely to be larger and statistically significant in warm season analyses.  The CD (p. 8-
78) concludes that these findings are highly suggestive that short-term Os exposure directly or
indirectly contributes to non-accidental and cardiopulmonary-related mortality, but additional
research is needed to more fully establish underlying mechanisms by which such effects occur.

       6.3.1.2  Exposure- and Risk-based Considerations
       In addition to the evidence-based considerations, staff has  also considered exposures and
health risks estimated to occur upon meeting the current O3 standard to help inform judgments
about the extent to which exposure and risk estimates may be judged to be important from a
public health perspective, taking into account key uncertainties associated with the estimated
exposures and risks.  For this review,  exposures have been estimated for people of all ages,
school age children (ages 5-18), and asthmatic school age children in  12 urban areas across the
U.S.3 In this discussion we focus particularly on the exposure estimates for all and asthmatic
school age children while at moderate or greater exertion levels since these groups are
particularly at risk for experiencing Os-related health effects due to the greater amount of time
spent outdoors during the 63 season engaged in relatively high levels of physical activity.  With
regard to the quantitative risk assessment, we have estimated the occurrences of moderate or
greater lung function decrements in all and asthmatic school age children, respiratory symptoms
in children with moderate to severe asthma, respiratory-related hospital admissions, and non-
accidental and cardiorespiratory mortality. We have selected these particular health endpoints
because they can be considered to be adverse to the health of individuals, either in single or with
repeated occurrences, and the requisite data for the assessment are available.
       In making judgments as to when various Os-related effects become regarded as adverse
to the health of individuals, staff has relied upon the guidelines  published by the American
       3As discussed in Chapter 4, since the proportion of children classified as "active" in the exposure analysis
has been overestimated, in part due to uncertainty in the CHAD MET values, we have not discussed this population
group in this chapter.
                                             6-17

-------
Thoracic Society (ATS) and the advice of CASAC. As discussed in Chapter 3, the ATS (1985)
has defined adverse respiratory health effects as medically significant physiologic changes,
which include acute changes such as those that interfere with the normal activity or acute
respiratory illness, and longer-term changes such as progressive respiratory dysfunction,
permanent respiratory injury, or incapacitating illness. Of the morbidity effects estimated in the
risk assessment, hospital admissions are clearly adverse to the health of the individual.
Recognizing that respiratory symptoms in moderate to severe asthmatic children are likely to
cause increased medication use, interference with normal  activities, or increased absences from
school, occurrences of such symptoms are considered to be adverse. For active healthy people,
moderate lung function decrements would likely interfere with normal activity for relatively few
individuals within this group who are particularly responsive to Os exposures; whereas large
functional responses would likely interfere with normal activities for a greater proportion of
individuals within this group, such that even single occurrences of large functional responses
would be considered adverse under ATS guidelines.  In judging the extent to which moderate
lung function decrements are adverse, especially in healthy people,  an additional factor that has
been considered is whether these occur on a single occasion or repeatedly over the course of an
63 season. It has been judged that repeated occurrences of moderate responses, even in
otherwise healthy individuals, may be considered to be adverse since they could well set the
stage for more serious illness.  For people with lung disease, even moderate lung function
decrements would likely interfere with normal activity for many individuals within this group,
and would likely result in additional and more frequent use of medication; large functional
responses would likely interfere with normal activity for most individuals in this group and
would increase the likelihood that these individuals would seek medical treatment. Thus,
occurrences of either moderate or large functional responses in people with lung disease would
be considered to be adverse to the health of individuals experiencing these effects.
       Beyond the health effects discussed above that are included  in the risk assessment,
Chapter 3 discusses a broader array  of Os-related health endpoints that are representative of a
"pyramid of effects" that include various  indicators of morbidity that could not be included in the
risk assessment (e.g., school absences, increased medication use, ED visits). Ozone-related
effects that are judged to be important indicators of this broader array of health endpoints, and
are thus potentially adverse to the individuals experiencing such effects, include: (1) increased
nonspecific airway responsiveness which is related, for example, to aggravation of asthma,
potentially leading to increased medication use, increased school and work absences, increased
visits to doctors' offices and EDs, and increased admissions to hospitals; (2) decreased
pulmonary defense mechanisms which are suggestive of increased susceptibility to respiratory
infection, potentially leading to increased school and work absences, increased visits to doctors'
                                            6-18

-------
offices and EDs, and increased admissions to hospitals; and (3) indicators of pulmonary
inflammation which are related to aggravation of asthma, potentially leading to increased
medication use, increased school and work absences, increased visits to doctors' offices and
EDs, and increased admissions to hospitals; increased cellular permeability associated with
inflammation may be a mechanism by which Os exposure can lead to cardiovascular system
effects; and potential chronic effects such as chronic bronchitis or long-term damage to the lungs
that can lead to reduced quality of life.  Some perspective on the O3-related public health
impacts of these types of effects are characterized based on the results of the exposure analysis in
terms of estimates of the number of occurrences of "exposures of concern," as discussed below.
       We estimated exposures and risks for the three most recent years (2002 - 2004) for which
data were available at the time of the  analyses, as discussed in Chapters 4 and 5. Within this 3-
year period, 2002 was a year with generally poorer air quality in most, but not all,  areas and
provides a generally more upper-end  estimate of exposures and risks, while 2004 was a year with
generally better air quality in most, but not all, areas and provides a generally more lower-end
estimate of exposures and risks. Exposure and risk estimates for the year 2003 generally fall
between the estimates for 2002 and 2004. In presenting these results, we note that the exposure
analysis and risk assessment discussed in Chapters 4 and 5, respectively, identify a number of
uncertainties, as highlighted below.
       With respect to the exposure analysis, the exposure modeling approach accounts for
variability in ambient 63 levels, demographic characteristics, physiological attributes, activity
patterns, and factors affecting microenvironmental concentrations.  In our judgment the most
important uncertainties affecting the exposure estimates are related to the modeling of activity
patterns over an 63 season, modeling micro-scale variations in ambient concentrations (e.g., near
roadways), and modeling air exchange rates in microenvironments. Another important
uncertainty that does not directly affect estimates of exposure, but affects the characterization of
how many exposures are associated with moderate or greater exertion, is the characterization of
energy expenditure (i.e., measured in terms of METS - metabolic equivalents of work) for
children engaged  in various activities. As discussed in section 4.3.4.7, the uncertainty in METS
values carries over to the uncertainty  of the modeled ventilation rates, which are important since
they are used to classify exposures of potentially greater risk.
       A comprehensive picture of the uncertainty of the exposure model estimations has been
developed through sets of complementary analyses addressing these different aspects of the
overall uncertainty. An analysis was  performed which accounts for the uncertainties associated
with the microenvironment models and the ambient air quality data. Analyses have been
conducted to address the remaining significant sources of uncertainty: near-road exposures and
the activity data.  The uncertainty of the model structure (as distinct from uncertainty driven by
                                            6-19

-------
uncertain model inputs) is judged to be less important than the uncertainties of the model inputs
and parameters. Based on these analyses, relatively small uncertainties are associated with the
estimation of ambient concentrations and microenvironment model parameters.  Relatively larger
uncertainties are associated with the representativeness of the CHAD activity diaries with respect
to the specific cities and time periods modeled. The APEX model significantly underestimates
the frequency of occurrences of individuals experiencing repeated 8-hour average exposures
greater than 0.06,  0.07, and 0.08 ppm.  While the frequency of repeated occurrences is
significantly underestimated, we have more confidence that the estimates for total number of
occurrences (i.e., person days with exposures greater than 0.06, 0.07, and 0.08 ppm) are not
biased.  Section 4.6 provides a summary of the exposure modeling uncertainties; the details of
this uncertainty analysis are described in Langstaff (2007).  It is important to note that there have
been significant improvements in several components of the exposure model and in the inputs to
the model (e.g., better characterization of the ambient air quality surface across each area, more
complete data on air exchange rates, much larger human activity database) relative to the
exposure analysis conducted for the 1997 review.  Thus, while we recognize and have considered
the kind and degree of uncertainties associated with the exposure estimates, we believe,
consistent with CAS AC's views (Henderson, 2006c), that the exposure analysis represents  a
state-of-the-art modeling  approach and the quality of the estimates is such that they are suitable
to be used as an input to the decisions on the Os standard.
       As discussed in Chapter 5, uncertainties related to the air quality data affect both the
controlled human exposure studies-based and the epidemiological studies-based parts of the risk
assessment.  These include uncertainties associated with the air quality adjustment procedure that
was used to  simulate just  meeting the current and alternative 8-hr standards, and the uncertainties
associated with estimating background Os concentrations for each location.  Based on sensitivity
analyses conducted in the prior review, alternative air quality adjustment procedures had only a
modest  impact on the risk estimates. With respect to uncertainties about estimated background
concentrations,  as discussed in section 5.4.3, alternative assumptions about PRB levels have a
variable impact depending on the location and standard analyzed in terms of the absolute
magnitude of the risk estimates.  Alternative assumptions about PRB levels have a greater impact
on risk estimates associated with more stringent 8-hr standards.  However, the overall pattern of
reductions, expressed in terms of percentage reduction relative to the current standard, is
generally unaffected by alternative assumptions for PRB for most of the standards analyzed.
       With respect to the lung function part of the health risk assessment, key uncertainties
include  uncertainties in the  exposure estimates for children engaged in moderate or greater
exertion (noted  above) and uncertainties associated with the shape of the exposure-response
relationship, especially at levels below 0.08 ppm, 8-hr average, where only limited data are
                                            6-20

-------
available down to 0.04 ppm and there is an absence of data below 0.04 ppm.  The uncertainty
associated with the shape of the exposure-response relationship was considered using a Bayesian
Markov Chain Monte Carlo approach as recommended by members of the CAS AC panel
(Henderson, 2006c).  As discussed in section 5.3.2.5, the impacts of changing the functional
form of the exposure-response relationship varied substantially. Changing from the 90%
logistic/10% linear base case to the 80%/20% split generally had only a small impact for all
school age children, with most risk estimates being within 5% of the base case estimates. The
impacts of changing from the base case to the 50% logistic/50% linear case were generally
(although not always) larger.  We  observed greater changes for all school age children between
the 50/50 split and the base case in terms of percent change in risk for two more stringent
alternative standards relative to the current standard. With respect to the lung function risk
estimates for asthmatic children, there were relatively small changes observed between the  50/50
split and the base case in the percent changes associated with two more stringent alternative
standards relative to the current standard.
       With respect to the part of the health risk assessment based on effects reported in
epidemiological studies, an important uncertainty for the mortality risk estimates is the extent to
which the associations reported between 63 and non-accidental and cardiorespiratory mortality
actually reflect causal relationships. Other important uncertainties for this part  of the risk
assessment include uncertainties (1) surrounding estimates of Os coefficients in concentration-
response functions used in the assessment, (2) concerning the specification of the concentration-
response model (including the shape of the relationship) and whether or not a population
threshold or non-linear relationship exists within the range of concentrations  examined in the
studies, (3) related to the extent to which concentration-response relationships derived from
studies in a given location and time when Os levels were higher or behavior and/or housing
conditions were different provide accurate representations of the relationships for the same
locations with lower air quality distributions and different behavior and/or housing conditions,
and (4) concerning the possible role of co-pollutants which also may have varied between the
time of the  studies and the current assessment period.  For both parts of the risk assessment,
statistical uncertainty due to sampling error has been characterized.  As discussed in section
5.4.5, there are additional unquantified uncertainties including model uncertainty noted above.
       While we and CASAC have recognized the various uncertainties that are inherent in
conducting such risk assessments, CASAC found the health risk assessment to be "well done,
balanced, and reasonably communicated" (Henderson, 2006c). We have considered the kind and
extent of uncertainties in the health risk estimates, but judge that these estimates, discussed in
Chapter 5 and in this chapter, are appropriate for consideration as an input to the decisions on the
63 standard.
                                            6-21

-------
       The exposure and risk assessments estimated occurrences of exposures of concern and
lung function decrements, respectively, in school-age children ages 5 to 18. Due to the increased
amount of time spent outdoors engaged in relatively high levels of physical activity, school-age
children, as a group, are particularly at risk for experiencing CVrelated health effects.  We report
results for school-age children down to age five, but there is a trend for younger children to
attend school. We are not taking these younger children into account in our analysis due to a
lack of information which would let us characterize this group of children.  Some states allow 4-
year-olds to attend kindergarten, and more than 40 states have preschool programs for children
younger than five (Blank and Mitchell, 2001). In 2000, six percent of U.S.  children ages 3 to  19
who attend school were younger than five years old (2000  Census Summary File 3, Table QT-
P19: School Enrollment). Clearly the estimates of exposures of concern and lung function
decrements in school-age children would be higher had we been able to include this group of
children in the exposure and risk assessments.
       Exposure Assessment Results
       The results of exposure assessments, which provide estimates of the number of people
exposed to different levels of ambient 63 while at prescribed exertion levels, serve two purposes.
First, the entire range of modeled personal exposures to ambient Os is an essential input to the
risk assessments based on exposure-response functions from controlled human exposure studies.
Secondly, estimates of personal exposures to ambient O3 levels at and above specific benchmark
concentrations provide some perspective on the public health impacts of health effects that we
cannot currently evaluate in quantitative risk assessments that may occur at current air quality
levels, and the extent to which such impacts might be reduced by meeting the current and
alternative standards.  This is especially true when there are exposure levels at which we know or
can reasonably infer that specific (Vrelated health effects  are occurring. We refer to exposures
at and above these benchmark concentrations as "exposures of concern."
       Estimating exposures of concern is important because it provides some indication of the
potential public health impacts of a range of (Vrelated health outcomes, such as  pulmonary
inflammation, increased airway responsiveness, or changes in host defenses.4  These particular
       4 We note that estimates of the number of people likely to experience exposures of concern can not be
directly translated into quantitative estimates of the number of people likely to experience specific health effects,
since sufficient information to draw such comparisons is not available ~ if such information were available, we
would have included these health outcomes in the quantitative risk assessment  As discussed in section 3.3.1.1.3, the
studies reporting inflammatory responses and markers of lung injury have clearly acknowledged that there is
significant variation in responses of subjects exposed, especially to 6.6 hour O3 exposures at 0.08 and 0.10 ppm. To
provide some perspective on the public health impacts of these health effects, we note that one study (Devlin et al.,
1991, Figure 5), for example, showed that roughly 10 to 50% of the 18 young healthy adult subjects in the study
experienced notable increases (i.e., > 2 fold increase) in most of the inflammatory and cellular injury indicators

                                             6-22

-------
effects have been demonstrated in controlled human exposure studies of healthy individuals to
occur at levels as low as 0.08 ppm Os, but have not been evaluated at lower levels in controlled
human exposure studies.  We have not estimated these effects in a quantitative risk assessment
due to the lack of adequate information on the overall exposure-response relationships. For the
exposure analysis in the second draft Staff Paper, an exposure of concern was defined the same
way as it was in the 1997 review, i.e., an 8-hr average exposure > 0.08 ppm Os while engaged in
intermittent moderate or greater exertion levels (62 FR 38860). However, at the August 2006
CAS AC meeting, the CAS AC Os Panel encouraged staff to also consider exposures of concern
at levels lower than 0.08 ppm, expressing the view that there is a range of health effects that
likely occur below the lowest level tested for these effects (EPA 2006e, p. 104-105). There is no
reason to assume that there is a threshold for effects,  such as markers of inflammation, at 0.08
ppm 63.  Moreover, Panel members noted there is evidence of adverse effects that are strongly
correlated with pulmonary inflammation, such as ED visits observed in epidemiological studies,
at levels well below 0.08 ppm.
       We concur  with these views, and note that exposures of concern, and the health outcomes
they represent, likely occur across a range of Os exposure levels, such that there is no one
exposure level that addresses all relevant public health concerns.  Therefore, we have estimated
exposures of concern not only at 0.080 ppm Os, a level at which there are demonstrated effects,
but also at 0.070 and 0.060 ppm Os, recognizing that there is no apparent threshold for Os health
effects and that potentially adverse lung function decrements have been demonstrated in
controlled human exposure studies of healthy individuals at 0.060 + 0.003 ppm Os.  Moreover,
there will be varying degrees of concern about exposures at each of these levels, based in part on
the population subgroups experiencing them. For example, it is reasonable to conclude that a
high degree of protection is warranted against effects that have been clearly demonstrated in
healthy people, not only for the general public,  but especially for members  of sensitive
subgroups, such as children or people with asthma and other lung diseases. At levels where
effects have not been demonstrated in controlled human exposure  studies but there is reason to
infer that effects likely occur, or where the evidence is less clear, the appropriate level of
protection will depend on the strength of the evidence and the adversity of the effect.  At
comparable levels  of uncertainty in the evidence, it is important to provide  stronger protection
against effects that are more clearly adverse.  Given that there is clear evidence of inflammation,
increased airway responsiveness, and changes in host defenses in healthy people exposed to
0.080 + 0.004 ppm Os and reason to infer that such effects will continue at  lower exposure
analyzed associated with 6.6-hour exposures at 0.08 ppm. We also note that susceptible subpopulations such as
those with asthma may be even more affected.
                                           6-23

-------
levels, but with increasing uncertainty about the extent to which such effects occur at lower 63
concentrations, in the following discussion we present information on all three levels of
exposures of concern, with a focus on exposures of concern at levels > 0.070 and > 0.060 ppm
03.
       Exposure estimates for 6 scenarios (i.e., recent air quality for each of the three years
examined, 2002 to 2004, and just meeting the current standard in each of the same three years
based on adjusting air quality in this three-year period to determine the amount of adjustment
needed) aggregated across 12 urban areas, have been developed for each of these exposure-of-
concern levels. These estimates are shown in Table 6-1. The exposure estimates are for the
number and percent of people exposed, in each of the population subgroups, and the number of
person-days (occurrences) of exposures, with daily 8-hr maximum average exposures at or above
0.080 ppm (Table 6-la), 0.070 ppm (Table 6-lb), and 0.060 ppm (Table 6-lc), while at
intermittent moderate or greater exertion.5
       As shown in Table 6-1, the patterns of exposures in terms of percentages of the
population exceeding a given exposure level are very similar for the general population and for
all and asthmatic school age (5-18) children, although  children are about twice as likely to be
exposed, based on the percent of the population exposed, at any given level. Thus, in the
discussion below, we focus on the patterns observed for all school  age children, which includes
asthmatic children, with the recognition that these exposure patterns apply to this other
subpopulation. While Table 6-1 shows aggregate results, it is important to note that there is
substantial variability in the percent of the population subgroups estimated to experience
exposures of concern across the 12 urban areas. For example, in the case of exposures of
concern > 0.070 ppm Os, for 2002 when the  current standard is just met, while the aggregate
estimate is 16% (Table  6-lb), the estimates of exposures for all children range from about 1% to
more than 35% of the population  across the 12 urban areas analyzed (see Figure 4-7).
Variability in the degree of health protection offered across urban areas is an important factor to
consider in evaluating the adequacy of the current standard.
       Information in Table 6-1 is drawn from Appendices 4A.
                                           6-24

-------
Table 6-la. Summary of Estimates of Number of People Exposed and Number of Occurrencesl at Moderate Exertion!
           Associated with 8-Hour Daily Maximum Ozone Concentrations > 0.080 ppm for 12 Urban Areas in the U.S.
Air Quality
Scenario
General Population
(88.5 million people)
Persons
(% of population)
Person Days
[% change from
recent air quality]
All Children (5-1 8 years old)
(18.3 million children)
Persons
(% of population)
Person Days
[% change from
recent air quality]
Asthmatic Children (5-18 years old)
(2.6 million children)
Persons
(% of population)
Person Days
[% change from
recent air quality]
2002 Air Quality
Recent Air Quality
Just Meeting
Current Standard
9,020,000
(10%)
1,550,000
(2%)
11,670,000
1,670,000
[86% reduction]
3,660,000
(20%)
600,000
(3%)
4,930,000
660,000
[87% reduction]
570,000
(22%)
100,000
(4%)
760,000
110,000
[85% reduction]
2003 Air Quality
Recent Air Quality
Just Meeting
Current Standard
5,030,000
(6%)
360,000
(0.4%)
6,900,000
360,000
[95% reduction]
2,040,000
(11%)
130,000
(0.7%)
2,920,000
140,000
[95% reduction]
320,000
(12%)
20,000
(0.9%)
440,000
20,000
[95% reduction]
2004 Air Quality
Recent Air Quality
Just Meeting
Current Standard
2,000,000
(2%)
50,000
(0.1%)
2,520,000
50,000
[98% reduction]
780,000
(4%)
20,000
(0.1%)
1,000,000
20,000
[98% reduction]
110,000
(4%)
3,000
(0.1%)
150,000
3,000
[98% reduction]
Estimates for persons and person days are rounded to the nearest 10,000. Percentages greater than or equal to 1 are rounded to the nearest percent and
percentages less than one are rounded to the nearest tenth.
 Moderate exertion is defined as having an 8-hr average equivalent ventilation rate (EVR) in the range 13-27 1-min/m2.
                                         6-25

-------
Table 6-lb.  Summary of Estimates of Number of People Exposed and Number of Occurrences1 at Moderate Exertion2
           Associated with 8-Hour Daily Maximum Ozone Concentrations > 0.070 ppm for 12 Urban Areas in the U.S.
Air Quality
Scenario
General Population
(88.5 million people)
Persons
(% of population)
Person Days
[% change from
recent air quality]
All Children (5-1 8 years old)
(18.3 million children)
Persons
(% of population)
Person Days
[% change from
recent air quality]
Asthmatic Children (5-18 years old)
(2.6 million children)
Persons
(% of population)
Person Days
[% change from
recent air quality]
2002 Air Quality
Recent Air Quality
Just Meeting
Current Standard
20,220,000
(23%)
7,260,000
(8%)
34,740,000
9,170,000
[74% reduction]
7,860,000
(43%)
2,920,000
(16%)
15,040,000
3,830,000
[75% reduction]
1,180,000
(46%)
450,000
(18%)
2,280,000
600,000
[74% reduction]
2003 Air Quality
Recent Air Quality
Just Meeting
Current Standard
12,140,000
(14%)
1,910,000
(2%)
20,150,000
2,040,000
[90% reduction]
4,840,000
(27%)
750,000
(4%)
8,760,000
800,000
[91% reduction]
720,000
(28%)
130,000
(5%)
1,290,000
140,000
[89% reduction]
2004 Air Quality
Recent Air Quality
Just Meeting
Current Standard
6,560,000
(7%)
520,000
(0.6%)
9,990,000
550,000
[94% reduction]
2,610,000
(14%)
190,000
(1%)
4,220,000
200,000
[95% reduction]
370,000
(14%)
30,000
(1%)
590,000
30,000
[95% reduction]
Estimates for persons and person days are rounded to the nearest 10,000. Percentages greater than or equal to 1 are rounded to the nearest percent and
percentages less than one are rounded to the nearest tenth.
 Moderate exertion is defined as having an 8-hr average equivalent ventilation rate (EVR) in the range 13-27 1-min/m2.
                                         6-26

-------
Table 6-lc. Summary of Estimates of Number of People Exposed and Number of Occurrences1 at Moderate Exertion2
           Associated with 8-Hour Daily Maximum Ozone Concentrations > 0.060 ppm for 12 Urban Areas in the U.S.
Air Quality
Scenario
General Population
(88.5 million people)
Persons
(% of population)
Person Days
[% change from
recent air quality]
All Children (5-1 8 years old)
(18.3 million children)
Persons
(% of population)
Person Days
[% change from
recent air quality]
Asthmatic Children (5-18 years old)
(2.6 million children)
Persons
(% of population)
Person Days
[% change from
recent air quality]
2002 Air Quality
Recent Air Quality
Just Meeting
Current Standard
34,060,000
(39%)
19,480,000
(22%)
91,420,000
34,580,000
[62% reduction]
12,200,000
(67%)
7,370,000
(40%)
40,540,000
14,800,000
[63% reduction]
1,800,000
(70%)
1,130,000
(44%)
6,110,000
2,290,000
[63% reduction]
2003 Air Quality
Recent Air Quality
Just Meeting
Current Standard
25,080,000
(28%)
8,580,000
(10%)
57,080,000
10,860,000
[81% reduction]
9,370,000
(52%)
3,380,000
(19%)
25,060,000
4,390,000
[82% reduction]
1,380,000
(54%)
520,000
(21%)
3,660,000
680,000
[81% reduction]
2004 Air Quality
Recent Air Quality
Just Meeting
Current Standard
17,530,000
(20%)
3,680,000
(4%)
36,810,000
4,320,000
[88% reduction]
6,810,000
(37%)
1,450,000
(8%)
16,280,000
1,710,000
[90% reduction]
970,000
(38%)
220,000
(8%)
2,260,000
250,000
[89% reduction]
Estimates for persons and person days are rounded to the nearest 10,000. Percentages greater than or equal to
percentages less than one are rounded to the nearest tenth.
                                                                                      1 are rounded to the nearest percent and

Moderate exertion is defined as having an 8-hr average equivalent ventilation rate (EVR) in the range 13-27 1-min/m2.
                                         6-27

-------
       In addition to city-to-city variability, substantial year-to-year variability in exposure
estimates is observed. For example, about 2.6 million children (14% of all school age children
for the 12 urban areas) to more than 7.8 million children (43% of all school age children for the
12 urban areas), are estimated to experience one or more exposures of concern (i.e., > 0.070
ppm, while engaged in moderate or greater exertion) for 2004 and 2002, respectively.6 When air
quality is adjusted to simulate just meeting the 8-hr standard, the estimated number of children
exposed is substantially reduced. Depending on which year is adjusted for just meeting the
current standard, approximately 190,000 children (1% of all school age children) (based on 2004
air quality data) to more than 2.9 million (16% of all school age children) (based on 2002 air
quality data) are estimated to experience one or more exposures of concern.  These results
suggest reductions of approximately 95%, based on 2004 air quality, to 75%, based on 2002 air
quality, in the number of children estimated to experience one or more 8-hr average exposures
above 0.070 ppm when the current 8-hr Os standard is just met.
       We have also examined the  extent to which individuals are likely to experience repeated
exposures of concern, which is an important aspect to  consider when making judgments  about
the extent to which these exposures can be considered adverse to individuals. However, based
on the analysis described in section 4.6.4, staff concludes that the APEX exposure model
significantly underestimates the frequency of occurrences of individuals (adults and children)
experiencing repeated 8-hour average exposures at the three levels.  As discussed in Chapter 4,
this underestimation results from the way that people's activities are modeled in APEX using
CHAD, which does not properly account for repeated behavior of adults and children, and may
be the greatest source of uncertainty in the exposure estimates.  Thus, it is likely that the number
of repeated exposures, which are estimated here to be relatively small, is significantly
underestimated. As seen in Table 6-1, under any of the air quality scenarios, the  estimated
number of occurrences is only slightly larger than the estimated number of people exposed,
indicating that the estimated number of repeated exposures is relatively small.  Moreover, due to
limitations in the CHAD database, the exposure assessment does not include outdoor workers,
some proportion of whom are likely to be exposed repeatedly day after day to elevated ambient
Os levels while at work.
       Risk Assessment Results
       Turning to risk-based considerations, as discussed in Chapter 5, risk estimates have been
calculated and are discussed below  for several important health endpoints, including:
       6Unless otherwise noted, estimates for 2003 fall between the estimates for 2002 and 2004.
                                         6-28

-------
   •   Lung function decrements (i.e., > 15% and > 20% reductions in FEVi) in all school age
       children for 12 urban areas;
   •   Lung function decrements (i.e., > 10% and > 20% reductions in FEVi) in asthmatic
       school age children for 5 urban areas (a subset of the 12 urban areas);
   •   Respiratory symptoms (i.e., chest tightness, shortness of breath, wheeze) in moderate to
       severe asthmatic children for the Boston area;
   •   Respiratory-related hospital admissions for 3 urban areas;
   •   Non-accidental and cardiorespiratory mortality for 12 urban areas.
       In the sections on lung function  decrements in children (all and asthmatic children), we
describe the scope  of the assessments, the number and percent of children experiencing moderate
and large lung function decrements, the number of occurrences of moderate and large lung
function decrements, and what risk is estimated to remain after the current 8-hr standard is met.
For effects such as lung  function decrements, which are transient and reversible, aspects such as
the likelihood that these effects would interfere with normal  activities or occur repeatedly  are
important to consider in making judgments about adversity to individuals and are described
below.  There are discussions about year-to-year variability in the results and variability in
results across the urban areas, which are important for making judgments about public health
impacts. These estimates indicate that there are substantial differences in the natural fluctuation
of air quality levels from year to year. This can result in significant variability in the number of
children affected by moderate or greater lung function decrements and the number of
occurrences they experience during a 3-year period that is adjusted to just meet the current 8-hr
standard. For example, the number of children affected, and the number of occurrences, can
increase by more than 100% for a year with generally poorer air quality compared to a year with
better air quality, within a three year period.
       Other aspects of the risk information presented are important to consider.  The first is that
there is some degree of consistency in the estimated population risk across the 12 urban areas, as
indicated by the percent of the population estimated to be affected, which describes the risk
normalized across the urban  areas with very different population sizes. In Table 6-2, the percent
of all  children likely to experience one or more moderate or greater lung function responses (>
15% reduction in FEVi) under recent air quality and when air quality just meets the current 8-hr
standard are 7% and 3% (based on 2002 air quality), respectively. The range across the  12 urban
areas, from Table 5-6 is  approximately 5% to 9% under recent (2002) air quality, and about 1%
to 5% when air quality is adjusted to just meet the current 8-hr standard based on that year. The
pattern across the 12 urban areas is similar for the risk estimates based on 2004 air quality.
       The remaining sections on respiratory symptoms in moderate to severe asthmatic
children, respiratory related hospital admissions, and non-accidental and cardiorespiratory

                                         6-29

-------
mortality include discussions of the scope of the assessments, provide estimates of the incidence,
incidence per 100,000 and percent of total incidence of the effects, and discussion of year-to-year
variability in the estimates and the risk that remains after the current standard is met.

Table 6-2.  Summary of Number and Percent of All School Age Children (5-18) in 12
            Urban Areas Estimated to Experience Lung Function Responses and the
            Number of Occurrences1 Associated with 8-Hour Ozone Exposures While
            Engaged in Moderate Exertion2 for 2002, 2003, and 2004 Air Quality and Just
            Meeting the Current 8-Hour Standard3
Air Quality
Scenario
Recent Air
Quality (2002)
Just Meeting
Current Standard
Recent Air
Quality (2003)
Just Meeting
Current Standard
Recent Air
Quality (2004)
Just Meeting
Current Standard
FEVi
Children
(% of Children,
5-18)
1,240,000
(7%)
570,000
(3%)
900,000
(5%)
310,000
(2%)
570,000
(3%)
210,000
d%)
> 15%
Occurrences
[% reduction
from recent air
quality]
5,670,000
2,900,000
[49% reduction]
4,190,000
1,760,000
[58% reduction]
3,360,000
1,520,000
[55% reduction]
FEVi
Children
(% of Children,
5-18)1
490,000
(3%)
150,000
(0.8%)
330,000
(2%)
70,000
(0.4%)
170,000
(0.9%)
40,000
(0.2%)
> 20%
Occurrences
[% reduction
from recent air
quality]
1,210,000
430,000
[64% reduction]
820,000
200,000
[76% reduction]
530,000
150,000
[72% reduction]
Estimates for persons and person days greater than 10,000 were rounded to the nearest 10,000. Estimates for
persons and person days less than 10,000 were rounded to the nearest thousand. Percentages less than 1 are rounded
to the nearest tenth, percentages greater than or equal to 1 are rounded to the nearest percent.

2Moderate exertion is defined as having an 8-hr average equivalent ventilation rate > 13 1-min/m2.

3Estimates are the aggregate results based on 12 urban areas (Atlanta, Boston, Chicago, Cleveland, Detroit, Houston,
Los Angeles, New York, Philadelphia, Sacramento, St. Louis, and Washington, D.C.) based on Tables 5-6 and 5-7
for estimates based on 2002 and 2004 air quality and tables provided in Post (2007) for estimates based on 2003 air
quality. Estimates are for the ozone season which is all year in Houston, Los Angeles and Sacramento and March or
April to September or October for the remaining urban areas.


       Lung function  decrements in all school-age children.  Tables 5-6 and 5-7 in Chapter 5

display risk estimates for all school age children (ages 5-18) for moderate or greater lung

function decrement responses for the 12 urban areas.  As with the exposure estimates, the risk

estimates are associated with three years of recent air  quality (i.e., 2002, 2003,  2004) and  air

quality based on adjusting these same three years to simulate just meeting the current 0.08 ppm,

8-hr Os standard based on the three-year design value. All estimates in both tables reflect
                                           6-30

-------
responses associated with exposure to 63 in excess of exposures associated with policy relevant
background (PRB) Oj concentrations.7 Table 5-6 shows the number and percent of all school
age children estimated to have at least one moderate or greater lung function response (i.e., >
15% decrement in FEVi) during the 63 season. Table 5-7 displays the total number of
occurrences for the moderate or greater lung function responses during the Os season.  Table 6-2
draws upon the risk estimates for all school age children contained in tables in Chapter 5 and the
Risk Assessment TSD, and provides the number of children estimated to experience one or more
occurrences of moderate or greater (i.e., > 15% decrement in FEVi) and large or greater (i.e., >
20% decrement in FEVi) lung function responses, and the number of total days of occurrences,
aggregated across all 12 urban areas.
       As shown in Table 6-2, for the three recent years (2002 - 2004), from 570,000 to over 1.2
million school age children (3 to 7% of all school age children) are estimated to experience 1 or
more moderate lung function responses (i.e., > 15% reduction in FEVi) in the 12 urban areas
combined. Similar to the exposure estimates discussed above, when air quality is adjusted to
simulate just meeting the current 8-hr standard, there are significant reductions in estimated
health outcomes. Depending on which year is adjusted to just meet the current 8-hr standard,
from about 210,000 to about 570,000 children (1 to 3% of all school age children) are estimated
to experience moderate (i.e., > 15% reduction in FEVi) lung function responses in these 12 urban
areas combined upon just meeting the current standard. Among all school age children, these
estimates  indicate that the percent of children likely to experience one or more moderate or
greater lung function decrements (i.e., > 15% FEVi decrement) is reduced by  about 57 to 66%
when the current standard is just met based on adjusting for 2002, 2003, and 2004 air quality.
The analogous reduction in the number of occurrences of lung function decrements > 15% is 49
to 58%.
       It is also important to note that many of these children will experience repeated
occurrences of moderate or greater lung function responses. These results indicate that among
all school age  children, on average an individual is likely to experience 5 to 7  occurrences of
moderate  or greater lung function responses during an Oj season.8 However, based on the
distribution of exposures estimated from the 1997 review, it is reasonable to expect that many
       7 With respect to the impact of uncertainties about estimated background concentrations on the risk
estimates presented in Tables 5-6 and 5-7, as discussed in section 5.4.3, alternative assumptions about PRB levels
had a variable impact depending on the location and standard analyzed in terms of the absolute magnitude of the risk
estimates.  However, the overall pattern of reductions, expressed in terms of percentage reduction relative to the
current standard, is not impacted by alternative assumptions for PRB for most of the alternative standards analyzed.

       8This number is estimated for example for all children, by dividing the estimated number of children into
the estimated number of occurrences, resulting in an average of about 5 to 7 occurrences per child.
                                          6-31

-------
children will experience one or just a few moderate or greater lung function responses, while a
smaller number of children will experience large numbers of such responses.  These ranges of
estimated number of occurrences of moderate or greater lung function decrements in an O?,
season are important in considering the implications for the health status of individuals likely to
experience these effects.
       As discussed in section 3.6.3, for active healthy people, large lung function responses
(i.e., > 20% decrement in FEVi), would likely interfere with normal activities in many sensitive
individuals, therefore single occurrences would be considered to be adverse under ATS
guidelines, and are the appropriate indicator to consider.  As shown in Table 6-2, for the three
recent years (2002 - 2004), there were estimated to be from about 1,210,000 to about 530,000
occurrences of large lung function decrements (i.e., > 20% reduction in FEVi) in school age
children in the 12 urban areas combined.  Similar to the exposure estimates discussed above,
when air quality is adjusted to simulate just meeting the current 8-hr standard, there are
significant reductions in estimated health outcomes.  Depending on which year is considered,
upon just meeting the current 8-hr standard, there are estimated to be from about 430,000 to
about 150,000 occurrences of large lung function decrements (i.e., > 20% reduction in FEVi) in
school age children in these 12 urban areas combined.  Among all school age children, these
estimates indicate that occurrences of large lung function decrements (i.e.,  > 20% FEVi
decrement) are reduced by about 64 to 76% when the current standard is just met based on
adjusting 2002, 2003, and 2004 air quality.
       Lung function decrements in asthmatic school age children.  As discussed in greater
detail in section 3.6.3, FEVi decrements > 10% but < 20% have been judged to represent
moderate levels of functional responses for active healthy people and would likely interfere with
normal activity for relatively few sensitive individuals. However, for persons with lung disease,
such as asthma, lung function decrements at the lower end of the moderate range (i.e., FEVi
decrements > 10%) would likely interfere with normal activity for many individuals and would
likely result in additional and more frequent use of medication.  We also note that new evidence
described above indicates that children with asthma, particularly those with moderate-to-severe
asthma, are more likely to have lung function and symptomatic responses, and have bigger
responses, from Os exposure than children who do not have asthma. Studies discussed in section
3.3.1.1 that show increased lung function responses, inflammation, and increased airway
responsiveness in asthmatics indicate that the risk estimates for lung function decrements derived
from controlled exposures of healthy adult volunteers likely underestimate the percent of
asthmatic school age children that would experience decrements in FEVi.  In this final Staff
Paper, we use a > 10% decrease in FEVi as a benchmark for moderate functional responses in
asthmatic children.  The CASAC endorsed this approach.  Thus, as discussed in  section 5.4.1,

                                         6-32

-------
consistent with the CASAC recommendation (Henderson, 2006c) that for asthmatic children a
lung function response defined in terms of FEVi decrement > 10% serves as an indicator of
potential adverse health effects for this group, risk estimates have been calculated for lung
function decrements (i.e., > 10% and > 20% reductions in FEVi) in asthmatic school age
children for 5 urban areas that are a subset of the 12 urban areas included for all children.
       As shown in Table 6-3, for the three recent years (2002 - 2004), from about 150,000 to
about 240,000 asthmatic school age children (9 to 15% of asthmatic school age children) are
estimated to experience 1 or more moderate lung function responses (i.e., > 10% reduction in
FEVi) in the 5 urban areas  combined.  Similar to the risk estimates for all school age children
discussed above, when air quality is adjusted to simulate just meeting the current 8-hr standard,
there are significant reductions in estimated health outcomes.  Depending on which year is
adjusted to just meet the current 8-hr standard, from about 60,000 to about 120,000 children (4 to
7% of asthmatic school age children) are estimated to experience moderate (i.e., > 10% reduction
in FEVi) lung function responses in these 5 urban areas combined upon just meeting the current
standard. Among asthmatic school age children, these estimates indicate that the number of
children likely to experience one or more moderate or greater lung function decrements (i.e., >
10% FEVi decrement) drops by about 50 to 60% when the current standard is just met based on
adjusting 2002, 2003, and 2004 air quality.
       Asthmatic school age children are estimated to experience a greater number of repeated
occurrences of moderate or greater lung function responses  per individual responding compared
to all school age children. The results in Table 6-3 indicate that among asthmatic school age
children, on average an individual is likely to experience  8 to  11 occurrences of moderate or
greater lung function responses during an Os  season. As  discussed above, the more likely
distribution is that many children will  experience one or only a few occurrences of moderate or
greater lung function decrements (> 10% decrement in FEVi), while some may experience a
very large number, based on these estimates.  Recognizing that nationally over 14% of school
age children have asthma, these numbers raise concern about the potential number of children
with asthma who could experience a large number of occurrences of moderate or greater lung
function decrements (> 10% decrement in FEVi) even with air quality just meeting the current 8-
hr standard.
                                        6-33

-------
Table 6-3. Summary of Number and Percent of Asthmatic School Age Children (5-18) in 5
           Urban Areas Estimated to Experience Lung Function Responses and the
           Number of Occurrences1 Associated with 8-Hour Ozone Exposures While
           Engaged in Moderate Exertion2 for 2002, 2003, and 2004 Air Quality and Just
           Meeting the Current 8-Hour Standard3
Air Quality
Scenario
Recent Air
Quality (2002)
Just Meeting
Current Standard
Recent Air
Quality (2003)
Just Meeting
Current Standard
Recent Air
Quality (2004)
Just Meeting
Current Standard
FEVi>
Children
(% of Asthmatic
Children in)1
240,000
(15%)
120,000
(7%)
200,000
(13%)
90,000
(5%)
150,000
(9%)
60,000
(4%)
10%
Occurrences1
[% reduction
from recent air
quality]1
1,820,000
990,000
[46% reduction]
1,590,000
730,00 0
[54% reduction]
1,360,000
660,000
[51% reduction]
FEVi>
Children
(% of Asthmatic
Children)1
40,000
(3%)
10,000
(0.7%)
40,000
(2%)
6,000
(0.3%)
20,000
(1%)
3,000
(0.2%)
20%
Occurrences1
[% reduction
from recent air
quality]1
110,000
30,000
[73% reduction]
90,000
20,000
[78% reduction]
60,000
10,000
[83% reduction]
Estimates for persons and person days greater than 10,000 were rounded to the nearest 10,000. Estimates for
persons and person days less than 10,000 were rounded to the nearest thousand. Percentages less than 1 are rounded
to the nearest tenth, percentages greater than or equal to 1 are rounded to the nearest percent.
2Moderate exertion is defined as having an 8-hr average equivalent ventilation rate > 13 1-min/m2

3Estimates are the aggregate results based on 5 urban areas (Atlanta, Chicago, New York, Houston, and Los
Angeles) based on Tables 5-8 and 5-9 for estimates based on 2002 and 2004 air quality and Tables 3-8, 3-9, 3-24, 3-
26, and 3-27 in the Risk Assessment TSD for estimates based on 2003 air quality. Estimates are for the ozone
season which is all year in Houston, and Los Angeles, and March or April to September or October for the
remaining urban areas.
       As discussed in section 3.6.3, for people with asthma, large lung function responses (i.e.,
> 20% decrement in FEVi), would likely interfere with normal activities for most individuals
and would also increase the likelihood that these individuals would use additional medication or
seek medical treatment.  Single occurrences would be considered to be adverse to the individuals
and would be cause for concern.  As shown in Table 6-2, for the three recent years (2002 - 2004),
there were estimated to be from about  110,000 to about 60,000 occurrences of large lung
function responses (i.e., > 20% reduction in FEVi) in asthmatic school age children in the 5
urban areas combined. When air quality is adjusted to simulate just meeting the current 8-hr
standard, there are significant reductions in estimated health outcomes. Depending on which
                                           6-34

-------
year is adjusted to just meet the current 8-hr standard, there are estimated to be from about
30,000 to about 10,000 occurrences of large lung function responses (i.e., > 20% reduction in
FEVi) in asthmatic children in these 5 urban areas combined.  Among asthmatic school age
children, these estimates indicate that occurrences of large lung function decrements (i.e., > 20%
FEVi decrement) in these 5 urban areas are reduced by about 73 to 83% when the current
standard is just met based on  adjusting 2002, 2003, and 2004 air quality.9
       Respiratory symptoms in moderate to severe asthmatic children. Risk estimates were
developed for several respiratory symptoms (i.e., chest tightness, shortness of breath, and
wheeze) during the 63 season in children of age 0 to  12 years with moderate to severe asthma (as
defined by the use of maintenance asthma medications), living in the Boston area.10 About 40%
of the children with asthma in the Boston area are estimated to be on controller medications and
would be considered moderate-to-severe asthmatics.11 In this population of 25,000 children with
moderate-to-severe asthma, as shown in Tables 5-10 and 5-11, the estimated incidence of
symptom days of chest tightness (across 4 models reflecting 2 different lags and Os alone vs.
inclusion of PM2.5 in the model) ranges from almost 6,900 to 10,800 based on a year (2002) with
poorer air quality, and from 5,300 to 8,400 based on a year with better (2004) air quality.
       As indicated in Figure 5-10, the current standard reduces the incidence of symptom days
for chest tightness by relatively small and consistent amounts across the 4 models specified.
Risk estimates for the other symptom endpoints, shortness of breath and wheeze, show similar
patterns as the risk estimates  for chest tightness. The reduction of risks across the 4 models for
chest tightness is shown in Table 6-4. Averaging the median estimates of symptom days
indicates that just meeting the current 8-hr standard is estimated to reduce the total number of
symptom days for chest tightness in children with moderate to severe asthma by 11% (8,700 to
7,700) based  on a year (2002) with generally poorer air quality and by 15% (from 6,700 to
5,700) based  on a year (2004) with generally better air quality. The current standard clearly does
not provide the same degree of protection against respiratory symptoms in moderate to severe
       9Risk estimates for all and asthmatic children school age children (i.e., > 15% and > 10% reductions in
FEVi, respectively) are discussed in more detail in Chapter 5 and in the Risk Assessment TSD. Risk estimates for
large lung function decrements in all and asthmatic children (i.e., > 20% reductions in FEVi), are included in the
Risk Assessment TSD.

       10To minimize uncertainty, this risk assessment was performed for the Boston area because that is the urban
area nearest to where the epidemiological study was conducted that is the basis for the exposure-response function
used in the assessment.

       nThe estimated percent of asthmatic children using maintenance medications (40%) was obtained via email
4-05-06 from Jeanne E. Moorman, Survey Statistician, National  Center for Environmental Health, CDC.  The email
communication has been placed in the docket for this review.
                                          6-35

-------
asthmatic children as it provides against moderate or greater lung function decrements in all
children.
       Looking at percent of total incidence of symptom days, even after the current 8-hr
standard is met in a year with generally better air quality, among children with moderate to
severe asthma in the Boston area, as many as one symptom day in 8 during the Os season is
estimated to be attributable to 63 exposure. In a year with generally poorer air quality, as many
as one symptom day in 6 is estimated to be attributable to O3 exposure.  These results support the
human clinical and epidemiological evidence that people with asthma are more likely to
experience effects related to 63 exposure than the general population, and provide evidence that
the current 8-hr Os standard is not as protective for children with moderate to severe asthma in
the Boston area as it is for all school age children in the 12 urban areas evaluated.
       Respiratory-related hospital admissions in New York City.  For unscheduled hospital
admissions, risk estimates for the New York City area12 associated with Oj levels above
background for the period from April to September are shown in Table 6-5 for recent air quality
(2002, 2004), and for just meeting the current 8-hr  standard based  on adjusting a recent 3-year
period (2002-2004). The current 8-hr standard reduces the incidence of respiratory-related
hospital admissions by about 16% in a year with poorer air quality (2002) and about 18% in a
year with better air quality (2004).  The incidence of asthma-related hospital admissions (which
are a subset of total respiratory hospital admissions) were reduced by about the same amount in
each of the two scenarios. This results in an incidence per 100,000 of 4.6 to 6.4 for respiratory-
related hospital admissions, and 3.9 to 5.5  for asthma-related hospital admissions, based on two
air quality years, after the current standard is met.
       12To minimize uncertainty, this risk assessment was performed for the New York City area because that is
where the epidemiological study was conducted that is the basis for the exposure-response function used in the
assessment.
                                          6-36

-------
Table 6-4. Incidence of Respiratory Symptom Days for Chest Tightness Associated with
           Recent (2004, 2002) Air Quality and Just Meets the Current Standard Based on
           Adjusting 2004 and 2002 Air Quality in Moderate to Severe Asthmatic Children
           in Boston, MA
Respiratory
Symptoms in
Moderate to Severe
Asthmatic Children
on Controller Meds
Incidence
Incidence per
100,000
Percent of Total
Incidence
Year
2002
2004
2002
2004
2002
2004
Average Incidence of Chest Tightness Associated with
Air Quality
(range of median estimates)1'2
[% reduction from recent air quality]
Recent Air Quality
8,700
(6,900 - 10,800)
6,700
(5,300 - 8,400)
34,400
(27,200 - 42,700)
26,400
(20,700-33,100)
12% - 20%
9%- 15%
Just Meets 0.08 ppm
7,700 [11% reduction]
(6,100-9,600)
5,700 [15% reduction]
(4,500 - 7,200)
30,600
(24,100-38,100)
22,600
(17,700-28,400)
11%- 17%
8% -13%
1 Incidence rounded to nearest 100. Percentages rounded to nearest tenth.
2 Range of median estimates across models using lag 0 and lag 1 day and O3 only and including PM2 5 in the model.
       Total non-accidental and cardiorespiratory mortality in 12 urban areas.  Table 6-6
summarizes risk estimates for non-accidental mortality in 12 urban areas associated with Os
levels above background for the period from April to September based on the 95-city function
reported in Bell et al. (2004) for non-accidental mortality. This table includes risks for two
recent years of air quality (2002 and 2004) and risks associated with just meeting the current 8-hr
standard over a recent 3-year period (2002-2004).13 We chose to present the multi-city function
risk estimates here because they are available for all 12 urban areas,  while single-city estimates
       13The information presented in Table 6-6 is based on Tables 5-10 and 5-11 in this Staff Paper which
summarize the risk estimates for non-accidental mortality in 12 urban areas for recent air quality (2002, 2004) and
Tables 5C-13 to 5C-16 in Appendix 5C of this Staff Paper and Tables 4-15 and 4-18 of the Risk Assessment TSD
which present risk estimates for just meeting the current 8-hr standard based on adjusting the 3-year period (2002-
2004).
                                           6-37

-------
are only available for 7 of the 12 urban areas and because the multi-city risk estimates are less
subject to publication bias.  In comparing estimates between recent air quality and just meeting
the current standard, similar patterns are seen in terms of relative reductions regardless of
whether single- or multi-city functions are used.14  Across the 12 urban areas, the estimates of
mortality incidence per 100,000 relevant population range from 0.4 to 2.8 (for 2002) and from
0.5 to 1.5  (for 2004).  Meeting the current standard results in  a reduction of the incidence per
100,000 to a range of 0.3 to 2.4 based on adjusting 2002 air quality and a range of 0.3 to 1.2
based on adjusting 2004 air quality. Estimates for cardiorespiratory mortality show similar
patterns (Tables 5-14,  5-15).
       14Additional risk estimates for cardiorespiratory mortality are included in the Risk Assessment TSD for 8 of
the 12 urban areas based on Huang et al. (2005).
                                           6-38

-------
Table 6-5. Risks of Respiratory- and Asthma-related Hospital Admissions Associated with Recent (2004, 2002) Air Quality
           and Air Quality Adjusted to Just Meets Current Standard Based on Adjusting 2004 and 2002 Air Quality in New
           York City, NY
Unscheduled
Hospital
Admissions
Respiratory
Asthma
(subset of
respiratory)
Air
Quality
Scenario
Recent
Just Meets
0.08 ppm
Recent
Just Meets
0.08 ppm
Incidence1
(Range)2
[% reduction from recent]
2002
610
(150-1070)
510
(120-900)
[16% reduction]
520
(110-930)
440
(90 - 780)
[16% reduction]
2004
450
(110-790)
370
(90 - 640)
[18% reduction]
380
(80-680)
310
(70 - 560)
[18% reduction]
Incidence per 100,000
(Range)2
2002
7.6
(1.8-13.3)
6.4
(1.5-11.3)
6.5
(1.4-11.6)
5.5
(1.2-9.8)
2004
5.6
(1.4-9.8)
4.6
(1.1-8)
4.8
(1-8.5)
3.9
(0.8 - 7)
Percent Total Incidence
(Range)2
2002
1.7%
(0.4 - 3%)
1.5%
(0.4 - 2.6%)
4%
(0.8-7.1%)
3.3%
(0.7 - 6%)
2004
1.3%
(0.3 - 2.2%)
1%
(0.3 - 1.8%)
2.9%
(0.6-5.2%)
2.4%
(0.5 - 4.3%)
Incidence rounded to the nearest 10. Incidence per 100,000 and percent of total incidence rounded to the nearest tenth.

2Numbers in parentheses are 95% confidence or credible intervals based on statistical uncertainty surrounding the O3 coefficient
                                           6-39

-------
Table 6-6. Risks of Non-accidental Mortality Associated with Recent (2004, 2002) Air Quality and Air Quality Adjusted to
          Just Meets Current Standard Based on Adjusting 2004 and 2002 Air Quality
Location
Atlanta
Boston
Chicago
Cleveland
Detroit
Houston
Air Quality
Scenario
Recent
Just Meets 0.08 ppm
Recent
Just Meets 0.08 ppm
Recent
Just Meets 0.08 ppm
Recent
Just Meets 0.08 ppm
Recent
Just Meets 0.08 ppm
Recent
Just Meets 0.08 ppm
Estimated Risk of Non-accidental Mortality1'2
Incidence
franeeY3
2002 2004
17
(6 - 29)
14
(5 - 23)
10
(3 - 17)
9
(3 - 15)
69
(23-115)
55
(18-91)
38
(13-64)
31
(10-52)
29
(10-48)
24
(8 - 39)
14
(5-24)
9
(3 - 15)
12
(4 - 20)
9
(3 - 15)
7
(2 - 12)
6
(2-9)
49
(16-81)
33
(11-55)
17
(6-28)
12
(4-20)
17
(6-28)
12
(4-20)
17
(6- 28)
11
(4 - 18)
Incidence per 100,000
franeeY3
2002 2004
1.2
(0.4 - 1.9)
0.9
(0.3 - 1.6)
1.5
(0.5-2.5)
1.3
(0.4-2.1)
1.3
(0.4-2.1)
1
(0.3 - 1.7)
2.8
(0.9-4.6)
2.2
(0.8-3.7)
1.4
(0.5-2.3)
1.1
(0.4 - 1.9)
0.4
(0.1-0.7)
0.3
(0.1-0.4)
0.8
(0.3 - 1.4)
0.6
(0.2 - 1)
1.0
(0.3 - 1.7)
0.8
(0.3 - 1.4)
0.9
(0.3 - 1.5)
0.6
(0.2 - 1)
1.2
(0.4-2)
0.9
(0.3 - 1.4)
0.8
(0.3 - 1.4)
0.6
(0.2 - 1)
0.5
(0.2 - 0.8)
0.3
(0.1-0.5)
Percent of Total Incidence
(raneeY3
2002 2004
0.4%
(0.1-0.6%)
0.3%
(0.1-0.5%)
0.4%
(0.1-0.7%)
0.3%
(0.1-0.6%)
0.3%
(0.1-0.5%)
0.3%
(0.1-0.4%)
0.5%
(0.2 - 0.9%)
0.4%
(0.1-0.7%)
0.3%
(0.1-0.5%)
0.3%
(0.1-0.4%)
0.2%
(0.1-0.3%)
0.1%
(0% - 0.2%)
0.3%
(0.1-0.4%)
0.2%
(0.1-0.3%)
0.3%
(0.1-0.5%)
0.2%
(0.1-0.4%)
0.2%
(0.1-0.4%)
0.2%
(0.1-0.3%)
0.2%
(0.1-0.4%)
0.2%
(0.1-0.3%)
0.2%
(0.1-0.3%)
0.1%
(0 - 0.2%)
0.2%
(0.1-0.3%)
0.1%
(0% - 0.2%)
                                      6-40

-------
Location
Los Angeles
New York
Philadelphia
Sacramento
St. Louis
Washington, DC
Air Quality
Scenario
Recent
Just Meets 0.08 ppm
Recent
Just Meets 0.08 ppm
Recent
Just Meets 0.08 ppm
Recent
Just Meets 0.08 ppm
Recent
Just meets 0.08 ppm
Recent
Just Meets 0.08 ppm
Estimated Risk of Non-accidental Mortality1'2
Incidence
franeeY3
2002 2004
110
(37 - 184)
52
(17 - 86)
105
(35 - 174)
84
(28 - 139)
37
(12-62)
30
(10-50)
23
(8 - 39)
18
(6 - 30)
6
(2 - 10)
5
(2-8)
15
(5 - 25)
14
(5 - 23)
133
(45-221)
67
(22-111)
60
(20 - 100)
43
(15-72)
23
(8 - 38)
17
(6 - 28)
18
(6 - 29)
12
(4-21)
3
(1-5)
2
(1-4)
8
(3 - 14)
7
(2 - 12)
Incidence per 100,000
(raneeY3
2002 2004
1.2
(0.4 - 1.9)
0.5
(0.2-0.9)
1.2
(0.4 - 2)
0.9
(0.3 - 1.6)
2.4
(0.8-4.1)
2
(0.7-3.3)
1.9
(0.6-3.2)
1.5
(0.5-2.4)
1.7
(0.6-2.8)
1.4
(0.5-2.3)
2.6
(0.9-4.4)
2.4
(0.8-3.9)
1.4
(0.5-2.3)
0.7
(0.2 - 1.2)
0.7
(0.2- 1.1)
0.5
(0.2 - 0.8)
1.5
(0.5 - 2.5)
1.1
(0.4 - 1.8)
1.4
(0.5 - 2.4)
1
(0.3 - 1.7)
0.9
(0.3 - 1.5)
0.7
(0.2- 1.1)
1.5
(0.5 - 2.4)
1.2
(0.4-2.1)
Percent of Total Incidence
(ranee)
2002 2004
0.4%
(0.1-0.7%)
0.2%
(0.1-0.3%)
0.3%
(0.1-0.6%)
0.3%
(0.1-0.4%)
0.5%
(0.2 - 0.8%)
0.4%
(0.1-0.6%)
0.6%
(0.2 - 0.9%)
0.4%
(0.1-0.7%)
0.3%
(0.1-0.5%)
0.2%
(0.1% -0.4%)
0.6%
(0.2 - 0.9%)
0.5%
(0.2 - 0.8%)
0.5%
(0.2 - 0.8%)
0.2%
(0.1-0.4%)
0.2%
(0.1-0.3%)
0.2%
(0.1-0.3%)
0.3%
(0.1-0.5%)
0.2%
(0.1-0.3%)
0.4%
(0.1-0.7%)
0.3%
(0.1-0.5%)
0.2%
(0.1-0.3%)
0.1%
(0%- 0.2%)
0.3%
(0.1-0.5%)
0.3%
(0.1-0.4%)
1 All results are for mortality (among all ages) associated with short-term exposures to O3.  All results are based on single-pollutant model from Bell et al. (2004)
95-cities model.
Incidence was quantified down to estimated policy relevant background levels. Incidences are rounded to the nearest whole number; incidences per 100,000
relevant population and percents are rounded to the nearest tenth.
3Note: Numbers in parentheses are 95% confidence or credible intervals based on statistical uncertainty surrounding the O3 coefficient.
                                                 6-41

-------
       6.3.1.3   CASAC and Public Commenters' Views on the Adequacy of the Current
              Standard
       Beyond the evidence- and risk/exposure-based information discussed above, staff has
also taken into account the comments and advice of CASAC (Henderson, 2006a, b, c), based on
their review of the CD and earlier drafts of this document and the related technical support
documents, as well as comments on earlier drafts of these documents provided by public
commenters.15  The range of views summarized here generally reflects differing judgments as to
the relative weight to place on various types of evidence, the exposure- and risk-based
information, and the associated uncertainties, as well as differing judgments about the
importance of various Os-related health effects from a public health perspective.
       In a letter to the Administrator (Attachment 1), the CASAC O3 Panel, with full
endorsement of the chartered CASAC,  unanimously concluded that there is "no scientific
justification for retaining" the current primary Os standard, and the current standard "needs to be
substantially reduced to protect human health, particularly in sensitive subpopulations"
(Henderson, 2006c, pp.  1-2).16 The Panel's rationale for this conclusion is outlined in their
letter, beginning with their conclusion that "new evidence supports and build-upon key, health-
related conclusions drawn in the 1997  Ozone NAAQS review." (id., p. 3). The Panel points to
studies discussed in Chapter 3 and Appendix 3B of this document in noting that several new
single-city  studies and large multi-city  studies have provided more evidence for adverse health
effects at concentrations lower than the current standard, and that these epidemiological studies
are backed-up by evidence from controlled human exposure studies.  The Panel  specifically
noted evidence from the recent Adams  (2006) study that reported statistically significant
decrements in the lung function of healthy, moderately exercising adults at a 0.08 ppm exposure
level, and importantly, also reported  adverse lung function effects in some individuals at 0.06
ppm.  In concluding that these results indicate that the current standard "is not sufficiently
health-protective with an adequate margin of safety," the Panel noted that that while similar
studies in sensitive groups such as asthmatics have yet to be conducted, "people with asthma,
and particularly children, have been found to be more sensitive and to experience larger
       15All written comments submitted to the Agency are available in the docket for this rulemaking, as are
transcripts of the public meetings held in conjunction with CASAC's review of earlier drafts of this document and of
draft and final versions of the CD on which this document is based.
       16Comments of individual Panel members are available as Attachment D to the CASAC O3 Panel letter
(Henderson, 2006c); the letter without its attachments is reproduced as Attachment 1 to this document and the
attachments to the letter, including individual Panel member comments, can be found in the docket and online at
http://www.epa.gov/sab/pdf/casac-07-001 .pdf.
                                          6-42

-------
decrements in lung function in response to ozone exposures than would healthy volunteers
(Mortimer etal, 2002)." (Henderson, 2006c, p. 4).
       The CASAC Panel also highlighted a number of adverse health effects, beyond lung
function decrements, that are attributable to low-concentration exposure to ambient 63, below the
level of the current standard, based on a broad range of epidemiological and controlled exposure
studies (id.).  These adverse health effects include increases in school absenteeism, respiratory
hospital emergency department visits among asthmatics and patients with other respiratory
diseases, hospitalizations for respiratory illnesses, symptoms associated with adverse health
effects (including chest tightness and medication usage, and mortality (non-accidental,
cardiorespiratory deaths) reported at exposure levels well below the current standard.  "The
CASAC considers each of these findings to be an important indicator of adverse health effects"
(id.).  The Panel further noted that the risk assessment (discussed above in chapter 5) estimated
that beneficial reductions in some adverse health effects would occur upon meeting the lowest
standard level (0.064 ppm) considered in the assessment.
       With regard to the justification discussed in the second draft of this Staff Paper for
consideration of retaining the current standard,17 the CASAC Panel felt that more emphasis
should be placed on numbers of subjects in controlled human exposure studies with FEVi
decrements greater than 10%, which can be clinically significant, rather than on the relatively
small average decrements.  The Panel also emphasized significant  Os-related inflammatory
responses and markers of injury to the epithelial lining of the lung  that are independent of
spirometric responses. Further, the Panel expressed the view that the justification for
considering retaining the current standard discussed in the earlier draft of this document did not
place enough emphasis on serious morbidity (e.g., hospital admissions) and mortality observed
in epidemiology studies.  On the basis of the large amount of recent data evaluating adverse
health effects at levels at and below the current 63 standard, it was the unanimous opinion of the
CASAC that the current primary O3 standard is not adequate to protect human health, that the
relevant scientific data do not support consideration of retaining the current standard, and that the
current standard needs to be substantially reduced to be protective  of human health, particularly
in sensitive subpopulations (id., pp. 4-5).
       Further, the CASAC letter noted that "there is no longer significant scientific uncertainty
regarding the CASAC's conclusion that the current 8-hr primary NAAQS must be lowered" (id.,
p. 5). The Panel noted that a "large body of data clearly demonstrates adverse human health
effects at the current level" of the standard, such that "[Retaining  this standard would continue
       17See second draft O3 Staff Paper (U.S. EPA, 2006c, e.g., p. 6-50) for the justification referred to by
CASAC. In the second draft Staff Paper staff concluded that "consideration could be given" to retaining the current
standard, as discussed below in section 6.3.1.4 which presents staff's final conclusions.
                                          6-43

-------
to put large numbers of individuals at risk for respiratory effects and/or significant impact on
quality of life including asthma exacerbations, emergency room visits, hospital admissions and
mortality" (id.).  The Panel also noted that "scientific uncertainty does exist with regard to the
lower level of ozone exposure that would be fully protective of human health," concluding that
"it is possible that there is no threshold for an ozone-induced impact on human health and that
some adverse events may occur at policy-relevant background" (id.).
       Consistent with the advice of CASAC, several public commenters supported revising the
primary Oj standard to provide increased public health protection.18 In considering the available
evidence as a basis for their views, these commenters generally noted that the controlled human
exposure studies, showing statistically significant declines in lung function, increases in
respiratory symptoms, airway inflammation and airway responsiveness at a 0.08 ppm exposure
level, were conducted with healthy adults, not members of sensitive groups including people
with asthma and active children generally.  Further, recognizing the substantial variability in
response between subjects, some of these commenters felt that the number of subjects included
in these studies was too small to ascertain the full range of responses, especially for sensitive
groups.  Such considerations in part were the basis for these commenters' view that an Os
standard set at 0.08  ppm is not protective of public health and has no margin of safety for
sensitive groups.
       In considering the results of the human exposure and health risk assessment, this group of
commenters generally expressed the view that these assessments substantially underestimate the
public health impacts of exposure to Os.  For example, several commenters noted that the
assessments are done for a limited number of cities, they  do not address risks to important
sensitive subpopulations (e.g., outdoor workers, active people who spend their summers
outdoors, children up to 5 years of age), and they do not include many health effects that are
important from a public health perspective (e.g., school absences, restricted activity days).
Further, some of these commenters expressed the view that the primary O3 standard should be set
to protect the most exposed and most vulnerable groups,  and the fact that some children are
frequently indoors, and thus at lower risk, should not weigh against setting a standard to protect
those children who are active outdoors.  To the extent the exposure and risk estimates are
considered, some of these commenters felt that primary consideration should be given to the
estimates based on 2002 air quality, for which most areas had relatively higher 63 levels than in
2004, so as to ensure public health protection even in years with relatively worse Os air quality
levels.  Some commenters also felt that the exposure analysis should focus on exposures of
       18 This group of commenters included a public health advocacy group, a medical association, a State
agency, and a regional State organization.
                                         6-44

-------
concern down to at least 60 ppb, the lower end of the range of alternative standards advocated by
the CAS AC Panel during their public meeting in August 2006.
       In contrast to the views discussed above, several other public commenters supported
retaining the current standards.19 In considering the available evidence as a basis for their views,
these commenters challenged a number of aspects of the interpretation of the evidence presented
in the CD.  For example, some of these commenters asserted that EPA generally overestimates
the magnitude and consistency of the results of short-term exposure epidemiological studies
(e.g., for respiratory symptoms, school absences, hospital admissions, mortality), mistakenly
links statistical significance and consistency with strength of associations, and underestimates the
uncertainties in interpreting the results of such studies. Further, these commenters generally
express the view that there is significant uncertainty related to the reliability of estimates from
time-series studies, in that ambient monitors do not provide reliable estimates of personal
exposures,  such that the small reported morbidity and mortality risks are unlikely to be
attributable to people's exposures to 63. Rather, these commenters variously attribute the
reported risks to the inability of time series studies to account for key model specification factors
such as smoothing for time-varying parameters, meteorological factors,  removal of Os by
building ventilation systems, and confounding by co-pollutants. In particular,  these commenters
generally asserted that reported associations between short-term Os exposure and mortality  are
not causal,  in that the reported relative risks are too small to provide a basis for inferring
causality and the associations are most likely due to confounding, inappropriately specified
statistical models,  or publication bias.
       In considering the results of the human exposure and  health risk assessment, this group of
commenters generally expressed the view that these assessments are based on  a number of
studies that should not be used in quantitative risk assessment.  For example, some commenters
variously asserted that the results of time-series studies should not be used at all in quantitative
risk assessments, that risk estimates from single city time-series studies  should not be used since
they are highly heterogeneous and influenced by publication  bias, and that risk estimates from
multi-city studies should not be used in estimating risk for individual  cities. This group of
commenters also generally expressed the view that the assessments generally overestimate the
public health impacts of exposure to 03. Noting that the risk assessment used  a nonlinear
exposure-response function to estimate decreased lung function risks, some commenters
expressed the view that a nonlinear approach should also be used to assess other acute morbidity
effects and mortality. This view was in part based on judgments that it is not possible to
determine if thresholds exist using time-series analyses and that the lack of association  of O3 to
       19This group of commenters included industry associations and corporations.
                                          6-45

-------
mortality in the winter season is highly supportive of the likelihood of the existence of an effect
threshold. With regard to the risk assessment based on controlled human exposure studies of
lung function decrements, some commenters expressed the view that the assessment should not
rely on what they characterized as "outlier" information to define exposure-response
relationships, with reference to the data in the Adams (2006) study at the 0.06 and 0.04 ppm
exposure levels, but rather should focus on group central tendency response levels. Further,
some commenters expressed the view that the air quality rollback algorithm used may result in
overestimates in benefits from emission reductions.  Some commenters noted that potential
beneficial effects of 63 in shielding from UV-B radiation are not quantified in the assessment,
and that the assessment should discuss the evidence for both adverse and beneficial effects with
the same objectivity. Finally, some of these commenters asserted that since estimates of
exposures of concern (defined as 0.080 ppm) and lung function decrements are substantially
below the estimates available when the current Os standard was set in 1997, retaining the current
standard is the most appropriate policy alternative.

       6.3.1.4  Staff Conclusions  on the Adequacy of the Current Standard
       As discussed above, we have considered new evidence from controlled human exposure,
lexicological, and epidemiological studies as well as estimates of Os-related exposures of
concern and risks upon meeting the current 63 standard in many urban areas across the U.S.,
together with associated uncertainties.  As an initial matter, we note that there is general
agreement among staff, CASAC, and all interested parties who commented on earlier drafts of
this document that this information  supports consideration of a primary 63 standard that is at
least as protective as the current standard, with no one supporting consideration of an Os
standard that is any less protective.
       In considering whether the current standard is adequate or should be revised to provide
increased public health protection, we first note that in the second draft of this Staff Paper,
retention of the current standard was included among the policy options that we identified for
consideration.  At that time, while we recognized that there was substantial evidence  that could
be interpreted as calling into question the adequacy of the current standard, we chose not to
exclude the option of retaining the current standard from the provisional conclusions  presented in
that draft document. We wanted to have the benefit of receiving CASAC views and public
comments on the strength of the available evidence, the results of the exposure/risk assessments,
and their interpretation of the evidence with regard to judging the adequacy of the current
standard before reaching final conclusions. As discussed below, based on the available
information and taking into account the views of CASAC and public comments, we now
conclude that the overall body of evidence clearly calls into question the adequacy of the  current
                                         6-46

-------
standard in protecting sensitive groups, notably including asthmatic children and other people
with lung disease, as well as all children and older adults, especially those active outdoors, and
outdoor workers,20 against an array of adverse health effects that range from decreased lung
function to serious indicators of respiratory morbidity including ED visits and hospital
admissions for respiratory causes, and possibly cardiovascular effects and mortality. We believe
the available information provides strong support for consideration of an 63 standard that would
provide increased health protection for these sensitive groups.
       In discussing information related to the adequacy of the current standard, we have noted
that evidence of a range of respiratory-related morbidity effects  seen in the last review has been
strengthened, both through toxicological and controlled human exposure (see Chapter 3, e.g.,
Table 3-1 and Appendix 3C) studies as well as through many new panel and epidemiological
studies (see Chapter 3, e.g., Figure 3-4 and Appendix 3B). In addition, new evidence identifies
people with asthma as an important susceptible population for which estimates of respiratory
effects in the general population may underestimate the magnitude or importance of the effect.
New evidence about mechanisms of toxicity helps to explain the biological plausibility of O3-
induced respiratory effects and is beginning to suggest mechanisms that may link Os exposure to
cardiovascular effects.  Further, there is now relatively strong evidence for associations between
Os and total nonaccidental and cardiopulmonary mortality, even after adjustment for the
influence of season and PM. Relative to the information that was available to inform the
Agency's 1997 decision to set the current standard, the newly available evidence increases our
confidence that a broad array of adverse health effects, especially indicators of respiratory
morbidity, are causally related to Os exposures, and that mortality is likely associated with Os
exposures during the 63 season.
       In examining the entire body of evidence and considering CASAC's views and advice on
interpreting the evidence with regard to the adequacy of the current standard, we conclude that
there is important new evidence demonstrating that exposures to O3 at levels below the level of
the current standard cause or are clearly associated with  a broad array of adverse health effects in
sensitive populations. For example, we note new direct  evidence of transient and reversible lung
function effects and respiratory symptoms in some healthy individuals at exposure levels below
the level of the current standard. In addition, there is now epidemiological evidence of
statistically  significant (Vrelated associations with lung function and respiratory symptom
effects, respiratory-related ED visits and hospital admissions, as well as possibly increased
mortality, in areas that likely would have met the current standard.  There are also many
       20 In defining sensitive groups this way we are including both groups with greater inherent sensitivity and
those more likely to be exposed.
                                          6-47

-------
epidemiological studies done in areas that likely would not have met the current standard but
which nonetheless report statistically significant associations that generally extend down to
ambient Os concentrations that are well below the level of the current standard. Further, there
are a few studies that have examined subsets of data that include only days with ambient 63
concentrations below the level of the current standard, or below even much lower Os
concentrations, and continue to report statistically significant associations.  Our level of
confidence in the findings from these studies is not related to whether they were done in areas
that likely would or would not have met the current standard.  However, we agree with the views
expressed by the CASAC 63 Panel21 that uncertainty in epidemiological findings increases at the
low end of the ranges of concentrations observed in these studies, which are generally well
below the level of the current standard, because of limitations in interpreting the results that
relate, for example, to poor correlations between ambient concentrations and personal exposure
and to questions of plausibility that are more salient at relatively low concentrations.
       Based on the strength of the currently available evidence of adverse health effects,
especially indicators of respiratory morbidity, and on the extent to which the evidence indicates
that such effects likely  result from exposures to ambient Os concentrations well below the level
of the current standard, we conclude that the available evidence clearly calls into question the
adequacy of the current standard and provides strong support for giving consideration to revising
the standard to provide increased protection, especially for sensitive groups, against a broad
array of adverse health effects. As discussed below, we have also considered the results of the
exposure and risk assessments conducted for this review to provide some quantitative
perspective on the extent to which sensitive groups are likely to experience exposures of concern
and on the risk of experiencing various adverse health effects when air quality is adjusted to
simulate meeting the current standard in a number of urban areas in the U.S.
       In considering the results of the exposure and risk assessments, we first note that the
CASAC Panel has expressed the view that the exposure analysis represents a  state of the art
modeling approach, that the risk  assessment is well-done and balanced, and that the results of
both are appropriate input to the  decision on the 63 NAAQS.  Moreover, the additional
uncertainty and sensitivity analyses conducted after CASAC review of the second draft Staff
Paper have increased our overall confidence in the results of these assessments. Accordingly, in
considering the adequacy of the current standard, we have placed substantial weight on these
results, both as  direct measures of a limited set of Os-related risks to  public health and as
indicators of the potential for a range of other types of adverse health effects for which currently
available information is too limited to allow for direct estimates of risk.
       21 See, for example, the written comments of Dr. Vedal (Henderson, 2006c, Appendix D, p. D-74).
                                          6-48

-------
       Turning to the results of the exposure assessment, we note that estimating exposures of
concern provides an important indication of the potential magnitude of the incidence of health
outcomes that we cannot currently evaluate in a quantitative risk assessment, such as, increased
airway responsiveness, increased pulmonary inflammation, including increased cellular
permeability, and decreased pulmonary defense mechanisms.  These physiological effects, which
have been demonstrated to occur in healthy people at 63 exposures as low as 0.080 ppm, are
associated with aggravation of asthma, increased medication use, increased school and work
absences, increased susceptibility to respiratory infection, increased visits to doctors' offices and
EDs, increased admissions to hospitals, and possibly to cardiovascular system effects and
chronic effects such as  chronic bronchitis or long-term damage to the lungs that can lead to
reduced quality of life.  In considering whether the current standard provides a margin of safety
against such serious respiratory morbidity effects not just in healthy adults but in sensitive
groups, such as people  with asthma or other lung diseases, children, and older adults, we believe
it is appropriate to consider the extent to which the current standard reduces exposures  of
concern not only at and above the 0.080 ppm benchmark level, but more importantly also at
lower benchmark levels.  Therefore, we have focused  on the extent to which the current standard
reduces exposures of concern at the 0.070 and 0.060 ppm benchmark  levels, noting that 0.060
ppm is the lowest level at which potentially adverse lung function decrements have been
observed in healthy people. While we believe that exposures of concern at these lower
benchmark levels are an important indicator of the potential for adverse health effects especially
in sensitive groups, as discussed above, we note that due to individual variability in
responsiveness only a subset of individuals in these groups with exposures of concern can be
expected to experience such adverse health effects.
       Based on the aggregate estimates summarized  above in Table  6-la-c for the  12 U.S.,
urban areas included in the exposure analysis, we first note that there  is substantial year-to-year
variability across the three years included in this analysis, ranging up  to an order of magnitude or
more, in estimates of the number of people and the number of occurrences of exposures of
concern at each of the benchmark levels. We believe it is appropriate to consider not just the
average estimates across all years, but also to consider public health impacts in years with
relatively poorer air quality. In so doing, we note that even when considering the benchmark
level of > 0.080 ppm, an exposure level at which adverse respiratory effects have been
demonstrated in healthy adults, approximately 100,000 asthmatic children (and over 600,000
total children) in these  12 cities alone are estimated to experience such exposure levels in the
worst of the three years when the current standard is met. In looking  at the lower benchmark
levels that are more relevant to providing a margin of  safety for sensitive groups, over 400,000
thousand (-18%) asthmatic children (and close to 3 million total children) in these 12 cities are
                                         6-49

-------
estimated to experience exposures > 0.070 ppm in the worst of the three years; even in the mid-
year, over 100,000 (-5%) asthmatic children (and over 700,000 total children) are estimated to
experience such exposures.  These estimates are roughly 2 to 4 times higher when considering
exposures at the benchmark level of > 0.060 ppm.
       We also note that there is substantial city-to-city variability in these estimates, as
summarized in Table 6-7, and we believe it is appropriate to consider not just aggregate
estimates across all cities, but also to consider public health impacts in cities that receive
relatively less protection from the current standard.  For example, in considering the benchmark
level of > 0.070 ppm, while the aggregate percentage of asthmatic children estimated to
experience such exposures of concern across all 12 cities is ~ 5% in the mid-year and -18% in
the worst year when the current standard is met, these estimates range up to 12% and 38%,
respectively, in the city with the least degree of protection from the current standard. As seen in
Table 6-7, such percentages are substantially higher when considering exposures at the
benchmark level of > 0.060 ppm, ranging up to 37 to 65% in the mid to worst years. Estimates
of the percent of all children exposed are generally similar or slightly lower than those for
asthmatic children.

Table 6-7.  Estimates  of Percent of Children Exposed While at Moderate Exertion to 8-
           Hour Daily Maximum Ozone Concentrations > 0.070 ppm  and > 0.060 ppm
           Combined for 12 Urban Areas in the U.S., and the Range of Estimates for Each
           of the 12 Cities - Just Meeting Current Standard
Exposure of
Concern
Benchmark
Level
> 0.070 ppm
> 0.060 ppm
Percent of All Children (5-18 yrs old)
(18.3 million children) -
Aggregated across 1 2 cities
(Range for each of 1 2 cities)
2002
16
(1-37)
40
(7-64)
2003
4
(1-10)
19
(8 - 38)
2004
1
(0-5)
8
(1-22)
Percent of Asthmatic Children (5-18 yrs old)
(2.6 million children) -
Aggregated across 1 2 cities
(Range for each of 1 2 cities)
2002
18
(1 - 38)
44
(7-65)
2003
5
(2-12)
21
(8-37)
2004
1
(0-5)
8
(1 - 22)
       With regard to estimates of risks of health effects in sensitive populations likely to remain
upon meeting the current standard, we note that some such estimates related to relatively less
serious lung function effects are now appreciably lower than in the last review, whereas risk
                                        6-50

-------
estimates related to more serious effects, such as hospital admissions, are as high or higher than
previously estimated.  In addition, unlike in the last review, there is now evidence that supports
estimating risks for respiratory symptoms in asthmatic children and Os-related mortality.
       Based on Tables 6-2 and 6-3, we note that meeting the current 63 standard substantially
reduces the estimated risk of moderate lung function decrements (i.e., > 15% FEVi decrement)
in all school age children across 12 urban areas.  In asthmatic children the reduction in the
estimated risk of moderate lung function decrements (i.e., > 10% FEVi decrement) is not as
large, with about 4 to 7% of asthmatic school age children estimated to experience one or more
occurrences of moderate lung function decrements even when the current standard is met,
resulting in almost 1 million occurrences just in 5 urban areas in a year with relatively poorer air
quality (2002).  Moreover, the estimated number of occurrences of moderate or greater lung
function decrements per child is on average approximately 5  to 7 in all children and 8 to 11 in
asthmatic children in an Os season, even when the current standard is met.  In the 1997 review  of
the Os standard a general consensus view of the adversity of such moderate responses emerged
as the frequency of occurrences increases, with the judgment that repeated occurrences of
moderate responses, even in otherwise healthy individuals, may be considered adverse since they
may well set the stage for more serious illness.
       Large lung function decrements (i.e., > 20% FEVi decrement) would likely interfere with
normal activities in many healthy individuals, therefore single occurrences would be considered
to be adverse. In people with asthma, large lung function responses (i.e., > 20% FEVi
decrement), would likely interfere with normal activities for most individuals and would also
increase the likelihood that these individuals would use additional medication or seek medical
treatment. Not only would single occurrences be considered to be adverse to asthmatic
individuals under the ATS definition, but they also would be cause for medical concern. While
the current standard reduces the occurrences of large lung function decrements in all children and
asthmatic children overall from about 60 to 80%, in a year with relatively poorer air quality
(2002) there are estimated to be more than 400,000 occurrences in all school children across  12
urban areas, and more than 30,000 occurrences in asthmatic children across just 5 urban areas.
As noted above,  it is clear that even when the current standard is met over a three-year period, air
quality in each year can vary considerably, as evidenced by relatively large differences between
risk estimates based on 2002 to 2004 air quality.  We believe it is appropriate to consider this
yearly variation in air quality allowed by the current standard in judging the extent to which
impacts on members of sensitive groups in a year with relatively poorer air quality remains of
concern from a public health perspective.
       As seen in Tables 6-4 through 6-6, risks of respiratory symptom days in moderate to
severe asthmatic children, respiratory-related hospital admissions, and non-accidental and
                                         6-51

-------
cardiorespiratory mortality, respectively, are not reduced to as great an extent by meeting the
current standard as are lung function decrements. For example, just meeting the current standard
reduces the estimated average incidence of chest tightness in moderate to severe asthmatic
children living in the Boston urban area by 11 to 15%, based on adjusting 2002 and 2004 air
quality, respectively, resulting in an incidence per 100,000 relevant population of approximately
23,000 to 31,000 children, attributable to 63 exposure (Table 6-4). The current standard reduces
the estimated incidence of respiratory-related hospital admissions in the New York City urban
area by 16 to 18%, based on adjusting 2002 and 2004 air quality, respectively, resulting in an
incidence per 100,000 population of approximately 4.6 to 6.4, respectively (Table 6-5). Across
the 12 urban areas considered in this assessment, the estimates of non-accidental mortality
incidence per 100,000 relevant population range from 0.4 to 2.6 (for 2002) and 0.5 to 1.5 (for
2004) (Table 6-6). Meeting the current standard results in a reduction of the estimated incidence
per 100,000 population to a range of 0.3 to 2.4 based on adjusting 2002 air quality and a range of
0.3 to 1.2 based on adjusting 2004 air quality. Estimates for cardiorespiratory mortality show
similar patterns.
       Staff notes that in considering the estimates of the proportion of population affected and
the number of occurrences of the health effects that are included in the risk assessment, these
limited estimates are indicative of a much broader array of Os-related health endpoints that are
part of a "pyramid of effects" that include various indicators of morbidity that could not be
included in the risk assessment (e.g., school absences, increased medication use, ED visits) and
which primarily  affect members of sensitive groups. While we had sufficient information to
estimate and consider the number of symptom days in children with moderate to severe asthma,
we recognize that there are many other effects that may be associated with symptom days, such
as increased medication use, school and work absences, or visits to doctors' offices, that we did
not have sufficient information to estimate but are important to consider in assessing the
adequacy of the current standard. The same is true  for more serious, but less frequent effects.
We estimated hospital admissions, but we did not have sufficient information to estimate ED
visits in a quantitative risk assessment.  Consideration of such unquantified risks for this array of
health effects, in conjunction with risk estimates for health effects that we did quantify, leads us
to conclude that they are indicative of risks to sensitive groups that can reasonably be judged to
be important from a public health perspective.  These risk-based considerations reinforce our
conclusion that consideration should be given to revising the standard so as to provide increased
public health protection, especially for sensitive groups such as people with asthma or other lung
diseases, as well as children and older adults, particularly those active outdoors, and outdoor
workers.
                                          6-52

-------
       Taking into account the above evidence- and exposure/risk-based considerations, staff
concludes that the body of information that is now available supports consideration of revising
the current primary O?, standard so as to afford greater public health protection, especially to
sensitive groups, and that it does not support retention of the current standard.  The following
sections on indicator, averaging time, level, and form are intended to help inform consideration
of an appropriate range of alternative standards.
       6.3.2   Indicator
       In the last review EPA focused on a standard for O3 as the most appropriate surrogate for
ambient photochemical oxidants. In this review, while the complex atmospheric chemistry in
which 63 plays a key role has been highlighted, no alternatives to 63 have been advanced as
being a more appropriate surrogate for ambient photochemical oxidants.
       It is generally recognized that control of ambient O?, levels provides the best means of
controlling photochemical oxidants of potential health concern. Further, among the
photochemical oxidants, the acute exposure chamber, panel and field epidemiological human
health database provides evidence only for 63 at levels of photochemical oxidants commonly
reported  in the ambient air, in part because few other photochemical oxidants are routinely
measured. However, recent investigations on copollutant interactions have used simulated urban
photochemical oxidant mixes.  These investigations suggest the need for similar studies to help
in understanding the biological basis for effects observed in epidemiological studies that are
associated with air pollutant mixtures, where Os is used as the surrogate for the mix of
photochemical oxidants. Meeting the 63 standard can be expected to provide some degree of
protection against potential health effects that may be independently associated with other
photochemical oxidants but which are not discernable from currently available studies indexed
by 63 alone.  Since the precursor emissions that lead to the formation of 63 generally also lead to
the formation of other photochemical oxidants,  measures leading to reductions in population
exposures to 63 can generally be expected to lead to reductions in population exposures to other
photochemical oxidants.

       6.3.3   Averaging Time

       6.3.3.1  Short-Term and Prolonged (1 to 8 Hours)
       The current 8-hr averaging time for the primary Os NAAQS was set in 1997.  The
decision  to revise the averaging time of the primary standard from 1  to 8 hr was supported by the
following key observations and conclusions (62 FR 38861):
       (1) The 1-hr averaging time of the previous NAAQS was originally selected on the basis
of health effects associated with short-term (i.e., 1- to 3-hr) exposures.
                                         6-53

-------
       (2)  Substantial health effects information was available for the 1997 review that
demonstrated associations between a wide range of health effects (e.g., moderate to large lung
function decrements, moderate to severe symptoms and pulmonary inflammation) and prolonged
(i.e., 6- to 8-hr) exposures below the level of the NAAQS.
       (3)  Results of the quantitative risk analyses showed that reductions in risks from both
short-term and prolonged exposures could be achieved through a primary standard with an
averaging period of either 1 or 8 hr.
       (4)  The 8-hr averaging time is more directly  associated with health effects of concern at
lower 63 concentrations than the 1-hr averaging time. It was thus the consensus of CAS AC "that
an 8-hour standard was more appropriate for a human health-based standard than a 1-hour
standard." (Wolff, 1995)
       In looking at the new information that is discussed in section 7.6.2 of the CD,
epidemiological studies have used various averaging periods for Os concentrations, most
commonly  1-hr, 8-hr and 24-hr averages. As described more specifically below,  in general the
results presented from U.S. and Canadian studies (Appendix 3B) show no consistent difference
for various averaging times in different studies.
       Only a few studies presented results for different 63 averaging periods using the same
data set.  Two of the recent multi-city mortality studies reported associations for multiple
averaging times (Bell et al., 2004; Gryparis et al., 2004). Both reported that the effect estimates
for different averaging times were not statistically different, though the effect estimates for
associations with  1-hr daily maximum Os concentrations were somewhat larger than those for
longer averaging times, especially 24-hr average 03.  In addition, Gent et al., (2003) reported that
associations for 1-hr and 8-hr average 63 with respiratory  symptoms were not significantly
different.
       Among the single-city epidemiological studies, Peters et al. (2001) reported positive, but
not statistically  significant associations  between O3 and the incidence of myocardial infarction
(CD, p. 7-55); this study differs from most since the short-term Os concentration used was the
time period preceding the health event,  not the highest daily short-term average concentration.
The effect estimate for the association with Os averaged over a 2-hr period prior to the
myocardial infarction was substantially larger than that reported for an association with 24-hr
average 63 (Peters et al., 2001). The CD reports results for a number of single-city results that
generally reported effect estimate sizes  that were larger when comparing 1-hr or 8-hr daily
maximum 63 concentrations with the 24-hr concentration, but the results did not differ
statistically (CD, p. 7-120).  The CD observes that the various O3 average concentrations were
generally very highly correlated with one another, so it is not surprising that effect estimates
                                         6-54

-------
would be similar. The CD concludes that the epidemiological study results were generally
comparable for the three Os averaging times (CD, p. 7-120).
       Because the 8-hr averaging time continues to be more directly associated with health
effects of concern from controlled human exposure studies at lower concentrations than do
shorter averaging periods, we have not focused on alternative averaging times in this review and
have not conducted exposure or risk assessments for standards with averaging times other than 8
hours. In its letter to the Administrator, the CASAC O3 Panel supported the continued use of an
8-hr averaging time for the primary Os standard (Henderson, 2006c, p. 2), as did many
commenters.
       Some other commenters expressed the view that consideration should be given to setting
or reinstating a 1-hr standard, in addition to maintaining the use of an 8-hr averaging time, to
protect people in those parts of the country with relatively more "peaky" exposure profiles.
These commenters point out that when controlled exposure studies using triangular exposure
patterns (with relatively higher 1-hr peaks) have been compared to constant exposure patterns
with the same aggregate O3 dose (in terms of concentration x time), "peaky" exposure patterns
are  seen to lead to higher risks. The California Air Resources Board made particular note of this
point, expressing the view that a 1-hr standard would more closely represent actual exposures, in
that many people spend only 1  to 2 hours a day outdoors, and that it would be better matched to
Os concentration profiles along the coasts where Os levels are typically high for shorter
averaging periods than  8 hours.
       In considering the information discussed above, CASAC views and public comments on
the  earlier draft of this Staff Paper, staff concludes that the 8-hr averaging time remains the most
appropriate averaging time for a human health-based standard.  This conclusion is based on  the
observations summarized above, particularly:  (1) the fact that the 8-hr averaging time is more
directly associated with health effects of concern at lower 63 concentrations than are averaging
times of shorter duration and (2) results from quantitative risk analyses showing that attaining  an
8-hr standard reduces the risk of experiencing health effects associated with both 8-hr and shorter
duration exposures. Furthermore, the CASAC 63 Panel unanimously agreed that the health-
based standard should be an 8-hr  average in  1995 (Wolff, 1995) and made no comment in 2006
(Henderson, 2006c) to suggest that any averaging time other than 8-hr was appropriate for the
health-based standard.
       In addition to quantitative risk analyses, we conducted an analysis  of a recent three-year
period of air quality data (2002 to 2004) to determine whether the comparative 1- and 8-hr air
quality patterns that were observed in the last review continue to be observed based on more
recent air quality data.  This updated air quality analysis (McCluney, 2007) is very consistent
with the analysis done in the last review in that it indicates that only two urban areas of the U.S.
                                         6-55

-------
have such "peaky" air quality patterns that the ratio of 1-hr to 8-hr design values is greater than
1.5.  This suggests that based on recent air quality data, it is reasonable to again conclude that an
8-hr average standard at or below the current level would generally be expected to provide
protection equal to or greater than the previous 1-hr standard of 0.12 ppm in almost all urban
areas. Thus, staff concludes that setting a standard with an 8-hr averaging time can effectively
limit both 1- and 8-hr exposures of concern and is appropriate to provide adequate and more
uniform protection of public health from both short-term and prolonged exposures to O3 in the
ambient air. Therefore, we recommend that the 8-hr averaging time be retained and do not
recommend consideration of a separate 1-hr standard at this time.

       6.3.3.2  Long-Term
       During the last review, there was a large animal toxicological  database for consideration
that provided clear evidence of associations between long-term (e.g.,  from several months to
years) exposures and lung tissue damage, with additional evidence of reduced lung elasticity and
accelerated loss of lung function. However, there was no corresponding evidence for humans,
and the state of the science had not progressed sufficiently to allow quantitative extrapolation of
the animal study findings to humans.  For these reasons, consideration of a separate long-term
primary Os standard was not judged to be appropriate at that time, recognizing that the 8-hr
standard would act to limit long-term exposures as well as short-term and prolonged exposures.
       In the current review, long-term animal toxicological studies continue to support the
relationship between Os exposure and structural alterations in several regions of the respiratory
tract and identify the CAR as the most affected region.  In addition, animal toxicological studies
that utilized exposure regimens to simulate seasonal exposure patterns also report increased lung
injury compared to conventional long-term, stable exposures.  (CD, p. 8-85) Collectively, the
evidence from animal studies strongly suggest that 63 is capable of damaging the distal airways
and proximal alveoli, resulting in lung tissue remodeling leading to apparently irreversible
changes. Compromised pulmonary function  and structural changes due to persistent
inflammation may exacerbate the progression and development of chronic lung disease (CD, p.
8-70). Recent epidemiological studies observed that reduced lung function growth in children
was associated with seasonal exposure to 63; however, cohort studies investigating  the effect of
annual or multiyear Os exposure observed little clear evidence for impacts of longer-term,
relatively low-level Os exposure on lung  function development in children.
       Collectively, the epidemiological studies are inconclusive, but suggestive of respiratory
health effects from long-term Os exposure. While there continues to be evidence of structural
changes in the respiratory tract in animal studies, with some very weak support from
epidemiological studies in children, it is highly uncertain as to what long-term patterns of
                                         6-56

-------
exposure or 63 concentrations in humans may be required to produce the morphological changes
found in the animal studies and it is not currently possible to characterize the possible magnitude
or severity of any such effects occurring in humans in response to ambient O?, exposures at levels
observed in the U.S.. Further, to the extent that meeting an 8-hr 63 standard in some cases is
expected to result in lower long-term average concentrations, the 8-hr standard would provide
some protection against effects that may be associated with long-term 63 exposures.
       In its letter to the Administrator, the CASAC Oj Panel offered no views on the long-term
exposure evidence, nor did it suggest that consideration of a primary Os standard with a long-
term averaging time was appropriate (Henderson, 2006c, p. 2). Similarly, no commenters
expressed support for considering such a standard.
       Staff concludes that a health-based standard with a longer-term averaging time than 8
hours is not warranted at this time.  While potentially more serious health effects have been
identified as being associated with longer-term exposure studies of laboratory animals and in
epidemiology studies, there remains substantial uncertainty regarding how these data could be
used quantitatively to develop a basis for setting a long-term health standard. Because long-term
air quality patterns would be improved in areas  coming into attainment with an 8-hr standard, the
potential risk of health effects associated with long-term exposures would be reduced in any area
meeting an 8-hr standard.  Furthermore, the CASAC Os Panel offered no advice either in 1995
(Wolff, 1995) or in 2006 (Henderson, 2006c) that a long-term health-based standard should be
considered. Thus, staff does not recommend consideration of a long-term, health-based standard
at this time.

       6.3.4  Level
       In considering alternative Os standard levels that would provide greater protection against
the array of (Vrelated adverse health effects than that afforded by the current standard, staff has
taken into account both evidence- and exposure/risk-based considerations, as well as the
comments and advice of CASAC and public commenters'  views. The discussion of alternative
levels in this section builds upon the information presented above in the discussion of the
adequacy of the current standard (section 6.3.1) to help inform staffs evaluation of the range of
levels that would be appropriate for consideration.
       As an initial matter, we have considered whether it is appropriate to continue to specify
the level of the Os standard to the nearest hundredth (two decimal places) ppm, or whether the
precision with which ambient 63  concentrations are measured supports specifying the standard
level to the nearest thousandth ppm (i.e.,  to the nearest part per billion (ppb)).  As discussed
above in Chapter 2 (section 2.4.2), staff conducted an analysis to determine the impact of
ambient O3 measurement error on calculated 8-hr average O3 design value concentrations, which
                                         6-57

-------
are compared to the level of the standard to determine whether the standard is attained (Cox and
Camalier, 2006).  The results of this analysis suggest that instrument measurement error, or
possible instrument bias, contribute very little to the uncertainty in design values. More
specifically, measurement imprecision was determined to contribute less than 1 ppb to design
value uncertainty, and a simulation study indicated that randomly occurring instrument bias
could contribute approximately 1 ppb. Staff has interpreted this analysis as being supportive of
specifying the level of the standard to the nearest thousandth ppm. This information was
provided to the  CASAC 63 Panel and made available to the public at the August 24-25, 2006
public meeting. The Panel concluded that current monitoring technology "allows accurate
measurement of Os concentrations with a precision of parts per billion" and recommended that
the specification of the level of the Os standard should reflect this degree of precision
(Henderson, 2006c).22  Based on these considerations, staff recommends that consideration be
given to specifying the level of an alternative 8-hr Os standard to the nearest thousandth ppm. If
the current standard were to be specified to this degree of precision, the current standard would
effectively be at a level of 0.084 ppm, reflecting the data rounding conventions that are part of
the definition of the current 0.08 ppm 8-hr standard.

       6.3.4.1  Evidence-based Considerations
       In taking into account evidence-based considerations,  staff has evaluated available
evidence from controlled human exposure studies and epidemiological studies, as well  as the
uncertainties and limitations in that evidence. In so doing, we focused primarily but not
exclusively on U.S. and Canadian studies. In particular, we have considered  the extent to which
controlled human exposure studies provide evidence of lowest-observed-effects levels and the
extent to which epidemiological studies provide evidence of associations that extend down to the
lower levels of 63 concentrations observed in the studies or some indication of potential effect
thresholds in terms of 8-hr average  Os concentrations.
       In considering the available controlled human exposure studies, as discussed above in
Chapter 3 (section 3.3.1.1), we note that FEVi decrements and various measures of respiratory
symptoms were observed in some healthy adults following a 6.6-hr exposure level of 0.06 ppm
(reflecting exposures of 0.060 + 0.003 ppm) during moderate exertion. More specifically,
Adams (2002) reports that in an earlier study (Adams, 1998) 20% of subjects (6 of 30 subjects)
had notable responses (FEVi decrements > 10%) at the 0.06 ppm exposure level, and data
underlying the Adams (2006) study show that 7% of healthy adult subjects had > 10% FEVi
       22 We also note that the 8-hr O3 standard adopted by the state of California in 2006 is specified to the
nearest thousandth part per million (at a level of 0.070 ppm) (http://www.arb.ca.gov/research/aaqs/ozone-rs/ozone-
rs.htm).
                                         6-58

-------
decrements at the 0.06 ppm exposure level in that study.  Notably, in Adams (2006), total
respiratory symptoms (which includes pain on deep inspiration, shortness of breath, and cough)
following 5.6 and 6.6-hr exposures at 0.06 ppm (during a triangular exposure pattern, that is
more representative of those encountered in summer air pollution episodes than a square-wave
exposure pattern) reached statistical significance.
       In considering the controlled human exposure study results discussed above in the
context of the broader body of controlled human exposure studies, we conclude that these studies
provide evidence of a lowest-observed-effects level of 0.060 ppm for potentially adverse lung
function decrements and respiratory symptoms in some healthy adults while at prolonged
moderate exertion. We further conclude that since people with asthma, particularly children,
have been found to be more sensitive and to experience larger decrements in lung function in
response to 63 exposures than would healthy adults, the 0.060 ppm exposure level also can be
interpreted as representing a level likely to cause adverse lung function decrements and
respiratory symptoms in children with asthma and more generally in people with respiratory
disease.
       In considering controlled human exposure studies of pulmonary inflammation, airway
responsiveness, and impaired host defense capabilities, we note that these studies provide
evidence of a lowest-observed-effects level for such effects in healthy adults at prolonged
moderate exertion of 0.08 ppm (generally reflecting exposures of 0.080 ppm + 0.004 ppm).  As
discussed above, these physiological effects have been linked to aggravation of asthma and
increased  susceptibility to respiratory infection, potentially leading to increased medication use,
increased  school and work absences, increased visits to doctors' offices and EDs, and increased
hospital admissions.  Further, pulmonary inflammation is related to increased cellular
permeability in the lung, which may be a mechanism by which Os exposure can lead to
cardiovascular system effects, and to potential chronic effects such as chronic bronchitis  or long-
term damage to the lungs that can lead to reduced quality of life. These are all  indicators of
adverse Os-related morbidity effects, which are consistent with and lend plausibility to the
adverse morbidity effects and mortality effects observed in epidemiological studies.
       In considering epidemiological studies, we first recognize that the available evidence
neither supports nor refutes the existence of effect thresholds at the population  level for
morbidity and mortality effects. As discussed above, based on a consideration  of studies that
have explored the question of potential thresholds and of seasonal studies that show no consistent
(Vrelated effects during the cold season when 63 concentrations are generally low, we conclude
that if a population threshold level does exist, it would likely be well below the level of the
current Os standard and possibly within the range of background levels. More  specifically, as
discussed above in Chapter 3 (section 3.4.5) and more fully in the CD (Chapter 7, section 7.6.5),
                                         6-59

-------
a number of studies reported some suggestive evidence of possible thresholds for morbidity and
mortality outcomes in terms of 24-hr, 8-hr, and 1-hr averaging times. These results, taken
together, provide some indication of possible 8-hr average threshold levels from below about 25
to 35 ppb up to approximately 50 ppb. Other studies, however, observe linear concentration-
response functions suggesting no effect threshold. In considering this information, we conclude
that the statistically significant associations between ambient O3 concentrations and lung
function decrements, respiratory symptoms, indicators of respiratory morbidity including
increased ED visits and hospitals admissions, and possibly mortality reported in a large number
of studies likely extend down to ambient O3 concentrations that are well below the level of the
current standard. Toward the lower end of the range of O3 concentrations observed in such
studies, ranging down to background levels, however, we conclude that there is increasing
uncertainty as to whether the observed associations remain plausibly related to exposures to
ambient O3, rather than to the broader mix of air pollutants present in the ambient atmosphere.
       We have also considered studies that did subset analyses that include only days with
ambient O3 concentrations below the level of the current standard, or below even lower O3
concentrations, and continue to report statistically significant associations. Notably, as discussed
above, Bell et al. (2006) conducted a subset analysis that continued to show statistically
significant associations even when only days with a maximum 8-hr average O3 concentration
below  a value of approximately 61 ppb were included.23 Also of note is the large multi-city
NCICAS (Mortimer et al., 2002) that reported statistically significant associations between
ambient O3 concentrations and lung function decrements even when days with 8-hr average O3
levels greater than 80 ppb were excluded (which consisted of less than 5% of the days in the
eight urban areas in the study).
       Being mindful of the uncertainties and limitations inherent in interpreting the available
evidence, staff believes that the range of alternative O3 standards appropriate for consideration in
this review should take into account information on lowest-observed-effects levels in  controlled
human exposure studies as well as indications of possible effects thresholds reported in some
epidemiological studies and questions of biological plausibility in attributing associations
observed down to background levels to O3 exposures alone.  Based on the evidence and these
considerations,  we conclude that the upper end of the range of consideration should be somewhat
below  0.080 ppm, the lowest-observed-effects level for effects such as pulmonary inflammation,
increased airway responsiveness and impaired host-defense capabilities in healthy adults while at
prolonged moderate exertion. As discussed above, these physiological  effects have been linked
       23 Bell et al. (2006) referred to this level as being approximately equivalent to 120 ug/m3, daily 8-hr
maximum, the World Health Organization guideline and European Commission target value for O3.
                                         6-60

-------
to aggravation of asthma and increased susceptibility to respiratory infection, potentially leading
to increased medication use, increased school and work absences, increased visits to doctors'
offices and EDs, and increased hospital admissions.  Further, pulmonary inflammation is related
to increased cellular permeability in the lung, which may be a mechanism by which 63 exposure
can lead to cardiovascular system effects, and to potential chronic effects such as chronic
bronchitis or long-term damage to the lungs that can lead to reduced quality of life. These are all
indicators of adverse O3-related morbidity effects, which are consistent with and lend plausibility
to the adverse morbidity effects and mortality effects observed in epidemiological studies
reporting statistically significant associations with ambient 63 concentrations that range down to
levels well below 0.080 ppm. Based on the evidence, we also conclude that the lower end to the
range of alternative Os standards appropriate for consideration should be at least as low as the
lowest-observed-effects level for potentially adverse lung function decrements and respiratory
symptoms in healthy adults, 0.060 ppm, which is also a level likely to cause adverse effects in
sensitive groups, and above the level where there is some indication of possible effects
thresholds in epidemiological studies. In considering a lower end of the range for consideration,
we also recognize that control strategies designed to attain an Os standard set at a particular level
within an urban area, as measured at the monitor reporting the highest 63 design value, would
cause the entire distribution of Os concentrations across the area to be reduced, thus lowering not
only concentrations above the level of the standard but also those below that level as well. Thus,
we believe that it is appropriate to also consider the results of the exposure and risk assessments
that are based on modeling changes in the entire distribution of ambient Os concentrations to
simulate just meeting alternative standards,  discussed below, in reaching conclusions about an
appropriate lower end of the range for consideration.

       6.3.4.2  Exposure/Risk-based Considerations
       In addition to the evidence-based considerations, staff has also considered quantitative
exposures and health risks estimated to occur upon meeting the current and alternative standards
to help inform judgments about a range of standard levels for consideration that could provide an
appropriate degree of public health protection. In so doing, we are mindful of the important
uncertainties and limitations that are associated with the exposure and risk assessments, as
discussed above in section 6.3.1.2 and more fully in Chapters 4 and 5. For example, important
uncertainties affecting the exposure estimates are related to modeling human activity patterns
over an 63 season (especially repetitive exposures), modeling microscale variations in ambient
concentrations, and modeling building air exchange rates. With regard to the risk assessment,
important uncertainties include, for example, those related to exposure estimates (for children
engaged in moderate or greater exertion), as well as those related to estimation of concentration-
                                          6-61

-------
response functions, specification of concentration-response models, the possible role of
copollutants in interpreting reported associations with Os, and inferences of a likely causal
relationship between Os exposure and non-accidental mortality (for risk estimates based on
epidemiological studies).  As noted above, after considering the key uncertainties, the CASAC
Panel expressed the view that the exposure analysis represents a state of the art modeling
approach, that the risk assessment is well-done and balanced, and that the results of both are
appropriate input to the decision on the O3 NAAQS.
       Beyond these uncertainties, we and CASAC also recognize important limitations to the
exposure and risk analyses.  For example, we did not have sufficient information to evaluate all
relevant sensitive groups (e.g., outdoor workers) or all Os-related health outcomes (e.g.,
increased medication use, school absences, ED visits), and the scope of our analyses was
generally limited to estimating exposures and risks in 12 urban areas across the U.S., and to only
five or just one area for some risk analyses.  Thus,  it is clear that national-scale public health
impacts of ambient 63 exposures are much larger than the quantitative estimates of (Vrelated
incidences of adverse health effects and the numbers of children likely to experience exposures
of concern associated with meeting the current or alternative standards.  Taking these limitations
into account, the CASAC advised us not to rely solely on the results of the exposure and risk
assessments in considering alternative standards, but also to place significant weight on the body
of evidence of Os-related health effects in drawing conclusions about an appropriate range of
levels for consideration.  We concur with this important caveat.
       Turning to the results of the exposure assessment, we examine the extent to which
alternative standard levels below the current standard are estimated to reduce exposures of
concern at the 0.070 and 0.060 ppm benchmark levels, for all and asthmatic school age children
in the 12 urban areas included in the assessment. The alternative  standard levels evaluated
include standards set at: 0.080 ppm, 4th daily maximum (i.e., the  80/4 scenario); 0.074 ppm, 5th
daily maximum  (i.e., the 74/5 scenario); 0.074 ppm, 4th daily maximum (i.e., the 74/4 scenario);
0.074 ppm, 3rd daily maximum  (i.e., the 74/3 scenario); 0.070 ppm, 4th daily maximum (i.e., the
70/4 scenario); and, 0.064 ppm, 4th daily maximum (i.e., the 64/4 scenario).24 Exposure
estimates for 14 scenarios are examined (i.e., meeting 6 alternative standard level/form
combinations, based on adjusting 2002 and 2004 air quality for all 6 alternative standards, and
       24
         The abbreviated notation used to identify the current and alternative standards in the figures showing
reductions in risk estimates in this chapter is in terms of ppm and the nth highest daily maximum. For example, the
current standard is identified as "0.084/4." This notation is equivalent to the abbreviated labeling used in Chapters 4
and in the text and tables in this chapter which is in terms of ppb and the nth highest daily maximum (e.g., the
current standard is labeled "84/4").
                                          6-62

-------
adjusting 2003 air quality for the 74/4 and 64/4 alternative standards25) at the 0.060 and 0.070
ppm Os benchmark levels, for all children and asthmatic children. Individual city estimates of
the percent of all children likely to experience exposures of concern are given in the exhibits in
Chapter 4.  Estimates of the percent of asthmatic children, the number of all children and
asthmatic children, and the number of occurrences of exposures of concern, both the aggregate
estimates across the 12 urban areas and the individual city estimates, are shown in Appendix 4A.
The estimates are for the number and percent of all children and asthmatic children exposed, and
the number of person-days (occurrences) of exposures, with daily 8-hr maximum exposures at or
above the 0.060 ppm and the 0.070 ppm benchmark levels while at intermittent moderate or
greater exertion. For the purpose of this discussion, recommending an appropriate range of
levels for consideration, we will focus on scenarios with the same form as the current O?,
standard (i.e. the 80/4, 74/4, 70/4 and 64/4 scenarios) and will address consideration of
alternative forms in the next section.
       As shown in the exhibits in Chapter 4 and Appendix 4A, the percent of population
exposed at any given level is very similar for all and asthmatic school age children.  Substantial
year-to-year variability in exposure estimates is observed, ranging to over an order of magnitude
at the higher alternative standard levels, in estimates of the number of children and the number of
occurrences of exposures of concern at both of the benchmark levels.  For example,  for the 80/4
scenario, almost 2 million children (-10%) (and almost 300,000 asthmatic children) based on
2002 air quality, to more than 80,000 children (<0.5%) (and about 10,000 asthmatic children)
based on 2004 air quality, are estimated to experience one or more exposures of concern at the
benchmark level of > 0.070 ppm 03. For the 74/4 and 64/4 scenarios, we estimated  exposures of
concern at the two benchmark levels, based on air quality in 2003, which was intermediate
between 2002 and 2004. Estimates of exposures of concern for this year are between the
estimates for 2002 and 2004.  Across the alternative standard levels, in the year with poorer air
quality (2002) estimates of the number of all children exposed one or more times ranges from 1.8
million (80/4 scenario) to 23,000 (64/4 scenario); for asthmatic children the range is almost
300,000 (80/4 scenario) to about 4,000 (64/4 scenario), at the benchmark level of > 0.070 ppm
Os. These results suggest reductions of approximately 90% (80/4 scenario) to about 100% (64/4
scenario) across the range of alternative standards in the number of all children  and asthmatic
children at this exposure of concern level.
       The estimates of exposures of concern are considerably larger at the benchmark level of >
0.060 ppm 63, and the pattern of year-to-year variability remains. For example, for  the 80/4
       25 Estimates for exposures of concern for the year 2003 were developed since the second draft of this Staff
Paper for only 2 alternative standard levels (i.e., 74/4 and 64/4) due to time constraints.
                                         6-63

-------
scenario, more than 6 million children (-34%) (> 900,000 asthmatic children) based on 2002 air
quality, to more than 800,000 children (-5%) (and > 100,000 asthmatic children) based on 2004
air quality, are estimated to experience one or more exposures of concern at the benchmark level
of > 0.060 ppm 63. In the year with poorer air quality (2002), across the alternative standard
levels, estimates of the number of all children exposed one or more times ranges from 6.1 million
(80/4 scenario) to almost 900,000 children (64/4 scenario); for asthmatic children the range is
more than 900,000 (80/4 scenario) to more than 100,000 (64/4 scenario) at the benchmark level
of > 0.060 ppm Os. At this benchmark level, in the year with poorer air quality, these results
suggest reductions of about 66% (80/4 scenario) to about 95% (64/4 scenario) in the percent of
children estimated to be exposed across the range of alternative standards.
       We also note that there is substantial city-to-city variability in these estimates, as
summarized in Table  6-8, and we believe it is appropriate to consider not just the aggregate
estimates across all cities, but also to consider the public health impacts in cities that receive
relatively less protection from the alternative standards. For example, in considering the
benchmark level of > 0.070 ppm, for the 74/4 scenario, while the aggregate percentage of all or
asthmatic children estimated to experience such exposures  of concern across all 12 cities is about
4% in the worst year, it ranges up to 13% in the city with the least degree of protection from that
alternative standard.  This pattern of city-to-city variability also occurs at the benchmark level of
> 0.060 ppm 03. While the aggregate percentage of all and asthmatic children estimated to
experience one or more exposures of concern across all 12  cities, for the 74/4 scenario, is about
22 to 25% in the worst year, it ranges up to approximately 48% in the city with the least degree
of protection from that alternative standard.
       Turning to the estimates from the risk assessment, Figures 6-1  through  6-6 show the
percent reduction in risk estimates from just meeting the current standard (the 84/4 scenario) to
just meeting the same alternative standards discussed above, based on adjusting 2002 and 2004
air quality data.  These figures also provide perspective on  the extent to which  the risks in these
recent years (i.e., 2002 and 2004) are greater than those estimated to occur upon meeting the
current standard (in terms of a negative percent reduction relative to the 84/4 scenario). Figures
6-1 and 6-2 show the percent reduction in the numbers of school age children estimated to
experience at least one decrement in FEVi > 15% in each of the 12 urban areas for 2002 and
2004, respectively, and Figure 6-3 and 6-4 show the percent reduction in the number of
asthmatic school age children estimated to experience at least one decrement in FEVi  ^ 10% in
only 5 urban areas for 2002 and 2004, respectively.26 Figures 6-5 and 6-6 show the percent
       26 The health risk assessment for lung function decrements for asthmatic school age children was conducted
for 5 of the 12 urban areas and for a more limited set of alternative standards due to time constraints. The areas
                                          6-64

-------
Table 6-8. Daily Maximum Ozone Concentrations > 0.07 ppm and > 0.06 ppm Combined
           for 12 Urban Areas in the U.S., and the Range of Estimates for Each  of the 12
           Cities - Just Meeting Alternative Standards
Exposure
of Concern
Benchmark
Level
> 0.07 ppm
> 0.06 ppm
Alternative
Standard
Level/
Form
80/4
74/5
74/4
74/3
70/4
64/4
80/4
74/5
74/4
74/3
70/4
64/4
Percent of All Children (5-18 yrs
old)
(18.3 million children) -
Aggregated across 1 2 cities
(Range for each of 1 2 cities)
2002
10
(1 - 27)
6
(0-15)
4
(0-12)
2
(0-8)
1
(0-5)
0
(0-1)
34
(5 - 59)
25
(1-51)
22
(1 - 46)
18
(1-41)
14
(1 - 35)
5
(0-15)
2003


1
(0-2)


0
(0-0)


5
(2-14)


0
(0-1)
2004
0
(0-3)
0
(0-1)
0
(0-1)
0
(0-0)
0
(0-0)
0
(0-0)
5
(1-15)
2
(0-9)
1
(0-7)
1
(0-5)
0
(0-4)
0
(0-1)
Percent of Asthmatic Children (5-18 yrs
old)
(2.6 million children) -
Aggregated across 1 2 cities
(Range for each of 1 2 cities)
2002
11
(1-27)
c
(0-16)
4
(0-13)
2
(0-8)
2
(0-6)
0
(0-1)
37
(5-61)
28
(2 - 52)
25
(1 - 48)
20
(1 - 42)
16
(0 - 36)
5
(0-16)
2003


1
(0-2)


(0-0)


6
(2-14)


0
(0-1)
2004
0
(0-3)
0
(0-1)
0
(0-1)
0
(0-0)
0
(0-0)
0
(0-0)
5
(0-14)
2
(0-8)
1
(0-7)
1
(0-4)
0
(0-2)
0
(0-1)
were selected to be geographically diverse including urban areas that are not meeting the current O3 standard in the
northeast, southeast, deep south, Midwest, and southern California.
                                          6-65

-------
reduction in risk estimates of non-accidental mortality for the 12 urban areas, for 2002 and 2004,
respectively.  The legend under each figure lists the estimated number of cases (and 95%
credible intervals) when Os concentrations just meet the current standard next to the name of
each location.  There were two health outcomes that we evaluated in one city only, respiratory
symptom days in moderate to severe asthmatic children (Boston, Table 6-9) and respiratory-
related hospital admissions (New York City, Table 6-10), because the concentration-response
functions were developed in these cities and we did not want to introduce additional uncertainties
by applying these functions to other locations.  We believe, however, that it is reasonable to
assume that these results would be generally applicable to other locations.
       As shown in Figures 6-1 and 6-2, we first note that just meeting  the 80/4 scenario is
estimated to result in about a 20% reduction in the number of all school age children estimated to
experience moderate lung function decrements (> 15% reduction in FEVi) relative to the current
standard. Reducing the level of the standard to the 74/4 scenario results  in about a 40 to 50%
reduction in the number of all school age children estimated to experience moderate lung
function decrements, depending on whether 2002 or 2004 air quality is the basis for adjustment.
As shown in Figures 6-3 and 6-4, for asthmatic school age children, reducing the level of the
standard to the 74/4 scenario results in about a 25 to 45% reduction in estimated risks across the
5 urban areas relative to the current standard.  An alternative standard set at the 64/4  scenario
provides an appreciably greater reduction of about 65 to 80% in the number of all school age
children estimated to experience moderate lung function decrements (> 15% reduction in FEVi)
relative to the current standard (depending on the year adjusted and the urban area).  This same
64/4 scenario reduces estimates of moderate lung function decrements (> 10% reduction in
FEVi) in asthmatic school age children by about 55 to 65% in most of the areas, with 1 area
having reductions of about 75%.
       As shown in Figures 6-5 and 6-6, we first note that just meeting the 80/4 scenario is
estimated to result in about a 5 to 15% reduction in estimated incidences of Os-related non-
accidental mortality relative to the current standard.  Reducing the level of the standard to the
74/4 scenario results in the estimated incidences of non-accidental mortality being reduced by
about 15 to nearly 40% (depending on the year adjusted and the urban area) relative to the
current standard. Just meeting the 64/4 scenario is estimated to provide appreciably greater
reduction relative to the current standard in some areas, with an estimated reduction of about 30
to 40% in most areas and about 60 to 70% in two areas (depending on the year adjusted and the
urban area) in the estimated incidences of O3-related non-accidental mortality.
                                         6-66

-------
Figure 6-1. Percent Changes in Numbers of School Age Children Experiencing at Least
            One Decrement in FEV1 >15% when O3 Concentrations are Reduced from
            Those Just Meeting the Current Standard to Those that Would Just Meet Each
            Alternative Standard, Based on Adjusting 2002 Data*
             2002 air     0.084/4
              quality
                   0.084/3    0.080/4    0.074/5    0.074/4    0.074/3    0.070/4    0.064/4
                                                Alternative Standard
-Atlanta: 31 (17-47); 3.3% (1.8%-5%)
 Chicago: 67 (37-100); 3.4% (1.9%-5.2%)
-Detroit: 45 (27-66); 4% (2.4%-5.9%)
- Los Angeles: 35 (8 - 63); 1% (0.2%-1.7%)
 Philadelphia: 58 (37-83); 4.9% (3.1%-7%)
 St. Louis: 30 (20-43); 5.2% (3.4%-7.4%)
                                                      -+- Boston: 47 (29 - 68); 4.3% (2.6% - 6.2%)
                                                        <- Cleveland: 27 (17 - 39); 4.6% (2.9% - 6.6%)
                                                      H-Houston: 22 (9-36); 2% (0.8%-3.3%)
                                                      —NewYork: 131 (70-200); 3.1% (1.7%-4.8%)
                                                          Sacramento: 9 (4-14); 2.2% (1%-3.5%)
                                                          Washington, DC:  64 (39 - 93); 4.3% (2.6% - 6.3%)
*The numbers in the box below the figure show for each urban area the number of children estimated to experience
moderate lung function decrements (FEVi > 15%), in thousands (and 95% credible interval) and the percent of
children (and 95% credible interval) estimated to experience these effects when O3 concentrations just meet the
current 0.084/4 8-hr standard. The 8-hr average standards shown in this figure, denoted m/n, are characterized by a
concentration of m ppm and an nth-highest daily maximum form. For example, the current standard is 0.084/4 ~
0.084 ppm, 4th-highest daily maximum 8-hr average.  The figure also compares the current standard to a recent year
of air quality.  The percent change from the current standard (0.084/4) to a recent year of air quality was omitted for
Los Angeles because it was so large in magnitude (-528% in 2002).
                                              6-67

-------
Figure 6-2.  Percent Changes in Numbers of School Age Children Experiencing at Least
            One Decrement in FEV1 >15% when O3 Concentrations are Reduced from
            Those Just Meeting the Current Standard to Those that Would Just Meet Each
            Alternative Standard, Based on Adjusting 2004 Data*
       2004 air    0.084/4    0.084/3    0.080/4     0.074/5     0.074/4     0.074/3    0.070/4
        quality
                                         Alternative Standard
      +Atlanta: 18 (6-31); 1.9% (0.7%-3.3%)
         Chicago: 14 (0-30); 0.7% (0%-1.5%)
      HH Detroit 12(2-23);1.1%(0.2%-2.1%)
      -+- Los Angeles: 33 (5 - 61); 0.9% (0.1 % -1.7%)
      —Philadelphia: 17 (4-30); 1.4% (0.3%-2.5%)
         St. Louis: 7 (1-13);  1.2% (0.2%-2.3%)
-+- Boston: 13(3-25); 1.2% (0.3%-2.3%)
-x-Cleveland:6(1-12);1%(0.1%-1.9%)
+Houston: 21 (8-35); 1.9% (0.8%-3.2%)
—New York: 39 (4 - 77); 0.9% (0.1 % -1.9%)
   Sacramento:4(1-7);1%(0.1%-1.8%)
   Washington, DC: 24 (8 - 42); 1.6% (0.5% - 2.8%)
* The numbers shown in the box below the figure for each urban area represent the number of children estimated to
experience moderate lung function decrements (FEVi > 15%), in thousands (and 95% credible interval) and the
percent of children (and 95% credible interval) estimated to experience these effects when O3 concentrations just
meet the current 0.084/4 8-hr standard. The 8-hr average standards shown in this figure, denoted m/n, are
characterized by a concentration of m ppm and an nth-highest daily maximum form.  For example, the current
standard is 0.084/4 ~ 0.084 ppm, 4th-highest daily maximum 8-hr average. The figure also compares the current
standard to a recent year of air quality. The percent change from the current standard (0.084/4) to a recent year of
air quality was omitted for Los Angeles because it was so large in magnitude (-553% in 2004).
                                             6-68

-------
Figure 6-3. Percent Changes in Numbers of Asthmatic School Age Children Experiencing
           at Least One Decrement in FEV1 >10% when O3 Concentrations are Reduced
           from Those Just Meeting the Current Standard to Those that Would Just Meet
           Each Alternative Standard, Based on Adjusting 2002 Data*
       100%
      -200%
                 2002 air quality
0.084/4            0.074/4
   Alternative Standards
0.064/4
                               •Atlanta: 11  (8 -16); 9.6% (7.2% -13.9%)
                               •Chicago: 26 (20 - 38); 9.4% (7% -13.5%)
                                Houston: 8 (6 -13); 6.2% (4.4% - 9.5%)
                                Los Angeles: 16  (11 - 24); 3.4% (2.5% - 5.3%)
                               •New York: 58 (43 - 85); 9.1% (6.7% -13.3%)
*The numbers shown in the box below the figure show for each urban area the number of asthmatic children
estimated to experience moderate lung function decrements (FEVi > 10%), in thousands (and 95% credible interval)
and the percent of asthmatic children (and 95% credible interval) estimated to experience these effects when O3
concentrations just meet the current 0.084/4 8-hr standard. The 8-hr average standards shown in this figure, denoted
m/n, are characterized by a concentration of m ppm and an nth-highest daily maximum form. For example, the
current standard is 0.084/4 ~ 0.084 ppm, 4th-highest daily maximum 8-hr average.  The figure also compares the
current standard to a recent year of air quality. The percent change from the current standard (0.084/4) to a recent
year of air quality was omitted for Los Angeles because it was so large in magnitude (-275% in 2002).
                                           6-69

-------
   Figure 6-4. Percent Changes in Numbers of Asthmatic School Age Children Experiencing
              at Least One Decrement in FEV1 > 10% when O3 Concentrations are Reduced
              from Those Just Meeting the Current Standard to Those that Would Just Meet
              Each Alternative Standard, Based on Adjusting 2004 Data*
      100%
 O)
 1_
 1_
 3
O

 E
   re
 0)1
   re
 re'
 O)
 if   -150%
 O)
Q_
      -50%
     -100%
     -200%
               2004 air quality
0.084/4            0.074/4
   Alternative Standards
0.064/4
                           -•-Atlanta: 7 (5 -11); 6.2% (4.2% - 9.8%)
                           -•-Chicago: 8 (5 -13); 3% (1.7% - 4.8%)
                           -A-Houston: 8 (6 -13); 6.1% (4.3% - 9.4%)
                           -H- Los Angeles: 16 (11 - 25); 3.4% (2.4% - 5.4%)
                           -*-New York: 24 (14 - 39); 3.7% (2.2% - 6%)
   *The numbers in the box below the figure show for each urban area the number of asthmatic children estimated to
   experience moderate lung function decrements (FEVi > 10%), in thousands (and 95% credible interval) and the
   percent of asthmatic children (and 95% credible interval) estimated to experience these effects when O3
   concentrations just meet the current 0.084/4 8-hr standard.  The 8-hr average standards shown in this figure, denoted
   m/n, are characterized by a concentration of m ppm and an nth-highest daily maximum form.  For example, the
   current standard is 0.084/4 ~ 0.084 ppm, 4th-highest daily maximum 8-hr average.  The figure also compares the
   current standard to a recent year of air quality.  The percent change from the current standard (0.084/4) to a recent
   year of air quality was omitted for Los Angeles because it was so large in magnitude (-281% in 2004).
                                              6-70

-------
Figure 6-5. Percent Changes in OS-Related Non-Accidental Mortality Incidence when O3
           Concentrations are Reduced from Those Just Meeting the Current Standard to
           Those that Would Just Meet Each Alternative Standard, Based on Adjusting
           2002 Data* (Using Bell et al., 2004 - 95 U.S. Cities)
andard
OT
i.
Percent Change from Cu







80% -
60% -
40%
20%
0%
-20%
-40%
-60%
-80%
-100%
-120% -

^X\j»
_^^/
^nn^^^^^^
^^^^^^^^^^^








2002 air 0.084/4 0.084/3 0.080/4 0.074/5 0.074/4 0.074/3 0.070/4 0.064/4



quality
Alternative Standard
—*— Atlanta: 14 (5 - 23); 0.3% (0.1% - 0.5%) — •— Boston: 9 (3 - 15); 0.3% (0.1% - 0.6%)
Chicago: 55 (18 - 91); 0.3% (0.1% - 0.4%) — K— Cleveland: 31 (10 - 52); 0.4% (0.1% - 0.7%)
— *— Detroit: 24 (8 - 39); 0.3% (0.1 % - 0.4%) — •— Houston: 9 (3 - 1 5); 0.1 % (0% - 0.2%)
— 1 — Los Angeles: 52 (17 - 86); 0.2% (0.1% - 0.3%) — New York: 84 (28 - 139); 0.3% (0.1% - 0.4%)
	 Philadelphia: 30 (10 - 50); 0.4% (0.1% - 0.6%) Sacramento: 18 (6 - 30); 0.4% (0.1% - 0.7%)
St Louis: 5 (2 - 8); 0.2% (0.1% - 0.4%) —A— Washington: 14 (5 - 23); 0.5% (0.2% - 0.8%)
*The numbers in the box below the figure show for each urban area the number of cases (and 95% credible interval)
and the percent of total incidence (and 95% credible interval) of O3-related non-accidental mortality when O3
concentrations just meet the current standard (0.084/4).  The 8-hr average standards shown in this figure, denoted
m/n, are characterized by a concentration of m ppm and an nth-highest daily maximum form. For example, the
current standard is 0.084/4 ~ 0.084 ppm, 4th-highest daily maximum 8-hr average. The figure also compares the
current standard to a recent year of air quality.
                                            6-71

-------
Figure 6-6. Percent Changes in OS-Related Non-Accidental Mortality Incidence When O3
            Concentrations are Reduced from Those Just Meeting the Current Standard to
            Those that Would Just Meet Each Alternative Standard, Based on Adjusting
            2004 Data* (Using Bell et al., 2004 - 95       U.S. Cities)
            80%
            60%
     ra
     55
     o
     E
     01
     O)
     c
     ra
     ^
     O
     Ol
     a
          -120%
                 2004 air  0.084/4  0.084/3  0.080/4  0.074/5  0.074/4  0.074/3  0.070/4  0.064/4
                  quality
                                          Alternative Standard
             -Atlanta: 9 (3 -15); 0.2% (0.1% - 0.3%)
             Chicago: 33 (11 - 55); 0.2% (0.1% - 0.3%)
             - Detroit: 12 (4 - 20); 0.1 % (0% - 0.2%)
             - Los Angeles: 67 (22 - 111); 0.2% (0.1 % - 0.4%)
             -Philadelphia: 17 (6 - 28); 0.2% (0.1% - 0.3%)
             St Louis: 2 (1 - 4); 0.1% (0% - 0.2%)
- Boston: 6 (2 - 9); 0.2% (0.1 % - 0.4%)
-Cleveland: 12 (4 - 20); 0.2% (0.1% - 0.3%)
- Houston: 11 (4 - 18); 0.1 % (0% - 0.2%)
- New York: 43 (15 - 72); 0.1 % (0% - 0.2%)
 Sacramento: 12 (4 - 21); 0.3% (0.1% - 0.5%)
-Washington: 7 (2 - 12); 0.3% (0.1% - 0.4%)
   **The numbers in the box below the figure show for each urban area the number of cases (and 95% credible
   interval) and the percent of total incidence (and 95% credible interval) of O3-related non-accidental mortality
   when O3 concentrations just meet the current standard (0.084/4). The 8-hr average standards shown in this
   figure, denoted m/n, are characterized by a concentration of m ppm and an nth-highest daily maximum form.
   For example, the current standard is 0.084/4 ~ 0.084 ppm, 4th-highest daily maximum 8-hr average The figure
   also compares the current standard to a recent year of air quality.
                                             6-72

-------
       With regard to respiratory symptom days for chest tightness in moderate to severe
asthmatic children in the Boston area, as shown in Table 6-9, the alternative standards provide
incremental protection beyond that offered by the current standard.  From the 80/4 scenario to
the 64/4 scenario, the estimated incidence of respiratory symptom days is reduced by 5 to 25% in
the worst year, and by approximately 7 to 31% in the best year. In the worst of the two years, the
estimated percent of total incidence,  or the percent of respiratory-symptom days attributable to
O3 exposure ranges from about 14% for just meeting the current standard to about 10% for the
64/4 scenario.  This means that even under the most stringent alternative standard evaluated, as
many as one symptom day in 10 would be estimated to be attributable to 63 exposure in the 63
season.
       Risk estimates for respiratory-related hospital admissions attributable to Os exposure in
New York City are shown in Table 6-10.  Across the range of alternative standards from the 80/4
scenario to the 64/4 scenario, the estimated number of Os-related hospital admissions declines by
about 6 to 29% in the worst year (2002), and about 7 to 34% in the best year (2004).  The percent
of total respiratory-related hospital admissions attributable to O3 exposure declines from  about
1.5% for the current standard to about 1% or less for the 64/4 scenario.
                                         6-73

-------
Table 6-9. Risks of Respiratory Symptom Days for Chest Tightness Associated with Just
           Meeting the Current and Alternative Ozone Standards Based on Adjusting 2002
           and 2004 Air Quality in Moderate to Severe Asthmatic Children in Boston, MA1
Current and
Alternative
Standards
Current
Standard
(84/4)
80/4
74/4
70/4
64/4
Risk Metric
Incidence
Percent of Total
Incidence
Incidence
Percent of Total
Incidence
Incidence
Percent of Total
Incidence
Incidence
Percent of Total
Incidence
Incidence
Percent of Total
Incidence
Average Risks of Chest Tightness Associated
with Air Quality2'3
[percent reduction from current standard]
2002
7800
14%
7400
[5% reduction]
13%
6800
[12% reduction]
12%
6400
[17% reduction]
12%
5900
[25% reduction]
10%
2004
5700
10%
5400
[7% reduction]
10%
4900
[15% reduction]
9%
4500
[22% reduction]
8%
3900
[31% reduction]
7%
   :It is estimated that there are 25,000 children with moderate to severe asthma in the Boston area.
   Incidence rounded to nearest 100.
   3Average of median estimates for models using lag 0 and lag 1 day and O3 only and including PM2 5 in the
   model.
                                          6-74

-------
Table 6-10. Risks of Hospital Admissions for Respiratory Illness Associated with Just
           Meeting the Current and Alternative Ozone Standards Based on Adjusting 2002
           and 2004 Air Quality in New York, NY
Current and
Alternative
Standards
Current
Standard
(84/4)
80/4
74/4
70/4
64/4
Risk Metric
Incidence
Percent of Total
Incidence
Incidence
Percent of Total
Incidence
Incidence
Percent of Total
Incidence
Incidence
Percent of Total
Incidence
Incidence
Percent of Total
Incidence
Hospital Admissions for Respiratory Illness
Associated with Ozone Exposures1'2
[percent reduction from current standard]
2002
513
1.5%
483
[6% reduction]
1.4%
439
[14% reduction]
1.2%
410
[20% reduction]
1.2%
365
[29% reduction]
1.0%
2004
366
1.0%
341
[7% reduction]
1.0%
304
[17% reduction]
0.9%
278
[24% reduction]
0.8%
241
[34% reduction]
0.7%
  Incidence rounded to nearest whole number.
  295 % credible intervals based on statistical uncertainty surrounding the O3 coefficient are presented in tables in
  Appendix 5C of this Staff Paper.
                                          6-75

-------
       6.3.4.3  CASAC and Public Commenters' Views on the Level of the Standard
       As noted above in section 6.3.1.4, staff recognizes that the exposure- and risk-based
information can be considered both in terms of whether the risks estimated to remain upon
attaining the current standard are important from a public health perspective  and/or whether
additional reductions in risk estimated to be associated with alternative, more protective
standards are important from a public health perspective.  Judgments about the importance of the
estimates of exposures and risks need to take into account the important uncertainties associated
with such estimates. We recognize that public health policy judgments, including the weight to
place on various types of evidence and how to weigh the importance of estimated risks in a
public health perspective, are ultimately decisions left to the Administrator.  To help inform
those judgments with regard to the level of the primary Os standard, the views expressed by
CASAC as well as the views of interested parties who have commented on earlier drafts of this
document are summarized here.
       As stated in its letter to the Administrator, "the CASAC unanimously recommends that the
current primary ozone NAAQS be revised and that the level that should be considered for the
revised standard be from 0.060 to 0.070 ppnf (Henderson, 2006c, p. 5).  The CASAC  coupled
this recommended range of levels with a range of forms, as discussed in the next section below.
This recommendation follows from their more general recommendation, discussed above in
section 6.3.2, that  the current standard of 0.08 ppm needs to be substantially  reduced to be
protective of human health, particularly in sensitive subpopulations.  The lower end of this range
reflects CASAC's views that "[W]hile data exist that adverse health effects may occur  at levels
lower than 0.060 ppm, these data are less certain and achievable gains in protecting human
health can be accomplished through lowering the ozone NAAQS to a level between 0.060 and
0.070 ppm." (id.).
       The same group of commenters that expressed the view that the current primary Os
standard is not adequate also submitted comments that supported revising the level of the
primary O3 standard to within the same or even lower range of levels than the range
recommended by CASAC.  The basis for these commenters' views on the level of the standard is
generally reflected in the discussion above on the basis for their views on the adequacy of the
current standard and in the  rationale given by CASAC. In addition, some of these commenters
also noted that the World Health Organization's guidelines for OB air quality are in the  range of
51 to 61 ppb. The other group of commenters who expressed the view that the current  standard
is adequate did not provide any provisional views on alternative levels that would be appropriate
for consideration should the Administrator consider revisions to the standard.
                                        6-76

-------
       6.3.4.4  Staff Conclusions on the Level of the Standard
       Staffs consideration of alternative levels of the primary O3 standard builds upon our
conclusion, discussed above in section 6.3.1, that the overall body of evidence clearly calls into
question the adequacy of the current standard in protecting sensitive groups, notably including
asthmatic children and other people with lung disease, as well as all children and older adults,
especially those active outdoors, and outdoor workers, against an array of adverse health effects
that range from decreased lung function to serious indicators of respiratory morbidity including
ED visits and hospital admissions for respiratory causes, and possibly cardiovascular-related
effects and mortality. Thus, we believe that the available information provides strong support for
consideration of a range of standard levels that is below the level of the current standard so as to
afford increased health protection for these sensitive groups.  We have also concluded, as an
initial matter, that it is appropriate to consider specifying the level of the 63 standard to the
nearest thousandth ppm.
       As discussed above in section 6.3.4.1., based on the evidence, we conclude that it is
appropriate to consider a range of levels for the primary 63 standard from somewhat below  0.080
ppm down to at least as low as 0.060 ppm. This evidence-based recommendation takes into
account information on lowest-observed-effects levels in controlled human exposure studies as
well  as indications of possible effects thresholds reported in some epidemiological studies and
questions of biological plausibility in attributing associations observed down to background
levels to 63 exposures alone. The upper end of this range is somewhat below the lowest-
observed-effects level for effects such as pulmonary inflammation, increased airway
responsiveness and impaired host-defense capabilities in healthy adults while at prolonged
moderate exertion. These effects have been linked to aggravation of asthma and increased
susceptibility to respiratory infection, potentially leading to increased medication use, increased
school and work absences, increased visits to doctors' offices and EDs, and increased hospital
admissions, and pulmonary inflammation is also related to increased cellular permeability in the
lung, which may be a mechanism  by which Os exposure can lead to cardiovascular system
effects, and to potential chronic effects such as chronic bronchitis or long-term damage to the
lungs that can lead to reduced quality of life. These indicators of adverse O3-related morbidity
effects lend plausibility to the adverse morbidity effects and mortality effects observed in
epidemiological studies reporting  statistically significant associations with ambient 63
concentrations that range down to levels well below 0.080 ppm.  The lower end of this range
reflects the lowest-observed-effects level for potentially adverse lung function decrements and
respiratory symptoms in some healthy adults, 0.060 ppm, which is also a level likely to cause
these adverse effects in sensitive groups, and is above the level where there is some indication of
possible effects thresholds in epidemiological studies.
                                         6-77

-------
       Having reached this evidence-based conclusion on an appropriate range of levels for
consideration, we have also focused on considering the public health implications of selecting
different levels within this range (< 0.08 to 0.060 ppm O^). In so doing, we have looked to the
results of the analyses of exposure and risk for the 74/4 scenario to represent the public health
impacts of selecting a standard level in the upper part of the range, the results of the analyses of
the 70/4 scenario to represent the impacts in the middle part of the range, and the results of the
analyses of the 64/4 scenario to represent the lower part of the range.
       As discussed above in  section 6.3.4.2, for each of these alternative standard levels, we
have considered exposures of concern at the two benchmark levels discussed  above (i.e., 0.070
ppm and 0.060 ppm), that serve as indicators of health outcomes for which there is insufficient
information to do quantitative  risk assessments.  We have also considered the quantitative
estimates of risk for moderate  lung function decrements in all and asthmatic children, respiratory
symptom days in moderate to severe asthmatic children, respiratory-related hospital admissions,
and non-accidental mortality.  In considering both exposures of concern and quantitative risk
estimates, we again note that there is substantial year-to-year variability across the three years
included in this analysis (2002 to 2004) in the estimates of the number of children and the
number of occurrences of exposures of concern at the benchmark levels and in the quantitative
risk estimates. We also note the substantial city-to-city variability in these estimates of
exposures of concern and quantitative risk. We believe that it is appropriate and important to
consider not just the average estimates across all years or all cities included in the analyses, but
to consider the public health impacts in years and locations with relatively poorer air quality and
in locations receiving relatively less protection from any alternative standard.
       We turn now to considering the public health implications of setting the standard in the
upper, middle and lower parts  of the range. A standard set in the upper part of this range (e.g.,
the 74/4 scenario) would result in an aggregate estimate of about 4% of all school age children27
(~ 700,000 children in 12 urban areas) likely to experience exposures of concern at the > 0.070
ppm benchmark level in the worst (2002) of the 3 years evaluated, while the estimates range up
to 12% of all school age children (~ 130,000 children) in the single city with the least degree of
protection from this standard.  In the mid-year (2003), in aggregate about 1% of all school age
children (-93,000 children) are estimated to experience exposures of concern at this level; in the
city with the least degree of protection from this  standard the estimate is less than 2% of all
school age children (~ 24,000  children).  At the benchmark level of > 0.060 ppm,  in aggregate in
       27 We note that the percent of all school age children and asthmatic school age children estimated to
experience exposures of concern (aggregate and individual city estimates) are very similar, and the results for all
school age children are presented in the exhibits in Chapter 4, thus for ease of discussion we present results for all
school age children here.
                                          6-78

-------
the worst year about 22% of all school age children (~ 4 million children) are estimated to
experience exposures of concern; this estimate ranges up to about 46% of all school age children
(~ 1 million children) in the single city with the least degree of protection from this standard.
Even in the mid-year, in aggregate about 5% of all school age children (~ 1 million children) are
estimated to experience exposures of concern, ranging up to 14% of all school age children
(-240,000 children) in the single city with the least degree of protection from this standard.  A
standard set at this level would reduce the number of all and asthmatic school age children
estimated to experience one or more moderate lung function decrements by about 25 to 50%
relative to the current standard (Figures 6-1 through 6-4), with city-to-city  differences accounting
for most of the variability in estimates.  A standard set at this level would reduce non-accidental
mortality by about 10 to 40%, with most of the variability occurring across the 12 city estimates
(Figures 6-5 and 6-6).
       There were two health outcomes that we evaluated in one city  only, respiratory symptom
days in moderate to severe asthmatic children (Boston, Table 6-9) and respiratory-related
hospital admissions (New York City, Table 6-10). In the worst year, a standard set at this level
(the 74/4 scenario) is estimated to reduce the incidence of symptom days in children28 with
moderate to severe asthma in the Boston area to 6,800 days, a 12% reduction relative to the
current standard. Even with this reduction, it is estimated that 1 respiratory symptom day in 8
during the O?, season would be attributable to Os exposure.  Estimated incidence of respiratory-
related hospital admissions was reduced by 14 to 17% by a standard set at  this level relative to
the current standard, in the year with worst and best air quality respectively. A standard set  at
this level reduces exposures of concern at the > 0.070 benchmark level much more  than
exposures of concern at the > 0.06 ppm benchmark level, placing relatively less weight on the
evidence from the controlled human exposure studies showing lung function decrements and
respiratory symptoms in some healthy adults at 0.060 ppm 63, as well as evidence from
epidemiological studies showing an array of respiratory morbidity effects occurring at levels well
below the current standard. It would place relatively more weight on the uncertainties associated
with the exposure and risk estimates, suggesting less importance for the implications of
exposures at the 0.060 ppm benchmark level from a public health policy perspective.
       A standard  set in the middle part of this range  (e.g., the 70/4 scenario) would result in an
aggregate  estimate of about 1% of all school age children (> 200,000 children in 12 urban areas)
likely to experience exposures of concern at the > 0.070 ppm benchmark level even in the worst
year (2002); in the city  with the least protection about 5% of all school age children (~ 56,000
       28 Since there are estimated to be about 25,000 moderate to severe asthmatic children in the Boston area,
this incidence rate is per 25,000 children.
                                         6-79

-------
children) are estimated to experience exposures of concern at this level.29 At the benchmark
level of > 0.060 ppm, in aggregate in the worst year about 15% of all school age children (~ 2.6
million children) are estimated to experience exposures of concern; this number ranges up to
35% of all school age children (~ 530,000 children) in the single city with the least degree of
protection from this standard. A standard set at this level would reduce the number of all school
age children30 estimated to experience one or more moderate lung function  decrements by about
50 to 65% over the current standard, with city-to-city differences  accounting for most of the
variability in estimates.  A standard set at this level would reduce non-accidental mortality by
about 20 to 55%, with most of the variability occurring across the 12 city estimates. In the worst
year, a standard set at this level is  estimated to reduce the incidence of symptom days in children
with moderate to severe asthma in the Boston area only slightly over the standard set at the upper
end of the range, to 6,400 days (a  12% reduction). With this reduction, it is estimated that about
1 respiratory symptom day in 8 during the Os season would be attributable to Os exposure.
Estimated incidence of respiratory-related hospital admissions was reduced by about 20 to 24%
in the year with worst and best air quality, respectively.
       A standard set in the middle part of the recommended range, as indicated by the estimates
for the 70/4  scenario, would reduce the exposures of concern at the 0.070 ppm level substantially
over the current standard, even in the worst of the three years and in the city with the least degree
of protection. However, it reduces exposures of concern at the 0.060 ppm benchmark level much
less so, leaving relatively large percentages of all school age children unprotected in the worst
year or the city with the least protection from this standard.  It provides incremental additional
protection for members of sensitive groups, over the current Os standard, against respiratory
morbidity effects such as lung function decrements, respiratory symptom days and  hospital
admissions,  as well as non-accidental mortality.
       A standard set in the lower part of this range (e.g., the 64/4 scenario) would result in an
aggregate estimate of less than 0.5% of all school age children (~ 23,000 children) likely to
experience exposures of concern at the 0.070 ppm benchmark level in the worst year and only
1% of all school age children  (9,000 children) in the city with the least degree of protection from
this standard. In the mid-year (2003), estimates of exposures of concern go close to zero, even in
the city with the least degree of protection. At the benchmark level of 0.060 ppm, in aggregate
in the worst year about 5% of all school age children (~ 873,000 children) are estimated to
experience exposures of concern; this number ranges up to 15% of all school age children
       29 Estimates were not developed for the mid-year (2003) for this alternative standard.

       30 Estimates for asthmatic children were not developed for this alternative standard.
                                         6-80

-------
(-179,000 children) in the city with the least degree of protection from this standard.  In the mid-
year exposures of concern at this level are reduced substantially, resulting in an aggregate
estimates of less than 0.5% of all school age children (~ 58,000 children), ranging up to only 1%
of all school age children (~ 16,000 children) in the city with the least degree of protection from
this standard. A standard set at this level would reduce the number of all and asthmatic school
age children estimated to experience one or more moderate lung function decrements by about
50 to 80% over the current standard, and non-accidental mortality by about 25 to 75%, with most
of the variability occurring across the 12 city estimates. In the worst year, a standard set at this
level is estimated to reduce the incidence of symptom days in children with moderate to severe
asthma in the Boston area to 5,900 days, about a 25% reduction over the current standard. But
even with this reduction, it is estimated that 1 respiratory symptom day in 10 during the O?,
season is attributable to 63 exposure. Estimated incidence of respiratory-related hospital
admissions was reduced by 30 to 35% over the current standard, in the year with worst and best
air quality respectively.
       These results indicate that setting a standard in the lower part of the range would
essentially eliminate exposures of concern at the benchmark level of 0.070 ppm, even in the
worst of the three  years and in the city with the least degree of protection. It would also
substantially reduce exposures of concern at the benchmark level of 0.060 ppm, especially in the
mid-year of the three years evaluated. It provides additional incremental protection for members
of sensitive groups over the current 63 standard and the alternative standards at the upper to
middle part of the range, against respiratory morbidity effects such as lung function decrements,
respiratory symptom days and hospital admissions, as well as non-accidental mortality.  A
standard set in the lower part of the range would place relatively more weight on the evidence
from the controlled human exposure studies showing lung function decrements and respiratory
symptoms in some healthy adults at 0.060 ppm 63, as well as evidence from epidemiological
studies showing an array of respiratory morbidity effects occurring at levels below the current
standard. It would place relatively less weight on the uncertainties associated with the exposure
and risk estimates, and reflect the greater importance, from a policy perspective, of the public
health implications of exposures at the 0.060 ppm benchmark level.
       The CAS AC recommended a range down to 0.060 ppm for the level of the Os standard,
noting that "achievable gains in protecting public health" (Henderson 2006c, p. 5) can be
accomplished by setting the level of the  standard down to 0.060 ppm Os. The results of the
exposure and risk assessments support this recommendation. Staff concludes that important
improvements in protecting the health of sensitive groups can be made by setting the level of the
Os standard within the range of < 0.080  ppm to 0.060 ppm Os. Standard levels within this range
considered in staff exposure and risk assessments include 0.074, 0.070, and 0.064 ppm, which
                                         6-81

-------
are representative of levels within the upper, middle, and lower parts of this range, respectively.
Moreover, these assessment results indicate that even in the lower end of the range there are
benefits to the health of sensitive groups that warrant consideration.
       To provide some perspective on the implications of alternative 8-hr primary standards
(within the range of levels recommended above and within the range of forms discussed in the
next section below), staff assessed (based on 2002 and 2004 air quality data) the percentage of
counties, and the populations in those counties, that likely would not attain various 8-hr O3
standards.  This assessment, shown in Appendix 6B for various forms  and levels of the 8-hr
standards, was not considered as a basis for the above staff conclusions and recommendations.

       6.3.5  Form
       In 1997 the primary Oj NAAQS was changed from a "1-expected-exceedance" form31 to
a concentration-based statistic, specifically the 3-year average of the annual fourth-highest daily
maximum 8-hr concentrations.  The principal advantage of the concentration-based form is that it
is more directly related to the ambient Os concentrations that are associated with the  health
effects. With a concentration-based form, days on which higher 63 concentrations occur would
weigh proportionally more than days with lower concentrations, since  the actual concentrations
are used in determining whether the standard is attained. That is, given that there is a continuum
of effects associated with exposures to varying levels of 63, the extent to which public health is
affected by exposure to ambient Os is related to the actual magnitude of the Os concentration, not
just whether the concentration is above a specified level.
       In evaluating alternative forms for the primary standard in conjunction with specific
standard levels, staff considers the adequacy of the public health protection provided by the
combination of the level and form to be the foremost consideration. In addition, we  recognize
that it is important to have a form of the standard that is  stable and insulated from the impacts of
extreme meteorological events that are conducive to Os formation. Such instability can have the
effect of reducing public health protection, because frequent shifting in and out of attainment due
of meteorological conditions can disrupt an area's ongoing implementation plans and associated
control programs.  Providing more stability is one of the reasons that EPA moved to  a
concentration-based form in 1997.
       During the 1997 review, consideration was given to a range of alternative forms,
including the second-, third-, fourth- and  fifth-highest daily maximum 8-hr concentrations in an
63  season, recognizing that the public health risks associated with exposure to a pollutant
       31The 1-expected-exceedance form essentially requires that the fourth-highest air quality value in 3 years,
based on adjustments for missing data, be less than or equal to the level of the standard for the standard to be met at
an air quality monitoring site.
                                          6-82

-------
without a clear, discernable threshold can be appropriately addressed through a standard that
allows for multiple exceedances to provide increased stability, but that also significantly limits
the number of days on which the level may be exceeded and the magnitude of such exceedances.
Consideration was given to setting a standard with a form that would provide a margin of safety
against possible, but uncertain chronic effects, and would also provide greater stability to
ongoing control programs.  The fourth-highest daily maximum was selected because it was
decided that the differences in the degree of protection against potential chronic effects afforded
by the alternatives within the range were not well enough understood to use any such differences
as a basis for choosing the most restrictive forms.  On the other hand, the relatively large
percentage of sites that would experience Os peaks well above 0.08 ppm and the number of days
on which the level of the standard may be exceeded even when attaining a fifth-highest 0.08 ppm
concentration-based standard, argued against choosing that form.
        In selecting alternative standards to include in our exposure and risk analyses, we
considered two concentration-based forms, the nth-highest maximum concentration and a
percentile-based form. A percentile-based statistic is useful for comparing datasets of varying
length because it samples approximately the same place in the distribution of air quality values,
whether the dataset is several months or several years long. However, a percentile-based form
would allow more days with higher air quality values in locations with longer Os seasons relative
to places with shorter Os seasons. An nth-highest maximum concentration form would more
effectively ensure that people who live in areas with different length 63 seasons receive the same
degree of public health protection. For this reason, our exposure and risk analyses were based on
a form specified in terms of an nth-highest concentration, with n ranging from 3 to 5.
       The results of some of these analyses are shown in Figures 6-1 through 6-4, discussed
above in section 6.3.4.2. These figures illustrate the estimated percent change in risk estimates
for the incidence of moderate or greater decrements in lung function (> 15% FEVi) in all school
age children and moderate or greater lung function decrements (> 10% FEVi) in asthmatic
school age children, associated with going from meeting the current standard to meeting
alternative standards with alternative forms. Figures 6-5 and 6-6 illustrate the estimated percent
change in the estimated incidence of non-accidental mortality, associated with going from
meeting the current standard to meeting alternative standards.  These results are generally
representative of the patterns found in all of the analyses.  The estimated reductions in risk
associated with different forms of the standard, ranging from third- to fourth-highest daily
maximum concentrations at 0.084 ppm, and from third- to fifth-highest daily maximum
concentrations at 0.074 ppm, are  generally less than the estimated reductions associated with the
different levels that were analyzed. As seen in these figures, there is much city-to-city
variability, particularly in the percent changes associated with going from a fourth-highest to
                                         6-83

-------
third-highest form at the current level of 0.084 ppm, and with estimated reductions associated
with the fifth-highest form at a 0.074 ppm level. In most cities, there are generally only small
differences in the estimated reductions in risks associated with the third- to fifth-highest forms at
a level of 0.074 ppm.
       In their letter to the Administrator, CASAC recommended that "a range of
concentration-based forms from the third- to the fifth-highest daily maximum 8-hr average
concentration" be considered (Henderson, 2006c, p. 5). Further, CASAC  recommended that the
Agency conduct a broader evaluation of alternative concentration-based forms to evaluate the
implications of a broader range of alternative forms on public health protection and stability (i.e.,
with respect to yearly variability to ensure a stable target for control programs).
       The same group of commenters that expressed the view that the current primary Os
standard is not adequate also submitted comments that supported a more health-protective form
of the standard than the current form (e.g., a second- or third-highest daily maximum form). The
other group of commenters who expressed the view that the current standard is adequate did not
provide any provisional views on alternative forms that would be appropriate for consideration
should the Administrator consider revisions to the standard.
       Staff notes that there is not a clear health-based threshold for selecting a particular nth-
highest daily maximum form of the standard from among the ones analyzed.32  We further note
that the changes in the form considered in our analyses result in only small differences in the
estimated reductions in risks in most cities, although in some cities larger differences are
estimated.
       Staff concludes that a range of concentration-based  forms from the third- to the fifth-
highest daily maximum 8-hr average concentration is appropriate for consideration in setting the
standard. Given that there is a continuum of effects associated with exposures to varying levels
of 63, the extent to which public health is affected by exposure to ambient 63 is related to the
actual magnitude of the O3 concentration, not just whether the concentration is above a  specified
level. The principal advantage of a concentration-based form is that it is more directly related to
the ambient 63 concentrations that are associated with health effects. Robust, concentration-
based forms,  in the range of the third- to fifth-highest daily  maximum  8-hr average
concentration, including the current 4th-highest daily maximum form, minimize the inherent lack
of year-to-year stability of exceedance-based forms and provide insulation from the impacts of
       32 Staff consideration of an alternative form based on looking at the nth-highest daily maximum 8-hr
average concentration over three years (specifically the ^-highest value in three years), rather than the current
form that is based on the 3-year average of annual nth-highest concentrations, did not identify a specific alternative
form that was appreciably more consistent across areas than the range of forms previously considered.
                                          6-84

-------
extreme meteorological events.  Such instability can have the effect of reducing public health
protection by disrupting ongoing implementation plans and associated control programs.

       6.3.6   Summary of Staff Conclusions and Recommendations on the Primary Os
              NAAQS
       Staff conclusions and recommendations on the elements of the primary Os standard for
the Administrator's consideration in making decisions on the primary 63 standard are
summarized below, together with supporting conclusions from sections 6.3.1 to 6.3.5 above.
These standard elements, including indicator, averaging time, level, and form, collectively
determine the public health protection afforded by the standard.
       We recognize that selecting from among alternative standards will necessarily reflect
consideration of qualitative and quantitative uncertainties inherent in the relevant evidence and in
the assumptions of the quantitative exposure and risk  assessments. Any such standard should
protect public health against health effects associated  with exposure to Os, alone or in
combination with related photochemical oxidants, taking into account both evidence-based and
exposure- and risk-based considerations, and the nature and degree of uncertainties in such
information. In recommending these ranges of alternative standards for consideration, we are
mindful that the Act requires standards that, in the judgment of the Administrator, are requisite to
protect public health with an adequate margin of safety.  The standards are to be neither more nor
less stringent than necessary. Thus, the Act does not require that NAAQS be set at zero-risk
levels, but rather at levels that reduce risk sufficiently to protect public health with an adequate
margin of safety.

(1)    It is appropriate to continue to use 63 as the indicator for a standard that is intended to
       address effects associated with exposure  to Os, alone or in combination with related
       photochemical oxidants.  Based on the available information, and consistent with the
       views of CASAC and public commenters, we  concluded that there is no basis for
       considering any alternative indicator at this time. Staff notes that while the new body of
       time-series epidemiological evidence cannot resolve questions about the relative
       contribution of other photochemical oxidant species to the range of morbidity and
       mortality effects associated with Os in these types of studies, control of ambient Os levels
       is generally understood to provide the best means of controlling photochemical oxidants
       in general, and thus of protecting against effects that may be associated with individual
       species and/or the broader mix of photochemical oxidants, independent of effects
       specifically related to O3.
                                         6-85

-------
(2)     It is appropriate to continue to use an 8-hr averaging time for the primary 63 standard.
       We conclude that a standard with an 8-hr averaging time can effectively limit both 1- and
       8-hr exposures of concern and that an 8-hr averaging time is appropriate to provide
       adequate and more uniform protection of public health from both short-term (1- to 3-hr)
       and prolonged (6- to 8-hr) exposures to Os in the ambient air. Therefore, we recommend
       retaining the 8-hr averaging time and do not recommend consideration of a separate 1-hr
       standard at this time.  We also conclude that consideration of a standard with a longer-
       term averaging time (e.g., annual)  is not warranted at this time.

(3)     We conclude that the overall body of evidence clearly calls into question the adequacy of
       the current standard and provides strong support for consideration of an O?, standard that
       would provide increased health protection for sensitive groups, including asthmatic
       children and other people with lung disease, as well as all children and older adults,
       especially those active outdoors, and outdoor workers, against an array of adverse health
       effects that range from decreased lung function and respiratory symptoms to serious
       indicators of respiratory morbidity including ED visits and hospital admissions for
       respiratory causes, and possibly cardiovascular-related effects and mortality. We also
       conclude that risks projected to remain upon meeting the current standard, based on the
       exposure and risk assessment, are indicative of risks to sensitive groups that can
       reasonably be judged to be important from a public health perspective, which reinforces
       our conclusion that consideration should be given to revising the level of the standard so
       as to provide increased public health protection.
       (a)    We recommend that consideration be given to  a standard level within the range of
             somewhat below 0.080 ppm to 0.060 ppm,  reflecting our judgment that a standard
             set within this range could  provide an appropriate degree of public health
             protection and would result in important improvements in protecting the health of
             sensitive groups. Standard levels within this range that were considered in staff
             analyses of air quality, exposure, and risk include 0.074, 0.070, and 0.064 ppm,
             representative of levels within the upper, middle, and lower parts of this range,
             respectively.
       (b)    We further recommend that consideration be given to specifying the level of the
             primary  standard to the nearest thousandth ppm, reflecting the degree of precision
             with which ambient 63 concentrations can be measured and design values can be
             calculated.
                                         6-86

-------
(4)    We conclude that it is appropriate to consider a form in the range of the annual third- to
       fifth-highest daily maximum 8-hr average concentration, which includes the current form
       of the annual fourth-highest daily maximum 8-hr average concentration, averaged over
       three years. It is appropriate to consider a form within this range in conjunction with a
       standard level within the recommended range, so as to provide an appropriate degree of
       increased public health protection.

6.4    SUMMARY OF KEY UNCERTAINTIES AND RESEARCH
       RECOMMENDATIONS RELATED TO SETTING A PRIMARY O3 STANDARD
       We believe it is important to continue to highlight  the uncertainties associated with
establishing standards for Os during and after completion  of the NAAQS review process.
Research needs go beyond what is necessary to understand health and welfare effects, population
exposures, and the risks of exposure for purposes of setting standards. Research can also support
the development of more efficient and effective control strategies. It should be noted, however,
that a thorough discussion of research needs related to control strategy development is beyond
the scope of this document.
       Following completion of the 1996 Ozone Staff Paper (U.S. EPA, 1996), the EPA held a
research needs workshop and produced a draft document33 for review by the CASAC at a public
meeting held November 16, 1998. Based on our review of scientific information contained in
the 2006 CD, we have concluded that 63 health research needs and priorities have not changed
substantially since the above document was written.  Key  uncertainties and research needs that
continue to be high priority for future reviews of the health-based primary standards are
identified below:

(1)    An important aspect of risk characterization and decision making for air quality standard
       levels for the Os NAAQS is the characterization of the shape of exposure-response
       functions for 63, including the identification of potential population threshold levels.
       Recent controlled human exposure studies conducted at levels below 0.08 ppm O3
       provide evidence that measurable lung function effects occur in some individuals for 6-8
       hr exposures in the range of 0.08 to as low as 0.04  ppm.  A major limitation of these data
       is that they were collected in one laboratory located in an area of the U.S. that typically
       experiences higher ambient air levels of O3; therefore, prior attenuation of subject
       response may have been  a factor in the responses observed.  Considering the importance
       of estimating health risks in the range of 0.04 to 0.08 ppm O3, additional research is
       33"Ozone Research Needs to Improve Health and Ecological Risk Assessment" (U.S. EPA, 1998).
                                         6-87

-------
       needed to evaluate responses in healthy and asthmatic individuals in the range of 0.04 to
       0.08 ppm for 6-8 hr exposures while engaged in moderate exertion.

(2)     Similarly, for health endpoints reported in epidemiological studies such as hospital
       admissions, ED visits, and premature mortality, an important aspect of characterizing risk
       is the shape of concentration-response functions for O3, including identification of
       potential population threshold levels. Most of the recent studies and analyses continue to
       show no evidence for a clear threshold in the relationships between Os levels and these
       health endpoints or have suggested that any such thresholds must be at very low levels
       approaching policy relevant background levels.  Whether or not exposure errors,
       misclassification of exposure, or potential impacts of other copollutants may be obscuring
       potential population thresholds is still unknown.

(3)     The extent to which the broad mix of photochemical oxidants and more generally other
       copollutants in the ambient air (e.g.,  PM, NC>2, 862, etc.) may play  a role in modifying or
       contributing to the observed associations between ambient Os and various morbidity
       effects and mortality continues to be an important research question. Ozone has long
       been known as an indicator of health effects of the entire photochemical oxidant mix in
       the ambient air and has served as a surrogate for control purposes.  A better
       understanding of sources of the broader pollutant mix, of human exposures, and of how
       other pollutants may modify or contribute to the health effects of O3 in the ambient air,
       and vice versa, is needed to better inform future NAAQS reviews.

(4)     As epidemiological research has become a more important factor in assessing the public
       health impacts of 63, methodological issues in epidemiological studies have received
       greater visibility and scrutiny. Investigations of questions on the use of generalized
       additive models in time-series epidemiological studies have raised model specification
       issues. There remains a need for further study on the selection of appropriate modeling
       strategies and appropriate methods to control for time-varying factors, such as
       temperature, and to better understand the role of copollutants in the ambient air.

(5)     Limited controlled human exposure and epidemiology research has provided suggestive
       evidence of both direct and indirect effects of Os on the cardiovascular system,
       cardiovascular hospital admissions, and cardiovascular mortality. However, additional
       work will be needed to examine biologically plausible mechanisms of cardiovascular
       effects and to determine the extent to which Os is directly implicated or works together

-------
       with other pollutants in causing adverse cardiovascular effects in sensitive individuals
       and in the general population.

(6)     Most epidemiological studies of short-term exposure effects have been time-series studies
       in large populations. Time-series studies remain subject to uncertainty due to use of
       ambient fixed-site data  serving as a surrogate for ambient exposures, to the difficulty of
       determining the impact of any single pollutant among the mix of pollutants in the ambient
       air, to limitations in existing statistical models, or to a combination of all of these factors.
       Independent variables for air pollution have generally been measurements made at
       stationary outdoor monitors, but the accuracy with which these measurements actually
       reflect subjects' exposure is not yet fully understood. Also, additional research is needed
       to improve the characterization of the degree to which discrepancy between stationary
       monitor measurements  and actual pollutant exposures introduces error into statistical
       estimates of pollutant effects in time-series studies.

(7)     Improved understanding of human exposures to ambient Os and to related copollutants is
       an important research need. Population-based information on human exposure for
       healthy adults and children and susceptible or at-risk populations including asthmatics to
       ambient Os concentrations, including exposure information in various
       microenvironments, is needed to better evaluate current and future 63 exposure models.
       Such information is needed for sufficient periods to facilitate evaluation of exposure
       models throughout the Os season.

(8)     Information is needed to improve inputs to current and  future population-based Os
       exposure and health risk assessment models.  Collection of time-activity data over longer
       time periods is needed to reduce uncertainty in the modeled exposure distributions that
       form an important part  of the basis for decisions regarding air quality standard for Os and
       other air pollutants. Research addressing energy expenditure and associated breathing
       rates in various population groups, particularly healthy  and asthmatic children, in various
       locations, across the spectrum of physical  activity, including sleep to vigorous physical
       exertion is needed.

(9)     An important consideration in the Os NAAQS review is the characterization of policy
       relevant background levels. There still remain significant uncertainties in the
       characterization of 8-hr daily maximum O3 background concentrations. Further research
       to improve the evaluation of the GEOS-CHEM model which has been used to
       characterize estimates of policy relevant background levels would help reduce

                                         6-89

-------
uncertainties in estimating health risks relevant for standard setting (i.e., those risks
associated with exposure to Os in excess of policy relevant background levels) and would
aid in the development of associated control programs.
                                   6-90

-------
REFERENCES


Abt Associates Inc. (2006). Ozone Health Risk Assessment for Selected Urban Areas: Draft Report. Prepared for
      Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency, Research Triangle
      Park, NC. June 2006.  Available electronically on the internet at:
      http://www.epa.gOv/ttn/naaqs/standards/ozone/s  o3crtd.html.

Adams, W.C. (2002).  "Comparison of Chamber and Face-Mask 6.6-Hour Exposures to Ozone on Pulmonary
      Function and Symptoms Responses." Inhalation Toxicology 14:745-764.

Adams, W.C. (2003).  "Comparison of Chamber and Face Mask 6.6-Hour Exposure to 0.08 ppm Ozone via Square-
      Wave and Triangular Profiles on Pulmonary Responses." Inhalation Toxicology 15: 265-281.

Adams, W.C. (2006).  "Comparison of Chamber 6.6-h Exposures to 0.04-0.08 ppm Ozone via Square-Wave and
      Triangular Profiles on Pulmonary Responses." Inhalation Toxicology 18: 127-136.

Bell, M. L.; McDermott, A.; Zeger, S. L.; Samet, J. M.; Dominici, F. (2004) Ozone and short-ter5m mortality in 95
      US urban communities, 1987-2000. JAMA J. Am. Med. Assoc. 292: 2372-2378.

Bell, M,A. R.D. Peng, and F. Dominici (2006). "The Exposure-Response Curve for Ozone and Risk of Mortality
      and the Adequacy of Current Ozone Regulations." Environmental Health Perspectives.  Available online at:
      http://dx.doi.org/

Blank, H. and Mitchell, A. (2001).  The Status of Preschool Policy in the States.  Children's Defense Fund and Early
      Childhood Policy Research.  Available at:
      www.earlychildhoodfinance.org/handouts/StatusOfPreschoolPolicyInTheStates.doc

Brauer, M.; Blair, J.; Vedal, S. (1996) Effect of ambient ozone exposure on lung function in farm workers. Am. J.
      Respir. Crit. Care Med. 154: 981-987.

Burnett, R. T.; Brook, J. R.; Yung, W. T.; Dales, R. E.;  Krewski, D. (1997a) Association between ozone and
      hospitalization for respiratory diseases in 16 Canadian cities. Environ. Res. 72:  24-31.

Burnett, R. T.; Brook, J. R.; Yung, W. T.; Dales, R. E.;  Krewski, D. (1997b) Association between ozone and
      hospitalization for respiratory diseases in 16 Canadian cities. Environ. Res. 72:  24-31.

Burnett, R. T.; Smith-Doiron, M.; Stieb, D.; Raizenne, M. E.; Brook, J. R.; Dales, R. E.; Leech, J. A.; Cakmak, S.;
      Krewski, D. (2001) Association between ozone and hospitalization for acute respiratory diseases in children
      less than 2 years of age. Am. J. Epidemiol. 153:  444-452.

Chock, D. P.; Winkler, S. L.; Chen, C. (2000) A study of the association between daily mortality and ambient air
      pollutant concentrations in Pittsburgh, Pennsylvania. J. Air Waste Manage. Assoc.  50: 1481-1500.

Delfino, R. J.; Murphy-Moulton, A. M.; Burnett, R. T.; Brook, J. R.; Becklake, M. R. (1997) Effects of air pollution
      on emergency room visits for respiratory illnesses in Montreal, Quebec. Am. J.  Respir. Crit. Care Med. 155:
      568-576.

Delfino, R.J.; Zeiger, R.S.; Seltzer, J.M.; Street, D.H. (1998) Symptoms inpediatric asthmatics and air pollution:
      differences in effects by symptom severity, anti-inflammatory medication use and paniculate averaging time.
      Environ. Health Perspect.  106: 751-761.

Dockery, D. W.; Schwartz, J.; Spengler, J. D. (1992) Air pollution and daily mortality: associations with
      particulates and acid aerosols. Environ.  Res. 59:  362-373.
                                                6-91

-------
Fairley, D. (1999) Daily mortality and air pollution in Santa Clara County, California: 1989-1996. Environ. Health
      Perspect. 107: 637-641.

Fairley, D. (2003) Mortality and air pollution for Santa Clara County, California, 1989-1996. In:  Revised analyses
      of time-series studies of air pollution and health. Special report. Boston, MA:  Health Effects Institute;
      pp. 97-106. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdf [18 October, 2004].

Federal Register (1997) National Ambient Air Quality Standards for Ozone; Final Rule. 40 CFR 50; Federal
      Register 62: 38856.

Gamble, J. L. (1998) Effects of ambient air pollution on daily mortality: a time series analysis of Dallas, Texas,
      1990-1994. Presented at: 91st 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.

Gent, J. F.; Triche, E. W.; Holford, T. R.; Belanger, 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

Gryparis, A.; Forsberg, B.; Katsouyanni, K.; Analitis, A.; Touloumi, G.; Schwartz, J.; 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.

Henderson, R. (2006a) Letter from CASAC Chairman Rogene Henderson to EPA Administrator Stephen Johnson,
      February 16, 2006, EPA-CASAC-06-003.

Henderson, R. (2006b) Letter from CASAC Chairman Rogene Henderson to EPA Administrator Stephen Johnson,
      June 5, 2006, EPA-CASAC-06-007.

Henderson, R. (2006c) Letter from CASAC Chairman Rogene Henderson to EPA Administrator Stephen Johnson,
      October 24, 2006, EPA-CASAC-07-001.

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.

Ito, K. (2003) Associations of paniculate 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],

Ito, K.; De Leon, S. F.; Lippmann, M. (2005) Associations between ozone and daily mortality, analysis and meta-
      analysis. Epidemiology 16: 446-457.

Kinney, P. L.; Ito, K.; Thurston, G. D. (1995) A sensitivity analysis of mortality/PM10 associations in Los Angeles.
      In: Phalen, R. F.; Bates, D. V., eds. Proceedings of the colloquium on paniculate air pollution and human
      mortality and morbidity; January 1994; Irvine, CA. Inhalation Toxicol. 7:  59-69.

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.

Korrick, S. A.; Neas, L. M.; Dockery, D. W.; Gold, D. R.; Allen, G. A.; Hill, L. B.; Kimball, K. D.; Rosner, B. A.;
      Speizer, F. E. (1998) Effects of ozone and other pollutants on the pulmonary function of adult hikers.
      Environ. Health Perspect. 106:  93-99.

Langstaff, J.  (2007) Analysis of Uncertainty in Ozone Population Exposure Modeling, Draft Memorandum to the
      Ozone NAAQS Review Docket (OAR-2005-0172), July 24, 2006.
                                                6-92

-------
Levy, J. I.; Chemerynski, S. M; Sarnat, J. A. (2005) Ozone exposure and mortality, an empiric Bayes
      metaregression analysis. Epidemiology 16: 458-468.

Linn, W. S.; Shamoo, D. A.; Anderson, K. R.; Peng, R.-C.; Avol, E. L.; Hackney, J. D.; Gong, H., Jr. (1996) Short-
      term air pollution exposures and responses in Los Angeles area schoolchildren. J. Exposure Anal. Environ.
      Epidemiol. 6: 449-472.

Lippmann, M.; Ito, K.; Nadas, A.; Burnett, R. T. (2000) Association of paniculate matter components with daily
      mortality and morbidity in urban populations. Cambridge, MA: Health Effects Institute; research report
      no. 95.

Moolgavkar, S. H.; Luebeck, E. G.; Hall, T. A.; Anderson, E. L. (1995) Air pollution and daily mortality in
      Philadelphia. Epidemiology 6:  476-484.

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.

Naeher, L. P.; Holford, T. R.; Beckett, W. S.; Belanger, K.; Triche, E. W.; Bracken, M. B.; Leaderer, B. P. (1999)
      Healthy women's PEF variations with ambient summer concentrations of PM10, PM25, SO42~, H+, and O3.
      Am. J. Respir. Crit. Care Med. 160: 117-125.

Neas, L. M.; Dockery, D. W.; Koutrakis, P.; Tollerud, D. J.; Speizer, F. E. (1995) The association of ambient air
      pollution with twice daily peak expiratory flow rate measurements in children. Am.  J. Epidemiol. 141: 111-
      122.

Ostro, B. (1995)  Fine paniculate air pollution and mortality in two Southern California counties. Environ. Res.
      70: 98-104.

Ostro, B.; Lipsett, M.; Mann, J.; Braxton-Owens, H.; White, M. (2001) Air pollution and exacerbation of asthma in
      African-American children in Los Angeles. Epidemiology 12: 200-208.

Ostro, B. D.; Broadwin, R.; Lipsett, M. J. (2003) Coarse particles and daily mortality in Coachella Valley,
      California. In: Revised analyses of time-series studies of air pollution and health. Special report. Boston, MA:
      Health Effects Institute; pp.  199-204. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdf [18
      October, 2004].

Peters, A.; Dockery, D.  W.; Muller, J. E.; Mittleman, M. A. (2001)  Increased paniculate air pollution and the
      triggering of myocardial infarction. Circulation 103: 2810-2815.

Post, E. (2007).  Memorandum - Subject: Additional Tables of Lung Function Risk Estimates Associated with
      Ozone Exposures Based on 2003 Air Quality Data. January 30.  Available at:
      http://www.epa.gov/ttn/naaqs/standards/ozone/s_o3_cr_td.html.

Ross, M. A.; Persky, V. W.; Scheff, P. A.; Chung, J.; Curtis, L.; Ramakrishnan, V.; Wadden, R. A.; Hryhorczuk,
      D. O. (2002) Effect of ozone and aeroallergens on the respiratory health of asthmatics. Arch. Environ. Health
      57: 568-578.

Schwartz, J. (1996) Air pollution and hospital admissions for respiratory disease. Epidemiology 7: 20-28.

Schwartz, J. (2005) How sensitive is the association between ozone and daily deaths to control for temperature?
      Am. J. Respir. Crit. Care Med. 171: 627-631.

Sheppard, L. (2003) Ambient air pollution and nonelderly asthma hospital admissions in Seattle, Washington, 1987-
      1994. In: Revised analyses of time-series studies of air pollution and health. Special report. Boston, MA:
                                                 6-93

-------
      Health Effects Institute; pp. 227-230. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdf
      [18 October, 2004].

U.S. Environmental Protection Agency (1996) Review of the national ambient air quality standards for ozone:
      assessment of scientific and technical information. OAQPS staff paper. Research Triangle Park, NC: Office
      of Air Quality Planning and Standards; EPA report no. EPA-452/R-96-007. Available from: NTIS,
      Springfield, VA; PB96-203435. http://www.epa.gov/ttn/naaqs/standards/ozone/s_o3_pr_sp.html

U.S. Environmental Protection Agency (1998) Ozone Research Needs to Improve Health and Ecological Risk
      Assessments (External Review Draft). Research Triangle Park, NC:  Office of Research and Development;
      report no. EPA/600/R-98/031. March 31, 1998.

U.S. Environmental Protection Agency (2006a) Air quality criteria for ozone and related photochemical oxidants
      (Final).  Research Triangle Park, NC: National Center for Environmental Assessment; report no. EPA/600R-
      05/004aB-cB, 3v.  Available: http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid= 137307 [March 2006]

U.S. Environmental Protection Agency (2006b) Ozone Population Exposure Analysis for Selected Urban Areas
      (draft).  Office of Air Quality Planning and Standards, Research Triangle Park, NC. June. Available
      electronically on the internet at: http://www.epa.gOv/ttn/naaqs/standards/ozone/s o3crtd.html.

U.S. Environmental Protection Agency (2006c) Review of the national ambient air quality standards for ozone:
      policy assessment of scientific and technical information. OAQPS staff paper- Second draft. Research
      Triangle Park, NC: Office of Air Quality Planning and Standards; EPA report no. EPA-452/D-05-002, July
      2006. Available electronically at: http://www.epa.gOv/ttn/naaqs/standards/ozone/s o3crsp.html.

U.S. Environmental Protection Agency (2006d) Transcript of the August 24, 2006 Public Advisory Meeting to
      Conduct a Peer Review of EPA's 2nd Draft Ozone Staff Paper and Related Draft Technical Support
      Documents, Science Advisory Board, August. Docket number EPA-HQ-OAR-2005-0172-0044.

U.S. Environmental Protection Agency (2006e) Transcript of the August 25, 2006 Public Advisory Meeting to
      Conduct a Peer Review of EPA's 2nd Draft Ozone Staff Paper and Related Draft Technical Support
      Documents, Science Advisory Board, August. Docket number EPA-HQ-OAR-2005-0172-0045.

Vedal, S.; Brauer, M.; White, R.; Petkau, J. (2003) Air pollution and daily mortality in a city with low levels of
      pollution. Environ. HealthPerspect. Ill:  45-51.

Villeneuve, P.J.; Burnett,  R.T.; Shi, Y.; Krewski, D.; Goldberg, M.S.; Hertzman, C.; Chen, Y.; Brook, J.  (2003) A
      time-series study of air pollution, socioeconomic status, and mortality in Vancouver, Canada. J. Exposure
      Anal. Environ. Epidemiol. 13: 427-435.

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.
                                               6-94

-------
   7.     POLICY-RELEVANT ASSESSMENT OF WELFARE EFFECTS
                                     EVIDENCE
7.1    INTRODUCTION
       This chapter presents information critical to the review of the secondary NAAQS for Os.
Welfare effects addressed by a secondary NAAQS 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.  Of these welfare effects categories,
the effects of O3 on vegetation, including agricultural crops, trees in managed and unmanaged
forests, and herbaceous and woody species growing in natural settings are of most concern at
concentrations typically occurring in the U.S. As stated in earlier reviews, "of the phytotoxic
compounds commonly found in the ambient air, O3 is the most prevalent, impairing crop
production and injuring native vegetation and ecosystems more than any other air pollutant"
(U.S. EPA, 1989,  1996b).
       Ozone can also affect other ecosystem components such as soils, water, wildlife, and
habitat, either directly, or indirectly, through its effects on vegetation.  These individual
ecosystem components are associated with one or more of six essential ecological attributes
(EEAs) recently described in A Framework for Assessing and Reporting on Ecological
Condition: an SAB report (Young and Sanzone, 2002) as part of a conceptual framework useful
for assessing and reporting on ecological condition  (see Figure 7-21 and discussion in section
7.7). This framework can be used to link O3 effects at the species level to potential impacts at
higher levels in the hierarchy (e.g., EEAs).  Some of these species level impacts have direct,
quantifiable economic value, while others are currently not quantifiable,  but still have societal
value. In the absence of sufficient research to allow quantification of O3 impacts at the
ecosystem level, including impacts on ecosystem goods and services, only a qualitative
discussion is included. However, the staff infers, based on the linkages described in the SAB
framework, that increasing protection for vegetation from O3 related effects would also improve
the protection afforded to ecosystems and their related public welfare categories.
       Other O3 related welfare effects categories include damage to certain manmade materials
(e.g., elastomers, textile fibers, dyes, paints, and pigments) and climate interactions. The amount
of damage to actual in-use materials and the economic consequences of that damage are poorly
characterized, however, and the scientific literature contains very little new information to
adequately quantify estimates of materials damage from photochemical oxidants (U.S. EPA,
                                          7-1

-------
1996a, b, 2006). Therefore, staff judges that there is insufficient information in the materials
damage literature to inform secondary standard setting and so it will not be discussed further.
Interested readers are referred to Chapter 11 in the CD (U.S. EPA, 2006).  In contrast, the
welfare impact of 63 on local, regional and global climates has received more attention in recent
years.  Ozone enhances the heat capacity of the atmosphere.  The overall body of scientific
evidence suggests that high concentrations of 63 on a regional scale could have a discernable
influence on climate, leading to surface temperature and hydrological cycle changes.  However,
the CD stated that confirming this effect will require further advances in monitoring and
improvement in chemical transport and regional-scale modeling.  Thus, staff concludes that
insufficient information is available at this time to quantitatively inform the secondary NAAQS
process with regard to this aspect of the Os-climate interaction and will not address it further.
Another aspect, is the potential modification of plant response to 63 under conditions of
changing climate, is included in the discussion of factors that can modify the predicted
vegetation responses (see section 7.4.2).
       To summarize, this chapter includes an integrated discussion of the key policy relevant
science regarding Os-related effects on vegetation (sections 7.2 through 7.4) and terrestrial
ecosystems (section 7.7),  as described in the previous CD (U.S. EPA, 1996a) and reiterated in
the current CD (U.S. EPA, 2006).  The remaining sections (7.5 and 7.6) of this chapter are
focused on a discussion of the analyses that have been conducted in support  of this current
NAAQS review that update and expand upon the exposure, risk and benefits assessments
conducted in the last review (U.S. EPA, 1996b).  These updated assessments incorporate newer
data, models, and approaches, and take into account alternative Os air quality scenarios under
consideration. The environmental assessment technical support document, Technical Report on
Ozone Exposure, Risk, and Impacts Assessments for Vegetation (Abt,  2007)  (hereafter cited as
"Environmental Assessment TSD"), presents  a detailed description of the exposure, risk and
impacts analysis methodology. Results from these assessments, along with key uncertainties and
limitations, are also described in sections 7.5 and 7.6.  This information forms the basis for a
discussion in Chapter 8 of staff conclusions and recommendations with respect to the secondary
O3 NAAQS.

7.2    MECHANISMS  GOVERNING PLANT RESPONSE TO OZONE
       The interpretation of predictions of risk associated with vegetation response at ambient
Os exposure levels can be informed by scientific understanding regarding Os impacts at the
genetic, physiological, and mechanistic levels. In most cases, the mechanisms of response are
similar regardless of the degree of sensitivity of the species. The information assessed in the
1996 CD (U.S. EPA 1996a) regarding the fundamental hypotheses concerning Os-induced
                                           7-2

-------
changes in physiology continues to be valid.  However, during the last decade, understanding of
the cellular processes within plants has been further clarified and enhanced. Therefore, this
section reviews the key scientific conclusions identified in 1996 O3 CD (U.S. EPA, 1996a), and
incorporates new information from the current CD (U.S. EPA, 2006).  This section describes:
(1) O3 uptake, (2) cellular to systemic O3 response, (3) plant compensation and defense
mechanisms, (4) O3-induced changes to plant metabolism, and (5) plant response to chronic O3
exposures.

       7.2.1   Ozone Uptake: Canopy, Plant and Leaf
       To cause injury, O3 must first enter the plant through the stomata of the leaves.  Leaves
exist in a three dimensional environment called the plant canopy, where each leaf has a unique
orientation and receives a different exposure to ambient air, microclimatological conditions, and
sunlight. In addition, a plant may be located within a stand of other plants which further
modifies ambient air exchange with individual leaves. Not all O3 entering a plant canopy is
absorbed into the leaves, but may be adsorbed to other surfaces e.g., leaf cuticles, stems, and soil
(termed non-stomatal deposition) or scavenged by reactions with intra-canopy biogenic VOCs
and naturally occurring NOx emissions from  soils. Because O3 does not penetrate the leafs
cuticle, it must reach the stomatal openings in the leaf for absorption to occur. The movement of
O3 and other gases such as CC>2 into and out of leaves is controlled primarily through the
stomata.  The aperture of the stomata are controlled by guard cells, which respond to a variety of
internal species-specific factors as well  as external site specific environmental factors such as
light, humidity, CC>2 concentration, soil fertility and water status, and in some cases the presence
of other air pollutants, including O3 (see section 7.4.2).  These modifying factors produce
stomatal conductances that vary across the diurnal cycle, days and seasons. Once O3 is inside the
leaf, a phytotoxic effect will only occur if sufficient amounts of O3 reach sensitive cellular sites
that are subject to the various  physiological and biochemical controls within the leaf cells (see
the discussion in section 7.2.3 below - Compensation and Detoxification).
       A measure of O3 flux is attractive because it incorporates both relevant environmental
factors and physiological processes, and is considered the measure that most closely links
exposure to plant response.  Unfortunately, measurement of flux is very complex, making it
difficult to extrapolate uptake from an individual leaf to that of a whole plant or canopy. Since
the last review, interest has been increasing, particularly in Europe, in using mathematically
tractable flux models for O3 assessments at the regional and national scale  (Emberson et al.,
2000a, b). Though significant new research has been done with respect to  flux model
development, it has  still not advanced to a point of being generally applicable across a range of
                                           7-3

-------
species and environments at a national scale.  These topics are discussed in more detail in
Appendix 7 A of this document and in the current CD (U.S. EPA, 2006).

       7.2.2   Cellular to Systemic Response
       Once Os diffuses into the leaf air spaces it can react with a variety of biochemical
compounds that are exposed to the air (path 1) or it can be solubilized into the water lining the
cell wall of the air spaces (path 2). Having entered the aqueous phase, Os can be rapidly altered
to form oxidative products that can diffuse more readily into and through the cell and react with
many biochemical compounds.  The initial  sites of membrane reactions seem  to involve transport
properties and, possibly, the external signal transducer molecules (U.S. EPA,  2006). This
alteration in plasma membrane function is clearly an early step in a series of Os-induced events
that leads to leaf injury.
       Under certain circumstances, Os reacts with organic molecules to generate peroxides,
including hydrogen peroxide (H2O2). The role of hydrogen peroxide as a signaling molecule in
plants is now better understood. The primary set of metabolic reactions that Os triggers clearly
includes those typical of "wounding" responses generated by cutting of the leaf or by
pathogen/insect attack.  One aspect of this total response is the production of O2 and H2O2 by the
cell (Lamb and Dixon, 1997). The presence of higher-than-normal levels of H2O2 within the
apoplastic space is a potential trigger for the normal, well-studied pathogen defense pathway.
       Ethylene is another compound produced when plants are subjected to biotic or abiotic
stressors.  Increased ethylene production by plants exposed to Os stress was identified as a
consistent marker for Os exposure decades ago (Tingey et al., 1976).  These studies suggested
that increased production of stress-ethylene correlated well with the degree of foliar injury that
developed within hours or days after Os exposure. Thus,  one could postulate  that Os generates a
wounding response with the production of ethylene, which would, in turn, generate a change in
stomatal conductance and photosynthesis.

       7.2.3   Compensation and Detoxification
       Ozone injury will not occur if (1) the rate and amount of Os uptake is small enough for
the plant to detoxify or metabolize O3 or its metabolites or (2) the plant is able to repair or
compensate for the Os impacts (Tingey and Taylor,  1982; U.S. EPA, 1996a).  Leaves may
physically exclude Os from sensitive tissues.  A few studies have documented a direct stomatal
closure or restriction in response to the presence of Os ranging from within minutes to hours or
days of exposure (Moldau et al., 1990; Dann and Pell, 1989; Weber et al.,  1993). However,
exclusion of Os also restricts the uptake of CO2, thus limiting photosynthesis and growth.
       In addition, plants can also effectively protect tissue against damage by dissipating excess
oxidizing power using antioxidants.  Since  1996, the role  of detoxification in providing a level of
                                           7-4

-------
resistance to 63 has been further investigated.  A number of antioxidants, including ascorbate,
glutathione peroxidase, and sulfuroxide dimutase, which are highly reactive, can detoxify the
chemicals generated by 03. The pattern of changes in these antioxidant proteins varies greatly
among different species and conditions. Most recent reports indicate that ascorbate within the
cell wall provides the first significant opportunity for detoxification to occur. The balance
between the total 63 flux and the detoxification process has been defined as the "effective flux"
(Dammgen et al.,  1993; Griinhage and Haenel, 1997; Musselman and Massman, 1999).
       In spite of the new research, however, it is still not clear as to what extent detoxification
protects against 63 injury.  Specifically, data are needed on the potential rates of antioxidant
production and on the subcellular location of the antioxidants.  Potential rates of antioxidant
production are needed to assess whether they are sufficient to detoxify the Os as it enters the cell.
Data on the subcellular location(s) are needed to assess whether the antioxidants are in the cell
wall or plasmalemma locations that permit contact with the Os  before it has a chance to damage
subcellular systems. In addition, generation of these antioxidants in response to (Vinduced
stress potentially diverts resources away from other sinks and expends energy.  Thus, scientific
understanding of the detoxification mechanisms is not yet complete and requires further
investigation (U.S. EPA, 2006).
       Once Os injury has occurred in leaf tissue, some plants  are able to repair or compensate
for the impacts (Tingey and Taylor, 1982).  In general, plants have a variety of compensatory
mechanisms for low levels of stress including reallocation of resources, changes in root/shoot
ratio,  production of new tissue, and/or biochemical shifts, such as increased photosynthetic
capacity in new foliage and changes in respiration rates, indicating possible repair or replacement
of damaged membranes or enzymes. Since these mechanisms are genetically determined, not all
plants have the same complement or degree  of tolerance, nor are all stages of a plant's
development equally  sensitive to 63. It is not yet known to what degree or how the use of plant
resources for repair processes affects the overall carbohydrate budget or subsequent plant
response to O3 or other stresses (U.S. EPA, 1996a, U.S. EPA, 2006).

       7.2.4   Changes to Plant Metabolism
       Ozone inhibits photosynthesis, the process by which plants produce energy rich
compounds (e.g., carbohydrates) in the leaves. This impairment can result from direct impact to
chloroplast function and/or Os-induced stomatal closure resulting in reduced uptake of CO2. A
large body of literature published since 1996 has further elucidated the mechanism of effect of
Os within the chloroplast.  Pell et al. (1997)  showed that Os exposure results in a loss of
Ribulose-1, 5-bisphosphate carboxylase/oxygenase (RuBisCo), the central carboxylating enzyme
that plays an important role in the production of carbohydrates. Due to its central importance,
                                           7-5

-------
any decrease in RuBisCo may have severe consequences for the plant's productivity.  Several
recent studies have found that Os has a greater effect as leaves age, with the greatest impact of
Os occurring on the oldest leaves (Fiscus et al., 1997; Reid and Fiscus, 1998; Noormets et al.,
2001; Morgan et al., 2004).  The loss of RuBisCo and its messenger RNA as a function of
increasing Os exposure is also linked to an early senescence or a speeding up of normal
development leading to senescence. If total plant photosynthesis is sufficiently reduced, the
plant will respond by reallocating the remaining carbohydrate at the level of the whole organism
(see section 7.3 below) (U.S. EPA, 1996a, 2006).

       7.2.5   Plant Response to Chronic/Long-term Os Exposures
       Many changes  that occur with 63 exposure can be observed within hours, or perhaps
days, of the exposure,  including those connected with wounding and elicitor-induced changes in
gene expression. Other effects due to Os, however, take longer to occur and tend to become
most obvious after long exposures to low-Os concentrations. These have been linked to
senescence or some other physiological response very closely linked to senescence. The
understanding of how  Os affects long-term growth and resistance to other biotic and abiotic
insults in long-lived trees is unclear.  Often, the conditions to which a tree is subjected to in one
year will affect, or "carry-over", the response of that tree into the next year (U.S. EPA, 2006). In
other words, a condition in an earlier year sets the stage for a reaction in the next year; thereby
giving a "cause-effect" scenario (U.S. EPA, 2006).  In perennial plant species, growth affected
by a reduction in carbohydrate storage may result in the limitation of growth the following year
(Andersen et al., 1997).  Carry-over effects have been documented in the growth of tree
seedlings (Hogsett et al., 1989; Sasek et al., 1991; Temple et al., 1993; U.S. EPA, 1996a) and in
roots (Andersen et al.,  1991; U.S. EPA, 1996a).  Accumulation of carry-over effects overtime
will affect survival and reproduction. Understanding of how Os interacts with the plant at a
cellular level has dramatically improved in recent years.  However, additional work remains to
more fully elucidate the translation of those cellular mechanisms into altered cell metabolism,
whole plant productivity, and other physiological effects.

7.3    NATURE OF  EFFECTS ON VEGETATION
       Studies published since the conclusion of the 1996 review continues to support and
strengthen key conclusions regarding O3 effects on  vegetation and ecosystems found in the
previous CD (U.S. EPA, 1996a) and reiterated in the current CD (U.S. EPA, 2006). For
additional detail the reader is referred to Chapter 9 and Annex 9 in the current CD (U.S. EPA,
2006)
                                          7-6

-------
       7.3.1   Ozone Sensitive Plants and Their Relationship to Public Welfare
       Of all the plant species growing within the U.S. (over 43,000 species have been
catalogued in the USDA PLANTS database, USD A, NRCS, 2006), only a small percentage have
been studied with respect to Os sensitivity.  Most of the studied species were selected because of
their commercial importance or observed Os-induced visible foliar injury in the field. Given that
Os impacts to vegetation also include less obvious but often more significant impacts, such as
reduced annual growth rates and below ground root loss (see following sections), the paucity of
information on other species means the true range of both inter- and intra-species Os sensitivity
that exists within U.S. vegetation is unknown. Even so, plant species/genotypes with known Os
sensitivity span a broad range of vegetation types and public use categories. These use
categories include food production for human and domestic animal consumption, fiber and
materials production, and urban/private landscaping. In addition to these direct uses, a number
of Os sensitive species have specific relevance to public welfare based on their aesthetic,
medicinal, or habitat value.  Table 7J-1 in Appendix 7J presents a list of Os sensitive species that
occur in Federal Class I areas and Table 7J-2 presents Os sensitive species in each state.

       7.3.2   Vegetation Effects Endpoints
       Ozone injury at the cellular level,  when it has accumulated sufficiently, will be
propagated to the level of the whole leaf or plant.  These larger scale effects can include: visible
foliar injury and premature senescence; reduced carbohydrate production and reallocation;
reduced growth or reproduction; and reduced plant vigor. Much of what is now 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 (U.S. EPA, 1978; 1986; 1996a). Further,  research results
published since 1996 have supported earlier EPA conclusions (U.S. EPA, 2006) and in some
cases have expanded and strengthened those conclusions. The sections below describe the
current understanding of the physiological effects of Os on  vegetation.

       7.3.2.1    Visible Foliar Injury and Premature Senescence
       Cellular injury can and often does become visible. Acute injury usually appears  within
24 hours after exposure to Os and, depending on species, can occur under a range of exposures
and durations from 0.04 ppm for a period of 4 hours  to 0.41 ppm for 0.5 hours for crops  and 0.06
ppm for 4 hours to 0.51 ppm for 1 hour for trees and shrubs (Jacobson, 1977).  Chronic injury
may be mild to severe.  In some cases,  cell death or premature leaf senescence may occur. The
significance of Os injury at the  leaf and whole plant levels depends on how much of the  total leaf
area of the plant has been affected, as well as the plant's age, size, developmental stage,  and
degree of functional redundancy among the existing  leaf area.  Previous CDs have noted the
                                          7-7

-------
difficulty in relating visible foliar injury symptoms to other vegetation effects such as individual
tree growth, stand growth, or ecosystem characteristics (U.S. EPA, 1996a) and this difficulty
remains to the present day (See discussion in section 7.6.3.2). As a result, it is not presently
possible to determine, with consistency across species and environments, what degree of injury
at the leaf level has significance to the vigor of the whole plant.
       However, visible foliar injury by itself, can impact the public welfare.  The presence of
visible symptoms due to O3 exposures can reduce the market value of certain leafy crops (such as
spinach, lettuce) and impact the aesthetic value of ornamentals (such as petunia, geranium, and
poinsettia) in urban landscapes and scenic vistas in protected natural areas such as national parks
and wilderness areas. Though not quantified, there is likely some level of economic impact to
businesses and homeowners from Os-related injury on sensitive ornamental species due to the
cost associated with more frequent replacement and/or increased maintenance (fertilizer or
pesticide application). In addition, because Os not only results in discoloration of leaves but can
lead to more rapid  senescence (early shedding of leaves), there is a potential for some lost tourist
dollars at sites where fall foliage is less available or attractive.
       In recent years, field surveys of visible foliar injury symptoms have become more
common, with greater attention to the standardization of methods and the use of reliable
indicator species (Campbell et al., 2000; Smith et al., 2003). Specifically, the Unites States
Forest Service (USFS) through the Forest Health Monitoring Program (FHM) (1990 - 2001) and
currently the Forest Inventory and Analysis (FIA) Program collects data regarding the incidence
and severity of visible foliar injury on a variety of Os sensitive plant species throughout the U.S.
(Coulston et al. 2003, 2004; Smith et al. 2003). Section 7.6.3.2 contains additional information
on the use of visible foliar injury incidence on bioindicator species as a measure of the
occurrence of phytotoxic levels of Os in the ambient air.
       To view pictures of (Vinduced foliar injury symptoms to selected sensitive species go to
the USDA Agricultural Research Service website
(http://www.ars.usda.gov/Main/docs.htm?docid=12463) or the USDA Forest Service  Ozone
Biomonitoring website (http://www.fiaozone.net/species/index.html).

       7.3.2.2   Carbohydrate Production and Allocation
       When total  plant photosynthesis is sufficiently reduced, the plant will respond by
reallocating the remaining carbohydrate at the level of the whole organism. Many studies have
demonstrated that root growth is more sensitive to 63 exposure than stem or leaf growth (U.S.
EPA, 2006).  When fewer carbohydrates are present in the roots, less energy will be available for
root-related functions such as acquisition of water and nutrients. Mycorrhizal fungi in the soil
form a symbiotic relationship with many terrestrial plants. For host plants, these fungi improve

-------
the uptake of nutrients, protect the roots against pathogens, produce plant growth hormones, and
may transport carbohydrates from one plant to another (U.S. EPA, 1996a).  Ozone can disrupt
the association between mycorrhizal fungi and host plants by inhibiting photosynthesis and the
amount of carbohydrates available for transfer to the roots. This effect has recently been
documented in the field. Data from a long-studied pollution gradient in the San Bernardino
Mountains of southern California suggest that 63 substantially reduces root growth in natural
stands of ponderosa pine (Pinusponderosd).  Root growth in mature trees was decreased at least
87% in a high-pollution site as compared to a low-pollution site (Grulke et al., 1998), and a
similar pattern was found in a separate study with whole-tree harvest along this gradient (Grulke
and Balduman, 1999). Though effects on other ecosystem components were not examined, a
reduction of root growth of this magnitude could have significant implications for the below
ground communities at those sites.  In contrast, a study in Great Smoky Mountains National Park
in Tennessee found no statistically significant effects of Os exposure on stem or root biomass for
several tree species (Neufeld et al., 2000). The difference in the results from these two studies
may reflect the species specific nature of the symbiont-host relationship.
       Unlike root systems, effects on leaf and needle carbohydrate content under Os stress
range from a reduction (Barnes et al., 1990; Miller et al., 1989), to no effect (Alscher et al.,
1989), to an increase (Luethy-Krause and Landolt, 1990). Therefore,  studies that examine only
above-ground vegetative components may miss important Os-induced changes below ground.
These below-ground changes could signal a shift in nutrient cycling with significance at the
ecosystem level (Young and Sanzone, 2002).

       7.3.2.3   Growth and Reproduction
       Studies of the growth response  of trees to Os have established that individual deciduous
trees are generally less sensitive to 63 than most annual plants, with the exception of a few
genera such as Populus, which are highly sensitive and in some cases (for instance, quaking
aspen and black cherry), are as sensitive to 63 as annual plants. The 63 sensitivity of seedlings
and mature trees within species and between species varies widely. In general, mature deciduous
trees are likely to be more sensitive to Os compared to seedlings, while mature evergreen trees
are likely to be less sensitive than seedlings. Based on these results, stomatal conductance, 63
uptake, and Os effects cannot be assumed to be equivalent in seedlings and mature trees.
       Depending on exposure duration, concentrations of Os currently in the United States are
sufficient to affect the growth of a number of tree species during the annual growing season.
However, these conclusions do not take into account the possibility of "carry over" effects on
growth in subsequent years, an important consideration in the case of long-lived species. Given
that multiple-year exposures may cause a cumulative effect on the growth of some trees (Hogsett
                                           7-9

-------
et al. 1989; Simini et al., 1992; Temple et al., 1993), it is likely that a number of species
currently are being impacted.
       Other research in the U.S. in the last 10 years has focused on perennial forage crops (U.S.
EPA, 2006). 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 as ruminant animal feed at 63 exposures that occur in some years over large areas of the U.S.
Ozone may also reduce the quality or nutritive value of annual species.
       Recent studies have also further demonstrated Os effects on different stages of plant
reproduction. Effects of 63 have been observed on pollen germination, pollen tube growth,
fertilization, and abortion of reproductive structures, as reviewed by Black et al. (2000). For
seed-bearing plants, reproductive effects will culminate in seed production. The recent scientific
literature supports the conclusions of the 1996 CD that ambient 63 concentrations are reducing
the yield of major crops in the U.S. For example, the yield reductions for soybean are generally
similar to those reported previously (U.S. EPA, 2006).

       7.3.2.4   Reduced Plant Vigor
       Though O3 levels over most of the U.S. are not high enough to kill vegetation directly,
current levels have been shown to reduce the ability of many sensitive species and genotypes
within species to adapt to or withstand other environmental stresses. These may include
increased susceptibility to  freezing temperatures, pest infestations and/or root disease, and
compromised ability to compete for available resources.  For example, when different species are
grown together, 63 exposure can increase the growth of (Vtolerant species while exacerbating
the growth decrease of Os-sensitive species. In the long run, the result of this loss in vigor may
be plant death.

7.4    IMPACTS ON PUBLIC WELFARE
       7.4.1   What Constitutes an Adverse Vegetation Impact from Ozone Exposure?
       Ozone  can cause a  variety of effects, beginning at the level of the individual  cell and
accumulating up to the level of whole leaves, plants, plant populations, communities and whole
ecosystems.  Not all Os-related effects, however, have been classified as "adverse" to public
welfare.  Previous reviews have classified 63 vegetation effects as either "injury"  or "damage" to
help in determining adversity. Specifically, "injury" is defined as encompassing 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 the intended use or value of the
plant (Guderian,  1977). In contrast, "damage" includes those injury effects that also reduce or
impair the intended use or value of the plant.  Damage includes reductions in aesthetic values
                                          7-10

-------
(e.g., visible 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.
While this construct has proved useful in the past,  it appears most useful in the context of
evaluating effects on single plants or species grown in monocultures, such as agricultural crops
or managed forests. It is less clear how it might apply to potential effects on natural forests or
entire ecosystems, such as shifts in species composition or nutrient cycling where the intended
use or value of the system is not specifically quantified.
       A more recent construct for assessing risks to forests described in Hogsett et al. (1997)
suggests that "adverse effects could be classified into one or more of the following categories:
(1) economic production, (2) ecological structure,  (3) genetic resources, and (4) cultural values."
This expands the context for evaluating the adversity of (Vrelated effects beyond the species
level.  In another recent publication, A Framework for Assessing and Reporting on Ecological
Condition: an SAB report (Young and Sanzone, 2002), additional support is provided for
expanding the consideration of adversity by making explicit the linkages between stress- related
effects (e.g., Os exposure) at the species level and  at higher levels within an ecosystem hierarchy.
Staff suggests that consideration of adverse effects undertaken within the context of such a
broader paradigm is appropriate in the context of this secondary NAAQS review.

       7.4.2    Factors That Modify Functional  and Growth Response
       The caveat that must be placed on results from any experimental study on the response of
living organisms to a stressor in a specific setting is that uncertainty is introduced when
attempting to extrapolate or apply those results outside that specific setting (e.g., to a different set
of organisms,  scales, or exposure/growing conditions).  The description of plant response to Os
exposure is no different.  Because staff must necessarily rely on experimental data produced
under very specific sets of conditions in conducting this assessment, it is important to understand
the range of factors that can influence plant response to 63 and the magnitude and direction of
that response,  in order to better assess the likelihood of observing the experimentally predicted
response in the ambient environment.
       Plant response to  63 exposure is a function of the plant's ongoing integration of genetic,
biological,  physical and chemical factors both within and external to the plant.  The corollary is
also true that Os exposure can modify the plant's subsequent integrated response to other
environmental factors, both by influencing the plant response directly, and by contributing to
altered climatic factors that influence plant response through its greenhouse gas forcing
properties.
                                           7-11

-------
       The 1996 63 CD (U.S. EPA, 1996a) concluded with a statement that our understanding
regarding modifying factors was too fragmented to permit drawing many general conclusions.
Unfortunately, in the interval  since the 1996 CD, little additional information has become
available, thus, this earlier conclusion remains unchanged.  Therefore, only a brief overview of
the current understanding from this research is provided. The reader is referred to the 1996 CD
(U.S. EPA 1996a) and the current 2006 CD (U.S. EPA 2006) for further information.

       7.4.2.1   Genetics
       Plant response to O3 is determined by genes that are directly related to oxidant stress and
to an unknown number of genes that are not specifically related to oxidants but instead control
leaf and cell wall thickness, stomatal conductance, and the internal architecture of the air spaces.
It is unlikely that single genes are responsible for Os tolerance, except in rare cases (Engle and
Gabelman, 1966).  Recent studies using molecular biological tools and transgenic plants have
begun to positively verify the role of various genes and gene products in 63 tolerance and are
beginning to increase the understanding of Os toxicity and differences in Os sensitivity.
Specifically, 63 has been shown to trigger the production of a number of compounds (e.g.,
ethylene) and the signaling of these molecules determines, in some cases, the O3 susceptibility of
plants (U.S. EPA, 2006).  Because the genetic code is species specific, species vary greatly in
their responsiveness to 63.  Even within a given species, individual genotypes or populations can
also vary significantly with respect to Os sensitivity. Thus, caution should be taken when
ranking species categorically  as having an absolute degree of sensitivity to 03.

       7.4.2.2   Biological Factors
       The biological factors within the plant's environment that may directly or indirectly
influence its response to Os in a positive or negative manner encompass insects, other animal
pests, diseases, weeds, and other competing plant species. Ozone and other photochemical
oxidants may influence the severity of a disease or infestation either by direct effects on the
causal species, or indirectly by affecting the host, or both. Likewise, mutually  beneficial
relationships or symbioses involving higher plants and bacteria or fungi may also be affected by
O3. Ozone can also have indirect  effects on herbivorous animals due to O3-induced changes in
feed quality.
       New evidence with regard to insect pests and diseases has done little to remove the
uncertainties noted in the 1996 CD (U.S. EPA 1996a).  Most of the large numbers of such
interactions that may affect crops, forest trees, and other natural vegetation have yet to be
studied. With respect to any particular Os-plant-insect interaction, it is still not possible to predict
its likelihood, or its severity.  The situation is only a little clearer with respect to interactions
involving facultative necrotrophic plant pathogens, with Os generally leading to increased
                                           7-12

-------
disease. In contrast, with obligate biotrophic fungal, bacterial, and nematode diseases, there are
twice as many reports indicating Os-induced inhibitions than enhancements.  At this time,
therefore, although some diseases may become more widespread or severe as a result of
exposure to 63, it is still not possible to predict which diseases are likely to present the greatest
risks to crops and forests.
       The latest studies on 63 interactions with root symbionts present a more complex picture
than was described in the last review. In addition to adverse effects of O3 on the functioning of
tree root symbioses with mycorrhizae (discussed in section 7.3.1 above), there is also evidence
that the presence of mycorrhizae may help plants overcome root diseases stimulated by 63 and/or
encourage the spread of mycorrhizae to the roots of uninfected trees.
       The few recent studies of the impact of Os on intraspecific plant competition have again
confirmed that grasses frequently show greater resilience than other types of plants. In grass-
legume pastures, the leguminous species suffer greater growth inhibition.  Separately, the
suppression of ponderosa pine seedling growth by blue wild-rye grass was markedly increased
by O3 (Andersen et al. 2001). Due to the limited number of species studied under competitive
situations to date, however, it is still not possible to predict the outcome of Os exposure on other
specific competitive situations, such as  successional plant communities or  crop-weed
interactions.  Clearly, however, Os stress creates a selective pressure in some vegetative
communities that can lead to a shift in community composition. This community change may be
undesirable in some settings.

       7.4.2.3   Physical Factors
       The interaction of a plant with its physical environment (e.g., light, temperature, relative
humidity, soil moisture and wind speed/turbulence) influences the degree and/or nature of the
plant response to 63 exposure.  Light is an essential "resource" whose energy content drives
photosynthesis and CC>2 assimilation. It has been suggested that increased light intensity may
increase the sensitivity of light-tolerant species to 63 while decreasing the  63 sensitivity of
shade-tolerant species, but this appears to be an oversimplification with many exceptions.
       Temperature affects the rates of all physiological processes based on enzyme-catalysis
and diffusion, and each process and overall growth  (the integral of all processes) has a distinct
optimal temperature range. Although some recent field studies have indicated that Os impact
significantly increases with increased ambient temperature, other studies have revealed little
effect of temperature.  Temperature is unquestionably an important variable affecting plant
response to Os in the presence of the elevated CC>2 levels contributing to global climate change
(see below).  In contrast, evidence continues to accumulate to indicate that exposure to 63
sensitizes plants to low temperature stress by reducing below-ground carbohydrate reserves,
                                           7-13

-------
possibly leading to responses in perennial species ranging from rapid demise to impaired growth
in subsequent seasons.
       High relative humidity of the ambient air has generally been found to increase the
adverse effects of 63 by increasing stomatal conductance and thereby increasing 63 flux.
Similarly, abundant evidence indicates that the ready availability of soil moisture results in
greater sensitivity to 63.  The opposite condition, drought, has been observed in field
experiments and modeled in computer simulations to provide partial "protection" against the
adverse effects of Os as would be expected. This is because, in the short-term, drought causes
stomates to close and thus, decrease the exposure to 63. However, there is also compelling
evidence that Os can predispose plants to drought stress. Hence, the response will depend to
some extent upon the sequence in which the stresses occur, and the species-specific nature of the
response.  Regardless of the interaction, however, the net result of drought on growth in the
short-term is negative, although in the case of tree species, other responses such as increased
water use  efficiency could be a benefit to long-term survival.
       Wind speed and air turbulence affect the thickness of the boundary layers over leaves and
canopies and, hence, affects gas exchange rates. These factors can have a significant impact on
the relationship between ambient air exposures and actual exposure concentrations at the leaf or
canopy  surface.

       7.4.2.4   Chemical Factors
       Mineral nutrients in the soil, other gaseous air pollutants, and agricultural chemicals
constitute chemical factors in the environment. The evidence regarding interactions with
specific nutrients is still too contradictory to permit any sweeping conclusions. Somewhat
analogous with temperature, it appears that any shift away from the nutritional optimum may
lead to greater sensitivity, but the shift would have to be substantial before a significant effect on
response to Os was observed.
       Interactions of 63 with other air pollutants have received relatively little recent attention.
The  situation with SO2 remains inconsistent, but seems unlikely to pose any additional risk to
those related to the individual pollutants. With NO and NC>2, the situation is complicated by
their nutritional value as N sources.  In leguminous species,  it appears that NC>2 may reduce the
impact of Os on growth, with the reverse in other species, but the nature of the exposure pattern,
i.e., sequential or concurrent, also determines the outcome. Much more investigation is needed
before it will be possible will be able to predict the outcomes of different Os-NO-NCh exposure
scenarios. The latest research into Os x acid rain interactions has confirmed that, at realistic
acidities, significant interactions are unlikely. A continuing lack of information precludes
offering any generalizations about interactive effects of O3 with NH3, HF, or heavy metals.
                                           7-14

-------
More evidence has been reported that the application of fungicides affords some protective
effects against Os.
       Over the last decade, considerable emphasis has been placed on research into Os
interactions with two components of global climate change: increased atmospheric CCh and
increased mean global temperature. Most of these studies, however, have tended to regard
increased CC>2 levels and increased mean temperatures as unrelated phenomena, in spite of the
crucial role of temperature as a climatic determinant (Monteith and Elston,  1993). Thus,
experiments that examine the effects of doubled CC>2 levels at the current mean ambient
temperatures are not particularly helpful in trying to assess the impact of climate change on
responses to Os, since most of the biotic and chemical interactions with oxidants may be
modified by these climatic changes. Though it is now known from limited  experimental
evidence and evidence obtained by computer simulation that an atmosphere sufficiently enriched
with CC>2 (e.g., 600 + ppm) would more than offset the impact of Os on responses as varied as
wheat yield or the growth of young Ponderosa pine trees, the concurrent increase in temperature
would reduce, but probably not eliminate, the net gain.
       Little, if any, experimental evidence exists related to three-way interactions, such as Os  x
CC>2 x disease or 63 x  CC>2 x nutrient availability. Increased use of computer simulations may
be important in suggesting outcomes of the many complex interactions of Os and various
combinations of environmental factors. However, the results obtained will  only be as reliable as
the input data used for parameterization. Thus, additional data from organized, systematic
studies are needed.
       It is important to recognize that wide variations in net impacts of climate change in
different geographic areas are expected.  Many regions are predicted to experience severe,
possibly irreversible, adverse effects due to climate change. The EPA is currently leading a
research effort that uses regional-scale climate models with the goal of identifying changes to 63
and PM concentrations that may  occur in a warming climate.  An assessment of the results of this
effort is expected to be available for consideration in the next review of the Os NAAQS.

7.5    CHARACTERIZATION OF VEGETATION EXPOSURES TO OZONE
       7.5.1   Key Considerations in Vegetation Exposure Characterization
       In the last review,  the Administrator chose to make the secondary Os NAAQS equal to
the primary standard set as the three year average of the annual 4th-highest daily maximum 8-hr
average concentrations at the level of 0.08 ppm. While recognizing this as  a reasonable policy
choice, she also recognized that "a SUM06 seasonal standard is more biologically relevant and,
therefore, ... also appropriate to consider" (62 FR 38877). This conclusion by the Administrator
in 1997 is again supported by the recent body of science reviewed in the 2006 Os CD (U.S. EPA,
                                          7-15

-------
2006). Staff, therefore, continues to express hourly 63 monitoring data in terms of both the 8-hr
average and seasonal cumulative index forms for comparison. Staff considers the cumulative,
concentration weighted SUM06 and W126 index forms discussed in the 1996 Staff Paper (U.S.
EPA, 1996b). Staff rationale for including the W126 will emerge from the discussions of current
patterns of Os air quality and of policy-relevant background (PRB) in the remainder of this
section. Further, in a letter to the Administrator on October 24, 2006, CASAC indicated a
preference for the W126  form (Henderson, 2006c). Below are the definitions of the three index
forms considered in this review and how they will be referred to in the rest of this document:

   •   8-hr average form: 4th-highest daily maximum 8-hr average over the Os  season.

   •   12-hr SUM06:  3-month sum of all hourly Os concentrations greater than or equal to 0.06
       ppm observed during the daily 12-hr period between 8 am and 8 pm.  The 3 months are
       the maximum consecutive 3 months during the Os season.

   •   12-hr W126: Sigmoidally weighted 3-month sum of all hourly Os concentrations
       observed during the daily 12-hr period between 8 am to 8 pm. The 3 months are the
       maximum consecutive 3 months during the Os season.

       More specifically, W126 is  defined in Lefohn et al., 1988 as:
        i
-------
of 31 ppm-hr that approximates the upper end of the SUM06 range analyzed in the last review
(U.S. EPA, 1996b) and which was associated with up to 17% yield loss in 50% of crop cases.
All approximate equivalency calculations between the 12-hr W126 and 12-hr SUM06 metrics
discussed in Chapter 7 and Chapter 8 of this document were done based on NCLAN data (See
Appendix 7B).
       Since the conclusion of the last review, improvements in 63 air quality have occurred in
some areas of the U.S. In the eastern U.S., these improvements may be attributable in part to the
reductions in NOx emissions resulting from the initiation of Phase II of the acid rain program
(U.S. EPA, 2004) and the NOx SIP call in 2002 (see Chapter 2 of this SP). In addition, efforts to
attain the current NAAQS have no doubt contributed to some air quality improvements,
including lower hourly maximum values and fewer occurrences of those maximum values at
some sites.  One example of this is at the Crestline site in California, where the number of days
with concentrations > 95 ppb has been declining steadily over the last decade, matched by a
decline in peak 1-hr concentrations and 12-hr SUM06 values. These declines match a similar
trend in NOx and reactive organic gases (U.S. EPA, 2006, section Annex  9-207, Figure AX9-17)
(U.S. EPA 2006; Lee et al 2003). However, not all areas in the U.S. show this declining trend.
       The 1997 final rule recognized that "it remained uncertain as to the extent to which air
quality improvements designed to reduce  8-hr Os concentrations would reduce Os exposures
measured by a seasonal SUM06 index" (62 FR 38876). In the last review, staff undertook an
analysis to explore the relationship between the 8-hr average form and the seasonal SUM06
index.  Results of that analysis suggested that many areas that were above the proposed SUM06
standard of 25 ppm-hr were also above the 0.08 primary standard. However, considerable
uncertainty remained as to the  strength of the relationship, especially between urban Os air
quality and distributions that occur in non-monitored rural or remote areas. Using recent (2001-
2004) county-level air quality data, staff has performed a similar analysis  to compare the degree
of overlap between the current level of the 8-hr average form and exposure levels of concern for
vegetation expressed in terms of the 12-hr W126.  Figure 7-1 depicts county Os air quality in
terms of both the current secondary standard 8-hr average form (Y-axis) and the 3-month 12-hr
W126 form (X-axis).  This graph shows the relationship between these two forms averaged over
three recent years (2002-2004). Both the 21 and 13 ppm-hr levels for the  12-hr W126 are
indicated on the graph.  For the 3-year average of 2002-2004, only a few counties would have a
12-hr W126 higher than a level of 21 ppm-hr while meeting the 0.08 level of the current 8-hr
average form. At the lower 12-hr W126 level of 13 ppm-hr, approximately 135 counties would
have a 12-hr W126 higher than a level of 13 ppm-hrs while meeting the 0.08 level of the  current
8-hr average form. Based on this comparison, air  quality levels associated with adverse
                                         7-17

-------
Figure 7-1. The 3-year average (2002-2004) of the 4* -highest maximum 8-hr average (current standard form) versus the 3-

          year average of the highest 3-month 12-hr W126, by county.
 (U
 O)
    0.14-
 (0
CO

£0.12H

CO


f 0.10-1
(U
O) 0.08 -
CO
cB

< 0.06 -
.c
CO

X 0.04 -
CO
    0.02
 O)

 X
    0.00
                                               -    *
                                 Current standard
                                                Alternative standard
                                   -hr W126 options analyzed
           0      10      20      30      40      50      60      70      80      90

                Highest 3-month 12-hrW126 (ppm-hrs), 3-year average
                                                      7-18

-------
vegetation response can occur in some areas that meet the current 8-hr average secondary
NAAQS. The number of counties meeting the current 8-hr average standard and above a 12-hr
W126 standard would obviously depend greatly on the level of 12-hr W126 and 8-hr average
form selected. In addition, the number of counties also varies depending on the air quality of a
particular year or set of years. Figures of the relationship between the current 8-hr average form
and 12-hr W126 for 2002 and 2004 are presented in Appendix 7B. These figures demonstrate
that the relationship between the current 8-hr average form and 12-hr W126 is not constant and
can vary between years.  Thus, staff suggests caution should be used in evaluating the likely
vegetation impacts associated with a given level of air quality expressed in terms of the 8-hr
average form in the absence of parallel 12-hr SUM06 or W126 information.  Unfortunately,
much of the data published both in this review and in other Agency reports only depicts trend
information in terms of the 8-hr average form.
       National parks represent areas of nationally recognized ecological and public welfare
significance, which are afforded a higher level of protection.  Therefore, staff has also focused on
air quality in the subset of national park  sites and important natural areas. Two recent reports
presented some discussion of Os trends in a subset of national parks (See discussion in The
Ozone  Report: Measuring Progress through 2003 (U.S. EPA, 2004) and 2005 Annual
Performance and Progress Report: Air Quality in national parks (NPS, 2005). Unfortunately,
much of this information is presented  only in terms of the current 8-hr average form. Therefore,
staff has selected a subset of national parks and other significant natural areas representing 4
general regions of the U.S to analyze available air quality data in terms of the 12-hr W126 levels
from 2001 to 2005  (Figures 7-2 and 7-3). These graphs show that many national parks and
natural areas have monitored 63 levels above concentrations that have been shown to decrease
plant growth and above the 12-hr W126 levels analyzed in this review. For example, one park in
the east and four parks in the west had more than one year with a 12-hr W126 above 21 ppm-hr.
This level of exposure has been estimated to cause a 9% biomass loss in 50% of the 49 tree
seedling cases studied (Lee and Hogsett,  1996). Sensitive tree seedling species such as black
cherry  (Primus serotina) and ponderosa  pine (Pinusponderosd) have been reported to have 10%
biomass losses at levels as low as 5 and  11 ppm-hr (Lee and Hogsett, 1996).  Impacts on
seedlings may potentially affect long-term growth and survival, ultimately affecting the
competitiveness of sensitive species and individual  trees.
       Another key aspect of evaluating exposure levels of concern to vegetation is
distinguishing between pollution levels that can be controlled by U.S. regulations (or through
international agreements with neighboring countries) from levels that are generally considered
uncontrollable by the U.S., e.g., policy-relevant-background (PRB). As described in Chapter 2
of this  SP, the global photochemical transport model GEOS-CHEM (Fiore et al., 2003) was used
                                          7-19

-------
Figure 7-2. Highest 3-month 12-hr W126 values from monitors in National Parks and
           other natural areas in the Southeast (A) and Northeast (B). Monitors
           designated as GSMNP are found in different areas of the Great Smoky
           Mountain National Park.
                       A.
                              70 -
                           CO  50
                           CM
                              40 -
                           o  20
                           CO
                           x  10 H
                              0 -
GSMNP-Purchase Knob
GSMNP-Clings. Dome
GSMNP-Look Rock
GSMNP-Cades Cove
Shenandoah NP
                                   2001
                                            2002
                                                     2003

                                                     Year
                                                              2004
                                                                       2005
                      B.
                             80 -,
                             70 -
                          1.  60
                             50 -
                             40 -
                          o
                          CO  20
                             10 -

	 o 	
	 V 	
Base Whiteface Mtn.NY
Summit.Whiteface Mtn.NY
Mt. Washington, NH
Mt. Greylock, MA
Acadia NP, ME
                                  2001
                                           2002
                                                    2003

                                                    Year
                                                             2004
                                                                      2005
                                            7-20

-------
Figure 7-3. Highest 3 month 12-hr W126 values from monitors in National Parks in the
          Mountain West (A) and California (B).
 A.
B.
80


70
       0. 60
       CD
       ™  50
       CN
       X
       co
          40 -
       ~  30 H
       o
       E
       co  20
           10 -
          80



       ? 70

       Q.
       S 60
       CD
       CN
       5  50

       1_

       c5  40

       .c
          30


	 o 	
	 T 	
	 v 	
Rocky Mtn, NP
Yellowstone NP
Grand Canyon NP
Glacier NP
y —
2001

	 V 	
	 v 	
2002 2003
Year
"""••o 	
-v
2004

	 o
2005

o
co 20 -
X
CO
10 -
0 -
o


• Sequoia NP
o Yosemite NP



2001 2002 2003 2004 2005
Year
                                         7-21

-------
to estimate PRB levels. This model shows that PRB O3 concentrations, which vary as a function
of season, altitude and total surface O3 concentration, are generally predicted to be in the range
of 0.015 to 0.035 ppm at the surface in the afternoon, and they decline under conditions
conducive to O3 episodes. They are highest during spring and decline into summer. Higher
values tend to occur at higher elevations during spring due to contributions from hemispheric
pollution and stratospheric intrusions. The stratospheric contribution to surface O3 is typically
well below 0.020 ppm and only rarely elevates O3 concentrations at low-altitude sites and only
slightly more often elevates them at high-altitude sites (U.S. EPA, 2006, AX3-148).
      The modeled range of 0.015 to 0.035 ppm in the 2006 CD is lower than the 0.03 to 0.05
ppm range used as background O3 in the 1996 O3 NAAQS review (U.S. EPA, 1996a, 2006).
This is significant for the secondary standard review because the higher end of the range (0.05
ppm)  provided an important policy consideration for staff in 1996 for selecting the cumulative
SUM06 exposure index that did not weight concentrations below 0.06 ppm.  Thus, SUM06 was
not influenced by concentrations thought to be at background.
      Partially on the basis of these lower estimates of PRB, as well as declining peak O3 levels
at some  sites, staff has re-evaluated the usefulness of using the sigmoidally weighted W126
index to capture more of the vegetation relevant exposures below 0.06 ppm.  Though the W126
index weights all concentrations, the concentrations below 0.04 ppm receive substantially
smaller weights (3 percent or less) so as not to contribute significantly to the value of the index
(Lefohn et al. 1988). Indeed, a constant concentration of the highest estimated PRB (0.035  ppm)
would only add up to a 3-month  12-hr W126 of less than 1 ppm-hr. In addition,  because the
W126 form does not contain an absolute threshold like the SUM06 form, it is more in keeping
with scientific consensus that there is no threshold for exposures that cause effects on vegetation
(Heck and Cowling  1997, U.S. EPA 2006). Further, CAS AC has indicated a preference for the
12-hr W126 metric (Henderson, 2006c). Therefore, staff has continued to include the 12-hr
W126 in the vegetation risk analyses. Figure 7-4 shows the relationship between 12-hr W126
and SUM06 as measured at O3 monitors in 2001. The metrics, as calculated at the monitors, are
highly correlated. A similar correlation was seen with other years (2002-2004).  Because the
inflection point of W126 is approximately 0.06 ppm, the SUM06 metric is essentially a simple
approximation of the sigmoidally weighted W126 form and it is not surprising that the two
metrics measure O3  exposures in a very similar way at most monitoring stations  (Lee et al.
1989). Finally, the W126 metric should also be easier to model than SUM06 since small errors
in prediction of hourly concentrations around 0.06 ppm could cause variations in the SUM06
metric.  This issue is avoided in the continuous weighting of the W126.
                                          7-22

-------
Figure 7-4. Maximum 3-month 12-hr SUM06 plotted against maximum 3-month 12-hr
          W126.  Data points are from the AQS and CASTNET O3 monitors for the year
          2001.
    90
 Q.
CD
CN
 O
 E
 co
                           proposed 1996 SUM06 standard
                                                   r2 = 0.98
           10    20    30    40    50     60     70

                    3-month 12-hrSUM06 (ppm-hr)
80
90
                                       7-23

-------
       7.5.2   Monitor Networks: National Coverage
       Hourly Os monitoring data are available from two national networks: (1) Air Quality
System (AQS; http://www.epa.gov/ttn/airs/airsaqs) and (2) Clean Air Status and Trends Network
(CASTNET; http://www.epa.gov/castnet/). The locations of these monitors are presented in
Figure 7-5 and are described in sections 2.3.1 and 2.3.2 of this document.  The AQS monitoring
network currently has over 1100 active 63 monitors which are generally sited near population
centers. However, this network also includes approximately 36 monitors located in national
parks.  CASTNET is the nation's primary source for data on dry acidic deposition and rural,
ground-level 63. It consists of over 80 sites across the eastern and western U.S. and is
cooperatively operated and funded with the National Park Service. Due to the overall stability in
these monitoring networks and standardized, rigorous QA/QC and data handling protocols, they
provide useful information regarding long-term trends in air quality across regions and at
specific sites. For more on the AQS protocols, see section 2.3.1 of this Staff Paper or Code of
Federal Regulations, Title 40, Part 58 (40 CFR Part 58). CASTNET, in terms of data quality,
achieved 98% to 99% of all precision and accuracy audits being within the ±10% criteria for
both precision and accuracy. Overall, CASTNET Os monitors are stable and show only very
small variation (U.S. EPA 2003b, p.22). Both networks take Os measurements on an hourly time
step which allows for quick comparisons between different air quality index forms and different
averaging times.
       In spite of the size and quality of these monitoring networks, however, vast rural areas of
the U.S.,  where important crops and natural vegetation occur, still do not have Os monitor
coverage (Figure 7-5). As was the case in the 1996 review, staff found it necessary to select a
method that could be used to characterize 63 air quality over broad geographical areas of
concern (see sections 7.5.3 and 7.5.4 below) to support a national scale risk assessment of the
effects of ambient 63 exposures on vegetation and ecosystems.  Staffs review of the monitoring
data showed that within the five most recent years available (2000 to 2004), 2001 was a fairly
moderate Os year. Based on this information, and because it coincided with the most recently
available air quality model data (see section 7.5.3. below), 2001 was selected as the initial (base)
air quality year for most of the quantitative vegetation risk analyses conducted in this review. In
a few cases (e.g., visible foliar injury and tree growth modeling),  monitoring data from other air
quality years were used.
                                          7-24

-------
Figure 7-5. Locations of AQS monitors (top) and CASTNET monitoring stations (bottom)
                                             A  *
                                      \:±iM
                                7-25

-------
       7.5.3   Community Multi-scale Air Quality Model (CMAQ)
       Staff investigated the appropriateness of using the 63 outputs from the EPA/NOAA
Community Multi-scale Air Quality (CMAQ) model system
(http://www.epa.gov/asmdnerl/CMAQ, Byun and Ching, 1999; Arnold et al. 2003, Eder and Yu,
2005) to improve spatial interpolations based on the regionally limited and unevenly distributed
Os monitoring network in the western U.S. (see section 7.5.2). The CMAQ model is a multi-
pollutant, multiscale air quality model that contains state-of-the-science techniques for
simulating all atmospheric and land processes that affect the transport, transformation, and
deposition of atmospheric pollutants and/or their precursors on both regional and urban scales. It
is designed as a science-based modeling tool for handling many major pollutants (including
photochemical oxidants/Os, particulate matter, and nutrient deposition) holistically. The CMAQ
model can generate estimates of hourly O3 concentrations for the contiguous U.S., making it
possible to express model outputs in terms of a variety of exposure indices (e.g., W126, 8-hr
average). Due to the significant resources required to run CMAQ, however, model outputs are
only available for a limited number of years. For this review, 2001 outputs from CMAQ version
4.5 were the most recent available. This version of CMAQ utilizes the more refined 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.  The 12
km x 12 km domain covers an area from roughly central Texas, north to North Dakota, east to
Maine, and south to central Florida. More detailed information on CMAQ can be found in
Appendix 7C.  Section 7.5.4 below describes the very limited capacity in which staff used the
CMAQ results. As explained below, in the final analysis, staff opted not to use 63 values
calculated from the CMAQ model, but instead only used model results to scale  interpolations  in
the western U.S.

       7.5.4   Generation of Potential Ozone Exposure  Surfaces (POES)
       Staff evaluated ten approaches for interpolating 63 air quality across the U.S. which
included (1) use of the CMAQ model  alone; (2) use of only monitoring data with the Voronoi
Neighbor Averaging (VNA) technique; and (3) use of a combination of monitoring data and
scaling from CMAQ called enhanced Voronoi Neighbor Averaging (eVNA). The evaluations
were based on how well the CMAQ model or interpolation techniques were able to predict the
12-hr SUM06, 12-hr W126 and the current 8-hr average form measured at each monitor.  For
VNA and eVNA evaluations, each monitor was dropped out sequentially and a  value for that
monitor was interpolated using the remaining monitors. At each monitor site, Normalized Mean
Bias (NMB), Normalized Mean Error (NME), Absolute Mean Bias (AMB) and Absolute Mean
Error (AME) were calculated (Table 7-1).  For more details see discussions in section 7.5.5
below and in the Environmental Assessment TSD  (Abt, 2007). From the results of these
                                          7-26

-------
evaluations, the eVNA and VNA performed equally well in many cases.  The CMAQ model
alone did not perform as well as the VNA and eVNA methods. The staff chose to use separate
interpolation techniques in the east and the west.  The simpler VNA approach was chosen for the
eastern U.S. since it was determined that enhancing the interpolation with CMAQ did not add
much information to the eastern U.S. interpolation where the monitoring network has greater
coverage than in the west (Figure 7-5). In the west, eVNA was chosen because of the sparse
monitoring network in those states.  Although the VNA and eVNA interpolation approaches are
not as complex or sophisticated as some techniques (e.g., Bayesian methods), they have the
advantages of relying on readily available data, being relatively inexpensive to run, and being
able to quickly produce estimates of any exposure index, for multiple months or years, and for
different air quality scenarios.
       To generate the POES, a set of geographical locations for which 63 data would be
interpolated was needed. Ideally these locations would be regularly spaced, cover the
continental US, and be close enough to each other to provide a good spatial resolution.  Staff
chose to use the regularly spaced grid structure of the CMAQ model as the basis for these
locations.  Specifically,  the center of each grid cell was identified both for cells in the 12 km x 12
km grid (which covers only the Eastern U.S.), and the 36 km x 36 km grid (the Western US).
This approach produced the densest possible non-redundant "composite" grid of 44432 regularly
spaced grid cell center locations throughout the U.S. Using VNA in the eastern U.S. and eVNA
in the West, 63 values were interpolated for each grid cell center in the composite grid (see
Environmental Assessment TSD for more details, Abt, 2007).
       To support the vegetation exposure and risk assessments, ambient Os exposures were
projected using seasonal 63 air quality for the 2001 base year in terms of the 3-month 12-hr
W126 (Figure 7-6) and  12-hr SUM06 exposure indices (Figure 7D-1 in Appendix 7D).  The
uncertainties of this interpolation are discussed below (section 7.5.5).  Taking the uncertainties
into account and given the absence of more complete monitoring data in  rural areas, staff finds
the POES serves as a useful tool for identifying areas across the country where Os exposure
levels would be expected to exceed those known to produce yield loss or biomass loss at given
levels for crops and trees, respectively.
       Figure 7-6 suggests that under the base year (2001) air quality, a large portion of
California had a 12-hr W126  above 31 ppm-hr which has been reported to produce 14% biomass
loss in 50% of tree seedlings  studies by National Health and Environmental Effects Research
Lab, Western Ecology Division (NHEERL-WED). Broader multistate regions in the east and
west were predicted to have 12-hr W126 above 21 ppm-hr, which is approximately equal to the
secondary standard proposed  in 1996. A 12-hr W126 ppm-hr of 21 is associated with a 9%
biomass loss in 50% of  tree seedlings studied (Lee and Hogsett,  1996). Much of the east and
                                          7-27

-------
Figure 7-6. Estimated 12-hr W126 Ozone Exposure - Max 3-months for 2001: "As Is" scenario
 W126 (ppm-h)
 I   I <=7 (min = 0.5)
 I   |731 (max = 62)
                                                           7-28

-------
Arizona and California have seasonal W126 values above 13 ppm-, which has been associated
with a 7% biomass loss in 25% of tree seedlings studied (Lee and Hogsett,  1996). This indicates
that current air quality levels could result in significant impacts to vegetation in some areas.
However, these exposures may be uncertain in some cases with respect to vegetation with
canopy heights below monitor inlet heights, e.g., crops and tree seedlings.  In the crop and tree
seedling risk/benefit assessments, staff incorporated an adjustment of monitored 63 to take into
account the uncertainty associated with a potential vertical O3 gradient from the height of the
monitoring probe (~4 meters) to the approximate canopy height of crops and seedlings (see
section 7.6.2.3).
       To evaluate changing vegetation exposures and risks under changing air quality, maps
were generated for selected "just meet" scenarios (Figures 7-7, 7-8, 7-9, 7-10) by analytically
adjusting air quality distributions with the quadratic method to reflect "just meeting" the level of
various alternative standards (see Horst and Duff, 1995; Rizzo, 2005 & 2006). This technique
combines both linear and quadratic elements to reduce larger 63 concentrations more than
smaller ones. In this regard, the quadratic method attempts to account for reductions in
emissions without greatly affecting lower concentrations near ambient background levels.  The
following "just meet" air quality scenarios were generated:
     •   4th-highest daily maximum 8-hr average of 0.084 ppm (current EPA standard) and
         0.070 ppm
     •   3-month, 12-hr SUM06 of 25 ppm-hr (alternate standard proposed in the 1996 NAAQS
         review) and 15 ppm-hr
     •   3-month, 12-hr W126 of 21 ppm-hr and 13 ppm-hr

Maps generated for the SUM06 25 and 15 ppm-hr scenarios were nearly identical to the maps of
12-hr W126 levels of 21  and 13 ppm-hr and thus, only maps of the SUM06 25 and 15 ppm-hr
scenarios are displayed.  When 2001 air quality was rolled back to meeting the level of the
current 8-hr standard (0.08 ppm), the overall seasonal 12-hr W126 exposures did not improve
very much (Figure 7-7).  Under this scenario, some areas in the east improve, but there are still
many areas of the country that have seasonal Os levels above 12-hr W126 of 21 ppm-hr.
       The exposure maps generated for the 0.070 ppm level of the 8-hr average form (Figure 7-
8), 12-hr SUM06 of 25 and 15 ppm-hr alternatives (Figures 7-9 and 7-10) and 12-hr W126 of 21
and 15 ppm-hr showed a markedly improved picture of 63 air quality compared to Figure 7-7.
Thus, the staff observes that, except for just meeting the current form, all other alternative
standards, when met at all locations, would be expected to provide improved protection of
                                          7-29

-------
                                                                                                         ,th
Figure 7-7. Estimated 12-hr W126 Ozone Exposure - Max 3-months for 2001: Quadratic Rollback to just meet 4  -Highest 8-
          hr Maximum of >0.084
 W126 (ppm-h)
 |   | <=7 (min = 0.5)
 I   |731 (max = 41)
                                                         7-30

-------
Figure 7-8. Estimated 12-hr W126 Ozone Exposure - Max 3-months for 2001:  Quadratic Rollback to just meet 4th Highest 8-
          hr Maximum of >0.070
 W126 (ppm-h)
 |   | <=7 (min = 0.5)
 I   |721 (max = 23)
                                                         7-31

-------
Figure 7-9. Estimated 12-hr W126 Ozone Exposure - Max 3-months for 2001:  Quadratic Rollback to just meet 12-hr SUM06
          of 25 ppm-hr, secondary standard proposed in 1996
 W126(ppm-h)
 I   I <=7 (min = 0.5)
 I   1?21 (max = 21 3)
                                                        7-32

-------
Figure 7-10. Estimated 12-hr W126 Ozone Exposure - Max 3-months for 2001: Quadratic Rollback to just meet 12-hr
          SUM06 of 15 ppm-hr
 W126 (ppm-h)
 I   I <=7 (min = 0,5)
 I   |713 (max= 15)
                                                        7-33

-------
vegetation from seasonal 63 exposures of concern over the current standard. As expected,
however, the greatest improvements in air quality and estimated exposures to sensitive
vegetation were observed when just meeting the lower 12-hr W126 of 13, 12-hr SUM06 of 15
and 0.07 ppm, 8-hr average scenarios.

       7.5.5   Uncertainties in the Os Exposure Analysis
       Staff recognizes there are inherent uncertainties in using an interpolation that must rely
on sparse data that, for the most part, are representative of urban and near-urban areas. This
network could bias the picture of the O3 exposure estimate especially in the western U.S. where
monitoring sites can be very far apart.  Intuitively, it is expected that the eVNA approach with
spatial scaling from CMAQ approach would be an improvement over a simple interpolation in
the West. However, it is difficult to test for this because of the paucity of monitoring sites in the
western U.S.
       To quantify the uncertainty associated with the exposure surface, each monitoring site
was sequentially dropped out of the interpolation and recalculated with the remaining monitoring
sites.  At each monitoring site, Normalized Mean Bias (NMB), Normalized Mean Error (NME),
Absolute Mean Bias (AMB) and Absolute Mean Error (AME) were calculated. These statistics
are defined below:
                           „ predictedMETRIC  - actualMETRIC ^
  NMB = average Jedrovouts (100 * ^	'•	'-)
              5 
-------
large while the absolute difference is small.  For example, if a monitor with a W126 of 4 ppm-hr
is measured and the interpolation predicts a W126 of 6 ppm-hr then the NME would be 50%.
Therefore, staff thought it was useful to also report the absolute mean bias and error. In absolute
terms, the average bias for SUM06 was slightly low (-1.83 ppm-hr in the East and -2.62 ppm-hr
in the West).  CASTNET monitors are also presented to illustrate how well the interpolation
techniques predicted air quality in that rural monitoring network. In general, the interpolations in
the East and West under-predicted the  12-hr SUM06 values. This under-prediction is likely a
result of the averaging inherent in the interpolation. Similar results are seen for the 12-hr W126
(Table 7-lb).  However, in almost all cases, the interpolation was able to predict monitored
W126 slightly better than monitored SUM06.  The calculation of error and bias metrics for the
interpolation represents a notable improvement over the 1996 assessment which did not have an
evaluation of the error and bias associated with the exposure surface.
       Figure 7-11 also depicts predicted W126 values, from the sequential drop-out exercise,
against the actual W126 values measured at CASTNET monitors and AQS monitors designated
as "rural." This figure gives a graphical representation of how well the O3 exposures were
predicted in the rural monitors away from urban areas.  A perfect prediction would result in all
points aligning on the black "one-to-one" line.  In general, these graphs indicate that the
interpolation technique slightly overestimated W126 exposure at the low levels and
underestimated W126 exposure at the high levels.  Biologically, the more significant error is at
the high exposures, since vegetation responds more at high exposures. Figure 7-11 indicates, in
general, that the most relevant high exposures were underestimated. This may have implications
for the subsequent calculation of crop yield and tree seedling biomass loss, potentially resulting
in an underestimation of risk in some areas.  More detailed information from this analysis is
presented in the Environmental Assessment TSD (Abt, 2007).
                                          7-35

-------
Table 7-la. Evaluation statistics for the 3-month 12-hr SUM06 interpolations of the Eastern and Western U.S. domains.
          NMB is Normalized Mean Bias, NME is Normalized Mean Error, AMB is Absolute Mean Bias and AME is
          Absolute Mean Error.  An explanation of these metrics is given in section 7.5.5.
Region
Eastern US
Eastern US
Western US
Western US
Monitors
All monitors
CASTNET only
All monitors
CASTNET only
NMB (%)
-0.04
-8.84
16.46
-6.03
NME (%)
25.78
20.76
62.39
42.12
AMB (ppm-hr)
-1.83
-2.95
-2.62
-2.15
AME (ppm-hr)
4.07
4.79
6.05
7.98
Table 7-lb. Evaluation statistics for the 3-month 12-hr W126 interpolations of the Eastern and Western U.S. domains.
Region
Eastern US
Eastern US
Western US
Western US
Monitors
All monitors
CASTNET only
All monitors
CASTNET only
NMB (%)
-1.06
-8.43
14.57
0.67
NME (%)
21.92
17.44
48.38
41.47
AMB (ppm-hr)
-1.21
-2.00
-1.50
-0.60
AME (ppm-hr)
2.97
3.22
4.27
5.21
                                                        7-36

-------
Figure 7-11. Comparison of predicted versus observed 12-hr W126 at CASTNET and
          "rural" AQS monitors. Monitor data was predicted by dropping out each
          monitor sequentially and interpolated with the all remaining monitors.
    35
    30 -
 CD
 (N

 §  25-
o
E
ro
T3
    20 -
    15 -
 ^  10"
 l_
 CL
     5 H
        East
                                    One-to-one line
                                    	Best fit line  r2 = 0.60
              5      10      15      20       25      30

                  Observed 3-month 12-hr W126 (ppm-h)
                                                           35
             10     20      30      40      50     60

                  Observed 3-month 12-hr W126 (ppm-h)
                                                        70
                                         7-37

-------
7.6    CHARACTERIZATION OF VEGETATION RISKS
       7.6.1   Scope of Vegetation Risk Assessment
       The vegetation impact assessment conducted for the current review (see Figure 7-12a-c),
consists of exposure, risk and benefits analyses and improves and builds upon the similar
analyses performed in support of the 1996 secondary NAAQS review (U.S. EPA 1996b). The
vegetation exposure assessment was discussed above in section 7.5. The organization of this
section reflects the remaining risk and benefit components of the assessment.  The vegetation
risk discussion which follows is divided between the crop and tree analyses.  The crop analysis
discussed in section 7.6.2 includes estimates of the risks to crop yields from current and
alternative Os exposure conditions and the associated change in economic  benefits expected to
accrue in the agriculture sector upon meeting the levels of various alternative standards.  The tree
risk analysis described in section 7.6.3 includes three distinct lines of evidence: (1) estimates of
seedling growth loss under current and alternative Os exposure conditions; (2) observations of
visible foliar injury in the field linked to recent monitored Os air quality for the years 2001 -
2004; and (3) simulated mature tree growth reductions using the TREGRO model to simulate the
effect of meeting alternative air quality standards on the predicted annual growth of a single
western species (ponderosa pine) and two eastern species (red maple and tulip poplar). Both
quantitative and qualitative discussions of known sources and ranges of uncertainties associated
with the components of this assessment are also discussed.
                                          7-38

-------
Figure 7-12 (a-c).   Major Components of Vegetation Risk Assessment
                                   Interpolation q
                                  regional potential
                                   ozone exposu
                                     surface
        Crop yield
County USDA
crop planting
data

NCLAN I
C-R
functions J
Seedling growth
                                   Calculation
                                   of seedling
                                    rowth loss
County
USFS foliar
injury data

County 'As
is' air
quality
                                                                               'As is' &
                                                                              'just meet'
                                                                              air quality
                                                                              from east
                                                                               Swest
                                                                                sites
   (c) Economic Benefits
                                                7-39

-------
       7.6.2   Characterization of Crop Risks and Associated Economic Benefits
       7.6.2.1   Exposure Methodologies Used in Vegetation Research
       In the 1996 review, Os exposure studies were dominated by the use of various versions of
the open-top chamber (OTC), first described by Heagle et al. (1973) and Mandl et al. (1973).
Hogsett et al. (1985, 1987) described in detail many of the subsequent modifications to the
original OTC design. The OTC method continues to be a widely used technique in the U.S. and
Europe for exposing plants to varying levels of Os (U.S. EPA, 2005b).
       Chambered systems, including OTCs, have several advantages. For instance, they can
provide a range of treatment levels including charcoal-filtered (CF), clean-air control, and above
ambient concentrations for Os experiments. Depending on experimental intent, a replicated,
clean-air control treatment is an essential component in many experimental designs. The OTC
can provide a consistent, definable exposure because of the constant wind speed and delivery
systems. From a policy perspective, the statistically robust concentration-response (C-R)
functions developed using such systems are necessary for evaluating the implications of various
alternative air quality scenarios on vegetation response.
       Nonetheless, there are several characteristics of the OTC design and operation that can
lead to exposures that might differ from those experienced by plants in the field. First,  the OTC
plants are subjected to constant turbulence, which, by lowering the boundary layer resistance to
diffusion, which may result in increased uptake.   This may lead to an overestimation of effects in
areas with less turbulence (Krupa et al., 1995; Legge et al.,  1995). As with all methods that
expose vegetation to modified O3 concentrations  in chambers, OTCs create internal
environments that differ from ambient air.  This so-called "chamber effect" refers to the
modification of microclimatic variables, including reduced  and uneven light intensity, uneven
rainfall, constant wind speed, reduced dew formation, and increased air temperatures (Fuhrer,
1994; Manning and Krupa, 1992).  However, staff notes that the uncertainties associated with the
influence of other modifying factors occurring in the field such as water and nutrient availability
(see discussion above in section 7.4.2), are likely to be greater than the uncertainties in  the data
due to the influence of OTCs. Because of the standardized  methodology and protocols used in
National Crop Loss Assessment Network (NCLAN) and other programs, the database can be
assumed to be internally consistent.
       While it is clear that OTCs can alter some aspects of the microenvironment and plant
growth, the question to be answered is whether or not these differences affect the relative
response of a plant to Os.  As noted in the 1996 Os CD (U.S. EPA, 1996a), evidence from a
number of comparative studies  of OTCs and other exposure systems suggested that responses
were essentially the same regardless of exposure  system used and chamber effects did not
                                          7-40

-------
significantly affect response.  For example, a study of chamber effects examined the responses of
tolerant and sensitive white clover clones (Trifolium repens) to ambient Os in greenhouse, open-
top, and ambient plots (Heagle et al., 1996). The response found in OTCs was the same as in
ambient plots.  The California Air Resources Board (CARB), during its recent 63 standard
review, came to a similar conclusion about the usefulness of OTC data. Its review states "there
is little scientific justification for the categorical discounting of 63 yield-response relationships
obtained using the OTC technology" (CEPA, 2005).
       In recent years, a few studies have employed a modified Free Air CC>2 Enrichment
(FACE) methodology to expose vegetation to elevated 63 without using chambers.  This
exposure method was originally developed to expose vegetation to elevated levels of CC>2, but
has been modified to include Os exposure in Illinois (SoyFACE) and Wisconsin (AspenFACE)
for soybean and deciduous trees, respectively (Dickson et al., 2000; Morgan et al., 2004). The
FACE method releases gas (e.g., CC>2, Os) from a series of orifices placed along the length of the
vertical pipes surrounding a circular field plot and uses  the prevailing wind to distribute it. This
exposure method may more closely replicate conditions in the field and, more importantly for
forest research, has the benefit of being able to expand vertically with the growth of the trees,
allowing for exposure experiments to span numerous years.
       The FACE methodology has a different set of limitations than those of the OTC.  Most
importantly, it  is not possible with FACE to produce  a number of replicated treatment levels,
including 63 concentrations below ambient levels that are needed to build the statistically robust
C-R functions possible with OTCs.  One also must recognize the potential for significant
gradients of exposure gas concentrations throughout the FACE exposure rings. While the FACE
protocols minimize exposure concentration gradients, plants near the gas emitters will be
exposed to larger concentrations than centrally located plants near the air monitoring point.
There is little information on within-plot Os concentrations in FACE-type exposures and this
issue needs to be addressed more fully to understand  O3 exposure and response data from FACE
studies. Despite the differences in these two exposure methods, recent evidence obtained using
FACE and OTC systems appear to support the results observed in OTC studies used in the 1996
review. For example, a series of studies undertaken at AspenFACE (Isebrands et al., 2000,
2001) showed that Os-symptom expression was generally similar in OTCs, FACE, and ambient-
Os gradient sites, and supported the previously observed variation among trembling aspen clones
(Populus tremuloides L.) using OTCs (Karnosky et al.,  1999).
       In the SoyFACE experiment in Illinois,  soybean (Pioneer 93B15 cultivar) yield loss data
from a two-year study was recently published (Morgan et al., 2006). This cultivar is a recent
selection and, like most modern cultivars, has been selected with an already high current Os.  It
was found to have average sensitivity to Os compared to 22 other cultivars tested at SoyFACE.
                                         7-41

-------
In this experiment, ambient hourly 63 concentrations were increased by approximately 20% and
measured yields were decreased by 15% in 2002, as a result of the increased Os exposure
(Morgan et al., 2006).  To compare these results to chamber studies, Morgan et al. (2006)
calculated the expected yield loss from a linear relationship constructed from chamber data using
7-hr seasonal averages (Ashmore, 2002).  They calculated an 8% expected yield loss from the
2002 63 exposure which was surprisingly less than the measured 15% yield loss.  Staff believes
that the expected yield loss may have been closer if the authors used C-R functions based on the
W126 metric. Nonetheless, the results from this study suggest that C-R relationships developed
from chambers are not overestimating response of recently developed soybean cultivars to
elevated Os exposure.  As more FACE data become available, a more quantitative comparison of
findings from the SoyFACE and AspenFACE systems would be useful.
      Other exposure methods described both in the 1996  and 2006 O3 CDs (U.S. EPA, 1996a;
U.S. EPA 2006) also provided useful information on plant responses to Os exposure. For
example, Gregg et al. (2003), found significant effects of 63 on the growth of cottonwood
saplings along an ambient O3 gradient in the New York City area, similar to those reported in
OTCs (see section 7.6.3. Other exposure methods include but are not limited to chemical
protectants (e.g., ethylenediurea [EDU]) and 63 exclusion.  Nonetheless, given a continued
policy need for robust C-R functions to evaluate vegetation response under alternative air quality
scenarios and the apparent consistency between plant responses using OTC and other methods,
staff concludes that the robust C-R functions derived using the OTC methodology are currently
the most useful in a policy context and we continue to rely on them in the  following analyses.

      7.6.2.2   Basis for C-R Functions
      The 1996 crop assessment was built upon the NCLAN Os C-R functions. Since very few
new studies have published C-R functions that would be useful in an updated assessment, C-R
functions from NCLAN remain the best data available for a national assessment of crop loss
under various Os air quality scenarios. The NCLAN protocol was designed to produce crop C-R
functions representative of the areas in which the crops were typically grown.  The U.S. was
divided into 5 regions over which a network of field sites was established. In total, 15 crop
species (corn, soybean, winter wheat, tobacco, sorghum, cotton, barley, peanuts, dry beans,
potato, lettuce, turnip, and hay [alfalfa, clover, and fescue]), were studied. The first 12 of these
15 listed species were analyzed for the 1996 review and included 38 different cultivars studied
under a variety of unique combinations of sites, water regimes, and exposure conditions,
producing a total of 54 separate cases. Figure 7-13 uses the regression equations for each of the
54 cases to graph predicted relative yield loss at various exposure levels in terms of a 12-hr
W126 (Figures 7E-1, 2, 3 present similar figures with the 8-hr average and 12-hr SUM06 forms).
                                          7-42

-------
Figures 7-14 (a-d) show composite graphs for some individual crops from NCLAN and the
variations in sensitivity between important crops.  According to the most recent USDA National
Agricultural Statistical Survey (NASS) data, the 12 species analyzed in the last O3 NAAQS
review account for greater than 70% of principal crops acreage planted in the U.S. in 2004.1
Corn, soybean, and winter wheat alone accounted for 62% of 2004 principal crop acreage
planted.  For the economic analysis described in section 7.6.2.4, a reduced list of 9 species (69%
of 2004 principal crops) were included (e.g., cotton, field corn, grain sorghum, peanut, soybean,
winter wheat, lettuce, kidney bean, potato), with tobacco, turnip and barley not evaluated.
       Since the NCLAN studies were performed during the years 1980 to 1988, there is some
uncertainty whether the crop cultivars tested in NCLAN are representative of crops grown today.
In general, new crop varieties are not specifically bred for O3 tolerance and the cultivars used
today were bred from the same very narrow genetic stock available in the 1980's.  Thus, it is not
expected that there would be much difference in O3 tolerance between cultivars used today and
when the NCLAN studies were done.  Since the last review, there has been no evidence that
crops are becoming more tolerant of O3 (U.S. EPA, 2006). For cotton, some newer varieties
have been found to have higher yield loss due to O3 compared to older varieties (Olszyk et al.,
1993, Grantz and McCool, 1992). In a meta-analysis of 53 studies, Morgan et al. (2003) found
consistent deleterious effects of O3 exposures on soybean from studies published between 1973
and 2001. Further, early results from the SoyFACE experiment in Illinois indicate a lack of any
apparent difference in the O3 tolerance of old and recent cultivars of soybean in a study of 22
soybean varieties (Long et al., 2002).
       1 Principal crops as defined by the USDA include corn, 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 were taken from the USDA NASS
2005 Acreage Report (http://usda.mannlib.cornell.edu/reports/nassr/field/pcp-bba/acrg0605.pdf).

                                           7-43

-------
Figure 7-13.  Median crop yield loss from NCLAN crops characterized with the 12-hr
            W126
                                                                 75th Percentile
         -r
20       30       40

     12-hr W126(ppm-hr)
                                                    1
                                                    50
 1 - 1
60
Distribution of yield loss predictions from Weibull exposure-response models that relate yield to O3 exposure
characterized with the 12-hr W126 statistic using data from 31 crop studies from National Crop Loss Assessment
Network (NCLAN). Separate regressions were calculated for studies with multiple harvests or cultivars, resulting in
a total of 54 individual equations from the 31 NCLAN studies. Each equation was used to calculate the predicted
relative yield orbiomass loss at a 12-hr W126 of 10, 20, 30, 40, 50, and 60 ppm-hr, and the distributions of the
resulting loss were plotted. The solid line represents the 50th percentile. Source: U.S. EPA, 1996a; Lee and Hogsett
1995.
                                                7-44

-------
Figure 7-14 (A-D). Median soybean (A), wheat (B), cotton (C) and corn (D) yield
             loss from NCLAN crops characterized with the 12-hr W126
        A. Soybean
               20   30    40   50   60
                 12-hrW126(ppm-hr)
100 -
90 -
£ 80 -
J3 70 -
•a
« 60-
5=
> 50 -
15
« 40-
S> 30 -
H 20 •
0-
10 -
n -
B. Wheat 	 75th Percentile




| 	 1
Lx--
	

	

S
	

^

x



^





7
50th Percentile
/
| 	 25th Percentile



20    30   40   50

   12-hrW126 (ppm-hr)
                               75th Percentile
                                50th Percentile
                                25th Percentile
                                                                   	O	
               12-hr W126 (ppm-hr)
                                                                   12-hr W126 (ppm-hr)
Distribution of yield loss predictions from Weibull exposure-response models that relate yield to O3
exposure characterized with the 12-hr W126 statistic using data from 22 soybean, 7 wheat, 9 cotton and 2
corn studies from National Crop Loss Assessment Network (NCLAN). Separate regressions were
calculated for studies with multiple harvests or cultivars. Each equation was used to calculate the predicted
relative yield loss at a 12-hr W126 of 10, 20, 30, 40, 50, and 60 ppm-hr, and the distributions  of the
resulting loss were plotted. Source: U.S. EPA, 1996a; Lee and Hogsett 1995.
                                              7-45

-------
       7.6.2.3   Considerations for Exposures at Crop Canopy Height
       An important consideration when predicting crop yield and/or tree seedling biomass loss
using monitored Os exposure levels is the potential positive exposure bias associated with the
height at which the measurement is taken.  Ambient monitor inlets are typically at heights of 3 to
5 meters, and thus are located in the inner part of the planetary boundary layer (U.S. EPA,
2005b). It is well known that within this layer Os reacts with vegetation, other surfaces and
volatile compounds and can create a vertical gradient of decreasing 63 concentration from the
inlet height of the monitors to the canopies of short vegetation.  The magnitude of the gradient is
determined in large part by the intensity of turbulent mixing in the surface layer. During daytime
hours, the vertical 63 gradient is relatively small because turbulent mixing maintains the
downward flux of Os.  For example, Horvath et al. (1995) calculated a 7% decrease in Os going
from a height of 4 meters down to 0.5 meters above the surface during unstable (or turbulent)
conditions in  a study over low vegetation in Hungary [see section AX3.3.2. of the 2006 CD (U.S.
EPA, 2006)].  This is compared to a 20% decrease during stable conditions, which usually occur
during the night. The average decrease for all times measured was 10%. The daytime versus
nighttime bias is an important distinction since the assessments outlined below rely heavily  on
daytime metrics, such as the!2-hr SUM06 and W126.  Thus, staff selected 10% as a daytime
downward adjustment factor to apply to hourly monitor-derived exposures (including
interpolated values) prior to estimating crop yield and tree seedling biomass loss values. We
consider this  10% adjustment at the upper-end of the differences between the monitor height and
top of the canopy of low vegetation in the daytime.
       Staff recognizes that a 10% adjustment to hourly monitoring data across the country is a
very simple method to deal with a complicated issue.  The exchange of Os between the
atmosphere and vegetation is controlled by complex interactions of meteorological and
biological processes. Ideally one should account for the exact height of each monitor, canopy
roughness for each vegetation type and the seasonal and diurnal nature of turbulence. This was
not possible in our analyses.  To bound the uncertainty associated with applying a 10%
adjustment to all monitors and short canopies, staff performed a sensitivity analysis by also
calculating crop and tree seedling assessments without an adjustment.  Staff agrees with CASAC
comments that these calculations will provide a bracket of responses within which the reality
probably lies  for the true exposure of Os to short vegetation (Henderson, 2006c).  For brevity,
staff has presented the 10% adjusted figures in the main body of the Staff Paper and have placed
companion figures without the 10% adjustment in the Appendices 7-G and 7-H.  However, both
sets of results are discussed in this chapter.
                                          7-46

-------
       The inclusion of a 10% hourly adjustment had a substantial effect on the predicted 12-hr
SUM06 and W126 exposures. Reducing each hourly value by 10% over the entire interpolated
surface resulted in an average reduction of the 3-month 12-hr SUM06 by 53% and an average
reduction of 42% in the 3-month 12-hr W126. These large reductions in the SUM06 and W126
exposures are most likely a result of many monitored hourly concentrations occurring near the
SUM06 threshold and the inflection point for W126 (approximately 0.06 ppm). When these
"mid-level" hourly O3 values are reduced by 10%, many fall below 0.06 ppm, decreasing the
amount of hourly values counted in (SUM06) or contributing to (W126) these metrics.
       Given the somewhat lesser impact of the 10% adjustment on exposures using the W126
and the lack of evidence for a biological threshold for effects at 0.06 ppm, staff considered the
W126 index form more appropriate for conducting the crop yield and tree seedling biomass loss
risk assessment.  Other information that supports this decision includes: 1) studies that document
effects on crops and other sensitive vegetation at Os concentrations below 0.06 ppm [e.g.,
exposures as low as a 0.04 ppm 7-hr seasonal average (U.S. EPA, 2006)]; and 2) the high degree
of correlation between both forms when describing ambient exposures (see Figure 7-4) and their
similar predictive power of NCLAN crop data results in retrospective analyses (Lee et al., 1989;
U.S. EPA, 1996a, 2006).

       7.6.2.4   Quantifiable Risk of Yield Loss In Select Commodity, Fruit and Vegetable
                Crops
       The 2001 county-level crop planting data were  obtained for the 9 commodity crops (corn,
soybean, winter wheat,  sorghum, cotton, peanuts, kidney bean, potato & lettuce) from USDA-
NASS  (National Agricultural Statistics Service; http://www.usda.gov/nass). The appropriate
NCLAN C-R functions (available in the 12-hr W126 format) for each of the nine commodity
crops were identified from the analysis done for the 1996 Staff Paper (U.S. EPA 1996b, Table
7F-1).  The C-R functions for six fruit and vegetable species (tomatoes-processing, grapes,
onions, rice, cantaloupes, Valencia oranges) were identified from the 1996 California fruit and
vegetable analysis (Table 7F-2). Staff notes that fruit and vegetable studies were not part of the
NCLAN program and C-R functions were available only in terms of seasonal 7 hr or 12-hr mean
index.  This index form is considered less effective in predicting plant response for a given
change in air quality than the cumulative form used with other crops.  Therefore, staff considers
the fruit and vegetable C-R functions more uncertain than those for commodity crops.  Staff
combined the C-R functions with the crop planting information and with projections of 2001 Os
exposure based on a 12-hr W126 calculated for the 3 months prior to the harvest date for each
commodity crop and the appropriate growing season 7-hr or 12-hr average used for some fruits
and vegetables. Calendar periods used for computing W126, 7-hr and  12-hr exposure statistics,
                                          7-47

-------
are based on the harvest date and are done on a state-specific basis.  This allows for geographic
variation and better reflects actual Os exposure during the true growing period of the crop so that
calculated expected yield change for each crop, fruit and vegetable is specific to where they were
planted (Abt, 2007).
       Some of the results of this risk assessment are presented in Appendix 7F in Table 7F-4.
This table depicts the maximum county-level relative change in crop yield loss under air quality
scenarios of just meeting various alternative standard options under consideration using the
median C-R functions.  Maps of predicted yield loss for selected major crops are presented in
Appendix 7G.  Figure 7-15 shows a map of predicted yield loss for soybean from 2001 using the
10% adjusted "as is" estimated Os exposure scenario. Soybean is predicted to have the largest
yield loss in southwestern Pennsylvania, southern New Jersey  and east Texas. However, these
areas are not places of high soybean production.  In a high soybean producing state, such as
Illinois, yield loss was predicted to reach  a maximum range of 1-2% with a 10% adjusted Os
exposure (Figure 7-15)  and 3-4% without a 10%  adjusted 63 exposure (Figure 7G-1 in Appendix
7G). Corn, another major commodity crop, was not predicted  to have any loss in 2001. This is
because the two corn cultivars studied in NCLAN were not sensitive to Os. In contrast, cotton, a
more sensitive crop, had predicted yield loss above 10% in southern California (see Appendix
7G, Figures 7G-3 & 4).
                                          7-48

-------
Figure 7-15.  Estimated soybean yield loss based on interpolated 2001 3-month 12-hr W126 with a 10% downward adjustment
           of hourly Os concentrations.
                                                                       i
                                                       -  •  •    -   • : - -,,- V- -
Yield (% loss)
I    l<=1 (min = 0.1)
I    11 < Yield <=2
I    | >2 (max = 3.4)
   _ No production value reported
                                                           7-49

-------
       7.6.2.5   Economic Benefits Assessment - AGSIM
       This section presents results of the quantitative economic benefits analysis associated
with just meeting alternate standards. Adequate data are currently available to assess economic
benefits for 9 of the commodity crops studied in the NCLAN project and 6 fruit and vegetable
species.  Fruits and vegetables were evaluated in the 1996 review using a separate regional
benefits model separate from the national commodity crop model (U.S. EPA 1996b). This was
due to the fact that only regional planting data were available at the time for those fruits and
vegetables.  In the current benefits assessment, both commodity crops and fruits and vegetables
were evaluated together in the same national scale model.  Fruit and vegetables are a large part of
the U.S. agricultural sector and may be especially susceptible to Os pollution because much of
the production is located in the San Joaquin Valley region of California, which has very high
levels of 63 exposure (CEP A, 2005).  Because 6 of fruits and vegetables were not a part of the
NCLAN program and the uncertainties inherent in those experiments are less well known,
information on fruits and vegetables is presented separately in this document. Nonetheless, fruits
and vegetables are large portion of the U.S. agricultural economy. For example, in 2004, cash
income from California fruit and nut production was worth more than 9.6 billion dollars and over
7.2 billion dollars for vegetable crops (California Agricultural Resource Directory, 2005,
http://www.cdfa.ca.gov/).
       The Agriculture Simulation Model (AGSIM) (Taylor, 1994; Taylor, 1993) has been
utilized recently in many major policy evaluations.2 AGSEVI is an econometric-simulation
model used to calculate agricultural benefits of changes in Os exposure and is based on a large
set of statistically estimated demand and supply equations for agricultural commodities produced
in the U.S. A number of updates to AGSIM were performed before running this analysis: (1)  an
update of the commodity data for 2001, (2) incorporation of the most recent version of the
official USDA baseline model, (3) an econometric component added to AGSEVI to compute total
farm program payments for different levels of farm program parameters, and (4) a farm payment
program  component was added to the economic surplus module.  The AGSIM model was run to
provide benefit estimates for nine major commodity crops  (soybeans, corn, winter wheat, cotton,
peanuts,  sorghum, potato, lettuce, kidney bean) and six fruits and vegetables mainly grown in
California (tomatoes-processing, grapes, onions, rice, cantaloupes, Valencia oranges). As
       2 For example, AGSIM© has been used in EPA's prospective study of the benefits derived 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.

                                          7-50

-------
described earlier, hourly 63 exposures were adjusted downward by 10% before calculating the
W126, 7-hr or 12-hr exposure metrics.
       Percent relative yield losses (PRYL) calculated as the change in yield occurring between
just meeting 'as is' air quality and various alternative standard scenarios were the relevant input
parameters to the AGSEVI model. The AGSIM model predicted acreage, production, supply and
price parameters for each crop for each year, as well as yield-per-harvested acre, based on
calculated new yield-per-planted acre values, as well as on lagged price data, ending stocks from
the previous year and other variables. From these results and demand relationships embedded in
the model, AGSIM calculated the utilization of each crop (i.e., exports, feed use, other domestic
use,  etc), as well as change in consumer surplus, net crop income, deficiency payments and other
government support payments. Total undiscounted economic surplus was  calculated as the crop
income plus consumer surplus. For more detail on the AGSIM model see Appendix I of the
Environmental TSD (Abt, 2007). The AGSIM model was run for 14 years for each scenario in
order for the model parameters to adjust to the initial change in yield. Annual changes in total
undiscounted economic surplus were calculated for each of the 14 years. The annual averages
for the 14 years are reported in Tables 7-3 A-B.
       The results from applying the AGSIM model to determine commodity crop and fruit and
vegetable benefits based on meeting the level  of the current 8-hr average standard and five
alternative standards are presented in Tables 7-3A-B. Note that Table 7-3A presents results with
the 10% downward hourly adjustment and Table 7-3B presents results without the adjustment.
In summary, this analysis estimated a range of benefits using both the available minimum and
maximum yield loss equations for each crop.  Results are presented in annual 2000 dollars for
the commodity crops, fruits and  vegetables and total agricultural sector.  Overall, benefits from
the fruit and vegetable species in this analysis accounted for a relatively large portion of the total
agricultural benefits compared with the commodity crops. This is likely because many of the
fruits and vegetables are grown in parts of California with high O3 exposures and any rolling
back of air quality produced greater changes in Os levels, resulting in higher changes in yield.
All of the alternative standards analyzed showed positive incremental benefits greater than those
associated with just meeting the  level of the current 8-hr average standard.  Including a 10%
downward adjustment the hourly monitoring did not have a large effect on the overall benefits
calculated for each standard. Not surprisingly, not adjusting the hourly monitoring data
downward resulted in slightly higher benefits.  Meeting the SUM06 of 25 ppm-hr proposed in
the last review and the approximate equivalent W126 of 21 ppm-hr produced benefits of
approximately $140-$260 million for the total agricultural sector.  Of all the scenarios, W126 of
13 ppm-hr, SUM06 of 15 ppm-hr and 8-hr average of 0.07 ppm had the largest economic benefit.
Meeting the alternative W126 of 13 ppm-hr and approximate equivalent of SUM06 of 15 ppm-hr
                                          7-51

-------
produced benefits of approximately $290-$630 million for the total agricultural sector. It is
important to note that these results represent a macro-analysis of the U.S. agricultural economy.
Farmers in areas that have higher Os levels are more adversely affected than farmers that are in
areas with low 63 levels.  These important effects are difficult to quantify in a macro-analysis.
       The current CD reports very few new studies conducted on the economic effect of Os on
U.S. agriculture (U.S. EPA, 2006).  A study by Murphy et al. (1999) confirmed the general
magnitude of economic effects reported by two key studies performed a decade earlier (Adams,
1986; Adams et al.,  1985). Specifically, Murphy et al. (1999) evaluated benefits to eight major
crops associated with several scenarios concerning the reduction or elimination of 63 precursor
emissions from motor vehicles in the U.S. Their analysis reported a $2.8 to 5.8 billion (1990
dollars) benefit from complete elimination of Os exposures from all sources, i.e., ambient Os
reduced to a background level assumed to be 0.025 to 0.027 ppm. In comparison,  AGSIM
calculates $300 million to 2.5 billion (2000 dollars) in economic benefit for 9 major commodity
crops when 63 levels are reduced to near background.  These AGSIM results are without any
downward adjustment to the O3 monitoring data and without subtracting out farm payments.
With a 10% adjustment and subtracting farm payments, AGSIM calculates substantially lower
benefits ($200-800 million) for the same 9 major commodity crops.  The Murphy et al. (1999)
analysis and the current AGSIM analysis are quite difficult to compare for many reasons:
different economic models, different air quality years, different treatment of government farm
payment programs, dollar value unadjusted for inflation, different assumptions, etc. However,
these comparisons point out that initial assumptions about Os exposure and crop payments have
large implications when calculating agricultural benefits for reducing O3 to background levels.

       7.6.2.6   Uncertainties In the Crop Risk and Benefit Analyses
       The crop risk assessment utilized the C-R relationships developed in OTC experiments
performed between  1980 and!988 in the NCLAN program and in other experiments done on
fruits and vegetables. As discussed earlier, fruit and vegetable studies were not part of the
NCLAN program and C-R functions were available only in terms of a seasonal 7-hr or 12-hr
mean index. This mean index form is considered less effective in predicting plant response for a
given change in air quality than the cumulative form used with other crops.  Two of the
uncertainties using the OTC C-R functions in the crop risk assessment are chamber effects (see
section 7.6.2.1) and sensitivity of current crops (see section 7.6.2.2).  Staff qualitatively
addressed these uncertainties citing studies with recent cultivars and studies not using chambers.
However, it was not possible to perform a quantitative assessment of these uncertainties.
Therefore, despite support in the scientific literature for the magnitude of yield effects of 63
exposure on crops from OTCs, staff cannot quantify how these uncertainties would affect
                                          7-52

-------
estimated crop risks and benefits.  Staff notes that the predicted yield losses calculated from the
OTC studies represent losses for crops that were not being affected by other stressors.  Stressors
such as drought may decrease the yield response to Os exposure while insect or disease damage
to crops may be exacerbated by 63 exposure.
       An additional source of uncertainty not described or accounted for in the last review is
that associated with the presence of a decreasing 63 gradient from the height of the monitor
probe down to the canopy heights for most crops.  The presence of this gradient makes less
certain the predictions of current crop exposures and the associated yield losses based on ambient
monitor data. Staff selected a 10% reduction factor to represent the maximum gradient believed
to occur for daylight hours. However, recognizing that the actual downward adjustment value
varies depending on interactions between numerous plant and site-specific factors, staff chose to
present estimates of yield loss for each crop as a range, with non-adjusted and 10%-adjusted air
quality as the upper and lower bounds (see section 7.6.2.3 for a detailed discussion).
       It is important to restate the uncertainties associated with the results of the AGSIM
economic analysis presented in section 7.6.2.5. Uncertainties are introduced by: (1) the
interpolation of air quality monitoring data to estimate 2001 national  Os exposures; (2) the use of
C-R functions from OTC studies to estimate relative yield losses from 2001 exposures; (3) the
use of a quadratic rollback method to project the "just meet" air quality scenarios without a direct
link to an emissions control strategy; and (4) the inherent uncertainties associated with use of an
economic model such as AGSIM.  It is also important to note that the range of results from this
analysis represents impacts associated only with available NCLAN experimental  data and a
limited number of fruits and vegetable studies. Not all crops have been subjected to  exposure-
response experiments and effects on those crops would be missed.  Despite the amount of
uncertainty, staff concludes that this analysis provides useful insights for comparing  the relative
benefits obtained as a result of meeting alternative regulatory scenarios.
                                           7-53

-------
Table 7-3 A-B. Agricultural model results with (A) and without (B) a 10% adjustment of
          hourly Os exposures.
A.
Average Annual Changes in Total Undiscounted Economic Surplus for the
Current 8hr Standard (0.08) and Alternative Standards (millions $; 2000)
Standard
0.08 4th-highest
0.07 4th-highest
W126 = 21
W126=13
SUM06 = 25
SUM06= 15
Commodity Crops
$10-20
$50 - 200
$10-40
$30-140
$10-50
$60 - 200
Fruits & Vegetables
$60 - 80
$310-360
$130-140
$260 - 300
$160-180
$290 - 330
Total Ag.
$70-100
$360 - 560
$140-180
$290 - 440
$170-230
$350 - 530
B.
Average Annual Changes in Total Undiscounted Economic Surplus for the
Current 8hr Standard (0.08) and Alternative Standards (millions $; 2000)
Standard
0.08 4th-highest
0.07 4th-highest
W126 = 21
W126=13
SUM06 = 25
SUM06= 15
Commodity Crops
$10-30
$70 - 280
$20 - 40
$60 - 1 90
$20 - 60
$70 - 260
Fruits & Vegetables
$70 - 80
$350-410
$140-160
$280 - 340
$180-200
$320 - 370
Total Ag.
$80- 110
$420 - 690
$160-200
$340 - 530
$200 - 260
$390 - 630
                                       7-54

-------
       7.6.3   Tree Risk Assessments
       In the last review (U.S. EPA, 1996b), analyses of the effects of O3 on trees were limited
to 11 tree species for which C-R functions for the seedling growth stage had been developed
from OTC studies conducted by the National Health and Environmental Effects Research Lab,
Western Ecology Division (NHEERL-WED). Figure 7-16A uses the regression equations for
each of the 26 studies (49 cases) to graph predicted relative yield loss at various exposure levels
in terms of a 12-hr W126.  Figures 7-16B-C show composite graphs for the intensively studied
quaking aspen and ponderosa pine.  Work done since the 1996 review at the AspenFACE site in
Wisconsin (see section 7.6.2.1) on quaking aspen has confirmed the detrimental effects of 63
exposure on tree growth in a field study without chambers (Karnosky et al., 2005). Since the
1996 review, only a few new studies have developed C-R functions for additional tree seedling
species (U.S. EPA, 2006).  One such study of eastern cottonwood (Populus deltoides) saplings
was done without chambers or Os FACE-type fumigation (Gregg et al., 2003). Eastern
cottonwood is a fast growing tree that is important ecologically along streams and commercially
for pulpwood, manufacturing furniture  and a possible source for energy biomass (Burns and
Hankola, 1990).  Gregg et al. (2003) found that the cottonwood saplings grown in urban New
York City grew faster than saplings grown in more rural areas where Os was higher.  The
secondary nature of the reactions of O3 formation and NOx titration reactions within the city
center resulted in significantly higher cumulative Os exposures in the rural sites. After carefully
considering many factors, they concluded the major explanation for the difference in growth was
the gradient of O3 exposure between urban and rural sites. This explanation was also confirmed
with an OTC study (Gregg et al., 2003). Figure 7-17 shows the biomass growth of cottonwood
plotted against the monitored 12-hr W126 at the sites the trees were planted (Gregg et al., 2003).
Staff notes that the responses of natural populations of cottonwood to Os may vary because of
precipitation patterns and differences in native soils. The Gregg et al. (2003) study is important
because it demonstrated that growth effects of Os exposure could be documented in field without
chambers or fumigation and the growth decreases were as great as to seen in previous OTC
studies. The evidence from the AspenFACE results and Gregg et al. (2003) provide support for
the continued use of NHEER-WED OTC studies to estimate risk to seedlings in the U.S.  Section
7.6.3.1 describes how staff updated the tree seedling risk analysis performed in the last review.
       Section 7.6.3.2 discusses the approach for assessing Os effects on vegetation in natural
settings using visible foliar injury data.  Section 7.6.3.3 discusses the analysis and results for
modeling Os impacts on mature trees in the Eastern and Western U.S.  The tree and/or forest
analyses  outlined below will enable staff to begin to assess important long-term effects of
various secondary standard levels on forest ecosystem health and services.
                                          7-55

-------
Figure 7-16.  Median tree seedling biomass loss for all 49 cases (A), quaking aspen
            (B), and ponderosa pine (C) characterized with the 12-hr W126
          „  90 -
                            20     30     40     50

                               12-hrW126(ppm-hr)
                                                         75th Percentile
                                                         50th Percentile

                                                         25th Percentile
100 -
_ 90 :
£^ :
w 80 -
(/)
ii ™-
w
E 60 •
o
CO „

-------
Figure 7-17. Cottonwood (Populus deltoides) shoot biomass (mean ± s.e.) at urban (filled)
          and rural (open) sites in the vicinity of New York City versus ambient Os
          exposure (growing period 12-hr W126, July 7 - Sept. 20). Squares, circles and
          triangles represent responses in 1992,1993 and 1994, respectively.  Cottonwood
          saplings were grown in potting soil under well watered conditions.  (Modified
          from Gregg et al., 2003)
    90

    80

    70

~  60
3
?  50
 re
 o  40
m
    30

    20

    10
                y = -6.1902x + 87.149  R2 = 0.8622
      0.0     2.0     4.0    6.0     8.0
                        Ozone Exposure
                        (W126, ppm-hrs)
10.0   12.0    14.0
                                       7-57

-------
       7.6.3.1   Quantifiable Risk of Biomass Loss In Select Tree Seedling Species
       In a process similar to that used for crops above (7.6.2.4), C-R functions for biomass loss
for a subset of seedling tree species taken from the CD (Table 7F-3) and information on tree
growing regions derived from the U.S. Department of Agriculture's Atlas of United States Trees
(Little, 1971) were combined with  projections of air quality based on 2001 POES, to produce
estimated biomass loss for each of the seedling tree species individually. Some of the results for
the highest areas of risk to tree seedlings are presented in Table 7F-5 in Appendix 7F.  In
addition, maps depicting these results for selected tree seedling species are found in Appendix
7H.
       Figure 7-18 shows an example of the quaking aspen seedling biomass loss with an hourly
Os exposures adjusted down by 10%. Figure H-l in Appendix H shows the quaking aspen
without the 10% hourly adjustment. The quaking aspen maps show significant variability in
projected seedling biomass loss across its range for 2001.  Quaking aspen seedling biomass loss
(with thelO% adjustment) was projected to be greater than 4% over much of its geographic
range, though it can reach as high as 12% in some areas.  In Appendix 7H, there are additional
maps of ponderosa pine and black cherry along with maps of seedling biomass loss with and
without a 10% adjustment of the monitoring data. Further, in Chapter 5 of the Environmental
TSD, a series of maps are presented showing seedling biomass gain when various standard levels
are met. These biomass gain maps indicate that substantial improvements in seedling growth
may be achieved when the alternative standards are met, especially the 0.07 ppm 4th-highest
max., SUM06 of 15 ppm-hr and W126 of 13 ppm-hr.  It should be noted that the species mapped
are generally sensitive and they are also important tree species in ecosystems across vast areas of
the U.S.  Though each map shows the geographical range for a species,  it does not indicate that
an individual of that species will be found at every point within its range. It should also be
recognized that the production of these maps incorporates several separate sources of
uncertainty, beginning with the C-R functions produced for seedlings in OTCs to the
uncertainties associated with the inputs used to generate the POES. Furthermore, percent
biomass loss in tree seedlings is not intended to be a surrogate for expected biomass loss in
mature trees of the same species (see section 7.6.3.3 for modeling of mature tree growth).
Studies indicate that mature trees can be more or less sensitive than seedlings depending on the
species. Further, seedling biomass loss cannot be considered comparable to percent yield loss in
agricultural crops. This is because a small biomass loss per year in a perennial tree species, if
compounded over multiple years of exposure could have a large effect on the growth of that tree,
while yield loss in annual  crops is only affected by the O3 exposure for that year.  In summary,
                                          7-58

-------
Figure 7-18.  Estimated aspen seedling annual biomass loss based on interpolated 2001 maximum 3-month 12-hr W126 with a
           10% downward adjustment of hourly Os concentrations. This map indicates the geographic range for quaking
           aspen (Populus tremoloides), but it does not necessarily indicate that aspen will be found at every point within its
           range.
  Biomass (% loss)
  I   |<1 (min = 0.1)
  '|   11 < Biomass ==2
  I   | 2 < Biomass <=4
  I   I 4 < Biomass <=6
  HI 6 < Biomass <=10
  JHT>1° (max = 12)
     No production value reported
                                                          7-59

-------
 this analysis indicates that current air quality can produce significant seedling biomass loss in
the areas which those trees grow. Meeting the level of alternative standards is expected to
improve biomass growth in the seedlings analyzed.

       7.6.3.2   Visible Foliar Injury Incidence
       The use of sensitive plants as biological indicators to detect phytotoxic levels of 63 is a
longstanding and effective methodology (Chappelka and  Samuelson,  1998; Manning and Krupa,
1992). Some well defined bioindicators for ambient 63 include blackberry, black cherry, green
ash, milkweed, quaking aspen, sassafras, yellow poplar, and white ash. Each of these
bioindicators exhibits typical Os injury symptoms when exposed under appropriate conditions.
These symptoms are considered diagnostic as they have been verified in exposure-response
studies under experimental conditions.  Typical visible injury symptoms on broad leaved plants
include: 1) acute exposure (pigmented lesions (stippling), flecking, surface bleaching, and/or
bifacial necrosis); 2) chronic exposure (pigmentation (bronzing), chlorosis or premature
senescence). Typical visible injury symptoms for conifers include: 1) chlorotic banding or
tipburn (acute  exposure); 2) flecking or chlorotic mottling, premature senescence of needles
(chronic exposure).  Though common patterns of injury develop within a species, these foliar
lesions can vary considerably between and within taxonomic groups.  Furthermore, the degree
and extent of visible foliar injury development varies from year to year and site to site, even
among co-members of a population exposed to similar Os levels, due  to the influence of co-
occurring environmental and genetic factors.  It is important to note that the visible foliar injury
occurs only when sensitive plants are exposed to elevated 63 concentrations in a predisposing
environment.  Thus, great care must be taken when assessing the response of bioindicators to
ambient O3 (Flagler, 1998).
       The Unites States Forest Service (USFS) through the Forest Health Monitoring Program
(FHM) (1990 - 2001) and currently the Forest Inventory and Analysis (FIA) Program has been
collecting data regarding the incidence and severity  of visible foliar injury on a variety of 63
sensitive plant species throughout the U.S. (Coulston et al.,  2003, 2004; Smith et al., 2003). The
plots where these data are taken are known as biosites. These biosites are located throughout the
country and analysis of visible foliar injury within these sites follows  a set of established
protocols (for more details see http://fiaozone.net/).  Since the conclusion of the  1996 NAAQS
review, the FIA monitoring program network and database has continued to expand.  The visible
foliar injury indicator has been identified as a means to track 63 exposure stress trends in the
nation's natural plant communities as highlighted in EPA's  most recent Report on the
Environment (U.S. EPA, 2003a; http://www.epa.gov/indicators/roe).  EPA staff also considers it
important to assess the  degree to which  O3-induced  visible foliar injury observed in situ,
                                          7-60

-------
corresponds with monitored Os air quality in recent years.  In a collaborative effort with FIA
staff, EPA staff conducted an analysis to compare the incidence of visible foliar injury at
different levels of air quality (e.g., the current standard and alternative levels under
consideration) by county throughout the U.S.  This analysis potentially provides a measure of the
effectiveness and degree of protection provided by the current form/level of the secondary
standard for this welfare effect.
       The major confounding effect for O3 induced visible foliar injury is the amount of soil
moisture (local rainfall) available to a plant during the year that the visible foliar injury is being
assessed.  This is because lack of soil moisture decreases stomatal conductance of plants and,
therefore, limits  the amount of Os entering the leaf that can cause injury. Many researchers have
shown that dry periods in local areas tend to decrease the incidence and severity of visible foliar
injury caused by Os in plants measured by the USFS (Smith et al., 2002).  Therefore, the
incidence of visible foliar injury is not always higher in years with higher Os, especially when
there is drought in areas where visible foliar injury is assessed.
       Due  to a  congressional requirement that the USFS protect landowner privacy, FIA cannot
publicize the exact locations of the biosites. As a result, all data in this analysis are reported on a
county-level.  County-level visible foliar injury data were available for the years 2001 to 2004
for all areas  of the U.S. except the Mountain West region.  However, according to the FIA staff,
no Os injury was reported at any site in that region.  Figure 7-19, shows that the incidence of
visible foliar injury  in 2001 was widespread across the eastern and western U.S. The 2001 data
are indicative of the incidence of visible foliar injury in the years 2001 to 2004. (see Appendix
71 for 2002). This indicates that Os levels are above phytotoxic levels sufficient to cause adverse
effects in natural plant populations in many areas. It is important to note that direct links
between Os  induced visible foliar injury symptoms and  other adverse effects (e.g., biomass loss),
are not always found.  However, in some cases, visible foliar symptoms have been correlated
with decreased vegetative growth (Karnosky et al., 1996; Peterson et al., 1987; Somers et al.,
1998) and with impaired reproductive function (Black et al., 2000; Chappelka, 2002).  Though
visible injury is a valuable indicator of the presence  of phytotoxic concentrations of Os in
ambient air it is not always a reliable indicator of damage or other injury endpoints. The lack of
visible injury does not indicate a lack of phytotoxic concentrations of Os or a lack of non-visible
Os effects.
       In an attempt to assess how meeting various  Os standard levels affected the incidence of
visible foliar injury, staff matched up county-level Os monitoring data with counties that had US
Forest Service biosites.  In counties containing multiple O3 monitors,  staff used the monitor
measuring the highest Os to characterize county air quality. Because visible foliar injury
symptoms reflect the Os stress of the year in which they are observed, staff looked at yearly
                                           7-61

-------
Figure 7-19. 2001 County-level incidence of visible foliar injury in the eastern and western U.S. as measured by the US Forest
           Service FIA program.

                            Is Foliar Injury Present or Absent?, 2001
                       Foliar Injury
Absent
Present
                                                         7-62

-------
snapshots of county-level air quality data.  Between 235 and 286 counties had EPA 63
monitoring and at least one USFS FIA biosite surveyed for visible foliar injury in the years 2001
- 2004 (see Table 7-4). However, because the specific locations of the USFS biosite are not
publicly available, staff was unable  to determine how close the biosites within each county were
to the Os monitor selected to represent that county.  Air quality was evaluated in terms of both
the current 8-hr, average and 12-hr SUM06 forms, using a number  of different levels.  Table 7-4
shows the percentage and number of counties with and without visible foliar injury at or below
various standard levels for the 2001-2004 period. Because the FIA program reorganized the
locations of biosites in 2002 and expanded the number of biosites in 2003 and 2004, the total
number of counties containing both an Os monitor and an FIA biosite changed each year and it is
difficult to interpret changes in the number of counties in different  categories between years.
Therefore, staff found it more informative to present results in terms of percent of total counties
with or without injury under different levels of air quality. First, this table illustrates that visible
foliar injury is occurring in areas that are meeting the current 8-hr average 63 standard (0.084
ppm).  Second, the table illustrates that the secondary standard option of a SUM06 of 25 ppm-hr
proposed in 1996  did not appear to offer more protection from visible foliar injury than the
current 8-hr average standard form.  By comparison, the SUM06 of 15 ppm-hr and the 8-hr
average of 0.074 ppm provided more protection across all years than either the 0.084 ppm  8-hr
average or SUM06 of 25 ppm-hr standards. At the 0.084 ppm, 8-hr average, the percent of
counties showing injury ranged from 21% to 39%. Under a SUM06 of 25 ppm-hr, the percent of
counties with injury was 26% to 49%. For the two lower air quality alternatives (0.074 ppm 8-hr
average and SUM06 of 15 ppm-hr), values ranged from 12% injured to 30% and 35%,
respectively.
       In summary, this analysis indicates that incidence of Os induced visible foliar injury is
widespread across the eastern and western U.S. Visible foliar injury was observed in counties
that are meeting the current level of the 8-hr standard and an alternative secondary standard
option of a SUM06 of 25  ppm-hr proposed in 1996.  Lower standards in the 8-hr average and
SUM06 forms would be expected to have lower incidences of visible foliar injury. However, the
level of protection would  depend heavily on local environmental variable such as soil  moisture.
Finally, in the consensus workshop  held on the secondary O?, standard, researchers were in
agreement that a 3 month 12-hr SUM06 value of 8 to 12 ppm-hr should be considered for
protection from visible foliar injury to natural ecosystems (Heck and Cowling, 1997).  The
analysis above supports this recommendation that these levels would reduce the incidence  of
visible foliar injury to natural ecosystems.
                                          7-63

-------
Table 7-4. Percentage and number of counties with visible foliar injury (injured) and without injury (not injured) below
          various standard levels for the years 2001-2004. Each county had an Os monitor and a USDA forest service FIA
          plot tracking visible foliar injury due to O3 exposure.
Year
2001
2002
2003
2004

# of counties
injured
not injured
# of counties
injured
not injured
# of counties
injured
not injured
# of counties
injured
not injured
£0.084*
(ppm)
99
39% (39)
61% (60)
89
21% (19)
79% (70)
185
28% (52)
72% (133)
260
35% (91)
65% (169)
£0.074*
(ppm)
36
25% (9)
75% (27)
43
12% (5)
88% (38)
61
11% (7)
89% (54)
159
30% (47)
70% (11 2)

-------
       7.6.3.3   Modeled Mature Tree Growth Response: Eastern and Western Species
                Case Studies
       In the 1996 Os Staff Paper, evaluations of Os impacts on tree growth were limited to the
seedling growth stage (U.S. EPA, 1996b).  At that time, robust C-R functions were available
only for 11 tree seedlings developed from OTC data. Few studies had been done comparing
seedling sensitivity to that of a mature tree of the same species.  Recent experiments using the
FACE methodology have been able to expose 3 tree species to O3 beyond the seedling growth
stage. However, this methodology has not yielded C-R functions at this time, due to the limited
number of exposure regimes used. Findings from FACE publications, however, do show
decreased biomass growth under elevated Os in trees beyond the seedling stage (King et al.,
2005). In order to better characterize the potential Os effects on mature tree growth, staff used a
tree growth model (TREGRO) as a tool to evaluate the effect of changing 63 air quality
scenarios from just meeting alternative Os standards on the growth of mature trees.
       TREGRO is a process-based, individual tree growth simulation model (Weinstein et al,
1991) and has been used to evaluate the effects of a variety of O3 scenarios and linked with
concurrent climate data to account for Os and climate/meteorology interactions on several
species of trees in different regions of the U.S. (Tingey et al., 2001; Weinstein et al., 1991;
Retzlaff et al., 2000; Laurence et al., 1993; Laurence et al., 2001; Weinstein et al., 2005). The
model provides an analytical framework that accounts for the nonlinear relationship between Os
exposure and response.  The interactions between 63 exposure, precipitation and temperature are
integrated as they affect vegetation, thus providing an internal consistency for comparing effects
in trees under different exposure scenarios and climatic conditions (see the Environmental
Assessment TSD for more details on TREGRO).  An earlier  assessment of the effectiveness of
national ambient air quality standards in place since the early 1970s took advantage of 40 years
of air quality and climate data for the Crestline site in the San Bernardino Mountains of
California to simulate Ponderosa pine growth over time with the improving air quality using
TREGRO (Tingey et al., 2004).
       Staff collaborated with the EPA NHEERL-WED laboratory to use the TREGRO model
to assess growth of Ponderosa pine (Pinusponderosd) in the San Bernardino Mountains of
California (Crestline) and the growth of yellow poplar (Liriodendron tulipiferd) and red maple
{Acer rubrum) in the Appalachian mountains of Virginia and North Carolina, Shenandoah
National Park (Big Meadows) and Linville Gorge Wilderness Area (Cranberry), respectively.
Total tree growth associated with 'as is' air quality, and air quality adjusted to just meet
alternative O3 standards was assessed (Table 7-5).  Ponderosa pine is one of the most widely
distributed pines in western North America, a major source of timber, important as wildlife
                                          7-65

-------
Table 7-5. Relative increase in total annual tree biomass growth, simulated with the TREGRO model, if the level of the
          current (0.08 ppm) and alternative standards are met.
Species
Site
0.08 4th-highest
0.07 lst-highest
0.07 4th-highest
SUM06 = 253
SUM06= 154
red maple
Big Meadows, VA
(1993-1995)
0.41%
2.71%
2.24%
0.34%
4.49%
red maple
Cranberry, NC
(1993-1995)
no rollback
2.31%
1.38%
no rollback1
2.99%
yellow poplar
Big Meadows, VA
(1993-1995)
0.03%
0.38%
0.34%
0.07%
0.60%
yellow poplar
Cranberry, NC
(1993-1995)
no rollback
6.54%
3.91%
no rollback1
8.26%
ponderosa pine
Crestline, CA
(1995-2000)
8.63%
10.81%
n.a.2
10.33%
n.a.2
1A rollback was not necessary for the Cranberry site for the 0.08 ppm 4th-highest and SUM06 = 25 ppm-hr scenarios since air quality
was at or below the levels of those scenarios.
2 TREGRO was not run for ponderosa pine for the 0.07 ppm 4th-highest scenario.
3The roll-back to a SUM06 of 25 ppm-hr was a W126 of approximately 18 ppm-hr at Cranberry, Big Meadows and Crestline.
4The roll-back to a SUM06 of 15 ppm-hr was a W126 of approximately 13 ppm-hr at Cranberry and Big Meadows
                                                        7-66

-------
habitat, and valued for aesthetics (Burns and Honkala, 1990). Red maple is one of the most
abundant species in the eastern U.S. and is important for its brilliant fall foliage and highly
desirable wildlife browse food (Burns and Honkala, 1990). Yellow poplar is an abundant species
in the southern Appalachian forest.  It is 10% of the cove hardwood stands in southern
Appalachians which are widely viewed as some of the country's most treasured forests because
the protected, rich, moist set of conditions permit trees to grow the largest in the eastern U.S.
The wood has high commercial value because of its versatility and as a substitute for
increasingly scarce softwoods in furniture and framing construction. Yellow poplar is also
valued as a honey tree, a source of wildlife food, and a shade tree for large areas (Burns and
Honkala, 1990).
      At the western site, staff and NHEERL-WED scientists used Crestline, CA air quality
and climate data from the years 1995 to 2000 to run TREGRO, while at the eastern sites, staff
used Big Meadows, VA and Cranberry, NC air quality and climate data from the years  1993 to
1995. These three years were the only years in the east with readily  available 63 and climate
data that could be used in TREGRO. The years chosen to run the TREGRO at each site appear
to have annual Os exposures typical of the last 15 years (Figure 7-20). Air quality from each site
and year was adjusted using the quadratic roll-back method to 'just meet' the current 8-hr
average secondary standard (0.084 ppm), a 12-hr SUM06  of 25  ppm-hr, and 1st highest max
average of 0.07 ppm. Staff also tested the 4th-highest 0.07 ppm  level on the Cranberry and Big
Meadows sites. For the ponderosa pine at Crestline, TREGRO was run for "as is" and "just
meet" air quality conditions in four 3 year increments to increase the accountability of climate
variability and the annual average biomass determined from these 4 simulations to yield an
annual average biomass change over the 6 years of Os exposure. For the yellow poplar and red
maple, two sites (Big Meadows, VA and Cranberry, NC) were chosen to run TREGRO to
increase the variability in climate since there were only 3 years of data available at each site.
The differences between growth under "just meet" and "as is" air quality conditions were
compared to evaluate the effectiveness of the current secondary standard and alternative
standards in  protecting these three tree species.
      Results of the TREGRO simulations are presented in Table 7-5. Clearly, the greatest
simulated growth benefits in the scenarios are seen in ponderosa pine at the Crestline site in
California. As shown in Figure 7-20, Os levels are much higher at Crestline than the sites in the
eastern US.  Meeting the level of the current standard was simulated to result in an 8.63%
increase annual growth and a SUM06 of 25 ppm-hr is expected  increase growth 10.33% in
ponderosa pine. In the eastern sites (Cranberry and Big Meadows), O3 levels are much lower
(Figure 7-20) and had less of an effect on the simulated growth of red maple and yellow poplar.
In fact, the Cranberry, NC site was below the level of the current 8-hr average standard and the
                                          7-67

-------
Figure 7-20.  Historical Os data as measured in the 3-month 12-hr SUM06 and 4th highest
           8-hr metrics for the 3 sites used to run the TREGRO model. For Big Meadows,
           VA and Cranberry, NC, climate and Os data from 1993 to 1995 was used to run
           TREGRO and for Crestline, CA, 1995 to 2000 was used.  Missing data points in
           the top panel indicate incomplete data to calculate a SUM06. * indicates which
           years of data were used in the TREGRO model at each site.
    120
  E
  Q.
  O.
 CD
 O
    100 -
    80 -
 Z)   60
 CO
 CN
     40 -
o
E
co   20
 E
 0.
    0.18 -
    0.16 -
 w  o.14 -

 00

 (0  0.12 H
 E
 O)
    0.10 -
    0.08 -
    0.06
                                           —•— Crestline, CA
                                           • •••<>•• Big Meadows, VA
                                           —^T— Cranberry, NC
             proposed 1996 SUM06 standard
                        O
                              ,0
                                           D
                                           ^-^
  D
V
                                                        o
                                           —•— Crestline, CA
                                           ••••(>••• Big Meadows, VA
                                           —-T— Cranberry, NC
       1988   1990   1992   1994   1996   1998   2000   2002   2004

                                  Year
                                          7-68

-------
SUM06 of 25 ppm-hr scenarios and, therefore, no benefits were calculated for those levels.  At
Big Meadows, VA, the current 8-hr average standard and SUM06 scenarios resulted in relatively
small growth increases for yellow poplar (0.03-0.07%) and red maple (0.34-0.41%). This was
mostly because the Big Meadows site was close to meeting those levels in 1993-1995 (Figure 7-
20).  Red maple was simulated to have a similar response (-2%) to the 0.07 ppm 1st and 4th-
highest 8-hr max in Big Meadows and Cranberry.  For the same scenarios, yellow poplar had a
very different response to O3 reduction at Big Meadows (0.34-0.38%) compared to Cranberry
(3.91-6.54%).  The climate at Cranberry is much more ideal for yellow poplar than under the
cool temperatures of Big Meadows, making it much more likely that its growth would be
suppressed by Os and that, conversely, it would respond much more to Os reductions.  Red
maple has a much larger geographical distribution, so that the temperature differences between
Big Meadows and Cranberry are less likely to affect the growth response. This phenomenon was
reflected in the simulations.
       The effect of 63 on an individual tree may be quite different than the predicted effect on a
forest stand after many years. Some researchers have used the ZELIG model, a forest stand
simulator, to predict stand growth using growth rates of individual species from TREGRO
scenarios (Laurence et al., 2001; Weinstein et al., 2005).  Small changes in growth of an
individual tree over a short period of time have sometimes been simulated to have large changes
in basal area as it develops over a long time period. For example, Weinstein et al. (2005) found a
simulated  63 effect on an individual ponderosa pine at Crestline to reduce growth by 6.7% in
three years under normal precipitation, yet stand basal area was calculated to be reduced by 29%
after 100 years. Similarly, Laurence et al. (2001) found individual yellow poplar in NC with an
63 induced growth loss of 1.7% which was then calculated to reduce basal area by 14% after 100
years.  This suggests that small effects on individual tree growth may result in substantial effects
on forest stand growth after many years.

       7.6.3.4   Uncertainties In the Tree Risk Analyses
       It should be recognized the seedling risk assessment incorporates several sources of
uncertainty that have been previously discussed.  Specifically, uncertainties associated with the
development C-R functions using OTCs and uncertainties associated with the inputs used to
generate the POES (see sections 7.6.2.1 and 7.5.5). As with crops, the potential differences
between exposures measured above seedling canopies and actual exposure at the top of the
canopy is an important uncertainty. As explained in section 7.6.2.3, it is impossible to fully
account for these potential differences throughout the U. S. Therefore, staff calculated risks using
a 10%  adjustment of hourly exposures and no adjustment of hourly exposures. These
                                          7-69

-------
calculations provide a bracket of responses within which the reality probably lies for the actual
Os exposures tree seedling canopies.
       The visible foliar injury risk assessment contains several sources of uncertainty. First,
due to the major confounding effect of soil moisture (local rainfall) in determining the level of
observed symptom expression, the incidence and degree of visible foliar injury is not always
higher in years with higher Os, especially when there is drought in areas where foliar injury is
assessed. Second, the lack of visible injury does not indicate a lack of phytotoxic concentrations
of Os or a lack of non-visible Os-induced effects, since it is not  always a reliable indicator of
other Os-related injury and damage endpoints.  Finally,  due to the change in FIA protocols in
2002 and the unavailability of specific biosite locations, staff was unable to determine the degree
to which county level monitored Os values reflect the actual Os exposure conditions at the
biosites within those counties.
       As with every model, TREGRO has many known and unknown sources of uncertainty.
Because TREGRO only models individual trees, the effects of competition are not factored in.
There is genetic variability within species so that the values produced for an individual tree may
not reflect the variability within the species as a whole.  Only a  few species have been
parameterized in TREGRO. Due to the limited number of species tested and included in this
assessment, it is unclear to what degree these results apply to Os impacts on mature trees in
general. For further discussion of uncertainties see Appendix J  in the Environmental Assessment
TSD (Abt, 2007).

7.7    QUALITATIVE RISK: ECOSYSTEM CONDITION, FUNCTION AND
       SERVICES
       Ecosystems are  comprised of complex assemblages of organisms that provide distinct
ecological attributes, many of which may be adversely affected by Os (U.S. EPA, 2006).  A new
effort has been initiated within the Agency to identify indicators of ecological condition whose
responses can be clearly linked to changes in air quality and be  used to track improvements in
environmental protection attributable to environmental program actions/implementation.
Moreover, a recent critique of the secondary NAAQS review process published in the report by
the National Academy of Sciences on Air Quality Management in the United States (NRC, 2004)
stated that "EPA's current practice for setting secondary standards for most criteria pollutants
does not appear to be sufficiently protective of sensitive crops and ecosystems..."  This report
made several specific recommendations for improving the secondary NAAQS process and
concluded that "There is growing evidence that tighter standards to protect sensitive ecosystems
in the United States  are needed..."  However, the vast majority of information regarding the
effects of Os involves the sensitivity of individual species. Therefore, this section lays out some
                                          7-70

-------
examples of our current understanding of how 63 may be affecting ecosystems and identifies
areas of research needed to address this issue.
       An ecosystem is defined as comprising all of the organisms in a given area interacting
with the physical environment, so that a flow of energy leads to a clearly defined trophic
structure, biotic diversity, and cycling of materials between living and nonliving parts (Odum,
1963). Individuals within a species and populations of species are the building blocks from
which communities and ecosystems are constructed.  Classes of natural ecosystems, e.g., tundra,
wetland, deciduous forest, and conifer forest, are distinguished by their dominant vegetation
forms. Ecosystem boundaries are delineated when an integral unit is formed by their physical
and biological parts. Defined pathways for material transport and cycling and for the flow of
energy are contained within a given integrated unit.
       Each level of organization within an ecosystem has functional and structural
characteristics.  At the ecosystem level, functional characteristics include, but are not limited to,
energy flow; nutrient, hydrologic, and biogeochemical cycling; and maintenance of food chains.
The sum of the functions carried out by ecosystem components provides many benefits to
humankind, as in the case of forest ecosystems (Smith, 1992).  Some of these benefits include
food, fiber production, aesthetics, genetic diversity, and energy exchange.
       A conceptual framework for discussing the effects of Os on ecosystems was developed
by the EPA Science Advisory Board (Young and Sanzone, 2002). Their six essential ecological
attributes (EEAs) include landscape condition, biotic condition, organism condition, ecological
processes, hydrological and geomorphological processes, and natural  disturbance regimes.
Figure 7-21 outlines how common anthropogenic stressors, including tropospheric Os, might
affect the essential ecological attributes outlined by the SAB.
       There is evidence that tropospheric Os is an important stressor of ecosystems, with
documented impacts on the biotic condition, ecological processes, and chemical/physical nature
of natural ecosystems (U.S. EPA, 2006). Most of the effects on ecosystems must be inferred
from Os exposure to individual plants and processes that are scaled up through the ecosystem
affecting processes such as energy and material flow, inter- and intraspecies competition, and net
primary productivity (NPP).  Thus, effects on individual keystone species and their associated
microflora and fauna, which have been shown experimentally, may cascade through the
ecosystem to the landscape level. By affecting water balance, cold hardiness, tolerance to wind
and by predisposing plants to insect and disease pests, Os may even impact the occurrence and
impact of natural disturbance (e.g., fire, erosion).
       Another approach to assessing O3 effects on ecosystems is the identification and use of
indicators. For example, the main indicators of phytotoxic Os exposures used for forest
ecosystems are visible foliar injury (as described in section 7.6.3.2 above) and radial growth of
                                          7-71

-------
Figure 7-21. Common anthropogenic stressors and the essential ecological attributes they
              affect.  Modified from Young and Sanzone (2002)
       Hydrologic alteration
       Habitat conversion
       Habitat fragmentation
       Climate (flange
       Invasive non-native species
       Turbidity/sedimentation
       Pesticides
       Disease/pest outbreaks
       Nutrient pukes
Hydrologic alteration
Habitat conversion
Habitat fragtnentj tion
donate change
Over-harvesting of vegetation
Large-scale invasive
 species introductions
Large-scale disease/pest outbreaks
       Dissolved oxygen depletion
       Oione (tropospheri-c)
       Landscape
       Condition
       Hydrologic alteration
         Habitat conversion
            Qimate change
Over-harvesting of vegetation
      Disease/pest outbreaks
         Altered fire regime
        Altered flood regime

                                      Bio tic
                                    Condition
       Hydrologic alteration
       Habitat conversion
       Climate change
       Turbidity/sedimentation
       Pesticides
       Nutrient pulses
       Dissolved oxygen depletion
       Ozone (tropospherk)
       Nitrogen oxides
                                    Chera
                                     Physical
                   Natural
                 Disturbance
                 Hydrology/
                Geomorptology

       Ecological
       Processes

       Hydrologic alteration
         HabitoS conversion
      Habitat fctgttMH&ttHM
            Climate change
     Turbidity/sedimentation
Hydrologic alteration
Habitat c onversion
Clitnate change
Pesticides
Disease/pest outbreaks
Nutrient pulses.
Dissolved oxygen depletion
Nitrogen oxides
                                                      7-72

-------
trees. Systematic injury surveys demonstrate that foliar injury occurs on (Vsensitive species in
many regions of the United States. However, there is not always a direct relationship between
the severity of the visible foliar symptoms and growth.  This essentially means it is difficult to
quantify or characterize the degree which EEAs may be impacted when visible foliar injury is
found in the field. Investigations of the relationship between changes in radial growth of mature
trees and ambient 63, in combination with data from many controlled studies with seedlings,
suggest that ambient O3 is reducing the growth of mature trees in some locations. However,
definitively attributing growth losses in the field to Os in a wide array of ecosystems is often
difficult because of confounding factors with other pollutants, climate, insect damage and
disease.
       The CD (U.S. EPA, 2006) outlines seven case studies where Os effects on ecosystems
have either been documented or are suspected.  However, in most cases, only a few components
in each of these ecosystems have been examined and characterized for Os effects and, therefore,
the full extent of ecosystem changes in these example ecosystems is not fully understood.
Clearly, there is a need for highly integrated ecosystem  studies that specifically investigate the
effect of Os on ecosystem structure and function in order to fully determine the extent to which
63 is altering ecosystem services.

       7.7.1   Evidence of Potential Ozone Alteration of Ecosystem Structure and
              Function
       The seven case studies listed in the 2006 CD demonstrate the potential for 63 to alter
ecosystem structure and function (U.S. EPA, 2006). The oldest and clearest example involves
the San Bernardino Mountain forest ecosystem. In this  example, O?, appeared to be a
predisposing factor leading to increased drought stress, windthrow, root diseases, and insect
infestation (Takemoto et al., 2001). Increased mortality of susceptible tree species, including
ponderosa and Jeffrey pine, resulting from these combined stresses has shifted community
composition towards white fir and incense cedar and has altered forest stand structure (Miller et
al., 1989). A shift of community composition towards white fir may make this ecosystem more
susceptible to fire.  Although the role of 63 was extremely difficult to separate from other
confounding factors, such as high N deposition, there is evidence that this shift in species
composition has altered trophic structure and food web dynamics (Pronos et al.,  1999) and C and
N cycling (Arbaugh et al., 2003). Ongoing research in this important ecosystem will reveal the
extent to which ecosystem services have been affected.
       One of the best-documented studies of population and community response to Os effects
are the long-term studies of common plantain (Plantago major) in native plant communities in
the United Kingdom (Davison and Reiling, 1995; Lyons et al., 1997; Reiling and Davison,
                                          7-73

-------
1992c). Elevated 63 significantly decreased the growth of sensitive populations of common
plantain (Pearson et al., 1996; Reiling and Davison, 1992a, b; Whitfield et al., 1997) and reduced
fitness as determined by decreased reproductive success (Pearson et al., 1996; Reiling and
Davison, 1992a).  While spatial comparisons of population responses to 63 are complicated by
other environmental factors, rapid changes in Os resistance were imposed by ambient levels and
variations in 63 exposure (Davison and Reiling, 1995). At the site of plantain seed collection,
the highest correlations occurred between O3 resistance and ambient O3 concentrations (Lyons et
al., 1997).  In this case study, it appeared that Os-sensitive individuals are being removed by Os
stress and the genetic variation represented in the population could be declining. If genetic
diversity and variation is lost in ecosystems, there may be increased vulnerability of the system
to other biotic and abiotic stressors, and ultimately a change in the services provided by those
ecosystems.
       Reconstructed ecosystems in artificial exposure experiments have also provided new
insight into how 63 may be altering ecosystem structure and function (Karnosky et al., 2005).
For example, the Aspen Free-Air CO2 Enrichment facility was designed to examine the effects of
both elevated CC^and Oson aspen (Populus tremuloides), birch (Betulapapyrifera), and sugar
maple (Acer saccharum) in a simple reconstructed plantation characteristic of Great Lakes
aspen-dominated forests (Karnosky et al., 2003; Karnosky et al., 1999). They found evidence
that the effects on above- and below-ground growth and physiological processes have cascaded
through the ecosystem, even affecting microbial communities (Larson et al., 2002; Phillips et al.,
2002). This study also confirmed earlier observations of  Os-induced changes in trophic
interactions involving keystone tree species, as well as important insect pests and their natural
enemies (Awmack et al., 2004; Holton et al., 2003; Percy et al., 2002).
       Collectively these examples suggest that Os is an important stressor in natural
ecosystems, but it is difficult to quantify the contribution  of 63 due to the combination of stresses
present in ecosystems.  Continued research, employing new approaches, will be necessary to
fully understand the extent to which Os is affecting ecosystem services.

       7.7.2   Effects on Ecosystem Services and Carbon Sequestration
       Since it has been established that 63 affects photosynthesis and growth of plants, 63 is
most likely affecting the productivity of forest ecosystems.  Therefore, it is desirable to link
effects on growth and productivity to essential ecosystem services.  However, it is  very difficult
to quantify ecosystem-level productivity losses  because of the amount of complexity in scaling
from the leaf-level or individual plant to the ecosystem level, and because not all organisms in an
ecosystem are equally affected by 63.
                                          7-74

-------
       Terrestrial ecosystems are important in the Earth's carbon (C) balance and could help
offset emissions of CO2 by humans if anthropogenic C is sequestered in vegetation and soils.
The annual increase in atmospheric CC>2 is less than the total inputs from fossil fuel burning and
land use changes (Prentice et al., 2001) and much of this discrepancy is thought to be attributable
to CC>2 uptake by plant photosynthesis (Tans & White, 1998). Temperate forests of the northern
hemisphere have been estimated to be a net sink of 0.6 to 0.7 Pg of C per year (Goodale et al.
2002). Ozone interferes with photosynthesis, causes some plants to senesce leaves prematurely
and in some cases, reduces allocation to stem and root tissue.  Thus, Os decreases the potential
for C  sequestration. For the purposes of this discussion, we define C sequestration as the net
exchange of carbon by the terrestrial biosphere. However, long-term storage in the soil organic
matter is considered to be the most stable form of C storage in ecosystems.
       In a study including all ecosystem types, Felzer et al. (2004), estimated that U.S. net
primary production (net flux of C into an ecosystem) was decreased by 2.6-6.8% due to Os
pollution in the late 1980's to early 1990's.  Ozone not only reduces C sequestration in existing
forests, it can also affect reforestation projects (Beedlow et al. 2004).  This  effect, in turn, has
been found to ultimately inhibit C  sequestration in forest soils which act as  long-term C storage
(Loya et al., 2003; Beedlow et al. 2004).  The interaction of rising Os pollution and rising CO2
concentrations in the coming decades complicates predictions of future sequestration potential.
Models generally predict that, in the future, C sequestration will increase with increasing CO2,
but often do not account for the decrease in productivity due to the local effects of tropospheric
Os.  In the presence of high Os levels, the stimulatory effect of rising CO2 concentrations on
forest productivity has been estimated to be reduced by more that 20% (Tingey et al., 2001;
Ollinger et al. 2002; Karnosky et al., 2003).
       In summary, it would be anticipated that meeting lower Os standards would increase the
amount of CO2  uptake by many ecosystems in the U.S.  However, the amount of this
improvement would be heavily dependent on the species composition of those ecosystems.
Many ecosystems in the U.S. do have Os sensitive plants.  For, example forest ecosystems with
dominant species such as aspen or ponderosa pine would be expected to increase CO2 uptake
more with lower Os than forests with more Os tolerant species.
                                          7-75

-------
REFERENCES


Abt Associates, Inc. (1995). Ozone NAAQS benefits analysis: California crops. Report to U.S. EPA, July.

Abt Associates Inc. (2007).  Technical Report on Ozone Exposure, Risk, and Impacts Assessments for Vegetation:
        Final Report.  Prepared for Office of Air Quality Planning and Standards, U.S.  Environmental Protection
        Agency, Research Triangle Park, NC. January 2007. Available electronically on the internet at:
        http://www.epa.gov/ttn/naaqs/standards/ozone/s_o3_cr_td.html.

Adams, R. M.; Hamilton, S. A.; McCarl, B. A. (1985) An assessment of the economic effects of ozone on U.S.
        agriculture. J. AirPollut. Control Assoc. 35: 938-943.

Adams, R. M.; Hamilton, S. A.; McCarl, B. A. (1986) The benefits of pollution control:  the case of ozone and U.S.
        agriculture. Am. J. Agric. Econ. 68: 886-893.

Alscher, R. G.; Amundson, R. G.; Gumming, J. R.; Fellows, S.; Fincher, J.; Rubin, G.; vanLeuken, P.; Weinstein, L.
        H. (1989) Seasonal changes in the pigments, carbohydrates and growth of red spruce as affected by ozone.
        NewPhytol.  113:211-223.

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://www.cmascenter.org/PDF/CMAQ  Scd Peer  Rev  July 5.pdf)

Amar, P., R. Bornstein, 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.; Hogsett, W. E.; Plocher, M.; Rodecap, K.; Lee, E. H. (2001) Blue wild-rye grass  competition
        increases the effect of ozone on ponderosa pine seedlings. Tree Physiol. 21: 319-327.

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 the 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, (2003) 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.

Ashmore M.R. (2002). Effects of oxidants at the whole plant and community level. In: Bell JNB, Treshow M, eds.
        Air pollution and plants. London, UK: J Wiley, 89-118.
                                                 7-76

-------
Ashmore, M.; Emberson, L.; Karlsson, P. E.; Pleijel, H. (2004) New directions: a new generation of ozone critical
        levels for the protection of vegetation in Europe (correspondence). Atmos. Environ. 38: 2213-2214

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, I; Palmer-Brown, D.; Fuhrer, I; 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.

Barnes, J. D.; Eamus, D.; Brown, K. A. (1990) The influence of ozone, acid mist and soilnutrient status on Norway
        spruce [Picea abies (L.) Karst.]: II. photosynthesis, dark respiration and soluble carbohydrates of trees
        during late autumn. New Phytol. 115: 149-156.

Beedlow P.A., Tingey D.T., Phillips D.L., Hogsett W.E. & Olszyk D.M. (2004) Rising atmospheric CO2 and carbon
        sequestration in forests. Frontiers in Ecology and the Environment 2, 315-322.

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.

Black, V. J.; Black, C. R.; Roberts, J. A.; Stewart, C. A. (2000) Impact of ozone on the reproductive development of
        plants. New Phytol. 147: 421-447.

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.

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.

Burns, R. M., B. H. Honkala,  tech. coords. (1990). Silvics of North America: 1. Conifers; 2. Hardwoods. Agriculture
        Handbook 654. U.S.  Department of Agriculture, Forest Service, Washington, DC. vol.2, 877 p.

Byun, D.W., Ching, J.K.S. (Eds.), 1999. Science Algorithms of the EPAModels-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.

Campbell, S.; Temple, P.; Pronos, J.; Rochefort, R.; Andersen, C. (2000) Monitoring for ozone injury in west coast
        (Oregon, Washington, California) forests in 1998. Portland, OR: U.S. Department of Agriculture, Forest
        Service, Pacific Northwest Research Station; general technical report no. PNW-GTR-495. Available:
        http://www.fs.fed.us/pnw/gtrs.htm [11 April, 2003].

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.
                                                 7-77

-------
CEPA. (2005) Review of the California Ambient Air Quality Standard for Ozone. Staff Report: Initial Statement of
        Reasons for Proposed Rulemaking. California Environmental Protection Agency, Air Resources Board

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.

Chappelka, A. H. (2002) Reproductive development of blackberry (Rubus cuneifolius) as influenced by ozone. New
        Phytol. 155: 249-255.

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 Assessment83: 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.

Dammgen, U.; Griinhage, L.; Haenel, H. D.; Ja'ger, H.  J. (1993) Climate and stress  in ecotoxicology: a coherent
        system of definitions and terms. Angew Bot. 67: 157-162.

Dann, M. S.;Pell,E. J. (1989) Decline of activity and quantity of ribulosebisphosphatecarboxylase/oxygenase and
        net photosynthesis in ozone-treated potato foliage. PlantPhysiol. 91: 427-432.

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.

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.

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 Karnosky D.F. (2000) Forest atmosphere carbon transfer
        storage-II (FACTS II) - The aspen free-air CO2 and O3 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. PlantPhysiol. 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

Elkiey, T.; Ormrod, D. P. (1979) Leaf diffusion resistance responses of three petunia cultivars to ozone and/or sulfur
        dioxide. J. Air Pollut. Control Assoc. 29: 622-625.

Emberson, L.; Ashmore, M. R.; Cambridge,  H. M.; Simpson, D.; Tuovinen, J. P. (2000a) Modelling stomatal ozone
        flux across Europe. Environ. Pollut. 109: 403-413.
                                                 7-78

-------
Emberson, L. D.; Wieser, G.; Ashmore, M. R. (2000b) Modelling of stomatal conductance and ozone flux of
        Norway spruce: comparison with field data. Environ. Pollut. 109: 393-402.

Engle, R. L.; Gabelman, W. H. (1966) Inheritance and mechanism for resistance to ozone damage in onion, Allium
        cepaL. Proc. Am. Soc. Hortic. Sci. 89: 423-430.


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.

Federal Register (1996) National Ambient Air Quality Standards for Ozone, Proposed Decision. 40 CFR 50; Federal
        Register 61: 65716.

Federal Register (1997) National Ambient Air Quality Standards for Ozone; Final Rule. 40 CFR 50; Federal
        Register 62: 38856.

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.

Finnan, J. M.; Burke, J. L.; Jones, M. B. (1997) An evaluation of indices that describe the impact of ozone on the
        yield of spring wheat (Triticum aestivum L). Atmos. Environ. 31: 2685-2693.

Fiore, A. M.; Jacob, D. J.; Mathur, R.; Martin, R. V. (2003) Application of empirical orthogonal functions to
        evaluate ozone simulations with regional and global models. J. Geophys. Res. (Atmos.) 108(D14):
        10.1029/2002JD003151.

Fiscus, E. L.; Reid, C. D.; Miller, J. E.; Heagle, A. S. (1997) Elevated CO2 reduces O3 flux and O3-inducedyield
        losses in soybeans: possible implications for elevated COa 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.

Flagler, R. B. (1998) Recognition of air pollution injury to vegetation: A pictorial atlas; second edition. Pittsburgh,
        PA: Air & Waste Management Association.

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.
                                                 7-79

-------
Goodale, C. L., Apps, M. I, Birdsey, R. A., Field, C. B., Heath, L. S., Houghton, R. A., Jenkins, J. C., Kohlmaier,
        G. H., Kurz, W., Liu, S., Nabuurs, G.-J., Nilsson, S. and Shvidenko, A. Z. (2002) Forest carbon sinks in the
        northern hemisphere. Ecol. Appl. 12, 891-899.

Grantz, D. A., McCool, P.H. (1992) Effect of ozone on Pima and Acala cottons in the San Joaquin Valley. In:
        Herber,  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.

Grantz, D. A.; Yang,  S. (2000) Ozone impacts on allometry and root hydraulic conductance are not mediated by
        source limitation nor developmental age. J. Exp. Bot. 51: 919-927.

Gregg, J. W., C.G.  Jones and T.E. Dawson (2003). "Urbanization effects on tree growth in the vicinity of New York
        City." Nature 424:  183-187.

Griinhage, 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.

Griinhage, L.; Haenel, H. D. (1997) PLATIN (PLant-ATmosphere-INteraction) I: a model of plant-atmosphere
        interaction for estimating absorbed doses of gaseous air pollutants. Environ. Pollut. 98: 37-50.

Griinhage, L.; Krupa, S. V.; Legge, A. H.; Jager, H. J. (2004) Ambient flux-based critical values of ozone for
        protecting vegetation: differing spatial scales and uncertainties in risk assessment. Atmos. Environ.  38:
        2433-2437.

Grulke, N. E.; Andersen, C.  P.; Fenn, M. E.; Miller, P. R. (1998) Ozone exposure and nitrogen deposition lowers
        rootbiomass of ponderosa pine in the  San Bernardino Mountains, California. Environ. Pollut. 103: 63-73.

Grulke, N. E.; Balduman, L. (1999) Deciduous conifers: high N deposition and O3 exposure effects on growth and
        biomass allocation in ponderosa pine.  Water Air Soil Pollut. 116: 235-248.

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.

Grulke, N. E.; Preisler, H. K.; Rose, C.; Kirsch, J.; Balduman, L. (2002) O3 uptake and drought stress effects on
        carbon acquisition of ponderosa pine in natural stands. New Phytol. 154: 621-631.

Guderian, R. (1977) Discussion of the suitability of plant responses as a basis for air pollution control measures. In:
        Billings, W.  D.; Golley, F.; Lange, O.  L.; Olson, J. S., eds. Air pollution: phytotoxicity of acidic gases and
        its significance in air pollution control. Berlin, Federal Republic of Germany: Springer Verlag; pp. 75-97.

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 mbra 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. Qual. 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.
                                                  7-80

-------
        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.; Wellburn, 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.; Cowling, E. B. (1997) The need for a long term cumulative secondary ozone standard - an ecological
        perspective. EM (January): 23-33

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.

Henderson, R.  (2006c) Letter from CASAC Chairman Rogene Henderson to EPA Administrator Stephen Johnson,
        October 24, 2006, EPA-CASAC-07-001.

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.

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.; Ormrod, D. P.; Taylor, G. E., Jr.; Tingey, D. T. (1987) 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).
                                                 7-81

-------
Horvath, L.; Nagy, Z.; Weidinger, T.; Artz, R.; Luke, W. T.; Valigura, R.; Pinto, J. P.; Womack, J. (1995)
        Measurement effluxes of trace gases (O3, NOx, SO2, CO2, HNO3), paniculate 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.

Isebrands, J. G.; Dickson, R. E.; Rebbeck, J.; Karnosky, 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.; Karnosky, D. F.
        (2001) Growth responses of Populus tremuloides clones to interacting carbon dioxide and tropospheric
        ozone. Environ. Pollut. 115: 359-371.

Jacobson, J. S. (1977) The effects of photochemical oxidants on vegetation. In: Ozon und Begleitsubstanzen im
        photochemischen Smog: das Kolloquium [Ozone and related  substances in photochemical smog: the
        colloquium]; September 1976; Dusseldorf, Federal Republic of Germany. Dusseldorf, Federal Republic of
        Germany: VDI-Verlag GmbH; pp.  163-173. (VDI-Berichte nr. 270).

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.

Karnosky, D.F., B. Mankovska, K. Percy, R.E. Dickson, O.K. Podila, J. Sober, A. Noormets, 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 O3-gradient and a FACE experiment." J. Water, Air and
        Soil Pollut.  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. (2003) Tropospheric O3 moderates responses of temperate hardwood forests to elevated
        CO2:  A synthesis of molecular to ecosystem results from the  Aspen FACE project. Funct. Ecol. 17: 289-
        304.

Karnosky, D.F., Pregitzer, K.S., Zak, D.R., Kubiske, M.E., Hendrey, G.R., Weinstein, D., Nosal, M. & Percy, K.E.
        2005  Scaling ozone responses of forest trees to the ecosystem level in a changing climate. Plant Cell
        Environ.28, 965-981

Kats, G.; Thompson, C. R.; Kuby, W. C. (1976) Improved ventilation of open top greenhouses. J. Air Pollut.
        Control Assoc. 26: 1089-1090.

Kats, G.; Olszyk, D. M.; Thompson, C. R. (1985) Open top experimental chambers for trees. J. Air Pollut. Control
        Assoc. 35: 1298-1301.
                                                 7-82

-------
King, J.S., M. E. Kubiske, K. S. Pregitzer, G. R. Hendrey, E. P. McDonald, C. P. Giardina, V. S. Quinn, D. F.
        Karnosky. (2005) Tropospheric O3 compromises net primary production in young stands of trembling
        aspen, paper birch and sugar maple in response to elevated atmospheric CO2. New Phytologist. 168:623-
        636.

Kraft, M.; Weigel, H. I; Mejer, G. I; 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. New Phytol. 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.; Kerner, 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.,Kohut, R.J., Amundson, R.G., (1993). Use of TREGRO to  simulate the effects of ozone on the
        growth of red spruce seedlings. Forest Science. 39: 453-464.

Laurence, J. A.; Retzlaff, W. A.; Kern, 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 ZELG models. For. Ecol. Manage. 146: 247-263.

Laurence, J. A.; Andersen, C. P. (2003) Ozone and natural systems: understanding exposure, response, and risk.
        Environ. Int. 29: 155-160.

Lee, E. H.;  Tingey, D. T.; Hogsett, W. E.; Laurence, J. A.  (2003) History of tropospheric ozone forthe San
        Bernardino Mountains of southern California,  1963-1999. Atmos. Environ. 37: 2705-2717.

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.;  Hogsett, W. E. (1999) Role of concentrations and time of day in developing ozone exposure indices for a
        secondary standard. J. Air Waste Manage. Assoc. 49: 669-681

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.
        Environmental Protection Agency, Environmental Research Laboratory; EPA report no.
        EPA-600/3-89-047. Available from: NTIS, Springfield, VA; PB89-195036.
                                                 7-83

-------
Lefohn, A. S.; Laurence, J. A.; Kohut, R. J. (1988) A comparison of indices that describe the relationship between
        exposure to ozone and reduction in the yield of agricultural crops. Atmos. Environ. 22: 1229-1240.

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.

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.
        http://www.ars.usda.gov/research/publications/publications.htm?SEQ_NO_l 15=142752

Loya W.M., Pregitzer K.S., Karberg N.J., King J.S. & Giardina C.P. (2003) Reduction of soil carbon formation by
        tropospheric ozone under elevated carbon dioxide. Nature 425, 705-707.

Luethy-Krause, B.; Landolt, W. (1990) Effects of ozone on starch accumulation in Norway spruce (Picea abies).
        Trees 4: 107-110.

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 Physiol. 101: 969-976.

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.

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.

Mandl, R. H.; Laurence, J. A.; Kohut, R. J. (1989) Development and testing of open-top chambers for exposing
    large, perennial plants to air  pollutants. J. Environ.  Qual. 18: 534-540.

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.

Massman, W. J.; Musselman, R.  C.; Lefohn, A. S. (2000) A conceptual ozone dose-response model to develop a
        standard to protect vegetation. Atmos. Environ. 34: 745-759.

Massman, W. J. (2004) Toward an ozone standard to protect vegetation based on effective dose: a review of
        deposition resistance and a possible metric. Atmos. Environ. In press.

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.

Matyssek, R.; Gunthardt-Goerg, M. S.; Maurer, S.; Keller, T. (1995) Nighttime exposure to  ozone reduces
        wholeplant production in Betula pendula. Tree Physiol. 15: 159-165.

McAinsh, M. R.; Evans, N. G.; Montgomery, L. T. (2002) Calcium signalling in stomatal responses to pollutants.
        New Phytol. 153:441-447.
                                                 7-84

-------
McLaughlin, S. B.; McConathy. (1983) Effects of SO2 and O3 on alocation of 14C-labeled photosynthate in
        Phaseolus vulgaris. Plant Physiol. 73: 630-635.

Mehlhorn, 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.; Sober, J.; Sober, A. (1990) Differential sensitivity of stomata and mesophyll tosudden exposure of bean
        shoots to ozone. Photosynthetica 24: 446-458.

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.

Monteith, J. L.; Elston, J. (1993) Climatic constraints on crop production. In: Fowden, L.; Mansfield, T.; Stoddart,
        J., eds. Plant adaptation to environmental stress.  London, United Kingdom: Chapman and Hall; pp. 3-18.

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.I, 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.

Morgan, P.B, Mies T.A., Bollero G.A., Nelson R.L., Long, S.P.  (2006) Season-long elevation of ozone
        concentration to projected 2050 levels under fully open-air conditions substantially decreases the  growth
        and production of soybean. New Phytologist  2006 170: 333-343.

Murphy, J. J.; Deluki, M. A.; McCubbin, D. R.; Kim, H. J. (1999) The cost of crop damage caused by ozone air
        pollution from motor vehicles. J.  Environ. Manage. 55: 273-289.

Musselman, R. C.; Massman, W. J. (1999) Ozone flux to vegetation and its relationship to plant response and
        ambient air quality standards. Atmos. Environ. 33: 65-73.

Musselman, R. C.; Minnick, T. J. (2000) Nocturnal stomatal conductance and ambient air quality standards for
      ozone. Atmos. Environ. 34: 719-733.

Neufeld, H. S.; Lee, E. H.; Renfro, J. R.; Hacker, W. D. (2000) Seedling insensitivity to ozone for three conifer
      species native to Great Smoky Mountains National Park. Environ. Pollut. 108: 141-151.

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.

NFS (2005) 2005 Annual Performance & Progress Report: Air Quality in National Parks. National Park Service.
      http://www2.nature.nps.gov/air/Pubs/pdf/gpra/Gpra2005_Report_03202006_Final.pdf

NRC (2004). Air quality management—United States. I. National Research Council (U.S.). Committee on Air
      Quality Management in the United  States. TD883.2.A64325 2004 363.739'25'0973—dc222004014594
      http://www.nap.edu/openbooy0309089328/html/
                                                 7-85

-------
Odum, E. P. (1963) Ecology. New York, NY: Holt, Rinehart and Winston. (Modern biology series).

Ollinger, S. V., Aber, J. D., Reich, P. B. and Freuder, R. J. (2002) Interactive effects of nitrogen deposition,
      tropospheric ozone, elevated CO2 and land use history on the carbon dynamics of northern hardwood
      forests.  Glob. Change Biol. 8(6), 545-562.

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.

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. New Phytol. 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.

Percy, K. E.; Awmack,  C. S.; Lindroth, 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.; Karnosky, D. F. (2002)
      Altered performance of forest pests under atmospheres enriched with CO2 and O^. Nature (London) 420:
      403-407.

Peterson, D. L.; Arbaugh, M. J.; Wakefield, V. A.; Miller, P. R. (1987) Evidence of growth reduction in ozone-
        injured Jeffrey pine (Pinus jeffreyi Grev and Balf) in Sequoia and Kings Canyon National Parks. JAPCA
        37: 906-912.

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.; Karlsson, 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.

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.
                                                  7-86

-------
Pitelka, L. F. (1988) Evolutionary responses of plants to anthropogenic pollutants. Trends in Ecology and
        Evolution 3: 233-236.

Prentice 1C, Farquhar GD, Fasham MJR, et al. (2001) The Carbon Cycle and Atmospheric Carbon Dioxide. In
        Climate Change 2001: The Scientific Basis. Contribution of Working Group I to the Third Assessment
        Report of the Intergovernmental Panel on Climate Change, (ed J. T. Houghton YD, D. J. Griggs, M.
        Noguer, P. J. van der Linder, X. Dai, K. Maskell, and C. A. Johnson), pp. 241-280. Cambridge
        University Press, Cambridge, United Kingdom and New York, NY, USA.

Rao, M. V.; Davis, K. R. (2001) The physiology of ozone induced cell death. Planta 213: 682-690.

Rao, M. V.; Lee, H. L; 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
        ofPlantago 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 ofPlantago 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 Physiol. 20: 195-202.

Rizzo, M (2005). Evaluation of a quadratic approach for adjusting distributions of hourly ozone concentrations to
        meet air quality standards. November 7, 2005.

Rizzo, M.  (2006).  A distributional comparison between different rollback methodologies applied to ambient ozone
        concentrations. May 31,2006

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 flushes of loblolly pine seedlings. For. Sci. 37: 1078-1098.
                                                  7-87

-------
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.

Simini, M.; Skelly, J. M.; Davis, D. D.; Savage, J. E.; Comrie, A. C. (1992) Sensitivity of four hardwood species to
        ambient ozone in north central Pennsylvania. Can. J. For. Res. 22: 1789-1799.

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.

Smith, G., Coulston J., Jepsen, J. and Prichard, T. (2003) "A national ozone biomonitoring program: Results from
        field surveys of ozone sensitive plants in northeastern forest (1994-2000)" Environmental Monitoring and
        Assessment^''(3): 271-291.

Somers, G. L.; Chappelka, A. H.; Rosseau, P.; Renfro, J. R. (1998) Empirical evidence of growth decline related to
        visible ozone injury. For. Ecol. Manage. 104: 129-137.

Stitt, M. (1996) Plasmodesmata play an essential role in sucrose export from leaves: a step towards an integration of
        metabolic biochemistry and cell biology. Plant Cell 8: 565-571.

Takemoto, B. K.; Bytnerowicz, A.; Fenn, M. E. (2001) Current and future effects of ozone and atmospheric nitrogen
        deposition on California's mixed conifer forests. For. Ecol. Manage. 144:  159-173.

Tans PP, White JWC (1998) In balance, with a little help from the plants. Science, 281, 183-184.

Taylor R. (1994) "Deterministic versus stochastic evaluation of the aggregate economic effects of price support
        programs" Agricultural Systems 44: 461-473.

Taylor, C.R. "AGSIM: Model Description and Documentation." Agricultural Sector Models for the United States.
        C.R. Taylor, K.H. Reichelderfer, and S.R. Johnson, eds. Ames IA: Iowa State University Press, (1993).

Temple, P. J.; Riechers, G. H.; Miller, P. R.; Lennox, R. W. (1993) Growth responses of ponderosa pine to longterm
        exposure to ozone, wet and dry acidic deposition, and drought. Can. J. For. Res. 23: 59-66.

Tingey, D. T.; Files, R. C.; Wickliff, C. (1975) Activity changes in selected enzymes from soybean leaves following
        ozone exposure. Physiol. Plant. 33: 316-320.

Tingey, D. T.; Standley, C.; Field, R. W. (1976)  Stress ethylene evolution: a measure of ozone effects on plants.
        Atmos. Environ. 10: 969-974.

Tingey, D. T.; Taylor, G. E., Jr. (1982) Variation in plant response to ozone: a conceptual model of physiological
        events. In: Unsworth, M. H.; Ormrod, D. P., eds. Effects of gaseous air pollution in agriculture and
        horticulture. London, United Kingdom: Butterworth Scientific; pp. 113-138.

Tingey, D. T.; Laurence, J. A.; Weber, J. A.; Greene, J.; Hogsett, W. E.; Brown, S.; Lee, E. H. (2001) Elevated CO2
        and temperature alter the response of Pinusponderosa to ozone: A simulation analysis. Ecol. Appl.
        11: 1412-1424.

Tingey, D. T.; Hogsett, W. E.; Lee, E. H.; Laurence, J. A. (2004) Stricter ozone ambient air quality standard has
        beneficial effect on Ponderosa pine in California. Environ. Manage. 34: 397-405.

Unsworth, M. H.; Heagle, A. S.; Heck, W. W. (1984a) Gas exchange in open-top field chambers -1 measurement
        and analysis of atmospheric resistances to gas exchange. Atmos.  Environ. 18: 373-380.

Unsworth, M. H.; Heagle, A. S.; Heck, W. W. (1984b) Gas exchange in open-top field chambers - II resistances to
        ozone uptake by soybeans. Atmos. Environ. 18: 381-385.
                                                  7-S

-------
Urban, D.L., G.B. Bonan, T.M. Smith, H.H. Shugart (1991). "Spatial applications of gap models." For. Ecol. Mgmt.
        42:95-110.

USD A, NRCS. 2006. The PLANTS Database (http://plants.usda.gov, December 2006). National Plant Data Center,
        Baton Rouge, LA 70874-4490 USA.

U.S. Environmental Protection Agency (1978) Air quality criteria for ozone and other photochemical oxidants.
        Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and
        Assessment Office; report no. EPA-600/8-78-004. Available from: NTIS, Springfield, VA; PB80-124753.

U.S. Environmental Protection Agency (1986) Air quality criteria for ozone and other photochemical oxidants.
        Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and
        Assessment Office; report nos. EPA-600/8-84-020aF-eF. 5v. Available from: NTIS, Springfield, VA;
        PB87-142949.

U.S. Environmental Protection Agency (1989). Review of National Ambient Air Quality Standards for Ozone:
        Assessment of Scientific and Technical Information - OAQPS Staff Paper. Office of Air Quality Planning
        and Standards, Research Triangle Park, NC.

U.S. Environmental Protection Agency (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.

U.S. Environmental Protection Agency (1996b). Review of National Ambient Air Quality Standards for Ozone:
        Assessment of Scientific and Technical Information - OAQPS Staff Paper. EPA/452/R-96-007. Office of
        Air Quality Planning and Standards, Research Triangle Park, NC.

U.S. Environmental Protection Agency (2003a). Ozone Injury to Forest Trees.  In:  EPA's Draft Report on the
        Environment Technical Document. EPA-600-R-03-050. U.S. EPA, ORD, Washington, DC, page 5-19

U.S. Environmental Protection Agency (2003b).Clean Air Status and Trends Network (CASTNet) 2001 Quality
        Assurance Report; Research Triangle Park, NC: Office of Air Quality Planning and Standards. Report from
        EPA Contract No. 68-D-98-112.

U.S. Environmental Protection Agency (2004). The Ozone Report: Measuring Progress through 2003. EPA/454/K-
        04-001.  Office of Air Quality Planning and Standards, Research Triangle Park, NC.

U.S. Environmental Protection Agency (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 http://www.epa.gov/ttn/naaqs/standards/ozone/s_O3_cr_pd.html.

U.S. Environmental Protection Agency (2005b). Technical Support Document for the Final Clean Air Interstate
        Rule: Air Quality Modeling. Office of Air Quality Planning and Standards, Research Triangle Park, NC.

U.S. Environmental Protection Agency (2006) Air Quality Criteria for Ozone and Related Photochemical Oxidants
        (Final). Office of Research and Development, National Center for Environmental Assessment, Research
        Triangle Park, NC. EPA/600/R-05/004aF-cF, 2006.

USDA Forest Service: 1999, 'Forest Health Monitoring 1999 Field Methods Guide. USDA Forest Service',
        National Forest Health Monitoring Program, Research Triangle Park, NC.

Vahala, I; Ruonala, R.; Keinanen, M.; Tuominen, H.; Kangasjarvi, J. (2003) Ethylene insensitivity modulates
        ozone-induced cell death in birch. Plant Physiol. 132: 185-195.

Walcek, C. J.; Yuan, H. H.; Stockwell, W. R. (1997) The influence of aqueous-phase chemical reactions on ozone
        formation in polluted and nonpolluted clouds. Atmos. Environ. 31: 1221-1237.
                                                 7-89

-------
Weber, J. A.; Clark, C. S.; Hogsett, W. E. (1993) Analysis of the relationship(s) among O3 uptake, conductance, and
        photosynthesis in needles of Pinus ponderosa. Tree Physiol. 13: 157-172.

Weinstein, D.A., Beloin, R.M., R.D. Yanai (1991). "Modeling changes in red spruce carbon balance and allocation
        in response to interacting ozone and nutrient stress." Tree Physiology 9:  127-146.

Weinstein, D.A., J.A. Laurence, W.A. Retzlaff, J.S. Kern, E.H. Lee, W.E. Hogsett, J. Weber (2005). Predicting the
        effects of tropospheric ozone on regional productivity of ponderosa pine and white fir. Forest Ecology and
        Management 205: 73-89.

Wellburn, A. R. (1990) Why are atmospheric oxides of nitrogen usually phytotoxic and not alternative fertilizers?
        NewPhytol. 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. E.; 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.

Winner, W. E.; 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.

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.pdft9December, 2003].
                                                  7-90

-------
    8.   STAFF CONCLUSIONS AND RECOMMENDATIONS ON THE
                           SECONDARY O3 NAAQS
8.1    INTRODUCTION
       This chapter presents staff conclusions and recommendations regarding an appropriate
range of options for the Administrator to consider in selecting a pollutant indicator,  averaging
time, form, and level for the secondary 63 standard.  In so doing, this chapter describes the
results and conclusions of staff assessments of scientific evidence presented in the CD and of air
quality, exposure, and risk analyses presented in Chapters 2 and 7 herein. Comments and advice
received from CAS AC in their review of earlier drafts of this document, as well as comments on
earlier drafts submitted by  interested parties, that have significantly informed the  development of
staffs views, are also discussed.
       In presenting policy options for the Administrator's consideration, we note that the final
decision on retaining or revising the current secondary Os standard is largely a public welfare
policy judgment.  A final decision should draw upon scientific information and analyses about
welfare effects, exposure and risks,  as well as judgments about the appropriate response to the
range of uncertainties that are inherent in the scientific evidence and analyses. Our approach to
informing these judgments, discussed more fully  below, is consistent with the requirements of
the NAAQS provisions of the Act and with how EPA and the courts have historically interpreted
the Act. These provisions require the Administrator to establish secondary standards that, in the
Administrator's judgment,  are requisite to protect public welfare from any known or anticipated
adverse environmental effect. In so doing, the Administrator seeks to establish standards that are
neither more nor less stringent than necessary for this purpose.

8.2    APPROACH
       Welfare effects, as defined in section 302(h) (42 U.S.C. 7602(h)) of the Clean Air Act
include, but are not limited to, "effects on soils, water, crops, vegetation, manmade materials,
animals, wildlife, weather,  visibility, and climate, damage to and deterioration of property, and
hazards to transportation, as well as effects on economic values and on personal comfort and
well-being." As in the last review, this review has focused on vegetation and crops, since effects
on these public welfare categories are well-studied and currently known to be of most concern at
Os concentrations typically occurring in the U.S.  Further, by adversely affecting  natural
vegetation and commercial crops, O3 may  also indirectly adversely affect natural  ecosystems and
their components (e.g., soils, water, animals, and wildlife). Therefore, these important but less
well-studied indirect effects will be qualitatively  discussed. As discussed above in Chapter 7, for
                                         J-l

-------
other welfare effects categories, insufficient new information was available to inform the
selection of an indicator, form, averaging time or level for a distinct secondary standard and they
are not discussed further, except in terms of research needs.
       In evaluating whether the current secondary standard is adequate or whether
consideration of revisions is appropriate, we adopted an approach in this review that builds upon
the general approach used in the last review and reflects the broader body of evidence now
available. In developing conclusions and recommendations for the Administrator to consider in
this review, staff presents effects-, exposure- and risk-based considerations.  We have expanded
and modified the exposure and risk assessments to reflect the availability of new tools,
assessment methods, and a larger and more diverse body of evidence. We have taken a weight
of evidence approach that evaluates information across the variety of vegetation-related research
areas described in the CD (e.g., seedling and mature forest tree species and commodity, fruit,
vegetable and forage crop species), and includes assessments of air quality, exposures, and
qualitative and quantitative risks associated with alternative air quality scenarios.
       With respect to vegetation effects information,  we have evaluated the conclusions drawn
at the end of the last review in light of more recent evidence from chamber, free air, gradient,
model and field-based observation studies for a variety of vegetation effects endpoints. We
place greater weight on U.S. studies due to the often species-, site-, and climate-specific nature
of Os-related vegetation response. With respect to quantitative exposure- and risk-based
considerations, we have relied on both monitored and interpolated 63 exposures as described in
section 7.5. of Chapter 7. Several alternative air quality scenarios were selected for evaluation to
reflect a range of alternative standards under consideration.  These scenarios include current "as
is" air quality (2001), as well as six "just meet" scenarios for which interpolated 63 air quality is
adjusted using a rollback method to simulate just meeting a range of alternative standards.
Uncertainties associated with the exposure and risk assessments are also discussed, including,
where possible, some sense of the direction and/or magnitude of the uncertainties  that should be
taken into account as one considers these estimates. With regard to the use of the TREGRO
model for estimating mature tree risks, staff acknowledges the presence of unknown and
unquantifiable sources  of uncertainty, as is typical with all  such models.
       In this review, a series of general questions frames our approach to informing conclusions
and the identification of an appropriate range of policy options for consideration by the
Administrator regarding the current secondary Os standard. Our consideration of the adequacy
of the current standard  begins in section 8.3.1  by addressing questions such as the following:
      •   To what extent does newly available information reinforce or call into question
         evidence of associations with effects identified in the last review?
                                          8-2

-------
      •   To what extent does newly available information reinforce or call into question any of
         the basic elements of the current standards?
      •   To what extent have important uncertainties identified in the last review been reduced
         and have new uncertainties emerged?
To the extent that the available information suggests that revision of the current standards may
be appropriate to consider, we explore whether the currently available information supports
consideration of a standard that is either more or less protective by addressing the following
questions:
      •   Is there evidence that vegetation effects extend to ambient 63 concentration levels that
         are as low as or lower than had previously  been observed, and what are the important
         uncertainties associated with that evidence?
      •   Are exposures and vegetation risks of concern estimated to occur in areas that meet the
         current standard; are they important from a public welfare perspective; and what are
         the important uncertainties associated with the estimated risks?
To the extent that there is support for consideration of revised standards, we then identify a range
of alternative standards (in terms of an indicator for photochemical oxidants, averaging time,
level, and form in sections 8.3.2 through 8.3.5 below, respectively) that staff feels are appropriate
for the Administrator to consider and that reflect staff conclusions and recommendations on the
science, taking into account other public welfare policy considerations.  In so doing, staff
addresses the following questions:
      •   Does the evidence provide support for considering a different O3 indicator?
      •   Does the evidence provide support for considering different averaging times?

      •   What ranges of levels and forms of alternative standards are supported by the evidence,
         and what are the uncertainties and limitations in that evidence?
      •   To what extent do specific levels and forms of alternative standards reduce the
         estimated exposures of concern and risks attributable to O3 and other photochemical
         oxidants, and what are the uncertainties in  the estimated exposure and risk reductions?
       A summary of staff conclusions and recommendations regarding a range of policy
options identified for the Administrator's consideration, as well as key CAS AC and  public
commenter views concerning whether, and if so how, to revise the current secondary 63 standard
is presented in section 8.3.6 below.  This chapter concludes with a discussion of key
uncertainties and recommendations for additional research related to setting a secondary Os
NAAQS in section 8.4.
                                          8-3

-------
8.3    SECONDARY O3 STANDARD
       The current secondary standard is a 3-year average of the annual 4th-highest maximum 8-
hr average set at a level of 0.08 ppm.  This standard was selected to provide protection to the
public welfare, especially agricultural crops and other at-risk sensitive plant species, against a
wide range of Os-induced effects.  As an introduction to our discussion in this section of the
adequacy of the current Os standard, it is useful to summarize the key factors that formed the
basis of the decision in the last review to revise the averaging time, level and  form of the then
current 1-hr secondary standard.
       In the 1996 proposal notice (61 FR 65716), the Administrator proposed to replace the
then existing  1-hr O3 secondary NAAQS with one of two alternative new standards:  a standard
identical to the proposed 0.08 ppm, 8-hr primary standard (described above),  or alternatively, a
new seasonal standard expressed as a sum of hourly concentrations greater than or equal to 0.06
ppm, cumulated over 12 hours per day during the maximum 3-month period during the Os
monitoring season (SUM06), set at a level of 25 ppm-hr. At the time, this latter standard was
considered to be an annual standard. This proposal was based on a thorough review of the latest
scientific information  available and described in the 1996 Os CD, as well as (1) staff assessments
of the policy-relevant information  in the 1996 Os CD presented in the 1996 Os Staff Paper
including air quality, vegetation exposure and risk, and economic values; (2) consideration of the
degree of protection to vegetation potentially afforded by the proposed 0.08 ppm, 8-hr primary
standard; (3) CASAC  advice and recommendations; and (4) public comments.
       In the final rule for the O3 NAAQS published in July  1997 (62 FR 38877), the
Administrator decided to replace the then current 1-hr, 0.12-ppm secondary NAAQS with a
standard that was identical in every way to the new revised primary standard of an 0.08 ppm
annual 4th-highest maximum 8-hr average standard averaged over 3 years. Her decision was
based on her judgment that: (1) the then existing secondary standard did not provide adequate
protection for vegetation against the adverse welfare effects of 63; (2) reflected CASAC advice
"that a secondary NAAQS, more stringent than the present primary standard, was necessary to
protect vegetation from 63" (Wolff, 1996); (3) the new 8-hr average standard would provide
substantially improved protection for vegetation from O3-related adverse effects as compared to
the level of protection provided by the then current 1-hr, 0.12-ppm secondary standard; (4)
significant uncertainties remained  with respect to exposure dynamics, air quality relationships,
and the exposure,  risk, and monetized valuation analyses presented in the proposal, resulting in
only rough estimates of the increased public welfare likely to be afforded by each of the
proposed alternative standards; (5) there was value in allowing more time to obtain additional
information to better characterize Os-related vegetation effects under field conditions from
additional research and to develop a more complete rural monitoring network and air quality
                                         8-4

-------
database from which to evaluate the elements of an appropriate seasonal secondary standard; and
(6) there was value in allowing more time to evaluate more specifically the improvement in rural
air quality and in Os-related vegetation effects resulting from measures designed to attain the
new primary standard (62 FR 38877-78).
       The Administrator further concluded (62 FR 38877-78) that continued research on the
effects of 63 on vegetation under field conditions and on better characterizing the relationship
between O3 exposure dynamics and plant response would be important in the next review
because:
      •   The available biological database highlighted the importance of cumulative, seasonal
         exposures as a primary determinant of plant responses.
      •   The association between daily maximum 8-hr Os concentrations and plant responses
         had not been specifically examined in field tests.
      •   The impacts of attaining an 8-hr, 0.08 ppm primary standard in upwind urban areas on
         rural air quality distributions could not be characterized with confidence due to limited
         monitoring data and air quality modeling in rural and remote areas.

       8.3.1  Adequacy of Current Os Standard
       The new evidence available in this review continues to support and strengthen key
policy-relevant conclusions drawn in the previous review (U.S. EPA, 2006).  Based on this new
evidence, the current CD once more concludes that:  (1) a plant's response to Os depends upon
the cumulative nature of ambient exposure (e.g., concentration times duration) as well as the
temporal dynamics of those concentrations; (2) current ambient concentrations in many areas of
the country are sufficient to impair growth of numerous common and economically valuable
plant and tree species; (3) the entrance of 63 into the leaf through the stomata is the critical step
in O3  effects; (4) effects can occur with only a few hourly concentrations above 80 ppb; (5) other
environmental biotic and abiotic factors are also influential to the overall impact of Os on plants
and trees; and (6)  a high degree of uncertainty remains in our ability to assess the impact of 63
on ecosystem services.  The effects-based evidence described in the CD underlying the
reaffirmation of these conclusions will be discussed in more detail in the sections that follow.
Based on the above policy-relevant findings from the CD, and while recognizing that important
uncertainties and research questions remain, we also conclude that progress has been made since
the last review and thus, we generally find support in the available effects-based evidence for
consideration of an 63 standard that is at least as protective as the current standard and do not
find support for consideration of an Os standard that is less protective than the current standard.
This general conclusion is consistent with the advice and recommendations of CASAC and with

-------
the views expressed by all the interested parties who provided comment on the previous draft of
this document.
       Having reached this general conclusion, we then evaluated the adequacy of the current
standard by considering to what degree risks to vegetation and ecosystems would be expected to
occur after just meeting the current as well as a range of alternative secondary standards. As
discussed in Chapter 7 and in greater detail below, staff conclusions regarding the adequacy of
the current standard are based on the available vegetation effects, exposure and risk-based
evidence (section 8.3.1.1) and CASAC and public commenter views (section 8.3.1.2) in
conjunction with the additional policy-relevant considerations presented under the discussions on
indicator, averaging time, form, and level (sections 8.3.2 through 8.3.5). In evaluating the
strength of this information, staff has taken into account the uncertainties and limitations in the
scientific evidence and analyses as well as considered the views of CASAC and other interested
parties provided on the second draft of this document.

       8.3.1.1 Vegetation Evidence-, Exposure- and Risk-Based Considerations
       In the last review, crop yield and seedling biomass loss open-top chamber (OTC) data
provided the basis for staff analyses, conclusions, and recommendations (U.S. EPA, 1996b).
Since then, several additional lines of evidence have progressed sufficiently to provide staff with
a more complete and coherent picture of the scope of (Vrelated vegetation risks, especially
those currently faced by seedling, sapling and mature tree species growing in field settings, and
indirectly, forested ecosystems.  Specifically, new research reflects an increased emphasis on
field-based exposure methods (e.g., free air exposure and ambient gradient), improved field
survey biomonitoring techniques, and mechanistic tree process models. Findings from each of
these research areas are discussed separately below.  However, in reaching conclusions regarding
the adequacy of the current standard, staff has considered the combined information from all
these areas together, using an integrated, weight of evidence approach.
       In evaluating the degree to which the current standard is adequate in protecting
vegetation at the national  scale, staff has relied on both measured and modeled air quality
information. For some effects, like visible foliar injury and modeled mature tree growth
response, only monitored air quality information was used. For other effects categories (e.g.,
crop yield and tree seedling growth), staff relied on interpolated Os exposures.  Staff recognizes
that exposures predicted by this interpolation method are more uncertain.  The uncertainties
associated with this approach are discussed under the exposure assessment discussion below.
Additional sources of uncertainty associated with the risk assessment are described in the section
preceding the discussion of seedling and mature tree biomass-loss risk results.

-------
       Visible Foliar Injury Evidence
       Recent systematic injury surveys continue to document visible foliar injury symptoms
diagnostic of phytotoxic Os exposures on sensitive bioindicator plants. These surveys produce
more expansive evidence than that available at the time of the last review that visible foliar
injury is occurring in many areas of the U.S. under current ambient conditions.  Staff performed
an assessment combining recent U.S. Forest Service Forest Inventory and Analysis (FIA)
biomonitoring site data with the county level air quality data for those counties containing the
FIA biomonitoring sites.  This assessment showed that incidence of visible foliar injury ranged
from 21 to 39% during the four-year period (2001-2004) across all counties with air quality
levels at or below that of the current 8-hr standard.  The magnitude of these percentages suggests
that phytotoxic exposures sufficient to induce visible foliar injury would still occur in many areas
after meeting the level of the current secondary standard.  Additionally, the data show that
visible foliar injury occurrence is geographically widespread and is occurring on a variety of
plant species in forested and other natural systems (see Figure 7-19 in section 7.6.3.2). Linking
visible foliar injury to other plant effects is still problematic. However, its presence indicates
that other Os-related vegetation effects could  also be present.
       The presence of visible foliar injury can adversely impact the public welfare.  For
example, visible foliar injury in national parks and wilderness areas can impact the aesthetic
experience for both outdoor enthusiasts and the occasional park visitor.  In  addition, because
these areas are afforded a higher degree of protection, the presence of (Vinduced vegetation
effects, including visible foliar injury, can take on increased significance. Specifically, federal
land managers (FLMs) ".. .have determined that given the high ecological,  aesthetic, and
intrinsic value of federal lands,  all native species are significant and warrant protection" (NFS,
2000). As a result, FLMs have  identified visible foliar injury, along with other Os-induced
vegetation effects,  as air quality related values (AQRV) of concern (NFS, 2000).  As shown in
Appendix 7J, numerous O3 sensitive species are found on Class I federal lands.  In addition, the
presence of visible foliar injury also has the potential to economically impact for those who rely
on healthy looking vegetation for their livelihood (e.g., horticulturalists, farmers of leafy crops,
landscapers, Christmas tree growers).  Many ornamental species have been listed as sensitive to
Os (Abt,  1993). Similarly, early senescence of fall foliage could also diminish the time available
for viewing fall foliage, important in some regions of the country in drawing tourists.  Although
data are not available to allow the quantification of these impacts, the potential for their existence
should not be overlooked.
       Exposure-Based Considerations
       As described in Chapter 7, due to the paucity of rural 63 monitoring data, it was
necessary to select an interpolation method that could be used to characterize Os air quality over

                                         8-7

-------
broad geographic areas. Staff recognizes there are inherent uncertainties in the interpolation that
must rely on sparse data that, for the most part, are representative of urban and near-urban areas.
The interpolation method used for the western U.S. contains additional sources of uncertainty
associated with the use of CMAQ model outputs to develop scaling factors for the interpolation.
See section 7.5 of chapter 7 for details on how the interpolation was constructed and how staff
quantified the uncertainties (error and bias) associated with the interpolation.  This quantification
of exposure uncertainty for the interpolation represents a notable improvement over the 1996
assessment which did not have an evaluation of the exposure surface.  In general, this
interpolation method under-predicts higher 12-hr W126 exposures. Due to the important
influence of higher exposures in determining risks to plants, this feature of the interpolated
surface could result in an under-estimation of risks to vegetation in some areas.  Taking these
uncertainties into account, and given the absence of more complete rural monitoring data, staff
judged that this approach was appropriate to use in developing national vegetation exposure and
risk assessments that estimate relative changes in risk for the various alternative standards
analyzed.
       To evaluate changing vegetation exposures and risks under selected "just meet"
scenarios, staff analytically adjusted 2001 base year air quality distributions with a rollback
method (Horst and Duff,  1995; Rizzo, 2005 & 2006) to reflect "just meeting" the current and
alternative secondary standard options.  This technique combines both linear and quadratic
elements to reduce higher 63 concentrations more than lower ones. In this regard, the rollback
method attempts to account for reductions  in emissions without greatly affecting lower
concentrations near ambient background levels. The following "just meet" air quality scenarios
were generated along with maps for several scenarios (see Figures 7-7, 7-8, 7-9, 7-10):
      •  4th-highest daily maximum 8-hr average: 0.084 ppm (the effective level of the current
         standard) and 0.070 ppm levels
      •  3-month, 12-hr. SUM06: 25 ppm-hr (proposed in the 1996 review) and 15 ppm-hr
         levels
      •  3-month, 12-hr. W126: 21 ppm-hr and 13 ppm-hr levels
       Staffs rationale for selecting these  six alternative standards for evaluation is presented
here and in section 7.5.1 of Chapter 7. The two 8-hr average levels were chosen as possible
alternatives  of the current form for comparison with the cumulative seasonal alternative forms.
For both the SUM06 and W126 forms, the two levels were selected on the basis of the associated
levels of crop yield loss protection described in the last review. Specifically, both the upper
levels of SUM06 (25 ppm-hr) and W126 (21  ppm-hr) were associated with a level of crop
protection of approximately no more than 10% yield loss in 50% of crop cases studied in the

-------
National Crop Loss Assessment Network (NCLAN) experiments (section 7.6.2.2.).
Alternatively, the lower levels of both SUM06 (15 ppm-hr) and W126 (13 ppm-hr) were
associated with a level of crop protection of approximately no more than 10% yield loss in 75%
of NCLAN cases. Another level to note is the upper level benchmark of W126 of 31 ppm-hr that
approximates the upper end of the SUM06 range analyzed in the last review (U.S. EPA, 1996b)
and which was associated with no more than 17% yield loss in 50% of crop cases as described in
the last review.  The above levels have also been associated with varying levels of tree seedling
biomass loss protection based on a similar set of tree seedling studies performed by scientists in
the National Health and Environmental Effects Research Lab, Western Ecology Division
(NHEERL-WED).
      Under the base year (2001) air quality, a large portion of California had 12-hr W126
above 31 ppm-hr, which has been reported to produce approximately 14% biomass loss in 50%
of NHEERL-WED tree seedling studies. Broader multi-state regions in  the east and west are
predicted to have levels of air quality above the W126 level of 21 ppm-hr, which is
approximately equal to the secondary standard proposed in 1996 and is associated with
approximately 9% biomass loss in 50% of tree seedlings studied.  Much of the east and Arizona
and California have 12-hr W126 values above 13 ppm-hr which has been reported to allow
approximately 7% biomass loss in 25% of tree seedlings studied.  Although there is appreciable
uncertainty associated with these exposure estimates, the results of the exposure assessment
indicates that current air quality levels could result in significant impacts to vegetation in some
areas.
      When 2001 air quality is rolled back to just meet the current 8-hr secondary standard, the
overall 3-month 12-hr W126  exposures do not improve by much (Figure 7-7). Under this
scenario, there are still many  areas of the country that have seasonal Os levels above the 12-hr
W126 level of 21 ppm-hr. The exposure maps generated for the 0.070 ppm, 4th-highest
maximum 8-hr average alternative standard, the SUM06 alternatives of 25 and 15 ppm-hr, and
the W126 alternatives  of 21 and 13 ppm-hr (Figures 7-8, 7-9, 7-10), all showed a markedly
improved picture of 63 air quality compared to the current standard (Figure 7-7). Thus, the staff
observes that all other alternative standards, when met at all locations, would be expected to
provide  improved protection of vegetation from seasonal Os exposures of concern over the
current standard. As expected, however, the greatest improvements in air quality and estimated
exposures to sensitive vegetation were observed when just meeting the lower W126 alternative
of 13 ppm-hr, the SUM06 alternative of 15 ppm-hr, and the 0.07 ppm, 8-hr alternative standard.
      Risk-Based Considerations
      This review continues to rely upon the concentration-response (C-R) functions developed
from OTC exposure systems (also relied upon in the 1996 review).  Due to what has been

                                        8-9

-------
described in the scientific literature as the "chamber effect," some continue to express concern as
to the appropriateness of applying OTC generated C-R functions to non-chambered
environments. A shift toward the use of more field-based approaches (e.g., free air exposure and
ambient gradient) in recent research has occurred, providing information in the peer-reviewed
literature that at least qualitatively informs how one might weigh this concern.  These new field-
based studies, conducted on a limited number of crop and tree seedling species to date,
demonstrate plant growth and visible foliar injury responses similar in nature and magnitude to
those observed previously under OTC exposure conditions. These findings lend qualitative
support to the conclusion that OTC conditions do not fundamentally alter the nature of the Os-
plant response.  A related concern with respect to the use of the OTC C-R functions for crops is
the concern that the crop varieties grown today may have Os sensitivities significantly different
than those used to derive the NCLAN crop and OTC tree seedling C-R functions which are relied
upon in this review. Nothing in the recent literature, however,  suggests that the Os sensitivity of
crop or tree species studied in the last review and for which C-R functions were developed has
changed significantly in the intervening period.  Indeed, in the few recent studies where this is
examined, Os sensitivities are found to be as great or greater than those observed in the last
review. As a result, staff continues to rely on the C-R functions available in the last review for
predicting relative crop yield and tree seedling biomass loss potentials across a range of possible
ambient Os exposures.
       An additional source of uncertainty not described or accounted for in the last review is
that associated with the presence of a decreasing Os gradient from the height of the monitor
probe down to the lower plant canopy heights for most crop and seedling trees. The presence of
this gradient makes less certain the predictions of current crop and tree seedling exposures and
the associated yield and biomass losses, respectively, based on ambient monitor data.  Staff
selected a 10% reduction factor to represent the maximum gradient believed to occur for daylight
hours.  However, recognizing that the actual downward adjustment value varies depending on
interactions between numerous plant and site-specific factors, staff chose to present estimates  of
yield and biomass loss for each crop and tree seedling species,  respectively, as a range, with non-
adjusted and 10%-adjusted air quality as the upper and lower bounds (See Chapter 7, section
7.6.2.3 for a detailed discussion).
       Seedling and Mature Tree Biomass Loss. Biomass loss in  sensitive tree seedlings is
predicted to occur under Os exposures that just meet the level of the current secondary standard
(see  Table 7F-5 in Appendix 7F). For instance, black cherry, ponderosa pine, eastern white pine,
and aspen had estimated median seedling biomass losses as high as 24, 11, 6, and 6%,
respectively, when air quality was rolled back to just meet the current 8-hr standard with the 10%
adjustment applied. Staff notes that these results are for tree seedlings and that mature trees of
                                         8-10

-------
the same species may have more or less of a response to 63 exposure. Due to the potential for
compounding effects over multiple years, a consensus workshop on Os effects reported that a
biomass loss greater than 2% annually can be significant (Heck and Cowling, 1997). Decreased
seedling root growth and survivability could affect overall  stand health and composition in the
long term.
       Our analysis using modeled mature tree growth response under different air quality
scenarios for a western species (ponderosa pine) and two eastern species (red maple and tulip
poplar) projected that just meeting the current standard would likely continue to allow Os-related
reductions in annual net biomass gain in these species (see Table 7-5 in Chapter 7). This
judgment is based, in part, on model outputs that estimate that as Os levels are reduced below
those of the current standard, significant improvements in growth would occur.  For instance,
estimated growth in red maple increased by 4% and 3% at Big Meadows and Cranberry sites,
respectively, when air quality was rolled back to just met a SUM06 value of 15 ppm-hr
(approximately equivalent to a W126 value of 13 ppm-hr). Yellow poplar was projected to have
a growth increase between 0.6 and 8% under the  same scenarios at the two sites.
       Though there is significant uncertainty associated with this analysis, we judge that this
information  should be given careful consideration in light of several other pieces of evidence.
Specifically, limited evidence from experimental studies that goes beyond the seedling growth
stage continues to show decreased growth under elevated Os (King et al. 2005).  Some mature
trees such as red oak have shown an even greater sensitivity of photosynthesis to 63 than
seedlings of the same species (Hanson et al., 1994). As indicated above, smaller growth loss
increments may be significant for perennial species. The potential for cumulative "carry over"
effects as well as compounding must be considered. The accumulation of such "carry-over"
effects over time may affect long-term survival and reproduction of individuals and ultimately
the abundance of sensitive tree species in forest stands.
       Crop Yield Loss.  Staff exposure and risk assessments estimate that just meeting the
current 8-hr standard would still allow Os-related yield loss to occur in several fruit and
vegetable species and major commodity crop species  currently grown in the U.S. (see Table 7F-4
in Appendix 7F). These estimates are substantially lower than those estimated in the last review
as a result of several factors. First, Os air quality has improved in many areas of the country
since the last review. Secondly, staff has factored in an 63 adjustment for the height gradient, as
described above, and will present results for both non-adjusted and adjusted exposure levels to
approximate upper and lower bounds of predicted yield loss.
       Several sources of uncertainty should be taken into account when evaluating the
significance of these findings. First, yield loss estimates were generated using the median C-R
function when more than one function was available for a given species.  For some species,
                                         8-11

-------
however, only one C-R function was available. In this latter case, there is more uncertainty
regarding the range of variability in Os sensitivity within each crop.  Secondly, six of the fruit
and vegetable species were not part of the NCLAN program and C-R functions were available
only in terms of seasonal 7-hr or  12-hr mean indices.  These indices are considered less effective
in predicting plant response for a given change in air quality than cumulative forms used with
other crops. Therefore, staff places less weight on the fruit and vegetable yield loss numbers
than those for commodity crops, even though the magnitude of the fruit and vegetable effect was
much greater.  Finally, staff recognize that agricultural systems are heavily managed  and
vulnerable to adverse impacts from a variety of other factors (e.g., weather, insects, disease),
which can overshadow the magnitude of yield impacts predicted for a given Os exposure.
However, it should also be recognized that, in some experimental cases, exposure to  Os has made
plants more sensitive or vulnerable to other important stressors such as disease, insect pests, and
harsh weather (U.S. EPA, 2006).  Due to the significant impact these  other stressors can have on
crop production in some areas, staff recommends that additional research be done to  better
understand the nature and significance of these interactive effects of O3 with other plant
stressors.
       Keeping these uncertainties in mind, the results of the risk assessment show that when air
quality is rolled back to just meet the current standard, yield loss is still estimated to occur in
several fruit and vegetable species and major commodity crop species currently grown in the
U.S. (see 7.6.2.4 of Chapter 7). For example, based on median C-R function response, in
counties with the highest Os levels, potatoes and cotton had estimated yield losses of 9-15% and
5-10%, respectively,  when air quality just met the level of the current standard. Estimated yield
improved in these counties when the alternative SUM06 and W126 standard levels were met.
The very important soybean crop had generally small yield losses throughout the country under
current air quality (0-6%) and just meeting the current standard (0-4%).
       Another group of crops, multiple year forage crops, have also received additional study
since the last review. Based on these new studies, the yields and quality of multiple-year forage
crops have  also been shown to be sufficiently reduced as to have nutritional and  possibly
economic implications for their use as ruminant animal feed at Os exposures that occur in some
years over large areas of the U.S. However, it is not clear at this time to what degree they are
impacted at lower levels of air quality, since the studies were not designed  to address this
question.
       Summary
       In summary, O?, levels that would be expected to remain after  meeting the level of the
current secondary standard are sufficient to cause visible foliar injury, seedling and mature tree
growth, and reduce crop yields.  Other Os-induced effects described in the literature include an

                                         8-12

-------
impaired ability of many sensitive species and genotypes within species to adapt to or withstand
other environmental stresses such as freezing temperatures, pest infestations and/or root disease,
and reduced ability to compete for available resources. In the long run, the result of these
impairments (e.g., loss in vigor) may be premature plant death.  Though effects on other
ecosystem components have not been examined, except in isolated cases,  effects such as those
described above could have significant implications for plant community and associated species
biodiversity and the structure and function of whole ecosystems (Young and Sanzone, 2002).


       8.3.1.2 CASAC and Public Commenter Views on the Adequacy of the Current
              Standard
       Staff recognizes that the exposure-and risk-based information can be considered both in
terms of whether the risks estimated to remain upon attaining the current standard are important
from a public welfare perspective and/or whether additional reductions in risk estimated to be
associated with alternative, more protective standards are also important from  a public welfare
perspective. Judgments about the importance of the estimates of exposure and risks need to take
into account the important uncertainties associated with such estimates.
       There is general recognition among staff, CASAC, and all interested parties that public
welfare policy judgments, including the weight to place on various types of evidence and how to
weigh the importance of estimated risks in a public welfare perspective, are ultimately decisions
left to the Administrator. To help inform those judgments with regard to the adequacy of the
current secondary Os standard, the views  expressed by CASAC as well as the views of other
interested parties who have commented on earlier drafts of this document are summarized here.
The range of views generally reflects differing judgments as to the relative weight to place on
various types of exposure- and risk-based information, and the associated uncertainties, as well
as differing judgments about the importance of various (Vrelated vegetation effects from a
public welfare perspective.
       In a letter to the Administrator (Henderson, 2006c), the CASAC O3 Panel, with full
endorsement of the chartered CASAC,  unanimously concluded that "despite limited recent
research, it has become clear since the last review that adverse effects on a wide range of
vegetation including visible foliar injury are to be expected and have been observed in areas that
are below the level of the current 8-hour primary and  secondary ozone standards..."  Therefore,
"based on the Ozone Panel's review of Chapters 7 and 8, the CASAC unanimously agrees that it
is not appropriate to try to protect vegetation from the substantial, known or anticipated, direct
and/or  indirect, adverse effects of ambient ozone by continuing to promulgate  identical primary
and secondary standards for ozone. Moreover, the members of the Committee and a substantial
                                         8-13

-------
majority of the Ozone Panel agrees with EPA staff conclusions and encourages the
Administrator to establish an alternative cumulative secondary standard for ozone and related
photochemical oxidants that is distinctly different in averaging time, form and level from the
currently existing or potentially revised 8-hour primary standard" (Henderson, 2006c).
       In contrast to the views of CASAC discussed above, others submitted comments that
supported retaining the current standards.1 In considering the available evidence as a basis for
their views, these commenters identified a number of key concerns that, in their view, make it
inappropriate to revise the secondary standard at this time.  For example, they assert:  1) The key
uncertainties cited by the Administrator in the 1997 review as reasons for deciding it was not
appropriate to move forward with a seasonal secondary, (e.g., uncertainties in the exposure, risk
and valuation analyses and the lack of air  quality data in rural and remote areas), have not been
materially reduced in this current review;  and 2) The exposure assessment is inaccurate and too
uncertain due to the use of low estimates of policy-relevant background (PRB), an arbitrary
rollback method that is uninformed by atmospheric chemistry from photochemical models, and
the use of the CMAQ model in the west, whose biases and uncertainties are insufficiently
characterized and evaluated.

       8.3.1.3 Staff Conclusions on the Adequacy of the Current Standard
       On the basis of the vegetation effects that have been observed to still occur under current
ambient exposure conditions and those predicted to occur under the scenario of just meeting the
current secondary NAAQS, staff concludes  that the current secondary NAAQS is inadequate to
protect the public welfare from known and anticipated adverse welfare  effects.  As discussed
above, this conclusion derives from several  lines of evidence.
       First, visible foliar injury observations for the years 2001 to 2004 at USDA FIA
biomonitoring sites show widespread Os-induced effects occurring in the field, including in
forested ecosystems. For a  few studied species, it has been further shown that the presence of
visible foliar injury is linked to the presence of other vegetation effects  (e.g., reduced plant
growth and impaired below ground root development) (U.S. EPA, 2006), though for most
species, making this linkage remains problematic.  Nevertheless, when  visible foliar injury is
present, the possibility that  other Os-induced vegetation effects could also be present should be
considered. Staff recognizes that it is not  possible at this time to quantitatively assess the degree
of visible foliar injury that should be judged adverse in all settings and across all species, and
that other environmental factors can mitigate or exacerbate the degree of Os-induced visible
foliar injury expressed at any given concentration of Os. However, recognizing that the presence
        This group of commenters included industry associations, corporations, and individuals.
                                         8-14

-------
of visible foliar injury alone can be adverse to the public welfare (see foliar injury discussion in
8.3.1.1), and on the basis of the above considerations, staff concludes that the current standard
continues to allow levels of visible foliar injury that could reasonably be considered to be
adverse from a public welfare perspective.
       Second, a recent ambient gradient study and a free air Os enrichment (FACE) experiment
have supported earlier findings from 63 experiments conducted in OTC. Studies conducted at
the AspenFACE site in  Wisconsin (see section 7.6.2.1 of chapter 7) on quaking aspen has
confirmed the detrimental effects of Os exposure on tree growth in a field study without
chambers (Isebrands et  al., 2000, 2001). The recent ambient gradient study (Gregg et al, 2003)
evaluated biomass loss  in cottonwood along an urban-to-rural gradient at several  locations.
Study results found that conditions in the field were sufficient to produce substantial biomass
loss in cottonwood, with larger impacts observed in downwind rural areas due to  the presence of
higher Os concentrations (See Section 7.6.3). Staffs inclusion and emphasis on these two field-
based lines of evidence  is consistent with the Administrator's conclusion at the end of the last
review (62 FR 38877-78), that continued research on the effects of Oj on vegetation  under field
conditions would be important in this next review.  Staff feels that the expanded field-based
evidence provides qualitative support for the continued usefulness of findings obtained from
chamber studies.
       Staffs conclusion is further strengthened by evidence of remaining impacts on tree
seedling biomass loss when the current 8-hr standard is met. Staff estimated annual biomass loss
up to 6-24% for some sensitive species in areas of high Os exposure. Because of the potential
for indirect effects on plant vigor from even small incremental biomass or growth reductions in
the field, staff observes  that these levels of tree seedling growth reduction are well above the 1-
2% range of concern identified by the 1997 consensus workshop (Heck and Cowling, 1997).
Staff also took into account modeled mature tree growth loss estimates and commodity crop and
fruit and vegetable yield loss in arriving at these conclusions. Linkages across ecosystem
hierarchies (Young and Sanzone, 2002) make indirect impacts to ecosystems another welfare
effects category of concern even after attaining the  current secondary standard.

       8.3.2   Pollutant Indicator
       The staff concludes that Os remains the appropriate pollutant indicator for use in a
secondary NAAQS that provides protection for public welfare from exposure to all
photochemical oxidants. This conclusion is based on the same rationale presented in the
previous Staff Paper (U.S. EPA, 1996b), which recognizes that among the other photochemical
oxidants, the database for vegetation effects only raises concern at levels found in the ambient air
                                         8-15

-------
for 63 and, therefore, control of ambient 63 levels provides the best means of controlling other
photochemical oxidants of potential welfare concern.
       8.3.3  Averaging Times
       Plants, unlike people, are exposed to ambient air 24 hours a day, every day for their entire
life.  For annual species, this is for only a period within one year; for perennials, exposures are
for multiple years, decades or centuries. Regardless of plant type, it has been well established in
the literature that 63 effects are cumulative, and that longer exposure durations have a greater
impact than shorter durations, all else being equal (U.S. EPA, 2006). Air quality indices that
account for the exposure duration overall do a better job predicting plant response than short- or
long-term averages.  However,  63 levels are not continuously elevated and plants are not equally
sensitive to Os over the course of a day, season or lifetime. Thus, it becomes necessary to
identify periods of exposure that have the most relevance for plant response.

       8.3.3.1 Seasonal Window
       Many recent studies described in the 2006 CD have specifically selected exposure indices
that take into account the cumulative, concentration-weighted impact of Os-induced effects
throughout the growing season when measuring growth and yield impacts and have substantiated
the 1996 CD and 1996 Staff Paper conclusions on the importance of cumulative, seasonal
exposures (U.S. EPA, 2006). Annual crops are typically grown for periods  of two to three
months before being harvested. In contrast, perennial species may be photosynthetically active
longer (up to 12 months each year for a few species) depending on the species and where it is
grown.  In general, the period of maximum physiological activity and thus, potential Os uptake
for annual crops, herbaceous species, and deciduous trees and shrubs coincides with some or all
of the intra-annual period defined as the Os season, which varies on a state-by-state basis. This
is because the high temperature and high light conditions that promote the formation of
tropospheric 63 also promote physiological activity in vegetation.
       In the 1996 Staff Paper  and proposal notice, we noted that the selection of any single
averaging time for a national standard would represent a compromise, given the significant
variability in growth patterns and lengths of growing seasons among the wide range of
vegetation species that may experience adverse effects associated with Os exposure. However,
we concluded, based on the information available at that time, that selection of the maximum
consecutive 3-month period within the  Os season was reasonable, and in most cases, would most
likely coincide with the periods of greatest plant sensitivity on an annual basis. Based on the
information assessed in the current CD (U.S. EPA, 2006) and Chapter 7 above, we again
conclude the maximum consecutive 3-month period within the Os season is  a reasonable
averaging time for vegetation.
                                         8-16

-------
       8.3.3.2 Diurnal Window
       Stomata are the entry points for O3 into plant leaves.  Over the course of a day, plant
stomatal conductance varies along with light level, soil moisture and other factors. In general,
stomata are most open during daylight hours in order to allow sufficient CC>2 uptake for use in
carbohydrate production through the light driven process of photosynthesis. At most locations,
Os concentrations are also highest during the daytime, potentially coinciding with maximum
stomatal uptake. Ozone uptake during daylight hours impairs the light-driven process of
photosynthesis, which can then lead to impacts on carbohydrate production, plant growth,
reproduction (yield) and root function. Thus, in the last review, staff selected the 12-hr daylight
window (8 am to 8 pm) to capture the diurnal window with most relevance to the photosynthetic
process. Since that time, some limited work has been done by Musselman and Minnick (2000)
to more fully characterize O3 uptake at night and its potential contribution to total plant uptake
and response.  This work reports that some species do take up O3 at night, but that the degree of
nocturnal stomatal conductance varies widely between species and its relevance to overall O3-
induced vegetation effects remain unclear. We conclude that such information continues to be
preliminary and not generalizable at this time (see also Appendix 7A of Chapter 7). Staff,
therefore, again concludes that the daytime 12-hr window is the most appropriate period over
which to cumulate diurnal O3 exposures, specifically those most relevant to plant growth and
yield responses.

       8.3.3.3 Alternative Views and Staff Conclusions
       The CAS AC expressed views in agreement with staff with respect to both seasonal and
diurnal averaging times. Specifically, CASAC states "the suggested approach to the secondary
standard is a cumulative seasonal growing standard such as the indices SUM06 or W126
aggregated over at least the three summer months exhibiting the highest cumulative ozone levels
and includes the ozone exposures from at least 12 daylight hours."
       In contrast, some commenters  pointed to new information on nocturnal conductance as
evidence for the need for a 24-hour  diurnal window. Specifically, they  state "an extensive
review of the literature reported that a large number of species had varying degrees of nocturnal
stomatal conductance." Based on this review, Musselman and Minnick (2000) recommend that
any O3 exposure index used to relate air quality to plant response should use the 24-hour
cumulative exposure period.  No commenters addressed the adequacy of the three month
seasonal window. In examining the available information on nocturnal conductance (See
Appendix  7A), staff concludes that it remains unclear to what extent nocturnal uptake contributes
to the vegetation effects of yield loss, biomass loss or visible foliar injury. Due to the many
species- and site-specific variables that influence the potential for and significance of nocturnal
                                         8-17

-------
uptake, staff concludes that additional research needs to be done before considering whether this
component of vegetation exposure should be addressed with a different averaging time.
       Based on these considerations, as well as information assessed in the current CD (U.S.
EPA, 2006) and Chapter 7 above, we again conclude that a 12-hr (8:00 am to 8:00 pm) diurnal
window remains appropriate for a secondary NAAQS designed to protect a wide range of
vegetation growing in environmental conditions found across the U.S.

       8.3.4   Form of the Standard
       The 2006 O3 CD states, "In the 1996 O3 CD..., it was concluded, based on the best
available data, that those O3 exposure indices that cumulated differentially weighted hourly
concentrations were the best candidates for relating exposure to plant growth response.... The
few studies that have been published since the 1996 O3 CD continue to support the earlier
conclusions..." (U.S. EPA, 2006, pg. 9-12). The following selections taken from the 1996 CD
(see U.S. EPA, 1996a, pgs. 5-88/89, 5-95/96), further elucidate the depth and strength of these
conclusions. "When O3 effects are the primary cause of variation in plant response, plants from
replicate studies of varying duration showed greater reductions in yield or growth when exposed
for the longer duration." "The mean exposure index of unspecified duration could not account
for the year-to-year variation in response."  "Because the mean exposure index treats all
concentrations equally and does not specifically include an exposure duration component, the
use of a mean exposure index for characterizing plant exposures appears inappropriate for
relating exposure with vegetation effects"
       Though the scientific justification for a cumulative, seasonal form was generally accepted
in the last review, an analysis undertaken by EPA at that time had showed that there was
considerable overlap between areas that would be expected not to meet the range of alternative
8-hr  standards being considered for the primary NAAQS and those expected not to meet the
range of values (expressed in terms of the seasonal SUM06 index) of concern for vegetation.
This result suggested that improvements in national air quality expected to result from attaining
an 8-hr primary standard within the recommended range of levels would also be expected to
reduce levels below those of concern for vegetation in those same areas. Thus, in the proposal
notice, the Administrator proposed two alternatives for consideration:  one alternative was to
make the secondary  standard equal in every way to the proposed 8-hr, 0.08 ppm primary
standard; and the second was to establish a 3-month, 12-hr SUM06 seasonal secondary standard
(set at a level of 25 ppm-hr) as also appropriate to protect public welfare from known or
anticipated adverse effects given the available scientific knowledge and that such a seasonal
standard ".. .is more biologically relevant..." (61 FR 65716). In the 1997 final rule, the decision
was made to make the secondary identical  to the primary standard. It acknowledged, however,
                                         8-18

-------
that "it remained uncertain as to the extent to which air quality improvements designed to reduce
8-hr average Os concentrations averaged over a 3-year period would reduce Os exposures
measured by a seasonal SUM06 index." (62 FR 38876) In other words, it was uncertain as to
whether the 8-hr average form would, in practice, provide sufficient protection for vegetation
from the seasonal, cumulative and concentration-weighted exposures described in the scientific
literature as of concern.
       On the basis of that history, Chapters 2 and 7 of this Staff Paper revisited the issue of the
appropriateness of using an 8-hr average standard form to provide the requisite protection
required for vegetation.

       8.3.4.1 Comparison of 8-Hour Average and Cumulative Seasonal Forms
       Staff performed an analysis to evaluate the extent to which there appears to be a
relationship between county level air quality measured in terms of the current 8-hr average form
and that measured in terms of an alternative cumulative, seasonal form  (e.g. 12-hr W126).  Staff
determined it was most useful to begin by comparing the 3-year averages of each form, since the
current 8-hr average secondary form is a 3-year average. However, in recognition that some
vegetation effects (e.g. crop yield and foliar injury) are driven solely by annual O3 exposures,
and that typically the cumulative forms are  defined in terms of the annual growing season, staff
also performed a comparison of the current 8-hr form to the annual W126 air quality values for
both 2002 and 2004 (see Appendix 7B).
       Staff performed this analysis using recent (2002-2004) county-level air quality data from
AQS sites and the subset of CASTNET sites having the highest 63 levels for the counties in
which they  are located. Due to the lack  of more complete monitor  coverage in many rural areas,
staff acknowledges that this analysis may not be an accurate reflection of the situation in non-
monitored,  rural counties. Results of the 3-year average comparison showed that after meeting
the current  3-year average form of the 0.08-ppm, 8-hr average standard, only a few counties
showed 3-year average W126 values above the upper level (21 ppm-hr) evaluated (see Figure 7-
1).  This result, taken alone, might suggest that areas that met the current level and form would
typically overlap with the areas that met the analyzed alternative cumulative level and form.
However, at the lower W126 level of 13 ppm-hr (see discussion on level in section 8.3.5 below),
many more counties that meet the current 8-hr standard level and form  no longer meet the
alternative W126 form at this level. When individual years are compared, this  lack of a
relationship becomes clearer. For example, the relatively high 2002 air quality year, showed a
greater degree of overlap between those areas that would meet the levels analyzed for the current
8-hr and alternative W126 forms than did the relatively low 2004 air quality year (See Appendix
7B). It is clear from this analysis that the degree to which the current 8-hr standard form and
                                         8-19

-------
level would overlap with areas of concern for vegetation expressed in terms of the 12-hr W126
standard is inconsistent and would depend greatly on the level of the 12-hr W126 and 8-hr forms
selected and the distribution of hourly O?, concentrations within the annual and/or 3 year average
period.  It is not clear how this relationship would change due to the change in 63 patterns
resulting from control strategies put in place to attain different levels of standards.
       This view is consistent with those of CASAC who unanimously agree that it is not
appropriate to try to protect vegetation from the substantial, known or anticipated, direct and/or
indirect, adverse effects of ambient ozone by continuing to promulgate identical primary and
secondary standards for ozone.  Moreover, the members of CASAC and a substantial majority of
the CASAC Os Panel agree with staff conclusions and encourage the Administrator to establish
an alternative cumulative  secondary standard for ozone  and related photochemical oxidants that
is distinctly different in averaging time, form, and level  from the currently existing or potentially
revised 8-hour primary  standard.  The suggested approach to the secondary standard is a
cumulative seasonal growing standard such as the indices SUM06 or W126 aggregated over at
least the three summer months exhibiting the highest cumulative Oj levels and includes the Oj
exposures from at least  12 daylight hours (Henderson, 2006c).
       Some  other public commenters agreed that "direct!onally a cumulative form of the
standard may better match the underlying data." However, they believe further work is needed
to determine whether a  cumulative exposure index for the form  of the secondary standard is
needed. Specifically, a few commenters were of the view that a W126 (or SUM06) was not
sufficient in and of itself but should be combined with a measure of the number of peaks above
100 ppb (N100). Some of these same commenters also  felt a  24-hr averaging time was
supported by the data on nocturnal stomatal conductance.
       Staff recognizes that the relationship between Os exposure and plant response is more
complex than described by the single component indices and  supports the need for further
research into improvements on such indices to better capture factors that influence flux.  Staff
also recognizes that meeting the current 8-hr standard would result in air quality improvements
that could potentially benefit vegetation in some areas.  However, at this time, based on the
weight of evidence in the  scientific literature demonstrating the  cumulative nature of Os-induced
plant effects and the need to give greater weight to higher concentrations,  the advice of CASAC
consistent with this view,  and the results of the above analysis, staff again concludes that a
secondary standard should, at a minimum, be defined in terms of a form that reflects the two
components of exposure known to influence plant response, i.e. differentially weighted peak
concentrations and cumulative seasonal exposures.  Further, staff suggests caution should be
used in evaluating the likely vegetation impacts associated with a given level of air quality
expressed in terms of the 8-hr form in the absence of parallel  SUM06 or W126 information.
                                         8-20

-------
       Selecting a more biologically-relevant secondary standard form would also:  1) be directly
responsive to the recommendation of the 2004 National Research Council's report titled Air
Quality Management in the United States (NRC, 2004) which encourages the Agency to evaluate
its historic practice of setting the secondary NAAQS equal to the primary; 2) provide support to
important new Agency initiatives to enhance ecosystem-related program tracking and
accountability; and 3) potentially spur more policy relevant vegetation effects research in the
future.

       8.3.4.2 Comparison of SUM06 and W126 Cumulative, Concentration-Weighted
              Forms
       In addition to evaluating the 8-hr average form, we evaluated the appropriateness of the
SUM06 alternative proposed in the last review by comparing it to another cumulative,
concentration-weighted form discussed in the 1996 Staff Paper, the W126 index.  In the 1996
Staff Paper, the preference for the SUM06 over  other cumulative forms was based on the
following science and policy considerations:

      •   All  cumulative, peak-weighted exposure indices considered, including W126 and
         SUM06, were about equally good as exposure measures to predict exposure-response
         relationships reported in the NCLAN crop studies.

      •   The SUM06 form would not be influenced by PRB Os concentrations (defined at the
         time as 0.03 to 0.05 ppm) under many typical air quality distributions.
       In the current review, we have reconsidered whether the SUM06 form should still be
judged the most appropriate cumulative form for a secondary NAAQS protective  of vegetation
and ecosystems, based on the following:

      •   Model predictions of PRB in the range of 0.015 to 0.035 ppm for the current review are
         below the PRB range of 0.03 to 0.05 ppm described in the 1996 review, making PRB
         contributions much less of a factor influencing the choice of an appropriate cumulative
         index.

      •   There is no evidence in the extensive vegetation effects literature of a biological
         exposure threshold applicable across the broad array of Os-sensitive species found
         growing in the U.S. The SUM06 index, with a threshold set at 0.06 ppm,  artificially
         truncates exposures that have been shown to produce vegetation effects  of concern
         given sufficient duration.  The W126 index, on the other hand, cumulates  all O?,
         concentrations. However, because concentrations below 0.04 ppm, receive
         substantially smaller weights (3 percent or less), those concentrations within the range
         of PRB levels would not contribute significantly to the value of the index.
       The CASAC Ozone Panel also views the 3-month growing season W126 index ".. .as a
potentially more biologically-relevant index than the 3-month growing season SUM06 index.
This is because the W126 index has no absolute minimum ozone concentration threshold and
                                         8-21

-------
only lightly weights the lower ozone concentrations. Therefore, a three-month seasonal W126
range that is the approximate equivalent of the SUM06 range of 10 to 20 ppm-hr is preferred"
(Henderson, 2006c).
       On the basis of the information highlighted above, staff concludes that the W126 form is
a more appropriate biologically based and policy-relevant cumulative, concentration-weighted
form and recommends the Administrator consider the W126 as a more appropriate form for the
secondary standard. This recommendation is consistent with the views of CASAC.  Given the
legitimate policy interest in having a more stable standard form, the Administrator may want to
give consideration to using a 3-year average of the 12-hr W126, in addition to consideration of
an annual form.

       8.3.5  Level of the Standard
       The level at which a secondary standard should be set depends on a blending of science
and policy judgments by the Administrator as to the level of air quality which is requisite to
protect the public welfare from any known or anticipated adverse effects associated with the
pollutant in the ambient air. The exposure and risk assessments conducted in Chapter 7 and
summarized briefly above, provide information regarding the effects associated with a number of
different welfare endpoints at different levels of air quality, often expressed in terms of both the
current 8-hr average form and the W126 (or SUM06) seasonal  form(s).
       At the end of the last review, we identified  a range for a 3-month, 12-hr SUM06 standard
form of 25 to 38 ppm-hr, for the Administrator's consideration. These levels were estimated to
allow 10% to 20% yield loss, respectively, to occur in no more than 50% of the studied NCLAN
agricultural crop cases. These levels were also estimated to provide an increased level of
protection for other categories of vegetation such as tree seedlings and mature trees  in
commercial, Class I, and other forested areas in urban, rural, and remote environments. It was
recognized, however, that a standard set within this range would not protect the most sensitive
species or individuals within a species from  all potential effects related to 63  exposures. The
Administrator proposed the lower end of the range (e.g., 25 ppm-hr) as necessary to provide a
requisite level of protection for vegetation against the adverse effects  of Os.  Staff believes that
this level is an appropriate upper bound for a range of levels recommended for consideration in
this review, as it would continue to provide a level of crop and  tree protection judged requisite by
the Administrator in the last review.  In addition, this level derives from the extensive and
quantitative historic and recent crop effects database, as well as current staff exposure and risk
analyses.
       In identifying a lower bound for the range of alternative standard levels appropriate for
consideration, staff concludes that several lines of evidence point to the need  for greater
                                         8-22

-------
protection for tree seedlings, mature trees, and associated forested ecosystems. Tree growth is an
important endpoint to consider because it can be related to other aspects of societal welfare such
as sustainable production of timber and related goods, recreation, and carbon (CC^)
sequestration.  Equally important, impacts on tree growth can also affect ecosystems through
shifts in species composition and the loss of genetic diversity due to the loss of Os sensitive
individuals or species. To help inform staff judgment about an appropriate level of protection to
consider for trees, staff considered the results of a consensus-building workshop on the need for
a long-term cumulative secondary Os standard.  At this workshop, expert scientists expressed
their judgments on what standard form(s) and level(s) would provide vegetation with adequate
protection from Os-related adverse effects.  Consensus was reached with respect to selecting
appropriate ranges of levels in terms of a 3-month,  12-hr SUM06 standard for a number of
vegetation effects endpoints. These ranges are identified below, with the estimated approximate
equivalent W126 shown in parentheses (See Appendix 7B for explanation of SUM06 to W126
equivalents).  For yield reductions in agricultural crops - a range of 15 to 20 (13 to 17) ppm-hr;
for growth  effects to tree seedlings in natural forest stands - a range of 10 to 15 (7 to 13) ppm-hr;
for growth  effects to tree seedlings and saplings in plantations - a range of 12 to 16 (9 to 14)
ppm-hr; and for visible foliar injury to natural ecosystems - a SUM06 range of 8 to 12 (5 to 9)
ppm-hr (Heck and Cowling, 1997). In the 1997 final rule, the Administrator had pointed to the
results of this workshop  as providing important support to her view that the then current
secondary standard was not adequately protective of vegetation, contributing to her rationale that
revision of the secondary standard was needed (62 FR 38877)
       In its October 24, 2006 letter to the Administrator, CASAC expressed its view regarding
the appropriate form and range of levels for the Administrator to consider. The CASAC
preferred a seasonal 3-month W126 standard in a range that is the approximate equivalent of the
SUM06 at  10 to 20 ppm-hrs. Staff has determined  that the approximate equivalent 3-month
W126 range is 7 to 17 ppm-hrs.  The lower end of this range (7 ppm-hr) is the same as the lower
end of the range identified in the 1997 Consensus Workshop as protective of tree seedlings in
natural forest stands from growth effects (Heck and Cowling, 1997).
       Staff believes that Os-related effects on forest tree species are an important public welfare
effect of concern.  Therefore staff concludes that it is appropriate to include as the lower bound
of the recommended range,  the lower end of the approximate range recommended by CASAC
(Henderson, 2006c).  Based on our analyses of risks of tree seedling biomass loss and mature
tree growth reductions and on the basis of the scientific effects literature, we anticipate that the
lower end of the range identified for the Administrator's consideration would  substantially
decrease the adverse effects of Os on forested ecosystems. Additionally, it is anticipated that the
lower end of this range would provide increased protection from the more subtle impacts of 63

-------
acting in synergy with other natural and man-made stressors to adversely affect individual plants,
populations and whole systems. By disrupting the photosynthetic process, decreasing carbon
storage in the roots, increasing early senescence of leaves and affecting water use efficiency in
trees, 63 exposure could potentially disrupt or change the nutrient and water flow of an entire
system.  Weakened trees can become more susceptible to other environmental stresses such as
pest and pathogen outbreaks or harsh weather conditions. Though it is not possible to quantify
all the ecological and societal benefits associated with varying levels of alternative secondary
standards, we conclude that this information should be weighed in considering the extent to
which a secondary standard should be precautionary in nature in protecting against effects that
have not yet been adequately studied  and evaluated.
      Based on all the above considerations, staff concludes that a 3-month, 12-hr W126 range
of 21-7 ppm-hr is appropriate for consideration, with the upper bound  equivalent to that
proposed in the last review and with the lower bound being that recommended by CAS AC.
      In the absence of any information regarding a threshold of 63 exposures for vegetation,
staff recognizes that the level selected is largely a policy judgment as to the requisite level of
protection needed.  In determining the requisite level of protection for crops and trees, the
Administrator will  need to weigh the  importance of the predicted risks of these effects in the
overall context of public welfare protection, along with a determination as to the appropriate
weight to place on the associated uncertainties and limitations of this information.

      8.3.6   Summary of Staff Conclusions and Recommendations on the Secondary Os
              Standard
       Staff conclusions and recommendations on the elements of the secondary Os standard for
the Administrator's consideration in making decisions on the secondary 63 standard are
summarized below, together with supporting conclusions from sections 8.3.3 to 8.3.5 above. We
recognize that selecting from among alternative policy options will necessarily reflect
consideration of qualitative and quantitative uncertainties inherent in the relevant evidence and in
the assumptions of the quantitative exposure and risk assessments. Any such standard should
protect public welfare from any known or anticipated adverse effects associated with the
presence of the pollutant in the ambient air. In recommending these options for consideration,
we are mindful that the Act requires standards that, in the judgment of the Administrator, are
requisite to protect public welfare. The standards are to be neither more nor less stringent than
necessary.
      In the last review, the Administrator took into account the  following in reaching her final
decision: 1) the varying degrees of protection afforded by the alternative primary standards
recommended; 2) the incremental protection associated with alternative cumulative, seasonal
                                         8-24

-------
secondary standards under consideration; 3) the value of establishing a seasonal form for the
secondary standard that is more representative of biologically relevant exposures; and 4) the
extent to which a secondary standard should be precautionary in nature, given the possibility of
63 impacts acting in synergy with other natural and manmade stressors to impact climate and
other environmental endpoints, particularly given the potential significance at a regional scale
and in Class I areas.
       Staff notes that since the last review, several additional policy-relevant developments
have occurred that may also warrant consideration by the Administrator when making decisions
about what is requisite to protect public welfare.  First, the Agency has undertaken a number of
activities geared toward improving ecosystem-related program tracking and accountability and is
currently engaged in efforts to identify relevant indicators for that purpose.  Having a more
biologically-relevant air quality index would allow the Agency to better track improvements in
vegetation protection on the ground with specific program  actions aimed  at accomplishing that
end. Second, the NRC report (described above) states: "Whatever the reason that led EPA to use
identical primary and secondary NAAQS in the past, it is becoming increasingly evident that a
new approach will be needed in the future.  There is growing evidence that  the current forms of
the NAAQS are not providing adequate protection to sensitive ecosystems and crops" (NRC,
2004).
       The following secondary standard recommendations encompass the breadth of policy-
relevant considerations described above:

(1)    It is appropriate to continue to use Os as the indicator for a standard that is intended to
       address  effects associated with exposure to 63, alone or in combination with related
       photochemical oxidants. Based on the available information, we conclude that there is no
       basis for considering any alternative indicator at this time.

(2)    It is not  appropriate to continue to use an 8-hr averaging time for the secondary Os
       standard. The 8-hr average form should be replaced with a cumulative, seasonal,
       concentration weighted form.  Given the reasons stated in earlier discussions herein, staff
       concludes that the W126 form is more appropriate than the SUM06  form recommended
       in the last review.

(3)    It is appropriate to consider the maximum consecutive 3 month period within the Os
       season as the seasonal averaging time over which to cumulate hourly 63 exposures for
       the daily 12-hr daylight (8 am to 8 pm) window.  Though the length of time in the
       growing season varies significantly  between species, staff concludes that the 3-month
                                         8-25

-------
       period of maximum 63 exposure generally coincides with maximum biological activity
       for most plants. Staff notes that for certain welfare effects of concern (e.g., foliar injury,
       yield loss for annual crops, growth effects on other annual vegetation and potentially tree
       seedlings), an annual standard form may be more appropriate, while for other welfare
       effects (e.g., mature tree biomass loss), a 3-year average form may be more appropriate.
       Staff concludes it is appropriate to consider both the annual and 3-year average forms.

(4)     It is appropriate to consider a range of levels when making a determination regarding
       what is requisite public welfare protection. Staff concludes that an appropriate upper
       bound of this range is 21 ppm-hrs, expressed in terms of the W126 index, which is
       roughly equivalent to that proposed by the Administrator in the last review as able to
       provide a requisite level of protection to vegetation. Our analyses indicate that this level
       will provide protection against Os-related adverse impacts on vegetation such as tree
       growth and crop yield beyond that afforded by the current 8-hr standard. In large part,
       the basis for selecting the level in the last review was a judgment as to what was an
       appropriate level of protection against annual crop yield loss. Though crop data are still
       useful as a potential indicator of risk to other sensitive annual herbaceous plants, staff
       recognizes that agricultural  systems are heavily managed. In addition, the annual
       productivity of agricultural  systems is vulnerable to disruption from many other stressors
       (e.g., weather,  insects, disease),  whose impact in any given year can greatly outweigh the
       direct reduction in annual productivity resulting from elevated O3 exposures. On the
       other hand, O3 can also more subtly impact crop and forage nutritive quality and
       indirectly exacerbate the severity of the impact from other stressors. These latter effects
       cannot currently be quantified and deserve further study.  Taking all of the above
       considerations into account, staff concludes that from a public welfare perspective,
       greater concern should be placed on the impacts of O3 exposures on vegetation in less
       heavily managed and unmanaged ecosystems such as tree seedlings, mature trees, and
       forested ecosystems in general.  Thus, staff concludes that the lower end of the range
       should incorporate the lower end of the range expressed by  CASAC of a 3-month 12-hr
       W126 approximately equal to 7  ppm-hrs.  This lower level will increase protection for
       the most sensitive tree species and the ecosystems where  they are found.

       Several additional factors should be considered when selecting an appropriate level for a
       secondary standard. These include 1) the fact that O3 effects are cumulative and have
       been shown to have carry over effects from one year to the next; 2) some seedling tree
       species have sensitivities as great as annual crops and the importance of protecting
                                         8-26

-------
       against small percentages of biomass loss on an annual basis has been expressed by some
       within the scientific community; 3) visible foliar injury impacts can occur within a
       growing season at very low levels of O?, exposure; and 4) the extent to which a secondary
       standard should be precautionary in nature, given the possibility of 63 impacts acting in
       synergy with other natural and manmade stressors. Should a 3-year average of a 12-hr
       W126 be selected, the level chosen should reflect the fact that annual impacts are still  a
       concern for visible foliar injury, tree seedling biomass loss, and crop yield loss, so that a
       potentially lower level might be considered to reduce the potential of adverse impact
       from the single high 63 year that could still occur while attaining a 3-year average.

8.4    SUMMARY OF KEY UNCERTAINTIES AND RESEARCH
       RECOMMENDATIONS RELATED TO SETTING A SECONDARY O3
       STANDARD
       Staff has identified the following key uncertainties and research questions that have been
highlighted in Chapter 9 of the CD and Chapter 7 herein, associated with this review of the
welfare-based secondary standards. The first set of key uncertainties and research
recommendations discussed below is that associated with the extrapolation to plant species and
environments outside of specific experimental or field study conditions. The second set of key
uncertainties and research recommendations pertain to our ability to assess the impact of 63 on
other welfare effects categories such as climate, ecosystem components such as wildlife, and
whole ecosystem structure and function.  Third, we identify research recommendations related to
the development of approaches, tools, or methodologies useful in characterizing the relationship
between Os and public welfare in a policy context.  These three  areas are described below.
(1)    Plant Species-Level Research Needs:
     •   To reduce uncertainties associated with extrapolating plant response for a given level
         of Os using composite response functions across differing regions and climates, studies
         using large numbers of plant species across regions where those species are indigenous
         are recommended.  In addition, to better understand the full range of response of plant
         species to 63, research on more species is recommended.
     •   To reduce uncertainty associated with estimating the risk to vegetation of differing
         amounts of O3-induced visible foliar injury over the plant's leaf area, research to
         explore the relationship between visible foliar injury and other (Vrelated effects is
         recommended.

     •   To reduce uncertainty associated with the impact of differing levels of Os on the
         nutritive quality of forage and other crops, additional research is needed.
                                         8-27

-------
     •   To reduce uncertainty associated with estimated or modeled flux into plants, research is
         recommended to evaluate the factors that affect Os flux into plants, including the
         genetic determinants of O3 sensitivity and the range of variability among species with
         respect to detoxification/compensation and nocturnal uptake and response. Research
         that explores the relative importance of flux rate versus total cumulative flux or dose,
         and that leads to a database of Os flux-response relationships for vegetation; similar to
         the extensive concentration-response database that currently exists is recommended to
         further reduce existing uncertainties.

     •   To reduce uncertainties in extrapolating from Os effects on juvenile to mature trees and
         from trees grown in the open versus those in a closed forest canopy in a competitive
         environment, additional research is recommended.

     •   To reduce uncertainties in extrapolating individual plant response spatially or to higher
         levels of biological organization, including ecosystems, research that explores and
         better quantifies the nature of the relationship between 63, plant response and multiple
         biotic and abiotic stressors, including those associated with climate change, is
         recommended.

(2)     Ecosystem Level Impacts:

     •   To reduce uncertainties associated with projections of the effects of 63 on the
         ecosystem processes of water, carbon, and nutrient cycling, particularly at the stand
         and community levels, research is needed on the effects on below ground ecosystem
         processes in response to O3 exposure alone and in combination with other stressors.
         These below ground processes include interactions of roots with the soil or
         microorganisms, effects of Os on structural or functional components of soil food webs
         and potential impacts on plant species diversity, changes in the water use  of sensitive
         trees, and if the sensitive tree species is dominant, potential changes to the hydrologic
         cycle at the watershed and landscape level.

     •   To conclusively show whether Os affects biodiversity or genetic diversity, research on
         competitive interactions under elevated O3 levels are recommended. This research
         could be strengthened by modern molecular methods to quantify impacts  on diversity.

     •   To fill the data gaps regarding interactions and potential feedback mechanisms between
         Os and Os precursor (e.g., volatile organic carbons) production, atmospheric processes,
         and climate change variables, research is recommended to evaluate whether 63 will
         negate the positive effects of an elevated CC>2 environment on plant carbon and water
         balance, whether the likelihood of various biotic stressors such as pest epidemics and
         insect outbreaks would be expected to increase in the future

     •   To reduce uncertainties associated with scaling 63 effects up from the responses of
         single or a few plants to effects on communities and ecosystems, additional research is
         recommended.  Because these uncertainties  are multiple and significant due to the
         complex interactions involved, new research will likely require a combination of
         manipulative experiments with model ecosystems, community and ecosystem studies
                                         8-28

-------
         along natural 63 gradients, and extensive modeling efforts to project landscape-level,
         regional, national and international impacts of O^.

     •   To fill the data gaps regarding Os impacts to other non-plant welfare effects categories
         such as climate, as well as potential direct impacts of 63 on some sensitive species of
         animals and wildlife, more research is needed.

(3)     Approaches, Tools, Methodologies:

     •   To reduce uncertainties associated with valuing improved vegetation and ecosystem
         function from improved 63 air quality, research is needed on methodologies to
         determine the values associated with important services and benefits derived from
         natural ecosystems such that these could be used in comprehensive risk and benefits
         assessments for O3 effects on natural ecosystems.

     •   To reduce uncertainties associated with evaluating the performance of different
         exposure indices given different patterns of Os exposures, experiments would need to
         be designed to specifically test the performance of different indices in predicting plant
         response under different exposure regimes.

     •   To reduce uncertainties associated with the generation of rural Os exposures, improved
         model capabilities are needed, including a more refined spatial grid for the western
         U.S., better handling of 63 movement in complex terrain and predicting nocturnal
         concentrations. Further, research is needed regarding whether strategic placement of
         passive or mobile monitors might benefit the estimation  of impact to particular
         resources of concern.
                                         8-29

-------
REFERENCES

Abt Associates, Inc. (1993) Urban ornamental plants: sensitivity to ozone and potential economic losses. Report to
        U.S. EPA, December

Abt Associates Inc. (2007). Technical Report on Ozone Exposure, Risk, and Impacts Assessments for Vegetation:
        Final Report.  Prepared for Office of Air Quality Planning and Standards, U.S. Environmental Protection
        Agency, Research Triangle Park, NC. January 2007.  Available electronically on the internet at:
        http://www.epa.gOv/ttn/naaqs/standards/ozone/s o3crtd.html.

Federal Register (1996) National Ambient Air Quality Standards for Ozone; Proposed Rule. 40 CFR 50; Federal
        Register 61: 65716.

Federal Register (1997) National Ambient Air Quality Standards for Ozone; Final Rule. 40 CFR 50; Federal
        Register 62: 38856.

Gregg, J. W., Jones, C.G., Dawson, T.E. (2003). "Urbanization effects on tree growth in the vicinity of New York
        City." Nature 424: 183-187.

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 mbra L. foliage:
        differential sensitivity to ozone exposure." Tree Physiology 14:1351-1366.

Heck, W. W. and Cowling, E. B. (1997) The need for a long term cumulative secondary ozone standard - an
        ecological perspective. EM (January): 23-33

Henderson, R. (2006c) Letter from CASAC Chairman Rogene Henderson to EPA Administrator Stephen Johnson,
        October 24, 2006, EPA-CASAC-07-001.

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.

Isebrands, J. G.; Dickson, R. E.; Rebbeck, J.; Karnosky, 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.; Karnosky, D. F.
      (2001) Growth responses of Populus tremuloides clones to interacting carbon dioxide and tropospheric
      ozone. Environ. Pollut. 115: 359-371.

King, J.S., M. E. Kubiske, K. S. Pregitzer, G.  R. Hendrey, E. P. McDonald, C. P. Giardina, V. S. Quinn, D. F.
      Karnosky. (2005) Tropospheric O3 compromises net primary production in young stands of trembling
      aspen, paper birch and sugar maple in response to elevated atmospheric CO2.  New Phytologist.
      168:623-636.

Musselman, R. C.; Minnick, T. J. (2000) Nocturnal stomatal conductance and ambient air quality standards for
      ozone. Atmos. Environ. 34: 719-733.

NRC (2004). Air quality  management—United States. I. National Research Council (U.S.). Committee on Air
      Quality Management in the United States. TD883.2.A64325 2004 363.739'25'0973—dc222004014594
      http://www.nap.edu/openbook/0309089328/html/

National Park Service (2000). Federal Land Managers' Air Quality  Related Values Workgroup (FLAG) Phase I
        Report. (http://www2.nature.nps.gov/air/Pubs/pdf/flag/FlagFinal.pdf).
                                                8-30

-------
Rizzo, M (2005). Evaluation of a quadratic approach for adjusting distributions of hourly ozone concentrations to
        meet air quality standards. November 7, 2005.

Rizzo, M. (2006).  A distributional comparison between different rollback methodologies applied to ambient ozone
        concentrations. May 31,2006

U.S. Environmental Protection Agency (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.

U.S. Environmental Protection Agency (1996b). Review of National Ambient Air Quality Standards for Ozone:
        Assessment of Scientific and Technical Information - OAQPS Staff Paper.  EPA/452/R-96-007.  Office of
        Air Quality Planning and Standards, Research Triangle Park, NC.

U.S. Environmental Protection Agency (2006) Air Quality Criteria for Ozone and Related Photochemical Oxidants
        (Final). Office of Research and Development, National Center for Environmental Assessment, Research
        Triangle Park, NC.  EPA/600/R-05/004aF-cF, 2006.

Wolff, G.T. (1996) Letter from Chairman of Clean Air Scientific Advisory Committee to the EPA Administrator,
        dated April 4, 1996. EPA-SAB-CASAC-LTR-96-006.

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.pdft9December, 2003].
                                               8-31

-------
                                ATTACHMENT
                    UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
                                   WASHINGTON D.C. 20460
                                                               OFFICE OF THE ADMINISTRATOR
                                                                 SCIENCE ADVISORY BOARD
                                   October 24, 2006
EPA-CASAC-07-001

Honorable Stephen L. Johnson
Administrator
U.S. Environmental Protection Agency
1200 Pennsylvania Avenue, NW
Washington, DC 20460

       Subject:  Clean Air Scientific Advisory Committee's (CASAC) Peer Review of the
                Agency's 2nd Draft Ozone Staff Paper

Dear Administrator Johnson:

       EPA is in the process of reviewing the national ambient air quality standards (NAAQS)
for ozone (Os) and related photochemical oxidants, which the Agency most recently revised in
July 1997. As part of its ongoing review of the ozone NAAQS, EPA's Office of Air Quality
Planning and Standards (OAQPS) developed a 2nd Draft Ozone Staff Paper, entitled, Review of
the National Ambient Air Quality Standards for Ozone: Policy Assessment of Scientific and
Technical Information (July 2006).  At the request of the Agency, EPA's Clean Air Scientific
Advisory Committee (CASAC or Committee), supplemented by subject-matter-expert panelists
— collectively referred to as the CASAC Ozone Review Panel (Ozone Panel) — met in a public
meeting in Durham, NC, on August 24-25, 2006, to conduct a peer review of this draft Ozone
Staff Paper and three related draft technical support documents.

       In its summary of EPA staff conclusions on the primary (health-related) ozone NAAQS
found in Chapter 6 of the 2nd Draft Ozone Staff Paper, OAQPS set-forth two options with regard
to revising the level and the form of the standard: (1) retain the current primary eight-hour (8-hr)
NAAQS of 0.08 parts per million (ppm); or (2) consider a reduction in the level of the primary
Os NAAQS within the range of alternative 8-hr standards included in Staffs exposure and risk
assessments (which included a range from 0.064 to 0.084 ppm) with primary focus on an Os
level of 0.07 ppm with a range of forms from third- through fifth-highest daily maximum.  The
Ozone Panel unanimously concludes that:

    1.  There is no scientific justification for retaining the current primary 8-hr NAAQS of 0.08
       parts per million (ppm), and

-------
    2.  The primary 8-hr NAAQS needs to be substantially reduced to protect human health,
       particularly in sensitive subpopulations.

       Therefore, the CASAC unanimously recommends a range of 0.060 to 0.070ppm for the
primary ozone NAAQS. With regard to the secondary (welfare-related) ozone NAAQS, the
Ozone Panel is in strong agreement with the scientific and technical evidence presented in the
summary of EPA staff conclusions on the secondary ozone NAAQS found in Chapter 8 of the
draft Staff Paper in support of the alternative secondary standard of cumulative form that
extends over an entire growing season.

       The Ozone Panel members agree that this letter adequately represents their views.  The
chartered Clean Air Scientific Advisory Committee fully endorses the Panel's letter and hereby
forwards it to you as the Committee's consensus report on this subject. A discussion of each
chapter in the 2nd Draft Ozone Staff Paper follows this letter, and the comments of individual
Panel members on the 2nd Draft Ozone  Staff Paper and three related draft technical support
documents are attached as Appendix D.
1. Background

       Section 109(d)(l) of the CAA requires that the Agency periodically review and revise, as
appropriate, the air quality criteria and the NAAQS for the "criteria" air pollutants, including
ambient ozone. Pursuant to sections 108 and 109 of the Act, EPA is in the process of reviewing
the ozone NAAQS.  OAQPS, within the Office of Air and Radiation (OAR), developed the 2nd
Draft Ozone Staff Paper as part of this activity. In February 2006, the Agency's National Center
for Environmental Assessment, Research  Triangle Park, NC (NCEA-RTP), within the Agency's
Office of Research and Development (ORD), released its final Air Quality Criteria for Ozone
and Related Photochemical Oxidants, Volumes I, II, and III, (EPA/600/R-05/004aF-cF, Final
Ozone Air Quality Criteria Document) for this current review cycle for the ozone NAAQS.  The
2nd Draft Ozone Staff Paper evaluates the policy implications of the key scientific and technical
information contained in the Final Ozone  AQCD and identifies critical elements that the Agency
believes should be considered in its  review of the ozone NAAQS. The Ozone Staff Paper is
intended to "bridge the gap" between the scientific review contained in the Ozone AQCD and
the public health and welfare policy judgments required of the EPA Administrator in reviewing
the ozone NAAQS.

       The Ozone Panel met in a public meeting on December 8, 2005 to conduct a consultation
on EPA's 1st Draft Ozone Staff Paper and two related technical support documents. However,
given that the OAQPS' first draft Staff Paper did not contain Agency staff conclusions about
whether to retain or revise the existing primary and secondary Ozone standards, the CAS AC's
activity only amounted to a technical assessment of that document.  The Committee's letter to
you from that meeting (EPA-CASAC-CON-06-003), dated February 16, 2006, is posted at URL:
http://www.epa.gov/sab/pdf/casac con  06 003.pdf.

-------
2. CASAC Ozone Review Panel's Peer Review of the 2nd Draft Ozone Staff Paper and
Related Technical Support Documents

       The Ozone Panel reviewed the 2nd Draft Ozone Staff Paper and found it improved over
the earlier version that had been reviewed as part of a consultation process. However, the Panel
did not agree with the EPA staff conclusions that it was appropriate to consider retaining the
current NAAQS as an option that would be protective of public health and welfare.  The Ozone
Panel's recommendations for reducing the level of the primary ozone standard, and its rationale
for these recommendations, are provided immediately below. Following a detailed discussion on
the primary and secondary NAAQS are the Panel's major, chapter-specific comments. Finally,
the individual written comments of Ozone Panel members on the 2n Draft Ozone Staff Paper
and the three related draft technical support documents are attached in Appendix D. Panelists'
responses to the Agency's charge questions are included in these individual review comments.
                                 Primary Ozone NAAQS

       New evidence supports and build-upon key, health-related conclusions drawn in the 1997
Ozone NAAQS review.  Indeed, in the 2nd Draft Ozone Staff Paper, EPA staff themselves arrived
at this same conclusion:
       "Based on the above considerations and findings from the [Final Ozone AQCD], while being
   mindful of important remaining uncertainties, staff concludes that the newly available
   information generally reinforces our judgments about causal relationships between O3 exposure
   and respiratory effects observed in the last review and broadens the evidence of O3 -related
   associations to include additional respiratory-related endpoints, newly identified cardiovascular-
   related health endpoints, and mortality. Newly available evidence also has identified increased
   susceptibility in people with asthma. While recognizing that important uncertainties and research
   questions remain, we also conclude that progress has been made since the last review in
   advancing our understanding of potential mechanisms by which ambient O3, alone and in
   combination with other pollutants, is causally linked to a range of respiratory- and cardiovascular-
   related health endpoints." (Pages 6-6 and 6-7)

       Several new single-city studies and large multi-city studies designed specifically to
examine the effects of ozone and other pollutants on both morbidity and mortality have provided
more evidence for adverse health effects at concentrations lower than the current standard.  (See
the numerous ozone epidemiological single-city studies shown in Figure 3-4 on page 3-53 of the
2nd Draft Staff Paper and, in addition, Appendix 3B of the staff paper, which contains the
summary of effect estimates and air quality data for these studies and multi-city epidemiological
studies.) These studies are backed-up by evidence from controlled human exposure studies that
also  suggest that the current primary ozone NAAQS is not adequate to protect human health
(Adams, 2002; McDonnell, 1996).

       Furthermore, we have evidence from recently reported controlled clinical studies of
healthy adult human volunteers exposed for 6.6 hours to 0.08, 0.06,  or 0.04 ppm ozone,  or to
filtered air alone during moderate exercise (Adams, 2006).  Statistically-significant decrements
in lung function were observed at the 0.08 ppm exposure level.  Importantly, adverse lung
function effects were also observed in some individuals at 0.06 ppm (Adams, 2006). These

-------
results indicate that the current ozone standard of 0.08 ppm is not sufficiently health-protective
with an adequate margin of safety.  It should be noted these findings were observed in healthy
volunteers; similar studies in sensitive groups such as asthmatics have yet to be conducted.
However, people with asthma, and particularly children, have been found to be more sensitive
and to experience larger decrements in lung function in response to ozone exposures than would
healthy volunteers (Mortimer et a/., 2002).

       Going beyond spirometric decrements, adverse health effects due to low-concentration
exposure to ambient ozone (that is, below the current primary 8-hour NAAQS) found in the
broad range of epidemiologic and controlled exposure studies cited above include: an increase in
school absenteeism; increases in respiratory hospital emergency department visits among
asthmatics and patients with other respiratory diseases; an increase in hospitalizations for
respiratory illnesses; an increase in symptoms associated with adverse health effects, including
chest tightness and medication usage; and an increase in mortality (non-accidental,
cardiorespiratory deaths) reported at exposure levels well below the current standard. The
CASAC considers each of these findings to be an important indicator of adverse health effects.
As demonstrated in Chapter 5 of the 2nd Draft Ozone Staff Paper (specifically, Figures 5.5, 5.7,
5.8, and 5.9), a significant decrease in adverse effects due to ozone exposures can be achieved by
lowering the exposure concentrations below the current standard, which is effectively 0.084
ppm. Beneficial effects in terms of reduction of adverse health effects were calculated to occur
at the lowest concentration considered (i.e., 0.064 ppm). (See also Figure 3-4, "Effect estimates
(with 95% confidence intervals) for associations between short-term ozone exposure and
respiratory health outcomes," on page 3-53.)

       The justification provided in the 2nd Draft Ozone Staff Paper for retaining the current
level of the primary ozone standard as an option for the Administrator was based on results of
controlled human exposure studies measuring modest declines in FEVi after exposures to 0.08
ppm ozone. However, as stated in the Staff Paper (page 3-6), while average decrements in the
FEVi were relatively small, 26%  of the subjects had greater than 10% decrements, which can be
clinically significant.  Also, while measures of FEVi  are quantitative and readily obtainable in
humans, they are not the only measures — and perhaps not the most sensitive measures — of the
adverse health effects induced by ozone exposure.  As stated on page 6-32 of the Final Ozone
AQCD, "Spirometric responses to ozone are independent from inflammatory responses and
markers of epithelial injury (Balmes et a/., 1996; Bloomberg et a/., 1999; Hazucha etal., 1996;
Torres et a/., 1997). Significant inflammatory responses to ozone exposures that did not elicit
significant spirometric responses  have been reported  (Holz et a/., 2005; McBride et a/., 1994)."
Agency staffs analyses placed most emphasis on spirometric evidence and not enough emphasis
on serious morbidity (e.g., hospital admissions) and mortality observed in epidemiology studies
(see page 6-44).

       Therefore, on the basis of the large amount of recent data evaluating adverse health
effects at levels at and below the current NAAQS for ozone, it is the unanimous opinion of the
CASAC that the current primary ozone NAAQS is not adequate to protect human health.
Furthermore, the Ozone Panel is  in complete agreement both that: the EPA staff conclusion in
Section 6.3.6 arguing that "consideration could be given to retaining the current 8-hr ozone
standard" is not supported by the relevant scientific data; and that the current primary 8-hr

-------
standard of 0.08 ppm needs to be substantially reduced to be protective of human health,
particularly in sensitive subpopulations.

       Additionally, we note that the understanding of the associated science has progressed to
the point that there is no longer significant scientific uncertainty regarding the CASAC 's
conclusion that the current 8-hr primary NAAQS must be lowered. A large body of data clearly
demonstrates adverse human health effects at the current level of the 8-hr primary ozone
standard.  Retaining this standard would continue to put large numbers of individuals at risk for
respiratory effects and/or significant impact on quality of life including asthma exacerbations,
emergency room visits, hospital admissions and mortality.  (Scientific uncertainty does exist with
regard to the lower level of ozone exposure that would be fully-protective of human health. The
Ozone Panel concludes that it is possible that there is no threshold for an ozone-induced impact
on human health and that some adverse events may occur at policy-relevant background.)

       Moreover, EPA staff concluded that changes in the concentration-based form of the
standard (i.e., whether to use the third-, fourth-, or fifth-highest daily maximum 8-hr average
concentration) should also be considered.  The analysis found in the 2nd Draft Ozone Staff Paper
indicates that modest changes in the form of the standard can have substantial impacts on the
frequency of adverse health effects.  Therefore, the CASAC recommends that the Agency
conduct a broader evaluation of alternative concentration-based forms of the primary 8-hr ozone
standard and the implications of those alternative forms on public-health protection and stability
(i.e., with respect to yearly variability to ensure a stable target for control programs).

       The CASAC further recommends that the ozone NAAQS should reflect the capability of
current monitoring technology, which allows accurate measurement of ozone concentrations
with a precision of parts per billion, or equivalently to the third decimal place on the parts-per-
million scale. In addition,  given that setting a level of the ozone standard to only two decimal
places inherently reflects upward or downward "rounding," e.g., 0.07 ppm includes actual
measurements from 0.0651 ppm to 0.0749 ppm, the CASAC chooses to express its
recommended level, immediately below, to the third decimal place.

       Accordingly, the CASAC unanimously recommends that the current primary ozone
NAAQS be revised and that the level that should be considered for the revised standard be from
0.060 to 0.070 ppm, with a range of concentration-based forms from the third- to the fifth-
highest daily maximum 8-hr average concentration.  While data exist that adverse health effects
may occur at levels lower than 0.060 ppm, these data are less certain and achievable gains in
protecting human health can be accomplished through  lowering the ozone NAAQS to a level
between 0.060 and 0.070 ppm.
                                Secondary Ozone NAAQS

       An important difference between the effects of acute exposures to ozone on human health
and  the effects of ozone exposures on welfare is that vegetation effects are more dependent on
the cumulative exposure to, and uptake of, ozone over the course of the entire growing season
(defined to be a minimum of at least three months).  Therefore,  there is a clear need for a

-------
secondary standard which is distinctly different from the primary standard in averaging time,
level and form.  Developing a biologically-relevant ozone air quality index would be directly
responsive to the 2004 National Research Council (NRC) recommendations on Air Quality
Management in the United States (NAS, 1994) and will help support important new Agency
initiatives to enhance ecosystem-related program tracking and accountability.

       In its 1996 review of the ozone NAAQS, EPA staff proposed several  cumulative seasonal
ozone exposure indices, including SUM06, the concentration-weighted metric (i.e., the seasonal
sum of all hourly average concentrations > 0.06 ppm), and W126, the integrated exposure index
with a sigmoidal weighting function, as candidates for a secondary standard.  The Administrator
considered a three-month, 12-hr SUM06 secondary standard at a level of 25 ppm-hr as an
appropriate, biologically-relevant secondary standard, but ultimately rejected this option in favor
of simply setting the secondary standard equal to the primary. It was rationalized that efforts to
attain the new 8-hr primary standard would also eliminate most adverse effects on vegetation,
and at that time there were uncertainties in how cumulative seasonal exposures would change
with efforts to reduce peak 8-hour concentrations. Additionally, it was assumed that future
ozone/vegetation effects research over the coming years would clarify the very uncertain
quantitative relationships between ozone exposures and vegetation/ecological responses under
ambient field conditions.

       Unfortunately, however, the Agency has supported very little new vegetation/ecological
ozone effects research over the past decade.  The net result is that the quantitative evidence
linking specific ozone concentrations to specific vegetation/ecological effects must continue to
be characterized as having high uncertainties due to the lack of data for verification of those
relationships. It is not surprising that substantial research needs remain, as indicated both in
Chapter 8  and in individual reviewer comments.  The quantitative evidence linking specific
ozone concentrations to specific vegetation effects — especially at the complex ecosystem level
— must continue to be characterized as having high uncertainties due to the lack of data for
verification of those relationships. To a large extent, this is an unavoidable consequence of the
inherent complexities of ecosystem structure and function, interactions among biotic and abiotic
stressors and  stimuli, variability among species and genotype, detoxification  and compensatory
mechanisms,  etc. Nevertheless, the compelling weight of evidence provided in Chapter 7 of the
2nd Draft Ozone Staff Paper results from the convergence of results from many various and
disparate assessment methods including chamber and free air exposure, crop  yield and tree
seedling biomass experimental  studies, foliar injury data from biomonitoring plots, and  modeled
mature tree growth.

       Despite  limited recent research, it has become clear since the last review that adverse
effects on a wide range of vegetation including visible foliar injury are to be expected and have
been observed in areas that are  below the level of the current 8-hour primary and secondary
ozone standards. Such effects are observed in areas with seasonal 12-hr SUM06 levels below 25
ppm-hr (the lower end of the range of a SUM06 secondary standard suggested in the 1996
review and the upper end  of the range suggested in Chapter 8 of the 2nd Draft Ozone Staff
Paper).  Seasonal SUM06 (or equivalent W126) ranges well below 25 ppm-hr were
recommended for protecting various managed and unmanaged crops and tree seedlings in the
1997 workshop on secondary ozone standards (Heck and Cowling, 1997).  The absence of clear-

-------
cut lower effects thresholds for sensitive vegetation combined with the lower recent estimates of
policy-relevant background (typical range of 0.015 to 0.035 ppm) emphasizes the importance of
efforts to reduce low- to mid-range environmental exposures below 0.060 ppm.

       Based on the Ozone Panel's review of Chapters 7 and 8, the CASAC unanimously agrees
that it is not appropriate to try to protect vegetation from the substantial, known or anticipated,
direct and/or indirect, adverse effects of ambient ozone by continuing to promulgate identical
primary and secondary standards for ozone.  Moreover, the members of the Committee and a
substantial majority of the Ozone Panel agrees with EPA staff conclusions and encourages the
Administrator to establish an alternative cumulative secondary standard for ozone and related
photochemical oxidants that is distinctly different in averaging time, form and level from the
currently existing or potentially revised 8-hour primary standard.  The suggested approach to the
secondary standard is a cumulative seasonal growing standard such as the indices SUM06 or
W126 aggregated over at least the three summer months exhibiting the highest cumulative ozone
levels and includes the ozone exposures from at least 12 daylight hours. The CASAC suggests a
range of 10 to 20 ppm-hours for the three-month growing season SUM06 index for agricultural
crops rather than the 15-25 ppm-hours proposed in Chapter 8.

       However, the Ozone Panel views the three-month growing season W126 index as a
potentially more biologically-relevant index than the 3-month growing season SUM06 index.
This is because the W126 index has no absolute minimum ozone concentration threshold and
only lightly weights the lower ozone concentrations. Therefore, a three-month seasonal W126
that is the approximate equivalent of the SUM06 at 10 to 20 ppm-hr is preferred.  As shown by
the references cited at the end of Chapter 8, the consensus view among expert persons in the
ecological communities of both this country and elsewhere around the world is that a secondary
standard of cumulative form and extending over an entire growing season will be far more
effective than a secondary standard that is not cumulative inform and does not include the whole
growing season.

       In conclusion, the Clean Air Scientific Advisory Committee is pleased to provide its
scientific advice and recommendations to the Agency on the primary and secondary ozone
NAAQS.  We recognize that our recommendation of lowering of the current primary ozone
standard would likely result in a large portion of the U.S. being in non-attainment. Nevertheless,
we take very seriously the statutory mandate  in the Clean Air Act not only for the Administrator
to establish, but also for the CASAC to recommend to the Administrator, a primary standard that
provides for an "adequate margin of safety ... requisite to protect the public health. "

       Finally, as announced during the Ozone Panel's August meeting, once the Agency
releases the Final Ozone Staff Paper in early January 2007, the CASAC intends to hold a public
teleconference in late January or early February 2007 for the members of the Ozone Panel to
review — and, prospectively, to offer additional, unsolicited advice to the Agency concerning —
Chapter 6 (Staff Conclusions on Primary Os NAAQS) and Chapter 8 (Staff Conclusions on
Secondary Os NAAQS) in that final Agency document. The purpose of such advice would be to

-------
inform EPA's efforts as it develops the forthcoming, proposed rule for ozone and related
photochemical oxidants. As always, the CASAC wishes EPA well in this important endeavor.

                                            Sincerely,
                                                   /Signed/
                                            Dr. Rogene Henderson, Chair
                                            Clean Air Scientific Advisory Committee
Appendix A - Clean Air Scientific Advisory Committee Roster (FY 2006)
Appendix B - CASAC Ozone Review Panel Roster
Appendix C - Charge to the CASAC Ozone Review Panel
Appendix D - Review Comments from Individual CASAC Ozone Review Panel Members

-------
               CASAC Chapter-Specific Discussion Comments on
                        EPA's 2nd Draft Ozone Staff Paper
       Sub-groups of the CASAC Ozone Review Panel who led the discussion on individual
chapters of the Staff Paper summarized their comments in the following paragraphs:

       Chapter 2 (Air Quality Characterization): A better introduction to the central role of
photochemical oxidation reactions as the key reactions governing the behavior of air pollutants
in the atmosphere would improve this chapter.  Ozone is the key indicator of the extent of
oxidative chemistry and serves to integrate multiple pollutants.  Oxidation in the atmosphere
leads to the formation of particulate matter from SO2, NOx, and volatile organic compounds
(VOCs) as well as gas phase irritants (formaldehyde, acrolein, etc). Thus, although ozone itself
has direct effects on human health and ecosystems, it can also be considered as indicator of the
mixture of photochemical oxidants and of the oxidizing potency of the atmosphere. Section 2.2.6
only briefly covers the relationship of ozone to other photochemical oxidants.  It would be
beneficial to add a short paragraph outlining the role of ozone and other photochemical oxidants
in the atmospheric transformation processes that may results in the formation of more toxic
products (both in an outdoor and indoor environment), as provided in the individual comments
appended to this letter.

       The section on policy-relevant background (2.7) continues to have problems. Although
the section briefly cites the results of comparison of different models and measurements, it does
not adequately address the uncertainties of the global GEOS-CHEM model, and how these
uncertainties are reflected in the health risk analysis. Since ozone health effects are observed
down to concentrations of the order of 0.04-0.05 ppm, it is important to know how the PRB is
related to the considered primary ozone standard and what uncertainties there are in the risk
attributed to controllable sources.

       Chapter 3 (Policy-Relevant Assessment of Health Effects Evidence): The latest draft
of Chapter 3 is much  improved over the previous draft. Efforts  to respond to some of the earlier
concerns expressed by the CASAC are appreciated. While in general this chapter is well written,
and is a credible basis for the risk analyses that follow, there are inconsistencies and inaccuracies
that still need to be addressed.  Typically, there is appropriate use of cautionary phrases when the
data are not as strong as they might be, but this use is inconsistent across the chapter, and there
are instances where EPA staff appear to be stretching to infer that data support their statement.
While the individual comments of Ozone Panel members attached to this letter provide specifics
on these points, some of the Panel's more significant concerns are discussed briefly below.

       Discussion of measurement error is convoluted, confusing, and contains some mistakes.
The primary issue in the use of central ambient monitors for ozone in time-series
epidemiological studies is whether they provide any information at all that reflects daily personal
ozone exposure in susceptible  populations.  The discussion  on p. 3-37 of the impact of various
types of exposure measurement error is incorrect; the difference between true and measured
ambient concentrations is an example of classical measurement  error that results in bias of effect

-------
estimates to the null, not just an increase in standard error. Claiming that the difference between
average personal exposure and ambient concentrations results in "attenuation of risk" is not
appropriate.

       The Ozone Panel does not completely agree with staffs conclusion that "the use of
routinely monitored ambient ozone concentrations as a surrogate for personal exposures is not
generally expected to change the  principal conclusions from ozone epidemiological studies."
Indeed, Panel members have little insight as to what we would find if we had actual exposure
measurements.  Personal exposures most likely correlate better with central site values for those
subpopulations that  spend a good deal of time outdoors, which coincides, for example, with
children actively engaged in outdoor activities, and which happens to be a group that the ozone
risk assessment focuses upon.

       Some statements about which individuals are at greatest risk of ozone-induced effects are
not adequately supported by the information discussed in the chapter.  Individuals with chronic
obstructive pulmonary disease (COPD) and cardiovascular disease (CVD) are likely to be at
increased risk, but the hypothesis that such "hyper-responsiveness" can be used to identify
individuals with COPD or CVD who are at greatest risk of Os-induced health effects has not
been confirmed. A more appropriate conclusion would be that individuals with COPD and CVD
are at increased risk of O3-induced health effects.

       The discussion of the ranges for changes in FEVi that are considered to be small,
moderate, or large for persons with impaired respiratory systems is not consistent.  While EPA
staff state that the table values for the ranges do not need to be changed, staff indirectly
acknowledge that a  10% reduction in this variable in asthmatics  could have serious
consequences, an interpretation that is used in Chapters 4-6.

       The 30 subjects studied by Adams had a great influence on the analyses presented in
Chapters 5 and 6. While the discussion of the low-level exposures used in the controlled human
studies by Adams and colleagues is technically correct that no statistically significant changes
were found in FEVi for ozone at  40 to 60 ppb compared to filtered air, there were clearly  a few
individuals who experienced declines in lung function at these lower concentrations.  These were
healthy subjects, so  the percentage of asthmatic subjects, if they  had been studied, would most
likely be considerably greater.

       The lack of statistical power is consistently offered in Chapter 3 for why there appears to
be an inconsistent effect seen for  COPD mortality. Coherence of respiratory effects for ozone
suffers from neither no more nor  no less power considerations that do those for particulate matter
(PM).  Yet the Agency did not argue a lack of power when assessing PM risks, so consistency is
needed here relative to ozone effect estimates for COPD mortality.

       The relatively strong and relatively consistent effect of ozone on emergency department
visits for respiratory disease, especially asthma, as evidenced in Figure 3-4 is misrepresented in
several places in the Chapter (and in Chapters 5 and 6) as "inconclusive" or "inconsistent."  This
should be corrected.
                                           10

-------
       Chapter 4 (Characterization of Human Exposure to Ozone):  The second draft of
Chapter 4 has responded to many of the comments made on the first draft, and is thus clearer
than before. The panel was pleased to see the reanalysis for 2002 in addition to 2004.

       It would be helpful to have the estimated exposures for current (2002 and 2004) levels
displayed in Tables 4-8 & 4-9 (p. 4-32) and Figures 4-4 to 4-21 (pp. 4-33 to 4-41), in addition to
only those for just meeting the current standard and alternative more stringent standards.  This
would be analogous to the way estimated effects are displayed in Chapter 5  (Figures 5-5 to 5-9
[pp.5-58 to 5-65]).

       On the whole, Chapter 4 provides a clear "road map" for what was done to characterize
available knowledge about human exposure to ozone in the framework of generally accepted
modeling approaches of appropriately selected populations in 12 urban areas of the U.S. Much
of the text reads like a basic textbook on human exposure assessment using  state-of-the-art
modeling approaches, such as the Air Pollutants Exposure Model (APEX), including adjustments
for lung ventilation of delivered ozone dose. This extension, beyond exposure characterization,
is particularly important for ozone where the extent of measurable human responses is very
sensitive to the amount of ozone inhaled and to where it deposits along the respiratory tract.
Further extension of the methodology to estimate dose would have important implications and
should be discussed.

       There is an  explicit discussion of the limitations of the APEX model in terms of
variability and the quality of the input data, which is appropriate and fine as far as it goes.  There
are good reasons presented for selection of urban areas and the time periods to be modeled.
However, there was inadequate consideration of the populations selected for modeling. Those
selected were appropriate, but the omission of the elderly, the population most at risk for ozone-
associated premature daily mortality, was notable and not even mentioned in terms of why it was
not considered.

       The chapter was very good at exposition and clear presentation of modeling results, but
was deficient it its discussion of seemingly counterintuitive results, and of a potentially large
influence of measurement biases. As an example of the first of these issues, the children in LA
& Houston are estimated to have far fewer exposures above 0.07 ppm (8-hr) than in most other
cities with lower ozone concentrations and fewer children.  This was likely due to the greater
within-day and sampler-to-sampler variations in concentration within these two cities than in the
others, the fact that the entire year was modeled while for other sites the winter was not included
and/or the greater extent of air conditioning, especially in Houston. Whatever the reasons, there
should have been some discussion of the causes.  The quadratic rollback methodology should
have been better described since this strategy has important consequences for the modeled
results.

       The second issue that was presented, but left hanging without an adequate discussion is at
the bottom of page 4-47, where it was simply stated that "in general, APEX systematically
under-predicts the measured values by 0.001 to 0.02 ppm (zero to 50 percent)."  If this is so, is it
due to a really serious failure of the APEX model, or to unreliable measurements?  The
measurements at issue were six-day average concentrations based on the use of passive
                                           11

-------
(diffusion) samplers, which are known to be subject to significant errors when the air velocity
across the inlet is variable.  The comparison of measured and modeled concentrations depicted in
Figure 4-22 is certainly worthy of further analysis and discussion.

       Chapter 5 (Characterization of Health Risks):  Generally the panel found Chapter 5
and its accompanying risk assessment to be well done, balanced and reasonably communicated.
Additional text is needed at the beginning and end of the  chapter to put the limited risk
assessment into the context of the much larger body of evidence of ozone health effects.  The
discussion of uncertainty in these risk estimates is expanded in section 5.3.2.5.  Although a
number of issues are raised, their impacts on the estimates have not been thoroughly explored.
Additional sensitivity analyses seem warranted.  In particular, it is essential that the sensitivity of
the risk  assessment to the shape of the dose-response curve for FEVi be evaluated. Although the
3 parameter logistic (3PL) model emulates the pattern seen in the five "data points," these points
are aggregates of the original data, and may give a misleadingly optimistic picture of the quality
of the fit.  More importantly, although the problem of model uncertainty is noted it has not been
addressed even though methods exist for doing so. Even if only the linear and logistic models
were included in the analysis, the error bands around the  estimated response probabilities would
likely increase to better reflect that uncertainty. In addition, a suggestion to deal with the
uncertainties surrounding estimation of PRB, particularly as related to Table  5.5 (for lung
function) and Table 5.11 (mortality), would be to change the form of the analyses to assess the
impact of the concentration change in the expected number of health effects relative to the
current standard. The key advantage of estimating the effect of concentration change is that it
does not depend on the choice of the PRB.

      With regard to the controlled human exposure studies, Ozone Panel members believe that
the selection of changes in pulmonary function expressed as percent change in FEVi in children
is a fair  indicator of an adverse effect at 15% change in all active children; and, in asthmatic
children, a 10% change is indicative of adverse effects. However, the presentation of the figures
showing these effects needs to be revised to indicate the uncertainties in the results used,
particularly at the lower levels of exposure. The potential mechanisms whereby these changes
are a reflection of both pain on breathing, partial inflammation of smaller airways, other effects
on airways, and potentially triggers for more significant respiratory morbidity, particularly in
asthmatic children, are not adequately discussed. In addition, some added discussion is
necessary to indicate that these measures are generally taken in areas with relatively high
background levels of ozone exposure, and that the role that tolerance may play in minimizing the
degree of adverse effect observed needs to  be considered.

      From the perspective of the epidemiological data, the Ozone Panel judged the selection
of: respiratory symptoms in moderate/severe asthmatic children (ages zero [birth] to 12); hospital
admissions for respiratory illness among asthmatic children; and premature total non-accidental
and cardiorespiratory mortality for inclusion in the quantitative risk assessment to be appropriate.
However, the CASAC believes that several other endpoints should be discussed qualitatively to
support  the findings that these endpoints indicate that significant adverse effects are occurring at
exposure concentrations well below the current standard.  Other endpoints deemed worthy of
additional discussion included respiratory emergency department visits among asthmatics and
patients with other respiratory diseases, increased medication usage, and increased
                                            12

-------
symptomatology reported at exposure levels well below the current standard. Taken together,
members of the Ozone Panel felt strongly that these findings preclude including the current
standard as a scientifically defensible option for the Administrator (see discussion about Chapter
6 found in the main portion of the letter above).

       Another problem in the health effects calculations (see Table 5-5 and 5-11) is that they
are based on computations of the form Rx - RPRB, where Rx is the risk at a given concentration x
of O3 and RPRB is the corresponding risk at policy-relevant background (PRB) for O3.  As
discussed at the Ozone Panel's August meeting, the PRB is highly-problematic to calculate and
is, in some sense, "unknowable."  One can avoid this problem by calculating the A = R0.8 - Rx
for various concentrations x.  This form would allow focus on the question, "What is the
difference in the expected number of health effects that will occur at various concentrations of
O3, relative to the current standard of 0.08?" A key advantage of A is that it does not depend on
the choice of PRB, and thus is free of the uncertainties surrounding estimation of PRB.

       Chapter 6 (Staff Conclusions on Primary Os NAAQS): See the discussion on Chapter
6 found in the main portion of the letter above. It would  also be helpful to  have the estimated
exposures for current (2002 and 2004) levels displayed in figures 6-1 to 6-6 (pp. 6-34 to 6-39), in
addition to only those for just meeting the current standard and alternative more stringent
standards. This would be analogous to the way estimated effects are displayed in Chapter 5
(Figures 5-5 to 5-9 [pp.5-58 to 5-65]).

       Chapters 7 (Policy-Relevant Assessment of Welfare Effects Evidence) and 8 (Staff
Conclusions on Secondary Os NAAQS): Chapter 7 is a well-developed and persuasively
presented assessment of the welfare effects of ozone on vegetation, which forms the technical
basis for the range of secondary standards recommended in Chapter 8. That having been said,
the potential for significant propagation of error/uncertainty in the underlying technical
documentation draws into question the conclusions drawn by EPA Staff. As observed in the
Agency's  1989  and 1996 Ozone Staff Papers, ozone remains the most prevalent phytotoxic
compound in the ambient air "impairing crop production and injuring native vegetation and
ecosystems more than any other air pollutant" (USEPA 1989, 1996).  Furthermore, as has been
noted in the current assessment of human health effects, there also appears to be no safe
threshold concentration below which ozone effects on sensitive vegetation  are eliminated.  See
the additional discussion on Chapter 8 found in the main portion of the letter above.
                                           13

-------
   Appendix A - Clean Air Scientific Advisory Committee Roster (FY 2006)
                     U.S. Environmental Protection Agency
                   Science Advisory Board (SAB) Staff Office
              Clean Air Scientific Advisory Committee (CASAC)


CHAIR
Dr. Rogene Henderson, Scientist Emeritus, Lovelace Respiratory Research Institute,
Albuquerque, NM

MEMBERS
Dr. Ellis Cowling, University Distinguished Professor-at-Large, North Carolina State
University, Colleges of Natural Resources and Agriculture and Life Sciences, North Carolina
State University, Raleigh, NC

Dr. James D. Crapo, Professor, Department of Medicine, National Jewish Medical and
Research Center, Denver, CO

Dr. Frederick J. Miller, Consultant, Cary, NC

Mr. Richard L. Poirot, Environmental Analyst, Air Pollution Control Division, Department of
Environmental Conservation, Vermont Agency of Natural Resources, Waterbury, VT

Dr. Frank Speizer, Edward Kass Professor of Medicine, Channing Laboratory, Harvard
Medical School, Boston, MA

Dr. Barbara Zielinska, Research Professor, Division of Atmospheric Science, Desert Research
Institute, Reno, NV
SCIENCE ADVISORY BOARD STAFF
Mr. Fred Butterfield, CASAC Designated Federal Officer, 1200 Pennsylvania Avenue, N.W.,
Washington, DC, 20460, Phone: 202-343-9994, Fax: 202-233-0643 (butterfield.fred@epa.gov)
(Physical/Courier/FedEx Address: Fred A. Butterfield, III, EPA Science Advisory Board Staff
Office (Mail Code 1400F), Woodies Building, 1025 F Street, N.W., Room 3604, Washington,
DC 20004, Telephone: 202-343-9994)
                                       A-l

-------
               Appendix B - CASAC Ozone Review Panel Roster
                     U.S. Environmental Protection Agency
                   Science Advisory Board (SAB) Staff Office
              Clean Air Scientific Advisory Committee (CASAC)
                          CASAC Ozone Review Panel


CHAIR
Dr. Rogene Henderson*, Scientist Emeritus, Lovelace Respiratory Research Institute,
Albuquerque, NM

MEMBERS
Dr. John Balmes, Professor, Department of Medicine, University of California San Francisco,
University of California - San Francisco, San Francisco, California

Dr. Ellis Cowling*, University Distinguished Professor-at-Large, North Carolina State
University, Colleges of Natural Resources and Agriculture and Life Sciences, North Carolina
State University, Raleigh, NC

Dr. James D. Crapo*, Professor, Department of Medicine, National Jewish Medical and
Research Center, Denver, CO

Dr. William (Jim) Gauderman, Associate Professor, Preventive Medicine, Medicine,
University of Southern California, Los Angeles, CA

Dr. Henry Gong, Professor of Medicine and Preventive Medicine, Medicine and Preventive
Medicine, Keck School of Medicine, University of Southern California, Downey, CA

Dr. Paul J. Hanson, Senior Research and Development Scientist, Environmental Sciences
Division, Oak Ridge National Laboratory (ORNL), Oak Ridge, TN

Dr. Jack Harkema, Professor, Department of Pathobiology, College of Veterinary Medicine,
Michigan State University, East Lansing, MI

Dr. Philip Hopke, Bayard D. Clarkson Distinguished Professor, Department of Chemical
Engineering, Clarkson University, Potsdam, NY

Dr. Michael T. Kleinman, Professor, Department of Community & Environmental Medicine,
University of California - Irvine, Irvine, CA

Dr. Allan Legge, President, Biosphere Solutions, Calgary, Alberta, Canada
                                        B-l

-------
Dr. Morton Lippmann, Professor, Nelson Institute of Environmental Medicine, New York
University School of Medicine, Tuxedo, NY

Dr. Frederick J. Miller*, Consultant, Cary, NC

Dr. Maria Morandi, Assistant Professor of Environmental Science & Occupational Health,
Department of Environmental Sciences, School of Public Health, University of Texas - Houston
Health Science Center, Houston, TX

Dr. Charles Plopper, Professor, Department of Anatomy, Physiology and Cell Biology, School
of Veterinary Medicine, University of California - Davis, Davis, California

Mr. Richard L. Poirot*, Environmental Analyst, Air Pollution Control Division, Department of
Environmental Conservation, Vermont Agency of Natural Resources, Waterbury, VT

Dr. Armistead (Ted) Russell, Georgia Power Distinguished Professor of Environmental
Engineering, Environmental Engineering Group,  School of Civil and Environmental
Engineering, Georgia Institute of Technology, Atlanta, GA

Dr. Elizabeth A. (Lianne) Sheppard, Research Professor, Biostatistics and Environmental &
Occupational Health Sciences, Public Health and Community Medicine, University of
Washington, Seattle, WA

Dr. Frank Speizer*, Edward Kass Professor of Medicine, Channing Laboratory, Harvard
Medical School, Boston, MA

Dr. James Ultman, Professor, Chemical Engineering, Bioengineering Program, Pennsylvania
State University, University Park, PA

Dr. Sverre Vedal, Professor of Medicine, Department of Environmental and Occupational
Health Sciences, School of Public Health and Community Medicine, University of Washington,
Seattle, WA

Dr. James (Jim) Zidek, Professor, Statistics,  Science, University of British Columbia,
Vancouver, BC, Canada

Dr. Barbara Zielinska*, Research Professor, Division of Atmospheric Science, Desert Research
Institute, Reno, NV

SCIENCE ADVISORY BOARD STAFF
Mr. Fred Butterfield, CASAC Designated Federal Officer, 1200 Pennsylvania Avenue, N.W.,
Washington, DC, 20460, Phone: 202-343-9994, Fax: 202-233-0643 (butterfield.fred@epa.gov)
* Members of the statutory Clean Air Scientific Advisory Committee (CASAC) appointed by the EPA
  Administrator (FY 2006)
                                         B-2

-------
United States                              Office of Air Quality Planning and Standards              Publication No. EPA 452/R-07-003
Environmental Protection                   Air Quality Strategies and Standards Division                                  January 2007
Agency                                           Research Triangle Park, NC

-------